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 sincerely thank both reviewers for taking the time to examine our work, for the cogent points they have raised, and their constructive attitude aimed at improving our manuscript.

Reviewer #1

• Single-cell RNAseq (Fig. 1B) and in situ (Fig. 3D) results both indicate that SNHG7 is broadly expressed in multiple epidermal layers but more enriched in the spinous layer. Although most assays, such as colony formation and Ki67 staining, did not specifically examine the role of SNHG7 in the spinous layer, the raft culture experiment seemed to indicate specific reduction of the spinous layer (Fig. 3H), which was more prominent than basal defects. The authors should examine the defects more carefully in the raft culture system by using basal, spinous and granular markers. It is possible that SNHG7 functions to maintain limited cell proliferation while restrict premature differentiation. In addition, they should perform serial passage experiments to distinguish whether overexpression of SNGH7 can indeed confer self-renewal in long-term experiments.*

We have included staining of a range of epidermal markers in the raft cultures (ITGB1, K14, K10 and IVL) in the revised manuscript (Fig. S6B). We do not observe major changes in the distribution of any of the differentiation markers.

In terms of serial passage, we have cultured SNHG7-overexpressing or control cells for multiple passages until their growth capacity was exhausted. The SNHG7-overexpressing cells grew for approximately seven more passages than the control cells. We have added this information to the revised manuscript (Fig. S6L).

• The main proposed mechanism is the sequestration of miR34 by SNHG7. While miR34 is well known for its function in inhibiting cell proliferation, the ability of coding or noncoding RNA to sequestrate miRNAs is highly dependent on the stability and copy number of these RNAs. Since they have single-cell data with UMI information, they should estimate the copy number of SNHG7 in epithelial cell populations, and this could provide a range for the "buffering" capacity of SNHG7. They should also examine, ideally by in situ hybridization, the expression patterns of miR34 in human vs mouse skin. While miR34 expression can be induced by p53 activation, it is possible that its expression varies in different species. It'll be interesting to determine whether the lack of miR34 expression in mouse keratinocyte or mouse skin could explain the insensitivity of mouse keratinocytes to SNHG7. Finally, to further demonstrate the competition between SNHG7 and miR34 targets, they can use a heterologous luciferase reporter system with a canonical miR34 targeting site in the 3'UTR and quantify luciferase activities with or without SNHG7 (or SNHG7 mut34 variant). This assay could quantify the impact of SNHG7 on individual miR34 targets.*

We will include the analysis of the scRNAseq data to estimate the copy number of SNHG7 in the epidermal populations.

We will also perform in situ hybridization staining for miR-34 in human and mouse epidermis as well as mouse keratinocytes.

Finally, we will carry out the luciferase reporter experiments.

Reviewer #2

Major comments:

1) The premise of the study relies on the observation that SNHGs have low levels of sequence conservation. Indeed the authors aim to prove that biological functions can be identified even in the absence of evolutionary conservation. However, the extent of evolutionary conservation strongly depends on the phylogenetic scale at which it is analyzed. Here, the authors evaluate sequence conservation using PhastCons and PhyloP scores determined using alignments of human and 99 other vertebrate species. These scores reflect the extent of long-term sequence conservation. At this scale, only a small percentage of the human genome can be considered to be "conserved". It is thus not surprising that SNHG and other lncRNAs are not conserved at this scale, even if they carry some biological functions in the human genome. Here, it would be useful to redo the sequence conservation analyses using PhastCons and PhyloP scores computed on less distant species. Pre-computed scores exist for alignments containing human and other mammalian species, mainly primates (UCSC genome browser). It would also be good to provide comparisons of sequence conservation levels on the snoRNA genes and on the non-snoRNA parts of SNHGs. In addition to protein-coding genes, pseudogenes and lncRNAs, it would be good to add a perfectly neutral control for the sequence conservation in Figure 1C - for example, flanking intergenic regions for SNHGs. It might also be a good idea to analyze the GC content of SNHGs compared to other lncRNAs, since GC content can be correlated with sequence conservation levels, in particular in noncoding regions.

Importantly, the SNHG selected for detailed investigation (SNHG7) appears to be more conserved than the bulk of human lncRNAs, given that it is found in another primate and in mouse. It would be interesting to analyze in details what sequence features of this lncRNAs are conserved among species - for example, are the SNHG7 splice sites and promoter regions conserved? Are the snoRNA genes always located in the introns?

The reviewer raises very good points. We have added the conservation data of snoRNA genes (all or SNHG-resident, both significantly more conserved than SNHGs or lncRNA), and a “true neutral” control (we used the introns of 10,000 randomly sampled genes) to our PhastCons analysis (Fig.1C). We also added a conservation analysis of the promoters (defined as the 500 bp upstream of the transcription start site) of coding genes, pseudogenes, lncRNA and SNHGs (Fig, S1B). We have now performed all our conservation analyses using both PhastCons scores generated from the 100-vertebrate alignment and PhastCons scores generated from the 30-mammals (28 primate) alignment (Fig. S1E-G). We do not detect any marked difference between the two alignment sets. We have also added a comparison of the GC content in lncRNA, SNHGs, coding genes and introns (Fig.S1C).

Due to the nature of the PhyloP scores, the 30-mammal PhyloP track (phyloP30) would be unsuitable to detect additional conservation in the primate lineage using the thresholding analysis we employed in Fig. 1E. The PhyloP track gives base-wise p-values for conservation (positive values) or accelerated evolution (negative values). Using alignments of genomes that are overall more similar to each other (as in the 30-mammal alignment set) makes it more difficult to distinguish between conserved and neutrally evolving regions, because even segments that are not under constraint will look relatively similar due to the evolutionary proximity of the species in the set. For the same reason this alignment set is quite sensitive to accelerated evolution, as it contains many relatively similar genomes it the alignment.

This causes the PhyloP30 scores to be very asymmetrical around zero: the conservation (positive) scores never reach 2 (p-value of 0.01) in the whole track (not even coding regions of very well-conserved genes), while acceleration scores can reach very significant values, down to -20. Conversely, the PhyloP100 track (used in Fig.1E) is quite symmetrical around 0 and is thus better suited for the purposes of the analysis in Fig.1E, which are to detect both conserved and accelerated portions of SNHGs. We have however also inspected the PhyloP30 track manually and do not observe any clear evidence of presence of additional conserved elements in SNHG7. We have added all conservation tracks for SNHG7 to Fig. 3A.

While lncRNA orthologs can be identified by using a combination of sequence conservation, conserved synteny with surrounding genes and in some cases conserved gene structure, SNHG orthologs can additionally be identified by the embedded (conserved) intronic snoRNA sequence, which makes them easier to find even when the transcript sequence bears no similarity across species. The mouse SNHG7 sequence, for example, does not match with human SNHG7 even using the least stringent BLAST parameters. The monkey sequence is similar enough to match with the human in the 3’ of the gene, but the intron-exon structure of the 5’ is completely different. We agree with the reviewer’s assessment, however, when it comes to identification of SNHG orthologs in more evolutionary distant species, closer to the root of the vertebrate clade.

Regarding splice sites, they are often conserved among lncRNA in general (gene structure conservation occurs more frequently than sequence conservation, see Ulitsky, Nat Rev Genet, 2016). In the case of SNHG7 the structure of the gene appears conserved in the mouse (though this annotation has likely not been fully confirmed experimentally), and in the monkey based on genome alignments. However, our RACE experiments show that the 5’ end of SNHG7 in the monkey has a radically different splicing pattern when compared to human, so it is difficult to assess splicing conservation in the absence of full isoform characterization.

2) I am not perfectly convinced by the enrichment of miRNA target genes among the genes that are downregulated upon SNHG knockdown. The methods do not clearly explain how this enrichment is calculated. What is the background used for this enrichment analysis? In Figure 5C, we see that the genes predicted to be targeted by the top 10 microRNAs tend to have negative fold changes in the differential expression analysis (downregulation following knockdown). However, from Figure 5A it seems that the great majority of significantly differentially expressed (DE) genes have negative fold changes. How do the miRNA target genes differ from all other DE genes? What proportions of all predicted miRNA target genes (expressed in keratinocytes) are DE following knockdown, and how does this compare with the target genes of other miRNAs?

We have added a description of the statistical test used by MIENTURNET for the enrichment analysis to the methods section. More details can be found in the original publication. The significance of the enrichment is calculated by performing a hypergeometric test using as background the total number of miRNA-target interactions in the database and the total interactions the individual miRNA being tested engages in.

Figure 5C only includes the genes we used in our enrichment analysis (i.e. the significantly downregulated genes, not all DE genes) and it’s meant to show the extent of downregulation exhibited by the targets of the most significantly enriched miRNAs within this group of genes.

The reviewer is correct in pointing out that the imbalance between downregulated and upregulated genes (which we now further highlight in the plots in Fig. 5A-B) will tend to skew any group of genes towards having a relatively large number of downregulated genes. However, we found this bias to be particularly strong in the case of our candidate miRNAs. We now show this in volcano plots for validated targets of the candidate miRNAs and a control miRNA (Fig. S8E). In a similar way, when looking at the cumulative distributions of the fold changes of all predicted targets for a certain miRNA and comparing it to the fold change cumulative distribution of all other genes, our candidate miRNAs displayed a more pronounced shift towards downregulation than the control miR-21-5p (which we now added in Fig. S8F).

3) If keratinocyte RNA-seq data is available for other species (for example mouse), it would be interesting to test whether the high expression levels of SNHG7 and the other analyzed SNHGs are also conserved in the other species.

We will include the RNA-seq data, if available.

Minor comments:

1) The AD (atopic dermatitis) abbreviation should be explained the first time it is mentioned in the text.

We thank the reviewer for pointing out the missing abbreviation, we have now added it.

2) More details are needed regarding the MIENTURNET analyses in the methods and in the main text

We have added more details about the statistics involved to the methods (see above).

*3) Figure 5C, it is not clear how to interpret the color code for the boxplot. Does this represent the median or mean FDR of the target genes? Are only genes with FDRThe genes included in the enrichment analysis are all downregulated genes with an adjusted p-value (or FDR adjustment after the Wald test) adj”. The color scale in Fig 5C reports the significance of the enrichment for targets of the single miRNAs within the significantly downregulated genes list (FDR adjustment after the hypergeometric test), not the significance of the downregulation itself. FDR values are also used for all other enrichment analyses (GO terms, REACTOME Pathways). We apologise to the reviewer for the confusion, we have now modified the text and figures to make this clearer.

4) Figure 6D, I am not sure how to the D panel. Do the gray rectangles represent the exonic length of the SNHGs? Do the dots correspond to the positions of the miRNA target sites? Here, a more quantitative comparison with the extent of sequence conservation of miRNA binding sites in SNHGs, other lncRNAs and in protein-coding genes would be perhaps better suited.

The grey rectangles in Fig. 6D represent the total exonic length of the SNHGs (basically all exons are “stitched together” head to tail irrespective of the actual isoforms) and the dots represent the positions of the miRNA binding sites within this “maximum exonic coverage”. Since not all individual isoforms are analysed it is possible that some additional miRNA sites can be created at alternatively spliced junctions, however we would estimate the number of such sites to be small. We have added this caveat to the methods section.

The degree of conservation that is estimated by TargetScan to underlie a functional MRE in coding genes is taken into account in this analysis, as sites within SNHGs that pass this threshold are highlighted with yellow borders in the figure. We have now added a plot of the distribution of Branch length scores for MREs in SNHGs and the distribution of Branch length scores for MREs in a random sample of 250 Coding genes UTRs (Fig. S9E). A similar comparison for lncRNA is more challenging as the data is not readily available and is likely to be confounded by the nuclear localisation of a majority of lncRNA species.

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 submitted by M. Vietri Rudan and co-authors presents a functional analysis of a specific class of long non-coding RNA transcripts, namely the small nucleolar RNA host genes (SNHGs). These genes are defined by the fact that snoRNA molecules are embedded within their loci, often in the SNHGs introns. The rationale presented by the authors for studying this specific class of lncRNA genes is the fact that they display very low levels of evolutionary sequence conservation (even compared to the generally low conserved lncRNA transcripts) and that they are expressed at remarkably high levels (contrary to most lncRNA genes, which are very weakly expressed). The overarching goal of this study thus appears to be determining whether functional transcripts can exist in the absence of evolutionary conservation.

The authors study SNHGs in the human keratinocytes model system. They analyze the expression levels of SNHGs in skin affected by atopic dermatitis and psoriasis, compared to normal skin. They identify several SNHGs that are differentially expressed between diseased and normal skin, and which are also detected at strong expression levels in single cell RNA-seq assays of human keratinocytes. The authors use knockdown assays to investigate the potential roles of these SNHGs in keratinocytes and show that the knockdown of each of 5 selected SNHGs results in a loss of clonogenicity.

The last part of the manuscript focuses on the functional characterization of SNHG7, chosen because it is dysregulated in both atopic dermatitis and psoriasis and because its knockdown strongly affects clonogenicity. Additional assays showed that SNHG7 knockdown results in a reduced rate of proliferation and an increase in the fraction of differentiating cells. To ensure that the effects of the knockdown are not simply due to the absence of the snoRNA molecules embedded in the SNHG7 locus, the authors overexpressed the spliced form of SNHG7 (which lacks the snoRNA genes), successfully rescuing the cellular phenotype. They also verified that the knockdown does not affect the abundance of the corresponding snoRNA molecules.

To propose a potential mechanism for the involvement of SNHG7 in keratinocyte proliferation, the authors investigated its capacity to act as a miRNA "decoy". They searched for an enrichment of miRNA binding sites among the genes that are downregulated following SNHG7. They identified several miRNAs which are predicted to target SNHG7 as well as (a substantial fraction of) the genes that are downregulated following SNHG7 knockdown. The transfection of two of these miRNAs (miR-193-3p and miR-34-5p) has a negative effect on keratinocyte proliferation, supporting the authors's hypothesis that SNHG7 may act via "sponging" these miRNAs, thereby positively contributing to the control of the expression of the other miRNA target genes. Consistent with this hypothesis, the authors show that overexpression of a SNHG7 mutant sequence that lacks the miRNA binding sites is not able to rescue the knockdown phenotype.

However, this hypothesis is not supported (or at leat not further reinforced) by the evolutionary analysis performed by the authors. They were able to identify homologues for SNHG7 in another primate species (night monkey) and in the mouse. Transfection of the human SNHG7 sequence was able to increase clonogenicity in night monkey cells, but not in mouse cells. Given that miR-34 is also present in mouse and was previously shown to affect keratinocyte proliferation, it is not clear why the human SNHG7 sequence is not able to act as a miRNA sponge in this species. Likewise, only a slight effect on clonogenicity is observed upon SNHG7 transfection in night monkey. The authors conclude (correctly, in my view) that further investigations are needed to confirm the potential functions of SNHG7.

Overall, I find that this study is interesting and carefully conducted. Nevertheless, I have several comments that I hope can improve this manuscript.

Major comments:

1. The premise of the study relies on the observation that SNHGs have low levels of sequence conservation. Indeed the authors aim to prove that biological functions can be identified even in the absence of evolutionary conservation. However, the extent of evolutionary conservation strongly depends on the phylogenetic scale at which it is analyzed. Here, the authors evaluate sequence conservation using PhastCons and PhyloP scores determined using alignments of human and 99 other vertebrate species. These scores reflect the extent of long-term sequence conservation. At this scale, only a small percentage of the human genome can be considered to be "conserved". It is thus not surprising that SNHG and other lncRNAs are not conserved at this scale, even if they carry some biological functions in the human genome. Here, it would be useful to redo the sequence conservation analyses using PhastCons and PhyloP scores computed on less distant species. Pre-computed scores exist for alignments containing human and other mammalian species, mainly primates (UCSC genome browser). It would also be good to provide comparisons of sequence conservation levels on the snoRNA genes and on the non-snoRNA parts of SNHGs. In addition to protein-coding genes, pseudogenes and lncRNAs, it would be good to add a perfectly neutral control for the sequence conservation in Figure 1C - for example, flanking intergenic regions for SNHGs. It might also be a good idea to analyze the GC content of SNHGs compared to other lncRNAs, since GC content can be correlated with sequence conservation levels, in particular in noncoding regions.

Importantly, the SNHG selected for detailed investigation (SNHG7) appears to be more conserved than the bulk of human lncRNAs, given that it is found in another primate and in mouse. It would be interesting to analyze in details what sequence features of this lncRNAs are conserved among species - for example, are the SNHG7 splice sites and promoter regions conserved? Are the snoRNA genes always located in the introns? 2. I am not perfectly convinced by the enrichment of miRNA target genes among the genes that are downregulated upon SNHG knockdown. The methods do not clearly explain how this enrichment is calculated. What is the background used for this enrichment analysis? In Figure 5C, we see that the genes predicted to be targeted by the top 10 microRNAs tend to have negative fold changes in the differential expression analysis (downregulation following knockdown). However, from Figure 5A it seems that the great majority of significantly differentially expressed (DE) genes have negative fold changes. How do the miRNA target genes differ from all other DE genes? What proportions of all predicted miRNA target genes (expressed in keratinocytes) are DE following knockdown, and how does this compare with the target genes of other miRNAs? 3. If keratinocyte RNA-seq data is available for other species (for example mouse), it would be interesting to test whether the high expression levels of SNHG7 and the other analyzed SNHGs are also conserved in the other species.

Minor comments:

1. The AD (atopic dermatitis) abbreviation should be explained the first time it is mentioned in the text.
2. More details are needed regarding the MIENTURNET analyses in the methods and in the main text
3. Figure 5C, it is not clear how to interpret the color code for the boxplot. Does this represent the median or mean FDR of the target genes? Are only genes with FDR<5% included in this analysis?
4. Figure 6D, I am not sure how to the D panel. Do the gray rectangles represent the exonic length of the SNHGs? Do the dots correspond to the positions of the miRNA target sites? Here, a more quantitative comparison with the extent of sequence conservation of miRNA binding sites in SNHGs, other lncRNAs and in protein-coding genes would be perhaps better suited.

Significance

Overall, this study is highly relevant in the field of lncRNA functionality and evolution. It presents evidence for a potential involvement in the regulation of cell proliferation for SNHGs, a role that appears to be independent of the snoRNAs produced by these loci. This study reinforces the current body of work suggesting that lncRNAs and other noncoding transcripts can sometimes function as miRNA "decoy" targets. This study will be of interest for a specialized audience, oriented towards understanding lncRNA biological functions.

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

Vietri Rudan and colleagues present an interesting study to examine the function and mechanism of small nucleolar RNA Host Gene 7 (SNHG7) in human keratinocytes. The key findings are that a widely expressed SNHG7, a long noncoding RNA hosting two snoRNAs in the introns, promotes keratinocyte proliferation and inhibits differentiation likely by sequestrating miR34a from its targets. The experiments are generally well-executed with proper controls. Notably, they leveraged human, night monkey and mouse keratinocytes to reveal primate specific functions of SNHG7 in a miR34 dependent manner. I have a few comments and suggestions that should be addressed to further strengthen the study.

1. Single-cell RNAseq (Fig. 1B) and in situ (Fig. 3D) results both indicate that SNHG7 is broadly expressed in multiple epidermal layers but more enriched in the spinous layer. Although most assays, such as colony formation and Ki67 staining, did not specifically examine the role of SNHG7 in the spinous layer, the raft culture experiment seemed to indicate specific reduction of the spinous layer (Fig. 3H), which was more prominent than basal defects. The authors should examine the defects more carefully in the raft culture system by using basal, spinous and granular markers. It is possible that SNHG7 functions to maintain limited cell proliferation while restrict premature differentiation. In addition, they should perform serial passage experiments to distinguish whether overexpression of SNGH7 can indeed confer self-renewal in long-term experiments. Based on these results, they may need to refine their hypothesis/conclusion whether SNHG7 functions primarily on stem cell self-renewal or transiently maintain proliferation in transitioning cells.
2. The main proposed mechanism is the sequestration of miR34 by SNHG7. While miR34 is well known for its function in inhibiting cell proliferation, the ability of coding or noncoding RNA to sequestrate miRNAs is highly dependent on the stability and copy number of these RNAs. Since they have single-cell data with UMI information, they should estimate the copy number of SNHG7 in epithelial cell populations, and this could provide a range for the "buffering" capacity of SNHG7. They should also examine, ideally by in situ hybridization, the expression patterns of miR34 in human vs mouse skin. While miR34 expression can be induced by p53 activation, it is possible that its expression varies in different species. It'll be interesting to determine whether the lack of miR34 expression in mouse keratinocyte or mouse skin could explain the insensitivity of mouse keratinocytes to SNHG7. Finally, to further demonstrate the competition between SNHG7 and miR34 targets, they can use a heterologous luciferase reporter system with a canonical miR34 targeting site in the 3'UTR and quantify luciferase activities with or without SNHG7 (or SNHG7 mut34 variant). This assay could quantify the impact of SNHG7 on individual miR34 targets.

Significance

This study reveals a potential function and mechanism of primate specific noncoding RNA for its role in modulating gene expression and cellular functions in the skin. It provides a new paradigm for identifying molecular functions of poorly conserved RNAs.

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-2023-02281

Corresponding author(s): Maurizio Molinari

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 (Evidence, reproducibility and clarity (Required)):

In this manuscript from Fasana et al., the authors present data that investigates potential compensatory degradation pathways for misfolded glycoproteins in the ER - postulating that the ER-to-lysosome associated degradation (ERLAD) pathway becomes employed in the absence of a path for substrates to reach the ER-associated degradation (ERAD) mechanism. Using the classic ERAD substrate alpha1-antitrypsin NHK variant (NHK), the authors first demonstrate that pharmacologically preventing access of NHK to ERAD either with KIF (early) or PS-341 (late) elevates the number of LAMP-1 positive endolysosomes also immunoreactive for NHK (via HA), similar to what is observed for the ATZ variant that forms polymers in the ER (Fig 2). The authors next use shRNAs that silence essential ERAD factors (EDEM1, OS-9) involved in glycan recognition to demonstrate comparable enrichment of NHK in endolysosomes through genetic disruption (Fig 3). Next, the authors employ FAM134B-deficient MEFs to demonstrate the requirement for this ER-phagy receptor when ERAD is unavailable (Fig 4). Reconstituting FAM134B-/- MEFs treated with KIF/PS-341 + Baf, with a full length FAM134B rescue plasmid restores endolysosomal accumulation of NHK while a FAM134B-∆LIR does not, providing supporting evidence for substrate rerouting to ERLAD. Finally, the authors use knockouts of Atg7 and Atg13 to demonstrate dependence on LC3 lipidation and independence from macro-ERphagy (Fig 6), that points towards a pathway that is like that used to remove ATZ polymers. From these data, the authors conclude that ERLAD is increasingly engaged for substrate degradation when ERAD is impaired.

MAJOR COMMENTS 1. All assays rely on quantification of the NHK-HA substrates by microscopy. Would it be possible for the authors to also include biochemical analysis of NHK - potentially including data assessing its changing abundance and glycosylation state?

To consider this, and other comments, the new submission includes biochemical data (pulse-chase analyses) on NHK (new panels A-D in Fig. 2) and on BACE457delta, an additional ERAD substrate (new Fig. 6). Please also refer to Comment 3.

In Figure 3D, the knockdown of OS-9.1/2 is modest compared to that of EDEM1 (Fig 3A). Moreover, there is only data from single shRNAs presented. Could the authors please at least include another shRNA to confirm and demonstrate whether the targeting to ERLAD is accordingly scaled to loss of access to ERAD (based on the degree of OS-9 or EDEM1 remaining)?

__The reviewer is right. The phenotype (i.e., lysosomal delivery of NHK, Figs. 3B, 3C) is quite modest upon EDEM1 silencing. However, one has to consider that in contrast to OS9 lectins, EDEM1 is an enzyme, and residual protein may partially facilitate NHK de-mannosylation and access to the ERAD pathways and therefore reduce the ERLAD contribution for NHK clearance in these cells. Moreover, cells also express EDEM2 and 3 that may partially compensate the loss of EDEM1. __

While degradation is implied, it is not specifically demonstrated at any point in the manuscript. Perhaps the authors might include some demonstration of NHK stabilization in one of the figures via a translational shutoff or pulse-chase assay.

---

__In the new submission, we show biochemical analyses (pulse-chase) that reveal the decay of radiolabeled NHK (Fig. 2A, lanes 1-3) and BACE457delta (Fig. 6A, lanes 1-3), the inhibition by PS341 (lanes 4, 5) and by KIF (lanes 8, 9), and the intervention of lysosomal enzymes when ERAD is inhibited (lanes 6, 7 and 10, 11). Moreover, we confirm that the protein delivered to the endolysosome is eventually degraded by performing a Bafilomycin washout experiment (new Fig. 2J-2O). __

10-30% of NHK-HA positive endolysosomes are detected even with Baf alone (e.g. Fig 2E)? Does this mean that Baf impairs ERAD to some extent since or is it evidence for continuous ERLAD involvement when ERAD is intact? If so, how much is its contribution?

Pulse-chase analyses (new Fig. 2D) and published data show that BafA1 or chloroquine do not inhibit clearance of the ERAD substrates NHK and BACE457delta (e.g., Liu et al 1999, Molinari et al 2002, references in the manuscript). A basal level of endolysosomal delivery between the 20 and 30% as quantified with LysoQuant is observed in all experiments (Figs. 2I, 2O, 3C, 3F, 4C, 4K, 5H, 6G, 6O), which have been performed in 3 different cell lines (3T3, HEK293, MEF). We measure similar basal levels also when ER-phagy is monitored on quantification of lysosomal delivery of endogenous ER marker proteins (e.g., CNX), possibly to be ascribed to constitutive ER phagy that controls physiologic ER turnover.

---

An accounting of how much ERLAD is contributing to NHK degradation with or without ERAD impairment is not really present.. Effectively, how much degradation capacity is ERLAD making up? These would be interesting data to include if possible as they would speak to the "division of labour" for ER substrate degradation its potentially dynamic nature.

The biochemical analyses show the contribution of ERLAD on NHK (new Figs 2B, 2C, grey zones) and BACE457delta (new Figs. 6B,C, grey zones) clearance, when ERAD is dysfunctional.

MINOR COMMENTS 1. In Figure 4, an increase is observed for the rescue of FAM134B-/-MEFs with WT FAM134B that is 50% greater that of WT MEFs, suggesting that its availability might be rate limiting. Could the authors compare the relative levels of FAM134B for the WT and KO-rescue MEFs to address this possibility?

__The referee is right in assuming that FAM134B, expressed at low levels in these cells, is limiting. We now show the levels of endogenous FAM134B and of recombinant FAM134B in WB (new Fig. 4A). __

In Figures 1 and 6, the terms siOS9 and siEDEM1 are used but Figure 3 shows data from shRNAs and not siRNAs.

We apologize for the mistake. We have corrected this in the new Figures 1 and 7.

Samples from Figure 3 treated with Baf but this is not indicated in the figure or figure legend.

---

We have corrected this, thank you.

VCP/p97 inhibitors typically stabilize ERAD glycoprotein substrates better than proteasome inhibitors do. Is the same degree of endolysosomal targeting present ?

---

__For the convenience of the reviewer (we did not put these data in the new manuscript). In our experiments, the p97 inhibitor DBeQ is less efficient in deviating NHK to the endolysosomal degradative compartments, if compared with KIF (see below). At higher doses, DBeQ also inhibits other AAA-ATPases (e.g., VPS4, which plays a role in certain types of autophagy). This, or other cross-reactivities of DBeQ could explain the moderate capacity to activate ERLAD pathways as a response of ERAD inhibition, if compared with the phenotypes observed when ERAD is inhibited with KIF or PS341. __

Reviewer #1 (Significance (Required)):

Deconvolution of the different pathways taken by misfolded proteins to escape the ER is of great interest not only to the ER community but also represents consequences to consider for those interested in therapeutics involving UPS inhibition. While concise, this manuscript does a good job of trying to demonstrate the principal of substrate rerouting and the prioritisation of degradation pathways. Overall, the manuscript is well written, the experiments presented are performed to a sufficient standard, the data are lean but of good quality, and the appropriate statistical analyses have mostly been included where necessary and are described. The Methods and Materials is brief but describes the experiments that have been performed. The manuscript is brief in its results and would obviously benefit from additional complementary assays that would strengthen and broaden the authors arguments for rerouting. But too their credit, the authors do not grossly overstate their findings and merely present the culmination of a set of experiments to answer a single question - what happens to a misfolded glycoprotein substrate when ERAD is impaired. This is a key question with broad implications.

While their limited data clearly demonstrates an acquired dependence on ERLAD, one can't help but wonder how broadly these findings hold true, as only a single glycoprotein substrate example is used.

We have now added a complete set of experiments (imaging + biochemical to monitor clearance of the model polypeptides by pulse-chase analyses) performed with a second ERAD substrate (BACE457delta, Fig. 6). These data fully recapitulate the results obtained with NHK.

---

Moreover, it is not clear what percentage ERLAD contributes to overall NHK degradation (with or without ERAD) as the total NHK amount remaining is not assessed or measured.

---

Pulse-chase analyses (new Fig. 2D) and published data (e.g., Liu et al 1999, Molinari et al 2002, references in the manuscript) show that BafA1 or chloroquine do not inhibit clearance of the ERAD substrates NHK and BACE457delta. The biochemical analyses now show the contribution of ERLAD on NHK (new Figs 2B, 2C, grey zones) and BACE457delta (new Figs. 6B,C, grey zones) clearance, when ERAD is dysfunctional.

Nevertheless, the manuscript is an advancement of understanding of the fate of substrates unable to access ERAD and raises many future questions of interdependency between the ERAD and ERLAD pathways. The data just need a bit of shoring up.

Expertise - ERAD, UPS, protein quality control

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

The endoplasmic reticulum (ER) is a crucial site for protein synthesis and folding within the cell, and strict protein quality control is essential for maintaining ER homeostasis. In this context, ER-associated degradation (ERAD) and the unfolded protein response (UPR) play pivotal roles. Recent researches have highlighted the significance of ER-phagy in protein quality control. In this manuscript, the authors demonstrate the role of FAM134B in degrading misfolded proteins such as ATZ through the ER-phagy pathway when the ERAD pathway is obstructed. This work partially addresses a prominent issue in the field, unveiling the interconnections between different regulatory pathways in maintaining ER homeostasis.

Major issues: 1: In a multitude of experiments, the authors employed Bafilomycin A1 (BafA1) to block the fusion between autophagosomes and lysosomes, attempting to demonstrate that the clearance of misfolded proteins mediated by FAM134B is independent of autolysosomes. However, in Figure 4, the lack of rescue of FAM134B knockout by overexpressing FAM134B△LIR suggests a dependence on the interaction between FAM134B and LC3. The conclusions drawn before and after appear contradictory.

We apologize if our explanations were unclear. We have now modified the text and performed new experiments to clarify these issues.

__The inhibitor of the V-ATPase BafA1 is used here to inhibit the activity of lysosomal hydrolases and to accumulate undegraded material in the LAMP1-endolysosomes (note that these endolysosomes also display RAB7 at their limiting membrane) (Fregno et al 2018, Forrester et al 2019, Fregno et al 2021, …). __

__In Figs. 2A-2D, we now monitor the lack of NHK stabilization by cell exposure to BafA1 (Fig. 2D), which correlates with lack of accumulation of NHK in the LAMP1-positive compartment (e.g., Fig. 2F, 2J, and quantifications in 2I and 2O). The biochemical data also show that BafA1 stabilizes NHK in cells where ERAD has been inactivated with PS341 or KIF (Fig. 2A, lanes 6, 7, 10, 11 and grey zones in Figs. 2B and 2C), which correlates with accumulation of NHK in LAMP1-positive organelles (Figs. 2G, 2H, 2I, 2K, 2M, 2O). __

__In Figs. 2J-2O, we have now added panels showing that NHK clearance from the LAMP1-positive endolysosome lumen is restored upon BafA1 washout. __

Importantly, the involvement of the lipidation machinery, of the ER-phagy receptor FAM134B and of the LC3-binding function of FAM134B (the LIR), does not necessarily imply the involvement of autophagosomes in the process under investigation, as the comment by the referee seems to suggest. For example, both the clearance from the ER of ATZ polymers and of mutant forms of procollagen rely on the LC3 lipidation machinery and on the LC3-binding function of FAM134B, but ERLAD of ATZ polymers does not rely on autophagosomes intervention (new Fig. 1B, arrow 1 and Fregno et al 2018), whereas ERLAD of procollagen relies on intervention of autophagosomes (new Fig. 1B, arrow 2 and Forrester et al 2019).

2: Some Western blot data are insufficient to substantiate the author's conclusions. For instance, in Figure 5D, the ATG7 KO line is inadequately supported

The WB show____s the absence of ATG7 in the ATG7-KO cells (a well-established cell line generated in the lab of Masaaki Komatsu (____Komatsu M, et al. J Cell Biol 169: 425-434_) and used in many_ laboratories, including our lab in Fumagalli et al 2016, Fregno et al 2018, Fregno et al 2021, Loi et al 2019, Kucinska et al 2023). We agree with the reviewer that the anti-Atg7 shows cross-reactions. We have now added a WB showing the lack of LC3 lipidation in the Atg7-KO cells exposed to nutrient deprivation (new Fig. 5D).

3: The author employed Lamp1 antibody for lysosomal staining in cells and observed a significant abundance of lysosomes in some experiments, as depicted in Figure 2C, 2D, 4I, etc. Is the phenomenon of lysosomes extensively filling the entire cell a common occurrence? Is it indicative of a normal physiological state?

There may be variations depending on the cell type used for the experiments. In the new version of the manuscript, we now present imaging data for 3 cell lines (NIH 3T3 with stable expression of NHK and ATZ (Figs. 2E-2H), MEF (Figs. 2J-2N, 4, 5, 6) and HEK293 with transient expression of ERAD clients (Figs. 3).

Minor issues： 1: Some immunofluorescence experimental data are unclear. Please request the authors to replace these with more distinct images, as seen in Figure 3B and 3E.

---

We hope that the quality of the new images will be considered sufficient for publication.

2: Some expressions appear to be questionable. For instance, the necessity of utilizing endolysosomes requires clarification.

For the use of endolysosomes (lysosome would be incorrect in our opinion to indicate these LAMP1/RAB7-positive degradative organelles), we now refer to the papers by Bright et al ____Endolysosomes Are the Principal Intracellular Sites of Acid Hydrolase Activity_ Curr Biol 2016, and the original definition by Huotari and Helenius _Endosome maturation EMBO J 2011 (Introduction, page 2).

3: Some writing lacks precision, such as referring to FAM134B as FAM134.

__Corrected, thank you____ __ Reviewer #2 (Significance (Required)):

o General assessment: o Advance: provide an meaningful evidence that how two degradative pathways are coordinated in maintaining ER homeostasis. o Audience: cell biologist o Reviewer's expertise: autophagy, vesicle trafficking, organelle biolgy Reviewer #3 (Evidence, reproducibility and clarity (Required)):

In their study, Fasana and colleagues investigate protein quality control in the ER. Specifically, they test whether folding-incompetent proteins that are normally cleared by ER-associated degradation (ERAD) can also be targeted for degradation by direct vesicular transport from the ER to lysosomes in case ERAD is blocked. They show that blocking ERAD pharamacologically or genetically indeed leads to re-rerouting of an ERAD model substrate (the NHK variant of alpha-antitrypsin) to lysosomes and that this pathway requires the reticulon-like protein FAM134B, the ability of FAM134B to interact with the ubiquitin-like protein LC3 and the machinery for LC3 lipidation.

The paper is, for the most part, easy to follow. There are, however, a few minor issues and I think the authors could do more to connect their work with similar studies in the literature. Accordingly, I have some general and specific suggestions to make the manuscript more accessible for the reader.

General suggestions

1. To avoid confusion, it would be helpful to more clearly distinguish between vesicular transport to endolysosomes and autophagy. Previous work by the authors has defined a trafficking pathway from the ER to endolysosomes that appears to rely on conventional vesicle-mediated transport (Fregno et al, EMBO J 2018). This pathway delivers material from the ER lumen to the lumen of endolysosomes, which are both topologically equivalent to the extracellular space. Hence, this pathway is distinct from autophagy, which is the transport of cytoplasmic components to endolysosomes and thus the transport of material from intracellular to extracellular space. This distinction is particularly important as both vesicular ER-to-lysosome transport and autophagy of the ER involve LC3 and FAM134B, which is typically referred to as an ER-phagy receptor. To make this less confusing, it may be helpful to explain that FAM134B appears to be a multifunctional molecule that can function as a receptor for macroautophagy but also in the vesicular transport pathway studied here. In addition, it would be helpful to point out that LC3 appears to also have roles unrelated to autophagosome formation.

The reviewer is referring to the original definition of ERLAD to describe the mechanisms of clearance of ATZ polymers (Fregno et al 2018). The definition of ERLAD has now been expanded and is given, for example, in Klionsky DJ, et al (2021) Guidelines for the use and interpretation of assays for monitoring autophagy (4th edition). Autophagy 17: 1-382 and is explained in detail in our recent review Rudinskiy M, Molinari M (2023) ER-to-lysosome-associated degradation in a nutshell: mammalian, yeast, and plant ER-phagy as induced by misfolded proteins. Febs Letters: 1928-1945.

__Notably, the acronym ERAD for ER-associated degradation has originally been used to describe ____the proteasomal clearance from the ER of misfolded pro-alpha factor in a reconstituted yeast system in McCracken AA, Brodsky JL (1996) Assembly of ER-associated protein degradation in vitro: dependence on cytosol, calnexin, and ATP. The Journal of cell biology 132: 291-298. Only later on, the acronym has been used as an umbrella term that now covers all the pathways that control proteasomal clearance of misfolded proteins from the ER. A short historical excursus is presented in the new introduction to better explain these issues. __

It is well established that LC3 and the LC3 lipidation machinery have functions that go beyond macroautophagy (which involves double membrane autophagosomes). Micro-autophagy (or micro-ER-phagy to remain on the topic of our paper) is an example of autophagic pathway relying on ER-phagy receptor that engage LC3, on the LC3 lipidation machinery, without involving autophagosomes. This is schematically represented in the new Fig. 1B.

Several recent papers that appear relevant to the present study are not mentioned. In particular, Sun et al., Dev Cell 2023 (PMID: 37922908) appears worthy of discussion, as does Gonzalez et al., Nature 2023 (PMID: 37225996).

Thank you. Both papers are not directly linked to our study addressing the intervention of ERLAD pathways when ERAD activity is impaired. In particular the work of Gonzales et al describes post-translational modification of ER-phagy receptors for their activation. The Sun et al paper is not really related to the topic covered in our manuscript, but we cite it as an alternative pathway that removes ATZ from the ER (page 8).

Specific suggestions

1. Abstract: The abstract begins with "About 40% of the eukaryotic cell's proteome is synthesized ... in the ER." Similar statements can be found in many papers and purportedly reflect common knowledge. However, it is unclear where the figure of 'about 40%' comes from. It would be proper to provide a reference and demonstrate that giving such a fairly precise estimate is supported by experimental data. Alternatively, the statement could be modified to avoid being precise than is justified.

No reference is allowed in the abstract. We therefore modified the sentence as suggested by the reviewer.

1. p2: "The ER is site of gene expression in nucleated cells and ... native proteins to be delivered at their site of activity ...". There is something missing at the beginning of this sentence. Also, it should be 'delivered to their site of activity', not 'delivered at'.

Thank you

1. p2: "... by mechanistically distinct ER-phagy pathways collectively defined as ER-to-lysosome-associated degradation ERLAD." This statement suggests that all pathways subsumed under the term ERLAD are ER-phagy pathways, which I believe is misleading (see comment above on the distinction between autophagy and vesicular transport pathway).

---

See point 1.

1. p2: "KIF selectively ...". Please spell out KIF and explain what kind of compound it is.

Thank you, we changed to “_The alkaloid kifunensine (KIF) is a cell permeable selective inhibitor of the members of the glycosyl hydrolase 47 family of a____1,2-mannosidases_”____ __ 5. p3: "Notably, ERAD inhibition delays, rather than blocking degradation of ERAD clients ...". Please correct, for example: Notably, ERAD inhibition delays rather than blocks degradation of ERAD clients ...

Thank you

Figures 2 - 5: The number of quantified cells is given but it is not clear if experiments were done once or in biological replicates. Please indicate this in the figure legends.

__N is now given for all panels in the corresponding figure legends.____ __ 7. p4: "To verify if ERAD inactivation ..." sounds odd. Less ambiguous would be 'To test whether' or 'To ask if'.

Thank you

1. p7, beginning of discussion: Please correct "delivered at" to 'delivered to'.

Thank you

Reviewer #3 (Significance (Required)):

This is a concise and convincing manuscript with a clear message. The idea that proteins that cannot be processed by ERAD can be eliminated by other means, for instance by autophagy, is not new. Similarly, the FAM134B- and LC3-dependent pathway for ER-to-lysosome transport has been described by the authors before (Fregno et al, EMBO J 2018). Furthermore, the study exclusively relies on microscopy and does not attempt to tackle new mechanistic questions. Still, this study presents a definite functional advance in our understanding of the interplay of various ER quality control pathways.

The findings presented here will be of interest mainly to molecular cell biologists working on protein quality control and organelle homeostasis. However, given the disease-relevance of misfolded proteins, and alpha-antitrypsin in particular, the impact of this study may eventually go beyond basic research and may also interest translational researchers.

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 their study, Fasana and colleagues investigate protein quality control in the ER. Specifically, they test whether folding-incompetent proteins that are normally cleared by ER-associated degradation (ERAD) can also be targeted for degradation by direct vesicular transport from the ER to lysosomes in case ERAD is blocked. They show that blocking ERAD pharamacologically or genetically indeed leads to re-rerouting of an ERAD model substrate (the NHK variant of alpha-antitrypsin) to lysosomes and that this pathway requires the reticulon-like protein FAM134B, the ability of FAM134B to interact with the ubiquitin-like protein LC3 and the machinery for LC3 lipidation.

The paper is, for the most part, easy to follow. There are, however, a few minor issues and I think the authors could do more to connect their work with similar studies in the literature. Accordingly, I have some general and specific suggestions to make the manuscript more accessible for the reader.

General suggestions

1. To avoid confusion, it would be helpful to more clearly distinguish between vesicular transport to endolysosomes and autophagy. Previous work by the authors has defined a trafficking pathway from the ER to endolysosomes that appears to rely on conventional vesicle-mediated transport (Fregno et al, EMBO J 2018). This pathway delivers material from the ER lumen to the lumen of endolysosomes, which are both topologically equivalent to the extracellular space. Hence, this pathway is distinct from autophagy, which is the transport of cytoplasmic components to endolysosomes and thus the transport of material from intracellular to extracellular space. This distinction is particularly important as both vesicular ER-to-lysosome transport and autophagy of the ER involve LC3 and FAM134B, which is typically referred to as an ER-phagy receptor. To make this less confusing, it may be helpful to explain that FAM134B appears to be a multifunctional molecule that can function as a receptor for macroautophagy but also in the vesicular transport pathway studied here. In addition, it would be helpful to point out that LC3 appears to also have roles unrelated to autophagosome formation.
2. Several recent papers that appear relevant to the present study are not mentioned. In particular, Sun et al., Dev Cell 2023 (PMID: 37922908) appears worthy of discussion, as does Gonzalez et al., Nature 2023 (PMID: 37225996).

Specific suggestions

1. Abstract: The abstract begins with "About 40% of the eukaryotic cell's proteome is synthesized ... in the ER." Similar statements can be found in many papers and purportedly reflect common knowledge. However, it is unclear where the figure of 'about 40%' comes from. It would be proper to provide a reference and demonstrate that giving such a fairly precise estimate is supported by experimental data. Alternatively, the statement could be modified to avoid being precise than is justified.
2. p2: "The ER is site of gene expression in nucleated cells and ... native proteins to be delivered at their site of activity ...". There is something missing at the beginning of this sentence. Also, it should be 'delivered to their site of activity', not 'delivered at'.
3. p2: "... by mechanistically distinct ER-phagy pathways collectively defined as ER-to-lysosome-associated degradation ERLAD." This statement suggests that all pathways subsumed under the term ERLAD are ER-phagy pathways, which I believe is misleading (see comment above on the distinction between autophagy and vesicular transport pathway).
4. p2: "KIF selectively ...". Please spell out KIF and explain what kind of compound it is.
5. p3: "Notably, ERAD inhibition delays, rather than blocking degradation of ERAD clients ...". Please correct, for example: Notably, ERAD inhibition delays rather than blocks degradation of ERAD clients ...
6. Figures 2 - 5: The number of quantified cells is given but it is not clear if experiments were done once or in biological replicates. Please indicate this in the figure legends.
7. p4: "To verify if ERAD inactivation ..." sounds odd. Less ambiguous would be 'To test whether' or 'To ask if'.
8. p7, beginning of discussion: Please correct "delivered at" to 'delivered to'.

Significance

This is a concise and convincing manuscript with a clear message. The idea that proteins that cannot be processed by ERAD can be eliminated by other means, for instance by autophagy, is not new. Similarly, the FAM134B- and LC3-dependent pathway for ER-to-lysosome transport has been described by the authors before (Fregno et al, EMBO J 2018). Furthermore, the study exclusively relies on microscopy and does not attempt to tackle new mechanistic questions. Still, this study presents a definite functional advance in our understanding of the interplay of various ER quality control pathways.

The findings presented here will be of interest mainly to molecular cell biologists working on protein quality control and organelle homeostasis. However, given the disease-relevance of misfolded proteins, and alpha-antitrypsin in particular, the impact of this study may eventually go beyond basic research and may also interest translational researchers.

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 endoplasmic reticulum (ER) is a crucial site for protein synthesis and folding within the cell, and strict protein quality control is essential for maintaining ER homeostasis. In this context, ER-associated degradation (ERAD) and the unfolded protein response (UPR) play pivotal roles. Recent researches have highlighted the significance of ER-phagy in protein quality control. In this manuscript, the authors demonstrate the role of FAM134B in degrading misfolded proteins such as ATZ through the ER-phagy pathway when the ERAD pathway is obstructed. This work partially addresses a prominent issue in the field, unveiling the interconnections between different regulatory pathways in maintaining ER homeostasis.

Major issues:

1. In a multitude of experiments, the authors employed Bafilomycin A1 (BafA1) to block the fusion between autophagosomes and lysosomes, attempting to demonstrate that the clearance of misfolded proteins mediated by FAM134B is independent of autolysosomes. However, in Figure 4, the lack of rescue of FAM134B knockout by overexpressing FAM134B△LIR suggests a dependence on the interaction between FAM134B and LC3. The conclusions drawn before and after appear contradictory.
2. Some Western blot data are insufficient to substantiate the author's conclusions. For instance, in Figure 5D, the ATG7 KO line is inadequately supported
3. The author employed Lamp1 antibody for lysosomal staining in cells and observed a significant abundance of lysosomes in some experiments, as depicted in Figure 2C, 2D, 4I, etc. Is the phenomenon of lysosomes extensively filling the entire cell a common occurrence? Is it indicative of a normal physiological state?

Minor issues:

1. Some immunofluorescence experimental data are unclear. Please request the authors to replace these with more distinct images, as seen in Figure 3B and 3E.
2. Some expressions appear to be questionable. For instance, the necessity of utilizing endolysosomes requires clarification.
3. Some writing lacks precision, such as referring to FAM134B as FAM134.

Significance

• General assessment:
• Advance: provide an meaningful evidence that how two degradative pathways are coordinated in maintaining ER homeostasis.
• Audience: cell biologist
• Reviewer's expertise: autophagy, vesicle trafficking, organelle biolgy

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 from Fasana et al., the authors present data that investigates potential compensatory degradation pathways for misfolded glycoproteins in the ER - postulating that the ER-to-lysosome associated degradation (ERLAD) pathway becomes employed in the absence of a path for substrates to reach the ER-associated degradation (ERAD) mechanism. Using the classic ERAD substrate alpha1-antitrypsin NHK variant (NHK), the authors first demonstrate that pharmacologically preventing access of NHK to ERAD either with KIF (early) or PS-341 (late) elevates the number of LAMP-1 positive endolysosomes also immunoreactive for NHK (via HA), similar to what is observed for the ATZ variant that forms polymers in the ER (Fig 2). The authors next use shRNAs that silence essential ERAD factors (EDEM1, OS-9) involved in glycan recognition to demonstrate comparable enrichment of NHK in endolysosomes through genetic disruption (Fig 3). Next, the authors employ FAM134B-deficient MEFs to demonstrate the requirement for this ER-phagy receptor when ERAD is unavailable (Fig 4). Reconstituting FAM134B-/- MEFs treated with KIF/PS-341 + Baf, with a full length FAM134B rescue plasmid restores endolysosomal accumulation of NHK while a FAM134B-∆LIR does not, providing supporting evidence for substrate rerouting to ERLAD. Finally, the authors use knockouts of Atg7 and Atg13 to demonstrate dependence on LC3 lipidation and independence from macro-ERphagy (Fig 6), that points towards a pathway that is like that used to remove ATZ polymers. From these data, the authors conclude that ERLAD is increasingly engaged for substrate degradation when ERAD is impaired.

Major comments:

1. All assays rely on quantification of the NHK-HA substrates by microscopy. Would it be possible for the authors to also include biochemical analysis of NHK - potentially including data assessing its changing abundance and glycosylation state?
2. In Figure 3D, the knockdown of OS-9.1/2 is modest compared to that of EDEM1 (Fig 3A). Moreover, there is only data from single shRNAs presented. Could the authors please at least include another shRNA to confirm and demonstrate whether the targeting to ERLAD is accordingly scaled to loss of access to ERAD (based on the degree of OS-9 or EDEM1 remaining)?
3. While degradation is implied, it is not specifically demonstrated at any point in the manuscript. Perhaps the authors might include some demonstration of NHK stabilization in one of the figures via a translational shutoff or pulse-chase assay.
4. 10-30% of NHK-HA positive endolysosomes are detected even with Baf alone (e.g. Fig 2E)? Does this mean that Baf impairs ERAD to some extent since or is it evidence for continuous ERLAD involvement when ERAD is intact? If so, how much is its contribution?
5. An accounting of how much ERLAD is contributing to NHK degradation with or without ERAD impairment is not really present.. Effectively, how much degradation capacity is ERLAD making up? These would be interesting data to include if possible as they would speak to the "division of labour" for ER substrate degradation its potentially dynamic nature.

Minor comments:

1. In Figure 4, an increase is observed for the rescue of FAM134B-/-MEFs with WT FAM134B that is 50% greater that of WT MEFs, suggesting that its availability might be rate limiting. Could the authors compare the relative levels of FAM134B for the WT and KO-rescue MEFs to address this possibility?
2. In Figures 1 and 6, the terms siOS9 and siEDEM1 are used but Figure 3 shows data from shRNAs and not siRNAs.
3. Samples from Figure 3 treated with Baf but this is not indicated in the figure or figure legend.
4. VCP/p97 inhibitors typically stabilize ERAD glycoprotein substrates better than proteasome inhibitors do. Is the same degree of endolysosomal targeting present

Significance

Deconvolution of the different pathways taken by misfolded proteins to escape the ER is of great interest not only to the ER community but also represents consequences to consider for those interested in therapeutics involving UPS inhibition. While concise, this manuscript does a good job of trying to demonstrate the principal of substrate rerouting and the prioritisation of degradation pathways. Overall, the manuscript is well written, the experiments presented are performed to a sufficient standard, the data are lean but of good quality, and the appropriate statistical analyses have mostly been included where necessary and are described. The Methods and Materials is brief but describes the experiments that have been performed. The manuscript is brief in its results and would obviously benefit from additional complementary assays that would strengthen and broaden the authors arguments for rerouting. But too their credit, the authors do not grossly overstate their findings and merely present the culmination of a set of experiments to answer a single question - what happens to a misfolded glycoprotein substrate when ERAD is impaired. This is a key question with broad implications.

While their limited data clearly demonstrates an acquired dependence on ERLAD, one can't help but wonder how broadly these findings hold true, as only a single glycoprotein substrate example is used. Moreover, it is not clear what percentage ERLAD contributes to overall NHK degradation (with or without ERAD) as the total NHK amount remaining is not assessed or measured. Nevertheless, the manuscript is an advancement of understanding of the fate of substrates unable to access ERAD and raises many future questions of interdependency between the ERAD and ERLAD pathways. The data just need a bit of shoring up.

Expertise - ERAD, UPS, protein quality control

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 used hIPSCs to generate spheroids capable of elongation when cultured in KSR and exposed to BMP4. After examining all three developmental lineages of these treatments, the spheroids were embedded in different hydrogels to restrict movement and observed how this affected elongation and lineage differentiation. This manuscript was difficult to review because of the the style of the writing, the sometimes confused narrative of results, the problematic statistics and questionable interpretation of the literature. Authors should actively address the issues presented below with a significant overhaul of the text and presentation.

Overall comments

Major concerns - The body of the manuscript should be re-written to offer a coherent narrative that delivers a clear and condensed message . In fact, it was quite surprising given the well written and concise format of the introduction. However, there is some confusion about the interpretation of the first reference (https://doi.org/10.1038/s41586-020-2383-9) in which it was already presented that BMP4 treatment did not induce elongation in a hIPSC-derived gastruloid model. This is not to say that there is anything wrong with repeated experiments and different results, but the results from this reference were interpreted as if BMP4 treatment had induced elongation. The use of the methodology present in this manuscript as opposed to the validated model of using Wnt agonism to induce elongation is not convincingly justified. The results themselves have potential if reworded and condensed but in their current format they are not convincing. Statistics have substantial problems. According to the methodology, only one-way ANOVAs were used, when many of the circumstances would call for t-tests (if parametric) or at least mention the normality tests used to justify the ANOVAs. There is no clear mention of the number of independent differentiations and only one cell line was used. As a result of this, and the state of the statistics and interpretation of the data, the conclusions presented in this paper are inconclusive or misrepresented. The article should be re-written in a coherent scientific style, including not writing in the first person/active voice, and addressing the grammatical errors mentioned below.

Minor concerns - Throughout the methods there are issues with inconsistent notation, e.g. acronyms aren't always in brackets (HBSS, PBS, KSR), units are sometimes spaced from the value (10ng/mL or 10 µg/mL), units are incorrectly written (u instead of µ). There are also typographical errors such as basic grammar (capital letters and full stops), double spacing, typos, strange sentence construction, as well as interchangeable use of commas and semi colons when listing antibodies. These issues were not confined to the methods, but were particularly noticeable there, and so the whole manuscript requires thorough proofreading. Positive comments - The introduction was well written and concise, the images were mostly clear and easy to interpret, illustrating a result even if not directly the one described.

Specific Comments

Introduction - Comments

The introduction was well written and easy to follow. However, certain aspects should be refined in order to highlight the argument as to why the mechanical forces and environment are of importance to lineage determination in hIPSCs. For example, the paragraph describing 2D micropattern cultures could be reduced or integrated into the introduction of 3D gastruloid models. In its place, further exploration of examples of mechanical force's impact on stem cell/embryonic differentiation, both in vitro and in vivo, would be beneficial (perhaps including the geometric shape aspect mentioned in the 2D micropattern paragraph here instead). This would emphasise the necessity for investigation of environmental forces on 3D in vitro differentiation, tying it back to the first paragraph's broader developmental questions. Currently, it lacks the reiteration of the overarching purpose of this investigation.

Materials and Methods - Comments

hIPSC culture - The paper suggests that only one hIPSC line was used for this research; considering the variability innate to each stem cell line and their differentiation capabilities, at least two cell lines should be used. Please reference that stem cell validation was conducted by HipSci. It is concerning that antibiotics were used for tissue culture on such a short protocol and should not be used in future. Similarly, were these hIPSCs tested regularly for mycoplasma? Reference briefly the method in which hIPSCs were passaged and what range of passage number was used.

Spheroids derivation from hIPSCs - There is no mention of hIPSC confluency upon differentiation. Based on the brief 4 minutes of Accutase exposure, colonies would either be small to achieve single cell dissociation (which is not in the best interest of hIPSC colonies) or that single cell dissociation was not achieved. Seeding density is also close to double seeding density for currently circulating gastruloid protocols (see https://doi.org/10.1242/dev.150391) with no explanation as to why this cell count was chosen. Make mention of whether or not they were fed during the initial 48 hours. There is also no reference to how many independent differentiations were executed and therefore replicability is of significant concern considering the use of an unvalidated protocol.

Gastrulation-like induction of 3D hIPSCs - Difficult to follow timeline of treatments without reference to Figure 1 and therefore requires a rewrite for clarity. In addition, media 3 and 4 are formatted differently from 1 and 2, making it more confusing. Do not rely on Figure 1 to explain the treatment course, it should be clear in the methodology.

Live Imaging - No need to state each media, particularly if it is out of numerical order as described previously. Can be written as "Live hIPSC spheroids were imaged for 96 hours....". Also include magnification and time frame of each hour, e.g. between 48 and 96 hours. Immunofluorescence staining - Other than the grammatical errors referenced above, as there are numerous in this paragraph in particular, the catalogue number for each antibody should also be supplied. Considering the number of antibodies, perhaps this would be best supplied as a supplemental table. No mention either here or previously as to how non-encapsulated spheroids were mounted and if cleared.

Results - Comments

As mentioned previously, the results are written in a style which is not sufficiently rigorous for a scholarly article and should therefore be edited to reflect the standard expected quality. BMP4 signalling induces axial elongation in 3D gastrulation-like models -

Figure 1 - Nice and clean diagram but unnecessarily large, could be shrunk and placed at the top of Figure 2. Also is lacking key methodological details e.g. seeding density, reference to media changes in time line, etc.

An example of a statement that lacks rigor and specificity is the following: "We used defined medium conditions that provided consistent shape variation in spheroid morphology. E8 medium promotes self-renewing conditions, while KSR BMP4 triggers differentiation [31,32]". Self-renewing conditions of what? Differentiation of what? Into what cells?

These statements are too vague and remain without connection to the research This kind of issues are unfortunately recurrent throughout the manuscript. They are easily rectifiable with concise text rather than expecting the reader to search through references. "Consistent with our previous results..." - not referenced. Mention of size is brief in text and instead description of results is found in Figure 2 legend.

Figure 2 - Although Figure 2.A and 2.B are clear images, Figure 2.B is unnecessary and doesn't add anything to the message. Figure 2.C: The statistics for this table are hard to believe, except between E8 and KSR BMP4, due to the spread of the data. Change references to "size" to "sphericity", as size would arguably be better investigated using area as a metric, again be precise in choice of language. There isn't even a description of what Figure 2.D is portraying, nor error bars or notation of statistical testing.

Figure 3 - Figure 3.A Writing on scale bars is too small to be useful and unnecessary if mentioned in the legend. The images highlight the issue with the size/sphericity issue, although the images themselves are reasonably clear and will highlighted (if slightly overexposed and not cleared). The title of this figure is also misleading - E8 BMP4 is not emulating gastruloid development, and the gastruloid-like entity of KSR BMP4 has yet to be validated. A more appropriate title would be "Exposure to BMP4 in spheroid and elongated spheroid culture increases SOX17 expression". There is no mention of how cells were counted in methods and considering the variable fluorescence observed in Figure 3.A OCT4 for E8, this could easily be misconstrued. Inappropriate dictation of p value, should only be reporting the alpha. Much like the similar issue with sphericity, the error bars in Figure 3.E in particular make it difficult to believe the statistical significance achieved. Overall it seems inappropriate to focus solely on endodermal lineage and leave ectoderm and mesoderm to supplementals, when classically axial elongation in gastruloids is punctuated not just by SOX17 but also by BRACH expression throughout the extended region. This suggests that the authors may not have sufficiently evaluated the literature referenced. The justification provided by the authors for choosing SOX17 reinforces this fact by declaring that SOX17 is expressed during early development, as if SOX2 and BRACH isn't! Most concerningly, claims of SOX2's absence in E8 media spheroids as if to be a positive should actually be worrisome, considering SOX2 is a Yamanaka factor. Arguably, without further validation, this result undermines the foundation of this work.

PEG-peptide hydrogel encapsulation disturbs SOX17 patterning - The interchangeable use of "control" and E8 makes for difficult reading. Although no further substantial issues were taken with Figure 4 that had not already been addressed (scale bars, overexposure), the final comment of this section is simply inaccurate: "Overall, this suggests that embedding hiPSCs in a confining environment blocks morphogenesis, and despite the addition of BMP4, this was not sufficient to enable SOX17 expression.". It does not define that the hIPSCs in this case are spheroids and determines this assumption based off data in Figure 6.B and Figure 6.C which could have been easily integrated into Figure 4. Rather than comparing non-deg PEG to deg-PEG, in order to make this statement, it is necessary to compare the non-embedded spheroids to the embedded spheroids. There is also reference to a Figure 3.C. that does not exist.

Modulating PEG-peptide degradability promotes SOX17 expression - This section had similar issues to the one above, predominantly flitting between Figure 5 and 6 when it could have been condensed. The microscope images are also particularly poor, with high background in the 488 channel, overexposure for Hoechst and incomplete light penetrance in the centre for both E8 and E8 BMP4. Despite the lack of methodological explanation of how counting was conducted, these aspects are still likely to have had significant impact on cell count regardless of how the count was conducted. The last comment of this section is again questionable, as there is no definitive comparison as to what the deg-PEG is improving upon with regards to SOX17 expression.

Biochemical cues regulate morphogenesis via cell proliferation and cellular tension - There is no explanation as to why the manuscript pivots away from hydrogels to look at proliferation and F-actin. It does not appear to benefit the overarching goal of this manuscript (i.e. investigating the impact of confinement on 3D hIPSC-derived elongation models), with proliferation investigations feeling like an afterthought and F-actin investigations being placed about 3 sections too late. It was surprising to see EdU to check cell proliferation in this manner; similar can and should be achieved by using KI67 staining. This was a significantly missed opportunity to do pulse-chase experiments to understand potential cell cycle changes, and suggests the experimental design was not thought through. However, it is more reasonable to examine the F-actin network as shown by the manuscript. Although the pipeline in Figure 7.D is an interesting take on quantifying F-actin staining, no statistical analysis was undertaken to confirm a link between conditions and F-actin, nor F-actin and SOX17 expression, despite what this manuscript suggests. This should be coupled with information prior to the introduction to hydrogels, and then re-evaluated in those confined to hydrogels.

Overall these results need to be substantially edited for clarity, rigor and completeness. The narrative of the results is lost half way through, the statistics are questionable or non-existent and the logic behind certain experiments are debatable.

Discussion - Comments

To consider the models generated in this manuscript as gastruloids is misleading and inaccurate and should be replaced with something that reflects what was actually generated, e.g., BMP4-induced elongated spheroids. A quick search of key gastruloid papers, such as https://doi.org/10.1242/dev.113001, indicates how far detached the spheroids examined in this paper are, only sharing similar polarised SOX17 expression. In fact, the above referenced paper already examined aspects of classic gastruloid differentiation when exposed to BMP4, albeit using mESCs, and indicated that elongation only occurred in a proportion of the population. Similarly in the first reference which used hIPSCs specifically, https://doi.org/10.1038/s41586-020-2383-9, it was reported that BMP4 treatment did not induce elongation. At no point in this manuscript were these results discussed, which is concerning. Instead, this manuscript claims to have "robustly" concluded many of the results which have been questioned in detail above, with little reference to the variability across multiple differentiations or adequate statistical analysis. Noticeably they have also not mentioned their results on proliferation, which again makes the question of its relevance to the paper's aims.

Significance

The authors used hIPSCs to generate spheroids capable of elongation when cultured in KSR and exposed to BMP4. After examining all three developmental lineages of these treatments, the spheroids were embedded in different hydrogels to restrict movement and observed how this affected elongation and lineage differentiation. This manuscript was difficult to review because of the the style of the writing, the sometimes confused narrative of results, the problematic statistics and questionable interpretation of the literature. Authors should actively address the issues presented here with a significant overhaul of the text and presentation.

Our expertise is in stem cell biology, 2D and 3D model systems, microscopy, and single cell analysis.

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

Within the manuscript "Morphogen-driven human iPSCs differentiation in 3D in vitro models of gastrulation is precluded by physical confinement" by Alsehli et al, the authors aim to dissect the effects of biochemical cues and physical confinement using gastruloid models. Using this model, the authors find some indications that physical confinement prevents morphological changes and reduces lineage commitment of cells (as assayed by Sox 17).

While the manuscript addresses an interesting and important question, I do not think that the setup and data as present in the manuscript provide conclusive insight in any of the questions studied. First of all, the authors use a very different system (BMP4 induction) as compared to mainstream conditions for gastruloid formation. Based on the images and movies within the manuscript, symmetry breaking and elongation seem to be very sub-optimal under the conditions used by the authors, which makes any further functional assaying highly challenging. Importantly, it remains unclear why the authors did not (at least) include mainstream conditions (Chiron pulse) within their assays for comparative purposes, I feel this is a missed chance. On top of that, KSR is known to be poorly compatible with gastrulation and further development, at least in mouse (https://pubmed.ncbi.nlm.nih.gov/35988542/), so it remains unclear why the authors include a KSR condition.

Referees cross-commenting

Like me, both other reviewers indicated (major) shortcomings of the current paper, and indicated potential directions for improvement. Both these other reviews are very balanced as well. Altogether, there are significant changes required for the manuscript, and it depends on the authors which direction they would like to take this manuscript. Hence, I do not know what else to conclude than that in the current format there too many shortcomings to be of added value to current literature (similar to reviewer 3, and in line w/ reviewer 1 but who is somewhat more mild in their evaluation). If the authors decide to move on with this story, this paper requires a re-evaluation of a much improved version.

Significance

Secondly, I am not convinced by the data quality and representation in the manuscript. As outlined above, symmetry breaking and elongation seem to be very sub-optimal. Also, Fig 4b and Fig 5b hardly show any visible SOX17, in particular also not in the deg-PEG E8 BMP4 condition. Yet, the authors claim high expression of SOX17 in this condition in Fig 6C, as well as large and significant differences between conditions. To my feeling, Fig 6C does not represent the Sox17 observations of the IF. Similarly, I am not convinced by the EdU staining in Fig 7A. Importantly, the authors base their conclusions on only a single cell line, very few spheroid of this cell line, and an unknown number of biological experimental replicated (which to me seems single experiments). Altogether, the setup and data lacks reliability and robustness.

Currently, the efficiency of gastruloid formation is a important discussion in the field. Therefore, it is important to report on this in the current manuscript.

Altogether, while the manuscript makes a few interesting observations, it is very preliminary and not-well worked out or validated. Without being more robust and conclusive, I do not think the manuscript is of much added value to current literature.

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: This manuscript developed a BMP-based culturing protocol for gastruloid-like tissue that eexpresses markers from ectoderm, mesoderm, and endoderm. Through modulating the mechanical properties of extracellular environments, the expression of endodermal marker, SOX17, was found to correlate with the mechanical properties of the extracellular environment. However, the data presented only constitute a crude characterization of the cell lineages within the tissue and, therefore, the biological relevance of the gastruloid-like tissue, together with the associated discoveries, is not immediately evident.

Major comments

1. Extra explanation should be provided to justify why mechanical property of extracellular environments is a relevant/important factor for either early embryo development or gastruloid technology.
2. (OPTIONAL) To demonstrate the impact of mechanical property of extracellular environments, the authors compared two different gels, a stiffer gel with non-significant stress relaxation, and a softer gel with non-trivial stress relaxation. However, such comparison can be somewhat non-intuitive since a variety of mechanical properties are modulated simultaneously between two groups, including both tissue stiffness and its time-dependency. Since it has been well established that both elasticity and viscoelasticity can affect cell behavior, respectively, I would suggest the authors consider adding an extra comparison between a stiff elastic gel and a soft elastic gel.

Minor comments

1. The author provided individual staining of SOX2, BRA and SOX17, it would be more helpful to co-stain these markers to demonstrate their relative spatial distribution.
2. The quantification method for Fig. 7D&E seems missing from the method section, and more explanation for the method would be helpful. What is the area ratio between the inner and outer ROI? What does the y-axis number mean in Fig.7E?
3. In the last 2nd paragraph of "Biochemical cues regulate morphogenesis via cell proliferation and cellular tension", it was mentioned that "spheroids cultured in E8 medium exhibited organised and tightly packed F-actin with a homogenous network orientation", yet in the next paragraph, it was claimed that "F-actin orientation in round spheroids in E8 medium is mostly distributed in the inner core". Aren't these statements contradictory? Also in this paragraph, it was mentioned "Whereas in KSR conditions, F-actin was distributed in the periphery without significant differences compared to KSR medium conditions", I assume the authors intended to say "KSR BMP4" here?
4. The authors mentioned that "cells in the KSR BMP4 spontaneously elongate mirroring A-P elongation", yet based on the data presented, the elongation orientation appears more consistent with the D-V embryo axis?

Significance

This manuscript demonstrates a rudimentary investigation towards how mechanical properties of the extracellular environment may affect the marker expression within a gastruloid-like tissue. Such information may provide certain useful knowledge for the field of gastruloid. However, compared with the established gastruloid models in the field, the advantage of the tissue model developed in this study isn't very clear. Further, the physiological relevance of extracellular mechanics (and particularly, matrix viscoelasticity, which is studied in this manuscript) is not immediately evident.

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

This is a manuscript that focuses on understanding the function of BLTP2 protein in regulation of ciliation. BLTP family of proteins recently emerged as a lipid transporters in several contexts, such as formation and extension of autophagic membrane. The entire study is build on very intriguing observation that BLTP2 depletion leads to enhanced ciliation in RPE1 cells grown in the presence of serum. That makes BLPT2 a negative regulation of cilia formation that appears to mediate serum-dependent inhibition of ciliation in RPEs. Since molecular machinery governing serum-dependent inhibition of cilia formation remains poorly understood, this study does have a potentially high significance and interest. Unfortunately, most of the study (besides Figure 1) is done using Hela cells. It is a puzzling choice since Hela cells do not form cilia. Additionally, mots analyses are fairly descriptive and do not really lead to any specific hypothesis. As the result author's conclusion is very vague, specifically stating that "BLTP2 may suppress ciliogenesis by altering the lipid dynamics and/or densities of unidentified integral membrane proteins that suppress ciliogenesis". Thus, as it stands, this study does not really lead to new insights in BLPT2-dependent regulation of cilia formation.

Few specific other comments are listed below.

1. Figure 1. Two different individual siRNAs for each target should be used (unless authors show rescues) to minimize the possibility of off target effects.
2. Figure 2-4. Not quite sure what was the rationale to study BLTP2 localization in HeLa cells instead of RPE1. Considering that the main focus of this manuscript is ciliation, one would want to see extensive analysis of localization of BLPT2 in RPE1 cells.
3. The key to this manuscript (considering its focus on ciliation) would be to look at BLPT2 dynamics in the RPE1 cells in the presence and absence of serum, especially during ciliary vesicle and cilia formation. Instead, authors do most of their analysis in HeLa cells that do not even form cilia.

Significance

This is a manuscript that focuses on understanding the function of BLTP2 protein in regulation of ciliation. BLTP family of proteins recently emerged as a lipid transporters in several contexts, such as formation and extension of autophagic membrane. The entire study is build on very intriguing observation that BLTP2 depletion leads to enhanced ciliation in RPE1 cells grown in the presence of serum. That makes BLPT2 a negative regulation of cilia formation that appears to mediate serum-dependent inhibition of ciliation in RPEs. Since molecular machinery governing serum-dependent inhibition of cilia formation remains poorly understood, this study does have a potentially high significance and interest. Unfortunately, most of the study (besides Figure 1) is done using Hela cells. It is a puzzling choice since Hela cells do not form cilia. Additionally, mots analyses are fairly descriptive and do not really lead to any specific hypothesis. As the result author's conclusion is very vague, specifically stating that "BLTP2 may suppress ciliogenesis by altering the lipid dynamics and/or densities of unidentified integral membrane proteins that suppress ciliogenesis". Thus, as it stands, this study does not really lead to new insights in BLPT2-dependent regulation of cilia formation.

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 Parolek at al demonstrates the role of Bridge-like lipid transfer protein family member 2 (BLTP2) as a negative regulator of ciliogenesis in cultured retinal pigmental epithelial (RPE-1) cells. Based on genetic interaction of BLTP2 with another ciliogenesis negative regulator, WDR44, a RAB11 effector, the authors mechanistically study the regulation of cilia by these proteins. The authors nicely show Hela cell specific colocalization of BLTP2 and WDR44 in the ER-tubular endosome network membrane contact sites (ER-TEN MCS). They demonstrate that the C-terminus of BLTP2 is necessary for localization to the TEN. They also find that the WDR44 network to be present in higher proportion of Hela cells lacking BLTP2. Interestingly, BLTP2 and WDR44 also localize at tips of GFP-Rab8 and GFP Rab10 tubular membrane network in Hela cells. However, RPE-1 cells do not possess WDR44- and BLTP2-associated tubules, so the context of the TEN with the ciliogenesis phenotype in RPE cells, if any, is presently unclear. The rigor and reproducibility of the presented experiments is high and the presented results are novel for cilia, membrane trafficking, memembrane contact sites and lipid transfer fields. Overall, the authors suggest that BLTP2 and WDR44 are common components of a pathway that suppresses ciliogenesis in serum-fed RPE-1 cells.

Major comments

How general are the effects of BLTP2 on ciliogenesis? Effects on ciliogenesis in other cells would consolidate the current results in RPE-1 cells.

Are the effects on ciliogenesis in RPE cells mediated by the role of BLTP2 and WDR44 in TEN network shown in Hela cells? If not, can the authors demonstrate the subcellular localization of these proteins with respect to ciliogenesis, and optionally if possible with the Rab8-Rabin8-Rab11 cascade? For eg., WDR44 knockdown has been shown to increase peri-centrosomal localization of Rabin8 in serum-fed conditions. Does BLTP2 knockdown have similar effects?

The authors show that the C-term of BLTP2 is necessary for its recruitment into TEN network, irrespective of WDR44 in Hela cells. Does the C-term of BLTP2 also regulate ciliogenesis?

Optional: Interactions between BLTP2 and WDR44 are currently unclear other than their colocalization. Can the authors further address if there are close physical interactions between these proteins and/or Rab11?

Minor comments

Fig 3C. In merge/inset: please check the white bar.

If possible, show rescue of ciliogenesis phenotype from BLTP2 knockdown or knockout.

Significance

The roles of lipid transfer proteins in ciliogenesis are currently unknown. By showing the function of BLTP2 in ciliogenesis, the authors suggest the possibility of lipid transfer between the ER and the endo-lysosomal network in regulating ciliogenesis pathways. The results showing co-localization of BLTP2 and WDR44 in ER-tubular endosome network membrane contact sites (ER-TEN MCS) is nicely done and the effects on TEN upon WDR44 loss is also very intriguing. Recent papers suggest a role of WDR44 in ciliogenesis and in ciliopathies. The current data will be of broad interest to readers in ciliogenesis, intracellular trafficking, and lipid transport fields. From my own expertise in cilia biology, I think the results are exceptionally novel but needs more context to further solidify the role of BLTP2 in ciliogenesis.

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

This manuscript by Parolek and Burd addresses the role of the evolutionarily conserved Bridge-like lipid transfer protein 2 (BLTP2) VPS13 family member in negatively regulating ciliogenesis. BLTP proteins are thought to mediate the transfer of lipids between a variety of different organelles, and while BLTP2 loss-of-function in plants and multicellular organism including flies and mice has been connected to various phenotypes, the role of human BLTP2 is poorly understood. In the current study, the authors initially used the Cancer Dependency Map Portal to find a very high mutual co-dependency between the BLTP2 gene and WDR44, a previously well-established negative regulator of primary ciliogenesis. This led the authors to ask whether BLTP2 (similar to WDR44) affects ciliogenesis and indeed they demonstrated that knock-down of either protein in serum-containing RPE1 cells (conditions that normally do not facilitate robust cilia formation) led to an increased percentage of ciliated cells. The authors next addressed the localization of human BLTP2, finding it on ER membranes as well as in linear/tubular structures that correlate with ER subdomains, at least in HeLa cells. Moreover, BLTP2 localization overlapped with that of WDR44. The authors then demonstrated that in BLTP2 knockout cell lines a significantly greater percentage of cells displayed a robust tubular endosome network than in wild-type cells. Finally, the authors show that both BLTP2 and WDR44 tend to localize to the tips of the endocytic tubular network, suggesting that these proteins may localize to contact sites between the ER and (tubular) endosomes, and further, that these contact sites may be required for primary ciliogenesis.

Major Comments

This is an interesting manuscript that addresses a poorly understood protein, and potentially provides a novel connection between ER-recycling endosome contact and ciliogenesis. The experimentation is well executed, and the presented data are clear. The strengths of the manuscript lie in partially unveiling new function for BLTP2, functionally associating BLTP2 with WDR44, and the potentially novel relationship between ER-endosome contact sites and ciliogenesis. Moreover, the tubular network of endosomes (which has been primarily implicated in recycling), remains an incompletely understood organelle and this manuscript suggests a new role for it. The weaknesses lie in the preliminary nature of the studies (which in the current form are largely correlative/speculative), and the somewhat descriptive nature of the study. For example, while the proposed connection between ER-endosome contacts, TEN, and cilia is intriguing, there is little attempt to address how such contact sites and TEN might affect ciliogenesis. Developing these ideas would greatly increase the significance of the findings, which as presented are mainly correlative. Other key questions include whether these proteins affect functions previously attributed to tubular endosomes, such as receptor recycling and endocytosis? What is the function of the TEN in maintenance or generation of ER contact sites? There is some literature suggesting that TEN serve in the sorting and/or fission process that might separate Rab5 anterograde transport from Rab4 recycling endosomes, but this is not addressed in the context of the current manuscript. How do ER contact sites and TEN impact ciliogenesis at a molecular level? How might a lipid transfer protein affect primary ciliogenesis? Is the TEN network connected to lipid transfer? Reports hold that PI4,5P2, PI4P, PA and other lipids might be enriched in the TEN; are these (or other lipids in the TEN) affected by BLTP2? In addition, while BLTP2 and WDR44 are linked by localization (and their mutual co-dependency), little attempt has been made to understand how these proteins function together. Do they physically interact? Are Rab8/10 responsible for their recruitment to the tips of endosomes?

Minor comments:

Could endogenous Rab proteins be used to address localization of BLTP2 and WDR44 to tubule tips?

Could the authors please explain how the R=0.61 mutual co-dependency was arrived at? This reviewer used the portal website to find an R value of 0.58. Admittedly these numbers are both very high and the difference between them not significant, but nonetheless they are not identical.

Significance

The novelty of this study is that it identifies a new player, BLTP2, involved in the regulation of primary ciliogenesis. Since ciliogenesis impacts so many physiological events and underlies over 100 distinct ciliopathgies, there is clear significance in identifying new proteins that regulate this process (even if hundreds have already been identified). Perhaps the most (potentially) novel finding is the relationship between the ER contact sites, the tubular network of endosomes and ciliogenesis. Few papers have made any connection between tubule generation and ciliogenesis, with the exception of Insinna et al. (Nat Comm., 2019), a paper reporting that the tubular endosome scaffold MICAL-L1 is required for ciliogenesis (Xie et al., J. Cell Sci., 2019), and the recent Biorxiv manuscript by Ott and colleagues suggesting that tubules allow formation of "deconstructed" cilia, none of these connect ciliogenesis to ER contact sites. The limitations of the study are the preliminary nature; as noted, while the individual experiments are clear, substantial data that clearly connect ER contact sites, tubular endosomal networks and ciliogenesis remain largely uncharacterized in this study. In terms of the audience, if a more tangible relationship between ER contact sites/TEN and cilia could be derived, the study would be cross-disciplinary and of interest to a wide audience.

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

---

1. General Statements [optional]

Response: We would like to thank both reviewers for their insightful comments. We have addressed most comments in the transferred manuscript and are willing to perform additional experiments to respond to some remaining points, as detailed in the following sections.

1. Description of the planned revisions

Insert here a point-by-point reply that explains what revisions, additional experimentations and analyses are planned to address the points raised by the referees.

Reviewer #1:

Fig. 5, It would also be interesting to explore the involvement of IFN-I receptors (IFNAR1 vs IFNAR2) by dissecting IFN-α from IFN-β responses.

Response: Our understanding is that IFNAR1 and IFNAR2 form a heterodimer, which can be activated by both IFN-α and IFN-β. It is thus difficult to dissect the role of individual receptors or cytokines from each other. However, to confirm that IFN-α and IFN-β are both acting through the IFNAR1 receptor to prime for IL-1b activation and release in human neutrophils, we will perform additional experiments by treating neutrophils isolated from healthy individuals with both cytokines individually (as opposed to using them synergistically as in the current manuscript) and blocking their effects by a commercial IFNAR1 blocking antibody. 

Reviewer #2:

- The authors are encouraged to specify the number of independent experiments conducted. For mouse model studies, it is recommended to include results that are either representative of, or aggregated from, a minimum of two independent experiments to ensure robustness of the data.

Response: We did three independent animal experiments in total. The first two experiments involved two time points (2 and 4 days post infection) with different downstream experiments: The first experiment was used for the detection of viral loads by PCR as well as IHC-based quantification of viral antigen and neutrophil influx into the lungs (Exp1, new suppl. Fig 6) while the second experiment (Exp2) was used to isolate neutrophils from the lungs in order to assess neutrophil caspase1 activity and to obtain samples for RNAseq (new Fig.7 A-D). The third experiment (Exp3) was to perform anti-IFNAR and isotype treatments for infected mice; here, all animals were at day 2 post infection. For this experiment we prepared lung samples for IHC and PCR (new suppl Fig. 7) and isolated neutrophils for the assessment of caspase1 activity and for PCR assays (new Fig. 7E-G). We have now improved the text and supplementary Table S2 to indicate these separate experiments more clearly.  In our experience, this approach, i.e. making use of the lungs of oe animals for the determination of several parameters, is of benefit not only because we can directly compare these parameters in a given animal and can thereby reduce the number of animals used for the project (following the 3R principles).

However, in order to increase sample size and to respond to the various issues raised by the reviewer regarding our mouse work, we have decided to undertake another mouse infection experiment to analyze the effect of the anti-IFNAR treatment on viral loads by assaying for viral titers in the lung sample (which will hopefully directly answer the reviewer’s concerns raised below regarding the effect of anti-IFNAR treatment on viral replication). We will also confirm the negative effect of anti-IFNAR treatment on the expression of IFN-responsive genes by measuring OAS2 mRNA levels by PCR in the lungs of the anti-IFNAR treated as compared to isotype-treated mice. Finally, we will isolate neutrophils from the lungs to repeat the experiment showing the effect of anti-IFNAR treatment on neutrophil inflammasome activity as shown in Fig 7. E-G and will take a sample each for the histological and immunohistological analysis to complement the other tests.

• The authors are advised to employ more quantitative methods, such as flow cytometry, to measure neutrophil recruitment in mice. Additionally, it should be clarified how many tissue sections from each mouse were assessed for every experimental condition to ensure the reproducibility and statistical validity of the results. 

Response: While we are not aware that flow cytometry can be considered as a “more quantitative method” than morphometry, we agree that it is an alternative, i.e. complementary quantitative method. Furthermore, we feel that direct quantification of isolated Ly-6G+ neutrophils which are obtained from homogenized lung tissue by magnetic beads (as we have done for some mouse experiments of the current manuscript) is another quantitative approach and a method comparable to flow cytometry. Therefore, we are willing to repeat the mouse infection experiments and quantify the isolated neutrophils in parallel to IHC-based morphometry in order to determine the robustness of our morphometrical neutrophil quantification but are not inclined to undertake flow cytometry, in particular since this would not allow the assessment of all other relevant parameters in the same lung.

The reviewer asked for information regarding the number of tissue sections that were assessed from each mouse. In our opinion, for quantitative purposes (i.e. morphometry), it is more meaningful to determine the total tissue area that is examined, as different pathologists take different approaches to trim the lung for histological examination (cross sections vs. longitudinal section of a lobe). In our study we examined a tissue area of 19.5 ± 6 mm2 for each lung, which is stated now also in the revised manuscript.

Revised manuscript, methods section, line 307:

“The average total tissue area used for the quantification was 19.5 ± 6 mm2”.

• The authors need to address discrepancies in their text regarding the effects of anti-IFNAR1 blockade on viral titers and neutrophil recruitment in SARS-CoV-2 infected mice. While they state there is no change, Supplementary Figure 5D suggests increased SARS-CoV-2 NP staining with anti-IFNAR1 treatment, and there appears to be a lack of quantitative data on lung neutrophils to substantiate the claim that neutrophil recruitment remains unaffected. It is necessary for the authors to provide a more detailed explanation or additional data to resolve these inconsistencies *

Response: We did not observe any significant differences in the antigen expression between anti-IFNAR treated and isotype-treated mice. The lack of differences in viral loads was also shown by PCR (suppl Fig. 6B). However, since we are suggesting to repeat the anti-IFNAR experiment which will also include the quantification of viral titers in lung tissue with and without anti-IFNAR treatment, we will gain further insight whether and how IFN-I activity could regulate infection kinetics. We will also repeat the quantification of neutrophils in the lungs of anti-IFNAR and isotype-treated mice by IHC-based morphometry as well as by determining the number of isolated LY-6G+ neutrophils. 

• The authors should demonstrate the effectiveness of anti-IFNAR1 blockade in mice by providing evidence of sustained inhibition of IFN-I signaling throughout the duration of the experiment to validate the treatment protocol used. 

Response: We feel the observed inhibition of inflammasome-related pathways by anti-IFNAR treatment strongly argues that blockade of IFNAR activity was successful during the 2 day time course of the direct experiment (Exp3). However, the reviewer’s comment is valid since we have not shown the effect of anti-IFNAR treatment on specific IFN-induced genes. Therefore, we will repeat the anti-IFNAR treatment in infected mice and confirm its negative effect on the expression of IFN-responsive genes by detecting the expression of OAS2 mRNA in lung samples by PCR (OAS2 is one of the mostly upregulated genes by SARS-CoV-2 infection based on our neutrophil transcriptomics analysis). We will also assess whether the treatment directly affects the levels of infectious virus by quantifying viral titers in lung tissue.

