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 their manuscript the authors show the involvement of the AAA ATPase Atad1 in Desmin degradation. They identify PLAA and Ubxn4 as partners of Atad1 that participate to its function in desmin degradation.

A general comment is that some conclusions are overstated. The authors mention several times that Atad1 depolymerises desmin filaments. The data show that Atad1 participates to the degradation of Desmin and to its solubilization. "Depolymerisation" should be kept for the model presented in figure 8 but not used in the result section. Major comments:

1. It would be useful to visualize Atad1 and partners localization in muscle fibers in immunofluorescence. Do they colocalize with desmin filaments, with calpain?
2. In the same line, interactors were obtained from large crosslinked complexes. It would make the model more convincing if direct interactions with Atad1 were shown, for example using Proximity Ligation Assays.
3. Evaluation of atrophy is made on cross-sections of muscles electroporated with shRNAs. Histology pictures should be shown.
4. What is the percentage of electroporated fibers? To evaluate the effect of shRNAs it is important to have this information. For example, if the efficiency is 50% it means that the reduction in expression of the target in electroporated fibers is twice the value reported for the whole muscle. Alternatively, immunofluorescence could be provided to see the decrease in targeted proteins in electroporated fibers.
5. The same is true for all the experiments quantifying the effect of shRNAs in western blot. Since quantifications are probably made on whole muscles (ie a mix between electroporated and non electroporated fibers) and since the percentage of electroporated fibers is not given it is not possible to estimate the efficiency of the shRNAs in electroporated fibers.
6. Figure 2C: by decreasing solubilization of desmin, one would expect a decrease in the levels of soluble desmin. Conversely the authors observe an increase in both insoluble and soluble desmin. Of course, this can be explained by reduced desmin degradation once solubilized but this should be demonstrated at least by showing that UPS inhibitors induces an increase in soluble ubiquitinated Desmin.
7. Figure 2E: the levels of Atad1 in the insoluble fraction seem to be the same in the shLacZ and GSK3DN conditions, whereas the phosphor Ser is different. In other words, there should be more Atad1 in the insoluble fraction with shLacZ than with GSAK3DN since the phosphorylation level with shLacZ is significantly higher.
8. Figure 4E: the authors state that phosphorylation decreases because of increased degradation (lanes 6-8). However, Calpain also increases degradation and phosphorylation is increased (lanes 2-4), so increasing degradation does not systematically cause a decrease in phosphorylation. Similarly, lane 5 Atad1 induces less degradation than Calpain, however, it causes a decrease in phosphorylation. Explain.
9. The AAA ATPase VCP shares partners with Atad1 and is involved in muscle atrophy. It would greatly add to the manuscript if the authors inhibited VCP to compare its effect to Atad1

Minor comments:

1. The soluble fraction contains a large number of ubiquitinated proteins. Please explain how it can be stated that an increase in total soluble polyubiquitinated proteins corresponds to an increase in ubiquitinated desmin.
2. Page 11: the authors conclude that denervation enhance the interactions with Atad1. Figure 3D indeed show an increase for Ubxn4, but it is not clear for the other proteins.
3. Figure 4 F: show muscle sections
4. Page 21 in vivo transfection: it is stated "see details under immunofluorescence" but there is no immunofluorescence section in materials and methods.
5. The authors show that Atad1 inhibition in innervated muscle is sufficient to induce muscle hypertrophy (Figure 4E). They conclude that the hypertrophic effect of Atad1 is due to the inhibition of Desmin degradation. However, this hypertrophic effect could be independent of the action of Atad1 on Desmin.

Significance

This is new information in the field since calpain cannot hydrolyze desmin insoluble filaments and that the mechanisms that give calpain access to desmin are not known.

The authors already made important contribution in the study of muscle atrophy and especially in desmin degradation. This work constitutes a new advance in their attempts to understand the molecular mechanisms leading to desmin degradation and muscle atrophy.

Audience: desmin is the main intermediate filament in skeletal muscle. This work will therefore interest scientists working on skeletal muscle.

Expertise of the reviewer: molecular and cellular biology of skeletal muscles, muscle atrophy.

Referee Cross-commenting

I fully agree with reviewer 1.

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 reports the identification of a novel protein complex involved in denervation-induced desmin degradation. The first protein to be identified was the ATPAse Atad1. A clever isolation strategy was based on the fact that the ATPAse p97/VCP is involved in the extraction of ubiquitinated myofibrillar proteins but is not required for the removal of ubiquitinated desmin filaments. The authors reasoned that a related ATPAse might be specifically required for desmin filaments. Atad1 was identified by treating desmin filaments with a nonhydolyzable ATP analog and looking for ATPases that are associated with desmin filaments by proteomics. Knockdown of Atad1 causes a loss of desmin degradation and led to a loss of denervation-induced muscle atrophy. It seems that Atad1 binds desmin in a phsphorlation-dependent manner, although the binding maybe mediated by a protein that hasn't yet been identified. The authors went on and identified two additional proteins which together with Atad1 form a protein complex involved in recruiting calpain for desmin degradation.

Overall, this study is very convincing providing novel important insight. I have only some minor comments

Minor comments

1. I wondered whether Aatd1 is expressed at higher-than-normal levels in muscle and heart. I looked that expression pattern up and it seems that they are especially abundant in muscle and heart and expressed at lesser levels in smooth muscle and overall have a restricted expression. Maybe you have some data on their expression in muscle tissue. Did you perform some staining of muscle tissue at baseline and after denervation with regard to the protein localization by immunostaining?
2. The string data presented in Figure 3C needs some further explanation with regard to the colors used for the different proteins. While the authors explained the meaning of the proteins labeled in red, there is no explanation for the other colors.
3. Molecular weights in Fig. 2E, 3D needs to be 'repaired' and additional MW information is required in case of the ubiquitin blot shown in 3D.
4. Fibre size distributions shown in Fig. 1D and 4F. Have the differences been statistically tested?
5. For my taste the referral to the individual data (Fig. numbers) in the discussion section is too detailed and becomes a second results section. This should be substituted by a summary paragraph before the implications are discussed.
6. The summary slide is very good. However, could you please add information, which protein of the three in the Atad1 complex is depicted by each symbol?

Significance

Novel insight into the proteins involved in desmin filament degradation. Since this is an important subject both in muscle and heart and plays an important role in muscle and heart disease, it is of significant clinical importance. Currently it has only been implicated in denervation-induced skeletal muscle atrophy, but it is likely that desmin filament metabolisms is also similarly regulated in the heart.

I am a researcher mainly focusing on the cardiac biology with some expertise also on muscle, however no specific knowledge about desmin filament biology.

Referee Cross-commenting

Overall, I think all three reviewers agree that this is a significant and important paper. I think that the comments made by the reviewers are fair and probably add to the quality of the manuscript.

Thus, both myself and reviewer 2 agree that it would be useful to visualize Atad1 and partners localization in muscle fibers by immunofluorescence. These data would provide independent support to the model the authors are proposing, which currently is only based on biochemical analysis.

I also support the proposed use of proximity ligation to provide further evidence of the presence of the Atad1, Ubxn4 and PLAA in a complex. However, this experiment depends on the quality of the available antibodies and I would consider this not absolutely required.

I also agree that some further information on the proteomics data (as suggested by reviewer 3) is required with regard to the method of filtering for UPS components was performed.

The proposed request for further information on the electroporation approach is a valid comment and if the authors have this information, it would be good to provide. However, I do not recommend further experiments as overall the data are very consistent and the findings are very significant and represent a major advance in our understanding of desmin degradation.

Note: This rebuttal 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

RESPONSE TO REVIEWER #1:

We wish to express our appreciation to Reviewer #1 for his or her insightful comments, which will significantly improve this paper. We thank the reviewers for giving us the opportunity to improve the manuscript. We have responded to all the comments pointed out. The revised sections are highlighted in red characters and yellow backgrounds in the preliminary revised manuscript.

Reviewer #1 (Evidence, reproducibility and clarity (Required)): This manuscript "Histidine-rich protein 2: a new pathogenic 1 factor of Plasmodium falciparum malaria" by Iwasaki, et al. reports effects of recombinant HRP2 protein on various mammalian cell lines. The MS clearly demonstrates that recombinant HRP2 enters into HT1080 cells, causes inhibition of autolysosome fusion, increases lysosomal Ca ion concentration and reduces general autophagic degradation. The authors also show that the presence of FBS or metal chelators like EDTA and EGTA mitigate toxicity of HRP2, as the former traps HRP2 and the latter compete with HRP2 for Ca binding. The experiments are appropriately carried out with suitable controls in most of the cases. There are some concerns as listed below:

**Major concerns:** 1.HRP2 has been shown to be associated with virulence and causes vascular leakage, particularly cerebral malaria (references 37 and 38 ). Plasmodium falciparum histidine-rich protein II has been demonstrated to exacerbate experimental cerebral malaria in mice, which has been proposed to be associated with vascular leakage, activation of inflammasome and cytokine production (references 37, 38 and PMID: 31858717). This study complements the previous findings of the effect of HRP2 on mammalian cells. However, this study reveals another mechanism by which HRP2 might cause toxicity, which is inhibition of general autophagy and increase in lysosomal Ca concentration. However, whether these in vitro effects would translate in vivo needs to be shown.

Response: We sincerely appreciate the reviewer's effort to evaluate our work. As the reviewer pointed out, this is an in vitro study, so further in vivo validation is essential in the future. However, it is also true that we discovered new findings that have been overlooked because we conducted an artificial and simple in vitro experiment. In the future, it is necessary to demonstrate the cytotoxicity, autophagy inhibition, and lysosomal calcium concentration variation of PfHRP2 by in vivo studies using model animals. Concretely, we need to confirm whether PfHRP2 behaves as a similar virulence factor in vivo by animal experiments using PfHRP2-administrated or PfHRP2-overexpressing/deficient P. falciparum-infected mouse models. These future tasks have been added to the Discussion (page 9, lines 294–297 and 309–310; page 10, lines 339–342). We have also added the study (PMID: 31858717) reporting PfHRP2 elicits pro-inflammatory effect and induces vascular permeability as reference 40.

Furthermore, the title of the original paper was vague and gave the impression that it included in vivo experiments. Therefore, to avoid misunderstanding, we modified the paper's title to be more concrete, "Plasmodium falciparum histidine-rich protein II exhibits cell penetration and cytotoxicity with autophagy dysfunction".

Reference

P. Dinarvand, L. Yang, I. Biswas, H. Giri, A. R. Rezaie, Plasmodium falciparum histidine rich protein HRPII inhibits the anti-inflammatory function of antithrombin. J. Thromb. Haemost. 18, 1473–1483 (2020).

2.All the experiments are done with recombinant HRP2 and BSA as a control. The authors should show if similar effects happen with infected parasites.

Response: As the reviewer pointed out, it is required to perform in vivo experiments, i.e., to clarify whether the same phenomenon observed in the present study occurs in PfHRP2-administrated or P. falciparum-infected mouse models. However, in vivo studies are not possible immediately because we do not have the research facilities to carry out in vivo experiments. Therefore, we have added statements (page 9, lines 294–297 and 309–310; page 10, lines 339–342) to emphasize that the present findings are limited to in vitro and that further in vivo studies described above will be required in the future.

3.HRP2 is released in circulation, making it accessible to endothelial cells and immune cells. How would it reach to the equivalents of these cells in the human body?

Response: Since PfHRP2 induces vascular permeability as described in References 37–40, we propose that PfHRP2 can reach and contact cells in the human body after causing vascular leakage. I have added this possibility of contact between PfHRP2 and cells in the human body to Discussion (page 9, lines 287–290).

**Minor concerns** 1.p62 is an appropriate marker to assess autophagy cargo degradation. If possible, it would be good to support this with LC3 processing as well.

Response: Following the reviewer's advice, we will use LC3 as an autophagy marker as well as p62 to evaluate the autophagy inhibition of PfHRP2. Concretely, we plan to treat HT1080 cells with PfHRP2 (1 μM) for 12–60 hours and quantify the amount of LC3 protein by Western blotting. The results of this experiment will be added to Fig. 5 in the main manuscript.

2.HRP2 might affect general lysosomal degradation process. The authors can also check whether HPR2 affects degradation of a lysosomal substrate.

Response: Following the reviewer's advice, we will determine the effect of PfHRP2 on lysosomal degradation activity using the plasmid-based lysosomal-METRIQ (MEasurement of protein Transporting integrity by RatIo Quantification) probe, reported in a previous study (https://doi.org/10.1038/s41598-019-48131-2), to quantify lysosomal activity. The results of this experiment will be added to Fig. 5 in the main manuscript.

Reviewer #1 (Significance (Required)): This study compelements previous findings (references 37, 38 and PMID: 31858717). It identifies a new mechanism by which HRP2 might cause toxicity. However, it is completely an in vitro study, and the previous studies (references 37 and 38) have used in vivo models as well.

Response: We wish to thank the reviewer for this comment. As the reviewer pointed out, this study is completely in vitro, and further in vivo studies are essential in the future. Therefore, we have added statements (page 9, lines 294–297 and 309–310; page 10, lines 339–342) to emphasize that the present findings are limited to in vitro and that further in vivo studies are required in the future. We have also added the study (PMID: 31858717) reporting PfHRP2 elicits pro-inflammatory effect and induces vascular permeability as reference 40.

Furthermore, the title of the original paper was vague and gave the impression that it included in vivo experiments. Therefore, to avoid misunderstanding, we modified the paper's title to be more concrete.

Reference

P. Dinarvand, L. Yang, I. Biswas, H. Giri, A. R. Rezaie, Plasmodium falciparum histidine rich protein HRPII inhibits the anti-inflammatory function of antithrombin. J. Thromb. Haemost. 18, 1473–1483 (2020).

We thank you again for giving us the opportunity to improve our paper, and we hope that the changes are satisfactory.

RESPONSE TO REVIEWER #2:

We wish to express our appreciation to Reviewer #2 for his or her insightful comments, which will significantly improve this paper. We thank the reviewers for giving us the opportunity to improve the manuscript. We have responded to all the comments pointed out. The revised sections are highlighted in red characters and yellow backgrounds in the preliminary revised manuscript.

Reviewer #2 (Evidence, reproducibility and clarity (Required)): This paper showed that recombinant Plasmodium falciparum HRPII generated in E. coli is internalized by human tumor derived cells lines and at high concentrations, induces calcium-dependent cell death. The authors propose that HRPII inhibits autolysosome formation and autophagy. Of major concern is the use of E. coli generated HRP2 without addressing the inherent confounders of copurified bacterial components, namely endotoxin LPS. It is crucial for validation of their conclusions that the authors address steps taken to remove endotoxin which is known to bind poly-histidine and HRPII, the quantification of endotoxin bound to purified protein, and the LPS sensitivity of model cell lines. Even small quantities of LPS have been shown to potentially inhibit endosome maturation (https://doi.org/10.1074/jbc.M114.611442). Would recommend caution with conclusions regarding cytotoxicity and autophagy inhibition without addressing this issue.

Response: We sincerely appreciate the reviewer's effort to evaluate our work. The reviewer points out that the endotoxin LPS may also affect the cytotoxicity and autophagy inhibition of PfHRP2 in this study. The reviewer's point is crucial, and we agree with the reviewer. In our study, recombinant PfHRP2 was captured by anti-FLAG antibody-immobilized affinity gel (Medical & Biological Laboratories Co., Ltd., Nagoya, Japan) and washed with 20-bed volumes of washing buffer (20 mM Tris-HCl pH7.4, 500 mM NaCl, 0.1% Triton X-100) to remove contaminants including endotoxin LPS according to the manufacturer's protocol (https://ruo.mbl.co.jp/bio/dtl/dtlfiles/3328R-ver4.0.pdf). After washing, affinity gel was equilibrated with 10-bed volumes of washing buffer without Triton X-100, and recombinant PfHRP2 was eluted by 10-bed volumes of elution buffer (20 mM Tris-HCl pH7.4, 500 mM NaCl, 0.1 mg/mL FLAG peptide: DYKDDDDK). However, we did not determine the residual endotoxin LPS bound to purified PfHRP2. To address the reviewer's concern, we will follow the reviewer's suggestion and quantify the residual endotoxin LPS in the purified PfHRP2 using the LAL Endotoxin Assay Kit. We also plan to test whether the same amount of endotoxin LPS alone as the residual endotoxin LPS affects cytotoxicity and autophagy inhibition. The results of additional experiments on endotoxin LPS will be added to Supplementary Information as Fig. S2. Furthermore, we have added additional information on the purification and washing of PfHRP2 to the Materials and methods section (page 11, lines 356–362).

Additional concerns for specific experiments are as follows: Figure 2A. There is an increase in BSA penetration at lower pH as well which suggests nonspecific increased cell permeability.

Response: As pointed out by the reviewer, the cell membrane permeability of BSA was enhanced at low pH (pH less than 5.8), and this result implies an increase in nonspecific cell permeability. Since we have reported in another study (https://doi.org/10.1093/bbb/zbab221) that BSA shows cell penetration to human gastric cancer cell lines at pH 5.0, the cell membrane permeability of BSA at low pH in this study is satisfactory. However, comparing pH 7.4 and pH 5.6, the net charge of BSA increased by 21.9 from -14.0 (pH7.4) to +7.9 (pH5.6), and the cell penetration increased by 34%. On the other hand, the net charge of PfHRP2 increased by 79.4 from -19.2 (pH7.4) to +60.2 (pH5.6), and the cell penetration increased by 246%. This suggests that the increase in cell membrane permeability of PfHRP2 under low pH conditions is due to the increase in net charge, not to the non-specific increase in cell permeability as seen in BSA. The above explanation has been added to lines 97–103.

Figure 3A, 3B, and 4C. There is inconsistency between the cell viability data. For example, in panel A, 1 μM of HRPII for 24 h showed 84% cell viability whereas in panel B, the cell viability is 61% for 1 μM HRP2 by 24 hours. Figure 3A and 4C (full length) differ at cell viability for 5 μM HRP2.

Response: We thank the reviewer for the critical remarks. There was an error in the time condition described in the graph of Fig. 3A. Correctly, Fig. 3A is the result of cell viability treated with 1 μM PfHRP2 for 3 hours, so we have corrected the time condition described in Fig. 3A. Namely, Fig. 3A and 3B show that a 3-hour treatment with 1 μM PfHRP2 results in 84% cell viability, but a 24-hour treatment with 1 μM PfHRP2 decreases cell viability to 61%. These results are correctly described in lines 119–120, highlighted in yellow.

On the other hand, as the reviewer points out, in Fig. 3A and Fig. 4C (full-length PfHRP2), the cell viability treated with 5 μM PfHRP2 for 24 hours was 5% and 26%, respectively. We believe that the discrepancy in these values is an experimental error. However, both Fig. 3A and Fig. 4C (full-length PfHRP2) agree that 5 μM PfHRP2 is statistically and significantly cytotoxic, which should not affect the claims of this study.

Figure 5C. It would be more informative if the cell viability data at 1 μM of HRP at timepoints beyond 60 hours and for bafilomycin treatment is also presented.

Response: We thank the reviewer for their suggestions. However, the purpose of the experiment in Figure 5C is to prove that PfHRP2 induces autolysosomal dysfunction. Since we confirmed that treatment of cells with 1 µM PfHRP2 for 60 hours resulted in accumulation of p62 in the same amount as the positive control, Bafilomycin A1, we believe that no further additional experiments are necessary.

Figure 3D. (Minor) Consider additional experimental detail regarding maintenance of cell cultures for 5 day. Are there interval media changes or supplement additions?

Response: We apologize for the insufficient information in the description of the experimental procedure in Fig. 3D. In the experiment in Fig. 3D, cell culture was maintained for 5 days by changing a fresh medium containing each concentration of proteins every day. We have added this information to the legends of Figure 3 (page 23, lines 653–655) and Figure S2 (page 4, lines 28–29).

Reviewer #2 (Significance (Required)): The authors present the novel finding of HRP2 permeability into human cells. The significance of these findings is limited given the major confounder with endotoxin and also since the experiments were conducted in tumor-derived cells lines with supraphysiologic concentrations of HRPII. Although the authors showed cell viability effects with lower concentrations over 3 and 5 days, the bulk of the experiments were at more than 10-fold mean physiological concentrations. Also, since these are early findings in tumor-derived cell lines, it is difficult to extrapolate the physiological relevance of these findings and use of calcium chelators as therapeutics. Several studies have proposed a pathogenic role for HRP2 including those cited in the paper regarding blood-brain barrier disruption (references 37 and 38), coagulation disruption (DOI: 10.1182/blood-2010-12-326876), and pro-inflammatory signaling (DOI: 10.1111/jth.14713). The challenge with all these studies is establishing the clinical relevance of the multitude of HRPII effects. If the issue of endotoxin is addressed, this paper could establish an interesting mechanism for further study in more clinically representative systems. Our lab has studied the many functions of HRPII including catalysis of heme polymerization, inhibition of antithrombin, brain endothelial disruption using tissue culture and mouse models.

Response: As pointed out by the reviewer, this study must clear up the effect of endotoxin LPS. In this regard, as mentioned above, we plan to quantify the residual endotoxin LPS in the purified PfHRP2 using the LAL Endotoxin Assay Kit. We will also check the effect of the endotoxin LPS itself on cytotoxicity and autophagy inhibition.

Furthermore, as the reviewer pointed out, this is an in vitro study using high concentrations of PfHRP2 and a tumor-derived cell line, so further in vivo validation is essential in the future. However, it is also true that we discovered new findings that have been overlooked because we conducted an artificial and simple in vitro experiment. In the future, it is necessary to demonstrate the cytotoxicity and autophagy inhibition of PfHRP2 by in vivo studies using model animals. Concretely, we need to confirm whether PfHRP2 behaves as a similar virulence factor in vivo by animal experiments using PfHRP2-administrated or PfHRP2-overexpressing/deficient P. falciparum-infected mouse models. We also need to demonstrate that calcium chelators such as EDTA have an in vivo therapeutic effect. These future tasks have been added to the Discussion (page 9, lines 294–297 and 309–310). We have also added the studies (DOI: 10.1182/blood-2010-12-326876, DOI: 10.1111/jth.14713) reporting PfHRP2 elicits pro-inflammatory effect and induces vascular permeability as reference 37 and 40.

Furthermore, the title of the original paper was vague and gave the impression that it included in vivo experiments. Therefore, to avoid misunderstanding, we modified the paper's title to be more concrete, "Plasmodium falciparum histidine-rich protein II exhibits cell penetration and cytotoxicity with autophagy dysfunction".

References

M. Ndonwi, et al., Inhibition of antithrombin by Plasmodium falciparum histidine-rich protein II. Blood 117, 6347–6354 (2011). P. Dinarvand, L. Yang, I. Biswas, H. Giri, A. R. Rezaie, Plasmodium falciparum histidine rich protein HRPII inhibits the anti-inflammatory function of antithrombin. J. Thromb. Haemost. 18, 1473–1483 (2020).

We thank you again for giving us the opportunity to improve our paper, and we hope that the changes are satisfactory.

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

Learn more at Review Commons

---

Referee #2

Evidence, reproducibility and clarity

This paper showed that recombinant Plasmodium falciparum HRPII generated in E. coli is internalized by human tumor derived cells lines and at high concentrations, induces calcium-dependent cell death. The authors propose that HRPII inhibits autolysosome formation and autophagy.

Of major concern is the use of E. coli generated HRP2 without addressing the inherent confounders of copurified bacterial components, namely endotoxin LPS. It is crucial for validation of their conclusions that the authors address steps taken to remove endotoxin which is known to bind poly-histidine and HRPII, the quantification of endotoxin bound to purified protein, and the LPS sensitivity of model cell lines. Even small quantities of LPS have been shown to potentially inhibit endosome maturation (https://doi.org/10.1074/jbc.M114.611442). Would recommend caution with conclusions regarding cytotoxicity and autophagy inhibition without addressing this issue.

Additional concerns for specific experiments are as follows:

Figure 2A. There is an increase in BSA penetration at lower pH as well which suggests nonspecific increased cell permeability.

Figure 3A, 3B, and 4C. There is inconsistency between the cell viability data. For example, iIn panel A, 1 uM of HRPII for 24 h showed 84% cell viability whereas in panel B, the cell viability is 61% for 1 uM HRP2 by 24 hours. Figure 3A and 4C (full length) differ at cell viability for 5 uM HRP2.

Figure 5C. It would be more informative if the cell viability data at 1 uM of HRP at timepoints beyond 60 hours and for bafilomycin treatment is also presented.

Figure 3D. (Minor) Consider additional experimental detail regarding maintenance of cell cultures for 5 day. Are there interval media changes or supplement additions?

Significance

The authors present the novel finding of HRP2 permeability into human cells. The significance of these findings is limited given the major confounder with endotoxin and also since the experiments were conducted in tumor-derived cells lines with supraphysiologic concentrations of HRPII. Although the authors showed cell viability effects with lower concentrations over 3 and 5 days, the bulk of the experiments were at more than 10-fold mean physiological concentrations. Also, since these are early findings in tumor-derived cell lines, it is difficult to extrapolate the physiological relevance of these findings and use of calcium chelators as therapeutics.

Several studies have proposed a pathogenic role for HRP2 including those cited in the paper regarding blood-brain barrier disruption (references 37 and 38), coagulation disruption (DOI: 10.1182/blood-2010-12-326876), and pro-inflammatory signaling (DOI: 10.1111/jth.14713). The challenge with all these studies is establishing the clinical relevance of the multitude of HRPII effects. If the issue of endotoxin is addressed, this paper could establish an interesting mechanism for further study in more clinically representative systems.

Our lab has studied the many functions of HRPII including catalysis of heme polymerization, inhibition of antithrombin, brain endothelial disruption using tissue culture and mouse models.

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

Learn more at Review Commons

---

Referee #1

Evidence, reproducibility and clarity

This manuscript "Histidine-rich protein 2: a new pathogenic 1 factor of Plasmodium falciparum malaria" by Iwasaki, et al. reports effects of recombinant HRP2 protein on various mammalian cell lines. The MS clearly demonstrates that recombinant HRP2 enters into HT1080 cells, causes inhibition of autolysosome fusion, increases lysosomal Ca ion concentration and reduces general autophagic degradation. The authors also show that the presence of FBS or metal chelators like EDTA and EGTA mitigate toxicity of HRP2, as the former traps HRP2 and the latter compete with HRP2 for Ca binding. The experiments are appropriately carried out with suitable controls in most of the cases. There are some concerns as listed below:

Major concerns:

1.HRP2 has been shown to be associated with virulence and causes vascular leakage, particularly cerebral malaria (references 37 and 38 ). Plasmodium falciparum histidine-rich protein II has been demonstrated to exacerbate experimental cerebral malaria in mice, which has been proposed to be associated with vascular leakage, activation ofinflamosome and cytokine production (references 37, 38 and PMID: 31858717). This study complements the previous findings of the effect of HRP2 on mammalian cells. However, this study reveals another mechanism by which HRP2 might cause toxicity, which is inhibition of general autophagy and increase in lysosomal Ca concentration. However, whether these in vitro effects would translate in vivo needs to be shown.

2.All the experiments are done with recombinant HRP2 and BSA as a control. The authors should show if similar effects happen with infected parasites.

3.HRP2 is released in circulation, making it accessibele to endothelial cells and immune cells. How would it reach to the equivalents of these cells in the human body?

Minor concerns

1.p62 is an appropriate marker to assess autophagy cargo degradation. If possible, it would be good to support this with LC3 processing as well.

2.HRP2 might affect general lysosomal degradation process. The authors can also check whether HPR2 affects degradation of a lysosomal substrate.

Significance

This study compelements previous findings (references 37, 38 and PMID: 31858717). It identifies a new mechanism by which HRP2 might cause toxicity. However, it is completely an in vitro study, and the previous studies (references 37 and 38) have used in vivo models as well.

Note: This rebuttal 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

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

In this paper, Harterink and colleagues investigate the establishment of minus-end-out microtubule polarity in the anterior dendrite of C. elegans PVD neurons. These neurons offer an excellent model system due to their simplicity and well-defined microtubule polarity. The authors investigate the role of two proteins in particular, the well-studied Patronin protein and a newly identified homologue of Ninein (Noca-2). They show that these proteins are redundantly required for correct minus-end-out polarity. Absence of one of these proteins results in a low penetrant phenotype, but absence of both results in a strongly penetrant phenotype. Interestingly, in all cases the neurons display either almost fully retrograde or almost fully anterograde microtubule polarity, and not a mix of retrograde and anterograde microtubules. This is probably linked to the fact that the authors show that endosomes at the distal tip of the dendrite (that are known to mediate retrograde microtubule nucleation events) are either present or absent in these mutants (to differing degrees that reflect the polarity phenotypes of each mutant type). The authors further show that Noca-2, but not Patronin, is required for proper localisation of γ-tubulin to the distal endosomes, suggesting that the proteins influence microtubule polarity in different ways. They provide some evidence that Patronin clusters, while not colocalized to the distal endosomes, are somehow connected. The paper and figures are clear and the work should be reproducible.

Most conclusions are supported by the data, except for when the authors say: "Taken together, these results show that PTRN-1 (CAMSAP) and NOCA-2 (NINEIN) act in parallel in the PVD neuron during early development to establish minus-end out microtubule organization, and that this organization is important for proper dendritic morphogenesis." But the authors show that removal of Patr results in some neurons having a complete anterograde phenotype in the anterior genotype, but that no Patr neurons have a severe morphology defect (Fig 2). This would suggest that the severe morphology defects in Patr/Noca-2 double mutants are not simply due to the reversal of polarity in the anterior dendrite. This should be discussed.

We agree with the comment, and we will discuss this more clearly in a revised manuscript.

The paper could be strengthened with some biochemistry showing that Noca-2 can associate with γ-TuRCs i.e. do purified fragments of Noca-2 pull out γ-TuRCs from a cell extract (not necessarily a neuron cell extract)? This should be possible within 1 month.

We thank the reviewer for this suggestion. We will perform some biochemistry experiment to probe the association of NOCA-2 with γ-TuRCs. However, instead of doing the IP by overexpression of NOCA-2 and γ-TuRCs in cells, we will use the CRISPR knockin animals for NOCA-2 and γ-TuRCs, to exclude potential overexpression artifacts.

Minor comments

1) "However, in polarized cells such as neurons, most microtubules are organized in a non-centrosomal manner (Nguyen et al., 2011)." Need more up to date reference here, such as a recent review from Jens Lüders.

We will update the references in the revision version of the manuscript.

2) "and also in Drosophila Patronin was found important for dendritic microtubule polarity (Feng et al., 2019)." Also Wang et al., 2019 in eLife.

We will add this reference.

3) "In the non-ciliated PHC neuron or the ciliated URX neuron we did not observe microtubule organization defects in the ptrn-1 mutant (Supplemental figure 1A-B), which suggests that these neurons do less or do not dependent on PTRN-1." End of sentence needs re-phasing

We will rephrase the text.

Reviewer #1 (Significance (Required)):

Overall, the paper adds some interesting information to the field but does not make a conceptual advance that would make it attractive to a wide audience. It will, however, be of interest to those studying mt regulation in neurons. It is a shame that the molecular mechanism that allow Noca-2, and particularly Patronin, to establish microtubule polarity remain to be determined. Figuring out these mechanisms would significantly strengthen the paper.

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

Harterink comments:

In this manuscript He et al investigate the role of two key microtubule minus end regulators, Patronin/CAMSAP and NOCA-2/ninein, in establishing dendrite microtubule organization. The authors use a well-characterized branched sensory neuron in C. elegans for their analysis and make significant contributions to our understanding of neuronal microtubule organization. First, they show that C. elegans has not one, but two, ninein-like proteins, NOCA-1 and NOCA-2. Previously only NOCA-1 had been identified, and neuronal functions of ninein have remained elusive, perhaps in part because NOCA-2 had been missed. It had previously been shown that in epithelial cells NOCA-1 acts with gamma-tubulin as one arm of a microtubule minus end organizing pathway, while Patronin acts in parallel on minus ends. The current manuscript very nicely extends this functional map to neurons. The authors show that NOCA-2 helps recruit the gamma-tubulin ring complex (g-TuRC) to Rab11 endosomes that are important for microtubule nucleation at developing dendrite tips. As in epithelial cells, Patronin seems to act in parallel to this pathway and rather than being involved in recruiting the g-TuRC to Rab11 endosomes, is instead important for allowing the Rab11 endosomes to be transported to developing dendrite tips. In total this analysis not only identifies a new player in dendritic microtubule organization (NOCA-2), but also helps synthesize the functions of other players (g-TuRC, Patronin) into a model that makes sense in the broader context of microtubule organization across species and cell types.

• Specific points *

1. • NOCA-2 is described as a previously unidentified member of the ninein family. In order to evaluate this claim critically, it would be helpful to have a figure showing how similar NOCA-1 and NOCA-2 are to mammalian ninein. It is also critical to include a phylogeny to get a better feel for how NOCA-2 fits into the evolutionary history of the family. * We agree with the suggestion. We will perform this analysis and add it to a revised version of the manuscript.

• The nucleation assay used throughout is not very clear as one reads through the manuscript. The color-coding of Fig 1G could be better defined in the legend, and it would be helpful to have more information in the legend or results about what is meant here by microtubule nucleation. Is it simply initiation of new microtubule growth events? If so, how are these distinguished from catastrophe rescue? It would be good to use the same color coding in 2E. *

We thank the reviewer for pointing this out. In Fig 1G and 2E we indeed quantified EBP-2::GFP growth events. Although we later show that the microtubule nucleator gamma-tubulin localized to the distal segment where we observe increased microtubule growth events, we agree that we cannot distinguish microtubule nucleation from regrowth after catastrophe. Therefore, we will describe this more accurately in the text, legend and in the figure.

• The colocalization of Rab-11 and NOCA-2 seems to be supported only by a single overlapping puncta in a neuron before the anterior dendrite extends (Fig 4). It would be good to flesh out this data set more as it is an important part of the argument that NOCA-2 is involved in recruiting g-TuRC to Rab11 endosomes. *

We thank the reviewer for pointing this out. We will flesh out this data either by adding several examples and/or a movie to show the localization of Rab-11 and NOCA-2 in the revised version of the manuscript.

• Summary diagrams of results either as conclusions are made in the individual figures or synthesized at the end would help readers to understand the evidence that NOCA-2 and PTRN-1 function at different steps in establishment of MT polarity *

We agree that a summary diagram could be helpful, and we will consider adding this to the revised version of the manuscript.

• NOCA-1 is introduced at the beginning of the manuscript and appears to act in parallel to both NOCA-2 and PTRN-1. One is left with many questions, for example, is it also required to recruit g-TuRCs to Rab11 vesicles, or does it have some other role? However, I appreciate that it is beyond the scope of a single manuscript to answer all questions and the authors state a clear rationale for focusing on NOCA-2. *

We agree that the function of NOCA-1 is interesting to be investigated in the future, since we found it acts redundantly to PTRN-1 and NOCA-2. As NOCA-1 is an essential gene this brings along some technical difficulties to properly address its function and would require generating novel tools. We appreciate the reviewers understanding that this is beyond the scope of the current manuscript.

• It would be helpful to clearly state at some point which aspects of localization that are described are seen only in developing dendrites and which are seen in both developing and mature dendrites. For example, is there any similarity in localization of PTRN-1 NOCA-2 and gip2 in mature dendrites to that shown in immature? Is there any sign of continued localization to RAB11 vesicles, or is this only transient? Perhaps a diagram to summarize these findings would also be helpful. *

We thank the reviewer for pointing this out. We will better explain the localization of NOCA-2, PTRN-1, GIP-2 and RAB-11 vesicles in developing neurons vs mature neurons in a revised version of the manuscript.

• The authors propose several different ideas about how Patronin might contribute to Rab11 vesicle localization. However, I am not sure that they really describe the simplest one: that Patronin helps minus ends grow out from the cell body as shown in Drosophila (and Fig S6 here), and that these minus-end-out microtubules could be the tracks used to transport Rab11 into dendrites. Have I missed some reason why this model is not presented as a good fit for the data? *

We thank the reviewer for pointing this out. Feng et al indeed showed that EB proteins can track microtubule plus- and minus-end growth in the sensory neurons of Drosophila. Since the slower event co-localize with Patronin they suggested that these help to populate the minus-end out microtubules in the drosophila dendrites (Feng et al., 2019).

Although we do not have strong data against this model for the PVD dendrites in C. elegans, there are several observations that to us suggest that it is unlikely that minus-end growth is the driving force for the forward movement of the MTOC vesicles. These include: the MT being mixed in the distal segment, therefore it is hard to imagine how specifically one pool is growing; we do not see EBP-2 localize to the Camsap puncta as was seem in Drosophila; the Camsap dynamics at the growth cone seem very different (less processive) to the dynamics in the shaft (which indeed could be minus-end growth). We will make this reasoning more clear in the revised manuscript.

• There are some grammatical errors throughout, as well as a few typos (like PTNR-1 for PTRN-1). *

We will correct the text grammar and typos in the revision version of manuscript.

Reviewer #2 (Significance (Required)):

This analysis will help synthesize a more complete and meaningful understanding of how non-centrosomal microtubules are organized. The authors not only identify a new player in non-centrosomal microtubule organization, but also help fit together several existing players into a framework that brings together observations from other model systems and cell types into a more coherent whole.

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

Harterink comments:

In this manuscript He et al investigate the role of two key microtubule minus end regulators, Patronin/CAMSAP and NOCA-2/ninein, in establishing dendrite microtubule organization. The authors use a well-characterized branched sensory neuron in C. elegans for their analysis and make significant contributions to our understanding of neuronal microtubule organization. First, they show that C. elegans has not one, but two, ninein-like proteins, NOCA-1 and NOCA-2. Previously only NOCA-1 had been identified, and neuronal functions of ninein have remained elusive, perhaps in part because NOCA-2 had been missed. It had previously been shown that in epithelial cells NOCA-1 acts with gamma-tubulin as one arm of a microtubule minus end organizing pathway, while Patronin acts in parallel on minus ends. The current manuscript very nicely extends this functional map to neurons. The authors show that NOCA-2 helps recruit the gamma-tubulin ring complex (g-TuRC) to Rab11 endosomes that are important for microtubule nucleation at developing dendrite tips. As in epithelial cells, Patronin seems to act in parallel to this pathway and rather than being involved in recruiting the g-TuRC to Rab11 endosomes, is instead important for allowing the Rab11 endosomes to be transported to developing dendrite tips. In total this analysis not only identifies a new player in dendritic microtubule organization (NOCA-2), but also helps synthesize the functions of other players (g-TuRC, Patronin) into a model that makes sense in the broader context of microtubule organization across species and cell types.

Specific points

1. NOCA-2 is described as a previously unidentified member of the ninein family. In order to evaluate this claim critically, it would be helpful to have a figure showing how similar NOCA-1 and NOCA-2 are to mammalian ninein. It is also critical to include a phylogeny to get a better feel for how NOCA-2 fits into the evolutionary history of the family.
2. The nucleation assay used throughout is not very clear as one reads through the manuscript. The color-coding of Fig 1G could be better defined in the legend, and it would be helpful to have more information in the legend or results about what is meant here by microtubule nucleation. Is it simply initiation of new microtubule growth events? If so, how are these distinguished from catastrophe rescue? It would be good to use the same color coding in 2E.
3. The colocalization of Rab-11 and NOCA-2 seems to be supported only by a single overlapping puncta in a neuron before the anterior dendrite extends (Fig 4). It would be good to flesh out this data set more as it is an important part of the argument that NOCA-2 is involved in recruiting g-TuRC to Rab11 endosomes.
4. Summary diagrams of results either as conclusions are made in the individual figures or synthesized at the end would help readers to understand the evidence that NOCA-2 and PTRN-1 function at different steps in establishment of MT polarity
5. NOCA-1 is introduced at the beginning of the manuscript and appears to act in parallel to both NOCA-2 and PTRN-1. One is left with many questions, for example, is it also required to recruit g-TuRCs to Rab11 vesicles, or does it have some other role? However, I appreciate that it is beyond the scope of a single manuscript to answer all questions and the authors state a clear rationale for focusing on NOCA-2.
6. It would be helpful to clearly state at some point which aspects of localization that are described are seen only in developing dendrites and which are seen in both developing and mature dendrites. For example, is there any similarity in localization of PTRN-1 NOCA-2 and gip2 in mature dendrites to that shown in immature? Is there any sign of continued localization to RAB11 vesicles, or is this only transient? Perhaps a diagram to summarize these findings would also be helpful.
7. The authors propose several different ideas about how Patronin might contribute to Rab11 vesicle localization. However, I am not sure that they really describe the simplest one: that Patronin helps minus ends grow out from the cell body as shown in Drosophila (and Fig S6 here), and that these minus-end-out microtubules could be the tracks used to transport Rab11 into dendrites. Have I missed some reason why this model is not presented as a good fit for the data?
8. There are some grammatical errors throughout, as well as a few typos (like PTNR-1 for PTRN-1).

Significance

This analysis will help synthesize a more complete and meaningful understanding of how non-centrosomal microtubules are organized. The authors not only identify a new player in non-centrosomal microtubule organization, but also help fit together several existing players into a framework that brings together observations from other model systems and cell types into a more coherent whole.

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 paper, Harterink and colleagues investigate the establishment of minus-end-out microtubule polarity in the anterior dendrite of C. elegans PVD neurons. These neurons offer an excellent model system due to their simplicity and well-defined microtubule polarity. The authors investigate the role of two proteins in particular, the well-studied Patronin protein and a newly identified homologue of Ninein (Noca-2). They show that these proteins are redundantly required for correct minus-end-out polarity. Absence of one of these proteins results in a low penetrant phenotype, but absence of both results in a strongly penetrant phenotype. Interestingly, in all cases the neurons display either almost fully retrograde or almost fully anterograde microtubule polarity, and not a mix of retrograde and anterograde microtubules. This is probably linked to the fact that the authors show that endosomes at the distal tip of the dendrite (that are known to mediate retrograde microtubule nucleation events) are either present or absent in these mutants (to differing degrees that reflect the polarity phenotypes of each mutant type). The authors further show that Noca-2, but not Patronin, is required for proper localisation of γ-tubulin to the distal endosomes, suggesting that the proteins influence microtubule polarity in different ways. They provide some evidence that Patronin clusters, while not colocalized to the distal endosomes, are somehow connected. The paper and figures are clear and the work should be reproducible.

Most conclusions are supported by the data, except for when the authors say: "Taken together, these results show that PTRN-1 (CAMSAP) and NOCA-2 (NINEIN) act in parallel in the PVD neuron during early development to establish minus-end out microtubule organization, and that this organization is important for proper dendritic morphogenesis." But the authors show that removal of Patr results in some neurons having a complete anterograde phenotype in the anterior genotype, but that no Patr neurons have a severe morphology defect (Fig 2). This would suggest that the severe morphology defects in Patr/Noca-2 double mutants are not simply due to the reversal of polarity in the anterior dendrite. This should be discussed.

The paper could be strengthened with some biochemistry showing that Noca-2 can associate with γ-TuRCs i.e. do purified fragments of Noca-2 pull out γ-TuRCs from a cell extract (not necessarily a neuron cell extract)? This should be possible within 1 month.

Minor comments

1. "However, in polarized cells such as neurons, most microtubules are organized in a non-centrosomal manner (Nguyen et al., 2011)." Need more up to date reference here, such as a recent review from Jens Lüders.
2. "and also in Drosophila Patronin was found important for dendritic microtubule polarity (Feng et al., 2019)." Also Wang et al., 2019 in eLife.
3. "In the non-ciliated PHC neuron or the ciliated URX neuron we did not observe microtubule organization defects in the ptrn-1 mutant (Supplemental figure 1A-B), which suggests that these neurons do less or do not dependent on PTRN-1." End of sentence needs re-phasing

Significance

Overall, the paper adds some interesting information to the field but does not make a conceptual advance that would make it attractive to a wide audience. It will, however, be of interest to those studying mt regulation in neurons. It is a shame that the molecular mechanism that allow Noca-2, and particularly Patronin, to establish microtubule polarity remain to be determined. Figuring out these mechanisms would significantly strengthen the paper.

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 rebuttal 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

Manuscript number: RC-2021-01129

Corresponding author(s): Koji Kikuchi

---

Reviewer #1

Evidence, reproducibility and clarity (Required):

In this manuscript, Kikuchi et al describe the characterization of MAP7D2 and MAP7D1, two MAP7 family members in mouse with specific expression patterns. Focusing mostly on MAP7D2, they assess its expression pattern across the body and find that it is mostly expressed in certain neuronal subsets. They then characterize the MT-related properties of MAP7D2 based on previous knowledge of other MAP7 family members. They show that MAP7D2 binds MTs (via the N-terminus), determine the binding affinity, and show that it can stimulate MT polymerization (or stabilization) both in vitro and in vivo. Using a specific antibody, they localize MAP7D2 to centrosomes, midbody and neurites in N1-E115 cells. Functionally, they show that loss of MAP7D1/2 mildly affects microtubule stability as judged by acetyl-tubulin staining, and properties of these cells that rely on cytoskeletal elements such as cell migration and neurite growth. Interestingly, there might be a feedback loop regulating MAP7D1/2 expression, as knockdown of MAP7D1 upregulates MAP7D2.

Overall, the experiments and conclusions are very solid and convincing, such that I would not ask for further experiments. This is in part because the experiments are largely based on previous characterizations of other MAP7 family members, which are largely confirmed. The presentation of the data is also very clear.

Significance (Required):

I see the value of the study in the fact that it provides solid and specific research tools for MAP7D1/2 which could be very useful for the microtubule/neuronal cytoskeleton community.

Response: We thank the reviewer very much for appreciating the content of our manuscript.

\*Referees cross-commenting***

Reviewers 2 and 3 criticize that the evidence for an effect of MAP7D1/2 on MT dynamics is weak. I would agree in that ac-tub stainings and in vitro experiments are rather indirect. The experiments suggested by reviewer 2 should clarify this (esp. nocodazole should be easy). I also agree that an experiment addressing the potential involvement of kinesin-1 would help, the involvement of which seems to have been omitted by the authors. A kinesin-binding deficient mutant would add another MAP7D1/2 tool and increase the value for the community.

Response: As for the reviewer’s suggestions listed above, please refer to our responses to the comments of Reviewer #2.

Reviewer #2

Evidence, reproducibility and clarity (Required):

In this study, the authors investigate 2 members from the MAP7 family Map7D2 and Map7D1. They first address the tissue distribution of Map7D2, by northern blotting using a variety of rat tissues. To complement their analysis, they also raised an antibody to look at the protein distribution. From their studies, they concluded that Map7D2 is abundantly expressed in the brain and testis. The authors went on to perform a series of functional assays. First, they biochemically demonstrated that rat Map7D2 directly binds to MTs by MT co-sedimentation assay. The MT binding domain was mapped to the N-terminal half. They performed MT turbidity assay to demonstrate enhanced MT polymerisation in the presence of Map7D2, suggesting that this Map stabilises MTs. The authors went on to characterise in detail the subcellular localisation of Map7D2 which was predominantly present in the centrosome and partially localised to MTs including within neurites from N1-E115 cells. Kikuchi et al. further revealed the overlap in expression between Map7D2 and another family member, Map7D1. The authors continued these studies by a series of functional studies in N1-E115 cells where they performed single or combined knock-downs of Map7D2 and Map7D1 and studied the levels of acetylated and detyrosinated tubulins and the effect of the knock-downs on migration and neurite extension. The main conclusion from this work was that Map7D2 and Map7D1 facilitate MT stabilization through distinct mechanisms which are important in controlling cell motility and neurite outgrowth. Map7D2 is proposed to stabilise MTs by direct binding whereas Map7D1 does it indirectly by affecting acetylation.

Major comments:

The main conclusion from this work that Map7D2 and Map7D1 facilitate MT stabilization and that this is necessary for correct migration and neurite extension has not been convincingly demonstrated. In my opinion, a more detailed study of MT properties to demonstrate a role in MT stabilisation would greatly benefit the work, eg. experiments using MT destabilising agents such as nocodazole. In addition, a series of experiments aiming to study MT dynamics would help to understand the function of these MT regulators. The authors proposed an elevation in microtubule dynamics to explain the increase in migration and neurite extension but no experimental proof was provided.

Response: According to the reviewer’s suggestion, we plan to assess the role of MT stabilization in greater detail by analyzing the sensitivity to the MT-destabilizing agent, nocodazole.

To study MT dynamics, methods such as analyzing the velocity and direction of an EB1-GFP comet are commonly used. We have previously analyzed the roles of Map7 and Map7D1 in MT dynamics using HeLa cells stably expressing EB1-GFP (Kikuchi et al., EMBO Rep., 2018). However, no such tools have been developed for analyzing MT dynamics in N1-E115 cells, which were used in this study. In addition, it is difficult to analyze MT dynamics by transient expression of EB1-GFP because of the low plasmid transfection efficiency. Therefore, we instead plan to assess the effect on MT dynamics by measuring the EB1 comet length by immunofluorescence, referring to Fig. 7D in EMBO J. 32:1293–1306, 2013.

Moreover, considering the possibility that the Map7D2 dynamics are altered when MT stability is changed, e.g., before and after differentiation induction, we analyzed the Map7D2 dynamics at the centrosome by fluorescence recovery after photobleaching (FRAP) using N1-E115 cells stably expressing EGFP-rMap7D2. We found that the dynamics were altered between the proliferative and differentiated states (see the figure below). Compared to the proliferative state, the recovery rate of EGFP-Map7D2 was reduced (lower left panel), and the immobile fraction of Map7D2 was increased in the differentiated state (lower right panel). As these data suggest that the increase in immobile Map7D2 may enhance MT stabilization, we will present them in a new figure in our manuscript along with the results of the above two experiments.

It has been previously demonstrated that loss of MAP7D2 leads to a decrease in axonal cargo entry to axons resulting in defects in axon development and neuronal migration. The C-terminus is necessary for this function as it mediates interaction with Kinesin-1 (Pan et al., 2019). Such mechanisms could also explain the defects in migration and neurite growth that the authors observed. This possibility has not been considered but instead, the subtle changes in total α-tubulin led to suggest MT stabilisation as a key function without proof of causation. Could the authors provide some further experimental evidence to demonstrate that stability is the main contributor to the phenotypes observed? Eg. by rescuing migration and neurite phenotypes with a variant of MAP7D2 which cannot bind kinesin1.

Response: The reviewer states “Such mechanisms could also explain the defects in migration and neurite growth that the authors observed;” however, our results showed that loss of Map7D2 elevated the rates of both cell motility and neurite outgrowth (original Fig. 5). In contrast, it has been reported in several papers that when Kinesin-1 function is impaired, both cell motility and neurite outgrowth are reduced (Curr. Biol., 23: 1018–1023, 2013; Mol. Cell. Biol., 39: e00109–19, 2019; etc.). Therefore, it is likely that the phenotypes we observed are independent of the functions associated with Kinesin-1 in N1-E115 cells. It is indeed possible that the experiment suggested by the reviewer may reveal relationships between Map7D2 and kinesin-1 in terms of cell motility and neurite outgrowth, however, it is difficult to conduct such an experiment because transient expression of Map7D2 induces MT bundling, as shown in original Fig. 2F. Based on the above, we plan to add a discussion of the relationship between Map7D2 and Kinesin-1.

A key conclusion proposed by the authors is that Map7D2 and Map7D1 facilitate MT stabilization through distinct mechanisms. Such different roles in MT stabilisation are important in controlling cell motility and neurite outgrowth. In my opinion, their data does not fully support this statement and the findings using MT readouts do not match the defects in migration and neurite growth. Loss of Map7D2 leads to a very subtle phenotype on α-tubulin, while Map7D1 decreases both α-tubulin and acetylated tubulin, but Map7D1 seems to have a milder or similar effect on migration and neurite growth than Map7D2. Furthermore, it would be expected that the combined loss of function would lead to a stronger phenotype in cell migration when compared to the single loss of functions due to their distinct roles on MT stability, however, this seems not to be the case.

Response: The fact that no stronger phenotype was observed may be because, besides Map7D2 and Map7D1, other molecules are involved in MT stabilization. Another possible explanation is that the increases in both cell motility and neurite outgrowth caused by decreased MT stabilization are offset by Kinesin-1 dysfunction. We plan to add a discussion of the above two possibilities.

Minor comments:

1) In the first result section, the author refers to Fig. S3 to suggest the expression of MAP7D2 in the cerebral cortex, however, there are no transcripts in the cerebral cortex according to the figure. Similarly, the immunofluorescence analysis done by the authors shows marginal expression of MAP7D2 in the cerebral cortex.

Response: According to the reviewer’s comment, we have changed the order of the data shown in Fig. 1C, top panels. The data from the olfactory bulb, cerebellum, and hippocampus, in which Map7D2 expression was detected in the database, were arranged in the top three rows, and the data from the cerebral cortex, in which Map7D2 expression was not detected in the database, were moved to the bottom row as a negative control. In addition, we have revised the relevant part of the Results section as follows: “Based on RNA-seq CAGE, RNA-Seq, and SILAC database analysis (Expression Atlas, https://www.ebi.ac.uk/gxa/home/), Map7D2 expression was detected in the cerebellum, hippocampus, and olfactory bulb, and not in the cerebral cortex (Fig. S3). We further confirmed Map7D2 expression in the above four brain tissue regions of postnatal day 0 mice by immunofluorescence. Among these regions, Map7D2 was the most highly expressed in the Map2-negative area of the olfactory bulb, i.e., the glomerular layer (Fig. 1C). Weak signals were detected in the cerebellum, and marginal signals were observed in the hippocampus and cerebral cortex (Fig. 1C).” (page 5, lines 4–11)

2) The authors use γ-Tubulin as a housekeeping gene in Fig. 3D, since Map7D2 is enriched in centrosomes this may not be the most appropriate choice.

Response: γ-Tubulin is abundant in both the cytosol and the nuclear compartments of cells (Sig. Transduct. Target Ther. 3: 24, 2018). As it has been used for similar purposes in several other studies (Cancer Res., 61: 7713–7718, 2001; J. Biol. Chem., 291: 23112–23125, 2016; etc.), we considered it acceptable for use as a loading control for immunoblotting.

3) According to the authors, knockdown of Map7D2 leads to a decrease in the intensity of α-tubulin and Map7D1 (Fig. 4C and D). This data doesn't agree with the previous statement made by the authors where they show that Map7D2 knockdown or knockout did not affect Map7D1 expression by Western Blot Analysis (Fig. S2C and S5B)

Response: The immunoblotting results indicate that the total amount of Map7D1 in the cells is not affected by loss of Map7D2. In contrast, the immunofluorescence results indicate that the amount (distribution) of Map7D1 localized around the centrosome is decreased by loss of Map7D2, presumably due to a reduction in the number of MT structures that can serve as scaffolds for Map7D1. We plan to add this interpretation in the Results section.

4) Line 6 page 7 "Endogenous Map7D2 expression is suppressed in N1-E115 cells stably expressing EGFP-rMap7D2 and was restored by specific knock-down of EGFP-rMap7D2 using gfp siRNA (Fig. 3D)". No quantifications and stats are shown. Also, endogenous Map7D2 after knock-down of EGFP-rMap7D2 is not comparable to the control.

