Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

Reviewer #1 (Evidence, reproducibility and clarity (Required)):

Summary Maintenance of the histone H3 variant CENP-A at centromeres is necessary for proper kinetochore assembly and correct chromosome segregation. The Mis18 complex recruits the CENP-A chaperone HJURP to centromeres to facilitate CENP-A replenishment. Here the authors characterise the Mis18 complex using hybrid structural biology, and determine complex interface separation-of-function mutants.

Major Comments The SAXS and EM data on the full-length Mis18 components must be included in the main Figures, either as an additional figure or by merging/rearranging the existing figures. The authors discuss these results in three whole paragraphs, which are a very important part of the paper.

We thank the reviewer for this constructive suggestion. We have now included an additional figure (new Fig. 2, attached below), that highlights the fit of the integrative model against the SAXS and EM data.

Could the authors also compare the theoretical SAXS scattering curves generated by their final model(s) with the experimental SAXS curves? This would provide some additional evidence for the overall shape of their complex model beyond the consistency with the Dmax/Rg.

We acknowledge the importance of this suggestion. We have now compared the theoretical SAXS scattering curve of the Mis18a/b core complex (named Mis18a/b DN), which lacks the flexible elements (disordered regions and the helical region flexibility connected to the Yippee domains). The theoretically calculated SAXS scattering curve of the model matches nicely with the experimental data with c2 value of 1.36. This data is now included in new Fig. 2 (Fig. 2f) and is referenced on page 9 line 21.

Minor Comments

While the introduction is clearly written, an additional cartoon schematic, representing the system/question would be helpful to a non-specialist reader to interpret the context of the study.

We have now included a cartoon in the revised Fig. 1 to support the introduction on centromere maintenance and the central role of the Mis18a/b/BP1 complex in this process. Please find the new Fig. 1 below.

No doubt the authors had a reason for choosing their figure allocation, but I wonder if more material couldn't be brought from the supplementary into the main figures?

As addressed in our response to one of the major comments, we have now moved key CLMS, SAXS and EM data from the supplemental figure into the main figure, new Fig. 2.

Page 6 "Mis18-alpha possesses an additional alpha-helical domain" - please make it clear in addition to what (I assume it's in addition to Mis18-beta).

Apologies for the lack of clarity. We have now rephrased this sentence to highlight that this difference is in comparison with Mis18b on page 6 line 15.

Page 7 - Report the RMSD of the Pombe vs. Human Mis18-alpha yipee structures?

The S. pombe Mis18 Yippee structure superposes on to the Human Mis18a Yippee domain with an RMSD of 0.92 angstroms with is now mentioned on page 7 line 9.

Page 7 - "We generated high-confidence structural models...." is there a metric for the confidence as reported by RaptorX? Perhaps includinging the PAE plots in the supplementary for the AlphaFold generated models would be useful?

We thank the reviewer for the valid suggestion. We have now included the PAE plot corresponding to the AlphaFold model in the supplementary Fig. S1d and reference on page 7 line 18. RaptorX ranks models based on estimated error. We have now included this information in the new figure legend for Supplementary Fig. S1.

Figure 1 - Perhaps label figure 1b as being experimentally determined, with the R values (as for Figure 1d), and 1c being a predicted model.

We have included Rfree and Rwork values for the Mis18a Yippee homo dimer structure and labelled Mis18a/b Yippee hetero-dimer as the predicted model in Fig. 1c and 1d.

Page 8 "This observation is consistent with the theoretically calculated pI of the Mis18alpha helix" This is a circular argument, of course this region has a low pI due to the amino acid composition. Please remove this statement.

We have now removed this statement as suggested.

Page 8 "...reveals tight hydrophobic interactions" these are presumably shown in Figure 1d rather than in the referenced 1e.

We apologise for the oversight. We have now referred to the correct figure (Fig. 1f in the revised Fig. 1).

Page 8 - The authors should briefly somewhere discuss why there is a difference between their results and those in Pan et al 2009. As I understand it, the Pan et al paper was based in part on modelling with CLMS data as restraints.

We thank the reviewer for this suggestion. According to Pan et al., 2009, the model shown by them was generated using CCBuilder, and their CLMS data could not differentiate the two models with the 2nd Mis18a C-terminal helix in either parallel or anti-parallel orientation. We now briefly discuss this on page 8 and line 22 as follows: "Although the Pan et al., 2019 model presented the 2nd Mis18a in a parallel orientation, they did not rule out the possibility of this assembling in an anti-parallel orientation within the Mis18a/b C-terminal helical assembly (Pan et al., 2019)."

Figure 1 - The labelling of the residues for Mis18-alpha in Figure 1d is problematic, they are black on dark purple (might be my printer/screen/eyes) suggest amending.

We have now rearranged the label positions to overcome this issue. For clarity, the labels that could not be moved appropriately are shown in white.

Figure S3a - Do the authors have some data to show the mass of the cross-linked complex that was loaded onto grids is consistent with what is expected?

Unfortunately, the amount of material that we recover after performing GraFix is not sufficient enough to determine the molecular weight of the crosslinked sample by techniques such as SEC-MALS. However, GraFix fractions were analysed by SDS PAGE, and fractions that ran around the expected molecular weight were selected for EM analysis. We have now included the corresponding SDS-PAGE showing the migration of the crosslinked sample analysed by EM (Supplementary Fig. S3a).

Figure S3b - scale bar

Revised Fig. 2d now includes the scale bar shown.

Figure S3c - Could the authors show or explain the differences between these different 3D reconstructions?

The models mainly differ in the relative orientations of the bulkier structural features that are referred to as 'ear' and 'mouth' pieces of a telephone handset. This has been mentioned in the text, but we note that the figure is not referenced right next to this statement. We have now amended this (Page 9 line 19), and to make it clear, we have also highlighted the difference using an arrowhead in Fig. 2e and S3b. The different orientations are also stated in the corresponding figure legends.

Page 9 - The use of "AFM" for AlphaFoldMultimer" is a little confusing since AFM is the established acronym for Atomic Force Microscopy. Perhaps AF2M?

We have now replaced AFM with AF2M on page 9 to avoid confusion.

Figure S4a - Control missing for Mis18-alpha wild-type

Apology for the confusion, this control is present in Fig. 4a. We have now stated this in the figure legend of S4a for clarity.

Figure S4 d and e - The contrast between the bands and the background is very bad (at least in my copy).

We have now adjusted the contrast of the blots in Fig. S4d and S4e response to this comment.

Page 13 "Our structural analysis suggests that two Mis18BP1 fragments.....". How did you arrive at this conclusion? Is this based on the AlphaFold/RaptorX model? What additional evidence do you have that the positioning of the Mis18BP1 is correct? Does the CLMS data support this?

We confirm that this statement is based on AlphaFold model. We have now explicitly highlighted this on page 14, line 5. As noted in the same paragraph (page 14, line 19), this model agrees with the contacts suggested by the cross-linking mass spectrometry data presented here.

Figure 4a - Would the authors like to consider using a different colour for Mis18BP1? The contrast is not great, especially in the electrostatic surface inset.

In response to this suggestion, the Mis18BP1 helix is now shown in grey in the inset of Fig. 5a.

Reviewer #1 (Significance (Required)):

General Assessment The paper is extremely clearly written. Likewise the figures are beautifully presented and the data extremely clean and fully supportive of the authors conclusions. Indeed it is seldom that one sees the depth of the structural approaches (X-ray, CLMS, EM, SAXS) in one paper which is a huge strength of the manuscript. In addition the translation of this data into very clean cell biological experiments, makes the paper truly outstanding.

Advance The authors provide the first model of the Mis18 complex, with extensive evidence to back up this model. The authors provide additional evidence as to how the deposition/renewal of CENP-A might be mediated by the Mis18 complex. The advance comes from both the level of clarity, detail, and scope achieved in this paper.

Audience This will likely be of great interest to anyone with an interest in chromosome biology, plus be of interest to structural biologists as an outstanding example of hybrid structural biology.

Expertise I am a biochemist with a background in structural biology with some familiarity with centromere biology

Reviewer #2 (Evidence, reproducibility and clarity (Required)):

Summary: The manuscript "structural basis for Mis18 complex assembly: implications for centromere maintenance" by Thamkachy and colleagues describes a study that uses structural analysis to test essential candidate residues in Mis18 complex components in CENP-A loading. For chromosomes to faithfully segregate during cell division, CENP-A levels must be maintained at the centromere. How CENP-A levels are maintained is therefore important to understand at the mechanistic level. The Mis18 complex has been found to be important, but how exactly the various Mis18 complex components interact and how they regulate new CENP-A loading remains not fully understood. This study set out to characterize the critical residues using X-ray crystallography, negative staining EM, SEC analysis, molecular modeling (Raptorx, AlphaFold2, and AlphaFold-multimer) to identify the residues of Mis18a and Mis18b that are critical for the formation of the Mis18a/b hetero-hexamer and which residues are important for Mis18a and Mis18BP1 interactions. A complex beta-sheet interface dictates the Mis18a and Mis18b interactions. Mutating the Mis18a residues that are important for the Mis18a/b interactions resulted in impaired pull-down of Mis18b and reduced centromeric levels of mutated Mis18a. The functional consequences of mutating residues that impair Mis18a/b interactions is that with reduced centomeric levels of Mis18a, also impaired new CENP-A loading. Interestingly, mutated Mis18b did not impact centromeric Mis18a levels and only modestly impaired new CENP-A loading. These data were interpreted that Mis18a is critical for new CENP-A loading, whereas Mis18b might be involved in finetuning how much new CENP-A is loaded. Overall, it is a very well described and well written study with exciting data.

Major comments:

• Overall, the structural data and the IF data support the importance of Mis18a residues 103-105 are critical for centromeric localization and new CENP-A loading, whereas Mis18b residues L199 and I203 are critical for centromeric localization, but only very modestly impair centromeric Mis18a localization and new CENP-A loading. In the discussion the authors argue that the N-terminal helical region of Mis18a mediate HJURP binding. This latter is postulated based on published work, but not tested in this work. This should be clarified as such.

We thank the reviewer for this comment. Our very recent study aimed at understanding the licencing role of Plk1, independent of the work reported here, serendipitously has now validated this suggestion and demonstrates that a Plk1-mediated phosphorylation cascade activates the Mis18a/b complex via a conformational switch of the N-terminal helical region of Mis18a, which facilitates a robust HJURP-Mis18a/b interaction (Parashara et al. bioRxiv 2024). An independent study from the Musacchio lab (Conti et al. bioRxiv, 2024) also reports similar findings, mutually strengthening our independent conclusions. Overall, these studies highlight the importance of the critical structural insights into the Mis18 complex this study reports. We now explicitly discuss the validation of our original hypothesis by citing our recent work along with that of the Musacchio lab. The corresponding section of the last paragraph now reads as follows (page 17 line 10): "Previously published work identified amino acid sequence similarity between the N-terminal region of Mis18a and R1 and R2 repeats of the HJURP that mediates Mis18a/b interaction (Pan et al., 2019). Deletion of the Mis18a N-terminal region enhanced HJURP interaction with the Mis18 complex (Pan et al., 2019). Here, we show that the N-terminal helical region of Mis18a makes extensive contact with the C-terminal helices of Mis18a and Mis18b, which had previously been shown to mediate HJURP binding by Pan et al., 2019. Collectively these observations suggest that the N-terminal region of Mis18a might directly interfere with HJURP - Mis18 complex interaction. Two independent recent studies (Parashara et al., 2024, Conti et al., 2024) reveal that this is indeed the case and a Plk1-mediated phosphorylation cascade involving several phosphorylation and binding events of the Mis18 complex subunits relieve the intramolecular interactions between the Mis18a N-terminal helical region and the HJURP binding surface of the Mis18a/b C-terminal helical bundle. This facilitates robust HJURP-Mis18a/b interaction in vitroand efficient HJURP centromere recruitment and CENP-A loading in cells. Overall, these studies also highlight the importance of the critical structural insights into the Mis18 complex we report here."

• Overall, the authors clearly describe their data and methodology and use adequate statistical analyses. The structural data of the Mis18a/b complex being a hetero-hexamer is convincing, but the validation in vivo is missing. As structural experiment are not performed under physiological conditions, it is important to establish the stoichiometry in vivo to further support the totality of the findings of the structural experiments and modeling. The data for the hierarchical assembly of Mis18a and Mis18b at the centromere and its importance in new CENP-A loading is convincing. An additional open question is whether "old" centromeric CENP-A or HJURP:new CENP-A complex is needed to recruit Mis18a to the centromere and whether the identified residues have a role in Mis18a centromeric localization. These data would provide a solid link between the Mis18 complex and how it is directly linked to new CENP-A loading.

We agree that establishing the stoichiometry of Mis18 subunits of the Mis18 complex in vivo would be insightful. However, considering that the Mis18 complex assembles in a specific window of the cell cycle (late Mitosis and early G1), we think characterising the stoichiometry in cells is extremely difficult and technically challenging. However, consistent with our structural model, several lines of independent evidence (Pan et al., 2017 and Spiller et al., 2017) using different biophysical methods (Analytical Ultra Centrifugation (Pan et al., 2017), SEC-MALS (Spiller et al., 2017)) showed that recombinantly purified Mis18 complex (irrespective of the expression host, from both E. Coli or insect cells) is a hetero-octamer made of a hetero-hexameric Mis18a/b (4 Mis18a and 2 Mis18 b) complex bound to two copies of Mis18BP1. These observations suggested that hetero-hexamerisation of the Mis18a/b complex may be needed to bind and dimerise Mis18BP1 in cells. Previously published cellular studies support the in vivo requirement of the hetero-octameric Mis18 assembly as: (i) Perturbing the hetero-hexamerisation of the Mis18a/b complex (by introducing mutations at the Mis18a/b Yippee dimerisation interface, which while did not disrupt Mis18a/b complex formation, perturbed its hetero-hexamerisation and resulted in a hetero-trimeric Mis18a/b complex made of 2 Mis18aand 1 Mis18b) abolished Mis18BP1 binding in vitro and in cells, consequently abolished CENP-A deposition (Spiller et al., 2017) and (ii) artificial dimerisation of Mis18BP1, by expressing Mis18BP1 as a GST-tagged protein, enhanced the centromere localisation of Mis18BP1 highlighting the requirement of Mis18a/b hexameric assembly mediated dimerization of Mis18BP1 in cells (Pan et al., 2017). While these studies highlighted the importance of maintaining the right stoichiometry (hetero-octamer of 4 Mis18a, 2 Mis18b and 2 Mis18BP1), lack of structural information on how this essential biological assembly is established remained a major knowledge gap. Our work presented here fills this critical knowledge gap by showing that a segment of Mis18BP1 (aa 20-51) also binds at the Yippee dimerisation interface. To highlight this, we have included the following statements in the introduction on page 5 and 20 "Perturbing the Yippee domain-mediated hexameric assembly of Mis18a/b (that resulted in a Mis18a/b hetero-trimer, 2 Mis18a and 1 Mis18b) abolished its ability to bind Mis18BP1 in vitro and in cells (Spiller et al., 2017), emphasising the requirement of maintaining correct stoichiometry of Mis18a/b subunits. Consistent with this, artificial dimerisation of Mis18BP1, by expressing Mis18BP1 as a GST-tagged protein, enhanced the centromere localisation of Mis18BP1 (Pan et al., 2017)." and in the Results section on page 14 line 12: "Mis18BP120-51 contains two short b strands that interact at Mis18a/b Yippee interface extending the six-stranded-b sheets of both Mis18a and Mis18b Yippee domains. This provides the structural rationale for why Yippee domains-mediated Mis18a/b hetero-hexamerisation is crucial for Mis18BP1 binding (Spiller et al., 2017)."

Regarding the question "whether 'old' centromeric CENP-A or HJURP:new CENP-A complex is needed to recruit Mis18a centromere localisation and whether identified residues have a role in Mis18a centromere localisation": According to the published literature, the Mis18 complex associates with centromeres through interaction with CCAN components CENP-C and CENP-I (Shono et al., 2015, Dambacher et al., 2012, Moree et al., 2011, Hoffmann et al., 2020). Considering CCAN assembles on CENP-A nucleosomes, and HJURP:new CENP-A centromere recruitment depends on the Mis18 complex, it will be reasonable to argue that the 'old' centromeric CENP-A contributes to the centromere localisation of the Mis18 complex. Amongst the components of the Mis18 complex, Mis18BP1 and Mis18bhave previously been suggested to interact with CENP-C. Within the Mis18 complex, we (Spiller et al., 2017) and others (Pan et al., 2017) have shown that Mis18a can directly interact with Mis18BP1, but it does so more efficiently when Mis18a hetero-oligomerises with Mis18b via their Yippee domains. Here, our structural analysis mapped the interaction interfaces and showed that Mis18a residues E103, D104 and T105 contribute to Mis18BP1 binding, as mutating these residues abolishes centromere localisation of Mis18a (Fig. 5c and 5d). To accentuate our findings, we have now included the following paragraph in the discussion section (page 17 line 26): "One of the key outstanding questions in the field is how does the Mis18 complex associate with the centromere. Previous studies identified CCAN subunits CENP-C and CENP-I as major players mediating the centromere localisation of the Mis18 complex mainly via Mis18BP1 (Shono et al., 2015, Dambacher et al., 2012, Moree et al., 2011), although Mis18b subunit has also been suggested to interact with CENP-C (Stellfox et al., 2016). Within the Mis18 complex, we and others have shown that the Mis18a/b Yippee hetero-dimers can directly interact with Mis18BP1. Here our structural analysis allowed us to map the interaction interface mediating Mis18a/b-Mis18BP1 binding. Perturbing this interface on Mis18a completely abolished Mis18a centromere localisation and reduced Mis18BP1 centromere levels. These observations show that Mis18a associates with the centromere mainly via Mis18BP1, and assembly of the Mis18 complex itself is crucial for its efficient centromere association, as previously suggested. Future work aimed at characterising the intermolecular contact points between the subunits of the Mis18 complex, centromeric chromatin and CCAN components and understanding if the Mis18 complex undergoes any conformational and/or compositional variations upon centromere association and/or during CENP-A deposition process, will be crucial to delineate the mechanisms underpinning the centromere maintenance."

Minor comments:

• The bar graphs shown ideally also show the individual data points for the authros to appreciate the spread of the data. These figures can be replicated in the Supplemental to avoid making the main figures look too busy.

We thank the reviewers for this suggestion. Reviewer #3 made a similar comment and suggested we use Superplot, which allows visualisation of individual data points of independent experiments. We have now revised all bar graphs using Superplot to address both reviewers' suggestions.

Reviewer #2 (Significance (Required)):

• This study uses a broad range of structural techniques, including molecular modeling which were subsequently validated by in vitro pull-down assays, co-IP, and IF. This combination of these techniques is important because many structural techniques cannot be performed under physiological conditions. Validating the main findings of the structural results by IF and co-IP is therefore critical.
• This work greatly advances our structural understanding how Mis18a, Mis18b, and Mis18BP1 form the Mis18 complex and how the critical residues in especially Mis18a help the Mis18 complex localize to the centromere and influence new CENP-A loading. This study also provides the first strong evidence in hierarchical assembly of the Mis18 complex.
• How centromere identity is maintained is a critical question in chromosome biology and genome integrity. The Mis18 complex has been identified as an important complex in the process. Several structural and mutational studies (all adequately cited in this manuscript) have tried to address which residues guide the assembly and functional regions of the Mis18 complex. This work builds and expands our understanding how especially Mis18a holds a pivotal role in both Mis18 complex formation and its impact on maintaining centromeric CENP-A levels.
• This work will be of interest to the chromosome field in general and anyone studying the mechanism of cell division.
• Chromatin, centromere, CENP-A, cell division. This reviewer has limited expertise in structural biology.

Reviewer #3 (Evidence, reproducibility and clarity (Required)):

Centromere identity is defined by CENP-A loading to specific sites on genomic DNA. CENP-A loading is known to rely on the Mis18 complex, and several regulators are known; yet how the Mis18 complex achieves this complex process has remained puzzle. By elucidating the structural basis of Mis18 complex assembly using integrative structural approaches the authors show that multiple homo and heterodimeric interfaces of Mis18alpha, beta and Mis18BP1 are involved in centromere maintenance. The authors show that Mis18alpha can associate with centromeres and deposit CENP-A independent of Mis18 β. Mis18α functions in CENP-A deposition at centromeres independent of Mis18β. Mis18β is required for maintaining a specific level of CENP-A occupancy at centromeres. Thus, using structure-guided and separation-of-function mutants the study reveals how Mis18 complex ensures centromere maintenance. Major comments: This is an excellent study on centromere inheritance, combining structural and cell biology techniques. The comments here primarily refer to Cell biology aspect of the work.

Figures show that new CENP-A deposits in Mis18βL199D/I203D mutants, but the level was reduced moderately. Based on this observation, the authors make a strong conclusion that Mis18β licenses the optimal levels of CENP-A at centromeres. Mis18α may be essential for both CENP-A incorporation and depositing a specific amount of CENP-A, as Mis18α and CENP-A levels are both reduced in Mis18βL199D/I203D mutants which failed to form the triple helical assembly with Mis18α as shown in Figure 3B and 3C. The authors may want to qualify some of these claims as preliminary or speculative.

We thank the reviewer for this suggestion. We agree that although the reduction in CENP-A levels upon replacing WT Mis18b with Mis18b L199D/I203D is more prominent than the reduction in centromere localised Mis18a, one cannot completely rule out the contribution of reduced Mis18a on CENP-A loading. This also raises an interesting possibility where Mis18b ensures the correct amount of CENP-A deposition by facilitating the optimal level of Mis18a at centromeres. We now explicitly discuss this in the discussion as follows (page 16 line 26): "Whilst proteins involved in CENP-A loading have been well established, the mechanism by which the correct levels of CENP-A are controlled is yet to be thoroughly explored and characterised. The data presented here suggest that Mis18b mainly contributes to the quantitative control of centromere maintenance - by ensuring the right amounts of CENP-A deposition at centromeres - and maybe one of several proteins that control CENP-A levels. We also note that the Mis18b mutant, which cannot interact with Mis18a, moderately reduced Mis18a levels at centromeres, and hence, it is possible that Mis18b ensures the correct level of CENP-A deposition by facilitating optimal Mis18a centromere recruitment. Future studies will focus on dissecting the mechanisms underlying the Mis18b-mediated control of CENP-A loading amounts along with any other mechanisms involved."

This work and others show that phosphorylation of Mis18BP1 by CDK1 can interfere with complex function (Spiller et al., 2017, Pan et al., 2017). Does the structure provide any insight into PLK1-mediated phosphorylation surfaces for activation of the complex? If yes, a brief discussion would help to link CDK1 and PLK1 mediated opposing actions will strengthen the work.

As described in our response to the first major comment of Reviewer 2, our very recent study aimed at understanding the licencing role of Plk1, independent of the work reported here, identified and evaluated the functional contribution of Plk1 phosphorylation on the subunits of the Mis18 complex (Parashara et al., bioRxiv 2024). Serendipitously, this recent work has now validated our hypothesis proposed based on the structural characterisation reported here and demonstrates that a Plk1-mediated phosphorylation cascade activates the Mis18a/b complex via a conformational switch of the N-terminal helical region of Mis18a which facilitates a robust HJURP-Mis18a/b interaction (Parashara et al. bioRxiv 2024). An independent study from the Musacchio lab (Conti et al., bioRxiv 2024) also reports similar findings, mutually strengthening our independent conclusions. Overall, these studies highlight the importance of the critical structural insights into the Mis18 complex this study reports. We now explicitly discuss the validation of our original hypothesis by citing our recent work along with that of the Musacchio lab. The corresponding section of the last paragraph now reads as follows (page 17 line 10): "Previously published work identified amino acid sequence similarity between the N-terminal region of Mis18a and R1 and R2 repeats of the HJURP that mediates Mis18a/binteraction (Pan et al., 2019). Deletion of the Mis18a N-terminal region enhanced HJURP interaction with the Mis18 complex (Pan et al., 2019). Here, we show that the N-terminal helical region of Mis18a makes extensive contact with the C-terminal helices of Mis18a and Mis18b, which had previously been shown to mediate HJURP binding by Pan et al., 2019. Collectively these observations suggest that the N-terminal region of Mis18a might directly interfere with HJURP - Mis18 complex interaction. Two independent recent studies (Parashara et al., 2024, Conti et al., 2024) reveal that this is indeed the case and a Plk1-mediated phosphorylation cascade involving several phosphorylation and binding events of the Mis18 complex subunits relieve the intramolecular interactions between the Mis18a N-terminal helical region and the HJURP binding surface of the Mis18a/b C-terminal helical bundle. This facilitates robust HJURP-Mis18a/b interaction in vitro and efficient HJURP centromere recruitment and CENP-A loading in cells. Overall, these studies also highlight the importance of the critical structural insights into the Mis18 complex we report here."

I am happy with the way cell biology data and the methods are presented so that they can be reproduced. The experiments are adequately replicated and the statistical analysis adequate. It will help to include sample size of cells or centromeres used for building the graphs.

We have now included this information in figure legends of Fig. 3a, 3c, 4b, 4c, 5b, 5c and 5d.

This is a strong interdisciplinary study using a variety of in vitro and in vivo techniques. Can the authors discuss if they expect chromatin associated Mis18 complex to host a similar structure as the soluble one? In other words, are they able to comment on any key differences between chromatin and non-chromatin associated Mis18 complexes.

We thank the reviewer for the suggestion. We agree that one cannot rule out the possibility of the Mis18 complex undergoing compositional and/or conformational variations during the processes of CENP-A loading at centromeres. We now explicitly discuss this possibility in the last paragraph of the discussion section (page 18 line 10): "Future work aimed at characterising the intermolecular contact points between the subunits of the Mis18 complex, centromeric chromatin and CCAN components and understanding if the Mis18 complex undergoes any conformational and/or compositional variations upon centromere association and/or during CENP-A deposition process, will be crucial to delineate the mechanisms underpinning the centromere maintenance."

Minor comments: -

In cell biology experiments, fluorescence intensities could be presented as a superplot for added value across cells and repeats (instead of bar graphs). More on superplot:https://doi.org/10.1083/jcb.202001064.

We thank the reviewers for this kind suggestion. We have now included graphs made using 'superplot' as suggested.

In general, ACA levels do not appear to change significantly between WT and mutant expressing cells although new CENP-A loading is significantly absent in the presence of a few mutants - please comment if ACA used here can recognise CENP-A. Would this mean that old CENP-A remains normally?

We thank the reviewer for this comment. While new CENP-A incorporated at centromeres is selectively labelled using the SNAP-tag, the ACA antibody used in these experiments can recognise CENP-A, CENP-B and CENP-C, with CENP-B being the primary target (Kallenberg, Clinical Rheumatology,1990). We would also like to note that ACA has commonly been used to locate the centromere in CENP-A loading assays where new CENP-A levels are assessed via selective labelling (e.g. McKinley 2014).

It is unclear whether any of the mutant acted in a dominant negative fashion in the presence of endogenous Mis18 proteins. It would have been useful to test this particularly in the context of mis18alpha mutants that seem to fully abolish new CENP-A recruitment.

As Mis18 subunits oligomerise (homo and hetero), we thought expressing these mutants in the presence of endogenous proteins might interfere with endogenous protein in a heterogenous manner and might make the interpretation difficult. Hence, we did not test this. Instead, as described in the manuscript we have tested these mutants in siRNA rescue experiments (Fig. 3, 4 and 5).

In figure 3a, GFP panel (input lane, 1) is shown to mark a band corresponding to GFP. Is this expected? Please comment.

Yes, as a control, an empty vector was transfected to express just GFP along with Mis18a-mCherry. These were used to show that there was no unspecific interaction between the beads used for IP or Mis18a-mCherry and GFP tag, and that any interaction seen was due to Mis18b. A similar control was used in S4b, where mCherry was expressed along with Mis18b-GFP. We have now clarified this in the corresponding legends of Fig. 4a and S4b.

Would be useful to have the scale for the cropped images presented as insets. Figure 4B should read YFP and not YPF.

We apologise for this typographical error. We have now corrected this.

The authors may want to explain whether the tag differences matter for their study (Case in point: His-SUMO-Mis18a191-233 WT and mutant His-MBP-Mis18b188-229 proteins).

The MBP tag was chosen to perform amylose pull-down assays, whereas the SUMO tag was chosen to increase the protein size. This is crucial as the C-terminal fragments of Mis18a and Mis18b are less than 50 amino acids long and are not easy to visualise by the band intensity in the Coomassie-stained SDS PAGE gels.

Reviewer #3 (Significance (Required)):

This work elucidates the structural basis of Mis18 complex assembly and the intermolecular interfaces essential for Mis18 functions. This is a significant advance in the field as it helps researchers in the field better understand CENP-A deposition and mechanism underpinning the maintenance of centromere identity. This is a broad area of research benefitting those studying cell division, genome stability, centromere identity and epigenetics might all be interested in and influenced by these findings. Novelty and strength lies in combining structural and cell biology work. Strengths of the work are structural details of the Mis18 complex. Minor weakness is the link between Mis18 structure and Centromere inheritance is limited to one immunostaining assay (I have mentioned this as a minor comment because addressing this may not be within the scope of this manuscript and is likely to require a repeat of a vast majority of the work with additional reagents which may not directly add value to the current manuscript).

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #3

Evidence, reproducibility and clarity

Centromere identity is defined by CENP-A loading to specific sites on genomic DNA. CENP-A loading is known to rely on the Mis18 complex, and several regulators are known; yet how the Mis18 complex achieves this complex process has remained puzzle. By elucidating the structural basis of Mis18 complex assembly using integrative structural approaches the authors show that multiple homo and heterodimeric interfaces of Mis18alpha, beta and Mis18BP1 are involved in centromere maintenance. The authors show that Mis18alpha can associate with centromeres and deposit CENP-A independent of Mis18 β. Mis18α functions in CENP-A deposition at centromeres independent of Mis18β. Mis18β is required for maintaining a specific level of CENP-A occupancy at centromeres. Thus, using structure-guided and separation-of-function mutants the study reveals how Mis18 complex ensures centromere maintenance.

Major comments:

This is an excellent study on centromere inheritance, combining structural and cell biology techniques. The comments here primarily refer to Cell biology aspect of the work.

1. Figures show that new CENP-A deposits in Mis18βL199D/I203D mutants, but the level was reduced moderately. Based on this observation, the authors make a strong conclusion that Mis18β licenses the optimal levels of CENP-A at centromeres. Mis18α may be essential for both CENP-A incorporation and depositing a specific amount of CENP-A, as Mis18α and CENP-A levels are both reduced in Mis18βL199D/I203D mutants which failed to form the triple helical assembly with Mis18α as shown in Figure 3B and 3C. The authors may want to qualify some of these claims as preliminary or speculative.
2. This work and others show that phosphorylation of Mis18BP1 by CDK1 can interfere with complex function (Spiller et al., 2017, Pan et al., 2017). Does the structure provide any insight into PLK1-mediated phosphorylation surfaces for activation of the complex? If yes, a brief discussion would help to link CDK1 and PLK1 mediated opposing actions will strengthen the work.
3. I am happy with the way cell biology data and the methods are presented so that they can be reproduced. The experiments are adequately replicated and the statistical analysis adequate. It will help to include sample size of cells or centromeres used for building the graphs.
4. This is a strong interdisciplinary study using a variety of in vitro and in vivo techniques. Can the authors discuss if they expect chromatin associated Mis18 complex to host a similar structure as the soluble one? In other words, are they able to comment on any key differences between chromatin and non-chromatin associated Mis18 complexes.

Minor comments:

In cell biology experiments, fluorescence intensities could be presented as a superplot for added value across cells and repeats (instead of bar graphs). More on superplot: https://doi.org/10.1083/jcb.202001064. In general, ACA levels do not appear to change significantly between WT and mutant expressing cells although new CENP-A loading is significantly absent in the presence of a few mutants - please comment if ACA used here can recognise CENP-A. Would this mean that old CENP-A remains normally?

It is unclear whether any of the mutant acted in a dominant negative fashion in the presence of endogenous Mis18 proteins. It would have been useful to test this particularly in the context of mis18alpha mutants that seem to fully abolish new CENP-A recruitment.

In figure 3a, GFP panel (input lane, 1) is shown to mark a band corresponding to GFP. Is this expected? Please comment. Would be useful to have the scale for the cropped images presented as insets.

Figure 4B should read YFP and not YPF.

The authors may want to explain whether the tag differences matter for their study (Case in point: His-SUMO-Mis18a191-233 WT and mutant His-MBP-Mis18b188-229 proteins).

Significance

This work elucidates the structural basis of Mis18 complex assembly and the intermolecular interfaces essential for Mis18 functions. This is a significant advance in the field as it helps researchers in the field better understand CENP-A deposition and mechanism underpinning the maintenance of centromere identity. This is a broad area of research benefitting those studying cell division, genome stability, centromere identity and epigenetics might all be interested in and influenced by these findings. Novelty and strength lies in combining structural and cell biology work.

Strengths of the work are structural details of the Mis18 complex. Minor weakness is the link between Mis18 structure and Centromere inheritance is limited to one immunostaining assay (I have mentioned this as a minor comment because addressing this may not be within the scope of this manuscript and is likely to require a repeat of a vast majority of the work with additional reagents which may not directly add value to the current manuscript).

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

Summary:

The manuscript "structural basis for Mis18 complex assembly: implications for centromere maintenance" by Thamkachy and colleagues describes a study that uses structural analysis to test essential candidate residues in Mis18 complex components in CENP-A loading. For chromosomes to faithfully segregate during cell division, CENP-A levels must be maintained at the centromere. How CENP-A levels are maintained is therefore important to understand at the mechanistic level. The Mis18 complex has been found to be important, but how exactly the various Mis18 complex components interact and how they regulate new CENP-A loading remains not fully understood. This study set out to characterize the critical residues using X-ray crystallography, negative staining EM, SEC analysis, molecular modeling (Raptorx, AlphaFold2, and AlphaFold-multimer) to identify the residues of Mis18a and Mis18b that are critical for the formation of the Mis18a/b hetero-hexamer and which residues are important for Mis18a and Mis18BP1 interactions. A complex beta-sheet interface dictates the Mis18a and Mis18b interactions. Mutating the Mis18a residues that are important for the Mis18a/b interactions resulted in impaired pull-down of Mis18b and reduced centromeric levels of mutated Mis18a. The functional consequences of mutating residues that impair Mis18a/b interactions is that with reduced centomeric levels of Mis18a, also impaired new CENP-A loading. Interestingly, mutated Mis18b did not impact centromeric Mis18a levels and only modestly impaired new CENP-A loading. These data were interpreted that Mis18a is critical for new CENP-A loading, whereas Mis18b might be involved in finetuning how much new CENP-A is loaded. Overall, it is a very well described and well written study with exciting data.

Major comments:

• Overall, the structural data and the IF data support the importance of Mis18a residues 103-105 are critical for centromeric localization and new CENP-A loading, whereas Mis18b residues L199 and I203 are critical for centromeric localization, but only very modestly impair centromeric Mis18a localization and new CENP-A loading. In the discussion the authors argue that the N-terminal helical region of Mis18a mediate HJURP binding. This latter is postulated based on published work, but not tested in this work. This should be clarified as such.
• Overall, the authors clearly describe their data and methodology and use adequate statistical analyses. The structural data of the Mis18a/b complex being a hetero-hexamer is convincing, but the validation in vivo is missing. As structural experiment are not performed under physiological conditions, it is important to establish the stoichiometry in vivo to further support the totality of the findings of the structural experiments and modeling. The data for the hierarchical assembly of Mis18a and Mis18b at the centromere and its importance in new CENP-A loading is convincing. An additional open question is whether "old" centromeric CENP-A or HJURP:new CENP-A complex is needed to recruit Mis18a to the centromere and whether the identified residues have a role in Mis18a centromeric localization. These data would provide a solid link between the Mis18 complex and how it is directly linked to new CENP-A loading.

Minor comments:

• The bar graphs shown ideally also show the individual data points for the authors to appreciate the spread of the data. These figures can be replicated in the Supplemental to avoid making the main figures look too busy.

Significance

• This study uses a broad range of structural techniques, including molecular modeling which were subsequently validated by in vitro pull-down assays, co-IP, and IF. This combination of these techniques is important because many structural techniques cannot be performed under physiological conditions. Validating the main findings of the structural results by IF and co-IP is therefore critical.
• This work greatly advances our structural understanding how Mis18a, Mis18b, and Mis18BP1 form the Mis18 complex and how the critical residues in especially Mis18a help the Mis18 complex localize to the centromere and influence new CENP-A loading. This study also provides the first strong evidence in hierarchical assembly of the Mis18 complex.
• How centromere identity is maintained is a critical question in chromosome biology and genome integrity. The Mis18 complex has been identified as an important complex in the process. Several structural and mutational studies (all adequately cited in this manuscript) have tried to address which residues guide the assembly and functional regions of the Mis18 complex. This work builds and expands our understanding how especially Mis18a holds a pivotal role in both Mis18 complex formation and its impact on maintaining centromeric CENP-A levels.
• This work will be of interest to the chromosome field in general and anyone studying the mechanism of cell division.
• Chromatin, centromere, CENP-A, cell division. This reviewer has limited expertise in structural biology.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

Summary

Maintenance of the histone H3 variant CENP-A at centromeres is necessary for proper kinetochore assembly and correct chromosome segregation. The Mis18 complex recruits the CENP-A chaperone HJURP to centromeres to facilitate CENP-A replenishment. Here the authors characterise the Mis18 complex using hybrid structural biology, and determine complex interface separation-of-function mutants.

Major Comments

The SAXS and EM data on the full-length Mis18 components must be included in the main Figures, either as an additional figure or by merging/rearranging the existing figures. The authors discuss these results in three whole paragraphs, which are a very important part of the paper.

Could the authors also compare the theoretical SAXS scattering curves generated by their final model(s) with the experimental SAXS curves? This would provide some additional evidence for the overall shape of their complex model beyond the consistency with the Dmax/Rg.

Minor Comments

While the introduction is clearly written, an additional cartoon schematic, representing the system/question would be helpful to a non-specialist reader to interpret the context of the study.

No doubt the authors had a reason for choosing their figure allocation, but I wonder if more material couldn't be brought from the supplementaries into the main figures?

Page 6 "Mis18-alpha possesses an additional alpha-helical domain" - please make it clear in addition to what (I assume it's in addition to Mis18-beta).

Page 7 - Report the RMSD of the Pombe vs. Human Mis18-alpha yipee structures?

Page 7 - "We generated high-confidence structural models...." is there a metric for the confidence as reported by RaptorX? Perhaps includinging the PAE plots in the supplementary for the AlphaFold generated models would be useful?

Figure 1 - Perhaps label figure 1b as being experimentally determined, with the R values (as for Figure 1d), and 1c being a predicted model.

Page 8 "This observation is consistent with the theoretically calculated pI of the Mis18alpha helix" This is a circular argument, of course this region has a low pI due to the amino acid composition. Please remove this statement.

Page 8 "...reveals tight hydrophobic interactions" these are presumably shown in Figure 1d rather than in the referenced 1e.

Page 8 - The authors should briefly somewhere discuss why there is a difference between their results and those in Pan et al 2009. As I understand it, the Pan et al paper was based in part on modelling with CLMS data as restraints.

Figure 1 - The labelling of the residues for Mis18-alpha in Figure 1d is problematic, they are black on dark purple (might be my printer/screen/eyes) suggest amending.

Figure S3a - Do the authors have some data to show the mass of the cross-linked complex that was loaded onto grids is consistent with what is expected?

Figure S3b - scale bar

Figure S3c - Could the authors show or explain the differences between these different 3D reconstructions?

Page 9 - The use of "AFM" for AlphaFoldMultimer" is a little confusing since AFM is the established acronym for Atomic Force Microscopy. Perhaps AF2M?

Figure S4a - Control missing for Mis18-alpha wild-type

Figure S4 d and e - The contrast between the bands and the background is very bad (at least in my copy).

Page 13 "Our structural analysis suggests that two Mis18BP1 fragments.....". How did you arrive at this conclusion? Is this based on the AlphaFold/RaptorX model? What additional evidence do you have that the positioning of the Mis18BP1 is correct? Does the CLMS data support this?

Figure 4a - Would the authors like to consider using a different colour for Mis18BP1? The contrast is not great, especially in the electrostatic surface inset.

Significance

General Assessment

The paper is extremely clearly written. Likewise the figures are beautifully presented and the data extremely clean and fully supportive of the authors conclusions. Indeed it is seldom that one sees the depth of the structural approaches (X-ray, CLMS, EM, SAXS) in one paper which is a huge strength of the manuscript. In addition the translation of this data into very clean cell biological experiments, makes the paper truly outstanding.

Advance

The authors provide the first model of the Mis18 complex, with extensive evidence to back up this model. The authors provide additional evidence as to how the deposition/renewal of CENP-A might be mediated by the Mis18 complex. The advance comes from both the level of clarity, detail, and scope achieved in this paper.

Audience

This will likely be of great interest to anyone with an interest in chromosome biology, plus be of interest to structural biologists as an outstanding example of hybrid structural biology.

Expertise

I am a biochemist with a background in structural biology with some familiarity with centromere biology

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

Reply to the Reviewers

We would like to thank the reviewers for their insightful comments, which will significantly guide us in enhancing our manuscript. We are capable of addressing most of the concerns raised by the reviewers. However, we encounter limitations in addressing Reviewer 2's comment regarding the delineation between cell autonomy and non-autonomy. As highlighted by the reviewer, ideally, we would dissect the mechanisms underlying cell turnover within the M/+ wing pouch, particularly in terms of cell autonomy and non-autonomy. Unfortunately, clonal analysis is not feasible due to the 'salt and pepper' distribution pattern of cell death and subsequent proliferation within the M/+ wing pouch. Despite these challenges, we commit to making efforts to address these issues to the best of our ability. Furthermore, we will aim to present our findings with greater precision and without speculation, an approach we believe will significantly enhance the quality of our manuscript.

In this manuscript, we extend the insights gained from our previous study (Akai et al., PLOS Genetics, 2021) by uncovering a novel mechanism within the context of the M/+ mutant. We demonstrate that JNK-dependent exocytosis in dying cells is crucial for cell turnover in the M/+ wing pouch. This phenomenon of cell turnover, initiated by cell death and the following proliferation, is fundamental to a variety of processes in multicellular organisms, such as normal development, tissue homeostasis, wound healing, tumor development, and potentially tumor recurrence. While our analysis is specific to the M/+ mutant, the underlying mechanisms of cell-cell communication between dying and proliferating cells, which are still not fully understood, may have broader implications. Our findings offer significant contributions to the field of cell competition and suggest a framework for understanding the operating principles of multicellular communities through cell-cell communications. Although it is yet to be determined how these insights apply in various contexts, they possess the potential to guide future research.

Referee #1

It would be important to check how much JNK is sufficient to trigger the exocytosis upregulation. Is the accumulation of Cyt1 and CD63 vesicles in apoptotic cell common to any JNK dependent death or does it require the Minute background? Could the authors check whether clones expressing HepCA transiently in WT background also accumulate the same vesicles?

Response:

Following the reviewer's suggestion, we plan to examine whether the accumulation of Syt1- and CD63-positive vesicles in apoptotic cells is a general characteristic of JNK-dependent cell death or requires the Minute background. Specifically, we intend to check whether mitotic clones expressing HepCA in the wild-type eye-antennal disc to see if these vesicles also accumulate. Additionally, we will assess mitotic clones expressing Eiger in the wild-type eye-antennal disc, which activates JNK signaling. This is due to concerns that HepCA may induce cell death too strongly, potentially resulting in clones too small for effective analysis. Moreover, we also plan to express HepCA transiently only in the posterior region of the wing disc to further explore its effects.

1. It would be relevant to have the status of JNK in Minute disc upon downregualtion of exocytosis. One could imagine some positive feedback between JNK activation, exocytosis, cell death and further JNK activation.

Response:

We thank the reviewer for the comment. As pointed out by the reviewer, there could be a positive feedback loop involving JNK activation, exocytosis, cell death, and further JNK activation. Our observations lend support to this hypothesis, as we specifically noted a significant reduction in the number of JNK-activating cells following targeted rab3 knockdown in these cells using puc-gal4 driver within the RpS3/+ wing disc (Fig R1, below). To further confirm the positive feedback loop between JNK activation and exocytosis, we plan to investigate the effects of downregulating unc-13 on JNK activation within the M/+ wing pouch, employing the JNK reporter TRE-DsRed for this purpose.

Additionally, we need to elucidate the molecular mechanism linking cell death and exocytosis. We are currently exploring two potential relationships between JNK activation, caspase signaling, and exocytosis, as depicted in Fig R2 below. To assess these possibilities, we aim to investigate if inhibiting apoptosis-either by introducing the H99/+ mutation or by overexpressing DroncDN or mirRHG-can affect JNK-mediated exocytosis in the RpS3/+ wing pouch.

So far, the evidence for the epistatic link between exocytosis, Wg and cell turnover is mostly based on colocalization and the similarity of the phenotype but I believe this may need some additional evidences. Ideally one would need to be able to enhance exocytosis and test whether Wg downregulation suppress the phenotype, but I am not sure that upregulation of core exocytosis genes will be sufficient to do this. Alternatively, if Wg is indeed downstream of the upregulation of exocytosis, the reduction of cell turnover upon Wg flattening (e.g. : ftz2-/+ background) should not be enhanced by the reduction of exocytosis. Moreover, could the authors test the status of Wg downstreams targets upon inhibition of exocytosis in the Minute background (for instance, do they see a supression of the nmo-LacZ upregulation that they previously characterised in Minute wing disc in 2021) ?

Response:

We thank the reviewer for the comment. In line with the reviewer's suggestion, exploring the enhancement of exocytosis would be valuable to elucidate the epistatic relationship between exocytosis, Wg signaling, and cell turnover. Therefore, we plan to attempt the overexpression of exocytosis-related genes to check if we can indeed enhance exocytosis. Additionally, in response to the reviewer's comment, we intend to investigate whether the reduction in cell turnover observed upon Wg attenuation (ftz2-/+background or Wg-/+ background) is not exacerbated by further reducing exocytosis. Furthermore, we intend to investigate the status of the Wg downstream target, specifically using nmo-lacZ, when exocytosis is inhibited in the RpS3/+ wing pouch.

Since Cyt1 and CD63 seem to mostly accumulate in apoptotic cells, it would be interesting to check their status in Minute wing disc upon apoptosis inhibition (e.g. : with H99 or mirRHG).

Response:

We thank the reviewer for the comment. Following the reviewer's suggestion, we intend to investigate whether inhibiting apoptosis, either through the introduction of H99/+ or by overexpressing DroncDN or mirRHG, could suppress the increase in CD63- or Sty1-positive vesicles in the M/+ wing pouch.

I would remain cautious about some of the statements, notably in the abstract, since some of them are mostly speculative and not really based on any experiments. For instance, the statement "This interaction between dying cells and their neighboring living cells is pivotal in determining cell fate, dictating which cells will undergo apoptosis and which cells will proliferate" is not backed up by any experiment (which would require to show that exocytosis and Wg from the dying cell specifically is required for the survival and proliferation of their neighbours, and/or showing that cell death occurs specifically in cells with local differences in Wg signaling). I would recommend to be more cautious here and us a clear conditional statement.

• *Response:

We thank the reviewer for their insightful comment. In response, we aim to present our findings with greater precision and to avoid speculative interpretations. At this stage, we have focused on revising the Abstract to include clear conditional statements, as suggested. Furthermore, we plan to comprehensively update the remaining sections of the manuscript to reflect the additional experiments requested by the reviewers. These updates will ensure a more cautious approach throughout the paper, aligning with the reviewer's recommendations.

__(page 2, line 34-39 in the "Abstract") __

"Our data also suggest a potential role for the Wg receptor Frizzled-2 (Fz2) in inducing cell-turnover within the M/+ wing pouch. Overall, our findings provide mechanistic insights into robust tissue growth through the orchestration of cell-turnover, which is primarily governed by JNK-mediated exocytosis in the context of Drosophila Minute/+ wing morphogenesis."

Other minor point:

The authors document Wg localisation in Minute wing disc upon expression of P35. It would be interesting to describe what is the status of Wg in Rps3+/- compared to WT without p35 (if I am correct, this was done in their previous article, and in that case it would be relevant to describe these former results in the main text).

Response:

We thank the reviewer for the comment. In the RpS3/+ wing pouch without p35, we were unable to detect any upregulation of Wg using the anti-Wg antibody (Fig R3, below). However, we observed an increase in GFP-Wg-positive puncta originating from a knock-in allele (McGough et al., Nature, 2020) within areas of massive cell death in the RpS3/+ wing pouch lacking p35, compared to the wild-type control (Fig 3B, compared to Fig 3A in the transferred manuscript). This increase is similar to the phenotype observed in the RpS3/+ wing pouch expressing p35, where Wg-positive puncta are significantly elevated (Fig R4B below, corresponding to Fig 3B in the original manuscript). Moreover, the increase in GFP-Wg-positive puncta in regions of massive cell death in the RpS3/+ wing pouch becomes more pronounced when utilizing a membrane-tethered anti-GFP nanobody (Vhh4-CD8) (McGough et al., Nature, 2020) (Fig 3G, compared to Fig 3F in the transferred manuscript). These observations indicate that Wg-positive puncta are indeed upregulated in the RpS3/+ wing pouch without p35 compared to the wild-type control.

In the transferred manuscript, we have now made modifications to Fig. 3 by replacing images of the RpS3/+ wing pouch overexpressing p35, which were stained with the anti-Wg antibody, with new images. These new images depict the RpS3/+ wing pouch that harbors the GFP-Wg knock-in allele, which was stained with both anti-GFP and anti-cDCP1 antibodies.

To clarify this point, we have now modified the sentence as follows:

(page 7, line 204-210)

Interestingly, we found that the RpS3/+ wing pouch expressing the cell death inhibitor p35 (which allows dying RpS3/+ cells to survive) exhibited elevated levels of Wg protein, compared to the wild-type control (Fig 3A and 3B). This increase in Wg protein was significantly diminished by overexpressing the JNK inhibitor Puc (Fig 3C), suggesting that Wg expression is upregulated via JNK signaling in the M/+ wing pouch, similar to apoptotic cells in which JNK signaling induces the production of secreted growth factor, including Wg (18, 53, 54). In addition, ____W____e found t____hat GFP-Wg-positive puncta, derived from a knock-in allele ____(34)____, ____were more abundant in the RpS3/+ wing pouch compared to the wild-type control (Fig 3A and 3B). This increase in GFP-Wg-positive puncta, ____especially in the area with massive cell death within the RpS3/+ wing pouch (Fig S3C-S3D'')_, was more evident when using a membrane-tethered anti-GFP nanobody (Vhh4-CD8), which immobilizes GFP-Wg on the cell surface _(34)____ (Fig 3F and 3G, quantified in Fig 3J).

Referee #2

Primary Concerns:

A significant challenge arises concerning the delineation of cell autonomy/non-autonomy. This study focuses on two distinct cell types, namely dying cells and proliferating cells. However, the consistent use of nub-gal4, a wing pouch driver, and the heterozygous minute mutant throughout the paper impedes the ability to conclusively analyze the autonomy of events. The authors previously posited that caspase-induced cell death triggers non-autonomous proliferation, but recent studies also suggested caspase-induced autonomous proliferation in both flies and mammals (Yosefzon et al. Mol. Cell 2018, Shinoda et al., PNAS 2019). Therefore, a meticulous distinction between autonomous and non-autonomous events, particularly through experimentation involving clones, is imperative. This necessity is particularly evident in Fig 4.

Response:

• As pointed out by the reviewer, it is ideal to dissect the mechanisms underlying cell turnover within the M/+ wing pouch, especially concerning cell autonomy and non-autonomy. However, clonal analysis is not feasible due to the pattern of cell death and subsequent proliferation occurring in a 'salt and pepper' distribution within the M/+* wing pouch. In response to this challenge, we found that almost dying cells do not undergo proliferation, as assessed by staining with the M phase marker phospho-Histone H3 (Fig R5, below).

We also plan to investigate whether JNK-activating cells similarly refrain from proliferating, which will be assessed by staining with the anti-phospho-Histone H3 antibody.

Additionally, about Fig 4, considering the evidence presented in this study, which demonstrates that dying cells increase Wg secretion through JNK-dependent exocytosis (Fig 3G-3H'), and given that Fz2 is expressed in adjacent cells within the M/+ wing pouch (Fig 4B-C'' and Fig S4B-C''), it is plausible that these neighboring cells could receive Wg via Fz2, leading to the upregulation of Wg signaling in these cells in the M/+ wing pouch. This upregulation of Wg signaling is supported by Fig S3A in the transferred manuscript. However, we were unable to delineate the specific role of Fz2, particularly in terms of cell autonomy and non-autonomy. In response to this challenge, we aim to explore whether cell turnover can be inhibited by enhancing Wg signaling exclusively in JNK-activated cells through the overexpression the active form of ArmadilloS10 (Baena-Lopez LA et al., Sci Signal., 2009), utilizing the puc-gal4 driver. Should cell turnover be inhibited under these conditions, it would indicate that differences in Wg signaling activity between dying cells and their neighboring cells drive cell turnover. While the exact mechanism of Fz2 remains unclear, the inhibition of cell turnover under these conditions clearly demonstrates that differences in Wg signaling activity play a significant role in cell turnover. Additionally, in our response to Reviewer 2's comment No. 2, we outline our intention to investigate whether the increase in Wg signaling could be induced by JNK-dependent exocytosis. Accordingly, we plan to assess the effects of downregulating JNK signaling or exocytosis on the elevated Wg signaling activity observed in the RpS3/+ wing pouch.

Furthermore, we intend to present our findings with greater precision, steering clear of speculative interpretations. At this stage, we have focused on revising the Abstract to include clear conditional statements, as detailed below. We plan to comprehensively update the remaining sections of the manuscript to reflect the additional experiments requested by the reviewers.

(page 2, line 34-39 in the "Abstract")

"Our data also suggest a potential role for the Wg receptor Frizzled-2 (Fz2) in inducing cell-turnover within the M/+ wing pouch. Overall, our findings provide mechanistic insights into robust tissue growth through the orchestration of cell-turnover, which is primarily governed by JNK-mediated exocytosis in the context of Drosophila Minute/+ wing morphogenesis."

Closely tied to the issue of cell autonomy/non-autonomy is the question of how cells differentiate Wg from dying cells and the dorsal-ventral boundary.

Response:

Wg is expressed at the dorsal-ventral boundary in both wild-type and M/+ wing discs. However, we observed in our previous study that Wg signaling activity was significantly more elevated in the RpS3/+ pouch compared to the localized activation in wild-type controls at the same developmental stage, as assessed by the nmo-lacZ reporter (Fig 3C-D' in Akai et al., PLOS Genetics, 2021, as also shown in Fig R6A-B' below). Interestingly, the areas of massive cell death in the RpS3/+ wing pouch always corresponded to the areas of relatively lower Wg signaling activity (Fig S3A in the transferred manuscript). Moreover, our previous study revealed that decreasing or increasing Wg signaling activity, thus reducing the aberrant Wg signaling gradient, significantly inhibited cell death in the M/+ wing pouch (Fig 3I-K in Akai et al., PLOS Genetics, 2021, as also shown in Fig R6C-E' below). This suggest that the aberrant Wg signaling gradient is crucial for massive cell-turnover in the M/+ wing pouch. As also described above, considering the evidence presented in this study, which demonstrates that dying cells increase Wg secretion through JNK-dependent exocytosis (Fig 3G-3H'), and given that Fz2 is expressed in adjacent cells (Fig 4B-C'' and Fig S4B-C'') within the M/+ wing pouch, it is plausible that these neighboring cells could receive Wg via Fz2, leading to the upregulation of Wg signaling in these cells within the M/+ wing pouch. To further investigate whether the increase in Wg signaling could be induced by JNK-dependent exocytosis, we plan to examine the effects of downregulating JNK signaling or exocytosis on the elevated Wg signaling activity observed in the RpS3/+ wing pouch. Should Wg signaling activity decrease as a result of these genetic interventions, it would imply that the enhanced Wg signaling in the M/+ wing pouch depends on JNK-mediated exocytosis.

In Fig 1, the authors interpret the upregulation of exocytosis-related genes as indicative of increased exocytosis. However, this interpretation lacks direct evidence and overlooks the possibility of opposing effects. For example, autophagosome accumulation means either activation or inhibition of autophagy. Or, in case of Dilp secretion, absence of vesicles indicates upregulation of secretion. To substantiate their claim, the authors must provide more conclusive evidence of increased exocytosis. Fig 2 suggests that inhibiting exocytosis-related genes suppresses caspase activation, favoring the proposition that exocytosis is upregulated. However, demonstrating a direct increase in exocytosis in minute cells would bolster their argument. Higher resolution imaging of the exosome marker, with overlayed images, would enhance clarity too.

Response:

We have demonstrated an increase in the number of vesicles positive for EGFP-CD63 (an exosome marker) and Syt1-EGFP (a vesicle marker) in the RpS3/+ wing pouch compared to the wild-type control (Fig 1A, 1B, 1G, and 1H in the transferredmanuscript). To more directly investigate whether exocytosis is indeed elevated in the M/+ wing pouch, we plan to assess if cells activating JNK signaling upregulate the production of extracellular vesicles. This will be done by specifically expressing the EGFP-CD63 probe in JNK-activating cells, utilizing the puc-gal4 driver for targeted expression. Additionally, following the reviewer's suggestion, we plan to prepare high resolution images of the exosome marker, with overlayed images in the revised manuscript.

Fig 3 introduces ambiguity regarding the relationship between cell death and exocytosis. The authors assert that dying cells exhibit elevated Wg protein levels compared to the wild-type control but omit an important comparison to the minute disc. Moreover, while Figs 1-2 propose a signaling cascade involving minute>JNK>exocytosis>cell death, Fig 3 implies that cell death regulates exocytosis. The coherence of their model and logic requires clarification - specifically, elucidating the molecular coupling mechanism between cell death and exocytosis.

Response:

We observed an increase in GFP-Wg-positive puncta originating from a knock-in allele (McGough et al., Nature, 2020) in the area of massive cell death within the RpS3/+ wing pouch, compared to the wild-type control (Fig 3A and 3B in the transferredmanuscript). However, as pointed out by the reviewer, it is challenging to determine whether the GFP-Wg-positive puncta are emanating from dying cells, given that Wg is secreted from the cells that produce it. To address this issue, we employed a membrane-tethered anti-GFP nanobody (Vhh4-CD8 morphotrap) designed to capture GFP-Wg on the cell surface, thereby preventing the diffusion of GFP-Wg from its producing cells. By utilizing the Vhh4-CD8 morphotrap, we found that GFP-Wg levels are indeed elevated in areas of cell death compared to adjacent regions within the RpS3/+ wing pouch (Fig 3G in the transferred manuscript).

To clarify this point, we have now modified the sentence as follows:__ __

(page 7, line 204-210)

In addition, ____W____e found t____hat GFP-Wg-positive puncta, derived from a knock-in allele ____(34)____, ____were more abundant in the RpS3/+ wing pouch compared to the wild-type control (Fig 3A and 3B). This increase in GFP-Wg-positive puncta, ____especially in the area with massive cell death within the RpS3/+ wing pouch (Fig S3C-S3D'')_, was more evident when using a membrane-tethered anti-GFP nanobody (Vhh4-CD8), which immobilizes GFP-Wg on the cell surface _(34)____ (Fig 3F and 3G, quantified in Fig 3J).

Additionally, as pointed out by the reviewer, we need to elucidate the molecular mechanism linking cell death and exocytosis. We are currently exploring two potential relationships between JNK activation, caspase signaling, and exocytosis, as depicted in Fig R2 below. To assess these possibilities, we aim to investigate if inhibiting apoptosis-either by introducing the H99/+ mutation or by overexpressing DroncDN or mirRHG-can affect JNK-mediated exocytosis in the RpS3/+ wing pouch.

Specific Points:

In Fig S1D, contrary to the authors' claim, cadp2 is not upregulated in a JNK-dependent manner.

Response:

We apologize for any confusion caused. We found that Cadps expression in the RpS3/+ wing pouch was increased compared to both the wild-type control (2.29-fold increase, RpS3/+ compared to wild-type) and the RpS3/+ wing pouch expressing Puckered (Puc), driven by the nub-gal4 driver (3.33-fold increase, RpS3/+ compared to RpS3/+ + Puckered). This suggests that the increase in Cadps expression in the RpS3/+ wing pouch is dependent on JNK signaling. We have now revised Fig. S1D for clearer representation. We also have made modifications in the transferred manuscript as follows:

(page 5, line 107-113)

"Mining the list of genes differentially expressed in the RpS3/+ wing pouch cells dependent on JNK signaling (Fig S1C and S2 Table), we noticed that among the genes associated with the "secretion by cell" GO term (Fig S1B), the evolutionarily conserved exocytosis-related genes unc-13, SNAP25, and cadps (Calcium-dependent secretion activator) were upregulated in a JNK-dependent manner (2.29-fold increase, RpS3/+ compared to wild-type; 3.33-fold increase, RpS3/+ compared to RpS3/+ +_ Puckered)_ (Fig S1D)."

When detecting multiple proteins in the same tissue, it is advisable for the authors to present overlayed images to enhance the clarity of their findings. Many pictures require higher magnification too.

Response:

Following the reviewer's suggestion, we have overlayed images in Fig 1A-D, Fig 1G-J, and Fig S1E in the transferredmanuscript. Additionally, following the reviewer's suggestion, we plan to prepare high resolution images in the revised manuscript.

In Fig 1A, the observed upregulation of EGFP-CD63 and Syt1-EGFP may potentially result from an artifactual effect of apoptosis. To validate their findings are specific, the authors should include negative controls that do not exhibit upregulation in dying cells.

Response:

Following the reviewer's suggestion, we used the CD8-PARP-Venus probe as a negative control. We observed that CD8-PARP-Venus-positive puncta were minimally present (Fig R7, below), indicating that the upregulation of EGFP-CD63 and Syt1-EGFP in the RpS3/+ wing pouch is indeed occurring, rather than being an artifactual effect of apoptosis. We intend to incorporate this negative control into the revised manuscript.

Referee #3

Major comments:

Figure 1A-H, S1E. The authors used EGFP-CD63 and Syt1-EGFP as markers of exocytosis. They see an increased number of puncta in apoptotic cells. Is this a specific effect in the dying cells in M mutant discs, or is it a general effect in apoptotic cell death? This should be examined in a condition where apoptosis is induced independently of M mutants such as nub-reaper or nub-hid.

Response:

We thank the reviewer for the comment. To ascertain whether the increase in EGFP-CD63-/Sty1-EGFP-positive puncta is specific to dying cells in M/+ mutants or a general characteristic of apoptotic cell death, we examined the presence of EGFP-CD63-positive puncta in the wing pouch, where Reper (Rpr) was expressed under the control of the nub-gal4 driver. Unfortunately, the expression of Rpr driven by nub-gal4 resulted in significant cell death, preventing us from drawing a definitive conclusion (Fig R8, below). Nonetheless, we did observe EGFP-CD63-positive puncta under these conditions, as indicated by the arrowheads in Fig R8 below. To further investigate, we plan to induce temporary expression of Rpr in the wing pouch using a combination of the temperature-sensitive Gal80 and the nub-gal4 driver.

Is exocytosis actually upstream or downstream of cell death, or both? The authors are kind of vague about it. On one hand, they say the dying cells induce exocytosis and secrete Wg. On the other hand, unc13RNAi can suppress cell death (Figure 1D', J'). Please clarify.

Response:

As pointed out by the reviewer, we need to clarify whether exocytosis actually occurs upstream or downstream of cell death. We are currently exploring two potential relationships between JNK activation, caspase signaling, and exocytosis, as depicted in Fig R2 below. To assess these possibilities, we aim to investigate if inhibiting apoptosis-either by introducing the H99/+ mutation or by overexpressing DroncDN or mirRHG-can affect JNK-mediated exocytosis in the RpS3/+ wing pouch.

Figure 1L-P. The observation of Ca++ flashes is very interesting. However, are they important for exocytosis, cell death and compensatory proliferation? Right now, this is just a stand-alone observation. Can mutants affecting Ca++ signaling block exocytosis, cell death and comp prol?

Response:

We thank the reviewer for the comment. we plan to investigate whether the downregulation of Ca++ signaling impacts exocytosis, cell death, and compensatory proliferation in the M/+ wing pouch. This will be achieved by downregulating genes essential for sensing intracellular calcium concentrations (Rizo J and Rosenmund C, Nat Struct Mol Biol., 2008; Sudhof TC, Annu Rev Neurosci., 2004) and genes encoding voltage-gated calcium channels (Kuromi H et al., Neuron, 2004).

Figure 3B'. Can you visualize aberrant wg expression in RpS3/+ wing discs only in the presence of p35? The expression of p35 in apoptotic cells generates undead cells which by itself induce Wg expression in a JNK-dependent manner (Perez-Garijo et al 2004; Ryoo et al 2004). Also, in Figure 3B', Wg is upregulated in the ventral half of the pouch, whereas in other figures (4B for example) apoptosis is strongly induced in the dorsal half. Does that make sense?

Response:

We thank the reviewer for the comment. In the RpS3/+ wing pouch without p35, we were unable to detect any upregulation of Wg using the anti-Wg antibody (Fig R3, below). However, we observed an increase in GFP-Wg-positive puncta, derived from a knock-in allele (McGough et al., Nature, 2020), in areas of massive cell death within the RpS3/+ wing pouch without p35, relative to the wild-type control (Fig 3B, compared to Fig 3A in the transferred manuscript). This increase is similar to the phenotype observed in the RpS3/+ wing pouch expressing p35, where Wg-positive puncta are significantly elevated (Fig R4B below, corresponding to Fig 3B in the original manuscript). Moreover, the increase in GFP-Wg-positive puncta in regions of massive cell death in the RpS3/+ wing pouch becomes more pronounced when utilizing a membrane-tethered anti-GFP nanobody (Vhh4-CD8) (McGough et al., Nature, 2020) (Fig 3G, compared to Fig 3F). These observations indicate that Wg-positive puncta are indeed upregulated in the RpS3/+ wing pouch without p35 compared to the wild-type control.

In the transferred manuscript, we have now made modifications to Fig. 3 by replacing images of the RpS3/+ wing pouch overexpressing p35, which were stained with the Wg antibody, with new images. These new images depict the RpS3/+ wing pouch that harbors the GFP-Wg knock-in allele, which was stained with both anti-GFP and anti-cDCP1 antibodies.

To clarify this point, we have now modified the sentence as follows:

(page 7, line 204-210)

Interestingly, we found that the RpS3/+ wing pouch expressing the cell death inhibitor p35 (which allows dying RpS3/+ cells to survive) exhibited elevated levels of Wg protein, compared to the wild-type control (Fig 3A and 3B). This increase in Wg protein was significantly diminished by overexpressing the JNK inhibitor Puc (Fig 3C), suggesting that Wg expression is upregulated via JNK signaling in the M/+ wing pouch, similar to apoptotic cells in which JNK signaling induces the production of secreted growth factor, including Wg (18, 53, 54). In addition, ____W____e found t____hat GFP-Wg-positive puncta, derived from a knock-in allele ____(34)____, ____were more abundant in the RpS3/+ wing pouch compared to the wild-type control (Fig 3A and 3B). This increase in GFP-Wg-positive puncta, ____especially in the area with massive cell death within the RpS3/+ wing pouch (Fig S3C-S3D'')_, was more evident when using a membrane-tethered anti-GFP nanobody (Vhh4-CD8), which immobilizes GFP-Wg on the cell surface _(34)____ (Fig 3F and 3G, quantified in Fig 3J).

Additionally, we apologize for any confusion caused by Fig 3B'. It is noteworthy that cell death sometimes occurs more prominently in the ventral half of the RpS3/+ wing pouch than in the dorsal half (Fig R9, below). Furthermore, the increase in Wg level can be more pronounced on the ventral side, or at times, it is equally strong on both the dorsal and ventral sides (Fig S3C-D in the transferred manuscript).

Can wg RNAi in cells destined to die (puc-Gal4 UAS-wgRNAi) suppress apoptosis and comp prol?

Response:

Following the reviewer's suggestion, we intend to investigate whether expressing Wg-RNAi in JNK activated cells, using the puc-gal4 driver, could suppress apoptosis and compensatory proliferation.

Figure 3D,E. Use cDcp1 as apoptotic marker instead of CD63-mCherry which is not an apoptotic marker.

Response:

Following the reviewer's suggestion, we have now utilized the cDCP1 antibody as an apoptotic marker in Fig S3B-D'' in the transferred manuscript, as described in the Reviewer 3's comment No.4.

Figure 4A' and B'. I am not sure there is much of a difference in the expression of fz2-lacZ in wild-type and RpS3/+ discs. The quantification in 4F shows there is no difference in the dorsal half of the pouch where massive apoptosis occurs. Can you generate a condition in which only one compartment (anterior or posterior) is mutant for RpS3/+, while the other compartment is wt? That would allow direct side-by-side comparison. Comparisons between different discs is problematic.

Response:

We thank the reviewer for the comment. Following the reviewer's suggestion, we conducted a more detailed investigation into the differences in fz2-lacZ expression levels between wild-type and RpS3/+ wing discs. We found that the expression level of fz2-lacZwas indeed elevated in the RpS3/+ wing pouch. This elevation in expression is demonstrated by the results of the genetic rescue experiment in the posterior compartment of RpS3/+ wing discs, where overexpression of RpS3 using the engrailed-gal4 driver led to a reduction in fz2-lacZ expression compared to the anterior RpS3/+ control (Fig R10, below).

After conducting statistical analysis, we plan to include Fig R10 in the revised manuscript.

There is some inconsistency in the presence of apoptotic cells and presence of cells which secrete Wg. Apoptotic cells occur in large cell clusters (Fig. 2G, 3H', S3A', S4B), while markers of exocytosis and Wg are present in individual puncta (Fig 3D', E', S3C,E). That does not seem to fit with the authors' conclusion that dying cells secrete Wg by exocytosis. Are all dying cells secreting Wg through exocytosis? Please explain.

Response:

We thank the reviewer for the comment. As highlighted by the reviewer, it appears that not all dying cells secrete Wg through exocytosis. Beyond exosomes, various mechanisms for Wg/Wnt transport have been proposed, including those mediated by lipoprotein particles, the cell-surface proteoglycan Dally-like protein (Dlp), and filopodia-like cellular extensions known as cytonemes. Therefore, it's plausible that dying cells may also utilize these additional mechanisms for Wg transport alongside exosomes. We have incorporated this explanation into the transferred manuscript as follows:

(page 9, line 272-286)

Previous studies have identified exosomes as carriers of Wnt/Wg in the extracellular space of both mammalian and Drosophila cells, including wing disc cells (28, 49, 50, 51, 52, 64). Concurrently, alternative mechanisms for Wg transport, such as those involving lipoprotein particles or a lipocalin Swim in the Drosophila wing disc, have been reported (29, 65). Additionally, the cell-surface proteoglycan Dally-like-protein (Dlp) has been reported to enable long-range signaling of the palmitoylated Wg _(51)._ Intriguingly, Wnt/Wg transport has also been observed through filopodia-like cellular extensions known as cytonemes (66, 67, 68, 69). In our study of the M/+_ wing disc, a model characterized by massive cell-turnover, we observed partial colocalization of Wg-GFP puncta with exosomal markers such as CD63-mCherry and Hrs._ Our data also suggest that Wg secretion through exocytosis may not uniformly occur among all dying cells within this context. It is noteworthy that whereas dying cells frequently form large clusters, exosomal markers and Wg typically localize within individual puncta. This disparity suggests that while exosomes from dying cells significantly contribute to Wg transport within the ____M/+____ wing pouch, other pathways may also be operative.

I was a bit confused by the conclusion of the authors about the results in Figure 4G-I. What do they mean by "difference in Fz2 expression"? Are they referring to the gradient that they described in their previous work? Or is it just the absolute level of Fz2 that determines apoptosis and proliferation? If the latter, is the gradient still present (at least in fz2/+), just at a lower level?

Response:

I apologize for any confusion caused by the way we presented our conclusions regarding the results depicted in Figure 4G-I in the original manuscript.

In our previous study, we observed that Wg signaling activity was elevated much more broadly in the RpS3/+ pouch compared to the localized activation observed in the wild-type control at the same developmental stage, as assessed by the nmo-lacZ reporter(Fig 3D in Akai et al., PLOS Genetics, 2021, as also shown Fig. R6 below). Interestingly, the areas of massive cell death in the RpS3/+ wing pouch always corresponded to the areas of relatively lower Wg signaling activity (as also shown in Fig S3A in the transferred manuscript). Moreover, our previous study revealed that decreasing or increasing Wg signaling activity, thus reducing the aberrant Wg signaling gradient, significantly inhibited cell death in the M/+ wing pouch (Fig 3I-K in Akai et al., PLOS Genetics, 2021, as also shown in Fig R6C-E' below). This suggest that the aberrant Wg signaling gradient is crucial for massive cell-turnover in the M/+ wing pouch.

Considering the evidence presented in this study, which demonstrates that dying cells increase Wg secretion through JNK-dependent exocytosis (Figure 3G-3H'), and given that Fz2 is expressed in adjacent cells within the M/+ wing pouch (Figure 4B-C'' and Fig S4B-C''), it is plausible that these neighboring cells could receive Wg via Fz2, leading to the upregulation of Wg signaling in these cells in the M/+ wing pouch.

However, we were unable to delineate the specific role of Fz2, particularly in terms of cell autonomy and non-autonomy as highlighted by Reviewer 2, through clonal analysis due to the occurrence of cell death and subsequent proliferation in a salt-and-pepper pattern within the M/+ wing pouch. In response to this challenge, we aim to explore whether cell turnover can be inhibited by enhancing Wg signaling exclusively in JNK-activated cells through the overexpression the active form of ArmadilloS10(Baena-Lopez LA et al., Sci Signal., 2009), utilizing the puc-gal4 driver. Should cell turnover be inhibited under these conditions, it would indicate that differences in Wg signaling activity between dying cells and their neighboring cells drive cell turnover. While the exact mechanism of Fz2 remains unclear, the inhibition of cell turnover under these conditions clearly demonstrates that differences in Wg signaling activity play a significant role in cell turnover. Additionally, in our response to Reviewer 2's comment No.2, we outline our intention to investigate whether the increase in Wg signaling could be induced by JNK-dependent exocytosis. Accordingly, we plan to assess the effects of downregulating JNK signaling or exocytosis on the elevated Wg signaling activity observed in the RpS3/+ wing pouch.

