A genetic screen to uncover molecular mechanisms underlying lipid transfer protein function at membrane contact sites and neurodegeneration

  1. Shirish Mishra
  2. Vaishnavi Manohar
  3. Shabnam Chandel
  4. Tejaswini Manoj
  5. Subhodeep Bhattacharya
  6. Nidhi Hegde
  7. Vaisaly R Nath
  8. Harini Krishnan
  9. Corinne Wendling
  10. Thomas Di Mattia
  11. Arthur Martinet
  12. Prasanth Chimata
  13. Fabien Alpy
  14. Padinjat Raghu

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Lipid transfer proteins mediate the transfer of lipids between organelle membranes in eukaryotes and loss of function in these has been linked to neurodegenerative disorders. However, the mechanism by which loss of lipid transfer protein function leads to neurodegeneration is not understood. In Drosophila photoreceptors, depletion of Retinal Degeneration B (RDGB), a phosphatidylinositol transfer protein localized to endoplasmic reticulum-plasma membrane contact sites leads to defective phototransduction and retinal degeneration but the mechanism by which RDGB function is regulated and the process by which loss of this activity leads to retinal degeneration is not understood. RDGB is localized to membrane contact sites (MCS) and this depends in the interaction of its FFAT motif with the ER integral protein VAP. To identify regulators of RDGB function in vivo , we depleted more than 300 VAP interacting proteins and identified a set of 52 suppressors of rdgB . The molecular identity of these suppressors indicates a role for novel lipids in regulating RDGB function and for transcriptional and ubiquitination processes in mediating retinal degeneration in rdgB . The human homologs of several of these molecules have been implicated in neurodevelopmental diseases underscoring the importance of VAP mediated processes in these disorders.

Article activity feed

  1. Review Commons

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

    Learn more at Review Commons

    Reply to the reviewers

    We thank the reviewers for their comments and constructive criticisms of the manuscript. We thank them for positive comments on the high quality of the genetic screen and recognizing our contributions in respect of RDGB work in Drosophila.

    In general, there is one important comment by the reviewers about the candidates identified in the proteomic screen as potential VAP interactors which have then been tested in the genetic screen. Reviewers have noted that many proteins in the proteomics study identified as VAP interactors do not have the classical FFAT motif that mediates VAP interaction. Therefore, what is the significance of such genes?

    Response: It is important to reflect on the fact that while VAP interacts with FFAT motifs in proteins , a VAP immunoprecipitation will identify two classes of proteins (i) those with classical FFAT motifs (ii) those proteins without FFAT motifs that interact indirectly with VAP via proteins which themselves have FFAT motifs. We have already depicted this in Fig 2A as category C proteins.

    We believe that the in vivo genetic screen does in fact serve the specific purpose of testing the functional significance of such non-FFAT containing proteins identified in the proteomic screen by functional validation of their ability to modulate rdgB degeneration.

    Key modifications to the text and a few experiments planned are listed in the next sections against pointwise response to reviewer comments. We believe that this will strengthen the manuscript.

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

    Mishra and colleagues have conducted a large genetic screen to identify modulators of a Drosophila model for retinal degeneration. Using biomolecular techniques, they selected a few hundred proteins that interact with an ER bound protein VAP, to further test them in the retinal degeneration model. This was done by downregulating their expression using interference RNA (RNAi). Degeneration was first measured through a pseudopupil analysis, then suppressors of degeneration were further tested in a different retinal degeneration model, and finally through an ERG experiment. Finally, they focused on a strong suppressor of degeneration, dCert, by using a mutant allele to confirm the findings from the RNAi. The results suggest that a handful of these candidates are suppressors in this model of retinal degeneration, and identify at which stage of retinal degeneration these proteins may be involved in. These proteins may have a significance in forms of neurodegeneration.

    Major comment:

    • Perhaps a larger number of replicates could be done in the optical neutralization experiment as well as in the ERG. Figure 4.A(i) and (ii), please clearly state n values. I would suggest this as optional, but perhaps ut could help to increase n?

    Each optical neutralization experiment was done using 5 independent animals and 10 ommatidia were scored from each animal. For the optical neutralization experiment in 4.A(i) and (ii) we did 5 independent animals with 10 ommatidia/ animal for the statistical score. Based on our past experience, this number is sufficient to capture the intra-ommatidial variability in each eye and the inter animal variability between animals. This information will be added to the figure legend.