Response: According to the reviewer’s suggestion, we have quantified the amount of endogenous Map7D2 or EGFP-rMap7D2, normalized it to the amount of γ-tubulin, and calculated relative values to endogenous Map7D2 in the parental control. The amount of endogenous Map7D2 was decreased to 53% in N1-E115 cells stably expressing EGFP-rMap7D2, suggesting that EGFP-rMap7D2 expression suppressed endogenous Map7D2 expression. In this cell line, the total amount of Map7D2 (EGFP-rMap7D2 + endogenous Map7D2) was increased, however, when EGFP-rMap7D2 was depleted using sigfp in this cell line, endogenous Map7D2 was expressed to the same level as EGFP-rMap7D2 before knock-down. Together with the finding that Map7d1 knock-down increased the amount of Map7D2, these findings indicate that the amount of Map7D2 in the cells is regulated in response to the amount of Map7D1 and exogenous Map7D2. We have added this interpretation in the Results section. (page 7, lines 8–15)

In addition, we have changed the legend of the original Fig. 3D to clarify the quantification method, as follows: “(D) Generation of N1-E115 cells stably expressing EGFP-rMap7D2. To check the expression level of EGFP-rMap7D2, lysates derived from the indicated cells were probed with anti-GFP (top panel) and anti-Map7D2 (middle panel) antibodies. The blot was reprobed for γ-tubulin as a loading control (bottom panel). The amount of endogenous Map7D2 or EGFP-rMap7D2 was normalized to the amount of γ-tubulin, and the value relative to endogenous Map7D2 in the parental control was calculated.” (page 22, lines 18–20)

5) Line 8 page 7 "These results suggest that the expression of Map7D2 was influenced by changes in that of Map7D1" This statement seems in the wrong place, after the Map7D2 and EGFP-rMap7D2 experiment. Instead for clarity, it would be better placed after line 5 where the authors explain the effect of Map7D1 knock-down on the levels of Map7D2.

Response: According to the reviewer’s suggestion, we have rephrased the relevant sentence as “Interestingly, Map7d1 knock-down upregulated Map7D2 expression, as confirmed with three different siRNAs (Fig. S2C), suggesting that Map7D2 expression is affected by changes in Map7D1 expression, not by off-target effects of a particular siRNA.” (page 7, lines 7, 8)

6) Line 8 page 8 "Although the physiological role of the C-terminal region of Map7D2 is currently unknown..." This statement seems not adequate as there are several studies reporting the role of the C-terminal region of Map7D2 in Kinesin1- mediated transport. The authors mention such studies in the discussion.

Response: According to the reviewer’s suggestion, we plan to add a discussion of the relationship between Map7D2 and kinesin-1.

7) Line 6 page 9 " Further, the knock-down of either resulted in a comparable reduction of MT intensity (Fig. 4C and D) ..." This is not visible and/or justified by the images provided and would benefit from some sort of quantification at other regions such as neurites.

Response: Considering the cell motility, quantification of α-tubulin/Ace-tubulin/Map7D1/Map7D2 intensities in neurites is not appropriate. Instead, we have added arrowheads indicating α-tubulin/Ace-tubulin/Map7D1/Map7D2 in Fig. 4C, for better understanding.

8) In Fig. 2B, a band corresponding to his6-rMAP7D2 of molecular weight >97 kDa co-sedimented with the microtubules. However, the cloned rMAP7D2 had a molecular weight of 84.82 kDa and the addition of 6XHis-Tag would add another 2-3 kDa, therefore, the final protein band observed should be less than 90 kDa. It would be beneficial if the authors could specify the molecular weight of the purified protein after the addition of the V5-his tag and/or if there was addition of amino acids due to cloning strategy.

Response: In Fig. 2B, we used full-length GST-tagged rMap7D2, like in Fig. 2E and D; therefore, we have corrected His6-rMap7D2 as GST-rMap7D2. We apologize for the mistake.

9) In Fig. 2C, there is misalignment of the western blot with the panel or text underneath.

Response: We thank the reviewer for pointing this out; we have corrected the misalignment of the CBB staining in Fig. 2C.

10) In Fig. 3C the inset from the first panel seems to correspond to a different focal plane than the main image.

Response: We have revised the relevant part of the figure legend as follows: “In C, images of differentiated cells were captured by z-sectioning, because the focal planes of the centrosome and neurites are different. Each inset shows an enlarged image of the region indicated with a white box at each focal plane. Arrowheads indicate the centrosomal localization of Map7D2.”

11) In Fig. 4A, the cell type is not specified and is referred as "indicated cells", also the material and methods section seems to omit the specific cells used.

Response: We have added “in N1-E115 cells treated with each siRNA” in the legend of Fig. 4A.

12) Fig. S6 is not mentioned in the results.

Response: We apologize for having referred to Fig. S6 only in the Discussion section in the original manuscript. We plan to describe the findings shown in the original Fig. S6 to the Results section and renumber the figures accordingly.

Significance (Required):

MTs play essential roles in practically every cellular process. Their precise regulation is therefore crucial for cellular function and viability. MAPs are specialised proteins that interact with MTs and regulate their behaviour in different manners. Understanding their precise function in different cellular contexts is of utmost importance for many biological and biomedical fields.

MAPs are well known for their ability to promote MT polymerization, bundling and stabilisation in vitro (Bodakuntla et al., 2019). Several members of the Map7 family have been shown to regulate microtubule stability. For instance, MAP7 can prevent nocodazole-induced MT depolymerization and maintain stable microtubules at branch points in DRG neurons (Tymanskyj & Ma, 2019). Ensconsin, the Drosophila Map, is required for MT growth in mitotic neuroblasts by regulating the mean rate of MT polymerization (Gallaud et al., 2014). However, this family of Maps seems to have diverse functions encompassing a variety of mechanisms, as exemplified by a series of studies demonstrating the involvement of MAP7 family proteins in the recruitment and activation of kinesin1 (Hooikaas et al., 2019; Pan et al., 2019) and in microtubule remodelling and Wnt5a signalling (Kikuchi et al., 2018). Further understanding of this family of Maps and how its members differ in their function is important and will help to advance the field.

Response: We appreciate the reviewer’s comments. We believe that our revision plan will greatly improve the quality of our manuscript.

Reviewer #3

Evidence, reproducibility and clarity (Required):

Summary:

Microtubule Associated Proteins (MAPs) are important regulators of microtubule dynamics, microtubule organization and vesicular transport by modulating motor protein recruitment and processivity. In the current manuscript the authors have characterized 2 members of the MAP7 protein family, MAP7D1 and MAP7D2. The authors characterized MAP7D2 expression pattern in the brain and its microtubule binding properties in vitro and in cells. In cells both proteins localize to the centrosome and to microtubules and upon depletion centrosome localized microtubules seem reduced, and cell migration and neurite outgrowth are increased. Surprisingly, they find that microtube acetylation (a common marker for stable microtubules) is reduced upon MAP7D1 depletion but not MAP7D2 depletion. Based on this finding the authors conclude that these proteins have a distinct mechanism in stabilizing MTs to affect cell migration and neurite outgrowth; MAP7D2 stabilizes by binding to MTs, whereas MAP7D1 stabilizes MTs by acetylation.

Main comments:

- Both MAP7 proteins show strong localization to the centrosome and to a lesser degree to MTs. Knockdown of either protein leads to reduced MTs around the centrosome, which lead the authors to conclude the MAP7s are stabilizing the MTs. However, the effect could just as well be an indirect effect due to a function of these MAPs at the centrosome. To address this authors could e.g. quantify microtubule properties in postmitotic cells. In addition, antibody specificity should be tested using knockdown of knockout cells, as this centrosome localization was not observed in Hela cells (Hooikaas, 2019; Kikuchi, 2018). Maybe this localization is specific to rat MAP7s or to the cell line used.

Response: We think that this comment partly overlaps with the comments raised by Reviewer #2. We plan to assess the role of MT stabilization in greater detail by analyzing the sensitivity to the MT-destabilizing agent, nocodazole, and the effect on MT dynamics by measuring the EB1 comet length by immunofluorescence.

Regarding the reviewer’s concern about antibody specificity, we had carefully confirmed the antibody specificity, as shown in Fig. S2 of the original manuscript. Subsequently, Map7D2 localization was confirmed in N1-E115 cells stably expressing EGFP-rMap7D2, as shown in Fig. 3D, E of the original manuscript. In addition, we are currently conducting analyses using Map7d1-egfp knock-in mice, which confirmed that Map7D1 localizes around the centrosome in cortical neurons, as shown below (we would like to disclose these unpublished data to the reviewers only). Therefore, it is thought that the localization pattern of Map7D2 and Map7D1 differs depending on the cell type and cell line. We plan to add this interpretation to the Results section.

- Centrosome nucleated microtubules are typically highly dynamic and little modified. Therefore is the Ac-tub staining at the centrosome really MTs? I cannot identify MTs in the fluorescent images in 4C. Maybe authors could consider ac-tub/alpha-tub ratio in non centrosomal region (e.g. neurites). Moreover, as both Acetylation and detyrosination are associated with long-lived/stable MTs, it is surprising that only acetylated tubulin goes down on WB. Does this suggest that long-lived MTs are still present to normal level? If so, can one still argue that the loss of acetylation is the cause of the lower MT levels? This should at least be discussed.

Response: As for the reviewer’s statement “Centrosome nucleated microtubules are typically highly dynamic and little modified. Therefore is the Ac-tub staining at the centrosome really MTs?”, it has been previously reported that tubulin acetylation is observed around the centrosome in some cell lines (J. Neurosci., 30: 7215–7226, 2010; PLoS One, 13: e0190717, 2018; etc.). N1-E115 is one of the cell lines in which tubulin acetylation is observed around the centrosome.

It is not surprising that “only acetylated tubulin goes down on WB,” as it has been previously reported that acetylated and detyrosinated tubulins are sometimes not synchronous (J. Neurosci., 23: 10662–10671, 2003; J. Neurosci., 30: 7215–7226, 2010; J. Cell Sci., 132: jcs225805, 2019., etc.). For instance, Montagnac et al. (Nature, 502: 567–570, 2013) showed that defects in the α-tubulin acetyltransferase αTAT1-clathrin-dependent endocytosis axis reduce only tubulin acetylation, resulting in a shift from directional to random cell migration. Although the details of the molecular function of Map7D1 are beyond the main purpose of this study, we plan to add a discussion of the reduced tubulin acetylation by Map7d1 knock-down based on the above.

- MAP7D1 and MAP7D2 depletion leads to subtle defect in cell migration and neurite outgrowth, which the author suggest is caused by reduced MT stability. However, MAP7 proteins have well characterized functions in kinesin-1 transport, and thus the phenotypes may well be caused by defects in kinesin-1 transport. Ideally the authors would do rescue experiments with FL or just the MT binding N-termini to separate these functions. Moreover this is needed to substantiate the claim of the authors that MAP7D1 effect on MT stability is not mediated by direct binding.

Response: As this comment largely overlaps with the comments raised by Reviewer #2, please refer to our responses to the comments of Reviewer #2.

- The authors do not refer well to published work. Several papers have published very similar work (especially to Fig1+2) and it would help the reader much if this would be discussed/compared along the results section and not briefly mention these in the results section. In addition, authors overstate the novelty of their results e.g. page 3: these proteins are not "functionally uncharacterized" nor are their expression patter and biochemical properties analyzed for the first time in this manuscript; page 8 "Although the physiological role of the C-terminal region of Map7D2 is currently unknow, ..." There is a clear function for the C-terminus for the recruitment/activation of kinesin-1.

Response: According to the reviewer’s suggestion, we plan to add a comparison with data on the Map7 family members presented in previous papers in the Results section and rephrase the relevant part regarding the physiological role of the C-terminal region of Map7D2.

Minor comments

- P6 Map7D3 also binds with its N-terminus to MTs, like other MAP7s (Yadav et al)

Response: According to the reviewer’s comment, we have revised this as “Map7D3 binds through a conserved region on not only the N-terminal side, but also the C-terminal side (Sun, 2011; Yadav et al., 2014).” (page 6, lines 4, 5)

- P7 "As Map7D2 has the potential to functionally compensate for Map7D1 loss" where is this based on?

Response: For clarity, we have rephrased this as “As Ma7D2 expression was upregulated upon suppression of Map7D1 expression, Map7D2 has the potential to functionally compensate for Map7D1 loss.” (page 7, line 17, 18)

- Fig2F quality of black-white images is low potentially due to conversion issues

Response: We thank the reviewer for pointing out these conversion issues, and we have made the necessary corrections.

Significance (Required):

At this stage the conceptual advance is limited. Part of the findings are not novel. The finding that MAP7s depletion have a different effect on MTs acetylation may be interesting to cytoskeleton researchers, although the potential mechanism has not been addressed experimentally or textually.

However, their conclusion that this leads to reduced MTs and then to cellar migration and neurite formation defects is not sufficiently supported by experimental evidence.

Response: We appreciate the reviewer’s comments. We believe that our revision plan will greatly improve the quality of our manuscript.

\*Referees cross-commenting***

I completely agree with reviewer #2: At this stage the paper's conclusions are not sufficiently supported by the data. Important will be to further characterize the effect om the MTs (do they really have a different effect) and to look at the possible involvement of the motor recruitment. Maybe that a 3 to 6 months revision time would have been more accurate.

Response: Please refer to our responses to the comments of Reviewer #2.

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:

Microtubule Associated Proteins (MAPs) are important regulators of microtubule dynamics, microtubule organization and vesicular transport by modulating motor protein recruitment and processivity. In the current manuscript the authors have characterized 2 members of the MAP7 protein family, MAP7D1 and MAP7D2. The authors characterized MAP7D2 expression pattern in the brain and its microtubule binding properties in vitro and in cells. In cells both proteins localize to the centrosome and to microtubules and upon depletion centrosome localized microtubules seem reduced, and cell migration and neurite outgrowth are increased. Surprisingly, they find that microtube acetylation (a common marker for stable microtubules) is reduced upon MAP7D1 depletion but not MAP7D2 depletion. Based on this finding the authors conclude that these proteins have a distinct mechanism in stabilizing MTs to affect cell migration and neurite outgrowth; MAP7D2 stabilizes by binding to MTs, whereas MAP7D1 stabilizes MTs by acetylation.

Main comments:

• Both MAP7 proteins show strong localization to the centrosome and to a lesser degree to MTs. Knockdown of either protein leads to reduced MTs around the centrosome, which lead the authors to conclude the MAP7s are stabilizing the MTs. However, the effect could just as well be an indirect effect due to a function of these MAPs at the centrosome. To address this authors could e.g. quantify microtubule properties in postmitotic cells. In addition, antibody specificity should be tested using knockdown of knockout cells, as this centrosome localization was not observed in Hela cells (Hooikaas, 2019; Kikuchi, 2018). Maybe this localization is specific to rat MAP7s or to the cell line used.
• Centrosome nucleated microtubules are typically highly dynamic and little modified. Therefore is the Ac-tub staining at the centrosome really MTs? I cannot identify MTs in the fluorescent images in 4C. Maybe authors could consider ac-tub/alpha-tub ratio in non centrosomal region (e.g. neurites). Moreover, as both Acetylation and detyrosination are associated with long-lived/stable MTs, it is surprising that only acetylated tubulin goes down on WB. Does this suggest that long-lived MTs are still present to normal level? If so, can one still argue that the loss of acetylation is the cause of the lower MT levels? This should at least be discussed.
• MAP7D1 and MAP7D2 depletion leads to subtle defect in cell migration and neurite outgrowth, which the author suggest is caused by reduced MT stability. However, MAP7 proteins have well characterized functions in kinesin-1 transport, and thus the phenotypes may well be caused by defects in kinesin-1 transport. Ideally the authors would do rescue experiments with FL or just the MT binding N-termini to separate these functions. Moreover this is needed to substantiate the claim of the authors that MAP7D1 effect on MT stability is not mediated by direct binding.
• The authors do not refer well to published work. Several papers have published very similar work (especially to Fig1+2) and it would help the reader much if this would be discussed/compared along the results section and not briefly mention these in the results section. In addition, authors overstate the novelty of their results e.g. page 3: these proteins are not "functionally uncharacterized" nor are their expression patter and biochemical properties analyzed for the first time in this manuscript; page 8 "Although the physiological role of the C-terminal region of Map7D2 is currently unknow, ..." There is a clear function for the C-terminus for the recruitment/activation of kinesin-1.

Minor comments:

• P6 Map7D3 also binds with its N-terminus to MTs, like other MAP7s (Yadav et al)
• P7 "As Map7D2 has the potential to functionally compensate for Map7D1 loss" where is this based on?
• Fig2F quality of black-white images is low potentially due to conversion issues

Significance

At this stage the conceptual advance is limited. Part of the findings are not novel. The finding that MAP7s depletion have a different effect on MTs acetylation may be interesting to cytoskeleton researchers, although the potential mechanism has not been addressed experimentally or textually.

However, their conclusion that this leads to reduced MTs and then to cellar migration and neurite formation defects is not sufficiently supported by experimental evidence.

Referees cross-commenting

I completely agree with reviewer #2: At this stage the paper's conclusions are not sufficiently supported by the data. Important will be to further characterize the effect om the MTs (do they really have a different effect) and to look at the possible involvement of the motor recruitment. Maybe that a 3 to 6 months revision time would have been more accurate.

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

Learn more at Review Commons

---

Referee #2

Evidence, reproducibility and clarity

In this study, the authors investigate 2 members from the MAP7 family Map7D2 and Map7D1. They first address the tissue distribution of Map7D2, by northern blotting using a variety of rat tissues. To complement their analysis, they also raised an antibody to look at the protein distribution. From their studies, they concluded that Map7D2 is abundantly expressed in the brain and testis. The authors went on to perform a series of functional assays . First, they biochemically demonstrated that rat Map7D2 directly binds to MTs by MT co-sedimentation assay. The MT binding domain was mapped to the N-terminal half. They performed MT turbidity assay to demonstrate enhanced MT polymerisation in the presence of Map7D2, suggesting that this Map stabilises MTs. The authors went on to characterise in detail the subcellular localisation of Map7D2 which was predominantly present in the centrosome and partially localised to MTs including within neurites from N1-E115 cells. Kikuchi et al. further revealed the overlap in expression between Map7D2 and another family member, Map7D1. The authors continued these studies by a series of functional studies in N1-E115 cells where they performed single or combined knock-downs of Map7D2 and Map7D1 and studied the levels of acetylated and detyrosinated tubulins and the effect of the knock-downs on migration and neurite extension. The main conclusion from this work was that Map7D2 and Map7D1 facilitate MT stabilization through distinct mechanisms which are important in controlling cell motility and neurite outgrowth. Map7D2 is proposed to stabilise MTs by direct binding whereas Map7D1 does it indirectly by affecting acetylation.

Major comments:

The main conclusion from this work that Map7D2 and Map7D1 facilitate MT stabilization and that this is necessary for correct migration and neurite extension has not been convincingly demonstrated. In my opinion, a more detailed study of MT properties to demonstrate a role in MT stabilisation would greatly benefit the work, eg. experiments using MT destabilising agents such as nocodazole. In addition, a series of experiments aiming to study MT dynamics would help to understand the function of these MT regulators. The authors proposed an elevation in microtubule dynamics to explain the increase in migration and neurite extension but no experimental proof was provided.

It has been previously demonstrated that loss of MAP7D2 leads to a decrease in axonal cargo entry to axons resulting in defects in axon development and neuronal migration. The C-terminus is necessary for this function as it mediates interaction with Kinesin-1 (Pan et al., 2019). Such mechanisms could also explain the defects in migration and neurite growth that the authors observed. This possibility has not been considered but instead, the subtle changes in total -tubulin led to suggest MT stabilisation as a key function without proof of causation. Could the authors provide some further experimental evidence to demonstrate that stability is the main contributor to the phenotypes observed? Eg. by rescuing migration and neurite phenotypes with a variant of MAP7D2 which cannot bind kinesin1.

A key conclusion proposed by the authors is that Map7D2 and Map7D1 facilitate MT stabilization through distinct mechanisms. Such different roles in MT stabilisation are important in controlling cell motility and neurite outgrowth. In my opinion, their data does not fully support this statement and the findings using MT readouts do not match the defects in migration and neurite growth. Loss of Map7D2 leads to a very subtle phenotype on -tubulin, while Map7D1 decreases both -tubulin and acetylated tubulin, but Map7D1 seems to have a milder or similar effect on migration and neurite growth than Map7D2. Furthermore, it would be expected that the combined loss of function would lead to a stronger phenotype in cell migration when compared to the single loss of functions due to their distinct roles on MT stability, however, this seems not to be the case.

Minor comments:

1. In the first result section, the author refers to Fig. S3 to suggest the expression of MAP7D2 in the cerebral cortex, however, there are no transcripts in the cerebral cortex according to the figure. Similarly, the immunofluorescence analysis done by the authors shows marginal expression of MAP7D2 in the cerebral cortex.
2. The authors use -Tubulin as a housekeeping gene in Fig. 3D, since Map7D2 is enriched in centrosomes this may not be the most appropriate choice.
3. According to the authors, knockdown of Map7D2 leads to a decrease in the intensity of -tubulin and Map7D1 (Fig. 4C and D). This data doesn't agree with the previous statement made by the authors where they show that Map7D2 knockdown or knockout did not affect Map7D1 expression by Western Blot Analysis (Fig. S2C and S5B)
4. Line 6 page 7 "Endogenous Map7D2 expression is suppressed in N1-E115 cells stably expressing EGFP-rMap7D2 and was restored by specific knock-down of EGFP-rMap7D2 using gfp siRNA (Fig. 3D)". No quantifications and stats are shown. Also, endogenous Map7D2 after knock-down of EGFP-rMap7D2 is not comparable to the control.
5. Line 8 page 7 "These results suggest that the expression of Map7D2 was influenced by changes in that of Map7D1" This statement seems in the wrong place, after the Map7D2 and EGFP-rMap7D2 experiment. Instead for clarity, it would be better placed after line 5 where the authors explain the effect of Map7D1 knock-down on the levels of Map7D2.
6. Line 8 page 8 "Although the physiological role of the C-terminal region of Map7D2 is currently unknown..." This statement seems not adequate as there are several studies reporting the role of the C-terminal region of Map7D2 in Kinesin1- mediated transport. The authors mention such studies in the discussion.
7. Line 6 page 9 " Further, the knock-down of either resulted in a comparable reduction of MT intensity (Fig. 4C and D) ..." This is not visible and/or justified by the images provided and would benefit from some sort of quantification at other regions such as neurites.
8. In Fig. 2B, a band corresponding to his6-rMAP7D2 of molecular weight >97 kDa co-sedimented with the microtubules. However, the cloned rMAP7D2 had a molecular weight of 84.82 kDa and the addition of 6XHis-Tag would add another 2-3 kDa, therefore, the final protein band observed should be less than 90 kDa. It would be beneficial if the authors could specify the molecular weight of the purified protein after the addition of the V5-his tag and/or if there was addition of amino acids due to cloning strategy.
9. In Fig. 2C, there is misalignment of the western blot with the panel or text underneath.
10. In Fig. 3C the inset from the first panel seems to correspond to a different focal plane than the main image.
11. In Fig. 4A, the cell type is not specified and is referred as "indicated cells", also the material and methods section seems to omit the specific cells used.
12. Fig. S6 is not mentioned in the results.

Significance

MTs play essential roles in practically every cellular process. Their precise regulation is therefore crucial for cellular function and viability. MAPs are specialised proteins that interact with MTs and regulate their behaviour in different manners. Understanding their precise function in different cellular contexts is of utmost importance for many biological and biomedical fields.

MAPs are well known for their ability to promote MT polymerization, bundling and stabilisation in vitro (Bodakuntla et al., 2019). Several members of the Map7 family have been shown to regulate microtubule stability. For instance, MAP7 can prevent nocodazole-induced MT depolymerization and maintain stable microtubules at branch points in DRG neurons (Tymanskyj & Ma, 2019). Ensconsin, the Drosophila Map, is required for MT growth in mitotic neuroblasts by regulating the mean rate of MT polymerization (Gallaud et al., 2014). However, this family of Maps seems to have diverse functions encompassing a variety of mechanisms, as exemplified by a series of studies demonstrating the involvement of MAP7 family proteins in the recruitment and activation of kinesin1 (Hooikaas et al., 2019; Pan et al., 2019) and in microtubule remodelling and Wnt5a signalling (Kikuchi et al., 2018). Further understanding of this family of Maps and how its members differ in their function is important and will help to advance 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 #1

Evidence, reproducibility and clarity

In this manuscript, Kikuchi et al describe the characterization of MAP7D2 and MAP7D1, two MAP7 family members in mouse with specific expression patterns. Focusing mostly on MAP7D2, they assess its expression pattern across the body and find that it is mostly expressd in certain neuronal subsets. They then characterize the MT-related properties of MAP7D2 based on previous knowledge of other MAP7 family members. They show that MAP7D2 binds MTs (via the N-terminus), determine the binding affinity, and show that it can stimulate MT polymerization (or stabilization) both in vitro and in vivo. Using a specific antibody, they localize MAP7D2 to centrosomes, midbody and neurites in N1-E115 cells. Functionally, they show that loss of MAP7D1/2 mildly affects microtubule stability as judged by acetyl-tubulin staining, and properties of these cells that rely on cytoskeletal elements such as cell migration and neurite growth. Interestingly, there might be a feedback loop regulating MAP7D1/2 expression , as knockdown of MAP7D1 upregulates MAP7D2.

Overall, the experiments and conclusions are very solid and convincing, such that I would not ask for further experiments. This is in part because the experiments are largely based on previous characterizations of other MAP7 family members, which are largely confirmed. The presentation of the data is also very clear.

Significance

I see the value of the study in the fact that,it provides solid and specific research tools for MAP7D1/2 which could be very useful for the microtubule/neuronal cytoskeleton community.

Referees cross-commenting

Reviewers 2 and 3 criticize that the evidence for an effect of MAP7D1/2 on MT dynamics is weak. I would agree in that ac-tub stainings and in vitro experiments are rather indirect. The experiments suggested by reviewer 2 should clarify this (esp. nocodazole should be easy). I also agree that an experiment addressing the potential involvement of kinesin-1 would help, the involvement of which seems to have been omitted by the authors. A kinesin-binding deficient mutant would add another MAP7D1/2 tool and increase the value for the community.

Note: This rebuttal 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

Reviewer #1:

This is the first such piece of data to come from human infective parasites in the field. Technically this is a feat - because the small number of parasites that are present per mL of human blood at any given time during infection with T gambiense. Nevertheless they manage to identify up to 14 unique VSGs per patient sample. And this raises the first theoretical question: can they extrapolate to the average diversity load per human?

This is an intriguing question that we would like to eventually answer, but we do not believe we can make this estimate from the data we currently have. We know our sampling is insufficient based on the correlation between parasitemia and diversity, and we do not have sufficiently precise estimates of parasitemia that could be used to extrapolate total diversity in the blood. Moreover, our analysis was only performed on RNA extracted from whole blood samples. Recent studies indicate that significant populations of parasites reside in extravascular tissue spaces, and our analysis did not address antigenic diversity in these spaces. We believe it is unlikely that the blood alone reflects the full diversity of VSG expression in an infection, and an estimate based only on blood-resident parasites (if possible) could be misleading.

this is important because the timing of sample collection (ie that it occurred within a period of weeks) suggesting that an initial group of infected tsetse infected these patients (rather than a small number of interactions between a bloodmeal and a new infection - generally in itself on the order of 1 month or so). If parasitemia is low and diversity limited, this would explain both why CATT works as well as it does (because really it shouldn't at all!) and perhaps even the chronicity of infection (in the sense that the organism is unlikely to "run out" even of complete VSGs, never mind mosaics). The paper would benefit from a direct discussion on this.

Indeed, the timing of sample collection could inform our interpretation of the data. However, sample collection occurred over a period of six months. More importantly, patients were in both early and late stage disease at the time of sample collection, so we cannot estimate how long any individual patient had been infected. We have added text (line 180) to highlight this fact. Because some patients were infected at least 6 months apart (if not much more than that), it is unlikely that patients were infected around the same time by a small group of infected tsetse flies. Reviewer #1 introduces an interesting point about the efficacy of the CATT diagnostic test as it relates to antigenic diversity. We discuss CATT sensitivity in the introduction (lines 115-120) as well as the discussion, where regional sensitivity differences are mentioned (lines 715-718). Given uncertainty about total diversity and time since initial infection, we have refrained from speculating about how diversity/timing could affect CATT sensitivity.

An interesting feature of this new study is the apparent bias to type B N-terminal domain VSGs as well as the discovery that two patients share a specific VSG isolate (though it is not mentioned whether they are related by distance etc). This raises the possibility of substrains with different VSG archives that vary by geography.

We found two VSGs which were expressed in more than one patient. One was expressed in two patients from the same village (village C) while the other VSG was common between two cases originating from villages C and D, some 40 km apart. We agree that our data generally support the possibility that the VSG archive might vary geographically. We have performed additional analyses suggested by reviewer #2 that support this idea: we have now shown that Tbg patient VSGs classified in this study, which originated from the DRC, are distinct from the VSGs encoded by the reference strain Tbg DAL 972 which was isolated in Cote d’Ivoire. We mention this possibility on lines 721-724.

Alternatively it suggests that perhaps type B VSGs are picked up differentially by serology (and there the one feature of type B VSGs that could be shared, with regards to detection, is the O-hexose decoration on a number of type B VSG surfaces. Could CATT be detecting elements common to sugar decorated VSGs? Experimentally this is something that can be tested even with mouse infection materials.

This is indeed an intriguing possibility. We mention this in the discussion (lines 772-778): “In T. brucei, several VSGs have evolved specific functions besides antigenic variation [74]. Recently, the first type B VSG structure was solved [75], revealing a unique O-linked carbohydrate in the VSG’s N-terminal domain. This modification was found to interfere with the generation of protective immunity in a mouse model of infection; perhaps structural differences between each VSG type, including patterns of glycosylation, could influence infection outcomes.” While this is an experimentally tractable explanation for the type B VSG bias we observe, we believe such experiments are beyond the scope of the current paper.

Side comment: are the common VSGs mutated between patient samples?

We classified VSGs as common between patient samples if they had >98% nucleotide sequence identity as well as meeting the other quality cutoffs such as 1% expression level and consistency across technical replicates. This identity cutoff still allows for several mismatches between sequences, which we do occasionally observe. However, we cannot confidently rule out that the “mutations” we observe are sequencing or PCR errors. Thus, we cannot say for sure if there are mutations between common VSGs.

Reviewer #2

---

1.Throughout the manuscript you observe 'diversity' in expressed VSG and its existence becomes a principal conclusion. I feel that the meaning of diversity and its significance is not sufficiently explained for the reader. In the abstract (l48) you say that there is 'marked diversity' in parasite populations. Presumably you mean parasite infrapopulations, i.e. within patients, not across the DRC? In any case, what is 'marked' about it, and relative to what? Why does it matter that there are multiple expressed VSG in a single patient at one time? Is this not a reasonable expectation for a population of (presumably) clones capable of switching the expressed VSG? How is this different to the view typical of the literature since 1970 that one VSG dominates while others wait in the background at low frequencies. If 'diversity' is the conclusion, then you need to define it and explain its significance more.

When we refer to diversity, we do mean infrapopulations of parasites within patients, or individual animals in this case, rather than across the DRC. We have edited the text to make this clear (see below). However, the study which benchmarked the application of VSG-seq to quantify VSG expression in vivo during mouse did not support the previously-held view that one VSG dominates while others wait in the background at low frequencies. Frequently we observe a handful of VSGs present at 10-20% of the population at any timepoint, and many VSGs (~50% of all detected variants) present at “In a proof-of-principle study, we used VSG-seq to gain insight into the number and diversity of VSGs expressed during experimental mouse infections [30]. This proof-of-principle study revealed significant VSG diversity within parasite populations in each animal, with many more variants expressed at a time than the few thought to be sufficient for immune evasion. This diversity suggested that the parasite’s genomic VSG repertoire might be insufficient to sustain a chronic infection, highlighting the potential importance of recombination mechanisms that form new VSGs.

2.Following on from 1., why does the analysis deal in counts of distinct VSG or N-terminal domains, and not then progress to their relative expression? The expression data are in Supp Table 3 and they show that, in most cases where many VSG are observed in the same patient, 1-3 of these are 'dominant', i.e. they account for >50% of the population.

The VSG-seq analysis pipeline does estimate the relative expression level of each identified variant in the population, and this information is available in the supplemental data (Supplemental Figure 1, Supplemental Table 3). However, we chose not to rely on these measurements too heavily because there was some variation between Tbg technical replicates, which is shown in the supplemental heatmap (Supplemental Figure 1). Replicate three tends to not agree with the first two replicates. We suspect that this was due to the order of sample processing and the fact that the parasite-enriched cDNA sample was repeatedly freeze-thawed between library preparations for technical replicates. Additionally, because our sampling did not reach saturation, some VSGs are not detected in all replicate libraries, making it difficult to estimate their abundance.

We have added a discussion of these issues to the text on lines 431-433: “Because our sampling did not reach saturation, resulting in some variability between technical replicates, we chose to focus only on the presence/absence of individual VSGs rather than expression levels within parasite populations.”

Figure 1 deals in VSG counts, but I would then expect another figure to illustrate the reality that only a minority of these observed VSG are likely to be clinically relevant (i.e. the subject of the immune response). This impacts the 'diversity' conclusion, as given in the discussion (ll 657-9), because you cannot afford to treat all these VSG equally when their abundances are quite different.

---

We agree that relative expression level is a useful metric, but absent longitudinal sampling it is impossible to determine which VSGs are clinically relevant as defined by the reviewer: low abundance VSGs at one time point may be the predominantly expressed variant at another. Moreover, the threshold for triggering an anti-VSG antibody response remains unknown. Thus, we have chosen to treat all detected variants equally.

3.How related are these VSG? Were you able to ensure unique read mapping to the VSG assembly? Can you show that reads mapped to a single VSG only and therefore, that the RPKM values are reliable?

Our analysis accounts for the fact that VSGs can be very similar. We only considered uniquely mapping reads in our VSG-seq analysis. We also account for mappability in our quantification, so VSG sequences that are less unique (and thus have fewer uniquely mapping reads) are not artificially underrepresented in estimates of relative expression. We have specified the parameters used for alignment (line 274) in the methods.

4.The authors observed no orthology between expressed VSG and DAL972 genes. This is really interesting and deserves closer attention. Presumably there is microhomology? For T. brucei VSG, with constant recombination, we would predict that a comparison of the VSG in West and Central Africa would reveal a pattern of mosaicism, such that individual sequences in DRC would break down into motifs present in multiple genes in the West African reference. Question is, how many genes? What does that distribution look like? What is the smallest homology tract? There is an opportunity here to comment on how VSG repertoires diverge under recombination. How much of the expressed VSG sequence is truly unrepresented in the West African reference (or other T.b.gambiense genome sequences available in ENA). I can believe that none of the N-terminal domains in these data are present intact in DAL972, but I cannot believe that their components are not present without evidence.

We appreciate the reviewer’s suggestion to look at this more closely. We have performed additional analyses to address sequence similarity, or lack thereof, between the assembled DRC patient VSG and the West African reference TbgDAL972. We ran a nucleotide BLAST of expressed VSGs against the TbgDAL972 genome reference sequence pulled from TriTrypDB.org (release 54). We have added a supplemental figure depicting the results of this analysis (Supplemental Figures 6 and 7). Briefly, our analysis shows that most of the N-termini we identified have no significant similarity to DAL972 VSGs, even with very permissive search parameters. There are frequent hits in the VSG C-termini, however, which might be expected. Most BLAST hits are short spans 98% identity are short 20-25 bp regions. Given the large divergence from the reference, we were unable to infer any patterns of recombination in the VSGs. However, we believe this analysis supports our claim that the N-termini of VSGs assembled from DRC patients are novel, with their component parts largely unrepresented in the West African reference genome.

Figure 4 compares NTD type composition in the DRC data with previously published mouse experiments. The latter take place over very short timescales in maladapted hosts, while the timescales of the latter in natural hosts are unknown but plausibly very much longer. So are these data really comparable and are we learning anything from their comparison, given that the most likely explanation for the NTD bias in expressed VSG is the underlying genomic composition?

---

Indeed, this is our intended conclusion from figure 4. The figure is meant to illustrate our claim that the expressed VSGs in each experimental set reflect the underlying genomic composition of their corresponding reference strains, despite fluctuations over time. The language and legend for Figure 4 has been clarified to emphasize this point. We have emphasized in the text that it is unknown whether these fluctuations occur over time in much longer natural infections.

---

6.Please comment on the technical reproducibility of the data, there are multiple instances in Supp Table 3 where technical replicates expressed different VSG.

Three RNA-seq library technical replicates were prepared for each individual gHAT patient RNA sample. Replicates were prepared in batches together so all 1’s were done on the same day, for example. The original parasite-enriched cDNA sample was frozen and thawed between each batch. We suspect that the cDNA degraded after repeated freeze-thaw cycles, which is why replicate three tends to not agree with the other two as can be seen on the heatmap in supp fig. 1 and the expression data in supp table 3. We also suspect the fact that our sampling did not reach saturation resulted in the detection of different VSGs in individual replicate preps. We have edited the methods and mentioned this variability in the results section to communicate this issue more transparently.

• Lines 395-397 “Using RNA extracted from 2.5 mL of whole blood from each patient, we prepared libraries for VSG-seq in three separate batches for each technical replicate.”
• Lines 431-433: “Because our sampling did not reach saturation, resulting in some variability between technical replicates, we chose to only focus on the presence/absence of individual VSGs rather than relative expression levels within the population”

---

Reviewer #3

1. In line 499, the authors conclude the due to the expressed VSGs being different in the blood and CSF being difference it may indicate that different organs harbor different VSG sets. Given that this is n=1 for patient samples I think this is too speculative a statement. There is also no indication as to whether the samples were taken at the same time or not.

This is absolutely correct. The precise timing of CSF sample collection is unknown for these samples. It likely occurred within hours to days after blood collection, but even on this short time scale, the unique CSF repertoire could represent the antibody-mediated clearance of one VSG population and replacement with another. We have scaled back our language and only point out that there are unique VSGs in this space (Lines 522 – 524).

I think that the authors need to be very careful as to the conclusions drawn about VSG expression over time in terms of hierarchy and N-terminal fluctuations. For any conclusions to be drawn on the hierarchy of VSG expression more data points are needed taken over time (this is obviously challenging when looking at patient samples). I find it too speculative to draw any conclusions when single time points are assessed and the assumption on the progression of the infection depends on whether it is a Tb or Tbr.

Reviewer #2 also pointed this out. We agree and have attempted to limit definitive conclusions in the text and instead discuss multiple possible explanations behind our observations.

I found some of the figure legends a bit terse. For example, in Figure 1 C, what do the black circles and lines represent? Perhaps a little more detail would help the reader.

Clarified legends for UpSet plots in figures 1C and 3C as follows: “The intersection of expressed VSG sets in each patient. Bars on the left represent the size of the total set of VSGs expressed in each patient. Dots represent an intersection of sets with bars above the dots representing the size of the intersection.”

In figure 2, I found it difficult to distinguish between the orange and dark red in (A) and the two lighter blue colors.

We have changed N-terminal type color palette for all plots to make red and blue hues more distinctive.

In line 389 – estimate

Corrected

In line 498 - should be reference been to figure 2C?

This should be a reference to Figure 3B. We have corrected the reference.

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 work, So and Sudlow et al have used an established methodology - VSG-seq to assess the expressed VSG diversity in 12 patients infected with T. brucei gambiense. As with what is seen in mouse models, there is a diversity in VSG expression seen in patients. The application of this technology has not previously been used on patient samples and is now validated as a valuable tool to study antigenic variation in human populations. The authors have found that in addition to the VSG diversity seen there was a significant bais towards B type N-terminal domains and a restricted C-terminal types. This work, although on a small sample group, is an important step forward to applying this technology to understanding trypanosome immune evasion in the field.

Major comments:

I think that overall, the key conclusions on the expressed VSG diversity and that there are geographical variations are convincing and would agree with the conclusions that it is now feasible to study antigenic variation in the field. But given the sample size the I feel that two of the findings are overstated and should at least be qualified as speculative.

1.In line 499, the authors conclude the due to the expressed VSGs being different in the blood and CSF being difference it may indicate that different organs harbor different VSG sets. Given that this is n=1 for patient samples I think this is too speculative a statement. There is also no indication as to whether the samples were taken at the same time or not.

2.I think that the authors need to be very careful as to the conclusions drawn about VSG expression over time in terms of hierarchy and N-terminal fluctuations. For any conclusions to be drawn on the hierarchy of VSG expression more data points are needed taken over time (this is obviously challenging when looking at patient samples). I find it too speculative to draw any conclusions when single time points are assessed and the assumption on the progression of the infection depends on whether it is a Tb or Tbr. I don't believe that any other experiments are needed and the statistical analysis is adequate.

Minor comments:

I found some of the figure legends a bit terse. For example, in Figure 1 C, what do the black circles and lines represent? Perhaps a little more detail would help the reader.

In figure 2, I found it difficult to distinguish between the orange and dark red in (A) and the two lighter blue colors.

In line 389 - estimate

In line 498 - should be reference been to figure 2C?

Significance

Overall, this is an interesting study and shows the practical application of VSG-seq on the study of human infections. There is clearly interesting biology to be learned about both Tbg and Tbr infections and immune evasion by these parasites - which can now be done with the development and application of these technologies. I am a molecular cell biologist who specialises in trypanosome 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

So et al. have analyzed the expression profiles of T.b.gambiense VSG genes in 12 natural human infections in DRC during a six month period of 2013, and compared these results to existing data for T.b.rhodesiense VSG and previously published data from mice. They use the VSGseq approach developed by the Mugnier lab over the last few years to good effect and provide a description of the expression profiles using phylogenetic and network approaches. The main conclusions are that parasite infrapopulations in each patient expression largely mutually exclusive VSG cohorts, with a couple of exceptions where patients 'shared' identical VSG transcripts. The authors note that these congolese VSG are not comparable with the West African T.b.gambiense reference sequence, and there is a pronounced bias in the systematic composition of expressed VSG (towards 'B-type VSG') that is not observed in other T. brucei subspecies. These latter observations lead to the suggestion that there may be substantial variation in expressed VSG repertoire among T. brucei populations that could have important consequences for pathology, although the spatial or temporal scale upon which this variation could be expected to occur cannot be inferred from these data. Overall, a competent study and a welcome addition to, if not extension of, recent work describing the dynamics of VSG expression in multiple African trypanosomes.

Major points:

1.Throughout the manuscript you observe 'diversity' in expressed VSG and its existence becomes a principal conclusion. I feel that the meaning of diversity and its significance is not sufficiently explained for the reader. In the abstract (l48) you say that there is 'marked diversity' in parasite populations. Presumably you mean parasite infrapopulations, i.e. within patients, not across the DRC? In any case, what is 'marked' about it, and relative to what? Why does it matter that there are multiple expressed VSG in a single patient at one time? Is this not a reasonable expectation for a population of (presumably) clones capable of switching the expressed VSG? How is this different to the view typical of the literature since 1970 that one VSG dominates while others wait in the background at low frequencies. If 'diversity' is the conclusion, then you need to define it and explain its significance more.

2.Following on from 1., why does the analysis deal in counts of distinct VSG or N-terminal domains, and not then progress to their relative expression? The expression data are in Supp Table 3 and they show that, in most cases where many VSG are observed in the same patient, 1-3 of these are 'dominant', i.e. they account for >50% of the population. Figure 1 deals in VSG counts, but I would then expect another figure to illustrate the reality that only a minority of these observed VSG are likely to be clinically relevant (i.e. the subject of the immune response). This impacts the 'diversity' conclusion, as given in the discussion (ll 657-9), because you cannot afford to treat all these VSG equally when their abundances are quite different.

3.How related are these VSG? Were you able to ensure unique read mapping to the VSG assembly? Can you show that reads mapped to a single VSG only and therefore, that the RPKM values are reliable?

4.The authors observed no orthology between expressed VSG and DAL972 genes. This is really interesting and deserves closer attention. Presumably there is microhomology? For T. brucei VSG, with constant recombination, we would predict that a comparison of the VSG in West and Central Africa would reveal a pattern of mosaicism, such that individual sequences in DRC would break down into motifs present in multiple genes in the West African reference. Question is, how many genes? What does that distribution look like? What is the smallest homology tract? There is an opportunity here to comment on how VSG repertoires diverge under recombination. How much of the expressed VSG sequence is truly unrepresented in the West African reference (or other T.b.gambiense genome sequences available in ENA). I can believe that none of the N-terminal domains in these data are present intact in DAL972, but I cannot believe that their components are not present without evidence.

5.Figure 4 compares NTD type composition in the DRC data with previously published mouse experiments. The latter take place over very short timescales in maladapted hosts, while the timescales of the latter in natural hosts are unknown but plausibly very much longer. So are these data really comparable and are we learning anything from their comparison, given that the most likely explanation for the NTD bias in expressed VSG is the underlying genomic composition?

6.Please comment on the technical reproducibility of the data, there are multiple instances in Supp Table 3 where technical replicates expressed different VSG.

Minor points:

1. Type 'estimates' line 389

Significance

The significance of this work relates to the application of VSG expression profiling to natural human infections, something not previously done largely because human infections are rare and materials difficult to obtain. The approach and the conclusions are not novel and do not represent substantial advances on previous efforts, but have an important aspect in confirming for natural infections what has been observed in quite artificial experimental settings. Sample size is small and this means that the conclusions remain speculative and cannot readily be extended to all HAT settings. This is not a criticism, since the analysis of any human samples is progress, but it does mean that the study raises interesting questions (e.g. variation across the population in N-terminal domain usage) rather than providing definitive conclusions. It is likely to interest trypanosome biologists with a specific interest in antigenic variation.

My own field concerns trypanosome genomics and the evolutionary dynamics of variant antigen genes.

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 paper by Mugnier and colleagues describe the repertoire of VSGs present within a cohort of human HAT cases that occurred at relatively close geographical distance.

VSG repertoires were first described by the senior author a few years ago already, from mouse infection data. This is the first such piece of data to come from human infective parasites in the field. Technically this is a feat - because the small number of parasites that are present per mL of human blood at any given time during infection with T gambiense. Nevertheless they manage to identify up to 14 unique VSGs per patient sample. And this raises the first theoretical question: can they extrapolate to the average diversity load per human? this is important because the timing of sample collection (ie that it occurred within a period of weeks) suggesting that an initial group of infected tsetse infected these patients (rather than a small number of interactions between a bloodmeal and a new infection - generally in itself on the order of 1 month or so). If parasitemia is low and diversity limited, this would explain both why CATT works as well as it does (because really it shouldn't at all!) and perhaps even the chronicity of infection (in the sense that the organism is unlikely to "run out" even of complete VSGs, never mind mosaics). The paper would benefit from a direct discussion on this.

An interesting feature of this new study is the apparent bias to type B N-terminal domain VSGs as well as the discovery that two patients share a specific VSG isolate (though it is not mentioned whether they are related by distance etc). This raises the possibility of substrains with different VSG archives that vary by geography. Alternatively it suggests that perhaps type B VSGs are picked up differentially by serology (and there the one feature of type B VSGs that could be shared, with regards to detection, is the O-hexose decoration on a number of type B VSG surfaces. Could CATT be detecting elements common to sugar decorated VSGs? Experimentally this is something that can be tested even with mouse infection materials.

Side comment: are the common VSGs mutated between patient samples?

Significance

Significance: high in the sense that this is the first in human field study of a disease that has been studied quite a lot in mouse models. Clearly from this work, there is still a lot to be learned from studying a disease in context.

Audience: parasitologists

My own expertise: parasitology and immunology

Referees cross-commenting

Nothing substantial to add. From the comments (all of which are worthwhile) I would suspect this would require minor revision.

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Note: This rebuttal 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

Reviewer #1 (Evidence, reproducibility and clarity): Thank you for the opportunity to review "Population-level survey of loss-of-function mutations revealed that background dependent fitness genes are rare and functionally related in yeast" by Caudal et al. This manuscript reports on the genetic background-dependent traits resulting from natural variation. Authors use 39 natural isolates of the budding yeast (S. cerevisiae) and apply transposon saturation mutagenesis approach to analyze fitness due to loss of function mutations. They identified background and environment dependent genes. They estimate that background specific rewiring is rare and represents instances of bridging between bioprocesses as well as connecting functional related genes. Major comments

1. Authors filtered strains based on whole chromosome aneuploidies, but what about chromosome arm aneuploidies. Were they detected and if so how were they handled? This should be discussed.

We did not detect any chromosome arm aneuploidies. In fact, if any significant segmental duplication were present in any of the tested strains, we would have observed changes of gene essentiality for multiple successive ORFs, which was not the case.

How does chromatin structure variation across different genetic backgrounds affect the results of the screen? Is this a confounding variable? This should be discussed.

We thank the author for raising this interesting point. There are two aspects to take into consideration. First, transposon insertion is biased by nucleosome occupation, as is more or less expected. In previous screens and in our data, this bias is translated by the lower insertion density in the promoter in addition to the ORF for essential genes. If the nucleosome occupancy were conserved across different genetic background, this insertion bias won’t be a confounding factor as the same gene will share the same bias across different genetic backgrounds. Second, if the nucleosome occupancy is variable across different genetic backgrounds, it could potentially lead to some background-specific insertion biases, however it is difficult to know whether it would be the cause or the consequence of the mutation. In any case, currently there is no chromatin structure data across different genetic backgrounds available and this could be a direction for future research.

On page 7 authors discuss the involvement of other biological processes in addition to respiration and mitochondrial function. It is not clear what they are referring to. This should be clarified in the main text.

We clarified this point in the modified ms.

It would be useful to annotate the functional information discussed in the text directly on the network in Fig. 4 A and B.

We included annotations on the networks (see Fig. 4 and Fig. S4) as suggested in the modified ms.

On page 9, authors should comment on the origin of ACP and CLG strain that would result in the similarity of their fitness profile to S288C which they note as an exception.

ACP is an isolate from Russian wine and CLG is a clinical isolate from UK. In terms of the overall genetic diversity, these two strains are not closely related to the reference strain S288C. As for other profiles, no correlations were observed between the background-dependent mutant fitness variation and their genetic origins.

On page 10 authors discuss that background-specific fitness genes can belong to protein complexes. Can authors test this formally by looking at the overlap with the protein complex standard or protein interaction standard? This would strengthen this statement.

Due to the low number of cases, it is impossible to test this using protein complex standards as the size of the terms are too small as well as the sample size. However, the enriched SAFE terms are in general representative of biological processes which includes multiple protein complexes with similar functions. The genes enriched for each SAFE term is further broken down to specific GO terms, as indicated in Table S4.

Authors should discuss the reasons why transcription & chromatin remodeling and nuclearcytoplasmic transport, are anticorrelated with genes involved in mitochondrial translation in terms of their fitness profiles and the implications for the evolution of environment-dependent fitness genes.

These observations were new and we are currently looking for potential explanations to this effect. Unfortunately, there is no obvious explanation we can think of and discuss at this point. More data and further experiments are needed to have some clues about this observation.

Authors discuss the limitation of the Hermes system however couldn't they test this system with a different inducible promoter such as estradiol regulated promoter to remove the effect of galactose metabolism?

For the Hermes system to work effectively, we need a highly expressed promoter system that is also inducible and GAL1 is the strongest available. As for the estradiol system, first it requires the induction machinery to be integrated in the strain and that will significantly limit the scaling of the project, and second, the maximum induction level is significantly lower than that from the GAL1 system, as is recently shown in Arita et al., MSB 2021. For these reasons, the effect of galactose metabolism is inevitable using any transposon system at present.

Minor comments All figures should contain the appropriate colour bars and legends. For example, Figure S5B relies on the colour bar in Figure 5C but it should have its own colour bar.

We modified the figures as suggested.

Reviewer #1 (Significance): This work provides a comprehensive survey of the variation in natural isolates of yeast and would be interesting to a broad audience studying the genotype-to-phenotype relationship. It is the first study that systematically assessed the fitness effect of loss of function mutations across a large panel of natural isolates providing novel insight into the background specific and environment dependent genes. This represents a valuable resource for the community to ask questions about natural variation in yeast. My expertise is in complex genetic networks in yeast and genome evolution.

Reviewer #2 (Evidence, reproducibility and clarity): For decades, geneticists have used loss of function (LoF) mutations to unravel the molecular bases of phenotypic variability. However, a common concern is to what extent the phenotypes observed in a strain or accession recapitulates what happens at the species level. In not few cases, anecdotal evidence show that an observed mutant phenotype is not recapitulated in another strain, presumably due to the "strain background". Recent efforts using different strains of Saccharomyces cerevisiae have addressed the problem, but the number has been limited. Here, Elodie Caudal et al. use an ingenious transposon-saturation strategy to carry out a large-scale, genome-wide screen of LoF mutations in 39 strains. Based on a competitive-pooling strategy, authors estimate the probability of 4,469 genes being essential, compared to the reference S288C laboratory strain. Background-specific effects were in general rare. Around 15% of these genes show an essential phenotype which is dependent on the strain background, most of them showing continuous variation across all backgrounds and one third being specific of only one strain. Such background specific genes are functionally related and are under relaxed purifying selection and show "intermediate" integration in genetic-interaction networks compared to essential and non-essential genes. The manuscript is very easy to follow, and the experimental/statistical procedures are transparent and in general well described. Major comments

1. A limitation of the transposon saturation strategy is the need of galactose as the carbon source, which confounds scoring of genetic background effects. The study would highly benefit from any kind of orthogonal validation or phenotype predictions, beyond the BMH1 case presented (Fig S5). Few options would be direct testing of lethal/sick phenotypes of clean gene knockouts for discussed hits (Fig 5) in several strains and conditions including galactose, testing few of the transposon libraries under different conditions to validate the environment nature of the continuous behavior, or testing the predictive power of the method using data or strains used in Galardini 2019 (ref. 25).

As all three reviewers suggested that validation of our predicted probability score should be supported by experimental data, we performed orthogonal validations for 8 genes across 17 backgrounds. We have included the new results in the revised ms.

Showing the degree of replicability of the entire procedure would also help, form transposon insertion to phenotypic comparisons. If we understood correctly, this was indeed done for isolate AKE. What is the correlation of their probability scores?

The AKE strain was done twice due to the mixed haploid/diploid profile, as mentioned in the text. In this case the reproducibility in terms of probability scores is expected to be lower. We plotted the predicted probability values for the two reps (attached below) and calculated the Pearson’s correlation. The correlation coefficient is 0.86 (P-value

The use of "fitness genes" is confusing, since the main phenotypic output here scored for each gene is LoF lethality, or more specifically the probability of being lethal or non-essential. Lethality or essentiality would be a more appropriate concept throughout. A next step would indeed be to quantify the phenotypic effects in a more quantitative manner (which is generally used while referring to a gene's fitness effect).

We clarified this point in the revised version and use “predicted fitness variation” instead of “fitness genes”.

Some minor comments -Considering that part of the signal is coming from the specific environment tested, one would expect some degree of clustering among related strains based on their gene-essentiality probability (Fig2), given that growth phenotypes correlate well with strain origins when tested under different environments (Warringer et al., 2011). Please discuss.

In Warringer et al. 2011, the correlation was more pronounced between species (S. paradoxus vs. S. cerevisiae) than intraspecifically. Moreover, it was based on a very small sample size. In fact, multiple more recent studies have shown that the growth phenotypes across a large number of conditions between strains in S. cerevisiae is not correlated with their genetic origins (Peter et al. Nature 2018). Indeed, it is not unexpected that the gene-essentiality probability profiles are not correlated with their origins.

-Galactose is not a non-fermentable carbon source (pg 11, pg12). It is true that flux trough the fermentative pathways is lower and that the respiratory pathways are induced in galactose, when compared to growth on glucose, but galactose is readily fermented under low oxygen conditions. Indeed, variation in the regulation of these pathways could explain the environmental effects detected.

The reviewer raised a good point. While galactose is not a non-fermentable carbon source, the entry of galactose into glycolysis requires the respiration pathways and rho-/rho0 yeast mutants are unable to grow on galactose as the sole carbon source. We clarified this in the new version of the ms.

-Examples on FigS3 were useful for a better intuition of how the actual data looks like. Perhaps some of this belongs in Fig1.