Furthermore, we intend to present our findings with greater precision, steering clear of speculative interpretations. At this stage, we have focused on revising the Abstract to include clear conditional statements, as detailed below. We plan to comprehensively update the remaining sections of the manuscript to reflect the additional experiments requested by the reviewers.

(page 2, line 34-39 in the "Abstract")

"Our data also suggest a potential role for the Wg receptor Frizzled-2 (Fz2) in inducing cell-turnover within the M/+ wing pouch. Overall, our findings provide mechanistic insights into robust tissue growth through the orchestration of cell-turnover, which is primarily governed by JNK-mediated exocytosis in the context of Drosophila Minute/+ wing morphogenesis."

Regarding point 8, given that reduction of Fz2 can suppress apoptosis in RpS3/+ wing discs, how does it regulate apoptosis when it is down-regulated anyway in apoptotic cells (Fig 4B', S4B')?

Response:

As described in our response to point 9, our previous study suggested that we observed that Wg signaling activity was elevated much more broadly in the RpS3/+ pouch compared to the localized activation observed in the wild-type control at the same developmental stage, as assessed by the nmo-lacZ reporter (Fig 3D in Akai et al., PLOS Genetics, 2021, as also shown Fig. R6 above). This elevation leads to the formation of an ectopic cell population with high Wg signaling activity in the wing pouch, potentially causing non-autonomous cell death among cells exhibiting lower Wg signaling activity, a process akin to cell competition. Notably, the significant reduction of apoptosis in RpS3/+ wing discs following Fz2 downregulation suggests that Fz2 may play a crucial role in the massive cell turnover, through a mechanism similar to cell competition, even though the exact mechanism remains unclear. We plan to incorporate such discussions into the discussion section of the revised manuscript.

Figure S1D. It seems that cadps is further up-regulated if JNK signaling is inhibited in RpS3/+ cells. Why did the authors select this gene?

Response:

We apologize for any confusion caused. We found that Cadps expression in the RpS3/+ wing pouch was increased compared to both the wild-type control (2.29-fold increase, RpS3/+ compared to wild-type) and the RpS3/+ wing pouch expressing Puc, driven by the nub-gal4 driver (3.33-fold increase, RpS3/+ compared to RpS3/+ +Puc). This suggests that the increase in Cadpsexpression in the RpS3/+ wing pouch is dependent on JNK signaling. We have now revised Fig S1D for clearer representation. We also have made modifications in the transferred manuscript as follows:

(page 5, line 107-113)

"Mining the list of genes differentially expressed in the RpS3/+ wing pouch cells dependent on JNK signaling (Fig S1C and S2 Table), we noticed that among the genes associated with the "secretion by cell" GO term (Fig S1B), the evolutionarily conserved exocytosis-related genes unc-13, SNAP25, and cadps (Calcium-dependent secretion activator) were upregulated in a JNK-dependent manner (2.29-fold increase, RpS3/+ compared to wild-type; 3.33-fold increase, RpS3/+ compared to RpS3/+ +_ Puckered)_ (Fig S1D)."

Figure S1H'. I don't see that Hrs is upregulated in this panel.

Response:

We thank the reviewer for the comment. Our quantitative analysis showed an increase in Hrs-positive puncta in the RpS3/+ wing pouch compared to the wild-type control (Fig S1I). We have substituted the relevant Figure with new images that more clearly demonstrate the elevation of Hrs-positive puncta (Fig S1G-H' in the transferred manuscript).

Minor points:

Describe in more detail, what are unc13, SNAP25 and cadps.

Response:

We thank the reviewer for the comment. We have now included an explanation for unc-13, SNAP25 and cadps in the transferredmanuscript as follows:

(page 5, line 117-126)

"It has been shown that unc-13, SNAP25, and cadps collectively regulate the docking process of secretory vesicles to the plasma membrane during Ca2+-mediated exocytosis ____((21, 23, 24), Reviewed in (22)). ____Specifically, UNC-13, a conserved presynaptic protein with calcium-binding domains, interacts with syntaxin to prime vesicles for fusion, crucial for calcium-regulated exocytosis (____23____). SNAP-25, in conjunction with syntaxin-1 and synaptobrevin, forms the pivotal SNARE complex for neuronal exocytosis, assembling into a four-helix bundle that is essential for drawing vesicle and plasma membranes close together to enable membrane fusion (____24, 25____). Like UNC-13, CAPS possesses conserved C-terminal domains that are instrumental in the assembly of SNARE complexes, thus priming vesicles for Ca2+-induced exocytosis (____26)____.____"

I wondered about the use of Hrs as exosome marker. To my knowledge, it is an endosomal marker. Same with Alix. Please clarify.

Response:

As pointed out by the reviewer, ESCRT-0 component Hrs is also localized to endosome. However, Hrs is also used as an exosome marker in the previous manuscript (McGough IJ et al., Nature, 2020), alongside Alix and Tsg101, which are widely recognized as exosome markers (Willms E et al., Sci Rep., 2016; Dear JW et al., Proteomics, 2013). Indeed, Hrs is a key factor for mediating ESCRT-dependent exosome secretion in mammals and flies (Tamai K et al., Biochem Biophys Res Commun., 2010; Gross JC et al., Nat Cell Biol., 2012; Colombo M et al., J Cell Sci., 2013; Vietri M et al., Nat Rev Mol Cell Biol., 2020). We have incorporated this explanation into the revised manuscript as follows:

(page 6, line 139-141)

"____Furthermore, we noted an increase in vesicles positive for the ESCRT protein Hrs, ____an additional exosome marker crucial for exosome secretion (____34, 35____), in the RpS3/+ wing pouch relative to the wild-type control (Fig S1G and S1H, quantified in Fig S1I)."

In the context of references 16 and 17, Huh et al. (2004), Current Biology needs also to be cited.

Response:

We thank the reviewer for the comment. We have referenced the paper by Huh et al. (2004) published in Current Biology in the transferred manuscript as follows:

(page 3, line 68-71)

"Apoptotic cells, for example, can secrete mitogens su____ch as Wingless ____(Wg; a Wnt homolog), dpp (a BMP homolog), and Hh, which could promote the proliferation of nearby cells in the Drosophila epithelium (15, _16_, 17, 18, 19)."

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #3

Evidence, reproducibility and clarity

Summary: In previous work, the authors showed that in the pouch of wing imaginal discs in Minute (M) mutants apoptosis and compensatory proliferation are dramatically increased in a Wingless (Wg) and JNK-dependent manner. Here, they show that JNK-induced exocytosis is mediating the secretion of Wg in apoptotic cells. Wg in turn stimulates the Fz2 receptor in neighboring surviving cells promoting their proliferation.

Major comments:

1. Figure 1A-H, S1E. The authors used EGFP-CD63 and Syt1-EGFP as markers of exocytosis. They see an increased number of puncta in apoptotic cells. Is this a specific effect in the dying cells in M mutant discs, or is it a general effect in apoptotic cell death? This should be examined in a condition where apoptosis is induced independently of M mutants such as nub-reaper or nub-hid.
2. Is exocytosis actually upstream or downstream of cell death, or both? The authors are kind of vague about it. On one hand, they say the dying cells induce exocytosis and secrete Wg. On the other hand, unc13RNAi can suppress cell death (Figure 1D', J'). Please clarify.
3. Figure 1L-P. The observation of Ca++ flashes is very interesting. However, are they important for exocytosis, cell death and compensatory proliferation? Right now, this is just a stand-alone observation. Can mutants affecting Ca++ signaling block exocytosis, cell death and comp prol?
4. Figure 3B'. Can you visualize aberrant wg expression in RpS3/+ wing discs only in the presence of p35? The expression of p35 in apoptotic cells generates undead cells which by itself induce Wg expression in a JNK-dependent manner (Perez-Garijo et al 2004; Ryoo et al 2004). Also, in Figure 3B', Wg is upregulated in the ventral half of the pouch, whereas in other figures (4B for example) apoptosis is strongly induced in the dorsal half. Does that make sense?
5. Can wg RNAi in cells destined to die (puc-Gal4 UAS-wgRNAi) suppress apoptosis and comp prol?
6. Figure 3D,E. Use cDcp1 as apoptotic marker instead of CD63-mCherry which is not an apoptotic marker.
7. Figure 4A' and B'. I am not sure there is much of a difference in the expression of fz2-lacZ in wild-type and RpS3/+ discs. The quantification in 4F shows there is no difference in the dorsal half of the pouch where massive apoptosis occurs. Can you generate a condition in which only one compartment (anterior or posterior) is mutant for RpS3/+, while the other compartment is wt? That would allow direct side-by-side comparison. Comparisons between different discs is problematic.
8. There is some inconsistency in the presence of apoptotic cells and presence of cells which secrete Wg. Apoptotic cells occur in large cell clusters (Fig. 2G, 3H', S3A', S4B), while markers of exocytosis and Wg are present in individual puncta (Fig 3D', E', S3C,E). That does not seem to fit with the authors' conclusion that dying cells secrete Wg by exocytosis. Are all dying cells secreting Wg through exocytosis? Please explain.
9. I was a bit confused by the conclusion of the authors about the results in Figure 4G-I. What do they mean by "difference in Fz2 expression"? Are they referring to the gradient that they described in their previous work? Or is it just the absolute level of Fz2 that determines apoptosis and proliferation? If the latter, is the gradient still present (at least in fz2/+), just at a lower level?
10. Regarding point 8, given that reduction of Fz2 can suppress apoptosis in RpS3/+ wing discs, how does it regulate apoptosis when it is down-regulated anyway in apoptotic cells (Fig 4B', S4B')?
11. Figure S1D. It seems that cadps is further up-regulated if JNK signaling is inhibited in RpS3/+ cells. Why did the authors select this gene?
12. Figure S1H'. I don't see that Hrs is upregulated in this panel.

Minor points:

1. Describe in more detail, what are unc13, SNAP25 and cadps.
2. I wondered about the use of Hrs as exosome marker. To my knowledge, it is an endosomal marker. Same with Alix. Please clarify.
3. In the context of references 16 and 17, Huh et al. (2004), Current Biology needs also to be cited.

Referees cross-commenting

I think all three reviewers agree in their assessment of this manuscript. Some of the comments by the reviewers address the same concerns. I liked the comment by reviewer 2 about the autonomy/non-autonomy of the signaling events in this model. I was thinking that, too, but didn't express it in my review. So, thanks reviewer 2, for bringing this up.

Significance

General assessment. A major strength of the work is that the experiments are done in a very good manner. The authors used multiple assays to come to the same conclusions. Limitations and weaknesses are mentioned in "major comments".

Advance. In my mind, one problem is novelty of this work. The authors showed before that Wg is secreted by dying cells in Minute mutants. It was also shown that Wg can be secreted by exocytosis in many studies from different authors. Here, the authors basically combine both observations and show that Wg is secreted by exosomes in Minute mutants.

Audience. This work likely addresses a specialized audience involved in Minute-induced cell competition. I don't think it will be of interest beyond this specific field.

Background of the reviewer. I have an interest in cell competition, apoptosis and neurodegeneration.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

The authors initiate their study by illustrating that minute cells undergo an augmentation in exocytosis-related proteins and calcium flashes (Fig 1). Subsequently, they establish a correlation between JNK-mediated exocytosis and the regulation of caspase activation in minute cells (Fig 2). Further, the authors unveil a dependency of Wg secretion on exocytosis, elucidated in Fig 3. Finally, they discern the involvement of one of the Wg receptors, Fz2, in the cell death/proliferation phenotype of minute cells. Despite the formulation of an intriguing hypothesis, the presented data falls short of robustly supporting their assertions.

Primary Concerns:

1. A significant challenge arises concerning the delineation of cell autonomy/non-autonomy. This study focuses on two distinct cell types, namely dying cells and proliferating cells. However, the consistent use of nub-gal4, a wing pouch driver, and the heterozygous minute mutant throughout the paper impedes the ability to conclusively analyze the autonomy of events. The authors previously posited that caspase-induced cell death triggers non-autonomous proliferation, but recent studies also suggested caspase-induced autonomous proliferation in both flies and mammals (Yosefzon et al. Mol. Cell 2018, Shinoda et al., PNAS 2019). Therefore, a meticulous distinction between autonomous and non-autonomous events, particularly through experimentation involving clones, is imperative. This necessity is particularly evident in Fig 4.
2. Closely tied to the issue of cell autonomy/non-autonomy is the question of how cells differentiate Wg from dying cells and the dorsal-ventral boundary.
3. In Fig 1, the authors interpret the upregulation of exocytosis-related genes as indicative of increased exocytosis. However, this interpretation lacks direct evidence and overlooks the possibility of opposing effects. For example, autophagosome accumulation means either activation or inhibition of autophagy. Or, in case of Dilp secretion, absence of vesicles indicates upregulation of secretion. To substantiate their claim, the authors must provide more conclusive evidence of increased exocytosis. Fig 2 suggests that inhibiting exocytosis-related genes suppresses caspase activation, favoring the proposition that exocytosis is upregulated. However, demonstrating a direct increase in exocytosis in minute cells would bolster their argument. Higher resolution imaging of the exosome marker, with overlayed images, would enhance clarity too.
4. Fig 3 introduces ambiguity regarding the relationship between cell death and exocytosis. The authors assert that dying cells exhibit elevated Wg protein levels compared to the wild-type control but omit an important comparison to the minute disc. Moreover, while Figs 1-2 propose a signaling cascade involving minute>JNK>exocytosis>cell death, Fig 3 implies that cell death regulates exocytosis. The coherence of their model and logic requires clarification - specifically, elucidating the molecular coupling mechanism between cell death and exocytosis.

Specific Points:

1. In Fig S1D, contrary to the authors' claim, cadp2 is not upregulated in a JNK-dependent manner.
2. When detecting multiple proteins in the same tissue, it is advisable for the authors to present overlayed images to enhance the clarity of their findings. Many pictures require higher magnification too.
3. In Fig 1A, the observed upregulation of EGFP-CD63 and Syt1-EGFP may potentially result from an artifactual effect of apoptosis. To validate their findings are specific, the authors should include negative controls that do not exhibit upregulation in dying cells.

Referees cross-commenting

I agree with comments by other reviewers.

Significance

The authors initiate their study by illustrating that minute cells undergo an augmentation in exocytosis-related proteins and calcium flashes (Fig 1). Subsequently, they establish a correlation between JNK-mediated exocytosis and the regulation of caspase activation in minute cells (Fig 2). Further, the authors unveil a dependency of Wg secretion on exocytosis, elucidated in Fig 3. Finally, they discern the involvement of one of the Wg receptors, Fz2, in the cell death/proliferation phenotype of minute cells. Despite the formulation of an intriguing hypothesis, the presented data falls short of robustly supporting their assertions.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

Summary:

In this study, Akai and colleagues study the cellular mechanism leading to high cell turnover in the Minute mutant background in Drosophila wing disc. Previously, the same group characterised an unexpected high rate of proliferation in the wing pouch of the Minute heterozygous larvae, which have been long known to have slower development time. This high rate of proliferation is driven by high apoptosis rate, JNK activation, the release of Dilp8 hormone, and the upregulation of Wg signaling (downstream of JNK) necessary for increased apoptosis and proliferation ( Akai et al., Plos Genetics 2021). Here, the authors address now the molecular link connecting JNK activation and the increased turnover. Using RNAseq and comparing WT, M+/- wing disc with M+/- wing disc upon JNK inhibition, they found a number of genes associated with exocytosis that were upregulated in Minute disc which was lost upon JNK inhibition. The authors first confirm the existence of higher number of exocytic vesicles in Minute wing disc as well as higher calcium activity that were all reduced upon JNK inhibition. Moreover, downregulation of several components of the exocytosis machinery abolished the high turnover rate of Minute wing disc (reducing both apoptosis and proliferation) while leading to the appearance of morphological defects in adult wing, while having no effect in a WT background. Interestingly, Wg accumulates in these exocytosis vesicles, and this accumulation relies on JNK and core exocytosis machinery. The role of Wg is confirmed by reducing the concentration (using heterozygous mutant or RNAi) of the Wg receptor Fz2, which is also specifically downregulated in the cluster of apoptotic cells. Altogether, this study suggests that upregulation of exocytosis by JNK activation is a central regulator of the higher cell turnover in Minute background.

The article is well written and the data overall convincing, including a lot of genetic backgrounds confirming the impact of exocytosis, as well as all the necessary quantifications. Some additional epistatic experiments may help to clearly test to which extend exocytosis is the major downstream target of JNK, and additional control may also help to clarify the sufficiency of JNK for generating such phenotype. Finally, I believe the epistatic link between exocytosis and Wg would deserve more experiments to be definitly proven.

Major comments:

1. It would be important to check how much JNK is sufficient to trigger the exocytosis upregulation. Is the accumulation of Cyt1 and CD63 vesicles in apoptotic cell common to any JNK dependent death or does it require the Minute background ? Could the authors check whether clones expressing HepCA transiently in WT background also accumulate the same vesicles ?
2. It would be relevant to have the status of JNK in Minute disc upon downregualtion of exocytosis. One could imagine some positive feedback between JNK activation, exocytosis, cell death and further JNK activation.
3. So far, the evidence for the epistatic link between exocytosis, Wg and cell turnover is mostly based on colocalization and the similarity of the phenotype but I believe this may need some additional evidences. Ideally one would need to be able to enhance exocytosis and test whether Wg downregulation suppress the phenotype, but I am not sure that upregulation of core exocytosis genes will be sufficient to do this. Alternatively, if Wg is indeed downstream of the upregulation of exocytosis, the reduction of cell turnover upon Wg flattening (e.g. : ftz2-/+ background) should not be enhanced by the reduction of exocytosis. Moreover, could the authors test the status of Wg downstreams targets upon inhibition of exocytosis in the Minute background (for instance, do they see a supression of the nmo-LacZ upregulation that they previously characterised in Minute wing disc in 2021) ?
4. Since Cyt1 and CD63 seem to mostly accumulate in apoptotic cells, it would be interesting to check their status in Minute wing disc upon apoptosis inhibition (e.g. : with H99 or mirRHG).
5. I would remain cautious about some of the statements, notably in the abstract, since some of them are mostly speculative and not really based on any experiments. For instance, the statement "This interaction between dying cells and their neighboring living cells is pivotal in determining cell fate, dictating which cells will undergo apoptosis and which cells will proliferate" is not backed up by any experiment (which would require to show that exocytosis and Wg from the dying cell specifically is required for the survival and proliferation of their neighbours, and/or showing that cell death occurs specifically in cells with local differences in Wg signaling). I would recommend to be more cautious here and us a clear conditional statement.

Other minor point:

The authors document Wg localisation in Minute wing disc upon expression of P35. It would be interesting to describe what is the status of Wg in Rps3+/- compared to WT without p35 (if I am correct, this was done in their previous article, and in that case it would be relevant to describe these former results in the main text).

Referees cross-commenting

I overall agree with all the other comments. It seems that the overall assessment and criticisms (cell autonomy, better characterisation of the status of exocytsosi) are along the same line

Significance

This article is mostly a follow up of a former study published by the same authors in Plos Genetics in 2021. The high turnover (high proliferation and high apoptosis rate) in the minute background was the most surprising observations that was performed previously by the authors, and the role of JNK and Wg was already quite extensively explored in this previous article. The main novelty here is to provide a systematic analysis of expression profile of Minute disc upon JNK inhibition and identify the strong contribution of exocytosis for increased apoptosis and proliferation in Minute wing disc downstream of JNK. In that sense, I believe these are interesting results, but novelty remains a bit limited (at least relative to their previous study). Still, these results could be interesting for the community studying cell competition, ribosomopathies, and growth regulation specially in Drosophila (so mostly for a specialised readership).

I have expertise in epithelial cell death, apoptosis, and cell competition, specially in Drosophila. I feel confident to evaluate every part of the article.

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

Manuscript number: RC-2024-02352

Corresponding author(s): Elise, Belaidi

1. General Statements

We would like to thank the reviewers for their constructive suggestions and comments. We hope that the “point by point answer” and the revision plan proposed below will convince the reviewers and the editor.

We thank the reviewer 1 for his/her comments. We would like to specify that the involvement of HIF-1 in IH-induced mitochondrial remodeling has indeed been initiated by RNA-seq analysis and confirmed in a cell-based model as well as in wild-type and HIF-1a+/- heterozygous mice subjected to intermittent hypoxia (IH). In vivo, we originally demonstrated that Metformin reversed IH-induced increase in myocardial infarct size through AMPKa2 and, we proposed that metformin could modify HIF-1 activity. Then, we validated our hypothesis in an in vitro model allowing to demonstrate that Metformin, by increasing HIF-1a phosphorylation decreases its activity. We acknowledge that we used several models and this is the reason why we detailed as much as possible the Materials and Methods section including all models, experimental sets designed and methods details. We hope that the point by point response that we made for the reviewer 1 will increase the clarity of our work and we hope that the new results provided will strengthen the evidences concerning the mechanisms by which metformin can inhibit and modulate the deleterious impact of HIF-1 on IH-induced an increase in myocardial infarct size.

We thank the reviewer 2 for his/her conclusion highlighting that “our work opens new avenues for exploring the potential effects of metformin as a modulatory of HIF-1𝛂 activity in obstructive sleep apnea syndrome”. We hope that the clarifications and/or justifications brought will convince him/her.<br /> We thank the reviewer 3 for having underlined that “metformin induces HIF-1α phosphorylation, decreases its nuclear localization and subsequently HIF-1 transcriptional activity are very much interesting” and for having highlighting that “our study is convincing”. We hope that the justification and the corrections brought in the point by point answer will convince him/her.

Alltogether, as underlined by the 3 reviewers, our study is very interesting for translational science in the fields of cardiovascular, respiratory and sleep medicine. We hope that the point by point answer and the revision plan proposed will allow the publication of our article in EMBO Molecular Medicine.

2. Description of the planned revisions

• *

Please find below the revision that we plan to address to answer to the questions of the reviewer 1.

Figure 1 - it would be helpful to list all of the DEGs (what genes are changed?). Including the expression of HIF-1α and PHD isoforms would be informative. If there is a robust HIF-1α signal, changes in the expression of HIF and PHD isoforms would be anticipated. Fig 1F - with regards to glycolysis and hypoxia pathway analysis, most of the DEGs are not canonical HIF-1α/hypoxia targets.

Figure 1 aimed at better understanding and manipulate the well-recognized involvement of HIF-1 in response to our specific IH stimulus (Semenza, Physiology 2009; Belaidi, Pharmacol & Ther. 2016). __The results provided by the RNA-seq analysis shows that IH induces cardiac oxidative and metabolic stress which are inter-related with HIF-1 activation. __We did not claim that these genes are HIF-1 targets genes. The RNA seq analysis did not allow to reveal HIF-1a and PHD1-3 transcript as the most dysregulated genes of the panel. In case of publication, bulk data and DEGS will be provided in an online file. We agree with the reviewer that the list of the 40 up and down-regulated genes would be very informative and would increase the value of the paper. Thus, we plan to add the name of the 40 up and down-regulated genes on Figure 1B.

Figure 5G-I, show cytoplasmic HIF1a as well as nuclear.

Alternatively, why not use IHC for subcellular localization?

We think that the comments of the reviewer 1 concern Fig.4G-I and not 5G-I. In this figure, we showed that IH increases nuclear HIF-1____a____ expression compared to N condition and that this IH-effect is abolished in mice treated with Metformin, suggesting that, upon IH, Metformin impacts HIF-1__a __nuclear content and subsequently, its activity. The nuclear localization of HIF-1a is the most relevant mean to indicate its activation. We agree with the reviewer that IHC also allows for the indication of the nuclear localization of HIF-1a. Indeed, we previously performed IHC on nuclear HIF-1a localization and demonstrated that IH increased HIF-1a nuclear localization by IHC that was corroborated by Western-blot (Moulin S, TACD, 2020). Western-blot and IHC are both semi-quantitative techniques with different process of analyses. In this study, we choose Western-blot because we have the material to perform this technique and because IHC is associated with an analysis process (size of a slice, areas to analyze, colorimetry…) that is more complex than the analysis process of Western-blot (densitometry solely).

While the nuclear localization of HIF-1a is the most relevant mean to indicate its activation; it could be interesting to see that HIF-1a cytosolic content was neither modify by IH nor by Metformin. This would also corroborate the results of the RNA-seq that did not demonstrate any difference in DEGs of HIF-1a or of other members of the HIF family. This would also confirm that Metformin plays a major role on HIF-1 activaty regulation (and not transcription) in the context of IH.

Thus, we plan to perform a Western-blot of HIF-1a on cytosolic extracts of hearts from mice exposed to N or IH and treated or not with Metformin. These extracts are already available and Western-blot would be performed and replicated in 3 weeks. We could also provide a Western-blot in order to show the purity of our extraction protocol (nucleus vs cytosol).

Figure 5F, it would be important to show the levels of expression of HIF1a in these experiments. Are there positive and negative controls that the authors could use for HIF21a activity in this experiment?

In our manuscript, we aimed at demonstrating that Metformin decreases HIF-1 activity in a context of strong HIF-1____a____expression and/or stabilizion those mimics what happens after chronic IH in mice (Belaidi E, Int J Cardiol 2016, Moulin S, Ther Adv Chronic Dis 2020) and in apneic patients (Moulin S, Can J Cardiol 2020). Thus, we used a transfection allowing to overexpress HIF-1a that is one of the best means to increase HIF-1 activity. In the Figure 1 below, HA-HIF-1α-WT Addgene AmpR and 5 HRE GFP AmpR plasmids co-transfection induced a decrease in H9c2 viability and an increase in GFP-positive cells that were not observed in H9c2 transfected with pcDNA 3.1 HA-C AmpR (negative control). __This validates our in vitro model as a good positive control to mimic IH consequences. __ However, we agree with the reviewer that we could add a supplemental figure or a panel demonstrating that our transfection induced an increase in HIF-1a expression. Thus, we will perform a Western-blot targeting HIF-1a on H9c2 transfected with the control plasmid (pcDNA 3.1 HA-C AmpR) or the plasmid allowing the overexpression of HIF-1a (HA-HIF-1α-WT AmpR). This work would be performed in 2 months.

Moreover, we already improved the lisibility of the Figure 5F to clarify the experimental conditions (table inserted under the graphic); we also completed the Materials and Methods section to specify the plasmid used (modifications are in red in the manuscript).

Figure to see on the downloaded file.

Figure 1 : GFP fluorescence in H9c2 cells transfected with pcDNA 3.1 HA-C AmpR (control condition) or HA-HIF-1α-WT AmpR (positive control, overexpression of HIF-1a) and 5 HRE-GFP AmpR plamids and treated with CoCl2 (1mM, 2h); magnification x100.

• *

This paragraph concerns only the point 4 of the fifth question.

In these experiments, as well as subsequent studies, it would be very informative to use a specific AMPK activator e.g. MK-8772, to compare with metformin. It is well known that metformin has a number of other targets in addition to AMPK.

We agree with the reviewer that metformin has pleiotropic effect. Very interestingly, we demonstrated that the reduced-infarct size is not related to the metabolic systemic effect of metformin since it failed to improve the IH-induced insulin resistance while it improves the answer to insulin in normoxic mice (supplemental Figure S3B). This demonstrates that in our model, the cardioprotective effects of Metformin are independent of a potential systemic effect. Then, we demonstrated that metformin protects the heart against ischemia-reperfusion through AMPK____a____2 activation by using AMPK____a____2KO exposed to IH in which Metformin failed to decrease infarct size (Fig.4N). MK-8772 is not widely used in vivo models. Moreover a recent study indicates that chronic treatment with MK-8772 (14 days 1 month in mice and rats, respectively) induces cardiac hypertrophy characterized by an increase in heart weight (Myers R, Science 2017). In vivo experiments with MK-8772 would be not clinically relevant as the use of metformin that is already used in clinic. However, in order to improve the mechanistic investigation concerning the role of AMPKa2 activation on inhibiting HIF-1 activity, we propose to perform the in vitroexperiments performed in Figure 5 with a specific allosteric small-molecule activator of AMPKa2 such as 991.

We plan to:

-Expose H9c2 to CoCl2 and treat them with 991 in order to measure HIF-1a phosphorylation.

-Transfect H9c2 with our plasmids HA-HIF-1α-WT AmpR and 5 HRE GFP AmpR and treat them or not with 991 in order to measure HIF-1 activity (GFP fluorescence). These experiments would be performed in 2 months.

__Please find below the revision that we plan to address to answer to the second question of the reviewer 2. __

• *

2) The WB images were cut and pasted. Please add the original images

We acknowledge the reviewer's comment and will address it by submitting a supplementary file containing the uncropped immunoblot images. Since this file already exists, our plan is to standardize it by providing, for each slide (immunoblot), all relevant information pertaining to our experiments, including groups, molecular weight markers, cutting, membrane stripping, and other pertinent details.

3. Description of the revisions that have already been incorporated in the transferred manuscript

• *

__Please find below the answers or the revisions that have already been incorporated in response to the comments of the reviewer 1. Please note that we provided new results and new figures at the discretion of the reviewer, but we are ready to insert them as figures or supplemental figures in a new revised manuscript if the reviewers and the editor think that it would improve our message. __

Was mitochondrial content in the hearts after IH experiment measured e.g. mtDNA measurements? IH results in mitochondrial dysfunction/reduced mitochondrial content. It would have been good to show mitochondrial dysfunction by doing basic functional experiments (e.g. TMRM/MitoROS imaging etc.) by isolating cardiomyocytes from the N and IH experiments.

We thank the reviewer for these questions about the mitochondrial function and content. The impact of IH on mitochondrial function has already been demonstrated in heart (Moulin S, Antioxydants 2022, Wei Q, Am J Physiol 2012). __Indeed, we previously showed that mitochondria isolated from hearts of mice exposed to IH had a decrease in maximal respiration in complex I and II that was not observed in HIF-1_a_+/- ____mice (Moulin S, Antioxidants 2022), indicating that HIF-1 is responsible for IH-induced mitochondrial dysfunction. __

Figure 4 shows that Metformin abolished IH-induced mitochondrial remodeling similar to what we observed in HIF-1a+/-(Figure 2). This means that treating with Metformin or partially deleting the gene encoding for HIF-a induce the same impact on IH. Then, we demonstrated that Metformin can control HIF-1 activation and we concluded that metformin could be cardioprotective through inhibiting HIF-1 activation and subsequent mitochondrial stress and remodeling. In this study, we focused on the effects of Metformin on HIF-1 and we did not aim at directly test the effect of metformin on mitochondrial function. Actually, metformin exhibits biphasic effects on bioenergetics of cardiac tissue depending on the modality of administration (i. e. single injection, time of administration during an ischemia-reperfusion procedure); the dose administered and the tissue studied (i. e. hiPSC-CMs, isolated mitochondria…) (Emelyanova N, Transl Res 2021). But, we collected some data that we would like to submit at the discretion of the reviewer. Using oximetry, we measured maximal respiration in complex 1 and 2 on isolated mitochondria from hearts of mice exposed to N, IH and treated or not with metformin during the exposure__. While we observed that IH decreases maximal respiration in complex 1 and 2, we did not find any effect of metformin on mitochondrial respiration alteration induced by IH (Figure 2A, B). Using spectrofluorometry, we measured the mitochondrial membrane potential using TMRM; __we did not find any modification of membrane potential in IH or Metformin-treated mice (Figure 2C). Because we previously did not observe any impact of IH on mtDNA/gDNA ratio (Figure 2D), we did not test metformin on this parameter.

To conclude, we think that these results are not directly in the scope of our work but if the reviewer thinks that they deserve to be discussed, we could add them in a supplementary figure.

Please, see the figure on the dowloaded file

Figure 2 : Mice were exposed to 21 days of Normoxia (N) or Intermittent Hypoxia (IH) (1-min cycle of FiO2 5%-21%) and treated with vehicle (Vh, CmCNa 0.01%, 0,1ml.10g-1) or Metformin (Met, 300mg.kg-1.d-1). (A, B) Mitochondrial function __was measured by oximetry with sequential addition of substrate (state 2), ADP (200mM, state 3, maximal respiration) and oligomycin (12.5 mM, state 4) ; quantification of O2 consumption for NADH-linked mitochondrial respiration (complex I-glutamate-malate, GM, 20mM) (A), and for FADH2-linked mitochondrial respiration (complex II-succinate, S, 5mM, in presence of complex I inhibition by rotenone, 6.25mM) (B) (n=8). (C) Mitochondrial membrane potential measured by spectrofluorometry after Tetramethylrhodamine Methyl Ester (TMRM, 0.2mM) in presence of GM (basal condition), maximal respiration (ADP) and uncoupling condition Carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone, FCCP, 3mM); fluorescence intensity is expressed relative to fuorescence at baseline (before GM) (n=8). (D__) Mitochondrial content assessed by the expression of mitochondrial DNA (mtDNA, COX1) relative to genomic DNA (gDNA, ApoB) measured after PCR (n=6); *p

Fig 2A-D - CoCl2 is not a good model to mimic hypoxia due its effect on disrupting iron homeostasis in cells, which can mean that some of the effects are due to changes in iron levels and not HIF stabilisation.

The capacity of CoCl2 to chelate iron is the main property of CoCl2 that we used in order to stabilize HIF-1____a____. Actually, prolyl-4-hydroxylases need Fe2+ to hydroxylate HIF-1____a____ and induce its degradation. __Then, intermittent hypoxia (IH) is characterized by very rapid changes in PO2. This stimulus was designed to reproduce sleep apnea syndrome and its associated disorders (i. e. insulin-resistance, hypertension, increase in myocardial infarct size). This model was firstly developed and validated in rodents (Dematteis M, ILARJ 2008, Belaidi E, Eur Resp Rev 2022, Harki , Eur Resp J 2022). Compelling evidence indicate that the involvement of HIF-1 in IH-deleterious consequences is related to the repetitive phases of oxygenation and especially to IH-induced oxidative stress (Semenza Physiology 2009, Belaidi Pharmacol. & Ther 2016). In order to increase the level of mechanistic insights on HIF-1, we next attempted to optimize in vitro models. A device was developed by Minoves et al. (Minoves M, Am J Physiol, 2017) to expose endothelial and cancer cells to IH. __However, as illustrated below, this device does not mimic efficient rapid hypoxia-reoxygenation cycles able to induce cardiac cell death (Figure 3A). However, CoCl2 decreases H9C2 viability by 60% (Figure 3B) that is associated with a sustained stabilization of HIF-1____a (Figure 3C,D). Thus, we choose this in vitro model as it replicates cardiac cell death and HIF-1_a_overexpression or stabilization which we similarly observe in our in vivo model and in apneic patients (Moulin S, Can J Cardiol 2020).

Please see the figure on the dowloaded file

Figure 3 : (A-B) Cell viability measured by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H- tetrazolium bromide (MTT) of H9c2 cells exposed to 6 hours (h) of repetitive cycles of Intermittent hypoxia (2 minutes (min) PO2 16% - 2 min PO2 2%) (n=3) (A) or treated with CoCl2 (1mM, 2h) (n=6) (B). (C-D), __quantification of __total ____HIF-1____𝛂 expression relative to tubulin (C) and representative image of Wetsern-blot (D) (n=2-3); *p****p*

Fig. 2K, change in BNIP3 expression is modest, but change in Parkin is very dramatic. BNIP3 is a HIF-1α target but Parkin is not, so it is plausible that mitophagy could be occurring through a HIF-1α independent mechanism.

Fig. 2K is a representative panel of 2-3 independent experiments. The quantification reported in Figures 2I and 2J demonstrated a significant decrease in BNIP3 and Parkin expressions in HIF-1a heterozygous mice exposed to Intermittent Hypoxia (IH) compared to HIF-1a+/+ mice exposed to IH. While we acknowledge that only BNIP3 is a direct target of HIF-1, the role of HIF-1 in IH-induced auto/mitophagy is demonstrated by our experiments performed in HIF-1____a____heterozygous mice. This shows an important role for HIF-1 without excluding any impact of HIF-1a independent mechanisms.

What are we meant to be looking at in Fig 2L?

Figure 2L aims at illustrating mitochondrial remodeling under IH. Stars indicate that mitochondria have abnormal fate in IH conditions and arrows point autophagosomal membrane and formation. This figure was magnified to be clearer (please see new Figure 2L).

Figure 3B-C, the reduction in pT172 on AMPK is modest. It would be good to include pACC as a downstream target for AMPK.

As recommended by the reviewer, we inserted in the manuscript the quantification of 79Ser-P-ACC/ACC western-blot as well as a representative image of the Western-blot (See new Figure 3). We also modified the legend of the figure. 79Ser-P-ACC is an important target of AMPK; however, in our experimental conditions, its phosphorylation is not associated to the decrease in AMPK phosphorylation. This could be explained by many points. First, Metformin was administered every day and hearts were harvested 24 hours later after the last administration. Most studies demonstrating a modification of AMPK and ACC phosphorylation are experiments performed in vitro or directly (less than 1 hour) after a single dose of Metformin administration. In the context of myocardial ischemia-reperfusion, Yin et al. showed an increase in P-AMPK/AMPK directly after Metformin treatment without showing any data on P-ACC/ACC (Yin M; Am J Physiol 2011); similar data were published in models of chronic cardiac diseases (Soraya H, Eur J Pharmacol, Gundewar S, Circ Res 2009). Second, in line with the previous explanation, the lack of effect of metformin on P-AMPK and/or P-ACC in rodent models could be explained by its rapid distribution (Sheleme T Clin Pharmacokinetics of Metformin 2021) and its short half-life that is around 3.7 hours in mice (Junien N, Arch Int Pharmacodyn Ther 1979).

To conclude, since we performed all our analysis 24h after the last treatment and exposure to hypoxia, we argue that the slight but significative decrease in AMPK phosphorylation that we observed in our study highlight a robust impact of chronic IH. However, this would be elegant to confirm this result by measuring AMPK through its phosphorylation capacity (Cool B, Cell Metab 2006, Ducommun S, Am J Physiol 2014). We already sent hearts from mice exposed to Normoxia or Intermittent Hypoxia to Luc Bertrand’s lab (IREC, Belgium) where they used to perform this assay.

Fig. E-G, show data for mice treated with vehicle.

In Figure 4 I-J­­, we demonstrated that Metformin significantly decreases infract size in IH condition only and this validates our main hypothesis regarding the specific beneficial effect of this drug in the context of chronic IH. __In order to show that the cardioprotective effect of Metformin is relative to AMPKa2 activation, we first showed that 79Ser-PACC/ACC, one of the main downstream targets of AMPKa2 was increased (Fig. E-G). We did not find it necessary to does not exhibit cardioprotective effects. However, as shown in Figure 3 below, __Metformin also increases 79Ser-PACC/ACC in Normoxic mice validating the treatment. Thus, in normoxic conditions, AMPK____a____2 activation does not exert any cardioprotective effect. We acknowledge that this reinforces our result about the specificity of AMPK____a____2 activation by Metformin under chronic IH condition. We could add this Figure in supplemental results.

Please, see the figure on the dowloaded file

Figure 4 : AMPK activation in Normoxic mice treated with vehicle (Vh, CmCNa 0.01%, 0,1ml.10g-1) or __ __metformin (Met, 300mg.kg-1.d-1 : __ __172Thr-P-AMPK/AMPK (A) and 79Ser-P-ACC/ACC (B) ratio and representative image of Western-blot (C) (n=3-6); *p

Fig. 3K - what cre is used for the a2 KO mice?

As written in the Materials and methods section, AMPKa2KO mice are not inducible Knock-out mice. Constitutive AMPKα2 knockout mice were kindly generated by Benoit Viollet (Viollet B, JCI, 2003).

Include normoxia data for the a2 KO mice studies.

The question of the reviewer concerning the cardioprotective effects of metformin is interesting but is not aligned with the objectives of the study. Indeed, we did not treat normoxic mice with Met for several reasons. First, the objective of the study was to find a cardioprotective strategy against IH-induced an increase in infarct size. Second, Fig. 3I shows that Met significantly reduced infarct size upon IH only; this suggests that AMPK____a____2 activation is specifically involved in IH-induced increase in infarct size but not in reducing infarct size in normoxic mice. Moreover, the beneficial impact of metformin in standard models of myocardial ischemia-reperfusion is controversial and has been extensively discussed (Foretz et al. Cell Metab. 2014). Overall, using AMPKa2 mice was legitimated in the context of IH only. We validated the involvement of AMPKa2 in the cardioprotective effect of metformin especially in IH conditions.

Figure 5G, what is the rationale for switching to CoCl2 in the mice to prove metformin reduced HIF-1α expression? Why not use reduced O2 tension in mice.

We respectfully disagree with the reviewer since mice were exposed to N and IH and treated or not with Metformin to demonstrate that this drug abolished IH-induced increase in HIF-1a nuclear expression (Figure 4 H, I). The same model was used in Figure 3 to demonstrate the impact of Metformin on infarct size. Fig. 5G was conducted to demonstrate the potential link between AMPKa2 and HIF-1a phosphorylation in basal conditions of AMPKa2 content or in absence of AMPKa2 (AMPKa2-/- mice). The single presence of AMPK____a____2 demonstrates an increase in HIF-1____a____ phosphorylation if its stabilization is increased by CoCl2; this was not observed in AMPK____a____2-/- mice highlighting that AMPK____a____2 plays an important role in HIF-1____a phosphorylation.

Please find below the answer to the first question asked by the reviewer 2.

1) Why did authors choose the IH protocol illustrated in Fig. S1A

The choice of the hypoxic stimulus was based on literature and mainly on our recognized expertise in preclinical studies aiming at better understanding obstructive sleep apnea syndrome (OSA); a chronic pathology associated with several comorbidities such as diabetes, hypertension… We are conscious that the hypoxic stimulus used in this study is very severe, with a nadir arterial oxygen saturation (SaO2) around 60%. However, this experimental design is required to induce detrimental cardiovascular effects __in the absence of any confounding factors (i.e., obesity) or genetic susceptibility for complications (i. e. genetic susceptibility to hypertension) (Dematteis M, ILARJ 2008). Especially in the context of myocardial infarction, exposing rodents to 14 to 21 days of IH at 5% and subjected them to a myocardial ischemia-reperfusion protocol allows us to reproduce the increase in infarct size in rats (Belaidi E, J. Am. Coll. Cardiol., 2009; Bourdier G, Am J Physiol, 2016) and in mice (Belaidi E, Int J Cardiol 2016; Moulin S, Can J Cardiol 2020) similar to what has been observed in apneic patients (Buchner S, EHJ 2014). __Moreover, we recently conducted a meta-analysis based on 23 preclinical studies aiming at investigating the impact of the IH pattern (duration, FiO2, repetition of cycles…) on infarct size and cardiomyocyte death (Belaidi E, Eur. Resp. J 2022). We showed that IH significantly increases infarct size when IH is applied several days (especially 14 to 21 days) and when FiO2 is around 5%; whereas IH decreases infarct size when it is applied a single day at a FiO2 at 10%. This meta-analysis provided the confirmation that we need to apply a chronic and severe stimulus to reproduce an increase in infarct size that is observed in apneic patients which are exposed every day, during several days to a decrease in SaO2. If the reviewers and the editor consider that this point should be discussed in the discussion section, we will be happy to include it.

• *

Please find below the answers or the revisions that have already been incorporated in response to the comments of the reviewer 3. They appear in red in the new manuscript except the modifications performed in the “references” section which appear in black.

1 The authors used H9c2 rat cardiac cells in vitro experiments although they used mouse model in vivo experiments. Using mouse P19.CL6 cardiac cells instead of rat H9c2 cells may much clearer. Why the authors did not use P19.CL6 cells should be explained.

We thank the reviewer for his/her suggestion. P19CL6 cell line has been isolated from pluripotent P19 embryonal carcinoma (EC) cells after long term culture under conditions for mesodermal differentiation (Habara-Ohkubo A, Cell Struc Funct 1996). Therefore, these cells are mostly used to ____study the differentiation of cardiac muscle. Indeed, they were recognized to avoid large variations in the differentiation rates which were extensively reported (Mueller I, J Biomed Biotechnol 2010). To our knowledges no ventricular non-beating mice cell line. In this study, we used H9c2 which are extensively used and recognized as a gold standard cellular model to study the biology of cardiomyocytes including mechanisms involved in cardiac ischemia-reperfusion injury (Paillard M, Circulation, 2013; Zhang G Circulation 2021__), cardiac hypertrophy__ (Zhang N, Cell Death Diff 2020; Hu H, Cardiovasc Res 2020), intra-organites calcium exchanges (Moulin S, Antioxydants, 2022, Paillard M, Circulation, 2013) as drug testing (Beshay NM, J Pharm and Tox Methods, 2007). Recently, H9c2 and P19.CL6 were exposed to intermittent hypoxia (70 cycles of FiO2 1% (5 min) - FiO2 21% (10%)) in order to “mimic OSA” and investigate the transcription level of a pool of genes. The authors show some similiraties and differences of mRNA expression between the two cell lines that, indeed, could be attributed to variations in the cell origin (Takasawa S, IJMS 2022). However, in this study, there are no experiment allowing to assess the state of cardiac cells (apoptosis, life, metabolism, remodeling) questioning the pathophysiologic transposability of the model. Moreover, the number of experiments conducted on H9C2 (pubmed references : 7000 vs 100 for P19.CL6) to understand the mechanism involved in acute and chronic cardiac pathologies makes our choice confident and relevant.

2 The authors described, "The protocol was approved by the French minister (APAFIS#23725-2020012111137561.v2)." in Animals (page 16) without showing approval date, the authors should clearly show the approval date together with their approval numbers.

We added the approvement date in the materials and methods section. It was approved on February 20, 2020.

3 In Figure 2 L, scale bar(s) should be added because figures are magnified and/or reduced by printer.

We agree with the reviewer that scale bars were not visible; we highlighted them.

4 In Figure 3J and 3N, scale bar(s) should be added.

Scale bars have been now added on figures 3J and 3N. All pictures were acquired at the maximal zoom of a camera placed at an equal distance from the slices. Then, analyses were performed, slice per slice, with Image J with the same zoom (x5 to get an image at 100%). In this context, scale was added based on a photo of slice taken close to a ruler.

5 In introduction, "HIF-1" should be changed to "hypoxia-inducible factor 1 (HIF-1)".

Thank you, we replaced HIF-1 by Hypoxia Inducible Factor-1 (HIF-1) in the introduction section.

6 In Results, "Angpt1, Txnip, Nmrk2, Nuak1 or Pfkfb1" should be changed to "Angiopoietin 1 (Angpt1), Thioredoxin-interacting protein (Txnip), Nicotinamide riboside kinase 2 (Nmrk2), NUAK family SNF1-like kinase 1 (Nuak1) or 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 1 (Pfkb1)".

We did the modification and let the abbreviation in italic since it concerns genes name.

7 In Chronic intermittent hypoxia, "FiO2" should be changed to "FiO2".

Thank you, we changed it in the materials and methods section.

8 In Western-blot, "Bio-Rad, California, USA" should be changed to "Bio-Rad, Hercules, CA".

Thank you, we did it.

9 In Western-blot, what "tubulin" (α-tubulin or β-tubulin) should be clarified.

We agree with the reviewer that this point should be specified; α-tubulin was stained, we added the “α”.

As mentioned below, we have done all the modifications required by the reviewer in the references section.

10 In Ref. 8, "Antioxidants (Basel) 11 (2022)" should be changed to "Antioxidants (Basel) 11, 2326 (2022)".

Done

11 In Ref.10, "Pharmacol Ther (2016)" should be changed to "Pharmacol Ther 168, 1-11 (2016)".

Done

12 In Ref.13, "Antioxidants (Basel) 11 (2022)" should be changed to "Antioxidants (Basel) 11, 1462 (2022).

Done

13 In Ref. 21, "Diabetes (2017)" should be changed to "Diabetes 66, 2942-2951 (2017)".

Done

14 In Ref. 28, "Adv Biol (Weinh), e2300292 (2023)" should be changed to "Adv Biol (Weinh), 8, 2300292 (2023)".

Done

15 In Ref. 37, "J Am Heart Assoc 6 (2017)" should be changed to "J Am Heart Assoc 6, e006680 (2017)".

Done

16 In Ref. 41, "Eur Respir Rev 32 (2023)" should be changed to "Eur Respir Rev 32, 230083 (2023)".

Done

17 In Ref. 46, "Int J Mol Sci 22 (2020)" should be changed to "Int J Mol Sci 22, 268 (2021)".

Done

18 In Ref. 47, "Int J Mol Sci 21 (2020)" should be changed to "Int J Mol Sci 21, 2428 (2020)". Done

4. Description of analyses that authors prefer not to carry out

__Please find below the answers to the comment of the reviewer 3 that we cannot provide and that is not in the scope of the study. __

• *

Figure 5, multiple phosphorylation sites have been identified on HIF1a. What is the nature of the Thr/SerP-HIF1a antibody? It would be far more preferable (essential?) to identify the site(s) within HIF1a that are phosphorylated by AMPK.

The antibody was provided by Cell Signalling, ref. 9631. Phospho-(Ser/Thr) Phe Antibody detects phospho-serine or threonine in the context of tyrosine, tryptophan or phenylalanine.

The identification of the Phosphorylation sites will require a long-time consuming phosphoproteomic analysis and subsequent functional validation in vivo and in vitro (directed mutagenesis, knock-in mice, …) which are out of the scope of our paper.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #3

Evidence, reproducibility and clarity

Although the findings that metformin inhibits intermittent hypoxia (IH)-induced mitophagy in myocardium and decreases hypoxia-inducible factor 1-α (HIF-1α) nuclear expression in mice subjected to IH. In vitro demonstrated that metformin induces HIF-1α phosphorylation, decreases its nuclear localization and subsequently HIF-1 transcriptional activity are very much interesting, numbers of points need clarifying and certain statements require further justification. These are given below.

Point

1. The authors used H9c2 rat cardiac cells in vitro experiments although they used mouse model in vivo experiments. Using mouse P19.CL6 cardiac cells instead of rat H9c2 cells may much clearer. Why the authors did not use P19.CL6 cells should be explained.
2. The authors described, "The protocol was approved by the French minister (APAFIS#23725-2020012111137561.v2)." in Animals (page 16) without showing approval date, the authors should clearly show the approval date together with their approval numbers.
3. In Figure 2 L, scale bar(s) should be added because figures are magnified and/or reduced by printer.
4. In Figure 3J and 3N, scale bar(s) should be added.
5. In introduction, "HIF-1" should be changed to "hypoxia-inducible factor 1 (HIF-1)".
6. In Results, "Angpt1, Txnip, Nmrk2, Nuak1 or Pfkfb1" should be changed to "Angiopoietin 1 (Angpt1), Thioredoxin-interacting protein (Txnip), Nicotinamide riboside kinase 2 (Nmrk2), NUAK family SNF1-like kinase 1 (Nuak1) or 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 1 (Pfkb1)".
7. In Chronic intermittent hypoxia, "FiO2" should be changed to "FiO₂".
8. In Western-blot, "Bio-Rad, California, USA" should be changed to "Bio-Rad, Hercules, CA".
9. In Western-blot, what "tubulin" (α-tubulin or β-tubulin) should be clarified.
10. In Ref. 8, "Antioxidants (Basel) 11 (2022)" should be changed to "Antioxidants (Basel) 11, 2326 (2022)".
11. In Ref.10, "Pharmacol Ther (2016)" should be changed to "Pharmacol Ther 168, 1-11 (2016)".
12. In Ref.13, "Antioxidants (Basel) 11 (2022)" should be changed to "Antioxidants (Basel) 11, 1462 (2022).
13. In Ref. 21, "Diabetes (2017)" should be changed to "Diabetes 66, 2942-2951 (2017)".
14. In Ref. 28, "Adv Biol (Weinh), e2300292 (2023)" should be changed to "Adv Biol (Weinh), 8, 2300292 (2023)".
15. In Ref. 37, "J Am Heart Assoc 6 (2017)" should be changed to "J Am Heart Assoc 6, e006680 (2017)".
16. In Ref. 41, "Eur Respir Rev 32 (2023)" should be changed to "Eur Respir Rev 32, 230083 (2023)".
17. In Ref. 46, "Int J Mol Sci 22 (2020)" should be changed to "Int J Mol Sci 22, 268 (2021)".
18. In Ref. 47, "Int J Mol Sci 21 (2020)" should be changed to "Int J Mol Sci 21, 2428 (2020)".

Significance

the study is convincing

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

In the submitted paper by Moulin et al., the authors investigated the effect of metformin treatment (an oral anti-diabetic drug and AMPK activator) on cardiac response to ischemia-reperfusion in mice exposed to intermittent hypoxia (IH), a major component of obstructive sleep apnea (OSA). The data demonstrates that IH alters the expression profile of a subset of genes involved in myocardial mitochondrial dysfunction, such as Angpt1, Txnip, Nmrk2, Nuak1 or Pfkfb1, hallmark genes of hypoxia and glycolysis. In H9c2 cells treated with CoCl2 with or without chloroquine, the authors claim that IH induces mitophagy through HIF-1 activation, associated with an increase in autophagic flux. This effect is abolished in HIF-1𝛂+/- mice. Metformin treatment in mice reverses IH-increased in infarct size through AMPK activation, decreases IH-induced mitophagy and HIF-1𝛂 nuclear localization.

The paper is interesting but needs some clarifications:

1. Why did authors choose the IH protocol illustrated in Fig. S1A)
2. The WB images were cut and pasted. Please add the original images

Significance

This work opens new avenues for exploring the potential effects of metformin as a modulatory of HIF-1𝛂 activity in obstructive sleep apnea syndrome.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

Summary: The manuscript by Moulin and colleagues begins by exploring changes in gene expression in mice subjected to intermittent hypoxia (IH). Based on these results, the authors postulate a link with HIF1a, and so use a cell-based model to examine this link. Next, the authors turn to AMPK showing that IH decreases Thr172 phosphorylation on AMPK. In view of this, the authors use metformin as a way to activate AMPK during IH and monitor necrosis in hearts following ischemia-reperfusion. Metformin treatment reduces necrosis in mice subjected to IH. Mechanistically, the authors link decreased HIF1a activity following AMPK activation in response to metformin and suggest that this is due to phosphorylation of HIF1a by AMPK. Overall, this reviewer found some of these connections to be quite weak and in some cases the evidence supporting the claims was equivocal. In some instances, the lack of appropriate controls or the use on non-specific reagents, undermined the findings. Adding to these issues, there was a lack of experimental detail both in the Figure legends and in the M&M section. Below I have listed some of the major concerns that if addressed satisfactorily would greatly strengthen the manuscript.

Major Comments

Major Points

1. Figure 1 - it would be helpful to list all of the DEGs (what genes are changed?). Including the expression of HIF-1α and PHD isoforms would be informative. If there is a robust HIF-1α signal, changes in the expression of HIF and PHD isoforms would be anticipated. Fig 1F - with regards to glycolysis and hypoxia pathway analysis, most of the DEGs are not canonical HIF-1α/hypoxia targets.
2. Was mitochondrial content in the hearts after IH experiment measured e.g. mtDNA measurements? IH results in mitochondrial dysfunction/reduced mitochondrial content. It would have been good to show mitochondrial dysfunction by doing basic functional experiments (e.g. TMRM/MitoROS imaging etc.) by isolating cardiomyocytes from the N and IH experiments.
3. Fig 2A-D - CoCl2 is not a good model to mimic hypoxia due its effect on disrupting iron homeostasis in cells, which can mean that some of the effects are due to changes in iron levels and not HIF stabilisation. Fig. 2K, change in BNIP3 expression is modest, but change in Parkin is very dramatic. BNIP3 is a HIF-1α target but Parkin is not, so it is plausible that mitophagy could be occurring through a HIF-1α independent mechanism. What are we meant to be looking at in Fig 2L?
4. Figure 3B-C, the reduction in pT172 on AMPK is modest. It would be good to include pACC as a downstream target for AMPK. Fig. E-G, show data for mice treated with vehicle. Fig. 3K - what cre is used for the a2 KO mice? Include normoxia data for the a2 KO mice studies. In these experiments, as well as subsequent studies, it would be very informative to use a specific AMPK activator e.g. MK-8772, to compare with metformin. It is well known that metformin has a number of other targets in addition to AMPK.
5. Figure 5G-I, show cytoplasmic HIF1a as well as nuclear. Alternatively, why not use IHC for subcellular localisation?
6. Figure 5, multiple phosphorylation sites have been identified on HIF1a. What is the nature of the Thr/SerP-HIF1a antibody? It would be far more preferable (essential?) to identify the site(s) within HIF1a that are phosphorylated by AMPK.
7. Figure 5F, it would be important to show the levels of expression of HIF1a in these experiments. Are there positive and negative controls that the authors could use for HIF21a activity in this experiment?
8. Figure 5G, what is the rationale for switching to CoCl2 in the mice to prove metformin reduced HIF-1α expression? Why not use reduced O2 tension in mice.

Minor Comments

Nothing at this stage of the process. The authors need to focus on the major points to improve the quality of the manuscript.

Significance

Overall, the lack of robust mechanistic studies makes it difficult to fully interpret the impact of the study.

This feels like a preliminary exploration of a potentially important biological system, but the manuscript seems incomplete and requires more attention to details.

Upon revision, I think the study would be of interest to basic and clinical researchers working in the filed of cardiac metabolism and hypoxia.

My expertise is in metabolism and cell signalling.

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

The authors do not wish to provide a response at this time

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #3

Evidence, reproducibility and clarity

Here, Ahel et al characterize complex composiiton, chromatin binding, and respressive function of ADNP2. First they show that ADNP2 forms a complex with CHD4 and HP1b, very similar to, although biochemically distinct from, the ChAHP complex formed by ADNP. This complex is prevalently bound at repeats, and in particular at LTR retrotranspoons. Recruitment of ADNP2 is mostly mediated by HP1b binding to H3K9 and a PxVxL mutant that cannot bind to HP1 does not localize properly to chromatin. While ADNP has its own specific targets in SINE elements, it also binds some ADNP2 targets in an HP1-dependent manner. Depletion of ADNP or ADNP2 results in upregulation of some shared and some distinct transposons, indicating distinct roles but also partial redundancy.

This is a solid study that adds to our knowledge of these repressive complexes and their transposable element targets. I feel that the data could be analyzed a bit more in depth, especially in the comparisons of ChIP-seq vs. RNA-seq.

Major points

• Fig. 5: it would be interesting to analyze differences between repeats that are only under ADNP2 control vs. those that are sensitive to additional loss of ADNP. Presumably they both have K9me3 so why some are also repressed by ADNP?
• Fig. S5B, S7A: degrons are known to destabilize some proteins even before induction of degradation. These western blots should include a WT line to compare abundance of the endogenous protein with the tagged version.
• FIg. 5: it would be good to show a bit more overlap between RNA-seq and ChIP-seq. How many of the bound transposons are derepressed? Can co-occupancy by ADNP and ADNP2 explain which transposons will display synergistic effets from removal of both proteins?

Minor points

• Fig. 2B: Upset plot seems like a strange choice for visualization. I would show as stacked bar plot compared to genome distribution. I.e. are peaks enriched on repeats or is it just that a larger genomic space is occupied by repeats compared to TSSs?
• Fig. 3A: if ADNP2 only binds to a subset of K9me+ retrotransposons, why arent' there regions of K9me+ ADNP2- in this plot?
• Check figure calls for 3B and 3C, some of them seem off.
• Fig. S7B: the text says several dozen genes but I only see 21 up and 15 down upon ADNP with ADNP2 present.

Significance

This is a solid study that adds to our knowledge of these repressive complexes and their transposable element targets. I feel that the data could be analyzed a bit more in depth, especially in the comparisons of ChIP-seq vs. RNA-seq.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

In this work, Ahel, Pandey and Schwaiger et al. examine the function of ChAHP complexes in transposon silencing in mouse embryonic stem cells. This works builds on a proteomics approach previously developed in the Buhler lab, which identified ADNP1-ChAHP as a repressor complex of SINE retrotransposons. In their previous work they also identified the zinc-finger protein ADNP2 as an interactor of the ChAHP component CHD4. Here, the authors test the hypothesis of whether ADNP2 is part of an alternative ChAHP, which they termed ChAHP2. They first establish, through proteomics/biochemistry experiments, that the ChAHP2 complex can form independently of ChAHP. By ChIP-sequencing they then show that the ChAHP2 member ADNP2 predominantly occupies retrotransposons of the LTR families (e.g., ERV elements). They further show that recognition of target transposable elements (TEs) is conferred by the ChAHP2 complex member HP1 and its affinity for H3K9me3. Using RNA-seq they then addressed the repressive activities of ChAHP and ChAHP2 on TEs through degradation of ADNP, ADNP2 or both. The experimental approach is very well thought out and experiments are done with extensive controls.

I have the following comments/suggestions:

• Regarding the ChIP-seq experiments, it is not entirely clear from the manuscript text and also not from the Material and Methods whether ChIP-seq was done with a tagged-version of ADNP2, which is what I assume, or an endogenous ADNP2 antibody. Could you please state this clearly?
• The authors map short read sequencing data to individual TE insertions and conclude that ADNP2 binds retrotransposons both internally and at their terminal sequences. Particularly for the internal parts, how sure can the authors be that their approach, mapping to repeat consensus sequences, is not confounded by e.g., genetic differences between their ESC lines and the references they map to, or simply by the limitations of short read data?
• Page 7, the first sentence starting with "Contrary to the expected behavior of a TF, ...". Could the authors please elaborate on what they mean with "the expected behavior of a TF"?
• In Figure S4B, could the authors please clarify if HP1, CHD4 and Tubulin are high/low exposure?
• SETDB1 is the H3K9me3 methyltransferase responsible for LTR/ERV silencing in mESCs and its disruption leads to a pronounced upregulation of TEs (Karimi et al., 2011). Could the authors comment on the effect of their degron-mediated SETDB1 depletion on de-repression of LTR/ERV elements? At least for the elements shown to lose H3K9me3 upon SETDB1 depletion (Figure S5D - strong depletion replicate)? Could you please also provide information about morphological/pluripotency characteristics of the 2HA-FKBPSETDB1 ESC lines before and after the 48h treatment with 500nM dTAG13?
• Maybe I missed this, but are H3K9me3 levels affected in the ADNP and/or ADNP2 degron lines where transcriptional dysregulation of TEs is observed (Figure 5)? If not H3K9me3, is TE de-repression accompanied by reduction in repressive histone marks in ADNP2 degraded lines?
• When looking at TE de-repression, the upregulation seems very modest (Figure 5) and Log2FoldChanges >0.9 (FDR<0.05) were considered as statistically significant. Is this the result of using mESCs that "approximate a constitutive KO situation through inducible ADNP2 degradation" and 14 days of treatment with 250nM dTAG13? As mention above, could you please also provide information about morphological/pluripotency characteristics?

Significance

The findings are interesting and expand on our understanding of how complexes of chromatin bound proteins control epigenetic silencing of individual retrotransposon classes/families in mouse embryonic stem cells. More specifically, this study adds a new player, ChAHP2, a ChAHP alternative that associates with ERV and LINE1 retrotransposons via HP1-mediated binding of H3K9me3 and is required for transcriptional repression of LTR class retrotransposons. Although, in my opinion, it remains somewhat uncertain how depletion of ADNP2 results in TE upregulation.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

The Buhler group previously identified the ChAHP complex and showed it to be made up of the transcription factor ADNP, chromatin remodeler CHD4, and HP1 proteins. They found that genetic removal of ADNP leads to increased expression of SINE B2 elements in mouse ES cells (mESCs), and an increase in chromatin accessibility and CTCF binding.

In the current manuscript they describe a related protein complex they name ChAHP2. ChAHP2 has a vertebrate paralogue of ADNP, a transcription factor they call ADNP2 and the same chromatin remodeler CHD4, and HP1 proteins. By investigating chromatin occupancy they find that ChAHP2 chromatin binding specificity is distinct from ChAHP, and predominantly associates with ERV and LINE1 retrotransposons via HP1beta-mediated binding of H3K9 trimethylated histones. The ChAHP and ChAHP2 complexes control a wide variety of molecularly disparate retrotransposons, including SINEs, LINEs, and ERVs.

The findings are of significance as the ChAHP2 complex is novel - it has not previously been identified or characterized and its description is therefore worthy of publication. Overall this study was well-performed and nicely-presented. In general, the text is clear, the flow of experiments seemed logical, easy to follow and supported by the presented evidence without the findings being overstated.

The data is mainly descriptive and it was at times hard to identify what were the clear conclusions from their study. Having identified the new ChAHP2 complex they show that its chromatin binding characteristics are distinct from ChAHP, which predominantly binds SINE elements. In contrast, ChAHP2 binds different classes of retrotransposons and recruitment to H3K9me3 heterochromatin is dependent on HP1-Beta. This was all relatively clear. It became a little more confusing trying to unravel the functional consequences of the loss of ChAHP2 - either alone, or in combination with ChAHP.

Although they see changes - both up- and down-regulation of a large number of genes, they say that: 'None of the changing gene categories showed strong and significant patterns either in terms of GO term enrichment (FigureS7C), distance between the promoter and the nearest ADNP/ADNP2 peaks'- does this mean that none of the upregulated genes following ADNP1/2 deletion show ADNP/ADNP2 peaks over the specific gene in question - i.e. do they think that the majority of effects here are indirect?

When it came to the 'repeat' families of genes - the situation was also not entirely clear.

They say: All these differentially expressed repeats belonged to the LTR class of retrotransposons, including families identified as ADNP2-bound in ChIP sequencing - it would be helpful to know which familes are bound and maybe show snapshots of ADNP2 binding. Are these differentially expressed repeats directly regulated by ADNP2? - they seem like a fairly heterogenous group of repeat elements? - do they have any shared features? They say that expression of two LINE1 subclasses were significantly upregulated - do these L1s (L1Mdas) have common features - are they old or young - it would be helpful if these different issues could be clarified.

In the discussion they are upfront about their inability to identify any specific sequence motifs required for ADNP2 occupancy. Since the repression is not obviously H3K9Me3 dependent, then what is the role of H3K9me3 in repressing here? They surmise that repression is likely to be regulated through chromatin remodeling by CHD4 - would ATAC-seq help clarify this suggestion? - particularly as they suggest that having CHD4 as an integral component of the repressor complex is a unique feature of ChAHP1/2?

In this context, do they not consider MORC2 to be an essential component of HUSH-dependent silencing?

Other points

The language in the introduction somewhat confusing and could be improved: They jumble ZNFs/TRIM28/HUSH/HP1 as if there is no specificity here - there clearly is - needs a more balanced and informative description.

What is firstly referring to? Firstly, histone deacetylation and the removal of activating histone methylations disfavor transcription13,14. Secondly, the underlying DNA is extensively methylated, further repressing the locus

They state: Like ERVs, LINE1 retrotransposons are autonomous- would help if they could clarify exactly what is meant here.

Significance

The findings are of significance as the ChAHP2 complex is novel - it has not previously been identified or characterized and its description is therefore worthy of publication. Overall this study was well-performed and nicely-presented. In general, the text is clear, the flow of experiments seemed logical, easy to follow and supported by the presented evidence without the findings being overstated.

The findings will be of broad interest to chromatin biologists - particularly those interested in the regulation of retroelements and repeat regions

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

We would like to thank the reviewers for carefully reviewing our manuscript submitted to Review Commons. Their constructive comments have led us to identify key additional experiments to perform (testing endothelial cells, investigating the role of CXCR4/Sdf1 axis in the asymmetric division). They have shed light on specific points of the manuscript that had to be amended (see below).

First, we are pleased that the reviewers acknowledged the quality of the experiments performed (Reviewer #1: “In general, is a well-done work” ; Reviewer #2: “the evidence is convincing, experiments are rigorously performed with adequate replicates and reproducibility”), as well as interest and large scope our study (Reviewer #1: * “it gives a few more details describing that when HSPCs divide asymmetrically it seems there is an association between centrosome and lysosome distribution”, Reviewer #2: “These results collectively contribute to a deeper understanding of how the hematopoietic niche and interactions with neighboring cells influence HSPC behavior and their commitment to distinct cell fates during division”, “The findings have significance in understanding how the microenvironment influences HSPC behavior. This is broadly significant for stem cell biology”. Reviewer #3 “The manuscript is likely to have broader interest to not only HSPC researchers, but stem cell biologists and even engineers due to the technologies used”).*

Our detailed answers to the reviewers’ concerns are listed below.

---

Reviewer #1 (Evidence, reproducibility and clarity):

In this manuscript, Candelas and colleagues investigated asymmetric cell division of human hematopoietic stem and progenitor cells (HSPCs) upon heterotypic interactions with osteoblasts. They developed a new in vitro system to test this and found that upon interaction, HSPCs polarized in interphase with centrosome and the Golgi apparatus and lysosomes positioned close to the site of contact. In addition, during mitosis, HSPCs were found to orient their spindle perpendicular to the plane of contact. This division gave rise to siblings with unequal amounts of lysosomes and CD34. In general, is a well-done work and it gives a few more details describing that when HSPCs divide asymmetrically it seems there is an association between centrosome and lysosome distribution.

Authors: We thank the reviewer for this positive assessment of our work.

Reviewer #1 (Significance (Required)):

However, all of these features (described above) have already been independently described even in human HSPCs thus this work does not represent a major advancement in this field (e.g. on the potential molecular mechanism by which heterotypic interactions with osteoblasts promote such behaviours). Overall, this stage is descriptive in nature.

Authors: We respectfully disagree with the reviewer on this point. The main finding of our work does not simply concern the ability of HSPCs to undergo asymmetric division. This ability has indeed been previously shown in mouse and human models, and we refer to the princeps works in our introduction (page 2 “In vitro HSPCs can undergo asymmetric divisions 11 12 13 14 15 16 17 ”). Our main finding is about the instrumental role of heterotypic interactions with stromal cells of the niches. This had not been shown previously, mostly due to the limitations of classical co-culture systems and of in vivo tracking of HSPCs. We have engineered artificial niches in microwells specifically to overcome these limitations. We are confident that this finding is of interest for of hematopoietic niches biology and more globally for the field of stem cell biology. In that sense, the corresponding biorXiv preprint has just been quoted by the seminal review by Hans Clevers “Hallmarks of stemness in mammalian tissues” (doi: 10.1016/j.stem.2023.12.006).

One major caveat of this work is that CD34+ HSPCs represent a very heterogeneous population. Although the underlying features described in this work could be similar between different cell populations, the frequency of these events (e.g. % magnupodium; polarization index, % of asymmetric inheritance, etc) is likely to be different between stem and progenitor cells. This may also explain the wide spread of their data points. Experiments should also be conducted with different cell populations, in particular with HSCs (either with the CD34+CD38-CD45RA-CD90+ or CD34+CD38-CD45RA-CD90+CD49f+ HSC enriched fraction or the highly CD34+CD38-CD45RA-EPCR+ HSCs).

Authors: We agree that we are dealing with an heterogenous population of stem and progenitor cells; we cannot exclude that more pronounced effects could be obtained by analyzing separately stem cells, lymphoid and myeloid progenitors. The use of CD34+ heterogenous population has been a strategical choice. We are working with human primary cells, harvested from cord blood, a rare and precious material, who’s access has been in addition reduced during the COVID period and since. Working with the global CD34+ population was giving us the ability to work with a higher number of cells and avoid material limitations to perform experiments. Importantly, in our initial publication (Bessey et al doi:10.1083/jcb.202005085), we both used CD34+, or CD34+/CD38- versus CD38+/CD34+ subpopulations (see figure 4): the effects we described using global CD34+ population were significant, validating a posteriori this choice. Similarly, the effects observed in the present work with CD34+ population have always been validated by appropriate statistic tests. So, we are confident that the use of the CD34+ population, despite its heterogeneity, did not alleviate the validity and pertinence of our analyzes and conclusions.

Results shown in Figure 2 were obtained with a very limited number of cells; also data obtained for Figure 4; this should be substantiated;

Authors: We acknowledge that the analyses performed on fixed mitotic cells are based on relatively small samples. We have decided to fix and analyze the cells at 40h of culture which corresponds to the peak of mitosis, instead of synchronizing HSPCs in mitosis with available drugs, in order to avoid potential perturbations of these drugs. The consequence was the scarcity of the mitotic HSPCs in each experiment. This limitation has also been encountered by other laboratories working on HSPCs : our samples sizes are in fact within the range of the samples analyzed in other works (see Hinge et al., 2020, doi:10.1016/j.stem.2020.01.016: around 20 cells; Florian, et al., 2018: doi: 10.1371/journal.pbio.2003389. from 8 to 40 cells). However, in order to improve this limitation, few additional experiments and analyses have been performed to increase the number of mitotic cells (Figure 3: from 27 to 38 cells).

In addition, data from Figure 2 adds little and should be combined with Figures 3 and 4 to make a single stronger message.

Authors : We agree on this comment: Figures 2 and 3 have now been merged

It is/was unclear why the authors investigated the distribution of CD34 and CD33 and the major and important question remained to be answered: whether the symmetric cell divisions were/are differentiating or self-renewal in nature hence, the asymmetric cell divisions promoted by the osteoblasts may represent asymmetric self-renewal ones; this needs to be investigated further and potential molecular mechanisms for such promotion should be highlighted.

Authors: CD34 and CD33 were chosen as our “gold standards” according to the princeps work of Loeffler et al (Loeffler et al., 2019, and 2022). Our selection is based on the need for markers that encompass this diverse population. We focused on CD34 and CD33 as they are among the most prevalent CD markers on HSPCs, particularly when working with the CD34+ population, which is broad and necessitates widely expressed CD markers. Importantly, previous works have reported that CD34 anti-correlates with lysosome asymmetric inheritance after the first division, whereas CD33 does not (see Fig3E, Loeffler et al., 2022). So using these markers we can compare two markers: one that follows lysosome inheritance after the first division, and one which doesn’t. We found that osteoblasts inducing asymmetric lysosome inheritance in HSPC do increase asymmetric production of CD34 but not CD33, in concordance with the literature. This has been now explained in more detail in the text.

In addition, higher lysosome inheritance has been previously demonstrated to be associated with the most stem cell (Loeffler et al., 2019), implying a more stem-like profile for the proximal cell during division in our work. However, recent studies using single-cell sequencing approaches challenge the notion of discrete differentiation within this stem cell model (reviewed by Zhang et al., 2018; Laurenti and Gottgens, 2018; Cheng et al., 2020; Liggett and Sankaran, 2020). Our study contributes to this discussion by providing insights into the relationship between CD markers, lysosome inheritance, and the stem cell profile during asymmetric division, and presenting this heterogeneity observed in single-cell sequencing works.

Reviewer #2 (Evidence, reproducibility and clarity (Required)):

The paper titled "Heterotypic Interaction Promotes Asymmetric Division of Human Hematopoietic Stem and Progenitor Cells" presents research on how interactions between Hematopoietic Stem and Progenitor Cells (HSPCs) and osteoblasts impact cell division symmetry. The use of polyacrylamide microwells to simulate the hematopoietic niche and study osteoblast interactions is a novel and valuable approach. The paper effectively explains the experimental design, making it easy for others to replicate and verify the results. The paper provides detailed insights into how interactions with osteoblasts influence HSPCs, particularly in terms of cell polarization, spindle orientation, and organelle distribution during cell division. The evidence is convincing, experiments are rigorously performed with adequate replicates and reproducibility.