*Referee Cross-Commenting*

• I agree with Reviewer#1 about trying to dissect the role of IFNAR-1 vs IFNAR-2. Authors partially look at this in the in vivo mice experiments using the anti-IFNAR1 blocking Ab, but it would reinforce the study to see if this holds with human cells. 

Response: To confirm that IFN-α and IFN-β are both acting through the IFNAR1 receptor to prime for IL-1b activation in human neutrophils, we will perform additional experiments by treating neutrophils isolated from healthy individuals with both cytokines individually (as opposed to using them synergistically as done for the current manuscript) and blocking their effects by a commercial IFNAR1 blocking antibody. 

The study presents an investigation into the role of neutrophils and inflammasome formation in COVID-19 pathology, contributing to the field with transcriptomic profiling of neutrophils from varying severity of patient cases and a SARS-CoV-2 mouse model. A significant IFN-I gene signature in severe cases was confirmed, and differences in inflammasome response were identified, adding to our understanding of disease mechanisms.

Strengths of the paper include the comprehensive analysis of neutrophil maturation states and the novel insights into the priming of inflammasome activation by IFN-I. However, limitations were noted in the purity of samples for RNAseq analysis and the lack of conclusive in vivo evidence for the direct role of IFN-I in neutrophil inflammasome priming. The study's implications suggest potential avenues for targeted therapies, but the authors were advised to moderate their conclusions without stronger in vivo evidence and to clarify the potential therapeutic implications of their findings.

Additional suggestions for improvement include the use of more quantitative methods like flow cytometry for neutrophil recruitment measurements, clarification on experimental replication, and resolving discrepancies in data presentation regarding anti-IFNAR1 blockade effects. Furthermore, the paper would benefit from discussing the relevance of autoantibodies against IFN-I in the context of their findings and from exploring the causal relationship between inflammasome patterns and disease severity.

Response: We would like to thank the reviewer for highly impactful comments. As suggested, we have redone the RNAseq analysis including only samples of higher neutrophil content, with similar conclusions as those previously made, and have amended the discussion based on the reviewer’s comments. Furthermore, to address the remaining questions raised by the reviewer, we will undertake the following additional animal experiments:

• Five groups of animals (n = 4);  

• PBS-inoculated animals 

• Infected with SARS-CoV-2 for 2 days, treated with isotype 
• Infected with SARS-CoV-2 for 2 days, treated with anti-IFNAR
• Infected with SARS-CoV-2 for 2 days 

• Infected with SARS-CoV-2 for 4 days 

• Measurements from lung tissue  

• Quantification of LY-6G neutrophils by morphometry 

• RT-PCR for OAS2 mRNA (assessment of the successful blockade of IFN-I signaling by IFNAR antibody) 

• Viral titers by quantification of infectious virus in cell culture (assessment of the successful blockade of IFN-I signaling by IFNAR antibody) 

• Isolation of LY-6G neutrophils from lung tissue 

• Quantification of the number of isolated LY-6G neutrophils 

• Repeating measurement of caspase1 activity 

• Repeating RT-PCR for caspase1 and IL-1b mRNA

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

Please insert a point-by-point reply describing the revisions that were already carried out and included in the transferred manuscript. If no revisions have been carried out yet, please leave this section empty.

Reviewer #1:

IFN-I promotes immune responses to diverse viruses. Some of them have evolved to dampen IFN-I responses in order to weaken or delay antiviral responses. A number of studies support the idea that this could be the case of SARS-CoV-2 (0.1038/s41586-022-04447-0, 10.1016/j.it.2021.02.003, 10.1038/s41467-022-34895-1, to cite a few). The authors should refer to some of these studies to counterbalance the statements of increased IFN-I response in severe COVID-19 patients in some parts of the manuscript, such as in the Discussion.

Response: This is a good point and we added discussion on this topic.

Revised manuscript, discussion section, between lines 634-648:

“It should be noted that several SARS-CoV-2 encoded proteins have been shown to inhibit IFN-I signaling (60). However, no evidence suggests that neutrophils can be infected by SARS-CoV-2 and therefore it seems unlikely that such direct virus mediated effects could play a role in the observed neutrophil unresponsiveness to IFN-I.  

The dualistic nature of the IFN-I response in COVID-19 has been recognized previously. It seems that a strong initial IFN-I response to SARS-CoV-2 is more likely to result in asymptomatic or mild COVID-19 whereas a decreased initial IFN-I activity, due to e.g. genetic defects or increased levels of IFN-I autoantibodies, can lead to more severe COVID-19 (61). This initial beneficial effect of IFN-I is probably due to its ability to limit viral replication at early stages of the infection. However, at later stages of the disease IFN-I can be detrimental by promoting inflammatory pathways instead of direct antiviral effects (62). Thus, similarly to the IFN-I response in general, the role of neutrophil inflammasomes in development and severity of COVID-19 might be dualistic in nature with an initial protective effect while damaging when sustained for prolonged periods.”

1. Rashid F, et al. Roles and functions of SARS-CoV-2 proteins in host immune evasion [preprint]. Front Immunol. 2022;13. https://doi.org/10.3389/fimmu.2022.940756.
2. da Silva RP, et al. Circulating Type I Interferon Levels and COVID-19 Severity: A Systematic Review and Meta-Analysis. Front Immunol. 2021;12. https://doi.org/10.3389/fimmu.2021.657363.
3. Smith N, et al. Defective activation and regulation of type I interferon immunity is associated with increasing COVID-19 severity. Nat Commun. 2022;13(1). https://doi.org/10.1038/s41467-022-34895-1.

In Fig 5, the authors stimulate PMNs from healthy controls in vitro with IFN-I and use mainly IL-1b as a readout for neutrophil priming. The authors should analyse whether IFN-I-mediated priming ultimately leads to NETosis, given the relevance of NETs to COVID-19 pathology (10.1016/j.tips.2023.06.007), which is acknowledged by the authors in the Introduction.

Response: We agree that NETosis is a potential outcome of neutrophil activation. This is in part why we assayed for the release of MPO, which can be used as a marker of NETosis, in the ex vivo activated PMNs. However, while a clear increase in MPO levels was observed in response to nigericin treatment, no increase was seen with IFN-I or LPS treatments alone in COVID-19 or HC PMNs (Fig. 4C). We slightly modified the text to make it clearer that MPO served to detect PMN degranulation and NETosis in our study. 

Revised manuscript, results section, lines 440-444:

“However, the release of myeloperoxidase (MPO), used as a marker of degranulation and/or NETosis, in response to nigericin was similar between COVID-19 PMNs and HC PMNs, and therefore the observed diminished IL-1β release by COVID-19 PMNs is not due to general cellular inertia but may be specific to the ex vivo induced inflammasome pathway.”

In Methods (Histology and immunohistochemistry), the authors mention that histone H3 is a NET marker and use this in Supp Fig 5 to support evidence of NETosis. The authors should state which antibody clone/company was used. Histone H3 is expressed in high levels by non-NETotic neutrophils. Citrullinated histone H3 (CitH3), on the other hand, is detectable by few commercially available antibodies and can be used as a NET marker in conjunction with other markers such as DNA staining (10.1084/jem.20201129).

Response: For immunohistochemical detection of NETosis, we have used the following antibody: rabbit anti-histone H3 (citrulline R2 + R8 + R17; Abcam). We have provided a reference to a publication of a co-author (Schmid AS, et al. Antibody-based targeted delivery of interleukin-4 synergizes with dexamethasone for the reduction of inflammation in arthritis. Rheumatology (United Kingdom). 2018;57(4):748–755. Doi: 10.1093/rheumatology/kex447) which includes this information and fully describes the staining protocol. We admit that our labelling of the figure and the text was misleading by stating that the used antibody was anti-histone H3 instead of indicating that an antibody detecting citrullinated form of histone H3 was used. We have therefore relabeled Supplementary Fig. 7 and rewritten the text to indicate this more clearly.

In the publication kindly referred to by the reviewer, the following antibody was used: rabbit anti-histone H3 (H3Cit; Abcam; cat. ab5103; 1:500) in immunofluorescence, where DAPI fluorescence served to highlight the DNA. We generally work with immunohistochemistry instead of immunofluorescence as it allows better alignment with histopathological features. However, we have now applied the anti-H3cit antibody in a fluorescence protocol, using the same tissues and antibody as shown in the immunohistochemistry image of suppl. Fig7, to indicate that the used antibody works equally well in immunofluorescence and immunohistochemistry.

In the PDF version of this revision there is a figure plate that shows: NET IHC, showing abundant expression in the lumen of a bronchiole (top) and NET IF, showing part of a bronchiole with NET (green) and nuclei (DAPI; blue) and a closer view (bottom) with NET expression in a cell with the nuclear morphology of a neutrophil (arrowhead).

In Fig 8, the authors show that neutrophils migrate to the lungs of infected animals. This finding is showed in many previous studies and does not seem to favour the structure of the manuscript. Insteand, it seems it would fit a Supp Fig better, or a portion of the following figure.

Response: We are aware that neutrophil recruitment into the lungs with SARS-CoV-2 infection has been shown previously. However, this was so far not done in the model that we have used, mouse-adapted SARS-CoV-2 infection in wild type (BALB/C) mice in which the infection is short-lived and wanes off after 4 days (Gawish et al., 2022). We are happy to move this figure plate to the Supplements as Reviewer #2 shares this opinion. In the revised manuscript, it now features as Supplementary Fig. 6.

Reviewer #2:

Major points:

-Figure 2A: for this RNAseq analysis, the authors claim that "This analysis also demonstrated that cells in the LDG fraction were predominantly immature neutrophils, meanwhile PMNs were composed of mainly mature neutrophils (Figure 2A)". Nevertheless, there are 2 samples in the "Severe COVID-19" group that show a fraction of neutrophils {less than or equal to}0.6, which indicates low levels of purity. In addition, there is one "PMN" sample with high LDG fraction (around 40%). Authors should remove from this analysis samples with such a low purity since the big fraction ({greater than or equal to}40%) of contaminating cells could introduce a bias in this group. Are the differences observed still present in the absence of these low-purity samples?

Response: For Figure1 of the original manuscript, we first performed the analyses including all the samples, and after assessing their purity, included only the samples with highest purity in the following figures. However, we followed the reviewer’s recommendation and redefined the purity of the samples as >65% of total neutrophils, independent of their maturity. Of note, unlike the reviewer suggests, we did not remove the one “PMN” sample with high degree of immaturity, since PMNs and LDGs are defined based on their isolation method and not their degree of maturity. Along these lines, we have decided to move the new unsupervised heatmap and present it together with the RNA deconvolution plot as supplementary Figure 1. Therefore, Figure 1 now contains the samples with high purity based on the redefined criteria, with PCA in panel A, pathway analyses in panel B (PMN vs LDG) and interferon-related genes heatmap in panel C. We also redid suppl Fig. 2 to include samples with the redefined criteria and modified the results text accordingly.

Revised manuscript, results section, lines 333-380:

“Unsupervised RNA-seq analysis reveals an antiviral gene expression signature of circulating neutrophils in COVID-19 that is strongly influenced by maturity

With our recent findings on increased frequencies of low-density granulocytes (LDGs, isolated from the PBMC fraction) during COVID-19 and their likely relevant role in disease progression (7), we sought to understand in more detail how the transcriptomic profile of LDGs differs from their higher “normal” density counterpart, the circulating polymorphonuclear cells (PMNs) (31), typically consisting mainly of mature neutrophils. Neutrophils isolated from different cohorts comprised three PMN groups (severe COVID-19, mild COVID-19, and healthy controls), and one LDG group. Initial deconvolution of the RNA sequencing (RNA-seq) data allowed us to gain a comprehensive understanding of the cellular composition within PMN and LDG fractions and verified that most cells present in the samples were neutrophils (Supplementary Figure 1A). This analysis also demonstrated that cells in the LDG fraction were predominantly immature neutrophils, meanwhile PMNs were composed of mainly mature neutrophils.

The samples with predominant neutrophil cell populations were selected for subsequent gene expression analysis (neutrophils ≥ 65 %). The high variance in gene expression between PMNs and LDGs was confirmed by principal component analysis (PCA) (Figure 1A), which revealed that the gene expression patterns of COVID-19 LDGs differed from those of all PMNs regardless of the patients’ disease state. Functional enrichment analyses through gene overrepresentation (ORA) and gene-set enrichment analyses (GSEA) (Figure 1B) compared PMNs with LDGs from severe COVID-19 patients. The most statistically significant result was an overrepresentation of the NOD-like receptor signaling pathway in PMNs in contrast with LDGs, highlighting that the different neutrophil fractions have a distinct inflammatory profile. This was supported by GSEA, where the most obvious increases in fold changes were the enrichment of the interferon signaling pathways. Another relevant difference was the cell cycle and DNA replication pathways, identified by both ORA and GSEA, which supported our previous findings suggesting LDGs to be predominantly immature cells (7). Furthermore, a heatmap of selected type I IFN (IFN-I) related genes confirmed a robust IFN-I gene signature in severe COVID-19 PMNs, while LDGs from severe COVID-19 distinctively lacked this signature (Figure 1C). Unsupervised clustering analysis, namely Iterative Clustering and Guide Gene Selection (ICGS) using the AltAnalyze software, supported these findings by identifying the top 118 differentially expressed (DE) genes, including several IFN-related genes (Supplementary Fig. 1B). Similarly to the selected samples included in Figure 1, this analysis classified the samples into two major clusters: a first one containing all isolated LDG samples, and a second one comprising all isolated PMN samples. The former cluster consisted of neutrophil antimicrobial and granule marker genes (e.g. DEFA3, DEFA4, SERPINB10, CTSG), while in the latter cluster the most significantly upregulated genes in the PMNs from severe COVID-19 subgroup were mainly interferon inducible (e.g. IFI44L, IFI6, GBP3, IRF7). These differences were supported by a detailed gene analysis (Supplementary Fig. 2A).

Inflammasomes are activated in severe COVID-19 PMNs, but not directly by SARS-CoV-2

Looking more closely into PMN fractions, pathway analyses identified the inflammasome related NOD-like and RIG-like receptor signaling pathways among the most significantly overrepresented pathways, differentially expressed in severe COVID-19 PMNs versus HC PMNs (Figure 2A and Supplementary Fig. 2B-C) or mild COVID-19 PMNs (Figure 2B and Supplementary Fig. 2D, E). However, mild COVID-19 PMNs did not significantly differ from HC PMNs in their inflammatory profile (Supplementary Fig. 2F).”

Finally, to respond to the reviewer’s specific question concerning whether the observed differences are still present in the absence of the newly defined low-purity samples, we can conclude that the results continue to highlight the differences we had previously described (increased cell cycle and metabolism-related pathways in LDGs and a distinct IFN-I signature in severe COVID-19 PMNs). 

- While the in vitro data suggest that IFN-I may prime the inflammasome response in neutrophils, in vivo evidence remains inconclusive. The systemic blockade of IFNAR1 with antibodies in infected mice does not confirm that IFN-I directly primes neutrophil inflammasomes, as other cells could initially sense IFN-I and subsequently produce neutrophil-activating stimuli. In the absence of in vivo experiments utilizing conditional IFNAR1 knockout models, such as Mrp8-Cre x IFNAR1 fl/fl mice, the authors should consider moderating the stated significance of these findings in the discussion about the limitations of the study.

Response: We agree with the reviewer on this point. Thus, we modified the “limitations of study” paragraph in the discussion.

Revised manuscript, discussion section, lines 674-677:

“Furthermore, the observed inhibitory effects on neutrophil inflammasome activity by IFNAR blockade does not exclude the possibility that IFN-I could promote neutrophil inflammasome formation by indirect effects such as stimulating the release of pro-inflammatory cytokines by other cell types.” 

- Given the reports of life-threatening COVID-19 infections occurring in conjunction with autoantibodies against type I IFNs (DOI: 10.1126/science.abd4585), the authors should explore how this intersects with their findings. A discussion is needed on whether patients with such autoantibodies may exhibit inflammasome activation patterns similar to the severe cases in this study, which could provide valuable insights into patient stratification and treatment approaches.

Response: This is a good point, thank you. Based on our data it is unlikely that patients with IFN-I autoantibodies or genetic defects in the production of IFN-I would show significant neutrophil inflammasome activation. Like with IFN-I response in general, neutrophil inflammasomes can probably be either protective or damaging to the host, depending on the context and durability of the response. We have discussed this topic further in the revised manuscript, in response to this concern and the first point of reviewer 1.

Revised manuscript, discussion section, lines 639-648:

“The dualistic nature of the IFN-I response in COVID-19 has been recognized previously. It seems that a strong initial IFN-I response to SARS-CoV-2 is more likely to result in asymptomatic or mild COVID-19 whereas a decreased initial IFN-I activity, due to e.g. genetic defects or increased levels of IFN-I autoantibodies, can lead to more severe COVID-19 (61). This initial beneficial effect of IFN-I is probably due to its ability to limit viral replication at early stages of the infection. However, at later stages of the disease IFN-I can be detrimental by promoting inflammatory pathways instead of direct antiviral effects (62). Thus, similarly to the IFN-I response in general, the role of neutrophil inflammasomes in development and severity of COVID-19 might be dualistic in nature with an initial protective effect while damaging when sustained for prolonged periods.” 

- The authors need to delve deeper into whether the inflammasome patterns observed in severe COVID-19 cases are a contributing factor to the disease's progression or a result of the infection's severity, thereby clarifying the causality in their discussion.

Response: This is an important aspect of the study and we agree that the relationship between neutrophil inflammasome activity and disease severity could be highlighted better. However, understanding causality by analyzing clinical patient samples is difficult due to the typical lack of patient samples from the early phase of the disease. Hospitalization and thereby patient sample collection typically occurs when the patients are already experiencing the peak of symptomatic phase of the disease. To highlight the link between neutrophil inflammasome formation and disease severity more clearly, we have marked the statistically significant correlations between inflammasome activation and disease severity parameters by an asterisk in the correlation plot (Fig. 5) and added text in the discussion accordingly.

Revised manuscript, discussion section, lines 649-654:

“Our study demonstrated a strong association between PMN caspase1 activity and plasma levels of calprotectin, a marker of neutrophil activation. Additionally, increased disease severity, as assessed by the WHO ordinal scale, was significantly linked to PMNs being less responsive to ex vivo IFN-induced inflammasome activation, which is suggestive of prior in vivo inflammasome activation. Thus, these results suggest that neutrophil inflammasomes play a potential role in disease severity rather than being protective in COVID-19.” 

This point links to the previous comment by the reviewer regarding the role of IFN-I in disease severity. We have added text in the discussion highlighting the dual nature of the IFN-I response and neutrophil inflammasome activation in COVID-19 disease severity (see response above).     

-Authors should clarify if and how their findings may lead to any therapeutic advantage for severe COVID-19 patients.

Response: The text has been adapted accordingly, and more details have been added into the revised manuscript.

Revised manuscript, discussion section, lines 685-687:

“For example, pharmacologically targeting the inflammasome pathway in neutrophils with novel inhibiting molecules, may help mitigate the exaggerated inflammatory response observed in severe cases.”

Reference added: Mangan, M. S., Olhava, E. J., Roush, W. R., Seidel, H. M., Glick, G. D., & Latz, E. (2018). Targeting the NLRP3 inflammasome in inflammatory diseases. Nature reviews Drug discovery, 17(8), 588-606.

Minor points:

- Authors should increase the font in all the figures as most of them are difficult to be read

Response: Thank you for your suggestion. The font sizes have been increased, but their size will ultimately depend on the requirements of the journal where is the work will be published.

Thank you for your suggestion. The font sizes have been increased, but their size will ultimately depend on the requirements of the journal where the work will be published.

- Authors should better clarify how the choice for statistics tests was conducted "depending on sample distribution and the number of groups analyzed" in the Methods.

Response: We added additional details to the methods.

Revised manuscript, methods section, lines 329-331:

“To elaborate, nonparametric tests like Mann-Whitney and Kruskall-Wallis were employed when the data violated assumptions of normality, while ANOVA tests were applied when the data met parametric assumptions”. 

Referee Cross-Commenting*

-I agree with Reviewer#1 with adding relevant references about the evolution of some viruses, including SARS-COV-2, in evading type I IFN response.

Response: We added this information into the manuscript.

Revised manuscript, discussion section, lines 635-638:

“It should be noted that several SARS-CoV-2 encoded proteins have been shown to inhibit IFN-I signaling. However, no evidence suggests that neutrophils can be infected by SARS-CoV-2 and therefore it seems unlikely that such direct virus mediated effects could play a role in the observed neutrophil unresponsiveness to IFN-I".   

Reference added: Rashid F, et al. Roles and functions of SARS-CoV-2 proteins in host immune evasion [preprint]. Front Immunol. 2022;13.

-I agree with Reviewer#1 about Figure 8 not delivering a novel message. I also suggest to move this to the supplementary section.

Response: We have moved the figure plate to the Supplements. 

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

Please include a point-by-point response explaining why some of the requested data or additional analyses might not be necessary or cannot be provided within the scope of a revision. This can be due to time or resource limitations or in case of disagreement about the necessity of such additional data given the scope of the study. Please leave empty if not applicable.

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 of this study seek to examine the role of neutrophils, particularly their inflammasome formation potential, in the pathophysiology of COVID-19. They conducted transcriptomic profiling of neutrophils from both mild and severe COVID-19 patients, as well as SARS-CoV-2 infected mice, and compared these profiles to non-infected healthy controls. Their analysis confirmed a prominent IFN-I gene signature in severe cases previously reported by others. Furthermore, they observed that neutrophils from severe COVID-19 patients have an altered response to inflammasome activation and that IFN-I can serve as a priming stimulus for neutrophil inflammasomes, a finding further supported by a COVID-19 mouse model. Last, in this study the authors revealed that in severe COVID-19, LDGs show gene upregulation indicating immaturity, while PMNs exhibit enhanced pathogen-responsive NLR signaling.

Major points:

• Figure 2A: for this RNAseq analysis, the authors claim that "This analysis also demonstrated that cells in the LDG fraction were predominantly immature neutrophils, meanwhile PMNs were composed of mainly mature neutrophils (Figure 2A)". Nevertheless, there are 2 samples in the "Severe COVID-19" group that show a fraction of neutrophils {less than or equal to}0.6, which indicates low levels of purity. In addition, there is one "PMN" sample with high LDG fraction (around 40%). Authors should remove from this analysis samples with such a low purity since the big fraction ({greater than or equal to}40%) of contaminating cells could introduce a bias in this group. Are the differences observed still present in the absence of these low-purity samples?

• While the in vitro data suggest that IFN-I may prime the inflammasome response in neutrophils, in vivo evidence remains inconclusive. The systemic blockade of IFNAR1 with antibodies in infected mice does not confirm that IFN-I directly primes neutrophil inflammasomes, as other cells could initially sense IFN-I and subsequently produce neutrophil-activating stimuli. In the absence of in vivo experiments utilizing conditional IFNAR1 knockout models, such as Mrp8-Cre x IFNAR1 fl/fl mice, the authors should consider moderating the stated significance of these findings in the discussion about the limitations of the study.

• The authors are encouraged to specify the number of independent experiments conducted. For mouse model studies, it is recommended to include results that are either representative of, or aggregated from, a minimum of two independent experiments to ensure robustness of the data

• The authors are advised to employ more quantitative methods, such as flow cytometry, to measure neutrophil recruitment in mice. Additionally, it should be clarified how many tissue sections from each mouse were assessed for every experimental condition to ensure the reproducibility and statistical validity of the results.

• The authors need to address discrepancies in their text regarding the effects of anti-IFNAR1 blockade on viral titers and neutrophil recruitment in SARS-CoV-2 infected mice. While they state there is no change, Supplementary Figure 5D suggests increased SARS-CoV-2 NP staining with anti-IFNAR1 treatment, and there appears to be a lack of quantitative data on lung neutrophils to substantiate the claim that neutrophil recruitment remains unaffected. It is necessary for the authors to provide a more detailed explanation or additional data to resolve these inconsistencies

• The authors should demonstrate the effectiveness of anti-IFNAR1 blockade in mice by providing evidence of sustained inhibition of IFN-I signaling throughout the duration of the experiment to validate the treatment protocol used.

• Given the reports of life-threatening COVID-19 infections occurring in conjunction with autoantibodies against type I IFNs (DOI: 10.1126/science.abd4585), the authors should explore how this intersects with their findings. A discussion is needed on whether patients with such autoantibodies may exhibit inflammasome activation patterns similar to the severe cases in this study, which could provide valuable insights into patient stratification and treatment approaches

• The authors need to delve deeper into whether the inflammasome patterns observed in severe COVID-19 cases are a contributing factor to the disease's progression or a result of the infection's severity, thereby clarifying the causality in their discussion.

• Authors should clarify if and how their findings may lead to any therapeutic advantage for severe COVID-19 patients.

Minor points:

• Authors should increase the font in all the figures as most of them are difficult to be read

• Authors should better clarify how the choice for statistics tests was conducted "depending on sample distribution and the number of groups analyzed" in the Methods.

Referee Cross-Commenting

• I agree with Reviewer#1 with adding relevant references about the evolution of some viruses, including SARS-COV-2, in evading type I IFN response.

• I agree with Reviewer#1 about trying to dissect the role of IFNAR-1 vs IFNAR-2. Authors partially look at this in the in vivo mice experiments using the anti-IFNAR1 blocking Ab, but it would reinforce the study to see if this holds with human cells.

• I agree with Reviewer#1 about Figure 8 not delivering a novel message. I also suggest to move this to the supplementary section.

Significance

• The study presents an investigation into the role of neutrophils and inflammasome formation in COVID-19 pathology, contributing to the field with transcriptomic profiling of neutrophils from varying severity of patient cases and a SARS-CoV-2 mouse model. A significant IFN-I gene signature in severe cases was confirmed, and differences in inflammasome response were identified, adding to our understanding of disease mechanisms.

• Strengths of the paper include the comprehensive analysis of neutrophil maturation states and the novel insights into the priming of inflammasome activation by IFN-I. However, limitations were noted in the purity of samples for RNAseq analysis and the lack of conclusive in vivo evidence for the direct role of IFN-I in neutrophil inflammasome priming. The study's implications suggest potential avenues for targeted therapies, but the authors were advised to moderate their conclusions without stronger in vivo evidence and to clarify the potential therapeutic implications of their findings.

• Additional suggestions for improvement include the use of more quantitative methods like flow cytometry for neutrophil recruitment measurements, clarification on experimental replication, and resolving discrepancies in data presentation regarding anti-IFNAR1 blockade effects. Furthermore, the paper would benefit from discussing the relevance of autoantibodies against IFN-I in the context of their findings and from exploring the causal relationship between inflammasome patterns and disease severity.

• The study is of relevance to immunologists specializing in viral infections and researchers focused on neutrophil 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

• IFN-I promotes immune responses to diverse viruses. Some of them have evolved to dampen IFN-I responses in order to weaken or delay antiviral responses. A number of studies support the idea that this could be the case of SARS-CoV-2 (0.1038/s41586-022-04447-0, 10.1016/j.it.2021.02.003, 10.1038/s41467-022-34895-1, to cite a few). The authors should refer to some of these studies to counterbalance the statements of increased IFN-I response in severe COVID-19 patients in some parts of the manuscript, such as in the Discussion.

• In Fig 5, the authors stimulate PMNs from healthy controls in vitro with IFN-I and use mainly IL-1b as a readout for neutrophil priming. The authors should analyse whether IFN-I-mediated priming ultimately leads to NETosis, given the relevance of NETs to COVID-19 pathology (10.1016/j.tips.2023.06.007), which is acknowledged by the authors in the Introduction. It would also be interesting to explore the involvement of IFN-I receptors (IFNAR1 vs IFNAR2) by dissecting IFN-a from IFN-b responses.

• In Methods (Histology and immunohistochemistry), the authors mention that histone H3 is a NET marker and use this in Supp Fig 5 to support evidence of NETosis. The authors should state which antibody clone/company was used. Histone H3 is expressed in high levels by non-NETotic neutrophils. Citrullinated histone H3 (CitH3), on the other hand, is detectable by few commercially available antibodies and can be used as a NET marker in conjunction with other markers such as DNA staining (10.1084/jem.20201129).

• In Fig 8, the authors show that neutrophils migrate to the lungs of infected animals. This finding is showed in many previous studies and does not seem to favour the structure of the manuscript. Instead, it seems it would fit a Supp Fig better or a portion of the following figure.

Significance

By conducting transcriptomics analyses, Cabrera LE et al. elucidate the inflammasome formation in neutrophils during SARS-CoV-2 infection. The authors have found a strong IFN-I signature in neutrophils from severe COVID-19 patients and show that IFN-I functions as a priming stimulus for neutrophil inflammasome.

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

---

1. General Statements

We thank the reviewers for their excellent work that greatly improved our work. We are very content that reviewer #1 considered our work to be “novel, interesting and important for understanding the mitochondrial biology of PD”. This reviewer also valued our work as “a significant advancement” and suggested further study of the relationship of CISD1 (dimerization) to general mitophagy/autophagy. We addressed this in the already transferred revision (version 1, v1).

Also reviewer #2 considered our work to be “an exciting and well-executed piece of research focusing on the defects in iron homeostasis observed in Parkinson's disease which a wide audience will appreciate”. This reviewer had a very specific suggestion on how to improve our manuscript which makes a lot of sense and is feasible. As the suggested experiments include fly breeding and behavioral analysis, these experiments will be included in the second revision to be uploaded as soon as possible (version 2, v2).

Finally, reviewer #3 gathered that parts of our results “are confirmatory to recently published work” but also appreciated that our results established that iron-depleted apo-Cisd is an important determinant of toxicity which has not been shown before. I would like to comment here, that in contrast to the paper mentioned by this reviewer, our contribution includes data from dopaminergic neurons obtained from human patients suffering from familial Parkinson’s disease that demonstrate the same increase in apo-Cisd levels as the flies. This reviewer mainly suggested that the manuscript would be improved by a more balanced discussion of the strengths and weaknesses of the study and more circumspection in interpretation of data which we did in the revised version of our manuscript. We also added data on the expression levels of Cisd and apo-Cisd in transgenic flies as also suggested.

2. Description of the planned revisions

Reviewer #1 (Evidence, reproducibility and clarity):

Summary: The manuscript focuses on mitochondrial CISD1 and its relationship to two Parkinson's disease (PD) proteins PINK1 and Parkin. Interestingly, CISD1 is a mitochondrial iron sulfur binding protein and an target of Parkin-mediated ubiquitinylation. Disruption of iron metabolism and accumulation of iron in the brain has long since been reported in PD but the involvement of iron sulfur binding is little studied both in vivo and in human stem cell models of PD. This work addresses the relationship between CISD1 and two mitochondrial models of PD (PINK1 and Parkin) making use of in vivo models (Drosophila), PINK1 patient models (iPSC derived neurons) and Mouse fibroblasts. The authors report a complex relationship between CISD1, PINK1 and Parkin, where iron-depleted CISD1 may illicit a toxic gain of function downstream of PINK1 and Parkin.

Major comments:

The conclusions are overall modest and supported by the data. One question remains unaddressed. Is mitochondrial CISD1 a downstream target that specifically mediates PINK1 and Parkin loss of function phenotypes or are the phenotypes being mediated because CISD1 is downstream of mitophagy in general?

It would be interesting to know what happens to CISD1 (dimerization?) upon initiation of mitophagy in wild type cells? Would dissipation of mitochondrial membrane potential be sufficient to induce changes to CISD1 in wild type cells or PINK1 deficient cells? Since iron chelation is a potent inducer of mitophagy (Loss of iron triggers PINK1/Parkin-independent mitophagy. George F G Allen, Rachel Toth, John James, Ian G Ganley. EMBO Reports (2013)14:1127-1135) it would be useful to show one experiment addressing the role of CISD1 dimerization under mitochondrial depolarizing and non-depolarizing conditions in cells.

Based on the overall assumption of the reviewer that our work is “novel, interesting and important for understanding the mitochondrial biology of PD” and “a significant advancement” we understand the word “modest” here as meaning “not exaggerated”. To address this question, we studied CISD1 dimerization in response to more classical activators of mitophagy namely FCCP and antimycin/oligomycin which had no significant effect on dimerization suggesting that this phenotype is more pronounced under iron depletion. These data are shown in the new Fig. 2c.

Alternatively, the authors should discuss the topic of mitophagy (including PINK1-parkin independent mitophagy), the limitation of the present study not being able to rule out a general mitophagy effect and previous work on the role of iron depletion on mitophagy induction in the manuscript.

The data and the methods are presented in such a way that they can be reproduced.

The experiments are adequately replicated and statistical analysis is adequate.

Minor comments:

Show p values even when not significant (ns) since even some of the significant findings are borderline < p0.05.

Here, I decided to leave it as it is, because the figures became very cluttered and less easy to understand. Borderline findings are however indicated and mentioned in the text.

Because the situation for CISD1 is complicated (overexpression, different models etc.) it would be helpful if in the abstract the authors could summarize the role. E.g. as in the discussion that iron-depleted CISD1 could represent a toxic function.

The abstract has been completely rewritten and now mentions the potential toxic function of iron-depleted CISD1.

If there is sufficient iron (accumulation in PD) why would CISD1 be deactivated? Perhaps that could be postulated or discussed in a simplified way?

We actually think that apo-CISD1 without its iron/sulfur cluster is incapable of transferring its Fe/S cluster to IRP1 and IRP2. This then results in increased levels of apo-IRP1/2 and subsequent changes that lead to iron overload. Such a sequence of events would place CISD1 upstream of the changes in iron homeostasis observed in PD and models of PD. This is now discussed in more detail.

In the methods section both reducing and non-reducing gel/Western blotting is mentioned but the manuscript only describes data from blots under reducing conditions. Are there blots under non-reducing conditions that could be shown to see how CISD1 and dimerized CISD1 resolve?

We now show these blots as supplemental data in new supplemental Figure 2.

In the results section, PINK1 mutant flies, it is said that the alterations to CISD1 (dimerization) are analogous to the PINK1 mutation patient neurons. The effect is seen in old but not young flies. Since iPSC-derived neurons are relatively young in the dish, would one not expect that young flies and iPSC-derived neurons have similar CISD1 phenotypes? Could the authors modify the text to reflect that? or discuss the finding in further context.

We only studied one time point in PINK1 mutation patient neurons and controls. It would indeed be interesting whether neuronal aging (as far as this can be studied in the dish) would result in increased CISD1 dimerization. This is now discussed.

Reviewer #1 (Significance):

The strengths of this work are in the novelty of the topic and the use of several well established in vivo and cell models including patient-derived neurons. The findings discussed in the text are honest and avoid over-interpretation. The findings are novel, interesting and important for understanding the mitochondrial biology of PD.

We thank the reviewer for their kind words.

Limitations include the lack of strong phenotypes in the CISD1 models and the lack of robust, sustained and consistent increase in CISD1 dimers in the patient and fly models (just significant because of variability). The relationship of CISD1 (dimerization) to general mitophagy/autophagy is not shown here.

We do not completely agree with the assumption that all CISD1 models lack a strong phenotype. At least the CISD1-deficient fibroblasts exhibit a strong phenotype consisting of fragmented mitochondria and increased oxidative stress. The lack of a strong phenotype in Cisd-deficient flies could actually hint to a potential compensatory mechanism that could also protect the Pink1 mutant x Cisd-deficient double-knockout flies. It is correct that the increase in CISD1/Cisd dimers in the PD models are not overwhelming but – as also mentioned by the reviewer – this could be increased in “older” cultures. This is now discussed in more detail. As suggested by the reviewer, we have now added experiments that study the relationship between CISD1 dimerization and conventional mitophagy as described above.

There is a significant advancement. So far researchers were able to describe the importance of iron metabolism in PD (For example refer to work from the group of Georg Auburger such as PMID 33023155 and discussion of therapeutic intervention such as reviewed by Ma et al. PMID: 33799121) but few papers describe involvement of iron sulfur cluster proteins specifically (such as Aconitase) in relation to PINK1 and parkin (these are cited). The fact that CISD1 is a protein of the mitochondrial outer membrane makes it particularly interesting and further studies looking more closely at the interaction of CISD1 with mitochondrial proteins associated with PD will be of interest.

We thank the reviewer for pointing out these excellent publications. Key et al present an enormous wealth of data on protein dysregulation of wildtype and Pink1-/- fibroblast cell lines upon perturbation of the iron homeostasis (Key et al, 2020). Both cell lines exhibit a downregulation of CISD1 levels upon iron deprivation with the agent 2,2′ -Bipyridine possibly as a compensatory mechanism to limit the toxic gain of function of iron-depleted CISD1. The other paper, Ma et al. is a recent review on changes in iron homeostasis in PD and PD models (Ma et al, 2021). Both papers are now cited in the manuscript.

This paper describes CISD1 as a new and relevant player in PINK1 and Parkin biology. Further work could lead to exploration of whether CISD1 could be a therapeutic target, considering its role in maintaining mitochondrial redox and mitochondrial health. This is of particular interest to mitochondrial biologists and pre-clinical research in PD.

This preprint was reviewed by three scientists whose research focus in the mitochondrial biology underlying Parkinson's disease. The group has a special interest in the functions of the mitochondrial outer membrane. We work with several cell models of Parkinson's disease and work with patient donated samples. We do not have expertise in Drosophila models of PD nor the quantification of iron described in the manuscript.

Reviewer #2 (Evidence, reproducibility and clarity):

Summary: In the paper entitled 'Mitochondrial CISD1 is a downstream target that mediates PINK1 and Parkin loss-of-function phenotypes', Bitar and co-workers investigate the interaction between CISD1 and the PINK1/Parkin pathway. Mutations in PINK1 and PARKIN cause early onset Parkinson's disease and CISD1 is a homodimeric mitochondrial iron-sulphur binding protein. They observed an increase in CISD1 dimer formation in dopaminergic neurons derived from Parkinson's disease patients carrying a PINK1 mutation. Immuno-blots of cells expressing CISD1 mutants that affects the iron sulphur cluster binding and as well as cells treated with iron chelators, showed that the tendency of CISD1 to form dimers is dependent on its binding to iron-sulphur clusters. Moreover, the Iron-depleted apo-CISD1 does not rescue mitochondrial phenotypes observed in CISD1 KO mouse cells. Finally, In vivo studies showed that overexpression of Cisd and mutant apo-Cisd in Drosophila shortened fly life span and, using a different overexpression model, apo-Cisd caused a delay in eclosion. Similar as patient derived neurons, they observed an increase in Cisd dimer levels in Pink1 mutant flies. Additionally, the authors showed that double mutants of Cisd and Pink1 alleviated all Pink1 mutant phenotypes, while double mutants of Prkn and Cisd rescued most Prkn mutant phenotypes.

Major comments:

1) The authors observed an increase in the levels of Cisd dimers in Pink1 mutant flies and removing Cisd in Pink1 mutant background rescues all the mutant phenotypes observed in Pink1 mutant flies, suggesting that the Cisd dimers are part and partial of the Pink1 mutant phenotype. The authors also generated a UAS_C111S_Cisd fly which can overexpress apo-Cisd. Overexpression of the C111S_Cisd construct with Tub-Gal4 showed a developmental delay. Since apo-Cisd forms more dimeric Cisd, my question is: does the strong overexpression (e.g. with Tub-Gal4) of the C111S_Cisd in wild type flies shows any of the Pink1 mutant phenotypes? If not, the authors should mention this and elaborate on it.

We thank the reviewer for their comments. In fact, we only observed very few flies ecclosing after overexpression of wildtype Cisd or C111S Cisd using the strong tubP-Gal4 driver during development. We considered these very few flies to be escapees (also indicated by the rather low induction of Cisd mRNA suggesting compensatory downregulation) and only used them to conduct the analysis shown in Figure 4c-e. This is now mentioned in more detail in the manuscript.

2) Figure 6g: Shows the TEM pictures of the indirect flight muscles of Pink1 mutant flies and Pink1, Cisd double mutants. To me, the Picture of Pink1 mutant mitochondria is not very convincing. We expect swollen (enlarged) mitochondria with disrupted mitochondrial matrix. However, this is not clear in the picture. Moreover, in my opinion, Figure6 g, is missing an EM Picture of the Cisd mutant indirect flight muscles.

We now show exemplary pictures from Pink1 mutant and DKO in a higher magnification which better demonstrate the rounded Pink1 mutant mitochondria and the disrupted cristae structure. EM pictures of all four genotypes in different magnifications are now shown in new supplemental Figure 6.

3) OPTIONAL: The authors suggest that most probably apo-Cisd, assumes a toxic function in Pink1 mutant flies and serves as a critical mediator of Pink1-linked phenotypes. If this statement is correct, we can hypothesize that increasing apo-Cisd in Pink1 mutant background should worsen the pink1 mutant defects.

Therefore, I suggest overexpressing Cisd1 wild type (and/or C111S Cisd) in pink1 mutant flies, as pink1 is on the X chromosome, and mild overexpression of Cisd1 with da is not lethal, these experiments could be done in 3-4 fly crosses and hence within 1.5 - 2 months.

We have set up this experiment and will report in the second revision (v2) of our manuscript.

Since Pink1 mutant flies contain higher levels of endogenous Cisd dimers, we can expect that overexpression of wild type Cisd will result in an even stronger increase of dimers. If these dimers indeed contribute to Pink1 mutant phenotypes we can expect that overexpression of Cisd will result in a worsening of the Pink1 mutant phenotypes.

We have set up this experiment and will report in the second revision of our manuscript.

Minor Comments:

-) In the Introduction (Background) there are some parts without references:

E.g., there is not a single reference in the following part between

'However, in unfit mitochondria with a reduced mitochondrial membrane potential ...&... compromised mitochondria safeguards overall mitochondrial health and function.'

We thank the reviewer for pointing out this flaw. We have now added a suitable reference to the introduction.

-) In the introduction there is some confusion about the nomenclature used in the article: e.g. following comments are made in the text: Cisd2 (in this publication referred to as Dosmit) or fly Cisd2 (in this publication named MitoNEET).

However, the names Dosmit and MitoNEET do not appear in the manuscript (except in references)

The literature and nomenclature for CISD1 are indeed confusing. We have now revised the introduction.

-) Figure 1: I am not sure why some gels are shown in this figure. The two last lanes of figure 1c are redundant and Figure 1c' which is also not mentioned in the text, is also a repetition of figure 1c.

The blots in 1c and 1c’ represent all data points (different patients and different individual differentiations) shown in the quantification in 1d. This is now explained better in the revised manuscript.

-) The authors mention in material and methods that T2A sites are used at the C-terminus of CISD1 to avoid tagging of CISD1. However, this is not entirely true as T2A will leave some amino acids (around 20) after the self-cleaving and therefore CISD1 will be tagged.

This is indeed true and we have now changed the wording in the revised manuscript.

-) In figure 5 P1 is used to abbreviate Pink1 mutants, however P1, to me, refers to pink1 wild type. It would be clearer to abbreviate Pink1 mutants as P1B9 in the graphs as B9 is the name of the mutant pink1 allele.

We thank the reviewer for pointing out this flaw. We have now altered Fig. 5 to be clearer.

-) In figure 7: Parkin is abbreviated both as Prkn and as Park

We thank the reviewer for pointing out this flaw, we indeed mixed up both names because it is complicated. The gene symbol is Prkn, the fly line is called Park25. We have now clarified this in the text and Fig. 7.

-) I suggest changing the title. Recently an article (Ham et al, 2023 PMID: 37626046) was published showing similar genetic interactions between Pink1/Prkn and Cisd. However, the article of Ham et al, 2023 was focused on Pink1/Prkn regulation of ER calcium release, while this article is more related to iron homeostasis. I suggest that the title shows this distinction.

This is indeed a very good suggestion. We have now altered the title to “Iron/sulfur cluster loss of mitochondrial CISD1 mediates PINK1 loss-of-function phenotypes”.

Reviewer #2 (Significance):

In general, this is an exciting and well-executed piece of research focusing on the defects in iron homeostasis observed in Parkinson's disease which a wide audience will appreciate. Very recently, a similar genetic interaction between Cisd and Pink1/Prkn in flies was published (Ham et al, 2023 PMID: 37626046) however, from a different angle. While, Ham et al focused on the role of Pink1/Prkn and Cisd in IP3R related ER calcium release, this manuscript approaches the Pink1/Prkn - Cisd interaction from an iron homeostasis point of view. Since, iron dysregulation contributes to the pathogenesis of Parkinson's disease, the observations in this manuscript are relevant for the disease. Hence, the work is sufficiently novel and deserves publication. However, additional experiments are suggested to strengthen the authors' conclusions.

We thank the reviewer for their kind words. As mentioned above, these additional experiments are on their way and will be included in version 2 of our revised manuscript (v2).

I work on Drosophila models of Parkinson's disease

Referees cross-commenting

I agree with the reviewer number 1 that it would be interesting to investigate CISD1 dimerisation status during mitophagy.

As mentioned above, we now studied CISD1 dimerization in response to more classical activators of mitophagy namely FCCP and antimycin/oligomycin which had no significant effect on dimerization suggesting that this phenotype is more pronounced under iron depletion. These data are shown in the new Fig. 2c.

Reviewer #3 (Evidence, reproducibility and clarity):

Here the authors provide evidence that Cisd is downstream of Parkin/Pink1 and suggest that the levels of apo-Cisd correlate with neurotoxicity. The data presented generally supports the conclusions of the authors and will be useful to those in the field. The manuscript would be improved by a more balanced discussion of the strengths and weaknesses of the study and more circumspection in interpretation of data.

We thank the reviewer for their comments aimed to improve our manuscript. We have now discussed the strengths and weaknesses of our study in more detail.

Introduction. While iron has been implicated in Parkinson's disease, it is an overstatement to say that disruption in iron metabolism contributes significantly to the pathogenesis of the disease.

There is certainly a plethora of data implicating perturbed iron homeostasis in PD as also pointed out by reviewer #1. We have tried to tone down our wording in the text and added a recent review on the topic (Ma et al, 2021) as also suggested by reviewer #1.

Introduction. The discussion of the various names for Cisd2 is important, but confusing as written. Specifically, the use of "this" makes the wording unclear.

We thank the reviewer for pointing out this flaw. We have altered the wording in the introduction.

Methods. It would be preferable to use heterozygous driver lines or a more similar genetic control rather than w-1118.

The exact controls were indeed not well explained in the Methods section, this has been corrected in the revised version. In brief, homozygous driver and UAS lines were indeed used in Fig. 4, this will be addressed in the second revision of our manuscript together with the experiments reviewer #2 suggested. The data shown in Fig. 5, 6, and 7 all used w1118 as control because all other fly strains are on the same genetic background.

Page 10. It appears that the PINK1 lines have been described previously. The authors should clarify this point and ensure that the new data presented in the current manuscript (presumably the mRNA levels, Fig. 1a) is indicated, as well as data that is confirmatory of prior findings (Fig. 1b).

Yes, these PINK1 lines have been described previously as pointed out in the manuscript. The original paper did not quantify the PINK1 mRNA levels shown in Fig. 1a. The blots shown in Fig. 1b are from new differentiations and have also not been shown before but confirm findings published in Jarazo et al. (Jarazo et al, 2022). This has been clarified in the revised version of our manuscript.

Fig. 3 legend. There is a typographical error, "ne-way ANOVA."

We thank the reviewer for pointing out this flaw. This has been corrected in the revised version.

Page 15. The nature of the Pink1-B9 mutant should be specified.

We now added a supplemental Figure 1 that depicts the specific mutation in these flies.

Fig. 4. Levels of mutant and wild type Cisd should be compared in transgenic flies.

We now added a quantification of mutant and wildtype Cisd levels to the new Figure 4d.

Fig. 5b,d. The striking change seems to be the decrease in dimers in young Pink1 mutant animals, not the small increase in dimers in the older Pink1 mutants.

It is always difficult to find a “typical” picture that reflects all changes observed in quantitative data. This Figure actually shows a decrease of total Cisd levels in young flies in Fig. 5c but no difference of the dimer/monomer ratio in Fig. 5d.

Fig. 5f. Caution should be used in interpreting the results. Deferiprone has toxicity to wildtype flies (trend) and may simply be making sick Pink1 mutants sicker.

There is certainly a tendency for wildtype flies to thrive less in food containing deferiprone. To make this more obvious, we have now added the exact p value (0.0764, which we don’t consider borderline but a tendency) to this figure and mention this fact in the text.

Fig. 5e. The data are hard to interpret. The number of animals is very small for a viability study and the strains are apparently in different genetic backgrounds, though this is not clearly specified. The experiment in Supplementary Fig. 1 appears better controlled and supports the Pink1 data; however, a similar concern pertains to Fig. 7. The authors may thus wish to be more circumspect in their interpretation, especially of the Parkin data.

In Fig 5e we quantified total iron levels and the Fe3+/Fe2+ ratio using capillary electrophoresis-inductively coupled plasma mass spectrometry (CE-ICP-MS). Although indeed not so many flies were used in this quantification, the results are highly significant. If the reviewer was referring to Fig. 5f, we agree that this experiment was not well (to be honest, even wrongly explained) which we corrected in the revised version of this manuscript. We thank the reviewer for pointing out this flaw.

Reviewer #3 (Significance):

The major significance of the study is in putting downstream of Parkin/Pink1 (largely confirmatory to recently published work) and suggesting that the levels of apo-Cisd are an important determinant of toxicity. The work will be of interest to those in the field.

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

The changes already carried out and included in the transferred manuscript (v1) are indicated above in bold orange. All changes pending on ongoing experiments to be included in the second revision of the manuscript are indicated above in bold magenta.

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

All changes suggested by the reviewers were addressed (v1) or will be addressed (v2).

References

Jarazo J, Barmpa K, Modamio J, Saraiva C, Sabaté-Soler S, Rosety I, Griesbeck A, Skwirblies F, Zaffaroni G, Smits LM, et al (2022) Parkinson’s Disease Phenotypes in Patient Neuronal Cultures and Brain Organoids Improved by 2-Hydroxypropyl-β-Cyclodextrin Treatment. Mov Disord 37: 80–94

Key J, Sen NE, Arsović A, Krämer S, Hülse R, Khan NN, Meierhofer D, Gispert S, Koepf G & Auburger G (2020) Systematic Surveys of Iron Homeostasis Mechanisms Reveal Ferritin Superfamily and Nucleotide Surveillance Regulation to be Modified by PINK1 Absence. Cells 9

Ma L, Gholam Azad M, Dharmasivam M, Richardson V, Quinn RJ, Feng Y, Pountney DL, Tonissen KF, Mellick GD, Yanatori I, et al (2021) Parkinson’s disease: Alterations in iron and redox biology as a key to unlock therapeutic strategies. Redox Biol 41: 101896

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 the authors provide evidence that Cisd is downstream of Parkin/Pink1 and suggest that the levels of apo-Cisd correlate with neurotoxicity. The data presented generally supports the conclusions of the authors and will be useful to those in the field. The manuscript would be improved by a more balanced discussion of the strengths and weaknesses of the study and more circumspection in interpretation of data.

• Introduction. While iron has been implicated in Parkinson's disease, it is an overstatement to say that disruption in iron metabolism contributes significantly to the pathogenesis of the disease.

Introduction. The discussion of the various names for Cisd2 is important, but confusing as written. Specifically, the use of "this" makes the wording unclear.<br /> - Methods. It would be preferable to use heterozygous driver lines or a more similar genetic control rather than w-1118.<br /> - Page 10. It appears that the PINK1 lines have been described previously. The authors should clarify this point and ensure that the new data presented in the current manuscript (presumably the mRNA levels, Fig. 1a) is indicated, as well as data that is confirmatory of prior findings (Fig. 1b).<br /> - Fig. 3 legend. There is a typographical error, "ne-way ANOVA."<br /> - Page 15. The nature of the Pink1-B9 mutant should be specified.<br /> - Fig. 4. Levels of mutant and wild type Cisd should be compared in transgenic flies.<br /> - Fig. 5b,d. The striking change seems to be the decrease in dimers in young Pink1 mutant animals, not the small increase in dimers in the older Pink1 mutants.<br /> - Fig. 5f. Caution should be used in interpreting the results. Deferiprone has toxicity to wildtype flies (trend) and may simply be making sick Pink1 mutants sicker.<br /> - Fig. 5e. The data are hard to interpret. The number of animals is very small for a viability study and the strains are apparently in different genetic backgrounds, though this is not clearly specified. The experiment in Supplementary Fig. 1 appears better controlled and supports the Pink1 data; however, a similar concern pertains to Fig. 7. The authors may thus wish to be more circumspect in their interpretation, especially of the Parkin data.

Significance

The major significance of the study is in putting downstream of Parkin/Pink1 (largely confirmatory to recently published work) and suggesting that the levels of apo-Cisd are an important determinant of toxicity. The work will be of interest to those in the field.

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 the paper entitled 'Mitochondrial CISD1 is a downstream target that mediates PINK1 and Parkin loss-of-function phenotypes', Bitar and co-workers investigate the interaction between CISD1 and the PINK1/Parkin pathway. Mutations in PINK1 and PARKIN cause early onset Parkinson's disease and CISD1 is a homodimeric mitochondrial iron-sulphur binding protein. They observed an increase in CISD1 dimer formation in dopaminergic neurons derived from Parkinson's disease patients carrying a PINK1 mutation. Immuno-blots of cells expressing CISD1 mutants that affects the iron sulphur cluster binding and as well as cells treated with iron chelators, showed that the tendency of CISD1 to form dimers is dependent on its binding to iron-sulphur clusters. Moreover, the Iron-depleted apo-CISD1 does not rescue mitochondrial phenotypes observed in CISD1 KO mouse cells. Finally, In vivo studies showed that overexpression of Cisd and mutant apo-Cisd in Drosophila shortened fly life span and, using a different overexpression model, apo-Cisd caused a delay in eclosion. Similar as patient derived neurons, they observed an increase in Cisd dimer levels in Pink1 mutant flies. Additionally, the authors showed that double mutants of Cisd and Pink1 alleviated all Pink1 mutant phenotypes, while double mutants of Prkn and Cisd rescued most Prkn mutant phenotypes.

Major comments:

1. The authors observed an increase in the levels of Cisd dimers in Pink1 mutant flies and removing Cisd in Pink1 mutant background rescues all the mutant phenotypes observed in Pink1 mutant flies, suggesting that the Cisd dimers are part and partial of the Pink1 mutant phenotype. The authors also generated a UAS_C111S_Cisd fly which can overexpress apo-Cisd. Overexpression of the C111S_Cisd construct with Tub-Gal4 showed a developmental delay. Since apo-Cisd forms more dimeric Cisd, my question is: does the strong overexpression (e.g. with Tub-Gal4) of the C111S_Cisd in wild type flies shows any of the Pink1 mutant phenotypes? If not, the authors should mention this and elaborate on it.
2. Figure6 g: Shows the TEM pictures of the indirect flight muscles of Pink1 mutant flies and Pink1, Cisd double mutants. To me, the Picture of Pink1 mutant mitochondria is not very convincing. We expect swollen (enlarged) mitochondria with disrupted mitochondrial matrix. However, this is not clear in the picture. Moreover, in my opinion, Figure6 g, is missing an EM Picture of the Cisd mutant indirect flight muscles.
3. OPTIONAL: The authors suggest that most probably apo-Cisd, assumes a toxic function in Pink1 mutant flies and serves as a critical mediator of Pink1-linked phenotypes. If this statement is correct, we can hypothesize that increasing apo-Cisd in Pink1 mutant background should worsen the pink1 mutant defects.

Therefore, I suggest overexpressing Cisd1 wild type (and/or C111S Cisd) in pink1 mutant flies, as pink1 is on the X chromosome, and mild overexpression of Cisd1 with da is not lethal, these experiments could be done in 3-4 fly crosses and hence within 1.5 - 2 months.

Since Pink1 mutant flies contain higher levels of endogenous Cisd dimers, we can expect that overexpression of wild type Cisd will result in an even stronger increase of dimers. If these dimers indeed contribute to Pink1 mutant phenotypes we can expect that overexpression of Cisd will result in a worsening of the Pink1 mutant phenotypes.

Minor Comments:

1. In the Introduction (Background) there are some parts without references:<br /> E.g., there is not a single reference in the following part between<br /> 'However, in unfit mitochondria with a reduced mitochondrial membrane potential ...&... compromised mitochondria safeguards overall mitochondrial health and function.'
2. In the introduction there is some confusion about the nomenclature used in the article: e.g. following comments are made in the text: Cisd2 (in this publication referred to as Dosmit) or fly Cisd2 (in this publication named MitoNEET).<br /> However, the names Dosmit and MitoNEET do not appear in the manuscript (except in references)
3. Figure 1: I am not sure why some gels are shown in this figure. The two last lanes of figure 1c are redundant and Figure 1c' which is also not mentioned in the text, is also a repetition of figure 1c.
4. The authors mention in material and methods that T2A sites are used at the C-terminus of CISD1 to avoid tagging of CISD1. However, this is not entirely true as T2A will leave some amino acids (around 20) after the self-cleaving and therefore CISD1 will be tagged.
5. In figure 5 P1 is used to abbreviate Pink1 mutants, however P1, to me, refers to pink1 wild type. It would be clearer to abbreviate Pink1 mutants as P1B9 in the graphs as B9 is the name of the mutant pink1 allele.
6. In figure 7: Parkin is abbreviated both as Prkn and as Park
7. I suggest changing the title. Recently an article (Ham et al, 2023 PMID: 37626046) was published showing similar genetic interactions between Pink1/Prkn and Cisd. However, the article of Ham et al, 2023 was focused on Pink1/Prkn regulation of ER calcium release, while this article is more related to iron homeostasis. I suggest that the title shows this distinction.

Referees cross-commenting

I agree with the reviewer number 1 that it would be interesting to investigate CISD1 dimerisation status during mitophagy.

Significance

In general, this is an exciting and well-executed piece of research focusing on the defects in iron homeostasis observed in Parkinson's disease which a wide audience will appreciate. Very recently, a similar genetic interaction between Cisd and Pink1/Prkn in flies was published (Ham et al, 2023 PMID: 37626046) however, from a different angle. While, Ham et al focused on the role of Pink1/Prkn and Cisd in IP3R related ER calcium release, this manuscript approaches the Pink1/Prkn - Cisd interaction from an iron homeostasis point of view. Since, iron dysregulation contributes to the pathogenesis of Parkinson's disease, the observations in this manuscript are relevant for the disease. Hence, the work is sufficiently novel and deserves publication. However, additional experiments are suggested to strengthen the authors' conclusions.<br /> I work on Drosophila models of Parkinson's disease

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 focuses on mitochondrial CISD1 and its relationship to two Parkinson's disease (PD) proteins PINK1 and Parkin. Interestingly, CISD1 is a mitochondrial iron sulfur binding protein and an target of Parkin-mediated ubiquitinylation. Disruption of iron metabolism and accumulation of iron in the brain has long since been reported in PD but the involvement of iron sulfur binding is little studied both in vivo and in human stem cell models of PD. This work addresses the relationship between CISD1 and two mitochondrial models of PD (PINK1 and Parkin) making use of in vivo models (Drosophila), PINK1 patient models (iPSC derived neurons) and Mouse fibroblasts. The authors report a complex relationship between CISD1, PINK1 and Parkin, where iron-depleted CISD1 may illicit a toxic gain of function downstream of PINK1 and Parkin.

Major comments:

• The conclusions are overall modest and supported by the data. One question remains unaddressed. Is mitochondrial CISD1 a downstream target that specifically mediates PINK1 and Parkin loss of function phenotypes or are the phenotypes being mediated because CISD1 is downstream of mitophagy in general?<br /> It would be interesting to know what happens to CISD1 (dimerization?) upon initiation of mitophagy in wild type cells? Would dissipation of mitochondrial membrane potential be sufficient to induce changes to CISD1 in wild type cells or PINK1 deficient cells?

• Since iron chelation is a potent inducer of mitophagy (Loss of iron triggers PINK1/Parkin-independent mitophagy. George F G Allen, Rachel Toth, John James, Ian G Ganley. EMBO Reports (2013)14:1127-1135) it would be useful to show one experiment addressing the role of CISD1 dimerization under mitochondrial depolarizing and non-depolarizing conditions in cells.

• Alternatively, the authors should discuss the topic of mitophagy (including PINK1-parkin independent mitophagy), the limitation of the present study not being able to rule out a general mitophagy effect and previous work on the role of iron depletion on mitophagy induction in the manuscript.

• The data and the methods are presented in such a way that they can be reproduced.<br /> The experiments are adequately replicated and statistical analysis is adequate.

Minor comments:

• Show p values even when not significant (ns) since even some of the significant findings are borderline < p0.05.

• Because the situation for CISD1 is complicated (overexpression, different models etc.) it would be helpful if in the abstract the authors could summarize the role. E.g. as in the discussion that iron-depleted CISD1 could represent a toxic function.

• If there is sufficient iron (accumulation in PD) why would CISD1 be deactivated? Perhaps that could be postulated or discussed in a simplified way?

• In the methods section both reducing and non-reducing gel/Western blotting is mentioned but the manuscript only describes data from blots under reducing conditions. Are there blots under non-reducing conditions that could be shown to see how CISD1 and dimerized CISD1 resolve?

• In the results section, PINK1 mutant flies, it is said that the alterations to CISD1 (dimerization) are analogous to the PINK1 mutation patient neurons. The effect is seen in old but not young flies. Since iPSC-derived neurons are relatively young in the dish, would one not expect that young flies and iPSC-derived neurons have similar CISD1 phenotypes? Could the authors modify the text to reflect that? or discuss the finding in further context.

Significance

• The strengths of this work are in the novelty of the topic and the use of several well established in vivo and cell models including patient-derived neurons. The findings are discussed in the text are honest and avoid over-interpretation. The findings are novel, interesting and important for understanding the mitochondrial biologyof PD.
• Limitations include the lack of strong phenotypes in the CISD1 models and the lack of robust, sustained and consistent increase in CISD1 dimers in the patient and fly models (just significant because of variability). The relationship of CISD1 (dimerization) to general mitophagy/autophagy is not shown here.
• There is a significant advancement. So far researchers were able to describe the importance of iron metabolism in PD (For example refer to work from the group of Georg Auburger such as PMID 33023155 and discussion of therapeutic intervention such as reviewed by Ma et al. PMID: 33799121) but few papers describe involvement of iron sulfur cluster proteins specifically (such as Aconitase) in relation to PINK1 and parkin (these are cited). The fact that CISD1 is a protein of the mitochondrial outer membrane makes it particularly interesting and further studies looking more closely at the interaction of CISD1 with mitochondrial proteins associated with PD will be of interest.
• This paper describes CISD1 as a new and relevant player in PINK1 and Parkin biology. Further work could lead to exploration of whether CISD1 could be a therapeutic target, considering its role in maintaining mitochondrial redox and mitochondrial health. This is of particular interest to mitochondrial biologists and pre-clinical research in PD.
• This preprint was reviewed by three scientists whose research focus in the mitochondrial biology underlying Parkinson's disease. The group has a special interest in the functions of the mitochondrial outer membrane. We work with several cell models of Parkinson's disease and work with patient donated samples. We do not have expertise in Drosophila models of PD nor the quantification of iron described in the manuscript.

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 [optional]

__We thank all the reviewers for their time and their constructive criticism, based on which we will revise our manuscript. All our responses are indicated in red. __

2. Description of the planned revisions

Insert here a point-by-point reply that explains what revisions, additional experimentations and analyses are planned to address the points raised by the referees.

__Reviewer #1 (Evidence, reproducibility and clarity (Required)): __

The manuscript by Nguyen and Cheng is investigating the timing and mechanism of cessation of neuroblasts in the pupal optic lobe. Previous studies by several groups have determined the spatial and temporal factors required for the neuroepithelial to neuroblast transition and neuroblast to neural/glycogenesis in third instar larvae such that neuroblasts are eliminated. The mechanism of elimination of neuroblasts in the VNC or mushroom bodies have been investigated, but the mechanism(s) and the timing of elimination of medulla neuroblasts has not been investigated. The authors suggest that medulla neuroblasts are eliminated via a combination of mechanisms including apoptosis, prospero induced size symmetric terminal differentiation and a switch to gliogenesis by gcm expression. Expression of Tailless also was found to affect the timing of medulla neuroblast termination. They also ruled out several mechanisms such as ecdysone pulses.

Major comments

Clearly written and logical flow to experiments and results not over interpreted.

Clearly show that the neuroblast number and size decrease (12 to 18 hrs) and are eliminated by 30 hours

Figure 2a Marking of the Neuroepithelium. Would be more convincing if shown by PatJ expression and is clonal analysis. While the following panels use PatJ in clones suggesting are NE and NBs present it is more difficult to put into the context in the higher magnification images (Figure 2 D- M) and the Miranda expression in F' seems to be the entire lobe and it is not clear if would be any NE which does not agree with what is shown in panel A.

We will perform clonal analysis using MARCM to show that the elimination of medulla NBs (marked by Dpn) is accompanied by the depletion of NE (marked by PatJ). For Figure 2 D, E, I, L, we will change the images to the whole lobes to clearly show the shift in the NE-NB transition upon Notch OE/KD.

Is difficult to see the neuroblasts in Figure 2 D D" and E. The figure does not match what is stated in the results in that the neuroblasts are difficult to observe. If the point is that there is fewer NE cells and more neuroblasts then this is hard to see. It has been previously shown that with Notch RNAi clones prematurely extrude form the NE (Egger 20210; Keegan 2023) and could be expressing more Neuroblast markers but this is not visible in the panels as shown. Are the images single focal plane or maximum projections? Imaging more deeply in the brain or viewing in cross section would account for these possibilities. The possibility that are more neuroblasts but not all at the surface of the OL should be addressed as this could also alter the overall results.

Figure 2 is key to first point of the paper so needs to be addressed.

The images are single focal plane of the superficial layer of the medulla. We will specify this information in the figure legends. We will include cross-section of the notch RNAi clones to show the delamination of precocious NBs.

Minor comments

Why express volume of DPN in clone volume. Would make the point more clear and more strong be to express as number of NB in the 3-D volume of the clone. This measurement occurs in several figures.

We will redo the quantification as suggested.

Use of Miranda to mark NBs is unclear in Figure 2. Perhaps more clear in B&W.

We will redo the staining with Dpn instead of Mira to mark medulla NBs. Figures will be presented in B&W as suggested.

Make clear in figures (or figure legend) if single focal plane or projections.

We will do so.

It is unclear what percentage of NB the Gal4 line eyR16F10 are expressed in. Veen 2023 state that the GAL4 is also expressed in neurons and at different levels whether deeper within the brain or superficially on the surface of the brain. At 16 APF it is expressed but it is not clear whether it is in all cells at a low level or only within a few cells

We will further characterize the expression of eyR16F10-GAL4 in the pupal medulla as suggested.

Some RNAi lines referenced as previously validated and other are not. For example: EcR, Oxphos, Med27, Notch need references or confirmation of specificity to the intended target (qRT)

We will perform RT-qPCR to validate the use of UAS-med27 RNAi. For RNAi stocks such as UAS-EcR RNAi, UAS-Atg1 RNAi, UAS-notch RNAi that have been previously used in other publications, we will provide appropriate references.

At least 2 animals per genotype were used. While I appreciate the technical difficulty of working in pupae this seems a bit low in terms of number of samples and data would be more robust with more numbers.

Any experiments in which less than 3 animals were used, we will redo the experiments.

Reviewer #1 (Significance (Required)):

This provides mechanism and timing for the elimination of neuroblasts (NE to NB) that arise from the medulla. As these are most similar to mammalian brain development (Radial glial to NSC) this information provides more context to interpret the formation of glial and neurons in the adult optic lobe given the effect on timing and mechanisms of elimination.

This paper would be of interest to developmental biologist who work with Drosophila or mice who are looking at neural development. An understanding of how neural diversity is achieved and the mechanisms behind this that can be dysfunctional in terms of etiology of neural diseases. Is a well done study for the most part that would be improved by clarifying some data and provided more replicates for robustness of the data.

I am a developmental biologist working with Drosophila in larval and adult neural development.

__Reviewer #2 (Evidence, reproducibility and clarity (Required)): __

Lineages of neural stem cells are of great interest to understand how many neural types are generated. They produce very diverse neurons, often in a highly stereotyped series. However, they must terminate their life when the animal becomes functional or if neurons need time to become mature before birth.

In the Drosophila optic lobes, neural stem cells are produced over a period of several days by a wave of neurogenesis that transforms a neuroepithelium into neural stem cells that undergo a series of temporal patterning steps. It has been reported that they finish their life when a symmetric division generates glial cells. The authors however analyze the end of a particular lineage, that of the latest born neural stem cells of the medulla.

The paper shows that neural stem cells stop being produced when the neuroepithelium is consumed. But how do the latest born neural stem cells stop their lineage?

The results show that they do so by several means, which is quite unexpected: they may die from apoptosis, or autophagy, by becoming glioblasts or by a terminal symmetric division.

There are no major issues affecting the conclusions

• The paper shows that the end of production of neural stem cells occurs the neuroepithelium is completely transformed. The experiments performed by the authors are fine and show that, if the transition is delayed, neural stem cells terminate their life later, and vice versa. However, the lifespan of the neural stem cells is not affected by the timing of the transition. Therefore, these experiments do not tell us how neural stem cells terminate their life, which is the central question of the study. The discussion should be written accordingly and the title and the model in Fig 6 modified to reflect the importance of the end of life of the stem cells, the main theme of the paper.

We agree that our said experiments did not elucidate how NBs terminate at the end of neurogenesis. Nevertheless, our aim is to show that the timing of NB termination in the medulla is dependent on the timing of the NE-NB transition.

In Supplementary Figure 1, we showed that factors previously shown to be involved in NB termination in other lineages did not play similar roles in the medulla NBs. Thus, we think that NB termination in the medulla is likely regulated at the levels of the NE, but not the NBs themselves. Although we have briefly mentioned this in our manuscript, we hope by conducting the experiments suggested by the reviewer (see below), we can subsequently modify our model in Figure 6 and our discussion.

• The authors talk about Pros-dependent symmetric division and gliogenic switch as two separate processes, but these may be two sides of the same phenomenon. Tll+ gcm+ neural stem cells undergo Pros-dependent cell cycle exit, generating glial progeny. If the authors agree with this, could they update their model (and discussion) to reflect the fact that gliogenic switch occurs via a Pros-dependent symmetric division, and these are not two separate processes independently contributing to the depletion of the neural stem cell pool? Ideally, a triple staining between Dpn, Pros, and gcm would show that the symmetrically dividing cells seen by the authors are committed to the glial fate.

We will further test how gliogenesis is affected in pros RNAi clones. The results may shed light on whether Pros-mediated symmetric division is required for Gcm-mediated gliogenesis in the medulla. Regarding the model, we have summarized our findings and suggestions in Figure 5K, however, we will integrate this information into our final model.

In Figure 5C, we showed that at 12h APF, there are Dpn+ NBs in the medulla that expressed both Pros and Gcm, suggesting that it is very likely that Pros is upstream of Gcm to induce the glial cell fate switch of the medulla NBs.

• Why were Notch RNAi experiments assessed for the presence of neural stem cells at P12 and gcm RNAi experiments at P24? Given that most optic lobe neural stem cells disappear between P12-18, a subtle effect of gcm RNAi may have been missed. Do the authors have data for gcm RNAi at P12?

We hypothesized that the timing of NE-NB transition affects the timing of NB termination in the medulla. Because Notch KD was previously shown to induce precocious NE-NB transition in the OL, meaning that medulla NBs are born prematurely, we expected that this manipulation will lead to a corresponding premature elimination of the NBs. In contrast, gcm RNAi which inhibits the switch into the glial cell fate of the NBs, is expected to prolong the neurogenic phase of the NBs, and thereby, their persistence by 24h APF when WT NBs are eliminated.

• The authors should acknowledge that the inhibition of either apoptosis or autophagy alone may not be fully sufficient to prevent the death of NBs. In mushroom body neural stem cells, both processes must be inhibited simultaneously to produce a strong effect on their survival (Pahl et al. 2019, PMID 30773368).

We will add this information in our discussions.

• There is an important missing point that should be addressed: is there a specific point in time when all neural stem cells must stop their lineage wherever they are in the temporal series and either die or divide symmetrically? One possibility that is not discussed is that most neural stem cells end their life through a gliogenic symmetric division while those that were generated late must stop en route and die by apoptosis and/or autophagy. This would solve the strange diversity of end-of-life, which could be easily addressed by identifying the temporal stage of the neural stem cells that undergo apoptosis

We agree that it would be of interest to understand how there are diverse mechanisms by which medulla NBs terminate during pupal development. To address if temporal progression is involved in apoptosis of the medulla NBs, we will first characterize the expression of some temporal TFs (e.g., Ey, Slp, Tll) at 12h APF when we found a subset of medulla NBs undergo apoptosis in the wildtype animals.

Minor suggestions:

We agree with these minor modifications.

• Line 46: Specify that there are 8 type II neural stem cells in each hemisphere*.

• The statement in lines 181-182 that "cell death, and not autophagy, makes a minor contribution to..." should be replaced with "apoptosis, and not autophagy," as autophagy is also a type of cell death.

• The authors should adjust the logic of the section "Medulla neuroblasts terminate during early pupal development": Describe the wild-type pattern first (the decrease in the number of neural stem cells and their size with age) and then describe the perturbations aimed at disrupting the number and the size of neural stem cells

• Line 151 should refer to Fig. 2I-K, not Fig. 2J-K.

**Referees cross-commenting**

How can NBs die by different mechanisms?? This might only happen is they are in a different states, an issue that is not addressed.

it has been shown that optic lobe NBs end their life by a symmetric, gliogenic last division at the end of the last temporal window, and not by PCD.

It is likely, and the authors do hint at it, that NBs only die by PCD when they prematurely interrupt the temporal series in early pupation when neurons synchronously start undergoing maturation.

I believe that the authors should explain this, if this is indeed their model, and show that NBs die while still in early temporal windows.

Reviewer #2 (Significance (Required)):

Lineages of neural stem cells are of great interest to understand how many neural types are generated. They produce very diverse neurons, often in a highly stereotyped series. However, they must terminate their life when the animal becomes functional or if neurons need time to become mature before birth.

In the Drosophila optic lobes, neural stem cells are produced over a period of several days by a wave of neurogenesis that transforms a neuroepithelium into neural stem cells that undergo a series of temporal patterning steps. It has been reported that they finish their life when a symmetric division generates glial cells. The authors however analyze the end of a particular lineage, that of the latest born neural stem cells of the medulla.

The paper shows that neural stem cells stop being produced when the neuroepithelium is consumed. But how do the latest born neural stem cells stop their lineage?

The results show that they do so by several means, which is quite unexpected: they may die from apoptosis, or autophagy, by becoming glioblasts or by a terminal symmetric division.

There are no major issues affecting the conclusions

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

Summary

In this manuscript, the authors address the timing and mechanisms responsible for the termination of medulla neuroblasts in Drosophila visual processing centres, also known as optic lobes. Through time course experiments the authors demonstrate the medulla NBs are completely eliminated by 30h APF during early pupal development. By manipulating the Notch signalling pathway as well as proneural genes such as lethal of scute, the authors show that altering the NE-NB transition is sufficient to change the timing of NB termination. In contrast, ecdysone signalling and components of the mediator complex, known to terminate proliferation of central brain NBs, are not required for the termination of medulla NBs. Medulla NBs sequentially express a variety of temporal transcription factors to promote cellular diversity, however, the authors demonstrate that altering temporal factors such as Ey, Sco or Hth, does not affect the timing of the medulla NBs termination. Interestingly however overexpression of the transcription factor tailless can cease medulla NB termination via the conversion of type I to type II NB fate. They further go on to show the importance of the differentiation factor, Prospero, in promoting the differentiation of medulla NBs as well as terminating medulla neurogenesis during pupal development. Finally, in addition to differentiation, the authors show another mechanism responsible for the cessation of neurogenesis which is the commencement of gliogenesis. Through manipulation of the neurogenic to gliogenic switch by knockdown or overexpressing the glial regulatory gene, gcm, the authors show that even though the downregulation of gcm is is not sufficient to induce NB persistence, gcm overexpression can cause premature termination of NBs.

Major comments:

• Are the key conclusions convincing?

Yes, the key conclusions are convincing with proper controls, quantifications and statistical analyses.

• Should the authors qualify some of their claims as preliminary or speculative, or remove them altogether?

The conclusion that temporal transcription factors (TTF) do not affect the timing of medulla NB termination is somewhat preliminary. The authors investigated a simplified temporal series including Homothorax, Eyeless, Sloppy-paired, Dichaete and Tailless. However, there are additional temporal factors that have not been examined for their potential involvement in medullar NB termination. Previous reports have identified several other temporal factors that play a role in medulla TTF cascade, such as, SoxNeuro (SoxN) and doublesex-Mab related 99B (Dmrt99B) that start their expression in the NE similar to Hth, however, Dmrt99B is likely to be repressed much later than Hth (Li, Erclik et al. 2013, Zhu, Zhao et al. 2022). At this point, it remains challenging to completely rule out the possibility that other temporal factors play a role in medullar NB termination or have redundant functions in regulating the timing of medulla NB cessation. It is suggested to tone down this claim and provide a brief discussion on alternative possibilities, citing relevant papers on the functions of other temporal factors in medullar NBs.

We agree.

• 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.

Loss of pros by RNAi caused the formation of ectopic NBs and the NBs persist even at 24h APF. Do these NBs persist at 30h or 48h APF? Does overexpression of Pros result in early termination of medulla NBs?

We will do these experiments in clones as suggested.

• 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.

Yes, I believe the suggested experiments are realistic in terms of time and resources, with an estimation of 3 months to complete the experiments.

• Are the data and the methods presented in such a way that they can be reproduced?

Yes.

• Are the experiments adequately replicated and statistical analysis adequate?

The experiments are straight forward and were performed with proper controls, supported by quantifications and proper statistical analyses. However, there is no mention about how many replicates were used.

We will add this information in our Material and Methods section.

Minor comments:

1. The authors use the eyR6F10-Gal4 driver in certain experiments. The eyR6F10-Gal4 driver is however expressed only in a subset of medulla NBs. Can the authors comment on what percentage of medulla NBs is the driver expressed in? We will characterize this.

Does the EGFR signalling pathway or JAK/STAT pathway affect the timing of termination of medulla NBs? Experiments are not necessary. The author can speculate on their roles.

We will modify our discussion accordingly.

Figure 1C has a p value of only 0.03 (*) but shows a strong reduction in the number of Dpn+ cells from 12h to 18h, etc. Is this correct? Also, is the p value the same for the comparison between 12h and 24h as well as 12h and 30h APF?

Yes. P-values showed no significant differences between 28-24h and 24-30h APF.

The controls in figure 2B and to some extent figure 2H show one major outlier (much higher than the other brain lobes in the control). Will the removal of this outlier affect the significance/ p-value of the experiment?

No, removing the outliers do not change the statical results.

In figure 2B what is the p-value between 12h and 18h APF? Is it *** as well?

No, it’s not significant.

Line 84 of the introduction introduces Tll, Gcm and Pros for the first time in the manuscript and should be written out in full.

We will change this.

• Are prior studies referenced appropriately?

Yes.

• Are the text and figures clear and accurate?

Yes.

• Do you have suggestions that would help the authors improve the presentation of their data and conclusions?

Quite a few of data mentioned in the manuscript have been described as data not shown. I think it would be nice to show quantifications or representative images in the supplementary figures.

We will add the data which was previously not shown.

Reviewer #3 (Significance (Required)):

Since the mechanisms by which medulla NBs are terminated are currently unknow, this is an important and interesting study to understand how medulla neuroblasts in the optic lobe are terminated. The balance between stem cell maintenance and differentiation is critical for proper brain development and the results presented in this paper are impactful. Furthermore, Drosophila melanogaster is an excellent model to study stem cell niches and neuroblast temporal patterning. The authors provide key mechanisms namely cell death, Pros-mediated differentiation and the gliogenic switch that contribute to a better understanding of how the NB progenitor pool can be terminated in the Drosophila OL, which is largely supported by the data.

• Place the work in the context of the existing literature (provide references, where appropriate).

So far, most work in this field has focused on the regulation of the temporal factors to promote the progression of the TTF transcriptional cascade and thereby diversity of the neural progenitors (Li, Erclik et al. 2013, Naidu, Zhang et al. 2020, Ray and Li 2022, Zhu, Zhao et al. 2022). Furthermore, work on pathways such as EGFR and Notch signalling that allows the proneural wave to progress and subsequently induce neuroblast formation in a precise and orderly manner have also been studied (Yasugi, Umetsu et al. 2008, Yasugi, Sugie et al. 2010). Here, considering previous literature, the authors move one step forward to determine how and when these neuroblast progenitors cease proliferation during development thus providing mechanisms for the regulation of the neuroepithelial stem cell pool, its timely conversion into NSCs and the switch from neurogenesis to gliogenesis thus providing important implications for brain size determination and function.

• State what audience might be interested in and influenced by the reported findings.

Stem cell research, neurobiologists and developmental biologists.

• Define your field of expertise

Stem cells, developmental 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 #3

Evidence, reproducibility and clarity

Summary

In this manuscript, the authors address the timing and mechanisms responsible for the termination of medulla neuroblasts in Drosophila visual processing centres, also known as optic lobes. Through time course experiments the authors demonstrate the medulla NBs are completely eliminated by 30h APF during early pupal development. By manipulating the Notch signalling pathway as well as proneural genes such as lethal of scute, the authors show that altering the NE-NB transition is sufficient to change the timing of NB termination. In contrast, ecdysone signalling and components of the mediator complex, known to terminate proliferation of central brain NBs, are not required for the termination of medulla NBs. Medulla NBs sequentially express a variety of temporal transcription factors to promote cellular diversity, however, the authors demonstrate that altering temporal factors such as Ey, Sco or Hth, does not affect the timing of the medulla NBs termination. Interestingly however overexpression of the transcription factor tailless can cease medulla NB termination via the conversion of type I to type II NB fate. They further go on to show the importance of the differentiation factor, Prospero, in promoting the differentiation of medulla NBs as well as terminating medulla neurogenesis during pupal development. Finally, in addition to differentiation, the authors show another mechanism responsible for the cessation of neurogenesis which is the commencement of gliogenesis. Through manipulation of the neurogenic to gliogenic switch by knockdown or overexpressing the glial regulatory gene, gcm, the authors show that even though the downregulation of gcm is is not sufficient to induce NB persistence, gcm overexpression can cause premature termination of NBs.

Major comments:

• Are the key conclusions convincing?

Yes, the key conclusions are convincing with proper controls, quantifications and statistical analyses. - Should the authors qualify some of their claims as preliminary or speculative, or remove them altogether?

The conclusion that temporal transcription factors (TTF) do not affect the timing of medulla NB termination is somewhat preliminary. The authors investigated a simplified temporal series including Homothorax, Eyeless, Sloppy-paired, Dichaete and Tailless. However, there are additional temporal factors that have not been examined for their potential involvement in medullar NB termination. Previous reports have identified several other temporal factors that play a role in medulla TTF cascade, such as, SoxNeuro (SoxN) and doublesex-Mab related 99B (Dmrt99B) that start their expression in the NE similar to Hth, however, Dmrt99B is likely to be repressed much later than Hth (Li, Erclik et al. 2013, Zhu, Zhao et al. 2022). At this point, it remains challenging to completely rule out the possibility that other temporal factors play a role in medullar NB termination or have redundant functions in regulating the timing of medulla NB cessation. It is suggested to tone down this claim and provide a brief discussion on alternative possibilities, citing relevant papers on the functions of other temporal factors in medullar NBs. - 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.

Loss of pros by RNAi caused the formation of ectopic NBs and the NBs persist even at 24h APF. Do these NBs persist at 30h or 48h APF? Does overexpression of Pros result in early termination of medulla NBs? - 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.

Yes, I believe the suggested experiments are realistic in terms of time and resources, with an estimation of 3 months to complete the experiments. - Are the data and the methods presented in such a way that they can be reproduced?

Yes. - Are the experiments adequately replicated and statistical analysis adequate?

The experiments are straight forward and were performed with proper controls, supported by quantifications and proper statistical analyses. However, there is no mention about how many replicates were used.

Minor comments:

1. The authors use the eyR6F10-Gal4 driver in certain experiments. The eyR6F10-Gal4 driver is however expressed only in a subset of medulla NBs. Can the authors comment on what percentage of medulla NBs is the driver expressed in?
2. Does the EGFR signalling pathway or JAK/STAT pathway affect the timing of termination of medulla NBs? Experiments are not necessary. The author can speculate on their roles.
3. Figure 1C has a p value of only 0.03 (*) but shows a strong reduction in the number of Dpn+ cells from 12h to 18h, etc. Is this correct? Also, is the p value the same for the comparison between 12h and 24h as well as 12h and 30h APF?
4. The controls in figure 2B and to some extent figure 2H show one major outlier (much higher than the other brain lobes in the control). Will the removal of this outlier affect the significance/ p-value of the experiment?
5. In figure 2B what is the p-value between 12h and 18h APF? Is it *** as well?

6. Line 84 of the introduction introduces Tll, Gcm and Pros for the first time in the manuscript and should be written out in full.

7. Are prior studies referenced appropriately?

Yes. - Are the text and figures clear and accurate?

Yes. - Do you have suggestions that would help the authors improve the presentation of their data and conclusions?

Quite a few of data mentioned in the manuscript have been described as data not shown. I think it would be nice to show quantifications or representative images in the supplementary figures.

Significance

Since the mechanisms by which medulla NBs are terminated are currently unknow, this is an important and interesting study to understand how medulla neuroblasts in the optic lobe are terminated. The balance between stem cell maintenance and differentiation is critical for proper brain development and the results presented in this paper are impactful. Furthermore, Drosophila melanogaster is an excellent model to study stem cell niches and neuroblast temporal patterning. The authors provide key mechanisms namely cell death, Pros-mediated differentiation and the gliogenic switch that contribute to a better understanding of how the NB progenitor pool can be terminated in the Drosophila OL, which is largely supported by the data.

• Place the work in the context of the existing literature (provide references, where appropriate).