Schematic presentation of the insertion profiles is already shown in Figure 1C. Due to the limited size of Figure 1, we kept Fig S3 as it is in the new version of the ms.

Fig2, restrict the #insertions label to the actual limits for the set of 39 strains. Currently, it seems there are strains with fewer than 100K and no strain with 300K insertions.

We thank the reviewer for pointing this out, it was a scaling problem and we fixed it.

-pg5 paragraph 2, a line on how representative is the set of 106 isolates and again later for the final data set of 39. Which main clades are missing or perhaps overrepresented?

Compared to the original 106 isolates, the final 39 isolates are still broadly representative of the species diversity, albeit some of the most divergent clusters, such as isolates from the French Guiana and from China, are underrepresented. We included this comment in the revised version.

-pg6 paragraph 1, should be 106 or 107?

It was 106 plus the reference strain. This point is clarified now in the new ms.

-pg14 line2, is OD of 0.5 correct or was also 0.05 as in galactose? This is relevant, since it would change the competitive selection regime under galactose or glucose (more generations under glucose in the latter case). For clarity, authors could here state an approximate number of cell divisions in each medium.

The OD of 0.5 was correct as this step was only intended as a “recovery phase” and was used to increase the mutant pool for sequencing. We also clarified this point in the text. -pg14 line 2, correct wording "to enrich for cells the transposon.."

We clarified this point in the revised version.

Reviewer #2 (Significance): While recent previous studies have measured genetic background dependent effects of gene mutations at the genome-wide level, this is the first study addressing the problem at the broader population level. Confirming that such effects are in general rare, even at this broad level, is a significant advance in the field. It is limited in the number of environmental conditions and subsequent insights (as in Galardini 2019, ref #25) and in more mechanistic views of specific allele interactions (as in Mullis 2018, ref #5). We feel, however, that these directions would already be out of the scope of the well-framed question here addressed. Because of the problem addressed and tackled in an ingenious and comprehensive manner, this manuscript will attract the attention of a broad audience of geneticists, genome and systems biologists. Our main expertise is in yeast genetics and functional genomics. **Referee Cross-commenting** Reviewer #1 commented the possibility that insertion density could be determined by local chromatin status instead of gene essentiality, given that transposon insertion occurs more often at nucleosome free sites (point 2). While the insertion pattern around the essential gene's vicinity is convincing, we agree that it would help to show that these phenomena are independent from one another, or that this issue must at least be discussed. The seeming need of further experimental or analytical validation was raised by reviewers #1 and #3. As mentioned above, we performed orthogonal validations for 8 genes across 17 backgrounds and we included the results in the revised ms.

Reviewer #3 (Evidence, reproducibility and clarity): In this manuscript, Caudal et al tested differences in gene knockout phenotypes across genetically diverse yeast strains using a transposon system. After initially querying 106 strains, most tested strains were removed from further consideration due to low transposon insertion numbers, aneuploidies, or other issues. The authors used the remaining 39 strains to identify a set of 632 genes that are required for normal growth in some genetic backgrounds but not in others. These context-dependent fitness genes are enriched for genes with a role in respiration, which could be because the experiment is performed using galactose as carbon source. Further analysis of potential environment-dependent fitness genes revealed two separate groups of genes that were anti-correlated in their fitness profiles. I found this an interesting paper, that explores differential gene essentiality (fitness) across diverse yeast strains. The authors give a detailed description of their findings, thereby differentiating between "environmental" and "genetic background" factors. The paper is well-written and the results are clearly presented. I have only two main concerns, both regarding the quality of the produced data: Major comments:

• Looking at the differential fitness scores in the supplementary data, none of the 57 genes that are known to show differential essentiality between S288C and Sigma1278b (Dowell et al., 2010, Science) appear to be identified as having differential fitness in the transposon screen. The authors mention that some of these genes have a severe fitness defect when deleted in the nonessential background and that some are only partially essential. Although this is certainly true for specific cases, deletion mutants of most of these 57 genes show a large difference in fitness between S288C and Sigma, and this thus doesn't sufficiently explain the complete lack of validation of 57 known positive cases. I think the authors need to further clarify why these known positive controls are not identified in their screen.

In Dowell et al. 2010, the essentiality was determined by tetrad dissection comparing S288C and Sigma, and as shown in the supplemental data, ~1/3 out of the 57 are in fact extremely sick in one background and non-viable in the other. This strong fitness defect cannot be distinguished using the transposon method. More recently in Hou et al. PNAS 2019, it has been shown that ~15 out of the 57 original cases were due to chromosomal genetic modifiers, which again, mainly concerned the “domain essential” effect that we also captured in our data. An addition, 8 hits out of the 57 were shown to be related to mitochondrial genomes in Edward et al. PNAS 2014, and due the galactose condition we used, these cases were not detected. Other undetected cases were due to the low coverage in the corresponding regions in either one or both backgrounds.

• Related to the previous point, the authors perform no secondary validation of identified context-dependent essential genes. They show that they can recapitulate known sets of essential and nonessential genes in S288c, but given my previous point, it is not clear how well their logistic model works for predicting differential gene essentiality/fitness. In my opinion, experimental validation of a subset of the identified differential fitness genes is needed to be able to be confident about the results.

As already mentioned above, new experiments were performed in order to validate a subset of the identified differential fitness genes. The results were included in the revised version of the ms.

Minor comments:

• The authors provide lots of data spread over many columns in the supplementary tables. However, a description of what is in each column is missing, and without it, it is not always possible to understand the data.

We added column annotations in the spread sheets as suggested.

• I didn't understand the sentence at the bottom of page 5: "the number of insertion drops from -100 bp prior to CDS and extends to - 100 bp until the terminator region". Perhaps the authors can rephrase.

We clarified this point in the revised version.

Reviewer #3 (Significance): To my knowledge, this is the first paper exploring gene essentiality across a large number of genetically diverse yeast strains. This paper will be of interest to a broad range of geneticists.

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, Caudal et al tested differences in gene knockout phenotypes across genetically diverse yeast strains using a transposon system. After initially querying 106 strains, most tested strains were removed from further consideration due to low transposon insertion numbers, aneuploidies, or other issues. The authors used the remaining 39 strains to identify a set of 632 genes that are required for normal growth in some genetic backgrounds but not in others. These context-dependent fitness genes are enriched for genes with a role in respiration, which could be because the experiment is performed using galactose as carbon source. Further analysis of potential environment-dependent fitness genes revealed two separate groups of genes that were anti-correlated in their fitness profiles.

I found this an interesting paper, that explores differential gene essentiality (fitness) across diverse yeast strains. The authors give a detailed description of their findings, thereby differentiating between "environmental" and "genetic background" factors. The paper is well-written and the results are clearly presented. I have only two main concerns, both regarding the quality of the produced data:

Major comments:

• Looking at the differential fitness scores in the supplementary data, none of the 57 genes that are known to show differential essentiality between S288C and Sigma1278b (Dowell et al., 2010, Science) appear to be identified as having differential fitness in the transposon screen. The authors mention that some of these genes have a severe fitness defect when deleted in the nonessential background and that some are only partially essential. Although this is certainly true for specific cases, deletion mutants of most of these 57 genes show a large difference in fitness between S288C and Sigma, and this thus doesn't sufficiently explain the complete lack of validation of 57 known positive cases. I think the authors need to further clarify why these known positive controls are not identified in their screen.
• Related to the previous point, the authors perform no secondary validation of identified context-dependent essential genes. They show that they can recapitulate known sets of essential and nonessential genes in S288c, but given my previous point, it is not clear how well their logistic model works for predicting differential gene essentiality/fitness. In my opinion, experimental validation of a subset of the identified differential fitness genes is needed to be able to be confident about the results.

Minor comments:

• The authors provide lots of data spread over many columns in the supplementary tables. However, a description of what is in each column is missing, and without it, it is not always possible to understand the data.
• I didn't understand the sentence at the bottom of page 5: "the number of insertion drops from -100 bp prior to CDS and extends to - 100 bp until the terminator region". Perhaps the authors can rephrase.

Significance

To my knowledge, this is the first paper exploring gene essentiality across a large number of genetically diverse yeast strains. This paper will be of interest to a broad range of geneticists.

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

For decades, geneticists have used loss of function (LoF) mutations to unravel the molecular bases of phenotypic variability. However, a common concern is to what extent the phenotypes observed in a strain or accession recapitulates what happens at the species level. In not few cases, anecdotal evidence show that an observed mutant phenotype is not recapitulated in another strain, presumably due to the "strain background". Recent efforts using different strains of Saccharomyces cerevisiae have addressed the problem, but the number has been limited. Here, Elodie Caudal et al. use an ingenious transposon-saturation strategy to carry out a large-scale, genome-wide screen of LoF mutations in 39 strains. Based on a competitive-pooling strategy, authors estimate the probability of 4,469 genes being essential, compared to the reference S288C laboratory strain. Background-specific effects were in general rare. Around 15% of these genes show an essential phenotype which is dependent on the strain background, most of them showing continuous variation across all backgrounds and one third being specific of only one strain. Such background specific genes are functionally related and are under relaxed purifying selection and show "intermediate" integration in genetic-interaction networks compared to essential and non-essential genes. The manuscript is very easy to follow, and the experimental/statistical procedures are transparent and in general well described.

Major comments

1. A limitation of the transposon saturation strategy is the need of galactose as the carbon source, which confounds scoring of genetic background effects. The study would highly benefit from any kind of orthogonal validation or phenotype predictions, beyond the BMH1 case presented (Fig S5). Few options would be direct testing of lethal/sick phenotypes of clean gene knockouts for discussed hits (Fig 5) in several strains and conditions including galactose, testing few of the transposon libraries under different conditions to validate the environment nature of the continuous behavior, or testing the predictive power of the method using data or strains used in Galardini 2019 (ref. 25).
2. Showing the degree of replicability of the entire procedure would also help, form transposon insertion to phenotypic comparisons. If we understood correctly, this was indeed done for isolate AKE. What is the correlation of their probability scores?
3. The use of "fitness genes" is confusing, since the main phenotypic output here scored for each gene is LoF lethality, or more specifically the probability of being lethal or non-essential. Lethality or essentiality would be a more appropriate concept throughout. A next step would indeed be to quantify the phenotypic effects in a more quantitative manner (which is generally used while referring to a gene's fitness effect).

Some minor comments

-Considering that part of the signal is coming from the specific environment tested, one would expect some degree of clustering among related strains based on their gene-essentiality probability (Fig2), given that growth phenotypes correlate well with strain origins when tested under different environments (Warringer et al., 2011). Please discuss.

-Galactose is not a non-fermentable carbon source (pg 11, pg12). It is true that flux trough the fermentative pathways is lower and that the respiratory pathways are induced in galactose, when compared to growth on glucose, but galactose is readily fermented under low oxygen conditions. Indeed, variation in the regulation of these pathways could explain the environmental effects detected.

-Examples on FigS3 were useful for a better intuition of how the actual data looks like. Perhaps some of this belongs in Fig1.

Fig2, restrict the #insertions label to the actual limits for the set of 39 strains. Currently, it seems there are strains with fewer than 100K and no strain with 300K insertions.

-pg5 paragraph 2, a line on how representative is the set of 106 isolates and again later for the final data set of 39. Which main clades are missing or perhaps overrepresented?

-pg6 paragraph 1, should be 106 or 107?

-pg14 line2, is OD of 0.5 correct or was also 0.05 as in galactose? This is relevant, since it would change the competitive selection regime under galactose or glucose (more generations under glucose in the latter case). For clarity, authors could here state an approximate number of cell divisions in each medium.

-pg14 line 2, correct wording "to enrich for cells the transposon.."

Significance

While recent previous studies have measured genetic background dependent effects of gene mutations at the genome-wide level, this is the first study addressing the problem at the broader population level. Confirming that such effects are in general rare, even at this broad level, is a significant advance in the field. It is limited in the number of environmental conditions and subsequent insights (as in Galardini 2019, ref #25) and in more mechanistic views of specific allele interactions (as in Mullis 2018, ref #5). We feel, however, that these directions would already be out of the scope of the well-framed question here addressed.

Because of the problem addressed and tackled in an ingenious and comprehensive manner, this manuscript will attract the attention of a broad audience of geneticists, genome and systems biologists. Our main expertise is in yeast genetics and functional genomics.

Referee Cross-commenting

Reviewer #1 commented the possibility that insertion density could be determined by local chromatin status instead of gene essentiality, given that transposon insertion occurs more often at nucleosome free sites (point 2). While the insertion pattern around the essential gene's vicinity is convincing, we agree that it would help to show that these phenomena are independent from one another, or that this issue must at least be discussed.

The seeming need of further experimental or analytical validation was raised by reviewers #1 and #3.

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

Thank you for the opportunity to review "Population-level survey of loss-of-function mutations revealed that background dependent fitness genes are rare and functionally related in yeast" by Caudal et al. This manuscript reports on the genetic background-dependent traits resulting from natural variation. Authors use 39 natural isolates of the budding yeast (S. cerevisiae) and apply transposon saturation mutagenesis approach to analyze fitness due to loss of function mutations. They identified background and environment dependent genes. They estimate that background specific rewiring is rare and represents instances of bridging between bioprocesses as well as connecting functional related genes.

Major comments

1. Authors filtered strains based on whole chromosome aneuploidies, but what about chromosome arm aneuploidies. Were they detected and if so how were they handled? This should be discussed.
2. How does chromatin structure variation across different genetic backgrounds affect the results of the screen? Is this a confounding variable? This should be discussed.
3. On page 7 authors discuss the involvement of other biological processes in addition to respiration and mitochondrial function. It is not clear what they are referring to. This should be clarified in the main text.
4. It would be useful to annotate the functional information discussed in the text directly on the network in Fig. 4 A and B.
5. On page 9, authors should comment on the origin of ACP and CLG strain that would result in the similarity of their fitness profile to S288C which they note as an exception.
6. On page 10 authors discuss that background-specific fitness genes can belong to protein complexes. Can authors test this formally by looking at the overlap with the protein complex standard or protein interaction standard? This would strengthen this statement.
7. Authors should discuss the reasons why transcription & chromatin remodeling and nuclearcytoplasmic transport, are anticorrelated with genes involved in mitochondrial translation in terms of their fitness profiles and the implications for the evolution of environment-dependent fitness genes.
8. Authors discuss the limitation of the Hermes system however couldn't they test this system with a different inducible promoter such as estradiol regulated promoter to remove the effect of galactose metabolism?

Minor comments

All figures should contain the appropriate colour bars and legends. For example, Figure S5B relies on the colour bar in Figure 5C but it should have its own colour bar.

Significance

Significance

This work provides a comprehensive survey of the variation in natural isolates of yeast and would be interesting to a broad audience studying the genotype-to-phenotype relationship. It is the first study that systematically assessed the fitness effect of loss of function mutations across a large panel of natural isolates providing novel insight into the background specific and environment dependent genes. This represents a valuable resource for the community to ask questions about natural variation in yeast. My expertise is in complex genetic networks in yeast and genome evolution.

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Note: This rebuttal 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

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

This is a very interesting paper with novel observations. The authors find that, in yeast, Rvb1/2 AAA+ ATPases couple transcription, mRNA granular localization, and mRNAs translatability during glucose starvation. Rvb1 and Rvb2 were found to be enriched at the promoters and mRNAs of genes involved in alternative glucose metabolism pathways that are transcriptionally upregulated but translationally downregulated during glucose starvation.

The following are some comments

Introduction

1. "Structural studies have shown that they form a dodecamer comprised of a stacked Rvb1 hexametric ring and a Rvb2 hexametric ring." o Rvb1 and Rvb2 form a heterohexameric ring with alternating arrangement (not homohexamers that stack on top of each other as suggested by this sentence) o In yeast, they oligomerize mostly as single hexametric rings, with dodecamers reported being less than 10% in frequency in vivo (eg Jeganathan et al. 2015 https://doi.org/10.1016/j.jmb.2015.01.010)

   Results Section: Rvb1/Rvb2 are identified as potential co-transcriptionally loaded protein factors on the alternative glucose metabolism genes

2. "These two proteins are generally thought to act on DNA but have been found to be core components of mammalian and yeast cytoplasmic stress granules" • These two papers extensively show Rvb1/Rvb2 localization to granules/condensates under stress/nutrient starvation conditions and should be cited. The Rvb1/2 foci were named Rbits: i. Rizzolo et al. 2017 https://doi.org/10.1016/j.celrep.2017.08.074 ii. Kakihara et al. 2014 https://doi.org/10.1186/s13059-014-0404-4
3. "a portion of them becomes localized to cytoplasmic granules that are not P-bodies in both 15-minute and 30-minute glucose starvation conditions (Figure 1-figure supplement 2)" • Supplement figure 2 only includes results under 30-min glucose starvation, no 15-min data was shown

4. Figure 1C, unclear whether p-value here is for FC of GLC3 over HSP or FC of GLC3 over CRAPome. In addition, both FC datasets should have p-values.

   Section: Rvb1/Rvb2 are enriched at the promoters of endogenous alternative glucose metabolism genes

5. "Here, we performed ChIP-seq on Rvb1, Rvb2, and the negative control Pgk1 in 10 minutes of glucose starvation (Figure 2-figure supplement 3, left)" • Unclear what figure is being referred to, panel A or panel B?

6. "Structural studies have shown that Rvb1/Rvb2 can form a dodecamer complex. Their overlapped enrichment also indicates that Rvb1 and Rvb2 may function together." • They function together regardless of forming a dodecamer or not, as they assemble as heterohexamers

   Section: Engineered Rvb1/Rvb2 tethering to mRNAs directs the cytoplasmic localization and repressed translation

7. Does binding of any protein to PP7 loop in this construct alter cytoplasmic fate? A control such as GFP-CP or any other protein attached to CP should be used.
8. No statistical analysis was done for Figure 4E quantification

9. "Results showed that after replenishing the glucose to the starved cells, the translation of those genes is quickly induced, with an ~8-fold increase in ribosome occupancy 5 minutes after glucose readdition for Class II mRNAs (Figure 4-figure supplement 9)" o Would be important to see this recovery (increase in translation after glucose replenishment) in one of the reporter constructs used in the paper, such as GL3 promoter driven CFP.

   Section: Engineered Rvb1/Rvb2 binding to mRNAs increases the transcription of corresponding genes

10. How many biological replicates is in Figure 5B? There does not seem to be any error bars/gray sections indicating sample variation. P-value was also not calculated.

Reviewer #1 (Significance (Required)):

This is a very interesting manuscript that ascribes yet another function of the highly conserved RVB1/2 AAA+ ATPases.

**Referee Cross-commenting**

All reviewers agree that this an interesting paper. However, the reviewers do suggest specific experiments to verify some of the results. Carrying out these experiments will definitely improve the paper.

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

In their manuscript entitled "Rvb1/Rvb2 proteins couple transcription and translation during glucose starvation", Chen and co-authors use genetics and microscopy to demonstrate how budding yeast regulate cytoplasmic translation by their promoter sequences by two conserved ATPases Rvb1 and Rvb2 during nutrient stress. The authors show that these two ATPases repress translation of target mRNAs and then propose that these two proteins also recruit mRNAs to P bodies. The authors show that Rvb1/2 preferentially binds in the presence of Class II promoters using CoTrIP, that Rvb1/2 binds specifically at Class II promoters using ChIP-Seq, that Rvb1/2 are bound to transcripts with Class II promoters using RIP-Seq, that tethering of Rvb1/2 to a transcript decreases its translatability and that Rvb1/2 binding to a transcript increases its transcript levels by increasing transcription and not slowing mRNA decay.

The CoTrIP experiment is clever and for the most part well executed. The key conclusions are largely convincing but some clarifications are nevertheless needed (see below). Overall, this paper is well written with well executed experiments that largely support the authors' model. No major additional experiments are needed to support the claims of the paper. There are a few minor concerns that should be addressed before this manuscript gets published. These are: Minor comments: 1) Are Rvb1/2 components (enriched in) of P bodies? The model proposed by the authors suggests this but no data is show. 2) Fig. 1A: The model proposed by the authors indicates that Rvb1/2 and other proteins are recruited to the mRNAs in a promoter-dependent manner and not mRNA sequence dependent manner. This is largely supported by the data presented in the paper. However the authors should also discuss the possibility that RNA sequences could nevertheless contribute as only a uniform ORF has been tested. Could the promoter recruit Rvb1/2 similarly regardless of the ORF sequence tested? Please provide a sequence of the uniform ORF, discuss what this "uniformity" means and how a change in RNA sequence could affect the outcome of the experiment outlined in Fig. 1A. 3) Fig. 2: The authors use Pgk 1 in their ChIP control but this is not the appropriate control for the experiment as Pgk 1 is not nuclear and thus cannot demonstrate non-specific interaction with genetic regions of tested genes. Regardless, the data is convincing enough to support the model that Rvb1/2 are specifically recruited to the promoters of Class II stress-induced genes and not Class I stress-induced genes. GFP-NLS would be a better control. The authors should discuss in their materials and methods section why they chose a cytoplasmic protein for their normalization control but preferably perform ChIP with GFP-NLS or other nuclear protein that could bind to chromatin non-specifically to further demonstrate the specificity of Rvb1/2 enrichment at Class II promoters. 4) The authors claim that Rvb1/Rvb2 binding to transcripts leads to formation of granules that are non-colocalized with P-bodies and instead co-localized to SGs, but no SG fluorescent marker is used to demonstrate this claim. The authors should provide this data or remove this claim from their manuscript. 5) Fluorescent images are fuzzy, very small and difficult to interpret. mRNA puncta are difficult to observe and it is hard to conclude which green puncta colocalize with P bodies and which do not (and how frequently). It is difficult to differentiate between the cytoplasm and nucleus. Consider adding DAPI overlay. 6) The relevance of Figure 2B is not clear - please discuss. 7) Fig 5A modeling adds little supporting evidence to the entire figure. The experimental results are more convincing. Consider moving to the Supplement. 8) Fig. 4 and 3B. The authors suggest that Rvb1/2 loaded by the promoters onto the mRNA determine accumulation of mRNAs to P bodies. To test this model, the authors tether Rvb1/2 onto the mRNA using MS2-MCP system and then look for co-localization of the mRNA with P bodies. However, if the authors' model is correct, this experiment could have been achieved already using the constructs in Fig. 3B. The authors should look at the P body localization pattern using chimeras used in Fig. 3B. 9) Fig. 6: The authors present a model where mRNAs transcribed from Class II promoters are decorated with Rvb1/2 co-transcriptionally, exported into the cytoplasm, recruited to P bodies and translationally repressed. However, this model is not fully supported by the data shown. Specifically, the authors have not shown that localization of mRNAs to P bodies induces translational repression or whether the recruitment is a consequence of this repression. The authors should revise their model to reflect this uncertainty. Also, the numbering of steps 1,2 3 is confusing. Does it imply a temporal sequences? Some of these steps could be occurring simultaneously (like 1 and 3). How does step 3 lead from step 2? Please clarify this model. 10) Consider showing data-points in Fig 1 figure supplement 1. The box/whisker plot doesn't give a good sense of the enrichment alone 11) Figure 1 Fig supplement 2 shows that the fluorophore seems to influence the % of cells with foci. Why is this the case? 12) List gene names in Fig 2 fig supp 5. 13) Throughout the paper the graph axis labels are very small and difficult to read. 14) Figure 4 fig supplement 7C and 8E: on the y-axis the legend says proportion of cells (%), so the value on the y-axis might be 25, 50, 75 and not 0.25, 0.50 and 0.75. 15) The last paragraph of the Introduction (page 2) detailed how Rvb1/Rvb2 are core components of the stress granule. Yet most experiments were conducted to relate Rvb1/Rvb2 with P-bodies. Maybe some information about the known roles Rvb1/Rvb2 play in the P-bodies in the Introduction section could help.

Reviewer #2 (Significance (Required)):

Ruvb helicase has been shown to regulate the formation of stress granules in human U2OS cells during oxidative stress (Parker lab, Cell, 2016). Thus, the authors suggest that Rvb proteins could have a broad and conserved role in the formation of RNA granules, which advances our understanding of how biomolecular condensates could form. In addition, translationally-repressed mRNAs have been shown to preferentially recruit to diverse RNA granules, from stress granules P bodies in human cells as well as germ granules in C. elegans and Drosophila. These observations have gained considerable attention in the past 5 years and exact molecular principles behind this phenomenon are not entirely clear. Long and exposed RNA sequences are thought to be sufficient for this enrichment. The authors however suggest that specific proteins (Rvb1/2) could also trigger enrichment either directly by interacting with P bodies or indirectly by repressing translation and exposing RNA sequences. This finding will be particularly relevant to the field of biomolecular condensates. My expertise is in the area of RNA biology, mRNA decay, RNA granules and mRNA localization.

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

Dr. Brian Zid has previously published in Nature that, in response to glucose starvation, promoters of some genes ("class II") can control synthesis of mRNAs that are sequestered in cytoplasmic P bodies or Stress granules, away from the translation apparatus. In this paper, his group reports about the underlying mechanism. They have found proteins that bind preferentially class II promoters as well as their transcripts and are capable of repressing their translation and stimulating their assembly with P bodies. They found a correlation between the capacity of Rvb1/2 binding to promoters and binding to mRNAs. Using a tethering technique, they found that Rb1/Rvb2 recruitment to reporter mRNA (not class II) led to the association of the transcript with PBs and its translation repression. Interestingly, Binding of Rvb1/Rvb2 to the studied transcript increased transcription of its own gene, probably by remodeling the nearby chromatin. The paper uncovers a mechanism to sequester mRNAs as translationally repressed in RNA granules during starvation and warrants a publication in a good journal, after responding to various comments below.

1. CoTrIP is a method to identify proteins that differentially bind plasmids carrying different promoters/genes. However, the claim that it identifies proteins bound to nascent mRNAs is an overreach, as the proteins bind both DNA and RNA and the purified plasmid contains both types of nucleic acids. Therefore, the title of section 1 ("Rvb1/Rvb2 are identified as potential co-transcriptionally loaded protein factors on the alternative glucose metabolism genes") should be changed to something like: Rvb1/Rvb2 are identified as proteins that are co-purified with a plasmid expressing alternative glucose metabolism genes. Description of CoTrIP and its results should be discussed throughout the manuscript accordingly.

2. The engineered Rvb1/Rvb2 tethering to mRNAs of choice is a potentially convincing way to show the causative effect of Rvb1/Rvb2 on RNA performance. Using this method, the authors show that attachment of Rvb1/Rvb2 to an engineered mRNA mediate its association with granules and inhibits its translation. However, this experiment takes Rvb1/2 out of its natural context such that its behavior in this case may not be exemplative of its endogenous function. The authors are encouraged to support their results by depleting Rvbs with AID and examine the outcome of this depletion on PBs formation and translation of class II genes (and class I as controls).

3. The tethering experiments, shown in Fig. 4, would be more convincing by including an additional control. To rule out the possibility that any bulky protein that is recruited to the 3'-UTR by the PP7 element affects translation (not an unlikely possibility), they want to consider fusing irrelevant protein (e.g., Pgk1p) to CP, in place of Rvb1/2.

1. The proposal that Rvb1 binds class II transcripts during transcription is a plausible possibility (which I personally believe to represent the reality), but by no means demonstrated. This should be clearly addressed in the paper.

2. An optional suggestion: The paper can be upgraded by performing ribosome profiling, as shown in Supplemental Fig. 9, after a short depletion of Rvb1/2 by AID (see comment 2). This, in combination with the results already shown in Supp Fig. 9, can demonstrate the role of Rvb1/2 in mRNA storage in granules and in translation shortly after glucose refeeding. The large data sets thus produced (in particular the ratio between depleted and non-depleted signal per each gene) can be used to try correlate the extent of ribosome occupancy (or the above mentioned ratio) with cis-element(s) or known trans-acting elements within the promoters. This may identify elements within the promoters that recruit (directly or indirectly) Rvb1/2. If successful, it can pave the way to demonstrate co-transcriptional RNA binding. I also suggest moving Supp Fig. 9 as an additional panel of the main Fig. 4. Minor point:

3. The original reference about "mRNA imprinting" was published by Choder in Cellular logistics 2011.
4. The graph in 5B does not have error bars and the number of replicates is unclear.

Reviewer #3 (Significance (Required)):

The paper uncovers a mechanism to sequester mRNAs as translationally repressed in RNA granules during starvation. This significantly advances our understanding of how gene expression in yeast responds to the environment and warrants a publication in a good journal, after responding to the various comments, indicated above. My expertise is regulation of gene expression.

**Referee Cross-commenting**

In general all reviewers feel that the paper deals with a significant issue, each from his/her point of view, and is basically of high quality.

I concur with all the comments of Reviewer 1 and 2. In particular, two comments drove my attention. Reviewer 1: Would be important to see increase in translation after glucose replenishment in one of the reporter constructs used in the paper, such as GL3 promoter driven CFP. Reviewer 2: The authors should look at the P body localization pattern using chimeras used in Fig. 3B.

There are comments common to more than one reviewer.

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

Dr. Brian Zid has previously published in Nature that, in response to glucose starvation, promoters of some genes ("class II") can control synthesis of mRNAs that are sequestered in cytoplasmic P bodies or Stress granules, away from the translation apparatus. In this paper, his group reports about the underlying mechanism. They have found proteins that bind preferentially class II promoters as well as their transcripts and are capable of repressing their translation and stimulating their assembly with P bodies. They found a correlation between the capacity of Rvb1/2 binding to promoters and binding to mRNAs. Using a tethering technique, they found that Rb1/Rvb2 recruitment to reporter mRNA (not class II) led to the association of the transcript with PBs and its translation repression. Interestingly, Binding of Rvb1/Rvb2 to the studied transcript increased transcription of its own gene, probably by remodeling the nearby chromatin.<br> The paper uncovers a mechanism to sequester mRNAs as translationally repressed in RNA granules during starvation and warrants a publication in a good journal, after responding to various comments below.

1. CoTrIP is a method to identify proteins that differentially bind plasmids carrying different promoters/genes. However, the claim that it identifies proteins bound to nascent mRNAs is an overreach, as the proteins bind both DNA and RNA and the purified plasmid contains both types of nucleic acids. Therefore, the title of section 1 ("Rvb1/Rvb2 are identified as potential co-transcriptionally loaded protein factors on the alternative glucose metabolism genes") should be changed to something like: Rvb1/Rvb2 are identified as proteins that are co-purified with a plasmid expressing alternative glucose metabolism genes. Description of CoTrIP and its results should be discussed throughout the manuscript accordingly.
2. The engineered Rvb1/Rvb2 tethering to mRNAs of choice is a potentially convincing way to show the causative effect of Rvb1/Rvb2 on RNA performance. Using this method, the authors show that attachment of Rvb1/Rvb2 to an engineered mRNA mediate its association with granules and inhibits its translation. However, this experiment takes Rvb1/2 out of its natural context such that its behavior in this case may not be exemplative of its endogenous function. The authors are encouraged to support their results by depleting Rvbs with AID and examine the outcome of this depletion on PBs formation and translation of class II genes (and class I as controls).
3. The tethering experiments, shown in Fig. 4, would be more convincing by including an additional control. To rule out the possibility that any bulky protein that is recruited to the 3'-UTR by the PP7 element affects translation (not an unlikely possibility), they want to consider fusing irrelevant protein (e.g., Pgk1p) to CP, in place of Rvb1/2.
4. The proposal that Rvb1 binds class II transcripts during transcription is a plausible possibility (which I personally believe to represent the reality), but by no means demonstrated. This should be clearly addressed in the paper.
5. An optional suggestion: The paper can be upgraded by performing ribosome profiling, as shown in Supplemental Fig. 9, after a short depletion of Rvb1/2 by AID (see comment 2). This, in combination with the results already shown in Supp Fig. 9, can demonstrate the role of Rvb1/2 in mRNA storage in granules and in translation shortly after glucose refeeding. The large data sets thus produced (in particular the ratio between depleted and non-depleted signal per each gene) can be used to try correlate the extent of ribosome occupancy (or the above mentioned ratio) with cis-element(s) or known trans-acting elements within the promoters. This may identify elements within the promoters that recruit (directly or indirectly) Rvb1/2. If successful, it can pave the way to demonstrate co-transcriptional RNA binding. I also suggest moving Supp Fig. 9 as an additional panel of the main Fig. 4.

Minor point:

1. The original reference about "mRNA imprinting" was published by Choder in Cellular logistics 2011.
2. The graph in 5B does not have error bars and the number of replicates is unclear.

Significance

The paper uncovers a mechanism to sequester mRNAs as translationally repressed in RNA granules during starvation. This significantly advances our understanding of how gene expression in yeast responds to the environment and warrants a publication in a good journal, after responding to the various comments, indicated above.

My expertise is regulation of gene expression.

Referee Cross-commenting

In general all reviewers feel that the paper deals with a significant issue, each from his/her point of view, and is basically of high quality.

I concur with all the comments of Reviewer 1 and 2. In particular, two comments drove my attention. Reviewer 1: Would be important to see increase in translation after glucose replenishment in one of the reporter constructs used in the paper, such as GL3 promoter driven CFP. Reviewer 2: The authors should look at the P body localization pattern using chimeras used in Fig. 3B.

There are comments common to more than one reviewer.

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 their manuscript entitled "Rvb1/Rvb2 proteins couple transcription and translation during glucose starvation", Chen and co-authors use genetics and microscopy to demonstrate how budding yeast regulate cytoplasmic translation by their promoter sequences by two conserved ATPases Rvb1 and Rvb2 during nutrient stress. The authors show that these two ATPases repress translation of target mRNAs and then propose that these two proteins also recruit mRNAs to P bodies. The authors show that Rvb1/2 preferentially binds in the presence of Class II promoters using CoTrIP, that Rvb1/2 binds specifically at Class II promoters using ChIP-Seq, that Rvb1/2 are bound to transcripts with Class II promoters using RIP-Seq, that tethering of Rvb1/2 to a transcript decreases its translatability and that Rvb1/2 binding to a transcript increases its transcript levels by increasing transcription and not slowing mRNA decay.

The CoTrIP experiment is clever and for the most part well executed. The key conclusions are largely convincing but some clarifications are nevertheless needed (see below). Overall, this paper is well written with well executed experiments that largely support the authors' model. No major additional experiments are needed to support the claims of the paper. There are a few minor concerns that should be addressed before this manuscript gets published. These are:

Minor comments:

1) Are Rvb1/2 components (enriched in) of P bodies? The model proposed by the authors suggests this but no data is show.

2) Fig. 1A: The model proposed by the authors indicates that Rvb1/2 and other proteins are recruited to the mRNAs in a promoter-dependent manner and not mRNA sequence dependent manner. This is largely supported by the data presented in the paper. However the authors should also discuss the possibility that RNA sequences could nevertheless contribute as only a uniform ORF has been tested. Could the promoter recruit Rvb1/2 similarly regardless of the ORF sequence tested? Please provide a sequence of the uniform ORF, discuss what this "uniformity" means and how a change in RNA sequence could affect the outcome of the experiment outlined in Fig. 1A.

3) Fig. 2: The authors use Pgk 1 in their ChIP control but this is not the appropriate control for the experiment as Pgk 1 is not nuclear and thus cannot demonstrate non-specific interaction with genetic regions of tested genes. Regardless, the data is convincing enough to support the model that Rvb1/2 are specifically recruited to the promoters of Class II stress-induced genes and not Class I stress-induced genes. GFP-NLS would be a better control. The authors should discuss in their materials and methods section why they chose a cytoplasmic protein for their normalization control but preferably perform ChIP with GFP-NLS or other nuclear protein that could bind to chromatin non-specifically to further demonstrate the specificity of Rvb1/2 enrichment at Class II promoters.

4) The authors claim that Rvb1/Rvb2 binding to transcripts leads to formation of granules that are non-colocalized with P-bodies and instead co-localized to SGs, but no SG fluorescent marker is used to demonstrate this claim. The authors should provide this data or remove this claim from their manuscript.

5) Fluorescent images are fuzzy, very small and difficult to interpret. mRNA puncta are difficult to observe and it is hard to conclude which green puncta colocalize with P bodies and which do not (and how frequently). It is difficult to differentiate between the cytoplasm and nucleus. Consider adding DAPI overlay.

6) The relevance of Figure 2B is not clear - please discuss.

7) Fig 5A modeling adds little supporting evidence to the entire figure. The experimental results are more convincing. Consider moving to the Supplement.

8) Fig. 4 and 3B. The authors suggest that Rvb1/2 loaded by the promoters onto the mRNA determine accumulation of mRNAs to P bodies. To test this model, the authors tether Rvb1/2 onto the mRNA using MS2-MCP system and then look for co-localization of the mRNA with P bodies. However, if the authors' model is correct, this experiment could have been achieved already using the constructs in Fig. 3B. The authors should look at the P body localization pattern using chimeras used in Fig. 3B.

9) Fig. 6: The authors present a model where mRNAs transcribed from Class II promoters are decorated with Rvb1/2 co-transcriptionally, exported into the cytoplasm, recruited to P bodies and translationally repressed. However, this model is not fully supported by the data shown. Specifically, the authors have not shown that localization of mRNAs to P bodies induces translational repression or whether the recruitment is a consequence of this repression. The authors should revise their model to reflect this uncertainty. Also, the numbering of steps 1,2 3 is confusing. Does it imply a temporal sequences? Some of these steps could be occurring simultaneously (like 1 and 3). How does step 3 lead from step 2? Please clarify this model.

10) Consider showing data-points in Fig 1 figure supplement 1. The box/whisker plot doesn't give a good sense of the enrichment alone.

11) Figure 1 Fig supplement 2 shows that the fluorophore seems to influence the % of cells with foci. Why is this the case?

12) List gene names in Fig 2 fig supp 5.

13) Throughout the paper the graph axis labels are very small and difficult to read.

14) Figure 4 fig supplement 7C and 8E: on the y-axis the legend says proportion of cells (%), so the value on the y-axis might be 25, 50, 75 and not 0.25, 0.50 and 0.75.

15) The last paragraph of the Introduction (page 2) detailed how Rvb1/Rvb2 are core components of the stress granule. Yet most experiments were conducted to relate Rvb1/Rvb2 with P-bodies. Maybe some information about the known roles Rvb1/Rvb2 play in the P-bodies in the Introduction section could help.

Significance

Ruvb helicase has been shown to regulate the formation of stress granules in human U2OS cells during oxidative stress (Parker lab, Cell, 2016). Thus, the authors suggest that Rvb proteins could have a broad and conserved role in the formation of RNA granules, which advances our understanding of how biomolecular condensates could form.

In addition, translationally-repressed mRNAs have been shown to preferentially recruit to diverse RNA granules, from stress granules P bodies in human cells as well as germ granules in C. elegans and Drosophila. These observations have gained considerable attention in the past 5 years and exact molecular principles behind this phenomenon are not entirely clear. Long and exposed RNA sequences are thought to be sufficient for this enrichment. The authors however suggest that specific proteins (Rvb1/2) could also trigger enrichment either directly by interacting with P bodies or indirectly by repressing translation and exposing RNA sequences. This finding will be particularly relevant to the field of biomolecular condensates.

My expertise is in the area of RNA biology, mRNA decay, RNA granules and mRNA localization.

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 a very interesting paper with novel observations. The authors find that, in yeast, Rvb1/2 AAA+ ATPases couple transcription, mRNA granular localization, and mRNAs translatability during glucose starvation. Rvb1 and Rvb2 were found to be enriched at the promoters and mRNAs of genes involved in alternative glucose metabolism pathways that are transcriptionally upregulated but translationally downregulated during glucose starvation.

The following are some comments

Introduction

1. "Structural studies have shown that they form a dodecamer comprised of a stacked Rvb1 hexametric ring and a Rvb2 hexametric ring." o Rvb1 and Rvb2 form a heterohexameric ring with alternating arrangement (not homohexamers that stack on top of each other as suggested by this sentence) o In yeast, they oligomerize mostly as single hexametric rings, with dodecamers reported being less than 10% in frequency in vivo (eg Jeganathan et al. 2015 https://doi.org/10.1016/j.jmb.2015.01.010)

Results Section: Rvb1/Rvb2 are identified as potential co-transcriptionally loaded protein factors on the alternative glucose metabolism genes

1. "These two proteins are generally thought to act on DNA but have been found to be core components of mammalian and yeast cytoplasmic stress granules" • These two papers extensively show Rvb1/Rvb2 localization to granules/condensates under stress/nutrient starvation conditions and should be cited. The Rvb1/2 foci were named Rbits: i. Rizzolo et al. 2017 https://doi.org/10.1016/j.celrep.2017.08.074 ii. Kakihara et al. 2014 https://doi.org/10.1186/s13059-014-0404-4
2. "a portion of them becomes localized to cytoplasmic granules that are not P-bodies in both 15-minute and 30-minute glucose starvation conditions (Figure 1-figure supplement 2)" • Supplement figure 2 only includes results under 30-min glucose starvation, no 15-min data was shown
3. Figure 1C, unclear whether p-value here is for FC of GLC3 over HSP or FC of GLC3 over CRAPome. In addition, both FC datasets should have p-values.

Section: Rvb1/Rvb2 are enriched at the promoters of endogenous alternative glucose metabolism genes

1. "Here, we performed ChIP-seq on Rvb1, Rvb2, and the negative control Pgk1 in 10 minutes of glucose starvation (Figure 2-figure supplement 3, left)" • Unclear what figure is being referred to, panel A or panel B?
2. "Structural studies have shown that Rvb1/Rvb2 can form a dodecamer complex. Their overlapped enrichment also indicates that Rvb1 and Rvb2 may function together." • They function together regardless of forming a dodecamer or not, as they assemble as heterohexamers

Section: Engineered Rvb1/Rvb2 tethering to mRNAs directs the cytoplasmic localization and repressed translation

1. Does binding of any protein to PP7 loop in this construct alter cytoplasmic fate? A control such as GFP-CP or any other protein attached to CP should be used.
2. No statistical analysis was done for Figure 4E quantification
3. "Results showed that after replenishing the glucose to the starved cells, the translation of those genes is quickly induced, with an ~8-fold increase in ribosome occupancy 5 minutes after glucose readdition for Class II mRNAs (Figure 4-figure supplement 9)" o Would be important to see this recovery (increase in translation after glucose replenishment) in one of the reporter constructs used in the paper, such as GL3 promoter driven CFP.

Section: Engineered Rvb1/Rvb2 binding to mRNAs increases the transcription of corresponding genes

1. How many biological replicates is in Figure 5B? There does not seem to be any error bars/gray sections indicating sample variation. P-value was also not calculated.

Significance

This is a very interesting manuscript that ascribes yet another function of the highly conserved RVB1/2 AAA+ ATPases.

Referee Cross-commenting

All reviewers agree that this an interesting paper. However, the reviewers do suggest specific experiments to verify some of the results. Carrying out these experiments will definitely improve the paper.

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Transparent Peer Review

---

Download the complete Review Process [PDF] including:

• reviews
• authors' reply
• editorial decisions

</br>

Note: This rebuttal 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

RC-2021-00739

“Plasma membrane damage limits replicative lifespan in yeast and induces premature senescence in human fibroblasts”

Kono et al.

Point-by-point response

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

*In this article, Kono et al worked on cellular outcomes induced by plasma membrane damage (PMD) in yeast and in human cells. Plasma membrane damage is induced by some stresses and alteration of its repair can lead to some diseases. Globally little is known about PMD. Authors observed that PMD-induced by low concentration of SDS in yeast and in human cells can limit their replicative lifespan. A genetic screen in yeast has identified the endosomal sorting complexes required for transport (ESCRT) genes as required for PMD response. In human cells, the authors observed that PMD-induced premature senescence is dependent of p53 activity but independent of DNA damage. This work sounds novel and interesting in the context of senescence on human cells. Nevertheless, they are some limits and questions that should be addressed to strongly improve this interesting work.**

*

Thank you very much for reviewing our manuscript. We are delighted to know that reviewer #1 thinks our work is novel and interesting.

**\*Major comments:****

*- can the authors describe and explain what are common and divergent betweenreplicative lifespan in yeast and human cells, for instance on telomere biology? It is particularly important as the authors jumped from replicative lifespan in yeast to replicative senescence in human cells.

Thank you for raising this point. The telomere biology in yeast and human cells share at least three central mechanisms but obviously there are limitations of using yeast as a model. We included this point in discussion (page 12, line 10-22).

- a better characterization of premature senescence induced by SDS is required to delineate this new type of senescence: for instance, SASP content characterization and EdU incorporation assays to properly demonstrate the proliferation arrest.

According to the reviewer’s suggestion, we added SASP qPCR results (Fig. 3I and J). We also performed EdU incorporation assays and included in the revised manuscript (Fig. 3F).

- the authors claimed that PMD-induced senescence is DNA damage-independent and that PMD could occur during replicative senescence. As mentioned in some references cited by the authors, replicative senescence normally occurs in response to telomere shortening and this shortening results in a DNA damage response which initiates senescence (ref 23). So authors should formulate their conclusions and discussion in the light of these well described results and tone down some of their conclusions.

We agree with the reviewer’s point that the best-studied mechanism underlying replicative senescence is telomere shortening (Blackburn, 2001; Shay and Wrightas, 2001) and telomere-dependent replicative senescence is mediated by the DNA repair pathway (d'Adda di Fagagna et al., 2003). We changed the title, abstract, and introduction (title, “Plasma membrane damage limits replicative lifespan in yeast and induces premature senescence in human fibroblasts”, abstract page 2 line 12-13, introduction page 4 line 1-2). We hope new sentences describe our findings more precisely.

In that context it will be also interesting to investigate whether PMD occurs in other types of cellular senescence (different inducers and different cell types).

Thank you very much for the suggestion. We performed the experiment. The results indicate that PMD does not occur in DNA damage (doxorubicine)-dependent premature senescence (Fig. S8A and B).

- this story will be strongly improved if the authors provide some mechanistic insights. In particular if they can link their observations in yeast to their observation in human cells. For instance, does ESCRT impact SDS-induced senescence in human cells? Can this be linked to p53 activity?

Thank you very much for the suggestion. According to the comment, we tested whether VPS4A/B overexpression extends replicative lifespan in human cells analogous to what we observed in yeast. Unfortunately, VPS4A/B overexpression from CMV promoter gradually decreased cell viability within several days. Therefore, we could not conclude their functions on lifespan extension.

*

• in the discussion section, the authors discuss calcium signaling as a possible actor of PMD-induced p53 activation, can they show some data in that direction at least by measuring cytosolic calcium levels during PMD-induced senescence.*

According to the reviewer’s suggestion, we measured cytosolic calcium levels and included them in our revised manuscript (Fig. 5A-C). Our new results indicate that the cytosolic Ca2+ is increased after SDS treatment. We also added new figures confirming the previously reported result that KCl-dependent Ca2+-influx is sufficient for senescence induction (Fig. 5D-F). To test whether Ca2+ is required for PMDS, we treated the cells with both SDS and Ca2+ chelators but the cells ruptured immediately due to the failure of membrane resealing. Therefore, although it is likely that Ca2+ is required for PMDS, we could not dissect Ca2+’s function in membrane resealing and premature senescence. We will intensively analyze this point in our next paper.

*- ESCRT is involved in nuclear envelope repair. Can the authors ruled out any effects of SDS on nuclear envelopes as nuclear envelope alterations can be involved in cellular senescence?**

*

We appreciate reviewer #1 for raising an important point. We can rule out the possibility based on the following evidence. Nuclear deformation and subsequent upregulation of DNA damage signaling is a striking feature of nuclear envelope damage as observed in premature aging diseases Laminopathies (Eriksson et al., 2003; De Sandre-Giovannoli et al, 2003; Earle et al., 2019). We found that SDS treatment did not induce nuclear envelope deformation (Fig. 1F and Fig. S2A). Moreover, ESCRT did not accumulate at the nuclear membrane after SDS treatment (Fig. S3D, green). These results suggest that the SDS-dependent cellular senescence cannot be attributed to the nuclear envelope damage. We added sentences in discussion of the revised manuscript (page 12, line 1-5).

\*Minor comments:***　

*

• images are used twice between Figure 1F and S2A, please replaced images to avoid this.*

According to the reviewer’s comment, we replaced the images.

- in Figure 3 it will be better to present cumulative population doublings which is a more classical way to present these results.

According to the reviewer’s comment, we replaced the graphs.

*- several human cell lines are used but in most of time for different experiments. It will be good to show that at least one of them display the expected results with the different assays.**

*

According to the reviewer’s comment, we added Fig. S7 to show that WI-38 cells also show PMDS. Thank you again for reviewing our manuscript despite your hectic schedule.

* Reviewer #1 (Significance (Required)):

see above.*

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

**Summary:**

Makoto Nakanishi and co-workers use SDS (and EGTA) to induce plasma membrane damage (PMD) on budding yeast cells and human fibroblast. Their results correlate SDS induced PMD with reduced the replicative lifespan of budding yeast and p53 mediated senescence in human fibroblast.

Using genetic screens in budding yeast, 48 SDS sensitive mutants were identified, including a large set of ESCRT mutants, V-ATPase mutants, and several mutants deficient in metabolic enzymes (amino acid metabolism and lipid metabolism). Three of the SDS sensitive yeast mutants showed a reduced replicative lifespan.

SDS induced PMD on human fibroblast triggered p53 induction (without concomitant DNA damage) and subsequent p53 mediated senescence. SDS induced PMD also induced phosphatidyl-serine (PS) externalization of PM projections that co-localized with the ESCRT-III subunit CHMP4a.

These results describe a potentially interesting and novel pathophysiological effect of PMD.

*

Thank you very much for serving as a reviewer. We are delighted that the reviewer #2 considers our work to be novel and interesting.

\*Major points.***

While the description of the PMD induced phenotypes in yeast and fibroblast are interesting, mechanistic insight is not provided. Perhaps the phenotypic description could be solidified by addressing the following points: *

1. Quantification of PMD using state-of-the-art FACS analysis in yeast cells and human fibroblasts e.g. using PI together with Annexin V.*

Thank you so much for the valuable suggestion. According to the comment, we performed these experiments. We could successfully quantify the DAPI penetration in normal human fibroblasts by FACS (added to the revised manuscript as Fig. S2D). In contrast, we failed to detect the increase of Annexin V (PS externalization signals) by FACS, probably due to the detection limit of the FACS machine we used (please see below). Let me remind you that the signal at the PS externalizing spots after PMD are extremely weak; the signals cannot be compared with massive PS externalization during apoptosis. Instead, we quantified the Annexin V signals of entire cells using Zeiss inverted confocal microscope (LSM780) and Zen blue software and included them in Fig. S3B. We hope these new data serve as objective evidence supporting our conclusion.

• The results from the yeast screens should be better characterized and explained. *

Thank you very much for the suggestion. According to the reviewer’s comment, we performed characterization of the screening hits and identified four novel mechanisms involved in PMD response in budding yeast (Fig. S5, S6, and Supplementary texts).

Why do the authors focus on 'replicative lifespan' rather than on e.g. 'nutrient-utilization'.

Thank you for the comment. Indeed, we are also interested in the relation of PMD and other cellular processes, including nutrient utilization. The project is on-going. In this manuscript, we would like to focus on the point that the PMD response and the replicative lifespan regulation share some key regulators.

In principle, this is fine with me, given that there are only 48 hits, but then the authors could rather argue e.g.: that they look into ESCRT mutants because the ESCRTs have been already implicated in resealing the PM in a Ca2+ dependent manner.

Thank you for the comment. In the revised manuscript, we edited the text and emphasized that ESCRT was known to be involved in membrane repair in higher eukaryotes (page 6 line 25-page 7 line 2). Here, we looked into ESCRT to test our working hypothesis that the PMD responses and the replicative lifespan regulation could share part of the fundamental mechanisms.

To drive home the point the ESCRTs (but also Vps34 and Erg2) limit the replicative life span of budding yeast due to the accumulation of PMD, this should be experimentally tested (e.g. compare replicative life span of the mutants +/- SDS to WT cells +/- SDS). Snf7, Vps34 and Erg2 mutants could affect the replicative life-span in a number of ways that is independent from PMD.

Thank you very much for raising this point. We performed the experiment. The result was that all mutants (snf7, vps34, and erg2) did not divide at all in the presence of SDS (replicative lifespan=0), consistent with the screening strategy that we isolated the mutants with absolutely no growth on SDS plates (Fig. S4). These results were added to the result section of the revised manuscript (page 7, line 11-13).

• The rational for over-expressing Vps4 is not clear to me? Vps4 is most likely not the rate limiting factor for the ESCRT machinery under these conditions.*

Thank you for asking this question. Vps4 is a AAA-ATPase promoting disassembly of the structural components (ESCRT-III filaments) and thus critical for pinching off the membrane. The most straightforward rate-limiting factor could be ATP but obviously it is nonspecific, having too many downstream consequences. Therefore, we decided to mildly overexpress VPS4 from TEF1 promoter and luckily the strategy worked well.

Perhaps it would be more telling to overexpress Vps4 in a snf7 mutant and test if it still improves the replicative life-span?

Thank you for the comment. According to the comment, we constructed pTEF1-VPS4 in a snf7 mutant and found that the strain is lethal. Thus, the lifespan extension by pTEF1-VPS4 is at least partly mediated by SNF7. In addition, the synthetic lethality suggests that pTEF1-VPS4 also does some deleterious function to a part of the ESCRT functions. That makes sense because ESCRT is involved in many cellular processes including nuclear membrane repair, lysosome repair, multivesicular body formation, cytokinesis, and exosome production.

• The finding that PMD induces p53 mediated senescence in fibroblast is an important initial finding, as is the observation of the formation of PM extrusion that contain ESCRTs and externalize PS. Unfortunately, also these experiments remain rather descriptive. Many questions remain open: a. How is p53 activated? b. Are these 'protrusion' formed by the ESCRTs? c. Are the protrusions essential for entry into senescence or a consequence?

  *

We cannot thank more for these fascinating suggestions. We are thrilled to tackle these questions. Using mRNA seq and pathway analysis, we identified upstream regulators of p53 during PMDS. We are ready to submit it as an independent manuscript because it involves large datasets.

\*Minor points:****

I understand that the author can use FIB-SEM as a very powerful technique for volumetric ultrastructural analysis. I'm wondering why it was used in Figure 5c? Would 'simple' SEM not yield exactly the same results but given the relative ease of SEM, many more cells could be quantified...? FIB-SEM would actually be great for the analysis of PMD more directly, right after SDS treatment in both yeast cells (were the entire volume of the cell could be analyzed) and in human cells.

*

Thank you very much for a valuable advice. As reviewer #2 may know very well, SEM requires dehydration of cells, and the data acquisition is performed under high-vacuum condition. These two treatments significantly alter the structure of the plasma membrane of human normal fibroblasts. In contrast, for FIB-SEM observation, the cells in a culture dish can be directly fixed and embedded in resin, which preserves fine structures of the plasma membrane including soft and tiny projections (280-2500 nm). Based on these reasons, we decided to utilize FIB-SEM in Fig. 5C (now Fig. 6C in the revised manuscript).

* Reviewer #2 (Significance (Required)):

The authors report very exciting observations that describe novel effects of plasma membrane damage (PMD) on cell (patho)physiology. Unfortunately, I find it difficult to connect the yeast part to the studies using human fibroblast (expect that SDS is used to cause PMD). While the description of the PMD induced phenotypes in yeast and fibroblast are interesting, mechanistic insight (e.g. the role of the ESCRTs in PMD and induction of p53 mediated senescence) is largely lacking at the moment. Provided that a more through phenotypic description (see major points) and perhaps some mechanistic insight can be provided, this work will be of interest to a wide audience in molecular cell biology.*

Thank you very much for the encouraging comment. We are delighted to know that reviewer #2 highly evaluates the potential impact of this work. Here, we would like to report that 1) the PMD limits replicative lifespan in two independent eukaryotic cell types, and 2) the PMD response and the replicative lifespan regulations partly share their fundamental mechanisms, especially the mechanisms underlying cell cycle checkpoint activation. This work opens up many exciting future directions and we are extensively following them up. We hope we will be able to report detailed mechanisms very soon. Thank you again for reviewing our manuscript despite your hectic schedule.