Authors: We thank the reviewer for this positive evaluation of our work.

• The study acknowledges that not all HSPCs respond to heterotypic interaction, suggesting that individual variability in HSPC behavior plays a role. Future research could explore the factors that determine which HSPCs undergo asymmetric divisions. This raises questions about the heterogeneity in HSPC responses. The researcher starts culturing CD34+ which is a heterogeneous group with osteoblast. This heterogeneity by default will give different cells progeny which mimics the in vivo BM status. But, to minimize the invitro experiment variables examining a more pure HSC population such as isolated CD34+CD38-CD90+ cells would strengthen the findings?

Authors: We agree that we are dealing with an heterogenous population of stem and progenitor cells; we cannot exclude that more pronounced effects could be obtained by analyzing separately stem cells, lymphoid and myeloid progenitors. The use of CD34+ heterogenous population has been a strategical choice. We are working with human primary cells, harvested from cord blood, a rare and precious material, who’s access has been in addition reduced during the COVID period and since. Working with the global CD34+ population gave us the ability to work with a higher number of cells and avoid material limitations to perform experiments. Importantly, in our initial publication by Bessy et al (doi:10.1083/jcb.202005085), we used both CD34+, or in some cases CD34+/CD38- versus CD38+/CD34+ subpopulations (see figure 4), to investigate HSPC mode of polarization. The effects observed in the case of total CD34+ were significant, validating a posteriori this choice. Similarly, the effects observed in the present work with CD34+ population have always been validated by appropriate statistic tests. Our study indicates that the mechanism we described is conserved throughout the entire HSPC population.

In conclusion, we are confident that the use of the CD34+ population, despite its heterogeneity, did not alleviate the validity and pertinence of our analyzes and conclusions.

What was the Osteoblast/HSCs ratio used in the experimental setup? Does this ratio affect the results?

Authors: The loading of HSPC has been optimized to get one HPSC per microwell, as indicated in the result section (page 3 line 3). This has now been clarified in the material and methods section. However, we observe some wells with more than one HSPC on the stromal cell, but we did not notice any clear impact on cell division asymmetry and did not investigate further this parameter.

• The paper primarily focuses on lysosome inheritance and CD34 expression, leaving room to explore other lineage-specific markers and their correlation with asymmetric division. Do osteoblast/ HSCPs interactions have also an effect on mitochondrial inheritance?

Authors: We fully agree that other markers could have been analyzed to comfort our observations.

A far as mitochondria are concerned, recent publications suggest that lysosomes can be considered as a “inversed” readout of mitochondria inheritance: (1) mitophagy has been shown to take place in quiescent hematopoietic stem cells (Ito et al., 2016), suggesting that lysosomes do participate in the regulation mitochondria levels in HSPCs. (2) Loeffler et al. (2019) has shown an opposite pattern between mitochondria and lysosomes, lysosomes co-inheriting with mitophagosomes. Considering that this mutual or opposite inheritances were already described, we did not investigate them further. However we agree that it could be the focus for future works.

• While the study hints at the clinical implications of its findings, it does not show the practical applications in detail. How the knowledge gained from these interactions could be translated into clinical practices or therapies is not explicitly discussed. Bridging this gap is necessary for understanding the practical utility of these results.

Authors: we are confident that microwell technology could find, in the long term, more medical applications. However, we did not wish to speculate on this point in the present manuscript. The words “clinical” and “therapy” show no occurrence in our text. The potential implications of our findings in the field of leukemia are now mentioned in the discussion section with the corresponding references (page 6 line 10) “Such spatial control of the division is a conserved feature of asymmetric divisions in other stem cell niches 6. It may account for in vivo observations of organization into clusters of blood cell differentiation 44, and may play an important role in competition mechanisms at play within the hematopoietic niches in physiological 45 and pathological 46 contexts.”

__Reviewer #2 (Significance (Required)): __

• Investigating the process of HSPC asymmetric division is highly important to understand HSPC fate decisions and functions. The findings have significance in understanding how the microenvironment influences HSPC behavior. This is broadly significant for stem cell biology.
• The paper's observation of uneven lysosome distribution and its potential impact on differentiation markers, including CD34 and CD33, contribute to our understanding of HSPC division and lineage determination.
• Upon interaction with osteoblasts, HSPCs polarize during interphase, with centrosomes, the Golgi apparatus, and lysosomes positioned close to the site of contact. This reorganization of intracellular architecture plays a significant role in asymmetric division.
• These results collectively contribute to a deeper understanding of how the hematopoietic niche and interactions with neighboring cells influence HSPC behavior and their commitment to distinct cell fates during division.

Authors: We thank the reviewer for this quite positive evaluation of our work.

Reviewer #3 (Evidence, reproducibility and clarity):

Summary: In this study, the authors systematically work to identify the contribution of hematopoietic niche stromal cells in the control of HSPC asymmetric divisions. Using live cell imaging techniques, the authors show that heterotypic interaction with osteoblasts promotes asymmetric division of human HSPCs. The authors find that HSPCs polarize in interphase with centrosome, the Golgi apparatus and lysosomes positioned close to the site of contact. Moreover, during mitosis, HSPCs orient their spindle perpendicular to the plane of osteoblast contact and the subsequent division gives rise to siblings with unequal amounts of lysosomes and differentiation markers such as CD34. Taken together, the authors suggest that this asymmetric inheritance generates heterogeneity in progeny, which is likely to contribute to plasticity during the early steps of hematopoiesis.

Major Comments:

1. The study that uses polyacrylamide microwells as minimalist niches to evaluate the role of heterotypic interactions in the asymmetric division of HSPCs. Using this model system, the data clearly indicate that osteoblasts promote HSPC polarization when compared to fibronectin or skin fibroblasts. However, osteoblasts weren't compared to other HSPC stromal niche components. For example, is the same polarization observed with endothelial cells? This would help to determine whether this observation is specific to only osteoblasts in the niche or if there's a broader role for heterotypic interactions, which is the claim made in the discussion.

Authors: We have now performed additional experiments using endothelial cells (see Figure 1 and 2, Page 3 and 4) and found that endothelial cells do also promote asymmetric division.

This is primarily a descriptive study that doesn't attempt to address or even postulate the type of receptors potentially responsible for the polarization-driving heterotypic interactions. Perhaps inhibitors to candidate receptors could be incorporated or at least candidate receptors could be mentioned and described in the discussion as future directions.

Authors: We have now performed additional experiments to address this point: We have analyzed the effect of the CXCR4 antagonist AMD3100 on HSPC division upon interaction with osteoblast and found that blocking these receptors reduce the asymmetry of daughter cells to the level of isolated HSPC (see Figure 3 E -G; Figure S3 H-J; page 4).

__Minor Comments: __

1. There is no reference to Figures 2C and D in the text. Authors : This oversight has been corrected.

Page 3: “HSPC cultured on fibronectin exhibited spindle oriented parallel to the well bottom, (Figure 2A-C and S3C and D)”.

Page 4: In contrast, both equal and unequal LysoBriteTM segregations could be observed (Figure 2D and S3E).

The different circle colors in Figure 4C and D are a bit difficult to distinguish.

Authors: The figure has been modified accordingly

It's difficult to identify where the sites of cell contact are in Figure 2. Perhaps contact sites could be indicated in the images.

Authors: The sites of contact have now been systematically indicated using dashed lines, in Figure 2 and 3.

It's not clear how spindle orientation is being identified in Figure 3.

Authors: We are aware of the fact that spindle orientation is not easy to visualize in the still images presented in Figure 3A and B. These images are extracted from the movies that were used to determine the cell division plane visually. This plane, following the long-axis rule, is perpendicular to the spindle orientation, which could therefore be determined. Movies 6 and 7 are shown to support these observations and illustrate how amenable these movies are to support spindle orientation analysis.

Some grammatical errors need to be addressed throughout the manuscript.

Authors: We thank the reviewer for drawing our attention on this point: the modified text has been carefully examined, and hopefully improved.

Reviewer #3 (Significance):

Significance: Overall, this is straightforward study that presents data illustrating the polarization driving capacity of osteoblast interactions with HPSC. The findings are important to the field, as little is known about asymmetric vs symmetric division in HSPCs. However, in its current form, the manuscript lacks clarity with respect to whether specifically osteoblasts promote HSPC polarization, or any stromal niche player has this capacity. Moreover, there is a lack of mechanistic data and/or discussion about how osteoblasts may be driving the observed polarization.

Authors: We have now added data showing the similar role of endothelial cells and the implication of CXCR4 receptors.

The manuscript is likely to have broader interest to not only HSPC researchers, but stem cell biologists and even engineers due to the technologies used.

Authors: We thank the reviewer for his/her enthusiasm for our work

---

---

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #3

Evidence, reproducibility and clarity

Title: Heterotypic interaction promotes asymmetric division of human hematopoietic stem and progenitor cells

Authors: Adrian Candelas, Benoit Vianay, Matthieu Gelin, Lionel Faivre, Jerome Larghero, Laurent Blanchoin, Manuel Théry, Stephane Brunet

Summary: In this study, the authors systematically work to identify the contribution of hematopoietic niche stromal cells in the control of HSPC asymmetric divisions. Using live cell imaging techniques, the authors show that heterotypic interaction with osteoblasts promotes asymmetric division of human HSPCs. The authors find that HSPCs polarize in interphase with centrosome, the Golgi apparatus and lysosomes positioned close to the site of contact. Moreover, during mitosis, HSPCs orient their spindle perpendicular to the plane of osteoblast contact and the subsequent division gives rise to siblings with unequal amounts of lysosomes and differentiation markers such as CD34. Taken together, the authors suggest that this asymmetric inheritance generates heterogeneity in progeny, which is likely to contribute to plasticity during the early steps of hematopoiesis.

Major Comments:

1. The study that uses polyacrylamide microwells as minimalist niches to evaluate the role of heterotypic interactions in the asymmetric division of HSPCs. Using this model system, the data clearly indicate that osteoblasts promote HSPC polarization when compared to fibronectin or skin fibroblasts. However, osteoblasts weren't compared to other HSPC stromal niche components. For example, is the same polarization observed with endothelial cells? This would help to determine whether this observation is specific to only osteoblasts in the niche or if there's a broader role for heterotypic interactions, which is the claim made in the discussion.
2. This is primarily a descriptive study that doesn't attempt to address or even postulate the type of receptors potentially responsible for the polarization-driving heterotypic interactions. Perhaps inhibitors to candidate receptors could be incorporated or at least candidate receptors could be mentioned and described in the discussion as future directions.

Minor Comments:

1. There is no reference to Figures 2C and D in the text.
2. The different circle colors in Figure 4C and D are a bit difficult to distinguish.
3. It's difficult to identify where the sites of cell contact are in Figure 2. Perhaps contact sites could be indicated in the images.
4. It's not clear how spindle orientation is being identified in Figure 3.
5. Some grammatical errors need to be addressed throughout the manuscript.

Significance

Overall, this is straightforward study that presents data illustrating the polarization driving capacity of osteoblast interactions with HPSC. The findings are important to the field, as little is known about asymmetric vs symmetric division in HSPCs. However, in its current form, the manuscript lacks clarity with respect to whether specifically osteoblasts promote HSPC polarization, or any stromal niche player has this capacity. Moreover, there is a lack of mechanistic data and/or discussion about how osteoblasts may be driving the observed polarization.

The manuscript is likely to have broader interest to not only HSPC researchers, but stem cell biologists and even engineers due to the technologies used.

I have experience with questions of HSPC regulation.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

The paper titled "Heterotypic Interaction Promotes Asymmetric Division of Human Hematopoietic Stem and Progenitor Cells" presents research on how interactions between Hematopoietic Stem and Progenitor Cells (HSPCs) and osteoblasts impact cell division symmetry. The use of polyacrylamide microwells to simulate the hematopoietic niche and study osteoblast interactions is a novel and valuable approach. The paper effectively explains the experimental design, making it easy for others to replicate and verify the results. The paper provides detailed insights into how interactions with osteoblasts influence HSPCs, particularly in terms of cell polarization, spindle orientation, and organelle distribution during cell division. The evidence is convincing, experiments are rigorously performed with adequate replicates and reproducibility.

• The study acknowledges that not all HSPCs respond to heterotypic interaction, suggesting that individual variability in HSPC behavior plays a role. Future research could explore the factors that determine which HSPCs undergo asymmetric divisions. This raises questions about the heterogeneity in HSPC responses. The researcher starts culturing CD34+ which is a heterogeneous group with osteoblast. This heterogeneity by default will give different cells progeny which mimics the in vivo BM status. But, to minimize the invitro experiment variables examining a more pure HSC population such as isolated CD34+CD38-CD90+ cells would strengthen the findings? What was the Osteoblast/HSCs ratio used in the experimental setup ? Does this ratio affect the results?
• The paper primarily focuses on lysosome inheritance and CD34 expression, leaving room to explore other lineage-specific markers and their correlation with asymmetric division. Do osteoblast/ HSCPs interactions have also an effect on mitochondrial inheritance ?
• While the study hints at the clinical implications of its findings, it does not show the practical applications in detail. How the knowledge gained from these interactions could be translated into clinical practices or therapies is not explicitly discusssed. Bridging this gap is necessary for understanding the practical utility of these results.

Significance

• Investigating the process of HSPC asymmetric division is highly important to understand HSPC fate decisions and functions. The findings have significance in understanding how the microenvironment influences HSPC behavior. This is broadly significant for stem cell biology.
• The paper's observation of uneven lysosome distribution and its potential impact on differentiation markers, including CD34 and CD33, contribute to our understanding of HSPC division and lineage determination.
• Upon interaction with osteoblasts, HSPCs polarize during interphase, with centrosomes, the Golgi apparatus, and lysosomes positioned close to the site of contact. This reorganization of intracellular architecture plays a significant role in asymmetric division.
• These results collectively contribute to a deeper understanding of how the hematopoietic niche and interactions with neighboring cells influence HSPC behavior and their commitment to distinct cell fates during division.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

In this manuscript, Candeleas and colleagues investigated asymmetric cell division of human hematopoietic stem and progenitor cells (HSPCs) upon heterotypic interactions with osteoblasts. They developed a new in vitro system to test this and found that upon interaction, HSPCs polarized in interphase with centrosome and the Golgi apparatus and lysosomes positioned close to the site of contact. In addition during mitosis, HSPCs were found to orient their spindle perpendicular to the plane of contact. This division gave rise to siblings with unequal amounts of lysosomes and CD34. In general, is a well-done work and it gives a few more details describing that when HSPCs divide asymmetrically it seems there is an association between centrosome and lysosome distribution.

Significance

However, all of these features (described above) have already been independently described even in human HSPCs thus this work does not represent a major advancement in this field (e.g. on the potential molecular mechanism by which heterotypic interactions with osteoblasts promote such behaviours). Overall, this stage is descriptive in nature.

One major caveat of this work is that CD34+ HSPCs represent a very heterogeneous population. Although the underlying features described in this work could be similar between different cell populations, the frequency of these events (e.g. % magnupodium; polarization index, % of asymmetric inheritance, etc) is likely to be different between stem and progenitor cells. This may also explain the wide spread of their data points. Experiments should also be conducted with different cell populations, in particular with HSCs (either with the CD34+CD38-CD45RA-CD90+ or CD34+CD38-CD45RA-CD90+CD49f+ HSC enriched fraction or the highly CD34+CD38-CD45RA-EPCR+ HSCs).

Results shown in Figure 2 were obtained with a very limited number of cells; also data obtained for Figure 4; this should be substantiated; in addition, data from Figure 2 adds little and should be combined with Figures 3 and 4 to make a single stronger message.

It is/was unclear why the authors investigated the distribution of CD34 and CD33 and the major and important question remained to be answered: whether the symmetric cell divisions were/are differentiating or self-renewal in nature hence, the asymmetric cell divisions promoted by the osteoblasts may represent asymmetric self-renewal ones; this needs to be investigated further and potential molecular mechanisms for such promotion should be highlighted.

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

Reviewer #1

• Albeit the link between CSA and NE integrity the work is in my eyes too preliminary. Although the data presented are well done and carefully evaluated they mostly (except Fig 1A) rely on direct comparisons of one patient cell line (CS-A or CS-B) to the same cells expression the wildtype protein. It remains thus open whether the effects seen on LEM2 expression, LEM2-LaimA/C interaction, stress fibre formation, cGAS/STING signaling pathway activation in the CS-A cells are representative for a number of different CS patient derived cells. This is especially important given the small changes observed. Please note that there are alos clear differences between the CSA-wt and CSB-wt cells. Would the HPA CSA KO cells show in addition to NE irregularities (not even quantified) the same phenotypes and can they be reverted by re-expression of the wildtype CSA protein? We would like to thank the reviewer for this comment. Indeed we have previously observed nuclear circularity defects in CSA KO HAP1 cells but we haven’t investigated the other phenotypes in this cell line. One of the main reasons behind this is the technical difficulties associated with performing immunofluorescence with HAP1 cells who are very small and tend to grow in aggregates.

Proposed experimental plan: To address this reviewer’s comment, we will attempt again to use HAP1 cells (WT and CSA KO) and look at nuclear circularity (with quantification), stress fiber formation and cGas foci. If we don’t succeed, we will use an alternative isogenic cell model consisting of fibroblasts in which we will knock out CSA using CRISPR/Cas9. We will then repeat the same experiments as proposed above for the HAP1 cells.

• The link between CSA and SUN1 is not well worked out. What is the effect of SUN2 downregulation and that of nespirins? It remains unclear whether the observed effects are indeed LINC mediated. Proposed experimental plan: To address this point, we will downregulate SUN2 and nesprins using siRNAs in two different cell models (as described above) and assess nuclear shape as well as cGas/STING pathway activation.

Minor comments:

* Fig 1B: Why is HA-Tagged CSA not shown on the CSA western? This would be helpful to compare to the endogenous levels at least in CSB cells. A western showing an housekeeping marker would allow better comparison. Judging from the proteins markers HA-tagged CSA seems much larger as endogenous CSA (first versus second row). Again, less cropped western blots would help.*

We are sorry for the confusion and we have realised that the molecular weights on the western blots were incorrectly labeled on this figure. This will be modified, and the full, uncropped WB will be provided as a supplementary file.

Fig 3A: Is CSA FLAG or HA-tagged? Or both? If both are expressed the question raises of why the CSA-LEM 2 interaction is only seen in an overexpression situation.

• *

CSA is HA tagged. To address this reviewer’s comment we will try performing immunoprecipitation on endogenous proteins.

Fig 5: Inconsistency between figure and figure legend: 20 vs 25 nM Jasplakinolide. I assume Latrunculin A should read Cytochalasin D?

Thank you for pointing this out, and yes this is indeed a mistake on our labeling. We will rectify this on the figure legend.

Fig5B: Not clear why in "CSA-wT cells" Cytochalasin D and Jasplakinolide have the same effect on nuclear envelope shape yet only Jasplakinolide increases the number of blebs.

Cyt D inhibits actin polymerization while jasplakinolide increases polymerization. Likely actin polymerization increases blebs through extra force being put on the nucleus through actin cables/ actin based motility. Both drugs decrease nuclear roundness as they disrupt the normal actin network leading to worsening of nuclear shape through different mechanisms.

Page 10: Method for IF: 4% (v/v) paraformaldehyde and 2% /v/v) should likely read (w/v). Page 19: replace "withl" by "with".

We will rectify both these points.

Reviewer #2

The paper is well-written and for the most part, the data support the conclusions of the authors. Some minor caveats could be addressed to improve the quality of the manuscript.

We would like to thank the reviewer for their positive feedback on our manuscript.

• The phenotype of decreased LEMD2 incorporation into the NE in CS-A cells is minor. Only ~20% and thus, it is not clear whether this is causal of any of the NE abnormalities. It should be better explained how these data add to the story.

To address this point, we will overexpress LEMD2 in CS-A cells and assess whether the NE phenotype can be significantly rescued. This will add value to this part of story.

• Inducing actin polymerization and depolarization impact nuclear morphological abnormalities and nuclear blebbing. Do these treatments impact nuclear fragility and cGAS accumulation at NE break sites?* This is a good point indeed. To address this question, we will include cGas foci staining and quantification upon treatment with these chemicals.

• Depletion of SUN1 in CS-A cells increased nuclear circularity, decreased blebbing, and phosphorylation of TBK1. The impact of SUN1 depletion in cGAS foci formation at NE break sites and phosphorylation of STING is not shown. Such experiments will provide stronger evidence that CS-A activates the cGAS-STING pathway in a SUN1 (mechanical stress)-dependent manner.

* We will address this question by analysing cGas foci and cGas-STING pathway activation upon SUN1 depletion by siRNA.

Reviewer #3

• *

The data are generally clear, well performed and well interpreted with some exceptions:*

1) I appreciate the use of isogenic cell lines (a big plus when dealing with patient-derived cell lines). However, these lines were established 30 years ago and the reported phenotypes might be due to genetic drifts. To exclude this, I suggest to complement the HAP-1 ERCC8 KO cell line with exogenously expressed CSA and assess if this rescues the phenotypes reported. Validation of the KO in these lines, either by western blotting or sequencing is needed.*

This point has also been raised by the first reviewer, and will be addressed as described above (and pasted below):

We would like to thank the reviewer for this comment. Indeed we have previously observed nuclear circularity defects in CSA KO HAP1 cells but we haven’t investigated the other phenotypes in this cell line. One of the main reasons behind this is the technical difficulties associated with performing immunofluorescence with HAP1 cells who are very small and tend to grow in aggregates.

Proposed experimental plan: To address this reviewer’s comment, we will attempt again to use HAP1 cells (WT and CSA KO) and look at nuclear circularity (with quantification), stress fiber formation and cGas foci. If we don’t succeed, we will use an alternative isogenic cell model consisting of fibroblasts in which we will knock out CSA using CRISPR/Cas9. We will then repeat the same experiments as proposed above for the HAP1 cells.

2) Related to the complementation of patient cell lines, the exogenous HA-CSA is not recognised by the anti-CSA in the CSA-null patient cell lines (Fig 1B, second blot). Shouldn't you be able to see this exogenous protein? HA-GFP-CSB in the complemented CSB-null patient cell line runs at the same weight as endogenous CSB (Fig 1B, fourth blot). This is also unexpected. I think you need better characterisation of your cell lines and need to demonstrate the level of exogenous transgenes that have been used to complement the cells and that they localise appropriately, presumably to the nucleus. You should also make sure to cite the paper where they were isolated and describe that they were immortalised (Troelstra et al., 1992) and the paper in which transgenes were stably overexpressed (Qiang et al., 2021).*

*

As mentioned above, we have realised that we made some mistakes with the labeling of the molecular weight on this western blot. This will be corrected and the full uncropped western blot will be provided as a supplementary figure.

We will also cite the suggested papers accordingly.

3) Immunolocalisation of INM proteins is notoriously tricky and the permeabilisation steps include only 0.2% Tx100, which can be insufficient to permeabilise the INM. I appreciate the Emerin and Lamin immunostaining seems to have worked, but in many cases successful immunostaining can be antibody-specific. Can you try harsher permeabilisation to expose LEM2 epitopes? I'm somewhat uncomfortable with the suggestion that there is a cytosolic (ER?) pool of endogenous LEM2 as this runs counter to the literature and feel that your antibody or fixation conditions are illuminating a non-specific protein. The WB in Fig 2E shows that there is virtually no LEM2 in the "soluble" fraction. I would be more cautious on this cytoplasmic/nuclear pool interpretation. Biochemical nuclear and cytoplasmic fractionation would help clarify the signal in a NE vs a non-NE pool.*

*

As suggested by the reviewer, we will try harsher permeabilisation conditions to test the LEMD2 antibody. As we suggest in the manuscript however, we think that the “cytoplasmic” LEMD2 pool we observed by IF in the absence of pre-extraction is indeed unspecific. This is why we have performed the rest of the experiments with a pre-extraction step, that we have shown to give a specific LEMD2 signal that disappear upon depleting LEMD2 by siRNA.

4) Page 15: "Using a Proximity Ligation Assay (PLA), we showed a significant reduction in the number of PLA foci in CS-A cells compared to the WT(HA-CSA) cells, reflecting a reduced number of LEMD2-lamin A/C complexes (Figure 2G, 2H). This data suggests defects in the incorporation of LEMD2 into the NE and lamin protein complexes in CS-A cells". If you have less LEM2 in the NE, it is quite expected that you will have less "LEM2-laminA/C" complexes. To me the logic doesn't hold and this data does not suggest that there is an underlying defect in LEM2-lamin interaction. To ascertain whether there is such a defect one could perform an IP against LEM2 and quantify laminA/C, normalizing by the amount of LEM2 in the input.

We feel we may not have been clear in how we interpreted this data. What we mean is that in each individual cell, the number of Lamin-LEMD2 complexes is decreased, probably indeed due to the fact that there is less LEMD2 altogether within the nucleus in the absence of CSA. We will clarify this in the text.

5) "We overexpressed LEMD2-GFP and Flag-CSA constructs, followed by GFP pulldown in WT(HA-CSA) cells". Since the co-IP data are obtained in overexpression conditions (of both HA-CSA and Flag-CSA?), the authors should validate the interaction between LEM2 and CSA using an orthogonal approach. Perhaps anti-HA capture of the WT(HA-CSA) cells would allow you to immunoblot for endogenous LEM2?*

*

To address this point, we will try to immunoprecipitate HA-CSA and look at endogenous LEMD2.

6) Related to the CSA-LEM2 binding in the above experiment, the procedure involves combining a native detergent-extracted cytoplasmic pool with a denatured (RIPA-extracted) nuclear pool for performing the GFP-trap. From which pool was the tagged CSA bound to LEM2 in?*

*

We are sorry about the confusion. We didn’t try to run the IP from the different pools but instead from the combined pools, to ensure we were looking in the whole cell extract. We would expect however that the interaction occurs in the nuclear pool as both CSA and LEMD2 are nuclear proteins.

7) "The absence of CSA in CS-A patient cells does not affect the mobility of LEMD2 at the NE but instead decreases its interaction with A-type lamins". To me the fact that loss of CSA decreases LEM2-lamin interaction is not well supported (see point 3).

• *

See our response to point 4

8) "Here, we showed by immunoprecipitation that LEMD2 also interacts with CSA. This suggests that the recruitment and stabilization of LEMD2 to the NE is mediated by an interaction with CSA, although the mechanism remains unclear". I think this is an overstatement: there are no data suggesting that CSA recruits or stabilises LEM2 at the NE.

* *We will tone down this statement in the text

9) As the authors suggest in the discussion, it would be worth checking whether LEM2 overexpression is able to rescue some of the NE defects reported, strengthening the hypothesis that LEM2 levels are at least in part responsible for the phenotypes reported.

To address this point, we will perform LEMD2 overexpression in the CSA cells, and analyse the nuclear envelope defects and ruptures (shape and cGas foci quantification)

10) To me it is not clear how the reported phenotypes are interrelated. The first part of the manuscript shows that CSA interacts with LEM2, and that loss-of-function CSA impacts on LEM2 levels and LEM2-lamin interaction, suggesting a direct role for CSA at the nuclear envelope. The second part of the manuscript shows that cells with defective CSA have more actin stress fibres and releasing the cytoskeleton-nuclear tethering is able per se to rescue the nuclear membrane and cGAS phenotypes. How do the authors reconciliate these two parts? Is CSA directly involved in both inner nuclear membrane homeostasis and actin cytoskeleton modulation or is this latter role upstream and the NE defects a mere consequence of increased cytoskeleton rigidity?

At this point indeed we cannot draw definitive conclusions as to whether the two described phenotypes are inter-related. However, by addressing the other points raised by the reviewers, we hope this will help clarifying the mechanism.

11) It is not clear how or why actin stress fibres are elevated in the CS-A cells. Can the authors provide any insight based on their RNAseq analysis? Demonstrating a link to ROCK, LIMK or Rho signalling would be interesting and verifying ppMLC2 levels would help explain why contractility is enhanced. Additionally, is the increase in contractility dependent upon any of the genes identified as up- or downregulated in RNAseq? Presently, the manuscript is missing a link between its two halves.*

*

We would like to reiterate that the RNASeq analysis we performed was done on previously published data from another group (as described in the text). To address the point raised by the reviewer, we will look more specifically into our analysis to look at ROCK, LIMK or Rho signalling to see if any of these pathways appear to be modulated by the absence of CSA.

12) Related to point 1, the RNAseq comparison was performed on patient cells lacking CS-A and patient cells lacking CS-A and later over-expressing HA-CSA, and this comparison is used extensively for phenotype description in the manuscript. In isn't clear to me that this is the most insightful comparison to make; the rescue by overexpression is not as elegant as CRISPR reversion and the ko fibroblasts have presumably been surviving well in culture without CS-A before this protein was overexpressed. Can you validate the differential expression of any identified proteins in the acute HAP1 ko? Can you validate any of the differentially expressed proteins in comparison to normal fibroblasts (e.g., 13O6, as per Qiang et al., 2021)?

As we will validate our experiments in an additional cell model (as described above), we will also indeed validate the level of expression of cytoskeletal proteins upon CSA KO/rescue.

Minor comments

* - Page 14: "To characterize the NE phenotypes further, we obtained CS patient-derived cell lines carrying loss-of-function mutations in CSA (CS-A cells) or CSB (CS-B cells), and their respective isogenic control cell lines (WT(HACSA) and WT(HA-GFP-CSB))." What type of loss-of-function? Is the mutant protein still produced? In Fig 6A there seem to be a band in the CS-A blot (second lane), but in Fig 1B, there isn't. I think this is important to know to interpret the phenotype related to LEM2 interaction.*

We can clarify that in the text. Indeed, the loss of function mutation leads to the absence of CSA protein.

- Figure 1B is poorly annotated. What do - and + stand for? In general, I find a bit confusing how the WB are presented throughout the manuscript, specifically how the antibodies are reported (e.g., HA-CSA instead of HA). Please mark up all western blots with antisera used. Please make sure all expected bands are within the crops - e.g., Fig 3B, the anti-LEM2 blot should be expanded vertically to show the LEM2-GFP relative to endogenous LEM2.

We will correct these on the figures

- From the methods, it appears that you obtained a Please provide clarity on which construct was used in which figure, and verify that an N-terminally tagged LEM2 still localises to the NE.

We actually cloned LEMD2 into an empty pEGFP vector but still maintained LEM2-GFP. We will remove the C1plasmid from the methods to avoid confusion as we removed the MCS and GFP and just used the blank vector and inserted lem2-gfp as we obtained it.

- Fig 1I: there is some text on top of the upper panels (DAPI, cGAS, Merge).

• *

* - "Through gene ontology analysis, we found that genes involved in endoplasmic reticulum (ER) stress were differentially expressed (Figure 4B)". I don't think that the way data are shown in Fig 4B is effective. Since GO has been performed, I would replace the table with a GO enrichment analysis graph. Ensure to report all the data in a supplementary .xls so that others can see and reuse it. Is there a mandated repository that accepts RNAseq data?*

The RNAseq experiment and data was performed by another group and reported in a previous study, as referenced in the main text of the manuscript (Epanchintsev A, Costanzo F, Rauschendorf MA, Caputo M, Ye T, Donnio LM, et al. Cockayne’s Syndrome A and B Proteins Regulate Transcription Arrest after Genotoxic Stress by Promoting ATF3 Degradation. Mol Cell. 2017 Dec;68(6):1054-1066.e6.). Here, we only re-analysed their data using STRING pathway analysis, as detailed in the Material and Methods. However, as suggested by the reviewer, we will replace the table by a GO enrichment graph.

- The volcano plot looks weird with many values at the maximum log10 (P-value) - is the data processed appropriately?

As mentioned above, the RNA Seq analysis was performed and published in a different study. We think this is because the Y axis shows adjusted P values.

- Figure 5B: the legend says "Latrunculin A". Please correct.

We will correct this

- For a Wellcome funded researcher, I'm surprised that the mandated OA statement and RRS is absent from the acknowledgements.

We will of course comply with the open access policy of the Wellcome Trust. However, and based on the WT requirements detailed on their website, we believe the acknowledgement section complies with the funder’s policy: “All research publications must acknowledge Wellcome's support and list the grant reference number which funded the research reported.”

Maybe mention the changes in nuclear shape is not a causative of nuclear blebbing. But maybe not say that they are completely mutually exclusive phenotype to each other.

suggestion

Maybe say that we will overexpress LEMD2 in CS-A cells and show that the NE phenotype can be significantly rescued. This will add value to this part of story. I remember when I did the FRAP experiment, CS-A cells with expression of LEMD2-GFP (that doesn’t form aggregates) looks better in term of shape.

I think Anne, please check the plasmid map? According to the lab inventory (Plasmids Anne), it is LEMD2-GFP. So probably GFP is at C-terminus.

I think there was a part in discussion was LEMD2-GFP was mistakenly written as GFP-LEMD… But I am sure I used LEMD2-GFP throughout the work

We cloned it into an empty pEGFP vector but still maintained LEM2-GFP. Maybe remove the C1 in the methods to avoid confusion as we removed the MCS and GFP and just used the blank vector and inserted lem2-gfp as we obtained it.

Same construct was used for GFP pulldown and for FRAP. And we can see in FRAp that they localise to the NE. SO it should localise to the NE. Maybe mention that we will do a LEMD2-GFP over expression experiment in CS-A cells and show that they do localise to the NE.

I don’t remember fully if Denny did this and what came out. I thought he did and ER stress and cytoskeleton regulation came out as enriched terms?

Denny will have to check this but I think this is because the Y axis shows adjusted P values?? I have the same in my data and Jack told me this is an artefact of the analysis if you adjust for multiple comparisons and is something more often seen in mass spec data

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #3

Evidence, reproducibility and clarity

Summary

In this manuscript, Yang and colleagues report that the absence of CSA (a protein with a well-characterised role in DNA damage repair and whose mutations cause a premature ageing syndrome) leads to defects in nuclear envelope (NE) integrity. Using a Cockayne syndrome patient-derived cell line they show that CSA deficiency correlates with the presence of an aberrant nuclear morphology, nuclear blebbing and increased cGAS-signalling. Moreover, they suggest that this CSA defective cell line presents decreased LEM2 at the NE and a consequent reduction in LEM2-laminA/C interaction. They were able to observe interaction betweenCSA and LEM2, andsuggested that CSA might interfere with LEM2 functionality at the nuclear envelope. Starting from the analysis of an already published RNAseq dataset, the authors go on investigating the cytoskeleton status in the CSA mutant cell line and identify an increase in actin, which they suggest leads to an increase in actin stress fibres. By promoting general actin depolymerisation and by disanchoring the cytoskeleton from the nuclear envelope, Yang et al are able to rescue the dysfunctional phenotypes reported above, suggesting that increased mechanical forces transmitted from the actin cytoskeleton to the nuclear envelope might cause defects in nuclear envelope integrity. Finally, the authors report that alleviating the nuclear envelope defects in CSA mutant cells does not decrease their sensitivity to a UV mimetic, hinting that CSA role in maintaining nuclear integrity is independent from its DNA damage repair function.

Major comments

The key conclusions of the manuscript are convincing and supported by the data. However, there are some conceptual concerns that limit my enthusiasm.

Firstly, that the actin cytoskeleton can transmit forces to the NE via the LINC complex is established. That these forces that lead to NE deformation, and in some cases, rupture, is somewhat unsurprising. The interesting observation in this manuscript is that CSA mutant patient cells exhibit more stress fibres. This leads to more LINC-dependent transmission of forces to the NE, with consequential effects on NE morphology and rupture. However, the mechanism by which actomyosin contractility is elevated in CSA-mutant patient cells is unexplored.

Secondly, it is unclear whether CSA controls the quality, or simply the number, of LEM2/Lamin interactions. As such, a mechanistic link between elevated stress fibres and CSA-dependent NE/lamina interactions is not provided and the paper sits as two presently unconnected observations.

The data are generally clear, well performed and well interpreted with some exceptions:

1. I appreciate the use of isogenic cell lines (a big plus when dealing with patient-derived cell lines). However, these lines were established 30 years ago and the reported phenotypes might be due to genetic drifts. To exclude this, I suggest to complement the HAP-1 ERCC8 KO cell line with exogenously expressed CSA and assess if this rescues the phenotypes reported. Validation of the KO in these lines, either by western blotting or sequencing is needed.
2. Related to the complementation of patient cell lines, the exogenous HA-CSA is not recognised by the anti-CSA in the CSA-null patient cell lines (Fig 1B, second blot). Shouldn't you be able to see this exogenous protein? HA-GFP-CSB in the complemented CSB-null patient cell line runs at the same weight as endogenous CSB (Fig 1B, fourth blot). This is also unexpected. I think you need better characterisation of your cell lines and need to demonstrate the level of exogenous transgenes that have been used to complement the cells and that they localise appropriately, presumably to the nucleus. You should also make sure to cite the paper where they were isolated and describe that they were immortalised (Troelstra et al., 1992) and the paper in which transgenes were stably overexpressed (Qiang et al., 2021).
3. Immunolocalisation of INM proteins is notoriously tricky and the permeabilisation steps include only 0.2% Tx100, which can be insufficient to permeabilise the INM. I appreciate the Emerin and Lamin immunostaining seems to have worked, but in many cases successful immunostaining can be antibody-specific. Can you try harsher permeabilisation to expose LEM2 epitopes? I'm somewhat uncomfortable with the suggestion that there is a cytosolic (ER?) pool of endogenous LEM2 as this runs counter to the literature and feel that your antibody or fixation conditions are illuminating a non-specific protein. The WB in Fig 2E shows that there is virtually no LEM2 in the "soluble" fraction. I would be more cautious on this cytoplasmic/nuclear pool interpretation. Biochemical nuclear and cytoplasmic fractionation would help clarify the signal in a NE vs a non-NE pool.
4. Pag 15: "Using a Proximity Ligation Assay (PLA), we showed a significant reduction in the number of PLA foci in CS-A cells compared to the WT(HA-CSA) cells, reflecting a reduced number of LEMD2-lamin A/C complexes (Figure 2G, 2H). This data suggests defects in the incorporation of LEMD2 into the NE and lamin protein complexes in CS-A cells". If you have less LEM2 in the NE, it is quite expected that you will have less "LEM2-laminA/C" complexes. To me the logic doesn't hold and this data does not suggest that there is an underlying defect in LEM2-lamin interaction. To ascertain whether there is such a defect one could perform an IP against LEM2 and quantify laminA/C, normalizing by the amount of LEM2 in the input.
5. "We overexpressed LEMD2-GFP and Flag-CSA constructs, followed by GFP pulldown in WT(HA-CSA) cells". Since the co-IP data are obtained in overexpression conditions (of both HA-CSA and Flag-CSA?), the authors should validate the interaction between LEM2 and CSA using an orthogonal approach. Perhaps anti-HA capture of the WT(HA-CSA) cells would allow you to immunoblot for endogenous LEM2?
6. Related to the CSA-LEM2 binding in the above experiment, the procedure involves combining a native detergent-extracted cytoplasmic pool with a denatured (RIPA-extracted) nuclear pool for performing the GFP-trap. From which pool was the tagged CSA bound to LEM2 in?
7. "The absence of CSA in CS-A patient cells does not affect the mobility of LEMD2 at the NE but instead decreases its interaction with A-type lamins". To me the fact that loss of CSA decreases LEM2-lamin interaction is not well supported (see point 3).
8. "Here, we showed by immunoprecipitation that LEMD2 also interacts with CSA. This suggests that the recruitment and stabilization of LEMD2 to the NE is mediated by an interaction with CSA, although the mechanism remains unclear". I think this is an overstatement: there are no data suggesting that CSA recruits or stabilises LEM2 at the NE.
9. As the authors suggest in the discussion, it would be worth checking whether LEM2 overexpression is able to rescue some of the NE defects reported, strengthening the hypothesis that LEM2 levels are at least in part responsible for the phenotypes reported.
10. To me it is not clear how the reported phenotypes are interrelated. The first part of the manuscript shows that CSA interacts with LEM2, and that loss-of-function CSA impacts on LEM2 levels and LEM2-lamin interaction, suggesting a direct role for CSA at the nuclear envelope. The second part of the manuscript shows that cells with defective CSA have more actin stress fibres and releasing the cytoskeleton-nuclear tethering is able per se to rescue the nuclear membrane and cGAS phenotypes. How do the authors reconciliate these two parts? Is CSA directly involved in both inner nuclear membrane homeostasis and actin cytoskeleton modulation or is this latter role upstream and the NE defects a mere consequence of increased cytoskeleton rigidity?
11. It is not clear how or why actin stress fibres are elevated in the CS-A cells. Can the authors provide any insight based on their RNAseq analysis? Demonstrating a link to ROCK, LIMK or Rho signalling would be interesting and verifying ppMLC2 levels would help explain why contractility is enhanced. Additionally, is the increase in contractility dependent upon any of the genes identified as up- or downregulated in RNAseq? Presently, the manuscript is missing a link between its two halves.
12. Related to point 1, the RNAseq comparison was performed on patient cells lacking CS-A and patient cells lacking CS-A and later over-expressing HA-CSA, and this comparison is used extensively for phenotype description in the manuscript. In isn't clear to me that this is the most insightful comparison to make; the rescue by overexpression is not as elegant as CRISPR reversion and the ko fibroblasts have presumably been surviving well in culture without CS-A before this protein was overexpressed. Can you validate the differential expression of any identified proteins in the acute HAP1 ko? Can you validate any of the differentially expressed proteins in comparison to normal fibroblasts (e.g., 13O6, as per Qiang et al., 2021)?

Minor comments

• Pag 14: "To characterize the NE phenotypes further, we obtained CS patient-derived cell lines carrying loss-of-function mutations in CSA (CS-A cells) or CSB (CS-B cells), and their respective isogenic control cell lines (WT(HACSA) and WT(HA-GFP-CSB))." What type of loss-of-function? Is the mutant protein still produced? In Fig 6A there seem to be a band in the CS-A blot (second lane), but in Fig 1B, there isn't. I think this is important to know to interpret the phenotype related to LEM2 interaction.
• Figure 1B is poorly annotated. What do - and + stand for? In general, I find a bit confusing how the WB are presented throughout the manuscript, specifically how the antibodies are reported (e.g., HA-CSA instead of HA). Please mark up all western blots with antisera used. Please make sure all expected bands are within the crops - e.g., Fig 3B, the anti-LEM2 blot should be expanded vertically to show the LEM2-GFP relative to endogenous LEM2.
• From the methods, it appears that you obtained a LEM2-GFP, and cloned it into an expression vector (pEGFPC1) to make GFP-LEM2. Please provide clarity on which construct was used in which figure, and verify that an N-terminally tagged LEM2 still localises to the NE.
• Fig 1I: there is some text on top of the upper panels (DAPI, cGAS, Merge).
• "Through gene ontology analysis, we found that genes involved in endoplasmic reticulum (ER) stress were differentially expressed (Figure 4B)". I don't think that the way data are shown in Fig 4B is effective. Since GO has been performed, I would replace the table with a GO enrichment analysis graph. Ensure to report all the data in a supplementary .xls so that others can see and reuse it. Is there a mandated repository that accepts RNAseq data?
• The volcano plot looks weird with many values at the maximum log10 (P-value) - is the data processed appropriately?
• Figure 5B: the legend says "Latrunculin A". Please correct.
• For a Wellcome funded researcher, I'm surprised that the mandated OA statement and RRS is absent from the acknowledgements.

Referees cross-commenting

I think the other reviews are fair and accurate

Significance

This work provides interesting insights on a possible new moonlighting role of the CSA protein. This could enhance the pathophysiological comprehension of some clinical manifestations in CS patients. The nuclear envelope defects are well described and convincing; however, there is no clear understanding of how (or even if) the reported cytoskeleton and nuclear envelope defects depend on CSA. Better characterising mechanistic roles of CSA in both nuclear envelope integrity and in stress fibre formation will boost the overall impact of the manuscript. At this stage, the manuscript could be of interest for a specialised audience (premature ageing syndromes) but with more mechanistical dissection it could become of broader interest (basic research, nuclear architecture/integrity readership).

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

The manuscript by Yang et al. shows that ERCC8 (Cockayne syndrome A protein) mutations/loss of function causes nuclear envelope (NE) abnormalities and fragility, in addition to the well described deficiencies in transcription-coupled nucleotide excision repair (TC-NER). Cells from CS-A patients show decreased LEMD2-lamin A/C complexes at the NE and increased actin stress fibers that cause mechanical stress in the NE, in addition to more blebbing and ruptures of NE. This is turn causes activation of the innate/immune cGAS/STING pathway. Importantly, disrupting the LINC complex rescued NE problems and activation of the cGAS/STING pathway. This effect of ERCC8 dysfunction on NE integrity and activation of the cGAS/STING pathway may be behind patients' phenotypes of neuroinflammation, but not UV sensitivity.

The paper is well-written and for the most part, the data support the conclusions of the authors. Some minor caveats could be addressed to improve the quality of the manuscript.

• The phenotype of decreased LEMD2 incorporation into the NE in CS-A cells is minor. Only ~20% and thus, it is not clear whether this is causal of any of the NE abnormalities. It should be better explained how these data add to the story.
• Inducing actin polymerization and depolarization impact nuclear morphological abnormalities and nuclear blebbing. Do these treatments impact nuclear fragility and cGAS accumulation at NE break sites?
• Depletion of SUN1 in CS-A cells increased nuclear circularity, decreased blebbing, and phosphorylation of TBK1. The impact of SUN1 depletion in cGAS foci formation at NE break sites and phosphorylation of STING is not shown. Such experiments will provide stronger evidence that CS-A activates the cGAS-STING pathway in a SUN1 (mechanical stress)-dependent manner.

Referees cross-commenting

I am more positive than the other reviewers. I agree with reviewer 1 that another patient line would be great, and that some of the westerns need improvement. However, I still find that the phenotypes found in this cell line, which are rescued by the wild-type protein, are interesting and worth reporting. I do not fully agree with technical recommendations from reviewer 3, or suggesting that that the authors address questions outside the focus of their study.

Significance

The significance of the study is that shows for the first time the nuclear envelop defects and fragility in cells from Cockayne Syndrome patients (mutations in ERCC8) are due to mechanical stress that is alleviated by depletion of the LINC complex. In addition, the study shows activation of the cGAS-STING pathway in these cells, although the evidence about this phenotype could be strengthen.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

Summary:

Cockayne syndrome (CS), an autosomal recessive premature ageing/progeria disease, is caused by mutations in the genes coding for CSA and CSB. These play a well-studied role in transcription-coupled nucleotide excision repair. In this manuscript, CSA but not CSB dysfunction is associated with nuclear envelope defects, a typical phenotype of cells from progeria diseases caused by mutations in NE proteins. HPA model cells lacking CSA show an irregular nuclear envelope not observed in cells lacking CSB. Patient cells with defective CSA, compared to isogenic cells expressing the wild-type CSA protein, show more severe nuclear envelope deformation and blebbing and activation of the cGAS/STING signaling pathway, indicating nuclear envelope damage, a slight reduction in the expression of LEM2, a protein of the inner nuclear membrane, and a reduced interaction of LEM2 with the lamina component lamin A/C. Patient cells with defective CSA show aberrant acThe proposed link between CSA and nuclear envelope integrity is interesting. CSA would in this case a factor involved in DNA repair as well as nuclear envelope integrity, two processes linked to progeria. However, the work relies on a single patient derived cell line which is compared to the same cell line expressing the wild type protein. To consolidate, more cells form different CSA patients need to be included. If worked out this idea will be interesting all cell biologists interested in DAN damage repair and nuclear envelope structure and function as well as researchers interested in the molecular mechanisms of progeria syndromes. My expertise: nuclear envelope structure and function, nuclear transport tin stress fiber and their nuclei show different effects toward actin polymerization inhibitors or stabilizers. Downregulation of SUN1, one of the nuclear envelope proteins linking to the cytoskeleton, improves nuclear envelope blebbing phenotypes in CSA patient cells

Major comments:

Albeit the link between CSA and NE integrity the work is in my eyes too preliminary. Although the data presented are well done and carefully evaluated they mostly (except Fig 1A) rely on direct comparisons of one patient cell line (CS-A or CS-B) to the same cells expression the wildtype protein. It remains thus open whether the effects seen on LEM2 expression, LEM2-LaimA/C interaction, stress fibre formation, cGAS/STING signaling pathway activation in the CS-A cells are representative for a number of different CS patient derived cells. This is especially important given the small changes observed. Please note that there are alos clear differences between the CSA-wt and CSB-wt cells. Would the HPA CSA KO cells show in addition to NE irregularities (not even quantified) the same phenotypes and can they be reverted by re-expression of the wildtype CSA protein? The link between CSA and SUN1 is not well worked out. What is the effect of SUN2 downregulation and that of nespirins? It remains unclear whether the observed effects are indeed LINC mediated.

Minor comments:

Fig 1B: Why is HA-Tagged CSA not shown on the CSA western? This would be helpful to compare to the endogenous levels at least in CSB cells. A western showing an housekeeping marker would allow better comparison. Judging from the proteins markers HA-tagged CSA seems much larger as endogenous CSA (first versus second row). Again, less cropped western blots would help.

Fig 3A: Is CSA FLAG or HA-tagged? Or both? If both are expressed the question raises of why the CSA-LEM 2 interaction is only seen in an overexpression situation.

Fig 5: Inconsistency between figure and figure legend: 20 vs 25 nM Jasplakinolide. I assume Latrunculin A should read Cytochalasin D?

Fig5B: Not clear why in "CSA-wT cells" Cytochalasin D and Jasplakinolide have the same effect on nuclear envelope shape yet only Jasplakinolide increases the number of blebs.

Page 10: Method for IF: 4% (v/v) paraformaldehyde and 2% /v/v) should likely read (w/v).

Page 19: replace "withl" by "with".

Significance

The proposed link between CSA and nuclear envelope integrity is interesting. CSA would in this case a factor involved in DNA repair as well as nuclear envelope integrity, two processes linked to progeria. However, the work relies on a single patient derived cell line which is compared to the same cell line expressing the wild type protein. To consolidate, more cells form different CSA patients need to be included. If worked out this idea will be interesting all cell biologists interested in DAN damage repair and nuclear envelope structure and function as well as researchers interested in the molecular mechanisms of progeria syndromes.

My expertise: nuclear envelope structure and function, nuclear transport

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

The authors do not wish to provide a response at this time.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #3

Evidence, reproducibility and clarity

The authors conducted a comprehensive investigation into the spatial and temporal expression patterns of Pcdh10 during brain development and employed two deficient mouse models to delve into its neurobehavioral functions, with a specific emphasis on ultrasonic vocalizations (USV). Interestingly, the authors found that heterozygous cKO pups showed an exaggerated effect on USV as compared to the ubiquitous heterozygous KO pups. These observations in general agree with previous studies of Pcdh10's complicated function in ASD-related neurodevelopment, as well as USV, indicating its important function in neurodevelopment. However, several key concerns warrant attention:

Major comments:

1. The authors analyzed published scRNAseq adult mouse dataset and found that Pcdh10-expressing cells differentially expressed genes involved in vocalization behaviors, including Foxp2, Cntnap2, Nrxn1 and Nrxn3, as compared to cells that did not express Pcdh10. Given that the authors also performed bulk RNA-seq experiments using the sorted cKO cells, it would be interesting and important to include these data in the analysis and validate whether the above genes are differentially expressed in the animal model in the current study. This may provide additional molecular mechanisms of the behavioral abnormality observed in the animal models.
2. In Figure 1S, the synapse assembly is one of the most significant GO-terms. Does Pcdh10 deficiency affect neuromophology and synaptic density/assembly? Characterizing the neuroanatomy of the animal model will provide a neuronal explanation for the behavioral abnormalities observed in this study.
3. Could the authors explain why cHE mice display a stronger phenotype than cKO at P6? Also, why cHE and cKO show stronger phenotypes than whole-body KO? Does this data indicate that loss of Pcdh10 in the inhibitory neurons only resulted in E-I in-balance? The authors should at least discuss the possibilities of this result they observed.

Minor comments:

1. It would be interesting and important to know the level of PCDH10 at the adult stage after P7 to learn whether this gene also plays an important role beyond early development.
2. Besides USV, does Pchd10 deficiency in mice show other autistic phenotypes in adolescence and adulthood, such as social interaction and repetitive behavior?

Significance

The experiments are well designed and the detailed characterization of USV in the animal models would be informative for the audience who are interested in Pcdh10's function in neurodevelopment and Social Communication. However, because previous studies already showed the function of PCDH10 in multiple mouse models, the novelty of this study is limited. The observation that cKO/cHE mice show stronger phenotype than whole-body KO mice could be potentially interesting, it would be nice if the authors could provide some explanation of this result with additional experimental evidence.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

Summary

In this paper the authors present a new cKO mouse model for PCDH10-related ASD. This model consists of the ablation of PCDH10 specifically in interneurons of the basolateral complex. Interestingly, the use of this mouse model together with a complete KO adds evidence towards an excitatory/inhibitory imbalance causing the ASD phenotype, rather than the complete ablation of a protein (in this case PCDH10). Firstly, the developmental dynamics of PCDH10 are measured with diverse techniques. The presence of PCDH10 at embryonic stages in different areas of the basal forebrain. Here it is made clear why the specific downregulation of PCDH10 in Gsh2- lineage interneurons. The cKO mouse model is validated with bulk RNA sequencing fluorescence imaging of the eGFP reporter in the telencephalon. The KO mouse model is validated with WB assay showing the gradual decrease of PCDH10 in Het and Ho mice. Secondly, the USV emitted by isolated pups throughout development (P3, P6, P9 and P12) are analysed with different parameters in both mouse models.

Major Comments

1. The role of anxiety in the phenotype described in this work should be supported by behavioural experiments in adulthood (Open field/light/dark/Plus Maze test)2-3M for mice to reach the appropriate age + 2 weeks for performing the experiments and extracting results.
2. The WB in panel F1C&E should be done with non-pooled biological replicates to be informative 3 weeks
3. The statistical test used for F2B and F3D-I needs to be specified.
4. The reduced GABA input in the amygdala that is hypothesized to be causing the phenotype could be studied by iPSP analysis through LFP (OPTIONAL) 1M

Minor comments

1. The results section should be subdivided in sections corresponding to each figure
2. Detail the pinhole opening in M&M used for the imaging of the images in panel of Figure 1 M-R
3. The group size and the power calculation used to determine it should be detailed in M&M
4. The WB membrane image in Panel 1F has saturated pixels, the image needs to be changed
5. Instead of asterisks, writing the exact p-value is more informative in the graphs
6. Figure 1B & D: detail what the Pcdh10 levels are normalised to. In the legend there's a typo "no-way ANOVA"
7. Figure 1F: specify what it means P17.5 (norm)
8. Figures 3-4: choose higher contrast colours for an easier readability and more accessibility.
9. Figure 3B: the WB image needs to be at a higher resolution
10. Figure 3C: the colour coding for dBFS in the spectrogram needs to be specified in the maximum and minimum number.
11. Figure 3D-I & 6G: results would be more clear if shown as a ratio of the WT (would be also more evident the mouse model differences). Also the titles of the graphs are misleading, as the graphs are showing data from all the genotypes of the mouse models.
12. Figure 2B: specify what it is normalised to
13. Figure 4 E-H: it is not described what the dotted lines correspond to.
14. In all the figures, the panels are excessively subdivided, the following panels should be grouped in one:
    • a. Figure 1: i. C & E are showing the same data

ii. G-I are showing the same data

iii. J-L are showing the same data

iv. M-R are showing the same data - b. Figure 2: panel 2C-G are showing the same data - c. Figure 3: i. A-B are showing the same data

ii. D-I are showing the same data - d. Figure 4: i. A-C are showing the same data

ii. E-H are showing the same data - e. Figure 5: the frequency parameters (A-E) should be all 1 panel f. Figure 6: B-E are showing the same data

Significance

The detailed study of the socio-affective communication of these mouse models is accurate and quite informative. There is still a big body of work to do for classifying and using pup USV as biomarkers for mouse model phenotyping, and this thorough work is a step forward. This type of work will be of interest to neurodevelopmental neuroscientists interested in behaviour and mouse model phenotyping. However, the claim of an autistic-like phenotype would be much stronger with additional behavioural assessments in adulthood (as USV in mating behaviour, social novelty recognition and stereotypical behaviours). In addition, the hypothesis of an excitation/inhibition imbalance is interesting, but correlational in this work. Further experiments would need to be done to prove causality.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

This paper provides evidence that Pcdh10 (associated with autism spectrum disorder) is involved in anxiety-like behavior and socio-affective communication in developing mouse pups. Using a specific Cre-driver line targeting Gsh2-lineage interneurons, the authors performed a series of behavioral analyses, including isolation-induced ultrasonic vocalization. Particularly, the authors provided detailed analyses to provide distinct clusters that might correspond to identified call types. This work is excellent and I do not have any further comments, except one minor comment: why did the authors use CD1 line to generate the mouse lines, instead of C57BL/6 substrains?

Significance

This is of paramount significance, such that Pcdh10 expressed in Gsh2-lineage interneurons (a subpopulation in the basolateral complex of the amygdala) is required for isolation-induced ultrasonic vocalization.

Note: This response was posted by the corresponding author to Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

General Statements

We thank the reviewers for their constructive feedback on our study. We have updated the manuscript according to their recommendations.

Two of the three reviewers raised concerns about the quality of some of our data, in particular of the DP population. Since most DP samples suffered from low coverage, introducing bias when being compared to the higher quality data, we moved all results involving DP to the Supplement. We mention this in the Discussion. The main text and its figures were updated to focus on the remaining three wild populations and two laboratory strains.

We rewrote parts of the Abstract, Introduction and Discussion to clarify the comparison between the NLR families of Arabidopsis thaliana and Danio rerio. In regard to this, we also added an extra panel to figure 3.

Reply to the reviewers

Reviewer #1 (Evidence, reproducibility and clarity):

General comments:<br /> 1) Odd statement in abstract comparing plants to fishes - why not put in context of animals in general, or vertebrates in particular? Same in the introduction and discussion. Either a more broad comparative framework should be set - from plants, through invertebrates to vertebrates, or it should be shortened and perhaps discussed in the light of diversity present in e.g., fish or just vertebrates. A persistent focus on plants is peculiar and does not seem to provide adequate background/context._

We have made an effort to clarify the reason for mentioning NLRs from A. thaliana in the Abstract, Introduction and Discussion. We added a plot to figure 3, which is analogous to figure 3A from Van de Weyer et al., 2019 where the pan-NLRome of A. thaliana is analysed.

Furthermore, we added a paragraph discussing copy number variation of another immune gene family (MHC genes) in vertebrates.

2) Methods seem overall well described, and the analyses seem to have been performed diligently-although I am not an expert in the field. One aspect that would be important to evaluate methodology is the repeatability of NLR detection - If we perform the procedure twice on DNA sample from a single individual, how repeatable collection do we get?

It is possible that we might have failed to identify a few NLRs from some individuals. Unfortunately, sequencing could not be repeated due to the limited amount of raw tissue. The population DP in particular is problematic since we had low sequencing depths across all of its samples. We have now excluded it from the main results and moved it to the Supplement. Still, we are confident that the general feature of widespread copy number variation both within and between populations persists even if the experiments were repeated. One argument for this is that the non-linear functions describing pan-NLRome sizes are fairly consistent among wild populations. Also, sample sizes of about 20 individuals per (wild) population are large enough to compensate for poor data quality of one or two indviduals. Finally, we found a substantial fraction of NLRs to be present in all wild populations, including DP. It would be unlikely to detect so many shared genes if the amplified sequences were spurious or the NLR identification procedure faulty.

We have expanded the discussion of the caveats of our approach.

3) Discussion is quite speculative, and some claims seem exaggerated; in conclusions, eg: "This study advances our understanding of the evolutionary dynamics affecting very large gene families." - as the study is mostly descriptive documentation od CNV of gene (or, gene fragments) I am not convinces how it really advances "understanding of the evolutionary dynamics".

We toned down Discussion and Conclusions.

Minor points:

At this stage the journal and so the audience is unknown, so perhaps this will not be an issue - but for a broader audience, a better explanation of what PRY/SPRY/B30.2 are would be useful._

We added an explanation.

English could be smoothed, certain sentences sound odd: eg. lines 87-89; or 90-93 - Studies have shown that viral and bacterial infections can induce the expression of specific fish NLRs (reviewed in (24)). Some have PYD or CARD domains and can even form inflammasomes similar to mammalian NLRs (25, 26). - could be read as if infections had PYD or CARD domains, not NLRs. Lines 239-240 - not sure what does "presence/absence variation" mean here.

We made an effort to improve language style.

Sequence of baits used should be provided in some supplement or repository.

The sequences are now attached to the manuscript as a supplementary dataset.

Reviewer #1 (Significance):

The article tackles an overall interesting subject: an expansion of a relevant group of genes in one of the major model species of biomedical importance. The main strength of the study is putting in context variation found in the lab strains - via comparison to wild populations. A limited and "haphazard" genetic variation (especially of genes involved in immune processes) of laboratory models can have paramount implications for the interpretation of experimental studies._

We expand on this in the Discussion._

That being said, expansion of NLR family in teleosts in general, and in zebrafish in particular, has been previously described (eg, https://bmcecolevol.biomedcentral.com/articles/10.1186/1471-2148-8-42_), and this study mainly expands description of this phenomena.

The expansion of NLRs has indeed been described before and in great detail. The manuscript contains multiple references on this topic, including the one recommended by the reviewer (e.g. Stein et al., 2007, Laing et al., 2008, Howe et al., 2016). The previously reported expansion of NLRs in the reference genome was a description, but not an interpretation "in the light of evolution". We demonstrate that gene birth, death, and presence/absence variation are ongoing processes and active on a population-genetic time scale.

Given the methodology, little can be said about possible functions of the discovered diversity. In particular, a technical aspect imposing limitation of sequence length resulted in a collection of exons, but not full genes - precluding e.g., deepened analysis of domain architecture.

We agree that our study does not reveal much about the function of the NLR-C genes. The focus on function in the discussion was disproportionate to our findings and we have reduced it to avoid speculation.

Selection analysis is also quite limited in scope; there are several, more sophisticated tools to infer balancing or purifying selection, either pervasive or episodic (see eg. excellent tools of_ https://www.datamonkey.org__/_).

Yes. However, with our limited amount of data, we deliberately do not want to speculate too much about selection. As mentioned in the text, many genes are monomorphic in our sample. To reliably infer the action of selection, and more so to distinguish it from signatures of demographic history, a much broader basis of sequence data is needed.

Nonetheless, the study certainly highlights the possible importance of such an expanded group of genes in this species, with a potential to inspire further - more mechanistic/functional research in this area. The article will likely interest a few groups of rather specialized audience - e.g., those working in NLR genes in particular, or in immunity of zebrafish (or fish in general). Another, somewhat broader angle, able to attract a wider audience would be subjects of general evolution of multigenic families (birth-and-death models, trade-offs at different CNV etc.) and comparative analysis to other groups of animals - or genes - evolving in a similar way. My field of expertise is diversity and evolution of adaptive branch of the immune system in vertebrates, in particular non-model vertebrates.

Reviewer #2 (Evidence, reproducibility and clarity):

Summary:

The authors have examined the variation of the immune related gene family "Nucleotide-binding domain Leucine-rich Repeat containing" (NLR) in the zebrafish Danio rerio. These proteins, while highly divergent, are conserved across animal and plant species. In humans or rodents, the number of these genes is relatively small (20-30) but other species such as Arabidopsis appear to have >10,000 NLR genes.<br /> Schäfer et al. collected 67 wild-type zebrafish from 4 independent sites near the Bay of Bengal and added 8 fish each from the laboratory strains Tübingen (TU) and Cologne (CGN). Using DNA capture and sequencing technology, they identified 1,560 unique "FISNA-NACHT" domains representing different NLR genes and 574 NLR-associated B30.2 domains. A subset, 714 and 229 respectively, were identified in all of the fish populations representing a "core" set common NLR genes. A significant subset of the genes could not be aligned to the GRCz11 reference suggesting significant variation across subpopulations of zebrafish.

Major comments:<br /> None. The work is straight-forward, carefully done and conservatively interpreted.

Minor comments:<br /> GRCz11 is still a very fragmented assembly, particularly chr4 where heterochromatin repeats on the long arm were intractable to short read sequences. There is a more recent assembly generated by the Tree of Life initiative (GCA_944039275.1) from the SAT fish line that may allow a more robust alignment and placement of the NLR genes, perhaps even some phasing of unassembled fragments._

We have examined some of the existing long-read based zebrafish assemblies and found that even the length of chromosome 4 that contains most of the NLRs can significantly differ between different strains and different genome assemblies. This is also in agreement with the recent findings of McConnell et al., 2023 on large structural differences on this chromosome between three strains derived from the AB genetic background. We now mention this finding and the alternative existing zebrafish assemblies in the Discussion.

Reviewer #2 (Significance):

General assessment:<br /> Obtaining more data on the inherent variation within species of NLR genes as well as collecting more cross-species data for evolutionary comparisons is valuable for our understanding of this interesting but poorly understood class of immune response genes as well as the dynamics of gene duplication/deletion in complex, repeated arrays.<br /> -The major strength of the study is the efforts put into capturing wild-type zebrafish from multiple different locations to maximize the diversity of the NLR sequences.<br /> -The primary limitation is targeted capture is always contingent on having enough homology in the probes to capture all the desired genes in roughly even proportions. Some of the more interesting NLR genes might have diverged too much to be properly captured but could provide critical information on evolutionary functional adaptation. Similarly, while based on cost, it is understandable why sequence capture was chosen, but phasing information across clusters could help explain (or discover) very diverse haplotype sequences in the introns of many of these gene arrays. As sequencing costs drop, this is an important aspect to the evolution of zebrafish chromosome 4 (as well as the rest of the genome). "Pangenome" differences in the zebrafish genome might be quite radically different.

Yes, we are aware of this possibility and discussed it now in more detail._

Advance:<br /> The study represents an important characterization of a poorly understood but important class of immune genes. A very similar (but more detailed) characterization of Arabidopsis NLR genes by Van de Weyer et al., Cell 2019, is considered an important advance to the plant biology community and has been highly cited since its publication. The collected data is potentially quite useful for future studies that want to understand the complex dynamics of highly repetitive, yet functional regions of the genome and how it can result in the creation and extinction of new genes.

Audience:<br /> There are two main likely audiences, those interested in how NLR genes are involved immune protection and those interested in genomes and evolution. It probably doesn't reveal a fundamental biological principle that would be of broad, general interest to the research community.

Reviewer #3 (Evidence, reproducibility and clarity):

Summary:<br /> The authors performed exon capture sequencing of FISNACHT and B30.2 domains from individuals among various inbred and wild zebrafish populations. This study provides average numbers of domains per fish and thus a valuable estimate for the size of the pan-NLRome in zebrafish.

Major comments:<br /> - Are the key conclusions convincing?_ Yes, and overall their findings are also consistent with what has been found in other model systems, most notably Arabidopsis (Weyer 2019, https://www.cell.com/cell/fulltext/S0092-8674(19)30837-2).<br /> - Should the authors qualify some of their claims as preliminary or speculative, or remove them altogether?<br /> 'Complex patterns of inheritance' is unclear from significance statement.

We have reformulated this sentence.

For mechanisms driving copy number differences, haplotypes/segregation need not be presented as a model opposed to birth and death gene evolution/tandem duplication. These may both operate at different scales.

Yes, we reformulated.

• Are the data and the methods presented in such a way that they can be reproduced?<br /> Yes, overall, though the authors could provide sequences for their adapters, similar to Weyer 2019. It seems that raw data for this study is not yet available in NCBI: https://www.ncbi.nlm.nih.gov/bioproject/966920

The data will be made available at the time of publication. The sequences for all baits used, including non-NLR-ones, have been added as a supplementary .fasta-formatted text file.

Minor comments:<br /> - Specific experimental issues that are easily addressable.<br /> Their targeted exon capture approach cannot ensure that all NLR genes will be sequenced. This method also does not sequence across other exons from NLR genes, providing only a partial view of NLR gene structure and evolution. This may miss for example additional integrated domain architectures or evidence of physical clustering in the genome (Thatcher 2023, https://bsppjournals.onlinelibrary.wiley.com/doi/10.1111/mpp.13319). Nevertheless, focusing on capturing and enumerating FISNACHT domains as the most conserved and characteristic domain for NLRs appears to be a reasonable approach for estimating rough copy numbers.

We agree with the reviewer and have now added explanatory text on this topic to the discussion.

Unfortunately, mapping these domains back to only a single reference genome also means that the accuracy of their predictions for these domains, including which belong to genes or pseudogenes, is likely to be more error-prone (Wang 2019,https://www.pnas.org/doi/10.1073/pnas.1910229116), particularly for more divergent sequences. This may lead to additional limitations for assigning one-to-one relationships between FISNACHT domains, which they measure, and genes, which they seek to enumerate.

Thank you for pointing out the above reference. This is a limitation that we realized early on in the study after which we decided to not rely on the reference genome at all, instead opting to assemble and cluster all data de novo. In the manuscript we only use the reference genome for checking which of our identified NLRs have a clear homologue to satisfy potential interest by the biomedical community. All the other analyses were conducted based on the de novo assemblies and orthologous clusters._

In addition, many of their samples appear to be of poor DNA quality, meaning that

modeling estimates of domain copy number for the population can appear erroneous, even perhaps off by as much as a factor of 10 for the DP strain. This may contribute to some curious 'artifact' in some of the figures, such as Figs S2B1, S3B, S4C2, particularly for the B30.2 sequences in DP. The authors mention 'low sequencing depths' and 'low coverage' as contributing factors for this 'artifact' but also invoke possible 'evolutionary factors' in the discussion. Additional discussion for the apparent experimentally-induced effects on underestimating B30.2 sequences (perhaps due to increased DNA fragmentation and smaller domain size?), rather than coincidence of low coverage and actual biological strain-specific loss of B30.2 domains, appears warranted.

As mentioned in the text, although samples from the population CHT were older than the others and slightly degraded, they were rescued by applying the PreCR Repair Mix from New England Biolabs. We incorporated incubation with the reagent to the standard protocol that we used for all samples. The other samples had no quality problem.

Furthermore, the reviewer raises concerns regarding data quality for the DP population samples. After careful consideration, we decided to exclude the DP data and its analysis from the main text and deferred it to the supplementary material with the necessary remarks of caution. The main text now focuses on the other three wild populations, which did not have those issues.

Additionally, as per the reviewer’s suggestion, we discuss the additional caveats of bait-based targeted sequencing, especially for B30.2 domains._

• Do you have suggestions that would help the authors improve the presentation of their data and conclusions?<br /> Statement in the abstract about fewer (not 'less') NLRs in lab strains vs wild: consider revising to average # NLRs per individual._

We changed the respective sentence in the abstract.

Revise Fig 3C legend: 'Totally discovered NLR genes', e.g., to 'total # NLR genes'.

The recommended change was made to the figure legend.

Reviewer #3 (Significance):

• Describe the nature and significance of the advance (e.g. conceptual, technical, clinical) for the field.<br /> This study elucidates the pan-NLRome of zebrafish via application of a somewhat new technique.

• Place the work in the context of the existing literature (provide references, where appropriate).<br /> The authors apply a targeted capture and long read technology previously used in plants (Weyer 2019) to fish, thereby elucidating the pan-NLRome of zebrafish, a vertebrate with a large # of NLR genes. Previous studies has shown NLR gene variation in zebrafish, but had not compared levels within and between strains, or estimated a total NLR # across populations._

• State what audience might be interested in and influenced by the reported findings.<br /> Those interested in immune function, genetic diversity and genome evolution.

• Define your field of expertise with a few keywords to help the authors contextualize your point of view. Indicate if there are any parts of the paper that you do not have sufficient expertise to evaluate.<br /> Immune genes and evolution. The specific statistical approach_

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #3

Evidence, reproducibility and clarity

Summary:

The authors performed exon capture sequencing of FISNACHT and B30.2 domains from individuals among various inbred and wild zebrafish populations. This study provides average numbers of domains per fish and thus a valuable estimate for the size of the pan-NLRome in zebrafish.

Major comments:

• Are the key conclusions convincing?

Yes, and overall their findings are also consistent with what has been found in other model systems, most notably Arabidopsis (Weyer 2019, https://www.cell.com/cell/fulltext/S0092-8674(19)30837-2).<br /> - Should the authors qualify some of their claims as preliminary or speculative, or remove them altogether?

'Complex patterns of inheritance' is unclear from significance statement.

For mechanisms driving copy number differences, haplotypes/segregation need not be presented as a model opposed to birth and death gene evolution/tandem duplication. These may both operate at different scales.<br /> - Would additional experiments be essential to support the claims of the paper? Request additional experiments only where necessary for the paper as it is, and do not ask authors to open new lines of experimentation.

Not necessary<br /> - Are the suggested experiments realistic in terms of time and resources? It would help if you could add an estimated cost and time investment for substantial experiments.

NA<br /> - Are the data and the methods presented in such a way that they can be reproduced?

Yes, overall, though the authors could provide sequences for their adapters, similar to Weyer 2019. It seems that raw data for this study is not yet available in NCBI: https://www.ncbi.nlm.nih.gov/bioproject/966920<br /> - Are the experiments adequately replicated and statistical analysis adequate?

Yes

Minor comments:

• Specific experimental issues that are easily addressable.

Their targeted exon capture approach cannot ensure that all NLR genes will be sequenced. This method also does not sequence across other exons from NLR genes, providing only a partial view of NLR gene structure and evolution. This may miss for example additional integrated domain architectures or evidence of physical clustering in the genome (Thatcher 2023, https://bsppjournals.onlinelibrary.wiley.com/doi/10.1111/mpp.13319). Nevertheless, focusing on capturing and enumerating FISNACHT domains as the most conserved and characteristic domain for NLRs appears to be a reasonable approach for estimating rough copy numbers.

Unfortunately, mapping these domains back to only a single reference genome also means that the accuracy of their predictions for these domains, including which belong to genes or pseudogenes, is likely to be more error-prone (Wang 2019, https://www.pnas.org/doi/10.1073/pnas.1910229116), particularly for more divergent sequences. This may lead to additional limitations for assigning one-to-one relationships between FISNACHT domains, which they measure, and genes, which they seek to enumerate.

