Author Response

Reviewer #1 (Public Review):

This paper evaluates the effect of knocking out CST7(Cystatin 5) on the APPNL-G-F Alzheimer's disease mouse model. They found sexually dimorphic outcomes, with differential transcriptional responses, increased phagocytosis (but interestingly a higher plaque burden) in females and suppressed inflammatory microglial activation in males (but interestingly no change in plaque burden). This study offers new insight into the functional role of CST7 that is upregulated in a subset of disease- associated microglia in AD models and human brain. Despite the discovery of disease-associated microglia several years ago, there has been little effort in understanding the function of the different genes that make up this profile, making this paper especially timely. Overall, the experiments are well-controlled and the data support the main conclusions and the manuscript could be strengthened by addressing the below comments and clarifying questions that could impact the interpretation of their data/ findings.

1) In the first section discussing CST7 expression levels in AD models, it would be good to involve a discussion of levels of CST7 change in human AD samples. There are sufficient available datasets to look at this, and it would help us understand how comparable the animal models are to human patients. For example, while in mice CST7 is highly enriched in microglia/macrophages, in human datasets it seems like it is not quite so specific to microglia - it is equally expressed in endothelial cells. This might have a significant impact on the interpretation of the data, and it would be good to introduce and assess the findings in mice through the human subjects lens. There is a discussion of the human data in the discussion section, but it would be more appropriately assessed in the same way as the mouse data and comparatively presented in the results section. The authors could also include the data from Gerrits et al. 2021 in their first figure.

We agree with the reviewer on the importance of considering the work in the context of human disease. While CST7 is not as strongly upregulated in human AD brain as it is in mouse expression is observed predominantly in myeloid cells in the brain with very minimal expression detected in endothelial cells (see screenshots in Author response image 1 from Brain Myeloid Landscape platform (http://research-pub.gene.com/BrainMyeloidLandscape/BrainMyeloidLandscape2/) and is enriched in AD clusters vs homeostatic in scRNASeq studies (Gerrits et al., 2021). We attempted immunostaining for human CF (CST7) in AD brains to assess expression and co-localisation with microglial markers but failed to validate any of the antibodies tested. Additionally, King et al., 2023 (PMID: 36547260) recently showed increase in CST7 expression in bulk hippocampal RNASeq in AD vs mid-life controls suggesting an ageing/AD mechanism. CST7 has also been shown to be expressed following overexpression of TREM2 in human microglia in vitro and that siRNA-mediated knockdown of expression leads to an increase in phagocytosis (Popescu et al., 2023 - PMID: 36480007), mirroring our data and suggesting a conserved role in human cells. Overall, we believe that, even in the context of mouse models, the understanding of the function of genes upregulated in disease is of importance to the field and that this study paves the way for further work investigating human CST7 in disease. We have added this (with citations to the datasets mentioned) to the discussion (highlighted).

Author response image 1

2) The differential RNAseq data is perhaps one of the most striking results of this paper; however it is difficult to see exactly how similar the male v female APPNL-G-F profiles are, in addition to the genes shared or not between the KO condition. Venn diagrams, in addition to statistical tests, would enhance this part of the paper and add more clarity.

We have added Venn diagrams to show DEGs between male and female AppNL-G-F microglia vs WT control to show how similar the male v female APPNL-G-F profiles are. Additionally, to exemplify the Cst7KO-Sex interaction, a Venn showing DEGs between male and female AppNL-G-F microglia vs. AppNL-G-FCst7-/- microglia (Fig. 2 – Fig. supplement 3). We confirm we have derived all differential gene expression changes reported (including those represented in the Venn diagrams) using appropriate Padj statistical approaches (see Methods).

3) A major argument in the paper is a continuation of Sala-Frigerio 2019 which says that the female phenotype is an acceleration of the male phenotype. Does this mean that if males were assessed at later timepoints, they would be more similar to the females? Or are there intrinsic differences that never resolve? It would be helpful to see a later timepoint for males to get at the difference between these two options

This is an interesting question and while we acknowledge that empirically addressing with a later timepoint could add insight, we believe it would actually need multiple closely-spaced timepoints as choosing what single later timepoint would be optimal is difficult to judge (and likely not possible at all) for reasons below. We also believe data already published combined with our observations show it is most-likely a cell-intrinsic effect that explains our sex-specific differences.

First, we emphasize the acceleration of the microglial phenotype in female AppNL-G-F mice previously published is fairly subtle and relative rather than absolute e.g. the DAM/ARM microglia state represents ~50% of all microglia in male and ~55% of all microglia in females at 12 months old therefore both sexes have similarly abundant microglia in the state that most highly express Cst7. Indeed, after the age at which DAM/ARM state microglia appear in appreciable numbers (~ 6 months), both females and males both have an abundance of them. It is important to note that a 12-month male is far more “progressed” than a 6-month female hence the stepped age effect is temporally short.

Second, Cst7 deletion in the AppNL-G-F mice condition caused qualitative differences affecting distinct genes and/or overlapping genes moving in different directions between female and male mice - if a stepped age effect explained sex differences from Cst7 deletion, given that it could only be stepped by a very short timeframe (several weeks maximum) from reasoning above, we would expect to see similar qualitative changes but of different magnitude in female and male mice arising from Cst7 deletion; this is not the pattern we see.

Third, beyond 12 months old, regression from ARM/DAM actually occurs, again making it unlikely males would “catch up” with females to show the same profile from Cst7 deletion but just at an older age – practically, this also complicates choosing a single later timepoint (and age-related systemic morbidity emerges as a potential confounder as well).

In summary, while the acceleration of the DAM signature in female microglia offers an intriguing possible explanation to our observation of sexual dimorphism in response to deletion of one of the key genes in this signature, we believe it more likely that intrinsic effects are responsible for the Cst7 deletion sex-related impact. Taking the alternative perspective, even if a stepped age effect in the underlying progression of the model could explain our findings, this would need multiple timepoints with short gaps between (e.g. monthly at 12, 13, 14, 15 months old) to provide the temporal resolution to expose this pattern; we would not have the resources to conduct such a resource-intensive and lengthy study. We hope this reasoning appears logical and conscious of the importance to convey this in our manuscript we have revised the Discussion to as concisely as possible capture some key points outlined above.

4) If the central argument is that CST7 in females decreases phagocytosis and in males increases microglia activation, are there changes in amyloid plaque burden or structure in the APPNL-G-F /CST 7 KO mice compared to APPNL-G-F/CST7 WT that reflect these changes? Please address. If not, how does this affect the functional interpretation of differential expression observed in phagocytic/reactive microglia genes? Pieces of this are discussed but it could be clearer.

We emphasise the data already presented in Fig 6 and Fig. 6 – Fig. Supplement 2 showing altered Aβ burden (6E10 staining) and plaque count (MeX04) but no change in plaque area. Regarding the functional interpretation of Cst7-dependent gene changes in microglia beyond the endolysosomal function we present in figures 3-5, we have included additional data using simple immunohistochemistry, as suggested by the reviewer, to assess synapse abundance. We show loss of Sy38 coverage around plaques (Fig. 6I) and a moderate but significant decrease in coverage between AppNL-G-F/Cst7-/- vs AppNL-G-F brains only in females (Fig. 6J). This reflects the effect observed with plaque coverage whereby we observe increased burden in AppNL-G-F/Cst7-/- vs AppNL-G-F females but not males (Fig. 6B-F) suggesting the increased plaque burden in Cst7-/- female mice may lead to increased synapse loss. We would also emphasise that altered expression of phagolysosomal genes could affect disease in ways beyond interactions with amyloid and synapses.

5) It is confusing that increased phagocytosis in the APPNL-G-F/CST7 KO females leads to greater plaque burden, considering proteolysis is not affected. What might explain this observation? Additionally, it is interesting that suppression of microglial activation doesn't lead to an increase in plaques in the male APPNL-G-F/CST7 KO mice. How does the profile of phagocytic microglia in the male APPNL-G-F/CST7 KO mice differ from the APPNL-G-F males?

We emphasize our comments on this topic in the discussion where we speculate that the greater plaque burden in females is linked to increased uptake of Aβ (which we observe in Fig. 4B&C) and deposition into plaques as suggested by Huang et al., 2021 (PMID: 33859405), d’Errico et al., 2022 (PMID: 34811521) and Shabestari et al., 2022 (PMID: 35705056). Regarding the lack of effect in males despite the suppression of inflammatory genes, we agree this is a curious observation, although may point to as yet ill-defined mechanisms for how inflammatory pathways influence plaque pathology. Unfortunately, we were not able to specifically compare the profile of phagocytic microglia in AppNL-G-F vs AppNL-G-FCst7-/- as we did not perform single-cell RNASeq. However, our bulk RNASeq profiling suggests modest downregulation of phagocytic/endolysosomal genes (eg Lilrb4a, Fig. 2I) and reduced expression of LAMP2 in microglia by immunostaining. We have added further comment on this in the discussion.

6) Seems that the authors have potentially discovered an unusual mechanism for how CST7 could regulate cell autonomous function without impacting its canonical protease target. The authors deal with this extensively in the discussion but an ELISA or ICC to localize CST7 to microglia in vitro or in vitro would help address this point.

We have added FISH data localising Cst7 expression to IBA1+ cells specifically around plaques in App brains (Fig. 1B-E). We agree that assessing the subcellular localisation and any non-microglial expression of Cystatin-F (the protein coded by Cst7) would offer valuable insight into the protease target and may reveal details on the precise mechanism by which CF deletion leads the phenotype we observe in this study. However, despite attempting numerous commercially available and gifted antibodies to detect CF we were unable to validate (using Cst7-/- as controls) any methods other than FISH.

7) The authors focus on plaques in their final figure, however dysregulated microglial phagocytosis could impact many other aspects of brain health. Simple immunohistochemistry for synapses and myelin/oligodendrocytes (especially given the results of the in vitro phagocytosis assay) could provide more insight here.

We fully agree with the reviewer. As also outlined in our responses elsewhere, phagocytic changes could have multiple consequences, and we have included additional data using immunohistochemistry as advised for synapses in WT, AppNL-G-F, and AppNL-G-F/Cst7-/- brains. We show loss of Sy38 coverage around plaques (Fig. 6I) and a moderate but significant decrease in coverage between AppNL-G-F/Cst7-/- vs AppNL-G-F brains only in females (Fig. 6J). This reflects the effect observed with plaque coverage whereby we observe increased burden in AppNL-G-F/Cst7-/- vs AppNL-G-F females but not males (Fig. 6B-F) suggesting the increased plaque burden in Cst7-/- female mice may lead to increased synapse loss.

We also performed immunohistochemistry for myelin makers MAG and MBP but found no plaque-associated pathology. Finally, we searched for dystrophic neurites using LAMP1 but found that the antibody stained microglial lysosomes rather than dystrophic neurites in this model (see Author response image 2), an observation that has been made by others (Sharoar et al., 2021 - PMID: 34215298).

Overall, our data suggest Cst7 may play a protective role in females, limiting phagocytosis, reducing plaque burden and blunting synapse loss.

Author response image 2.

Reviewer #3 (Public Review):

In this manuscript, Daniels et al explored the role of Cystatin F in an A-driven mouse model of Alzheimer's disease. By crossing a constitutive knockout mouse lacking the gene that encodes Cystatin F, Cst7, to the AppNL-G-F mouse line, the authors describe impairments in microglial gene expression and phagocytic function that emerge more prominently in females versus males lacking Cst7. A strength of the study is its focus: given mounting evidence that microglia are a hub of neurological dysfunction with particular potential to trigger or exacerbate neurodegenerative disorders, it is essential to determine the changes in microglia that occur pathologically to promote disease progression. Similarly, the wide-spread identification of the gene in question, Cst7, as upregulated in AD models makes this gene a good target for mechanistic studies.

The paper in its current form also has several weaknesses which limit the insights derived, weaknesses that are largely related to the experimental tools and approaches chosen by the authors to test their hypotheses. For example, the paper begins with a figure replotting data from previous studies showing that Cst7 is upregulated in mouse models of Alzheimer's disease. Though relevant to the current study, there are no new insights provided here. Next, the authors perform bulk RNA-sequencing on microglia isolated from male and female mice in the Cst7-/-; AppNL-G-F mouse line. In the methods, it is unclear whether the authors took precautions to preserve the endogenous transcriptional state of these cells given evidence that microglia can acquire a DAM-like signature simply due to the process of dissociation (Marsh et al, Nature Neuroscience, 2022). If the authors did not control for this, their results may not support the conclusions they draw from the data. Relatedly, it appears the authors pooled all microglia together here, instead of just isolating DAMs specifically or analyzing microglia at single-cell resolution, which could reveal the heterogeneous nature of the role of Cst7 in microglia. In addition to losing information about heterogeneity, another concern is that they could be diluting out the major effects of the model on microglial function by including all microglia. Overall, the biggest issue I have with the RNA-sequencing data is the lack of validation of the gene expression changes identified using a different method that does not require dissociation, like immunohistochemistry or fluorescence in situ hybridization. Especially given the limited number of genes they found to be mis-regulated (see Fig. 2 E and G), I worry that these changes might simply be noise, especially since the authors provide no further evidence of their mis-regulation. Without further validation, the data presented are not sufficient to support the authors' claims.

We believe we have addressed this comment in the “Essential Revisions (for the authors)” section above. Please see again below:

We took standard precautions to minimise the risk of aberrant ex vivo cell activation, including maintaining cells on ice during non-enzyme steps of the procedure and carrying out preps in small batches to minimise time taken from removal of brain to purification of microglial RNA. Importantly, we also validated key expression data by in situ methods such as RNA FISH for Cst7 and Lilrb4a (Fig. 1B-E, Fig 2. - Fig. supplement 3) thus eliminating dissection-induced effects. Additionally, when performing qPCR on microglia from non-disease mice to test the disease-specific role of Cst7-dependent gene regulation we did not observe the same gene changes (Fig 2. - Fig. supplement 4) which, if such changes were dependent on tissue dissociation, we would expect to observe in WT or disease animals. We utilised the resources provided by Marsh et al. 2022 to search for overlap between enzyme-induced genes and our DEG lists from our key comparisons. We found the enzyme-induced gene set had very minimal overlap with any of our comparisons with overlap of only 4 genes between enzyme-induced genes and Cst7-dependent genes in males and no overlap between enzyme-induced genes and Cst7-dependent genes in females. We would further point out that the disease-induced microglial RNAseq profile in the AppNL-G-F Cst7+/+ (i.e. disease WT) condition mirrors those observed previously by multiple methods including in situ profiling (Zeng et al 2023 - PMID: 36732642) and RiboTag approaches (Kang et al 2018 - PMID: 30082275). We believe these combined approaches provide convincing validation of the RNAseq data.

In assessing the changes in microglial function and A pathology that occur in males and females of the Cst7-/-; AppNL-G-F line, the authors identify some differences between how females and males are affected by the loss of Cst7. While the statistical analyses the authors perform as given in the figure legends appear to be correct, the plots do not show significant changes between males and females for a given parameter. Take for example Figure 3H. Loss of Cst7 decreases IBA+Lamp+ microglia in males but increases this parameter in females. However, it does not appear that there is a significant difference in IBA+Lamp+ microglia in male versus female mice lacking Cst7. If there is no absolute difference between males and females, can the differential effects of Cst7 knockout on the sexes really be so relevant to the sexual dimorphism observed in the disease? I question this connection, but perhaps a greater discussion of what the result might mean by the authors would be helpful for placing this into context.

We understand the reviewer’s perspective and we agree that the interpretations could be presented and explained better in the text - we have updated the discussion as suggested to address this.

We designed our study initially to search for sex-specific effects of Cst7. Therefore, whilst our ANOVA does include main effects analysis for disease or sex, we carried out post-hoc analysis primarily to investigate effects of Cst7 deletion within sex. In the case of Fig. 3H pointed out by the reviewer, we observe a main effect for disease in the ANOVA and for disease-sex interaction but not for sex. Post-hoc analysis revealed the sex-specific effects of Cst7 we describe in the manuscript. This approach on analysis was also taken by Hoghooghi et al. (2020 - PMID: 33027652) who show related pathway gene Cstc is detrimental in EAE in females but not males (included in the discussion in this manuscript). The observation in Fig. 3H that there appears to be a Cst7 effect in males and females but not a sex effect in Cst7-/- is accurate but a relative anomaly in this study. Generally, we find that, alongside Cst7 deletion affecting females differently to males, we also see a sex effect in Cst7-/- animals but not in Cst7+/+ animals i.e. absolute levels in disease condition as well as relative changes from control to disease condition are different between males and females. This is exemplified in Fig. 4B&C where we observe increased microglial Aβ in female Cst7-/- animals vs male Cst7-/- animals and in Fig. 6D where we observe increased Aβ plaque burden in female Cst7-/- animals vs male Cst7-/- animals. This is most strikingly demonstrated in the case of our RNASeq data where we observe a difference in sex-dependent genes in AppNL-G-F vs AppNL-G-F/Cst7-/- (Fig. 2 – Fig. supplement 3B) implying removal of the Cst7 gene led to an ‘unlocking’ of sexual dimorphism in our cohort which we comment on in the discussion.

Finally, the use of in vitro assays of microglial function can be helpful as secondary analyses when coupled with in vivo or ex vivo approaches, but are not on their own sufficient to support the authors' conclusions. Quantitative engulfment assays (see Schafer et al, Neuron, 2012) on brain tissue showing that male and female microglia lacking Cst7 engulf different amounts of material (e.g. plaques, synapses, myelin) in the intact brain would be more convincing.

We agree that in vitro assays for microglial function are not always sufficient as standalone methods to support conclusions on functions in disease. The reviewer may have missed our in vivo MeX04 uptake assays (Fig 4A-D) which use measurements by flow cytometry on isolated microglia, this is a reflection of the microglial uptake in vivo following MeX04 injection pre-mortem – this experiment showed increased microglial Aβ in female Cst7-/- animals vs male Cst7-/- animals (Fig. 4B&C). Our in vitro assays complement and extend insight in ways not possible in vivo, for example they offer key insight into uptake/degradation kinetics that would be extremely challenging to carry out in vivo.

In general, a major limitation to the insights that can be derived in the study is the decision of the authors to perform all experiments at a single late-stage time point of 12 months of age. As this is quite far into disease progression for many AD models, phenotypic changes identified by the authors could arise due to the downstream effects of plaque deposition and therefore may not implicate Cst7 as a mechanism driving neurodegeneration rather than one of many inflammatory changes that accompany AD mouse models nearing the one-year time point. A related problem is that the study uses a constitutive KO mouse that has lacked Cst7 expression throughout life, not just during disease processes that increase with aging. In summary, the topic of the article is important and timely, but the connection between the data and the authors' conclusions is not as strong as it could be.

As described above, Cst7 expression is absent at steady-state and low until 6-12 months. Therefore, we predict that deletion would have little effect until 12+ months whereby cells expressing Cst7 have had the temporal window to affect disease pathology, as we find in the current study. This was a key part of the reasoning in our choice of the 12-month age for analyses. The negligible expression of Cst7 at baseline/early stages of disease suggests constitutive KO of the gene will not impact the phenotype until disease onset. This is substantiated by the lack of any genotype-related differences in the WT vs Cst7-/- comparisons in the non-disease condition.

Author Response

Reviewer #1 (Public Review):

This paper presents an interesting data set from historic Western Eurasia and North Africa. Overall, I commend the authors for presenting a comprehensive paper that focuses the data analysis of a large project on the major points, and that is easy to follow and well-written. Thus, I have no major comments on how the data was generated, or is presented. Paradoxically, historical periods are undersampled for ancient DNA, and so I think this data will be useful. The presentation is clever in that it focuses on a few interesting cases that highlight the breadth of the data.

The analysis is likewise innovative, with a focus on detecting "outliers" that are atypical for the genetic context where they were found. This is mainly achieved by using PCA and qpAdm, established tools, in a novel way. Here I do have some concerns about technical aspects, where I think some additional work could greatly strengthen the major claims made, and lay out if and how the analysis framework presented here could be applied in other work.

clustering analysis

I have trouble following what exactly is going on here (particularly since the cited Fernandes et al. paper is also very ambiguous about what exactly is done, and doesn't provide a validation of this method). My understanding is the following: the goal is to test whether a pair of individuals (lets call them I1 and I2) are indistinguishable from each other, when we compare them to a set of reference populations. Formally, this is done by testing whether all statistics of the form F4(Ref_i, Ref_j; I1, I2) = 0, i.e. the difference between I1 and I2 is orthogonal to the space of reference populations, or that you test whether I1 and I2 project to the same point in the space of reference populations (which should be a subset of the PCA-space). Is this true? If so, I think it could be very helpful if you added a technical description of what precisely is done, and some validation on how well this framework works.

We agree that the previous description of our workflow was lacking, and have substantially improved the description of the entire pipeline (Methods, section “Modeling ancestry and identifying outliers using qpAdm”), making it clearer and more descriptive. To further improve clarity, we have also unified our use of methodology and replaced all mentions of “qpWave” with “qpAdm”. In the reworked Methods section mentioned above, we added a discussion on how these tests are equivalent in certain settings, and describe which test we are exactly doing for our pairwise individual comparisons, as well as for all other qpAdm tests downstream of cluster discovery. In addition, we now include an additional appendix document (Appendix 4) which, for each region, shows the results from our individual-based qpAdm analysis and clustering in the form of heatmaps, in addition to showing the clusters projected into PC space.

An independent concern is the transformation from p-values to distances. I am in particular worried about i) biases due to potentially different numbers of SNPs in different samples and ii) whether the resulting matrix is actually a sensible distance matrix (e.g. additive and satisfies the triangle inequality). To me, a summary that doesn't depend on data quality, like the F2-distance in the reference space (i.e. the sum of all F4-statistics, or an orthogonalized version thereof) would be easier to interpret. At the very least, it would be nice to show some intermediate results of this clustering step on at least a subset of the data, so that the reader can verify that the qpWave-statistics and their resulting p-values make sense.

We agree that calling the matrix generated from p-values a “distance matrix” is a misnomer, as it does not satisfy the triangle inequality, for example. We still believe that our clustering generates sensible results, as UPGMA simply allows us to project a positive, symmetric matrix to a tree, which we can then use, given some cut-off, to define clusters. To make this distinction clear, we now refer to the resulting matrix as a “dissimilarity matrix” instead. As mentioned above, we now also include a supplementary figure for each region visualizing the clustering results.

Regarding the concerns about p-values conflating both signal and power, we employ a stringent minimum SNP coverage filter for these analyses to avoid extremely-low coverage samples being separated out (min. SNPs covered: 100,000). In addition, we now show that cluster size and downstream outlier status do not depend on SNP coverage (Figure 2 - Suppl. 3).

The methodological concerns lead me to some questions about the data analysis. For example, in Fig2, Supp 2, very commonly outliers lie right on top of a projected cluster. To my understanding, apart from using a different reference set, the approach using qpWave is equivalent to using a PCA-based clustering and so I would expect very high concordance between the approaches. One possibility could be that the differences are only visible on higher PCs, but since that data is not displayed, the reader is left wondering. I think it would be very helpful to present a more detailed analysis for some of these "surprising" clustering where the PCA disagrees with the clustering so that suspicions that e.g. low-coverage samples might be separated out more often could be laid to rest.

To reduce the risk of artifactual clusters resulting from our pipeline, we devised a set of QC metrics (described in detail below) on the individuals and clusters we identified as outliers. Driven by these metrics, we implemented some changes to our outlier detection pipeline that we now describe in substantially more detail in the Methods (see comment above). Since the pipeline involves running many thousands of qpAdm analyses, it is difficult to manually check every step for all samples – instead, we focused our QC efforts on the outliers identified at the end of the pipeline. To assess outlier quality we used the following metrics, in addition to manual inspection:

First, for an individual identified as an outlier at the end of the pipeline, we check its fraction of non-rejected hypotheses across all comparisons within a region. The rationale here is that by definition, an outlier shouldn’t cluster with many other samples within its region, so a majority of hypotheses should be rejected (corresponding to gray and yellow regions in the heatmaps, Appendix 4). Through our improvements to the pipeline, the fraction of non-rejected hypotheses was reduced from an average of 5.3% (median 1.1%) to an average of 3.8% (median 0.6%), while going from 107 to 111 outliers across all regions.

Second, we wanted to make sure that outlier status was not affected by the inclusion of pre-historic individuals in our clustering step within regions. To represent majority ancestries that might have been present in a region in the past, we included Bronze and Copper Age individuals in the clustering analysis. We found that including these individuals in the pairwise analysis and clustering improved the clusters overall. However, to ensure that their inclusion did not bias the downstream identification of outliers, we also recalculated the clustering without these individuals. We inspected whether an individual identified as an outlier would be part of a majority cluster in the absence of Bronze and Copper Age individuals, which was not the case (see also the updated Methods section for more details on how we handle time periods within regions).

In response to the “surprising” outliers based on the PCA visualizations in Figure 2, Supplement 2: with our updated outlier pipeline, some of these have disappeared, for example in Western and Northern Europe. However, in some regions the phenomenon remains. We are confident this isn’t a coverage effect, as we’ve compared the coverage between outliers and non-outliers across all clusters (see previous comment, Figure 2 - Suppl. 3), as well as specifically for “surprising” outliers compared to contemporary non-outliers – none of which showed any differences in the coverage distributions of “surprising” outliers (Author response images 1 and 2). In addition, we believe that the quality metrics we outline above were helpful in minimizing artifactual associations of samples with clusters, which could influence their downstream outlier status. As such, we think it is likely that the qpAdm analysis does detect a real difference between these sets of samples, even though they project close to each other in PCA space. This could be the result of an actual biological difference hidden from PCA by the differences in reference space (see also the reply to the following comment). Still, we cannot fully rule out the possibility of latent technical biases that we were not able to account for, so we do not claim the outlier pipeline is fully devoid of false positives. Nevertheless, we believe our pipeline is helpful in uncovering true, recent, long-range dispersers in a high-throughput and automated manner, which is necessary to glean this type of insight from hundreds of samples across a dozen different regions.

Author response image 1.

SNP coverage comparison between outliers and non-outliers in region-period pairings with “surprising” outliers (t-test p-value: 0.242).

Author response image 2.

PCA projection (left) and SNP coverage comparison (right) for “surprising” outliers and surrounding non-outliers in Italy_IRLA.

One way the presentation could be improved would be to be more consistent in what a suitable reference data set is. The PCAs (Fig2, S1 and S2, and Fig6) argue that it makes most sense to present ancient data relative to present-day genetic variation, but the qpWave and qpAdm analysis compare the historic data to that of older populations. Granted, this is a common issue with ancient DNA papers, but the advantage of using a consistent reference data set is that the analyses become directly comparable, and the reader wouldn't have to wonder whether any discrepancies in the two ways of presenting the data are just due to the reference set.

While it is true that some of the discrepancies are difficult to interpret, we believe that both views of the data are valuable and provide complementary insights. We considered three aspects in our decision to use both reference spaces: (1) conventions in the field (including making the results accessible to others), (2) interpretability, and (3) technical rigor.

Projecting historical genomes into the present-day PCA space allows for a convenient visualization that is common in the field of ancient DNA and exhibits an established connection to geographic space that is easy to interpret. This is true especially for more recent ancient and historical genomes, as spatial population structure approaches that of present day. However, there are two challenges: (1) a two-dimensional representation of a fairly high-dimensional ancestry space necessarily incurs some amount of information loss and (2) we know that some axes of genetic variation are not well-represented by the present-day PCA space. This is evident, for example, by projecting our qpAdm reference populations into the present-day PCA, where some ancestries which we know to be quite differentiated project closely together (Author response image 3). Despite this limitation, we continue to use the PCA representation as it is well resolved for visualization and maximizes geographical correspondence across Eurasia.

On the other hand, the qpAdm reference space (used in clustering and outlier detection) has higher resolution to distinguish ancestries by more comprehensively capturing the fairly high-dimensional space of different ancestries. This includes many ancestries that are not well resolved in the present-day PCA space, yet are relevant to our sample set, for example distinguishing Iranian Neolithic ancestry against ancestries from further into central and east Asia, as well as distinguishing between North African and Middle Eastern ancestries (Author response image 3).

To investigate the differences between these two reference spaces, we chose pairwise outgroup-f3 statistics (to Mbuti) as a pairwise similarity metric representing the reference space of f-statistics and qpAdm in a way that’s minimally affected by population-specific drift. We related this similarity measure to the euclidean distance on the first two PCs between the same set of populations (Author response image 4). This analysis shows that while there is almost a linear correspondence between these pairwise measures for some populations, others comparisons fall off the diagonal in a manner consistent with PCA projection (Author response image 3), where samples are close together in PCA but not very similar according to outgroup-f3. Taken together, these analyses highlight the non-equivalence of the two reference spaces.

In addition, we chose to base our analysis pipeline on the f-statistics framework to (1) afford us a more principled framework to disentangle ancestries among samples and clusters within and across regions (using 1-component vs. 2-component models of admixture), while (2) keeping a consistent, representative reference set for all analyses that were part of the primary pipeline. Meanwhile, we still use the present-day PCA space for interpretable visualization.

Author response image 3.

Projection of qpAdm reference population individuals into present-day PCA.

Author response image 4.

Comparison of pairwise PCA projection distance to outgroup-f3 similarity across all qpAdm reference population individuals. PCA projection distance was calculated as the euclidean distance on the first two principal components. Outgroup-f3 statistics were calculated relative to Mbuti, which is itself also a qpAdm reference population. Both panels show the same data, but each point is colored by either of the two reference populations involved in the pairwise comparison.

PCA over time

It is a very interesting observation that the Fst-vs distance curve does not appear to change after the bronze age. However, I wonder if the comparison of the PCA to the projection could be solidified. In particular, it is not obvious to me how to compare Fig 6 B and C, since the data in C is projected onto that in Fig B, and so we are viewing the historic samples in the context of the present-day ones. Thus, to me, this suggests that ancient samples are most closely related to the folks that contribute to present-day people that roughly live in the same geographic location, at least for the middle east, north Africa and the Baltics, the three regions where the projections are well resolved. Ideally, it would be nice to have independent PCAs (something F-stats based, or using probabilistic PCA or some other framework that allows for missingness). Alternatively, it could be helpful to quantify the similarity and projection error.

The fact that historical period individuals are “most closely related to the folks that contribute to present-day people that roughly live in the same geographic location” is exactly the point we were hoping to make with Figures 6 B and C. We do realize, however, that the fact that one set of samples is projected into the PC space established by the other may suggest that this is an obvious result. To make it more clear that it is not, we added an additional panel to Figure 6, which shows pre-historical samples projected into the present-day PC space. This figure shows that pre-historical individuals project all across the PCA space and often outside of present-day diversity, with degraded correlation of geographic location and projection location (see also Author response image 5). This illustrates the contrast we were hoping to communicate, where projection locations of historical individuals start to “settle” close to present-day individuals from similar geographic locations, especially in contrast with pre-historic individuals.

Author response image 5.

Comparing geographic distance to PCA distance between pairs of historical and pre-historical individuals matched by geographic space. For each historical period individual we selected the closest pre-historical individual by geographic distance in an effort to match the distributions of pairwise geographic distance across the two time periods (left). For these distributions of individuals matched by geographic distance, we then queried the euclidean distance between their projection locations in the first two principal components (right).

Author Response

Reviewer #1 (Public Review):

“The authors use hM4Di to "silence" Fos-tagged neurons in the basal forebrain, but they have not validated the efficiency or the possible various effects of this reagent.

It is possible that hM4Di actually has a relatively small effect on suppressing the AP activity of neurons. Nevertheless, hM4Di might still be an effective manipulation, because it was shown to additionally reduce transmitter release at the nerve terminal (see e.g. Stachniak et al. (Sternson) 2014, Neuron). Thus, the authors should evaluate in control experiments whether hM4Di expression plus CNO actually electrically silences the AP-firing of ChAT neurons in the BF as they seem to suggest, and/or if it reduces ACh release at the terminals. For example, one experiment to test the latter would be to perfuse CNO locally in the BLA; after expressing hM4Di in the cholinergic neurons of the BF. At the very least, the assumed action of hM4Di, and the possible caveats in the interpretation of these results should be discussed in the paper.”

We find that activation of hM4Di with clozapine in basal forebrain cholinergic neurons results in clear alterations to neuronal activation in projection targets and in behavior (Figures 3, Figure 3-Supplement 1, Figure 5, Figure 5-Supplement 1, Figure 5-Supplement 2, Figure 6-Supplement 1 and Figure 8). Previous studies demonstrated that activation of hM3Dq or hM4di in cholinergic neurons results in changes to electrical activity and behavioral response (Zhang et al. 2017 & Jin et al. 2019). Though we are unable to distinguish whether the effects on behavior in our experiments are a result of decreases in ACh release at terminals, inhibition of action potential firing, or both, our behavioral findings are consistent with demonstrations that inhibition of basal forebrain cholinergic neurons can alter behavior. See Page 17 Lines 488-493 for a discussion.

“The names of brain areas like "NBM/SIp" and "VP-SIa" need to be better introduced, and somehow contextualized (in the Introduction, and also at first reading in the Results).”

We agree that our prior presentation of these regions was confusing and in general the boundaries of these regions are not well-defined in the field. We have included a description of anatomical landmarks and bregma coordinates to clarify our definitions of the regions NBM/SIp (Page 4 Line 103-104) and VP/SIa (Page 4 Line 107-108).

“Figure 3C: Application of CNO on the memory recall day leads to a strong reduction in CS-driven freezing. However, in this experiment, and also in Fig. S7, the pre-tone value of freezing is also strongly reduced. This would indicate that the activity of NBM/SIp cells (or else, ACh-release from these cells - see also Major point 1), also influences contextual learning. The authors should, first, statistically, test these effects (I am not sure this was done). If these differences are significant, a possible role of ACh in contextual fear learning should be discussed. Has it been shown before whether ACh is involved in contextual fear learning? Does this indicate the involvement of another target area of ACh neurons (e.g., the hippocampus?).”

We statistically compared the pre-tone freezing response between Sham and hM4Di groups across our experiments and found no significant differences in pre-tone freezing between the groups (Figure 3D- Sham vs. ADCD-hM4Di, Pre-tone p=0.3544; Figure 5B- Sham vs. hM4di, Pre-tone p=0.0679; Figure 5C- Sham vs. hM4Di, Pre-tone p=0.0966; Figure 5-Supplement 2A- Sham vs. hM4Di, Pre-tone p>0.99). These comparisons can also be reviewed in the statistical reporting table uploaded along with the manuscript.

“The discussion could be improved by better comparing what they found, to the wider literature. For example, previous papers studying other neuromodulatory systems found evidence for a modulation of neuromodulator release after learning, e.g. see Martins and Froemke 2015 Nat. Neuroscience for the noradrenergic system, Tang et al. (Schneggenburger lab) 2020 J. Neuroscience for the dopaminergic system and fear learning; and Uematsu et al., 2017, Nat. Neuroscience for the noradrenergic system and fear learning. Maybe the authors could include these and similar references when revising their discussion to take into account a broader view of previous findings related to other neuromodulatory systems.”

Our study joins the growing body of literature demonstrating stimulus-encoding and rapid stimulus-contingent responses in various neuromodulatory systems in learning and memory recall. We have now added a substantial discussion, detailing both the similarities and differences between our findings and those found in the dopaminergic, serotonergic, noradrenergic, and oxytocinergic systems in fear learning. See Pages 20-21 Lines 575-605.

Reviewer 2 (Public Review):

“Throughout the paper, the authors use comparisons of cell activity between groups to address questions about projection-specific and cue-specific cell activation and reactivation. However, statistical comparisons are sometimes done between biological replicates (e.g. Fig. 5A), whereas a lot of them are done between technical replicates (e.g. Fig. 2B, 5B, 7B). Adding statistics that compare biological replicates would help increase confidence in the results.”

We have replotted our data as a comparison of biological replicate (by individual animal) in new versions of Figures 1-8, and Figure 1-Supplements 1-3, Figure 5-Supplements 1 & 2, Figure 6-Supplements 1 & 2, Figure 7-Supplement 1, and Figure 8-Supplement 1. Correspondingly, all statistical analyses have been conducted comparing biological replicates. To note, these changes have not changed the overall conclusions of each figure. The sample size, statistical test and p-values for our comparisons are included in the figure legends and in the newly included statistical reporting table.

"To demonstrate engram-like specificity, in figure 4C the authors show fold change in cholinergic reactivation in low and high responders (animals that show low and high defensive freezing upon cue presentation) as normalized by cell activity while sitting in the home cage. However, the authors also collected a better control for this comparison, which is shown in figure S4, where the animals were exposed to an unconditioned tone cue. Comparing fold change to this tone-alone condition would provide stronger evidence for the authors' point, as this would directly compare the specificity of cholinergic reactivation to a conditioned vs an unconditioned cue. A discussion of the same comparison is relevant for figure 2 (and is shown in figure S4) but is not mentioned in the text.”

We have evaluated the cholinergic response to the tone using GRABACh3.0 as a readout of ACh release in the BLA, and using IEG expression as a readout of cholinergic neuron activation. We find no significant increase in ACh release in the BLA in response to tone presentation (Figure 1C-left, 1D-left) and no significant increase in tone associated reactivation of cholinergic neurons (using IEG as a readout, 2C/D, Figure 1-Supplement 2, Figure 1-Supplement 3, Figure 6-Supplement 1A) unless the tone has been previously paired with a foot shock(see Figure 1C-right, 2C, 3D). In addition, we find no statistically significant differences between home cage and tone alone conditions (Figure 2C – home cage-home cage condition vs. tone-tone condition, p=0.5012; Based on these analyses, we use the home cage group as our control group for comparison.

“The significant correlation between cue-evoked percent change in defensive freezing from pretone and fold change in cholinergic cell activity relative to the home cage that is shown in figure 4D is somewhat confusing. Is the correlation considering all the points shown (high and low responders as depicted by black and grey points)? It's first reported as one correlation but then is discussed as two populations that have different results. Further, is the average amount of reactivation for the home-cage controls used here the same denominator for each reported animal? Similarly to the point above, a correlation looking at fold change from tonealone would also be helpful to determine the degree to which cholinergic reactivation is specific to threat-association learning versus the more general attentional component that this system is known for.”

We have substantially modified this figure, now new Figure 6, to clarify our point. Along with this revision, we have removed the correlation plots and corresponding analyses from the revised version of the manuscript and figures.

Figure 6 now begins with behavior data from a distinct cohort of mice outlining our criteria for high vs. low responders (Figure 6A/B). In Figure 6C, conducted in a separate cohort of mice that only underwent behavioral testing to clarify the definition of high vs. low responders, we note via schematic that ADCD labeling was carried out during the recall session (unlike Figure 2). In panel D, we show fold change of activated cholinergic neurons stratified by High vs. Low responder status. This fold change is normalized to the average activation from the home cage control animals in each experimental cohort. Taken together we find animals with a ~2 fold increase in activation of cholinergic neurons display significant, distinguishable freezing in response to the tone as compared to pretone freezing. We find that this cluster of activated neurons is segregated to the anterior NBM/SIp (Figure 6E).

Regarding the involvement of cholinergic reactivation tone response (attention) rather than learning - in Figure 1-Supplement 3, we evaluate ACh release and behavioral response in mice that were exposed to three shocks alone (no tone) on day 1 and then exposed to a single (novel) tone on day 2. In these mice we find no significant change in ACh release in the BLA in response to tone, and no significant increase in freezing behavior in response to the tone. In Figure 2D, we evaluate reactivation of cholinergic neurons in a similar context and find that this group does not significantly differ from the home cage → home cage group. Further, we present that this home cage group does not significantly differ from Low Responders. As such, we find significant reactivation of cholinergic neurons in animals with increased responsiveness to the CS tone during the recall session (High Responders).

“The compelling argument of this paper is that the authors are separating out the general attention role typically attributed to the cholinergic system from a more specific, engram-based role. Given the importance of untangling this, it would useful to see the recorded traces and behavioral scoring for the data shown in figure S2B. For example, was the higher slope in the recorded cholinergic response during unconditioned tone 1 also accompanied by an increase in freezing, which later went away with additional non-reinforced tones? Given that the animals were not habituated to tones (according to the Methods), this activity could be related to a habituation/general attention response, which may then be weaker than the learned response.”

We include individual traces of GRABACh3.0 release in the BLA in response to the unconditioned tone from a protocol with 3x tone presentation on Day 1 and tone presentation on Day 2 (Figure 1-Supplement 2C). We have also included average + SEM traces for the entire duration of the tone presentation for the three unconditioned tones in this paradigm along with an inset showing 1s before and after tone onset (Figure 1Supplement 2D). Finally, we include individual traces of GRABACh3.0 release in the BLA in response to the first (naïve) tone from mice that underwent the training (tone + shock) followed by recall (tone) paradigm in Figure 1-Supplement 4C, left. None of the unconditioned tone responses were statistically significantly different from the preceding baseline. Instead, we find the learned response is significantly higher than the response baseline (Figure 1D).

Author Response

Reviewer #1 (Public Review):

The authors used MD simulations to investigate the role of N-terminal myristoylation and the presence of two SH domains on the allosteric regulation of c-Abl kinase. Standard established MD simulation methods and analyses were applied, including the force distribution analysis (FDA) method developed by Grater et al. some time ago.

The system is large and the conformational changes are complicated. In light of this, and aggravated by the fact that direct comparison with - and critical testing against - experimental data is not possible in the present case, I consider the overall simulation times to be rather short (several repeats, but only 500 ns). So there might be statistical convergence issues. Especially also because at least some of the starting structures were generated from available experimental structures after some modifications/modelling, and they might thus be out of equilibrium and need some time to fully relax during the MD simulations.

Unfortunately, I cannot find any convergence tests concerning the length of the simulations, which are usually considered to be standard analyses (Appendix Fig. 5 shows the effect of different thermostats and capping of the peptide chain, but no tests concerning simulation time). This could be critical in the present case, where the authors acknowledge themselves (e.g., on p. 4) that there are only subtle differences between the different simulation systems and the variations within a given system are larger than the relevant (putative) differences between systems (Fig. 1 C, D, E).

We thank the reviewer for taking the time and critically assessing our manuscript. We appreciate and have addressed the raised concerns as follows. We have quadrupled the simulation time to 2 µs for 20 out of the 30 replicates and show the updated results for these. We refer the reviewer to the modified Fig. 2 and 3 (former Fig. 1 and 2) with the updated data. Our main conclusions remained unchanged, namely that Myr unbinding shifts the overall kinase domain dynamics towards an active state. We furthermore still observe allosteric signal propagation from the Myr binding site to the active site along the alpha_F helix and a collaborative effect of Myr and the SH domains. Only some minor points were not confirmed after analyzing the longer simulations, for example the force differences transmitted to the A-loop upon SH domain binding/unbinding (former Fig. 2D), and changes in amplitude of N- and C-lobe opening upon Myr unbinding (former Fig. 1E). Furthermore, to demonstrate convergence, we added block and autocorrelation analyses for Fig. 1 (now Fig. 2) to Fig. 2 – fig supplement 3, and observed good convergence across all systems. Finally, we also increased simulation times of the umbrella sampling from 50ns to 200ns, again without that the quantitative trends and our conclusions have changed (see also next point).

Issues with statistical convergence are expected not only for the standard MD simulations but also for the umbrella sampling simulations, as 50 ns sampling per window is nowadays not considered state of the art and is likely insufficient for quantitative binding free energy calculation, especially for membranes (see, e.g., DOI 10.1021/ct200316w). However, worrying about this latter aspect might neither be useful nor needed, because in our view the statement that myristoyl groups can bind to the membrane and that they can compete with binding in the hydrophobic protein pocket can hardly be considered a surprise and would not have required any simulation at all in my view because the experimental K_D values are available (Table 1). The very unfavourable K_d values for unbinding of Myr from both the hydrophobic protein pocket as well as from the membrane in fact show that this is not how it is expected to work in reality. The fully solvated state will be avoided due to its high free energy. Instead, isn't the myristoyl expected to directly transition from the pocket into the membrane, after membrane binding of the kinase in a proper orientation?

The experimental values were determined with different methods, i.e. estimated from zeta potential measurements in case of the membrane and calorimetry, which only considered the kinase domain instead of the SH3-SH2-kinase complex, in case of Abl. We thus found it appropriate to perform Umbrella Sampling simulations to ensure comparability. Additionally, these allowed us to study the effects of different alpha_I helix conformations, which had a significant impact on the free energy of Myr unbinding, precisely Abl with a partially unfolded helix reflected the experimental energy better than the crystal structure with a kinked helix. We highlight this more explicitly in the corresponding Discussion section. Regarding the simulation time per sampling window, we did a block analysis (Fig. 5 – fig supplement 1) as suggested in the cited reference and also extended the time of each sampling window from 50 ns to 200 ns. This did not significantly alter the results and, importantly, the relative differences between Abl and the membrane stayed the same and are in good agreement with the experimental values.

Concerning the metadynamics simulations, these are usually done to obtain a free energy landscape. Why was this not attempted here? In the present case, the authors seemed to have used metadynamics only for generating starting structures, with different degrees of helicity of the alpha_I part, for subsequent standard MD simulations. Not surprisingly, nothing much happened during the latter, and conformers with kinked/partially unfolded alpha_I as well as conformers with straight alpha_I were both found to be "stable", at least on the short simulation time scale. It could also not be expected that the SH domain would spontaneously detach in response to helix straightening - again, this would require much longer simulation times than 500 ns. Nevertheless, alpha_I straightening might very well reduce the binding affinity towards SH - this can only be explicitly studied with free energy simulations, however.

Our main goal was indeed to achieve different alpha_I helix conformations for subsequent Umbrella Sampling simulations, and found that helix formation is in principle possible without SH2 domain unbinding. We would like to emphasize the impact of the different helix conformations on the free energy of Myr unbinding, which further highlights the need to investigate these structures. We chose Metadynamics to obtain them because it only facilitates the transition away from the kinked conformation without biasing towards certain end structures or transition pathways, which we found advantageous compared to alternative methods such as targeted MD. The reason for not reporting a free energy surface is that we considered the helicity of all seven residues making up the kink within a single CV, which smeared the energy landscape to the point that it is almost completely flattened. Furthermore, orthogonal CVs such as new interactions between the alpha_I helix with the SH2 domain or positional adjustments of the SH2 domain would have to be considered for a reliable quantitative result. We nevertheless observed transient SH2 domain unbinding during the applied time scale and added histograms to Fig. 4 – fig supplement 1 (former appendix Fig. 4) to make this more obvious.

Reviewer #2 (Public Review):

The manuscript aims at understanding how the fatty acid ligand MYR inhibits the activity of Abl kinase. Despite a wealth of structural and biochemical data, a key mechanistic understanding of how MYR binding could inactive Abl was missing.

The authors used equilibrium and enhanced molecular dynamics (MD) simulations to masterfully answer open questions left by extensive experimental data in the mechanistic understanding of this system. The authors took advantage of several state-of-the-art simulation techniques and carefully planned simulations to extract a coherent understanding from a wealth of experimental facts.

The manuscript convincingly identifies an allosteric regulation by MYR. Allostery is often a source of confusion and sometimes is used as a magic catch-it-all explanation for poorly understood phenomena. Here, the authors show very compelling evidence of the existence of an allosteric mechanism. Also, they identify the physical origin of the allosteric pathway, providing a clear mechanistic understanding at the residue-level resolution. This is an impressive achievement.

We thank the reviewer for appreciating our work and its significance for understanding Abl regulation.

By leaving a pocket in the protein, MYR enables the protein's activation. But MYR is a highly hydrophobic molecule surrounded by water. Where could it go rather than quickly binding back to the protein pocket? By asking this reasonable question, the authors propose an exciting mechanistic hypothesis. The physical proximity of Abl kinase to a cellular membrane could lead to a competition between the protein and the membrane for MYR, leading to a novel layer of regulation for this kinase. Free energy calculations performed by the authors show that this hypothesis is reasonable from the thermodynamic point of view.

From a broader perspective, this manuscript is an important contribution to the discussion of four outstanding topics. 1) myristoylation is an example of lipidation, a post-translational modification where an acyl chain is covalently linked to a protein. The role of post-translational modifications has been greatly underappreciated and investigated in the MD community. However, as all the work on Sars-Cov2 and this contribution show, post-translational modifications can be crucial to understanding function. Ignoring them could lead to severely biased results. 2) the debate on the nature of allostery is still on the rage. Some authors claim that looking for a residue-level mechanistic chain of events that explains the allosteric action does not make sense and that the only way of thinking about allostery is as a sudden global change of the conformational landscape. Here, the authors show that instead, it is possible and leads to an essential understanding. 3) The authors hypothesize a novel crosstalk between the Abl and cellular membranes mediated by MYR. This exciting and far-reaching hypothesis opens the door to new complex layers of regulation. I suspect that these crosstalks between cytosolic proteins, or the soluble domain of membrane-tethered proteins and membranes, are much more ubiquitous than what has been appreciated so far. 4) From a methodological point of view, this manuscript represents a masterful use of simulations to put existing experimental data in a coherent picture. It is an example of the use of MD simulations at its best, where the simulations make sense of experiments, integrate existing data into a unified picture, and lead to new hypotheses that can be tested in future experiments.

We thoroughly appreciate the reviewers positive feedback and the valuable suggestions for improvement below.

It would be superb if the authors could propose precise predictions that could inspire future experiments. Now that they present a residue-resolution allosteric pathway, can they suggest point mutations that would interrupt it?

We have added a short segment to the end of the discussion proposing possible experiments.

Author Respones

Reviewer #1 (Public Review):

The manuscript by Hekselman et al presents analyses linking cell-types to monogenic disorders using over-expression of monogenic disease genes as the signal. The manuscript analyses data from 6 tissues (bone marrow, lung, muscle, spleen, tongue and trachea) together with ~1,000 rare diseases from OMIM (with ~2,000 associated genes) to identify cell-type of interest for specific disease of choice. The signal used by the approach is the relative expression of OMIM-genes in a particular cell type relative to the expression of the gene in the tissue of interest identifying celltype-disease pairs that are then investigated through literature review and recapitulated using mouse expression. A potentially interesting finding is that disease genes manifesting in multiple tissues seem to hit same cell-types. Overall this important study combines multiple data analyses to quantify the connection between cell types and human disorders. However whereas some of the analyses are compelling, the statistical analyses are incomplete as they don't provide full treatment of type I error.

Statistical analyses were changed to include permutation testing and a different threshold (Results, page 6, 1st paragraph; Methods, page 21-22, ‘PrEDiCT score calculation and significance assessment’; Figure 1–figure supplement 2). Assessments of type I error were based on literature text-mining and expert curation, and showed that false-positive rates were low in both (0.01 and 0.07, respectively; Figure 1F and Figure 1–figure supplement 4A).

Reviewer #2 (Public Review):

This study identifies 110 disease-affected cell types for 714 Mendelian diseases, based on preferential expression of known disease-associated genes in single-cell data. It is likely that many or most of the results are real, and the results are biologically interesting and provide a valuable resource. However, updates to the method are needed to ensure that inference of statistical significance is appropriately stringent and rigorous.

Strengths: a systematic evaluation of disease-affected cell types across Mendelian diseases is a valuable addition to the literature, complementing systematic evaluations of common disease and targeted analyses of individual Mendelian diseases. The validation via excess overlap with diseasecell type pairs from literature co-appearance provides compelling evidence that many or most of the results are real. In addition, many of the results are biologically interesting. In particular, it is interesting that diseases with multiple affected tissues tend to affect similar cell types in the respective tissues.

Limitations: the main limitation of the study is that, although many or most of the results are likely to be real, the criteria for statistical significance is probably not stringent enough, and is not welljustified. For diseases with only 1 disease-associated gene, the threshold is a z-score>2 for preferential expression in the cell type, but this threshold is likely to be often exceeded by chance. (For diseases with many disease-associated genes, the threshold is a median (across genes) zscore>2 for preferential expression in the cell type, which is less likely to occur by chance but still an arbitrary threshold.) Thus, there is a good chance that a sizable proportion of the reported disease-affected cell types might be false positives. The best solution would be to assess statistical significance via empirical comparison with results for non-disease-associated control genes, and assess the statistical significance of the resulting P-values using FDR.

We thank the reviewer for the valuable insights and suggestions. We revised the method to assess statistical significance by using empirical comparison followed by FDR correction, as suggested by the reviewer (Results, page 6, 1st paragraph; Methods, page 21-22, ‘PrEDiCT score calculation and significance assessment’; Figure 1–figure supplement 2).

The re-analysis using mouse single-cell data adds an interesting additional dimension to the study, with the small caveat that mouse single-cell data does not provide statistically independent information across genes (for the same reason that adding data from independent human individuals would not provide statistically independent information across genes, given that human and mouse expression are partially correlated).

We acknowledge this caveat in the text (Discussion, page 17, 2nd paragraph, lines 8-11).

Reviewer #3 (Public Review):

The authors describe the method, PrEDiCT, which helps identify disease affected cell types based on gene sets. As I understand it, the method is based on finding which "disease genes" (from an annotation) are relatively highly expressed. The idea is nice, however, I have concerns about how "significance" is assessed and the relative controls.

Overall, I find the idea interesting, but the execution raises some concerns.

1) From a causal perspective, there is an association of high expression of these genes within these cell types, but without also assessing individuals with those specific diseases, I do not it is fair to say "disease affected" cell types. It is possible that these genes might behave completely fine but are highly expressed in those cell types while being affected another in other cell types.

We agree with the reviewer. We changed the terminology to "likely disease-affected cell types” and added this caveat to the Discussion, page 16, 2nd paragraph.

2) It is unclear to me what the "null" comparison is in the method and if there is one. For example, by chance, would I expect this gene to be highly expressed because other genes are also highly expressed in this cell type? Some way to assess "significance" or "enrichment" beyond simply using ranks and thresholds would be helpful in deciding whether these associations are robust.

We revised the procedure for assessing statistical significance to include permutation tests. Specifically, given a disease D with n disease-associated genes, the null hypothesis was that the PrEDiCT score of these genes is not significantly different from the PrEDiCT score of a random set of n genes. To test this, we randomly selected n genes expressed in any cell type, and computed the PrEDiCT score for this random gene set in each cell type of the disease-affected tissue (referred to as ‘random score’). We repeated this procedure 1,000 times, resulting in 1,000 random scores per disease and cell type. The p-value of the PrEDiCT score of disease D in cell type c was set to the fraction of random scores in c that were at least as high as the original PrEDiCT score of D in c. The acquired p-values were adjusted for multiple hypothesis testing per disease using the Benjamini-Hochberg procedure. To increase stringency, we treated only statistically significant disease–cell-type pairs with PrEDiCT score≥1 as 'likely affected'. The procedure is detailed in Results, page 6, 1st paragraph; Methods, page 21-22, ‘PrEDiCT score calculation and significance assessment’; Figure 1–figure supplement 2. Additionally, we estimated type I error by using literature text-mining or expert curation (Results, page 7, 2nd paragraph; Methods, page 22, ‘Textmining of PubMed records’, and page 23, ‘Expert curation and assessment of disease-affected cell types’; Figure 1F and Figure 1–figure supplement 4A).

3) Additionally, it is unclear to me, but I suspect that there are unequal cell numbers in the scores computed as well as between relevant tissues. This is related to point (2) above, but as a result, the estimates of the scores will inherently have different variances, thus making comparisons between them difficult/unreliable unless accounted for. If I understand correctly, the score is first the average expression within a tissue, then, the Z-score? If so, my comment applies.

To clarify, the PrEDiCT score of a disease D in cell type c was set to the median preferential expression P of its disease genes (Equation 1 below). The preferential expression of each gene in c was computed as a Z-score, by comparing the average expression of the gene in c to its average expression in all cell types of the tissue, divided by the standard deviation (SD, Equation 2 below). Tissues indeed had unequal numbers of cell types, however, the distribution of PrEDiCT scores were similar between tissues (now in Supplementary File 13). We revised this part of Methods and added Equations 1 and 2 (Methods, page 21-22, ‘PrEDiCT score calculation and significance assessment’) and Supplementary File 13.

4) There is a large set of work done in gene enrichment sets which appears to not be mentioned (e.g. GSEA and other works by the Price group). It would be helpful for the authors to summarize these methods and how their method differs.

We added work done in gene enrichment sets (including two relevant and recent studies from the Price group) and summarized these methods in the Introduction (page 2-3).

5) Additionally, it should be noted that a caveat of this analysis is that the comparisons are all done only relative to the cell types sampled and the diseases which have Mendelian genes associated with them. I would expect these results to change, possibly drastically, if the sampled cell types and diseases were to be changed.

We agree with the reviewer and now discuss the generalizability of our results, relating to the extent of the sampled cell types (Discussion, page 18, 1st paragraph).

6) Finally, I would appreciate a more detailed explanation in the methods of how the score is computed. Some equations and the data they are calculated from would be helpful here.

We now provide a detailed explanation of how the score and its statistical significance were computed and added Equations 1 and 2 (Methods, page 21-22, ‘PrEDiCT score calculation and significance assessment’).

In summary, the general idea is an interesting one, but I do think the issues above should be addressed to make the results convincing.

We thank the reviewer for the important feedback which helped us strengthen our analyses.

Author Response

Thank you for providing us with the reviewer comments. We will provide the revised manuscript at a later stage as recommended.

Reviewer #1 (Public Review):

Summary:

In this manuscript, the authors used machine learning algorithm to analyze published exosome datasets to find biomarkers to differentiate exosomes of different origin.

Strengths:

The performance of the algorithm are generally of good quality.

Weaknesses:

The source datasets are heterogeneous as described in Figure 1 and Figure 2, or Line 72-75; and therefore questionable.

We thank the reviewer for this assessment. The commonly used biomarkers of exosomes exhibit heterogeneous presence and abundance within the exosomes derived from different cell lines, tissue, and biological fluids. The primary goal of this study was to identify universal exosomal biomarkers that remain consistent across different sources of exosomes, unaffected by potential isolation and quantification bias. This objective was achieved through an integration of datasets from different sources, which allowed for the subsequent identification of common proteins associated with exosomes. Among the 18 protein markers identified, it is noteworthy that they are universally abundant in all cell lines and their exosomes. We believe that despite the heterogeneity of the datasets used here, the identification of 18 universal protein markers in exosomes from diverse sources is a strength of this analysis.

Reviewer #2 (Public Review):

Summary:

This is a fine work on the development of computational approaches to detect cancer through exosomes. Exosomes are an emerging biomarker resource and have attracted considerable interests in the biomedical field. Kalluri and co-workers collected a large sample pool and used random forest to identify a group of protein markers that are universal to exosomes and to cancer exosomes. The results are very exciting and not only added new knowledge in cancer research but also a new and advanced method to detect cancer. Data was presented very nicely and the manuscript was well written.

Strengths:

Identified new biomarkers for cancer diagnosis via exosomes.

Developed a new method to detect cancer non-invasively.

Results were presented nicely and manuscript were well written.

Weaknesses:

N/A.

We appreciate the the enthusiastic assessment of our study by the reviewer.

Reviewer #3 (Public Review):

In the current study, Li et al. address the difficulty in early non-invasive cancer diagnosis due to the limitations of current diagnostic methods in terms of sensitivity and specificity. The study brings attention to exosomes - membrane-bound nanovesicles secreted by cells, containing DNA, RNA, and proteins reflective of their originating cells. Given the prevalence of exosomes in various biological fluids, they offer potential as reliable biomarkers. Notably, the manuscript introduces a new computational approach, rooted in machine learning, to differentiate cancers by analyzing a set of proteins associated with exosomes. Utilizing exosome protein datasets from diverse sources, including cell lines, tissues, and various biological fluids, the study spotlights five proteins as predominant universal exosome biomarkers. Furthermore, it delineates three distinct panels of proteins that can discern cancer exosomes from non-cancerous ones and assist in cancer subtype classification using random forest models. Impressively, the models based on proteins from plasma, serum, or urine exosomes achieve AUROC scores above 0.91, outperforming other algorithms such as Support Vector Machine, K Nearest Neighbor Classifier, and Gaussian Naive Bayes. Overall, the study presents a promising protein biomarker signature tied to cancer exosomes and proposes a machine learning-driven diagnostic method that could potentially revolutionize non-invasive cancer diagnosis.

We appreciate this positive assessment of our work.

Author Response

The following is the authors’ response to the original reviews.

eLife assessment

The study by O'Reilly and Delis provides a valuable data-driven framework for extracting task-related muscle synergies in a step towards the understanding and practical use of synergies in real scenarios (e.g., evaluation of patients in a clinical environment). The approach is incomplete since the authors did not compare their method with classical physiologically grounded approaches for assessing muscle synergies. In this sense, the comparisons with classical approaches would clarify if physiological assemblies were preserved and were not altered to incorporate task space variables. Despite limitations, the proposed framework would interest motor control and neural engineering researchers.

We thank the editors for the positive assessment of our work and appreciate their constructive feedback. In our revised manuscript, we believe we have sufficiently addressed the identified limitations by a) comparing our approach to existing physiologically-based methods, providing thorough comparisons of their respective outputs, b) applying it to a dataset of post-stroke participants to demonstrate that it can identify physiologically-interpretable markers of motor recovery and c) providing examples to demonstrate how readers can interpret the novel perspective introduced.

Reviewer #1 (Public Review):

The proposed study provides an innovative framework for the identification of muscle synergies taking into account their task relevance. State-of-the-art techniques for extracting muscle interactions use unsupervised machine-learning algorithms applied to the envelopes of the electromyographic signals without taking into account the information related to the task being performed. In this work, the authors suggest including the task parameters in extracting muscle synergies using a network information framework previously proposed. This allows the identification of muscle interactions that are relevant, irrelevant, or redundant to the parameters of the task executed.

The proposed framework is a powerful tool to understand and identify muscle interactions for specific task parameters and it may be used to improve man-machine interfaces for the control of prostheses and robotic exoskeletons.

With respect to the network information framework recently published, this work added an important part to estimate the relevance of specific muscle interactions to the parameters of the task executed. However, the authors should better explain what is the added value of this contribution with respect to the previous one, also in terms of computational methods.

We thank the reviewer for their constructive comments. We have adjusted the introduction section of the manuscript to better explain the added value of this framework over previous work. Specifically, we draw the reviewer’s attention to the following updated section of the introduction:

“In [11], we considered, key limitations among current approaches to muscle synergy analysis in extracting functionally relevant and interpretable patterns of muscle activity [12]. We proposed a combinatorial approach based on information- and network-theory and dimensionality reduction (the network-information framework (NIF)) that significantly improved the generalisability of the extraction process by, among others, removing restrictive model assumptions (e.g. linearity, same mixing coefficients) and the reliance on variance-accounted-for (VAF) metrics [12]. By determining the pairwise mutual information between muscles, this innovation paved the way for the appropriate mapping of muscular interactions to the task space. To elaborate on the significance of this development, the extraction of motor patterns in isolation of the task space comes at the expense of both functional and physiological relevance [12,13]. Furthermore, effective methods for mapping large-scale physiological dynamics to behaviour is a current gap across the neurosciences [14]. Thus, here we build on this work by, for the first time, directly including task space parameters during muscle synergy extraction. In doing so, we address these current research gaps, progressing muscle synergy research and successful engineering applications in a fruitful direction [12,15,16]. This enables us, in a novel way, to dissect the concept of the muscle synergy and therefore quantify interactions between muscle activations with shared or complementary functional roles. “

In general, the method proposed relies on several hyperparameters and cost functions that have been optimized for the specific datasets. A sensitivity analysis should be performed, varying these parameters and reporting the performance of the framework.

We thank the reviewer for this comment which enabled us to clarify a potential misunderstanding. Our proposed framework does not require setting or varying hyperparameters to optimise cost functions.

For model-rank specification, a modularity maximising cost-function is used which determines what partitioning of the networks results in maximal modularity. We have offered two alternative approaches using this cost-function which consistently converge on the same solution. To further ensure the representativeness of this solution, we also offer a consensus-based approach where we apply these alternative approaches to individual participant or task data, then group the collective partitions together and re-apply the approaches. One of these approaches (Equation 2.2) requires two hyperparameters, γ and ω, which adjust the intra- and inter- network layer resolutions. As stated in the manuscript, we set both of these parameters to 1, thus nullifying their presence in the cost-function and aligning our work with the classical notion of modularity. Across the two alternative approaches to model-rank specification, the solution is unique and data-driven and has a demonstratable generalisability across datasets.

The only other cost-function present in the framework is during dimensionality reduction, which is a standard loss function used across the muscle synergy analysis literature. Thus, the approach is essentially parameter-free and we now have mentioned this more explicitly in the manuscript:

“To empirically determine the number of components to extract in a parameter-free way, we then concatenated these adjacency matrices into a multiplex network and employed network community-detection protocols to identify modules across spatial and temporal scales (fig.3(D)) [29–32,44].”

“In its generalised multilayer form, the Q-statistic is given an additional term to consider couplings between layers l and r with intra- and inter-layer resolution parameters γ and ω (Equation 2.2). Here, μ is the total edge weight across the network and γ and ω were set to 1 in the current study for classical modularity [30], thus removing the need for any hyperparameter tuning.”

It is not clear how the well-known phenomenon of cross-talk during the recording of electromyographic muscle activity may affect the performance of the proposed technique and how it may bias the overall outcomes of the framework.

Indeed artifacts such as crosstalk are a standard issue across the EMG literature and may impact the performance of subsequent analyses where prevalent in the dataset. Crosstalk is expected to be present irrespective of the task and so should not affect redundant and synergistic muscle representations, however it could be present in the task-irrelevant muscle interactions extracted. Due to the prominence of long-range functional connections with the task-irrelevant representations extracted, we suggest that such artifacts are unlikely to have played a prominent role in the extracted patterns. Nonetheless, we have recognised this possibility with the following updated sentence in the Discussion section:

“Although distinguishing task-irrelevant muscle couplings may capture artifacts such as EMG crosstalk, our results convey several physiological objectives of muscles including gross motor functions [65], the maintenance of internal joint mechanics and reciprocal inhibition of contralateral limbs [20,50].”

Reviewer #2 (Public Review):

This paper is an attempt to extend or augment muscle synergy and motor primitive ideas with task measures. The authors idea is to use information metrics (mutual information, co-information) in 'synergy' creation including task information directly. My reading of the paper is that the framework proposed radically moves from attempts to be analytic in terms of physiology and compositionality with physiological bases, instead into more descriptive ML frameworks that may not support physiological work easily.

We thank the reviewer for taking the time to provide a thorough commentary on this manuscript. An overall aim in developing this framework is to build on other recent developments in providing a more fine-grained functional architecture underlying movement control [1,2]. It is a requirement for the successful communication and introduction of this toolbox to the field to provide readers with an understanding of how to use the framework and an intuition on how to interpret the results. Thus, we agree with the reviewer that functional interpretations are of crucial use.

We also agree with the reviewer that maintaining a physiological underpinning is a desirable direction for the field and should not be made secondary to functional descriptions. In our updated version of this manuscript, we have therefore included direct comparisons with the gold-standard in the field for muscle synergy extraction, namely non-negative matrix factorisation based muscle synergy extraction (see ‘Building on current approaches to muscle synergy analysis’ and fig.5-6 of revised manuscript) [3,4]. In these comparison, we show how our framework goes beyond this current approach in terms of functional insight while still maintaining physiological relevance. Indeed, in the revised manuscript we also include a fourth dataset comprising post-stroke participants and healthy controls (Fig.6). We demonstrate, through a simple example application to this dataset, how our proposed framework can produce more predictive representations of motor impairment than the gold-standard approach. The representations we identified were discriminative of motor impairment measured via the Fugl-Meyer assessment using just one trial per participant. This improves considerably upon the sensitivity of the current approach to altered motor patterns which have predominantly required many trials and participants to gain significance [5,6]. Thus, the patterns we extract are a more comprehensive representation of the actual underlying physiological state of the participants.

This approach is very different from the notions of physiological compositional elements as muscle synergies and motor primitives, and to me seems to really be striving to identify task relevant coordinative couplings. This is a meta problem for more classical analyses. Classical analyses seek compositional elements stable across tasks. These elements may then be explored in causal experiments and generative simulations of coupling and control strategies. The present work does not convince me that the joint 'meta' analysis proposed with task information added is not unmoored from physiology and causal modeling in some important ways. It also neglects publications and methods that might be inconvenient to the new framework.

We would be very interested in receiving the reviewer’s suggestions of existing approaches that we have not incorporated here and would be happy to discuss these in the revised manuscript.

Information based separation has been used in muscle synergy analyses using infomax ICA, which is information not variance based at core. Though linear mixing of sources is assumed, minimized mutual information is the basis.

We agree with the reviewer that ICA relies on information measures, however it does not incorporate task-space information. The novelty of our approach lies in the characterisation of muscle interactions with respect to the task at hand. If the reviewer could provide references to this statement, we would be able to consider this further.

Physiological causal testing of synergy ideas is neglected in the literature reviews in the paper. Although these are in animal work, the clear connection of muscle synergy choices and analyses to physiology is important and needs to be managed in the new methods proposed. Is any correspondence assumed? Possible?

We agree with τhe reviewer that this a crucial element of muscle synergy research and will aim to address it in our future work. However, we would like to point out that the current manuscript is a “tools and resources” article aiming to introduce a new framework. In our revised manuscript, we have incorporated an application of the framework to a dataset from post-stroke patients to demonstrate the use of the framework in clinical settings to identify biomarkers and use them to make predictions of motor recovery (see Fig.6 of updated manuscript).

Questions and concerns with the framework as an overall tool:

First, muscle based motor information sources have influences on different time scales in the task mechanics. Analyses of synergies in the methods proposed will be very much dependent on the number and quality of task variables included and how these are managed. Standardizing and comparing among labs, tasks sets and instrumentation differences is not well enough considered as a problem in this new proposed method toolset, at least in my reading. Will replication, and testing across groups ever be truly feasible in this framework?

We agree with the reviewer that this important point can be a limitation of the applicability of the framework. For this reason, we chose a “holistic” approach, applying the framework to several datasets collected in different settings, and selecting different kinds of task variables to extract muscle networks from. Crucially, we used a leave-one-task-out and leave-one-participant-out cross validation procedure to specifically address this point. Our results showed that the extracted couplings are robust irrespective of the task variable and/or participant excluded and this lends credit to the generalisability of the framework.

Muscle based motor information sources have influences on different time scales in the task mechanics. Kinematic analyses, dynamic analyses and force plate analyses of the same task may provide task variables that alter the results in the proposed framework it seems.

As we have mentioned above, here we used all the above types of task variables together to illustrate the range of measures that can be included in the proposed framework and showed that the outputs are robust to the exclusion of any task/participant. This point is especially evident for dataset 3 results, where high levels of generalisability were found despite the inclusion of kinematic, dynamic and IMU data (see Table 1. of original submission and updated manuscript). We believe that this is an advantage of the approach as it allows researchers to apply the method to different kinds of measurements they may have collected and gain insights into the relationships of muscle couplings with kinematic/dynamic/force parameters. This will also enable scientists to attribute different functional roles to the identified couplings and it is something we plan to do in future applications of the framework.

Second, there is a sampling problem in all synergy analyses. We cannot record all muscles or all task parameters. Examining synergies across multiple tasks seeks 'stationary' compositionality. Including task specific elements may or may not reinforce or give increased coordinative precision to the stationary compositionality.

We fully agree that this is a limitation of all synergy analyses and aimed to consider this study a step in the direction of addressing this limitation by providing the research community with a toolbox that can be used to quantify muscle couplings that can have different levels of task specificity.

To me the new methods proposed seem partly orthogonal to the ideas of stable compositionality. The 'synergies' obtained will likely differ, and are more likely to be coordinative control groupings of recurrent task and muscle motifs (based on instrumentation) which may or may not relate to core compositionality in physiology. Is there any expectation that the framework should relate to core compositionality and physiology. This is not clear in the paper as written.

In our new analysis, we have compared the proposed approach to existing physiologically-based methodologies and showed that the new framework can capture several salient physiological features of movement that the current NMF-based approach cannot. For example, as we have moved away from optimising variance accounted for metrics, our framework can identify subtle muscle couplings that have important functional roles. These subtle couplings are often not captured in current muscle synergy analysis as, against physiological relevance, higher amplitude muscles often take prominence. Further, by directly including task parameters during extraction, we can determine the muscles that have a functional role concerning the included task parameter rather than inferring this relationship indirectly using knowledge about the task executed. In our updated manuscript, by applying the framework to post-stroke participants (see Fig.6), we were also able to demonstrate that the extracted couplings are associated with functional parameters of motor recovery and have a clear link with the physiological state of individual participants.

It would be useful to explore the approach with a range of neuromechanical models and controllers and simulated data to explore the issues I am raising and convince readers that this analysis framework adds clarity rather than dissolving the generalizability and interpretability of analyses in terms of underlying causal mechanisms.

The authors need to better frame their work in relation to causal analyses if they are claiming links to muscle synergies analyses and claim extension/refinement. Alternatively, these may not be linked, and instead parallel approaches exploring different hypotheses and goals using different organizational data descriptors.

To address the reviewers concerns here, we have included in the updated manuscript a toy example simulating situations in which pairs of muscles would have a redundant or synergistic functional relationship (see Fig.2). This simulation gives clear intuition on situations where two muscles (e.g. an antagonist-agonist pair) may share functionally similar or complementary information about task direction (left vs right). In particular, within the main text describing this figure, we state how current NMF based approaches consider muscles functionally equivalent when they share similar magnitude activations, whereas our framework captures muscles with identical task information. Thus, our work is an extension of current approaches towards understanding causal mechanisms. The suggestion to use neuromechanical models is valuable, however we consider it beyond the scope of this work. This “Tools and Resources” paper is aimed at introducing the computational framework for the analysis of large-scale muscle couplings in task space. Our future work will use this framework to address unanswered questions in the field and we hope that it will be helpful for other scientists in testing their hypotheses.

To me this appears a data science tool that may not help any reductionist efforts and leads into less interpretable descriptions of motor control. Not invalid, but sufficiently different that common term use muddies the water.

We believe that the novel evidence we provided both on simulated and real data have contributed to a better interpretability of the approach outcomes. Specifically, we have introduced examples showing the functional roles of the different types of interactions as well as the predictive power of the outputs. Concerning the use of the term synergy, we have provided a clear description throughout the manuscript regarding the interpretation of synergy vs redundancy in the novel perspective we propose. For example in the discussion section:

“ We thus sought to provide greater nuance to the notion of ‘working together’ by defining motor redundancy and synergy in information-theoretic terms [6,56]. In our framework, redundancy and synergy are terms describing functionally similar and complementary motor signals respectively, introducing a new perspective that is conceptually distinct from the traditional view of muscle synergies as a solution to the motor redundancy problem [3,6,7]. In this new definition of muscle interactions in the task space, a group of muscles can ‘work together’ either synergistically or redundantly towards the same task. In doing so, the perspective instantiated by our approach provides novel coverage to the partitioning of task-relevant and -irrelevant variability implemented by the motor system along with an improved specificity regarding the functional roles of muscle couplings [20–22]. Our framework emphasises not only the role of functionally redundant muscle couplings that result from the underlying degeneracy of the motor system, but also of complementary, synergistic dependencies that are important for communication and integration across specialised neural circuitry [57,58]. Thus, the present study aligns the muscle synergy concept with the current mechanistic understanding of the nervous system whilst offering an analytical approach amenable to the continued advances in large-scale data capture [14,59].”

Reviewer #3 (Public Review):

In this study, the authors developed and tested a novel framework for extracting muscle synergies. The approach aims at removing some limitations and constrains typical of previous approaches used in the field. In particular, the authors propose a mathematical formulation that removes constrains of linearity and couple the synergies to their motor outcome, supporting the concept of functional synergies and distinguishing the task-related performance related to each synergy. While some concepts behind this work were already introduced in recent work in the field, the methodology provided here encapsulates all these features in an original formulation providing a step forward with respect to the currently available algorithms. The authors also successfully demonstrated the applicability of their method to previously available datasets of multi-joint movements.

Preliminary results positively support the scientific soundness of the presented approach and its potential. The added values of the method should be documented more in future work to understand how the presented formulation relates to previous approaches and what novel insights can be achieved in practical scenarios and confirm/exploit the potential of the theoretical findings.

Strengths:

This work proposes a novel framework that addresses physiologically non-verified hypothesis of standard muscle synergy methods: it removes restrictive model assumptions (e.g. linearity, same mixing coefficients) and the reliance on variance-accounted-for (VAF) metrics.

The method is solid and achieves the prescribed objectives at a computational level and in preliminary laboratory data.

A toolbox is available for testing the methods on a larger scale.

The paper is well written and shows a high level of innovation, original content and analysis

Weaknesses:

Task performance variables could be specified in more quantitative definition in future work (e.g.: articular angles rather than a generic starting point- end point).

We agree with this point and will incorporate it in future work. Our aim here was to show that the framework would work with any task variable and that scientists can use it to identify the relevance of muscle interactions to different types of task parameters.

The paper does not show a comparison with previous approaches (e.g.: NMF) or recently developed approaches (such as MMF).

We have now illustrated such a comparison on two datasets and explained more how the new framework can dissect the different types of muscle groupings (see ‘Building on current approaches to muscle synergy analysis’ section and Fig.5-6 of revised manuscript).

A discussion of the likely impact of the work on the field, and the utility of the methods and data to the community.

In our revised manuscript, we have introduced 2 new applications of the framework to real data to exemplify its use for a) functional interpretability and b) identification of biomarkers (see ‘Building on current approaches to muscle synergy analysis’ section and Fig.5-6 of revised manuscript). We also point towards its use in movement restoration and augmentation devices and in the clinical setting in the discussion section:

“The separate quantification of these muscle interaction types opens up novel opportunities in the practical application of muscle synergy analysis, as demonstrated in the current study through the identification of a significant predictor of motor impairment post-stroke from single-trials [5,12,65]. For instance, these distinct representations may encapsulate different neural substrates that can be specifically assessed at the muscle-level for the purpose of bodily restoration and augmentation [66]. Uncovering their neural underpinnings is an interesting topic for future research.”

In this work, the effort of the authors aimed at developing the field is clear. It is fundamental to develop novel frameworks for synergy extraction and use them to make them more interpretable and applicable to real scenarios, as well as more adherent to recent findings achieved in motor control and neuroscience that are not reflected in the standard models. At the same time, muscle synergies are being used more and more in research but their impact in practical scenarios is still limited, probably because synergies have rarely been analyzed in a functional context. This paper shows a very in-depth analysis and a novel framework to interpret data that links to the task space from a functional perspective. I also found that the results on the datasets are very well commented but could expand more to show why using this framework is advantageous.

There are some key points for discussion that follow from this paper which can be described more, maybe in future work, and that might contribute to major developments in the field, including:

The understanding of how the separation between relevant (redundant and synergistic) and irrelevant synergies impact on synergy analysis in practical works;

We have now introduced new figures (Fig. 5 and 6) to the revised manuscript, demonstrating simple applications of the framework and providing intuition regarding the outputs. We have also added points to the Discussion commenting on the differences between types of couplings and how they can be interpreted in future works:

“Our framework emphasises not only the role of functionally redundant muscle couplings that result from the underlying degeneracy of the motor system, but also of complementary, synergistic dependencies that are important for communication and integration across specialised neural circuitry [57,58]. Thus, the present study aligns the muscle synergy concept with the current mechanistic understanding of the nervous system whilst offering an analytical approach amenable to the continued advances in large-scale data capture [14,59].”

“Although distinguishing task-irrelevant muscle couplings may capture artifacts such as EMG crosstalk, our results convey several physiological objectives of muscles including gross motor functions [64], the maintenance of internal joint mechanics and reciprocal inhibition of contralateral limbs [19,49]. Thus, task-irrelevant muscle interactions reflect both biomechanical- and task-level constraints that provide a structural foundation for task-specific couplings. The separate quantification of these muscle interaction types opens up novel opportunities in the practical application of muscle synergy analysis, as demonstrated in the current study through the identification of a significant predictor of motor impairment post-stroke from single-trials [5,12,65]. For instance, these distinct representations may encapsulate different neural substrates that can be specifically assessed at the muscle-level for the purpose of bodily restoration and augmentation [66]. Uncovering their neural underpinnings is an interesting topic for future research.”

Interpreting how different synergistic organizations described in this work allows to better describe data from real scenarios (e.g.: motor recovery of patients after neurological diseases);

We have now added an example application of the framework to a dataset of stroke patients (Fig.6) and identified a redundant muscle patterns that are predictive of functional measures.

Discussing in detail how the presented findings compare with standard algorithms such as NMF to determine the added value provided with this approach;

As indicated above, we have now shown such a comparison on two new datasets (see Fig.5-6 of revised manuscript).

Describe how redundant synergies reflect real neural organization and - if their "existence" is confirmed - how they contribute to redesign the concept of muscle synergies and of modular/synergistic control in general.

This is an important point that we have now addressed more in our Discussion by relating redundant muscle couplings to degeneracy in the motor system and synergistic couplings to integrative dynamics by higher-level processes. We have also added a simple simulation illustrating how synergistic and redundant interactions co-exist and represent different contributions to task performance (see Fig.2 of revised manuscript).

Author Response

The following is the authors’ response to the original reviews.

Summary of changes

I thank the reviewers for their thorough feedback on this paper and providing me with such a detailed list of recommendations. I have been able to incorporate many of their suggestions, which I believe has greatly improved this paper.

The most important changes:

• I added comparisons to the lexicon- and rule-based sentiment algorithms TextBlob and VADER to Supplementary Fig. 4. This shows the superiority of ChatGPT in scoring the sentiment of scientific texts compared to existing and already-validated tools for sentiment analysis based on natural language processing. [Suggestion Reviewer 2]

• I added the measure intra-class correlation to Fig. 3b, emphasizing the inconsistency in sentiment scores across different reviews of the same paper. [Suggestion Reviewer 3]

• I added Supplementary Fig. 6, in which I directly propose different experiments to test the causes of the observed gender effects on peer review. [Suggestion Reviewer 3]

• I further studied the issue of variability in responses by ChatGPT (Supplementary Fig. 2), and learned that this has greatly improved in the latest version of ChatGPT (for Version Aug 3, 2023, R2 values of 0.99 (sentiment) and 0.86 (politeness) were reached). I show these findings in Supplementary Fig. 2. [Suggestions Reviewers 1 and 3]

• Throughout the manuscript (most notably in the Abstract and Discussion), I emphasize that this is a proof-of-concept study, and make suggestions on how to scale this up across journals and fields. I also toned down certain claims given the relatively small sample size of this study, including in the abstract. I also more prominently and elaborately discuss the limitations of the study in the Discussion section. [Suggestions Reviewers 1, 2 and 3]

• I made many smaller changes to text, figures and references on the basis of the reviewers’ comments. [Suggestions Reviewers 1, 2 and 3]

Notably, Reviewer 3 has provided me with a very detailed list of recommendations for follow-up experiments. I appreciate their ideas, and I am currently considering different options for future work. Specifically I am looking to team up with a journal to perform the experiments laid out in Supplementary Fig. 6 of the new paper, to study whether I can find evidence of bias across rejected and accepted papers. As suggested by this reviewer, I am also looking into ways to automate data collection using APIs, and by utilizing the rapidly expanding databases for transparent peer review.

Based on this preprint, I have received messages from academics that are interested in using generative AI to study scientific texts. By revising this manuscript, I hope to provide them with the tools to concurrently expand the analysis of peer review into different scientific disciplines and journals.

Reviewer #1 (Public review)

Strengths:

The innovative method is the biggest strength of this article. Moreover, the method can be implemented across fields and disciplines. I myself would like to see this method implemented in a grander scale. The author invested a lot of effort in data collection and I especially commend that ChatGPT assessed the reviews twice, to ensure greater objectivity.

I want to thank this reviewer for commending the innovative methodology of this study. I appreciate that this reviewer would like to see this methodology implemented at a grander scale, which is a view that I share. I initially only included Neuroscience papers, because I was uncertain whether I would be able to properly assess the reviews from different scientific disciplines (and thus judge whether ChatGPT was able to provide plausible scores).

The reviewers have provided me with a list of potential follow-up experiments, and I am currently considering different options for future work. Specifically I am looking to team up with a journal to perform the experiments laid out in (the new) Supplementary Fig. 6 of the new paper, to study whether I can find evidence of bias across rejected and accepted manuscript of a journal. In addition, as suggested by Reviewer #3, I am looking into ways to automate data collection using APIs, and by utilizing the rapidly expanding databases for transparent peer review. Importantly, based on this preprint, I have received messages from academics that are interested in using generative AI to study scientific texts. By revising this manuscript now, I hope to provide them with the tools to concurrently expand the analysis of peer review into different scientific disciplines and journals.

The comments I received from the different reviewers made me realize that I did not describe the intent of this paper well enough in the original submission. I rewrote much of the Abstract, to emphasize the proof-of-concept nature of this study, and rewrote the Discussion to focus more on the limitations of the study.

Weaknesses:

I have several concerns regarding the methodology of the article. The first relates to the fact that the sample is not random. The selection of journal and inclusion and exclusion criteria do not contribute well to the strength of the evidence.

Indeed, the inclusion of only accepted manuscript from a single journal is the biggest caveat of this paper. I have re-written much of the Abstract to emphasize that this is a proof-of-concept paper, hoping that other researchers concurrently expand this method to larger and more diverse datasets.

An important methodological fact is that the correlation between the two assessments of peer reviews was actually lower than we would expect (around 0.72 and 0.3 for the different linguistic characteristics). If the ChatGPT gave such different scores based on two assessments, should it not be sound to do even more assessments and then take the average?

This was a great recommendation by this reviewer, and a point also raised by Reviewer #3. Based on their suggestion, I looked into how each additional iteration of scoring would reduce the variability of scoring for a subset of papers (thus being able to advice users on an optimal number of iterations).

Interestingly, I observed that ChatGPT has become significantly more reliable in providing sentiment and politeness scores in recent versions. For the latest version (ChatGPT Aug 3, 2023), R2 = 0.992 for sentiment and R2 = 0.859 for politeness were reached for two subsequent iterations of scoring. Unfortunately, OpenAI does not allow access to previous version of ChatGPT, so the current dataset could not be re-scored. Yet, based on these data, there may no longer be a need for people to perform repeated scoring. I show these data in Supplementary Fig. 2, as I believe this is very useful information for people who are interested in using this tool.

Reviewer #1 (Recommendations to author)

I had some difficulties reading the article, so it would maybe help to structure the article more (e.g. In the introduction there are three aims stated, so the Statistical Analysis section could be divided in three sections, and instead of the link to figures, the author could state which variables were analysed in a specific manner) to be easier to comprehend the details. Also, I found on one place that the sample consisted of 572 reviews, and on other that it was 558.

These are very good points. I re-wrote the statistical analysis for clarity (Page 7 of the manuscript). The 558 reviews was a mistake from my part, as I forgot to include the fourth review for the 14 papers that received four reviews in the histograms of Fig. 2b and the accompanying text. This has been updated.

For figures 1a and 1b it could be considered to enter the table instead of several figures.

I thank the reviewer for pointing this out. I tried this suggestion, but I found it to reduce the readability of the paper. As an alternative, I now provide an Excel spreadsheet with all the raw data, so people can find all the characteristics of the included papers.

99.8% of the reviews analysed were assessed as polite. This is, in my opinion, extremely important finding, which shows that reviewers are still holding to certain degree of standards in communication, and it can be mentioned in the abstract.

I very much agree with this reviewer; this has now been added to the Abstract.

In results you state that QS World Ranking is "imperfect" measure. When stating that in the results section, it poses the question why it is used in the study, so maybe it is more suitable for the discussion.

This point is well taken. Even though the QS World Ranking score is imperfect, I still think it can be useful, as a rough proxy of perceived prestige of an institution. I now removed this “imperfect measure” statement from the Results section, and moved it to the Discussion (Page 5).

In the Results section, instead of using only p values, please add measures of effect (correlations, mean differences), to make it easier to place in the context.

For the significant effects of Fig. 4, I have added these to the figure legends. Please note that the used statistical tests are non-parametric, so I reported the Hodges-Lehmann differences (which is the median of all possible pairwise differences between observations from the two groups).

I think the results interpretation should be softened a bit, or the limitations of the study should be placed as the second paragraph in the discussion, since this was only specific journal with specific subfield.

I agree with this reviewer that the relatively small sample size of this paper demands more careful wording. Throughout the manuscript, I have toned down claims, and emphasized the “proof of concept” nature of this study (for example in the Abstract). I also moved the limitations section to the second paragraph of the Discussion, and elaborate more on the study’s caveats.

Methods:

The measure Review time was assessed from submission to acceptance, but this does not need to be review time since it takes a lot of time sometimes to find reviewers. that needs to be stated as the limitation.

This point is well taken. I changed this to “Paper acceptance time” in Fig. 3 and the accompanying text.

Gender name determination methods differed between the assessment of the first authors and the last authors, and that needs stronger explanation.

I appreciate this reviewer raising this point, which has also been raised by Reviewer #3. For this paper, I have carefully weighed the pros and cons of automated versus manual gender determination. Initially, my intention was to rely only on a programmatic method to identify authors' names. However, I came to realize that there were inaccuracies in senior author gender predictions made by ChatGPT/Genderize. This was evident to me due to my personal familiarity with some of these authors, either because they are famous or through personal interactions. It seemed problematic to me to proceed with this analysis knowing that these misclassifications would introduce unnecessary variability to the dataset.

The advantage of the relatively small sample size in this study was the opportunity to manually perform this task, rather than being fully dependent on algorithms. While I attempted manual gender identification for the first author as well, this was way more challenging due to their limited online presence. The discrepancy in gender identification accuracy between first and senior authors did not go unnoticed, and I acknowledge the issue it presents. I also recognize that, unlike senior authors, reviewers may not necessarily be familiar with the first authors of the papers they evaluate, as indicated in the original submission of this paper. In light of this, I sought input from several PIs who often serve as reviewers. Their feedback confirmed that they typically possess knowledge of senior authors' identities, for example through conferences, whereas the same is not true for first authors. Yet, this may be different for other scientific disciplines, where the pool of reviewers might be bigger.

Notably, for future studies I may make a different decision, especially when I use larger datasets that require me to automate the process.

I also realize that my rationale for the different methods of gender determination was not explained well enough in the original submission; I now explain my reasoning more elaborately on Page 7 on the manuscript.

For sentiment analysis: Please state based on what the GPT made a decision? Which program? (e.g. for gender it used genderize.io)

This has been added to Page 7.

Finally, your entire analysis can be made reproducible (since everything is publicly available). You can share ChatGPT chats as online materials with variables entered with the dataset analysed and the code. This would increase the credibility of the findings.

I will make the entire raw dataset available through the eLife website, including all reviews and their scores.

Reviewer #2 (Public review)

Strengths include:

1) Given the variability in responses from ChatGPT, the author pooled two scores for each review and demonstrated significant correlation between these two iterations. He confirmed also reasonable scoring by manipulating reviews. Finally, he compared a small subset (7 papers) to human scorers and again demonstrated correlation with sentiment and politeness.

2) The figures are consistently well presented and informative. Figure 2C nicely plots the scores with example reviews. The supplementary data are also thoughtful and include combination of first/last author genders. It is interesting that first author female last author male has the lowest score.

3) A series of detailed analysis including breaking down reviews by subfield (interesting to see the wide range of reviewer sentiment/politeness scores in computational papers), institution, and author's name and inferred gender using Genderize. The author suggests that peer review to blind the reviewers to authors' gender may be helpful to mitigating the impoliteness seen.

Thank you.

Weaknesses include:

1) This study does not utilize any of the wide range of Natural Language Processing (NLP) sentiment analysis tools. While the author did have a small subset reviewed by human scorers, the paper would be strengthened by examining all the reviews systematically using some of the freely available tools (for example, many resources are available through Hugging Face [https:// huggingface.co/blog/sentiment-analysis-python ]). These methods have been used in previous examinations of review text analysis (Luo et al. 2022. Quantitative Science Studies 2:1271-1295). Why use ChatGPT rather than these older validated methods? How does ChatGPT compare to these established methods? See also: colab.research.google.com/drive/ 1ZzEe1lqsZIwhiSv1IkMZdOtjPTSTlKwB?usp=sharing

This was a great recommendation by this reviewer, and I have tested ChatGPT against TextBlob and VADER, the two algorithms also used by the Luo et al. study — see Supplementary Fig. 4. Perhaps unsurprisingly, these algorithms performed very poorly at scoring sentiment of the reviews. Please note that I also tested these two algorithms at scoring individual sentences, Tweets and Amazon reviews, which it did very well (i.e., the software package was working correctly). Thus, ChatGPT is better at scoring scientific texts than TextBlob and VADER, likely because these algorithms struggle with finding where in the review the sentiment is conveyed. I now discuss this on Pages 1, 3 and 4 of the manuscript.

2) The author's claim in the last paragraph that his study is proof of concept for NLP to analyze peer review fails to take into account the array of literature already done in this domain. The statement in the introduction that past reports (only three citations) have been limited to small dataset sizes is untrue (Ghosal et al. 2022. PLoS One 17:e0259238 contains over 1000 peer review documents, including sentiment analysis) and reflects a lack of review on the topic before examining this question.

I thank this reviewer for pointing me to this very useful study. I regret missing this one in my initial submission; I now discuss this paper in Pages 1 and 5 of the manuscript.

3) The author acknowledges the limitation that only papers under neuroscience were evaluated. Why not scale this method up to other fields within Nature Communications? Cross-field analysis of the features of interest would examine if these biases are present in other domains.

I share this reviewer’s opinion that it would be very interesting to expand this analysis to different subfields. I initially only included Neuroscience papers, because I was uncertain whether I would be able to properly assess the reviews from different scientific disciplines (and thus judge whether ChatGPT was able to provide plausible scores). The different reviewers have provide me with a list of potential follow-up experiments, and I am currently considering different options for future work, including expanding into different fields within Nature Communications. Additionally, I am looking to team up with a journal to perform the experiments laid out in (the new) Supplementary Fig. 6 of the new paper, to study whether I can find evidence of bias across rejected and accepted manuscript papers of a journal. I am also looking into ways to automate data collection using APIs, and by utilizing the rapidly expanding databases for transparent peer review. Yet, based on this preprint, I have received messages from academics that are interested in using generative AI to study scientific texts. By revising this manuscript now, I hope to provide them with the tools to concurrently expand the analysis of peer review into different scientific disciplines and journals.

The comments I received from the different reviewers made me realize that I did not describe the intent of this paper well enough in the original submission. I rewrote much of the Abstract, to emphasize the proof-of-concept nature of this study, and rewrote the Discussion to focus more on the limitations of the study.

Reviewer #3 (Public review)

Strengths:

On the positive side, I thought the use of ChatGPT to score the sentiment of text was novel and interesting, and I was largely convinced by the parts of the methods which illustrate that the AI provides broadly similar sentiment and politeness scores to humans who were asked to rank a sub-set of the reviews. The paper is mostly clear and well-written, and tackles a question of importance and broad interest (i.e. the potential for bias in the peer review process, and the objectivity of peer review).

Thank you.

Weaknesses:

The sample size and scope of the paper are a bit limited, and I have written a long list of recommendations/critiques covering diverse aspects including statistical/inferential issues, missing references, and suggestions for other material that could be included that would greatly increase the usefulness of the paper. A major limitation is that the paper focuses on published papers, and thus is a biased sample of all the reviews that were written, which prevents the paper properly answering the questions that it sets out to answer (e.g. is peer review repeatable, fair and objective).

I very much appreciate this reviewer taking the time to provide me with such a detailed list of recommendations. Below, I will respond to this list in a point-by-point manner.

Reviewer #3 (Recommendations to author)

My main issues with the paper are that it is not very ambitious, and gave me the impression the aim was to write the first paper using ChatGPT to address this question, rather than to conduct the most thorough and informative investigation that would have been feasible (many obvious questions that could be addressed are not tackled, since the sample size is small and restricted). There are also issues with selection bias, and the statistical analysis, that have possibly led to erroneous inferences and greatly limit what conclusions can be drawn from the analysis. I hope my comments of use in further improving the paper.

The repeatability of ChatGPT when calculating the two linguistic characteristics is low. Taking the average of multiple assessments is one way to deal with this. To verify that taking the average of, say, 5 scores gives a repeatable score, the author could consider calculating 10 scores for a set of 20-30 reviews, calculating two scores for each review using the first 5 and second 5 ChatGPT ratings, and then calculating repeatability across the 20-30 reviews. It is important to demonstrate that ChatGPT is sufficiently repeatable for this new method to be useful.<br /> Also, it might be possible to automate this process a bit to save time - e.g. the author could change the ChatGPT prompt, like "please rate the politeness of this review from -100 to +100, do it 10 times independently, and print your 10 ratings as well as their average". Hopefully the AI is smart enough to provide 10 independently-computed ratings this way, saving the need to copypaste the prompt into the chat box 10 times per review.

This was a great recommendation by this reviewer, and a point also raised by Reviewer #1. Based on their suggestion, I looked into how each additional iteration of scoring would reduce the variability of scoring for a subset of papers (thus being able to advice users on an optimal number of iterations). I also tested this Reviewer’s suggestion to ask ChatGPT to score many times, and give separate scores for each iteration — this worked very well.

Interestingly, I observed that ChatGPT has become significantly more reliable in providing sentiment and politeness scores in recent versions. For the latest version (ChatGPT Aug 3, 2023), R2 = 0.992 for sentiment and R2 = 0.859 for politeness were reached for two subsequent iterations of scoring. Unfortunately, OpenAI does not allow access to previous version of ChatGPT, so the current dataset could not be re-scored. Yet, based on these data, there may no longer be a need for people to perform repeated scoring. I show these data in Supplementary Fig. 2, as I believe this is very useful information for people who are interested in using this tool.

To my mind, the main reason to use an AI instead of one or more human readers to rank the sentiment/politeness of peer reviews is to save time, and thereby allow this study to have a larger sample size than would be feasible using human readers. With this in mind, why did you choose to download only 200 papers, all from the discipline of Neuroscience, and only from Nature Communications? It seems like it would be relatively easy to download papers from many more journals, fields of research, or time periods if using AI-based methods, and in fact it would have been feasible (though fairly laborious) for one person to read and classify the sentiment of the reviews for 200 papers.

As well as providing more precise estimates of the parameters you are interested in (e.g. the consistency of reviews, and the size of the difference in reviewer sentiment between author genders), expanding the sample beyond this small set of papers would allow you to address other interesting questions. For example, you could ask whether the patterns observed for neuroscience are similar to those in other research disciplines, whether Nature Comms is representative of all journals (given there are other journals with public reviews), and you could test whether the male-female differences have become greater or smaller over time (e.g. by comparing the male-female differences observed in the past to the effect size observed in 2022-23). Additionally, the main analyses in this paper would have higher statistical power - for example, you only include 53 papers with a female senior author, giving you quite low power/ precision to estimate the gender difference in the average sentiment of reviews (given the high variance in sentiment between papers).

I want to thank this reviewer for taking the time about possible ways to increase the impact of this work. I agree, these are all great suggestions, and there are many possibilities to apply ChatGPTbased natural language processing to scientific peer review. Respectfully, I chose to continue with publishing this work in the form of a proof-of-concept paper, because I currently do not have the resources to perform this (quite labor intensive) study. Below I will explain my reasoning, that I also shared with Reviewers #1 and #2.

I initially only included Neuroscience papers, because I was uncertain whether I would be able to properly assess the reviews from different scientific disciplines (and thus judge whether ChatGPT was able to provide plausible scores). The different reviewers have provide me with a list of potential follow-up experiments, and I am currently considering different options for future work, including expanding into different fields within Nature Communications. Additionally, I am looking to team up with a journal to perform the experiments laid out in (the new) Supplementary Fig. 6 of the new paper, to study whether I can find evidence of bias across rejected and accepted manuscript papers of a journal. I am also looking into ways to automate data collection using APIs, and by utilizing the rapidly expanding databases for transparent peer review. Yet, based on this preprint, I have received messages from academics that are interested in using generative AI to study scientific texts. By revising this manuscript now, I hope to provide them with the tools to concurrently expand the analysis of peer review into different scientific disciplines and journals. The comments I received from the different reviewers made me realize that I did not describe the intent of this paper well enough in the original submission. I rewrote much of the Abstract, to emphasize the proof-of-concept nature of this study, and rewrote the Discussion to focus more on the limitations of the study.

Also, if you could include some reviews of papers that were reviewed double-blind, you could test whether the gender-related differences in peer reviews are ameliorated by double-blind reviewing. Nature Comms (and many other journals with open review) do have some double-blinded papers, and there is evidence that that double-blinding is preferentially selected by authors who think they will experience discrimination in the peer review process (DOI: 10.1186/s41073-018-0049-z), and also that double-blinding does ameliorate bias (DOI: 10.1111/1365-2435.14259), so this seems very relevant to the ideas under study here.

I note that the PLOS journals allow open peer review, and there is an API for PLOS which one can use to download the reviews for a given paper (e.g. try this query to get to the XML file of a paper which has open peer review: http://journals.plos.org/plosone/article/file?id=10.1371/ journal.pone.0239518&type=manuscript). Using an API could allow this project to be scaled up, because you can programmatically search for the papers with open reviews, download those reviews using the API and some code, and then score them using the same ChatGPT-based methods used for Nature Comms. Also, Publons recently merged with Web of Science (Clarivate), and you can now read all the open peer reviews on Web of Science for papers which had open review (e.g. for this paper: https://www-webofscience-com.napier.idm.oclc.org/wos/woscc/fullrecord/WOS:000615934800001). It would be possible to write to Web of Science, request access to their data or search engine, and programmatically download many thousands of papers and their associated reviews, and then use ChatGPT or a similar AI to score them all (especially if you can pass the reviews to ChatGPT for scoring programmatically, instead of manually copy-pasting the reviews into the chat box one at a time as it appears was done in the present study).

These are great suggestions, and I have different plans for follow-up studies, including the use of APIs to download large batches of peer reviews. The analyses in this paper have been performed in February of this year, even before the ChatGPT API had been released, which did not let me automate the process at that time. As a result, these analyses have been performed manually. I realize that the field is moving rapidly, and that there are now different options to scale this up quickly.

I plan on using the suggestions from this Reviewer for follow-up experiment in a next paper, and publish this revision as a proof-of-concept paper. In this way, different researchers can optimally use ChatGPT-based sentiment analyses for similar studies without a delay.

As you acknowledge, there is a selection bias in this study, since you only include papers that were ultimately published in Nature Comms (missing reviews of papers that were rejected). This is a really big limitation on the usefulness of some of your analyses. For example, you found no relationship between author institutional prestige and reviewer sentiment. This could be evidence of a fair and impartial review process (which seems unlikely!), or it could be a direct result of selection bias (specifically a "collider bias", like the famous example involving height and skill among professional basketball players). The likelihood that a paper is published is positively related both to its quality and the prestige held by the authors, we might expect a flatter (or even negative) correlation between prestige and reviewer sentiment among papers that were published than among the whole set of papers (like how the correlation between height and speed/skill is less positive among NBA players than among the general population, since both height and speed/skill provide advantages in basketball).