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:

Makoto Nakanishi and co-workers use SDS (and EGTA) to induce plasma membrane damage (PMD) on budding yeast cells and human fibroblast. Their results correlate SDS induced PMD with reduced the replicative lifespan of budding yeast and p53 mediated senescence in human fibroblast.

Using genetic screens in budding yeast, 48 SDS sensitive mutants were identified, including a large set of ESCRT mutants, V-ATPase mutants, and several mutants deficient in metabolic enzymes (amino acid metabolism and lipid metabolism). Three of the SDS sensitive yeast mutants showed a reduced replicative lifespan.

SDS induced PMD on human fibroblast triggered p53 induction (without concomitant DNA damage) and subsequent p53 mediated senescence. SDS induced PMD also induced phosphatidyl-serine (PS) externalization of PM projections that co-localized with the ESCRT-III subunit CHMP4a.

These results describe a potentially interesting and novel pathophysiological effect of PMD.

Major points.

While the description of the PMD induced phenotypes in yeast and fibroblast are interesting, mechanistic insight is not provided. Perhaps the phenotypic description could be solidified by addressing the following points:

1. Quantification of PMD using state-of-the-art FACS analysis in yeast cells and human fibroblasts e.g. using PI together with Annexin V.
2. The results from the yeast screens should be better characterized and explained. Why do the authors focus on 'replicative lifespan' rather than on e.g. 'nutrient-utilization'. In principle, this is fine with me, given that there are only 48 hits, but then the authors could rather argue e.g.: that they look into ESCRT mutants because the ESCRTs have been already implicated in resealing the PM in a Ca2+ dependent manner.
3. To drive home the point the ESCRTs (but also Vps34 and Erg2) limit the replicative life span of budding yeast due to the accumulation of PMD, this should be experimentally tested (e.g. compare replicative life span of the mutants +/- SDS to WT cells +/- SDS). Snf7, Vps34 and Erg2 mutants could affect the replicative life-span in a number of ways that is independent from PMD.
4. The rational for over-expressing Vps4 is not clear to me? Vps4 is most likely not the rate limiting factor for the ESCRT machinery under these conditions. Perhaps it would be more telling to overexpress Vps4 in a snf7 mutant and test if it still improves the replicative life-span?
5. The finding that PMD induces p53 mediated senescence in fibroblast is an important initial finding, as is the observation of the formation of PM extrusion that contain ESCRTs and externalize PS. Unfortunately, also these experiments remain rather descriptive. Many questions remain open: a. How is p53 activated?<br> b. Are these 'protrusion' formed by the ESCRTs? c. Are the protrusions essential for entry into senescence or a consequence?

Minor points:

I understand that the author can use FIB-SEM as a very powerful technique for volumetric ultrastructural analysis. I'm wondering why it was used in Figure 5c? Would 'simple' SEM not yield exactly the same results but given the relative ease of SEM, many more cells could be quantified...? FIB-SEM would actually be great for the analysis of PMD more directly, right after SDS treatment in both yeast cells (were the entire volume of the cell could be analyzed) and in human cells.

Significance

The authors report very exciting observations that describe novel effects of plasma membrane damage (PMD) on cell (patho)physiology. Unfortunately, I find it difficult to connect the yeast part to the studies using human fibroblast (expect that SDS is used to cause PMD). While the description of the PMD induced phenotypes in yeast and fibroblast are interesting, mechanistic insight (e.g. the role of the ESCRTs in PMD and induction of p53 mediated senescence) is largely lacking at the moment. Provided that a more through phenotypic description (see major points) and perhaps some mechanistic insight can be provided, this work will be of interest to a wide audience in molecular cell 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

In this article, Kono et al worked on cellular outcomes induced by plasma membrane damage (PMD) in yeast and in human cells. Plasma membrane damage is induced by some stresses and alteration of its repair can lead to some diseases. Globally little is known about PMD. Authors observed that PMD-induced by low concentration of SDS in yeast and in human cells can limit their replicative lifespan. A genetic screen in yeast has identified the endosomal sorting complexes required for transport (ESCRT) genes as required for PMD response. In human cells, the authors observed that PMD-induced premature senescence is dependent of p53 activity but independent of DNA damage. This work sounds novel and interesting in the context of senescence on human cells. Nevertheless, they are some limits and questions that should be addressed to strongly improve this interesting work.

Major comments:

• can the authors describe and explain what are common and divergent between replicative lifespan in yeast and human cells, for instance on telomere biology? It is particularly important as the authors jumped from replicative lifespan in yeast to replicative senescence in human cells.
• a better characterization of premature senescence induced by SDS is required to delineate this new type of senescence: for instance, SASP content characterization and EdU incorporation assays to properly demonstrate the proliferation arrest.
• the authors claimed that PMD-induced senescence is DNA damage-independent and that PMD could occur during replicative senescence. As mentioned in some references cited by the authors, replicative senescence normally occurs in response to telomere shortening and this shortening results in a DNA damage response which initiates senescence (ref 23). So authors should formulate their conclusions and discussion in the light of these well described results and tone down some of their conclusions. In that context it will be also interesting to investigate whether PMD occurs in other types of cellular senescence (different inducers and different cell types).
• this story will be strongly improved if the authors provide some mechanistic insights. In particular if they can link their observations in yeast to their observation in human cells. For instance, does ESCRT impact SDS-induced senescence in human cells? Can this be linked to p53 activity?
• in the discussion section, the authors discuss calcium signaling as a possible actor of PMD-induced p53 activation, can they show some data in that direction at least by measuring cytosolic calcium levels during PMD-induced senescence.
• ESCRT is involved in nuclear envelope repair. Can the authors ruled out any effects of SDS on nuclear envelopes as nuclear envelope alterations can be involved in cellular senescence?

Minor comments:

• images are used twice between Figure 1F and S2A, please replaced images to avoid this.
• in Figure 3 it will be better to present cumulative population doublings which is a more classical way to present these results.
• several human cell lines are used but in most of time for different experiments. It will be good to show that at least one of them display the expected results with the different assays.

Significance

see above.

Note: This rebuttal 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

We thank the reviewers for their helpful comments. We believe that we will be able to address all of their concerns and suggestions. We have highlighted our responses in the revision plan and the changes we have already made to the manuscript in blue text. For figures where we have added data or analyses at the request of reviewers, we have highlighted the corresponding text in the figure legends.

2. Description of the planned revisions

Reviewer #1

2- In Figure 4, the two mutations appear to have statistically differential effects on Rab5 and Rab7 puncta even though the data mean and distribution seem very similar. Interestingly, in each case the non-significant effect is associated with a smaller sample size. Given that the overall sample sizes used are rather small for such highly variable data, this could easily cause a statistical anomaly due to sampling bias. The sample size should be made uniform across all genotypes and should ideally be at least doubled.

We will repeat this staining to increase the n to at least double this number, and adjust our conclusions if need be, in the revised manuscript..

3- Perhaps the most important issue related to Figure 6 where the authors find that there is no sterol accumulation at 96h APF in the Vps50 mutant. However, even that the dendritic phenotype is slower to appear in this mutant compared to the Vps54, are the authors sure that the accumulation is not just slower? This should be examined using the same temporal sequence used for Vps54 shown in Future 6 C. In addition, the fact that sterol accumulation returns to normal in the Vps54 mutant at 1 day, supports the notion of a delay phenotype (see point 1 above). These issues should be experimentally addressed to see if the data fully support the initial conclusions, or if the conclusions should be modified to suggest differential contribution of the two complexes to the process being studied and to a developmental delay phenotype.

We had included the filipin staining for Vps50KO/KO at 1 day in Figure S4 A (which did not show a significant difference from control). We did not collect data for this genotype at 72hrs APF because the dendritic length phenotype didn’t appear until later, and so we did not include Vps50KO/KO in the full time-course in Fig 6 C. We will collect additional data so that we can include Vps50KO/KO at all timepoints in this figure in the revised manuscript.

. Reviewer #2

It is stated that loss of VPS50 and VPS54 only causes dendrite morphogenesis defects. However, the corresponding supplemental figure S2c (which is not referenced in the text), is not suited to address this question. Axonal arborization, in particular terminal arbors, are not visible in samples where multiple/all c4da axons are labeled simultaneously (Fig. S2c). Analogous to the dendrite analysis of c4da neurons single cell resolution is essential to examine this in a meaningful way. Likely, however, c4da neurons may not be a good choice to address this question.

We should be able to get single cell resolution of the c4da axon terminals using MARCM. We already have two of the knockout lines recombined with FRTs (Vps53 and Vps54) for this analysis and we will make the third recombinant line so that we can use MARCM for all three lines to examine single-cell axon morphology, as suggested.

Overall, I am concerned whether the data shown here can be generalized. The cd4a neurons are rather extreme cell types due to their very large dendritic compartment. It seems quite possible that many other neurons may not have a comparable sensitivity to the supply of lipids/sterols. This type of question can only be addressed if other types of neurons/dendrites are examined. Are class 2 or class 3 da neurons showing any defects in VPS mutants?

Given that we see the phenotype emerge during the pupal stage, we want to analyze neurons that persist from the larval to adult stages. However, not all of the dendritic arborization neurons survive into adulthood- class I and II persist, while class III die during metamorphosis (Shimono et al., 2009). As we do not have adequate tools to for studying the class II neurons, we will examine dendrite morphology of the class I neurons in larvae and adults in our knockout lines. We would be happy to look at class III neurons at the reviewers request, but our analysis will necessarily be limited to the larval stage.

Reviewer #3

• Some of the experiments include multiple genotypes and so it would be important to show all in all figures. For example, figure 4B,D show four groups but figure 4F, presumably from the same set of animals, shows only three. Addition of the rescue genotype to 4F is particularly important here so should be shown. The same concern is valid for figure 5, where puncta number and area must be available.

The data from Fig. 4 F (using a genetically encoded marker for lysosomes, UAS-spin-RFP) are not from the same samples as Fig. 4 B and D (staining). We did not include the rescue for Fig 4. F because the lysosome marker, the rescue transgenes and the neuronal membrane marker are all on the third chromosome. We will build additional fly stocks so that we can include the rescue in experiments looking at lysosome morphology.

• This concern is amplified by the images in figure 6 of the filipin staining, that are more obviously perinuclear. However, the two sets of images in 6A and 6D, where co-staining with Golgin245 is shown, look very different. Improved images are required and it may be helpful to use supplementary information to show additional examples of the staining.

The images in Fig. 6 A are maximum projections of z stacks while Fig. 6 D shows single confocal planes, making it easier to see the perinuclear Golgi ring. Because other reviewers wanted some additional experiments related to Fig. 6 that we plan to incorporate into this figure in the revised manuscript, we will address this comment in a future revision and include additional images in the supplement.

• For the lipid regulation experiments in figure 7, please use an orthogonal approach to show that the Osbp and fwd RNAi had the expected effects on lipid accumulation.

In addition to sterol, Osbp and fwd both affect levels of PI4P at the Golgi. We have obtained a transgenic PI4P sensor that we can use to show the effect of these manipulations on this lipid as well.

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

Reviewer #1 While the data presented clearly support a role for GARP in regulating sterol levels to support dendritic growth, they do not inter current for suffice to exclude a role for EARP as important analyses to allow such a clear cut conclusion are either insufficient or missing. If the authors wish to maintain this claim - as suggested by the title of the manuscript - further analyses are essential.

We don’t mean to argue the EARP complex doesn’t contribute to dendrite development at all – we do show it contributes to development in Fig 3, and as we discuss in the text.. We want to argue that the GARP and EARP complexes contribute to dendrite development by distinct mechanisms. Losing the GARP complex inhibits dendrite development by means of sterol accumulation at the TGN, which is what we are trying to highlight with our title. The reduced dendrite growth that we observe in EARP deficient neurons must occur by some other as yet unknown means. We apologize for the confusion and have reworded the title to read “Sterol accumulates at the trans-Golgi in GARP complex deficient neurons during dendrite remodeling.”

1- Figure 3E shows that whereas both Vps50 and Vps54 mutations reduce dendritic complexity, the Vps54 phenotype appears earlier (96h APF). Furthermore, at 7 days dendrites appear to grow again but at a slower rate than controls. This begs the question of whether these mutations are causing a delay rather than a block in the regrowth after pruning and whether the growth will eventually be normal a few days later or whether it will stop at some point.

We have included data for an additional adult timepoint (21 days) in the new Fig. 3 E. We also included graphs in which we show the statistics for each genotype over time (new Fig. S2 D-F), and discuss this analysis in the text (lines 186-195). We have also included a table of the p-values for each comparison in the Supplemental Materials (Table S2). From this analysis, we conclude that there is not a developmental delay in the knockouts, but rather a decrease in growth during the 72-96hrs APF and 1-7 day windows when the control neurons grow. We are unable to draw conclusions about the rate of growth as we analyzed neurons from different samples at each developmental timepoint, and not the same neurons over time.

Reviewer #2

It would be important to know, whether the dendrite morphogenesis defect is indeed a developmental patterning defect or rather a "scaling" defect due to the fact that da neurons increase their size (but not necessarily their projection pattern) during larval maturation.

We have analyzed the larval data for coverage index – neuron area/hemisegment (receptive field) area as defined in (Parrish et al., 2009) to determine if there is a scaling defect at this stage in development. We do not observe a defect in scaling (Fig. S2 C) and discussed in lines 175-182.

Reviewer #3

• The statistical analyses generally look appropriate but it would be critical to clarify what N means in every case. For example in figure 2 the authors state n=8 without clarifying if this is n=8 animals or n=8 neurons. N should always be the number of animals, but then the n of independent cells counted should also be indicated. Typically, one would either pre-average per genotype or use a mixed model that includes N of animals and n of cells (or similar).

For experiments analyzing dendrite morphology, n represents the number of neurons, as we have clarified in our figure legends. As per another reviewer’s request, we will increase the n for the organelle and filipin staining in our planned revision and specify fly and cell number at that time.

• Please add details of how experiments were blinded to genotype

The researcher was blinded to genotype during analysis. We have included that detail in our Methods section (line 566).

• Some of the experiments include multiple genotypes and so it would be important to show all in all figures. For example, figure 4B,D show four groups but figure 4F, presumably from the same set of animals, shows only three. Addition of the rescue genotype to 4F is particularly important here so should be shown. The same concern is valid for figure 5, where puncta number and area must be available.

We address the first portion of this comment in section 2, for additional experiments involving generating new fly lines. We have included data on puncta area, and mean fluorescence intensity for Rab5 and Rab7 in the supplement (Fig S3). We had already included the data on puncta number and area in Fig 5, but we have added the data on mean fluorescence intensity as well.

• Related to figure 5, please provide validation of the staining of the TGN. Typically, one would expect trans Golgi to be close to the nucleus with at least some extended stacks. A Golgin245 knockout would be ideal.

The Golgi in most Drosophila cells is typically found as discrete puncta dispersed throughout the cytosol like what we see in the Golgin245 staining, as opposed to the ribbon “stack of pancake” morphology typically seen near the nucleus in mammalian cells. For reference, please see Figure 6D in (Ye et al., 2007), Figures 2,4,5 in (Rosa-Ferreira et al., 2015), and observations reviewed in (Kondylis and Rabouille, 2009).

The Golgin245 antibody was well characterized in the paper first describing it (Riedel et al., 2016) (colocalization with other Golgi markers, decreased staining with Golgin245 RNAi), but we would be happy to repeat this validation in the c4da neurons at the reviewer’s request. There do not appear to be Golgin245 mutant or KO lines available, so we would also use the Golgin245 RNAi.

• For figures 6F, G please show examples of staining for late endosomes and lysosome with appropriate validation.

Because several of our planned revisions relate to Fig. 6, we will include images for Fig. 6 F and G when we remake this figure to incorporate those planned revisions. To clarify, we used the same reagents to mark late endosomes and lysosomes in both Fig. 4 and Fig. 6. Like the Golgin245 antibody, the Rab7 antibody was developed by the Munro lab and characterized in (Riedel et al., 2016) (partial colocalization with the endosomal marker Hrs and with the lysosomal marker Arl8). Spinster (aka benchwarmer) is a known lysosomal transmembrane protein that colocalizes with Lamp1 (Dermaut et al., 2005; Rong et al., 2011). The fluorescently tagged spin transgenes were developed by the Bellen lab and have been frequently used to mark lysosomes. We would be happy to carry out additional validation experiments at the reviewer’s specification.

• The title of figure 2 is inaccurate, at least if I understand the experiment, as it does not show neuron-specific knockout but instead whole body knockout with neuron rescue. Please rephrase.

Because of the lethality of whole body Vps53KO/KO in adult flies, we analyze MARCM clonal neurons that are Vps53KO/KO in flies that are otherwise heterozygous (Vps53KO/+). To clarify this experiment, we have changed the title of Fig. 2 from “Neuron-specific knockout of Vps53 results in smaller dendritic arbors” to “Vps53KO/KO MARCM clonal neurons have smaller dendritic arbors”.

• Figure 8 needs examples of the TGN and late endosome morphology.

We have included these images in Figure

The order appears different in Fig. 4 B & D because we only included the rescue for the KO that shows a phenotype for each staining. The genotypes included in Fig. 4 B are: +/+, Vps50KO/KO, Vps50KO/KO + rescue, and Vps54KO/KO. The genotypes included in Fig. 4 D are +/+, Vps50KO/KO, Vps54KO/KO, Vps54KO/KO + rescue. We have changed the shading of the bars corresponding to these rescue genotypes throughout the manuscript to make it easier to distinguish the two rescue conditions.

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

References Cited

Dermaut, B., K.K. Norga, A. Kania, P. Verstreken, H. Pan, Y. Zhou, P. Callaerts, and H.J. Bellen. 2005. Aberrant lysosomal carbohydrate storage accompanies endocytic defects and neurodegeneration in Drosophila benchwarmer. Journal of Cell Biology. 170:127–139. doi:10.1083/jcb.200412001.

Kondylis, V., and C. Rabouille. 2009. The Golgi apparatus: Lessons from Drosophila. FEBS Letters. 583:3827–3838. doi:10.1016/j.febslet.2009.09.048.

Parrish, J.Z., P. Xu, C.C. Kim, L.Y. Jan, and Y.N. Jan. 2009. The microRNA bantam Functions in Epithelial Cells to Regulate Scaling Growth of Dendrite Arbors in Drosophila Sensory Neurons. Neuron. 63:788–802. doi:10.1016/j.neuron.2009.08.006.

Riedel, F., A.K. Gillingham, C. Rosa-Ferreira, A. Galindo, and S. Munro. 2016. An antibody toolkit for the study of membrane traffic in Drosophila melanogaster. Biology Open. 5:987–992. doi:10.1242/bio.018937.

Rong, Y., C.K. McPhee, S. Deng, L. Huang, L. Chen, M. Liu, K. Tracy, E.H. Baehrecke, L. Yu, and M.J. Lenardo. 2011. Spinster is required for autophagic lysosome reformation and mTOR reactivation following starvation. Proceedings of the National Academy of Sciences. 108:7826–7831. doi:10.1073/pnas.1013800108.

Rosa-Ferreira, C., C. Christis, I.L. Torres, and S. Munro. 2015. The small G protein Arl5 contributes to endosome-to-Golgi traffic by aiding the recruitment of the GARP complex to the Golgi. Biology Open. 4:474–481. doi:10.1242/bio.201410975.

Shimono, K., A. Fujimoto, T. Tsuyama, M. Yamamoto-Kochi, M. Sato, Y. Hattori, K. Sugimura, T. Usui, K. Kimura, and T. Uemura. 2009. Multidendritic sensory neurons in the adult Drosophila abdomen: origins, dendritic morphology, and segment- and age-dependent programmed cell death. Neural Dev. 4:37. doi:10.1186/1749-8104-4-37.

Ye, B., Y. Zhang, W. Song, S.H. Younger, L.Y. Jan, and Y.N. Jan. 2007. Growing Dendrites and Axons Differ in Their Reliance on the Secretory Pathway. Cell. 130:717–729. doi:10.1016/j.cell.2007.06.032.

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

O'Brien et al report how deficiency in GARP specific protein VPS54 or the EARP specific protein VPS50 affects the developmental dendritic remodeling of multidendritic class IV da (c4da) neurons in Drosophila. The main findings are that while both complexes play a role in dendritic remodeling, VPS54 deficiency leads to lipid accumulation in the trans-Golgi network (TGN). Manipulating sterols at the TGN affects dendritic remodeling suggesting that lipid accumulation is responsible for control of neuron morphology in this model. Overall, the data is interesting and the authors develop the experiments enough to be convincing on their major claims. However, a few aspects need clarification and perhaps revisiting conclusions.

Major comments

• The statistical analyses generally look appropriate but it would be critical to clarify what N means in every case. For example in figure 2 the authors state n=8 without clarifying if this is n=8 animals or n=8 neurons. N should always be the number of animals, but then the n of independent cells counted should also be indicated. Typically, one would either pre-average per genotype or use a mixed model that includes N of animals and n of cells (or similar).
• Please add details of how experiments were blinded to genotype
• Some of the experiments include multiple genotypes and so it would be important to show all in all figures. For example, figure 4B,D show four groups but figure 4F, presumably from the same set of animals, shows only three. Addition of the rescue genotype to 4F is particularly important here so should be shown. The same concern is valid for figure 5, where puncta number and area must be available.
• Related to figure 5, please provide validation of the staining of the TGN. Typically, one would expect trans Golgi to be close to the nucleus with at least some extended stacks. A Golgin245 knockout would be ideal.
• This concern is amplified by the images in figure 6 of the filipin staining, that are more obviously perinuclear. However, the two sets of images in 6A and 6D, where co-staining with Golgin245 is shown, look very different. Improved images are required and it may be helpful to use supplementary information to show additional examples of the staining.
• For figures 6F, G please show examples of staining for late endosomes and lysosome with appropriate validation.
• For the lipid regulation experiments in figure 7, please use an orthogonal approach to show that the Osbp and fwd RNAi had the expected effects on lipid accumulation.
• Figure 8 needs examples of the TGN and late endosome morphology.

Minor comments

• The title of figure 2 is inaccurate, at least if I understand the experiment, as it does not show neuron-specific knockout but instead whole body knockout with neuron rescue. Please rephrase.
• For ease of reading, it would be helpful to show genotypes in the same order in all figures (see 4B, 4D)

Significance

The advance here is to nominate lipid accumulation at the trans Golgi network (TGN) is sufficient to affect dendritic remodeling during development. Although the work is performed in a model system, it may have relevance to human neurodevelopmental disorders caused by mutations in the orthologous genes. The work will be of highest relevance to developmental neurobiologists, particularly those working on GARP or EARP mutations and those who use Drosophila as an appropriate model for neurodevelopment.

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 presents a solid genetic analysis of components of the GARP and EARP complexes. The analysis is focused on a specialized type of sensory neurons i.e. class IV da neurons in Drosophila larvae. The authors show that loss of multiple components (VPS50-54) disrupt dendrite morphogenesis in c4da neurons in distinct ways. Additional genetic interaction studies further support the notion of functional differences of GARP and EARP in vivo.

Overall this is a solid study and with one exception (see below) I have little concern regarding the presented experiments. I do, however, find the exclusive focus on a highly specialized cell type c4da somewhat problematic.

Concerns:

Experimental concern: It is stated that loss of VPS50 and VPS54 only causes dendrite morphogenesis defects. However, the corresponding supplemental figure S2c ( which is not referenced in the text), is not suited to address this question. Axonal arborization, in particular terminal arbors, are not visible in samples where multiple/all c4da axons are labeled simultaneously (Fig. S2c). Analogous to the dendrite analysis of c4da neurons single cell resolution is essential to examine this in a meaningful way. Likely, however, c4da neurons may not be a good choice to address this question.

It would be important to know, whether the dendrite morphogenesis defect is indeed a developmental patterning defect or rather a "scaling" defect due to the fact that da neurons increase their size (but not necessarily their projection pattern) during larval maturation.

Overall, I am concerned whether the data shown here can be generalized. The cd4a neurons are rather extreme cell types due to their very large dendritic compartment. It seems quite possible that many other neurons may not have a comparable sensitivity to the supply of lipids/sterols. This type of question can only be addressed if other types of neurons/dendrites are examined. Are class 2 or class 3 da neurons showing any defects in VPS mutants?

Significance

At this point i am not convinced that the findings can be generalized. The c4da neuron is really an extreme cell type with a massive disproportionate increase in membrane extensions. This is rather unusual and other neuron types should be tested.

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:

O'Brien and colleagues use Drosophila dendrite development to dissect the roles of the GARP and EARP vesicular trafficking complexes in the development of neuronal morphology. By making complex-specific KOs they investigate the role of each complex in the growth, pruning and re-growth of sensory dendrites and conclude that the GARP, but not EARP, complex is required for proper dendrite development by limiting sterol accumulation in the neuronal TGN.

Major comments:

While the data presented clearly support a role for GARP in regulating sterol levels to support dendritic growth, they do not inter current for suffice to exclude a role for EARP as important analyses to allow such a clear cut conclusion are either insufficient or missing. If the authors wish to maintain this claim - as suggested by the title of the manuscript - further analyses are essential.

1- Figure 3E shows that whereas both Vps50 and Vps54 mutations reduce dendritic complexity, the Vps54 phenotype appears earlier (96h APF). Furthermore, at 7 days dendrites appear to grow again but at a slower rate than controls. This begs the question of whether these mutations are causing a delay rather than a block in the regrowth after pruning and whether the growth will eventually be normal a few days later or whether it will stop at some point.

2- In Figure 4, the two mutations appear to have statistically differential effects on Rab5 and Rab7 puncta even though the data mean and distribution seem very similar. Interestingly, in each case the non-significant effect is associated with a smaller sample size. Given that the overall sample sizes used are rather small for such highly variable data, this could easily cause a statistical anomaly due to sampling bias. The sample size should be made uniform across all genotypes and should ideally be at least doubled.

3- Perhaps the most important issue related to Figure 6 where the authors find that there is no sterol accumulation at 96h APF in the Vps50 mutant. However, even that the dendritic phenotype is slower to appear in this mutant compared to the Vps54, are the authors sure that the accumulation is not just slower? This should be examined using the same temporal sequence used for Vps54 shown in Future 6C. In addition, the fact that sterol accumulation returns to normal in the Vps54 mutant at 1 day, supports the notion of a delay phenotype (see point 1 above).

These issues should be experimentally addressed to see if the data fully support the initial conclusions, or if the conclusions should be modified to suggest differential contribution of the two complexes to the process being studied and to a developmental delay phenotype.

Significance

The study advances our understanding of the role of regulation of lipid storage in sculpting neuronal morphology during 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>

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 rebuttal 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

Response to reviewers

Reviewer #1

I believe that this is a very sound and authoritative study. The analysis of all data seems appropriate and robust, and many connections between the data (and subsets of data) and their possible interpretations have been considered. In fact, in the massive Results section, some interpretations are supported by cited references (this is not meant as a critique). However, I wonder about the length of the Results section, and the balance between it and the relatively short Discussion section. It is difficult for me to nail down any part of Results that might be shortened, as I could not find clear redundancies. I also think that the level of speculation is absolutely warranted, and I did not find excessive claims being made to this or that end. Rather, I suggest to broaden the perspective somewhat (in their Discussion; see below under Significance), which might allow people with a less mechanistic perspective to grasp the potential relevance of this work for non-model plant systems studied mostly by evolutionary geneticists.

Response: We thank the reviewer for their kind remarks. We have spent a very large amount of time trying to streamline the results section and we are not sure if it would be possible to shorten it any further without removing critical details.

We appreciate the reviewer’s comment to add more detail to the discussion to make it more appealing to evolutionary geneticists and we have added the following lines to the discussion section: “The WISO or “weak inbreeder/strong outcrosser” model (Brandvain & Haig, 2005) emerges from the dynamics of parental conflict and parent-of-origin effects. Under this model, a parent from populations with higher levels of outcrossing is exposed to higher levels of conflict and can thus dominate the programming of maternal resource allocation in a cross with an individual from a population with lower levels of outcrossing. Such a phenomenon has been observed in numerous clades including Dalechampia, Arabidopsis, Capsella and Leavenworthia (Brandvain & Haig, 2018; İltaş et al., 2021; Lafon-Placette et al., 2018; Raunsgard et al., 2018). Intriguingly, loss of function phenotypes in the RdDM pathway are more severe in recently outcrossing species than in A.thaliana (Grover et al., 2018; Wang et al., 2020) and suggests that RNA Pol IV functions are more elaborate and important in these species. This raises the possibility that the role for RNA Pol IV and RdDM in parental conflict that we describe in A.thaliana here is likely heightened in and mediates the elevated level of parental conflict in species that are currently or have been recently outcrossing.”

One aspect that might warrant more scrutiny is the mapping of sRNA reads to the reference genome. I found the short section of this (M&M section, page 20, lines 23-25) to be too brief. It is not clear to me which of ShortStack's v3 weighting scheme the authors used, which is relevant for multi-mapping reads (see NR Johnson et al. 2016, G3). In addition, it is not mentioned whether zero mismatches were allowed. Perhaps this is described in more detail in Erdmann et al. (2017), but even if so, it deserves to be clarified here.

Response: Small RNA reads were aligned after allowing two mismatches. This was indicated in the bowtie command (‘bowtie -v 2’ where v 2 indicates two mis-matches). We have added text to expand on the meaning of the commands.

We have also expanded the commands used for ShortStack. We used the “Placement guided by uniquely mapping reads (-u)” option to divide the multi-mapping reads.

The manuscript is well-written and concise, despite the length of the Results section. The verbal clarity and absence of typos or grammatical issues is superb. I did find some of the Figures to be somewhat "un-intuitive", in the sense that it takes acute concentration for an outsider (of sorts) to gather and interpret the underlying data. This is probably due to the many cross-comparisons of differences between two genotypes on one axis and those of a different pair of genotypes on the other axis. I am not sure how this issue can be ameliorated (nor whether this is really necessary); however, from a technical point of view, all Figures and Suppl. Figures are really well-done.

Response: We thank the reviewer for their kind remarks. We have strived to make the figures easier to understand but we are aware that the figures do require a lot of concentration. We haven’t found an easy way to fix this. We thank the reviewer for patiently going through the figures.

The list of references seems adequate in terms of citing relevant (both older and very recent) publications. However, almost all cited papers concern Arabidopsis or other model species; I suggest to consider adding a few relevant studies on non-Brassicaceae (whether considered model taxa or not), in conjunction with my suggestion (in Significance) to potentially broaden the scope by searching for natural phenomena that also involve parent-of-origin effects on endosperm/seed development. Curiously, many of the references are "incomplete" in the sense of stopping with the journal's name, then stating the doi, i.e. they lack volume numbers and page/article numbers. This should be harmonized throughout.

Response: We have added references to non-Brassicaceae species and have also fixed the references.

Reviewer #2: This manuscript provides evidence that a loss of either the maternal or paternal copy of NRPD1 have different, and sometime opposite, effects on the accumulation of small RNAs and on expression of a subset of genes, with a loss of the maternal copy having more substantial effects. The manuscript is well written, and the conclusions, as far as they go, are justified by the data, which are effectively presented. The overall effect is subtle but informative and according to the authors support a parental conflict model for imprinting. The experiments failed to find a smoking gun in the form of a mechanism to explain how or why the maternal and paternal alleles have different effects and the explanation for a lack of clear phenotypic differences was reasonable, but untested. I would have like to see it tested by looking in a plant species that is outcrossing and highly polymorphic. However, I do appreciate that the observation that the male and female alleles can have distinct effect when mutant is an important clue. My specific comments below may reflect confusion on my part, rather than real issues. If that is that I hope that confusion can aid in clarifying what are in places quite subtle points.

Response: We thank the reviewer for their comments. We agree that it would be potentially informative to do similar experiments in an outcrossing species but that this is beyond the scope of this manuscript. Additionally, loss of NRPD1 or other components of the RdDM pathway has dramatic effects on gametogenesis in some examined outcrossing species(Grover et al., 2018; Wang et al., 2020), which could prevent the detection of subtle parent-of-origin effects on seed development.

Page 6, last paragraph: "Because the endosperm is triploid, in these comparisons there are 3 (wild-type), 2 (pat nrpd1+/-), 1 (mat nrpd1+/-) and 0 (nrpd1-/-) functional NRPD1 alleles in the endosperm. However, NRPD1 is a paternally expressed imprinted gene in wild-type Ler x Col endosperm and the single paternal allele contributes 62% of the NRPD1 transcript whereas 38% comes from the two maternal alleles (Pignatta et al., 2014). Consistent with paternal allele bias in NPRD1 expression, mRNA-Seq data shows that NRPD1 is expressed at 42% of wild- type levels in pat nrpd1+/- and at 91% of wild-type levels in mat nrpd1+/- (Supplementary Table 6)". I would think this would really complicate the analysis. Should all of the dosage values include NPRD1 imprinting values? That is to say, expressed in terms of expression values? This is also a bit confusing. The maternal copies together express 38% of the transcript, so why isn't the mat nrpd1 at 68%, rather than 91%? In any event, given this imprinting and differences in dosage of the male and female it appears that two variables, parental origin and expression levels are being compared. Since 91% is awfully close to 100%, are the mat pat comparisons really just comparing low with nearly normal expression of NRPD1? And actually, given that, the outsized effect of the mat nrpd1 +/- is even more striking.

Response: We included the details of dosage rather than imprinting values because the potential for buffering of expression upon loss of one allele could not be discounted. Indeed, we do find that the endosperm transcriptome buffers against the loss of the maternal or paternal alleles (Supplementary Table 6). The reviewer is correct in pointing out that the outsized effect of mat nrpd1+/- on gene expression is even more striking, and strongly supports our view that these effects are parental rather than endospermic.

To reduce confusion in this section, we removed the details about 38% maternal allele transcripts obtained from our previous study, and instead report only the observed values from this study (which are also consistent with the previously reported paternally-biased expression of NRPD1 in endosperm).

Page 4, Line 16. I'm afraid it's still a bit difficult to understand what was being compared what in this section. Please clarify.

Response: The authors in this previously published study compared sRNAs obtained from wild-type whole seeds (which consists of three different tissues, including endosperm) with mutant endosperm. We are pointing out that the difference in tissue composition makes the effect of nrpd1 mutation hard to disentangle from the tissue differences between the two genotypes.

Page 5, Line 5. I'm sure this is fine, but it's not entirely clear what is from the previously published paper and what is reanalysis here. All the crosses and measurements were made then, but not organized in this way?

Response: This data was indeed previously published. In that analysis, we had pooled results from different crosses and calculated significance between genotypes using chi-square tests. During a later study (Satyaki and Gehring, 2019), we realized that we were losing information by ignoring the seed abortion values per cross. So, a reanalysis of that data on a cross by cross basis allowed us to find strong evidence for maternal and paternal effects.

Page 6, Line 26. This is an excellent dosage series, but it's complicated by imprinting. So it's not 3, 2, 1, 0 effective copies. If we set the paternal copy at ~1 and each maternal at ~0.1, then it's 1.2 (wild type), 0.20 (pat nrpd1+/-), 1 (mat nrpd1+/-), and 0 (nrpd1-/-).

Response: At the genomic DNA level, its 3, 2,1 and 0 doses. The reviewer’s comment on the transcriptional dose is not clear to us. Based on measured gene expression levels, relative wild-type NRPD1 transcriptional dose =1, pat nrpd1+/- is 0.42, and mat nrpd1+/- is 0.91.

Page 6, line 31. Is the main thing we are comparing the difference between expression at 42% verses 91% of wild type?

Response: We are using the small RNA-seq data alongside the mRNA-seq data to argue that loss of mat and pat nrpd1+/- have no impact on overall Pol IV activity in endosperm (as measured by small RNA production). A nrpd1 heterozygous endosperm has almost the same small RNA profile as a wild-type endosperm. Thus any effects seen in the endosperm, including the effects on mRNA expression described later in the manuscript, are likely parental rather than zygotic endospermic effects.

Page 7, line 11. So, the overall effect in either direction on smRNA gene targets was really quite small, and I'm guessing the effect on gene expression was even smaller.

Response: The effects of loss of maternal or paternal Pol IV on sRNAs was indeed small (Fig. 1/Fig. S3). Effect of loss of maternal Pol IV on gene expression was substantially large and distinct from the relatively small impacts observed upon loss of paternal Pol IV (Fig. 3) This observation supports the view that Pol IV mediates parent-of-origin effects on gene expression.

Page 7, line 17. I take it that it is this difference, rather than the overall numbers that is of interest.

Response: Correct. The lack of a relationship between sRNAs impacted upon loss of mat and pat nrpd1 is additionally suggestive of parent-of-origin effects

Page 9, line 2. Really interesting, since one might expect that these are methylated loci that would be expected to be fed into any existing embryo maintenance methylation pathway. Surprising that they are maintained independently.

Response: It is indeed surprising that Pol IV activity in parents can have different impacts on sRNAs in the endosperm. It should be noted though, that as described in Erdmann et al 2017 and in this paper later on, many endosperm sRNA loci are in fact not associated with endosperm DNA methylation. In addition, sRNA loci that are dependent on paternal Pol IV activity are more likely to be associated with DNA methylation than are sRNA loci associated with maternal Pol IV activity. These points have been described in Figure S8.

Page 9, line 22. Proportion of total imprinted genes? Did the mutant obviate/enhance the imprinting?

Response: We have modified the manuscript to describe effects on imprinted genes: “ The expression of imprinted genes is known to be regulated epigenetically in endosperm. In mat nrpd1+/- imprinted genes were more likely to be mis-regulated than expected by chance (hypergeometric test p-15) – 15 out of 43 paternally expressed and 45 out of 128 maternally expressed imprinted genes were mis-regulated in mat nrpd1+/- while two maternally expressed imprinted genes but no paternally expressed imprinted genes were mis-regulated in pat nrpd1+/- (Table S6).” We have also added a new supplementary figure (Fig. S6) that describes the impacts of NRPD1 loss of imprinted gene expression.

Page 9, line 27. How could 2) occur in the homozygous mutant?

Response: Loss of NRPD1 may impact gene expression in both parents. When the nrpd1-/- mutant endosperm is investigated, we are also examining the consequences of the inheritance of these disrupted gene expression states. We refer to this as epistatic interactions of mat and pat nrpd1.

Page 10, line 9. Interesting!

Response: We strongly agree!

Page 10, line 11. Is this 2.7 versus 2.18 significant because it's statistically significant, or because it's conceptually significant?

Response: We are pointing out that the 2.7-fold value is quite similar to the predicted value of 2.18-fold, which is arrived at by simply summing the effects of mat nrpd1 and pat nrpd1. This is a conceptually significant point.

Are the examples in 3D representative, or the most convincing examples? And a big difference in ROS1 is of some concern, since that may well be expected to affect imprinting indirectly. I know I'm being picky here, but the pattern is so intriguing I'd be worried about confirmation bias.

Response: The examples in 3D are representative for those genes with significant changes in expression in both mat and pat nrpd1, and other genes also behave similarly. The antagonistic effect described for 3D can also be observed as a much broader trend affecting hundreds of genes to varying extents in Fig 3C and 3E-H. The concern about ROS1 is not clear to us but we agree that an effect of ROS1 may be one way that NRPD1 controls gene expression.

Page 10, line 18. Ok, but 0.123 is a pretty subtle negative correlation. Although I do appreciate that it clearly is not a positive correlation. If I'm understanding correctly, this was the "aha" moment, because it's exactly what one might expect of NRPD1 from the mother and father or working at cross purposes. But the numbers are getting awfully small here.

Response: It is unclear how to calibrate our expectations of effect sizes considering that our study is the first (to our knowledge) to make such a measurement involving gene expression in parental conflict. A review of the few empirical examples of parental conflict’s impact on seeds shows that parental conflict may drive small changes in seed size (Brandvain and Haig, 2018).

The evolution of quantitative traits maybe driven by selection for large effects at a small number of loci and/or by selection of small effects at a large number of loci. In a similar vein, parental conflict can impact seed phenotypes either via large effects at a few loci or via small effects at a large number of loci. Our analysis described in Fig 3D-H can fit either possibility. Large effects can be found at a few loci such as SUC2 and PICC (Fig. 3D). Smaller antagonistic effects can be found at hundreds of loci as shown in Figure 5A. The negative correlation described in this figure can be observed even upon dropping the genes that show a statistically significant differential expression in both mat and pat nrpd1+/- (slope after dropping genes significantly mis-regulated in both mat and pat nrpd1+/- is -0.126). In summary, a correlation of -0.123 strongly supports the existence of a widespread antagonistic regulatory effect.

Page 10, line 29. The point simply being that that other phenomenon is also significant even if the differences are that large?

Response: We are pointing out that the magnitude of the effects we see are similar to that observed for phenomenon such as dosage compensation.

Page 12. So, there is no effect on cleavage and no obvious effect on flanking siRNA clusters. The suspense is building...

Page 12, line 24. And not in potential regulatory regions? CNSs?

Response: We did not identify a significant enrichment for differentially methylated regions in regulatory regions. We used the relative distance function in bedtools (https://bedtools.readthedocs.io/en/latest/content/tools/reldist.html) to calculate the relationship between the genomic location of DMRs and genomic location of a differentially expressed gene. This analysis was chosen as it does not make a priori assumptions about the size of the regulatory region of a gene. A broad association between DMRs and differentially expressed genes would be indicated by a frequency far greater than 0.02. We show the results of this analysis in Fig. S8F; we find no evidence for significant enrichment of DMRs in the regulatory regions of differentially expressed genes.

Page 12, line 28. I guess it depend on whether or not the changes are in regulatory sequences no immediately apparent as part of the gene, doesn't it?

Response: We examined DNA methylation over genes here because in endosperm, unlike in other tissues, many small RNAs are genic. Moreover, DNA methylation within the gene may control transcript abundance (Eimer et al., 2018; Klosinska et al., 2016). We have also examined regulatory regions adjacent to genes in Fig S8F and found no effect.

Line 13, line 2. Not sure it's that important, but couldn't you chop all of these genes in half and see if they are no longer significant collectively?

Response: We do not think that this will provide a useful insight.

Page 14, line 15. I'm afraid I'm getting confused here with the terms cis and trans here. Just to be clear, cis means a direct effect of small RNAs that are dependent on NRPD1 on a gene and trans means anything else? But in this context, it's not clear that is what is meant. Do you mean that gene expression is determined and preset in the gametophyte? What are the levels of expression of NRPD1 in the two gametophytes?

Response: The reviewer’s interpretation of cis and trans is correct. However, the cis imprints may be preset in gametophytes or in the sporophytic tissues that surround or give rise to the gametophyte. Pol IV is known to be active either in gametophyte or in related sporophytic tissues in both the mother and the father(Kirkbride et al., 2019; Long et al., 2021; Olmedo-Monfil et al., 2010).

Page 14, line 19. Prior to fertilization?

Response: Yes, that is the idea. As described in the manuscript, Pol IV activity in either the parental sporophyte or gametophyte prior to fertilization could impact gene expression in the endosperm after fertilization.

Page 14, line 27. Do you mean driven by, or just associated with?

Response: In response to the comment, we have replaced the phrase “driven by” with “due to” for increased clarity. In wild-type, DOG1 is predominantly expressed from the paternal allele. In mat nrpd1+/-, the paternal allele is somewhat upregulated but the maternal allele, which is almost silent in wild-type, is highly expressed in mat nrpd1+/-.

Page 15, line 26. And this is really the issue. The primary conclusion, backed up by the lack (I'm assuming) of phenotypic differences between mat NRPD1 -/+ and pat NRPD1 +/- suggests that the observed differences in expression are not particularly important, unless the exceptional cases are informative.

Response: We are not sure whether the reviewer means “issue” in a negative, neutral, or positive light. Seed phenotypes are often subtle and we have not examined phenotypic differences in sufficient detail to comment.

Page 15, line 12. Yes, but I'm not at all clear what the mechanism for this is.

Response: We have tested and falsified multiple hypotheses to explain how Pol IV can regulate gene expression in endosperm. Considering the complex genetics and the difficulty of isolating endosperm, we have concluded that this is a matter for a future study. The point of this study is the discovery of Pol IV’s parental effects.

Page 15, line 23. Since this is a very small subset of genes, are these genes that you might expect to play a role in parental conflict?

Response: The functions of most genes in endosperm remain unknown. However, some have a likely role in conflict. SUC2 is antagonistically regulated by parental Pol IV (Fig. 3D). SUC2 transports sucrose, the key form of carbon imported into seeds from the mother (Sauer & Stolz, 1994).

Page 15, line 33. Indeed, these could be the informative exceptions.

Response: We believe the reviewer means that the identify of strongly antagonistically regulated genes may be informative in terms of thinking about these results in the context of parental genetic conflict, which we agree with.

Page 15, line 29. Hardly surprising, given that the paternal copy of NRPD1 is expressed at a higher level than the maternal copies, is it?

Response: It is actually somewhat surprising since we show in Fig. 2 that the sRNA production in mat and pat nrpd1+/- are comparable to that of wild-type. The higher contribution of NRPD1 from the paternal copy does not really explain the methylation differences

Page 16, line 1. So this is what you mean by in cis. Presetting?

Response: The reviewer’s previous interpretation of cis (acting directly at a target gene) is correct. However, the cis imprints may be preset in gametophyte or in the sporophytic tissues that surround or give rise to the gametophyte. Pol IV is known to be active in gametophytes and in related sporophytic tissues in both the mother and the father.

These are intriguing results that would benefit from a test of the hypothesis by comparing these result with those obtained in an outcrossing plant species.

Response: We agree that it would interesting and informative to perform similar experiments in an outcrossing species. However, loss of NRPD1 or other components of the the RdDM pathway have dramatic effects on gametogenesis in outcrossing species (Grover et al., 2018; Wang et al., 2020), preventing the detection of subtle parent-of-origin effects on seed development. Additionally, this would be a separate study.

Reviewer #3

We thank the reviewer for their comments.

• Expression of NRPD1 was 42% of WT in paternal nrpd1 and 91% of WT in maternal nrpd1, yet throughout the paper the effect of maternal nrpd1 was far stronger than paternal nrpd1. The authors may also want to confirm that protein levels follow the same pattern, in case protein degradation or post-transcriptional regulation may play a role.

Response: We show in Fig. 2 that sRNA production in mat and pat nrpd1+/- are similar to wild-type endosperm. This strongly suggests that NRPD1 protein is produced at functionally equivalent levels in wild-type, mat and pat nrpd1+/-. The finding that mat nrpd1+/- has a stronger effect on gene expression and small RNAs, despite having higher levels of NRPD1 transcript in endosperm, is consistent with our conclusion that the effects we are observing in heterozygous endosperm are due to NRPD1 action before fertilization.

P. 9 line 1 - this only seems to be true for maternal ISRs, not paternal ISRs; this claim should be narrowed.

Response: Accordingly, we have modified the text here to : “In summary, these results indicate that most maternally and some paternally imprinted sRNA loci in endosperm are dependent on Pol IV activity in the parents and are not established de novo post-fertilization.”

A small number of sRNA loci become highly depleted in maternal nrpd1 but not paternal nrpd1 (Fig. 1D, F, Fig. 2C) - are these siren loci?

Response: This is an interesting question. Siren loci have not been defined in Arabidopsis but are described as loci with high levels of sRNAs in ovules, seed coat, endosperm and embryo (Grover et al., 2020). Loci losing sRNAs in maternal nrpd1+/- include a large number of maternally expressed imprinted sRNAs (mat ISRs). We do not know if mat ISR loci are expressed in the ovule. In Erdmann et al (2017), we excluded loci that were also expressed in the seed coat from mat ISRs. Thus, these loci meet only some of the conditions for being defined as siren loci.

Fig. 2 suggests that many of the downregulated sRNA regions in maternal nrpd1 are maternally biased to begin with. Related, are genic sRNAs more likely to be affected by maternal or paternal nrpd1 than non-genic or TE sRNAs?

Response: As described in Fig. 1B and S3, loss of maternal NRPD1 has more impacts on the sRNA landscape. As a percentage of total loci, genes are more likely to be affected than TEs.

For the sRNA loci shown in Fig. 2C, how is % maternal affected in maternal vs. paternal nrpd1? These ISRs are normally maternal or paternal biased, does this change in maternal or paternal nrpd1?

Response: We assess the allelic bias of ISRs only when they have at least ten reads in the genotypes being compared. In mat nrpd1+/-, most mat ISRs lose almost all their reads (Fig. 2) and we can assess allelic bias only at 107/366 mat ISRs. As seen in the Rev. comment. Fig1, these 107 lose their maternal bias. In pat nrpd1+/-, loci with maternally biased sRNAs show somewhat increased expression (Fig 2E) but do not show an appreciable change in maternal bias (Figure Review 1). All paternal ISRs do not show any dramatic impacts on allelic bias in mat or pat nrpd1+/-. We have not added this additional datapoint to our paper because we were worried that the paper was becoming too dense – a concern also voiced by reviewer 1. However, we can add this to the manuscript if the reviewer prefers.

• Might have missed this, but I didn't see the gene ontology results (p9 line 16) shown anywhere? Would like to see significance values, fold enrichments, etc. In particular, the group of paternal nrpd1 up-regulated genes seems too small to have much confidence for GO enrichment analysis.

Response: We have added a Supplementary Table 7 with outputs of GO analyses.

• I would suggest expanding the analysis in Fig. 3D-H to explore whether the additive model is more predictive of nrpd1-/- expression levels than other potential models (epistatic, etc.) in general at all genes, or only at the subsets of genes shown, independently of whether the effects are large enough to pass the arbitrary significance cutoffs used in E-H. Identifying specifically which genes do and don't follow this additive pattern could help dissect mechanism. For example, genes following this pattern might share a TF binding site for a TF that is regulated by Pol IV.

Response: While we are interested, we currently cannot explore other models such as epistasis as this would require knock-down of NRPD1 in the endosperm and we plan to do this as part of a future study.

1. 13 line 26 - how do changes in CG methylation in maternal or paternal nrpd1 compare to changes in dme or ros1? Do either set of DMRs significantly overlap dme or ros1 DMRs? Could some of these be explained by changes in ROS1 expression, since ROS1 is a Pol IV target?

Response: Yes. It’s entirely possible that a subset of observed gene expression changes are linked to changes in ROS1 expression. However, there are no comparable methylation data for ROS1 in the endosperm. A potential role for ROS1 has been discussed on Page 11, line 4. Comparison with DMRs in the dme endosperm is difficult. dme mutant endosperm has low non-CG methylation (Ibarra et al., 2012). We have unpublished data showing that the expression of genes involved in RNA-directed DNA methylation (RdDM) is reduced in the dme endosperm. It is therefore difficult to understand if and how DME-mediated demethylation may impact RdDM.

P. 10 line 3 - is the overlap of 36 out of 51 genes unlikely to occur by chance

Response: A hypergeometric test indicates that this is indeed significant. We have added it to text on Page 9, line 34.

In sRNA and mRNA-seq libraries, what was the overall maternal/paternal ratio in each library? Did loss of Pol IV affect this?

The graphs above show the maternally derived fraction of mRNA and sRNA libraries for different genotypes. Please note that the Ler nrpd1 mutant was generated by backcrossing Col-0 nrpd1+/- into Ler. Some Col-0 regions remain in this background and are called “hold-outs”. Reads mapping to these hold-outs have been excluded while calculating the maternal fraction of each library described in the graph above. We cannot confidently judge if the overall maternal fraction of the mRNA transcriptome is affected by loss of NRPD1 as we likely need more replicates. However, we find that loss of all NRPD1-dependent sRNAs (as in the nrpd1 null mutant) leaves behind sRNAs that roughly reflect the genomic 2:1 ratio.

P. 9 line 22 - how many paternally and maternally expressed imprinted genes were considered? Were imprinted genes statistically more likely to be misregulated in mat nrpd1?

Response: We considered 128 maternally and 43 paternally expressed genes that had been previously been identified as imprinted in Col x Ler crosses (Pignatta et al 2014). We have modified the manuscript to describe effects on imprinted genes: “ The expression of imprinted genes is known to be regulated epigenetically in endosperm. In mat nrpd1+/- imprinted genes were more likely to be mis-regulated than expected by chance (hypergeometric test p-15) – 15 out of 43 paternally expressed and 45 out of 128 maternally expressed imprinted genes were mis-regulated in mat nrpd1+/- while two maternally expressed imprinted genes but no paternally expressed imprinted genes were mis-regulated in pat nrpd1+/- (Table S6). “ We have also added a supplementary figure (Figure S6) that focuses on genic mRNA imprinting in NRPD1 heterozygotes and homozygous mutants.