In addition, many of their samples appear to be of poor DNA quality, meaning that modeling estimates of domain copy number for the population can appear erroneous, even perhaps off by as much as a factor of 10 for the DP strain. This may contribute to some curious 'artifact' in some of the figures, such as Figs S2B1, S3B, S4C2, particularly for the B30.2 sequences in DP. The authors mention 'low sequencing depths' and 'low coverage' as contributing factors for this 'artifact' but also invoke possible 'evolutionary factors' in the discussion. Additional discussion for the apparent experimentally-induced effects on underestimating B30.2 sequences (perhaps due to increased DNA fragmentation and smaller domain size?), rather than coincidence of low coverage and actual biological strain-specific loss of B30.2 domains, appears warranted.<br /> - Are prior studies referenced appropriately?

yes<br /> - Are the text and figures clear and accurate?

yes<br /> - Do you have suggestions that would help the authors improve the presentation of their data and conclusions?

Statement in the abstract about fewer (not 'less') NLRs in lab strains vs wild: consider revising to average # NLRs per individual.

Revise Fig 3C legend: 'Totally discovered NLR genes', e.g., to 'total # NLR genes'.

Significance

• Describe the nature and significance of the advance (e.g. conceptual, technical, clinical) for the field.

This study elucidates the pan-NLRome of zebrafish via application of a somewhat new technique.<br /> - Place the work in the context of the existing literature (provide references, where appropriate).

The authors apply a targeted capture and long read technology previously used in plants (Weyer 2019) to fish, thereby elucidating the pan-NLRome of zebrafish, a vertebrate with a large # of NLR genes. Previous studies has shown NLR gene variation in zebrafish, but had not compared levels within and between strains, or estimated a total NLR # across populations.<br /> - State what audience might be interested in and influenced by the reported findings.

Those interested in immune function, genetic diversity and genome evolution.<br /> - Define your field of expertise with a few keywords to help the authors contextualize your point of view. Indicate if there are any parts of the paper that you do not have sufficient expertise to evaluate.

Immune genes and evolution. The specific statistical approach

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

Summary:

The authors have examined the variation of the immune related gene family "Nucleotide-binding domain Leucine-rich Repeat containing" (NLR) in the zebrafish Danio rerio. These proteins, while highly divergent, are conserved across animal and plant species. In humans or rodents, the number of these genes is relatively small (20-30) but other species such as Arabidopsis appear to have >10,000 NLR genes.<br /> Schäfer et al. collected 67 wild-type zebrafish from 4 independent sites near the Bay of Bengal and added 8 fish each from the laboratory strains Tübingen (TU) and Cologne (CGN). Using DNA capture and sequencing technology, they identified 1,560 unique "FISNA-NACHT" domains representing different NLR genes and 574 NLR-associated B30.2 domains. A subset, 714 and 229 respectively, were identified in all of the fish populations representing a "core" set common NLR genes. A significant subset of the genes could not be aligned to the GRCz11 reference suggesting significant variation across subpopulations of zebrafish.

Major comments:

None. The work is straight-forward, carefully done and conservatively interpreted.

Minor comments:

GRCz11 is still a very fragmented assembly, particularly chr4 where heterochromatin repeats on the long arm were intractable to short read sequences. There is a more recent assembly generated by the Tree of Life initiative (GCA_944039275.1) from the SAT fish line that may allow a more robust alignment and placement of the NLR genes, perhaps even some phasing of unassembled fragments.

Significance

General assessment:

Obtaining more data on the inherent variation within species of NLR genes as well as collecting more cross-species data for evolutionary comparisons is valuable for our understanding of this interesting but poorly understood class of immune response genes as well as the dynamics of gene duplication/deletion in complex, repeated arrays.

• The major strength of the study is the efforts put into capturing wild-type zebrafish from multiple different locations to maximize the diversity of the NLR sequences.
• The primary limitation is targeted capture is always contingent on having enough homology in the probes to capture all the desired genes in roughly even proportions. Some of the more interesting NLR genes might have diverged too much to be properly captured but could provide critical information on evolutionary functional adaptation. Similarly, while based on cost, it is understandable why sequence capture was chosen, but phasing information across clusters could help explain (or discover) very diverse haplotype sequences in the introns of many of these gene arrays. As sequencing costs drop, this is an important aspect to the evolution of zebrafish chromosome 4 (as well as the rest of the genome). "Pangenome" differences in the zebrafish genome might be quite radically different.

Advance:

The study represents an important characterization of a poorly understood but important class of immune genes. A very similar (but more detailed) characterization of Arabidopsis NLR genes by Van de Weyer et al., Cell 2019, is considered an important advance to the plant biology community and has been highly cited since its publication. The collected data is potentially quite useful for future studies that want to understand the complex dynamics of highly repetitive, yet functional regions of the genome and how it can result in the creation and extinction of new genes.

Audience:

There are two main likely audiences, those interested in how NLR genes are involved immune protection and those interested in genomes and evolution. It probably doesn't reveal a fundamental biological principle that would be of broad, general interest to the research community.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

Summary:

Study by Schäfer et al. investigates diversity of NLR family genes in zebrafish (Danio rerio). Genes belonging to this vast family are implicated in various defense/immune functions across studied domains of life, and form extensive families of often highly duplicated genes. Here, repertoire of NRL genes was studied in 93 individuals, coming from both laboratory strains and natural populations. Authors found a relevant, intra-specific copy number variation in the NLR genes in zebrafish, and a higher diversity of these genes in wild animals, compared to their laboratory counterparts.

General comments:

1. Odd statement in abstract comparing plants to fishes - why not put in context of animals in general, or vertebrates in particular? Same in the introduction and discussion. Either a more broad comparative framework should be set - from plants, through invertebrates to vertebrates, or it should be shortened and perhaps discussed in the light of diversity present in e.g., fish or just vertebrates. A persistent focus on plants is peculiar and does not seem to provide adequate background/context.
2. Methods seem overall well described, and the analyses seem to have been performed diligently-although I am not an expert in the field. One aspect that would be important to evaluate methodology is the repeatability of NLR detection - If we perform the procedure twice on DNA sample from a single individual, how repeatable collection do we get?
3. Discussion is quite speculative, and some claims seem exaggerated; in conclusions, eg: "This study advances our understanding of the evolutionary dynamics affecting very large gene families." - as the study is mostly descriptive documentation od CNV of gene (or, gene fragments) I am not convinces how it really advances "understanding of the evolutionary dynamics".

Minor points:

At this stage the journal and so the audience is unknown, so perhaps this will not be an issue - but for a broader audience, a better explanation of what PRY/SPRY/B30.2 are would be useful.

English could be smoothed, certain sentences sound odd: eg. lines 87-89; or 90-93 - Studies have shown that viral and bacterial infections can induce the expression of specific fish NLRs (reviewed in (24)). Some have PYD or CARD domains and can even form inflammasomes similar to mammalian NLRs (25, 26). - could be read as if infections had PYD or CARD domains, not NLRs. Lines 239-240 - not sure what does "presence/absence variation" mean here.

Sequence of baits used should be provided in some supplement or repository.

Significance

The article tackles an overall interesting subject: an expansion of a relevant group of genes in one of the major model species of biomedical importance. The main strength of the study is putting in context variation found in the lab strains - via comparison to wild populations. A limited and "haphazard" genetic variation (especially of genes involved in immune processes) of laboratory models can have paramount implications for the interpretation of experimental studies.

That being said, expansion of NLR family in teleosts in general, and in zebrafish in particular, has been previously described (eg, https://bmcecolevol.biomedcentral.com/articles/10.1186/1471-2148-8-42), and this study mainly expands description of this phenomena. Given the methodology, little can be said about possible functions of the discovered diversity. In particular, a technical aspect imposing limitation of sequence length resulted in a collection of exons, but not full genes - precluding e.g., deepened analysis of domain architecture. Selection analysis is also quite limited in scope; there are several, more sophisticated tools to infer balancing or purifying selection, either pervasive or episodic (see eg. excellent tools of https://www.datamonkey.org/). Nonetheless, the study certainly highlights the possible importance of such an expanded group of genes in this species, with a potential to inspire further - more mechanistic/functional research in this area.

The article will likely interest a few groups of rather specialized audience - e.g., those working in NLR genes in particular, or in immunity of zebrafish (or fish in general). Another, somewhat broader angle, able to attract a wider audience would be subjects of general evolution of multigenic families (birth-and-death models, trade-offs at different CNV etc.) and comparative analysis to other groups of animals - or genes - evolving in a similar way.

My field of expertise is diversity and evolution of adaptive branch of the immune system in vertebrates, in particular non-model vertebrates.

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

My response to the reviewers appears in the uploaded "Revision Plan" PDF file

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #3

Evidence, reproducibility and clarity

This study by Chourasia et al determined the effects of MTCH2 deletion on the total metabolite content (polar and lipid metabolites) per total protein of HeLa cells, analyzed at different consecutive times after adding fresh and complete media (high glucose, 10%FBS). They also analyzed the effects of nutrient depletion (no FBS In addition, authors assessed the effects of MTCH2 deletion on mitochondrial morphology, lipid droplet number and size in HeLa cells, as well as in the differentiation of mouse fibroblasts to adipocytes. From the metabolite snaphots under htese differents times and conditions, authors conclude that MTCH2 deletion increases mitochondrial oxidative function to induce a catabolic state, which impedes lipid synthesis and, as a result, adipocyte differentiation. The major concerns are that it is unclear whether the metabolic phenotype observed is a consequence of MTCH2 deletion inducing a decrease in proliferation of HeLa, as well as of fibroblasts that need to reach confluence to differentiate. In this regard, it is also unclear whether MTCH2 deletion increases ATP demand and/or promotes a catabolic program, as metabolic flux analyses are missing. A minor concern is the use of computer (processing system), antenna and wifi analogies to describe the role of MTCH2 in mitochondrial function, which is confusing.

Significance

This study represents a thorough characterization of the metabolite content in proliferating HeLa cells in the absence of MTCH2 expression. The changes observed in polar and lipidic metabolites are novel, interesting and contribute to our understanding on the role of MTCH2 function in cellular metabolism. The main limitation of the study is that the levels of most metabolites are normalized by protein content, comparing conditions in which cell number and protein synthesis have changed. Thus, it is unclear whether some of the effects observed are a consequence of the reported role of MTCH2 supporting the proliferation of different tumors and cell lines, or whether it is a direct effect of MTCH2 increasing ATP demand and/or being a direct activator of mitochondrial catabolism. Related to this point, it is unclear whether the defect in adipocyte differentiation induced by MTCH2 KO in NIH3T3 fibroblasts might be caused by an inability of MTCH2 KO to reach confluency at day 0, needed for differentiation.Finally, respirometry and mitochondrial ROS content analyses would be needed to confirm that the changes in the metabolite levels induced by MTCH2 are caused by an increase in mitochondrial oxidation leading to nutrient depletion, as authors conclude. For example, an increase in the ADP/ATP ratio could also be caused by an inhibition of mitochondrial ATP synthase in the mitochondria, concurrent to an increase in ROS production, which would decrease NADH and NADPH content.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

Mitochondrial carrier homolog 2 (MTCH2, SLC25A50) loss induces alterations in mitochondrial dynamics and energy utilization. However, the molecular mechanisms underlying these changes are still unknown. The study employs temporal metabolomic and lipidomic analyses, uncovering heightened catabolism, increased lipid storage, and disrupted adipogenesis in MTCH2 KO cells. The manuscript provides a comprehensive metabolic profile, revealing ATP demand increase, oxidized cellular environment, and adaptive changes in MTCH2 KO cells. Notably, in line with the fundamental role of fatty acid biosynthesis and anabolism in adipogenesis, the authors demonstrates that MTCH2 loss inhibits adipocyte differentiation. This work offers novel insights into the broader metabolic consequences of MTCH2 depletion.

Major comments:

• The key conclusions of the paper align with the conducted experiments, but a few additional experiments are necessary to state some claims and provide more robust conclusions.

• The paper could benefit from the inclusion of specific experiments, particularly those that address the following aspects:

• Validate MTCH2 ablation in HeLa and NIH3T3L1 through sequencing of clonal lines, Western blot analysis to confirm the absence of the protein, and real-time PCR to assess whether the mechanism involves mRNA decay.

• Provide a more detailed rationale for their temporal metabolomics approach, elucidating the choice of the media and the timepoints of cell collection. The method involves an initial culture of the cells in DMEM medium, followed by a switch to complete medium (CM) for overnight cell growth, and subsequent refreshment with CM for different timepoints before the metabolomics analyses. Authors should articulate the reasoning for opting for CM. Furthermore, authors should explicitly explain the rationale behind selecting specific timepoints for cell collection after the addition of the fresh medium.
• In Figure 1, authors conclude that MTCH2 ablation stimulates oxidative metabolism and ATP production to fulfill increased cellular ATP demands. However, this conclusion is based only on metabolomic analyses of the ADP/ATP ratio. To comprehensively assess the impact on cellular respiration, the authors should monitor the Oxygen Consumption Rate (OCR) and report the Respiratory Control Ratio (RCR).
• NAD+/NADH ratio: authors should measure NADH levels in both mitochondria and cytosol. This can be accomplished through NADH autofluorescence (recommended) or commercially available kits. This additional analysis would contribute to a more comprehensive interpretation of the observed changes in oxidative metabolism. They should also include measurement of mitochondrial membrane potential using TMRM. Suggested experiment: measuring NADH autofluorescence. The autofluorescence of mitochondrial NADH can be distinguished from cytosolic NADH by optimizing substrate consumption followed by the complete inhibition of electron feeding to the ETC. The redox state of NADH reflects the equilibrium between mitochondrial ETC activity and the rate of substrate supply. After acquiring basal autofluorescence levels through live imaging, max signal is obtained by stimulating maximal respiration (FCCP), and min signal is obtained by inhibiting respiration (NaCN or Rot+AA). Subsequently, "NADH redox indexes" are generated by expressing the basal NADH levels as a percentage of the difference between the oxidized and reduced signals. Furthermore, by examining the fluorescence signal increase after NaCN addition, the rate of NADH production can be monitored. This rate serves as a proxy of TCA efficiency.
• Authors observe a reduction in the levels of various amino acids and TCA cycle intermediates, indicative of an increased flux through the TCA cycle. This proposition could be further supported by measuring the kinetics of NADH autofluorescence. Additionally, a decrease in metabolites associated with the urea cycle, such as citrulline and ornithine, is observed, yet this observation remains uncommented and warrants discussion. Intriguingly, an elevation in Branched-Chain Amino Acids (BCAAs) and unsaturated acyl carnitines is noted, leading to the hypothesis of an increased transport and breakdown of fatty acids in the mitochondria to meet the heightened cellular demand for ATP in MTCH2 KO cells. To substantiate this, and to quantitatively measure mitochondrial fuel utilization in live cells, authors shall perform a Mitofuel Flex Test by measuring the Oxygen Consumption Rate (OCR) in cells treated with inhibitors of each mitochondrial oxidative pathway including etomoxir. This approach would enable the measurement of the dependency, capacity, and flexibility of cells concerning the pathway of interest in meeting ATP demand. It is also recommended to perform MitoStress test in cells supplemented with only one of the carbon sources (such as Glucose, Glutamine, Long chain and Short Chain Fatty acids).
• In Fig 3, a reduction in membrane lipids, free fatty acids, and non-esterified fatty acids is observed, while there is an increase in esterified fatty acids, storage lipids like Triacylglycerols (TAG) and Cholesterol Esters (CE), and lipid droplet number and size. Notably, these lipid droplets are positioned closer to mitochondria in MKO cells. The authors propose that MKO results in enhanced transfer and metabolism of lipid moieties at the mitochondria to generate ATP. To provide insights into the molecular mechanisms underlying the observed lipid changes in MTCH2 KO cells, the following experiments are recommended: Employ Western blot and real-time PCR to measure the levels of enzymes crucial in TAG and CE formation and accumulation (e.g., Long-chain acyl-CoA synthetase (Acsl), Stearoyl-CoA desaturase (SCD) or others). Evaluate the enzymatic activity of these identified enzymes to understand their functional role in lipid metabolism in MTCH2 KO cells.
• The suggested experiments are realistic in terms of time and resources, ensuring practical feasibility.
• The data and methods are presented in a clear and reproducible manner.
• The experiments appear adequately replicated, and the statistical analysis seems OK.

Minor comments:

• There are no specific experimental issues that require addressing.
• Prior studies are appropriately referenced
• In general, both the text and figures are clear and accurate. The significant alteration of metabolites found in their metabolomic dataset should be plotted using the online tool MetaboAnalyst to analyze metabolic pathways and generate better visualizations.
• Overall, the presentation is satisfactory with only minor language adjustments recommended. A minor suggestion for improvement involves refining the language used in the text. Instead of consistently using the term "produce energy," please use "conversion of energy".

Significance

General Assessment: The study, through the integration of metabolomic and lipidomic data in MTCH2 KO cells, provide a comprehensive overview of the metabolic rewiring of these cells. This metabolic change is particularly interesting in the context of adipogenesis, offering valuable insights into the interconnectedness of a mitochondrial solute carrier, cellular metabolism and adipogenesis.

Comparison and Advance: The current study significantly advances our understanding of the mitochondrial carrier homolog 2 (MTCH2) by uncovering its intricate roles in metabolism, and adipogenesis. While prior research identified MTCH2 as a regulator of apoptosis and mitochondrial dynamics, the present study expands our knowledge by elucidating its involvement in cellular metabolism and adipocyte differentiation. The major advance lies in the detailed exploration of MTCH2's impact on cellular metabolism through temporal metabolomic and lipidomic analyses. The study reveals that MTCH2 deletion leads to heightened ATP demand, an oxidized cellular environment, and alterations in lipid, amino acid, and carbohydrate metabolism. Additionally, the adaptive response in MTCH2 knockout cells involves a strategic decrease in membrane lipids and an increase in storage lipids. Furthermore, the study unveils a novel connection between the imbalance in energy metabolism triggered by MTCH2 deletion and the inhibition of adipocyte differentiation-a process that demands substantial energy and reductive biosynthetic activities. This mechanistic insight provides a conceptual advance, indicating how MTCH2, beyond its known role in apoptosis and mitochondrial dynamics, plays a pivotal role in orchestrating cellular metabolism and adipogenesis. Importantly, this work aligns with prior observations that hinted at MTCH2's involvement in fatty acid synthesis, storage, and use through his identified interactome. In summary, the study advances our knowledge of MTCH2 by providing a more comprehensive understanding of its roles in cellular metabolism and adipocyte differentiation, shedding new light on its multifaceted functions beyond its originally identified roles.

Audience: This research will appeal to a broad audience, ranging from specialists in cellular metabolism to those with a general interest in mitochondrial dynamics and biochemistry.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

In the present study, Chourasia et al. describe the effects of MTCH2 deficiency on various metabolic parameters. Using temporal metabolomics in HeLa cells, they show an increase in ATP demand in cells lacking MTCH2. They also show altered lipid metabolism in NIH3T3L1 preadipocytes lacking MTCH2, associated with impaired maturation.

This study is mainly descriptive, containing large amount of information that would be of interest to the understanding of the role of MTCH2 in cell metabolism. The manuscript would benefit from thorough editing to make it more focused and accurate. This is challenging since there is a very large amount of data. Below are some suggestions, along with a couple of experiments that I suggest to add in order to clarify the mechanism through which MTCH2 acts.

Major Comments:

1. The ratio ADP/ATP as well as AMP/ATP as well as the decrease in TCA metabolites are indications of ATP demand. Nevertheless, these are not fluxes. It would therefore be worthy to complement these results with respirometry.
2. The dramatic increase in AMP/ATP ratio suggests an increase in AMPK activity. Testing it (by measuring AMPK phosphorylation or ACC phosphorylation) could further strengthen the results but it is not a must-have.
3. The cause of the increase in AMP in the NIH3T3L1 is not addressed. The involvement of acyl-CoA synthetase is worth discussing or investigating.
4. Fig. 2F shows MKO and WT, but it doesn't show MKO-R. This is an important control since lactate doesn't seem to be affected in the MKO-R.

Minor comments:

In the intro there are a few instances that could benefit from some more accuracy:

1. In the abstract there is no mention of the cells that are used.
2. Line 56: "...(OXPHOS) converts nutrients into adenosine triphosphate (ATP)". It would be more accurate to write that OXPHOS converts the chemical energy that is stored in nutrients into ATP.
3. Line 58: "The mitochondrial NAD+/NADH pool are substrates for OXPHOS". It would be more accurate to write that NADH is the substrate. (NAD+ is after all the product of oxidation)
4. Line 59-60: "Along with ADP, NAD+ also plays an important role in the regulation of the Krebs cycle". Instead of "regulation" I think it is more specific to write "stimulates" (otherwise add ATP and NADH which are also involved in regulation. 5. Line 65: "...changing metabolic states. In addition, mitochondria..." A link between the two sentences seems to be missing.
5. Figure 1 is a heavy figure. Some results are significant and some show only a tendency. The description (Lines 119-125) is too general. It addresses only the "trends". A bit more specificity as to the metabolites or ratios of metabolites and the time points that are significant would be in place.
6. Lines 136-139 "The metabolomics analyses revealed additional important changes in many more nutrient substrates, which included a decrease in most amino acids (Fig. 2A and Fig. S2A). Notably, the most significant change was seen in glutamine (Fig. 2A, left top graph), one of the major amino acid-nutrient sources" Glutamine is indeed an important amino acid and the effect is strong, but in this case it's increased. From the first sentence one would think it's decreased. Again, be more specific in your sentencing.
7. Lines 157-159: "Thus, the acyl carnitine profile suggests that 1- to 12-hrs post media change the MKO cells use BCAAs as a nutrient source, and later shift to unsaturated acyl carnitines, specifically to the C16:1 and C18:1 forms". This conclusion does not derive from the description by the result - address the difference between the different carnitines that occurs at different times.
8. Lines 167-168: "These results suggest that there is higher metabolism of acetyl CoA in the MKO cells leading to a bell-shape dynamics (low-high-low levels)." The interpretation is unclear. I understand that you mean that there is fluctuation within the group, however the acetyl-CoA levels remain lower MKO than in the control at all time points. This further suggests a decrease in TCA cycle.
9. Lines 172-173 addresses the bell shape of lactate, yet the most prominent result is the 3-fold increase in lactate levels compared to WT which is not mentioned. Unfortunately, the MKO-R shows a similar increase. Still, this should be addressed in the text.
10. 182-184: "Taken together, the results presented above are consistent with the idea that the increased amino acid/lipid/carbohydrate metabolism and substantial decrease of many metabolites in MKO cells is most likely due to their increased utilization to meet themincreased cellular energy demand." I'm not sure how this conclusion is reached. From metabolite levels alone, it's difficult to conclude about fluxes. Also, in a previous conclusion the authors wrote that the TCA cycle is probably reduced; this contradicts the above conclusion. If the authors mean that the FFAs and amino acids are used for anabolism and glucose for meeting energy demand, they should state so more clearly.
11. Figure 3A can be split into 2 or 3 subfigures. I think it would make it more comprehensible.
12. Line 210-211: "Notably, we also found that MTCH2 knockout cells showed accelerated mitochondria elongation (Fig. S3D, top panels), which was further pronounced when cells were grown in HBSS". The increase in mitochondrial elongation comes after fragmentation in MTCH2 KO. Although this is a known phenotype, it is good to address it shortly in the text.
13. Line 208-209: "LDs from dispersed to a highly clustered distribution that was often observed in close proximity to mitochondria..." The proximity of mitochondria to LD suggests the possibility that there is an increase in peridroplet-mitochondria, which have been shown to be involved in biogenesis LD. It might be interesting to investigate this path as an explanation to the observed phenotype.
14. Lines 244-245: "These results suggest that the MTCH2 knockout preadipocytes face a cellular energy crisis that is similar to the one seen in the MTCH2 knockout HeLa cells presented earlier" It's true that NAD+ and AMP (as well as AMP/ATP ratio) are increased but in view of the high ATP and NADH, it's difficult reach the conclusion that there's an energy crisis.
15. In the discussion- Lines 267-269: "Thus, MTCH2 might act like a "relay station" by sensing and connecting between metabolic intermediates/pathways and dynamic changes in mitochondria morphology/energy production by receiving and sending Wi-Fi signals." It's difficult to raise such specific hypothesis from the results. Use milder terms.

Significance

Great study

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

The authors do not wish to provide a response at this time.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

Summary

This study highlights the role of the septin cytoskeleton in plasma membrane repair in HeLa cells perforated by the pore-forming toxin listeriolysin O (LLO). The authors performed a silencing RNA screen targeting protein-coding genes involved in endocytosis, exocytosis and intracellular trafficking. Besides the recovery of proteins that were previously identified to be part of the membrane repair machinery, they uncovered novel plasma membrane repair candidates, including septin 7 (SEPT7).

They found that upon LLO treatment, septins redistribute from actin stress fibers to the cell surface where they form knobs and loops together with F-actin, Myosin-IIA and Annexin A2 (ANXA2). Using super resolution microscopy and 3D reconstruction, they showed that these structures often protruding from the cell surface are formed by septins and F-actin that are organized in intertwined filaments associated with Annexin A2, and that they are functionally correlated with plasma membrane repair efficiency. Silencing SEPT7 further revealed that the remodeling of the repair protein ANXA2 at the cell surface is greatly decreased in LLO-injured cells, whereas the down regulation of ANXA2 had no impact on the arrangement of septins and F-actin into knobs and loops. Altogether, their results evidenced that the septin cytoskeleton triggers the organization of membrane domains containing the actomyosin cytoskeleton and ANXA2, that are essential for the repair to occur.

Major comments:

• The authors show that silencing SEPT6 or SEPT7, but not SEPT2 or SEPT9, perturbed plasma membrane repair of LLO-injured cells. The authors explain this result by indicating that the reduced expression of SEPT7 and SEPT6 (according to the siRNA), but not that of SEPT2 results in a reduced expression of septins from other groups. This could have been an explanation but, in Fig. S2B, downregulating SEPT2 clearly seem to impact the expression of SEPT6 and SEPT7 (except for siRNA#3) once normalized with the loading control tubulin. Moreover, it is well accepted in the literature and has been observed in many cell types, including HeLa cells, that knocking down a septin from one group (with sometimes the exception of septins of Group 3) induces the downregulation of septins of the other groups, and that it consistently results in the loss of septin filaments. Therefore, the fact that silencing SEPT2 does not perturb plasma membrane repair is quite surprising. This could suggest that SEPT6 and SEPT7, independently of their filament organization, play a role in membrane repair after LLO treatment. Nevertheless, the SEPT2 staining to study the fate of septin filaments following LLO exposure indicated that it is the redistribution of septin filaments that is crucial in this repair process. Interestingly, BORG proteins which are involved in the association of septin filaments to the actin cytoskeleton in interphase cells bind to the SEPT6/SEPT7 coiled-coil region of septin polymers. Could these proteins be involved, knowing that they are Cdc42 effector proteins, and that links exist between Cdc42 activation and Ca2+ entry? OPTIONAL: silencing BORG proteins (BORG2 for example) and studying septin and F-actin remodeling following LLO exposure could help the authors to understand the reason of such a redistribution.
• What about the terms "knob" and "loop": Are they structurally related to the "specks" described in other papers? Or are they new structures that no one observed before? Nobody has never looked at septins in this repair process before, but actin has long been described to be involved.
• It seems that knobs are formed before loops take over. This would deserve further investigation. Is that a reality? Or are they two independent structures? OPTIONAL: it would be interesting to do time-lapse video microscopy to follow the fate of a knob. Related to the previous point: why to show 3 sets of images in the LLO condition in Fig. 2A? Does the top b-panel represent the knob stage? Where there are still many stress fibers indicating that septins have not yet fully redistributed? And when septins are fully dissociated from actin cables, which are then lost, loops are forming (middle c-panel) and then increase in size (bottom d-panel)?
• Even though, some information is given in the discussion section, it would be helpful to mention in the introduction section the different pathways that cells activate to repair plasma membrane defects, and to precise which one(s) has(ve) already been described in the literature to be switched on in response to the LLO toxin.
• Some experiments are not rigorous enough: Sometimes, they have not been repeated, as exemplified in Fig. 7B. Count less cells but repeat the experiment at least three times. Sometimes, one condition is missing, as in Fig. 6C. Where is the DMSO condition? What about the statistics? Fig. 6A: In the calcium free condition, it seems that the two cells that are illustrated depict a telophase. The subcellular organization is obviously different at the end of cell division. Show only one interphase cell as in the top panels. Fig. 6B and C: It is mentioned in the figure legend: "Cells were treated as indicated in (A) and (D)". But the "C" condition is not mentioned anywhere: Does "C" stand for no DMSO or FCF treatment, in the presence of calcium, but under LLO treatment? Likewise, it would be very helpful to indicate in each figure panel whether cells have been treated with LLO or FCF. Please help the reader. Fig. 7A: Whatever SEPT7 is expressed or downregulated, the actin stress fibers are still present. If these cells were not well transfected, replace the images.
• Concerning the FCF experiments: The FCF cytokinin has been used by many authors to perturb septin dynamics. It induces the stabilization of septin polymers, thus promoting the formation of thick ectopic fibers. It is a potent inducer of septin polymerization and acts as a stabilizer. Fig. 6D: In the Ctr condition, a DMSO condition is needed to visualize the impact of FCF on septin filaments. Does FCF stabilize septins and induce the formation of thick filaments? The SEPT2 image in the FCF condition without LLO is of bad quality (see above remark). Also in Fig. 6D (FCF condition without LLO), the F-actin staining revealed that there are no stress fibers!!??? Usually, the more septins are associated with actin, the thicker stress fibers you get, since septins stabilize actin cables. FCF treatment often induces thick ectopic septin filaments that are not associated with stress fibers (which are therefore lost). Was it the case in all FCF-treated cells? Does FCF treatment really mimic what happens physiologically in the cell? Many off-target effects have been observed with this molecule in non-plant cells.
• The image quality in Figs 3A and B, and 6A and D needs to be improved regarding the septin staining. In control conditions, septin filaments cannot be clearly distinguished.
• Fig. 3B: It seems that ANXA2 is overexpressed in LLO-injured cells. Its accumulation level between both conditions should be compared by immunoblot. ANXA2 is indeed recovered on loops, but it is difficult to consider whether it is a redistribution.
• Fig. 7D: Compared to the control condition (we have to refer to Fig. S5D), ANXA2 again seems to be overexpressed under LLO treatment. To affirm that ANXA2 remodeling in LLO-injured cells requires the formation of septin/F-actin knobs and loops, data in Fig. 7D must be quantified.
• Fig. 6 (B-D): In panel B, there is a significant difference between the "C" and "FCF" conditions regarding the number of knobs + loops per cell. Where are the images corresponding to the "C" condition?

Minor comments:

• Fig. 2A: Report the white squares (selected enlarged areas) in all panels (SEPT2 and overlay). In panels b, do not place an arrowhead where we are supposed to observe an enlarged area. Also, from panels b, it would be worth showing an enlarged area including a knob. Show enlarged areas also from panels d.
• Fig. 3B'i: Septins are not on stress fibers. Select a transfected cell where septins still coalign with actin fibers, not a cell that was impaired by the transfection.
• Fig. 3C: Add the time point "0min". What was the % of colocalization before LLO treatment? and in DMSO condition? What about ALIX at 5 and 10min? Again, it's only one experiment.
• Figs 4 and 5: Very nice images but obtained following FCF exposure. Hopefully FCF would not have induced an aberrant organization!
• The "Ctr" abbreviation is often used, in different conditions, and may be confusing. Precise in the figure (not in the figure legend) whether it is siRNA ("Ctr siRNA"). Mention "DMSO" for the controls of your drugs (like in Figs S4 and S5).
• Fig. S4C: How is this figure different or does it provide additional information compared to Fig. 2C?
• Fig. S5D: It is hard to know that the ANXA2 siRNA worked, since no difference of staining between the Ctr and the transfected cells can be observed. Were these cells really transfected? It would have been helpful to use fluorescent siRNAs. The same applies to Fig. S5C: Silencing SEPT7 supposedly greatly reduces the level of expression of all septins. The SEPT2 staining is still high, and many actin stress fibers are still observable (whereas the loss of septin filaments results in the loss of actin stress fibers, as observed by many authors, including in HeLa cells). Same remark for Fig. S6, regarding the SEPT7 silencing in the Ctr condition (no LLO). No impact on stress fibers! Are these cells transfected? The authors themselves mention that sometimes cells are less effectively silenced (like in Fig. 7A, B). Why not to show cells effectively silenced!!
• In the abstract, it is specified that SEPT7 also plays a role in membrane repair after mechanical wounding. Based only on one type of experiment (SEPT7 silencing, Fig. 1H), this statement should only be mentioned in the text or used to discuss the putative repair mechanisms that septins are involved in, but not stated in the abstract as a main conclusion.

Significance

Strengths:

Despite septins have been involved in endocytosis, exocytosis, membrane protrusions, cell junction integrity or actomyosin constriction at cytokinesis, the involvement of the septin cytoskeleton in the plasma membrane repair machinery has, to my knowledge, never been reported before. The authors not only showed that septins are present in specific membrane protrusions (knobs and loops) but also evidenced that septin filaments trigger the formation of these plasma membrane repair domains by recruiting F-actin and ANXA2, essential for the repair to occur. The novelty of this study has therefore to be acknowledged, and these data will benefit the scientific community, and the septin community in particular.

This is a descriptive paper that nevertheless clearly shows, by different means, the reorganization of the septin cytoskeleton in LLO-injured cells. The use of high-resolution microscopy coupled to 3D reconstruction which enables to easily appreciate the organization of septins, F-actin and ANXA2 in the knobs and loops is a true strength of the paper.

Limitations:

The authors mention in the abstract that septins act as scaffolds to recruit contractile actin fibers and ANXA2. Biochemical experiments such as co-immunoprecipitations could strengthen this notion. The molecular mechanism by which septins are involved in this repair process has not been addressed at all in the paper. Even though the silencing RNA screen highlighted several proteins involved in known membrane repair mechanisms, the authors just presented a few data concerning ALIX, a component of the ESCRT-III machinery. A % of colocalization of SEPT2 and SEPT7 with ALIX is reported in Fig. 3C but this experiment has only been done once (n=1) and only following 15-min exposure to LLO. Is that too late? Immunofluorescence images of SEPT2 and ALIX with or without LLO (15min) are also provided in Fig. S5A but no quantification is reported. Is it sufficient to say that the ESCRT machinery is not involved?

My field of expertise:

Cytoskeleton, Septin, Actin, Microtubule, Signaling pathways

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

Review of "The septin cytoskeleton is required for plasma membrane repair" by Prislusky et al.

Eukaryotic cells rapidly repair damage to their plasma membrane and underlying cortical cytoskeleton. Such repair is increasingly recognized as being of major importance to human health (PMID: 33849525). Two broadly conserved cell damage responses have been described: a very rapid membrane resealing response which commences within a second or so following damage, and a cortical cytoskeletal response which commences within ~15-30s and which is based on activation of the Rho GTPases. However, our understanding of either of these responses is extremely limited, a situation which has engendered considerable debate about not only the mechanistic bases of these responses but also their relative roles and the extent to which they may be interdependent.

In the current study, the authors use an siRNA screen to identify Septin7 (hereinafter SEPT7) as a critical participant in the cell repair response. They further demonstrate that cell damage, as induced either by bacterial pore-forming proteins or by mechanical abrasion results in accumumulation of septins (including SEPT7) in curious ring-like structures associated at the plasma membranes; these structures are often associated with plasma membrane protrusions, which are a common feature of damaged cells. Additionally, the authors show that the septins colocalize with F-actin, myosin-2 (an F-actin-based motor protein) and annexin-2A, a protein previously implicated in cell repair. Lastly, the authors show that depletion of septins reduces the recruitment of annexin-2A to the plasma membrane in wounded cells, implying that the septins are upstream of the annexin in the wound response.

This is an exciting study that is also very well-documented. The excitement is provided by the following observations: first, septins have not previously been implicated in cell repair; second, the association of the septins with F-actin and myosin-2 in ring-like structures at the plasma membrane is suggestive of the possibility that local contraction may promote healing, a long-standing idea derived from studies of frog oocyte healing (PMID: 10359696; PMID: 11502762) which has proven controversial for healing of other cell types (see below); third, a link between septins and annexins in cell repair or, for that matter, any other process, is novel. With respect to the support for their claims, the authors go above and beyond to make their case-every point is supported by multiple approaches-for example the importance of septins is shown via siRNA, shRNA, and inducible depletion-and the imaging is very, very nice.

The potential role for actomyosin-powered contraction in healing of wounds made in cultured mammalian cells has been largely discounted because of studies wherein cells are wounded after pharmacological treatment with actin poisons have shown that healing is actually improved. The problem with such studies, is that depolymerization of actin prior to cell damage will dramatically alter the response to damage due to loss of cortical tension (PMID: 19846787). Thus, besides being important in its own right, the current study opens new doors for experimental assessment of the possible roles for cortical actomyosin in cell repair.

I have only minor concerns or questions:

1. What is the spatial relationship between the septin rings and actual damage sites? This could be addressed by wounding in the presence of a lysine fixable dextran.
2. The information in table 1 could be made more reader-friendly. In particular, it is not clear how the authors are getting their gene/protein names for their hits and what they correspond to. This was most noticeable for IQSEC1, ABI1, and GBF1 which the authors describe in the text as "genes that control the actin cytoskeleton" but in the table are listed as "Signaling proteins". I may have the abbreviations wrong (which is more reason for additional clarity) but GBF1 is the abbreviation for a protein involved in intracellular trafficking; IQSEC1 is a GEF for Arf proteins, and ABI1 is best known as a subunit of the WAVE complex.
3. The statement that begins the abstract "Mammalian cells are frequently exposted to mechanical and biochemical stresses..." could just as easily be "Eukaryotic cells..." or even "Cells..." as the membrane repair response is apparently universal and, indeed, was first described in nonmammalian cells. Similarly, the introduction begins "The plasma membrane of mammalian cells forms a biophysical barrier that separates the cell from its external environment". As far as I know, this is not a specific feature of mammalian plasma membranes but rather all plasma membranes. I don't know if it is the author's intention to imply their work is only relevant to mammals, but that is certainly not the case and they end up reducing the impact of their work by making it sound like cell repair is a phenomenon specific to mammalian cells.
4. The word "subplasmalemmal" is likely to be confusing for those who are not aware that plasmalemma is an antiquated term for the plasma membrane. It might be easier for the reader if the authors refer to "subdomains of the plasma membrane".

Significance

This is an exciting study that is also very well-documented. The excitement is provided by the following observations: first, septins have not previously been implicated in cell repair; second, the association of the septins with F-actin and myosin-2 in ring-like structures at the plasma membrane is suggestive of the possibility that local contraction may promote healing, a long-standing idea derived from studies of frog oocyte healing (PMID: 10359696; PMID: 11502762) which has proven controversial for healing of other cell types (see below); third, a link between septins and annexins in cell repair or, for that matter, any other process, is novel. With respect to the support for their claims, the authors go above and beyond to make their case-every point is supported by multiple approaches-for example the importance of septins is shown via siRNA, shRNA, and inducible depletion-and the imaging is very, very nice.

The potential role for actomyosin-powered contraction in healing of wounds made in cultured mammalian cells has been largely discounted because of studies wherein cells are wounded after pharmacological treatment with actin poisons have shown that healing is actually improved. The problem with such studies, is that depolymerization of actin prior to cell damage will dramatically alter the response to damage due to loss of cortical tension (PMID: 19846787). Thus, besides being important in its own right, the current study opens new doors for experimental assessment of the possible roles for cortical actomyosin in cell repair.

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

General Statements

We are grateful to the reviewers for reviewing our manuscript. In general, the reviewers agree that our method presents a new approach of Nanopore direct RNA-seq that is not limited to the standard use of only the adenylated fraction of RNAs within a sample but they have also requested more evidence about the effectiveness and usefulness of this approach. Reviewer 1 notes that NERD-seq "extends nanopore direct RNA-Sequencing to beyond the poly(A) fraction." Reviewer 2 notes that "This manuscript has the potential of having major significance to researchers in the field of transcriptomics". At the same time Reviewer 1 remarks that, "the authors need to expand more on why this is useful, and what scenarios this would be used in" and need "to demonstrate that NERD-Seq is more than an incremental improvement to existing approaches". Reviewer 2 notes that "This technique has great potential, as ONT direct RNA sequencing can be used to detect RNA modifications... There are however some issues that need to be addressed before the manuscript is suitable to publication" and agrees with Reviewer 1 that the authors need to demonstrate "that their newly sequencing techniques could indeed improve RNA detection beyond the current techniques." We are grateful to the reviewers for these comments, and we fully agree that the manuscript in its initially submitted form falls "short from providing strong arguments supporting the fidelity, accuracy and coverage of the new technique." and that more "proof of the increased accuracy or utility of the new technique" was needed.

We attach a substantially revised and expanded version of the manuscript that includes additional data needed to ensure the methodology is replicable and further supports the rationale why NERD-seq is a useful addition to the current direct RNA sequencing methodology repertoire. We now provide:

• the repetition of all performed NERD-seq runs with a new enzyme (commercially available), as the one used in our initial study (Omniamp polymerase) is not anymore commercially available

• 6 revised main figure panels based on the new sequencing runs and a new main figure (Fig.7),

• 20 new supplementary figures, and

• 1 new supplementary table (Suppl. Tables 1),

that correspond to the points raised by the reviewers. We have also revised the main text accordingly.

Please find below the reviewers 'comments and a detailed point by point response to these comments. A document with the changes from the initially submitted manuscript being highlighted is attached at the end of this response.

Point-by-point description of the revisions

Reviewer 1

*------------------------------------------------------------------------------ *

*Reviewer #1 (Evidence, reproducibility and clarity (Required)): *

• *

***Summary:** *

• *

This reviewer generally remarks that: "Saville et al detail a new method, NERD-Seq, which extends nanopore direct RNA-Sequencing to beyond the poly(A) fraction. In principle, this allows the capture of additional RNA types in a single sequencing reaction, albeit at the expense of sequencing depth. However, the authors need to expand more on why this is useful, and what scenarios this would be used in. The rationale for many of their experimental choices are not explained/contextualized appropriately. The manuscript also suffers from excessive jargon and grammatical errors."

• *

We appreciate these comments made by the reviewer. We now expand more on why and how NERD-seq is useful, and what scenarios this could be used in. In particular:

In the introduction, we now denote that a significant number of the most well described RNA modification positions are located in classes of RNAs such as tRNAs, snoRNAs and other ncRNAs like 7SK RNA. In the results section, we build on the second reviewer's suggestion that "This technique has great potential, as ONT direct RNA sequencing can be used to detect RNA modifications...", and we first show in revised Figures 3-5 that the standard approach misses the above classes of RNAs. Subsequently, in two new figures (Fig. 7 and Suppl.Fig. S23) we present an example of the ability of NERD-seq to decipher known RNA modifications in a well-studied ncRNA, 7SK, compared with the inability of the standard approach to do so. Widely studied RNAs such as 7SK have the potential and thus often serve as controls for known positions of certain RNA modifications when validating them for novel positions in other RNAs. This however means that the application of the standard approach, which is not able to detect RNAs and thus their modifications for RNAs such as 7SK, makes it difficult to use them as controls. In other words, the ability to perform genome wide studies of RNA modifications using the direct RNA-seq approach relies on the ability to be able to confirm at the same time these findings in already known control positions located in the above classes of RNAs. If these known controls such as short ncRNAs are missing, as we show that it is the case with standard RNA-seq, it is difficult to perform such genome wide epitranscriptome studies. Thus, the usefulness of our approach is not only that it can sequence certain classes of shorter RNAs beyond mRNAs, but that it can do this without affecting the ability to study mRNAs simultaneously or use targeted sequencing adapters. This is now discussed in the text. We have also tried to avoid the excessive use of jargon language and correct any grammatical errors.

*------------------------------------------------------------------------------ *

***Major comments:** *

• *

Point 1. The reviewer mentions:* "To my understanding, the primary purpose of NERD-Seq is to allow the sequencing of the non-adenylated fraction of RNAs within sample with a size filter step designed to exclude ncRNAs larger than ~200 nt. While this allow for subsequent polyadenylation of the small fraction, it remains quite likely that larger non-adenylated non-rRNAs are being missed by the protocol. It is also not clear why NERD-Seq should be considered the optimal strategy. The authors should show that they have considered/evaluated other strategies such as: *

*- Depletion/Seperation of the poly(A) fraction prior to size selection performed on non-adenylated fraction. *

- Targeted degradation of the rRNA fraction using rRNA depletion kits."

• *

We thank the reviewer for this comment. The primary purpose of NERD-seq is not only to allow the sequencing of the non-adenylated fraction of shorter RNAs but to do this while maintaining the ability to sequence longer polyadenylated RNAs. In particular, we are not sequencing only the short RNA fraction but also the longer RNA fraction of naturally polyadenylated RNAs. This includes both mRNAs and ncRNAs that are polyadenylated. We apologize that this may have not be presented clearly in our initial text, and we have now revised the respective results part (page 8, lines 13-14 and 30-31). As mentioned below, we now also provide evidence that our approach does not affect the ability to efficiently sequence mRNAs and decipher effectively their isoforms (new Suppl. Fig. S19). Only large non poly-A ncRNAs still evade our detection. Now, we discuss this limitation in the last paragraph of the discussion section (page 19, lines 12-20). We have now also evaluated other strategies such as those mentioned by the reviewer (depletion of rRNA, size selection using magnetic beads, different enzymes) and in the new Suppl. Figure S5 we show that they were suboptimal in providing sufficient reads that pass the quality thresholds in base calling compared to our NERD-seq approach.

• *

To sum up, we certainly don't claim to be able to capture longer ncRNAs, due to their lack of polyadenylated tails. We show though that while the ability to capture simultaneously all reads under the current Nanopore sequencing protocols may be unattainable, the NERD-seq methodology demonstrates a useful addition to the Nanopore sequencing repertoire due to its ability, in addition to polyadenylated transcripts, to simultaneously capture multiple classes of short ncRNAs, hitherto until now only achievable one transcript at a time with custom adaptor ligation.

Point 2. The reviewer mentions that "Related to this, the authors note that NERD-Seq is designed to allow the sequencing of both adenylated and (short) non-adenylated fractions of RNAs within a study. What is the actual value of this versus, say, simply targeting the short non-adenylated fraction directly? "

Please see our response to the general comment by this reviewer and the new figures Fig. 7 and Suppl. Fig. S23 and the completely new results section at pages 15 and 16 about examples on how NERD-seq expands the study of epitranscriptomic signatures to additional RNA classes while maintaining the ability to efficiently assess the protein coding transcriptome (new Suppl. Fig. S19).

Point 3. The reviewer notes that "The manuscript in general is written in a somewhat subversive style, seemingly focused on highlighting the 'failings' of the standard ONT protocol. This is somewhat disingenuous as these are not 'failings' per se given the design objective of the standard DRS protocol (i.e. to capture and sequence the poly(A) fraction of RNAs) works well. The authors would be better off highlighting the situations (i.e. study questions) where NERD-Seq would provide a measurable benefit over the standard DRS strategy. "

• *

We apologize if this was indicated in our initial submission, as this was not our intention. We now focus on highlighting what NERD-seq can do, and this is reflected in the changes we have made in the introduction and discussion section.

Point 4. The reviewer asks "The authors state that 1.5ug total RNA is used as input for NERD-Seq but how much input (poly-A RNA) actually goes into the short- and long-fraction parts of the DRS protocol? This is important to know. "

We have now included data in the manuscript for polyadenylation signals in the reads themselves (new Suppl. Fig S5F) and view it as a useful addition to the manuscript as it seems to show there is little difference in the relationship of length of transcript and polyadenylation tail length between the two methods.

RNA RIN values differ across samples so within different samples the portions of poly-A RNA, non poly-A RNA and former poly-A RNA that has lost the poly-A due to degradation may also vary. This however has not been shown to affect the reproducibility of the standard direct RNA-seq methodology, so it should not affect also that of NERD-seq. Nevertheless, we searched the literature for a well characterized methodology for quantifying this further and found little on poly(A) specific quantification approaches that we could use. In fact, poly(A) selection has also been shown to have variable efficiency and could thus produce inaccurate measurements that would make it difficult for us to assess explicitly the exact portions of poly-A vs non poly-A portions. Thus, we feel that in the absence of a well characterized methodology to make these quantifications, developing a new one may have been beyond the scope of this manuscript.

Point 5. The reviewer recommends: "The authors show that NERD-Seq performs well on a tissue that is generally enriched in non-coding RNA activity but without the inclusion of biology replicates or tissue samples from other sources, it is impossible to assess (1) the reproducibility/robustness of the methodology and (2) whether NERD-Seq would produce useable data from other tissues with lower non-coding RNA activity. These experiments are required. "

• *

We thank the reviewer for this suggestion and we have now included an additional biological replicate for the mouse tissue and data from another tissue in a separate organism (5-person pool source human cerebral cortex RNA).

Point 6. The reviewer notes: "Several figures are poorly explained. For instance, Does Figure 2A show data only from the short RNA fraction? If not then this suggests incredibly high levels of mRNA degradation in the 'standard' ONT fraction - far beyond what is seen in other studies. The legends for all figures should be far more specific and detailed."

• *

We have now replaced figure 2A with a different mapping strategy and a whole sample assessment of mapped reads lengths. We have also edited the figure legends to include more information.

The reviewer made also the following comments:

Minor comments:

• *

": Please refrain from using 'next-generation' in a sequencing context (abstract, introduction). Having a 'new' next-generation sequencing is confusing in regard to the old 'next generation' sequencing. "

• *

The instances mentioned have been removed as per the reviewer's suggestion.

"Introduction should acknowledge that targeted sequencing of (individual) non-adenylated RNAs is possible (i.e. there is an ONT protocol for this)."

• *

We now include mention of this in the introduction, results and discussion.

"Several figure elements not referenced in the text appropriately (e.g. red bars in Figure 2A). "

• *

We have now addressed this and thank the reviewer for making us aware of this.

"The authors mention profiling of the epitranscriptome several times in the introduction and discussion but do not include any work geared toward looking for RNA modifications. Indeed it is entirely unclear whether NERD-Seq produces the depth of sequencing (on non-adenylated RNAs) required for current RNA modification detection tools. "

• *

We have now included a new main figure (Fig. 7) and a new supplementary figure (Suppl. Fig. S23) which profile elements of the epitranscriptome and show that indeed NERD-seq produces the depth of sequencing necessary for RNA modification detection.

Reviewer 1's summary:

• *

"NERD-Seq presents a modified method for DRS of both short non-adenylated and the standard adenylated fraction of RNAs within a sample. This utility of this approach appears rather niche and it is hard to determine situations in which this approach would be generally useful versus many of the existing (Illumina-based) approaches. The burden to demonstrate that NERD-Seq is more than an incremental improvement to existing approaches lies with the authors. "

• *

We appreciate the reviewer's comments. While we agree with them that for various applications (such as differential gene expression, quantification of lowly expressed genes) "the existing (Illumina-based) approaches" Illumina RNA-seq or short-RNA-seq would be more appropriate due to higher sequence yields, the main advantage of nanopore direct RNA-seq is the ability to identify RNA modifications directly at the same time (see new Figs 7 and S23). While we also agree that the field of high throughput sequencing epitranscriptomics is quite nascent, it is quickly gaining interest. Based on the reviewer's suggestions we believe that we have now modified the manuscript substantially to provide justification for NERD-seq's feasibility as a methodology to capture multiple classes of ncRNAs for the study of their sequence, quantity and their epitranscriptomic signatures.

We would also like to note that even as a BioRxiv preprint, our manuscript has received already a number of citations, including from an article published in Nature Biotechnology and feel this demonstrates interest in NERD-seq from the scientific community.

Reviewer 2

Reviewer #2 (Evidence, reproducibility and clarity (Required))

• *

This reviewer generally remarks that "In this manuscript Saville et al. present NERD-seq, a novel method to enrich and directly sequence non-coding RNA using the Oxford Nanopore Technology (ONT). The technique is based on the addition of a poly-A tail to small non-coding RNAs which enables their sequencing by a conventional ONT direct RNA-seq protocol using a poly(T)-tethering adaptor. This technique has great potential, as ONT direct RNA sequencing can be used to detect RNA modifications. Furthermore, non-coding RNAs are known to harbor a lot of these modifications and that they are important to their function. In addition, non-coding RNAs aren't generally poly-adenylated and therefore are underrepresented when using standard ONT direct RNA-seq. The authors have shown that using NERD-seq, they are able to significantly improve the number of reads on several subsets of non-coding RNAs : tRNAs, snoRNAs, snRNAs, scRNAs and rRFs. There are however some issues that need to be addressed before the manuscript is suitable to publication."

• *

We thank the reviewer for their encouraging comments.

Major comments:

• *

Point 1. The reviewer recommends that "Data on RNA modifications present in the sequenced non-coding RNA would greatly improve the impact of the manuscript, as the method is presented as read tool to study RNA modifications. If the authors can't include these data, they should at least explain why in the discussion section."

• *

We have now included a new main figure (Fig.7) and a supplemental figure (Suppl. Fig. S23) which show the ability of NERD-seq to identify previously described RNA modifications and its superiority in the case of non-coding RNAs such as 7SK. We thank the reviewer for this recommendation that significantly improved our manuscript better revealing the rationale and impact of our approach.

Point 2. The reviewer remarks that "The authors did not discuss possible biases of their method. For example, is the poly-A tail ligation efficiency affected by sequence and or structural differences between RNA? Is there detectable differences in the efficiency of sequencing small RNA with structured 3'end when compared to mRNA? Are the proportions of reads obtained per non-coding RNA sub-category and RNA rank supported by other biochemical data (e.g. Northern blots, primer extension, RT-qPCR? "

• *

We have included commentary in the discussion on these biases (page 19, lines 7-11). We also include a new supplementary figure (Suppl. Fig. S5F) regarding polyadenylation in which we see similar transcript length to polyadenylation tail length dynamics between the two methodologies.

Point 3. The reviewer notes that "It is essential to include external size and sequence markers spike-in to directly evaluate the quality and fidelity of the newly develop sequencing technique. "

• *

We have now repeated the sequencing, including sequencing with the well described RNA sequins mix B. We include a new supplementary figure with this data (Suppl. Fig. S10) that shows that NERD-seq does not differ from the standard approach regarding its ability to effectively evaluate the quantity and complexity of the transcripts present.

Point 4. The reviewer mentions that "The author compared NERD-seq to ONT standard direct RNA-seq but did not compare it with other methods that are currently used for non-coding RNA sequencing (e.g. Illumina based sequencing, TGIRTseq). There is a need for side by side comparison at least by using data available in the literature if not experimentally. "

• *

We now include a comparison with Illumina sequencing data for both long and short fractions from each biological replicate (Suppl.Figs S12-15,17,19 and 21). This data shows that NERD-seq combines the advantages of both standard and short RNA-seq with Illumina, enriching both long and short RNAs, while simultaneously as mentioned above, offering the benefits of direct RNA-seq. We feel this also demonstrates the utility of the unique chemistry of Nanopore sequencing where fragmenting of long RNAs is not required for sequencing, allowing the combination of long RNAs and short RNAs in the same library to reproduce what is essentially enriched in short RNA libraries as well as poly(A) sequencing libraries. Since the enzyme included in our initial submission is not anymore widely accessible, we also tested a series of other enzymes, including the commercial replacement for omniamp (the originally used enzyme), lavalamp, and other small tweaks to the library methodology, until finding the GSP SSD2.0 enzyme to be the most suitable for our research goals. The summary of our findings is in Figure S5. Initially, we considered TGIRT as a potential enzyme for the sequencing reaction because of its use for sequencing snoRNAs and other highly structured RNAs but because of its strand switching properties, we felt it would be poorly suited for a nanopore direct RNA sequencing approach.

The reviewer made also the following comments/suggestions:

Minor comments

• *

"Page 5 line 3, NERD-seq is misspelled. "

• *

This is now fixed.

"Throughout this section µL and µg are written as uL and ug."

• *

This is now fixed.

"Several temperatures are missing the o symbol between the number and C."

• *

This is now fixed.

Final comments:

• *

"This manuscript has the potential of having major significance to researchers in the field of transcriptomics if the author demonstrated that their newly sequencing techniques could indeed improve RNA detection beyond the current techniques. Unfortunately, the manuscript only show the capacity of the technique to detect non-coding RNA but fall short from providing strong arguments supporting the fidelity, accuracy and coverage of the new technique. The use of Nanopore sequencing is not unique and was used in the past for sequencing non-coding RNA. What is needed now is a proof of the increased accuracy or utility of the new technique to justify yet another publication about Nanopore sequencing paper. The manuscript would potentially be of interest to researcher working on different type of non-coding RNA and transcriptomics.* *

• *

I am in the field of non-coding RNA sequencing and functional analysis and very familiar with this approach."

• *

We thank the reviewer again for their encouraging comments and valid concerns. We agree that the initial manuscript fell short of providing evidence of its utility as an addition to the Nanopore sequencing repertoire. We hope the additional data we provide (rerunning with an additional enzyme, use of replicates, use of standard spike in controls, testing an additional organism and tissue, testing and comparing different enzymes and other sequencing platforms) provide the necessary assurances about reproducibility and accuracy, while the new data concerning identification of RNA modifications and detection of important RNA classes missed by the standard protocol provides the necessary assurances about the utility of our approach.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

In this manuscript Saville et al. present NERD-seq, a novel method to enrich and directly sequence non-coding RNA using the Oxford Nanopore Technology (ONT). The technique is based on the addition of a poly-A tail to small non-coding RNAs which enables their sequencing by a conventional ONT direct RNA-seq protocol using a poly(T)-tethering adaptor. This technique has great potential, as ONT direct RNA sequencing can be used to detect RNA modifications. Furthermore, non-coding RNAs are known to harbor a lot of these modifications and that they are important to their function. In addition, non-coding RNAs aren't generally poly-adenylated and therefore are underrepresented when using standard ONT direct RNA-seq. The authors have shown that using NERD-seq, they are able to significantly improve the number of reads on several subsets of non-coding RNAs : tRNAs, snoRNAs, snRNAs, scRNAs and rRFs. There are however some issues that need to be addressed before the manuscript is suitable to publication.

Major Comments:

• Data on RNA modifications present in the sequenced non-coding RNA would greatly improve the impact of the manuscript, as the method is presented as read tool to study RNA modifications. If the authors can't include these data, they should at least explain why in the discussion section.

• The authors did not discuss possible biases of their method. For example, is the poly-A tail ligation efficiency affected by sequence and or structural differences between RNA? Is there detectable differences in the efficiency of sequencing small RNA with structured 3'end when compared to mRNA? Are the proportions of reads obtained per non-coding RNA sub-category and RNA rank supported by other biochemical data (e.g. Northern blots, primer extension, RT-qPCR?

• It is essential to include external size and sequence markers spike-in to directly evaluate the quality and fidelity of the newly develop sequencing technique.

• The author compared NERD-seq to ONT standard direct RNA-seq but did not compare it with other methods that are currently used for non-coding RNA sequencing (e.g. Illumina based sequencing, TGIRTseq). There is a need for side by side comparison at least by using data available in the literature if not experimentally.

Minor Comments:

• Page 5 line 3, NERD-seq is misspelled.

• Throughout this section µL and µg are written as uL and ug.

• Several temperatures are missing the o symbol between the number and C.

We recommend that all these issues should be addressed before publication.

Significance

This manuscript has the potential of having major significance to researchers in the field of transcriptomics if the author demonstrated that their newly sequencing techniques could indeed improve RNA detection beyond the current techniques. Unfortunately, the manuscript only show the capacity of the technique to detect non-coding RNA but fall short from providing strong arguments supporting the fidelity, accuracy and coverage of the new technique. The use of Nanopore sequencing is not unique and was used in the past for sequencing non-coding RNA. What is needed now is a proof of the increased accuracy or utility of the new technique to justify yet another publication about Nanopore sequencing paper. The manuscript would potentially be of interest to researcher working on different type of non-coding RNA and transcriptomics.

I am in the field of non-coding RNA sequencing and functional analysis and very familiar with this approach.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

Summary:

Saville et al detail a new method, NERD-Seq, which extends nanopore direct RNA-Sequencing to beyond the poly(A) fraction. In principle, this allows the capture of additional RNA types in a single sequencing reaction, albeit at the expense of sequencing depth. However, the authors need to expand more on why this is useful, and what scenarios this would be used in. The rationale for many of their experimental choices are not explained/contextualized appropriately. The manuscript also suffers from excessive jargon and grammatical errors.

Major comments:

1: To my understanding, the primary purpose of NERD-Seq is to allow the sequencing of the non-adenylated fraction of RNAs within sample with a size filter step designed to exclude ncRNAs larger than ~200 nt. While this allow for subsequent polyadenylation of the small fraction, it remains quite likely that larger non-adenylated non-rRNAs are being missed by the protocol. It is also not clear why NERD-Seq should be considered the optimal strategy. The authors should show that they have considered/evaluated other strategies such as: - Depletion/Seperation of the poly(A) fraction prior to size selection performed on non-adenylated fraction. - Targeted degradation of the rRNA fraction using rRNA depletion kits.

2: Related to this, the authors note that NERD-Seq is designed to allow the sequencing of both adenylated and (short) non-adenylated fractions of RNAs within a study. What is the actual value of this versus, say, simply targeting the short non-adenylated fraction directly?

3: The manuscript in general is written in a somewhat subversive style, seemingly focused on highlighting the 'failings' of the standard ONT protocol. This is somewhat disingenuous as these are not 'failings' per se given the design objective of the standard DRS protocol (i.e. to capture and sequence the poly(A) fraction of RNAs) works well. The authors would be better off highlighting the situations (i.e. study questions) where NERD-Seq would provide a measurable benefit over the standard DRS strategy.

4: The authors state that 1.5ug total RNA is used as input for NERD-Seq but how much input (polyA RNA) actually goes into the short- and long-fraction parts of the DRS protocol? This is important to know.

5: The authors show that NERD-Seq performs well on a tissue that is generally enriched in non-coding RNA activity but without the inclusion of biology replicates or tissue samples from other sources, it is impossible to assess (1) the reproducibility/robustness of the methodology and (2) whether NERD-Seq would produce useable data from other tissues with lower non-coding RNA activity. These experiments are required.

6: Several figures are poorly explained. For instance, Does Figure 2A show data only from the short RNA fraction? If not then this suggests incredibly high levels of mRNA degradation in the 'standard' ONT fraction - far beyond what is seen in other studies. The legends for all figures should be far more specific and detailed.

Minor quibbles:

1: Please refrain from using 'next-generation' in a sequencing context (abstract, introduction). Having a 'new' next-generation sequencing is confusing in regard to the old 'next generation' sequencing.

2: Introduction should acknowledge that targeted sequencing of (individual) non-adenylated RNAs is possible (i.e. there is an ONT protocol for this).

3: Several figure elements not referenced in the text appropriately (e.g. red bars in Figure 2A).

4: The authors mention profiling of the epitranscriptome several times in the introduction and discussion but do not include any work geared toward looking for RNA modifications. Indeed it is entirely unclear whether NERD-Seq produces the depth of sequencing (on non-adenylated RNAs) required for current RNA modification detection tools.

Significance

NERD-Seq presents a modified method for DRS of both short non-adenylated and the standard adenylated fraction of RNAs within a sample. This utility of this approach appears rather niche and it is hard to determine situations in which this approach would be generally useful versus many of the existing (Illumina-based) approaches. The burden to demonstrate that NERD-Seq is more than an incremental improvement to existing approaches lies with the authors.

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

The authors do not wish to provide a response at this time.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #3

Evidence, reproducibility and clarity

Summary:

This manuscript examines the effects of fibroblast-derived extracellular vesicles (EV) on axon outgrowth in primary neurons, and investigates potential underlying mechanisms. The authors show that fibroblast EV increase axon outgrowth, which is dependent on components of the Wnt-PCP pathway in neurons. They further show that axonal outgrowth is not affected by neuron- or astrocyte-derived EV and that EV from activated astrocytes inhibit axon elongation. Although several experiments are performed thoroughly, major revision is required to substantiate the main claims of the manuscript.

Major comments:

1. Even though the authors made a good effort to characterize fibroblast-derived EV, the data so far does not indicate that strictly 'exosomes' are implicated in axon outgrowth, as it is currently virtually impossible to isolate a pure population of exosomes.
2. In the methods the authors state that the conditioned media was stored for up to 8 weeks at 4C. As long-term storage of EV was shown to decrease their activity, the authors should specify whether they took steps to test the effect of storage time on EV concentration and activity.
3. In Fig 3F, the authors claim that Vangl2 re-localizes from proximal to distal end of the axon. However, in the representative image the PBS-treated axon is much shorter, and Vangl2 can also be also detected in the growth cone. Therefore, it is not clear how neurons were classified in the analysis. As this is one of the main claims of the paper, the analysis should be performed in a more quantitative manner, such as quantification of intensity and volume of Vangl2 in the soma, proximal / distal axon and growth cone, while accounting for changes in axon length.
4. The claim that exosomes mobilize neuronal Wnt to promote axon growth is unsubstantiated. Co-localization between Wnt7b and GFP-CD81 is not convincing given the low magnification and broad distribution of Wnt7b. It is also unclear how EV internalized in the soma would have this effect. Additional experiments to prove the direct influence of fibroblast EV on Wnt-PCP signaling (and optionally, how this is unique for fibroblast EV) would increase the validity of these claims.
5. Neuronal EV were isolated at a different developmental time point (8DIV), which most likely has an effect on EV composition. Therefore, the claim that neuronal EV do not promote neurite outgrowth is not convincing. In addition, AraC or LPS used to treat neurons and astrocytes respectively, could be co-purified with EV and therefore have adverse effects on recipient neurons.
6. The authors claim that this effect is unique to fibroblast EV. This claim is not valid without a full characterization of EV derived from multiple cell types and is therefore misleading.

Minor comments:

1. As in point 1 above, it is recommended to replace the term 'Exosome' in the figures and refer instead to the method of purification (eg. 100k). The term 'Exosome markers' in the text should also be replaced accordingly, as it is not currently clear whether CD81, TSG101 or Flotilin 1 are strictly on exosomes.
2. Experiments using EV purified using gradient centrifugation (Fig S3) should be repeated at least once, such that statistical significance is calculated and results are shown as in other functional experiments. Alternatively, experiments could be performed using purified EV treated with a proteinase in the presence or absence of detergent, to verify whether it is EV cargo or extracellular proteins co-isolated with EV that mediate the observed phenotype.
3. The authors should show that porcupine knockdown results in decreased Wnt secretion in fibroblast EV using western blotting.

Significance

Overall, the main take home message of this study is that fibroblast EV could have the common feature of upregulating axon outgrowth in neurons during development. While this is not relevant for CNS development per se, it could ultimately be of interest to a specialized audience investigating the translational relevance of fibroblast-derived EV.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

The definition of axonal fate has been under a lot of scrutiny. However nowadays it is not clear how the axonal fate could be distinguished from axon elongation. In other words, how one the neurite is selected to eventually grow as an axon versus the mechanism sustaining axonal growth once the fate is established. In addition, it is not clear, once the axon is growing, what is the mechanism that precludes the other neurites (future dendrites) from growing.

The authors of this manuscript claim that fibroblast-derived exosomes promote axon growth; meanwhile, exosomes from activated astrocytes inhibit axon elongation. Moreover, the authors claim that exosomes mediate axon elongation throughout the PCP and Wnts pathways.

I have several concerns regarding the data and concept presented in this manuscript that I consider precluding its publication in the current form.

First, the claim that axon growth is promoted by exosomes is not well supported by the quantifications. It will be more convincing to show the frequency distribution for each experiment rather than the average growth per experiment (e.g., Figure 1F). In the way the data is presented now, we do not learn the real effect of the treatment on axonal growth. For instance, how is the variability in axon length in each experiment? the way the data is presented now might mask variability that could reduce the effect of their treatment

Second, I find it difficult to explain this concept in neurons differentiating in the developing cortex. Which cell type is the source of exosomes mediating axon elongation? Is this cellular mechanism an artifact of the culture condition? In other words, are exosomes relevant for axonal growth in situ?

In the model presented in Figure 7, authors show that fibroblast might mediate axon extension meanwhile activated astrocytes preclude axon extension. Authors do not consider that in the developing cortex, neurons are formed and elongating an axon (while they migrate to the cortical plate) before astrocytes are produced (Noctor et al Nature Neurosci. 2004). How do the authors reconcile this conceptual discrepancy with her in vitro studies? How do the authors reconcile this discrepancy with their studies in situ?

Significance

The authors of this manuscript claim that fibroblast-derived exosomes promote axon growth; meanwhile, exosomes from activated astrocytes inhibit axon elongation. Moreover, the authors claim that exosomes mediate axon elongation throughout the PCP and Wnts pathways.

I have several concerns regarding the data and concept presented in this manuscript that I consider precluding its publication in the current form.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

Ahmad et al. report that exosomes isolated from fibroblast cell lines stimulate neurite extension in cultured neurons while those from neurons or astrocytes do not. They investigate the role of different components from the planar cell polarity (PCP) pathway and conclude that exosomes stimulate autocrine signaling by Wnts in a PCP pathway-dependent manner. However, the inactivation of most PCP components severely reduces neurite length already in untreated controls suggesting that they are required for neurite growth in general rather than specifically for the response to exosomes. The manuscript reports an extensive set of interesting results but does not provide sufficient evidence for the proposed mechanism.

Major points:

1. Knockout, knockdown, pharmacological inhibition or blocking antibodies were used to inactivate multiple components of the PCP pathway in cultured neurons. In most cases, the inactivation resulted in a reduction of neurite length both in control cultures and in cultures with exosomes. Because neurons lack the capacity to extend longer neurites after the inactivation of PCP components, it is not possible to determine if they are required for the response to exosomes. Only Fzd7 appears to be specifically required for the effect of exosomes since its knockdown does not reduce neurite length in controls.
2. The physiological relevance of the stimulation of neurite growth by exosomes is unclear. It is not explained how cortical neurons come into contact with fibroblasts or the exosomes produced by them as the authors acknowledge at the end of the discussion.
3. The mechanism how exosomes stimulate neurite growth remains unclear. The authors suggest that exosomes modulate autocrine Wnt signaling but do not provide sufficient evidence for this. The knockdown of Wntless or Wnts and blocking Wnt secretion by inhibiting Porcupine suppress neurite extension. This phenotype could results from a defect in autocrine signaling but also from a reduced secretion of Wnts into the medium.
4. The authors claim that the cell body takes up exosomes that then acquire endogenous Wnt7b. The co-localization of the signals in Fig. 5H is not informative because Wnt7b shows uniform distribution while the CD81-EYFP signal is present in distinct structures that do not show a stronger EYFP signal.
5. The specificity of the siRNAs has to be verified by rescue experiments in neurons.

Minor points:

It is not explained why the authors tested exosomes from fibroblasts.

Fig. 7: The graphical summary shows that exosomes somehow enter the soma of neurons. It is not clear if this happens by endocytosis or fusion with the membrane. In either case, the topology of exosomes in the soma is incorrect.

Significance

Ahmad et al. report that exosomes isolated from fibroblast cell lines stimulate neurite extension in cultured neurons while those from neurons or astrocytes do not. They investigate the role of different components from the planar cell polarity (PCP) pathway and conclude that exosomes stimulate autocrine signaling by Wnts in a PCP pathway-dependent manner. However, the inactivation of most PCP components severely reduces neurite length already in untreated controls suggesting that they are required for neurite growth in general rather than specifically for the response to exosomes. The manuscript reports an extensive set of interesting results but does not provide sufficient evidence for the proposed mechanism.

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

Manuscript number: RC-2024-02335

Corresponding authors: Pöhlmann, Stefan & Karsten, Christina B.

1. General Statements

We would like to express our gratitude to Reviewers 1, 2 and 3 for dedicating their time to assess this manuscript and for sharing their invaluable expertise on the subject matter. We have incorporated most of the suggestions made by reviewers but regrettably we were unable to conduct any additional experiments. This is because we chose to work with a single pooled virus stock for both CD4+ T cell and macrophage-derived simian immunodeficiency virus (SIV, T-SIV, M-SIV) throughout the whole study. While this heightened our data quality by omitting donor-induced variations into our virus productions, accurately estimating the total amount of viruses needed for this project was difficult, and at this point we have depleted our reserves of M-SIV and T-SIV. Nevertheless, we are confident that our manuscript has been substantially improved in response to the reviewer’s feedback, and we firmly believe that this study holds considerable implications for the development of new biomedical interventions against the human immunodeficiency virus (HIV).

Please see below our point-by-point responses to the reviewer’s comments and concerns. All manuscript references refer to the revised manuscript, in which the key changes have been tracked.

2. Point-by-point description of the revisions

Reviewer #1:

Major comments:

1 - As noted above, the differences in glycosylation are difficult to understand without more background and perhaps a figure, but it is also not clear that the changes described between lines 119 and 137 are biologically or statistically meaningful. For example, does it matter that more M-SIV virions have glycans with four antennae than T-SIV? Are there other data that show this or could experiments be done to specifically cleave these glycans at certain points to reduce their complexity and show that the infectivity differences between M-SIV and T-SIV disappear? Further, it is difficult to confirm the statement on lines 124-125 that "profiles of complex-type N-glycans differed between the two viruses (Fig. 2C)", as no statistical tests were done to compare the glycosylation being detailed in the M-SIV and T-SIV. It is more appropriate to make note that there are minor between M-SIV and T-SIV or run specific statistical tests on the data.

These are very good remarks. We have introduced a new figure (Fig. 1A) to illustrate the various glycan types, moieties, and structures discussed in the paper. Additionally, we have modified Fig. 1D-G with visualizations of the investigated N-glycans and adjusted and clarified the text to enhance accessibility for experts not specializing in glycobiology (line 62-66, 204-210, 269-273). With our current dataset, we are unable to establish a direct correlation between glycosylation and functional outcomes. Consequently, we can only speculate about the potential impact of certain glycan characteristics on SIV viral functions, a field that remains largely unexplored. Furthermore, due to the absence of remaining virus samples, we are unable to conduct experiments to validate any potential direct relationship between glycosylation patterns and viral functions. It's important to note that no statistical tests were applied to the glycan analysis data, we have now emphasized this in the revised text (line 13-16, 145-149, 250-254).

2 - On line 161, the authors note that the results showing that "the virus-producing cell has a broader impact on SIV infectivity beyond its influence on Env incorporation." This is certainly one possibility that is suggested by these and prior data, there are also other possibilities. For example, the impact of Env is not linear and perhaps a certain number of Env need to be engaged, creating some kind of threshold effect that means that the virions with fewer Env just have less infectivity. Given that there is significant data that virions are generated in different locations in macrophages and T-cells, this could also be a function of which specific membrane areas in different cell types that Env embeds in, or it could be something else associated with Env that is not cell type specific.