    For the ERGs minimum 5 animals were used per experiment which is already mentioned in the figure legend and conforms to the standards of analysis in the field for such experiments.

    • For human orthologs (Table 1), it could be worthwhile to add alignment scores between fly and human?

    We will add the table with the alignment score

    Minor comment:

    • Clarify the purpose in focusing on dCert specifically in the last results section and discussion

    • Several typos

    • Affect vs effect

    • Following the initial genetic screen, it was necessary to characterize a genes to understand in detail the temporal and spatial aspects of it role in modulating degeneration. Dcert was chosen for several reasons (i) a classical germ line mutant allele was available (ii) Prior papers had established its role as a protein that functions at contact sites. We will clarify our purpose of including dCert as proof of principle in the discussion part.

    -Typos will be corrected.

    Reviewer #1 (Significance (Required)):

    General assessment:

    The main significance of this study comes from the focus on proteins that are known to interact with VAP. This implies that the suppressors of degeneration that they have identified in the RdgB9 model may have an effect in other neurodegenerative models, namely in ALS models. This could have a very high significant potential in therapeutic avenues for neurodegenerative diseases.

    Among six candidates that had an effect with knocked down through RNAi, they pursued a single one (dCert) as proof of principle. It would help to add a justification for this choice in the main text and whether the authors have performed or intend to perform experiments using mutant forms of the other candidate proteins.

    Although six candidate genes were available for analysis, there were no mutants available in two of them (SET and CG3071). Mutants in Yeti are homozygous lethal making it difficult to work on it in this setting. However a viable mutant in APC is now available and a CG9205 CRISPR germ line deletion mutant has recently been generated in our lab. We will use these two alleles to test their, interaction with rdgB like we did for dcert. Since dCert and CG9205 have membrane interacting domains we prefer to focused on these two genes for this study as proof of principle.

    The work from Raghu and his team have been leading the research surrounding this model of degeneration in Drosophila. This study naturally further extends their field of research, identifying more candidates that modulate this form of degeneration, and helping elucidate the pathways leading to cellular degeneration.

    These results will be of high interest for specialized researchers studying the molecular pathways that lead to cellular degeneration, both in the context of retinal degeneration as well as neurodegeneration. Specifically, researchers that may be interested in these candidate proteins and how they may play a role in the pathogenesis of various degenerative diseases.

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

    Summary: In this study, Mishra and colleagues combine proteomic techniques with Drosophila genetics to identify interactors of the endoplasmic reticulum (ER) protein VAP and analyse their indirect implication in the light-dependent retinal degeneration phenotype caused by misfunction of the lipid transfer protein RdgB, another VAP interactor. This is an ideal model system to study neurodegeneration in vivo via which the authors aim to further understand the molecular mechanisms involved in VAP-RdgB's function in maintenance of membrane lipid homeostasis. Of an initial list of 403 VAP mammalian interactors found via Immunoprecipitation - Mass Spectromety performed in a human cell line, 52 homologous Drosophila genes are found to suppress the RdgB degeneration phenotype upon knockdown. The authors then test a series of genetic interactions to dissect the potential role of these genes in the degeneration phenotype and identify six genes that likely suppress the phenotype by directly acting on the same molecular processes as RdgB, two of which (Cert and CG9205) could have a direct mechanistic role in the lipid homeostasis maintained by the VAP-RdgB complex.

    Major comments:

    • The model suggested by the authors in Fig 2A is one by which the tested genes influence the VAP-RdgB function via direct binding to VAP. The direct interaction with VAP is a core aspect of the study as the IP-MS experiment biases the list of genes tested in Drosophila, and this list is often referred to as "VAP interacting proteins". However, the interactions between the proteins coded by the most relevant genes in the genetic screen and VAP are not tested. In fact, of the six genes that are found to likely modulate the same processes as RdgB, four probably do so via affecting gene expression, as discussed by the authors, therefore making it unlikely that they are true VAP interactors (unless they shuttle from the ER to the nucleus). Additionally, it would seem that most of the 52 genes found to suppress the degeneration phenotype are not necessarily VAP interactors either as only a handful of genes in this list have predicted strong FFAT motifs. In general, the authors should provide additional comments/evidence on the interaction, or likelihood of the interaction, of these proteins with VAP, that should include: o The IP-MS data should be made available (at least on my end, in the current submission the supplementary lists available cover only the 52 genes found to suppress the phenotype, I found the list of 403 genes in the biorXiv submission but this also does not include data on the enrichment of each hit in the IP-MS). It should be made clear how enriched were each of the proteins in the analysis, and how high the most relevant genes in the genetic screen rank within this interactor list.