References cited in the response

Brandvain, Y., & Haig, D. (2005). Divergent Mating Systems and Parental Conflict as a Barrier to Hybridization in Flowering Plants. The American Naturalist, 166(3), 330–338. https://doi.org/10.1086/432036

Brandvain, Y., & Haig, D. (2018). Outbreeders pull harder in a parental tug-of-war. Proceedings of the National Academy of Sciences, 115(45), 11354–11356. https://doi.org/10.1073/pnas.1816187115

Eimer, H., Sureshkumar, S., Singh Yadav, A., Kraupner-Taylor, C., Bandaranayake, C., Seleznev, A., Thomason, T., Fletcher, S. J., Gordon, S. F., Carroll, B. J., & Balasubramanian, S. (2018). RNA-Dependent Epigenetic Silencing Directs Transcriptional Downregulation Caused by Intronic Repeat Expansions. Cell. https://doi.org/10.1016/j.cell.2018.06.044

Grover, J. W., Burgess, D., Kendall, T., Baten, A., Pokhrel, S., King, G. J., Meyers, B. C., Freeling, M., & Mosher, R. A. (2020). Abundant expression of maternal siRNAs is a conserved feature of seed development. Proceedings of the National Academy of Sciences of the United States of America, 117(26), 15305–15315. https://doi.org/10.1073/pnas.2001332117

Grover, J. W., Kendall, T., Baten, A., Burgess, D., Freeling, M., King, G. J., & Mosher, R. A. (2018). Maternal components of RNA ‐directed DNA methylation are required for seed development in Brassica rapa. The Plant Journal, 94(4), 575–582. https://doi.org/10.1111/tpj.13910

Ibarra, C. A., Feng, X., Schoft, V. K., Hsieh, T.-F., Uzawa, R., Rodrigues, J. A., Zemach, A., Chumak, N., Machlicova, A., Nishimura, T., Rojas, D., Fischer, R. L., Tamaru, H., & Zilberman, D. (2012). Active DNA Demethylation in Plant Companion Cells Reinforces Transposon Methylation in Gametes. Science, 337(6100), 1360–1364. https://doi.org/10.1126/science.1224839

İltaş, Ö., Svitok, M., Cornille, A., Schmickl, R., & Lafon Placette, C. (2021). Early evolution of reproductive isolation: A case of weak inbreeder/strong outbreeder leads to an intraspecific hybridization barrier in Arabidopsis lyrata. Evolution, 75(6), 1466–1476. https://doi.org/10.1111/evo.14240

Kirkbride, R. C., Lu, J., Zhang, C., Mosher, R. A., Baulcombe, D. C., & Chen, Z. J. (2019). Maternal small RNAs mediate spatial-temporal regulation of gene expression, imprinting, and seed development in Arabidopsis. Proceedings of the National Academy of Sciences, 116(7), 2761–2766. https://doi.org/10.1073/pnas.1807621116

Klosinska, M., Picard, C. L., & Gehring, M. (2016). Conserved imprinting associated with unique epigenetic signatures in the Arabidopsis genus. Nature Plants, 2, 16145. https://doi.org/10.1038/nplants.2016.145

Lafon-Placette, C., Hatorangan, M. R., Steige, K. A., Cornille, A., Lascoux, M., Slotte, T., & Köhler, C. (2018). Paternally expressed imprinted genes associate with hybridization barriers in Capsella. Nature Plants, 4(6), 352–357. https://doi.org/10.1038/s41477-018-0161-6

Long, J., Walker, J., She, W., Aldridge, B., Gao, H., Deans, S., Vickers, M., & Feng, X. (2021). Nurse cell­–derived small RNAs define paternal epigenetic inheritance in Arabidopsis. Science, 373(6550). https://doi.org/10.1126/science.abh0556

Olmedo-Monfil, V., Durán-Figueroa, N., Arteaga-Vázquez, M., Demesa-Arévalo, E., Autran, D., Grimanelli, D., Slotkin, R. K., Martienssen, R. A., & Vielle-Calzada, J.-P. (2010). Control of female gamete formation by a small RNA pathway in Arabidopsis. Nature, 464(7288), 628–632. https://doi.org/10.1038/nature08828

Raunsgard, A., Opedal, Ø. H., Ekrem, R. K., Wright, J., Bolstad, G. H., Armbruster, W. S., & Pélabon, C. (2018). Intersexual conflict over seed size is stronger in more outcrossed populations of a mixed-mating plant. Proceedings of the National Academy of Sciences, 115(45), 11561–11566. https://doi.org/10.1073/pnas.1810979115

Sauer, N., & Stolz, J. (1994). SUC1 and SUC2: two sucrose transporters from Arabidopsis thaliana; expression and characterization in baker’s yeast and identification of the histidine-tagged protein. The Plant Journal, 6(1), 67–77. https://doi.org/10.1046/j.1365-313X.1994.6010067.x

Wang, Z., Butel, N., Santos-González, J., Borges, F., Yi, J., Martienssen, R. A., Martinez, G., & Köhler, C. (2020). Polymerase IV Plays a Crucial Role in Pollen Development in Capsella. The Plant Cell, 32(4), 950–966. https://doi.org/10.1105/tpc.19.00938

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

Short Summary:

In this study, Satyaki and Gehring investigate the role of RNA Pol IV in Arabidopsis endosperm, focusing on parent-of-origin-specific functions and potential mechanisms. Using a combination of gene expression, sRNA profiling, and DNA methylation data from reciprocal crosses, they find that maternal loss of Pol IV has distinct, and in some cases opposite, effects on gene expression compared to paternal loss of Pol IV. This is also true to a lesser extent for sRNAs and DNA methylation, consistent with the function of RNA Pol IV in driving 24nt sRNA production and targeted DNA methylation through the RdDM pathway. DNA methylation was more strongly affected by paternal Pol IV loss while expression was much more affected in maternal Pol IV loss. Surprisingly, the authors consistently find no evidence that the minor changes sRNA production or DNA methylation in maternal/paternal nrpd1/+ heterozygotes are correlated with gene expression changes in either heterozygote. However, while the mechanism remains unclear, evidence presented here that maternal and paternal Pol IV can have opposite, additive effects on gene/sRNA expression and phenotype (rescue of paternal excess crosses) is convincing and an interesting finding, potentially consistent with the idea that Pol IV helps mediate parental conflict in endosperm.

Major Comments/suggestions:

• Expression of NRPD1 was 42% of WT in paternal nrpd1 and 91% of WT in maternal nrpd1, yet throughout the paper the effect of maternal nrpd1 was far stronger than paternal nrpd1. The authors may also want to confirm that protein levels follow the same pattern, in case protein degradation or post-transcriptional regulation may play a role.
• P. 9 line 1 - this only seems to be true for maternal ISRs, not paternal ISRs; this claim should be narrowed.
• A small number of sRNA loci become highly depleted in maternal nrpd1 but not paternal nrpd1 (Fig. 1D, F, Fig. 2C) - are these siren loci? Fig. 2 suggests that many of the downregulated sRNA regions in maternal nrpd1 are maternally biased to begin with. Related, are genic sRNAs more likely to be affected by maternal or paternal nrpd1 than non-genic or TE sRNAs?
• For the sRNA loci shown in Fig. 2C, how is % maternal affected in maternal vs. paternal nrpd1? These ISRs are normally maternal or paternal biased, does this change in maternal or paternal nrpd1?
• Might have missed this, but I didn't see the gene ontology results (p9 line 16) shown anywhere? Would like to see significance values, fold enrichments, etc. In particular, the group of paternal nrpd1 up-regulated genes seems too small to have much confidence for GO enrichment analysis.
• I would suggest expanding the analysis in Fig. 3D-H to explore whether the additive model is more predictive of nrpd1-/- expression levels than other potential models (epistatic, etc.) in general at all genes, or only at the subsets of genes shown, independently of whether the effects are large enough to pass the arbitrary significance cutoffs used in E-H. Identifying specifically which genes do and don't follow this additive pattern could help dissect mechanism. For example, genes following this pattern might share a TF binding site for a TF that is regulated by Pol IV.
• P. 13 line 26 - how do changes in CG methylation in maternal or paternal nrpd1 compare to changes in dme or ros1? Do either set of DMRs significantly overlap dme or ros1 DMRs? Could some of these be explained by changes in ROS1 expression, since ROS1 is a Pol IV target?
• P. 10 line 3 - is the overlap of 36 out of 51 genes unlikely to occur by chance?

Minor Comments:

• In sRNA and mRNA-seq libraries, what was the overall maternal/paternal ratio in each library? Did loss of Pol IV affect this?
• P. 9 line 22 - how many paternally and maternally expressed imprinted genes were considered? Were imprinted genes statistically more likely to be misregulated in mat nrpd1?

Significance

Significance:

PolIV is a plant-specific polymerase that functions part of the plant-specific RNA-directed DNA methylation pathway, which has been very well characterized in Arabidopsis. Mutations in PolIV were previously shown to rescue paternal excess crosses when inherited paternally (Erdmann et al. 2017), and this study extended that finding to show that maternal vs. paternal loss of Pol IV has opposite effects on seed survival in paternal excess crosses. Only one other example (met1) of opposed paternal vs. maternal effects on seed development is known, making Pol IV a useful tool for studying why and how these effects occur. As the authors note, the dominant theory on 'why' involves Pol IV mediating parental conflict over resource allocation in the seed, and the opposite effects of Pol IV maternal/paternal loss at some genes support this hypothesis. The 'how' remains unclear, although this study eliminates several possibilities, and the most likely remaining model is that Pol IV parent-of-origin specific effects occur mostly in trans. Future work can build on these findings to identify the mechanism by which Pol IV achieves these parent-of-origin specific effects.

My background is mostly in plant epigenetics and genomics.

Referee Cross-commenting

The other referee comments seem fair, and I have not comments at this time.

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

Learn more at Review Commons

---

Referee #2

Evidence, reproducibility and clarity

This manuscript provides evidence that a loss of either the maternal or paternal copy of NRPD1 have different, and sometime opposite, effects on the accumulation of small RNAs and on expression of a subset of genes, with a loss of the maternal copy having more substantial effects. The manuscript is well written, and the conclusions, as far as they go, are justified by the data, which are effectively presented. The overall effect is subtle but informative and according to the authors support a parental conflict model for imprinting. The experiments failed to find a smoking gun in the form of a mechanism to explain how or why the maternal and paternal alleles have different effects and the explanation for a lack of clear phenotypic differences was reasonable, but untested. I would have like to see it tested by looking in a plant species that is outcrossing and highly polymorphic. However, I do appreciate that the observation that the male and female alleles can have distinct effect when mutant is an important clue. My specific comments below may reflect confusion on my part, rather than real issues. If that is that I hope that confusion can aid in clarifying what are in places quite subtle points.

Specific comments:

Page 6, last paragraph: "Because the endosperm is triploid, in these comparisons there are 3 (wild-type), 2 (pat nrpd1+/-), 1 (mat nrpd1+/-) and 0 (nrpd1-/-) functional NRPD1 alleles in the endosperm. However, NRPD1 is a paternally expressed imprinted gene in wild-type Ler x Col endosperm and the single paternal allele contributes 62% of the NRPD1 transcript whereas 38% comes from the two maternal alleles (Pignatta et al., 2014). Consistent with paternal allele bias in NPRD1 expression, mRNA-Seq data shows that NRPD1 is expressed at 42% of wild- type levels in pat nrpd1+/- and at 91% of wild-type levels in mat nrpd1+/- (Supplementary Table 6)".

I would think this would really complicate the analysis. Should all of the dosage values include NPRD1 imprinting values? That is to say, expressed in terms of expression values? This is also a bit confusing. The maternal copies together express 38% of the transcript, so why isn't the mat nrpd1 at 68%, rather than 91%? In any event, given this imprinting and differences in dosage of the male and female it appears that two variables, parental origin and expression levels are being compared. Since 91% is awfully close to 100%, are the mat pat comparisons really just comparing low with nearly normal expression of NRPD1? And actually, given that, the outsized effect of the mat nrpd1 +/- is even more striking.

Page 4, Line 16. I'm afraid it's still a bit difficult to understand what was being compared what in this section. Please clarify.

Page 5, Line 5. I'm sure this is fine, but it's not entirely clear what is from the previously published paper and what is reanalysis here. All the crosses and measurements were made then, but not organized in this way?

Page 6, Line 26. This is an excellent dosage series, but it's complicated by imprinting. So it's not 3, 2, 1, 0 effective copies. If we set the paternal copy at ~1 and each maternal at ~0.1, then it's 1.2 (wild type), 0.20 (pat nrpd1+/-), 1 (mat nrpd1+/-), and 0 (nrpd1-/-).

Page 6, line 31. Is the main thing we are comparing the difference between expression at 42% verses 91% of wild type?

Page 7, line 11. So, the overall effect in either direction on smRNA gene targets was really quite small, and I'm guessing the effect on gene expression was even smaller.

Page 7, line 17. I take it that it is this difference, rather than the overall numbers that is of interest.

Page 9, line 2. Really interesting, since one might expect that these are methylated loci that would be expected to be fed into any existing embryo maintenance methylation pathway. Surprising that they are maintained independently.

Page 9, line 22. Proportion of total imprinted genes? Did the mutant obviate/enhance the imprinting?

Page 9, line 27. How could 2) occur in the homozygous mutant?

Page 10, line 9. Interesting!

Page 10, line 11. Is this 2.7 versus 2.18 significant because it's statistically significant, or because it's conceptually significant? Are the examples in 3D representative, or the most convincing examples? And a big difference in ROS1 is of some concern, since that may well be expected to affect imprinting indirectly. I know I'm being picky here, but the pattern is so intriguing I'd be worried about confirmation bias.

Page 10, line 18. Ok, but 0.123 is a pretty subtle negative correlation. Although I do appreciate that it clearly is not a positive correlation. If I'm understanding correctly, this was the "aha" moment, because it's exactly what one might expect of NRPD1 from the mother and father or working at cross purposes. But the numbers are getting awfully small here.

Page 10, line 29. The point simply being that that other phenomenon is also significant even if the differences are that large?

Page 12. So, there is no effect on cleavage and no obvious effect on flanking siRNA clusters. The suspense is building...

Page 12, line 24. And not in potential regulatory regions? CNSs?

Page 12, line 28. I guess it depend on whether or not the changes are in regulatory sequences no immediately apparent as part of the gene, doesn't it?

Line 13, line 2. Not sure it's that important, but couldn't you chop all of these genes in half and see if they are no longer significant collectively?

Page 14, line 15. I'm afraid I'm getting confused here with the terms cis and trans here. Just to be clear, cis means a direct effect of small RNAs that are dependent on NRPD1 on a gene and trans means anything else? But in this context, it's not clear that is what is meant. Do you mean that gene expression is determined and preset in the gametophyte? What are the levels of expression of NRPD1 in the two gemetophytes?

Page 14, line 19. Prior to fertilization?

Page 14, line 27. Do you mean driven by, or just associated with?

Page 15, line 26. And this is really the issue. The primary conclusion, backed up by the lack (I'm assuming) of phenotypic differences between mat NRPD1 -/+ and pat NRPD1 +/- suggests that the observed differences in expression are not particularly important, unless the exceptional cases are informative.

Page 15, line 12. Yes, but I'm not at all clear what the mechanism for this is.

Page 15, line 23. Since this is a very small subset of genes, are these genes that you might expect to play a role in parental conflict?

Page 15, line 33. Indeed, these could be the informative exceptions.

Page 15, line 29. Hardly surprising, given that the paternal copy of NRPD1 is expressed at a higher level than the maternal copies, is it?

Page 16, line 1. So this is what you mean by in cis. Presetting?

Page 16, line 9. So ideally, one would want to look at a highly polymorphic out-crosser. I'm not suggesting that for this paper, but would this be a good test of the hypothesis? How about maize?

Page 16, line 15. But the pat and mat heterozygotes looked the same. No differences in phenotype?

Page 17, line 22. I'm confused, since aren't most 24 nt smRNAs dependent on POLIV (Figure S2)? Do you mean differentially regulated smRNAs? Expression of POLIV specifically in one or the other parent?

Page 17, line 23. How are you defining important here? Important because at least in the female NPRD1 is not expressed in the central cell? But not important, since this mutant has no effect on phenotype except in an imbalanced cross.

Page 18, line 13. For this reason, it would be nice to know much more about these genes. Mutant phenotypes, for instance. And how many of these have this feature conserved?

Significance

These are intriguing results that would benefit from a test of the hypothesis by comparing these result with those obtained in an outcrossing plant species.

Referee Cross-commenting

I agree that the other comments seem both fair and reasonable.

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 study addresses key aspects of gene regulation in the developing endosperm of flowering plants. The endosperm is the product of the fusion of the (normally diploid) female central cell with one of the sperm cells, and is indispensable for nourishing the developing embryo, among other important functions. Evolutionary models predict that in flowering plants, the endosperm ought to be the tissue in which parental conflict over the allocation of (female) resources to progeny should manifest. Consequently, endosperm gene expression (including the phenomenon of genomic imprinting) and developmental trajectories have been studied from various perspectives, including the possibility of the fast build-up of reproductive barriers due to failing endosperm (and thus seed) development.

More specifically, this study utilizes knock-out mutants of the NRPD1 gene, which codes for the largest subunit of RNA Polymerase IV (Pol IV), which is part of the RNA-directed DNA methylation (RdDM) pathway. It builds on previous work by the authors that suggested an important role for Pol IV in mediating allelic dosage in developing endosperm, that small interfering RNAs are produced from both paternally and maternally-derived alleles (Erdmann et al. 2017; contra purported claims by other labs), and that normally inviable seeds from paternal-excess crosses (2n x 4n) can be largely rescued by knocking out individual (paternal) components of the RdDM pathway (Satyaki & Gehring 2019).

Here, Satyaki & Gehring characterize a variety of expression responses in reciprocal heterozygotes, i.e. products of crosses between a homozygous WT and a homozygous nrpd1 mutant (all with the Ler and Col-0 accessions of A. thaliana). The resulting heterozygotes differ in whether the maternal parent (mat nrpd1+/-) or the paternal parent (pat nrpd1+/-) contributed the nrpd1 allele. In addition, mRNA and sRNA expression was also assessed for the WT (+/+) and the homozygous nrpd1 lines (-/-).

Key findings of this work are that the loss of Pol IV in maternal and paternal parents has different consequences for endosperm gene expression, some of which appear to be antagonistic. In other words, the presence of a functioning Pol IV in the mother and father have parent-of-origin effects on the resulting endosperm. Furthermore, one parent's copy of NRPD1 was found to be sufficient for the production of most Pol IV-dependent sRNAs, yet with a fairly small number of mostly non-overlapping loci losing sRNAs upon loss of either maternal or paternal NRPD1. Pol IV activity in the father and mother is shown to have distinct impacts on the endosperm's DNA methylation landscape.

Interestingly, while the proportion of mis-regulated genes seems small in both heterozygotes, it is much more restricted in pat nrpd1+/-. Jointly, the authors' results suggest that paternal and maternal Pol IV are genetically antagonistic and that their effects on endosperm transcription in heterozygotes is established before fertilization.

Major Comments:

I believe that this is a very sound and authoritative study. The analysis of all data seems appropriate and robust, and many connections between the data (and subsets of data) and their possible interpretations have been considered. In fact, in the massive Results section, some interpretations are supported by cited references (this is not meant as a critique). However, I wonder about the length of the Results section, and the balance between it and the relatively short Discussion section. It is difficult for me to nail down any part of Results that might be shortened, as I could not find clear redundancies. I also think that the level of speculation is absolutely warranted, and I did not find excessive claims being made to this or that end. Rather, I suggest to broaden the perspective somewhat (in their Discussion; see below under Significance), which might allow people with a less mechanistic perspective to grasp the potential relevance of this work for non-model plant systems studied mostly by evolutionary geneticists.

One aspect that might warrant more scrutiny is the mapping of sRNA reads to the reference genome. I found the short section of this (M&M section, page 20, lines 23-25) to be too brief. It is not clear to me which of ShortStack's v3 weighting scheme the authors used, which is relevant for multi-mapping reads (see NR Johnson et al. 2016, G3). In addition, it is not mentioned whether zero mismatches were allowed. Perhaps this is described in more detail in Erdmann et al. (2017), but even if so, it deserves to be clarified here.

All in all, I find this work to be meticulously presented and the data to be thoughtfully interpreted. The major conclusions seem to be convincing and adequate, given the underlying data. I have no qualms about replication issues, nor about statistics.

Minor Comments:

The manuscript is well-written and concise, despite the length of the Results section. The verbal clarity and absence of typos or grammatical issues is superb. I did find some of the Figures to be somewhat "un-intuitive", in the sense that it takes acute concentration for an outsider (of sorts) to gather and interpret the underlying data. This is probably due to the many cross-comparisons of differences between two genotypes on one axis and those of a different pair of genotypes on the other axis. I am not sure how this issue can be ameliorated (nor whether this is really necessary); however, from a technical point of view, all Figures and Suppl. Figures are really well-done.

The list of references seems adequate in terms of citing relevant (both older and very recent) publications. However, almost all cited papers concern Arabidopsis or other model species; I suggest to consider adding a few relevant studies on non-Brassicaceae (whether considered model taxa or not), in conjunction with my suggestion (in Significance) to potentially broaden the scope by searching for natural phenomena that also involve parent-of-origin effects on endosperm/seed development. Curiously, many of the references are "incomplete" in the sense of stopping with the journal's name, then stating the doi, i.e. they lack volume numbers and page/article numbers. This should be harmonized throughout.

Significance

Significance:

While part of the earlier data from Erdmann et al. (2017) were re-analyzed in the present study, the vast amount of data are new and concern the expression consequences at the diploid level (2n x 2n crosses), and thus may prove to be more relevant for future comparisons with non-model flowering plants, either for normal intraspecific seed development or (partly) failing crosses between slightly diverged evolutionary lineages. In my view, this study presents a significant advance in understanding the downstream consequences (endosperm mRNA and sRNA expression levels, levels and patterns of DNA methylation) of a perturbed Pol IV expression in both parents or the female vs. male parent. Much of the emphasis in the field has been on paternal-excess crosses, within the larger realm of the "triploid block" or the reproductive barriers between plants of different ploidy (typically diploids x tetraploids in both cross directions).

The fact that the molecular consequences of disabled Pol IV in one or both parents were assessed in balanced crosses (and not interploidy crosses) may allow an easier connection to natural phenomena such as (partial) hybrid seed lethality between closely related lineages of flowering plants, where parent-of-origin effects have emerged in the recent literature, both at the phenotypic level of endosperm/seed growth and at the molecular level (perturbed imprinting in the endosperm, mRNA and sRNA expression levels, etc.). It thus might prove worthwhile to screen recent papers in Solanum, Mimulus, and Capsella to evaluate the possibility of "connections" between the current data and recent, admittedly more descriptive, findings in diverse taxa that don't offer the same genomic resources as Arabidopsis.

What the current version of this work already does is relating the finding of partly antagonistic influences of pat nrpd1+/- versus mat nrpd1+/- on endosperm mRNA expression to evolutionary models championed by D. Haig ("parental conflict", "kinship theory"). My above suggestions would strengthen such connections and likely would broaden the appeal of this work to scientists with diverse backgrounds outside the core expertise of plant molecular/developmental biologists. In any case, I see the prime scientific audience as the latter group, but I see potential to intrigue people with a more evolutionary background.

Field of expertise:

population genomics, hybrid seed lethality, speciation, genomic imprinting, evolutionary models.

Referee Cross-commenting

I agree that the other referee comments (while being quite complementary to mine due to differences in main expertise) seem both fair and reasonable.

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 rebuttal 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

Description of the planned revisions

From Review Commons: Please find below our 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)):

---

Comment #1: In this manuscript, the authors follow up on an interesting finding that varA null mutants of V. cholerae form spherical cells in stationary phase. The authors determine that this cell rounding is due to weakening of the cell wall via less production of tetrapeptide cross links. Mutation of the regulator csrA and the enzyme aspA lead to a model in which a varA mutant cell lacks aspartate leading to low cross-linked cell wall that is unable to hold the typical curved V. cholerae shape. The data are robust, and the manuscript is clearly written.

Authors’ reply #1: We very much appreciate the reviewer’s accurate summary and the appraisal of the robust data and the clearly written manuscript.

• *

Comment #2: I think the finding is quite interesting, even though it is not clear to me if this observed cell morphology has a biological function or if it is an artifact of completly removing VarA. However, this manuscript builds the foundation to further test this question.

Authors’ reply #2: We agree with the reviewer. However, it is worth mentioning that this two-component system (TCS) has been first described in 1998 yet the input signal (or repression of the signal under certain conditions) still remains elusive. Maybe, this isn’t too surprising, given that studies on V. cholerae are strongly biased towards its pathogenic lifestyle, while the varA/varS system is highly conserved among Gram-negative bacteria including non-pathogenic environmental V. cholerae strains. These strains can live under very diverse conditions of slow or fast growth, including long starvation periods. Unfortunately, we still lack significant insight into this part of the V. choleraebiology. We therefore believe that the current study is very important, as the elucidation of the molecular mechanism of the observed shape transition within the varA mutant will foster fresh hypotheses on the role that the system plays in V. cholerae and what signals might be sensed.

We would also like to remind the editor and reviewer(s) that a plethora of studies have been published based on varA and varS (and csrA) deletion mutants of V. cholerae with various readouts ranging from transcriptomics to quorum sensing defects, impairment of virulence, etc. Thus, the argument that the complete removal of varA might cause an artifact seems equally valid for previous work by others and, maybe, even the vast majority of studies in which TCS are investigated for which the sensed signal has yet to be identified.

In conclusion, we propose to address this valid point of critique in the revised manuscript by clearly stating the caveat of the gene deletion(s). However, as the reviewer correctly stated, “this manuscript builds the foundation to further test this question.”

Comment #3: The data all support the conclusions, but I do think the authors could have really confirmed their model by connecting mutations in csrA and aspA to restoration of high cross-linked cell well similar to the WT strain as done in Fig. 2. As it stands, this is still somewhat hypothetical and has not been directly demonstrated, although I do think their model is correct and these experiments will be conformation of it.

Authors’ reply #3: We thank the reviewer for their comment and the assumption that our model might be correct. It is very unlikely that the csrA suppressor mutant(s) or the ∆varA∆aspA mutant maintain the low level of cross-links and the high level of dipeptides that we observed for the ∆varA mutant. Indeed, it would be unclear how the cells could restore the Vibrio shape that we visualized in the phase contrast image under such conditions. However, as this point seems very important to the reviewer (see also comment #7 below), we will perform the suggested cell wall analysis of these mutants and include the new data in the revised manuscript.

Comment #4: I also have a few other suggestions to improve the manuscript, but in sum I think it is a well-done research study that will be interesting to research in V. cholerae and other gamma proteobacteria.

Authors’ reply #4: Once again, we thank the reviewer for their kind words.

Major comments:

Comment #5: 1. The enrichment for suppressors is very creative and connected the varA impact on cell morphology to misregulation of csrA as 10/10 mutants were ultimately linked to this gene. However, insertion in aspA should also suppress this phenotype, and I am curious why this gene was not identified in the transposon suppressor screen.

Authors’ reply #5: This is a very relevant and important comment. The reason why we did not isolate ∆varA-aspA::Tn mutants is most likely due to a growth defect that we observed for the double ∆varA∆aspA mutant compared to the ∆varA-csrA suppressor mutant(s). In the figure on the right, respective growth curves are shown [∆varA∆aspA in orange and the ∆varA-csrA suppressor mutant ∆varA-Tn A in gray]. Any ∆varA-aspA::Tn mutant is therefore expected to be outcompeted by the ∆varA-csrA suppressor mutants during the enrichment process. We will include this information and the corresponding data (e.g., final growth curves after 3 biologically independent experiments) in the revised manuscript.

[figure not shown in online form]

Comment #6: 2. The authors should complement at least one of their varA/csrA mutants with csrA.

Authors’ reply #6: Agreed. We are in the process of performing the suggested experiment and will include the results in the revised manuscript.

Comment #7: 3. The changes in cell wall structure are not directly connected to the genetic identification of csrA and aspA. Yes, I agree their model makes sense, but to really nail it down they should analyze the cell wall composition in the varA/csrA and varA/aspA double mutants and show it returns to WT levels of crosslinking.

Authors’ reply #7: As mentioned above, we will perform those cell wall analyses (see also authors’ reply #3 above), as requested.

Comment #8: 4. Does deletion of aspA in the WT or varA mutant impact the growth rate?

Authors’ reply #8: This is indeed the case in the ∆varA background but not in the WT background (as shown under authors reply #5). These data will be included in the revised manuscript.

• *

Minor comments

Comment #9: 5. Are the round cells able to divide? The data in Fig. S2 would suggest they can based on the increase in CFUs from hour 6 to hour 8, but the authors never comment on this point and it might be worth addressing in the discussion.

Authors’ reply #9: We never observed diving round cells. Indeed, the increase of the optical density and CFUs from 6h to 8h is most likely based on those bacteria that have not yet changed their cellular morphology and therefore keep dividing (see below the imaging/quantification for this timeframe taken from Fig. 2). What we observed though is that upon dilution into fresh medium, the round cells start to elongate and then divide resulting in newborn Vibrio shaped cells. We will include these new data in the revised manuscript.

[figure not shown in online form]

• *

Comment #10: 6. Fig. S2-Why does the OD600 increase from 8 to 24 hours but the CFUs decrease in the varA mutant?

Authors’ reply #10: This is an interesting observation that might reflect the presence of dead but not yet lysed cells in these cultures. Indeed, while it looks as if the OD600 values are still increasing for the ∆varA mutant at 24h, we cannot exclude at this point that the OD600 values increased during the 16h-time interval and went again down at 24h (e.g., like shifting the WT peak to later time points/the right of the X-axis). Notably, the purpose of this figure was mostly to i) indicate the slower growth of the ∆varA mutant while ii) emphasizing that late during growth (e.g., 24h here) the strain can still reach similar OD600 values as well as CFUs/ml as the WT strain. We will change Fig. S2 to better emphasize these two points in the revised manuscript.

Comment #11: 7. Lines 309-A little bit more detail here would help the reader. Are the authors examining whole cell lysates or lysates from specific cellular components? I am actually very surprised this worked as there are so many proteins in crude cell lysates.

Authors’ reply #11: Indeed, these are whole cell lysates, which were prepared as described in the methods section (lines 482 onwards; “SDS-PAGE, Western blotting and Coomassie blue staining”). We fully agree that there are many proteins in the crude cell lysates and realized that we might not have explained well enough that only the gel region containing the overproduced band in the ∆varA strain and the same location in the WT sample were analyzed by mass spectrometry (even though we were referring to the “gel pieces” in line 498 onwards). Please accept our sincerest apologies for this neglect. During the revision, we will ensure that this information is explicitly stated and that these details are included in the main text and the methods section.

---

Comment #12: 8. Lines 320-321-I don't think there is evidence that CsrA enhances aspA RNA translation, merely that CsrA enhances AspA protein production. It is likely through increasing translation, but this cannot be concluded without direct evidence.

Authors’ reply #12: We fully agree and thank the reviewer for this important comment. Indeed, we meanwhile know that the aspA mRNA levels also increase in the ∆varA mutant strain (which might or might not be linked to enhanced translation). We will add these transcript level data to the revised manuscript and discuss all possibility that could explain the AspA overproduction.

Comment #13: 9. Line 348-350-I do not understand the logic of this sentence stating that the "..until now, the signal that abrogates VarA phosphorylation..." as this manuscript does not contribute to our understanding of the VarS signal.

• *Authors’ reply #13: We apologize that this sentence or the logic behind it wasn’t clear. As this is a combined result and discussion section, the aim of the sentence was to put the observed shape transition of the bacteria into a broader context, which required us to mentioned that the input signal is still unknown. We will make sure that this becomes more obvious in the revised manuscript by rephrasing this sentence.

Comment #14: 10. I am curious if the total volume of the round versus curved cells is constant at 20 hours. This should be easy to determine using ImageJ and worth reporting.

Authors’ reply #14: We are not entirely sure how this question is relevant for the study (e.g., for this report on the observed shape transition phenotype it doesn’t matter if the cells maintain the same volume or not). However, given the importance for the reviewer, we will perform these volume measurements on our images and add a sentence to the revised manuscript on the analysis’ outcome (plus include the data as a supplementary panel).

• *

Reviewer #1 (Significance (Required)):

Comment #15: Understanding changes to cell morphology and their biological implications is a growing area of microbiology. This study makes a new contribution to this area by demonstrating a round, spherical form of V. cholerae that is driven by alterations to the cell that decrease cell-wall cross linking.

Authors’ reply #15: Once again, we thank the reviewer for this summary and for placing our study into context. We agree that cell morphological changes and the underlying molecular mechanism(s) are an exciting and growing area of microbiology.

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

Comment #16: In this manuscript, Rocha et al studied the effect of the VarA response regulator on cell shape of Vibrio cholerae. VarA is part of a two-component system that also includes the histidine kinase VarS. It has previously been shown that VarA activates the expression of three redundantly acting regulatory RNAs called CsrB, CsrC, and CsrD. All three Csr RNAs share the same regulatory principle, which is to sequester the activity of the RNA-binding protein CsrA. CsrA in turn can bind hundreds of mRNA species in the cell, which in the majority of cases results in reduced translation of these mRNAs (in addition to various other modes of action that have been reported). Here, the authors discovered that deletion of the varA gene results in an abnormal, spherical cell shape in stationary phase grown V. cholerae cells. Biochemical analysis revealed an unusual peptidoglycan composition in varA-deficient cells, showing increased levels of dipeptides, a reduction of tetrapeptides, and an overall decrease in peptide-cross linkage. Interestingly, the varA phenotype was complemented by the addition of conditioned medium from wild-type cells, which are likely to provide peptidoglycan building blocks in trans. The authors further discover that varA-deficiency results in AspA over-production, which could be linked to the activity of CsrA. The authors speculate that high AspA levels deplete the cell of aspartate, which is required to produce peptidoglycan precursors. The manuscript is interesting, well-written, and the rationale of the experiments is easy to follow.

Authors’ reply #16: We thank the reviewer for this excellent summary and the kind words on the quality of the manuscript.

Comment #17: However, I have two major points of criticism, which reduce my enthusiasm for this work. First, the molecular pathways that links varA-deficiency to increased AspA levels is incomplete: please clarify how CsrA activates AspA levels and if this phenotype is linked to direct binding of CsrA to the aspA mRNA and if so how is activation is achieved at the molecular level.

Authors’ reply #17: We thank the reviewer for the comment. However, we never had the intention to decipher the entire pathway and it is indeed possible that intermediate regulators might be involved. Notably, the first part of the signaling pathway (VarA -> CsrB,C,D -> CsrA) seemed well established in the literature and we truly believe that our work supports this part of the pathway (given the numerous csrA suppressor mutants that we obtained in the varA-minus background). For the link between CsrA and AspA, we indeed do not provide direct evidence. Nonetheless, we discuss recent work in Salmonella by mentioning “Interestingly, previous studies identified the aspA mRNA amongst hundreds of direct CsrA targets in Salmonella using the CLIP-seq technique to identify protein-RNA interactions (32)”. Notably, this finding by Holmqvist et al. (2016, EMBO J.) has been reproduced for E. coli by Potts et al. (2017, Nat. Commun.; see Supplementary Data file 1, CsrA CLIP-seq data; with three highly significant peaks corresponding to aspA mRNA binding), an information that we will add to the revised manuscript. Of course, neither Salmonella nor E. coli belongs to the same genus as V. cholerae (though, of course, all are gamma-Proteobacteria). Thus, to accommodate the reviewer’s comment, we will revise the manuscript to include the caveat that direct CsrA binding of the aspA mRNA has been shown in both Salmonella and E. coli but that it is still feasible that intermediate regulatory proteins might be involved in the case of V. cholerae. We will also revise the model to show such potential intermediate steps between CsrA and AspA.

Comment #18: Second, I am not convinced about the biological relevance of the findings. The authors speculate in the discussion section that the VarA-pathway could modulate cell shape under physiological conditions, however, I am not sure such conditions exist given that VarA activity is not only controlled by VarS, but rather integrates information from multiple histidine kinases. I have a several additional comments, which I listed below.

Authors’ reply #18: We regret the referee’s personal opinion that our findings might not be of biological relevance.

However, we respectfully disagree with the notion that such physiological conditions would never occur just because several histidine kinases can feed into VarA signaling. Indeed, as discussed above under authors’ reply #2, the (VarS/)VarA-CsrA pathway is highly conserved in Vibrio species and other proteobacteria. Yet, for V. choleraemost studies have focused on virulence-inducing conditions, while the species’ environmental lifestyle has been vastly neglected in the past. Indeed, even our own work on several V. cholerae’s phenotypes (natural competence for transformation [Meibom*, Blokesch* et al., 2005, Science]; T6SS production in pandemic strains [Borgeaud et al., 2015, Science]; pilus-mediated aggregation [Adams et al., 2019, Nat. Microbiol]; etc.) has remained unknown for decades, given the chitin dependency for their induction – a substrate not commonly studied in lab settings. Interestingly, several of these findings have also initially been considered biologically irrelevant, “artifacts”, or even non-reproducible by reviewers in the past, while nowadays all these phenotypes have been extensively reproduced by many different research groups and are well accepted in the field as biologically highly relevant. Thus, we truly believe that one should be open to new phenotypes and, as reviewer #1 rightfully acknowledged, consider that “this manuscript builds the foundation to further test this question.”

It should also be noted that the (VarS/)VarA-CsrA system has been studied for >15 years based on deletion strains, as we did in this study, and the readouts of these studies have been well accepted in the field without provision of the physiological conditions that would mimic the situation of these knock-out strains.

Collectively, we truly believe that there are still many understudied physiological conditions for V. cholerae; however, finding the right conditions could take years and is therefore beyond the scope of the current study.

---

Major points:

Comment #19: - Figs. 1 and S1: I think it is interesting that the varS mutant strain does not share the cell shape phenotype with the varA mutant. As pointed out by the authors, this result indicates that varA activity is controlled by another histidine kinase. While I believe it might be beyond the scope of this manuscript to determine which other histidine kinases signal towards VarA, I think it would be useful to measure and compare CsrB/C/D levels in WT, DvarA, and DvarS cells.

Authors’ reply #19: Thanks for this comment. We fully agree that finding the secondary histidine kinase is beyond the scope of this study. In the revised manuscript, we will, however, include the CsrB/C/D levels of the WT, ∆varA, and ∆varS strains, as suggested by the reviewer.

Comment #20: - Figs. 1, S1, and 4C: The regulatory logic implied by these results suggest that deletion of varA results in reduced CsrB/C/D levels, which in turn leads to higher activity of CsrA in the cell. Thus, it would be useful to test if A) over-production of CsrB, CsrC, or CsrC can rescue the phenotype of an varA mutant and if B) combined deletion of csrB/C/D will phenocopy the mutation of varA.

Authors’ reply #20: These are also very good suggestions. Notably, this has been done in the past for the V. cholerae system (Lenz et al., 2005, Mol. Microbiol.). Indeed, after receiving the reviewer’s comment, we immediately asked these authors to kindly share their csrB,C,D overproduction plasmids as well as the triple knock-out strain with us (as all of these constructs have been extensively verified in their published work). Unfortunately, we are not entirely sure whether it will be possible to receive these constructs any time soon, as we were told that such shipment might take >1 year (though, upon further discussion, this timeframe was lowered to ~3 months). If we manage to receive these published constructs in a reasonable timeframe, we will certainly perform the suggested experiments.

---

Comment #21: - Figs. 4B & 5C: I was somewhat surprised by these results. Given that AspA overproduction is suggested to cause cell shape abnormalities in the varA mutant, I would have expected additional transposon insertion in aspA. The fact that mutations only occurred in csrA could indicate that additional (CsrA-controlled) could be involved in the phenotype.

Authors’ reply #21: See authors’ reply #5 above, where we explain that the ∆varA∆aspA strain has a slight growth disadvantage. For this reason, any ∆varA-aspA::Tn transposon mutant would likely be outcompeted by the csrAsuppressor mutants in our genetic screen.

---

Minor points:

Comment #22: - Throughout study: italicize gene names

Authors’ reply #22: Gene names have been italicized in the initial manuscript; however, strain names - such as strain ∆varA – haven’t been italicized, in accordance with several of our previous publications. However, for the revision, we will italicize all strain names to accommodate the reviewer’s request.

Comment #23: - Figs. 1C and S3C: please quantify the results of these western blots and indicate how many replicates were performed.

Authors’ reply #23: We apologize for this oversight – indeed, all Western Blot were performed three independent times, as is good scientific practice. We will add this information into the methods section of the revised manuscript.

Concerning the quantification: the primary claim of these figures is that the HapR protein is still produced in the ∆varA mutant in the different pandemic strain backgrounds, while the luxO-mutated strains have a significant defect in HapR production (as we have previously reported; Stutzmann and Blokesch, 2016, mSphere). These data are qualitatively very clear in the Western Blots and can be considered as “black or white” results.

However, for the revision we will quantify the bands’ intensities of the performed Western blots and provide these quantitative data, as requested by the reviewer.

Comment #24: - Fig. 5A and B: in order to properly quantify the levels of AspA in the cell (and link them to CsrA activity in the transposon mutants), I think it would be better to add a tag to the chromosomal aspA gene and perform quantitative Western blot analysis.

Authors’ reply #24: We respectfully disagree. Firstly, this is not a subtle difference that we observe in these cell lysates/the corresponding stained gel bands but a rather strong difference when WT is compared to the mutants (see, for instance, a copy of panel 5B below as a kind reminder). Together with the genetic experiments that follow afterwards, the link seems very solid to us. Secondly, adding a tag could change the proteins abundance (change of the protein’s production/degradation dynamics) and/or activity, which could cause more confusion than needed (and a loss of the spherical cell shape if the enzyme loses its activity through the tagging).

However, as mentioned above under authors’ reply#12, we meanwhile observed that the aspA mRNA levels also increase in the ∆varA mutant. Thus, we will provide qRT-PCR data in the revised manuscript (and discuss all options on how the increase of the transcript and subsequently the protein might be caused, as mentioned above under authors’ reply #12), which we truly believe will fulfill the reviewer’s request for quantification.

[figure not shown in online form]

Reviewer #2 (Significance (Required)):

Comment #25: I think this manuscript starts with an interesting observation, which is that varA mutant cells of V. cholerae display an aberrant cell shape. The manuscript also provides several important findings explaining the molecular basis of this phenotype.

Authors’ reply #25: Once again, we thank the reviewer for the kind words.

Comment #26: However, as pointed out in my report, I think the manuscript is yet incomplete in connecting this information to identify the underlying regulatory mechanism.

Authors’ reply #26: As mentioned above, the focus of this study was never on the elucidation of the entire regulatory pathway. Instead, we aimed at deciphering the molecular mechanism behind an observed phenotype - that is, the cell wall modification in the varA-deficient strain that leads to the bacterium’s spherical shape, which can be restored to the WT Vibrio shape by peptidoglycan precursor cross-feeding from neighboring cells – followed by the identification of several regulators and enzymes that trigger these phenotypes. Overall, we consider this a very complete study. However, as mentioned above, we will certainly discuss in the revised manuscript that the step between CsrA and AspA could be indirect in V. cholerae, in contrast to what was experimentally shown for Salmonella and E. coli.

**Referee Cross-commenting**

Comment #27: As pointed out in my review, I think this manuscript is well written and easy to follow. However, I agree with reviewer #1 that the underlying phenotype is most likely an artifact, which limits the biological relevance of this study. In addition, I am missing the molecular mechanism that connects CsrA with AspA production in V. cholerae.

Authors’ reply #27: See authors’ reply #18 above. We disagree that there is any strong indication that the observed phenotype is an artifact. Given that it is state-of-the-art to study TCS by deleting their genes, our study isn’t any more prone to being an artifact than any other study on TCSs.

We truly believe that it is also important to not take reviewer #1’s comment out of context by stating “I agree with reviewer #1 that the underlying phenotype is most likely an artifact”. Indeed, he/she provided a rather encouraging statement in which he/she mentions the possibility of an artifact but also clearly states that this study is interesting and builds the foundation to further investigate the newly observed phenotype(s): “I think the finding is quite interesting, even though it is not clear to me if this observed cell morphology has a biological function or if it is an artifact of completly removing VarA. However, this manuscript builds the foundation to further test this question.”

Moreover, whether there is a direct (as in Salmonella and E. coli) or indirect connection between CsrA and AspA production is not a key aspect of the current study, as discussed above.

• *

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

Learn more at Review Commons

---

Referee #2

Evidence, reproducibility and clarity

In this manuscript, Rocha et al studied the effect of the VarA response regulator on cell shape of Vibrio cholerae. VarA is part of a two-component system that also includes the histidine kinase VarS. It has previously been shown that VarA activates the expression of three redundantly acting regulatory RNAs called CsrB, CsrC, and CsrD. All three Csr RNAs share the same regulatory principle, which is to sequester the activity of the RNA-binding protein CsrA. CsrA in turn can bind hundreds of mRNA species in the cell, which in the majority of cases results in reduced translation of these mRNAs (in addition to various other modes of action that have been reported). Here, the authors discovered that deletion of the varA gene results in an abnormal, spherical cell shape in stationary phase grown V. cholerae cells. Biochemical analysis revealed an unusual peptidoglycan composition in varA-deficient cells, showing increased levels of dipeptides, a reduction of tetrapeptides, and an overall decrease in peptide-cross linkage. Interestingly, the varA phenotype was complemented by the addition of conditioned medium from wild-type cells, which are likely to provide peptidoglycan building blocks in trans. The authors further discover that varA-deficiency results in AspA over-production, which could be linked to the activity of CsrA. The authors speculate that high AspA levels deplete the cell of aspartate, which is required to produce peptidoglycan precursors. The manuscript is interesting, well-written, and the rationale of the experiments is easy to follow. However, I have two major points of criticism, which reduce my enthusiasm for this work. First, the molecular pathways that links varA-deficiency to increased AspA levels is incomplete: please clarify how CsrA activates AspA levels and if this phenotype is linked to direct binding of CsrA to the aspA mRNA and if so how is activation is achieved at the molecular level. Second, I am not convinced about the biological relevance of the findings. The authors speculate in the discussion section that the VarA-pathway could modulate cell shape under physiological conditions, however, I am not sure such conditions exist given that VarA activity is not only controlled by VarS, but rather integrates information from multiple histidine kinases. I have a several additional comments, which I listed below.

Major points:

• Figs. 1 and S1: I think it is interesting that the varS mutant strain does not share the cell shape phenotype with the varA mutant. As pointed out by the authors, this result indicates that varA activity is controlled by another histidine kinase. While I believe it might be beyond the scope of this manuscript to determine which other histidine kinases signal towards VarA, I think it would be useful to measure and compare CsrB/C/D levels in WT, DvarA, and DvarS cells.
• Figs. 1, S1, and 4C: The regulatory logic implied by these results suggest that deletion of varA results in reduced CsrB/C/D levels, which in turn leads to higher activity of CsrA in the cell. Thus, it would be useful to test if A) over-production of CsrB, CsrC, or CsrC can rescue the phenotype of an varA mutant and if B) combined deletion of csrB/C/D will phenocopy the mutation of varA.
• Figs. 4B & 5C: I was somewhat surprised by these results. Given that AspA overproduction is suggested to cause cell shape abnormalities in the varA mutant, I would have expected additional transposon insertion in aspA. The fact that mutations only occurred in csrA could indicate that additional (CsrA-controlled) could be involved in the phenotype.

Minor points:

• Throughout study: italicize gene name
• Figs. 1C and S3C: please quantify the results of these western blots and indicate how many replicates were performed.
• Fig. 5A and B: in order to properly quantify the levels of AspA in the cell (and link them to CsrA activity in the transposon mutants), I think it would be better to add a tag to the chromosomal aspA gene and perform quantitative Western blot analysis.

Significance

I think this manuscript starts with an interesting observation, which is that varA mutant cells of V. cholerae display an aberrant cell shape. The manuscript also provides several important findings explaining the molecular basis of this phenotype. However, as pointed out in my report, I think the manuscript is yet incomplete in connecting this information to identify the underlying regulatory mechanism.

Referee Cross-commenting

As pointed out in my review, I think this manuscript is well written and easy to follow. However, I agree with reviewer #1 that the underlying phenotype is most likely an artifact, which limits the biological relevance of this study. In addition, I am missing the molecular mechanism that connects CsrA with AspA production in V. cholerae.

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 follow up on an interesting finding that varA null mutants of V. cholerae form spherical cells in stationary phase. The authors determine that this cell rounding is due to weakening of the cell wall via less production of tetrapeptide cross links. Mutation of the regulator csrA and the enzyme aspA lead to a model in which a varA mutant cell lacks aspartate leading to low cross-linked cell wall that is unable to hold the typical curved V. cholerae shape. The data are robust, and the manuscript is clearly written. I think the finding is quite interesting, even though it is not clear to me if this observed cell morphology has a biological function or if it is an artifact of completly removing VarA. However, this manuscript builds the foundation to further test this question. The data all support the conclusions, but I do think the authors could have really confirmed their model by connecting mutations in csrA and aspA to restoration of high cross-linked cell well similar to the WT strain as done in Fig. 2. As it stands, this is still somewhat hypothetical and has not been directly demonstrated, although I do think their model is correct and these experiments will be conformation of it. I also have a few other suggestions to improve the manuscript, but in sum I think it is a well-done research study that will be interesting to research in V. cholerae and other gamma proteobacteria.

Major comments:

1. The enrichment for suppressors is very creative and connected the varA impact on cell morphology to misregulation of csrA as 10/10 mutants were ultimately linked to this gene. However, insertion in aspA should also suppress this phenotype, and I am curious why this gene was not identified in the transposon suppressor screen.
2. The authors should complement at least one of their varA/csrA mutants with csrA.
3. The changes in cell wall structure are not directly connected to the genetic identification of csrA and aspA. Yes, I agree their model makes sense, but to really nail it down they should analyze the cell wall composition in the varA/csrA and varA/aspA double mutants and show it returns to WT levels of crosslinking.
4. Does deletion of aspA in the WT or varA mutant impact the growth rate?

Minor comments

1. Are the round cells able to divide? The data in Fig. S2 would suggest they can based on the increase in CFUs from hour 6 to hour 8, but the authors never comment on this point and it might be worth addressing in the discussion.
2. Fig. S2-Why does the OD600 increase from 8 to 24 hours but the CFUs decrease in the varA mutant?
3. Lines 309-A little bit more detail here would help the reader. Are the authors examining whole cell lysates or lysates from specific cellular components? I am actually very surprised this worked as there are so many proteins in crude cell lysates.
4. Lines 320-321-I don't think there is evidence that CsrA enhances aspA RNA translation, merely that CsrA enhances AspA protein production. It is likely through increasing translation, but this cannot be concluded without direct evidence.
5. Line 348-350-I do not understand the logic of this sentence stating that the "..until now, the signal that abrogates VarA phosphorylation..." as this manuscript does not contribute to our understanding of the VarS signal.
6. I am curious if the total volume of the round versus curved cells is constant at 20 hours. This should be easy to determine using ImageJ and worth reporting.

Significance

Understanding changes to cell morphology and their biological implications is a growing area of microbiology. This study makes a new contribution to this area by demonstrating a round, spherical form of V. cholerae that is driven by alterations to the cell that decrease cell-wall cross linking.

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 rebuttal 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

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

**Summary**

The authors have performed highly quantitative analyses of GPCR signaling to reveal heterogenous ERK and Akt activation patterns by using kinase translocation reporters. Using a massive number of single-cell imaging data, the authors show heterogeneous responses to GPCR agonists in the absence or presence of inhibitors. By cluster analysis, the responses of ERK and Akt were classified into eight and three patterns. This paper is clearly written with sufficient information for the reproducibility. However, the conclusion may not be necessarily supported by the provided data as described below.

**Major comments:**

This work has been well done in an organized way and adds new insight into the regulation of protein kinases by GPCRs. The conclusion will be of great interest in the field of single-cell signal dynamics and quantitative biology. On a bit negative note, considering the complexity of the downstream of GPCRs, some of the conclusions may need revision. *

We thank the reviewer for the evaluation and for raising a number of comments that have helped us to strengthen the manuscript and that will be addressed below.

1. The conclusion of the title that "Heterogeneity and dynamics of ERK/Akt activation by GPCR depend on the activated heterotrimeric G proteins," may not be supported by the data. The authors compared just one pair each of GPCR and ligand. The heterogeneity may come from the nature of the ligand or the characteristics of the single clone chosen for this study. The title may suggest that the heterogeneity depends only on the G-protein (although that is not what the title says). Instead, we mean that G-proteins play a role in the heterogeneity, as we infer from the experiments with the G-protein inhibitors. If the reviewer feels strongly about this, we are open to changing the title, for instance to:

“Kinase translocation reporters reveal the single cell heterogeneity and dynamics of ERK and Akt activation by G protein-coupled receptors”

• The obvious question is that why the authors did not analyze the correlation between ERK and Akt activity more extensively. Cell Profiler will be able to extract multiple cellular features. Linking the heterogeneous signals to cellular features will benefit readers in the broad cell biology field. If the authors wish to write another paper with that data, it should be at least discussed. *

We agree that we can add more information on the correlation between ERK and Akt activity and we have added a plot that shows the co-incidence of the ERK and Akt clusters. This is now panel C of figure 8. We have no wish of writing another paper and we have made the data and code available, so anyone can do a more detailed analysis if desired.

We appreciate the suggestion to correlate activities with cellular features, such as cell area and shape. However, in our analysis we use nuclear fluorescence to segment the nuclear and cytoplasmic fluorescence (as generally done in studies that use KTRs). Therefore, the information on cellular features is not readily available. Such analysis would require a marker for the cytoplasm or membrane (or yet another image analysis procedure).

Another apparent flaw of this work is that YM was not challenged to UK-stimulated cells. The authors probably assumed lack of effect. Nevertheless, I believe it is required to show. Or, remove the PTx data from the Histamine-stimulated cell data.

We agree that this is valuable data to include. Unfortunately, this experiment was done in a slightly different condition than the other experiments (different spacing of the time intervals) and we initially skipped the data for these reasons. After careful examination of the data, we have decided to include these data (added to figures 3 & 5).

We note that we still miss the data from the YM+PTx data for UK and we have currently no way to carry out these experiments (mainly due to lack of funding). In our opinion, the absence of this data is not critical for the interpretation of the results. We prefer to show the YM+PTx data for the other two conditions.

The most interesting response is that of S1P. ERK is biphasically activated. Combined inhibition of Gq and Gi failed to suppress ERK activity. It may be discussed why the biphasic activation pattern was not identified by the classification.

We think that the biphasic activation pattern is reflected by cluster 7 and 8 and we now mention this in the text: “The biphasic ERK activation pattern, which is specific for stimulation with S1P are reflected by cluster 7 and 8.”

For clarity, we now added the dynamics for each cluster to figure 9.

*The authors argue that the brightness of the KTR reporter was not correlated with the dynamic range of ERK or Akt reporter (Supplementary Figure 3), but it is not clear. I had an impression that ERK-KTR brightness (Supplementary Figure 3A) has a slightly negative correlation with "maximum change in CN ratio" (Supplementary Figure 3B) (e.g., A6>B3>B5 in brightness and A6

We thank the reviewer for the suggestion and have now added this data to supplemental figure 3 as panel C.

The authors have shown cluster analyses for the temporal patterns in kinase activations. However, the only difference of cluster 3 and 5 (Figure 7) seem to be amplitude. The authors have also shown the amplitude is dependent on the dose of the activators, which together makes it difficult to see the biological meaning of discriminating the two patterns in comparing different agonists, e.g., Histamine, UK, and S1P. The authors should discuss their views on how the clustering analyses will benefit biological interpretations together with possible limitations.

This is a valid point, and it is a consequence of clustering method. We have added text to the discussion to explain our view: “The clustering is a powerful method for the detection of patterns and simplification of large amounts of data. Yet, it should be realized that clustering is mathematical procedure that is not necessarily reflecting the biological processes. One example is the graded response of ERK and Akt activities to ligands, whereas cells are grouped in weak, middle and strong responders. This may be solved by developing and using clustering methods that take the underlying biological processes into account.”

Considering the importance of the content, the supplemental note 2 may be included in the main text.

We appreciate this suggestion, and we have incorporated supplemental note 2 in the main text.

\*Minor comments:**

1. The authors should clarify the cell type they used (HeLa cells) in the main text and figure legends. *

This information is now indicated in the first paragraph of the results section and in the legend of figure1.

Supplementary note1: The data-not-shown data (no correlation of KTR expression and its response to serum) should be very informative for the readers. The data should be shown as an independent supplementary figure.

This relates to major point 5 and we agree that this is valuable. The data of the expression and the maximum response has been added to supplementary figure 3 as panel C.

Supplementary Figure S2: The authors should clarify this image processing is about background subtraction. Also, the authors should clearly note "rolling ball with a radius of 70 pixels" is about an ImageJ function, "Subtract Background".

We added text to highlight that the processing is a background subtraction and noise reduction. We added text to explain it is a FIJI function.

• 1. Supplementary Figure S5: Figure labels are "A, A, B, B" not "A, B, C, D". Also the top two figures are lacking Y axis labels. *

Thanks for pointing this out. We the labels are corrected.

Page6 (top): The authors should mention the description is about Supplementary Figure S5 (UK) and Supplementary Figure S6 (S1P).

This is an accidental omission, it is corrected.

Figure 3: the figures are lacking x-axis labels (probably uM, nM and pM from left).

Well spotted, this is fixed by adding the units to the labels for each ligand.

Values in tables: The significant figure must be 2, at best. This should be consistent throughout the text. For example, "The EC50 values for histamine, S1P and UK were respectively 0.3 μM, 63.7 nM and 2.5 pM." This is somewhat awkward.

This has been fixed in the text and in the table.

Page 7, the first paragraph: No comments on S1P!

We added our observation that: “The response to S1P is hardly affected by YM, but the amplitude is reduced by PTx.”

Fig. 3: 100 mM must read as 100 micromolar.

We do not understand this comment, but the units of figure 3 are now corrected (see also point 6).

• Fig. 9: Concentration unit is missing.*

Thanks for pointing this out, units are added.

• Page 11, line 4: EKR should read as ERK. *

Fixed

• Page 13: "So far, only a couple of studies looked into kinase activation by GPCRs and these studies used overexpressed receptors [32,33]." Please describe precisely. Protein kinase activation by GPCR has been studied more than 20 years. Why are these two recent papers cited here? *

We updated the text to explain that: “So far, only a couple of studies looked into kinase activation by GPCRs in single cells with KTRs and these studies used overexpressed receptors”.

"This is in marked contrast to other fluorescent biosensors that typically require an overexpressed receptor for robust responses [34]." Following words should be included in the end: "in our hands".