I agree with this reviewer that the selection bias is a major limitation of this study. I rewrote much of the Abstract and Discussion to tone down claims, and more prominently discuss the limitations of this study. I also made several suggestions for follow-up experiments.

In the section "Consistency across reviewers", you write that there was little similarity between review sentiment scores from different reviewers from the same paper, and then write "This surprising result indicates high levels of disagreement between the reviewers' favorability of a paper, suggesting that the peer review process is subjective." However I disagree with this conclusion for three reasons:

• Firstly, your dataset only includes papers that were published, and thus there is a selection bias against manuscripts where both/all reviewers disliked the paper - the removal of this (probably large) set of reviews will add a (potentially very strong) downward bias to your estimate of how consistent the review process is (since you are missing all those papers where the reviewers agreed). I think that one cannot properly answer the question "are reviewers consistent in their appraisals" without having access to papers that were rejected as well as those that were accepted.

I agree with this reviewer that there is a selection bias in this study, which I acknowledged throughout the initial submission of this manuscript. Indeed, having access to reviews of rejected papers will greatly increase my confidence in this finding. However, if there is consistency across reviewers in the entire pool of (post-review rejected+accepted) manuscripts, some of that has to trickle down into the pool of accepted papers. The correlation between sentiment scores of the different reviewers is so strikingly low (or even absent) that I simply cannot envision a way in which there is consistency across reviewers in the pre-editioral decision stage. Yet, I realize that this point is debatable. Therefore, I changed the phrasing of the Discussion section, including the following sentence:

That being said, the extremely low (or even absent) relation between how different reviewers scored the same paper was striking, at least to this author.

• Secondly, the method used to assess whether the reviews for each paper tend to be similar (shown in Figure 3b) does not fully utilize the information contained in the data and could be replaced with another method. (In the paper 3 univariate regressions compare the sentiment scores for R1 vs R2, R1 vs R3, and R2 vs R3, which needlessly splits up the data in the case of papers with more than 2 reviewers, reducing power.) You could instead calculate the intraclass correlation coefficient (aka 'repeatability'), to determine what proportion of the variance in sentiment scores is between vs within papers (I suggest using the excellent R package rptR for this). Note that the sentiment scores are not normally distributed, and so regular regression (as you used) or one-way ANOVA (which you might be tempted to use for the ICC calculation) are not ideal - consider using a GLM or transformation (the rptR package automates the tricky calculation of repeatability for generalized models).

I thank this reviewer for pointing me towards this option. I added this analysis to Fig. 3b, which confirmed the inconsistency in sentiment scores for reviews of the same paper (ICC = 0.055). As suggested by this reviewer, I decided to perform the ICC on log-transformed data, as ICC calculation is very sensitive to non-normally distributed data.

• Thirdly, an alternative and very plausible hypothesis for this lack of similarity (besides peer review being highly subjective) is that ChatGPT is estimating the "true sentiment" of a review (i.e. what the reviewer intended to say) with some amount of error (e.g. due to limitations/biases in the AI, or reviewers struggling to make themselves understood due to issues such as writing in a second language, typos, or writing under time pressure), which dilutes the similarly in the estimated sentiment of the reviews. In other words, if the true sentiment values are strongly correlated, but there is random error in how those values are estimated by ChatGPT, then the correlation between reviewer scores for each paper will tend to zero as the error tends to infinity. Furthermore a nebulous quality like "sentiment" cannot be fully summarised in a single variable running from -100 to +100, and if you had used a more multi-dimensional classification system for the reviews (or qualitative assessment by human readers) you might have found that there is a bit more correspondence (I'm speculating here, but I think you cannot really exclude this and the paper doesn't mention this limitation).

This point is well taken. I added caveats to the Discussion section on Page 5. Altogether, after taking these caveats into account, I do believe that this analysis convincingly demonstrates subjectivity in the peer review of this subset of papers. That said, I hope that my re-written discussion and additional analysis have added the necessary nuance to this point.

In Figure 3C, you write "Contribution of paper scores to review time". This strongly implies to the reader that the sentiment scores inferred for the reviews have a causal effect on the review time. This is imprecise writing (since the scores were calculated by you after the papers were published, and thus cannot be causal - you mean that the actual reviews affected the review time, not the scores), but more importantly you cannot infer any causality here since your dataset is observational/correlational. You could fix this by re-phrasing to emphasise this, e.g. "Statistical associations between paper scores and review time".

This is a very good point raised by this reviewer. I have corrected the phrasing so it no longer implies causality.

For the analysis shown in Figure 4d and Figure 4e, I am not certain what you mean by "data split per lowest/median/highest sentiment score". This is ambiguous, and I am also not sure what the purpose of this analysis is or what it shows - I suggest re-writing for greater clarity (and ideally providing the code used in all your analyses) and perhaps revising the analysis. Additionally, an important missing piece of information from this analysis (and most analyses in the paper) is the effect size. For example, you don't report what is the difference in politeness score and sentiment score between male and female authors, and what is the SE and 95% CIs for this difference. From eyeballing the figure, it looks like the difference in politeness is about 4 points on your 200point scale - this is small in absolute terms, but might be quite large in relative terms given that "politeness score" usually hovered around a small part of the full 200-point scale. What is this as a standardised effect size (i.e. in terms of standard deviations, as captured by effect sizes like Cohen's d and Hedges' g)? Calculating this (and its 95% CIs) would allow you to say whether the difference between genders is a "big effect", and give an idea of your confidence in your effect size estimate and any inferences drawn from it. You even discuss the effect size in your discussion, so it would help to calculate the standardised effect size. If you're not familiar with effect size and why it's useful, I found this paper very instructive: https://onlinelibrary.wiley.com/ doi/abs/10.1111/j.1469-185X.2007.00027.x

I agree with this reviewer that this phrasing was ambiguous. I now rephrased this on Page 4 of the manuscript:

To study whether these more impolite reviews for female first authors were due to an overall lower politeness score, or due to one or some of the reviewers being more impolite, I split the reviews for each paper by its lowest/median/highest politeness score. I observed that the lower politeness scores for first authors with a female name was driven by significantly lower low and median scores (Fig. 4d, bottom panel). Thus, the least polite reviews a paper received were even more impolite for papers with a female first author.

I also added effect sizes of the significant effects from Fig. 4 to its figure legend. Please note that the used statistical tests are non-parametric, so I reported the Hodges-Lehmann differences (which is the median of all possible pairwise differences between observations from the two groups).

"Double-blind peer review has been debated before, but has come under scrutiny for various reasons" - this is vague and unhelpful. I think it's worthwhile to properly engage with the debate and the substantial body of evidence in your paper, given your main focus is on potential bias in the review process based on authors' identities (e.g. gender, institutional prestige).

I thank the reviewer for pointing this out. I rephrased this sentence to indicate that there is evidence that it helps to remove certain forms of bias (Page 5):

To address this issue, double-blind peer review, where the authors' names are anonymized, could be implemented. Evidence suggests that this is useful in removing certain forms of bias from reviewing8,9, but has thus far not been widely implemented, perhaps because some studies have cast doubt on its merits21,22.

I have also added a Supplementary Fig. 6 to this paper, in which I lay out how my tool can be used to study bias by applying it to single- and double-blinded reviews (see also my answer to the other question about this topic below).

On a related note, in the first paragraph, when discussing the potential of single-blind review to allow reviewers to essentially discriminate against papers by women, there is a key missing citation. This year, the first truly experimental test of this hypothesis was published (DOI: 10.1111/1365-2435.14259); a journal conducted a randomised controlled trial in which submitted manuscripts were reviewed either single- or double-blind. They found no effect of author gender on reviewer ratings or editorial decisions (though there was an effect of review type on success rate of authors from different countries). It would be better to cite this instead of reference 6, which as you acknowledge is methodologically flawed. This paper is also worth a read given your focus on Nature journals: DOI: 10.1186/s41073-018-0049-z.

This point is well taken. I now cite this paper (citation #8) and rephrased this part of the Introduction (Page 1).

"Another - arguably more simple - solution [compared to double-blind peer review] could be for reviewers to be more mindful of their language use." Here, you seem to be saying that we don't need to blind author names during peer reviewers, because it would simpler if all reviewers were simply nicer! I object to this because A) double-blind review is easy to implement, and greatly reduces the opportunity to tune the review to the author's identity (and there is some experimental evidence that it works in this regard), and B) it seems like wishful thinking to say that we don't need to implement measures that reduce the scope for bias, because all reviewers could instead stop using impolite language.

This is a very valuable comment. I rephrased this to emphasize that this is an additional measure.

"reviewers may want to use ChatGPT to extract a politeness score for their review before submitting" Yes, that's an interesting idea, and I can imagine that some (probably small) proportion of reviewers will be interested in doing this. But I think you should think bigger about wholesale changes to the review system that are possible because of AI like ChatGPT. For example, the submission platforms where reviewers submit their reviewers (e.g. ScholarOne, Manuscript Central) could be updated to use AI to pre-screen draft reviews, and issue a warning to reviewers, like "Our AI assistant has indicated that the writing in this review might be impolite (example phrases here) - would you like to edit your review before you submit it?" Also, reviewcredit platforms like Publons could display not only the number of reviews that someone wrote, but an AI-generated assessment of how constructive, detailed, and polite their reviews are (this would help nudge people into writing better reviews, and also give credit where it's due to careful reviewers, which is part of the aim of Publons and similar platforms). This is just off the top of my head - there are many other good ideas about how AI could transform the peer review process. Indeed, AI is already good enough to generate quite useful peer reviews and constructive criticism of draft papers, and will surely get better at this... this surely has lots of implications for science publishing over the coming decades.

These are great suggestions for implementation of this tool. I now end the first paragraph of the Discussion (Page 4) with the following sentence:

Such an automated language analysis of peer reviews can be used in different ways, such as afterthe-fact analyses (as has been done here), providing writing support for reviewers (for example by implementation in the journal submission portal), or by helping editors pick the best papers or most constructive reviewers.

"Further research is required to investigate the reasons behind this effect and to identify in what level of the academic system these differences emerge." Here you could mention what this research would be - I think you'd need the full sample of reviewed papers, not just those that were accepted. Spell out what analyses would be required to test and falsify the various (very plausible and interesting) competing hypotheses that you mention for the male-female difference in sentiment scores.

Great point. I added a Supplementary Fig. 6, in which I show a visual depiction of the experiments that can be performed to answer these questions.

"areas of concern were discovered within the academic publishing system that require immediate attention. One such area is the inconsistency between the reviews of the same paper, highlighting the need for greater standardization in the peer review process." I disagree here. I think it is natural for there to sometimes be differences in how two or more reviewers rate the quality of a paper, even if the peer review process were carefully standardised (e.g. via the use of a detailed "peer review form", which helps guide reviewers to comment on all important aspects of the paper - some journals use these). This is because reviewers differ in their experience, expertise, or interests, and so some reviewers will catch mistakes that others miss, or request stylistic changes that others would not. More broadly, it's often not possible to write a version of the paper that satisfies all possible reviewers.

I re-phrased part of the Discussion on Page 5 to indicate other sources of inter-reviewer variability. Specifically, I mention that some variability in sentiment can be expected based on the different backgrounds of the reviewers:

Notably, some level of variability may be expected, for example due to different backgrounds, experiences, and biases of the reviewers. In addition, ChatGPT may not always reliably assess a reviews sentiment, adding some spurious inter-reviewer variability.

Yet, as also mentioned in my response to one of the previous questions, I still find the the extremely low levels of consistency striking, even after taking these possible sources of interreviewer variability into account.

"the maximum score an institution could receive was 100 (in 2023 this was Massachusetts Institute of Technology)" - this seems unnecessary information (just mention the score runs from 0-100).

I agree with this reviewer that this was unnecessary information. This has been removed.

"reviewers are generally familiar with the senior author of papers they review and thus are likely aware of their gender identity." This seems like a strong assumption, and you don't provide any evidence for it Speaking personally, as a reviewer and journal editor I am often not familiar with the senior author, or I am familiar with the first author - I am not sure how often I know the senior author but not the first author or vice versa. It's also not always the case that the first author is a junior scientist and the last author a senior, famous one, as you imply. I suggest that you use the same approach to score the gender of both author positions, namely inferring their gender programmatically from their name (I agree that generally the important thing for the purposes of this study is the gender that reviewers will infer from the name, not the author's actual gender, and so gender estimation from first names is the correct approach).

I appreciate this reviewer raising this point, and I have carefully weighed the pros and cons of both approaches. Initially, my intention was to rely only on a programmatic method to identify authors' names. However, I came to realize that there were inaccuracies in senior author gender predictions made by ChatGPT/Genderize. This was evident to me due to my personal familiarity with some of these authors, either because they are famous or through personal interactions. It seemed problematic to me to proceed with this analysis knowing that these misclassifications would introduce unnecessary variability to the dataset.

The advantage of the relatively small sample size in this study was the opportunity to manually perform this task, rather than being fully dependent on algorithms. While I attempted manual gender identification for the first author as well, this was way more challenging due to their limited online presence. The discrepancy in gender identification accuracy between first and senior authors did not go unnoticed, and I acknowledge the issue it presents. I also recognize that, unlike senior authors, reviewers may not necessarily be familiar with the first authors of the papers they evaluate, as indicated in the original submission of this paper. In light of this, I sought input from several PIs who often serve as reviewers. Their feedback confirmed that they typically possess knowledge of senior authors' identities, for example through conferences, whereas the same is not true for first authors. Yet, this may be different for other scientific disciplines, where the pool of reviewers might be bigger.

Notably, for future studies I may make a different decision, especially when I use larger datasets that require me to automate the process. I now more elaborately explain why I made this decision on Page 7 of the manuscript.

In the Abstract, you write "suggesting a gender disparity in academic publishing". This part of the sentence contains no information about what you think is the cause of the male/female difference, and no further interpretation of its ramifications, so I think you can just remove it (because "disparity" just means a difference, so you are effectively saying something redundant like "there was a difference between papers with male and female senior authors, suggesting there is a difference")

I thank the reviewer for pointing this out. I replaced the latter part of this sentence with “(…) for which I discuss potential causes.”, which I think is better than a short summary of potential causes which may lack the nuance that such a topic deserves.

Author Response

The following is the authors’ response to the original reviews.

First of all, we would like to again thank the reviewers for their work. We appreciate the constructive review comments and useful suggestions to further improve our article. With those comments in mind, we have now revised our manuscript. Please see below for a point-by-point response (our responses in green) to all comments.

Reviewer #1 (Recommendations For The Authors):

Sun and colleagues outline structural and mechanistic studies of the bacterial adhesin PrgB, an atypical microbial cell surface-anchored polypeptide that binds DNA. The manuscript includes a crystal structure of the Ig-like domains of PrgB, cryo-EM structures of the majority of the intact polypeptide in DNA-bound and free forms, and an assessment of the phenotypes of E. faecalis strains expressing various PrgB mutants.

Generally, the study has been conducted with a good level of rigor, and there is consistency in the findings. However, I do have some specific technical concerns relating to the study that necessitate the undertaking of additional experiments. These are summarized as follows:

1) Recombinant PrgB188-1233 produced in the study purifies as a mixture of monomeric and dimeric species separatable by SEC. There is very limited discussion in the text re. the significance and/or implications of this. Is it feasible that the dimeric form is biologically relevant in the context of the in vivo situation? Or alternatively, is this simply an artifact of protein production?

Experimental data that we published in 2018 indeed indicates that the dimer is relevant in the in vivo situation. We did not discuss this here since this was discussed in detail in the previous paper: Schmitt et al, 2018. We have now added a bit more information on this in the results section, highlighting this, so that it is clearer to the reader (lines 114-116).

2) The authors see no evidence of the adhesive domain of PrgB in their PX structure highlighting that this must have been cleaved during crystallisation. Is this claim supported by an inspection of the crystal packing? It could be that this region of the protein is dynamic within the context of the crystal and is thus not observed. This should be clarified in the text either way.

The crystal packing does not provide any space for the PAD. We have added this to the results section. We have added a sentence describing this in lines 122-124.

3) The Cryo-EM structures reported are both at ~10-angstrom resolution. Are the authors truly confident in the placement of their crystal structures on these maps? Visual inspection indicates that their positioning of the PrgB domains into the EM envelopes is somewhat questionable. The authors need to provide some quantitative measures of the quality of their domain fitting. The narrative of the manuscript very much hinges on this being correct.

This is something that the other reviewer also commented on. The fitting of the crystal structures in the maps are indeed not optimal, but was the best we could do with the available data. In line with point #6, we have now constructed new protein variants of the stalk domain (the four Ig-like domains) alone, and have assayed it’s interaction with the PAD in vitro using native gels and size exclusion chromatography. The outcome of these experiments is that the two domains do not interact in any substantial way on their own. Thus, the added experiments do not support the hypothesis that the PAD interacts with the Ig-like domains, at least not without the local high concentration provided by the linker region in the in vivo situation.

To account for these new experiments, we have moved the cryo-EM structure to the supplement, and rewritten this part of the manuscript to say that the cryo-EM data indicated that there might be an interaction, but that we have not been able to verify this in vitro, indicating that if the interaction at all exists it must have a low affinity and is likely not physiologically relevant. In line with this, we have also further modified the text throughout the manuscript to account for this.

4) The manuscript would be significantly strengthened if the authors could include confirmatory hydrodynamic data in support of the observed conformational reorganization of PrgB in the presence of DNA. SAXS analysis of the DNA-free and bound complexes would be ideal for this and would also help address the issues raised above in pt 3.

To analyze PrgB radius with and without DNA, we tried both SEC-MALS and DLS experiments. It proved difficult to obtain precise and reproducible values, but the initial data indicated that no large changes were observed upon DNA binding. As we could also not measure specific interaction between the PAD and the stalk in vitro, we did not perform SAXS experiments. As mentioned in the response to point #3, we have modified the results and discussion regarding the potential interaction of th PAD and Stalk domains.

5) The authors present binding studies of various PrgB mutant-expressing strains. A number of the mutations generated delete significant portions of the polypeptide. Can the authors confirm that these mutant proteins are correctly folded despite the introduced mutations? It could be that loss of function is simply a consequence of mutation-induced misfolding. I would like to see some confirmatory data (CD, SEC, etc.) in support of the foldedness of the mutant proteins.

We cannot completely rule out that the folding of some of the variants is affected in E. faecalis. However, CD or SEC experiments would only give indications of the contrary if the overall fold had been majorly affected in an in vitro situation where the protein is not anchored to the E. faecalis cell wall.

To alleviate this valid concern, we probed if all variants are correctly exported and linked to the cell-wall. Therefore we have now extracted the cell wall of E. faecalis producing wild-type or variant PrgB and performed Western blot . The results of the Western blot with cell wall extract largely matches the whole cell experiments that were in the initial manuscript. If a protein variant was largely misfolded, it would likely not be targeted and linked to the cell-wall, nor would it be stable in vivo. We have added this new data as a new fig 3 – figure supplement 1 and on lines 201-214

6) The authors suggest a direct interaction between the PAD and the stalk domains in PrgB. The discussion of this is very generic and no evidence to support this is provided other than the 10-angstrom resolution EM map. If they believe this to be the case, then additional evidence should be provided.

Answer: As mentioned previously, we have now performed additional in vitro experiments to probe this potential interaction, but conclude that this indication from the EM data is likely not a real high affinity interaction. In line with this, we have modified the results and discussion regarding this point, see also response to point #3 and 4.

---

Reviewer #2 (Recommendations For The Authors):

As currently presented, I don't feel that the cryoEM data support the authors' proposed model, largely because the fit of the crystal structures to the EM volumes does not seem entirely reasonable for the apo- dataset and because the EM volume for the ssDNA bound dataset is not even contiguous. For me to believe the model as it is currently built, I would want to see a dataset with the PAD deleted, showing that its proposed density disappears, or a dataset with a PAD-specific antibody as a fiducial marker. It would be nice to see some goodness of fit metric with a comparison to other crystal structures fit such low-resolution data as well. At the very least, the authors must include the standard cryoEM workflow supplementary figure showing representative micrographs, 2Ds, and 3Ds along with particle numbers.

In line with the comments raised by reviewer #1, we have now added more experiments where we have analyzed the potential interaction between PAD and the stalk domain. From this new data, it looks like they do not interact with any substantial affinity, at least not on their own without any linker region holding them together, and that this interaction if it all exist likely is not physiologically relevant. The cryo-EM data has been moved to the supplement as we agree with both reviewers that the resolution, and the fitted model, is not good enough to draw any hard conclusions. The standard table for the cryoEM workflow was present as supplementary table 2, where eg particle numbers etc are described, but we have now also added a new supplementary fig 2 – figure supplement 2 that shows the EM processing workflow, including representative micrographs, 2D and 3D classes. We debated whether we should remove the EM data, but decided against it in line of transparency and to explain why the interaction studies with the PAD and stalk domains were performed.

The X-ray crystallographic structure is very nice, but I was a bit surprised by the R factors in Table 1. After downloading the structure factors and coordinates from the PDB (thank you for depositing before submission!) I was able to see quite a few positive peaks in the difference map that could probably use some cleaning up. I realize I may just be a bit of a masochist when it comes to adding/deleting waters and moving around side chains to get things just right, but for such lovely data, I would have liked to see the model polished up a bit more. I was going to say that the isopeptide bond should be modelled, but I can see from a cursory Google that the authors did in fact try to find a way to model this and that it is indeed a bit of a pain.

The model refinement proved surprisingly recalcitrant with regards to the remaining difference density, so we took the decision to only model what was solidly there (which leads to slightly higher R factors). We did indeed try to model the isopeptide bond, but we did not find a good way to do so (despite trying quite extensively), and ended up determining them as a linker in the PDB file, so that the bond shows up when one opens the structure in eg. Pymol.

For protein production/purification in general I would have liked to see actual traces for the gel filtration and pure protein on a gel in a supplementary figure. I strongly believe that this type of information is so critical for future researchers looking to replicate or build upon published work so that they have some sense that what they are doing is working in the way it should be.

We have now added a supplementary figure (as new Fig. 1 – figure supplement 1) that shows SEC and SDS-PAGE for the purification of PrgB188-1233.

Finally, I think for the in vivo data it only makes sense to show the reader whether any or all the differences measured across your different mutants are statistically significant. Having done the graphing and analysis in GraphPad this should be a simple thing to achieve.

We have now added statistical test (One way Anova) that show the statistical significance between the mutants, and show that in Fig 3 and Fig 4.

Overall, I think it's a very nice paper and while I feel that the cryoEM data in its current form doesn't support the model of occlusion from PrgA, I also don't think that removing the cryoEM data and that specific mechanistic idea from the paper detracts from its overall message and impact.

Thank you for those comments.

Author Response

The following is the authors’ response to the original reviews.

Reviewer #1 (Recommendations For The Authors):

p. 5, l. 87-90: The control of flgM by OmrA/B (PMID 32133913) and the antisense RNA to flhD (PMID 36000733) are other examples of known regulatory RNAs that impact the flagellar regulon.

We thank the reviewer for pointing out these references and have added citations to them (page 5, lines 87-91).

p.11/Fig. 3: it is intriguing that ArcZ and RprA, two of the rpoS-activating sRNAs, repress lrhA. I realize that it is outside of the scope of this study, but have the authors considered the possibility that ArcZ or McaS could have a role in the previously reported repression of rpoS by LrhA (PMID 16621809)?

We agree that it is intriguing that ArcZ and RprA, two of the rpoS-activating sRNAs, repress lrhA, and added mention of this regulatory connection (page 12, lines 247-250).

p. 13/l. 272: I do not understand why the authors say that "r-proteins were almost exclusively found in chimeras with MotR and FliX and no other sRNAs...", given that several other chimeras between r-prot and other sRNAs are found

While some r-proteins encoding genes were found with other sRNAs in RIL-seq datasets, MotR and FliX generally had the highest numbers. The text was revised to better describe the RIL-seq data for r-proteins interaction partners (page 14, lines 291-295), and a new panel showing the S10 operon with all the interacting sRNAs was added to Figure 3—figure supplement 1B.

Fig. 4 and 5: One possible improvement would be to more systematically assess the effect of base-pairing mutants of the sRNAs, such as MotRM1 or FliXM1 on fliC and rps/rpl genes in vivo. This is especially important for the mutants that affected the sRNA effects in the in vitro probing assays, such as UhpU-M2, MotR-M1 and FliX-S-M1 on fliC (Fig. S7)

As suggested, we examined fliC mRNA levels across growth in motR-M1 and fliX-M1 chromosomal mutants. The results of these northern assays, now shown in Figure 8—figure supplement 1, are consistent with our model as we observed delayed expression of fliC mRNA in motR-M1 background and premature expression in fliX-M1 background (page 21, lines 444446, 449-453).

Fig. 5: it may be worth including a schematic of the whole S10 operon to highlight its length and its organization?

As suggested, a schematic representation of the S10 operon was added to Figure 3—figure supplement 1 with a summary of the RIL-seq data for this operon.

Probing data (Fig. 5, S7 and S9): in general, it is difficult to differentiate the thin and thick brackets, and what is indicated by the dashed brackets is not always clear. Maybe using a color-code instead could help? Highlighting the predicted pairing regions on the different gels could be useful as well.

We thank the reviewer for this suggestion and color-coded the brackets (Figure 5, Figure 4figure supplement 2, and Figure 5-figure supplement 2). The correspondences to regions of predicted pairing are described in the figures legends.

Fig. S10: The experimental evidence used to support FliX-dependent degradation of the rpsS mRNA is indirect (primer extension to observe higher levels of cleavage intermediates). It would be nice to be able to observe a decrease in the mRNA levels as well, either by Northern, or primer extension from a region more distant to the FliX pairing site.

The S10 operon is long (~5 KB). We have tried multiple probes for this mRNA and detect many bands with each, likely due to extensive regulation of this operon. We think teasing out the origin of the different bands to appropriately interpret changes in patterns will require a significant amount of work.

legend of Fig. S10: from the gel, it seems that only the plasmids differ in the samples, and it is not clear where the data corresponding to the WT strain mentioned in the legend is shown

The samples shown in this figure are all for the indicated plasmids in the WT strain. We corrected the figure legend.

Table S1: please define the NOR (normalized odds ratio?)

The definition of Normalized Odds Ratio was added to the legend of Supplementary file 1.

Reviewer #2 (Recommendations For The Authors):

Major comments:

Figure 1B. Please add a negative control (which could be in the supplementary section) from a large section showing transcripts that are not directly influenced by Hfq.

We think the flgKLO browser in this figure serves as a negative control; flgK and flgL clearly are not enriched on Hfq in contrast to FlgO. Figure 1B was generated using published datasets that are easily accessible to the readers at a genome browser and show many other examples of transcripts that are not influenced by Hfq: https://genome.ucsc.edu/cgi-bin/hgTracks?hubUrl=https://hpc.nih.gov/~NICHD- core0/storz/trackhubs/ecoli_rilseq/hub.hub.txt&hgS_loadUrlName=https://hpc.nih.gov/~NICHDcore0/storz/trackhubs/ecoli_rilseq/session.txt&hgS_doLoadUrl=submit

Line 158. MotR* is a more abundant version of [the constitutively overexpressed] MotR. Is there a Northern or qPCR to confirm this? While I understand the relevance of these mutated constructs, their high expression can lead to artefactual effects.

This is a valuable point and therefore we provided a northern blot to document the relative levels of MotR and MotR* (Figure 2—figure supplement 1A).

Figure 2. The overexpression of MotR/MotR* from a plasmid is increasing the number of flagella. However, when the MotR gene is deleted, is there a reduction of the number of flagella? Same question with FliX: what happens when the fliX gene is deleted? According to the model described in the manuscript, we should expect fewer flagella in ΔmotR background and an increased number of flagella in ΔfliX background. Both Figure 2 and Figure 8 would benefit from additional experiments with deleted motR and fliX genes.

We agree that experiments regarding the endogenous effects of endogenous sRNAs are important. We provided such data in Figure 8 and Figure 8—figure supplement 1 for MotR and FliX in a variety of assays: flagella numbers by electron microscopy, motility and competition assays, expression of flagellar genes by RT-qPCR and western analysis. The chromosomallyexpressed MotR-M1 and FliX-M1 base pairing mutants did show the expected phenotypes of reduced and increased numbers of flagella, respectively (Figure 8A-B). As suggested by reviewer 1, we added northern analysis that examined fliC mRNA levels across growth in motRM1 and fliX-M1 chromosomal mutants. The results of these northern assays are consistent with our model as we observed delayed expression of fliC mRNA in motR-M1 background and premature expression in fliX-M1 background. We went to the trouble of constructing strains carrying point mutations in the chromosomal copies of these genes rather than deletions to avoid interfering with the expression of motA and fliC given that MotR and FliX encompass the 5’ and 3’ UTRs, respectively.

Figure 3 is key to demonstrating the sRNAs pairing with their specific targets and potential effect on bacterial swimming. However, these results would be more relevant with endogenous expression of the sRNAs and demonstration of their effects on the same targets. A Northern blot showing the overproduced sRNA level compared to endogenous sRNA level could help us appreciate the expression ratio.

The levels of the UhpU, MotR and FliX expressed from the overexpression plasmids are at least 100-fold higher than the endogenous levels. Thus, we agree that assays of chromosomal deletion/point mutants are important experiments. We did construct chromosomal uhpU-M1 and uhpU∆seed sequence mutants. However, under the conditions assayed, the uhpU chromosomal mutations did not result in observable effects on motility or FlhD-SPA protein levels. It is possible we would be able to detect differences between the wild type and uhpU chromosomal mutant strains under different growth conditions or in different assays, but this would require a significant amount of work. For many other sRNA chromosomal mutations have no or only subtle effects, suggesting redundancy between sRNAs or sRNA roles in fine tuning gene expression.

Figure 4. In panel B, the empty plasmid pZE alone seems to positively affect the flagellin expression when compared to the WT background. This can also be seen in Figure 4C. There is no fliC signal with empty plasmid pBR* but a strong fliC signal with empty plasmid pZE. Maybe the authors can explain this in the manuscript.

With respect to panel B and Figure 4—figure supplement 1A, we agree that there is some variation between the levels of flagellin in the WT and pZE control samples, possibly due to the addition of antibiotic to the pZE culture. We added quantification of the bands in Figure 4— figure supplement 1 to better document the changes in flagellin levels.

With respect to panel C, the pBR samples were collected in crl+ background while the pZE samples were collected in crl- background, which explains the lack of fliC signal in the pBR control sample. This is now noted in the figure legend.

In lines 154-157, the justification for using two plasmids is described. An IPTG-inducible Plac promoter, the pBR*, is used because the constitutive overexpression of UhpU is resulting in mutated UhpU clones. These observations suggest a toxic expression level of UhpU that the cell can only tolerate when the UhpU RNA is somewhat deactivated by mutations. This does not seem like a detail and could be discussed further.

We agree with the reviewer that this observation is important and now mention that it suggests at a critical UhpU role (page 8, lines 160-163).

Figure 5E and I. While the bindings of MotR on rpsJ and Flix-S on rpsS are clear, the resolution of both gels in the areas of binding (upper part of both gels) could be improved.

We found it tricky to choose the mRNA fragments for the in vitro structure probing for the regions of predicted pairing internal to CDSs. Given that we hoped to retain native RNA folding, we chose long fragments; for rpsJ, we started with the +1 of S10 leader and for rpsS, we started 147 nt into the CDS, a region that overlaps the region that was cloned to the rpsS-rplV-gfp fusion. Consequently, the region of base pairing is in the upper part of both gels. The gels were already run for an unusually long time. Thus, we do not think the resolution could be improved further. Nevertheless, we think the region of protection is evident for both mRNAs.

Minor comments:

Fig 1B. The promoter symbols are extremely small, please increase the size.

As suggested, we have enlarged the promoter symbols in Figure 1B as well as in Figure 3A.

Line 211. "the lrhA mRNA has an unusually long 5´ UTR". How long exactly?

The 5’ UTR of the lrhA mRNA is 371 nt long. This is now mentioned in the text (page 11, line 224)

Line 320. Should "Fig 9C" be "Fig S9C" instead?

We thank the reviewer for noticing this typo. Callouts to supplementary figures have now been renumbered per eLife format.

Line 384. Something seems to be missing in the sentence "a representative combined class 2 and 3 promoter".

The sentence has been modified to clarify the designation (page 19, lines 409-411).

Reviewer #3 (Recommendations For The Authors):

Recommendation to clarify/strengthen the presentation of science in the paper:

Lines 102-103: Can the authors provide some more information on how the sRNAs were initially discovered to be potentially sigma-28 dependent and selected?

As suggested, we expanded the section discussing the discovery and the selection of these sRNAs (page 6, lines 104-109).

Lines 192-193: It would be helpful to provide a bit more information in the main text about what are the different RIL-seq data sets (18 in total).

As suggested, we now provide more details about the different RIL-seq datasets we used in the analysis (page 10, lines 202-205).

It would be helpful to specify the criteria for "top" interactions in targets retrieved from RIL-seq data (Table S1 and text, e.g., line 273): e.g. number of conditions, number of chimeras, etc.

As suggested, we now more explicitly specify the criteria for selecting targets to characterize (page 10, lines 205-206).

Fig. 4B/ S6 and line 242: The flagellin amount in the empty vector control (pZE) looks higher than in WT, and the stated effect of MotR/MotR* OE on flagellin is not very clear from the blot. The "cross-reacting band" above flagellin also seems to vary among strains. Could the authors include a quantification of flagellin protein amount and normalize relative to a housekeeping protein (e.g., GroEL), instead of Ponceau S as loading control?

We agree that there is some variation between the levels of flagellin in the WT and pZE control sample, possibly due to the addition of antibiotic to the pZE culture. We added quantification of the bands in Figure 4—figure supplement 1 to better document the changes in flagellin levels.

Figure legends: It would be helpful to have a bit more information about the method used/displayed image rather than stating results in the legends.

As suggested, we now provide a bit more information about the methods used/displayed image in the figure legends to allow for easier comprehension of the data presented in the figures (while trying to balance this with the length of the legends).

Fig. 2: Please include a scale for all electron microscopy images or, if it is the same for all panels, state it in the figure legend. Moreover, the same image is used for the pZE control in panel C, E and Figure S4A/C. It would be better to show different fields of bacteria for the pZE sample.

As is now mentioned in the legends to Figure 2, Figure 2—figure supplement 2, and Figure 8, the same scale was used for all panels. We thought it was better to show the same image for the pZE control in the different panels to emphasize that these samples were all analyzed on the same day.

Fig. 2: The sRNA OE strains seem to show some heterogeneity in cell length (pZE-MotR) or width (pZE-FliX). The authors could, e.g., check whether this is a phenotype correlated to sRNA OE by quantifying these parameters for different fields and comparing to WT or comment on this in the text if this is not consistently seen.

We also were intrigued by the slightly different sizes and widths of cells in the EM images. However, our statistical analysis did not reveal significant differences between the different samples. We now comment on this (page 53, lines 1178-1179).

As a follow-up to this study, it would be interesting to assess the impact of MotR and FliX regulation of ribosomal protein synthesis on overall ribosome activity (e.g., via Ribo-seq), also considering that antitermination regulates rRNA transcription. In the case of MotR, the authors suggest that MotR upregulation of S10 protein might not only impact antitermination, but also lead to the formation of more active ribosomes that would increase flagellar protein synthesis (lines 359-362). However, in the RNA-seq performed in OE MotR* several transcripts encoding rRNA and ribosomal proteins are significantly downregulated compared to EVC (Supplementary Table S2). Could the authors comment on this?

We share the reviewer’s enthusiasm for follow-up work and thank for the suggested experiments. We hope we will be able to decipher the full mechanism of MotR and FliX action on ribosomal protein synthesis in future experiments. The observation that some ribosomal protein-coding gene levels are reduced in the RNA-seq experiment with overexpression of MotR* is interesting but we do not have an explanation other than the fact that the samples were collected early in exponential growth. We now mention the observation in the text (page 19, lines 404-407).

Considering that OE of the WT MotR appears to increase fliC mRNA abundance but has no strong impact on flagellin protein levels, can the authors speculate what is the physiological relevance of MotR* for flagellin production?

We agree that while we do see significant increases in the flagella number and fliC mRNA abundance with MotR and MotR* overexpression, the western analysis did not reveal a striking increase in flagellin levels and also wonder how MotR strongly increases the flagella number, which requires flagellin subunits, but only has a weak effect on the intercellular levels of flagellin. One possibility explanation is that it is more difficult to see significant increases for a protein whose levels are high to begin with. These points are now discussed (page 13, lines 264-269).

Fig. 4C: The pZE samples seem to show variable expression of fliC mRNA although the samples are collected at the same timepoints. Try to clarify in the text.

The northern membrane on the bottom was exposed for a longer time due to the lower fliC mRNA levels in the samples with FliX overexpression. We now note these differences in the legends to Figure 4 and Figure 4—figure supplement 1.

Fig. 7/S13: While a volcano plot for MotR is shown in Fig. 7A, quantification of GFP reporter fusion regulation is shown for MotR. Quantifications of MotR are shown in Fig. S13. Maybe swap the figures.

Given that the data for MotR are in the supplement figures for all other figures we would also like to retain this distribution for Figure 7 (aside from the volcano plot since this experiment was only carried out for MotR).

Lines 135-136 (Fig. S1B): on the northern blots, only sRNA levels of MotR are comparable between rich and minimal media (excluding M63 G6P and M63 gal). Most other sRNA seem to be more abundantly expressed in minimal media conditions compared to LB. Maybe rephrase.

As suggested, the text was revised to point out the differences in the sRNA levels for cells grown in different growth media (page 7, lines 140-144).

Lines 229-234: this paragraph seems not directly connected to the aims of the study (i.e., no effect on motility tested of these other sRNAs) and could be removed (or moved to discussion).

We appreciate the reviewer’s suggestion but, considering Reviewer 1’s comments, think that showing the regulation of lrhA by other sRNAs has value in highlighting the complexity of the regulatory circuit. We have revised the text to incorporate Reviewer 1’s suggestions and better explain why these results are intriguing (page 12, lines 247-250).

Line 200 and Fig. S5: For FlgO sRNA only one target was identified in RIL-seq. This gene could be specified and labeled in Fig. S5 and the text. Does FlgO also bind ProQ?

We now mention the single FlgO target (gatC) detected in four datasets (page 10, lines 213215). In Figure 3—figure supplement 1, we labeled only targets that we followed up with in the current study. Therefore, to be consistent, we prefer not to label gatC in the FlgO plot. FlgO was found to co-immunoprecipitate with ProQ but at much lower levels than with Hfq, and to have very few RNA partners (Melamed et al., 2020).

Lines 493-498: It is mentioned that the four sRNAs were also detected in recent RIL-seq experiments of Salmonella and EPEC. Are any of the here identified targets also found in other species or was none detected as analyses were carried out under conditions that do not favor flagella expression?

The targets identified in this study were not detected in the Salmonella and EPEC RIL-seq datasets. However, the Salmonella and EPEC experiments were carried out under different growth conditions. Based on the sequence conservation of the Sigma 28-dependent sRNAs across several bacterial species (Figure 8—figure supplement 2), we do think overlapping targets will be found in other bacterial species under the appropriate growth conditions.

The strongest evidence of MotR dependent target regulation is the one on rpsJ, which does not necessarily require the additional experiments with MotR. Since the authors were able to show upregulation of the rpsJ-gfp reporter upon OE of MotR WT, it would have strengthened the results if they performed the experiments in Fig. S8C with MotR WT. Similary as an increase of flagella number was seen with OE of MotR WT in Fig. 2A, the effect of the OE S10∆loop could be compared to OE MotR instead of OE MotR (Fig. 6A). At least if would be helpful, to briefly comment on why MotR* was used instead of MotR WT for these experiments.

As suggested, we state MotR was used in some assays given the stronger effects for some phenotypes (page 10, lines 196-197). We think, given that we established MotR and MotR cause the same effects, with increased intensity for the latter, it is reasonable to use MotR* in some of the experiments.

p. lines 482-491 and 508-511: The authors discuss that both UhpU sRNAs and RsaG sRNA from S. aureus are derived from the 3'UTR of uhpT, but conclude there is no overlap regarding flagella regulation, suggesting independent evolution of these sRNAs. However, the authors also mention that UhpU sRNA has many additional targets beyond LhrA involved in carbon and nutrient metabolism. Thus, maybe regulation of metabolic traits could be a conserved theme and function for UhpU and RsaG? Maybe try to comment on or better connect these two parts in the discussion.

As suggested, we now comment on the possibility of the regulation of metabolic traits being a conserved theme and function for UhpU and RsaG (page 24, lines 520-527).

Check the text for consistency regarding the use of italics for gene names (e.g., legend of Figs. 7 and 8)

The text was corrected.

Please introduce abbreviations, e.g., G6P (line 139), REP (line 150), ARN (line 258), NOR/U (Table S1 legend)

As suggested, we now introduce the abbreviations for G6P (page 7, line 142), REP (page 8, lines 155-156), and NOR (Supplementary file 1 legend). Regarding ARN, these sequences are already written in parentheses in the same sentence. However, we revised this to “ARN motif sequences” (page 13, line 278).

Fig. S1A: Highlight REP sequence mentioned in text (line 150).

REP sequences are now highlighted in gray in Figure 1—figure supplement 1A.

Fig. S1C: It would be helpful to list number nt positions on the sRNAs based on full-length transcripts.

The corresponding positions based on the full-length transcripts have also been added to this figure.

Fig. S2: Adjust the position of UhpU-S label.

UhpU-S label position was adjusted.

Fig. S6: Include UhpU in the figure title.

UhpU was added to the title.

Fig. S10: It would be helpful to indicate on the figure (or state more clearly in the legend) which RNA was extracted from WT or ΔfliCX background.

The samples shown in the Figure are all in a WT strain. We corrected the figure legend accordingly.

Line 290: the effect is on flagella number, not motility.

This typo is now corrected (page 15, line 312).

Fig. S8: One-way ANOVA (panel A legend)

This typo is now corrected (page 64, line 1433).

Line 320: Fig. S9C instead of 9C

We thank the reviewer for noticing the typo. The numbering of the supplementary figures has now been changed to the eLife format.

It would be helpful to add reference for statement in line 57.

A reference to (Fitzgerald et al., 2014) was added as suggested.

Add PMID:32133913 as reference for post-transcriptional regulation of the flagellar regulon in the introduction (lines 87-91)

The indicated reference was added as suggested (page 5, lines 87-91).

Legend Fig. S6: expand view -> expanded view

This typo is now corrected (page 63, line 1406).

line 513: sRNA -> sRNAs

This typo is now corrected (page 25, line 549).

Fig. 8G: Maybe include lrhA as target of UhpU sRNA at top of the cascade.

As suggested lrhA has been added as a target of UhpU at the top of the cascade.

Author Response

The following is the authors’ response to the original reviews.

Reviewer #1 (Recommendations For The Authors):

• The improvement of the gene annotations of the ferret genome was an important part of this study, and so I would recommend that the authors have a results section and figure dedicated to documenting this.

Thank you so much for appreciating our efforts on improving gene models, which was indeed a critical part in this study. According to the reviewer’s suggestion, we added a new section to the main text, “Improvement of the gene model for scRNA-seq of ferrets” with a figure (Fig.1 C, D, E).

• Are the references to figure S8A, B alright (line 306)? In fact, that entire figure was not well described or out of place. In general, unlike the rest of the manuscript, the section dealing with the human-ferret comparison was a little bit confusing, and the figure legends were not extremely helpful. Could the authors please revisit the main text and figure legends of this section for clarity?

We agree with the reviewer’s recommendation. We removed references to Figure S8A, B. In place of that, we explained the reason more carefully; “We chose a recently published human dataset (Bhaduri et al, 2021) for comparison, because this study containing GW25 dataset which included more tRG cells than previous studies that did not contain GW25 data. Furthermore, we used only data at GW25”

We also revised several parts in this section to understand more easily by additional explanations as well as in the legends of Fig. 7 and Fig. S8.

Reviewer #2 (Recommendations For The Authors):

I have a few very minor comments on the manuscript.

• I would caution the authors against claiming that they have demonstrated bona fide generation of ependymal cells from tRG cells. While the expression of FOXJ1 is a very good indication, they have not demonstrated the morphological transformation of a tRG cell into an ependymal cell.

We agree the reviewer’s opinion. We have never thought that we proved that tRG differentiates ependymal cells, but we consider that this is highly likely the case (We use the term “suggest” in the abstract). To prove this genetically, we extensively tried to knock the EGFP gene into the CRYAB gene by the CRISPR/Cas9 method, to be able to show the lineage relationship between tRG and ependymal cells. However, we have so far failed to do this for a year trial. We also tried to just label tRG with EGFP and follow it in the slice culture.

However, we failed to keep the slice in the culture until we observed the transition from tRG shape to the ependymal shape. It seems to be a slow process. What we could do was to observe the transition from single cilia to multi-cilia, which is part of the morphological transition from epithelial neural stem cells such as Radial Glia to an ependymal-like sheet form. To prove this transition from tRG to ependymal cells (and also astrocytes) is one of the most important issue which needs some new idea, technique or strategy.

• There are several typos throughout the manuscript that I would recommend fixing for example, page 5 line 123 says "OLIGO2" instead of "OLIG2"

Thank you so much. We carefully read and corrected typos. We wish we corrected all of them.

Besides these two points, the manuscript is already prepared to a high standard.

I really appreciate reviewersʼ efforts to finish reviews in a short time, responding to our request related to the first authorʼs thesis application.

Author Response

Summary of reviewers recommendations.

Reviewer 1

Point# 1. Make a new section in the text with a figure about the improvement of the genomic information (gene modeling) of ferrets ".

Point# 2. the references to figure S8A, B alright (line 306)?

Point# 3. Revise the main text and figure legends of the section dealing with the human-ferret comparison for clarity.

Reviewer 2

Point# 4. Weaken (change the text from “conclusive” to suggestive” ) the expression that we identified that tRG become ependymal cells, because we have not demonstrated the morphological transformation of a tRG cell into an ependymal cell, which is practically difficult although we have shown morphological change in terms of the single-cilia to multi-cilia form transition (Fig. S6A).

Point# 5. Correct several typos throughout the manuscript that I would recommend fixing for example, page 5 line 123 says "OLIGO2" instead of “OLIG2.

Provisional revision plan and our responses.

Point #1 The new section for the improvement of gene models will be made by transferring the part of methods to the main text and Fig S2B,C to new Figure 1 with one schematic panel.

Point #2; We cited (Bhaduri et al., 2020) as a reference in the figure S8A , while "Bhaduri et, al, 2021” was cited in the text. Which is correct? We will correct this, by choosing the correct one. Descriptions are indeed poor regarding Fig. S8A and S8B in the text as well as in the legends.

Point #3 : We will describe the methods of comparison between ferrets and humans more thoroughly, by adding definition of words such as gene scores, subtype scores in the main text. (as well, the explanation of (Figure S3C) will be improved. ). Legends for Fig. 6 are too simple. So we would explain more in these legends. Explanations of analysis and figures, which we made, responding to the reviewer comments of “review commons” are generally not easy to understand with too short explanations, comparing with complexity of figures and contents, let’s say, Figure S8A-D. We will give more explanations for each of panel in Figure S8A-D, and E and F.

Point #4; The authors' response to this point goes like this; we totally agree that we need to genetically labeling (knocking in the Cryab gene) to prove “tRG cells differentiate ependymal cells”. We tried many times but eventually failed. We have partially show single-cilia to multi-cilia transition which is characteristic to epithelial-ependymal transition. This process appears to take a long time and therefore, morphological tracing by time-lapse imaging in tissue culture is not a realistic way, Therefore, we weakened the conclusion; it is "highly likely" that tRG cells differentiate to be ependymal cells.

Point#5: We will survey typos-> correct them, by all authors read the manuscript carefully again.

Author Response

The following is the authors’ response to the original reviews.

eLife assessment

This is a valuable investigation of the chromatin dynamics throughout the cell cycle by using fluorescence signals and patterns of GFP-PCNA and CY3-dUTP, which labels newly synthesized DNA. The authors report reduced chromatin mobility in S relative to G1 phase. The technology and methods used are solid, but the significance of the work is reduced by the model system employed, the HeLa cell line, which has a greatly abnormal genome.

We have obtained data from a diploid human cell that validates the reduction of S-phase chromatin mobility.

Public Review:

The manuscript presented by Pabba et al. studied chromatin dynamics throughout the cell cycle. The authors used fluorescence signals and patterns of GFP-PCNA (GFP tagged proliferating cell nuclear antigen) and CY3-dUTP (which labels newly synthesized DNA but not the DNA template) to determine cell cycle stages in asynchronized HeLa (Kyoto) cells and track movements of chromatin domains. PCNA binds to replication forks and form replication foci during the S phase. The major conclusions are: (1) Labeled chromatin domains were more mobile in G1/G2 relative to the S-phase. (2) Restricted chromatin motion occurred at sites in proximity to DNA replication sites. (3) Chromatin motion was restricted by the loading of replisomes, independent of DNA synthesis. This work is based on previous work published in 2015, entitled "4D Visualization of replication foci in mammalian cells corresponding to individual replicons," in which the labeling method was demonstrated to be sound. Although interesting, reduced chromatin mobility in S relative to G1 phase is not new to the field.

It was first shown in yeast (Heun et al. 2001; DOI:10.1126/science.1065366) that the S-phase mobility is reduced compared to the G1 phase. This was followed by other papers showing the same in yeast [(Gasser 2002; DOI: 10.1126/science.1067703), (Smith et al. 2019; DOI: 10.1091/mbc.E19-08-0469)]. The relation between chromatin motion and cell cycle progression in the mammalian genome is less studied. Over recent years there have been a few studies that addressed chromatin mobility and cell cycle progression but from a different perspective. In the publication Nozaki et al. (2017; DOI:10.1016/j.molcel.2017.06.018) chromatin motion analysis was performed on single histones. The study did not find a significant change of histone/nucleosome mobility measured during cell cycle progression. Using CRISPR/dCas9 to label random DNA loci, Ma et al. (2019; DOI:10.1083/jcb.201807162) found that chromatin motion in S-phase was significantly lower than in the G1 phase. However, most of the studies measure the chromatin motion using either insertion of ectopic loci or proteins marking the loci (dCas9) or histones. Using either ectopic loci addition or CRISPR/dCas9 might have an effect on the chromatin mobility itself and measuring single histone motion is not equivalent to measuring the motion of DNA segments. We, therefore, opted to label the DNA directly using the replication of the DNA. In this manner we preserve the native chromatin structure and, thus, motion.

Importantly, in addition to measuring decreased DNA motion in S-phase, our study indicates that it is not the DNA synthesis per se but the loading of replisomes onto chromatin that slows down its motion. This allowed us to propose a mechanism on how chromatin motion is affected by DNA replication in S-phase.

The genome in HeLa cells is greatly abnormal with heterogeneous aneuploidy, which makes quantification complicated and weakens the conclusions.

We agree that the HeLa cells are aneuploid and we have addressed the heterogeneity of HeLa Kyoto within our detection methods (for clarification see point 3). To validate our conclusions in normal diploid human cells, we performed the chromatin mobility analysis using human fibroblasts (IMR90 cells in figures 2, 3 and S2) and plotted the MSD curves for different cell cycle stages. The outcome of this analysis showed that the mobility of chromatin in diploid fibroblasts in S-phase is lower than in G1 and G2. In fact, this effect is stronger in IMR90 cells than in HeLa Kyoto cells. Hence, this is not an aneuploid tumor cell phenomenon.

The manuscript is difficult to follow in places due to insufficient clarity. The manuscript should be written in a way that can be understood without referencing previous articles. Overall, the work is moderately impactful to the field.

Major recommendations:

1) In Figure 1B, the illustration and images for S phase are confusing. The author should specify which is early S and which is late S. Do the yellow circles represent GFP-PCNA foci? How did the authors distinguish mid S from early S and late S (in Figure 2)? Are all images in Figure 1 scaled to the same contrast threshold?

The yellow circles correspond to the colocalized signal of GFP-PCNA and Cy3-dUTP that overlap and represent the labeled chromatin sites that are replicated in the next cell cycle.

We clarified all the points mentioned above and updated figure 1 and figure 2 accordingly.

2) In Figure 2B, the y-axis is marked as "Frequency of cells" but the equation listed below is counting DNA (per focus). How to convert DNA (per focus) to DNA (per cell)? The x-axis is marked as "Genome size" without any unit (e.g., kb? Mb?) The x-axis seems to be the C factor, not the genome size.

To determine the amount of DNA present in each labeled DNA focus, we first segmented the whole nucleus and measured the total intensity of DAPI (DNA amount) which is called IDNA TOTAL. Then the labeled replication foci are segmented and the intensity of label present in each segmented foci is measured (IRFi). Throughout the S-phase progression the amount of DNA increases twofold from early to late S-phase. The cells at each cell cycle stage were determined using the PCNA pattern. By plotting the frequency (number of cells) and the relative genome content normalized to the G1 stage we calculated the relative genome size otherwise called cell cycle correction factor for each stage from G1 to G2. The ratio of DNA intensity in labeled replication (IRFi)/ to the total DNA intensity of DAPI (IDNA total) gives the fraction of DNA present in each foci compared to the whole nucleus. This ratio was then multiplied by the genome size (Kbp) of HeLa Kyoto cells which was measured and published in Chagin et al. (2016; DOI:10.1038/ncomms11231). This gives us the approximate amount of DNA present in each labeled replication foci in Kbp. Since the genome duplicates over cell cycle stages, the measured DNA content in IRFi was corrected to the cell cycle stage (determined by PCNA) by multiplying the cell cycle correction factor.

3) HeLa cells are known to be highly heterogeneous and heavily aneuploidy. Cells in one sample have different numbers of chromosomes ranging from 50 - 80. Therefore, GS (genome size) for each cell should not be the same. Using one constant GS in the equation for every cell introduces errors. Has the cell-to-cell variation been considered and corrected in the data? If not, the authors should provide information regarding cell-to-cell variations, such as the intensity variation of nuclear DAPI signals in synchronized cells.

It is true that the HeLa genome is aneuploid. However, the heterogeneity of the genome is true, if one compares different HeLa strains as studied in Frattini et al. (2015; DOI:10.1038/srep15377), where they show the variability of genome and RNA expression profiles and small genomic rearrangements among different HeLa strains. However, to our knowledge, it is not studied extensively or shown whether the heterogeneity and aneuploidy would also be a cell to cell variation. Therefore, we performed a control experiment to verify the variability between HeLa Kyoto cells, where we either synchronized or not and stained with DAPI and the DNA content profiles of all cells were plotted as a histogram (supplementary figure 1B) to show that cell to cell variations is not present and by synchronizing, we see that the cell population in G1, has similar DNA content showing that the cell to cell variability is negligible in our detection methods. Nonetheless, we have obtained data using normal diploid human fibroblasts, which validated our outcome.

STABLE:

Macville, Merryn, et al. "Comprehensive and definitive molecular cytogenetic characterization of HeLa cells by spectral karyotyping." Cancer research 59.1 (1999): 141-150.

UNSTABLE:

Liu, Yansheng, et al. "Multi-omic measurements of heterogeneity in HeLa cells across laboratories." Nature biotechnology 37.3 (2019): 314-322.

Landry, Jonathan JM, et al. "The genomic and transcriptomic landscape of a HeLa cell line." G3: Genes, Genomes, Genetics 3.8 (2013): 1213-1224.

4) The chromatin foci are in a variety of sizes and intensities. How were boundaries of foci determined? Weak foci were picked up in one image but not in another. This is a concern because the size of the chromatin domain could influence mobility measurement. The authors should provide control experiments or better explanations for detecting and selecting chromatin foci.

The method for detecting chromatin foci is described in “Materials and Methods” section “Automated tracking of chromatin structures in time-lapse videos”. “Chromatin structures are detected by the spot-enhancing filter (SEF) (Sage et al., 2005; doi:10.1109/TIP.2005.852787) which consists of a Laplacian-of-Gaussian (LoG) filter followed by thresholding the filtered image and determination of local maxima. The threshold is automatically determined by the mean of the absolute values of the filtered image plus a factor times the standard deviation.” For reasons of consistency, we used the same threshold factor for all images of an image sequence. Therefore, depending on the intensity distribution in an image, it can happen that weak foci are not detected in some images. Alternatively, one could manually adapt the threshold factor for all single images, which, however, would be subjective. We now added the information that we used the same threshold factor for all images of an image sequence.

5) In Figure 3, the authors combined MSD from G1 and G2 in one group. Has any published data suggested that chromatin dynamics are the same in G1 and G2?

To clarify this we separated G1 and G2 mobility measurements in supplementary figure S2 and updated the figures and text accordingly.

6) In Figure 3B, cytoplasmic CY3-dUTP foci are found in the G1/G2 and S images. Are these CY3-dUTP aggregates? If so, are they also found in the nucleus? What is the mobility of the cytoplasmic CY3-dUTP foci?

These are aggregates and not found in the nucleus. These foci were excluded from the analysis by using a nuclear mask based on the PCNA signal. This information was added to the figure 3B legend.

7) In Figure 4, how is colocalization defined? 1.8 um is approximately the size of a chromosome territory, which is much larger than 0.5 Mb. Two foci that are 1.8 um apart should not be considered in the same chromosome.

We agree that colocalized would indeed mean that the signals are overlapping. Therefore, we updated the figures and text as center to center distance or proximity analysis.

Minor comments:

1) Figure 3D should be presented by a box and whisker plot. The histogram does not show an actual distribution of the data.

The histograms shown in figure 3D is the average mean square displacement measurement value for different cell cycle stages. These are the same data shown in the table. Therefore, the histogram is removed and the table in figure 3C is retained.

2) Please explain Figure 3C error bars in the figure legend. Are they SD?

The error bars of the MSD curves (highlighted in bright color around the curves) in figure 3C show the standard error of the mean (SEM) representing the deviations between the MSD curves for an image sequence. We clarified this in the legend of Figure 3C.

3) In Figure 5C, some western blotting results seem to be assembled from replicate experiments. Comparing signals from one experiment with the same background is suggested.

We made sure that the western blots from the same replicates are cropped and the information is also added to the respective figure legends.

Author Response

The following is the authors’ response to the original reviews.

We are very grateful to the reviewers for their thorough assessment of our study, and their acknowledgment of its strengths and weaknesses. We did our best below to address the weaknesses raised in their public review, and to comply with their recommendations.

Reviewer #1 (Public Review):

Segas et al. present a novel solution to an upper-limb control problem which is often neglected by academia. The problem the authors are trying to solve is how to control the multiple degrees of freedom of the lower arm to enable grasp in people with transhumeral limb loss. The proposed solution is a neural network based approach which uses information from the position of the arm along with contextual information which defines the position and orientation of the target in space. Experimental work is presented, based on virtual simulations and a telerobotic proof of concept

The strength of this paper is that it proposes a method of control for people with transhumeral limb loss which does not rely upon additional surgical intervention to enable grasping objects in the local environment. A challenge the work faces is that it can be argued that a great many problems in upper limb prosthesis control can be solved given precise knowledge of the object to be grasped, its relative position in 3D space and its orientation. It is difficult to know how directly results obtained in a virtual environment will translate to real world impact. Some of the comparisons made in the paper are to physical systems which attempt to solve the same problem. It is important to note that real world prosthesis control introduces numerous challenges which do not exist in virtual spaces or in teleoperation robotics.

We agree that the precise knowledge of the object to grasp is an issue for real world application, and that real world prosthesis control introduces many challenges not addressed in our experiments. Those were initially discussed in a dedicated section of the discussion (‘Perspectives for daily-life applications’), and we have amended this section to integrate comments by reviewers that relate to those issues (cf below).

The authors claim that the movement times obtained using their virtual system, and a teleoperation proof of concept demonstration, are comparable to natural movement times. The speed of movements obtained and presented are easier to understand by viewing the supplementary materials prior to reading the paper. The position of the upper arm and a given target are used as input to a classifier, which determines the positions of the lower arm, wrist and the end effector. The state of the virtual shoulder in the pick and place task is quite dynamic and includes humeral rotations which would be challenging to engineer in a real physical prosthesis above the elbow. Another question related to the pick and place task used is whether or not there are cases where both the pick position and the place position can be reached via the same, or very similar, shoulder positions? i.e. with the shoulder flexion-extension and abduction-adduction remaining fixed, can the ANN use the remaining five joint angles to solve the movement problem with little to no participant input, simply based on the new target position? If this was the case, movements times in the virtual space would present a very different distribution to natural movements, while the mean values could be similar. The arguments made in the paper could be supported by including individual participant data showing distributions of movement times and the distances travelled by the end effector where real movements are compared to those made by an ANN.

In the proposed approach users control where the hand is in space via the shoulder. The position of the upper arm and a given target are used as input to a classifier, which determines the positions of the lower arm, wrist and the effector. The supplementary materials suggest the output of the classifier occurs instantaneously, in that from the start of the trial the user can explore the 3D space associated with the shoulder in order to reach the object. When the object is reached a visual indicator appears. In a virtual space this feedback will allow rapid exploration of different end effector positions which may contribute to the movement times presented. In a real world application, movement of a distal end-effector via the shoulder is not to be as graceful and a speed accuracy trade off would be necessary to ensure objects are grasped, rather than knocked or moved.

As correctly noted by the reviewer and easily visible on videos, the distal joints predicted by the ANN are realized instantaneously in the virtual arm avatar, and a discontinuity occurs at each target change whereby the distal part of the arm jumps to the novel prediction associated with the new target location. As also correctly noted by the reviewer, there are indeed some instances where minimal shoulder movements are required to reach a new target, which in practice implies that on those instances, the distal part of the arm avatar jumps instantaneously close to the new target as soon as this target appears. Please note that we originally used median rather than mean movement times per participant precisely to remain unaffected by potential outliers that might come from this or other situations. We nevertheless followed the reviewer’s advice and have now also included individual distributions of movement times for each condition and participant (cf Supplementary Fig. 2 to 4 for individual distributions of movement time for Exp1 to 3, respectively). Visual inspection of those indicates that despite slight differences between participants, no specific pattern emerges, with distributions of movement times that are quite similar between conditions when data from all participants are pooled together.

Movement times analysis indicates therefore that the overall participants’ behavior has not been impacted by the instantaneous jump in the predicted arm positions at each of the target changes. Yet, those jumps indicate that our proposed solution does not satisfactorily reproduce movement trajectory, which has implications for application in the physical world. Although we introduced a 0.75 s period before the beginning of each trial for the robotic arm to smoothly reach the first prediction from the ANN in our POC experiment (cf Methods), this would not be practical for a real-life scenario with a sequence of movements toward different goals. Future developments are therefore needed to better account for movement trajectories. We are now addressing this explicitly in the manuscript, with the following paragraph added in the discussion (section ‘Perspectives of daily-life applications’):

“Although our approach enabled participants to converge to the correct position and orientation to grasp simple objects with movement times similar to those of natural movements, it is important to note that further developments are needed to produce natural trajectories compatible with real-world applications. As easily visible on supplementary videos 2 to 4, the distal joints predicted by the ANN are realized instantaneously such that a discontinuity occurs at each target change, whereby the distal part of the arm jumps to the novel prediction associated with the new target location. We circumvented problems associated with this discontinuity on our physical proof of concept by introducing a period before the beginning of each trial for the robotic arm to smoothly reach the first prediction from the ANN. This issue, however, needs to be better handled for real-life scenarios where a user will perform sequences of movements toward different objects.”

Another aspect of the movement times presented which is of note, although it is not necessarily incorrect, is that the virtual prosthesis performance is close too perfect. In that, at the start of each trial period, either pick or place, the ANN appears to have already selected the position of the five joints it controls, leaving the user to position the upper arm such that the end effector reaches the target. This type of classification is achievable given a single object type to grasp and a limited number of orientations, however scaling this approach to work robustly in a real world environment will necessitate solving a number of challenges in machine learning and in particular computer vision which are not trivial in nature. On this topic, it is also important to note that, while very elegant, the teleoperation proof of concept of movement based control does not seem to feature a similar range of object distance from the user as the virtual environment. This would have been interesting to see and I look forward to seeing further real world demonstrations in the authors future work.

According to this comment, the reviewer has the impression that the ANN had already selected a position of the five joints it controls at the start of each trial, and maintained those fixed while the user operates the upper arm so as to reach the target. Although the jumps at target changes discussed in the previous comment might give this impression, and although this would be the case should we have used an ANN trained with contextual information only, it is important to stress that our control does take shoulder angles as inputs, and produced therefore changes in the predicted distal angles as the shoulder moves.

To substantiate this, we provide in Author response image 1 the range of motion (angular difference at each joint between the beginning and the end of each trial) of the five distal arm angles, regrouped for all angles and trials of Exp1 to 3 (one circle and line per participant, representing the median of all data obtained by that participant in the given experiment and condition, as in Fig. 3 of the manuscript). Please note that those ranges of motion were computed on each trial just after the target changes (i.e., after the jumps) for conditions with prosthesis control, and that the percentage noted on the figure below those conditions correspond to the proportion of the range of motion obtained in the natural movement condition. As can be seen, distal angles were solicited in all prosthesis control conditions by more than half the amount they moved in the condition of natural movements (between 54 and 75% depending on conditions).

Author response image 1.

With respect to the last part of this comment, we agree that scaling this approach to work robustly in a real world environment will necessitate solving a number of challenges in machine learning and in particular computer vision. We address those in a specific section of the discussion (‘Perspectives for daily-life application’) which has been further amended in response to the reviewers’ comments. As also mentioned earlier and at the occasion of our reply to other reviewers’ comments, we also agree that our physical proof of concept is quite preliminary, and we are looking forward to conduct future work in order to solve some of the issues discussed and get closer to real world demonstrations.

Reviewer #2 (Public Review):

Segas et al motivate their work by indicating that none of the existing myoelectric solution for people with transhumeral limb difference offer four active degrees of freedom, namely forearm flexion/extension, forearm supination/pronation, wrist flexion/extension, and wrist radial/ulnar deviation. These degrees of freedom are essential for positioning the prosthesis in the correct plan in the space before a grasp can be selected. They offer a controller based on the movement of the stump.

The proposed solution is elegant for what it is trying to achieve in a laboratory setting. Using a simple neural network to estimate the arm position is an interesting approach, despite the limitations/challenges that the approach suffers from, namely, the availability of prosthetic hardware that offers such functionality, information about the target and the noise in estimation if computer vision methods are used. Segas et al indicate these challenges in the manuscript, although they could also briefly discuss how they foresee the method could be expanded to enable a grasp command beyond the proximity between the end-point and the target. Indeed, it would be interesting to see how these methods can be generalise to more than one grasp.

Indeed, we have already indicated those challenges in the manuscript, including the limitation that our control “is suitable to place the hand at a correct position and orientation to grasp objects in a wide workspace, but not for fine hand and grasp control ...” (cf 4th paragraph of the ‘Perspectives for daily-life applications’ section of the discussion). We have nevertheless added the following sentence at the end of this paragraph to stress that our control could be combined with recently documented solutions for multiple grasp functions: “Our movement-based approach could also be combined with semi-autonomous grasp control to accommodate for multiple grasp functions39,42,44.”

One bit of the results that is missing in the paper is the results during the familiarisation block. If the methods in "intuitive" I would have thought no familiarisation would be needed. Do participants show any sign of motor adaptation during the familiarisation block?

Please note that the familiarization block indicated Fig. 3a contains approximately half of the trials of the subsequent initial acquisition block (about 150 trials, which represents about 3 minutes of practice once the task is understood and proficiently executed), and that those were designed to familiarize participants with the VR setup and the task rather than with the prosthesis controls. Indeed, it is important that participants were made familiar with the setup and the task before they started the initial acquisition used to collect their natural movements. In Exp1 and 2, there was therefore no familiarization to the prosthesis controls whatsoever (and thus no possible adaptation associated with it) before participants used them for the very first time in the blocks dedicated to test them. This is slightly different in Exp3, where participants with an amputated arm were first tested on their amputated side with our generic control. Although slight adaptation to the prosthesis control might indeed have occurred during those familiarization trials, this would be difficult in practice to separate from the intended familiarization to the task itself, which was deemed necessary for that experiment as well. In the end, we believe that this had little impact on our data since that experiment produced behavioral results comparable to those of Exp1 and 2, where no familiarization to the prosthesis controls could have occurred.

In Supplementary Videos 3 and 4, how would the authors explain the jerky movement of the virtual arm while the stump is stationary? How would be possible to distinguish the relative importance of the target information versus body posture in the estimation of the arm position? This does not seem to be easy/clear to address beyond looking at the weights in the neural network.

As discussed in our response to Reviewer1 and now explicitly addressed in the manuscript, there is a discontinuity in our control, whereby the distal joints of the arm avatar jumps instantaneously to the new prediction at each target change at the beginning of a trial, before being updated online as a function of ongoing shoulder movements for the rest of that trial. In a sense, this discontinuity directly reflects the influence of the target information in the estimation of the distal arm posture. Yet, as also discussed in our reply to R1, the influence of proximal body posture (i.e., Shoulder movements) is made evident by substantial movements of the predicted distal joints after the initial jumps occurring at each target change. Although those features demonstrate that both target information and proximal body posture were involved in our control, they do not establish their relative importance. While offline computation could be thought to quantify their relative implication in the estimation of the distal arm posture, we believe that further human-in-the-loop experiments with selective manipulation of this implication would be necessary to establish how this might affect the system controllability.

I am intrigued by how the Generic ANN model has been trained, i.e. with the use of the forward kinematics to remap the measurement. I would have taught an easier approach would have been to create an Own model with the native arm of the person with the limb loss, as all your participants are unilateral (as per Table 1). Alternatively, one would have assumed that your common model from all participants would just need to be 'recalibrated' to a few examples of the data from people with limb difference, i.e. few shot calibration methods.

AR: Although we could indeed have created an Own model with the native arm of each participant with a limb loss, the intention was to design a control that would involve minimal to no data acquisition at all, and more importantly, that could also accommodate bilateral limb loss. Indeed, few shot calibration methods would be a good alternative involving minimal data acquisition, but this would not work on participants with bilateral limb loss.

Reviewer #3 (Public Review):

This work provides a new approach to simultaneously control elbow and wrist degrees of freedom using movement based inputs, and demonstrate performance in a virtual reality environment. The work is also demonstrated using a proof-of-concept physical system. This control algorithm is in contrast to prior approaches which electrophysiological signals, such as EMG, which do have limitations as described by the authors. In this work, the movements of proximal joints (eg shoulder), which generally remain under voluntary control after limb amputation, are used as input to neural networks to predict limb orientation. The results are tested by several participants within a virtual environment, and preliminary demonstrated using a physical device, albeit without it being physically attached to the user.

Strengths:

Overall, the work has several interesting aspects. Perhaps the most interesting aspect of the work is that the approach worked well without requiring user calibration, meaning that users could use pre-trained networks to complete the tasks as requested. This could provide important benefits, and if successfully incorporated into a physical prosthesis allow the user to focus on completing functional tasks immediately. The work was also tested with a reasonable number of subjects, including those with limb-loss. Even with the limitations (see below) the approach could be used to help complete meaningful functional activities of daily living that require semi-consistent movements, such as feeding and grooming.

Weaknesses:

While interesting, the work does have several limitations. In this reviewer's opinion, main limitations are: the number of 'movements' or tasks that would be required to train a controller that generalized across more tasks and limbpostures. The authors did a nice job spanning the workspace, but the unconstrained nature of reaches could make restoring additional activities problematic. This remains to be tested.

We agree and have partly addressed this in the first paragraph of the ‘Perspective for daily life applications’ section of the discussion, where we expand on control options that might complement our approach in order to deal with an object after it has been reached. We have now amended this section to explicitly stress that generalization to multiple tasks including more constrained reaches will require future work: “It remains that generalizing our approach to multiple tasks including more constrained reaches will require future work. For instance, once an intended object has been successfully reached or grasped, what to do with it will still require more than computer vision and gaze information to be efficiently controlled. One approach is to complement the control scheme with subsidiary movements, such as shoulder elevation to bring the hand closer to the body or sternoclavicular protraction to control hand closing26, or even movement of a different limb (e.g., a foot45). Another approach is to control the prosthesis with body movements naturally occurring when compensating for an improperly controlled prosthesis configuration46.”

The weight of a device attached to a user will impact the shoulder movements that can be reliably generated. Testing with a physical prosthesis will need to ensure that the full desired workspace can be obtained when the limb is attached, and if not, then a procedure to scale inputs will need to be refined.

We agree and have now explicitly included this limitation and perspective to our discussion, by adding a sentence when discussing possible combination with osseointegration: “Combining those with osseointegration at humeral level3,4 would be particularly relevant as this would also restore amplitude and control over shoulder movements, which are essential for our control but greatly affected with conventional residual limb fitting harness and sockets. Yet, testing with a physical prosthesis will need to ensure that the full desired workspace can be obtained with the weight of the attached device, and if not, a procedure to scale inputs will need to be refined.”

The reliance on target position is a complicating factor in deploying this technology. It would be interesting to see what performance may be achieved by simply using the input target positions to the controller and exclude the joint angles from the tracking devices (eg train with the target positions as input to the network to predict the desired angles).

Indeed, the reliance on precise pose estimation from computer vision is a complicating factor in deploying this technology, despite progress in this area which we now discuss in the first paragraph of the ‘Perspective for daily life applications’ section of the discussion. Although we are unsure what precise configuration of input/output the reviewer has in mind, part of our future work along this line is indeed explicitly dedicated to explore various sets of input/output that could enable coping with availability and reliability issues associated with real-life settings.

Treating the humeral rotation degree of freedom is tricky, but for some subjects, such as those with OI, this would not be as large of an issue. Otherwise, the device would be constructed that allowed this movement.

We partly address this when referring to osseointegration in the discussion: “Combining those with osseointegration at humeral level3,4 would be particularly relevant as this would also restore amplitude and control over shoulder movements, which are essential for our control but greatly affected with conventional residual limb fitting harness and sockets.” Yet, despite the fact that our approach proved efficient in reconstructing the required humeral angle, it is true that realizing it on a prosthesis without OI is an open issue.

Overall, this is an interesting preliminary study with some interesting aspects. Care must be taken to systematically evaluate the method to ensure clinical impact.

Reviewer #1 (Recommendations For The Authors):

Page 2: Sentence beginning: "Here, we unleash this movement-based approach by ...". The approach presented utilises 3D information of object position. Please could the authors clarify whether or not the computer vision references listed are able to provide precise 3D localisation of objects?

While the references initially cited in this sentence do support the view that movement goals could be made available in the context of prosthesis control through computer vision combined with gaze information, it is true that they do not provide the precise position and orientation (I.e., 6d pose estimation) necessary for our movementbased control approach. Six-dimensional object pose estimation is nevertheless a very active area of computer vision that has applications beyond prosthesis control, and we have now added to this sentence two references illustrating recent progress in this research area (cf. references 30 and 31).

Page 6: Sentence beginning: "The volume spread by the shoulder's trajectory ...".

• Page 7: Sentence beginning: "With respect to the volume spread by the shoulder during the Test phases ...".

• Page 7: Sentence beginning: "Movement times with our movement-based control were also in the same range as in previous experiments, and were even smaller by the second block of intuitive control ...".

On the shoulder volume presented in Figure 3d. My interpretation of the increased shoulder volume in Figure 3D Expt 2 shown in the Generic ANN was that slightly more exploration of the upper arm space was necessary (as related to the point in the public review). Is this what the authors mean by the action not being as intuitive? Does the reduction in movement time between TestGeneric1 and TestGeneric 2 not suggest that some degree of exploration and learning of the solution space is taking place?

Indeed, the slightly increased shoulder volume with the Generic ANN in Exp2 could be interpreted as a sign that slightly more exploration of the upper arm space was necessary. At present, we do not relate this to intuitiveness in the manuscript. And yes, we agree that the reduction in movement time between TestGeneric1 and TestGeneric 2 could suggest some degree of exploration and learning.

Page 7: Sentence beginning: "As we now dispose of an intuitive control ...". I think dispose may be a false friend in this context!

This has been replaced by “As we now have an intuitive control…”.

Page 8: Section beginning "Physical Proof of Concept on a tele-operated robotic platform". I assume this section has been added based on suggestions from a previous review. Although an elegant PoC the task presented in the diagram appears to differ from the virtual task in that all the targets are at a relatively fixed distance from the robot. In respect to the computer vision ML requirements, this does not appear to require precise information about the distance between the user and an object. Please could this be clarified?

Indeed, the Physical Proof of Concept has been added after the original submission in order to comply with requests formulated at the editorial stage for the paper to be sent for review. Although preliminary and suffering from several limitations (amongst which a reduced workspace and number of trials as compared to the VR experiments), this POC is a first step toward realizing this control in the physical world. Please note that as indicated in the methods, the target varied in depth by about 10 cm, and their position and orientation were set with sensors at the beginning of each block instead of being determined from computer vision (cf section ‘Physical Proof of Concept’ in the ‘Methods’: “The position and orientation of each sponge were set at the beginning of each block using a supplementary sensor. Targets could be vertical or tilted at 45 and -45° on the frontal plane, and varied in depth by about 10 cm.”).

Page 10: Sentence beginning: "This is ahead of other control solutions that have been proposed ...". I am not sure what this sentence is supposed to convey and no references are provided. While the methods presented appear to be a viable solution for a group of upper-limb amputees who are often ignored by academic research, I am not sure it is appropriate for the authors to compare the results obtained in VR and via teleoperation to existing physical systems (without references it is difficult to understand what comparison is being made here).

The primary purpose of this sentence is to convey that our approach is ahead of other control solutions proposed so far to solve the particular problem as defined earlier in this paragraph (“Yet, controlling the numerous joints of a prosthetic arm necessary to place the hand at a correct position and orientation to grasp objects remains challenging, and is essentially unresolved”), and as documented to the best we could in the introduction. We believe this to be true and to be the main justification for this publication. The reviewer’s comment is probably directed toward the second part of this sentence, which states that performances of previously proposed control solutions (whether physical or in VR) are rarely compared to that of natural movements, as this comparison would be quite unfavorable to them. We soften that statement by removing the last reference to unfavorable comparison, but maintained it as we believe it is reflecting a reality that is worth mentioning. Please note that after this initial paragraph, and an exposition of the critical features of our control, most of the discussion (about 2/3) is dedicated to limitations and perspectives for daily-life application.

Page 10: Sentence: "Here, we overcame all those limitations." Again, the language here appears to directly compare success in a virtual environment with the current state of the art of physical systems. Although the limitations were realised in a virtual environment and a teleoperation PoC, a physical implementation of the proposed system would depend on advances in machine vision to include movement goal. It could be argued that limitations have been traded, rather immediately overcome.

In this sentence, “all those limitations” refers to all three limitations mentioned in the previous sentences in relation to our previous study which we cited in that sentence (Mick et al., JNER 2021), rather than to limitations of the current state of the art of physical systems. To make this more explicit, we have now changed this sentence to “Here, we overcome those three limitations”.

Page 11: Sentence beginning: "Yet, impressive progresses in artificial intelligence and computer vision ...".

• Page 11: Sentence beginning: "Prosthesis control strategies based on computer vision ..."

The science behind self-driving cars is arguably of comparable computational complexity to the real-world object detection and with concurrent real-time grasp selection. The market for self-driving cars is huge and a great deal of R&D has been funded, yet they are not yet available. The market for advanced upper-limb prosthetics is very small, it is difficult to understand who would deliver this work.

We agree that the market for self-driving cars is much higher than that for advanced upper-limb prosthetics. Yet, as mentioned in our reply to a previous comment, 6D object pose estimation is a very active area of computer vision that has applications far beyond prosthesis control (cf. in robotics and augmented reality). We have added two references reflecting recent progress in this area in the introduction, and have amended the discussion accordingly: “Yet, impressive progress in artificial intelligence and computer vision is such that what would have been difficult to imagine a decade ago appears now well within grasp38. For instance, we showed recently that deep learning combined with gaze information enables identifying an object that is about to be grasped from an egocentric view on glasses33, and this even in complex cluttered natural environments34. Six-dimensional object pose estimation is also a very active area of computer vision30,31, and prosthesis control strategies based on computer vision combined with gaze and/or myoelectric control for movement intention detection are quickly developing39–44, illustrating the promises of this approach.”

Page 15: Sentence beginning: "From this recording, 7 signals were extracted and fed to the ANN as inputs: ...".

• Page 15: Sentence beginning: "Accordingly, the contextual information provided as input corresponded to the ...".

The two sentences appear to contradict one another and it is difficult to understand what the Own ANN was trained on. If the position and the orientation of the object were not used due to overfitting, why claim that they were used as contextual information? Training on the position and orientation of the hand when solving the problem would not normally be considered contextual information, the hand is not part of the environment or setting, it is part of the user. Please could this section be made a little bit clearer?

The Own ANN was trained using the position and the orientation of a hypothetic target located within the hand at any given time. This approach has been implemented to increase the amount of available data. However, when the ANN is utilized to predict the distal part of the virtual arm, the position and orientation of the current target are provided. We acknowledge that the phrasing could be misleading, so we have added the following clarification to the first sentence: "… (3 Cartesian coordinates and 2 spherical angles that define the position and orientation of the hand as if a hypothetical cylindrical target was placed in it at any time, see an explanation for this choice in the next paragraph)".

Page 16: Sentence beginning: "A trial refers to only one part of this process: either ...". Would be possible to present these values separately?

Although it would be possible to present our results separately for the pick phase and for the place phase, we believe that this would overload the manuscript for little to no gain. Indeed, nothing differentiates those two phases other than the fact that the bottle is on the platform (waiting to be picked) in the pick phase, and in the hand (waiting to be placed) in the place phase. We therefore expect to have very similar results for the pick phase and for the place phase, which we verified as follows on Movement Time: Author response image 2 shows movement time results separated for the pick phase (a) and for the place phase (b), together with the median (red dotted line) obtained when results from both phases are polled together. As illustrated, results are very similar for both phases, and similar to those currently presented in the manuscript with both phases pooled (Fig3C).

Author response image 2.

Page 19: Sentence beginning "The remaining targets spanned a roughly ...". Figure 2 is a very nice diagram but it could be enhanced with a simple visual representation of this hemispherical region on the vertical and horizontal planes.

We made a few attempts at enhancing this figure as suggested. However, the resulting figures tended to be overloaded and were not conclusive, so we opted to keep the original.

Page 19: Sentence beginning "The Movement Time (MT) ..."

• Page 19: Sentence beginning "The shoulder position Spread Volume (SV) ..." Would it be possible to include a traditional timing protocol somewhere in the manuscript so that readers can see the periods over which these measures calculated?

We have now included Fig. 5 to illustrate the timing protocol and the periods over which MT and SV were computed.

Reviewer #2 (Recommendations For The Authors):

Minor comments

Page 6: "Yet, this control is inapplicable "as is" to amputees, for which recording ..." -> "Yet, this control is inapplicable "as is" to amputees, for WHOM recording ... "

This has been modified as indicated.

Throughout: "amputee" -> "people with limb loss" also "individual with limb deficiency" -> "individual with limb difference"

We have modified throughout as indicated.

It would have been great to see a few videos from the tele-operation as well. Please could you supply these videos?

Although we agree that videos of our Physical Proof of Concept would have been useful, we unfortunately did not collect videos that would be suitable for this purpose during those experimental phases. Please note that this Physical Proof of Concept was not meant to be published originally, but has been added after the original submission in order to comply with requests formulated at the editorial stage for the paper to be sent for review.

Reviewer #3 (Recommendations For The Authors):

Consider using the terms: intact-limb rather than able-bodied, residual limb rather than stump, congenital limb different rather than congenital limb deficiency.

We have modified throughout as indicated.

Author Response

The following is the authors’ response to the original reviews.

REVIEWER #1:

The authors present a carefully controlled set of experiments that demonstrate an additional complexity for GPCR signaling in that endosomal signaling make be different when b-arrestin is or isn't associated with a G protein-bound V2R vasopressin receptor. It uses state of the art biosensorbased approaches and b-arrestin KO lines to assess this. It adds to a growing body of evidence that G proteins and b-arrestin can associate with GPCR complexes simultaneously. They also demonstrate the possibility that Gaq might also be activated by the V2R receptor. My sense is one thing they may need to be considered is the possibility of such "megacomplexes" might actually involve receptor dimers or oligomers.

1.1 Can the authors please review the data that describes the concept of "GPCR megacomplexes"? I feel this is missing from the introduction. The notion means different things to different people. As you will see from my other comments, you should especially focus on evidence at the level of the single receptor.

We appreciate the reviewer’s comments and have now included a more wholesome description of the GPCR megacomplex, or ‘megaplex’, concept in the introduction (page 2, 1st paragraph).

1.2 The authors use mini-G proteins to conclude that V2R receptors interact with Gaq (in addition to Gas). I would prefer if there were a more direct measure of this. Can the authors show that the receptor interacts with full length Gaq (and not the other G proteins in Figure)? Is there a signaling phenotype associated with Gaq coupling? Is it sensitive to Gaq inhibition?

Excellent point and we are happy to expand further on this. The ability of the V2R to activate Gq/11 has already been demonstrated before (Zhu, X. et al. Mol Pharmacol 46(3):460-9 (1994); Lykke, K. et al. Physiol Rep. 3(8):e12519 (2015); Avet, C. et al. eLife 11: e74101 (2022); Heydenreich, F.M. et al. Mol Pharmacol 102(3):139-49 (2022). Therefore, we did not attempt to document this activation using more traditional assays. On the other hand, to demonstrate an interaction between V2R and Ga subunit in cells is challenging for several reasons. First, the full-length Ga subunit is already located at the plasma membrane at basal state, and thus, generates high background signals in proximity assays. Second, upon receptor activation, the Ga subunit interaction with V2R is so transient that it is difficult, if not impossible, to catch this transient moment in a proximity assay. Although the miniG proteins are highly engineered, coupling specificity of the different subtypes (Gas, Gai/o, Gaq/11, and Ga12/13) to GPCRs is maintained. In addition, as they are homogenously expressed in the cytosol under basal states rather than at the membrane, they generate low background noise. Upon agonist stimulation, miniG proteins are recruited from the cytosol to the V2R at the plasma membrane, resulting in a robust signal in proximity assays. Thus, miniG proteins are unique in that they can actually detect GPCR–G protein interactions in cellular proximity assays, which is very challenging using full-length Ga subunits.

That being said, we fully understand the reviewer’s concern and greatly value the effort in enhancing robustness of our study. Therefore, we have now monitored downstream signaling events of Gaq/11 in the absence or presence of the selective Gaq/11 inhibitor YM-254890 as a secondary method of documenting Gaq/11 activity. Specifically, we used a newly developed biosensor to measure diacylglycerol (DAG) production, a downstream second messenger of Gaq/11 activation, at both the plasma membrane and endosomes. Using a second biosensor, we detect general protein kinase C (PKC) activation, which is another downstream signaling event of Gaq/11 activation. Together, we demonstrated that AVP-stimulation leads to DAG production at both the plasma membrane and endosomes (Fig. 1C-D) as well as PKC activation (Fig. 1E), which all are sensitive to YM-254890 inhibition (Fig. 1C-D and E). Together these results rigorously suggest that the V2R interacts with and activates Gaq/11.

1.3 I raise a similar concern with Gaq coupling in endosomes.

For similar reasons that miniG proteins are excellent tools for demonstrating V2R interaction with G proteins at the plasma membrane, miniG proteins can also be used to detect V2R interaction with G proteins at endosomes by measuring proximity between miniG and an endosomal marker in response to agonist challenge. However, to ensure that the endosomal recruitment of miniGsq to the V2R demonstrated in our study corresponds to endosomal Gaq/11 activation, we monitored the production of DAG at the early endosomes in a similar way to which we detected DAG production at the plasma membrane. As shown in Fig. 1D, stimulation of V2R with AVP induces recruitment of the DAG-binding biosensor to the early endosomal marker Rab5. Pre-treatment of the cells with the selective Gaq/11 inhibitor YM-254890 abrogated this response, confirming that V2R activation leads to production of DAG at the early endosomes in a Gaq/11-dependent manner (Fig. 1D).

1.4 Can the confocal data be shown for Gai and Ga12?

Yes, we can certainly show this data as negative control. We have now included the confocal data using Halo-mGsi as a negative control for confocal microscopy (Fig. 2). As seen on this figure, mGsi does not colocalize with Lck (plasma membrane), nor with EEA1 (early endosomes) upon stimulation of cells with AVP in line with a receptor that does not couple to Gai/o.

We did not include data using Halo-mG12, as this G protein subtype, similar to Gi/o, does not couple functionally to V2R. Therefore, it is highly unlikely we would obtain different results from the experiments using Halo-mGsi.

1.5 The authors want us to believe that there is simultaneous binding of G proteins and b-arrestin. This is never demonstrated and is at odds with the structural basis of G protein and b-arrestin binding. Have the authors considered that "simultaneous" occupancy might simply reflect binding at distinct GPCR monomers in the context of dimeric or oligomeric receptors? They could I suppose provide data at the level of a single receptor rather than using the bulk BRET approaches used.

We appreciate the comment and opportunity to highlight some of our previous work, which address the megacomplexes at the level of a single receptor. First, we have characterized the megacomplex biochemically and structurally at a low resolution (Thomsen ARB et al. 2016, Cell 166(4):907-19). The results unequivocally demonstrate that a single GPCR interacts simultaneously with heterotrimeric G protein, at the receptor core, and with b-arrestin via the phosphorylated receptor carboxy-terminal. We also documented functionality of the megacomplex as the receptor can interact with and activate the G protein, which were shown by 3 different biochemical approaches (Thomsen ARB et al. 2016, Cell 166(4):907-19). In addition, we solved a high-resolution cryo-EM structure of a megacomplex further highlighting the architecture of this complex (Nguyen AH et al. 2019, Nat Struct Mol Biol 26:1123-31). As both biochemical and structural analyses were done in vitro in which the receptor was embedded in a detergent micelle, we also confirmed that the megacomplex structural architecture fits naturally within the context of a membrane in molecular dynamics simulation experiments (Nguyen AH et al. 2019, Nat Struct Mol Biol 26:1123-31).

In cells, we and others have also showed that GPCRs such as the V2R can bind b-arrestins exclusively via the phosphorylated carboxy-terminal tail as it does in the megacomplex (Kumari P et al. 2016, Nat Commun 7:13416; Cahill III TJ et al. 2017, PNAS 114(10):2562-67; Kumari P et al. 2017, Mol Biol Cell 28(8):1003-10; Chen K et al. 2023, Nature (online doi: https://doi.org/10.1038/s41586-023-06420-x). In addition, we and others have used BRET and confocal microscopy to show that the V2R and other GPCRs recruit G protein and b-arrestin simultaneously and that the three components colocalize in endosomes upon prolonged agonist exposure (Thomsen ARB et al. 2016, Cell 166(4):907-19; Chen K et al. 2023, Nature (online doi: https://doi.org/10.1038/s41586-023-06420-x). As the reviewer correctly points out, in these cellular experiments (as well as in single molecule microscopy), the working resolution is not high enough to rule out that the receptors that co-recruit G protein and b-arrestin in endosomes could be dimeric instead of monomeric. Thus, we conducted a series of experiments with GPCR–b-arrestin fusions where the two proteins are covalently attached at the receptor carboxy-terminal tail. We showed that despite the GPCR–b-arrestin coupling being fully functional (in respect to b-arrestin promoting a highaffinity state of the receptor for agonist binding and constitutively internalizing the receptor) the receptor could still activate G proteins (Thomsen ARB et al. 2016, Cell 166(4):907-19; Nguyen AH et al. 2019, Nat Struct Mol Biol 26:1123-31), which demonstrates that the single receptor megaplex can physically form in cells.

We have now included an extra paragraph in the discussion to go over these megaplex-related considerations (5th paragraph in the discussion), and we thank the reviewer for raising this point.

1.6 Please introduce abbreviations when you first use this- this was not done consistently.

Thank you for noticing these errors, which we now have corrected.  

REVIEWER #2:

This manuscript by Daly et al., probes the emerging paradigm of GPCR signaling from endosomes using the V2R as a model system with an emphasis on Gaq/11 and b-arrestins. The study employs cellular imaging, enzyme complementation assays and energy transfer-based sensors to probe the potential formation of GPCR-G-protein-b-arrestin megaplexes. While the study is certainly very interesting, it appears to be very preliminary at many levels, and clearly requires further development in order to make robust conclusions. The authors should consider expanding on this work further to make the points more convincingly to make the work solid and impactful. The two corresponding authors are among the leaders in the field having demonstrated the existence of megaplexes, and building on the work in a systematic fashion should certainly move the paradigm forward. As the work presented in the current manuscript is already pre-printed, the authors should take this opportunity to present a completer and more comprehensive story to the field.

We are grateful for the time and efforts the reviewer has put into reviewing our work. We are certainly excited to learn that the reviewer finds our work “very interesting”. Regarding the robustness, we have added extra control experiments to increase the completeness of the study. These experiments include:

• Measurements of AVP-stimulated diacylglycerol production, a signaling event downstream of Gaq/11 activation. These measurements were conducted both at plasma membrane (Fig. 1C) and early endosomes (Fig. 1D) using a newly developed DAG-binding biosensor, and demonstrate that the V2R activates Gaq/11 at both of these subcellular locations.

• Monitoring AVP-promoted protein kinase C activation, another downstream signaling effect of Gaq/11 activation (Fig. 1E). The result of this approach shows in another way that V2R activates of Gaq/11.

• Inhibition of signaling events downstream of Gaq/11 activation using the selective of Gaq/11 inhibitor YM254890. YM-254890 inhibits both AVP-stimulated DAG production at plasma membrane and endosomes as well as PKC activation (Fig. 1C-E), which strongly confirms that these signaling outputs are results of Gaq/11 activation.

• We have also included the confocal data using Halo-mGsi as a negative control for confocal microscopy (Fig. 2). As seen in this figure, mGsi does not translocate to the plasma membrane or early endosomes upon stimulation with AVP, which validates that V2R activation does not couple to and activate Gai/o.

Finally, we would like to kindly remind the reviewer that the production of the pre-print manuscript is part of the peer-review process in eLife.

2.1 The use of miniG proteins in these experiments is a major concern as these are highly engineered and may not represent the true features of G proteins. While these have been used as a readout in other publications, their use in demonstrating megaplex formation is sub-optimal, and native, full-length G proteins should be used.

We are a bit unsure as to what the reviewer means by using native full-length G proteins. If the reviewer is suggesting to co-immunoprecipitate V2R with native unlabeled G protein and b-arrestin, it should be considered that the G protein interaction with the receptor is extremely transient and unlikely to survive the pull-down procedure unless stabilized by a nanobody or crosslinking. Although the b-arrestin interaction with the receptor is more stable of nature, co-immunoprecipitation with the receptor requires crosslinking or stabilization with a Fab/nanobody. Therefore, we do not think this approach can be used as a more accurate way of detecting native megaplexes.

If the reviewer is suggesting the use of full-length G proteins in our cell-based proximity assays instead of miniG proteins, we would like to highlight that this approach is somewhat prone to false-positive responses. The major reason behind this is that G proteins are located at regions in membranes close to the receptor whereas b-arrestins are distributed throughout the cytosol. Upon activation of the V2R, barrestins translocate to the receptor at the plasma membrane, which results in enhanced BRET between V2R-coupled G protein subtypes and b-arrestins (see Author response image 1 below of preliminary data). This translocation also results in non-specific BRET signals between b-arrestins and G protein subtypes at the plasma membrane that do not couple to V2R but are located in close proximity to the receptor. As these nonspecific BRET signals do not report on the formation of functional V2R megaplexes (see Author response image 1), we have purposely not used this approach.

Author response image 1.

To overcome this technical hurdle in detection of functional megaplexes, we have replaced full-length G proteins by miniG proteins as the latter are located in the cytosol at resting states and only translocate to the membrane area if a receptor adopts an active conformation. This replacement is advantageous since activation of megaplex-forming receptors such as the V2R results in simultaneous translocation of miniG proteins and b-arrestins from the cytosol to the receptor at the plasma membrane, which produces a highly specific proximity signal (see Author response image 2 below of preliminary data). When stimulating the V2R, we only observe increases in proximity between b-arrestin1 and miniG proteins that are activated by the V2R (miniGs and miniGsq) but not the miniG proteins that are not activated by this receptor (miniGsi and miniG12) (see Author response image 2). Therefore, usage of miniG proteins offers a more accurate experimental approach to detect functional megaplexes as compared to the usage of full-length G proteins.

Author response image 2.

2.2 The interpretation of complementation (NanoLuc) or proximity (BRET) as evidence of signaling is not appropriate, especially when overexpression system and engineered constructs are being used.

We thank the reviewer for raising this concern. We have previously demonstrated global Gas activation and Gas signaling in form of cAMP stimulated by internalized V2R (Thomsen ARB et al. 2016, Cell 166(4):907-19). As mentioned previously, in the current updated manuscript we have now included experiments to document downstream signaling events in response to Gaq/11 activation. These experiments include measurement of production of DAG at the plasma membrane (Fig. 1C) and early endosomes (Fig. 1D), as well as phosphorylation/activation of PKC (Fig. 1E). Pre-incubation with the selective Gaq/11 inhibitor YM-254890, abrogated all these downstream signals and confirms that the V2R stimulates Gaq/11 protein signaling at both the plasma membrane and endosomes (Fig. 1C-E).

2.3 After the original work from the same corresponding authors on megaplex formation, the major challenge in the field is to demonstrate the existence and relevance of megaplex formation at endogenous levels of components, and the current study focuses solely on showing the proximity of Gaq and b-arrestins.

We completely agree with the reviewer that it will be important to demonstrate functionality endogenous megaplexes and we are currently working on this in other studies using different receptor systems. However, doing this is not trivial and we will have to overcome major technical barriers that we feel is somewhat out of the scope of the current study. The goal of our V2R study is to demonstrate that V2R megaplexes form with Gaq/11 resulting to Gaq/11 activation at endosomes, and that endosomal G protein activation by the V2R can occur independently of b-arrestin, which we in our humble opinion accomplish.

2.4 The study lacks a coherent approach, and the assays are often shifted back and forth between the two b-arrestin isoforms (1 and 2), for example, confocal vs. complementation etc.

We understand the reviewer’s concern. However, as opposed to the β2-adrenergic receptor that binds βarrestin2 with higher affinity than β-arrestin1, V2R has a strong affinity for both β-arrestin1 and β-arrestin2 (Oakley et al. 2000, JBC 275(22):17201-10). The V2R’s almost identical affinity for β-arrestin1 and βarrestin2 is well illustrated in Fig. 3B. Thus, although different β-arrestin isoforms were used in some experiments, it is very unlikely that the overall results and conclusions from this study will change by adding extra experiments to ensure that both β-arrestin isoforms are used in every experiment.

2.5 In every assay, only the G proteins and b-arrestins are monitored without a direct assessment of the presence of receptor, and absent that data, it is difficult to justify calling these entities megaplexes.

Mini G proteins and b-arrestin come into close proximity upon agonist stimulation of the V2R. Using confocal microscopy, we observed this co-recruitment of miniGs/miniGsq and b-arrestin in response to prolonged V2R stimulation at endosomes specifically (Fig. 3D-F). In absence of GPCR stimulation, both miniG and b-arrestin would be homogenously distributed throughout the cytosol, and thus, the only reason to why both proteins have been recruited to endosomes in response to AVP challenge is that they are recruited to internalized and active V2R. This point was obviously not adequately described in the original manuscript, and thus, we have now clarified this further in the updated manuscript at the 8th sentence of the last paragraph of the "The V2R recruits Gas/Gaq and barrs simultaneously" section.

REVIEWER #3:

The manuscript by Daly et al. examines endosomal signaling of the vasopressin type 2 receptors using engineered mini G protein (mG proteins) and a number of novel techniques to address if sustained G protein signaling in the endosomal compartment is enhanced by b-arrestin. Employing these interesting techniques they have how V2R could activates Gas and Gaq in the endosomal compartments and how this modulation could occur in arrestin-dependent and -independent manner. Although the phenomenon of endosomal signaling is complex to address the authors have tried their best to examine these using a number of well controlled set of experiments. Though this is an interesting and well carried out study of endosomal signaling of G proteins, my concerns are:

3.1 The study is done in overexpressed HEK 293 cells with these engineered constructs making me wonder if the kinetics would be the same in primary cells?

The reviewer raises an interesting and valid point. It is possible that in the context of primary cells the kinetic would differ slightly and it would definitely be interesting to address this in a subsequent study. However, despite being an interesting aspect of our study, the kinetic itself is not our major take home message, but rather the subcellular localization of the G protein activation and the role of β-arrestin in these events. We have now highlighted this aspect in our updated manuscript (1st paragraph of the discussion) and we thank the reviewer for addressing this.

3.2 The use of the phrase "G protein activation independent of b-arrestins to a minor degree" would make me question its physiological relevance. The authors should discuss the relevance of their findings in physiological or pathological context.

We are glad that the reviewer focuses on this point, and we would like to highlight that other GPCRs including the glucagon-like peptide-1 receptor (GLP1R) internalizes in a β-arrestin-independent manner (Claing A et al. 2000 PNAS 97(3):1119-24), while signaling through Gas from endosomes. In the case of the GLP1R, this endosomal Gas signaling promotes glucose-stimulated insulin secretion in pancreatic βcells (Kuna RS et al. 2013 Am J Physiol Endocrinol Metab 305:E161-70). Consequently, β-arrestinindependent endosomal G protein signaling appears to have some physiological relevance. Similarly, in a very recent pre-print from the von Zastrow group (Blythe EE and von Zastrow M 2023 BioRxiv https://doi.org/10.1101/2022.09.07.506997), it was reported that endogenously-expressed vasoactive intestinal peptide receptor 1 (VIPR1), which regulates gastro-intestinal functions, promotes robust G protein signaling from endosomes in a completely β-arrestin-independent fashion. This again suggest that endogenously expressed GPCRs can internalize and activate G proteins from endosomes independently from β-arrestin to produce physiological responses. We have now discussed about these studies in the 6th paragraph of the discussion.

3.3 The confocal colocalization studies shown in Figure 2 and their conclusion "suggesting a certain level of endosomal Gas/Gaq signaling despite the absence of barr2" seems rather inconclusive.

As opposed to V2R a receptor that retains β-arrestin in endosomes upon internalization, β-arrestin quickly dissociates from V2b2AR after internalization due to the low affinity of the carboxy-terminal of β2AR for βarrestin. In the previous Fig. 2 (now Fig. 3), after 45 minutes of AVP stimulation, no β-arrestin is visible at endosomes in cells expressing V2b2AR as β-arrestin has already dissociated from the receptor and translocated back to the cytosol. However, clear green clusters of mGs and mGsq are still visible at endosomes indicating the presence of active receptor interacting with Gas or Gaq despite the fact that βarrestin is back to the cytosol. We quantified the percentage of the green mGs or mGsq clusters that do not colocalize with β-arrestin and have added this information to the updated version of the manuscript (Fig. 3G). In V2R-expressing cells, almost all active receptors that interact with Gas or Gaq/11 also associate with β-arrestin (Fig. 3G). In contrast, in V2b2AR-expressing cells, approximately 75% of the active receptors do not interact with β-arrestin (Fig. 3G). This suggests that β-arrestin binding to V2R is not an absolute requirement for endosomal Gas and Gaq activation by V2R. This point was obviously not addressed adequately in the original manuscript, and thus, we have now elaborated further on this in the updated version in the last paragraph of the "The V2R recruits Gas/Gaq and βarrs simultaneously" section.