So far, most work in this field has focused on the regulation of the temporal factors to promote the progression of the TTF transcriptional cascade and thereby diversity of the neural progenitors (Li, Erclik et al. 2013, Naidu, Zhang et al. 2020, Ray and Li 2022, Zhu, Zhao et al. 2022). Furthermore, work on pathways such as EGFR and Notch signalling that allows the proneural wave to progress and subsequently induce neuroblast formation in a precise and orderly manner have also been studied (Yasugi, Umetsu et al. 2008, Yasugi, Sugie et al. 2010). Here, considering previous literature, the authors move one step forward to determine how and when these neuroblast progenitors cease proliferation during development thus providing mechanisms for the regulation of the neuroepithelial stem cell pool, its timely conversion into NSCs and the switch from neurogenesis to gliogenesis thus providing important implications for brain size determination and function. - State what audience might be interested in and influenced by the reported findings.

Stem cell research, neurobiologists and developmental biologists. - Define your field of expertise

Stem cells, developmental 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 #2

Evidence, reproducibility and clarity

Lineages of neural stem cells are of great interest to understand how many neural types are generated. They produce very diverse neurons, often in a highly stereotyped series. However, they must terminate their life when the animal becomes functional or if neurons need time to become mature before birth.

In the Drosophila optic lobes, neural stem cells are produced over a period of several days by a wave of neurogenesis that transforms a neuroepithelium into neural stem cells that undergo a series of temporal patterning steps. It has been reported that they finish their life when a symmetric division generates glial cells. The authors however analyze the end of a particular lineage, that of the latest born neural stem cells of the medulla. The paper shows that neural stem cells stop being produced when the neuroepithelium is consumed. But how do the latest born neural stem cells stop their lineage?

The results show that they do so by several means, which is quite unexpected: they may die from apoptosis, or autophagy, by becoming glioblasts or by a terminal symmetric division.

There are no major issues affecting the conclusions

• The paper shows that the end of production of neural stem cells occurs the neuroepithelium is completely transformed. The experiments performed by the authors are fine and show that, if the transition is delayed, neural stem cells terminate their life later, and vice versa. However, the lifespan of the neural stem cells is not affected by the timing of the transition. Therefore, these experiments do not tell us how neural stem cells terminate their life, which is the central question of the study. The discussion should be written accordingly and the title and the model in Fig 6 modified to reflect the importance of the end of life of the stem cells, the main theme of the paper.
• The authors talk about Pros-dependent symmetric division and gliogenic switch as two separate processes, but these may be two sides of the same phenomenon. Tll+ gcm+ neural stem cells undergo Pros-dependent cell cycle exit, generating glial progeny. If the authors agree with this, could they update their model (and discussion) to reflect the fact that gliogenic switch occurs via a Pros-dependent symmetric division, and these are not two separate processes independently contributing to the depletion of the neural stem cell pool? Ideally, a triple staining between Dpn, Pros, and gcm would show that the symmetrically dividing cells seen by the authors are committed to the glial fate.
• Why were Notch RNAi experiments assessed for the presence of neural stem cells at P12 and gcm RNAi experiments at P24? Given that most optic lobe neural stem cells disappear between P12-18, a subtle effect of gcm RNAi may have been missed. Do the authors have data for gcm RNAi at P12?
• The authors should acknowledge that the inhibition of either apoptosis or autophagy alone may not be fully sufficient to prevent the death of NBs. In mushroom body neural stem cells, both processes must be inhibited simultaneously to produce a strong effect on their survival (Pahl et al. 2019, PMID 30773368).
• There is an important missing point that should be addressed: is there a specific point in time when all neural stem cells must stop their lineage wherever they are in the temporal series and either die or divide symmetrically? One possibility that is not discussed is that most neural stem cells end their life through a gliogenic symmetric division while those that were generated late must stop en route and die by apoptosis and/or autophagy. This would solve the strange diversity of end-of-life, which could be easily addressed by identifying the temporal stage of the neural stem cells that undergo apoptosis

Minor suggestions:

• Line 46: Specify that there are 8 type II neural stem cells in each hemisphere*.
• The statement in lines 181-182 that "cell death, and not autophagy, makes a minor contribution to..." should be replaced with "apoptosis, and not autophagy," as autophagy is also a type of cell death.
• The authors should adjust the logic of the section "Medulla neuroblasts terminate during early pupal development": Describe the wild-type pattern first (the decrease in the number of neural stem cells and their size with age) and then describe the perturbations aimed at disrupting the number and the size of neural stem cells
• Line 151 should refer to Fig. 2I-K, not Fig. 2J-K.

Referees cross-commenting

How can NBs die by different mechanisms?? This might only happen is they are in a different states, an issue that is not addressed. it has been shown that optic lobe NBs end their life by a symmetric, gliogenic last division at the end of the last temporal window, and not by PCD. It is likely, and the authors do hint at it, that NBs only die by PCD when they prematurely interrupt the temporal series in early pupation when neurons synchronously start undergoing maturation. I believe that the authors should explain this, if this is indeed their model, and show that NBs die while still in early temporal windows.

Significance

Lineages of neural stem cells are of great interest to understand how many neural types are generated. They produce very diverse neurons, often in a highly stereotyped series. However, they must terminate their life when the animal becomes functional or if neurons need time to become mature before birth.

In the Drosophila optic lobes, neural stem cells are produced over a period of several days by a wave of neurogenesis that transforms a neuroepithelium into neural stem cells that undergo a series of temporal patterning steps. It has been reported that they finish their life when a symmetric division generates glial cells. The authors however analyze the end of a particular lineage, that of the latest born neural stem cells of the medulla.

The paper shows that neural stem cells stop being produced when the neuroepithelium is consumed. But how do the latest born neural stem cells stop their lineage?

The results show that they do so by several means, which is quite unexpected: they may die from apoptosis, or autophagy, by becoming glioblasts or by a terminal symmetric division.

There are no major issues affecting the conclusions

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 Nguyen and Cheng is investigating the timing and mechanism of cessation of neuroblasts in the pupal optic lobe. Previous studies by several groups have determined the spatial and temporal factors required for the neuroepithelial to neuroblast transition and neuroblast to neural/glycogenesis in third instar larvae such that neuroblasts are eliminated. The mechanism of elimination of neuroblasts in the VNC or mushroom bodies have been investigated, but the mechanism(s) and the timing of elimination of medulla neuroblasts has not been investigated. The authors suggest that medulla neuroblasts are eliminated via a combination of mechanisms including apoptosis, prospero induced size symmetric terminal differentiation and a switch to gliogenesis by gcm expression. Expression of Tailless also was found to affect the timing of medulla neuroblast termination. They also ruled out several mechanisms such as ecdysone pulses.

Major comments

Clearly written and logical flow to experiments and results not over interpreted. Clearly show that the neuroblast number and size decrease (12 to 18 hrs) and are eliminated by 30 hours

Figure 2a Marking of the Neuroepithelium. Would be more convincing if shown by PatJ expression and is clonal analysis. While the following panels use PatJ in clones suggesting are NE and NBs present it is more difficult to put into the context in the higher magnification images (Figure 2 D- M) and the Miranda expression in F' seems to be the entire lobe and it is not clear if would be any NE which does not agree with what is shown in panel A.<br /> Is difficult to see the neuroblasts in Figure 2 D D" and E. The figure does not match what is stated in the results in that the neuroblasts are difficult to observe. If the point is that there is fewer NE cells and more neuroblasts then this is hard to see. It has been previously shown that with Notch RNAi clones prematurely extrude form the NE (Egger 20210; Keegan 2023) and could be expressing more Neuroblast markers but this is not visible in the panels as shown. Are the images single focal plane or maximum projections? Imaging more deeply in the brain or viewing in cross section would account for these possibilities. The possibility that are more neuroblasts but not all at the surface of the OL should be addressed as this could also alter the overall results. Figure 2 is key to first point of the paper so needs to be addressed.

Minor comments

Why express volume of DPN in clone volume. Would make the point more clear and more strong be to express as number of NB in the 3-D volume of the clone. This measurement occurs in several figures. Use of Miranda to mark NBs is unclear in Figure 2. Perhaps more clear in B&W. Make clear in figures (or figure legend) if single focal plane or projections. It is unclear what percentage of NB the Gal4 line eyR16F10 are expressed in. Veen 2023 state that the GAL4 is also expressed in neurons and at different levels whether deeper within the brain or superficially on the surface of the brain. At 16 APF it is expressed but it is not clear whether it is in all cells at a low level or only within a few cells Some RNAi lines referenced as previously validated and other are not. For example: EcR, Oxphos, Med27, Notch need references or confirmation of specificity to the intended target (qRT) At least 2 animals per genotype were used. While I appreciate the technical difficulty of working in pupae this seems a bit low in terms of number of samples and data would be more robust with more numbers.

Significance

This provides mechanism and timing for the elimination of neuroblasts (NE to NB) that arise from the medulla. As these are most similar to mammalian brain development (Radial glial to NSC) this information provides more context to interpret the formation of glial and neurons in the adult optic lobe given the effect on timing and mechanisms of elimination.

This paper would be of interest to developmental biologist who work with Drosophila or mice who are looking at neural development. An understanding of how neural diversity is achieved and the mechanisms behind this that can be dysfunctional in terms of etiology of neural diseases. Is a well done study for the most part that would be improved by clarifying some data and provided more replicates for robustness of the data. I am a developmental biologist working with Drosophila in larval and adult neural development.

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

Rebuttal for Review Commons to:

“A specific innate immune response silences the virulence of Pseudomonas aeruginosa in a latent infection model in the Drosophila melanogaster host”

We thank the reviewers for their careful scrutiny of our manuscript. We believe that we have addressed satisfactorily the points raised by the reviewers and that our revised manuscript is definitely improved. Our replies below are in blue and use a distinctive font.

Reviewer #1

__Evidence, reproducibility and clarity __

This works describes a latent Drosophila intestinal infection, which spreads systemically, with a direct systemic Drosophila infection using a common laboratory strain of Pseudomonas aeruginosa. The major observation of this study is that P. aeruginosa can cause a latent infection via its passage through the gut (as opposed to being injected). In doing so it exhibits cell rounding (instead of elongation), reduced cell motility, loss of O5-antigen, antibiotic resistance, ability to cross the intestinal barrier and circulate in the hemolymph and infiltrate the host tissue underneath the cuticle. In addition, latent infection bacteria induce all brunches of the systemic response: the Imd pathway, phagocytosis, and the melanization cascade. Moreover, the melanization pathway protects the host from a secondary systemic infection with various types of bacterial and fungal microbes.

An issue that needs to be clarified is the sensitivity of P. aeruginosa virulence to its biochemical environment. The authors note that. For example, liquid bacterial culture in BHI induces the latent form of bacteria. So the growth conditions and the infection media play a major role in the infection process. They authors need to clarify further the effect of media and infection vehicles, sucrose (high/low), LB, and BHI (as well as temperature) on the latent phenotype.

Temperature is definitely an important parameter and bacteria appear to be somewhat more virulent at 25°C. This point is now addressed in the Material and Methods section (lines 675-679) and in Fig. S6I.

As regards the influence of the composition of the infection solution, it does not seem to be a critical parameter that we have described in the context of continuous feeding on the bacterial solution (Limmer et al., PNAS, 2011). In preliminary experiments, we had tried LB or BHB medium to grow the bacteria and this did not make any difference (see Panel A below for LB [BHB used in our experiments]). As regards the sucrose concentration to the infection solution, we have tried two concentrations and did not observe any difference as regards the establishment of the latent infection. (see panel B below for 50 mM sucrose [100mM used in our experiments]). Of note, P. aeruginosa does not grow on sucrose solution alone. However, a latent infection was still established upon feeding the flies with PAO1 in sucrose alone, albeit likely with a mildly increased virulence, in the absence of any BHB medium (see panels C-D) below.

A) Comparison of LB vs. BHB B) Establishment of latent infection with 50mM sucrose

C) Establishment of a latent infection with a sucrose-only bacterial solution D) Colonization of host tissues by PAO1 ingested in a sucrose-only bacterial solution Minor issues: -Lines 579-581> How were the PAO1GFP/RFP constructed (details are needed)

Done; please, see lines 641-643.

-Figure 1D and other figures > CFUs given as Log2 are unconventional. One cannot easily deduce the burden unless e.g. translate 2e10 to ~1000 and 2e30 to ~10e9 CFUs.

True, but bacterial titer increases by a factor of two at each division cycle. Even though we have previously used a Log10 representation, we now prefer using a Log2 representation. This representation has also been used by other authors in the field, e.g., Duneau et al., eLife, 2019.

-Figire S1DB (now S1C)> "but from the outside of the gut". The given experiment does not prove that statement.

This issue has been already dealt with in the Nehme et al. PLoS Pathogens 2007 article, as cited in the manuscript. We further provide in Fig. S1B pictures documenting the presence of bacteria associated with visceral muscles. Finally, we also show that the gut lumen is essentially cleared of bacteria after a period of feeding on a sucrose solution or gentamicin. Hence, most bacterial colonies originate from the outer layer of the gut. We clarify the issue in the text (lines 154-158).

-Lines 146-7 > data are missing in support to the statement.

We have now added Fig. S1B to document that the gentamicin treatment does work, as actually does feeding on sucrose solution alone, as previously documented in Limmer et al., 2011 (Fig. S2B). Of note, we cannot exclude that a few bacteria remain, especially in the crop, but those would be at very low titer. Please, see also reply to Reviewer 2.

-Figure S1C > The effect of injury seems to be huge, and may account for much/most of the differences observed (including those between latent and active infection). This is further supported by Figure4A, injury may account for gut collapse and/or systemic stress.

It is well known that injury alone induces the systemic IMD pathway response 6 hours after injury but largely subsides by 24 hours. The point of Fig. S1C is that the level of induction reached during latent infection is very low as compared to that observed during a systemic infection, here obtained for reference with an Escherichia coli injection and to a lesser extent with a PBS injection. In our latent infection model, we do not perform any injury, except as noted by the reviewer in Fig. 4: the effects of an experimental injury are observed only while the bacteria are crossing the intestinal barrier and hardly any effect is observed when the injury is performed on day 10 (Fig. S4B).

-Figure S1D > How was "fated to die" assessed?

The fluorescent flies were sorted out and their subsequent survival was monitored. As compared to nonfluorescent flies from the same batch, they died within two days of sorting them.

-Figure 3B/10th day > Average line is misplaced.

We thank the reviewer for pointing out this problem. The line is not the average but the median. We have now added a precise description of the bars to all the figure legends.

-Lines 382-5 > what is the evidence of gut damage (or the absence of it)? How do the bacteria escape the gut?

The absence of major gut damages has been documented in Limmer et al, PNAS, 2011. How the bacteria escape the gut remains an open question (intracellular and/or paracellular route).

-Lines437-442 > The distinction between dormant P. aeruginosa in the fly tissues and persister cells (upon antibiotic treatment) cannot be justifies on the basis of relative bacterial numbers in the two systems. The extent of resistance to antibiotics though my serve that purpose.

In our latent infection model, most of the bacteria that have crossed the gut barrier become dormant and are associated with tissues, except at the beginning of the infection. In contrast, when a bacterial culture is treated with antibiotics, most of the bacteria are killed by the treatments and only a few ones persist, likely because of an inactive metabolism. Thus, the induction of dormancy in our latent infection model does not rely on the selection of a few metabolically-inactive bacteria able to withstand an immune response or an antibiotic treatment.

Significance

The study is a significant advance to our knowledge. Notwithstanding further explanations, it provides a solid basis of understanding active versus dormant bacteria. It further establishes a mode of intestinal to systemic infection as a tool for further explorations.

Reviewer #2

Evidence, reproducibility and clarity

Summary: In this study, Chen and colleagues investigated a new latent infection model for Pseudomonas aeruginosa using Drosophila melanogaster as a host. First, the authors established a new model for latent Pseudomonas infection. The key feature of this model is the translocation of P. aeruginosa from the gut to the hemolymph and the colonization of fly tissues by the dormant bacteria. Bacteria that translocated from the gut appeared strikingly different in morphology and resistance to antibiotics compared to bacteria that were directly injected. Next, the authors suggest that melanization but not the Imd pathway or hemocytes are necessary to promote dormancy and colonization of fly tissues. Finally, flies with latent P. aeruginosa infection exhibit improved survival after secondary infections in a melanisation-dependent manner. The study reports an interesting model for latent infection, provides insights into the host factors promoting latency and describes some of the consequences of such latent infection for the host. However, some of the conclusions are not fully supported by the data and need further experimental evidence.

Major comments: 1. The latent infection model requires some clarifications. First, temperature. Could the authors explain why they used 18 {degree sign}C and could low temperature contribute to the establishment of dormancy?

As shown in Fig.S6I, the latent infection model is less compelling at 25°C in terms of survival curves, which may reflect an increased rate of spontaneous reactivation of the virulence, a phenomenon we have not studied at 25°C. In another manuscript in preparation (Lin et al.), we actually show that a small heat shock does contribute to reactivation of the bacteria, an issue that is outside of the scope of the present study. Please, see also reply to reviewer 1.

Second, the use of gentamycin. How does gentamicin affect PAO1 outside the gut? From Fig.1C It looks like the cfus in the hemolymph diminished rapidly after gentamicin treatment (around day 3), suggesting the potential effect of the antibiotic. Once the bacteria have crossed the gut and entered the hemolymph, they could still be affected by feeding flies the antibiotic. Is there a possibility that gentamicin treatment is a stress factor that could trigger or facilitate the transition to dormancy? The authors could test this experimentally either by omitting the antibiotic and assessing dormancy or by feeding injected flies with gentamycin and scoring dormancy.

We had actually tested the issue about the potential role of gentamicin outside of the gut compartment. We have thus fed flies on different concentrations of gentamicin and monitored the survival of those flies to the injection of PAO1 bacteria (please, see Figure below). When flies were feeding on the highest concentration of gentamicin tested, 32 mg/mL, they were succumbing fast to the PAO1 challenge, but not as fast as nontreated positive control PAO1 injected flies. The use of lower concentrations (16, 8, and 4 mg/mL) led to a progressively stronger protection from PAO1 injection that inversely correlated with the gentamicin dose. We interpret the data with high gentamicin concentrations as an indication that gentamicin at such concentrations is likely directly toxic to the flies, an issue that could be experimentally tested but is not relevant to this study. Interestingly, lower doses led to a much-decreased protection from PAO1 (2mg/mL) to no protection at the dose we use to establish latent infection (100 µg/mL). Thus, these data show that gentamicin can pass the gut barrier when provided at high concentrations, down to 2 mg/mL. However, there is no proof of such a passage at the dose we use. In keeping with this latter possibility, we made a control experiment in which the gentamicin treatment step was replaced by simply feeding on the sucrose solution: the bacterial titer decreased in the hemolymph at the same rate as for gentamicin-treated flies (new Fig. S1F), demonstrating that ingested gentamicin does not contribute to the decreased titer. Rather, the likely depletion of the “source”, that is PAO1 in the gut lumen, best accounts for this phenomenon.

We have now cited references which document a lack of permeability of the gut barrier to ingested gentamicin in vertebrate animals (lines 130-133).

As regards the possibility that gentamicin acts as a stress factor on bacteria, our data do not support this possibility, as a latent infection is established in the absence of gentamicin by just feeding the flies on a sucrose solution. We had previously reported that flies fed with P. aeruginosa for up to three days do not succumb within the next two weeks when they are fed on a sterile sucrose solution after having ingested the bacterial solution (Limmer et al., PNAS, 2011; Fig. 1C). Under the conditions of two days of PAO1 ingestion, we document in novel Fig. S1G that the carcass is equally well colonized under these conditions.

Figure: impact of gentamicin ingestion at diverse concentrations on the survival of injected PAO1 bacteria. The ingested antibiotics can act on bacteria present in the hemocoel at concentrations over 2 mg/mL and not at that used in our experiments (100 µg/mL).

Does melanization really induce the dormant state of the bacteria? I am not sure the provided data fully support this claim. Addressing these questions might provide a stronger evidence: Fig. 2 A-F: What causes the morphological changes of the bacteria? Melanization or the passage through the gut? Do authors see the same changes in bacteria retrieved from PO-deficient mutant flies? Fig. 2G: Do the authors see the same resistance of PAO1 that colonized PO mutant flies to antibiotics?

In a novel Fig. S4, we now document comprehensively the physiological state of PAO1 bacteria fed to PO-deficient flies. We find that these bacteria are susceptible to antibiotics treatment as they can be rescued from death by the injection of antibiotics on day 3 (Fig. S4A-B). Second, they show a mixed phenotype in terms of colony morphology (Fig. S4C) and bacterial morphology and cell wall properties: even though most bacteria appeared to have kept a rounded morphology, they predominantly (about 75%) expressed the O5-LPS antigen. We interpret these data in terms of a slower transition to virulence than in a septic injury model. Melanization thus strongly contributes to the establishment of latency, even though it is likely that other factors contribute to the establishment of dormancy, but at best provide a minor contribution.

How do PO mutants behave after PAO1 injection? Are they similarly more susceptible?

PPO1-PPO2 mutants are not more susceptible to PAO1 injection than wt controls (new Fig. S3C).

Fig. 3F: PPO1 is believed to be the fast-acting PPO, whereas PPO2 is deployed later in infection.

This statement is based on experimental data gained on larvae, not adults. It is not really clear whether the about 10% adult hemocytes that express PPO2 actually contain crystals, in as much as the adult may be better oxygenated than larvae that grow in a hypoxic environment (description by the laboratory of Prof. Jiwon Shim of a role for PO in respiration at the latest EDRC meeting).

How does the Western blot look for PPO1? Will it show an early induction of melanization that could drive the change into the dormant state?

We provide below a characterization of the PPO antibody we use by Western blot analysis. This antibody had originally been raised by the late Dr. Hans-Michael Müller against a PPO from mosquito cell lines, hence explaining its cross-reaction to both * Drosophila PPO1 and PPO2 (Muller, H.M., Dimopoulos, G., Blass, C., and Kafatos, F.C. (1999)). A hemocyte-like cell line established from the malaria vector Anopheles gambiae expresses six prophenoloxidase genes. J Biol Chem 274*, 11727-11735.). It follows that at least one PO is partially cleaved at day 2 and that both are fully cleaved by day6 of the establishment of the latent infection (Fig. 3F, Fig. S3F).

Figure: characterization of the antibody raised against A. gambiae PPO

Alternatively, the induction of melanization could also be measured with an L-DOPA test.

This experiment is not needed given the explanation provided above.

Fig. 3E: Melanization prevents the growth of PAO1 adhering to tissues, as shown in Fig. 3E. One can see higher levels of cfus in the carcass in PO deficient flies compared to wt flies. However, after, 5 days, there is no difference in the cfus of wt and mutant flies anymore. If the growth inhibition was melanization mediated, would we not expect a consistent growth of bacteria in PO mutants? How to explain the drop in cfus in PO deficient mutants?

This observation is difficult to account for and the explanations we can put forward at this stage are somewhat speculative. It appears that bacteria found in the tissues in PO-deficient flies have a morphology found in in vitro culture and within the gut, which does not correlate with virulence but also not with the avirulence state since they are LPS O5 positive. Given the shallow survival curves, we envision that there is a progressive release of bacteria from the tissue and then quick proliferation in the hemolymph in a few flies that would then die, but at a frequency too low to reliably ascertain in our hemolymph titer data, with a few flies displaying a high titer (Fig. 3D). By day5, the decreased titer in the carcass may reflect the progressive depletion of tissue-associated bacteria as they progressively become planktonic.

Fig. 5D: How do PAO1 bacteria react to Levofloxacin treatment? Do they still go into the dormant state? Do they still attach to tissues? The authors should show that Levofloxacin treatment leads to the same dormant state as gentamycin before interpreting the results of this experiment.

Taken together, our data yield a mixed result. When levofloxacin was fed for two days to latently-infected flies, we found that colonization was not altered (Fig. S2D’), in contrast to a septic injury model in which injected bacteria were susceptible to the ingested antibiotics (Fig. S2D”). Following the reviewer’s query, we have further monitored survival and bacterial colonization in the levofloxacin ingestion model. Fig. S2D had already demonstrated that ingested levofloxacin protects the flies from injected PAO1. Fig. S6F shows that the double mutant PO bacteria are protected from ingested PAO1 by the ingestion of this antibiotics. When we monitored the bacterial burden, we found for both wild-type and double PO mutant flies that the bacteria had been cleared in some 50% of the flies. The exact interpretation of the wild-type data is not straightforward. On the one hand, the colonizing bacteria may have become susceptible to the antibiotics even though they remained dormant. On the other hand, they might have been reactivated in their virulence and thus become secondarily susceptible to the antibiotics. For the double PO mutants, the 40% bacteria remaining may witness the mixed bacterial state of PAO1 in these mutants, as documented in Fig. S4. Nevertheless, the important point is that bacteria are unlikely to contribute to the demise of secondarily infected flies since they have been cleared in at least 50% of the flies, yet the secondarily challenged flies become susceptible only when the relevant melanization genes are affected. The nonPAO1-infected controls succumb faster to the infection than infected ones: the protection against secondary infections is provided by the activation of the melanization cascade by colonizing PAO1 bacteria, even if the colonization is transient in the levofloxacin treatment.

We have altered the main text to reflect these novel data: lines 387-403.

Minor comments:

Lines 68-72. Mechanisms that are listed are not specific against Gram-negative bacteria but rather general. Please correct.

We are of course aware of this. If it is general, it also applies to Gram-negative bacteria that are the focus of this article. Actually, an earlier version of the manuscript just mentioned the IMD pathway, hence the reference to Gram-negative bacteria. However, the Toll pathway is also required in the host defense against some Gram-negative pathogens such as P. aeruginosa. We have now deleted “Gram-negative” in this corrected version.

Line 95. In - as?

We are not sure we understand this comment. We have now added a reference documenting that P. aeruginosa can be found in rotting fruits (line 97).

Lines 182-187. Some background information is needed. What is O5 LPS antigen? What motivated the authors to look at it specifically?

The O-antigen is a long-chain polysaccharide motif that constitutes the outermost part of the cell wall. It varies according to the strain. We have added a couple of references that refer to O-antigen (line 198). We had actually already found out this result (unpublished) with the Serratia marcescens Db11 O-antigen (O18) that was not found in bacteria that had crossed the gut. The loss of the O5 antigen changes the surface of the bacterium and likely its interactions with tissues and/or the immune system. In the case of Serratia, we suspect that the loss of its O-antigen allows the bacterium to be phagocytosed in an eater-dependent manner.

Fig. 3C: Why PPO1 and Hayan and PPO1,2 and Sp7 are compared but not mutant vs wild type?

The reason is that it was obviously significant. We have now added the comparisons to wild-type in the revised figure.

How precise is estimation of bacteria in the carcass?

Even though it is not possible to measure how precise these measures are, they are nevertheless reproducible making us confident that they provide an estimate of the rough number of these bacteria found associated to tissues.

How do the authors prevent dissemination of the bacteria during dissection? I wonder if some bacteria might by lost during the dissection (when removing the gut and ovaries) or if you carry over some bacteria from the hemolymph into the carcass measurement? How to make sure, that the bacteria you recover were really adherent and were not leftover from the hemolymph?

It is not possible to prevent dissemination as we cannot fix the tissues and bacteria if we make cfu counts. However, the finding that bacteria are found in the hemolymph only for the first three days, with a distinct morphology from tissue-associated bacteria, and not at later time points make us confident that this is not an issue, which suggests that the bacteria are rather tightly attached to the tissues. As regards contamination of tissues by hemolymph, it is also not an issue since the hemolymph titers are so low. However, when the bacteria are actively proliferating to high levels, this is a legitimate concern.

I am also curious how the differences in the cfu levels between whole fly and carcass can be explained (Fig. 1D). After day 5 there are almost no bacteria left in the hemolymph, however, if you compare cfus in the whole fly vs. the carcass, one can see that the whole fly cfus are rising from day 4 onwards. Where do these bacteria come from if not from the hemolymph?

To assess the statement of the reviewer, we now have included the numerical values of the medians of the bacterial burdens displayed in Fig. 1D. There is no increased bacterial burden in whole flies between days 5 to 12; however, the titer is increased at days 15 and 22. Whether this slight increase is biologically relevant is questionable given the spread of the data (see also reply to previous point on the precision of measures). We cannot rigorously exclude that there might be a low-level proliferation of colonizing bacteria late in the latent infection, which has been observed in specific conditions of reactivation of dormant bacteria (Lin et al., in preparation).

Fig. S4D: If the protection to secondary PAO1 infection is not mediated via Imd or phagocytosis, is it mediated via melanization? How do melanization mutants (increased or decreased) respond to PAO1 secondary infection?

We have performed the experiment (Fig. S6A-B) and found that the protection was abrogated. As noted in the main text, the interpretation is however difficult since the bacteria are no longer in a dormancy state in the PPO mutants.

Significance

This study suggests that host factors, particularly specific immune responses, could drive the latent infections. Hence, besides bacterial mechanisms that received significant attention, we should not underestimate the host's contribution to promoting the latent state in bacteria.

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, Chen and colleagues investigated a new latent infection model for Pseudomonas aeruginosa using Drosophila melanogaster as a host. First, the authors established a new model for latent Pseudomonas infection. The key feature of this model is the translocation of P. aeruginosa from the gut to the hemolymph and the colonization of fly tissues by the dormant bacteria. Bacteria that translocated from the gut appeared strikingly different in morphology and resistance to antibiotics compared to bacteria that were directly injected. Next, the authors suggest that melanization but not the Imd pathway or hemocytes are necessary to promote dormancy and colonization of fly tissues. Finally, flies with latent P. aeruginosa infection exhibit improved survival after secondary infections in a melanisation-dependent manner. The study reports an interesting model for latent infection, provides insights into the host factors promoting latency and describes some of the consequences of such latent infection for the host. However, some of the conclusions are not fully supported by the data and need further experimental evidence.

Major comments:

1. The latent infection model requires some clarifications. First, temperature. Could the authors explain why they used 18 {degree sign}C and could low temperature contribute to the establishment of dormancy? Second, the use of gentamycin. How does gentamicin affect PAO1 outside the gut? From Fig.1C It looks like the cfus in the hemolymph diminished rapidly after gentamicin treatment (around day 3), suggesting the potential effect of the antibiotic. Once the bacteria have crossed the gut and entered the hemolymph, they could still be affected by feeding flies the antibiotic. Is there a possibility that gentamicin treatment is a stress factor that could trigger or facilitate the transition to dormancy? The authors could test this experimentally either by omitting the antibiotic and assessing dormancy or by feeding injected flies with gentamycin and scoring dormancy.
2. Does melanization really induce the dormant state of the bacteria? I am not sure the provided data fully support this claim. Addressing these questions might provide a stronger evidence: Fig. 2 A-F: What causes the morphological changes of the bacteria? Melanization or the passage through the gut? Do authors see the same changes in bacteria retrieved from PO-deficient mutant flies? Fig. 2G: Do the authors see the same resistance of PAO1 that colonized PO mutant flies to antibiotics? How do PO mutants behave after PAO1 injection? Are they similarly more susceptible?
3. Fig. 3F: PPO1 is believed to be the fast-acting PPO, whereas PPO2 is deployed later in infection. How does the Western blot look for PPO1? Will it show an early induction of melanization that could drive the change into the dormant state? Alternatively, the induction of melanization could also be measured with an L-DOPA test.
4. Fig. 3E: Melanization prevents the growth of PAO1 adhering to tissues, as shown in Fig. 3E. One can see higher levels of cfus in the carcass in PO deficient flies compared to wt flies. However, after, 5 days, there is no difference in the cfus of wt and mutant flies anymore. If the growth inhibition was melanization mediated, would we not expect a consistent growth of bacteria in PO mutants? How to explain the drop in cfus in PO deficient mutants?
5. Fig. 5D: How do PAO1 bacteria react to Levofloxacin treatment? Do they still go into the dormant state? Do they still attach to tissues? The authors should show that Levofloxacin treatment leads to the same dormant state as gentamycin before interpreting the results of this experiment.

Minor comments:

Lines 68-72. Mechanisms that are listed are not specific against Gram-negative bacteria but rather general. Please correct.

Line 95. In - as?

Lines 182-187. Some background information is needed. What is O5 LPS antigen? What motivated the authors to look at it specifically?

Fig. 3C: Why PPO1 and Hayan and PPO1,2 and Sp7 are compared but not mutant vs wild type? How precise is estimation of bacteria in the carcass? How do the authors prevent dissemination of the bacteria during dissection? I wonder if some bacteria might by lost during the dissection (when removing the gut and ovaries) or if you carry over some bacteria from the hemolymph into the carcass measurement? How to make sure, that the bacteria you recover were really adherent and were not leftover from the hemolymph? I am also curious how the differences in the cfu levels between whole fly and carcass can be explained (Fig. 1D). After day 5 there are almost no bacteria left in the hemolymph, however, if you compare cfus in the whole fly vs. the carcass, one can see that the whole fly cfus are rising from day 4 onwards. Where do these bacteria come from if not from the hemolymph?

Fig. S4D: If the protection to secondary PAO1 infection is not mediated via Imd or phagocytosis, is it mediated via melanization? How do melanization mutants (increased or decreased) respond to PAO1 secondary infection?

Significance

This study suggests that host factors, particularly specific immune responses, could drive the latent infections. Hence, besides bacterial mechanisms that received significant attention, we should not underestimate the host's contribution to promoting the latent state in bacteria.

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 works describes a latent Drosophila intestinal infection, which spreads systemically, with a direct systemic Drosophila infection using a common laboratory strain of Pseudomonas aeruginosa. The major observation of this study is that P. aeruginosa can cause a latent infection via its passage through the gut (as opposed to being injected). In doing so it exhibits cell rounding (instead of elongation), reduced cell motility, loss of O5-antigen, antibiotic resistance, ability to cross the intestinal barrier and circulate in the hemolymph and infiltrate the host tissue underneath the cuticle.

In addition, latent infection bacteria induce all brunches of the systemic response: the Imd pathway, phagocytosis, and the melanization cascade. Moreover, the melanization pathway protects the host from a secondary systemic infection with various types of bacterial and fungal microbes.

An issue that needs to be clarified is the sensitivity of P. aeruginosa virulence to its biochemical environment. The authors note that. For example, liquid bacterial culture in BHI induces the latent form of bacteria. So the growth conditions and the infection media play a major role in the infection process. They authors need to clarify further the effect of media and infection vehicles, sucrose (high/low), LB, and BHI (as well as temperature) on the latent phenotype.

Minor issues:

• Lines 579-581> How were the PAO1GFP/RFP constructed (details are needed)
• Figure 1D and other figures > CFUs given as Log2 are unconventional. One cannot easily deduce the burden unless e.g. translate 2e10 to ~1000 and 2e30 to ~10e9 CFUs.
• Figire S1D > "but from the outside of the gut". The given experiment does not prove that statement.
• Lines 146-7 > data are missing in support to the statement.
• Figure S1C > The effect of injury seems to be huge, and may account for much/most of the differences observed (including those between latent and active infection). This is further supported by Figure4A, injury may account for gut collapse and/or systemic stress.
• Figure S1D > How was "fated to die" assessed?
• Figure 3B/10th day > Average line is misplaced.
• Lines 382-5 > what is the evidence of gut damage (or the absence of it)? How do the bacteria escape the gut?
• Lines437-442 > The distinction between dormant P. aeruginosa in the fly tissues and persister cells (upon antibiotic treatment) cannot be justifies on the basis of relative bacterial numbers in the two systems. The extent of resistance to antibiotics though my serve that purpose.

Significance

The study is a significant advance to our knowledge.

Notwithstanding further explanations, it provides a solid basis of understanding active versus dormant bacteria.

It further establishes a mode of intestinal to systemic infection as a tool for further explorations.

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 [optional]

All four reviewers have positive comments on the paper. We totally agree with their comments, and proposed controls and experiments. Most of them are already introduced in the present text and several new figures added, as we had the controls/experiments proposed. Few others are now being done and we hope to have the complete set of experiments ready in 2-3 months.

2. Description of the planned revisions

Insert here a point-by-point reply that explains what revisions, additional experimentations and analyses are planned to address the points raised by the referees.

Reviewer #1

Most comments of this reviewer have already been done and included in the transferred manuscript, except for part of the first comment:

1.1 b. Is it possible that the loss of function of Wengen on its own has a phenotype? If so, that would suggest that Wgn in addition to its role in regeneration might be implicated in pro-survival processes in homeostatic conditions?

This issue is very important to understand the differential role of Wgn and Grnd. First of all, Wengen knock out (wgnKO; Andersen et al., 2015) is viable in homozygosis. However, in this paper we have focused on inducible mutants. Therefore, we have now crossed the flies to get the genotype hh-Gal UAS RNAi wgn and we will check for apoptotic phenotype, as suggested. This will take us few weeks of work.

Reviewer #2 Most comments have been already carried out and included in the transferred manuscript, except these ones:

*2.3. Aside from wgn, other RNAi experiments are not validated through independent RNAi lines. I suggest expanding the Supplemental Figures to reproduce a few key findings with independent RNAi lines. *

We have recently received a set of independent RNAi line to repeat the experiments for Traf1, Traf2, Ask1 from Bloomington Stock Center. And We did not do it before mainly because we wanted to focus on wgn and grnd. However, we agree with the Reviewer 2 and we will do the experiments. Another RNAi from VDRC for grnd and Tak1 have been ordered. These experiments will take about 2 months from the crosses to the analysis of results (some flies still to arrive, and many crosses will be done at 17ºC).

*2 4. In Figure 1E, the authors show that wgn RNAi enhances cell death caused by hh>egr. What is missing here is a wgn RNAi control without hh>egr. Is there any cell death caused by the loss of wgn alone (without hh>egr)? *

This control is now in progress. Expected to have it complete in 2 weeks.

*2.5. If wgn RNAi causes some degree of cell death, is the observed effect with hh>egr a significant genetic interaction, or merely additive? *

The result from the previous comment will help us to respond this point.

• 1. Is the wgn-p38 pathway sufficient to block egr induced cell death? The authors could test this by having hh>egr in the licT1.1 background. The authors have a more complex experiment in Figure 3, where licT1.1 is introduced into the hh>egr, wgn RNAi background. However, testing the effect of licT1.1 without wgn would establish a more direct relationship between egr and wgn-p38. *

We have set the crosses for the experiment hh>egr and licT1.1 as suggested. The results will be included in the new version of the manuscript. 1 month.

Reviewer #3

All comments already carried out and included in the transferred manuscript. See next sections.

__Reviewer #4 __

Major comments:

*4.3 In Figure 5, the cells expressing Rpr appeared to be pulled/extruded basally as expected. It would be beneficial to quantify Wgn and Grnd signals along cross-sections and provide higher magnification images of domain boundaries to illustrate differences in TNFR localization and levels. ** The micrographs for Grnd Figure 5B,D, F capture substantial signal from the peripodial epithelium where the salE/Pv> driver is likely not active? *

We will do a thorough quantification of high-resolution stacks of images and include higher magnification of the analyzed stacks. To this aim, we need some more weeks to collect the images of each genotype, processed and quantify them. We propose to do have this work done in two months.

*4.4 The non-autonomous induction of Wgn seems stronger when facing dying Rpr overexpressing cells simultaneously depleted of Eiger compared to Rpr OE alone. Should this be a reproducible, could the authors discuss potential reason for this observation? *

It is difficult to respond this question, without quantification. The quantification suggested in the previous point, will allow us to state if Wgn is more accumulated in rpr +egr than rpr alone. Therefore, the previous point will tell us if there are significant differences and if, so it will help us to discuss it.

Timing: The entire plan can be executed in 2-3 month.

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

Please insert a point-by-point reply describing the revisions that were already carried out and included in the transferred manuscript. If no revisions have been carried out yet, please leave this section empty.

__Reviewer #1 __

1.1 a- *The result in Fig1.H is somehow surprising. How does the overexpression of Egr induce caspase activation in the absence of its receptor Grnd? *

The results of Fig. 1H, in which egr+grndRNAi+wgnRNAi results in high apoptosis indicates that wgn down regulation compromises survival even in the absence of grnd. The reviewer correctly points that “How does the overexpression of Egr induce caspase activation in the absence of its receptor Grnd?”.

There is evidence that Eiger is involved in the regulation of the pro-apoptotic gene head involution defective (hid) in primordial germ cells (Maezawa 2009 Dev. Growth Differ., 51 (4) (2009), pp. 453-461) and in the elimination of damaged neurons during development (Shklover et al., 2015). Moreover, Eiger is necessary for HID stabilization and regulates HID-induced apoptosis independently of JNK signaling (Shklover et al., 2015). Therefore, in our discs egr activation in the absence of grnd and wgn can still result in apoptosis because of the absence of wgn’s survival signal and, presumably, activation of hid.

We have introduced this issue in the text as:

“To check for epistasis between grnd and wgn, we activated hh> egrweak and knocked down both TNFRs. We found high levels of cell death compared to wgn RNAi alone (Fig. 1H and 1I), which demonstrates that wgn down-regulation is dominant over grnd. This is surprising as it is generally assumed that Egr interacts with Grnd to induce apoptosis via JNK, which in turn activates the proapoptotic gene hid (Andersen et al., 2015; Diwanji & Bergmann, 2020; Fogarty et al., 2016; Igaki et al., 2002; Moreno, Yan, et al., 2002; Sanchez et al., 2019; Shlevkov & Morata, 2012). Interestingly, Egr is necessary for HID stabilization and can regulate HID-induced apoptosis independently of JNK (Shklover et al., 2015). Therefore, cells egrweak that downregulate grnd and wgn can still be eliminated because the lack of both Wgn-survival signal and the pro-apoptotic Grnd/JNK signal could result in an alternative pathway of apoptosis.”

*1.2- In Fig.6, it would be relevant to include wengen inactivation within the domain where rpr is expressed to show that wengen is not required autonomously for regeneration (sal>rpr + wgn RNAi). What is the phenotype of the adult wing of sal-lexA>rpr + nub-gal4 >wgn RNAi animals.? *

We have already added a new figure (Fig. S4C) containing this data. As shown, both wgnRNAi alone and wgn RNAi + rpr do not show relevant anomalies and regenerate normally. Therefore, we conclude that wgn is not autonomously required for regeneration.

The adult wings sal-lexA>rpr + nub-gal4 >wgn RNA result in a strong aberration, as regeneration is inhibited. This experiment has been also added in another figure (Fig. S4B) it is done.

*1.4 Minor- In fig.1I, it is surprising that knockdown of neither Grnd nor dTRAF2 significantly affects Egr-induced apoptosis *

After applying a One-Way ANOVA test to compare all the groups against all the groups in fig. 1B no significative differences were detected between Control and RNAi grnd or RNAi dTRAF2 (p>0,05). But if we apply a Student’s T test, which is less restrictive, we obtain, indeed, significative differences:

Control vs. RNAigrnd p=9,48x10-7

Control vs. RNAi dTRAF2 p=2,47x10-7

We have now added in the text:

“Note that when egrweak cells downregulated dTRAF2 or grnd the cell death area ratio is slightly lower than egrweak alone (Fig. 1I), comfirming that dTRAF2 and Grnd contribute to apoptosis in egrweak cells.”

*1.5 Minor The ability of the wing disc to regenerate has been associated with the induction of a developmental delay mediated by Dilp8. Are the authors observing this developmental delay is the case of sal-lexA>rpr + Ap-gal4 >wgn RNAi or sal-lexA>rpr + Ci-Gal4>wgn RNAi *

The developmental delay due to Dilp8 has been observed by many laboratories, indeed. The question of the reviewer is relevant because if there is no delay in pupariation, regeneration could be compromised not because regeneration has been affected but because after pupariation regeneration is impeded.

However, delay in pupariation has been found in our experiments. Usually for 11hrs of heat shock (to induce apoptosis) we found 1-2 days of delay.

We have added the following text:

“The ability of the wing disc to regenerate after genetic ablation has been associated with the induction of a developmental delay (Colombani et al., 2012; Garelli et al., 2012; Jaszczak et al., 2015; Katsuyama et al., 2015; Smith-Bolton et al., 2009). All genotypes analyzed in figure 6 showed a similar developmental delay of 1-2 days (at 17ºC) after genetic ablation in comparison to the animals of the same genotype in which no genetic ablation was induced, i.e. developed continuously at 17ºC (Fig. S4A). After the adults emerged, the wings were dissected, and regeneration was analyzed.”

*1.7 Minor - The investigation of the evolutionary origin of TNFR in drosophila included in Fig.2 is cutting a bit the flow of the results. *

The evolutionary origin starts now with a sentence that can smoothen the flow and few changes in that paragraph have been made:

“Opposing roles between proteins of the TNFR superfamily suggests that they have an ancient origin and have followed divergent evolutionary paths. To track the differences observed between grnd and wgn, we decided to investigate the evolutionary origin of these two Drosophila genes.”

*1.8 Minor The authors could explain in more details the double transactivation system for non-fly specialists. *

The entire section has been re-written in Material and Methods.

*1.9 Minor - It can be interesting to include and/or discuss these few references: *

*PLoS Genet. 2019 Aug; 15(8): e1008133. ** PLoS Genet. 2022 Dec 5;18(12):e1010533. FEBS Lett. 2023 Oct;597(19):2416-2432. *

*Curr Biol. 2016 Mar 7;26(5):575-84. *

*Nat Commun. 2020 Jul 20;11(1):3631. **

*

All these references, and few others, have been introduced in the text.

__Reviewer #2 __ *2. 1. The authors find that wgn RNAi enhances hh>egr-induced apoptosis. They validate the results with two independent RNAi lines in Figure S1. However, Figure S1 is missing a control without wgn RNAi, and therefore, the results are difficult to assess. *

Fig S1A now contains this control.

• 1. Are the two independent wgn RNAi lines targeting different regions of the coding sequence? *

As the regions targeted by the 2 RNAi’s are different, see below, we have included in the text:

“This observation was corroborated with an independent RNAi-wgn strain targeting a different region in the coding sequence (Fig. S1A and S1B). “

Bloomington BL55275 (dsRNA-HMCO3962)

VDRC GD9152 (dsRNA-GD3427)

*2.7. In Figure 4, the authors show that egr expression induces ROS and performs anti-oxidant experiments. This part could be strengthened if they show that the ROS sensor signal disappears after Sod::Cat expression. *

We had done this experiment and there is a definitively drop in Mitosox in discs in which the weak allele of egr is active. We have included this new image in Figure 4G and in the text.

*2.8. How effective is egr RNAi? In Figure 5E, F, the authors knock down egr and obtain negative results. Based on this, the authors argue that Wgn localization occurs through an egr-independent mechanism. Drawing strong conclusions based on a negative result with egr RNAi is not a good practice since one cannot rule out residual egr activity that mediates the effect (of course , because there is cell death as well, death cells express egr). I suggest either finding ways to completely abolish egr function, or tone down the conclusion. *

We have used ‘after knocking down eiger’ instead of in the ‘absence’ or ‘abolish’ eiger.

• 1. Figure 6 shows that wgn RNAi aggravates the reaper phenotype. What's missing is a control that expresses wgn RNAi but not reaper. *

Control experiments using the UAS-wgnRNAi in the absence of rpr are now shown in figure S4C.

Reviewer #3 ____ 3.1.Minor Fig 6C-E would need a control disc without induced apoptosis (ie wildtype disc) stained for phospho-p38 as a baseline comparison. This is important to judge the significance of the remaining phospho-p38 in panel E where wgn is knocked down. The authors write ** " However, after knocking down wgn, phosphorylated p38 in the wing pouch ** surrounding the apoptotic cells was abolished (Fig. 6E)." *Depending on the amount of phospho-p38 in control discs, this may need to be rephrased to "strongly reduced" instead of "abolished". *

A control disc stained with P-p38 has been added in Figure 6.

We have changes ‘abolished’ by ‘strongly reduced’.

* 3.2. This sentence in the Intro needs fixing because TNFa doesn't transduce the signal from TNFR to Ask1 since it's upstream of TNFR: "TNFα can transduce the TRAF-mediated signal from TNFR to Ask1 to modulate its activity (Hoeflich et al., 1999; Nishitoh et al., 1998, p. 0; Obsil & Obsilova, 2017; Shiizaki et al., 2013)." *

We have rephrased this sentence by:

“TNFα binds to TNFRs which in turn interact with TRAFs to transduce the signal to Ask1 to modulate its activity”.

*3.3a In the results section, the authors start by ectopically overexpressing Eiger. Are there conditions where Eiger expression is induced in the wing? If yes, it would be helpful for the reader to mention that this system with the genetically GAL4-induced expression of Eiger aims to phenocopy one of these conditions. *

Eiger ectopic expression has been induced in the wing to generate apoptosis. This is a common technique in Drosophila, and the Reviewer3 is right that a sentence should be useful for the reader.

A sentence has been introduced at the beginning of the results section:

“Ectopic expression of egr in Drosophila imaginal discs results in JNK-dependent apoptosis (Brodsky et al., 2004; Igaki et al., 2002; Moreno, Yan, et al., 2002).”

*3.3b Fig 2C is not very self-explanatory: it is worth writing out what Hsa (H. sapiens), Bla and Sco stand for (there is plenty of space). *

We have re-designed figure 2 to make it more self-explanatory.

*3.4. This sentence is confusing: ** " ...Wgn localization were due to ROS or to the expression of egr, we used RNAi to knock down egr in the apoptotic cells and found that reduced Egr/TNFα had no effect on Wgn localization (Fig. 5E, 5F)." The authors may want to specify that Wgn is still accumulated even without Egr. ("No effect" is unclear). *

This sentence has been modifies as:

“Wgn localization were due to ROS or to the expression of egr, we used RNAi to knock down egr in the apoptotic cells and found that Wgn accumulation was not altered by the knocking down Egr/TNFα (Fig. 5E, 5F). “

*3.5 Comment. It discovers that Wengen is activated by ROS. In fact, since Wengen binds TNF with an affinity that is several orders of magnitude lower than Grindelwald, and since Wengen is not even located at the cell membrane, these data call into question whether Wengen is a TNF receptor, or a ROS receptor? Could the authors comment on this ? Could it be that the results obtained in the past showing that Wengen is activated by TNF were actually due to TNF inducing apoptosis, leading to production of ROS, leading to activation of Wengen?

*

We totally agree with Reviewer#3. We have added a final paragraph in the discussion section.

“We speculate that the subcellular location of Wgn and Grnd may contribute to the different functions of both receptors. Grnd is more exposed at the apical side of the plasma membrane, which makes this receptor more accessible for ligand interactions (Palmerini et al., 2021). Wgn, embedded in cytoplasmic vesicles, is less accessible to the ligand and could be more restricted to being activated by local sources of signaling molecules, such as ROS. In contrast to initial reports (Kanda et al., 2002; Kauppila et al., 2003), los-of-function of wgn does not rescue Egr-induced apoptosis in the Drosophila eye (Andersen et al., 2015), which supports our observation in the wing that Wgn is not required for Egr-induced apoptosis. Instead, Egr-induced apoptosis generates ROS which target intracellular Wgn to foster a cell survival program of cells close to the apoptotic zone.”

__Reviewer #4 __

*4.1 b Are phospho-p38 levels increased in cells expressing Egr[weak]? *

We have the results of these experiments. To respond to this point, a new figure has been added (Fig. S4) in which we show the P-p38 levels are increased (non-autonomously) in egrw, as previously found for reaper. In addition, we show that egrw + activation of p35 and egrw + activation of Sod1::Cat results in strong reduction of P-p38. This indicates that P-p38 is stimulated by the ROS produced by apoptotic cells.

The text now:

“It is worth noting that cells egrw induce phosphorylation of p38 in neighboring cells (Fig. S4A) and that, as previously found for rpr (REF), depends on ROS generated by egrw apoptotic cells (Fig. S4B, C).”

*4.2 In Figure 4C it appears that the Dcp-1 positive cells move apically rather than basally. Including nuclear staining would be very informative allowing assessment of tissue morphology. ** The authors focus on the pouch region of the wing imaginal disc, where phenotypes are strong and obvious. However, the hh-Gal4 driver also affects posterior cells in the hinge and notum, where the effects of Eiger[weak] overexpression seem weaker (e.g., minimal to no MitoSox signal in hinge and notum posterior cells). Could the authors explain this observation? *

Point 1: Actually, cells move more basally, though some move more apical as well. Depending on the section cells the image could be confusing. To solve that, we show now a plane on these discs at apical and a plane basal. Both high magnifications. There one can see that there is more concentration of pyknotic nuclei basally. We have added this observation in a new supplementary figure (Fig. S3) and the corresponding text in page 5: “Apoptotic cells in egrweak are characterized by pyknotic nuclei and are positive for Dcp1. These cells tend to concentrate in the basal side of the epithelium, although some are scattered apically (Fig. S3). Accumulation of Wgn was observed in healthy anterior cells adjacent to both apical and basal egrweak cells (Fig. 4, Fig. S3A, B).”

Point 2 Comment on MitoSOX in notum: At the stages of the imaginal discs used in this study, almost all notum cells are anterior compartment. The hh positive cells in notum much less abundant, therefore most of the staining was found in the posterior compartment of the wing pouch.

*4.5 Figure 6 C-E. Does WgnRNAi potentiates and GrndRNAi suppress Rpr-induced apoptosis similarly to their effects when knocked down in Eiger[weak]OE cells? *

The areas controlled by salE/Pv >rpr (dotted lines) are full of pycnotic nuclei, which indicates concentration of apoptotic cells in all genotypes shown.

Thus, in the conditions generated here, apoptosis is not inhibited and grnd RNAi does not interfere with the activation of P-p38. In wgn knock down, phospho-p38 is strongly inhibited, indicating the importance of wgn in phosphorylation of p38.

To clarify this point, we have added in the text: “Note that rpr-induced apoptosis is not suppressed after knocking down grnd or wgn.” Also in the figure legend we added: “White lines in the confocal images outline the salE/Pv-LHG,LexO-rpr dark area full of pyknotic nuclei of apoptotic cells.”

4.6 The activation of p38 following salE/Pv>rpr-mediated ablation as shown by immunostaining is noteworthy. While loss Grnd knockdown leads to phospho-p38 signal enrichment around the rpr-expressing cells, WgnRNAi results in reduced phospho-p38 signal in the wing pouch but also beyond the nub-expression domain. Do salE/Pv>rpr nub>WgnRNAi cells still generate ROS?

So far there is no evidence of Wengen as a ROS scavenger. We have evidence that ROS (using MitoSox probe) are produced in egrweak + Wgn RNAi cells. Thus, the inhibition of wgn expression does not block ROS production. A new figure shows this observation (Figure S4A).

4.7 Are ROS responsible for the long-range signaling and p38 activation, referring to authors' previous work Santabarbara-Ruiz et al., 2019, PLoS Genet 15(1): e1007926. https://doi.org/10.1371/journal. pgen.1007926, Figure 5G?

ROS are responsible for p38 activation as shown in a new figure (Fig. S4). In this new figure egrweak is activated in hh, and p38 is most of cells in the posterior compartment, and also anterior. However, after blocking apoptosis or ROS production, this P-p38 is reduced.

4.8 Minor I propose rephrasing the description of "UAS-Egr[weak] transgene, a strain that produces a reduced Egr/TNFα activity". It could imply a loss of function strain rather than a transgene that causes mild/moderate Egr overexpression.

The sentence has been rephrases as suggested (End of the first paragragraph in results section).

*4.9 Minor. I recommend the authors to revise the charts for improved clarity in genotype representation. For example, in Figure 1I, the label "control-GFP" might be misleading. It would be beneficial to specify that "control" refers to Eiger[weak] alone with other manipulations being done simultaneously with Eiger[weak] overexpression. *

All charts have been revised.

4.10 Minor. Additionally, considering that individuals with color blindness may struggle to differentiate between red and green colors, I strongly suggest using a color-blind-friendly palette, especially in Figure 4A, C, G, and Figure 4A, C, E." ** All images have been revised for color blind code.

• 11 Minor. Providing detailed information regarding the reagents used in the study, such as Catalogue Numbers or RRIDs, is beneficial for enhancing reproducibility. *

We have added the RRID and Cat #. If no ID was available, we added the reference or contact.

4.12 This reviewer points two limitations that we are now trying to solve:

*Limitations: *

*Quality of the imaging – higher magnification images and quantification would enhance the study. ** The summarizing model may contain excessive speculations that lack support from the data or references to the existing literature. *

Quality of imaging. We have now an extra supplemental figure with higher magnifications. Extra higher magnifications will be included in the next version as well as quantification, as exposed for the Revision Plan points 4.3 and 4.4.

Model: We have re-written the paragraph on the model, introduced references and drop some speculations. We hope the current version will be more convincing for the reader.

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

Please include a point-by-point response explaining why some of the requested data or additional analyses might not be necessary or cannot be provided within the scope of a revision. This can be due to time or resource limitations or in case of disagreement about the necessity of such additional data given the scope of the study. Please leave empty if not applicable.

Reviewer 1

*1.3. Is the overexpression of Wengen sufficient to increase tissue regeneration? *

The suggestion of the reviewer is a key point in regeneration biology: how to accelerate regeneration?

We have demonstrated that Wengen is upstream the Ask1-p38 axis that drives regeneration. The reviewer wonders if Wengen overexpression can result in increase in regeneration. In a previous work we have demonstrated that p38 activation is key for regeneration but its overexpression can be deleterious (Esteban-Collado et al., 2021). Only in discs that sensitized for low p38 (starvation, low Akt, Ask1S83A mutant), the overexpression rescues regeneration. Therefore, the levels of the Wgn-Ask1-p38 have to be very tightly controlled. An excess will be deleterious. We are aware of the importance of the question, but at this point we do not have the technology to finely control Wgn-Ask1-p38 levels to do this experiment.

1.6 Minor - It possible to test the induction of apoptosis in a wgn null mutant background to see if the phenotype is even stronger than the one observed with RNAi (the wgn RNAi is induced at the same time than egr or rpr overexpression).

Flies wgnKO survive, but they gave us problems when carrying transgenes for our design of genetic ablation. Indeed, we tried to generate wgnKO carrying Gal4+tubGal80+eigerweak without success.

In addition, the reason we have used an inducible mutant is because it allows us to work in time and space without altering expression in other cell types beyond wing discs. Wgn is required in other organs during development like gut, trachea and axon growth, etc.., and thus, we ensure the affected cells belong to the tissue analyzed.

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 #4

Evidence, reproducibility and clarity

The study by Florenci Serras and colleagues presents compelling evidence highlighting distinct functions of the two Drosophila TNFRs, Wengen (Wgn) and Grindewald (Grnd) in developing imaginal epithelia. The study shows that while Grnd-dTraf2-Tak1 module controls apoptosis in Eiger (Egr)-dependent manner, Wgn-dTraf1-Ask1 promotes survival likely via p38 signaling independent of Egr. Their phylogenetic analysis underscores the ancient origin of both receptors while revealing their divergent evolutionary path, manifested by markedly different CRD sequences. Moreover, Wgn shows higher degree of similarity to mammalian TNFRs. Using functional genetics and confocal imaging of immunostained wing imaginal discs, the authors confirm the differential localization of Wgn and Grnd in imaginal cells, consistent with several recent studies. Interestingly, they demonstrate distinct responses between Grnd that is internalized upon Eiger[weak] overexpression from the plasma membrane, and Wgn, which cytoplasmic levels decrease in Eiger[weak] OE cells but become enriched in neighboring wild type cells. The non-autonomous accumulation of Wgn required ROS but not Eiger production by dying cells. Finally, employing an elegant double-driver lexA/lexO and Gal4/UAS system, enabling independent gene manipulation in specific domains of the wing imaginal discs, the authors established the essential role of Wgn, but not Grnd, in the regenerative response to apoptosis, including the occurrence of phosphorylated p38.

Major comments:

The conclusion regarding the protective role of p38 in response to Egr[weak] should be supported by a p38 knockdown experiment. Are phospho-p38 levels increased in cells expressing Egr[weak]?

In Figure 4C it appears that the Dcp-1 positive cells move apically rather than basally. Including nuclear staining would be very informative allowing assessment of tissue morphology. The authors focus on the pouch region of the wing imaginal disc, where phenotypes are strong and obvious. However, the hh-Gal4 driver also affects posterior cells in the hinge and notum, where the effects of Eiger[weak] overexpression seem weaker (e.g., minimal to no MitoSox signal in hinge and notum posterior cells). Could the authors explain this observation?

In Figure 5, the cells expressing Rpr appeared to be pulled/extruded basally as expected. It would be beneficial to quantify Wgn and Grnd signals along cross-sections and provide higher magnification images of domain boundaries to illustrate differences in TNFR localization and levels. The micrographs for Grnd Figure 5B,D, F capture substantial signal from the peripodial epithelium where the salE/Pv> driver is likely not active?

The non-autonomous induction of Wgn seems stronger when facing dying Rpr overexpressing cells simultaneously depleted of Eiger compared to RprOE alone. Should this be a reproducible, could the authors discuss potential reason for this observation?

Figure 6 C-E. Does WgnRNAi potentiates and GrndRNAi suppress Rpr-induced apoptosis similarly to their effects when knocked down in Eiger[weak]OE cells? The activation of p38 following salE/Pv>rpr-mediated ablation as shown by immunostaining is noteworthy. While loss Grnd knockdown leads to phospho-p38 signal enrichment around the rpr-expressing cells, WgnRNAi results in reduced phospho-p38 signal in the wing pouch but also beyond the nub-expression domain. Do salE/Pv>rpr nub>WgnRNAi cells still generate ROS? Are ROS responsible for the long-range signaling and p38 activation, referring to authors' previous work Santaba ́rbara-Ruiz et al., 2019, PLoS Genet 15(1): e1007926. https://doi.org/10.1371/journal. pgen.1007926, Figure 5G?

Minor comments:

I propose rephrasing the description of "UAS-Egr[weak] transgene, a strain that produces a reduced Egr/TNFα activity". It could imply a loss of function strain rather than a transgene that causes mild/moderate Egr overexpression.

I recommend the authors to revise the charts for improved clarity in genotype representation. For example, in Figure 1I, the label "control-GFP" might be misleading. It would be beneficial to specify that "control" refers to Eiger[weak] alone with other manipulations being done simultaneously with Eiger[weak] overexpression. Additionally, considering that individuals with color blindness may struggle to differentiate between red and green colors, I strongly suggest using a color-blind-friendly palette, especially in Figure 4A, C, G, and Figure 4A, C, E."

Providing detailed information regarding the reagents used in the study, such as Catalogue Numbers or RRIDs, is beneficial for enhancing reproducibility.

Significance

This is a very solid study that uncovers unique roles of Drosophila TNFRs in regulating imaginal cell behaviors crucial for tissue regeneration. It expands our knowledge on processes controlled by TNFR-mediated signaling, highlighting the potential for ligand-independent regulation. The study nicely complements recent findings by several laboratories (Letizia et al., 2023; Loudhaief et al., 2023; Palmerini et al., 2021). Beyond its contribution to fundamental biology, the study has biomedical implication for regenerative medicine. It emphasizes the necessity of balancing TNFR activities, downstream signaling and their dependence on ligands, providing important insights for the development of receptor agonists or antagonists. The findings are relevant to audience interested in developmental and regenerative biology, gene evolution.

Strengths: functional genetics revealing distinctive roles for the two TNFRs in Drosophila and their dependency on ligand in the paradigm of tissue regeneration.

Limitations: quality of the imaging - higher magnification images and quantification would enhance the study. The summarizing model may contain excessive speculations that lack support from the data or references to the existing literature.

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:

TNF receptors have a broad range of possible function, from inducing apoptosis to promoting cell survival and proliferation. How this works is not completely understood. Drosophila has two TNF receptors, Wengen and Grindelwald. This manuscript nicely shows that Grindelwald is pro-apoptotic while Wengen promotes cell survival and proliferation. Strikingly, if TNF is expressed in Drosophila tissue, knockdown of the receptor Wengen leads to elevated levels of apoptosis, clearly showing its cell-protective function. Interestingly, the authors find that Wengen is activated by ROS produced by neighboring dying cells - regardless of whether they are dying due to TNF signaling or not - and that Wengen then activates p38 downstream to mediate a regenerative response.

Major comments:

Overall the conclusions are interesting, clear and convincing. The data are of very good quality. I only have a few minor comments below.

Minor comments:

1. Fig 6C-E would need a control disc without induced apoptosis (ie wildtype disc) stained for phospho-p38 as a baseline comparison. This is important to judge the significance of the remaining phospho-p38 in panel E where wgn is knocked down. The authors write " However, after knocking down wgn, phosphorylated p38 in the wing pouch surrounding the apoptotic cells was abolished (Fig. 6E)." Depending on the amount of phospho-p38 in control discs, this may need to be rephrased to "strongly reduced" instead of "abolished".
2. This sentence in the Intro needs fixing because TNFa doesn't transduce the signal from TNFR to Ask1 since it's upstream of TNFR: "TNFα can transduce the TRAF-mediated signal from TNFR to Ask1 to modulate its activity (Hoeflich et al., 1999; Nishitoh et al., 1998, p. 0; Obsil & Obsilova, 2017; Shiizaki et al., 2013)."
3. In the results section, the authors start by ectopically overexpressing Eiger. Are there conditions where Eiger expression is induced in the wing? If yes, it would be helpful for the reader to mention that this system with the genetically GAL4-induced expression of Eiger aims to phenocopy one of these conditions.
4. Fig 2C is not very self-explanatory: it is worth writing out what Hsa (H. sapiens), Bla and Sco stand for (there is plenty of space).
5. This sentence is confusing: " ...Wgn localization were due to ROS or to the expression of egr, we used RNAi to knock down egr in the apoptotic cells and found that reduced Egr/TNFα had no effect on Wgn localization (Fig. 5E, 5F)." The authors may want to specify that Wgn is still accumulated even without Egr. ("No effect" is unclear).

Significance

This manuscript makes several important discoveries:

1. it clearly shows that one TNF receptor, Grindelwald, is mainly pro-apoptotic, while the other, Wengen, is mainly pro-survival. This provides a mechanistic explanation for the dual role of the TNF, Eiger.
2. It discovers that Wengen is activated by ROS. In fact, since Wengen binds TNF with an affinity that is several orders of magnitude lower than Grindelwald, and since Wengen is not even located at the cell membrane, these data call into question whether Wengen is a TNF receptor, or a ROS receptor? Could the authors comment on this ? Could it be that the results obtained in the past showing that Wengen is activated by TNF were actually due to TNF inducing apoptosis, leading to production of ROS, leading to activation of Wengen?
3. It was previously shown that damage, for instance in the fly intestine, induces production of ROS, which then activates p38, leading to a proliferative/regenerative response. This manuscript provides a missing mechanistic link, showing that the ROS activates Wengen, which in turn activates p38. This thereby completes the mechanistic chain of events from damage to the regenerative response.

Hence, overall, this is a very interesting study. It will be of interest for a broad audience of people studying TNF signaling, stress signaling and stress response, tissue damage and repair, and regeneration.

My expertise: Drosophila, growth, 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 #2

Evidence, reproducibility and clarity

The Drosophila genome encodes a single TNFa ortholog, eiger (egr), and two TNF Receptros (TNFR), wengen (wgn) and grindelwald (grnd). While egr overexpression can cause apoptosis, the authors here report that wgn and grnd have opposing roles in cell death and survival. Specifically, the authors show evidence that grnd promotes cell death in response to egr expression, while wgn promotes cell survival through the p38 MAP Kinase pathway. They further show that apoptotic cells have high levels of ROS, which activates the wgn-p38 axis for tissue regeneration, independent of egr.

Overall, the manuscript is well written. At the same time, there are some technical concerns and missing controls that need to be addressed. Below are a few specific comments for the authors' consideration:

Major Comments

1. The authors find that wgn RNAi enhances hh>egr-induced apoptosis. They validate the results with two independent RNAi lines in Figure S1. However, Figure S1 is missing a control without wgn RNAi, and therefore, the results are difficult to assess.
2. Are the two independent wgn RNAi lines targeting different regions of the coding sequence?
3. Aside from wgn, other RNAi experiments are not validated through independent RNAi lines. I suggest expanding the Supplemental Figures to reproduce a few key findings with independent RNAi lines.
4. In Figure 1E, the authors show that wgn RNAi enhances cell death caused by hh>egr. What is missing here is a wgn RNAi control without hh>egr. Is there any cell death caused by the loss of wgn alone (without hh>egr)?
5. If wgn RNAi causes some degree of cell death, is the observed effect with hh>egr a significant genetic interaction, or merely additive?
6. Is the wgn-p38 pathway sufficient to block egr induced cell death? The authors could test this by having hh>egr in the licT1.1 background. The authors have a more complex experiment in Figure 3, where licT1.1 is introduced into the hh>egr, wgn RNAi background. However, testing the effect of licT1.1 without wgn would establish a more direct relationship between egr and wgn-p38.
7. In Figure 4, the authors show that egr expression induces ROS and performs anti-oxidant experiments. This part could be strengthened if they show that the ROS sensor signal disappears after Sod::Cat expression.
8. How effective is egr RNAi? In Figure 5E, F, the authors knock down egr and obtain negative results. Based on this, the authors argue that Wgn localization occurs through an egr-independent mechanism. Drawing strong conclusions based on a negative result with egr RNAi is not a good practice since one cannot rule out residual egr activity that mediates the effect. I suggest either finding ways to completely abolish egr function, or tone down the conclusion.
9. Figure 6 shows that wgn RNAi aggravates the reaper phenotype. What's missing is a control that expresses wgn RNAi but not reaper.

Significance

There is now a detailed understanding of mammalian TNFRs, which play pro-apoptotic and non-apoptotic roles depending upon the context. Previous studies had also reported that TNFR1 respond to ROS. By comparison, our understandings of the two TNFRs in Drosophila remain rudimentary. The two receptors have different loss-of-function phenotypes, some of which may be independent of egr signaling. The major significance of this work is in delineating the distinct behaviors of the two Drosophila TNFRs, centering around their pro-apoptotic, or pro-survival properties.

Audience: This study will draw the interest of Drosophila geneticists, those interested in Reactive Oxygen Species and cell death, and evolutionary biologists.

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

Tumor necrosis factor (TNF)-α stands out as a remarkably conserved pro-inflammatory cytokine that plays crucial roles in immunity, tissue repair, and cellular homeostasis. The Drosophila TNF-TNF receptor (TNFR) system, known for its simplicity, combined with a versatile genetic toolkit, has been instrumental in unraveling the intricate mechanisms governing both the physiological and pathological functions mediated by TNF. Recently, the fly TNFR Wengen has been described to have ligand independent functions in maintaining tissue homeostasis and tracheal remodeling. The current manuscript describes a novel TNF/Egr-independent function of Wengen in regulating tissue regeneration in imaginal discs. The authors identify both the upstream regulator (ROS) and the downstream signaling pathway through Ask1/p38 MAPK. The data presented are solid and support an interesting model where ROS emanating from damaged tissue triggers Wgn-dependent signaling in adjacent cells to promote regeneration. Few points could be addressed:

Major points:

• The result in Fig1.H is somehow surprising. How does the overexpression of Egr induce caspase activation in the absence of its receptor Grnd? Is it possible that the loss of function of Wengen on its own has a phenotype? If so, that would suggest that Wgn in addition to its role in regeneration might be implicated in pro-survival processes in homeostatic conditions?
• In Fig.6, it would be relevant to include wengen inactivation within the domain where rpr is expressed to show that wengen is not required autonomously for regeneration (sal>rpr + wgn RNAi). What is the phenotype of the adult wing of sal-lexA>rpr + nub-gal4 >wgn RNAi animals?
• Is the overexpression of Wengen sufficient to increase tissue regeneration?

Minor points:

• In fig.1I, it is surprising that knockdown of neither Grnd nor dTRAF2 significantly affects Egr-induced apoptosis
• The ability of the wing disc to regenerate has been associated with the induction of a developmental delay mediated by Dilp8. Are the authors observing this developmental delay is the case of sal-lexA>rpr + Ap-gal4 >wgn RNAi or sal-lexA>rpr + Ci-Gal4>wgn RNAi
• It possible to test the induction of apoptosis in a wgn null mutant background to see if the phenotype is even stronger than the one observed with RNAi (the wgn RNai is induced at the same time than egr or rpr overexpression).
• The investigation of the evolutionary origin of TNFR in drosophila included in Fig.2 is cutting a bit the flow of the results.
• The authors could explain in more details the double transactivation system for non-fly specialists.
• It can be interesting to include and/or discuss these few references:

PLoS Genet. 2019 Aug; 15(8): e1008133.

PLoS Genet. 2022 Dec 5;18(12):e1010533.

FEBS Lett. 2023 Oct;597(19):2416-2432.

Nat Commun. 2020 Jul 20;11(1):3631.

Curr Biol. 2016 Mar 7;26(5):575-84.

Significance

The understanding of the mechanistic interplay between TNFR in integrating TNF-dependent and independent signals to stimulate distinct downstream responses lays the foundation for investigating whether these insights can be generalized to other members within the TNFR superfamily in all organisms. This work is relevant for a large audience of researchers working in the field of inflammation and TNFR.

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 reading of our manuscript, which we believe has led to substantial improvements.

To aid clarity, we have split Fig. 1 into three separate figures.

For convenience, we have put all major changes in the text in blue.

Reviewer #1

Evidence, reproducibility and clarity

Summary: Hui et al. tackle a crucial question in biology: what factors influence the preference for carbon sources in yeasts?

They reveal that the growth rate on palatinose exceeds that on glucose,

The above statement is incorrect --- we think the reviewer may have confused sugars.

despite palatinose utilization being repressed in the presence of glucose. Consequently, the favored carbon source does not necessarily align with the one supporting the fastest growth rate. The study also delves into potential regulatory mechanisms governing carbon source preference and dismisses certain existing theories, such as the general carbon flux sensing mechanism proposed by Okano et al. [25].

Major comments: None

Minor comments:

The authors suggest that a higher growth rate implies a higher glycolytic flux (l63), a crucial assumption underpinning their interpretation of the absence of a ``general carbon flux sensing mechanism' (l65). To substantiate this significant conclusion, they could calculate the extracellular uptake fluxes (based on the time-course concentrations of biomass and substrates).

This suggestion is a good one, but unfortunately the number of data points in the new Fig. 3 are insufficient to estimate the uptake flux reliably.

To address whether glycolytic flux increases, we have added a new paragraph to the introduction explaining how all the sugars we consider feed upper glycolysis, providing either its first or second metabolite. We therefore think it highly likely that any differences in growth rate are generated by differences in glycolytic flux. Indeed, Hackett et al., 2016, showed that the glycolytic flux increases with growth rate when they changed extracellular glucose concentrations. We now include this reference in the Discussion.

The accumulation of certain by-products is known to be toxic, reducing cellular growth rate (e.g., acetate DOI: 10.1038/srep42135, ethanol DOI: 10.1016/B978-0-12-040308-0.50006-9, etc.), while they can also enhance growth under specific conditions (e.g., acetate DOI: 10.15252/embj.2022113079). Considering this is crucial to rule out certain hypotheses, such as the possibility that a by-product produced during growth on the first carbon source would not modulate growth on the second carbon source, potentially influencing the growth rate differentially in each phase. Although the authors use mutant strains to eliminate the role of some C2 compounds (acetate and ethanol), alternative pathways could be implicated in the (co-)utilization of these by-products. This aspect should be discussed, and ideally, the authors could quantify the time-course concentrations of by-products to assess their potential role.

We agree with the reviewer that extracellular acetate and ethanol may inhibit growth, although budding yeast might be less sensitive than E. coli, the subject of most of the studies provided.

Nevertheless, we think it unlikely that these chemicals modify the decision-making we see. First, the icl1Δ mutant we tested is unable to consume ethanol (Fernandez et al., 1992) or acetate (Lee et al., 2011) --- we now include these references in the SI --- and yet has wild-type behaviour (Fig. S2D). Second, we observe that isomaltase expression strongly decreases in the presence of galactose when we grow cells in a microfluidic device (Fig. S4), just like it does in batch culture (Fig. 3A), even though the constant flow of medium through the device removes any chemicals the cells excrete.

The general flux-sensing regulatory mechanism proposed by Okano et al. [25], which has been dismissed by this study, has recently been questioned, as discussed in DOI: 10.15252/embj.2022113079. This aspect should be included in the discussion.

Okano et al. studied E. coli while we study budding yeast. We therefore have shown that the understanding for that organism does not transfer to our eukaryotic example. We suspect that control in budding yeast combines both flux-sensing and specific regulation, as we say in the discussion, and so we consider our results to build on those of Okano et al.

Significance

Strengths & limitations: The work is robust, and the experiments in the study have been appropriately designed and conducted. The primary question of this study has been tackled using a combination of experimental and computational methods to thoroughly assess various regulatory and functional aspects. However, there are gaps in the data that could enhance key conclusions, notably the absence of glycolytic flux measurements. Moreover, further evidence is needed to substantiate the assertion that by-products do not play a role in carbon source preference.

Advance: This study represents a significant step forward in comprehending the nutritional strategy of microbes. The authors demonstrate that the preferred carbon source may not necessarily be the one supporting the fastest growth rate. Furthermore, they dismiss certain theories that have been proposed to explain the growth strategy of microbes on mixed carbon sources.

Audience: By addressing a fundamental question in life science, this work is important in the field of biology in general and of particular interest in systems biology, biotechnology, synthetic biology, and health. Consequently, it will be of interest to a broad audience.

Reviewer #2

Evidence, reproducibility and clarity

Summary: The authors have used microtiter plates to produce growth profiles on combinations of different sugars. From this data they have evaluated whether the sugars are co-consumed or if there is a preference for either sugar, seen as a diauxic shift. They found diauxie between galactose and the disaccharide palatinose, but co-consumption between palatinose and fructose. They further used strains with perturbations in their GAL regulon to attempt to explain this discrepency.

Major comments:

I unfortunately found a large portion of the present manuscript unintelligable.

Firstly, figures were incorrect to the point I could not dechiffre them: Figure 2A-C have black solid and dashed lines in the legend that are not found in the graph, instead there are orange and blue dashed lines in the graph with no legends. Figure 4C has no description of the y-axis. The growth rates in Figure 1C are very hard to follow, and there are definitely local maxima in both the blue and green profiles that are not being discussed (at 15-20 h). I cannot evaluate the conclusions drawn from the data until these issues have been resolved.

We apologise for the difficulties experienced by this reviewer.

The black lines in the old Fig. 2's legend, now Fig. 4, explain the different styles used: dashed lines are for single sugars regardless of their concentration and full lines are for mixtures regardless of their concentration. We now explicitly say this in the caption.

We have fixed the missing label in what is now Fig. 6C and have moved the statement that we are showing two biological replicates for each set of concentrations earlier in Fig. 2's caption.

We now explore the meaning of the shoulder for the fructose-palatinose mixture in Fig. 2B in the Discussion. This point is not a local maximum, unlike the case for diauxie, because the growth rate always decreases. The shoulder for the glucose-palatinose mixture was likely an artefact generated by measurement noise at low ODs because it was not present when we repeated the experiment. We now use that data for Fig. 2A & B. We also include a new Fig. S5 showing that there are sucrose-palatinose concentrations too that have a similar shoulder.

Secondly, the language in the Results and Discussion sections is confusing. Alternating between present and imperfect tense as well as active and passive form makes it hard to distinguish the authors own results from literature findings (Results are usually written in passive, imperfect tense). Examples are found on lines 24, 29, 37-38, 59, 84, 131, and 165.

We have made both sections flow more smoothly with substantial re-writing. As before, we cite all results that are not our own.

The authors also do not consider the differences and similarities in catabolic pathways for assimilation of galactose, fructose and palatinose. Even if they do not see a reason to continue that as a possible explanation for the co-consumption between fructose and palatinose a discussion of why it is disregarded would not be out of place here.

A good point, and we now state in the Introduction that all the sugars we study feed upper glycolysis.

Significance

There is some novelty to the authors findings, but I would argue it is being overstated in the present manuscript. Some examples of studies looking at catabolite repression, the main cause of diauxie, of sugars other than glucose can be found in: Simpson-Lavy and Kupiec (2019), Gancedo (1998), Prasad and Venkatesh (2008) and Borgstrom et al (2022).

We strongly disagree with this statement. The papers cited do not address, as we do, the co-consumption between two sugars neither of which is glucose. Where they study two sugars, they always study glucose.

Simpson-Lavy and Kupiec, 2019, investigate the interaction between acetate and ethanol, neither of which are sugars. Further, they are not independent carbon sources because cells convert ethanol into acetate when catabolising ethanol.

Gancedo, 1998, is a review of glucose repression and describes how glucose represses the expression of genes for other sugars. Although Gancedo mentions ``galactose repression', this repression is of genes encoding enzymes for gluconeogenesis and the TCA and glyoxylate cycles, not of other sugar regulons, our subject.

Prasad and Venkatesh, 2008, also focus on glucose and the well studied diauxie between glucose and galactose.

Borgstrom et al., 2022, focus too on glucose and growth on glucose and xylose in recombinant strains. The standard laboratory strains we study have not be artificially engineered to consume xylose. They do mention that galactose causes repression of TPS1, which encodes an enzyme that synthesises the storage carbohydrate trehalose. This repression is again not of a sugar catabolic regulon, our subject.

I would not say that the field would be significantly advanced by the publication of this manuscript, and the authors have themselves not explained the application of futhering the understanding palatinose metabolism in yeast. As mentioned above, the catabolite repression potential of galactose is already known, it just hasn't been shown for palatinose specifically before.

We again strongly disagree. Our findings are novel. The reviewer did not provide any evidence for galactose repression of other sugar regulons, which is not widely recognised as we emphasised in the Discussion. We believe that the reviewer has confused the known "galactose repression' of gluconeogenic or TCA-cycle genes with our new report of repression of other sugar regulons in the presence of the sugar catabolised by the regulon.

I would recommend a complete rewrite of the manuscript as presented, with a lower stated novelty, clearer language and comprehensible figures.

Reviewer #3

Evidence, reproducibility and clarity

Summary: Microbes grow at different growth rates in different carbon sources. When more than one carbon sources are present in the media microbes often show a preference over certain carbon sources, and 'non-preferred' carbons sources are used only when the preferred carbon source is exhausted in the media, this process called diauxic shift.

Why microbes exhibit such utilization preference over certain carbon sources, is an interesting question in microbiology and evolutionary biology, and the molecular mechanisms that enable microbes to preferentially use one carbon over another is worth investigating. It is intuitive to think that microbes will prefer to use a carbon source that confers maximum growth rate, but when tested experimentally it has been often observed that a carbon source in which microbes grow at sub optimal growth rate is actually preferentially used.

Although the reviewer states that "it has been often observed that a carbon source in which microbes grow at sub optimal growth rate is actually preferentially used“, we are unaware of this work and would appreciate references, particularly for budding yeast. The most systematic study we know, in E. coli by Aidelberg et al., 2014 --- reference 13, concludes that "the faster the growth rate, the higher the sugar on the hierarchy“, the opposite behaviour.

In this study authors demonstrate that budding yeast prefer to use galactose over palatinose, but not over sucrose or fructose where all three sugars can support faster growth rate compared to palatinose. Authors presented data where preferential galactose use and diauxic shift is observed in the growth curve when galactose and palatinose or glucose and palatinose combinations were used.

No diauxic shift was observed in the growth curve when fructose-palatinose, or sucrose-palatinose combination were used. In fructose-palatinose and sucrose-palatinose combinations growth curves agree more with co-utilization strategies. Authors used transcriptomics and genetic perturbations to decipher the molecular mechanism of such preferential carbon use, and reports preference of galactose over palatinose is achieved by preventing positive feedback of MAL regulon, which encodes the genes for palatinose catabolism. We found this observation is interesting and the molecular mechanism of such preferential carbon use is nicely described in this paper. We also find claims authors made are well supported by experiments. Although catabolite repression and diauxic transitions are known in yeast, and authors also pointed out such previous references, but preferential use of a slower carbon source, i.e. galactose over at least one other fast-growing carbon is interesting enough for publication. We would like to support the publication of this article, but we have major concerns about the data analysis and data presentation. Authors must address our concerns which are mentioned below.

Major comments:

1. This study mainly hinges on growth rate measurements, but we found growth rates are not properly represented in the figures. Growth curves are always shown in linear scale, which makes it almost impossible to compare fast and slow growth when presented in same plot. All growth curves must be shown on log scale.

We have changed all growth curves to log2 scale, following New et al., 2014, rather than Monod's choice of linear scale that we had originally.

Our conclusions are unaffected.

1. Growth rates of the Yeast strain growing individual single carbon sources (galactose, palatinose, sucrose and fructose) should be shown as a figure panel and t-test should be performed to conclude if the individual growth rates are significantly different or not.

We already showed these growth rates in their own panel in Fig. 1B. Following the reviewer's suggestion, we have now added their statistical significance to the caption.

1. Growth phase, lag phase, diauxic shift and post shift growth should be clearly shown in figure 2 and 4, each phase should be clearly marked, carbons used in each phase should be mentioned on the plot. Also, the growth curve must be plotted using log scale.

Although we have changed all growth curves to log scale, we decided against include this additional labelling for two reasons. First, we are presenting evidence that some of the growth we observe is diauxic and labelling the curves as diauxic before we discuss this evidence undermines that discussion. Second, any further labels would clutter the figures, and we believe would hinder rather than help the reader.

Instead we changed the colour scheme and the boldness of the diauxic growth curves in Fig. 2, which we hope the reviewer agrees adds the clarity they felt was missing.

1. Authors has taken in account that MAL12 gene overexpression causes long lag when cells need to switch to maltose from glucose, and shown deletion of IMA1 decreases the lag with subsequent 2% growth rate increase in palatinose. How significant is this increase?

We have confirmed the statistical significance through a t-test and added the results to the caption of Fig. 6C.

1. Authors have an interesting observation that in sucrose-palatinose and fructose palatinose combinations, most probably co utilization of the carbons is taking place. Authors should discuss this in more details. In galactose-palatinose scenario intracellular galactose-based repression of gal80 and subsequent lack of feed forward of the Mal regulon is expected to stop co-utilization of palatinose. As authors have RNA seq data, can they make predictions for other carbon pairs, where sequential utilization can occur based on their model?

We agree and have added more discussion of the fructose- and sucrose-palatinose mixtures to the Discussion and a new figure, Fig. S5.

Our RNAseq data reveals the difference in gene expression caused by an active versus an inactive GAL regulon. In Fig. S11, we show that the hexose transporters HXT2 and HXT7 are down regulated in 0.1% fructose when the GAL regulon is active, perhaps implying that cells are able to prioritise galactose over other hexoses. Nevertheless, to predict if particular carbon sources are therefore favoured, we would need to know whether cells use specific hexose transporters to drive growth on different carbon sources, which has been little investigated.

