Assembly of higher-order SMN oligomers is essential for metazoan viability and requires an exposed structural motif present in the YG zipper dimer

Article activity feed

1. Version published to 10.1093/nar/gkab508

   Jun 28, 2021
2. 

   eLife

   Jun 3, 2021

   Author Response:

   We first posted this manuscript on bioRxiv in Nov 2020. While it was under consideration, the bioRxiv posting caused another group to submit a competing manuscript to Nucleic Acids Research. Their manuscript was recently published in (PMID: 33754639 DOI: 10.1093/nar/gkab158). Although that manuscript is less comprehensive that ours (limited only to the study of the S.pombe SMN complex), the paper included a high-resolution crystal structure of a non-native fusion of SMN (involving a large internal deletion of the yeast protein). Publication of that structure caused us to pause submission of a revised version to another journal and to carry out a disulfide crosslinking experiment aimed at directly testing the two possible models of SMN oligomerization (parallel vs antiparallel). As shown in our revised …

   Author Response:

   We first posted this manuscript on bioRxiv in Nov 2020. While it was under consideration, the bioRxiv posting caused another group to submit a competing manuscript to Nucleic Acids Research. Their manuscript was recently published in (PMID: 33754639 DOI: 10.1093/nar/gkab158). Although that manuscript is less comprehensive that ours (limited only to the study of the S.pombe SMN complex), the paper included a high-resolution crystal structure of a non-native fusion of SMN (involving a large internal deletion of the yeast protein). Publication of that structure caused us to pause submission of a revised version to another journal and to carry out a disulfide crosslinking experiment aimed at directly testing the two possible models of SMN oligomerization (parallel vs antiparallel). As shown in our revised manuscript, the data are incompatible with the proposed antiparallel model suggested by 10.1093/nar/gkab158.

   Overall, the eLife referees were extremely thorough, and we thank them for their detailed comments that certainly improved the manuscript. The critiques were all readily addressible and no additional experiments were suggested, so its mostly a subjective decision. We do, however, disagree with one major point regarding overall impact of the study (see below).

   Summary:

   The manuscript describes a very detailed mutagenesis analysis of the dimerization / oligomerization behavior of the protein Survival Motor Neuron. Mutations in this protein cause Spinal Muscular Atrophy. Analysis of disease causing mutations show a correlation with their impact on oligomerization. A structural model that includes different domains of the protein involved in oligomerization is built from these analyses.

   This analysis is an excellent source for researchers working in the field of SMN proteins. A mechanistic interpretation of how changes in the oligomerization lead to the disease or impact the formation of membraneless organelles, is however missing. Thus, the manuscript provides an enormous amount of important mutational analysis data but does not lead to a significant advancement in our understanding of the disease mechanism.

   The referees feel that our study represents an incremental advance in terms of “our understanding of disease mechanism.” Why is the bar placed so high? We did not set out to solve the disease mechanism here. There is a great deal of misconception and misinformation regarding the molecular etiology of SMA. After >25 years of intensive study, researchers still have essentially no idea why low levels of SMN protein cause the disease.

   We know that SMN binds to itself (along with many other proteins) and that it forms a large, heterogeneous complex in vivo. The assumption has been that YG box-mediated oligomerization of SMN is important for its essential function. Is it? Nobody knows for sure. So we set out to ask and answer this question. As detailed below, we disagree with the contention that the manuscript fails to provide mechanistic evidence regarding the formation of SMN oligomers.

   Reviewer #1:

   Gupta et al. provide a very detailed and in depth analysis of the dimerization / oligomerization behavior of the protein Survival Motor Neuron (SMN). The protein is able to use a modified glycine zipper motif to form tightly packed dimers and additional hydrophobic amino acids for higher oligomeric states. Mutations in SMN cause Spinal Muscular Atrophy and the authors show that mutations leading to this disease affect the oligomerization state of the protein.

   Overall, this is a very detailed study using several biophysical techniques and extensive mutagenesis. The data are of high importance for researchers working in the field of SMN proteins.

   A mechanistic link of how these differences in oligomeric states changes the cellular behavior leading to Spinal Muscular Atrophy is unfortunately missing.

   There is no evidence that formation of SMN-containing membraneless organelles has anything to do with SMA. The mechanism whereby SMN forms nuclear and cytoplasmic foci is a separate but interesting biological question, well worth investigating. But first, we need to know more about how SMN makes oligomers. We identified the residues involved in this process and, as Referee 3 points out, we presented a “plausible model for higher-order oligomers.”