3.4 Though a novel observation it is not clear to me how V2R would internalize after activation without arrestin. Is it some sort of generalized microcytosis occurring in these overexpressed cells? Should discuss.

This is certainly a very interesting observation and something other research laboratories also have seen recently – in particular, in context to endosomal G protein signaling (Blythe EE and von Zastrow M 2023 BioRxiv https://doi.org/10.1101/2022.09.07.506997). The main and best characterized pathway for GPCR internalization is clathrin-dependent where receptors most commonly are associated with β-arrestins. However, for some GPCRs, the β-arrestin association is not required for clathrin-mediated internalization. One example is the apelin receptor that can internalize via clathrin-coated pits, but in β-arrestinindependent manner (Pope GR et al. 2016 Moll Cell Endocrinol. 437:108-19). Alternatively, GPCRs can also internalize independently of any clathrin and β-arrestin associations via caveolae or fast endophilinmediated endocytosis (FEME). We have now expanded our discussion of possible mechanisms for βarrestin-independent receptor internalization in the updated manuscript in the 6th paragraph of the discussion, and we thank the reviewer for the suggestion.

3.5 Is use of mini G protein a good representation? The authors should justify.

Excellent point and something we have comprehensively discussed in our response to reviewer 1 and 2 (points 1.2 and 2.1).

Author Response

Reviewer #1 (Public Review):

Like the "preceding" co-submitted paper, this is again a very strong and interesting paper in which the authors address a question that is raised by the finding in their co-submitted paper - how does one factor induce two different fates. The authors provide an extremely satisfying answer - only one subset of the cells neighbors a source of signaling cells that trigger that subset to adopt a specific fate. The signal here is Delta and the read-out is Notch, whose intracellular domain, in conjunction with, presumably, SuH cooperates with Bsh to distinguish L4 from L5 fate (L5 is not neighbored by signal-providing cells). Like the back-to-back paper, the data is rigorous, well-presented and presents important conclusions. There's a wealth of data on the different functions of Notch (with and without Bsh). All very satisfying.

Thanks!

I have again one suggestion that the authors may want to consider discussing. I'm wondering whether the open chromatin that the author convincingly measure is the CAUSE or the CONSEQUENCE of Bsh being able to activate L4 target genes. What I mean by this is that currently the authors seem to be focused on a somewhat sequential model where Notch signaling opens chromatin and this then enables Bsh to activate a specific set of target genes. But isn't it equally possible that the combined activity of Bsh/Notch(intra)/SuH opens chromatin? That's not a semantic/minor difference, it's a fundamentally different mechanism, I would think. This mechanism also solves the conundrum of specificity - how does Notch know which genes to "open" up? It would seem more intuitive to me to think that it's working together with Bsh to open up chromatin, with chromatin accessibility than being a "mere" secondary consequence. If I'm not overlooking something fundamental here, there is actually also a way to distinguish between these models - test chromatin accessibility in a Bsh mutant. If the author's model is true, chromatin accessibility should be unchanged.

I again finish by commending the authors for this terrific piece of work.

Thanks! It is a crucial question whether Notch signaling regulates chromatin landscape independently of a primary HDTF. We will include this discussion in the text and pursue it in our next project. We think Notch signaling may regulate chromatin accessibility independently of a primary HDTF based on our observation: in larval ventral nerve cord, all motor neurons are NotchON neurons while all sensory neurons are NotchOFF neurons; NotchON neurons share similar functional properties, despite expressing distinct HDTFs, possibly due to the common chromatin landscape regulated by Notch signaling.

Reviewer #2 (Public Review):

Summary:

In this work, the authors explore how Notch activity acts together with Bsh homeodomain transcription factors to establish L4 and L5 fates in the lamina of the visual system of Drosophila. They propose a model in which differential Notch activity generates different chromatin landscapes in presumptive L4 and L5, allowing the differential binding of the primary homeodomain TF Bsh (as described in the co-submitted paper), which in turn activates downstream genes specific to either neuronal type. The requirement of Notch for L4 vs. L5 fate is well supported, and complete transformation from one cell type into the other is observed when altering Notch activity. However, the role of Notch in creating differential chromatin landscapes is not directly demonstrated. It is only based on correlation, but it remains a plausible and intriguing hypothesis.

Thanks for the positive feedback!

Strengths:

The authors are successful in characterizing the role of Notch to distinguish between L4 and L5 cell fates. They show that the Notch pathway is active in L4 but not in L5. They identify L1, the neuron adjacent to L4 as expressing the Delta ligand, therefore being the potential source for Notch activation in L4. Moreover, the manuscript shows molecular and morphological/connectivity transformations from one cell type into the other when Notch activity is manipulated.

Thanks!

Using DamID, the authors characterize the chromatin landscape of L4 and L5 neurons. They show that Bsh occupies distinct loci in each cell type. This supports their model that Bsh acts as a primary selector gene in L4/L5 that activates different target genes in L4 vs L5 based on the differential availability of open chromatin loci.

Thanks!

Overall, the manuscript presents an interesting example of how Notch activity cooperates with TF expression to generate diverging cell fates. Together with the accompanying paper, it helps thoroughly describe how lamina cell types L4 and L5 are specified and provides an interesting hypothesis for the role of Notch and Bsh in increasing neuronal diversity in the lamina during evolution.

Thanks for the positive feedback on both manuscripts.

Weaknesses:

Differential Notch activity in L4 and L5:

● The manuscript focuses its attention on describing Notch activity in L4 vs L5 neurons. However, from the data presented, it is very likely that the pool of progenitors (LPCs) is already subdivided into at least two types of progenitors that will rise to L4 and L5, respectively. Evidence to support this is the activity of E(spl)-mɣ-GFP and the Dl puncta observed in the LPC region. Discussion should naturally follow that Notch-induced differences in L4/L5 might preexist L1-expressed Dl that affect newborn L4/L5. Therefore, the differences between L4 and L5 fates might be established earlier than discussed in the paper. The authors should acknowledge this possibility and discuss it in their model.

We agree. Historically, LPCs are thought to be homogenous; our data suggests otherwise. We now emphasize this in the Discussion as requested. We are also investigating this question using single cell RNAseq on LPCs to look for molecular heterogeneities. Thanks for the great comment!

● The authors claim that Notch activation is caused by L1-expressed Delta. However, they use an LPC driver to knock down Dl. Dl-KD should be performed exclusively in L1, and the fate of L4 should be assessed.

Dl is transiently expressed in newborn L1 neurons. To knock down Dl in L1, we need to express Dl-RNAi before Dl protein is expressed in newborn L1; the only known Gal4 line expressed that early is the LPC-Gal4 that we used. There is no L1-gal4 line expressed early enough to eliminate L1 expression of Dl.

● To test whether L4 neurons are derived from NotchON LPCs, I suggest performing MARCM clones in early pupa with an E(spl)-mɣ-GFP reporter.

We agree! Whether L4 neurons are derived from NotchON LPCs is a great question. However, MARCM clones in early pupa with an E(spl)-mɣ-GFP reporter will not work because E(spl)-mɣ-GFP reporter is only expressed in LPCs but not lamina neurons. We now mention this in the Discussion.

● The expression of different Notch targets in LPCs and L4 neurons may be further explored. I suggest using different Notch-activity reporters (i.e., E(spl)-GFP reporters) to further characterize these. differences. What cause the switch in Notch target expression from LPCs to L4 neurons should be a topic of discussion.

Thanks! It is a great question why Notch induces Espl-mɣ in LPCs but Hey in new-born neurons. However, it is not the question we are tackling in this paper and it will be a great direction to pursue in future. We will add this to our Discussion.

Notch role in establishing L4 vs L5 fates:

● The authors describe that 27G05-Gal4 causes a partial Notch Gain of Function caused by its genomic location between Notch target genes. However, this is not further elaborated. The use of this driver is especially problematic when performing Notch KD, as many of the resulting neurons express Ap, and therefore have some features of L4 neurons. Therefore, Pdm3+/Ap+ cells should always be counted as intermediate L4/L5 fate (i.e., Fig3 E-J, Fig3-Sup2), irrespective of what the mechanistic explanation for Ap activation might be. It's not accurate to assume their L5 identity. In Fig4 intermediate-fate cells are correctly counted as such.

Thanks for the comment! We will annotate Pdm3/Ap+ as L4/L5 fate in the corresponding figures.

● Lines 170-173: The temporal requirement for Notch activity in L5-to-L4 transformation is not clearly delineated. In Fig4-figure supplement 1D-E, it is not stated if the shift to 29{degree sign}C is performed as in Fig4-figure supplement 1A-C.

Thank you for catching this. We will correct it in the text.

● Additionally, using the same approach, it would be interesting to explore the window of competence for Notch-induced L5-to-L4 transformation: at which point in L5 maturation can fate no longer be changed by Notch GoF?

Our data show that Bsh with Notch signaling in newborn neurons specifies L4 fate while Bsh without Notch signaling in newborn neurons specifies L5 fate. Therefore, we think the window of fate competence is during newborn neurons. We will include the data to support this.

L4-to-L3 conversion in the absence of Bsh

● Although interesting, the L4-to-L3 conversion in the absence of Bsh is never shown to be dependent on Notch activity. Importantly, L3 NotchON status is assumed based on their position next to Dl-expressing L1, but it is not empirically tested. Perhaps screening Notch target reporter expression in the lamina, as suggested above, could inform this issue.

Our data show that the L4-to-L3 conversion in the absence of Bsh and in the presence of Notch activity while the L5-to-L1 conversion in the absence of Bsh and in the absence of Notch activity. Therefore, Notch activity is necessary for the L4-to-L3 conversion. Unfortunately, currently we only have Hey as an available Notch target reporter in new-born neurons. To tackle this challenge in the future, we will profile the genome-binding targets of endogenous Notch in newborn neurons. This will identify novel genes as Notch signaling reporters in neurons for the field.

● Otherwise, the analysis of Bsh Loss of Function in L4 might be better suited to be included in the accompanying manuscript that specifically deals with the role of Bsh as a selector gene for L4 and L5.

That is an interesting suggestion, but without knowing that Bsh + Notch = L4 identity the experiment would be hard to interpret. Note that we took advantage of Notch signaling to trace the cell fate in the absence of Bsh and found the L4-to-L3 conversion (see Figure 5G-K).

Different chromatin landscape in L4 and L5 neurons

● A major concern is that, although L4 and L5 neurons are shown to present different chromatin landscapes (as expected for different neuronal types), it is not demonstrated that this is caused by Notch activity. The paper proves unambiguously that Notch activity, in concert with Bsh, causes the fate choice between L4 and L5. However, that this is caused by Notch creating a differential chromatin landscape is based only in correlation. (NotchON cells having a different profile than NotchOFF). Although the authors are careful not to claim that differential chromatin opening is caused directly by Notch, this is heavily suggested throughout the text and must be toned down.e.g.: Line 294: "With Notch signaling, L4 neurons generate distinct open chromatin landscape" and Line 298: "Our findings propose a model that the unique combination of HDTF and open chromatin landscape (e.g. by Notch signaling)" . These claims are not supported well enough, and alternative hypotheses should be provided in the discussion. An alternative hypothesis could be that LPCs are already specified towards L4 and L5 fates. In this context, different early Bsh targets in each cell type could play a pioneer role generating a differential chromatin landscape.

We agree and appreciate the comment, it is well justified. We have toned down our comments and clearly state that this is a correlation that needs to be tested for a causal relationship. Thank you for requesting it!

● The correlation between open chromatin and Bsh loci with Differentially Expressed genes is much higher for L4 than L5. It is not clear why this is the case, and should be discussed further by the authors.

We agree, and think in L5 neurons, the secondary HDTF Pdm3 also contributes to L5 specific gene transcription during synaptogenesis window, in addition to Bsh. We will include this in the text.

Author Response

Reviewer #1 (Public Review):

In this very strong and interesting paper the authors present a convincing series of experiments that reveal molecular mechanism of neuronal cell type diversification in the nervous system of Drosophila. The authors show that a homeodomain transcription factor, Bsh, fulfills several critical functions - repressing an alternative fate and inducing downstream homeodomain transcription factors with whom Bsh may collaborate to induce L4 and L5 fates (the author's accompanying paper reveals how Bsh can induce two distinct fates). The authors make elegant use of powerful genetic tools and an arsenal of satisfying cell identity markers.

Thanks!

I believe that this is an important study because it provides some fundamental insights into the conservation of neuronal diversification programs. It is very satisfying to see that similar organizational principles apply in different organisms to generate cell type diversity. The authors should also be commended for contextualizing their work very well, giving a broad, scholarly background to the problem of neuronal cell type diversification.

Thanks!

My one suggestion for the authors is to perhaps address in the Discussion (or experimentally address if they wish) how they reconcile that Bsh is on the one hand: (a) continuously expressed in L4/L4, (b) binding directly to a cohort of terminal effectors that are also continuously expressed but then, on the other hand, is not required for their maintaining L4 fate? A few questions: Is Bsh only NOT required for maintaining Ap expression or is it also NOT required for maintaining other terminal markers of L4? The former could be easily explained - Bsh simply kicks of Ap, Ap then autoregulates, but Bsh and Ap then continuously activate terminal effector genes. The second scenario would require a little more complex mechanism: Bsh binding of targets (with Notch) may open chromatin, but then once that's done, Bsh is no longer needed and Ap alone can continue to express genes. I feel that the authors should be at least discussing this. The postmitotic Bsh removal experiment in which they only checked Ap and depression of other markers is a little unsatisfying without further discussion (or experiments, such as testing terminal L4 markers). I hasten to add that this comment does not take away from my overall appreciation for the depth and quality of the data and the importance of their conclusions.

Great suggestions, we will discuss these two hypotheses as requested.

Bsh initiates Ap expression in L4 neurons which then maintain Ap expression independently of Bsh expression, likely through Ap autoregulation. During the synaptogenesis window, Ap expression becomes independent from Bsh expression, but Bsh and Ap are both still required to activate the synapse recognition molecule DIP-beta. Additionally, Bsh also shows putative binding to other L4 identity genes, e.g., those required for neurotransmitter choice, and electrophysiological properties, suggesting Bsh may initiates L4 identity genes as a suite of genes. The mechanism of maintaining identity features (e.g., morphology, synaptic connectivity and functional properties) in the adult remains poorly understood. It is a great question whether primary HDTF Bsh maintains the expression of L4 identity genes in the adult. To test this, in our next project, we will specifically knock out Bsh in L4 neurons of the adult fly and examine the effect on L4 morphology, connectivity and function properties.

Reviewer #2 (Public Review):

Summary:

In this paper, the authors explore the role of the Homeodomain Transcription Factor Bsh in the specification of Lamina neuronal types in the optic lobe of Drosophila. Using the framework of terminal selector genes and compelling data, they investigate whether the same factor that establishes early cell identity is responsible for the acquisition of terminal features of the neuron (i.e., cell connectivity and synaptogenesis).

Thanks for the positive words!

The authors convincingly describe the sequential expression and activity of Bsh, termed here as 'primary HDTF', and of Ap in L4 or Pdm3 in L5 as 'secondary HDTFs' during the specification of these two neurons. The study demonstrates the requirement of Bsh to activate either Ap and Pdm3, and therefore to generate the L4 and L5 fates. Moreover, the authors show that in the absence of Bsh, L4 and L5 fates are transformed into a L1 or L3-like fates.

Thanks!

Finally, the authors used DamID and Bsh:DamID to profile the open chromatin signature and the Bsh binding sites in L4 neurons at the synaptogenesis stage. This allows the identification of putative Bsh target genes in L4, many of which were also found to be upregulated in L4 in a previous single-cell transcriptomic analysis. Among these genes, the paper focuses on Dip-β, a known regulator of L4 connectivity. They demonstrate that both Bsh and Ap are required for Dip-β, forming a feed-forward loop. Indeed, the loss of Bsh causes abnormal L4 synaptogenesis and therefore defects in several visual behaviors. The authors also propose the intriguing hypothesis that the expression of Bsh expanded the diversity of Lamina neurons from a 3 cell-type state to the current 5 cell-type state in the optic lobe.

Thanks for the excellent summary of our findings!

Strengths:

Overall, this work presents a beautiful practical example of the framework of terminal selectors: Bsh acts hierarchically with Ap or Pdm3 to establish the L4 or L5 cell fates and, at least in L4, participates in the expression of terminal features of the neuron (i.e., synaptogenesis through Dip-β regulation).

Thanks!

The hierarchical interactions among Bsh and the activation of Ap and Pdm3 expression in L4 and L5, respectively, are well established experimentally. Using different genetic drivers, the authors show a window of competence during L4 neuron specification during which Bsh activates Ap expression. Later, as the neuron matures, Ap becomes independent of Bsh. This allows the authors to propose a coherent and well-supported model in which Bsh acts as a 'primary' selector that activates the expression of L4-specific (Ap) and L5-specific (Pdm3) 'secondary' selector genes, that together establish neuronal fate.

Thanks again!

Importantly, the authors describe a striking cell fate change when Bsh is knocked down from L4/L5 progenitor cells. In such cases, L1 and L3 neurons are generated at the expense of L4 and L5. The paper demonstrates that Bsh in L4/L5 represses Zfh1, which in turn acts as the primary selector for L1/L3 fates. These results point to a model where the acquisition of Bsh during evolution might have provided the grounds for the generation of new cell types, L4 and L5, expanding lamina neuronal diversity for a more refined visual behaviors in flies. This is an intriguing and novel hypothesis that should be tested from an evo-devo standpoint, for instance by identifying a species when L4 and L5 do not exist and/or Bsh is not expressed in L neurons.

Thanks for the appreciation of our findings!

To gain insight into how Bsh regulates neuronal fate and terminal features, the authors have profiled the open chromatin landscape and Bsh binding sites in L4 neurons at mid-pupation using the DamID technique. The paper describes a number of genes that have Bsh binding peaks in their regulatory regions and that are differentially expressed in L4 neurons, based on available scRNAseq data. Although the manuscript does not explore this candidate list in depth, many of these genes belong to classes that might explain terminal features of L4 neurons, such as neurotransmitter identity, neuropeptides or cytoskeletal regulators. Interestingly, one of these upregulated genes with a Bsh peak is Dip-β, an immunoglobulin superfamily protein that has been described by previous work from the author's lab to be relevant to establish L4 proper connectivity. This work proves that Bsh and Ap work in a feed-forward loop to regulate Dip-β expression, and therefore to establish normal L4 synapses. Furthermore, Bsh loss of function in L4 causes impairs visual behaviors.

Thanks for the excellent summary of our findings.

Weaknesses:

● The last paragraph of the introduction is written using rhetorical questions and does not read well. I suggest rewriting it in a more conventional direct style to improve readability.

We agree, and will update the text as suggested.

● A significant concern is the way in which information is conveyed in the Figures. Throughout the paper, understanding of the experimental results is hindered by the lack of information in the Figure headers. Specifically, the genetic driver used for each panel should be adequately noted, together with the age of the brain and the experimental condition. For example, R27G05-Gal4 drives early expression in LPCs and L4/L5, while the 31C06-AD, 34G07-DBD Split-Gal4 combination drives expression in older L4 neurons, and the use of one or the other to drive Bsh-KD has dramatic differences in Ap expression. The indication of the driver used in each panel will facilitate the reader's grasp of the experimental results.

We agree, and will update the figure annotation.

● Bsh role in L4/L5 cell fate:

o It is not clear whether Tll+/Bsh+ LPCs are the precursors of L4/L5. Morphologically, these cells sit very close to L5, but are much more distant from L4.

Our current data show L4 and L5 neurons are generated by different LPCs. However, currently we don’t have tools to demonstrate which subset of LPCs generate which lamina neuron type. We are currently working on a followup manuscript on LPC heterogeneity, but those experiments have just barely been started.

o Somatic CRISPR knockout of Bsh seems to have a weaker phenotype than the knockdown using RNAi. However, in several experiments down the line, the authors use CRISPR-KO rather than RNAi to knock down Bsh activity: it should be explained why the authors made this decision. Alternatively, a null mutant could be used to consolidate the loss of function phenotype, although this is not strictly necessary given that the RNAi is highly efficient and almost completely abolishes Bsh protein.

The reason we chose CRISPR-KO (L4-specific Gal4, uas-Cas9, and uas-Bsh-sgRNAs) is that it effectively removed Bsh expression from majority of L4 neurons. However, it failed to knock down Bsh in L4 neurons using L4-split Gal4 and Bsh-RNAi because L4-split Gal4 expression depends on Bsh. We will include this explanation in the text.

o Line 102: Rephrase "R27G05-Gal4 is expressed in all LPCs and turned off in lamina neurons" to "is turned off as lamina neurons mature", as it is kept on for a significant amount of time after the neurons have already been specified.

Thanks; we will make that change.

o Line 121: "(a) that all known lamina neuron markers become independent of Bsh regulation in neurons" is not an accurate statement, as the markers tested were not shown to be dependent on Bsh in the first place.

Good point. We will rephrase it as “that all known lamina neuron markers are independent of Bsh regulation in neurons”.

o Lines 129-134: Make explicit that the LPC-Gal4 was used in this experiment. This is especially important here, as these results are opposite to the Bsh Loss of Function in L4 neurons described in the previous section. This will help clarify the window of competence in which Bsh establishes L4/L5 neuronal identities through ap/pdm3 expression.

Thanks! We will include Gal4 information in the text for every manipulation.

● DamID and Bsh binding profile:

○ Figure 5 - figure supplement 1C-E: The genotype of the Control in (C) has to be described within the panel. As it is, it can be confused with a wild type brain, when it is in fact a Bsh-KO mutant.

Great point! Thank you for catching this and we will update it.

○ It Is not clear how L4-specific Differentially Expressed Genes were found. Are these genes DEG between Lamina neurons types, or are they upregulated genes with respect to all neuronal clusters? If the latter is the case, it could explain the discrepancy between scRNAseq DEGs and Bsh peaks in L4 neurons.

We did not use “L4-specific Differentially Expressed Genes”. Instead, we used all genes that are significantly transcribed in L4 neurons (line 209-210).

● Dip-β regulation:

○ Line 234: It is not clear why CRISPR KO is used in this case, when Bsh-RNAi presents a stronger phenotype.

As we explained it above, the reason we chose CRISPR-KO (L4-specific Gal4, uas-Cas9, and uas-Bsh-sgRNAs) is that it effectively removed Bsh expression from majority of L4 neurons. However, it failed to knock down Bsh in L4 neurons using L4-split Gal4 and Bsh-RNAi because L4-split Gal4 expression depends on Bsh. We’ll include this explanation in the text.

○ Figure 6N-R shows results using LPC-Gal4. It is not clear why this driver was used, as it makes a less accurate comparison with the other panels in the figure, which use L4-Split-Gal4. This discrepancy should be acknowledged and explained, or the experiment repeated with L4-Split-Gal4>Ap-RNAi.

I think you mean 6J-M shows results using LPC-Gal4. We first tried L4-Split-Gal4>Ap-RNAi but it failed to knock down Ap because L4-Split-Gal4 expression depends on Ap. We will add this to the text.

○ Line 271: It is also possible that L4 activity is dispensable for motion detection and only L5 is required.

Thanks! Work from Tuthill et al, 2013 showed that L5 is not required for any motion detection. We will include this citation in the text.

● Discussion: It is necessary to de-emphasize the relevance of HDTFs, or at least acknowledge that other, non-homeodomain TFs, can act as selector genes to determine neuronal identity. By restricting the discussion to HDTFs, it is not mentioned that other classes of TFs could follow the same Primary-Secondary selector activation logic.

That is a great point, thank you! We will include this in the discussion.

Author Response

The following is the authors’ response to the original reviews.

This important study shows that two methods of sleep induction in the fly, optogenetically activation of the dorsal fan-shaped body (which is rapidly reversible and maintains a neuronal activity signature similar to wakefulness), and Gaboxadol-induced sleep (which shuts down neuronal activity), produce distinct forms of sleep and have different effects on brain-wide neural activity. The majority of the conclusions of the paper are supported by compelling data, but the evidence supporting the claim that the two interventions trigger distinct transcriptional responses is incomplete.

Thank you for the helpful and detailed reviews. We feel that these have improved the manuscript considerably, and hopefully the additional figures in this Reply letter will help further convince our readers.

Public Review

In this study, Anthoney and coworkers continue an important, unique, and technologically innovative line of inquiry from the van Swinderen lab aimed at furthering our understanding of the different sleep stages that may exist in Drosophila. Here, they compare the physiological and transcriptional hallmarks of sleep that have been induced by two distinct means, a pharmacological block of GABA signaling and optogenetic activation of dorsal fan-shaped-body neurons. They first employ an incredibly impressive fly-on-the-ball 2-photon functional imaging setup to monitor neural activity during these interventions, and then perform bulk RNA sequencing of fly brains at different stages. These transcriptomic analyses leads them to (a) knocking out nicotinic acetyl-choline receptor subunits and (b) knocking down AkhR throughout the fly brain testing the impact of these genetic interventions on sleep behaviors in flies. Based on this work, the authors present evidence that optogenetically and pharmacologically induced sleep produces highly distinct brain-wide effects on physiology and transcription. The study is of significant interest, is easy to read, and the figures are mostly informative. However there are features of the experimental design and the interpretation of results that diminish enthusiasm.

a- Conditions under which sleep is induced for behavioral vs neural and transcriptional studies

1- There is a major conceptual concern regarding the relationships between the physiological and transcriptomic effects of optogenetic and pharmacological sleep promotion, and the effects that these manipulations have on sleep behavior. The authors show that these two means of sleep-induction produce remarkably distinct physiological and transcriptional responses, however, they also show that they produce highly similar effects on sleep behavior, causing an increase in sleep through increases in the duration of sleep bouts. If dFB neurons were promoting active sleep, the sleep it produces should be more fragmented than the sleep induced by the drug, because the latter is supposed to produce quiet sleep. Yet both manipulations seem to be biasing behavior toward quiet sleep.

This is a correct observation, which is already evident in our sleep architecture data (Figure 2E-H): chronic optogenetic sleep induction promotes longer sleep bouts that are similar in structure (bout number vs bout duration) to those produced by THIP feeding. Since our plots in Figure 2E-H follow the 5min sleep criterion cutoff, upon the Reviewer’s advice we re-analyzed our optogenetic experiments for short (1-5min) sleep. These are graphed below in Author response image 1. As can be seen, and as suspected by the Reviewer, the optogenetic manipulation does not increase the total amount of short sleep; indeed, it decreases it compared to baseline (these are for the exact same data as in Figure 2). Optogenetic sleep induction does not create a bunch of short sleep bouts.

Author response image 1.

Short sleep in optogenetic experiments. A. Average baseline (±SEM) 1-5min sleep across a day and night. B. Average (±SEM) 1-5min sleep in optogenenetically-activated flies, across a day and night.

We agree with the reviewer that this observation might seem inconsistent with the idea that optogenetic activation promotes active sleep, and that short sleep is active sleep. However, it does not necessarily follow that optogenetic activation has to produce short sleep. Indeed, we know from our brain imaging data (and the associated behavioral analysis) that active sleep will persist for as long as we induce it with red light. While we have not induced it for longer than 15 minutes (Tainton-Heap et al, Current Biology, 2021; Troup et al, J. of Neuroscience, 2023), this is already clearly longer than a <5min sleep bout. So our interpretation is that the longer sleep bouts induced by optogenetic activation are prolonged active sleep, rather than quiet sleep. In other words, this artificial sleep manipulation induces prolonged active sleep, rather than many short sleep bouts. This is of course different than what happens during spontaneous sleep. We have tried to be clearer about sleep bout durations in the revised manuscript (e.g., the new Figure 3), and we now admit early in the results (lines 376-380) that that we don’t know what optogenetic activation looks like in the fly brain beyond 15 minutes.

2- The authors show that the pharmacological block of GABA signaling and the optogenetic activation of dorsal fan-shaped-body neurons cause different responses on brain activity. Based on these recordings and the behavioral and brain transcriptomic data they then claim that these responses correspond to different sleep states and are associated with the expression and repression of a different constellation of genes. Nevertheless, neural activity in animals was recorded following short stimulations whereas behavioral and transcriptomic data were obtained following chronic stimulation. In this regard, it would be interesting to determine how the 12-hour pharmacological intervention they employed for their transcriptomic analysis changes neural activity throughout the brain - 12 hours will likely be too long for the open-cuticle preps, but an in-between time-point (e.g. 1h) would probably be equally informative.

The longest we’ve imaged brain activity for optogenetic sleep induction is 15 minutes, as discussed above. We see no changes in activity across this time, which would normally have led to a quiet sleep stage in spontaneous sleep recordings. Whole-brain imaging after 10 hours of optogenetic sleep induction (our RNA collection timepoint) is not realistic, and even 1 hour is difficult. We have however conducted overnight electrophysiological recordings (with multichannel silicon probes), where we activated the same R23E10 neurons for successive 20-minute bouts (alternating with 20min of no red light). We are preparing this work for publication (Van De Poll, et al). We see no evidence of optogenetic activation of this circuit ever producing anything resembling quiet sleep. Since we are not in a position to provide this new electrophysiological data in the current study, we are careful to clarify that we have not investigated what brain imaging looks like after chronic optogenetic activation (lines 376-380). We are showing through diverse lines of evidence that what is called sleep can look different in flies.

b- Efficiency of THIP treatment under different conditions

1- There are no data to quantify how THIP alters food consumption. It is evident that flies consume it otherwise they would not show increased sleep. However, they may consume different amounts of food overall than the minus THIP controls. This might have an influence on the animal's metabolism, which could at least explain the fact that metabolism-related genes are regulated (Figure 5). Therefore, in the current state, it is not possible to be certain that gene regulation events measured in this experiment are solely due to THIP effects on sleep.

We have two arguments against this reasonable criticism. First, as discussed above, the optogenetic flies are sleeping at least as much as the THIP-fed flies, so in principle they also might be feeding less. But we see no metabolic gene downregulation in the optogenetic dataset. We include this counterargument in the discussion (lines 752-756). Then, together with our co-author Paul Shaw we have shown that THIP-fed flies are not eating less compared to controls (Dissel et al, Current Biology, 2015), by tracking dye consumption. We show those results again below in Author response image 2 to support our reasoning that feeding is not an issue.

Author response image 2.

Flies were fed blue dye in their food while being sleep deprived (SD), or while being induced to sleep with 0.1mg/ml THIP in their food, or both. Dye consumption was measured in triplicate for pooled groups of 16 flies. Average absorbance at 625nm (±stan dev) is shown. Experiments were not significantly different (ANOVA of means).

2- A similar problem exists in the sleep deprivation experiments. If flies are snapped every 20 seconds, they may not have the freedom to consume appropriate amounts of food, and therefore their consumption of THIP or ATR may be smaller than in non-sleep deprived controls. Thus, it would be crucial to know whether the flies that are sleep-deprived (i.e. shaken every 20 seconds for 12 hours) actually consume comparable amounts of food (and therefore THIP) as those that are undisturbed. If not, then perhaps the transcriptional differences between the two groups are not sleep-specific, but instead reflect varying degrees of exposure to THIP.

Please see our response to the similar critique above, and how Figure R2 addresses this concern.

3- The authors should further discuss the slow action of THIP perfusion vs dFB activation, especially as flies only seem to fall asleep several minutes after THIP is being washed away. Is it a technical artifact? If not, it may not be unreasonable to hypothesize that THIP, at the concentration used, could prevent flies from falling asleep, and that its removal may lower the concentration to a point that allows its sleep-promoting action. The authors could easily test this by extending THIP treatment for another 4-5 minutes.

The reviewer is partially correct in suggesting a technical artifact: THIP does not get washed away immediately after 5min of perfusion. The drip system we employ means that THIP concentration will slowly increase to the maximum concentration of 0.2mg/ml, and then slowly get diluted away at a rate of 1.25ml/minute (this is all in the Methods). In a previous study (Yap et al, Nature Communications, 2017) we used this exact same perfusion procedure to test a range of THIP concentrations, and settled on 0.2mg/ml as the lowest that reliably induced quiet sleep within 5 minutes. Higher concentrations induced quiet sleep faster, so the alternate explanation proposed by the Reviewer is not supported. We feel that our previous electrophysiological study provided the necessary groundwork for using the same approach and dosage here for our whole-brain imaging readout.

c- Comments regarding the behavioral assays

1- L319-322: the authors conclude that dFB stimulation and THIP consumption have similar behavioral effects on sleep. However, this is inaccurate as in Figure S1 they explain that one increases bout number in both day and night and the other one only during the day.

We have now added a caveat about night bout architecture being different (lines 353-356). Figure S1 is now Figure 3.

2- The behavioral definitions used for active and quiet sleep do not fit well with strong evidence that deep sleep (defined by lowered metabolic rates) is probably most closely associated with bouts of inactivity that are much longer than the >5min duration used here, i.e., probably 30min and longer (Stahl et al. 2017 Sleep 40: zsx084). Given that the authors are providing evidence that quiet sleep is correlated with changes in the expression of metabolism related genes, they should at least discuss the fact that reductions in metabolism have been shown to occur after relatively long bouts of inactivity and might reconsider their behavioral sleep analysis (i.e., their criteria for sleep state) with this in mind.

Interestingly, induced sleep bout durations are on average longer for the optogenetic manipulation (40min vs 25min); this was evident in Figure S1C vs S1F (now Figure 3). So as discussed above, this provides a counterargument for sleep bout duration alone being indicative of metabolic processes associated with quiet sleep: the optogenetic dataset did not uncover metabolic-related pathways as relevant to that sleep manipulation. We refer to Stahl et al, Sleep, 2017, in our discussion (lines 748-750), making exactly this point about metabolic rates being decreased in longer sleep bouts, and flowing up with our observation that optogenetic flies sleep just as much, and their bouts are actually longer. So clearly different processes must be involved.

d- Comments regarding the recordings of neuronal activity

1- There is an additional concern regarding the proposed active and quiet sleep states that rest at the heart of this study. Here these two states in the fly are compared to the REM and NREM sleep states observed in mammals and the parallels between active fly sleep and REM and quiet fly sleep and NREM provide the framework for the study. The establishment of such parallel sleep states in the fly is highly significant and identifying the physiological and molecular correlates of distinct sleep stages in the fly is of critical importance to the field. However, the proposal that the dorsal fan shaped body (dFB) neurons promote active sleep runs counter to the prevailing model that these neurons act as a major site of sleep homeostasis. If quiet sleep were akin to NREM, wouldn't we expect the major site of sleep homeostasis in the brain to promote it? Furthermore, the authors state that the effects of dFB neuron excitation on transcription have "almost no overlap" (line 500) with the transcriptomic effects of sleep deprivation (Supplementary Table 3), which is not what would be expected if dFB neurons are tracking sleep pressure and promoting sleep, as suggested by a growing body of convergent work summarized on page four of the manuscript. Wouldn't the 10h excitation of the dFB neurons be predicted to mimic the effects of sleep deprivation if these neurons "...serve as the discharge circuit for the insect's sleep homeostat..." (line 60)? Shouldn't their prolonged excitation produce an artificial increase in sleep drive (even during sleep) that would favor deep, restorative sleep? How do the authors interpret their results with regard to the current prevailing model that dFB neurons act as a major site of sleep homeostasis? This study could be seen as evidence against it, but the authors do not discuss this in their Discussion.

These are all excellent and thoughtful points, which have made us re-think parts of our discussion. First off, the potential comparison with REM and NREM is entirely speculative, and we have tried to make that more obvious in introduction) and the discussion (e.g, see lines 43, 708, 818). The evidence that the FB neurons (and maybe others) are involved in the homeostatic regulation of sleep is well-supported in the literature, so that part of the discussion holds. However, we concede that the timing of our sleep manipulations could benefit from more explanation. We conducted these during the flies’ subjective day, after the animals had presumably had a good night’s sleep. This means that we induced either kind of sleep for 10 daytime hours, which presumably replaced whatever behavioural states would ‘naturally’ be happening during the day. Female flies sleep less during the day than at night, and we have shown in previous work that daytime sleep quality is different than night-time sleep (van Alphen et al, Journal of Neuroscience, 2013), leading us to suggest that most ‘deep’ or quiet sleep happens at night, for flies. Following this reasoning, daytime optogenetic activation might not be depriving flies of much quiet sleep, or accumulating a deep sleep drive as the Reviewer proposes. Rather, both induced sleep manipulations could be providing 10 hours of either kind of sleep that the flies don’t really ‘need’. Why did we design it this way? Firstly, we were interested in simply asking what these chronic sleep manipulations do to gene expression in rested flies, and how they might be similar or different. We focussed on daytime manipulations to avoid precisely the confound of sleep pressure, and also because we observed red-light artifacts at night for our optogenetic experiments (which we reported). Our sleep deprivation strategy was designed specifically as a control for the THIP (Gaboxadol) experiments, to control for non-sleep related effects of the drug (see below our rationale for why this was less crucial for the optogenetic experiments). In conclusion, we had a logical rationale for how the experiments were done, centred on the straightforward question of whether these two different approaches to sleep induction were having similar effects in well-rested flies. In retrospect, we were not anticipating the Reviewer’s thoughtful logic regarding the dFB’s potential role in also regulating deep sleep homeostasis. We now provide some discussion along these lines to make readers aware of this line of reasoning, as well as our rationale for why prolonged optogenetic sleep induction was not sleep-depriving (lines 768-777).

2- Regarding the physiological effects of Gaboxadol, to what extent is the quieting induced by this drug reminiscent of physiology of the brains of flies spontaneously meeting the behavioral criterion for quiet sleep? Given the relatively high dose of the drug being delivered to the de-sheathed brain in the imaging experiments (at least when compared to the dose used in the fly food), one worries that the authors may be inducing a highly abnormal brain state that might bear very little resemblance to the deeply sleeping brain under normal conditions. As the authors acknowledge, it is difficult to compare these two situations. Comparing the physiological state of brains put to sleep by Gaboxadol and brains that have spontaneously entered a deep sleep state therefore seems critical.

As discussed above, our Gaboxadol (THIP) perfusion concentration (0.2mg/ml) was the minimal dosage that effectively induced sleep within 5 minutes, based upon previously published work (Yap et al, Nature Communications, 2017). Lower concentrations were unreliable, with some never inducing sleep at all. Comparisons with feeding THIP are tenuous, and we make that clear in our discussion (lines 731-735). Nevertheless, the Reviewer makes an excellent point about comparisons with spontaneous ‘quiet’ sleep. Here, we feel well supported (please see Author response image 3 below, comparing THIP-induced sleep (this work, B) and spontaneous sleep (A) from previous study). In our previous study (Tainton-Heap et al, 2021) we showed that neural activity and connectivity decreases during spontaneous quiet sleep. This is what we also see with THIP perfusion. In contrast, in Troup et al, J. of Neuroscience (2023) we confirm that neither neural activity nor connectivity changes during optogenetic R23E10 activation, and general anesthesia – unlike THIP – does NOT produce a quiet brain state. Our finding that THIP effects are nothing like general anesthesia (at the level of brain activity levels) suggests a physiological sleep state closer to spontaneous quiet sleep. We elaborate on this important observation in our results, also pointing to crucial differences with general anesthesia (lines 411-415).

Author response image 3.

THIP-induced sleep resembles quiet spontaneous sleep. A. Calcium imaging data from spontaneously sleeping flies, taken from Tainton-Heap et al, 2021. Left, percent neurons active; right, mean degree, a measure connectivity among active neurons. Both measures decrease during later stages of sleep. B. Calcium imaging data from flies induced to sleep with 5min of 0.2mg/ml THIP perfusion (this study). Left, percent neurons active; right, mean degree. Both measures are significantly decreased, resembling the later stages of spontaneous sleep, which we have termed ‘quiet sleep. Hence THIP-induced sleep resembles quiet sleep. Note that the genetic background is different in A and B, hence the different baseline activity levels.

3- There are some issues with Figure 3, in particular 3C-D. It is not clear whether these panels show representative traces or an average, however both the baseline activity and fluorescence are different between C and D, in particular in their amplitude. Therefore, it is difficult to attribute the differences between C and D to the stimulation itself or to the previously different baseline. In addition, the fact that flies with dFB activation seem to keep a basal level of locomotor activity whereas THIP-treated ones don't is quite striking, however it is not being discussed. Finally, the authors claim that the flies eventually wake up from THIP-induced sleep (L360-361), however there are no data to support this statement.

These are representative traces, which is a way of showing the raw calcium data (Cell ID) so readers can see for themselves that one manipulation silences whereas the other does not – even though flies become inactive for both. The Y-axis scale is standard deviation of the experiment mean. Since THIP decreases neural activity, then the baseline is comparatively higher. Since optogenetic activation does not change average neural activity levels, the baseline is centered on zero. This is an outcome of our analysis method and does not reflect any ‘true’ baseline. We have now clarified this in our figure legend. We now also confess that flies rendered asleep optogenetically can be ‘twitchy’ (line 374). Finally, we show data for 3 flies that were recorded until they woke up. The rest were verified behaviorally, after the experiment. This is now explained in the Methods.

4- In Figure 4C, it is strange that the SEM is always exactly the same across the whole experiment. Readers should be aware that there might have been an issue when plotting the figure.

This is not a mistake, the standard errors are just all quite close (between 0.17 and 0.22). This is because of the way we did the analysis, asking how many flies responded to each stimulus event, with incremental levels of responsiveness. This is explained in the Methods. The figure makes the important point of sleep and recovery.

e- Comments regarding the transcript analyses

1- General comment: the title of this manuscript is inaccurate - the "transcriptome" commonly refers to the entirety of all transcripts in a cell/tissue/organ/animal (including genes that are not differentially expressed following their interventions), and it is therefore impossible to "engage two non-overlapping transcriptomes" in the same tissue. Perhaps the word "transcriptional programs" or transcriptional profiles" would be more accurate here?

We thank the Reviewer for this advice and have changed the title as proposed.

2- Given the sensitivity of transcriptomic methods, there is a significant concern that the optogenetic experiments are not as well controlled as they could be. Given the need for supplemental all-trans retinal (ATR) for functional light gating of channelrhodopsins in the fly, it is convenient to use flies with Gal4-driven opsin that have not been given supplemental ATR as a negative control, particularly as a control for the effects of light. However, there is another critical control to do here. Flies bearing the UAS-opsin responder element but lacking the GAL4 driver and that have been fed ATR are critical for confirming that the observed effects of optogenetic stimulation are indeed caused by the specific excitation of the targeted neurons and not due to leaky opsin expression, or the effect of ATR feeding under light stimulation or some combination of these factors. Given the sensitivity of transcriptomic methods, it would be good to see that the candidate transcripts identified by comparing ATR+ and ATR- R23E10GAL4/UAS-Chrimson flies are also apparent when comparing R23E10GAL4/UAS-Chrimson (ATR+) with UAS-Chrimson (ATR+) alone.

We have not done these experiments on UAS-Chrimson/+ controls. Like many others in our field, we viewed non-ATR flies as the best controls, because this involves identical genotypes. Since we were however aware that ATR feeding itself could be affect gene expression, we specifically checked for this with our early (1hour) collection timepoint. We only found 26 gene expression differences between ATR and -ATR flies at this early timepoint, compared with 277 for the 10-hour timepoint. We detail this rationale in our results, explaining why this is a convincing control for ATR feeding. If there was leaky opsin expression / activity, this would have been evident in our design. Regarding the cumulative effect of light, this would also have been accounted in our design, as only 1 hour would have elapsed in our first timepoint compared to 10 hours in our second. While the Reviewer is correct in saying that parental controls are called for in many Drosophila experiments, this becomes quickly unmanageable in transcriptomic studies, which is exactly why well-designed +ATR vs -ATR comparisons in the exact same strain are most appropriate. We feel that our 1-hr timepoint mostly addresses this concern.

3- Figures about qPCR experiments (5G and 6G) are problematic. First, whereas the authors seem satisfied with the 'good correspondence' between their RNA-seq and qPCR results, this is true for only ~9/19 genes in 5G and 2/6 genes in 6G. Whereas discrepancies are not rare between RNA-seq and qPCR, the text in L460-461 and 540-541 is misleading. In addition, it is unclear whether the n=19 in L458 refers to the number of genes tested or the number of replicates. If the qPCR includes replicates, this should be more clearly mentioned, and error bars should be added to the corresponding figures.

We consider that our qPCR validations were convincing, as they were all mostly changed in the ‘right’ direction. We agree that are some discrepancies, so have modified our language to reflect this. We have also clarified that 19 refers to the number of genes validated by qPCR in that THIP dataset. All qPCRs involved three technical replicates. We prefer to keep these histograms the way they are to convey these simple trends. For complete transparency, we now provide a supplemental Excel worksheet with all of the qPCR data, alongside corresponding RNAseq data and stats for the selected genes (Supplementary Table 9).

4- There is a lack of error bars for all their RNAseq and qPCR comparisons, which is particularly surprising because the authors went to great lengths and analyzed an applaudably large amount of independent biological replicates, yet the variability observed in the corresponding molecular data is not reported.

The genes reported in each of our datasets and associated supplemental figures and tables were all significant, as determined by criteria outlined in the Methods. However, we appreciate that readers might want to get a sense of the values and variances involved, as well as access to the entire gene datasets. We now provide all of these as additional ‘sheets’ in our existing supplemental tables (S2-S7), so this should be very easy to navigate and evaluate. In addition to the previously provided lists for significant genes, in the second Excel sheet (‘All genes’) readers will be able to see the data for all 5 replicates, for the significant genes as well as all other ~15,000 genes (listed in alphabetical order). We feel that this will be a helpful resource, because admittedly significance thresholds can still be a little arbitrary and some readers might want to look up ‘their’ genes of interest.

Comments to authors

Other comments

1- Text in L441 & 606 is misleading. According to ref 52, AkhR is involved specifically in starvation-induced sleep loss, and not in general sleep regulation.

Corrected.

2- The language used in L568-570 and 573-574 is confusing. The authors should specify that the knock down of cholinergic subunits, rather than the subunits themselves is what causes sleep to increase or decrease.

Corrected.

3- The authors' investigation of cholinergic receptor subunits function is very preliminary, and it is difficult to draw any conclusion from what is presented here. In particular, their behavioral data is difficult to reconcile with the RNA-seq data showing overexpression of both short sleep increasing and short sleep decreasing subunits. Without knowing where in the brain these subunits are required for controlling sleep, the data in Figure 7 is difficult to appreciate.

We have now conducted additional experiments where we specifically knocked down these alpha receptor subunits (all 7 of them) in the R23E10 neurons. This seemed an obvious knockdown location, to determine if any of these subunits regulated activity in the same sleep promoting neurons that were the focus of this study. We found that alpha1 knockdown in these neurons had similar sleep phenotypes, which we believe is an important result. Since this functional localisation is a logical ending for the paper, we have now made it the final figure.

Suggestions & comments

1- It would be interesting if the authors could discuss their findings that metabolism genes are downregulated in THIP flies in the context of recent work that showed upregulation of mitochondrial ROS after sleep deprivation (Kempf et al, 2019).

We now add the Kempf 2019 reference and allude to how those findings could be consistent with ours.

2- The fact that THIP-induced sleep persists long after THIP removal (Fig 3D) is very intriguing and interesting. This suggests that the drug might trigger a sleep-inducing pathway that can continue on its own without the drug, once activated.

This is correct, and in stark contrast to the optogenetic manipulation we employ, which does not appear to show such sleep inertia. We have now added a sentence highlighting this interesting difference (lines 394-396).

3- The authors identify many new genes regulated in response to specific methods for sleep induction. These are all potentially interesting candidates for further studies investigating the molecular basis of sleep. It would be interesting to know which of these genes are already known to display circadian expression patterns.

By providing all of the gene lists, these are now available to ask questions such as these. We hesitate however to delve into this domain for this work, as our main goal was to compare these two kinds of sleep in flies.

4- The brain-wide monitoring of neural activity invites a number of very exciting follow-up experiments - most importantly, it would be fascinating to establish, which neurons are active in the different phases the authors describe! Are these neurons that are involved in transmitting external visual stimuli to the central brain? Do they also project into the central complex? They could make use of the large collection of existing driver lines in the fly and they could also exploit the extraordinary knowledge of the connectome and transcriptome of the fly brain.

Thank you for sharing our enthusiasm for these likely future directions.

5- The Dalpha2,3,4,6 and 7 Knock-out strains they generate will be a useful reagent for the Drosophila neuroscience community once the efficiency/success of the knock-out has been confirmed by qPCR.

These knockout strains have all been confirmed by our co-authors Hang Luong, Trent Perry, and Philip Batterham. These knockout confirmations are outlined in publications that we reference (Perry et al, 2021).

Materials and methods:

1- This study has employed custom-built apparatus and custom-written code/scripts, but these do not appear to be available to the reader. For the sake of replicability, the authors should make these available.

The code/scripts are available via the University of Queensland research data management system as described in the Methods, and can be sent by the Lead Contact. The imaging hardware and analysis code are identical to what was described in a previous publication, and available as directed therein (Tainton-Heap et al, 2021).

2- Also, the authors should give details on the food used to rear their flies. Fly media comes in several common forms and sleep is sensitive to diet.

This has now been elaborated in the beginning of the Methods.

3- The light regime used for optogenetic excitation of dFB neurons consists of 12h of uninterrupted bright red LED light. Most optogenetic stimulations consist of pulsed high frequency flashes interlaced with pauses in illumination. Can dFB neurons be driven constitutively with 12 hours of bright light?

We showed in Tainton-Heap (2021) that 7Hz pulsed red light had exactly the same effect on R23E10/Chrimson readouts as continuous red light, which is why we opted here to provide continuous red light. That optogenetic sleep induction can be driven continuously for 12 hours is evident by our 24-hour sleep profiles. However, we agree that one could question whether sleep quality is similar after 12 hours. To address this, we did an additional experiment where we stimulated the flies hourly, to determine if their behavioural responsiveness to mechanical stimuli changed over the course of continued sleep induction, for both optogenetic and THIP-induced sleep. We present the data below in Author response image 4. As can be seen in these new analyses, while optogenetic sleep induction persists across 12 daytime hours (speed is close to zero throughout), flies do indeed become more responsive later in the day. This could have two different interpretations: either some sleep functions are being satisfied over time, or the activation regime is becoming less effective over time. Either way, these data show that at our 10-hour daytime timepoint, unstimulated flies are still largely inactive, even though their arousal thresholds might have gradually changed; so the uninterrupted red-light regime is still effective. The comparison with THIP is interesting: here there does not seem to be a change in responsiveness over time; the drug just decreases behavioral responsiveness throughout. Together, these experiments support our view that both approaches are sleep-promoting throughout the 12-hour day, although we appreciate that sleep quality is not identical.

Author response image 4.

A) The average speed of baseline (grey) and optogenetically-activated flies (green) across 24 hours. Red dots indicate vibration stimulus times. B) The average speed of control (grey) and THIP-fed flies (blue) across 24 hours. Flies are all R23E10/Chrimson. N= 87 for optogenetic, n=88 for -THIP, n=85 for +THIP.

4- The authors use the SNAP apparatus to prevent THIP-treated flies from sleeping to tease out possible sleep-independent effects. This is an excellent control. Why have the authors not done the same with the optogenetic treatment? It's surprising not to see this control given the concern the authors express (lines 501 - 502) that the dFB manipulation might be paralyzing awake flies, which certainly seems possible given the light regimes used. Why not test this directly with SNAP?

We appreciate that this may have been a valuable additional control. However, we designed this control for the THIP experiments specifically because of concerns about THIP’s (yet unknown) mechanism of action in flies. THIP is a gabaergic drug with most likely many off-target effects that have little to do with sleep, hence the need for a control where we compare to flies that ingested THIP but have been prevented from sleeping. In contrast, R23E10-driven sleep induction is exactly that, a circuit when activated that induces sleep. Whatever specific neurons might really be involved, the Gal4 circuit is sleep-inducing. This is well supported by multiple publications. The most appropriate control for assessing transcriptomic effects during optogenetic sleep here is not preventing sleep, but rather no increased sleep in flies that have not ingested ATR, and comparing that to effects of ATR alone, which is what we have done. Adding a sleep-deprivation layer onto both of these analyses may have been interesting, but a lot more analyses and not strictly required to identify relevant sleep-related genes. We have rephrased the misleading sentence about paralyzing flies, to instead clarify that lack of overlap with the SD dataset suggests that optogenetic activation is not preventing sleep functions from being engaged.

5- A pairwise comparison of ZT01 and ZT10 does not address circadian expression cycles in a meaningful way. There will be strong effects of the LD cycle here. I suggest toning this down. (Though it is gratifying to see the expected changes in the core clock genes.)

We have changed the language from ‘circadian’ to ‘light-dark’ to address this, although have kept the word ‘circadian’ when referring specifically to genes such as per, clock, timeless, etc.

6- Line 109: There is a reference missing.

We now provide the relevant reference.

Results

1- General comment regarding the figures: a general effort could be made to improve the design and quality of the figures and make them more readable. There are a lot of issues such as stretched or misaligned text, badly drawn frames, etc.

We think we know which figures this might relate to (e.g., Figures 3,4B), so we have adjusted where appropriate.

2- Instead of 'dFB-induced' (e.g., L77) it would be more accurate to use 'optogenetically-induced'

Thank you for this helpful advice. We have changed our language throughout to say ‘optognetically-induced’

3- Figure S1 should be integrated in the main figure to make the quantification more easily 4accessible.

We have integrated Figure S1 into the main figures. It is now Figure 3.

5- It would be good to include red light controls in Figure 2C, E, G.

Making Figure S1 a main figure has better highlighted the fact that we have done red light controls (‘baseline’).

6- line 313: Fig2E-H - these graphs would benefit if the authors made it more obvious where the maximum sleep amount would fall - i.e. the combination of bouts and minutes that add up to 12 hours (and therefore the entire day/night)

If a fly were to sleep uninterrupted for all 12 hours of a day or night, that would amount to a sleep bout 720 minutes long. We do not feel that identifying this maximum on these graphs would be helpful. It should be clear from the data that a floor is reached with very few sleep bouts exceeding 60 minutes in our paradigm. To help orient the reader though, we now clarify in the figure legend that the maximum is 720 minutes or 12 hours.

7- Fig. 2B, D: It was not clear why the authors took the 3-day average here. Doesn't that lead to a whole range of very different behaviors? I could, perhaps naively, imagine that a fly's behavior changes after 2 days of almost-permanent sleep?

We took the 3-day average because the effect of THIP on each successive day was not significantly different (see Author response image 5, below). Flies wake up enough to have a good feed (see Author response image 2) and then go back to sleep. Since this is however an important point raised by the reviewer, we now mention in the Methods that sleep duration was not different among the 3 averaged days and nights (lines 193-195).

Author response image 5.

Data from THIP feeding experiment (Figure 2B) in manuscript, separated into 3 successive days and nights, with THIP-fed flies (blue) compared to controls (white). Averages  SD are shown, samples sizes are the same as in Figure 2D. No THIP data was significantly different across days and nights (ANOVA of means).

8- In Figure 2C the authors compare optogenetically induced to "spontaneous sleep," which I think refers to baseline sleep before stimulation, according to the figure. I think the proper comparison would be to the red light control (ATR-); though see the comment above regarding optogenetic controls).

This information was provided in Figure S1. We now provide it as a main Figure 3, as requested above.

We also made a point about red light having an effect at night, which is why we focussed on daytime effects for our transcriptomic comparisons. We feel that the ATR-fed flies (minus red light) are an appropriate control here for optogenetically-induced sleep: same exact genotype and ATR feeding, just no optogenetic activation. We therefor would prefer to keep these graphs as they are, especially since we show -ATR data subsequently.

9- Figures 3A and 4A are redundant; Figure 3B has some active ROIs that are outside of the brain. I am not sure how this is possible?

We have removed the redundant 4A and replaced it with the THIP molecule to clearly signal what this figure is focussed on. In Figure 3B (now 4B), the brain mask is a visual estimate made from the middle of the image stack. Some neurons in other layers are outside this single-layer estimate. All neurons were all accounted for.

10- Figure 4B is confusing. It took me a while to understand and so it can do with re-drawing in a more accessible way.

We agree that this was confusing, e.g. there were too many arrows. We have redrawn and simplified (Now 5A).

11- The authors state that flies wake up from THIP-induced sleep on the ball, but in Figure 4D there appears to be fewer samples for flies who have woken up from THIP (3) compared to those observed before THIP administration. Are flies dying?

None of the flies died. Most flies were removed from imaging to confirm recovery, while 3 were left in our imaging setup to measure brain activity upon recovery. These results are in Figure 5C and now clarified in the Methods.

12- Fig5C,D: I'm surprised that by far the most significant changes (in terms of log2-FC and p-val) occur in the sleep-deprived flies? It is not clear to me what the authors mean by effects that "relate waking process"? Perhaps they could elaborate on this?

We have removed the phrase ‘relates to waking processes’. We now also remark on the high level of fold-change in many of these genes but refrain from discussing this further in the results. It is interesting though.

13- The sentence in L425-428 is unclear - it would be good to rephrase this.

We have rephrased this sentence, hopefully it’s clearer now.

14- Text in L544-545 is confusing. What do you mean by 'less clear'?

We have replaced ‘less clear’ with ‘not dominated by a single category’.

15- It is unclear what is the control in Fig 7A. It would be good to mention what strain was used.

Different knockout strains had different controls. These are identified in the figure legend and Methods.

16- L579-581: it would be helpful to include this data in a supplementary figure.

We now provide this as a supplementary figure as requested (Supplementary Figure 6).

17- There is no information about R57C10 in the methods - it would be good to explain which neurons this line labels, and why you chose it.

We now clarify in the methods that R57C10-Gal4 is a pan-neural driver, and provide a reference.

18- Table S5 - If I'm not mistaken then the first line should say 1h, not 10h.

Corrected

Author Response

We are grateful for the constructive comments of the reviewers and for the succinct assessment of our work by the editors. Here we provide a brief summary of our response to answer the major criticism of our reviewers. We will give a detailed point-to-point response soon when we upload a revision of our paper.

1) The MATLAB code for the spatial autocorrelation analysis is now freely available at the following site: : https://github.com/dcsabaCD225/Moran_Matlab/blob/main/moran_local.m If any question arises during its implementation, please contact Csaba Dávid (david.csaba@koki.hu)

2) Concerning the computer resources and times required to perform Moran’s I image analysis, here we provide a brief description of the hardware and the calculations for images with different sizes.

Hardware used for performing the analysis:

Intel(R) Xeon(R) Silver 4112 CPU @ 2.60GHz, 2594 Mhz, 4 kernel CPU, 64GB RAM, NVIDIA GeForce GTX 1080 graphic card.

MATLAB R2021b software was used for implementation.

Computation times are shown in Author response table 1.

Author response table 1.

3) In response to the comment:

“While the method's avoidance of AI training appeals to those lacking computational know-how and shows improved accuracy over basic threshold-based techniques, there are valid concerns regarding its performance in comparison to advanced methodologies”.

Comparison of Moran’s I image analysis with AI based segmentations raises conceptual problems which will be addressed in detail in the revised version. Briefly, the basis of AI based analyses is that the ground truth is known and using a large teaching set AI learns to extract the relevant information for image segmentation. In several cases, however (like protein distribution in the membrane) the ground truth is not known and cannot be easily determined by any single observer. Defining spatial inhomogeneities in protein distribution, differentiating proteins involved vs not involved in clusters is highly subjective. Indeed, our analysis showed the 23 expert human observers varied hugely in establishing the boundaries of a protein cluster. As a consequence, establishing and using a teaching set would be highly contentious in these cases. In an average laboratory setting generating a teaching set using hundreds of images examined by two dozen people would not be impossible but not really plausible. The beauty of Moran’n I analysis is that it is able to extract the relevant signals from an image generated in different, often noisy condition using a simple algorithm that allows quantitative characterization and identification of changes in many biological and non-biological samples.

Author Response

Reviewer #2 (Public Review):

The authors describe the synthesis and testing of the anti-cancer activity of a new molecule CK21 against pancreatic cancer mouse models. This part of the study is very strong showing regression of pancreatic tumors at non-toxic concentrations, which is very hard to achieve for practically uncurable pancreatic cancer. Authors synthesized CK21 as an analog of a known inhibitor of RNA synthesis which is very toxic. The authors did very little attempt to understand whether the mechanism of anti-cancer efficacy of CK2 is similar to this known inhibitor of transcription or not. One cannot compare gene expression profiles between untreated and CK21-treated cells, taking into account that CK2 may inhibit the expression of all genes. The effect of CK2 on general transcription needs to be tested first, and then based on this data absolute changes in the expression of genes may be considered for the revealing of the mechanism of activity of CK21.

We also appreciated the toxicity concerns; thus, we designed the transcriptomic analysis on the human organoid cultured cells for early time points of 3, 6, 9 and 12 h, and with a CK21 concentration of 50nM, to ensure that at the time of harvest, the cells were ~100% viable. At these time points, many genes were upregulated but defined by IPA as enriched for cell death (apoptosis and necrosis), senescence and cell cycle arrest (Fig 5). This led us to hypothesize that the direct effect of CK21 on the tumor cells is the induction of apoptosis, but via multiple pathways.

Reviewer #3 (Public Review):

This manuscript describes CK21, a modified version of Triptolide, a natural compound with antcancer activities, to improve its bioavailability. The authors tested the compound in two human pancreatic cancer cell lines, in vitro and in vivo. The authors also use two human organoid lines derived from pancreatic cancer, and mouse KC and KPC cell lines. In all models, CK21 treatment induces dose-dependent cytotoxicity. In vivo, CK21 causes tumor regression. The authors perform gene expression analysis and show that treated organoids have generally lower transcription, consistent with cytotoxicity, and a reduction in the KFkB pathway activation.

Key experiments that would strengthen the current manuscript are: the inclusion of normal cell lines and organoids, too, presumably, show no cytotoxic effect. If that is the case, the authors would have the opportunity to compare responses and determine whether a tumor-specific mechanism can be defined.

Our in vivo studies suggest that CK21 is more specific to tumors, as CK21 ≤3 mg/kg treated mice were 100% viable and gained weight comparably to no treatment group (Fig.2d). Furthermore, in vitro studies with primary fibroblast cells indicate that comparable significant toxicity to CK21 after 72h culture was observed at 500 nM (Fig.s2). In contrast, CK21 induced significant toxicity in AsPC1 and Panc-1 cells at 50 nM (Fig. 1f.)

The authors observe that few gene changes - besides from overall lowering in transcription, occur upon treatment with CK21. They suggest that the drug acts through inhibition of the NFkB pathway and an increase in reactive oxygen species (ROS). However, no experiments to test whether either/both of these findings explain the cytotoxic effect (rescue experiments would be particularly valuable).

We performed a rescue study using an ROS inhibitor (acetylcysteine) but observed no significant effect (data not shown). We speculate that ROS and/or NF-B might function synergistically; additionally, it is possible that other mechanisms might be involved in the anti-tumor effects of CK21.

In the last figure, the authors text whether CK21 is immunosuppressive by testing immunity against a mis-matched tumor cell line (using KPC tumors, mixed strain, in mixed strain mice). The immunity against HLA mis-matched cells is a very strong immune reaction, and mild immune suppression might be missed, which diminishes the value of these findings.

KPC-960 tumor cells were derived from KPC (C57BL/6 background); therefore, KPC-960 tumors were HLA matched with host C57BL/6 mice. We were surprised to observe spontaneous rejection of the KPC-960 tumor line, since this contrasts with Torres et al. 2013. We speculate that this could be due to the increased number of passages resulting in antigenic drift, which may result in the accumulation of mutations that induce spontaneous rejection.

We agree that there might be mild immunosuppression that we did not detect; we have included this caveat in the discussion. KC-6141 tumor cells used as CTL targets were from KC mice (mixed background – B6.129).

Author Response

Reviewer #1:

This is a very timely paper that addresses an important and difficult-to-address question in the decision-making field - the degree to which information leakage can be strategically adapted to optimise decisions in a task-dependent fashion. The authors apply a sophisticated suite of analyses that are appropriate and yield a range of very interesting observations. The paper centres on analyses of one possible model that hinges on certain assumptions about the nature of the decision process for this task which raises questions about whether leak adjustments are the only possible explanation for the current data. I think the conclusions would be greatly strengthened if they were supported by the application and/or simulation of alternative model structures.

We thank the reviewer for this positive appraisal of our study. We now entirely agree with their central comment about whether leak adjustments are the only (or even the best) explanation for the current data. We hope that the additional modelling sections that we have discussed in response to main comment 1 above have strengthened the paper. We have responded point-by-point to their public review, as this contained their main recommendations for revision.

The behavioural trends when comparing blocks with frequent versus rare response periods seem difficult to tally with a change in the leak. […] Are there other models that could reproduce such effects? For example, could a model in which the drift rate varies between Rare and Frequent trials do a similar or better job of explaining the data?

We can see why the reviewer has advocated for a possible change of drift rate (or ‘gain’ applied to sensory evidence) between conditions to explain our behavioural findings. We found, however, that changes in drift rate could elicit qualitatively similar changes in integration kernels to changes in decision threshold:

Author response image 1.

Changes in gain applied to incoming sensory evidence (A parameter in model) have similar effects on recovered integration kernels from Ornstein-Uhlenbeck simulation as changes in decision threshold.

The likely reason for this is that the overall probability of emitting a response at any point in the continuous decision process is determined by the ratio of accumulated evidence to decision threshold. A similar logic applies to effects on reactions times and detection probability (main figure 2): increasing sensory gain/decreasing decision threshold will lead to faster reaction times and increased detection probability during response periods.

Both parameters may even have a similar effect on ‘false alarms’, because (as the reviewer notes below) false alarms in our paradigm are primarily being driven by the occurrence of stimulus changes as well as internal noise. In fact, the false alarm findings mean it is difficult to fully reconcile all of our behavioural findings in terms of changes in a single set of model parameters in the O-U process. It is possible that other changes not considered within our model (such as expectations of hazard rates of inter-response intervals leading to dynamic thresholds etc.) may have had a strong impact upon the resulting false alarm rates. A full exploration of different variations in O-U model (with varying urgency signals, hazard rates, etc.) is beyond the scope of this paper.

For this reason, we have decided in our new modelling section to focus primarily on a single, well-established model (the O-U process) and explore how changes in leak and threshold affect task performance and the resulting integration kernels. We note that this is in line with the suggestion of reviewer #2, who focussed on similar behavioural findings to reviewer #1 but suggested that we look at decision threshold rather than drift rate as our primary focus.

This ties in to a related query about the nature of the task employed by the authors. Due to the very significant volatility of the stimulus, it seems likely that the participants are not solely making judgments about the presence/absence of coherent motion but also making judgments about its duration (because strong coherent motion frequently occurs in the inter-target intervals). If that is so, then could the Rare condition equate to less evidence because there is an increased probability that an extended period of coherent motion could be an outlier generated from the noise distribution? Note that a drift rate reduction would also be expected to result in fewer hits and slower reaction times, as observed.

As mentioned above, the rare and frequent targets are indeed matched in terms of the ease with which they can be distinguished from the intervening noise intervals. To confirm this, we directly calculated the variance (across frames) of the motion coherence presented during baseline periods and response periods (until response) in all four conditions:

Author response image 2.

The average empirical standard deviation of the stimulus stream presented during each baseline period (‘baseline’) and response period (‘trial’), separated by each of the four conditions (F = frequent response periods, R = rare, L = long response periods, S = short). Data were averaged across all response/baseline periods within the stimuli presented to each participant (each dot = 1 participant). Note that the standard deviation shown here is the standard deviation of motion coherence across frames of sensory evidence. This is smaller than the standard deviation of the generative distribution of ‘step’-changes in the motion coherence (std = 0.5 for baseline and 0.3 for response periods), because motion coherence remains constant for a period after each ‘step’ occurs.

Some adjustment of the language used when discussing FAs seems merited. If I have understood correctly, the sensory samples encountered by the participants during the inter-response intervals can at times favour a particular alternative just as strongly (or more strongly) than that encountered during the response interval itself. In that sense, the responses are not necessarily real false alarms because the physical evidence itself does not distinguish the target from the non-target. I don't think this invalidates the authors' approach but I think it should be acknowledged and considered in light of the comment above regarding the nature of the decision process employed on this task.

This is a good point. We hope that the reviewer will allow us to keep the term ‘false alarms’ in the paper, as it does conveniently distinguish responses during baseline periods from those during response periods, but we have sought to clarify the point that the reviewer makes when we first introduce the term.

“Indeed, participants would occasionally make ‘false alarms’ during baseline periods in which the structure of the preceding noise stream mistakenly convinced them they were in a response period (see Figure 4, below). Indeed, this means that a ‘false alarm’ in our paradigm has a slightly different meaning than in most psychophysics experiments; rather than it referring to participants responding when a stimulus was not present, we use the term to refer to participants responding when there was no shift in the mean signal from baseline.”

And:

“The fact that evidence integration kernels naturally arise from false alarms, in the same manner as from correct responses, demonstrates that false alarms were not due to motor noise or other spurious causes. Instead, false alarms were driven by participants treating noise fluctuations during baseline periods as sensory evidence to be integrated across time, and the physical evidence preceding ‘false alarms’ need not even distinguish targets from non-targets.”

The authors report that preparatory motor activity over central electrodes reached a larger decision threshold for RARE vs. FREQUENT response periods. It is not clear what identifies this signal as reflecting motor preparation. Did the authors consider using other effectorselective EEG signatures of motor preparation such as beta-band activity which has been used elsewhere to make inferences about decision bounds? Assuming that this central ERP signal does reflect the decision bounds, the observation that it has a larger amplitude at the response on Rare trials appears to directly contradict the kernel analyses which suggest no difference in the cumulative evidence required to trigger commitment.

Thanks for this comment. First, we should simply comment that this finding emerged from an agnostic time-domain analysis of the data time-locked to button presses, in which we simply observed that the negative-going potential was greater (more negative) in RARE vs. FREQUENT trials. So it is simply the fact that it precedes each button press that we relate it to motor preparation; nonetheless, we note that (Kelly and O’Connell, 2013) found similar negative-going potentials at central sensors without applying CSD transform (as in this study). Like them, we would relate this potential to either the well-established Bereitschaftpotential or the contingent negative potential (CNV).

We agree that many other studies have focussed on beta-band activity as another measure of motor preparation, and to make inferences about decision bounds. To investigate this, we used a Morlet wavelet transform to examine the time-varying power estimate at a central frequency of 20Hz (wavelet factor 7). We repeated the convolutional GLM analysis on this time-varying power estimate.

We first examined average beta desynchonisation at a central cluster of electrodes (CPz, CP1, CP2, C1, Cz, C2) in the run-up to correct button presses during response periods. We found a reliable beta desynchonisation occurred, and, just as in the time-domain signal, this reached a greater threshold in the RARE trials than in the FREQUENT trials:

Author response image 3.

Beta desynchronisation prior to a correct response is greater over central electrodes in the RARE condition than in the FREQUENT condition.

We agree with the reviewer that this is likely indicative of a change in decision threshold between rare and frequent trials. We also note that our new computational modelling of the O-U process suggests that this in fact reconciles well with the behavioural findings (changes in integration kernels). We now mention this at the relevant point in the results section:

“As large changes in mean evidence are less frequent in the RARE condition, the increased neural response to |Devidence| may reflect the increased statistical surprise associated with the same magnitude of change in evidence in this condition. In addition, when making a correct response, preparatory motor activity over central electrodes reached a larger decision threshold for RARE vs. FREQUENT response periods (Figure 7b; p=0.041, cluster-based permutation test). We found similar effects in beta-band desynchronisation prior, averaged over the same electrodes; beta desynchronisation was greater in RARE than FREQUENT response periods. As discussed in the computational modelling section above, this is consistent with the changes in integration kernels between these conditions as it may reflect a change in decision threshold (figure 2d, 3c/d). It is also consistent with the lower detection rates and slower reaction times when response periods are RARE (figure 2 b/c).”

We did also investigate the lateralised response (left minus right beta-desynchronisation, contrasted on left minus right responses). We found, however, that we were simply unable to detect a reliable lateralised signal in either condition using these lateralised responses. We suspect that this is because we have far fewer response periods than conventional trialbased EEG experiments of decision making, and so we did not have sufficient SNR to reliably detect this signal. This is consistent with standard findings in the literature, which report that the magnitude of the lateralised signal is far smaller than the magnitude of the overall beta desynchronisation (e.g. (Doyle et al., 2005))

P11, the "absolute sensory evidence" regressor elicited a triphasic potential over centroparietal electrodes. The first two phases of this component look to have an occipital focus. The third phase has a more centroparietal focus but appears markedly more posterior than the change in evidence component. This raises the question of whether it is safe to assume that they reflect the same process.

We agree. We have now referred to this as a ‘triphasic component over occipito-parietal cortex’ rather than centroparietal electrodes.

Reviewer #2:

Overall, the authors use a clever experimental design and approach to tackle an important set of questions in the field of decision-making. The manuscript is easy to follow with clear writing. The analyses are well thought-out and generally appropriate for the questions at hand. From these analyses, the authors have a number of intriguing results. So, there is considerable potential and merit in this work. That said, I have a number of important questions and concerns that largely revolve around putting all the pieces together. I describe these below.

Thanks to the reviewer for their positive appraisal of the manuscript; we are obviously pleased that they found our work to have considerable potential and merit. We seek to address the main comments from their public review and recommendations below.

1) It is unclear to what extent the decision threshold is changing between subjects and conditions, how that might affect the empirical integration kernel, and how well these two factors can together explain the overall changes in behavior.

I would expect that less decay in RARE would have led to more false alarms, higher detection rates, and faster RTs unless the decision threshold also increased (or there was some other additional change to the decision process). The CPP for motor preparatory activity reported in Fig. 5 is also potentially consistent with a change in the decision threshold between RARE and FREQUENT. If the decision threshold is changing, how would that affect the empirical integration kernel? These are important questions on their own and also for interpreting the EEG changes.

This important comment, alongside the comments of reviewer 1 above, made us carefully consider the effects of changes in decision threshold on the evidence integration kernel via simulation. As discussed above (in response to ‘essential revisions for the authors’), we now include an entirely new section on how changes in decision threshold and leak may affect the evidence integration kernel, and be used to optimise performance across the different sensory environments. In particular, we agree with the reviewer that the motor preparatory activity that differs between RARE and FREQUENT is consistent with a change in decision threshold, and our simulations have suggested that our behavioural findings on evidence integration are also consistent with this change as well. These are detailed on pp.1-4 of the rebuttal, above.

2) The authors find an interesting difference in the CPP for the FREQUENT vs RARE conditions where they also show differences in the decay time constant from the empirical integration kernel. As mentioned above, I'm wondering what else may be different between these conditions. Do the authors have any leverage in addressing whether the decision threshold differs? What about other factors that could be important for explaining the CPP difference between conditions? Big picture, the change in CPP becomes increasingly interesting the more tightly it can be tied to a particular change in the decision process.

We fully agree with the spirit of this comment, and we’ve tried much more carefully to consider what the influences of decision threshold and leak would be on our behavioural analyses. As discussed in the response to reviewer 1, we think that the negative-going potential at the time of responses (which is greater in RARE vs. FREQUENT, main figure 7b, and mirrored by equivalent changes in beta desynchronisation, see Reviewer Response Figure 5 above) are both reflective of a change in decision threshold between RARE and FREQUENT conditions. We have tried to make this link explicit in the revised results section:

“As large changes in mean evidence are less frequent in the RARE condition, the increased neural response to |Devidence| may reflect the increased statistical surprise associated with the same magnitude of change in evidence in this condition. In addition, when making a correct response, preparatory motor activity over central electrodes reached a larger decision threshold for RARE vs. FREQUENT response periods (Figure 7b; p=0.041, cluster-based permutation test). We found similar effects in beta-band desynchronisation prior, averaged over the same electrodes; beta desynchronisation was greater in RARE than FREQUENT response periods. As discussed in the computational modelling section above, this is consistent with the changes in integration kernels between these conditions as it may reflect a change in decision threshold (figure 2d, 3c/d). It is also consistent with the lower detection rates and slower reaction times when response periods are RARE (figure 2 b/c).”

I'll note that I'm also somewhat skeptical of the statements by the authors that large shifts in evidence are less frequent in the RARE compared to FREQUENT conditions (despite the names) - a central part of their interpretation of the associated CPP change. The FREQUENT condition obviously has more frequent deviations from the baseline, but this is countered to some extent by the experimental design that has reduced the standard deviation of the coherence for these response periods. I think a calculation of overall across-time standard deviation of motion coherence between the RARE and FREQUENT conditions is needed to support these statements, and I couldn't find that calculation reported. The authors could easily do this, so I encourage them to check and report it.

See Author response image 2.

3) The wide range of decay time constants between subjects and the correlation of this with another component of the CPP is also interesting. However, in trying to interpret this change in CPP, I'm wondering what else might be changing in the inter-subject behavior. For instance, it looks like there could be up to 4 fold changes in false alarm rates. Are there other changes as well? Do these correlate with the CPP? Similar to my point above, the changes in CPP across subjects become increasingly interesting the more tightly it can be tied to a particular difference in subject behavior. So, I would encourage the authors to examine this in more depth.

Thanks for the interesting suggestion. We explored whether there might be any interindividual correlation in this measure with the false alarm rate across participants, but found that there was no such correlation. (See Author response image 4; plotting conventions are as in main figure 9).

Author response image 4.

No evidence of between-subject correlations in CPP responses and false alarm rates, in any of the four conditions.

We hope instead that the extended discussion of how the integration kernel should be interpreted (in light of computational modelling) provides at least some increased interpretability of the between-subject effects that we report in figure 9.

Reviewer #3 (Public Review):

The main strength is in the task design which is novel and provides an interesting approach to studying continuous evidence accumulation. Because of the continuous nature of the task, the authors design new ways to look at behavioral and neural traces of evidence. The reverse-correlation method looking at the average of past coherence signals enables us to characterize the changes in signal leading to a decision bound and its neural correlate. By varying the frequency and length of the so-called response period, that the participants have to identify, the method potentially offers rich opportunities to the wider community to look at various aspects of decision-making under sensory uncertainty.

We are pleased that the reviewer agrees with our general approach as a novel way of characterising various aspects of decision-making under uncertainty.

The main weaknesses that I see lie within the description and rigor of the method. The authors refer multiple times to the time constant of the exponential fit to the signal before the decision but do not provide a rigorous method for its calculation and neither a description of the goodness of the fit. The variable names seem to change throughout the text which makes the argumentation confusing to the reader. The figure captions are incomplete and lack clarity.

We apologise that some of our original submission was difficult to follow in places, and we are very grateful to the reviewer for their thorough suggestions for how this could be improved. We address these in turn below, and we hope that this answers their questions, and has also led to a significant improvement in the description and rigour of the methodology.

Author Response

Reviewer #3 (Public Review):

Dysbiosis has a substantial impact on host physiology. Using the nematode C. elegans and E.coli as a model of host-microbe interactions, Yang et al. defined a mechanism by which the host deals with gut dysbiosis to maintain fitness. They found that accumulation of E. coli in the intestine secreted indole, a tryptophan metabolite, and activated the transcription factor DAF-16. DAF-16 induced the expression of lys-7 and lys-8, which in turn limited E. coli proliferation in the gut of worms and maintained the longevity of worms. Finally, these authors demonstrated that indole-activated DAF-16 via TRPA-1 in neurons of worms.

This study revealed a new mechanism of host-microbe interaction. The concept of their work is of broad interest and the results they present are convincing. However, there are some issues that need to be addressed to support the conclusions.

Major issues

1) The authors isolated the crude extract from a high-performance liquid chromatograph (HPLC). A candidate compound was detected by activity-guided isolation and further identified as indole with mass spectrometry and NMR data. The HPLC fractionations and activity-guided isolation experiments should be described in more detail with a schematic figure to reveal how these experiments were performed and how indole was identified. Showing a chemical characterization of indole in Figure 2A is not sufficient for the evaluation of the results. Rather, a figure comparing the fraction 26th with standard indole by MS and NMR is more appealing.

We appreciate the concerns of the reviewer. Activity-guided isolation was performed as follows: The crude extract of E. coli supernatant metabolites was divided into 45 fractions according to polarity using Ultimate 3000 HPLC (Thermofisher, Waltham, MA) coupled with automated fraction collector. After freeze-drying each fraction, 1 mg of metabolites were dissolved in DMSO for DAF-16 nuclear localization assay in worms (Please see new Supplementary Table S2). The 26th fraction with DAF-16 nuclear translocation-inducing activity was then separated on silica gel column (200-300 mesh) with a continuous gradient of decreasing polarity (100%, 70%, 50%, 30%, petroleum ether/acetone) to yield four fractions (26a-d). Only the fraction of 26b could induce DAF-16 nuclear translocation. Then the fraction was further separated using a Sephadex LH-20 column to yield 32 fractions. The 26b-11th fraction with DAF-16 nuclear translocation-inducing activity contained a single compound identified by thin layer chromatography, mass spectrometry and nuclear magnetic resonance (NMR). The compound exhibited a quasimolecular ion peak at m/z 181.0782 [M+H]+ in the positive APCI-MS, and was assigned to a molecular formula of C8H7N. A comparison of these 1H NMR and 13C NMR spectra with the data reported in the literature revealed that the compound was indole (Yagudaev, 1986). The figure shows the comparison of the 26b-11 fraction with the standard indole by MS (Author response image 1).

Author response image 1.

High resolution mass spectrum of the candidate compound and indole.

2) DAF-16::GFP was mainly located in the cytoplasm of the intestine in worms expressing daf-16p::daf-16::gfp fed live E. coli OP50 on Day 1 (Figure 1A and 1B). The nuclear translocation of DAF-16 in the intestine was increased in worms fed live E. coli OP50 on Days 4 and 7, but not in age-matched WT worms fed heat-killed (HK) E. coli OP50 (Figure 1A and 1B). Since DAF-16 functions downstream of DAF-2, have the levels of DAF-2 been tested during aging on OP50 and (HK) OP50, or with and without indole supplementation?

In response to the reviewer’s suggestion, we carried out the RT-PCR experiment in 4-day-old and 7-day-old worms. It has been shown that DAF-2 initiates a kinase cascade that leads to the phosphorylation and cytoplasmic retention of DAF-16. By contrast, a reduction in the DAF-2 signaling leads to the dephosphorylation of DAF-16, allowing its nuclear translocation. In response to the reviewer’s suggestion, we tested the expression of daf-2 in 4-day-old and 7-day-old worms fed with OP50 and (HK) OP50. We found that the mRNA levels of daf-2 were significantly increased in worms on days 4 and 7 in the presence of either live or dead E. coli OP50, compared with those in worms on day 1 (Author response image 2A). In addition, supplementation with indole did not alter the mRNA levels of daf-2 in young adult worms (Author response image 2B). To conclude, the activation of DAF-16 is independent of DAF-2.

Author response image 2.

DAF-16 nuclear translocationisindependent of DAF-2.(A) The mRNA levelsof daf-2weregradually increasedin worms with age.P< 0.01;*P< 0.001; ns, not significant. (B)The mRNA levelsof daf-2were not alteredaftertreatment withindole for 24 hours.ns, not significant.

3) In lines 155-157, the author argued that the increase in the levels of indole in worms results from the intestinal accumulation of live E. coli OP50, rather than exogenous indole produced by E. coli OP50 on the NGM plates. However, the work also showed that supplementation with indole (50-200 μM) could significantly increase the indole levels in young adult worms on Day 1 (Figure 2-figure supplement 3B), which could induce nuclear translocation of DAF-16 in worms (Figure 2B). This result suggested that worms could take in indole from outside culturing environment. The concentration of indole in OP50 and (HK) OP50 could be measured.

We appreciate the concerns of the reviewer. Reviewer #2 also pointed out this problem. In this study, our data showed that the levels of indole were 30.9, 71.9, and 105.9 nmol/g dry weight in worms fed live E. coli OP50 on days 1, 4, and 7, respectively (Figure 2C). This increase in the levels of indole in worms was accompanied by an increase in CFU of live E. coli OP50 in the intestine of worms with age (Figure 2C). In addition, we determined the levels of indole in worms fed HK E. coli OP50, and found that the levels of indole were 28.2, 31.6, and 36.1 nmol/g dry weight in worms fed HK E. coli OP50 on days 1, 4, and 7, respectively (Figure 2-figure supplement 3A). It should be noted that the levels of indole in worms fed dead E. coli OP50 on day 1 were comparable of those in worms fed live E. coli OP50 on day 1 (30.9 vs 28.2 nmol/g dry weight). However, the levels of indole were not increased in worms fed HK E. coli OP50 on days 4 and 7. Furthermore, the observation that DAF-16 was retained in the cytoplasm of the intestine in worms fed live E. coli OP50 on day 1 (Figure 1A and 1B) also indicated that indole produced by E. coli OP50 on the NGM plates is not enough to induce DAF-16 nuclear translocation. By contrast, supplementation with indole (50-200 μM) significantly increased the indole levels in worms on day 1 (Figure 2-figure supplement 3B), which could induce nuclear translocation of DAF-16 in worms (Figure 2B). Thus, the increase in the levels of indole in worms with age results from intestinal accumulation of live E. coli OP50, rather than indole produced by E. coli OP50 on the NGM plates.

4) Recent work showed that the multicopy DAF-16 transgene acts differently from the single copy GFP knock in DAF-16 transgene. Which DAF-16 transgene was used in this work?

The strain we used is TJ356. Its genotype has been described as zIs356 [daf-16p::daf-16a/b::GFP+rol-6(su1006)] (Lee, Hench, & Ruvkun, 2001; Lin, Hsin, Libina, & Kenyon, 2001), from the Caenorhabditis Genetics Center (CGC).

5) In lines 190-193, the author argued that the supplementation with indole (100 M) inhibited the CFU of E. coli K-12 in WT worms, but not daf-16(mu86) mutants, on Days 4 and 7 (Figure 3H and 3I). These results suggest that endogenous indole is involved in maintaining a normal lifespan in worms. This is overstating. The data here more likely suggest that indole could inhibit the proliferation of E. coli through DAF-16.

We really appreciate this reviewer’s preciseness. In response to the reviewer’s suggestion, we had changed "...indole is involved in maintaining a normal lifespan in worms" to "...indole produced by bacteria in the gut could inhibit the proliferation of E. coli via DAF-16 in worms".

6) Sonowal (2017) reported that AHR mediates indole-promoted lifespan extension at 16 C. Yet this work argued that RNAi knockdown of ahr-1 did not affect the nuclear translocation of DAF-16 in worms fed E. coli K12 strain on Day 7 (Figure 4-figure supplement 1A) or young adult worms treated with indole (100 M) for 24 h. The difference between these two works should be discussed.

We really appreciate this reviewer’s preciseness. It has been shown that AHR-1 mediates indole-promoted lifespan extension in worms at 16 C (Sonowal et al., 2017). However, our data show that AHR-1 is not involved in activation of DAF-16 by indole-induced nuclear translocation of DAF-16 at 20 C. This means that AHR-1 and TRPA-1-lifespan extension by indole are essentially different. In our study, indole is added to NGM plates when worms reached the young adult stage. In the study by Sonowal et al., indole is supplemented at the stage of L1 larva. In addition, lifespan of C. elegans varies at different temperatures (Xiao et al., 2013). Thus, indole may promote lifespan extension via different mechanisms, which is dependent on exposure time and temperature.

7) Sonowal (2017) conducted mRNA profiling for worms growing on K12 and K12△tnaA. Is TRPA1 in their de-regulated gene list? Have other de-regulated genes been tested in this work?

We appreciate the concerns of the reviewer. We found that TRPA-1 is not included in the de-regulated gene list. Sonowal et al. focus on the gene expression profiles in worms from L1 larvae to young adults, whereas we pay attention to gene expression profiles in worms from young adults to aged worms. Thus, we did not test the de-regulated genes in their work.

8) How does indole activate TRPA1? In the absence of trpa1, what is the concentration of indole in worms? Since TRPA1 is a channel, is there any possibility that TRPA1 is involved in the transport of indole? It is really interesting and surprising that neuronal TRPA-1, but not intestinal TRPA-1, mediates the beneficial effect of indole. How does indole specifically activate TRPA-1 in neurons to preserve the longevity of worms?

We appreciate the concerns of the reviewer. TRPA1 is a nonselective cation channel permeable to Ca2+, Na+, and K+ (Zygmunt & Hogestatt, 2014). It is unlikely that TRPA1 is capable of transporting heterocyclic organic compounds, such as indole.

In response to the reviewer’s suggestion, we detected the content of indole in trpa-1(ok999) worms. We found that the levels of indole in trpa-1(ok999) worms were slightly increased in worms on days 4 and 7, compared to those in WT worms on days 4 and 7 (Author response image 3).

Recently, Ye et al. have demonstrated that indole and indole-3-carboxaldehyde (IAld) are agonists of TRPA1, which is conserved in vertebrates (Ye et al., 2021). Thus, it is mostly likely that indole acts as an agonist of TRPA-1 in C. elegans by directly binding to TRPA-1. One possibility is that activation of TRPA-1 in neurons by indole could induce a pathway that release a neurotransmitter, which in turn triggers a signaling pathway to extend lifespan of worms via activating DAF-16 in a non-cell autonomous manner. In contrast, the activation of TRPA-1 in the intestine by indole is unable to release such a neurotransmitter. Indeed, TRPA1 induces the releasing of calcitonin gene-related peptide in perivascular sensory nerves, leading to membrane hyperpolarization and arterial dilation on smooth muscle cells (Talavera et al., 2020). Moreover, the activation of TRPA1 by indole and IAld induces the secretion of the neurotransmitter serotonin in zebrafish (Ye et al., 2021).

Author response image 3.

The indole levels in trpa-1 mutants are increased on days 4 and 7, compared with those in WT worms. *P < 0.05.

9) How neuronal- and intestinal-specific knockdown of trpa-1 by RNAi was conducted? And what is the tissue-specific expression pattern of trap-1? Speculating how indole was transported to neuron cells is pretty appealing.

We appreciate the concerns of the reviewer. SID-1 is required cell-autonomously for systemic RNAi (Winston, Molodowitch, & Hunter, 2002). Thus, the sid-1 mutants are resistant to RNAi in the neuronal- and intestinal-specific RNAi strains, sid-1 was expressed under control of the neuronal-specific unc-119 and the intestinal-specific vha-6 promoters, respectively. Although it has been reported that TRPA-1 is expressed in neurons, muscles, hypodermal cells, and the intestine, Xiao et al. proved that only TRPA-1 expressed in the intestine and neurons contributes to life extension at low temperature (Xiao et al., 2013). The transporter of indole has not been identified. In Arabidopsis, ATP-binding cassette (ABC) transporter G family 37(ABCG37) has been reported to transport a range of indole derivatives (Ruzicka et al., 2010). However, all fifteen C. elegans ABC transporters share less than 30% sequence identity with ABCG37. Thus, it is impossible to determine which one is the transport channel for indole and indole derivatives in C. elegans.

10) Supplementation with indole only up-regulated the expression of lys-7 and lys-8 in worms subjected to intestinal-specific (Figure 7-figure supplement 2C), but not neuronal-specific, RNAi of trpa-1 (Figure 7-figure supplement 2D). If this is the case, should the addition of indole specifically induce the expression of lys-7p::gfp or lys-8p::gfp in neurons?

We really appreciate this reviewer’s preciseness. Indeed, lys-7 and lys-8 are expressed in both neurons and the intestine (Author response image 4A and 7B). However, the expression of lys-8p::gfp and lys-7p::gfp in neurons was not altered in worms after treatment with indole or knockdown of trpa-1 by RNAi (Author response image 4C and 4D).

Author response image 4.

The expression of LYS-7 and LYS-8 in neurons is not altered after treatment with indole or knockdown of trpa-1 by RNAi. (A and C) Representative images of lys-7p::gfp (A) and lys-8p::gfp (C). Both lys-7 and lys-8 could be expressed in neurons and the intestine. (B and D) Quantification of fluorescent intensity of lys-7p::gfp (B) and lys-8p::gfp (D) in neurons. These results are means ± SD of three independent experiments. ns, not significant.

11) The authors demonstrated that K-12△tnaA strain had undetectable tnaA mRNA or indole levels. Furthermore, the deletion of tnaA significantly inhibited the nuclear translocation of DAF-16 in worms. However, mutations in E. coli still have non-specific effects as there are several transposon insertions or polar mutations influencing downstream genes. The authors should demonstrate that only disruption of TnaA causes the failure of nuclear translocation of DAF-16.

In response to the reviewer’s suggestion, we rescued the expression of tnaA in the K-12 △tnaA strain. As expected, the indole level of from the supernatant in the K12 △tnaA::tnaA strain cultures was 34.1 μmol/L, which was comparable of that in the K12 strain cultures (42.5 μmol/L)（new Figure 2-figure supplement 4D). In addition, DAF-16 nuclear accumulation was increased in worms grown in the K12 △tnaA::tnaA strain on days 4 and 7 (new Figure 2-figure supplement 4E).

Author Response

Reviewer #1 (Public Review):

The study by Akter et al demonstrates that astrocyte-derived L-lactate plays a key role in schema memory formation and promotes mitochondrial biogenesis in the Anterior Cingulate Cortex (ACC).

The main tool used by the authors is the DREADD technology that allows to pharmacologically activate receptors in a cell-specific manner. In the study, the authors used the DREADD technique to activate appropriately transfected astrocytes, a subtype of muscarinic receptor that is not normally present in cells. This receptor being coupled to a Gi-mediated signal transduction pathway inhibiting cAMP formation, the authors could demonstrate cell-(astrocyte) specific decreases in cAMP levels that result in decreased L-lactate production by astrocytes.

Behaviorally this pharmacological manipulation results in impairments of schema memory formation and retrieval in the ACC in flavor-place paired associate paradigms. Such impairments are prevented by co-administration of L-lactate.

The authors also show that activation of Gi signaling resulting in L-lactate decreased release by astrocytes impairs mitochondrial biogenesis in neurons in an L-lactate reversible manner.

By using MCT 2 inhibitors and an NMDAR antagonist the authors conclude that the molecular mechanisms underlying the observed effects are mediated by L-lactate entering neurons through MCT2 transporters and involve NMDAR.

Overall, the article's conclusions are warranted by the experimental evidence, but some weak points could be addressed which would make the conclusions even stronger.

The number of animals in some of the experiments is on the low side (4 to 6).

In the revised manuscript, we have increased the animal numbers in two key experimental groups (hM4Di-CNO and Control groups) of behavioral experiments. Now the animal numbers in different groups are as follows:

• 15 rats in hM4Di-CNO group

o Further divided into two subgroups for probe tests (PT1-4) conducted during flavor-place paired associate training; 8 rats in the hM4Di-CNO (saline) and 7 rats in the hM4Di-CNO (CNO) subgroups receiving I.P. saline or I.P. CNO, respectively, before these PTs.

• 8 rats in the Control group

• 7 rats in the Rescue group (hM4Di-CNO+L-lactate)

• 4 rats in the Control-CNO group. Animal number in this group was not increased as it was apparent from these 4 rats that CNO alone was not impairing the PA learning and memory retrieval in these rats (AAV8-GFAP-mCherry injected). Their result was very similar to the control group. Additionally, in a previous study (Liu et al., 2022), we showed that CNO administration in the rats injected with AAV8-GFAP-mCherry into the hippocampus does not show any impairments in schema.

Also, in the newly added open field test experiments to investigate the locomotor activity as suggested by the Reviewer #2, 8 rats were used in each group.

The use of CIN to inhibit MCT2 is not optimal. Authors may want to decrease MCT2 expression by using antisense oligonucleotides.

In the revised manuscript, we have conducted the experiment using MCT2 antisense oligodeoxynucleotide (ODN) as suggested.

To test whether the L-lactate-induced neuronal mitochondrial biogenesis is dependent on MCT2, we bilaterally injected MCT2 antisense oligodeoxynucleotide (MCT2-ODN, n=8 rats, 2 nmol in 1 μl PBS per ACC) or scrambled ODN (SC-ODN, n=8 rats, 2 nmol in 1 μl PBS per ACC) into the ACC. After 11 hours, bilateral infusion of L-lactate (10 nmol, 1 μl) or ACSF (1 μl) was given into the ACC and the rats were kept in the PA event arena. After 60 mins (12 hours from MCT2-ODN or SC-ODN administration), the rats were sacrificed. As shown in Author response image 1B, SC-ODN+L-lactate group showed significantly increased relative mtDNA copy number compared to the SC-ODN+ACSF group (p<0.001, ANOVA followed by Tukey's multiple comparisons test). However, this effect was completely abolished in MCT2-ODN+L-lactate group, suggesting that MCT2 is required for the L-lactate-induced mitochondrial biogenesis in the ACC.

We have integrated this new data and results in the revised manuscript.

Author response image 1.

Mitochondrial biogenesis by L-lactate is dependent on MCT2 and NMDAR. A. Experimental design to investigate whether MCT2 and NMDAR activity are required for L-lactate-induced mitochondrial biogenesis. B and C. mtDNA copy number abundance in the ACC of different rat groups relative to nDNA. Data shown as mean ± SD (n=4 rats in each group). ***p<0.001, ANOVA followed by Tukey's multiple comparisons test.

The experiment using AVP to block NMDAR only partially supports the conclusions. Indeed, blocking NMDAR will knock down any response that involves these receptors, whether L-lactate is necessary or not.

In the current study we found that Astrocytic Gi activation in the ACC reduced L-lactate level in the ECF of ACC which was also associated with decreased PGC-1α/SIRT3/ATPB/mtDNA abundance suggesting downregulation of mitochondrial biogenesis pathway. We also found that exogenous administration of L-lactate into the ACC of astrocytic Gi-activated rats rescued this downregulation. In line with this, in a recently published study (Akter et al., 2023), we found upregulation of mitochondrial biogenesis pathway in the hippocampus neurons of exogenous L-lactate-treated anesthetized rats. Another recent study has demonstrated that exercise-induced L-lactate release from skeletal muscle or I.P. injection of L-lactate can induce hippocampal PGC-1α (which is a master regulator of mitochondrial biogenesis) expression and mitochondrial biogenesis in mice (Park et al., 2021). Together, these results provide compelling evidence that L-lactate promotes mitochondrial biogenesis.

L-lactate is known to promote expression of synaptic plasticity genes like Arc, c-Fos, and Zif268 in neurons (Yang et al., 2014). After entry into the neuronal cytoplasm, mainly through MCT2, it is converted into pyruvate by lactate dehydrogenase 1 (LDH1). This conversion also produces NADH, affecting the redox state of the neuron. NADH positively modulates the activity of NMDAR resulting in enhanced Ca2+ currents, the activation of intracellular signaling cascades, and the induction of the expression of plasticity-associated genes (Yang et al., 2014; Magistretti & Allaman, 2018). The study demonstrated that L-lactate–induced plasticity gene expression was abolished in the presence of NMDAR antagonists including D-APV (Yang et al., 2014). These results suggested that the MCT2 and NMDAR are key players in the regulation of L-lactate induced plasticity gene expression.

In the current study, we investigated whether similar mechanisms might be involved in L-lactate-induced neuronal mitochondrial biogenesis. We now used MCT2 antisense oligodeoxynucleotide to decrease the expression of MCT2 (as mentioned in the previous response and Author response image 1B) and showed that MCT2 is necessary for L-lactate-induced mitochondrial biogenesis to manifest, indicating that L-lactate’s entry into the neuron is required. As mentioned before, after entry into neuron, L-lactate is converted into pyruvate by LDH, which also produce NADH, which in turn potentiates NMDAR activity. Therefore, we investigated whether NMDAR activity is required for L-lactate-induced mitochondrial biogenesis. We used D-APV to inhibit NMDAR (Author response image 1C) and found that L-lactate does not increase mtDNA copy number abundance if D-APV is given, suggesting that NMDAR activity is required for L-lactate to promote mitochondrial biogenesis.

NMDAR serves diverse functions. Therefore, as mentioned by the reviewer, blocking NMDAR may knock down many such functions. While our current data only suggests the involvement of MCT2 and NMDAR in the upregulation of mitochondrial biogenesis by L-lactate, we have not investigated other mechanisms and pathways modulating mitochondrial biogenesis that are either dependent or independent of MCT2 and NMDAR activity. Further studies are needed in future to dissect and better understand this interesting observation. We have now clarified this in the discussion section of the manuscript.

Is inhibition of glycogenolysis involved in the observed effects mediated by Gi signaling? Indeed, L-lactate is formed both by glycolysis and glycogenolysis. The authors could test whether the glycogen metabolism-inhibiting drug DAB would mimic the effects of Gi activation.

In this study we have shown that astrocytic Gi activation in the ACC leads to a decrease in the cAMP and L-lactate. L-lactate is produced by glycogenolysis and glycolysis. cAMP in astrocytes acts as a trigger for L-lactate production (Choi et al., 2012; Horvat, Muhič, et al., 2021; Horvat, Zorec, et al., 2021; Zhou et al., 2021) by promoting glycogenolysis and glycolysis (Vardjan et al., 2018; Horvat, Muhič, et al., 2021; Horvat, Zorec, et al., 2021). Therefore, one promising explanation of reduced L-lactate level observed in our study is the reduction of L-lactate production in the astrocyte due to decreased glycogen metabolism as a result of decreased cAMP. We have now mentioned this in the discussion.

DAB is an inhibitor of glycogen phosphorylase that suppresses L-lactate production. It was shown to impair memory by decreasing L-lactate (Newman et al., 2011; Suzuki et al., 2011; Iqbal et al., 2023). As we found that the impairment in the schema memory and mitochondrial biogenesis was associated with decreased L-lactate level in the ACC and that the exogenous L-lactate administration can rescue the impairments, it is likely that DAB will mimic the effect of Gi activation in terms of schema memory and mitochondrial biogenesis. However, further study is needed to confirm this.  

Reviewer #2 (Public Review):

The manuscript of Akter et al is an important study that investigates the role of astrocytic Gi signaling in the anterior cingulate cortex in the modulation of extracellular L-lactate level and consequently impairment in flavor-place associates (PA) learning. However, whereas some of the behavioral observations and signaling mechanism data are compelling, the conclusions about the effect on memory are inadequate as they rely on an experimental design that does not allow to differentiate acute or learning effect from the effect outlasting pharmacological treatments, i.e. effect on memory retention. With the addition of a few experiments, this paper would be of interest to the larger group of researchers interested in neuron-glia interactions during complex behavior.

• Largely, I agree with the authors' conclusion that activating Gi signaling in astrocytes impairs PA learning, however, the effect on memory retrieval is not that obvious. All behavioral and molecular signaling effects described in this study are obtained with the continuous presence of CNO, therefore it is not possible to exclude the acute effect of Gi pathway activation in astrocytes. What will happen with memory on retrieval test when CNO is omitted selectively during early, middle, or late session blocks of PA learning?