Minor comments:

1. In figure 5, authors attempted to summarize the model, which is informative, but it will be more useful for non-specific reader if a cell-based cartoon, with transports on surface and catabolic enzymes inside is also added.

We have re-designed Fig. 5, now Fig. 7, following this suggestion and agree it improves clarity.

In this schematic diagram, switch from galactose (blue line) to red line (palatinose) shows a mixed color zone, it's a bit confusing, as this represents a bi-stable state. Authors should clearly comment on possibility of biostability while discussing their proposed mechanism.

In the new figure, this part has been removed.

1. The author may want to put their work in the context of other recent observations that bacteria do not try to maximize their growth rates in many conditions. Fast growth is often associated with expansive tradeoffs, and a carbon source which confers fast growth rate may confer selective disadvantage. Thus, there are evolutionary benefits of sub-optimal growth, which could be discussed in the manuscript. In this regard a recent study (bioRxiv (2023) doi:10.1101/2023.08.22.554312.) has established the link between resource allocation strategies, growth rates and tradeoffs, which may be taken in account while discussing. Are there any known tradeoffs, when galactose is used over palatinose and which is not the case sucrose or fructose?

This is an interesting reference looking at growth on a single carbon source. We are unaware of similar tradeoffs relevant to our study. For example, we see little evidence for a constraint on the proteome because in a strain with a constitutively active GAL regulon there is no change in phenotype if we delete the genes for the three highly expressed GAL enzymes (Fig. S6B). Nevertheless and as we state in the penultimate paragraph of the Discussion, we agree that such a constraint must exist, although perhaps this constraint is ecological.

Referees cross-commenting

As other reviewers pointed out, this study has merit and addressed interesting questions, but needed to be written well in a more understandable form, we agree with this assessment. Also figures must be made much clearer, as all of the reviewers pointed out. In summary, this is an interesting study, but needs some work before publication.

Significance

General assessment: Strength and limitations:

This study addressed an interesting question regarding resource preference and growth rate optimization in microbes. This is an important question in the field. Study is well designed and claims are backed up with experimental results. One of the limitations of the study is lack of predictability. Authors explained the mechanism for one pair of carbon sources, but how applicable that will be in general is not clear.

We would argue that one of our important findings is to demonstrate that the scientific community is missing the information needed to make such predictions. We provide a counter example to the generally accepted belief that accurate predictions can be made using growth rates. Our work poses the question: what then are the physiological variables required to predict how a cell will consume a pair of carbon sources?

Advance: This study helps to advance our knowledge. Their observation regarding preferential utilization of a carbon source which supports slower growth over a carbon source which can support faster growth, and the molecular mechanism provided will help researchers to understand resource allocation strategies better.

Audience: Microbiology, systems biology, evolutionary biology, fermentation and bio process engineering research.

Reviewer expertise: Biochemistry, systems biology, metabolic strategies and tradeoffs in microbes, microbial ecology.

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

Review of the paper by Yu Huo et al.

Summary:

Microbes grow at different growth rates in different carbon sources. When more than one carbon sources are present in the media microbes often show a preference over certain carbon sources, and 'non-preferred' carbons sources are used only when the preferred carbon source is exhausted in the media, this process called diauxic shift. Why microbes exhibit such utilization preference over certain carbon sources, is an interesting question in microbiology and evolutionary biology, and the molecular mechanisms that enable microbes to preferentially use one carbon over another is worth investigating. It is intuitive to think that microbes will prefer to use a carbon source that confers maximum growth rate, but when tested experimentally it has been often observed that a carbon source in which microbes grow at sub optimal growth rate is actually preferentially used. In this study authors demonstrate that budding yeast prefer to use galactose over palatinose, but not over sucrose or fructose where all three sugars can support faster growth rate compared to palatinose. Authors presented data where preferential galactose use and diauxic shift is observed in the growth curve when galactose and palatinose or glucose and palatinose combinations were used.

No diauxic shift was observed in the growth curve when fructose-palatinose, or sucrose-palatinose combination were used. In fructose-palatinose and sucrose-palatinose combinations growth curves agree more with co-utilization strategies. Authors used transcriptomics and genetic perturbations to decipher the molecular mechanism of such preferential carbon use, and reports preference of galactose over palatinose is achieved by preventing positive feedback of MAL regulon, which encodes the genes for palatinose catabolism. We found this observation is interesting and the molecular mechanism of such preferential carbon use is nicely described in this paper. We also find claims authors made are well supported by experiments. Although catabolite repression and diauxic transitions are known in yeast, and authors also pointed out such previous references, but preferential use of a slower carbon source, i.e. galactose over at least one other fast-growing carbon is interesting enough for publication. We would like to support the publication of this article, but we have major concerns about the data analysis and data presentation. Authors must address our concerns which are mentioned below.

Major comments:

1. This study mainly hinges on growth rate measurements, but we found growth rates are not properly represented in the figures. Growth curves are always shown in linear scale, which makes it almost impossible to compare fast and slow growth when presented in same plot. All growth curves must be shown on log scale.
2. Growth rates of the Yeast strain growing individual single carbon sources (galactose, palatinose, sucrose and fructose) should be shown as a figure panel and t-test should be performed to conclude if the individual growth rates are significantly different or not.
3. Growth phase, lag phase, diauxic shift and post shift growth should be clearly shown in figure 2 and 4, each phase should be clearly marked, carbons used in each phase should be mentioned on the plot. Also, the growth curve must be plotted using log scale.
4. Authors has taken in account that MAL12 gene overexpression causes long lag when cells need to switch to maltose from glucose, and shown deletion of IMA1 decreases the lag with subsequent 2% growth rate increase in palatinose. How significant is this increase?
5. Authors have an interesting observation that in sucrose-palatinose and fructose palatinose combinations, most probably co utilization of the carbons is taking place. Authors should discuss this in more details. In galactose-palatinose scenario intracellular galactose-based repression of gal80 and subsequent lack of feed forward of the Mal regulon is expected to stop co-utilization of palatinose. As authors have RNA seq data, can they make predictions for other carbon pairs, where sequential utilization can occur based on their model?

Minor comments

1. In figure 5, authors attempted to summarize the model, which is informative, but it will be more useful for non-specific reader if a cell-based cartoon, with transports on surface and catabolic enzymes inside is also added.

In this schematic diagram, switch from galactose (blue line) to red line (palatinose) shows a mixed color zone, it's a bit confusing, as this represents a bi-stable state. Authors should clearly comment on possibility of biostability while discussing their proposed mechanism. 2. The author may want to put their work in the context of other recent observations that bacteria do not try to maximize their growth rates in many conditions. Fast growth is often associated with expansive tradeoffs, and a carbon source which confers fast growth rate may confer selective disadvantage. Thus, there are evolutionary benefits of sub-optimal growth, which could be discussed in the manuscript. In this regard a recent study (bioRxiv (2023) doi:10.1101/2023.08.22.554312.) has established the link between resource allocation strategies, growth rates and tradeoffs, which may be taken in account while discussing. Are there any known tradeoffs, when galactose is used over palatinose and which is not the case sucrose or fructose?

Referees cross-commenting

As other reviewers pointed out, this study has merit and addressed interesting questions, but needed to be written well in a more understandable form, we agree with this assessment. Also figures must be made much clearer, as all of the reviewers pointed out. In summary, this is an interesting study, but needs some work before publication.

Significance

General assessment: Strength and limitations: This study addressed an interesting question regarding resource preference and growth rate optimization in microbes. This is an important question in the field. Study is well designed and claims are backed up with experimental results. One of the limitations of the study is lack of predictability. Authors explained the mechanism for one pair of carbon sources, but how applicable that will be in general is not clear.

Advance: This study helps to advance our knowledge. Their observation regarding preferential utilization of a carbon source which supports slower growth over a carbon source which can support faster growth, and the molecular mechanism provided will help researchers to understand resource allocation strategies better.

Audience: Microbiology, systems biology, evolutionary biology, fermentation and bio process engineering research.

Reviewer expertise: Biochemistry, systems biology, metabolic strategies and tradeoffs in microbes, microbial ecology.

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 used microtiter plates to produce growth profiles on combinations of different sugars. From this data they have evaluated whether the sugars are co-consumed or if there is a preference for either sugar, seen as a diauxic shift. They found diauxie between galactose and the disaccharide palatinose, but co-consumption between palatinose and fructose. They further used strains with perturbations in their GAL regulon to attempt to explain this discrepency.

Major comments: I unfortunately found a large portion of the present manuscript unintelligable.

Firstly, figures were incorrect to the point I could not dechiffre them: Figure 2A-C have black solid and dashed lines in the legend that are not found in the graph, instead there are orange and blue dashed lines in the graph with no legends. Figure 4C has no description of the y-axis. The growth rates in Figure 1C are very hard to follow, and there are definitely local maxima in both the blue and green profiles that are not being discussed (at 15-20 h). I cannot evaluate the conclusions drawn from the data until these issues have been resolved.

Secondly, the language in the Results and Discussion sections is confusing. Alternating between present and imperfect tense as well as active and passive form makes it hard to distinguish the authors own results from literature findings (Results are usually written in passive, imperfect tense). Examples are found on lines 24, 29, 37-38, 59, 84, 131, and 165.

The authors also do not consider the differences and similarities in catabolic pathways for assimilation of galactose, fructose and palatinose. Even if they do not see a reason to continue that as a possible explanation for the co-consumption between fructose and palatinose a discussion of why it is disregarded would not be out of place here.

Significance

There is some novelty to the authors findings, but I would argue it is being overstated in the present manuscript. Some examples of studies looking at catabolite repression, the main cause of diauxie, of sugars other than glucose can be found in: Simpson-Lavy and Kupiec (2019), Gancedo (1998), Prasad and Venkatesh (2008) and Borgstrom et al (2022).

I would not say that the field would be significantly advanced by the publication of this manuscript, and the authors have themselves not explained the application of futhering the understanding palatinose metabolism in yeast. As mentioned above, the catabolite repression potential of galactose is already known, it just hasn't been shown for palatinose specifically before.

I would recommend a complete rewrite of the manuscript as presented, with a lower stated novelty, clearer language and comprehensible figures.

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: Hui et al. tackle a crucial question in biology: what factors influence the preference for carbon sources in yeasts? They reveal that the growth rate on palatinose exceeds that on glucose, despite palatinose utilization being repressed in the presence of glucose. Consequently, the favored carbon source does not necessarily align with the one supporting the fastest growth rate. The study also delves into potential regulatory mechanisms governing carbon source preference and dismisses certain existing theories, such as the general carbon flux sensing mechanism proposed by Okano et al. [25].

Major comments: None

Minor comments:

• The authors suggest that a higher growth rate implies a higher glycolytic flux (l63), a crucial assumption underpinning their interpretation of the absence of a "general carbon flux sensing mechanism" (l65). To substantiate this significant conclusion, they could calculate the extracellular uptake fluxes (based on the time-course concentrations of biomass and substrates).
• The accumulation of certain by-products is known to be toxic, reducing cellular growth rate (e.g., acetate DOI: 10.1038/srep42135, ethanol DOI: 10.1016/B978-0-12-040308-0.50006-9, etc.), while they can also enhance growth under specific conditions (e.g., acetate DOI: 10.15252/embj.2022113079). Considering this is crucial to rule out certain hypotheses, such as the possibility that a by-product produced during growth on the first carbon source would not modulate growth on the second carbon source, potentially influencing the growth rate differentially in each phase. Although the authors use mutant strains to eliminate the role of some C2 compounds (acetate and ethanol), alternative pathways could be implicated in the (co-)utilization of these by-products. This aspect should be discussed, and ideally, the authors could quantify the time-course concentrations of by-products to assess their potential role.
• The general flux-sensing regulatory mechanism proposed by Okano et al. [25], which has been dismissed by this study, has recently been questioned, as discussed in DOI: 10.15252/embj.2022113079. This aspect should be included in the discussion.

Significance

Strengths & limitations: The work is robust, and the experiments in the study have been appropriately designed and conducted. The primary question of this study has been tackled using a combination of experimental and computational methods to thoroughly assess various regulatory and functional aspects. However, there are gaps in the data that could enhance key conclusions, notably the absence of glycolytic flux measurements. Moreover, further evidence is needed to substantiate the assertion that by-products do not play a role in carbon source preference.

Advance: This study represents a significant step forward in comprehending the nutritional strategy of microbes. The authors demonstrate that the preferred carbon source may not necessarily be the one supporting the fastest growth rate. Furthermore, they dismiss certain theories that have been proposed to explain the growth strategy of microbes on mixed carbon sources.

Audience: By addressing a fundamental question in life science, this work is important in the field of biology in general and of particular interest in systems biology, biotechnology, synthetic biology, and health. Consequently, it will be of interest to a broad audience.

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)):

This well done and interesting paper examining the connection between TXNIP and GDF15. The main thrust is that TXNIP upregulation chemotherapies, such as Oxa, results in an a down regulation of GDF15 early in tumorigenesis. Later in tumorigenesis, TXNIP upregulation is less pronounced, elevating GFP15 resulting in a blockage of tumor suppressive immune responses. Generally the work is convincing. For example, it's clear that TXNIP is up regulated by Oxa in an ROS and MondoA-dependent manner. Likewise its quite clear TXNIP loss reads to an upregulation of GDF15. However, it's also quite clear that Oxa suppresses GDF15 in a manner that appears to be completely independent of TXNIP. The writing in the paper implies strongly that there is a mechanistic connection between TXNIP and GDF15, but no experiments investigate this possibility.

We feel this is very fair and is reflective of a) perhaps an overemphasis of the TXNIP knockout observation and supportive tissue data, which suggests a relationship but not a mechanistic understanding b) an underemphasis of the data in Figure 3 that shows a decrease in GDF15 after oxaliplatin treatment in TXNIP knockout lines.

We have addressed these concerns in several ways:

1. We have carried out knockdown experiments using siRNA for ARRDC4, which we felt, given its regulation by MondoA and ROS, and homology to TXNIP, may also regulate GDF15. This was found to be the case and may explain the data in Figure 3. At the very least it shows that other factors involved in oxidative stress management may have similar impacts – a form of functional redundancy. Lines 553-559 “Finally, given our previous data (Figure S4) we looked to assess the role of ARRDC4 on GDF15 expression. In the absence of oxaliplatin, knocking down ARRDC4 in DLD1 and HCT15 cells drove an increase in GDF15. When challenged with oxaliplatin, both ARRDC4 and TXNIP expression increased and GDF15 decreased. When the ARRDC4 knockdown was challenged TXNIP increased further and GDF15 decreased further (Figure S6G-J). Given the common regulatory pathways and homology between TXNIP and ARRDC4, and their similar functional roles, we suggest these data are evidence of redundancy within this system. “

We have included some context in the discussion:

Lines 930-933: “Further support for both TXNIP and ARRDC4’s role in regulating GDF15 after the induction of ROS comes from a pan cancer meta-analysis assessing the impact of metformin (which has been reported to inhibit ROS) on gene expression. Here the top two downregulated genes were TXNIP and ARRDC4 and the top four upregulated genes were DDIT4, CHD2, ERN1 and GDF1572

We have tempered the text:

Lines 522-524 “It is important to note however that here we saw clear evidence that TXNIP was not solely responsible for the downregulation of GDF15 post oxaliplatin treatment, with decreased levels seen in knockout lines (Figure 3C-G, S5E).”

Lines 926-929 “It must be stressed that these data do not place TXNIP as the sole regulator of GDF15, for example ARRDC4 can also be seen to regulate GDF15. We envisage TXNIP as one of a number of ROS-dependent GDF15 regulators, with this redundancy potential evidence of the importance of this regulatory framework.”

We have carried out additional analysis detailed in the discussion and in Figure S12 which suggests TXNIP impacts MYC function, as reported elsewhere (detailed below). For ease, the key paper can be accessed through this link https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3001778

Lines 934-956: “The main shortcoming of this paper is the lack of mechanistic understanding linking TXNIP to GDF15. There are 650 transcription factors that have been shown, or are predicted, to bind to GDF15 promoter and/or enhancer regions. By assessing our list of differentially expressed genes (Suppl. Table 1-2) for the presence of these factors we identified 6 GDF15 binding TFs that show significantly decreased expression after oxaliplatin treatment in both cell lines (ATF4, MYC, SREBF1, PHB2, HBP1, KLF9). There was only one, MYC, that was downregulated by oxaliplatin treatment (validated; Figure S12A), and with this downregulation partially being rescued in a matched TXNIP knockout line (Figure S12B). We then observed that c-myc has been shown or is predicted to bind to promoter/enhancer regions of the top five transcriptomic and proteomic differentials in TXNIP knockout lines, including TXNIP itself (apart from C16orf90). Even with c-myc’s promiscuity (binds to 10-20% of all promoters/enhancers) this may be suggestive of a specific relationship. Finally, when looking at the correlations between these 6 TFs and TXNIP and GDF15 in the TCGA COAD dataset, MYC has the greatest and most significant negative correlation to TXNIP (r=-0.4631 p=1.42e-28) and the greatest and most significant positive correlation to GDF15 (r=0.4653 p=7.32e-29). ATF4 and PHB2 are the other TFs in the list, that show the same significant trends (Figure S12C), and therefore may play a role in the TXNIP-independent oxaliplatin-dependent regulation of GDF15. Further exploration of these additional TFs is outside the scope of the current manuscript.

MYC’s role in bridging from TXNIP to GDF15 is further supported by a recent paper which shows that TXNIP is “a broad repressor of MYC genomic binding” and that “TXNIP loss mimics MYC overexpression”73. Furthermore, the inter-dependent regulatory relationship between MondoA, TXNIP, and MYC has been seen in a variety of models74, whilst the impact of NAC on MYC-dependent pathways has been seen in lymphoma75. These studies lend credence to the idea that MYC is the most likely TXNIP-regulated TF that regulates GDF15 in our systems.”

It seems equally likely that TXNIP and GDF15 represent independent parallel pathways. Even if TXNIP is a direct regulator of GDF15, it's also clear that other "factors" up or down-regulated by Oxa also contribute to the regulation of GDF15. These are not explored and even though TXNIP is highly regulated genes shown Figure 2 that are not identified or discussed that may also be contributing to GDF15 regulation.

As mentioned above, the new data suggests that at least one other factor, ARRDC4, can regulate GDF15 (changes upon oxaliplatin treatment) and that MYC is a potential mechanistic bridge between TXNIP and GDF15. Whilst assessing for the transcription factor that may link TXNIP and GDF15 we found an additional 5 TXNIP-independent factors (ATF4, PHB2, SREBF1, HBP1, KLF9) that bind to GDF15 promoter/enhancer regions and are downregulated post-oxaliplatin treatment. When looking at correlations between these factors and GDF15 in the TCGA COAD dataset, ATF4 and PHB2 correlate most closely with GDF15 (when removing MYC) and so we would cautiously suggest that these may be the most pertinent. This data is now included.

Further, the experiments treating PBMCs with conditioned media contain other cytokines/factors, in addition to GDF15, that likely also contribute the observed effects on the different immune cells understudy. The conditioned media from GDF15 knock out cells are a good experiment, but the media is not rigorously tested to see what other cytokines/factors might have also been depleted.

The TXNIP knockout media is the same as that analysed by mass spec and the protein array, however as the reviewer states there is no analysis (excluding assessing for the presence or absence of GDF15) on the double knockout supernatant or over-expression supernatant. The text has been corrected as follows:

Lines 675-679. “In light of other secreted factors being seen to be regulated by TXNIP (Figure 3A-B), we included double knockouts (TXNIP and GDF15 knockout; GTKO) as well as an overexpression system (GDF15a) to test for GDF15 specific effects. However, we do not know the impact of knocking out or overexpressing GDF15 on the broader secretome.”

Perhaps a GDF15 complementation experiment would help here.

We felt that the association between GDF15 and Treg induction is reasonably well established in the literature, and so once we saw that the supernatant from our GDF15 overexpression system (+/- CD48 blockade) complemented what has already been demonstrated, we were encouraged. However we needed more – hence the TCGA data and IHC staining.

Finally, even if completely independent, a TXNIP/GDF15 ratio does seem to have utility in determining chemo-therapeutic response.

We agree – we feel that conceptually this may be the most interesting part of the project and is an example of what can be done with these tools.

Other major points: 1. Please label the other highly regulated genes shown in Fig 2A and B. Might they also explain some of the underlying biology. This could be on the current figures or in a supplement, though the former is preferred.

Many thanks – we have done this.

Please address why the TXNIP induction is so much less in patient-derived organoids vs. cell line spheroids (Fig S2). By the western blots, TXNIP inductions in the organoids looks quite modest. Further, the text is quite cryptic and implies that the "upregulation" is similar in both organoids and spheroids.

You are absolutely correct. Many apologies, the wording has changed:

Lines 320-323 “In both models we observed the upregulation of TXNIP mRNA (Figure S2E-H) and TXNIP protein (Figure S2I-L) after oxaliplatin treatment, with spheroids showing greater responsiveness. This difference is most likely due to culturing conditions or differences in the number and location of cycling cells.”

We have two possible explanations. Firstly the media in which the organoids are cultured contains a lower glucose concentration than that used for the spheroids. As per some of our new data (Figure S3 – later in the rebuttal), the upregulation of TXNIP after oxaliplatin is glucose dependant, with lower concentrations resulting in less of a differential. Secondly, while restricted to the periphery, the Ki67 signal in DLD1 spheroids is quite pronounced indicating that, within the outer zone, many cells (probably the majority) are in the S/G1/G2 phase of the cell cycle at any given point in time (figure below this text).

This is not the case for the organoids, where the Ki67 (and pCDK1) signal is quite weak, and only sporadic in the outer layer. So we believe that there are many more rapidly cycling cells in the most drug-exposed layer of spheroids when compared to the comparable region in organoids. As the spheroid cells are likely cycling more rapidly, they would also be expected to be more adversely affected by the drug within the finite drug treatment window. Indeed, these spheroids grow large, and quite quickly. If the organoid cells are cycling more slowly and if, within the cell layer most exposed to drug, these cycling cells are less abundant, then the TXNIP response may well be subdued in organoids when compared with spheroids.

We have decided to not include the above (full) explanation and figure within the new draft, as we feel it may distract from the central message. However do let ourselves and the editor know if you disagree.

What was the rationale of performing the MS experiment on control and TXNIP KO DLD1 cells in the absence of oxaliplatin? The other experiments in Fig 3 clearly show that Oxa can repress GDF15 even in the absence of TXNIP, which implicates other pathways. ARRDC4? Or something else? This needs to be addressed.

We adopted this approach because of the order in which the assays occurred and technical issues surrounding the use of post-oxaliplatin treated supernatant. By the time we moved to the proteomics we had already identified, and validated, GDF15 as our number one candidate (initially from the protein array), in terms of response to oxaliplatin and dependence on TXNIP. This led us to the next stage of the project – to assess the environmental impacts of this factor in vitro before validation in situ. To do this, aware of the issue of contaminated recombinant GDF15, we decided early on to use cell line supernatant. We carried out some pilot studies on immune cells using supernatant from oxaliplatin treated cell lines and we had several technical issues (difficulty in determining the correct controls, immune cell death…). This changed the emphasis to using supernatant from knockout models rather than knockout and treated models. Before we began these assays in earnest we wanted to assess exactly what was enriched in TXNIP knockout supernatant and so we turned to proteomics. When this further validated GDF15, we then generated GDF15 and TXNIP/GDF15 knockouts to further elucidate GDF15’s role specifically.

With regards the other pathways, as you correctly predicted, ARRDC4 also appears to regulate GDF15 – many thanks for helping with this line of enquiry. Please see earlier in the rebuttal for more details and the data.

The data in 3J with the MondoA knockdown is not convincing. The knockdown is weak and TXNIP goes down a smidge. Agree that GDF15 goes up

We agree. We have re-run this and pooled the densitometry data – see new figure below (Panel 3J).

Minor points 1. Line 79. The "loss" of TXNIP/GDF15 axis is confusing. It's really loss of TXNIP and upregulation of GDF15, right?

Absolutely - corrected to responsiveness.

Lines 144-147: “Intriguingly, multiple models including patient-derived tumor organoids demonstrate that the loss of TXNIP and GDF15 responsiveness to oxaliplatin is associated with advanced disease or chemotherapeutic resistance, with transcriptomic or proteomic GDF15/TXNIP ratios showing potential as a prognostic biomarker.”

Please provide an explanation for the different stages in tables 1 and 2. This will likely not be clear to non-clinicians.

Many thanks. The following has been added at the bottom of the second table.

Lines 304-309: “The TNM staging system stands for Tumor, Node, Metastasis. T describes the size of the primary tumor (T1-2; 5cm). N describes the presence of tumor cells in the lymph nodes (N0; no lymph nodes. N1-3 >0). M describes whether there are any observable metastases (M0; no metastases. M1; metastases). The clinical stage system is as follows: I/II; the tumor has remained stable or grown, but hasn’t spread. III/IV; the tumor has spread, either locally (III) or systemically (IV).”

Line 231 should probably read ...cysteine (NAC), a reactive oxygen species inhibitor,

Many thanks - corrected

Line 247, should be RT-qPCR I think.

Many thanks - corrected

Lines 343-345. I don't quite understand the wording. Does this mean to say that 675 soluble proteins were not changed between the condition media from both cell populations?

Yes, exactly this. We have removed as this is superfluous and confusing.

The data in FigS1 B and C don't seem to reach the standard p value of > 0.05

Very true – we have rewritten the text to make sure the reader knows there is no significance.

Lines 269-271. “High levels of both the protein (significantly) and the transcript (not significantly) were seen to be associated with favourable prognosis (Figure 1G,H and S1B,C).”

**Referee Cross-Commenting**

cross comment regarding referees 2 and 3 above. I'm am convinced that TXNIP is at least contemporaneously upregulated with GDF15 downregulation. However, the strong implication from the writing is that TXNIP regulates GDF15 directly. I agree with the comment above that exploring mechanisms may be open-ended especially as TXNIP has been implicated in gene regulation by several different mechanism. I'd be satisfied with a more open-minded discussion of potential mechanisms by which TXNIP may repress GDF15 and the possibility of other parallel pathways that likely contribute to GDF15 repression.

Many thanks, this is a generous and understanding approach. As described above we have carried out extra analysis and have found 6 differentially regulated transcription factors which have been shown to bind GDF15 promoter or enhancer regions with 1 of these, MYC, being significantly affected in the TXNIP knockout cell lines, which in combination with supportive literature suggests a degree of TXNIP dependence. We have also identified ARRDC4 as an additional regulator of GDF15 – again please see above.

Reviewer #1 (Significance (Required)):

This is an interesting contribution but the mechanistic connection between GDF15 and TXNIP is relatively weak. That said, even as independent variables they do seem to have utility in predicting therapeutic response.

Many thanks for the comment – we concur. We have reanalysed our data looking for relevant transcription factors (those that bind GDF15 promoter / enhancer regions) finding MYC as the most likely bridge. Please see above.

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

The manuscript by Deng et al. investigates a mechanistic link between TXNIP and GDF15 expression and oxaliplatin treatment and acquired resistance. They observe an upregulation in TXNIP expression in the tumors of patients who have previously received chemotherapy. They demonstrate oxaliplatin-driven MondoA transcriptional activity is what underlies the induction of TXNIP. They further demonstrate that TXNIP is a negative regulator of GDF15 expression. Together, oxaliplatin induces MondoA activity and TXNIP expression, resulting in a downregulation of GDF15 expression and consequently decreased Treg differentiation.

Major Comments

1. The authors suggest that TXNIP induction and GDF15 downregulation are a common effect of chemotherapies; however, the mechanistic studies were limited to oxaliplatin. The authors should clarify this point through further investigation using other commonly used CRC chemotherapies (5-FU, irinotecan, etc.),or through textual changes. To be clear, I think that the oxaliplatin results could potentially stand on their own but would require additional clarification. For example, regarding the patient samples analyzed in 1D and 4F, which patients received oxaliplatin? Could the analysis of publicly available molecular data be drilled down to just the patients who received oxaliplatin?

Many thanks – this is an excellent point. Firstly, all the patients in 1D and 4F received oxaliplatin. Secondly, we have included new data looking at the impact of other chemotherapies (FOLRIRI, FU-5 and SN-38) on aspects of the study, ultimately finding that these processes (especially an anti-correlation between GDF15 and TXNIP changes upon chemo treatment) appear to be specific to oxaliplatin. These data have been added (Figure S11) and throughout the emphasis has been switched from chemotherapeutic treatment to oxaliplatin treatment.

Lines 796-799: “To check if the pre-treatment GDF15/TXNIP ratio could be used for patients treated with FOLFIRI we performed the same analyses finding no significance (S11A-D). This oxaliplatin specificity was then confirmed by western blot analysis in DLD1 and HCT15 cells treated with 5-FU or SN38 (Figure S11E-F).

Example of change of emphasis from ‘chemotherapy’ to ‘oxaliplatin’ – lines 139-142: “Here, in colorectal adenocarcinoma (CRC) we identify oxaliplatin-induced Thioredoxin Interacting Protein (TXNIP), a MondoA-dependent tumor suppressor gene, as a negative regulator of Growth/Differentiation Factor 15 (GDF15).”

The data demonstrating the induction of MondoA transcriptional activity and TXNIP expression in response to oxaliplatin treatment is quite convincing. The data regarding ROS induction of TXNIP is interesting, especially in light of other studies arguing that ROS limits MondoA activity (PMID: 25332233). Given this apparent disparity, I think that this study could really be strengthened by also investigating other potential mechanisms of oxaliplatin induction of MondoA. In particular, given many studies arguing for direct nutrient-regulation of MondoA, the authors should address the potential for oxaliplatin regulation of glucose availability and a potential glucose dependence of oxaliplatin-induced TXNIP. 2

In line with the previous point, since MondoA activity and TXNIP expression are sensitive to glucose levels, the authors should investigate oxaliplatin-regulation of TXNIP under physiological glucose levels. No need to replicate everything, just key experiments.

We feel these are excellent point and really help the piece – many thanks. We have carried out assays around these points suggested and have included the findings in the new draft – see below.

Lines 332-339: “As such, we went back to first principles and assessed the impact of different concentrations of glucose on TXNIP induction +/- oxaliplatin treatment, finding a concentration dependent effect (Figure S3A). Intriguingly, high glucose alone was able to induce increased TXNIP expression. We then assessed if oxaliplatin treatment drove an increase in glucose uptake, with this seen at concentrations >10mM (Figure S3B). Next, to investigate the impact of glucose metabolism, and consequent ROS generation, on TXNIP induction we treated cells with Antimycin A, an inhibitor of oxidative phosphorylation, finding a complete block in oxaliplatin-induced TXNIP (Figure S3C).”

The authors did a good job of linking TXNIP and GDF15 in untreated conditions; however, the data arguing for oxaliplatin regulation of GDF15 through TXNIP is less clear. For example, in 3B-H, oxaliplatin treatment reduces GDF15 approximately to the same extent in the NTC and TKO cells, potentially in line with a mechanism of downregulation that doesn't involve TXNIP.

A very salient point and completely in line with the other reviewers. We have carried out a few additional analyses mentioned previously in this letter. The most pertinent for this specific point are the experiments around ARRDC4, where we found evidence to suggest that, like TXNIP, it regulates GDF15.

Minor Comments

1. The presentation of data in Figure 5 is confusing. A-B include raw cell numbers, whereas C-F show "normalized proliferation." What does this mean? And how was the normalization done?

Apologies for this. Legend test has been corrected to “Normalised proliferation (normalised to MFI from control: i.e. cells treated with supernatant from NTC cells) on gated CD3+CD8+ or CD3+CD4+ cells is shown. n=6. (G-H) Normalised IFNg concentrations (normalised to MFI from control: i.e. cells treated with supernatant from NTC cells) in the supernatant of cells from C-F.” (lines 727-729).

**Referee Cross-Commenting**

cross-comment regarding reviewer #1

I agree with the referee that the link between TXNIP and GDF15 is weak, though as I mentioned before, this is particularly true in the context of oxaliplatin-regulation of TXNIP. I agree that given all the presented data, it is likely that oxaliplatin-regulation of TXNIP and GDF15 are independent. In my opinion, the referee brought up all valid concerns, but this is by far the biggest concern that I share.

We agree that this is the weakest aspect of the paper, however our new analyses plus supportive literature, suggests that the relationship between TXNIP and GDF15 may be mediated by MYC (please see above)

cross-comment regarding reviewer #3

The major concern that this referee addresses is whether another transcription factor supersedes the proposed MondoA/TXNIP induction in regulating GDF15 expression in later stage CRC. In my opinion, this another other concerns of the referee are all valid, but still I remain unconvinced that TXNIP induction underlies the oxaliplatin-regulation of GDF15. I think fleshing out that aspect of the study would potentially help the authors tease apart how this potential MondoA-TXNIP-GDF15 axis is dysregulated later in CRC progression.

This is a great discussion. Interestingly enough, c-myc is seen at higher levels in late stage CRC (Hu X, Fatima S, Chen M, Huang T, Chen YW, Gong R, Wong HLX, Yu R, Song L, Kwan HY, Bian Z. Dihydroartemisinin is potential therapeutics for treating late-stage CRC by targeting the elevated c-Myc level. Cell Death Dis. 2021 Nov 5;12(11):1053. Doi: 10.1038/s41419-021-04247-w. PMID: 34741022; PMCID: PMC8571272.), is seen as an important factor in resistance, and as this review argues, is driven by stress (Saeed H, Leibowitz BJ, Zhang L, Yu J. Targeting Myc-driven stress addiction in colorectal cancer. Drug Resist Updat. 2023 Jul;69:100963. Doi: 10.1016/j.drup.2023.100963. Epub 2023 Apr 20. PMID: 37119690; PMCID: PMC10330748.). So it is very plausible that the partial TXNIP-mediated regulation of myc in early / sensitive CRCs that we may be observing, and has been reported recently (TXNIP loss expands Myc-dependent transcriptional programs by increasing Myc genomic binding Lim TY, Wilde BR, Thomas ML, Murphy KE, Vahrenkamp JM, et al. (2023) TXNIP loss expands Myc-dependent transcriptional programs by increasing Myc genomic binding. PLOS Biology 21(3): e3001778. https://doi.org/10.1371/journal.pbio.3001778) is lost in late stage / resistant CRCs. If this is the case, in effect what we would have observed is the loss of a stress-associated method (TXNIP) of controlling c-myc activity. What makes our collective lives difficult is that, as reported “this expansion of Myc-dependent transcription following TXNIP loss occurs without an apparent increase in Myc’s intrinsic capacity to activate transcription and without increasing Myc levels.” (TXNIP loss expands Myc-dependent transcriptional programs by increasing Myc genomic binding Lim TY, Wilde BR, Thomas ML, Murphy KE, Vahrenkamp JM, et al. (2023) TXNIP loss expands Myc-dependent transcriptional programs by increasing Myc genomic binding. PLOS Biology 21(3): e3001778. https://doi.org/10.1371/journal.pbio.3001778)

Reviewer #2 (Significance (Required)):

Generally speaking the experiments are well controlled and the findings are significant and novel. Though the link between MondoA activity and ROS could be strengthened, and the data could be validated under more physiological settings. Further, the authors should clarify their interpretations so as to not overstate the findings.

Many thanks for the comments. We have taken onboard the need for more physiological settings and have included varying levels of glucose to reflect concentrations in different environments. We have repeated the siMondoA work in 3J strengthening the conclusions wrt its impact on TXNIP and GDF15 expression (see above).

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

In this well-written manuscript, the authors show that chemotherapy increases a MondoA-dependent oxidative stress-associated protein, TXNIP, in chemotherapy-responsive colorectal cancer cells. They show that TXNIP negatively regulates GDF-15 expression. GDF-15, in turn, correlates with the presence of T cells (Treg), and inhibits CD4 and CD8 T cell stimulation. In advanced disease and chemo-resistant cancers, upregulation of TXNIP and downregulation of GDF-15 appear to get lost. Based on a somewhat smallish data set, the authors suggest that the pre-treatment GDF-15/TXNIP ratio can predict responses to oxaliplatin treatment. This is a very interesting, novel finding. In general, the quality of the experiments and the data are high and the conclusions appear sound. Still, there are a number of aspects that should still be improved:

The observed loss of the ROS - MondoA - TXNIP - GDF15 axis in chemoresistant and/or metastatic tumors implies that another transcription factor or pathway becomes dominant upon tumor progression. As this switch would be key to better understanding the mechanism underlying the prognostic role of the TXNIP/GDF15 ratio, the authors should at least do data mining followed by ChEA or Encode (or other) analysis to identify transcription factors or pathways that become activated in late-stage/metastatic CRC cells. There is a high likelihood that a transcription factor or pathway involved in GDF-15 upregulation in cancer (e.g. p53, HIF1alpha, Nrf2, NF-kB, MITF, C/EBPß, BRAF, PI3K/AKT, MAPK p38, EGR1) supersedes the inhibitory effect of the MondoA-TXNIP axis. As it stands, the proposed loss of function of the ROS - MondoA - TXNIP - GDF-15 axis is far less convincing than almost all other aspects of the study.

An extremely fair point. We adopted a similar approach to that suggested – as mentioned above, we looked at TFs that bind to GDF15 promoter/enhancer regions and then looked at the presence of these in our transcriptomic data – specifically any evidence of change post oxaliplatin treatment. We found 6 such TFs that were decreased post-oxaliplatin treatment. We then looked for any evidence of TXNIP dependence in these TFs by comparing post-oxaliplatin treatment across NTC and TXNIP knockout lines, when we did this we found only one GDF15 promoter/enhancer binding TF was significantly changed: MYC. We then looked at the relationship between MYC,TXNIP, and GDF15 against the other 5 ‘control’ TFs in the TCGA COAD dataset, we found that MYC showed the strongest correlations, in the ‘correct’ directions. This finding was further backed up in the literature where a TXNIP knockout in a breast cancer model drove c-myc-dependent transcription, whilst c-myc has been observed to increase in later stage CRC patients, is associated with cellular stress and resistance. The collective evidence therefore suggests that MYC is the factor that is initially at least partially regulated by TXNIP, before this regulation is lost in advanced / resistant disease. Continuing on this line, it is likely that the predictive GDF15/TXNIP ratio is at least in part, a measure of c-myc responsiveness to oxaliplatin. All the while we must bear in mind TXNIP-independent oxaliplatin-dependent regulation of GDF15, most likely ARRDC4, as described earlier in this document.

Using pathway analysis software to compare our transcriptomic data from cell lines treated with/without oxaliplatin, the most likely pathways upstream of MYC/c-myc that are negatively affected by chemotherapy are BAG2, Endothelin-1, telomerase, ErbB2-ErbB3 and Wnt/B-catenin. When looking at the comparison of UTC and resistant lines’ transcripts there is only one key component of these pathways which is upregulated in both lines - ERBB3 – which has already been shown to be important in CRC metastasis and resistance (Desai O, Wang R. HER3- A key survival pathway and an emerging therapeutic target in metastatic colorectal cancer and pancreatic ductal adenocarcinoma. Oncotarget. 2023 May 10;14:439-443. doi: 10.18632/oncotarget.28421. PMID: 37163206; PMCID: PMC10171365.). It is highly speculative, but our data suggests the most likely pathway to supersede TXNIP in its (partial) regulation of MYC is the ErbB2-ErbB3 pathway.

My further criticisms are mostly more technical:

Figure 2 I-L: What was the extent of MondoA downregulation achieved by siRNA treatment? Could the effects also be seen with the small molecule mondoA inhibitor SBI-477 (or a related substance)?

This experiment has been repeated. The pooled densiometric data is also now given (please see above).

How do you explain the different GDF-15 levels between untreated non-target control cells (NTC) and TXNIP knock-down cells (TKO) in Figures 3C-F?

The only way to interpret this is that there is a TXNIP-independent pathway regulating GDF15 expression after oxaliplatin treatment, as described this is most likely to be ARRDC4 - the text has been updated to:

Lines 522-524: “It is important to note, however, that we saw clear evidence that TXNIP was not solely responsible for the downregulation of GDF15 post oxaliplatin treatment (Figure 3C-G, S6E).”

In figures 3 E-G the dots for the individual measurements should be indicated. This would be more informative than just the bar graphs.

Completed.

Figure 4C,D and Table 3: Data on the role of GDF-15 in CRC are largely valedictory of previous work (e.g. Brown et al. Clin Cancer Res 2003, 9(7):2642-2650, Wallin et al., Br J Cancer. 2011 May, 10;104(10):1619-27). Therefore, the previous studies should be cited.

Apologies for the oversight and many thanks – this is an excellent addition.

Figure 5C-F: Please indicate in the figure legend how proliferation was assessed.

Many thanks. This was noticed by another reviewer also. We have changed the text to include how the data was normalised: “(C-F) Labelled PBMCs were stimulated with anti-CD3 and anti-CD28 for 4 days in the presence of fresh supernatant from indicated cell lines, before being stained with anti-CD3 and anti-CD8 (C-D) or anti-CD4 (E-F) antibodies and measured by flow cytometry. Normalised proliferation (normalised to MFI from control: i.e. cells treated with supernatant from NTC cells) on gated CD3+CD8+ or CD3+CD4+ cells is shown. n=6. (G-H) Normalised IFNg concentrations (normalised to MFI from control: i.e. cells treated with supernatant from NTC cells) in the supernatant of cells from C-F.” (lines 724-730)

Figure S8E-G: Please indicate the analysed parameters in the graphs. In Figure S8G, the legend just indicates that "aggression of tumour" is dichotomized and plotted. This clearly requires a better definition.

Many thanks, this has been changed as per the below.

Lines 862-868: “(E-G) Receiver operating characteristic (ROC) curves showing area under the curve and p values for the use of GDF15/TXNIP ratio in predicting origin of cell line (E; primary; DLD1, HCT15, HT29, SW48 [n=4] or secondary; DiFi, LIM1215 [n=2]), sensitivity to oxaliplatin (F; parental DLD1 (plus biological repeat), HCT15 [n=3] or resistant DLD1 (plus biological repeat), HCT15 [n=3]), aggression of tumor (G; non-aggressive; The authors propose a novel ROS - MondoA - TXNIP - GDF15 - Treg axis, where MondoA activation, TXNIP up- and GDF-15 downregulation enhance tumor immunogenicity. While this axis has been analyzed in some detail, GDF-15 is not only linked to induction of regulatory T cells. There has been a report showing that GDF-15/MIC-1 expression in colorectal cancer correlates with the absence of immune cell infiltration (Brown et al. Clin Cancer Res 2003, 9(7):2642-2650). The link between GDF-15 and immune cell exclusion has also been confirmed in other conditions, including different cancers (Kempf et al. Nat Med 2011, 17(5):581-588, Roth P et al. Clin Cancer Res 2010, 16(15):3851-3859, Haake et al. Nat Commun 2023, 14(1):4253). A key mechanism is the GDF-15 mediated inhibition of LFA-1 activation on immune cells. As the authors argue that the described pathways turns cold tumors hot in response to oxaliplatin-based chemotherapy, this GDF-15 dependent immune cell exclusion mechanism might be at least as relevant than induction of Treg. Likewise, inhibition of dendritic cell maturation by GDF-15 (Zhou et al. PLoS One 2013, 8(11):e78618) could explain why GDF-15high tumors are immunologically cold. Reviewed in 3

The authors propose that the pathways discovered by them contributed to the "heating up" of the tumor microenvironment after oxaliplatin-based chemotherapy. The authors should thus look in their data sets for the presence of cytotoxic T cells and their possible correlation with TXNIP and GDF-15 levels.

This is a wonderful explanation – many thanks. We have taken the opportunity to assess the impact of GDF15 expression on a variety of T cell markers (Figure S9). In this data a negative association between GDF15 and CD8 CTLs can clearly be seen, as predicted by the reviewer.

Lines 712-717: “To assess if the GDF15-dependent presence of Tregs may be associated with a decrease in activated cytotoxic CD8 T cells, we interrogated the TCGA COAD dataset. We found that low GDF15 tumors carried significantly higher levels of CD8, CD69, IL2RA, CD28, PRF1, GZMA, GZMK, TBX21, EOMES and IRF4 (Figure S9); transcripts indicative of activated cytotoxic CD8 T cells. High GDF15 tumors were enrichment for FOXP3 and, interestingly, RORC (Figure S9). These data support the hypothesis that GDF15 induces Foxp3+ve Tregs which inhibit CD8 T cell proliferation and activation in the TME.”

The paragraph on GDF-15 receptors needs to be corrected: The purported role of a type 2 transforming growth factor (TGF)-beta receptor in GDF-15 signalling had been due to a frequent contamination of recombinant GDF-15 with TGF-beta (Olsen et al. PLoS One 2017, 12(11):e0187349). There have been a number of screenings for GDF-15 receptors that have all failed to show an interaction between GDF-15 and TGF-beta receptors. Instead, only GFRAL was found in these large-scale screenings (Emmerson et al. Nat Med 2017, 23(10):1215-1219, Hsu et al. Nature 2017, 550(7675):255-259, Mullican et al. Nat Med 2017, 23(10):1150-1157, Yang et al. Nat Med 2017, 23(10):1158-1166). The one subsequent report that shows a link between GDF-15, engagement of CD48 on T cells and induction of a regulatory phenotype (Wang et al. J Immunother Cancer 2021, 9(9)) still awaits independent validation. Considering that CD48 lacks an intracellular signaling domain that would be critical for a classical receptor function, I recommend to be more cautious regarding the role of CD48 as GDF-15 receptor. Given the mechanism outlined by Wang et al. the word interaction partner might be more apt. Moreover, an anti-GDF-15 antibody would be a good control for the experiments involving an anti-CD48 antibody in Figure 5.

Thank you so much for this concise and highly informative paragraph. We have changed the text to read:

202-204: “As a soluble protein, GDF15 exerts its effects by binding to its cognate receptor, GDNF-family receptor a-like (GFRAL)44,45,46,47 or interaction partner, CD48 receptor (SLAMF2)43, with the latter still requiring additional verification.”

We would have ideally included an anti-GDF15 antibody in the CD48 assay at the time but didn’t have the foresight. We have included the additional text to temper any conclusions.

Lines 701-711: “Furthermore, when stimulating naïve CD4 T cells in the presence of GDF15 enriched supernatant we were able to both differentiate these cells into functional Tregs and also block the generation of this functionality using an anti-CD48 antibody (Figure 5M-N). However, it must be stressed that the binding and functional impacts of GDF15’s interaction with CD48 still require further verification.”

Cell surface externalization of annexin A1 has been described as a failsafe mechanism to prevent inflammatory responses during secondary necrosis (PMID: 20007579). Thus, I am surprised that the authors list annexin A1 among the immune-stimulatory molecules exposed or released in response to chemotherapy-induced cell death (line 103). Please clarify!

We agree – it shouldn’t be there!! Removed. Many thanks.

**Referee Cross-Commenting**

Regarding the cross-comment by referee 2: In my opinion, the data shown in Figure 3C-H clearly demonstrates that TXNIP can repress GDF-15 expression. I agree that there will likely be further regulators. The GDF-15 promoter is constantly regulated by a multitude of factors (which mostly induce transcription). As downregulation of GDF-15 in response to oxaliplatin is the opposite of the frequently described induction of GDF-15 upon chemotherapy, net effects may always be "smudged" by contributions from different pathways (e.g. by cell stress due to siRNA transfection). Therefore, I believe that the data are as good as it will get. Accordingly, I would not force the authors to further amplify the observed effect.

Many thanks for your understanding – yes, GDF15 has >650 TFs that bind its promoter/enhancer regions – a number we found rather daunting. Happily your comments and those of the other reviewers inspired us to dig and we now have data that is supportive of MYC’s and ARRDC4’s involvement – detailed throughout this reply.

cross comment regarding referee #1: I share the general assessment of the referee and recognize the very detailed mechanistic analysis. To further support the moderate effects of the MondoA knockdown, a small molecule inhibitor like SBI-477 might be useful. (I had already suggested using this inhibitor to support these data.)

Many thanks for the suggestion. We opted to increase the number of siRNA repeats instead – with the data included in Figure 3J (above).

Still, my view on the potential relevance of oxaliplatin-induced, TXNIP-independent downregulation of GDF-15 differs from that of referee 1. In the clinics, platinum-based chemotherapy is one of the strongest inducers of GDF-15 (compare Breen et al. GDF-15 Neutralization Alleviates Platinum-Based Chemotherapy-Induced Emesis, Anorexia, and Weight Loss in Mice and Nonhuman Primates. Cell Metabolism 32(6), P938-950, 2020.DOI:https://doi.org/10.1016/j.cmet.2020.10.023). I was thus surprised that the authors found a pathway, which leads to an outcome that an exactly opposite effect.

This is fascinating that oxaliplatin drives this increase in GDF15 – we were unaware of this paper. Looking at figure 2(H-K), GDF15 is being produced from multiple non-diseased tissues after systemic chemotherapy – even at day 19 post-treatment – this suggests that wrt this study, systemic GDF15 could not be used as a readout of success or otherwise – which is extremely helpful! Thank you.

Thus far, the only obvious reason for reduced GDF-15 secretion upon treatment with cytotoxic drugs was a reduction in tumor cell number due to cytotoxicity.

Please do not discount this. This study was focused on the cells which survived oxaliplatin treatment – the cells which did not were discarded. Our view, given your input, would be a complex picture where in early stages systemic GDF15 goes up, due to off-target effects, but locally levels drop owing to cell death and this, and other, stress-related pathways in the remaining tumor cells.

Still, the authors managed to convince me that the described pathway (ROS - MondoA - TXNIP - GDF-15) exists. (Here, I still largely concur with referee 1.) Moreover, as we have identified some factors required for GDF-15 biosynthesis that could easily interact with TXNIP, I find the proposed mechanism plausible.

Extremely encouraging for us to hear!

Nevertheless, as a downregulation of GDF-15 in response to chemotherapy is hardly ever observed in late-stage cancers, I believe that the observed switch in pathway activation between early- and late-stage cancers might be highly relevant - in particular, as there is so much evidence for platinum-based induction of GDF-15 in late-stage cancer patients. Emphasizing the divergent clinical observations (e.g. by Breen et al.) could thus help to put the finding into perspective.

Very much agree. We did see this phenomenon in LIM1215 cells (Figure 6B) and the resistant lines we generated continually produced higher levels.

Analysing TXNIP-independent mechanisms involved in the oxaliplatin-dependent repression of GDF-15, as suggested by referee #1, will require enormous efforts and resources, and may still turn out to be fruitless. Personally, I would thus be content if the authors just mentioned possible contributions from other pathways upon cancer progression. To me, the described pathway seems to be limited to early-stage cancers, and the actual finding that GDF-15 is downregulated is an interesting observation, irrespective of further involved pathways.

Many thanks – this is extremely fair. Happily we have managed to make some tentative steps forward in highlighting the potential role of MYC, and the suggestion of redundancy wrt ARRDC4, but as you say, much more work needs to be done to fully understand these processes.

cross comment regarding referee #2: I fully agree with the referee that activation of the pathway by further chemotherapeutic drugs could be a valuable addition. As Guido Kroemer´s lab has described oxaliplatin to induce a more immunogenic cell death compared to other platinum-based chemotherapies, even a rather limited comparison between oxaliplatin and cisplatin could be very interesting.

Absolutely agree – extra data on this has been included in Figure S11, which is included earlier in this letter. We also uncovered a meta-analysis using metformin, which has been seen to inhibit ROS, where TXNIP and ARRDC4 are the top two downregulated transcripts whilst GDF15 appears in the top four upregulated. This may suggest that chemotherapeutic immunogenicity, at least through the presence or absence of GDF15, may in part be driven by ROS.

Lines 930-933: “Further support for both TXNIP and ARRDC4’s role in regulating GDF15 after the induction of ROS comes from a pan cancer meta-analysis assessing the impact of metformin (which has been reported to inhibit ROS) on gene expression. Here the top two downregulated genes were TXNIP and ARRDC4 and the top four upregulated genes were DDIT4, CHD2, ERN1 and GDF1572 “

Reviewer #3 (Significance (Required)):

In general, this is a very interesting manuscript describing a cascade of events that may contribute to successful chemotherapy (which likely requires induction of an immune response against dying tumor cells.) The observation that this pathway is only active in early/non-metastatic cancer cells is striking. Unfortunately, the authors cannot explain inactivation of this pathway in later stage/ metastatic/ highly aggressive cancers. Understanding this switch could easily be the most important finding triggered by this report. Therefore, I highly recommend to make some effort in this direction. Strikingly, the authors find that disruption of TXNIP-mediated GDF-15 downregulation is strongly associated with worse prognosis. They also suggest that this ratio could indicate whether a patient will respond to oxaliplatin-based chemotherapy.

This is again very fair – we have posited a potential mechanism for the loss of this switch elsewhere in this reply– one which involves a change in TXNIP-mediated MYC regulation and/or increased HER2-HER3 signalling – but although reasonable for a rebuttal (and publication in that context) we do not feel we have the evidence to include this within the full manuscript.

Altogether, the findings described in manuscript are very novel and may have prognostic (or, in case of the presumed loss of the MondoA - TXNIP - GDF-15 pathway) therapeutic implications. Thus, the manuscript certainly fills various gaps and should be of major interest for cell biologists working on immunogenic cell death, or colorectal cancer, or MondoA, TXNIP or GDF-15. Still, due to its translational implications, it would also be worthwhile reading for a large number of researchers in the oncology field.

We are very grateful for your kind comments.

1 Sinclair, L. V., Barthelemy, C. & Cantrell, D. A. Single Cell Glucose Uptake Assays: A Cautionary Tale. Immunometabolism 2, e200029, doi:10.20900/immunometab20200029 (2020).

2 Yu, F. X., Chai, T. F., He, H., Hagen, T. & Luo, Y. Thioredoxin-interacting protein (Txnip) gene expression: sensing oxidative phosphorylation status and glycolytic rate. J Biol Chem 285, 25822-25830, doi:10.1074/jbc.M110.108290 (2010).

3 Wischhusen, J., Melero, I. & Fridman, W. H. Growth/Differentiation Factor-15 (GDF-15): From Biomarker to Novel Targetable Immune Checkpoint. Front Immunol 11, 951, doi:10.3389/fimmu.2020.00951 (2020).

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 well-written manuscript, the authors show that chemotherapy increases a MondoA-dependent oxidative stress-associated protein, TXNIP, in chemotherapy-responsive colorectal cancer cells. They show that TXNIP negatively regulates GDF-15 expression. GDF-15, in turn, correlates with the presence of T cells (Treg), and inhibits CD4 and CD8 T cell stimulation. In advanced disease and chemo-resistant cancers, upregulation of TXNIP and downregulation of GDF-15 appear to get lost. Based on a somewhat smallish data set, the authors suggest that the pre-treatment GDF-15/TXNIP ratio can predict responses to oxaliplatin treatment. This is a very interesting, novel finding. In general, the quality of the experiments and the data are high and the conclusions appear sound. Still, there are a number of aspects that should still be improved:

The observed loss of the ROS - MondoA - TXNIP - GDF15 axis in chemoresistant and/or metastatic tumors implies that another transcription factor or pathway becomes dominant upon tumor progression. As this switch would be key to better understanding the mechanism underlying the prognostic role of the TXNIP/GDF15 ratio, the authors should at least do data mining followed by ChEA or Encode (or other) analysis to identify transcription factors or pathways that become activated in late-stage/metastatic CRC cells. There is a high likelihood that a transcription factor or pathway involved in GDF-15 upregulation in cancer (e.g. p53, HIF1alpha, Nrf2, NF-kB, MITF, C/EBPß, BRAF, PI3K/AKT, MAPK p38, EGR1) supersedes the inhibitory effect of the MondoA-TXNIP axis. As it stands, the proposed loss of function of the ROS - MondoA - TXNIP - GDF-15 axis is far less convincing than almost all other aspects of the study.

My further criticisms are mostly more technical:

Figure 2 I-L: What was the extent of MondoA downregulation achieved by siRNA treatment? Could the effects also be seen with the small molecule mondoA inhibitor SBI-477 (or a related substance)?

How do you explain the different GDF-15 levels between untreated non-target control cells (NTC) and TXNIP knock-down cells (TKO) in Figures 3C-F?

In figures 3 E-G the dots for the individual measurements should be indicated. This would be more informative than just the bar graphs.

Figure 4C,D and Table 3: Data on the role of GDF-15 in CRC are largely valedictory of previous work (e.g. Brown et al. Clin Cancer Res 2003, 9(7):2642-2650, Wallin et al., Br J Cancer. 2011 May, 10;104(10):1619-27). Therefore, the previous studies should be cited.

Figure 5C-F: Please indicate in the figure legend how proliferation was assessed.

Figure S8E-G: Please indicate the analysed parameters in the graphs. In Figure S8G, the legend just indicates that "aggression of tumour" is dichotomized and plotted. This clearly requires a better definition.

The authors propose a novel ROS - MondoA - TXNIP - GDF15 - Treg axis, where MondoA activation, TXNIP up- and GDF-15 downregulation enhance tumor immunogenicity. While this axis has been analyzed in some detail, GDF-15 is not only linked to induction of regulatory T cells. There has been a report showing that GDF-15/MIC-1 expression in colorectal cancer correlates with the absence of immune cell infiltration (Brown et al. Clin Cancer Res 2003, 9(7):2642-2650). The link between GDF-15 and immune cell exclusion has also been confirmed in other conditions, including different cancers (Kempf et al. Nat Med 2011, 17(5):581-588, Roth P et al. Clin Cancer Res 2010, 16(15):3851-3859, Haake et al. Nat Commun 2023, 14(1):4253). A key mechanism is the GDF-15 mediated inhibition of LFA-1 activation on immune cells. As the authors argue that the described pathways turns cold tumors hot in response to oxaliplatin-based chemotherapy, this GDF-15 dependent immune cell exclusion mechanism might be at least as relevant than induction of Treg. Likewise, inhibition of dendritic cell maturation by GDF-15 (Zhou et al. PLoS One 2013, 8(11):e78618) could explain why GDF-15high tumors are immunologically cold.

The authors propose that the pathways discovered by them contributed to the "heating up" of the tumor microenvironment after oxalilatin-based chemotherapy. The authors should thus look in their data sets for the presence of cytotoxic T cells and their possible correlation with TXNIP and GDF-15 levels.

The paragraph on GDF-15 receptors needs to be corrected: The purported role of a type 2 transforming growth factor (TGF)-beta receptor in GDF-15 signalling had been due to a frequent contamination of recombinant GDF-15 with TGF-beta (Olsen et al. PLoS One 2017, 12(11):e0187349). There have been a number of screenings for GDF-15 receptors that have all failed to show an interaction between GDF-15 and TGF-beta receptors. Instead, only GFRAL was found in these large-scale screenings (Emmerson et al. Nat Med 2017, 23(10):1215-1219, Hsu et al. Nature 2017, 550(7675):255-259, Mullican et al. Nat Med 2017, 23(10):1150-1157, Yang et al. Nat Med 2017, 23(10):1158-1166). The one subsequent report that shows a link between GDF-15, engagement of CD48 on T cells and induction of a regulatory phenotype (Wang et al. J Immunother Cancer 2021, 9(9)) still awaits independent validation. Considering that CD48 lacks an intracellular signaling domain that would be critical for a classical receptor function, I recommend to be more cautious regarding the role of CD48 as GDF-15 receptor. Given the mechanism outlined by Wang et al. the word interaction partner might be more apt. Moreover, an anti-GDF-15 antibody would be a good control for the experiments involving an anti-CD48 antibody in Figure 5.

Cell surface externalization of annexin A1 has been described as a failsafe mechanism to prevent inflammatory responses during secondary necrosis (PMID: 20007579). Thus, I am surprised that the authors list annexin A1 among the immune-stimulatory molecules exposed or released in response to chemotherapy-induced cell death (line 103). Please clarify!

Referee Cross-Commenting

Regarding the cross-comment by referee 2: In my opinion, the data shown in Figure 3C-H clearly demonstrates that TXNIP can repress GDF-15 expression. I agree that there will likely be further regulators. The GDF-15 promoter is constantly regulated by a multitude of factors (which mostly induce transcription). As downregulation of GDF-15 in response to oxaliplatin is the opposite of the frequently described induction of GDF-15 upon chemotherapy, net effects may always be "smudged" by contributions from different pathways (e.g. by cell stress due to siRNA transfection). Therefore, I believe that the data are as good as it will get. Accordingly, I would not force the authors to further amplify the observed effect.

cross comment regarding referee #1: I share the general assessment of the referee and recognize the very detailed mechanistic analysis. To further support the moderate effects of the MondoA knockdown, a small molecule inhibitor like SBI-477 might be useful. (I had already suggested using this inhibitor to support these data.) Still, my view on the potential relevance of oxaliplatin-induced, TXNIP-independent downregulation of GDF-15 differs from that of referee 1. In the clinics, platinum-based chemotherapy is one of the strongest inducers of GDF-15 (compare Breen et al. GDF-15 Neutralization Alleviates Platinum-Based Chemotherapy-Induced Emesis, Anorexia, and Weight Loss in Mice and Nonhuman Primates. Cell Metabolism 32(6), P938-950, 2020.DOI:https://doi.org/10.1016/j.cmet.2020.10.023). I was thus surprised that the authors found a pathway, which leads to an outcome that an exactly opposite effect. Thus far, the only obvious reason for reduced GDF-15 secretion upon treatment with cytotoxic drugs was a reduction in tumor cell number due to cytotoxicity. Still, the authors managed to convince me that the described pathway (ROS - MondoA - TXNIP - GDF-15) exists. (Here, I still largely concur with referee 1.) Moreover, as we have identified some factors required for GDF-15 biosynthesis that could easily interact with TXNIP, I find the proposed mechanism plausible. Nevertheless, as a downregulation of GDF-15 in response to chemotherapy is hardly ever observed in late-stage cancers, I believe that the observed switch in pathway activation between early- and late-stage cancers might be highly relevant - in particular, as there is so much evidence for platinum-based induction of GDF-15 in late-stage cancer patients. Emphasizing the divergent clinical observations (e.g. by Breen et al.) could thus help to put the finding into perspective. Analysing TXNIP-independent mechanisms involved in the oxaliplatin-dependent repression of GDF-15, as suggested by referee #1, will require enormous efforts and resources, and may still turn out to be fruitless. Personally, I would thus be content if the authors just mentioned possible contributions from other pathways upon cancer progression. To me, the described pathway seems to be limited to early-stage cancers, and the actual finding that GDF-15 is downregulated is an interesting observation, irrespective of further involved pathways.

cross comment regarding referee #2: I fully agree with the referee that activation of the pathway by further chemotherapeutic drugs could be a valuable addition. As Guido Kroemer´s lab has described oxaliplatin to induce a more immunogenic cell death compared to other platinum-based chemotherapies, even a rather limited comparison between oxaliplatin and cisplatin could be very interesting.

Significance

In general, this is a very interesting manuscript describing a cascade of events that may contribute to successful chemotherapy (which likely requires induction of an immune response against dying tumor cells.) The observation that this pathway is only active in early/non-metastatic cancer cells is striking. Unfortunately, the authors cannot explain inactivation of this pathway in later stage/ metastatic/ highly aggressive cancers. Understanding this switch could easily be the most important finding triggered by this report. Therefore, I highly recommend to make some effort in this direction. Strikingly, the authors find that disruption of TXNIP-mediated GDF-15 downregulation is strongly associated with worse prognosis. They also suggest that this ratio could indicate whether a patient will respond to oxaliplatin-based chemotherapy.

Altogether, the findings described in manuscript are very novel and may have prognostic (or, in case of the presumed loss of the MondoA - TXNIP - GDF-15 pathway) therapeutic implications. Thus, the manuscript certainly fills various gaps and should be of major interest for cell biologists working on immunogenic cell death, or colorectal cancer, or MondoA, TXNIP or GDF-15. Still, due to its translational implications, it would also be worthwhile reading for a large number of researchers in the oncology field.

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 Deng et al. investigates a mechanistic link between TXNIP and GDF15 expression and oxaliplatin treatment and acquired resistance. They observe an upregulation in TXNIP expression in the tumors of patients who have previously received chemotherapy. They demonstrate oxaliplatin-driven MondoA transcriptional activity is what underlies the induction of TXNIP. They further demonstrate that TXNIP is a negative regulator of GDF15 expression. Together, oxaliplatin induces MondoA activity and TXNIP expression, resulting in a downregulation of GDF15 expression and consequently decreased Treg differentiation.

Major Comments

1. The authors suggest that TXNIP induction and GDF15 downregulation are a common effect of chemotherapies; however, the mechanistic studies were limited to oxaliplatin. The authors should clarify this point through further investigation using other commonly used CRC chemotherapies (5-FU, irinotecan, etc.), or through textual changes. To be clear, I think that the oxaliplatin results could potentially stand on their own but would require additional clarification. For example, regarding the patient samples analyzed in 1D and 4F, which patients received oxaliplatin? Could the analysis of publicly available molecular data be drilled down to just the patients who received oxaliplatin?
2. The data demonstrating the induction of MondoA transcriptional activity and TXNIP expression in response to oxaliplatin treatment is quite convincing. The data regarding ROS induction of TXNIP is interesting, especially in light of other studies arguing that ROS limits MondoA activity (PMID: 25332233). Given this apparent disparity, I think that this study could really be strengthened by also investigating other potential mechanisms of oxaliplatin induction of MondoA. In particular, given many studies arguing for direct nutrient-regulation of MondoA, the authors should address the potential for oxaliplatin regulation of glucose availability and a potential glucose dependence of oxaliplatin-induced TXNIP.
3. In line with the previous point, since MondoA activity and TXNIP expression are sensitive to glucose levels, the authors should investigate oxaliplatin-regulation of TXNIP under physiological glucose levels. No need to replicate everything, just key experiments.
4. The authors did a good job of linking TXNIP and GDF15 in untreated conditions; however, the data arguing for oxaliplatin regulation of GDF15 through TXNIP is less clear. For example, in 3B-H, oxaliplatin treatment reduces GDF15 approximately to the same extent in the NTC and TKO cells, potentially in line with a mechanism of downregulation that doesn't involve TXNIP.

Minor Comments

1. The presentation of data in Figure 5 is confusing. A-B include raw cell numbers, whereas C-F show "normalized proliferation." What does this mean? And how was the normalization done?

Referee Cross-Commenting

cross-comment regarding reviewer #1

I agree with the referee that the link between TXNIP and GDF15 is weak, though as I mentioned before, this is particularly true in the context of oxaliplatin-regulation of TXNIP. I agree that given all the presented data, it is likely that oxaliplatin-regulation of TXNIP and GDF15 are independent. In my opinion, the referee brought up all valid concerns, but this is by far the biggest concern that I share.

cross-comment regarding reviewer #3

The major concern that this referee addresses is whether another transcription factor supersedes the proposed MondoA/TXNIP induction in regulating GDF15 expression in later stage CRC. In my opinion, this another other concerns of the referee are all valid, but still I remain unconvinced that TXNIP induction underlies the oxaliplatin-regulation of GDF15. I think fleshing out that aspect of the study would potentially help the authors tease apart how this potential MondoA-TXNIP-GDF15 axis is dysregulated later in CRC progression.

Significance

Generally speaking the experiments are well controlled and the findings are significant and novel. Though the link between MondoA activity and ROS could be strengthened, and the data could be validated under more physiological settings. Further, the authors should clarify their interpretations so as to not overstate the findings.

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 well done and interesting paper examining the connection between TXNIP and GDF15. The main thrust is that TXNIP upregulation chemotherapies, such as Oxa, results in an a down regulation of GDF15 early in tumorigenesis. Later in tumorigenesis, TXNIP upregulation is less pronounced, elevating GFP15 resulting in a blockage of tumor suppressive immune responses. Generally the work is convincing. For example, it's clear that TXNIP is up regulated by Oxa in an ROS and MondoA-dependent manner. Likewise its quite clear TXNIP loss reads to an upregulation of GDF15. However, it's also quite clear that Oxa suppresses GDF15 in a manner that appears to be completely independent of TXNIP. The writing in the paper implies strongly that there is a mechanistic connection between TXNIP and GDF15, but no experiments investigate this possibility. It seems equally likely that TXNIP and GDF15 represent independent parallel pathways. Even if TXNIP is a direct regulator of GDF15, it's also clear that other "factors" up or down-regulated by Oxa also contribute to the regulation of GDF15. These are not explored and even though TXNIP is highly regulated genes shown Figure 2 that are not identified or discussed that may also be contributing to GDF15 regulation. Further, the experiments treating PBMCs with conditioned media contain other cytokines/factors, in addition to GDF15, that likely also contribute the observed effects on the different immune cells understudy. The conditioned media from GDF15 knock out cells are a good experiment, but the media is not rigorously tested to see what other cytokines/factors might have also been depleted. Perhaps a GDF15 complementation experiment would help here. Finally, even if completely independent, a TXNIP/GDF15 ratio does seem to have utility in determining chemo-therapeutic response.

Other major points:

1. Please label the other highly regulated genes shown in Fig 2A and B. Might they also explain some of the underlying biology. This could be on the current figures or in a supplement, though the former is preferred.
2. Please address why the TXNIP induction is so much less in patient-derived organoids vs. cell line spheroids (Fig S2). By the western blots, TXNIP inductions in the organoids looks quite modest. Further, the text is quite cryptic and implies that the "upregulation" is similar in both organoids and spheroids.
3. What was the rationale of performing the MS experiment on control and TXNIP KO DLD1 cells in the absence of oxaliplatin? The other experiments in Fig 3 clearly show that Oxa can repress GDF15 even in the absence of TXNIP, which implicates other pathways. ARRDC4? Or something else? This needs to be addressed.
4. The data in 3J with the MondoA knockdown is not convincing. The knockdown is weak and TXNIP goes down a smidge. Agree that GDF15 goes up

Minor points

1. Line 79. The "loss" of TXNIP/GDF15 axis is confusing. It's really loss of TXNIP and upregulation of GDF15, right?
2. Please provide an explanation for the different stages in tables 1 and 2. This will likely not be clear to non-clinicians.
3. Line 231 should probably read ...cysteine (NAC), a reactive oxygen species inhibitor,
4. Line 247, should be RT-qPCR I think.
5. Lines 343-345. I don't quite understand the wording. Does this mean to say that 675 soluble proteins were not changed between the condition media from both cell populations?
6. The data in FigS1 B and C don't seem to reach the standard p value of > 0.05

Referee Cross-Commenting

cross comment regarding referees 2 and 3 above. I'm am convinced that TXNIP is at least contemporaneously upregulated with GDF15 dowregulation. However, the strong implication from the writing is that TXNIP regulates GDF15 directly. I agree with the comment above that exploring mechanisms may be open-ended especially as TXNIP has been implicated in gene regulation by several different mechanism. I'd be satisfied with a more open-minded discussion of potential mechanisms by which TXNIP may repress GDF15 and the possibility of other parallel pathways that likely contribute to GDF15 repression.

Significance

This is an interesting contribution but the mechanistic connection between GDF15 and TXNIP is relatively weak. That said, even as independent variables they do seem to have utility in predicting therapeutic response.

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 are sincerely thankful to all reviewers for their work and constructive comments that allowed us to improve the quality of the present manuscript. We are very pleased to announce that we were able to tackle all the raised concerns (except the reporter assays which is a focus of future research for our laboratory, see below), and would like to briefly mention here three major improvements:

1) We have crossed our data in frog with available human ATAC-Seq datasets. We have followed a similar approach to the one we employed in Xenopus, by “subtracting” human osteoblastic ATAC-Seq peaks with human liver, heart and lung. This cross-species validation strategy led to the identification of osteoblast-specific NFRs in human that compare very well to the Xenopus osteoblastic regulatory landscape (new Fig 6). 2) We have included ChIP-Seq data that was performed by Patricia Hanna, a former PhD student from our laboratory, in collaboration with Laurent Sachs and Nicolas Buisine (these three researchers were incorporated as new co-authors). We were planning to publish this ChIP-Seq separately but find that it contributes very well to this manuscript (modified Fig 4, new Fig 7). 3) We have included in situ hybridization analyses on frog and shark performed by David Muñoz, a former PhD student from our laboratory, in collaboration with Melanie Debiais-Thibaud and Catherine Boisvert (these three researchers were incorporated as new co-authors). This data ends nicely the manuscript by providing a biological dimension and by strengthening our evolutionary model (See new Fig 7).

We hope that our responses match the quality criteria of Review Commons and of its affiliated journals, thank you very much once again and kind regards, Sylvain Marcellini

Point-by-point description of the revisions:

Reviewer #1 (Evidence, reproducibility and clarity (Required)): Summary: The paper investigates the genetic mechanisms driving osteoblast differentiation in Xenopus tropicalis, shedding light on bone diseases and early skeletal evolution. Through ATAC-seq analysis, the study identifies osteoblast-specific regulatory regions, confirming their role as osteogenic transcriptional enhancers. A substantial number of these enhancers are conserved in humans, potentially offering insights into skeletal disorders. Additionally, the research highlights an evolutionary perspective by revealing shared regulatory elements between Xenopus tropicalis and the elephant shark, suggesting an ancient origin for mineralized tissues in vertebrates.

Major comments:

Methodology of this paper is kinda vague and the paper seems to be fragmented and not logically organized in a linear fashion.

Reply: We have improved the methodology section. We provide the accession numbers for all raw sequencing datasets generated for this study have been submitted and linked to the NCBI BioProject database (page 27). The paper has been almost completely rewritten and the figures substantially modified. There are now less figure which contain more information presented in a friendly fashion. The logic of the paper is as follows: -Identification of enhancers and promoters (Figs 1 and 2) -Characterization of their nucleotide sequence and TFBSs (Fig 3) -Validation with RNA-Seq and ChIP-Seq (Fig 4) -Global sequence conservation (Fig 5) -Cross validation with ATAC-Seq in human (Fig 6) -Evolutionary model (Fig 7).

Authors could provide evidentiary support that the control tissues are non-mineralized (and exp tissues are) by simple calcein staining. Mineralization occurs during tadpole stage, and calcification of heart and lung tissue in amphibians is not well understood. This will strengthen the attestation of these tissues as controls and provide a useful diagram for exactly what tissues were used.

Reply: We have performed Alizarin reg staining on larval skull, liver, heart and lung and show that, like in mammals, only the calvaria is mineralized (see page 6 and new Supporting Information 1).

There appears to be no mention of osteocytes or other cell types. What measures were taken to ensure that osteoblasts are the principal cell type being described? The reference for bone tissue extraction refers to a cell culture technique in which it is likely no osteocytes would prevail.

Reply: This is an important point to clarify because osteoblasts and their osteocytic progeny harbour a completely different function, physiology and gene expression profile. Our laboratory has studied frog osteocytes in details (Fritz et al, 2018), and we have added the following sentence “Of note, this extraction procedure does not harvest osteocytes that lie embedded within the bone matrix, allowing us to exclusively study osteoblasts. As controls, we also included larval liver, heart and lung following the criteria that they are nonmineralized (Supporting information 1) and unrelated to skeletal tissues”. See page 6.

Minor comments:

Data on conservation of mentioned transcription factors could be easily added (NFAT, etc.)

Reply: We have performed extensive protein alignments showing broad conservation of the osteogenic transcription factors for which we detected binding site enrichment in osteoblast-specific enhancers (see page 10 and new Supporting Information 7).

The data presentation is poor, especially figure 2 and figure 4.

Reply: Following the reviewer’s advice these figures have been eliminated and replaced by Figures 2B and 2C, which, we believe, present the same information in a much clearer and friendly fashion.

Line 115-117: "By focusing on annotated Xt transcription start sites (TSSs), we found that the ATAC-Seq NFR and mononucleosome signals form two distinct clusters," it would be helpful to briefly explain the significance of these two clusters. What does it indicate about the regulatory regions associated with TSSs?

Reply: We have clarified this point by being more explicit: “The first cluster is composed of 5,949 promoters harbouring a robust NFR located immediately upstream of the TSS and flanked by two well-positioned nucleosomes (Fig 1B, left panel), likely corresponding to expressed genes. By contrast, the second cluster contains 16,947 promoters showing weak NFR and diffuse mononucleosome signals (Fig 1B, right panel), and is probably enriched in transcriptionally repressed genes or genes expressed at low levels”. See Page 6.

Line 133-139: When discussing hierarchical clustering and the similarity of NFR landscapes between different tissues, you could provide a sentence or two to speculate on the potential biological implications. For instance, why might heart and lung tissues exhibit more similarity in NFR landscapes compared to osteoblasts and liver?

Reply: This is an interesting point to raise because there is data in the literature supporting our findings. We have modified the following sentence on page 7: “Hierarchical clustering showed that the landscape of the NFRs from heart and lung are more similar to each other than to osteoblasts or liver, which is true both for TSS and non-TSS regions (Fig 1D) and which parallels data obtained in mouse [10]”. Our novel analysis with human ATAC-Seq data also leads to the same finding (Page 13): “Available human liver, heart and lung ATAC-Seq datasets were retrieved, and hierarchical clustering confirmed a higher similarity for heart and lung, and that the osteoblast sample substantially differs from the three other tissues (Supporting information 11), similarly to the situation in frog (Fig 1D) and mouse [10]”.

Line 134: To enhance clarity, you might consider using phrases like "Figure 3A" and "Figure 3B" instead of "Compare Fig 3A and B" to directly refer to the figures in the text.

Reply: This has been corrected has we have deeply improved the figures. See “Globally, the Pearson correlation coefficient was much higher for TSS than non-TSS peaks (Fig 1D), a finding consistent with previous studies showing that, between distinct cell types, histone marks are largely invariable at promoters while they display highly context-dependent patterns at enhancers [6, 7].” on page 7.

Line 142-144: Please consider briefly explaining why you chose liver, heart, and lung tissues as controls. What specific characteristics or functions of these tissues make them suitable for this comparative analysis?

Reply: We now mention “As controls, we also included larval liver, heart and lung following the criteria that they are nonmineralized (Supporting information 1) and unrelated to skeletal tissues.” on page 6.

When discussing the potential function of osteoblastic enhancers in cartilaginous fish, you might briefly mention the role of cartilage in these organisms and how these enhancers may have evolved to regulate cartilage-related processes.

Reply: We agree with the reviewer that this is an exciting point which is of high interest for our laboratory (see for instance our review, Cervantes et al, 2017). However, as we feel that the manuscript is already quite long and has many references, we preferred not to discuss this point and to simply focus on the osteoblast/odontoblast aspect of skeletal evolution.

Ensure that the formatting of your methods section is consistent. For example, consistently use italics for software/tool names (e.g., "SAMtools") and follow a standard format for listing parameters or options used in software/tools.

Reply: We have corrected these points.

Reviewer #1 (Significance (Required)): The paper's significance lies in its elucidation of osteoblast-specific regulatory regions in Xenopus tropicalis. By characterizing these regions and connecting them to specific genes and pathways, the study advances our understanding of osteogenesis. Additionally, the identification of conserved elements across vertebrates provides insights into the deep evolutionary origins of skeletal features, offering a unique perspective on vertebrate evolution. However, one of the main limitations of the study is the lack of extensive experimental validation for the identified regulatory regions, leaving a gap in confirming their functionality.

Reply: Thank you very much again for your helpful and constructive comments. As a functional validation, at least from the chromatin perspective, we have incorporated ChIP-Seq data (Fig 4) with four key histone marks present at active promoters (H3K4me3), active enhancers (H3K4me1), and at active chromatin (H3K27Ac) and repressed chromatin (H3K27me3). This ChIP-Seq was already available in our laboratory (thereby explaining the incorporation of three new co-authors, Dr Hanna, Dr Sachs and Dr Buisine), but we were planning to incorporate it in a different manuscript. However, we feel that it is important to include it in the present paper. Another functional validation lies in the identification of 138 conserved osteogenic enhancers harbouring a NFR both in frog and human (Fig 6). We do not intend to incorporate reporter assays at this stage, as this is a future direction of research for our laboratory, together with CRISPR mutagenesis.

Reviewer #2 (Evidence, reproducibility and clarity (Required)): In this study, Hector Castillo and the coauthors conducted ATAC-seq and RNA-seq analyses across several cell types in Xenopus tropicalis (Xt) to identify regulatory elements specific to osteoblasts. They explored the evolutionary conservation of the osteoblast regulatory elements across species. Their research encompassed the identification of osteoblast-specific regulatory elements through cross-tissue analysis, offering comprehensive insights into tissue-specific regulatory elements. These insights included cell-type-specific chromatin accessibility, biological functions predicted by gene ontology analysis, and potential transcriptional regulators associated with these regions. The cross-species analysis unveiled partial conservation of osteoblast-specific regulatory regions between the Xt and the human genome, with the shared genomic regions being linked to osteoblast-related genes. Additionally, the enriched transcription factors were identified in these regions. The study further explored comparative analyses involving multiple species, providing evolutionary insights into the gene regulatory mechanisms underlying osteoblast identity and pathology.

Major comment All the cross-species analyses in this study were primarily based on sequence conservation. However, since human osteoblast ATAC-seq data, as well as ChIP-seq and Hi-C data, are publicly available (PMID: 35906483), conducting a direct comparative analysis between Xenopus tropicalis (Xt) osteoblast ATAC-seq and human osteoblast ATAC-seq could provide more concrete evidence regarding the conservation of chromatin-accessible regions between these two species. This additional analysis has the potential to significantly strengthen the conclusions drawn in the study.

Reply: We are thankful to the reviewer for this insightful comment that dramatically improved the scope of our work. We have indeed incorporated available ATAC-Seq experiments performed on human osteoblasts (SRR12933513 and SRR12933514), liver (SRR21927033 and SRR21927032), heart (SRR21927531 and SRR21927534) and lung (SRR21927095 and SRR21927098). This is explained on pages 13-14 (results), pages 19-20 (discussion) and pages 22-24 (methods). Hence, we have uncovered 138 conserved enhancers that display an osteoblast-specific NFR both in frog and human (see new Fig 6). As the reviewer states, we believe that our conclusions have been significantly strengthened, allowing us to reformulate the manuscript title which now vehiculates a more functional message. Also, thanks to this comment, we were able to propose a more attractive title for our work: “Cross-validation of conserved osteoblast-specific enhancers illuminates bone diseases and early skeletal evolution”.

Minor comment 1 The authors made use of "annotated human enhancers" in their study; however, the specific definition or source of this annotation was not provided in the manuscript. It is crucial that the authors clarify the criteria or source used for annotating human enhancers to ensure transparency and allow readers to better understand the basis of their analyses and conclusions.

Reply: The reviewer is correct. These “annotated human enhancers” have now completely been eliminated for the study and replaced by the analysis shown in Fig 6 (and see our reply to the previous comment).

Minor comment 2 In relation to the association studies conducted between Xenopus tropicalis (Xt) osteoblast enhancers and genes related to human bone diseases, it's important for the authors to express their statements with caution. While the putative target genes may be potentially regulated by shared regulatory elements between Xt and humans, there exists no direct evidence demonstrating that these regulatory regions are the causative factors behind these diseases. It's worth noting that there are several other open chromatin regions in proximity to these putative target genes. As a result, the shared genomic regions may or may not have a direct relationship with human diseases. To establish a substantial linkage, more in-depth analyses would be required to provide evidence of a pathological connection.

Reply: This is an important point, on page 14 we now state “While the osteoblast-specific regulatory regions reported here might not be directly involved in the aetiology of the aforementioned diseases, their identification considerably improves our understating of the transcriptional control of these genes”.

Minor comment 3 In lines 394 to 397, the authors assert that the enrichment of TWIST1/2 transcription factor binding sites (TFBS) at Xenopus tropicalis (Xt) osteogenic enhancers is a novel finding. However, this claim lacks clarity regarding the novelty of this discovery, given that they reference previous literature (reference 42) that has already demonstrated the involvement of TWIST1/2 in osteoblast differentiation. The authors should provide a more precise explanation of how their specific findings related to TWIST1/2 TFBS enrichment contribute to existing knowledge or differ from previous studies to clarify the novelty of their results.

Reply: We now provide a clearer explanation by mentioning “In this respect, the reported enrichment in TWIST1/2 TFBS (Fig 3 and Supporting information 5) represents the first evidence that TWIST proteins might control the timing of osteoblastic differentiation through binding to hundreds of osteogenic enhancers, a possibility that could be confirmed by ChIP-Seq” on page 19.

Minor comment 4 Depositing the NGS data, including ATAC-seq and RNA-seq datasets, in a public database would be a valuable contribution to the research community.

Reply: Yes, this data has now been made available, see pages 26-27: “Data Availability. The raw sequencing datasets generated for this study have been submitted and linked to the NCBI BioProject database with the following accession numbers: PRJNA1011469 (ATAC-seq), PRJNA1021677 (RNA-seq), and PRJNA1056467 (ChIP-seq)”.

Reviewer #2 (Significance (Required)): The comparative analysis of ATAC-seq among different cell types in Xenopus tropicalis (Xt) provides a broad perspective on cell-type-specific chromatin accessible regions, which is a notable strength of the study. It's worth highlighting that, as far as known, this study represents the first report of ATAC-seq in Xt osteoblasts. However, it's important to acknowledge that the overall message of the study is consistent with previous findings in mammals. For example, the observation that non-transcription start site (TSS) regions were more cell-type-specific, correlating with cell-type distinct gene expressions, aligns with findings in mammalian systems. Additionally, many of the osteoblast regulators predicted from the data are already known osteogenic factors in mammals. The cross-species analysis provides valuable insights into the evolutionary aspects of putative enhancers in osteoblasts. The study identifies conserved gene regulatory regions and putative transcription factors associated with these genomic regions, shedding light on their potential roles in gene regulation. Moreover, the identification of conserved regions possibly linked to human skeletal diseases is a noteworthy aspect of the research, showcasing its strengths. However, it's essential to acknowledge a potential limitation related to this aspect of the study: the analyses conducted so far have been descriptive, primarily focusing on DNA sequence conservation. Given that several osteoblast ATAC-seq datasets from different species are publicly available, a more direct comparison between the Xt dataset and these other datasets could provide a deeper understanding of enhancer conservation and evolution. This study offers valuable resources for researchers in the field of skeletal biology and evolution. The comprehensive analysis of osteoblast-specific regulatory elements in Xenopus tropicalis, along with insights into their conservation and potential roles in human skeletal diseases, provides a foundation for further investigations in this area. Additionally, the evolutionary insights offered by the cross-species analysis contribute to the growing body of knowledge in evo-devo studies, shedding light on the evolution of gene regulatory mechanisms related to osteoblast identity. These resources and insights can serve as a valuable reference and guide for future research endeavors in both bone biology and evolutionary developmental biology. This reviewer specializes in the study of gene regulatory mechanisms in skeletal development and metabolism, primarily utilizing mouse and human tissues.