    We will provide more details on the MS data. We can add the ratio (VAP WT vs VAP mutant) and the PSM, number of peptides etc in the table, and we will deposit the raw data on PRIDE repository.

    It is true that some of the genetic interactors, especially those that do not have an obvious FFAT motif likely influence the retinal degeneration phenotype either by indirect interactions with VAP or functional interactions rather than structural interaction. This point can be emphasized in the discussion. It is important to note that this is in some senses a good reason to couple of protein interaction screen with a genetic screen., the genetic screen sometimes uncovering functional interactions via indirect mechanism. The likely mode of interaction can be made clear in a revision.

    o The authors should provide further detail on the rationale behind defining the list of 403 genes to be tested (for example, what was the threshold enrichment considered as interaction). Also in relation to this, the authors should at least provide speculation as to why more than half of the 403 genes defined do not have an FFAT motif, despite the fact that the proteomic data was normalized to a non-FFAT motif binding mutant of VAP which should be capable of maintaining non-FFAT mediated interactions of the protein.

    As the reviewer correctly mentions, the 403 genes from the proteomics screen used for the genetic screen all satisfied the criterion of being differentially enriched in binding seen to wild type VAP but not the non-binding version of VAP. This is despite that fact that many of these proteins do not have an identifiable FFAT motif. The reason for this is most likely that the candidates without the FFAT motif most likely bind to VAP indirectly via a protein which itself has an FFAT motif. This is depicted in Figure 2A. This has already been explained in the text but this can be elaborated further.

    The rationale for including candidates from the proteomic screen without an FFAT motif in the genetic screen is that we were interested in all candidates that might influence RDGB function whether they are direct or indirect binders of VAP.

    o The authors should acknowledge the limitations of their experimental design in regards to identifying real interactors of VAP in Drosophila and avoid referring to this set of genes as "VAP interacting proteins" and rather use a more accurate description such as "proteins enriched in the IP-MS" or at least "potential VAP interacting proteins".

    We agree that the use of ‘potential VAP interactors’ may be more appropriate.

    o Testing interaction between VAP and all the 52 genes found to suppress the phenotype would be a huge amount of work. But the finding most relevant to the initial premise of the study (i.e. "molecular mechanisms underlying lipid transfer protein function at membrane contact sites") is that Cert (a lipid transfer protein) and CG9205 (fly homolog of a mammalian lipid transfer protein) influence RdgB function. Demonstrating an interaction between these proteins and VAP would argue for the experimental design and support the hypothesis and model of the study. Cert is already a well established interactor of VAP, hence the authors would not need to add anything regarding this protein. Is CG9205 expected to be also a true interactor of VAP? Biochemical experiments could be used to test this idea, or even recently developed in silico modelling of interactions (i.e. AlphaFold Multimer) could be of help. If no interaction is observed/expected this should also be pointed out in the manuscript. Optionally, showing localization of Cert or CG9205 to the ER-PM interface would also greatly support the model of VAP-RdgB regulation suggested by the authors.

    We agree that further experimental evidence to support the interaction of CG9205 would add useful information. This can be attempted by co-IP or by in silico methods such as Alpha fold multimer.

    Minor comments:

    • In the model shown in Fig2A, it would seem that many proteins can bind VAP in addition to RdgB, however, VAP proteins have only one FFAT binding pocket. This model would only be possible if oligomerization of VAP is considered (oligomerization of VAP has been reported to occur, see for example PMID:20207736). The model should be redrawn considering this fact.

    We agree.

    • In line 161 of the text VPS13D is mentioned, however VPS13C is the gene indicated in Fig 1D.

    The text will be corrected.

    Reviewer #2 (Significance (Required)):

    Previous studies by some of the authors and others have shown that RdgB can transfer lipids between the ER, to which it binds via VAP, and the plasma membrane (PM), and is required for proper replenishment of PI(4,5)P2 in the PM which is in turn necessary for sustained PLC signaling in Drosophila photoreceptors. Lack of RdgB leads to light-dependent degeneration of the retina, and hence it is utilized by the authors as a model for neurodegeneration. Given the clear phenotype of RdgB loss-of-function and the ease of Drosophila genetics, this system represents an ideal model to perform screens for the identification of new genes involved in maintaining neuronal lipid homeostasis required for proper function of the photoreceptors in vivo, and this aspect is the main strength of this study. Importantly, the use of this system could also shed light on the mechanisms behind human neurodegenerative disorders, as many of these involve dysregulation of lipid signaling and lipid transfer at membrane contact sites. A novel and interesting finding is the identification of another lipid transfer protein, Cert, to be involved in the degeneration of photoreceptors.