We have now added 8 more rats to the hM4Di-CNO group (i.e., the group with astrocytic Gi activation) to clarify the memory retrieval. These rats underwent flavor-place paired associate (PA) training similar to the previously described rats (n=7) of this group, that is they received CNO 30 minutes before and 30 minutes after the PA training sessions (S1-2, S4-8, S10-17). However, contrasting to the previous rats of this group which received CNO before PTs (PT1, PT2, PT3), we omitted the CNO (instead administered I.P. saline) selectively on these PTs conducted at the early, middle, and late stage of PA training, as suggested by the reviewer. These newly added rats did not show memory retrieval in these PTs, suggesting that the rats were not learning the PAs from the PA training sessions. See Author response image 2C-E, where this subgroup is denoted as hM4Di-CNO (Saline).

We then continued more PA training sessions (S21 onwards, Author response image 2B) for these rats without CNO. They gradually learned the PAs. PTs (PT5, PT6, PT7; Author response image 2G-I) were done during this continuation phase of PA training; once without CNO (i.e., with I.P. saline instead), and another one with CNO. As seen in the Author response image 2H and 2I, they retrieved the memory when PT6 and PT7 were done without CNO. However, if these PTs were done with CNO, they could not retrieve the memory. Together these results suggest that ACC astrocytic Gi activation by CNO during PT can impair memory retrieval in rats which have already learned the PAs.

As shown in the Author response image 2B, we replaced two original PAs with two new PAs (NPA 9 and 10) at S34. This was followed by PT8 (S35). As seen in Author response image 2J, these rats retrieved the NPA memory if the PT is done without CNO. However, they could not retrieve the NPA memory if the PT was done with CNO. This result suggests that ACC astrocytic Gi activation by CNO during PT can impair NPA memory retrieval.

In summary, these data show that astrocytic Gi activation in the ACC can impair PA memory retrieval. We have integrated this new data and results in the revised manuscript.

Author response image 2.

A. PI (mean ± SD) during the acquisition of the six original PAs (OPAs) (S1-2, 4-8, 10-17) and new PAs (NPAs) (S19) of the control (n=8), hM4Di-CNO (n=15), and rescue (hM4Di-CNO+L-lactate) (n=7) groups. From S6 onwards, hM4Di-CNO group consistently showed lower PI compared to control. However, concurrent L-lactate administration into the ACC (rescue group) can rescue this impairment. B. PI (mean ± SD) of hM4Di-CNO group (n=8) from S21 onwards showing gradual increase in PI when CNO was withdrawn. C, D, and E. Non-rewarded PTs (PT1, PT2, and PT3 conducted on S3, S9, and S18, respectively) to test memory retrieval of OPAs for the control, hM4Di-CNO, and rescue groups. The percentage of digging time at the cued location relative to that at the non-cued locations are shown (mean ± SD). In both PT2 and PT3, the control group spent significantly more time digging the cued sand well above the chance level, indicating that the rats learned OPAs and could retrieve it. Contrasting to this, hM4Di-CNO group did not spend more time digging the cued sand well above the chance level irrespective of CNO administration before the PTs. The rescue group showed results similar to the hM4Di-CNO group if CNO is given without L-lactate. On the other hand, they showed results similar to the control group if L-lactate is concurrently given with CNO, indicating that this group learned OPAs and could retrieve it. p < 0.05, p < 0.01, p < 0.001, one-sample t-test comparing the proportion of digging time at the cued sand well with the chance level of 16.67%. F. Non-rewarded PT4 (S20) which was conducted after replacing two OPAs with two NPAs (NPA 7 & 8) in S19 for the control, hM4Di-CNO, and rescue groups. Results show that the control group spent significantly more time digging the new cued sand well above the chance level indicating that the rats learned the NPAs from S19 and could retrieve it in this PT. Contrasting to this, hM4Di-CNO group did not spend more time digging the new-cued sand well above the chance level irrespective of CNO administration before the PT. The rescue group showed results similar to the hM4Di-CNO group if CNO is given without L-lactate. On the other hand, they showed results similar to the control group if L-lactate is concurrently given with CNO indicating that this group learned NPAs from S19 and could retrieve it. p < 0.001, one-sample t-test comparing the proportion of digging time at the new cued sand well with the chance level of 16.67%. G, H, and I. Non-rewarded PTs (PT5, PT6, and PT7 conducted on S23, S27, and S33, respectively) to test memory retrieval of OPAs for the hM4Di-CNO group. In both PT6 and PT7, the rats spent significantly more time digging the cued sand well above the chance level if the tests are done without CNO, indicating that the rats learned the OPAs and could retrieve it. However, CNO prevented memory retrieval during these PTs. p < 0.001, one-sample t-test comparing the proportion of digging time at the cued sand well with the chance level of 16.67%. J. Non-rewarded PT4 (S35) which was conducted after replacing two OPAs with two NPAs (NPA 9 & 10) in S34 for the hM4Di-CNO group. Results show that the rats spent significantly more time digging the new cued sand well above the chance level if CNO was not given before the PT, indicating that the rats learned the NPAs from S34 and could retrieve it in this PT. However, if CNO is given before the PT, the retrieval is impaired. *p < 0.001, one-sample t-test comparing the proportion of digging time at the new cued sand well with the chance level of 16.67%.

• I found it truly exciting that the administration of exogenous L-lactate is capable to rescue CNO-induced PA learning impairment, when co-applied. Would it be possible that this treatment has a sensitivity to a particular stage of learning (acquisition, consolidation, or memory retrieval) when L-lactate administration would be the most efficacious?

The hM4Di-CNO group, when continued with PA training without CNO (S21-S32) (Author response image 2B), was able to learn the six original PAs (OPAs). In the PT7 done at S33 (Author response image 2I), this group of rats was able to retrieve the memory if the test was done without CNO but could not retrieve the memory if CNO was given. Similarly, the Rescue group (hM4Di-CNO+L-lactate) (Author response image 2A), which received both CNO and L-lactate during PA training sessions (S1-S17), they were able to learn the OPAs. And at PT3 done at S18 (Author response image 2E), these rats were able to retrieve the memory when the test was done with CNO+L-lactate but not if the test is done with only CNO. Together, these results clearly show that ACC astrocytic Gi activation with CNO impairs memory retrieval and exogenous L-lactate can rescue the impairment. Therefore, it can be concluded that the memory retrieval is sensitive to L-lactate.

The PA learning is hippocampus-dependent. Over the course of repeated PA training, systems consolidation occurs in the ACC, after which the already learned PA memory (schema) becomes hippocampus-independent (Tse et al., 2007; Tse et al., 2011). A higher activation (indicated by expression of c-Fos) in the hippocampus relative to the ACC during the early period of schema development, and the reverse at the late stage was observed in our previous study (Liu et al., 2022). However, rapid assimilation of new PA into the ACC requires simultaneous activation/retrieval of previous schema from ACC and hippocampus dependent new PA learning (Tse et al., 2007; Tse et al., 2011). During new PA learning, increase of c-Fos neurons in both CA1 and ACC was detected (Liu et al., 2022).

Our hM4Di-CNO group received CNO 30 mins before and after each PA training session in S1-S17 (Author response image 2A). Also, the Rescue group similarly received CNO+L-lactate before and after each PA training session in S1-S17. Therefore, while this study design allowed us to conclude that ACC astrocytic Gi activation impairs PA learning and that exogenous L-lactate can rescue the impairment, it does not allow clear differentiation of the effects of these treatments on memory acquisition and consolidation. Further studies are needed to investigate this.

• The hypothesis that observed learning impairments could be associated with diminished mitochondrial biogenesis caused by decreased l-lactate in the result of astrocytic Gi-DREADDS stimulation is very appealing, but a few key pieces of evidence are missing. So far, the hypothesis is supported by experiments demonstrating reduced expression of several components of mitochondrial membrane ATP synthase and a decrease in relative mtDNA copy numbers in ACC of rats injected with Gi-DREADDs. L-lactate injections into ACC restored and even further increased the expression of the above-mentioned markers. Co-administration of NMDAR antagonist D-APV or MCT-2 (mostly neuronal) blocker 4-CIN with L-lactate, prevented L-lactate-induced increase in relative mtDNA copy. I am wondering how the interference with mitochondrial biogenesis is affecting neuronal physiology and if it would result in impaired PA learning or schema memory.

The observation of diminished mitochondrial biogenesis in the astrocytic Gi-activated rats that showed impaired PA learning is exciting. However, our study does not provide experimental data on how mitochondrial biogenesis could be associated with impaired PA learning and schema memory. Results from several previous studies linked mitochondrial biogenesis and its regulators such as PGC-1α and SIRT3 to diverse neuronal and cognitive functions as described in the discussion section of the manuscript. In the revised manuscript, we have provided further discussion as follows to discuss potential mechanisms:

“In this study, we have demonstrated that ACC astrocytic Gi activation impairs PA learning and schema formation, PA memory retrieval, and NPA learning and retrieval by decreasing L-lactate level in the ACC. Although we have shown that these impairments are associated with diminished expression of proteins of mitochondrial biogenesis, the precise mechanisms of how astrocytic Gi activation affects neuronal functions and schema memory remain to be elucidated. We previously demonstrated that neuronal inhibition in either the hippocampus or the ACC impairs PA learning and schema formation (Hasan et al., 2019). In another recent study (Liu et al., 2022), we showed that astrocytic Gi activation in the CA1 impaired PA training-associated CA1-ACC projecting neuronal activation. Yao et al. recently showed that reduction of astrocytic lactate dehydrogenase A (an enzyme that reversibly catalyze L-lactate production from pyruvate) in the dorsomedial prefrontal cortex reduces L-lactate levels and neuronal firing frequencies, promoting depressive-like behaviors in mice (Yao et al., 2023). These impairments could be rescued by L-lactate infusion. It is possible that the impairment in PA learning and schema observed in our study might have involved a similar functional consequence of reduced neuronal activity in the ACC neurons upon astrocytic Gi activation.

Schema consolidation is associated with synaptic plasticity-related gene expression (such as Zif268, Arc) in the ACC (Tse et al., 2011). L-lactate, after entry into neurons, can be converted to pyruvate during which NADH is also produced, promoting synaptic plasticity-related gene expression by potentiating NMDA signaling in neurons (Yang et al., 2014; Margineanu et al., 2018). Furthermore, L-lactate acts as an energy substrate to fuel learning-induced de novo neuronal translation critical for long-term memory (Descalzi et al., 2019). On the other hand, mitochondria play crucial role in fueling local translation during synaptic plasticity (Rangaraju et al., 2019). Therefore, it could be hypothesized that the rescue of astrocytic Gi activation-mediated impairment of schema by exogenous L-lactate could have been mediated by facilitating synaptic plasticity-related gene expression by directly fueling the protein translation, potentiating NMDA signaling, as well as increasing mitochondrial capacity for ATP production by promoting mitochondrial biogenesis. Furthermore, the potential involvement of HCAR1, a receptor for L-lactate that may regulate neuronal activity (Bozzo et al., 2013; Tang et al., 2014; Herrera-López & Galván, 2018; Abrantes et al., 2019), cannot be excluded. Future research could explore these potential mechanisms, examining the interactions among them, and determining their relative contributions to schema. Our previous study also showed that ACC myelination is necessary for PA learning and schema formation, and that repeated PA training is associated with oligodendrogenesis in the ACC (Hasan et al., 2019). Oligodendrocytes facilitate fast, synchronized, and energy efficient transfer of information by wrapping axons in myelin sheath. Furthermore, they supply axons with glycolysis products, such as L-lactate, to offer metabolic support (Fünfschilling et al., 2012; Lee et al., 2012). The association of oligodendrogenesis and myelination with schema memory may suggest an adaptive response of oligodendrocytes to enhance metabolic support and neuronal energy efficiency during PA learning. Given the impairments in PA learning observed in the ACC astrocytic Gi-activated rats in the current study, it is reasonable to conclude that the direct metabolic support to axons provided by oligodendrocytes is not sufficient to rescue the schema impairments caused by decreased L-lactate levels upon astrocytic Gi activation. On the other hand, L-lactate was shown to be important for oligodendrogenesis and myelination (Sánchez-Abarca et al., 2001; Rinholm et al., 2011; Ichihara et al., 2017). Therefore, it is tempting to speculate that a decrease in L-lactate level may also impede oligodendrogenesis and myelination, consequently preventing the enhanced axonal support provided by oligodendrocytes and myelin during schema learning. Recently, a study has demonstrated that upon demyelination, mitochondria move from the neuronal cell body to the demyelinated axon (Licht-Mayer et al., 2020). Enhancement of this axonal response of mitochondria to demyelination, by targeting mitochondrial biogenesis and mitochondrial transport from the cell body to axon, protects acutely demyelinated axons from degeneration. Given the connection between schema and increased myelination, it remains an open question whether L-lactate-induced mitochondrial biogenesis plays a beneficial role in schema through a similar mechanism. Nevertheless, our results contribute to the mounting evidence of the glial role in cognitive functions and underscores the new paradigm in which glial cells are considered as integral players in cognitive functions alongside neurons. Disruption of neurons, myelin, or astrocytes in the ACC can disrupt PA learning and schema memory.”

Reviewer #3 (Public Review):

Akter et al. investigated how the astroglial Gi signaling pathway in the rat anterior cingulate cortex (ACC) affects cognitive functions, in particular schema memory formation. Using a stereotactic approach they intracranially introduced AAV8 vectors carrying mCherry-tagged hM4Di DREADD (Designer Receptor Exclusively Activated by Designer Drugs) under astrocyte selective GFAP promotor (AAV8-GFAP-hM4Di-mCherry) into the AAC region of the rat brain. hM4Di DREADD is a genetically modified form of the human M4 muscarinic (hM4) receptor insensitive to endogenous acetylcholine but is activated by the inert clozapine metabolite clozapine-N-oxide (CNO), triggering the Gi signaling pathway. The authors confirmed that hM4Di DREADD is selectively expressed in astrocytes after the application of the AAV8 vector by analysing the mCherry signals and immunolabeling of astrocytes and neurons in the ACC region of the rat brain. They activated hM4Di DREADD (Gi signalling) in astrocytes by intraperitoneal administration of CNO and measured cognitive functions in animals after CNO administration. Activation of Gi signaling in astrocytes by CNO application decreased paired-associate (PA) learning, schema formation, and memory retrieval in tested animals. This was associated with a decrease in cAMP in astrocytes and L-lactate in extracellular fluid as measured by immunohistochemistry in situ and in awake rats by microdialysis, respectively. Administration of exogenous L-lactate rescued the astroglial Gi-mediated deficits in PA learning, memory retrieval, and schema formation, suggesting that activation of astroglial Gi signalling downregulates L-lactate production in astrocytes and its transport to neurons affecting memory formation. Authors also show that expression level of proteins involved in mitochondrial biogenesis, which is associated with cognitive functions, is decreased in neurons, when Gi signalling is activated in astrocytes, and rescued when exogenous L-lactate is applied, suggesting the implication of astrocyte-derived L-lactate in the maintenance of mitochondrial biogenesis in neurons. The latter depended on lactate MCT2 transporter activity and glutamate NMDA receptor activity.

The paper is very well written and discussed. The conclusions of this paper are well supported by the data. Although this is a study that uses established and previously published methodologies, it provides new insights into L-lactate signalling in the brain, particularly in AAC, and further confirms the role of astroglial L-lactate in learning and memory formation. It also raises new questions about the molecular mechanisms underlying astrocyte-derived L-lactate-mediated mitochondrial biogenesis in neurons and its contribution to schema memory formation.

• The authors discuss astrocytic L-lactate signalling without considering the recently discovered L-lactate-sensitive Gs and Gi protein-coupled receptors in the brain, which are present in both astrocytes and neurons. The use of nonendogenous L-lactate receptor agonists (Compound 2, 3-chloro-5-hydroxybenzoic acid) would clarify the implication of L-lactate receptor signalling in schema memory formation.

In the revised manuscript, we have included this point in the discussion section to mention the potential role of HCAR1 in schema memory as follows:

“Schema consolidation is associated with synaptic plasticity-related gene expression (such as Zif268, Arc) in the ACC (Tse et al., 2011). L-lactate, after entry into neurons, can be converted to pyruvate during which NADH is also produced, promoting synaptic plasticity-related gene expression by potentiating NMDA signaling in neurons (Yang et al., 2014; Margineanu et al., 2018). Furthermore, L-lactate acts as an energy substrate to fuel learning-induced de novo neuronal translation critical for long-term memory (Descalzi et al., 2019). On the other hand, mitochondria play crucial role in fueling local translation during synaptic plasticity (Rangaraju et al., 2019). Therefore, it could be hypothesized that the rescue of astrocytic Gi activation-mediated impairment of schema by exogenous L-lactate could have been mediated by facilitating synaptic plasticity-related gene expression by directly fueling the protein translation, potentiating NMDA signaling, as well as increasing mitochondrial capacity for ATP production by promoting mitochondrial biogenesis. Furthermore, the potential involvement of HCAR1, a receptor for L-lactate that may regulate neuronal activity (Bozzo et al., 2013; Tang et al., 2014; Herrera-López & Galván, 2018; Abrantes et al., 2019), cannot be excluded. Future research could explore these potential mechanisms, examining the interactions among them, and determining their relative contributions to schema.”

• The use of control animals transduced with an "empty" AAV9 vector (AAV8-GFAP-mCherry) compared with animals transduced with AAV8-GFAP-hM4Di-mCherry throughout the study would strengthen the results of this study, since transfection itself, as well as overexpression of the mCherry protein, may affect cell function.

We thank the reviewer for pointing this. The schema experiment includes a control group (Control-CNO group) of rats injected with AAV8-GFAP-mCherry bilaterally into the ACC. As shown in Author response image 3, after habituation and pretraining, these rats were trained for PA learning similarly to the other groups. Before 30 mins and after 30 mins of each PA training session, they received I.P. CNO. The PA learning, schema formation, memory retrieval, NPA learning and retrieval, and latency (time needed to commence digging at the correct well) were similar to the control group of rats. This result is consistent with our previous study where rats bilaterally injected with AAV8-GFAP-mCherry into CA1 of hippocampus did not show impairments in PA learning and schema formation upon CNO treatment (Liu et al., 2022).

Author response image 3.

A. PI (mean ± SD) during the acquisition of the original six PAs (OPAs) (S1-2, 4-8, 10-17) and new PAs (NPAs) (S19) of the control (n=6) and control-CNO (n=4) groups. B. Non-rewarded PTs (PT1, PT2, and PT3 done on S3, S9, and S18, respectively) to test memory retrieval of OPAs for the control-CNO group. C. Non-rewarded PT4 (S20) which was done after replacing two OPAs with two NPAs (NPA 7 & 8) in S19 for the control-CNO group. D. Latency (in seconds) before commencing digging at the correct well for control and control-CNO groups. Data shown as mean ± SD.

References

Abrantes, H. d. C., Briquet, M., Schmuziger, C., Restivo, L., Puyal, J., Rosenberg, N., Rocher, A.-B., Offermanns, S., & Chatton, J.-Y. (2019). The Lactate Receptor HCAR1 Modulates Neuronal Network Activity through the Activation of Gα and Gβγ Subunits. The Journal of Neuroscience, 39(23), 4422-4433. https://doi.org/10.1523/jneurosci.2092-18.2019

Akter, M., Ma, H., Hasan, M., Karim, A., Zhu, X., Zhang, L., & Li, Y. (2023). Exogenous L-lactate administration in rat hippocampus increases expression of key regulators of mitochondrial biogenesis and antioxidant defense [Original Research]. Frontiers in Molecular Neuroscience, 16. https://doi.org/10.3389/fnmol.2023.1117146

Bozzo, L., Puyal, J., & Chatton, J.-Y. (2013). Lactate Modulates the Activity of Primary Cortical Neurons through a Receptor-Mediated Pathway. PLoS One, 8(8), e71721. https://doi.org/10.1371/journal.pone.0071721

Choi, H. B., Gordon, G. R., Zhou, N., Tai, C., Rungta, R. L., Martinez, J., Milner, T. A., Ryu, J. K., McLarnon, J. G., Tresguerres, M., Levin, L. R., Buck, J., & MacVicar, B. A. (2012). Metabolic communication between astrocytes and neurons via bicarbonate-responsive soluble adenylyl cyclase. Neuron, 75(6), 1094-1104. https://doi.org/10.1016/j.neuron.2012.08.032

Covelo, A., Eraso-Pichot, A., Fernández-Moncada, I., Serrat, R., & Marsicano, G. (2021). CB1R-dependent regulation of astrocyte physiology and astrocyte-neuron interactions. Neuropharmacology, 195, 108678. https://doi.org/https://doi.org/10.1016/j.neuropharm.2021.108678

Descalzi, G., Gao, V., Steinman, M. Q., Suzuki, A., & Alberini, C. M. (2019). Lactate from astrocytes fuels learning-induced mRNA translation in excitatory and inhibitory neurons. Communications Biology, 2(1), 247. https://doi.org/10.1038/s42003-019-0495-2

Endo, F., Kasai, A., Soto, J. S., Yu, X., Qu, Z., Hashimoto, H., Gradinaru, V., Kawaguchi, R., & Khakh, B. S. (2022). Molecular basis of astrocyte diversity and morphology across the CNS in health and disease. Science, 378(6619), eadc9020. https://doi.org/10.1126/science.adc9020

Fünfschilling, U., Supplie, L. M., Mahad, D., Boretius, S., Saab, A. S., Edgar, J., Brinkmann, B. G., Kassmann, C. M., Tzvetanova, I. D., Möbius, W., Diaz, F., Meijer, D., Suter, U., Hamprecht, B., Sereda, M. W., Moraes, C. T., Frahm, J., Goebbels, S., & Nave, K.-A. (2012). Glycolytic oligodendrocytes maintain myelin and long-term axonal integrity. Nature, 485(7399), 517-521. https://doi.org/10.1038/nature11007

Harris, R. A., Lone, A., Lim, H., Martinez, F., Frame, A. K., Scholl, T. J., & Cumming, R. C. (2019). Aerobic Glycolysis Is Required for Spatial Memory Acquisition But Not Memory Retrieval in Mice. eNeuro, 6(1). https://doi.org/10.1523/ENEURO.0389-18.2019

Hasan, M., Kanna, M. S., Jun, W., Ramkrishnan, A. S., Iqbal, Z., Lee, Y., & Li, Y. (2019). Schema-like learning and memory consolidation acting through myelination. FASEB J, 33(11), 11758-11775. https://doi.org/10.1096/fj.201900910R

Herrera-López, G., & Galván, E. J. (2018). Modulation of hippocampal excitability via the hydroxycarboxylic acid receptor 1. Hippocampus, 28(8), 557-567. https://doi.org/https://doi.org/10.1002/hipo.22958

Horvat, A., Muhič, M., Smolič, T., Begić, E., Zorec, R., Kreft, M., & Vardjan, N. (2021). Ca2+ as the prime trigger of aerobic glycolysis in astrocytes. Cell Calcium, 95, 102368. https://doi.org/https://doi.org/10.1016/j.ceca.2021.102368

Horvat, A., Zorec, R., & Vardjan, N. (2021). Lactate as an Astroglial Signal Augmenting Aerobic Glycolysis and Lipid Metabolism [Review]. Frontiers in Physiology, 12. https://doi.org/10.3389/fphys.2021.735532

Ichihara, Y., Doi, T., Ryu, Y., Nagao, M., Sawada, Y., & Ogata, T. (2017). Oligodendrocyte Progenitor Cells Directly Utilize Lactate for Promoting Cell Cycling and Differentiation. J Cell Physiol, 232(5), 986-995. https://doi.org/10.1002/jcp.25690

Iqbal, Z., Liu, S., Lei, Z., Ramkrishnan, A. S., Akter, M., & Li, Y. (2023). Astrocyte L-Lactate Signaling in the ACC Regulates Visceral Pain Aversive Memory in Rats. Cells, 12(1), 26. https://www.mdpi.com/2073-4409/12/1/26

Jourdain, P., Rothenfusser, K., Ben-Adiba, C., Allaman, I., Marquet, P., & Magistretti, P. J. (2018). Dual action of L-Lactate on the activity of NR2B-containing NMDA receptors: from potentiation to neuroprotection. Sci Rep, 8(1), 13472. https://doi.org/10.1038/s41598-018-31534-y

Kofuji, P., & Araque, A. (2021). G-Protein-Coupled Receptors in Astrocyte-Neuron Communication. Neuroscience, 456, 71-84. https://doi.org/10.1016/j.neuroscience.2020.03.025

Lee, Y., Morrison, B. M., Li, Y., Lengacher, S., Farah, M. H., Hoffman, P. N., Liu, Y., Tsingalia, A., Jin, L., Zhang, P. W., Pellerin, L., Magistretti, P. J., & Rothstein, J. D. (2012). Oligodendroglia metabolically support axons and contribute to neurodegeneration. Nature, 487(7408), 443-448. https://doi.org/10.1038/nature11314

Licht-Mayer, S., Campbell, G. R., Canizares, M., Mehta, A. R., Gane, A. B., McGill, K., Ghosh, A., Fullerton, A., Menezes, N., Dean, J., Dunham, J., Al-Azki, S., Pryce, G., Zandee, S., Zhao, C., Kipp, M., Smith, K. J., Baker, D., Altmann, D., Anderton, S. M., Kap, Y. S., Laman, J. D., Hart, B. A. t., Rodriguez, M., Watzlawick, R., Schwab, J. M., Carter, R., Morton, N., Zagnoni, M., Franklin, R. J. M., Mitchell, R., Fleetwood-Walker, S., Lyons, D. A., Chandran, S., Lassmann, H., Trapp, B. D., & Mahad, D. J. (2020). Enhanced axonal response of mitochondria to demyelination offers neuroprotection: implications for multiple sclerosis. Acta Neuropathologica, 140(2), 143-167. https://doi.org/10.1007/s00401-020-02179-x

Liu, S., Wong, H. Y., Xie, L., Iqbal, Z., Lei, Z., Fu, Z., Lam, Y. Y., Ramkrishnan, A. S., & Li, Y. (2022). Astrocytes in CA1 modulate schema establishment in the hippocampal-cortical neuron network. BMC Biol, 20(1), 250. https://doi.org/10.1186/s12915-022-01445-6

Magistretti, P. J., & Allaman, I. (2018). Lactate in the brain: from metabolic end-product to signalling molecule. Nat Rev Neurosci, 19(4), 235-249. https://doi.org/10.1038/nrn.2018.19

Margineanu, M. B., Mahmood, H., Fiumelli, H., & Magistretti, P. J. (2018). L-Lactate Regulates the Expression of Synaptic Plasticity and Neuroprotection Genes in Cortical Neurons: A Transcriptome Analysis. Front Mol Neurosci, 11, 375. https://doi.org/10.3389/fnmol.2018.00375

Netzahualcoyotzi, C., & Pellerin, L. (2020). Neuronal and astroglial monocarboxylate transporters play key but distinct roles in hippocampus-dependent learning and memory formation. Progress in Neurobiology, 194, 101888. https://doi.org/https://doi.org/10.1016/j.pneurobio.2020.101888

Newman, L. A., Korol, D. L., & Gold, P. E. (2011). Lactate produced by glycogenolysis in astrocytes regulates memory processing. PLoS One, 6(12), e28427. https://doi.org/10.1371/journal.pone.0028427

Park, J., Kim, J., & Mikami, T. (2021). Exercise-Induced Lactate Release Mediates Mitochondrial Biogenesis in the Hippocampus of Mice via Monocarboxylate Transporters. Front Physiol, 12, 736905. https://doi.org/10.3389/fphys.2021.736905

Peterson, S. M., Pack, T. F., & Caron, M. G. (2015). Receptor, Ligand and Transducer Contributions to Dopamine D2 Receptor Functional Selectivity. PLoS One, 10(10), e0141637. https://doi.org/10.1371/journal.pone.0141637

Rangaraju, V., Lauterbach, M., & Schuman, E. M. (2019). Spatially Stable Mitochondrial Compartments Fuel Local Translation during Plasticity. Cell, 176(1), 73-84.e15. https://doi.org/10.1016/j.cell.2018.12.013

Rinholm, J. E., Hamilton, N. B., Kessaris, N., Richardson, W. D., Bergersen, L. H., & Attwell, D. (2011). Regulation of oligodendrocyte development and myelination by glucose and lactate. J Neurosci, 31(2), 538-548. https://doi.org/10.1523/JNEUROSCI.3516-10.2011

Sánchez-Abarca, L. I., Tabernero, A., & Medina, J. M. (2001). Oligodendrocytes use lactate as a source of energy and as a precursor of lipids. Glia, 36(3), 321-329. https://doi.org/10.1002/glia.1119

Suzuki, A., Stern, S. A., Bozdagi, O., Huntley, G. W., Walker, R. H., Magistretti, P. J., & Alberini, C. M. (2011). Astrocyte-neuron lactate transport is required for long-term memory formation. Cell, 144(5), 810-823.

Tang, F., Lane, S., Korsak, A., Paton, J. F. R., Gourine, A. V., Kasparov, S., & Teschemacher, A. G. (2014). Lactate-mediated glia-neuronal signalling in the mammalian brain. Nature Communications, 5(1), 3284. https://doi.org/10.1038/ncomms4284

Tauffenberger, A., Fiumelli, H., Almustafa, S., & Magistretti, P. J. (2019). Lactate and pyruvate promote oxidative stress resistance through hormetic ROS signaling. Cell Death Dis, 10(9), 653. https://doi.org/10.1038/s41419-019-1877-6

Tse, D., Langston, R. F., Kakeyama, M., Bethus, I., Spooner, P. A., Wood, E. R., Witter, M. P., & Morris, R. G. (2007). Schemas and memory consolidation. Science, 316(5821), 76-82. https://doi.org/10.1126/science.1135935

Tse, D., Takeuchi, T., Kakeyama, M., Kajii, Y., Okuno, H., Tohyama, C., Bito, H., & Morris, R. G. (2011). Schema-dependent gene activation and memory encoding in neocortex. Science, 333(6044), 891-895. https://doi.org/10.1126/science.1205274

Vardjan, N., Chowdhury, H. H., Horvat, A., Velebit, J., Malnar, M., Muhič, M., Kreft, M., Krivec, Š. G., Bobnar, S. T., Miš, K., Pirkmajer, S., Offermanns, S., Henriksen, G., Storm-Mathisen, J., Bergersen, L. H., & Zorec, R. (2018). Enhancement of Astroglial Aerobic Glycolysis by Extracellular Lactate-Mediated Increase in cAMP [Original Research]. Frontiers in Molecular Neuroscience, 11. https://doi.org/10.3389/fnmol.2018.00148

Vezzoli, E., Cali, C., De Roo, M., Ponzoni, L., Sogne, E., Gagnon, N., Francolini, M., Braida, D., Sala, M., Muller, D., Falqui, A., & Magistretti, P. J. (2020). Ultrastructural Evidence for a Role of Astrocytes and Glycogen-Derived Lactate in Learning-Dependent Synaptic Stabilization. Cereb Cortex, 30(4), 2114-2127. https://doi.org/10.1093/cercor/bhz226

Wang, J., Tu, J., Cao, B., Mu, L., Yang, X., Cong, M., Ramkrishnan, A. S., Chan, R. H. M., Wang, L., & Li, Y. (2017). Astrocytic l-Lactate Signaling Facilitates Amygdala-Anterior Cingulate Cortex Synchrony and Decision Making in Rats. Cell Rep, 21(9), 2407-2418. https://doi.org/10.1016/j.celrep.2017.11.012

Yang, J., Ruchti, E., Petit, J. M., Jourdain, P., Grenningloh, G., Allaman, I., & Magistretti, P. J. (2014). Lactate promotes plasticity gene expression by potentiating NMDA signaling in neurons. Proc Natl Acad Sci U S A, 111(33), 12228-12233. https://doi.org/10.1073/pnas.1322912111

Yao, S., Xu, M.-D., Wang, Y., Zhao, S.-T., Wang, J., Chen, G.-F., Chen, W.-B., Liu, J., Huang, G.-B., Sun, W.-J., Zhang, Y.-Y., Hou, H.-L., Li, L., & Sun, X.-D. (2023). Astrocytic lactate dehydrogenase A regulates neuronal excitability and depressive-like behaviors through lactate homeostasis in mice. Nature Communications, 14(1), 729. https://doi.org/10.1038/s41467-023-36209-5

Yu, X., Zhang, R., Wei, C., Gao, Y., Yu, Y., Wang, L., Jiang, J., Zhang, X., Li, J., & Chen, X. (2021). MCT2 overexpression promotes recovery of cognitive function by increasing mitochondrial biogenesis in a rat model of stroke. Anim Cells Syst (Seoul), 25(2), 93-101. https://doi.org/10.1080/19768354.2021.1915379

Zhou, Z., Okamoto, K., Onodera, J., Hiragi, T., Andoh, M., Ikawa, M., Tanaka, K. F., Ikegaya, Y., & Koyama, R. (2021). Astrocytic cAMP modulates memory via synaptic plasticity. Proc Natl Acad Sci U S A, 118(3), e2016584118. https://doi.org/10.1073/pnas.2016584118

Zhu, J., Hu, Z., Han, X., Wang, D., Jiang, Q., Ding, J., Xiao, M., Wang, C., Lu, M., & Hu, G. (2018). Dopamine D2 receptor restricts astrocytic NLRP3 inflammasome activation via enhancing the interaction of β-arrestin2 and NLRP3. Cell Death Differ, 25(11), 2037-2049. https://doi.org/10.1038/s41418-018-0127-2

Author Response

Reviewer #2 (Public Review):

Zou et al. presented a comprehensive study where they generated single-cell RNA profiling of 138,982 cells from 13 samples of six patients including AK, squamous cell carcinoma in situ (SCCIS), cSCC, and their matched normal tissues, covering comprehensive clinical courses of cSCC. Using bioinformatics analysis, they identified keratinocytes, CAFs, immune cells, and their subpopulations. The authors further compared signatures within subpopulations of keratinocytes along with the clinical progression, especially basal cells, and identified many interesting genes. They also further validate some of the markers in an independent cohort using IHC, followed by some knockdown experiments using cSCC cell lines.

The strength of this study is the unique data set they have created, providing the community with invaluable resources to study and validate their findings. However, a lot of analyses were not robust enough to support the claims and conclusions in the paper. More clarification and cross-comparison with polished data are needed to further strengthen the study and claims.

1) Stemness markers were used. The authors used COL17A1, TP63, ITGB1, and ITGA3 to represent stemness markers. However, these were not common classic stemness markers used in cSCC. What is the source claiming these genes were stemness markers in cSCC? TP63 is a master regulator and early driver event in SCC, while COL17A1, ITGB1, and ITGA3 are all ECM genes. The authors need to use commonly well-known stem cell markers in cSCC, e.g., LGR5, to mark stem-like cells.

Thanks for raising this good point. We may not have provided a clear description of the markers COL17A1, TP63, ITGB1, and ITGA3 in the previous texts. We would like to clarify that these genes were used as the markers of epidermal stem cells in normal skin samples rather than tumor stem cells in cSCC. To avoid any possible misunderstanding, we revised the main text accordingly and added the references [4-11].

2) Cell proportion analysis. The authors used the mean proportions to compare different clinical groups for subpopulations of keratinocytes, e.g., Figure 2B, and Figure 5B. This is not robust, as no statistics can be derived from this. For example, from Fig 2A, it is clearly shown there is a high level of heterogeneity of cellular compositions for normal samples. One cannot say which group is higher or lower simply based on mean not variance as well.

We replotted the proportion analysis with statistics and presented the new graphs in Figure 2-figure supplement 1 for Figure 2B and Figure 5-figure supplement 1 for Figure 5B.

3) Basal tumour cells in SCCIS and SCC. To make the findings valid, authors need to compare these cells/populations with the keratinocyte cell populations defined by Ji et al. Cell 2020. Do basal-SCCIS-tumours cells, also in SCC samples, resemble any of the population defined in Ji et al. Ji et al. also had 10 match normal, thus the authors need to validate their findings of SCC vs normal analysis using the Ji et al. dataset.

Thanks for this valuable suggestion. We compared basal tumor cell in our study with the cell populations defined in Ji et al. Cell 2020 data using SingleCellNet [1]. The results showed that both the basal-SCCIS-tumor cells of SCCIS and basal tumor cells of cSCC in our study closely resemble the Tumor_KC_Basal subcluster defined in Ji et al’s paper (Figure 4-figure supplement 4, C and D). Tumor_KC_Basal highly expressed CCL2, CXCL14, FTH1, MT2A, which is consistent with our findings in basal tumor cells.

4) Copy number analysis. Authors used inferCNV to perform copy number analysis using scRNA-seq data and identified CNVs in subpopulations of keratinocytes in SCCIS and SCC. To ensure these CNVs were not artefacts, were some of the CNVs identified by inferCNV well-known copy number changes previously reported in cSCC?

In poorly-differentiated cSCC sample, the significant gains in chromosome 7, 9 and deletion in chromosome 10 were reported in previous study, indicating the reliability of the CNV analysis results (Figure 5-figure supplement 2) [12].

5) Pseudotime analysis lines 308-313. Not sure the pseudotime analysis added much as, as it is unclear two distinct subgroups were identified from this analysis. Suggest removing this to keep it neater

Thank you for this suggestion. We have deleted the result of pseudotime analysis.

6) Selection of candidate genes for validation using IHC and cell line work. For example, lines 205-206, lines 352-356 and lines 437-441, authors selected several genes associated with AK and SCC to further validate using IHC and cell line knockdown work. What are the criteria for selecting those genes for validation? It is unclear to readers how these were selected. It reads like a fishing experiment, then followed by a knockdown. Clear rationale/criteria need to be elaborated.

The first consideration of candidate gene selection is the fold change of expression. We have provided the statistical results of DEGs in Supplementary file 1b, 1h, 1j-1m. Then we selected top changed genes and conducted an extensive literature search on these genes. We prioritized genes that, although not directly associated with cSCC development, have a close relationship with related pathways, as determined through functional enrichment analysis. These genes were arranged for further verification experiments. We have added more details in main text and methods section.

7) TME. Compared to keratinocytes populations, the investigation of TME cells was weak. (a) can authors produce UMAP files just for T cells, DC cells, and fibroblasts separately? Figure 7B is not easy to see those subclusters. (b) similar to what was done for keratinocytes, can authors find differentially expressed clusters and genes among the different clinical groups, associated with disease progression? (c) where are the myeloid cell populations, also B cells?

Thank you for your suggestions. (a) We have added the UMAP files for T cells, DC cells and stromal cells separately in new Figure 7A. (b) We identified DEGs in TME cells among the different groups. Several key genes showed monotonically changing trends associated with disease progression. For example, with the increase of malignancy, FOS shows down-regulation while S100A8 and S100A9 monotonically increase in all three types of TME cells (Figure 7C). (c) We identified two types of myeloid cell populations, macrophage and monocyte derived DCs (MoDC). We didn’t find other myeloid cells, such as neutrophil. For B cells, there were only 28 B cells in poorly-differentiated cSCC sample, which didn’t meet the threshold for further cell-cell communication analysis.

8) Heat shock protein genes line 327-329. HSP signature was well-known to be induced via tissue dissociation and library prep during the scRNA experiment. How could the authors be sure these were not artefacts induced by the experiment? If authors regress their gene expression against HSP gene signatures, would this cluster still be identified?

Thank you for this valuable suggestion. It is important to note that the Basal-SCCIS-tumor cluster was identified through CNV analysis, rather than the HSP signature. To address this concern and further validate this result, “AddModuleScore” function in Seurat package was used to regress gene expression against HSP gene signatures for retrieved basal cells. Our result showed that Basal_SCCIS tumor population still can be identified after regression, even more clearly (Author response image 1).

Author response image 1.

The identity of Basal-SCCIS-tumor cluster considering regression against HSP signatures.

9) Cell-cell communication analysis. The authors claimed that that cell-to-cell interaction was significantly enhanced in poorly-differentiated cSCC, and multiple interaction pathways were significantly active. How was this kind of analysis carried out? How did the authors define significance? what statistical method was used? these were all unclear. Furthermore, it is difficult to judge the robustness of the cell-cell communication analysis. Were these findings also supported by another method, such as celltalker, and cellphoneDB?

To determine the significance of the increased overall cell-to-cell interaction strength between two groups, we utilized CellChat to obtain the communication strength in different samples. We combined the communication strength based on cell type pairs, where missing values were set to 0. We performed a paired Wilcoxon test to determine whether the enhancement of cell-to-cell interaction between samples was significant.

For the comparison of outgoing or incoming interaction strength of the same cell types between two groups, we first extracted the communication strength of each signal pathway contributing to outgoing or incoming strength, and then merged the strengths of signal pathways among samples, where the strength of non-shared pathways with missing value was determined to be 0. Subsequently, we performed a paired Wilcoxon test to define the significance.

For multiple groups comparisons, the Kruskal-Wallis rank sum test was first performed. If the p-value is less than 0.1, the pairwise Wilcoxon test was used for subsequent pairwise comparisons. The comparison of individual signaling pathways between groups is similar to the above. We defined p-value < 0.1 as significance threshold. We have added the significance test method in figure legend for Figure 7 and Figure 8 as well as and detailed statistical data in new Supplementary file 1q-1u.

As suggested, we also used the approach of CellPhoneDB based on CellChatDB database to verify our cell-cell communication results. There are 55-58% of the ligand-receptor interactions predicted by CellChat were also predicted by CellPhoneDB (Author response image 2). The enhancement of cell interaction through MHC-II, Laminin and TNF signaling pathways in poorly-differentiated cSCC sample compare to normal sample were consistent in both CellChat and CellPhoneDB (Figure 8C and Figure 8-figure supplement 1B).

Author response image 2.

The overlap of the predicted ligand-receptor interactions between CellChat and CellPhoneDB.

10) Statistics and significance. In general, the detail of statistics and significance was lacking throughout the paper. Authors need to specify what statistical tests were used, and the p-values. It is difficult to judge the correctness of the test, and robustness without seeing the stats.

We have included all statistics and significance values in the figure legend and supplemental tables, and described the statistical tests in the methods section. In this revision, we have added the necessary details of statistics and significance in the main text and figures.

11) Overall, this manuscript needs a lot of re-writing. A lot of discussion was also included in the results, making it really difficult to read overall. The authors should simplify the results sections, remove the discussion bits, and further highlight and streamline with the key results of this paper.

Thanks a lot for this advice. We have revised the paper thoroughly, removed discussion in results section to make the manuscript easier to read.

Author Response

Reviewer #1 (Public Review):

Zhao et al. investigated the molecular nature of the binding site for carbohydrates within the UDP-sugars known to activate the P2Y14 receptor. In order to do so, they built a molecular model of the hP2Y14, docked the corresponding agonists, and performed MD simulation on the resulting complexes. The modeling was used to identify the key molecular interactions with a cluster of charged residues in the extracellular side of the TM region of the receptor, which they show are conserved within the P2Y receptors. The binding site of the UDP region was, not surprisingly, overlapping with the analogous ADP binding site experimentally observed for the P2Y12 receptor, and consequently, the region that recognizes the sugars could be anticipated. Nevertheless, the detailed modeling and simulation work shows the consistency of this hypothesis and provides a quantification of the particular interactions involved, pinpointing specifically the residues candidate to be involved in the recognition of sugars.

It follows the characterization, by functional assays, of the effect of single-point mutations of these residues in the efficacy of the different UDP-sugars. Here the results show a tendency to correlate with the molecular models, however some of the data has very low statistical significance and consequently the interpretation and conclusions extracted from this data should be taken with caution. This pertains to the particular role of the identified residues in the binding of the different sugars, which in some cases should be taken as a suggestion rather than a proof, though the general conclusion of the identification of the binding region for the sugar, its conservation among P2Y receptors and the role of some specific residues in sugar recognition seems convincing and the data are conveniently presented.

Finally, the design of ADP-sugars that activate the P2Y12 receptor, based on the transferability of the observations with the UDP-sugars for the P2Y14 receptor, is a first indication that such a recognition is possible and should happen in an analogous binding region. However, the low potencies exhibited by the ADP-sugars, in the micromolar range, are too far from the ADP agonist and the relevance of this mechanism remains to be proved. The difference between P2Y12 and P2Y14, with the last one showing much higher potencies for UDP-sugar derivatives than P2Y12 for the corresponding ADP-sugars, remains an interesting question not explored in this manuscript.

Thanks for your valuable comments. We have revised the interpretation of the data that has relatively low statistical significance in the manuscript. The conclusions extracted from this data have also been modified as suggestions. In this work, to investigate whether sugar nucleotides can also activate human P2Y12, we tested three ADP-sugars for human P2Y12. Discovery of highly potent P2Y12 agonists requires screening of a large number of compounds. It is possible there are the other ADP-sugars, which are highly potent P2Y12 agonists. It is technically challenging to synthesize ADP-sugars. Currently, we can only obtain ADP-Glc, ADP-GlcA and ADP-Man. Once the other ADP-sugars are available for us, we will test them and try to discover highly potent agonists in the future work. The highly potent agonists will be useful chemical tools to unveil the relevance mechanism of P2Y12. To explore the nature of binding site of the P2Y12 and P2Y14, we performed more experiments of mutagenesis study and added relevant data in the revised manuscript.

Reviewer #2 (Public Review):

The manuscript employs multiple approaches, including molecular docking, molecular dynamic simulations, and functional experiments to uncover a distinct uridine diphosphate-sugar-binding site on P2Y14 - a key drug target for inflammation and immune responses. Overall, the manuscript is clearly written, and the experimental techniques are well-documented. However, it may benefit from further analysis, particularly in terms of validating the binding pose.

Thanks for your comments. We used MMPBSA to analyze the ligand-binding energy for each receptor residue using MD trajectories. To further characterize the ligand-binding pose, we calculated the percentage of occurrence of hydrogen binding between the ligand and the carbohydrate-binding site (K277, E278, R253 and K77). We also calculated the ligand RMSF and ligand RMSD to show the stability of the ligand-binding pose and the simulation convergence. These data have been included in the revised manuscript.

Author Response

Reviewer #3 (Public Review):

Seeking a selective inhibitor that precisely inhibits on-target activities and avoids side effects is a major challenge in the field of drug discovery and therapeutics. The authors proposed an alternative method that combines multiple inhibitors to maximize on-target inhibition and minimize off-target inhibition. Focusing on the kinase-inhibitor interaction dataset, the authors developed a quantitative way to measure the selectivity for mixtures of inhibitors by using the Jenson-Sahannon distance metric. The method sounds technical.

From their computation and assays, the multi-compound-multitarget scoring (MMS) method framework was validated to be able to select a combination of inhibitors that is more selective than a single highly selective inhibitor for one kinase target, or for multiple targets. The MMS method is a promising solution to reduce off-target effects and could be applicable to other inhibitor-target interactions. My suggestion is that a comparative analysis of MMS with other similar methods can be conducted to highlight the advantage of MMS over others.

We thank the reviewer for this excellent summary and their suggestions. We agree that comparing new methods to prior ones is an important step in benchmarking new approaches and methods. However, to our knowledge, no other method exists for calculating selective combinations of kinase inhibitors. We compare our JSD selectivity scoring metric to other representative target-specific and non target-specific selectivity metrics (Figure 2 Figure Supplement 2).

The paper is not well organized and not easily readable. For example, first, the captions of the figures are two long; some of these texts could be moved to methods or results sections. Second, the concept of "penalty distribution" or "penalty prior" is vital to understand the MMS method, thus, at least a brief definition and introduction should be put in the main text rather than supporting method, as well as the rationale to use it. Third, the method section can be divided into several subsections with clear organizations and connections. Fourth, what is the difference between "a less selective inhibitor profile" and "an even less selective inhibitor profile" in Figure 3? Overall, the details of the paper are difficult to understand in the current version. I suggest rewriting the paper in a more concise and logical style.

We appreciate these suggestions and have significantly edited and revised our manuscript in order to facilitate clear communication. Specifically:

1) We have added an additional description of the penalty distribution to the description of the MMS method in the main Results section of the manuscript as opposed to solely in the Materials and Methods section.

2) We have provided a high-level concise summary of the MMS method in the results section in order to help orient a reader to the method. This description follows the same order (1 to 5) as the associated Figure 2, we hope this helps more clearly communicate the method.

3) We have moved descriptive figure captions to the methods section and, in general, substantially reduce the size of figure captions.

4) We have subdivided the Materials and Methods section as suggested.

5) We now describe in our main text how the simulated profiles were generated by smoothing the PKIS2645-like profile with two restraints; non-zero activity for LS inhibitors, and similar on-target probability for PKIS2-645-like, RS, and LS inhibitors to facilitate direct comparisons. We provide a new figure to quantify the selectivity of these simulated inhibitors and their similarity with true compounds (Figure 3 Figure Supplement 1).

6) We have removed content from the introduction and results sections that was less important to communicate to a general audience in order to make the manuscript more concise. We have also removed or condensed extraneous supplemental figures that were not required to communicate the central results and findings of experiments (ex: supplemental figures for Figure 3 and Figure 4 from the prior submission).

Author Response

Joint Public Review

(1) The developed model considers the interaction of multiple signaling networks that are essential for morphogenesis and homeostasis in the intestinal tissue, as well as other elements that had been proposed as relevant in the literature. Nevertheless, the details of how these interactions are modeled couldn't be evaluated in the current revision as the model was not shared with the reviewers and it is not available yet online, nor specified in any detail in the current manuscript. Additionally, how quantitative information from Wnt and BMP signaling pathways is incorporated in a quantitative way in the model is not clear.

Model files are provided with this reply. These are ‘.jl’ files for use with Julia. The model (the files provided with this reply) will be freely publicly available through BioModels upon acceptance of this manuscript for publication.

The model includes abstracted values to reproduce Wnt and BMP signalling gradients and their effect on cell proliferation and differentiation to generate the three-dimensional crypt spatial cell distribution. To further clarify the implementation of the quantitative information from Wnt and BMP signalling pathways in the model, we have added the following paragraph in the Appendix Section 8) Cell fate: proliferation, differentiation, arrest, apoptosis

"…During this migration the Wnt content in absorptive progenitors is halved in each division and, away from Wnt sources, progressively decreases, while BMP signals increase, towards the villus. In our model, differentiation into enterocytes occurs when progenitors encounter a BMP signal level, higher that their Wnt signal content. For instance, in the ileal crypt in homeostasis this occurs approximately at cell position 16 from the crypt base, where progenitors migrating from the stem cell niche reach a reduced content of Wnt signals of about 8 a.u. On the other hand, the BMP signalling level has a maximum value of 64 at approximately cell position 23 from the crypt base, where BMP signals are generated by mature enterocytes. These BMP signals diffuse towards the crypt base and, hence, decrease exponentially to reach values of 8 a.u. at approximately position 16, which, hence, enable differentiation into enterocytes. Epithelial injuries resulting in a decreased number of enterocytes reduce BMP signal production and its diffusion range which results in the enlargement of the proliferation compartment as cells encounter the required level of BMP signals for differentiation only at higher positions in the crypt."

(2) Some conclusions by the authors are not properly justified in the text, as "Paneth cells are the main driver behind the differential mechanical environment in the niche", "Wnt-mediated feedback loop prevents the uncontrolled expansion of the niche", the specific effect of p27 in contrast with Wee1 phosphorylation over the cell cycle length, and "their recovery [absorptive progenitors] started before the end of the treatment, driven by a negative feedback loop from mature enterocytes to their progenitors".

We have reworded these statements as described below.

The paragraph “Paneth cells are the main driver behind the differential mechanical environment in the niche, where cells with longer cycles accumulate more Wnt and Notch signals. In agreement with experimental reports {Pin, 2015 #719}, in our model Paneth cells are assumed to be stiffer and larger than other epithelial cells, requiring higher forces to be displaced and generating high intercellular pressure in the region” has been modified and now reads as follows “In agreement with experimental reports {Pin, 2015 #719}, Paneth cells are assumed to be stiffer and larger than other epithelial cells, requiring higher forces to be displaced and generating high intercellular pressure in the niche. Due to this increased mechanical pressure, cells in the niche have longer division cycles and can accumulate more Wnt and Notch signals.”

The sentence “Wnt-mediated feedback loop prevents the uncontrolled expansion of the niche” has been deleted from paragraph, that now reads “To generate a niche of stable size, we implemented a negative Wnt-mediated feedback loop that resembles the reported stem cell production of RNF43/ZNRF3 ligands to increase the turnover of Wnt receptors in nearby cells {Hao, 2012 #2086;Koo, 2012 #2089;Clevers, 2013 #538;Clevers, 2013 #2098}. Similarly, in our model, a number of stem cells in excess of the homeostatic value reduces cell tethering of Wnt ligands and hence inhibits Paneth and stem cell generation (Figures 1A-B).”

Regarding the specific effect of p27 in contrast with Wee1 phosphorylation over the cell cycle length. We have simplified the text in the main manuscript that now reads “Using the model of Csikasz-Nagy et al. {Csikasz-Nagy, 2006 #1870}, we modulated the duration of G1 through the production rate of the p27 protein. The p27 protein has been reported to regulate the duration of G1 by preventing the activation of Cyclin E-Cdk2 which induces DNA replication and the beginning of S-phase {Morgan, 2007 #2073}. We, hence, hypothesized that rapid cycling absorptive progenitors located in regions of low mechanical pressure outside the stem cell niche have low levels of p27, which bring forward the start of S-phase to shorten G1 (Figures 2D). In support of this hypothesis, it has been demonstrated that p27 inhibition has no effect on the proliferation of absorptive progenitors {Zheng, 2008 #2074} (see the Appendix for a full description).

In the Appendix Section 2 we provide an extended explanation of the use of the p27 and Wee1 kinetic governing parameters to decrease the length of the cell cycle by decreasing mainly G1 but maintaining the length of S phase constant, which is as follows

"Regarding G1 phase, the p27 protein has been reported to regulate the duration of G1 by preventing the activation of Cyclin E-Cdk2 which induces DNA replication and defines the beginning of S-phase {Morgan, 2007 #2073}. We hypothesized that fast cycling cells have low levels of p27 which result in earlier DNA replication, bringing forward the start of S-phase and shortening the length of G1. In support of this hypothesis, it has been experimentally demonstrated that inhibiting p27 has no effect on the proliferation of absorptive progenitors {Zheng, 2008 #2074}. In the Csikasz-Nagy model {Csikasz-Nagy, 2006 #1870}, the duration of G1 can be modulated through the parameter V_si, which is the basal production rate of p21/p27 (in the Csikasz-Nagy model, the p21 and p27 proteins are represented by a single variable, here we refer to that model quantity as p21/p27).

Additionally, the end of S-phase is associated with the decrease of Wee1 to basal levels due to Cdc14 mediated phosphorylation of Wee1. In the Csikasz-Nagy model {Csikasz-Nagy, 2006 #1870}, this reaction is described by a Goldbeter-Koshland function, which includes the parameter KA_Wee1p to regulate the level of Cdc14 required for the phosphorylation of Wee1.

Therefore, we modified these two parameters, V_si and KA_Wee1p, to ensure that variations of the cycle duration mostly impact on G1 while the length of S phase remains constant. We assumed that the value of the two parameters scales linearly with the duration of the division cycle, t_cycle, between a lower and upper bound, which prevent aberrant behaviour of the cell cycle model in the dynamically changing conditions of the crypt."

The paragraph related to “their recovery started before the end of the treatment…” sentence has been amended in the text and now reads “Simulated proliferative absorptive progenitors were indirectly affected by stem cell ablation and their decrease was followed by a reduction in mature enterocytes. The progenitors recovered soon after treatment interruption to later reach values above baseline when responding to the negative feedback signalling from mature enterocytes (Figure 3A).”

(3) Only the results of the "main" model are shown, with no information about its sensitivity to parameter values, and how their conclusions depend on specific decisions on the model. For example, the authors said that "an optimal crypt cell composition is achieved when BMP and Wnt differentiation thresholds result in progenitors dividing approximately four times before differentiating into enterocytes", but the results of alternative scenarios are not shown.

To address this comment, we have included a new section in the Appendix, called “What-if Analysis”, and new figures (Figure S4-S8) with simulations of alternative scenarios affecting the main signalling pathways that govern crypt composition, in particular, we simulated stronger and weaker Wnt, BMP, Notch and ZNRF3/RNF43 signalling.

We attach the new section here:

"10) What-if Analysis

We investigated the effect on the simulated crypt of increasing and decreasing the strength of the main signalling pathways, Wnt, BMP and ZNRF3/RNF43 signalling, and modifying the Notch thresholds. For each alternative parameterisation, except when decreasing ZNRF3/RNF43 signalling, the simulation was run for 30 days to ensure stability was reached with the new parameter set and the final 10 days were included in the analysis. When decreasing ZNRF3/RNF43 signalling, we simulated 60 days to demonstrate the expansion of the niche and analysed the final 10 days. The reference parameter set used as baseline was the ileal mouse crypt parameter set reported in Appendix Table 1. In all cases, we only consider modifications of one signalling mechanism at a time.

To study alternative Wnt signalling scenarios, we used the WntRange parameter (Appendix Table 1), to double and halve the spreading area of Wnt signals emitted by Paneth cells while we maintained the original WntRange value for Wnt-emitting mesenchymal cells at the bottom of the crypt (Appendix Section 7.1) (Figures S4A-S4F). When WntRange was doubled, we observed increased number of stem and Paneth cells in a noticeably enlarged niche (Figures S4C-S4D), with cells choosing the stem cell fate instead of differentiating into absorptive progenitors. On the other hand, decreasing Wnt signalling, by halving WntRange in Paneth cells but maintaining its homeostatic value in mesenchymal cells, resulted in no apparent changes in the niche cell composition (Figures S4E-S4F) which resembled published experimental results of persisting functional stem cells after Paneth cell ablation {Durand, 2012 #434}.

The ZNRF3/RNF43-mediated negative feedback mechanism regulates the size of the niche by modulating Wnt signalling. We simulated increasing and decreasing the strength of the ZNRF3/RNF43, by doubling and halving, respectively, the parameter Z described in the Appendix Section 7.2 (Figures S5A-S5F). Following the increase of the intensity of ZNRF3/RNF43 signalling, we observed a decrease in the number of stem and Paneth cells together with relatively minor changes in the transit-amplifying region (Figures S5C-S5D). On the other hand, when decreasing ZNRF3/RNF43 signalling levels, the niche expanded , resulting in a crypt dominated by Paneth and stem cells (Figures S5E-S5F ) which replicates reported experimental phenotypes {Koo, 2012 #2089}.

To modify Notch signalling, we increased and decreased by 1 A.U. the Notch threshold required for lateral inhibition (Figures S6A-S6F). This Notch signalling threshold determines the number of contacting Notch-secreting cells (secretory lineage) to inhibit the differentiation of stem cells into the secretory lineage. Thus, increasing this Notch threshold enhances the production of secretory cells leading to the increase of Paneth, goblet and enteroendocrine cells (Figure S6C-S6D). Alternatively, decreasing the Notch threshold enhances differentiation into the absorptive lineage, reducing the number of Paneth and secretory cells (Figure S6E-S6F).

We modified the range of diffusion of BMP signals by doubling and halving the parameter A , (Figures S7A-S7F) which denotes the amount of diffusing BMP signals towards the base of the crypt (Appendix Section 7.4). When we increased the BMP signalling range, enterocytes differentiated at lower crypt positions effectively reducing the transit-amplifying zone (Figure S7A, Figure S7B). Decreasing BMP signalling strength by halving A resulted in the increase of proliferative absorptive progenitors, which reach higher positions in the crypt (Figure S7C-S7D). The niche was largely unaffected in both cases (Figure S7E-S7F)."

(4) Regarding the construction of the model, the authors used "counts of Ki-67 positive cells recorded by position" while the original data reported "overall cell counts per crypt and villus". Some explanation about how this conversion was made, why it is valid, as well as any potential problems, is needed. Additionally, the model is based on experiments done by others in mouse models; the similarity to the response in human intestinal crypts is not discussed.

Ki-67 immunostaining data during 5-FU treatment was derived from the same experiments. The overall cell counts per crypt and villus are published in {Jardi, 2022 #2416}. For this manuscript, we reanalysed the intestinal samples to estimate counts of cell types by position in the crypt.

We have clarified the text, which now reads …“The samples from this later study {Jardi, 2022 #2416} were analysed again to count Ki-67 positive cells at each position along the longitudinal crypt axis, for 30-50 individual hemi crypt units per tissue section per mouse as previously described {Williams, 2016 #2165}.”

We agree that the understanding of the translation of results derived from animal models into a human or clinical context is of high relevance. The mouse crypt is a model of choice to study epithelial biology and exhibits remarkable similarities with the human crypt. In our team, we are focussed on developing translational modelling strategies and have a version of the model that describes a human crypt. That model assumes mostly conserved crypt biology and structure across species and includes changes in parameter values needed to compensate reported differences in morphometrics and cell cycle duration. Due to the relevance and extent of this translational work, we chose to focus on the mouse crypt entirely in this manuscript. We think the translational modelling strategy to explore the quantitative translation between human and mouse and/or other species/settings merits a full report.

(5) The authors imply that their mathematical model of the intestinal crypt is an improvement over those already published but there is no direct comparison or review of the literature to substantiate this claim.

An extended literature review including more details of previous ABMs to enable a direct comparison with our model is now included in the manuscript and reads as follows:

“Several agent-based models (ABMs) have been proposed to describe the complexity and dynamic nature of the intestinal crypt. Early models were used as in silico platforms to study the dynamics and cellular organisation of the crypt. For instance, one of the pioneering ABMs was used to study the distribution and organisation of labelling and mitotic indices {Meineke, 2001 #326}. This model comprises a fixed ring of Paneth cells beneath a row of stem cells, which divide asymmetrically to produce a stem cell and a transit-amplifying cell that terminally differentiates after a fixed number of divisions. Some subsequent models are lattice-free, recapitulate neutral drift of equipotent stem cells and describe proliferation and cell fate regulated by a fixed Wnt signalling spatial gradient, which is defined by the distance from the crypt base, with proliferating cells progressing through discrete phases of the cell cycle and showing variable duration of the G1 phase {Pitt-Francis, 2009 #129}. Further model refinements can be seen in the model of Buske et al (2011), with stochastic cell growth and division time {Buske, 2011 #1}, Wnt levels defined by the fixed local curvature of the crypt and lateral inhibition driven by Notch signalling. Here, we present a lattice-free agent-based model that describes the spatiotemporal dynamics of single cells in the small intestinal crypt driven by the interaction of surface tethered Wnt signals, cell-cell Notch signalling, BMP diffusive signals, RNF43/ZNRF3-mediated feedback mechanisms and the cycle protein network responding to the crypt mechanical environment. We show that our computational model enables the simulation of the ablation and recovery of the stem cell niche as well as of how drug-induced molecular perturbations trigger a cascade of disruptive events spanning from the cell cycle to single cell arrest and/or apoptosis, altered cell migration and turnover and ultimately loss of epithelial integrity.”

(6) The authors claim that the simulated data and the available mouse data match up. Nevertheless, the data vs the model still appear both quantitatively and qualitatively different (as presented in Figures 2E, F, and 5C, D). This puts in doubt how much the model can actually reproduce the experimental data. In conclusion, the model would benefit from further refinement, particularly if the goal is to use the model for predicting the dynamics of oncogenic drug candidates.

To address this comment, we have made several adjustments: we refined the counting algorithm that determines cell position and improved the Ki67 and BrdU staining simulations by modifying the simulated staining criteria and adding an estimation of the experimental error to the simulated responses. A description of these changes is described in a new section in the appendix called “ABM simulation of Ki-67 and BrdU Staining”

With these changes we think we have achieved a more satisfactory agreement between observed and predicted results and updated all figures with Ki67 and BrdU staining simulated results.

Author Response

We are grateful to the editors and the reviewers for the thorough evaluation of our manuscript and their feedback, as it allows us to provide additional clarification of our findings and improve the manuscript.

In their evaluation reviewers raised a key conceptual point linked to the inhibitory mechanism that appeared to be insufficiently explained in the manuscript, leading to a misconception regarding the physiological relevance. They have also missed experimental data related to the concentrations of Aβ used and their relevance for Alzheimer’s disease (AD). We believe that our studies, although performed in vitro in model systems, provide novel conceptual framework and shed light on the unexplored mechanisms underlying AD.

We discuss these points below in a provisional response to their comments.

Reviewer #1 (Public Review):

Summary:

Human Abeta42 inhibits gamma-secretase activity in biochemical assays.

Strengths:

Determination of inhibitory concentration human Abeta42 on gamma-secretase activity in biochemical assays.

Weaknesses:

Human Abeta42 may concentrate up to microM order in endosomes.

This is correct.

If so, production of Abeta42 would be attenuated then lead to less Abeta deposition in the brain. The authors finding is interesting but does not fit the physiological condition in the brain.

We thank the reviewer for raising this key conceptual point, as this gives us the opportunity to clarify it for the future readers.

The characterized inhibitory mechanism is more complex than the reviewer’s interpretation, and a number of factors must be considered. Indeed, our data show that Aβ42 upon intracellular concentration inhibits γ-secretase activity, resulting in increased γ-secretase substrate (C-terminal fragment, CTF) levels. It is important however to highlight that this inhibition is competitive in nature, implying that it is partial, reversible, and regulated by the relative concentrations of the Aβ42 peptide (inhibitor) and the substrates. The model that we put forward is that cellular uptake and intracellular concentration of Aβ42 facilitates γ-secretase inhibition, which results in the accumulation of APP-CTFs (and γ-secretase substrates in general). However, as Aβ42 levels fall, the increased concentration of substrates shifts the equilibrium towards their processing and Aβ production. As Aβ42 concentration raises again, equilibrium is shifted back towards inhibition and so on. This inhibitory mechanism will translate into pulses of (partial) γ-secretase inhibition, which will alter γ-secretase mediated signalling (arising from increased CTF levels or decreased release of soluble intracellular domains from substrates). These alterations may affect the dynamics of systems oscillating in the brain, such as NOTCH signalling, implicated in memory formation (2), and potentially others (related to e.g. cadherins, p75 or neuregulins).

It is worth noting that oscillations in γ-secretase activity induced by treatment with a γ-secretase inhibitor (semagacestat) have been proposed to have contributed to the cognitive alterations observed in semagacestat treated patients in the failed Phase-3 IDENTITY clinical trial (2, 3); and that semagacestat, like Aβ42, acts as a high affinity competitor of substrates (Koch et al, 2023). We will include this clarification in the discussion of the revised manuscript and create an additional figure presenting the proposed mechanism.

It is not clear whether the FRET-based assay in living cells really reflect gamma-secretase activity.

The specificity of this assay is supported by the γ-secretase inhibitor treatment included as a positive control (Figure 3). In addition, the following literature supports that this assay truthfully assesses γ-secretase activity in cellular context (4-7).

Processing of APP-CTF in living cells is not only the cleavage by gamma-secretase.

This is correct, and therefore we have analysed the contribution of other APP-CTF degradation pathways by performing cycloheximide-based stability assay in the presence of γ-secretase inhibitor. Quantitative analysis of the levels of both APP-CTFs and APP-FL over the 5h time-course failed to reveal significant differences between Aβ42 treated cells and controls. As expected, Bafilomycin A1 treatment markedly prolonged the half-life of both proteins (Figure 7B & C). The lack of a significant impact of Aβ42 on the half-life of APP-CTFs under the conditions of γ-secretase inhibition is consistent with the proposed inhibitory mechanism. Finally, we note that the inhibition will not only affect APP-CTF, but also the processing of γ-secretase substrates in general.

Reviewer #2 (Public Review):

Summary:

In the current study, the authors tested the hypothesis that Aβ42 toxicity arises from its proven affinity for γ-secretases. Specifically, the increases in Aβ42, particularly in the endolysosomal compartment, promote the establishment of a product feedback inhibitory mechanism on γ-secretases, and thereby impair downstream signaling events. They showed that human Aβ42 peptides, but neither murine Aβ42 nor human Aβ17-42 (p3), inhibit γ-secretases and trigger accumulation of unprocessed substrates in neurons, including (CTFs of APP, p75 and pan-cadherin. Moreover, Aβ42 dysregulated cellular homeostasis by inducing p75-dependent neuronal death. Because γ-secretases process many other membrane proteins, including NOTCH, ERB-B2 receptor tyrosine kinase 4 (ERBB4), N-cadherin (NCAD) and p75 neurotrophin receptor (p75-NTR), revealing a broad range of downstream signaling pathways, including those critical for neuronal structure and function. Hence, they propose to identification of a selective role for the Aβ42 peptide, and raise the intriguing possibility that compromised γ-secretase activity against the CTFs of APP and/or other neuronal substrates contributes to the pathogenesis of AD. Overall, the data are not very convincing to support the main claim.

Strengths.

Different in vitro and cellular approaches are employed to test the hypothesis.

Weaknesses.

The experimental concentrations for Aβ42 peptide in the assay are too high, which are far beyond the physiological concentrations or pathological levels. The artificial observations are not supported by any in vivo experimental evidence.

It is correct that in the majority of the experiments we used low μM concentrations of Aβ42. However, we would like to note that we also performed experiments where conditioned medium collected from human APP.Swe expressing neurons was used as a source of Aβ. In these experiments total Aβ concentration was in low nM range (0.5-1 nM) (Figure 4G). Treatment with this conditioned medium led to the increase APP-CTF levels, supporting that low nM concentrations of Aβ are sufficient for partial inhibition of γ-secretase.

We would like to underline that Aβ is estimated to be present in the brain in concentration ranging from fM to mM, depending on the pool (soluble, aggregated, fibrillar, etc) that is considered (8, 9). However, it is rather the local than the global concentration of Aβ that is critical for the disease pathogenesis. In this regard, it is proposed that as AD progresses Aβ42 slowly accumulates in the endo-lysosomal system wherein it reaches μM concentrations that are required for aggregation and seeding (1, 10, 11). Our findings are consistent with the analysis showing that extracellular soluble Aβ42 peptide, at low nM concentrations, is taken up by cortical neurons and neuroblastoma (SH-SY5Y) cells, and concentrated in the endo-lysosomal system wherein effective peptide concentrations reach ~2.5 μM (1). Hence, a slow vesicular peptide accumulation and/or degradation imbalance (1, 11, 12) could lead to several order of magnitude increases in the effective concentration of Aβ42 over the span of years to decades in AD pathogenesis. We note that our experimental settings, using low μM concentrations of extracellular Aβ42 over 24h treatment, were designed to accelerate this 'peptide concentration’ process in vitro. As discussed in our report, a high μM Aβ peptide concentration in the endo-lysosomal system not only leads to aggregation but also facilitates γ-secretase inhibition. Of note, we are currently developing protocols and will undertake follow up studies to quantitatively define the Aβ concentration in synaptosomes and endosomes in AD brain, as well as in in vitro systems (i.e. cells treated with Aβ preparations obtained from AD brains).

Finally, we would like to highlight that analyses of the brains of the AD affected individuals have shown that APP-CTFs accumulate in both sporadic and genetic forms of the disease (13-15); and recently, Ferrer-Raventós et al have revealed a correlation between APP-CTFs and Aβ levels at the synapse (13).

To conclude, we would like to highlight that as clarified above, the Aβ peptide concentrations and the conditions tested fit well within pathophysiology, and that the data presented in our report collectively provide evidence in support of an Aβ42-mediated inhibitory effect on γ-secretase.

References:

1. X. Hu et al., Amyloid seeds formed by cellular uptake, concentration, and aggregation of the amyloid-beta peptide. Proc Natl Acad Sci U S A 106, 20324-20329 (2009).
2. B. De Strooper, Lessons from a failed γ-secretase Alzheimer trial. Cell 159, 721-726 (2014).
3. R. S. Doody et al., A phase 3 trial of semagacestat for treatment of Alzheimer's disease. N Engl J Med 369, 341-350 (2013).
4. M. C. Houser et al., A Novel NIR-FRET Biosensor for Reporting PS/γ-Secretase Activity in Live Cells. Sensors (Basel) 20, (2020).
5. M. C. Q. Houser et al., Limited Substrate Specificity of PS/γ-Secretase Is Supported by Novel Multiplexed FRET Analysis in Live Cells. Biosensors (Basel) 11, (2021).
6. M. Maesako et al., Visualization of PS/γ-Secretase Activity in Living Cells. iScience 23, 101139 (2020).
7. M. Maesako, M. C. Q. Houser, Y. Turchyna, M. S. Wolfe, O. Berezovska, Presenilin/γ-Secretase Activity Is Located in Acidic Compartments of Live Neurons. J Neurosci 42, 145-154 (2022).
8. B. R. Roberts et al., Biochemically-defined pools of amyloid-β in sporadic Alzheimer's disease: correlation with amyloid PET. Brain 140, 1486-1498 (2017).
9. J. A. Raskatov, What Is the "Relevant" Amyloid β42 Concentration? Chembiochem 20, 1725-1726 (2019).
10. M. P. Schützmann et al., Endo-lysosomal Aβ concentration and pH trigger formation of Aβ oligomers that potently induce Tau missorting. Nat Commun 12, 4634 (2021).
11. E. Wesén, G. D. M. Jeffries, M. Matson Dzebo, E. K. Esbjörner, Endocytic uptake of monomeric amyloid-β peptides is clathrin- and dynamin-independent and results in selective accumulation of Aβ(1-42) compared to Aβ(1-40). Sci Rep 7, 2021 (2017).
12. M. F. Knauer, B. Soreghan, D. Burdick, J. Kosmoski, C. G. Glabe, Intracellular accumulation and resistance to degradation of the Alzheimer amyloid A4/beta protein. Proc Natl Acad Sci U S A 89, 7437-7441 (1992).
13. P. Ferrer-Raventós et al., Amyloid precursor protein Neuropathol Appl Neurobiol 49, e12879 (2023).
14. M. Pera et al., Distinct patterns of APP processing in the CNS in autosomal-dominant and sporadic Alzheimer disease. Acta Neuropathol 125, 201-213 (2013).
15. L. Vaillant-Beuchot et al., Accumulation of amyloid precursor protein C-terminal fragments triggers mitochondrial structure, function, and mitophagy defects in Alzheimer's disease models and human brains. Acta Neuropathol 141, 39-65 (2021).

Author Response

The following is the authors’ response to the original reviews.

Reviewer #1

As written in my public review I consider the science of this work to be high quality. I have some suggestions for the write-up though. As a general comment, I think that too much has been put into the appendices. In particular, the main text could contain more details about the model.

We are pleased that this Reviewer feels that our work to be of “high quality”. We value the reviewer’s insightful suggestions and comments. Following this Reviewer’s suggestion we have moved certain sections to the main text.

In what follows, we provide responses to each of the reviewer’s inquiry, and indicate the appropriate changes in the revised version.

P2 -

ϕ is introduce as packing fraction - on p3 it’s called cell density. Also it is not clear whether it is an area fraction or a cell number density. Please define properly and I would suggest sticking to one notion.

ϕ is the cell packing fraction. In two dimensions (as is the case in our simulations) it is the area fraction. However, in order to stick to one general notation (independent of dimension) we use “packing fraction” to represent how densely the cells are packed. We changed it the revised manuscript to ensure uniformity.

P3 -

“which should and should slow down the overall dynamics” Typo?

Corrected it in the revised manuscript.

“One would intuitively expect that the ϕfree should decrease with increasing cell density” Please, define ϕfree

ϕfree is defined in Eqn. 4. We ought to have defined it in the introduction.

“When ϕ exceeds ϕS, the free area ϕfree saturates because the soft cells interpenetrate each other,” I suggest clearly distinguishing between biological cells and the agents (disks) used in the simulation. Please, also clarify What interpenetration of agents corresponds to in tissues?

We have rewritten the sentence as, ”The simulations show that when..” Soft disks used in the simulations seem to be not an unrealistic model for biological cells. The small deformations noted in our model is not that different from the cells in the tissues. For visual reference, please see Author response image 1. In the left panel of the figure, a 2D snapshot of the experimental zebrafish tissue, displays the deformation of cells labeled as 1 and 2. Likewise, the right panel illustrates the extent to which such deformations are replicated in the simulation by allowing two cells to overlap (the white area in the right panel of Author response image 1 represents the interpenetration). In the revised manuscript, we have made the necessary change from “soft cells” to “soft disks.”

Author response image 1.

Snapshots of zebrafish tissue (left panel) (Ref. [14] main text) and model two dimensional tissue (right). In the right panel the white area represents the overlap and the black vertical line represents the intersection.

“The facilitation mechanism, invoked in glassy systems [22] allows large cells to move with low mobility.” What is the facilitation mechanism?

Facilitation, which is an intuitive idea, that refers to a mechanism by which cells in a in highly jammed environment can only move if the neighboring cells get out of the way. In our case (as shown in the text (Fig.3 (A) and Fig. 13 (A) & (B)) the smaller cells move faster almost independent of ϕ. When a small cell moves, it creates a void which could facilitate neighboring cells (including big ones) to move.

“η (or relaxation time)” I suggest explaining the link between η and the relaxation time.

First, in making this point on aging we only showed that the relaxation time is independent of the waiting time. In the revised manuscript we deleted η.

Although not germane to this study, in the literature on glass transition, it is not uncommon to use relaxation time τα (as a proxy of viscosity η) to describe the dynamics. The relation between τα and η is given by

where G∞ is the “infinite frequency” shear modulus, which holds in unjammed or in liquids. This relation suggests that τα is proportional to η, which is almost never satisfied in glass forming systems.

P5 - “In addition, the elastic forces characterizing cell-cell interactions are soft, which implies that the cells can penetrate with rij − (Ri + Rj) < 0 when they are jammed.” Is this about the model or the biological tissue? Presumably the former, because real cells do not penetrate each other, right? What are rij, Ri and Rj?

This is about the model. The cells are sufficiently soft that they can be deformed, which allows for modest interpenetration. Real cells exhibit similar behavior (see Fig. 1). In inset of Fig. 4 (b) rij is the center to center distance between cells with radii Ri and Rj. It is better to use the word overlap instead of penetrate, which is what we have done in the revised version.

“we simulated a highly polydisperse system (PDs) in which the cell sizes vary by a factor of ∼ 8” Is it important to have a factor 8 - the zebra fish tissue presents a factor 5 − 6?

This is an important question, which is difficult to answer using analytic theory. It does require simulations unfortunately. We do not know a priori the polysipersity value needed to observe saturation in η at high value of ϕ. However, we have shown that the a system with one type of cell (monodisperse) crystallizes. Furthermore, mixtures of two cell types do not show any saturation in η over the parameter range that we explored. A systematic simulation study is needed to explore a range of parameter values to determine the minimum PD, which would match the experimental findings.

We performed 3D simulations to figure out if much less PD would yield saturation in η. Preliminary simulations in three dimensions with a lower value of PD (11.5% with a size variations by a factor of ≈ 2 ) exhibits saturation in the relaxation time. For comparison, the value of PD in the current work is ≈ 24% with a size variation by a factor of 8.

P6 -

“which is related to the Doolittle equation [26] for fluidity ( )” what is the Doolittle equation? Is it important here? Also: “VFT equation for cells”? Is it the same as given on p.2 - so nothing special for cells - or a different one?

Historically, the Doolittle equation was proposed to describe the change in η in terms of free volume in the context polymer systems over 60 years ago. The physics in the polymers is very different from the soft models for cells considered here. Nevertheless, the equations has meaning in the context as well. The Doolittle (other names associated with similar equations are Ferry, Flory... ) equation is given by

, where A and B are constants, V is the total volume and Vhc is the hardcore volume. Essentially, is the relative free volume. It can be shown that one can arrive at the VFT equation starting from the Doolittle equation.

The VFT equation for cells is same as given in page 2, which we restate for completeness. Here, we introduce the apparent activation energy.

“The stress-stress tensor” Why not simply stress tensor?

We have corrected it.

“shows qualitatively the same behavior as the estimate of viscosity (using dimensional arguments) made in experiments.” Where is this shown?

The dependence of viscosity as a function ϕ is shown in Figure 1 (c).

P7 -

Fig 2A caption “dashed line” Maybe full line?

This should be full line. It is fixed in in the revised manuscript.

P8 -

“a puzzling finding that is also reflected” Why is it puzzling?

In figure 2 (C), it shows that the increase in the duration in the plateau of Fs(q,t) ceases when ϕ exceeds ≈ 0.90. This to us is puzzling (always a matter of perspective) because we expected that the duration of Fs(q,t) plateau to increase as a function of ϕ based on the VFT behavior for ϕ ≤ ϕS. As a result, we imagined that the relaxation time τα would continue to increase beyond ϕS. However, the simulations show that the relaxation time is essentially a constant for ϕ > 0.90, which implies that the soft disk system (our model for the tissue) is an unusual with behavior that has no counter part in the material world.

“If the VFT relation continues” –“If the VFT relation continued”

We have fixed it.

First paragraph does not seem to be coherent

What is RS (or Rs)?

RS is the radius of the small cell. In the revised manuscript we have made this clear.

P10 -

Please, define the waiting time.

The waiting time refers to the period between sample preparation and data collection either in experiments or in simulations. In an ergodic system, the properties should not depend on the waiting time provided provided it is large. In other words, after the system reaches thermal equilibrium, the waiting time tω should not have an impact on the properties of the system.

“fully jammed” Please, define.

The term “fully jammed” refers to a state in which the constituent particles in a system do not move. For example, it a hard sphere system at a packing fraction of approximately 0.84 is fully jammed, which implies there is wiggle room for a particle move without violating the excluded volume restriction. At this specific packing fraction, the hard sphere system undergoes a jamming transition, resulting in the particles becoming completely immobile. The nonconfluent tissue modeled here is not fully jammed.

P11 -

Fig.4 it is hard to see that the width of P(hij) increases with ϕ.

Please see Author response image 2 with a less number of curves for a better visualization. We have replaced this figure in the revised version.

Author response image 2.

Probability of overlap (hij) between two cells, P(hij), for various ϕ values.

“Thus, even if the cells are highly jammed at ϕ ≈ ϕS, free area is available because of an increase in the overlap between cells.” This conclusion seems premature at this point.

The Referee is correct. This is shown in Fig. 5. We amended the ends of the sentence to reflect this observation.

P12 -

“as is the case when the extent of compression increases” extent of compression = density?

This is correct. Extent of compression corresponds to the packing fraction or the density.

“This effect is expected to occur with high probability at ϕS and beyond,” Why? What is special about ϕS.

To achieve high packing fractions beyond a certain value of ϕ soft cells have, which would occur at a certain value ϕS. In the system studied here, ϕ ≈ 0.90 = ϕS. Note that ϕS could be altered by changing the system parameters.

P15 -

“local equilibrium” In a thermodynamic sense? There is also cell migration, so thermodynamic equilibrium does not seem to be appropriate.

This is an important point. The observation that equilibrium concepts hold in what is manifestly a non-equilibrium system is a surprise. It is referred in a thermodynamic sense. We agree with the reviewer because of cell division (in Ref. [14] main text), cell death, thermodynamic equilibrium does not seems to be appropriate. This is exactly the point we raise in the introduction. However, considering the timescale of cell division and death it appears that there may be a local steady state, which we we call a “local equilibrium”. As a consequence phase transition ideas and Green-Kubo relations are applicable. Indeed, a surprise in the conclusion in Ref. [14] is that in the zebrafish morphogenesis equilibrium description seems adequate.

“number of near neighbor cells that is in contact with the ith cell. The jth cell is the nearest neighbor of the ith cell, if hij > 0” A neighbour cell or the nearest neihbor?

A neighbour cell is accurate.

P16 -

“In our model there is no dynamics with only systematic forces because the temperature is zero.” What is a systematic force? I do not understand the sentence.

Systematic force between two cells is defined in Eqn. 5 in the main text. Because temperature is not a relevant variable in our model, we want to emphasize that in the absence of self propulsion, the cells would not move at all.

Reviewer #2

Major comments:

A/ Role of size polydispersity

In the text, and also in the methods (Appendix A), the authors mention that they need large polydispersity of particle sizes to explain the viscous plateau, as the dynamics of small vs large cells are ”dramatically different” (Appendix G). They simulate a system where cell sizes vary by a factor 8, mentioning this is typical in tissues, but I found this quite surprising - this would be heterogeneities in cell volume of 500, many orders of magnitude above what has been measured in tissues. As far as I’m aware, divisions are quite symmetric and synchronous in early vertebrate embryogenesis, so volume variations are expected to be very small (similarly in epithelial tissues, where jamming has been looked at extensively, I’m not aware of examples with ratio of 8 between cell diameters). One question I had is that when the authors look at ”small polydispersity”, there are 50 − 50 mixtures. Would small polydispersity with continuous distributions change this picture? Could they take their current simulations but smoothly change the ratio of polydispersity from 8 to 0 to see exactly how much they need to explain viscosity plateauing, and at which point is the transition?

We thank the reviewer for raising this important question, which was also a concern for Reviewer #1. The value of polydispersity (PD) required to observe such behavior is not known a priori even within the simple model used. We selected a PD value, with a size variation of a factor of 8, guided in part by the experiment (projection onto 2D) shown in Figure 1(B) and Figure 6(D). We also showed that the monodisperse system crystallizes, and the binary system do not show signs of saturation within the explored range of parameter space and ϕ. This suggests that a certain degree of size dispersity is necessary to obtain saturation in η.

As discussed in Appendix B, the binary system is characterized by the variables , where RB and RS represent the radii of the big and small cells, respectively, and the packing fraction ϕ. By more fully exploring the parameter space encompassing λ and ϕ than we did, it maybe possible, as the Referee suggests, that a system with two different cell sizes would yield the experimentally observed dependence of η on ϕ.

As part of an answer to the Reviewer #1 on a the same issue, we mentioned results of preliminary simulations in three dimensions with reduced levels of polydispersity, and discovered that at lower levels of polydispersity (variation in size by a factor of ≈ 2 and polydispersity value 11.50%), the relaxation time does saturate beyond a certain packing fraction (see Fig. 3). We have not established if η, the key quantity of interest, would exhibit a similar behavior in 3D.

Author response image 3.

(A) τα as a function of ϕ for 11% polydispersity with size variation by a factor of ∼ 2 in the three dimensional system. (B) Same as (A) except polydispersity value is 24% and a size variation by a factor of ∼ 8.

B/ Role of fluctuations/self-propulsion in this system, and relationship to recent findings

“A priori it is unclear why equilibrium concepts should hold in zebrafish morphogenesis, which one would expect is controlled by non-equilibrium processes such as self-propulsion, growth and cell division. ”

This is raised as a key paradox, but is not very clear to me in the context raised by the authors. In particular, they use self-propulsion as a source of activity and explain the evolution of viscosity but a facilitation process involving re-arrangements/motility. But I don’t think self-propulsion has been argued to play a role in zebrafish blastoderm - Ref 14 argues that this is effectively a zerotemperature phenomenon and that cell motility/rearrangements do not show any correlation with viscosity. So this part of the model assumption was not clear to me in relationship with the proposed experimental system. Active noise has been proposed to play key roles in other systems, including motility-driven and tension fluctuation-driven unjamming (among many others Bi et al, PRX, 2016, Mitchel et al, Nat Comm, 2020, Pinheiro et al, Nat Phys, 2022 as well as Kim & Campas, Nat Physics, 2021) - maybe this is somewhere where the author model could fit? In Kim & Campas, Nat Phys, 2021 in particular, the authors develop simulations of non-confluent tissues with noise, that seems to bear some resemblance to the model developed here, so it would be important to discuss the similarities and distinctions (usually I think polydispersity is not considered indeed). In general, the authors look here at a particle based model, but cells have adhesions with well-defined contact angles, so there is a question of the cross-over between their findings and the large body of recent literature on active foams/vertex models (which are not really discussed there).

We appreciate the lengthy comment here, and there is a lot to unpack. We also thank the referee for the references, some of which we did not know about earlier.

The primary objective of our study is to determine the simplest minimal model that would explain the experimentally observed dependence of viscosity in zebrafish blastoderm tissue as ϕ is increased beyond a certain packing fraction during morphogenesis. In Reference 14, the authors analyzed the data using the framework of rigidity percolation theory and presented evidence of a genuine equilibrium phase transition. Consequently, one would that expect zebrafish blastoderm tissue to be in equilibrium, which is surprising from many perspectives. However, since the tissue is a growing system involving numerous cell divisions and cell death, it is not immediately evident whether the assumption of equilibrium is valid. Indeed, the same problem arises when considering the glass transition where rapid cooling drives the system out of equilibrium. Nevertheless, heat capacity and η are often analyzed using the notion of equilibrium. Hence, considering this issue within the context of our research appears to be reasonable.

To the best of our knowledge, the authors in Ref. 14 did not provide an explanation for the η behavior. The focus was, which was excellent and is the basis on which we initiated this study, was on the use of rigidity percolation theory to explain the results. Indeed, they performed an experiment by mildly reducing myosin II activity, which apparently affects cell motility. The quantitative effect was not reported.

We did not impose any requirement of cell rearrangements etc in the model. There is essentially one variable, free area available, that explains the η dependence on ϕ. It is possible that one can come up with other zero temperature models that could also explain the data. To the best of our knowledge, it has not been proposed.

It would be interesting to set our model in the context of other models that the referee points out. This would be an interesting research topic to explore. The only comment we would like to make is that it is unclear how vertex model for confluent tissues could explain the viscosity data.

C/ Calculation of the effective shear viscosity

The authors calculate viscosity from a Green-Kubo relation, although it would be good to clarify at which time scale (and maybe even shear amplitude) they expect this to be valid. These kinds of model would be expected to show plastic rearrangements for large deformations for instance, could the authors simulate realistic rheological deformations (e.g. Kim & Campas, 2021 applying external shear on the simulations) to see how much this matches both their expectation and the data?

Once it is established that there is local equilibrium (as implied by the use of phase transition ideas to analyse the experimental data in Ref. 14), it is natural to use the Green-Kubo relation to calculate transport properties. Hence, for our purposes, it is valid for all time scales and amplitude. The Reviewer also wonders if the model could be used to simulate response to shear in order to probe rheological properties. There is no conceptual issue here and indeed this is an excellent suggestion that we intend to pursue in the future.

D/ Role of cell adhesion

The authors consider soft elastic disks of different sizes but unless I missed it, there is no adhesion being considered. This is expected to play a key role in jamming and multicellular mechanics, so I think the authors should either look at what this changes in their simulations, or at least discuss why they are neglecting it. One reason I’m asking is that it’s not totally clear to me that the ”free space” picture, coming from the fact that cells can interpenetrate in their model would hold in a model of deformable cells adhering to each other with constant volume (leading to more equilibration of deformations it would seem?).

The referee raises another question regarding the lack of adhesion in the simulations. As pointed out before, we were trying to create a minimal model to account for the experimental observations for η upon changing the packing fraction. Thus, we a coarse-grained model where we considered poly-disperse cells with elastic interactions which recapitulates the experimental observations. The referee is correct that adhesion plays a role in jammed systems, and examination of how it would affect is an aspect that would be interesting to consider in the future. We hasten to add that even systems without attractive adhesion-type interaction become jammed. In principle, in many-body systems, the parameter space is large and one needs to carefully determine which parameter is important for the problem at hand. Therefore, in the first pass we did not find the need to consider the role of adhesion.

Minor comments:

The writing could be condensed in some places, with some details being moved to SI (for instance, section E on ageing is very short and seem more suited for supplements, or at least not as an independent section, note that the figure numbering also jumps to Fig. 9 there, although it’s Fig. 3 just before and Fig. 9 just after - re-ordering into main and supporting figures would be clearer.

We thank the Reviewer for this recommendation. The ageing section, although is short, it does provide a line of evidence that equilibrium approaches could be valid. We have modestly expanded the section by moving Appendix D to the main text, a general suggestion made by Referee 1. We have tried to be consistent in the numbering of figures in the revision.

Reviewer #3

I am very much in favor of the manuscript in its present form - I only suggest commenting (in the manuscript) on the issue described below.

Motivated by the fact that the experimental system consists of living, motile cells the authors use an active particle model (eq. 6) with stochastic selfpropulsion as the only source for noise (zero-temperature). It would be useful to elaborate briefly how important this stochastic self-propulsion is for the emergent rheological properties of the system (as summarized above): would these properties also be present in the “passive” version of the same model at “non-vanishing” temperature, and if not, why? Or analogously in a “passive” version which is “shaken”, reminiscent of shaken granular matter? To clarify these issues would relate this study to (or discriminate it from) passive, but complex, liquids or granular matter.

We appreciate the reviewer’s positive feedback on our work. The reviewer has raised an important question concerning our model in which self-propulsion serves as the source of noise. Without self-propulsion, the system would come to a stationary state after reaching mechanical equilibrium. As mentioned in Eqn. (6) (in the main text), we can define a characteristic time . It is possible that scaling the time t by τ would not alter the results.

The second question raised by the reviewer is also important. A passive version of the model would be to consider Eq. 6 in our article, and instead of using activity use the standard stochastic force. The resulting force would be at a finite temperature,. The coefficient of noise (a diffusion term) would be related to γi through the Fluctuation dissipation theorem(FDT)). Such a system of equations cannot ne mapped to Eq. 6 in which µ and γi are independently varied. It is unlikely that such a model, incorporating a “non-vanishing” temperature, would not result in the observed dependence of η on ϕ for the following reason. The passive model represents a polydisperse system, which would form a glass with η increasing with volume fraction, following the VFT law, as has been demonstrated in the glass transition literature for harmonic glasses. The other proposal whether the shaken version version would explain the experiments is also interesting. These are worth pursuing in future studies.

Author Response

The following is the authors’ response to the original reviews.

Thank you very much for the kind comments about our manuscript. We have improved the text to address all reviewers’ comments and suggestions. Additionally, we corrected and improved the supplementary tables.

Reviewer #1 (Public Review):

This paper provides new evidence on the relationship between genetic/chromosome divergence and capacity for asexual reproduction (via unreduced, clonal gametes) in hybrid males or females. Whereas previous studies have focussed just on the hybrid combinations that have yielded asexual lineages in nature, the authors take an experimental approach, analysing meiotic processes in F1 hybrids for combinations of species spanning different levels of divergence, whether or not they form asexual lineages in nature. As such, the findings here are a substantial advance towards understanding how new asexual lineages form.

The quality of the work is high, the analyses are sound, and the authors sensibly link their observations to the speciation continuum. I should also add that the cytogenetic work here is just beautiful!

A key finding is that the precondition for asexual reproduction - the formation of unreduced gametes - is not unusual among hybrid females, so that we have to consider other factors to explain the rarity of asexual species - a major unresolved issue in evolutionary biology. This work also highlights a previously overlooked effect of chromosome organisation on speciation.

Thank you for the nice comments about our work as well as for appreciating our cytogenetics work and figures.

Reviewer #2 (Public Review):

The authors investigate the origin of asexual reproduction through hybridization between species. In loaches, diploid, polyploid, and asexual forms have been described in natural populations. The authors experimentally cross multiple species of loaches and conduct an impressively detailed characterization of gametogenesis using molecular cytogenetics to show that although meiosis arrests early in male hybrids, a subset of cells in females undergo endoreplication before meiosis, producing diploid eggs. This only occurred in hybrids of parental species that were of intermediate divergence. This work supports an expanding view of speciation where asexuality could emerge during a narrow evolutionary window where genomic divergence between species is not too high to cause hybrid inviability, but high enough to disrupt normal meiotic processes.

Thank you.

I enjoyed reading this study and I appreciate the amount of work it takes to conduct these types of cytogenetic experiments. But, my main concern with this study is I was left wondering if the sample sizes are large enough to get a sense how variable endoreplication is in these loach species. Most of the hybrids between species are the result of crosses between 1-2 families. Within males and females, meiocyte observations are limited to a handful of pachytene and diplotene stages. I think it would be helpful to be more transparent about the sample sizes in the main text.

Thank you for raising this point. We have improved the Supplementary Tables S2 and S3 to clarify how many individuals we analyzed from each genetic family and added this information to the main text. In total we obtained 12 combinations with 19 F1 hybrid families. For the combination, C. elongatoides x C. taenia hybrids we obtained three families, for C. elongatoides x C. ohridana, C. elongatoides x C. tanaitica, C. elongatoides x C. bilineata and C. ohridana x C. bilineata, we obtained two families For the rest of the combinations of hybrids we obtained single family. From these families, 79 individuals were used for the analysis of the meiocites. Additionally, 24 parental individuals, males and females, were analysed. For the parental species, we analysed 852 cells, for hybrid males we investigated 244 cells, and 665 cells for hybrid females.

Along these lines, the authors argue against the possibility that endoreplication may be predisposed to occur at a higher rate in some species (line 291). Instead, they suggest that endoreplication is a result of perturbing the cell cycle by combining the genomes of two different species. Their main argument is based on gonocyte counts from parental females in a previous reference. It is essential to include counts from the parents used in this study to make a clear comparison with the F1s.

Thank you, we agree with your comment and included the observations of meiocytes from several parental species, i.e. C. elongatoides, C. taenia, C. pontica, C. tanaitica, and C. ohridana. Among 852 cells analyzed, we did not observe cells with duplicated genomes and abnormalities in chromosomal pairing. By contrast, among 665 pachytene cells of F1 hybrid females, we revealed altogether ~1% of endoreplicated ones. We tested these data by binomial GLM and found these differences to be significant, suggesting that sexuals, even if they may have some unnoticed duplication events, clearly have a significantly lower incidence of abnormal pachytene cells. We have now included this information in the main text.

In the discussion (lines 320-333), the authors postulate the sex-specific clonality they observe could be a result of Haldane's rule. Given these fish do not have known sex chromosomes, I do not find this argument strong. Haldane's rule refers to the exposure of recessive incompatibilities with the sex chromosomes in the hybrid heterogametic sex. This effect would therefore be limited to degenerated sex chromosomes where much of the sequence content on the Y or W has been lost. These species may have homomorphic sex chromosomes, but if this is the case, they likely are not very degenerated. Instead, it seems more plausible that the sex-specific effect the authors observe is due to intrinsic differences of spermatogenesis and oogenesis. Is there any information about sex-specific differences in the fidelity of gametogenesis from other species that would support a higher likelihood of endoreplication?

Thank you for this important question, however, we think it was a misunderstanding. We do not postulate that our observation conforms to Haldanes’ rule as, by contrast to this rule based on sex chromosomes, our previous publication demonstrated that whatever the gonadal sex differentiation is in our taxa, the ability to overcome sterility by asexual gametogenesis is always confined to female gonadal environment (or oogenesis in general), even in the transplanted spermatogonial cells (Tichopad et al. 2022). What we meant by our text is that our results do not fully conform to Haldane’s rule. We therefore reworded our text to rule out such a misconception.

Nonetheless, we note that it has been demonstrated that Haldanes’ rule is also applicable to species with little differentiated sex chromosomes (e.g. Presgraves and Orr 1998) and that recessive incompatibilities are not the only explanation as faster male theory or faster X may also apply in such cases (Dufresnes et al. 2016). Therefore, we have kept our remarks about Haldane’s rule here. Moreover, for several parental species, we preliminary found the occurrence of an XY gonadal sex differentiation system, albeit these are unpublished and need further validation.

The final thing I was left wondering about was this missing link between endoreplication and activating the embryonic development of the diploid egg. In these loach species, a sperm is required to activate egg development, but the sperm genome is discarded (line 100). What is the mechanism of this and how does it evolve concurrently during hybridization?

Thank you for the comment. There have been many speculations about why gynogens actually need sperm to activate their egg development, but to our knowledge, no explanation has been validated to date. Interestingly, a recent theoretical model by Fyon et al. BiorXiv 2023 suggested that the ability of sperm exclusion may evolve separately from the ability to produce clonal eggs. Hence, this topic is complex and remains unresolved, and we feel that it is out of the scope of the present MS. We have slightly modified the text and added 2 refs., to address your suggestion.

Reviewer #1 (Recommendations For The Authors):

The paper is well prepared - though the resolution of Fig 1 on the pdf is rather poor.

Thank you! We have now provided the high-resolution figures.

Overall, I have few suggestions for improvements:

Line 58. How does endoduplication itself "overcome accumulated incompatibilities" other than failure of synapsis? Perhaps by maintaining the F1 state, and so avoiding reduced fitness arising from recombination and disruption of coadapted gene combinations.

We have added a sentence to the main text “Premeiotic genome endoreplication thus not only ensures clonal reproduction but also allows hybrids to overcome problems in chromosome pairing that would otherwise lead to their sterility 15,17.” that we hope sufficiently addresses this issue.

Line 118 - please explain the AKD index here - as you have some in SI. Also please be clearer on how you measure genetic divergence as proportion of heterozygous SNPs - presumably this is via exon sequences from F1 females?

Please note that we have explained the AKD index in the relevant part of the Methods section already. However, we have now also added a brief explanation to the Results section, as suggested. We apologize for imprecise description of the genetic divergence measurements. As described in the Methods section, this is not measured by heterozygosity (as we wrongly stated here), but as p-distance among sequences of coding regions between parental species.

Lines 126 ff. It is unfortunate that the design of the crosses was not more balanced or extensive. Nonetheless, I do appreciate the effort involved here and think the results are solid as is.

Thank you.

Line 142. Please define PS and TB (and other acronyms) at first use.

We have added the definition for all acronyms at the first use.

Lines 192-193. What about EP and EN - as shown to have unreduced gametes in Fig. 2?

Thank you for this question. Based on analyses of the diplotene stage, we showed that EP and EN hybrids produced diploid eggs. However, in pachytene, we did not find duplicated oocytes due to the rarity of endoreplication. Similarly, the low incidence of duplicated pachytene cells was observed in natural as well as F1-hybrids in loaches and reptiles (Newton et al., 2016, Dedukh et al., 2021, 2022).

Lines 217-219. The observed correlation of chromosome divergence (AKD index) and numbers of bivalents in pachytene makes sense and is an important observation. Did this GLM simultaneously consider the effect of genetic divergence (as implied in methods)?

Thank you for this comment. We originally tested separately the fit of two models, one with AKD and the other with SNP divergence. Since the AKD model significantly outperformed the SNP-based one, we focused our interpretation on the former. However, as you suggested, we now re-calculated the model taking into account the joint effects of both predictors in a single model and indeed, this model outperformed both single predictors. In conclusion, while AKD is still the strongest single predictor for the observed amounts of bivalents, the additional effect of genetic distance still significantly improves the model fit. We have now included this result into the main text.

This finding does not alter our conclusions, it just suggests that the effect of chromosomal morphology is probably more complex, involving the role of more subtle sequence divergence or structural variants.

Line 242. The Discussion is a great read - careful interpretation and a really interesting interpretation in context of the broader literature.

Thank you for the appreciation. Your positive feedback and evaluation are highly motivating us to expand our work.

Line 396. Some references from book chapters (18, 52) are incomplete. Please fix.

We have now corrected these references accordingly.

Reviewer #2 (Recommendations For The Authors):

Transparency about meiocyte sample sizes: These counts are all in supplemental table 3. From this table, it is unclear if a majority of these meiocytes are from a single individual or from multiple males or females. Or, in the crosses where there are multiple families, are the meiocytes sampled from all families? I am trying to get a sense whether endoreplication and the fidelity of oogenesis could be influenced by genetic variants segregating within species. If the meiotcytes are only sampled from a single individual from a single cross, you may not see this variation. If this is the case, perhaps the correlation between genetic divergence and the formation of asexual clones may not be as strong. Additional replicates may not be feasible, but at a minimum I think it would be helpful to address whether endoreplication could or could not be variable and if the sample sizes are sufficient.