Reply: Thank you very much again for your helpful and constructive comments.

Reviewer #3 (Evidence, reproducibility and clarity (Required)): Summary Starting with ATAC-seq on the Xenopus tropicalis (Xt) genome, the authors tried to identify regulatory regions, which were evolutionally conserved and critical for osteoblasts, through computational approaches. They obtained profiles of the nucleosome-free regions (NFRs), i.e., open chromatin regions, in bone, liver, heart, and lung of Xt. The NFRs contain TSS-associated regions (TSS regions) and non-TSS regions. They then identified tissue-specific NFRs. Tissue-specific NFRs were predominantly located in introns and intergenic regions, and the trend was more highlighted in non-TSS regions. Regarding osteoblast-specific NFRs, non-TSS regions were associated with genes related to osteoblasts. Osteoblast-specific TSS- and non-TSS regions were enriched with motifs of osteoblast-related transcription factors (TFs), including Smad, AP-1, TEAD, Runx2, Nfic, Twist, and Nfat. By integrating ATAC-seq data with RNA-seq data, they found that osteoblast-specific NFRs were associated with transcriptionally active genes. When inter-species conservation of the Xt tissue-specific NFRs was analyzed, osteoblast-specific ones were well conserved in human, chick, and Callorhinchus milii (elephant shark). The authors further identified human homologous regions to Xt osteoblast-specific NFRs, which were enriched with binding motifs of osteoblast-related TFs, proposing putative osteogenic enhancers associated with skeletal diseases. Lastly, they identified a set of Xt osteoblast-specific NFRs that were conserved with the human, chick, and elephant shark genomes. The putative target genes of NFRs are enriched with osteogenesis-related TFs. Based on these data, they propose that evolutionary origins of osteoblast and odontoblasts are common, given that elephant shark is a cartilaginous fish, where bone is absent but odontoblast is present.

Major comments A major critical concern on this work is that their findings and claim fully rely on bioinformatic analyses. Bioinformatic prediction should be verified by wet-type experiments. Otherwise, it is quite difficult to draw definitive conclusions. In particular, it remains to be verified if the "putative enhancers" that they computationally identified have actual enhancer activities in in-vivo contexts. ATAC-seq alone identifies open chromatin regions on the genome and is not enough to define the location of enhancers and their activities. The authors need to perform ChIP-seq for enhancer marks and reporter assays for enhancer activities, in order to verify their prediction on at least several key regions they propose.

Reply: We have taken very seriously the reviewer´s comments and have incorporated three major experimental validations that go beyond bioinformatic analyses: -ChIP-Seq data on 4 key histone marks, previously performed in our laboratory, performed on Xenopus primary osteoblasts (see Fig 4). -Available human ATAC-Seq data for osteoblasts and control tissues (see new Fig 6). -In situ hybridization on elephant shark dental plates (see new Fig 7). We therefore have deeply modified the whole manuscript and now propose a more attractive title for our work: “Cross-validation of conserved osteoblast-specific enhancers illuminates bone diseases and early skeletal evolution”. We were not able to incorporate Reporter assays because (i) these experiments are lengthy, (ii) the current manuscript is already quite extensive and (iii) this is a major future research focus of our laboratory.

Minor comments Line 146: Fig S2 is unlikely to be provided.

Reply: We would like to keep this data available for readers, former Fig S2 is now “Supporting Information 3”.

Lines 158 to 163 and Fig. 4: GO analysis was performed only on non-TSS peaks. What about TSS peaks?

Reply: We now state on page 8 “Due to the low number of regions, no significant results were obtained with lung-specific non-TSS ATAC-Seq peaks, or with any category of TSS”.

Line 269: In the text, the authors describe that 48 osteoblast-specific TSS peaks are aligned to corresponding regions on the human genome. However, Fig. S7 shows 46 peaks are aligned. Please double-check.

Reply: This discrepancy has now been corrected.

Lines 289 to 296, Figs. 8, and S11: Although TRPS1 appears in Fig. S11, the authors did not mention it in the main text and Fig. 8. Why is the gene specifically excluded from the explanation?

Reply: This omission has now been corrected and now trps1 appears in Fig 6C, in Supporting Information 12, and is mentioned in the abstract and at pages 13-14 “Some cross-validated osteoblastic promoters and enhancers are located at loci of genes involved in skeletal diseases (See Supporting information 12 and Ref. [49]), such as osteoarthritis (adam12), osteoporosis (etv1), geroderma osteodysplasticum (gorab), keipert syndrome (gpc4), buschke-Ollendorff syndrome (lemd3), cleidocranial dysplasia (runx2) and trichorhinophalangeal syndrome type I (trps1).”.

Reviewer #3 (Significance (Required)): - This work is potentially interesting, not just leading to identification of regulatory regions critical for osteoblast biology, but also providing evolutionary insight into bone development. However, as mentioned, lack of validation of bioinformatic prediction is a major weakness of this work. This work's concept would engage the interest in the field of bone development and skeletal transcriptional programs. However, the reviewer is not sure how much this work engages general interest. - Expertise of the reviewer is mammalian skeletal development, particularly focusing on gene regulatory networks and epigenome during the process.

Reply: Thank you very much again for your helpful and constructive comments.

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

Starting with ATAC-seq on the Xenopus tropicalis (Xt) genome, the authors tried to identify regulatory regions, which were evolutionally conserved and critical for osteoblasts, through computational approaches. They obtained profiles of the nucleosome-free regions (NFRs), i.e., open chromatin regions, in bone, liver, heart, and lung of Xt. The NFRs contain TSS-associated regions (TSS regions) and non-TSS regions. They then identified tissue-specific NFRs. Tissue-specific NFRs were predominantly located in introns and intergenic regions, and the trend was more highlighted in non-TSS regions. Regarding osteoblast-specific NFRs, non-TSS regions were associated with genes related to osteoblasts. Osteoblast-specific TSS- and non-TSS regions were enriched with motifs of osteoblast-related transcription factors (TFs), including Smad, AP-1, TEAD, Runx2, Nfic, Twist, and Nfat. By integrating ATAC-seq data with RNA-seq data, they found that osteoblast-specific NFRs were associated with transcriptionally active genes. When inter-species conservation of the Xt tissue-specific NFRs was analyzed, osteoblast-specific ones were well conserved in human, chick, and Callorhinchus milii (elephant shark). The authors further identified human homologous regions to Xt osteoblast-specific NFRs, which were enriched with binding motifs of osteoblast-related TFs, proposing putative osteogenic enhancers associated with skeletal diseases. Lastly, they identified a set of Xt osteoblast-specific NFRs that were conserved with the human, chick, and elephant shark genomes. The putative target genes of NFRs are enriched with osteogenesis-related TFs. Based on these data, they propose that evolutionary origins of osteoblast and odontoblasts are common, given that elephant shark is a cartilaginous fish, where bone is absent but odontoblast is present.

Major comments

A major critical concern on this work is that their findings and claim fully rely on bioinformatic analyses. Bioinformatic prediction should be verified by wet-type experiments. Otherwise, it is quite difficult to draw definitive conclusions. In particular, it remains to be verified if the "putative enhancers" that they computationally identified have actual enhancer activities in in-vivo contexts. ATAC-seq alone identifies open chromatin regions on the genome and is not enough to define the location of enhancers and their activities. The authors need to perform ChIP-seq for enhancer marks and reporter assays for enhancer activities, in order to verify their prediction on at least several key regions they propose.

Minor comments

Line 146: Fig S2 is unlikely to be provided. Lines 158 to 163 and Fig. 4: GO analysis was performed only on non-TSS peaks. What about TSS peaks? Line 269: In the text, the authors describe that 48 osteoblast-specific TSS peaks are aligned to corresponding regions on the human genome. However, Fig. S7 shows 46 peaks are aligned. Please double-check. Lines 289 to 296, Figs. 8, and S11: Although TRPS1 appears in Fig. S11, the authors did not mention it in the main text and Fig. 8. Why is the gene specifically excluded from the explanation?

Significance

• This work is potentially interesting, not just leading to identification of regulatory regions critical for osteoblast biology, but also providing evolutionary insight into bone development. However, as mentioned, lack of validation of bioinformatic prediction is a major weakness of this work. This work's concept would engage the interest in the field of bone development and skeletal transcriptional programs. However, the reviewer is not sure how much this work engages general interest.
• Expertise of the reviewer is mammalian skeletal development, particularly focusing on gene regulatory networks and epigenome during the process.

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, Hector Castillo and the coauthors conducted ATAC-seq and RNA-seq analyses across several cell types in Xenopus tropicalis (Xt) to identify regulatory elements specific to osteoblasts. They explored the evolutionary conservation of the osteoblast regulatory elements across species. Their research encompassed the identification of osteoblast-specific regulatory elements through cross-tissue analysis, offering comprehensive insights into tissue-specific regulatory elements. These insights included cell-type-specific chromatin accessibility, biological functions predicted by gene ontology analysis, and potential transcriptional regulators associated with these regions.

The cross-species analysis unveiled partial conservation of osteoblast-specific regulatory regions between the Xt and the human genome, with the shared genomic regions being linked to osteoblast-related genes. Additionally, the enriched transcription factors were identified in these regions. The study further explored comparative analyses involving multiple species, providing evolutionary insights into the gene regulatory mechanisms underlying osteoblast identity and pathology.

Major comment

All the cross-species analyses in this study were primarily based on sequence conservation. However, since human osteoblast ATAC-seq data, as well as ChIP-seq and Hi-C data, are publicly available (PMID: 35906483), conducting a direct comparative analysis between Xenopus tropicalis (Xt) osteoblast ATAC-seq and human osteoblast ATAC-seq could provide more concrete evidence regarding the conservation of chromatin-accessible regions between these two species. This additional analysis has the potential to significantly strengthen the conclusions drawn in the study.

Minor comment 1

The authors made use of "annotated human enhancers" in their study; however, the specific definition or source of this annotation was not provided in the manuscript. It is crucial that the authors clarify the criteria or source used for annotating human enhancers to ensure transparency and allow readers to better understand the basis of their analyses and conclusions.

Minor comment 2

In relation to the association studies conducted between Xenopus tropicalis (Xt) osteoblast enhancers and genes related to human bone diseases, it's important for the authors to express their statements with caution. While the putative target genes may be potentially regulated by shared regulatory elements between Xt and humans, there exists no direct evidence demonstrating that these regulatory regions are the causative factors behind these diseases. It's worth noting that there are several other open chromatin regions in proximity to these putative target genes. As a result, the shared genomic regions may or may not have a direct relationship with human diseases. To establish a substantial linkage, more in-depth analyses would be required to provide evidence of a pathological connection.

Minor comment 3

In lines 394 to 397, the authors assert that the enrichment of TWIST1/2 transcription factor binding sites (TFBS) at Xenopus tropicalis (Xt) osteogenic enhancers is a novel finding. However, this claim lacks clarity regarding the novelty of this discovery, given that they reference previous literature (reference 42) that has already demonstrated the involvement of TWIST1/2 in osteoblast differentiation. The authors should provide a more precise explanation of how their specific findings related to TWIST1/2 TFBS enrichment contribute to existing knowledge or differ from previous studies to clarify the novelty of their results.

Minor comment 4

Depositing the NGS data, including ATAC-seq and RNA-seq datasets, in a public database would be a valuable contribution to the research community.

Significance

The comparative analysis of ATAC-seq among different cell types in Xenopus tropicalis (Xt) provides a broad perspective on cell-type-specific chromatin accessible regions, which is a notable strength of the study. It's worth highlighting that, as far as known, this study represents the first report of ATAC-seq in Xt osteoblasts. However, it's important to acknowledge that the overall message of the study is consistent with previous findings in mammals. For example, the observation that non-transcription start site (TSS) regions were more cell-type-specific, correlating with cell-type distinct gene expressions, aligns with findings in mammalian systems. Additionally, many of the osteoblast regulators predicted from the data are already known osteogenic factors in mammals.

The cross-species analysis provides valuable insights into the evolutionary aspects of putative enhancers in osteoblasts. The study identifies conserved gene regulatory regions and putative transcription factors associated with these genomic regions, shedding light on their potential roles in gene regulation. Moreover, the identification of conserved regions possibly linked to human skeletal diseases is a noteworthy aspect of the research, showcasing its strengths. However, it's essential to acknowledge a potential limitation related to this aspect of the study: the analyses conducted so far have been descriptive, primarily focusing on DNA sequence conservation. Given that several osteoblast ATAC-seq datasets from different species are publicly available, a more direct comparison between the Xt dataset and these other datasets could provide a deeper understanding of enhancer conservation and evolution.

This study offers valuable resources for researchers in the field of skeletal biology and evolution. The comprehensive analysis of osteoblast-specific regulatory elements in Xenopus tropicalis, along with insights into their conservation and potential roles in human skeletal diseases, provides a foundation for further investigations in this area. Additionally, the evolutionary insights offered by the cross-species analysis contribute to the growing body of knowledge in evo-devo studies, shedding light on the evolution of gene regulatory mechanisms related to osteoblast identity. These resources and insights can serve as a valuable reference and guide for future research endeavors in both bone biology and evolutionary developmental biology.

This reviewer specializes in the study of gene regulatory mechanisms in skeletal development and metabolism, primarily utilizing mouse and human tissues.

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 paper investigates the genetic mechanisms driving osteoblast differentiation in Xenopus tropicalis, shedding light on bone diseases and early skeletal evolution. Through ATAC-seq analysis, the study identifies osteoblast-specific regulatory regions, confirming their role as osteogenic transcriptional enhancers. A substantial number of these enhancers are conserved in humans, potentially offering insights into skeletal disorders. Additionally, the research highlights an evolutionary perspective by revealing shared regulatory elements between Xenopus tropicalis and the elephant shark, suggesting an ancient origin for mineralized tissues in vertebrates.

Major comments:

Methodology of this paper is kinda vague and the paper seems to be fragmented and not logically organized in a linear fashion. Authors could provide evidentiary support that the control tissues are non-mineralized (and exp tissues are) by simple calcein staining. Mineralization occurs during tadpole stage, and calcification of heart and lung tissue in amphibians is not well understood. This will strengthen the attestation of these tissues as controls and provide a useful diagram for exactly what tissues were used. There appears to be no mention of osteocytes or other cell types. What measures were taken to ensure that osteoblasts are the principal cell type being described? The reference for bone tissue extraction refers to a cell culture technique in which it is likely no osteocytes would prevail.

Minor comments:

Data on conservation of mentioned transcription factors could be easily added (NFAT, etc.) The data presentation is poor, especially figure 2 and figure 4. Line 115-117: "By focusing on annotated Xt transcription start sites (TSSs), we found that the ATAC-Seq NFR and mononucleosome signals form two distinct clusters," it would be helpful to briefly explain the significance of these two clusters. What does it indicate about the regulatory regions associated with TSSs? Line 133-139: When discussing hierarchical clustering and the similarity of NFR landscapes between different tissues, you could provide a sentence or two to speculate on the potential biological implications. For instance, why might heart and lung tissues exhibit more similarity in NFR landscapes compared to osteoblasts and liver? Line 134: To enhance clarity, you might consider using phrases like "Figure 3A" and "Figure 3B" instead of "Compare Fig 3A and B" to directly refer to the figures in the text. Line 142-144 :Please consider briefly explaining why you chose liver, heart, and lung tissues as controls. What specific characteristics or functions of these tissues make them suitable for this comparative analysis? When discussing the potential function of osteoblastic enhancers in cartilaginous fish, you might briefly mention the role of cartilage in these organisms and how these enhancers may have evolved to regulate cartilage-related processes. Ensure that the formatting of your methods section is consistent. For example, consistently use italics for software/tool names (e.g., "SAMtools") and follow a standard format for listing parameters or options used in software/tools.

Significance

The paper's significance lies in its elucidation of osteoblast-specific regulatory regions in Xenopus tropicalis. By characterizing these regions and connecting them to specific genes and pathways, the study advances our understanding of osteogenesis. Additionally, the identification of conserved elements across vertebrates provides insights into the deep evolutionary origins of skeletal features, offering a unique perspective on vertebrate evolution. However, one of the main limitations of the study is the lack of extensive experimental validation for the identified regulatory regions, leaving a gap in confirming their functionality.

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. Content has not been altered except for formatting.

Learn more at Review Commons

---

Reply to the reviewers

1. General Statements [optional]

We would like to extend our warmest thanks to the reviewers for their constructive comments and strong support for our study.

2. Point-by-point description of the revisions

Reviewer #1:

Table

1. It would be nice to have a table of isoform, dose, promoter, enhancer and other conditions tested and the brief summary of phenotype as Table.

We thank the reviewer for this valuable suggestion and have now included a summary Table (Table 1) cited in the last result section.

Discussion

1. This experiment was done on knockout condition but in real patient different form of mutant protein will exist in retinal tissue. Authors indicated that co‐expression of short and long form of FAM161A worked better to rescue function. How would authors cope with interfering endogenous mutant protein in real patients?

We thank the reviewer for raising this interesting point. Most mutations described so far are nonsense or frameshift mutations common to both long and short isoforms which, consequently, are not present at the protein level (Beryozkin et al 2020, doi.org/10.1038/s41598-020-72028-0, Matsevich et al 2022, doi.org/10.1016/j.xops.2022.100229). Thus, we don’t expect to have an imbalance between the remaining functional alleles and the therapeutic ones. However, we cannot exclude the discovery of missense mutations and the effect of such allele would have to be molecularly evaluated to determine if gene replacement is limited for this specific condition. This question could be assessed in cellular models by co-expression of both mutated and WT-tagged proteins or in organoid models.

1. Related to the first question, the expression of these retinal structural proteins will be different in mice and human. How would authors optimize the vector for human patient gene therapy?

Aware that the 60% homology between the human and mouse protein could cause important limitations for the evaluation of the vector in the mouse model, we are continuing the validation of our vectors in human retina organoïds. We plan to test both the reliable localization of the human isoforms in WT organoid and the rescue of structural photoreceptor defects of FAM161A-deficient human organoids. In parallel, vector-derived expression will also be validated in non-human primates.

Reviewer #2:

Scotopic and photopic ERG were performed to study retinal function. However, mouse behavior tests such as optomotor response should be employed to confirm vision restoration.

In our hand, we didn’t notice a significant modification of the optomotor response between 4 and 16 weeks (for figure on visual acuity changes with age in Fam161atmb/tmb mice (n=6-9), see uploaded word document), and consequently of the estimated visual acuity, in Fam161atmb/tmb mice at 3.5 months corresponding to the endpoint of our study (see figure below). In a separate study to this work, we are thus conducting a follow-up long term gene therapy study to be able to complete the functional analysis of the gene therapy rescue with the optomotor response at age with significant decreased visual acuity in untreated mice compared to WT. We will have to wait at least 6 months to expect to see a difference between groups.

The immunostaining in Figure 3 has some noise. Filtering the blocking solution before use could improve the quality of the staining.

We thank the reviewer for this suggestion. The blocking solution was already filtered and the limited success of the mouse FAM161A staining is due to the imperfect recognition of anti-human FAM161A antibodies to the mouse protein.

In Figure 5f, the data of wildtype mice should be included for comparison.

As noted by reviewer 3, in Fig5 F, the plain gray horizontal line surrounded by the 2 dotted ones are referring to the mean +/- SEM of the WT value respectively. We added “WT” on the right of the graph to highlight the plain line.

The cited paper, such as 'Garafalo AV, Cideciyan AV, Heon E, Sheplock R, Pearson A, WeiYang Yu C, Sumaroka A, Aguirre GD, and Jacobson SG. Progress in treating inherited retinal diseases: Early subretinal gene therapy clinical trials and candidates for future initiatives. Prog Retin Eye Res. 2020;77(100827),' should be an original research paper, not a review article.

As noted by reviewer 3, we think appropriate to cite this review which is a complete reference to the different gene therapy approaches developed for inherited retinal diseases.

Major:

Fig 1A‐B. Do hTERT‐RPE1 cells endogenously express FAM161A? This set of images lacks a negative control (i.e., no transfected RPE1 cells). Western blot of FAM161A is recommended, similar to Fig 1C.

We previously showed that hTERT-RPE1 cells express FAM161A in the basal body of the centriole (Di Gioia 2015), but we recognized that it is not apparent in Figure 1A and B, probably due to a limitation of the antibody reactivity which labeled only overexpressed proteins. We thus performed additional experiments using the human ARPE19 cell line to demonstrate endogenous FAM161A expression in untransfected cells and to perform a Western blot from human transfected cells. We observed that in untransfected cells FAM161A labeling is weak and is only revealed in the centriole labeled by centrin after a long exposure time (Figure 1A). When FAM161A HS or HL is overexpressed the FAM161A labeling is present in the cell body, very strong, and is observed with short exposure time (Figure 1A). We also extracted protein from untransfected and HS- or HL-transfected ARPE-19 cells to identify the FAM161A protein by Western blot (Figure 1B). Thus, we added the negative control and a western blot from human cells to answer reviewer comments.

Fig 1C. The authors noted in the discussion that HS isoform is more abundant than HL isoform from human retinal extract. Although this is from 661W, a mouse photoreceptor cell line, it seems this is aligned with the notion. To echo with the last comment, I am curious to see if under the same transfection, the HS isoform is preferentially expressed in hTERT‐RPE1 cells.

We do not think that transfection experiment is sufficient to prove that HS is preferentially express than HL. Even if we transfect the same amount of DNA, we would need an internal control for transfection to allow relative quantification of the protein expression after transfection. However, we performed an additional experiment in human RPE cells using the ARPE-19 cell line which is more efficiently transfected than hTERT-RPE1 in our hands. As shown in Figure 1B, we observed again more abundant expression of HS in these human transfected cells. However, we cannot exclude difference in transfection efficiency between HL and HS conditions that could explain the difference in the final amount of FAM161A protein.

Fig 3 and Fig 5: low mag WT images of FAM161A are the same. But higher mag images (presumably selected from ROIs in low mag) are not the same. Please make sure of no duplication images.

We are facing technical limits with the labeling of the mouse Fam161A. The antibodies available have limited affinity for the mouse Fam161A protein. While we were able to perform Uex-M from mouse tissue samples (flatmount retina) to study Fam161A expression in the connecting cilium (Mercey et al PLoS Biol 2022), it was more challenging to obtained low magnification picture from mouse retina sections. We propose to show in Figure 3 mouse Fam161A expression obtained from retina section and keep the low magnification from a flatmount for the figure 5. Thus, there will be no duplication of images as recommended by the reviewer.

Fig 4H. HS+HL combo, and HL alone, showed almost a polarized quantification, quite variable. Can the authors speculate the reason?

Despite the fact that injections are targeting similar retinal region in treated animals, there is still variation in the localization and extend of the gene transfer due to the surgical success. Indeed, the area of retinal detachment is hard to control in the mouse as of the quality of re-attachment. Moreover, the effective dose may lightly vary when some viral particles might be loss due to reflux. One would need to treat a larger number of eyes to really conclude that HS alone would be less variable than HL alone or HS+HL. However, we could also speculate that HS+HL and HL treatments being more efficient to rescue connecting cilium length compared to HS alone (Fig 5F) could, in the best injected eyes, have a better ONL thickness rescue than the limited ONL rescue induced by HS treatment.

Also can the authors comment on if there is any associated notable inflammation especially in high tier dosage (10^11 GC)?

We didn’t follow inflammation directly by fundus examination or OCT imaging following injection. However, despite the high dose used in our successful conditions (10E11 GC/eye), we didn’t notice any differences in the general mouse welfare after injection compare to lower doses. Systemic administration of Rimadyl (carprofen) was however adapted to each mouse during the 24 hrs post-surgery. In comparison to other groups with lower vector doses, no particular emergence of inflammatory cells or damages were observed by histology.

Can the authors comment on the difference in the injection time, PN14‐15 in this study vs. PN24‐29 in their previous study? Have the authors attempted to treat the older mice with the optimized vector?

The gene therapy study using the mouse cDNA was performed before establishing the time course of connecting cilia disruption in the Fam161atmb/tmb mouse (Mercey et al. 2022). Following the observation that CC develop similarly to healthy animal up to postnatal day 10, we decided to treat the mouse earlier for the second gene therapy study using human proteins. Nonetheless, the action of the vector occurred when the cilium is already disorganized as we expect expression of the WT Fam161A from 2 weeks post-injection. We are now testing treatments at different ages, including PN28, to determine the therapeutic window and if the optimal conditions (dose, ratio) may vary with the age at treatment.

Can the authors speculate on why IRBP‐GRK1 human FAM161A did not realize functional rescue (Fig 2) as it did with mouse FAM161A (previous work)?

Our hypothesis to explain the absence of functional rescue following IRBP-GRK1 vector injection is that the difference in human protein distribution compared to the mouse protein in the mouse retina could impact the function of the photoreceptor by interfering with physiological process such as transport. As mentioned in our discussion: “overexpression of these proteins could saturate the transport system impacting the cellular processes”.

As mentioned in the discussion, there is only 60% of homology between human and mouse proteins which could induce a major impact on protein localization and function. Post-translational modification which are also known to be crucial for modulating connecting cilium addressing (Rao et al. 2016) could also differ and impact both human protein distribution and function (for example 3 cysteines in the human protein sequence could be palmytoylated (C359, C366, C367) and are absent in the mouse sequence). Moreover, the exact role of the human long and short isoforms are unknown and their adaptability to the mouse system not yet identified. Further studies should be performed to understand the consequence of such differences on the function and to unravel the function of both long and short human isoforms in the retina.

Minor:

While the manuscript is overall well communicated, there are areas requiring further proofread. For example, in the Abstract section, "In 15 years" should be "For 15 years", "14‐days FAM161atm1b/tm1b mice" should be "14‐day old". In the Introduction, "... suggesting that protein miss‐localization" should be "mis‐localization". In the last paragraph before Discussion, "(iii) the restauration of CC..." should be "restoration", etc.

We corrected these errors and carefully proofread the whole manuscript to avoid typing mistakes.

I recommend the authors to use a table to summarize different promoters, titers and key findings (e.g., expression level, localization) used and refer back to each figure.

We thank the reviewer for this valuable suggestion and have now included a summary Table (Table 1) cited in the last result section.

Scale bars on all figures, or every set of images.

We added scale bars on figures containing microscopic images.

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 manuscript led by Arsenijevic and Chang is an important technical development to the ocular gene therapy space, and touches on the important aspect of structural protein restoration by gene therapy, that is, the precise control of localization and subsequent functional realization. Overall the manuscript is well written, and the experiments are technically sound, with limitations acknowledged.

To briefly summarize, the authors wanted to understand precise control of FAM161A expression and connecting cilium (CC) restoration. They built on, and extended their previous work that showed limited structural and functional rescue by photoreceptor expression of the longer isoform of mouse FAM161A in Fam161a KO driven by IRBP-GRK1 promoter. In the current work, the authors experimented with delivering human ortholog of FAM161A cDNA, short, or long, or both isoforms using newly devised, relatively weak promoters. The main readouts include retinal morphology (e.g., ONL thickness), ERG, and protein localization by IHC (e.g., correct location, no ectopic expression). It is worth noting that the authors highlighted the use of expansion microscopy technology to examine the connecting cilium (CC) organization and protein expression, which may minimize the use of TEM for CC structure determination and enable acceleration.

My enthusiasm for recommending it for publication is high. Nonetheless, I have the following comments, hoping the authors could address to further improve the manuscript.

Major:

Fig 1A-B. Do hTERT-RPE1 cells endogenously express FAM161A? This set of images lacks a negative control (i.e., no transfected RPE1 cells). Western blot of FAM161A is recommended, similar to Fig 1C.

Fig 1C. The authors noted in the discussion that HS isoform is more abundant than HL isoform from human retinal extract. Although this is from 661W, a mouse photoreceptor cell line, it seems this is aligned with the notion. To echo with the last comment, I am curious to see if under the same transfection, the HS isoform is preferentially expressed in hTERT-RPE1 cells..

Fig 3 and Fig 5: low mag WT images of FAM161A are the same. But higher mag images (presumably selected from ROIs in low mag) are not the same. Please make sure of no duplication images.

Fig 4H. HS+HL combo, and HL alone, showed almost a polarized quantification, quite variable. Can the authors speculate the reason? Also can the authors comment on if there is any associated notable inflammation especially in high tier dosage (10^11 GC)?

Can the authors comment on the difference in the injection time, PN14-15 in this study vs. PN24-29 in their previous study? Have the authors attempted to treat the older mice with the optimized vector?

Can the authors speculate on why IRBP-GRK1 human FAM161A did not realize functional rescue (Fig 2) as it did with mouse FAM161A (previous work)?

Minor:

While the manuscript is overall well communicated, there are areas requiring further proofread. For example, in the Abstract section, "In 15 years" should be "For 15 years", "14-days FAM161atm1b/tm1b mice" should be "14-day old". In the Introduction, "... suggesting that protein miss-localization" should be "mis-localization". In the last paragraph before Discussion, "(iii) the restauration of CC..." should be "restoration", etc.

I recommend the authors to use a table to summarize different promoters, titers and key findings (e.g., expression level, localization) used and refer back to each figure.<br /> Scale bars on all figures, or every set of images.

Referees cross-commenting

To reviewer #2, Fig5f - WT data was shown as the gray horizontal line. I had the same question but then saw they noted in the legends. I think it is fine to cite the PRER review article to make their point.

I agree with the comments addressed by Reviewer #1 and am glad we both raise the point of using table for summarization.

Significance

This well-drafted paper represents a technical development that could supplement current gene therapy strategies to certain ciliopathies. In this particular case, the authors chose FAM161A, a disease causal gene to retinitis pigmentosa-28 and encodes for a microtubule-associated ciliary protein involved in organizing the connecting cilium in photoreceptors. Of importance, the authors devised novel promoters to drive gene expression and took advantage of expansion microscopy for quickly examining cilia proteins and structures. Conceptually, the techniques developed in this manuscript could be applicable to several other inherited retinal dystrophies that share similar disease mechanisms.

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

Arsenijevic et al. investigated the therapeutic function of the FCBR1-F0.4 promoter-driven expression of both the short and long isoforms of human FAM161A. The results showed that this method not only repaired the disorganized connecting cilium but also restored the appropriate expression and localization of other proteins in the connecting cilium, thus restoring retinal function. Additionally, the study systematically evaluated the AAV dose, different promoters, and FAM161A isoforms' effects on retinal survival and function. Overall, the study is novel and robust. Here are some suggestions that may help improve the manuscript:

Scotopic and photopic ERG were performed to study retinal function. However, mouse behavior tests such as optomotor response should be employed to confirm vision restoration.

The immunostaining in Figure 3 has some noise. Filtering the blocking solution before use could improve the quality of the staining.

In Figure 5f, the data of wildtype mice should be included for comparison.

The cited paper, such as 'Garafalo AV, Cideciyan AV, Heon E, Sheplock R, Pearson A, WeiYang Yu C, Sumaroka A, Aguirre GD, and Jacobson SG. Progress in treating inherited retinal diseases: Early subretinal gene therapy clinical trials and candidates for future initiatives. Prog Retin Eye Res. 2020;77(100827),' should be an original research paper, not a review article.

Referees cross-commenting

Agree with the comments addressed by Reviewer #1 and #3

Significance

Overall, the manuscript is clear and interesting. I suggest a major resion for the 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 #1

Evidence, reproducibility and clarity

The manuscript "Fine-tuning FAM161A gene augmentation therapy to restore retinal function" submitted by Arsenijevic et al., describes gene therapy for RP28 caused by mutation in FAM161A in human. The authors worked on Fam161a-dficient mice by testing different isoforms, dose, promoter and enhancers to control the expression level and localization of the protein to functionally rescue the mice to prevent blindness. The tight control of protein expression is required for mutation in genes coding structural proteins in the retina.

The authors have clearly showed the optimized combination of conditions to restore function of Fam161atm1b/tm1b mice and also area of improvement to make.

Comments

Table

1. It would be nice to have a table of isoform, dose, promoter, enhancer and other conditions tested and the brief summary of phenotype as Table.

Discussion

1. This experiment was done on knockout condition but in real patient different form of mutant protein will exist in retinal tissue. Authors indicated that co-expression of short and long form of FAM161A worked better to rescue function. How would authors cope with interfering endogenous mutant protein in real patients?
2. Related to the first question, the expression of these retinal structural proteins will be different in mice and human. How would authors optimize the vector for human patient gene therapy?

Significance

This is an important and excellent work showing tight control of expression is required for future retinal gene therapy.

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 [optional]

We are thankful to the reviewers for the time and effort invested in assessing our manuscript and for their suggestions to improve it. We have now considered the points raised by them, carried out additional experiments, and modified the text and figures to address them. We feel that the new experiments and modifications have been able to solve all concerns raised by the reviewers and have improved the manuscript substantially, strengthening and extending our conclusions.

The main modifications include:

• We have extended the analysis of the overexpression strains to highly stringent conditions, which revealed a mild acidification defect for the strain overexpressing Oxr1. In addition, we have included in our analysis a strain in which both proteins are overexpressed, which resulted in a further growth defect.
• We have analyzed the recruitment of Rtc5 to the vacuole under additional conditions: deletion of the main subunit of the RAVE complex RAV1, medium containing galactose as the sole carbon source and pharmacological inhibition of the V-ATPase. These experiments allowed us to strengthen and extend our conclusions regarding the requirements for Rtc5 targeting to the vacuole.
• We have analyzed V-ATPase disassembly in intact cells, by addressing the localization to the vacuole of subunit C (Vma5) in glucose and galactose-containing medium. The results strengthen our conclusion that both Rtc5 and Oxr1 promote an in vivo state of lower V-ATPase assembly.
• We have extended our analyses of V-ATPase function to medium containing galactose as a carbon source, since glucose availability is one of the main regulators of V-ATPase function in vivo. The results are consistent with what we observed in glucose-containing medium.
• We have included a diagram of the structure of the V-ATPase for reference.
• We have included a diagram and a paragraph describing Oxr1 and Rtc5 regarding protein length and domain architecture and comparing them to other TLDc domain-containing proteins.
• We have made changes to the text and figures to improve clarity and accuracy, including a methods section that was missing. We include below a point-by-point response to the reviewers´ comments.

2. Point-by-point description of the revisions

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

__ __Suggestions:

1. The authors observed that knockout of Rtc5p or Oxr1p does not affect vacuolar pH. If Rtc5p and Oxr1p both cooperate to dissociate V-ATPase, the authors may wish to characterize the effect of a ∆Rtc5p∆Oxr1p double knockout on vacuolar pH. The double mutant ∆rtc5∆oxr1 was already included in the original manuscript (the growth test is shown in Figure 5 B and the BCECF staining is shown in Figure 5C). This strain behaved like wt in both of these assays. Of note, what we observe for the deletion strains is increased assembly (Figure 5 D - G), so we expect that it would be hard to observe a difference in vacuole acidity or growth in the presence of metals.

Therefore, we have now also included a strain with the double overexpression of Oxr1 and Rtc5. Since overexpression of the proteins results in decreased assembly, it is more likely that this strain will show impaired growth under conditions that strongly rely on V-ATPase activity. Indeed, we observed that the overexpression of Oxr1 alone resulted in a slight growth defect in media containing high concentrations of ZnCl2 and the double overexpression strain showed an even further defect (Figure 6 A and C).

The manuscript would benefit from a well-labelled diagram showing the subunits of V-ATPase (e.g. in Figure 2D).

We agree with the reviewer and we have now added a diagram of the structure of the V-ATPase labeling the different subunits in Figure 2B.

The images of structures, especially in Figure 1-Supplement 1B, are not particularly clear and could be improved (e.g. by removing shadows or using transparency).

We are thankful to the reviewer for this suggestion. To improve the clarity of the structures in Figure 1 C and Figure 1 – Supplement 1A, we are now presenting the different subunits in the structures with different shades of blue and grey.

The authors should clearly describe the differences between Rtc5p and Oxr1p in terms of protein length, sequence identity, domain structure, etc.

We are thankful for this suggestion and we have now included a diagram of the domain architecture and protein length of Rtc5 and Oxr1, comparing with two human proteins containing a TLDc domain in Figure 5A. In addition, we have added the following paragraph describing the features of the proteins.

“Rtc5 is a 567 residue-long protein. Analysis of the protein using HHPred (Zimmermann et al., 2018), finds homology to the structure of porcine Meak7 (PDB ID: 7U8O, (Zi Tan et al., 2022)) over the whole protein sequence (residues 37-559). For both yeast Rtc5 and human Meak7 (Uniprot ID: Q6P9B6), HHPred detects homology of the C-terminal region to other TLDc domain containing proteins like yeast Oxr1 (PDBID: 7FDE), Drosophila melanogaster Skywalker (PDB ID: 6R82), and human NCOA7 (PDB ID: 7OBP), while the N-terminus has similarity to EF-hand domain calcium-binding proteins (PDB IDs: 1EG3, 2CT9, 1S6C6, Figure 5A). HHPred analysis of the 273 residue long Saccharomyces cerevisiae Oxr1, on the other hand, only detects similarity to TLDc domain containing proteins (PDB IDs: 7U80, 6R82, 7OBP), which spans the majority of the sequence of the protein (residues 71-273). The overall sequence identity between Oxr1 and Rtc5 is 24% according to a ClustalOmega alignment within Uniprot. The Alphafold model that we generated for Rtc5 is in good agreement with the available partial structure of Oxr1 (7FDE) (root mean square deviation (RMSD) of 3.509Å) (Figure 5 - S1 A), indicating they are structurally very similar, in the region of the TLDc domain. Taken together, these analyses suggest that Oxr1 belongs to a group of TLDc domain-containing proteins consisting mainly of just this domain like the splice variants Oxr1-C or NCOA7-B in humans (NP_001185464 and NP_001186551, respectively), while Rtc5 belongs to a group containing an additional N-terminal EF-hand-like domain and a N-myristoylation sequence, like human Meak7 (Finelli & Oliver, 2017) (Figure 5 A).”

Minor:

1. The "O" in VO should be capitalized. This has been corrected.

In Figure 4 supplement 1, the labels "I", "S", and "P" should be defined.

This has been clarified in the figure legend.

Please clarify what is meant by "switched labelling"

This refers to the SILAC vacuole proteomics experiments, for which yeast strains are grown in medium containing either L-Lysine or 13C6;15N2- L-Lysine to produce normal (‘light’) or heavy isotope-labeled (‘heavy’) proteins. This allows comparing two conditions. To increase the robustness of the comparisons, the experiments are done twice with both possible labeling schemes (condition A – light, condition B – heavy + condition A – heavy + condition B – light), which is commonly described as switched labeling or label switching.

We have exchanged the original sentence in the manuscript for:

“Performing the same experiments but switching which strain was labeled with heavy and light amino acids gave consistent results.”

The meaning of the sentence "Indeed, this was the case for both of them" is ambiguous.

We have now replaced this sentence with the following:

“Indeed, overexpression of either Rtc5 or Oxr1 resulted in increased growth defects in the context of Stv1 deletion (Figure 7 H and I).”

For Figure 1-Supplement 1B it is hard to see the crosslink distances.

We have updated this figure to improve the visibility of the cross-links. In addition, we now include a supplemental table (supplemental table 5) with a list of the Cα- Cα distances measured for all the crosslinks we mapped onto high-resolution structures.

The statement "The effects of Oxr1 are greater than those caused by Rtc5" requires more context. Is there a way of quantifying this effect for the reader?

We agree that this sentence was too general and vague. The effects caused by one or the other protein depend on the condition and the assay. We have thus deleted this sentence, and we think it is better to refer to the description of the individual assays performed.

The phrase "negative genetic interaction" should be clarified.

We have included in the text the following explanation of genetic interactions:

“A genetic interaction occurs when the combination of two mutations results in a different phenotype from that expected from the addition of the phenotypes of the individual mutations. For example, deletion of OXR1 or RTC5 has no impact on growth in neutral pH media containing zinc in a control background but improves the growth of RAV1 deletion strains (Figure 7 E and F), so this is a positive genetic interaction. On the other hand, overexpression of either Rtc5 or Oxr1 results in a growth defect in a background lacking Rav1 in neutral media containing zinc, a negative genetic interaction.”

* * In the sentence "Isogenic strains with the indicated modifications in the genome where spotted as serial dilutions in media with pH=5.5, pH=7.5 or pH=7.5 and containing 3 mM ZnCl2", "where" should be "were".

This has been corrected.

Figure 2D: the authors should consider re-coloring these models, as it is challenging to distinguish Rtc5p from the V-ATPase.

We have changed the coloring of this structure and added a diagram of the V-ATPase structure with the same coloring scheme to improve clarity.

Reviewer #1 (Significance (Required)):

The vacuolar protein interaction map alone from this manuscript is a nice contribution to the literature. Experiments establishing colocalization of Rtc5p to the vacuole are convincing, as is dependence of this association on the presence of assembled V-ATPase. Similarly, experiments related to myristoylation are convincing. The observed mislocalization of V-ATPases that contain Stv1p to the vacuole (which is also known to occur when Vph1p has been knocked out) upon knockout of Oxr1p is also extremely interesting. Overall, this is an interesting manuscript that contributes to our understand of TLDc proteins.

We are thankful to the reviewer for their appreciation of the significance of our work, including the interactome map of the vacuole as a resource and the advances on the understanding of the regulation of the V-ATPase by TLDc domain-containing proteins.

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

Major points:

1. The evidence of Oxr1 and Rtc5 as V-ATPase disassembly factors is circumstantial. The authors base their interpretation primarily on increased V1 (but not Vo) at purified vacuoles from Oxr1- or Rtc5-deleted strains, which does not directly address disassembly. Of course, the results regarding Oxr1 confirm detailed disassembly experiments with the purified protein complex (PMID 34918374), but on their own are open to other interpretations, e.g. suppression of V-ATPase assembly. Of note, the authors emphasize that they provide first evidence of the in vivo role of Oxr1, but monitor V1 recruitment with purified vacuoles and do not follow V-ATPase assembly in intact cells. We are thankful to the reviewer for pointing this out. We did not want to express that the molecular activity of the proteins is the disassembly of the complex, as our analyses include in vivo and ex vivo experiments and do not directly address this. We rather meant that both proteins promote an in vivo state of lower assembly of the V-ATPase. We have modified the wording throughout the manuscript to be clearer about this.

In addition, we have added new experiments to monitor V-ATPase assembly in intact cells, as suggested by the reviewer. Previous work has shown that in yeast, only subunit C leaves the vacuole membrane under conditions that promote disassembly, while the other subunits remain at the vacuole membrane (Tabke et al 2014). Our own experiments agree with what was published (Figure 3 D). We have thus monitored Vma5 localization to the vacuole under glucose or after shift to galactose containing media in cells lacking or overexpressing Rtc5 or Oxr1. We observed that cells overexpressing either TLDc domain protein show lower levels of Vma5 recruitment to the vacuole in glucose (Figure 6 D and E). Additionally cells lacking either Rtc5 or Oxr1 contain higher levels of Vma5 at the vacuole after 20 minutes in galactose medium (Figure 5 F and G). Thus, these results re-inforce our conclusions that Rtc5 and Oxr1 promote states of lower assembly.

Oxr1 and Rtc5 have very low sequence similarity. It would be helpful if the authors provided more detail on the predicted structure of the putative TLDc domain of Rtc5 and its relationship to the V-ATPase - Oxr1 structure. Is Rtc5 more closely related to established TLDc domain proteins in other organisms?

We have now included a diagram of the domain architecture of Rtc5 and Oxr1, and comparison to the features of other TLDc domain containing proteins in Figure 5 A, as well as a paragraph describing them:

“Rtc5 is a 567 residue-long protein. Analysis of the protein using HHPred (Zimmermann et al., 2018), finds homology to the structure of porcine Meak7 (PDB ID: 7U8O, (Zi Tan et al., 2022)) over the whole protein sequence (residues 37-559). For both yeast Rtc5 and human Meak7 (Uniprot ID: Q6P9B6), HHPred detects homology of the C-terminal region to other TLDc domain containing proteins like yeast Oxr1 (PDBID: 7FDE), Drosophila melanogaster Skywalker (PDB ID: 6R82), and human NCOA7 (PDB ID: 7OBP), while the N-terminus has similarity to EF-hand domain calcium-binding proteins (PDB IDs: 1EG3, 2CT9, 1S6C6, Figure 5A). HHPred analysis of the 273 residue long Saccharomyces cerevisiae Oxr1, on the other hand, only detects similarity to TLDc domain containing proteins (PDB IDs: 7U80, 6R82, 7OBP), which spans the majority of the sequence of the protein (residues 71-273). The overall sequence identity between Oxr1 and Rtc5 is 24% according to a ClustalOmega alignment within Uniprot. The Alphafold model that we generated for Rtc5 is in good agreement with the available partial structure of Oxr1 (7FDE) (root mean square deviation (RMSD) of 3.509Å) (Figure 5 - S1 A), indicating they are structurally very similar, in the region of the TLDc domain. Taken together, these analyses suggest that Oxr1 belongs to a subfamily of TLDc domain-containing proteins consisting mainly of just this domain like the splice variants Oxr1-C or NCOA7-B in humans (NP_001185464 and NP_001186551, respectively) , while Rtc5 belongs to a subfamily containing an additional N-terminal EF-hand-like domain and a N-myristoylation sequence, like human Meak7 (Finelli & Oliver, 2017) (Figure 5 A).”

The authors conclude vacuolar recruitment of Rtc5 depends on the assembled V-ATPase, based on deletion of different V1 and Vo domain subunits. However, these genetic manipulations likely cause a strong perturbation of vacuolar acidification; indeed, the images show drastically altered vacuolar morphology. To strengthen their conclusion, it would be helpful to show that Rtc5 recruitment is not blocked by inhibition of vacuolar acidification, and that conversely it is blocked by deletion of rav1.

We are thankful to the reviewer for this insightful suggestion and we have now performed both experiments suggested. The experiment regarding rav1Δ is now Figure 3C, and we observed that this mutation also disrupts Rtc5 localization to the vacuole. In addition, we decided to include an experiment showing the subcellular localization of Rtc5 after shifting the cells to galactose containing medium for 20 minutes, as a physiologically relevant condition that results in disassembly of the complex (Figure 3D). We observed that under these conditions Rtc5 re-localizes to the cytosol. This result is particularly interesting given that in yeast only subunit C (but not other V1 subunits) re-localizes to the cytosol under these conditions. In addition, the experiment using Bafilomycin A to inhibit the V-ATPase shows that Rtc5 is still localized at the vacuole membrane under conditions of V-ATPase inhibition (Figure 3 F). Taken together these results allowed us to strengthen our original interpretation that Rtc5 requires an assembled V-ATPase for its localization and extend it to the fact that the V-ATPase does not need to be active.

Reviewer #2 (Significance (Required)):

This is an interesting paper that confirms and extends previous findings on TLDc domain proteins as a novel class of proteins that interact with and regulate the V-ATPase in eukaryotes. The title seems to exaggerate the findings a bit, as the authors do not investigate V-ATPase (dis)assembly directly and only phenotypically describe altered subcellular localization of the Golgi V-ATPase in Oxr1-deleted cells. A recent structural and biochemical characterization of Oxr1 as a V-ATPase disassembly factor (PMID 34918374) somewhat limits the novelty of the results, but the function of Oxr1 in regulating subcellular V-ATPase localization and the identification of a second potential TLDc domain protein in yeast provide relevant insights into V-ATPase regulation. This paper will be of interest to cell biologists and biochemists working on lysosomal biology, organelle proteomics and V-ATPase regulation.

We thank the reviewer for the assessment of our work, and for recognizing the novel insights that we provide. Regarding the previous biochemical work on Oxr1 and the V-ATPase, we have clearly cited this work in the manuscript. In our opinion, our results complement and extend this article, showing that the function in disassembly is relevant in vivo. Additionally, this is only one of five major points of the article, the other four being

• The interactome map of the vacuole as a resource
• The identification of Rtc5 as a second yeast TLDc domain containing protein and interactor of the V-ATPase.
• The identification of the role of Rtc5 in V-ATPase assembly.
• The identification of the role of Oxr1 in Stv1 subcellular localization. We believe these additional points add important insights to researchers interested in lysosomes, the V-ATPase, intracellular trafficking and TLDc-domain containing proteins.

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

Major comments

__1) Re: A cross-linking mass spectrometry map of vacuolar protein interactions (results) __ While XL-MS is a very powerful method, it is a high-throughput approach and there should be some kind of negative control in these experiments. In cross-linking experiments, non-cross-linked samples are usually used as negative controls. What was the negative control in cross-linking mass-spectrometry experiments here? If there was no negative control, how the specificity of interactions was evaluated? Maybe the authors analyzed the dataset for highly improbable interactions and found very few of them?

We fully agree that it is crucial to ensure the specificity of the interactions detected by XL-MS. To achieve this, one needs to control (1) the specificity of the data analysis (i.e. that the recorded mass spectrometry data are correctly matched to cross-linked peptides from the sequence database) and (2) the biological specificity (i.e. that the cross-linking captured natively occurring interactions).

To ascertain that criterion (1) is met, cross-link identifications are filtered to a pre-defined false-discovery rate (FDR) – an approach that the XL-MS field adopted from mass spectrometry-based proteomics. As a result, low-confidence identifications (e.g. cross-linked peptides that are only supported by a few signals in a given mass spectrum) are removed from the dataset. FDR filtering in XL-MS is a rather complex matter as it can be done at different points during data analysis and the optimal FDR cut-off depends on the specific scientific question at hand (for more details see for example Fischer and Rappsilber, Anal Chem, 2017). Generally speaking, an overly restrictive FDR cut-off would remove a lot of correct identifications, thereby greatly limiting the sensitivity of the analysis. On the other hand, a too relaxed FDR cut-off would dilute the correct identifications with a high number of false-positives, which would impair the robustness and specificity of the dataset. While many XL-MS study control the FDR on the level of individual spectrum matches, we opted for a 2% FDR cut-off on the level of unique residue pairs, which is more stringent (see Fischer and Rappsilber, Anal Chem, 2017). Our FDR parameters are described in the Methods section (Cross-linking mass spectrometry of isolated vacuoles - Data analysis). Of note, we have made all raw mass spectrometry data publicly available through the PRIDE repository (https://www.ebi.ac.uk/pride/ ; accession code PXD046792; login details during peer review: Username = reviewer_pxd046792@ebi.ac.uk, Password = q1645lTP). This will allow other researchers to re-analyze our data with the data analysis settings of their choice in the future.

To ascertain that criterion (2) is met, we mapped the identified cross-links onto existing high-resolution structures of vacuolar protein complexes. Taking into account the length of our cross-linking reagent, the side-chain length of the cross-linkable amino acids (i.e. lysines), and a certain degree of in-solution flexibility, cross-links can reasonably occur between lysines with a mutual Cα-Cα distance of up to 35 Å. Using this cut-off, the lysine-lysine pairs in the high-resolution structures we studied can be split into possible cross-linking partners (Cα-Cα distance 35 Å). Of all cross-links we could map onto high-resolution structures, 95.2% occurred between possible cross-linking partners. In addition, our cross-links reflect numerous known vacuolar protein interactions that have not yet been structurally characterized. These lines of evidence increase our confidence that our XL-MS approach captured genuine, natively occurring interactions. These analyses are described in more detail in the first Results sub-section (“A cross-linking mass spectrometry map of vacuolar protein interactions”).

In addition, the high purity of vacuole preparation is critical. How was it assessed by the authors?

We disagree that the purity of the vacuole preparation is critical for this analysis to be valid. The accuracy of the protein-protein interactions detected will depend on their preservation during sample preparation until the sample encounters the cross-linker, and the data analysis, as described above. The experiment would have been equally valid if performed on whole cell lysates without any enrichment of vacuoles, but the coverage of vacuolar proteins would have likely been very low. For this reason, we decided to use the vacuole isolation procedure to obtain better coverage of the proteins of this particular organelle. The use of the Ficoll gradient protocol (Haas, 1995) was based on that it is a protocol that yields strong enrichment of proteins annotated with the GO Term “vacuole” (Eising et al, 2019) and that it preserves the functionality of the organelle, as evidenced by its use for multiple functional assays (vacuole-vacuole fusion (Haas, 1995), autophagosome-vacuole fusion (Gao et al, 2018), polyphosphate synthesis by the VTC complex (Desfougéres et al, 2016), among others).

2) Re: Rtc5 and Oxr1 counteract the function of the RAVE complex (results)

Taken together, data, presented in this section of the manuscript, provide strong evidence that Rtc5 and Oxr1 negatively regulate V-ATPase activity, counteracting the V-ATPase assembly, facilitated by the activity of the RAVE complex. However, the complete deletion of the major RAVE subunit Rav1p was required to observe this effect in vivo in yeast. The other way to induce V-ATPase disassembly in yeast is glucose deprivation. It will be interesting to study if there is a synergistic effect between glucose deprivation and RTC5/OXR1 deletion on V-ATPase assembly, vacuolar pH, and growth of single oxr1Δ, rtc5Δ or double oxr1Δrtc5Δ mutants (OPTIONAL). Glucose deprivation is a more physiologically relevant condition than a deletion of an entire gene.

We would like to point out that an effect on assembly is observed without deleting the RAVE complex: deletions of Oxr1 or Rtc5 resulted in increased V-ATPase assembly in vivo in the presence of glucose and of the RAVE complex (Figures 5 D and E). We have now also added the experiments showing that the overexpression strains have a mild growth defect under conditions that force cells to strongly rely on V-ATPase activity (Figures 6 A and C).

Nevertheless, we agree that addressing the effect of changing the levels of Oxr1 and Rtc5 under low-glucose conditions is an interesting physiologically relevant question. We have now included growth assays and BCECF staining in medium containing galactose as the carbon source (Figures 5 – Supplement 1 B and C, and Figure 6 C and Figure 6- Supplement 1A). In addition, we have addressed the vacuolar localization of Vma5 in medium containing glucose or after shifting to medium containing galactose for 20 minutes, as a proxy for V-ATPase disassembly in intact cells (Figure 5 F and G, Figure 6 D and E). Taken together, these analyses reinforce our conclusions that both Rtc5 and Oxr1 promote an in vivo state of lower V-ATPase assembly, based on the following observations:

• Higher localization of Vma5 to the vacuole after 20 mins in galactose in cells lacking Oxr1 or Rtc5 (Figure 5 F and G).
• Lower localization of Vma5 to the vacuole in medium containing glucose in cells overexpressing Oxr1 or Rtc5 (Figure 6 D and E).
• Growth defect of the strain overexpressing Oxr1 in medium containing galactose with pH = 7.5 and zinc chloride, with a further growth defect caused by additional overexpression of Rtc5 (Figure 6 C). 3) Re: Figure 6 - supplement 1. The title is relevant to panel D only, it should be renamed to reflect the results of the disassembly of V-ATPase in rav1Δ mutant strains, while results about the stv1Δ-based strains (Panel D) should be shown together with similar experiments in Figure 7 - supplement 2 for clarity.

We have shifted the Panel D from the original Figure 6 – Supplement 1 to the main Figure (now Figure 7 – H and I). Regarding the title of the Figure, whether Supplemental Figures have titles or not will depend on the journal where the manuscript is published. For now, we have removed all titles from supplemental figures, as they are conceived to complement the main Figures.

4) Re: Figure 7 - supplement 1, Panel A. The proper assay to show that Stv1-mNeonGreen is functional is to express it in double mutant vph1Δstv1Δ to see if the growth defect is reversed. In addition, the vph1Δ growth defect is not changed (improved or worsened) in the presence of Stv1-mNeonGreen, so it means that the expression of Stv1-mNeonGreen does not further compromise the V-ATPase function, but it does not mean that it improves its function.

It is clear from the experiment suggested by the reviewer that they think that we have expressed Stv1-mNeonGreen from a plasmid. This was not the case, Stv1 was C-terminally tagged with mNeonGreen in the genome. It is thus the only expressed version in the strain. The experiment we have performed is thus equivalent to the one suggested by the reviewer, but for genomically expressed variants. For reference, the genotypes of all the strains used can be found in Supplemental Table 1.

5) Re: Figure 7 - supplement 2. This figure should be combined with Fig. 6- suppl 1, panel D as also mentioned above. The figure seems to lack some labels, and conclusions are not accurate as discussed below. However, this data provides important additional information about relationships between isoform-specific subunits of V-ATPase Vph1 and Stv1 and both Rtc5 and Oxr1 and should be repeated if it is not done yet to have a better idea about these relationships.

Panel B: Based on this picture, deletion of RTC5 has a negative genetic interaction with the deletion of VPH1, since double deletion mutant vph1Δ rtc5Δ grows worse than each individual mutant. Although it also means that there is no positive interaction, it is not the same.

Indeed, there is a negative genetic interaction between the deletion of RTC5 and VPH1. We have replaced the growth tests in this figure (Figure 8 – Supplement 2 A in the new manuscript) to show this negative genetic interaction better. This effect is reproducible, as shown in the repetitions of the experiments.

Panel C: Same as for panel B. Based on this picture, the deletion of OXR1 has a weak negative genetic interaction with the deletion of STV1, since double deletion mutant stv1Δ oxr1Δ grows worse than each individual mutant at 6 mM ZnCl2.

Panel D: Same as for panels B and C. Based on this picture, deletion of RTC5 has a negative genetic interaction with the deletion of STV1, since double deletion mutant stv1Δ rtc5Δ grows worse than each individual mutant at 6 mM ZnCl2. There is no label in the middle panel (growth conditions) and no growth assay data in the presence of CaCl2.

However, these results will be then in contradiction with the results from Figure 6 - Supplement 1, panel D, showing negative genetic interaction between the overexpression of Rtc5 or Oxr1 and deletion of Stv1, since both deletion and overexpression of Rtc5 or Oxr1 would have negative genetic interactions with Stv1.

For both Panels C and D (Now Figure 8 - Supplement 2 B and C). The effect pointed out by the reviewer (slightly stronger growth defect for the double mutants than for the single mutants) is very mild. We have attempted to make it more evident by assessing growth in medium with higher and lower concentrations of zinc and this was not possible. This is in contrast with the very clear positive genetic interaction that we observe between the deletion of OXR1 and VPH1 (Now Figure 8 H). This is the reason that we decided to report the lack of a positive genetic interaction instead of the presence of a negative one, as we do not want to draw conclusions based on results that are borderline detectable.

In addition, there is no label for the media in the middle panel, is it just YPAD pH=7.5, without the addition of any metals?

Indeed, the media is YPAD pH=7.5, without the addition of any metals. The line drawn above several images based on this media indicated this. Since this form of labeling appears to be confusing, we have now replaced it and placed the label directly above the image.

Why there is no growth assay in the presence of CaCl2, like in panels A and B?

Every growth test shown in the manuscript was performed including growth in YPD pH=5,5 as a control of a permissive condition for lack of V-ATPase activity, and then in YPD pH=7,5 including a broad range of Zinc Chloride and Calcium chloride concentrations. From all these pictures, the conditions where the differences among strains were clearly visible were chosen to assemble the figures. Conditions that did not provide any information for that particular experiment were not included in the figure to avoid making them unnecessarily large and crowded.

Re: Figure 7 - supplement 2, continued. How many times all these experiments were repeated? These experiments should be repeated at least 3 times, which is especially necessary for the experiments in panel C, because the effects are borderline. If results are reproducible and statistically significant, although small, the conclusion should be changed from "no positive genetic interactions" to "negative genetic interactions", which is more precise and informative.

All growth tests shown in the manuscript were repeated at least three times for the conditions shown. We are thankful to the reviewer for pointing out that this was not mentioned, and we have added this to the methods section. We have assembled a file with all repetitions of the shown growth tests and added it at the end of this file. In doing so, these are already available for the public. These repetitions show that all effects reported are reproducible. We will then discuss with the editors of the journal where this manuscript is published about the necessity of including it with the final article.

Regarding reporting the lack of a positive genetic interaction vs. a negative one, we have discussed this above. Shortly, for Panel B (Figure 8 – Supplement 2 A in the new manuscript) we have changed the conclusion to “negative genetic interaction” as adjusting the zinc chloride concentration allowed us to show this clearly and reproducibly, as shown by the repetitions of the experiments. For panels C and D (Now Figure 8 - Supplement 2 B and C), the effect is really mild and barely detectable, even when we tried a wide range of zinc chloride concentrations. For this reason, we would prefer to maintain the “no positive genetic interaction” conclusion.

Re: Methods. There is no description of yeast serial dilution growth assay at all. In addition, why the specific media (neutral pH, in the presence of high concentrations of calcium or zinc) was used is not explained either in the results or methods. Appropriate references should be included, for example, PMID: 2139726, PMID: 1491236.

We apologize for the oversight of the missing methods section, which we have now included.

Regarding the explanation of the media used, the following section was already a part of the results section, before the description of the first growth test:

“The V-ATPase is not essential for viability in yeast cells, and mutants lacking subunits of this complex grow similarly to a wt strain in acidic media. However, when cells grow at near-neutral pH or in the presence of divalent cations such as calcium and zinc, the mutants lacking V-ATPase function show a strong growth impairment (Kane et al, 2006).”

We have now replaced this with the following, more complete version:

“As a first approach for addressing the role of these proteins, we tested growth phenotypes related to V-ATPase function in strains lacking or overexpressing them. The V-ATPase is not essential for viability in yeast cells, and mutants lacking subunits of this complex grow similarly to a wt strain in acidic media, but display a growth defect at near-neutral pH the mutants (Nelson & Nelson, 1990). In addition, the proton gradient across the vacuole membrane generated by the V-ATPase energizes the pumping of metals into the vacuole, as a mechanism of detoxification. Thus, increasing concentrations of divalent cations such as calcium and zinc, generate conditions in which growth is increasingly reliant on V-ATPase activity (Förster & Kane, 2000; MacDiarmid et al, 2002; Kane, 2006).”

---

MINOR COMMENTS

Yeast proteins are named with "p" at the end, such as "Rtc5p".

This nomenclature rule is falling into disuse during the last decades, as the use of capitals vs lowercase and italics allows to distinguish between genes proteins and strains (OXR1 = gene, Oxr1 = protein, oxr1Δ = strain). As an example, I include a list of the latest papers by some of the major yeast labs around the world, all of which use the same nomenclature as we do (in alphabetical order). This list even includes some work in the field of the V-ATPase.

• Alexey Merz, USA. PMID: 33225520
• Benoit Kornmann, UK. PMID: 35654841
• Christian Ungermann, Germany. PMID: 37463208
• Claudio de Virgilio, Switzerland. PMID: 36749016
• Daniel E. Gottschling, USA. PMID: 37640943
• David Teis, Austria. PMID: 32744498
• Elizabeth Conibear, Canada. PMID: 35938928
• Fulvio Reggiori, Denmark. PMID: 37060997
• J Christopher Fromme, USA. PMID: 37672345
• Maya Schuldiner, Israel. PMID: 37073826
• Patricia Kane, USA. PMID: 36598799
• Scott Emr, USA. PMID: 35770973
• W Mike Henne, USA. PMID: 37889293
• Yoshinori Ohsumi, Japan. PMID: 37917025 In addition, we would prefer to keep the nomenclature that we already use, to keep consistency with other published articles from our lab.

Re: Introduction. In the introduction it should be indicated that Rtc5 was originally discovered as a "restriction of telomere capping 5", using screening of temperature-sensitive cdc13-1 mutants combined with the yeast gene deletion collection [PMID: 18845848]. A couple of sentences should be written about the RAVE complex and its role in V-ATPase assembly.

We are thankful for this suggestion and we have now included both pieces of information in the introduction.

*“The re-assembly of the V1 onto the VO complex when glucose becomes again available, is aided by a dedicated chaperone complex known as the RAVE complex, which also likely has a general role in V-ATPase assembly (Seol et al, 2001; Smardon et al, 2002, 2014).” *

“In our cross-linking mass spectrometry interactome map of isolated vacuoles we found that the only other TLDc-domain containing protein of yeast, Rtc5, is a novel interactor of the V-ATPase. Rtc5 is a protein of unknown function, originally described in a genetic screen for genes related to telomere capping (Addinall et al, 2008)”

Re: The TLDc domain-containing protein of unknown function Rtc5 is a novel interactor of the vacuolar V-ATPase (results)

1) It is important to understand, that Oxr1 was co-purified before with the V1 domain of V-ATPase from a certain mutant strain, not wild-type yeast [PMID: 34918374]. It may explain why the authors did not identify it in their original protein-protein interactions screen here.

The structural work on the V1 domain bound to Oxr1 (Khan et al, 2022) showed that the binding of Oxr1 prevented V1 from assembling onto the Vo. Since our experiments rely on the purification of vacuoles, they should contain mainly only V1 assembled onto the VO, and not the free soluble V1. This is likely the reason that we do not detect Oxr1, in addition to it being less abundant. We have clarified this now in the manuscript and added the fact that Oxr1 was co-purified with a V1 containing a mutant version of the H subunit.

“In a previous study, Oxr1 was co-purified with a V1 domain containing a mutant version of the H subunit, and its presence prevented the in vitro assembly of this V1 domain onto the VO domain and promoted disassembly of the holocomplex (Khan et al., 2022). This is likely the reason why we do not detect Oxr1 in our experiments, which rely on isolated vacuoles and thus would only include V1 domains that are assembled onto the membrane. In addition, Oxr1 is less abundant in yeast cells than Rtc5 according to the protein abundance database PaxDb (Wang et al, 2015).”

2) It is a wrong conclusion that because Rtc5 was co-purified with both V1 and V0 domain subunits it interacts with the assembled V-ATPase, this does not exclude a possibility that Rtc5 also interacts with separate V1 sector or separate V0 sector of V-ATPase.

We agree with the reviewer that the co-purification of Rtc5 with both V1 and VO domain subunits does not necessarily mean that it interacts with the assembled V-ATPase. Thus, we have modified the text in this part to:

“The fact that we can co-enrich Rtc5 both with Vma2 and with Vph1 indicates that it can interact either with both the VO and V1 domains or with the assembled V-ATPase.”

However, other results throughout the manuscript can be taken into account to strengthen this idea:

1. Rtc5 requires an assembled V-ATPase to localize to the vacuole membrane, and thus seems not to interact with free VO domains, which would be available when we delete V1 subunits or in medium containing galactose.
2. Rtc5 becomes cytosolic in galactose-containing media. This would indicate that it also does not interact with free V1 domains, which are still localized to the vacuole membrane under these conditions. Taken together with the pull-downs, these results suggest that Rtc5 interacts with the assembled V1-VO V-ATPase. Thus, we have included the following sentence after Figure 3, which shows the subcellular localization experiments.

*“Taking into account that Rtc5 is co-enriched with subunits of both the VO and V1 domain, and that it localizes at the vacuole membrane dependent on an assembled V-ATPase, we suggest that Rtc5 interacts with the assembled V-ATPase complex.” *

Re: Figure 1, Panel C. Is it possible to show individual proteins in different colors for clarity?

Panel D. How were cross-link distances measured? It is not obvious if you are not an expert in the field and it is not described in the methods.

We have modified Figure 1 C and Figure 1 – Supplement 1B (now Figure 1 – Supplement 1 A) to present the different subunits in the structures with different shades of blue and grey.

Furthermore, we have clarified the distance measurement approach in the methods section and in the legend of Fig 1D: “Ca-Ca distances were determined using the measuring function in Pymol v.2.5.2 (Schrodinger LLC).”

__Re: Figure 1 - Supplement 1, __

Panel A. What scientific information are we getting from this picture?

This panel was just a visual representation of the complexity of the protein network we obtained. Indeed, there was no specific scientific message, so we have decided to remove this panel from the revised manuscript.

Panel B. Why are these complexes shown separately from the complexes in Figure 1, panel C? Also, can individual proteins be colored differently here as well?

We did not want to overload Fig 1C, so we decided to show some of the protein complexes in Fig 1 – Supplement 1B. The most important information is the histogram showing that 95% of the mapped cross-links fall within the expected length range, and this is shown in the main Figure (Figure 1D). As stated above, we have adjusted the subunit coloring in Figure 1 C to improve clarity.

Re: Figure 3. It will be nice to show the localization of the untagged protein as well if antibodies are available (OPTIONAL).

Unfortunately, there are no available antibodies for either Rtc5 or Oxr1. This hinders us from detecting the endogenous untagged proteins. We would like to point out that we have been very careful in showing which tagged proteins are functional (C-terminally tagged Rtc5) and which are not (C-terminally tagged Oxr1), so that the reader can know how to interpret the localization data.

Re: Figure 4. Why different tags were used in panels A (GFP), C (msGFP2) and D

(mNeonGreen)?

In general, we prefer to use mNeonGreen as a tag for microscopy experiments because it is brighter and more stable, and msGFP2 as a tag for experiments involving Western blots because we have better antibodies available. There was a mistake in the labeling, and actually, all constructs labeled as GFP were msGFP2. We have now corrected this. Of note, we have tested the functionality of both tagged version (mNeonGreen and msGFP2).

Panels B and C. Were Rtc5 fusions detected using anti-GFP antibodies?

Indeed, Rtc5-msGFP2 was detected with an anti-GFP antibody. We have now indicated next to each Western blot membrane the primary antibody used. In addition, all antibodies are detailed in Supplemental Figure 3.

The authors should have full-size Western blots available, not just cut-out bands, as some journals and reviewers require them for publication.

For all western blots, we always showed a good portion of the membrane and not cut-out bands. The cropping was performed to avoid making figures unnecessarily large. The whole membranes are of course available and will be included in an “extended data file” if required by the journal.

Re: Figure 4 - Supplement 1, Panel A. Does "-" and "+" mean -/+ Azido-Myr?

Indeed. We have now added this label to the figure.

Panel B. There is no blot with a membrane protein marker (Vam3 or Vac8), it should be included.

We have replaced this western blot for a different repetition of this experiment in which a membrane protein marker was included. Of note, the two other repetitions of the experiment shown (Figure 4 – Supplement 1 panel C and Figure 4 panel C) also include both a membrane protein marker and a soluble protein marker.

Re: Figure 5. The title does not describe all results in this figure and should be modified accordingly.

The original data from Figure 5 is now separated into Figures 5 and 6 because of the additional experiments included during revisions. We have modified the Figure titles to be descriptive of the overall message of the Figures.

Panel C. Statistical significance value for *** should be indicated in the legend.

This has been indicated in the Figure legend.

It is not clear how many times yeast growth assays were repeated. Usually, all experiments should be done in triplicates or more.

All shown growth tests were performed at least three times for the conditions shown. We have now indicated this in the materials and methods section. In addition, we now provide in this response a file with all repetitions of growth tests, which will be appended to the article if deemed necessary by the editors.

Re: Figure 5 - supplement 1. No title

Re: Figure 5 - supplement 2. No title

Whether the supplemental Figures should have a title or not will depend on the style of the journal where the manuscript is finally published. The current idea of the supplemental Figures is that they complement the corresponding main Figure. For this reason, we have removed all titles from supplemental Figures.

Re: Figure 6. There is a typo on the second lane in the legend: "...the genome were", not "...the genome where".

This has been corrected.

Panel C. Why the analysis of BCECF vacuole staining of double mutants oxr1Δrav1Δ and rtc5Δrav1Δ is not shown? Was it done at all?

We had not included this piece of data, as we thought that the genetic interaction of RTC5 and OXR1 and rav1Δ was sufficiently well supported with the included data (growth tests in combination with the deletion, growth tests in combination with the overexpression, vacuole proteomics in combination with overexpression, and BCECF staining in combination with the overexpression). Because of the request of the reviewer, we have now included this experiment as Figure 7 G.

Re: Figure 6 - Supplement 2. Why were two different tags (2xmNG and msGFP2) used?

We tried both tags to see if one of them would be functional. Unfortunately, they both resulted in non-functional proteins, as shown by the corresponding growth tests.

Did the authors study N-terminally tagged Oxr1? Was it functional?

We have tagged Oxr1 N-terminally, and this unfortunately resulted in a protein that was not completely functional. We show below the localization of N-terminally mNeon-tagged Oxr1, under the control of the TEF1 promoter. The protein appears cytosolic (Panel A) but is not completely functional (Panel B). The localization of Oxr1 had already been misreported by using a tagged version that we now show to be non-functional. For this reason, we preferred not to include this data in the manuscript, to avoid again including in the literature subcellular localizations that correspond to non-functional or partially functional proteins.

Panel B. Results for the untagged TEF1pr-Oxr1 overexpression are not shown, thus tagged and untagged proteins can't be compared. Are they available? What is the promoter for the expression of 2xmNG fusion constructs?

Oxr1-2xmNG was C-terminally tagged in the genome, which means that the promoter is the endogenous one, it was not modified. For this reason, the correct controls are a strain expressing Oxr1 at endogenous levels (the wt strain) and a strain lacking Oxr1. Both controls were included in the Figure, and in all repetitions made of this experiment. For reference, all the genotypes of the strains used are found in Supplemental Table 1.

Re: Methods. Were vacuoles prepared differently for XL-MS and SILAC-based vacuole proteomics (there are different references) and why? Methods for XL-MS and quantitative SILAC-based proteomics can be placed together for clarity.

The basis for the method of vacuole purification is the same, from (Haas, 1995). This reference was included in both protocols that include vacuole purifications. However, modifications of this method were performed to fit the crosslinking method (higher pH, no primary amines) or to fit the SILAC labeling (combination of two differentially labeled samples in one purification). The reference for the vacuole proteomics (Eising et al 2022) corresponds to a paper in which the SILAC-based comparison of vacuoles from different mutant strains was optimized, and includes not only the vacuole purification but the growth conditions and downstream processing of the vacuoles.

Since both the SILAC-based vacuole proteomics and the XL-MS are multi-step methods, containing numerous parameters including the sample preparation, processing for MS, MS run and data analysis, we would prefer to keep them separate. We think this would allow a person attempting to reproduce these methods to go through them step by step.

What is CMAC dye? Why was it used to stain the vacuolar lumen?

We apologize for this oversight, we have included the definition of CMAC as 7-Amino-4-Chlormethylcumarin. It is a standard-used organelle marker for the lumen of the vacuole.

Some abbreviations (TEAB, ACN) are not explained.

We apologize for this oversight. We have now replaced these abbreviations with the full names of the compounds in the article.

What is 0% Ficoll?

We used the term 0% Ficoll, because this is the name given to the buffer in the original Haas 1995 paper on vacuole purifications. However, we agree that the term is misleading and we have now added the composition of the buffer (10 mM PIPES/KOH pH=6.8, 0.2 M Sorbitol).

Reviewer #3 (Significance (Required)):

The vacuolar-type proton ATPase, V-ATPase, is the key proton pump, that hydrolases ATP and uses this energy to pump protons across membranes. Amazingly, this proton pump and its function are conserved in eukaryotes from yeast to mammals. While V-ATPase structure and function have been studied for more than 30 years in various organisms, its regulation is not completely understood. The very recent discoveries of two new V-ATPase interacting proteins in yeast, first Oxr1 (OXidative Resistance 1), and now Rtc5 (Restriction of Telomere Capping 5), both the only two members of TLDc (The Tre2/Bub2/Cdc16 (TBC), lysin motif (LysM), domain catalytic) proteins in yeast, provide new insights in V-ATPase regulation in yeast, and because the interaction is conserved in mammals its relevance to mammalian V-ATPases regulation as well.

TLDc proteins are best known for their role in protection from oxidative stress, in particular in yeast and in the nervous system in mammals. The discovery of the novel Rtc5-V-ATPase interaction points to the role of V-ATPase not only in protection from oxidative stress but also in restriction of telomere capping in yeast and most likely higher species. The studies of other species also highlight the possible conserved role of V-ATPase in lifespan determination and Torc1 signaling, mediated through these interactions. Thus, the discovery of this new functionally important interaction between the second TLDc family member in yeast, Rtc5, and V-ATPase will shed light on the molecular mechanisms of all these essential biological processes and pathways.

In addition, because the authors performed a comprehensive proteomics protein-protein interaction study of the purified yeast vacuole it provides a valuable resource for all researchers who study vacuoles and/or related to them lysosomes.

The follow-up functional studies using the rav1Δ strain clearly demonstrated that Rtc5 and Oxr1 disassemble V-ATPase and counteract the function of V-ATPase assembly RAVE complex in vivo in yeast. Thus, they are essentially the first discovered endogenous eukaryotic protein inhibitors of V-ATPase. Moreover, because the authors obtained the evidence that Oxr1 is the regulator of the specific subunit isoform of V-ATPase Stv1p in vivo in yeast, it suggests that different TLDc proteins may regulate different specific V-ATPase subunit isoforms in cell- and tissue-specific manner in higher eukaryotes. The mechanism of this isoform-specific regulation in yeast and other species needs further investigation in the future.

Because of the conservation of the TLDc-V-ATPase interactions, all this information can be extrapolated to higher species, all the way to humans, in whom genetic mutations in various TLDc proteins are known to cause devastating diseases and syndromes.

We are thankful to the reviewer for their positive comments about the significance of 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

Summary

In this manuscript, the authors used a proteomics approach to comprehensively study yeast vacuole protein-protein interactions using cross-linking mass-spectrometry (XL-MS). They identified 16694 interactions between 2051 proteins. Many known vacuolar protein complexes were found and used as positive controls, confirming the high quality of the dataset, however, no negative controls were reported, and this issue is raised in the 'Major comments' section. The authors then focused on one particular previously unknown protein-protein interaction between the TLDc-domain containing protein of unknown function Rtc5 and the vacuolar-type proton ATPase, V-ATPase, which acidifies yeast vacuoles. The methods and results regarding Rtc5 discovery as a novel interactor of the V-ATPase, Rtc5 myristoylation, and its V-ATPase-dependent localization to the vacuole membrane are convincing. The authors then moved on to study the in vivo function of Rtc5 as well as Oxr1, the only other TLDc-domain-containing protein in yeast. Interestingly, they did not originally detect Oxr1 in their protein-protein interaction studies, apparently due to its very low abundance in yeast. However, they found that deletion of either RTC5 or OXR1 in vivo resulted in more assembled V-ATPase at the yeast vacuole and this effect was stronger in oxr1Δ cells. However, RTC5/OXR1 deletion or overexpression in parental yeast strains did not affect either vacuolar pH (a readout of functional V-ATPase) or yeast growth, including growth under specific conditions (neutral pH, in the presence of high concentrations of calcium or zinc), which is used to reveal a conditional lethal phenotype of unfunctional V-ATPase (the Vma− phenotype). Since they did not observe any in vivo phenotype in parental yeast strains, they subsequently studied the effects of RTC5/OXR1 deletion and overexpression in the 'sensitized' rav1Δ strain, lacking a specific assembly factor of V-ATPase, Rav1, one of the subunits of RAVE complex. In this strain, RTC5/OXR1 overexpression resulted in less acidic vacuolar pH and reduced growth of double mutant cells, compared to the single rav1Δ mutant. In addition, overexpression of Oxr1, but not Rtc5, caused disassembly of the V-ATPase in rav1Δ cells, noteworthy this effect was not detectable in the parent strain with intact Rav1p. Finally, they found that in oxr1Δ cells there is more Stv1 in the vacuole and concluded that Oxr1 is necessary for the retention of Stv1 containing V-ATPase at the vacuole. However, the mechanism seems to be complicated and remains to be elucidated. In summary, an impressive variety of methods from a technologically advanced XL-MS to classical yeast growth assays were used to identify Rtc5 interaction with V-ATPase and analyze its functional role in vivo in yeast, making the conclusions well justified overall.

---

Major comments

Re: A cross-linking mass spectrometry map of vacuolar protein interactions (results)

While XL-MS is a very powerful method, it is a high-throughput approach and there should be some kind of negative control in these experiments. In cross-linking experiments, non-cross-linked samples are usually used as negative controls. What was the negative control in cross-linking mass-spectrometry experiments here? If there was no negative control, how the specificity of interactions was evaluated? Maybe the authors analyzed the dataset for highly improbable interactions and found very few of them? In addition, the high purity of vacuole preparation is critical. How was it assessed by the authors? All this is important to know to use this dataset as a reliable resource in the future.

Re: Rtc5 and Oxr1 counteract the function of the RAVE complex (results)

Taken together, data, presented in this section of the manuscript, provide strong evidence that Rtc5 and Oxr1 negatively regulate V-ATPase activity, counteracting the V-ATPase assembly, facilitated by the activity of the RAVE complex. However, the complete deletion of the major RAVE subunit Rav1p was required to observe this effect in vivo in yeast. The other way to induce V-ATPase disassembly in yeast is glucose deprivation. It will be interesting to study if there is a synergistic effect between glucose deprivation and RTC5/OXR1 deletion on V-ATPase assembly, vacuolar pH, and growth of single oxr1Δ, rtc5Δ or double oxr1Δrtc5Δ mutants (OPTIONAL). Glucose deprivation is a more physiologically relevant condition than a deletion of an entire gene.

Re: Figure 6 - supplement 1. The title is relevant to panel D only, it should be renamed to reflect the results of the disassembly of V-ATPase in rav1Δ mutant strains, while results about the stv1Δ-based strains (Panel D) should be shown together with similar experiments in Figure 7 - supplement 2 for clarity.

Re: Figure 7 - supplement 1, Panel A. The proper assay to show that Stv1-mNeonGreen is functional is to express it in double mutant vph1Δstv1Δ to see if the growth defect is reversed. In addition, the vph1Δ growth defect is not changed (improved or worsened) in the presence of Stv1-mNeonGreen, so it means that the expression of Stv1-mNeonGreen does not further compromise the V-ATPase function, but it does not mean that it improves its function.

Re: Figure 7 - supplement 2. This figure should be combined with Fig. 6- suppl 1, panel D as also mentioned above. The figure seems to lack some labels, and conclusions are not accurate as discussed below. However, this data provides important additional information about relationships between isoform-specific subunits of V-ATPase Vph1 and Stv1 and both Rtc5 and Oxr1 and should be repeated if it is not done yet to have a better idea about these relationships. Panel B: Based on this picture, deletion of RTC5 has a negative genetic interaction with the deletion of VPH1, since double deletion mutant vph1Δ rtc5Δ grows worse than each individual mutant. Although it also means that there is no positive interaction, it is not the same. Panel C: Same as for panel B. Based on this picture, the deletion of OXR1 has a weak negative genetic interaction with the deletion of STV1, since double deletion mutant stv1Δ oxr1Δ grows worse than each individual mutant at 6 mM ZnCl2. In addition, there is no label for the media in the middle panel, is it just YPAD pH=7.5, without the addition of any metals? Why there is no growth assay in the presence of CaCl2, like in panels A and B? Panel D: Same as for panels B and C. Based on this picture, deletion of RTC5 has a negative genetic interaction with the deletion of STV1, since double deletion mutant stv1Δ rtc5Δ grows worse than each individual mutant at 6 mM ZnCl2. There is no label in the middle panel (growth conditions) and no growth assay data in the presence of CaCl2.

Re: Figure 7 - supplement 2, continued. How many times all these experiments were repeated? These experiments should be repeated at least 3 times, which is especially necessary for the experiments in panel C, because the effects are borderline. If results are reproducible and statistically significant, although small, the conclusion should be changed from "no positive genetic interactions" to "negative genetic interactions", which is more precise and informative. However, these results will be then in contradiction with the results from Figure 6 - Supplement 1, panel D, showing negative genetic interaction between the overexpression of Rtc5 or Oxr1 and deletion of Stv1, since both deletion and overexpression of Rtc5 or Oxr1 would have negative genetic interactions with Stv1. In addition, apparently, there is no data about genetic interaction between the overexpression of Rtc5 or Oxr1 and the deletion of Vph1. All this needs clarification, therefore repeating these experiments is essential. In conclusion, while genetic interactions between RTC5/OXR1 and RAV1 are straightforward, they seem to be more complex with STV1/VPH1.

Re: Methods. There is no description of yeast serial dilution growth assay at all. In addition, why the specific media (neutral pH, in the presence of high concentrations of calcium or zinc) was used is not explained either in the results or methods. Appropriate references should be included, for example, PMID: 2139726, PMID: 1491236.

Minor comments

Yeast proteins are named with "p" at the end, such as "Rtc5p".

Re: Introduction. In the introduction it should be indicated that Rtc5 was originally discovered as a "restriction of telomere capping 5", using screening of temperature-sensitive cdc13-1 mutants combined with the yeast gene deletion collection [PMID: 18845848]. A couple of sentences should be written about the RAVE complex and its role in V-ATPase assembly.

Re: The TLDc domain-containing protein of unknown function Rtc5 is a novel interactor of the vacuolar V-ATPase (results) 1) It is important to understand, that Oxr1 was co-purified before with the V1 domain of V-ATPase from a certain mutant strain, not wild-type yeast [PMID: 34918374]. It may explain why the authors did not identify it in their original protein-protein interactions screen here. 2) It is a wrong conclusion that because Rtc5 was co-purified with both V1 and V0 domain subunits it interacts with the assembled V-ATPase, this does not exclude a possibility that Rtc5 also interacts with separate V1 sector or separate V0 sector of V-ATPase.

Re: Figure 1, Panel C. Is it possible to show individual proteins in different colors for clarity? Panel D. How were cross-link distances measured? It is not obvious if you are not an expert in the field and it is not described in the methods.

Re: Figure 1 - Supplement 1, Panel A. What scientific information are we getting from this picture? Panel B. Why are these complexes shown separately from the complexes in Figure 1, panel C? Also, can individual proteins be colored differently here as well?

Re: Figure 3. It will be nice to show the localization of the untagged protein as well if antibodies are available (OPTIONAL).

Re: Figure 4. Why different tags were used in panels A (GFP), C (msGFP2) and D (mNeonGreen)? Panels B and C. Were Rtc5 fusions detected using anti-GFP antibodies? The authors should have full-size Western blots available, not just cut-out bands, as some journals and reviewers require them for publication.