    The main limitation lies in the experimental design proposed by the authors to define the genes that are studied in their system. These are identified as potential interactors of human VAP in a mammalian cell line. Despite the fact that VAP is a highly conserved protein, and the genes identified are present in Drosophila as well, there is no evidence that these interactions are in fact occurring in Drosophila photoreceptors, and in fact, based on the function and the lack of VAP-binding motif in many of the 52 genes identified to have an effect on the RdgB phenotype, it is likely that many of the interactions are purely genetic and indirect, and that the modulation of the phenotype could in fact be due to a wide variety of factors (including, as discussed by the authors, gene expression, post-translational modifications, trafficking of proteins, etc) unrelated to mechanisms of VAP-RdgB mediated lipid transfer at ER-PM membrane contact sites.

    A more unbiased screen could have been carried out to identify VAP interactors involved in this degeneration phenotype by testing all of the FFAT or FFAT-related motif containing proteins. Due to this initial bias in the selection of genes to be tested, it is possible that other important VAP interactors that play a role at the ER-PM interface of photoreceptors have not been identified.

    We agree that an alternative approach might have been to perform the VAP interaction proteomics in fly photoreceptors rather than start with a proteomics data set from mammalian cells. At this late stage in the project this will not be a feasible approach. However, we could consider testing any FFAT containing proteins, identified bioinformatically in the fly genome in the future.

    An initial bioinformatics analysis has revelated that there are only 51 genes in the entire fly genome with an identifiable conventional FFAT motif. Of these 7 are already part of the genetic screen already completed. Of the remaining 44 genes, 11 show no expression in the eye and 9 show very low expression. Thus, using the approach suggested by the reviewer ca. 24 genes with FFAT motifs could have been missed and therefore could be screened, subject to genetic tools, i.e RNAi lines being available for these.

    This study provides great functional advance in the understanding of genes implicated in photoreceptor degeneration, and in those regards it is a great resource for a specialized audience, as it enables further characterization by others of the different processes implicated in this neurodegeneration phenotype. However, the advance is small in regards to the core mechanism of RdgB function at VAP-mediated ER-PM, which was the main aim of the article and the most broadly interesting aspect of the study. Many of the VAP-interacting proteins identified in the proteomic approach were already expected to be VAP interactors as they contain FFAT motifs, and these FFAT-containing proteins do not seem to have a major role in VAP-RdgB maintenance of neuronal lipid homeostasis, with the exception of Cert. The implication of Cert in RdgB-mediated lipid homeostasis is certainly interesting as it touches on a current topic in the field of membrane contact sites related to how the multiple interactions of VAP, a universal contact site adaptor at the ER, are regulated and influenced by each other.

    Reviewer's field of expertise: Lipid transfer at membrane contact sites; membrane lipid homeostasis in neurons. All of the Drosophila data seem to be of good general quality to me, but I do not have any expertise in Drosophila work.

  2. Review Commons

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

    Learn more at Review Commons

    Referee #2

    Evidence, reproducibility and clarity

    Summary: In this study, Mishra and colleagues combine proteomic techniques with Drosophila genetics to identify interactors of the endoplasmic reticulum (ER) protein VAP and analyse their indirect implication in the light-dependent retinal degeneration phenotype caused by misfunction of the lipid transfer protein RdgB, another VAP interactor. This is an ideal model system to study neurodegeneration in vivo via which the authors aim to further understand the molecular mechanisms involved in VAP-RdgB's function in maintenance of membrane lipid homeostasis. Of an initial list of 403 VAP mammalian interactors found via Immunoprecipitation - Mass Spectromety performed in a human cell line, 52 homologous Drosophila genes are found to suppress the RdgB degeneration phenotype upon knockdown. The authors then test a series of genetic interactions to dissect the potential role of these genes in the degeneration phenotype and identify six genes that likely suppress the phenotype by directly acting on the same molecular processes as RdgB, two of which (Cert and CG9205) could have a direct mechanistic role in the lipid homeostasis maintained by the VAP-RdgB complex.