Thank you for raising this point. We have improved the Supplementary table to clarify how many individuals we analyzed from each family and added this information to the main text. Unfortunately, additional replicates are not feasible due to the long generation time of the fish. We otherwise agree with your comment and included this point in the Discussion.

Gonocyte counts from parental females: The authors say they "analysed hundreds of gonocytes of sexual females without a single incidence of genome endoreplication." I could not find a clear count in the references given. They note that the incidence of endoreplication was very low in pachytene cells in this study (0.7%).

Thank you, we agree with your comment and included the observations of meiocytes from several parental species, i.e. C. elongatoides, C. taenia, C. pontica, C. tanaitica, and C. ohridana. Among 852 cells analyzed, we did not observe cells with duplicated genomes and abnormalities in chromosomal pairing. By contrast, among 665 pachytenic cells of F1 hybrid females, we revealed altogether ~1% of endoreplicated ones. We tested these data by binomial GLM and found these differences to be significant, suggesting that sexuals, even if they may have some unnoticed duplication events, clearly have significantly lower incidence. of abnormal pachytene cells. We have now included this information in the main text.

They refer to supplemental table 4 (line 196), which does not exist in the supplement. The authors should report these numbers in the revised manuscript.

Thank you for pointing this out. We have corrected the name of the supplementary table, it actually is supplementary table S3.

Author Response

eLife assessment

The authors' finding that PARG hydrolase removal of polyADP-ribose (PAR) protein adducts generated in response to the presence of unligated Okazaki fragments is important for S-phase progression is potentially valuable, but the evidence is incomplete, and identification of relevant PARylated PARG substrates in S-phase is needed to understand the role of PARylation and dePARylation in S-phase progression. Their observation that human ovarian cancer cells with low levels of PARG are more sensitive to a PARG inhibitor, presumably due to the accumulation of high levels of protein PARylation, suggests that low PARG protein levels could serve as a criterion to select ovarian cancer patients for treatment with a PARG inhibitor drug.

Thank you for the assessment and summary. Please see below for details as we have now addressed the deficiencies pointed out by the reviewers.

We believe that PARP1 is one of the major relevant PARG substrates in S phase cells. Previous studies reported that PARP1 recognizes unligated Okazaki fragments and induces S phase PARylation, which recruits single-strand break repair proteins such as XRCC1 and LIG3 that acts as a backup pathway for Okazaki fragment maturation (Hanzlikova et al., 2018; Kumamoto et al., 2021). In this study, we revealed that accumulation of PARP1/2-dependent S phase PARylation eventually led to cell death (Fig. 2). Furthermore, we found that chromatin-bound PARP1 as well as PARylated PARP1 increased in PARG KO cells (Fig. S4A and Fig. 4A), suggesting that PARP1 is one of the key substrates of PARG in S phase cells. Of course, PARG may have additional substrates besides PARP1 which are required for its roles in S phase progression, as PARG is known to be recruited to DNA damage sites through pADPr- and PCNA-dependent mechanisms (Mortusewicz et al., 2011). Precisely how PARG regulates S phase progression warrants further investigation.

Reviewer #1 (Public Review):

I have a major conceptual problem with this manuscript: How can the full deletion of a gene (PARG) sensitize a cell to further inhibition by its chemical inhibitor (PARGi) since the target protein is fully absent?

Please see below for details about this point. Briefly, we found that PARG is an essential gene (Fig. 7). There was residual PARG activity in our PARG KO cells, although the loss of full-length PARG was confirmed by Western blotting and DNA sequencing (Fig. S9). The residual PARG activity in these cells can be further inhibited by PARG inhibitor, which eventually lead to cell death.

The authors state in the discussion section: "The residual PARG dePARylation activity observed in PARG KO cells likely supports cell growth, which can be further inhibited by PARGi". What does this statement mean? Is the authors' conclusion that their PARG KOs are not true KOs but partial hypomorphic knockdowns? Were the authors working with KO clones or CRISPR deletion in populations of cells?

The reviewer is correct that our PARG KOs are not true KOs. We were working with CRISPR edited KO clones. As shown in this manuscript, we validated our KO clones by Western blotting, DNA sequencing and MMS-induced PARylation. Despite these efforts and our inability to detect full-length PARG in our KO clones, we suspect that our PARG KO cells may still express one or more active fragments of PARG due to alternative splicing and/or alternative ATG usage.

As shown in Fig. 7, we believe that PARG is essential for proliferation. Our initial KO cell lines are not complete PARG KO cells and residual PARG activity in these cells could support cell proliferation. Unfortunately, due to lack of appropriate reagents we could not draw solid conclusions regarding the isoforms or the truncated PARG expressed in these cells (Please see Western blots below).

Are there splice variants of PARG that were not knocked down? Are there PARP paralogues that can complement the biochemical activity of PARG in the PARG KOs? The authors do not discuss these critical issues nor engage with this problem.

There are five reviewed or potential PARG isoforms identified in the Uniprot database. The sgRNAs used to generate initial PARG KO cells in this manuscript target all three catalytically active isoforms (isoforms 1, 2 and 3), while isoforms 4 and 5 are considered catalytically inactive according to the Uniprot database. However, it is likely that sgRNA-mediated genome editing may lead to the creation of new alternatively spliced PARG mRNAs or the use of alternative ATG, which can produce catalytically active forms of PARG. Instead of searching for these putative spliced PARG RNAs, we used two independent antibodies that recognize the C-terminus of PARG for WB as shown in Author response image 1. Unfortunately, besides full-length PARG, these antibodies also recognized several other bands, some of them were reduced or absent in PARG KO cells, others were not. Thus, we could not draw a clear conclusion which functional isoform was expressed in our PARG KO cells. Nevertheless, we directly measured PARG activity in PARG KO cells (Fig. S9) and showed that we were still able to detect residual PARG activity in these PARG KO cells. These data clearly indicate that residual PARG activity are present and detected in our KO cells, but the precise nature of these truncated forms of PARG remains elusive.

Author response image 1.

These issues have to be dealt with upfront in the manuscript for the reader to make sense of their work.

We thank this reviewer for his/her constructive comments and suggestions. We will include the data above and additional discussion upfront in our revised manuscript to avoid any further confusion by our readers.

Reviewer #2 (Public Review):

Summary:

In this manuscript, Nie et al investigate the effect of PARG KO and PARG inhibition (PARGi) on pADPR, DNA damage, cell viability, and synthetic lethal interactions in HEK293A and Hela cells. Surprisingly, the authors report that PARG KO cells are sensitive to PARGi and show higher pADPR levels than PARG KO cells, which are abrogated upon deletion or inhibition of PARP1/PARP2. The authors explain the sensitivity of PARG KO to PARGi through incomplete PARG depletion and demonstrate complete loss of PARG activity when incomplete PARG KO cells are transfected with additional gRNAs in the presence of PARPi. Furthermore, the authors show that the sensitivity of PARG KO cells to PARGi is not caused by NAD depletion but by S-phase accumulation of pADPR on chromatin coming from unligated Okazaki fragments, which are recognized and bound by PARP1. Consistently, PARG KO or PARG inhibition shows synthetic lethality with Pol beta, which is required for Okazaki fragment maturation. PARG expression levels in ovarian cancer cell lines correlate negatively with their sensitivity to PARGi.

Thank you for your nice comments. The complete loss of PARG activity was observed in PARG complete/conditional KO (cKO) cells. These cKO clones were generated using wild-type cells transfected with sgRNAs targeting the catalytic domain of PARG in the presence of PARP inhibitor.

Strengths:

The authors show that PARG is essential for removing ADP-ribosylation in S-phase.

Thanks!

Weaknesses:

1) This begs the question as to the relevant substrates of PARG in S-phase, which could be addressed, for example, by analysing PARylated proteins associated with replication forks in PARG-depleted cells (EdU pulldown and Af1521 enrichment followed by mass spectrometry).

We believe that PARP1 is one of the major relevant PARG substrates in S phase cells. Previous studies reported that PARP1 recognizes unligated Okazaki fragments and induces S phase PARylation, which recruits single-strand break repair proteins such as XRCC1 and LIG3 that acts as a backup pathway for Okazaki fragment maturation (Hanzlikova et al., 2018; Kumamoto et al., 2021). In this study, we revealed that accumulation of PARP1/2-dependent S phase PARylation eventually led to cell death (Fig. 2). Furthermore, we found that chromatin-bound PARP1 as well as PARylated PARP1 increased in PARG KO cells (Fig. S4A and Fig. 4A), suggesting that PARP1 is one of the key substrates of PARG in S phase cells. Of course, PARG may have additional substrates besides PARP1 which are required for its roles in S phase progression, as PARG is known to be recruited to DNA damage sites through pADPr- and PCNA-dependent mechanisms (Mortusewicz et al., 2011). Precisely how PARG regulates S phase progression warrants further investigation.

2) The results showing the generation of a full PARG KO should be moved to the beginning of the Results section, right after the first Results chapter (PARG depletion leads to drastic sensitivity to PARGi), otherwise, the reader is left to wonder how PARG KO cells can be sensitive to PARGi when there should be presumably no PARG present.

Thank you for your suggestion! However, we would like to keep the complete PARG KO result at the end of the Results section, since this was how this project evolved. Initially, we did not know that PARG is an essential gene. Thus, we speculated that PARGi may target not only PARG but also a second target, which only becomes essential in the absence of PARG. To test this possibility, we performed FACS-based and cell survival-based whole-genome CRISPR screens (Fig. 5). However, this putative second target was not revealed by our CRISPR screening data (Fig. 5). We then tested the possibility that these cells may have residual PARG expression or activity and only cells with very low PARG expression are sensitive to PARGi, which turned out to be the case for ovarian cancer cells. Equipped with PARP inhibitor and sgRNAs targeting the catalytic domain of PARG, we finally generated cells with complete loss of PARG activity to prove that PARG is an essential gene (Fig. 7). This series of experiments underscore the challenge of validating any KO cell lines, i.e. the identification of frame-shift mutations, absence of full-length proteins, and phenotypic changes may still not be sufficient to validate KO clones. This is an important lesson we learned and we would like to share it with the scientific community.

To avoid further misunderstanding, we will include additional statements/comments at the end of “PARG depletion leads to drastic sensitivity to PARGi” section and at the beginning of “CRISPR screens reveal genes responsible for regulating pADPr signaling and/or cell lethality in WT and PARG KO cells”. Hope that our revised manuscript will make it clear.

3) Please indicate in the first figure which isoforms were targeted with gRNAs, given that there are 5 PARG isoforms. You should also highlight that the PARG antibody only recognizes the largest isoform, which is clearly absent in your PARG KO, but other isoforms may still be produced, depending on where the cleavage sites were located.

The sgRNAs used to generate PARG KO cells in this manuscript target all three catalytically active isoforms (isoforms 1, 2 and 3), while isoforms 4 and 5 are considered catalytically inactive according to the Uniprot database. As suggested, we will modify Fig. S1D and the figure legends.

The manufacturer instruction states that the Anti-PARG antibody (66564S) can only recognize isoform 1, this antibody could recognize isoforms 2 and 3 albeit weakly based on Western blot results with lysates prepared from PARG cKO cells reconstituted with different PARG isoforms, as shown below. As suggested, we will add a statement in the revised manuscript and provide the Western blotting data in Author response image 2.

Author response image 2.

To test whether other isoforms were expressed in 293A and/or HeLa cells, we used two independent antibodies that recognize the C-terminus of PARG for WB as shown in Author response image 3. Unfortunately, besides full-length PARG, these antibodies also recognized several other bands, some of them were reduced or absent in PARG KO cells, others were not. Thus, we could not draw a clear conclusion which functional isoforms or truncated forms were expressed in our PARG KO cells.

Author response image 3.

4) FACS data need to be quantified. Scatter plots can be moved to Supplementary while quantification histograms with statistical analysis should be placed in the main figures.

We agree with this reviewer that quantification of FACS data may provide straightforward results in some of our data. However, it is challenging to quantify positive S phase pADPr signaling in some panels, for example in Fig. 3A and Fig. 4C. In both panels, pADPr signaling was detected throughout the cell cycle and therefore it is difficult to know the percentage of S phase pADPr signaling in these samples. Thus, we decide to keep the scatter plots to demonstrate the dramatic and S phase-specific pADPr signaling in PARG KO cells treated with PARGi. We hope that these data are clear and convincing even without any quantification.

5) All colony formation assays should be quantified and sensitivity plots should be shown next to example plates.

As suggested, we will include the sensitivity plot next to Fig. 3D. However, other colony formation assays in this study were performed with a single concentration of inhibitor and therefore we will not provide sensitivity plots for these experiments. Nevertheless, the results of these experiments are straightforward and easy to interpret.

6) Please indicate how many times each experiment was performed independently and include statistical analysis.

As suggested, we will add this information in the revised manuscript.

Reviewer #3 (Public Review):

Here the authors carried out a CRISPR/sgRNA screen with a DDR gene-targeted mini-library in HEK293A cells looking for genes whose loss increased sensitivity to treatment with the PARG inhibitor, PDD00017273 (PARGi). Surprisingly they found that PARG itself, which encodes the cellular poly(ADP-ribose) glycohydrolase (dePARylation) enzyme, was a major hit. Targeted PARG KO in 293A and HeLa cells also caused high sensitivity to PARGi. When PARG KO cells were reconstituted with catalytically-dead PARG, MMS treatment caused an increase in PARylation, not observed when cells were reconstituted with WT PARG or when the PARG KO was combined with PARP1/2 DKO, suggesting that loss of PARG leads to a strong PARP1/2-dependent increase in protein PARylation. The decrease in intracellular NADH+, the substrate for PARP-driven PARylation, observed in PARG KO cells was reversed by treatment with NMN or NAM, and this treatment partially rescued the PARG KO cell lethality. However, since NAD+ depletion with the FK868 nicotinamide phosphoribosyltransferase (NAMPT) inhibitor did not induce a similar lethality the authors concluded that NAD+ depletion/reduction was only partially responsible for the PARGi toxicity. Interestingly, PARylation was also observed in untreated PARG KO cells, specifically in S phase, without a significant rise in γH2AX signals. Using cells synchronized at G1/S by double thymidine blockade and release, they showed that entry into S phase was necessary for PARGi to induce PARylation in PARG KO cells. They found an increased association of PARP1 with a chromatin fraction in PARG KO cells independent of PARGi treatment, and suggested that PARP1 trapping on chromatin might account in part for the increased PARGi sensitivity. They also showed that prolonged PARGi treatment of PARG KO cells caused S phase accumulation of pADPr eventually leading to DNA damage, as evidenced by increased anti-γH2AX antibody signals and alkaline comet assays. Based on the use of emetine, they deduced that this response could be caused by unligated Okazaki fragments. Next, they carried out FACS-based CRISPR screens to identify genes that might be involved in cell lethality in WT and PARG KO cells, finding that loss of base excision repair (BER) and DNA repair genes led to increased PARylation and PARGi sensitivity, whereas loss of PARP1 had the opposite effects. They also found that BER pathway disruption exhibited synthetic lethality with PARGi treatment in both PARG KO cells and WT cells, and that loss of genes involved in Okazaki fragment ligation induced S phase pADPr signaling. In a panel of human ovarian cancer cell lines, PARGi sensitivity was found to correlate with low levels of PARG mRNA, and they showed that the PARGi sensitivity of cells could be reduced by PARPi treatment. Finally, they addressed the conundrum of why PARG KO cells should be sensitive to a specific PARG inhibitor if there is no PARG to inhibit and found that the PARG KO cells had significant residual PARG activity when measured in a lysate activity assay, which could be inhibited by PARGi, although the inhabited PARG activity levels remained higher than those of PARG cKO cells (see below). This led them to generate new, more complete PARG KO cells they called complete/conditional KO (cKO), whose survival required the inclusion of the olaparib PARPi in the growth medium. These PARG cKO cells exhibited extremely low levels of PARG activity in vitro, consistent with a true PARG KO phenotype.

We thank this reviewer for his/her constructive comments and suggestions.

The finding that human ovarian cancer cells with low levels of PARG are more sensitive to inhibition with a small molecule PARG inhibitor, presumably due to the accumulation of high levels of protein PARylation (pADPr) that are toxic to cells is quite interesting, and this could be useful in the future as a diagnostic marker for preselection of ovarian cancer patients for treatment with a PARG inhibitor drug. The finding that loss of base excision repair (BER) and DNA repair genes led to increased PARylation and PARGi sensitivity is in keeping with the conclusion that PARG activity is essential for cell fitness, because it prevents excessive protein PARylation. The observation that increased PARylation can be detected in an unperturbed S phase in PARG KO cells is also of interest. However, the functional importance of protein PARylation at the replication fork in the normal cell cycle was not fully investigated, and none of the key PARylation targets for PARG required for S phase progression were identified. Overall, there are some interesting findings in the paper, but their impact is significantly lessened by the confusing way in which the paper has been organized and written, and this needs to be rectified.

We believe that PARP1 is one of the major relevant PARG substrates in S phase cells. Previous studies reported that PARP1 recognizes unligated Okazaki fragments and induces S phase PARylation, which recruits single-strand break repair proteins such as XRCC1 and LIG3 that acts as a backup pathway for Okazaki fragment maturation (Hanzlikova et al., 2018; Kumamoto et al., 2021). In this study, we revealed that accumulation of PARP1/2-dependent S phase PARylation eventually led to cell death (Fig. 2). Furthermore, we found that chromatin-bound PARP1 as well as PARylated PARP1 increased in PARG KO cells (Fig. S4A and Fig. 4A), suggesting that PARP1 is one of the key substrates of PARG in S phase cells. Of course, PARG may have additional substrates besides PARP1 which are required for its roles in S phase progression, as PARG is known to be recruited to DNA damage sites through pADPr- and PCNA-dependent mechanisms (Mortusewicz et al., 2011). Precisely how PARG regulates S phase progression warrants further investigation.

As suggested, we will revise our manuscript accordingly and provide additional explanation/statement upfront to avoid any misunderstandings.

Author Response

The following is the authors’ response to the original reviews.

Reviewer #1:

1) Utilization of known AhR ligands as controls will strengthen the interpretation of the conclusions.

We agree with the reviewer that AhR ligands could be used as controls for delineating structure-activity relationships and cell context-specific effects. However, such studies are beyond the scope of the current manuscript. The AhR has many endogenous ligands, including several tryptophan derived metabolites, that have been shown to elicit different responses depending on the dose and cell type. Our unpublished data show that the expression of AhR target genes such as Cyp1a1, Cyyp2e1, and Tiparp were not modulated by I3A in RAW cells, which suggests that the observed effects may occur independent of the AhR.

Reviewer #2:

Specific comments:

1) The title is misleading "Microbially-derived indole-3-actate" suggests that this article is about the production of I3A by the gut microbiota, in fact this is a dietary supplementation article. The title needs to reflect this fact.

Our title reflects the natural source of I3A in mice. We used oral supplementation to study the effects of this metabolite. Per suggestion by the reviewer, we changed the title as follows: <br /> “Oral supplementation of gut microbial metabolite indole-3-acetate alleviates diet-induced steatosis and inflammation in mice”

2). The description of the amount of I3A in the drinking water is not properly described. The actual concentration in the drinking water should be given.

The concentration of I3A in drinking water was as follows: WD50 = 0.5mg/ml and WD100 = 1mg/ml. We added this information in the revised manuscript.

3) The serum concentration data of I3A is critical data and should be moved in Figure 1.

We have now included serum levels of I3A as part of Figure 1.

4) The authors should have determined the actual concentration of indole-3-actetate in serum by running a standard curve of I3A during the LC-MS analysis. Also, recovery and matrix effects should be determined. Without this information their data will be difficult to compare to other studies.

We agree with the reviewer that quantification of I3A in serum would be useful. However, we are unable to do so due to limited sample available as well as concerns with sample integrity after long-term storage.

5) In the data in Figure S1C, there appears to be only 2-3 mice out of nine that exhibit a difference in serum indole-3-acetate levels between the WD-50 and WD-100. Do the authors have an explanation for this small difference compared to the other endpoints assessed?

The serum I3A measurements at week 16 are a snapshot that may not reflect tissue levels due to differences in water intake, I3A metabolism in the body, and/or elimination of I3A. The other phenotypic assays are physiological measurements that reflect the result of sustained administration of I3A.

6) Since the Ah receptor may play a role in the results obtained CYP1A1 mRNA levels in the liver and intestinal tract should have been measured.

We measured alterations in Cyp1a1 mRNA in the liver and no significant change was observed in the WD50 and WD100 groups relative to controls. Also, see response to reviewer 1.

7) The main mechanistic experiment performed is shown in Figure 6 and the figure legend states that they are examining macrophages, but these are cell lines, they are macrophages models, and this should be clearly stated. The first two panels are liver data, so the title of the figure legend needs to reflect that fact.

We agree and have changed the title of Figure 6 to “I3A modulates AMPK phosphorylation and suppresses RAW 264.7 macrophage cell inflammation in an AMPK dependent manner”.

8) In Figure 6, 1 mM I3A is added to the cells, how is this very high concentration relevant to the concentrations observed in vivo? Does adding 1 mM acetate to the cell culture media lower the pH of the media and could this influence the results obtained? Would acetic acid yield the same results? Could treatment with an acid even explain in vivo results?

It is difficult to match the concentration of I3A in the in vitro experiments to liver tissue concentrations. Addition of 1 mM I3A did not lower the pH of cell culture media or reduce the viability of cultured RAW 264.7 macrophages. As I3A is not known to degrade into acetic acid and indole, we do not expect acetic acid to recapitulate the effects elicited by I3A.

Reviewer #3:

My primary concern with the manuscript is the organization and interpretation of the data. It appears that little effort was given by the authors on interpreting the data and digesting it for the reader into a coherent package. Rather, the authors have collected a vast amount of data and organized it without much thought about what the reader would take away from it. Furthermore, it seems the authors have taken this as an opportunity to overload this manuscript with data that are superfluous to the conclusions the authors draw at the end. Based on this, I think the authors need to invest more time into distilled their complex biological data into a unifying scientific interpretation for the readers that advances our understanding of I3A. My suggestions for the authors are described below.

1) The data lack a rationale behind how they are organized within the manuscript. For example, the authors will combine disparate biological pathways and lump data together without logic as in Figure 2. Why are inflammatory pathways and bile acid synthesis combined in a figure? What was the rationale?

We respectfully disagree that the data are presented without rationale. Both inflammation and bile acid dysregulation are commonly observed with NAFLD and thus are presented in two separate panels of Figure 2 (A, inflammatory cytokines, and B bile acids).

2) The authors give very little effort to performing integrative omics analysis even though multi-omics is provided. Example given, the authors provide proteomic data on the fatty acid metabolism pathway, however, no mention of this pathway within the metabolomic dataset. Vice versa, the authors provide in depth investigation in the metabolic changes within the tryptophan pathway, however, no investigation into the proteomic changes that may underlie this phenomenon. It would be recommended that the authors invest more energy into performing more in-depth analysis of their multi-omics data presented.

We attempted to co-analyze the proteomic and metabolomic data, but this analysis was not informative. Protein and metabolite abundances do not necessarily correlate, and the two types of omics data carry different observation biases. For example, label-free, untargeted proteomics data favor abundant proteins, whereas untargeted metabolomics data are influenced by concentration and ionization efficiency, among other factors. Therefore, we opted to analyze the two datasets independently, and then linked the findings from the two analyses using biological pathways as guides. For example, we describe changes in acyl-carnitine and discuss how this observation is consistent with changes in abundance of fatty acid metabolism enzymes.

3) Figures 1&2 shows that low dose treatment reduces inflammation but does not alter hepatic TG levels. This is in direct disagreement with the graphical model provided by the authors (Supp. Fig 9). In the author's model, I3A is directing hepatic lipid metabolism through modulation of macrophage inflammation. This interpretation is erroneous and needs to be reevaluated by the authors. Furthermore, the tryptophan pathway and bile acid pathways are not even represented in the model, which begs the question of why that data are included in the manuscript to begin with.

We would like to respectfully point out that Figure 1D does show a statistically significant (p < 0.05) difference in liver TG between the WD and WD100 groups. Supp. Figure S9 is meant to be a summary of the main biochemical changes elicited by I3A that we have shown in the current study (e.g., the involvement of AMPK) rather an atlas of all the changes detected in the metabolomics and proteomic data. Specifically, we have not included the tryptophan or bile acid pathways as we do not have mechanistic information on how these changes are mediated by I3A.

4) The authors switch from hepatocytes to macrophages without giving any rationale, The authors need to invest more time into describing a logical flow of thought when assembling the manuscript.

We mention the rationale for investigating the effect of I3A on macrophages in the introduction (last paragraph of the section): “In vitro, both I3A and TA attenuated the expression of inflammatory cytokines (Tnfα, Il-1β and Mcp-1) in macrophages exposed to palmitate and LPS.”. We also explain why we used an in vitro model, RAW cells, at the beginning of the corresponding Results section: “Since our previous study found that the metabolic effects of I3A in hepatocytes depend on the AhR, we tested if this was also the case in macrophages.” Moreover, the strong effects of I3A on liver inflammatory cytokines also motivates the macrophage experiments.

Author Response

We thank the Editors and the Reviewers for the time spent on our manuscript entitled “The CD4 transmembrane GGXXG and juxtamembrane (C/F)CV+C motifs mediate pMHCII-specific signaling independently of CD4-Lck interactions”. We appreciate the helpful feedback and the opportunity to participate in eLife’s new model for publishing.

We are writing to provide the following provisional author responses for posting with the first version of the reviewed preprint:

1) To address comments about the limited scope of this study and referencing of the Methods section to our prior study, we would like to note that we submitted the current study via the Research Advance mechanism. Our goal was to build upon the conclusions of our 2022 eLife publication (PMID: 35861317) and address an unresolved question from that study (as nicely summarized by Reviewer #2). In the current manuscript we present data from reductionist experiments that were designed specifically for this purpose and, as noted by the reviewers, we provide answers to the question being asked. We think that the Research Advance mechanism is an ideal opportunity to make these results available to the field given the stated purpose of such articles (for reference: “A Research Advance might use a new technique or a different experimental design to generate results that build upon the conclusions of the original research by, for example, providing new mechanistic insights or extend the pathway under investigation…”).

a. The Methods were not duplicated in this manuscript because we referenced our prior study as per instructions for the Research Advance mechanism.

2) The constituent residues of the motifs analyzed in this and our prior study were determined to be functionally significant in vivo through the computational reconstruction of CD4’s evolutionary history, which provided us with data from ~435 million years of natural experiments with CD4 in numerous jawed vertebrate species. We agree that having conditional knock-in mice of these CD4 mutants, and those characterized in our last study, would be useful for determining how these mutations impact T cell development, activation, differentiation, and effector function. Given the costs involved with making genetically engineered mouse model systems, the computational and experimental data we have generated in the current and prior study will help us prioritize next steps to dig deeper into the details of why the residues we are studying are under purifying selection (fail to propagate to progeny if mutated, meaning terminal). In short, only now, with the data in hand, can we prioritize mouse studies. We think it is important for the advancement of the field that we make these results available in a timely manner rather than waiting to report them together with the results of mouse models once generated and analyzed.

3) The reductionist experimental data presented here provide us with mechanistic insights into why the residues we are studying are functionally important. We therefore think it is of value to note that 58a-b- T cell hybridomas were used in seminal work that established a link between CD4Lck association, via motifs in the CD4 intracellular domain, and signaling output as measured by IL-2 production (Glaichenhaus, et al., 1991). Importantly, the impact of disrupting CD4-Lck interactions on proximal signaling were not interrogated until the work we describe here and in our preceding study, wherein we establish that CD4-Lck association does not regulate proximal signaling in 58a-b- T cell hybridomas. Given that this experimental system was used to help establish the dominant paradigm (i.e. the widely held view that CD4 recruits Lck to TCR-CD3 to initiate pMHCII-specific signaling), we think it is a legitimate system to directly test this model and further test core questions of CD4 function by employing more modern experimental techniques.

Author Response:

We would like to express our heartfelt gratitude for the reviewers’ scholarly and insightful reviews of our manuscript. The constructive comments and thought-provoking experimental proposals have been invaluable not only in improving the quality of this study but also in shaping the direction of future research. In revision, all comments will be addressed point-by-point, and the manuscript will be revised thoroughly. Here in this reply, we focus on the most critical issue regarding the source of noises during stability inference.

When faced a stack of objects, individuals are more likely to assess taller stacks of objects as being more unstable compared to shorter ones (Fig. 2b & 2d). This bias persists even when comparing single objects of different heights that share the same contact area with the supporting surface. Known as “stability inference bias,” this phenomenon challenges deterministic models with a single, fixed vector for the representation of gravity’s direction (i.e., directly downward). To reconcile this bias with deterministic models, previous studies (e.g., Allen et al., 2020; Battaglia et al., 2013; Kubricht et al., 2017) have incorporated external noises such as perceptual uncertainty and external force perturbations to increase their fit to human performance, also pointed out by Reviewer 1.

In this study, we introduced an alternative perspective through a stochastic model in which variability is instead embedded in the representation of gravity’s direction. In this framework, gravity’s direction is not a fixed vector but a distribution of possible vectors, with the vertical direction serving as the maximum likelihood. While the distinction between deterministic and stochastic models is conceptually clear, mathematically they are equivalent. In addition, our stochastic model does not negate the role of external noises in stability inference, because gravity is seldom the sole force acting upon a moving object in the physical world, as pointed out by Reviewer 1. Together, these two factors make it challenging to ascribe the source of variability to either external or internal noises (Smith & Vul, 2013). This is the major concern raised by all three reviewers.

To distinguish between the deterministic and stochastic models, we designed a series of experiments aimed at demonstrating that internal noises, rather than external noises such as perceptual uncertainty or external force perturbations, influences our inference about object stability. However, the supporting evidence was dispersed and at times implicit throughout the manuscript. In revision, we will thoroughly clarify the ambiguities. In this reply, we will consolidate and present the evidence comprehensively.

1. The examination of external noises.

1.1 External Force Perturbations. Deterministic models suggests that during object stability inference, individuals implicitly assume the presence of external forces (e.g., wind) that could destabilize stacks. While this assumption aligns with the omnipresence of such forces in natural settings, it overlooks a crucial variable: the directionality of these external forces. In psychological studies, individual differences are commonly observed, and the perceived force direction is not an exception. That is, some may assume that it comes from the left, while others from the right. In essence, if external forces were to play a significant role in stability inference, one would expect the perceived force directions to exhibit non-uniform distributions (i.e., anisotropy) in the horizontal plane within individuals and to show substantial variability between individuals.

Contrary to this expectation, our study revealed a different pattern. In the study, we specifically measured the distribution of 𝜑, the horizontal component reflecting the direction of object collapse. Our results indicated that all participants exhibited a uniform distribution of gravity’s directions in the horizontal plane (Fig. 1d right; Extended Data Fig. 2 and 3). This uniformity suggests that if external forces were a key determinant in stability inference, participants would have to assume a varying direction of external force in each trial—an assumption we consider unlikely. Instead, our RL model simulation suggests that the isotropy of 𝜑 arises from agent-environment interactions, notably in the absence of external forces (Extended Data Fig. 6).

In summary, the uniform distribution of horizontal direction component, 𝜑, observed in all participants, challenges the argument for the dominant role of external forces in stability inference. We are sorry that this aspect was not explicitly emphasized in the original text, and in revision we will explain why external forces are unlikely to substantially shape our perception of object stability.

1.2 Perceptual uncertainty. To assess the impact of perceptual uncertainty on stability inference, we examined whether the representation of gravity’s direction is cognitive impenetrable. Specifically, we posited that if noises are external (i.e., perceptual uncertainty), the inference bias should be modulated by task context; in contrast, if noises are internal, the stochastic representation of gravity’s direction will be encapsulated from the context. To test this idea, we inverted the virtual environment, making gravity appear to point upward (also see a similar idea by Reviewer 3). In this unfamiliar context, which diverges dramatically from daily experiences, one would expect heightened perceptual uncertainty, which according to deterministic models would result in a larger inference bias – manifested as an increased width of the distribution (𝜎) of gravity’s direction. Contrary to this prediction, we observed that the width of the distribution remained unchanged (Fig. 1d and 1f). Furthermore, there was a high correlation (r = 0.91) between widths in the upright and inverted conditions across participants (Extended Data Fig. 2 and 3).

In summary, this finding suggests that the manipulation of perceptual uncertainty is unable to cognitively penetrate the representation of gravity’s direction, casting doubt on its dominant role in stability inference. We are sorry that in the original text, we did not clarify the rationale for employing the approach of cognitive impenetrability. In revision, this will be clarified.

2. The origin of intrinsic noises in stability inference.

In deterministic models, either external force perturbations or perceptual uncertainty is often assumed but rarely empirically tested. Indeed, these external noises are introduced primarily to account for observed biases in stability inference. In this study, we explicitly examined the possible origin of the intrinsic noises embedded in the representation of gravity’s direction. Without assumed perceptual uncertainty and external perturbation of forces, the RL model simulation showed that the distribution could evolve naturally based mainly on the agent’s experience, as it used the mismatch between the expectation and the observed state of the stack under natural gravity to update its representation of gravity’s direction (Fig. 3a). Importantly, the width of the distribution for the agent was comparable to that of human participants as measured in the psychophysics experiments (Fig. 3b). Therefore, the experience alone may be sufficient to generate stochastic representation of gravity’s direction, obviating the need for external noises.

Taken together, these findings underscore the limitations of the combination of deterministic models and external noises in accounting for stability inference, and suggest that intrinsic noises embedded in the representation of gravity play a pivotal role in shaping our stability inference of the physical world.

3. Thought experiments.

Although the evidence shown above may provide valuable insights, our study does not definitively settle the debate between deterministic models and our proposed stochastic model. Specifically, our study only preliminarily investigates two sources of external noise, perceptual uncertainty and external force perturbations, leaving many other factors such as object mass and surface friction, unexplored (for studies on these factors, please see Hamrick et al., 2016). As such, the reviewers have proposed a series of thought experiments that warrant further investigation. Below, we enumerate some of them, followed by ours.

3.1 Experiment 1. Reviewer 3 proposed a thought experiment in which participants assess stability of a single block of varying heights. The reviewer argues that a block, regardless of its height, will remain stable on a horizontal surface unless externally disturbed. This assertion is perfectly true in the physical realm. However, in the cognitive domain, both deterministic models and our stochastic model predict differently. Take an extreme example of a standing needle: while it would remain upright in the physical world without external disturbances, both deterministic and stochastic models, which account for mental inference of physical events, will predict a likelihood of it falling, aligning with our subjective feelings. This is because in both models, noises are considered in the intuitive physics engine. In deterministic models, external force perturbations, as well as perceptual uncertainty, are assumed to be omnipresent noises in probabilistic reasoning. In our stochastic model, noises are embedded in the representation of gravity’s direction. Therefore, although this thought experiment, along with other thought experiments on object mass, surface friction (proposed by Reviewer 3), and falling trajectories behind an occlude (proposed by Reviewer 1), is insightful, but it cannot serve to differentiate deterministic and stochastic models. 3.2 Experiment 2. Reviewer 2 suggested constructing a wall on one side of the virtual scene to make it improbable that participants would infer an external force perturbation emanating from that direction. In this setting, deterministic models would predict a non-uniform distribution of the horizontal component, 𝜑, skewed away from the wall. In contrast, according to our stochastic model, the distribution of 𝜑 would remain unaffected, maintaining the uniform distribution observed in previous experiments. Extending this logic, another test scenario could contrast an indoor scene with an outdoor scene. In a confined and static indoor environment, the likelihood of external force perturbations should be much lower than in a dynamic, open outdoor setting. Here, deterministic models would predict an increase in the width of the distribution, 𝜎, in the outdoor environment, whereas our model would anticipate no such change. The underlying rationale for these experiments parallels that of our previous setup (figure 1e), where we inverted the virtual environment and reversed the direction of gravity. Indeed, they all aim to assess the extent to which manipulations of external factors can cognitively penetrate the representation of gravity’s direction.

3.3 Experiment 3: A noteworthy insight derived from our RL model simulation relates to variations in the number of blocks within the virtual worlds. Deterministic models would predict an enlarged bias in stability inference as the number of blocks increased, which is attributed to elevated levels of perceptual uncertainty and an expanded area susceptible to external force perturbations. However, the results from our RL model simulation contradict this prediction, revealing that an augmented number of blocks instead led to a narrowing of the width of the distribution. This decrease in width can be ascribed to richer information provided by a larger number of blocks for refining its representation of gravity’s direction. In line with this rationale, we propose a new experiment from the perspective of ecological psychology, which emphasizes that cognitive processes are shaped by our interactions with the environment. Specifically, we hypothesize that individuals raised in mountainous terrains may exhibit more accurate representations of gravity’s direction than those raised in flat terrains. This proposed experiment could not only help resolving the ongoing debate between two models to some extent, but also advocate future studies on intuitive physics within a more ecologically valid framework.

To conclude, both deterministic and stochastic models align closely with Bayesian principles, where stability inference is conceptualized as probabilistic reasoning. Nevertheless, the divergence between them is no trivial, as it hinges on distinct philosophical assumptions about the relationship between the inner mind and the external world. Deterministic models propose that the mind serves as a faithful reflection of the world; therefore, gravity’s direction is represented as a single, fixed vector directly downward, the same as that in the world. In these models, uncertainty for probabilistic reasoning emanates from factors external to the module of the intuitive physics engine. In contrast, our stochastic model underscores the notion that the mind is an active inference machine, continually reinterpreting inputs from outside world; therefore, the mind gains increased adaptability, allowing for a more nuanced accounting of uncertainty in the world – factors often crucial for survival. Such active inference necessitates flexible representations; accordingly, within the model of intuitive physics engine, variations are embedded into the representation of gravity’s direction. While resolving this philosophical debate is beyond the capacity of the present study, we contend that the field of intuitive physics offers a valuable lens through which to pry open the complex interplay between the mind and the world we live in.

References

• Allen, K. R., Smith, K. A., & Tenenbaum, J. B. (2020). Rapid trial-and-error learning with simulation supports flexible tool use and physical reasoning. Proceedings of the National Academy of Sciences, 117(47), 29302–29310.
• Battaglia, P. W., Hamrick, J. B., & Tenenbaum, J. B. (2013). Simulation as an engine of physical scene understanding. Proceedings of the National Academy of Sciences, 110(45), 18327–18332.
• Kubricht, J. R., Holyoak, K. J., & Lu, H. (2017). Intuitive physics: Current research and controversies. Trends in Cognitive Sciences, 21(10), 749–759.
• Smith, K. A., & Vul, E. (2013). Sources of uncertainty in intuitive physics. Topics in Cognitive Science, 5(1), 185–199.

Author Response:

Reviewer #1 (Public Review):

Summary: The authors made significant updates to Hippacampome.org including 50 new cell types.

Strengths: The authors have been thorough in basing their views on peer-reviewed literature. They have made the data highly accessible and the user has the ability to control what is included.

Weaknesses: There are many inconsistencies in the literature regarding cell types and how these are incorporated into hippocampome.org is not clear.

We agree with the Reviewer that there can be inconsistencies in the literature, especially when it comes to nomenclature. This is why for Hippocampome.org v1.0 we decided to focus on the morphologies, the distributions of axons and dendrites across the layers and parcels of the hippocampal formation, rather than the names authors have applied to the neurons they are studying. We have also clarified our stance on nomenclature in our Brain Informatics manuscript that accompanied v1.1. We will revise the manuscript to make these points explicit.

Properties are often a result of modeling and not biological data, and caveats to this approach, and other assumptions are unclear.

The foundation for Hippocampome.org has always been the data that are published in the literature. Those include, among others, the axonal and dendritic spans in each layer and subregion, the molecular expression patterns, the total neuron count by layer and subregion, the membrane properties, firing patterns, and experimental synaptic signals and corresponding covariates. For all of those, we do not depend on how the data are modeled, although there is always some level of interpretation of the data to make them machine readable and ready for incorporation into our database. However, some of the simulation-ready parameters now also included in Hippocampome.org are indeed the result of modeling, such as the neuronal input/output functions (Izhikevich model) and the unitary synaptic values (Tsodyks-Markram model). Other simulation-ready parameters are the result of specific analysis approaches, including the connection probabilities (axonal-dendritic spatial overlaps) and the neuron type census (numerical optimization of all constraints). We plan to explicitly distinguish among these various cases in the revised manuscript.

Several interneuron subtypes in the dentate gyrus do not appear to be listed, such as neurogliaform cells.

The neuron types listed in Figure 2 of the current manuscript are only the new additions to the catalog of neuron types at Hippocampome.org v2.0. DG Neurogliaform cells were included in our original eLife manuscript, which described the deployment of v1.0 of the website. We will clarify this in the revisions.

The nomenclature HIPROM should be distinguished or made synonymous with HIPP. Same for MOCAP and MOPP/HICAP.

The Reviewer has referred to 5 separate neuron types in Hippocampome.org. Each neuron type has a unique distribution of axonal and dendritic invasions of the 26 layers and parcels of the hippocampal formation. For example, HIPROM cells have dendrites in the inner one-third of stratum moleculare, stratum granulosum, and hilus and axons in all four layers of the dentate gyrus in addition to axonal projections into CA3 stratum radiatum, stratum lucidum, stratum pyramidale, and stratum oriens. HIPP cells in contrast have dendrites only in the hilus and axons only in the outer two-thirds of stratum moleculare with no cross-subregional projections. Similar considerations distinguish MOPP, MOCAP, and HICAP cells in Hippocampome.org. In expanding the nomenclature to include the neuron types we first described at Hippocampome.org, we attempted to mimic the styling of the already established neuron types of the DG: HIPROM (Hilar Interneuron with PRojections to the Outer Molecular layer), HIPP (HIlar Perforant Path-associated), MOCAP (MOlecular Commissural-Associational Pathway-related axons and dendrites), MOPP (MOlecular layer Perforant Path-associated), and HICAP (HIlar Commissural-Associational Pathway-related). We intend to insert a paragraph in the revised version to clarify these issues.

Dorsal ventral and sex differences are not mentioned.

We thank the Reviewer for pointing this out. As a result of the dearth of literature describing differences between dorsal and ventral hippocampus when we first assembled Hippocampome.org v1.0, we made the decision to focus solely on the distributions of the axons and dendrites along the depth, or layers, of the hippocampal formation. As the amount of literature concerning relating to the other axes of the hippocampus continues to grow, we will gradually incorporate information along the added dimensions into our knowledge base. In the revised manuscript we intend to note this, and also stress the fact that Hippocampome.org contains knowledge from a mixture of sexes, and that whenever the original papers report the animal sex, so does our knowledge base. The revised manuscript will also mention that, whenever possible (e.g. synaptic physiology parameters), values are reported separately for males and females.

Reviewer #2 (Public Review):

Summary and strengths: The authors have developed a helpful resource for the community regarding hippocampal cell types and their interactions from many perspectives. There have been many updates to hippocampome v1.0 to v1.12, that are nicely summarized and explained (e.g., Table 1). The content and impact are also presented (Fig. 4).

Weaknesses: My main comment is that it is not completely clear and/or it is a bit buried as to what makes this v2.0 (rather than v1.13). The title would seem to encompass it ('... enabling data-driven spiking neural network simulations...), but in the introduction, the authors seem to emphasize "50 newly identified neuron types...". Is it the case that launching network simulations (using CARLsim) was not possible up to v1.12? I don't think so? I think that this research advance is to announce and summarize the various updates and to demonstrate how network simulations can be easily done? If so, this should and could be made more clear so that the reader does not necessarily have to go through all the previous versions to understand what is 'special' or different about v2.0. This could perhaps be achieved by situating their tool and its goals relative to other efforts (e.g., blue brain project) that are mentioned in the Discussion?

We thank the Reviewer for their helpful suggestions. Hippocampome.org v1.12 included the final piece needed, the synaptic physiology parameter values, to start fully simulating the hippocampal formation. In the revised manuscript, we will endeavor to emphasize more the specialness of v2.0 over the various v1.X in the Abstract, Introduction, and Discussion, in part by more fully describing the differences between our work and that of other efforts, such as the Blue Brain Project.

Reviewer #3 (Public Review):

Summary: The authors aim to provide a multidisciplinary resource on the structural and physiological organization of the hippocampal system and make the available experimental data available for further theoretical work, providing tools to do so in a very flexible and user-friendly way. Since this is a new version of an already existing data-resource, the authors certainly reach their aim and fulfil expectations that the reader might have. The content of the database is as good as the original data, collected from the published knowledge-database, sometimes with the help of the original authors, and the overall quality depends further on how the data are curated by the team of authors and many others who helped them. That process is briefly described and more details are available in descriptions of previous versions and on the website. The data extraction, examples of how data can be used, and the part on attempts to model the hippocampus are exciting and open doors to new and exciting research opportunities.

Strengths: Excellent description with many outlined opportunities. Nicely illustrated and inviting to explore the online database.

Weaknesses: The figures are complex, containing a heavy information load with many abbreviations. You need some general knowledge of the system in order to grasp the enormous potential of what is provided.

We agree with the Reviewer that we generously used abbreviations throughout our figures as a means of conserving limited space. We have attempted to balance that by providing a complete glossary of all the abbreviations used throughout the manuscript. However, we will make an effort to supply definitions of the abbreviations in the figure captions and at their first use in the manuscript, or even replacing the abbreviations altogether in key places in the figures.

Author Response

We are very thankful for the editors' and reviewers' thoughtful feedback and criticisms on our manuscript. We have carefully considered all of the comments and will provide a revised manuscript with detailed responses as soon as we can. In the meantime, we will make our best effort to conduct additional experiments to further support our conclusions.We greatly appreciate the time and consideration given to improving our work.

Reviewer #1 (Public Review):

Summary:

The question at hand is whether astrocytes contribute to the mechanism of long-term synaptic potentiation (LTP) at synaptic contacts between excitatory glutamatergic neurons and inhibitory neurons (E-I synapses). This is a legitimate query considering the immense body of work that has now established synaptic plasticity (LTP, LTD and spike-timing dependent plasticity) as an astrocyte-dependent process at excitatory synapses and, by contrast, the lack of knowledge on whether and how astrocytes control IN activity. Taking direct inspiration from that same body of work, authors recapitulate a number of experiments and approaches from prior seminal studies and provide evidence that E-I synapses in the stratum radiatum of the hippocampus display NMDAR-dependent plasticity, which can be suppressed by pharmacologically hindering astrocytes physiology, preventing astrocyte Ca2+ transients or blocking endocannabinoid CB1 receptors. Under any of these conditions, LTP can still be rescued by exogenously applying D-serine, a naturally occurring co-agonist of NMDARs primarily released by astrocytes. Coincidently, authors show that the conditions used to elicit LTP also cause a transient increase in NMDAR co-agonist site occupancy. Lastly, based on some evidence that gamma-CaMKII is predominantly expressed in INs rather than excitatory neurons, authors conducted AAV-mediated IN-specific gamma-CaMKII shRNA experiments and found that this is sufficient to suppress LTP at E-I synapses. They found that this approach also impairs contextual fear learning in behaving mice. Authors conclude that astrocytes gate LTP at E-I synapses via a mechanism wherein neuronal depolarization during LTP induction elicits endocannabinoid release which drives CB1-dependent astrocyte Ca2+ activity, causing the release of the NMDAR co-agonist D-serine (required for NMDAR activation).

Strengths:

This is an important question and the experimental work seems to have been conducted at high standards. The electrophysiology traces are impeccable, the experiments are well powered, including the behavioral testing, and multiple controls and validations are provided throughout. The figures are clear and easy to understand. Overall, the conclusions from the study are consistent, or partially consistent, by the findings.

We greatly appreciate you taking the time to evaluate our study thoroughly and provide such thoughtful feedback.

Main Weaknesses:

1) A major point of concern is the lack of proper acknowledgment of the seminal studies that were mimicked in this manuscript, notably Henneberger et al, Nature 2010, Adamsky et al, Cell 2018; and Robin et al., Neuron 2017. The entire study design is a replica of these landmark studies: it isn't built upon or inspired from them, it exactly repeats the experiments and methods performed in them, coming dangerously close to being simply a hidden attempt to plagiarize published work. The resemblance goes as far as using an identical figure display (see Fig4.D vs Fig 2D of Ref#4). The issue is that authors frame the problem, scientist logic, reasoning, technical tricks, approaches, and interpretations as their own whereas, in reality, they were taken verbatim out of previous work and applied to a (shockingly) similar problem. The probity of the present study is thus in question. Authors need to clearly acknowledge, in all relevant instances, that the work presented here recapitulates the approach, reasoning and methodology used in past seminal studies that tackled the mechanisms of astrocyte regulation of LTP.

Thank you very much for your review and valuable comments on our manuscript. We greatly appreciate your concern regarding the proper acknowledgment of previous studies. We sincerely apologize for not adequately citing and acknowledging the seminal works in our manuscript. We highly value avoiding academic misconduct.

For the research design, although there are some similarities between our work and other studies, our key scientific questions and technical approaches are markedly different, as evidenced by our central hypothesis and experimental methods. We did not completely replicate their research design.

Regarding research methods, many basic techniques like electrophysiology, chemogenetic are common experimental methods, not patented by any one paper. Our choice of methods is based on the research needs, not to replicate a particular paper. But we recognize that there are similarities in our experimental methods, specifically the chemogenetic stimulation of astrocytes to induce de novo LTP, which has been inspired by previous studies (Van Den Herrewegen et al. Molecular Brain (2021), Adamsky et al. Cell (2018), Nam et al. Cell reports (2019)). We were also inspired by the previous work of Henneberger et al. in Nature (2010) to investigate whether stimulation, specifically we using TBS (theta burst stimulation), could transiently increase NMDA receptor-mediated synaptic responses.

For the similarity between our Fig. 4D and Fig. 2D of Ref. 4, it is primarily because both studies have the similar purpose(we monitored NMDA currents in interneurons, others monitored in pyramidal cells) using similar methods, but our figure layout follows a regular display pattern. Additionally, we would like to draw your attention to our previous studies, specifically Shen et al., Scientific Reports (2017), Supplementary figure 4, and Shen et al., Journal of Neurochemistry (2021), Supplementary figures 8 and 9. In these studies, we also employed a regular display pattern in our figure layouts. It is important to note that while there may be similarities in the figure arrangement, each study presents distinct findings and contributes to the broader understanding of the topic.Our use of a similar way to present data does not equal plagiarism. We apologize for any confusion caused by the lack of explicit citation and acknowledgment in our manuscript again. In the revised version, we will ensure to provide clear and detailed references to all relevant studies.

In terms of citations, we have cited Henneberger et al, Nature 2010, Adamsky et al, Cell 2018; and Robin et al., Neuron 2017.'s work in multiple places, indicating we have learned from their research ideas and findings. We will supplement any missing citations. But overall, our work has distinct differences and innovations.

We are not intended as a hidden attempt to plagiarize or simply replicate their methods. Rather, they are part of a deliberate effort to establish a comparable and reproducible experimental framework. Our study aims to validate and further explore the conclusions drawn by replicating the experiments of these seminal studies and deepening our understanding of the mechanisms of astrocyte regulation of LTPE-I.

We sincerely appreciate your review and guidance. We will carefully consider your criticism and incorporate more accurate and thorough citations in the revised version, ensuring proper respect and acknowledgment of the previous works.

2) Relatedly, in past work, field recordings were used to monitor LTP in hippocampal slices (refs 4, 26 and others). This method captures indiscriminately all excitatory synapses where glutamate is released to cause AMPAR-dependent (and NMDAR) transmembrane flux of cations in the postsynaptic element, including E-I synapses and not just E-E synapse like the authors claim. Therefore, a strong argument can be made that there is no actual ground to differentiate the present results from past ones.

Thank you for your thoughtful comments regarding the differentiation of our results from previous studies. We appreciate the opportunity to address this issue and provide further clarification.

Indeed, in past studies, field recordings were commonly utilized to monitor long-term potentiation (LTP) in hippocampal slices. It is true that this method captures all flux of cations in excitatory synapses, inhibitory synapses and glia. This includes both excitatory-excitatory (E-E) and excitatory-inhibitory (E-I) synapses.

When using the LTP recording protocol, one limitation is that the experimenter cannot determine the exact contribution of E-E and E-I currents to the recorded current. Additionally, it is not possible to know, with the same induction protocol, the specific effects on E-E synapses versus E-I synapses. It is plausible that E-E synapses could undergo LTP, while E-I synapses could undergo LTD, or vice versa.

Thus, it becomes crucial to carefully dissect the functioning of E-I synapses and investigate how astrocytes modulate these synapses. Past field recordings have provided important insights, our selective interrogation of the astrocyte-E-I synapse interface represents a conceptual advance to delineate the nuanced modulation of distinct synaptic connections by astrocytes. We specifically focus on studying the modulation of E-I synapses by astrocytes and aim to elucidate the intricate dynamics and underlying mechanisms. By untangling the complex contributions of astrocytes to E-I synapse function and plasticity, we can unveil novel aspects of neuroglial interactions and advance our understanding of the fundamental principles governing neural network activity.

3) There is a general lack of excitement about this study. One reason is that it replicates almost identically past work, as mentioned above. Another is that the scientific question and importance of the findings are not framed appropriately. The work is presented as an astrocyte-focused investigation, but it has very limited value to the astrocyte field. The findings are, in all accounts, identical to those unveiled by previous work especially because E-I synapses are, in fact, excitatory synapses. Where this study does bring value, however, is to the field of interneurons, but it would need to be reframed to shift the emphasis from astrocytes to E-I connections. Authors would need to elevate the text by framing their work around relevant considerations, such as IN diversity, mechanisms of LTP in IN subtypes, role of E-I connections in hippocampal circuit function, information processing, behavior, spatial learning, navigation, or grid cells activity etc...

We appreciate your insightful comments and concerns regarding the lack of excitement surrounding our study. We would like to clarify that while our study use similar certain methodologies, for example electrophysiology, chemogenetics and pharmacology, our research aims to provide a deeper understanding of the underlying mechanisms of how astrocytes regulate E-I synapses. We apologize if this replication aspect was not adequately highlighted in our manuscript, and we will make sure to emphasize the novel contributions of our study in the revised version.

Regarding the framing of our study, we recognize the importance of interneurons and the role of E-I connections in hippocampal circuit function, information processing, behavior, spatial learning, navigation, and other relevant aspects. However, the scientific question and scope of the study are to explore whether and how astrocytes modulate E-I synapses. We believe that this study brings value to the field of astrocyte-neuron interaction. Of course, this study also brings value to the field of interneurons. Perhaps the lack of excitement among audiences stems from the mechanisms for astrocytes modulating E-I and E-E synapses are the same.

4) A clear weakness of the study is that it fails to consider the molecular and functional diversity of interneurons in the stratum radiatum and provides no insights or considerations related to it. Authors provide no information on what type of IN were patched, or the location of their cell body in the s.r., effectively treating all patched IN as a homogeneous ensemble of cells - which they are not. Relatedly, the study is extremely evasive on the importance of the results in the context of inhibitory interneurons. This renders the significance of the insights highly uncertain and dampens both the impact of the study and the excitement it generates. Hippocampal interneurons are very diverse in molecular identity, sub-anatomical location, morphology, projections, connectivity and functional importance. Some experts go as far as recognizing 29 subtypes in the CA1, including 9 in the stratum radiatum alone (based on the location of their soma). However, this is neither addressed nor acknowledged by the authors, with the exception of a statement (line 659) where they claim to have "focused on a subpopulation of interneurons in the stratum radiatum" without providing any precision or evidence to corroborate this assertion. This diversity, alone, could explain why not all cells showed LTP, or why the mechanisms authors describe in the radiatum do not seem to be at play in the oriens. Hence, carefully considering the diversity of INs in the present work is necessary. It would refine and augment the conclusions of the paper. Instead of a sub-region specificity, the study might fuel the notion of an IN subtype specificity of LTP mechanisms, which is more useful to the field.

Thank you very much for your review and valuable comments on our study. We agree with the point you raised regarding a clear weakness in our study, specifically the lack of consideration the diversity of interneurons in the stratum radiatum.

As the reviewer notes, there are many subtypes of interneurons in hippocampal region CA1 that likely contribute in distinct ways to circuit function. Unfortunately we did not gather information on the specific molecular or morphological identity of the interneurons we recorded from.This is a limitation of our study. We will add discussion of this issue as a caveat, and highlighted it as an opportunity for future work to dissect how long-term potentiation in interneurons regulated by astrocytes may differ across interneuron subpopulations. Thank you once again for your insightful comments.

5) Authors take several shortcuts. Some of the conclusions are a leap from the experiments and are only acceptable due to the close analogy with very similar investigations conducted in the past that provided identical results. For instance, the present study provides no evidence of any sort that D-serine is involved - rather, it provides evidence that the pathway at hand contributes to increasing the occupancy of the co-agonist binding site of NMDARs. Considering the absence of work demonstrating that D-serine is the endogenous co-agonist of NMDARs at E-I synapses, most of the authors claims on D-serine are unfounded. This would necessitate using tools such as the canonical D-serine scavengers DAAS or DsDA, serine racemase KO mice etc. Similarly, authors provide no compelling evidence that endocannabinoid CB1 receptors involved in this pathway are located on astrocytes

Thank you for your insightful comments on our study. We appreciate your attention to detail and your concerns regarding our conclusions. We agree that further evidence is needed to establish the involvement of D-serine as the endogenous co-agonist of NMDARs at E-I synapses. We will take into consideration your suggestion of using tools such as D-serine scavengers to provide clearer evidence.

Regarding the involvement of endocannabinoid CB1 receptors on astrocytes in this pathway, we provide evidence that astrocytic calcium signaling could blocked by CB1 receptor antagonist AM251, as shown in figure 3.However, we agree that further research is necessary to accurately identify the localization of CB1 receptors. As part of our future investigations, we will take note of this limitation in our discussion and emphasize the need for additional studies to explore the precise location of CB1 receptors. In addition, we will endeavor to perform immunohistochemistry to identify the exact location of CB1 receptors in astrocytes.

Thank you once again for your valuable feedback. We will carefully address these concerns and make appropriate revisions to ensure the clarity and accuracy of our findings.

6) An important caveat in this study is the protocol employed to induce LTP, which includes steps of sustained depolarization of the patched IN to -10mV. Neuronal depolarization is known to induce endocannabinoids production. In several instances, this was shown to 'activate' astrocytes and elicit the release of astrocyte-derived transmitters at nearby synapses. This implies that the endocannabinoid-dependent pathway described in the study is, most likely, artificially engaged by the protocol itself. Hence, the present work only provides evidence that an astrocyte-dependent, CB1-D-serine-pathway can be artificially called upon with this specific LTP protocol, but does not convincingly demonstrate that it is naturally occurring or necessary for plasticity at E-I synapses. Authors would need to thoroughly address this caveat by replicating some of their key findings (AM251, calcium-clamp, D-serine and CaMKII shRNA) using a protocol that does not entail the artificial depolarization of the patched interneuron.

Thank you for raising this important point. We agree that the sustained depolarization protocol we used to induce LTP could potentially engage endocannabinoid signaling and astrocyte activation. However, we observed that preventing astrocyte Ca2+ transients or blocking endocannabinoid CB1 receptors prevented the induction of LTP by this depolarization protocol suggests that this astrocyte-endocannabinoid-dependent pathway is necessary,

Importantly, synaptic depolarization of neurons can occur naturally during learning and memory. Though ‘artificial’ here, our protocol may mimic aspects of natural activity patterns that engage ‘endocannabinoid release’ and astrocyte involvement in plasticity.

Another limitation of our study is that we currently cannot conclusively determine the source of the CB1. We cannot distinguish whether the CB1 originates from neurons or astrocytes based on our current experiments. We will explicitly acknowledge this caveat in the discussion, noting that further experiments are needed to clarify the cellular origin of the CB1. Thank you for drawing our attention to this critical issue - we will refine the manuscript accordingly to more comprehensively and accurately present the study conclusions and limitations. Your feedback helps improve the rigor of our research.

7) Reading and understanding are hindered by a rather vast array of issues with the text itself. It needs thorough editing for typos, misnomers, meaning-altering errors in syntax, and a number of issues with English.

Thank you very much for your review and feedback on our text. We highly appreciate your comments and take them seriously. We will carefully address the issues you mentioned and thoroughly edit the text to eliminate any typos, misnomers, syntax errors that may alter the meaning, and other English-related issues. We truly value your input and appreciate your patience as we work on these improvements.

Reviewer #2 (Public Review):

Summary:

This work explores the implication of astrocytes in the regulation of long-term potentiation of excitatory synapses onto inhibitory neurons in CA1 hippocampus. They found that astrocytes of a sub-region of CA1 regulate this plasticity through their activation of endocannabinoids that lead to the release of the NMDA receptor co-agonist, D-serine.

Strengths:

The experiments are well considered and conceptualized, and use appropriate tools to explore the role of astrocytes in the tripartite synapse. The results highlight a novel role of astrocytes in an important aspect of the synaptic regulation of the hippocampal circuit. There are extensive levels of analysis for each experimental group of evidence.

Thank you for your positive feedback on our study. We appreciate your recognition of the careful consideration and conceptualization of our experiments, as well as the use of appropriate tools to investigate the role of astrocytes in the tripartite synapse. We are pleased to hear that the results have highlighted a novel role of astrocytes in an important aspect of synaptic regulation in the hippocampal circuit.

Thank you for taking the time to review our work and for providing such positive feedback. We will continue to improve and refine our study based on your valuable comments.

Weaknesses:

The authors underscore and used an oversimplified view of the heterogeneity of interneuron populations and their selective roles in the hippocampal network. Also, there is an uneven level of astrocyte-selective tools used in the different experiments which creates an uneven strength of arguments and conclusions regarding the role of glial cells. Finally, the wording used by the authors often lead to some confusion or sense of overinterpretation

We appreciate the reviewer raising these important points about the characterization of interneuron and astrocyte populations in our study. We agree that oversimplifying or overlooking cellular heterogeneity could undermine the conclusions. In the revised manuscript, we will:

1) Add more detailed discussion of interneuron diversity. We will note this as an area for further study.

2) Review the wording used when describing results and conclusions, ensuring we avoid overstating interpretations of the data.

Thank you again for the thoughtful feedback.

Author Response

We thank the reviewers for truly valuable advice and comments. We have made multiple corrections and revisions to the original pre-print accordingly. Here we address 2 major points.

1) Regarding the genetic association of the common COL11A1 variant rs3753841 (p.(Pro1335Leu)), we do not propose that it is the sole risk variant contributing to the association signal we detected and have clarified this in the manuscript. We concluded that it was worthy of functional testing for reasons described here. Although there were several common variants in the discovery GWAS within and around COL11A1, none were significantly associated with AIS and none were in linkage disequilibrium (R2>0.6) with the top SNP rs3753841. We next reviewed rare (MAF<=0.01) coding variants within the COL11A1 LD region of the associated SNP (rs3753841) in 625 available exomes representing 46% of the 1,358 cases from the discovery cohort. The LD block was defined using Haploview based on the 1KG_CEU population. Within the ~41 KB LD region (chr1:103365089- 103406616, GRCh37) we found three rare missense mutations in 6 unrelated individuals, Author response table 1. Two of them (NM_080629.2: c.G4093A:p.A1365T; NM_080629.2:c.G3394A:p.G1132S), from two individuals, are predicted to be deleterious based on CADD and GERP scores and are plausible AIS risk candidates. At this rate we could expect to find only 4-5 individuals with linked rare coding variants in the total cohort of 1,358 which collectively are unlikely to explain the overall association signal we detected. Of course, there also could be deep intronic variants contributing to the association that we would not detect by our methods. However, given this scenario, the relatively high predicted deleteriousness of rs3753841 (CADD= 25.7; GERP=5.75), and its occurrence in a Gly-X-Y triplet repeat, we hypothesized that this variant itself could be a risk allele worthy of further investigation.

Author response table 1.

We also appreciate the reviewer’s suggestion to perform a rare variant burden analysis of COL11A1. We conducted pilot gene-based analysis in 4534 European ancestry exomes including 797 of our own AIS cases and 3737 controls and tested the burden of rare variants in COL11A1. SKATO P value was not significant (COL11A1_P=0.18) but this could due to lack of power and/or background from rare benign variants that could be screened out using the functional testing we have developed.

2) Regarding functional testing, by knockdown/knockout cell culture experiments, we showed for the first time that Col11a1 negatively regulates Mmp3 expression in cartilage chondrocytes, an AIS-relevant tissue. We then tested the effect of overexpressing the human wt or variant COL11A1 by lentiviral transduction in SV40-transformed chondrocyte cultures. We deleted endogenous mouse Col11a1 by Cre recombination to remove the background of its strong suppressive effects on Mmp3 expression. We acknowledge that Col11a1 missense mutations could confer gain of function or dominant negative effects that would not be revealed in this assay. However as indicated in our original manuscript we have noted that spinal deformity is described in the cho/cho mouse, a Col11a1 loss of function mutant. We also note the recent publication by Rebello et al. showing that missense mutations in Col11a2 associated with congenital scoliosis fail to rescue a vertebral malformation phenotype in a zebrafish col11a2 KO line. Although the connection between AIS and vertebral malformations is not altogether clear, we surmise that loss of the components of collagen type XI disrupt spinal development. in vivo experiments in vertebrate model systems are needed to fully establish the consequences and genetic mechanisms by which COL11A1 variants contribute to an AIS phenotype.

Author Response

The following is the authors’ response to the current reviews.

We thank both reviewers for their detailed and positive assessment of our work.

To Reviewer #2, we have now explicated the pattern -- (QXQXQX>3)4 where X>3 denotes any length of three or more residues of any composition -- in the first paragraph of the discussion.

To Reviewer #3, we have made slight modifications to the text in the “Q zippers poison themselves” results section, to attempt to further clarify the mechanism of self-poisoning.

Briefly, the reviewer questions if an alternative model -- where inhibition involves non-structured rather than Q-zipper containing oligomers -- better explains the data. We provided two lines of evidence that we believe exclude this alternative model. First, we point out in the first paragraph of the “Q zippers poison themselves” section that the cells that unexpectedly lack amyloid in the high concentration regime have negligible levels of AmFRET, indicating that the inhibitory oligomers themselves occur at low concentrations regardless of the total concentration, and are therefore limited by a kinetic barrier. Second, we point out in the third paragraph of the section that the severity of amyloid inhibition with respect to concentration has a sequence dependence that matches the expectation of converging phase boundaries for crystal polymorphs -- specifically, inhibition is most severe for sequences that have a local Q density just high enough to form a Q zipper on both sides of each strand. Inhibition relaxed for sequences having more or less Qs than that threshold. In contrast, disordered oligomerization is not expected to have such a dependence on the precise pattern of Qs and Ns.

---

The following is the authors’ response to the original reviews.

We are pleased that the editors find our study valuable. We find that the reviewers’ criticisms largely arise from misunderstandings inherent to the conceptually challenging nature of the topic, rather than fundamental flaws, as we will elaborate here. We are grateful for the opportunity afforded by eLife to engage reviewers in what we intend to be a constructive public dialogue.

Response to Reviewer 1

This review is highly critical but lacks specifics. The reviewer’s criticisms reflect a position that seems to dismiss a critical role for (or perhaps even the existence of) conformational ordering in polyQ amyloid, which is untenable.

The reviewer states that our objective to characterize the amyloid nucleus “rests on the assertion that polyQ forms amyloid structures to the exclusion of all other forms of solids”. We do not fully agree with this assertion because our findings show that detectable aggregation is rate-limited by conformational ordering, as evident by 1) its discontinuous relationship to concentration, 2) its acceleration by a conformational template, and 3) its strict dependence on very specific sequence features that are consistent with amyloid structure but not disordered aggregation).

We strongly disagree with the reviewer’s subjective statement that we have not critically assessed our findings and that they do not stand up to scrutiny. This statement seems to rest on the perceived contradiction of our findings with that of Crick et al. 2013. Contrary to the reviewer’s assessment, we argue here that the conclusions of Crick et al. do more to support than to refute our findings. Briefly, Crick et al. investigated the aggregation of synthetic Q30 and Q40 peptides in vitro, wherein fibrils assembled from high concentrations of peptide were demonstrated to have saturating concentrations in the low micromolar range. As explained below, this finding of a saturating concentration does not refute our results. More relevant to the present work are their findings that “oligomers” accumulated over an hours-long timespan in solutions that are subsaturated with respect to fibrils, and these oligomers themselves have (nanomolar) critical concentrations. The authors postulated that the oligomers result from liquid–liquid demixing of intrinsically disordered polyglutamine. However, phase separation by a peptide is expected to fix its concentration in both the solute and condensed phases, and, because disordered phase separation is faster than amyloid formation, the postulated explanation removes the driving force for any amyloid phase with a critical solubility greater than that of the oligomers. In place of this interpretation that truly does appear to -- in the reviewer’s words -- “contradict basic physical principles of how homopolymers self-assemble”, we interpret these oligomers as evidence of Q zipper-containing self-poisoned multimers, rounded as an inherent consequence of self-poisoning (Ungar et al., 2005), and plausibly akin to semicrystalline spherulites that have been observed in other polymer crystal and amyloid-forming systems (Crist and Schultz, 2016; Vetri and Foderà, 2015). Importantly, the physical parameters governing the transition between amyloid spherulites and fibrils have been characterized in the case of insulin (Smith et al. 2012), where it was found that spherulites form at lower protein concentrations than fibrils. This mirrors the observation by Crick et al. that fibrils have a higher solubility limit than the spherical oligomers. . Further rebuttal to the perceived incompatibility of monomeric nucleation with the existence of a critical concentration for amyloid

We appreciate that the concept of a monomeric nucleus can superficially appear inconsistent with the fact that crystalline solids such as polyQ amyloid have a saturating concentration, but this is only true if one neglects that polyQ amyloids are polymer crystals with intramolecular ordering. The perceived discrepancy is perhaps most easily dispelled by the fact that folded proteins can form crystals, and the folded state of the protein. These crystals have critical concentrations, and the protein subunits within them each have intramolecular crystalline order (in the form of secondary structure). When placed in a subsaturated solution, the protein crystals dissolve into the constituent monomers, and yet those monomers still retain intramolecular order. Our present findings for polyQ are conceptually no different.

To further extrapolate this simple example to polyQ, one can also draw on the now well-established phenomenon of secondary nucleation, whereby transient interactions of soluble species with ordered species leads to their own ordering (Törnquist et al., 2018). Transience is important here because it implies that intramolecular ordering can in principle propagate even in solutions that are subsaturated with respect to bulk crystallization. This is possible in the present case because the pairing of sufficiently short beta strands (equivalent to “stems” in the polymer crystal literature) will be more stable intramolecularly than intermolecularly, due to the reduced entropic penalty of the former. Our elucidation that Q zipper ordering can occur with shorter strands intramolecularly than intermolecularly (Fig. S4C-D) demonstrates this fact. It is also evident from published descriptions of single molecule “crystals” formed in sufficiently dilute solutions of sufficiently long polymers (Hong et al., 2015; Keller, 1957; Lauritzen and Hoffman, 1960).

In suggesting that a saturating concentration for amyloid rules out monomeric nucleation, the reviewer assumes that the Q zipper-containing monomer must be stable relative to the disordered ensemble. This is not inherent to our claim. The monomeric nucleating structure need not be more stable than the disordered state, and monomers may very well be disordered at equilibrium at low concentrations. To be clear, our claim requires that the Q zipper-containing monomer is both on pathway to amyloid and less stable than all subsequent species that are on pathway to amyloid. The former requirement is supported by our extensive mutational analysis. The latter requirement is supported by our atomistic simulations showing the Q zipper-containing monomer is stabilized by dimerization (included in our 2021 preprint). Hence, requisite ordering in the nucleating monomer is stabilized by intermolecular interactions. We provide in Author response image 1 an illustration to clarify what we believe to be the discrepancy between our claim and the reviewer’s interpretation.

Author response image 1.

That the rate-limiting fluctuation for a crystalline phase can occur in a monomer can also be understood as a consequence of Ostwald’s rule of stages, which describes the general tendency of supersaturated solutes, including amyloid forming proteins (Chakraborty et al., 2023), to populate metastable phases en route to more stable phases (De Yoreo, 2022; Schmelzer and Abyzov, 2017). Our findings with polyQ are consistent with a general mechanism for Ostwald’s rule wherein the relative stabilities of competing polymorphs differ with the number of subunits (De Yoreo, 2022; Navrotsky, 2004). As illustrated in Fig. 6 of Navrotsky, a polymorph that is relatively stable at small particle sizes tends to give way to a polymorph that -- while initially unstable -- becomes more stable as the particles grow. The former is analogous to our early stage Q zipper composed of two short sheets with an intramolecular interface, while the latter is analogous to the later stage Q zipper composed of longer sheets with an intermolecular interface. Subunit addition stabilizes the latter more than the former, hence the initial Q zipper that is stabilized more by intra- than intermolecular interactions will mature with growth to one that is stabilized more by intermolecular interactions.

We have added a new figure (Fig. 6) to the manuscript to illustrate qualitative features of the amyloid pathway we have deduced for polyQ.

Rebuttal to the perceived necessity of in vitro experiments

The overarching concern of this reviewer and reviewing editor is whether in-cell assays can inform on sequence-intrinsic properties. We understand this concern. We believe however that the relative merit of in-cell assays is largely a matter of perspective. The truly sequence-intrinsic behavior of polyQ, i.e. in a vacuum, is less informative than the “sequence-intrinsic” behaviors of interest that emerge in the presence of extraneous molecules from the appropriate biological context. In vitro experiments typically include a tiny number of these -- water, ions, and sometimes a crowding agent meant to approximate everything else. Obviously missing are the myriad quinary interactions with other proteins that collectively round out the physiological solvent. The question is what experimental context best approximates that of a living human neuron under which the pathological sequence-dependent properties of polyQ manifest. We submit that a living yeast cell comes closer to that ideal than does buffer in a test tube.

The reviewer’s statements that our findings must be validated in vitro ignores the fact -- stressed in our introduction -- that decades of in vitro work have not yet generated definitive evidence for or against any specific nucleus model. In addition to the above, one major problem concerns the large sizes of in vitro systems that obscure the effects of primary nucleation. For example, a typical in vitro experimental volume of e.g. 1.5 ml is over one billion-fold larger than the femtoliter volume of a cell. This means that any nucleation-limited kinetics of relevant amyloid formation are lost, and any alternative amyloid polymorphs that have a kinetic growth advantage -- even if they nucleate at only a fraction the rate of relevant amyloid -- will tend to dominate the system (Buell, 2017). Novel approaches are clearly needed to address these problems. We present such an approach, stretch it to the limit (as the reviewer notes) across multiple complementary experiments, and arrive at a novel finding that is fully and uniquely consistent with all of our own data as well as the collective prior literature.

That the preceding considerations are collectively essential to understand relevant amyloid behavior is evident from recent cryoEM studies showing that in vitro-generated amyloid structures generally differ from those in patients (Arseni et al., 2022; Bansal et al., 2021; Radamaker et al., 2021; Schmidt et al., 2019; Schweighauser et al., 2020; Yang et al., 2022). This is highly relevant to the present discourse because each amyloid structure is thought to emanate from a different nucleating structure. This means that in vitro experiments have broadly missed the mark in terms of the relevant thermodynamic parameters that govern disease onset and progression. Note that the rules laid out via our studies are not only consistent with structural features of polyQ amyloid in cells, but also (as described in the discussion) explain why the endogenous structure of a physiologically relevant Q zipper amyloid differs from that of polyQ.

A recent collaboration between the Morimoto and Knowles groups (Sinnige et al.) investigated the kinetics of aggregation by Q40-YFP expressed in C. elegans body wall muscle cells, using quantitative approaches that have been well established for in vitro amyloid-forming systems of the type favored by the reviewer. They calculate a reaction order of just 1.6, slightly higher than what would be expected for a monomeric nucleus but nevertheless fully consistent with our own conclusions when one accounts for the following two aspects of their approach. First, the polyQ tract in their construct is flanked by short poly-Histidine tracts on both sides. These charges very likely disfavor monomeric nucleation because all possible configurations of a four-stranded bundle position the beginning and end of the Q tract in close proximity, and Q40 is only just long enough to achieve monomeric nucleation in the absence of such destabilization. Second, the protein is fused to YFP, a weak homodimer (Landgraf et al., 2012; Snapp et al., 2003). With these two considerations, our model -- which was generated from polyQ tracts lacking flanking charges or an oligomeric fusion -- predicts that amyloid nucleation by their construct will occur more frequently as a dimer than a monomer. Indeed, their observed reaction order of 1.6 supports a predominantly dimeric nucleus. Like us and others, Sinnige et al. did not observe phase separation prior to amyloid formation. This is important because it not only argues against nucleation occurring in a condensate, it also suggests that the reaction order they calculated has not been limited by the concentration-buffering effect of phase separation.

While we agree that our conclusions rest heavily on DAmFRET data (for good reason), we do provide supporting evidence from molecular dynamics simulations, SDD-AGE, and microscopy.

To summarize, given the extreme limitations of in vitro experiments in this field, the breadth of our current study, and supporting findings from another lab using rigorous quantitative approaches, we feel that our claims are justified without in vitro data.

Rebuttals to other critiques

We do not deny that flanking domains can modulate the kinetics and stability of polyQ amyloid. However, as stated and referenced in the introduction, they do not appear to change the core structure. We have also added a paragraph concerning flanking domains to the discussion, and acknowledged that “the extent to which our findings will translate in these different contexts remains to be determined.” Nevertheless, that the intrinsic behavior of the polyQ tract itself is central to pathology is evident from the fact that the nine pathologic polyQ proteins have similar length thresholds despite different functions, flanking domains, interaction partners, and expression levels.

The reviewer states that we found nucleation potential to require 60 Qs in a row. Our data are collectively consistent with nucleation occurring at and above approximately 36 Qs, a point repeated in the paper. The reviewer may be referring to our statement, ”Sixty residues proved to be the optimum length to observe both the pre- and post-nucleated states of polyQ in single experiments”. The purpose of this statement is simply to describe the practical consideration that led us to use 60 Qs for the bulk of our assays. We do appreciate that the fraction of AmFRET-positive cells is very low for lengths just above the threshold, especially Q40. They are nevertheless highly significant (p = 0.004 in [PIN+] cells, one-tailed T-test), and we have modified the figure and text to clarify this.

The reviewer characterizes self-poisoning as the hallmark of crystallization from polymer melts, which would be problematic for our conclusions if self-poisoning were limited to this non-physiological context. In fact the term was first used to describe crystallization from solution (Organ et al., 1989), wherein the phenomenon is more pronounced (Ungar et al., 2005).

Response to Reviewer 2

We thank the reviewer for their detailed and helpful critique.

The reviewer correctly notes that the majority of our manipulations were conducted with 60-residue long tracts (which corresponds to disease onset in early adulthood), and this length facilitates intramolecular nucleation. However, we also analyzed a length series of polyQ spanning the pathological threshold, as well as a synthetic sequence designed explicitly to test the model nucleus structure with a tract shorter than the pathological threshold, and both experiments corroborate our findings.

The reviewer mentions “several caveats” that come with our result, but their subsequent elaboration suggests they are to be interpreted more as considerations than caveats. We agree that increasing sequence complexity will tend to increase homogeneity, but this is exactly the motivation of our approach. We explicitly set out to determine the minimal complexity sequence sufficient to specify the nucleating conformation, which we ultimately identified in terms of secondary and tertiary structure. We do not specify which parts of a long polyQ tract correspond to which parts of the structure, because, as the reviewer points out, they can occur at many places. Hence, depending on the length of the polyQ tract, the nucleus we describe may have any length of sequence connecting the strand elements. We do not think that the effects of N-residue placement can be interpreted as a confounding influence on hairpin position because the striking even-odd pattern we observe implicates the sides of beta strands rather than the lengths. Moreover, we observe this pattern regardless of the residue used (Gly, Ser, Ala, and His in addition to Asn).

We thank the reviewer for noting the novelty and plausibility of the self-poisoning connection. We would like to elaborate on our finding that self-poisoning inhibits nucleation (in addition to elongation), as this will be confusing to many readers. While self-poisoning is claimed to inhibit primary nucleation in the polymer crystal literature (Ungar et al., 2005; Zhang et al., 2018), the semantics of “nucleation” in this context warrants clarification. Technically, the same structure can be considered a nucleus in one context but not in another. The Q zipper monomer, even if it is rate-limiting for amyloid formation at low concentrations (and is therefore the “nucleus”), is not necessarily rate-limiting when self-poisoned at high concentrations. Whether it comprises the nucleus in this case depends on the rates of Q zipper formation relative to subunit addition to the poisoned state. If the latter happens slower than Q zipper formation de novo, it can be said that self-poisoning inhibits nucleation, regardless of whether the Q zipper formed. We suspect this to be the mechanism by which preemptive oligomerization blocks nucleation in the case of polyQ, though other mechanisms may be possible.

We believe the revised text also now incorporates the remaining suggestions of this reviewer, with two exceptions. 1) We retain the phrase “hidden pattern”, because we believe our data argue for a nucleus whose formation requires that Qs occur in a pattern that we now elaborate as (QXQXQX>3)4 where X>3 denotes any length of three or more residues of any composition. In amyloids formed from long polyQ molecules, the nucleus will involve any subset of 12 Qs that match this pattern. 2) We decided not to re-order the mansucript to discuss self-poisoning after establishing the monomer nucleus (even though we agree that doing so would improve the logical flow) because the interpretation of the data with respect to self-poisoning helps to establish critical strand lengths, and self-poisoning creates an anomaly in the DAmFRET data that is difficult to ignore. We add text clarifying that high local concentrations “effectively shifts the rate-limiting step to the growth of a higher order relatively-disordered species”.

Response to Reviewer 3

We thank the reviewer for their helpful comments.

We opted to retain Figures 1A and B because we think they are important for comprehending the subject and objectives of the study. We modified the former to attempt to make it more clear. We have also elaborated on DAmFRET as it is a relatively new approach that may be unfamiliar to many readers. Beyond this, we refer the reviewer and readers to our cited prior work describing the theory and interpretation of DAmFRET. Note that the y-axes of DAmFRET plots are not raw FRET but rather “AmFRET”, a ratio of FRET to total expression level. As explained thoroughly in our cited prior work, the discontinuity of AmFRET with expression level indicates that the high AmFRET-population formed via a disorder-to-order transition. When the query protein is predicted to be intrinsically disordered, the discontinuous transition to high AmFRET invariably (among hundreds of proteins tested in prior published and unpublished work) signifies amyloid formation as corroborated by SDD-AGE and tinctorial assays.

When performed using standard flow cytometry as in the present study, every AmFRET measurement corresponds to a cell-wide average, and hence does not directly inform on the distribution of the protein between different stoichiometric species. As there is only one fluorophore per protein molecule, monomeric nuclei have no signal. DAmFRET can distinguish cells expressing monomers from stable dimers from higher order oligomers (see e.g. Venkatesan et al. 2019), and we are therefore quite confident that AmFRET values of zero correspond to cells in which a vast majority of the respective protein is not in homo-oligomeric species (i.e. is monomeric or in hetero-complexes with endogenous proteins). The exact value of AmFRET, even for species with the same stoichiometry, will depend both on the effect of their respective geometries on the proximity of mEos3.1 fluorophores, and on the fraction of protein molecules in the species. Hence, we only attempt to interpret the plateau values of AmFRET (where the fraction of protein in an assembled state approaches unity) as directly informing on structure, as we did in Fig. S3A.

We believe that AmFRET decreases with longer polyQ because the mass fraction of fluorophore decreases in the aggregate, simply because the extra polypeptide takes up volume in the aggregate.

Yes, the fraction of positive cells in a discontinuous DAmFRET plot does increase with time. However, given the more laborious data collection and derivation of nucleation kinetics in a system with ongoing translation, especially across hundreds of experiments with other variables, ours is a snapshot measurement to approximately derive the relative contributions of intra- and intermolecular fluctuations to the nucleation barrier, rather than the barrier’s magnitude.

We have revised the tautological statement by removing “non-amyloid containing”.

Concerning the correlation of our data with the pathological length threshold -- as we state in the first results section, “Our data recapitulated the pathologic threshold -- Q lengths 35 and shorter lacked AmFRET, indicating a failure to aggregate or even appreciably oligomerize, while Q lengths 40 and longer did acquire AmFRET in a length and concentration-dependent manner”. Hence, most of our experiments were conducted with 60Q not because it resembles the pathological threshold, but rather because it was most convenient for DAmFRET experiments.

Self-poisoning is a widely observed and heavily studied phenomenon in polymer crystal physics, though it seems not yet to have entered the lexicon of amyloid biologists. We were new to this concept before it emerged as an extremely parsimonious explanation for our results. As described in the text, two pieces of evidence exclude the alternative mechanism suggested by the reviewer -- that non-structured oligomers form and subsequently engage and inhibit the template. Specifically, 1) inhibition occurs without any detectable FRET, even at high total protein concentration, indicating the species do not form in a concentration-dependent manner that would be expected of disordered oligomers; and 2) inhibition itself has strict sequence requirements that match those of Q zippers. Hence our data collectively suggest that inhibition is a consequence of the deposition of partially ordered molecules onto the templating surface.

We have softened the subheading and text of the relevant section in the discussion to more clearly indicate the speculative nature of our statements concerning the possible role of self-poisoned oligomers in toxicity.

We stand by our statement 'that kinetically arrested aggregates emerge from the same nucleating event responsible for amyloid formation', as this follows directly from self-poisoning.

Regarding the arguments for lateral and axial growth, we agree that the data are indirect. However, that polyQ forms lamellar amyloids both in vitro and in vivo is now established, so we do not feel it necessary to rigorously show that here. Nevertheless, we need to include this section primarily because it introduces the fact that ordering in polyQ amyloid occurs in the lateral as well as axial dimensions, and the onset of lateral ordering (lamellar growth) explains the very different behaviors of QU and QB sequences apparent on the DAmFRET plots. Ultimately, the two dimensions of growth are important to understand self-poisoning and maturation of the short nucleating zipper to amyloid.

References

Arseni D, Hasegawa M, Murzin AG, Kametani F, Arai M, Yoshida M, Ryskeldi-Falcon B. 2022. Structure of pathological TDP-43 filaments from ALS with FTLD. Nature 601:139–143. doi:10.1038/s41586-021-04199-3

Bansal A, Schmidt M, Rennegarbe M, Haupt C, Liberta F, Stecher S, Puscalau-Girtu I, Biedermann A, Fändrich M. 2021. AA amyloid fibrils from diseased tissue are structurally different from in vitro formed SAA fibrils. Nat Commun 12:1013. doi:10.1038/s41467-021-21129-z

Buell AK. 2017. The Nucleation of Protein Aggregates - From Crystals to Amyloid Fibrils. Int Rev Cell Mol Biol 329:187–226. doi:10.1016/bs.ircmb.2016.08.014

Chakraborty D, Straub JE, Thirumalai D. 2023. Energy landscapes of Aβ monomers are sculpted in accordance with Ostwald’s rule of stages. Sci Adv 9:eadd6921. doi:10.1126/sciadv.add6921 Crist B, Schultz JM. 2016. Polymer spherulites: A critical review. Prog Polym Sci 56:1–63. doi:10.1016/j.progpolymsci.2015.11.006

De Yoreo JJ. 2022. Casting a bright light on Ostwald’s rule of stages. Proc Natl Acad Sci USA 119. doi:10.1073/pnas.2121661119

Hong Y, Yuan S, Li Z, Ke Y, Nozaki K, Miyoshi T. 2015. Three-Dimensional Conformation of Folded Polymers in Single Crystals. Phys Rev Lett 115:168301. doi:10.1103/PhysRevLett.115.168301 Keller A. 1957. A note on single crystals in polymers: Evidence for a folded chain configuration. Philosophical Magazine 2:1171–1175. doi:10.1080/14786435708242746

Landgraf D, Okumus B, Chien P, Baker TA, Paulsson J. 2012. Segregation of molecules at cell division reveals native protein localization. Nat Methods 9:480–482. doi:10.1038/nmeth.1955

Lauritzen JI, Hoffman JD. 1960. Theory of Formation of Polymer Crystals with Folded Chains in Dilute Solution. J Res Natl Bur Stand A Phys Chem 64A:73–102. doi:10.6028/jres.064A.007

Navrotsky A. 2004. Energetic clues to pathways to biomineralization: precursors, clusters, and nanoparticles. Proc Natl Acad Sci USA 101:12096–12101. doi:10.1073/pnas.0404778101

Ohhashi Y, Ito K, Toyama BH, Weissman JS, Tanaka M. 2010. Differences in prion strain conformations result from non-native interactions in a nucleus. Nat Chem Biol 6:225–230. doi:10.1038/nchembio.306

Organ SJ, Ungar G, Keller A. 1989. Rate minimum in solution crystallization of long paraffins. Macromolecules 22:1995–2000. doi:10.1021/ma00194a078

Radamaker L, Baur J, Huhn S, Haupt C, Hegenbart U, Schönland S, Bansal A, Schmidt M, Fändrich M. 2021. Cryo-EM reveals structural breaks in a patient-derived amyloid fibril from systemic AL amyloidosis. Nat Commun 12:875. doi:10.1038/s41467-021-21126-2

Sahoo B, Singer D, Kodali R, Zuchner T, Wetzel R. 2014. Aggregation behavior of chemically synthesized, full-length huntingtin exon1. Biochemistry 53:3897–3907. doi:10.1021/bi500300c

Schmelzer JWP, Abyzov AS. 2017. How do crystals nucleate and grow: ostwald’s rule of stages and beyond In: Šesták J, Hubík P, Mareš JJ, editors. Thermal Physics and Thermal Analysis, Hot Topics in Thermal Analysis and Calorimetry. Cham: Springer International Publishing. pp. 195–211. doi:10.1007/978-3-319-45899-1_9

Schmidt M, Wiese S, Adak V, Engler J, Agarwal S, Fritz G, Westermark P, Zacharias M, Fändrich M. 2019. Cryo-EM structure of a transthyretin-derived amyloid fibril from a patient with hereditary ATTR amyloidosis. Nat Commun 10:5008. doi:10.1038/s41467-019-13038-z

Schweighauser M, Shi Y, Tarutani A, Kametani F, Murzin AG, Ghetti B, Matsubara T, Tomita T, Ando T, Hasegawa K, Murayama S, Yoshida M, Hasegawa M, Scheres SHW, Goedert M. 2020. Structures of α-synuclein filaments from multiple system atrophy. Nature 585:464–469. doi:10.1038/s41586-020-2317-6

Snapp EL, Hegde RS, Francolini M, Lombardo F, Colombo S, Pedrazzini E, Borgese N, Lippincott-Schwartz J. 2003. Formation of stacked ER cisternae by low affinity protein interactions. J Cell Biol 163:257–269. doi:10.1083/jcb.200306020

Törnquist M, Michaels TCT, Sanagavarapu K, Yang X, Meisl G, Cohen SIA, Knowles TPJ, Linse S. 2018. Secondary nucleation in amyloid formation. Chem Commun 54:8667–8684. doi:10.1039/c8cc02204f

Ungar G, Putra EGR, de Silva DSM, Shcherbina MA, Waddon AJ. 2005. The Effect of Self-Poisoning on Crystal Morphology and Growth Rates In: Allegra G, editor. Interphases and Mesophases in Polymer Crystallization I, Advances in Polymer Science. Berlin, Heidelberg: Springer Berlin Heidelberg. pp. 45–87. doi:10.1007/b107232

Vetri V, Foderà V. 2015. The route to protein aggregate superstructures: Particulates and amyloid-like spherulites. FEBS Lett 589:2448–2463. doi:10.1016/j.febslet.2015.07.006

Wild EJ, Boggio R, Langbehn D, Robertson N, Haider S, Miller JRC, Zetterberg H, Leavitt BR, Kuhn R, Tabrizi SJ, Macdonald D, Weiss A. 2015. Quantification of mutant huntingtin protein in cerebrospinal fluid from Huntington’s disease patients. The Journal of Clinical Investigation.

Yang Y, Arseni D, Zhang W, Huang M, Lövestam S, Schweighauser M, Kotecha A, Murzin AG, Peak-Chew SY, Macdonald J, Lavenir I, Garringer HJ, Gelpi E, Newell KL, Kovacs GG, Vidal R, Ghetti B, Ryskeldi-Falcon B, Scheres SHW, Goedert M. 2022. Cryo-EM structures of amyloid-β 42 filaments from human brains. Science 375:167–172. doi:10.1126/science.abm7285

Zhang X, Zhang W, Wagener KB, Boz E, Alamo RG. 2018. Effect of Self-Poisoning on Crystallization Kinetics of Dimorphic Precision Polyethylenes with Bromine. Macromolecules 51:1386–1397. doi:10.1021/acs.macromol.7b02745

Author Response

The following is the authors’ response to the original reviews.

Reviewer #1 (Public Review):

The current manuscript provides a timely contribution to the ongoing discussion about the mechanism of the apical sodium/bile acid transporter (ASBT) transporters. Recent structures of the mammalian ASBT transporters exhibited a substrate binding mode with few interactions with the core domain (classically associated with substrate binding), prompting an unusual proposal for the transport mechanism. Early structures of ASBT homologues from bacteria also exhibit unusual substrate binding in which the core substrate binding domain is less engaged than expected. Due to the ongoing questions of how substrate binding and mechanism are linked in these transporters, the authors set out to deepen our understanding of a model ABST homolog from bacteria N. meningitidis (ABST-NM).

The premise of the current paper is that the bacterial ASBT homologs are probably not physiological bile acid transporters, and that structural elucidation of a natively transported substrate might provide better mechanistic information. In the current manuscript, the authors revisit the first BASS homologue to be structurally characterized, ABST-NM. Based on bacteriological assays in the literature, the authors identify the coenzyme A precursor pantoate as a more likely substrate for ABSTNM than taurocholate, the substrate in the original structure. A structure of ASBT-NM with pantoate exhibits interesting differences in structure. The structures are complemented with MD simulations, and the authors propose that the structures are consistent with a classical elevator transport mechanism.

The structural experiments are generally solid, although showing omit maps would bolster the identification of the substrate binding site.

We have added an omit map in Fig S2.

One shortcoming is that, although pantoate binding is observed, the authors do not show transport of this substrate, undercutting the argument that the pantoate structure represents binding of a "better" or more native substrate. Mechanistic proposals, like the proposed role of T112 in unlocking the transporter, would be much better supported by transport data.

In the absence of being able to source radiolabelled pantoate at a reasonable cost, we decided to focus on binding studies, relying on the fact that pantoate/pyruvate uptake has been shown in other BASS transporters. While we agree that transport needs to be substantiated, our crystallographic and molecular dynamics studies combined provide a picture of sodium ions stabilising the substrate binding site to enable the binding of the substrate, which in turn induces further conformational changes. Such changes would be consistent with a mechanism of sodium driven transport with clear coupling of the sodium ions to substrate translocation. We are not saying this is a “better” substrate but rather that a substrate binding like this would be able to elicit the conformational changes necessary for transport – something that has been missing from previous studies.

Reviewer #2 (Public Review):

The manuscript starts with a demonstration of pantoate binding to ASBTnm using a thermostability assay and ITC, and follows with structure determinations of ASBTnm with or without pantoate. The structure of ASBTnm in the presence of pantoate pinpoints the binding site of pantoate to the "crossover" region formed by partially unwinded helices TMs 4 and 9. Binding of pantoate induces modest movements of side chain and backbone atoms at the crossover region that are consistent with providing coordination of the substrate. The structures also show movement of TM1 that opens the substrate binding site to the cytosol and mobility of loops between the TMs. MD simulations of the ASBT structure embedded in lipid bilayer suggests a stabilizing effect of the two sodium ions that are known to co-transport with the substrate. Binding study on pantoate analogs further demonstrates the specificity of pantoate as a substrate.

The weakness of the manuscript includes a lack of transport assay for pantoate and a lack of demonstration that the observed conformational changes in TM1 and the loops are relevant to the binding or transport of pantoate.

We agree that the manuscript would have been bolstered by transport data (see response to reviewer 1). The take-home message from the movement of TM1 and the loops is that they are flexible. It is probably unlikely that TM1 moves like this during the transport cycle and we have avoided overplaying the significance of this movement. Instead, we have focussed on the conformational changes in the pantoate binding site. We have made an additional movie concentrating on the binding site and not including TM1.

Overall, the structural, functional and computational studies are solid and rigorous, and the conclusions are well justified. In addition, the authors discussed the significance of the current study in a broader perspective relevant to recent structures of mammalian BASS members.

Reviewer #3 (Public Review)

The manuscript describes new ligand-bound structures within the larger bile acid sodium symporter family (BASS). This is the primary advance in the manuscript, together with molecular simulations describing how sodium and the bile acids sit in the structure when thermalized. What I think is fairly clear is that the ligands are more stable when the sodiums are present, with a marked reduction in RMSD over the course of repeated trajectories. This would be consistent with a transport model where sodium ions bind first, and then the bile acid binds, followed by a conformational change to another state where the ligands unbind.

While the authors mention that BASS transporters are thought to undergo an elevator transport mechanisms, this is not tested here. In my reading, all the crystal structures describe the same conformational state, and the simulations do not make an attempt to induce a transition on accessible simulation timescales. Instead, there is a morph between two states where different substrates are bound, which induces a conformational change that looks unrelated to the transport cycle.

To make our conclusions clearer we have added another movie showing a morph between the structure without substrate (instead of using the structure with taurocholate, which we were using as a representative of the unbound structure) and that with pantoate and have omitted the panel domain including TM1. While both of these structures are inward-facing, there are significant conformational changes within TM4 that we have described in the article.

Instead, the focus is on what kinds of substrates bind to this transporter, interrogating this with isothermal calorimetry together with mutations. With a Kd in the micromolar range, even the best binder, pantoate, actually isn't a particularly tight binder in the pharmaceutical sense. For a transporter, tight binding is not actually desirable, since the substrate needs to be able to leave after conformational change places it in a position accessible to the other side.

As the referee points out the Kd that we observe would be consistent with those for substrates of other transporters.

There is one really important point that readers and authors should be aware of. In Figure 2A, the names are not consistent with the chemical structure. "-ate" denotes when a carboxylic acid is in the deprotonated form, creating a charged carboxylate. What is drawn is pantoic acid, ketopantoic acid, and pantoethenic acid. Less importantly, the wedges and hashes for the methyl group are arguably not appropriate, since the carbon they are attached to is not a chiral center. For the crystallization, this makes no difference, since under near-neutral pKas the carboxylic acid will spontaneously deprotonate, and the carboxylate form will be the most common. However, if the structures in Figure 2A were used for classical molecular simulation, that would be a big problem, since now that would be modeling the much rarer neutral form rather than the charged state. I am reasonably sure based on Figure 5 that the MD correctly modeled the deprotonated form with a carboxylate, but that is inconsistent with Figure 2A. Otherwise, the structure and simulation analysis falls into the mainstream of modern structural biology work.

We have corrected the inconsistency of the protonaNon state in the naming of the molecular structures. Thank you for poinNng this out – though the names represented the predominant form in soluNon, the more aestheNcally pleasing protonated form got the beOer of us in our representaNons. The correct form was used in the MD.

Reviewer #1 (Recommendations For The Authors):

1) Omit maps (Fo-Fc) should be shown for pantoate and for the sodiums in the structure.

This has been added to supplementary Figure 2.

2) Line 86 - could you briefly describe the alternative mechanism proposed for the mammalian NTPCs?

We have added an extra line to describe this deviation from the classical alternating access model.

3) Line 124 - where is the lipid like molecule, and does it interact with either the kinked helix or the substrate? A supplemental figure would be helpful.

The lipid like molecule lies between the substrate and the kinked helix, but doesn’t interact strongly with either. It would appear that the lipid would bind in the crevice rather than causing the crevice. We add Author response image 1 here but have not added it to the supplementary figures. The maps and PDB file are available for download.

Author response image 1.

The 2mFo-DFc density is at 1σ, the mFo-DFc density is at 2.5σ.

4) I notice that the apo and pantoate structures are crystallized in different space groups. How does this compare to the original TCH structure? Is there any chance that crystal packing is altering the TM1 geometry or loop 1?

We cannot rule out the effect of the crystallisation conditions on the movement of the TM1. We have now solved a number of different structures of ASBTNM and this is the first time we observe TM1 in this conformation. As stated above we have refrained from overplaying the significance of the movement of TM1 to transport, other than to say that some adjustments need to be made to accommodate the pantoate.

Reviewer #2 (Recommendations For The Authors):

Minor comments:

Pg 3, "... with a 5-fold inverted repeat...", Should be 2-fold?

Changed, thank you.

Reviewer #3 (Recommendations For The Authors):

Is there any chance that the MD simulations (even in a reduced form) could be uploaded to Zenodo or a similar repository?

We have taken up this suggestion and added the information in the paper: MD trajectories in the GROMACS XTC format were deposited in the OSF.io repository under DOI 10.17605/OSF.IO/KFDT5 under the open CC-BY Attribution 4.0 International license. The trajectories contain all atoms and were subsampled at 5-ns intervals. GROMACS run input files (TPR format) and initial coordinate files (GRO format) together with topology files (GROMACS format) are also included.

Watch the "Å" symbol in Figures 5, S6, S7. This looks like they were made in matplotlib, and probably used something like: "$\AA$", which puts the symbol in math mode. This makes the Å symbol in italics. Matplotlib has gotten better UTF-8 support

Changed, thank you.

Your citation for LINCS duplicates the citation for PME. I think you want the Hess 1998 paper. 10.1002/(SICI)1096-987X(199709)18%3A12<1463%3A%3AAID-JCC4>3.0.CO%3B2-H

Changed, thank you

Author Response

The following is the authors’ response to the original reviews.

Reviewer #1 (Public Review):

The authors performed a meta-analysis of GC concentrations and metabolic rates in birds and mammals. They found close associations for all studies showing a positive association between these two traits. As GCs have been viewed with close links to "stress," authors suggest that this overlooks the importance of metabolism and perhaps GC variation does not relate to "stress" per se but an increase in metabolism instead.

This is an important meta-analysis, as most researchers acknowledge the link between GCs and metabolism, metabolism is often overlooked in studies. The field of conservation physiology is especially focused on GCs being a "stress" hormone, which overlooks the importance of GCs in mediating energy balance, i.e., an animal that has high GC concentrations may not be doing that poorly compared to an animal with low GC concentrations, it might just be expending more energy, e.g., caring for young. The results, with overwhelming directionality and strong effect sizes, support the link for a positive association with these two variables.

My main concern lies in that most of the studies come from a few labs, therefore there may be limited data to test this relationship. I would include lab as a random effect to see how strong this effect might be.

We think this is a good point, and we ran the main models included in the manuscript including Lab as random effect (N= 35 experiments, 21 studies, 16 labs). This did not affect the results, leading to negligible changes in the model parameters (alternative model tables are shown in Author response table 1 and 2). In the revised version of the manuscript we mention that we tested the effect of Lab but did not keep this variable in the models (lines 183-185)

Author response table 1.

Meta regression model testing the association between metabolic rate (MR) effect sizes and glucocorticoid effect sizes.

Author response table 2.