We’ve included the suggested line.

• "Histamine is reported to predominantly activate Gq in HeLa cells [36] and UK activates Gi [37]." Describe the name of receptors for the better understanding. *

We added names: “Histamine is reported to predominantly activate Gq in HeLa cells by the histamine H1 receptor [36] and UK activates Gi by α2-adrenergic receptors [37]”

• "S1P can activate a number of different GPCRs, all known to be expressed by HeLa cells [24]." Why is this paper chosen? The authors can easily find RNA-Seq data, if they wish to see the expression level. The cited paper did not scrutinize the S1P receptors expressed in HeLa cells. *

The S1PR levels are scrutinized in the cited paper, but it is ‘hidden’ in the supplemental figure S4A. We will clarify this and explicitly mention this supplemental figure: “The situation for S1P is different. S1P can activate a number of different GPCRs, all known to be expressed by HeLa cells as shown in the supplemental figure S4A of [24]”

*Reviewer #1 (Significance (Required)):

The authors used biosensors for ERK and Akt to examine the kinetics of activation by GPCR ligands. Technical advancement is in the massive analysis method and cluster analysis. This is an important direction for the quantitative biology. GPCR signaling is complex because of multiple receptors coupled with different G proteins. The simple ones such as histamine receptor and alpha2-adrenergic receptor can be easily analyzed as shown in this study. However, there are many S1P receptors, which make the interpretation difficult. If the authors could have shown interesting proposal on this data, the paper may interest many researchers in the field of cell biology and systems biology.

Expertise: Cell biology, signal transduction of protein kinases, fluorescence microscopy.

**Referee Cross-commenting**

1. I agree with the other two reviewers in that immunoblotting data is required to show the efficiency of P2A cleavage.
2. All reviewers think it looks strange that the authors did not show UK + YM data.

3. Showing the dynamic range of the biosensors will reinforce the data as Reviewer #3 states. ERK-KTR is quite sensitive and can be easily saturated. Ideally, the ratio of pERK vs ERK can be quantified by the different mobility in SDS-PAGE. But, I do not know how we can do it for Akt.

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

   **Summary**

   In this paper Chavez-Abiega and colleagues investigate the dynamics of ERK and Akt activity downstream of several G protein-couples receptors (GPCRs). Using drugs to block specific G-proteins, they probe the activation of ERK/Akt by different heterotrimeric G proteins with fluorescent biosensors at the single cell resolution. Main finding is that ERK/AKT can be activated by different G-proteins, depending on the receptor coupling to the G-protein subclass, and that the ERK/AKT dynamics for S1P are specifically heterogeneous. Moreover, it seems that the AKT signaling response is very similar to ERK after GPCR stimulation.

   **Major points:**

   1) For this paper, the authors produced a new construct to express simultaneously the nuclear marker, the Akt and the ERK biosensors. The tree parts are connected by P2A peptides that determine their separation. Although, the biosensors are based on existing ones, the connection between them by P2A might create artifacts if the separation of the two parts is incomplete. For that, important controls are missing, such as treatment with an ERK and an Akt inhibitor. If the two parts are well separated the inhibitors should block the cytosol translocation of one of the two components and not of the other. This control is also important to check if in HeLa cells the Akt biosensors is not phosphorylated by ERK as well, as described in other reports. Alternatively, P2A separation can be quantified on a protein blot. *

We agree that it is important to establish that the P2A sequence results in separation of the reporters. There are several observations that support our notion that the separation is efficient. First, we have been using the 2A-like sequences for over a decade in HeLa cells (first paper: doi:10.1038/nmeth.1415) and we have never encountered situations where the cleavage was problematic. Second, the distribution in signal of the nuclear Scarlet probe differs substantially from that of the mTurquoise2 and the mNeonGreen probe. Third, the dynamics of the ERK-KTR and Akt-KTR are different. Fourth, we have included new data with an ERK inhibitor, showing that the Akt-KTR responds independently of the ERK-KTR (figure S5). We have also added text to explain this: “Next, we examined the effect of the MEK inhibitor PD 0325901. Pre-incubation with the inhibitor for 20 minutes blocked the response of the ERK-KTR to FBS, but not that of Akt-KTR (Supplemental Figure S5). This supports previous observations [14] [15] that the P2A effectively separates the different components, since the Akt-KTR and ERK-KTR show independent relocation patterns.”

This latter point is also supported by the co-incidence plot of the ERK versus Akt clusters (figure 8C) showing that the probes act independently (which is the main reason for using this strategy).

Although any of the aforementioned points cannot exclude that a small fraction of the probe remains fused, we think that this potential issue is far outweighed by the benefits of the use of 2A peptides.

2) The description of ERK and Akt should be reported in a more uniform way, such as using the same representations for both (e.g. the equivalent of figure 2 for Akt is missing) or the same number of clusters.

We choose to concentrate first on ERK activity, that is why a similar plot for Akt activation is not shown. However, the Akt responses are detailed in figure 4 and supplemental figures S5 and S7.

For the cluster analysis, we looked into the optimal number of clusters (as explained in Supplemental note S2). This number differs for ERK and Akt, since the complexity of the responses is different. We move supplemental note 2 to the main text, which also clarifies the different number of clusters that we used for the analysis.

3) Figure 3 & Figure 5: It seems that the YM and YM+PTx data for the UK 14304 data is missing. This would be an interesting addition to the manuscript, and it is easy to add. A similar analysis for the Akt sensor is missing in figure 3 and should be added for consistency. Figure 4 shows data for Akt, but as timeseries and only for Histamine. See point 2, it would benefit the reader greatly if ERK and AKT are presented in a more uniform and complete fashion throughout the manuscript.

We agree that it is valuable to add data for UK with YM. This data has been added, see also reply to reviewer 1, major point 3

As for the Akt data, the response was largely similar albeit with less complexity and a lower amplitude. This is the reason to focus on ERK and this is explained in the discussion: “Therefore, the measurement of Akt does not add information. Moreover, the Akt response had a relatively poor amplitude.”

4) In the results text of figure 4, the authors state that "...as shown in Figure 4C-D, which is in line with the effect of histamine on ERK.". It is unclear what the authors mean with this statement, the effects of single/double inhibition of Histamine stimulation on ERK are not quantified or discussed. Both responses can be quantified more carefully and compared.

We agree that this is poorly formulated, and we rephrase it to make it clearer: “Inhibition of Gq (figure 4C) decreases the maximum activity up to ~70%, and simultaneous inhibition of Gq and Gi causes a decrease of the responses up to ~90%, as shown in Figure 4D. These Akt amplitudes and effects of inhibitors are largely similar to those observed for ERK.”

5) This paper would benefit from a mechanistic investigation. For instance, the authors could investigate the pathways that lead to the generation of the pulse of ERK and Akt. These (preliminary) results presented call for deeper investigation into the signaling pathway from Gai and Gaq to ERK and AKT, and the authors are in a great position to probe this. One simple approach is to explore the upstream pathway, such as the MAPK cascade, PI3K, RTKs by means of inhibitors.

We agree that there is much that can be done with the KTR technology. To this end, we deposit the probe and make all our data analysis methods available. We hope that others will benefit from our efforts and use the tools for mechanistic studies. 6) Since different G-proteins seem to elicit similar responses on ERK and especially for Akt, it is likely a B-arrestin / beta-gamma subunit mediated mechanism? It would be interesting to hear what the authors think of this, did they investigate/consider this possibility? E.g. Perhaps blocking RTK signaling / B-arrestin signaling would reduce heterogeneity?

We appreciate this suggestion and have added a statement to the discussion: “Based on our data, we cannot exclude that beta-arrestin or RTKs play a role in the activation of ERK and Akt. To study the role of non-classical routes to ERK activation, inhibitor studies, or probes that interrogate these processes would be useful.”

7) The authors should take a serious effort to summarize the data in the figures better. Many plots that can be merged/presented in a more concise way, which would improve the readability of the manuscript greatly.

We will take care to improve the data visualization during the revision. We will address any specific points that are raised.

\*Minor points:**

1) The authors should spell out in the legend of each figure if they are representing the absolute C/N or the normalized C/N *

Thanks for pointing this out. We added this information to the legends and it is also written in the materials and methods: "data was normalized by subtracting the average of two time points prior to stimulation (usually the 5th and 6th time point) from every data point."

2) In Figure 2 the authors should show the control with no stimulus. Also would be informative to inform the reader about the stimulation protocol used, or indicate the stimulation time and length in the figure.

We have added the no stimulus control and added the information to the legend.

3) Figure 3: This figure would benefit from a different presentation of the data, it is currently confusing. E.g. Average curves per drug condition in a single graph would present the point the authors make more clear and concise, and this single cell overview can be moved to supplements.

Our main focus is on single cell analysis and we think that the current plots convey the message in a clear and transparent fashion. It is in line with the recently proposed idea of “superplot” (https://doi.org/10.1083/jcb.202001064). We also provide scripts and data, enabling anyone to replot the data if that is desired.

4) Figure 4 legend states "CN ERK" and "ERK C/N", but is depicting only Akt responses? Only in 4c the axes are labeled, this together is very confusing.

Thanks for pointing this out. This is corrected

*5) Figure 5 is missing the controls with ERK and Akt inhibitors, to show the loss of correlation between the AUC of the two

*We have included data with a MEK inhibitor (new supplemental figure S5) to demonstrate the specificity of the probe and it also demonstrates that Akt can be independently activated

6) Figure 6, the presumed lack of correlation between baseline activity and response should be confirmed statistically.

We have improved the presentation of figure 6. We now show only the maximal response and how this varies between conditions. It is evident from the graphical representation that the curves are similar for the different start ratios. We feel that the use of statistics is not necessary here.

7) It seems that in S1P treated cells there is a second oscillation in ERK activity well visible in figure 2 and also in S10. Could the authors comment on that?

We add text to the discussion to address this: “We observed that activation of endogenous S1P receptors resulted in a strong, but highly heterogeneous ERK-KTR response, with two peaks in a population of cells.” and “When PTx is present, the biphasic response is abolished and the first peak of activation is reduced, suggesting that the initial response is due to Gi signaling.”

*8) In the abstract it is unclear what authors mean with "UK".

*Changed to brimonidine

9) Figure 9, it would be helpful to visually repeat the typical curve of the different clusters here, to guide the reader.

This is a good suggestion and we have added the typical curves for the different clusters to the plot.

10) The observed heterogeneity in responses might be related to different cell cycle stages, did the authors investigated/consider this possibility (e.g. with a cell cycle biosensor)?

This is a very valid comment. We do consider its importance, but we did not investigate the effects of cell cycle.

*Reviewer #2 (Significance (Required)):

The paper describes with high accuracy the dynamics of ERK and Akt biosensors downstream of several GPCRs.

However, it feels like this is a preliminary report that leaves many important questions still open. It does not provide mechanistic insight and doesn't fully exploit the potential of single-cell technologies. The authors have the tools to investigate several important questions that are left open in the manuscript (e.g. connection Gaq/Gai to ERK/AKT, B-arrestin/betagamma involvement). Moreover, some important controls are missing. The authors should also consider the data presentation in the figures, to improve readability and interpretation of the manuscript.

Properly revised, would be of interest for a broad audience in cell biology, specifically GPCR and RTK signaling fields.

Expertise in cell biology, gpcr and rtk signaling, fluorescent biosensors.

**Referee Cross-commenting**

I agree with the assessments by the other reviewers.

Indeed showing the dynamic range of the biosensors, as Reviewer #3 states, would strengthen the manuscript and put the S1P response heterogeneity in context.

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

This manuscript uses a live-cell biosensor approach to examine the activity kinetics of the ERK and Akt kinases in response to different GPCR ligands. The paper provides a detailed description of the development of a HeLa reporter cell line that expresses both Akt and ERK biosensors, along with a nuclear marker for use in cell tracking. The authors then catalog the individual responses from thousands of cells to three GPCR ligands. Individual cells show strong correlation in stimulated ERK and Akt activity. Using inhibitors for Gq and Gi proteins, it is shown that ERK and Akt activities are dependent on different G proteins. The authors also show that the heterogeneous responses within each population can be decomposed into several clusters representing similar dynamic behaviors; the frequencies of these clusters increase or decrease depending on treatments.

Overall, this is a well documented extension of an existing biosensor approach to examine GPCR signaling, and the approach is clearly described. There are however, some control experiments that are essential to support the conclusions.

**Major comments:**

1. The maximal responses of ERK and Akt biosensors in the selected cell clone are not adequately shown. Although FBS responsiveness is used as a validation and selection criterion, it would be much more informative to show the distribution of single-cell responses for defined activators of ERK and Akt, such as EGF and IGF-1, respectively. Without seeing the variability in these responses, it is difficult to put the heterogeneity observed in GPCR responses into context. *

The FBS is used as a (crude) way to examine responsiveness of the clones. We understand that treatment of the cells with growth factors would add more data and therefore more information to the manuscript. However, the main aim of the study is to examine whether KTR technology can be used to study endogenous GPCR signaling. It is clear that the answer is positive. Next, we asked whether we could detect differences for different GPCRs and that was the focus of this study. It is unclear how studies with EGF would add new information to our observations.

It is not clear whether the basal activity for the biosensors represents actual activity or simply the measurement floor. This should be established by using saturating treatment inhibitors for ERK and Akt to determine the biosensor readings in the absence of any activity. Ideally, an approach such as the one shown by Ponsioen et al. (PMID: 33795873) should be used to determine the dynamic range of the sensors.

We studied the basal levels and the effect of serum. We found that the basal levels are reduced by replacing the growth medium with serum free medium. The reduction in C/N ratio reaches a plateau after ~ 2hours of replacing the medium. This data is added as supplemental figure S4. Therefore, we have performed all experiments 2 hours after replacing the growth medium with serum free imaging medium.

Because the biosensors are separated by self-cleaving peptides, there is the potential that incomplete cleavage could complicate the results. Cleavage efficiency should be assessed by western blot or an equivalent method.

We agree that it is important to establish that the P2A sequence results in separation of the reporters. There are several observations that support our notion that the separation is efficient. First, we have been using the 2A-like sequences for over a decade in HeLa cells (first paper: doi:10.1038/nmeth.1415) and we have never encountered situations where the cleavage was problematic. Second, the distribution in signal of the nuclear Scarlet probe differs substantially from that of the mTurquoise2 and mNeonGreen probe. Third, the dynamics of the ERK-KTR and Akt-KTR are different. Fourth, we have included new data with an ERK inhibitor, showing that the Akt-KTR responds independently of the ERK-KTR (figure S5). We have also added text to explain this: “Next, we examined the effect of the MEK inhibitor PD 0325901. Pre-incubation with the inhibitor for 20 minutes blocked the response of the ERK-KTR to FBS, but not that of Akt-KTR (Supplemental Figure S5). This supports previous observations [14] [15] that the P2A effectively separates the different components, since the Akt-KTR and ERK-KTR show independent relocation patterns.”

This latter point is also supported by the co-incidence plot of the ERK versus Akt clusters (figure 8C) showing that the probes act independently (which is the main reason for using this strategy).

Although any of the aforementioned points cannot exclude that a small fraction of the probe remains fused, we think that this potential issue is far outweighed by the benefits of the use of 2A peptides.

Ideally, an alternate method such as immunofluorescence for phosphorylated ERK/Akt or their substrates could be used in a subset of the conditions to validate the heterogeneity observed by the biosensors.

We thank the reviewer for this suggestion. Since we see a lot of variability in the dynamics, which cannot be addressed by immunofluorescence, we do not think this will experiment be valuable. Of note, GPCR activity is known to induce ERK activity in a dose-dependent manner on a population level as determined with immunolabeling methods and that is what we observe with the ERK KTR as well.

\*Minor comments:**

1. In the introduction, more rationale and background could be provided for the examination of GPCR-stimulated ERK and Akt activity. There is not much information provided on why this is an interesting question. Other than the involvement of beta arrestin and RTK transactivation, which are mentioned, what mechanisms are known to be involved? Also, the importance of ERK and Akt in cancer is brought up, but it is not made clear how this approach or results would connect specifically to a cancer model. *

We think that the connections between heterotrimeric G-proteins and kinase activity are not well established. Except for the classical Gq -> PKC -> ERK pathway, not so much is known and we add this to the discussion: “The classic downstream effector of Gq is PKC, which can activate ERK. On the other hand, it is not so clear how Gq would affect Akt. The molecular network that connects the activity of Gi with kinases also not so clear.”

*It would be helpful to provide some explanation for why the UK+YM and UK+YM+PTx data are not shown in figure 3

*

We agree that this is valuable data to include. Unfortunately, this experiment was done in a slightly different condition than the other experiments (different spacing of the time intervals) and we initially skipped the data for these reasons. After careful examination of the data, we have decided to include these data (added to figures 3 & 5).

We note that we still miss the data from the YM+PTx data for UK and we have currently no way to carry out these experiments (mainly due to lack of funding). We prefer to show the YM+PTx data for the other two conditions.

• In the Abstract figure, it is not clear which samples "Inhibitor" and "Agonist" are referring to. **

  *

Thanks for this comment. We will remove the visual abstract when the preprint is submitted to a journal.

* Reviewer #3 (Significance (Required)):

While similar reporter approaches have been used in a number of papers to examine growth factor signaling dynamics of ERK and Akt, this manuscript is the first I have seen to examine the responses of these kinases to different GPCR ligands. In doing so, it adds significantly to the growing body of literature on single-cell signaling responses. The mechanisms of ERK and Akt activation by GPCRs remain somewhat ambiguous, and the data reported here will be helpful in refining models for this signal transduction process. The findings that the GPCR ligands examined show different G protein dependencies than anticipated is an interesting facet, as is the observation that, while ERK and Akt are generally correlated, inhibition of Gi preferentially blocks S1P-induced ERK activity more so than Akt activity. However, the main findings of heterogeneity in signaling, and the observation of clusters that describe the different dynamic behaviors present within a population, are highly consistent with what has been shown in other systems. Overall, this study is a useful confirmation that GPCR signaling to ERK and Akt follows a similar pattern to other forms of stimulation.

**Referee Cross-commenting**

Regarding the dynamic range, I don't think it is necessary to do a western blot (though this would be nice) - I think it would be sufficient to show maximal activation using EGF/IGF and full suppression using MEK/ERK and Akt inhibitors. I also agree that all the points raised by the other reviewers. In particular, a deeper exploration and better visualization of the relationship between ERK and Akt would be very useful, as noted by both Reviewers #1 and #2.*

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 uses a live-cell biosensor approach to examine the activity kinetics of the ERK and Akt kinases in response to different GPCR ligands. The paper provides a detailed description of the development of a HeLa reporter cell line that expresses both Akt and ERK biosensors, along with a nuclear marker for use in cell tracking. The authors then catalog the individual responses from thousands of cells to three GPCR ligands. Individual cells show strong correlation in stimulated ERK and Akt activity. Using inhibitors for Gq and Gi proteins, it is shown that ERK and Akt activities are dependent on different G proteins. The authors also show that the heterogeneous responses within each population can be decomposed into several clusters representing similar dynamic behaviors; the frequencies of these clusters increase or decrease depending on treatments.

Overall, this is a well documented extension of an existing biosensor approach to examine GPCR signaling, and the approach is clearly described. There are however, some control experiments that are essential to support the conclusions.

Major comments:

1. The maximal responses of ERK and Akt biosensors in the selected cell clone are not adequately shown. Although FBS responsiveness is used as a validation and selection criterion, it would be much more informative to show the distribution of single-cell responses for defined activators of ERK and Akt, such as EGF and IGF-1, respectively. Without seeing the variability in these responses, it is difficult to put the heterogeneity observed in GPCR responses into context.
2. It is not clear whether the basal activity for the biosensors represents actual activity or simply the measurement floor. This should be established by using saturating treatment inhibitors for ERK and Akt to determine the biosensor readings in the absence of any activity. Ideally, an approach such as the one shown by Ponsioen et al. (PMID: 33795873) should be used to determine the dynamic range of the sensors.
3. Because the biosensors are separated by self-cleaving peptides, there is the potential that incomplete cleavage could complicate the results. Cleavage efficiency should be assessed by western blot or an equivalent method.
4. Ideally, an alternate method such as immunofluorescence for phosphorylated ERK/Akt or their substrates could be used in a subset of the conditions to validate the heterogeneity observed by the biosensors.

Minor comments:

1. In the introduction, more rationale and background could be provided for the examination of GPCR-stimulated ERK and Akt activity. There is not much information provided on why this is an interesting question. Other than the involvement of beta arrestin and RTK transactivation, which are mentioned, what mechanisms are known to be involved? Also, the importance of ERK and Akt in cancer is brought up, but it is not made clear how this approach or results would connect specifically to a cancer model.
2. It would be helpful to provide some explanation for why the UK+YM and UK+YM+PTx data are not shown in figure 3.
3. In the Abstract figure, it is not clear which samples "Inhibitor" and "Agonist" are referring to.

Significance

While similar reporter approaches have been used in a number of papers to examine growth factor signaling dynamics of ERK and Akt, this manuscript is the first I have seen to examine the responses of these kinases to different GPCR ligands. In doing so, it adds significantly to the growing body of literature on single-cell signaling responses. The mechanisms of ERK and Akt activation by GPCRs remain somewhat ambiguous, and the data reported here will be helpful in refining models for this signal transduction process. The findings that the GPCR ligands examined show different G protein dependencies than anticipated is an interesting facet, as is the observation that, while ERK and Akt are generally correlated, inhibition of Gi preferentially blocks S1P-induced ERK activity more so than Akt activity. However, the main findings of heterogeneity in signaling, and the observation of clusters that describe the different dynamic behaviors present within a population, are highly consistent with what has been shown in other systems. Overall, this study is a useful confirmation that GPCR signaling to ERK and Akt follows a similar pattern to other forms of stimulation.

Referee Cross-commenting

Regarding the dynamic range, I don't think it is necessary to do a western blot (though this would be nice) - I think it would be sufficient to show maximal activation using EGF/IGF and full suppression using MEK/ERK and Akt inhibitors. I also agree that all the points raised by the other reviewers. In particular, a deeper exploration and better visualization of the relationship between ERK and Akt would be very useful, as noted by both Reviewers #1 and #2.

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

Learn more at Review Commons

---

Referee #2

Evidence, reproducibility and clarity

Summary

In this paper Chavez-Abiega and colleagues investigate the dynamics of ERK and Akt activity downstream of several G protein-couples receptors (GPCRs). Using drugs to block specific G-proteins, they probe the activation of ERK/Akt by different heterotrimeric G proteins with fluorescent biosensors at the single cell resolution. Main finding is that ERK/AKT can be activated by different G-proteins, depending on the receptor coupling to the G-protein subclass, and that the ERK/AKT dynamics for S1P are specifically heterogeneous. Moreover, it seems that the AKT signaling response is very similar to ERK after GPCR stimulation.

Major points:

1) For this paper, the authors produced a new construct to express simultaneously the nuclear marker, the Akt and the ERK biosensors. The tree parts are connected by P2A peptides that determine their separation. Although, the biosensors are based on existing ones, the connection between them by P2A might create artifacts if the separation of the two parts is incomplete. For that, important controls are missing, such as treatment with an ERK and an Akt inhibitor. If the two parts are well separated the inhibitors should block the cytosol translocation of one of the two components and not of the other. This control is also important to check if in HeLa cells the Akt biosensors is not phosphorylated by ERK as well, as described in other reports. Alternatively, P2A separation can be quantified on a protein blot.

2) The description of ERK and Akt should be reported in a more uniform way, such as using the same representations for both (e.g. the equivalent of figure 2 for Akt is missing) or the same number of clusters.

3) Figure 3 & Figure 5: It seems that the YM and YM+PTx data for the UK 14304 data is missing. This would be an interesting addition to the manuscript, and it is easy to add. A similar analysis for the Akt sensor is missing in figure 3 and should be added for consistency. Figure 4 shows data for Akt, but as timeseries and only for Histamine. See point 2, it would benefit the reader greatly if ERK and AKT are presented in a more uniform and complete fashion throughout the manuscript.

4) In the results text of figure 4, the authors state that "...as shown in Figure 4C-D, which is in line with the effect of histamine on ERK.". It is unclear what the authors mean with this statement, the effects of single/double inhibition of Histamine stimulation on ERK are not quantified or discussed. Both responses can be quantified more carefully and compared.

5) This paper would benefit from a mechanistic investigation. For instance, the authors could investigate the pathways that lead to the generation of the pulse of ERK and Akt. These (preliminary) results presented call for deeper investigation into the signaling pathway from Gai and Gaq to ERK and AKT, and the authors are in a great position to probe this. One simple approach is to explore the upstream pathway, such as the MAPK cascade, PI3K, RTKs by means of inhibitors.

6) Since different G-proteins seem to elicit similar responses on ERK and especially for Akt, it is likely a B-arrestin / beta-gamma subunit mediated mechanism? It would be interesting to hear what the authors think of this, did they investigate/consider this possibility? E.g. Perhaps blocking RTK signaling / B-arrestin signaling would reduce heterogeneity?

7) The authors should take a serious effort to summarize the data in the figures better. Many plots that can be merged/presented in a more concise way, which would improve the readability of the manuscript greatly.

Minor points:

1) The authors should spell out in the legend of each figure if they are representing the absolute C/N or the normalized C/N

2) In Figure 2 the authors should show the control with no stimulus. Also would be informative to inform the reader about the stimulation protocol used, or indicate the stimulation time and length in the figure.

3) Figure 3: This figure would benefit from a different presentation of the data, it is currently confusing. E.g. Average curves per drug condition in a single graph would present the point the authors make more clear and concise, and this single cell overview can be moved to supplements.

4) Figure 4 legend states "CN ERK" and "ERK C/N", but is depicting only Akt responses? Only in 4c the axes are labeled, this together is very confusing.

5) Figure 5 is missing the controls with ERK and Akt inhibitors, to show the loss of correlation between the AUC of the two

6) Figure 6, the presumed lack of correlation between baseline activity and response should be confirmed statistically.

7) It seems that in S1P treated cells there is a second oscillation in ERK activity well visible in figure 2 and also in S10. Could the authors comment on that?

8) In the abstract it is unclear what authors mean with "UK".

9) Figure 9, it would be helpful to visually repeat the typical curve of the different clusters here, to guide the reader.

10) The observed heterogeneity in responses might be related to different cell cycle stages, did the authors investigated/consider this possibility (e.g. with a cell cycle biosensor)?

Significance

The paper describes with high accuracy the dynamics of ERK and Akt biosensors downstream of several GPCRs.

However, it feels like this is a preliminary report that leaves many important questions still open. It does not provide mechanistic insight and doesn't fully exploit the potential of single-cell technologies. The authors have the tools to investigate several important questions that are left open in the manuscript (e.g. connection Gaq/Gai to ERK/AKT, B-arrestin/betagamma involvement). Moreover, some important controls are missing. The authors should also consider the data presentation in the figures, to improve readability and interpretation of the manuscript.

Properly revised, would be of interest for a broad audience in cell biology, specifically GPCR and RTK signaling fields.

Expertise in cell biology, gpcr and rtk signaling, fluorescent biosensors.

Referee Cross-commenting

I agree with the assessments by the other reviewers.

Indeed showing the dynamic range of the biosensors, as Reviewer #3 states, would strengthen the manuscript and put the S1P response heterogeneity in context.

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 authors have performed highly quantitative analyses of GPCR signaling to reveal heterogenous ERK and Akt activation patterns by using kinase translocation reporters. Using a massive number of single-cell imaging data, the authors show heterogeneous responses to GPCR agonists in the absence or presence of inhibitors. By cluster analysis, the responses of ERK and Akt were classified into eight and three patterns. This paper is clearly written with sufficient information for the reproducibility. However, the conclusion may not be necessarily supported by the provided data as described below.

Major comments:

This work has been well done in an organized way and adds new insight into the regulation of protein kinases by GPCRs. The conclusion will be of great interest in the field of single-cell signal dynamics and quantitative biology. On a bit negative note, considering the complexity of the downstream of GPCRs, some of the conclusions may need revision.

1. The conclusion of the title that "Heterogeneity and dynamics of ERK/Akt activation by GPCR depend on the activated heterotrimeric G proteins," may not be supported by the data. The authors compared just one pair each of GPCR and ligand. The heterogeneity may come from the nature of the ligand or the characteristics of the single clone chosen for this study.
2. The obvious question is that why the authors did not analyze the correlation between ERK and Akt activity more extensively. Cell Profiler will be able to extract multiple cellular features. Linking the heterogeneous signals to cellular features will benefit readers in the broad cell biology field. If the authors wish to write another paper with that data, it should be at least discussed.
3. Another apparent flaw of this work is that YM was not challenged to UK-stimulated cells. The authors probably assumed lack of effect. Nevertheless, I believe it is required to show. Or, remove the PTx data from the Histamine-stimulated cell data.
4. The most interesting response is that of S1P. ERK is biphasically activated. Combined inhibition of Gq and Gi failed to suppress ERK activity. It may be discussed why the biphasic activation pattern was not identified by the classification.
5. The authors argue that the brightness of the KTR reporter was not correlated with the dynamic range of ERK or Akt reporter (Supplementary Figure 3), but it is not clear. I had an impression that ERK-KTR brightness (Supplementary Figure 3A) has a slightly negative correlation with "maximum change in CN ratio" (Supplementary Figure 3B) (e.g., A6>B3>B5 in brightness and A6<B3<B5 in maximum change in CN ratio). The authors should show dot plots of average fluorescence vs. the maximum change in CN ratio.
6. The authors have shown cluster analyses for the temporal patterns in kinase activations. However, the only difference of cluster 3 and 5 (Figure 7) seem to be amplitude. The authors have also shown the amplitude is dependent on the dose of the activators, which together makes it difficult to see the biological meaning of discriminating the two patterns in comparing different agonists, e.g., Histamine, UK, and S1P. The authors should discuss their views on how the clustering analyses will benefit biological interpretations together with possible limitations.
7. Considering the importance of the content, the supplemental note 2 may be included in the main text.

Minor comments:

1. The authors should clarify the cell type they used (HeLa cells) in the main text and figure legends.
2. Supplementary note1: The data-not-shown data (no correlation of KTR expression and its response to serum) should be very informative for the readers. The data should be shown as an independent supplementary figure.
3. Supplementary Figure S2: The authors should clarify this image processing is about background subtraction. Also, the authors should clearly note "rolling ball with a radius of 70 pixels" is about an ImageJ function, "Subtract Background".
4. Supplementary Figure S5: Figure labels are "A, A, B, B" not "A, B, C, D". Also the top two figures are lacking Y axis labels.
5. Page6 (top): The authors should mention the description is about Supplementary Figure S5 (UK) and Supplementary Figure S6 (S1P).
6. Figure 3: the figures are lacking x-axis labels (probably uM, nM and pM from left).
7. Values in tables: The significant figure must be 2, at best. This should be consistent throughout the text. For example, "The EC50 values for histamine, S1P and UK were respectively 0.3 μM, 63.7 nM and 2.5 pM." This is somewhat awkward.
8. Page 7, the first paragraph: No comments on S1P!
9. Fig. 3: 100 mM must read as 100 micromolar.
10. Fig. 9: Concentration unit is missing.
11. Page 11, line 4: EKR should read as ERK.
12. Page 13: "So far, only a couple of studies looked into kinase activation by GPCRs and these studies used overexpressed receptors [32,33]." Please describe precisely. Protein kinase activation by GPCR has been studied more than 20 years. Why are these two recent papers cited here?
13. "This is in marked contrast to other fluorescent biosensors that typically require an overexpressed receptor for robust responses [34]." Following words should be included in the end: "in our hands".
14. "Histamine is reported to predominantly activate Gq in HeLa cells [36] and UK activates Gi [37]." Describe the name of receptors for the better understanding.
15. "S1P can activate a number of different GPCRs, all known to be expressed by HeLa cells [24]." Why is this paper chosen? The authors can easily find RNA-Seq data, if they wish to see the expression level. The cited paper did not scrutinize the S1P receptors expressed in HeLa cells.

Significance

The authors used biosensors for ERK and Akt to examine the kinetics of activation by GPCR ligands. Technical advancement is in the massive analysis method and cluster analysis. This is an important direction for the quantitative biology. GPCR signaling is complex because of multiple receptors coupled with different G proteins. The simple ones such as histamine receptor and alpha2-adrenergic receptor can be easily analyzed as shown in this study. However, there are many S1P receptors, which make the interpretation difficult. If the authors could have shown interesting proposal on this data, the paper may interest many researchers in the field of cell biology and systems biology.

Expertise: Cell biology, signal transduction of protein kinases, fluorescence microscopy.

Referee Cross-commenting

1. I agree with the other two reviewers in that immunoblotting data is required to show the efficiency of P2A cleavage.
2. All reviewers think it looks strange that the authors did not show UK + YM data.
3. Showing the dynamic range of the biosensors will reinforce the data as Reviewer #3 states. ERK-KTR is quite sensitive and can be easily saturated. Ideally, the ratio of pERK vs ERK can be quantified by the different mobility in SDS-PAGE. But, I do not know how we can do it for Akt.

Note: This rebuttal 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

We thank the reviewers for their appreciation of our study and for their comments, which we believe helped us to improve our manuscript.

We have carefully considered both the general comment on the significance of our work expressed by the Reviewers 1 and 3 and a few specific points requested by the Reviewer 2 and we believe we have answered all of the reviewers' concerns.

2. Point-by-point description of the revisions

Reviewers 1 and 3 agreed that our work was both original and well performed. Although they have not raised any specific issues, apart from minor editorial changes, they asked us to clarify the potential interest of our results in designing novel therapeutic interventions for hepatocellular carcinoma.

Reviewer 1: While studies are very elegant and results convincing, it is unclear how they might be deployed to therapeutic ends.

Reviewer 3: The authors should explain why this is an interesting finding. They mention in the abstract that this heterogeneity highlights potential vulnerabilities that could be therapeutically exploited. How do they envision this? Why is this not a trivial result and in what way can this observation help design new therapies?

We believe that our results suggest exciting opportunities in the search for novel therapeutic options and we agree further discussion on this important issue should be included in the revised manuscript. We have now expanded the discussion on these points and commented on the clinical relevance of our findings to answer the reviewers' concern (Discussion section page 15 lines 7-15).

Our data are consistent with the widely acknowledged role of NK cells in anti-tumour immunity (e.g. Pende et al, Frontiers Immunol. 2019, 10:1179, for review). NK activity is governed by engagement of a repertoire of activating and inhibitory receptors expressed on their surface (Shimasaki et al., Nature Rev Drug Discovery, 2020, 19, 200-218). Among the latter, homophilic interactions of CEACAM1 (CD66A) expressed on melanoma cells have been shown to protect the tumor from NK-mediated toxicity (Markel et al, J Immunol 2002; 168:2803-2810;), in a strict parallel to our interpretation.

NK cells are relatively sparse in the peripheral blood and abundant in a healthy liver. In patients with HCC, the numbers of peripheral, liver resident and tumor-infiltrating NK all drop significantly, mainly due to the disappearance of CD56dimCD16pos cell subset, corresponding to the cytotoxic NK population. Moreover, despite the continuous expression of activating receptors, the functionality of both the cytotoxic (and the cytokine producing (CD56brightCD16neg) remaining NKs is severely impaired (Cai et al Clinical Immunology (2008) 129, 428–437). The molecular mechanisms underlying NK anergy in the context of HCC have yet to be fully elucidated. However, CEACAM1 expression has been shown to suppress NK function in hepatitis C patients (Suda et al Hepatology Communications 2018;2:1247-1258) and there is ample evidence of CEACAM1 playing a major role in hepatic disease and in particular in protection against inflammation and immune-induced hepatitis (reviewed in Horst et al Int. J. Mol. Sci. 2018, 19, 3110). Thus, CEACAM1 is a bona fide regulator of NK function that is relevant in cancer and in non-cancerous liver pathology.

In this context, our data introduce an additional notion, namely the tumor-promoting effect of a strong ERK activation in HCC that leads to CEACAM1-mediated anergy of NK.

How might these findings be translated into future therapeutic options for HCC? Several scenarios can be envisaged, a very attractive being a cell-mediated immunotherapy, notably either autologous or allogeneic NK transfer. These therapies, which were initially developed for hematopoietic malignancies, are currently gaining momentum for solid tumors. Infusion of modified NK cells, including CAR-NK, presents major advantages over T-cell based therapies, mainly due to a very much diminished risk of GVDH in allogenic setting and of cytokine release syndrome and neurotoxicity for autologous transfer. Moreover, because NK-mediated cytotoxicity is HLA-independent, it does not require careful haplotype matching, thus greatly increasing the speed and availability of cellular preparations (recently reviewed in Xie et al. EBioMedicine 59 (2020) 102975).

Currently, there are 219 registered clinical trials, including 31 on HCC, for NK-mediated anti- solid tumor responses (clinicaltrials.gov). Although most of these are only in phase I or phase II, they bear a great promise for the future. Our data strongly suggest that a new combination therapy might have an improved efficiency in a subset of HCC characterized by a strong ERK activation. This would involve either activated NK or CAR-NK in combination with a FDA-approved inhibitor of the MAPK ERK, such as trametinib. Our data lead us to predict that even a partial decrease in the intensity of ERK signaling would be likely to significantly increase the efficacy of NK-mediated anti-tumor activity, at least in a subset of HCC. While we appreciate that this suggestion remains speculative at this point in time, we believe the strength and novelty of our data warrants an exploration of such novel therapeutic opportunity for this tumor type that dramatically lacks reliable treatment options.

• Specific points Reviewer 1

Minor edits: Editorial review for minor, infrequent word usage edits

We apologize for any English language mistakes in the manuscript. While the formulation of the remark makes us believe that our word usage does not impair the understanding of the text, we shall of course be willing to correct it.

Figure 1E: Not possible to read genes in left heatmap, middle heatmap very small. Figure 3D: Units at x and y axes not legible/small. 3E: scales not legible. Figure 4: typo in legend H-> G.

We apologize for not being more careful in preparing these figures, this has now been corrected. We realize that due to the high number or genes, their names are still in a very small print in the left panel of Fig. 1E, however, the complete list the genes is given in the supplementary table 1, and we added larger image of the heat map with the table.

Reviewer 2

1. Is there any significant change of EMT like status in BMEL cells having H-RAS (high) vs H-RAS (low)? Several EMT markers (e.g. vimentin or loss of E-cadherin) are induced by H-Ras in BMEL cells, as we have previously reported (Akkari et al. J Hepatol, 2012). Moderate levels of Ras expression appear to be sufficient for this phenotype, since we did not detect significant differences in their expression profiles either between RASHIGH and RASLOW populations or between cells isolated from the hepatic versus the peritoneal tumors. We conclude that the phenotype of a selective advantage afforded by a high RAS expression level is not due to the EMT.

There is no significant fold difference (MFI number) to put the sorting gates to enrich H-RAS high vs H-RAS low cells. The mRNA expression level was almost 3 fold difference. Is it correlated with protein expression level?

This is a very valid point. We agree that the MFI difference is not strong, although it is in fact statistically significant in three independent cell sorting experiments. We were confident that the differences in the H-Ras mRNA level were reflected in the level of protein expression, since we have observed distinct transcriptional signatures as well as significant phenotypic differences in RasHIGH and RasLOW cells (Fig. 1 B and D). Nevertheless, we quite agree that the difference in protein expression level needed to be confirmed. This has now been done by immunoblot analysis of protein extracts with an antibody specific to RasG12V (Cell signaling #14412). These data have now been included in Figure 1A, and the text modified accordingly (page 4 line 9-19).

Is there any translational relevance of these genes Al467606, Aim2, Dynap, Htra3, Itgb7, Tspan13 in HCC patients with poor survivability?

The expression of these genes positively correlated with the level of the Ras oncogene in the ex vivo cell culture model, thus providing a nice demonstration that variation in HRAS oncogenic dosage translates into differential transcriptomic outputs. The analysis of publicly available data from the cancer genome atlas (TCGA) also showed their expression in the HCC cohort (372 patients samples). The clinical outcome of the level of their expression (shown below) is somewhat ambiguous: strong expression of ITGB7 and C16ORF54 (Human ortholog of Al467606) correlated with a better prognosis, while expression of AIM2 and DYNAP had no impact on patient overall survival. Finally, HTRA3 and TSPAN13 were associated with worse outcomes and thus constitute particularly interesting candidates for future investigations. These somewhat unexpected divergent correlations likely reflect the fact that RAS/MAPK signalling is unlikely to be the sole regulator of their expression.

4.Is there any difference between survival curve upon grafting of H-RAS (high) vs H-RAS (low) cells in Fig.2A?

This experiment has not been performed for ethical reasons. Indeed, the difference in tumor growth upon injection of RasHIGH vs RasLOW is statistically significant 21 days after injection (Fig. 2A, p-value= 0.008). The size of the RasHIGH tumours is rather large and we chose to sacrifice the animals before they developed any signs of suffering.

Is there any difference of H-RAS expression between liver tumor and peritoneal tumors?

We have quantified H-Ras expression levels by RTqPCR in the flow cytometry sorted tumoral cells derived from the liver and peritoneal tumours (Fig. 3C). In the revised version of the manuscript we provide evidence that the mRNA expression levels of the oncogene correlate with the protein expression. Therefore, while the measurement of H-Ras protein has not been performed on the tumours, we would argue that it will indeed be different in the two tumour locations.

Please provide the data for pro-inflammatory cytokines in TME.

These data have been shown in the Suppl. Fig. 4C.

Please provide an explanation of the DC activation with antigen presentation though the tumor is non-necrotic or apoptotic.

While it is true that peritoneal tumors are less necrotic and have a lower apoptotic index than the matched hepatic primary ones (Fig. 3E), significant cell death can be detected at both locations. We assume that the released antigens are sufficient for presentation by the DC, as supported by the data in Fig. 4C, D and G.

Is the TAM showing M2 phenotypes at peritoneal tumors?

The reviewer correctly points out that the distinction between liver and peritoneal TAM polarization is not perfectly clear-cut, since some immunosuppressive but also some inflammatory markers are present at both tumor locations (Fig. 4B and Suppl Fig4). This is not unexpected, as the spectrum of activation macrophages can undertake in vivo is neither static nor fully faithful to the M1/M2 polarization extremes inducible in vitro (see e.g. Ringelhan et al., Nat Immunol. 2018;19(3):222-232 ; Ruffell et al., Trends Immunol. 2012 33(3):119-26). We thus integrate these results with our observations of other modulated immune cell phenotypes in these tumors. Indeed, in addition to the macrophage polarization markers, we noted a more mature, activated phenotype in the peritoneal TAMs. Together with the cytokine expression profile in the two tumor locations (which is included as a supplementary table in the revised version of the manuscript) our data argue for a less inflammatory environment in the peritoneal tumors.

Significance

The data showed pretty promising and has a seminal impact on H-RAS high expressing HCC patients. TAM and DC showed some important immune regulation to promote HCC.

We thank the reviewer for his appreciation of the significance of our study.

Reviewer 3

It is possible that RAS levels may not stay constant but dynamically go up and down. While this is a possibility that would complicate interpretations of the results, I am ok with the conclusions in the manuscript as it is, since there seems to be a significant difference between the different populations assayed.

This is a valid point that we have addressed by comparing the H-RAS expression level in the parental BMEL population (labelled “cells before injection” in Fig 2D) to those either freshly isolated from the tumors after a rapid cell-sorting by flow cytometry (“tumors” in Fig. 2D) and then to those isolated from tumors and kept in culture for 14 days (“tumoral cell lines” in Fig. 2D). Our conclusion was that the level of RAS expression was stable upon ex vivo culture. This result does not exclude a possibility of epigenetic regulation that operated in vivo and was maintained in the subsequent cell culture. However, even if this was the case, it would not alter the conclusion of distinct selective advantage of the HRAS expression levels in the two tumoral locations.

Significance

The authors should explain why this is an interesting finding. They mention in the abstract that this heterogeneity highlights potential vulnerabilities that could be therapeutically exploited. How do they envision this? Why is this not a trivial result and in what way can this observation help design new therapies?

This important point is very similar to the concern raised by the reviewer 1 and we have answered them together at the beginning of the rebuttal.

We would like to thank again the reviewers for raising this issue, which prompted us to include the considerations of potential usefulness of our findings in the revised discussion.

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

Learn more at Review Commons

---

Referee #3

Evidence, reproducibility and clarity

In this manuscript, Lozano et al. used primary hepatocyte precursor cells expressing high and low levels of oncogenic RAS to determine the dose dependent effects of RAS on tumor formation. The authors found that cells expressing different levels of RAS encounter different selection forces that modulate tumor growth. That is, high RAS expression was required in tumor cells growing in the liver, but not in the peritoneum. These findings suggest that differences in tumor microenvironment activate selection mechanisms that trigger tumor heterogeneity. The authors also show that different levels of RAS signaling cause resistance/sensitivity to NK cell attack.

The experiments are done well, and the conclusions follow from the data, with the challenge that the RAS high and low cells are not clones but are purified from a population of cells expressing a continuum of different levels of RAS. It is possible that RAS levels may not stay constant but dynamically go up and down. While this is a possibility that would complicate interpretations of the results, I am ok with the conclusions in the manuscript as it is, since there seems to be a significant difference between the different populations assayed.

Significance

I find this finding conceptually only somewhat interesting, the most interesting aspect being that the liver is a more selective environment than the peritoneum. Do the authors have an explanation for this? It is certainly no surprise that tumor cells require different signaling activities to survive and proliferate in different environments. Here it is different levels of RAS, why not.

The authors should explain why this is an interesting finding. They mention in the abstract that this heterogeneity highlights potential vulnerabilities that could be therapeutically exploited. How do they envision this? Why is this not a trivial result and in what way can this observation help design new therapies?

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

[NOTE FROM THE EDITOR: THIS REVIEWER INDICATED S/HE DOES NOT WISH TO BE CONTACTED AGAIN]

Anthony Lozano et al manuscript"Ras/MAPK signalling intensity defines subclonal fitness in a mouse model of primary and metastatic hepatocellular carcinoma" showed a mechanistic role of Ras/MAPK pathway in HCC with some immune mechanism. However the author needs to address the following concerns in the manuscript.

Major Comments-

1. Is there any significant change of EMT like status in BMEL cells having H-RAS (high) vs H-RAS (low)?
2. There is no significant fold difference (MFI number) to put the sorting gates to enrich H-RAS high vs H-RAS low cells. The mRNA expression level was almost 3 fold difference. Is it correlated with protein expression level?
3. Is there any translational relevance of these genes Al467606, Aim2, Dynap, Htra3, Itgb7, Tspan13 in HCC patients with poor survivability? 4.Is there any difference between survival curve upon grafting of H-RAS (high) vs H-RAS (low) cells in Fig.2A?
4. Is there any difference of H-RAS expression between liver tumor and peritoneal tumors? 6.Please provide the data for pro-inflammatory cytokines in TME.
5. Please provide an explanation of the DC activation with antigen presentation though the tumor is non-necrotic or apoptotic.
6. Is the TAM showing M2 phenotypes at peritoneal tumors?

Significance

The data showed pretty promising and has a seminal impact on H-RAS high expressing HCC patients. TAM and DC showed some important immune regulation to promote HCC.

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 a well-written, elegant manuscript where authors demonstrate that hepatocyte precursors (BMEL), when transformed with different copy numbers of oncologenic mutant H-Ras G12V, exhibit a dose-dependent fitness. Authors find that low dose H-ras mutant clones appear to be eliminated from primary tumors, but not from secondary tumours. Authors suggest that the different (primary v secondary) microenvironments, especially with respect to innate immunity, influence selection pressure differences. Specifically, investigators find that ceacam1-driven NK inhibition contributes to this clonal selection using murine models with and without adaptive immunity.

Minor Weakness: While RAS pathway activation is present in 40% of HCC, given that few HCC are driven by Ras mutations, relevance could be called into question. Authors offer reasonable explanations for why these used the models they did.

Minor edits: Editorial review for minor, infrequent word usage edits

Figure 1E: Not possible to read genes in left heatmap, middle heatmap very small. Figure 3D: Units at x and y axes not legible/small. 3E: scales not legible. Figure 4: typo in legend H-> G.

Significance

HCC treatment options remain limited and outcomes are poor. Defining which pathways are coopted by HCC to improve fitness as primary or secondary tumors may improve treatment options. While studies are very elegant and results convincing, it is unclear how they might be deployed to therapeutic ends.

Note: This rebuttal 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

We thank the reviewers for their helpful, detailed and insightful comments. We have modified the figures and rewritten large sections of the manuscript following the reviewers’ suggestions. In addition, we have incorporated new data throughout the manuscript and figures to clarify and better support our conclusions. All of these changes have significantly improved the coherence, consistency and clarity of our data, and have allowed us to better communicate the advance our findings represent for the fields of splicing and muscle development.

Please find a point-by-point response to the reviewers’ comments below. The reviewers’ comments are in black and italics.

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

Rbfox proteins regulate skeletal muscle splicing and function and in this manuscript, Nikonova et.al. sought to investigate the mechanisms by which Rbfox1 promotes muscle function in Drosophila.

Using a GFP-tagged Rbfox1 line, the authors showed that Rbfox1 is expressed in all muscles examined but differentially expressed in tubular and fibrillar (IFM)muscle types, and expression is developmentally regulated. Based on RNA-seq data from isolated muscle groups, the authors showed that Rbfox1 expression is much higher in TDT (jump muscle) than IFM.

Using fly genetics authors developed tools to reduce expression of Rbfox1 at different levels and the highest levels of muscle-specific Rbfox1 knockdown was lethal and displayed eclosion defects (deGradFP > Rbfox1-IRKK110518 > Rbfox1-RNAi > Rbfox1-IR27286). Consistently, Rbfox1 knockdown flies have reduced jumping and climbing phenotypes, due to tubular muscle defect where Rbfox1 is expressed at higher levels. Rbfox1 knockdown in IFM caused flight defects which have been shown previously. Further characterization of IFM and tubular muscles demonstrated a requirement of Rbfox1 for the development of myofibrillar structures in both fibrillar (IFM) and tubular fiber-types in Drosophila. Interestingly, knockdown or overexpression of Rbfox1 displayed hypercontraction phenotypes in IFMs which is often an end result of misregulation of acto-myosin interactions which was rescued by expression of force-reduction myosin heavy chain (Mhc, P401S), in the context of Rbfox1 knockdown (the rescue experiment could not be performed with Rbfox1 overexpression due to complex genetics).

Authors also performed computation analyses of the Rbfox binding motifs in the fly genome and identified GCAUG motif in 3,312, 683, and 1184 genes in the intronic, 5'UTR, and 3'UTR, respectively. These genes are enriched for factors that play important roles in muscle function including transcription factors (exd, Mef2, Salm), RNA-binding proteins (Bru1), and structural proteins (TnI, encoded by wupA). Many of these gene transcripts and proteins are affected in flies with reduction or overexpression of Rbfox1. Using fly genetics, authors propose and test different mechanisms (co-regulation of gene targets by Rbfox1 and Bru1), and regulators of muscle function (exd, Me2, Salm) and structural proteins (TnI, Mhc, Zasp52, Strn-Mlck, Sls) by which these changes could affect the muscle function.

*Overall, the characterization of Rbfox1 phenotypes and myofibrillar structure is very well elucidated, mechanisms by which Rbfox1 affects muscle function are not clear and remain largely speculative. We thank the reviewer for the positive evaluation of our phenotypic analysis of Rbfox1 knockdown in multiple muscle fiber types. This manuscript is the first detailed characterization of Rbfox1 in Drosophila muscle, extending far beyond our previous finding that Rbfox1-IR flies are flightless. Beyond behavioral and cellular phenotypes, we report that there are regulatory interactions between Rbfox1, Bruno1 and Salm and identify other Rbfox1 targets in flies. We acknowledge that there are molecular and biochemical details of specific regulatory mechanisms that remain to be elucidated, but this paper provides many foundational observations to guide future biochemical experiments and is thus important to the muscle field.

\*Major comments**

*1. The varying level of Rbfox1 knockdown (deGradFP > Rbfox1-IRKK110518 > Rbfox1-RNAi > Rbfox1-IR27286) was achieved by different strategies without validation at the protein level (likely due to lack of a Rbfox1 antibody). It is important to show different Rbfox1 protein level (at least with different RNAi), especially when authors propose that autoregulation of Rbfox1 causes increased level Rbfox1 transcript in case of Rbfox1-RNAi (mild knockdown). Autoregulation of Rbfox1 in mammalian cells may not be similar in flies.

To address this comment, we have toned-down the discussion of level-dependent regulation throughout the manuscript, and have removed claims of Rbfox1 autoregulation. We appreciate the reviewer’s point that it would be ideal to be able to determine the protein levels of Rbfox1 in the different knockdown conditions. We have tested the published antibody against DmRbfox1, but it is very dirty and we see multiple bands in Western Blot. This background partially obscures the bands from 80-90 kDa at the molecular weight where we expect Rbfox1, and prevents accurate quantification (see Reviewer Figure 1). Verification of protein levels of Rbfox1 will require generation of a new antibody which is beyond the scope of this study. As we do not have a good antibody, we performed two experiments to demonstrate our ability to tune knockdown efficiency. First, we crossed Rbfox1-IRKK110518 and Rbfox1-IR27286 to UAS-Dcr2, Mef2-Gal4 and demonstrated we could enhance the phenotype (Figure 2A, B). Second, we performed knockdown with the same hairpins at different temperatures and demonstrate that stronger knockdown at higher temperature leads to stronger phenotypes with the same hairpin

(Figure 2B). This data supports our knockdown series interpretation.

Reviewer Figure 1. Western Blot of whole fly with anti-Rbfox1 (A2BP1) (Shukla et al., 2017). Tubulin was blotted as a loading control.

• TnI and Act88F protein levels are inversely correlated with Rbfox1 level in IFM but did not correlate with the RNA level. Using RIP authors showed that Rbfox1 was shown to bound to wupA transcripts (has Rbfox binding sites) but not Act88F transcripts (does not have Rbfox binding sites). Authors performed Rbfox1 IP and identified co-IP of components of cellular translational machinery and propose that wupA (TnI) levels are regulated by translation or NMD (non-sense mediated decay). A follow up experiment was not performed to identify the mechanism by which TnI level is regulated by Rbfox1. *

Further biochemical and genetic verification of the underlying mechanisms of Rbfox1 regulation in Drosophila muscle will be addressed in a future manuscript, as in vivo modulation of translation or NMD in an Rbfox1 knockdown background involves recombination to coordinate multiple genetic elements. We have modified the text to reflect this hypothesis remains to be explored in future experiments (Line 473-474).

We have further added RT-PCR data for wupA transcript levels in IFM and TDT with Rbfox1-IRKK110518 knockdown (Figure S4 A), but as in Rbfox1-RNAi flies, there is not a significant change in expression. We do see significant downregulation of Act88F when we overexpress Rbfox1 in IFM (Figure S4 B), as well as in TDT when we knockdown Rbfox1 with either Rbfox1-IRKK110518 or Rbfox1-IR27286.

It was known that TnI mutations (affects splice site, fliH or Mef2 binding site, Hdp-3) led to a reduction in TnI level and hypercontraction. Authors showed rescue of hypercontraction phenotype in hdp-3 background by knocking down Rbfox1, likely due to increase in wupA transcription (Mef2-dependent or independent manner). However, no rescue was observed in the fliH background. Reduced level of Rbfox1 in fliH background would be expected to cause worsening of phenotype as splicing of remaining wupA transcripts would be affected with reduced Rbfox1 level. The splicing of wupA of exon 4 is not affected in Rbfox1 knockdown (fig. 6U), it's not clear if the splicing of exon 6b1 is affected in Rbfox1 knockdown.