    Major comments:

    • The model suggested by the authors in Fig 2A is one by which the tested genes influence the VAP-RdgB function via direct binding to VAP. The direct interaction with VAP is a core aspect of the study as the IP-MS experiment biases the list of genes tested in Drosophila, and this list is often referred to as "VAP interacting proteins". However, the interactions between the proteins coded by the most relevant genes in the genetic screen and VAP are not tested. In fact, of the six genes that are found to likely modulate the same processes as RdgB, four probably do so via affecting gene expression, as discussed by the authors, therefore making it unlikely that they are true VAP interactors (unless they shuttle from the ER to the nucleus). Additionally, it would seem that most of the 52 genes found to suppress the degeneration phenotype are not necessarily VAP interactors either as only a handful of genes in this list have predicted strong FFAT motifs. In general, the authors should provide additional comments/evidence on the interaction, or likelihood of the interaction, of these proteins with VAP, that should include:
    • The IP-MS data should be made available (at least on my end, in the current submission the supplementary lists available cover only the 52 genes found to suppress the phenotype, I found the list of 403 genes in the biorXiv submission but this also does not include data on the enrichment of each hit in the IP-MS). It should be made clear how enriched were each of the proteins in the analysis, and how high the most relevant genes in the genetic screen rank within this interactor list.
    • The authors should provide further detail on the rationale behind defining the list of 403 genes to be tested (for example, what was the threshold enrichment considered as interaction). Also in relation to this, the authors should at least provide speculation as to why more than half of the 403 genes defined do not have an FFAT motif, despite the fact that the proteomic data was normalized to a non-FFAT motif binding mutant of VAP which should be capable of maintaining non-FFAT mediated interactions of the protein.
    • The authors should acknowledge the limitations of their experimental design in regards to identifying real interactors of VAP in Drosophila and avoid referring to this set of genes as "VAP interacting proteins" and rather use a more accurate description such as "proteins enriched in the IP-MS" or at least "potential VAP interacting proteins".
    • Testing interaction between VAP and all the 52 genes found to suppress the phenotype would be a huge amount of work. But the finding most relevant to the initial premise of the study (i.e. "molecular mechanisms underlying lipid transfer protein function at membrane contact sites") is that Cert (a lipid transfer protein) and CG9205 (fly homolog of a mammalian lipid transfer protein) influence RdgB function. Demonstrating an interaction between these proteins and VAP would argue for the experimental design and support the hypothesis and model of the study. Cert is already a well established interactor of VAP, hence the authors would not need to add anything regarding this protein. Is CG9205 expected to be also a true interactor of VAP? Biochemical experiments could be used to test this idea, or even recently developed in silico modelling of interactions (i.e. AlphaFold Multimer) could be of help. If no interaction is observed/expected this should also be pointed out in the manuscript. Optionally, showing localization of Cert or CG9205 to the ER-PM interface would also greatly support the model of VAP-RdgB regulation suggested by the authors.

    Minor comments:

    • In the model shown in Fig2A, it would seem that many proteins can bind VAP in addition to RdgB, however, VAP proteins have only one FFAT binding pocket. This model would only be possible if oligomerization of VAP is considered (oligomerization of VAP has been reported to occur, see for example PMID:20207736). The model should be redrawn considering this fact.
    • In line 161 of the text VPS13D is mentioned, however VPS13C is the gene indicated in Fig 1D.

    Significance

    Previous studies by some of the authors and others have shown that RdgB can transfer lipids between the ER, to which it binds via VAP, and the plasma membrane (PM), and is required for proper replenishment of PI(4,5)P2 in the PM which is in turn necessary for sustained PLC signaling in Drosophila photoreceptors. Lack of RdgB leads to light-dependent degeneration of the retina, and hence it is utilized by the authors as a model for neurodegeneration. Given the clear phenotype of RdgB loss-of-function and the ease of Drosophila genetics, this system represents an ideal model to perform screens for the identification of new genes involved in maintaining neuronal lipid homeostasis required for proper function of the photoreceptors in vivo, and this aspect is the main strength of this study. Importantly, the use of this system could also shed light on the mechanisms behind human neurodegenerative disorders, as many of these involve dysregulation of lipid signaling and lipid transfer at membrane contact sites. A novel and interesting finding is the identification of another lipid transfer protein, Cert, to be involved in the degeneration of photoreceptors.