Meta regression model (quantitative approach) testing the effect of (a) Taxa, (b) Before / after effect, (c) Experiment / control effect, (d) Use of Metabolic Rate or Heart Rate as metabolic variable and (e) Treatment type, on the association between metabolic rate (MR) and glucocorticoid effect sizes across studies.

Furthermore, I would like to see a test of the directionality of the two variables. Authors suggest that changes in metabolism affect GC levels but likely changes in GC levels would affect metabolism. Why not look into studies that have altered GC levels experimentally and see the effect on metabolism? Based on the close link, authors suggest that GCs may not play a role outside of "stress" beyond the stressor's effect on metabolic rate. However, if they were to investigate manipulations of GCs on metabolic rate, the link may or may not be there, which would be interesting to look at. I firmly believe that GCs are tightly linked to metabolism; however, I also think that GCs have a range of effects outside of metabolism as well, depending on the course and strength of the stressor.

The directionality of the two variables is indeed a question of interest – we show that changes in metabolic rate affect GCs, but does the reverse also happen? In the schematic model we propose in Box 1, we propose that the effect is uni-directional, i.e. metabolic rate affects GC-levels, but GCs have no direct effect on metabolic rate. We note that there may however be an indirect effect, in that in the absence of a GC-response to an increase in metabolic rate the organism would after some time no longer be able to fuel the metabolic rate. Because we anticipate that more readers may raise this question, we have added the following paragraph to the discussion:

“We selected studies in which experimental treatments affected MR, leading us to conclude that the most parsimonious explanation of our finding is that GC levels were causally related to MR. Suppose however that instead we reported a correlation between MR and GCs, using for example unmanipulated individuals. The question would then be justified whether changes in GCs affected MR or vice versa. Direct effects of GCs could be studied using pharmacological manipulations. However, while many studies show that GC administration induces a cascade of effects, when the function of GCs is to facilitate a level of MR, as opposed to regulate variation in MR, we do not anticipate such manipulations to induce an increase in MR (Box 1). On the other hand, when MR is experimentally increased in conjunction with pharmacological manipulations that supress the expected GC-increase (an experiment that to our best knowledge has not yet been done), we would predict that the increase in MR can be maintained less well compared to the same MR treatment in the absence of the pharmaceutical manipulation. This result, we would interpret to demonstrate that maintaining a particular level of MR may be dependent on GCs as facilitator, but it would be misleading to interpret this pattern to indicate that GCs regulate MR, as is sometimes proposed. Additionally, it would be informative to investigate whether energy turnover immediately before blood sampling is a predictor of GC levels, as we would predict on the basis of the interpretation of our findings. Increasing the use of devices and techniques that monitor energy expenditure or its proxies (e.g. accelerometers) may be a way to increase our understanding of the generality of the GC-MR association. “

We based our hypotheses and searching criteria on the assumption that GCs induce physiological processes to help the organism facilitate energetic demands. Pharmacologically induced increases in GCs would lead to physiological responses and associations that we consider not comparable to the ones reported in this work, as we base our hypotheses on natural (i.e. non pharmacologically induced) GC and MR variation. This said, with exogenous GC administration, we may expect GC cascade effects, but not necessarily an increase in MR. Here - and acknowledging that the link between GCs and metabolic rate may entail complex steps - we predict that GC administration may lead to an increase in blood glucose and may affect energy allocation at a tissue-specific level. However, such increase may have no effect on whole-organism energy expenditure, unless energy expenditure is limited by glucose availability. We however acknowledge that it would be interesting to investigate the kind of associations between MR, GCs and other physiological variables (e.g. glucose) that appear when inducing an increase in GCs, as these would broaden our understanding of the mechanistic processes underlying these associations.

We show that variation in GC levels was explained by variation in MR, independent of the stimulus that caused the increase in MR. We propose that the most parsimonious interpretation of our findings is that GC variation is an indicator of variation in MR, independent of the cause of variation in MR. We do not intend to prove causality when making predictions on the co-dependency of metabolic rate and GCs. In fact, our predictions do not imply that one trait necessarily affects the other per se, as these interplay is likely to be shaped by the environmental or physiological context (Box 1). Thus, the specific mechanisms underlying how changes in metabolic rate induce changes in GCs - or the other way around - need to be investigated. One step to tackle this in upcoming research would indeed be studying the effects of exogenous GCs on metabolic rate.

In the manuscript, we clarify that GCs have a variety of cascade effects besides metabolism (Box 1). On the basis of our results, however, we suggest that many of the downstream effects of GCs may be interpreted as allocation adjustments to the metabolic level at which organisms operate (lines 235236), but we do acknowledge that these cascade effects are complex and affects many systems besides metabolism.

This work helps in the thinking that GCs are not the same as a "stress" hormone or labelling hormones with only one function. As hormones are naturally pleiotropic, the view of any one hormone being X is overly simplistic.

We fully agree, but stress that we focus on how GCs are regulated, which may be less complex than its pleiotropic functions. Indeed, we consider that the many functions of GCs have potentially clouded the question as to how GCs are regulated.

Reviewer #2 (Public Review):

Where this study is interesting is that the authors do a meta-analysis of studies in which metabolic rate was experimentally manipulated and both this rate and glucocorticoid levels were simultaneously measured. Unsurprisingly, there are relatively few such studies and many are from the lab of Michael Romero. While the results of the analysis are compelling, they are not surprising. That said, this work is important.

It is worth noting that in this analysis, the majority of the studies, if not all, are dealing with variation in baseline levels of glucocorticoids. That means the hormone is mostly acting metabolically at these lower levels and not as a stress response hormone as it does when levels are much higher. This difference is probably due to differences in receptors being activated. This could be discussed.

As mentioned in Box 1, within our hypothesis framework we make no distinction between baseline and stress-induced GC-levels, and thereby in effect assume these to be points in a continuum from a metabolic perspective. Our results support this view, as our sample includes baseline- and stressinduced –range GC values, and these are not distinguishable (Fig. 3). We do however recognize that we did not return to this issue in the Discussion, while the same issue may well occur to many readers familiar with the literature. We therefore added the following paragraph to the discussion:

“ Note that in the context of our analysis we made no distinction between ‘baseline’ and ‘stressinduced GC-levels (Box 1). Firstly, because these concepts are not operationally well defined – baseline GC-levels are usually no better defined than ‘not stress-induced’. Secondly, when considering the facilitation of metabolic rate as primary driver of GC regulation, there does not appear a need to invoke different classes of GC-levels instead of the more parsimonious treatment as continuum. This is not to say that this also applies to the functional consequences of GC-level variation: it is well known that receptor types differ in sensitivity to GCs (Landys et al. 2006; Sapolsky et al. 2000; Romero 2004), thereby potentially generating step functions in the response to an increase in GC-levels.”

We note further that to our best knowledge there are no standard or established thresholds that allow us to separate GC levels into “baseline” and “stress-induced”, and in any case these concentration ranges differ strongly among species and experimental set-ups (e.g. captive vs. free-living individuals). Consequently, many of the studies included in our work report what would typically be interpreted as “stress-induced” levels, and thus within the range of those reported by standardized stress protocols (e.g. levels above 20-30 ng/ml for corticosterone in bird species, Cohen et al. 2007, Jimeno et al. 2018; levels between 150-300 ng/ml in captive rats, Buwalda et al. 2012, Beerling et al. 2011; levels 2-10 times above baseline in humans, Sramek et al. 1999). We also want to note that we work with effect sizes, i.e. not GC levels, and that GC measurement units differ among studies. Mean GC values by study in the original units are shown in Table S3.

Reviewer #1 (Recommendations For The Authors):

L26: why is the causality in this direction? Not that I don't think that metabolic rate drives GC variation but the meta-analyses here could suggest the opposite direction as well? That GC phenotype could limit or promote metabolic activity? (In terms of the natural variation studies and not the experimental ones)

See our detailed response above, on the directionality of the association and the hypotheses underlying our searching criteria and the paragraph on this topic added to the discussion.

L27: again, I am not sure the meta-analyses can lead to this question. Although there is a tight link between GC and metabolic rate, there is still variation around that is unexplained.

See our detailed response above, on the directionality of the association and the hypotheses underlying our searching criteria and the paragraph on this topic added to the discussion.

L45: I think there is plenty of literature in the field that would say that GCs are linked to metabolism and don't define GCs as synonymous with stress. See MacDougall and others that you cite later in the paragraph: "GCs and stress are not synonymous." I think maybe shifting the strong language at the beginning might help with your argument later on.

We do not disagree, but two considerations made us retain the ‘strong language’. Firstly, while many authors mention links between GCs and metabolic rate, as we read the literature, the quantitative importance of this link to understand GC variation is underestimated in our view. Secondly, the literature is rife with articles that clearly do not consider metabolic rate variation as a driver of the GC variation they observe.

Box 1: on the diagram the link between GCs and learning is problematic as there are plenty of studies that show a negative effect on learning with GC exposure. It usually depends on the time course of GCs and learning outcomes.

We agree with the referee´s point. Learning was deleted from the diagram to avoid confusion.

The diagram also suggests that GCs in the blood decreases insulin. For Aves that are rather insulin insensitive, the evidence that GCs affect insulin concentrations are very limited, even in the poultry literature.

Indeed, and we now mention in box 1 that GC effects on insulin are primarily found in mammals, and less so in birds.

Box 1 at the end also makes a point about GCs having complex downstream effects at baseline and stressinduced levels, besides energy mobilization but the abstract seems to indicate that there are limited effects of GCs outside of metabolism. Hence why I also advocate being careful about the wording in the abstract.

The related abstract sentence has been rewritten to avoid this inconsistency (lines 17-18)

L107: "being or not significant" meaning significant or not? The wording is awkward

We reworded the sentence for clarity. We included studies reporting both significant and nonsignificant increases in metabolic rate.

L110: why not look at whether experimental increases in GCs also induce increases in metabolic rate, i.e., the directionality of the two variables. (point 2)

See our detailed response above, on the directionality of the association and the hypotheses underlying our searching criteria and the paragraph on this topic added to the discussion.

The studies, although there are ~30, are overlapping in terms of labs, i.e., a lot of them came from the same lab. Did you think to include lab as a random effect to see if there are effects of one or two labs doing work that strengthened the results?

We think this is a good point, and we ran the main models included in the manuscript including Lab as random effect (N= 35 experiments, 21 studies, 16 labs). Including Lab as random factor did not affect the results, leading to negligible changes in the model parameters. We provide tables with the model results in our previous response. In the text we now mention that we tested the effect of Lab but did not keep this variable in the models (lines 183-185)

L314: I think it depends on the time course and intensity of the stressor. I firmly believe that outside of metabolic demands, high levels of GCs chronically or the inability to mount a proper stress response is indicative of pathology or something outside of metabolism.

Whether the association between GCs and MR holds under a context of ‘chronic stress’ (i.e. understood as chronically elevated GCs) remains to be tested. We note, however, that chronically high levels of metabolic rate may potentially have pathological effects.

Reviewer #2 (Recommendations For The Authors):

I find the title a bit misleading. The conclusion from the study is that glucocorticoid levels can reflect metabolic rate, not that glucocorticoid levels do not indicate stress. Remember, stress can certainly affect metabolic rate.

We see the point but note that other drivers of variation in metabolic rate also increase GCs, as we show in our analysis, and hence we propose that GC variation always indicate variation metabolic rate, and only stress when stress is the cause of the increase in metabolic rate.

Author Response

The following is the authors’ response to the original reviews.

We are very grateful to the reviewers for their insightful and detailed analysis of our work, in particular to reviewer 2. We also would like to thank the Elife editorial team for organizing this form of public review and debate, which we believe will be of interest to the science community.

Reviewer #1 (Public Review):

Despite durable viral suppression by antiretroviral therapy (ART), HIV-1 persists in cellular reservoirs in vivo. The viral reservoir in circulating memory T cells has been well characterized, in part due to the ability to safely obtain blood via peripheral phlebotomy from people living with HIV-1 infection (PWH). Tissue reservoirs in PWH are more difficult to sample and are less well understood. Sun and colleagues describe isolation and genetic characterization of HIV-1 reservoirs from a variety of tissues including the central nervous system (CNS) obtained from three recently deceased individuals at autopsy. They identified clonally expanded proviruses in the CNS in all three individuals.

Strengths of the work include the study of human tissues that are under-studied and difficult to access, and the sophisticated near-full length sequencing technique that allows for inferences about genetic intactness and clonality of proviruses. The small sample size (n=3) is a drawback. Furthermore, two individuals were on ART for just one year at the time of autopsy and had T cells compatible with AIDS, and one of these individuals had a low-level detectable viral load (Figure S1). This makes generalizability of these results to PWH who have been on ART for years or decades and have achieved durable viral suppression and immune reconstitution difficult.

While anatomic tissue compartment and CNS region accompany these PCR results, it is unclear which cell types these viruses persist in. As the authors point out, it is possible that these reservoir cells might have been infiltrating T cells from blood present at the time of autopsy tissue sampling. Cell type identification would greatly enhance the impact of this work. Several other groups have undergone similar studies (with similar results) using autopsy samples (links below). These studies included more individuals, but did not make use of the near-full length sequencing described here. In particular, the Last Gift cohort, based at UCSD and led by Sara Gianella and Davey Smith, has established protocols for tissue sampling during autopsy performed soon after death. https://pubmed.ncbi.nlm.nih.gov/35867351/ https://pubmed.ncbi.nlm.nih.gov/37184401/

We agree with reviewer 1 that studies to identify specific cell types that harbor intact HIV-1 in individual tissue compartments would be very informative; our group has recently initiated such studies.

Overall, this small, thoughtful study contributes to our understanding of the tissue distribution of persistent HIV-1, and informs the ongoing search for viral eradication.

We thank reviewer 1 for these encouraging remarks.

Reviewer #2 (Public Review):

The manuscript by Sun et al. applies the powerful technology of profiling viral DNA sequences in numerous anatomical sites in autopsy samples from participants who maintained their antiviral therapy up to the time of death. The sequencing is of high quality in using end-point dilution PCR to generate individual viral genomes. There is a thoughtful discussion, although there are points that we disagree with. This is an important data set that increases the scope of how the field thinks about the latent reservoir with a new look at the potential of a reservoir within the CNS.

We greatly appreciate the comments by reviewer 2 and would like to thank them for their detailed and very knowledgeable analysis of this paper.

1) The participants are very different in their exposure to HIV replication and disease progression. Participant 1 appears to have been on ART for most of the time after diagnosis of infection (16 years) and died with a high CD4 T cell count. The other two participants had only one year on ART and died with relatively low CD4 T cell counts (under 200). This could lead to differences in the nature of the reservoir. In this regard, the amount of DNA per million cells appears to be about 10-fold lower across the compartments sampled for participant 1. Also, one might expect fewer intact proviruses surviving after 16 years on ART compared to only 1 year on ART. The depth of sampling may be too limited and the number of participants too few to assess if these differences are features of these participants because of their different exposures to HIV replication. On the positive side, finding similarities across these big differences in participant profiles does reinforce the generalizability of the observations.

Many thanks for pointing this out. We also noticed that the total number of HIV-1 proviruses is smaller in our study participant 1 (who had been on ART for 16 years), compared to study persons 2 and 3 with more limited treatment durations (1-2 years), however, due to the small number of study persons, we think we cannot use these results for inferring how treatment duration influences viral reservoir size in tissues.

2) The following analysis will be limited by sampling depth but where possible it would be interesting to compare the ratio of intact to defective DNA. A sanctuary might allow greater persistence of cells with intact viral DNA even without viral replication (i.e. reduced immune surveillance). Detecting one or two intact proviruses in a tissue sample does not lend itself to a level of precision to address this question, but statistical tests could be applied to infer when there is sampling of 5 or more intact proviruses to determine if their frequency as a ratio of total DNA in different anatomical sites is similar or different. This would allow adjustment for the different amount of viral DNA in different compartments while addressing the question of the frequency of intact versus defective proviruses. One complication in this analysis is if there was clonal expansion of a cell with an intact genome which would represent a fortuitous overrepresentation intact genomes in that compartment.

We have performed the analysis suggested by reviewer 2 and included a diagram reflecting the ratio of intact/defective proviruses as a new supplemental figure (Figure S2). Unfortunately, we do not feel comfortable to draw any real conclusions from this additional analysis; the sample sizes are simply too limited.

3) The key point of this work is that the participants were on therapy up to the time of death ("enforcing" viral latency). The predominance of defective genomes is consistent with this assumption. Is there data from untreated infections to compare to as a signature of whether the viral DNA population was under selective pressure from therapy or not? Presumably untreated infections contain more intact DNA relative to total DNA. This would represent independent evidence that therapy was in place.

We agree that an analysis of autopsy samples from untreated persons living with HIV-1 would be of great interest, and are actively collaborating with neuropathologists from multiple sites to obtain such samples. Yet, we are not convinced that selection pressure on reservoir cells during ART can be appropriately identified through quantitative virological assays. Rather, we feel that the selection of proviruses can be best assessed when qualitative parameters, including proviral integration sites and their position relative to host epigenetic chromatin features, are evaluated.

4) There are several points in Figure 5 to raise about V3 loop sequences. The analysis includes a large number of "undetermined" sequences that did not have a V3 loop sequence to evaluate. We would argue it is a fair assumption that the deleted proviruses have the same distribution of X4 and R5 sequences as the ones that have a V3 sequence to evaluate. In this view it would be possible to exclude the sequences for which there is no data and just look at the ratio of X4 and R5 in the different compartments, specifically does this ratio change in a statistically significant way in different compartments? The authors use "CCR5 and non-CCR5" as the two entry phenotypes. The evidence is pretty strong that the "other" coreceptor the virus routinely uses is CXCR4, and G2P is providing the FPR for X4 viruses. Perhaps the authors are trying to create some space for other coreceptors on microglia, but we are pretty sure what they are measuring is X4 viruses, especially in this late disease state of participant 2. Finally, we have previously observed that the G2P FPR score of <2 is a strong indicator of being X4, FPR scores between 2 and 10 have a 50% chance of being X4, and FPR scores above 10 are reliably R5 (PMID27226378). In addition, we observed that X4 viruses form distinct phylogenetic lineages. The authors might consider these features of X4 viruses in the evaluation of their sequences. Specifically, it would be helpful to incorporate the FPR scores of the reported X4 viruses.

Many thanks for these thoughts. We have now included FPR scores for all sequences and considered sequences with FPR score <2 as X4-tropic. Among 497 proviral sequences derived from all three participants, only 14 proviral sequences had FPR scores between 2 and 10 and their tropism was classified as CCR5 in the new Figure 5. We agree that viral tropism analysis of proviral sequences from the CNS would be of particular interest for study subject 2; however, most brain-derived sequences from that person had large deletions in the env region, precluding an analysis of viral tropism.

5) We have puzzled over the many reports of different cell types in the CNS being infected. When we examined these cell types (both as primary cells and as iPSC-derived cells), all cells could be infected with a version of HIV that had the promiscuous VSV-G protein on the virus surface as a pseudotype. However, only macrophages and microglia could be infected using the HIV Env protein, and then only if it was the M-tropic version and not the T-tropic version (PMID35975998). RNAseq analysis was consistent with this biological readout in that only macrophages and microglia expressed CD4, neurons and astrocytes do not. From the virology point of view, astrocytes are no more infectable than neurons.

We appreciate these comments. As described in our discussion, we agree that the role of astrocytes as target cells for HIV-1 infection is highly controversial; we look forward to future opportunities to evaluate HIV sequences in sorted astrocytes from autopsy tissues.

6) The brain gets exposed to virus from the earliest stages of infection but this is not synonymous with viral replication. Most of the time there is virus in the CSF but it is present at 1-10% of the level of viral load in the blood and phylogenetically it looks like the virus in the blood, most consistent with trafficking T cells, some of which are infected (PMID25811757). The fact that the virus in the blood is almost always T cell-tropic in needing a high density of CD4 for entry makes it unlikely that monocytes are infected (with their low density of CD4) and thus are not the source of virus found in the CNS. It seems much more likely that infected T cells are the "Trojan Horse" carrying virus into the CNS.

We appreciate the reviewer’s referral to Greek mythology and agree that the hypothesis of infected T cells acting as “Trojan horses” is more intuitive and better supported by available data. We have adjusted our discussion accordingly.

7) While all participants were taking antiretroviral therapy at the time of their death, they were not all suppressed when the tissues were collected. The authors are careful not to mention "suppressive ART" in the text, which is appreciated. However, the title should be changed to also reflect this fact.

Thanks for pointing this out. From our perspective, ART is never fully suppressive, as low-level viremia (below the detection threshold of commercial PCR assays) is detectable in almost all ART-treated persons. As such, it is not clear to us that “suppressive” necessarily implies suppression below the detection limits of commercial PCRs. We argue that ART can also be suppressive when plasma viral loads are in the range of 100 copies/ml, as they are in our study subject 3. Nevertheless, we have changed the title to avoid confusion.

Reviewer #1 (Recommendations For The Authors):

I encourage the authors to compare their autopsy and tissue sampling procedures to those used by The Last Gift researchers and consider including references to this ongoing study. If the authors plan to continue in this line of research, the field would greatly benefit from a collaboration that would bring together their excellent and advanced PCR technique with the larger sample size offered by The Last Gift. Lastly, is there some way to simultaneously determine cell type when NFL sequencing is performed?

We look forward to collaborating with investigators from the Last Gift Cohort in the future and have integrated additional references in the manuscript to acknowledge their work. At the current stage of technology development, we think that sorting of infected cells based on canonical markers of defined cell populations is the preferred approach for identifying phenotypic properties of infected cells; however, expansion of the PheP-Seq assay (Sun et al., Nature 2023), may facilitate this process in the future.

Reviewer #2 (Recommendations For The Authors):

1) The authors have chosen to lump all R5 viruses together in terms of their entry phenotype, giving all viruses an equal chance of infecting all potentially susceptible cell types. This ignores the fact that normal HIV is selected to infect cells, requiring a high density of CD4 as is found on T cells. We use the term R5 T cell-tropic to describe "normal" HIV. The ability to efficiently enter cells that have a low density of CD4, such as macrophages and microglia, involves the evolution of a distinct phenotype, termed macrophage tropism (PMID24307580, and work of others). This happens most often in the CNS where T cells are infrequent thus potentiating evolution to infect an alternative cell type. This change in entry phenotype is dramatic and, like X4 viruses, results in phylogentically distinct lineages (PMID22007152). There are no sequence signatures for M-tropic viruses as there are for X4 viruses, but the fact that there are sequences shared between the CNS and lymphoid tissue makes it much more likely that there are T cells migrating around the body, including into the CNS, that are carrying R5 T cell-tropic virus with them, with the cells potentially clonally expanding in situ in the CNS. The persistence of a potential CNS T cell reservoir was the point we were trying to make in our recent paper (ref. 38), not only that these CSF rebound viruses were R5 viruses but they were selected for replication in T cells as seen by their dependence of a high density of CD4 for entry. This is the conclusion one would reach if clonally expanded viral sequences were shared between two lymphoid compartments. It is not necessary to ascribe properties of infection and clonal amplification to microglia cells when a more parsimonious explanation is that there are low levels of T cells in the CNS, especially in the absence of entry phenotype data showing these sequences encode an M-tropic entry phenotype. As is the authors are just adding to the unproven belief that virus in the CNS must be in myeloid cells, which in this case in particular we suspect is the wrong interpretation.

We are impressed by reviewer 2’s recent work, suggesting the viral reservoir in the CNS may primarily consist of clonally-expanded R5 T-cell tropic viruses. We have adjusted our discussion to emphasize this possibility, and to highlight that viral entry phenotyping data will be informative for better understanding viral persistence in the brain.

2) The authors noted that the frequency of intact proviruses is highest in the lymph nodes of 2/2 participants for which they had lymph node samples, relative to the other tissues examined. They thus conclude, "Together, these results indicate that intact HIV-1 proviruses are preferentially detected in lymphoid and gastrointestinal (GI) tissues." However, an examination of Figure 2 reveals that the total HIV copy number is highest in the lymph nodes of these two people. Thus, it doesn't seem like HIV is preferentially intact in the lymph nodes as much as they sampled more provirus from that tissue and therefore were able to detect more intact proviruses.

We have adjusted our manuscript to indicate that the highest numbers of intact HIV-1 proviruses were present in lymph nodes, both in terms of absolute numbers and after normalization to the total numbers of cells analyzed.

3) In Figure 1A, the legend should be changed so that "PMSC" is spelled out as "premature stop codon" for ease of reading. This is done for Figure 1B.

We have corrected this issue as suggested by the reviewer.

4) The pie charts in Figure 5 could be better labeled for ease of interpreting. In Figure 5C, instead of just labeling it as "P2" it could be "Distribution of CXCR4-using proviruses, P2", as an example. As it stands, it is hard to know what the figure is describing without reading the text.

We have changed this accordingly.

5) While all participants were taking antiretroviral therapy at the time of their death, they were not all suppressed when the tissues were collected. The authors are careful not to mention "suppressive ART" in the text, which is appreciated. However, the title should be changed to also reflect this fact.

Thanks for pointing this out. From our perspective, ART is never fully suppressive, as low-level viremia (below the detection threshold of commercial PCR assays) is detectable in almost all ART-treated persons. As such, it is not clear to us that “suppressive” necessarily implies suppression below the detection limits of commercial PCRs. We argue that ART can also be suppressive when plasma viral loads are in the range of 100 copies/ml. Nevertheless, we have changed the title to avoid confusion.

Editorial comments:

In addition to the reviewers suggestion, we feel that adding more information on how you define intact proviral sequence, e.g. are only disrupted essential genes or also in accessory genes considered? Previous studies have shown that brain-derived HIV-1 strains are usually CCR5-tropic, show high affinity for the CD4 receptor and frequently contain defective vpu genes. Some information and discussion if the brainderived sequences confirm these previous finding seems of significant interest.

As described in our previous work (e. g. Lee et al, JCI 2017; Jiang et al, Nature 2020), accessory genes are not considered in our definition of “genome intactness”; this is consistent with approaches other investigators have chosen (e. g. Hiener et al, Cell Reports 2017). Within the genome intact sequences we identified in the CNS in our study persons, we found no evidence for deletions of vpu sequences; this has been emphasized in the revised manuscript.

Author Response

We thank the reviewers and editors for their deep, thoughtful and constructive assessment of our manuscript. We nevertheless would like to reply to the Reviewers reports.

Reviewer #1.

(...) The data can be well described by three components involving a closed state and two open states O1 and O2, in which the second component O2 is the one affected by the mutations and deletions

This statement is not completely clear to us. What we propose is that O1 is not visible in WT, only in the mutants. What would be affected is the access to O1 and the transition between O1 and O2, but not O2 itself.

From the beginning, it becomes challenging for non-experts to grasp the structural basis of the perturbations that are introduced (ΔPASCap and E600R), because no structural data or schematic cartoons are provided to illustrate the rationale for those deletions or their potential mechanistic effects. In addition, the lack of additional structural information or illustrations, and a somewhat confusing discussion of the structural data, make it challenging for a reader to reconcile the experimental data and mathematical model with a particular structural mechanism for gating, limiting the impact of the work.

Thank you very much for pointing this out and our apologies for the missing cartoon. It will be provided in the revised version.

There are several concerns associated with the analysis and interpretations that are provided. First, the conductance-voltage (G-V) relations for the mutants do not seem to saturate, and the absolute open probability is not quantified for any mutant under any condition. This makes it impossible to quantitatively compare the relative amplitudes of the two components because the amplitude of the second component remains undetermined. […] This reduces confidence in the parameters associated with G-V relations, as the shape and position of both components might change significantly if longer pulses were used.

We agree that the endpoint of activation is ill-defined in the cases where a steady-state is not reached. This does indeed hamper quantitative statements about the relative amplitude of the two components. However, while the overall shape does change, its position (voltage dependence) would not be affected by this shortcoming. The data therefore supports the claim of the “existence of mutant-specific O1 and its equal voltage dependence across mutants.”

Further, because the mutant channel currents do not saturate at the most positive potentials and time intervals examined, the kinetic characterization based on reaching 80% of the maximum seems inappropriate, because the 100% mark is arbitrary.

We agree that the assessment of kinetics by a t80% is not ideal. We originally refrained from exponential fits because they introduce other issues when used for processes that are not truly exponential (as is the case here). To address the concerns, we will add time constants from these fits in the revised version. Please note that in Figure 3, we do provide time constants, and they support the statement made.

Further, the kinetics for some of the other examined mutants (e.g. those in Fig. 2A) are not shown, making it difficult to assess the extent to which the data could be affected by having been measured before full equilibration.

This seems to be a misunderstanding. ∆2-10 kinetics is shown in Fig. 2c. ∆-eag is shown in Fig. 3. We will make sure to state this explicitly in the revised version.

For example, I would expect that the enhanced current amplitudes from Figure 5 are only transient, ultimately reaching a smaller steady-state current magnitude that depends only on the stimulation voltage and is independent of the pre-pulse. The entire time course including the rise-time and decay is not examined experimentally. This raises concern on whether occupancy of state O1 might be overestimated under some experimental conditions if a fraction of the occupancy is only transient. The mathematical model is not utilized to examine some of these slower relaxations - this may be because the model does not reproduce these slow processes, which would represent a serious shortcoming given that the slow kinetics appear to be intrinsic to transitions around state O1.

Thank you for thinking so deeply about the problem. We identified the same questions and did explore them using the model (Figure 8 c). Your intuition is confirmed there, the slow kinetics leads to a decrease of O1 occupancy after a transient accumulation. We intend to study this experimentally as well in the revised version.

The significance of the results with the Δ2-10.L341Split is unclear. First, structural as well as functional data has established that the coupling of the voltage sensor and pore does not entirely rely on the S4-S5 linker, and thus the Split construct could still retain coupling through other mechanisms, which is consistent with the prominent voltage dependence that is observed. If both state O1 and O2 require voltage sensor activation, it is unclear why the Split construct would affect state O1 primarily, as suggested in the manuscript, as opposed to decreasing occupancy of both open states.

Thank you for pointing out the unclear nature of our arguments. We rephrase in the following and will do so in the revised document: If, in non-split mutants, the upward transition of S4 allows entry to O1, it is reasonable to assume that the movement is not transmitted the same way in the split and the transition into O1 is less probable. The observation that, in the split, entry into O1 requires higher depolarization and appears to be less likely, suggests that downstream of S4 (beyond position 342), there is a mechanism to convey S4 motion to the gate of the mutants.

The figure legends and text do not describe which solutions exactly were utilized for each experiment, [...] Because no zero-current levels are shown on the current traces, it becomes very hard to determine which voltages correspond to each of the currents (see Fig. 1A).

Will be corrected.

… the rationale for choosing some solutions over others is not properly explained. […] The reversal potential for solutions used to measure voltage-activation curves falls right at the spot where occupancy of the first component peaks (e.g. see Figure 1B). […] It is unclear whether any artifacts could have been introduced to the mutant activation curves at voltages close to the reversal potential.

The high potassium extracellular solution was chosen to obtain tail currents of sufficient size, warranting precise determination of the reversal potential for every individual experiment. In this way, we ensured that there were no artifacts introduced to the activation curves. Tail currents were used when closing was reasonably fast (∆PASCapL322H and E600RL322H), but otherwise, we used the amplitude at the end of the pulse to get the reversal potential.

One key assumption that is not well-supported by the data pertains to the difference in single-channel conductance between states O1 and O2 - no analysis or discussion is provided on whether the data could also be well described by an alternative model in which O1 and O2 have the same conductance. No additional experimental evidence is provided related to the difference in conductance, which represents a key aspect of the mathematical model utilized to interpret the data.

We agree that the relative conductance of O1 and O2 is a key point. Our proposal mainly stems from the data presented in Fig. 4 and the amplitudes of the two components of the tail at potentials where both states are visible. We also agree that whole cell currents represent a product of occupancy and conductance and that only single channel recordings can produce unambiguous proof for the higher conductance of O1. We have embarked on a series of experiments directly addressing this in the mutants that will be reported in the revised version. Still, we did explore this issue with the model. Following the path of the least number of assumptions, we initially tested models with equal conductance for both states. None of these models was able to reproduce the shape of the tails and the prepulse-dependent increase.

The CaM experiments are potentially very interesting and could have wide physiological relevance. However, the approach utilized to activate CaM is indirect and could result in additional non-specific effects on the oocytes that could affect the results.

Thank you for the appreciative comments about the relevance of our results. We are aware of the potential side effects of the use of thapsigargin and ionomycin, but we still used this approach as an established method to raise intracellular Ca2+. This said, we would like to point out that the effects of Ca2+ increase on channel behavior do revert with a time course that mirrors the estimated time course of Ca2+ itself (supplement 1 to figure 7), suggesting that we are monitoring a Ca2+-dependent event.

The description of the mathematical model that is provided is difficult to follow, and some key aspects are left unclear, such as the precise states from which state O1 can be accessed, and whether there is any direct connectivity between states O1 and O2 - different portions of the text appear to give contradictory information regarding these points.

This seems to be a misunderstanding: supplement 1 to figure 8 graphically details the model’s layout and explicitly shows the connections to the two open states. It also shows that these are not connected. We will make sure that the text is more clearly stating this fact. We did explore models with one open state connected to more than one other state (loops) and found that none of these models can reproduce the large range of depolarizations for with conductance is reduced as compared to lower and higher depolarization (Figure 1).

Several rate constants other than those explicitly mentioned to represent voltage sensor activation are also assigned a voltage dependence - the mechanistic basis of that voltage dependence is unclear.

Some fundamental properties we observed in the mutants can be explained with constant, voltage-independent rate constants into and out of both open states. Specifically, it was possible to achieve behavior very close to that displayed in Figure 8c with constant η, θ, ε, and ζ. We then attempted to also reproduce the strong prepulse-dependence (Figure 6A and B) and found that we needed additional degrees of freedom to incorporate both behaviors with one parameter set. We could either add more states, and thereby rates, or introduce voltage dependence to η and θ. With already 32 states and 10 rates, we decided to adopt the less complex model variant. We agree that this probably reduced the interpretability of the model. As a rule, a transition with a voltage-dependence of the functional form of Eq.1 corresponds to the kinetic properties of two or three transitions, where one is voltage-independent (setting the maximal rate) and one has the classical exponential shape expected from truly molecular transitions.

We also agree that, conceptually, the transitions between the two layers – tentatively associated with a transition in the ring structure– should be voltage-independent. Interestingly, their voltage dependence is very similar to the voltage dependence of the early activation, i.e. centered at -100 and -120mV, similar to β. We therefore attempted to replace the voltage dependence of κ and λ with a state-dependence. To this end, we introduced a parameter that modified κ and λ depending on the state’s position along the α-β axis. While it seemed possible to include all desired features in a model with state-dependent κ and λ, it proved extremely difficult to tune the parameters. Eventually, we reverted to purely voltage-dependent and not state-dependent transition rates κ and λ. Nevertheless, we believe that their voltage dependence could be replaced by some form of state-dependence, i.e. by rates κ and λ that change systematically from the left-hand side of the scheme to its right-hand side.

Finally, a clear mechanistic explanation for the full range of effects that the ΔPASCap and E600R mutants have on channel function is lacking, as well as a detailed description of how those newly uncovered transitions would influence the activity of the WT channel.

We agree. Ultimate mechanistic explanations will have to await data from protein structures of intermediate states and in particular the mutant-specific open state.

…as well as a detailed description of how those newly uncovered transitions would influence the activity of the WT channel; this latter point is important when considering whether the findings in the manuscript advance our understanding of the gating mechanism of Kv10 channels in general, or are specific to the particular mutants that are studied.

We still do not know if the transitions to O1 are identical in the mutants and WT, although our data opens the path to dissecting the interplay of intracellular domains and voltage sensor. We think that the results are relevant for KCNH channels in general because we have made visible otherwise invisible states.

It is unclear, for example, how both the mutation or the deletion at the cytoplasmic gating ring enable conduction by state O1, especially when considering the hypothesis put forward in this study that transition to O1 exclusively involves transitions by the voltage sensor and not the cytoplasmic gating ring.

The transition to O1 is in our model made possible by a displacement of the voltage sensor. In our view, when this occurs with a properly folded and positioned intracellular ring, permeation (access to O1) is precluded. It is precisely the distortion in the intracellular ring induced by mutation or deletion what allows access to O1.

It is also not clearly described whether a non-conducting state with the equivalent state-connectivity as O1 can be accessed in WT channels, or if a state like O1 can only be accessed in the mutant channels. Importantly, if a non-conducting state with the same connectivity to O1 were to be accessed in WT channels, it would be expected that an alternating pulse protocol as in Fig. 4 would result in progressively decreasing currents as the occupancy of the non-conducting state equivalent to O1 is increased. Because this is not the case, it means that mutation and deletion cause additional perturbations on the gating energetics relative to WT, which are not clearly fleshed out.

Thank you for highlighting this important question. Following the arguments in the answer to the previous comment, our experiments cannot provide proof for the existence or accessibility of O1 in WT channels. We favor the interpretation that it is not accessible, because, as you point out, this is supported by the outcome of the alternating pulse on WT (figure 4A) and the paradoxical effect of CaM activation. However, this interpretation hinges on the hypothesis that the kinetics of entry into and departure from O1 would be the same in WT channels, as it is in the mutants. Because transitions into a non-conducting O1 would be only indirectly observable in the WT channel, this assumption would be extremely difficult to test.

Reviewer #2.

WT EAG currents are far right shifted compared to previously published data. It is not clear whether it is the recording conditions but at 0 mV very few channels are open. Compare this with recordings reported previously of the same channel hEAG1 by Gail Robertson's lab (Zhao et. al. (2017) JGP). In that case, most of the channels are open at 0 mV. There must be at least 25 mV shift in voltage-dependence. These differences are unusually large.

G-V curves presented in the literature show a large variability. Depending on the conditions, reported V1/2 values in Xenopus oocytes range from -43 mV (Schönherr et al., 2002 DOI: 10.1016/s0014-5793(02)02365-7) to +16 mV (Lörinczi et al, 2015 DOI: 10.1038/ncomms7672) through +4.1 mV (Lörinczi et al., 2016 DOI: 10.1074/jbc.M116.733576), or +10 mV (in the IUPHAR database). The results in the current manuscript are not significantly different from our previously published results on WT channels. In the report the reviewer is referring to, one source of the difference could be that Zhao et al. had no independent information about the reversal potential. In our experiments, we used solutions with high [K]ext. This places the reversal potential in a voltage range within measurable eag currents and thus allows direct determination of the reversal potential, together with the slow kinetics of the tails and the negative shift in the activation. We would argue that this makes the G-V curves less prone to assumptions, albeit for the price of large error bars around the reversal potential. Additionally, the presence of Mg2+ in the extracellular solutions can change the apparent V1/2 depending on the stimulation protocol.

In most of the mutants, O2 state becomes more prevalent at potentials above +50 mV. At these potentials, endogenous voltage-dependent currents are often observed in xenopus oocytes. The observed differences between the various mutants might simply be a function of the expression level of the channel versus endogenous currents.

Because we were aware of the potential issue of endogenous chloride currents in oocytes, we included data recorded in chloride-free solutions. Those show comparable results, and thus we conclude that endogenous currents are not the origin of the differences between mutants. We will clarify which solutions were used in the figure legends of the revised version and also include the argument against sizable endogenous current contributions in the revision. In a separate line of experiments, we expressed some of the mutants in HEK cells. Despite small current amplitudes, we were able to replicate the findings of two components, providing oocyte-independent evidence for the existence of a second open state.

Voltage-dependence of the kinetics of WT currents appears a bit strange. Why is the voltage-dependence saturated at 0 mV even though very few channels have activated at that point? I cannot imagine any kinetic model that can lead to such unusual voltage-dependence of kinetics.

The fact that voltage dependence of open probability and voltage dependence of activation time constant do not align reflects the multi-state nature of the underlying gating scheme. More than one of several sequential transitions limit the overall kinetics. In this case, the apparent kinetics can reflect a different “bottleneck” transition at different voltage ranges.

One of the other concerns I have is that in many cases, it is clear that the pulse is too short to measure steady-state voltage-dependence. For instance, the currents in -160 mV and -100 mV in Figure 6A and 6B are not saturated.

While we agree that steady-state curves can simplify quantitative evaluation – especially the normalization applied in the I/Imax curves in figure 6 – the conclusion of two components is independent of the absolute amplitude under steady state. The fact that in the raw current traces in Figure 6A, after a -160V prepulse, the same current amplitude is reached for two depolarizations (60 and 90 mV) but not for the intermediate depolarization, can only be explained by an I-V curve that has a minimum. Therefore, the raw data directly support the evidence of finding two components, even if the subsequent analysis is affected by insufficient test pulse durations.

Reviewer #3

Although very well established, the experimental conditions used in the present manuscript introduce uncertainties, weakening their conclusions and complicating the interpretation of the results. The authors performed most of their functional studies in Cl-based solutions that can become a non-trivial issue when the range of voltages explored extends to very depolarizing potentials such as +120mV. Oocytes endogenously express Ca2+-activated Cl- channels that will rectify Cl- at very depolarizing potentials -due to an increase in the driving force- and contribute dramatically to the current's amplitude observed at the test pulse in the voltage ranges where the authors identify the second open state.

As stated above, because we were aware of the potential issue of endogenous chloride currents in oocytes, we performed many of the experiments in chloride-free solutions. We conclude that endogenous currents are not the origin of the differences between mutants because the results were comparable regardless of the presence of chloride. We will clarify which solutions were used in the figure legends of the revised version and also include the argument against sizable endogenous current contributions in the revision. In a separate line of experiments, we expressed some of the mutants in HEK cells. Despite small current amplitudes, we were able to replicate the findings of two components, providing oocyte-independent evidence for the existence of a second open state.

The authors propose a two-layer Markov model with two open states approximating their results. However, the results obtained with the mutants suggest an inactivated state accessible from closed states and a change in the equilibrium between the close/inactivated/open states that could also explain the observed results; therefore, other models could approximate their data.

In the process of model development, we tested a large number of configurations. Those included models with a single open state which we connected to two closed (or inactivated) states that were not directly connected to each other and populated at different voltage ranges. In doing so, we attempted to allow access to the single open state from different regions of the “state-space”, reflecting the two voltage ranges of high conductance. However, in our hands, such a “loop” in the state-space inadvertently leads to a weak separation of the two states and a weak effect of prepulse potentials. The underlying reason is that given the short activation and deactivation time constants, a single open state in a loop provides an effective short-cut, linking otherwise separated parts of the state-space. To achieve the clear separation of the two component’s voltage dependence, two open states that are not connected to each other were essential. As we wrote in response to other comments above, the ultimate proof of two different open states cannot come from modeling, but from single channel measurements.

Author Response

The following is the authors’ response to the original reviews.

Reviewer #1 (Public Review):

In their manuscript, Brischigliaro et al. show that the disruption of respiratory complex assembly results in Drosophila melanogaster results in the accumulation of respiratory supercomplexes. Further, they show that the change in the supercomplex abundance does not impact respiratory function suggesting that the main role of supercomplex formation is structural. Overall, the manuscript is well written and the results and conclusion are supported. The D. melanogaster system, in which the abundance of supercomplexes can be altered through the genetic disruption of the assembly of the individual complexes, will be important for the field to discover the role of the supercomplexes. This manuscript will be of broad interest to the field of mitochondrial bioenergetics. The findings are valuable and the evidence is convincing.

Strengths

The system developed in which the relative levels of SCs can be varied will be extremely useful for studying SC physiology.

The experiments are clearly described and interpreted.

Weaknesses

The statement in the abstract regarding low amounts of SCs in "insect tissues" needs further support or should be narrowed. I am only aware of detailed characterization of the mitochondrial SC composition from D. melanogaster, which is insufficient to make a broad statement about the large and diverse category of insects. This should be rewritten.

Thank you for the comment. We have amended the text accordingly.

In the introduction (line 76) and discussion (line 283), the authors reference the CoQ binding sites in CI and CIII2 being "too far apart" to allow for substrate channeling. The distance between the active sites, though significant, is insufficient to rule out substrate channeling. A stronger argument arises from the fact that the CoQ sites of both CI and CIII2 are open to the membrane and that there are no clear barriers for the free exchange of CoQ with the membrane pool.

Thank you for the comment. We have modified both sentences accordingly.

Line 195, the slight elevation in CI amounts referred to here, does not appear to be statistically significant and therefore should not be discussed a being altered relative to the control.

To address this point of criticism we have revisited the statistical analysis, originally done by 2-way ANOVA and post-hoc test. After giving it some thought, we now consider that this might not have been the correct way to analyze either the mitochondrial respiratory chain (MRC) activity data or the densitometric quantifications. We have now used unpaired two-tailed Student’s t-test to compare the pairs of either KO or KD vs CTRL. The reason is that since the measurement of each individual MRC activity is actually an independent assay, it should be considered separately. The same applies to the densitometry because the absolute values of the intensity of individual CI and that within SCs largely differ. Therefore, we think that it is more correct to compare the abundance of individual CI in the WT vs. either KO or KD pairs and the abundance of the CI in SC independently using a t-test. With these new statistical analyses, the difference in the enzyme activity of CI reported in figure 4D is now significant, which we consider reflects better our observations. Also, with these new analyses, the difference in the amounts of CI+CIII are significantly higher in the Coa8 KD (Figure S1B). Therefore, the original affirmation is correct and we have left the sentence as it was.

Figure 4H, the assignments of the observed larger bands seem incorrect. The largest band (currently assigned as SC I1+III2+IV1) represents too large of a shift for only the addition of CIV and the band currently assigned at SC I1+III2 appears to also contain CIV. The identity of these bands should be reevaluated and additional experiments are needed to definitively prove their identity. This uncertainty should be addressed experimentally or made more explicit in the text.

Thank you for the comment. Taking a closer look at the images, we have to agree with the Reviewer that the assignment was incorrect. The higher band is too large indeed and the reviewer is correct that the band that we previously assigned as CI1+CIII2 does appear to contain CIV as well. Therefore, we have changed the labeling of that to CI1+CIII2+CIV1 because the stoichiometry is compatible with the apparent MW. Also, we have renamed the higher MW band to HMW-SC (high-MW SC) of uncertain nature (unknown stoichiometry) but clearly containing all three complexes I, III and IV. We amended the text (lines 219-221) plus figures 5H and S1 accordingly.

Line 302, the authors state that the structural basis for less SC in D. melanogaster is "due to a more stable association of the NDUFA11 subunit..." However, this would not result is a less stable SC association and only explains why NDUFA11 is more stably associated with CI in the absence of CIII2. The more likely structural reason for the observation of less SC in D. melanogaster is the N-terminal truncation of Dm-NDUFB4 relative to mammalian NDUFB4. This truncation results in the loss of a major SC interaction site between CI and CIII2 in the matrix.

Thank you for pointing this out. We have amended the text accordingly.

Reviewer #2 (Public Review):

Respiratory chain complexes assemble in higher-ordered structures termed supercomplexes or respirasomes. The functional significance of these assemblies is currently investigated, there are two main hypothesis tested, namely that supercomplexes provide kinetic advantages or structural stability. Here, the authors use the fruitfly to reveal that, while the respiratoy chain in the organism normally does not form higher-order assemblies, it does so under conditions when their assembly is impaired. Because the rather moderate increase in supercomplex formation does not change oxygen consumption stimulated by CI or CII substrate, the authors conclude that supercomplex formation has more a structural than a functional role. The main strength of this work is that the technical quality of the experiments is high and that the authors induced defects in respiratory chain assembly through sets of well-controlled genetic models. The obtained data are mostly descriptive using standard approaches and are very well executed. The authors claim that their experiments allow to conclude that the role of supercomplex formation is restricted to a structural role and, hence, exclude a function directly related to electron transport efficiency. However, while the authors can show convincingly that supercomplexes form in the mutants, but not in the wild type, their main claim is not well supported by data and both the structural mechanism of supercompelx formation and their significance remain unknown. While the supercomplex formation observed only in mitochondrial mutants per se is interesting, it would be good to great to define structural aspects of supercomplex formation and their potential impact on the stability of the respiratory chain complexes in these mutants.

We thank the Reviewer for the positive assessment of our work and the suggestions to improve the manuscript.

Reviewer #1 (Recommendations For The Authors):

The sentence on line 90, which starts "This is in contrast with..." is unclear and needs to be rewritten.

Thank you. We have modified the sentence to make it clearer.

Lines 153 and 155, reference is made to tissue specific expression patterns but no literature reference is provided.

Thank you for the comment. The tissue specific expression patterns of the different isoforms are reported in the FlyBase database. We added the link to website in the text.

Line 188, "...homogenates in presence of..." should read "homogenates in the presence of..."

Thank you. Amended.

Line 336, "...lower to the increase..." should read "...lower than the increase..."

Thank you. Amended.

Reviewer #2 (Recommendations For The Authors):

• In order to unravel the molecular mechanism by which supercomplexes form in the mutant, it would be important to identify the factor mediating this. Prime candidates would be additional proteins that co-purify of co-fractionate with the respiratory chain when they assemble into supercomplexes or changes in the lipid composition of the mitochondria, where cardiolipin has been shown to stabilize supercomplex formation. The inclusion and analysis of complexome data for all mutants would be excellent, plus an MS analysis of a purified supercomplex.

Thank you for the suggestion to which we completely agree. We have taken a closer look to the hierarchical clustering of peptide intensities in our complexome profiling data, which clusters the proteins according to their similarity in electrophoretic migration within the complexes. We have specifically looked for proteins in which the peptide intensity changed in a similar fashion as the complex I structural subunits. Among the four candidate proteins (Uniprot IDs Q8SXY6, Q95T19, Q9W0Y6, Q9VJQ3), only Q95T19 — Serine--tRNA synthetase-like protein Slimp is annotated as a mitochondrial protein. This protein is a Drosophila-specific paralog of the mitochondrial Serine-tRNA synthetase generated by gene duplication (PMID: 20870726), which carries out a function linking mitochondrial translation with mtDNA maintenance (PMID: 30943413). Therefore, in principle we would not consider it as a good candidate to be a ‘SC assembly factor’. The identification of factors promoting the formation of SC in Drosophila under these conditions is definitely an important point warranting future investigation.

• The authors could define the stability of the respiratory chain complexes through metabolic pulse-chase labeling experiments. This could reveal that the role of supercomplex formation is indeed structural, improving stability.

We agree that this would be an important piece of information to understand the phenomenon we have observed. Unfortunately, it is technically impossible to perform metabolic labeling of mitochondrial proteins in whole flies. It would be possible to perform in organello pulse-chase labelling, however our previous experience indicates that complex I does not completely assemble de novo in isolated mitochondria (PMID: 20385768).

• The authors should analyze oxygen consumption from mitochondria isolated from larvae as in the other experiments on enzyme activities or the (high-quality) BN-PAGE, and not from whole flies that are homogenized. Moreover, they need to determine the quantities of the complexes by complementary experiments (MS, Western blotting or spectroscopy).

Thank you for the comments. However, we believe that repeating the entire analyses with the larvae would not add significant information to the work and the main interpretation would not change, as the main claim of the paper is based on the data collected on adult flies. In addition, the band patterns of MRC complexes in the BNGE is the same between larvae and adults and therefore, does not depend on the developmental stage. Regarding the quantification of the complexes, we think that the data provided by using complementary approaches such as in gel activity assays (IGA), western blot (WB) and kinetic assays of MRC enzymatic activities, allowed us to confidently determine the amount of the individual complexes. Hence, we performed IGA assays and enzymatic activity assays (which reflect the amounts of fully assembled and functional complexes) in triplicate (independent samples). For the WB analyses, due to the scarcity of some of the antibodies available to detect the Dm MRC proteins, which were a kind gift of Dr. Edward Owusu-Ansah (Columbia University), we decided to pool the three independent samples of each group before running them through the Blue-Native gels. The densitometric curves of the WB bands (Figure S2) show the abundance of each individual MRC complex within the ‘free’ and SC forms. We prioritized the BN analyses over SDS-PAGE and WB analysis, as we consider that just measuring the steady-state levels of MRC subunits is not as informative, because it is possible that certain subunits are present in the mitochondrial membranes but not assembled into the final mature structures.

• Can changes in Coenzyme Q levels explain the absence of a defect on electron transport? This could be determined for the mutant as well as the wild type animals.

We agree that this would be a relevant aspect to investigate. For example, determining whether lower CoQ levels are able to maintain the same respiratory activities in the models in which higher amounts of SCs are formed, as it was proposed in Shimada et al. (PMID: 29191512) would be very interesting. However, the fact that the mild KD models show no MRC enzymatic defects whatsoever (Figure 4D, Figure 5I and Figure 6I), provides the most straightforward explanation to the observed absence of respiratory defects.

Author Response

The following is the authors’ response to the original reviews.

Reviewer #1 (Recommendations For The Authors):

Some sentences need to be clarified and some additional data and references could be added.

1) Line 18

SRY is the sex-determining gene

SRY is the testis-determining gene is more accurate as described in line 44

Modification done

2) Line 50

Despite losing its function in early testis determination in mice, DMRT1 retained part of this function in adulthood when it is necessary to maintain Sertoli cell identity.

Losing its function is misleading. The authors describe firstly that Dmrt1 has no obvious function in embryonic testis development but is critical for the maintenance of Sertoli cells in adult mice. The wording "losing its function in early testis" is confusing. Do the authors mean that despite the expression of Dmrt1 in early testis development, the function of Dmrt1 seems to be restricted to adults in mice? A comparison between the testis and ovary should be more cautious since GarciaAlonso et al (2022) have shown that the transcriptomics of supporting cells between humans and mice is partly different.

That’s what we thought, and the sentence has been changed as follow: “Although DMRT1 is not required for testis determination in mice, it retained part of its function in adulthood when it is necessary to maintain Sertoli cell identity.” (line 51 to 53)

3) Line 78

XY DMRT1-/- rabbits showed early male-to-female sex reversal.

Sex reversal indicates that there is no transient Sertoli cell differentiation that transdifferentiate into granulosa cells. This brings us to an interesting point. In the case of reprogramming, the transient Sertoli cells can produce AMH leading to the regression of the Mullerian ducts. In humans, some 9pdeleted XY patients have Mullerian duct remnants and feminized external genitalia. This finding indicates early defects in testis development.

Is there also feminized external genitalia in XY Dmrt1−/− rabbits. Can the authors comment on the phenotype of the ducts?

We proposed to add “and complete female genitalia” at the end of the following sentence: “Secondly, thanks to our CRISPR/Cas9 genetically modified rabbit model, we demonstrated that DMRT1 was required for testis differentiation since XY DMRT1-/- rabbits showed early male-tofemale sex reversal with differentiating ovaries and complete female genitalia.” (line 77 to 80)

Indeed, since the first stage (16 dpc) where we can predict the sex of the individual by observing its gonads during dissection, we always predict a female sex for XY DMRT1 KO fetuses. It is only genotyping that reveals an XY genotype. At birth, our rabbits are sexed by technicians from the facility and again, but now based on the external genitalia, they always phenotype these rabbits as female ones. In these XY KO rabbits, the supporting cells never differentiate into Sertoli, and ovarian differentiation occurs as early as in XX animals. Thus, these animals are fully feminized with female internal and external genitalia. Most of 9p-deleted patients are not homozygous for the loss-offunction of DMRT1, and the remaining wild-type allele could explain the discrepancy between KO rabbits and humans.

4) Line 53

In the ovary, an equivalent to DMRT1 was observed since FOXL2 (Forkhead family box L2) is expressed in female supporting cells very early in development.

Can the authors clarify what is the equivalent of DMRT1, is it FOXL2? DMRT1 heterozygous mutations result in XY gonad dysgenesis suggesting haploinsufficiency of DMRT1. However, to my knowledge, there is no evidence of haploinsufficiency in XX babies. Thus can we compare testis and ovarian genetics?

We agree, the term “equivalent” is ambiguous, and we changed the sentence as follows: “In ovarian differentiation, FOXL2 (Forkhead family box L2) showed a similar function discrepancy between mice and goats as DMRT1 in the testis pathway. In the mouse, Foxl2 is expressed in female supporting cells early in development but does not appear necessary for fetal ovary differentiation. On the contrary, it is required in adult granulosa cells to maintain female-supporting cell identity.” (line 53 to 56)

Regarding reviewer 2's question on haploinsufficiency in humans: the patient described in Murphy et al., 2015 is an XY individual with complete gonadal dysgenesis. But, it has been shown that the mutation carried by this patient leads to a dominant-negative protein, equivalent to a homozygous state (Murphy et al., 2022).

For FOXL2 mutation in XX females, haploinsufficiency does not affect early ovarian differentiation (no sex reversal) but induces premature ovarian failure.

We agree with the reviewer, we cannot compare testis and ovarian genetics considering two different genes.

5) Line 55

In mice, Foxl2 does not appear necessary for fetal ovary differentiation (Uda et al., 2004), while it is required in adult granulosa cells to maintain female-supporting cell identity (Ottolenghi et al., 2005). The reference Uhlenhaut et al (2009) reporting the phenotype of the deletion of Foxl2 in adults should be added.

The reference has been added.

6) Line 64<br /> These observations in the goat suggested that DMRT1 could retain function in SOX9 activation and, thus, in testis determination in several mammals.

Lindeman et al (2021) have shown that DMRT1 can act as a pioneer factor to open chromatin upstream and Dmrt1 is expressed before Sry in mice (Raymond et al, 1999, Lei, Hornbaker et al, 2007). Whereas additional factors may compensate for the absence of Dmrt1, these results suggest that DMRT1 is also involved in Sox9 activation.

Dmrt1 is indeed expressed before Sry/Sox9 in the mouse gonad. However, no binding site for DMRT1 could be observed at Sox9 enhancer 13 in mice. This does not support a role for DMRT1 in the activation of Sox9 expression in this species. Furthermore, in Lindeman et al 2021, the authors clearly state that DMRT1 acts as a pioneering factor for SOX9 only after birth. It does not appear to have this role before. One of the explanations put forward is that the state of chromatin is different during fetal development in mice: chromatin is more permissive and does not require a factor to facilitate its opening. This hypothesis is based in particular on the description of a similar chromatin profile in the precursors of XX and XY fetal supporting cells, where many common regions display an open structure (Garcia-Moreno et al., 2019). Once sex determination and differentiation are established, a sex-specific epigenome is set up in gonadal cells. Chromatin remodeling agents are then needed to regulate gene expression. We hypothesize that in non-murine mammals such as rabbits, the state of gonadal cell chromatin would be different in the fetal period, more repressed, requiring the intervention of specific factors for its opening, such as DMRT1.

7) Figure 1

Most of the readers might not be familiar with the developmental stages of the gonad in rabbits. A diagram of the key stages in gonad development would facilitate the understanding of the results.

Thank you, it has been added in Figure 1.

8) Figure 2

Arrowheads are difficult to spot, could the authors use another color?

Done

9) Line 117: can the authors comment on the formation of the tunica albuginea? Do the epithelial cells acquire some specific characteristics?

The formation of the tunica albuginea begins with the formation of loose connective tissue beneath the surface epithelium of the male gonad. The appearance of this tissue is concomitant with the loss of expression of DMRT1 in the cell of the coelomic epithelium. Our interpretation is that the contribution of the cells from the coelomic epithelium and their proliferation stops when the tunica begins to form because the structure of the tissue beneath the epithelium change, and the cellular interactions between the epithelium and the tissue below remain disrupted. By contrast, these interactions persist in the ovary until around birth for ovigerous nest formation.

10) The first part of the results described DMRT1 expression in rabbits. With the new single-cell transcriptomic atlas of human gonads, it would be important to describe the pattern of expression in this species. This could be described in the introduction in order to know the DMRT1 expression pattern in the human gonad before that of the rabbit.

A comment on the expression pattern of DMRT1 in human fetal gonads has been added in the discussion section: “In the human fetal testis, DMRT1 expression is co-detected with SRY in early supporting gonadal cells (ESCGs), which become Sertoli cells following the activation of SOX9 expression (Garcia-Alonso et al., 2022) » (line 222 to 224)

11) Figure 3 supplement 3

Dotted line: delimitation of the ovarian surface epithelium. Could the authors check that there is a dotted line?

Done

12) Figure 5 and Line 186

Quantification is missing such as the % of germ cells, % of meiotic germ cells.

Quantification is not easy to realize in rabbits because of the size and the elongated shape of the gonad. Indeed, it’s difficult to be sure that both sections (one from WT, the other from KO) are strictly in a similar region of the gonad and that the section is perfectly longitudinal or not. See also our answer to reviewer 3 (point 7) on this aspect. Actually, we are trying to make a better characterization of this XX phenotype and to find a marker of the pre-leptotene/leptotene stage susceptible to work in rabbits (SYCP3 will be the best, but we encountered huge difficulties with different antibodies and even RNAscope probe!). So actually, the most convincing indirect evidence of this pre-meiotic blockage (in addition to HE staining at 18 dpp in the new Figure 6) is the persistence of POU5F1 (pluripotency), specifically in the germinal lineage of KO XX and XY gonads. In addition to the new figure supplement 5, we can show you in Author response image 1: (i) the gonadal section at a lower magnification, where it is evident that there is a big difference between WT and KO germ cell POU5F1-stainings; and (ii) POU5F1 expression from a bulk RNA-seq realized the day after birth at 1 dpp where the difference is also transcriptionally very clear.

Author response image 1.

13) Line 186,

E is missing at preleptoten

Added

14) Figure supplement 7.

A magnification of the histology of the gonads is missing.

This figure is only for showing the gonadal size, and there are the same gonads as in the new Figure 6. So, the magnification is represented in Figure 6.

15)Discussion

Line 201

SOX9, well known in vertebrates,

The references of the human DSD associated with SOX9 mutations are missing. Thank you, references have been added.

16) Line 286

One of the targets of WNT signaling is Bmp2 in the somatic cells and in turn, Zglp1, which is required for meiosis entry in the ovary as shown by Miyauchi et al (2017) and Nagaoka et al (2020). Does the level of BMP pathway vary in DMRT1 mutants?

At 20 dpc, the expression level of BMP2 in XY and XX DMRT1 mutants gonads is similar to the one of XX control which is lower than in XY control (see the TMP values from our RNA-seq in Author response image 2).

Author response image 2.

Reviewer #2 (Recommendations For The Authors):

Here are my minor comments:

1) Line 106- You mention that coelomic epithelial cells only express DMRT1. Please add an arrow to highlight where you refer to.

Done

2) Line 112: In mice, the SLCs also express Sox9 but not Sry apart from Pax8. You mention here that the SLCs are expressing SRY and DMRT1 in addition to PAX8. Could you perhaps explain the difference? Please refer to that in the results or discussion.

We add a new sentence at the end of this paragraph on SLCs: “As in mice, these cells will express SOX9 at the latter stages (few of them are already SOX9 positive at 15 dpc), but unlike mice, they express SRY.” (line 114 to 115)

We already have collaborations with different labs on these SLC cells, and we will certainly come back later on this aspect, remaining slightly off-topic here.

3) Could you please explain why did you chose to target Exon 3 of DMRT1 and not exons 1-2 which contain the DM domain? Was it to prevent damaging other DMRT proteins? Is there an important domain or function in Exon 2?

Our choice was mainly based on technical issues (rabbit genome annotation & sgRNA design), but also we want to avoid targeting the DM domain due to its strong conservation with other DMRT genes. Due to the poor quality of the rabbit genome, exons 1 and 2 are not well annotated in this species. We have amplified and sequenced the region encompassing exons 1 & 2 from our rabbit line, but the software used for sgRNA design does not predict good guides on this region. The two best sgRNAs were predicted on exon 3, and we used both to obtain more mutated alleles.

4) Your scheme in Supp Figure 4 is not so clear. It is not clear that the black box between the two guides is part of Exon 3 (labelled in blue).

The scheme has been improved.

5) Did you only have 1 good founder rabbit in your experiment? Why did you choose to work with a line that had duplication rather than deletion?

Very good point! In the first version of this paper, we’d try to explain the long (around 2 years) story of breeding to obtain the founder animal. Here it is:

During the genome editing process, we generate 6 mosaic founder animals (5 males and 1 female), then we cross them with wild-type animals to isolate each mutated allele in F1 offspring used afterward to establish and amplify knockout lines. Unexpectedly, we observe a very slow ratio of mutated allele transmission (5 on 129 F1 animals), and only one mutated allele has been conserved from the unique surviving adult F1 animal. It consists of an insertion of the deleted 47 bp DNA fragment, flanked by the cutting sites of the two RNA guides used with Cas9.<br /> The main hypothesis to explain this mutation event is that in the same embryonic cell, the deletion occurs on one allele then the deleted fragment remains inserted into the other allele. Under this scheme, the embryonic cell carries a homozygous DMRT1 knockout genotype, albeit heterogeneous, with a deleted allele (del47) and the present allele (insertion of a 47 bp fragment leading to an in sense duplication). This may explain the very low frequency of transmission since all germ cells carrying a homozygous DMRT1-/- genotype will probably not be able to enter the meiotic process as suggested by our results on XX and XY DMRT1-/- ovaries. Finally, and under this hypothesis, the way we obtained this unique founder animal remains a mystery!

6) Figure 4- real-time data- where does it say what is a,b,c,d of the significance? It should appear on the figure itself and not elsewhere.

Modification done.

7) If I understand correctly, you were able to get the rabbits born and kept to adulthood (you show in supp figure 7 their gonads). What was the external phenotype of these rabbits? Did the XY mutant gonads have the internal and external genitals of a female (oviduct, uterus, vagina etc.)?

See our answer to Reviewer 1 on this question (point 3).

8) Line 20: It is more correct to write 46, XY DSD rather than XY DSD

Modification done.

9) Line 21: you can remove the "the" after abolished

Modification done.

10) Line 31: consider replacing the first "and" by "as well as" since the sentence sounds strange with two "and".

Modification done.

11) Line 212- Please check with the eLife guidelines if they allow "data not shown" in the paper.

This is unspecified.

Reviewer #3 (Recommendations For The Authors):

The following points should be addressed.

1) The in situ's in Fig 1 and 2 are very clear. Fig 1 and Fig 2, In situ hybridisation in tissue sections, it looked like DMRT1 could be expressed in some cells where SRY mRNA is absent @ E13.5dpc and 14.5 dpc. Do you think this is real, or maybe Sry is turned off now in those cells?

Based on the results of in situ hybridizations, DMRT1 appears to be expressed by both coelomic epithelium and genital crest medullar cells in a pattern that is actually broader than that of SRY. Moreover, in rabbits, SRY expression seems to start in the medulla of the genital ridge rather than in the surface epithelium, as described in mice (see Figure 1 at 12 and 13 dpc). Nevertheless, more detailed analyses are needed to ensure the lineage of cells expressing SRY and/or DMRT1, such as single-cell RNAseq at these key stages of sexual determination in rabbits (from 12 to 16 dpc).

2) It is curious that SRY expression is elevated in the DMRT1 KO (Knockout) rabbit gonads. Does this suggest feedback inhibition by DMRt1, or maybe indirect via effect on Sox9 (as I believe Sox9 feeds back to down-regulate Sry in mouse, for example).

The maintenance of SRY expression in the DMRT1 -/- rabbit testis seems to be linked to the absence of SOX9 expression. We believe that, as in mice, SOX9 would down-regulate SRY (even if, in rabbits, SRY expression is never completely turned off).

3) I suggest the targeting strategy and proof of DMRT1 knockout by sequencing etc. be brought out of the suppl. Data and shown as a figure in the text.

See also our answer to reviewer 2 (point 5). It has needed huge efforts to obtain these DMRT1 mutated rabbit line, and of course, it constitutes the basis of the study. But regarding the title and the main message of the article, we are not convinced that the targeting strategy should be moved into the main text.

4) Unless there are limitations imposed by the journal, I also feel that Suppl Fig 5 (the immunostaining) deserves to be in the paper text too. The Fig showing loss of DMRt1 by immunostaining is important.

We include the figure supplement 5 in the main text. So, Figure 4E and figure supplement 5 have been combined into a new Figure 5.

5) The RT-qPCR data should have the statistics clarified on the graphs. (e.g., it is stated that, although Sox9 mRNA is clearly down, there is a slight increase compared to control on KO XX gonads. Is this statistically significant? Figure legend states that the Kruskal-Wallis test is used, and significance is shown by letters. This is unclear. It would be better to use the more usual asterisks and lines to show comparisons.

Modification done.

6) Reference is made to DMRT1+/- rabbits having aberrant germ cell development, pointing to a dosage effect. This is interesting. Does the somatic part of the gonad look completely normal in the het knockouts?

DMRT1 heterozygous male rabbits have a phenotype of secondary infertility with aging, and we are trying now to better characterize this phenotype. The problem is complex because, as we cannot carry out conditional KO, it remains difficult to decipher the consequence of DMRT1 haploinsufficiency in the Sertoli cells versus the germinal ones. Anyway, the somatic part is sufficiently normal to support spermatogenesis since heterozygous males are fertile at puberty and for some months thereafter.

7) Can the authors indicate why meiotic markers were not used to explore the germ cell phenotype? It would be advantageous to use a meiotic germ cell marker to definitely show that the germ cells do not enter meiosis after DMRT1 loss. (Not just H/E staining or maintenance of POU). Example SYCP3, or STRA8 (as pre-meiotic marker) by in situ or immunostaining. Even though no germ cells were detected in adult KO gonads.

The expression of pre-meiotic or meiotic markers is currently under study in DMRT1 -/- females. Transcriptomic data (RNA-seq) are also being analyzed. We are preparing a specific article on the role of DMRT1 in ovarian differentiation in rabbits. We felt it was important to reveal the phenotype observed in females in this first article, but we still need time to refine our description and understanding of the role of DMRT1 in the female.

8) What future studies could be conducted? In the Discussion section, it is suggested that DMRT1 could act as a pioneering factor to allow SRY action upon Sox9. How could this be further explored?

To explore the function of DMRT1 as a pioneering factor, it now seems necessary to characterize the epigenetic landscapes of rabbit fetal gonads expressing or not DMRT1 (comparison of control and DMRT1-/- gonads). Two complementary approaches could be privileged: the study of chromatin opening (ATAC-seq) and the analysis of the activation state of regulatory regions (CUT&Tag). The study of several histone marks, such as H3K4me3 (active promoters), H3K4me1 (primed enhancers), H3K27ac (enhancers and active promoters), and H3K27me3 (enhancers and repressed promoters), would be of great interest. However, these techniques are only relevant for gonads that can be separated from the adjacent mesonephros, which is only possible from the 16 dpc stage in rabbits. To perform a relevant analysis at earlier stages, a "single-nucleus" approach such as ATAC-seq singlenucleus or multi-omic single-nucleus combining ATAC-seq and RNA-seq could be used.

Author Response

The following is the authors’ response to the previous reviews

Reviewer #1:

Comment on revised manuscript: Thank you for your responses - they have addressed most of my concerns.

We thank the reviewer again for their assistance in improving our manuscript.

Reviewer #2:

Additional context:

The sex differences between the samples are interesting as effects of sex are commonly found in AAC tasks. It would be interesting to look at the main model comparison with sex included as a covariate.

Firstly, we thank the reviewer for their re-evaluation of our manuscript.

To the reviewer’s comment, we apologise for the lack of clarity. The analyses included in our revision were indeed based on the main logistic regression model of choice, including sex and age as covariates. We have clarified this in the manuscript as follows:

While sex was significantly associated with choice in the hierarchical logistic regression in the discovery sample (β = 0.16 ± 0.07, p = 0.028) with males being more likely to choose the conflict option, this pattern was not evident in the replication sample (β = 0.08 ± 0.06, p = 0.173), and age was not associated with choice in either sample (p > 0.2).

As it is difficult to include sex as a covariate in the reinforcement learning models in the classical sense as in a linear regression, we assessed sex effects on the individual parameters produced by these models instead, as follows:

Comparing parameters across sexes via Welch’s t-tests revealed significant differences in reward sensitivity (t289 = -2.87, p = 0.004, d = 0.34; lower in females) and consequently reward-punishment sensitivity index (t336 = -2.03, p = 0.043, d = 0.22; lower in females i.e. more avoidance-driven). In the replication sample, we observed the same effect on reward-punishment sensitivity index (t626 = -2.79, p = 0.005, d = 0.22; lower in females). However, the sex difference in reward sensitivity did not replicate (p = 0.441), although we did observe a significant sex difference in punishment sensitivity in the replication sample (t626 = 2.26, p = 0.024, d = 0.18).

Could the authors double check the mean/SD of approach in each group for typos? The numbers are identical.

Thank you for spotting this – the means were indeed similar (discovery: 0.521, replication: 0.516), but the standard deviations were marginally different (discovery: 0.140, replication: 0.148). We have amended the manuscript to reflect this, as follows:

Across individuals, there was considerable variability in overall choice proportions (discovery sample: mean = 0.52, SD = 0.14, min/max = [0.03, 0.96]; replication sample: mean = 0.52, SD = 0.15, min/max = [0.01, 0.99]).

Reviewer #3:

The revised paper commendably adds important additional information and analyses to support these claims. The initial concern that not accounting for participant control over punisher intensity confounded interpretation of effects has been largely addressed in follow-up analyses and discussion.

I commend the authors on their revisions. My initial concerns have been largely addressed. Minor suggestions below.

We thank the reviewer again for their assistance in improving our analyses and manuscript.

Changing the visualisation of the logistic regression model in Figure 2 to tertiles instead of quartiles seems expedient, and does not properly address the points raised by the other reviewers. The argument that non-linear trends in the extreme bins are due to less data is plausible, but unsatisfying given how reliable the pattern seems to be (across samples, with small standard error) and . It is possible, albeit perplexing, that the influence of punishment probability on choice is non-linear. I think the current figure with tertiles is acceptable, but I would suggest including the figures with non-linear data as a supplementary figure, for sake of transparency and reader interest.

We agree that this is likely more complex than a simple linear effect (in the logistic space), especially given the concurrent reward probabilities which also fluctuate in the task. We also agree that the non-linear figures should be made available in the interests of transparency, and have included them in the Supplementary Materials.

We direct interested readers to the relevant section from the figure legend as follows:

"Figure 2. Predictors of choice in the approach-avoidance reinforcement learning task. … We show linear curves here since these effects were estimated as linear effects in the logistic regression models, however the raw data showed non-linear trends – see Supplementary Figure 15."

We have included the non-linear figures in Supplementary Section 9.11 Effects of outcome probabilities on choice in the task: non-linear effects as Supplementary Figure 15.

As an aside, the argument that approach-avoidance joystick tasks do not have a non-human counterpart misconstrues the translational root of these tasks, which was (at least in part) an attempt to model (successfully or not) general approach/avoidance processes measured in non-human tasks, e.g. appetitive/aversive runway tasks using rodents.

Our aim in this manuscript was to develop a task that was closely matched to non-human counterparts in both the experimental procedure (choice over reward/punishment outcomes) and cognitive process involved (simultaneous reward/punishment learning). With this in mind, we wanted to convey that non-human and human measures of approach/avoidance processes were historically distinct in terms of the procedures (e.g. using a joystick vs navigating a runway, due to ethological differences), and that this was potentially problematic with respect to computational validity. However, at this early point in the introduction, it was unnecessary to make a strong distinction between these tasks, which as the reviewer duly notes, follow similar approach/avoidance principles and share similar experimental roots. Therefore, we have opted to omit the reference to translational similarity in the relevant text, as follows:

In humans, on the other hand, approach-avoidance conflict has historically been measured using questionnaires such as the Behavioural Inhibition/Activation Scale (Carver and White 1994), or cognitive tasks that rely on motor/response time biases, for example by using joysticks to approach/move towards positive stimuli and avoid/move away from negative stimuli (Guitart-Masip, Huys et al. 2012, Phaf, Mohr et al. 2014, Kirlic, Young et al. 2017, Mkrtchian, Aylward et al. 2017).

Author Response

The following is the authors’ response to the original reviews.

Reviewer #1 (Public Review):

This manuscript by He et al. explores the molecular basis of the different stinging behaviors of two related anemones. The freshwater Nematostella which only stings when a food stimulus is presented with mechanical stimulation and the saltwater Exaiptasia which stings in response to mechanical stimuli. The authors had previously shown that Nematostella stinging is calcium-dependent and mediated by a voltage-gated calcium channel (VGCC) with very pronounced voltage-dependent inactivation, which gets removed upon hyperpolarization produced by taste receptors.

In this manuscript, they show that Exaiptacia and Nematostella differing stinging behavior is near optimal, according to their ecological niche, and conforms to predictions from a Markov decision model.

It is also shown that Exaiptacia stinging is also calcium-dependent, but the calcium channel responsible is much less inactivated at resting potential and can readily induce nematocyte discharge only in the presence of mechanical stimulation. To this end, the authors record calcium currents from Exaipacia nematocysts and discover that the VGCCs in this anemone are not strongly inactivated and thus are easily activated by mechanical stimuli-induced depolarization accounting for the different stinging behavior between species. The authors further explore the role of the auxiliary beta subunit in the modulation of VGCC inactivation and show that different n-terminal splice variants in Exaiptacia produce strong and weak voltage-dependent inactivation.

The manuscript is clear and well-written and the conclusions are in general supported by the experiments and analysis. The findings are very relevant to increase our understanding of the molecular basis of non-neural behavior and its evolutionary basis. This manuscript should be of general interest to biologists as well as to more specialized fields such as ion channel biophysics and physiology.

Some findings need to be clarified and perhaps additional experiments performed.

1) The authors identify by sequencing that the Exaiptacia Cav is a P-type channel (cacna1a). However, the biophysical properties of the nematocyte channel are different from mammalian P-type channels. The cnidarian channel inactivation is exceedingly rapid and activation happens at relatively low voltages. These substantial differences should be mentioned and commented on.

First, we thank Reviewer 1 for thoughtful and detail-oriented comments, as well as their shared appreciation for the molecular basis of unique behaviors. Indeed, Nematostella and rat CaV channels exhibit striking differences in inactivation (both fast and steady-state). We previously described this in Weir et al., 2020 and added additonal text to ensure that this result is clear.

2) The currents from Nematostella in Figure 3d seem to be poorly voltage-clamped. Poor voltage-clamp is also evident in the sudden increase of conductance in Figure 3C and might contribute to incorrect estimation of voltage dependence of activation and if present in inactivation experiments, also to incorrect estimation of the inactivation voltage range. This problem should be reassessed with new data.

Because it is necessary to use small-tipped pipettes to get recordings from small and technically challenging nematocytes, there is imperfect voltage clamp that is evident in the steep activation curves. This issue should have little effect on the inactivation curves determined with 1s pre-pulses because poor voltage control occurs transiently at the beginning of the pre-pulse. In our case, current is measured in response to a brief maximally activating pulse followed by a nearly 1s period. Thus, error should be minimal in inactivation curves if the test pulse is a maximally activating voltage. We ensured that these protocols are clearly described in the Methods to address this issue. In addition, we are confident in the described inactivation values because they are generally consistent with channel properties measured in a heterologous expression system in which we do not have this problem and see the same differences in inactivation (also see Weir et al., 2020).

3) While co-expression of the mouse Cav channel with the beta1 isoform from Exaiptacia indeed shifts inactivation to more negative voltages, it does not recapitulate the phenotype of the more inactivated Ca-currents in nematocytes (compare Figures 4d and 5d). It should be explained if this might be due to the use of a mammalian alpha subunit. Related to this, did the authors clone the alpha subunit from Exaiptacia? Using this to characterize the effect of beta subunits on inactivation might be more accurate.

While the cnidarian CaVβ subunits indeed shift inactivation consistent with native properties, we agree that using the Exaiptasia alpha subunit would be more accurate. We were unable to successfully clone and heterologously express this subunit, however, we did express all subunits from Nematostella and made chimeric channels in which alpha, alpha2d, or CaVβ were swapped between Nematostella and mammalian channels. These experiments demonstrated the requirement and sufficiency of the CaVβ subunit in altering inactivation (Weir et al., 2020). Furthermore, we were able to express CaVβ subunits from a variety of other cnidarians, all of which affected inactivation properties. Thus, we are confident in the conclusion that CaVβ subunits are major contributors to molecular tuning of cnidarian CaV channels. Future studies aim to incorporate describing properties of the alpha subunit from Exaiptasia and other cnidarians.

4) The in situ shown in Figure 4b are difficult to follow for a non-expert in cnidarian anatomy. Some guidance should be provided to understand the structures. Also, for the left panels, is the larger panel the two-channel image? If so, blue would indicate co-localization of the two isoforms and there seems to be a red mark in the same nematocyte.

We thank the reviewer for this important comment and have modified the figure to enhance visual guidance. We more clearly highlighted the nematocyte in the single and two-channel images and selected the clearest representative images. For additional reference, previous studies beautifully illustrate the unusual morphology of nematocytes, including the relative localization of the nematocyst and nucleus in the context of cnidarian tissues (Babonis and Martindale, 2017).

Reviewer #2 (Public Review):

This manuscript links the distinctive stinging behavior of sea anemones in different ecological niches to varying inactivation properties of voltage-gated calcium channels that are conferred by the identity of auxiliary Cavbeta subunits. Previous work from the Bellono lab established that the burrowing anemone, Nematostella vectensis, expresses a CaV channel that is strongly inactivated at rest which requires a simultaneous delivery of prey extract and touch to elicit a stinging response, reflecting a precise stinging control adapted for predation. They show here that by contrast, the anemone Exaiptasia diaphana which inhabits exposed environments, indiscriminately stings for defense even in the absence of prey chemicals, and that this is enabled by the expression of a CaVbeta splice variant that confers weak inactivation. They further use the heterologous expression of CaV channels with wild type and chimeric anemone Cavbeta subunits to infer that the variable N-termini are important determinants of Cav channel inactivation properties.

1) The authors found that Exaiptasia nematocytes could be characterized by two distinct inactivation phenotypes: (1) nematocytes with low-voltage threshold inactivation similar to that of Nematostella (Vi1/2 = ~ -85mV); and (2) a distinct population with weak, high-voltage threshold inactivation (Vi1/2 = ~ -48mV). What were the relative fractions of low-voltage and high-voltage nematocytes? Do the low-voltage Exaiptasia nematocytes behave similarly to Nematostella nematocytes with respect to requiring both prey extract and touch to discharge?

We thank Reviewer 2 for thoughtful comments and questions. Nematocyte patch clamp is technically challenging due to small size, large nematocyst, and, notably, the explosive discharge involved in stinging! Therefore, we only patch clamped a small number of cells. Despite this limitation, we were able to observe two distinct nematocyte populations based on physiological properties. Yet, we did not observe a correlation with morphology and cannot make broad comments on relative fractions. Because morphology was generally similar and Exaiptasia nematocytes discharge even from touch alone, it remains unclear whether the low-voltage population behaves similarly to Nematostella nematocytes that only discharge in response to chemicals and touch. Future in vivo approaches could be used to address this question.

2) The authors state in Fig 3 legend and in the results that Exaiptasia nematocyte voltage-gated Ca2+ currents have weak inactivation compared with Nematostella. This description is imprecise and inaccurate. Figure 3 in fact shows that Exaiptasia nematocyte voltage-gated Ca2+ currents display a faster rate of inactivation compared to Nematostella Ca2+ currents. A sub-population of Exaiptasia nematocytes does display less resting state (or steady-state) inactivation compared to Nematostella Ca2+ currents. The authors need to be more accurate and qualify what type of inactivation property they are talking about.'

We thank Reviewer 2 for this attention to detail and have defined this phrasing early in the text.

3) In a similar vein, the authors need to be more accurate when referring to 'rat beta' used in heterologous expression experiments. It should be made explicit throughout the manuscript that the rat beta isoform used is rat beta2a. Among the distinct beta isoforms, beta2a is unique in being palmitoylated at the N-terminus which confers a characteristic slow rate of inactivation and a right-shifted voltage-dependence of steady-state inactivation consistent with the data shown in Fig. 4D. Almost all other rat beta isoforms do not have these properties.

We used the rat CaVβ2a for comparison because it shares the highest homology with Nematostella CaVβ (Weir et al., 2020). We have now more clearly defined the rat subunit in the text and legends.

4) The profiling of the impact of different Cnidarian Cavbeta subunits on reconstituted Ca2+ channel current waveforms is nice (Fig 5 and Fig 5S1). The N-terminus sequence of EdCaVβ2 is different from palmitoylated rat beta2a, though both have similar properties in showing slow inactivation and a right-shifted voltage-dependence of steady-state inactivation. Does EdCaVβ2 target autonomously the plasma membrane when expressed in cells? If so, this would reconcile with what was previously known and provide a rational explanation for the observed functional impact of the distinct Cavbetas.

As far as we understand the question, our data support that Exaiptasia CaVβ2 targets the plasma membrane for a number of reasons: 1) Expressing Exaiptasia CaVβ2 produces consistent properties in comparison with other CaVβs, suggesting a homogenous population of channel complexes; 2) Distinct cnidarian-Exaiptasia CaVβ2 chimeras produce distinct and internally consistent properties; and 3) Expressing P/Q-type CaV alpha + alpha2d subunits without CaVβ in cell lines does not produce robust measurable voltage-gated currents. We further tested this in our case and found the same result: at an equivalent maximally activating step using the same protocol, we measured 458.68 ± 179.88pA average current amplitude for +Exaiptasia CaVβ2 (n = 6) and 43.03 ± 17.64pA average current amplitude for -CaVβ2 (n = 4).

Reviewer #3 (Public Review):

Summary:

The present article attempts to answer both the ultimate question of why different stinging behaviours have evolved in Cnidiarians with different ecological niches and shed light on the proximate question of which electro-physiological mechanisms underlie these distinct behaviours.

Account of major methods and results:

In the first part of the paper, the authors try to answer the ultimate question of why distinct dependencies of the sting response on internal starvation levels have evolved. The premise of the article that Exaiptasia's nematocyte discharge is independent of the presence of prey (Artemia nauplii) as compared to Nematostella's significant dependence of the discharge on the presence of actual prey, is shown be a robust phenomenon justified by the data in Figure 1.

The hypothesis that defensive vs. predatory stinging leads to different nematocyte discharge behaviours is analysed in mathematical models based on the suitable framework of optimal control/decision theory. By assuming functional relations between the:

1) cost of a full nematocyte discharge and the starvation level.

2) probability of successful predation/avoidance on the discharge level.

3) desirability/reward of the reached nutritional state.

Based on these assumptions of environmental and internal influences, the optimal choice of attack intensity is calculated using Bellman's equation for this problem. The model predictions are validated using counted nematocytes on a coverslip. The scaling of normalised nematocyte discharge numbers with scaled starvation time is qualitatively comparable to what is predicted from the models. The abundance of nematocytes in the tentacles was, on the other hand, independent of the starvation state of the animals.

Next, the authors turn to investigate the proximate cause of the differential stinging behaviour. The authors have previously reported convincing evidence that a strongly inactivating Cav2.1 channel ortholog (nCav) is used by Nematostella to prevent stinging in the absence of prey (Weir et al. 2020). This inactivation is released by hyperpolarising sensory inputs signalling the presence of prey. In this article, it is clearly shown by blocking respective currents that Exaiptasia, too, relies on extracellular Ca2+ influx to initiate stinging. Patch clamp data of the involved currents is provided in support. However, the authors find that in addition to the nCav with a low-inactivation threshold, Exaiptasia has a splice variant with a higher inactivation threshold expressed (Figure 3D).

The authors hypothesise that it is this high-threshold nCav channel population that amplifies any voltage depolarisation to release a sting irrespective of the presence of prey signals. They found that the β subunit that is responsible for Nematostella's unusually low inactivation threshold exists in Exaiptasia as two alternative splice isoforms. These N-terminus variants also showed the greatest variation in a phylogenetic comparison (Figure 5), rendering it a candidate target for mutations causing variation in stinging responses.

Appraisal of methodology in support of the conclusions:

The authors base their inference on a normative model that yields quantitative predictions which is an exciting and challenging approach. The authors take care in stating the model assumptions as well as showing that the data indeed does not contradict their model predictions. The interesting comparative nature of the modelling part of the study is complicated by slightly different cost assumptions for the two scenarios. Hence, Figure 2 needs to be carefully digested by readers.

We thank the reviewer for their careful revision of our work and excellent comments. We simplified Figure 2 considerably to make it easier to digest. We now compare the stinging response for predation vs defense under the same exact definition of cost per nematocyte for both models. You can find examples 1 and 2 in Figure 2 and examples 3 and 4 in Supplementary Figure 3 (see response below).

It would be even more prudent to analyse the same set of cost-of-discharge vs. starvation scenarios for both species. Specifically, for Nematostella the complete cost-of-discharge vs starvation-state curves as for Exaiptasia (Figure 2E, example 2-4) could be used. It is likely that the differential effect size of Nematostella and Exaiptasia behaviour is the strongest if only the flat cost-of-discharge vs starvation is used (Figure 2A) for Nematostella. But as a worst-case comparison the other curves, where the cost to the animal scales with starvation would be a good comparison. This could help the reader to understand when the different prediction of Nematostella's behaviour breaks down. In addition, this minor change could shed light on broader topics like common trade-offs in pursuit predation.

The results hold even when the cost increases moderately with starvation: Figure 2 now shows results with the same cost for predatory and defensive stinging (cost defined in Figure 2A, former examples 1 and 4). Predatory stinging robustly increases with starvation and defensive stinging remains constant or decreases. Interestingly, the fit between theory and data for both anemones improves by using the increasing cost (open circles in Figure 2E right). For other choices of increasing cost functions, defensive stinging will always decrease, and even more so if the cost increases dramatically (like for the former Examples 2 and 3). In contrast, predatory stinging will switch behavior if the cost increases too much with starvation (results with former Examples 2 and 3, now in Supplementary Figure 3 and theoretical arguments in Supplementary Information). Note however that these assumptions are less realistic because they necessitate that the cost of stinging for well-fed animals is negligible with respect to the cost for starved animals. A formal proof of the asymptotic solution for predatory stinging with varying cost is beyond the scope of this work and is subject of ongoing work where we consider implications for Markov Decision Processes in continuous space state.

The qualitatively similar scaling of the model-derived relation between starvation and sting intensity with the counted nematocytes for different feeding pauses is evidence that feeding has indeed been optimised for the two distinct ecological niches. To prove that Exaiptasia uses a similar Ca2+ channel ortholog as well as a different splice variant, the authors employed both clean electrophysiological characterisaiton (Figure 3) as well as transcriptomics data (Figure 4S1).

To strengthen the authors' hypothesis that variation in the N-termini leads to changes in Ca2+ channel inactivation and hence altered stinging, the response sequence variability of 6 Cnidaria was analysed.

Additional context:

Although, the present article focuses on nematocytes alone, currently, there has been a refocus in neurobiology on the nervous systems of more basal metazoans, which received much attention already in the works of Romanes (1885). In part, this is driven by the goal to understand the early evolution of nervous systems. Cnidarians and Ctenophors are exciting model organisms in this venture. This will hopefully be accompanied by more comparative studies like the present one. Some of the recent literature also uses computational models to understand mechanisms of motor behaviour using full-body simulations (Pallasdies et al. 2019; Wang et al. 2023), which can be thought of as complementary to the normative modelling provided by the authors.

Comparative studies of recent Cnidarians, such as the present article, can shed light on speculative ideas on the origin of nervous systems (Jékely, Keijzer, and Godfrey-Smith 2015). During a time (the Ediacarium/Cambrium transition) that has seen the genesis of complex trophic foodwebs with preditor-prey interaction, symbioses, but also an increase of body sizes and shapes, multiple ultimate causes can be envisioned that drove the increase in behavioural complexity. The authors show that not all of it needs to be implemented in dedicated nerve cells.

References:

Jékely, Gáspár, Fred Keijzer, and Peter Godfrey-Smith. 2015. "An Option Space for Early Neural Evolution." Philosophical Transactions of the Royal Society B: Biological Sciences 370 (December): 20150181. https://doi.org/10.1098/rstb.2015.0181.

Pallasdies, Fabian, Sven Goedeke, Wilhelm Braun, and Raoul-Martin Memmesheimer. 2019. "From Single Neurons to Behavior in the Jellyfish Aurelia Aurita." eLife 8 (December). https://doi.org/10.7554/elife.50084.

Romanes, G. J. 1885. Jelly-Fish, Star-Fish and Sea-Urchins: Being a Research on Primitive Nervous Systems. Appleton.

Wang, Hengji, Joshua Swore, Shashank Sharma, John R. Szymanski, Rafael Yuste, Thomas L. Daniel, Michael Regnier, Martha M. Bosma, and Adrienne L. Fairhall. 2023. "A Complete Biomechanical Model of hydra Contractile Behaviors, from Neural Drive to Muscle to Movement." Proceedings of the National Academy of Sciences 120 (March). https://doi.org/10.1073/pnas.2210439120.

Weir, Keiko, Christophe Dupre, Lena van Giesen, Amy S-Y Lee, and Nicholas W Bellono. 2020. "A Molecular Filter for the Cnidarian Stinging Response." eLife 9 (May). https://doi.org/10.7554/elife.57578.

We appreciate the excellent suggestion to further discuss non-neuronal adaptations in the context of studying the evolution of behavior. We have added additional text to the Discussion to cover this interesting field.

Author Response

The following is the authors’ response to the original reviews.

First and foremost, we would like to thank all the editors and reviewers for their thoughtful and thorough evaluations of our manuscript. We greatly appreciate their assessment about the novelty and strength in this study and have revised the manuscript according to their recommendations. Below are our detailed responses and revisions based on the reviewer recommendations.

Reviewer #1 (Recommendations For The Authors):

1) It is unclear the rationale for choosing the P35-42 adolescent window for stimulating the mesofrontal dopamine system.

The dopaminergic innervation in the mesofrontal circuit exhibits a protracted maturation from P21 to P56 (Kalsbeek, Voorn et al. 1988, Niwa, Kamiya et al. 2010, Naneix, Marchand et al. 2012, Hoops and Flores 2017). P35-42 is in the center of this period and captures the mid-adolescent stage in rodents (Spear 2000). We have previously shown that increasing dopamine neuron activity by wheel running or optogenetic stimulation during this period, but not adulthood, can induce formation of mesofrontal dopaminergic boutons and enhance mesofrontal circuit activity in wild-type mice (Mastwal, Ye et al. 2014). We therefore chose the P35-P42 adolescent window to stimulate the mesofrontal dopamine circuit and test the long-term effect of this intervention on the frontal circuit and memory-guided decision-making deficits in mutant mice. We have detailed this rationale in the revised manuscript when we first introduced this intervention.

2). Please provide a justification for choosing the optical recording M2 neuronal activity instead of the prelimbic prefrontal cortex, which has been known to show the highest levels of dopamine terminals.

While the prelimbic area has the highest level of dopamine terminals among frontal cortical regions, a robust presence of dopaminergic terminals and dopamine release in the M2 frontal cortex have been well documented (Berger, Gaspar et al. 1991, Mastwal, Ye et al. 2014, Aransay, Rodriguez-Lopez et al. 2015, Patriarchi, Cho et al. 2018). The M2 cortex plays an important role in action planning, generating the earliest neural signals among frontal cortical regions that are related to upcoming choice during spatial navigation (Sul, Kim et al. 2010, Sul, Jo et al. 2011). Our chemogenetic inactivation experiments (Supplementary Fig 1) has further confirmed the involvement of M2 in the memory-guided Y-maze navigation task used in this study. Technically, M2 has the advantage of being more amendable to optical recording of neuronal activity without the tissue damage caused by implanting a lens, which would be necessary for deeper areas such as the prelimbic cortex. We have provided this justification in the revised manuscript.

3). What was the rationale for using the 3-day chemogenetic stimulation paradigm?

Our previous work in wild-type adolescent mice showed that a single optogenetic stimulation session or a 2-hr wheel running session is sufficient to induce bouton formation in mesofrontal dopaminergic axons (Mastwal, Ye et al. 2014). In this study, we sought to rescue existing structural and functional deficits in the mesofrontal dopaminergic circuits due to genetic mutations. Because previous studies suggested that an optimal level of dopamine is important for normal cognitive function (Arnsten, Cai et al. 1994, Robbins 2000, Floresco 2013), we elected to do multiple stimulation sessions to boost the potential rescue effects. We tested both a 3-day and a 3-week stimulation paradigm, and found that the 3-day, but not the 3-week paradigm led to robust functional improvement (Fig. 5). These results indicate that moderate but not excessive stimulation of dopamine neurons can provide functional improvement of a deficient mesofrontal circuit. We have revised our text to clarify the rationale for these experiments.

4). A major maturational event occurring in the prefrontal cortex is the gain of local GABAergic transmission, which is crucial for sustaining proper levels of Y-maze tasks. I am wondering if the authors have any thoughts about what is really happening at the postsynaptic level following adolescent dopamine stimulation.

The developmental increases in dopaminergic innervation to the frontal cortex and local GABAergic transmission are likely synergistic processes, which both contribute to the maturation of high-order cognitive functions supported by the frontal cortex (Caballero and Tseng 2016, Larsen and Luna 2018). Previous electrophysiological studies have suggested that dopamine can act on five different receptors expressed in both excitatory and inhibitory postsynaptic neurons (Seamans and Yang 2004, Tseng and O'Donnell 2007, O'Donnell 2010). At the network level, dopaminergic signaling can increase the signal-to-noise ratio and temporal synchrony of neural activity during cognitive tasks (Rolls, Loh et al. 2008, Vander Weele, Siciliano et al. 2018, Lohani, Martig et al. 2019). As the frontal GABAergic inhibitory network undergoes major functional remodeling during adolescence (Caballero and Tseng 2016), adolescent stimulation of dopamine neurons may interact with this maturational process to promote a network configuration conducive for synchronous and high signal-to-noise neural computation (Porter, Rizzo et al. 1999, Murty, Calabro et al. 2016, Mukherjee, Carvalho et al. 2019). The microcircuit mechanisms underlying adolescent dopamine stimulation induced changes, particularly in the GABAergic inhibitory neurons, will be an exciting direction for future research. We have extended our discussion about these points in the revised manuscript.

5). A change in the density of dopamine boutons is unlikely to be limited to the M2 region in Arc-/- mice. The authors should provide some data illustrating that similar changes are widespread across the medial prefrontal cortex, and that the optical recording in the M2 region was preferred for technical limitations and to avoid damaging areas in the frontal cortex.

As discussed above, this study focused on the M2 region of the frontal cortex because it is functionally required for memory-guided Y-maze navigation, generates behavioral choice-related neural signals during spatial navigation, and is optically most accessible. The medial prefrontal regions (anterior cingulate, prelimbic and infralimbic) ventral to M2 also receive dense dopaminergic innervation and can act in concert with M2 in decision making (Sul, Kim et al. 2010, Sul, Jo et al. 2011, Barthas and Kwan 2017). As dopaminergic innervations to the frontal cortical regions progress in a ventral-to-dorsal direction during development (Kalsbeek, Voorn et al. 1988, Hoops and Flores 2017), how the changes induced by adolescent dopamine stimulation may proceed spatial-temporally across different frontal subregions requires more extensive investigation in the future. We have added this discussion into the revised manuscript.

Reviewer #2 (Public Review):

The manuscript by Mastwal and colleagues explores how transient adolescent stimulation of ventral midbrain neurons that project to the frontal cortex may help to improve performance on certain memory tasks. The manuscript provides an interesting set of observations that DREADD-based activation over only 3 days during adolescence provides a fast-acting and long-lasting improvement in performance on Y-maze spontaneous alternation as well as aspects of neuronal function as assessed using in vivo imaging methods. While interesting, there are several weaknesses. First and foremost, it is not clear that the effects the authors are observing are mediated by dopamine. It has been clearly documented that the DAT-Cre line provides a better representation of midbrain dopamine cells in the mouse, particularly near the midline of the ventral midbrain (Lammel et al., Neuron 2015). This is precisely where the cells that project to the frontal cortex are located. Therefore, the selection of TH-Cre is problematic. It is very likely that the authors are labeling a substantial number of non-dopaminergic cells.

We agree with Review 2 that the DAT-Cre line can provide specific labeling of midbrain dopamine neurons, particularly those projecting to the striatum, as discussed in the cited study (Lammel, Steinberg et al. 2015). DAT transports the extracellularly released dopamine back into presynaptic terminals, but it is not essential for dopamine synthesis and release (Sulzer, Cragg et al. 2016). Mesocortical dopamine neurons in the ventral tegmental area (VTA) express very little DAT (Sesack, Hawrylak et al. 1998, Lammel, Hetzel et al. 2008, Li, Qi et al. 2013), which limits the use of the DAT-Cre line to target these neurons (Lammel, Steinberg et al. 2015). Because mesocortical dopamine neurons have strong expression of TH, a key enzyme involved in dopamine synthesis, TH-Cre lines have been extensively used to study the mesocortical pathway (Lammel, Lim et al. 2012, Gunaydin, Grosenick et al. 2014, Ellwood, Patel et al. 2017, Vander Weele, Siciliano et al. 2018, Lohani, Martig et al. 2019). We provide more details below about our rationales for using TH-Cre rather than DAT-Cre mice in our study and the revisions we made in response to the reviewer’s specific recommendations.

Reviewer #2 (Recommendations For The Authors):

1). The authors should rigorously demonstrate that there is a reasonable midbrain DA projection to the coordinates that they are assessing and that their effects are due to DA release from these cells. It is not clear that there is a VTA dopaminergic projection to M2 - it does not appear for example in the Allen Mouse Brain Connectivity Atlas (https://connectivity.brainmap.org/projection/experiment/siv/160540751? imageId=160541123&imageType=TWO_PHOTON,SEGMENTATION&initImage=TWO_PHOTON&x=17321&y=15284&z=3). Though there is a projection to the mPFC, at the coordinates the authors report, there does not appear to be any signal from DAT-Cre mice. However, there is much more signal when expression is not restricted to dopamine cells (https://connectivity.brain-map.org/projection/experiment/siv/165975096? imageId=165975158&imageType=TWO_PHOTON,SEGMENTATION&initImage=TWO_PHOTON&x=17950&y=11504&z=3). The argument that these cells may express less TH is not relevant for this particular issue. Therefore, it is possible that the vast majority of observed effects are not in fact mediated by dopamine but another neurotransmitter such as glutamate. While the experiment using SCH23390 does suggest DA receptors may be involved, this result in isolation doesn't alleviate this caveat - there can be, for example, DA release from NE cells (e.g., Takeuchi et al., Nature 2016). While this does not entirely invalidate the authors' results, as their effects of stimulation of ventral midbrain cells to the forebrain don't necessarily have to occur via dopamine - the mechanism by how this is occurring needs to be clear.