Re: Figure 4 - Supplement 1, Panel A. Does "-" and "+" mean -/+ Azido-Myr? Panel B. There is no blot with a membrane protein marker (Vam3 or Vac8), it should be included.

Re: Figure 5. The title does not describe all results in this figure and should be modified accordingly. Panel C. Statistical significance value for *** should be indicated in the legend. It is not clear how many times yeast growth assays were repeated. Usually, all experiments should be done in triplicates or more.

Re: Figure 5 - supplement 1. No title

Re: Figure 5 - supplement 2. No title

Re: Figure 6. There is a typo on the second lane in the legend: "...the genome were", not "...the genome where". Panel C. Why the analysis of BCECF vacuole staining of double mutants oxr1Δrav1Δ and rtc5Δrav1Δ is not shown? Was it done at all?

Re: Figure 6 - Supplement 2. Why were two different tags (2xmNG and msGFP2) used? Did the authors study N-terminally tagged Oxr1? Was it functional? Panel B. Results for the untagged TEF1pr-Oxr1 overexpression are not shown, thus tagged and untagged proteins can't be compared. Are they available? What is the promoter for the expression of 2xmNG fusion constructs?

Re: Methods. Were vacuoles prepared differently for XL-MS and SILAC-based vacuole proteomics (there are different references) and why? Methods for XL-MS and quantitative SILAC-based proteomics can be placed together for clarity. What is CMAC dye? Why was it used to stain the vacuolar lumen? Some abbreviations (TEAB, ACN) are not explained. What is 0% Ficoll?

Referees cross-commenting

I agree with both reviewers, although I think that it is a pretty novel finding because while I was familiar with Oxr1 data I did not realize until now that there is a second protein in yeast. I think it is because homology between Oxr1 and Rtc5 is really low. I also agree that they should study more about what happens with V0 subunits.

Significance

Field of expertise keywords:

Protein-protein interactions, V-ATPase, TLDc

The vacuolar-type proton ATPase, V-ATPase, is the key proton pump, that hydrolases ATP and uses this energy to pump protons across membranes. Amazingly, this proton pump and its function are conserved in eukaryotes from yeast to mammals. While V-ATPase structure and function have been studied for more than 30 years in various organisms, its regulation is not completely understood. The very recent discoveries of two new V-ATPase interacting proteins in yeast, first Oxr1 (OXidative Resistance 1), and now Rtc5 (Restriction of Telomere Capping 5), both the only two members of TLDc (The Tre2/Bub2/Cdc16 (TBC), lysin motif (LysM), domain catalytic) proteins in yeast, provide new insights in V-ATPase regulation in yeast, and because the interaction is conserved in mammals its relevance to mammalian V-ATPases regulation as well.

TLDc proteins are best known for their role in protection from oxidative stress, in particular in yeast and in the nervous system in mammals. The discovery of the novel Rtc5-V-ATPase interaction points to the role of V-ATPase not only in protection from oxidative stress but also in restriction of telomere capping in yeast and most likely higher species. The studies of other species also highlight the possible conserved role of V-ATPase in lifespan determination and Torc1 signaling, mediated through these interactions. Thus, the discovery of this new functionally important interaction between the second TLDc family member in yeast, Rtc5, and V-ATPase will shed light on the molecular mechanisms of all these essential biological processes and pathways.

In addition, because the authors performed a comprehensive proteomics protein-protein interaction study of the purified yeast vacuole it provides a valuable resource for all researchers who study vacuoles and/or related to them lysosomes.

The follow-up functional studies using the rav1Δ strain clearly demonstrated that Rtc5 and Oxr1 disassemble V-ATPase and counteract the function of V-ATPase assembly RAVE complex in vivo in yeast. Thus, they are essentially the first discovered endogenous eukaryotic protein inhibitors of V-ATPase. Moreover, because the authors obtained the evidence that Oxr1 is the regulator of the specific subunit isoform of V-ATPase Stv1p in vivo in yeast, it suggests that different TLDc proteins may regulate different specific V-ATPase subunit isoforms in cell- and tissue-specific manner in higher eukaryotes. The mechanism of this isoform-specific regulation in yeast and other species needs further investigation in the future.

Because of the conservation of the TLDc-V-ATPase interactions, all this information can be extrapolated to higher species, all the way to humans, in whom genetic mutations in various TLDc proteins are known to cause devastating diseases and syndromes.

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

Using cross-linking proteomics, Klössel et al. identify the yeast TLDc domain protein Rtc5 as a novel interactor of yeast V-ATPase and characterize functions for Rtc5 and the TLDc domain protein Oxr1 in V-ATPase assembly and localization.

Major points:

1. The evidence of Oxr1 and Rtc5 as V-ATPase disassembly factors is circumstantial. The authors base their interpretation primarily on increased V1 (but not Vo) at purified vacuoles from Oxr1- or Rtc5-deleted strains, which does not directly address disassembly. Of course, the results regarding Oxr1 confirm detailed disassembly experiments with the purified protein complex (PMID 34918374), but on their own are open to other interpretations, e.g. suppression of V-ATPase assembly. Of note, the authors emphasize that they provide first evidence of the in vivo role of Oxr1, but monitor V1 recruitment with purified vacuoles and do not follow V-ATPase assembly in intact cells.
2. Oxr1 and Rtc5 have very low sequence similarity. It would be helpful if the authors provided more detail on the predicted structure of the putative TLDc domain of Rtc5 and its relationship to the V-ATPase - Oxr1 structure. Is Rtc5 more closely related to established TLDc domain proteins in other organisms?
3. The authors conclude vacuolar recruitment of Rtc5 depends on the assembled V-ATPase, based on deletion of different V1 and Vo domain subunits. However, these genetic manipulations likely cause a strong perturbation of vacuolar acidification; indeed, the images show drastically altered vacuolar morphology. To strengthen their conclusion, it would be helpful to show that Rtc5 recruitment is not blocked by inhibition of vacuolar acidification, and that conversely it is blocked by deletion of rav1.

Significance

This is an interesting paper that confirms and extends previous findings on TLDc domain proteins as a novel class of proteins that interact with and regulate the V-ATPase in eukaryotes. The title seems to exaggerate the findings a bit, as the authors do not investigate V-ATPase (dis)assembly directly and only phenotypically describe altered subcellular localization of the Golgi V-ATPase in Oxr1-deleted cells. A recent structural and biochemical characterization of Oxr1 as a V-ATPase disassembly factor (PMID 34918374) somewhat limits the novelty of the results, but the function of Oxr1 in regulating subcellular V-ATPase localization and the identification of a second potential TLDc domain protein in yeast provide relevant insights into V-ATPase regulation. This paper will be of interest to cell biologists and biochemists working on lysosomal biology, organelle proteomics and V-ATPase 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 #1

Evidence, reproducibility and clarity

Klössel et al. explore the role of the TLDc domain-containing proteins Oxr1p and Rtc5p in Saccharomyces cerevisiae. They performed cross-linking mass spectrometry and detected the interaction of Rtc5p with V-ATPase. TLDc domains have previously been found to serve as V-ATPase interacting domains. The authors find that both Oxr1p and Rtc5p induce dissociation of V-ATPase in vivo, an activity that was previously established for Oxr1p in vitro. They propose that this activity counteracts the activity of the V-ATPase assembling RAVE complex. They also find that Oxr1p is necessary for late Golgi retention of the Golgi form of the V-ATPase (i.e. containing the Stv1p isoform of subunit a). It is a little surprising that Oxr1p binding to V-ATPase was not detected by the cross-linking mass spectrometry, although the authors argue that this absence may be owing to the abundance of the proteins, which sounds reasonable.

Suggestions:

1. The authors observed that knockout of Rtc5p or Oxr1p does not affect vacuolar pH. If Rtc5p and Oxr1p both cooperate to dissociate V-ATPase, the authors may wish to characterize the effect of a ∆Rtc5p∆Oxr1p double knockout on vacuolar pH.
2. The manusript would benefit from a well-labelled diagram showing the subunits of V-ATPase (e.g. in Figure 2D).
3. The images of structures, especially in Figure 1-Supplement 1B, are not particularly clear and could be improved (e.g. by removing shadows or using transparency).
4. The authors should clearly describe the differences between Rtc5p and Oxr1p in terms of protein length, sequence identity, domain structure, etc.

Minor:

1. The "O" in VO should be capitalized.
2. In Figure 4 supplement 1, the labels "I", "S", and "P" should be defined.
3. Please clarify what is meant by "switched labelling"
4. The meaning of the sentence "Indeed, this was the case for both of them" is ambiguous.
5. For Figure 1-Supplement 1B it is hard to see the crosslink distances.
6. The statement "The effects of Oxr1 are greater than those caused by Rtc5" requires more context. Is there a way of quantifying this effect for the reader?
7. The phrase "negative genetic interaction" should be clarified.
8. In the sentence "Isogenic strains with the indicated modifications in the genome where spotted as serial dilutions in media with pH=5.5, pH=7.5 or pH=7.5 and containing 3 mM ZnCl2", "where" should be "were".
9. Figure 2D: the authors should consider re-coloring these models, as it is challenging to distinguish Rtc5p from the V-ATPase.

Significance

The vacuolar protein interaction map alone from this manuscript is a nice contribution to the literature. Experiments establishing colocalization of Rtc5p to the vacuole are convincing, as is dependence of this association on the presence of assembled V-ATPase. Similarly, experiments related to myristoylation are convincing. The observed mislocalization of V-ATPases that contain Stv1p to the vacuole (which is also known to occur when Vph1p has been knocked out) upon knockout of Oxr1p is also extremely interesting.

Overall, this is an interesting manuscript that contributes to our understand of TLDc proteins.

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

In this manuscript, Grossmann et al. present a new potential pathway that regulates PLK4 levels in cells mediated by the CRL4^DCAF1 E3 ubiquitin ligase complex (CUL4A/B-DDB1-DCAF1). PLK4 plays a crucial role in centriole assembly acting as a master regulator of the centriole biogenesis and thus contributes to centriole number control. Centriole numbers need to be tightly regulated as deviations could lead to aneuploidy and potentially cancer. At the onset of centriole assembly in G1/S, PLK4 is focusing into a single point on each parental centriole together with STIL and SAS6 defining the site of procentriole formation. For this process to happen, PLK4 trans-phosphorylates itself creating a binding site for SCF^β-TrCP E3 ubiquitin ligase that targets PLK4 for ubiquitination and degradation by the proteasome. The authors identified by co-IP and mass spectrometry the CRL4DCAF1 E3 ubiquitin ligase complex as a potential regulator of PLK4. They show that CRL4^DCAF1 E3 ubiquitin ligase complex binds to PLK4 and targets it for degradation. Furthermore, the authors present data where knockdown of DCAF1 leads to increased levels of PLK4 and centriole amplification. Using AlphaFold and followed by IPs with PLK4 point-mutants, they propose that DCAF1 binds to the dimer of PLK4 at PB1-PB2 at a similar site where Cep192/Cep152 bind. Then, they move on to show that CRL4^DCAF1 E3 ubiquitin ligase complex ubiquitylates PLK4 predominantly in G2 phase. Lastly, they propose that DCAF1 regulates the interaction of PLK4 with STIL and that it is required to prevent premature centriole disengagement in G2 phase. The manuscript is written in a clear and concise manner while the experimental approaches are sound and well described. The experimental data are well presented with a good number of replicates in most cases. However, some of the conclusions are drawn from marginal differences in the data and without statistical tests (cases indicated in detail below). I believe that this work is of interest to the scientific community, but it would require revisions to address the following major and minor comments.

Major comments:

• The key finding of the paper is that PLK4 co-IPs with DCAF1 and DDB1 that are core components of CUL4A/B E3 ubiquitin ligase. However, the only evidence that this interaction between PLK4 and DCAF1 is direct relies on the ubiquitylation assay performed in E. coli. This experiment was performed only once, and no quantitation is performed (Fig 4A). Given that the components are overexpressed in this heterologous system, it is very plausible to have a non-specific interaction between the DCAF1-Acidic domain and PLK4-PB1-PB2 considering that native binders to this region (Cep192/Cep152) are absent. The PLK4-DCAF1 model that was generated with AlphaFold suggests that this interaction is plausible but stronger verification that the interaction is direct are necessary in this reviewer's point of view. This could be performed for instance by purifying proteins (or fragments of them) to test binding in vitro, or through IPs of full-length proteins from bacterial extracts. If the interaction between PLK4 and DCAF1 is indeed direct, then the authors would need to provide an explanation of the feasibility of this interaction given that the binding site is occupied by Cep192 or Cep152 at the centrioles. Based on the current knowledge, PLK4 is loaded to the centriole through the interaction with Cep192 which is then switched to interaction with Cep152. For the DCAF1 to be able to bind to PLK4 it would need to outcompete Cep152. Thus, in order to prove that DCAF1 can control PLK4 at the centrosome, evidence would need to be provided that this interaction is possible. If that is not the case, then the most likely alternative is that DCAF1 interactions with cytoplasmic pool of PLK4, thus only indirectly controlling the PLK4 levels at centrioles. A plausible alternative interpretation of the data provided would be that DCAF1-Acidic domain could bind weakly and perhaps non-specifically to PLK4 but in human cells the interaction is mediated through another component such as Cep152 or Cep192 (which are also present in the MS data). Based on the AlphaFold model, the authors introduced point mutations that abolish the PLK4-DCAF1 interaction, but this effect could just as easily be an indirect effect due to abolishing of the PLK4-Cep152/Cep192 interaction.
• The authors state that DCAF1 depletion with siRNA or shRNA leads to increased level of PLK4 which triggers centriole overduplication. However, this statement is not entirely supported by the data provided. Firstly, in the western blots shown (Fig2A, 2D) the increase in PLK4 levels is hardly visible. Given that this is a key finding, stronger evidence would need to be provided. Furthermore, the quantification of the PLK4 levels upon siRNA mediated DCAF1 depletion are confusing as siDCAF1#2 leads to higher PLK4 levels than siDCAF1#1 despite being less effective in DCAF1 depletion (Fig 2A). More importantly, the quantification on the HeLa tet-on shDCAF1, that are used in many experiments, is missing an important statistical test (Fig 2D). Similarly, no statistical test is performed on the quantification of centriole numbers (Fig 2F) which puts to question the conclusion that "CRL4DCAF1 might function to keep PLK4 protein levels low, thus preventing centriole overduplication". Moreover, GFP-PLK4 levels shown in Fig 6A seem unaltered (if not lowered) upon DCAF1 depletion. Lastly, DCAF1 overexpression does not seem to decrease PLK4 levels as shown in Fig 6B. In that experiment, though, PLK4 is also overexpressed. In order to support the proposed function that CRL4DCAF1 keeps PLK4 levels low, it would be useful to also investigate whether overexpression of DCAF1 would lead to further decrease of PLK4 levels.
• In page 7, the authors mention: "A premature onset of centriole duplication in the absence of DCAF1 should also result in increased numbers of already disengaged centrioles in G2 phase." This premise is not correct as it is inverted to the current knowledge. It is the premature disengagement that licences for premature centriole duplication (or as often stated as re-duplication) rather than the premature onset of centriole duplication that causes disengagement. This is also what the authors correctly state in the discussion. In the data presented (Fig 6C) the authors observe centriole disengagement upon DCAF1 depletion using expansion microscopy, but no re-duplication is visible in the images provided. This is contrary to the overduplication claim made earlier on (Fig 2F). As such, the data presented do not fully support the drawn conclusion that DCAF1 controls PLK4 levels in G2 to prevent unscheduled centriole duplication. The authors would also need to investigate whether the prolonged use of Cdk1 inhibitor RO-3306 to synchronise the cells in G2 in addition to DCAF1 depletion contributes to the centriole disengagement that is observed, considering that Cdk1-Cyclin B acts also on PLK4-STIL complex.
• The mechanism proposed by the authors is that DCAF1 maintains PLK4 at low levels throughout G2 which prevents premature disengagement. Subsequently, low PLK4 levels prevent binding and activation of STIL impeding premature initiation of centriole duplication. However, this would not happen since centrioles remain engaged at this stage. Overall, some of the aspects of the proposed mechanism are not fully supported by the data presented. In addition, the proposed mechanism does not offer a suitable mechanistic explanation of how lower PLK4 levels by CRL4^DCAF1 mediated ubiquitylation and degradation prevent centriole disengagement.

Minor comments

• In Fig S2A authors need to indicate the expected size of the expressed protein. In its current form blot is difficult to be assessed. More specifically, it is unclear what is the result on the IP with the PLK4 fragment (1-879) since the more intense band in the input in not the same as in in the IP with Flag.
• In Fig 1C, S2B, S3B, it would be helpful to have a summary of the interactions observed next to each construct. This is commonly represented with (-, +, ++, +++) depending on the amounts present in the IP.
• In Fig 1D, even though not statistically significant, there seems to be a reduction in the IP of AA and PEST. Do the authors have some suggestion why that might be?
• Authors used two different cell lines in the experiments presented in Fig2A and Fig 2B. Given that depletion of siDCAF5 is provided as a control of having no effect in the PLK4 levels I would expect to have the experiment performed on same cell line.
• No statistical test is provided in the comparison on PLK4 levels upon siRNA treatment coupled with CHX (Fig 2C).
• In the quantification of the PLK4 levels at the centrosomes (Fig 2E), it is not specified whether a background subtraction step was performed prior to the normalisation to the untreated control.
• In the blot shown in Fig S3, no input is visible in the lane with expression of the Acidic domain.
• Authors claim that both WD40 and acidic domain contribute to binding of PLK4 because WD40-Acidic is more efficient in binding PLK4 that Acidic domain alone. However, in the blot provided, WD40 alone does not interact with PLK4. Thus, the most likely explanation would be that Acidic domain is the major interactor and WD40 has only minor contribution or it offers a stabilisation role to the acidic domain.
• Regarding the AlphaFold model provided, and in addition to the comments above, some further clarifications and controls would need to be provided. AlphaFold is a powerful tool but not without its caveats and needs to be used with caution. The authors need to provide a description on how they used AlphaFold to generate the model presented. Typically, AlphaFold produces 5 output models. At which site was DCAF1-Acidic domain positioned in the other output models? Based on what criteria the model shown was selected? Also, a confidence score for the model should be provided.
• The authors compare their PLK4-DCAF1 AlphaFold model with the structure of PLK4-CEP192 complex but not with the PLK4-Cep152. What is the explanation for this? Given that Cep152 is reported to have higher affinity than Cep192 (Park SY et al., 2014) it would be important to be included in the comparisons performed.
• The phrase "An overlay between the two structures revealed that ..." is not accurate as one is a merely a model. There are also other instances in the text that the model is referred to as 'structure' which is not correct.
• Please provide a citation for "Poisson-Boltzmann solver (APBS)".
• In Fig 3A ribbon representations are too small to see DCAF1 in the printout.
• The mutations designed might affect the folding of the PBs and thus no interaction is observed. Authors could test how the mutations would affect PB1-PB2 and also design one or two mutants that are in the vicinity but not in the interaction interface to serve as true negative controls in addition to the PLK4-WT. Do these mutants localise to centrioles or also the interaction with Cep192/Cep152 is affected?
• There is no statistical test in the quantification in Fig 3D, but it is not critical as the difference is very clear and certainly statistically significant.
• Authors state that DCAF1 strongly interacts with PLK4 during interphase but only weakly in mitosis with quantification in Fig 5A but there is no statistical test.
• Based on the data shown in Fig 5B, authors state that PLK4 is predominantly ubiquitylated by CRL4DCAF1 in G2 phase. However, in the blot shown, PLK4 seems to be in more abundance in G2 that might explain the apparent higher ubiquitylation. Furthermore, the experiment was performed once and no quantification of the ubiquitylation is performed. Lastly, there are no evidence that this apparent higher ubiquitylation in G2 is mediated by CRL4DCAF1.
• In Fig 6A, STIL levels upon DCAF1 depletion seem to be lower, is there any potential explanation for that? No statistical test is performed for the STIL/GFP-PLK4 levels difference in siGL2 versus siDCAF1. The authors should provide a justification for over-expressing PLK4 in this experiment. Similarly, in Fig 6B, the authors use overexpression of both PLK4 and DCAF1 and no statistical test is performed.
• Authors report in Fig 6C disengaged centrioles. How are disengaged centriole defined, is it based on a distance cut-off or loss of orthogonality? In the images provided, this reviewer's impression is that in the (+) Dox condition, there are two parental centrioles that have separated rather disengaged procentriole. Do the images come from the same cell?
• Based on the data presented, would overexpression of PLK4 in G2 would result in centriole disengagement? This is something that the authors would optionally check.
• The quantification of rootletin as an additional confirmation of centriole disengagement is puzzling to me as I would expect an increase rather than decrease of its levels. As centrioles disengage, a new link would need to form and thus the expected increase in its levels. However, new rootlet might form only later in mitosis. Also, given that the cells are synchronised in G2 the quantification is more complex. In late G2, centrioles separate in order to move to opposite poles to form the mitotic spindle. This would result in removal of the rootlet that might reflect the reduction the authors report. Ideally the quantification should be limited to cells in late G2 (that centrioles have separated) stained with Centrin 2 to allow for a quantification per centriole pair.
• In the discussion, authors state "It is conceivable that increasing amounts of PLK4 during mitosis, when the interaction between CRL4DCAF1 and PLK4 is weak, might capture STIL from binding to CDK1 initiating the interaction between PLK4 and STIL". In mitosis CDK1-Cyclin B binds to STIL and prevents formation of the PLK4-STIL complex, thus inhibiting untimely onset of centriole biogenesis (Zitouni et al., 2016). In addition, the authors show that total PLK4 levels are low in mitosis (Fig 5A). The conclusion drawn are not in line with the current literature.
• The addition of a graphical representation of the proposed mechanism would be beneficial to the readers.
• A reference for the Ac.Tubulin antibody used is missing.
• Please provide a citation for FiJi.

Referees cross-commenting

I find the comments by the other two reviewers to be valid, clear, insightful, and complementary to those made by this reviewer. There is a good convergence between the reviewers on the critical aspects in this manuscript that require attention. Following revisions this study will contribute to the understanding of regulatory mechanisms acting at the centrioles.

Significance

Centriole number control is an important aspect that is relevant not only to the centrosome research field but is also related to cilia, cells signaling, and cancer research. This work presents a novel pathway involved in the regulation of PLK4 levels in cells mediated by the CRL4^DCAF1 E3 ubiquitin ligase complex (CUL4A/B-DDB1-DCAF1). The authors present extensive data to characterise when and how DCAF1 interactions with PLK4 to lowers its levels through ubiquitination and subsequent degradation by the proteasome. However, the effects from various treatments are often minor. The study from Grossmann et al. comes to complement already known pathways of controlling centriole numbers, at G1/S through SCFβ-TrCP E3 ubiquitin ligase mediated PLK4 degradation, and in mitosis by CDK1-Cyclin B through STIL 'capturing' to block centriole reduplication. Given that certain aspects of the manuscript are revised, and an updated and more thorough mechanism is proposed and supported, it will contribute to the conceptual advancement or our understanding of centriole number control across the cell cycle. It could potentially also contribute to the ubiquitin research field of research, but it is hard for me to assess this as it is not my field of expertise.

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

Polo-like kinase 4 (Plk4) is the master regulator of centriole assembly and previous studies have shown that its level must be tightly regulated to ensure the precise duplication of centrioles during each cell cycle. It is well documented that the abundance of Plk4 is regulated by E3 ubiquitin ligases and in particular the SCF-TrCp ubiquitin ligase. However, in the absence of SCF-TrCp mediated regulation, PLk4 is still ubiquitylated suggesting that other ubiquitin ligases function to regulate Plk4 levels. Here Grossman and colleagues identify the CUL4-DDB1-DCAF1 (CRL4DCAF1) E3 ubiquitin ligase as a regulator of Plk4 levels and show that it functions predominantly in G2 phase to prevent centriole assembly in M phase. They propose a model whereby SCF-TrCp and CRL4DCAF1 cooperate to control the levels of PLK4 at different points in the cell cycle.

This study has the potential to yield some important and novel insights into the regulation of centriole assembly. However, in its current form the relationship between the increased Plk4 levels and the other effects described by the authors remain unclear. In particular, it is not clear to me that the small increase in Plk4 levels upon CRL4DCAF1 inhibition is responsible for the multipolar spindle phenotype. Nor is it clear how this increase in Plk4 is related to the premature disengagement defect. Finally, some of the experimental results could be made more convincing by including quantitation and/or additional controls. Major and minor issues are listed point-by-point below.

Major issues

Figure 2A and E. The authors report that depletion of DCAF-1 results in an increase in Plk4 levels. However, the actual increase is pretty small, about 1.5 fold for total levels (2A) and approximately 1.2 fold at centrosomes. How can the authors be sure this small increase in Plk4 levels is responsible for the multipolar spindle phenotype reported in figure 2F? It seems to me that CRL4DCAF1 could have other relevant substrates that are responsible for this defect. Related to this, can the authors show that the multipolar spindle phenotype is due to an overproduction of centrioles versus some other defect such as cytokinesis failure? Did the authors examine DCAF-1-depleted cells to if there are cell division defects that could explain the multipolar spindle defect?

Do the authors know if DCAF1 is operating within the context of the CRL4DCAF1 complex to control Plk4 levels? I know they showed that the entire complex is bound to Plk4 in pull down experiments, but have they tried to deplete other components of CRL4DCAF1 to see if they have the same effect on Plk4 levels?

Page 5 and Figure 3A. Th authors provide a model where the acidic domain of DCAF1 binds to a groove within the PB1-2 domain of Plk4. This is the same groove that binds CEP192, a protein that cooperates with Cep152 to recruit Plk4 to centrioles. Could it be that DCAF-1, at least in part, is competing with Cep192 and possibly cep152 for binding to Plk4? Thus, in the absence of DCAF1, Cep192 (and possibly Cep152) could recruit more Plk4. Can such a model be ruled out?

Figure 4. I don't find the results of the in vitro ubiquitin assays all that compelling. Here the authors are fusing DCAF1 to the E2 enzyme and show that this synthetic construct can ubiquitylate Plk4. I wonder in such a system if any protein could be ubiquitylated simply by tethering a binding domain for that protein to an E2 enzyme. So, I guess this is a question of specificity. Is there a control the authors can do to demonstrate specificity in this system?

Figure 5A and S5A. In figure 5A the authors use a flag-tagged Plk4 pulldown to show that DCAF1 strongly interacts with Plk4 during interphase and weakly during mitosis. In figure S5A, they perform the reverse experiment by pulling down endogenous DCAF1 and state that they obtained similar results. Looking at Figure S5A, this doesn't appear to be true. There is not much difference in the amount of Plk4 pulled down from interphase cells versus mitotic cells. The authors also do not indicate if any of the differences are significant.

Figure 5B. The authors investigate the cell-cycle-dependent pattern of Plk4 ubiquitination by co-expressing Flag-Plk4, HA-ubiquitin, and Myc-DCAF1 in HEk293 cells followed by a series of Flag IPs from cells arrested at different points in the cell cycle. They claim based on the retarded migration of Plk4, that CRL4DCAF1 ubiquitylates Plk4 specifically during G2 phase. It's hard to make any firm conclusions without quantitation. Furthermore, it's impossible to know how much of the ubiquitylation at any given cell cycle stage is dependent on DCAF1. The correct experiment would have been to have a no DCAF1 control for each cell cycle stage and to quantitate the differences. Since ubiquitin is tagged with HA, would it not be possible to probe the immunoprecipitate with an anti-HA antibody followed by quantitation.

Figures 6A and 6B. Why do the levels of Plk4 not respond to decreased or increased levels of DCAF1? In 6A for instance strong depletion of DCAF1 does not appear to affect the level of Plk4. Also, given that there is no change in Plk4 levels, the amount of STIL that is pulled down with PLK4 still increases upon DCAF1 knockdown. Does this mean that DCAF1 might function by directly inhibiting the Plk4-STIL interaction.

Figure 6C The authors find that upon DCAF1 knockdown, centrioles prematurely disengage during G2. They attribute this effect to the increased levels of Plk4. Is there any evidence that increased Plk4 levels lead to premature disengagement? Isn't it possible that this defect is independent of the increase in Plk4 protein? Either the authors should provide evidence of this or offer the possibility that the premature disengagement defect arises independently of the effect on Plk4 levels.

The authors should also consider exploring the possibility that CRL4DCAF1 functions semi-redundantly with the SCF. It would be interesting to see if there is a synergistic effect of knocking out both E3 ligases on Plk4 levels and centriole number. Such a finding would highlight the importance of the cooperative model the authors propose in this paper.

Minor issues

Page 5 typo: "in addition DCAF1 strongly binds to a WD40-acidic motif" I think you meant to say Plk4.

Figure 4. The terms l.e. and s.e. should be explicitly defined.

Many figures: Error bars are not defined. Do these represent SD or SE?

SCF-TrCp is not the only known E3 ligase that controls Plk4 levels. For instance Erich Nigg's group showed some time ago that the E3 ubiquitin ligase Mindbomb (Mib1) also regulates Plk4. (CAjanek et al 2015 J. Cell Sci. 128: 1674-82). This should also be mentioned in the introduction in order to paint a more complete picture of what is known about E3-based regulation of Plk4.

Significance

If my criticisms can be successfully addressed, this study has the potential to provide significant new insight into how centriole number is controlled. At least two E3 ligases have already been described that regulate Plk4 levels. This manuscript would provide a third. In an of itself, the discovery of a third E3 involved in the regulation of PLK4 levels would not have a major effect on the field. However if the authors can demonstrate how these two E3s are coordinated to control centriole assembly during the cell cycle that would be a great interest to those studying centriole assembly.

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 CUL4-DDB1-DCAF1 E3 ubiquitin ligase complex regulates PLK4 protein levels to prevent premature centriole duplication

Previously, it was thought that PLK4 is mainly regulated by autophosphorylation and degradation by the E3 ligase SCFSlimb/beta-TrCP in a phosphorylation dependent manner. In this manuscript by the Hoffmann group, the authors add an additional layer to the regulation of PLK4 as they identify the CRL4DCAF1 E3 ubiquitin ligase as a regulator of PLK4 that prevents PLK4 accumulation in G2 when beta-TrCP is low and therefore helps to restrict centrosome duplication to one event per cell cycle. More specifically, Grossmann et al. identified CRL4DCAF1 E3 ubiquitin ligase subunits in an immunoprecipitation mass spec approach. Using PLK4 kinase dead and phospho-mutants, they first show that CRL4DCAF1 binding is distinct from the SCFSlimb/beta-TrCP binding site. Depletion of DCAF1 leads to a modest increase in cellular PLK4 levels, PLK4 at centrosomes and cells with supernumerary centrosomes. Based on an IP experiments, they convincingly show that the acid C-terminus of DCAF1 interacts with PLK4 and they provide a model based on AlphaFold and analysis of mutations in the putative interaction interface how PLK4 and DCAF1 interact. They further provide evidences that DCAF1 directly ubiquitinates PLK4 in vitro. The interaction between DCAF1 and PLK4 is cell cycle dependent (peak in G2; Fig. 5) following the ubiquitination of PLK4. In Fig. 6 the authors analyze whether PLK4-STIL interaction is regulated by DCAF1. This is indeed the case and Fig. 6B likely indicates that DCAF1 functions as a competitive inhibitor for PLK4 and in this way blocks PLK4 binding to STIL. Finally, in Fig. 6C the authors analyze centrioles by expansion microscopy. The authors show mother-daughter centriole pair disengagement upon depletion of DCAF1 (on p. 7, bottom: "knockdown of DCAF1 leads to a significant higher number of disengaged centrioles"). This is similar to CEP57 depletion as shown by Kitagawa: JCB 2021 220: e202005153. Instead of analyzing centriole disengagement in further depth, the authors analyze in Fig. 6D centrosome separation, which is mechanistically quite distinct from centriole disengagement. Centrosome separation (mother-daughter pairs) in G2 is triggered by resolution of the rootletin linker through the action of the kinase Nek2A. Thus, Fig. 6 refers to two different events/mechanisms and it will be important to clarify whether DCFA1 depletion causes centriole disengagement or centrosome separation (e.g. by analyzing the centrin pattern and whether daughter centrioles mature). To my knowledge, there is no connection between PLK4/STIL and the centrosome linker. Thus, if DCFA1 regulates centrosome separation, Fig. 6 would be disconnected from the rest of the paper.

Main points

1. Fig. 2E: it would make sense to quantify the PLK4 signal at centrioles according to the cell cycle phase of the cell. G2 is probably the cell cycle phase when PLK4 is regulated by the CRL4DCAF1 E3 ubiquitin ligase.
2. It is known that PLK4 has a function in cytokinesis (i.e.: https://doi.org/10.1073/pnas.181882011). Thus, there is the possibility that the supernumerary centrosomes observed in Fig. 2F result from a cytokinesis defect and not from centriole over-duplication. To address this, the authors can use procentriole marker Sas6, and show that a newly disengaged centriole should still posseses Sas6. When the premature onset of centriole duplication happens to those newly disengaed centrioles, both mother and daughter centriole in the pair should posses Sas6 since Sas6 removal only happens in upcoming mitosis.
3. The authors suggest a competitive interaction between proteins DCAF1, PLK4 and STIL in Fig. 6A and Fig. 6B. However, they have not excluded direct binding of DCAF1 to STIL as an alternative explanation. Additionally, is the enhanced PLK4/STIL interaction in Fig. 6A G2 dependent?
4. The quality of the expansion microscopy in Fig. 6C could be improved.
5. The authors have to resolve whether Fig. 6C and D relate to centriole disengagement or centrosome separation and how this is connected to DCAF1, STIL and PLK4.

Minor points

1. P. 5: WD40-Acidic motif. This fragment needs to be described in the text and not just in Fig. S3.
2. Fig. 2E: The authors analyze the phenotype by combining all data points from three experiments. It would be better to show the average of the three independent experiments and do the statistics on the three data points.
3. Is the difference (> 4) in Fig. 2F significant?
4. Fig. 3A-C is difficult to follow. It is too small and DCAF1 and CEP192 are very difficult to see. I am sure that there are simple ways to improve this figure.
5. Define BP1 and BP2 in Fig. 3A. Does BP1 = PB1?
6. P. 5. "the first helix (D1420-E1436) of DCAF1 positioned .... (add DCAF1).
7. P. 20 Fig. 4B: 200 nm should be 200 nM.
8. The authors may want to test additional PLK4 mutations that are not localized in the predicted interaction interface with DCAF1 to show that these mutations do not affect binding.
9. In Fig. 4A the authors could IP GFP-PLK4 and show that a fraction of this protein carries His-Ubi conjugation using His antibodies.
10. Difference in quantification of Fig. 6D is not significant,
11. Fig. 4B (also Fig. 5B): Explain "l.e." and "s.e." in figure. Both blots are not the same (at least in case of Fig. 4B comparing the kDa numbers), thus l.e. = low exposure and s.e. = short exposure does not work. How was PLK4 detected in Fig. 4B?

Referees cross-commenting

I believe that all three reviewers have very similar concerns. I guess, everybody agrees that this manuscript, although potentialy very intresting, needs a substrainail amout of revision

Significance

The manuscript convincingly identifies CRL4DCAF1 E3 ubiquitin ligase as an PLK4 regulator and therefore is a very important contribution to the field. However, the impact of DCAF1 depletion is not too high. I therefore recommend double depletion of DCAF1 and SCFSlimb/beta-TrCP (not absolutely necessary but could increase impact). The interaction analysis of DCAF1 with PLK4 and the ubiquitination of PLK4 by the DCAF1 E3 ligase is convincing. I see a problems with the data in Fig. 6 that need to be revised.

Thus, key experiments that should be done are Main points 2), 3) and 5). The revision of the manuscript will take 3-6 months.

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)):

The manuscript develops the authors' previous work on the structure of the YeeE protein by presenting a co-structure with YeeD and investigating the role of certain key cysteine residues, especially C17 of YeeD. To this reviewer an entirely plausible mechanism for YeeD/E co-ordinated transport of thiosufate through the membrane and cleavage to sulfide and sulfite which are released into the cytoplasm is proposed on the basis of functional studies. The work is clearly described, the crystallography stats look good.

Thank you very much for your highly positive comments. We sincerely appreciate them.

Major comment: The 'cysteine relay' followed by a key role for C17 of YeeD in releasing a sulfide looks very plausible and makes the work of more general interest. An aspect that is not addressed is that of energetics. Moving thiosulfate into the cytoplasm as sulfide and sulfite means apparently that two negative charges net are generated in the cytoplasm for each thiosulfate taken up. This seems too simplistic (protons released as the bound sulfite is released b hydrolysis) but if thiosulfate were to be moved the whole way across there would be a divalent anion uniport which would work against the membrane potential negative inside (ie the main component of the protonmotive force). There is no mention in the paper of any pmf dependence and presumably the structure of YeeE shows no evidence of putative proton pathways? Some discussion of this and any wider implications could enhance the paper. In some ways the proposed transport scheme has some resemblance to Mitchells's old group translocation proposal for transport.

Thank you for highlighting the significance of the 'cysteine relay.' We also believe that this aspect is likely to interest a broad readership. Regarding protons, YeeE does not have apparent proton pathways inside, and we currently do not have data on its dependence on the pmf. Investigating pmf dependence falls beyond the scope of this study, hence we plan to explore this in future research. We appreciate you for pointing out that the YeeE-YeeD is a reminiscence of Mitchell’s original proposal of group translocation. This is a very intriguing point, and we have now included a discussion of this, along with a relevant citation, in the Discussion section (lines 356-357).

Reviewer #1 (Significance (Required)):

The subject of thiosulfate transport (movement) into bacteria is arguably of interest only to a narrow group of bacterial biochemists. However, the contents of this manuscript ought to be of wider interest because the YeeD/E system described is unusual in doing more that catalysing transport alone. Whether the authors' description in their title of 'sophisticated' is an appropriate adjective I am not sure. The term 'cofactor' applied to YeeD seems 'odd' to this reviewer. It is not a cofactor in the usual sense eg NADH.

We appreciate your comments. We have modified the title and avoided the unsuitable word 'cofactor' to describe YeeD.

reviewer's expertise: bactrial energetics but little knowledge of sulfur metabolism

---

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

Summary:

The publication "Structure and function of a YeeE-YeeD complex for sophisticated thiosulfate uptake" by Ikei et al. shows the protein-protein interaction of a thiosulfate transporter YeeE and a sulfur transferase YeeD, a TusA-family protein. The transporter YeeE has been structurally characterized previously, without showing its functional activity in a purified reconstituted system. This experiment complementing the previous publication is provided here, furthermore proving the functionality of the transporter. These experiments were further extended by the characterization of the cytoplasmic acceptor protein. This acceptor was proven to be YeeD, by structural characterization and biolayer interferometry. The binding kinetics between YeeD and YeeE were measured, quantifying the binding affinity between the two proteins. Furthermore, the surface residues of YeeD were specified by amino acid exchange mutants. Thus, the structure and essential residues were characterized protein. The interaction of sulfur transferase YeeD with the thiosulfate transporter YeeE is a novelty to the field. This illuminates the first time a specific function of YeeD in thiosulfate assimilation.

We appreciate your positive review and for recognizing the significance of our work in uncovering the functions of the YeeE and YeeD complex. We have addressed the following major and minor comments, thereby improving our manuscript. We appreciate the your constructive feedback.

Major comments:

I see the following major problem: The YeeD protein preparations used in the experiments contained several different protein species. Mass spectrometry showed the existence of the monomeric reduced protein, a TusA sulfinate and a TusA thiosulfonate. There is obviously an oxidation of cysteine to cysteine sulfinate, possibly due to the presence of oxygen as shown in Fig. 2D and stated in the text. The formation of sulfinates has to be avoided. This can be achieved by the use of stronger reducing agents or by purification under strict exlusion of oxygen. The formation of sulfenic, sulfinic and sulfonic acid on cysteines by oxidation has been reviewed by Ezraty et al 2017 Nat Rev Microbiol.

To answer these points, we have extensively several experiments and analyses, and modified the text. In the mass spectrometry analysis of purified StYeeD, three major peaks are observed (Fig. 2D), but they do not necessarily reflect actual relative abundances due to the nature of mass spectrometry analysis. Therefore, we also analyzed the purified StYeeD by non-reducing SDS-PAGE, which showed very few molecular species with S-S bonds, with over 90% existing as YeeD-SH (Fig. S2D). We considered this level of purity sufficient for conducting biochemical analyses. Furthermore, although a small amount of YeeD-SO2- was observed, this would be inactive and thus not impact the activity of StYeeD because a similar irreversible modification product, NEM-modified StYeeD(WT), was inactive (Fig. S2G).

We have also provided non-reducing SDS-PAGE results for each mutant StYeeD in Fig. S2F. All StYeeD mutants except for L45A showed a similar pattern to StYeeD(WT). Conducting experiments under anaerobic conditions is quite challenging in our laboratory facility, so we have displayed non-reducing SDS-PAGE profiles of all proteins used in order to avoid misunderstanding. We have also tried the purification in the presence of DTT, a stronger reducing agent, but the fraction of YeeD-SO2- was not significantly changed.

In the revised version, mass spectrometry analyses were reperformed using DTT-reduced YeeD, resulting in more precise data (Fig. 2D–H). Based on these results and your valuable comments, we have rewritten the paragraph entitled 'T____hiosulfate decomposition activity of YeeD and its catalytic center residue' to represent the reduction/oxidation forms accurately. We have also cited the Nat. Rev. Microbiol. review in the text (line 185).

In their in vitro assays, the authors use exceptionally high thiosulfate concentrations of 300 mM. This is so far from any physiologically relevant concentrations that strong doubt is shed the validity of any conclusions transferred from the in vitro to the in vivo situation.

In the revised version, the mass spectrometry analysis was reperformed with a thiosulfate concentration of 500 µM, which is the same concentration of thiosulfate used in the thiosulfate decomposition experiments. To clarify this, we have included the thiosulfate ion concentrations in the legend of Fig 2.

L247 and Fig5: The proposed mechanism cannot be true. Binding of thiosulfate to a reduced TusA protein is not possible without release of electrons. Where do these electrons go? In the proposed scheme, the number of electrons before and after the reaction steps is not equal (Fig. 5). A release of the sulfur atom between the cysteine sulfur atom and the oxidized sulfur atom is impossible.

Thank you for your insightful comments. We have revised Fig. 5B to represent a better model. However, elucidating the electron pathway falls outside the scope of this study, and we cannot offer a definitive explanation. We have addressed this limitation in the Discussion section and highlighted it as a topic for future research.

Have the authors checked whether TusA dimers are formed via disulfide bridges? If so, thiosulfate could resolve these disulfides leading to reduced TusA and thiosulfonated TusA (YeeD-S-S-YeeD + S2O32- → YeeD-S-S-SO3- + YeeD-S-).

It cannot be excluded that the YeeD-S-SO3- species is a result of removal of sulfite from the YeeD-S-S2O3- species (possibly by transfer to another YeeD molecule) resulting in YeeD-S-S- oxidized by molecular oxygen to YeeD-S-SO3-.

Upon answering to this comment, we have re-examined the gel filtration result using gel filtration markers. We found that a fraction of YeeD exists as dimers in solution, as shown in Fig. S2C. By performing non-reducing SDS-PAGE, it was shown that these YeeD dimers were not due to intermolecular disulfide bond (Fig. S2D). Following your valuable suggestion, we have introduced the possibility that YeeD can function as a dimer into our model, as presented in a box in Fig. 5B.

Sulfide may be formed by a reaction of YeeD-S- with S2O32- to YeeD-S-SO3- and S2- or reaction of YeeD-S-S- with S2O32- to YeeD-S-S2O3- and S2-. As there is the formation of sulfinic acid that prevents clear conclusions, I suggest repeating the experiments on thiosulfate decomposition under anaerobic conditions to clarify the reaction mechanism. Anoxic buffers and strong reducing agents may prevent chemical oxidation.

As described above, based on the non-reducing SDS-PAGE results (Fig. S2D), we believe that the low presence of oxidized species does not significantly affect our analysis. Moreover, the mass spectrometry analysis after DTT treatment yielded more precise results (Fig. 2D–H). As noted above, conducting experiments under anaerobic conditions is challenging in our facility, so we kindly request your understanding and consideration of the revisions made in this manuscript.

Minor comments:

In response to the minor comments, we have revised the manuscript.

L58 What is the nature of the binding of the thiosulfate ion during the transport via YeeE. Is it covalently bound? Please comment in the text.

In our previous study (Tanaka et al., Sci. Adv., 2020), we proposed that thiosulfate ions were transported via hydrogen bonds. Responding to your comment, we have included the explanation in the text and cited Tanaka et al., 2020 (lines 66-67).

L76-L77 Is there a publication on the functionality of the Corynebacterium YeeD-YeeE fusion? The term "cofactor" does not apply to YeeD, which is a 9-kDa protein.

Since the function of Corynebacterium YeeD-YeeE has not been reported, we have changed the sentence to "In some bacteria, such as Gram-positive Corynebacterium species, YeeE and YeeD are encoded as one polypeptide." We have also avoided the word "cofactor" in the revised text (lines 89-91).

L114 YeeD was probably accidentally lowercased here as Yeed

We have corrected this error (line 134).

L119 Please specify what the negative control consisted of.

We have elaborated on the conditions (lines 140-141).

L120-122 In Fig 2c, the mutations E19A, K21A, E26A, D31A, E32A and D38A are still shown, but an explanation or description of the results is missing. The reason for investigation of these mutations should be stated in the text.

We have added the requested mutation information (line 146).

L137 If thiosulfate was not added before the MALDI-TOF, where did the sulfonate S-SO3 originate from? Is this an artifact formed during the heterologous production or purification? Please comment on this possibility in the text.

We think that the -S-SO3- form arose during purification (Fig. 2D). The -S-SO3- form disappeared upon reduction by DTT (Fig. 2F). It is possible to consider it as an intermediate state in the catalytic cycle of YeeD. We commented on this in the section entitled "Thiosulfate decomposition activity of YeeD and its catalytic center residue."

L144 Please state in the text whether these experiments were performed under aerobic or anaerobic conditions. The sulfinic acid is likely a product of a spontaneous chemical reaction with molecular oxygen.

Thank you for your feedback. We have now included information about the aerobic conditions in the main text (line 166-167) and added comments regarding the mass spectrometry results at the end of the paragraph (lines 191-201).

L148 It should be stated in the text whether YeeD in Fig2G was reduced with DTT as in Fig 2F or non-reduced as in Fig. 2D before thiosulfate was added. Only the reduced YeeD can yield conclusive results on the loading with sulfur, as there is already a thiosulfonate bound to the protein after purification.

Thank you for pointing this out. For mass spectrometry analysis, data were re-obtained, and DTT-treated sample was used for the thiosulfate condition in this revised version. Furthermore, we performed mass spectrometry analysis for the hydrogen peroxide condition using DTT-treated sample. Figures were replaced with revised ones (Fig. 2D–H). The text in the section "Thiosulfat____e decomposition activity of YeeD and its catalytic center residue" was appropriately re-written. Detailed sample preparation is also described in MATERIALS AND METHODS section.

L154 The YeeD used for measurement of sulfide formation must be reduced before the experiments. It is not stated in the text if this is the case. Also, the release of sulfide requires electrons. It should be commented where these electrons originate from.

The sample in the purification process contains β-ME until just before the final column (gel filtration). As shown in Fig. S2D, more than 90% of the purified product is in a reduced state after gel filtration. For mass spectrometry analysis, data were re-obtained using DTT-treated samples, and the figures were replaced with new ones (Fig. 2D–H). Binding and activity measurements were conducted in the presence of β-ME. To avoid the confusion of the readers, the buffer conditions were included in the legends of both Fig. 2 and Fig. 4, along with the details in the MATERIALS AND METHODS section. Regarding electron origin, since the electron route remains unknown at this stage, we have added the explanation as a sentence in the Discussion section (lines 370-372).

L159-160 If the mutation of the non-conserved YeeD cysteine inhibits growth, can anything be said about its function?

Regarding the non-conserved Cys in EcYeeD, we added some sentences in the Discussion section (lines 393-397)

L214 Is it possible to provide the Kd and KD values for the mutant proteins?

The ka, kd and KD values the interactions between YeeE and YeeD proteins have been provided in Table 2. To provide these values for all the YeeD derivatives, the data was re-analyzed, and therefore, the value of the WT YeeD is slightly different from the original manuscript.

L229 Stating a need of YeeD for thiosulfate uptake by YeeE is somewhat misleading as thiosulfate was also imported into liposomes by YeeE alone. Maybe state that YeeD is a required component for growth when thiosulfate is imported via YeeE.

We have addressed the incorrect wording (lines 317-318).

Reviewer #2 (Significance (Required)):

The work of Ikei and colleagues significantly advances our understanding of thiosulfate import in Escherichia coli (E. coli) and prokaryotes in general. Sulfur metabolism as a field is generally considered to be underexplored, with a notable lack of biochemical and structural information on membrane transporters responsible for the movement of both inorganic and organic sulfur compounds. The mechanisms involved in sulfur transport are also relatively poorly understood.

The proteins of the TusA family in E. coli exhibit distinct functions, although the precise function has only been determined for the canonical and namesake protein TusA. The discovered genetic evidence and the interaction of YeeE and YeeD adds significantly to our understanding of sulfur transfer reactions.

The novelty of this reaction is of particular interest to researchers studying prokaryotic physiology, especially the synthesis of sulfur-containing cofactors such as coenzyme A (CoA), biotin, lipoate, thiamine, and iron-sulfur (FeS) clusters, as well as the biosynthesis of cysteine and methionine. In addition, recent findings related to the TusA family protein YeeD elucidate a novel mechanism for sulfur mobilization and transfer that will be of interest to researchers involved in the regulation of sulfur metabolism, sulfur dissimilation, and ecological studies focused on sulfur utilization. Thus, a wide range of studies could be influenced by this review.

Areas of expertise include dissimilatory sulfur oxidation, sulfur transfer reactions, and protein-protein interactions.

Thank you again for emphasizing the importance of our work. We also believe this study significantly advances the understanding of thiosulfate import in prokaryotes, shedding light on the underexplored field of sulfur metabolism. This has implications for various areas of study.

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

The manuscript "Structure and function of a YeeE-YeeD complex for sophisticated thiosulfate uptake" by Ikei et al., reports the enzymatic characterization, transport capability and concerted function of YeeE and YeeD. Moreover, the authors report the crystal structures of two mutant variants of the complex.

The present work fills an important gap in understanding thiosulfate uptake and the individual roles of the YeeE and YeeD proteins in this process. This Reviewer believes that the paper has the potential of becoming an important reference in the field. However, this Reviewer has two or three major comments, besides a couple of minor ones, that would like the authors to address.

We appreciate your valuable comments. We have addressed both major and minor comments in our revisions, improving our manuscript.

This Reviewer hypothesizes that some of the comments might derive from a poor understanding of the text, derived from the way the manuscript is written. So, this Reviewer urges the Authors to take these comments as positive feedback, and build on these to improve the manuscript (namely on English and grammar).

We have diligently revised the manuscript, addressing your major concerns related to sulfide terminology and explanations in crystal structure analysis as below. These revisions have enhanced clarity, and a native English speaker has reviewed and refined our text for language and grammar.

MAJOR CONCERNS

1. There is no clue on the title and, more importantly, on the Abstract, to which microorganism the Authors are reporting this work. Only later one we are introduced to Spirochaeta thermophila, but this information should be front and center (at least in the Abstract);

We recognize the importance of clearly indicating the microorganism in our work. In accordance with the comments, we have revised both the title and Abstract, ensuring that the species is clearly identified in the Abstract.

Also, in the Abstract, the Authors only mention the 2.6 A resolution structure, leaving behind the 3.34 A one. This becomes very confusing, especially once one gets to the Results section (more comments below);

We apologize for any confusion arising from the omission of the 3.34 A resolution structure in the Abstract. In the revised Abstract, we have included both the 2.60 A and 3.34 A resolution structures. As per your suggestion, we have also provided detailed information about the determination of these structures in the Results, minimizing potential confusion for readers (lines 217-233).

The Authors mention in line 137 and Fig. 2D that a "sulfonate" moiety is formed at C17. However, cysteine sulfonation is an irreversible process, so how would the enzyme recover from this modification to allow turnover of the mechanism?;

We apologize for the poorly written passage that led to confusion. This paragraph has been revised with the appropriate wording and a proper mention of the reduction and oxidation of the -SH group. We now use the appropriate terms, such as sulfinic acid (-S-O2-), sulfonic acid (-S-O3-), and perthiosulfonic acid (-S-SO3-) to describe the sulfur-related modification states. In contrast to sulfonic acid (-SO3-) formed by the oxidization of the cysteine residue that is an irreversible process, perthiosulfonic oxidization of cysteine residue (-S-SO3-) is a reversible process, as shown in (E. Doka et al., Sci Adv 6, eaax8358 (2020)). Therefore, the modified YeeD molecules should be able to recover to the original state.

If the "sulfonylation" reported in line 137 and Fig. 2D is not a sulfonylation of the cysteine (because the peak disappears upon reduction with DDT as visible in Fig. 2F), but rather a sulfonylation of the cysteine-persulfide version of C17, this was already reported previously and should be referenced [PDB ID 5LO9, Brito et al. (2016) J Biol Chem 291: 24804-24818];

Because there was a misleading statement, as replied above, we have rewritten this paragraph.

The perthiosulfonic acid (-S-SO3-) in Fig.2D is different from this -S-S2O3- in Brito et al., (2016), but consistent with Fig. 2G. This point is included in the text and the suggested paper has been cited, as requested. (lines 191-193)

Section "Crystal structure of the YeeE-YeeD complex" should be re-written. Not only it is confusing, but also undermines the tremendous amount of work done by the Authors. Please state clearle what was crystallized, how and why. Specify clearly the mutation introduced and complement Table 1 with this information;

Thank you for these comments. The determination of the structures was certainly challenging. We have restructured the first part of the section entitled "Crystal structure of the YeeE-YeeD complex". We have included a comprehensive explanation of the crystallization process and the construction of YeeE-YeeD. Additionally, we have updated Table 1 to provide more detailed information on the two structures.

Lines 403-407: are the crystallization conditions already cryo-protected or no cryo-protection was added before flash freezing? Please state clearly;

In response to your feedback, we have added the missing information in MATERIALS AND METHODS section.

Table 1:

• Is the multiplicity of PDB ID 8K1R correct? Is it really 321?? If so, is there any radiation damage to the crystal? If not, how?? Fine-fine-slicing during data collection, big crystals with elliptical data collection?? Pleas elaborate;

The multiplicity for PDB ID 8K1R is correct. We have provided detailed information on data collection in MATERIALS AND METHODS section.

• There are water molecules in the structure so please report number of atoms and B-factors for waters ("Solvent"), and ligands (e.g., thiosulfate, or others, if any), separately;

We have updated Table 1 to include the requested information.

• Please provide validation statistics for the structures, namely, rotamer outliers, clashscore and MolProbity score.

We have added the validation statistics to Table 1.

MINOR CONCERNS

1. Always reference paper and PDB ID for all structures. E.g., at line 181, only the paper is referenced;

We have ensured that all structures are properly referenced with both the paper and the corresponding PDB ID (lines 246, 250).

Remove "alpha" in line 199;

We have removed the "alpha" (line 268).

Add units to all concentrations. E.g., at lines 326 and 327, (w/V) and (V/V) are missing.

We have incorporated concentration units, (w/v) or (v/v), for percentages in the appropriate locations.

Reviewer #3 (Significance (Required)):

The scientific rationale is robust and the experimental approach is adequate and provide support to the conclusion drawn. However, there are some questions this Reviewer would like to see clarified, namely on the data collection and processing of PDB ID 8K1R.

We appreciate your feedback. These revisions enhance the clarity and accuracy of this 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

The manuscript "Structure and function of a YeeE-YeeD complex for sophisticated thiosulfate uptake" by Ikei et al., reports the enzymatic characterization, transport capability and concerted function of YeeE and YeeD. Moreover, the authors report the crystal structures of two mutant variants of the complex.

The present work fills an important gap in understanding thiosulfate uptake and the individual roles of the YeeE and YeeD proteins in this process. This Reviewer believes that the paper has the potential of becoming an important reference in the field. However, this Reviewer has two or three major comments, besides a couple of minor ones, that would like the authors to address. This Reviewer hypothesizes that some of the comments might derive from a poor understanding of the text, derived from the way the manuscript is written. So, this Reviewer urges the Authors to take these comments as positive feedback, and build on these to improve the manuscript (namely on English and grammar).

Major concerns

1. There is no clue on the title and, more importantly, on the Abstract, to which microorganism the Authors are reporting this work. Only later one we are introduced to Spirochaeta thermophila, but this information should be front and center (at least in the Abstract);
2. Also, in the Abstract, the Authors only mention the 2.6 A resolution structure, leaving behind the 3.34 A one. This becomes very confusing, especially once one gets to the Results section (more comments below);
3. The Authors mention in line 137 and Fig. 2D that a "sulfonate" moiety is formed at C17. However, cysteine sulfonation is an irreversible process, so how would the enzyme recover from this modification to allow turnover of the mechanism?;
4. If the "sulfonylation" reported in line 137 and Fig. 2D is not a sulfonylation of the cysteine (because the peak disappears upon reduction with DDT as visible in Fig. 2F), but rather a sulfonylation of the cysteine-persulfide version of C17, this was already reported previously and should be referenced [PDB ID 5LO9, Brito et al. (2016) J Biol Chem 291: 24804-24818];
5. Section "Crystal structure of the YeeE-YeeD complex" should be re-written. Not only it is confusing, but also undermines the tremendous amount of work done by the Authors. Please state clearle what was crystallized, how and why. Specify clearly the mutation introduced and complement Table 1 with this information;
6. Lines 403-407: are the crystallization conditions already cryo-protected or no cryo-protection was added before flash freezing? Please state clearly;
7. Table 1:

a. Is the multiplicity of PDB ID 8K1R correct? Is it really 321?? If so, is there any radiation damage to the crystal? If not, how?? Fine-fine-slicing during data collection, big crystals with elliptical data collection?? Pleas elaborate;

b. There are water molecules in the structure so please report number of atoms and B-factors for waters ("Solvent"), and ligands (e.g., thiosulfate, or others, if any), separately;

c. Please provide validation statistics for the structures, namely, rotamer outliers, clashscore and MolProbity score.

Minor concerns

1. Always reference paper and PDB ID for all structures. E.g., at line 181, only the paper is referenced;
2. Remove "alpha" in line 199;
3. Add units to all concentrations. E.g., at lines 326 and 327, (w/V) and (V/V) are missing.

Significance

The scientific rationale is robust and the experimental approach is adequate and provide support to the conclusion drawn. However, there are some questions this Reviewer would like to see clarified, namely on the data collection and processing of PDB ID 8K1R.

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 publication "Structure and function of a YeeE-YeeD complex for sophisticated thiosulfate uptake" by Ikei et al. shows the protein-protein interaction of a thiosulfate transporter YeeE and a sulfur transferase YeeD, a TusA-family protein. The transporter YeeE has been structurally characterized previously, without showing its functional activity in a purified reconstituted system. This experiment complementing the previous publication is provided here, furthermore proving the functionality of the transporter. These experiments were further extended by the characterization of the cytoplasmic acceptor protein. This acceptor was proven to be YeeD, by structural characterization and biolayer interferometry. The binding kinetics between YeeD and YeeE were measured, quantifying the binding affinity between the two proteins. Furthermore, the surface residues of YeeD were specified by amino acid exchange mutants. Thus, the structure and essential residues were characterized protein. The interaction of sulfur transferase YeeD with the thiosulfate transporter YeeE is a novelty to the field. This illuminates the first time a specific function of YeeD in thiosulfate assimilation.

Major comments:

I see the following major problem: The YeeD protein preparations used in the experiments contained several different protein species. Mass spectrometry showed the existence of the monomeric reduced protein, a TusA sulfinate and a TusA thiosulfonate. There is obviously an oxidation of cysteine to cysteine sulfinate, possibly due to the presence of oxygen as shown in Fig. 2D and stated in the text. The formation of sulfinates has to be avoided. This can be achieved by the use of stronger reducing agents or by purification under strict exlusion of oxygen. The formation of sulfenic, sulfinic and sulfonic acid on cysteines by oxidation has been reviewed by Ezraty et al 2017 Nat Rev Microbiol. In their in vitro assays, the authors use exceptionally high thiosulfate concentrations of 300 mM. This is so far from any physiologically relevant concentrations that strong doubt is shed the validity of any conclusions transferred from the in vitro to the in vivo situation. L247 and Fig5: The proposed mechanism cannot be true. Binding of thiosulfate to a reduced TusA protein is not possible without release of electrons. Where do these electrons go? In the proposed scheme, the number of electrons before and after the reaction steps is not equal (Fig. 5). A release of the sulfur atom between the cysteine sulfur atom and the oxidized sulfur atom is impossible. Have the authors checked whether TusA dimers are formed via disulfide bridges? If so, thiosulfate could resolve these disulfides leading to reduced TusA and thiosulfonated TusA (YeeD-S-S-YeeD + S2O32- → YeeD-S-S-SO3- + YeeD-S-). It cannot be excluded that the YeeD-S-SO3- species is a result of removal of sulfite from the YeeD-S-S2O3- species (possibly by transfer to another YeeD molecule) resulting in YeeD-S-S- oxidized by molecular oxygen to YeeD-S-SO3-. Sulfide may be formed by a reaction of YeeD-S- with S2O32- to YeeD-S-SO3- and S2- or reaction of YeeD-S-S- with S2O32- to YeeD-S-S2O3- and S2-. As there is the formation of sulfinic acid that prevents clear conclusions, I suggest repeating the experiments on thiosulfate decomposition under anaerobic conditions to clarify the reaction mechanism. Anoxic buffers and strong reducing agents may prevent chemical oxidation.

Minor comments:

L58 What is the nature of the binding of the thiosulfate ion during the transport via YeeE. Is it covalently bound? Please comment in the text.

L76-L77 Is there a publication on the functionality of the Corynebacterium YeeD-YeeE fusion? The term "cofactor" does not apply to YeeD, which is a 9-kDa protein.

L114 YeeD was probably accidentally lowercased here as Yeed

L119 Please specify what the negative control consisted of.

L120-122 In Fig 2c, the mutations E19A, K21A, E26A, D31A, E32A and D38A are still shown, but an explanation or description of the results is missing. The reason for investigation of these mutations should be stated in the text.

L137 If thiosulfate was not added before the MALDI-TOF, where did the sulfonate S-SO3 originate from? Is this an artifact formed during the heterologous production or purification? Please comment on this possibility in the text.

L144 Please state in the text whether these experiments were performed under aerobic or anaerobic conditions. The sulfinic acid is likely a product of a spontaneous chemical reaction with molecular oxygen.

L148 It should be stated in the text whether YeeD in Fig2G was reduced with DTT as in Fig 2F or non-reduced as in Fig. 2D before thiosulfate was added. Only the reduced YeeD can yield conclusive results on the loading with sulfur, as there is already a thiosulfonate bound to the protein after purification.

L154 The YeeD used for measurement of sulfide formation must be reduced before the experiments. It is not stated in the text if this is the case. Also, the release of sulfide requires electrons. It should be commented where these electrons originate from.

L159-160 If the mutation of the non-conserved YeeD cysteine inhibits growth, can anything be said about its function?

L214 Is it possible to provide the Kd and KD values for the mutant proteins?

L229 Stating a need of YeeD for thiosulfate uptake by YeeE is somewhat misleading as thiosulfate was also imported into liposomes by YeeE alone. Maybe state that YeeD is a required component for growth when thiosulfate is imported via YeeE.

Significance

The work of Ikei and colleagues significantly advances our understanding of thiosulfate import in Escherichia coli (E. coli) and prokaryotes in general. Sulfur metabolism as a field is generally considered to be underexplored, with a notable lack of biochemical and structural information on membrane transporters responsible for the movement of both inorganic and organic sulfur compounds. The mechanisms involved in sulfur transport are also relatively poorly understood.

The proteins of the TusA family in E. coli exhibit distinct functions, although the precise function has only been determined for the canonical and namesake protein TusA. The discovered genetic evidence and the interaction of YeeE and YeeD adds significantly to our understanding of sulfur transfer reactions. The novelty of this reaction is of particular interest to researchers studying prokaryotic physiology, especially the synthesis of sulfur-containing cofactors such as coenzyme A (CoA), biotin, lipoate, thiamine, and iron-sulfur (FeS) clusters, as well as the biosynthesis of cysteine and methionine. In addition, recent findings related to the TusA family protein YeeD elucidate a novel mechanism for sulfur mobilization and transfer that will be of interest to researchers involved in the regulation of sulfur metabolism, sulfur dissimilation, and ecological studies focused on sulfur utilization. Thus, a wide range of studies could be influenced by this review.

Areas of expertise include dissimilatory sulfur oxidation, sulfur transfer reactions, and protein-protein interactions.

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 develops the authors' previous work on the structure of the YeeE protein by presenting a co-structure with YeeD and investigating the role of certain key cysteine residues, especially C17 of YeeD. To this reviewer an entirely plausible mechanism for YeeD/E co-ordinated transport of thiosufate through the membrane and cleavage to sulfide and sulfite which are released into the cytoplasm is proposed on the basis of functional studies. The work is clearly described, the crystallography stats look good.

Major comment: The 'cysteine relay' followed by a key role for C17 of YeeD in releasing a sulfide looks very plausible and makes the work of more general interest. An aspect that is not addressed is that of energetics. Moving thiosulfate into the cytoplasm as sulfide and sulfite means apparently that two negative charges net are generated in the cytoplasm for each thiosulfate taken up. This seems too simplistic (protons released as the bound sulfite is released b hydrolysis) but if thiosulfate were to be moved the whole way across there would be a divalent anion uniport which would work against the membrane potential negative inside (ie the main component of the protonmotive force). There is no mention in the paper of any pmf dependence and presumably the structure of YeeE shows no evidence of putative proton pathways? Some discussion of this and any wider implications could enhance the paper. In some ways the proposed transport scheme has some resemblance to Mitchells's old group translocation proposal for transport.

Significance

The subject of thiosulfate transport (movement) into bacteria is arguably of interest only to a narrow group of bacterial biochemists. However, the contents of this manuscript ought to be of wider interest because the YeeD/E system described is unusual in doing more that catalysing transport alone. Whether the authors' description in their title of 'sophisticated' is an appropriate adjective I am not sure. The term 'cofactor' applied to YeeD seems 'odd' to this reviewer. It is not a cofactor in the usual sense eg NADH.

reviewer's expertise: bactrial energetics but little knowledge of sulfur metabolism

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

1. General Statements We thank the Editors and the Reviewers for their time and constructive criticism, which has allowed us to improve our manuscript. All of our responses are indicated in blue font. Revision Figures for the Reviewers are included just below the response. The line numbers given here refer to those in the revised manuscript, where we have marked the changes in red.
2. Description of the planned revisions If granted a full revision, we will experimentally address the following major points, which were raised by more than one Reviewer: ● Repeat experiment in Figure 4 C to assess statistical significance (Reviewer 1 and 3) ● Western blot analysis of HDV infected HLCs showing small and large delta antigens. We have already performed such an analysis on HLCs (see Revision Figure 2). In addition, we will perform a comparative analysis with common HDV infection models dHepaRG and Huh7-NTCP cells over time (Reviewers 2 and 3). ● Additional characterisation of the two HLC subpopulations at transcript and protein level (Reviewer 1 and 3). In addition, we planned to conduct the following experiments in response to the individual Reviewers: In response to Reviewer 1: We thank the Reviewer for their encouraging feedback on our model and for their helpful comments, allowing us to improve our manuscript. Figure 1: The observation of a denser subpopulation of hepatocytes more susceptible to HDV is interesting. Do you have more characterization of this cell subpopulation, by IFA, in term of hepatic maturation marker, known HDV host factors and particularly NTCP expression? We agree with the Reviewer that this is an interesting observation. We separated the two hepatocyte subpopulations to analyse the gene expression of the liver maturation markers NTCP and ALB by RT-qPCR (see Revision Figure 1A). Surprisingly, we found that the low-density population expressed higher levels of both ALB and NTCP, suggesting that they are more mature than the high-density population. In addition, we stained both markers by immunofluorescence and observed no apparent differences (see Revision Figures 1B & C). In contrast, the new host factor identified in our study, CD63, appeared to be more highly expressed in the high-density population compared to the low-density population (Fig. 6G). However, we cannot exclude the Revision Plan possibility that other factors play an additional role. As outlined in our response to Reviewer 3, we will separate the two populations and analyse the gene expression of other known HBV and HDV co-host factors to assess whether they play a role in addition to CD63 in conferring the higher susceptibility to HDV infection to the highly dense HLC population. Revision Figure 1: High-density HLCs population is not more mature than the low-density HLC population. (A) The low-density HLCs population was separated from the high-density HLC population by gentle dissociation. Total RNAs were isolated from both populations and Albumin and NTCP expression was analysed by RT-qPCR. (B & C) High-density HLCs (upper image) and low-density HLCs (bottom image) were stained with Albumin specific antibody. Shown are either images taken on an epifluorescence microscope (B) or single slices of confocal images acquired on a Airyscan confocal microscope (C). Fig 1B and C: Can a BLV control be included in the figure? Thank you for this suggestion, we will repeat the experiment for these panels and add BLV as control. Fig 1A-F: What is the overall level of NTCP between HLC, HepaRG, Huh7NTCP and HLCAAV- NTCP? Can NTCP and HDAg be stained simultaneously in your cells? This is an excellent question and we will compare the total NTCP levels between differentiated HepaRG, Huh7 NTCP, HLCs +/- AAV NTCP by Western blot analysis and immunofluorescence (IF) staining. Comparing NTCP expression in HLCs +/- AAV NTCP, we observed a strong upregulation of surface NTCP upon AAV transduction by IF staining (Figure 1D). Unfortunately, our initial attempts to simultaneously detect NTCP and HDAg were technically hampered. Since HDAg is mainly localised in the nuclei, we have to permeabilize the cells in a harsh manner, which interferes with the detection of membrane NTCP. The latter is further hampered by the availability of suitable anti-NCTP antibodies for IF staining. In our study, we used high doses of fluorescence-conjugated MyrB peptide to stain NTCP, but unfortunately it is very sensitive to the harsh permeabilization detergents mentioned above. However, since we have meanwhile optimised HDV infection, we will likewise try again to optimise the staining Revision Plan protocol. If we succeed, we will repeat the co-staining of NTCP and HDAg and include it in a revised manuscript. Figure 4: While the strategy is interesting, based on what has been previously shown for HCV in Wu et al., 2012, the lack of statistical data prevents the reader to really understand and see drastic difference in term of susceptibility to infection and level of expression of host genes. In panel C, is the difference between day 13 and 15 statistically significant? Same for panel D, day 17 vs 19?As a remark, day 19, the peak of susceptibility to HDV, seems to be also the peak of maturation, based on ALB RTqPCR (panel B). Thank you for this comment, and will perform another set of experiments allowing us to calculate statistical significance. The Reviewer correctly points out the correlation between HDV infection and hepatocyte maturity, which we find very intriguing. To identify potential host co- or restriction factors expressed in highly mature HLCs, we then performed the differential gene expression analysis (Figure 5). As shown in the new Figure 5A, GO analysis revealed that genes involved in pathways regulating viral entry into host cells were most significantly upregulated in mature HLCs and, as a probable consequence, they were more permissive to HDV infection. Indeed, among these factors, we identified CD63 as a novel host cofactor that renders mature HLCs susceptible to HDV infection (Figure 6). In response to Reviewer 2: We thank the Reviewer for their assessment of our study and for critically pointing out the increments over the previous study by Lange et al. We also appreciate their helpful suggestions, which allow us to improve the manuscript. The manuscript would benefit from a more detailed virological analysis, such as: •Determination of HDV genome and antigenome sequences and analysis of HDV editing. We thank the Reviewer for this comment. Accordingly, we will determine HDV genomes and antigenomes by Northern blot analysis and study HDV editing rates by sequencing in HDVinfected HLCs. •Analysis of HDV short and large antigens by western blot. We have already detected small and large HDAg in HDV-infected HLCs (see Revision Figure 2). To also satisfy Reviewer 3, we will additionally compare the S/L-HDAg ratios over time in HLCs, dHepaRGs, and Huh7-NTCP cells and include the results in a revised manuscript. Revision Figure 2: Detection of small and large delta antigen in HDV-infected HLCs. Mature HLCs were infected with HDV (MOI= 5 Int. Units/cell) and harvested 1 or 3 days post-infection. Cell lysates were analysed by Western blotting using antibodies against HDAg and b-actin. Revision Plan •Analysis of HBV-related virological parameters in monoinfected and co-infected cells. We agree with the Reviewer and we will include the characterisation of more HBV-related virological parameters in our mono- and co-infected HLCs. Accordingly, we will assess HBV cccDNA, RNA, and DNA by RT-qPCR, as well as released HBsAg and HBeAg via ELISA and add the results to the revised manuscript. In response to Reviewer 3: We thank the Reviewer for their positive evaluation, and we acknowledge their helpful comments, which will help us to improve our manuscript. Line 143: the authors describe two forms of HLCs (less and more confluent with differences regarding the susceptibility to HDV infection). The characteristics of the less and more confluent HLCs should be described in more detail-what is causative for the differences in susceptibility for HDV infection of these two forms? We thank the Reviewer for this comment. We likewise find this observation intriguing. As stated in our response to Reviewer 1, we have ruled out that NTCP and/or other mature markers such as ALB are differentially expressed between the two subpopulations. As one factor that could make a difference, we have identified CD63, which is highly expressed in the high-density HLC population and less so in the low-density HLC population (Figure 6G). Nevertheless, we will separate the two populations and analyse by RT-qPCR the expression of other known HBV and HDV host co-factors that may be additional factors governing the increased susceptibility of the highly dense HLC population. The statistical analyses should be improved: There are no p-values provided for the data presented in the supplement and a variety of figures lacks p-values We have added p-values to the Supplementary Figures (see revised Supplementary Fig. S2) and will repeat the experiments for Fig. 4 and Supplementary Fig. S1B and Fig S3 so that we can calculate the corresponding p-values. Kinetic of the infection: Here it would be interesting to see a comparative analysis by western blot investigating the ratio HBsAg/HDAg over the time in HLCs, HepaRGs and NTCP oe cells We thank the Reviewer for his comments. As stated in our response to Reviewer 2, we will perform this WB analysis to detect S/L-HDAg over time in infected HLCs, dHepaRG, and Huh7- NTCP cells. Line 157: What is the experimental evidence for the proper localization and functionality of the ectopically expressed NTCP in HLCs. Did the authors study the taurocholate transport after overexpression of NTCP? We thank the Reviewer for this comment. We analysed endogenous and ectopic NTCP expression by microscopy using a fluorescently conjugated peptide Atto-MyrB-565, which specifically binds to the ectodomain of human NTCP (Figure 2D) and found that both Revision Plan endogenously and ectopically expressed NTCP are located on the cell surface. To further confirm the correct localisation, we will perform NTCP co-staining with a cell membrane marker. We will also test the proper function of the ectopically expressed NTCP using a specific taurocholate transport assay as shown in our previous study (Ni et al, 2014, Gastroenterology). Line 169: The authors should include data comparing the number of double positive cells in HLCs, HepaRGs and Huh7NTCP o.e. expressing cells under the chosen experimental conditions We thank the Reviewer for this suggestion. We have already performed HBV/HDV co-infection of dHepaRG cells (Revision Figure 3) and we will perform the same experiment with Huh7-NTCP cells. Revision Figure 3: HBV/HDV co-infection of dHepaRG cells. Differentiated HepaRG were infected with HBV (MOI = 450 genome copies/cell) and HDV (MOI = 5). Cells were stained against HBV core (HBc), HDAg, and nuclei (DAPI) ten days p.i.. HBc- and HDAg-positive cells were counted using Cellprofiler imaging software to quantify HBV (pink) and HDV (green) single and co-infection (white) events. Images are representative of three independent differentiations. Line 291: expression analysis by RT-PCR is not sufficient. It will be important to study by CLSM if the identified factors are really present as proteins and properly localized. To satisfy this Reviewer, we will be happy to perform WB analysis of lysates from cells obtained at different stages of HLC differentiation to detect LDLR, LAMP1 and SR-B1 to further confirm our transcriptome analysis. As protein expression is easier to compare by WB analysis, we prefer this method to microscopic analysis. Regarding the role of CD63: what is the evidence for a direct role of CD63 for HDV entrycan the authors exclude that CD63 is relevant for targeting other factors to the surface? What is the impact of loss of CD63 on the functionality of the autophagosomal-MVB-EV system in HLCs? Since downregulation of CD63 before but not after impairs HDV infection, we conclude that CD63 is likely to be important for the early steps of the HDV life cycle, namely cell entry. Indeed, we speculate that CD63 may be critical for HDV trafficking to the vesicle, where fusion of the HBV glycoproteins is induced to allow capsid entry, based on the following observations: Although neither the precise site of HBV viral fusion nor the cues that induce fusion are currently fully understood, studies suggest that HBV can be co-transported with EGFR and NTCP to late endosomes for trafficking (Herrscher et al; 2020, Cells). We speculate that similar to what has been described for Lujo virus, CD63 may be involved in either HDV trafficking and/or virus fusion in the endosomal system (late endosome or lysosome) (Tominaga et al, 2014 Molecular cancer). Revision Plan CD63 is a ubiquitously expressed protein that localises to the endosomal system and, in its glycosylated form, to the cell surface. Non-glycosylated CD63 is not properly trafficked and aggregates at the nuclear periphery instead of the cell membrane (Tominaga et al., 2014, Molecular Cancer). According to the Western blot analysis in Figure 6, immature HLCs appear to express less glycosylated CD63 than mature HLCs. We will confirm the glycosylation by treating the cell lysates with PNGase F. Although AAV transduction enhanced CD63 expression of all three HLC stages tested (see new Supplementary Figure S6 in the revised manuscript), it only enhanced HDV infection of immature HLCs, in which the non-glycosylated form of CD63 appears to be the predominant form. To demonstrate that the glycosylated form of CD63 is involved in HDV entry, we will rescue WT CD63 in parallel with a glycosylation-deficient CD63 mutant (Yoshida et al., 2009, Microbiology and Immunology) in immature HLCs. We will also stain CD63 in both immature and mature HLCs to compare the subcellular localisation (plasma membrane/endosomes vs. nuclear membrane) of CD63 between the two stages.
3. Description of the revisions that have already been incorporated in the transferred manuscript Based on the constructive comments by the Reviewers we already made the following changes, which are highlighted in red in the revised manuscript. In response to Reviewer 1: Fig 1B-C: the comparison with dHepaRG is very interesting, and confirms the validity of SC derived hepatocytes as a model for HDV infection. dHepaRG can be heterogeneous. Do you also see the same phenotype of enriched HDV infection within a denser subpopulations of dHepaRG We thank the Reviewer for their comment. Undifferentiated bipotent HepaRG cells are not permissive for HDV infection due to the lack of surface NTCP expression. Due to their bipotent nature and upon differentiation, two morphologically distinct populations become apparent: hepatocyte-like cells and biliary epithelial-like cells (McGill et al., 2010, Hepatology). As shown in the Figure 1 of the study by Mesnage et al. (2018, Molecular Toxicology), dense hepatocyte-like colonies are surrounded by clear epithelial cells corresponding to primitive biliary cells. In agreement with other studies, we only observe that the ALB-positive hepatocyte-like cells are permissive to HBV and HDV infection (Hantz et al., 2009, Journal of General Virology), highlighting their specific hepatic tropism and the cellular determinants required. Fig 1I is confusing. Was BLV assay also performed on the HLC infection (Day 0), or only during the titration assay in Huh7NTCP? We apologise for the confusion in this panel. BLV was only added during the titration assay on Huh7NTCP cells to confirm new and productive infections and to rule out carry-over. We have changed the order of Figures 1I - 1K to make this clearer and explain this better in the new results section (line 171-179) and figure legend (line 797-806). Revision Plan Fig 1K: x-axis is confusing... is it number of HBV, HDV and HBV/HDV positive cells? Or number of infected cells upon inoculation with HBV, HDV, or both? Please clarify. We apologise for this additional confusion caused in this panel. We infected HLCs with both HBV and HDV simultaneously and then counted the number of positive cells that were either single infected with HBV (pink cells/column), single infected with HDV (green cells/column) or double infected with both viruses (white cells/column). We have clarified this in the revised Results section (line 172-176) and in the revised Figure Legend (line 798-803). Figure 2: The AAV based vector to over express HBsAg is a very interesting tool, and the figure convincingly show production of HDV progeny viruses in HLC-AAV-HBsAg. Results shown are in agreement with previous studies based on hepatoma cell lines. We thank the Reviewer for this positive comment and we agree that AAVs represent interesting tools to genetically manipulate HLCs and other hepatocyte culture systems. Figure 2B: What is IU/ml? Infectious Unit? International Unit? Are units in Fig 1B, 2B and 2C the same? We apologise for the lack of clarity. In Figures 1B and 2C, IU corresponds to infectious units of HDV, whereas in Figure 2B, IU corresponds to international units for the assessment of secreted HBSAg levels in the supernatant. To make the difference clearer, we have changed the unit on the y-axis in Figure 2B and explicitly stated the abbreviations in the corresponding revised Figure Legends (lines 785, 786, 794, 795, 816, and 819). Figure 3: What is the overall number of transmission events observed in the co-culture setup? Can you visually observed viral spreading? Panel A shows only 1 event, making it hard to assess its efficiency. Titration assay in Fig 2C show production of up to 4-5 log of infectious HDV. But HLCs susceptibility to HDV infection may change during time... Thank you for your comment and for raising this important issue. Panel A clearly and visually demonstrates that extracellular spread of HDV had occurred in the HLCs system, as initially only WT and non-GFP positive HLCs were infected with HDV. After co-culture, the progeny of WT HLCs were able to infect GFP-HLCs (Figure 3A). The overall efficiency of HDV spread/transmission in HLC efficiency is shown in Figure 3C. If we allow spread to occur (DMSOtreated condition), the total number of HDV-positive HLCs grown in a 24-well plate is approximately 1000. When we block secondary infection of progeny with BLV and thus spread, we count only about 500 HDV-positive HLCs in a well. In general, spreading in HLCs (Figure 3C) is not as efficient as retitration to Huh7-NTCP (Figure 2C) for the following reasons: In Figure 2C, we wanted to have an estimation of the maximum amount of secreted infectious progeny from HDV-producing HLCs. To this end, we did not want the re-infection itself to be a major bottleneck and used the most susceptible model Huh7-NTCP and infected them under the best conditions, which includes the addition of 4% PEG and 2% DMSO in the culture medium. For our spread assay in HLCs, we cannot add PEG to the cells over the course of the experiment and we also wanted to be as physiological as possible. PEG significantly enhances HDV infection Revision Plan of HLCs (Supplementary Fig. S2) and Huh7-NTCP cells (Revision Figure 4), which is in agreement with previous studies (Michailidis et al., 2017, Scientific reports). In addition, as the Reviewer correctly points out, similar to other primary hepatocyte culture models, the HLC system deteriorates over time. However, we have found that HLCs can be cultured for up to 3 weeks. Nevertheless, we believe that the efficiency of HDV spread in HLCs is sufficient for drug testing (Fig. 3C & D). Revision Figure 4: PEG enhances HDV infection of Huh7-NTCP cells. Huh7- NTCP cells were infected with HDV (MOI= 5 Int. Units/cell) in the absence or presence of PEG. Cells were harvested on D5 pi and HDV genome copies were quantified by RT-qPCR. Figure 5: In panel A, GO pathways should be sorted based on significance, not Number of genes. In panel B-D, what is the scale of the heatmap on figure 5: change in CPM values, however log2, log10? Thank you for this comment, we have sorted the GO pathways based on significance (new Figure 5A). For panels B-D, we did not calculate the fold change in CPM values and they were not log transformed. Instead, we calculated the z-scores of the genes shown by comparing the expression level of a given gene (in CPM) in a given sample with the expression level of that gene across all samples. To avoid further confusion, we have added "z-score" to the new Figure 5. Figure 6: Do you have info about CD63 in other mature model, like dHepaRG and PHHs? Is CD63 also limiting in these models? Our data in Figure 6 suggest that CD63 may be a limiting factor for HDV infection of immature HLCs but not mature HLCs. Both dHepaRG cells and PHHs are mature hepatocyte models and therefore we speculate that CD63 is not rate limiting. However, we will investigate whether CD63 is rate-limiting in undifferentiated HepaRG cells. In response to Reviewer 2: Additional information that needs to be added, better explained, or corrected: The authors should explain why they used different MOIs depending on the genotype. In our previous study by Wang et al. 2021 J Hepatol, we found that the different HDV genotypes are heterogeneous in their ability to infect Huh7 NCTP cells. For example, as shown in Figure 4B of Wang et al. 2021 J Hepatol, GT 4 and 5 are less infectious than other genotypes. Based on the different infectious titres of the genotypes obtained on Huh7 NTCP cells, we then decided to use different MOIs for infection of our HLCs. The aim of the present study by Chi et al. was not to Revision Plan compare the different HDV genotypes, but to analyse whether they can all infect HLCs. In order to obtain similar infection efficiencies of our HLCs with the different genotypes, we used higher MOIs for those genotypes that were less infectious in Huh7-NTCP cells compared to those genotypes that were more infectious in Huh7-NCTP cells. We apologise for not making this sufficiently clear and have added this information to the results section (line 167-170) and the corresponding figure legend (line 796) of the revised manuscript. In Figure 1, it is unclear on which day the HCLs were infected by HDV and on which day they were transduced with AAV-NTCP. We apologise for the lack of clarity in the experimental design. We transduced HLCs with AAV two days before HDV infection to ensure sufficient ectopic NTCP expression on the day of HDV infection to study its effect on HDV entry. We have clarified this in the results section (line 153, 156) and in the figure legend (line 788) in the revised manuscript. It is not very clear if the authors used AAV serotype 6 consistently to transduce the cells. It would be valuable to show the transduction efficiency of AAV at different time points of HLC maturation, as it might also be affected and could explain some results. For example, in Figure 6H, why does AAV-CD63 transduction increase HDV infectivity at day15 but not at day 10? It would be interesting to repeat the anti-CD63 western blot after AAV-CD63 transduction. Thank you for this comment. Yes, we have consistently used AAV 6 due to its relatively broad tissue tropism (Verdera et al., 2020, Molecular Therapy) and we have clarified this information in the revised manuscript (see line 331). We agree with the Reviewer's concerns regarding the variable transduction efficiency. We have previously tested different AAV capsids and found that AAV6 transduced mature HLCs at high levels (Zhang et al., 2022, Hepatol Commun). In this study, we also performed Western blot analysis to confirm successful CD63 overexpression by AAV transduction at different stages of hepatocyte differentiation. As shown in new Supplementare Figure 6, although there were some differences in transduction efficiency, the majority of all cells at each stage of differentiation were successfully transduced to ectopically express CD63. The authors claim that by using AAV to express HBsAg, they are mimicking the expression of HBsAg from the integrated sequence rather than cccDNA. However, it is the opposite, as AAV genomes, like cccDNA, remain as episomes in the cells. Yes, the Reviewer is conceptually correct and we apologise for the incorrect wording. In principle, we aim to trans-complement HBsAg in a setting outside of HBV infection and thus mimic the expression of antigen from integrated cells, although AAVs of course remain mostly episomal. We have clarified this in the revised manuscript (see lines 188 & 378). In response to Reviewer 3: Line 217: the complete inhibition of cell to cell spread by myrcludex suggests that there is no spread by cell-cell contact. This should be discussed. Revision Plan Yes, there is no evidence of HDV spread by cell-cell contact because, as the Reviewer correctly points out, BLV treatment almost completely blocked HDV de novo infection (Figure 2D & E). To our knowledge, cell-to-cell spread has not been demonstrated for HDV. According to our own studies by Zhang et al, 2021/2022, Journal of Hepatology, HDV spreads either extracellularly (which can be blocked by BLV) or by cell division (discussed in lines 362). Since HLC are similar to primary human hepatocytes and do not divide in vitro, we believe that extracellular spread is the predominant mode of spread in HLC (line 365). Line 210ff:Is there any evidence for syncytia formation in this system? No, we have not observed syncytia formation. Since HDV has no glycoproteins, we would not expect syncytia to form. Line 42: secrete should be replaced by release We thank the Reviewer for pointing out the inaccuracy in our terminology. We have replaced "secrete" with "release" (line 42). Line 241: proteins are not expressed, genes are expressed Thank you, we agree and changed the wording accordingly (line 246).
4. Description of analyses that authors prefer not to carry out In response to Reviewer 1: Fig 1B: Unit is confusing, using terms usually used for titration of infectivity, from the virus input point of view, not from the cellular point of view. Can you use % infected cells instead, or "HDV infection rate" like in Supp Fig 1B? We apologise for this confusion. For other viruses, such as but not limited to HCV or HEV, the most common method is to report focus forming units per ml (FFU/ml). HLCs do not divide and, in the absence of HBV S antigen, no cell-division mediated HDV spread can occur and only single infection events can be observed (hence infectious unit = IU/ml). Since differentiated, authentic hepatocyte culture models such as PHHs, HLCs or HepaRG cells are always characterised by strong cell heterogeneity, it is difficult to directly compare the overall percentage of infection with a homogeneous cell population such as Huh7-NTCP cells. Therefore, if the Reviewer allows us, we prefer to keep this unit in our main figures. However, and hopefully to the satisfaction of this Reviewer, we have also calculated the percentage of infected cells of this exact dataset and show it in the Supplementary figures (Suppl. Fig. S1 C). The proportion of infection efficiency comparing HLCs, dHepaRGs, and Huh7-NTCP cells does not differ when presented either as IU/ml or as percentage of infected cells.