We agree and have adapted the text to emphasize that our interpretation is just one possible interpretation, which assumes a linear relationship between Env incorporation and SIV infectivity (line 176-178).

*3 - For the studies in Figure 5, looking at the direct vs. indirect infectivity, it is not clear why CEMx174 R5 cells (a T-cell/B-cell hybrid line) were used instead of primary macrophages or T-cells, or macrophage or T-cell lines or a fully agnostic cell type. This would be more convincing tested on primary cells, or at least comparing in a myeloid lineage line as well. *

In this experiment, we followed a previously published protocol, which utilizes the CEMx174 R5 indicator cell line as target cell for an easy read-out of the results. We agree that the use of the suggested target cell types would be a useful extension of our existing work. Regrettably, due to the lack of remaining viruses, we are unable to fulfil this request.

*4 - In Figure 6, it is not clear that VSV-G pseudotyped virus is an appropriate control, as it enters via the acidified endosome pathway and not via similar processes as the T- and M-SIV derived virions. While this may show that the glycans can bind to CBA to inhibit entry, it could also mean that the general process of endocytosis is not as susceptible to CBA inhibition and this difference in pathways should be noted as a caveat. *

We acknowledge this caveat and have incorporated a statement into the results section to inform readers about the potential impact of differences in the viral uptake pathways between VSV and HIV/SIV (line 219-223).

*5 - A very large number of cell lines were used, and it is not clear why experiments were done using so many different indicator or target lines, instead of performing most assays in a single line or set of lines so that they are comparable across experiments. Some discussion of the rationale for this would be helpful. *

Thank you for bringing your perspective to our attention. By selecting the specific cell lines for our experiments, we have adhered to established norms within our research domain. Specifically, 293T cells, C8166, and TZM-bl cells are standard choices for virus production via transfection, SIV titration, and infectivity assays, respectively. Moreover, in conducting the transmission assay, we followed a previously published protocol. We have revised the text to elucidate the rationale behind our selection of cell lines for different assays, omitted the potentially perplexing reference to the C8166 cell line and included references to support our cell line choices (lines 160-163, 187-188, 215-217, 458-459, 476).

Minor comments

*1 - Inclusion of the p27 data characterizing the amount of virus in M-SIV and T-SIV stocks (line 95) should be shown as at least a supplemental figure or could easily be added to figure 1. *

Agreed. We have included the information about the p27-concentration, and additionally the information about the infectious units and RNA copy numbers of M-SIV and T-SIV into the results section (line 98-101).

3 - The figures are relatively thin and could be combined with other figures to better connect the experiments. For instance, Figure 1 could serve as panel A for what is currently listed as Figure 2 because it is a preliminary data to the experiments in Figure 2.

As suggested by the reviewer, we have combined Fig. 1-2 (now Fig. 1), and additionally Fig. 3-5 (now Fig. 2), and 6-7 (now Fig. 3) to enhance the connections between the experiments.

3 - The authors should include quantitation of the Western blot data in Figure 1 in an adjoining graph.

We appreciate the suggestion. However, the purpose of Fig. 1B (formerly Fig. 1) was to visually represent gp120 of M-SIV and T-SIV following all treatment conditions, which produce bands of varying intensities and widths. Consequently, the bands for PNGaseF-digested gp120 appear relatively thick, which typically hampers accurate quantification. Therefore, we made the decision not to quantify the results of these blots.

*4 - The legend states that the results in Figure 7 were obtained from two independent experiments (line 778), each with 3 technical replicates. As this represents only 2 biological replicates, and the experiments were performed in easily accessible cells (TZM-bl), they should be performed 1 - 3 more times to provide a more appropriate and robust data set for statistical analysis. *

We agree with the reviewer that this might further improve the results but unfortunately, due to the lack of remaining pooled virus from this study, we are unable to fulfill this request.

Reviewer #2:

Major comments:

*It appears to this assessor that some of the supplementary data can be brought to the front as part of the main figures for presentation. *

We appreciate that you consider our supplementary data worthy enough to be part of the main figure set. We have now included Table S1 into the main figure set as Table 1.

CURRENT figures 3, 5, and 7 can be combined into one figure. Similarly, CUREENT Figure 4, and 6 can also be grouped. Alternatively, incorporating additional approaches in each set of figures to tighten the claims.

Agreed. To address this suggestion, we have combined Fig. 1-2 (now Fig. 1), 3-5 (now Fig. 2), and 6-7 (now Fig. 3).

*Karsten et al pitched their story as glycosylation of SIV from different primary cells are linked to different functionality in its title and abstract, yet the authors then declared in discussion (line 318) that establishing a direct link between Env glycosylation and viral functions is technically challenging and beyond the scope of the study. This assessor feels that authors need to decide whether current manuscript should be a descriptive study (which is more fitting for a less impactful journal) or a study with further mechanistic insights. *

Thank you for bringing this to our attention. We have modified the text in the introduction, results, and discussion section to underscore that our work is only suggesting but not directly proving differential glycosylation as cause for functional differences between M-SIV and T-SIV (lines 71-75, 155-157, 211-215, 251-256, 320-322).

*Table S1 is highly important and should be part of the main figure. Specifically, authors took the opportunity to highlight the differential % of sialic acid terminal glycans in line 133. The charge of the sialic acids would be simple mechanism for these M-SIV particle to attach. Authors should consider some of the described nano-luciferase based viral particle attachment assays used in HIV-glycan biology. Authors should be able to treat SIV (or SIV VLPs) with sialidase to quantify the role of sialic acids on binding. *

Thank you for appreciating the details of our glycan analysis. We have now included Supplementary table 1 into the main figure set as Table 1. Due to the lack of remaining virus, we are unable to address the interesting suggestions for further experiments.

*As authors carefully pointed out (throughput the manuscript) that the identity (and biology) of the producer cells can have profound impacts on glycosylation events of viral particles that are being produced. This assessor was then interested to understand precisely how their simian PBMCs and monocytes derived macrophages were prepared. Additional details in M&M would be very helpful. *

We realized that indeed experimental details appeared to be missing in this section since it was not obvious that kits for magnetic bead isolation have been used to isolate the cells, and adapted the text to make this more clear (395-413).

*With the emphasis of cell type and glycosylation relationship, it is puzzling that authors would have chosen to use TZM-bl (artificially engineered cell line) and spinoculation (2hr to push the viruses down to cell surface with 870 x G force) in Figure 3 for comparison of M-SIV and T-SIV infectivity. To this assessor, this assay neglected the biological roles of SIV glycans. In context, 870 x G is ~150x higher than most human can withstand. *

While we appreciate the reviewer’s feedback, we respectfully disagree. The TZM-bl cell line has long been established as the standard cell line for SIV/HIV infectivity assays and neutralization assays in clinical trials (Sarzotti-Kelsoe et al., 2014). In this project, our initial aim was to conduct infectivity assays on a standard cell line before transitioning to more biologically relevant target cells. However, due to limited virus availability, these studies could not be completed and will be addressed in future studies.

The application of methods to increase virus-cell contact to increase cell infection of cell lines is wide-spread in HIV research. Larger virus quantities could have been used instead of spin infection, which would require the introduction of larger amounts of conditioned cell supernatant from the virus production into the experiment with potential influence on the outcomes. Another option would have been to expose the viruses to even stronger forces during ultracentrifugation to concentrate virus stocks, or to employ "sticky" reagents such as DEAE-dextran, which might generate virus aggregates (Davis et al., 2004). Considering these options, we deemed spin infection to have the smallest overall impact on our experiments, while delivering the most robust results.

Finally, we like to note that sensitivity to force between humans and cells is not comparable and cells can withstand much higher forces than humans. We kindly refer reviewer 2 to the work of Kodaka and colleagues. They carefully assessed the efficiency and impact of spin infection using retroviral constructs on primary human cells, and determined that the best conditions were 2,800xg for 90 min considering important parameter’s such as cell viability, proliferation and in vivo differentiation.

Sarzotti-Kelsoe M, Bailer RT, Turk E, Lin CL, Bilska M, Greene KM, Gao H, Todd CA, Ozaki DA, Seaman MS, Mascola JR, Montefiori DC. 2014. Optimization and validation of the tzm-bl assay for standardized assessments of neutralizing antibodies against hiv-1. J Immunol Methods, 409, 131-146. DOI:10.1016/j.jim.2013.11.022.

Davis HE, Rosinski M, Morgan JR, Yarmush ML. Charged polymers modulate retrovirus transduction via membrane charge neutralization and virus aggregation. Biophys J. 2004 Feb;86(2):1234-42. doi: 10.1016/S0006-3495(04)74197-1. PMID: 14747357; PMCID: PMC1303915.

Kodaka Y, Asakura Y, Asakura A. Spin infection enables efficient gene delivery to muscle stem cells. Biotechniques. 2017 Aug 1;63(2):72-76. doi: 10.2144/000114576. PMID: 28803542; PMCID: PMC5768144.

*Using a single antibody DA6 (in Figure 2, cited Edinger 2000) for Env incorporation estimation via Western seems to be crude and inadequate, even in the context of isogenic virus clone. As authors pointed out, different levels of glycosylation can affect protein folding, therefore affecting Env incorporation. By the same argument, differentially glycosylated Env protein can also impact on the ability of 'epitopes within Env protein' to be recognised by Ab. Therefore, virion incorporation of Env might not be affected, but just the detectability by a specific Ab. Western evaluation with a panel of anti-Env antibodies will help. Furthermore, quantitative proteomics coupling with glycomics would be highly useful. *

We respectfully disagree on some points of your assessment. The antibody DA6 specifically targets a linear epitope in gp120 C1 (amino acids 76 to 99 in SIVmac251, Edinger et al., 2000), and the proteins analyzed by Western blot are denatured. Therefore, changes in protein folding due to differential protein glycosylation of SIV in different target cells should not affect the results. We acknowledge the potential impact of differential N-glycan attachment to gp120 based on the virus-producing cell on the binding of primary antibodies in Western blot analysis and agree that this issue could be mitigated by employing a panel or mixture of primary antibodies. However, please note that in Fig. 1B the input virus of M-SIV and T-SIV was normalized based on the signal received for gp120 removed of N-glycans (PNGase F digested). If the differential glycosylation of M-SIV and T-SIV gp120 would interfere with DA6 binding, we should observed noticeable differences between M-SIV and T-SIV in the signal of the undigested and mannose-reduced gp120 (Endo H) but this is not the case. Thus, we believe it is in this case sufficient to use only the antibody DA6 for gp120 detection.

Edinger AL, Ahuja M, Sung T, Baxter KC, Haggarty B, Doms RW, Hoxie JA. 2000. Characterization and epitope mapping of neutralizing monoclonal antibodies produced by immunization with oligomeric simian immunodeficiency virus envelope protein. J Virol, 74(17), 7922-7935.

*It is understood that T-SIV were pooled from supernatant derived from 9 animals of PBMCs. Levels of p27 production (presumed as particles but including free p27 in reality) from each animal donor should be listed in supplement. Similar types of details should be made available for M-SIV that were derived from 8 animal donors of macrophages. qPCR estimations on the levels of viral particles production in T-SIV and M-SIV from primary cell culture amplifications appear to be already available, such information should be included in supplementary to strengthen the authors' estimated / relationships amongst glycosylation, virion Env incorporation levels, and viral particle productions are carefully controlled. *

Thank you for your suggestion. We opted to include only supernatants containing more than 10 ng p27/ml in the pooled virus that constituted M-SIV and T-SIV. Consequently, we did not determine the p27 concentrations of each virus harvests below this threshold. As a result, we are unable to present replication curves for every virus production. However, we are able to provide additional information on the pooled viruses and now included the information about the p27 concentration, infectious units, and genome copy number of M-SIV and T-SIV (line 98-101).

*Non-glycan biologists generally do not appreciate some of the fine details in glycan biology. The T-SIV and M-SIV system is a great model system to decode some of the functionality of glycan biology. The current team should have (in my opinion) a clear graphic representation on describing what types of different glycans in T-SIV and M-SIV are likely to contribute to the potential differences in biological outcomes. Such incorporation will guide non-glycan biologists to better appreciate the focus and the directions of authors, thereby further improving the citation of this work when it is published after peer reviewed. Importantly, focusing a specific question to be addressed may help to consolidate effort to accelerate publication of this work. A beautiful story line, just need to cross many 't' and dot a few 'i' in my view. *

Thank you for sharing the excellent suggestion. We have now incorporated a new figure (Fig-1A) to help the reader with understanding the glycan biology in our manuscript. We have further adapted Fig. 1D-G to include glycan structures to provide a visual representation of the assessed N-glycan subgroups. Finally, we adapted the text throughout the manuscript to improve the reader’s comprehension of our work (line 62-66, 204-210, 269-273).

To address your concerns regarding the clarity of the research question, we revised the text to become more specific (line 155-157, 277-283, 294-300, 317-319, 320-322, 329-338, 364-365).

Most primate centres often incorporate transcriptomic studies in their animal works. It will be helpful for the audience if the authors could provide additional transcriptomic data (with a focus on glycosylation related genes) of simian CD4+ T cells, simian macrophages, SIV infected simian CD4+ T cells, and SIV infected simian macrophages. These data will improve the comprehensiveness of this study (and should not require any major wet-lab studies) and add weight on the arguments of the authors.

This is an interesting suggestion and should be considered in future studies. Here, we choose to focus exclusively on the investigation of the viruses but not the host cell itself. Nevertheless, we discuss the existing transcriptomics data of glycosylation relevant genes in simian CD4+ T cells and macrophages published by Gaskill and colleagues (line 277-287) and find that their results provide explanations for the results of our N-linked glycan analysis of M-SIV and T-SIV.

Gaskill PJ, Zandonatti M, Gilmartin T, Head SR, Fox HS. 2008. Macrophage-derived simian immunodeficiency virus exhibits enhanced infectivity by comparison with t-cell-derived virus. J Virol, 82(3), 1615-1621.

Reviewer #3:

Major comments *1) Line 85 "These substitutions facilitate efficient utilization of CCR5 in the absence or at very low levels of CD4 expression (Puffer et al.2002). This makes the molecular clone studied rather unique and the authors should aim to address this throughout. It does not take away from the results presented but should be addressed. *

Thank you for this suggestion. We included the information about the M-tropism of our viral strain two more times into the manuscript to emphasize this information in the discussion and the limitations of the study (lines 247-249, 345-347).

*2) (Figure 1) State the predicted size differences between T-SIV and M-SIV stocks with EndoH digestion (and similar for all 3 runs?). *

We would have liked to address this suggestion but despite corresponding with other experts and literature research; we were unable to identify a tool, which would allow us to make such predictions.

*3) Line 95 Data should be shown. This describes infectivity and could incorporated within Figure 3 along with infection of the TZM-bl cells. Infectivity of T-SIV and M-SIV on primary CD4 T-cells and macrophages is of importance. This would only be possible I assume if p27 levels were measured at each time-point collected. *

We agree that the manuscript could be extended by virus replication curves on primary cells. However, we chose to include only supernatants with p27 concentrations exceeding 10 ng/ml in the pooled virus comprising M-SIV and T-SIV. Consequently, we did not determine p27-concentrations for virus harvests below this threshold in all cases, preventing us from presenting replication curves for every virus production. We have updated the results section to incorporate the total p27-concentration of M-SIV and T-SIV (lines 98-101). Prior to the submission of this manuscript, we had initiated replication assays of M-SIV and T-SIV on primary rhesus CD4+ T cells and macrophages were initiated but these could not be completed due to the depletion of available virus.

4) There is a lack of statistical difference in the results shown in Figure 2. I assume this is due to a single measurement, but can comment be made on likelihood of biological significance with such difference between values.

Indeed, we did not conduct statistical analysis on this dataset because, given that a single pooled virus for M-SIV and T-SIV was utilized, only one measurement exists. Although we agree with the assessor on the importance of discussing the potential biological significance of the identified glycosylation differences, we are of the opinion that there is currently insufficient evidence in the literature to make scientifically grounded arguments on this matter.

*5) (Figure 3) On TZM-bl cells the T-SIV stock shown 55-fold lower infectivity compared to M-SIV. This is the reverse as to what was found on macaque CD4 T-cells where T-SIV showed a 6.5-fold higher infectivity than M-SIV? Needs addressing. Again, this should be considered in context of the results with CEMx174 R5 cells where infection between the 2 stocks appears to be similar (Figure 6). *

Thank you for raising these points. We like to remark that p27-content and infectivity are not absolutely linked (Narayan et al. 2023). Virus productions also contain empty or non-functional virions and the proportions can differ depending on the virus producer cell. Thus, T-SIV having a lower infectivity per ng 27 is likely a result of cell type dependent variation in the proportion of non-functional virions and does not represent an inconsistent result.

Kedhar Narayan, Jeongpill Choi, Shreyas S. Gujar, Aidan McGraw, Hasset Tibebe, Lilia Mei Bose, Caroline N. Arnette, Taisuke Izumi, Identifying Discrepancies in HIV Infectivity and Virion Maturation Using FRET-Based Detection and Quantification, bioRxiv 2023.12.25.573317; doi: https://doi.org/10.1101/2023.12.25.573317

*6) (Figure 5) The result may look cleaner if No Lectin value is subtracted from the cell lines carrying the lectin expression? *

Agreed. We have adapted the figure (Fig. 2D) as suggested.

*7) A much clearer introduction to CBA's would be beneficial. *

We agree that this would improve our manuscript and have expanded the introduction on carbohydrate binding agents (lines 204-210).

8) A concern I have is the presentation of data in Figures 6 and 7, especially given that the cell line used is the TZM-bl cell which has been shown to be 55-fold less infectible with T-SIV. Plotting the results as % infectivity on the same graph could be somewhat misleading. Two graphs one panel for M-SIV and one for T-SIV may be easier to follow. The CEMx174 cell line may have been a better choice as similarity to infection was found? But assume those experiments were not performed?

We kindly ask reviewer 3 to note that the virus input in these experiments has been normalized for equal infectivity, as described in the figure legend. An example demonstrating the comparability of results obtained using this method can be observed in Figure 2A, right panel, which justifies our approach. While we agree that conducting neutralization assays on additional cell types might be a valuable extension of the existing work, we chose here to use the TZM-bl cell line, the standard cell line for neutralization assays in the field of HIV research (Sarzotti-Kelsoe, 2013).

Sarzotti-Kelsoe M, Bailer RT, Turk E, Lin CL, Bilska M, Greene KM, Gao H, Todd CA, Ozaki DA, Seaman MS, Mascola JR, Montefiori DC. 2014. Optimization and validation of the tzm-bl assay for standardized assessments of neutralizing antibodies against hiv-1. J Immunol Methods, 409, 131-146. DOI:10.1016/j.jim.2013.11.022.

*9) I do feel the Discussion is extremely long and could be stream-lined to make it clearer and to the point. *

To follow your valuable suggestion, we reduced the length of the discussion by approximately one-third and eliminated sections that were less directly related to our results.

10) Materials and Methods section. Is the first section on Animal studies required. Could this not just be cited if it has been previously published.

While we agree that a manuscript should be as compact as possible, we decided to include this information to ensure complete transparency regarding our animal experiments. Additionally, many journals request this information prior to publication. Consequently, we opt to retain the text in its current form.

Minor comments *1) A) needs to be removed from Figure legend 1. *

Please note that we have not addressed this suggestion since Fig. 1 is now a multi-graph figure.

*2) Line 160 I assume this the result from (Fig 3A). *

That is correct. We have included the figure reference to make this line of text more clear.

*3) Line 255-257 Difficult sentence to understand. *

Thank you for making this point. This sentence has been removed in our efforts to shorten the discussion.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #3

Evidence, reproducibility and clarity

Summary

In this manuscript the authors have studied a variety of phenotype of two generated viral stocks of SIVmac239/316 molecular cloned virus generated in either CD4 T-cells (T-SIV) and macrophages (M-SIV). These stocks are isogenic apart from the cell type produced in and are identical in Env amino acid sequence. Stocks generated from the different cell-types were found to alter EndoH digestion, mannose profiling, infectivity, Env incorporation and lectin interactions and sensitivity to CBS and monkey sera neutralisation. The claim is that producer cell-type can influence SIV biological properties that associate with viral transmission, dissemination and inhibition.

Major comments

1. Line 85 "These substitutions facilitate efficient utilization of CCR5 in the absence or at very low levels of CD4 expression (Puffer et al.2002). This makes the molecular clone studied rather unique and the authors should aim to address this throughout. It does not take away from the results presented but should be addressed.
2. (Figure 1) State the predicted size differences between T-SIV and M-SIV stocks with EndoH digestion (and similar for all 3 runs?).
3. Line 95 Data should be shown. This describes infectivity and could incorporated within Figure 3 along with infection of the TZM-bl cells. Infectivity of T-SIV and M-SIV on primary CD4 T-cells and macrophages is of importance. This would only be possible I assume if p27 levels were measured at each time-point collected.
4. There is a lack of statistical difference in the results shown in Figure 2. I assume this is due to a single measurement, but can comment be made on likelihood of biological significance with such difference between values.
5. (Figure 3) On TZM-bl cells the T-SIV stock shown 55-fold lower infectivity compared to M-SIV. This is the reverse as to what was found on macaque CD4 T-cells where T-SIV showed a 6.5-fold higher infectivity than M-SIV? Needs addressing. Again, this should be considered in context of the results with CEMx174 R5 cells where infection between the 2 stocks appears to be similar (Figure 6).
6. (Figure 5) The result may look cleaner if No Lectin value is subtracted from the cell lines carrying the lectin expression?
7. A much clearer introduction to CBA's would be beneficial.
8. A concern I have is the presentation of data in Figures 6 and 7, especially given that the cell line used is the TZM-bl cell which has been shown to be 55-fold less infectible with T-SIV. Plotting the results as % infectivity on the same graph could be somewhat misleading. Two graphs one panel for M-SIV and one for T-SIV may be easier to follow. The CEMx174 cell line may have been a better choice as similarity to infection was found? But assume those experiments were not performed?
9. I do feel the Discussion is extremely long and could be stream-lined to make it clearer and to the point.
10. Materials and Methods section. Is the first section on Animal studies required. Could this not just be cited if it has been previously published.

Minor comments

1. A) needs to be removed from Figure legend 1.
2. Line 160 I assume this the result from (Fig 3A).
3. Line 255-257 Difficult sentence to understand.

Significance

This is an interesting study and where there is significance for understanding how HIV/SIV viral phenotypes (those associated with transmission and emergence following transmssion) can be influenced by the cell type infected and modifications to glycosylation profiling).

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

Assessment for Karsten et al

In this manuscript, Karsten et al described the biology of simian macrophage derived SIV and simian CD4+ T cells derived SIV have different levels and types of glycosylation in their particles. The authors attributed that these differences in glycosylation are related to SIV function (infection / spread).

It appears to this assessor that some of the supplementary data can be brought to the front as part of the main figures for presentation.

CURRENT figures 3, 5, and 7 can be combined into one figure.

Similarly, CUREENT Figure 4, and 6 can also be grouped.

Alternatively, incorporating additional approaches in each set of figures to tighten the claims.

I would support the manuscript for its eventual publication, but I believe several major (but achievable) amendments are needed.

Major suggestions

Karsten et al pitched their story as glycosylation of SIV from different primary cells are linked to different functionality in its title and abstract, yet the authors then declared in discussion (line 318) that establishing a direct link between Env glycosylation and viral functions is technically challenging and beyond the scope of the study. This assessor feels that authors need to decide whether current manuscript should be a descriptive study (which is more fitting for a less impactful journal) or a study with further mechanistic insights.

Table S1 is highly important and should be part of the main figure. Specifically, authors took the opportunity to highlight the differential % of sialic acid terminal glycans in line 133. The charge of the sialic acids would be simple mechanism for these M-SIV particle to attach. Authors should consider some of the described nano-luciferase based viral particle attachment assays used in HIV-glycan biology. Authors should be able to treat SIV (or SIV VLPs) with sialidase to quantify the role of sialic acids on binding.

As authors carefully pointed out (throughput the manuscript) that the identity (and biology) of the producer cells can have profound impacts on glycosylation events of viral particles that are being produced. This assessor was then interested to understand precisely how their simian PBMCs and monocytes derived macrophages were prepared. Additional details in M&M would be very helpful.

With the emphasis of cell type and glycosylation relationship, it is puzzling that authors would have chosen to use TZM-bl (artificially engineered cell line) and spinoculation (2hr to push the viruses down to cell surface with 870 x G force) in Figure 3 for comparison of M-SIV and T-SIV infectivity. To this assessor, this assay neglected the biological roles of SIV glycans. In context, 870 x G is ~150x higher than most human can withstand.

Using a single antibody DA6 (in Figure 4, cited Edinger 2000) for Env incorporation estimation via Western seems to be crude and inadequate, even in the context of isogenic virus clone. As authors pointed out, different levels of glycosylation can affect protein folding, therefore affecting Env incorporation. By the same argument, differentially glycosylated Env protein can also impact on the ability of 'epitopes within Env protein' to be recognised by Ab. Therefore, virion incorporation of Env might not be affected, but just the detectability by a specific Ab. Western evaluation with a panel of anti-Env antibodies will help. Furthermore, quantitative proteomics coupling with glycomics would be highly useful.

It is understood that T-SIV were pooled from supernatant derived from 9 animals of PBMCs. Levels of p27 production (presumed as particles but including free p27 in reality) from each animal donor should be listed in supplement. Similar types of details should be made available for M-SIV that were derived from 8 animal donors of macrophages. qPCR estimations on the levels of viral particles production in T-SIV and M-SIV from primary cell culture amplifications appear to be already available, such information should be included in supplementary to strengthen the authors' estimated / relationships amongst glycosylation, virion Env incorporation levels, and viral particle productions are carefully controlled.

Non-glycan biologists generally do not appreciate some of the fine details in glycan biology. The T-SIV and M-SIV system is a great model system to decode some of the functionality of glycan biology. The current team should have (in my opinion) a clear graphic representation on describing what types of different glycans in T-SIV and M-SIV are likely to contribute to the potential differences in biological outcomes. Such incorporation will guide non-glycan biologists to better appreciate the focus and the directions of authors, thereby further improving the citation of this work when it is published after peer reviewed. Importantly, focusing a specific question to be addressed may help to consolidate effort to accelerate publication of this work. A beautiful story line, just need to cross many 't' and dot a few 'i' in my view.

Most primate centres often incorporate transcriptomic studies in their animal works. It will be helpful for the audience if the authors could provide additional transcriptomic data (with a focus on glycosylation related genes) of simian CD4+ T cells, simian macrophages, SIV infected simian CD4+ T cells, and SIV infected simian macrophages. These data will improve the comprehensiveness of this study (and should not require any major wet-lab studies) and add weight on the arguments of the authors.

Significance

General Assessment - The biological significance system these authors possess is highly valuable in virology and will reveal significant insights in the functions of glycans in infectious diseases. Authors are generally (in my opinion) correct with the big picture impacts / contributions of glycan biology. Presented experimentations need to be tighter controlled to avoid over-interpretations. A tighter focus of research question (or claim) will reduce levels of extra work prior to publication.

Advance - level of advance will be high regarding the role of glycan in biology and infectious diseases. The T-SIV and M-SIV system is a naturally relevant system with many prior works that lay the foundation to understand viral glycan biology.

Audience - with the right pitch and proper explanations, general audience will be highly interested. At the end of the day, glycan biology is shared amongst all living cells.

My expertise is in virology, particular in HIV. I also have a strong interest in glycan biology. I described the first glycan-glycan interaction in viral pathogenesis recently, explaining how these glycan-based interaction serves as a molecular Velcro for attachment, likely a shared mechanism in virology and biology in general.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

This is interesting data and will add to the relatively sparse literature in this area, but there are several issues limiting its utility and accessibility, and other data, experiments or explanations are needed to address these issues. The other major issue is that there is no direct experimental evidence showing specifically that the glycosylation being described, particularly the differences in T-SIV vs. M-SIV, as there had been in some of the prior studies. Specifically removing or adding particular glycans, and then showing changed effects on viral entry - particularly in primary cells - would add substantively to the findings and make this a much stronger study. In addition, there are a number of other more specific and/or minor changes that could be made to enhance the value of this manuscript.

1. As noted above, the differences in glycosylation are difficult to understand without more background and perhaps a figure, but it is also not clear that the changes described between lines 119 and 137 are biologically or statistically meaningful. For example, does it matter that more M-SIV virions have glycans with four antennae than T-SIV? Are there other data that show this or could experiments be done to specifically cleave these glycans at certain points to reduce their complexity and show that the infectivity differences between M-SIV and T-SIV disappear? Further, it is difficult to confirm the statement on lines 124-125 that "profiles of complex-type N-glycans differed between the two viruses (Fig. 2C)", as no statistical tests were done to compare the glycosylation being detailed in the M-SIV and T-SIV. It is more appropriate to make note that there are minor between M-SIV and T-SIV or run specific statistical tests on the data.
2. On line 161, the authors note that the results showing that "the virus-producing cell has a broader impact on SIV infectivity beyond its influence on Env incorporation." This is certainly one possibility that is suggested by these and prior data, there are also other possibilities. For example, the impact of Env is not linear and perhaps a certain number of Env need to be engaged, creating some kind of threshold effect that means that the virions with fewer Env just have less infectivity. Given that there is significant data that virions are generated in different locations in macrophages and T-cells, this could also be a function of which specific membrane areas in different cell types that Env embeds in, or it could be something else associated with Env that is not cell type specific.
3. For the studies in Figure 5, looking at the direct vs. indirect infectivity, it is not clear why CEMx174 R5 cells (a T-cell/B-cell hybrid line) were used instead of primary macrophages or T-cells, or macrophage or T-cell lines or a fully agnostic cell type. This would be more convincing tested on primary cells, or at least comparing in a myeloid lineage line as well.
4. In Figure 6, it is not clear that VSV-G pseudotyped virus is an appropriate control, as it enters via the acidified endosome pathway and not via similar processes as the T- and M-SIV derived virions. While this may show that the glycans can bind to CBA to inhibit entry, it could also mean that the general process of endocytosis is not as susceptible to CBA inhibition and this difference in pathways should be noted as a caveat.
5. A very large number of cell lines were used, and it is not clear why experiments were done using so many different indicator or target lines, instead of performing most assays in a single line or set of lines so that they are comparable across experiments. Some discussion of the rationale for this would be helpful.

Minor comments

1. Inclusion of the p27 data characterizing the amount of virus in M-SIV and T-SIV stocks (line 95) should be shown as at least a supplemental figure or could easily be added to figure 1.
2. The figures are relatively thin and could be combined with other figures to better connect the experiments. For instance, Figure 1 could serve as panel A for what is currently listed as Figure 2 because it is a preliminary data to the experiments in Figure 2.
3. The authors should include quantitation of the Western blot data in Figure 1 in an adjoining graph.
4. The legend states that the results in Figure 7 were obtained from two independent experiments (line 778), each with 3 technical replicates. As this represents only 2 biological replicates, and the experiments were performed in easily accessible cells (TZM-bl), they should be performed 1 - 3 more times to provide a more appropriate and robust data set for statistical analysis.

Significance

The manuscript Karsten et. al., 2024, discusses the differences in infectivity resulting from the cellular origin of SIV virions, specifically comparing virus generated in macrophages and T-cells. The primary focus of this comparison is the differences in glycosylation of the Env proteins on the virions with a T-cell origin (T-SIV) and a macrophage origin (M-SIV). These studies are expansions of a small number of prior publications that focus on this area and are cited extensively. The major innovation in this study were that the differences in glycan composition were assessed using xCGE-LIF glycan profiling, which is more detailed than the glycan profiling methods in the prior studies. Then the T-SIV and M-SIV were compared for their susceptibility to distinct carbohydrate binding agents, either ulex europaeus agglutinin (UEA), cyanovirin-N (CV-N), or galanthus nivalis agglutinin (GNA). These agents each bind different glycans, so differences in the capacity of these agents to bind T-SIV vs. M-SIV suggest confirming the presence of different types of glycans. Similar studies with serum from SIV infected macaques, the results suggesting some differences in susceptibility to neutralization in viruses of different cellular origin.

One primary issue is that many readers will not be familiar with the different types of glycosylation, the differences between subtypes of glycans, different branches, etc ... and what the biological relevance of these differences is to viral infection and/or immune activity. For example, the discussion refers to M5, M6, M8 and M9 on T-SIV vs. M-SIV (lines 234 - 5) but knowledge of these is not universal and there is no background to place these statements in context and understand why this might matter. To address this, additional language is needed, as well as the addition of a figure that helps to visualize the different glycans being discussed. Adding this to the beginning as part of the introduction, or at the end as a summary of the findings in the paper, would increase accessibility for a broader audience.

Transparent Peer Review

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

We would like to thank both reviewers for their thorough and constructive evaluation and comments on our manuscript. Following their suggestions, we have edited our manuscript to address all criticisms and comments from them. We hope that, with these introduced changes, this manuscript will be suitable for publication in an appropriate journal. Detailed point-to-point responses are shown below.

__Point-to-point response to reviewers’ critiques: __

Reviewer 1

Summary: Stadler and colleagues characterized CFAP410, using molecular structural biology, biophysics, biochemistry, cellular imaging and genetic engineering, to address molecular mechanism of ciliopathies caused by defect of this protein. They crystalized the C-terminal domain of CFAP410 and its homologue, from Trypanosoma, human and Chlamydomonas. All of them take tetrameric complex formation with four bundle helices in the center. Then they mutated highly conserved residues, L219 and L224, which are located at the helix bundle, and characterized its biophysical properties, demonstrating oligomerization defect and, in case of L224 mutation, collapse of secondary structure as well. This was supported by molecular dynamics. Next they examined the effect of these mutation in vivo using Trypanosoma brucei. They visualized the localization of CFAP410 at the postrior cell tip and its transfer to the basal body. By mutation, transfer of CFAP410 is prohibited and cytokinesis defect occurs. The experiments are logically designed and the results are clearly and convincingly stated. Undoubtedly this work deserves publication after minor revision of the manuscript.

Strength:* This is a compact paper, clearly stating the biological aim and experiments designed for that.

Limitation: While it is undoubtedly proved that the C-terminal domain of CFAP410 forms tetramer and L219 and L224 are key residues and that mutation of these residues causes severe defect at the localization of CFAP410, it is an overstatement to conclude the tetramer formation is essential for the localization. The authors experiments cannot exclude the possibility of another consequence of mutation (different from origomerization), which is the cause of abnormal localization.

*

: Indeed, the reason for the disrupted localization of the disease-causing mutation L224P to the basal body could be caused by mis-folding of TbCFAP410 if that is the only mutation we had tested. However, as shown in our results, the point mutation A267E of TbCFAP410 that breaks the tetramer into two dimers (so as Hs/Cr-CFAP410-A219E) did not change its folding or structural stability. Nevertheless, TbCFAP410-A267E lost its ability to localize to the basal body effectively (Figure 6A). It suggests that even a folded dimer is insufficient to correctly localize to the basal body. Therefore, we can confidently conclude that the fully assembled tetramer of CFAP410 is required for its localization to the basal body.

* Minor points: p.7: It would be interesting, if the authors, beside MD, attempt to predict conformation of the mutants using Alphafold2. *

: We tried to predict conformational changes of the point mutations of CFAP410 via AlphaFold2, but did not observe significant changes in the generated models. It was not so surprising to us though, because, as shown previously, although experimentally point mutations induced complete unfolding of some proteins, the AlphaFold2 models of the same point mutants folded similarly to the wild-type crystal structures (Buel & Walters, 2022, Nat. Struct. Mol. Biol. 29:1-2.)

* p.7-8: A cartoon to describe what proteins exist between the posterior tip and the basal body will help readers to understand. Do the authors have any thought how CFAP410 is transported to the basal body? *

: We thank the reviewer for the suggestion and have revised Figure 7 to illustrate the two distinct localization sites of CFAP410 in T. brucei.

We do not know the mechanism through which CFAP410 localizes to the posterior cell tip and the basal body. Both structures are microtubule based, with the basal body consisting of a barrel of microtubule triplets and the posterior cell tip is the site at which the ends of the microtubules which form the sub-pellicular array in T. brucei are located. This suggests that CFAP410 could interact with microtubules or microtubule binding proteins in these locations. Additionally, in human cells CFAP410 appears to interact with NEK1 and the equivalent interaction in trypanosomes may be important for CFAP410 localization. We have refined our ideas on this in the Discussion section (p.11): “It was shown previously that the L224P mutant of CFAP410 abolishes its interaction with NEK1 [4]. Given that the mutant L224P disassembles the tetramer of CFAP410-CTD (Figure 3), the tetrameric assembly of CFAP410 seems to play an essential role in its interaction with NEK1. Therefore, the disrupted location of TbCFAP410-L272P to the basal body we observed here could be attributed to its abolished interaction with the trypanosome equivalent of NEK1 as occurs in human cells. However, we cannot exclude another possibility that CFAP410 localizes to the basal body by interacting with an unidentified anchoring target there and NEK1 is subsequently recruited via its binding to CFAP410.”

It is worth mentioning though that the localization of TbCFAP410 to the posterior cell tip has only been reported in T. brucei and no other cellular localization sites have been reported for CFAP410 orthologs in other organisms including human. Moreover, in the genome-wide protein tagging project TrypTag many other proteins were found to localize to both the posterior tip and another site in the cell, including the basal body (Billington et al, 2023, Nat. Microbiol. 8: 533-547). The following paragraph discussing about this has been added to the Discussion section (p. 10): “Notably, recent genome-wide protein localizations revealed that the posterior cell tip in T. brucei has unexpectedly high complexity and contains many proteins that also localize to other organelles _[16]_. The tip may thus serve as a “moonlighting” site for those proteins. However, the extra localization site of CFAP410 at the cell tip has only been reported in T. brucei and no other cellular localization sites have been observed for CFAP410 in any other organisms.”

* p.10: Can the authors define "NTD linker" precisely (from which to which residues)? *

: We have defined the sequence range of both the NTD (aa1-160) and NTD-linker (aa1-254) of TbCFAP410 both in the text (p. 9 & p. 10) and legend of Figure 6.

* p.12: Alphafold-multimer may help to have information, which part of CFAP410 is likely interface to NEK1 and SPATA7. *

: We attempted to predict how CFAP410 interacts with NEK1 or SPATA7 by Alphafold-multimer. Results of the former prediction, which are consistent with previous studies (Gregorczyk et al, 2023, Life Sci. Alliance), have been added as a new figure (Figure S4). However, no convincing results were obtained for the latter pair.

* p.16: The paragraph starting with "DSF measurements were ..." is probably not necessary. Figure 4 caption: "HsCFP0-CTD" should be defined precisely. Or is it a typo of HsCFAP410-CTD? *

: We thank the reviewer for pointing out this mistake. This paragraph has been removed.

* Figure 5ad, Figure 6: blue is not defined.

*

: We have now defined blue in the figure legends for 5A, D and 6A. These cells have been counterstained with the DNA stain Hoechst 33342 to highlight the nucleus and kinetoplast (mitochondrial DNA) and this is the blue element in the images.

* Advances: This interdisciplenary work nicely characterized CFAP410 at atomic, molecular and cellular levels and acquired insight of its functional mechanism. *

: We appreciate the reviewer’s constructive comments and positive feedback.

Reviewer 2

*In this paper, the authors described 3 crystal structures of the CTD of CFAP410 from 3 different species. They explored the phenotype mutation L224P in humans which causes ciliopathies using in vitro and in vivo analysis. They were able to explain the oligomerization role of the L224P mutations and its importance for correct localization. In addition, using their structure, they also found A219 as an important residue for tetramerization as well. The paper is well-written and easier to read. *

* __There are some minor concerns: __

1. Fig. 5C: Why would the line reduce to lower value at 24hr, 48hr for both non-induced and induced one. *

: During routine culture of T. brucei, we had to split the cells to ensure they do not overgrow and are maintained in log phase growth. The reduction in value at each time point represents the cell splitting event and gives these characteristic “sawtooth” graphs for cell growth. At each time point the cell density is measured and then the cells (both non-induced and induced) are diluted to the same cell density, in this case 2x106 cells/ml, and grown for a further 24 hours before the next measurement.

*2. The authors wrote "Although we observed only little change in the average distance from the posterior cell tip to kinetoplast in 1K1N cytoskeletons after induction, there was a substantial increase in the range of these measurements, with cytoskeletons observed having a more reduced or increased distance from the kinetoplast to the posterior cell tip (Figure 5F)." *

*Why not back the wide range with a standard deviation calculation in the figure caption of 5F or display the std dev directly in the figure if it looks good? Also, the author can include legends for color dots in the figure as Replicate 1,2,3 for easy reading/comprehension. *

: We have included the standard deviations for each of the replicates in 5F in the figure legend. The spread of data is shown in the figure already with the individual points and the overlay of the standard deviation was not clear when we tried it. We have now included a legend for the dots in the figure as suggested.

*3.The information about the construct of Trypanosoma used for Figure 6 is not described at all. What exactly is the region of the construct of the NTD and NTD-linker? *

: The following sentence has been added to the legend of Figure 6: “Except for NTD (aa1-168) and NTD-linker (aa1-254), all other constructs are full-length proteins.”

*4. The author wrote "In the mutant A267E, the mNG::CFAP410 signal was exclusively found at the posterior of most cytoskeletons (63.5%), while full-length TbCFAP410-L272P and the two CTD-lacking constructs, NTD and NTD-linker ... This suggests that both the presence of the CTD and the integrity of its oligomerization are essential for the interaction of TbCFAP410 with the basal body and posterior cell tip." *

*This statement needs to be revised a bit. First, seems like the tetramerization is important for the localization, not dimerization. Second, is there any evidence that full-length TbCFAP410-L272P folds properly? Without the evidence that the NTD and NTD-linker region can fold properly in both full-length TbCFAP410-L272P and CTD truncation, it is not possible to exclude that the N-terminal is essential for the localization as well. *

: Thanks for the comments. We have managed to express and purify TbCFAP410-NTD, which shows that it folds properly on its own. We further checked whether NTD directly interacts with CTD, but ended up with negative results. The following part has been added to the last paragraph of the Results section to address the reviewer’s question. “We found that TbCFAP410-NTD folded properly on its own when expressed in bacteria, and no direct interaction between NTD and CTD was detected (data now shown). It suggests that the two structural modules of TbCFAP410 that are connected by a long disordered linker are folded independently. Therefore, we conclude that the localization of TbCFAP410 to the basal body and the posterior cell tip requires the CTD and its ability to oligomerize.”

* Small things: 1. It is worth define 1K1N, 2K1N and 2kK2N stages in the text for reader not in Trypanosoma field.

*: We have explained the KN nomenclature, and how during the cell cycle the kinetoplast and nucleus are duplicated and segregated in a defined order.

*2.Fig. 5C: label Non-induced black & induced orange as legends directly in the figure so readers don't have to read the Figure captions.

*: This has been done as suggested.

*3. In Methods, there is the part about DSF measurement in the Molecular Dynamic (MD) simulations *

: This paragraph is unnecessary and has been removed. We thank the reviewer for pointing it out.

4.* Abbreviation not defined: DSF (in Methods/MD simulations). Also, the molecular dynamics phrase appears well before the abbreviation MD in the Methods section. *

: As stated above, this paragraph has been removed.

*

Reviewer #2 (Significance (Required)):

Overall I found the methodology and results of the paper solid, and the interpretation and conclusion are sound. *

*The paper addresses the molecular mechanism of a mutation in CFAP410 resulting in severe spondylometaphyseal dysplasia, axial. *

* The audience of the paper should be the cilia field but also the paper is also good for other researchers as the paper is easy to read.*

: We appreciate the reviewer’s constructive feedback and positive evaluation.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

In this paper, the authors described 3 crystal structures of the CTD of CFAP410 from 3 different species. They explored the phenotype mutation L224P in humans which causes ciliopathies using in vitro and in vivo analysis. They were able to explain the oligomerization role of the L224P mutations and its importance for correct localization. In addition, using their structure, they also found A219 as an important residue for tetramerization as well. The paper is well-written and easier to read.

There are some minor concerns:

1. Fig. 5C: Why would the line reduce to lower value at 24hr, 48hr for both non-induced and induced one.
2. The authors wrote "Although we observed only little change in the average distance from the posterior cell tip to kinetoplast in 1K1N cytoskeletons after induction, there was a substantial increase in the range of these measurements, with cytoskeletons observed having a more reduced or increased distance from the kinetoplast to the posterior cell tip (Figure 5F)."

Why not back the wide range with a standard deviation calculation in the figure caption of 5F or display the std dev directly in the figure if it looks good? Also, the author can include legends for color dots in the figure as Replicate 1,2,3 for easy reading/comprehension. 3. The information about the construct of Trypanosoma used for Figure 6 is not described at all. What exactly is the region of the construct of the NTD and NTD-linker? 4. The author wrote "In the mutant A267E, the mNG::CFAP410 signal was exclusively found at the posterior of most cytoskeletons (63.5%), while full-length TbCFAP410-L272P and the two CTD-lacking constructs, NTD and NTD-linker ... This suggests that both the presence of the CTD and the integrity of its oligomerization are essential for the interaction of TbCFAP410 with the basal body and posterior cell tip."

This statement needs to be revised a bit. First, seems like the tetramerization is important for the localization, not dimerization. Second, is there any evidence that full-length TbCFAP410-L272P folds properly? Without the evidence that the NTD and NTD-linker region can fold properly in both full-length TbCFAP410-L272P and CTD truncation, it is not possible to exclude that the N-terminal is essential for the localization as well.

Small things:

1. It is worth define 1K1N, 2K1N and 2kK2N stages in the text for reader not in Trypanosoma field.
2. Fig. 5C: label Non-induced black & induced orange as legends directly in the figure so readers don't have to read the Figure captions.
3. In Methods, there is the part about DSF measurement in the Molecular Dynamic (MD) simulations
4. Abbreviation not defined: DSF (in Methods/MD simulations). Also, the molecular dynamics phrase appears well before the abbreviation MD in the Methods section.

Significance

Overall I found the methodology and results of the paper solid, and the interpretation and conclusion are sound.

The paper addresses the molecular mechanism of a mutation in CFAP410 resulting in severe spondylometaphyseal dysplasia, axial.

The audience of the paper should be the cilia field but also the paper is also good for other researchers as the paper is easy to read.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

Summary: Stadler and colleagues characterized CFAP410, using molecular structural biology, biophysics, biochemistry, cellular imaging and genetic engineering, to address molecular mechanism of ciliopathies caused by defect of this protein. They crystalized the C-terminal domain of CFAP410 and its homologue, from Trypanosoma, human and Chlamydomonas. All of them take tetrameric complex formation with four bundle helices in the center. Then they mutated highly conserved residues, L219 and L224, which are located at the helix bundle, and characterized its biophysical properties, demonstrating oligomerization defect and, in case of L224 mutation, collapse of secondary structure as well. This was supported by molecular dynamics. Next they examined the effect of these mutation in vivo using Trypanosoma brucei. They visualized the localization of CFAP410 at the postrior cell tip and its transfer to the basal body. By mutation, transfer of CFAP410 is prohibited and cytokinesis defect occurs.

The experiments are logically designed and the results are clearly and convincingly stated. Undoubtedly this work deserves publication after minor revision of the manuscript.

Significance

strength: This is a compact paper, clearly stating the biological aim and experiments designed for that.

Limitation: While it is undoubtedly proved that the C-terminal domain of CFAP410 forms tetramer and L219 and L224 are key residues and that mutation of these residues causes severe defect at the localization of CFAP410, it is an overstatement to conclude the tetramer formation is essential for the localization. The authors experiments cannot exclude the possibility of another consequence of mutation (different from origomerization), which is the cause of abnormal localization.

Minor points:

p.7: It would be interesting, if the authors, beside MD, attempt to predict conformation of the mutants using Alphafold2.

p.7-8: A cartoon to describe what proteins exist between the posterior tip and the basal body will help readers to understand. Do the authors have any thought how CFAP410 is transported to the basal body?

p.10: Can the authors define "NTD linker" precisely (from which to which residues)?

p.12: Alphafold-multimer may help to have information, which part of CFAP410 is likely interface to NEK1 and SPATA7.

p.16: The paragraph starting with "DSF measurements were ..." is probably not necessary.

Figure 4 caption: "HsCFP0-CTD" should be defined precisely. Or is it a typo of HsCFAP410-CTD?

Figure 5ad, Figure 6: blue is not defined.

advances: This interdisciplenary work nicely characterized CFAP410 at atomic, molecular and cellular levels and acquired insight of its functional mechanism.

audience: cilia community

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

1. General Statements

Regarding significance, we would like to highlight that the main finding and breakthrough of our manuscript is the discovery that intronic polyadenylation (IPA) isoforms are a source of microproteins (indeed, IPA was not known to induce sORF-encoded microproteins). We make the proof of principle of this concept (called miP-5’UTR-IPA) and of its functional relevance for one gene (PRKAR1B).

A second finding of this study is that IPA (including miP-5’UTR-IPA) isoforms are widely upregulated in cell response to cisplatin, and therefore we show the functional relevance of miP-5’UTR-IPA isoforms in this biological context.

Regarding the generality of the miP-5’UTR-IPA concept, we provide evidence that many genes generate miP-5’UTR-IPA isoforms, by crossing our 3’-seq data with available Ribo-Seq and mass spectrometry datasets, which were generated without cisplatin treatment. Also, the miP-5’UTR-IPA isoforms of PHF20 and PRKAR1B are detected both in the presence and absence of cisplatin. Thus, the novel concept of microprotein-coding IPA isoforms opens wide perspectives, way beyond cisplatin response.

2. Description of the planned revisions

REVIEWER #1

Evidence, reproducibility and clarity

Microporteins originating from coding and non-coding transcript are increasingly understood to control various cellular processes. In the present study, the authors investigated whether intronic polyadenylation (IPA) contributes to the formation of transcript isoforms encoding microproteins. Using genotoxic stress by cisplatin as a model in cell cultures, the authors detect abundant IPA. IPA in a subset of such transcripts leads to short 5'UTR transcript isoforms that are poorly associated with heavy polysomes and encode microproteins. For PRKAR1B, they demonstrate the expression of a corresponding microprotein and a function in modulating the cisplatin response. Based on depletion experiments of FANCD2 and STX1, the authors propose that impaired transcription processivity after cisplatin is one mechanism leading to IPA and microprotein production.

While this is an interesting manuscript, I felt the support for the claimed generalization falls a bit short.

Our response: The generality of the miP-5’UTR-IPA concept is supported by the large-scale analysis that we presented (Fig. 6): indeed, by crossing our 3’-seq data with Ribo-Seq and MS data (both of which originate from multiple cell types and tissues), we identified 156 genes with cisplatin-regulated miP-5’UTR-IPA isoforms. To strengthen this part and highlight the generality of the miP-5’UTR-IPA concept, we will provide the cell type/ tissue distribution of our set of 156 miP-5’UTR-IPA isoforms, by exploiting available 3’-seq datasets from various cells/tissues. (Please also see major point 1 below.)

Major:

1. • If I see it correctly, the authors mainly refer to existing riboSeq data and evidence from mass spectrometry/proteomics to infer the generality of the mechanism (beyond PRKAR1B). It is important to back this up with further experiments and validate this for the set-up used in this manuscript. This concerns the existence of the microproteins but also the downstream functional impact. Our response: In our study, we make the proof of principle of miP-5’UTR-IPA (that is, a microprotein-encoding IPA isoform) for the PRKAR1B gene and its sORF#2 (microprotein detection by WB and IF, functional evidence by siRNA and CRISPR of IPA site and sORF initiation codon). If I understand well (also based on minor point 3 below), this reviewer is requesting further evidence of microprotein existence (in addition to Ribo-Seq and mass spectrometry [MS] data) and function, for a second IPA-derived sORF that we study in this manuscript (either PHF20 sORF or PRKAR1B sORF#1). To the best of our knowledge, a proof of principle for a new concept is usually done on a single gene. Nevertheless, for the miP-5’UTR-IPA isoform of PHF20*, we already provided evidence for its function by using siRNAs (Fig. 3A-C) and for its translation by polysome profiling (Fig. 4C) in addition to Ribo-Seq and MS evidence (Fig. 4A). The fact that for PHF20 we did not detect the transfected Flag-tagged microprotein in HEK cells could be due to several reasons (as discussed on page 16); __we will __try this approach again with different biological conditions (cell lines, stress) or construct designs (as the sORF context may be important).

2. Also, I wonder is this limited to cisplatin-induced genotoxic stress and the specific cell line used or is this a more global mechanism?*

Our response: We provided evidence of IPA isoform regulation by cisplatin in two lung cancer cell lines (A549 and H358; Fig. 1A-B) but we agree that our analyses of miP-5’UTR-IPA were mainly done in A549 cells. We will: (i) clarify that we detected the miP-5’UTR-IPA isoforms of PRKAR1B and PHF20 in A549 cells (total cytosol and light polysomes) both in the presence and absence of cisplatin (Fig. 2D, 3A and S3D); (ii) add RT-qPCR validation of their cisplatin regulation in H358 cells; (iii) try to detect the PRKAR1B-encoded microprotein in a second cell line (Fig. 4); (iv) test the impact of PRKAR1B and PHF20 miP-5’UTR-IPA isoforms on cell survival in a second cell line and with a second genotoxic agent; (v) clarify in Fig. 6 that miP-5’UTR-IPA isoforms are regulated by cisplatin in both A549 and H358 cells (our 3’-seq data) and that the Ribo-Seq and MS datasets supporting their translation originate from multiple cell types and tissues without cisplatin treatment; and (vi) provide the cell type/ tissue distribution of our set of 156 miP-5’UTR-IPA isoforms (by exploiting available 3’-seq datasets from various cells/tissues).

Minor:

1. • While the rest of the paper reads well, the abstract could be improved/simplified to increase accessibility* Our response: We will improve the abstract.

Page 11: Pertaining to Figure 3 and the functional impact: The authors analyze the IPA effect by probing cell viability and cell survival. It would be important to define the effects in further detail, as the mere regulation of cell cycle and/or apoptosis could also result in such outcome (which is then not necessarily a direct cisplatin response). Does this also impact the response to other genotoxic stress (also pertains to the effects studied and shown in Fig. 5)?

Our response: Because cisplatin effects on cell growth are usually mediated by effects on cell cycle and cell death, we will determine which aspect is impacted by PRKAR1B and PHF20 miP-5’UTR-IPA isoforms, by carrying out FACS analysis of PI/BrdU and Annexin V (both in the presence and absence of cisplatin). As mentioned in major point 2, we will also test the impact of these isoforms on cell survival to a second genotoxic agent.

Page 12 concerning the microprotein expression: the authors refer to data from other resources to claim that the microproteins are expressed, however they fail to demonstrate this for their setup (at least for 2 out of three they study here). I think this is a weak point as it does not directly support the general claim.

Our response: Please see major point 1 above.

Also, I did not understand what the authors intended to demonstrate with the immunoflourescence (Fig. 4E). What should a defined nuclear expression imply versus the diffuse staining throughout after cisplatin? How does this relate to the functional effects?

Our response: We included in Fig. 4E the observation that the subcellular localization of the PRKAR1B-encoded micropotein is altered in response to cisplatin, because this supports the notion that this micropotein plays a role in cell response to cisplatin. We can remove this data if requested.

Page 13/Fig. 5E: the different clones of the mATG show very high variability. To my understanding it is difficult to draw a clear conclusion from this heterogeneity.

Our response: The statistical analysis shows a significant difference between the mATG and Control groups (p Page 15 on the mechanism: SETX has been demonstrated to control poly(A) site choice (PMID: 21700224, 32976578). However the quantitative role of SETX in poly(A) site choice regulation (compared to other regulators) seems to be rather marginal and not strictly unidirectional, i.e after SETX depletion also longer transcript isoforms can be detected (PMID: 32976578). How does this relate to the proposed mechanism of SETX-dependent processivity? Interestingly, from PMID: 32976578 it also appears that PRKAR1A has a 5'UTR poly(A) site that is regulated in a SETX-dependent manner.

Our response: We will add in the discussion statements that (i) the role of SETX in cisplatin regulation of IPA:LE isoform ratio and processivity might be different from its role in APA regulation in the absence of genotoxic treatment (citing PMID 32976578; keeping in mind that we did not compare them side by side on a genome-wide scale) and (ii) PRKAR1A seems to have a 5'UTR poly(A) site regulated by SETX in TREND-DB (PMID 32976578).

• Page 16, discussion first paragraph. While refs 1-4 are nice reviews that could be quoted here a study that appeared later represents the most comprehensive analysis to date covering the different facets from transcription to RNA processing and the resulting impact on poly(A) site choice (PMID: 30552333).*

Our response: We will cite PMID 30552333 and 32976578 as resources of APA regulation by various regulators of gene expression (keeping in mind, however, that for most factors these studies do not exclude indirect effects).

Significance

This could be a very significant report, provided the generality of the claims and mechanistic insigths are further strengthend.

Overall it targets a rather specialized readership. This could be improved by simplifing the abstract, additional experimental evidence for the generality of the proposed mechanism, and a stringent rewording of the main text drawing a clear line, omitting unnecessary details and focussing on the novel findings.

Our response: Please see our responses above. In addition, we will reword the main text where necessary.

REVIEWER #2

Evidence, reproducibility and clarity

Summary:

*In this manuscript, Devaux et al. report that the anti-cancer drug cisplatin upregulates intronic polyadenylation (IPA) isoforms in non-small cell lung cancer cell lines. Their finding was based on 3' end sequencing and long-read sequencing. Through polysome profiling they confirmed that many of the IPA isoforms are translated, despite being inefficient in most cases. *

Our response: There is some misunderstanding here. We will clarify in the text that inefficient association with heavy polysomes is observed for a minority (not the majority) of IPA isoforms. For this, in Fig. 2B and S2A, we will add the information that for the majority of IPA sites, the IPA:LE ratio is not significantly different (neither up or down) between total cytosol and heavy polysomes.

They validated functions of IPA isoforms from two genes, PHF20 and PRKAR1B, in cell survival upon cisplatin treatment, based on an array of methods, including siRNA knockdown, CRISPR knockout of IPA polyA site, and CRISPR mutation of the start codon. They further found that FANCD2 and Senataxin can regulate cisplatin-mediated IPA activation. The authors advocate a new paradigm of expression of IPA-encoding microproteins in cisplatin-treated cells.

Our response: We would like to point out that our data indicate that cisplatin upregulation of the IPA:LE isoform ratio is mediated at least in part by an inhibition of transcription processivity (explaining the decrease of LE isoforms), and that we do not claim an ‘IPA activation’ (that is, enhanced used of IPA sites) by cisplatin. This remark is also related to major point 1 below.

Major comments:

1. • While the phenomenon of IPA isoform upregulation by Cisplatin is quite convincing, the underlying mechanism is largely elusive. The authors indicated processivity as a potential mechanism and the effects of FACD2 and Senataxin appear in line with this hypothesis. However, they cannot rule out other possibilities based on the data presented in the manuscript. For example, it is not clear if the elongation rate of Pol II (distinct from its processivity) or nuclear RNA degradation is affected by cisplatin, which could also lead to increased expression of IPA isoforms. In addition, enhanced 3' end processing activity has been previously shown to activate IPA sites. Therefore, the underlying mechanism is mostly speculative. Our response: As explained on page 14, the reason why we focused on transcription processivity is that the cisplatin-induced upregulation of the IPA:LE isoform ratio was enriched in long genes and was accompanied by a decrease of LE isoform levels. Importantly, our data (e.g., for the PHF20 and PRKAR1B genes) indicate that the cisplatin-induced decrease of processivity explains –at least in part– the selective decrease of the LE but not IPA isoform levels and therefore the increase of the IPA:LE isoform ratio; we will clarify this in the manuscript (on pages 14 and 15). Our data also show that cisplatin effects on both processivity and IPA:LE isoform ratio are dependent on FANCD2 and SETX. We agree with the reviewer that we cannot exclude that IPA:LE isoform ratio upregulation by cisplatin might also be mediated in part by additional mechanisms (e.g.*, ‘factors involved in cleavage/ polyadenylation, splicing, transcription elongation and termination, and epigenetic marks’, as mentioned in the discussion on page 16) and we will add nuclear RNA degradation to the list of potential factors. However, we want to emphasize that the role of processivity is not speculative.

The authors used the polysome:cytosolic ratio to indicate translational efficiency. However, because the CDS size affects the number of ribosomes per mRNA, the translational efficiency should be based on polysome:cytosolic ratio normalized to CDS size. Ideally, the authors should calculate number of ribosome per transcript based on monosome, light polysome and heavy polysome.

Our response: We cannot normalize ribosome number by CDS size because (i) heavy polysomes are not a precise number of ribosomes and (ii) sORFs are not annotated as CDS.

The functions of PHF20 and PRKAR1B IPA isoforms are based on knockdown or knockout mutations. Because of its gain-of-function property, overexpression of the isoforms in cisplatin-treated cells would be necessary to definitively confirm their funcitons.

Our response: For PRKAR1B sORF#2, we ____will carry out overexpression of the sORF microprotein in A549 cells and CRISPR clones and analyze its effects on cell growth and cisplatin survival. We have appropriate constructs for this.

Minor:

1. • Fig. 1H, the numbers of IPA and LE transcripts should be provided. The statistical significance for the difference should also be included.* Our response: The numbers of IPA and LE transcripts were provided in Fig. S1I and we will provide the statistical significance (which is good), as requested.

Fig. 1I, the image should be accompanied with fold difference as indicated in the text. Some statistics for difference between vehicle only and CisPt only is necessary.

Our response: We will indicate the fold differences and provide the statistical significance, as requested.

• Fig. 6, the authors did data mining of ribo-seq data and mass-spec data and identified 156 genes whose IPA isoforms have potentials of protein expression. The enriched GO terms for the 156 IPA genes are different than the overall IPA isoforms shown in Fig. 1C. Does this mean some genes, like those in DNA damage stimulus, produce IPA isoforms with different consequences, such as to inhibit their expression? *

Our response: We think it is difficult to compare the enriched GO terms between overall IPA and miP-5’UTR-IPA. Indeed, differences could be due in part to trivial reasons (e.g., different number of genes in the lists). As suggested by this reviewer, it could be that for some gene sets enriched in particular functions, IPA may serve to downregulate the expression level of the full-length (canonical) mRNA. We discuss that this may be the case for the PRIM2 gene involved in DNA replication (page 17), but expanding on this would be speculative. Likewise, IPA isoforms encoding carboxy-terminal isoforms of canonical proteins, or IPA isoforms with a noncoding function (like ASCC3 or SPUD), might be enriched in particular gene functions, but again this idea is speculative and it goes beyond the scope of our manuscript.

In addition, the authors need to use ribo-seq and mass spec data as a validation tool for their polysome profiling data to indicate the reliability of using polysome data to call protein expression.

Our response: This comment seems to concern those IPA isoforms that are abundant in heavy polysomes. We do not wish to validate protein production from such isoforms, because they are not the focus of our study.

Significance

The significance of this work is its novelty in reporting IPA isoform activation by cisplatin. More importantly, some IPA isoform give rise to microproteins that have functional roles in cell survival upon cisplatin treatment.

Our response: We would like to highlight that the main finding of our manuscript is the discovery that IPA isoforms are a source of microproteins. Cisplatin response is the biological context in which we did the study, and therefore our functional and mechanistic analyses.

REVIEWER #3

Evidence, reproducibility and clarity

Devaux et al. report how cisplatin treatment changes the abundance of mRNA isoforms, favoring the expression of short transcripts originating from intronic polyadenylation (IPA) events relative to the expression of the corresponding mRNA isoform that includes the last annotated exon (LE). To detect IPA events the authors performed 3' end sequencing of polyadenylated mRNAs, long-read sequencing and conventional total RNA sequencing experiments in control and cisplatin treated cells. Analysis of the 3' end sequencing data revealed numerous genes showing an increase in the IPA:LE ratio upon cisplatin treatment, whereas few events with a decreased IPA:LE ratio were detected. Many of the identified events could be corroborated by the long-read sequencing data, sequencing of total RNA, and an existing polyA database. Furthermore, the authors validate IPA:LE ratios for a few selected genes using quantitative PCR. Subsequently, the authors continue to analyze if IPA isoforms are translated with a specific focus on IPA isoforms that do not contain any parts of the LE isoform coding sequence but terminated transcription in what is annotated as 5' untranslated region (UTR). These experiments show that IPA isoforms (including 5' UTR-IPAs) are translated but frequently associated with fewer ribosomes than the corresponding LE isoform. For two selected 5' UTR-IPA isoforms the authors identified potential small open reading frames (sORFs) that could give rise to microproteins with a potential function during cisplatin treatment. siRNA experiments targeting either the 5' UTR-IPA or the LE mRNA isoform of selected genes identified a small but significant differential effect on cell viability upon cisplatin treatment. Similar results were obtained when the endogenous IPA locus was deleted or the start codon of the potential sORF was mutated. Finally, the authors shed some light onto the molecular mechanisms of how cisplatin affects the IPA:LE ratio by decreasing transcription processivity.

*This is an interesting manuscript suggesting a link between IPA, sORFs and cancer treatment. The manuscript offers valuable datasets as a resource for the research community. While the authors generally present a well-analyzed and validated dataset supporting their claims, some aspects require further evidence or clearer presentation for robustness and reader comprehension. In addition, the manuscript would benefit from improving data visualization and we have several suggestions (see below) on how to make the representation of the data in the figures more appealing to the reader. We encourage the authors to reconsider several of their bar plots and instead plot their data on a continuous axis, e.g. using a scatter plot (fold change versus FDR) instead of a bar chart that can only represent up/down total numbers. *

Our response: Please see our responses below.

Main points:

1. • We disagree with one of the data interpretations concerning the high polysome (HP) versus total cytosolic polysomes (cytosol) localized IPA and LE mRNA isoforms in the paragraph "A subset of IPA isoforms are depleted in heavy polysomes and terminate in the annotated 5'UTR part of genes". Preferential IPA isoform localization to cytosol versus HP in comparison to the LE isoform does not mean that the IPA isoform translation efficiency is lower than that of the LE isoform. It just reflects the fact that IPA isoform coding sequence is considerably shorter than the coding sequence of the LE isoform (and thus can accommodate fewer ribosomes!). The authors mention that point later in the text but it should already be made clear at this point in the manuscript. They should make sure not to confuse translation efficiency (ribosome density across an open reading frame) and open reading frame length. * Our response: We will modify the text of this section (pages 10-11). We will __state that ‘the HP:cytosol ratio is usually considered as a proxy for translation efficiency’ and __we will only make conclusions in terms of ‘HP:cytosol ratio’ or ‘HP recruitment efficiency’, instead of ‘translation efficiency’ (we had used this term in a few sentences for the sake of simplicity). Please note that these changes will not alter the main conclusion of this part, because both the title (‘a subset of IPA isoforms are depleted in heavy polysomes and terminate in the annotated 5'UTR part of genes’) and the end of this section (page 11), as well as the legend of Fig. 2, were already written in such terms. Thus, in this section, we do not need to discuss ORF length (and we cannot, because sORFs are not annotated as CDS and we introduce sORFs only two sections later [Fig. 4]).

2. In Figure 5, the authors claim that the "cisplatin survival phenotype of the PRKAR1B 5'UTR-IPA isoform is attributable to its small ORF#2". This is an interesting phenotype but the authors only present a WST1 assay to support these claims. Given that it is an important Figure in their manuscript and links the observations made earlier to cisplatin-induced survival, it would be critical to bolster these claims with additional data, e.g. AnnexinV/PI staining and flow cytometry to distinguish changes in cisplatin-induced apoptosis from proliferation.*

Our response: We will make the requested experiments with FACS analysis of Annexin V and PI/BrdU to distinguish changes in cisplatin-induced apoptosis from proliferation (cell cycle).

• Along the same line, it would be important to test the overexpression of the sORF microprotein upon cisplatin treatment. Changes in the mRNA sequence (such as the AUG mutation) could potentially also alter the mRNA structure. It would therefore be critical to show that the sORF microprotein is indeed responsible for the changes in cisplatin-induced viability (for instance by expression of a sORF::P2A::GFP construct). *

Our response: As requested, we will test whether overexpression of the sORF microprotein can rescue the cisplatin survival phenotype of our PRKAR1B IPA and ATG mutants. We have appropriate constructs for this.

• Figure 5C: Please show the Western blot of PRKAR1B and GAPDH and not just the quantification. There is plenty of space in Figure 5. *

Our response: We will show the Western blots for PRKAR1B and GAPDH.

• In the following, we list suggestions to improve different figures where the data could be more adequately presented:*

- Figure 1A and B: We suggest representing the data in a scatter plot log fold change on the x-axis and FDR on the y-axis. The authors decided for an FDR cutoff of 10%. This is quite high. Why did the authors decide for this cutoff? How many genes would be identified with a more stringent cutoff (1% for example)? Please list the corresponding FDR values in TableS4.

Our response: We have never seen in the literature 3’-seq (or related) data of IPA:LE ratio regulation plotted as a scatter plot with log fold change on the x-axis and FDR on the y-axis. Instead, we propose to provide scatter plots with IPA fold change on the x-axis and LE fold change on the y-axis, as in many previous studies. We were not very stringent on the FDR or adjusted p values, in order to reduce the rate of false negatives, because we then cross our lists of regulated IPAs in different compartments (e.g., cytosol and heavy polysomes; Fig. 2C). We provided adjusted p values in Table S4; with an adjusted p value of 1%, we observe 1986 upregulated IPA sites and 33 downregulated ones.

*-Figure 1C: There are many ways to visualize fold change, p value and number of genes of a GO term analysis. The authors could choose one of the common ways to represent such data instead of just showing raw numbers in a table. *

Our response: We like showing GO terms as tables, but we can provide a figure if necessary.

-Figure 1E-G: Add to the figure that PRIM2 was assayed. It is only written in the figure legend.

Our response: We will write ‘PRIM2’ in the figure.

*-Figure 2A and B: Same suggestion as for Figure 1A and B, a scatter plot log fold change on the x-axis and FDR on the y-axis would visualize the data much better. *

Our response: Same response as for Fig 1A-B above.

-Figure S1B: Where does the number of 2118 cisplatin regulated genes come from? It was not described anywhere else. Should it not be 1987 regulated genes?

Our response: We will clarify that 2118 is the union of genes with cisplatin upregulated IPA:LE ratio in H358 and/or A549 cells.

-Figure S1H: Typo in the y-axis.

Our response: This typo will be corrected, thanks.

-Figure S2A: Same suggestion as for Figure 1A and B, a scatter plot log fold change on the x-axis and FDR on the y-axis.

Our response: Same response as for Fig 1A-B above.

-Figure S3C: If possible, show the plotted digital data of the polysome curves.

Our response: We do not have digital data for the polysome curves, just the printed graph shown at the bottom of the figure.

• Data availability: The provided UCSC genome browser link unfortunately does not load the data bam files. Please fix.*

Our response: We will fix this upon submission to journal.

Minor points:

• Please check the text for typos, e.g. page 8: artefacts instead of artifacts. *

Our response: We will check for typos.

Significance

The manuscript describes an interesting link between intronic polyadenylation, sORFs and cancer treatment and will be of interest to the gene expression regulation and RNA communities. As a relatively unknown mechanism to induce sORF-encoded microproteins, the study could lead to follow-up studies tackling intronic polyadenylation and their role in sORF expression.

Our response: We would like to highlight that IPA was not previously known to induce sORF-encoded microproteins.

While the authors generally present a well-analyzed and validated dataset, the link between sORF function and cisplatin response will require additional experiments to strengthen the sORF's impact for cellular survival.

Our response: Please see our responses to main points 2 and 3 above.

3. Description of the revisions that have already been incorporated in the transferred manuscript

None.

4. Description of analyses that authors prefer not to carry out

REVIEWER #2

Major point #2: The authors used the polysome:cytosolic ratio to indicate translational efficiency. However, because the CDS size affects the number of ribosomes per mRNA, the translational efficiency should be based on polysome:cytosolic ratio normalized to CDS size. Ideally, the authors should calculate number of ribosome per transcript based on monosome, light polysome and heavy polysome.

Our response: We cannot normalize ribosome number by CDS size because (i) heavy polysomes are not a precise number of ribosomes and (ii) sORFs are not annotated as CDS.

*Minor point #3: Fig. 6, the authors did data mining of ribo-seq data and mass-spec data and identified 156 genes whose IPA isoforms have potentials of protein expression. The enriched GO terms for the 156 IPA genes are different than the overall IPA isoforms shown in Fig. 1C. Does this mean some genes, like those in DNA damage stimulus, produce IPA isoforms with different consequences, such as to inhibit their expression? *

Our response: We think it is difficult to compare the enriched GO terms between overall IPA and miP-5’UTR-IPA. Indeed, differences could be due in part to trivial reasons (e.g., different number of genes in the lists). As suggested by this reviewer, it could be that for some gene sets enriched in particular functions, IPA may serve to downregulate the expression level of the full-length (canonical) mRNA. We discuss that this may be the case for the PRIM2 gene involved in DNA replication (page 17), but expanding on this would be speculative. Likewise, IPA isoforms encoding carboxy-terminal isoforms of canonical proteins, or IPA isoforms with a noncoding function (like ASCC3 or SPUD), might be enriched in particular gene functions, but again this idea is speculative and it goes beyond the scope of our manuscript.

In addition, the authors need to use ribo-seq and mass spec data as a validation tool for their polysome profiling data to indicate the reliability of using polysome data to call protein expression.

Our response: This comment seems to concern those IPA isoforms that are abundant in heavy polysomes. We do not wish to validate protein production from such isoforms, because they are not the focus of our study.

REVIEWER #3

-Figure S3C: If possible, show the plotted digital data of the polysome curves.

Our response: We do not have digital data for the polysome curves, just the printed graph shown at the bottom of the figure.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #3

Evidence, reproducibility and clarity

Devaux et al. report how cisplatin treatment changes the abundance of mRNA isoforms, favoring the expression of short transcripts originating from intronic polyadenylation (IPA) events relative to the expression of the corresponding mRNA isoform that includes the last annotated exon (LE). To detect IPA events the authors performed 3' end sequencing of polyadenylated mRNAs, long-read sequencing and conventional total RNA sequencing experiments in control and cisplatin treated cells. Analysis of the 3' end sequencing data revealed numerous genes showing an increase in the IPA:LE ratio upon cisplatin treatment, whereas few events with a decreased IPA:LE ratio were detected. Many of the identified events could be corroborated by the long-read sequencing data, sequencing of total RNA, and an existing polyA database. Furthermore, the authors validate IPA:LE ratios for a few selected genes using quantitative PCR. Subsequently, the authors continue to analyze if IPA isoforms are translated with a specific focus on IPA isoforms that do not contain any parts of the LE isoform coding sequence but terminated transcription in what is annotated as 5' untranslated region (UTR). These experiments show that IPA isoforms (including 5' UTR-IPAs) are translated but frequently associated with fewer ribosomes than the corresponding LE isoform. For two selected 5' UTR-IPA isoforms the authors identified potential small open reading frames (sORFs) that could give rise to microproteins with a potential function during cisplatin treatment. siRNA experiments targeting either the 5' UTR-IPA or the LE mRNA isoform of selected genes identified a small but significant differential effect on cell viability upon cisplatin treatment. Similar results were obtained when the endogenous IPA locus was deleted or the start codon of the potential sORF was mutated. Finally, the authors shed some light onto the molecular mechanisms of how cisplatin affects the IPA:LE ratio by decreasing transcription processivity.