    The main limitation lies in the experimental design proposed by the authors to define the genes that are studied in their system. These are identified as potential interactors of human VAP in a mammalian cell line. Despite the fact that VAP is a highly conserved protein, and the genes identified are present in Drosophila as well, there is no evidence that these interactions are in fact occurring in Drosophila photoreceptors, and in fact, based on the function and the lack of VAP-binding motif in many of the 52 genes identified to have an effect on the RdgB phenotype, it is likely that many of the interactions are purely genetic and indirect, and that the modulation of the phenotype could in fact be due to a wide variety of factors (including, as discussed by the authors, gene expression, post-translational modifications, trafficking of proteins, etc) unrelated to mechanisms of VAP-RdgB mediated lipid transfer at ER-PM membrane contact sites. A more unbiased screen could have been carried out to identify VAP interactors involved in this degeneration phenotype by testing all of the FFAT or FFAT-related motif containing proteins. Due to this initial bias in the selection of genes to be tested, it is possible that other important VAP interactors that play a role at the ER-PM interface of photoreceptors have not been identified.

    This study provides great functional advance in the understanding of genes implicated in photoreceptor degeneration, and in those regards it is a great resource for a specialized audience, as it enables further characterization by others of the different processes implicated in this neurodegeneration phenotype. However, the advance is small in regards to the core mechanism of RdgB function at VAP-mediated ER-PM, which was the main aim of the article and the most broadly interesting aspect of the study. Many of the VAP-interacting proteins identified in the proteomic approach were already expected to be VAP interactors as they contain FFAT motifs, and these FFAT-containing proteins do not seem to have a major role in VAP-RdgB maintenance of neuronal lipid homeostasis, with the exception of Cert. The implication of Cert in RdgB-mediated lipid homeostasis is certainly interesting as it touches on a current topic in the field of membrane contact sites related to how the multiple interactions of VAP, a universal contact site adaptor at the ER, are regulated and influenced by each other.

    Reviewer's field of expertise: Lipid transfer at membrane contact sites; membrane lipid homeostasis in neurons. All of the Drosophila data seem to be of good general quality to me, but I do not have any expertise in Drosophila work.

  3. Review Commons

    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

    Mishra and colleagues have conducted a large genetic screen to identify modulators of a Drosophila model for retinal degeneration. Using biomolecular techniques, they selected a few hundred proteins that interact with an ER bound protein VAP, to further test them in the retinal degeneration model. This was done by downregulating their expression using interference RNA (RNAi). Degeneration was first measured through a pseudopupil analysis, then suppressors of degeneration were further tested in a different retinal degeneration model, and finally through an ERG experiment. Finally, they focused on a strong suppressor of degeneration, dCert, by using a mutant allele to confirm the findings from the RNAi.

    The results suggest that a handful of these candidates are suppressors in this model of retinal degeneration, and identify at which stage of retinal degeneration these proteins may be involved in. These proteins may have a significance in forms of neurodegeneration.

    Major comment:

    • Perhaps a larger number of replicates could be done in the optical neutralization experiment as well as in the ERG. Figure 4.A(i) and (ii), please clearly state n values. I would suggest this as optional, but perhaps ut could help to increase n?
    • For human orthologs (Table 1), it could be worthwhile to add alignment scores between fly and human?

    Minor comment:

    • Clarify the purpose in focusing on dCert specifically in the last results section and discussion
    • Several typos
    • Affect vs effect

    Significance

    General assessment:

    The main significance of this study comes from the focus on proteins that are known to interact with VAP. This implies that the suppressors of degeneration that they have identified in the RdgB9 model may have an effect in other neurodegenerative models, namely in ALS models. This could have a very high significant potential in therapeutic avenues for neurodegenerative diseases.

    Among six candidates that had an effect with knocked down through RNAi, they pursued a single one (dCert) as proof of principle. It would help to add a justification for this choice in the main text and whether the authors have performed or intend to perform experiments using mutant forms of the other candidate proteins.

    The work from Raghu and his team have been leading the research surrounding this model of degeneration in Drosophila. This study naturally further extends their field of research, identifying more candidates that modulate this form of degeneration, and helping elucidate the pathways leading to cellular degeneration.

    These results will be of high interest for specialized researchers studying the molecular pathways that lead to cellular degeneration, both in the context of retinal degeneration as well as neurodegeneration. Specifically, researchers that may be interested in these candidate proteins and how they may play a role in the pathogenesis of various degenerative diseases.

  4. Version published to 10.1101/2023.07.28.550979v1 on bioRxiv