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 their manuscript entitled " An HBV/HDV Infection Model Using Human Pluripotent Stem 1 Cell-Derived Hepatocyte- Like Cells for Virus Host Interactions and Antiviral Evaluation" describe the use of HLCs derived from hPSCs as infection model for analysis of HDV life cycle. The ms is well written and clearly structured. It is easy to follow the concept of the study. The ms addresses a relevant topic and could help to overcome limitations in the analysis of HDV life cycle. The authors perform in many points a detailed characterization of this experimental system but there are still a variety of open points which must be addressed:

Specific points:

Line 143: the authors describe two forms of HLCs (less and more confluent with differences regarding the susceptibility to HDV infection). The characteristics of the less and more confluent HLCs should be described in more detail-what is causative fir the differences in susceptibitly for HDV infection of these two forms? The statistical analyses should be improved: There are no p-values provided for the data presented in the supplement and a variety of figures lacks p-values Kinetic of the infection: Here it would be interesting to see a comparative analysis by western blot investigating the ratio HBsAg/HDAg over the time in HLCs, HepaRGs and NTCP oe cells

Line 157: What is the experimental evidence for the proper localization and functionality of the ectopically expressed NTCP in HLCs. Did the authors study the taurocholate transport after overexpression of NTCP?

Line 169: The authors should include data comparing the number of double positive cells in HLs, HepaRGs and NTCP o.e. expressing cells under the chosen experimental conditions

Line 217: the complete inhibition of cell to cell spread by myrcludex suggests that there is no spread by cell-cell contact. This should be discussed.

Line 210ff:Is there any evidence for syncytia formation in this system?

Line 291: expression analysis by RT-PCR is not sufficient. It will be important to study by CLSM if the identified factors are really present as proteins and properly localized. Regarding the role of CD63: what is the evidence for a direct role of CD63 for HDV entry-can the authors exclude that CD63 is relevant for targeting other factors to the surface? What is the impact of loss of CD63 on the functionality of the autophagosomal-MVB-EV system in HLCs?

Minor points:

Line 42: secrete should be replaced by release Line 241: proteins are not expressed, genes are expressed

Significance

The manuscript describes the use of HLCs derived from hPSCs as infection model for analysis of HDV life cycle. The ms is well written and clearly structured. It is easy to follow the concept of the study.

The ms addresses a relevant topic and could help to overcome limitations in the analysis of HDV life cycle. The authors perform in many points a detailed characterization of this experimental system but there are still a variety of open points which must be addressed.

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 present work, Chi et al. demonstrated that Hepatocyte-like cells (HCLs) derived from human pluripotent cells (hPCs) can be infected by HBV. The development of new HDV cellular models is of great value for understanding HDV biology and developing new treatments. However, the relevance of the present work is limited by a recent publication by Lange et al., in which they also showed that HCLs derived from hPCs can be infected by HDV, inducing the activation of the innate immune response, as previously demonstrated in cells and mice.

The authors added new information to the work of Lange et al, including:

• HLCs derived from human pluripotent cells can be infected by different HDV genotypes.
• They proved that infectious HDV particles are formed.
• They identified CD63 as a potential HDV coreceptor.

The manuscript would benefit from a more detailed virological analysis, such as:

• Determination of HDV genome and antigenome sequences and analysis of HDV editing.
• Analysis of HDV short and large antigens by western blot.
• Analysis of HBV-related virological parameters in monoinfected and co-infected cells.

Additional information that needs to be added, better explained, or corrected:

The authors should explain why they used different MOIs depending on the genotype.

In Figure 1, it is unclear on which day the HCLs were infected by HDV and on which day they were transduced with AAV-NTCP.

It is not very clear if the authors used AAV serotype 6 consistently to transduce the cells. It would be valuable to show the transduction efficiency of AAV at different time points of HLC maturation, as it might also be affected and could explain some results.

For example, in Figure 6H, why does AAV-CD63 transduction increase HDV infectivity at day 15 but not at day 10? It would be interesting to repeat the anti-CD63 western blot after AAV-CD63 transduction.

The authors claim that by using AAV to express HBsAg, they are mimicking the expression of HBsAg from the integrated sequence rather than cccDNA. However, it is the opposite, as AAV genomes, like cccDNA, remain as episomes in the cells.

Significance

In the present work, Chi et al. demonstrated that Hepatocyte-like cells (HCLs) derived from human pluripotent cells (hPCs) can be infected by HBV. The development of new HDV cellular models is of great value for understanding HDV biology and developing new treatments. However, the relevance of the present work is limited by a recent publication by Lange et al., in which they also showed that HCLs derived from hPCs can be infected by HDV. The authors added new information to the work of Lange et al, including:

• HLCs derived from human pluripotent cells can be infected by different HDV genotypes.
• They proved that infectious HDV particles are formed.
• They identified CD63 as a potential HDV coreceptor.

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 the present manuscript, H Chi et al describe the infection of stem cell derived hepatocytes with HBV and HDV. They suggest that it could be used to validate antiviral treatment in a mature hepatocyte model. Moreover, they take advantage of the differentiation process of the cells to identify time points correlating with significant change in viral permissivity, and focus on one of these time points in an attempt to identify new host factors of HDV.

Overall, the manuscript is well written and brings interesting information toward the establishement of an efficient HBV HDV coinfection model in stem cell derived hepatocytes. Particularly, comparison to dHepaRG, another model relying on in vitro differentiation and commonly used to study HBV and HDV infection, reveals the potential of stem cell derived hepatocytes. While the efficiency of co infection in the stem cell derived hepatocytes may seem low, the manuscript goes in the direction of helping establishing a new mature in vitro model of infection.

Figure 1: The observation of a denser subpopulation of hepatocytes more susceptible to HDV is interesting. Do you have more characterization of this cell subpopulation, by IFA, in term of hepatic maturation marker, known HDV host factors and particularly NTCP expression?

Fig 1B-C: the comparison with dHepaRG is very interesting, and confirms the validity of SC derived hepatocytes as a model for HDV infection. dHepaRG can be heterogeneous. Do you also see the same phenotype of enriched HDV infection within a denser subpopulations of dHepaRG?

Fig 1B: Unit is confusing, using terms usually used for titration of infectivity, from the virus input point of view, not from the cellular point of view. Can you use % infected cells instead, or "HDV infection rate" like in Supp Fig 1B?

Fig 1B and C: Can a BLV control be included in the figure?

Fig 1A-F: What is the overall level of NTCP between HLC, HepaRG, Huh7NTCP and HLC-AAV-NTCP? Can NTCP and HDAg be stained simultaneously in your cells?

Fig 1I is confusing. Was BLV assay also performed on the HLC infection (Day 0), or only during the titration assay in Huh7NTCP?

Fig 1K: x-axis is confusing... is it number of HBV, HDV and HBV/HDV positive cells? Or number of infected cells upon inoculation with HBV, HDV, or both? Please clarify.

Figure 2: The AAV based vector to over express HBsAg is a very interesting tool, and the figure convincingly show production of HDV progeny viruses in HLC-AAV-HBsAg. Results shown are in agreement with previous studies based on hepatoma cell lines.

Figure 2B: What is IU/ml? Infectious Unit? International Unit? Are units in Fig 1B, 2B and 2C the same?

Figure 3: What is the overall number of transmission events observed in the co-culture setup? Can you visually observed viral spreading? Panel A shows only 1 event, making it hard to assess its efficiency. Titration assay in Fig 2C show production of up to 4-5 log of infectious HDV. But HLCs susceptibility to HDV infection may change during time...

Figure 4: While the strategy is interesting, based on what has been previously shown for HCV in Wu et al., 2012, the lack of statistical data prevents the reader to really understand and see drastic difference in term of susceptibility to infection and level of expression of host genes. In panel C, is the difference between day 13 and 15 statistically significant? Same for panel D, day 17 vs 19? As a remark, day 19, the peak of susceptibility to HDV, seems to be also the peak of maturation, based on ALB RTqPCR (panel B).

Figure 5: In panel A, GO pathways should be sorted based on significance, not Number of genes. In panel B-D, what is the scale of the heatmap on figure 5: change in CPM values, however log2, log10?

Figure 6: Do you have info about CD63 in other mature model, like dHepaRG and PHHs? Is CD63 also limiting in these models?

Significance

Overall, the manuscript brings interesting information toward the establishement of an efficient HBV HDV coinfection model in stem cell derived hepatocytes. Particularly, comparison to dHepaRG, another model relying on in vitro differentiation and commonly used to study HBV and HDV infection, reveals the potential of stem cell derived hepatocytes. While the efficiency of co infection in the stem cell derived hepatocytes may seem low, the manuscript goes in the direction of helping establishing a critical needed and long awaited mature in vitro model of HDV HBV infection.

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:

The authors identify ZMYND8 as a bromodomain protein: is there evidence the actions described in this paper involve interaction of ZMYND8 with acetylated lysines? Does this mechanism play a role in ZMYND8's transcriptional regulatory activities?

ZMYND8 is recruited to chromatin via its Bromo, PHD, and PWWP domains which recognize H3K4me1 and/or H3K14ac marks. Methyl marks on H3K4 are regulated by several lysine methyltransferases (e.g., MLL family and SETD1A/B) and demethylases (e.g., KDM5A-D) while H3K14ac is regulated by GCN5/PCAF, p300/CBP and/or Myst3. ZMYND8 also recruits histone deacetylases to chromatin including members of the highly conserved Nucleosome Remodeling and Deacetylase (NuRD) complex, HDAC1 and HDAC2. NuRD primarily deacetylates H3K27ac marks, however it is possible other acetyl moieties are affected by this complex.

Using ChIP-seq, we now show that Zmynd8-cKO cardiomyocytes retain H3K27ac marks at misexpressed genes. Interestingly, while some of these genes have altered H3K27ac at their promoters (and therefore have full-length misexpressed transcripts; i.e., Casq1, Cdh16) other genes (e.g., Lamb3, Chst3) show changes in H3K27ac in the middle of the gene and this tracks with gene expression changes. We interpret this unusual transcript and H3K27ac pattern as evidence of potential ZMYND8-regulated intragenic enhancer elements. We include the following in our resubmission:

1. Figure 5 which shows changes in H3K27ac levels at different genes, showing examples of genome browser tracks at the following genes Casq1, Cdh16, Camk1g, and Chst3.
2. Supplemental Figure S5 showing H3K27ac and H3K27me3 marks at the cardiac myosin locus (i.e., Myh6 and Myh7) and surrounding genes in control and Zmynd8-cKO * We also show retention of H3K27ac at the Zmynd8 gene in Zmynd8-cKO cardiomyocytes, again supporting an autoregulatory mechanism of Zmynd8 *expression.
3. An additional section in Results titled “H3K27 acetylation marks are retained at specific loci in Zmynd8-cKO cardiomyocytes”
4. New “ChIP-seq and Analysis” section in Materials and Methods
5. An updated model in Figure 6 that includes ZMYND8’s activities in modulating H3K27ac levels This first analysis on H3K27ac and H3K27me3 deposition in Zmynd8-cKO cardiomyocytes is not comprehensive and genome-wide analysis on these datasets will ultimately be performed in combination with additional datasets including ZMYND8 ChIP-seq from isolated cardiomyocytes. However, given the pertinence to ZMYND8’s transcriptional activities and in response to this reviewer’s critique, we include this pertinent H3K27ac and H3K27me3 ChIP-seq data.

Given the newness of this model and multiple isoform issues, the authors should show the entire gel for the westerns in SFigure 1C.

We now show the entire blots for all western blots in Supplementary Figure 1.

Nuclear staining is in SFigure 1E (typo in text): most of the staining in the control is non-myocyte and non-nuclear, making the statement about IHC showing depletion less convincing for Nkx lines.

We have fixed the typo in the text on page 5 line 128 and now correctly refer to this figure as Supplemental Figure S2. To better visualize nuclear ZMYND8 staining in this figure, we now show an adjusted image with increased contrast and brightness settings on both control and Zmynd8-cKO images and added arrowheads to indicate nuclei in the isolated cardiomyocytes. We also note that the flox sites only span the nuclear localization sequence for the protein so cytoplasmic ZMYND8 may still be present in Zmynd8-cKO cells.

Regarding perinuclear ZMYND8 staining: am I accurate in observing the perinuclear staining is still present in the KO? What do the authors make of this?

We do not observe perinuclear staining of ZMYND8 in KO cells. In Figure 1C, we believe the reviewer is observing potential staining in the cytoplasm, not perinuclear staining of ZMYND8 that we see in the control Myh6-CreTg/0 cardiomyocytes. We have added yellow arrowheads in Figure 1C to delineate perinuclear ZMYND8 staining we describe in the text.

What is the protein amount in the Zmynd8fl/+ mice? Do the hearts upregulate the protein to compensate?

We have added a gel in Supplemental Figure 1 that now shows protein isolated from Myh6-CreTg/0; Zmynd8fl/+ hearts and Myh6-CreTg/0 controls (Supplemental Figure 1C, right gel). It does not appear that Myh6-CreTg/0; Zmynd8fl/+ cardiomyocytes upregulate ZMYND8 to compensate for loss of one allele, as determined by Western blotting. However, our analysis shows differing ratios of the detected bands between conditional heterozygous mice, underscoring the need to further study the different ZMYND8 species present in cardiomyocytes. We state this in the results section (page 5, lines 123-124).

Do the individual cardiomyocytes hypertrophy in the Zymnd8 cKO mice? Do they proliferate?

Our analysis of cardiomyocyte morphology does reveal hypertrophy. The results we report include a new observation of variation in cell shape and are likely at least as sensitive as WGA staining which we find to be confounded by sectioning artifacts, cell identity, and position of the sections in the heart. We do not observe changes in H3S10ph staining between wild type and knockout hearts (data not shown) however we acknowledge further analysis of this may be warranted via other cell proliferation markers.

Regarding this statement: "These results show that ZMYND8 is necessary to prevent the onset of contractile dysfunction that leads to heart failure and death." I think what the authors showed is that loss of ZMYND8 causes contractile dysfunction, heart failure and death.

We acknowledge the difference in these statements and have now changed the text on page 7, lines 160-162 to “…these results show that loss of ZMYND8 from cardiomyocytes leads to contractile dysfunction, heart failure, and death.”

The switch like up regulation of skeletal muscle genes is an interesting observation. Do the authors have any evidence how this works? Other studies with EZH2 are mentioned, and if ZYMND8 is in fact acting as a bromodomain, the mechanism might involve regulation of enhancer methylation/acetylation at K27. This is testable, certainly at the target genes the investigators have identified (Casq1 and Tnni2), by ChIP-PCR.

As described above, we now include ChIP-seq data of H3K27ac and H3K27me3 marks in control and Zmynd8-cKO cardiomyocytes. As the reviewer suggests, there is retention of H3K27Ac marks in cKO cardiomyocytes, suggesting that ZMYND8 is necessary to recruit histone deacetylases to specific loci to remove acetyl moieties from H3K27. Regarding specific skeletal muscle genes, we do find a difference in histone acetylation at the promoter of the Casq1 gene and show this in Figure 5.

The model in Figure 4C makes sense, but the authors do not present any data to support this molecular mechanism. If the authors ChIP for localization of TFs in KO vs control and/or examine histone marks, they could build support for this model, particularly since they have already identified target genes.

We have now updated our model in Figure 6 to include ZMYND8’s role in modulating H3K27ac levels at target loci, leading to upregulation of mRNA transcripts. We add consideration of the implications of this in the Discussion.

Reviewer #1 (Significance (Required)):

The authors identify ZMYND8 as a bromodomain protein: is there evidence the actions described in this paper involve interaction of ZMYND8 with acetylated lysines? Does this mechanism play a role in ZMYND8's transcriptional regulatory activities? We include new data to demonstrate this. Please see above.

REVIEWER #2:

The study is reporting the role of ZMYND8 chromatin factor in the mouse heart. Mutations have been previously identified in genetic studies of atrioventricular septal defects and syndromic congenital cardiac abnormalities. Therefore the authors perform cardiomyocyte specific knockout of exon 4 (with the nuclear localisation signal) using Myh6 and Nkx2.5 cre. Full length protein seems to be removed from the nucleus. The knockout doesn't seem to affect embryonic development, but leads to hypertrophy and premature death. The authors perform transcriptome analysis and find 55 upregulated and 4 downregulated genes that are mainly related to contraction and ion transport. especially they find skeletal muscle proteins including fast-twitch troponin I upregulated. Tnni2 seems to be integrated into the sarcomeres, albeit the antibody staining is not in the expected location (see below). Shape of cardiomyocytes was apparently different, although this is seemingly not related to Tnni2 expression.

Specific points: - ZMYND8 has been previously linked to atrioventricular septal defects, but the authors do not explore if this is the case also in their model; could the authors please expand

We have not seen obvious septal defects in any Zmynd8-cKO mice. We now state this explicitly in the Results section on page 7, lines 159-160 and discuss this discrepancy from the observations in humans in our Discussion on page 12. The human study analyzing families carrying Zmynd8 mutations reported a variety of heart malformations in 7 of the 11 individuals. The septal defects observed in these individuals were not consistent and may be incidental to the molecular function of ZMYND8 within cardiomyocytes. One possibility is that these malformations are caused by stress during development, with Zmynd8 mutations sensitizing the heart to these defects. We acknowledge in the discussion that further analyses of septal defects in this knockout model could be useful in the future with more stringent stereoscopic techniques.

• the initial section is difficult to follow. Especially, the authors seem surprised regarding the size of the bands. They should make clear what the expected band size should be after removal of exon 4 and if this doesn't fit, explore the reasons experimentally if possible.

Rigorous analyses of the different Zmynd8 isoforms in cardiomyocytes will be a focus of future work as this may explain the mosaicism seen in cKO cardiomyocytes and the discrepancy between TNNI2 expression and cell shape (see below). We have reorganized the section and discuss potential explanations for our observed band sizes.

• the authors explore the shape of the cardiomyocytes and find cells that are shorter and thicker. It would be meaningful to include other metrics including, sarcomere length, contractility measurements and calcium transients (especially in light of the change ion transporters).

We agree that an investigation of the effects of the mutation and the skeletal muscle proteins on cellular contractility could be very interesting. Here we have contented ourselves with evaluating the effects at a physiological level through assessment of cardiac function.

• it is unclear why Tnni2 stains for the M-band (where in fact should be no actin and troponin) and not a typical double band with the H zone excluded (see here for good staining example: https://www.biorxiv.org/content/10.1101/2020.09.09.288977v1.full.pdf). also the staining looks very fuzzy. can the authors provide evidence that the antibody is staining troponin I in skeletal muscle at the correct localisation to demonstrate the specificity of the antibody?

We thank the reviewer for raising this point and do agree that there are instances where we observe TNNI2 staining colocalizing with MYOM1 staining. After closer examination of our images, we believe we do also see TNNI2 staining between M-lines and attribute this discrepancy to our antibody staining and/or biological differences between cells however, further analysis with better microscopy and immunostaining techniques is warranted. We have added an additional image to Figure 4A and have modified this results section on page 9, lines 217-222.

• it is interesting why Tnni2 is detectable only in a subfraction of cells, but this remains unexplored. Could this e.g. be right vs left ventricular cardiomyocytes? or is this related to the remaining isoforms of ZMYND8? The authors should try to identify the source of this variability

We agree that the TNNI2 mosaicism is an interesting phenotype and thank the reviewer for possible explanations. We favor the model of mosaicism being an effect of compensatory mechanisms by other ZMYND8 isoforms and discuss this in the discussion on page 8, line 228-229. This will be a focus of future work.

• if Tnni2 is unrelated to the changes in hypertrophic phenotype of the cardiomyocytes, then the authors should aim to identify if one of the other differentially regulated proteins might be related (e.g. ion transporter). The experiments above might help to identify this

We agree that identifying the causal agents of hypertrophy in this model would be interesting. It is however possible that we are simply seeing the expected effect of reduced contractility leading to hypertrophic compensation. Sorting this out will require additional mutant analyses and/or siRNA experiments all of which come with their own caveats and are outside of the scope of this initial analysis. Our aim for this manuscript was to report on the effects of ZMYND8 removal from cardiomyocytes. Additionally, it is certainly possible that phenotypes we report in this article are independent of the gene expression changes we have detected in the mutant and could be caused by other roles for ZMYND8 such as the DNA damage response. We include this possibility in our discussion.

Reviewer #2 (Significance (Required)):

Overall the manuscript is interesting in principle - it documents the role of a disease linked protein that hasn't been explored in the heart in detail, however at this point it seems premature and doesn't follow through on a solid detailed analysis.

The change in transcription profiles and especially the upregulation of skeletal muscle isoforms is intriguing, but should be further explored. There seems a lack of hypothesis and instead the authors analyse Tnni2 and cell shape, but while the cell shape is different they don't find a correlation with Tnni2. so if the authors suggest that cell shape is important (as indeed might be), how is this regulated?

Our goal for this initial paper is to describe the physiological and molecular phenotypes of the Zmynd8-cKO mouse model. It would be interesting to pursue a study directed at this question, perhaps of cell sorted "fat" and "thin" myocytes, but that would be beyond the scope of this report.

The study could be of interest to cardiovascular researchers, but needs to be expanded on the points above.

My expertise is in cardiovascular research

REVIEWER #3:

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

Summary:

Provide a short summary of the findings and key conclusions (including methodology and model system(s) where appropriate). Please place your comments about significance in section 2.

The authors found that Zmynd8-cKO mice develop dilated hearts, decreased cardiac function, and illegitimate expression of skeletal muscle genes. They concluded that ZMYND8 is necessary to maintain appropriate cardiomyocyte gene expression and cardiac function.

Major comments:

• Are the claims and the conclusions supported by the data or do they require additional experiments or analyses to support them? The claim that "Zmynd8 is dispensable for cardiac development" is not supported by the lethality of Zmynd8 D/D mice.

We interpret our observation that viable Nkx2.5-CreTg/0; Zmynd8fl/fl mice are born and grow to adulthood as evidence that Zmynd8 is not necessary for establishment of the cardiac lineage. However, we do agree that labeling Zmynd8 as dispensable is not supported by the experiments using Zmynd8D/D mice. We hypothesize that the lethality of the Zmynd8D/D mice is due to early embryonic events since empty egg sacs were observed at E8.0, however we do agree that ZMYND8’s role in cardiac development cannot be assessed using this line. We state that empty yolk sacs are found in mother uteri 8 days after mating on page 4, lines 94-96.

• Please request additional experiments only if they are essential for the conclusions. Alternatively, ask the authors to qualify their claims as preliminary or speculative, or to remove them altogether. The claim should be changed into "function of Zmynd8 in cardiac development can not be fully assessed in Zmynd8 D/D mice".

We agree that the lethality of Zmynd8D/D * mice prevents any analysis of early embryonic roles for the establishment of the cardiac lineage. This is additionally confounded by the fact that other partial-length isoforms of Zmynd8* may still be present in our knockout model. We have modified our interpretation and have further discussed the potential role of ZMYND8 in early cardiac development on page 4, line 96.

• If you have constructive further reaching suggestions that could significantly improve the study but would open new lines of investigations, please label them as "OPTIONAL". OPTIONAL: What about the phenotype of Nkx2-5 Cre mediated knockout of Zmynd8? Is it more severe than Myh6 Cre mediated knockout? At more earlier embryonic stage when cardiomyocytes are differentiated, are the skeletal muscle developmental genes ectopically upregulated in heart tube?

This is an interesting observation and deserves further investigation. Our initial analysis of Nkx2.5-CreTg/0; Zmynd8fl/fl mice reveals that these mice do not die earlier than Myh6-CreTg/0; Zmynd8fl/fl mice or have a more severe phenotype. In fact, mice with Nkx2.5-Cre mediated cKO mice live longer than Myh6-Cre mediated cKO mice. We show that these mice do have ZMYND8 depleted from their cardiomyocyte nuclei and ectopically express TNNI2.

This discrepancy in phenotype has been observed recently in mice lacking Kdm8 (Ahmed et al, 2023) and has been attributed to a lower efficiency of the Nkx2.5-Cre recombinase compared to Myh6-driven Cre.

• Are the suggested experiments realistic in terms of time and resources? It would help if you could add an estimated time investment for substantial experiments. Yes.

• Are the data and the methods presented in such a way that they can be reproduced? Yes.

• Are the experiments adequately replicated and statistical analysis adequate? Yes.

Minor comments:

• Specific experimental issues that are easily addressable. Have the female Zmynd8-cKO mice always died before their male siblings been pregnant with heart overload?

All lifespan data are of non-pregnant females. All mice (i.e., both males and females) used in these analyses were not used for mating. We now explicitly say this in the mouse husbandry section of our Materials and Methods section.

• Are prior studies referenced appropriately?

This paper "De Novo ZMYND8 variants result in an autosomal dominant neurodevelopmental disorder with cardiac malformations" should be referenced.

Thank you. We have referenced this paper (Dias et al. 2022) on page 3, line 61 as well as in the Discussion on page 9, line 211.

• Are the text and figures clear and accurate? Description of "cardiomegaly, preventing a compact myocardium phenotype, heart enlargement and thinning of the ventricular" should be more accurate and professional. We have changed the following in the text:

Page 6, line 150 “preventing a compact myocardium phenotype” to “during later stages of cardiac development” on

Page 6, line 153 “heart enlargement” to “The heart weight of Zmynd8-cKO mice”

Page 7, line 158 “thinning of the ventricular” to “dilated cardiomyopathy”

• Do you have suggestions that would help the authors improve the presentation of their data and conclusions? GSEA analysis of RNA-seq can be used to show the enrichment of cardiac and skeletal genes.

Because GSEA analysis requires at least three replicates per group to have the appropriate statistical power, we opted to show Gene Ontology analysis using DAVID software.

Reviewer #3 (Significance (Required)):

• General assessment: provide a summary of the strengths and limitations of the study. What are the strongest and most important aspects? What aspects of the study should be improved or could be developed? This study show that Zmynd8-cKO mice develop dilated hearts, decreased cardiac function, and illegitimate expression of skeletal muscle genes. However, the genes regulated by Zmynd8 during early developmental stage have not been identified and the functional mechanism of Zmynd8 during heart development remains unclear.

• Advance: compare the study to the closest related results in the literature or highlight results reported for the first time to your knowledge; does the study extend the knowledge in the field and in which way? Describe the nature of the advance and the resulting insights (for example: conceptual, technical, clinical, mechanistic, functional,...). Genetic mutations of Zmynd8 have been identified in congenital heart diseases with cardiac structural defects. And this study further shows that dysfunction/weaker mutations of Zmynd8 as a reason for dilated cardiomyopathy with decreased function.

• Audience: describe the type of audience ("specialized", "broad", "basic research", "translational/clinical", etc...) that will be interested or influenced by this research; how will this research be used by others; will it be of interest beyond the specific field? This study shows that dysfunction of Zmynd8 as a reason for dilated cardiomyopathy with decreased function. Researchers of "basic research" and "clinical" may be interested in this study.

• Please 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. heart development, dilated cardiomyopathy, epigenetics

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:

Provide a short summary of the findings and key conclusions (including methodology and model system(s) where appropriate). Please place your comments about significance in section 2.

The authors found that Zmynd8-cKO mice develop dilated hearts, decreased cardiac function, and illegitimate expression of skeletal muscle genes. They concluded that ZMYND8 is necessary to maintain appropriate cardiomyocyte gene expression and cardiac function.

Major comments:

• Are the claims and the conclusions supported by the data or do they require additional experiments or analyses to support them?

The claim that "Zmynd8 is dispensable for cardiac development" is not supported by the lethality of Zmynd8 / mice. - Please request additional experiments only if they are essential for the conclusions. Alternatively, ask the authors to qualify their claims as preliminary or speculative, or to remove them altogether.

The claim should be changed into "function of Zmynd8 in cardiac development can not be fully assessed in Zmynd8 / mice". - If you have constructive further reaching suggestions that could significantly improve the study but would open new lines of investigations, please label them as "OPTIONAL".

OPTIONAL: What about the phenotype of Nkx2-5 Cre mediated knockout of Zmynd8? Is it more severe than Myh6 Cre mediated knockout? At more earlier embryonic stage when cardiomyocytes are differentiated, are the skeletal muscle developmental genes ectopically upregulated in heart tube? - Are the suggested experiments realistic in terms of time and resources? It would help if you could add an estimated time investment for substantial experiments.

Yes. - Are the data and the methods presented in such a way that they can be reproduced?

Yes. - Are the experiments adequately replicated and statistical analysis adequate?

Yes.

Minor comments:

• Specific experimental issues that are easily addressable.

Have the female Zmynd8-cKO mice always died before their male siblings been pregnant with heart overload? - Are prior studies referenced appropriately?

This paper "De Novo ZMYND8 variants result in an autosomal dominant neurodevelopmental disorder with cardiac malformations" should be referenced. - Are the text and figures clear and accurate?

Description of "cardiomegaly, preventing a compact myocardium phenotype, heart enlargement and thinning of the ventricular" should be more accurate and professional. - Do you have suggestions that would help the authors improve the presentation of their data and conclusions?

GSEA analysis of RNA-seq can be used to show the enrichment of cardiac and skeletal genes.

Significance

• General assessment: provide a summary of the strengths and limitations of the study. What are the strongest and most important aspects? What aspects of the study should be improved or could be developed? This study show that Zmynd8-cKO mice develop dilated hearts, decreased cardiac function, and illegitimate expression of skeletal muscle genes. However, the genes regulated by Zmynd8 during early developmental stage have not been identified and the functional mechanism of Zmynd8 during heart development remains unclear.
• Advance: compare the study to the closest related results in the literature or highlight results reported for the first time to your knowledge; does the study extend the knowledge in the field and in which way? Describe the nature of the advance and the resulting insights (for example: conceptual, technical, clinical, mechanistic, functional,...). Genetic mutations of Zmynd8 have been identified in congenital heart diseases with cardiac structural defects. And this study further shows that dysfunction/weaker mutaions of Zmynd8 as a reason for dilated cardiomyopathy with decreased function.
• Audience: describe the type of audience ("specialized", "broad", "basic research", "translational/clinical", etc...) that will be interested or influenced by this research; how will this research be used by others; will it be of interest beyond the specific field? This study shows that dysfunction of Zmynd8 as a reason for dilated cardiomyopathy with decreased function. Researchers of "basic research" and "clinical" may be interested in this study.
• Please 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. heart development, dilated cardiomyopathy, epigenetics

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 study is reporting the role of ZMYND8 chromatin factor in the mouse heart. Mutations have been previously identified in genetic studies of atrioventricular septal defects and syndromic congenital cardiac abnormalities. Therefore the authors perform cardiomyocyte specific knockout of exon 4 (with the nuclear localisation signal) using Myh6 and Nkx2.5 cre. Full length protein seems to be removed from the nucleus. The knockout doesn't seem to affect embryonic development, but leads to hypertrophy and premature death. The authors perform transcriptome analysis and find 55 upregulated and 4 downregulated genes that are mainly related to contraction and ion transport. especially they find skeletal muscle proteins including fast-twitch troponin I upregulated. Tnni2 seems to be integrated into the sarcomeres, albeit the antibody staining is not in the expected location (see below). Shape of cardiomyocytes was apparently different, although this is seemingly not related to Tnni2 expression.

Specific points:

• ZMYND8 has been previously linked to atrioventricular septal defects, but the authors do not explore if this is the case also in their model; could the authors please expand
• the initial section is difficult to follow. Especially, the authors seem surprised regarding the size of the bands. They should make clear what the expected band size should be after removal of exon 4 and if this doesn't fit, explore the reasons experimentally if possible.
• the authors explore the shape of the cardiomyocytes and find cells that are shorter and thicker. It would be meaningful to include other metrics including, sarcomere length, contractility measurements and calcium transients (especially in light of the change ion transporters)
• it is unclear why Tnni2 stains for the M-band (where in fact should be no actin and troponin) and not a typical double band with the H zone excluded (see here for good staining example: https://www.biorxiv.org/content/10.1101/2020.09.09.288977v1.full.pdf). also the staining looks very fuzzy. can the authors provide evidence that the antibody is staining troponin I in skeletal muscle at the correct localisation to demonstrate the specificity of the antibody?
• it is interesting why Tnni2 is detectable only in a subfraction of cells, but this remains unexplored. Could this e.g. be right vs left ventricular cardiomyocytes? or is this related to the remaining isoforms of ZMYND8? The authors should try to identify the source of this variability
• if Tnni2 is unrelated to the changes in hypertrophic phenotype of the cardiomyocytes, then the authors should aim to identify if one of the other differentially regulated proteins might be related (e.g. ion transporter). The experiments above might help to identify this

Significance

Overall the manuscript is interesting in principle - it documents the role of a disease linked protein that hasn't been explored in the heart in detail, however at this point it seems premature and doesn't follow through on a solid detailed analysis.

The change in transcription profiles and especially the upregulation of skeletal muscle isoforms is intriguing, but should be further explored. There seems a lack of hypothesis and instead the authors analyse Tnni2 and cell shape, but while the cell shape is different they don't find a correlation with Tnni2. so if the authors suggest that cell shape is important (as indeed might be), how is this regulated?

The study could be of interest to cardiovascular researchers, but needs to be expanded on the points above.

My expertise is in cardiovascular research

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 authors identify ZMYND8 as a bromodomain protein: is there evidence the actions described in this paper involve interaction of ZMYND8 with acetylated lysines? Does this mechanism play a role in ZMYND8's transcriptional regulatory activities?

Given the newness of this model and multiple isoform issues, the authors should show the entire gel for the westerns in SFigure 1C. Nuclear staining is in SFigure 1E (typo in text): most of the staining in the control is non-myocyte and non-nuclear, making the statement about IHC showing depletion less convincing for Nkx lines.

Regarding perinuclear ZMYND8 staining: am I accurate in observing the perinuclear staining is still present in the KO? What do the authors make of this?

What is the protein amount in the Zmynd8fl/+ mice? Do the hearts upregulate the protein to compensate?

Do the individual cardiomyocytes hypertrophy in the Zymnd8 cKO mice? Do they proliferate?

Regarding this statement: "These results show that ZMYND8 is necessary to prevent the onset of contractile dysfunction that leads to heart failure and death." I think what the authors showed is that loss of ZMYND8 causes contractile dysfunction, heart failure and death.

The switch like up regulation of skeletal muscle genes is an interesting observation. Do the authors have any evidence how this works? Other studies with EZH2 are mentioned, and if ZYMND8 is in fact acting as a bromodomain, the mechanism might involve regulation of enhancer methylation/acetylation at K27. This is testable, certainly at the target genes the investigators have identified (Casq1 and Tnni2), by ChIP-PCR.

The model in Figure 4C makes sense, but the authors do not present any data to support this molecular mechanism. If the authors ChIP for localization of TFs in KO vs control and/or examine histone marks, they could build support for this model, particularly since they have already identified target genes.

Significance

The authors identify ZMYND8 as a bromodomain protein: is there evidence the actions described in this paper involve interaction of ZMYND8 with acetylated lysines? Does this mechanism play a role in ZMYND8's transcriptional regulatory activities?

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)):

In this manuscript, the authors report a novel and simple method to analyze the heterogeneity of various organelles. After imaging a large set of fluorescent-marker-labeled organelles, cluster analysis is adapted for illuminating the dynamics of organelles. Through this novel method, the authors are able to report organelle contact, which previously can only be observed by super-resolution imaging. This is method could significantly accelerate future discoveries at the cellular level. The manuscript is well written and has the potential be published in high-ranking journals, after a minor revision.

To further demonstrate the unique power of this new method, the authors should test cells under known stimulation altering the dynamics of organelles. For instance, wortmannin can blocks the conversion from early endosomes to late endosomes. By doing that, the potential of this new method will be endorsed.

Response:

We thank Reviewer #1 for the positive comments. We will add an experiment using wortmannin to block the process of endocytosis at a specific stage, as part of the experiments analyzing the process of endocytosis.

**Minor issue:** The authors should include more details about how to avoid signal crosstalk between adjacent fluorescent channels.

Response:

In the Methods section, we have added the following sentences to Lines 398-405.

“In order to avoid signal crosstalk between adjacent fluorescence channels, eight fluorophores with distinct spectral distances were selected, and the samples were irradiated sequentially with lasers in the order from the longest wavelength, i.e., fluorescence from 646 to 731 nm was excited by a 640 nm laser, fluorescence from 569 to 634 nm was excited by a 561 nm laser, fluorescence from 494 to 554 nm was excited by a 488 nm laser, and fluorescence from 411 to 481 nm was excited by a 405 nm laser, as shown in Extended Data Fig. 1b.”

Reviewer #1 (Significance (Required)):

The comprehensive monitoring of organelle dynamics through the integration of multi-dimensional parameters can proficiently evaluate the condition and prognosticate the destiny of living cells in response to external stimulations. This new multi-dimensional assay reported in this manuscript represents a huge step towards this goal. Since this new method is simple and powerful, cell biologists will quickly start to use this new method for the study of subcellular dynamics.

My lab is also developing a similar approach for organelles based on super-resolution imaging. I would like to congratulate the authors for this beautiful work.

Response:

We thank Reviewer #1 for the positive comment.

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

The manuscript reports a multi-parametric particle-based method for analysis of organelles. The method aims to resolve heterogeneous populations of organelles involved in various cellular processes. They propose to isolate organelles labelled with multiple markers, after homogenization and sonification of the cells, and analyse the resulting particles by fluorescence microscopy using spectral imaging. Afterwards, the authors visualize and analyse the obtained data with dimension reduction techniques.

Even though an interesting approach, the method and presented applications needs major improvisations before it can prove to be impactful for the field

I note some possible improvement points below:

• Initially, I think the current set of cell lines and labels should be extended also to include a wider set. The current limited set raises the question if the method authors report is also applicable to other cell lines, or if it only feasible with overexpressed markers. Including different cell lines with different labels would make the study more convincing and comprehensive.

Response:

We thank Reviewer #2 for this constructive comment. Regarding cell types, we will conduct experiments with HEK293T cells in addition to HeLa cells, labeling at least five different types of typical organelles. In our method, as shown in Figure 1a and 5a, we have already used not only overexpressed markers but also fluorescently labeled ligands (EGF-Alexa, transferrin-Alexa) and antibodies against endogenous proteins (anti-PMP70, anti-LAMP1), as well as direct labeling of cell membrane proteins (Alexa-NHS). Therefore, there are no significant limitations with respect to organelle labeling methods.

• It is surprising that the authors explicitly list already the limitations of fluorescence microscopy and super-resolution microscopy in the second paragraph of their introduction, however present a method fully dependent on fluorescence labelling and imaging methods. Actually their approach takes away the spatial information of FM approaches, and further makes the approach prone to the limitations they state.

They are also not fully fair about the limitation they state for Electron microscopy, as newly developed approaches (e.g. doi:10.1093/micmic/ozad067.1091; 　doi:10.1126/science.aay3134) widely extend the limited field of view and sampling capacity of EM. I recommend the authors to state the potential advantage/superiority of the reported method rather than stating the unclear limitations of the existing powerful methods.

Response:

Regarding fluorescence microscopy, it appears that our description was inadequate and misled the reviewers. There is no problem with fluorescence microscopy itself. What we intended to convey was that “when attempting to detect individual organelles ‘in cells’, there are limitations in the resolution of fluorescence microscopy because organelles are densely packed”. We have added this to the text on Line 49. Also, we thank Reviewer #2 for informing us about the high-speed 3D electron microscopy. We have cited the indicated papers in the text at Lines 54-55 and mention that “except for the recently developed high-throughput electron microscopy”.

• Most organelle markers the isolation of organelles are based on are overexpressed in the cells: endoplasmic reticulum (ER, mTagBFP2 (BFP)-SEC61B), mitochondria (GFP-OMP25 and SNAP-OMP25), and the Golgi (Venus-GS27). This raises significant questions about the native state relevance of the reported results, and how well they represent the endogenous processes.

Response:

We will add experiments analyzing the behavior of both endogenous and exogenous markers for the same organelles, for example, anti-LAMP1 antibody and VAMP7-GFP for lysosomes, and anti-PMP70 antibody and PEX16-GFP for peroxisomes.

• For the application on endosomes, can the authors state what is the new information enabled by their method? They study the very trafficking of EGF and Transferrin, 2 widely used endosomal cargoes with very well characterized trafficking steps, and show they are trafficked through Rab5/7 and Rab11 positive endosomes, respectively. This recapitulates the existing information, however falls short in delivering new insight. The authors can use these cargoes for proof-of-concept, but I would recommend to extend their study with less exploited cargoes to represent the potential of the reported method to deliver new information.

Response:

We thank Reviewer #2 for the positive suggestion about the potential of our method to provide new information. However, to demonstrate new biological insights, it would take a lot of time and delay the provision of our methodology, so we would like to submit this manuscript as a Methods paper with the proof-of-concept data.

Reviewer #2 (Significance (Required)):

The significance of biochemical and cellular processes being spatially regulated cellular organelles, and the roles of specific organelles in diseases from cancer to neurodegeneration are continuously being discovered and appreciated. Therefore development of methods reporting on the structure and function of organelles is important to accelerate these studies. In the reported method, however, the ultrastructure (as in Fib 1b) and the spatial information of the cellular organelles are inherently lost. The method falls in between a biochemical and a microscopic approach, however the advantages are not clearly portrayed. I recommend the authors to carefully and explicitly state where their method would be the method of choice rather than a biochemistry, mass spectroscopy, or microscopy approach. The authors should critically consider such an experiment as a proof-of-concept case.

Response:

We thank Reviewer #2 for the valuable suggestion. We added the following to the Discussion (Lines 267-277).

“A further potential application of our method would be to measure how the levels of key molecules in an organelle change during its differentiation or maturation. For example, the levels of PI4P and syntaxin 17 change during autophagosome maturation (Shinoda et al. eLife Preprint Review doi.org/10.7554/eLife.92189.1), which can be better demonstrated by this method using multiple markers for each stage of autophagosome formation and maturation, PI4P, and syntaxin17 because autophagosomes at different stages coexist in cells. In such cases, our single-particle analysis method, which examines the state of individual autophagosomes, would be more appropriate than biochemical methods that examine averages. In addition, it is difficult to quantitatively analyze many organelle structures in cells using fluorescence microscopy. Our particle-based analysis method can overcome this problem.”

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

**Comments, suggestions, and questions**

• I would like to start with a positive suggestion. The authors completely miss out on the opportunity to promote their approach by not relying on any type of fixation. In most multiplexing experiments, the first major challenge is to find antibodies that work well for imaging. The second challenge is then to find antibodies that work well under the same fixation conditions. The authors present a multiplexing approach that is completely independent of fixation. I suggest discussing this in the manuscript and promoting the approach in that regard.

Response:

We thank Reviewer #3 for pointing out the advantages of our method. We have added “Our method that is independent of fixation is advantageous for the optimization of the staining condition (Lines 298-299).

• I am wondering what defines the ‘resolution’ of this approach. I assume it is a combination of the sonication time -the longer the cell is sonicated, the smaller the fragments are - and the density of particles on the coverslip. What are the limits here? How does this affect the UMAP analysis? I would encourage the authors to discuss this in the manuscript.

Response:

The particle density on a coverslip can be easily reduced by simply diluting the particles in a buffer solution. Therefore, there is no density limit, which is an advantage of a cell-free system. To improve the resolution within a single organelle, for example, to separate distinct subdomains, as the reviewer mentioned, we can prolong the sonication time to make the particles smaller. However, since this will reduce the signal-to-background ratio and destroy organelle contacts, we used the sonication conditions as mild as possible. To investigate organelle subdomains and fragile contacts, the sonication conditions need to be optimized carefully, which should affect the UMAP analysis, but we think that these will be future work.

We do not think that prolonged sonication will affect the UMAP analysis because relative fluorescent signals of each particle would not change. However, as mentioned above, too strong sonication would worsen the signal-to-noise ratio, resulting in poor clustering.

We have added the above discussion to the Discussion (Lines 288-293).

“Also, to improve the resolution within a single organelle, for example, to separate distinct subdomains, we can prolong the sonication time to make the particles smaller. However, since this will reduce the signal-to-background ratio and may destroy organelle contacts, we used the sonication conditions as mild as possible. To investigate organelle subdomains and fragile contacts, the sonication conditions need to be optimized carefully.”

• The only real control the authors present are the correlative light and electron microscopy (CLEM) three images in Figure 1b, which seems very minimalistic for a very central and essential control experiment. How many of these control images did the authors take? Is there possibly a second method for a control experiment to link the fluorescence readout to an organelle fragment (e.g., purification or pulldown)?

Response:

Since all the markers we used are well-established, we believe that there is no concern about the fluorescence readouts to the organelle fragments. We have cited the following papers in Lines 84-85.

SEC61B: Rapoport, T. A., Jungnickel, B. & Kutay, U. Protein transport across the eukaryotic endoplasmic reticulum and bacterial inner membranes. Annu Rev Biochem 65, 271–303 (1996).

OMP25: Horie, C., Suzuki, H., Sakaguchi, M. & Mihara, K. Characterization of signal that directs C-tail-anchored proteins to mammalian mitochondrial outer membrane. Mol Biol Cell 13, 1615–1625 (2002).

GS27: Hay, J. C. et al. Localization, Dynamics, and Protein Interactions Reveal Distinct Roles for ER and Golgi SNAREs. J Cell Biol 141, 1489–1502 (1998).

Fusella, A., Micaroni, M., Di Giandomenico, D., Mironov, A. A. & Beznoussenko, G. V. Segregation of the Qb-SNAREs GS27 and GS28 into Golgi Vesicles Regulates Intra-Golgi Transport. Traffic 14, 568–584 (2013).

Although it is relatively easy to identify mitochondria-derived particles by EM based on their size and the presence of cristae-like structures (indeed we see many examples), it is more challenging for other organelles (because they appear simple vesicles). This is why we showed only mitochondria in Fig. 1b. Furthermore, the main purpose of this EM image is to show membrane contacts between the ER and mitochondria (related to Fig. 3).

• Line 37-41: Could the authors please strengthen these statements with an appropriate citation (e.g., a review)?

Response:

We have cited the textbook Molecular Biology of THE CELL (the 6th edition, Chapter 12 and Chapter 13) in Lines 37 and 41.

• Line 51: The statement, “Super-resolution microscopy could partially solve the resolution problem, but it is currently limited to four-color imaging,” is incorrect. Agasti et al. demonstrated up to nine target multiplexed super-resolved imaging with DNA-PAINT in 2017 (https://pubs.rsc.org/en/content/articlehtml/2017/sc/c6sc05420j). Additionally, two papers currently on Biorxiv demonstrate 12 target and 30 target multiplexed super-resolution imaging with FLASH-PAINT (https://www.biorxiv.org/content/10.1101/2023.05.17.541061v1.abstract) and SUM-PAINT(https://www.biorxiv.org/content/10.1101/2023.05.17.541210v1.abstract). Please cite these papers accordingly.

Response:

We thank Reviewer #3 for notifying us of these important studies. We have rewritten the sentence on Lines 51-52 to read “Although multicolor imaging has been attempted with super-resolution microscopy (references of the indicated papers), it only partially solves the issue of resolution.”

• The authors use spectral unmixing to overcome the limit of spectral multiplexing. While this has been demonstrated to work well for less than ten targets, it does not scale to multiplexing experiments with more than ten target species. I suggest that the authors discuss in the discussion part of the manuscript the potential of DNA-based multiplexed imaging, such as CODEX or DNA-PAINT, in combination with the presented approach.

Response:

In the Discussion (Lines 295-298), we have added the sentence “Current fluorescent particle detection uses spectral multiplexing, but this method has only been able to detect up to eight colors. Methods such as CODEX or DNA-PAINT with wide-field type illumination could significantly increase the number of targets”.

• Regarding the spectral unmixing approach, please cite previous work described in the literature (e.g.,https://www.nature.com/articles/nature22369, or earlier work).

Response:

We thank Reviewer #3 for informing us. We have cited it in Line 72.

• By using spectral unmixing for multiplexing, this method is limited to confocal due to spectral detection needs and therefore limited in throughput. It would be beneficial if it could work with wide-field type illumination. This could substantially increase the throughput, which is another reason why I think it would be important to discuss sequential multiplexing.

Response:

We agree with the Reviewer’s comment. We have added the discussion to Lines 295-298 as described in our response to Reviewer #3, Comment (6).

• To image contact sites, the authors use split GFP. There have been discussions that split GFP might, in some cases, facilitate the process that is supposed to be measured, in this case, the formation of contact sites. I suggest using at transient version of split GFP, called split fast, for follow-up experiments in the authors’ next papers (https://www.nature.com/articles/s41467-019-10855-0).

Response:

We thank Reviewer #3 for providing this information. We will do it as suggested in the next paper.

• Line 27 & 253: Please drop the term ‘intuitive’ or explain better what you mean by intuitive. For me, UMAP is certainly a very useful tool, but it is not at all what I would describe as intuitive.

Response:

We have deleted ‘intuitive’ in all seven places and rewritten them (Lines 27, 43, 58, 72, 180, 231, and 253).

• Lastly, I want to mention that the authors state they used ChatGPT, DeepL, and DeepL Write for translation from Japanese to English. I appreciate their honesty.

Response:

We thank Reviewer #3 for the comment.

Reviewer #3 (Significance (Required)):

In the manuscript titled “Organelle Landscape Analysis Using a Multi-parametric-Based Method,” Kurikawa et al.present a method for multi-parametric, particle-based analysis of cellular organelles. After lysing cells, the fractions of the organelles are partially labeled with fluorescently tagged antibodies, while others are already tagged with fluorescent proteins, using six to eight spectrally different fluorescent dyes/proteins. These fractions are subsequently immobilized on a poly-L-lysine-coated coverslip. The authors use spectral unmixing to distinguish these markers. The6-8 multiplexed imaging data is then presented in two-dimensional UMAP space. The authors then use this approach to visualize seven major organelles, transitional sites of endocytic organelles, and contact sites between the endoplasmic reticulum and mitochondria using split GFP.

The authors present, in my opinion, a conceptually new and interesting approach by combining spectral unmixing for imaging up to eight targets, with organelle fragment imaging, and presenting multidimensional data in two-dimensional Uniform Manifold Approximation and Projection (UMAP) space in this manuscript. They further validated this approach by linking the results of the experiments to results established or at least reported in the literature.

In general, the manuscript is, in my opinion, a good fit for publication as it presents a conceptionally new approach and an interesting example of applying the UMAP approach, which might be of interest to a broader readership. Therefore, after an appropriate response to my comments, suggestions, and questions (see below), I would recommend this manuscript for publication.

Response:

We thank Reviewer #3 for the positive comment.

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

Comments, suggestions, and questions

• I would like to start with a positive suggestion. The authors completely miss out on the opportunity to promote their approach by not relying on any type of fixation. In most multiplexing experiments, the first major challenge is to find antibodies that work well for imaging. The second challenge is then to find antibodies that work well under the same fixation conditions. The authors present a multiplexing approach that is completely independent of fixation. I suggest discussing this in the manuscript and promoting the approach in that regard.
• I am wondering what defines the 'resolution' of this approach. I assume it is a combination of the sonication time - the longer the cell is sonicated, the smaller the fragments are - and the density of particles on the coverslip. What are the limits here? How does this affect the UMAP analysis? I would encourage the authors to discuss this in the manuscript.
• The only real control the authors present are the correlative light and electron microscopy (CLEM) three images in Figure 1b, which seems very minimalistic for a very central and essential control experiment. How many of these control images did the authors take? Is there possibly a second method for a control experiment to link the fluorescence readout to an organelle fragment (e.g., purification or pulldown)?
• Line 37-41: Could the authors please strengthen these statements with an appropriate citation (e.g., a review)?
• Line 51: The statement, "Super-resolution microscopy could partially solve the resolution problem, but it is currently limited to four-color imaging," is incorrect. Agasti et al. demonstrated up to nine target multiplexed super-resolved imaging with DNA-PAINT in 2017 (https://pubs.rsc.org/en/content/articlehtml/2017/sc/c6sc05420j). Additionally, two papers currently on Biorxiv demonstrate 12 target and 30 target multiplexed super-resolution imaging with FLASH-PAINT (https://www.biorxiv.org/content/10.1101/2023.05.17.541061v1.abstract) and SUM-PAINT (https://www.biorxiv.org/content/10.1101/2023.05.17.541210v1.abstract). Please cite these papers accordingly.
• The authors use spectral unmixing to overcome the limit of spectral multiplexing. While this has been demonstrated to work well for less than ten targets, it does not scale to multiplexing experiments with more than ten target species. I suggest that the authors discuss in the discussion part of the manuscript the potential of DNA-based multiplexed imaging, such as CODEX or DNA-PAINT, in combination with the presented approach.
• Regarding the spectral unmixing approach, please cite previous work described in the literature (e.g., https://www.nature.com/articles/nature22369, or earlier work).
• By using spectral unmixing for multiplexing, this method is limited to confocal due to spectral detection needs and therefore limited in throughput. It would be beneficial if it could work with wide-field type illumination. This could substantially increase the throughput, which is another reason why I think it would be important to discuss sequential multiplexing.
• To image contact sites, the authors use split GFP. There have been discussions that split GFP might, in some cases, facilitate the process that is supposed to be measured, in this case, the formation of contact sites. I suggest using a transient version of split GFP, called split fast, for follow-up experiments in the authors' next papers (https://www.nature.com/articles/s41467-019-10855-0 ).
• Line 27 & 253: Please drop the term 'intuitive' or explain better what you mean by intuitive. For me, UMAP is certainly a very useful tool, but it is not at all what I would describe as intuitive.
• Lastly, I want to mention that the authors state they used ChatGPT, DeepL, and DeepL Write for translation from Japanese to English. I appreciate their honesty.

Significance

In the manuscript titled "Organelle Landscape Analysis Using a Multi-parametric-Based Method," Kurikawa et al. present a method for multi-parametric, particle-based analysis of cellular organelles. After lysing cells, the fractions of the organelles are partially labeled with fluorescently tagged antibodies, while others are already tagged with fluorescent proteins, using six to eight spectrally different fluorescent dyes/proteins. These fractions are subsequently immobilized on a poly-L-lysine-coated coverslip. The authors use spectral unmixing to distinguish these markers. The 6-8 multiplexed imaging data is then presented in two-dimensional UMAP space. The authors then use this approach to visualize seven major organelles, transitional sites of endocytic organelles, and contact sites between the endoplasmic reticulum and mitochondria using split GFP.

The authors present, in my opinion, a conceptually new and interesting approach by combining spectral unmixing for imaging up to eight targets, with organelle fragment imaging, and presenting multidimensional data in two-dimensional Uniform Manifold Approximation and Projection (UMAP) space in this manuscript. They further validated this approach by linking the results of the experiments to results established or at least reported in the literature.

In general, the manuscript is, in my opinion, a good fit for publication as it presents a conceptionally new approach and an interesting example of applying the UMAP approach, which might be of interest to a broader readership. Therefore, after an appropriate response to my comments, suggestions, and questions (see below), I would recommend this manuscript for publication.

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 reports a multi-parametric particle-based method for analysis of organelles. The method aims to resolve heterogeneous populations of organelles involved in various cellular processes. They propose to isolate organelles labelled with multiple markers, after homogenization and sonification of the cells, and analyse the resulting particles by fluorescence microscopy using spectral imaging. Afterwards, the authors visualize and analyse the obtained data with dimension reduction techniques.

Even though an interesting approach, the method and presented applications needs major improvisations before it can prove to be impactful for the field

I note some possible improvement points below:

• Initially, I think the current set of cell lines and labels should be extended also to include a wider set. The current limited set raises the question if the method authors report is also applicable to other cell lines, or if it only feasible with overexpressed markers. Including different cell lines with different labels would make the study more convincing and comprehensive.
• It is surprising that the authors explicitly list already the limitations of fluorescence microscopy and super-resolution microscopy in the second paragraph of their introduction, however present a method fully dependent on fluorescence labelling and imaging methods. Actuallt their approach takes away the spatial information of FM approaches, and further makes the approach prone to the limitations they state. They are also not fully fair about the limitation they state for Electron microscopy, as newly developed approaches (e.g. doi:10.1093/micmic/ozad067.1091; doi: 10.1126/science.aay3134) widely extend the limited field of view and sampling capacity of EM. I recommend the authors to state the potential advantage/superiority of the reported method rather than stating the unclear limitations of the existing powerful methods.
• Most organelle markers the isolation of organelles are based on are overexpressed in the cells: endoplasmic reticulum (ER, mTagBFP2 (BFP)-SEC61B), mitochondria (GFP-OMP25 and SNAP-OMP25), and the Golgi (Venus-GS27). This raises significant questions about the native state relevance of the reported results, and how well they represent the endogenous processes.
• For the application on endosomes, can the authors state what is the new information enabled by their method? They study the very trafficking of EGF and Transferrin, 2 widely used endosomal cargoes with very well characterized trafficking steps, and show they are trafficked through Rab5/7 and Rab11 positive endosomes, respectively. This recapitulates the existing information, however falls short in delivering new insight. The authors can use these cargoes for proof-of-concept, but I would recommend to extend their study with less exploited cargoes to represent the potential of the reported method to deliver new information.

Significance

The significance of biochemical and cellular processes being spatially regulated cellular organelles, and the roles of specific organelles in diseases from cancer to neurodegeneration are continuously being discovered and appreciated. Therefore development of methods reporting on the structure and function of organelles is important to accelerate these studies. In the reported method, however, the ultrastructure (as in Fib 1b) and the spatial information of the cellular organelles are inherently lost. The method falls in between a biochemical and a microscopic approach, however the advantages are not clearly portrayed. I recommend the authors to carefully and explicitly state where their method would be the method of choice rather than a biochemistry, mass spectroscopy, or microscopy approach. The authors should critically consider such an experiment as a proof-of-concept case.

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, the authors report a novel and simple method to analyze the heterogeneity of various organelles. After imaging a large set of fluorescent-marker-labeled organelles, cluster analysis is adapted for illuminating the dynamics of organelles. Through this novel method, the authors are able to report organelle contact, which previously can only be observed by super-resolution imaging. This is method could significantly accelerate future discoveries at the cellular level. The manuscript is well written and has the potential be published in high-ranking journals, after a minor revision.

To further demonstrate the unique power of this new method, the authors should test cells under known stimulation altering the dynamics of organelles. For instance, wortmannin can blocks the conversion from early endosomes to late endosomes. By doing that, the potential of this new method will be endorsed.

Minor issue: The authors should include more details about how to avoid signal crosstalk between adjacent fluorescent channels.

Significance

The comprehensive monitoring of organelle dynamics through the integration of multi-dimensional parameters can proficiently evaluate the condition and prognosticate the destiny of living cells in response to external stimulations. This new multi-dimensional assay reported in this manuscript represents a huge step towards this goal. Since this new method is simple and powerful, cell biologists will quickly start to use this new method for the study of sub-cellular dynamics.

My lab is also developing a similar approach for organelles based on super-resolution imaging. I would like to congratulate the authors for this beautiful work.

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 We appreciate the insightful reviewer comments. Both reviewers alluded to the logical lack of connection between two themes in the original paper. Specifically, we showed that N-cad differentially regulates migration in different environments, and that leader and follower cells differ phenotypically, but did not connect the two themes. In this revised version, we've performed additional experiments and undertaken a comprehensive reorganization of both the manuscript and figures. The major changes are outlined below:

2. Figure 4 A-C has been moved to Figure 6 F-H.

3. Figure 5 has been moved to Figure S3 F-H.
4. Figure 6 F has been moved to Figure 7 A.
5. Figure 6 G-H have been moved to Figure 7 D-E.
6. Figure 6 I-J have been moved to Figure S5 A-B.
7. Figure 7 C-F have been moved to Figure S5 C-F.
8. Added transcriptome profiling of control and N-cad-depleted cells and of leader and follower cells (Figures 6 E, S1 H and S4 C-D, Tables S2 and S3). We have incorporated additional figures (Figure 4 and 5 in the revised manuscript) to support the idea that the amount of N-cad at the cell surface is regulated by endocytic recycling, thereby stimulating glioma migration in the different local environments. Furthermore, our new findings showed that YAP1/TAZ regulates the surface level of N-cad during glioma migration (Figure 8). We trust that these additions contribute to the clarity and robust justification of our paper.

Similar to other types of tumors, our findings revealed that pediatric high-grade gliomas migrate collectively, possibly contributing to a more aggressive invasion than single cells. In this study, we found that N-cad mediates this collective glioma migration. Interestingly, within these migrating groups, leader and follower cells dynamically interchange positions during migration, accompanied by changes their phenotypic characteristics. This suggests that differences in phenotypes, including N-cad recycling, proliferation and YAP activation, may be predominantly regulated by cell-extrinsic factors rather than being predetermined by genetic or epigenetic factors. Moreover, our new RNA-sequencing results indicate minimal difference between leader and follower cells, except for the upregulation of YAP response and wound healing migration genes in leader cells. Although genomic alterations still possibly encode the leader-follower exchange, our findings strongly suggest that the activation of YAP1 and glioma migration are regulated by the cellular context, specifically where cells are located within the group.

Contrary to our initial findings suggesting a positive feedback loop between N-cad endocytosis and nuclear YAP1, our revised data indicates that nuclear YAP appears to be independent of N-cad. We observed that homotypic interactions with N-cad present in the surrounding environment, such as neurons (Figure 6 C-D) or N-cad extracellular domain-coated surface (Figure 7 B-C), did not affect nuclear YAP1. However, YAP1/TAZ depletion decreased N-cad expression and altered its localization at the surface (Figure 8). This leads us to propose a revised model where nuclear YAP1 stimulates surface N-cad, thereby facilitating the distinct modes of migration on ECM and neurons (Figure 8 I).

1. Point-by-point description of the revisions

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

In this manuscript, Kim and colleagues describe the role of N-Cadherin during pediatric glioma migration. They compare cell lines that have similar transcripts but different levels of N-Cadherin protein and find that N-Cadherin levels influence the route of migration - whether it be on ECM or other tissues. They also describe molecular feedback between N-Cadherin and YAP in leader vs follow cells of their systems. The data are clear, well presented, and convincing; and the conclusions described by the manuscript are mostly justified. My major criticism of the manuscript is that the line of questioning undertaken does not appear well justified. At many points, I was left asking "but why are they doing this?" and I could not understand the rationale for some of the experiments that were performed (even if they were performed well). The manuscript opens by validly describing how gliomas are highly invasive, poorly understood and that N-Cadherin was highly expressed in comparison to other adhesion proteins. This opened the path for the questions and experiments performed that contributed to Figures 1-3, which I thought were interesting. From there on, I found the logic of the story unclear and poorly justified. For example, I do not know why leader and follower cells were justified - when it had nothing to do with N-Cadherin which was the focus of the work prior. And then, having rightly concluded in Figure 4 that the data suggested that leader and follower cells dynamically exchange positions rather than being pre-determined, they went onto further figures focusing on differences between leader and follower cells, which left my quite confused. I am likewise confused by the model proposed in that, they authors describe that the difference between leader and follower cells contributes to a nuclear YAP/N-Cad endocytosis feedback loop that feeds into the speed of migration. Yet, the authors describe earlier that leader and follower cells frequently exchange positions, with no evidence that they are pre-determined. How do the authors square these seemingly conflicting points? And further, what is the relevance of this to understanding the differing modes of migration (on ECM or other tissues)? On this issue, I suggest authors re-consider whether the order of figures or logic of the story is appropriate (perhaps consider moving some figures to supplement?), and to clearly justify in the text the elements that are being addressed. Overall, I think the messaging, logic and justification could be use significant improvement; the experiments however are well performed, and the figures are very clear and nicely presented, and I don't have any qualms about them.

We appreciate your insightful comments, recognizing the need for logical and justifiable improvements in certain sections of our previous manuscript. Please see Section 1, General Statements, for an explanation of changes made.

Minor Comments

1. Not required, but the authors may wish to consider putting t=0 pictures of the experiments in the supplement as supportive evidence for the circles of the initial seeding location they show in Fig 1.

We provide the t=0 images in Figure S1 N and O.

1. I assume the title of the second results section should say "migration speed" rather than "speed migration"

The new title of the second results section is “N-cad stimulates and inhibits migration through intercellular homotypic interaction”.

1. Fig. 4D - Are both example cell pictures leaders? If so, I'm not sure why two have been provided; I'm guessing the bottom set are supposed to be follower cells. If so, please label as appropriate. (And if not, a representative set of pictures from a follower cell should be provided).

We have enhanced the clarity of the labels. We provide representative high magnification images of leader and follower cells. The updated figure can be found in Figure 5 A.

1. Figure 5 Legend - the title of this figure is too definitive, and exaggerates further than the main text does, which was correct in saying that the experiments only suggest that N-Cadherin endocytosis might regulate the localisation of b-catenin and p120-catenin. Probably I would go further and say that there is no experimental evidence provided that even suggests that in the first place, and that this is a hypothesis that remains to be tested. The authors should inhibit endocytosis specifically (rather than just depleting N-Cad) and see the effect, to justify their conclusion.

We appreciated your points and concerns. Following your earlier suggestion, we have moved the figure to the supplementary section (Figure S3 F-H). Moreover, we have addressed the reciprocal regulation of N-cad and catenins by knocking down p120-, β- or α-catenin. Our new findings showed that p120-, β- or α-catenin depletion decrease the amount of N-cad at the cell surface, not steady-state protein level, resulting in decreased migration on astrocytes but increased migration on ECM (see Figure 4). These findings support the idea that catenins play a role in glioma migration according to the environment by altering surface N-cad level. With that, we updated the figure title to “Catenins regulate N-cad surface levels to stimulate or inhibit migration.”

Reviewer #1 (Significance (Required)):

The manuscript provides a characterised of invasive glioma migration that was previously lacking. It also provides interesting observations related to the role of N-Cadherin for different modes of migration (on ECM or on tissues) that will be of interest for further exploration. It makes a good advance in terms of addressing a highly invasive cell type that has poor prognosis. I anticipate that now this initial characterisation has been performed, authors and others will be interested in gaining a deeper understanding as to how these two modes of migration are controlled, how there might be interplay between them and how such findings contribute to its highly invasive nature. I have expertise in collective cell migration and directed cell migration.

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

Summary In the submitted manuscript, Kim et al. describe various aspects of N-cadherin function in the collective migration of PBT-05 cells, a pediatric high-grade glioma line, on laminin, 3D-matrigel, neurons or astrocytes. N-cadherin promotes the collective migration on neurons or astrocytes, whereas it suppresses the migration on laminin or 3D-matrigel. The authors also show that N-cadherin is actively internalized and recycled in the leader, but not follower, cells of the collective, which induce the nuclear accumulation of YAP/TAZ proteins. YAP/TAZ proteins are shown to regulate the collective migration.

Thank you for the comments. Please see Section 1, General Statements, for a summary of changes made. Please also note that our new experiments failed to show that N-cad levels or traffic regulate YAP/TAZ nuclear accumulation. Rather, YAP/TAZ are regulated by cell density independent of N-cad, and YAP/TAZ regulate N-cad protein levels and traffic independent of changes in N-cad RNA levels

Major comments

1. In Fig. 1G, N-cadherin knockdown seems to affect the distribution of astrocytes. The authors should stain a marker for astrocytes, instead of actin, and the red alone images should be provided.

Astrocytes were cultured for 4 days to generate 3D scaffolds before adding the glioma spheroid, essentially as described (Gritsenko et al., Histochem Cell Biol, 2017). Co-cultures were stained for human-specific vimentin (glioma) or actin (glioma and astrocytes) (see Figure 1 G and separate channels in new Figure S1 P). There do not appear to be major changes in astrocyte organization outside the migration front, but we lack a way to stain for astrocytes specifically and cannot visualize astrocytes under the glioma cells. It remains possible that astrocytes may be affected differently by contact with control and N-cad-deficient glioma cells. However, we added a new experiment, assaying migration on decellularized astrocyte ECM. While N-cad stimulated migration on astrocytes it inhibited migration on astrocyte ECM (Figures 1 I and J and S1 Q). Thus N-cad stimulates glioma migration on astrocyte cells and not their ECM.

1. The colocalization between N-cadherin and Rab11 may not be high in Figs. 4F and S2B. It is unclear whether the majority of the internalized N-cadherin is recycled to the plasma membrane. In Fig. S2B, the internalized N-cadherin may be located mainly at the early endosomes before transported to the recycling endosomes (Is it 20 min after the N-cadherin antibody internalization?). First, the authors should analyze the colocalization between the N-cadherin and Rab11 at 30-40 min after the internalization. If the colocalization with Rab11 would be still low at that time point, some of the internalized N-cadherin might be degraded in the lysosomes. To test this possibility, the authors should analyze the colocalization between N-cadherin and LAMP1 under the treatment with a lysosome inhibitor.

At steady state, N-cad co-localized better with Rab5 than with Rab11 or LAMP1 (Figure 5 C-D). In kinetics experiments, N-cad antibodies were internalized for 40 min. They colocalized better with Rab5 or EEA1 than with Rab11 or LAMP1. When we allowed recycling for an additional 20 min, the surface level of N-cad antibodies partially recovered in leader cells more than follower cells (see Figures 5 G and S3 D). We tested whether treatment with lysosomal inhibitors would increase co-localization of N-cad with Rab11 in recycling endosomes. Surprisingly, however, Chloroquine or Bafilomycin A1 decreased the amount of internalized N-cad antibody in leader and follower cells, and long-term treatment did not increase total N-cad levels. Therefore, the fate of internalized N-cad in follower cells remains unclear.

1. When N-cadherin is depleted, dissociated single cells are increased, but these cells are not well characterized. A high magnification image of the dissociated single cells is required. In addition, the migration speed of the dissociated single cells should be measured.

We didn’t quantify single cell migration because only a minority of cells separate from the collective even when N-cad is depleted. Therefore, we quantified migration directionality and speed for cells at or near the front of collective migration (Figure 2 D-I). We have updated the image of single cells, providing representative high-magnification images in Figure S1 N and O.

1. In Fig. S2D, treatment with Pitstop-2 alone or Dyngo-4a alone is required. Dynamin is also involved in clathrin-independent endocytosis and N-cadherin is reported to be internalized via caveolin-1-mediated endocytosis as well as clathrin-mediated during neuronal migration. It would be better to clarify which type of endocytosis occurs in the leader cells.

We have removed results showing inhibition of cell migration and N-cad endocytosis by Pitstop-2 and Dyngo-4a from the paper. Treatment with either chemical alone had much less effect on internalization or migration than adding both together (see figure below). This is hard to explain. Pitstop-2 should inhibit clathrin-coated pit formation and Dyngo-4a should inhibit clathrin and caveolin-mediated endocytosis. Caveolin-1 and 2 transcripts were not detected in our cells (Table S2). There may be some other form of clathin-independent endocytosis. Interpretation is also challenging since these inhibitors will inhibit endocytosis of many receptors, not just N-cad. Accordingly, we have removed these results in the revised manuscript.

1. In Fig. 2, N-cadherin depletion disturbs the migration directionality. Is this a result from disruption of cell polarity? To test this, the position of centrosome or Golgi or lamellipodia in the leader cells should be analyzed. (OPTIONAL)

We elected not to perform this analysis.

1. I cannot understand the significance of Fig. 5F and 5G. If the authors would speculate that alpha- and beta-Catenins may transduce the intracellular signaling from the internalized N-cadherin, the authors should perform the knockdown experiments of the Catenins and analyze whether it may affect the nuclear accumulation of YAP/TAZ. (OPTIONAL)

We agree. In the initial manuscript, we showed that N-cad depletion altered the localization of p120-, β-, and α-catenin (previously shown in Figure 5 F-G). For better clarity and logic, these figures have been moved to Figure S2 H in the revised manuscript. Additionally, to test whether catenins regulate N-cad and YAP1, we depleted p120-, β-, or α-catenin using shRNA. We found that downregulation of p120-, β-, or α-catenin decreased N-cad surface levels, consequently slowing migration on astrocytes and stimulating migration on laminin (Figure 4). In other words, depleting catenins altered migration in parallel with the changes in N-cad surface level. Catenin depletion also increased single-cell dissociation, reduced the crowding of leader and follower cells, and increased nuclear YAP1 (see figure below). These findings suggest that the main role of p120-, β-, or α-catenin is to regulate surface N-cad. Since this result does not support a role for catenins transducing an N-cad signal to YAP1, we have not included it in the paper.

Minor comments

1. The quantitative data is required in Fig. 5E.

Quantitative data from three independent experiment are now presented in Figure S2 G.

1. Vinculin is associated with the cadherin-catenin complex and it may not be a good loading control (Fig. 3C and 3L).

The Western blot data has been updated and is now presented in new Figure 3 B and 3 F, with β-tubulin as a loading control.

**Referees cross-commenting**

I totally agree with the other Reviewers' comments and evaluation. As the reviewer-1 pointed out, I also think the experiments are well performed, but it would lack logic at least in part (see my comment-6). In addition, as the reviewer-3 pointed out, the linking mechanism of N-cadherin homophilic interaction with YAP/TAZ signaling is important to improve this manuscript

We hope the revisions have improved the logical flow. We have also added new results showing that YAP/TAZ regulate N-cad protein levels and localization but not N-cad RNA. N-cad is not needed for cell density-dependent regulation of YAP1 localization. The model is shown in Figure 8 I.

Reviewer #2 (Significance (Required)):

Strength N-cadherin has multiple function in cancer and neuronal migration, and both positive and negative effects of N-cadherin on cancer cell migration have been reported. In this regard, different behaviors of N-cadherin in the leader and follower cells of the collective are interesting and may explain the controversial previous results.

Limitation This study reveals various aspects of N-cadherin function in the collective migration of the glioma cell line, but it is unclear whether these findings are applied to pediatric high-grade gliomas in vivo.

Thus, this study is a potentially important and informative to cell biologists and researchers in cancer biology, although this reviewer also found several weak points that should be improved.

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

In this manuscript, the authors explore the role of N-cadherin in the migratory/infiltrative behavior of human pediatric brain tumor cells, in light of their surrounding microenvironment. Their in-depth phenotype analysis allows to document the behavior of migrating cells and revisit the concept of leading/follower migratory cells (somehow more commonly applied to endothelial cells). They suspected that the YAP/TAZ pathway might modulate N-cadherin endocytosis and vice versa, using imagery-based cell tracking.

Major comments

1. To control for co-culture models, migration should be evaluated on decellularized matrices from astrocyte and neuron cultures.

We thank for your suggestion. We tested glioma migration on astrocyte-derived decellularized matrices. The mouse astrocytes we used are known to produce various extracellular matrices with a composition similar to Matrigel, except for laminin α5. (Gritsenko et al., J Cell Sci, 2018). N-cad shRNA cells migrated faster on decellularized ECM than control (Figure 1 I-J and S1 Q). This result agrees with N-cad depletion increasing migration on ECM but is opposite to migration on astrocytes.

1. N-cadherin was stably knocked down with shRNA, which raises the question of adaptative/compensatory mechanisms. First, one could ask what happen in knockout conditions. Similarly, transient siRNA-mediated silencing might help to strengthen the findings. Second, is there any impact of Ncad knock down on alternate adhesive receptors (either cell-cell or cell-ECM). This should be verified with bulk RNAseq.

Transient knockdown with N-cad siRNA also increased migration on laminin-coated surface (Figure S1 L-M). Unfortunately, N-cad depletion with siRNA was short-lived, precluding its use for long-term assays, like coculture with neurons or astrocytes. To test whether there is any impact of N-cad knockdown on alternative adhesion receptors, we performed RNA-Seq (Figure S1 H, Table S2). We found N-cad depletion did not alter expression of other cell-cell and cell-ECM adhesive receptors except CDH3 (2.8-fold increase compared with 7-fold decrease in CDH2). Integrin expression was unchanged.

1. It would be interesting to evaluate the impact of N-cadherin/N-cadherin homotypic interactions on YAP/TAZ signaling, using for instance N-cad-coated surface.

We observed that the homotypic interaction of N-cad with surrounding neurons and astrocytes did not hinder the accumulation of nuclear YAP1 in leader cells (Figure 6 C-D). To further support the idea that N-cad does not directly regulate YAP1 signaling, we have now measured YAP1 localization in cells migrating over N-cad ECD. The new data confirms that N-cad does not directly regulate YAP1 localization (Figure 7 B-C).

1. along this line, the impact of mechanical cues (stiffness, 2D vs 3D) is not explored.

We appreciate your suggestion. It is possible that different mechanical and cytoskeletal cues between leader and follower cells affect YAP1 signaling. In this study, we would like to focus more on the role of N-cad-mediated cell adhesions in YAP signaling.

Minor comments

1. Levels of N-cadherin expression in normal Astro and Neurons to compare with pediatric brain cancer cells (S1C)

A new western blot analysis to show N-cad levels in DMG, PHGG and mouse cerebellar neurons and astrocytes has been added to Figure S1 F.

1. Low versus high density culture conditions should be controlled and its further impact on the YAP/Ncad endocytosis route should be supported experimentally, or to be omitted from their proposed model.

We previously used different size of micropattern discs to control low or high cell density. Smaller cell clusters, with more edge cells and hence fewer cell-cell interactions, had higher nuclear YAP1 (Figure 7 D-E). We have repeated this experiment, including N-cad ECD antibodies to measure N-cad endocytosis. Smaller cell clusters had higher N-cad antibody internalization (Figure 7 F). Together with our evidence that leader cells have higher YAP1 and more N-cad internalization than followers, and that YAP/TAZ knockdown inhibits N-cad internalization, these results high YAP/TAZ in leader cells regulates N-cad internalization.

Reviewer #3 (Significance (Required)):

This paper presents robust image analysis of human pediatric brain tumor migration in the context of the different microenvironment that they might encounter (matrices, neurons, astrocytes). This study brings new concepts on the way N-cadherin might contribute to tumor cell migratory behavior based on the nature of the interactions in which N-cadherin is involved. As a limitation, it remains unclear the mechanism by which N-cadherin endocytosis is driven.

We now discuss the limitations of the study as follows:

“The mechanisms by which YAP1 regulates N-cad levels and trafficking remain to be explored. YAP1 is widely expressed in human brain tumors and strongly associated poor survival. Leader cells expressed higher levels of YAP1-response and wound-healing gene transcripts, but transcript levels of N-cad and proteins known to regulate cadherin traffic, such as p120-catenin, Rab5/11 and Rac1, were similar. Therefore, N-cad is likely regulated at the level of protein synthesis or turnover. More endosomal N-cad recycled to the surface of leader than follower cells, implying that follower cells might divert more N-cad for lysosomal degradation, but our attempts to interfere with N-cad endocytosis or degradation specifically were unsuccessful. Further understanding of the mechanism and function of N-cad recycling for glioma cell migration will require cargo-specific ways to selectively regulate endocytosis and recycling”.

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, the authors explore the role of N-cadherin in the migratory/infiltrative behavior of human pediatric brain tumor cells, in light of their surrounding microenvironment. Their in-depth phenotype analysis allows to document the behavior of migrating cells and revisit the concept of leading/follower migratory cells (somehow more commonly applied to endothelial cells). They suspected that the YAP/TAZ pathway might modulate N-cadherin endocytosis and vice versa, using imagery-based cell tracking.

Major comments:

1. To control for co-culture models, migration should be evaluated on decellularized matrices from astrocyte and neuron cultures.
2. N-cadherin was stably knocked down with shRNA, which raises the question of adaptative/compensatory mechanisms. First, one could ask what happen in knockout conditions. Similarly, transient siRNA-mediated silencing might help to strengthen the findings. Second, is there any impact of Ncad knock down on alternate adhesive receptors (either cell-cell or cell-ECM). This should be verified with bulk RNAseq.
3. It would be interesting to evaluate the impact of N-cadherin/N-cadherin homotypic interactions on YAP/TAZ signaling, using for instance N-cad-coated surface.
4. along this line, the impact of mechanical cues (stiffness, 2D vs 3D) is not explored.

Minor comments:

1. Levels of N-cadherin expression in normal Astro and Neurons to compare with pediatric brain cancer cells (S1C)
2. Low versus high density culture conditions should be controlled and its further impact on the YAP/Ncad endocytosis route should be supported experimentally, or to be omitted from their proposed model.

Significance

This paper presents robust image analysis of human pediatric brain tumor migration in the context of the different microenvironment that they might encounter (matrices, neurons, astrocytes).

This study brings new concepts on the way N-cadherin might contribute to tumor cell migratory behavior based on the nature of the interactions in which N-cadherin is involved.

As a limitation, it remains unclear the mechanism by which N-cadherin endocytosis is driven.

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 the submitted manuscript, Kim et al. describe various aspects of N-cadherin function in the collective migration of PBT-05 cells, a pediatric high-grade glioma line, on laminin, 3D-matrigel, neurons or astrocytes. N-cadherin promotes the collective migration on neurons or astrocytes, whereas it suppresses the migration on laminin or 3D-matrigel. The authors also show that N-cadherin is actively internalized and recycled in the leader, but not follower, cells of the collective, which induce the nuclear accumulation of YAP/TAZ proteins. YAP/TAZ proteins are shown to regulate the collective migration.

Major comments

1. In Fig. 1G, N-cadherin knockdown seems to affect the distribution of astrocytes. The authors should stain a marker for astrocytes, instead of actin, and the red alone images should be provided.
2. The colocalization between N-cadherin and Rab11 may not be high in Figs. 4F and S2B. It is unclear whether the majority of the internalized N-cadherin is recycled to the plasma membrane. In Fig. S2B, the internalized N-cadherin may be located mainly at the early endosomes before transported to the recycling endosomes (Is it 20 min after the N-cadherin antibody internalization?). First, the authors should analyze the colocalization between the N-cadherin and Rab11 at 30-40 min after the internalization. If the colocalization with Rab11 would be still low at that time point, some of the internalized N-cadherin might be degraded in the lysosomes. To test this possibility, the authors should analyze the colocalization between N-cadherin and LAMP1 under the treatment with a lysosome inhibitor.
3. When N-cadherin is depleted, dissociated single cells are increased, but these cells are not well characterized. A high magnification image of the dissociated single cells is required. In addition, the migration speed of the dissociated single cells should be measured.
4. In Fig. S2D, treatment with Pitstop-2 alone or Dyngo-4a alone is required. Dynamin is also involved in clathrin-independent endocytosis and N-cadherin is reported to be internalized via caveolin-1-mediated endocytosis as well as clathrin-mediated during neuronal migration. It would be better to clarify which type of endocytosis occurs in the leader cells.
5. In Fig. 2, N-cadherin depletion disturbs the migration directionality. Is this a result from disruption of cell polarity? To test this, the position of centrosome or Golgi or lamellipodia in the leader cells should be analyzed. (OPTIONAL)
6. I cannot understand the significance of Fig. 5F and 5G. If the authors would speculate that alpha- and beta-Catenins may transduce the intracellular signaling from the internalized N-cadherin, the authors should perform the knockdown experiments of the Catenins and analyze whether it may affect the nuclear accumulation of YAP/TAZ. (OPTIONAL)

Minor comments

1. The quantitative data is required in Fig. 5E.
2. Vinculin is associated with the cadherin-catenin complex and it may not be a good loading control (Fig. 3C and 3L).