This is an interesting manuscript suggesting a link between IPA, sORFs and cancer treatment. The manuscript offers valuable datasets as a resource for the research community. While the authors generally present a well-analyzed and validated dataset supporting their claims, some aspects require further evidence or clearer presentation for robustness and reader comprehension. In addition, the manuscript would benefit from improving data visualization and we have several suggestions (see below) on how to make the representation of the data in the figures more appealing to the reader. We encourage the authors to reconsider several of their bar plots and instead plot their data on a continuous axis, e.g. using a scatter plot (fold change versus FDR) instead of a bar chart that can only represent up/down total numbers.

Main points:

1. We disagree with one of the data interpretations concerning the high polysome (HP) versus total cytosolic polysomes (cytosol) localized IPA and LE mRNA isoforms in the paragraph "A subset of IPA isoforms are depleted in heavy polysomes and terminate in the annotated 5'UTR part of genes". Preferential IPA isoform localization to cytosol versus HP in comparison to the LE isoform does not mean that the IPA isoform translation efficiency is lower than that of the LE isoform. It just reflects the fact that IPA isoform coding sequence is considerably shorter than the coding sequence of the LE isoform (and thus can accommodate fewer ribosomes!). The authors mention that point later in the text but it should already be made clear at this point in the manuscript. They should make sure not to confuse translation efficiency (ribosome density across an open reading frame) and open reading frame length.
2. In Figure 5, the authors claim that the "cisplatin survival phenotype of the PRKAR1B 5'UTR-IPA isoform is attributable to its small ORF#2". This is an interesting phenotype but the authors only present a WST1 assay to support these claims. Given that it is an important Figure in their manuscript and links the observations made earlier to cisplatin-induced survival, it would be critical to bolster these claims with additional data, e.g. AnnexinV/PI staining and flow cytometry to distinguish changes in cisplatin-induced apoptosis from proliferation.
3. Along the same line, it would be important to test the overexpression of the sORF microprotein upon cisplatin treatment. Changes in the mRNA sequence (such as the AUG mutation) could potentially also alter the mRNA structure. It would therefore be critical to show that the sORF microprotein is indeed responsible for the changes in cisplatin-induced viability (for instance by expression of a sORF::P2A::GFP construct).
4. Figure 5C: Please show the Western blot of PRKAR1B and GAPDH and not just the quantification. There is plenty of space in Figure 5.

5. In the following, we list suggestions to improve different figures where the data could be more adequately presented:

   • Figure 1A and B: We suggest representing the data in a scatter plot log fold change on the x-axis and FDR on the y-axis. The authors decided for an FDR cutoff of 10%. This is quite high. Why did the authors decide for this cutoff? How many genes would be identified with a more stringent cutoff (1% for example)? Please list the corresponding FDR values in TableS4.
   • Figure 1C: There are many ways to visualize fold change, p value and number of genes of a GO term analysis. The authors could choose one of the common ways to represent such data instead of just showing raw numbers in a table.
   • Figure 1E-G: Add to the figure that PRIM2 was assayed. It is only written in the figure legend.
   • Figure 2A and B: Same suggestion as for Figure 1A and B, a scatter plot log fold change on the x-axis and FDR on the y-axis would visualize the data much better.
   • Figure S1B: Where does the number of 2118 cisplatin regulated genes come from? It was not described anywhere else. Should it not be 1987 regulated genes?
   • Figure S1H: Typo in the y-axis.
   • Figure S2A: Same suggestion as for Figure 1A and B, a scatter plot log fold change on the x-axis and FDR on the y-axis.
   • Figure S3C: If possible, show the plotted digital data of the polysome curves.
   • Data availability: The provided UCSC genome browser link unfortunately does not load the data bam files. Please fix.

Minor points:

1. Please check the text for typos, e.g. page 8: artefacts instead of artifacts.

Significance

The manuscript describes an interesting link between intronic polyadenylation, sORFs and cancer treatment and will be of interest to the gene expression regulation and RNA communities. As a relatively unknown mechanism to induce sORF-encoded microproteins, the study could lead to follow-up studies tackling intronic polyadenylation and their role in sORF expression.

While the authors generally present a well-analyzed and validated dataset, the link between sORF function and cisplatin response will require additional experiments to strengthen the sORF's impact for cellular survival.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

Summary:

In this manuscript, Devaux et al. report that the anti-cancer drug cisplatin upregulates intronic polyadenylation (IPA) isoforms in non-small cell lung cancer cell lines. Their finding was based on 3' end sequencing and long-read sequencing. Through polysome profiling they confirmed that many of the IPA isoforms are translated, despite being inefficient in most cases. They validated functions of IPA isoforms from two genes, PHF20 and PRKAR1B, in cell survival upon cisplatin treatment, based on an array of methods, including siRNA knockdown, CRISPR knockout of IPA polyA site, and CRISPR mutation of the start codon. They further found that FANCD2 and Senataxin can regulate cisplatin-mediated IPA activation. The authors advocate a new paradigm of expression of IPA-encoding microproteins in cisplatin-treated cells.

Major comments:

• While the phenomenon of IPA isoform upregulation by Cisplatin is quite convincing, the underlying mechanism is largely elusive. The authors indicated processivity as a potential mechanism and the effects of FACD2 and Senataxin appear in line with this hypothesis. However, they cannot rule out other possibilities based on the data presented in the manuscript. For example, it is not clear if the elongation rate of Pol II (distinct from its processivity) or nuclear RNA degradation is affected by cisplatin, which could also lead to increased expression of IPA isoforms. In addition, enhanced 3' end processing activity has been previously shown to activate IPA sites. Therefore, the underlying mechanism is mostly speculative.
• The authors used the polysome:cytosolic ratio to indicate translational efficiency. However, because the CDS size affects the number of ribosomes per mRNA, the translational efficiency should be based on polysome:cytosolic ratio normalized to CDS size. Ideally, the authors should calculate number of ribosome per transcript based on monosome, light polysome and heavy polysome.
• The functions of PHF20 and PRKAR1B IPA isoforms are based on knockdown or knockout mutations. Because of its gain-of-function property, overexpression of the isoforms in cisplatin-treated cells would be necessary to definitively confirm their funcitons.

Minor:

Fig. 1H, the numbers of IPA and LE transcripts should be provided. The statistical significance for the difference should also be included.

Fig. 1I, the image should be accompanied with fold difference as indicated in the text. Some statistics for difference between vehicle only and CisPt only is necessary.

Fig. 6, the authors did data mining of ribo-seq data and mass-spec data and identified 156 genes whose IPA isoforms have potentials of protein expression. The enriched GO terms for the 156 IPA genes are different than the overall IPA isoforms shown in Fig. 1C. Does this mean some genes, like those in DNA damage stimulus, produce IPA isoforms with different consequences, such as to inhibit their expression? In addition, the authors need to use ribo-seq and mass spec data as a validation tool for their polysome profiling data to indicate the reliability of using polysome data to call protein expression.

Significance

The significance of this work is its novelty in reporting IPA isoform activation by cisplatin. More importantly, some IPA isoform give rise to microproteins that have functional roles in cell survival upon cisplatin treatment.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

Microporteins originating from coding and non-coding transcript are increasingly understood to control various cellular processes. In the present study, the authors investigated whether intronic polyadenylation (IPA) contributes to the formation of transcript isoforms encoding microproteins. Using genotoxic stress by cisplatin as a model in cell cultures, the authors detect abundant IPA. IPA in a subset of such transcripts leads to short 5'UTR transcript isoforms that are poorly associated with heavy polysomes and encode microproteins. For PRKAR1B, they demonstrate the expression of a corresponding microprotein and a function in modulating the cisplatin response. Based on depletion experiments of FANCD2 and STX1, the authors propose that impaired transcription processivity after cisplatin is one mechanism leading to IPA and microprotein production.

While this is an interesting manuscript, I felt the support for the claimed generalization falls a bit short.

Major:

If I see it correctly, the authors mainly refer to existing riboSeq data and evidence from mass spectrometry/proteomics to infer the generality of the mechanism (beyond PRKAR1B). It is important to back this up with further experiments and validate this for the set-up used in this manuscript. This concerns the existence of the microproteins but also the downstream functional impact.

Also, I wonder is this limited to cisplatin-induced genotoxic stress and the specific cell line used or is this a more global mechanism?

Minor:

• While the rest of the paper reads well, the abstract could be improved/simplified to increase accessibility
• Page 11: Pertaining to Figure 3 and the functional impact: The authors analyze the IPA effect by probing cell viability and cell survival. It would be important to define the effects in further detail, as the mere regulation of cell cycle and/or apoptosis could also result in such outcome (which is then not necessarily a direct cisplatin response). Does this also impact the response to other genotoxic stress (also pertains to the effects studied and shown in Fig. 5)?
• Page 11: Pertaining to Figure 3 and the functional impact: The authors analyze the IPA effect by probing cell viability and cell survival. It would be important to define the effects in further detail, as the mere regulation of cell cycle and/or apoptosis could also result in such outcome (which is then not necessarily a direct cisplatin response). Does this also impact the response to other genotoxic stress (also pertains to the effects studied and shown in Fig. 5)?
• Page 12 concerning the microprotein expression: the authors refer to data from other resources to claim that the microproteins are expressed, however they fail to demonstrate this for their setup (at least for 2 out of three they study here). I think this is a weak point as it does not directly support the general claim.
• Also, I did not understand what the authors intended to demonstrate with the immunoflourescence (Fig. 4E). What should a defined nuclear expression imply versus the diffuse staining throughout after cisplatin? How does this relate to the functional effects?
• Page 13/Fig. 5E: the different clones of the mATG show very high variability. To my understanding it is difficult to draw a clear conclusion from this heterogeneity.
• Page 15 on the mechanism: SETX has been demonstrated to control poly(A) site choice (PMID: 21700224, 32976578). However the quantitative role of SETX in poly(A) site choice regulation (compared to other regulators) seems to be rather marginal and not strictly unidirectional, i.e after SETX depletion also longer transcript isoforms can be detected (PMID: 32976578). How does this relate to the proposed mechanism of SETX-dependent processivity? Interestingly, from PMID: 32976578 it also appears that PRKAR1A has a 5'UTR poly(A) site that is regulated in a SETX-dependent manner.
• Page 16, discussion first paragraph. While refs 1-4 are nice reviews that could be quoted here a study that appeared later represents the most comprehensive analysis to date covering the different facets from transcription to RNA processing and the resulting impact on poly(A) site choice (PMID: 30552333).

Significance

This could be a very significant report, provided the generality of the claims and mechanistic insigths are further strengthend.

Overall it targets a rather specialized readership. This could be improved by simplifing the abstract, additional experimental evidence for the generality of the proposed mechanism, and a stringent rewording of the main text drawing a clear line, omitting unnecessary details and focussing on the novel findings.

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

Manuscript number: RC-2023-02232

Corresponding author(s): Shinji, Saiki and Nobutaka, Hattori

1. General Statements [optional]

Thank you for the review of our paper entitled “Identification of novel autophagy inducers by accelerating lysosomal clustering against Parkinson's disease” (RC-2023-02232). We have carefully read the critiques and planed experiments. Below we include point-by-point responses to the questions raised by the reviewers. We have also carried out some experiments and highlighted the revised sentences in the transferred manuscript in red. The numbers of pages and lines are indicated based on the MS Word transferred manuscript. We believe this revision plans appropriately addresses the issues raised by Reviewers. Finally, all the authors would like to thank again the Editor and Reviewers for improving our manuscript by providing their invaluable comments and suggestions.

Point-by-point description of the revisions

• *

Reviewer #1 (Evidence, reproducibility and clarity (Required)):

The manuscript by Date et al employed a cell model by stably expressing LGP120-mCherry and GFP-gamma-tubulin to carry out high-content screening in search of chemical compounds that enhance lysosomal clustering and autophagy. They found 6 clinically approved drugs categorized as topoisomerase II inhibitors and the benzimidazole class. They further validated these compounds by a set of well-designed experiments including autophagy flux assays and mTOR dependence. In the mechanistic study, they demonstrated the compounds induce lysosomal clustering in a JIP4-TRPML1-dependent manner. In a PD cell model, one of the compounds albendazole exhibited the effect on boosting the degradation of insoluble alpha-synuclein. The study is of interest, and the cell model and the approach generated by the authors would be transferable for future studies of other high-content imaging screening. Most of the data is clear and convincing.

Major comment

1) In addition to its role in facilitating a-syn turnover by autophagy, Is the chemical protective against a-syn toxicity?

RESPONSE:

As suggested by the Reviewer, we examined the cytotoxicity of aSyn aggregates in SH-SY5Y cells overexpressing aSyn-GFP by LDH assay. As shown in the revised version of Fig. 1, aSyn aggregates induced by introducing aSyn fibrils into SH-SY5Y cells overexpressing aSyn-GFP did not exhibit any cytotoxicity. In addition, we observed no significant change in cell death after 8 hours of treatment with albendazole compared with DMSO.

Previous studies have reported that induced pluripotent stem cells (iPSCs) derived from patients with PD with a triplication of the human SNCA genomic locus exhibited reduced capacity for differentiation into dopaminergic or GABAergic neurons, decreased neurite outgrowth, and lower neuronal activity compared with control cultures, albeit without showing cytotoxicity (Cell Death and Disease 6: e1994, Oliveira et al., 2015). Given this context, we were thus unable to conduct the suggested assessment due to technical limitations. Therefore, we consider the evaluation of the recovery of aSyn toxicity by drug treatment challenging in this cellular model using fibril aSyn.

__Revised Fig 1. __

SH-SY5Y cells overexpressing aSyn-GFP were transfected with aSyn fibril for 48 h and treated with the indicated albendazole concentrations for 8 h. The cytotoxicity was measured by using Cytotoxicity LDH Assay Kit-WST kit.

2) Please elaborate why albendazole does not change the levels of soluble a-syn, but those of insoluble, as shown Fig 8D.

RESPONSE:

The unchanged aSyn-GFP levels in the soluble fraction (Fig. 8D) are likely due to the abundance of soluble aSyn-GFP. To evaluate the autophagic degradation of aSyn monomers, we used SH-SY5Y cells stably expressing aSyn-Halo and measured aSyn degradation by quantifying cleaved Halo. As shown in the revised version of Fig. 2, albendazole treatment induced a higher cleavage rate of Halo than DMSO treatment for 8 h, suggesting that albendazole degrades both aSyn monomers and aSyn aggregates. We have added the data in Fig. S7A, and the description of these experiments in the Results section (page 10, lines 359 to 364).

__Revised Fig. 2. __

SH-SY5Y cells expressing aSyn-Halo were labeled for 20 min with 100 nM of tetramethylrhodamine-conjugated ligand in a nutrient-rich medium. After washing with phosphate-buffered saline and incubating in normal medium for 30 min, the cells were treated with 10 µM albendazole for 8 h. The experiments were performed in triplicate. Cell lysates were separated by electrophoresis and analyzed by in-gel fluorescence detection (left). The HaloTMR band intensity was normalized by the sum of the band intensities of HaloTMR-aSyn and HaloTMR. The vertical axis of the graph represents the intensity multiplied by 100. Mean values of data from five or three experiments are shown. The graph data are expressed as mean ± standard deviation. ****P 

3) Fig 6A shows that some of the compounds (Teniposide, Amsacrine) affect the levels of JIP4. Can albendazole also reduce JIP4 levels. It might be interesting to test this, as JIP4 is important for lysosomal clustering.

RESPONSE:

As the Reviewer pointed out, JIP4 is essential for lysosome accumulation. However, our data showed decreased JIP4 levels with the addition of lysosomal-clustering compounds. We hypothesized that this response was caused by the autophagy-induced degradation of JIP4. The decrease in JIP4 levels was detected by western blot after 4 h of treatment with 10 μM of teniposide. Moreover, the decrease in JIP4 levels induced by teniposide was suppressed by co-treatment with bafilomycin A1, indicating that JIP4 was degraded by teniposide-induced autophagy, as shown in the revised version of Fig. 3. We have added the data in Fig. S6 and the related description of these experiments in the Results section (page 8, lines 289 to 293).

__Revised Fig 3. __

SH-SY5Y cells were treated with 10 µM teniposide and with or without 30 nM bafilomycin A1 for 4 h. Cell lysates were immunoblotted with anti-JIP4 and actin antibodies.

Minor comments: The writing is good generally. Please tide up the text in a few occasions to make the expressions more formal.

RESPONSE: We have revised our manuscript to adopt a more formal tone.

Reviewer #1 (Significance (Required)):

Significance: The study generated a new approach for high-throughput screening of compounds to enhance lysosomal clustering. Audience: Basic and clinical research Expertise: Programmed cell death, neurodegenerative diseases

• *

Reviewer #2 (Evidence, reproducibility and clarity (Required)):

In this study, the authors focused on lysosome positioning and autophagy activity to search for novel agents effective against Parkinson's disease. As a result, several compounds were successively identified, including Topoisomerase inhibitors and Benzimidazole. Authors showed that these agents regulate lysosomal positioning through different pathways but commonly require JIP4 to regulate lysosomal positioning and subsequent autophagy. They also showed that albendazole treatment promoted the degradation of insoluble ubiquitinated proteins and αSyn in cultured cells.

Major Comments.

1) Two compounds, for instance teniposide and albendazole both requires JIP4 and/or TRPML1 to regulate lysosomal positioning and autophagy but their action seems different. What is the actual mechanism by which these compounds require JIP4/TRPML1. How inhibition of Topoisomerase leads to increase of JIP4 phosphorylation? Do teniposide and albendazole both affect calcium release from TRPML1?

RESPONSE:

We previously reported that acrolein/H2O2 accelerates lysosomal retrograde trafficking by TRPML1 and phosphorylated JIP4. Mechanistically, JIP4 was phosphorylated by CaMK2G activated by Ca2+ released from TRPML1 (EMBO J 41: e111476, Sasazawa et al., 2022). TRPML1 acts as a reactive oxygen species (ROS) sensor in lysosomes (Nat Commun 7: 12109, Zhang et al., 2016). We concluded that acrolein induces ROS production, which then activates TRPML1. (EMBO J 41: e111476, Sasazawa et al., 2022). Therefore, topoisomerase inhibitors (topo-i) may induce ROS and stimulate TRPML1. We examined intracellular ROS levels in response to topo-i. As shown in revised Fig. 4A, the topo-i teniposide, etoposide, and amsacrine significantly increased ROS levels. Moreover, N-acetyl-L-cysteine, an ROS scavenger, partially attenuated lysosomal clustering induced by topo-i (revised Fig. 4B). In addition, Ca2+ imaging showed that teniposide, but not albendazole, upregulates Ca2+ flux (revised Fig. 4C). Based on the activity of CaMK2G siRNA as shown in Fig. 5D, 5E, and S5, topo-i may activate TRPML1 in a ROS-dependent manner and increase PI(3,5)P2 binding with TRPML1 (Nat Commun 1, 38, Dong et al., 2010). Consecutive Ca2+ release via TRPML1 activated CaMK2G and is followed by enhanced lysosomal transport toward the MTOC via JIP4 phosphorylation.

We have added the revised Fig.4A and 4B data in Fig. S8A and S8B, and the related description of these experiments in the Discussion section (page 11, lines 401 to 409). We have also added the data in revised Fig. 4C to Fig. S6 and the related description of these experiments in the Results section (page 7, lines 266 to 267).

Conversely, we showed that benzimidazoles, including albendazole, induce lysosomal clustering mediated by JIP4, TRPML1, ALG2, and Rab7. Moreover, benzimidazoles showed lysosomal clustering activity within a narrow concentration range, as shown in Fig. S7D. Benzimidazoles inhibit tubulin polymerization (Int J Paras 18:885–936. Lacey et al., 1988). We hypothesized that the effect of tubulin polymerization induced by benzimidazole plays a key role in the induction of lysosomal clustering as described in the Discussion section. To clarify this, we observed the behavior of tubulin filaments in response to various albendazole concentrations under confocal microscopy. As shown in revised Fig. 4D, conditions where albendazole was administered to induce lysosomal clustering, tubulin filaments were observed only near the MTOC, and the filaments in the cell periphery were disassembled. In contrast, when exposed to higher albendazole concentrations, tubulin filaments throughout the cell were disassembled, resulting in the inhibition of lysosomal clustering This would explain why benzimidazole exerts lysosomal clustering activity within a narrow concentration range. Under JIP4, TRPML1, ALG2 and Rab7 silencing, lysosomes may fail to interact with microtubules, resulting in the inhibition of lysosomal clustering. We postulated that albendazole-induced lysosomal clustering is not mediated by factors activated by specific stimuli in lysosomal transport but, rather, is induced by spatially constraining conventional lysosomal transport mediated by various adaptors (i.e., JIP4, TRPML1, ALG2, and Rab7) through tubulin disassembly. We have added the data in Fig. S9C and the related description of these experiments in the Discussion section (page 12, lines 428 to 436).

A B

C

D

__Revised Fig. 4. __

1. SH-SY5Y cells were treated with the indicated compounds (10 µM) for 4 h. The amount of intracellular reactive oxygen species (ROS) is examined by ROS Assay Kit -Highly Sensitive DCFH-DA (Dojindo) and the normalized pixels above threshold as measured using an INCellAnalyzer 2200 and ImageJ.
2. SH-SY5Y cell lines were pretreated with 0.1 mM N-acetyl-L-cysteine (NAC) for 24 h and then treated with the indicated compound (10 µM) for an additional 4 Cells were fixed and stained with anti-g-tubulin (green) and anti-LAMP2 (red) antibodies. Lysosomal distribution was examined using an INCellAnalyzer 2200 and quantified using ImageJ software.
3. SH-SY5Y cells were treated with teniposide, amsacrine, etoposide, albendazole (1, 5, 10 µM), oxibendazole (0.1, 0.5, and 1 µM), or and mebendazole (0.5,1, and 5 µM) for 4 h, and stained with Fluo4-AM for 30 min. The fluorescence intensity was measured using a plate reader.
4. SH-SY5Y cells were treated with albendazole (10 and 100 µM) or nocodazole (0.5 and 10 µM) for 4 h. Cells were fixed and stained with LAMP1 (red) and a-tubulin (green) antibodies.

   2) The authors should clarify the functional advantage of these drugs identified in this study as drugs for Parkinson's disease by comparing with known autophagy inducers such as Torin1 or rapamycin.　

RESPONSE:

To evaluate the functional advantage of lysosome-clustering compounds over Torin1, we evaluated the degradation activity of insoluble aSyn induced by the addition of aSyn fibrils to aSyn-GFP cells. Torin1 induced the degradation of insoluble aSyn by autophagy, as shown in revised Fig. 5A. However, the degradation activity of albendazole was more vigorous, as shown in revised Fig. 5B. In contrast, we observed that Torin1 exhibited more autophagic induction activity than albendazole, as assessed using Halo-LC3. Similar results were obtained with teniposide (revised Fig. 5C). These results suggest that albendazole, with its ability to concentrate lysosomes around the degradation substrate, facilitates more effective degradation of insoluble aSyn than Torin1. This presents a significant advantage in the development of therapeutics for Parkinson's Disease. Moreover, Torin1 acts on the upstream signals of autophagy by inhibiting mTORC1, potentially impacting diverse cellular responses. Conversely, compounds that induce lysosomal clustering target the final step of autophagic degradation, which may have fewer side effects. We have added the description of these experiments in the Results section (page 10, lines 366 to 380) and the Discussion section (page 11 lines 410 to 412) and presented the data in Fig. S7B–S7E and Fig. 6D.

A____ ____B

C

__Revised Fig. 5. __

1. SH-SY5Y cells overexpressing aSyn-GFP were transfected with aSyn fibril (0.2 µg/mL) using Lipofectamine 3000. After 48 h, the transfection reagent was washed out, and the SH-SY5Y cells were treated with 100 nM Torin1 with or without 100 nM bafilomycin A1 for 8 h (B). Cell lysates were separated into Triton X-100–soluble (soluble) and pellet fractions (insoluble), then subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotting with the indicated antibody(left). The amount of insoluble aSyn was quantified using Image J software (C).
2. SH-SY5Y cells overexpressing aSyn-GFP were transfected with aSyn fibril (0.2 µg/mL) using Lipofectamine 3000. After 48 h, and washing out the transfection reagent, SH-SY5Y cells were treated with albendazole (10μM) with or without 100 nM Torin1 for 8 h. Cell lysates were separated into Triton X-100–soluble (soluble) and pellet fractions (insoluble), then subjected to SDS-PAGE and immunoblotting with the indicated antibody (left). The amount of insoluble aSyn was quantified using Image J software.
3. SH-SY5Y cells stably expressing Halo-LC3 were labeled for 20 min with 100 nM TMR-conjugated ligand in a nutrient-rich medium. After washing with PBS and incubating the cells in normal medium for 30 min, cells were treated with DMSO, teniposide (10 μM), albendazole (10 µM), and/or Torin1 (100 nM) for 8 h. Cell lysates were immunoblotted with the indicated antibody and analyzed by in-gel fluorescence detection (left). The HaloTMR band intensity was normalized by the sum of the band intensities of HaloTMR-LC3B and HaloTMR (right).

   3) Related to the previous question, in Fig.6A and B additional data comparing novel compounds with established autophagy inducers, such as torin1 and rapamycin, should be included and discussed.

RESPONSE:

As indicated in a previous response, we evaluated the autophagic induction activity of Torin1, and the results have been added to Fig. 6D. In addition, co-treatment with Torin1 and teniposide or albendazole induced autophagy more effectively than Torin1 treatment alone, without affecting mTOR inhibition activity (revised Fig. 4C). These findings indicate that the induction of autophagy by lysosomal clustering compounds is not caused by autophagosome formation but by the formation of autolysosomes. We have added a description of these experiments in the Results section (page 9, lines 316 to 322) and have added the data in Fig. 6D.

4) The authors should examined whether increased degradation of insoluble proteins and αSyn are dependent on JIP4. 　

RESPONSE:

As the Reviewer suggested, we have examined whether lysosomal accumulation through the JIP4-TRPML1 pathway is crucial for the degradation of aSyn aggregates. We evaluated the degradation activity of insoluble aSyn induced by the addition of aSyn fibrils to aSyn-GFP cells when JIP4, TMEM55B, or TRPML1 were knocked down. Interestingly, the insoluble fraction assay showed that JIP4 and TRPML1 knockdown regulated the decrease of aSyn-GFP and p-aSyn levels in the insoluble fraction for both DMSO and albendazole treatments. The results were particularly more pronounced with TRPML1 knockdown. However, the knockdown of TMEM55B did not produce such findings (revised Fig. 6). These data suggest that lysosomal clustering via the JIP4–TRPML1 pathway plays a significant role in aSyn degradation. We have added a relevant description in the Results section (page 10, lines 373 to 380) and have added the data in Fig. S7F and S7G.

__Revised Fig. 6. __

SH-SY5Y cells over-expressing aSyn-GFP were transfected with the indicated siRNAs for 24 h and then transfected with aSyn fibril (0.2 µg/mL) using Lipofectamine 3000 for 48 h. After washing out the transfection reagent, the SH-SY5Y cells were treated with dimethyl sulfoxide or albendazole (10 μM) for 8 h. Cell lysates were separated into Triton X-100–soluble (soluble) and pellet fractions (insoluble) and subjected to SDS-PAGE and immunoblotting with the indicated antibody. The bar graph presents the ratio of the insoluble aSyn-GFP to the soluble GAPDH or insoluble p-aSyn to the soluble GAPDH of the intensity of the data in panel F. Data are expressed as mean ± standard deviation.

5) Authors only utilized. SH-SY5Y cells in this study. It is important to examine whether these compounds also regulate lysosomal positioning and autophagy in other cell lines.

RESPONSE:

As per the Reviewer’s suggestion, we evaluated the lysosomal-clustering activity induced by topo-i and benzimidazole in human adenocarcinoma HeLa cells. As shown in revised Fig. 7A and 7B compounds do not induce lysosomal clustering or autophagy in HeLa cells. Furthermore, in the case of benzimidazole, they transport lysosomes to the cell periphery. Previously, we found that oxidative stress accumulates lysosomes in a neuroblastoma-specific manner through the TRPML1–phosphoJIP4-dependent mechanism (EMBO J 41: e111476, Sasazawa et al., 2022). Since we have demonstrated that topo-i-mediated lysosomal trafficking is dependent on the TRPML1–phosphorylated JIP4 complex, we hypothesized that several molecules involved in lysosomal trafficking are absent in HeLa cells.

In contrast, we showed that albendazole-induced lysosomal clustering is due to tubulin depolymerization. Therefore, we examined the relationships between tubulin depolymerization and lysosomal clustering induced by albendazole in HeLa cells and found that albendazole did not induce lysosomal clustering but rather inhibited it at higher concentrations (revised Fig. 7B). Interestingly, similar to SH-SY5Y cells, a low albendazole concentration (10 μM) induced tubulin depolymerization only at the cell periphery, whereas a high concentration (100 μM) depolymerized the entire cell (revised Fig. 7C). However, unlike SH-SY5Y cells, no characteristic accumulation of tubulin filaments was observed near the MTOC under low albendazole concentration (10 µM); instead, they were arranged around the nucleus. Concurrently, lysosomes were around these dispersed tubulin filaments. Therefore, the differences in the effects of benzimidazole in HeLa and SH-SY5Y cells lies in the dose-dependent effects on the state of tubulin filaments. We have added a relevant description in the Discussions section (page 12, lines 419 to 422, and 437 to 453).

__ A__

__ B__

C D

__Revised Fig. 7. __

HeLa cells were treated with teniposide (10 μM), amsacrine (10 μM), etoposide (10 μM), albendazole (10 μM), oxibendazole (1 μM), or mebendazole (5 μM). Images were captured using an INCellAnalyzer2200. INCellAnalyzer2200 images were analyzed using ImageJ for lysosomal clustering. The graph presents the lysosomal clustering values (n > 30). Data are expressed as mean ± standard deviation (SD). ****P HeLa cells were treated with teniposide (10 μM), amsacrine (10 μM), etoposide (10 μM), albendazole (10 μM), oxibendazole (1 μM), or mebendazole (5 μM) for 4 h. Cell lysates were immunoblotted with the indicated antibodies. The amount of LC3II was estimated using Image J software (bottom panel). HeLa cells were treated with albendazole at specified concentrations (in µM). After treatment, cells were fixed and stained with anti-g-tubulin (green) and anti-LAMP2 (red) antibodies, followed by imaging with an INCellAnalyzer2200. INCellAnalyzer2200 images were processed and analyzed using ImageJ for lysosomal clustering. The graph presents the lysosomal clustering values (n > 30). Data are expressed as mean ± SD. *P  HeLa cells were treated with albendazole (10 and 25 µM) for 4 h. Cells were fixed and stained with LAMP1 (red) and a-tubulin (green) antibodies.

6) The authors conclude that the six compounds do not mediate mTOR signaling in Fig. 3, but should more carefully describe in the manuscript why they performed this experiment and what the results mean for.

RESPONSE:　

As per the Reviewer’s advice, we have changed the description in the manuscript as follows:

Previous studies have shown that lysosomal retrograde transport regulates autophagic flux by facilitating autophagosome formation by suppressing mTORC1 and expediting fusion between autophagosomes and lysosomes (Kimura et al, 2008; Korolchuk et al, 2011). Conversely, we recent found that acrolein/H2O2 induces lysosomal clustering in an mTOR-independent manner (Sasazawa et al., 2022). In this study, we aimed to identify pharmacologic agents that act downstream rather than upstream in the autophagy pathway, with the goal of minimizing side effects. Therefore, we evaluated the effects of the compounds on the mTOR pathway. As shown in Fig. 3, these compounds induced lysosomal clustering without affecting mTOR activity, indicating their potential as promising candidates for PD therapy. We have added the description of these experiments in the Results section (page 6, lines 202 to 208 and line 217).

Minor comments. 1) The name of the compound should be written in the red point of Fig.2A.

RESPONSE:

We have included the names of the six compounds identified and are listed in Fig. 2A.

2) Regarding images of Fig.2B, the magnified images and quantitative data should be added.

RESPONSE:

We have included magnified images, as well as the quantitative results of lysosome clustering analysis using INCellAnalyzer2200 in Fig. 2B.

3) The results of Fig.2C need to be explained more carefully. A quantitative data is missing.

RESPONSE:

We have included the quantitative results of western blot in Fig. 2B.

4) Fig.S2, which compares autophagy activity with conventional agents, should be quantified and added to the Fig.3.

RESPONSE:

We have presented the results of RFP/GFP quantification performed by FACS analysis using SH-SY5Y cells stably expressing RFP-GFP-LC3 in Fig. S2, which is equivalent to the quantification of the data in the Fig. S2 image. These data are now presented as Fig. S2B. Since Fig. 3 focuses on mTOR signaling, we preferred to retain the figure number.

5) In the statistical analysis of Fig.4B, the clustering value was increased by siRILP, which should be briefly described in the manuscript.

RESPONSE:

On the contrary, the enhancement of lysosomal retrograde transport in RILP knockdown cells in Fig. 4B suggests the potential involvement of RILP in anterograde transport. However, to the best of our knowledge, no reports have investigated this matter. We presume that negative feedback mechanisms may be present. We have added this description to the Results section (page 7 lines 238 to 241).

6) In Fig.4A and B, it is possible that the knockdown efficiency of siRILP and siTMEM55B was not sufficient to observe the effect on lysosomes, and this concern should be described in the manuscript.

RESPONSE:

We established starvation conditions, which induce TMEM55B-dependent lysosomal retrograde transport, as a positive control and evaluated the lysosomal induction activity of compounds when TMEM55B was knocked down. As shown below, lysosome accumulation was suppressed only when subjected to starvation treatment, indicating sufficient knockdown efficiency of TMEM55B. These compounds induced lysosomal clustering independently of TMEM55B, unlike under starvation conditions. We have added a description of these experiments in the Results section and presented the data in Fig. S4A (page 7, lines 232 to 237).

On the other hand, we were unable to establish a positive control for RILP knockdown experiments because conditions that regulate RILP-dependent lysosomal distribution dependent are not understood. While we cannot completely rule out the possibility of insufficient knockdown efficiency, considering that RILP knockdown appears to paradoxically enhance lysosomal induction, as mentioned above, it is reasonable to assume that the knockdown effect has occurred.

__Revised Fig. 8. __

SH-SY5Y cells were transfected with TMEM55B siRNA for 48 h and then treated with teniposide (10 μM), albendazole (10 μM), or starvation medium for 4 h. Cells were fixed and stained with anti-g-tubulin (green) and anti-LAMP2 (red) antibodies. Images were captured using an INCellAnalyzer2200. INCellAnalyzer2200 images were analyzed using ImageJ for lysosomal clustering. The graph presents the lysosomal clustering values (n > 30). Data are expressed as mean ± standard deviation (SD). ****P

7) The authors should add the results of the WB experiment showing the amount of JIP4 protein in Fig.5G.　

RESPONSE:

We have added western blot data that introduce flag-JIP4 into JIP4KO SH-SY5Y cells, which are presented in Fig. 5G.

8) In Fig.5F, images of JIP4KO cells that do not express FLAG-JIP4 should be added as controls, and further quantification should be done on cells in all three conditions.

RESPONSE:

We have added immunofluorescence data that do not express flag-JIP4 in Fig. 5F, which had been obtained simultaneously during the acquisition of other images. Furthermore, we quantified lysosomal distribution, which is shown in Fig. 5E. Using ImageJ, we automatically delineated approximately 70% of the cell area toward the cell center and designated the region excluded from this area as the cellular peripheral region (revised Fig. 9A). Subsequently, we quantified the proportion of lysosomes contained within that region in cells expressing flag-JIP4 (revised Fig. 9B). We have added this experimental data in Fig. 5E.

A B

__Revised Fig. 9. __

1. Approximately 70% of the cell area toward the cell center was automatically delineated using ImageJ, with the region excluded from this defined as the cellular peripheral region.
2. JIP4 KO cells were transfected with flag-tagged JIP4 (wild-type and T217A) for 24 h and treated with teniposide (10 μM) for 4 h. Cell lysates underwent SDS-PAGE and were immunoblotted with anti-JIP4 and anti-actin The graph displays the percentage of cells with peripheral lysosomes. Data are expressed as mean ± standard deviation. *P 20).

   9) In Fig.6A, the total amount of JIP4 seems to change in some agent treatments, which needs to be explained.

RESPONSE:

As per our response to Reviewer 1, we evaluated the decrease in JIP4 expression by WB after 4 h of treatment with 10 μM teniposide. The teniposide-induced decrease of JIP4 was suppressed by bafilomycinA1 co-treatment, indicating that JIP4 was degraded by teniposide-induced autophagy (revised Fig. 3). We have added the data in Fig. S6, and the related description of these experiments have been added to the Results section (page 8, lines 289 to 293).

10) In Fig.7C and D, the effect of drug treatment on the amount of ubiquitinated proteins should also be checked.

RESPONSE:

We have included ubiquitin protein blots in Fig. 7C and 7D.

11) In Fig.8B, it is described that lysosomes are more localized in αSyn by drug treatment, but more convincing images and quantitative data are needed.

RESPONSE:

. The colocalization of LAMP2 and aSyn-GFP aggregates was assessed by measuring the fluorescence values of lysosomes in contact within the aSyn-GFP aggregation area using ImageJ. We have added this quantified data in Fig. 8D.

Reviewer #2 (Significance (Required)): Although the reviewer appreciates the discovery of novel drugs to induce autophagy through regulating lysosomal positioning, the detailed action of these compounds and their superiority in the field are not clear.

__Reviewer #3 (Evidence, reproducibility and clarity (Required)):____ __ In this manuscript, Date et al. sought to identify compounds that promote protein aggregates clearance - in particular those formed by mutant alpha synuclein. Briefly the authors screened a library of clinically approved compounds for inducers of lysosomal clustering followed by a secondary screen for autophagy inducers. By this two-step procedure, the authors identified three topoisomerase inhibitors and three anthelmintics as hits. Next, the authors unveiled that lysosomal clustering induced by these compounds is independent of mTORC1 but requires TRPML1 and JIP4. Moreover, the topoisomerase inhibitors hits involved phosphorylation of JIP4 while the anthelmintics additionally required Rab7 and ALG2. Intriguingly, the authors found that lysosomal clustering was prerequisite to autophagy induction. Focusing on the class of anthelmintics (i.e. albendazole) the authors showed that these induce autophagy to degrade aggregates formed upon proteasome inhibition. Lastly, the authors demonstrated that albendazole also led to increased degradation of αSyn aggregates through autophagy induction.

Major points 1) Most importantly, the authors need to tone down the significance of their findings throughout the manuscript. For examples, they should restrain from using "nullified" when it is really reduced only by 10-25 %.

RESPONSE:　

We have changed the description in the manuscript according to the Reviewer’s suggestion.

2) The authors claim that the topoisomerase inhibitors led to JIP4 phosphorylation while Figure 5C actually shows the opposite (partially reduced phosphorylation compared to DMSO treatment) and the Jak3 inhibitor has no obvious effect. The authors should quantify the phostag results.

RESPONSE:

We agree with the Reviewer that the Phos-tag PAGE results of JIP4 in Fig. 5C is complicated, and the bands were not clear. We have replaced these with more robust data (Fig. 5C).

3) Figure 6A/B: Why do all compounds except Mebendazole affect the abundance of JIP4?

RESPONSE:

As per our response to Reviewer 1, we evaluated the decrease in JIP4 expression by WB after 4 h of treatment with 10 μM teniposide. The teniposide-induced decrease of JIP4 was suppressed by bafilomycinA1 co-treatment, indicating that JIP4 was degraded by teniposide-induced autophagy (revised Fig. 3). We have added the data in Fig. S6, and the related description of these experiments have been added to the Results section (page 8, lines 289 to 293).

4) Figure 7C: The blot is not convincing. The authors should quantify this effect.

RESPONSE:

We evaluated and confirmed the degradation of p62 by albendazole, as shown in Fig. 7C.

Reviewer #3 (Significance (Required)):

Overall, the work of Date and colleague highlights the role of lysosomal clustering in clearing protein aggregates. Importantly, the identified classes of compounds might open new avenues for rationalizing treatment strategies for neurodegenerative diseases. However, several critical points remain.

• *

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #3

Evidence, reproducibility and clarity

In this manuscript, Date et al. sought to identify compounds that promote protein aggregates clearance - in particular those formed by mutant alpha synuclein. Briefly the authors screened a library of clinically approved compounds for inducers of lysosomal clustering followed by a secondary screen for autophagy inducers. By this two-step procedure, the authors identified three topoisomerase inhibitors and three anthelmintics as hits. Next, the authors unveiled that lysosomal clustering induced by these compounds is independent of mTORC1 but requires TRPML1 and JIP4. Moreover, the topoisomerase inhibitors hits involved phosphorylation of JIP4 while the anthelmintics additionally required Rab7 and ALG2. Intriguingly, the authors found that lysosomal clustering was prerequisite to autophagy induction. Focusing on the class of anthelmintics (i.e. albendazole) the authors showed that these induce autophagy to degrade aggregates formed upon proteasome inhibition. Lastly, the authors demonstrated that albendazole also led to increased degradation of αSyn aggregates through autophagy induction.

Major points

1. Most importantly, the authors need to tone down the significance of their findings throughout the manuscript. For examples, they should restrain from using "nullified" when it is really reduced only by 10-25 %.
2. The authors claim that the topoisomerase inhibitors led to JIP4 phosphorylation while Figure 5C actually shows the opposite (partially reduced phosphorylation compared to DMSO treatment) and the Jak3 inhibitor has no obvious effect. The authors should quantify the phostag results.
3. Figure 6A/B: Why do all compounds except Mebendazole affect the abundance of JIP4?
4. Figure 7C: The blot is not convincing. The authors should quantify this effect.

Significance

Overall, the work of Date and colleague highlights the role of lysosomal clustering in clearing protein aggregates. Importantly, the identified classes of compounds might open new avenues for rationalizing treatment strategies for neurodegenerative diseases. However, several critical points remain.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

In this study, the authors focused on lysosome positioning and autophagy activity to search for novel agents effective against Parkinson's disease. As a result, several compounds were successively identified, including Topoisomerase inhibitors and Benzimidazole. Authors showed that these agents regulate lysosomal positioning through different pathways but commonly require JIP4 to regulate lysosomal positioning and subsequent autophagy. They also showed that albendazole treatment promoted the degradation of insoluble ubiquitinated proteins and αSyn in cultured cells.

Major Comments.

• Two compounds, for instance teniposide and albendazole both requires JIP4 and/or TRPML1 to regulate lysosomal positioning and autophagy but their action seems different. What is the actual mechanism by which these compounds require JIP4/TRPML1. How inhibition of Topoisomerase leads to increase of JIP4 phosphorylation? Do teniposide and albendazole both affect calcium release from TRPML1?
• The authors should clarify the functional advantage of these drugs identified in this study as drugs for Parkinson's disease by comparing with known autophagy inducers such as Torin1 or rapamycin.
• Related to the previous question, in Fig.6A and B additional data comparing novel compounds with established autophagy inducers, such as torin1 and rapamycin, should be included and discussed.
• The authors should examined whether increased degradation of insoluble proteins and αSyn are dependent on JIP4.
• Authors only utilized. SH-SY5Y cells in this study. It is important to examine whether these compounds also regulate lysosomal positioning and autophagy in other cell lines.
• The authors conclude that the six compounds do not mediate mTOR signaling in Fig. 3, but should more carefully describe in the manuscript why they performed this experiment and what the results mean for.

Minor comments.

• The name of the compound should be written in the red point of Fig.2A.
• Regarding images of Fig.2B, the magnified images and quantitative data should be added.
• The results of Fig.2C need to be explained more carefully. A quantitative data is missing.
• Fig.S2, which compares autophagy activity with conventional agents, should be quantified and added to the Fig.3.
• In the statistical analysis of Fig.4B, the clustering value was increased by siRILP, which should be briefly described in the manuscript.
• In Fig.4A and B, it is possible that the knockdown efficiency of siRILP and siTMEM55B was not sufficient to observe the effect on lysosomes, and this concern should be described in the manuscript.
• The authors should add the results of the WB experiment showing the amount of JIP4 protein in Fig.5G.
• In Fig.5F, images of JIP4KO cells that do not express FLAG-JIP4 should be added as controls, and further quantification should be done on cells in all three conditions.
• In Fig.6A, the total amount of JIP4 seems to change in some agent treatments, which needs to be explained.
• In Fig.7C and D, the effect of drug treatment on the amount of ubiquitinated proteins should also be checked.
• In Fig.8B, it is described that lysosomes are more localized in αSyn by drug treatment, but more convincing images and quantitative data are needed.

Significance

Although the reviewer appreciates the discovery of novel drugs to induce autophagy through regulating lysosomal positioning, the detailed action of these compounds and their superiority in the field are not clear.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

The manuscript by Date et al employed a cell model by stably expressing LGP120-mCherry and GFP-gamma-tubulin to carry out high-content screening in search of chemical compounds that enhance lysosomal clustering and autophagy. They found 6 clinically approved drugs categorized as topoisomerase II inhibitors and the benzimidazole class. They further validated these compounds by a set of well-designed experiments including autophagy flux assays and mTOR dependence. In the mechanistic study, they demonstrated the compounds induce lysosomal clustering in a JIP4-TRPML1-dependent manner. In a PD cell model, one of the compounds albendazole exhibited the effect on boosting the degradation of insoluble alpha-synuclein. The study is of interest, and the cell model and the approach generated by the authors would be transferable for future studies of other high-content imaging screening. Most of the data is clear and convincing.

Major comments:

1. In addition to its role in facilitating a-syn turnover by autophagy, Is the chemical protective against a-syn toxicity?
2. Please elaborate why albendazole does not change the levels of soluble a-syn, but those of insoluble, as shown Fig 8D.
3. Fig 6A shows that some of the compounds (Teniposide, Amsacrine) affect the levels of JIP4. Can albendazole also reduce JIP4 levels. It might be interesting to test this, as JIP4 is important for lysosomal clustering.

Minor comments:

The writing is good generally. Please tide up the text in a few occasions to make the expressions more formal.

Significance

Significance: The study generated a new approach for high-throughput screening of compounds to enhance lysosomal clustering.

Audience: Basic and clinical research

Expertise: Programmed cell death, neurodegenerative diseases

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

#Reviewer 1 (Evidence, reproducibility and clarity):

This manuscript by Deshmukh et al is aimed at generating chimeric antigens that can be useful for making next generation vaccines that block blood stage infection by malaria parasite. Given that there is no blood stage vaccine against malaria and available liver stage vaccine shows only limited efficacy that too only in Africa, there is dire need for having novel approaches to generate successful vaccines. In the past attempts have been made to make multivalent vaccines but have not been successful. Nevertheless, it is still a good option as single target blood-stage vaccines have failed. Authors propose to target cytoadhesion and host erythrocyte invasion. For this purpose, they have selected epitopes from PfEMP1/VarB family members, which poses a major challenge as at least 60 genes encode them and they exhibit variations which facilitate the escape from the immune system. The other two chimeras target invasion related proteins like MSPs and adhesins shed by micronemes and rhoptries, which are critical for invasion. The reported work is interesting and provides a useful approach towards developing vaccines against blood stage infection.

We appreciate the time and effort given by our reviewer in thoroughly reading the manuscript. We are thankful for all the comments and suggestions for better shaping the article.

Comments:

1. __ The peptides used in InvB chimera did not show good reactivity especially when compared to VarB or MSP peptides. Please discuss the possible reasons.__

Response: Thank you for pointing out the difference in the explanation. With chimeric InvP, we see a strong response against a few peptides of SERA-5 and RH-5, while other peptides, in comparison, have lesser antibody responses. We have now included the following statement detailing this difference with possible explanations in the revised manuscript (Page 8, Line 25 to 30).

The IgG responses to chimeric InvP were slightly different from those to chimeric varB and MSP. The intensity of IgG to peptides of SERA-5 and RH-5 was very high in comparison to the rest of the peptides used in the construct, whereas in chimeric varB and MSP, the IgG titers were comparable between the peptides. This could be a result of antigen exposure in the cohort of 19 patient samples that we used, and may change when a larger sample size is considered.

__ It will be interesting to determine if blocking a specific VarB/PfEMP1 alters expression of other members. Based on the data provided in Fig. 4E, can a chimera be designed which only includes PfEMP1 that are represented well in HBEC-5i population?__

Response: We agree that observing the altered expression of PfEMP1 would be an interesting phenomenon to study. The blocking of PfEMP1 using anti-chimeric varB antibodies is a transient process in our assays (just enough to quantify the cytoadhesion). It may take multiple cycles with negative selection pressure on parasites for the switching to take place. Also, it will be interesting to design chimeras based on the HBEC-5i binding PfEMP1. We can certainly plan these as prospective future experiments.

__ Some of the invasion related proteins like RH5 and EBA175 are not present at parasite surface, instead, secreted from rhoptries and micronemes. It will be nice to perform Western blots on condition medium and see if InvP (or even MSP and VarB) antibodies recognizes the secreted version of these proteins.__

Response: We thank the reviewer for this valuable comment and the suggestive experiment. We will perform a western blot on spent media and probe using anti-chimeric MSP and InvP antibodies to detect the proteins selected in chimeric MSP and InvP antigens.

__ Fig. 6E- Statistics need to be provided for inhibition at 12.3 and 25ug.__

Response: We apologize for the missed statistics. It is now included in the figure panel.

__ Plasmodium uses multiple ligand-receptor interaction, which could depend (e.g. EBA-glycohophorins) or operate independent (e.g. RH5-basigin) of sialic acid. While there is representation from candidates from both of these families, most studies especially growth rate assays (Fig. 6E) have been carried using 3D7 strain, which does not require sialic acid. It is possible that if similar experiments were performed using sialic acid-sensitive strains, InvP and MSP antibodies may cause greater inhibition of parasite growth, which may be worth testing.__

Response: We are grateful for the suggestion of using a sialic acid-dependent strain. Indeed, the pathway of reinvasion chosen by the parasite may determine the growth inhibition assay (GIA) outcome. We will perform the GIA assay on the Dd2 strain and 3D7 with neuraminidase treatment (Sialic acid-dependent invasion). We will also note the difference in growth inhibition potential of chimeric antibodies in sialic-acid dependent and independent pathways.

__ The direct effect of InvP and MSP Abs should be tested directly on host erythrocyte invasion.__

Response: We thank the reviewer for this comment. We certainly can determine the inhibitory potential of anti-chimeric MSP and InvP antibodies through invasion assays. We will include the invasion inhibition potential of these antibodies in 3D7, Dd2, with neuraminidase treatment along with GIA data.

Reviewer #1 Significance:

Present study proposes novel strategies for the development of anti-malarial vaccine.

#Reviewer 2 (Evidence, reproducibility and clarity):

The manuscript describes the vaccine potential of unstructured P. falciparum merozoite protein fragments 25 amino acid long belonging to 3 different protein families. The work is well performed, easily reproducible and clearly described.

We appreciate the time and effort given by our reviewer in thoroughly reading the manuscript. We are thankful for all the comments and suggestions for better shaping the article.

Reviewer #2 (Significance):

1. The use of protein fragments whose structure can be predicted by their sequence has been exploited in many studies for the development of vaccines or other biologicals. In this studies the authors selected 3 different families belonging to the red blood stage of the parasite. The table showing the sequences selected is not readable and should be clearly provided in the supplementary section.

Response: We apologize for the readability of the sequences. The supplementary Table 1 has the proteins selected, the sequences taken, and the precise order for the stitching.

In addition, polymorphic residues should be highlighted.

Response: We thank the reviewer for pointing this out. We will analyze and compile the protein sequences in 3D conformation, highlighting polymorphic residues and the peptides selected in our study.

In addition, it is not to mention why the authors used immune rabbit sera obtained by injection of the 3 poly-epitopes instead of obtaining by affinity chromatography antigen specific human antibodies from sera of individuals living in endemic regions which could provide a direct and clear answer whether a protective vaccine could be obtained.

Response: We agree that the clear answer to the protective function of antibodies could have been answered using human antibodies. However, we did not have a sufficient volume of patient sera to perform affinity enrichment. The use of rabbits here was to ensure the generation of antigen-specific antibody responses in ample amounts. The patient sera in quantities available were used in ELISA, epitope mapping, and IP, followed by mass-spectrometry. The IP-MS clearly shows the presence of antibodies against the proteins taken in the generation of chimeric antigens (Supplementary Figure 1 D).

#Reviewer 3 (Evidence, reproducibility and clarity):

Multi-protein chimeric antigens... by: Deshmukh et al

This article addresses an extremely important objective, the development of an effective prophylactic vaccine for Malaria. The disease continues to be widespread claiming the lives of hundreds of thousands of people annually, many of them children. Despite efforts towards producing Malaria vaccines, none thus far have been sufficiently protective or long term. As the authors point out vaccines can target the parasite per se, and possibly more attractive would be to focus on parasite derived antigens expressed on the surface of infected erythrocytes, hence targeting the Blood stage of the infection, which is most directly associated with Malaria pathogenesis. The authors propose a somewhat novel approach in which they have selected an array of short (25 amino acids) segments of Plasmodium derived proteins stitched together to produce 3 chimeric recombinant proteins as potential immunogens. Although a considerable amount of work is described, the results are not compelling in proving the efficacy or advantage of using chimeric antigens as worthy vaccine candidates for Malaria.

Unfortunately, the rationale behind the experiments are not clearly defined which is a matter of concern. In addition, details of the work done and the technical aspects needs to better explained to fully understand how and why the target segments were selected and the chimeras produced. This review focuses first on scientific issues and then format and editing, both aspects demonstrate that the manuscript in its present form requires major changes for it to be of relevance to the field. This review focuses first on issues of substance and then format and editing, both aspects disqualify the publication of the manuscript in its present form.

We appreciate the time and effort given by our reviewer in thoroughly reading the manuscript. We are thankful for all the comments and suggestions for better shaping the article.

Experiments and Results:

1. The underlying proposal claims that chimeric antigens might be advantageous in eliciting protective antibodies. The authors produced three chimeras: var, MSP and InvP. __The var chimera contains 29 segments of PfEMP1 derived from 8 alleles. The hypothesis is that by expressing 29 different segments one will produce antibodies that can better cope with the antigenic diversity of this target. Indeed, serial monoallelic expression of anyone of the 60 PfEMP1 variants of a given P. falciparum strain has been thought to mediate immune evasion. The parasite is presumed to be able to escape immune defenses, by switching and serially expressing PfEMP1 alleles. Hence, one might assume that by introducing different segments, derived from different alleles, one will gain better protection. The authors have not really tested this idea. They have produced a single chimera and tested it without controlled comparison of performance to any single segment, or for that matter compared to alternative structural domain(s) of PfEMP. This brings me to the question of how the segments were selected and why. The authors implement IEDB-AR to identify presumably preferred B-cell epitopes. The methodology relies on a number of computational methods that predict the propensity of linear segments of proteins to have, for example, secondary structures, or be surface accessible, or relatively hydrophilic or flexible, etc. IEDB-AR is a tool to assist the identification of segments (5-25 amino acids in length), that might be associated with B-cell epitopes, or at least segments comprising linear aspects of B-cell epitopes. The input is a linear sequence of an antigen, proposing linear aspects of what could be associated with B-cell epitopes. B-cell epitopes, however, are typically conformational and discontinuous. They certainly can and do contain linear segments, but even these may require 3D conformations dictated by spatial constraints imposed by the native surrounding aspects of the natural antigen. It is hard to assume that by simply stitching 29 segments, one after the other, one can provide them with the native environment for them to assume a somewhat physiologically relevant conformation. Unfortunately, the authors have not addressed the unique characteristics of the antigen they have selected. PfEMP1, for example, is a family of antigens with discrete sub-domain structures and features (DBL and CIDR for example). It would be relevant and useful to relate the segments that they chose to the natural unique domains of the antigen and how they might best present common vs variant aspects of the antigen. There are at least 30 crystal atomic structures for PfEMP1 in complex with various physiologically relevant proteins (eg ICAM etc). The authors might have considered the 3D structure of PfEMP in their analyses and at least indicated on an atomic structure where the 29 segments lie. __

More concerning is the fact that the expression of the chimera does not produce a crisp single protein, but rather a complex of products as illustrated in the Supplement Figure 1 B. The authors simply claim that they produce the antigen for immunization of rabbits (or one rabbit?) and they collect gel-derived band(s) of what MW?? Assuming that a 25aa segment should be about 2500-2800 daltons and so 29 such segments strung together should be about 80kDa. The gel shows bands at 124kDa, and a slew of bands shorter than 71kDa. There is no mention what the expected MW should be and there is no explanation why the protein pattern contains so many bands of different sizes and what exact bands were taken for the immunogen or why.

Response: We thank the reviewer for this comment, as it tells us the reader’s perspective on how the chimeric construct part is underexplained. We have now expanded the section on chimeric construct design, the sequences used, the functional domains they belong to in the PfEMP1 protein (Supplementary Tables 1 and 2), and the expected sizes of the proteins created. As for the B-cell epitope prediction, we have used the linear epitope prediction tool. However, we will include a 3D conformational study highlighting the placement of peptides that we have used to generate chimeric antigens.

The sequences for chimeric constructs were synthesized commercially and confirmed using Sanger sequencing. The antigens run higher than their expected molecular weights, and we have confirmed them through western blot and mass spectrometry (Supplementary Figure 1 B and C). The chimeric varB antigen specifically shows a cleaving pattern, hence the multiple bands in western blotting (we have considered the top-most band with the highest anti-his intensity). After these confirmations, the antigens were independently injected in rabbits to generate antibodies.

Similar considerations can be made regarding the selection of the segments for the two other chimeras, although they seem to produce a single polypeptide.

Response: The antigens were confirmed using Sanger sequencing, expression using anti-his western blot, and proteins were confirmed using mass spectrometry for all three chimeric constructs (Supplementary Figure 1 B and C).

If the point was to test a "chimera" modality as an improved vaccine, it would have been more useful to focus on one chimera and carefully characterize it and compare it to its components used separately.

Response: The idea of chimera arises from the fact that individual proteins/components are insufficient to generate optimal responses. The proteins considered in our study have already been validated in the field (as separate components) and show that the efficacy observed was sub-optimal. Since our rationale is to include multiple proteins to tackle the redundancy and parasite virulence, we have focused on generating three chimeric constructs covering the entire blood stage of Plasmodium falciparum. Our objective is to demonstrate that a multi-protein, multi-factorial vaccine, as a proof of concept, works better in tackling malaria. We believe that in proving so, a comparison of chimera with their individual components is an unnecessary and economically unviable.

The authors devote much effort to the fluidics system and their assay. This might warrant a paper dedicated to the methodology they have developed.

Response: The Plasmodium virulence genes are extensively studied for their interactions with human endothelial receptors. Unfortunately, these studies fail to take human physiological conditions into account. We wanted to test our anti-chimeric varB antibodies in the best mimicking environment possible. Hence, the efforts were devoted to developing, standardizing, and quantifying the fluidic cytoadherence system. We thank the reviewer for their kind words of encouragement on our methodology.

Format and Editing:

1. The manuscript is very poorly written with multiple errors throughout. The authors use abbreviations that are not defined, eg iRBC (pg 5 line 22) or sometimes incorrectly defined, eg MSP ("merozoite-specific proteins - pg 6 line 18).

Response: We apologize for the abbreviation error. The abbreviation for iRBC is defined in the introduction section (page no 4, Line 15); hence, it is not redefined on page 5, line 22. We have corrected merozoite-specific proteins on page 6, line 18.

The Figures are of low resolution to the extent that they can not be read (for example Figure 3 pg 34). Figure 1 is somewhat useless and misleading. In Fig1 C - the diagram illustrates 5 hypothetical chimeras where in fact only three were produced. There really is no detail or explanation as to how the chimeras were produced.

Response: We apologize for the low resolution of the images. We have now improved the image quality. Figure 1C represents the idea of designing the construct, not the number of chimeras we generated. We apologize for this confusion and have explicitly mentioned this in the figure panel for Figure 1C. As for the design and generation of chimeric antigens, we understand that the materials and methods section is underexplained, and we have now expanded on it with all details included.

In the construction of the chimeras there is no mention as to whether short linkers were introduced between the segments or not. What was the expected weight of the chimera? Was the order of segments random or precise and consistent? Were the constructs sequence validated in addition to the MassSpec?

Response: We understand that the section on the chimeric construct is underexplained for the readers, and we thank the reviewer for pointing it out. We have now expanded the section on chimeric antigen design and included the details. Chimera was tested with GSGSGS linkers and without linkers for expression. The final antigen injected in rabbits was serially attached peptides without linkers. The segments stitched were in precise order, as mentioned in Supplementary Sheet 1. The construct was commercially synthesized and sequence validated along with the anti-his western blot and mass spectrometry analysis.

The figures of the Supplement are not numbered.

Response: We thank the reviewer for pointing this out. The figures are now numbered.

Note that the headings in Supplement Figure 1 B and C have overlapping text.

Response: Thank you for pointing this out. We have now rearranged the supplementary figures 1B, and 1C.

Most disturbing is that multiple references that are incomplete. For example: in References 15, 16, 25, 26, 27 there is no indication of the Journal.

Response: We apologize for the mistakes in referencing. These references did not have full citations in Endnote. We have now manually checked all the references and corrected the incomplete formats of the references.

The authors mention reference 13 [2006] in claiming that the antibodies can be protective, and then support this by referring to refs 14, 15 and 16 published in 1961, 1963 and 1962 respectively. Although, old articles can be useful, but the authors should attempt to provide current proof of such basic claims.

Response: We thank the reviewer for pointing this out. We have now separated these two statements and not mentioned the latter as a support to the former. As for references 14, 15, and 16, these were the early studies in the field that show the protective nature of antibodies through the passive immunization process and are foundations for the idea of blood stage vaccination. Current proofs of antibodies against blood-stage antigens are included for blood-stage vaccine candidates.

Reviewer #3 (Significance):

The goal of the study is very important.

The hypothesis that a chimeric presentation of select peptides could be advantageous was not rigorously tested nor well controlled in a meaningful evaluation and thus no conclusion can be made. There are no comparative analyses to test their hypothesis.

The method for selection of epitope segments is not well justified. There is little attempt to provide rationale or description of the segments chosen and how they fit within the antigens, thus justifying segments over multiple antigens.

The grammatical errors, lack of clarity accompanied by little attention to style and readability render the manuscript quite illegible.

There is no excuse for so many errors in the references.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #3

Evidence, reproducibility and clarity

Multi-protein chimeric antigens... by: Deshmukh et al

This article addresses an extremely important objective, the development of an effective prophylactic vaccine for Malaria. The disease continues to be widespread claiming the lives of hundreds of thousands of people annually, many of them children. Despite efforts towards producing Malaria vaccines, none thus far have been sufficiently protective or long term. As the authors point out vaccines can target the parasite per se, and possibly more attractive would be to focus on parasite derived antigens expressed on the surface of infected erythrocytes, hence targeting the Blood stage of the infection, which is most directly associated with Malaria pathogenesis. The authors propose a somewhat novel approach in which they have selected an array of short (25 amino acids) segments of Plasmodium derived proteins stitched together to produce 3 chimeric recombinant proteins as potential immunogens. Although a considerable amount of work is described, the results are not compelling in proving the efficacy or advantage of using chimeric antigens as worthy vaccine candidates for Malaria.

Unfortunately, the rationale behind the experiments are not clearly defined which is a matter of concern. In addition, details of the work done and the technical aspects needs to better explained to fully understand how and why the target segments were selected and the chimeras produced. This review focuses first on scientific issues and then format and editing, both aspects demonstrate that the manuscript in its present form requires major changes for it to be of relevance to the field. This review focuses first on issues of substance and then format and editing, both aspects disqualify the publication of the manuscript in its present form.

Experiments and Results:

The underlying proposal claims that chimeric antigens might be advantageous in eliciting protective antibodies. The authors produced three chimeras: var, MSP and InvP.

The var chimera contains 29 segments of PfEMP1 derived from 8 alleles. The hypothesis is that by expressing 29 different segments one will produce antibodies that can better cope with the antigenic diversity of this target. Indeed, serial monoallelic expression of anyone of the 60 PfEMP1 variants of a given P. falciparum strain has been thought to mediate immune evasion. The parasite is presumed to be able to escape immune defenses, by switching and serially expressing PfEMP1 alleles. Hence, one might assume that by introducing different segments, derived from different alleles, one will gain better protection. The authors have not really tested this idea. They have produced a single chimera and tested it without controlled comparison of performance to any single segment, or for that matter compared to alternative structural domain(s) of PfEMP. This brings me to the question of how the segments were selected and why. The authors implement IEDB-AR to identify presumably preferred B-cell epitopes. The methodology relies on a number of computational methods that predict the propensity of linear segments of proteins to have, for example, secondary structures, or be surface accessible, or relatively hydrophilic or flexible, etc. IEDB-AR is a tool to assist the identification of segments (5-25 amino acids in length), that might be associated with B-cell epitopes, or at least segments comprising linear aspects of B-cell epitopes. The input is a linear sequence of an antigen, proposing linear aspects of what could be associated with B-cell epitopes. B-cell epitopes, however, are typically conformational and discontinuous. They certainly can and do contain linear segments, but even these may require 3D conformations dictated by spatial constraints imposed by the native surrounding aspects of the natural antigen. It is hard to assume that by simply stitching 29 segments, one after the other, one can provide them with the native environment for them to assume a somewhat physiologically relevant conformation. Unfortunately, the authors have not addressed the unique characteristics of the antigen they have selected. PfEMP1, for example, is a family of antigens with discrete sub-domain structures and features (DBL and CIDR for example). It would be relevant and useful to relate the segments that they chose to the natural unique domains of the antigen and how they might best present common vs variant aspects of the antigen. There are at least 30 crystal atomic structures for PfEMP1 in complex with various physiologically relevant proteins (eg ICAM etc). The authors might have considered the 3D structure of PfEMP in their analyses and at least indicated on an atomic structure where the 29 segments lie. More concerning is the fact that the expression of the chimera does not produce a crisp single protein, but rather a complex of products as illustrated in the Supplement Figure 1 B. The authors simply claim that they produce the antigen for immunization of rabbits (or one rabbit?) and they collect gel-derived band(s) of what MW?? Assuming that a 25aa segment should be about 2500-2800 daltons and so 29 such segments strung together should be about 80kDa. The gel shows bands at 124kDa, and a slew of bands shorter than 71kDa. There is no mention what the expected MW should be and there is no explanation why the protein pattern contains so many bands of different sizes and what exact bands were taken for the immunogen or why.

Similar considerations can be made regarding the selection of the segments for the two other chimeras, although they seem to produce a single polypeptide.

If the point was to test a "chimera" modality as an improved vaccine, it would have been more useful to focus on one chimera and carefully characterize it and compare it to its components used separately. The authors devote much effort to the fluidics system and their assay. This might warrant a paper dedicated to the methodology they have developed.

Format and Editing:

The manuscript is very poorly written with multiple errors throughout. The authors use abbreviations that are not defined, eg iRBC (pg 5 line 22) or sometimes incorrectly defined, eg MSP ("merozoite-specific proteins - pg 6 line 18).

The Figures are of low resolution to the extent that they can not be read (for example Figure 3 pg 34). Figure 1 is somewhat useless and misleading. In Fig1 C - the diagram illustrates 5 hypothetical chimeras where in fact only three were produced. There really is no detail or explanation as to how the chimeras were produced.

In the construction of the chimeras there is no mention as to whether short linkers were introduced between the segments or not. What was the expected weight of the chimera? Was the order of segments random or precise and consistent? Were the constructs sequence validated in addition to the MassSpec?

The figures of the Supplement are not numbered.

Note that the headings in Supplement Figure 1 B and C have overlapping text.

Most disturbing is that multiple references that are incomplete. For example: in References 15, 16, 25, 26, 27 there is no indication of the Journal.

The authors mention reference 13 [2006] in claiming that the antibodies can be protective, and then support this by referring to refs 14, 15 and 16 published in 1961, 1963 and 1962 respectively. Although, old articles can be useful, but the authors should attempt to provide current proof of such basic claims.

Significance

The goal of the study is very important.

The hypothesis that a chimeric presentation of select peptides could be advantageous was not rigorously tested nor well controlled in a meaningful evaluation and thus no conclusion can be made. There are no comparative analyses to test their hypothesis.

The method for selection of epitope segments is not well justified. There is little attempt to provide rationale or description of the segments chosen and how they fit within the antigens, thus justifying segments over multiple antigens.

The grammatical errors, lack of clarity accompanied by little attention to style and readability render the manuscript quite illegible.

There is no excuse for so many errors in the references.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

The manuscript describes the vaccine potential of unstructured P. falciparum merozoite protein fragments 25 amino acid long belonging to 3 different protein families. The work is well performed, easily reproducible and clearly described.

Referees cross-commenting

The polymorphic residues should be highlighted in the supplementary figure.

Significance

The use of protein fragments whose structure can be predicted by their sequence has been exploited in many studies for the development of vaccines or other biologicals. In this studies the authors selected 3 different families belonging to the red blood stage of the parasite. The table showing the sequences selected is not readable and should be clearly provided in the supplementary section. In addition, polymorphic residues should be highlighted. In addition, it is not to mention why the authors used immune rabbit sera obtained by injection of the 3 poly-epitopes instead of obtaining by affinity chromatography antigen specific human antibodies from sera of individuals living in endemic regions which could provide a direct and clear answer whether a protective vaccine could be obtained.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

This manuscript by Deshmukh et al is aimed at generating chimeric antigens that can be useful for making next generation vaccines that block blood stage infection by malaria parasite. Given that there is no blood stage vaccine against malaria and available liver stage vaccine shows only limited efficacy that too only in Africa, there is dire need for having novel approaches to generate successful vaccines. In the past attempts have been made to make multivalent vaccines but have not been successful. Nevertheless, it is still a good option as single target blood-stage vaccines have failed. Authors propose to target cytoadhesion and host erythrocyte invasion. For this purpose, they have selected epitopes from PfEMP1/VarB family members, which poses a major challenge as at least 60 genes encode them and they exhibit variations which facilitate the escape from the immune system. The other two chimeras target invasion related proteins like MSPs and adhesins shed by micronemes and rhoptries, which are critical for invasion. The reported work is interesting and provides a useful approach towards developing vaccines against blood stage infection.

Comments:

1. The peptides used in InvB chimera did not show good reactivity especially when compared to VarB or MSP peptides. Please discuss the possible reasons.
2. It will be interesting to determine if blocking a specific VarB/PfEMP1 alters expression of other members. Based on the data provided in Fig. 4E, can a chimera be designed which only includes PfEMP1 that are represented well in HBEC-5i population?
3. Some of the invasion related proteins like RH5 and EBA175 are not present at parasite surface, instead, secreted from rhoptries and micronemes. It will be nice to perform Western blots on condition medium and see if InvP (or even MSP and VarB) antibodies recognizes the secreted version of these proteins.
4. Fig. 6E- Statistics need to be provided for inhibition at 12.3 and 25ug.
5. Plasmodium uses multiple ligand-receptor interaction, which could depend (e.g. EBA-glycohophorins) or operate independent (e.g. RH5-basigin) of sialic acid. While there is representation from candidates from both of these families, most studies especially growth rate assays (Fig. 6E) have been carried using 3D7 strain, which does not require sialic acid. It is possible that if similar experiments were performed using sialic acid-sensitive strains, InvP and MSP antibodies may cause greater inhibition of parasite growth, which may be worth testing.
6. The direct effect of InvP and MSP Abs should be tested directly on host erythrocyte invasion.

Significance

Present study proposes novel strategies for the development of anti-malarial vaccine.

Transparent Peer Review

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

We thank the reviewers for their kind works and helpful insights and suggestions. Below, we have pasted the reviews (in italics), with our responses:

---

Reviewer #1 (Evidence, reproducibility and clarity (Required)):

The study provides insights into how polyploidization via endomitosis may arise in human hepatocytes by studying fetal liver cell line-derived organoids. Using live cell imaging and LSM microscopy, binculeation was consistently observed in two independent cell line systems, at frequencies seen in human liver and sensitive to pharmacological inhibition (GSK3i) and genetic manipulation (E2F7 & E2F8 editing). The findings presented are in line with earlier data, largely gathered studying rodents. The data is convincing and robust indicating that these systems can be used to study cause and consequences of polyploidy in human hepatocytes.

1. While the authors do suggest that they provide a mechanisms how polyploidy is initiated in human hepatocytes undergoing endomitosis, ie. loss of membrane association of membrane-anchoring proteins at the midbody (e.g. Anillin, RacGAP1), I do feel that the data provided is rather descriptive and does not address a particular mechanism that may account for loss of membrane anchoring. As such, the title is making a too strong point, as, in my point of view, it associates with loss of membrane anchorage, but may not drive endomitosis. Whether this is a "passive" process in response to changes in physical forces and tension, or regulated via signalling intermediates to initiate regression of the cleavage furrow is not addressed experimentally (mislocalizing these proteins on a larger scale). Discussion seems warranted.

We agree with the reviewer that our mechanistic insights into the molecular mechanisms of endomitosis are limited, and we cannot currently prove that the loss of membrane-anchoring drives endomitosis. We have therefore toned down this conclusion and changed the title to “Binucleated human hepatocytes arise through late cytokinetic regression during endomitosis M phase”. Furthermore, we have expanded the Discussion to reflect on the gaps in knowledge and speculate about possible molecular mechanisms of endomitosis, see pages 12-16 (in particular, lines 404-423, lines 433-443, and 445-472.

I do not see the need for additional experiments, as I believe the data is robust and introduces an interesting new model where the role of ploidy can be studied in human hepatocytes ex vivo. However, if the authors wish to extend their studies and document further similarities with pathways engaged in rodents, some E2F7/8 targets relevant for ploidy control such as Anillin or PIDDosome components, or, maybe MDM2 processing for p53 activation, could be tested in wt and E2F mutant cell lines.