We thank the reviewer for pointing out our lack of clarity regarding exon 6b1 and IFM-specific isoform 6b1. To address this comment and validate our previous data, we performed additional Sanger sequencing on RT-PCR products, added a diagram of the wupA gene region in Figure 4 A and improved the clarity of our discussion of the fliH and hdp3 alleles and our results in the text.

To directly respond to the reviewer, first, it is unclear if the reduced level of Rbfox1 in a fliH background should actually cause a more severe phenotype. Our data suggests that Rbfox1 represses TnI expression through binding the 3’-UTR, and can likely indirectly regulate wupA expression level via Mef2. Thus, arguably, the reduced level of Rbfox1 in the fliH background might not affect splicing, as the mutations in the regulatory element should rather make wupA insensitive to increased Mef2 expression in the Rbfox-RNAi background.

Second, we confirmed via Sanger sequencing of RT-PCR products that both IFM and TDT in control and Rbfox1-IR flies use exon 6b1 (current exon 7). The IFM isoform contains exon 3, 6b1 and 9, while the TDT isoform contains exon 3 and 6b1, but skips exon 9 (see Figure 4 A). In other tubular muscles, wupA isoforms skip exons 3 and 9, and use exon 6b2 instead of 6b1. Thus, to directly answer the reviewer’s question, no, splicing of exon 6b1 itself is not affected by Rbfox1. However, Rbfox1 does influence expression of the ”6b1 isoform”, or the wupA isoforms in IFM and TDT containing exon 6b1 and exon 3. Additionally, our data shows that Bru1, not Rbfox1, regulates alternative splicing of wupA exon 9 (Fig. S6 T).

What the reviewer has correctly identified with this comment is that the effect on splicing in the hdp-3 allele also appears to be complex and to have not been fully clarified. Although hdp-3 results from mutation of a splice site in exon 6b1 (which based on (Barbas et al., 1993) results in aberrant use of 6b2 in IFM), it also results in a near complete absence of the longer isoform containing exon 3 in adult flies. hdp-3 is reported in the same paper to affect both IFM and TDT, which both express isoforms containing exon 3 and 6b1. It is not known how mis-splicing of exon 6b1 leads to loss of isoforms containing exon 3, but our data indicate that Rbfox1 is somehow involved. It is purely speculation and beyond the scope of this manuscript, but perhaps selection of alternative exons in wupA are not independent events (ie that the splicing of exon 3 depends on correct splicing of exon 6b1). This could be mediated with interactions with chromatin, the PolII complex or through a larger splicing factor complex (something like LASR, for example (Damianov et al., 2016)), that restricts choice in alternative events through higher-order interactions. Another possible mechanism is that a second mutation exists in the hdp-3 allele that affects splicing of exon 3, although this was not indicated in the extensive sequencing data in (Barbas et al., 1993).

Bruno1 was identified as a co-regulator of Rbfox1 in different IFM and tubular muscle types. However, except Mhc, other Rbfox1 targets seem to be regulated by either Rbfox1 or Bruno1, not both. Analyses of RNA-seq datasets from single and double knockouts should identify additional targets to support the claim that - Rbfox1 and Bruno1 co-regulate alternative splice events in IFMs. Phenotypic changes with reduced Rbfox1 and Bruno1 double knockdowns are very severe, but the mechanistic basis of such genetic interaction resulting in synergistic phenotypes in IFMs is lacking as splicing changes in single vs double knockout is similar.

We agree with the reviewer that RNA-seq data would be useful to obtain a genome-wide perspective on the regulatory interactions between Rbfox1 and Bru1, and we plan to generate this data as part of a future manuscript. However, the tissue-specific dissections to isolate enough material from all of the necessary genotypes will take months to complete, and are not realistic to wait to include in this manuscript. Instead, to address the reviewer’s question, we have expanded our RT-PCR experiments to cover a wider panel of events in 12 sarcomere genes (see new data in Figures 6 and S6 and summary in Figure 8). We now can show that splice events in Fhos and Zasp67 are Rbfox1 dependent, while events in sls, Strn-Mlck and wupA are Bru1 dependent. An event in Zasp66 responds to both Rbfox1 and Bru1, but in opposite directions. Events in Mhc, Tm1 and Zasp52 are regulated by both Rbfox1 and Bru1 (or are sensitive to changes in Bru1 expression in the Rbfox1 background), and change in the same direction. This data provides a clearer mechanistic basis for the synergistic phenotype observed between Rbfox1 and Bru1 in IFM.

Rbfox1 is expressed at a high level in tubular muscle whereas Bruno1 is expressed at a high level in IFM. Rbfox1 binds to Bruno1 transcript and inversely regulates Bru1-RB level but knockdown of Bru1 does not affect Rbfox1 level (Fig. S5 G,I,J). Overexpression of Bruno1 decreased the Rbfox1 level, however, it's difficult to interpret these results as overexpression of Bruno1 may have other effects on IFM gene expression.

The reviewer correctly pointed out that we did not observe significant changes in Rbfox1 mRNA levels in the mutant bru1M3 background, however, in the original version of this manuscript, we also showed a significant decrease in Rbfox1 expression in IFM from the bru1-IR background at both 72 h APF and 1 d adult in mRNA-Seq data. To clarify differences in Rbfox1 levels between bru1-IR and our bru1 mutant backgrounds, we have performed additional RT-PCR experiments. We examined Rbfox1 levels after knockdown of bru1 (bru1-IR), and we now show that Rbfox1 levels are significantly decreased in IFM and TDT after bru1-IR (Fig. 5S, Fig S5 I). We see a weaker effect in the bru1M2 hypomorphic mutant, which likely reflects differences in Bru1 expression levels in bru1-IR and the bru1M2 allele. These results are consistent with the mRNA-seq data we presented previously (now in Fig. 5R). These additional data suggest that loss as well as gain of Bru1 affects Rbfox1 expression levels.

A dose-dependent effect of Rbfox1 knockdown was shown to regulate the expression of transcription factors that are important for muscle type specification and function including exd, Mef2, and Salm. However, it is not clear how Rbfox1 mechanistically regulates the expression of these transcription factors.

We present two pieces of data suggesting possible regulatory mechanisms for Mef2. First, RIP data suggest Rbfox1 can directly bind the 3’-UTR region of Mef2, and this region contains two binding motifs identified in both the oRNAment database and in our PWMScan dataset. Second, we show that use of the 5’-UTR regions of Mef2 is altered in Rbfox1-IR muscle. Although not definitive, this suggests that regulation of alternative 5’-UTR use may influence transcript stability or translation efficiency. We feel the many experiments to elucidate the detailed mechanism of regulation (and indeed to determine the likely contribution of multiple, layered regulatory processes) are beyond the scope of this paper, and are better left for future studies. This manuscript is the first in-depth characterization of Rbfox1 function in Drosophila muscle, and we provide multiple lines of evidence suggesting that different regulatory mechanisms exist as a basis for future experiments to explore these interesting and important regulatory interactions.*

**Minor comments**

1. It is not described if the rescue of Rbfox1 knockout by expression of force-reduction myosin heavy chain (Mhc, P401S) led to rescue of phenotypes (jumping, climbing, flight). *

Force-reduction myosin heavy chain MhcP401S is a mutation at the endogenous Mhc locus that results in a headless myosin and was previously characterized to be flightless (Nongthomba et al., 2003). It is however able to rescue jumping and walking defects observed with the hdp2 TnI allele, and supports largely normal myofibril assembly (Nongthomba et al., 2003). It is also important to note that fibrillar muscle function is very finely tuned, such that alterations that result in flightlessness in many cases do not alter myofibril structure as detected by confocal microscopy (Schnorrer et al., 2010). We therefore looked at myofiber and sarcomere structure as a more sensitive read-out of the rescue ability in the Rbfox1 knockdown, to be able to detect a partial-rescue of myofibrillar structure that may not be evident in a behavioral assay.

Immunofluorescence (IF) and Western blotting are different techniques, and Bruno1 antibody was validated for specificity in IF but not in Western blots. Figure 5L and S5 E should include muscle samples from Bru1M2.

We have added a Western Blot panel in Figure S5 D including bru1-IR, bru1M2 and samples of different wild-type tissues including abdomen, ovaries, testis and IFM.

To quantify alternative splicing or percent spliced in (PSI), primers are typically designed in the exons flanking the alternative exons. A better primer design along with PSI calculation by RT-PCR will robustly validate alternative splicing changes in different genetic background (Fig 6U and S6 U).

We do not yet have RNA-Seq data from these Rbfox1 knockdown samples to facilitate calculation of transcriptome-wide PSI values; thus, we rely on the results from our RT-PCR experiments. Our primers used to detect alternative splice events are indeed located within flanking exons or as close to the alternative exons as possible based on sequence design limitations (see schemes in Figure 6 and Figure S6). Many of the events we are detecting are complex, and not a simple “included” or “excluded” determination, and are therefore not amenable to RT-qPCR. To increase the robustness of our validation, we now provide RT-PCR gel-based quantification of exon use for the events we tested in Zasp52, Zasp66, Zasp67, wupA and Mhc (Figure 6 U-W and Figure S6 T-U).*

Reviewer #1 (Significance (Required)):

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

Understanding how muscle fiber type splicing and gene expression is regulated will conceptually move the field forward. How transcriptional and posttranscriptional programs coordinate to specify muscle fiber type gene expression is still lacking.

Place the work in the context of the existing literature (provide references, where appropriate). Multiple RNA binding proteins and splicing factors have been shown to affect muscle function along with hundreds of gene expression and splicing changes in a complex fashion. Linking phenotypes with gene expression changes is still challenging as RNA binding proteins or RBPs are multifunctional and affect the function of other regulators that are important for muscle biology. *We thank the reviewer for recognizing the conceptual advance our findings represent, as well as the complexity in the regulatory network we are seeking to understand. A detailed understanding of the coordination of transcriptional and posttranscriptional programs is enabled by our work and will be the subject of future investigation.

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

Fly genetics, alternative splicing regulation, muscle specification and function.

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.

Regulation and function of alternative splicing in muscle. I do not have a thorough knowledge of Drosophila genetics.

---

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

**Summary**

This paper reports analysis of the function of RbFox1, an RNA-binding protein, best known for roles in the regulation of alternative splicing. It uses Drosophila as its in vivo model system, one that is highly suited to the analysis in vivo of complex biological events. In general, the authors present a very thorough approach with an impressive range of molecular analysis, genetic experiments and phenotypic assays. *We thank the reviewer for recognizing the suitability of our model organism as well as the time investment and diversity of experiments that were performed in this work. We have added and revised multiple experiments during this revision, which has greatly improved the manuscript.

* The authors report that Rbfox1 is expressed in all Drosophila muscle types, and regulated in both a temporal and muscle type specific manner. Using inhibitory RNA to knock down gene function, they show that Rbfox1 is required in muscle for both viability and pupal eclosion, and contributes to both muscle development and function. A Bioinformatic approach then identifies muscle genes with Rbfox1-binding motifs. They show Rbfox1 regulates expression of both muscle structural proteins and the splicing factor Bruno1, interestingly preferentially targeting the Bruno1-RB isoform. They report functional interaction between Rbfox1 and Bruno1 and that this is expression level-dependent. Lastly, they report that Rbfox1 regulates transcription factors that control muscle gene expression.

They conclude that the effect on muscle function of RbFox1 knock down is through mis-regulation of fibre type specific gene and splice isoform expression. Moreover, "Rbfox1 functions in a fibre-type and level-dependent manner to modulate both fibrillar and tubular muscle development". They propose that it does this by "binding to 5'-UTR and 3'-UTR regions to regulate transcript levels and binding to intronic regions to promote or inhibit alternative splice events." They also suggest that Rbfox1 acts "also through hierarchical regulation of the fibre diversity pathway." They provide further evidence to the field that Rbfox1's role in muscle development is conserved.

**MAJOR COMMENTS**

Are key conclusions convincing?

In terms of presentation, I suggest ensuring a clear demarcation throughout of the evidence behind the main conclusions. This can get somewhat lost as a great deal of information is presented, including all the parallels with prior findings in other systems. I am not saying this is a major problem, just highlighting the importance of clarity. Conclusions to clearly evidence include: Rbfox1 functions in a fibre-type manner to modulate both fibrillar and tubular muscle development (e.g. L664); Rbfox1 functions in a level-dependent manner (e.g. L664); Rbfox1 functions by binding to 5'-UTR and 3'-UTR regions to regulate transcript levels (e.g. L670); Rbfox1 functions by binding to intronic regions to promote or inhibit alternative splice events" (e.g. L670); "Bru1 can regulate Rbfox1 levels in Drosophila muscle, and likely in a level-dependent manner" (L488) - Clearly evidence the level effect; "first evidence for negative regulation for fine tuning acquisition of muscle-type specific properties. Depending on its expression level, Rbfox1 can either promote or inhibit expression of" muscle regulators (L797). Lastly, the controlled stoichiometry of muscle structural proteins is known to be important, but all mechanisms are not known, so again make the supporting evidence as clear as possible for the interesting point of a role for Rbfox1 in this (e.g. L787). *Using the above comments from the reviewer as a guide, we have rewritten the manuscript, including large portions of the discussion, introduction and results. We thank the reviewer for pointing out where we could more effectively communicate our results, support our conclusions and highlight the significance of our findings.

* Should some claims be qualified as preliminary or removed?

P301 "complicated genetic recombination" - seems a bit weak to include. Either do it or don't include? *

We have removed this statement from the text.*

*

Also, see section below on "adequate replication of experiments"

Are additional exps essential? (if so realistic in terms of time and cost) None essential in my view. It depends on the authors' goals, but for the most impact of the project then following up these suggestions are possible. L369-372: mutate putative Rbfox1 binding site and ask does binding still occur or not. If it doesn't, then ask if this mutation affects the expression of the putative target gene. L775-777 "Our data thus support findings that Rbfox1 modulates transcription, but introduce a novel method of regulation, via regulating transcription factor transcript stability." It would be good to demonstrate this.

We thank the reviewer for these suggestions, and agree they are indeed interesting experiments, but beyond the scope of this manuscript. We plan to pursue the detailed molecular and biochemical mechanisms of regulation in a future project including exploring Rbfox1 binding through use of reporters, identification of direct targets via CLIP and investigation of post-transcriptional regulation of translation or NMD.*

Presented in such a way as to be reproduced

Yes

Are exps adequately replicated?

A main area I would address is the authors frequent use of "may", "tend", "trend". This is confusing the picture they present. What is statistically significant and what is not? Only the former can be used as evidence. Examples include: L170: "may display preferential exon use" - does it or doesn't it? L272: "myofibrils tended to be thicker" - were they or weren't they? L350 "wupA mRNA levels tend towards upregulation in Rbfox1-RNAi". L353 "but tended towards upregulation (Fig. S4A)" L466 "Correspondingly, we see a trend towards increased protein-level expression of Bru1-PA" L474 "both Bru1-PA and Bru1-PB tend to increase" L485 "Overexpression of Bru1 in TDT with Act79B-Gal4 also tends to reduce Rbfox1" L595 "Rbfox1-IR27286 tended towards increased exd levels in IFM (Fig. 7A)" L614 "and a trend towards increased use of Mef2-Ex20 " Also, L487 "suggesting that Bru1 can also negatively regulate Rbfox1" - one cannot use a non-significant observation to suggest something. *

We have modified the text to limit use of “may”, “tend” and “trend”, and have removed discussion of non-significant results. We thank the reviewer for the very helpful and detailed list of sentences to modify.

\*MINOR COMMENTS**

*

Although individual samples are not significant, in aggregate there is a trend….

* Specific exp issues that are easily addressable

L162: "dip in Rbfox1 expression levels around 50h APF". The Fig indicates as early as 30h. Is this significantly less than the 24h data point? Comparisons in Figure 1G that are significant based on DESeq2 differential expression analysis with an adjusted p-value L427 "this staining was lost after Rbfox1 knockdown". This conflicts with Fig 5K which says no significant difference. Again in L429 "Rbfox1 knockdown leads to a reduction of Bru1 protein levels in IFMs and TDT." Fig says no significant difference in TDT. *

We thank the reviewer for pointing out this inconsistency. We have revised the text accordingly. Our Western Blot (Figure 5L, M) and RT-PCR (Figure 5N, O) do show changes of Bru1 protein and mRNA expression levels after knockdown of Rbfox1KK110518. *

Are prior studies referenced appropriately?

This m/s is an authoritative presentation of the field as a whole with a comprehensive, impressive reference list. However, a point related to this area is one of the main things I would consider tackling. This is to have more clarity in the demarcation of what this study has found that adds to prior knowledge. It is worthwhile in itself to demonstrate the many similarities with previous work in other systems, as part of establishing the Drosophila system with all its analytical advantages for in vivo molecular genetics as an excellent model for future study in this area of research. However, the impact/strength of this m/s would be enhanced by clarity in presenting what is new to the field in all organisms. *We thank the reviewer for this suggestion. We have rewritten large portions of the manuscript, including the introduction and discussion, to improve the clarity of our findings and their importance to the field.

* Are the text and Figs clear and accurate?

TEXT

L156: more precise language than "in a pattern consistent with the myoblasts" - maybe a simple co-expression with a myoblast marker? *

We have revised this phrasing in the text. Rbfox1 expression in myoblasts was previously reported by (Usha and Shashidhara, 2010). *

L181: at first use define difference between RNAi and IR*

We use IR as an abbreviation for RNAi. In particular, we are trying to distinguish the two hairpins obtained from stock centers (27286 and KK110518) from the third, homemade RNAi hairpin, originally named UAS-dA2BP1RNAi, that was generated by Usha and Shashidhara (Usha and Shashidhara, 2010). We have better defined this in the text and methods. *

L205: maybe clearly explain the link between eclosion and tubular muscle?? *

We have added a sentence explaining the link between eclosion and tubular muscle (see Line 331).*

L231: "Sarcomeres were not significantly shorter at 90h APF with the stronger Mef2-Gal4" - not clear why this is the case when the less strong knockdown conditions have shorter sarcomeres. *

We have modified the text as well as the figure labeling to clarify that the other samples were tested in 1 d adult, while the KK110518 hairpin was tested at 90 h APF. This likely indicates that the short sarcomeres observed in 1 d adults reflect hypercontraction, which in IFM is classically first apparent after eclosion when the flies actively try to use the flight muscles. The difference in timing is due to pupal lethality of the KK110518 hairpin line, so we could not evaluate adult flies.*

L234: "classic hypercontraction mutants in IFMs display a similar phenotype" - presumably not similar to the not significantly shorter sarcomeres of the previous sentence. *

We have modified the text to clarify this statement. The change in sarcomere length from 90 h APF to 1 d adult is actually the relevant observation, as this reflects the progressive shortening of sarcomeres observed in classic hypercontraction mutants.*

L244: "90h", should be "90h APF"? *

Yes, we have modified the text.*

L273: "Myofibrils in Act88F-Gal4 mediated knockdown only showed mild defects (Fig. 3 G, H, Fig. S2 C, D) despite adult flies being flight impaired". This seems worthy of discussion - the functional defect is not due to overt structure change? *

In our own experience as well as observations included in a genome-wide RNAi screen in muscle (Schnorrer et al., 2010), there are a rather large number of knockdown conditions where few if any structural defects are observed at the level of light microscopy, but flies are completely flightless. We interpret this to reflect the narrow tuning of IFM function, where slight alterations in calcium regulation or sarcomere gene isoform expression result in dysfunction and a lack of flight. Ultrastructural evaluation might reveal defects in these cases, but the defect could also be with the dynamics of tropomyosin complex function, calcium regulation, mitochondrial function or even neuro-muscular junction structure. We have added a sentence to the text to discuss and clarify the Act88F result.*

L281 "also known as Zebra bodies" - helpful to indicate these on the Fig, they are not. *

We have added arrows to the figure to mark the Zebra bodies, and updated the figure legend.*

L282: "we were unable to attempt a rescue of these defects" - I may have missed something, but what about rescue undertaken of the defects on previous pages? *

This is the first point in the text where we introduced overexpression of Rbfox1, as preceding experiments where knockdown or using a GFP-tagged protein trap line at the endogenous locus. We have revised the sentence to focus on the overexpression phenotype with UH3-Gal4.*

L283: "Over-expression of Rbfox1 from 40h APF" - this is the first over-expression experiment, so introduce why done now (and perhaps not earlier), and also explain the use of a different Gal4 driver.*

We have reworded this section of the text. The UH3-Gal4 driver is restricted to expressing in IFM from 40h APF, so is first expressed after myofibrils have been generated and selectively in IFM. This avoids lethality observed from pan-muscle expression with Mef2-Gal4 (presumably due to severe defects in tubular muscles), and also allows us to image IFM tissue from adult flies. Later experiments with Mef2-Gal4 were performed with a later temperature shift to avoid this early lethality.*

L290 "Interestingly, both Rbfox1 knockdown and Rbfox1 over-expression produce similar hypercontraction defects" - this could be interesting, worthy of discussion/explanation. *

The most logical explanation is that Rbfox1 regulates the balance in fiber-type specific isoform expression. Loss of Rbfox1 would cause a shift in the relative ratio of the isoforms of structural genes, and overexpression of Rbfox1 would likely cause a similar shift in the opposite direction. This is supported by our RT-PCR panel, where we see co-regulation of different events with Bru1, and we see fiber-type specific difference in regulation of alternative splicing (Figure 8). Overexpression of Rbfox1 would be expected to make IFM look more like TDT, which would result in an isoform imbalance and lead to the observed hypercontraction phenotype. Interestingly, loss and overexpression of Bru1 also result in the same hypercontraction phenotype, similar to what we observe with Rbfox1. We have added a paragraph in the discussion about level-dependent regulation, to address this reviewer comment.*

P305: Bioinformatic analysis. It is not clear what is taken as a potentially interesting result. On average a specific 5 base motif is found every 1000bps - so what is being looked for? How many sites in what length or position? A range of examples are described in the next pages of the m/s. For example: L337 "Bruno1.... contains 42 intronic and 2 5'-UTR Rbfox1 binding motifs" and L591 "exd contains three Rbfox1 binding sites," *

We have redone the bioinformatic analysis completely, relying on data from oRNAment and the in-vitro determined PWM. We have also rewritten all portions of the text related to this analysis and no longer focus on the number of observed motifs in a given gene. As we unfortunately do not have RNA CLIP data, we do not know genome-wide which motifs are bound in muscle. Clustering of motifs may reflect binding, but a single, strong motif can also be bound, as we demonstrate via RIP of the wupA transcript. Thus, we identified interesting targets to test based on 1) a previously described role in the literature in myofibril assembly or contractility and 2) the presence of any Rbfox1 motif in that gene. A more elegant selection method of direct and indirect target exons will be designed for a future manuscript after integrating CLIP and mRNA-Seq data that have not yet been collected.

L315: "many of these genes have binding or catalytic activity". "catalytic activity" seems very vague.

For the original supplemental figure panel, we relied on Panther high-level ontology terms, which can unfortunately be rather vague, ie “catalytic activity” or “binding activity”. We have redone this analysis and rely rather on GO terms in the biological process and molecular function categories (Figure S3 B).

L317 "When we look in previously annotated gene lists" - be more specific. What are they?

This section of the text has been rewritten, and the “previously annotated gene lists” are described in greater detail in the Methods. *

L327 "may also affect the neuro-muscular junction" - maybe better left for the Discussion? *

We have removed this sentence from the Results.*

L333 "extradenticle (exd) and Myocyte enhancer factor 2 (Mef2) contain 3 and 7 Rbfox1 motifs," Discuss the number and position of multiple motifs found in known targets? *

We have removed the discussion of the number of binding sites for different target genes, instead incorporating this information graphically in Figure S3 C. It is not clear that the number of binding sites per gene has any influence on whether it is regulated in Rbfox1 knockdown. Thus, we have de-emphasized discussion of the number of binding sites throughout the text.*

L350 "wupA mRNA levels " - clearer to stick to using TroponinI or WupA? *

We have updated instances throughout the text to consistently refer to the protein as Troponin-I (TnI) and the gene or mRNA as wupA. *

L376 "To check whether Rbfox1 regulates some target mRNAs such as wupA....." The suggestion here is more of a further indication than a "check". *

We have reworded this section of the results to make the link between post-transcriptional regulation and our mass spectrometry results more salient.*

L544 "In IFMs, knockdown of Rbfox1 and loss of Bru1 results in...." clarify if this is the two genes separately or the two genes together? *

We have rewritten this entire section and present an expanded list of tested alternative events. We have taken care in this revision to clearly denote if the genotype is Rbfox1-IR or bru1M2 or a double knockdown background.*

L580 "Our bioinformatic analysis identified Rbfox1 binding motifs in more than 40% of transcription factors genes" - is this all TFs or just "muscle" TF genes? *

We have redone this analysis and changed this sentence in the text.*

L598, what would be the mechanism of some decrease in Rbfox1 increasing mRNA levels and more of a decrease resulting in a decrease of the mRNA? The authors say "the nature of this regulation requires further investigation". *

We have added more data to this section of the manuscript and repeated several of these experiments. After adding more biological replicates and additional data points, we have more consistent results that also demonstrate the variability in bru1 expression levels after Rbfox1 knockdown. Overall levels of bru1 assayed with a primer set in exons 14 and 17 now consistently show an increase in bru1 expression after Rbfox1 knockdown between all three hairpins (Rbfox1-RNAi, Rbfox1-IRKK110518 and Dcr2, Rbfox1-IR27286) (Figure 5 N).

The relationship between expression level of Rbfox1 and expression level of bru1 and Bru1 protein isoforms is more complex. We now report a novel splice event in the annotated isoform bru1-RB that skips exon 7, resulting in a frame shift and generation of a protein that lacks all RRM domains, which we call bru1-RBshort (Figure S5). This short isoform is preferentially used in TDT, while the long isoform encoding the full-length protein is preferentially used in IFM (Figure 5 P). Presumably, this provides a mechanism, in addition to the use of different promoters, for muscle cells to regulate expression levels of different Bru1 isoforms. Knockdown of Rbfox1 in IFM results in a significant increase in the use of the long mRNA isoform, but paradoxically a decrease in the corresponding protein isoform (Figure 5, S5). We interpret this to mean that Rbfox1 regulates alternative splicing of Bru1, and likely independently a translational/post-translational mechanism regulates the expression level of Bru1-RB. This in theory could be mediated by interaction with translational machinery, post-translational modification, increased P-granule association, etc., and given the depth and breadth of experiments (as well as the multitude of isoform-specific expression reagents) required to isolate the responsible pathway, we deem it beyond the scope of this manuscript to biochemically demonstrate this specific regulatory mechanism. *

L609 "The short 5'-UTR encoded by Mef2-Ex17". Ensure all abbreviations are defined. What does "Ex" mean here? Not straightforward to relate to the diagram in the Supplemental material that indicates the Mef2 gene has many fewer than 17 exons. In Fig7 legend too. *

We have changed “Ex” to “exon” in the text. We apologize for the confusion. We have also added a diagram to Figure 7 E of the 5’-UTR region of Mef2, and a complete diagram of the locus in Figure S3 C. Based on the current annotation, Mef2 exons are numbered 1 to 21, corresponding to at least 16 distinct regions of the genome (18 if you include the variable 3’-UTR lengths). Exons sometimes will have more than one number in the annotation if a particular splice event causes a shift in the ORF, or if alternative splice sites or poly-adenylation sites are used. Mef2 is also on the minus strand, so as exons are numbered based on the genome scaffold, the exon numbering goes in reverse (ie exon 1 is the 3’-UTR).

We strongly believe in following the numbering provided in the annotation, to increase reproducibility and transparency in working with complex gene loci for many different genes. Another researcher can go to Flybase, look-up the exon number from a given gene from a specific annotation, and get the exact location and sequence of the exons we name. It is incredibly challenging and time intensive to go through older papers and figure out which exon or splice event corresponds to those in the current annotation, and we aim to alleviate this difficulty (we illustrate this in Figure 4 A for the wupA locus, where we verified exon numbers in annotation FB2021_05 by BLASTing each individual sequence and primer provided in (Barbas et al., 1993).*

L617 "Levels of Mef2 are known to affect muscle morphogenesis but not production of different isoforms" - clarify what is meant here by "different isoforms". *

We have revised this section of the text. This statement was meant to reflect that Mef2 affects muscle morphogenesis through regulation of transcription levels, but not at the level of alternative splicing.*

L638 "Salm levels were significantly increased in IFM from Rbfox1-RNAi animals, but significantly decreased in IFMs from flies with Dcr2 enhanced Rbfox1-IR27286 or Rbfox1-IRKK110518". This is worth discussion or further analysis. Normally would expect an allelic series, with an effect becoming more apparent with increased loss-of-function. *

Dcr2, Rbfox1-IR27286 and Rbfox1-IRKK110518 produce a stronger knockdown than Rbfox1-RNAi, and indeed produce significantly decreased levels of salm, thus following the allelic series. We repeated this experiment, but obtained the same results. *

L641 "This suggests that Rbfox1 can regulated Salm". How, if there are no Rbfox1 binding sites? Deserves further analysis? *

Our new bioinformatic analysis suggests a possible answer, in that it identified possible Rbfox1 motifs in a salm exon and a site in an intron. Previously, we had focused on introns and UTR regions. In addition, using the PWM we now recover Rbfox1 binding sites of the canonical TGCATGA as well as AGCATGA sites. The intron site in salm is an AGCATGA site. Further experiments will be required to determine if Rbfox1 directly binds to salm mRNA, if it interacts with the transcriptional machinery to regulate salm expression, or if this regulation occurs through yet a different mechanism, and are beyond the scope of this manuscript.*

L674: "We found the valence of several regulatory interactions..." I'm not sure the meaning of "valence" here and elsewhere will be readily understood. *

Thank you for pointing this out. We have used a different phrasing throughout the text.*

FIGURES

Fig 1 it is difficult to see the green in A-F. Can this be improved? It is clearer in I-L. *

We have replaced the images with better examples and increased the levels to make the green channel better visible. *

Fig 2 legend (others too), say what the clusters of small black ellipses in P and Q are. *

Thank you for pointing out this oversight. All boxplots are plotted with Tukey whiskers, such that they are drawn to the 25th and 75th percentile plus 1.5 the interquartile range. Dots represent outlying datapoints outside of this range. We have added statements in the relevant figure legends, as well as a more detailed explanation in the Methods. *

Fig 3 it is not easy to see a shorter sarcomere in D, as the arrow partially obscures what is being indicated. Also, the data in G indicates that sarcomeres are not shorter in Mef2 GAL4 > KK110518, although the legend says this is shown in D. *We have rephrased the statement in the legend. The arrows are pointing to frayed or torn myofibrils.

Fig 5 legend "-J). Bru1 signal is reduced with Rbfox1-IRKK110518 (C, F, I)". Clarify that this is only in IFM. It is not significant in TDT or Abd-M.

Done.*

Fig 7 legend "quantification of the fold change in exd transcript levels" - only KK110518 in IFM is significant. *

This panel was moved to Figure S7. The relevant regions of the text and figure legend were modified to reflect that only Rbfox1-IRKK110518 results in a significant change in exd levels. C - "indicates Rbfox1 binds to Mef2 mRNA" - it is not easy to see the band.

We replaced the image and adjusted the levels to make the band more visible. D - what do the different lanes on the gel below the histogram in D correspond to? We adjusted the labeling on the figure panel. The gel is a representative image of RT-PCR results that are quantified above in the histogram.

*Suggestions that would help the presentation of their data and conclusion **

There is a lot of good, thorough work here, but overall there is the impression that some of the presentation/writing could be improved (also see the above lists on clarity and accuracy). I admire the authors for their comprehensive presentation of what has already been found out in this field. As the authors summarise, a lot is already known in many other species, so (as also indicated above) it is crucial to emphasise what new is found in this work that advances overall knowledge in this field. This can be obscured in many places where they say because of what was found in vertebrate systems we looked in Drosophila. These include: L417: "This led us to investigate if Rbfox1 might regulate Bru1 in Drosophila." L452: "and we were curious if these interactions are evolutionarily conserved in flies." L528 "Thus, we next checked if Rbfox1 and Bru1 co-regulate alternative splicing in Drosophila muscle." L677 "Moreover, as in vertebrates, Rbfox1 and Bru1 exhibit cross-regulatory interactions" L683 "Rbfox1 function in muscle development is evolutionarily conserved" L697 "Here we extend those findings and show that as in vertebrates......" L702 "our observations are consistent with observations in vertebrates" L707 "Studies from both vertebrates and C. elegans suggest that Rbfox1 modulates developmental isoform switches." L746 "We see evidence for similar regulatory interactions between Rbfox1 and the CELF1/2 homolog Bru1 in our data from Drosophila." *We thank the reviewer for this honest and helpful assessment of the manuscript. Upon rereading the original text and with the guidance of the list of sentences above, we agreed with the reviewer and we have rewritten large segments of the manuscript. In particular in the introduction and discussion, we now better emphasize what is new in our findings and how they advance overall knowledge in this field.

L185 paragraph. The knockdown series is important for the study. A lot is presented in this paragraph, especially for a non-specialist and it could be easier to follow. Perhaps present the four genetic conditions in the order of the severity of their phenotype on viability. Also, clearly state what each Gal4 driver is used for. What is the nature of the RNAi/IR lines such that Dcr2 could enhance their action? Also comment on off targets - are any predicted?

We have rewritten this paragraph as the reviewer requested. The hairpins are ordered by decreasing phenotypic severity, and we have more clearly described each Gal4 driver as well as Dicer2. This information is also available in the Methods, along with the off targets for the hairpins. KK110518 has one predicted off-target ichor, but this gene is not expressed in IFM, TDT or leg based on mRNA-Seq data. 27286 has no predicted off-targets. *

L227: "In severe examples". Be as clear as possible. Are the "severe examples" using the stronger RNAi line or are they the most severe examples with a single line? I'd suggest including the result in the main Fig rather than in the Supplemental. However, as I read more of the m/s I realise there is a great deal of important information in the Supplemental Figs, and so the case is not much stronger for this example than many others. The balance of what is included where could be looked at, because it is not straightforward for the reader to read the paper and quickly flick between the main and supplemental Figs. Later in the m/s is a substantial section that starts L450 (finishes L489) and which only refers to Supplemental Figs. L503 is another area where it is necessary, and difficult, for the reader to move between main Figs and supplemental Figs. *We have reorganized the figure panels in several figures, notably Figures 4, 5, 6, 7 and 8 and the corresponding supplementary figures, including moving panels from the supplemental figures to the main figures and generating more comprehensive quantification panels. In the specific case referenced here for Fig. S1 P and Q, we chose to keep the most representative images of the phenotype in the main figure (Fig. 2 I, N), and have reworded the text to reflect that the most severe phenotypic instances are in the supplement. As we do not have CLIP data, we chose to keep the bioinformatics analysis in the supplement and have shortened the paragraph in the results devoted to Figure S3. We hope our reorganization and rewriting have better streamlined the text and figures.

L258: - perhaps a Table summarising this and other phenotype trends with the different RNA conditions might be helpful. It gets quite difficult to follow.

We have revised the text and several figure panels to make the phenotypic trends with the different RNAi conditions easier to follow.*

Reviewer #2 (Significance (Required)):

The advance reported is mechanistic.

The authors already do a very good job of placing their work in the context of prior research (see comment is Section A).

Muscle biologists interested in its development and function will be interested in this work. More broadly, those intrigued by alternative splicing will be interested. Despite its very widespread occurrence, much about alternative splicing is still poorly understood in terms of regulation and significance. This is especially the case in vivo, and this paper uses an excellent in vivo model system (Drosophila) for the genetic and mechanistic analysis of complex biological problems. My field of expertise: cell differentiation, gene expression, muscle development, Drosophila.

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

**SUMMARY**

This manuscript characterizes the role of splicing factor Rbfox1 in Drosophila muscle and explores its ability to modulate expression of genes important for fibrillar and tubular muscle development. The authors hypothesize that Rbfox1 binds directly to 5'-UTR and 3'-UTR regions to regulate transcript levels, and to intronic regions to promote or inhibit alternative splicing events. Because some of the regulated genes encode transcriptional activators and other splicing factors such as Bru1, the effects of Rbfox1 may encompass a complex regulatory network that fine-tunes transcript levels and alternative splicing patterns that shape developing muscle. Most likely the authors' hypothesis is correct that Rbfox1 is critical for muscle development in Drosophila, but overall the interesting ideas presented here are too often based only on correlations without further experimental validation. *

We respectfully disagree with the reviewer that our hypothesis that Rbfox1 is critical for muscle development in Drosophila is based only on correlation without further experimental validation. In this manuscript we extensively characterize the knockdown phenotype of 3 RNAi hairpins against Rbfox1 as well as a GFP-tagged Rbfox1 protein in both fibrillar flight muscle and tubular abdominal and jump muscle. All hairpins produce similar phenotypes with defects in myofiber and myofibril structure and result in behavioral defects in climbing, flight and jumping, confirming this phenotype is due to loss of Rbfox1 and not a random off-target gene. We also convincingly demonstrate that Rbfox1 regulates Bru1, another splicing factor known to be critical for fibrillar specific splice events in IFM. Moreover, Rbfox1 and Bru1 genetically interact selectively in IFM and our RT-PCR data for 12 select structural genes reveals fiber-type specific alternative splicing defects regulated by Rbfox1 selectively, by Bru1 selectively, or by both Rbfox1 and Bru1. Thus, we conclude that Rbfox1 is indeed critical for muscle development, and this is the first report to demonstrate this requirement in Drosophila.*

**MAJOR COMMENTS**

The hypothesis that Rbfox1 plays an important role in regulating muscle development is based on previous studies in other species and supported by much new data in this manuscript. Initial bioinformatic analysis showed that many Drosophila genes, including 20% of all RNA-binding proteins, 40% of transcription factors, etc. have the motifs in introns or UTR regions. However, I think a deeper analysis is required. Any hexamer might be present about once every 4kb, and we do not expect all UGCAUG motifs are necessarily functional, so one might ask whether the association of Rbfox motifs with muscle development genes is statistically significant? Are the motifs conserved in other Drosophila species, which might support a functional role in muscle? Are the intronic motifs located as expected for regulatory effects, that is, proximal to alternative exons that exhibit changes in splicing when Rbfox1 expression is decreased or increased? *

We appreciate the point of the reviewer that it would be ideal to distinguish genome-wide motifs that are actually bound directly by Rbfox1 from those that are unused, but our behavioral and phenotypic characterization of the knockdown phenotype in this manuscript is also valid without this data. The most effective approach to identify direct targets is to perform cross-linking immunoprecipitation, or CLIP, but we unfortunately do not have CLIP data from Drosophila muscle and it is beyond the scope of the current study to generate this data. It is not trivial to obtain the amount of material necessary to identify tissue-specific binding sites, as we would also likely expect differences in targeting specificity between tubular and fibrillar muscle. Genome-wide analysis of the evolutionary conservation of binding site motifs is also not trivial and is beyond the scope of this paper.

Despite these limitations and to address the reviewer’s comment, we have done the following:

1. We have completely redone our bioinformatic analysis using transcriptome data from the oRNAment database (Benoit Bouvrette et al., 2020), as well as searching genome-wide for instances of the in vitro determined PWM using PWMScan, to capture possible sites in introns (Figure S3). The oRNAment database was shown to reasonably predict peaks identified in eCLIP from human cell lines, which we assume would translate to a similar predictive capacity in the Drosophila
2. We have calculated the expected distribution of Rbfox1 sites in a random gene list for Figure S3, and indeed the number of Rbfox1 sites in sarcomere genes is significantly enriched.
3. We have looked more carefully at the distribution of Rbfox1 and Bru1 motifs in the transcriptome (in the oRNAment data), and find not only that these motifs frequently occur in the same muscle phenotype genes, but also that they are closer together than is expected by chance (Fig. S4 J).
4. We marked the location of Rbfox1 and Bru1 motifs in the vicinity of select alternative splice events we tested via RT-PCR on the provided summary diagrams (Fig. 6, Fig. S6).
5. We have tested additional alternative splice events in total from 12 structural genes, and of the 9 events misregulated after Rbfox1 or Bru1 knockdown, all but 1 are flanked by Rbfox1 or Bru1 binding motifs. This indicates that the motifs are indeed located as expected for a regulatory effect. Is it possible to knock out an Rbfox motif and show that splicing of the alternative exon is altered, or regulation of transcript levels is abrogated?


The construction and mutation of reporter constructs is possible, but would take longer than the recommended revision time-frame, in particular to generate reporters that can be evaluated in vivo. We intend to address the biochemical mechanism(s) of Rbfox1 regulation with future experiments in a separate manuscript.

Also, what was the background set of genes used for the GO enrichment analysis? Genes expressed in muscle or all genes?

The background set of genes for GO enrichment (now Figure S3 B) was all annotated genes for the “all genes” label and all muscle phenotype genes for the “Muscle phenotype” label.

The data on cross regulation between Rbfox1 and Bru1 are confusing and inconsistent, since mild knockdown and stronger knockdown of Rbfox1 seem to have different effects on Bru1 expression. New data suggest that Rbfox1 can positively regulate Bru1 protein levels (Fig.5), but this seems inconsistent with the lab's earlier studies indicating opposite temporal mRNA expression profiles for Rbfox1 and Bru1 across IFM development. 


We apologize for the confusion, but the relationship between Rbfox1 and bru1 levels across IFM development has not been published previously. We previously generated that mRNA-Seq data, but presented here (now in Figure 5Q) is a new analysis of that data, specifically focused on Rbfox1 and bru1 expression. We have corrected the phrasing in the text.

To address this comment, along with points raised above by Reviewer 2, we have revised this part of the manuscript, added more data to this section of the manuscript and repeated several of these experiments. After adding more biological replicates and additional data points, we have more consistent results that also demonstrate the variability in bru1 expression levels after Rbfox1 knockdown. Overall levels of bru1 assayed with a primer set in exons 14 and 17 now consistently show an increase in bru1 expression after Rbfox1 knockdown between all three hairpins (Rbfox1-RNAi, Rbfox1-IRKK110518 and Dcr2, Rbfox1-IR27286) (Figure 5 N). This is consistent with our observations of inversely correlated mRNA levels during IFM development, as when Rbfox1 levels decrease, bru1 transcripts increase.

We agree with the reviewer that the relationship between the expression level of Rbfox1 and expression level of bru1 mRNA and Bru1 protein isoforms is more complex. We now report a novel splice event in the annotated isoform bru1-RB that skips exon 7, resulting in a frame shift and generation of a protein that lacks all RRM domains, which we call bru1-RBshort (Figure S5). Unknowingly, we had previously used a primer set from exon 7 to exon 8 as “common”, which lead to some confusion. This short isoform is preferentially used in TDT, while the long isoform encoding the full-length protein is preferentially used in IFM (Figure 5 P). Presumably, this provides a mechanism, in addition to the use of different promoters, for muscle cells to regulate expression levels of different Bru1 isoforms. Knockdown of Rbfox1 in IFM results in a significant increase in the use of the long mRNA isoform, but paradoxically a decrease in the corresponding protein isoform (Figure 5, S5). We interpret this to mean that Rbfox1 regulates alternative splicing of Bru1, and likely independently a translational/post-translational mechanism regulates the expression level of Bru1-RB. This in theory could be mediated by interaction with translational machinery, post-translational modification, increased P-granule association, etc., and given the depth and breadth of experiments (as well as the multitude of isoform-specific expression reagents) required to isolate the responsible pathway, we deem it beyond the scope of this manuscript to biochemically demonstrate this specific regulatory mechanism. *

*

Both Rbfox1 and Bru1 gene have many Rbfox motifs, but they are both large genes (>100kb) and would be expected to have many copies of all hexamers. How do we know whether any of them are functional?

We do not know if all of the Rbfox1 binding sites in the Bru1 and Rbfox1 loci are bound, but the CLIP data required to assess this is beyond the scope of this manuscript, as discussed above. We do show, however, that changes in the expression level of Rbfox1 affect the expression of Bru1 on both the mRNA transcript and protein level, and changes in the expression level of Bru1 also can affect the expression level of Rbfox1. The direct or indirect nature of this regulation remains to be fully elucidated, although we do provide RIP data showing we can detect bru1 transcript bound to Rbfox1-GFP (Figure S4 I). We have modified the text to address this comment.

Figure S4, section I, J: if changes in Bru1-RB isoform expression are correlated with Rbfox1 knockdown, it seems reasonable to test whether the Bru1-RB promoter can drive expression of GFP in an Rbfox1-dependent manner. But if I understand correctly, the assay as described on p. 19 uses the promoter region upstream of Bru1-RA. What is the logic for this experiment? It is not surprising that no effect was observed. The end result is that we have no idea whether Rbfox1 directly regulates bru1-RB. Even if it does, bru-Rb appears to be a minor component of Bru expression in IFM.

Upon reevaluating this experiment and with respect to the reviewer’s comment, we have removed it from the manuscript to avoid confusion. Our new data indicate a switch in use of the bru1-RBlong and bru1-RBshort isoforms (Figure 5 N-P), suggesting that Rbfox1 regulation is on the level of splicing.

Further experiments will be necessary to refine the indirect versus direct regulatory effects of Rbfox1 on Bru1, but our data do demonstrate that Bru1 levels are regulated in Rbfox1 knockdown conditions. We also provide a RIP experiment (Figure S4 I) showing that Rbfox1-GFP does directly bind bru1 mRNA, but we did not determine if this was isoform-specific. Multiple additional experiments would be necessary to distinguish between regulation of alternative splicing, direct binding to regulate transcript translation or stability, or transcriptional regulation via regulation of Salm, or some combination of these possible mechanisms. The data presented here are important to the field as they are the first report of isoform-specific regulation of Bru1 in muscle, even if we do not conclusively show if this regulation by Rbfox1 is direct or indirect.

In the section "Rbfox1 and Bruno1 co-regulate alternative splice events in IFMs", the data show that splicing of several genes is altered by knockdown or over-expression of Rbfox1 and Bru1. The interesting conclusion is for a complex regulatory dynamic where Rbfox1 and Bru1 co-regulate some alternative splice events and independently regulate other events in a muscle-type specific manner. However, if we are to conclude that these activities are due to direct binding of Rbfox1 and Bru1 to the adjacent introns, we need information about the location of flanking Rbfox and/or Bru1 motifs. Do upstream or downstream binding sites correlate with enhancer or silencer activity, as reported in previous studies of these splicing factors in other species? For wupA, Figure S3 shows an intronic Rbfox site, but exon 4 is not labeled so the reader cannot correlate this information with the diagram in Figure 6U.

As mentioned above, we have marked the location of Rbfox1as well as Bru1 binding motifs in the diagrams in Figure 6 and Figure S6. We have tested additional alternative splice events, and can now show events regulated only in the Rbfox1 knockdown, only after bru1 knockdown, or in double knockdown flies (Figure 8). 8 out of 9 events where we see clear changes in splicing are flanked by potential Rbfox1 or Bru1 motifs. Demonstration of direct binding and assay of genome-wide binding sites through CLIP studies is beyond the scope of this manuscript and will be pursued in the future.

The evidence that Rbfox1 directly affects expression of transcription factor Exd seems to be based only a correlation between Rbfox1 knockdown and decreased expression of Exd. The observation that binding of Rbfox1 to the Exd 3'UTR in RIP experiments further weakens the case.

We agree with the reviewer and have moved the data related to exd to the supplement (Figure 7 and S7). We still mention exd in the text as it is significantly decreased after knockdown with Rbfox1-IRKK110518, but we have removed it from larger claims of transcriptional regulation as well as from the summary in Figure 8. Also, just to note that although we failed to detect Rbfox1-GFP bound to exd, this experiment was performed with adult flies. Since Exd is functionally important early in pupal development during fate specification of the IFMs, it is possible we might detect binding to exd mRNA at a different developmental timepoint.

Similarly, there is a correlation of Rbfox1 knockdown with expression of alternative 5'UTRs in the Mef2 gene. However, the changes in UTR expression appear mostly not statistically significant. Do the authors have a model to explain what mechanism might allow Rbfox to regulate expression of alternative 5'UTRs, which would seem to be a transcriptional process?

Mef2 transcript levels are significantly increased after knockdown with Rbfox1-RNAi and decreased after overexpression of Rbfox1, and we can detect direct binding of Rbfox1-GFP to Mef2 RNA via RIP. This establishes Mef2 as a likely direct target of Rbfox1 regulation, likely through the two Rbfox1 motifs in the 3’-UTR (Figure S3 C). In addition to this regulation, we made an observation that has not been previously reported in the literature, that IFM expresses a particular isoform of Mef2 that uses a short promoter encoded by Exon 17. We see both tissue-specific use of Exon 17 (Figure 7 F) as well as developmental regulation of Exon 17 use in IFM (Figure S7 C). Surprisingly, we saw that use of exon 17 in the Mef2 promoter is altered in Rbfox1 knockdown muscle. We now provide a quantification of this data, to show the change is statistically significant. We also provide a scheme of the Mef2 locus and RT-PCR primers with exons 17, 20 and 21 labelled (Figure 7 E). We have also rewritten this section of the text to increase the impact and clarity of our finding.

For Salm, there apparently are no Rbfox motifs in the gene, and there are statistically significant but apparently inconsistent changes in Salm expression when it is knocked down in IFM by Rbfox1-RNAi (Salm increases) vs knockdown by Rbfox1-IR27286 or Rbfox1-IRKK110518 (Salm decreases). These are potentially interesting observations but more data would be needed to make stronger conclusions. How would regulation occur in the absence of Rbfox motifs?


The best explanation we can provide for why salm expression is increased with the weak hypomorph Rbfox1-RNAi condition, but decreased with the stronger hypomorph Rbfox1-IRKK110518 or Dcr2, Rbfox1-IR27286 conditions is that salm regulation is sensitive to Rbfox1 expression or activity level. We now discuss this in a new section of the discussion. We further attempted several experiments to address this question, including obtaining an endogenously tagged Salm-GFP line, as well as a UAS-Salm line (kindly provided by F. Schnorrer). Disappointingly, there is no GFP expressed in the Salm-GFP line, either live, by immunostaining or in Western Blot of multiple developmental stages, indicating that the line has fallen apart and we have not yet redone the CRISPR targeting to generate a new line. The UAS-Salm construct works (too well), in that overexpression with Mef2-Gal4 results in early lethality and we have not yet managed to optimize the experiment and obtain enough pupal muscle where we can evaluate the effect on Bru1 or Rbfox1 levels.

Our new bioinformatic analysis further revealed possible Rbfox1 motifs in a salm exon and a site in an intron. Previously, we had focused on introns and UTR regions. Now, using the in vitro determined PWM, we can recover Rbfox1 binding sites of the canonical TGCATGA as well as AGCATGA sites. The intron site in salm is an AGCATGA site. Further experiments will be required to determine if Rbfox1 directly binds to salm pre-mRNA, if it interacts with the transcriptional machinery to regulate salm expression, or if this regulation occurs through yet a different mechanism. We feel the many required experiments are beyond the scope of the current manuscript. Our data provides an experimental basis for future studies on this topic.

\*MINOR COMMENTS**

1. In several figures there is a misalignment of the transcriptional driver information with the phenotype data in the bar graphs above. Please correct the alignments to make interpretation easier. *

We have revised the layout of labels for many plots throughout the manuscript to avoid a category label associated with a genotype label at a 45-degree angle, and to make interpretation easier.

On p. 14 Brudno et al. is cited as ref for Fox motifs near muscle exons, but this paper only focused on brain-specific exons.

In addition to brain-specific exons, Brudno et al. also analyzed a set of muscle-specific exons, and thus this is the appropriate reference. For instance, from the Brudno paper, “As an additional control in some experiments we analyzed a smaller sample of muscle-specific alternative exons that were collected exactly as described above for the brain-specific exons” and “UGCAUG was also found at a high frequency downstream of a smaller group of muscle-specific exons.” Further details of the muscle-specific exon analysis can be found in (Brudno et al., 2001).

For Mef2, why do exons described as 5'UTR have numbers 17, 20, and 21? One would normally expect these to be exon 1, 2 or 1A, 1B, etc.

We rely on the Flybase annotation and numbering system to refer to exons. Per Flybase, all exons are labeled in the 5’ to 3’ direction of the sequenced genome, even for genes, such as Mef2 or wupA, that are encoded on the reverse strand. We strongly believe in following the numbering provided in the annotation, to increase reproducibility and transparency in working with complex gene loci for many different genes. Another researcher can go to Flybase, look-up the exon number from a given gene from a specific annotation, and get the exact location and sequence of the exons we name. It is incredibly challenging and time intensive to go through older papers and figure out which exon or splice event corresponds to those in the current annotation. We illustrate this in Figure 4 A for the wupA locus, where we verified exon numbers in annotation FB2021_05 by BLASTing each individual sequence and primer provided in (Barbas et al., 1993). The Mhc locus is even more complex, in particular regarding alternative 3’-UTR regions and historic versus current exon designations (Nikonova et al., 2020). For clarity and reproducibility, we therefore rely on the current Flybase designations.

Fig 8: "regulation of regulators" seems to imply the Rbfox1 is impacting transcription?? Is there precedence for this type of regulation by Rbfox1? Yes, indeed, there is precedence for Rbfox1 impacting transcription, as we presented in the Discussion. Rbfox2 is reported to interact with the Polycomb repressive complex 2 to regulate gene transcription in mouse (Wei et al., 2016) and in flies Rbfox1 interacts with transcription factors including Cubitus interruptus and Suppressor of Hairless to regulate transcription downstream of Hedgehog and Notch signaling (Shukla et al., 2017; Usha and Shashidhara, 2010). In addition, Rbfox1 regulates splicing of Mef2A and Rbfox1 and Rbfox1 cooperatively regulate splicing of Mef2D during C2C12 cell differentiation (Gao et al., 2016). Our results provide a further piece of evidence implicating Rbfox1 either directly or indirectly in transcriptional regulation as well as regulation of alternative splicing.

* Reviewer #3 (Significance (Required)):

**SIGNIFICANCE**

These studies of a major tissue-specific RNA binding protein, Rbfox1, are definitely important for our understanding of functional differences between muscle subtypes, and between muscle and nonmuscle tissues. The broad outlines of Rbfox1 alternative splicing regulation are known, but there is very little specific detail about the important targets in muscle subtypes that might help explain functional differences between subtypes. If more experimental validation can be obtained for regulation of transcript levels by binding 3'UTRs, this would also represent new information. *

We thank the reviewer for recognizing the significance of our work and our detailed analysis of Rbfox1 phenotypes in different muscle fiber-types. Experimental validation of 3’-UTR binding will be a significant time investment in terms of building and testing in-vivo reporter constructs, assaying NMD and translation effects and performing the CLIP studies necessary for identification of directly-bound 3’-UTR regions, extending beyond the scope of this manuscript and the time allotted for revision. The data we present here represent an important advance in our understanding how Rbfox1 contributes to muscle-type specific differentiation, and form the basis for future experiments to explore the molecular and biochemical mechanisms underlying this regulation. *

I am reviewing based on my experience studying alternative splicing in vertebrate systems, with an emphasis on Rbfox genes. Therefore I am unable to evaluate the functional data on different subtypes of muscle in Drosophila.

*

Reviewer Response References

Barbas, J. A., Galceran, J., Torroja, L., Prado, A. and Ferrús, A. (1993). Abnormal muscle development in the heldup3 mutant of Drosophila melanogaster is caused by a splicing defect affecting selected troponin I isoforms. Mol Cell Biol 13, 1433–1439.

Benoit Bouvrette, L. P., Bovaird, S., Blanchette, M. and Lécuyer, E. (2020). oRNAment: a database of putative RNA binding protein target sites in the transcriptomes of model species. Nucleic Acids Research 48, D166–D173.

Brudno, M., Gelfand, M. S., Spengler, S., Zorn, M., Dubchak, I. and Conboy, J. G. (2001). Computational analysis of candidate intron regulatory elements for tissue-specific alternative pre-mRNA splicing. Nucleic Acids Res 29, 2338–2348.