Referees cross-commenting

I totally agree with the other Reviewers' comments and evaluation. As the reviewer-1 pointed out, I also think the experiments are well performed, but it would lack logic at least in part (see my comment-6). In addition, as the reviewer-3 pointed out, the linking mechanism of N-cadherin homophilic interaction with YAP/TAZ signaling is important to improve this manuscript

Significance

Strength

N-cadherin has multiple function in cancer and neuronal migration, and both positive and negative effects of N-cadherin on cancer cell migration have been reported. In this regard, different behaviors of N-cadherin in the leader and follower cells of the collective are interesting and may explain the controversial previous results.

Limitation

This study reveals various aspects of N-cadherin function in the collective migration of the glioma cell line, but it is unclear whether these findings are applied to pediatric high-grade gliomas in vivo.

Thus, this study is a potentially important and informative to cell biologists and researchers in cancer biology, although this reviewer also found several weak points that should be improved.

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

Kim et al review

In this manuscript, Kim and colleagues describe the role of N-Cadherin during pediatric glioma migration. They compare cell lines that have similar transcripts but different levels of N-Cadherin protein and find that N-Cadherin levels influence the route of migration - whether it be on ECM or other tissues. They also describe molecular feedback between N-Cadherin and YAP in leader vs follow cells of their systems. The data are clear, well presented, and convincing; and the conclusions described by the manuscript are mostly justified. My major criticism of the manuscript is that the line of questioning undertaken does not appear well justified. At many points, I was left asking "but why are they doing this?" and I could not understand the rationale for some of the experiments that were performed (even if they were performed well). The manuscript opens by validly describing how gliomas are highly invasive, poorly understood and that N-Cadherin was highly expressed in comparison to other adhesion proteins. This opened the path for the questions and experiments performed that contributed to Figures 1-3, which I thought were interesting. From there on, I found the logic of the story unclear and poorly justified. For example, I do not know why leader and follower cells were justified - when it had nothing to do with N-Cadherin which was the focus of the work prior. And then, having rightly concluded in Figure 4 that the data suggested that leader and follower cells dynamically exchange positions rather than being pre-determined, they went onto further figures focusing on differences between leader and follower cells, which left my quite confused.

I am likewise confused by the model proposed in that, they authors describe that the difference between leader and follower cells contributes to a nuclear YAP/N-Cad endocytosis feedback loop that feeds into the speed of migration. Yet, the authors describe earlier that leader and follower cells frequently exchange positions, with no evidence that they are pre-determined. How do the authors square these seemingly conflicting points? And further, what is the relevance of this to understanding the differing modes of migration (on ECM or other tissues)? On this issue, I suggest authors re-consider whether the order of figures or logic of the story is appropriate (perhaps consider moving some figures to supplement?), and to clearly justify in the text the elements that are being addressed. Overall, I think the messaging, logic and justification could be use significant improvement; the experiments however are well performed, and the figures are very clear and nicely presented, and I don't have any qualms about them.

Minor Comments

• Not required, but the authors may wish to consider putting t=0 pictures of the experiments in the supplement as supportive evidence for the circles of the initial seeding location they show in Fig 1.
• I assume the title of the second results section should say "migration speed" rather than "speed migration"
• Fig. 4D - Are both example cell pictures leaders? If so, I'm not sure why two have been provided; I'm guessing the bottom set are supposed to be follower cells. If so, please label as appropriate. (And if not, a representative set of pictures from a follower cell should be provided).
• Figure 5 Legend - the title of this figure is too definitive, and exaggerates further than the main text does, which was correct in saying that the experiments only suggest that N-Cadherin endocytosis might regulate the localisation of b-catenin and p120-catenin. Probably I would go further and say that there is no experimental evidence provided that even suggests that in the first place, and that this is a hypothesis that remains to be tested. The authors should inhibit endocytosis specifically (rather than just depleting N-Cad) and see the effect, to justify their conclusion.

Significance

The manuscript provides a characterised of invasive glioma migration that was previously lacking. It also provides interesting observations related to the role of N-Cadherin for different modes of migration (on ECM or on tissues) that will be of interest for further exploration. It makes a good advance in terms of addressing a highly invasive cell type that has poor prognosis. I anticipate that now this initial characterisation has been performed, authors and others will be interested in gaining a deeper understanding as to how these two modes of migration are controlled, how there might be interplay between them and how such findings contribute to its highly invasive nature.

I have expertise in collective cell migration and directed cell migration.

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

1. General Statements

We appreciate the reviewers’ thoughtful feedback and thank them for their valuable suggestions to improve the manuscript. We have endeavored to respond to all their comments, with many of their concerns already incorporated in the manuscript. Validations for the additional experiments to be incorporated into the manuscript have been performed and show that all the plans outlined in Section 2 are highly feasible and will be added for the full revision. We believe that the incorporated and planned revisions contribute to a significant improvement on the original manuscript.

2. Description of the planned revisions

Reviewer 1

Major comments:

Point 3. p. 5. The authors do not describe any relationship to notch signaling. But notch signaling is the mechanism by which a sprout is selected. The CA phenotype shows no selection, and every sprout can continue migration. Did the authors check for any relationship between notch signaling c-Src activation? Does upregulation of C-Src downregulate notch?

In previous unpublished results examining the impact of the loss of endothelial c-Src on notch signaling, we observed no alteration in DLL4 expression in the sprouting retina on postnatal day 5. Furthermore, no change in tip cell number was observed in mice with a loss of endothelial c-Src, suggesting c-Src depletion does not impact notch activity (Schimmel et al., Development, 2020, Figure 1M). We have started additional preliminary experiments performing immunostaining with a DLL4 antibody in migrating c-Src-CA cells to assess activation of notch signaling upon c-Src activation. We will continue these experiments for the full revision and will confirm the results via further analysis of notch activation by assessing DLL4 expression in the c-Src mutant cells using Western blot.

Reviewer 2

Major comments:

Point 1. The authors have only used one type of vein endothelial cells from one single donor but they conclude that is effect is general for all endothelial cells. Endothelial cells are very heterogeneous, not only depending on their function and localization, vein, artery or capillary, but also between different organs and in disease (PMID: 22315715, PMID: 28775214, PMID: 31944177, PMID: 33514719). The authors, should either repeat some of the key experiments in other type of endothelial cells, maybe arterial or microvasculature cells which are commercially available or at least state that the observations presented in this manuscript apply to HUVECs and discuss whether this would also apply for other cell types.

We agree it would be highly beneficial to assess whether c-Src-CA induces vascular expansion in other endothelial cell types. We have successfully transduced human arterial endothelial cells (HAEC) with empty vector and c-Src-CA lentivirus and are able to grow HAECs in 3D vessels. This demonstrates that introducing the c-Src constructs into other endothelial cells and putting them in 3D assays is highly feasible. We have also used human microvascular endothelial cells (HMVEC) in 3D vessels in previous studies (Schimmel et al., Clin Trans Immunol, 2021). Therefore, we will perform experiments introducing the full set of c-Src mutations in HAEC and/or HMVEC in 3D vessels for the revision to strengthen our findings.

Reviewer 3

Major comments:

Point 1. "This was further supported by our observation that there were no changes in proliferation in c-Src mutant cells grown in a 2D monolayer". Figure 1A appears to have increased number of cells in the c-Src-CA condition compared to the control condition. Could the authors quantify the number of cells/area as they did for their 3D vessel model? This would reinforce the idea that the ballooning phenotype they observe is not due to differences in proliferation.

We have started quantification on the number of cells per bead for the 3D bead sprouting experiments shown in Figure 1. We will complete this quantification for 3 independent experiments and the results will be added for the full revision.

Point 2. Would be strengthened with analysis of another proliferation marker, such as EdU label, which is incorporated only during S phase of the cell cycle. Comparing ki67 staining and EdU staining would provide more insights. Also, using their 3D vessel model for this analysis would increase its relevance.

We agree that showing proliferation in a 3D setting would be highly beneficial. We tested proliferation marker Ki67 in 3D vessels to ensure this analysis will be possible. We will perform full analysis of proliferation across c-Src mutations in 3D for the revision. We have started with BrdU labelling in 2D, and we will perform full analysis of proliferation with BrdU across c-Src mutations for the revision.

Point 3. In Figure 1E', cells expressing the constitutively active form of cSrc appear to detach, giving the impression of cell death. Have the authors tested the viability/apoptosis of c-Src-CA cells, particularly in their 3D model?

We agree that showing cell death in our model, especially in a 3D setting, would be highly beneficial. We have tested cell death marker Cleaved Caspase 3 (CC-3) in 3D vessels to ensure this analysis is feasible. We will perform full analysis of cell death across c-Src mutations in 3D for the revision.

Point 4. "Therefore, reduction of endothelial cell-cell contacts in c-Src-CA cells may be due to elevated VE-cadherin phosphorylation and subsequent internalisation", "As reduction in cell-cell junction integrity has been shown to increase migratory capacity and sprouting angiogenesis [38], our data suggest that a balanced control of both cell-matrix and cell-cell junctions is essential for mediating migration." In general, it's not clear how constitutively active cSrc affects focal adhesions and cell-cell adhesion and how this is responsible for their ballooning phenotype. The role of the phosphorylation of the VE-Cadherin and cell-cell junctions in this process is not clear either. Further analysis of cell-cell junctions and focal adhesions (co-staining of phosphorylated paxillin and VE-Cadherin) and focal adhesions/fibronectin (like in figure 4C) in the context of cell migration (scratch wound assay) would provide important information to strengthen this notion of balanced control of both cell-matrix and cell-cell junctions.

We will perform experiments on migrating cells in 2D, co-staining for p-paxillin and VE-cadherin, and p-paxillin and Fibronectin, to address the role of balanced cell-matrix and cell-cell junction adhesion, and how they influence Fibronectin deposition in migrating cells.

Point 6. "Taken together, these results reveal that proteases produced by c-Src-CA cells are locally secreted at FAs but are membrane bound." The claim that proteases are membrane-bound is not convincingly demonstrated. Could the authors assess whether the constitutive form of cSrc activates the expression of specific genes encoding MMPs by qPCR? Or is it more a matter of the effect of c-Src on the transport of MMPs by microtubules?

We would like to clarify the content of Figure 5, which presents two distinct sets of experiments supporting the assertion that the proteases under investigation are membrane-bound. Firstly, the transfer of conditioned medium from c-Src mutant cells demonstrated no degradation of fibronectin fibrils. Secondly, in the bead sprouting assay, a mixed culture of untransduced and c-Src-CA expressing cells was utilised. The results revealed that only c-Src-CA cells formed balloons, while untransduced cells sprouted normally right next to or sometimes even through a balloon.

Recognising the need for a more in-depth understanding, we acknowledge the importance of analysing specific MMP gene expression. To this end, we have ordered qPCR primers for distinct MMPs, namely MMP2, MMP7, MMP9, and MT1-MMP. These forthcoming experiments are not only highly feasible but will also contribute valuable insights. The results of this gene expression analysis will be incorporated into the revision, shedding light on whether constitutively active c-Src induces MMP gene expression or influences MMP transport.

Minor comments:

Point 2. The lab already showed in a previous study that mice lacking c-Src specifically in endothelial cells have reduced blood vessel sprouting, leading to the expectation that the constitutively active form of cSrc would increase sprout number in the sprouting assay. Could the authors explain why the constitutively active form of cSrc induces this vascular ballooning and not an increase in the number of sprouts?

In line with analysis to be performed on notch activity and DLL4 expression (Reviewer 1 point 3), we will provide additional discussion on the role of notch signalling and tip cell identity with the full revision.

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

R____eviewer 1

Major comments:

Point 1. p5. Fig 1: The sentence that the dominant negative completely abrogated 'this' phenotype implies that the dominant negative was put into the same cells as the constitutively active mutation. 'Abrogated' means it stops the phenotype, and the phenotype in the sentence prior was constitutively active. It is more accurate to say that the dominant negative was not distinguishable from wild type, which is what the statistics show. No double transfection (DN-CA) was performed.

We have changed the wording in the manuscript accordingly to ‘The c-Src-DN mutation showed no phenotype distinguishable from Ctrl (Fig 1A-D).’ on page 5.

Point 2. p.5. Fig 1: the phenotype of the CA cells is fascinating. They expand far beyond their normal territory, but they are held together in a lacy bubble. To me, this looks like a different phenotype from the ballooning that might occur in an arteriovenous malformation in vivo, as in vivo malformations are continuously covered by cells. I understand why the authors might use the term ballooning but given that the cells expand without continuously touching each other, I do not think this is the correct term. Would blebbing, or radial migration in a lace-like discontinuous pattern describe it better?

We have changed the phrasing from ‘ballooning morphology’ to ‘radial migration in a lace-like discontinuous pattern’ on page 5. For brevity, this has been referred to as ‘ballooning’ for the remainder of the manuscript, as noted on page 5.

Point 4. The statistical methods are not described in the methods (GraphPad?). These need to be added. Are only significant comparisons plotted? In Fig 6 and 7 only pairwise statistics are shown. If all significant comparisons are plotted, then this means that the comparison between the rescued CA and the treated or untreated control is not significant. This can be thought of as a partial rescue towards a wild type, but it is definitely not a full rescue. None of the statistical comparisons in Figure 6 or 7 show significant comparisons to wildtype. This needs more discussion.

We have now added additional clarification on statistical methods. Details on the statistical tests for each figure are mentioned in the figure legends. A general section on the statistical methods is now added to the methods section on page 18. Only significant comparisons are displayed in the graphs, but as mentioned by reviewer 2 (minor point 2), we have added additional information for transparency. Each of the different comparisons that were made, and their precise p value, have been compiled a table which has been added as Supplementary Table 1 to the manuscript.

In Figures 6 and 7, we exclusively plotted pairwise comparisons to assess the impact of Marimastat treatment. As outlined in Supplementary Table 1, there is still a statistical significance when comparing Marimastat-treated c-Src-CA with either Marimastat-treated Ctrl or Marimastat-treated c-Src-WT. This suggests a partial rescue. For clarity, we kept only pairwise comparisons in the graphs, but discussed the partial rescue due to remaining significant difference between Marimastat-treated c-Src-CA and Ctrl or c-Src-WT cells in the results, referring to Supplementary Table 1 for p values. An important sidenote: c-Src-CA treated cells cannot exhibit complete rescue since they are initially seeded without Marimastat, and have already initiated ballooning by the time treatment commences.

Point 5. Mmp activity is inferred, but not measured. This is a limitaion as the assumption is that marimostat acting through the expected pathway.

Marimastat is one of the most commonly used broad spectrum MMP inhibitors, with potent activity against major MMPs, including MMP1, MMP3, MMP2, MMP9, MMP7 and MMP14. This is outlined in the existing reference (Rasmussen and McCann, 1997). We have adjusted phrasing to clarify the potency of Marimastat and have emphasised this is an MMP targeting drug which has been widely utilised in oncology clinical trials (page 8).

Minor comments:

Point 1. Fig 5D. The presentation of the data in this graph is difficult to understand. It is trying to show the proportion of mScarlet in sprouts or balloons a percentage of all the scarlet cells. It would be better to have all cells represented in one bar, distributed between sprout and balloon in that one bar. i.e., for the control and dominant negative, the bars would be all black and then for the CA it would be all white. The zero data points are confusing. A proportions graph should be investigated here.

We have changed the graph in Figure 5D, which now represents the % of the outgrowth area, sprouts for Ctrl, c-Src-WT and c-Src-DN and balloon for c-Src-CA, that are mScarlet positive. Resulting in all black bars for Ctrl, c-Src-WT and c-Src-DN and all white bar for c-Src-CA, as the reviewer predicted.

Point 2. The methods for vessel coverage for quantification in figs 1 and 7 are missing.

We have added details of how quantification of vessel coverage in Figure 1 and 7 was performed to the methods section on page 17/18 as follow: ‘Microfluidic vessel coverage was measured by tracing any holes in the vessel wall (inverse of cell area marked by phalloidin) and dividing this by the total cell area per image.’

Reviewer 2

Minor comments:

Point 1. Although the methods are well written and can be understood. To improve transparency, the authors should reduce the referring to other papers to describe the methods they perform and at least some kind of brief description should be included.

We have added a brief description of the methods that included references to other papers; lentiviral transduction and microfluidic devices. More details about the lentivirus transduction were added on page 15 and a short description about the fabrication of the microfluidic devices was added on page 15/16.

Point 2. The authors should report the real p value for their tests. Also, when the test is not significant.

To provide more transparency about all of the different comparisons that were made and their precise p value, we have compiled a table listing all the p values and which is added as Supplementary Table 1 to the manuscript.

Reviewer 3

Minor comments:

Point 3. In Figure 1A, it would be beneficial to include images from orthogonal views. Indeed, in the c-Src-CA condition, it's not clear whether the vascular ballooning observed represents a cluster of cells or an empty space between the bead and the endothelial cells. (Supp movie 1 helps, but it would be useful to add orthogonal views to the figure)

For clarity, we have added single Z plane image for cross sectional views of the bead sprouts in Figure 1A to show that the c-Src-CA cells have an empty space inside the balloon, rather than being a big cluster of cells.

Point 4. In Figure 1D, the method used to analyze sprout shape is not clear, especially for the c-Src-CA condition where the number of sprouts is close to 0. The figure legend indicates that this measurement corresponds to the shape of the sprouting area. Could the authors clarify and explain their quantification method?

The shape of the sprouting area refers to the circularity index of the vascular area, measured by tracing the perimeter of the cell area in a minimum Z-projection of brightfield images and subtracting the area of the bead. For better clarity, we have adjusted the title of Figure 1D and Figure 6D to ‘Vascular area shape’ and added details of the quantification method in the methods section on page 17.

Point 5. "however cells within the vessel still maintained some connections (Fig 1E')": The connections between cells are difficult to see in the images in Figure 1E'. Could the authors provide higher magnification images of the VE-cadherin staining to illustrate these connections between cells?

For improved clarity, we have added high magnification images of the VE-cadherin channel only in black and white (Figure 1E’’) and indicated some of the maintained cell-cell connections in the c-Src-CA cells with black arrowheads.

Point 6. "The reduction in migration correlated with an increase in FA size c-Src-CA expressing cells.": Could the authors give more explanation?

We have adjusted phrasing to provide additional information (page 6/7) as follows: ‘The reduction in migration velocity in c-Src-CA cells coincides with an increase in FA size, number and density (Fig 2A-D). This suggests that the reduction of migration velocity is due to increased cellular adhesion via FAs.’

Point 7. Could the authors widen the cell trajectory trace in Supplementary Figure 3A?

We have adjusted the trajectory traces in Supplementary Figure 3A with wider lines for improved visibility.

Point 8. it is very difficult to distinguish fibronectin fibrils on the images shown in figure 4C. it would be beneficial to change the images.

We have enlarged the zoomed areas for better visibility of the focal adhesions and fibronectin degradation underneath those areas in the c-Src-CA cells. Additionally, arrows are added to indicate fibronectin fibrils.

Point 9. "Treatment of ECs with Marimastat in a fibrin bead sprouting assay resulted in a rescue of the ballooning morphology observed in the c-Src-CA cells" Based on the images displayed in the figure and the associated quantifications, it still appears that c-Src-CA+Marimastat induces a vascular ballooning even if it is less pronounced than in the DMSO condition. Hence, it would be more accurate to describe the observed effect as a "partial rescue". In the microfabricated 3D vessel, in the figure 7A, cell-cell junctions still appear altered by c-Src-CA after the treatment with Marimastat, compared to the c-Src-WT-Marimastat, it would be more appropriate to talk about "partial rescue".

We have changed ‘rescue’ to ‘partial rescue’ when referring to results in Figure 6 and 7 (page 8).

Point 10. In Figure 6A, it seems that there is a decrease in the number of sprouts in the c-Src-DN condition compared to the control condition after the DMSO treatment, which is not observed in Figure 1, could the authors explain why?

In Figure 1C, the number of sprouts is also reduced in the c-Src-DN condition compared to c-Src-WT, but this is not significant when compared to control (see Supplementary Table 1 for p values of all comparisons). However, it is true that the number of sprouts in the c-Src-DN condition is significantly reduced compared to both control and c-Src-WT upon DMSO treatment (Fig 6C). Reduction of sprouts in c-Src-DN cells was expected due to the dysfunctional kinase domain, as mentioned on page 5 and shown in reference 30 (Shvartsman, D.E., et al., J Cell Biol, 2007. 178(4): p. 675-86.). Why DMSO treatment seems to enhance the effects of dominant negative c-Src expression on sprouting behaviour remains unclear. However, DMSO has adverse effects on sprouting shown by reduction of sprouts in both control and c-Src-WT cells (comparing untreated condition in Fig 1C with DMSO treated condition in Fig 6C). We believe that DMSO treatment is an extra challenge for cells on top of c-Src-DN expression, which therefore display reduced sprouting compared to control and c-Src-WT.

Point 11. There is no statistical paragraph in the method section.

As pointed out by reviewer 1 and 2, we have now added a general section on the statistical methods to the method section on page 18. Additional details on the tests used for each specific graph can be found in the figure legends and Supplementary Table 1.

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

Reviewer 3

Major comments:

Point 5. It is not clear how the constitutive activation of c-Src affects both cell-cell junction and focal adhesion morphology. Did the authors study signaling pathways downstream of c-Src such as the PI3K-AKT pathway?

c-Src is well known to regulate a multitude of signalling pathways, which was definitively shown in analysis by Ferrando et al. using phosphoproteomics (Ferrando, I.M., et al., Mol Cell Proteomics, 2012. 11(8): p. 355-69.) In this manuscript, our primary emphasis is on elucidating the role of c-Src in governing cell-matrix adhesions and the degradation of the extracellular matrix. We delve into the nuanced connection between focal adhesions (FAs) and VE-cadherin through the actin framework in the discussion (see page 10). Additionally, we highlight that beyond its recognised direct targets in FAs and adherens junctions (AJs), c-Src exerts regulatory influence on these structures through its effects on the actin cytoskeleton.

The PI3K/AKT pathway is implicated in the progression of vascular malformations in Hereditary Hemorrhagic Telangiectasia (HHT), where patients exhibit rapid vasculature expansion akin to the observed effects upon introducing the c-Src-CA mutation. In HHT, PTEN inhibition triggers heightened activity of VEGFA/VEGFR2 and subsequent AKT kinase activation. Although we have conducted preliminary analysis revealing elevated phospho-AKT, we contend that an in-depth examination of each signaling pathway perturbed downstream of c-Src-CA is beyond the current scope of this manuscript. Our future studies will specifically address this, providing a meticulous exploration of c-Src activity in HHT and its intricate interaction with the AKT pathway.

Minor comments:

Point 1: General comment: The authors have predominantly presented composite images with overlapping staining, making it challenging to differentiate between different labels. It would be beneficial if the authors could provide individual channel images along with a merge.

Given the large numbers of multi-channel composite images, we believe it is not feasible to show each individual channel of every merged image in the manuscript. We have included individual channel images where we believe is appropriate. For example, p-paxillin Y118 (Figure 2), Fibronectin (Figure 4). We are happy to provide individual channel images for any image, where specifically requested, such as in Figure 1E’’ where VE-cadherin channel was added.

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 study, Essebier et al., investigated the impact of constitutive activation of cSrc on endothelial cell behavior during vascular sprouting and homeostasis. The authors generated various mutant versions of cSrc to enable the expression of wild type cSrc, constitutively active cSrc, or cSrc with a dysfunctional kinase domain in HUVEC. They used a range of in vitro methods, including traditional 2D culture techniques and cutting-edge approaches like microfabricated vessels for 3D cell culture. They showed that the constitutive activation of c-Src resulted in a vascular ballooning phenotype both in a 3D angiogenic sprouting assay and in microfabricated blood vessels subjected to shear stress. The expression of this mutant form of c-Src was associated with an increase of focal adhesion size and number and an increase of extracellular matrix degradation. The vascular ballooning phenotype induced by constitutive activation of c-Src was partially rescued by the pharmacological inhibition of the matrix metalloproteinase (MMPs).

Major:

• "This was further supported by our observation that there were no changes in proliferation in c-Src mutant cells grown in a 2D monolayer".
  • Figure 1A appears to have increased number of cells in the c-Src-CA condition compared to the control condition. Could the authors quantify the number of cells/area as they did for their 3D vessel model? This would reinforce the idea that the ballooning phenotype they observe is not due to differences in proliferation.
  • Would be strengthened with analysis of another proliferation marker, such as EdU label, which is incorporated only during S phase of the cell cycle. Comparing ki67 staining and EdU staining would provide more insights. Also, using their 3D vessel model for this analysis would increase its relevance.
  • In Figure 1E', cells expressing the constitutively active form of cSrc appear to detach, giving the impression of cell death. Have the authors tested the viability/apoptosis of c-Src-CA cells, particularly in their 3D model?
• "Therefore, reduction of endothelial cell-cell contacts in c-Src-CA cells may be due to elevated VE-cadherin phosphorylation and subsequent internalisation", "As reduction in cell-cell junction integrity has been shown to increase migratory capacity and sprouting angiogenesis [38], our data suggest that a balanced control of both cell-matrix and cell-cell junctions is essential for mediating migration." In general, it's not clear how constitutively active cSrc affects focal adhesions and cell-cell adhesion and how this is responsible for their ballooning phenotype. The role of the phosphorylation of the VE-Cadherin and cell-cell junctions in this process is not clear either.
  • Further analysis of cell-cell junctions and focal adhesions (co-staining of phosphorylated paxillin and VE-Cadherin) and focal adhesions/fibronectin (like in figure 4C) in the context of cell migration (scratch wound assay) would provide important information to strengthen this notion of balanced control of both cell-matrix and cell-cell junctions.
  • It is not clear how the constitutive activation of c-Src affects both cell-cell junction and focal adhesion morphology. Did the authors study signaling pathways downstream of c-Src such as the PI3K-AKT pathway?
• "Taken together, these results reveal that proteases produced by c-Src-CA cells are locally secreted at FAs but are membrane bound." The claim that proteases are membrane-bound is not convincingly demonstrated. Could the authors assess whether the constitutive form of cSrc activates the expression of specific genes encoding MMPs by qPCR? Or is it more a matter of the effect of c-Src on the transport of MMPs by microtubules?

Minor:

• General comment: The authors have predominantly presented composite images with overlapping staining, making it challenging to differentiate between different labels. It would be beneficial if the authors could provide individual channel images along with a merge.
• The lab already showed in a previous study that mice lacking c-Src specifically in endothelial cells have reduced blood vessel sprouting, leading to the expectation that the constitutively active form of cSrc would increase sprout number in the sprouting assay. Could the authors explain why the constitutively active form of cSrc induces this vascular ballooning and not an increase in the number of sprouts?
• In Figure 1A, it would be beneficial to include images from orthogonal views. Indeed, in the c-Src-CA condition, it's not clear whether the vascular ballooning observed represents a cluster of cells or an empty space between the bead and the endothelial cells. (Supp movie 1 helps, but it would be useful to add orthogonal views to the figure)
• In Figure 1D, the method used to analyze sprout shape is not clear, especially for the c-Src-CA condition where the number of sprouts is close to 0. The figure legend indicates that this measurement corresponds to the shape of the sprouting area. Could the authors clarify and explain their quantification method?
• "however cells within the vessel still maintained come connections (Fig 1E')": The connections between cells are difficult to see in the images in Figure 1E'. Could the authors provide higher magnification images of the VE-cadherin staining to illustrate these connections between cells?
• "The reduction in migration correlated with an increase in FA size c-Src-CA expressing cells.": Could the authors give more explanation?
• Could the authors widen the cell trajectory trace in Supplementary Figure 3A?
• it is very difficult to distinguish fibronectin fibrils on the images shown in figure 4C. it would be beneficial to change the images.
• "Treatment of ECs with Marimastat in a fibrin bead sprouting assay resulted in a rescue of the ballooning morphology observed in the c-Src-CA cells" Based on the images displayed in the figure and the associated quantifications, it still appears that c-Src-CA+Marimastat induces a vascular ballooning even if it is less pronounced than in the DMSO condition. Hence, it would be more accurate to describe the observed effect as a "partial rescue". In the microfabricated 3D vessel, in the figure 7A, cell-cell junctions still appear altered by c-Src-CA after the treatment with Marimastat, compared to the c-Src-WT-Marimastat, it would be more appropriate to talk about "partial rescue".
• In Figure 6A, it seems that there is a decrease in the number of sprouts in the c-Src-DN condition compared to the control condition after the DMSO treatment, which is not observed in Figure 1, could the authors explain why?
• There is no statistical paragraph in the method section.

Referees cross-commenting

Agree that the comments of the reviews all seem reasonable. Since cultured EC do not retain very specialized characteristics, perhaps repeating experiments with many other ECs would not be helpful, but suggest some key experiments be performed with one other type of EC.

Significance

General assessment:

The authors generated different mutant forms of c-Src and used them in innovative 3D endothelial cell culture models. The vascular ballooning phenotype induced by constitutive activation of c-Src is particularly interesting and impressive, especially as it can be reproduced in 2 different culture models. The model of cSrc inducing extracellular matrix degradation specifically at the level of focal adhesions is compelling, although it lacks rigorous support in the 3D model. Further analysis of signaling pathways downstream of c-Src would strengthen the work. The link and the necessity of a balance between cell adhesion and cell-cell junctions are mentioned and have started to be explored, particularly through the phosphorylation of Ve-Cadherin, and more in-depth analysis would strengthen this aspect of the work.

Advance:

This study provides new insight on the role of c-Src in vascular homeostasis and during sprouting angiogenesis and starts to explore cross-talk between EC cell junctions and focal adhesions. This study also provides new elements crucial for our understanding of vascular malformations and the implication of cell-adhesion to the extracellular matrix in this process. This study may lead to further investigations into the role of c-Src in tumor angiogenesis.

Audience:

Basic research / Specialized

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, Essebier and colleagues have shown that the upregulation of c-Src in endothelial cells results in vascular dilation independently of growth factors or shear stress. The authors have shown that this effect is driven by alteration in the number of focal adhesion and the secretion of matrix metalloproteinases responsible for extracellular matrix remodeling as the inhibition of the MMPs rescues the observed effects.

This is an elegant work, with well-designed experiments and nice images to illustrate them. Congratulations. Nevertheless, the results not really support the conclusions drawn by the authors. The authors have only used one type of vein endothelial cells from one single donor but they conclude that is effect is general for all endothelial cells. Endothelial cells are very heterogeneous, not only depending on their function and localization, vein, artery or capillary, but also between different organs and in disease (PMID: 22315715, PMID: 28775214, PMID: 31944177, PMID: 33514719).

The authors, should either repeat some of the key experiments in other type of endothelial cells, maybe arterial or microvasculature cells which are commercially available or at least state that the observations presented in this manuscript apply to HUVECs and discuss whether this would also apply for other cell types. Minor. Although the methods are well written and can be understood. To improve transparency, the authors should reduce the referring to other papers to describe the methods they perform and at least some kind of brief description should be included.

The authors should report the real p value for their tests. Also when the test is not significant.

Referees cross-commenting

I agree with reviewer #1. Description of the statistical methods should be described in the methods. I have nothing else to add to the comments from the other reviewers.

Significance

The work presented here by Essebier and colleagues is very well designed and performed. The main strength of the manuscript is the study of the molecular mechanism that regulate the relationship between cells and the extracellular matrix. This is not very well studied in the context of disease. Although all the assays have been performed elegantly, the main limitation of this study is that it has been performed in only one type of endothelial cell. For this reason, it is not possible to extrapolate the conclusions drawn to all endothelial cells like the authors do.

This work advances our knowledge of endothelial cell biology and it will be of special interest for the vascular biology and development 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 #1

Evidence, reproducibility and clarity

The manuscript demonstrates the effects of overexpression of c-Src variants in HUVEC endothelial cells. The c-Src kinase interacts with cell adhesion machinery, and the manuscript dissects relationships downstream of c-Src with respect to cellular behavior. Transduced wild type, constituent active, dominant negative c-Src is assayed by sprouting in 3D using a bead system, growth of microfabricated vessels under oscillatory flow, focal adhesion analysis, migration analysis, ECM analysis, and by rescue with a matrix Metalloprotease inhibitor.

Major comments:

1. p5. Fig 1: The sentence that the dominant negative completely abrogated 'this' phenotype implies that the dominant negative was put into the same cells as the constitutively active mutation. 'Abrogated' means it stops the phenotype, and the phenotype in the sentence prior was constitutively active. It is more accurate to say that the dominant negative was not distinguishable from wild type, which is what the statistics show. No double transfection (DN-CA) was performed.
2. p.5. Fig 1: the phenotype of the CA cells is fascinating. They expand far beyond their normal territory, but they are held together in a lacy bubble. To me, this looks like a different phenotype from the ballooning that might occur in an arteriovenous malformation in vivo, as in vivo malformations are continuously covered by cells. I understand why the authors might use the term ballooning but given that the cells expand without continuously touching each other, I do not think this is the correct term. Would blebbing, or radial migration in a lace-like discontinuous pattern describe it better?
3. p. 5. The authors do not describe any relationship to notch signaling. But notch signaling is the mechanism by which a sprout is selected. The CA phenotype shows no selection, and every sprout can continue migration. Did the authors check for any relationship between notch signaling c-Src activation? Does upregulation of C-Src downregulate notch?
4. The statistical methods are not described in the methods (GraphPad?). These need to be added. Are only significant comparisons plotted? In Fig 6 and 7 only pairwise statistics are shown. If all significant comparisons are plotted, then this means that the comparison between the rescued CA and the treated or untreated control is not significant. This can be thought of as a partial rescue towards a wild type, but it is definitely not a full rescue. None of the statistical comparisons in Figure 6 or 7 show significant comparisons to wildtype. This needs more discussion.
5. Mmp activity is inferred, but not measured. This is a limitaion as the assumption is that marimostat acting through the expected pathway.

Minor concerns:

1. Fig 5D. The presentation of the data in this graph is difficult to understand. It is trying to show the proportion of mScarlet in sprouts or balloons a percentage of all the scarlet cells. It would be better to have all cells represented in one bar, distributed between sprout and balloon in that one bar. i.e., for the control and dominant negative, the bars would be all black and then for the CA it would be all white. The zero data points are confusing. A proportions graph should be investigated here.
2. The methods for vessel coverage for quantification in figs 1 and 7 are missing.

Referees cross-commenting

The comments from the other reviewers seem reasonable.

Significance

The work is well executed and takes a mechanistic approach. The images are well put together and the movies significantly add to the manuscript. The phenotype describes highly unusual endothelial behavior, which is of interest, and an advance in the field for its novelty. Linking cSrc to downstream signalling including mmps and demonstrating a rescue is also novel and a strength. This is a conceptual advance in the relationship between a kinase and cell behaviour in 3D.

Understanding this mechanism may be useful in understanding enlarged vessels in vascular malformations, although the direct relevance is not clear due to limitations of using cultured cells in artificial environments, lacking, for instance, support by secondary cells and ECM that might be contributed by support cells and perhaps modulate the phenotype.

The audience would be specialized in the basic research 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

The authors will submit a complete point-by-point response to the reviewer's comments when submitting a fully revised version of the 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 #4

Evidence, reproducibility and clarity

This is a short report featuring an innovative proximity labeling approach to the identification of proteins enriched in distinct types of chromatin domains. The domains compared are centromeric heterochromatin and X-linked hyperactive chromatin in Drosophila cells. These are relatively well-described domains, thus serving as an excellent test for the targeting of biotinylation in the permeabilized nucleus via interaction of specific antibodies with ProteinA-Apex2 provided exogenously. In parallel with the signature chromatin proteins CID or MSL2 as baits, the authors also target proteins in proximity to specific histone tail PTMs. Taking the work one step further, they compare the recovery of proteins +/- pretreatment of nuclei with RNase. They conclude that in each case selective interactions are specifically lost with pre-treatment of RNase.

Major comment:

As mentioned above, the approach is innovative and raises the possibility of a simpler MS method to identify protein-protein interactions. The RNase result is also provocative. However, in each case the specificity of potentially novel results are not explored further. Thus, the work is of interest but clearly still preliminary.

Significance

Did the authors dig deeper into novel interactions without obtaining convincing validation? Did they conclude that the MS approach is worth pursuing further or not? Admittedly the RNase result is difficult to follow up, but additional discussion of prior related work as well as consideration of future experiments would help improve the 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

The manuscript entitled 'The role of RNA in the maintenance of chromatin domains as revealed by antibody mediated proximity labelling coupled to mass spectrometry' by Choudhury et al. describe a new method, which they termed AMPL-MS (Antibody mediated proximity labelling mass spectrometry). The technique is based on proximity labelling but uses antibodies instead of fusion proteins. They use this method to characterize chromatin domains containing specific signature proteins or histone modifications and focus on the composition of chromocenter as well as the chromosome territory containing the hyperactive X-chromosome in Drosophila. Last but not least they include data that show that RNA is involved in maintaining the integrity of chromatin domains by RNAse treatment and mass spec analysis. The technique works well and the results are very clear. I therefore expect that, in the right hands, it is very reproducible.

There are a few points that the authors may want to address:

1. Title 'The' role of RNA in the maintenance of chromatin domains as..., seems too much of a statement. The title is therefore an overstatement that needs to be fixed.
2. Figure 1 In Figure 1 the authors show very convincingly that the methods works well in their hands. They report on 172 proteins that localized in proximity to CID containing centromeric chromatin but do not provide the list of proteins as far as I can tell. Especially the RNA binders should be named.
3. Figure 2 Using the hyperactive X is very clever when addressing RNA function but it should be stated in the discussion that there may be certain aspects that are specific to the male x and that is impossible to discriminate general and specific effects uncovered by this method.
4. Figure 3 The authors should state more clearly the new findings of this figure since it is not fully obvious from its current representation.
5. Figure 4 These are certainly interesting data but the authors remain in the very descriptive state. This is fine for a methods paper but then, the authors should hypothesize more on what the results mean. Are certain RNA dependent factors specific or general and they then recruit a specific set of factors that fall off upon RNAse treatment as a secondary effect or because they bind RNA directly. I feel like there may be more information that they authors got get out of there data than what they currently provide.
6. Discussion The authors state: 'While we have not identified the RNAs responsible for the formation of theses domains, we clearly observe that they do confer specificity for the domains as we observe very little overlap in the factors lost from the corresponding domains (Fig 4h). the 'specificity' is hard to determine since factors bound to these regions are different, and therefore different factors will fall off, regardless of whether the RBP are specific unless the RNA is involved in recruiting the factors specifically, which the authors have not shown. Therefore, this result is suggestive and interesting but the statement is too strong and not backed by their results.

Significance

Overall, this is an interesting method that has been used in the past to identify protein modifications with high quality antibodies available. The authors show here that the method can also be used to different nuclear proteins and detect changes in protein complex composition. As it is it is primarily a methods paper, and for that the results are very clear. Gain of new info is not large but it is a useful technique to continue research on this subjectand is a nice start of many new avenues into how RNA effects chromatin,

My expertises are in epigenetics, chromatin biology, RNA and Drosophila genetics

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 manuscript by Choudhury et al. describes a new method for antibody-mediated proximity labeling and applies it in the cell nucleus. In short, nuclei are isolated, fix and permeabilized, proteins are labeled with primary antibodies, a bacterially expressed/purified protein-A-APEX2 fusion protein is added, conventional H2O2/biotin phenol labeling of proximate protein is performed, proteins are un-crosslinked and biotin-affinity captured for MS analysis. The application to nuclear proteins and results seems appropriate. The method is highly similar to and more complicated than prior methods as described in more detail below. I would focus the impact of this paper towards its biological results and not the novelty of the methods used.

Prior methods that effectively accomplish the same outcome (fixed cells/tissues, antibodies and proximity labeling for AP-MS) have been published before. Perhaps most recently it was reinvented as the so called BAR method in PMID 29256494. That paper was cited here but incorrectly as BirA-related, which it is not. Of course that prior manuscript itself ignored prior methods from years back (2008, 2012, 2014, 2015, PMID 18495923, 22936677, 24706754, 25829300) using the same approaches of antibody targeted peroxidase for the same purposes of proximity labeling.

This method seems a somewhat Rube Golderbergian approach to antibody-mediated proximity labeling, which has been performed previously in multiple reports. APEX/2was developed to function inside of living cells since HRP does not. The value of doing the proximity labeling in living cells was either to capture protein associations over time, as with BioID/TurboID, or to get snapshots of protein associations in living cells with APEX/2. HRP does however function quite well for proximity labeling outside of cells, or in fixed/permeabilized cells, as has been demonstrated in the prior methods/papers that are referenced above. Replacing commercially available secondary antibodies fused to HRP with homemade protein-A-fused to APEX2 seems counterintuitive and/or unnecessary.

Could the authors explain the mechanisms that underly the reported enhanced sensitivity of AMPL-MS compared to conventional APEX2 in living cells. Is there something about the nuclear isolation that reduces interfering background, the loss of small soluble molecules in the nucleus after isolation and/or permeabilization that enhance the proximity labeling, penetration issues with the biotin-phenol in living cells, and/or something else?

There seems to be the use of various controls based on the figures and legends, but they are not clearly described in the results or methods.

All MS results should be provided, preferably in an Excel file format.

Significance

This manuscript by Choudhury et al. describes a new method for antibody-mediated proximity labeling and applies it in the cell nucleus. In short, nuclei are isolated, fix and permeabilized, proteins are labeled with primary antibodies, a bacterially expressed/purified protein-A-APEX2 fusion protein is added, conventional H2O2/biotin phenol labeling of proximate protein is performed, proteins are un-crosslinked and biotin-affinity captured for MS analysis. The application to nuclear proteins and results seems appropriate. The method is highly similar to and more complicated than prior methods as described in more detail below. I would focus the impact of this paper towards its biological results and not the novelty of the methods used.

Prior methods that effectively accomplish the same outcome (fixed cells/tissues, antibodies and proximity labeling for AP-MS) have been published before. Perhaps most recently it was reinvented as the so called BAR method in PMID 29256494. That paper was cited here but incorrectly as BirA-related, which it is not. Of course that prior manuscript itself ignored prior methods from years back (2008, 2012, 2014, 2015, PMID 18495923, 22936677, 24706754, 25829300) using the same approaches of antibody targeted peroxidase for the same purposes of proximity labeling.

This work may be of interest to investigators studying the nuclear proteins/structures to which the APML-MS was applied.

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:

This manuscript describes the application of a new variation of proximity (biotin) labelling (antibody mediated proximity labelling coupled to mass spectrometry, AMPL-MS). Combining protein- or histone variant-specific antibodies with a APEX2-proteinA fusion protein, they characterise the proteome of nuclear subdomains and demonstrate that RNA is important for the integrity of two tested domains, the Drosophila chromocenter and the chromosomal territory of the hyperactive X chromosome.

Major comments:

The vast majority the experimental results, statistical analysis, and conclusions drawn by the authors appear sound and are described in way that should allow reproduction (however, see my comments below for some suggestions for minor improvements). The authors rigorously test their method, using the Drosophila chromodomain as 'playground', before applying it to other chromosomal areas and histone variants/modifications. Besides providing proteomes of the targeted nuclear subcompartments, they show that RNase treatment of the cells radically changes the proteome(s) and conclude a role for RNA in the integrity of the corresponding compartments. This is shown by immunofluorescence staining as well as proteomic analysis of the biotinylated proteins. The images in figure 4b (and to lesser extent 4c) show an increased intensity and more diffuse labelling. Can the authors exclude that RNase treatment simply leads to an increase in accessibility for the biotin-phenol, hence a visibly higher biotinylation? Along these lines, have the authors maybe observed an increase in overall labelling/pulldown efficiency or for biotinylated proteins in their proteomic data?

Minor comments:

1. In figures 1a and 4a (as well as in the Methods section), the authors use the term 'biotin-tyramide' as labelling agent, but in the main text and figure legends 'biotin-phenol' is used. For clarity, only one term should be used.
2. Figure 2a shows a magnified cell/nucleus in the last column. To what cells do the magnifications in this last column refer to? Maybe these cells could be boxed in the second last column?
3. In figures 4b + c:, the figure legend mentions the individual rows as '(I)' and '(II)' but no such label seen in the corresponding panel(s).
4. The Quantification method for co-localization (e.g. 1c and 2b) is insufficiently described to the reader (reference simply relates to Fiji package). What module/script within the Fiji package has been used?
5. The RNase treatment is not described at all in the methods section or the supplementary information and should be added.
6. The sentence on page 6 ('As expected, neither the targeted signature factor or proteins that mainly interact with them protein-protein interactions such as MSL1,3 and MOF for MSL2 or Cenp-C for Cid are not affected by RNAase treatment') should be rephrased as it is not comprehensible in the current form.

Significance

The findings in this manuscript advance our portfolio of proximity labelling techniques although this advancement is not a major step forward. As the authors state themselves, antibody-based proximity labelling has already been introduced, even in the context of chromosomal proteomes (e.g. Gan et al., 2022; https://doi.org/10.1016/j.gpb.2021.09.003). One major technical advance is the finding that modifications or protein variants can now reliably be targeted for proximity labelling, using their method. Furthermore, the number of cells that are required for proximity labelling and detection of biotinylated proteins could be significantly reduced compared to previous approaches (although this might simply be due to the use of a more advanced proximity labelling enzyme). I should state her that as I am not an expert in the field of chromatin domains, I cannot be certain if the proteomes and changes of proteomes the authors report are providing a significant increase in our knowledge on these domains, especially related to their individual functions.

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)):

In the paper entitled GOT1 primes the cellular response to hypoxia by supporting glycolysis and HIF1α stabilization, Grimm and co-authors investigate the metabolic adaptations of cancer cells upon acute hypoxia. By measuring metabolite levels at early time points upon hypoxia, they observe the accumulation of lactate and depletion of aspartate, along with other TCA cycle metabolites. Importantly, they demonstrate that these metabolic changes are independent of the HIF alpha-dependent transcriptional response. The authors investigate the role of aspartate during these initial phases of hypoxia. To this aim, they characterize cells devoid of glutamate oxaloacetate transaminase (GOT1), in which aspartate accumulates and can no longer be used for replenishing NAD+ via the downstream conversion of oxaloacetate to malate, via malate dehydrogenase. These cells have lower cytosolic NAD+ which affects glycolytic flux through the rate-limiting, NAD+-dependent enzyme GAPDH. GOT1 KO cells have a decrease in glucose consumption, lactate secretion and metabolite levels downstream of GAPDH upon early hypoxia, however ATP levels and viability are only affected with additional lactate dehydrogenase (LDH) impairment. Finally, the authors demonstrate that GOT1 KO cells have higher alpha-ketoglutarate (aKG) levels during early hypoxia, which could contribute to higher prolyl-hydroxylation and subsequent degradation of HIF, regulating the transcriptional response mediated by transcription factor.

* Major comments *

1. The authors claim that they were unable to supplement cells with aspartate (Figure S3), (even though an increase of aspartate is instead observed in cells treated with sodium aspartate) and had to resort to the GOT1 knock-out model to "prevent aspartate from decreasing in hypoxia". This approach implicitly assumes that Got1 is the main driver of aspartate depletion upon hypoxia. However, although steady-state levels of aspartate are indeed higher in these cells, there is still a strong decrease upon hypoxia, which the authors acknowledge but merely ascribe to "attenuated production from glutamine". This seems an insufficient explanation, considering the very fast depletion upon hypoxia originally observed. The authors should provide further information regarding why aspartate is depleted in these conditions and consider other aspartate-consuming enzymes such as GOT2, ASNS, or even nucleotide biosynthesis and urea cycle enzymes. These observations could be made using the labeling experiments already acquired. In addition, to corroborate their hypothesis, the authors could supplement 13 C-aspartate at a supraphysiological concentration (i.e. 5-10 mM) to determine to what extent it is consumed by GOT1 or other pathways. > We thank the reviewer for this comment that helped us to recognise, in retrospect, that by focusing on GOT1ko as a means to rescue aspartate levels detracted from our main finding and extensive mechanistic insights into the role of GOT1 in sustaining the increase in glycolysis in early hypoxia. As we detail in our response to the Reviewer’s point 2, we have now re-written our results section to better clarify why we focused on GOT1 (lines 175-223 of the revised manuscript – please note that line numbering corresponds to the word document with the track changes off). However, we also agree that, because the motivation that led us to GOT1 was the counter-correlation between aspartate and lactate, expanding on the pathways that determine aspartate levels in hypoxia would be useful to the reader.

2. To address the reviewer’s point, in revised Fig. S3E, we present new data where we incubated cells in normoxia or hypoxia for 3h in the presence of 1.5 mM 13C-aspartate. We opted for an intermediate aspartate concentration which was enough to observe intracellular labelling while minimising significant perturbation to cells. We found that the amount of labelled aspartate that accumulates intracellularly is not significantly different between normoxia and hypoxia. At the same time, we observe a vast depletion of unlabelled aspartate. We accept that aspartate labelling may not have reached isotopic steady state within the 3h time point we are confined to for our experiments. However, if increased consumption contributed significantly to aspartate depletion within this timeframe, the amount of labelled aspartate that accumulated would be lower in hypoxia compared to normoxia. Therefore, the data in Fig. S3E indicate that, at least within the timeframe of our experiments, the magnitude of aspartate consumption is not likely to increase to such an extent that could significantly contribute to the depletion in aspartate.

We had, indeed considered other aspartate-consuming pathways, however, in light of the above results and our subsequent finding that GOT1 is needed for increased glycolysis, we did not pursue these investigations any further and focused on the role of GOT1 instead.

• In revised Figure S3, and also in response to one of the Reviewer’s other comments below, we have now replotted the data from the experiment in the original manuscript to show both absolute and fractional isotopologue abundances of TCA intermediates from cells labelled with 13C-glucose or 13C-glutamine. Based on these re-plotted data, we find that the amounts of labelled intermediates from both labels decreases; the apparent decrease from glutamine appears greater than that from glucose, likely because glutamine labels more rapidly a greater fraction of TCA intermediates. Moreover, glutamate fractional labelling from glutamine decreases, but modestly increases from glucose over time in hypoxia compared to normoxia. These data raise the possibility that TCA intermediates are diverted to glutamate synthesis. However, as we point out in the revised text, the fact that only glutamine has reached an isotopic steady state by the end of the time course precludes us from making a more accurate quantitative statement and therefore we have refrained from further elaborating on these observations.

Taking the above observations together, in the revised text we do not dismiss increased consumption as a factor in decreased aspartate levels and rather state that “within the timeframe tested, decreased production is a significant contributor to the low aspartate levels in early hypoxia.” (lines 187-188).

In line with the previous comment, the conclusion that "GOT1 activity, rather than a decrease in aspartate concentration itself, is required to sustain the increase in glycolysis in early hypoxia." seems questionable, especially considering the failed aspartate supplementation. The authors suspect low expression of plasma membrane aspartate transporters as the reason and quote Garcia-Bermudez et al.2018 (PMID: 29941933). This paper contains ranked SLC1A2 mRNA expression data from the Cancer Cell Line Encyclopedia (CCLE). The authors may apply aspartate supplementation and "early hypoxia" to a cancer cell line expressing SLC1A2 or other aspartate transporters. Alternatively, they could try introducing the transporter by overexpression.

> We concede that the way we phrased this statement was not ideal and has rightly led to the reviewer’s criticism. In particular, referring to a “decrease in aspartate concentration”, could mislead the reader into thinking that we were referring to the process of aspartate consumption, rather than the low aspartate levels themselves, which is what we aimed to explore. In the revised text, we now carefully make this distinction; we show new data (Figure S3G) supporting the idea that low aspartate levels are not necessary for increased lactate; we explain that, given the known role of the malate-aspartate shuttle in coordinating redox balance and potentially affecting glycolytic flux, the fact that aspartate didn’t appear to be limiting was surprising and we therefore asked whether GOT1, which depends on aspartate, had a role in the increased glycolysis in early hypoxia. Given that GOT1ko attenuated the increase in glycolysis we subsequently focused on the mechanism underlying this observation. In more detail:

As Reviewer 2 noted in point 1 of their review, the increase in lactate became more apparent after 2 h, when aspartate levels had almost reached their minimum. This successive timing of abundance changes raised the possibility that low aspartate levels precede, and possibly drive, the increased lactate. Therefore, we sought to test whether this was the case by preventing depletion of aspartate in hypoxia with exogenous aspartate. We agree that, to address the comment of Reviewer 1 here, overexpression of an aspartate transporter would have been a good way to overcome poor aspartate uptake by MCF7 cells, however, at the time we initiated this study, SLC1A2 was not known as an aspartate transporter. We, therefore, cultured MCF7 cells for several weeks in media containing 0.5 mM aspartate (which is normally absent in our standard media formulation) because we expected that cells would adapt to take up more aspartate. We, thereby, obtained a derivative cell line that we called MCF7Asp. In new Figure S3G, we show that addition of 0.5 mM aspartate in the media of MCF7Asp cells largely prevented the decrease in intracellular aspartate seen in parental MCF7 cells after 3h in 1% O2; however, the increase in lactate was similar between MCF7 and MCF7Asp cells. These data are consistent with the idea that the low aspartate levels in hypoxia are not the likely cause for the increase in lactate.

As the Reviewer notes in point 3 below, production of malate m+1 from 2H-glucose does not decrease below the levels found in normoxia (Fig. 4H), even though aspartate levels are depleted (Fig. 1C). Together with the fact that maintaining aspartate levels to near-normoxic levels does not further boost lactate levels (Figure S3G), these findings speak against the notion that the lack of increased GOT1-MDH1 flux is due to insufficient aspartate and are aligned with the idea that the malate-aspartate shuttle is saturated (PMID: 35973426, 21982705).

• The observation that labelled m+1 malate produced from [4-2H]-glucose is similar in normoxia and hypoxia (Figure 4G), does not support the notion that GOT1-MDH axis is increased at low oxygen and seems to suggest that the depletion of aspartate observed in early hypoxia is unrelated to this axis. The authors should resolve this discrepancy.*

> In our manuscript, we do not claim that the flux through the GOT1-MDH1 axis is increased but, instead, we emphasise the fact that, as the reviewer observed, malate labelling from 2H-glucose is unchanged (e.g. see text in our original manuscript - lines 519-522 of the revised version: “Importantly, a model where increased upper glycolysis due to the Pasteur effect overwhelms GAPDH capacity also elucidates the apparent increase in the reliance of glycolysis on GOT1-MDH1 in hypoxia, even though flux through this pathway is not elevated.”). As we also detail in our responses to comments 1 and 2, above, in the revised manuscript, we have re-written the discussion to better explain that the reliance on GOT1 in hypoxia is not driven by increased flux through this pathway (which is likely saturated as outline in our response to point 2, above), but rather from the increased demand imposed by the elevation in incoming glucose carbons due to the Pasteur effect (lines 504-531). This is akin to a situation where increased demand for a product drives its price up if the manufacturer does not boost production to increase supply. We hope that the reviewed discussion makes this clearer and addresses the reviewer’s comment.

• The alpha-KG level regulation by Got1 and the subsequent HIF1alpha "priming" seem quite promising and likely the most novel part of the manuscript. However, further proof should be added to support this strong claim. First, aKG to succinate ratio, rather than aKG alone, is a better indicator of aKG-dependent dioxygenases activity. So. the authors should provide this measurement. *

In line with the reviewer’s excellent suggestion, in the revised manuscript, we added new panel in Figure 6F (discussed in lines 457-458) that shows αKG levels alongside the corresponding αKG/succinate ratios. These data agree with our original interpretation that cofactor levels in GOT1ko cells favour increased dioxygenase activity.

*Second, the authors should rule out the possibility that the differential hydroxylation of HIF is due to the redistribution of intracellular oxygen due to alterations in mitochondrial function. To do this, they could determine whether cytosolic oxygen levels differ in the two conditions. *

The reviewer raises the interesting hypothesis that, given the decreased respiration in hypoxic GOT1ko cells, one could expect increased availability of oxygen that could contribute to the destabilisation of HIF1α. To the best of our knowledge, measuring absolute cytosolic O2 concentration, particularly in hypoxia, would require specialised equipment [e.g. phosphorescence lifetime imaging (PMID: 26065366), or phosphorescence quenching oxymetry (PMID: 21912692); unfortunately, we do not have access to such equipment. In the revised manuscript, we acknowledge the reviewer’s point with added new text in the discussion (lines 576-577).

Finally, the authors could test whether α-ketoglutarate or 2-hydroxyglutarate supplementation affects HIF stability in their experimental conditions.

> We thank the reviewer for this suggestion. In the revised manuscript (new Figure S6H and lines 453-455) we show that addition of DM-αKG, a cell-permeable form of αKG, to the media of MCF7 cells incubated at 1% O2, decreases HIF1α protein levels in a dose-dependent manner and, at the highest dose, to a degree comparable to that of GOT1ko cells.

Minor comments:

- The glycerol-3-phosphate shuttle is another means of re-oxidizing NADH and α-GP is indeed higher in GOT1 KO. According to this, in Fig 5C a clear increase in a-GP is observed in LDH KO cells. Would the phenotype be stronger upon additional GPD1 knockout or inhibition?

> The main phenotype of combined LDHA/GOT1 inhibition is a deficit in ATP and decreased cell survival. While increased flux through GPD1 could, indeed, provide more NAD+, this would come at the expense of glucose carbons that would otherwise need to flow into lower glycolysis to produce ATP. Consistent with this idea, our data show that, even if GPD1 or other dehydrogenases reoxidise NADH, as would be the case in both the LDHAko and GOT1ko cells where α-GP is elevated, they are not sufficient to compensate for the decrease in LDH and GOT1 activity. Therefore, we did not pursue this hypothesis further.

* - Aspartate and lactate levels appear unchanged in MDA-MB231 upon hypoxia. Can these changes be ascribed to a pseudohypoxic state? The authors should comment on this observation.*

> In Figure S2A, we show that MDA-MB-231 cells have increased basal levels of HIF1α compared to the almost undetectable HIF1α seen in BT474 (same figure, adjacent panel) or MCF7 cells (Figure 2A). We, therefore, agree with the reviewer’s hypothesis that the attenuated changes in aspartate or lactate levels in MDA-MB-231 cells are likely due to a pseudohypoxic state. As this is speculative, we have refrained from elaborating on this point further in the manuscript.

* - Figure S3B: The authors do not provide information on the length of hypoxia for these experiments. *> The data shown in original Figure S3B (new Fig. S3A-B) are a time course. Cells were incubated at 21% or 1% O2 with the respective isotope label for increasing lengths of time, with the longest time point shown (6h) being the longest time we incubated cells in hypoxia. If the reviewer meant another panel, the length of hypoxia would be 3h unless otherwise stated.

- Glucose and glutamine isotopic labelling should be accompanied by graphs showing the total pool levels of these metabolites, and also the uptake of glucose and glutamine (and their specific isotopologue distribution). It would be important to show the isotopologue distribution of aKG in all the conditions tested, in particular, because of its proposed regulation by Got1.

> In the revised manuscript, new Fig. S3 panels A-D, we now show absolute and fractional isotopologue distributions for TCA intermediates for both glucose and glutamine labelling. We have omitted showing αKG in this figure as we could not reliably quantify it in the glutamine-labelling experiment. Also, unfortunately, quantification of glutamine in our GC-MS datasets is not reliable due to conversion to 5-oxoproline.

- Malate generated by MDH1 can be converted by ME1 into Pyruvate, which could be further processed by LDH. Have the authors measured this conversion in their dataset.

> In the figure below we labelled cells with [U-13C]-glutamine for 3 h at 21% or 1% O2 and plotted the fractional labelling for all observable isotopologues in malate, pyruvate and lactate. These data show that there is minimal labelling in pyruvate and lactate (- Aspartate absolute levels across cell lines appear different. Is this due to differences in cell volume? Can the authors comment on this observation?

> To address the reviewer’s hypothesis, we focused on MCF7 and MDA-MB-231, the two cell lines with the highest and lowest aspartate levels, respectively. The volume of MCF7 is approx. 19% higher than that of MDA-MB-231 (calculated based on cell size data from PMID: 31015463). Based on this calculation, and bearing in mind that cell volume is a good predictor of biomass content (PMID: 18595067), cell volume differences may contribute to, but cannot fully account for the one order of magnitude difference in aspartate abundance we see between these cell lines (Figures 1C and S1A).

The cell lines we used in this manuscript (MCF7, BT474, MDA-MB-231, MCF10A) represent different breast cancer (or untransformed, in the case of MCF10A) cell types, with different oncogenic mutation content (PMID: 17157791, 22460905) and proliferation rates (PMID: 22628656); all these factors can be related to steady-state cellular metabolite levels (PMID: 31015463). In the figure below, we have plotted aspartate abundance data (from PMID: 31068703) in 928 cell lines of various origins. These data show that aspartate levels can differ as much as 2 orders of magnitude between cancer cell lines and about half an order of magnitude between MCF7 and MDA-MB-231 or BT474 (MCF10A was not present in this dataset); they also show that aspartate levels in the three cell lines rank in the same order as in our manuscript (MCF7>BT474>MDA-MB-231), although, it is unclear if cells in this dataset were also cultured in dialysed serum as in ours, so we cannot confidently compare the absolute aspartate measurements between our studies.

In conclusion, we suspect that cell volume differences together with other factors, such as proliferation rates and metabolic network differences may account for the differences in intracellular aspartate levels.

- Under hypoxia the contribution of glutamine (labelled fraction, Fig. S3) to TCA cycle intermediates decreases. However, this is not paralleled by an increase in the contribution of glucose, as also supported by an increase in the m+0 in the glutamine labeling but not in the glucose one. How do the authors explain this apparent inconsistency? Are there sources of unlabelled TCA cycle during the hypoxic experiment?

> While glucose and glutamine are the major carbon sources in many cultured cancer cell lines, incl. MCF7 as indicated by the data in Figure S3A-D, other nutrients (such as amino acids, other than glutamine, and fatty acids) can also provide carbons at various points of the TCA cycle. The fact that fractional labelling of glutamate from glutamine is decreased in hypoxia would suggest that the source of decreased contribution of glutamine into the TCA is unlabelled glutamate. We can exclude uptake of exogenous glutamate, because all our metabolic measurements are performed with cells incubated in media without glutamate and supplemented with dialysed serum. However, we observe a modest increase in the fractional labelling from glucose into glutamate (Figure S3A). As glucose labelling into the TCA cycle is not at steady-state even after 5h, it is hard to assess whether, increased labelling from glucose suffices to explain the dilution of glutamine-derived labelling into glutamate a quantitative conclusion but it points to efflux of intermediates out of the TCA cycle (discussed in lines 181-183 of the revised manuscript).

We thank the reviewer for their time and thoughtful comments that helped us improve the presentation of our work.

**Referees cross-commenting**

Referee 2 raises important questions that are in part aligned with referee 1 and are reasonable and doable is the time frame proposed. These are all important questions and comments to consolidate the central hypothesis of the work and I believe are required for publication.

*

Reviewer #1 (Significance (Required)):*

Overall, this is an exciting and well-executed piece of work focusing on the early biochemical consequences of hypoxia that the wide metabolism/biochemistry audience will appreciate. While most of these observations are not entirely unexpected, the work brings a sufficiently novel perspective and insights to the field and deserves publication. However, some conclusions are not fully supported by the data and some additional experiments are suggested to bring clarification and strengthen the authors' conclusions.

We are a lab expert in cancer metabolism.

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

Summary ** This manuscript represents an interesting and novel description of the role of a cytosolic transaminase, glutamic-oxaloacetate transaminase 1 (GOT1) on both cytosolic redox (and therefore glycolysis through its functional linkage with malate dehydrogenase 1) and the availability of alpha-ketoglutarate for stabilisation of HIF1a in hypoxia. Some of the most interesting data are the evidence for increased cytosolic NAD+ regeneration through the combined action of LDHA (known) and GPD1 (less well-described increase in activity in hypoxia). The manuscript as a whole describes the multiple systems required for the early response to hypoxia, but the focus of the title and way the article is written do not entirely reflect this. For example, the title focuses on GOT1 as the enzymes whose activity is responsible for the early response to hypoxia. However, this is not reflected in some of the data - the deuteron labelling in particular - which shows that LDH and GPD1 are responsible for the biggest redox activity (i.e. support of glycolysis). A degree of reframing of the article may therefore be of benefit.

We thank the reviewer for their constructive suggestions. In the revised manuscript, we have re-written the title and the relevant parts of the results section, and we have significantly re-structured the discussion section to reflect the fact that multiple enzyme systems, one of which is GOT1, converge to support the glycolytic increase and cell survival in early hypoxia. Furthermore, in our point-by-point responses, below, we highlight in detail how we have streamlined the way we present our results.

*Major comments. *

In Figure 1 C and D, the data suggest significant changes in the decrease in cellular aspartate between 1-2 hours, which then slow. This is followed by a change in lactate concentrations from 2 hours onwards, which is observed in the cells (D) and media (F). The rapid decrease in aspartate concentration suggests a relatively large change, which does not correspond to the later lack of alteration in deuteron labelling from d4-glucose (Figure 4H-J) in m+1 malate. This therefore suggests that the biggest determinant of decreased aspartate is not coupled to MDH1 activity directly. If the manuscript is focused on the relevance of GOT1 activity to the early hypoxic response, this should be better resolved. Given that this could undermine the strength of the case being made for GOT1 activity playing a significant role (through MDH1), could the authors perform the same experiments but in the GOT1KO cells to show how NADH is handled under these conditions by LDHA and GPD1? If the focus of the manuscript is shifted, these experiments would likely not be necessary.

> We thank the reviewer for these comments, which, together with those by Reviewer 1, highlighted that the way we presented our results warranted improvement. First, we would like to clarify that by referring to a “decrease in aspartate concentration”, we may have misled the reader into thinking that we were referring to the process of aspartate consumption; rather we wanted to explore whether the low aspartate level itself could be causing the increase in lactate. This is because, as the Reviewer points out, the rate of lactate accumulation picked up after aspartate had almost reached its minimum. Furthermore, by not elaborating on the cause of decreased aspartate and by focusing on GOT1ko as a means to rescue aspartate levels implied a hypothesis whereby GOT1 was the main aspartate consumer, thereby detracting from our main finding and extensive mechanistic insights into the role of GOT1 in sustaining the increase in glycolysis in early hypoxia (regardless the contribution of GOT1 activity in the observed depletion of aspartate).

In the revised text, we have re-written parts of the results section to better clarify these points (e.g. lines 175-223 - please note that line numbering corresponds to the word document with the track changes off). In summary, and as detailed below, we explore the glucose and glutamine data further and present new data with 13C-Asp, which, together support the idea that decreased aspartate in early hypoxia is largely attributable to decreased synthesis and, to a lesser extent, if at all, to increased degradation. We then explain that, given the known role of the malate-aspartate shuttle in coordinating redox balance and potentially affecting glycolytic flux, we asked whether GOT1, which depends on aspartate, still had a role in the increased glycolysis vis-à-vis the low aspartate levels in early hypoxia. Given that GOT1ko did attenuate the increase in glycolysis we subsequently focused on the mechanism underlying this observation. We have re-structured the discussion, to highlight that GOT1 is one the multiple systems required for survival in early hypoxia. We also explain that the reliance on GOT1 in hypoxia is not driven by increased flux through the GOT1-MDH1 axis (which is likely saturated), but rather from the increased demand imposed by the elevation in incoming glucose carbons due to the Pasteur effect (lines 504-531). A relatable situation is when increased demand for a product drives its price up if the manufacturer does not boost production to increase supply. We hope that the revised text better clarifies these points.

Below, we detail the new experimental evidence/analyses we referred to above:

• In revised Figure S3A-D, we have now replotted the data from the experiments in the original manuscript to show both absolute and fractional isotopologue abundances of TCA intermediates from cells labelled with 13C-glucose or 13C-glutamine. Based on these re-plotted data, we find that the amounts of labelled intermediates from both labels decreases; the apparent decrease from glutamine appears greater than that from glucose, likely because glutamine labels more rapidly a greater fraction of TCA intermediates. Moreover, glutamate fractional labelling from glutamine decreases, but modestly increases from glucose over time in hypoxia compared to normoxia. These data raise the possibility that TCA intermediates are diverted to glutamate synthesis. However, as we point out in the revised text, the fact that only glutamine has reached an isotopic steady state by 5h precludes us from making a more accurate quantitative statement and therefore we have refrained from further elaborating on these observations.

• In revised Fig. S3E, we present new data where we incubated cells in normoxia or hypoxia for 3h in the presence of 1.5 mM 13C-aspartate. We found that the amount of labelled aspartate that accumulates intracellularly is not significantly different between normoxia and hypoxia. At the same time, we observe a vast depletion of unlabelled aspartate. We accept that aspartate labelling may not have reached isotopic steady state within the 3h time point we are confined to for our experiments. However, if increased consumption contributed significantly to aspartate depletion within this timeframe, the amount of labelled aspartate that accumulated would be lower in hypoxia compared to normoxia. Therefore, the data in Fig. S3E indicate that, at least within the timeframe of our experiments, the magnitude of aspartate consumption is not likely to increase to such an extent that could significantly contribute to the depletion in aspartate.

Together with the data in Fig. S3A-D, these findings suggest that decreased aspartate in early hypoxia is to a great degree driven by decreased production.

• The authors present data in Figure 1 and 3 using 2DG as a surrogate for glucose uptake. 2DG has been previously shown not to always be a surrogate for glucose uptake (Sinclair et al. Immunometabolism 2020). Given that this paper highlighted warns in particular about assuming SLC2A1 and SLC2A3 activities based on 2DG uptake, and that these two transporters are the major glucose transporters regulated by hypoxia, a cautious approach to these data is recommended. Assuming that 2DG uptake is a surrogate for glucose in this system (panel C), the effect of GOT1 appears to be at the level of glucose uptake even at 3 hours - it has been marked as being significant by the authors. This suggests that loss of GOT1 has an effect on glucose uptake prior to any transcriptional response is observed. Is the plasma membrane occupancy by the SLC2A1 or SLC2A3 been reduced after GOT1 KO? The same is true for Figure 1 - as intracellular aspartate and lactate and extracellular lactate is shown, could change in extracellular glucose not be presented as a direct measure?*

The reviewer raises two points: (a) that using 2DG may not faithfully report transporter-mediated glucose uptake and (b) that, if our observations with 2DG are valid, they could point to the possibility that attenuation of glycolysis in GOT1ko cells may be attributable to effects in glucose uptake. In brief, we cannot use glucose measurements in media as an indicator of glucose uptake rates because we do not observe measurable glucose depletion from media within the relevant timeframe (3h) of our experiments.

(a) Given that we did not have access to a set up for using radionuclides, we explored both 2DG-based and glucose depletion from media as potential means to assess glucose uptake. We found that, over 24h, MCF7 cells deplete glucose faster than cells incubated in normoxia for the same amount of time (figure below, A). The magnitude of this increase is similar to that we report using 2-DG (~3-fold, Fig. 1E and 3C). However, we observed only minimal depletion of glucose in the first 3-5 h of culturing cells with fresh media (figure below, B). This is perhaps not surprising given that studies that look at metabolite exchange rates (incl. glucose) typically sample over a period of one to several days rather than hours (e.g. PMID: 31015463, 22628656). In conclusion, we reasoned that detecting a positive change in signal (intracellular 2DG) would provide a more sensitive means than a decrease in extracellular glucose to enable assessment of glucose use within the early time-points that our manuscript is mainly concerned with.

(b) Indeed, we were initially intrigued by the decrease in glucose uptake by GOT1ko cells as it could explain decreased lactate production. However, the upregulation of upstream glycolytic intermediates in GOT1ko cells in both normoxia and hypoxia (Figure 4A) together with the evidence of increased α-GP production from glucose (Figure 4K-L) suggested that, even if less glucose is taken up by GOT1ko cells, there is still a bottleneck at the GAPDH step that prevents maximal flow of glycolytic intermediates to lower glycolysis. We therefore did not pursue further the cause of decreased glucose uptake by GOT1ko cells at this stage.

• The data shown in Figure 2D suggests that there is little change in overall contribution to citrate from glucose in hypoxia compared to normoxia, and that HIF1 is does not play a role in the hypoxic response at this point. However, the data presented are overall fractional labelling, and therefore do not focus on the main hypoxia-dependent point of control highlighted before this by the authors - pyruvate oxidation through PDH. Could the authors consider plotting m+2 isotopomer of citrate either alongside or instead of the total fractional label (which includes hypoxia-independent PC activity and cycling carbons). *

We agree with the reviewer’s suggestion. In the revised manuscript, we added a new panel in Fig. 2D that shows the m+2 citrate isotopologue alongside the original fractional labelling data. This new panel is shown as a bar graph to enable the presentation of individual datapoints and statistical test results.

Additionally, the experimental set-up means that average incorporation over the time shown is represented - i.e. the 3h timepoint is incorporation over the first two hours, while the 24 hour timepoint is averaged over the whole period. It is therefore likely under-representing the decrease in glucose contribution to citrate at 24 hours - the authors could point this out, or OPTIONALLY perform a more time-resolved experiment where flux over shorter periods is assessed for each of the timepoints (i.e. 0-1, 2-3, 5-6, 23-24).

Indeed, we did consider a more time-resolved labelling experiment as the reviewer suggests, however, we decided against this approach as we were concerned that even if we pre-equilibrated the labelling media in hypoxia, it would be challenging to avoid perturbations associated with handling of the cells during addition of the isotopically labelled compound. The new panel in Fig. 2D that shows absolute citrate m+2 abundances should address this point, however, in the revised text (lines 162-164) we added new text that points out this issue.

• Figure 3 data are key for the GOT1 theme of the manuscript, as the authors show that loss of GOT1 increases cellular aspartate in both normoxia and hypoxia - suggesting that GOT1 is an aspartate-consuming enzyme in both conditions. Indeed the magnitude of the change in aspartate after GOT1 knockdown appears similar in both conditions (Panel B). These are interesting data, as they contrast with a recently published study (Altea-Manzano et al. Molecular Cell 2022) suggesting that in respiration-deficient cells (a condition with parallels with hypoxia), GOT1 activity may be aspartate producing to supply aspartate to the mitochondria for GOT2. It would be important for the authors to discuss the differences between studies.*

Following the reviewer’s suggestion, in the revised manuscript (lines 547-556), we have now expanded our previous discussion on the functions of GOT1 in cells with respiration defects.

• Panel E shows data at 5 hours, while the rest of the panels here are a mix of 1 and 3h timepoints. Equally panel E also presents concentration, while D presents relative abundance of lactate - could a consistent approach to presenting the results be taken?*

We agree. Taking into consideration that the data in this panel show one time point of the full time-course in Figure S3F, and to streamline the presentation of these data, in the revised manuscript, we have moved the time-course graph to the main figure.

• In Figure S3, the authors show the lack of direct aspartate uptake, or supplementation through the use of an esterified form. OPTIONAL: they could consider using the expression of SLC1A3 (Tajan et al. Cell Metabolism 2018; Hart et al eLife 2023) to increase aspartate uptake in order to test their hypothesis. *

We agree that, to address this point, overexpression of an aspartate transporter would have been a good way to overcome poor aspartate uptake by MCF7 cells, however, at the time we initiated this study, SLC1A2 was not known as an aspartate transporter. We, therefore, cultured MCF7 cells for several weeks in media containing 0.5 mM aspartate (which is normally absent in our standard media formulation) because we expected that cells would adapt to take up more aspartate. We, thereby, obtained a derivative cell line that we called MCF7Asp. In new Figure S3G, we show that addition of 0.5 mM aspartate in the media of MCF7Asp cells largely prevented the decrease in intracellular aspartate seen in parental MCF7 cells after 3h in 1% O2. However, the increase in lactate was similar between MCF7 and MCF7Asp cells. These data are consistent with the idea that the low aspartate levels in hypoxia are not the likely cause for the increase in lactate.

*Figure S3B-E - the authors suggest based on these data that aspartate decrease in hypoxia is through decreased glutamine contribution. Indeed they could also interrogate the data further, as the defect is observed in glutamate, perhaps suggesting that glutamine metabolism through glutaminase is altered. *

To address the Reviewer’s point, in revised Figure S3, we have now replotted the data from the experiment in the original manuscript to show both absolute and fractional isotopologue abundances of TCA intermediates from cells labelled with 13C-glucose or 13C-glutamine. We have elaborated on these results in our response to point 1, and we re-iterate our conclusions here for the Reviewer’s convenience: Based on these re-plotted data, we find that the amounts of labelled intermediates from both labels decreases; the apparent decrease from glutamine appears greater than that from glucose, likely because glutamine labels more rapidly a greater fraction of TCA intermediates. Moreover, glutamate fractional labelling from glutamine decreases but modestly increases from glucose over time in hypoxia compared to normoxia. These data raise the possibility that TCA intermediates are diverted to glutamate synthesis. However, as we point out in the revised text, the fact that only glutamine has reached an isotopic steady state by 5h precludes us from making a more accurate quantitative statement and therefore we have refrained from further elaborating on these observations.

*Figure S3D and E - the authors show data from 3 hours of labelling, which is not at steady-state (observable from the timecourse also shown in B and C). To be able to compare the glucose and glutamine labelling, a timepoint in which (pseudo)steady-state is achieved would be better chose. *

In the revised manuscript, this concern is now addressed by showing both absolute and relative isotopologue abundances for all available time points. We agree that quantitative comparison of labelling must be done at steady-state conditions, however, as we also point out in the revised text (lines 180-181), only glutamine reaches isotopic steady state by 5h whereas glucose hasn’t.

Additionally, within the aspartate isotopomers arising from glutamine, there is an odd m+1 for aspartate not observed in the other proximal metabolites. Is this a technical defect or is there a biological reason for the significant fractional amount in normoxia?

We thank the reviewer for pointing this irregularity, which we should have clearly identified as such during proofreading of the manuscript. Probed by the reviewer’s comment, we reviewed the corresponding data tables used to plot these data and found that M+1 had exactly the same values as M+0. We then inspected the original data and confirmed that this resulted from an error during the copying of the data from the R-script output data table to GraphPad Prism for plotting (the line containing the replicates for the m+0 isotopologue was pasted again in the line of the M+1 isotopologues). This issue is now obsolete, as, in the revised manuscript Fig S3 new panels A-D, we have replaced the fractional data with detailed absolute and fractional labelling showing all isotopologues. We apologise for this error.

• Figure S6F - all samples from GOT1 KO cells have less actin - could an appropriately loaded western blot be presented?*

In the revised manuscript, we added a new panel with the Ponceau (27/02/2018) staining of the same membrane used for immunoblotting. This staining shows equal loading between all lanes. It is unclear why despite equal loading, the actin signal differs between the two lines.

• In Figure S5B, the authors present ATP data in wild-type control cells, and LDHA-KO with LDHA re-expression. These should be phenotypically similar, but clearly are not. It suggests that there is something not correct with the system being used.*

As shown in the western blot of this figure, expression of exogenous LDH only reaches a fraction of endogenous levels, which likely explains the partial, albeit significant, rescue of the ATP depletion observed in the LDHAko cells. We have not been able to achieve higher LDH expression in our cell preparations that would enable us to address this point further.

*

*

Minor comments

1. • PHDs need iron, alpha-ketoglutarate, oxygen and critically ascorbate (Introduction page 2)*

     We thank the reviewer for highlighting this critical omission. In the revised manuscript, we have now added this information (line 58).

* PDK1 phosphorylation of PDH leads to a reduction in pyruvate oxidation, rather than entry of glucose carbons to the TCA cycle (Introduction page 3)*

We agree with the reviewer that our wording was not accurate, and, in the revised text, we have re-written this part (lines 72-74): “…[PDK1] catalyses the inhibitory phosphorylation of pyruvate dehydrogenase (PDH), leading to attenuated pyruvate oxidation and, consequently, decreased contribution of glucose-derived carbons into the tricarboxylic acid (TCA) cycle.”

* SLC25A51 has been identified as being required for NAD transport into the mitochondria (Kori et al. Science Advances 2020), so it is incorrect to say that the inner mitochondrial membrane is impermeable to this metabolite (page 7)*

We agree that, in light of the Kori et al. study, the phrasing in our text presented an outdated view of pyridine nucleotide compartmentalisation. The data in Kory et al. support SLC25A51 as a mitochondrial NAD+ transporter, however, it is not clear if NADH is also a substrate. Furthermore, as the authors also point out, SLC25A51 has a relatively low affinity for NAD+ and therefore unlikely to interfere with the functions of the malate-aspartate shuttle. Taking all this into consideration, in the revised text (line 249), we acknowledge the existence of a low-affinity mitochondrial NAD+ transporter and retained the statement about impermeability specifically for NADH.

* Figure S6D - authors shows a highly significant increase in the mRNA for EGLN3, which is a HIF1 target gene, as well as encoding PHD3, which acts to hydroxylate HIF1a alongside PHD2. This should be commented on in the text.*

In the revised discussion (lines 577-578), we acknowledge that increased PHD3 (together with increased oxygen availability, related to Reviewer 1’s comment), may additionally contribute to HIF1α destabilisation. Please note that we have also added new data (Figure S6H) in response to Reviewer 1, where we show that exogenous αKG causes HIF1α destabilisation in hypoxia, further supporting the notion that boosting intracellular αKG, alone, can destabilise HIF1α.

* Figure S5G - could it be made clear on the graph whether this is at 21% or 1% O2?*

We thank the reviewer for pointing out this omission. We now state clearly both in the revised corresponding legend (line 937) and revised figure that these data are at 1% O2.

• Figure 5I shows ATP level against % labelling of alpha-GP. It isn't clear whether this is abundance or fractional label, but if the latter this it potentially misleading, as if the concentration of alpha-GP increases as fractional label decreases, there is effectively no change. Could the authors extract the steady-state data from the analysis and use this to calculate amount of m+3 label instead of fraction? Similarly for Figure S1H showing fractional labelling of lactate from glucose. It is likely that the title of this graph is a typo, and that m+3 instead was meant. Additionally, measurement of fractional labelling does not demonstrate increased concentrations of the metabolite, but the glucose carbons making up this isotopomer in the pool.*

For Figure 5I, we confirm that what we show is based on abundance of α-GP m+3 labelling from glucose and, in the revised manuscript (line 895), we amended the legend to clarify this important point.

We concede that the way we had originally written this sentence, suggested that we derived our conclusion that increased lactate in media was due to increased glycolysis based solely on the fractional data in Fig. S1H. In the revised manuscript, we have re-phrased the relevant sentence (lines 136-137) to indicate that our conclusion is based on the fractional data, together with the total lactate data that we show in Fig. 1F.

For all our GC-MS experiments we used ions that we detected reliably in all our experiments – in the case of lactate this is m/z 117. This is a 2-carbon fragment as indicated in the original legend; the molecular formula of the derivatised fragment is shown in Table S2. In the revised manuscript (line 671) we clarify that this fragment contains carbons 2 and 3 of lactate (which we concluded from experiments where labelling with 3,4-13C-glucose (which labels lactate at C1) led to partial decrease in this isotopologue); therefore changes in 117 m+2 indicate changes in glycolysis rather glycolysis and the PPP.

* Figure S2G - the purpose of the measurement of cysteine is unclear; measurement of NAC directly within cells would be a clearer demonstration of its uptake, and to demonstrate direct contribution to antioxidant response would instead require measurement of cellular antioxidants rather than cysteine itself.*

We agree with the reviewer’s comment that, ideally, we would have measured antioxidants, however, unfortunately our GC-MS experiments do not detect glutathione; we, therefore, opted to show cysteine as the best available proof that NAC was added to these cells from the same experiments where we measured aspartate and lactate.

* There is no Figure S3F (page 6 of text)*

In the original version of our manuscript we had awkwardly placed Figure S3F at the top right side of the figure due to space limitations, so, understandably, the reviewer may have missed it. In the revised manuscript, we have now moved this panel to the main Figure 3E, to also address the reviewer’s point 5, above (presentation of lactate data).

* Figure 2E, lactate excretion into the media is presenting an odd profile, suggesting that between 3 and 6 hour there is uptake by cells. Equally, the 24 hour timepoint is being presented as p

The overlap of the error bars arises from error propagation as we report the values at each time point relative to t=0h. The statistical difference we reported was calculated on the original values at 24 h alone, so to avoid this discrepancy we have opted for removing the results of this statistical test altogether.

*

Reviewer #2 (Significance (Required)):*

* The data throughout this paper provide some strong evidence for an early and likely HIF-independent metabolic response - while this is understood, detailed studies have not been performed into the various redox balancing cytosolic pathways, which are presented here. The focus on GOT1 is also interesting and novel, but represents part of a larger overall picture presented, which is not reflected in the title.*

* This is suitable for a relatively broad audience, as the phenotype is likely not cancer specific.

*__Reviewer #3 (Evidence, reproducibility and clarity (Required)): __

* Here, Grimm and colleagues investigate the immediate cellular response to hypoxia, prior to onset of HIF1a stabilization/activity. Consistent with established findings they describe that glycolysis is rapidly upregulated under hypoxia, in a HIF1 alpha independent manner, this correlates with an decreased aspartate levels. From this basis, they describe a key role for GOT1 activity in regulating the early hypoxic response, demonstrating its requirement for glycolysis, maintaining the NAD/NADH balance and - in combination with LDHA - maintaining ATP homeostasis in hypoxia. Finally they describe a role for GOT1 (though alpha KG depletion) in contributing to HIF1 alpha stabilization.*

* In sum, the authors present a compelling study investigating the mechanistic basis of early response to hypoxia, placing GOT1 as a key metabolic regulator of this response. The question of how cell metabolically adapt in the short term to hypoxia is, in my view, an often overlooked area of investigation but clearly has importance across biology, not least in cancer biology - thus the area of investigation is topical. The authors conclusions are supported by their data, often in multiple cell lines and/or through orthologous methods. I would support publication of this study as is.*

* Reviewer #3 (Significance (Required)):*

* Significance is stated in my review above, an understudied area of investigation (early hypoxic responses) but clearly important since without a transient response, the long-term impact of HIF1 stress responses would not be possible*

We thank the reviewer for their time assessing our manuscript and for their positive feedback.