While the prelimbic area has the highest level of dopaminergic terminals among frontal cortical regions, a robust presence of midbrain dopaminergic projections and dopamine release in the M2 frontal cortex have been well established by immunostaining, viral labeling, single-cell axon-tracing, and in vivo imaging of recently developed dopamine biosensors (Berger, Gaspar et al. 1991, Mastwal, Ye et al. 2014, Aransay, Rodriguez-Lopez et al. 2015, Ye, Mastwal et al. 2017, Patriarchi, Cho et al. 2018). It has also been reported repeatedly that mesocortical dopamine neurons in the VTA express very little DAT, which is different from mesostriatal dopamine neurons (Sesack, Hawrylak et al. 1998, Lammel, Hetzel et al. 2008, Li, Qi et al. 2013). This limitation in the use of the DAT-Cre line to target mesocortical dopamine neurons has been acknowledged in previous studies (Lammel, Steinberg et al. 2015) and is consistent with the reviewer’s observation of DAT-Cre labeling in the Allen Brain Mouse Connectivity atlas. Additionally, and interestingly, recent extensive evaluation of the DAT-Cre line reported ectopic labeling of multiple non-dopaminergic neuronal populations (Soden, Miller et al. 2016, Stagkourakis, Spigolon et al. 2018, Papathanou, Dumas et al. 2019). Our own evaluation of the DAT-Cre line’s utility for cortical imaging also revealed sparse axonal labeling and sporadic ectopic labeling of cortical cell somas. We have included representative DAT-Cre images in Author response image 1 to highlight the limitations of this line in the study of the dopaminergic mesocortical circuit.

Author response image 1.

Example images from DAT-Cre/Ai14 mice. Left most panel shows little axonal labeling in Layer 5/6 of M2. The center panel shows sparse axonal label in Layer 1/2 of M2, but also ectopic labeling of cell soma. The right panel shows a lack of labeling in L1/2 of prelimbic cortex as well. Scale bars 50um.

We as well as others have confirmed that TH immunoreactivity in the frontal cortex can label dopaminergic axons originated from the VTA, and ablation of VTA dopaminergic neurons removes this labeling (Niwa, Jaaro-Peled et al. 2013, Ye, Mastwal et al. 2017). Because mesocortical dopamine neurons have much stronger TH expression than DAT expression (Sesack, Hawrylak et al. 1998, Lammel, Hetzel et al. 2008, Li, Qi et al. 2013, Lammel, Steinberg et al. 2015), TH-Cre lines have been frequently used to label these neurons and study the mesocortical pathway (Lammel, Lim et al. 2012, Gunaydin, Grosenick et al. 2014, Ellwood, Patel et al. 2017, Vander Weele, Siciliano et al. 2018, Lohani, Martig et al. 2019). While TH-Cre expression itself is not restricted to dopaminergic neurons, we targeted our viral injections to the VTA and optogenetic stimulation to the cortical dopaminergic projection target area in M2 (Patriarchi, Cho et al. 2018) to specifically modulate mesofrontal dopaminergic axons. In addition, we tested D1 antagonist’s effects in our manipulations. Although we targeted dopamine neurons in our adolescent stimulation, the final behavioral outcome likely includes contributions from co-released neurotransmitters such as glutamate and non-dopaminergic neurons via network effects (Morales and Margolis 2017, Lohani, Martig et al. 2019), which will be interesting directions for future research. We have revised our results and discussion sections to highlight our rationales for using the TH-Cre line and the open mechanistic questions for future studies.

2) SSFOs don't increase excitability like DREADDs, but rather, cause long-lasting hyperactivity through continuous passage of cations. What the actual firing properties are of these cells over a long period of time is not clear.

We did not measure the precise firing patterns of the dopaminergic neurons targeted by SSFOs but evaluated the effects of SSFO activation on the frontal cortex. Similar to our DREADD-Gq mediated activity changes in the mesofrontal circuit, we found increased frontal cortical activity post-light stimulation of frontal dopamine axons in our SSFO treated animals (Fig 6a-c, S6e). While quantitatively the firing patterns of DREADD-Gq and SSFO activated dopaminergic neurons likely differ, qualitatively both of these manipulations lead to increased mesofrontal circuit activity and improvements in cognitive behaviors. In our previous work with wild-type adolescent mice, both wheel running and a single 10-min session of phasic optogenetic stimulation of the VTA resulted in dopaminergic bouton outgrowth in the frontal cortex (Mastwal, Ye et al. 2014). Taken together, these results suggest that adolescent dopaminergic mesofrontal projections are highly responsive to neural activity changes and a variety of adolescent stimulation paradigms are sufficient to elicit lasting changes in this circuit. We have added this discussion of the limitations and implications of our study into the revised manuscript.

3) It is not clear what the increase in boutons means, given that DA release is thought to largely occur via non-synaptic release.

Although many of dopamine boutons are not associated with defined postsynaptic structures, these axonal boutons and the active zones they contain are the major release sites for dopamine (Goldman-Rakic, Leranth et al. 1989, Arbuthnott and Wickens 2007, Sulzer, Cragg et al. 2016, Liu, Goel et al. 2021). Past studies have established a consistent association between increased dopaminergic innervation in the frontal cortex and an increase in dopamine levels (Niwa, Kamiya et al. 2010, Naneix, Marchand et al. 2012). Our previous work also found that increasing dopaminergic boutons through adolescent VTA stimulation led to prolonged frontal local field potential responses with high-frequency oscillations (Mastwal, Ye et al. 2014), which is characteristic of increased dopaminergic signaling (Lewis and O'Donnell 2000, Gireesh and Plenz 2008, Wood, Kim et al. 2012, Lohani, Martig et al. 2019). Importantly, in our quantification of the structural changes in this study, we evaluated boutons which were labeled with synaptophysin, a molecular marker indicating the presence of synaptic vesicle release machinery (Li, Tasic et al. 2010, Oh, Harris et al. 2014). Thus, our study, taken in the context of the previous work, suggests the increased number of boutons signifying an increase in dopaminergic signaling within the mesofrontal circuit. We have added this discussion into the revised manuscript.

4) The use of Arc and DISC mutants as models of schizophrenia is perhaps a bit overstated - while deficits in prefrontal innervation certainly occur, there are many differences between these models and the human disease states. Language should be toned down accordingly, particularly in the introduction.

We strived to avoid overstating the extent to which the mouse lines are models for specific diseases, but we can appreciate that this may not have been clear in our original writing. We have adjusted our language to better distinguish between the utility of the animal models for the purposes of our study and their relationship to specific human disease states. Particularly in the introduction, we stated that: “Genetic disruptions of several genes involved in synaptic functions related to psychiatric disorders, such as Arc and DISC1, lead to hypoactive mesofrontal dopaminergic input in mice (Niwa, Kamiya et al. 2010, Niwa, Jaaro-Peled et al. 2013, Fromer, Pocklington et al. 2014, Purcell, Moran et al. 2014, Wen, Nguyen et al. 2014, Manago, Mereu et al. 2016). Although there are many differences between these mouse lines and specific human disease states, these mice offer opportunities to test whether genetic deficits in frontal cortex function can be reversed through circuit interventions.”

5) Some experiments are missing proper controls, e.g., Figure 3g-I where a WT mouse should be used as a positive control.

The goal of this experimental design (Fig 3g-i) was to evaluate the potential effects of chemogenetic VTA stimulation in the Arc-/- mice. We used Arc-/- mice with mCherry injections to control for the potential effects of CNO administration. While WT mice could be used to determine if adolescent VTA stimulation would lead to long-lasting enhancement of VTA-to-Cortical transmission, this wouldn’t necessarily be a positive control for our experiments, but rather a separate line of inquiry. As dopamine’s effects often display an inverted-U dose-response curve (Vijayraghavan, Wang et al. 2007, Floresco 2013), evaluating the effects adolescent VTA stimulation in the absence of underlying dopamine deficiency could be an interesting future research direction. We have added this discussion into the revised manuscript.

Reviewer #3 (Recommendations For The Authors):

1) Did the SSFO stimulation of the TH+ axons in PFC during adolescence lead to the same long-term change in DA bouton number the authors saw with DREADDs?

We did not examine the degree of bouton growth in the SSFO cohort, which is a limitation of this study. Accurate quantification of dopamine boutons requires the co-injection of another AAV vector encoding Synaptophysin-GFP to label the boutons. Because we used light to directly stimulate SSFO-labeled dopaminergic axons in the frontal cortex, we were concerned that co-injecting another AAV vector may dilute SSFO-labeling of axons and reduce the efficacy of optogenetic stimulation. Given the behavioral benefits we observed, we would expect an increase in bouton density after optogenetic stimulation. A systematic optimization of viral co-labeling and optogenetic stimulation protocols will facilitate examination of the impact of SSFO stimulation at the structural level in future studies. We have added a discussion of the limitation of this study in the revised manuscript.

2) The DISC1 section is far less detailed than the Arc section, and it was not completely clear to me that the mechanisms of dysfunction and rescue were the same in these mice compared with the Arc mice. For example, there was no mention of DA bouton density or the patterned firing of the PFC neurons at the time of decision making.

The initial motivation of this study was to test if adolescent dopamine stimulation can rescue the deficits in the mesofrontal dopaminergic circuit and cognitive function of Arc-/- mice, which were identified in our previous studies (Manago, Mereu et al. 2016). We first conducted multiple levels of analyses including viral tracing, in vivo calcium imaging, and behavioral tests to establish the coherent impacts of adolescent dopamine neuron stimulation on circuits and behaviors. We then examined a range of stimulation protocols to assess the efficacy requirements for cognitive improvement, which is our primary goal. Finally, we included DISC1 mice in our study to test if adolescent dopamine stimulation can also reverse the cognitive deficit in another genetic model for mesofrontal dopamine deficiency. By demonstrating a similar cognitive recuse effect of adolescent VTA stimulation in an independent mouse model, this study provides a foundation for future research to compare the detailed cellular mechanisms that underlie the functional rescue in different genetic models. We have added the discussion of the scope and limitation of this study to the revised manuscript.

References

Aransay, A., C. Rodriguez-Lopez, M. Garcia-Amado, F. Clasca and L. Prensa (2015). "Long-range projection neurons of the mouse ventral tegmental area: a single-cell axon tracing analysis." Front Neuroanat 9: 59.

Arbuthnott, G. W. and J. Wickens (2007). "Space, time and dopamine." Trends Neurosci 30(2): 62-69.

Arnsten, A. F., J. X. Cai, B. L. Murphy and P. S. Goldman-Rakic (1994). "Dopamine D1 receptor mechanisms in the cognitive performance of young adult and aged monkeys." Psychopharmacology (Berl) 116(2): 143-151.

Barthas, F. and A. C. Kwan (2017). "Secondary motor cortex: where ‘sensory’meets ‘motor’in the rodent frontal cortex." Trends in neurosciences 40(3): 181-193.

Berger, B., P. Gaspar and C. Verney (1991). "Dopaminergic innervation of the cerebral cortex: unexpected differences between rodents and primates." Trends Neurosci 14(1): 21-27.

Caballero, A. and K. Y. Tseng (2016). "GABAergic Function as a Limiting Factor for Prefrontal Maturation during Adolescence." Trends Neurosci 39(7): 441-448.

Ellwood, I. T., T. Patel, V. Wadia, A. T. Lee, A. T. Liptak, K. J. Bender and V. S. Sohal (2017). "Tonic or Phasic Stimulation of Dopaminergic Projections to Prefrontal Cortex Causes Mice to Maintain or Deviate from Previously Learned Behavioral Strategies." J Neurosci 37(35): 8315-8329.

Floresco, S. B. (2013). "Prefrontal dopamine and behavioral flexibility: shifting from an "inverted-U" toward a family of functions." Front Neurosci 7: 62.

Fromer, M., A. J. Pocklington, D. H. Kavanagh, H. J. Williams, S. Dwyer, P. Gormley, L. Georgieva, E. Rees, P. Palta, D. M. Ruderfer, N. Carrera, I. Humphreys, J. S. Johnson, P. Roussos, D. D. Barker, E. Banks, V. Milanova, S. G. Grant, E. Hannon, S. A. Rose, K. Chambert, M. Mahajan, E. M. Scolnick, J. L. Moran, G. Kirov, A. Palotie, S. A. McCarroll, P. Holmans, P. Sklar, M. J. Owen, S. M. Purcell and M. C. O'Donovan (2014). "De novo mutations in schizophrenia implicate synaptic networks." Nature 506(7487): 179-184.

Gireesh, E. D. and D. Plenz (2008). "Neuronal avalanches organize as nested theta- and beta/gamma-oscillations during development of cortical layer 2/3." Proc Natl Acad Sci U S A 105(21): 7576-7581.

Goldman-Rakic, P. S., C. Leranth, S. M. Williams, N. Mons and M. Geffard (1989). "Dopamine synaptic complex with pyramidal neurons in primate cerebral cortex." Proc Natl Acad Sci U S A 86(22): 9015-9019.

Gunaydin, L. A., L. Grosenick, J. C. Finkelstein, I. V. Kauvar, L. E. Fenno, A. Adhikari, S. Lammel, J. J. Mirzabekov, R. D. Airan, K. A. Zalocusky, K. M. Tye, P. Anikeeva, R. C. Malenka and K. Deisseroth (2014). "Natural neural projection dynamics underlying social behavior." Cell 157(7): 1535-1551.

Hoops, D. and C. Flores (2017). "Making Dopamine Connections in Adolescence." Trends Neurosci 40(12): 709-719.

Kalsbeek, A., P. Voorn, R. M. Buijs, C. W. Pool and H. B. Uylings (1988). "Development of the dopaminergic innervation in the prefrontal cortex of the rat." J Comp Neurol 269(1): 58-72.

Lammel, S., A. Hetzel, O. Hackel, I. Jones, B. Liss and J. Roeper (2008). "Unique properties of mesoprefrontal neurons within a dual mesocorticolimbic dopamine system." Neuron 57(5): 760-773.

Lammel, S., A. Hetzel, O. Haeckel, I. Jones, B. Liss and J. Roeper (2008). "Unique properties of mesoprefrontal neurons within a dual mesocorticolimbic dopamine system." Neuron 57(5): 760-773.

Lammel, S., B. K. Lim, C. Ran, K. W. Huang, M. J. Betley, K. M. Tye, K. Deisseroth and R. C. Malenka (2012). "Input-specific control of reward and aversion in the ventral tegmental area." Nature 491(7423): 212-217.

Lammel, S., E. E. Steinberg, C. Foldy, N. R. Wall, K. Beier, L. Luo and R. C. Malenka (2015). "Diversity of transgenic mouse models for selective targeting of midbrain dopamine neurons." Neuron 85(2): 429-438.

Larsen, B. and B. Luna (2018). "Adolescence as a neurobiological critical period for the development of higher-order cognition." Neurosci Biobehav Rev 94: 179-195.

Lewis, B. L. and P. O'Donnell (2000). "Ventral tegmental area afferents to the prefrontal cortex maintain membrane potential 'up' states in pyramidal neurons via D(1) dopamine receptors." Cereb Cortex 10(12): 1168-1175.

Li, L., B. Tasic, K. D. Micheva, V. M. Ivanov, M. L. Spletter, S. J. Smith and L. Luo (2010). "Visualizing the distribution of synapses from individual neurons in the mouse brain." PLoS One 5(7): e11503.

Li, X., J. Qi, T. Yamaguchi, H. L. Wang and M. Morales (2013). "Heterogeneous composition of dopamine neurons of the rat A10 region: molecular evidence for diverse signaling properties." Brain Struct Funct 218(5): 1159-1176.

Liu, C., P. Goel and P. S. Kaeser (2021). "Spatial and temporal scales of dopamine transmission." Nat Rev Neurosci 22(6): 345-358.

Lohani, S., A. K. Martig, K. Deisseroth, I. B. Witten and B. Moghaddam (2019). "Dopamine Modulation of Prefrontal Cortex Activity Is Manifold and Operates at Multiple Temporal and Spatial Scales." Cell Rep 27(1): 99-114 e116.

Manago, F., M. Mereu, S. Mastwal, R. Mastrogiacomo, D. Scheggia, M. Emanuele, M. A. De Luca, D. R. Weinberger, K. H. Wang and F. Papaleo (2016). "Genetic Disruption of Arc/Arg3.1 in Mice Causes Alterations in Dopamine and Neurobehavioral Phenotypes Related to Schizophrenia." Cell Rep 16(8): 2116-2128.

Mastwal, S., Y. Ye, M. Ren, D. V. Jimenez, K. Martinowich, C. R. Gerfen and K. H. Wang (2014). "Phasic dopamine neuron activity elicits unique mesofrontal plasticity in adolescence." J Neurosci 34(29): 9484-9496.

Morales, M. and E. B. Margolis (2017). "Ventral tegmental area: cellular heterogeneity, connectivity and behaviour." Nat Rev Neurosci 18(2): 73-85.

Mukherjee, A., F. Carvalho, S. Eliez and P. Caroni (2019). "Long-Lasting Rescue of Network and Cognitive Dysfunction in a Genetic Schizophrenia Model." Cell 178(6): 1387-1402 e1314. Murty, V. P., F. Calabro and B. Luna (2016). "The role of experience in adolescent cognitive development: Integration of executive, memory, and mesolimbic systems." Neurosci Biobehav Rev 70: 46-58.

Naneix, F., A. R. Marchand, G. Di Scala, J. R. Pape and E. Coutureau (2012). "Parallel maturation of goal-directed behavior and dopaminergic systems during adolescence." J Neurosci 32(46): 16223-16232.

Niwa, M., H. Jaaro-Peled, S. Tankou, S. Seshadri, T. Hikida, Y. Matsumoto, N. G. Cascella, S. Kano, N. Ozaki, T. Nabeshima and A. Sawa (2013). "Adolescent stress-induced epigenetic control of dopaminergic neurons via glucocorticoids." Science 339(6117): 335-339.

Niwa, M., A. Kamiya, R. Murai, K. Kubo, A. J. Gruber, K. Tomita, L. Lu, S. Tomisato, H. Jaaro-Peled, S. Seshadri, H. Hiyama, B. Huang, K. Kohda, Y. Noda, P. O'Donnell, K. Nakajima, A. Sawa and T. Nabeshima (2010). "Knockdown of DISC1 by in utero gene transfer disturbs postnatal dopaminergic maturation in the frontal cortex and leads to adult behavioral deficits." Neuron 65(4): 480-489.

O'Donnell, P. (2010). "Adolescent maturation of cortical dopamine." Neurotox Res 18(3-4): 306-312.

Oh, S. W., J. A. Harris, L. Ng, B. Winslow, N. Cain, S. Mihalas, Q. Wang, C. Lau, L. Kuan, A. M. Henry, M. T. Mortrud, B. Ouellette, T. N. Nguyen, S. A. Sorensen, C. R. Slaughterbeck, W. Wakeman, Y. Li, D. Feng, A. Ho, E. Nicholas, K. E. Hirokawa, P. Bohn, K. M. Joines, H. Peng, M. J. Hawrylycz, J. W. Phillips, J. G. Hohmann, P. Wohnoutka, C. R. Gerfen, C. Koch, A. Bernard, C. Dang, A. R. Jones and H. Zeng (2014). "A mesoscale connectome of the mouse brain." Nature 508(7495): 207-214.

Papathanou, M., S. Dumas, H. Pettersson, L. Olson and A. Wallen-Mackenzie (2019). "Off-Target Effects in Transgenic Mice: Characterization of Dopamine Transporter (DAT)-Cre Transgenic Mouse Lines Exposes Multiple Non-Dopaminergic Neuronal Clusters Available for Selective Targeting within Limbic Neurocircuitry." eNeuro 6(5).

Patriarchi, T., J. R. Cho, K. Merten, M. W. Howe, A. Marley, W. H. Xiong, R. W. Folk, G. J. Broussard, R. Liang, M. J. Jang, H. Zhong, D. Dombeck, M. von Zastrow, A. Nimmerjahn, V. Gradinaru, J. T. Williams and L. Tian (2018). "Ultrafast neuronal imaging of dopamine dynamics with designed genetically encoded sensors." Science 360(6396): 1420-+.

Porter, L. L., E. Rizzo and J. P. Hornung (1999). "Dopamine affects parvalbumin expression during cortical development in vitro." J Neurosci 19(20): 8990-9003.

Purcell, S. M., J. L. Moran, M. Fromer, D. Ruderfer, N. Solovieff, P. Roussos, C. O'Dushlaine, K. Chambert, S. E. Bergen, A. Kahler, L. Duncan, E. Stahl, G. Genovese, E. Fernandez, M. O. Collins, N. H. Komiyama, J. S. Choudhary, P. K. Magnusson, E. Banks, K. Shakir, K. Garimella, T. Fennell, M. DePristo, S. G. Grant, S. J. Haggarty, S. Gabriel, E. M. Scolnick, E. S. Lander, C. M. Hultman, P. F. Sullivan, S. A. McCarroll and P. Sklar (2014). "A polygenic burden of rare disruptive mutations in schizophrenia." Nature 506(7487): 185-190.

Robbins, T. W. (2000). "Chemical neuromodulation of frontal-executive functions in humans and other animals." Exp Brain Res 133(1): 130-138.

Rolls, E. T., M. Loh, G. Deco and G. Winterer (2008). "Computational models of schizophrenia and dopamine modulation in the prefrontal cortex." Nat Rev Neurosci 9(9): 696-709.

Seamans, J. K. and C. R. Yang (2004). "The principal features and mechanisms of dopamine modulation in the prefrontal cortex." Prog Neurobiol 74(1): 1-58.

Sesack, S. R., V. A. Hawrylak, C. Matus, M. A. Guido and A. I. Levey (1998). "Dopamine axon varicosities in the prelimbic division of the rat prefrontal cortex exhibit sparse immunoreactivity for the dopamine transporter." J Neurosci 18(7): 2697-2708.

Soden, M. E., S. M. Miller, L. M. Burgeno, P. E. M. Phillips, T. S. Hnasko and L. S. Zweifel (2016). "Genetic Isolation of Hypothalamic Neurons that Regulate Context-Specific Male Social Behavior." Cell Rep 16(2): 304-313.

Spear, L. (2000). "Modeling adolescent development and alcohol use in animals." Alcohol Res Health 24(2): 115-123.

Stagkourakis, S., G. Spigolon, P. Williams, J. Protzmann, G. Fisone and C. Broberger (2018). "A neural network for intermale aggression to establish social hierarchy." Nat Neurosci 21(6): 834-842. Sul, J. H., S. Jo, D. Lee and M. W. Jung (2011). "Role of rodent secondary motor cortex in value-based action selection." Nat Neurosci 14(9): 1202-1208.

Sul, J. H., H. Kim, N. Huh, D. Lee and M. W. Jung (2010). "Distinct roles of rodent orbitofrontal and medial prefrontal cortex in decision making." Neuron 66(3): 449-460.

Sulzer, D., S. J. Cragg and M. E. Rice (2016). "Striatal dopamine neurotransmission: regulation of release and uptake." Basal Ganglia 6(3): 123-148.

Tseng, K. Y. and P. O'Donnell (2007). "Dopamine modulation of prefrontal cortical interneurons changes during adolescence." Cereb Cortex 17(5): 1235-1240.

Vander Weele, C. M., C. A. Siciliano, G. A. Matthews, P. Namburi, E. M. Izadmehr, I. C. Espinel, E. H. Nieh, E. H. S. Schut, N. Padilla-Coreano, A. Burgos-Robles, C. J. Chang, E. Y. Kimchi, A. Beyeler, R. Wichmann, C. P. Wildes and K. M. Tye (2018). "Dopamine enhances signal-to-noise ratio in cortical-brainstem encoding of aversive stimuli." Nature 563(7731): 397-401.

Vijayraghavan, S., M. Wang, S. G. Birnbaum, G. V. Williams and A. F. Arnsten (2007). "Inverted-U dopamine D1 receptor actions on prefrontal neurons engaged in working memory." Nat Neurosci 10(3): 376-384.

Wen, Z., H. N. Nguyen, Z. Guo, M. A. Lalli, X. Wang, Y. Su, N. S. Kim, K. J. Yoon, J. Shin, C. Zhang, G. Makri, D. Nauen, H. Yu, E. Guzman, C. H. Chiang, N. Yoritomo, K. Kaibuchi, J. Zou, K. M. Christian, L. Cheng, C. A. Ross, R. L. Margolis, G. Chen, K. S. Kosik, H. Song and G. L. Ming (2014). "Synaptic dysregulation in a human iPS cell model of mental disorders." Nature 515(7527): 414-418.

Wood, J., Y. Kim and B. Moghaddam (2012). "Disruption of prefrontal cortex large scale neuronal activity by different classes of psychotomimetic drugs." J Neurosci 32(9): 3022-3031.

Ye, Y., S. Mastwal, V. Y. Cao, M. Ren, Q. Liu, W. Zhang, A. G. Elkahloun and K. H. Wang (2017). "Dopamine is Required for Activity-Dependent Amplification of Arc mRNA in Developing Postnatal Frontal Cortex." Cereb Cortex 27(7): 3600-3608.

Author Response

We thank the editors for their care in handling our manuscript. We also thank the reviewers, especially reviewer 2, for their thorough comments. We will work to address their concerns in a revised version and provide some initial comments below.

A major concern of two reviewers was that odour profiles were not quantified rigorously. We acknowledge that our study does not achieve the level of quantitative rigour standard in most chemical ecology work. We plan to conduct a few additional analyses to help address this shortcoming. We will also adjust the text to clarify the semi-quantitative nature of the data.

Reviewers also suggested using several different analytical approaches (e.g., different column, different sorbent) to broaden the type and number of detectable compounds. The reviewers rightly point out that such choices strongly affect which compounds we are likely to sample. No single approach is comprehensive, and ours is no exception. We will work to ensure that the appropriate caveats are included prominently in the text.

However, we believe this concern in fact underscores a special strength of our study: analysing the odour of a large number of species in a single study using the same analytical approach, so that the inherent biases of different approaches do not complicate cross-species comparisons. We are aware of very few such large-scale studies in any system and welcome suggestions from reviewers or readers of any we might have overlooked.

In general, we believe many of the reviewers’ methodological concerns reflect standards in the field of chemical ecology established for studies that aim to describe the odour of one or a few species as comprehensively as possible with a high level of quantitative rigour. This was not our goal, and we will temper our language in the revised paper to make that clear. Instead, we aimed to sample as broadly as possible across species to gain insight into the general statistics of a large 'odour landscape' or 'odour space' — an endeavour that, to our knowledge, is less common in the chemical ecology literature. In doing so, we prioritized breadth over depth. We believe the resulting dataset provides solid evidence for our major conclusions, though we will revisit our analyses and conduct a small number of additional experiments to further substantiate our claims.

Author Response:

We thank the reviewers and editors for their constructive and encouraging feedback on our manuscript. We have carefully studied the reviewer comments and found that we agree with almost all of them; we will implement these suggestions and prepare a revised submission. In particular, we will aim to address the reviewers’ valid concerns regarding metagenomic detection limits via a high-sensitivity re-analysis of the data based on metagenomic read mapping, orthogonal to our current analyses based on read mapping to mOTU single copy marker genes. Moreover, we will revise the manuscript text for clarity and streamline the phrasing on some observations and claims. We are confident that our work will improve as a result and look forward to future feedback and interactions.

Sincerely, for the authors,

Sebastian Schmidt & Peer Bork

Author Response

Joint Public review

The manuscript by Mitra and coworkers analyses the functional role of Orai in the excitability of central dopaminergic neurons in Drosophila. The authors show that a dominant-negative mutant of Orai (OraiE180A) significantly alters the gene expression profile of flight-promoting dopaminergic neurons (fpDANs). Among them, OraiE180A attenuates the expression of Set2 and enhances that of E(z) shifting the level of epigenetic signatures that modulate gene expression. The present results also demonstrate that Set2 expression via Orai involves the transcription factor Trl. The Orai-Trl-Set1 pathway modulates the expression of VGCC, which, in turn, are involved in dopamine release. The topic investigated is interesting and timely and the study is carefully performed and technically sound; however, there are several major concerns that need to be addressed:

1) In Figure S2E, STIM is overexpressed in the absence of Set2 and this leads to rescue. It is presumed that STIM overexpression causes excess SOCE, yet this is rarely the case. Perhaps the bigger concern, however, is how excess SOCE might overcome the loss of SET2 if SET2 mediates SOCE-induced development of flight. These data are more consistent with something other than SET2 mediating this function.

Our statement that STIM overexpression overcomes deficits in SOCE is based on the following published work:

1. Studies of SOCE in wildtype cultured larval Drosophila neurons demonstrated that overexpression of STIM raised SOCE to the same extent as co-expression of STIM and Orai in the WT background (Chakraborty et al, 2016; Figure 1D).

2. Both Carbachol-induced IP3-mediated Ca2+ release and SOCE (measured by Ca2+ add back after Thapsigargin-induced store depletion) were rescued in primary cultures of IP3R hypomorphic mutant (itprku) Drosophila neurons by overexpression of STIM (Agrawal et al., 2010; Figure 8A-G).

3. Deb et al., 2016 (Supplementary Figure 2h,i) reaffirmed that overexpression of STIM significantly improves SOCE after Thapsigargin-induced passive store-depletion in Drosophila neurons expressing IP3RRNAi.

4. Consistent with the cellular rescue of SOCE, defects in flight initiation and physiology observed in the heteroallelic IP3R hypomorphic background (itprku) could be rescued by overexpression of STIM (Agrawal et al., 2010; Figure 3A-E) as well as Orai (Venkiteswaran and Hasan, 2009; Figure 3).

5. In Figure S2E, we show that flight deficits arising from THD’> Set2RNAi are rescued upon overexpression of STIM (i.e. THD’>Set2RNAi; STIMOE). Here and in another recent publication (Mitra et al., 2021) we show that neurons expressing Set2RNAi exhibit reduced expression of the IP3R and reduced ER-Ca2+ release presumably leading to reduced SOCE. As mentioned above we have consistently found that STIM overexpression raises both IP3-mediated Ca2+ release and SOCE in Drosophila neurons.

In this study, we propose that Ca2+ release through the IP3R followed by SOCE are part of a positive feedback loop driving expression of Set2 which in turn upregulates expression of mAChR and IP3R (Figure 3F) to regulate dopaminergic neuron function. Our observation that loss of Set2 (THD’>Set2RNAi) can be rescued by STIM overexpression is consistent with this model because:

1. Loss of Set2 (THD’>Set2RNAi) results in downregulation of several genes including mAChR and IP3R leading to decreased SOCE.

2. As evident from our previous studies increased STIM expression in the Set2RNAi background (THD’>Set2RNAi; STIMOE) is expected to enhance SOCE which we predict would rescue Set2 expression leading to rescue of other Set2 dependent downstream functions like flight (Figure 2D).

2) In Figure 3, data is provided linking SET2 expression and Cch-induced Ca2+ responses. The presentation of these data is confusing. In addition, the results may be a simple side effect of SET2-dependent expression of IP3R. Given that this article is about SOCE, why isn't SOCE shown here? More generally, there are no measurements of SOCE in this entire article. Measuring SOCE (not what is measured in response to Cch) could help eliminate some of this confusion.

We will re-write this section in the revised version for better clarity and explain how Set2-dependent IP3R expression is an important component of Orai-mediated Ca2+ entry in fpDANs. Here, we propose that IP3-mediated Ca2+ release and SOCE, through Orai, are together part of a positive feedback loop driving transcription of Set2 which in turn upregulates mAChR and IP3R expression (Figure 3F). We hypothesized that the observed loss of CCh-induced Ca2+ response in the Set2RNAi background (Figure 3B-D; THD’>Set2RNAi) results from decreased itpr and mAChR expression and verified this in Figure 3E. This is further validated by the rescue of CCh-induced Ca2+ response and itpr/mAChR expression in the OraiE180A background upon Set2 overexpression (Figure 3B-E; THD’>OraiE180A; Set2OE). We were constrained to measure CCh-induced Ca2+ responses in OraiE180A expressing neurons for the following reasons:

1. SOCE measurements through Tg mediated store Ca2+ release followed by Ca2+ add back require a 0 Ca2+ environment that can only be achieved in culture. The Drosophila brain is bathed in hemolymph which contains Ca2+ and there do not exist any methods to readily deplete Ca2+ from the tissue to create a 0 Ca2+ environment without also effecting the health of the neurons.

2. Cultures of the subset of dopaminergic neurons (THD’) we have focused on in this study were not feasible due to the small number of neurons being studied from the total number of dopaminergic neurons in the brain (~35/400). In previous studies we have shown that SOCE post-Tg induced store depletion is abrogated in cultured dopaminergic neurons from Drosophila upon expression of OraiE180A (Pathak et al., 2015).

Furthermore, Carbachol-induced IP3-mediated Ca2+ release is tightly coupled to SOCE in Drosophila neurons (Venkiteswaran and Hasan, 2009) and Ca2+ release from the IP3R is physiologically relevant for flight behavior in THD’ neurons (Sharma and Hasan, 2020).

3) A significant gap in the study relates to the conclusion that trl is a SOCE-regulated transcription factor. This conclusion is entirely based on genetic analysis of STIMKO heterozygous flies in which a copy of the trl13C hypomorph allele is introduced. While these results suggest a genetic interaction between the expression of the two genes, the evidence that expression translates into a functional interaction that places trl immediately downstream of SOCE is not rigorous or convincing. All that can be said is that the double mutant shows a defect in flight which could arise from an interruption of the circuit. Further, it is not clear whether the trl13C hypomorph is only introduced during the critical 72-96 hour time window when the Orai1E180E phenotype shows up. The same applies to the over-expression of Set2 and the other genes. If the expression is not temporally controlled, then the phenotype could be due to the blockade of an entirely different aspect of flight neuron function.

The idea that Trl functions downstream of Orai-mediated Ca2+ entry in THD’ neurons is based on the following genetic evidence:

1. In Figure 4D, we show evidence of genetic interaction between trl-STIM and trl-Set2. The rescue of trl13c/STIMKO with STIM overexpression in THD’ neurons indicates that excess SOCE (driven by STIMOE) may activate the residual Trl (there exists a WT Trl copy in this genetic background) to rescue THD’ flight function. This is further supported by the rescue of trl/STIMKO with Set2 overexpression in THD’ neurons, which is consistent with the feedback loop model proposed in Figure 5C - where we propose that reduced SOCE leads to reduced ‘activated’ Trl and thus reduced Set2 expression, and the latter is rescued by SET2OEThe manner in which SOCE ‘activates’ Trl is the subject of ongoing investigations.

2. The trl hypomorphic alleles (including trl13C) exist as genetic mutants and they affect Trl function in all tissues throughout development. While we concede that these mutant alleles would affect multiple functions at other stages of development, which may impinge on the phenotypes noted in Figure S4B, we have used a targeted RNAi approach to validate Trl function specifically in the THD’ neurons (Figure 4C).

3. Overexpression mediated rescues (including Set2) were not induced only during the critical 72-96 hrs APF developmental window. Having established that Orai function drives critical gene expression during this window (Figure 1), it is reasonable to assume that Set2 rescue of loss of flight in OraiE180A occurs in the same time window where flight is disrupted.

4- In Figure 4, data is shown that SOCE compensates for the loss of Trl, the presumed mediator of SOCE-dependent flight. The fact that flight deficits are rescued by raising SOCE in the absence of Trl is very inconsistent with this conclusion.

We apologise for this confusion and will clarify in the revision. trl13c is a recessive allele of Trl and should be written as such throughout the text and in the figures (i.e trl13c and NOT Trl13c). In all cases of Trl mutant rescue by STIMOE and Set2OE there exists residual Trl that can be activated by excess SOCE thus leading to the rescue. This is true for trl13C/ STIMKO where each mutant is present as a heterozygote (the complete genotype of this strain is STIMKO/+; trl13c/+; this will be corrected in the revision). Similarly, for TrlRNAi we expect reduced levels (but not complete loss) of Trl. Thus the SOCE rescue of loss of Trl occurs in conditions where Trl levels are reduced but NOT absent. Homozygous trl null mutants are lethal.

5- In Figure 5 (A-C), data is provided that Trl transcripts are unaffected by loss of SOCE and that overexpression cannot rescue flightlessness. From this, the authors conclude that this gene "must" be calcium responsive. While that is one possibility, it is also possible that these genes are not functionally linked.

The idea that Trl is functionally linked to SOCE is based on the following evidence:

1. In Figure 4C we show that flight defects caused by partial loss of Trl (THD’>TrlRNAi) were rescued by STIM overexpression (THD’>TrlRNAi; STIMOE). As mentioned above we have found that STIM overexpression raises SOCE.

2. Heteroalleles of the trl13C hypomorph exhibit a strong genetic interaction with a single copy of the null allele of STIMKO as shown by the flight deficit of trl13c/+; STIMKO/+ (trl13C/STIMKO ) flies (Figure 4D). The genotypes will be corrected in the revision.

3. Flight defects in trl13C/STIMKO flies could be rescued by STIM overexpression in the THD’ neurons (trl13C/STIMKO; THD’>STIMOE)

4. In Figure 4E, we show that partial loss of Trl in THD’ neurons (THD’>TrlRNAi) leads to decreased expression of the Ca2+ responsive genes mAChR, itpr, and Set2 genes indicating that Trl is a constituent of the SOCE-driven transcriptional feedback loop (Figure 5C).

Since we could not detect a well-defined Ca2+ binding domain in Trl, we hypothesize that it could be activated by a Ca2+ dependent post-translational modification. Phosphoproteome analysis of Trl demonstrated that it does indeed undergo phosphorylation at a Threonine residue (T237; Zhai et al., 2008), which lies within a potential site for CaMKII. Independently, CaMKII has been identified as a binding partner of Trl from a Trl interactome study (Lomaev et al., 2018). Past work from our group (Ravi et al., 2018) identified a role for CaMKII in THD’ neurons in the context of flight. We are currently testing if CaMKII functions downstream of SOCE in THD’ neurons to mediate flight and will update this information in the next version of the manuscript.

6) There is no characterization of SOCE in fpDANs from flies expressing native Orai or the dominant negative OraiE180A mutant. While the authors refer to previous studies, as the manuscript is essentially based on Orai function thapsigargin-induced SOCE should be tested using the Ca2+ add-back protocol in order to assess the release of Ca2+ from the ER in response to thapsigargin as well as the subsequent SOCE.

The fpDANs consist of 16-19 neurons in each hemisphere (PPL1 are 10-12 and PPM3 are 6-7 cells; Pathak et al., 2015). Measuring SOCE from these neurons in vivo is not possible due to the presence of abundant extracellular Ca2+ in the brain. Given their sparse number, it proved technically challenging to isolate the fpDANs in culture to perform SOCE measurements using the Ca2+ add back protocol. Due to these reasons, we have relied upon using Carbachol to elicit IP3-mediated Ca2+ release and SOCE as a proxy for in vivo SOCE. In previous studies we have shown that Carbachol treatment of cultured Drosophila neurons elicits IP3-mediated Ca2+ release and SOCE (Agrawal et al., 2010; Figure 8). Moreover, expression of OraiE180A completely blocks SOCE as measured in primary cultures of dopaminergic neurons (Pathak et al., 2015; Figure 1E). Hence we have not repeated SOCE measurements from all dopaminergic neurons in this work. In the revised version we will explicitly state this weakness of our study and the reasons for it.

7) In the experiments performed to rescue flight duration in Set2RNAi individuals the authors overexpress STIM and attribute the effect to "Excess STIM presumably drives higher SOCE sufficient to rescue flight bout durations caused by deficient Set2 levels.". This should be experimentally tested as the STIM:Orai stoichiometry has been demonstrated as essential for SOCE.

The assumption that STIM overexpression drives higher SOCE is based upon previously published work from Drosophila neurons (Agrawal et al., 2010; Chakraborty et al, 2016; Deb et al., 2016) which demonstrates that excess WT STIM overcomes IP3R deficiencies (RNAi or hypomorphic mutants) to rescue SOCE. We agree that STIM-Orai stoichiometry is essential for SOCE, and propose that the rescue backgrounds possess sufficient WT Orai, which is recruited by the excess STIM to mediate the rescue. We will reference the earlier work to validate our use of STIMOE for rescue of SOCE.

Here, we propose that Set2 is part of a positive feedback loop driving transcription of mAChR and IP3R (Figure 3F). In keeping with this hypothesis, we posit that the phenotypes observed in the Set2RNAi background (Figure 2D) result from decreased itpr and mAChR expression (validated in Figure 3E). This is further validated by the Set2 overexpression mediated rescue of OraiE180A (Figure 2D) and rescue of itpr/mAChR expression in the OraiE180A background (Figure 3B-E; THD’>OraiE180A; Set2OE).

8) The authors show that overexpression of OraiE108A results in Stim downregulation at a mRNA level. What about the protein level? And more important, how does OraiE108A downregulate Stim expression? Does it promote Stim degradation? Does it inhibit Stim expression?

We hypothesize that changes in STIM mRNA observed in the THD’ > OraiE180A neurons stems from an overall reduction in IP3-mediated Ca2+ release and SOCE due to loss of Trl-Set2 driven gene expression detailed in our transcriptional feedback loop model (Figure 5C). We will attempt to explain this aspect more clearly in the next version of the manuscript. While we agree that measuring levels of STIM protein would be helpful, estimation of protein levels from a limited number of neurons (~35 cells per brain) is technically challenging. The STIM antibody does not work well in immunohistochemistry. In the absence of any experimental evidence we cannot comment on how expression of OraiE180A might affect STIM protein turnover.

9) Lines 271-273, the authors state "whereas overexpression of a transgene encoding Set2 in THD' neurons either with loss of SOCE (OraiE180A) or with knockdown of the IP3R (itprRNAi), lead to significant rescue of the Ca2+ response". This is attributed to a positive effect of Set2 expression on IP3R expression and the authors show a positive correlation between these two parameters; however, there is no demonstration that Set2 expression can rescue IP3R expression in cells where the IP3R is knocked down (itprRNAi). This should be further demonstrated.

The rescue of IP3R expression by Set2 overexpression in itprRNAi was demonstrated in a different set of Drosophila neurons in an earlier study (Mitra et al., 2021) and has not been repeated specifically in THD’ neurons. Similar to the previous study, here we tested CCh stimulated Ca2+ responses of THD’ neurons with itprRNAi and itprRNAi; SetOE (Fig S3), which are indeed rescued by SET2OE.

10) The data presented in Figure 3E should be functionally demonstrated by analyzing the ability of CCh to release Ca2+ from the intracellular stores in the absence of extracellular Ca2+.

CCh-mediated Ca2+ release from the intracellular stores in the absence of extracellular Ca2+ has been described in primary cultures of Drosophila neurons in previously published work (Venkiteswaran and Hasan, 2009; Agrawal et al., 2010) This work focuses on a set of 16-19 dopaminergic neurons in a hemisphere of the Drosophila central brain. It is technically challenging to generate a 0 Ca2+ environment in vivo, which is essential for measuring store Ca2+ release. Given their meagre numbers, primary cultures of these neurons is not readily feasible.

11) The conclusion that SOCE regulates the neuronal excitability threshold is based entirely on either partial behavioral rescue of flight, or measurements of KCl-induced Ca2+ rises monitored by GCaMP6m in DAN neurons. The threshold for neuronal excitability is a precise parameter based on rheobase measurements of action potentials in current-clamp. Measurements of slow calcium signals using a slow dye such as GCaMp6m should not be equated with neuronal excitability. What is measured is a loss of the calcium response in high K depolarization experiments, which occurs due to the loss of expression of Cav channels. Hence, the use of this term is not accurate and will confuse readers. The use of terms referring to neuronal excitability needs to be changed throughout the manuscript. As such, the conclusions regarding neuronal excitability should be strongly tempered and the data reinterpreted as there are no true measurements of neuronal excitability in the manuscript. All that can be said is that expression of certain ion channel genes is suppressed. Since both Na+ channels and K+ channel expression is down-regulated, it is hard to say precisely how membrane excitability is altered without action potential analysis.

The claim that SOCE influences neuronal excitability is based on the following observations:

1. Interruption of the transcriptional feedback loop involving SOCE, Trl, and Set2 through loss of any of its constituents, results in the downregulation of VGCCs (Figure 5G, 6H), which are essential components of action potentials.

2. OraiE180A mediated loss of SOCE in THD’ neurons abrogates the KCl-evoked depolarization response (Figure 6B, C) measured using GCaMP6m. We verified that this response requires VGCC function using pharmacological inhibition of L-type VGCCs (Figure 6E, F).

3. SOCE deficient THD’ neurons, which were presumably compromised in their ability to evoke action potentials could be rescued to undergo KCl-evoked depolarisation by expression of NachBac, which lowers the depolarization threshold (Figure 7C, D) or through optogenetic stimulation using CsChrimson (Figure 7F).

We agree that ‘neuronal excitability threshold’ is a precise electrophysiological parameter that has not been directly investigated here by measurement of action potentials. Therefore, references to neuronal excitability will be tempered throughout the revised manuscript and be replaced with a more generic reference to ‘neuronal activity’. In this context we propose to include further evidence supporting reduced excitability of THD’ neurons upon loss of SOCE in the revision.

Since one of the key functional outcomes of activity during critical developmental periods such as the 72-96 hrs APF developmental window identified in this study, is remodelling of neuronal morphology, we decided to investigate the same in our context. Neuronal activity can drive changes in neurite complexity and axonal arborization (Depetris-Chauvin et al., 2011) especially during critical developmental periods (Sachse et al., 2007). To understand if Orai mediated Ca2+ entry and downstream gene expression through Set2 affects this activity-driven parameter, we investigated the morphology of fpDANs, and specifically measured the complexity of presynaptic terminals within the 2’1 lobe MB using super-resolution microscopy. We found striking changes in the neurite volume upon expression of OraiE180A which could be rescued by restoring either Set2 (OraiE180A; Set2OE) or by inducing hyperactivity through NachBac expression (OraiE180A ; NachBacOE). These data will be included in the revised manuscript.

12) Related, since trl does not contain any molecular domains that could be regulated by Ca2+ signaling, it is unclear whether trl is directly regulated by SOCE or the regulation is highly indirect. Reporter assays evaluating trl activation upon Ca2+ rises would provide much stronger and more direct evidence for the conclusion that trl is a SOCE-regulated TF. As such the evidence is entirely based on RNAi downregulation of trl which indicates that trl is essential but has no bearing on exactly what point of the signaling cascade it is involved.

We agree that luciferase Trl reporters would provide a direct method to test SOCE-mediated activation. Future investigations will be targeted in this direction. Regarding possible mechanisms of Trl activation - since we could not detect a well-defined Ca2+ binding domain in Trl, we hypothesize that it may be phosphorylation by a Ca2+ sensitive kinase. Phosphoproteome analysis of Trl indicates that it does indeed undergo phosphorylation at a Threonine reside (T237; Zhai et al., 2008), which may be mediated by the Ca2+ sensitive kinase-CaMKII based on binding partners identified in the Trl interactome (Lomaev et al., 2018). Past work (Ravi et al., 2018) has indeed demonstrated a requirement for CaMKII in THD’ neurons for flight. We are currently testing whether CaMKII functions downstream of SOCE in these neurons to mediate flight, and will be updating this information in the next version of the manuscript.

13) Are NFAT levels altered in the Orai1 loss of function mutant? If not, this should be explicitly stated. It would seem based on previous literature that some gene regulation may be related to the downregulation of this established Ca2+-dependent transcription factor. Same for NFkb.

As mentioned in the text in lines (307-309), Drosophila NFAT lacks a calcineurin binding site and is therefore not sensitive to Ca2+ (Keyser et al., 2007). In the past we tested if knockdown of NF-kB in dopaminergic neurons gave a flight phenotype and did not observe any measurable deficit. From the RNAseq data we find a slight downregulation of NFAT (0.49 fold, p value=0.048) and NF-kb (0.26 fold, p value =0.258) the significance of which is unclear at this point. We did not find any consensus binding sites for these two factors in the regulatory regions of downregulated genes from THD’ neurons.

14) Does over-expression of Set2 restore ion channel expression especially those of the VGCCs? This would provide rigorous, direct evidence that SOCE-mediated regulation of VGCCs through Set2 controls voltage-gated calcium channel signaling.

Set2 overexpression in the OraiE180A background indeed restores the expression of VGCC genes (Figure 6H).

15) All 6 representative panels from Figure 3B are duplicated in Figure 4G. Likewise, 2 representative panels from Figure 5H are duplicated in Figure 6D. Although these panels all represent the results from control experiments, the relevant experiments were likely not conducted at the same time and under the same conditions. Thus, control images from other experiments should not be used simply because they correspond to controls. This situation should be clarified.

We regret the confusion caused by the same representative images for the control experiments. These will be replaced by new representative images for Figure 5H in the next updated version of the manuscript.

16) The figures are unusually busy and difficult to follow. In part this is because they usually have many panels (Fig. 1: A-I; Fig. 2, A-J, etc) but also because the arrangement of the panels is not consistent: sometimes the following panel is found to the right, other times it is below. It would help the reader to make the order of the panels consistent, and, if possible, reduce the number of panels and/or move some of the panels to new figures (eLife does not limit the number of display items).

The image panels will be rearranged for ease of reading in the next updated version of the manuscript.

17) As a final recommendation, the reviewers suggest that the authors a- Reword the text that refers to membrane excitability since membrane excitability was not directly measured here. b-Explain why STIM1 rescues the partial loss of flight in Set2 RNAi flies (Fig. S2E); and c- Explain how/why trl is calcium regulated and test using luciferase (or other) reporter assays whether Orai activation leads to trl activation.

a. Textual references to membrane excitability will be appropriately modified.

b. We have provided a detailed explanation for how STIM overexpression might rescue the phenotypes caused by Set2RNAi in Point 1. In short, these phenotypes depend upon IP3R mediated Ca2+ entry driving a transcriptional feedback loop. We relied upon past reports that STIM overexpression upregulates IP3R-mediated Ca2+ release and SOCE in Drosophila itpr mutant neurons (Agrawal et al., 2010; Chakraborty et al, 2016; Deb et al, 2016). We therefore propose that STIM overexpression in the Set2RNAi background rescues IP3R mediated Ca2+ release followed by SOCE, which drives enhanced Set2 transcription, counteracting the effects of the RNAi. We will explain this more clearly with past references in the next revision.

c. We have provided a detailed response to this comment in Point 12. Briefly, we agree that building luciferase reporters for Trl could be an ideal strategy to test for its responsiveness to SOCE and needs to be done in future. As an alternate strategy, we have looked at data from existing studies of interacting partners of Trl (Lomaev et al., 2017) and identified CamKII, which is both Ca2+ responsive (Braun and Schulman, 1995; Yasuda et al., 2022), and thus might activate Trl through a phosphorylation-switch like mechanism. Moreover, a previous publication identified a requirement for CamKII in THD’ neurons for Drosophila flight (Ravi et al., 2018). We are testing the ability of a dominant active version of CamKII to rescue THD’>E180A flight deficits and will include this information in the next version of the manuscript.

References

1. Agrawal N, Venkiteswaran G, Sadaf S, Padmanabhan N, Banerjee S, Hasan G. Inositol 1,4,5-Trisphosphate Receptor and dSTIM Function in Drosophila Insulin-Producing Neurons Regulates Systemic Intracellular Calcium Homeostasis and Flight. J Neurosci. 2010;30:1301-1313. doi:10.1523/jneurosci.3668-09.2010
2. Braun AP, Schulman H. A non-selective cation current activated via the multifunctional Ca(2+)-calmodulin-dependent protein kinase in human epithelial cells. J Physiol. 1995. 488:37-55. doi:10.1113/jphysiol.1995.sp020944
3. Chakraborty S, Deb BK, Chorna T, Konieczny V, Taylor CW, Hasan G. Mutant IP3 receptors attenuate store-operated Ca2+ entry by destabilizing STIM-Orai interactions in Drosophila neurons. J Cell Sci. 2016. 129:3903-3910. doi:10.1242/jcs.191585
4. Deb BK, Pathak T, Hasan G. Store-independent modulation of Ca2+ entry through Orai by Septin 7. Nat Commun. 2016. 7:11751. doi:10.1038/ncomms11751
5. Depetris-Chauvin A, Berni J, Aranovich EJ, Muraro NI, Beckwith EJ, Ceriani MF. Adult-specific electrical silencing of pacemaker neurons uncouples molecular clock from circadian outputs. Curr Biol. 2011. 21:1783-1793. doi: 10.1016/j.cub.2011.09.027.
6. Keyser P, Borge-Renberg K, Hultmark D. The Drosophila NFAT homolog is involved in salt stress tolerance. Insect Biochem Mol Biol. 2007. 37:356-362. doi:10.1016/j.ibmb.2006.12.009
7. Kilo L, Stürner T, Tavosanis G, Ziegler AB. Drosophila Dendritic Arborisation Neurons: Fantastic Actin Dynamics and Where to Find Them. Cells. 2021. 10:2777. doi:10.3390/cells10102777
8. Lomaev D, Mikhailova A, Erokhin M, et al. The GAGA factor regulatory network: Identification of GAGA factor associated proteins. PLoS One. 2017. 12:e0173602. doi:10.1371/journal.pone.0173602
9. Mitra R, Richhariya S, Jayakumar S, Notani D, Hasan G. IP3/Ca2+ signals regulate larval to pupal transition under nutrient stress through the H3K36 methyltransferase dSET2. Development. 2021. 148:dev199018. doi:10.1101/2020.11.25.399329
10. Pathak T, Agrawal T, Richhariya S, Sadaf S, Hasan G. Store-Operated Calcium Entry through Orai Is Required for Transcriptional Maturation of the Flight Circuit in Drosophila. J Neurosci. 2015. 35:13784-13799. doi:10.1523/jneurosci.1680-15.2015
11. Ravi P, Trivedi D, Hasan G. FMRFa receptor stimulated Ca2+ signals alter the activity of flight modulating central dopaminergic neurons in Drosophila melanogaster. Barsh GS, ed. PLOS Genet. 2018. 14:e1007459. doi:10.1371/journal.pgen.1007459
12. Sachse S, Rueckert E, Keller A, Okada R, Tanaka NK, Ito K, Vosshall LB. Activity-dependent plasticity in an olfactory circuit. Neuron. 2007. 56:838-50. doi: 10.1016/j.neuron.2007.10.035.
13. Sharma A, Hasan G. Modulation of flight and feeding behaviours requires presynaptic IP3Rs in dopaminergic neurons. Elife. 2020;9. e62297.doi:10.7554/elife.62297
14. Venkiteswaran G, Hasan G. Intracellular Ca2+ signalling and store operated Ca2+ entry are required in Drosophila neurons for flight. Proc Natl Acad Sci. 2009.106:10326-10331. doi: 10.1073/pnas.0902982106
15. Yasuda R, Hayashi Y, Hell JW. CaMKII: a central molecular organizer of synaptic plasticity, learning and memory. Nat Rev Neurosci. 2022. 23: 666-682 doi:10.1038/s41583-022-00624-2
16. Zhai B, Villén J, Beausoleil SA, Mintseris J, Gygi SP. Phosphoproteome Analysis of Drosophila melanogaster Embryos. J Proteome Res. 2008. 7:1675-1682. doi:10.1021/pr700696a

Author Response

Reviewer #1 (Public Review):

This manuscript describes conditions under which "Self-inactivating Rabies" (SiR) can be grown to limit mutations that would allow the virus to replicate in the absence of TEV protease. It is also shown that neurons directly infected with a non-mutated virus remain healthy and that the virus does not mutate in the brain in vivo. Remarkably there is nothing in the manuscript to address the obvious question that is raised by the observation that such mutations were occurring around the time of the initial description of circuit tracing with this virus. Can the transsynaptic tracing experiments in the absence of TEV expression (as described in their original Neuron paper) be replicated with SiR that is not mutated? This obvious omission suggests that the authors might have conducted such experiments and were unable to replicate their published results. It is imperative that the authors be forthcoming about whether they have conducted such experiments and what were the results. If they have not conducted such experiments, they should do them and include the results here. Regardless of the outcome, the results should be published. If they cannot replicate their results, then the reliability of the Neuron paper is in doubt.

How do the results presented here relate to the results published in the Neuron paper and why are they not definitive with respect to the utility of SiR? The original publication in Neuron presents results that do not appear to be plausible and are best explained by the possibility that some experiments described in that manuscript were conducted using mutated SiR. This became most apparent when shortly after the Neuron publication, the Tripodi lab shared SiR as well as TEV expressing cell lines for propagation with other labs. Several of those groups observed that when they progagated the SiR received from the Tripodi lab, there was a mutation that removed the linkage of the PEST targeting sequence to N. This would be expected to allow the virus to replicate and spread without the need for TEV protease to remove the PEST sequence - precisely the phenotype observed in the trans-synaptic tracing experiments described in the Neuron paper. In the Neuron paper, culture experiments showed that the N-PEST (SiR) rabies could not replicate in the absence of TEV. And additional experiments showed that the virus is not toxic to neurons directly infected. These are the same experiments that are replicated in this submission. But then (in the Neuron paper) comes the unlikely report that this virus can spread trans-synaptically in vivo, in the absence of TEV expression. An alternative explanation would be that the virus used for those experiments was mutated and that is why TEV expression was not needed. There are no experiments in the original Neuron paper that address this possibility. Specifically, the experiments in Neuron describing cell survival during trans-synaptic tracing are not adequate to rule this out. This is because the two timepoints during which neurons were counted correspond to an early time when labeled neurons would be expected to still be accumulating and a later time that might be past the peak and represent a time when many neurons have died. To quantify proportions of neurons that survive, it is necessary to follow the same neurons over time, as has been done to demonstrate that only about half of neurons infected with G-deleted rabies die (half survive). Until tests are conducted testing whether TEV expression is required to obtain trans-synaptic labeling with an SiR that is known to not be mutated, it is irrelevant whether mutations can be prevented under particular culture conditions. The utility of this virus depends on whether it can be used for trans-synaptic tracing without toxicity and this manuscript presents no experiments to address that. Further, the omission of such experiments is glaring, as it is difficult to imagine that they have not been attempted.

We thank the reviewer for giving us the opportunity to improve on this point. We have performed additional experiments to confirm the ability of revertant-free SiR virus to spread transsynaptically in vivo. Our data shows that non-mutated SiR spreads transsynpatically in the mouse brain when complemented with G. In addition, we also tested the effect of the addition of TEVp to the starter neuronal population and found that it can significantly improve spreading efficiency. These data confirm the transsynaptic spreading capabilities of unmutated SiR in line with our original report. Furthermore, the data show the enhancing effect on the spreading efficacy of supplementing TEVp to the starter cells, broadly in line with what was recently reported by Jin et al., 2023. We have discussed the implications of these findings and suggested future directions in the main text and discussion.

Additionally, for completeness, we also assessed the spread efficiency of the recently generated SiR-N2c (based on the CVS-N2c rabies strain) in presence and absence of TEVp. We found that SiR-N2c spreads significantly better in the BLA-> NAc circuit than the original SiR (based on the SAD-B19 strain), and that the same spreading efficiency is not achieved by complementing SiR-B19 with the G from CVS_N2c Rabies strain. Interestingly, we found only a very small effect of the addition TEVp to the starting cells on the number of transsynaptically labelled cells with SiR-N2c. We have discussed the implications of these findings in the main text and discussion.

Changes in the manuscript: We have updated Figure 1 with the addition of a 6-month time point and update the main text accordingly. The updated paragraph is provided here:

"Results, SiR transsynaptic spreading.

We then tested the ability of revertant-free SiR to trace neural circuits transsynaptically in the mouse brain. ΔG-Rabies vectors can be pseudotyped with the chimeric EnvA glycoprotein to selectively infect neurons expressing the TVA receptor, which is not endogenously expressed by mammalian cells (Wickersham et al., 2007b). We injected the nucleus accumbens (NAc) of CRE-dependent tdTomato reporter mice with an AAV expressing either TVA and the rabies G or TVA only. After 3 weeks, we re-injected the NAc with EnvA-pseudotyped revertant-free SiR-CRE or EnvA-pseudotyped SiR-G453X-CRE and assessed the CRE-dependent tdTomato expression presynaptically, in the basolateral amygdala (BLA). At 1 month post SiR injection, we detected no tdTomato+ cells in the BLA in TVA-only-injected animals, confirming the G-dependency for SiR transsynaptic spreading (Fig 5B-C). In contrast, as expected, transsynaptic spreading was apparent in the TVA+G condition. We observed similar numbers of presynaptically traced neurons in both SiR-CRE and SiR-G453X-CRE injected brains (169 ± 24 and 190 ± 36 tdTomato+ neurons, respectively; two-tailed t-test, P = 0.64; Fig 5B-C). However, tdTomato+ microglial cells were only detected in the SiR-G453X-CRE condition indicating the re-emergence of toxicity of the revertant mutants (Fig 5B). We also tested the effect of supplying TEV protease to the starting cells, as this has been suggested to be a necessary step to ensure transsynapitc spreading. While the previous experiments unambiguously show that TEVp is not necessary for the transsynaptic spreading of SiR, the injection of an AAV expressing TEVp in the NAc did lead to an increase in the number of transsynaptically labelled BLA neurons (366 ± 69 tdTomato+ neurons; two-tailed t-test, P = 0.04; Fig 5C), indicating that TEVp-dependent SiR reactivation in starter cells can improve its spreading (Jin et al., 2023).

We recently showed that a novel SiR-N2c vector, derived from the neurotropic CVS-N2c Rabies strain, displays enhanced transsynaptic spreading and improved peripheral neurotropism over the original SAD B19-derived SiR (Lee et al., 2023). Hence, for completeness, we compared the transynaptic spreading efficacty of EnvA-pseudotyped revertant-free SiR-N2c and the original SiR. SiR-N2c labelled a greater number of BLA neurons at 1 month p.i. than what was detected with SiR (1691 ± 112 tdTomato+ neurons traced by SiR-N2c; two-tailed t-test, P = 2x105; Fig 5D-E). Additionally, TEVp expression in the starter cells in SiR-N2c tracing experiments had a negligible effect on the overall transsynaptic spreading (1934 ± 135 tdTomato+ neurons traced by SiR-N2c in presence of TEVp; two-tailed t-test, P = 0.24; Fig 5D-E). Since the use of G from the CVS-N2c Rabies strain (G_N2c) has been shown to improve ΔG-Rabies (SAD-B19) retrograde tracing (Zhu et al., 2020), we tested if complementing EnvA-pseudotyped SiR with G_N2c in the NAc could increase its spreading. While we detected more BLA tdTomato+ neurons than in our previous experiments, complementing SiR with G_N2c still labelled less neurons than SiR-N2c, even when TEVp was provided to the starter cells (487 ± 164 and 844 ± 14 tdTomato+ neurons traced by SiR in absence or presence of TEVp, respectively; Fig 5D-E)."

Discussion

"ΔG-Rabies vectors are powerful tools for the dissection of neural circuit organization thanks to their ability to spread retrogradely to synpatically-connected neurons. Here, we show that EnvA-pseudotyped revertant-free SiR vectors effectively spread transsynpatically in the mouse brain. Importantly, the co-delivery of an AAV expressing TEVp in addition to G increase the number of traced neurons in presynaptic areas, likely due to the TEVp-dependent reactivation of SiR in vivo (Ciabatti et al., 2017), in line with recent results (Jin et al., 2023). This should be considered when planning transsynaptic tracing experiments using SiR. To improve SiR spreading efficiency, further studies should investigate the use of inducible TEVp, as we previously showed (Ciabatti et al., 2017), that could maximise spreading efficiency while minimising possible side effects of prolonged protease expression.

Interestingly, we found that the recently developed SiR-N2c vector, generated by applying the same proteasome-targeting modification to the genome of the CVS-N2c ΔG-Rabies strain (Lee et al., 2023), show a higher number of retrogradely labelled neurons compared to the original SiR (SAD-B19) (Fig 5). Additionally, the co-delivery of TEVp had a smaller effect on the number of neurons transsynaptically-traced by SiR-N2c. Interestingly, the gap in trassynaptic spreading efficacy between SiR (SAD-B19) and SiR-N2c could not be filled by complementing the SiR with the neurotropic G_N2c. This could be linked to a more efficient packaging of SiR-N2c by G_N2c (Reardon et al., 2016; Sumser et al., 2022) or by the particularly high speed of CVS-N2c strain propagation (~12hrs)(Callaway, 2008; Hoshi et al., 2005). These results point to SiR-N2c as the vector of choice for transsynaptic experiments."

Other comments:

"A recently developed engineered version of the ΔG-Rabies, the non-toxic self-inactivating (SiR) virus, represents the first tool for open-ended genetic manipulation of neural circuits." It is not clear what the authors intend to be claiming with respect to "open-ended genetic manipulation of neural circuits" but it is clear that this assertion is overblown. There are numerous tools that are available for genetic manipulation of neural circuits. This is not the first, won't be the last, and it is arguably not the best.

We have rephrased this sentence.

Changes in the manuscript: The updated paragraph and figure panel is provided here:

Abstract

"A recently developed engineered version of the ΔG-Rabies, the non-toxic self-inactivating (SiR) virus, allows the long term genetic manipulation of neural circuits."

"Interestingly, a fraction of tdTomato+ neurons survived in ΔG- Rab-CRE-injected brains, differing from what we observed when injecting ΔGRab-GFP, where no cells were detected at 3 weeks p.i. (Fig 3CD) (Ciabatti et al., 2017). " This is a known result (same as Chatterjee et al., 2018) with a known mechanism. GFP expression is not observed because the rabies virus transitions from transcription to replication resulting in the termination of GFP expression. But Cre-recombination of the genome permanently labels cells with TdTomato. This is how Chatterjee et al. demonstrated that half of the neurons infected with G-deleted rabies survive. They imaged cells and saw that the GFP disappeared but the cells marked by Cre-recombination and RFP expression remained healthy indefinitely. The consideration of this in the Introduction is strange. There is no reason to suppose that Cre expression would somehow protect cells from rabies infection and there is no need to propose any such mechanism to explain the observed results.

This consideration is a response to the suggestion, proposed in Matsuyama et al 2019, that the toxicity reduction observed in ΔG-Rab-CRE could be linked to the expression of Cre recombinase compared to a cytosolic protein.

"Here we show that revertant-free SiR-CRE efficiently traces neurons in vivo without toxicity in cortical and subcortical regions for several months p.i.."

This wording is disingenuous and appears to be intentionally misleading. "Trace" implies that circuits were traced by transynaptic labeling, which they were not.

To avoid any misunderstanding, we have now changed trace to infect.

Changes in the manuscript: The updated sentence is provided here:

Abstract

"Here we show that revertant-free SiR-CRE efficiently infect neurons in vivo without toxicity in cortical and subcortical regions for several months p.i.."

Reviewer #2 (Public Review):

The study by Ciabatti et al examined the mutation issue for self-inactivating rabies (SiR), which was found by other labs. The authors identified the mutations in the rabies genome and showed that this mutation occurred more frequently after multiple passage of production cell lines with suboptimal TEVp expressions. The authors further showed that such mutation did not accumulate in vivo and that SiR-labeled cells remained alive across longitudinal imaging in vivo.

In this study, the rabies genome is rigorously examined by sequencing many viral particles from independent preparations. The rabies with point mutation in the PEST domain is directly engineered for sequencing and infection test. Overall, the mutation issue is well addressed by the authors and the conclusions are well supported, but some more aspects of discussion and data analysis need to be extended for an easier production of SiR in a condition not that optimal.

1) The authors stated that one should produce SiR from cDNA in order to avoid the potential mutation in SiR. From a practical point of view, it would be much better to amplify the rabies from a stock virus directly in the production cell lines. Any discussion or exploration on this direction would be appreciated in the field.

We thank the reviewer for giving us the opportunity to improve on this point. We have added in the discussion a paragraph suggesting the number of passages to be used during production for the packaging cells and viral stocks, referring to the equivalent passage in our experiments.

Changes in the manuscript: The updated paragraph is provided here:

Discussion

"Notably, we found that TEVp activity inevitably decreases after several passages of amplification of HEK-TTG, thus fresh low passage packaging cells should always be used to produce SiR preparations. Our results suggest that stock for packaging cells should be made within a couple of passage after selection is established, and then used freshly defrosted to produce SiR viruses (equivalent to P0 cells in Fig 2B-C). Similarly, SiR supernatant stocks should be made directly from cDNA transfection and amplified for a maximum of 2 passages (equivalent to SiR P0 in Fig 2E) before being used for large scale SiR productions."

2) 6 passages of production cell lines are not that extensive. In Fig.2C, there was already some level of TEVp activity reduction at 2nd passage. It is not clear to me that how the TEVp activity reduction naturally happens. Is there some room to play around puromycin concentration to maintain high TEVp activity?

As mentioned in the previous point, we have added in the discussion a paragraph describing the recommended number of passages to be used during production of the packaging cells and viral stocks, referring to the equivalent passage in our experiments. We clarified that our starting P0 conditions for packaging cells and stock SiR viruses were equivalent to already amplified stocks ready for viral production, which would add only 1-2 passages.

Reviewer #3 (Public Review):

This paper is a response to the report by Lin et al., bioRxiv 2022 (DOI: https://doi.org/10.1101/550640) that mutations in the genome of SiR were identified, which could result in a canonical G-deleted Rabies virus.

Strengths:

First, the authors found that SiR production from cDNA leads to revertant-free viruses by analyzing a total of 400 individual viral particles obtained from 8 independent viral productions with Sanger sequencing. Next, they identified the molecular mechanisms of mutations in the SiR; they found that extensive amplification of packaging cells HEK-TGG leads to the selection of clones with suboptimal TEVp expression level, which leads to the accumulation of revertant mutants, where, as the authors discuss, the revertant mutants have a specific replication advantage. Based on these observations, the authors recommend producing SiR freshly from cDNA with low passage packaging cells. Lastly, the authors observed that SiR-infected hippocampal and cortical neurons can survive for longer periods of time than the neurons infected with revertant mutants or a canonical G-deleted Rabies virus by combining next-generation sequencing of RNAs isolated from infected tissue and 2-photon in vivo longitudinal imaging of infected cortical neurons. Together, these findings support the idea that the degradation of N by PEST-mediated cellular mechanism results in the self-inactivation of SiR as suggested in the original SiR manuscript (Ciabatti et al., Cell 2017). Thus, SiR remains a powerful viral tool for the chronic investigation of neuronal circuitry and function as long as the virus is prepared in a way the authors recommend.

Weaknesses:

While most of the findings are solid, some conclusions are not fully supported by the data presented. The authors need to address the following points: Reviewer #3

1) In Figure 3B-D, the authors concluded that SiR-CRE -infected cells did not show cell death in contrast to Rab-CRE and SiR-G453X, but it cannot be fully supported only by this experiment. The authors should consider the potential variance in infection efficiency in each experimental animal and show evidence of suppressed cell death. In addition, it needs to be confirmed that SiR-Cre is diminished in infected cells at later times. The authors should explain and address these concerns by conducting additional experiments, for example, cleaved caspase-3 staining and quantification of virus RNA levels in each time point as performed in their previous study Ciabatti et al., Cell 2017 (DOI: 10.1016/j.cell.2017.06.014).

We thank the reviewer for the suggestion and give the opportunity to strengthen our work. We have added an analysis of the rabies transcripts over time in SiR-infected hippocampi (Fig S4). The drastic decrease of SiR RNA, along with the finding that the numbers of tdTomato-positive cells remain comparable at each time points support the reduction in mortality in SiR infected cells. We have added this data and clarified this point in the text..

Changes in the manuscript: The updated paragraph is provided here:

Results: Difference in cytotoxicity between ΔG-Rabies, PEST-mutant SiR and SiR

"We detected no decrease of tdTomato+ neurons in SiR-infected hippocampi (4109 ± 266 tdTomato+ neurons at 1 week p.i.; 4458 ± 739 tdTomato+ neurons at 2 months p.i.; one-way ANOVA, F = 0.08, p = 0.92, Fig 3C-D) while only 44% of tdTomato+ neurons were detected in Rabies-targeted and 60% in SiR-G453X-targeted hippocampi at 2 months p.i. (1422 ± 184 at 1 week versus 624 ± 114 at 2 months p.i. for ΔGRab; one-way ANOVA, F = 11.55, p = 0.003; 3052+508 at 1 week versus 1829+198 at 2 months p.i. for SiR-G453X; one-way ANOVA, F = 4.27, p = 0.05; Fig 3C-D). Additionally, we confirmed inactivation of revertant-free SiR by analysing the decrease of Rabies transcripts in the infected hippocampi over times (Fig S4). These results support the lack of toxicity of SiR on the infected neurons, in line with our previous findings (Ciabatti et al., 2017). Moreover, these data confirm the requirement for an intact PEST sequence to sustain the self-inactivating behaviour of SiR and suggest that PEST-targeting mutations do not occur in vivo."

2) In Figure 3E-F, to ensure the long-term stability of SiR-Cre in the vivo mouse brain, authors conducted SMRT sequencing 1 week after the virus infection. To test the potential slow accumulation of mutations at 1-month and 2-month, the authors should perform the same experiment at these time points. Only when SiR-Cre was undetected at 1-month and 2-month, would it be reasonable to show only 1-week data, however, such data is not presented.

We thank the reviewer for the suggestion. We have added an analysis of the Rabies transcript in the infected Hippocampi showing a drastic decrease of SiR RNA over time. This result, along with the finding that similar numbers of tdTomato-positive cells are detected in the infected hippocampi over time, support our choice of an early time point to find emergence and accumulation of revertant mutations.

3) In figure 4, the authors used only 2 mice for this experiment, although this is one of the most important experiments to ensure SiR-infected cells stay alive for the long term in vivo animals. It should be confirmed whether the conclusion remains the same by increasing the number of animals.

While we understand why the reviewer put forward this suggestion, we believe that our choice of number of animals is appropriate as the investment in time and resources to adding further animals would not strengthen our conclusion (which we have indirectly assessed previously (Ciabatti et al 2017) and here in Fig 3). For completeness, we have added a Fig4_S1with the images of all the ROI at every time points used in Fig 4.

4) The legend in Table 3 doesn't match the contents.

We thank the reviewer for pointing this out, in response we have now updated Table 3.

Author Response

The following is the authors’ response to the original reviews.

Thank you for reviewing our manuscript " Energy Coupling and Stoichiometry of Zn2+/H+ Antiport by the Cation Diffusion Facilitator YiiP". After carefully considering the reviewer's comments, we have made substantial changes to the manuscript, which we believe is now much improved. In addition to clarifying various points raised by the reviewers, we have also added a variety of new data from both experimental and computational studies. We hope that these changes will satisfy the reviewers such that we can move forward towards finalizing the publication process.

New data added to this revision includes

• SEC profiles comparing D287A and D287/H263A before and after complex formation with Fab to illustrate formation of higher order oligomerization (Suppl. Fig. 6).

• Control trace from MST using Mg2+ to illustrate reproducibility (Suppl. Fig. 6).

• Results from MD simulation of D72A mutant to explore the Asp72-Arg210 salt bridge as a stabilizing element (Fig. 4)

• Analysis of cavities in WT and D70A_asym structures to illustrate occlusion of site A (Suppl. Fig. 13).

• In addition, we have redone MD simulations for YiiP with site B empty. These simulations were originally done (3 x 1 μs) with a modified version of the zinc dummy model and we have redone them (6x1 μs) using our previously published zinc dummy model to be strictly consistent with other simulations on holo, apo and D72A structures. The new results are qualitatively consistent with the previous simulations and our conclusions remain unchanged.

In addition, the text has been modified and several figures have been updated to address concerns of the reviewers as described below.

Although figures will ultimately be renumbered to conform with eLife formatting, they have retained their original numbers for this revision to prevent confusion, except that Suppl. Fig. 13 is a new figure added at the request of reviewer 2.

Reviewer #1 (Recommendations For The Authors):

I have only a few comments that might need clarification from the authors:

• If the unbinding of Zn2+ to site B triggers the occlusion (and maybe the OF state) and the external pH does not affect that binding, how is it prevented from being always bound to Zn2+ and thus occluded also while it should be transporting protons (B to C panels in Figure 5)? Are there some other factors that I am missing?

Our data shows that the affinity of site B is low (micromolar), especially relative to the concentration of free Zn within the cytosol (picomolar - nanomolar). Therefore, we would expect that site B is normally empty and that the resting state would be represented by panel D in Figure 5. An elevation of Zn concentration, or delivery of Zn to the transporter by some as yet uncharacterized binding protein, would initiate the cycle starting with panel E.

It is notable that the TM2/TM3 loop adopts a novel conformation in the occluded state, in which it extends to interact with the CTD (panel G in Figure 5). In this conformation, the Zn binding site is disrupted, thus preventing binding of additional Zn ions to the TM2/TM3 loop. Although we do not know how this loop behaves as the protein transitions to the outward-facing state (panels A & B), it is tempting to speculate that it retains the extended conformation until the protein returns to the inward-facing, resting conformation in panel D. This idea has been added to the revised manuscript (line 464).

In addition, we have added a sentence (line 507) to explicitly state our assumption that Zn only binds to site B in the IF state.

• I am not an expert on experiments, but the results for mutants that abolish site C are difficult to understand. For D287A/H263A, the SEC columns data suggest a population of higher oligomers. Still, for the D70A/D287A/H263A and D51A/D287A/H263A, they showed a native dimer. I understand your suggestion that the Fab induces the domain swap, but how do you explain the double mutant SEC column result? Please elaborate.

The unexpected behavior of site C mutants certainly introduces complexity into our study. Considering all the ins and outs of our analyses, we are confident that site C is a high-affinity site that is constitutively occupied and serves as a structural site to stabilize the architecture of the native homodimer. In the original submission, we included SEC profiles for D287A and D287A/H263A in Suppl. Fig. 4 as well as profiles for D70A/D287A/H263A and D51A/D287A/H263A in Suppl. Fig. 6. The former in Suppl. Fig. 4 characterize the complex between mutant YiiP and Fab (for cryo-EM), whereas the latter in Suppl. Fig. 6 represent YiiP in the absence of Fab (for MST). In the absence of Fab, the mutations do not alter the elution volume at ~12 ml, consistent with the conclusion that the native YiiP homodimer remains unperturbed. In the presence of Fab, mutations affect the SEC profile in two ways: a shift in the main peak to ~11 ml, and appearance of a subsidiary peak at ~10 ml. The shift of the main peak can be explained by formation of a complex between YiiP and Fab. Presence of the subsidiary peak - seen for D70A, D287A, and D287A/H263A mutants - can be explained by formation of a dimer of dimers (4 YiiP + 4 Fab), which could be isolated as a subpopulation of particles during the processing of cryo-EM images. For D70A and D287A, the individual dimers were unperturbed in this dimer-of-dimers. In fact, we used masking and signal subtraction to isolate the individual dimers and included them in the final reconstruction together with the more prevalent dimeric species (2 YiiP + 2 Fab).

The D287A/H263A-Fab complex behaved differently. The main peak of the SEC profile was shifted to 10 ml, indicating that a dimer of dimers was the prevalent complex; absence of a peak at 11 ml indicated that isolated dimeric complexes were no longer present in the solution. Furthermore, the subsidiary peak was at ~9 ml, indicating an even larger complex not seen in the other preparations. The appearance of particles in cryo-EM images were distinct from the other mutants (e.g., compare 2D classes shown in panels C and D in Suppl. Fig. 4). 3-D structures revealed dimer-of-dimers with the domain swap as well as larger linear oligomers. Although not well resolved due to preferred orientation, it appears that these linear oligomers consist of a propagated domain swap.

We have included some new data to bolster our conclusion that, although the D287A/H263A mutant destabilized site C, Fab binding was responsible for inducing the domain swap. The new data, presented in Suppl. Fig. 6, shows an SEC profile for a preparation of D287A/H263A both before and after formation of the complex with Fab. In addition to including this new data, we have amplified our description of these SEC profiles under the heading "Zn2+ binding affinity" in the paragraph starting on line 289 to try to clarify this complex issue for the reader.

• Since in the D287A mutant, you are disrupting the preferred tetrahedral coordination of Zn2+, but it still binds, do you observe any waters that compensate for the missing aspartate? Maybe in the MD simulations?

Unfortunately, the resolution of the cryo-EM maps are not high enough to resolve water molecules that we assume are present at sites B and C. For the MD simulations, we did not use mutants, but simply removed Zn from each of the sites. So we are unfortunately not able to answer this question with the available data.

Reviewer #2 (Recommendations for The Authors):

1) It is no doubt that cryo-EM structures of four types of zinc-binding site mutants of a bacterial Zn2+/H+ antiporter YiiP provide important insight into distinct structural/functional roles of each of the binding sites. However, overall resolution of the cryo-EM maps presented in this paper is not high enough to address the Zn2+ coordination structures, the kinked TM5 segment seen in a D51A mutant, and the extended conformation of TM2/TM3 loop seen in the D70A asymmetric dimer. It would be better to highlight the density of the above regions and discuss the vitality of their structure models. Similarly, the presence of additional water molecules at sites B and C (line 117) do not seem convincing.

We are completely sympathetic with the recommendation of illustrating the map quality as thoroughly as possible. We hope that interested readers will download map and model from the respective PDB and EMDB repositories and see for themselves. Nevertheless, we have provided several new figure panels to illustrate explicitly the densities associated with the kinked TM5 segment in the D51A mutant (Suppl. Fig. 2) and the extended TM2/TM3 loop in the D70A mutant (Suppl. Fig. 5) and have referred to them at appropriate places in the text (line 128 and line 151). In Suppl. Fig. 5, we also included figure panels to show densities for this loop in WT and D287A/H263A mutants.

It is true that the maps are generally of insufficient resolution to clearly define the coordination of Zn. The relevant densities are shown for all sites in all mutants in Suppl. Fig. 2. Despite this shortcoming, the coordination geometry is well established by the previous, higher resolution X-ray crystal structure as well as by MD simulations. Each site is shown in the insets of Fig. 1b, c and d. The new cryo-EM densities and resulting models are consistent with this coordination, which we have now pointed out in the legend to Fig. 1. The important point is that the new cryo-EM maps document the occupancy of ions at the individual sites as well as the large scale conformational changes associated with this occupancy, which was the main goal of the study.

Finally, we agree that the presence of additional water molecules at the sites is not well supported; because this issue has little bearing on our analysis, these comments have been removed.

2) Identification of the occluded state in D70A asymmetric dimer is exciting, hence this reviewer recommends the authors to highlight the structure of this state more effectively in comparison with the IF/OF states. It would be better to show the side views of the superimposition between the occluded and IF/OF states, and the pore profile and radius in the TM domain of these three states. The authors should also show the density map of site A (including M2 and M5) in the occluded protomer of the asymmetric dimer in Suppl. Fig. 2. Additionally, the authors should include information regarding the cytosolic or periplasmic view in the legend of Figure 3A, B, D, F, G, and H.

As suggested, we have prepared a new supplemental figure juxtaposing the IF and occluded states and depicting differences in pore radius and accessibility of site A (Suppl. Fig. 13, initially referred to on line 152 and various other locations in the manuscript with methods described on line 680). However, we unfortunately do not have a structure in the OF state to complete this comparison.

The density map for site A including M2 and M5 of the occluded protomer is shown in Suppl. Fig. 2 in which density thresholds have been adjusted to show the helices.

We have updated the figure legend for Figure 2 (referred to as Figure 3 by the reviewer) with the orientation of view, which are all from the cytoplasm looking toward the membrane.

3) MST analyses using the YiiP mutants with a single Zn2+-binding site at different pH are useful, and the data interpretation in combination with computational approaches of CpHMD and MST inference are nice challenges, indeed. However, it may, in a sense, appear that the MD simulations have been carried out intentionally and/or forcibly so that the outcomes are compatible with the experimental MST data. Although this is not unusual or unacceptable, this reviewer is concerned that the determined pKa values of some residues, especially Asp residues at Site A, are unusually high. The validity of this outcome should be discussed from physicochemical viewpoint; what factors raise the pKa of Asp51 and Asp159 so high. In this context, the MST inference titration curve seems unusually steep for D159 (and H155), of which validity needs to be discussed. This reviewer is also concerned about the large variations per measurement in the MST experiments (Suppl. Fig 6 E, F, and G). Are such large variations common to this experiment? Optimization of the measurement conditions such as protein concentration, and/or increase of AlexaFluor-488 labeling efficiency might greatly improve the reproducibility per measurement. The authors should include information on which residue(s) is labeled with AlexaFluor 488 in YiiP (line 641).

One of the outcomes of our so-called MST-inference algorithm was the conclusion that protonation states for H155 and D159 were coupled. The basis for this conclusion is described in some detail in the Methods section (paragraph starting on line 1025) and results in cooperativity in the protonation state of these two residues. This cooperativity explains the unusually steep binding curve in Suppl. Fig. 10e. We added a couple of sentences to explain this result in the Results under "Zn2+ binding affinity", line 352.

There is indeed precedent for increased pKas of acidic residues based on experimental measurements for Glu and inferred for Asp, both in membrane proteins. Computational approaches similar to the ones we use (including some of our own earlier work) have also pointed to elevated pKas by 1-3 units for Asp residues. We included a paragraph in the Discussion of Stoichiometry and energy coupling (line 537) citing these references and explaining that such pKa shifts reflect strong Coulomb interactions of titratable residues in close proximity in the low dielectric environment of the membrane.

We believe there is a misunderstanding about our presentation of raw data for the MST experiments in Suppl. Fig. 6. Panels E, F and G show an overlay of data from the entire Zn titration, which is therefore expected to change according to the Zn concentration in each capillary. We have revised the corresponding legend to clarify the plots. We have also included traces from a Mg2+ titration as a negative control that better illustrates the reproducibility of these measurements.

The AlexaFluor dye contained the reactive NHS group which preferentially targets the N-terminus of the polypeptide chain. Although labeling of lysine side chains is possible, we do not expect much given the low labeling stoichiometry of ~1:1 used for our experiments. We updated the Methods section under MST experiments (line 689) with this information.

Reviewer #3 (Recommendations For The Authors):

By measuring the binding affinity of site A using the D70A mutant that retains site C at pH 5.6 is should be possible to verify if the affinity reported in Table 2 is affected by the quaternary structure of the system. The 40-fold difference in affinity between site A and site C at pH 5.6 should be sufficiently large to permit a meaningful measurement.

To address this suggestion, we have included additional data in Table 2 from the D70A/D287A mutant. Based on the cryo-EM structure of D287A, we expect that site C is still intact, which is why it was omitted from the original manuscript. However, the affinities measured at pH 6 and 7 are very consistent with those from the triple mutant (D70A/D287A/H263A), supporting the idea that complete abolishment of site C does not affect measurement of affinity at sites A or B. This additional data is presented in the section on "Zn2+ binding affinity" on line 304. We also note that the SEC profiles in the absence of Fab are consistent with formation of the native homodimer for all the mutants, as described in our response to reviewer 1 and now shown in Suppl. Fig. 6.

More details should be provided on the force field used for zinc(II) ions in MD simulations. Currently, there is only a reference to another article, where this info is in the caption of a supplementary figure.

We added a summary of our previous work to develop a non-bonded dummy model for Zn(II) on line 727 in the Methods section entitled "Overview of the MD simulations. However, we would like to point out that all details on the parameter development and the parameters themselves are stated in the Methods section “Classical force field model for Zn(II) ions” in our previous paper [Lopez-Redondo et al, J Gen Physiol 143 (2021)] and parameter files are available as package 2934 in the Ligandbook repository https://ligandbook.org/package/2934 .

We also realized that in the originally submitted version of this manuscript we reported “empty site B” simulations with an updated and experimental non-bonded Zn(II) dummy model that has close to experimental first-solvation shell water residence times but slightly worse solvation free energy. Although that does not really matter for these simulations because there was no Zn2+ ion in site B, we nevertheless performed a new set of 6 x 1 µs simulations with our published (J Gen Physiol 2021) Zn(II) model to make all simulations fully consistent with each other. The results remained qualitatively the same, with a lack of zinc ions in site B leading to increased flexibility in the TM2/3 loop and ultimately destabilization of the TMD-CTD interaction.

Author Response

The following is the authors’ response to the original reviews.

We sincerely appreciate the opportunity to revise the manuscript and the reviewers' critical comments and valuable suggestions. After carefully revising the manuscript, we strongly believe that the reviewers' comments are invaluable and will significantly enhance the quality of the manuscript and contribute to our future research. Following the reviewers' comments, we conducted a comprehensive and meticulous review, addressing each point individually and making extensive modifications and corrections. The responses to each question are provided in a point-by-point manner as follows:

Reviewer 1:

This study delves into the impact of imidacloprid, an insecticide documented for its toxicity towards honeybees, on the development of bee larvae. The investigation involved exposing bee larvae to various concentrations of imidacloprid, and observing the resultant effects.

The findings of this study revealed that imidacloprid exerted a dose-dependent delay in the development of bee larvae, marked by reductions in body mass, width, and an overall decline in the growth index. Moreover, at elevated concentrations, imidacloprid was observed to impair neural transmission, induce oxidative stress, inflict damage to the gut, and inhibit hormones and genes essential for development. The larvae were found to engage antioxidant defense systems and deploy detoxification mechanisms to mitigate these effects.

However, the manuscript could be significantly enhanced through several improvements. Firstly, the structure of the manuscript warrants refinement to foster coherence and clarity. Additionally, there is a need for careful reevaluation of the concentrations of imidacloprid employed in the study, to ensure their relevance and applicability. In terms of references, greater attention to accuracy in citation is imperative.

Furthermore, while the authors have provided an overview of the general effects of imidacloprid on both vertebrates and invertebrates, the inclusion of a more exhaustive literature review with a specific focus on honey bees and other insects would bolster the context and significance of this research. This would be particularly beneficial in the introduction section, which should be subjected to a major revision.

In summary, this study offers preliminary evidence of the detrimental effects of imidacloprid on the development of bee larvae by interfering with molting and metabolism. This research holds potential as a valuable resource for assessing the risks posed by pesticides to juvenile stages of various animal species.

On behalf of all the authors, I express our most sincere gratitude for your critical comments and suggestions. Following your suggestions, we have thoroughly reviewed and revised the entire manuscript, including the issues of imidacloprid concentration and citation accuracy you raised. More importantly, we have significantly revised the structure and content of the introductory section of the manuscript to include many more detailed reviews of critical literature, with particular attention to the overview of relevant research on honey bees, Drosophila, and other insects, to promote coherence and clarity of the introduction and to enhance the context and importance of this research. We hope that these changes meet with your approval. Overall, your valuable comments have greatly improved the quality of the manuscript and will facilitate our future research.

Q1: Line 48, "Adults exposed to high doses of imidacloprid experience", please provide a more precise value for the high doses.

Thank you very much for your comments. Following your suggestion, we have provided precise values for high doses of imidacloprid for adult exposure based on the study by Dr. Wu et al. 2001.

Q2: Line 82, There are several larvae effect reports using next generation sequencing approach. The authors should include those related references in this section.

Thank you for your comments. We have included relevant references in our revised manuscript.

Q3: Line 394, for the concentration design, the maximum concentration of imidacloprid used in this study is 377 ppb, which is from the imidacloprid residue level in beeswax. Bees don't consume beeswax, and the reference is wrong.

Thank you for raising this critical issue. As you point out, bees do not eat beeswax, but it is important to stress that this may well mean that the bee larvae themselves are exposed to higher doses. Therefore, in this study, we ultimately designed for the worst-case scenario of 377 ppb of imidacloprid residues in beeswax. We would like your agreement on this point. In addition, we have corrected the citation errors in the references here and included them in the revised manuscript.

Reviewer 2:

This study provides evidence on the ability of sublethal imidacloprid doses to affect growth and development of honeybee larva. While checking the effect of doses that do not impact survival or food intake, the authors found changes in the expression of genes related to energy metabolism, antioxidant response, and P450 metabolism. The authors also identified cell death in the alimentary canal, and disturbances in levels of ROS markers, molting hormones, weight and growth ratio. The study strengths come from applying these different approaches to investigate the impacts of imidacloprid exposure. The study weaknesses are not providing an in-depth investigation of the mechanisms behind the impacts observed and not bringing the results in light of the current literature. For instance, the authors' hypothesis is based on two main points, the generation of ROS that leads to gut cell death and energy dysfunction, and the increased P450 expression. They propose this increases P450 expression which in turn increases energy consumption and could contribute to developmental retardation. There is however no investigation on the mechanisms of ROS generation (it could be through mitochondrial damage, Nox/ Duox activity, NOS activity, P450s activity, etc). A link between higher P450 expression and increased energy consumption leading to energy deprivation is also missing. It would also be important for the authors to provide a more complete literature review as previous works have investigated imidacloprid sublethal dose impacts in larval stages for bees and other insect models.

I greatly appreciate your insightful comments and valuable suggestions on behalf of all the authors. Thank you for identifying the limitations of this study and providing valuable comments and suggestions. These comments and suggestions have significantly improved the quality of the paper and will facilitate our future research. Following your comments, we have revised and corrected the manuscript point by point. We hope that these corrections meet with your approval.

Q1: Abstract: It would be important to rephrase the abstract to make it clear when authors are talking about gene expression results or functional assays.

Thank you for your comment. Following your suggestion, we have revised the abstract to make it clearer, especially the description of the gene expression results. Please see lines 15-34 in our revised manuscript.

“Abstract Imidacloprid is a global health threat that severely poisons the economically and ecologically important honeybee pollinator, Apis mellifera. However, its effects on developing bee larvae remain largely unexplored. Our pilot study showed that imidacloprid causes developmental delay in bee larvae, but the underlying toxicological mechanisms remain incompletely understood. In this study, we exposed bee larvae to imidacloprid at environmentally relevant concentrations of 0.7, 1.2, 3.1, and 377 ppb. There was a marked dose-dependent delay in larval development, characterized by reductions in body mass, width, and growth index. However, imidacloprid did not affect larval survival and food consumption. The primary toxicological effects induced by elevated concentrations of imidacloprid (377 ppb) included inhibition of neural transmission gene expression, induction of oxidative stress, gut structural damage, and apoptosis, inhibition of developmental regulatory hormones and genes, suppression of gene expression levels involved in proteolysis, amino acid transport, protein synthesis, carbohydrate catabolism, oxidative phosphorylation, and glycolysis energy production. In addition, we found that the larvae may use antioxidant defenses and P450 detoxification mechanisms to mitigate the effects of imidacloprid. Ultimately, this study provides the first evidence that environmentally exposed imidacloprid can affect the growth and development of bee larvae by disrupting molting regulation and limiting the metabolism and utilization of dietary nutrients and energy. These findings have broader implications for studies assessing pesticide hazards in other juvenile animals”

Q2: Line 55-58: rephrase the sentences to make it clear that imidacloprid was not created in 1925, but only in the 90's.

Thank you for pointing out this error. We have corrected the citation. Please see the line 58 in our revised version.

Q3: Line 88: typo: " remain to be systematically investigated"

Thank you for pointing out this error. We have rewritten the sentence. Please see lines 121-122 in our revised manuscript.

Q4: Introduction is lacking important citations, a few of the important ones are: Farooqui 2013 (doi: 10.1016/j.neuint.2012.09.020.) - hypothesis linking neonic exposure, nAChRs receptors, and ROS in honeybees; Ihara et al 2020 (https://doi.org/10.1073/pnas.2003667117) - the targets of imidacloprid in honeybees; Martelli et al 2020 (https://doi.org/10.1073/pnas.2011828117) - mechanistic investigation of imidacloprid sublethal damage in Drosophila; Whitehorn et al 2018 (doi: 10.7717/peerj.4772) - investigation of imidacloprid sublethal dose impact on growth and development of butterflies; Chen et al 2021 (doi: 10.3390/ijms222111835) - sublethal effects of imidacloprid exposure on gene expression in honeybees at different life stages. It is important that the authors perform a more complete literature search to compare their work to previous ones, drawing conclusions and highlighting their novelties.

We greatly appreciate your insightful comments and valuable suggestions. Following your suggestions, we have made significant revisions to the structure and content of the Introduction section. We have incorporated the critical literature you provided and other relevant literature reviews, with a particular emphasis on studies of bees, fruit flies, and other insects. These revisions aim to improve the coherence, clarity, background, and significance of the Introduction. We hope that these modifications meet with your approval. Please see the red text in the Introduction section in our revised version.

Q5: Line 104: Explanation on the doses used should be included here, not later in the methods. Also, important to highlight that whereas the doses tested were found in bee products, they likely mean that the bees themselves were being exposed to even higher doses.

Thank you for your comment. Following your suggestions, we have moved the explanation of the imidacloprid doses used in this study to the Results section, as you mentioned. Please see lines 138-142 in our revised manuscript.

Q6: Line 112: It is important to identify the neuronal targets of imidacloprid in honeybees. Many are known. Some of the nAChRs targets were not investigated in this study (such as subunit alpha8 and beta1). Plus, is alpha2 an imidacloprid target? How does the expression of other nAChRs subunits compares? Importantly, these genes are expressed mostly in the nervous system, so a more correct approach would be a tissue specific analysis. The lack of tissue specific analysis is a consistent flaw throughout the methodological design.

Thank you very much for your important comment. Bees have more than ten nAChR subunit members. Imidacloprid inhibits acetylcholinesterase activity by competitively binding to acetylcholinesterase receptors. As you noted, this study did not investigate the expression of all nAChR subunits, including the alpha8 and beta1 subunits, in different tissues, which is a shortcoming of our study. We have always failed to make a technological breakthrough and cannot dissect to obtain important tissues from developing larvae alone. We have therefore had to abandon this design and use the whole larva as a sample for measurement. We are aware that this is a shortcoming of this research. In the future, we will make a breakthrough in technology and conduct a comparative analysis of all nAChR subunit genes in different organizations and developmental stages to obtain more comprehensive and accurate data. Thank you again for raising this important issue and for your valuable suggestions.

Q7 ~ Q9: Line 125: P450s expression may have opposite behavior when exposed to insecticides depending on tissue (such as brain and fat body). When checking whole larva gene expression, the tissue specific profiles become diluted and thus less reliable (for reference, check: https://doi.org/10.1073/pnas.2011828117); Line 131: Again, for the analysis of oxidative stress it would be important to investigate a tissue specific expression pattern and measurement of ROS markers. Investigating different time points during the exposure also adds to the mechanistic understanding. Do all tissues respond in the same way? In which tissue does an increase in ROS generation start? How? Does it spread to other tissues? By which mechanisms is it generated; Results in general: Tissue specific analyses and more time points can provide a better understanding of how sublethal imidacloprid doses impact growth and survival. Thinking about the doses of choice in light of what bees might be exposed is also important. The mechanistic understanding is missing in the paper, and without it the study does not add much in comparison to previous ones.

Thank you very much for your valuable comments. As you pointed out, the intensity of P450 detoxification and oxidative stress varies considerably between tissues. When checking whole larva gene expression, the tissue-specific profiles become diluted, which is detrimental to elucidating mechanisms. In this study, we encountered technical barriers in obtaining independent samples of specific tissues for anatomical sampling. As a result, we had to forego analysis of some specific tissues, including the tissue-differentiated analyses of P450 gene expression patterns and ROS markers that you mentioned. We only examined larval overall detoxification and antioxidant responses to imidacloprid toxicity. While we do not believe that data from specific tissues are fully representative of the complex overall picture of larvae, there is no doubt that the decision to study larvae as a whole does not contribute to our complete understanding of the mechanisms by which imidacloprid causes larval developmental retardation and larval responses to imidacloprid toxicity. In addition, the fact that this study only analyzed one-time points during imidacloprid exposure and did not design and comparatively analyze different time points limits our complete understanding of the above mechanisms. In summary, as you have pointed out, tissue-specific analyses and more time points could better understand how sublethal doses of imidacloprid affect growth and survival. In future studies, we will overcome the technical challenges and refer to your suggestions for further systematic and in-depth mechanistic studies specifically targeting imidacloprid toxicity in different tissues at different exposure times and incorporate your suggestions, such as whether the response is consistent across all tissues, the origin of the increase in ROS production, how it increases, whether it spreads to other tissues, and the underlying mechanisms into the next experimental design. Again, Thank you for your constructive and valuable comments, which have provided valuable insight for our study on mechanisms. Undoubtedly, these comments will enhance the innovativeness of our study and greatly facilitate our future research.

Q10: Line 236: The conclusion that mitochondrial dysfunction is taking place is not well corroborated. Are there changes in mitochondrial aconitase activity to suggest the mitochondrial origin of ROS? How do mitochondria look like under electron microscopy? Evidence for mitochondrial damage from functional assays? Could the ATP reduced levels be caused by increased consumption by other systems, instead of reduced production? Without functional assays to demonstrate mitochondrial dysfunction the indirect measurements of gene expression at most suggest expression perturbations in mitochondria for the point in time when gene profiles were examined.

Thank you for the comments. Based on the data of the present study, i.e., suppression of mitochondrial oxidative phosphorylation (COX17, NDUFB7) and expression of genes of its alternative glycolytic pathways (Gapdh, Oscillin), as well as a decrease in the ATP content, suggests that imidacloprid exposure leads to impaired energy metabolism in larvae and not to mitochondrial dysfunction. We have corrected this uncritical language presentation error. Please see the lines 267 and 275 red text in our revised version. We hope that this correction will meet with your approval.

Q11: Though not the aim of the study, an important step forward would be to investigate whether these doses that do not impact survival but cause growth retardation could affect the many stereotypical behaviors displayed by the worker bees when they reach the adult life. Without this sort of analysis, it is difficult to stablish whether the doses tested will impact the colony health.

Thank you very much for your valuable suggestions, which give us broader ideas for our subsequent, more in-depth work on the mechanism of toxicity. Inspired by your suggestion, we plan to conduct further studies to investigate the effects of different levels of imidacloprid exposure on the developmental process of bee larvae and the underlying mechanism of toxicity. We will also investigate the intrinsic link between this juvenile toxicity and behavioral and physiological defects in adult individuals.

Q12: Line 376: the authors do not provide a link to their hypothesis that increased P450, and antioxidant response is reducing larvae nutrient supply.

Thank you for your comment. I apologize for not fully understanding your point. If you mean that the hypothesis proposed in this study that increased P450 and antioxidant responses reduce larval nutrient energy supply is not well-founded, we have already addressed this in the previous paragraph. See Figure 7 and lines 395-399 for more details in our revised manuscript.

Q13: Line 393: Were the colonies single-cohort? Were the frames from different hives mixed together to create the experimental groups? Or each experimental group comes from a different frame/colony? This information is important to establish how much genetic variation might exist between the different experimental groups.

Thank you for your comment. In this study, the selected colonies were healthy and not exposed to pathogens or pesticides. Two-day-old larvae from the same frames of the same hive were individually transferred to sterile 24-well cell culture plates. The plates contained a standard diet containing royal jelly, glucose, fructose, water, and yeast extract. We have included the above text in our revised manuscript. Please see the lines 430-432 red text in the revised manuscript.

Author Response

The following is the authors’ response to the original reviews.

Please find enclosed our revised manuscript entitled “An unconventional gatekeeper mutation sensitizes inositol hexakisphosphate kinases to an allosteric inhibitor”. We would like to thank the editorial team and the reviewers for carefully reading the manuscript and for raising a number of valuable points. We have included additional data and discussion to address the questions raised. Please find the point-by-point responses below.

Reviewer #1:

1) While I understand that FMP-201300 is a tool (proof-of-concept) compound it would be useful to know if it has activity against IP6K1 (or IP6K2) in cells.

We were of course curious about this as well. Unfortunately, our attempts to generate cell lines in which IP6K1 or IP6K2 carry the gatekeeper mutation using CRISPR/Cas editing have not been successful so far. Nevertheless, to obtain information on the permeability and cellular activity of FMP-201300, we decided to treat wt cells, since the compound also inhibited IP6K1-wt and IP6K2wt at higher concentrations.