Unfortunately, we have not been able to look at E2F7/8 targets and their expression in E2F mutant Hep-Orgs. We performed qPCRs for some cytokinesis regulators such as Ect2, RacGap1 and Mklp1 in Hep-Orgs, however these genes are so lowly expressed that we can hardly detect them. This is likely because these transcripts are only expressed in a short period of the cell cycle during S/G2 phase, whereas the vast majority of cells in Hep-Orgs are in G1. Therefore, differences in gene expression are very difficult (if not impossible) to detect by qPCR. We also tried to perform single molecule FISH on Hep-Orgs, which would allow us to quantify lowly expressed transcripts in single cells, however despite that the smFISH stainings work well on cholangiocyte organoids and intestinal organoids, we could not get good signals in Hep-Orgs. Taken together, we are unable at this point to look into downstream targets of E2F7/8.

A minor suggestion is to clarify the term M-CDK activity in the introduction, as it may not be fully intuitive to all readers; similarly, ploidy reversal is still controversial in the field, but it is stated as a given fact.

Thank you for these suggestions, we have clarified the term M-CDK on page 3, lines 60-61, and have rephrased the sentence on ploidy reversal on page 3, lines 81-82.

Reviewer #1 (Significance (Required)):

Polyploidy at the cellular and nuclear level is a key feature of hepatocytes albeit the physiological significance of the process is not entirely clear. Increased ploidy has been linked to cancer resistance in the liver, but may pose a threat to hepatocyte survival under conditions of repeated compensatory proliferation cycles. Curiously, during normal regeneration after single surgical intervention liver regeneration is not compromised, even though it may recover faster starting when starting from higher ploidy levels. Mechanistically, most data has been generates studying rodents where it is documented that the proliferation behaviour changes around the time of weaning in mice when hepatocytes start to fail cytokinesis and undergo endomitosis, leading to cellular and nuclear polyploidy. In rodents, insulin signalling / AKT appears involved as is the E2F network and p53, activated by the caspase-2-PIDDosome.

The model system introduced here will allow mechanistic studies in human organoids and help to increase our understanding of this process in steady state and under conditions of stress.

---

Reviewer #2 (Evidence, reproducibility and clarity (Required)):

Summary:

Polyploid cells arise within various human tissues by multiple different mechanisms. Here, Darmasaputra et al present a study of one such mechanism, endomitosis, in liver cells using fetal-derived human hepatocyte organoids. In this model, they demonstrate that binucleated cells arise through the late regression of the cytokinetic furrow prior to abscission. They identify a rare event in cytokinetic cells - loss of midbody association with the plasma membrane - that could explain the cytokinesis failure observed in a proportion of these cells. Finally, they show that loss of Wnt signalling increases the number of binucleation events in a manner that depends on E2F7 and E2F8, similar to what has been observed in murine hepatocytes.

Major comments:

This is a compelling and well-presented study. The data presented are high quality, the experiments are well described and controlled and the conclusions are convincing. I am particularly impressed by the technical effort that the authors must have put into obtaining high quality live and IF images of dividing cells within organoids and their careful documentation of what are very rare mitotic events. In addition, the manuscript is extremely well written and I found it a pleasure to read.

1. I do not think that there are additional experiments that are essential to justify the conclusions of the paper. However, I do have suggestions that I think would strengthen this work and increase its significance. As is, the authors present findings in two different areas: the documentation of cytokinesis failure in hepatocyte organoids and the role of Wnt and E2F7/8 on binucleation. It would be really nice if the two parts could be linked. For example, the authors could examine cell divisions in the organoids without Wnt either live or fixed and show that they have a higher proportion of cells undergoing cytokinetic regression or with membrane-midbody attachment defects. Alternatively, they could look at whether the expression levels of key cytokinetic genes are changed in the Wnt and E2F7/8 organoids. As I said, these experiments are not required for or the publication of this work and I will leave it up to the authors to decide if they have the time or capacity to add additional data.

We thank the reviewer for this suggestion. Unfortunately, despite substantial effort, we have been unable to perform successful live imaging of Hep-Orgs under CHIR99021 removal conditions: these organoids become very sensitive to live imaging and they also proliferate very slowly. We have tried to look at the expression of cytokinetic genes by qPCR, however these experiments were inconclusive (see also our response to reviewer #1, point 2). Thus, we cannot rule out that the increase in binucleation that we see upon CHIR99021 removal is not due to increased endomitosis, but rather occurs independently, for example by an increased survival rate of binucleated cells upon WNT removal. We have now discussed this issue and explained the limitations of our study in the discussion, pages 14-15, lines 451-460.

Finally, before publication, the authors should discuss further the mechanisms by which loss of membrane attachment during cytokinesis could occur - there is quite a lot of literature in this area on the role of RacGAP1 and Ect2 in membrane attachment that is not discussed, particularly from the lab of Mark Pentronczki (eg Kotynkova 2026 PMID: 27926870, Lekmotsev PMID: 23235882). It's surprising that the authors haven't mentioned (or looked at) Ect2 at all, especially since Ect2 levels have been shown to control polyploidy in cardiomyocytes (Liu 2019 PMID: 31597755). This at least warrants some discussion.

We thank the reviewer for pointing us to these articles. We have elaborated the discussion to include the work on rodent and human cardiomyocytes, and to explain why we think that there is no defect in ECT2 and RhoA signaling in human hepatocytes undergoing endomitosis, see pages 13-14, 404-423 and 433-443.

Minor comments:

Table 1 would be more striking as a graphical representation. I appreciate that the n numbers in the regressed cells means that statistical comparisons is not possible, but some kind of colour coding or graph would make this part clearer

We agree that Table 1 was difficult to read – we now show the data schematically in a new figure, Fig.4.

It's not clear what the difference between Hep-Org 1 and Hep-Org 2 are. Are these from different donors?

Indeed Hep-Org1 and Hep-Org2 are from different donors. We have clarified this in the text, see page 5, lines 131-133.

Reviewer #2 (Significance (Required)):

This study is an important technical development in that it reports a new system to study in depth cell biology of liver endomitosis in non-transformed and, crucially, human 3D hepatocyte organoids. The findings reported using this system are potentially interesting although they could be further developed if they were mechanistically linked together (see major comments). This work is likely to be highly interesting to scientists studying cell division, cytokinesis and hepatocyte biology. It also has wider implications for liver biology and particularly liver regeneration. Additionally, given the role of polyploidisation in many different tissues, it will likely be of interest to scientists studying polyploidy and endomitosis more generally.

My area of expertise is in cytokinesis and cell division in general, although not specifically in hepatocytes. I am not an expert in organoids.

---

Reviewer #3 (Evidence, reproducibility and clarity (Required)):

In this manuscript, Darmasaputra and colleagues took advantage of human hepatocyte organoids (Hep-Org) to investigate the formation of binucleated cells that naturally occurs in liver. So far, the mechanism of hepatocyte binucleation has been studied in rodents, where binucleated hepatocytes arise upon weaning through an insulin/akt pathway that inhibits furrow contraction in a fraction of cells (Ref. 21, 22). In addition, it is known that E2F7 and E2F8 downstream of the Wnt signaling repress the expression in mouse hepatocytes of several key cytokinetic proteins (AuroraB, Mklp1, Ect2, Racgap1) and thereby promote binucleation (Ref. 23).

Advances:

As seen in vivo, the authors first show that a fraction (5-15%) of cells are binucleated in two independently derived human Hep-Orgs. Live cell imaging reveals that binucleation is not due to furrow ingression defects after anaphase but rather arises from post-furrowing intercellular bridge regression. Fixed data suggest that the cytokinetic midbody formed normally but lost its anchorage to the bridge membrane. Activation of the Wnt signaling resulted in a modest but significant increase in the proportion of binucleated cells (4.5 to Major comments

1. An outstanding question is whether human Hep-Orgs represent a bona-fide model to study the process of human liver binucleation. The absence of cholangiocytes, vascularization, other cell types and physiological hormones etc. might impact on the mechanism of binucleation, which is the main focus of this study. Since the mechanism of binucleation in human Hep-Orgs appears radically different from what has been reported in vivo in rodents, the authors should reproduce the lack of furrow ingression in mouse Hep-Orgs (that they were able to generate in Ref. 44). This could be done in fixed cells as in Fig. 3. Alternatively, they could use live cell imaging and chemical dyes such as SiR-Tubulin and Cell Mask to label microtubules and the plasma membrane, respectively, without the need of creating genome-edited reporter lines.

The mechanism of endomitosis that we observe in human hepatocyte organoids is indeed different from what has been observed in mouse hepatocytes. Unfortunately, mouse Hep-Orgs are more difficult to generate as they require a two-step perfusion protocol from live mice (described in Hu et al., 2018). Additionally, mouse Hep-Orgs do not survive freezing, so to be able to perform the suggested experiments, we would need to generate new mouse Hep-Org lines. As our collaborators are currently not performing any experiments with mouse livers, we would need to request an ethical permit to generate these organoids, which would take several months. We have seriously considered this option, however due to the substantial investment in time and resources, we feel these experiments would be more suited for a follow-up study.

To nonetheless better clarify the differences between what has been observed in rodents, and what we see in the Hep-Orgs, we have added a paragraph in the discussion, see pages 14-15 lines 433-460.

The videos acquired in Fig. 2 contain much more information than presented. The authors should measure the rate of furrow ingression, the extend of spindle elongation, the time of MT severing and the time of furrow/bridge regression after cytokinesis onset. All these parameters are important since spindle elongation and furrow ingression are altered in rodents. Is this also the case in human Hep-orgs? Furthermore, the spindle seems very different (bent bridges) in endomitotic compared to canonical cytokinesis (Fig. 2A). Finally, the authors should provide more time points during the time of furrow regression to better show how this phenomenon occurs. It seems, based on fixed images, that the midbody stays attached to the plasma yhmembrane in an asymmetric manner (i.e. does not fully detach, contrary to what is stated in the text). 3D reconstructions in fixed cells and a further characterization of the movies would clarify this point.

We thank the reviewer for this suggestion. Although there are some technical limitations that pose some restrictions (explained below), we have extended our analyses where possible. In our live imaging, we use 5-minute time intervals with 4 mm z-slices, which allows a delicate balance between having enough frames in M phase, and imaging for at least 48 hours, which is required to catch enough divisions. We are unable to image with smaller time intervals or smaller z-slices, as this leads to phototoxicity. Nonetheless, using these settings, we can get an indication of the rate of furrow ingression, time of severing and the time of furrow regression:

• We find that the time of furrowing onset and the rate of furrow ingression is very similar between canonical M phases and endomitosis M phases: we have now added this data in the results section, page 7, lines 192-199 and Fig. 2D.
• The time of cytokinetic regression is more variable between endomitoses events, and can range between 30 minutes and 2,5 hours. We have also added this information to page 7, lines 199-202and Fig. 2E
• The time of MT severing is similar between endomitosis M phases, as we show in Fig. 2C
• Unfortunately, we cannot accurately measure the extent of spindle elongation, as the divisions occur in 3D and our Z resolution is not good enough. Regarding the observation that the spindle looks different in the endomitosis example in Fig. 2A: we have quantified how often we observe bent midzones in endomitosis versus canonical M phases, and this occurs in 60% of canonical (n=12/20) and 83% of endomitosis M phases (n=15/18). We have now added this information in the results section, page 6, lines 1862-185.
• We have quantified how often we see the midbody remaining attached to one side of the plasma membrane versus fully detaching: we find that in 6 out of 9 late stage endomitotic regressions, the membrane is detached from both sides, and in 3 out of 9, it remains attached to one side. We have added this information to the results section, page 8 line 249-251.

  DAPI staining is not sensitive enough to detect thin chromatin bridges. To rule out that post-furrowing regression is not merely due to the present of DNA bridges, the authors should confirm their results with LAP2b staining (see PMID 19203582).

To exclude the presence of ultrafine DNA bridges during anaphase, we have performed a staining for RIF1, a factor that localizes to ultrafine DNA bridges in anaphase and is required for their resolution (Hengeveld et al, 2015, PMID: 26256213). In early anaphase, we find many RIF1-positive thread-like structures, as has been described before in other non-transformed and non-stressed cells. However, in late anaphase and telophase, we never observe these fibers (n=57/57), suggesting that they are fully resolved and are not the cause of cytokinetic regression. We have added this data to the results section, see page 8, lines 226-234, and Fig. S1.

The authors shows that binucleation results from defective anchorage of the bridge membrane to the midbody, but the molecular mechanism remains elusive and should be further probed. In Fig. 3, there is no obvious changes in the investigated markers. Are the intensities of RACGAP1, Anillin, CIT-K reduced in regressing cells? Are ECT2, activated (phospho) Myosin II, CEP55/ESCRT-III, (activated) AuroraB and MKLP1 normally localized/concentrated? ECT2, AuroraB and MKLP1 are regulated by E2F7/8 (Ref. 23) and AuroraB inactivation after bridge formation leads to late regression (PMID 19203582).

We agree with the reviewer that the molecular mechanism by which midbodies lose their attachment to the membrane is currently unclear. We do not see any clear differences in the intensities of RACGAP1, Anillin, or CIT-K in cells undergoing endomitotic regression. We also do not expect large differences in localization or abundancies of ECT2, AuroraB or MKLP1, because if this were the case, you would expect differences in early cytokinesis in endomitosis, such as a delay or a slower rate of furrow ingression. We did perform additional IF experiments to investigate the localization of SEPT9, a septin that is expressed in human hepatocytes and that has essential functions in membrane anchorage during cytokinesis. Although we find that SEPT9 exhibits more variable localizations than RACGAP1, Anillin, and CIT-K, we find that in the majority of endomitotic regressions, it is also absent from the regressed membrane (n=5/7 cells). We have added this data to the results section on page 9, and in the figures Fig. 3C and Fig.4C.

The results of Fig. 4F indicate that the increased proportion of binucleated cells upon CHIR99021 removal depends on E2F7/8. Without live cell imaging (or FISH experiments) the authors cannot conclude that conclude that the increase in endomitosis is dependent on E2F7/8. A decrease in binucleation could indeed not imply a reduced occurrence of endomitosis. For instance, it is possible that E2F7/8 KO induces the formation of mononucleated 4n cells due to early mitotic failure. This issue should be clarified.

The reviewer raises an important point. Unfortunately, we were unable to generate E2F7/8 KO lines containing fluorescent nuclear and membrane markers, which would allow us to perform live-imaging and confirm that these organoids perform less endomitosis. As an alternative, we tried to use SiR-Tubulin dyes for live imaging, but even at very low concentrations these dyes are toxic for the organoids. To exclude the possibility that E2F7/8 KO induces the formation of mononucleated 4n cells, we have measured the DNA content in wildtype, E2F7 and E2F8 lines, and found that the distribution of ploidies is very similar between these lines, both in normal growth conditions as well as upon removal of CHIR99021 (see the new supplemental figure, Fig. S3). We thus think it is unlikely that E2F7/8 KO induces the formation of mononucleated 4n, however it remains possible that the differences in percentage of binucleated cells arise independently of endomitosis. We have now toned down our conclusions on the function of WNT signaling and E2F7/8 in endomitosis, and discussed alternative explanations for our findings in the discussion, see page 14, lines 451-460.

Binucleation increases with age both in humans and rodents. Could this feature be mimicked in the human Hep-Org by leaving the organoids longer in culture? (optional but would reinforce the value of the model).

We do not see an increase in binucleation percentages in organoids that are kept longer in cultures, and we have now also tested the effect of growing the organoids in a “differentiation medium”, which was previously described to give rise to more mature hepatocyte gene expression (Hu et al. 2018), however we see no significant differences in the percentages of binucleated cells per organoid. We have now included this data, as well as our analyses of the effect of insulin in the growth medium (see our response to point 12 below) in the results section on page 11 lines 341-353 and we further discuss this point in the Discussion, pages 12-13, lines 389-397.

Minor comments

The results of Table 1 are based on very few fixed cells (3 to 6). The authors should consider increasing the number of regressing cells.

We are aware that the number of endomitotic regressions is very low. Unfortunately, it is extremely challenging to catch these events by IF: cells in Hep-Orgs cycle very slowly (they divide once every ±50 hours), and thus very few cells are in M phase at any given moment (only 1 or 2 cells per organoid) – the chance that this cell is then also in telophase is even lower, and then only ± 5% of the telophase are actually undergoing endomitosis. Due to technical limitations of the organoid IF staining protocol, it is not trivial to scale up these experiments, making it very difficult to find more than 3-5 endomitotic regressions per condition. Despite the low numbers of endomitotic regressions that we have identified, we find that RacGAP1, Anillin and CIT-K localize in a very similar manner in cells undergoing endomitosis. For SEPT9, we see a little bit more variation in the localizations, but also here the majority of cells in undergoing endomitosis have lost SEPT9 membrane association on the regressed membrane (see Fig. 4C).

Is WNT signaling modified by E2F7/8 mutations? To conclude that "WNT signaling inhibits binucleation in an E2F7/8-dependent manner", the authors should check that E2F7/8 KO does not impair the increase of WNT signaling upon CHIR99021 removal.

We had not thought of this option, but it is indeed possible that E2F7/8 influences the ability of cells to respond to CHIR99021 removal. WNT regulators are not known to be targets of E2F7 or E2F8 in mice (see PMIDs: 22180533, 18194653, and 23064264), however as we have not analyzed the gene expression changes in E2F7 or E2F8 mutant organoids, we cannot exclude the possibility that CHIR99021 has different effects in E2F7/E2F8 knock-out cells. We now discuss this possibility in the discussion, page 15, lines 459-460.

Please provide movies of the cells presented in Fig. 2A.

We have included movies of these cells, see Supplemental Movie 1 and Supplemental Movie 2.

1. Removal of CHIR99021 induces major shape changes and lumen formation (rather than "exhibited some morphological changes" as stated). Could the author speculate on this?

WNT signaling is likely important for many aspects of hepatocyte growth and differentiation, and it is possible that upon CHIR99021 removal, Hep-Orgs are starting to differentiate and become more secretory, which would explain why they start forming larger lumens. We now discuss this in more detail in the final part of the results section, see page 11 lines 341-353, and in the discussion, page 15 lines 462-472.

1. Fig. 4: Why do the authors use the cell line-1 that has the lowest level of binucleation in this experiment? Would the results be the same in cell line 2? (optional)

We perform most experiments in Hep-Org line 1 because this line is easier to maintain in culture, and we have been unable to generate CRISPR knock-outs in Hep-Org line 2.

1. Would insulin increase the proportion of binucleated cells, as in rodents? (optional)

We have tested this, but do not see a difference in the percentage of binucleated cells when we either increase or decrease the concentration of insulin in the growth medium. We have now added this data in the results section, see page 11, lines 347-350 and Fig. 5J.

Reviewer #3 (Significance (Required)):

Strengths and limitations:

The manuscript is well written, easy to follow, and the quality of the data is overall high. A clear strength of this study is the use of state-of-the-art human hepatocyte organoids and genome editing (to generate reporter lines and to KO E2F7/8). This allows the authors to address the mechanism of binucleation in a human context. Interestingly, it revealed both similarities (e.g. E2F7/8 depends for binucleation) and striking mechanistic differences (e.g. post-furrowing regression) between rodent and human systems. The study is rather descriptive -which is fine- but deeper mechanistic insights would strengthen the conclusions of the manuscript. For instance, "our results identify how human hepatocytes inhibit cell division in endomitosis" appears as an overstatement since the molecular reason of midbody anchorage defects remains elusive.

We thank the reviewer for their kind words. Unfortunately, we have been unable to gain deeper mechanistic insights into the molecular mechanism of membrane regression in endomitosis. We have therefore toned down our conclusions, see the new concluding sentence in the abstract, page 2, lines 35-36.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #3

Evidence, reproducibility and clarity

In this manuscript, Darmasaputra and colleagues took advantage of human hepatocyte organoids (Hep-Org) to investigate the formation of binucleated cells that naturally occurs in liver. So far, the mechanism of hepatocyte binucleation has been studied in rodents, where binucleated hepatocytes arise upon weaning through an insulin/akt pathway that inhibits furrow contraction in a fraction of cells (Ref. 21, 22). In addition, it is known that E2F7 and E2F8 downstream of the Wnt signaling repress the expression in mouse hepatocytes of several key cytokinetic proteins (AuroraB, Mklp1, Ect2, Racgap1) and thereby promote binucleation (Ref. 23).

Advances:

As seen in vivo, the authors first show that a fraction (5-15%) of cells are binucleated in two independently derived human Hep-Orgs. Live cell imaging reveals that binucleation is not due to furrow ingression defects after anaphase but rather arises from post-furrowing intercellular bridge regression. Fixed data suggest that the cytokinetic midbody formed normally but lost its anchorage to the bridge membrane. Activation of the Wnt signaling resulted in a modest but significant increase in the proportion of binucleated cells (4.5 to <8% or 3 to <6% depending on the subfigures). This increase depended on the presence of E2F7 and E2F8. This study represents the first description of binucleation in a human organoid context.

Major comments

1. An outstanding question is whether human Hep-Orgs represent a bona-fide model to study the process of human liver binucleation. The absence of cholangiocytes, vascularization, other cell types and physiological hormones etc. might impact on the mechanism of binucleation, which is the main focus of this study. Since the mechanism of binucleation in human Hep-Orgs appears radically different from what has been reported in vivo in rodents, the authors should reproduce the lack of furrow ingression in mouse Hep-Orgs (that they were able to generate in Ref. 44). This could be done in fixed cells as in Fig. 3. Alternatively, they could use live cell imaging and chemical dyes such as SiR-Tubulin and Cell Mask to label microtubules and the plasma membrane, respectively, without the need of creating genome-edited reporter lines.
2. The videos acquired in Fig. 2 contain much more information than presented. The authors should measure the rate of furrow ingression, the extend of spindle elongation, the time of MT severing and the time of furrow/bridge regression after cytokinesis onset. All these parameters are important since spindle elongation and furrow ingression are altered in rodents. Is this also the case in human Hep-orgs? Furthermore, the spindle seems very different (bent bridges) in endomitotic compared to canonical cytokinesis (Fig. 2A). Finally, the authors should provide more time points during the time of furrow regression to better show how this phenomenon occurs. It seems, based on fixed images, that the midbody stays attached to the plasma membrane in an asymmetric manner (i.e. does not fully detach, contrary to what is stated in the text). 3D reconstructions in fixed cells and a further characterization of the movies would clarify this point.
3. DAPI staining is not sensitive enough to detect thin chromatin bridges. To rule out that post-furrowing regression is not merely due to the present of DNA bridges, the authors should confirm their results with LAP2b staining (see PMID 19203582).
4. The authors shows that binucleation results from defective anchorage of the bridge membrane to the midbody, but the molecular mechanism remains elusive and should be further probed. In Fig. 3, there is no obvious changes in the investigated markers. Are the intensities of RACGAP1, Anillin, CIT-K reduced in regressing cells? Are ECT2, activated (phospho) Myosin II, CEP55/ESCRT-III, (activated) AuroraB and MKLP1 normally localized/concentrated? ECT2, AuroraB and MKLP1 are regulated by E2F7/8 (Ref. 23) and AuroraB inactivation after bridge formation leads to late regression (PMID 19203582).
5. The results of Fig. 4F indicate that the increased proportion of binucleated cells upon CHIR99021 removal depends on E2F7/8. Without live cell imaging (or FISH experiments) the authors cannot conclude that conclude that the increase in endomitosis is dependent on E2F7/8. A decrease in binucleation could indeed not imply a reduced occurrence of endomitosis. For instance, it is possible that E2F7/8 KO induces the formation of mononucleated 4n cells due to early mitotic failure. This issue should be clarified.
6. Binucleation increases with age both in humans and rodents. Could this feature be mimicked in the human Hep-Org by leaving the organoids longer in culture? (optional but would reinforce the value of the model).

Minor comments

1. The results of Table 1 are based on very few fixed cells (3 to 6). The authors should consider increasing the number of regressing cells.
2. Is WNT signaling modified by E2F7/8 mutations? To conclude that "WNT signaling inhibits binucleation in an E2F7/8-dependent manner", the authors should check that E2F7/8 KO does not impair the increase of WNT signaling upon CHIR99021 removal.
3. Please provide movies of the cells presented in Fig. 2A.
4. Removal of CHIR99021 induces major shape changes and lumen formation (rather than "exhibited some morphological changes" as stated). Could the author speculate on this?
5. Fig. 4: Why do the authors use the cell line-1 that has the lowest level of binucleation in this experiment? Would the results be the same in cell line 2? (optional)
6. Would insulin increase the proportion of binucleated cells, as in rodents? (optional)

Significance

Strengths and limitations:

The manuscript is well written, easy to follow, and the quality of the data is overall high. A clear strength of this study is the use of state-of-the-art human hepatocyte organoids and genome editing (to generate reporter lines and to KO E2F7/8). This allows the authors to address the mechanism of binucleation in a human context. Interestingly, it revealed both similarities (e.g. E2F7/8 depends for binucleation) and striking mechanistic differences (e.g. post-furrowing regression) between rodent and human systems. The study is rather descriptive -which is fine- but deeper mechanistic insights would strengthen the conclusions of the manuscript. For instance, "our results identify how human hepatocytes inhibit cell division in endomitosis" appears as an overstatement since the molecular reason of midbody anchorage defects remains elusive.

Audience:

broad, basic research.

Field of expertise:

cell biology of cytokinesis

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

Summary:

Polyploid cells arise within various human tissues by multiple different mechanisms. Here, Darmasaputra et al present a study of one such mechanism, endomitosis, in liver cells using fetal-derived human hepatocyte organoids. In this model, they demonstrate that binucleated cells arise through the late regression of the cytokinetic furrow prior to abscission. They identify a rare event in cytokinetic cells - loss of midbody association with the plasma membrane - that could explain the cytokinesis failure observed in a proportion of these cells. Finally, they show that loss of Wnt signalling increases the number of binucleation events in a manner that depends on E2F7 and E2F8, similar to what has been observed in murine hepatocytes.

Major comments:

This is a compelling and well-presented study. The data presented are high quality, the experiments are well described and controlled and the conclusions are convincing. I am particularly impressed by the technical effort that the authors must have put into obtaining high quality live and IF images of dividing cells within organoids and their careful documentation of what are very rare mitotic events. In addition, the manuscript is extremely well written and I found it a pleasure to read.

I do not think that there are additional experiments that are essential to justify the conclusions of the paper. However, I do have suggestions that I think would strengthen this work and increase its significance. As is, the authors present findings in two different areas: the documentation of cytokinesis failure in hepatocyte organoids and the role of Wnt and E2F7/8 on binucleation. It would be really nice if the two parts could be linked. For example, the authors could examine cell divisions in the organoids without Wnt either live or fixed and show that they have a higher proportion of cells undergoing cytokinetic regression or with membrane-midbody attachment defects. Alternatively, they could look at whether the expression levels of key cytokinetic genes are changed in the Wnt and E2F7/8 organoids. As I said, these experiments are not required for or the publication of this work and I will leave it up to the authors to decide if they have the time or capacity to add additional data.

Finally, before publication, the authors should discuss further the mechanisms by which loss of membrane attachment during cytokinesis could occur - there is quite a lot of literature in this area on the role of RacGAP1 and Ect2 in membrane attachment that is not discussed, particularly from the lab of Mark Pentronczki (eg Kotynkova 2026 PMID: 27926870, Lekmotsev PMID: 23235882). It's surprising that the authors haven't mentioned (or looked at) Ect2 at all, especially since Ect2 levels have been shown to control polyploidy in cardiomyocytes (Liu 2019 PMID: 31597755). This at least warrants some discussion.

Minor comments:

• Table 1 would be more striking as a graphical representation. I appreciate that the n numbers in the regressed cells means that statistical comparisons is not possible, but some kind of colour coding or graph would make this part clearer
• It's not clear what the difference between Hep-Org 1 and Hep-Org 2 are. Are these from different donors?

Significance

This study is an important technical development in that it reports a new system to study in depth cell biology of liver endomitosis in non-transformed and, crucially, human 3D hepatocyte organoids. The findings reported using this system are potentially interesting although they could be further developed if they were mechanistically linked together (see major comments). This work is likely to be highly interesting to scientists studying cell division, cytokinesis and hepatocyte biology. It also has wider implications for liver biology and particularly liver regeneration. Additionally, given the role of polyploidisation in many different tissues, it will likely be of interest to scientists studying polyploidy and endomitosis more generally.

My area of expertise is in cytokinesis and cell division in general, although not specifically in hepatocytes. I am not an expert in organoids.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

The study provides insights into how polyploidization via endomitosis may arise in human hepatocytes by studying fetal liver cell line-derived organoids. Using live cell imaging and LSM microscopy, binculeation was consistently observed in two independent cell line systems, at frequencies seen in human liver and sensitive to pharmacological inhibition (GSK3i) and genetic manipulation (E2F7 & E2F8 editing). The findings presented are in line with earlier data, largely gathered studying rodents. The data is convincing and robust indicating that these systems can be used to study cause and consequences of polyploidy in human hepatocytes.

While the authors do suggest that they provide a mechanisms how polyploidy is initiated in human hepatocytes undergoing endomitosis, ie. loss of membrane association of membrane-anchoring proteins at the midbody (e.g. Anillin, RacGAP1), I do feel that the data provided is rather descriptive and does not address a particular mechanism that may account for loss of membrane anchoring. As such, the title is making a too strong point, as, in my point of view, it associates with loss of membrane anchorage, but may not drive endomitosis. Whether this is a "passive" process in response to changes in physical forces and tension, or regulated via signalling intermediates to initiate regression of the cleavage furrow is not addressed experimentally (mislocalizing these proteins on a larger scale). Discussion seems warranted.

I do not see the need for additional experiments, as I believe the data is robust and introduces an interesting new model where the role of ploidy can be studied in human hepatocytes ex vivo. However, if the authors wish to extend their studies and document further similarities with pathways engaged in rodents, some E2F7/8 targets relevant for ploidy control such as Anillin or PIDDosome components, or, maybe MDM2 processing for p53 activation, could be tested in wt and E2F mutant cell lines.

A minor suggestion is to clarify the term M-CDK activity in the introduction, as it may not be fully intuitive to all readers; similarly, ploidy reversal is still controversial in the field, but it is stated as a given fact.

Significance

Polyploidy at the cellular and nuclear level is a key feature of hepatocytes albeit the physiological significance of the process is not entirely clear. Increased ploidy has been linked to cancer resistance in the liver, but may pose a threat to hepatocyte survival under conditions of repeated compensatory proliferation cycles. Curiously, during normal regeneration after single surgical intervention liver regeneration is not compromised, even though it may recover faster starting when starting from higher ploidy levels. Mechanistically, most data has been generates studying rodents where it is documented that the proliferation behaviour changes around the time of weaning in mice when hepatocytes start to fail cytokinesis and undergo endomitosis, leading to cellular and nuclear polyploidy. In rodents, insulin signalling / AKT appears involved as is the E2F network and p53, activated by the caspase-2-PIDDosome. The model system introduced here will allow mechanistic studies in human organoids and help to increase our understanding of this process in steady state and under conditions of stress.

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

*Reviewer #1 (Evidence, reproducibility and clarity (Required)):

Deletion of CerS4 in the entire mouse epidermis throughout development via the K14-Cre results in enlarged sebaceous glands and perturbed HFSC molecular phenotype. There is low or no expression of CD34, a known marker of the HFSCs along with apparent reduction of several other HFSC markers and acquisition of a more differentiated cell phenotype in these cells. Interestingly, skin and hair follicles seem to remain normal otherwise up to advanced age, though this contradicts the notion that HFSC were indeed affected at the functional level. The data does not demonstrate 'gradual decline' in the HFSC compartment, as claimed by the authors, but rather seem to indicate that the adult HFSC compartment is not properly established in its molecular signatures. Organoid cultures document defects in HFSC, which included reduced proliferation in the CerS4 KO cells. Lipid composition in plasma membranes was also affected by the CerS4 KO. Associated with this, Wnt signal transduction is also affected according to experiments that enhance the strength of wnt signals via a specific small molecular agonist of the pathway. Finally, the authors discover a resemblance of the mouse KO immune-phenotype, with human atopic dermatitis. The study is likely of interest to a specialized readership in skin biology and dermatology and adds to previous studies on CerS4 in skin that erroneously placed its role in the sebaceous gland. [The authors here demonstrate that deletion of CerS4 in the sebaceous glands via SCD3-Cre led to no phenotype, contradicting the previous assessment that CerS4 is important in sebaceous glands.] The study would need to be corrected in a few of its interpretations regarding stem cells to better match the data, as indicated below. *

We thank the reviewer for the constructive comments that will help us to improve the manuscript. In particular, it is clear that we have not been sufficiently clear in the data presentation.

Firstly, contrary to what the reviewer states, the CerS4epi-/- mice have a very strong hair follicle phenotype that results in complete hair loss. Also the epidermis is not normal as an inflammatory phenotype develops later, after the hair follicle architecture and function has been disrupted. Thus, there are clear functional consequences to the hair follicle and epidermis that arise from the dysfunction of the HFSC compartment. We will edit the manuscript and add photodocumentation of the macroscopic phenotype to ensure clarity.

We fully agree with the reviewer that the initial phenotype is inability to establish the adult hair follicle stem cell niche, as shown by the single cell sequencing data and as also stated in the manuscript title. We will further edit the manuscript to clarify this conclusion. Importantly, however, some hair follicle stem cells are generated but these become gradually depleted. So there is a dual phenotype: an inability to efficiently establish and maintain the hair follicle stem cell population. We will clarify this in the text.

Finally, we want to emphasize that the main finding of this manuscript is that hair follicle stem cells contain a unique lipid profile and perturbing this profile by deleting CerS4 leads to profound defects in stem cell fate regulation through Wnt. This is a completely new finding that has implications far beyond dermatology.

Major revisions: Fig1B - the data seems to simply shows that bulge cells express less or no CD34 and not that ' CerS4epi-/- mice showed reduced HFSC numbers'; the primary FACS data should be shown somewhere too.

Outer bulge hair follicle stem cells are defined as a population of cells that expresses CD34 and integrin-a6. The quantifications in Fig 1B show the quantitative FACS analyses of the size of this population and indicate less CD34+/integrin-a6+ cells in CerS4epi-/- epidermis. The mean fluorescence intensity of CD34 and integrin-a6 was not reduced in these CerS4epi-/- stem cells. This FACS analysis therefore allows the conclusion that there are less CD34+/integrin-a6+ cells in CerS4epi-/- epidermis. We will include the original FACS plot data to support this notion and the quantifications.

The conclusion that stemness is affected, and HFSCs lose their normal gene expression signature is at more convincing after looking at other HFSC markers down the road in the paper. However, in the absence of functional assays that would demonstrate stem cell function is lacking and seeing that hair follicles are maintained and grow in long-term, the notion that stem cells are lacking in these conditions is not supported by the data.

We appreciate that the reviewer finds the marker gene analysis convincing. To assay stem cell functionality, we have used the organoid assays (spheroid formation is classical, widely used assay for stemness). Using these functional assays we observe impaired self-renewal of stem cells (Fig. 3D), enhanced differentiation (Fig. 3I), and altered Wnt responsiveness (Fig. 5 E, F), all indicative of stem cell dysfunction and explaining the in vivo phenotypes of altered stem cell differentiation and inability to establish and maintain the stem cell population.

In the revised manuscript we will also include measurements of stem cell self-renewal in vivo using BrdU incorporation and provide more detailed description on the hair loss phenotype of the mice to further strengthen this conclusion.

*The conclusion after figure 2: "Collectively, these data indicate that CerS4-deficiency triggers ... ... gradual depletion of the quiescent HFSC compartment." There is no data showing gradual depletion of the quiescent HFSC compartment. We would need to see a gradual activation of HFSCs with over proliferation to conclude this. There is some data albeit not always convincing (see NFAC1 staining in Fig. 5C) indicating loss of markers associated with quiescence but there is no data indicating 'gradual' loss of markers. *

We agree with the reviewer that showing gradual activation of the HFSCs in vivo is important to conclude loss of quiescence. We will include in situ stainings of in vivo BrdU labeling and quantify proliferation in the hair follicle bulge stem cell region. Preliminary data of P47 mice already shows a clear increase in BrdU+ cell in the stem cell compartment in CerS4epi-/- skin . Further analysis at P21 will be carried out during the revision.

*Minor revisions: *

• Figure 1C legend - please spell out what are the abbreviations for the different subpopulations; please show these populations as % as opposed to absolute numbers. *

We will edit the Figure 1C legend for clarity and express the populations as %.

*

*

*Figure 1D - please make it clear in the cartoon what the different sub-populations listed are; *

We will edit the cartoon for clarity.

Is OB1 a CD34- HFSC population?

The outer bulge 1 (OB1) is a population of cells that expresses hair follicle stem cell markers, including CD34. We will clarify this in the legend.

Fig. 3B - the colors of the legend do not match the colors in the data so it is confusing as to which one is which!

We will alter the colors to match the data and thank the reviewer for pointing this out.

Fig 5C - the differences in NFATc1 are not visible in the images shown

We apologize for the suboptimal quality of these images and will replace them with higher resolution images to more clearly demonstrate the difference.

*

Reviewer #2 (Evidence, reproducibility and clarity (Required)):

The manuscript authored by Peters et al. titled "Sphingolipid metabolism orchestrates the establishment of the adult hair follicle stem cell niche to control skin homeostasis" elucidates the critical role of ceramide synthase 4 (CerS4) in the epidermal stem cell niche, particularly in regulating hair follicle bulge stem cells (HFSCs). Using epidermal specific CerS4 knockout mice as an in vivo model and hair follicle organoid culture as an ex vivo model, the authors conducted a comprehensive analysis, which includes cutting edge approaches such as scRNA-seq, proteomics, and lipidomics. The results highlight CerS4's function in the establishment/maintenance of the HFSC niche, as absence of CerS4 changes HFSCs' number and differentiation state. Potential underlying mechanisms identified include altered membrane lipid profiles and Wnt signaling responsiveness. Possible link to a chronic inflammatory skin disease, atopic dermatitis, is also implicated. The data presented are generally of high quality, and the work is significant as it uncovers a new regulator of HFSC fate with mechanistic connection to lipid metabolism. *

We thank the reviewer for the positive assessment of our work and finding it to be of high quality and significance. We further appreciate the constructive comments that will further help us to improve the manuscript.

*However, some issues were identified, most of them having to do with in vivo characterization and data interpretation:

*

*Major: 1. The in vivo HFSC phenotype can be better characterized. "Collective, these data show that CerS4 in HFSCs is essential to establish the adult stem cell compartment and to assure lineage fidelity." - this statement premature based on order of the data shown. Also the trajectory difference shown in Figure 2A is not striking. Subclustering out the relative cell subsets and redo the analysis might help to tease out the difference. Additional experiments such as lineage tracing would be useful to support the notion that there is lineage fidelity issue in the mutant - though it is understood that this is quite involved and may lie outside the scope of the current study. Are bulge cells in the mutant proliferative? - the authors should consider in vivo Edu labelling experiment or the like to assess the quiescence/proliferation of the bulge cells. Finally, analyzing hair follicles at earlier stages might help to clarify when and where the bulge and sebaceous gland changes start - is possible that aberrant divergence of bulge/sebaceous fates occur prior to the establishment of a stable bulge fate? *

We thank the reviewer for suggesting additional analyses of the single cell sequencing data. We have performed subclustering of the relevant populations for the trajectory analyses to more clearly demonstrate the altered lineage trajectories. We will include these new analyses in the manuscript. Importantly, the in vitro organoids show abnormal differentiation (Fig. 3H, 3I, Supplementary Fig. 2G), closely resembling the in vivo phenotypes, thereby strengthening the conclusion of cell-autonomously altered lineage trajectories of the hair follicle stem cells.

We will further preform in vivo BrdU labeling as suggested. These experiments have already been initiated and preliminary data show increased proliferation in the bulge stem cell region of CerS4epi-/- mice in older mice (P47). The data will be included to emphasize long term loss of quiescence in the stem cell compartment in CerS4epi-/- mice.

Understanding the early development of bulge and sebaceous fates is indeed an interesting question. This will be addressed by detailed analyses of stem cell fate at early stages (P17-P21) using key markers of stem cell state and sebaceous linages (CD34, Krt15, Lhx2, Nfatc1, SCD1 and FASN).

Finally, we have initiated lineage tracing experiments using the stem cell-specific Lgr5-Cre to conclusively demonstrate that Cers4-deletion leads to altered routing of hair follicle stem cells into upper hair follicle and sebaceous gland fates. This notion is supported by the preliminary analyses of these experiments. We will finalize these analyses and include them in the manuscript.

*2. The exclusion of IFE contribution is not backed up by data. Figure 6D - model emphasizes HFSC involvement in atopic dermatitis, but this could be due to epidermal barrier defect. Barrier defect could already be present even though IFE morphology appears normal. Maybe TEWL is measured at the time of analysis and shows no change - if so, this data should be included. HFSC changes might contribute but the involvement of IFE cannot be excluded. The conclusion that "CerS4 expression was restricted to the hair follicle" is not supported by data. IFE expression is apparent in Figure S1C. Along this line, there is also an apparent expansion of IFE basal II in the mutant (Figure 1C). *

We acknowledge that we have not been clear enough with the evidence that allowed us to exclude the involvement of an IFE-mediated barrier defect in the early skin inflammation phenotype. To address a potential barrier defect early on, we have performed careful analysis of TEWL. In Peters at al., 2020 we demonstrate no changes in TEWL at P0, a reduced TEWL at P21 and an increased TEWL in adult CerS4epi-/- mice starting only at P33. The reduction of the TEWL in adolescent CerS4epi-/- mice (P21) is likely linked to an increased production of sebaceous lipids lubricating the skin surface at this time point (Peters et al., 2020). Thus, defects in the hair follicle stem cell compartment, present at adolescence (P21) arise prior to defects in the adult (P33) IFE barrier function. We will clarify this in the manuscript.

Cers4 expression is overall low in skin, as is typical for enzymes. In situ stainings of Cers4mRNA (Fig.S1C) indeed show a sparse signal also in the IFE. This signal is also detected in CerS4-/- sections, although the KO skin cannot be conclusively used to control background as these mice were generated by deletion of exon 3 only, and Cers4 RNAscope probes might detect remnant Cers4 RNA in these mice. Importantly, our data on FACS sorted basal cells of the IFE shows no substantial Cers4 mRNA expression in IFE progenitors (Fig. S1D) and no mRNA is detected in the IFE in the single cell sequencing. Thus, while we cannot fully exclude low levels of CerS4 expression in the IFE, the levels are substantially lower than in the HFSC and SG compartments, and the phenotype, including the slight expansion of the IFE basal II population, is very minor compared to the hair follicle phenotype. However, to avoid overinterpreting our data, we will carefully edit the conclusions to be less strong on the involvement of the IFE. Furthermore, we will perform hair follicle stem cell lineage tracing experiments as outlined in resspose to the previous point to strengthen the conclusion on the hair follicle stem cell-autonomous phenotype.

3. Figure 1 - single cell analysis was done using only 2 pairs of mice, and data in E lack statistical assessment. At the very least, data for individual pairs should be shown in supplemental data to ensure that changes are consistent in both mutant mice rather than being dominated by dramatic alteration in only 1 mutant mouse.

We naturally have rigorously analyzed the replicates to ensure that the phenotype is consistently present in both. We will include the separate analysis of the mice to document this and include statistical analysis.

Minor: 1. CerS4SCD3-/+ nomenclature is mis-leading.

We will edit this for clarity

• 1. Figure 2- "Furthermore, we observed expansion of the inner bulge identity marker Krt6 protein expression into outer bulge stem cells and along the infundibulum in CerS4epi-/- hair follicles, whereas in control mice Krt6 was restricted to the inner bulge (Fig. 2C)." - Krt6 staining is presented in Fig 2D, not 2C. *

We thank the reviewer for pointing out this mistake that will correct.

• 1. Figure 3C - size of the organoids should be quantified with statistics. The images shown do not support the statement that "Strikingly, CerS4epi-/- organoids showed altered morphology characterized by smaller size and loss of cohesion of peripheral cells from the organoid clusters (Fig. 3C), ...".*

We will include quantifications.

• 1. Section titled "CerS4 regulates HFSC differentiation in a stem cell autonomous manner": "CD34- integrin- a6+ cells, which based on extensive transcriptome and marker expression analyses represent a mixture of HFSCs, hair follicle outer root sheath (ORS) cells and inner bulge cells (collectively termed non-HFSCs)." - shouldn't the CD34- integrin- a6+ population also contain IFE stem/progenitor cells? Are hair follicles micro-dissected out for FACS? *

The hair follicles are not micro-dissected out for FACS, and the entire basal cell population is initially isolated. However the organoid culture conditions speficically promote the expansion of the hair follicle linage, whereas cells of the IFE are not expanded and long term maintained as extensively documented in previous publications using this organoid system (see for example Kim et al., Cell Metabolism 2020; Chacon-Martinez EMBOJ 2016).

*5. Figure 5D - please provide the working concentration of Chir99021. *

We will provide the working concentration.

*6. Figure 5F - explain what arrows mean in legends. *

We will define the arrows.

• 1. Figure 6A - no significant changes in Th2 and ILC2 were observed at a 95% confidence interval. Increasing mouse number will help to increase statistical power*.

We agree with the reviewer and acknowledge that this experiment was unfortunately underpowered. We will repeat it with a larger cohort.

• 1. Additional Wnt target genes such as Axin 2 should be looked at.*

We thank the reviewer for this suggestion, we will include analyses of additional Wnt target genes.

• 1. The increased BMP signaling and decreased Nfatc1 expression are seemingly contradictory.*

We apologize for the lack of clarity here. Single cell sequencing showed increased BMP-signaling of outer bulge cell cells to inner bulge cells on mRNA level (Figure S4A). No alteration in BMP signaling was detected within the outer bulge stem cell compartment (Figure 5A). Nfatc1 protein expression was analyzed in the upper bugle (Figure 5C). The data indicate no differential gene expression of ligand receptor pairs mediating BMP-signaling within the outer bulge. A decrease in Nfact1 protein expression (Fig. 5C) together with an increased proliferation (see above) and loss of label retention (Peters et al., 2015) indicates loss of quiescence in this compartment. This data does not contradict an increased BMP signaling in the inner bulge (Figure S4A). An increase in BMP signaling in the inner bulge is in line with reduced inner bulge cell cluster detected in CerS4epi-/- skin via single cell sequencing, likely contributing to the hair loss observed. We will edit this paragraph to make this more clear.

• 1. Paragraph starting with "It is interesting to note that ceramide availability was shown to regulate Wnt signaling in Drosophila through strong effects on recycling endocytosis of the receptors (Pepperl et al., 2013)." Is redundant in the manuscript.*

We apologize for the accidental duplication of this paragraph and thank the reviewer for noticing this mistake.

*

Reviewer #3 (Evidence, reproducibility and clarity (Required)):

The authors created CerS4 mutant mice to test the role of sphingolipids in hair follicle stem cells (HFSCs) and the hair cycle. This work extends previous studies that show that loss of this enzyme leads to defects in the hair cycle and eventually hair loss. In this study the authors look early on in the course of the deletion in an attempt to understand why loss of this enzyme leads to the phenotype described previously. They use single cell profiling, proteomics, and in situ imaging to pinpoint issues in the stem cell niche that drive phenotypes and propose novel interactions between sphingolipid metabolism, Wnt signaling, and inflammation in regulation of HFSC homeostasis. The data are nicely presented, and the text is well written. The conclusions are clearly defined.*

We thank the reviewer for the positive assessment of our work and finding it well written and presented. We further appreciate the constructive comments that will further help us to improve the manuscript.

*While the data are clearly presented, there are numerous issues that are confusing to this reviewer. In addition, some of the phenotypes described are subtle, and thus do not make a convincing case.

1, In figure 1C, the cell proportion analysis suggests there are no OBII or SG in WT. I am not sure how this could be possible. In addition, there appears to be almost no sebaceous cells in either, but the mutant supposedly has much larger sebaceous glands (in Fig 2). In Fig S1I, there is no change in bulge cells? In Fig 1B, there is less HFSCs in the mutant than in the WT, but in 1C, there is more OBI in the mutant. The results in Fig 1B and C are confusing. Also, the schematic in Fig 1 is hard to read, the authors should color code the text with the image.*

It is important to emphasize that the single cell RNA sequencing was carried out at P19 when the bulge stem cell compartment only starts to be established. This explains why only few bulge stem cells are detected at this point. Nevertheless, the OBII and SG cluster is visible also in the wt in Figure 1D. We will include subclustering of the relevant subpopulations to make these populations more clearly visible also in the wt. We will also edit the labels for clarity.

Mature sebocytes are very large cells, and inherent to the single cell sequencing workflows these large cells are excluded from the sequencing libraries. Importantly, we do not detect a change in bulge stem cells in a mouse line in which CerS4 was specifically deleted only in sebocytes (Figure S1I). This analysis was carried out to exclude a sebocyte intrinsic effect on the hair follicle stem cell state and fate. The data does not contradict Fig. 1, as data in Fig. 1 was generated using a different mouse line in which CerS4 is deleted in the entire epidermal stem cell population using K14Cre. We will edit the manuscript to make this more clear.

Data presented in Fig 1B and C focus on two different aspects. Fig 1B shows the inefficient establishment and maintenance of CD34+/integrin-a6+ bulge hair follicle stem cells. The quantification is based on FACS analyses of cells expressing these cell surface molecules/stem cell markers. Fig 1C shows the quantification of the various cell states based on single cell RNA expression and subsequent clustering of the control and CerS4epi-/- epidermal cells together. The “outer bulge” cluster was annotated based on these cells expressing hair follicle stem cell markers. While the CerS4epi-/- epidermis shows increased number of cells in this cluster, the expression of all key stem cell genes (CD34, Sox9, Krt15, Lhx2) is reduced in CerS4epi-/- outer bulge 1 compartment compared to control. Thus, while this “outer bulge” population is expanded in the KO, the stem cell properties of this population are clearly attenuated, as defined by decreased expression of key stem cell transcription factors and increased expression of differentiation genes. We will clarify this in the revised version of the manuscript and also rename this cluster “outer bulge-like” to highlight that these cells are not necessarily bona fide stem cells and might not express high levels of CD34+/integrin-a6+ protein.

*2, In Figure 5, the signaling chart shows a strong upregulation of non-canonical Wnt signaling in the mutant bulge. Canonical Wnt signaling appears to be unchanged between wt and ko. Thus, it is not clear why the authors came to the conclusion that Wnt signaling is induced in the mutant. They further show expression of Lef1 and Nfatc1, but these are not typical markers used to denote canonical wnt activation, as implied. In fact, the data in Fig S4B suggest the induction of Lef1 and Tcf4 is actually very subtle. Instead, the authors should use nuclear b-catenin or transcriptional targets such as Axin or CyclinD. The authors should in fact explore the observation of Wnt5, as that appears to be the most dramatic change. In addition, the authors should use an ontological analysis with the single cell data from the tissue in the same manner that they did for organoids to take another look at molecular consequences of loss of CerS4. *

We agree with the reviewer that further analysis of canonical and non-canonical Wnt signaling will strengthen this conclusion. In our experience, nuclear b-catenin is very difficult to detect in the skin even when Wnt is highly active, but we will investigate Axin2 and CyclinD1 expression. We will also investigate Wnt5a signaling by analyzing its expression as well as its downstream target genes. We will further perform additional ontological analyses from the single cell sequencing data to strengthen the conclusions on the signaling alterations.

* 3, The authors suggest that much of the phenotype is due to inflammation. In Fig 6A, they showed analysis of CD45 cells in the skin. However, the only change was a very subtle change in Th2 cells, while no other CD45+ cells were altered.*

We agree with the reviewer and acknowledge that this experiment was unfortunately underpowered. We will repeat these analyses with a larger cohort.

4, The authors showed upregulation of Immune response in Fig 6C, but then in Fig S2, the genes downregulated are also related to immune response...how do the authors reconcile this?

We apologize for this confusion. Importantly, keratinocyte-intrinsic downregulation of homeostatic immune modulating activity is a key driver of allergic disorders, like atopic dermatitis. This barrier intrinsic immune modulation is distinct from immune cell-mediated inflammation. There is a strong overlap of genes constituting the term “Inflammatory abnormality of the skin” (Human phenotype ontology terms) Fig 6C and “Immune system process” (GOBP terms) Fig S2F. To name some, i.e. Adam-, ALOX-, ASXL- family members are annotated by both terms. Mutations in these genes are known to cause skin diseases associated with immune dysregulation but are likewise known to regulate immune responses.

Data in Figure 6C shows enrichment of this term from both up and downregulated proteins in the CerS4epi-/- condition compared to control, indicating that proteins involved in “Inflammatory abnormality of the skin” are dysregulated in CerS4epi-/- organoids. Data in Figure S2F shows the downregulation of these proteins in CerS4epi-/- organoids compared to control. We will clarify this in the text and figure legends.

5, The author propose that the phenotype in CerS4 null mice is due to disruption of the stem cell Niche. However, the authors have not shown evidence for such an effect through any in situ analysis. The single cell approaches are valuable, but in that case the niche is dissociated. The organoid work is also nice, but not exactly a stem cell niche either. The authors should instead test their hypothesis through an in situ analysis.

We have used the term niche to describe the cellular interactions between stem cells and the other niche resident cells such as the Krt6+ inner bulge cells that have been analyzed here. We will edit the conclusions for clarity. We will further include additional immunofluorescence analyses of the bulge compartment in situ, as suggested (including markers for quiescence and activation).

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #3

Evidence, reproducibility and clarity

The authors created CerS4 mutant mice to test the role of sphingolipids in hair follicle stem cells (HFSCs) and the hair cycle. This work extends previous studies that show that loss of this enzyme leads to defects in the hair cycle and eventually hair loss. In this study the authors look early on in the course of the deletion in an attempt to understand why loss of this enzyme leads to the phenotype described previously. They use single cell profiling, proteomics, and in situ imaging to pinpoint issues in the stem cell niche that drive phenotypes and propose novel interactions between sphingolipid metabolism, Wnt signaling, and inflammation in regulation of HFSC homeostasis. The data are nicely presented, and the text is well written. The conclusions are clearly defined.

While the data are clearly presented, there are numerous issues that are confusing to this reviewer. In addition, some of the phenotypes described are subtle, and thus do not make a convincing case.

1. In figure 1C, the cell proportion analysis suggests there are no OBII or SG in WT. I am not sure how this could be possible. In addition, there appears to be almost no sebaceous cells in either, but the mutant supposedly has much larger sebaceous glands (in Fig 2). In Fig S1I, there is no change in bulge cells? In Fig 1B, there is less HFSCs in the mutant than in the WT, but in 1C, there is more OBI in the mutant. The results in Fig 1B and C are confusing. Also, the schematic in Fig 1 is hard to read, the authors should color code the text with the image.
2. In Figure 5, the signaling chart shows a strong upregulation of non-canonical Wnt signaling in the mutant bulge. Canonical Wnt signaling appears to be unchanged between wt and ko. Thus, it is not clear why the authors came to the conclusion that Wnt signaling is induced in the mutant. They further show expression of Lef1 and Nfatc1, but these are not typical markers used to denote canonical wnt activation, as implied. In fact, the data in Fig S4B suggest the induction of Lef1 and Tcf4 is actually very subtle. Instead, the authors should use nuclear b-catenin or transcriptional targets such as Axin or CyclinD. The authors should in fact explore the observation of Wnt5, as that appears to be the most dramatic change. In addition, the authors should use an ontological analysis with the single cell data from the tissue in the same manner that they did for organoids to take another look at molecular consequences of loss of CerS4.
3. The authors suggest that much of the phenotype is due to inflammation. In Fig 6A, they showed analysis of CD45 cells in the skin. However, the only change was a very subtle change in Th2 cells, while no other CD45+ cells were altered.
4. The authors showed upregulation of Immune response in Fig 6C, but then in Fig S2, the genes downregulated are also related to immune response...how do the authors reconcile this?
5. The author propose that the phenotype in CerS4 null mice is due to disruption of the stem cell Niche. However, the authors have not shown evidence for such an effect through any in situ analysis. The single cell approaches are valuable, but in that case the niche is dissociated. The organoid work is also nice, but not exactly a stem cell niche either. The authors should instead test their hypothesis through an in situ analysis.

Significance

The authors created CerS4 mutant mice to test the role of sphingolipids in hair follicle stem cells (HFSCs) and the hair cycle. This work extends previous studies that show that loss of this enzyme leads to defects in the hair cycle and eventually hair loss. In this study the authors look early on in the course of the deletion in an attempt to understand why loss of this enzyme leads to the phenotype described previously. They use single cell profiling, proteomics, and in situ imaging to pinpoint issues in the stem cell niche that drive phenotypes and propose novel interactions between sphingolipid metabolism, Wnt signaling, and inflammation in regulation of HFSC homeostasis. The data are nicely presented, and the text is well written. The conclusions are clearly defined.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

The manuscript authored by Peters et al. titled "Sphingolipid metabolism orchestrates the establishment of the adult hair follicle stem cell niche to control skin homeostasis" elucidates the critical role of ceramide synthase 4 (CerS4) in the epidermal stem cell niche, particularly in regulating hair follicle bulge stem cells (HFSCs). Using epidermal specific CerS4 knockout mice as an in vivo model and hair follicle organoid culture as an ex vivo model, the authors conducted a comprehensive analysis, which includes cutting edge approaches such as scRNA-seq, proteomics, and lipidomics. The results highlight CerS4's function in the establishment/maintenance of the HFSC niche, as absence of CerS4 changes HFSCs' number and differentiation state. Potential underlying mechanisms identified include altered membrane lipid profiles and Wnt signaling responsiveness. Possible link to a chronic inflammatory skin disease, atopic dermatitis, is also implicated. The data presented are generally of high quality, and the work is significant as it uncovers a new regulator of HFSC fate with mechanistic connection to lipid metabolism.

However, some issues were identified, most of them having to do with in vivo characterization and data interpretation:

Major:

1. The in vivo HFSC phenotype can be better characterized. "Collective, these data show that CerS4 in HFSCs is essential to establish the adult stem cell compartment and to assure lineage fidelity." - this statement premature based on order of the data shown. Also the trajectory difference shown in Figure 2A is not striking. Subclustering out the relative cell subsets and redo the analysis might help to tease out the difference. Additional experiments such as lineage tracing would be useful to support the notion that there is lineage fidelity issue in the mutant - though it is understood that this is quite involved and may lie outside the scope of the current study. Are bulge cells in the mutant proliferative? - the authors should consider in vivo Edu labelling experiment or the like to assess the quiescence/proliferation of the bulge cells. Finally, analyzing hair follicles at earlier stages might help to clarify when and where the bulge and sebaceous gland changes start - is possible that aberrant divergence of bulge/sebaceous fates occur prior to the establishment of a stable bulge fate?
2. The exclusion of IFE contribution is not backed up by data. Figure 6D - model emphasizes HFSC involvement in atopic dermatitis, but this could be due to epidermal barrier defect. Barrier defect could already be present even though IFE morphology appears normal. Maybe TEWL is measured at the time of analysis and shows no change - if so, this data should be included. HFSC changes might contribute but the involvement of IFE cannot be excluded. The conclusion that "CerS4 expression was restricted to the hair follicle" is not supported by data. IFE expression is apparent in Figure S1C. Along this line, there is also an apparent expansion of IFE basal II in the mutant (Figure 1C).
3. Figure 1 - single cell analysis was done using only 2 pairs of mice, and data in E lack statistical assessment. At the very least, data for individual pairs should be shown in supplemental data to ensure that changes are consistent in both mutant mice rather than being dominated by dramatic alteration in only 1 mutant mouse.

Minor:

1. CerS4SCD3-/+ nomenclature is mis-leading.
2. Figure 2- "Furthermore, we observed expansion of the inner bulge identity marker Krt6 protein expression into outer bulge stem cells and along the infundibulum in CerS4epi-/- hair follicles, whereas in control mice Krt6 was restricted to the inner bulge (Fig. 2C)." - Krt6 staining is presented in Fig 2D, not 2C.
3. Figure 3C - size of the organoids should be quantified with statistics. The images shown do not support the statement that "Strikingly, CerS4epi-/- organoids showed altered morphology characterized by smaller size and loss of cohesion of peripheral cells from the organoid clusters (Fig. 3C), ...".
4. Section titled "CerS4 regulates HFSC differentiation in a stem cell autonomous manner": "CD34- integrin- a6+ cells, which based on extensive transcriptome and marker expression analyses represent a mixture of HFSCs, hair follicle outer root sheath (ORS) cells and inner bulge cells (collectively termed non-HFSCs)." - shouldn't the CD34- integrin- a6+ population also contain IFE stem/progenitor cells? Are hair follicles micro-dissected out for FACS?
5. Figurer 5D - please provide the working concentration of Chir99021. Figure 5F - explain what arrows mean in legends.
6. Figure 6A - no significant changes in Th2 and ILC2 were observed at a 95% confidence interval. Increasing mouse number will help to increase statistical power.
7. Additional Wnt target genes such as Axin 2 should be looked at.
8. The increased BMP signaling and decreased Nfatc1 expression are seemingly contradictory.
9. Paragraph starting with "It is interesting to note that ceramide availability was shown to regulate Wnt signaling in Drosophila through strong effects on recycling endocytosis of the receptors (Pepperl et al., 2013)." Is redundant in the manuscript.

Significance

The work uncovers a new regulator of HFSC fate with mechanistic connection to lipid metabolism and development signaling. The same group previously reported epidermal and hair cycling phenotypes of the same mutant mice, but this work now identifies a specific defect in HFSCs and present evidence for cellular, molecular and biochemical changes. Linking stem cell regulation to lipid metabolism is conceptually novel, and should have a broad audience. However, the study does have some limitations, such as lack of definitive evidence that CerS4 function in HFSCs is responsible for all the defects reported here, and that lipid alterations have a causal relationship with altered Wnt signaling.

My expertise is in skin biology, stem cell control, and developmental signaling.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

Deletion of CerS4 in the entire mouse epidermis throughout development via the K14-Cre results in enlarged sebaceous glands and perturbed HFSC molecular phenotype. There is low or no expression of CD34, a known marker of the HFSCs along with apparent reduction of several other HFSC markers and acquisition of a more differentiated cell phenotype in these cells. Interestingly, skin and hair follicles seem to remain normal otherwise up to advanced age, though this contradicts the notion that HFSC were indeed affected at the functional level. The data does not demonstrate 'gradual decline' in the HFSC compartment, as claimed by the authors, but rather seem to indicate that the adult HFSC compartment is not properly established in its molecular signatures. Organoid cultures document defects in HFSC, which included reduced proliferation in the CerS4 KO cells. Lipid composition in plasma membranes was also affected by the CerS4 KO. Associated with this, Wnt signal transduction is also affected according to experiments that enhance the strength of wnt signals via a specific small molecular agonist of the pathway. Finally, the authors discover a resemblance of the mouse KO immune-phenotype, with human atopic dermatitis. The study is likely of interest to a specialized readership in skin biology and dermatology and adds to previous studies on CerS4 in skin that erroneously placed its role in the sebaceous gland. [The authors here demonstrate that deletion of CerS4 in the sebaceous glands via SCD3-Cre led to no phenotype, contradicting the previous assessment that CerS4 is important in sebaceous glands.] The study would need to be corrected in a few of its interpretations regarding stem cells to better match the data, as indicated below.

Major revisions:

Fig1B - the data seems to simply shows that bulge cells express less or no CD34 and not that ' CerS4epi-/- mice showed reduced HFSC numbers'; the primary FACS data should be shown somewhere too. The conclusion that stemness is affected, and HFSCs lose their normal gene expression signature is at more convincing after looking at other HFSC markers down the road in the paper. However, in the absence of functional assays that would demonstrate stem cell function is lacking and seeing that hair follicles are maintained and grow in long-term, the notion that stem cells are lacking in these conditions is not supported by the data.

The conclusion after figure 2: "Collectively, these data indicate that CerS4-deficiency triggers ... ... gradual depletion of the quiescent HFSC compartment." There is no data showing gradual depletion of the quiescent HFSC compartment. We would need to see a gradual activation of HFSCs with over proliferation to conclude this. There is some data albeit not always convincing (see NFAC1 staining in Fig. 5C) indicating loss of markers associated with quiescence but there is no data indicating 'gradual' loss of markers.

Minor revisions:

Figure 1C legend - please spell out what are the abbreviations for the different subpopulations; please show these populations as % as opposed to absolute numbers.

Figure 1D - please make it clear in the cartoon what the different sub-populations listed are;

Is OB1 a CD34- HFSC population?

Fig. 3B - the colors of the legend do not match the colors in the data so it is confusing as to which one is which!

Fig 5C - the differences in NFATc1 are not visible in the images shown

Significance

Deletion of CerS4 in the entire mouse epidermis throughout development via the K14-Cre results in enlarged sebaceous glands and perturbed HFSC molecular phenotype. There is low or no expression of CD34, a known marker of the HFSCs along with apparent reduction of several other HFSC markers and acquisition of a more differentiated cell phenotype in these cells. Interestingly, skin and hair follicles seem to remain normal otherwise up to advanced age, though this contradicts the notion that HFSC were indeed affected at the functional level. The data does not demonstrate 'gradual decline' in the HFSC compartment, as claimed by the authors, but rather seem to indicate that the adult HFSC compartment is not properly established in its molecular signatures. Organoid cultures document defects in HFSC, which included reduced proliferation in the CerS4 KO cells. Lipid composition in plasma membranes was also affected by the CerS4 KO. Associated with this, Wnt signal transduction is also affected according to experiments that enhance the strength of wnt signals via a specific small molecular agonist of the pathway. Finally, the authors discover a resemblance of the mouse KO immune-phenotype, with human atopic dermatitis. The study is likely of interest to a specialized readership in skin biology and dermatology and adds to previous studies on CerS4 in skin that erroneously placed its role in the sebaceous gland. [The authors here demonstrate that deletion of CerS4 in the sebaceous glands via SCD3-Cre led to no phenotype, contradicting the previous assessment that CerS4 is important in sebaceous glands.]

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

Manuscript number: RC-2023-02132R

Corresponding author(s): Halyna, Shcherbata

Point-by-point description of the revisions

We would like to sincerely thank the reviewers for the positive evaluation of our work, careful reading of our manuscript, and helpful suggestions. In the revised version of our manuscript, we have introduced the proposed changes and added the new data based on the suggested experiments to address the reviewers’ concerns. We hope that this modified version of the manuscript is now acceptable for publication.

• *

Reviewer #1 (Evidence, reproducibility and clarity (Required)):

Summary Elucidating the cellular and molecular mechanisms underlying age-related neurodegeneration remains a key challenge for neurobiologists. In this manuscript, Mariana Tsap and colleagues in the team of Halyna Shcherbata focus on the function of the neuropathy target esterase NTE/Swiss Cheese (Sws) in the Drosophila brain. The authors use an elegant combination of genetics, light and electron microscopy, RT-qPCR and GS-MS mass spectrometry to determine the complex role of Sws in cellular blood brain barrier (BBB) integrity, the brain inflammatory response and fatty acid metabolism. The study provides a detailed characterisation as to how the loss of sws affects glial cell morphology in the BBB revealing abnormal membrane accumulations and tight junctions, and in consequence causing permeability issues. Importantly, they observed the upregulation of antimicrobial peptides in the brain, indicative of neuroinflammation, as well as of fatty acids, equally connected with the inflammatory response.

Major comments

The study provides a detailed and comprehensive characterization of the sws mutant phenotype, and in particular the role of this gene in blood-brain barrier forming glia.

The study connects neurodegeneration and inflammation, but also makes a particular point about "inflammaging". However, the age contribution has not been studied in detail. Indeed, the flies analyzed are 15 days old (according to the Material and Methods section, with the exception of Figure 1 where flies are 30 days old), and hence have not been compared with younger or older flies to make a point of age as evoked in the abstract, introduction or discussion. The authors should either add experiments comparing differently aged flies or de-emphasize this point to a brief consideration in the discussion. Instead, it would be very helpful to provide concise information about the current knowledge concerning the inflammatory response in the Drosophila brain. We thank the reviewer for raising this point. The decision to use 15-day-old flies was made due to the high mortality of swsmutants after two weeks and because age-dependent character of sws neurodegeneration has been previously well described. As the reviewer suggested, now we also included old animals in our experiments to show a connection between age-dependent neurodegeneration and inflammation. We measured and compared the mRNA levels of expression of the antimicrobial peptides (AMPs) Attacin A, Cecropin A, and Diptericin in the heads of 15- and 30-day-old sws loss-of-function mutants, in the heads of flies that had sws downregulation only in SPG cells (moody>swsRNAi) and in the heads of flies expressing NTE/SWS in SPG cells in sws mutant background. We found that the expression levels of the antimicrobial peptides are increased in the age-dependent manner in the tested mutants. In addition, we found that the expression of NTE/SWS in SPG in sws mutant background reduces inflammatory response in aging animals (see Figure 5D). Also, as the reviewer suggested, we provide brief information on the current understanding of the inflammatory response in the Drosophila brain in the Introduction and Results sections.

Related to this point, the authors convincingly show that sws is required in surface glia using rescue experiments. Nevertheless, all experiments rely on drivers and mutants that could cause the emergence of phenotypes during development. Thus, to strengthen the causative link between the breakdown of the BBB and the neuroinflammatory response, it would be helpful to consider an acute knock-down in adults after BBB formation has been completed. To strengthen the causative link between the breakdown of the BBB and the neuroinflammatory response during adulthood, we performed qPCR analysis and measured the mRNA levels of the antimicrobial peptides Attacin A, Cecropin A, and Diptericin in the heads of flies with sws downregulation in glia cells induced after the blood-brain barrier was formed using the Gal80ts tool. We found that sws downregulation in glial cells during adulthood, after the BBB is formed, leads to the increased inflammatory response (new Supplementary Figure 4E).

To test the brain permeability barrier, the study uses a 10 KDa dextran permeability assay. Almost 25% of brain in controls show a leaky barrier. It would be helpful to describe the causes for this relatively high occurrence. The observed relatively high occurrence of a leaky barrier phenotype in our control group may be attributed to our experimental procedure. We injected flies peritoneally and waited for over 12 hours before dissecting their brains for the permeability assay. Typically, such analyses are conducted after shorter periods, often around 2 hours. Additionally, we used Dextran with the smallest molecular weight (10kDa). The blood-brain barrier (BBB) is not 100% impermeable, and small molecules can gradually enter the brain over time. Recent studies have shown that this entry could be facilitated by endocytosis (Artiushin et al, 2018), which could partially explain the presence of Dextran 10kDa in control brains. Considering this, using a larger Dextran (70kDa) in our experiments could have been more accurate. Importantly, we always compared mutants and controls that underwent identical treatment, dissection, and analysis. We conducted experiments in multiple biological replicates to accurately assess the significance of the differences between mutants and controls. Therefore, we are confident that the differences we observed between controls and mutant flies in the BBB permeability are significant. We included all relevant numbers and statistics for these experiments in Supplementary Table 4.

An important point in the study concerns the increase of free fatty acids as cause of the inflammatory response. The measurements were based on measurements of whole heads, which could include the hemolymph and fat body within the head in addition to brain. However, the causative relationship remains unclear and the question why a leaky blood brain barrier would increase the free fatty acid levels in the body or brain remains mainly an observation at the descriptive level. Here, it would be helpful to design an experiment, which could test the causative links or to modify the interpretation in scheme 6D and adjust the wording in the text. We agree that the causative relationship between a leaky blood-brain barrier and increased free fatty acid levels in the body or brain is currently an observation at the descriptive level and that it would be important to investigate the correlation between a leaky blood-brain barrier, inflammation, and increased free fatty acid levels in greater detail in future studies. In the modified manuscript, we have changed the scheme in Figure 5G and adjusted the wording in the text.

Related to this, how do the levels of AMP caused by a leaky BBB would compare to an elicited neuroinflammation by the presence of bacteria? The neuroinflammatory response can be accompanied by macrophage entry into the brain following AMP induction. Could the authors detect this response (which could be envisioned as manipulations include pupal development, provided macrophages would persist into adulthood)? This would make a strong point regardless of the outcome. We thank the reviewer for suggesting this excellent experiment. To detect macrophage entry into the mutant brains, we used antibodies (NimC1) and srp(Hemo)>mCherry that label the macrophage cells. We found macrophages in the larval and adult sws mutant brains and also in adult brains upon downregulation of sws in SPG cells (Figure 5E-F and Supplementary Figure 4F-IG. These data additionally support our hypothesis that a leaky BBB in sws mutants induces neuroinflammation, which is accompanied by macrophage entry into the brain following AMP expression.

Expression of sws is determined using sws-Gal4 driving membrane-tethered GFP. As sws is expressed very widely and classical Gal4 lines tend to be active in the BBB, it is important to provide the exact information about the nature of this driver. We appreciate the reviewer for bringing this to our attention. We have now included information about the line we used to express transgenes in a sws-dependent manner. Specifically, we utilized the y*w*P{GawB}swsNP4072/FM7c line (Kyoto Stock Center 104592), which was generated using the Gal4 enhancer trap element P{GawB} insertion strategy.

The Material and Methods section should contain a proper Quantification and Statistical analysis section. In the Figures, it would be helpful to refer to the Table reporting sample numbers. As the reviewer suggested, we have now included a Quantification and Statistical analysis section in the Materials and Methods. Additionally, we ensured that all figure legends include a reference to the corresponding tables reporting sample numbers and statistics.

In Figure 5, it would be important to indicate sample numbers, the nature of the error bar, and show data points together with columns. We agree with the reviewer that it is important to report all sample numbers and statistics. We generated a new Supplementary Table 1 for all qRT-PCT data, and Supplementary Tables 5 containing all "n" values and corresponding p-values. In the Figure Legends, we denoted the type of error bars and deviations, included p-values, and referred to the relevant tables for comprehensive numerical data.

Minor comments

On page 8, cell death is visualized using "the apoptotic marker Cas3". It should be Caspase-3. Moreover, it is not clear whether this antibody (directed against vertebrate Caspase-3) recognizes indeed Caspase-3 in Drosophila? This should be formulated more carefully. As the reviewer correctly noted, the Caspase-3 antibody is designed for human Caspase-3. While it has been employed in Drosophila apoptosis research, its specificity for Caspase-3 in Drosophila is unclear. Given the very well-documented apoptosis in sws mutants (Kretzschmar et al, 1997; Muhlig-Versen et al, 2005) and the non-focus on neuronal cell death in this research, we have opted to exclude this information from the supplementary figure. We appreciate the reviewer for bringing this to our attention and for the valuable suggestion.

On Page 9 (3rd paragraph), the authors report that they "want to understand what signaling pathway is activated." However, the described experiments do not lead to a signaling pathway, but conclude that an antiflammatory response is evoked. This should thus be reworded. Thank you for pointing this out. In the revised version, we state that we wanted to understand whether the compromised brain barrier in sws mutants triggers the activation of any cellular stress pathways, including apoptosis, ferroptosis, oxidative stress, ER stress, and inflammation.