Damianov, A., Ying, Y., Lin, C.-H., Lee, J.-A., Tran, D., Vashisht, A. A., Bahrami-Samani, E., Xing, Y., Martin, K. C., Wohlschlegel, J. A., et al. (2016). Rbfox Proteins Regulate Splicing as Part of a Large Multiprotein Complex LASR. Cell 165, 606–619.

Gao, C., Ren, S., Lee, J.-H., Qiu, J., Chapski, D. J., Rau, C. D., Zhou, Y., Abdellatif, M., Nakano, A., Vondriska, T. M., et al. (2016). RBFox1-mediated RNA splicing regulates cardiac hypertrophy and heart failure. J Clin Invest 126, 195–206.

Nikonova, E., Kao, S.-Y. and Spletter, M. L. (2020). Contributions of alternative splicing to muscle type development and function. Semin. Cell Dev. Biol.

Nongthomba, U., Cummins, M., Clark, S., Vigoreaux, J. O. and Sparrow, J. C. (2003). Suppression of muscle hypercontraction by mutations in the myosin heavy chain gene of Drosophila melanogaster. Genetics 164, 209–222.

Schnorrer, F., Schönbauer, C., Langer, C. C. H., Dietzl, G., Novatchkova, M., Schernhuber, K., Fellner, M., Azaryan, A., Radolf, M., Stark, A., et al. (2010). Systematic genetic analysis of muscle morphogenesis and function in Drosophila. Nature 464, 287–291.

Shukla, J. P., Deshpande, G. and Shashidhara, L. S. (2017). Ataxin 2-binding protein 1 is a context-specific positive regulator of Notch signaling during neurogenesis in Drosophila melanogaster. Development 144, 905–915.

Usha, N. and Shashidhara, L. S. (2010). Interaction between Ataxin-2 Binding Protein 1 and Cubitus-interruptus during wing development in Drosophila. Dev Biol 341, 389–399.

Wei, C., Xiao, R., Chen, L., Cui, H., Zhou, Y., Xue, Y., Hu, J., Zhou, B., Tsutsui, T., Qiu, J., et al. (2016). RBFox2 Binds Nascent RNA to Globally Regulate Polycomb Complex 2 Targeting in Mammalian Genomes. Mol Cell 62, 875–889.

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

Learn more at Review Commons

---

Referee #3

Evidence, reproducibility and clarity

SUMMARY

This manuscript characterizes the role of splicing factor Rbfox1 in Drosophila muscle and explores its ability to modulate expression of genes important for fibrillar and tubular muscle development. The authors hypothesize that Rbfox1 binds directly to 5'-UTR and 3'-UTR regions to regulate transcript levels, and to intronic regions to promote or inhibit alternative splicing events. Because some of the regulated genes encode transcriptional activators and other splicing factors such as Bru1, the effects of Rbfox1 may encompass a complex regulatory network that fine-tunes transcript levels and alternative splicing patterns that shape developing muscle. Most likely the authors' hypothesis is correct that Rbfox1 is critical for muscle development in Drosophila, but overall the interesting ideas presented here are too often based only on correlations without further experimental validation.

MAJOR COMMENTS

The hypothesis that Rbfox1 plays an important role in regulating muscle development is based on previous studies in other species and supported by much new data in this manuscript. Initial bioinformatic analysis showed that many Drosophila genes, including 20% of all RNA-binding proteins, 40% of transcription factors, etc. have the motifs in introns or UTR regions. However, I think a deeper analysis is required. Any hexamer might be present about once every 4kb, and we do not expect all UGCAUG motifs are necessarily functional, so one might ask whether the association of Rbfox motifs with muscle development genes is statistically significant? Are the motifs conserved in other Drosophila species, which might support a functional role in muscle? Are the intronic motifs located as expected for regulatory effects, that is, proximal to alternative exons that exhibit changes in splicing when Rbfox1 expression is decreased or increased? Is it possible to knock out an Rbfox motif and show that splicing of the alternative exon is altered, or regulation of transcript levels is abrogated?
 Also, what was the background set of genes used for the GO enrichment analysis? Genes expressed in muscle or all genes?

1. The data on cross regulation between Rbfox1 and Bru1 are confusing and inconsistent, since mild knockdown and stronger knockdown of Rbfox1 seem to have different effects on Bru1 expression. Both Rbfox1 and Bru1 gene have many Rbfox motifs, but they are both large genes (>100kb) and would be expected to have many copies of all hexamers. How do we know whether any of them are functional? New data suggest that Rbfox1 can positively regulate Bru1 protein levels (Fig.5), but this seems inconsistent with the lab's earlier studies indicating opposite temporal mRNA expression profiles for Rbfox1 and Bru1 across IFM development. 

2. Figure S4, section I, J: if changes in Bru1-RB isoform expression are correlated with Rbfox1 knockdown, it seems reasonable to test whether the Bru1-RB promoter can drive expression of GFP in an Rbfox1-dependent manner. But if I understand correctly, the assay as described on p. 19 uses the promoter region upstream of Bru1-RA. What is the logic for this experiment? It is not surprising that no effect was observed. The end result is that we have no idea whether Rbfox1 directly regulates bru1-RB. Even if it does, bru-Rb appears to be a minor component of Bru expression in IFM.
3. In the section "Rbfox1 and Bruno1 co-regulate alternative splice events in IFMs", the data show that splicing of several genes is altered by knockdown or over-expression of Rbfox1 and Bru1. The interesting conclusion is for a complex regulatory dynamic where Rbfox1 and Bru1 co-regulate some alternative splice events and independently regulate other events in a muscle-type specific manner. However, if we are to conclude that these activities are due to direct binding of Rbfox1 and Bru1 to the adjacent introns, we need information about the location of flanking Rbfox and/or Bru1 motifs. Do upstream or downstream binding sites correlate with enhancer or silencer activity, as reported in previous studies of these splicing factors in other species? For wupA, Figure S3 shows an intronic Rbfox site, but exon 4 is not labeled so the reader cannot correlate this information with the diagram in Figure 6U.
4. The evidence that Rbfox1 directly affects expression of transcription factor Exd seems to be based only a correlation between Rbfox1 knockdown and decreased expression of Exd. The observation that binding of Rbfox1 to the Exd 3'UTR in RIP experiments further weakens the case.
5. Similarly, there is a correlation of Rbfox1 knockdown with expression of alternative 5'UTRs in the Mef2 gene. However, the changes in UTR expression appear mostly not statistically significant. Do the authors have a model to explain what mechanism might allow Rbfox to regulate expression of alternative 5'UTRs, which would seem to be a transcriptional process?
6. For Salm, there apparently are no Rbfox motifs in the gene, and there are statistically significant but apparently inconsistent changes in Salm expression when it is knocked down in IFM by Rbfox1-RNAi (Salm increases) vs knockdown by Rbfox1-IR27286 or Rbfox1-IRKK110518 (Salm decreases). These are potentially interesting observations but more data would be needed to make stronger conclusions. How would regulation occur in the absence of Rbfox motifs?


MINOR COMMENTS

1. In several figures there is a misalignment of the transcriptional driver information with the phenotype data in the bar graphs above. Please correct the alignments to make interpretation easier.
2. On p. 14 Brudno et al. is cited as ref for Fox motifs near muscle exons, but this paper only focused on brain-specific exons.
3. For Mef2, why do exons described as 5'UTR have numbers 17, 20, and 21? One would normally expect these to be exon 1, 2 or 1A, 1B, etc.

4. Fig 8: "regulation of regulators" seems to imply the Rbfox1 is impacting transcription?? Is there precedence for this type of regulation by Rbfox1?

5. The data on cross regulation between Rbfox1 and Bru1 are confusing and inconsistent, since mild knockdown and stronger knockdown of Rbfox1 seem to have different effects on Bru1 expression. Both Rbfox1 and Bru1 gene have many Rbfox motifs, but they are both large genes (>100kb) and would be expected to have many copies of all hexamers. How do we know whether any of them are functional? New data suggest that Rbfox1 can positively regulate Bru1 protein levels (Fig.5), but this seems inconsistent with the lab's earlier studies indicating opposite temporal mRNA expression profiles for Rbfox1 and Bru1 across IFM development. 


6. Figure S4, section I, J: if changes in Bru1-RB isoform expression are correlated with Rbfox1 knockdown, it seems reasonable to test whether the Bru1-RB promoter can drive expression of GFP in an Rbfox1-dependent manner. But if I understand correctly, the assay as described on p. 19 uses the promoter region upstream of Bru1-RA. What is the logic for this experiment? It is not surprising that no effect was observed. The end result is that we have no idea whether Rbfox1 directly regulates bru1-RB. Even if it does, bru-Rb appears to be a minor component of Bru expression in IFM.

7. In the section "Rbfox1 and Bruno1 co-regulate alternative splice events in IFMs", the data show that splicing of several genes is altered by knockdown or over-expression of Rbfox1 and Bru1. The interesting conclusion is for a complex regulatory dynamic where Rbfox1 and Bru1 co-regulate some alternative splice events and independently regulate other events in a muscle-type specific manner. However, if we are to conclude that these activities are due to direct binding of Rbfox1 and Bru1 to the adjacent introns, we need information about the location of flanking Rbfox and/or Bru1 motifs. Do upstream or downstream binding sites correlate with enhancer or silencer activity, as reported in previous studies of these splicing factors in other species? For wupA, Figure S3 shows an intronic Rbfox site, but exon 4 is not labeled so the reader cannot correlate this information with the diagram in Figure 6U.

8. The evidence that Rbfox1 directly affects expression of transcription factor Exd seems to be based only a correlation between Rbfox1 knockdown and decreased expression of Exd. The observation that binding of Rbfox1 to the Exd 3'UTR in RIP experiments further weakens the case.

9. Similarly, there is a correlation of Rbfox1 knockdown with expression of alternative 5'UTRs in the Mef2 gene. However, the changes in UTR expression appear mostly not statistically significant. Do the authors have a model to explain what mechanism might allow Rbfox to regulate expression of alternative 5'UTRs, which would seem to be a transcriptional process?

10. For Salm, there apparently are no Rbfox motifs in the gene, and there are statistically significant but apparently inconsistent changes in Salm expression when it is knocked down in IFM by Rbfox1-RNAi (Salm increases) vs knockdown by Rbfox1-IR27286 or Rbfox1-IRKK110518 (Salm decreases). These are potentially interesting observations but more data would be needed to make stronger conclusions. How would regulation occur in the absence of Rbfox motifs?


MINOR COMMENTS

1. In several figures there is a misalignment of the transcriptional driver information with the phenotype data in the bar graphs above. Please correct the alignments to make interpretation easier.

2. On p. 14 Brudno et al. is cited as ref for Fox motifs near muscle exons, but this paper only focused on brain-specific exons.

3. For Mef2, why do exons described as 5'UTR have numbers 17, 20, and 21? One would normally expect these to be exon 1, 2 or 1A, 1B, etc.

4. Fig 8: "regulation of regulators" seems to imply the Rbfox1 is impacting transcription?? Is there precedence for this type of regulation by Rbfox1?

Significance

SIGNIFICANCE

These studies of a major tissue-specific RNA binding protein, Rbfox1, are definitely important for our understanding of functional differences between muscle subtypes, and between muscle and nonmuscle tissues. The broad outlines of Rbfox1 alternative splicing regulation are known, but there is very little specific detail about the important targets in muscle subtypes that might help explain functional differences between subtypes. If more experimental validation can be obtained for regulation of transcript levels by binding 3'UTRs, this would also represent new information.

I am reviewing based on my experience studying alternative splicing in vertebrate systems, with an emphasis on Rbfox genes. Therefore I am unable to evaluate the functional data on different subtypes of muscle in Drosophila.

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

Learn more at Review Commons

---

Referee #2

Evidence, reproducibility and clarity

Summary

This paper reports analysis of the function of RbFox1, an RNA-binding protein, best known for roles in the regulation of alternative splicing. It uses Drosophila as its in vivo model system, one that is highly suited to the analysis in vivo of complex biological events. In general, the authors present a very thorough approach with an impressive range of molecular analysis, genetic experiments and phenotypic assays.

The authors report that Rbfox1 is expressed in all Drosophila muscle types, and regulated in both a temporal and muscle type specific manner. Using inhibitory RNA to knock down gene function, they show that Rbfox1 is required in muscle for both viability and pupal eclosion, and contributes to both muscle development and function. A Bioinformatic approach then identifies muscle genes with Rbfox1-binding motifs. They show Rbfox1 regulates expression of both muscle structural proteins and the splicing factor Bruno1, interestingly preferentially targeting the Bruno1-RB isoform. They report functional interaction between Rbfox1 and Bruno1 and that this is expression level-dependent. Lastly, they report that Rbfox1 regulates transcription factors that control muscle gene expression.

They conclude that the effect on muscle function of RbFox1 knock down is through mis-regulation of fibre type specific gene and splice isoform expression. Moreover, "Rbfox1 functions in a fibre-type and level-dependent manner to modulate both fibrillar and tubular muscle development". They propose that it does this by "binding to 5'-UTR and 3'-UTR regions to regulate transcript levels and binding to intronic regions to promote or inhibit alternative splice events." They also suggest that Rbfox1 acts "also through hierarchical regulation of the fibre diversity pathway." They provide further evidence to the field that Rbfox1's role in muscle development is conserved.

MAJOR COMMENTS

Are key conclusions convincing?

In terms of presentation, I suggest ensuring a clear demarcation throughout of the evidence behind the main conclusions. This can get somewhat lost as a great deal of information is presented, including all the parallels with prior findings in other systems. I am not saying this is a major problem, just highlighting the importance of clarity. Conclusions to clearly evidence include: Rbfox1 functions in a fibre-type manner to modulate both fibrillar and tubular muscle development (e.g. L664); Rbfox1 functions in a level-dependent manner (e.g. L664); Rbfox1 functions by binding to 5'-UTR and 3'-UTR regions to regulate transcript levels (e.g. L670); Rbfox1 functions by binding to intronic regions to promote or inhibit alternative splice events" (e.g. L670); "Bru1 can regulate Rbfox1 levels in Drosophila muscle, and likely in a level-dependent manner" (L488) - Clearly evidence the level effect; "first evidence for negative regulation for fine tuning acquisition of muscle-type specific properties. Depending on its expression level, Rbfox1 can either promote or inhibit expression of" muscle regulators (L797).

Lastly, the controlled stoichiometry of muscle structural proteins is known to be important, but all mechanisms are not known, so again make the supporting evidence as clear as possible for the interesting point of a role for Rbfox1 in this (e.g. L787).

Should some claims be qualified as preliminary or removed?

P301 "complicated genetic recombination" - seems a bit weak to include. Either do it or don't include? Also, see section below on "adequate replication of experiments"

Are additional exps essential? (if so realistic in terms of time and cost)

None essential in my view. It depends on the authors' goals, but for the most impact of the project then following up these suggestions are possible.

L369-372: mutate putative Rbfox1 binding site and ask does binding still occur or not. If it doesn't, then ask if this mutation affects the expression of the putative target gene.

L775-777 "Our data thus support findings that Rbfox1 modulates transcription, but introduce a novel method of regulation, via regulating transcription factor transcript stability." It would be good to demonstrate this.

Presented in such a way as to be reproduced

Yes

Are exps adequately replicated?

A main area I would address is the authors frequent use of "may", "tend", "trend". This is confusing the picture they present. What is statistically significant and what is not? Only the former can be used as evidence.

Examples include:

L170: "may display preferential exon use" - does it or doesn't it?

L272: "myofibrils tended to be thicker" - were they or weren't they?

L350 "wupA mRNA levels tend towards upregulation in Rbfox1-RNAi".

L353 "but tended towards upregulation (Fig. S4A)"

L466 "Correspondingly, we see a trend towards increased protein-level expression of Bru1-PA"

L474 "both Bru1-PA and Bru1-PB tend to increase"

L485 "Overexpression of Bru1 in TDT with Act79B-Gal4 also tends to reduce Rbfox1"

L595 "Rbfox1-IR27286 tended towards increased exd levels in IFM (Fig. 7A)"

L614 "and a trend towards increased use of Mef2-Ex20 "

Also, L487 "suggesting that Bru1 can also negatively regulate Rbfox1" - one cannot use a non-significant observation to suggest something.

MINOR COMMENTS

Specific exp issues that are easily addressable

L162: "dip in Rbfox1 expression levels around 50h APF". The Fig indicates as early as 30h. Is this significantly less than the 24h data point?

L427 "this staining was lost after Rbfox1 knockdown". This conflicts with Fig 5K which says no significant difference. Again in L429 "Rbfox1 knockdown leads to a reduction of Bru1 protein levels in IFMs and TDT." Fig says no significant difference in TDT.

Are prior studies referenced appropriately?

This m/s is an authoritative presentation of the field as a whole with a comprehensive, impressive reference list. However, a point related to this area is one of the main things I would consider tackling. This is to have more clarity in the demarcation of what this study has found that adds to prior knowledge. It is worthwhile in itself to demonstrate the many similarities with previous work in other systems, as part of establishing the Drosophila system with all its analytical advantages for in vivo molecular genetics as an excellent model for future study in this area of research. However, the impact/strength of this m/s would be enhanced by clarity in presenting what is new to the field in all organisms.

Are the text and Figs clear and accurate?

TEXT

L156: more precise language than "in a pattern consistent with the myoblasts" - maybe a simple co-expression with a myoblast marker?

L181: at first use define difference between RNAi and IR

L205: maybe clearly explain the link between eclosion and tubular muscle??

L231: "Sarcomeres were not significantly shorter at 90h APF with the stronger Mef2-Gal4" - not clear why this is the case when the less strong knockdown conditions have shorter sarcomeres.

L234: "classic hypercontraction mutants in IFMs display a similar phenotype" - presumably not similar to the not significantly shorter sarcomeres of the previous sentence.

L244: "90h", should be "90h APF"?

L273: "Myofibrils in Act88F-Gal4 mediated knockdown only showed mild defects (Fig. 3 G, H, Fig. S2 C, D) despite adult flies being flight impaired". This seems worthy of discussion - the functional defect is not due to overt structure change?

L281 "also known as Zebra bodies" - helpful to indicate these on the Fig, they are not.

L282: "we were unable to attempt a rescue of these defects" - I may have missed something, but what about rescue undertaken of the defects on previous pages?

L283: "Over-expression of Rbfox1 from 40h APF" - this is the first over-expression experiment, so introduce why done now (and perhaps not earlier), and also explain the use of a different Gal4 driver.

L290 "Interestingly, both Rbfox1 knockdown and Rbfox1 over-expression produce similar hypercontraction defects" - this could be interesting, worthy of discussion/explanation.

P305: Bioinformatic analysis. It is not clear what is taken as a potentially interesting result. On average a specific 5 base motif is found every 1000bps - so what is being looked for? How many sites in what length or position? A range of examples are described in the next pages of the m/s. For example: L337 "Bruno1.... contains 42 intronic and 2 5'-UTR Rbfox1 binding motifs" and L591 "exd contains three Rbfox1 binding sites,"

L315: "many of these genes have binding or catalytic activity". "catalytic activity" seems very vague.

L317 "When we look in previously annotated gene lists" - be more specific. What are they?

L327 "may also affect the neuro-muscular junction" - maybe better left for the Discussion?

L333 "extradenticle (exd) and Myocyte enhancer factor 2 (Mef2) contain 3 and 7 Rbfox1 motifs," Discuss the number and position of multiple motifs found in known targets?

L350 "wupA mRNA levels " - clearer to stick to using TroponinI or WupA?

L376 "To check whether Rbfox1 regulates some target mRNAs such as wupA....." The suggestion here is more of a further indication than a "check".

L544 "In IFMs, knockdown of Rbfox1 and loss of Bru1 results in...." clarify if this is the two genes separately or the two genes together?

L580 "Our bioinformatic analysis identified Rbfox1 binding motifs in more than 40% of transcription factors genes" - is this all TFs or just "muscle" TF genes?

L598, what would be the mechanism of some decrease in Rbfox1 increasing mRNA levels and more of a decrease resulting in a decrease of the mRNA? The authors say "the nature of this regulation requires further investigation".

L609 "The short 5'-UTR encoded by Mef2-Ex17". Ensure all abbreviations are defined. What does "Ex" mean here? Not straightforward to relate to the diagram in the Supplemental material that indicates the Mef2 gene has many fewer than 17 exons. In Fig7 legend too.

L617 "Levels of Mef2 are known to affect muscle morphogenesis but not production of different isoforms" - clarify what is meant here by "different isoforms".

L638 "Salm levels were significantly increased in IFM from Rbfox1-RNAi animals, but significantly decreased in IFMs from flies with Dcr2 enhanced Rbfox1-IR27286 or Rbfox1-IRKK110518". This is worth discussion or further analysis. Normally would expect an allelic series, with an effect becoming more apparent with increased loss-of-function.

L641 "This suggests that Rbfox1 can regulated Salm". How, if there are no Rbfox1 binding sites? Deserves further analysis?

L674: "We found the valence of several regulatory interactions..." I'm not sure the meaning of "valence" here and elsewhere will be readily understood.

FIGURES

Fig 1 it is difficult to see the green in A-F. Can this be improved? It is clearer in I-L.

Fig 2 legend (others too), say what the clusters of small black ellipses in P and Q are.

Fig 3 it is not easy to see a shorter sarcomere in D, as the arrow partially obscures what is being indicated. Also, the data in G indicates that sarcomeres are not shorter in Mef2 GAL4 > KK110518, although the legend says this is shown in D.

Fig 4 A - Western blot. Looks over-exposed. Is this in a linear response region?

Fig 5 legend "-J). Bru1 signal is reduced with Rbfox1-IRKK110518 (C, F, I)". Clarify that this is only in IFM. It is not significant in TDT or Abd-M.

Fig 7 legend "quantification of the fold change in exd transcript levels" - only KK110518 in IFM is significant. C - "indicates Rbfox1 binds to Mef2 mRNA" - it is not easy to see the band. D - what do the different lanes on the gel below the histogram in D correspond to?

Suggestions that would help the presentation of their data and conclusion

There is a lot of good, thorough work here, but overall there is the impression that some of the presentation/writing could be improved (also see the above lists on clarity and accuracy). I admire the authors for their comprehensive presentation of what has already been found out in this field. As the authors summarise, a lot is already known in many other species, so (as also indicated above) it is crucial to emphasise what new is found in this work that advances overall knowledge in this field. This can be obscured in many places where they say because of what was found in vertebrate systems we looked in Drosophila. These include:

L417: "This led us to investigate if Rbfox1 might regulate Bru1 in Drosophila."

L452: "and we were curious if these interactions are evolutionarily conserved in flies."

L528 "Thus, we next checked if Rbfox1 and Bru1 co-regulate alternative splicing in Drosophila muscle."

L677 "Moreover, as in vertebrates, Rbfox1 and Bru1 exhibit cross-regulatory interactions"

L683 "Rbfox1 function in muscle development is evolutionarily conserved"

L697 "Here we extend those findings and show that as in vertebrates......"

L702 "our observations are consistent with observations in vertebrates"

L707 "Studies from both vertebrates and C. elegans suggest that Rbfox1 modulates developmental isoform switches."

L746 "We see evidence for similar regulatory interactions between Rbfox1 and the CELF1/2 homolog Bru1 in our data from Drosophila."

L185 paragraph. The knockdown series is important for the study. A lot is presented in this paragraph, especially for a non-specialist and it could be easier to follow. Perhaps present the four genetic conditions in the order of the severity of their phenotype on viability. Also, clearly state what each Gal4 driver is used for. What is the nature of the RNAi/IR lines such that Dcr2 could enhance their action? Also comment on off targets - are any predicted?

L227: "In severe examples". Be as clear as possible. Are the "severe examples" using the stronger RNAi line or are they the most severe examples with a single line? I'd suggest including the result in the main Fig rather than in the Supplemental. However, as I read more of the m/s I realise there is a great deal of important information in the Supplemental Figs, and so the case is not much stronger for this example than many others. The balance of what is included where could be looked at, because it is not straightforward for the reader to read the paper and quickly flick between the main and supplemental Figs. Later in the m/s is a substantial section that starts L450 (finishes L489) and which only refers to Supplemental Figs. L503 is another area where it is necessary, and difficult, for the reader to move between main Figs and supplemental Figs.

L258: - perhaps a Table summarising this and other phenotype trends with the different RNA conditions might be helpful. It gets quite difficult to follow.

Significance

The advance reported is mechanistic.

The authors already do a very good job of placing their work in the context of prior research (see comment is Section A).

Muscle biologists interested in its development and function will be interested in this work. More broadly, those intrigued by alternative splicing will be interested. Despite its very widespread occurrence, much about alternative splicing is still poorly understood in terms of regulation and significance. This is especially the case in vivo, and this paper uses an excellent in vivo model system (Drosophila) for the genetic and mechanistic analysis of complex biological problems. My field of expertise: cell differentiation, gene expression, muscle development, Drosophila.

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

Rbfox proteins regulate skeletal muscle splicing and function and in this manuscript, Nikonova et.al. sought to investigate the mechanisms by which Rbfox1 promotes muscle function in Drosophila.

Using a GFP-tagged Rbfox1 line, the authors showed that Rbfox1 is expressed in all muscles examined but differentially expressed in tubular and fibrillar (IFM)muscle types, and expression is developmentally regulated. Based on RNA-seq data from isolated muscle groups, the authors showed that Rbfox1 expression is much higher in TDT (jump muscle) than IFM.

Using fly genetics authors developed tools to reduce expression of Rbfox1 at different levels and the highest levels of muscle-specific Rbfox1 knockdown was lethal and displayed eclosion defects (deGradFP > Rbfox1-IRKK110518 > Rbfox1-RNAi > Rbfox1-IR27286). Consistently, Rbfox1 knockdown flies have reduced jumping and climbing phenotypes, due to tubular muscle defect where Rbfox1 is expressed at higher levels. Rbfox1 knockdown in IFM caused flight defects which have been shown previously. Further characterization of IFM and tubular muscles demonstrated a requirement of Rbfox1 for the development of myofibrillar structures in both fibrillar (IFM) and tubular fiber-types in Drosophila. Interestingly, knockdown or overexpression of Rbfox1 displayed hypercontraction phenotypes in IFMs which is often an end result of misregulation of acto-myosin interactions which was rescued by expression of force-reduction myosin heavy chain (Mhc, P401S), in the context of Rbfox1 knockdown (the rescue experiment could not be performed with Rbfox1 overexpression due to complex genetics).

Authors also performed computation analyses of the Rbfox binding motifs in the fly genome and identified GCAUG motif in 3,312, 683, and 1184 genes in the intronic, 5'UTR, and 3'UTR, respectively. These genes are enriched for factors that play important roles in muscle function including transcription factors (exd, Mef2, Salm), RNA-binding proteins (Bru1), and structural proteins (TnI, encoded by wupA). Many of these gene transcripts and proteins are affected in flies with reduction or overexpression of Rbfox1. Using fly genetics, authors propose and test different mechanisms (co-regulation of gene targets by Rbfox1 and Bru1), and regulators of muscle function (exd, Me2, Salm) and structural proteins (TnI, Mhc, Zasp52, Strn-Mlck, Sls) by which these changes could affect the muscle function.

Overall, the characterization of Rbfox1 phenotypes and myofibrillar structure is very well elucidated, mechanisms by which Rbfox1 affects muscle function are not clear and remain largely speculative.

Major comments

1. The varying level of Rbfox1 knockdown (deGradFP > Rbfox1-IRKK110518 > Rbfox1-RNAi > Rbfox1-IR27286) was achieved by different strategies without validation at the protein level (likely due to lack of a Rbfox1 antibody). It is important to show different Rbfox1 protein level (at least with different RNAi), especially when authors propose that autoregulation of Rbfox1 causes increased level Rbfox1 transcript in case of Rbfox1-RNAi (mild knockdown). Autoregulation of Rbfox1 in mammalian cells may not be similar in flies.
2. TnI and Act88F protein levels are inversely correlated with Rbfox1 level in IFM but did not correlate with the RNA level. Using RIP authors showed that Rbfox1 was shown to bound to wupA transcripts (has Rbfox binding sites) but not Act88F transcripts (does not have Rbfox binding sites). Authors performed Rbfox1 IP and identified co-IP of components of cellular translational machinery and propose that wupA (TnI) levels are regulated by translation or NMD (non-sense mediated decay). A follow up experiment was not performed to identify the mechanism by which TnI level is regulated by Rbfox1.
3. It was known that TnI mutations (affects splice site, fliH or Mef2 binding site, Hdp-3) led to a reduction in TnI level and hypercontraction. Authors showed rescue of hypercontraction phenotype in hdp-3 background by knocking down Rbfox1, likely due to increase in wupA transcription (Mef2-dependent or independent manner). However, no rescue was observed in the fliH background. Reduced level of Rbfox1 in fliH background would be expected to cause worsening of phenotype as splicing of remaining wupA transcripts would be affected with reduced Rbfox1 level. The splicing of wupA of exon 4 is not affected in Rbfox1 knockdown (fig. 6U), it's not clear if the splicing of exon 6b1 is affected in Rbfox1 knockdown.
4. Bruno1 was identified as a co-regulator of Rbfox1 in different IFM and tubular muscle types. However, except Mhc, other Rbfox1 targets seem to be regulated by either Rbfox1 or Bruno1, not both. Analyses of RNA-seq datasets from single and double knockouts should identify additional targets to support the claim that - Rbfox1 and Bruno1 co-regulate alternative splice events in IFMs. Phenotypic changes with reduced Rbfox1 and Bruno1 double knockdowns are very severe, but the mechanistic basis of such genetic interaction resulting in synergistic phenotypes in IFMs is lacking as splicing changes in single vs double knockout is similar.
5. Rbfox1 is expressed at a high level in tubular muscle whereas Bruno1 is expressed at a high level in IFM. Rbfox1 binds to Bruno1 transcript and inversely regulates Bru1-RB level but knockdown of Bru1 does not affect Rbfox1 level (Fig. S5 G,I,J). Overexpression of Bruno1 decreased the Rbfox1 level, however, it's difficult to interpret these results as overexpression of Bruno1 may have other effects on IFM gene expression.
6. A dose-dependent effect of Rbfox1 knockdown was shown to regulate the expression of transcription factors that are important for muscle type specification and function including exd, Mef2, and Salm. However, it is not clear how Rbfox1 mechanistically regulates the expression of these transcription factors.

Minor comments

1. It is not described if the rescue of Rbfox1 knockout by expression of force-reduction myosin heavy chain (Mhc, P401S) led to rescue of phenotypes (jumping, climbing, flight).
2. Immunofluorescence (IF) and Western blotting are different techniques, and Bruno1 antibody was validated for specificity in IF but not in Western blots. Figure 5L and S5 E should include muscle samples from Bru1M2.
3. To quantify alternative splicing or percent spliced in (PSI), primers are typically designed in the exons flanking the alternative exons. A better primer design along with PSI calculation by RT-PCR will robustly validate alternative splicing changes in different genetic background (Fig 6U and S6 U).

Significance

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

Understanding how muscle fiber type splicing and gene expression is regulated will conceptually move the field forward. How transcriptional and posttranscriptional programs coordinate to specify muscle fiber type gene expression is still lacking.

Place the work in the context of the existing literature (provide references, where appropriate). Multiple RNA binding proteins and splicing factors have been shown to affect muscle function along with hundreds of gene expression and splicing changes in a complex fashion. Linking phenotypes with gene expression changes is still challenging as RNA binding proteins or RBPs are multifunctional and affect the function of other regulators that are important for muscle biology.

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

Fly genetics, alternative splicing regulation, muscle specification and function.

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.

Regulation and function of alternative splicing in muscle. I do not have a thorough knowledge of Drosophila genetics.

Note: This rebuttal 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

Reviewer 1

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

*The manuscript by Wibisana et al. describes an impressive set of experiments that analyse the NFkB response at the single-cell level, using a variety of cutting-edge techniques (live cell imaging, single-cell RNA-seq, single-molecule RNA FISH, and single-cell ATAC-seq) in chicken DT40 B-cells.

In the fist half of the paper, the authors perform a detailed characterization of the cell-to-cell variation arising from a homogeneous stimulation with various doses of anti-IgM. They observe that the NFKB TF RelA forms clear nuclear 'foci' upon stimulation in DT40 cells: this was anecdotally shown in a different cell-type by the same authors in ref 7, but (to my knowledge) has never been systematically studied. This allows them to quantitatively analyse the foci formed in response to stimulation, and they show that this is dose-dependent, heterogeneous and biomodal, and exhibits properties of cooperativity. In parallel, the authors analyse the resulting stimulus-driven changes in gene expression, first using single-cell RNA-seq, and then, elegantly, using RNA FISH, which allows them to directly compare the number of RelA foci to gene expression in individual cells. Like the RelA foci, they find that cell-to-cell gene expression is heterogeneous and bimodal (this has been described before). Interestingly, though, they are able to show that individual stimulus-responsive genes exhibit distinct patterns of cell-to-cell hetereogeneity: they can categorize 4 clusters of responding genes according to different patterns of cell-to-cell variation at distinct stimulus doses, and moreover they show that while the heterogeneity of NFKBIA arises due to bimodal expression levels, that of CD83 is simply due to broad variation between cells. Although focused on NFkB, there is a lot of information here with some important (and non-intuitive) implications that could apply to many other stimulus-driven or developmental responses that exhibit heterogeneous patterns of gene expression. A more in-depth analysis of the single-cell datasets would certainly be very worthwhile and fruitful.

In the second half of the paper, the authors attempt to use their single-cell data, alongside ATAC-seq genomic analyses, to draw inferences about how or whether the model genes NFKBIA and CD83 are regulated by super-enhancers (SEs). Both of these genes are associated with SEs that gain accessibility upon stimulation (recapitulating the authors' findings in ref 8 in a different cell-type), and the CD83 promoter exhibits co-accessibility with two regions within an adjacent SE. The authors also show that both genes are sensitive to treatment with 1.6-HD, a compound that disrupts liquid-like condensates (a characteristic that has been reported for SEs), and CD83 is sensitive to an inhibitor of Brd4 (which has been associated to SE function). However, while these findings could be considered to be suggestive of regulation by SEs, they are clearly not definitive (nor do the authors claim so).

Finally, the authors show (figure 4a-c) that while the level of stimulus-driven gene upregulation correlates with co-accessibility with both SEs and typical enhancers (TEs), the cell-to-cell heterogeneity of gene expression correlates only with co-accessibility with SEs. This would agree with a model in which SE-regulated gene regulation may generally impart heterogeneous or switch-like gene expression. *

**Specific comments**

• The experiments are adequately presented, and the authors indicate that not only the sequencing data but also the analysis code is available. Nevertheless, the methods section is rather terse, and could benefit from more detail to understand the various analyses, particularly concerning the analyses of SEs in figures 3 and S7, where it is often difficult to understand how peaks or genes are categorized.

Response: We thank the Reviewer for pointing this out and we agree that the Methods section was not described in detail, particularly in how the SEs were analyzed and categorized. Therefore, we have added more details on how SEs were categorized in the Methods section as follows:

“ Peak calling and enhancer identification from ATAC-seq data were performed using Homer v4.10.4 (http://homer.ucsd.edu/homer/) using the bam files generated from the Cell Ranger pipeline. Tag directories were created for the bam file from each condition using the “makeTagDirectory” program with the “--sspe -single -tbp 1” option. Peak calling was performed using the “findPeaks” program with the “-style super -typical -minDist 5000 -L 0 -fdr 0.0001” option. This procedure stitches peaks within 5 kb and ranks regions by their total normalized number reads and classifies TE and SE by a slope threshold of 1. Peak annotation was subsequently performed using the “annotatePeaks.pl” program with the GRCg6a.96 annotation file. The consequent peak files were merged between each stimulation condition for the SE and TE peaks separately using the “mergeBed” program of bedtools. Peak annotation was performed for the second time for the merged peaks to create the final SE and TE peaks. ATAC fold-change was then calculated between both conditions for the merged peaks separately for SE and TE. Genes associated with both SE and TE were assigned only to the SE.”

Similarly, we have added more details for other analyses in the Method section and the main sentences.

• The imaging, scRNA-seq and RNA-FISH experiments are well-presented, although the supplementary figures 4 and 5 include key results that would merit inclusion within the main figures. *

Response: We thank the Reviewer for this comment. We have included supplementary figures 4b and 5d in the main figures (new Fig. 2g) since both of these figures represent the raw data revealing the differences between smFISH counts and RNA-seq derived gene expression.

• It is strking that although all the conclusions about SEs are drawn almost exclusively from analysis of ATAC-seq data, no raw ATAC-seq data is directly shown in any figure (even in the browser snapshots of figure 4d & e). It would be important to show the actual ATAC data from which the inferences of figures 3 and 4 are drawn, especially so that it is possible to visualize the implication of a particular 'ATAC fold-change' or of 'ATAC-gained enhancers'. Response: We have added a browser snapshot of the ATAC-seq data, presenting the super-enhancer region assigned to both CD83 and NFKBIA* (new Fig. 3c).

Reviewer #1 (Significance (Required)):

• This manuscript can be considered as a follow-up of the authors' previous paper (Michida 2020, ref 8), here focusing on cell-to-cell heterogeneity rather than on the overall magnitude of the stimulus-induced response. Overall, the experiments are well-performed and bring new data to an interesting angle of gene regulation. However, the analyses presented do not seem to fully exploit the data, and the authors do not manage to present any strong conclusions, particularly relating to the possible involvement of super enhancers.

Response: To strengthen our conclusions about the possible involvement of super-enhancers in regulating heterogeneity, we performed additional analyses on the properties of the SE including the number of transcription factors, NF-κB and PU.1 binding motifs and the length of the enhancers, according to a previous report (Michida et al., 2020, Cell Rep). This was also conducted to confirm whether the ATAC-seq-based SE identification method presents results consistent with those provided by H3K27Ac-ChIP-based methods utilized in the previous study (Michida et al., 2020, Cell Rep). SEs revealed longer genomic length (new Supplementary Fig. 8a) and this length was positively correlated with the ATAC signal (new Supplementary Fig. 8b). Furthermore, gained and lost SE revealed a correlation with enhanced gene expression upregulation and downregulation, respectively, compared to TE (new Fig. 3g). We also demonstrated that SE-regulated genes have a higher Fano factor change, which is consistent with the state of an SE whether it is gained or lost (new Fig. 5a, 5b). For binding motif analysis, we observed a slightly higher PU.1 motif density at SEs (new Supplementary Fig. 11), corresponding to the results of the previous study (Michida et al., 2020, Cell Rep). Interestingly, only the density of NF-κB and not PU.1 was correlated with ATAC signal change in SE (new Fig. 4a), suggesting that those SEs were controlled by nuclear translocation of NF-κB.

As a mechanism to produce gene expression heterogeneity in phenotypically identical cells, we observed that co-accessibility, which has been reported to be concordant with genomic contacts is correlated to Fano factor change, indicating that gene expression heterogeneity possibly stems from cis-regulatory interactions. NF-κB activation has been reported to increase the heterogeneity in some genes and is attributed to the accumulation of Ser5p RNAPII (Wong et al., 2018, Cell Rep). Additionally, Ser5p RNAPII has been reported to accumulate at enhancer regions (Koch et al., 2011, Nat Struct Mol Biol), and that the accumulation of RNAPII is suggested to assist in gene expression activation through enhancer-promoter contact (Thomas et al., 2021, Mol Cell). Our results support these conclusions since co-accessibility or putative cis-regulatory interactions correlate to Fano factor changes. SE can form phase-separated transcription hubs containing multiple enhancers and/or promoters, which may enable the higher diffusion rate of active enhancers; therefore, it may induce a higher possibility of genomic DNA interactions (Gu et al., 2018, Science). In contrast, the enrichment of TATA motif has also been proposed to generate transcriptional heterogeneity (Faure et al., 2017, Cell Syst). Therefore, we examined this possibility with our data. However, we observed a higher occurrence of TATA box in genes associated with lost SE (new Supplementary Fig. 18) which might have caused gene expression heterogeneity in unstimulated cells. This heterogeneity might be due to the differences in Pol II loading intervals (Tunnacliffe & Chubb, 2020, Trends Genet) however the noise associated with gained SE is possibly generated by the fluctuation of high-order biomolecular assembly. Therefore, we believe that the source of heterogeneity in these conditions were different.

Additionally, we performed Hill function analysis to reveal the threshold behavior of gene expression in our analysis since previously gained SEs were associated with threshold gene expression (Michida et al., 2020, Cell Rep). In this study, we presented that threshold behavior in gained SE is related to motif density of NF-κB (Fig. 4d), however, threshold behavior does not seem to be related to heterogeneous gene expression.

Following these results, we concluded that NF-κB activated SE has two closely related but distinct functions for gene control: (1) enhanced heterogeneity and fold-changes and (2) switch-like expression. These are controlled by different mechanisms stemming from chromatin status: (1) frequency of cis-regulatory genomic interactions possibly mediated by phase separation and (2) cooperative binding of NF-κB to DNA. These differences were well represented by expression profiles of CD83 (higher heterogeneity and weak bimodal expression) and NFKBIA (lower heterogeneity and strong bimodal expression).

• For instance, the existence of multiple gene clusters that exhibit distinct patterns of heterogeneity implies that switch-like gene activation occurs on a per-gene basis, rather than corresponding to an all-or-nothing activation of individual cells. This would be an exciting finding, and the authors have the data to test this. Likewise, the division of heterogeneous gene expression into bimodal (like NFKBIA) or unimodal (like CD83) distributions could be a nice paradigm if systematically applied to the other 1335 differentially-expressed genes identified by the authors. * Response: We appreciate this comment. Following your comment, we analyzed the relationship between heterogeneity and bimodality (switch-like expression or high Hill coefficient) for the remaining genes. We observed that SE having a high Hill coefficient contained a higher number of NF-κB motif in SE (new Fig. 4), indicating that cooperative binding of NF-κB to DNA shaped non-linear gene expression profiles as we indicated in a previous paper (Michida et al., 2020, Cell Rep). Additionally, as described in the earlier section, we observed that heterogeneity arises from cis-regulatory genomic interaction. We compared these gene groups and observed that these properties were not completely shared (new Supplementary Fig. 15), indicating that bimodality and heterogeneity originated from different mechanisms. We assume that those differences are mediated through a combination of chromatin accessibility and the biophysical properties of NF-κB.

• In contrast, although the authors try to use their data to investigate gene regulation by SEs, these inferences are all somewhat indirect, and the authors themselves do not manage to draw any definitive conclusions. Response: We appreciate this comment. We performed the additional computational analysis and carefully interpreted the data. Additionally, we have now concluded that SEs have two major biological functions: (1) gene expression heterogeneity, which is mediated via cis-regulatory interactions (Fig. 5) and (2) bimodal gene expression, which is mediated by NF-κB binding (new Fig. 4). The latter finding has also been reported in a mouse primary B cell, albeit the mechanism causing heterogeneity was a novel conclusion of this study.

• I feel that the authors are under-selling their data here. As-is, the data represents more of a resource than a study with a clear message, but I believe that with more in-depth analysis the authors could make a much more significant advance, particularly concerning the cell-to-cell heterogeneity of gene expression. I would be very enthusiastic to review the same data again with a more detailed analysis, which I believe would enormously improve the manuscript. Response: We appreciate this comment. As described in this report and the revised manuscript, we performed a considerably detailed computational analysis and gained several novel insights to answer the question regarding the functional roles of SE. We are grateful to learn that gene expression patterns may be estimated from ATAC-seq profiles and that they may even be controlled. We hope that this Reviewer would observe the scientific value of our study and provide us with your valuable feedback on our revised manuscript.

Reviewer #2

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

Imaging and single cell sequencing analyses of super-enhancer activation mediated by NF-κB in B cells" by Wibisana et al. examined the relationship between super-enhancers, NF-κB nuclear aggregation, and target gene regulation. The authors have generated a large amount of data from fluorescent microscopy, scRNA-seq, scATAC-seq, smRNA FISH. While this is an impressive dataset in terms of diverse technically advanced methods employed, it is not clear what to take as a main conceptual advance. What could be the functional implications of observed cell-cell variability in B cell transcriptional responses to environmental stimuli? In addition to this general point, the following are specific comments that could improve the manuscript.

2. In Figure 2, smRNA FISH foci of CD83 and NFKBIA are quantified as # of spots per cell (Supplementary figure 5). But it is difficult to see in Figure 2 the colocalization of any mRNA spots with RelA foci. Ideally, it will be convincing to show by DNA FISH that these target loci are indeed located within NF-κB occupied super-enhancer puncta. Even with the current RNA FISH data, some colocalization analysis could have been performed. * Response: In Figure 2, we were unable to perform accurate colocalization analysis with the current smFISH data as the probes used by us map to exons. Moreover, we have also previously performed DNA-FISH; nevertheless, it was difficult to assess co-localization between the DNA and RelA proteins secondary to the degradation of RelA-GFP proteins. Therefore, we decided to perform intronic smRNA-FISH, which can be used to pinpoint the site of active transcription (Levesque and Raj, 2013, Nat Methods). The results, along with the quantification results, are presented in the new Fig. 2f.

• Supplementary Figure 5a shows lower correlations of # GFP-RelA foci to CD83 transcripts in comparison to NFKBIA. Even though the foci and smRNA FISH spots are derived from high resolution imaging data, we should remember that any snapshot measurements have limited information content for gene regulatory relationships. Live cell studies (for example, from the groups of Suzanne Gaudet, Kathryn Miller-Jensen, and Myong-Hee Sung) have shown that time-integrated measures (e.g. maximum fold change and area under the curve of RelA signaling time course in single cells) are better correlates to transcriptional output of target genes (Lee REC et al 2014 Mol Cell; Wong VC et al 2019 Biophysical J; Sung MH et al. 2014 Science Signal; Martin EW et al. 2020 Science Signal). *

Response: We thank the Reviewer for this valuable comment. One of the reasons for a lower correlation between GFP-RelA foci and CD83 transcripts compared to NFKBIA may be the difference in expression timing of CD83 and NFKBIA and the timing of nuclear localization of GFP-RelA. RelA localizes in the nucleus 10−30 mins after cell stimulation, and NFKBIA is an early responsive gene, however, CD83 is expressed later (new Supplementary Fig. 17). Therefore, this time difference possibly affects correlation accuracy. Although we agree that high-throughput time-course measurement of RelA-GFP combined with smFISH measurements, such as that reported in Wong VC et al., 2019, will be ideal, it is technically difficult since DT40 are suspension cells and the smFISH protocol requires multiple washing and centrifugation steps. Thus, with this experimental setup, we were unable to perform the time-course analysis.

Nonetheless, we measured the time-course foci formation at the same single-cells (new Supplementary Fig. 1b) and observed that it effectively represents Figure 1a, which is a snapshot of the population dynamics of RelA foci across time. Additionally, the observed dynamics, which revealed a steep initial increase and slight decrease with time, effectively recapitulates the previous reports (Lee et al., 2014, Mol Cell; Wong et al., 2019, Biophys J).

In our analysis, we performed imaging analysis to demonstrate that NF-κB foci formation is switch-like, and this formation might be involved in the formation of phase-separated condensates enhancing DNA to DNA contact. The number of foci may depend upon the intracellular concentration of NF-κB, and fold change in the RelA signal may be correlated with gene expression as previously reported (Lee et al., 2014, Mol Cell; Wong et al., 2019, Biophys J). However, there is another report presenting that promote/enhancer proximity is not related to gene expression (Alexander et al. 2019, eLife). Although we were unable to perform this analysis owing to the limitations stated above, we tried to find the relationship between RelA foci and gene expression by performing biochemical perturbations (Fig 1e-f, Fig 5h) and presented that these foci are related to gene expression.

• The analyses have been performed using DT40 cells. In the Methods section, no description was provided about what type of B cells DT40 is, even though few outside of the field may not know that the cells were immortalized from chicken. This is an important consideration, because some nuclear bodies and genome organization features are different between host species and they also depend on whether the cells are primary or transformed. Because the authors do not discuss this point, it seems possible that the findings about NF-κB aggregates and super-enhancers may not necessarily hold true for primary B cells. *

Response: We thank the Reviewer for pointing out these issues. We have added the following description on DT40 cells in the Methods section describing that DT40 cells are chicken bursal lymphoma cells.

DT40 B lymphocytes have been widely used as a B cell model for studying B cell receptor signaling (Mori et al., 2002, J. Exp. Med.; Patterson et al., 2002, Cell; Saeki et al., 2003, EMBO J.) due to its high gene targeting efficiency. We also previously confirmed that anti-IgM stimulation induces the NF-κB signaling pathway in mouse primary splenic B cells and DT40 and that the signaling molecules and dynamics in these cells are well conserved (Shinohara et al., 2014, Science; Shinohara et al., 2016, Sci. Rep.; Inoue et al., 2016, NPJ Syst. Biol. Appl.). However, we understand the Reviewer’s concerns. Therefore, we have provided the track view of primary B cell ATAC-seq data to demonstrate that the chromatin accessibility changes upon anti-IgM stimulation in CD83 and NFKBIA were similarly observed in primary B cell data (new Supplementary Fig. 9b) and that the upregulation and association with SE of CD83 and NFKBIA were also observed in primary B cell (new Supplementary Fig. 9a).

• Similarly, the GFP-RelA expressing DT40 cell generation should be described with more detail (beyond "provided by ..."). N-terminal or C-terminal fusion? Did the fusion construct contain an artificial promoter (e.g. CMV) or an upstream fragment of the genomic Rela locus (chicken or human)? Methods of transfection and cloning of stable lines? These choices affect the interpretation of the data, so they must be fully described and justified. *

Response: We thank you for pointing this out. We have added the following details on the RelA-GFP construct in the Methods section:

Mouse RelA-eGFP with eGFP on the C terminal was cloned into a pGAP vector containing Ecogpt resistance gene targeting endogenous GAPDH locus. This construct was further electroporated into wild-type cells and selected using Ecogpt to produce RelA-GFP-expressing DT40 cells.

• DT40 cells were cultured in 39 degrees. Michael White and colleagues have shown that high temperatures can alter NF-kappaB dynamics and function (https://www.pnas.org/content/115/22/E5243). Did the authors try lower temperatures to ascertain that the NF-kB aggregates and other major findings are still observed in 37 degrees? *

Response: We performed the experiments at 39 degrees to mimic the natural body temperature of chicken since DT40 cells were derived from chicken bursal lymphoma (Saribasak and Arikawa, 2006, Subcell Biochem). Previously, we cultured DT40 cells at 37 degrees and observed that the cell growth was inhibited, and thus, we believed that it was not ideal to perform experiments of DT40 cells at 37 degrees.

Reviewer #2 (Significance (Required)):

It is not clear what to take as a main conceptual advance.

Response: Considering the original manuscript, we agree with the Reviewer on the lack of strong emphasis on the conclusions of our study. Therefore, in this revised manuscript, we have focused on the comprehensive mechanism of heterogeneity and switch-like activation in gene expression control. As we described in the comments to Reviewer #1, we performed an additional in-depth computational analysis on SE and TE. Consequently, we demonstrated that enhanced heterogeneity and expression fold-changes mediated by SE are defined by the number of cis-regulatory genomic interactions in open chromatin regions (Figure 5), however, switch-like expression (bimodal patterns) is determined by the number of NF-κB binding in SE (new Figure 4). The latter finding has also been reported in a mouse primary B cell in our previous study (Michida et al. 2020, Cell Rep.). However, the mechanism causing heterogeneity is a novel conclusion obtained in this study. We also concluded that these similar, albeit quantitatively and slightly different characteristics in gene control can be achieved through a combination of chromatin accessibility of host cells and biophysical properties of NF-κB molecule, which is involved in phase separation.

What could be the functional implications of observed cell-cell variability in B cell transcriptional responses to environmental stimuli?

Response: We performed gene ontology analysis to reveal how the heterogeneously expressed genes (cluster 4) (Fig. 2d) presented enrichment for immune-related functions (Supplementary Fig. 5b). This result supports a previous study, which stated that variability in gene expression is related to function (Osorio et al., 2019, Cells).

This discussion is incorporated in the manuscript as follows:

“We observed that genes with an increased heterogeneity upon increasing stimulation dose are enriched with cell-type-specific immune regulatory genes (Supplementary Fig. 5b), supporting a previous report where heterogeneity in gene expression is tied to biological functions and may be used by cells as a bet-hedging or a response distribution mechanism (Osorio et al., 2019, Cells), where cells exhibit heterogeneity to enable response to changing environment and also allowing dose-dependent fractional activation respectively. This was observed in CD83, a B cell activation marker, demonstrating the involvement of heterogeneity in B cell development.”

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

"Imaging and single cell sequencing analyses of super-enhancer activation mediated by NF-κB in B cells" by Wibisana et al. examined the relationship between super-enhancers, NF-κB nuclear aggregation, and target gene regulation. The authors have generated a large amount of data from fluorescent microscopy, scRNA-seq, scATAC-seq, smRNA FISH. While this is an impressive dataset in terms of diverse technically advanced methods employed, it is not clear what to take as a main conceptual advance. What could be the functional implications of observed cell-cell variability in B cell transcriptional responses to environmental stimuli? In addition to this general point, the following are specific comments that could improve the manuscript.

1. In Figure 2, smRNA FISH foci of CD83 and NFKBIA are quantified as # of spots per cell (Supplementary figure 5). But it is difficult to see in Figure 2 the colocalization of any mRNA spots with RelA foci. Ideally, it will be convincing to show by DNA FISH that these target loci are indeed located within NF-κB occupied super-enhancer puncta. Even with the current RNA FISH data, some colocalization analysis could have been performed.
2. Supplementary Figure 5a shows lower correlations of # GFP-RelA foci to CD83 transcripts in comparison to NFKBIA. Even though the foci and smRNA FISH spots are derived from high resolution imaging data, we should remember that any snapshot measurements have limited information content for gene regulatory relationships. Live cell studies (for example, from the groups of Suzanne Gaudet, Kathryn Miller-Jensen, and Myong-Hee Sung) have shown that time-integrated measures (e.g. maximum fold change and area under the curve of RelA signaling time course in single cells) are better correlates to transcriptional output of target genes (Lee REC et al 2014 Mol Cell; Wong VC et al 2019 Biophysical J; Sung MH et al. 2014 Science Signal; Martin EW et al. 2020 Science Signal).
3. The analyses have been performed using DT40 cells. In the Methods section, no description was provided about what type of B cells DT40 is, even though few outside of the field may not know that the cells were immortalized from chicken. This is an important consideration, because some nuclear bodies and genome organization features are different between host species and they also depend on whether the cells are primary or transformed. Because the authors do not discuss this point, it seems possible that the findings about NF-κB aggregates and super-enhancers may not necessarily hold true for primary B cells.
4. Similarly, the GFP-RelA expressing DT40 cell generation should be described with more detail (beyond "provided by ..."). N-terminal or C-terminal fusion? Did the fusion construct contain an artificial promoter (e.g. CMV) or an upstream fragment of the genomic Rela locus (chicken or human)? Methods of transfection and cloning of stable lines? These choices affect the interpretation of the data, so they must be fully described and justified.
5. DT40 cells were cultured in 39 degrees. Michael White and colleagues have shown that high temperatures can alter NF-kappaB dynamics and function (https://www.pnas.org/content/115/22/E5243). Did the authors try lower temperatures to ascertain that the NF-kB aggregates and other major findings are still observed in 37 degrees?

Significance

It is not clear what to take as a main conceptual advance.

What could be the functional implications of observed cell-cell variability in B cell transcriptional responses to environmental stimuli?

Referee cross-commenting

I concur with Reviewer #1's comments about systematic grouping of 1335 differentially expressed genes based on heterogeneity, and also about showing raw ATAC-seq data tracks and plots. We both commented that the study lacks a significant conclusion in its current form.