In a previous study, we could show that reduced intracellular 5PP-InsP5 levels lead to a decrease in rRNA synthesis (https://doi.org/10.1101/2022.11.11.516170). We now repeated this experiment with FMP-201300, along-side the known IP6K inhibitors TNP and SC-919, and could show that FMP-201300 it is able to reproduce this phenotype, strongly suggesting it is capable to diffuse through the cell membrane and act on IP6Ks. We have included this data as a new Figure (Figure S10) and in the discussion part of the manuscript.

2) Did the authors try docking studies to gain insight into the binding site of FMP-201300?

The reviewer raises an important point, and we indeed strongly considered docking studies during the progress of the project. However, given that the HDX-MS data show that the region around the αC-helix becomes much more flexible upon introducing the gatekeeper mutation, we were concerned that docking studies (which would be based on the static wt structure) may not accurately reflect the more dynamic state of the mutated IP6K.

Upon consulting with our colleagues with expertise in docking and molecular dynamics simulations, we believe that MD simulations would need to be performed to obtain a more realistic picture of this protein ligand interaction, which we would like to pursue in the future.

3) Regarding the SAR, it would be useful to know if both carboxylic acids are required for allosteric inhibition.

Given the available data, it appears very likely that both carboxylic acids are required for the inhibitor to unfold its potency. Compound A2, which only contained one carboxylate group, showed drastically reduced potency. We have altered the text in the main manuscript to get this point across more clearly.

4) It would be helpful if the authors presented a model for how they think the Leu210 to Valine mutation sensitizes IP6K1 to FMP-201300.

We agree that it is important to better visualize the structural factors that play a role in the sensitization towards the compound. We have generated a new Figure 5 (and the old Figure 5 is now Supplementary Figure 9), and added a section to demonstrate how we propose the mutation leads to the sensitization of IP6K1 to FMP-201300. For a better understanding, we have also included a depiction how the mutation already affects the apo structures. Furthermore, we have added some text in the HDX section, to better describe the proposed mechanism.

Minor:

1) Figure 4: The authors should use the same units in panels a and b.

Thank you for pointing this out, the figure was edited accordingly.

2) In the supplementary Excel file, it would be helpful to include a tab that contains a legend.

A contents page was added to help describe the layout of the supplementary Excel file.

Reviewer #2:

Overall, this is an excellent study of high quality. The identified FMP-201300 has the potential for further compound and probe development. My only minor comment is that the authors could spend more time discussing the proposed allosteric binding mode of FMP-201300 and provide more detailed figures to highlight the proposed interactions with the protein and the conformational changes that must ultimately take place to accommodate the allosteric modulator. I appreciate that the co-crystallization experiments did not yield bound inhibitor structures, but perhaps the authors could consider MD simulations to complete their study. However, that could be a story in itself and should not be a must for the publication of this great work.

We agree with the reviewer (and also reviewer 1) that it is important to better visualize the structural factors that play a role in the sensitization towards the compound. We have generated a new Figure 5 (and the old Figure 5 is now Supplementary Figure 9), and added a section to demonstrate how we propose the mutation leads to the sensitization of IP6K1 to FMP-201300. For a better understanding, we have also included a depiction how the mutation already affects the apo structures. Furthermore, we have added some text in the HDX section, to better describe the proposed mechanism. In brief, we propose that the mutation leads to increased flexibility of the region in the mutation, allowing accommodation of FMP-201300 and ATP. These same regions are also the regions that have large decreases in deuterium exchange upon addition of the inhibitor.

We also appreciate the comment about using computational methods, to predict the binding site (also a remark from reviewer 1). We strongly considered docking studies during the progress of the project. However, given that the HDX-MS data show that the region around the αC-helix becomes much more flexible upon introducing the gatekeeper mutation, we were concerned that docking studies (which would be based on the static wt structure) may not accurately reflect the more dynamic state of the mutated IP6K. As the reviewer points out, MD simulations would likely be needed to obtain a more realistic picture of this protein ligand interaction, which we would like to pursue in the future.

Author Response

The following is the authors’ response to the original reviews.

Reviewer #1 (Public Review)

Overall, I quite enjoyed reading the manuscript and found it very well-structured and organized. I congratulate the authors for building this nice research. I do have a few major points to raise, but probably they would not affect the general message of the manuscript.

Thank you for taking the time to review our manuscript and the positive feedback. Following your suggestions, we have corrected some mistakes and added clarifications and a few of the suggested quality checks on the models. However, we decided not to run new analyses as: i) we believe there would be minor changes to the general message of the manuscript; and ii) while some suggested analyses are compelling, they are difficult to implement for different reasons or are outside the scope of the paper (clarified below).

I was confused about how IUCN data were used. The IUCN predictors are not mentioned in the model equations presented in the manuscript, but their effect size is reported in Figure 2.

Thank you for highlighting this issue. This was a typo: we forgot to mention the variable in both equations 1 and 2. Changed accordingly.

In the manuscript Methods, it is said that IUCN data was classified into 3 categories. I believe there was a mix of mechanisms in measuring it this way since at least two processes might be underlying IUCN data. First, one can inspect whether there is an effect on "scientific/societal interest" for assessed vs non-assessed species. This would not have any relationship with the assessed status itself. Assessed species are any with LC, NT, VU, EN, CR, EW, EX statuses, whereas non-assessed species might include DD and NE. Second, one may observe an effect of threat status itself, with threatened species being more researched than non-threatened species, this would only be possible for assessed species, although there are methods out there to impute missing statuses. By inspecting Figure 2, I got the feeling that only the second option was explored, but this would need to be confirmed.

We couldn’t test the effect of single categories (LC, NT, VU, EN, CR, EW, EX) because observations within factor levels were unbalanced. So, we re-grouped the different categories into three levels: “Threatened” (EX, EW, CR, EN and VU), “Non-Threatened” (NT and LC), and “Unknown” (DD and NE) and only tested this variable (your second option). Note that the effect size of the level Unknown is not shown in Figure 2 as this is the reference category. This is clarified in the caption of Figure 2.

In Figure 2, I was confused about the presence of three categories of domain. In the text, it states that four categories have been used. I believe these domains are non-mutually exclusive, that's why there is a fourth category. Would it not be better to assess the influence of domain through three dummy variables (terrestrial, marine, freshwater), where multiple presences (1's) would indicate the "multiple" category?

We opted for a categorical variable (rather than a dummy) to have the same number of variables in the two groups (‘species’ vs ‘culture’). This is needed for the variance partitioning analysis (VPA), because an unbalanced number of variables in one group of a VPA can artificially inflate R2 (see, e.g., this source: https://www.davidzeleny.net/anadat-r/doku.php/en:varpart). As for Figure 2, the level “Multiple”, being the reference category, is not shown. This is clarified in the caption: “Baseline levels for multilevel factor variables are: Domain [Multiple]”.

At present, I felt that the spatial components of your data were unexplored. Since you have centroids representing species distribution, it could be interesting to explore the presence of the species within protected areas or biodiversity hotspots. That might be something triggering at least scientific interest. Also, one can derive information about the major habitat of species occurrence (either using IUCN Major Habitat classification) or extracting overlap of species centroids with WWF biomes (e.g., simplified to just forested vs non-forested habitats; https://ecoregions.appspot.com/). Another point very common to research exploring biodiversity shortfalls is the proximity to research institutions (https://doi.org/10.1111/2041-210X.13152). And since societal interest is also being explored, what about the proximity to major cities (doi:10.1038/nature25181). Finally, other metrics derived from species centroids could inform "tropicality", if the species is tropical or not. Most often, the tropics species are neglected in comparison with those occurring in temperate regions.

We thank the reviewer for this suggestion, and we are aware that there are important spatial drivers of interest as highlighted in earlier research. Indeed, the spatial aspects of the data were somewhat underexplored as a deliberate choice because we hope to carry out additional work to explore these aspects in more detail. Nevertheless, we included the centroid of each species range as a broad proxy of its distribution, to help explore, for example, the role of species latitudinal distribution in driving interest metrics. We have also considered the suggestions provided as additional analyses, but we find these may be challenging to implement with the current data for a few reasons. First, each species centroid was calculated based on GBIF occurrences and therefore represents the midpoint of all locations, but not necessarily an area that is known to be occupied by the species. Using the centroid to assess whether a species is located in a given biome or within protected areas using this approach would therefore be potentially misleading (for example, for some terrestrial species it may fall in the sea, and vice versa). Also, for the same reasons, taking the centroid to estimate the species accessibility or proximity to research institutions may be misleading. We find that while important, these spatial aspects require a more nuanced approach to be explored in detail.

I was also thinking about the influence of time on the models. Species described long ago are often more known to people and scientists and had more "time" to be researched. Although metrics of societal interest were restricted to the last decade here, that does not necessarily mean that peoples' interest is not affected by their accumulated experiences. Similar reasoning applies to scientific interests, which have a lengthier time frame (~80 years). That said, the year of description or time since description could be added to capture some metric of time.

This is a good point, which we discussed prior to running the analysis. Indeed, there is evidence that such accumulated experiences can drive species interest as our own research has also previously highlighted (e.g. see Ladle et al. 2017 doi: 10.1002/pan3.10053). However, we felt that comparing the date of description as a proxy of accumulated human experiences with species was only fair within major biological groups and not between them. This is because taxonomic practices, definitions, and methods vary widely between biological groups. We therefore decided not to include time since description as a variable driving the measures of scientific and societal interest in this work. Nevertheless, we recognize the importance of the history of such experiences in driving human interest in species, and the consequences emerging from the loss of such links, and have thus included a brief discussion of this topic in the manuscript (see lines 177-182).

Model residuals could be checked for phylogenetic or spatial autocorrelation. I am aware there is no phylogenetic tree used, but the hierarchical taxonomy could be used (Phylum / Class / Order / Family / Genus) as a proxy for phylogenetic relationship.

We agree. Indeed, the hierarchical taxonomy was already included as a random factor (Phylum / Class / Order) in eq. 1. Note that we excluded Family and Genus from the random structure because in most Phyla a single genus and family has been sampled (as well as due to model convergence problems).

Concerning the spatial autocorrelation, one could check whether model residuals and their respective coordinate centroids of each species range. It is stated that GLMM has been used to avoid these non-independence issues, but it would be interesting to check whether residuals remained free of them.

Good suggestion, although the use of centroids may not be the most appropriate since it is only a rough approximation of each species distribution (see previous answer). Still, out of curiosity, we checked whether the random factor on biogeography was enough to capture residual spatial autocorrelation in the models. For this, we used the R package DHARMa, which performs a Moran's I test for distance-based autocorrelation. Given that some coordinates were duplicated, we grouped residuals by biogeographic regions (DHARMa requires all coordinates to be unique). Neither the Web of Science nor the Wikipedia models had spatial autocorrelation in the residuals:

Web of Science model: observed = –0.20482, expected = –0.14286, sd = 0.10682, p-value = 0.561

Wikipedia model: observed = –0.180820, expected = –0.142857, sd = 0.055513, p-value = 0.4941

A last point, it would be interesting to provide some sort of inset plots, such as barplots or donut plots (within the current plots), showing the proportion of species with respect to major clades and biogeographical regions.

This is a good suggestion, but we couldn’t find a good way to show this as an inset. We added a barchart showing the number of species in each Phyla/Division in the supplementary materials (Figures S2C). As for the proportion of species in each region, we thought it would be redundant with Figure S1 (summarizing spatial information in sampled species).

Reviewer #2 (Public Review):

Using standard and widely used tools, the authors revealed the factors (cultural, phenotypic, phylogenetic, etc.) shaping societal and scientific interest in natural species around the globe. The strength of this manuscript (and the authors') lies in its command of the available literature, database and variable management and analysis, and its solid discussion. The authors thus achieved a manuscript that was pleasant to read.

Thank you for taking the time to review our manuscript and the positive feedback.

While I agree that doing a global study requires losing details of local patterns, maybe this is exactly the biggest shortcoming of the manuscript, oblivious to how different cultures (compare USA to PNG, for example) are reflected in these global patterns.

Related to this previous point, my only other comment is about using English as a reference of societal interest (i.e., the presence of a common name in English). While English may be widespread in Academia, it is still not that common in other societal circles, especially those not using Wikipedia for lack of internet access.

We acknowledge the limitation of this choice, as well as our limited capacity to represent specific cultural contexts with our approach. Our decision to consider only the existence of English common names as a variable was partly driven by practical reasons, and partly by the very factors the reviewer highlights. Indeed, many cultures, communities and social circles do not use English frequently and also do not use the internet frequently. One consequence of this is also that the information compiled for species in other languages is more restricted than that available in English, including the existence of vernacular names. In languages other than English, it may even be the case that several common language names exist in reference to the same species, and this number may be an even better reflection of their cultural importance, but sadly this information is not comprehensively indexed across languages and biological groups which prevented us from considering it. On the other hand, most species have been attributed English common names as part of legislative, scientific and other societal processes, and it is therefore likely that if they are important in any specific cultural setting, they will probably also have a vernacular English language name. Ultimately, while we recognize the potential limitations of this decision, we felt that considering English common names was the simplest and less biassed approach to represent the degree with which a species is individually recognized nowadays. We now better expose the reasons for the decision to consider only English common names, and the limitations associated with it in the manuscript (lines 178-193).

Author Response

eLife assessment

This study reports the fundamental discovery of a novel structure in the developing gut that acts as a midline barrier between left and right asymmetries. The evidence supporting the dynamics, composition, and function of this novel basement membrane in the chick is in parts solid and in others convincing, but the investigation of its origin and impact on asymmetric organogenesis is not yet conclusive. This careful work is of broad relevance to anyone interested in patterning mechanisms, the importance of the extracellular matrix, and laterality disorders.

We extend our sincere gratitude to the editors at eLife for their meticulous evaluation of our manuscript, as well as the valuable insights shared in this Public Review. We also wish to convey our appreciation to the reviewers for their thought-provoking suggestions, which we are enthusiastic about integrating into our revised work. In this provisional response, our primary focus is to address the two main concerns raised: the necessity for functional data to elucidate the importance of the barrier, and the imperative to resolve uncertainties regarding its origin. We are dedicated to addressing these important points, and believe they will greatly enhance the quality and significance of our manuscript.

Joint Public Review:

When the left-right asymmetry of an animal body is established, a barrier that prevents the mixing of signals or cells across the midline is essential. Such a midline barrier preventing the spreading of asymmetric Nodal signaling during early left-right patterning has been identified. However, midline barriers during later asymmetric organogenesis have remained largely unknown, except in the brain. In this study, the authors discovered an unexpected structure in the midline of the developing midgut in the chick. Using immunofluorescence, they convincingly show the chemical composition of this midline structure as a double basement membrane and its transient existence during the left-right patterning of the dorsal mesentery, which authors showed previously to be essential for forming the gut loop and guiding local vasculogenesis. Labelling experiments suggest a physical barrier function, to cell mixing and signal diffusion in the dorsal mesentery. Cell labelling and graft experiments rule out a cellular composition of the midline from dorsal mesenchyme or endoderm origin and rule out an inducing role by the notochord. Based on laminin expression pattern and Ntn4 resistance, the authors propose a model, whereby the midline basement membrane is progressively deposited by the descending endoderm.

Laterality defects encompass severe malformations of visceral organs, with a heterogenous spectrum that remains poorly understood, by a lack of knowledge of the different players of left-right asymmetry. This fundamental work significantly advances our understanding of left-right asymmetric organogenesis, by identifying an organ-specific and stage-specific midline barrier. The complexities of basement membrane assembly, maintenance, and function are of importance in several other contexts, as for example in the kidney and brain. Thus, this original work is of broad interest.

Overall, reviewers refer to a strong and elegant paper discovering a novel midline structure, combining classic but challenging techniques, to show the dynamics, chemical, and physical properties of the midline. However, reviewers also indicate that further work will be necessary to conclude on the origin and impact of the midline for asymmetric organogenesis. Three issues have been raised to strengthen the claims:

1) The function of the midline as a physical barrier requires clarification. Dextran injection here seems to label cells and not the extracellular space. By counting the proportion of dextran-labeled cells rather than dextran intensity itself, the authors do not measure diffusion per se, but rather cell mixing.

We agree that an additional means of showing the barrier function is important. We are currently addressing this using a fluorescently tagged derivative of the drug AMD3100 that we recently synthesized, per Poty et al. 2015. We previously showed that AMD3100 perturbs left sided CXCR4-dependent vasculogenesis when introduced on the left side of the dorsal mesentery (DM), but not when introduced on the right (Mahadevan et al. 2014). These data suggest that a midline barrier prevents diffusion of AMD3100 across the DM. We are currently characterizing the extracellular diffusion of this fluorescent derivative through the DM to complement our previous dextran data.

Additionally, we should emphasize that the dextran-injected embryos shown in Fig. 6 D-F were isolated two hours post-injection, a timeframe insufficient for cell migration to occur across the DM (Mahadevan et al., 2014). We also collected additional post-midline stage embryos ten minutes after dextran injections - too short a timeframe for significant cellular migration (Mahadevan et al., 2014). Importantly, the fluorescent signal in those embryos was comparable to that observed in the embryos in Fig. 6. Thus, we believe the movement of fluorescent signal across the DM when the barrier starts to fragment (HH20-HH23) is unlikely to represent cell migration. More than a decade of DNA electroporation experiments of the left vs. right DM by our laboratory and others have never indicated substantial cell migration across the midline (Davis et al., 2008; Kurpios et al., 2008; Welsh et al., 2013; Mahadevan et al., 2014; Arraf et al. 2016; Sivakumar et al., 2018; Arraf et al. 2020; and Sanketi et al., 2022). This is also shown in our current GFP/RFP double electroporation data in Fig. 2 G-H, and DiI/DiO labeling data in Fig. 2 E-G. Collectively, our experiments suggest that the dextran signal we observed at HH20 and HH23 is likely not driven by cell mixing.

2) The descending endoderm zippering model for the formation of the midline lacks direct evidence. The claim of an endoderm origin is based on laminin expression, but the laminin observed in the midline with an antibody may not necessarily correspond to the same subtype assessed by in situ hybridization.

We have attempted to address this important issue by introducing several tagged laminin constructs, LAMB1-GFP, LAMB1-His, and LAMC1-His, to the endoderm via DNA electroporation to try to label the source of the basement membrane. However, despite endogenous laminin production and export within the endoderm, there appeared to be no export of any of the tagged proteins to the endodermal basement membrane. This experiment was further complicated by the necessarily large size of these constructs at 10-11kb due to the size of laminin subunit genes, resulting in low electroporation efficiency. Although we have not yet determined an alternative way to directly test the endodermal origin hypothesis, we are committed to exploring specific methods to help us test this in future experiments.

The midline may be Ntn4 resistant until it is injected in the relevant source cells.

Ntn4 has been shown to disrupt both nascently assembling and preformed mature basement membranes (Reuten et al., 2016). As such, we feel that this particular membrane’s resistance to degradation is likely not predicated by its stage of assembly.

Alternative origins could be considered, from the bilateral dorsal aortae or the paraxial mesoderm, which would explain the double layer as a meeting point of two lateral tissues.

We agree that alternate origins of the midline basement membrane cannot be ruled out from our existing data. We have indeed considered the bilateral dorsal aortae and the paraxial mesoderm as possibilities. However, at the earliest stages of midline basement membrane emergence, the dorsal aortae are already significantly distant from the nascent basement membrane, as are the somites, which have not yet undergone epithelial-to-mesenchymal transition. Fig. S2 G provides an example of a very early midline basement membrane without dorsal aortae or somite contact. Because this particular image is from a section that is fairly posterior in the HH12-13 embryo, it is thus less developed in pseudo-time and gives a window on midline formation in even earlier stage embryos. This is in contrast to the spatially close relationship of the midline basement membrane with the notochord and endoderm. In the context of potential dorsal aortae contributions, it is worth noting that the basement membrane of vascular endothelial cells has a distinct composition from a non-vascular basement membrane. For example, vascular endothelial cells produce only alpha 4 and alpha 5 laminin subunits but contain no alpha 1 subunit in any known species (reviewed in DiRusso et al., 2017). Thus, endothelial cell-derived basement membranes would not contain the alpha 1 laminin subunit that we used in our studies as a robust marker of the midline basement membrane. Note in Fig. 3 E-H and J-J’’’ the absence of dorsal aortae labeling using our laminin alpha 1 antibody. The dorsal aortae are also richer in fibronectin, as seen in Fig. S2, while the midline ECM exhibits far less fibronectin staining. While it may be possible that the converging aortae compress the midline ECM into a more compact structure, we feel direct contribution of basement membrane components is unlikely.

3) The title implies a role of the midline in left-right asymmetric gut development. However, the importance of the midline is currently inferred from previously published data and stage correlations and will require more direct evidence.

We agree that we have not fully and directly demonstrated the extent of the role of the midline in enabling the asymmetry of DM compartments during gut development. We propose the following revised title: “An atypical basement membrane forms a midline barrier during left-right asymmetric gut development”. It is important to note that we have made diligent efforts to investigate the functionality of the midline basement membrane through various methods in which we are highly experienced. However, while targeting either the left or right side of the DM is relatively straightforward, accessing the midline presents substantial challenges. We attempted physical perturbation using in vivo laser ablation, but we observed no significant effect or stable disruption of the midline. Additionally, our attempts at ablation using diphtheria toxin proved to be too harsh on the endoderm, preventing reliable and consistent data interpretation. We have tried electroporating MMP9 and MMP2 into the DM, but these did not produce any appreciable effect on the midline. We are also concerned that directly injecting MMPs or other enzymes may lead to injection-related tissue damage to the embryo that may be difficult to separate from direct MMP digestion of the matrix. However, we firmly believe that our inference regarding the involvement of the midline ECM in the asymmetry of DM compartments is robust, based on the functionally distinct yet closely positioned cell populations of the DM, and the timing of the midline in relation to the establishment of these asymmetric compartments. Notably, recent research conducted in our laboratory has highlighted the vital necessity of maintaining the separation of diffusible signaling molecules, such as Bmp4, from these neighboring cell populations, which would otherwise be in direct contact if not for the presence of the midline basement membrane (Sanketi et al., 2022). We will continue developing specific methods to perturb the midline in preparation of a revised manuscript.

Author Response

The following is the authors’ response to the previous reviews

We thank the Reviewers and Editors for the evaluation of our revised manuscript.

We especially value the careful assessment of Reviewer 1; at the same time, we clearly disagree with the reviewer’s statement that the revised manuscript “is essentially unchanged”. As appreciated by the other Reviewers, we performed a key experiment (in our opinion the only conclusive experiment) to further solidify that FK506-treatment kills parasites in a FK506-independent manner. Of note, however, Reviewer 1 made us aware of an error in the legend of Figure 4F, which likely contributed to the confusion regarding the antiplasmodial effect of FK506: Unfortunately, we missed updating this legend to appropriately imbed the new experiment. We therefore incorrectly stated that parasites were exposed to FK506 for 48 hours after FK506 treatment at 4-10 hpi and 36-42 hpi in G1. In contrast to the experiments described in the initial submission, parasite survival was not measured 48 h later, but in G2 ring stage parasites, i.e. at a time point during which parasitemia is not affected by the knockout of PfFKBP35. We have now corrected this. As pointed out correctly by Reviewer 1, it would otherwise not be possible to disentangle the effects of the gene knockout and the drug. The setup we now present in Figure 4F, however, is clearly able to do so.

We apologize for the inaccuracy and hope this resolves the ambiguities regarding the FKBP35-independent antimalarial effect of FK506. In line with the comments of Reviewers 2 and 3, we believe that our findings on FK506 activity are of particular importance for the malaria research community. We therefore hope that the final eLife assessment will reflect this.

Author Response

The following is the authors’ response to the original reviews.

On behalf of my co-authors, we thank you very much for giving us the opportunity to revise our manuscript entitled “A positive feedback loop between ZEB2 and ACSL4 regulates lipid metabolism to promote breast cancer metastasis” (manuscript number: eLife-RP-RA-2023-87510).

We would like to convey our appreciation to you and the expert reviewers for your valuable time and effort in reviewing and improving our work. We are grateful for the constructive comments raised by the six expert reviewers. We have studied the reviewer’s comments carefully and have accordingly conducted additional experiments as recommended. We have made the following revisions point by point. We found that our work was substantially strengthened by addressing these points.

Reviewer #1 (Public Review):

In this study, Jiamin Lin et al. investigated the potential positive feedback loop between ZEB2 and ACSL4, which regulates lipid metabolism and breast cancer metastasis. They reported a correlation between high expression of ZEB2 and ACSL4 and poor survival of breast cancer patients, and showed that depletion of ZEB2 or ACSL4 significantly reduced lipid droplets abundance and cell migration in vitro. The authors also claimed that ZEB2 activated ACSL4 expression by directly binding to its promoter, while ACSL4 in turn stabilized ZEB2 by blocking its ubiquitination. While the topic is interesting, there are several major concerns with the study and its conclusions are not convincing.

1) Figure 1A, the clinical relevance or biological significance of drug-resistant luminal breast cancer cell lines with metastatic cancer is questionable. Additionally, the RNA-seq analysis lacked multiple test correction for differential gene expression analysis, and no fold-change cut-off was used, leading to incorrect thresholds and wrongly identified significant signals.

We appreciate the reviewer’s valuable questions to improve our manuscript. We identified many EMT related transcription factors such as ZEB2, SNAIL, TWIST, etc. was up-regulated in drug-resistant cells, so we hypothesized that drug-resistant cells may undergone EMT and acquire metastatic capability. The drug-resistant cells used in this study had already been proved and examined in the previous studies of our research team as follows:

(1) Zheng FM, Long ZJ, Hou ZJ et al., A novel small molecule aurora kinase inhibitor attenuates breast tumor-initiating cells and overcomes drug resistance. Mol Cancer Ther. 2014 Aug;13(8):1991-2003.

(2) Yang N, Wang C, Wang Z, et al., FOXM1 recruits nuclear Aurora kinase A to participate in a positive feedback loop essential for the self-renewal of breast cancer stem cells. Oncogene. 2017 Jun 15;36(24):3428-3440.

For the second question, we used the fold-change cut-off in RNA-seq analysis and the fold change was over 1.5-fold and the adjust P value is less than 0.05. To make it more clearly, we have reset the cut off with a |log2FC|2 and p<0.05 and generated the volcano Plot using R4.3.0 software for differentially expressed genes as follows in Author response image 1. The results showed 3217 and 3035 up-regulated genes in TAXOL-resistant and EPI-resistant cells respectively, along with 2427 (TAXOL) and 2901 (EPI) down-regulated genes. Both ACSL4 and ZEB2 were up-regulated in two cell lines. We have put the figure in the new supplementary Fig S2.

Author response image 1.

2) Figure 1D-E, the clinical associations between ACSL4 and ZEB2 overexpression and poor patient survival are not justified. The authors used an old web tool, the Kaplan-Meier plotter database, based on microarray data, to perform the analysis. The reviewer repeated the analysis and found that multiple microarray probes for ZEB2 were available, leading to opposite results when different probes were selected. The reviewer also repeated the analysis using more reliable TCGA RNA-seq data and found no correlation between ASCL4 or ZEB2 expression and post-progression survival.

We appreciate the reviewer’s thoughtful questions. The Kaplan-Meier plotter database (http://kmplot.com/analysis/) we used is handled by a PostgreSQL server, which integrates gene expression and clinical data simultaneously including GEO, EGA and TCGA data. We used auto-select best cutoff for the the Kaplan-Meier analysis. Due to the web tool is old, we repeated the Kaplan-Meier survival analysis using R4.3.0 software and split the patients in TCGA database according to the third quartile expression (new Fig. 1D-F). The results also show that patients with high expression of ACSL4 and/or ZEB2 have relatively worse prognosis as follows in Author response image 2 (p<0.01):

Author response image 2.

3) Figure 1I relied on IHC to support the negative correlation between ACSL4 and Erα expression, but the small sample size limits the power to establish the relationship and the results are not definitive without further replication or biological investigation. The authors should provide more detailed and comprehensive analysis, including appropriate statistical tests, to ensure the findings are robust and reliable.

We appreciate the reviewer’s suggestion. To better understand the positive correlation between ACSL4 and ZEB2 expression, we add up to 45 breast cancer cases for IHC analysis and the correlation is shown as follows in Author response image 3 (new Fig. 1 H):

Author response image 3.

4) Figure 3B-C lacks justification of the differences by showing only one field without any internal control for exposure. The reviewer suggests to show additional fields where cells with both efficiently and inefficiently knocked-down are present, to justify the robustness of the results. This can also be achieved by mixing control and knockdown cells.

We totally understand the reviewer's concern. Thank you for pointing out this problem. The lower magnification field of view is shown as follows and it includes both efficiently and inefficiently knocked-down cells. We have changed the Fig. 3B and C as follows in Author response image 4:

Author response image 4.

5) Figure 4A-D, oleate-induced cell migration is a well-documented feature across different cancer types. To make it more relevant to the current study, the authors should examine multiple cell lines with high and low ZEB2/ACSL4 expression to determine the underlying relevance.

We appreciate the reviewer’s comments and performed the suggested experiments. To better determine the role of oleic acid and ACSL4 on cell migration, we use MCF-7 cell line, which has low ZEB2/ACSL4 expression, to test the influence of oleate on the cell migration. Transwell and Wound healing assays revealed that oleic acid treated MCF-7 cells also exhibited enhanced invasive and metastatic capacities compared with control cells. This results indicates that oleate induces cell migration in MCF-7 cells may via mechanisms other than ACSL4. We have added the results to the new Supplementary Fig. 8 as follows in Author response image 5.

Author response image 5.

6) Figure 4E, it is difficulty to conclude that cancer cells utilize stored lipids during migration to fuel metastasis based on current data. Do you see any evidence of lipid signal decreasing in the leading edge of the scratch wound-healing migration assay? The authors should also compare signals between unmigrated and migrated cells in the transwell assay.

We appreciate the reviewer’s constructive suggestion. We performed the wound-healing migration assay and observed that the lipid signal was obviously decreased in the leading edge of the scratch, as shown in the Author response image 6 (New Fig. 4E). In the transwell experiment, the cells which migrated to the lower side of the chamber after 24 hours showed decreased lipid signals (Fig. 4F). All these results indicates that lipid is utilized during migration.

Author response image 6.

7) Figure 6 warrants a genome-wide ChIP-seq to justify direct regulation of ASCL4 promoter by ZEB2. The reviewer’s analysis of publicly available ZEB2 ChIP-seq in multiple cell types detected no ZEB2 binding signaling within {plus minus} 5 kb of ASCL4 promoter.

We thank the reviewer for the concern. We found that the breast cancer cells are not included in some data base, such as Cistrome Data Browser, which is a resource of human and mouse cis-regulatory information derived from ChIP-seq, DNase-seq and ATAC-seq chromatin profiling assays. Due to that different cell type may have totally different mechanisms, that’s why the ZEB2 binding signaling cannot be found within ASCL4 promoter in some cells.

We searched JASPAR data base (https://jaspar.genereg.net/), which is an open-access database of non-redundant transcription factor (TF) binding profiles, and found the consensus binding sequences (CACCT) of ZEB (zinc finger E-box binding homeobox) transcription family were within the 2kb of ASCL4 promoter as follows in Author response image 7.

Author response image 7.

8) Figure 7 presents a series of self-contradictory results. Figure 7C, why no significant change in ZEB2-MYC expression was observed in the presence of ACSL4 and/or HA-Ubi? In Figure 7 E&G, why robust ACSL4 expression is present in the control group in E but not in (G)? Additionally, why there is no degradation in ZEB2 baseline level over time in the shACSL4 group in E? These raise severe concerns about the data quality.

We appreciate the reviewer to point out these problems.

Response to question 1: In fig. 7C, we used 293T cell for the ubiquitin assay and it is not a breast cancer cells. The efficiency of over-expression is different between ZEB2 and ACSL4 in 293T cell lines.

Response to question 2: Because the expression of ACSL4 is low in MCF-7 and is high in MDA-MB-231 cells. In Figure 7E (New Fig. 7G), we used MDA-MB-231 cells for the control and ACSL4 knockdown cells. In Figure 7G (New Fig. 7I), we used MCF-7 cells for the control and ACSL4 over-expressed cells. We have also revised the figure legend of Fig.7 as follows:

I, The stability of ZEB2 protein was detected by CHX treatment assay in control or ACSL4 over-expressed MCF-7 cells. GAPDH was used as the internal loading control.

Response to question 3: Because knockdown of ACSL4 also significantly decreased the mRNA level of ZEB2 (New Fig. 7A), so the baseline levels of ZEB2 in the shACSL4 group (New Fig. 7G) were very low and degradation is not obvious.

9) Figure 7D, the IP result of ACSL4 is not justified as there is no enrichment of ACSL4 in the IP compared to input. With the current data, it is hard to justify that there is any direct interaction. Moreover, based on IF data in Figure 3B-C, ACSL4 is exclusively localized in the cytoplasm, while ZEB2 is exclusively localized in the nucleus. It is hard to believe there is any direct interaction and mutual regulation.

We appreciate the reviewer’s thoughtful questions. We have repeated the IP assay and found that the enrichment of ACSL4 was observed in the IP process and added to new Fig. 7E as follows in Author response image 8. We also repeated the immunofluorescence assay in the MDA-MB-231 cells. We observed that ZEB2 can also be found in the cytoplasm and co-localized with ACSL in some certain regions of the cytoplasm as follows in Author response image 9 (Supplementary Fig. S11):

Author response image 8.

Author response image 9.

Reviewer #2 (Public Review):

In this study, the authors validated a positive feedback loop between ZEB2 and ACSL4 in breast cancer, which regulates lipid metabolism to promote metastasis.

Overall, the study is original, well structured, and easy to read. Despite the reliability of the data discussed in this article, there are still some deficiencies that need to be addressed through further explanation.

Major issues:

1) The authors demonstrated that ACSL4 regulates ZEB2 not only via a post-transcriptional mechanism but also via a transcriptional mechanism. The authors have not provided a comprehensive explanation of the specific mechanism in this paper. Therefore, it is recommended that the author delve into the potential mechanisms in the discussion section. For example, related mechanisms affecting ZEB2 ubiquitination degradation, as well as factors affecting ZEB2 upstream transcriptional regulation, etc.

We appreciate the positive comments and constructive suggestion from the reviewer. We have added the following paragraph in the second paragraph of the discussion section :

Interestingly, our RNA-seq data revealed that some ubiquitin E3 ligases, such as FBXO4, UBE3C, NEDD4, RBX1 etc. were significantly reduced in ACSL4 knockdown cells (Fig. S12). This result indicated that ACSL4 may reduce the ubiquitin of ZEB2 via down-regulating ubiquitin E3 ligase. Additionally, we found that ACSL4 promoted ZEB transcription as the mRNA level of ZEB2 was significantly reduced after ACSL4 knockdown. A recent study reported that LD-derived lipolysis provide acetyl-CoA for the epigenetic regulation of gene transcription. We observed that ACSL4 can also promote FAO, which generates acetyl-CoA for the epigenetic regulation. It is likely that ACSL4 regulates the ZEB2 mRNA level via lipid-epigenetic reprogramming mechanism, which is worth studying in the future.

2) To further clarify the interaction of ZEB2 and ACSL4, it is best to perform in vitro glutathione-S-transferase (GST) pulldown assay and immunofluorescence assay.

We appreciate the reviewer’s suggestion. We performed GST pull-down assay to examine whether ZEB2 and ACSL4 form a complex. GST pull-down assay confirmed the interaction of ZEB2 and ACSL4 as follows in Author response image 10 (Supplementary Fig. S10). We also performed immunofluorescence assay and found that ZEB2 was co-localized with ACSL in some certain regions of the cytoplasm as follows in Author response image 11. (Supplementary Fig. S11):

Author response image 10.

Author response image 11.

3) In Figure 7B, the protein level of ZEB2 seems not to be altered in BT549 BCSC cell line after the depletion of ACSL4.

We appreciate the reviewer to point out this problem. The protein level of ZEB2 in BT549 BCSC cell is not abundant as MDA-MB-231. We repeated the experiment and found that ZEB2 was reduced after the depletion of ACSL4 in BT549. We have replaced the Fig.7B as follows in Author response image 12:

Author response image 12.

4) EMT is characterized by changes in cell morphology, so the staining of cytoskeletons with Phalloidin is needed.

We appreciate the reviewer’s suggestion and performed the staining. The results show that the ACSL4 knockdown cells had a significantly smaller length to width ratio, which indicates the reversion of EMT process, than those of the control group (p<0.05). We have put the results in Supplementary Fig. S4 as follows in Author response image 13:

Author response image 13.

5) Additional breast cancer cases or cohorts (such as TMA) should be used to validate the positive correlation between ACSL4 and ZEB2 expression through IHC analysis.

We thank the reviewer for the suggestion. To better understand the positive correlation between ACSL4 and ZEB2 expression, we added more breast cancer cases up to 45 for IHC analysis and validated the positive correlation between ACSL4 and ZEB2. We have put the results into Fig 1 H and I as follows in Author response image 14:

Author response image 14.

Reviewer #3 (Public Review):

The manuscript by Lin et al. reveals a novel positive regulatory loop between ZEB2 and ACSL4, which promotes lipid droplets storage to meet the energy needs of breast cancer metastasis. It is of interest, however, some concerns should be addressed to strengthen the finding.

Major concerns:

1) The effect of ZEB2 overexpression is not fully demonstrated in the whole study. This point should be addressed.

We appreciate the positive comments and constructive suggestion from the reviewer. We have performed ZEB2 over-expressed MCF7 cell line. Over-expression of ZEB2 significantly enhanced the metastatic and invasive capacities of MCF7 cells. (Supplementary Fig. S5A and 5B).

Author response image 15.

1. Does the addition of oleate restore the ability of migration or invasion in ACSL4 knockdown cells?

We thank the reviewer for the question. To address this point, the oleate was added in the culture medium of ACSL4 knockdown cells. As expected, the addition of oleate obviously restores the invasive and metastatic capacities of ACSL4 knockdown cells by 33.12% and 18.61% respectively. We have added the results in the new Fig. 4D as follows in Author response image 16:

Author response image 16.

3) Which cellular compartment does ACSL4 localize in and interact with ZEB2 to stabilize ZEB2?

We thank the reviewer for the question. We have repeated the immunofluorescence assay in the MDA-MB-231 cells. We observed that ZEB2 can also be found in the cytoplasm and co-localized with ACSL in some certain regions of the cytoplasm (Supplementary Fig. S11):

4) The ubiquitination assay and Co-IP assay are just performed in HEK293T cells. This result should be confirmed in MDA-MB-231 cells or Taxol-resistant MCF-7 cells.

We appreciate the reviewer’s suggestion. We performed the ubiquitination assay and IP assay in MDA-MB-231 cells as follows. The results confirm that knockdown of ACSL4 obviously enhanced the ubiqutination of ZEB2. We have put the results into Fig. 7D and 7F as follows in Author response image 17:

Author response image 17.

5) How does ACSL4 regulate ZEB2 at the mRNA level？Please verify.

We thank the reviewer for the thoughtful question. A recent study reported that LD-derived lipolysis provide acetyl-CoA for the epigenetic regulation of gene transcription. We observed that ACSL4 can promote FAO, which generates acetyl-CoA for the epigenetic regulation. It is likely that ACSL4 regulates the ZEB2 mRNA level via lipid-epigenetic reprogramming mechanism, which is worth studying in the future and we had added the following sentences into the second paragraph in the discussion section :

Additionally, we found that ACSL4 promoted ZEB2 transcription as the mRNA level of ZEB2 was significantly reduced after ACSL4 knockdown. A recent study reported that LD-derived lipolysis provide acetyl-CoA for the epigenetic regulation of gene transcription. We observed that ACSL4 can also promote FAO, which can generate acetyl-CoA for the epigenetic regulation. It is likely that ACSL4 regulates the ZEB2 mRNA level via lipid-epigenetic reprogramming mechanism, which is worth studying in the future.

6) In Fig. 2F, the silencing efficiency for ACSL4 and ZEB2 should be shown. In addition, the protein level of ZEB2 or ACSL4 in shZEB2 and shZEB2+ACSL4 groups should also be addressed.

We appreciate the reviewer's suggestions. We have added the protein levels in Fig 2F and 2H as follows in Author response image 18:

Author response image 18.

7) What is the survival status of patients with both high expression of ACSL4 and ZEB2 in TCGA. In addition, more survival data from databases especially patients with both high expression of ACSL4 and ZEB2 are needed to analyze to support the finding.

We thank the reviewer for the constructive suggestion. We repeated the Kaplan-Meier survival analysis of TCGA RNA-seq data by using R4.3.0 software. The survival data show that the patients with both high expression of ACSL4 and ZEB2 have the worst prognosis in the four groups (P<0.05) ( New Fig. 1D-F).

Reviewer #1 (Recommendations For The Authors):

10) Only one siRNA/shRNA was used for knockdown in one cell line. Different siRNAs/shRNAs and multiple cell lines should be used to rule out off-target effects.

We appreciate the reviewer’s suggestion. We have test three siRNA and shRNA for the knockdown efficiency (negative control siRNA or ACSL4 and ZEB2 siRNA were from the company of GenePharma), we chose one sequence with the best knock-down effect.

Author response image 19.

11) Western blot data are required to justify the overexpression or knockdown efficiency of ACSL4 in cells in Figure 2 A-C.

We thank the reviewer for the suggestion. we have added the following western blot data in Figure 2:

Author response image 20.

12) In Figure 1G, there is a huge variation of the protein input, which makes the results not justified. The authors should repeat the experiments to ensure consistency and reproducibility of the results.

We appreciate the reviewer to point out this problem. Because this is the tissue samples of breast cancer patients. The results are affected by the tumor tissue composition between different patient sample, and it is difficult to obtain fresh tissues. In our paper, paraffin specimens have been used for IHC staining, and the results confirmed that ACSL4 and ZEB2 are positively correlated. We have put the results in the supplementary data.

Reviewer #2 (Recommendations For The Authors):

1) Data from Figure 1A showed the EMT transcription factor SNAIL was also among the top upregulated genes. Please explain why the association between ACSL4 and ZEB2 was studied instead of ACSL4 and SNAIL.

We appreciate the reviewer’s question. We had calculated the correlation between the ACSL4 and SNAIL by Pearson’s correlation test. The correlation of ACSL4 and SNAIL is 0.33, less than that of ZEB2. Bedsides, the binding motif analysis reveals that the consensus sequence of ZEB transcription family is within the ACSL4 promoter. Thus, we investigated the relationship between ACSL4 and ZEB2 in breast cancer cells.

Author response image 21

2) What is the limitation of your study? Please add some relevant description in the part of discussion.

We appreciate the reviewer’s suggestion. We have added the description of limitation of our study in the last paragraph of discussion section as follows:

The limitation of this study is the clinical samples is only 45. The future study should expand the clinical samples and cases to provide more clinical evidence for the crucial role of ACSL4 in breast cancer metastasis.

3). In Figure 3 Figure Legends part, the authors used the word "knockout", which is a description error.

We appreciate the reviewer’s advice. We have corrected "knockout" into "knockdown".

Reviewer #3 (Recommendations For The Authors):

Minor concerns:

1) In line 352-353, the statement about whether the high or low expression of ACSL4 and ZEB2 or the advanced breast cancer affects prognosis is inaccurate.

We appreciate the reviewer to point out this problem. We have corrected the statement into “We found that patients with higher ACSL4 or ZEB2 expression, especially those with simultaneous high expression had worse prognosis than those with lower expression ”.

2) The title of the seventh part of your results contains a logical error that is opposite to the experimental conclusion.

We truly appreciate the reviewer to point out this problem. We have changed the title of the seventh part of results to “ACSL4 regulates ZEB2 mRNA expression and protein stabilization”.

Author Response

The following is the authors’ response to the original reviews.

Reviewer #1

1) Overall, this is a useful tool, the data is well-presented, and the paper is well-written. However, the predictions are only compared with two existing reconstruction tools though more have been recently published

The aim of this work was to facilitate high-throughput generation of strain-specific metabolic models e.g. at the scale of 100s -1000s as indicated throughput the introduction (see lines 74-82, 91-94), and therefore we only compared tools which were capable of high-throughput analysis via command line and excluded others (e.g. merlin). We have now tested this against the other recent command line tool, gapseq which had escaped our gaze. Thank you for bringing this to our attention. Additionally, we have included KBase (ModelSEED, a web-based app that does not support high-throughput analysis) to allow for readers to interpret the results in the context of community standard approaches, since KBase is a popular tool.

We have added an explicit statement about the choice of approaches, now at lines 194-199 as follows:

“We compared the output and performance of Bactabolize to the two previously published tools that can support high-throughput analyses i.e. CarveMe (30) and gapseq (31). To aid interpretation in the context of community standard approaches, we also include a comparison to the popular web-based reconstruction tool, KBase (ModelSEED), and a manually curated metabolic reconstruction of K. pneumoniae strain KPPR1 (also known as VK055 and ATCC 43816, metabolic model named iKp1289) (15).”

The methods section was updated accordingly (now lines 552-558):

“A draft model was generated using gapseq version 1.2 with the ‘doall’ command using the unannotated genome (as gapseq does not take annotated input files). Gap-filling was subsequently performed using the ‘fill’ command and a custom M9 media file to match the nutrient list found in Bactabolize (https://github.com/kelwyres/Bactabolize/blob/main/data/media_definitions/m9 _media.json).

Finally, a draft model was constructed using the annotated genbank K. pneumoniae KPPR1 file and the KBase narrative (15) (https://narrative.kbase.us/narrative/ws.14145.obj.1).”

The results section (now lines 193-308) and Figure 2 have also been updated / restructured to reflect the new analyses, and include a comparison of the relative compute times for the construction of models (lines 281-291) as follows:

“While model features and accuracy are essential metrics for comparison, computation time is also a key consideration for high-throughput analyses. We recorded the time required for each tool to build draft models for 10 of the completed KpSC genomes used in the quality control framework (see below) on a high-performance computing cluster (Intel Xeon Gold 6150 CPU @ 2.70GHz and 155 GB of requested memory on a CentOS Linux release 7.9.2009 environment. CarveMe KpSC pan was the fastest with a mean of 20.04 (range 19.90 - 20.18) seconds, followed by CarveMe universal at 30.28 (range 29.20 - 31.80) seconds, then Bactabolize KpSC pan at 98.05 (range 92.19 - 100.4) seconds. KBase took 183.50 (range 120.00 - 338.00) seconds per genome via batch analysis, including genome upload time and queuing. gapseq took 5.46 (range 4.55 - 6.28) hours to produce draft models (not including the required gap-filling), consistent with previous reports (37).”

Finally, the whole discussion has been updated and substantially restructured (lines 472-474, 475-493, 494-512). Specific mentions to the new analyses are at lines:

472-474: “Consistent with this assertion, our draft KPPR1 model constructed with KBase (without manual curation) was an outlier in terms of the very low number of genes, reactions and metabolites that were included.”

475-493: “CarveMe with universal model (30) and gapseq (31) are the current gold standard automated approaches for model reconstruction, and we show that a draft KpSC model generated by Bactabolize with the KpSC pan v1 reference resulted in similar or better accuracy for phenotype prediction (Figure 2). Both the CarveMe universal and gapseq models resulted in high numbers of true-positive and true-negative growth predictions. However, these were also accompanied by comparatively higher numbers of false-positive predictions that resulted in a lower overall accuracy for substrate usage analysis compared to Bactabolize with the KpSC-pan v1 reference (Figure 2), and comparatively lower precision and specificity for the gene essentiality analysis. False-positive predictions may indicate that the relevant metabolic machinery are present in the cell but were not active during the growth experiments (e.g. due to lack of gene expression). In this regard, false-positives are not always a sign of model inaccuracy. However, false-positive predictions can also occur from incorrect gene annotations e.g. due to reduced specificity of ortholog assignment resulting from the use of the universal model without manual curation. Given a key objective here is to facilitate high-throughput analysis for large numbers of genomes, it is not feasible to expect that all models will be manually curated, and therefore we believe that identifying fewer genes with lower overall error rates provides greater confidence in the resulting draft models. We also note that the BiGG universal reference model which CarveMe leverages is no longer being actively maintained. In contrast, user defined reference models can be iteratively curated and updated to incorporate new knowledge and data as they become available.”

510-512: “However, gapseq’s long compute time makes it inappropriate for application to datasets comprising 100s-1000s of genomes (such as have become increasingly common in the bacterial population biology literature).”

2) My understanding is that the tool requires a set of reference reconstructions for other strains of the target species. If no reference reconstruction is available for another strain of the target species, can this species not be reconstructed?

Any input reference can be used to generate models however, single strain models matching the target species, or ideally a species-specific panreference, are recommended for best results. We have added a discussion on these points at lines 128-133:

“For optimum results we suggest using a pan-model that captures as much diversity as possible for the target species or group of interest, because Bactabolize’s reconstruction method is reductive i.e. each output strainspecific model will include only genes, reactions and metabolites that are present in the reference or a subset thereof (although novel genes, reactions and metabolites can be added via manual curation).”

We expand on these points further in the discussion:

494-512: “Bactabolize’s reference-based reconstruction approach is reductive, meaning the resultant draft models will comprise only the genes, reactions and metabolites present in the reference, or a subset thereof, and will not include novel reactions unless they are manually identified and curated by the user. This is an important caveat that should be considered carefully for application of Bactabolize to large genome data sets, particularly for genetically diverse organisms such as those in the KpSC. For optimum results we suggest using a curated pan-model that captures as much diversity as possible for the target species or group of interest. While we acknowledge that a reasonable resource investment is required to generate a high-quality reference, we have shown that a pan-model derived from just 37 representative strains can be sufficient to support the generation of highlyaccurate draft models (Figure 2 and 5). Additionally, we note that it is possible to use a single strain reference model, which should ideally represent the same or closely related species to that of the input genome assemblies, in order to facilitate accurate identification of gene orthologs. It is technically possible to use an unrelated reference model, but this is expected to result in inaccurate and/or incomplete outputs and has not been tested in this study. In circumstances were no high quality closely-related reference model is available, we recommend alternative reconstruction approaches that leverage universal databases e.g. CarveMe (30) or gapseq (31). However, gapseq’s long compute time makes it inappropriate for application to datasets comprising 100s-1000s of genomes (such as have become increasingly common in the bacterial population biology literature).”

3) How do the reconstructions generated by Bactabolize compare to those generated by other reconstruction tools besides CarveMe and ModelSEED, e.g., gapseq (Zimmermann et al, Genome Biology 2021. 22:81) or merlin (Capela et al, Nucleic Acids Res 2022, 50(11):6052-6066?

See response to rev 1 point 1.

4) How are the accuracy, specificity, and sensitivity of the pan-models calculated? Is the compared experimental data on the species level?

We used the pan-model as a reference from which we generated a strain-specific model for K. pneumoniae KPPR1 (using Bactabolize and CarveMe). This strain-specific metabolic model was then used to simulate growth phenotypes and compared to published experimental data for KPPR1. This was described in the methods section, including the calculations for the metrics (lines 589-593); however, we have also expanded the description within the results section to clarify the approach (lines 201-209):

“De novo draft models for strain KPPR1 were built using; i) Bactabolize with the KpSC pan v1 reference; ii) CarveMe, with its universal reference model (CarveMe universal); iii) CarveMe, with KpSC-pan v1 reference (CarveMe KpSC pan); iv) gapseq; and v) KBase (ModelSEED). ….. Subsequently, each model was used to predict growth phenotypes; i) in M9 minimal media with different sole sources of carbon, nitrogen, phosphorus and sulfur; and ii) for all possible single gene knockouts in LB under aerobic conditions. The predicted phenotypes were compared directly to the published phenotype data.” [Note the published data are cited in the previous manuscript sentence, not shown here].

5) The link https://github.com/rrwick/GFA-dead-end-counter, in line 286 does not work.

Link regenerated – now at line 451-452 and 604

Reviewer #2

1) KpSC pan-metabolic reference model is provided. Are they required as input for Bactabolize? Are the gene, metabolite information open accessible by users? o See response to reviewer 1 point 2 above and;

All data for the KpSC pan-model described in this work are accessible in the model files and amino acid + nucleotide files + data table at https://github.com/kelwyres/KpSC-pan-metabolic-model. This is also linked in the manuscript at line 631 and in the Data availability statement at line 661.

2) In the results section "description of Bactabolize", the authors present technical details on how to generate a metabolic model. For the input and output, please provide concrete examples to show the functionality of Bactabolize.

Detailed instructions, example code and example input/output files are available via the Bactabolize GitHub repository: https://github.com/kelwyres/Bactabolize.<br /> Instructions and example code can be found on the wiki: https://github.com/kelwyres/Bactabolize/wiki Test data and example files are at: https://github.com/kelwyres/Bactabolize/tree/main/data/test_data

The Github repository is linked in the manuscript at lines 95, 124, 552, and 667, and we have added a further reference at line 124, which mentions the example code/data: “Full documentation, including example code and test data are available at the Bactabolize code repository (https://github.com/kelwyres/Bactabolize).”

3) To generate metabolic models, the authors present comparison results with other methods. However, the authors only present the numbers in genes, metabolites and substrates. Since the interactions between gene, metabolite, and substrate are also critical, if possible, please provide the coverage details about these interactions. Venn diagram is recommended to compare these coverage differences.

Two additional supplementary figures have been generated (Figures S5 and 6) showing Venn diagrams of metabolites and reactions for the highthroughput analysis approaches that are most relevant to this work (see also response to rev 1, point 1). These are discussed at lines 224-237:

“Figures S5 and S6 show the overlaps of metabolites and reactions between the high-throughput reconstruction methods after processing with MetaNetX (59) to standardise the reaction and metabolite nomenclatures (excluding CarveMe pan for simplicity and given the likely problems of reaction oversubscription). The majority of the reactions included in the Bactabolize model were conserved in either the CarveMe universal model (n = 1225, 53.2%), gapseq model (n = 54, 2.3%) or both (n = 665, 28.9%). The reaction overlap was skewed to the CarveMe universal model which shared 1225 reactions that were conserved in the Bactabolize model but absent from the gapseq model. Notably, the gapseq model contained a large number (2200) of unique reactions (70.4% of those in the model). Similarly, the vast majority of metabolites in the Bactabolize model were conserved in one or both of the other models (n = 917, 85.6%). However, it is likely that true overlaps between methods are underrepresented due to the different reaction identifiers and chemical synonyms used within the BiGG (Bactabolize, CarveMe) vs ModelSEED nomenclatures (gapseq), which are difficult to harmonise in an automated manner even after the application of MetaNetX.”

Figure 2 shows not only the model numbers but also includes benchmarking to real phenotypic data in 2DEFG as the key mode of comparison between models. This encompasses meaningful interactions between gene, metabolic and substrate. The results are discussed at length in text at lines 253-271:

“We assessed the performance of each model for in silico prediction of growth phenotypes compared to the previously published experimental data (15). Accuracy, sensitivity, specificity, precision and F1 scores were calculated (60). Note that the specific set of growth substrates and gene knockouts that can be simulated is determined by the sets of genes and metabolites captured by each model and is therefore model-dependent (Data S1 and S2). Among those with matched experimental phenotype data, the Bactabolize and CarveMe universal models were able to predict growth for a greater number of carbon, nitrogen, phosphorous and sulfur substrates than gapseq, CarveMe KpSC pan, KBase and iKp1289 models (Figure 2C, Data S1). While the CarveMe universal model had the highest number of truepositive growth predictions overall (n = 132 of 617 total predictions), it also had a comparably high number of false-positive predictions (n = 39 of 617 total predictions, Figure 2D). Similarly, the gapseq and iKp1289 models resulted in 31 (262 total predictions) and 50 (513 total predictions) falsepositive predictions, respectively. In contrast, the Bactabolize model had fewer false-positive predictions (n = 21 of 505 total predictions) alongside a high number of true-positive predictions (n = 117 of 505 total predictions), resulting in the highest overall accuracy metrics (Figure 2E, Data S1). The KBase model was a notable outlier, associated with a high number of falsenegative predictions (n = 31 of 103 total predictions) and low false-positive predictions (n = 3 of 103 total predictions), presumably resulting from the very low number of genes and reactions included in the model, driving low sensitivity and accuracy.”

Lines 272-280:

“The gene essentiality results showed that gapseq produced the highest absolute number of true-positive gene essentiality predictions (n = 79 of), followed by Bactabolize KpSC pan (n = 44 of 1220 total predictions), then CarveMe universal (n = 39 of 1951 total predictions). CarveMe universal had the largest number of true-negatives by a wide margin (n = 1599 of 1951 total predictions), followed by gapseq (n = 1085 of 1403 total predictions), then Bactabolize KpSC pan (n = 939 of 1220 total predictions), driving their high accuracies (83.96%, 82.96% and 80.57%, respectively). The Bactabolize model was associated with the greatest overall precision and specificity (Figures 2F & 2G) while the gapseq model resulted in the highest F1-score and sensitivity.”

4) Are quality control and gap-filling needed to be processed when constructing a new metabolic model?

Our goal here was to implement an approach to support high-throughput analyses (see response to rev 1 point 1), including leveraging draft genome assemblies as the bases for the construction of strain-specific metabolic models. As part of this work, we have described a robust quality control (QC) framework for screening draft K. pneumoniae genomes i.e. to identify genome assemblies that should not be used. We developed this framework by comparison to models generated for matched completed genomes. Our analyses demonstrate the importance of applying QC to the input draft genome assemblies. When appropriate QC is applied to the input genomes, the resultant draft models show a high degree of completeness compared to the matched models derived from complete genomes. The draft models can also be used to simulate growth phenotypes with high accuracy as compared to those simulated for the matched complete genome models.

No specific QC was applied to the draft models themselves, other than confirmation of positive growth prediction in m9 minimal media plus glucose (which is expected to support growth of all K. pneumoniae). In cases where the input assembly passed our QC criteria but the resultant model was unable to simulate growth in m9 minimal media plus glucose, gap-filling may be optionally applied. Again, by comparison to the simulated phenotypes from matched complete genome models, we show that these gap-filled draft models can produce accurate phenotype predictions. See lines 396-404:

“Of the 901 draft genome assemblies which passed our QC criteria (≤200 assembly graph dead ends), 23 of the resulting draft models failed to simulate growth in M9 minimal media with glucose (despite capturing ≥99% of the genes and reactions in the corresponding complete models). It is expected that all KpSC models should be able to simulate growth on M9 media with glucose as a sole carbon source, as this central metabolism is universal amongst KpSC. To replace missing, critical reactions required for growth on M9 with glucose, we investigated model gap-filling using the patch_model command of Bactabolize. We then assessed the accuracy of the gap-filled models for prediction of growth on the full range of substrates, as compared to the predictions from the corresponding complete models.” Lines 409-413: “Substrate usage predictions from the 21 successfully gap-filled models were highly accurate, with 18/21 having a prediction concordance of ≥99% across all 846 growth conditions (12/21 had 100% concordance) (Figure S9). We therefore conclude that models generated for genome assemblies passing our QC criteria, which have been gap-filled to successfully simulate growth on minimal media plus glucose, are suitable for the prediction of growth across a range of substrates.”

5) Are there any visualization results to check the status of the generated draft model?

No. This is a tool for large-scale and rapid production of metabolic models, and phenotype prediction and we have not included visualisation tools. Third party tools are available e.g. https://fluxer.umbc.edu/. We do provide optional generation of MEMOTE reports at lines 136-138:

“Draft genome-scale metabolic models are output in both SMBL v3.1 (41) and JSON formats (one pair of files for each independent strain-specific model), along with an optional MEMOTE quality report (42)”.

Reviewer #3

1) The justification and evaluation of the generated models are inadequate and onedimensional. The authors only focus on statistics such as the number of reactions and genes in the models, which does not accurately depict the completeness of the model.

The reviewer has misunderstood how we have used ‘completeness’ in this manuscript. In the section describing our novel QC framework, we use this term to refer to the relative completeness of draft models generated from draft genome assemblies as compared to curated models generated from complete genome assemblies for the same strains. The latter were considered as the ‘complete’ models for this purpose. We are not referring to any measure of network or metabolic pathway completeness. We specifically refer to gene and reaction capture compared to the ‘complete’ models because these features directly reflect the problem we are trying to address i.e. that draft genome assemblies may not contain the complete set of genes that are truly present in the underlying genome. We have updated the manuscript text to further clarify the problem we aim to address in this section and justify the use of gene and reaction capture metrics:

Lines 310-319: “There are now thousands of bacterial genomes available in public databases, the majority of which are in draft form, comprising 10s to 1000s of assembly contigs. This fragmentation of the genome is caused by repetitive sequences that cannot be resolved by the assembly algorithm and/or sequence drop-out. The latter can result in the loss of genetic information such that some portion of genes present in the underlying genome are lost from the genome assembly (either completely or partially). This in turn, poses a limitation for the reconstruction of metabolic models using these assemblies, since most published approaches use sequence searches to predict the presence/absence of genes and their associated enzymatic reactions. Therefore, if we are to use public genome data for high-throughput metabolic modelling studies, it is essential to evaluate the expected model accuracies and understand the minimum input genome quality requirements.”

The biological accuracy of the curated ‘complete’ models has been described previously, and this is now noted in the text at lines 320-324:

“Here we performed a systematic analysis leveraging our published curated KpSC models (n=37, (14)), which were generated using completed genome sequences and were therefore considered to represent ‘complete’ models for which the underlying genome sequence contains all genes that are truly present in the genome (note the biological accuracy of these models was reported previously (14) and is not the subject of the current study).”

Throughput the manuscript we not only compare models in terms of the numbers of genes and reactions, but through comparison of binary growth predictions. Specifically, in the Performance Comparison section (Bactabolize vs other approaches) we use comparison of predicted to experimental phenotypes for strain KPPR1 (see response to rev 1 point 4 for details). In the QC Framework section we compare the predictions derived from draft models generated from draft genome assemblies to those derived from the matched ‘complete’ models, and report the concordance as a measure of impact of input assembly quality (lines 309-394). In the final results section (Predictive accuracy of draft models), we generate 10 additional models and compare the growth predictions to matched experimental data (lines 414-433). We view these phenotype prediction comparisons as the ultimate measure of ‘completeness’ with which to assess our models, because these data have direct biological meaning.

2) The authors have not provided evidence or discussion on the accuracy of any metabolic fluxes, which are considered to be crucial for reconstructing metabolic models. Additionally, the authors have not mentioned the importance of non-growth associated maintenance and the criticality of biomass composition analysis, both of which significantly determine the fluxes in the system.

We acknowledge the importance of flux calculations and accurate biomass compositions when using genome-scale models to quantitatively predict growth rates. However, at this stage, the reconstructions developed using Bactabolize are intended for binary predictions and comparisons of growth capabilities on various substrates. The accuracies we report are based on measures of network completion (presence/absence of relevant reactions leading to growth or no-growth phenotypes) rather than specific growth rates. Thus, the models generated by Bactabolize can be used to explore diversity at the strain level in terms of growth capabilities and can serve as a scaffold for building detailed (customized biomass), strain-specific models. Measuring biomass composition and metabolic flux analysis require significant experimental comparisons that are outside the scope of the current study but could be performed for target strains based on reconstructions developed using Bactabolize.

3) It would be interesting to compare the accuracy of the models generated using Bactabolize with those manually curated.

We did exactly this. We compared the manually curated model iKp1289 as part of our benchmarking. Lines 194 – 199:

“We compared the output and performance of Bactabolize to the two previously published tools that can support high-throughput analyses i.e. CarveMe (30) and gapseq (31). To aid interpretation in the context of community standard approaches, we also include a comparison to the popular web-based reconstruction tool, KBase (ModelSEED), and a manually curated metabolic reconstruction of K. pneumoniae strain KPPR1 (also known as VK055 and ATCC 43816, metabolic model named iKp1289) (15).”

Unfortunately, as far as we aware there are currently no other published manually curated models for strains with matched phenotype data that are also not included as part of our pan-reference model (the latter is a key point to ensure a fair comparison of models generated using our pan-reference vs those generated with a universal reference).

4) The authors have not provided evidence or discussion on the accuracy of any metabolic fluxes, which are considered to be crucial for reconstructing metabolic models.

See response to rev 3, point 2.

5) The justification regarding the completeness of the models requires further discussion.

See response to rev 3, point 1.

6) A detailed discussion on the importance of manually curated models would significantly enhance the quality of the manuscript.

This has been added at lines 458-474:

“Traditionally, genome-scale metabolic reconstruction approaches have relied upon significant manual curation efforts. While there will always remain a need for high quality curated models, such resource intensive approaches preclude their application at scale, and have therefore limited analyses to small numbers of individual strains (15, 16). However, automated reconstruction approaches can support the generation and comparison of multiple strain-specific draft models from which meaningful biological insights can be derived (61). Additionally, the quality of curated models is likely to vary depending on their age, level and type of curation, as well as the approach used for preliminary drafting. Indeed it is possible for automated approaches to outperform manually curated models; a draft model for K. pneumoniae KPPR1 generated using Bactabolize with the KpSC pan-v1 reference model outperformed the manually curated iKp1289 model representing the same strain (15). iKp1289 was published in 2017 (6 years prior to this study) and was initially drafted via the KBase pipeline (33), which uses RAST to annotate the sequences with Enzyme Commission numbers. It has been demonstrated several times that the Enzyme Commission scheme has systematic errors (62, 63), leading to a loss in accuracy when compared to the ortholog identification methods used by automated approaches. Consistent with this assertion, our draft KPPR1 model constructed with KBase (without manual curation) was an outlier in terms of the very low number of genes, reactions and metabolites that were included.”

Author Response

The following is the authors’ response to the previous reviews

Reviewer #2 (Recommendations For The Authors):

1) Overall, the novel phylogenetic analyses presented are satisfactory. With this new piece of information in hand, I would suggest using maximum-likelihood analyses as the major evidence supporting ortholog annotations. In fact, it would be best advised to add the bootstrap support analyses (perhaps over new trees) to the phylogenies presented in the supplement.

Thank you for suggestion. Although it would make sense to present phylogenetic trees constructed by maximum-likelihood analyses, we decided to keep the original trees (for CDCA7 and HELLS) in supplemental figures for an aesthetic reason. For example, for CDCA7/zf-4CXXC_R tree made by maximum likelihood method *Hif2_data2_zf4CXXC_R1_iqtree.txt), it would have been easier to visualize if the plant CDCA7 clade was positioned at the bottom, not the top, of the tree, as the topology was identical in both cases. Unfortunately, as the calculated result randomly put plant CDCA7 clade at the top, plant CDCA7 clade appears to be separated from the clades representing the rest of CDCA7 homologs. While we could manually adjust this in the final drawing, we wanted to avoid that.

2) There are still a few places in the main text where RBH - and is associated E-value - is used as evidence of orthology. As mentioned in my original review, this is evidence for homology, not orthology. Please make sure to amend the final text (for example in the first paragraph of the result section).

We concurred and amended the manuscript following this recommendation.

3) We agree with reviewer 1 that part of the functional considerations outside of the human and frog example should be softened, or clearly labelled as an hypothesis - which is now supported by this interesting study

I assume that this is related to Introduction of CDCA7. As this study defined CDCA7 homologs in result section We believe that this point has been addressed in our last revision.

4) In addition, make sure to indicate in the main text state the point about DNMT3 nomenclature (w.r.t. DRM).

In page 10, we added a sentence below to clarify this point.

“In this report, we call a protein DNMT3 if it clusters into the clade including metazoan DNMT3, plant DNMT3, and DRM.”

Author Response

The following is the authors’ response to the original reviews.

eLife assessment

This important manuscript reveals signatures of co-evolution of two nucleosome remodeling factors, Lsh/HELLS and CDCA7, which are involved in the regulation of eukaryotic DNA methylation. The results suggest that the roles for the two factors in DNA methylation maintenance pathways can be traced back to the last eukaryotic common ancestor and that the CDC7A-HELLS-DNMT axis shaped the evolutionary retention of DNA methylation in eukaryotes. The evolutionary analyses are solid, although more refined phylogenetic approaches could have strengthened some of the claims. Overall, this study should be useful for researchers studying DNA methylation pathways in different organisms, and it should be of general interest to colleagues in the fields of evolutionary biology, chromatin biology and genome biology.

We sincerely appreciate constructive comments and suggestions by the reviewers and a fair and accurate summary by the monitoring editor. Below we made point-by-point responses to reviewers’ comments.

Reviewer #1 (Public Review):

Overall, I find the work performed by the authors very interesting. However, the authors have not always included literature that seems relevant to their study. For instance, I do not understand why two papers Dunican et al 2013 and Dunican et al 2015, which provide important insight into Lsh/HELLS function in mouse, frog and fish were not cited. It is also important that the authors are specific about what is known and in particular about what is not known about CDCA7 function in DNA methylation regulation. Unless I am mistaken, there is currently only one study (Velasco et al 2018) investigating the effect of CDCA7 disruption on DNA methylation levels (in ICF3 patient lymphoblastoid cell lines) on a genome-wide scale (Illumina 450K arrays). Unoki et al 2019 report that CDCA7 and HELLS gene knockout in human HEK293T cells moderately and extremely reduces DNA methylation levels at pericentromeric satellite-2 and centromeric alpha-satellite repeats, respectively. No other loci were investigated, and it is therefore not known whether a CDCA7-associated maintenance methylation phenotype extends beyond (peri)centromeric satellites. Thijssen et al performed siRNA- mediated knockdown experiments in mouse embryonic fibroblasts (differentiated cells) and showed that lower levels of Zbtb24, Cdca7 and Hells protein correlate with reduced minor satellite repeat methylation, thereby implicating these factors in mouse minor satellite repeat DNA methylation maintenance. Furthermore, studies that demonstrate a HELLS-CDCA7 interaction are currently limited to Xenopus egg extract (Jenness et al 2018) and the human HEK293 cell line (Unoki et al 2019). Whether such an interaction exists in any other organism and is of relevance to DNA methylation mechanisms remains to be determined. Therefore, in my opinion, the conclusion that "Our co- evolution analysis suggests that DNA methylation-related functionalities of CDCA7 and HELLS are inherited from LECA" should be softened, as the evidence for this scenario is not very compelling and seems premature in the absence of molecular data from more species.

We appreciate this reviewer’s thorough reading of our manuscript.

Regarding the citation issues, we will cite Dunican 2013 and Dunican 2015. In addition, we went through the manuscript to update the citations.

As pointed out by the reviewer, the role of CDCA7 in genome DNA methylation was extensively studied in Velasco et al 2018. The result, together with Thijssen et al (2015), and Unoki et al. (2018), supports the idea that ZBTB24, CDCA7 and HELLS act within the same pathway to promote DNA methylation, the pattern of which is overlapping but distinct from DNMT3B-mediated methylation. This observation suggests that a ZBTB24- CDCA7-HELLS mechanism for DNA methylation may involve an alternative DNMT. Interestingly, our analysis of the gene presence-absence pattern revealed that the presence of CDCA7 coincides with DNMT1 more than DNMT3 genes. Indeed, while CDCA7 is lost from diverse branches of eukaryote species, genomes encoding CDCA7 always encode HELLS, and almost always encode DNMT1. Based on this observation, we speculate the role of CDCA7 is tightly linked to HELLS and DNA methylation throughout evolution.

As pointed out by Reviewer 1, the link between CDCA7, HELLS and DNA methylation has not been determined experimentally across these species. However, based on our previously published and unpublished data, we are confident about the functional interaction between CDCA7 and HELLS in Xenopus laevis and Homo sapiens.

Furthermore, the importance of HELLS homologs in DNA methylation has been extensively studied in human, mice and plants. We hope our current study will motivate the field to experimentally test the evolutionary conservation of HELLS-CDCA7 interaction, as well as their importance in DNA methylation, in other species.

The authors used BLAST searches to characterize the evolutionary conservation of CDCA7 family proteins in vertebrates. From Figure 2A, it seems that they identify a LEDGF binding motif in CDCA7/JPO1. Is this correct and if yes, could you please elaborate and show this result? This is interesting and important to clarify because previous literature (Tesina et al 2015) reports a LEDGF binding motif only in CDCA7L/JPO2.

We searched for a LEDGF binding motif ({E/D}-X-E-X-F-X-G-F, also known as IBM described in Tesina et al 2015) in vertebrate CDCA7 proteins, and reported their positions in Figure 2A. Examples of identified LEDGF-binding motifs are now presented in Fig. 2C.

To provide evidence for a potential evolutionary co-selection of CDCA7, HELLS and the DNA methyltransferases (DNMTs) the authors performed CoPAP analysis. Throughout the manuscript, it is unclear to me what the authors mean when referring to "DNMT3". In the Material and Methods section, the authors mention that human DNMT3A was used in BLAST searches to identify proteins with DNA methyltransferase domains. Does this mean that "DNMT3" should be DNMT3A? And if yes, should "DNMT3" be corrected to "DNMT3A"? Is there a reason that "DNMT3A" was chosen for the BLAST searches?

As described in the Methods section, both Human DNMT1 and DNMT3A were used to initially identify any proteins containing a domain homologous to the DNA methyltransferase catalytic domain. Within Metazoa, if their orthologs exist, the top hit from BLAST search using human DNMT1 and DNMT3A show E-value 0.0, and thus their orthology is robust. This is even true for DNMT1 and DNMT3 homologs in the sponge Amphimedon queenslandica, which is one of the earliest-branching metazoan species. For other DNMTs, such as DNMT2, DNMT4, DNMT5, DNMT6, we conducted separate BLAST searches using those proteins as baits as described in Methods. The methyltransferase domain was then isolated using the NCBI conserved domains search. The selected DNMT domain sequences were aligned with CLUSTALW to generate a phylogenetic tree to further classify DNMTs. In response to reviewer #2’s comments, we also generated another multi-sequence alignment of DNMTs using MUSCLE v5 and conducted maximum-likelihood-based phylogenetic tree assembly using IQ-TREE (new Fig. S6). The overall topology of these trees is consistent except for orphan DNMTs. It has been suggested that vertebrate DNMT3A and DNMT3B are derived from duplication of a DNMT3 gene of chordates ancestor (e.g., Liu et al 2020, PMID 31969623). As such many invertebrates encode only one DNMT3. As previously shown (Yaari et al., 2019, PMID 30962443), plants have two distinct DNMT3-like protein family, the ‘true DNMT3’ and DRM, the plant specific de novo DNMT that is often considered to be a DNMT3 homolog (see Reviewer 2’s comment). Our phylogenetic analysis successfully deviated the clade of DNMT3 and DRM from the rest of DNMTs (Figure S6). Yaari et al noted that PpDNMT3a and PpDNMT3b, the two DNMT3 orthologs encoded by the basal plant Physcomitrella patens, are not orthologs of mammalian DNMT3A and DNMT3B, respectively. Therefore, to minimize such nomenclature confusions, any DNMTs that belong to either the DNMT3 or DRM clades indicated in Figure S6 are collectively referred to as ‘DNMT3’ throughout the paper (see Figure S2 for overview).

CoPAP analysis revealed that CDCA7 and HELLS are dynamically lost in the Hymenoptera clade and either co-occurs with DNMT3 or DNMT1/UHRF1 loss, which seems important. Unfortunately, the authors do not provide sufficient information in their figures or supplementary data about what is already known regarding DNA methylation levels in the different Hymenoptera species to further consider a potential impact of this observation. What is "the DNA methylation status" of all these organisms? This information cannot be easily retrieved from Table S2. A clearer presentation of what is actually known already would improve this paragraph.

As the DNA methylation status of the species in the Hymenoptera clade has not been comprehensively tested, we initially did not include this information to Figure 7. However, during the course of the revision, we realized that Bewick et al.2017 (PMID 28025279) reported that DNA methylation is absent from the braconid wasp Aphidius ervi. We originally conducted synteny analysis on Aphidius gifuensis, which has a chromosome-level genome assembly with annotated proteins available in NCBI, whereas annotated proteins for Aphidius ervi protein are not available in NCBI. By conducting tBLASTn search against the Aphidius ervi genome, we now found that the presence/absence pattern of CDCA7, HELLS, DNMT1, DNMT3 and UHRF1 in Aphidius ervi is identical to that of Aphidius gifuensis, with a caveat that genome assembly of Aphidius ervi is at scaffold-level. In other words, DNA methylation, DNMT1 and CDCA7 are absent in Aphidius ervi, where 5mC is undetectable. Additionally, we also realized that the DNA methylation status reported for some species in Bewick et al. 2017 was inferred from the CpG frequency instead of the direct experimental detection of methylated cytosines. Therefore, we have amended Table S3 to indicate the presence of 5mC only for those species where this was experimentally tested. As such, we now consider the DNA methylation status of Fopius arisanus, which lacks DNMT1 and CDCA7, to be unknown.

Altogether, among the 17 Hymenoptera species that we analyzed (listed in the amended Table S3), the 8 species that have detectable DNA methylation all encode CDCA7, whereas the 2 species that do not have detectable DNA methylation lack CDCA7. We will note this finding in the revised text, and include the known 5mC status in the new Figure 7.

Furthermore, A. thaliana DDM1, and mouse and human Lsh/Hells are known to preferably promote DNA methylation at satellite repeats, transposable elements and repetitive regions of the genome. On the other hand, DNA methylation in insects and other invertebrates occurs in genic rather than intergenic regions and transposable elements (e.g. Bewick et al 2017; Werren JH PlosGenetics 2013). It would be helpful to elaborate on these differences.

We were aware of this interesting point, which was discussed in the third paragraph of the Discussion. To better illustrate this point, we now expanded the Discussion (page 14) to speculate about the role of DNA methylation in insects, where emerging evidence indicates the importance of DNMT1 in meiosis. It should be noted that, in the Arabidopsis ddm1 mutant, reduction of CG methylation of gene bodies is common (50% of all methylated euchromatic genes) (Zemach et al, 2013). In addition, hypomethylation is not limited to satellite repeats and transposable elements in ICF patients defective in HELLS or CDCA7 (Velasco et al., 2018).

Reviewer #2 (Public Review):

In this manuscript, Funabiki and colleagues investigated the co-evolution of DNA methylation and nucleosome remolding in eukaryotes. This study is motivated by several observations: (1) despite being ancestrally derived, many eukaryotes lost DNA methylation and/or DNA methyltransferases; (2) over many genomic loci, the establishment and maintenance of DNA methylation relies on a conserved nucleosome remodeling complex composed of CDCA7 and HELLS; (3) it remains unknown if/how this functional link influenced the evolution of DNA methylation. The authors hypothesize that if CDCA7-HELLS function was required for DNA methylation in the last eukaryote common ancestor, this should be accompanied by signatures of co-evolution during eukaryote radiation.

To test this hypothesis, they first set out to investigate the presence/absence of putative functional orthologs of CDCA7, HELLS and DNMTs across major eukaryotic clades. They succeed in identifying homologs of these genes in all clades spanning 180 species. To annotate putative functional orthologs, they use similarity over key functional domains and residues such as ICF related mutations for CDCA7 and SNF2 domains for HELLS. Using established eukaryote phylogenies, the authors conclude that the CDCA7-HELLS-DNMT axis arose in the last common ancestor to all eukaryotes. Importantly, they found recurrent loss events of CDCA7-HELLS-DNMT in at least 40 eukaryotic species, most of them lacking DNA methylation.

Having identified these factors, they successfully identify signatures of co-evolution between DNMTs, CDCA7 and HELLS using CoPAP analysis - a probabilistic model inferring the likelihood of interactions between genes given a set of presence/absence patterns. As a control, such interactions are not detected with other remodelers or chromatin modifying pathways also found across eukaryotes. Expanding on this analysis, the authors found that CDCA7 was more likely to be lost in species without DNA methylation.

In conclusion, the authors suggest that the CDCA7-HELLS-DNMT axis is ancestral in eukaryotes and raise the hypothesis that CDCA7 becomes quickly dispensable upon the loss of DNA methylation and/or that CDCA7 might be the first step toward the switch from DNA methylation-based genome regulation to other modes.

The data and analyses reported are significant and solid. However, using more refined phylogenetic approaches could have strengthened the orthologous relationships presented. Overall, this work is a conceptual advance in our understanding of the evolutionary coupling between nucleosome remolding and DNA methylation. It also provides a useful resource to study the early origins of DNA methylation related molecular process. Finally, it brings forward the interesting hypothesis that since eukaryotes are faced with the challenge of performing DNA methylation in the context of nucleosome packed DNA, loosing factors such as CDCA7-HELLS likely led to recurrent innovations in chromatin-based genome regulation.

Strengths:

• The hypothesis linking nucleosome remodeling and the evolution of DNA methylation.

• Deep mapping of DNA methylation related process in eukaryotes.

• Identification and evolutionary trajectories of novel homologs/orthologs of CDCA7.

• Identification of CDCA7-HELLS-DNMT co-evolution across eukaryotes.

Weaknesses:

• Orthology assignment based on protein similarity.

• No statistical support for the topologies of gene/proteins trees (figure S1, S3, S4, S6) which could have strengthened the hypothesis of shared ancestry.

We appreciate the reviewers’ accurate summary, nicely emphasizing the importance of the our study. We agree that better phylogenetic analysis for orthology assignment will strengthen our conclusion. Having anticipated this weakness, however, we specifically conducted a CoPAP analysis exclusively for Ecdysozoa specieswhich supported our major conclusion, as orthology assignment is straightforward in these species. For example, if we conduct BLAST search against the clonal raider ant Oocerea biroi protein dataset using human HELLS as a query, top 1 hit is a protein sequence annotated as one of three isoforms of ‘lymphoid-specific helicase” (i.e., HELLS), with E value 0.0. Similarly, the top BLAST hit from the Oocerea biroi dataset using human DNMT1 as a query also returns with isoforms of DNMT1 with E value 0.0. As such, there are little disputes in orthology assignment in Ecdysozoa. Outside of Chordata, classification of DNMTs, particularly in Excavata and SAR, require more extensive identification in these supergroups. Our current orthology assignment for the major targets in this study (HELLS, DNMT1, DNMT3, DNMT5) is largely consistent with published results (Ponger et al., 2005 PMID 15689527; Huff et al, 2014 PMID 24630728; Yaari et al., 2019 PMID 30962443; Bewick et al., 2019 PMID 30778188). However, while we are preparing this response and re-crosschecking our assignments with these references, we realized that we had erroneously missed DNMT5 orthologs in Leucosporidium creatinivorum, Postia placenta, Armillaria gallica and Saitoella complicata, and a DNMT6 ortholog in Fragilariopsis cylindrus. We also recognized that DNMT4 orthologs were identified in Fragilariopsis cylindrus and Thalassiosira pseudonana in Huff et al 2014 (PMID 24630728), but in our phylogenetic analysis, these proteins form a distinct clade between DNMT1/Dim-2 and DNMT4 (original Figure S6), although the confidence level of this classification by Huff et al was not strong. To resolve this potential confusion in DNMT annotations, we generated new multiple sequence alignments with MUSCLE v5 and IQ-TREE 2 (maximum likelihood-based method, coupled with selection of optimal substitution model and bootstrapping). The tree topology was not significantly altered between the two methods, except for the unambiguous location of orphan DNMTs and DNMT4-related proteins. To avoid unnecessary confusion in the DNMT annotations, we decided to present MUSCLE-IQ- TREE for the DNMT phylogenetic tree and classification (new Fig. S6). The raw results of IQ-TREE analysis for CDCA7/zf-4CXXC_R1, HELLS SNF2 domain, and DNMTs are included as Dataset S1-S3. We then conducted CoPAP analysis using the corrected classification. As it is not clear a priori if fungal specific CDCA7-like proteins (now referred to as CDCA7F with class II zf-4CXXC_R1) should be considered CDCA7 orthologs, we conducted CoPAP against two lists; the first list includes CDCA7F in the CDCA7 group, whereas the second list includes a separate category of class II zn-4CXXC_R1, which includes CDCA7F. Both results show slightly different topology in the coevolutionary linkages but support our major conclusion that CDCA7 coevolved with DNMT1-UHRF1 and HELLS. These new CoPAP results are shown in Fig. S7.

Reviewer #1 (Recommendations For The Authors):

Summary

Last sentence: "...a unique specialized role of CDCA7 in HELLS-dependent DNA methylation maintenance...". What do the authors mean?

Our analysis strongly indicates that CDCA7 is dispensable in systems lacking HELLS and DNMT (particularly DNMT1). In other words, species preserve CDCA7 only if it has both HELLS and DNMT1 (or in some cases DNMT5). The importance of HELLS homologs in DNA methylation has been extensively studied in human, mouse and plants. However, in these studies, substantial DNA methylation remains despite the defective HELLS/DDM1 (especially in euchromatic regions). Additionally, there are species (e.g., Bombyx mori) that have DNMT1 and detectable DNA methylation but lacks HELLS and CDCA7. These observations suggest that the role of CDCA7 must be unique and specialized in a way that it is strongly coupled to HELLS-dependent DNA methylation (but not HELLS-independent DNA methylation), and that this function of CDCA7 seems to be inherited from the last eukaryotic common ancestor.

Introduction

• page 3: "DNMTs are largely subdivided into maintenance and de novo DNMTs" - Which species are the authors referring to?

As described in the cited reference (Lyko 2018), maintenance DNA methylation and de novo DNA methylation are well accepted functional classification of DNA methylation. It is also currently accepted that distinct DNMTs execute maintenance DNA methylation or de novo DNA methylation, although crosstalk between these processes has been reported. Therefore, we stated, “DNMTs are largely subdivided into maintenance DNMTs and de novo DNMTs”, and this subdivision is species independent.

• page 3" "Maintenance DNMTs recognize hemimethylated CpGs. " - Can the authors please define the species and/or literature they are referring to? This seems important to clarify. For instance, mammalian DNMT1 requires a co-factor, UHRF1, which recognizes hemimethylated DNA and H3K9me3 (Bostick et al 2007).

We meant to describe, “Maintenance DNMTs directly or indirectly recognize hemimethylated CpGs…”. The specific requirement of UHRF1 for DNMT1-mediated maintenance DNA methylation is explained in the subsequent sentence “In animals…”. In the case of Cryptococcus neoformans, DNMT5 recognizes hemimethylated DNA independently of UHRF1 in vitro to execute maintenance methylation.

• page 3: The authors may want to mention that A. thaliana also has a de novo DNA methyltransferase, DRM2, a homolog of the mammalian DNMT3 methyltransferases. This seems important, since they show in Figure 1 that a de novo methyltransferase is found in A. thaliana. Also, later in their manuscript they mention plant de novo DNA methylation.

Thanks for pointing this out. As shown in Figure 5, we classified plant DRMs as DNMT3-like proteins, but we now note this in the Introduction.

• page 3: Sentence starting "In about 50% of ICF patients,..." - Why is DNMT3B referred to as "de novo", is it not a de novo DNA methyltransferase?

You are correct. Quotation marks are now removed to avoid unnecessary confusion.

• page 4: Sentence starting "Indeed, the importance of HELLS/CDCA7 in DNA methylation maintenance...", - Which references (Han et al., 2020; Ming et al., 2021; Unoki, 2021; Unoki et al., 2020) provide experimental evidence for a role of CDCA7 in DNA methylation maintenance by DNMT1?

Thanks for pointing out the typo. “/CDCA7” is now removed.

• page 5: Sentence starting "Indeed, it has been shown that DNMT3A..." - Should DNMTB be DNMT3B?

Yes. This is now corrected.

Results

• Page 5: Sentence starting "However, we identified a protein..." - No A. thaliana reference?

We added Zemach et al 2010, and Chan et al 2005.

• Figure 2B: "ICF4 mutations" should this be "ICF3 mutations"?

• Figure 3: "ICF4 mutations" should this be "ICF3 mutations"?

• Figure 4: "ICF4 mutations" should this be "ICF3 mutations"?

• Figure S1: Orange colored "CDC7L (fish), CDC7e, CDC7, CDC7L" is there an "A" missing?

• Figure S5: "ICF4 mutations" should this be "ICF3 mutations"?

These typos are now corrected. Thank you.

• Figure S7: What is "CDCA7(II)" referring to, "zf-4CXXC_R1 class II (plants)"?

The original CDCA7 (II) included proteins with class II zf-4CXXC_R1, which are found in plants, fungi, Acanthamoeba castellanii and Amphimedon. Among those species, the prototypical CDCA7 orthologs are absent only in fungi. It has been a priori unclear if fungal proteins with class II zf-4CXXC_R1 (now we term CDCA7F) should be included in CDCA7 for CoPAP analysis. Although we originally included CDCA7F in CDCA7, we now show the results of two analyses. In the first one (Fig. S7A) CDCA7F was included in CDCA7, whereas in in the second one (Fig. S7B) CDCA7F was included in the separate category of class II zf-4CXXC_R1. Topologies of two results are slightly different, but they both show coevolutionary linkage between the CDCA7 and DNMT1- UHRF1 cluster.

• Figure 4 and 5: In the case of preliminary genome assemblies what is the difference between empty squares with dotted lines and filled squares without dotted lines?

As it is difficult to be certain of a gene’s absence (did the species lose the gene or is it simply not annotated due to incomplete genome coverage?), we illustrated the absence of a gene in preliminary genome assemblies with an empty square with dotted outline. Since the presence of a gene is evident regardless of the level of genome assembly, the presence of a gene is represented with filled squares with solid lines, even for preliminary genome assemblies.

• Figure 1: Why was Mus musculus - one of the main model organisms used for many DNA methylation studies not included? Also what are empty and filled squares?

Filled and empty squares indicate the presence and absence of the indicated genes, respectively. Clarifying statement is now added in the figure legends. Mus musculus is now included in the figure.

• Figure S2: Adding the existence of DNA methylation and DNMT3 in the bottom right part of the figure (overall no of species) would make this panel more informative

We included this overview to summarize the co-retention of CDCA7, HELLS and maintenance DNMTs across the analyzed species. We decided not to include DNA methylation, since DNA methylation status is known for only a fraction of the listed species. Inclusion of DNMT3 will introduce too many possible gene presence-absence combinations to convey a clear message. However, we now mention in the revised text (page 11, second paragraph) that unlike the prevalent co-retention of DNMT1 in species with CDCA7, we identified several species that possess CDCA7, HELLS and DNMT1 but lack DNMT3. These examples include insects such as the bed bug Cimex lectularius and the red paper wasp Polistes canadensis.

• Page 6: Sentence starting "This leucine zipper sequence is highly conserved..." - Figure/Reference missing?

The sequence alignment of the leucine zipper is now shown in Fig. 2C.

• page 6: Sentence starting "In contrast to zf-4CXXC_R1 motif-containing proteins..." - The authors may want to mention the role of the CXXC zf domain in KDM2A/B, DNMT1, MLL1/2 and TET1/3 and what the CDCA7 CXXC zf domain is/could be required for.

The notion that zf-CXXC binds to nonmethylated CpG is now included. Due to the substantial difference between zf-CXXC and zf-4CXXC_R1, we hesitated to relate the function of zf-4CXXC_R1 with zf-CXXC, but we now discuss a potential role of zf- 4CXXC_R1 in sensing DNA methylation status in Discussion (Page 13).

• page 7: Sentence starting "Second, the fifth cysteine is replaced..."- Zoopagomycota" - Figure 4A does not have this labeling, one has to deduce this from Figure 4B.

We fixed this by including the list of Zoopagomycota species in the main text.

• page 7: Sentence containing "Neurospora crassa DMM-1 does not directly regulate DNA methylation or demethylation but rather..." - How does the information about DMM- 1 relate to what is shown in Figure 4B, to CDCA7, HELLS and DNMTs? Please clarify.

Both Neurospora DMM-1 and Arabidopsis IBM1 contain the JmjC domain and are implicated in an indirect control mechanism of DNA methylation. Since it has never been pointed out that they have a divergent zf-4CXXC_R1 domain, which clearly shares the origin with CDCA7 proteins, we thought that this is important to note. We realized that we did not clearly mark Neurospora XP-956257 as DMM-1 in Fig. 4B. This is now fixed.

• Heading "Systematic identification of CDCA7, HELLS and DNMT homologs in eukaryotes". When mentioning CDCA7 the authors may want to decide on the use of one consistent definition of "prototypical (Class I) CDCA7-like proteins (i.e. CDCA7 orthologs)" "Class I CDCA7 proteins". Constantly changing the way how they refer to these proteins is very confusing.

We now make it clear that we call proteins with class I zf-CXXC_R1 motif CDCA7 orthologs. We also define class II zf-4CXXC_R1 (as those with a substitution at ICF- associated glycine residue). Since no clear CDCA7 orthologs can be found in fungi, we now call fungi proteins with class II zf-4CXXC_R1 “CDCA7F”, implying its ambiguous orthology assignment.

Under this heading there is also no mention of DNMTs. Instead, the authors introduce DNMTs under the heading "Classification of DNMTs in eukaryotes" - Please clarify.

This is now corrected.

• page 9: Sentence containing "... presence of DNMT1, UHRF1 and CDCA7 outside of Viridiplantae and Opisthokonta is rare". What does "rare" mean? How is UHRF1 relevant here?

Among the 32 species outside of Viridiplantae and Opisthokonta, only the Acanthamoeba castellanii genome encodes clear orthologs of DNMT1, UHRF1 and CDCA7. Although it is often difficult to deduce if the selected panel of species is a reasonable representation, we think that it is not unreasonable to state that Acanthamoeba is a rare case to encode this set of proteins outside of Viridiplantae and Opisthokonta. We include UHRF1 since it is a well-established activator of DNMT1, and indeed our CoPAP analysis showed a tight coevolution of UHRF1 with DNMT1. Outside of Viridiplantae and Opisthokonta, only Acanthamoeba castellanii and Naegleria gruberi encode UHRF1. Interestingly, these two species also encode CDCA7 and HELLS.

Having said that, we rephrased this sentence, which reads; “Species that encode a set of DNMT1, UHRF1, CDCA7 and HELLS are particularly enriched in Viridiplantae and Metazoa.”

• page 11: Sentence containing "..., that the function of CDCA7-like proteins is strongly linked to HELLS and DNMT1,..." What do the authors mean with "the function of CDCA7-like proteins"? And what happened to DNMT3?

Our observation that almost all species that contain CDCA7 (including fungal CDCA7F) also have DNMT1 and HELLS, despite the frequent loss of these genes in species that do not contain CDCA7, indicates “that the function of CDCA7-like proteins is strongly linked to HELLS and DNMT1”. We found only 2 species that possesses CDCA7 (class I or class II) but not DNMT1 among the panel of 180 species. These 2 exceptional species, Naegleria gruberi and Taphrina deformans, do encode UHRF1-like proteins and a DNMT (an orphan DNMT in N. gruberi and DNMT4 in T. deformans). In contrast, we found 26 species that possess CDCA7 (or CDCA7F) but not DNMT3 (Table S1), so the linkage between CDCA7 and DNMT3 is weaker.

• page 11: Sentence containing "..., CDCA7 is lost from this gene cluster in parasitoid wasps, including Ichneumonoidea wasps and chalcid wasps". This sentence is confusing because already in an earlier paragraph the authors say that "Microplitis demolitor lost CDCA7" and in the following sentence they say "...among Ichneumonoidea wasps, CDCA7 appears to be lost in the Braconidae clade, ...". It would greatly help this reader if the authors could streamline these sentences and also decide on whether CDCA7 is lost in M. demolitor or CDCA7 appears to be lost in M.demolitor.

The confusion was in part due to the difficulty in differentiating between the true loss of a gene versus its apparent absence in a species due to an incomplete genome assembly, including for of M. demolitor. To verify that the loss of CDCA7 was not due to gaps in the genome assembly, we executed the synteny analysis. However, we edited this section to improve the readability (Page 12-13).

What could be the role for HELLS/CDCA7 in insect DNA methylation? In several cases, the authors analyses reveal co-evolutionary links between DNMT3 (DNMT3A?) and CDCA7/HELLS. I do not understand why this finding is not really discussed by the authors. Instead there is a strong focus on replication-uncoupled DNA methylation maintenance. Could the authors elaborate why?

The role of DNA methylation in insects is largely unclear, so discussion must be highly speculative. A recent finding in the clonal raider ant, showing that DNMT1 is not essential for development but is critical for oogenesis, pointed toward a possible more universal role of DNA methylation in meiosis. Stimulated from a finding in Neurospora, where DNA methylation is required for homolog pairing during meiosis, we discuss a speculative model that DNA methylation status acts as a hallmark to distinguish between healthy/young DNA and old/mutated (or competitive/pathogenic) DNA at homolog pairing during meiosis (page 14).

Regarding the cases where CDCA7 and DNMT3 are co-lost, we had discussed about this phenomenon at the last section of Result, stating, “This co-loss of CDCA7 and DNA methylation (together with either DNMT1-UHRF1or DNMT3) in braconid wasps suggests that evolutionary preservation of CDCA7 is more sensitive to DNA methylation status per se than to the presence or absence of a particular DNMT subtype.” Please note that we found several lineages that lacks CDCA7 but has DNMT1 (and DNMT3), whereas almost all species that has CDCA7 also has DNMT1 (but not necessarily DNMT3). Supported with our CoPAP analyses, our results indicate the tight functional link between CDCA7 and DNMT1, but it does not necessarily mean that CDCA7 does not play any role related to DNMT3 or de novo methylation. Clarification of this point and our speculation of how CDCA7 loss is linked to reduced requirement of DNA methylation are discussed in page 13 and 14 with additional texts.

Discussion

• page 12: Where is the data supporting. "... the red flour beetle Tribolium castaneum possesses DNMT1 and HELLS, but lost DNMT3 and CDCA7"?

Figure 5, Figure S2 and Table S1. This is now noted in the text.

• page 14: Based on which parts of their analyses or evidence from the literature can the authors speculate that "...the evolutionary arrival of HELLS-CDCA7 in eukaryotes might have been required to transmit the original immunity-related role of DNA methylation from prokaryotes to nucleosome-containing (eukaryotic) genomes"? Please clarify.

This is inferred from the well-known role of DNA methylation in bacteria for defending against phage viruses. However, it was not correct to state that such a function was inherited from prokaryotes. It should be stated that it was inherited from the last universal common ancestor (LUCA). We also admit that it is not clear if such an immunity-related role was inherited from LUCA, or if it emerged through convergent evolution. Therefore, we amended this description to emphasize our hypothesis that the advent of CDCA7 was “a key step to transmit the DNA methylation system from the LUCA to the eukaryotic ancestor with nucleosome-containing genomes”.

Supplementary Figures/Tables

• page 26: Table S2 and Table S3, it seems that these tables show data that supports what is shown in Figure 7 and not Figure 5.

You are correct. Thank you for pointing out the typos.

Has the methylation status been assessed in C. glomerata, C. typhae, Chelonus insularis, Diachasma alloeum or Aphidius gifuensis? Please clarify in Table S2.

Not to our knowledge. However, as we realized that absence of DNA methylation in Aphidius ervi was previously reported (Bewick et al 2017), we now included this data together with presence/absence analysis of DNMT1, UHRF1, DNMT3, CDCA7 and HELLS. Known presence/absence of DNA methylation is now shown in Fig.7.

Reviewer #2 (Recommendations For The Authors):

Recommendation to strengthen the paper:

1) Phylogenetics:

• Test and report the appropriateness of the substitution model used in protein alignments/trees.

• Use Maximum likelihood methods and/or MCM Bayesian inference to build and report trees with well supported topologies. This is required to properly assign orthology (shared ancestry). This will avoid false interpretation due to technical limitation of similarity-based phylogenies (without statistical support). Figure S1, S3, S4 and S6.

To address these points, we made new multisequence alignments using MUSCLE v6 and generated phylogenetic trees using the maximum likelihood-based IQ-TREE 2, where multiple models were screened. A consensus tree was generated after 1000 bootstrap replicates from the best alignment and model. The topology and assignment of these new trees were largely consistent with the original trees, except for some corrections in DNMT assignment as discussed below.

1. We realized that we erroneously missed DNMT5 orthologs of Leucosporidium creatinivorum, Postia placenta, Armillaria gallica and Saitoella complicata., and DNMT6 orthologs from Fragilariopsis cylindrus reported in Huff et al 2014 (PMID 24630728). They are now included in the new list and CoPAP analysis.

2. DNMT4 orthologs were identified in Fragilariopsis cylindrus and Thalassiosira pseudonana by Huff et al 2014 (PMID 24630728), but in our original phylogenetic analysis, these proteins form a distinct clade between DNMT1/Dim-2 and DNMT4. The new tree and classification are more consistent with Huff et al, so we present the new tree in Fig. S6 and conducted the classification based on this tree.

Beside Fig. S6, we decided to maintain original Fig. S1, S3 and S4 (with a few adjustments) for better visibility, but we included the results of IQ-TREE analysis as Dataset S1-S3.

The CoPAP analysis based on the revised assignment slightly changed the topology of coevolutionary linkages. In addition, we obtained a slightly different result depending on whether fungal specific CDCA7 with class II zn-4CXXC_R1 (now referred to as CDCA7F) is included as a CDCA7 ortholog or not. Despite this difference, we reproducibly observed the coevolutionary linkage between CDCA7 and DNMT1- UHRF1.

• Be more careful with wording: RBH is not sufficient to call gene/proteins orthologs (e.g. Page 8). The above mentioned method will help you support this claim (+ synteny when you can).

We were aware of this issue. This is why we conducted phylogenetic tree building based on sequence alignment of full-length HELLS (Fig. S3) and SNF2 domain only (Fig. S4), as explained in the text. We found that the RBH criterion is robust in Metazoa; orthologs are easily recognizable with very low E-value (0.0) and extensive homology over the full length of the protein, while synteny is not practical to employ in the diverse set of species.

• Also, use "co-retention" or "co-evolution" but not "co-selection" when describing CoPAP results - as CoPAP does not test for signature of natural selection.

This is a good point and is now corrected.

• The statistics (p-val...) underlying the CoPAP analyses should be explained.

The explanation is now added in Methods section.

“A method to calculate p-value for CoPAP was described previously (Cohen et al., 2012, PMID 22962457). Briefly, for each pair of tested genes, Pearson's correlation coefficient was computed. Parametric bootstrapping was used to compute a p-value by comparing it with a simulated correlation coefficient calculated based on a null distribution of independently evolving pairs with a comparable exchangeability (a value reporting the likelihood of gene gain and loss events across the tree).”

2) Figure S2 and S3 could be improved for readability

After consideration of this criticism, we decided to keep their original formats for following reasons.

Figure S2. The purpose of this list is to better visualize the comprehensive list shown in Table S2. A consolidated list is already shown in Figure 5. An alternative choice is to make a diagram where individual species names are unreadable. This kind of presentation is seen in many published papers, but we found that they are not helpful to check the details. As this is a supplementary figure, we prefer to show the detailed data that can be visible without a specialized software.

Figure S3. This figure is included to show which SNF2 family proteins are more likely to be misassigned as HELLS/DDM1 orthologs. We believe that the figure serves this purpose.

3) What is the meaning of the coloring patterns of ICF residues in znf?

ICF residues are highlighted as light blue in the schematics to indicate its conservation. In the alignment, the coloring reflects the level of conservation within the shown set of proteins, and the choice of coloring was set by Jalview.

4) To improve clarity: the introduction could be more focused on evolutionary considerations and functional link between CDCA7-HELLS and DNMTs.

We revised the first paragraph of the introduction to illustrate this point.

5) Could indicate the CDC7A loss / DNA methylation hypothesis in the abstract.

We now included this hypothesis in the Abstract.