Figure 1 reports the expression pattern and phenotype of sws; thus, the title of the figure should be extended. Thank you for the suggestion. We have updated the title of Figure 1 to more accurately reflect its content. The revised title is now: NTE/SWS is expressed in Drosophila brain and its loss leads to severe neurodegeneration.

Concerning the description of phenotypes, the authors use the term "clumps", but it is not clear what this entails (e.g., Page 6, or Figure 6). For the reader, it is also necessary to refer to original studies of moody to understand the septate junction phenotype represented in the figure. As the reviewer suggested, we changed the word “clumps” to “clusters”. We also agree with the reviewer’s recommendation to cite the original work on Moody to acknowledge previous research and enhance the understanding of moody phenotypes. We have now included the relevant citations in the manuscript.

**Referees cross-commenting**

I fully agree with the comments of the other two reviewers, as they were complementary and overlapping with mine (e.g. the contribution of age).

Reviewer #1 (Significance (Required)):

This study provides a detailed cellular and functional characterization of the swiss cheese phenotype in the blood-brain barrier so far not reported in previous studies, including the team's own earlier publications (e.g., Kretzschmar et al., 1997; Melentev et al., 2021 and Ryabova et al., 2021). Furthermore, it uses cutting-edge technology to provide links to neuroinflammation and neurodegeneration, Previous studies explored neuroinflammation in the brain of Drosophila by challenging the organism with bacteria to mount an inflammatory response (Winkler et al., 2021). Intriguingly, this current study provides evidence, that a leaky blood brain barrier alone could lead to an inflammatory response, and that in turn, treatment with anti-inflammatory agents could reduce the cellular defects in glia and in consequence neurodegeneration. This represents an important conceptual advance that will be of wide interest to neurobiologists interested in glial biology, neuroinflammation and neurodegeneration in Drosophila and in vertebrates. One possible limitation of the study may be that while complex cellular processes have been pinpointed, some of the causative links of the BBB with neuroinflammation remain unexplored, in particular the aspect of elevated free fatty acids/antimicrobial peptides.

We appreciate the reviewer's recognition of the conceptual significance of our study, revealing that a leaky blood-brain barrier alone can induce an inflammatory response, with subsequent treatment using anti-inflammatory agents and the importance of these findings for neurobiologists. We also thank the reviewer for thorough examination and insightful suggestions. Given that prior studies have demonstrated the induction of neurodegeneration by the overactivation of innate immune-response pathways, especially elevated expression of antimicrobial peptides (Cao et al, 2013), our new experimental data showing increased levels of antimicrobial peptides in aging flies with a defective BBB further strengthen the connection between the BBB, AMPs and neuroinflammation. This link is even more enhanced by the rescue experiments and the detection of macrophage entry in the mutant brains. We trust that the implemented revisions, accompanied by supplementary experimental data, enhance the suitability of our manuscript for publication.

Reviewer #2 (Evidence, reproducibility and clarity (Required)):

The manuscript by Tsap et al describes a role of NTE/SWS in forming the BBB in Drosophila. Disruption of the BBB in SWS mutants and knockdown flies results in morphological changes of the glia forming the BBB, increased brain permeability, altered lysosomes, and an upregulation of innate immune genes. The experiments to show a function of SWS in surface glia and the resulting changes in permeability are well supported by the experiments and the statistics appears appropriate.

The authors also show changes in innate immune genes and some fatty acids and that similar changes are found in another mutant affecting the BBB. They discuss that these changes are a consequence of the disruptions of the BBB but also that these changes induce changes in the BBB. To address this and confirm that the changes in immune genes and fatty acids is a consequence of the altered BBB, they should include experiment expressing SWS in the surface glia and measure if that normalizes these changes. Another major aspect that should be addressed is the effect of aging. As the authors point out, loss of SWS causes age-dependent phenotypes (shown by the author and others) and with the exception of figure 3F, the age isn't even mentioned in any of the other figures. Furthermore, at least some of the experiments should be done at different ages to determine whether the phenotype is progressive; this includes the permeability assays and the measurements of immune genes (the latter could also support whether changes in the immune genes affect the BBB or vice versa the BBB changes cause the upregulation of immune genes).

As the reviewer suggested, in order to establish a connection between age-dependent correlation between neurodegeneration and inflammation, we analyzed the mRNA expression levels of antimicrobial peptides in the heads of both 15- and 30-day-old sws loss-of-function mutants, as well as in flies with sws downregulation specifically in SPG cells (moody>swsRNAi). We found that the expression levels of the antimicrobial peptides are increased in the age-dependent manner in the tested mutants (Figure 5D, red and orange bars). Following the reviewer’s recommendation, we also performed an experiment where we expressed NTE/SWS in the surface glia in a sws mutant background (sws1; moody>sws, rescue). We measured mRNA levels of Attacin A, Cecropin A, and Diptericin in the heads of 15- and 30-day-old flies (Figure 5D, blue bars). The results showed that the levels of all three AMPs were not significantly different or slightly upregulated in the heads of “rescue” animals compared to Oregon R controls (Figure 5D, compare green and blue bars, and see Supplementary Table 1). Importantly, the levels of all AMPs were significantly lower in the heads of 30-day-old rescue animals than in the heads of the same age sws1mutants (Figure 5D, compare red and blue bars, green stars, see also Supplementary Table 1). These findings further support our hypothesis that sws deficit in the surface glia induces an immune response in age-dependent manner.

We did not conduct the Dextran permeability assay in older flies because approximately 90% of the 15-day-old flies with swsderegulation already exhibited impaired permeability of the BBB. This suggests that the phenotype is quite severe and may not show significant age-dependent progression. Moreover, older mutant flies were extremely weal, and it is likely that they would not have survived the peritoneal injection procedure.

Lastly, the authors claim that septate junctions are defective in sws mutants. However, this should be confirmed by EM studies (which the authors have already done) besides immunohistochemistry which doesn't provide enough resolution.

As the reviewer suggested, for a more detailed detection of septate junctions, we conducted additional electron microscopy experiments. The images included in Figure 6D-F show irregular aggregates and disruptions in the structures of septate junctions and membranes in sws mutants compared to controls. Additionally, we display the appearance of tight junctions in moody mutants (Supplementary Figure 5E-F), which look dramatically different compared to sws junctions and, as previously described, appear overgrown.

Reviewer #2 (Significance (Required)):

A role of SWS in maintaining the BBB and what consequences this has provides another insight how this protein (and its homolog NTE) affects brain health. Although a function of SWS in glia (as well as in neurons) has previously been described, changes in the surface glia and the BBB is a novel aspect. However, the causative role of SWS on some of the described consequences (see above) should be confirmed. Although the manuscript can add to a better understanding of the connection between disruptions of the BBB and neurodegenerative diseases, which is of interest for a broader field of researchers, the discussion of the results is quite speculative.

We appreciate the recognition of the novelty of this work and its potential contribution of our manuscript to a better understanding of the connection between disruptions of the BBB and neurodegenerative diseases. We thank the reviewer for the constructive feedback and hope that introduced changes, along with additional experimental data that address the concerns raised, strengthen the proposed role of sws in the formation of tight junctions in the BBB and its age-dependent maintenance.

Reviewer #3 (Evidence, reproducibility and clarity (Required)):

Summary: The study of the formation and maintenance of the blood-brain barrier (BBB) is a growing field of study, partly due to its strong link with neurological disorders. The BBB depends on the role of multiple cell types and mechanisms. Mutations in the conserved phospholipase NTE/SWS can lead to neurodegeneration, and previous work from the authors shows that SWS loss leads to abnormal glial morphology. In this work, authors use Drosophila to further study this phenotype, showing that SWS is mostly expressed in the BBB-related glia and that its loss leads to abnormal BBB permeability, increased inflammatory response and neural cell death. Interestingly, authors observed a dependence for the BBB-defective phenotype on aging, with important implications for SWS/NTE and neurodegeneration. Overall, the work represents a clear advance in the poorly explored role of NTE/SWS in neurodegeneration, with a broad impact on the understanding of BBB maintenance. This work shows a combination of multiple and appropriate experimental approaches, including confocal microscopy, EM, RT-qPCR, or gas chromatography-mass spectrometry among others.

Major comments:

The use of sws1 and sws1/sws4 transheterozygous animals, together with the use of sws RNAi is a solid approach to validate that the reported phenotypes are due to SWS loss. Using these models, the authors performed a convincing structural analysis of the subperineurial glia phenotype, and showed that it is accompanied by a defective BBB, inflammation and neuronal cell death. The key conclusions are properly supported by the data. However, there are some claims in the text that are not supported by any data in the Figures, but only qualifications. This needs to be fixed:

-Page 6, third paragraph:

"...we specifically downregulated sws in the nervous system using the double driver line that allows downregulation of sws in glia and neurons (repo, nSyb-Gal4, Suppl. Fig. 2C-Cʹ). Since these animals had the same disorganized structure of brain surface as the loss-of-function mutant..." Supp. Fig. 2C-C' only shows expression of CD8:GFP and nlacZ reporters by repo and nSyb-Gal4, but there is no data showing sws RNAi expression by these drivers.

We thank the reviewer for noticing these referencing mistakes. We have corrected the references to the expression patterns of the glial and/or neuronal Gal4 drivers (Supplementary Figure 1D, E and F). Bar graph in Supplementary Figure 1C shows RT-qPCR analysis of sws mRNA levels from flies with glial and/or neuronal sws downregulation (repo>swsRNAi, nSyb>swsRNAiand repo, nSyb>swsRNAi), and the images of mutant brains in Supplementary Figure 2 and Figure 2A-C show the surface glia phenotypes in these mutants.

"...Moreover, downregulation of sws in all glial cells (repo>swsRNAi) resulted in the same phenotype. At the same time, upon sws downregulation in neurons,... (Suppl. Fig. 4)..." Suppl. Fig. 4 only shows nsyb>swsRNAi data but not repo>swsRNAi

We show now both repo>swsRNAi and nSyb>swsRNAi (Supplementary Figure 2C and 2E, respectively).

-Page 6, fourth paragraph:

"Importantly, expression of Drosophila or human NTE in these glia cells rescued this phenotype (Fig. 2H)"

In addition to the indicated quantifications, it is essential to show some representative data showing the phenotype when Drosophila or human NTE are expressed in glial cells of sws mutant animals. We agree with the reviewer that it is important to show the rescue phenotypes. We have included images of the brain surface of sws mutants that have Drosophila or human NTE expressed in glial cells (Figure 2D and Supplementary Figure 2F).

-Page 9, last paragraph: "We found that in moody mutants, the surface glia phenotype analyzed using CoraC as a marker could also be suppressed by NSAID and rapamycin (Fig. 5A)." In addition to the indicated quantifications, it is essential to show some representative data showing the phenotype with and without treatments.

We appreciate the reviewer's suggestion, and as recommended, we have included representative data showing the phenotype with and without treatments in Supplementary Figure 4C-D.

A more detailed analysis of two aspects of the data would clearly improve the manuscript, whose findings are a bit superficial in the current state:

• The exact mechanism by which BBB permeability leads to brain inflammation remains unknown. Authors show that accumulation of polyunsaturated fatty acids (known to regulate inflammation) occurs in sws-depleted animals. However, they only observed a correlation between this phenotype and the inflammatory response, while is not clear whether the accumulation of polyunsaturated fatty acids causes inflammation in this model or is a consequence of it. An attempt to rescue the accumulation of polyunsaturated fatty acids (i.e., knocking down a required enzyme for their production) in sws mutants might help to understand this. Also, the fact that the defective BBB phenotype observed in either sws KO and glia-specific KD can only be partially rescued by the use of inflammation inhibitors, suggests that other pathways are involved.

We agree with reviewer that since the use of inflammation inhibitors only partially rescue the defective BBB phenotype in swsmutants, it implies the involvement of additional pathways. While our data reveal a correlation between the accumulation of polyunsaturated fatty acids and the inflammatory response, whether this accumulation causes inflammation in our system remains to be studied. We have revised the text to ensure that this explanation is clearly stated without overemphasis.

• While the differences between the phenotypes caused by sws or moody loss are well characterized, it would be key for this work to further study the mechanisms by which sws controls septate junctions. The authors propose the organization of lipid rafts, but some experiments in that direction to check this hypothesis. For example, can authors reproduce the septate junction phenotype of sws mutant (Fig. 6C) by using a different approach to induce defective lysosomes in subperineurial glia?

We appreciate the reviewer's suggestion for such an insightful experiment. To investigate whether the septate junction phenotype observed in sws mutants can be replicated in mutants with defective lysosomes in subperineurial glia, we downregulated several key lysosomal genes in SPG cells: moody>DysbRNAi, moody>Npc1aRNAi, moody>PldnRNAi, andmoody>spinRNAi (Supplementary Figure 6A-E). We were happy to see that downregulation of any of these genes resulted in abnormal formation of SJs and membrane organization in SPG cells. These additional experiments strongly support our hypothesis that lysosomal control of membrane homeostasis significantly impacts the appearance of SJs. Thank you for this excellent idea.

The attempt of the proposed approaches above should require about 3-6 months of investment, with limited economic effort, given the availability and diversity of lines found in the existing stock centres such as Bloomington or Vienna.

The data is presented very clearly, and the methods are adequately detailed, and the experiments and statistical analysis are adequate.

Minor comments:

Prior studies are referenced appropriately, but there is a case that should be addressed. On Page 3, first paragraph, regarding the sentence: "However, the molecular mechanisms underlying inflammaging remain unclear". I recommend specifying what is known and what is unknown in the field. Ideally describing (briefly) the knowledge about lipids, inflammaging and neurodegeneration, which are the specific topics of the research. Otherwise, the current sentence is too vague, while there is a lot of work published about it.

As the reviewer suggested we have extended the first part of our introduction to briefly describe how inflammaging is connected with the BBB, fatty acid metabolism and lysosomal functions.

The text and figures are clear and accurate. The logic of the experiments and the results are exposed very clearly (for example, the Suppl. Tables are very helpful). There are a few minor issues, however, that should be addressed:

• Page 4, first paragraph: regarding the sentence: "For various obvious reasons, humans are not ideal subjects for age-related research.", I recommend specifying the main reasons (i.e. life cycle, ethical issues, etc.?).

Thank you, the main reasons are specified now.

• I would recommend moving the text "For various obvious reasons...disrupted upon ageing." From its current position to just before "Drosophila melanogaster is an excellent...". This would keep a better logic in the text by explaining NTE first and later introducing the models to study its function. Presenting then Drosophila.

Thank you, done.

• To support the sentence "Together, Drosophila satisfies...neurodegeneration during aging", instead of citing so many papers, I recommend citing just one current review about it, since the amount of literature supporting the claim is huge and should not be limited to a few "random" articles. An alternative might be indicating that the lab has used Drosophila for this aim before, and then citing the examples from the literature.

Thank you for this suggestion, now we referenced few recent reviews and referred to our previous work on the topic.

• Page 4, second paragraph: if NTE/SWS is going to be used as a synonym for NTE/SWS loss of function (or other type) model, it needs to be specified. Otherwise, refers to the proteins and sentences like "NTE/SWS has been shown to result in lipid droplet accumulation..." are misleading. Thank you for the suggestion; we have now specified that NTE/SWS is used as a synonym for the SWS protein in Drosophila and corrected this throughout the manuscript.

• Page 4, last paragraph: the first time that "BBB" is used, its meaning should be specified. And three lines below use "BBB" instead of blood-brain barrier.

Thank you, corrected.

**Referees cross-commenting**

I agree with the comments provided by the other reviewers. They are well reasoned and cover some aspects of the work that I did not see. Regarding the main issue, the three revisions point at the same direction, that is the limited analysis about the mechanism underlying the phenotypes.

Reviewer #3 (Significance (Required)):

• Describe the nature and significance of the advance (e.g. conceptual, technical, clinical) for the field. This work represents a substantial advance in the understanding of NTE/SWS function in the context of neurodegeneration, and opens potential approaches to treat related disorders (they successfully use anti-inflammatory compounds to ameliorate some of the key phenotypes). However, the findings are a bit superficial in terms of mechanisms, and further analysis (see major comments) would notably improve the significance of the manuscript. This should be realistic and suitable, given the advantages of the Drosophila model and the availability of tools.

• Place the work in the context of the existing literature.

The role of SWS in regulating lysosomal function is potentially supported by NTE-deficient mice data (Akassoglou et al., 2004; Read et al., 2009), where different types of neurons show similar dense bodies containing concentrically laminated and multilayered membranes than those observed in this work in Drosophila sws mutant. Potentially, the rest of the work has a translation to mammals, which is supported by the fact that ectopic expression of NTE rescues some of the key phenotypes described in the manuscript.

• State what audience might be interested in and influenced by the reported findings. Neuroscience in general, since the study of BBB and neurodegeneration has a clear general interest in the whole field.

• Define your field of expertise with a few keywords to help the authors contextualize your point of view. Indicate if there are any parts of the paper that you do not have sufficient expertise to evaluate. Drosophila; Neurodegeneration; Hereditary Spastic Paraplegia; Alzheimer's disease; Motor neurons; Microglia; Endoplasmic reticulum; Mitochondria.

Lipid metabolism is the part of the manuscript where I have less expertise to evaluate, only having general knowledge about it.

We appreciate the positive evaluation of our work, the careful reading, and the valuable suggestions provided by the reviewer, including recommendations for additional experiments and changes in the text. We believe that the implemented changes, combined with the new experimental data, have improved the manuscript, making it ready for publication.

References

Artiushin G, Zhang SL, Tricoire H, Sehgal A (2018) Endocytosis at the Drosophila blood-brain barrier as a function for sleep. Elife 7

Cao Y, Chtarbanova S, Petersen AJ, Ganetzky B (2013) Dnr1 mutations cause neurodegeneration in Drosophila by activating the innate immune response in the brain. Proc Natl Acad Sci U S A 110: E1752-1760

Kretzschmar D, Hasan G, Sharma S, Heisenberg M, Benzer S (1997) The swiss cheese mutant causes glial hyperwrapping and brain degeneration in Drosophila. J Neurosci 17: 7425-7432

Muhlig-Versen M, da Cruz AB, Tschape JA, Moser M, Buttner R, Athenstaedt K, Glynn P, Kretzschmar D (2005) Loss of Swiss cheese/neuropathy target esterase activity causes disruption of phosphatidylcholine homeostasis and neuronal and glial death in adult Drosophila. J Neurosci 25: 2865-2873

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

'The authors do not wish to provide a response at this time.'

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #3

Evidence, reproducibility and clarity

Summary:

In this manuscript the authors are interested in understanding how fission yeast respond to a Nitrogen Signaling Factor (NSF) that has previously been shown to allow Leucine auxotrophs to grow in the presence of Leucine when Nitrogen Catabolite Repression (NCR) is triggered by the presence of a high quality Nitrogen source such as Ammonium Chloride (NH4Cl).

The authors begin with a screen to identify genes that affect the ability of wild type cells grown near cells with leucine auxotrophy to enhance or abolish NCR phenotype. They screened the non-essential gene deletion library which they manipulate so that it only contains a leucine auxotrophy (unlike the original gene deletion library which contains additional auxotrophies). They identify 137 genes whose deletion allows growth of Leu auxotrophs in the presence of Leucine and Ammonia without the presence of WT cells. These genes are required for NCR. They further identify 203 genes which do not bypass NCR even in the presence of wild type cells, and are thus important for bypassing NCR in the presence of WT cells.

They then conduct a second screen to identify which of these genes are important for bypassing NCR in response to the Synthetic NSF, 10(R)-hydroxy-8(Z)-octadecenoic acid, by looking for genes which grow in the presence of leucine when ammonia is not present, but do not grow in the presence of leucine when ammonia is present, even when NSF is added. This second screen identifies 117 strains carrying deletions in a gene set enriched for genes related to cellular respiration and mitochondria. They then show that the NSF bypass of NCR is linked to respiration by showing that it is abolished in the presence of the respiration inhibitor Antimycin A, that growth in low levels of glucose can bypass NCR in the absence of NSF< and that cells supplemented with NSF have a higher oxygen consumption rate.

To gain insight into how the cell responds to NSF, the authors then gather RNA expression data from cells grown in high ammonium concentrations following treatment with NSF relative to a negative control treated only with Methanol (the vehicle into which NSF is dissolved). They argue that the gene expression pattern resembles gene expression data from cells undergoing respiration in glycerol relative to cells undergoing fermentation in glucose. They show that the upregulated genes relate to trehalose synthesis, detoxification of Reactive Oxygen Species, and cellular fusion and the downregulated genes are related to cellular adhesion and flocculation.

They validate their RNA-seq measurements by showing that the two most highly induced and two most highly repressed genes respond to NSF addition in a dose dependent manner and do not respond oleic acid which is chemically similar to NSF. The most highly responsive gene they identify is an uncharacterized gene, SPBPB2B2.01, which they suggest naming "NSF-responsive amino acid transporter 1" (nrt1). They also show that the nrt1 response is dependent on the culture density, and that the response is present (though the magnitude varies) in YES and in EMM under varying nitrogen concentrations, and that yfp driven by the nrt1 promoter is induced by NSF.

The authors then investigate the 8 transcription factors that were present in their list of genes required for NSF-mediated adapted growth. They note that Hsr1 was the only one of these transcription factors, indeed the only gene, that was a hit in their screen for NSF-mediated adapted growth and whose expression was induced upon NSF treatment. To see if the activity of the other transcription factors changed in response to NSF treatment, the authors then gathered ChIP-seq data using 6 of these transcription factors as targets for IP. They saw that for Hsr1 and Php3, targets that had increased RNA-seq expression showed an increase in promoter occupancy while for Hsr1, Php3, Adn2, and Atf1, genes that had decreased RNA-seq expression showed a decrease in promoter activity.

Finally the authors attempt to identify the mode of action of NSF by generating a functionalized NSF with an alkyne tag (AlkNSF) which they then use as a probe to identify NSF binding partners. They first show that AlkNSF does allow bypass of NCR, although at 30-fold higher concentration. Also AlkNSF induces nrt1 expression in a dose dependent manner, although the expression saturates at a lower level and requires a much higher concentration for induction. They then look for proteins that co-purify with AlkNSF compared to a control that was pre-incubated with NSF which was expected to compete off AlkNSF. The only significant protein they saw was Ayr1, which was not identified in their screen and which did not abrogate NSF bypass of NCR when deleted independantly. They saw that Ayr1 deletion actually increases the response of nrt1 and mei2 targets to NSF, and speculate that Ayr1 metabolises NSF and reduces the cell's ability to respond to NSF to bypass NCR.

They then repeat the affinity purification / mass spec protocol in an Ayr1 delete cells to identify other interaction partners, this time incubating with a higher concentration of NSF, and also comparing to an experiment using Alkeyne Oleic Acid as a control for non-specific binding. The top two specific hits from this assay are Hmt2 and Gst3. NSF was still able to rescue NCR in gst3 deletes, indicating that it was not relevant for the phenotype. Cells lacking hmt2 did not grow in EMM, but did grow in YES when not supplemented with ammonium and when supplemented with ammonium did not grow, and addition of NSF did not rescue growth. They also see that nrt1 and mei2 gene induction in response to NSF is abolished when hmt2 is deleted. They then argue that hmt2, a sulfide:quinone oxidoreductase localized in the inner membrane of mitochondria is a direct target of NSF that triggers a switch to respiratory metabolism and allows bypass of NCR.

Below are comments that I think ought to be addressed prior to publication (Major comments)

1. In line 70, the authors state that "S. pombe cells rely on their own BCAA synthesis to sustain growth" when grown alongside Leucine when ammonium is supplied in the media. If prototrophs can inhibit NCR via NSFs in neighboring auxotrophic cells on the same plate, couldn't they also inhibit NCR within their own colony? How do we know that prototrophic cells grown in high quality nitrogen sources along with, say leucine, are not taking up leucine? The fact that leucine auxotrophs cannot grow in high quality nitrogen sources when leucine is present does not imply that wild type cells must use be synthesizing BCAAs rather than importing them. In a recent paper (Kamrad et al Nat. Microbiol. 2023, https://www.nature.com/articles/s41564-022-01304-8), it was shown that S. cerevisiae cells grown in lysine and in high concentrations of ammonium uptake lysine rather than synthesize it as lysine concentrations in the media are increased. I am aware via unpublished results that this is the case for Leucine as well. I would be surprised if the same isn't true in S. pombe. The authors should caveat or remove this assertion.
2. It is important for the authors to put their observation linking respiration to rescue from NCR in context with findings from a closely related study (Chiu et al 2022) which included some authors from this manuscript and which the authors cite. In that paper, it was shown that the siderefore ferrichrome can also rescue NCR in fission yeast. That paper stated "It is likely that ferrichrome increased mitochondrial activity, which enabled efficient utilization of glucose downstream of the glycolytic pathway" based on experiments in different concentrations of glucose. This evidence seems to support the link between respiration and rescue from NCR proposed by the authors of this manuscript. The authors should acknowledge this closely related and earlier work as it strengthen's the case they are trying to make. They could even test if ferrichrome addition makes cells sensitive to antimycin A (as in fig 1E), but that extra experiment would be optional in my opinion.
3. In figure 1B for the second screen I do not understand what the photos represent. For the photos, two rows are meant to have no NH4 and also no NSF and the label on that image makes no mention of Leucine supplementation. In the diagram there are two rows that have NH4 and leucine and one row that has no NH4 but does have leucine. I assume the diagram is correct and the labels on the images are incorrect.
4. It would be important for the authors to put their observation linking respiration to rescue from NCR in context with findings from Chiu et al 2022 which the authors cite. In that paper, it was shown that the siderefore Ferrichrome can also rescue NCR in fission yeast which the authors site which found that a siderephore rescues NCR. Also the authors of that paper stated "It is likely that ferrichrome increased mitochondrial activity, which enabled efficient utilization of glucose downstream of the glycolytic pathway." based on experiments in different concentrations of glucose. This evidence seems to support the link between respiration and rescue from NCR proposed by the authors of this manuscript.
5. In line 133. The authors state that the 29 mutants that didn't grow under Leucine supplementation either without NH4CL or with NH4Cl whether or not NSF was present were "related to EMM Growth, leucine uptake, or utilization of ammonium as the sole nitrogen source." The first two make sense, but I can't see why a a strain with deletion of a gene related to utilization of ammonium as a sole nitrogen source wouldn't grow when supplemented with leucine. In fact for all the leucine auxotrophs in the screen, if one was to try to grow them with ammonium as the sole nitrogen source they would not grow, so it isn't clear that this screen can identify genes responsible for utilization of ammonium as a sole nitrogen source. The authors should clarify or remove this point.
6. 203 strains are important for avoidance of NCR (because in the presence of Ammonium and Leucine, as well as a WT strain, they cannot grow). Of these 57 strains can't grow in the presence of a WT strain but they can grow in the presence of NSF. The authors conclude in line 138 that these strains are "likely to respond to a transmissible signal that is different from NSF". This is confusing because deletion of these genes still does allow cells to respond to NSF, however when these cells are growing in the presence of wild type cells (which in their model are releasing NSF), the cells don't grow. I am confused about the nature of the transmissible signal that the authors suggest. It would appear that when these genes are deleted and grown next to a wild type cell which sends the alternative signal and the NSF, the other transmissible signal would inhibits the ability of NSF to release NCR (as NSF can still rescue the gene). It is not clear how the other transmissible signal would work when the gene is present as it is clearly not necessary to rescue growth.

A simpler explanation might be that there was contamination in the second screen, or that there was a threshold effect - perhaps in the first screen the strains grew just below a threshold and in the second screen it grew just above that level.

The authors should clarify their interpretation for these strains, and acknowledge any alternative technical explanations.<br /> 7. The authors' efforts to removed confounding effects that might stem from additional auxotrophic alleles made the screen more convincing. However, Fig 1E, 1F, 5B, and 5E were done with EMM+Leu+Ade+Ura, while the initial strain was just done in the presence of additional Leucine. It is unclear why this was done from the text and captions, but I assume it was because they used a strain that was ade- and ura- in addition to being leu-. Given that they had strains without these additional mutations, this seems like a strange choice. The authors should acknowledge that there are possible confounding effects of adding adenine and uracil to the media, and, if they did have additional metabolic deletions, acknowledge that that could possibly be confounding.<br /> 8. Fig 1E, it appears that cells can grow without NSF in the presence of ammonium and additional amino acids after 10 days (although NSF is required for growth at 5 days). This is not a problem for the screen as that was taken at 5-6 days, but it appears as though NSF does not rescue growth so much as speed it up. The authors should acknowledge this when describing the phenotype. It also argues for a quantitative time course growth experiment to compare growth over the course of 10 days with and without NSF, although this would not be necessary to the paper's main argument.<br /> 9. In line 191 and 192, the authors suggest that the "downregulation of flocculation/adhesion related genes by NSF could serve to avoid undesirable mating during growth". If this is the case, I don't understand why mating genes and cellular fusion genes would be upregulated. What do the authors mean by undesirable mating? Wouldn't flocculation increase desirable mating as well? If all mating is undesirable, wouldn't upregulation of mating and cellular fusion genes be detrimental? 10. The authors mention that trehalose is an antioxidant, for which they reference Malecki 2019, however that paper shows no direct evidence of trehalose functioning as an antioxidant under respiratory conditions. It only shows that some trehalose synthesis genes are upregulated when cells are grown under glucose. The authors should identify primary literature to back this statement up, or soften the wording. Also trehalose is known to be a storage metabolite (which is mentioned in Malicki et al 2019, but not in this manuscript). In fact work in budding yeast has show that trehalose can be a shared metabolite that can be produced by respiring cells and used as a fermentable carbon source in communities of budding yeast cells that consist of fermenting and non-fermenting cells (Varahan et al, eLife 2019 https://doi.org/10.7554/eLife.46735). It seems that this role should be considered as an alternative explanation for the induction of trehalose in respiratory cells.<br /> 11. Line 208: The stimulatory effect of NSF on NRT1 decreased with cell density, thus cell density is likely to be an important factor in terms of gene expression. The methods section, text and figure captions do not mention the density at which cells were inoculated/harvested for RNA-seq and other experiments. If that density was more than OD 0.1, then this would be inconsistent with the measurements from Fig 3. Also in fig 3D, The culture density is not mentioned in the figure or the caption, even though the text suggests that for that experiment cells were grown at low density (Lines 212-213). The authors should provide information on density for their experiments in order for them to be reproducible, as they show it is a key factor. 12. In suggesting a name for NRT1 (NSF-responsive amino acid transporter 1), the authors assume that the gene has a role in amino acid transmembrane transport, but they have no experiments showing this phenotype. They mention that it is Inferred from homology with other amino acid transporters. I presume this name has already been approved by Pombase and is not provisional, but it seems that including phenotypes inferred from homology, rather than from experiments is unwise. Do the authors have any other direct evidence that this is a bona fide Amino Acid Transporter? Perhaps a name like "NSF-responsive gene" would be more appropriate.

Related to this, it appears that the expression level of Nrt1 may be very low (see Fig S2B in which the scale of the RNA-seq track is very small [-1,1] and the amount of expression is very small even when NSF is added). Looking at Fig 2A, the total transcript abundance did not appear to be very low in terms of counts per million (over 100) is this a discrepancy in fig S2B? Perhaps the large fold change is the result of counts very close to zero in the control condition? Also in Fig 3 the nrt1 expression levels did not appear to be especially low and they appeared repeatable. Is the RNA-seq data shown in fig S2B for nrt1 a fluke or am I misinterpreting it? <br /> 13. To show that their Chip-seq worked, the authors showed specific examples of Chip-seq reads for target genes Line 240, "Previously determined target genes of these TFs were significantly enriched in our data set, demonstrating that the experiment has worked (Figure S2A)." Is the significance here, the threshold from fig S2B? If so that threshold should be clearly stated here in the text. If it is the fact that asn1 shows up as "Fil1 bound" is strange as there are no genes that had significant changes in ChIP-seq signals for fig S2B. If there is another threshold the authors should describe it. While some of the examples they showed were convincing (e.g. php3-flag for the php3 regulated gene gln1 and the increased reads for srw1 for the reb1 target srw1), there were some targets that didn't seem to be especially enriched for their designated transcription factor. For example, the gene trx1 which was identified as an Hsr1 binding target had some binding from Hsr1, but more from Php3 and equivalent amounts for many of the other transcription factors. A clear description of how genes are chosen to be significant in the text, alongside references/selection criteria the authors used to select the specific genes shown should be provided to improve reproducability. <br /> 14. In lines 244-246 the authors state that "These differences in TF occupancy were positively correlated with target gene expression changes. That is, individual genes that were upregulated by NSF tended to be more strongly bound by the TFs, whereas downregulated genes were less occupied by the respective TFs (Figure 4A)." This is far from a general trend. The trend is not there for reb1 and fil1. In fact fil1 looks to the eye like it shows a decrease in occupancy for genes with increased expression, and I worry that the authors did a one sided test for significance that would have missed this, although the variability of the genes that don't change in this case is very high, so there could be no significant effect. The authors elaborate on some of the detail in following statements, but they should soften or remove this statement.

Related to this, in line 254, the authors state: "These results imply that NSF exposure rewires the recipient cell's transcriptional program, for which the TFs Atf1, Adn2, Adn3, Fil1, Hsr1, Php3, Php5, and Reb1 are indispensable (Table S3)." While I am convinced from the RNA-seq evidence and some of the chip-seq evidence that NSF exposure rewires cell's transcriptional program, I am not convinced that the 8 transcription factors they mention are indespensable for rewiring the transcriptional program. While they may be indespensible for the phenotype itself, Reb1, and Fil1 show no no siginificant enrichment in occupancy of upregulated or downregulated targets (Fig 4A) and, along with Atf1, Reb1, and Fil1, have very few genes in which ocupancy is changed significantly (Fig S2B), while no chip-seq experiments were shown for Php5 and Adn3.

The more specific summary of the data (Lines 250-253) from Fig S2B describing how hsr1 and adn2 have the strongest effects of the transcription factors required for NSF-mediated NCR bypass is a much stronger message for this section. 15. In line 335, the authors state that "in contrast to other communication systems, NSF does not induce noticeable changes in S. pombe's morphology", referring to changins in mating, filamentation, and bacterial biofilm formation. However they do show very clearly that NSF does cause a large decrease in expression in flocculation/adhesion genes. The fact that they do not see a change in morphology is likely due to the fact that the lab strain in the conditions used for this assay do not flocculate. We have recently identified conditions and strains which do exhibit flocculation in this preprint [https://www.biorxiv.org/content/10.1101/2023.12.15.571870v2]. It is likely that if they had a strain and conditions that did flocculate addition of NSF would break up flocculation and thus change the morphology based on their evidence. The authors should remove or caveat this point.<br /> 16. Line 270 Fig 5B: The concentration of NH4Cl listed in the text (374mM) does not match the concentration shown on the figure (748mM). I assume this is a typo but it should be corrected prior to publication.

Also I have several minor comments to help improve the manuscript.

m1: Lines 66-70- state that "uptake of the branched-chain amino acids (BCAA) isoleucine (Ile), leucine (Leu), and valine (Val) is suppressed in the presence of high-quality nitrogen sources such as ammonium or glutamate, because the expression of transporters or permeases that are needed for the uptake of poorer nitrogen sources are down regulated (Zhang et al, 2018)." This reference is for S. cerevisiae and is a review. The authors should cite original results in S. pombe if possible, and if that is not available, alert the reader that this result is from a different species.

m2: It is unclear from the methods section how the images taken for the screens were analyzed. Were they analayzed and scored by hand, or using custom image analysis software. Either way, when publishing the authors should publish the scores for each deletion mutant in their screen. If there was custom image analysis, the authors should mention in their methods the cutoffs which they used to score growth, and consider plotting the data as a supplement so readers can get a sense of how sensitive the screen was.

m3: The authors identify 137 mutants that did not require NSF signaling to bypass NCR and claimed these genes were required for NCR. It would be helpful and give more confidence in this screen to demonstrate the extent to which the genes identified in this study overlap with any previous genes required for NCR, and whether there was any GO-term enrichment in this set.

m4: It would be interesting if the authors could speculate a bit in their discussion on why mitochondrial respiration counteracts NCR. Is there something about cells undergoing respiration that would make it easier for them to use BCAAs than to produce them, or conversely something about fermenting cells that makes it easier for them to produce BCAAs rather than importing them?

m5: It is unclear why Figure 1F has 'MP biomedicals TM' listed in the figure. It doesn't seem to be listed in the caption or the methods. Is this different media than in other experiments? If so, the authors should add that information to the methods or the caption.

m6: In Line 160, positively influenced is strange wording, do the authors mean "induced"?

m7: In the section on gene expression change upon exposure to NSF, the authors use a + after each gene name. My understanding is that that notation is meant to refer to strains with the wild type genotype of that gene, and not the gene itself. Shouldn't the gene be italicised in lower case to represent the gene? See: Lera-Ramirez et al 2023 https://doi.org/10.1093/genetics/iyad143.

m8: In Fig 2A, genes are displayed on a plot that depicts level vs log2FC, but a comparison between the fold change and p-value would be more useful, and I believe DESeq2 should provide an adjusted p-value for these genes. A related issue is that it appears as though there were no biological replicates, though there was data gathered at different time points. In these genome wide experiments, replicates can give confidence to data and help distinguish true change from intrinsic variability of expression in specific genes. Though the authors did qPCR to validate specific results, it would have improved the quality of their systems-level data to have replicates for these and other key experiments (Chip-seq, affinity purification and even the screen).

m9: Supp Fig S1: To show that similar gene expression profiles exist for other time points, it would be more convincing to show Log fold change 2h vs 4h and 2h vs 6h and show correlation, or else to make a heat map with all genes to see that genes that go up in one condition go up in the other conditions. It is not clear if the red and blue colors are defined for the 2h dataset and then mapped onto the 4 and 6h dataset, or if they are independently assigned for each plot.

m10: Mbx2 is a key transcription factor related to flocculation and adhesion genes, and its expression is correlated with expression of its targets. If this transcription factor's expression levels decreased in response to NSF, that might strengthen and help explain the decrease in expression the authors observe in flocculation/adhesion genes when cells encounter NSF. If it it does not change, it might also be interesting for readers interested in these phenotypes.

m11: In Fig 3D, The notation for the Ammonium concentrations for EMM and YES are inconcistent (+ vs parentheses), also the units (mM from the caption) are not on the figure, but the abbreviation "N" is which is confusing and inconsistent with the other plots in which NH4CL is not abbreviated. Additionally, the caption lists additional nutrients in the media for the EMM conditions (Leu, Ade, Ura) which ought to also be listed.

m12: In lines 233-235, the authors say "One possibility is that they remain bound to their target genes but become activated or deactivated by NSF directly, or posttranslational modification, such as phosphorylation in the case of Atf1". I don't think the authors intend this, but this sentence could be taken to mean that Atf1 has been shown to be phorphorylated by NSF in the reference they site. I think the authors should clarify, i.e. by saying "..such as phophorylation which is known to regulate activity of Aft1 in response to oxidative and osmotic stress [Lawrence et al 2009]".

m13: In Fig 4B and Fig S2A, there are grey and colored tracks for the chip-seq (- and + NSF), but they are very difficult to see. If grey is in front it is hard to tell how close the colored peak wehn the colored peak is lower. For example, grey is in front for pex7 while color is in front for yhb1. Could the authors add some transparancy so that the data for both conditions could be seen at once? Also there is little information on the control. My assumption for the input(ChIP) sample was that it was cross-linked and sonicated but not immunoprecipitated, but it is not clear what conditions it was in. I would assume it was done without NSF treatment in WT cells, but those details should be added in the caption or methods. In particular, in the input there is a large spike for Gsf2. Do the authors have any explanation for this and does it have anything to do with that gene's NSF responsiveness?

m14: The authors might consider putting something like Fig S2B (or even a corresponding volcano plot) as a main figure for Fig 4 in addition to the other two panels, as the individual examples from fig 4B are nice to see, but do not give a broad overview of the data.

m15: In line 348, the wording "Would score" might be better replaced by "would be identified."

Significance

Assessment:

In general I find the authors arguments compelling and their experiments convincing. The initial and follow on screens were well designed and the authors linked respiration and the action of NSF in a convincing way. The analysis of RNA-seq data was also convincing, especially regarding the decreased expression of flocculation and adhesion genes, and the follow up of specific targets gives confidence in the data (though see Major point 12 below regarding the naming and expression levels of nsf1). The identification of hmt2 as a functional target of NSF was compelling and rigorous, and the authors offer an interesting hypothesis to connect this to respiration that could form the basis of future studies.

At times I thought that some of the interpretation of the results was hard to follow, poorly worded, or off the mark (see comments below). The presentation of the CHiP seq data also felt incomplete, though the influence of Hsr1 and Adn2 on expression of NSF1 targets was convincing. The genome wide assays (RNA-seq, CHiP seq, screen and pull-down/mass spec) could have done with replicates which would have improved statistics and reliability of the results presented for those experiments, although for key messages, the authors followed up with convincing targeted experiments.

The study represents an advance on recent work in NCR in fission yeast in linking this with the broad metabolic switch between fermentation and respiration, and in that sense makes this of interest to a broader swathe of the microbiology community, outside those interested in metabolic regulation in microbes. In addition to being of interest to applied researchers interested in producing metabolites with yeast and other microbes, the link to cell signaling and, via flocculation and adhesion genes, to microbial multicellular-like phenotypes would make this work of interest to those interested in microbial communities.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

This paper uses the model system Schizosaccharomyces pombe to investigate how the oxylipin nitrogen signaling factor functions to send signals and adapt the metabolism upon a change in nutrients in the environment. Combining genome-wide screens, RNA-sequencing and chemical biology, the authors find that the nitrogen signaling factor triggers a change from fermentation to a respiratory metabolism, through a direct interaction with a mitochondrial oxidoreductase Hmt2.

Major comments:

Overall, the manuscript lacks readability and coherence. Quite a lot of genes/TFs and proteins are mentioned, it is difficult to find a coherent story and clear overview and connection between these subparts. The manuscript would benefit from a general proposed scheme/working mechanism in the discussion and streamlining the results and data into a single biological storyline.

Several statements or results are not sufficiently clear, elaborated or nuanced. The paper would benefit from more explanation and discussion.

In the discussion section, the authors are not consistently referencing figures.

179-181: 'GO enrichment analysis of the 92 downregulated genes' but on line 167, it is '74 downregulated genes' that are mentioned. It is unclear where this difference in number of downregulated genes comes from. Similarly, for the upregulated genes. '156 genes' are mentioned on line 181 but only '98' on line 167.

189: The statement that the downregulation of flocculation could serve to avoid mating, though sounding logical, is undermined by the finding that mating-related processes are upregulated in the experiment. I find this statement rather speculative

247-249: The statement is too broad, the effects are visible for maybe 3 TFs, the others don't seem to make a difference in occupancy. Also, why are these two highlighted genes of importance (pex7+, yhb1+), this is the first and last time they are mentioned?

254-256: The statement that these TFs are indispensable may be too strong. Right before, the authors showed that most of these TFs don't change occupancy (and especially Fil1 and Reb1 do not show a correlation with up- and down-regulated genes, nor does Fil1 in FigS2B show a changed ChIP-seq signal).

365: 'independently of the carbon source'. As far as we can see, all experiments were performed using glucose as the carbon source, so this statement seems too strong as there is no clear proof for this. This could be an easy extra experiment to perform these tests on media with other carbon sources than glucose?

Fig1E: It is not clear if the experiment was performed with the Wild-Type or a deletion strain. In the case of the WT, colonies grew in the media not containing NSF but in Fig5E and Fig5I, the WT did grow in the media not containing NSF. It could be more relevant to plate out 1 colony like in the second screen. Thus, unless different strains were used for both experiments, the results seem inconsistent with each other, which is not mentioned in the manuscript.

Fig1E, Fig5B, Fig5E, and Fig5I: For these experiments, different nitrogen concentrations were used depending on the media, but this has not been addressed/mentioned in the manuscript.

Minor comments:

53-57: I would like more elaboration on why CCR and NCR are important for virulence of human pathogens or relevant for industrial applications, and link back to this in the discussion. Otherwise, it is superfluous to include this in the introduction.

108: Does having a h- library have any impact on the outcome compared to the original h+ library?

170: Why would only one of the 117 NSF-linked genes change expression in the RNA-sequencing experiment? Any explanation as to why the expression remains unchanged for the 116 other NSF-linked genes?

212: Please elaborate the discussion of these results. I understand the point that at low cell densities, the cells do not produce NSF as much, and thus adding NSF induces nrt1+. However, the added value of testing this in different media is unclear, especially when the results of strength of increase in nrt1+ show the opposite trend for the two different media between low and high nitrogen content.

216: Why was the ADH1+ promoter chosen as a 'negative control'?

284-288: Fig5E: To test whether AYR1 is indeed metabolizing NSF (and thus supporting this statement), an overexpression strain of AYR1 could be made to see if it grows on the EMM + NH4Cl without NSF added.

385: 'NSF would not strictly revoke NCR only, but also CCR': the authors should try to provide experimental evidence, citation(s), or clearly state it to be a hypothesis. This comment links back to the major comment on line 365.

405: typo: strains were validation, should be 'validated'

644: typo: '+' sign not in superscript

Fig1G: Could the differences in OCR be due to differences in growth rate or remaining glucose? It could for instance be that the culture in the control condition grew less fast, thus still having glucose and therefore still in fermentative metabolism. Showing or mentioning growth rates, nutrient concentrations could help to strengthen this finding.

Fig2A: The top two 'most' upregulated genes (nrt1 and mei2) were taken along for additional experiments. However, one gene with a significant upregulation labeled in red on the left seemingly shows stronger induction than the second gene (mei2). Why was this gene not taken along?

Fig2C: the x-axis label is not immediately clear to the reader.

Fig4B: typo: 'non treatmentt'

Significance

This manuscript advances our understanding of nitrogen signaling pathways and nitrogen catabolite repression in the model organism S. pombe. Specifically, it shows how a nitrogen signaling factor functions to send signals and adapt the metabolism upon a change in nutrients and reveals that this nitrogen signaling factor triggers a change from fermentation to a respiratory metabolism. These findings are relevant for the broad fields of applied microbiology, signal transduction and metabolic regulation. Relevant literature is appropriately cited, although the links with Crabtree repression in S. cerevisiae are perhaps not fully supported.

This manuscript was reviewed by experts with expertise in S. cerevisiae, Crabtree effect, respiration-fermentation balance, adaptation to changing environments.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

Summary

We have now reviewed the manuscript by the groups of Dr. Bühler and Dr Yashiroda entitled: "Nitrogen signaling factor triggers a respiration-like gene expression program". We have enjoyed the topic, the experiments and the science behind it all. The authors study here one part of the fission yeast 'quorum sensing'-like mechanism of counteracting NCR, mediated by the small molecule NSF: they identify pathways required to respond to NSF, and more specifically determine the mechanism by which NSF counteracts NCR: triggering respiration. This is a very interesting manuscript, with nicely executed experiments, and the topic is of great interest. Regarding the major comments, they are specific to the current data. The minor comments are questions raised in light of the present set of data, which should be appropriate for future research and future manuscripts.

Major comments

1. Consistency with the numbers of mutants/genes should be improved. Line 119: 203 genes, 206 in 4th datasheet of Table S1; line 139: 117 mutants but 119 analyzed for GO analysis (1st datasheet in Table S2); lines 179 and 181: where are these numbers (92 down and 156 up) coming from? (compare with 74 and 98 in line 167) (maybe they come from the merge of 2, 4 and 6 h, but it is not indicated).
2. Lists of genes up- and down-regulated from the RNA seq data should be provided. GO terms are not useful. Add supplementary table, please.
3. Comparing the transcriptomic response to that of Malecki et al 2020 in response to Antimycin A (EMBO Rep. 21:e50845) would be useful.
4. The optical densities and whether NCR has been induced has to be clearly specified in each experiment. For instance, RNA seq data. Line 165: for the transcriptome experiment, NSF is added or not to low density cells (not indicated in results, figure legend nor materials and methods). Should addition of NSF to wild-type strain en MM trigger the same transcriptomic changes?
5. Fig. 1G: Addition of NSF can enhance oxygen consumption at any cell density? And in prototrophs? And without NCR? Add in figure legend that this has been done at OD600 of 0.01.
6. Fig. 1 E and F: why 14 d after growth there is not growth at 2% glucose in panel F, but it is 10 d after in panel E? What is EMM Biomedical?
7. Fig. 2BC: Venn diagrams should be more useful to demonstrate overlap withn the Malecki data.
8. Fig. 4A: not very useful
9. Is Hsr1 required for some of the RNA seq changes upon NSF addition? Same with other TFs
10. Line 287: '...it is tempting to speculate that Ayr1 dampens adaptive responses by metabolizing NSF'. Calculating MEC for NSF in delta ayr1 and in cells over-expressing Ayr1 would be required to confirm this speculation. According to Pombase, cells lacking Ayr1 have their respiratory functions compromised (no growth in galactose, glycerol...), why is so? The opposite should be expected, if NSF-mediated respiration is enhanced in this background.
11. Regarding the two pull-down experiments, one to identify Ayr1 and the second Hmt2, why different negative controls are used? Is addition of NSF to WCE prior to pull-down also used in the second experiment (with delta ayr1 and AlkOle)?
12. The data regarding Hmt2 is very interesting. As for delta ayr1, delta hmt2 cells cannot grow in glycerol nor galactose according to Pombase. Is the result shown in Fig. 5J (lack of NSF-dependent activation of nrt1 and mei2 in delta hmt2) a consequence of the absence of the NSF receptor, or is it due to the lack of respiration of this background? Is delta hmt2 really auxotrophic for Cys? Why? In this background, H2S should be enhanced, and Cys and Met biosynthesis improved. In fact, in one manuscript these cells grow fine in SG minimal media (Mol Microbiol 01 42:29), while another report indicates they are auxotrophic for Cys (Genes to Cells 2016, 21:530).
13. M&M: regarding RNA isolation and sequencing: add info about OD of cultures, genotype (leu1-32?), growth media; also, number of replicates and filtering (fold-change used, Q value...)
14. M&M, ChIP seq: same as above. Also, MACS2 can be used for the unbiased identification of bona fide TF targets, by using a quantification tool reporting percentage of occupancy upstream the TSS (callpeak function).

Minor comments

1. Who triggers NCR? Analysis of 137 genes in Figure 1b.
2. Synthesis of NSF: how is it regulated, where does it come from?
3. NCR impairs import of BCAA. How are the aa importers such as Cat1 or Agp3 eliminated from the plasma membrane? Transporter internalization, degradation, transcriptional repression... And how does NSF block the NCR regarding aa uptake? Or aa usage?
4. How can enhanced respiration by NSF counteract all of the above? How can now leu1-32 cells grow?
5. Addition of NSF to any cell type would do the same, enhance respiratory rates? With or without previous NCR? Should this signaling molecule also drive different respiratory rates in a cell density-independent manner regarding glucose catabolite repression?

Significance

Within this manuscript, the authors study a cell-to-cell communication process, by which nitrogen catabolite repression can be counteracted by a small molecule called NSF. Specifically, the authors demonstrate here that NSF up-regulates respiratory metabolism as a mechanism to overcome the repression of amino acid internalization, which was blocked by excess nitrogen. This is a wonderful manuscript, with splendid data, on a very interesting topic.

Transparent Peer Review

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

The authors do not wish to provide a response at this time.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #3

Evidence, reproducibility and clarity

Summary:

In this manuscript the authors study a unique membraneless organelle (MLO) in the developing germline of the wasp (Nasonia vitripennis) and highlights the differences with polar granules, homologous membraneless assemblies formed in the Drosophila germline. They identify that in contrast to Drosophila, the wasp utilises an alternatively spliced isoform of the conserved RNA Helicase, Vasa, where the longer isoform harbours FG-repeats characteristic of nucleoporins. Additionally, they observe striking differences in the assembly of the perinucelar 'nuage' where the nuage components are heavily enriched at the anterior half as a mechanism to effectively silence transposon activity in the anterior nurse cells which are characterised by a high degree of DNA double strand breaks. In the course of oogenesis to embryogenesis, the authors observe that the oosome is dynamic and conserved germ plasm proteins (Osk, Vas, Tud) transition from diffuse distribution (in the oocytes) to a dense Tud shell surrounding there oosome filled with Osk-Vas granules (in the embryo). <br /> While the study provides insights into a novel germline condensate, there are some key questions which need to addressed to support publication.

Major Comments:

1. Nasonia expresses two distinct Vasa isoforms differing by 96 amino acids close to the N-terminus. The authors claim that the 96 amino acid insertion is FG-rich and intrinsically disordered providing experimental evidence with Circular Dichroism of the purified 96-amino acid fragment. However the amino acid sequence of the remaining N-terminal region upstream of the folded RecA domain is low complexity with an apparent over-representation of G, R, D, N as well as several FG repeats. Any computational disorder prediction tool (such as, IUPred, D2P2, etc) can be used to predict the sequence disorder of the entire N-terminus. Therefore, it is unclear why the authors claim that the 96 amino acid insertion exclusively confers special advantages and contribute to mesh-like properties to the oosome. Does the Long isoform provide specific advantages? This needs to be addressed. "Computational analysis predicted two alternatively spliced Nv-vas mRNAs, that should result in 92.3 kD and 82 kD proteins ": Please explain the analysis and tools used in methods. Fig. 1d : Please discuss the source and identity of the multiple non-specific bands of the RT-PCR experiment in the figure legends. Fig. 1f : For ease of readers, it is recommended to label the figure panels with stages as well as proteins probed.
2. Are there sequence similarities between the novel 57-residue NTD of Nv-Osk and the Drosophila NTD (138 amino acid long) present exclusively in Drosophila Long Oskar?
   • Fig. 2c : Consider including corresponding micrographs of the oocyte oosome imaged with AiryScan.
   • Fig 2d,e : Is Osk-GFP cytoplasmic or form peri-nuclear condensates? Is Vasa-mCherry nuclear or cytoplasmic or both? Please include DAPI channels and consider drawing outlines of the cell membrane.
   • Are these S2 cell images completely representative of the localisation patterns of Osk and Vas in the cells or do the authors only observe other phenotypic classes as well? If yes, please include the different classes observed with relevant statistics.
3. The differential assembly of the 'nuage' between anterior and posterior nurse cells is intriguing and well addressed. The higher degree of nuage assembly coincides with higher amount of DSBs in anterior nurse cells. Is this a cause or a consequence? Exposure to mutagens can induce DSBs uniformly in all nurse cells; will this lead to up regulation of nuage assembly uniformly in all nurse cells?
4. Dynamic sub-compartmentalization of Tudor to the periphery of the embryonic oosome suggests remodelling of the proteins components post-fertilization.
   • Can the authors perform live-imaging to track oosome dynamics from oogenesis to embryogenesis using FP tagged-germ plasm components? This would provide valuable insights into the assembly mechanism of this giant membraneless organelle.
   • Fig. 4 a, b: Line profiles required to show redistribution of Tudor. What do the three panels per condition indicate? What makes the authors conclude that the Tudor shell is "fibrillar" in nature? Is there any evidence?
5. Maternal mRNAs form homotypic clusters in Drosophila germ granules (Treck et al., 2015). Where do the maternal mRNAs localise in the oosome? In case they are recruited by Oskar (in line with Drosophila germ granules assembly model), are they diffusely distributed in the oocyte oosome and do they redistribute in the Osk-Vas granules that form at the core of the embryonic oosome? RNA in situ hybridization with a few maternal mRNAs can be done to understand how RNAs are distributed within this giant condensate.
6. The 'dense-shell liquid core' architecture of the oosome as proposed by the authors lacks any concrete proof. The migration of the nuclei (lines 342-345) can be facilitated by changes in physical properties of the Tudor shell in that particular embryonic stage, promoted in turn by key PTMs, for instance. Moreover, there is no evidence that the core oosome has liquid-like properties. In absence of live imaging and FRAP data, the 'dense-shell liquid core' architecture can not be addressed.

Minor Comments:

Line 40: "small spherical or amorphous cytoplasmic granules"; the terms "spherical" and "amorphous" have very different implications-one is for shape and the other indicates molecular organization. Consider re-phrasing.

Lines 55-57: mention embryonic stages as the Tud re-organization within oosome is a dynamic process.

Line 63: Is this really addressed in the manuscript?

Lines 95-97: What about Long Oskar in Drosophila? There is a 138 amino acid extension 5' of the LOTUS domain.

Line126: "proteome of the oosome". Proteome analysis would require isolation of the oosome followed by mass spectrometric identification of constituent proteins. Here the authors investigate the conserved germ plasm proteins and not the whole proteome. Please re-phrase accordingly.

Lines 308-310: "protein-free channels or cavities"- how do the authors know that they are protein free by probing for only three proteins? Consider re-phrasing.

Significance

The study describes a giant membraneless organelle in the developing wasp germline by developing an important set of tools and reagents necessary to study this novel organelle in greater depths in future. Experiments are well designed and validations of the tools developed are adequately carried out. However, the study is largely descriptive and the suggested experiments need to be performed to provide deeper insights into the significance of the work.

Considering the existing resources on assembly of germ granules by liquid-liquid phase separation, the 'oosome' represents another class of germ granules whose mechanism of assembly, dynamics and physical properties appear to be distinct from Drosophila polar granules as well as P-granules in C. elegans. Therefore, the work is significant in the fields of condensate biology and germline development as further studies focussing on the oosome can elucidate not only the molecular principles underlying oosome assembly but also address the plasticity in assembly of homologous MLOs across evolution.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

Summary

In this study, Kharel et al use endogenous antibodies to characterize the dynamics of germ granule components throughout development in the wasp Nasonia vitripennis. The authors observe several key differences in Nasonia germ granules as compared to the more well-studied germ granules of Drosophila. Kharel et al describe a novel isoform of the conserved granule component Vasa, which in Nasonia contains an FG domain. The authors use super-resolution microscopy to describe a core/shell architecture for the oosome, the single large germ granule that forms in the oocyte cytoplasm. Additionally, the authors find that a subset of Nasonia nurse cells have comparatively higher levels of perinuclear nuage which they hypothesize is related to high levels of DNA double strand breaks. The authors propose that these observed differences in conserved germ granule components may support the unique demands of germ granules in Nasonia.

Major comments

1. In many instances the authors make strong statements that are not directly supported by the presented data. For example, in the Abstract (line 62) the authors write "Our results point to the high degree of plasticity in the assembly of membraneless organelles, which adapt to specific developmental needs of different organisms, and suggest that novel molecular features of conserved proteins result in the unique architecture of the oosome in the wasp." Indeed, the authors have described several differences between germ granules in Drosophila versus Nasonia, but they have not presented data indicating that these differences are functionally relevant. They have also not shown that these differences result in the unique architecture of the oosome. The authors may of course speculate about the functional significance of their observations in the Discussion (emphasizing that certain statements are speculative), but should tone down or limit such statements in the Abstract and Results.

A few more examples of over interpreted data:

a. In line 111 of the Results the authors write: "Our data suggest that the Nv-Tud shell provides mechanical stability to contain a less dense oosome core during its migration in the early embryo." The authors have only observed an enrichment of Nv-Tud at the periphery of the oosome, which is not quantified. The authors have not performed experiments to test whether this enrichment is required for oosome integrity, or whether the core of the oosome is less dense.

b. In line 113 of the Results, the authors state: "Nasonia egg chambers have distinct subset of nurse cells in anterior that show evidence of double-strand DNA breaks and assemble higher amounts of perinuclear nuage than their posterior counterparts, indicating a higher demand for anterior nurse cells to silence transposable elements." Indeed, a subset of nurse cells have strikingly higher levels of nuage, and a subset have significantly higher levels of gH2Av staining. However, a link between high levels of nuage and a need to silence transposable elements seems speculative.

c. In line 421, the authors summarize their conclusions: "the assembly of the oosome...relies on the combination of highly conserved components...as well as a suite of novel features, including a novel Nv-Vas isoform, an unusual shell of Nv-Tud protein demarcating the edges of the oosome, and unusual distribution of nuage in the ovary." As the data presented in this study is largely descriptive, the authors have not directly tested whether any of these features are required for oosome assembly. 2. Throughout the study, the authors often show single western blots or representative images. To determine reproducibility, the authors should quantify their data whenever possible or indicate the number of independent experiments used to generate a figure. Sample size should be included in the Legends.

Some examples:

a. In Figure 1a, the authors show that the short Nv-Vas isoform is decreased in the embryo. How many times was this western performed, and is the short isoform always similarly decreased in the embryo?

b. In Figure 1d, the authors use RT-PCR to show that multiple vas RNA isoforms are present in ovaries, but only the long isoform is detected in embryos. How many times was the RT-PCR performed and how reproducible was this result? Also, there appears to be a third major RNA species present in ovaries, do the authors think this is relevant?

c. In Figure 2d, Nv-Osk forms granules when expressed in S2 cells. What fraction of S2 cells expressing Nv-Osk had granules?

d. In Figures 4 and 6, the authors should quantify core versus shell enrichment for Nv-Tud and other germ granule components (see Major comment 3 below). 3. A main finding of this study is that Nv-Tud forms a fibrillar shell by concentrating at the periphery of the oosome. The authors propose that this shell "fulfills the role of a membrane" (line 371) to protect the integrity of the oosome. This model is based on a handful of images, some of which are not entirely convincing. For example, Nv-Tud does not appear to form a shell in the middle image of Figure 4b. In order to strengthen their model, the authors should quantify the enrichment of different germ granule components in the core versus shell of Nasonia oosomes. Optionally, the authors could directly test a role for Tud in oosome integrity by observing the fate of the oosome core components following tud mutation or RNAi.

Minor comments

1. It might be helpful to add the protein structure for Drosophila Vasa for comparison in Figure 1c. Similarly, Drosophila Oskar could be added to Figure 2b.
2. The authors mis-reference Extended Data Fig 3 as Extended Data Fig 2 (starting in line 198).
3. In the Figures throughout, it would be helpful to label each panel with the antibodies used. Relatedly, in both the Figures and text the authors should always clarify whether they used antibodies recognizing the long Nv-Vas isoform or the entire Nv-Vas. Also make sure to include MWs for all blots.
4. For Extended Data Fig. 3d, I'm not sure that migration of an in vitro transcribed Osk "confirms" that the NTD is responsible for the higher molecular weight of Nasonia Osk. Is this experiment is needed?
5. In Figure 2e, the authors don't observe recruitment of Nv-Vas to Nv-Osk granules in S2 cells, leading to the proposal that "contrary to Drosophila, Nv-Osk does not directly recruit or associate with Nv-Vas in Nasonia" (line 231). It's possible that Nv-Osk is necessary but not sufficient to recruit Nv-Vas, and the authors might consider directly testing this by RNAi depletion of Osk in Nasonia.
6. The authors should mention and discuss the high level of gH2Av staining in the oocyte nucleus (Figure 3b). Has this been reported before?
7. The authors find strong gH2Av staining in anterior nurse cells, leading them to write in line 277: "indicating that the same population of nurse cells that assemble high amounts of nuage, shows high level of DSBs." While it is likely that this is the same population, without co-staining of germ granules and gH2Av in the same egg chmaber the authors cannot conclude that this is the same population of cells.
8. In Figure 3a, the authors note that Nv-Osk is produced in the cytoplasm of nurse cells, where it assembles into granules. It might be worthwhile to use osk RNAi as a control to make sure that the granular Osk signal in their IF is specific.
9. Unless I've missed it, the authors never reference Extended data Figures 6a and b in the text.
10. The authors write in line 362: "The spherical shape of these granule, point to their liquid characteristics and their formation inside the oosome core via liquid-liquid phase separation mechanism." Spherical shape alone is not sufficient to conclude liquid-like character or assembly via LLPS.

Referees cross-commenting

I might clarify that analyzing maternal RNA localization should be optional (Reviewer 3 Major Comment 5).

Significance

In this study Kharel et al use endogenous antibodies to observe the dynamics of conserved germ granule components in Nasonia. This approach allowed the authors to uncover several key differences between germ granules in Nasonia versus Drosophila. While these differences have the potential to be of interest to the specialized germ cell community, the data presented in this study are largely descriptive and not quantified. Therefore, the functional relevance of these differences remains uncertain.

The finding that a subset of nurse cells have high levels of nuage is quite striking. Furthermore, the authors write in line 405: "The occurrence of DSBs in a distinct population of nurse cells in any organisms has not been reported before to our knowledge." Therefore, this population of nurse cells may be unique to Nasonia and would be of interest to be explored further in future studies. The authors could use granule mutants to test their hypothesis that high levels of nuage are required to silence transposable elements.

A main goal of this study is to compare germ cell assembly between Nasonia and Drosophila, and thereby identify how unique features of Nasonia contribute to germ granule dynamics and function. Indeed, the oosome is quite unique as an enormous, solitary granule (though perhaps reminiscent of a Balbiani body?), and how it maintains integrity is an open question. The authors propose that Nv-Tud acts as a shell to stabilize the oosome, which would be remarkable for such a large germ granule. This finding could be of interest to a broader field of condensate researchers. Therefore, future studies should directly test whether Nv-Tud is required for oosome integrity. Finally, it would be helpful to the reader to more fully discuss direct comparisons between Nasonia and Drosophila germ granule components. For example, the authors should comment on what is known about isoforms of Drosophila Vasa, and whether Drosophila Vasa may have the potential to be alternatively spliced.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #1

Evidence, reproducibility and clarity

The manuscript entitled "Dynamic protein assembly and architecture of the large solitary membraneless organelle during germline development in the wasp Nasonia vitripennis" by Kharel et al. aims to examine structural features of the wasp oosome. Specifically, the authors look at the expression of the Vasa and Oskar protein isoforms, their accumulation in the oosome and nuage during oogenesis and early embryogenesis and aim to understand possible functional roles of the structure of these organelles in female germline development.

However, several conclusions in this paper are not fully supported by the data and some of the experiments need additional experimentation and controls. Below I am listing my concerns (listed in the order they appear in the manuscript):

Major concerns:

Lane 141: Extended Data Fig. 1b: Control demonstrating that CIP treatment worked. Lack of change could be due to the enzyme not working.

Lane 149: To demonstrate a differential splicing pattern the authors need to show PCR using primers spanning the intron that is retained or spliced out. These PCR products should also be quantified using qRT-PCR. The authors should explain the multitude of bands on the PCR gel - the gel presented by the authors is not convincing and shows issues with annealing efficiency (non-specificity) of primers.

Lane 168-169: "we detected Nv-Vas outside the oosome, distributed in the embryos' cytoplasm (Fig. 1g)." The authors should show a control of the embryo/oocyte stained the same way but without the primary antibody to evaluate background fluorescence in the staining. The images should be quantified and imaged/displayed using the same imaging and normalization parameters as the one shown in Fig. 1g.

Lane 171: Data showing the result of this mass spec experiment is not shown.3

Lane 181: Extended Data Fig. 2: This graph is hard to interpret. A control is missing that shows how a curve of the structured protein or an FG-repeat containing fragment of the nucleoporin would look like. As it stands now, it is just a curve that cannot be interpreted by someone who has never used CD spectroscopy to study IDRs.

Lane 202-203: "The Nv-osk mRNA 3' RACE mapping was consistent with the previously identified 3'-end of Nv-osk mRNA indicating that Nv-Oskis not extended at its C-terminus." Data for this is not shown in this manuscript.

Lane 204: "However, 5'-end mapping revealed that Nv-osk RNA starts at more than 800 bps upstream of previously predicted Nv-osk transcription start site (Extended Data Fig. 2c)." The authors show a schematic which is not data. Instead, they should show raw dat.

Lane 213-215: "in addition to major 51 kD band, a less intense ~100 kD band is detected, suggesting the formation of Osk dimers (Extended Data Fig. 3e) consistent with previous finding that LOTUS domain of Nv-Osk dimerizes" The authors are (presumably) running denaturing PAGE gels and they add beta-mercaptoethanol to the samples before they load them on the gel. Therefore, dimerization should not be detected on the gel. The band the authors see must be a contaminant.

Lane 231: "that, contrary to Drosophila, Nv-Osk does not directly recruit or associate with Nv-Vas in Nasonia (Fig. 2e)." The authors cannot make this conclusion. It is possible that Osk and Vasa interact directly but that one of the proteins requires a post-translational modification for this interaction and this modification does not happen in S2 cells on Nasonia proteins.

Lane 237: Lack of these specific amino acids in Nasonia proteins does not support the argument that Nasonia Osk and Vasa do not interact. Perhaps changes in amino acids in Vasa are compensated by changes in amino acids in Oskar.

Lane 243: "co-expressed in S2 cells, fail to form granules (Fig. 2e). Overall, our data suggest that while Nv- Osk has the intrinsic ability to condense into spherical granules..." If the expression of Vasa is not high enough, then Vasa will not phase separate in S2 cells. It is possible that both proteins phase separate but at different critical concentrations, which would explain the lack of granule formation of Vasa in S2 cells.

Lane 286: "transposon-related gene in Nasonia ovaries, we found no evidence that transposable elements are selectively upregulated in anterior nurse cells (Extended Data Fig. 5b), suggesting that high assembly of anterior nuage is needed to effectively silence transposable elements despite the prevalence of DSBs in anterior" This is an overstatement and incorrect interpretation. Since the staining is not mutually exclusive, the authors can conclude that there is no correlation between dsDNA breaks, nuage and transposon expression and that therefore nuage is not required for the regulation of transposon expression or dsDNA formation. Regardless, the data is correlative and existence of direct connection has not been tested. Lane 293: Control of IF staining without the primary antibody is missing to evaluate background fluorescence in the staining. The images should be quantified and imaged/displayed using the same imaging and normalization parameters. Also, the authors should do a western on oocytes that do not yet form germplasm to demonstrate Oskar protein expression in early oocytes.

Lane 311: There is no data demonstrating that the oosome has migrated - just two images of an oosome in embryos of different ages. The developmental changes (progression) of the embryo are also not evident. The data currently presented are not evidence of migration. The authors should avoid interpretations connected with migration using the data correctly presented.

Lane 314: the evidence that Tudor makes a shell is weak and only displayed in an image co-stained for Tud and Osk in Figure 4b. Co-staining of Tud and Vasa in the same panel does not display the shell convincingly. More data is needed to show a shell.

Lane 314: There is no data showing that the shell is fibrillar.

Lane 341-345: The authors overinterpret the data. Germ plasm is a cytoplasm and everything in a cell moves through the cytoplasm. This is not in itself evidence of a liquid nature of the oosome.

Lane 363: Round granule shape is consistent with LLPS but is not evidence for it. The authors should fix their statement.

Minor concerns:

Lane 95: Statement" In particular, we provide evidence for the presence of a novel N-terminal segment of Nasonia Oskar (Nv-Osk) adjacent to the conserved LOTUS domain that is absent in other insect Osk proteins." Is not true as written. The Drosophila melanogaster (D mel) Oskar has a N-terminal extension which forms the Long Oskar isoforms. In fact, what the authors report here is that Nosonia Oskar protein is a lot more similar to the D mel Oskar than previously reported by the authors, and that both organisms express long and short Oskar isoforms that with very similar protein structures. The authors should correct their statement.

Lane 134: Where is this data shown (Subsequently, we were able to confirm Nv-Vas identity of both IP-ed proteins with mass spectrometry.)- no mass spec data is reported in this manuscript

Lane 171: " ass spectrometry..." is missing an M.

Lane 200: Extended Data Fig 2b is referencing the wrong figure panel.

In general, western blots are missing molecular weight standards.

Lane 265: Tudor is not a component of nuage in D mel.

Significance

Overall, I found this manuscript interesting. It provided new insights in the expression of Vasa and Oskar and as well as new models of how these two proteins are regulated and how they accumulate in the oosome. Some of the features, like the newly identified N-terminal extension of Nasonia Oskar protein, appear to be shared with those of Drosophila melanogaster Oskar. This is an important finding because it indicates that mechanism by which Osk functions in the female germline might be conserved in insects.

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

Manuscript number: RC-2023-02270

Corresponding author(s): Usha Vijayraghavan

General Statements

We thank all three Reviewers for their thorough assessment of our manuscript and their constructive feedback and comments.

Point-by-point description of the revisions

This section is mandatory. *Please insert a point-by-point reply describing the revisions that were already carried out and included in the transferred manuscript. *

Reviewer #1

We are encouraged by the very positive comments made on the significance of our study that it provides convincing insights on alternative modes of nuclear positioning and division which is an important question in cell biology. We also took all possible suggestions to improve the interpretation of our results, have also added some newer data to address the constructive points raised by the reviewer.

Major comments:

1. A) I am concerned about the lethal phenotype caused by slu7 deprivation. Slu7 deficiency causes defective nuclear positioning at the bud in late G2. This phenotype per se should not cause defective mitosis, so slu7 deficiency may also be interfering with other aspects of mitosis which might indeed impinge on cell viability.

Response: Our data indeed show Slu7 knockdown has severe growth defect when grown on non-permissive media (YPD) where a two-fold difference in O.D. was seen by 12 hours (Supplementary figure 2.B).

We agree with the reviewer that defective mitosis, arises from several aspects of cell cycle including those in mitosis. The data we present show G2 arrest, small-budded cells with unsegregated nuclei and large-budded cells with segregated nuclei, all which do not progress through cell cycle phases and contribute to the severe growth defect. Further, GO enrichment analysis of deregulated pathways on knockdown of Slu7 support the above findings as various cell cycle related pathways are abnormal in their expression levels. In this study, we have focused on an in depth analysis of the role of Slu7 in a particular window and uncover how it controls nuclear position for progress G2-M phase cell cycle progression. The likely targets and mechanisms by which Slu7 regulates other phases of the cell cycle which needs similar other deeper investigations in future. Our detailed analysis of nuclear movement in Slu7 knockdown cells grown in YPD for 12 hours showed no nuclear movement (Supplementary figure 3B) which is the terminal phenotype. To examine events that lead to nuclear mispositioning phenotype we investigated the dividing slu7kd cells grown in non-permissive media for only 6 hours; under these conditions Slu7 protein is still detected at lower amount (Supplementary figure 1D). From the studies of nuclear position, mitotic spindle position and dynein distribution in mother and daughter cell, we propose that in the dividing cells, the nucleus does not experience enough force to move inside the daughter bud during mitosis. Further, we delineate the role of Slu7 in the splicing of transcripts for PAC1 encoding a protein whose homolog in S. cerevisiae has a proven role in nuclear migration. In live imaging of slu7kd cells that show nuclear segregation at the start of live imaging, new bud was not formed till the end of 60 minutes, implying that are arrested after transition to mitosis. We could speculate a role for Slu7 through regulation of genes involved in mitotic exit or cytokinesis.

1. B) Supp. Fig4 shows defective mitosis in TBZ, so TBZ may be exacerbating defective mitosis of slu7-deficient cells.