   The authors stress several times that SMN is part of membraneless organelles. Multivalent interactions are characteristic of such organelles, although they are typically based on "fuzzy" interactions involving low complexity regions (and not all dimerization / oligomerization events can be classified as liquid-liquid phase separation). This limits the impact of this detailed analysis.

   We do not wish to de-emphasize the fact that SMN forms membraneless organelles. Our findings are certainly pertinent to this subject because we know so little about how SMN or any other nuclear body protein (e.g. Coilin, NPAT) self-interact and multimerize. However, so as not to distract the reader in the Introduction, we have moved the subject of biomolecular condensation to the Discussion.

   As for this Referee’s other comments, we agree that not all oligomerization events can be classified as LLPS. We note that the sequence of SMN also contains large regions of low-complexity. Particularly in the area located between the YG box and Tudor domains of metazoan SMN proteins. The extent to which any of these domains (including the more structured ones) participate in biomolecular condensation is not known. Therefore, understanding the mechanism whereby the SMN YG box forms oligomers represents a critical first step in the exploration of many downstream aspects of SMN function, including LLPS.

   We note that the opto-droplet technique most investigators use to study LLPS inside mammalian cells involves the use of a very well structured self-interaction motif (Cry2) that is then tethered to a low-complexity “tester domain” (together with a fluorescent reporter). This is not unlike the natural arrangement of domains within SMN. Although we did not focus on the “fuzzy” parts of SMN in this manuscript, we feel that the current study provides an important mechanistic foundation for other researchers working on nuclear and cytoplasmic bodies whose proteins may well employ similar strategies.

   While this very detailed analysis is an excellent source for researchers working in this field the interest beyond SMN proteins will be limited. The paper could also be written in a less dense manner, which would make its reading easier. The main weakness is a missing mechanistic model that can explain how differences in the oligomerization behavior relates to the function of the protein and causes Spinal Muscular Atrophy. The impact of oligomerization on the formation of membraneless organelles would be important.

   This summary paragraph is a re-emphasis of the same points. See our responses above.

   Reviewer #2:

   The current study nicely demonstrates that high-order assembly of SMN protein oligomerization is necessary for animal survival and is dependent on a motif exposed to YG zipper dimers. Mutations in the human SMN1 gene have been shown to cause a neurodegenerative disease named Spinal Muscular Atrophy (SMA). About 50% of the SMA-causing mutations are located in the YG zipper domain. The authors used multi-disciplinary approaches such as biophysical, bioinformatic, computational and genetic approaches to demonstrate that a set of YG box amino acids in SMN protein are not involved in dimerization process and formation high-order oligomers is dependent on these residues. Importantly, mutating key residues within this new structural domain impairs SMN dimerization and causes motor dysfunction as well as viability defects in Drosophila. Overall, this is a well-written paper that offers new insights into the structural and functional aspects of SMN protein. The authors should consider addressing the following issues:

   1. The authors should discuss the impact of the YG zipper domain mutations on snRNP biogenesis. SMN protein is a master regulator of snRNP biogenesis. It is a little surprising that the authors did not mention snRNP biogenesis in the whole manuscript.

   Apologies if we did not mention it prominently enough; the biogenesis of small nuclear RNPs is an area of great interest to our group. However, using the same Drosophila model system, we have already devoted several entire manuscripts to this very subject (e.g. Garcia et al. 2016; Garcia et al. 2013; Praveen et al. 2012). Using qRT-PCR and RNA-seq we showed that hypomorphic point mutations that cause milder forms of the human disease show few overall defects in pre-mRNA splicing in flies. They do however, cause SMA-like phenotypes (e.g. locomotor defects in larvae and adults, reduced adult lifespan, etc). For additional details, see Spring et al. (2019).

   Thus, to the best of our current ability, we feel we have already addressed the question of snRNP biogenesis. That is: the pre-mRNA splicing-related phenotypes seen in the null mutants can indeed be separated from the neuromuscular defects observed in the hypomorphs.

   1. The authors should provide evidence that their transgenic lines express the desired transgene. A WB or qPCR would be great (even as supplementary data).

   Again, this is an extremely well characterized model system. We have already published WB and even RNA-seq data showing that the transgenic lines express the desired transgenes. See Praveen et al. 2014; 2012. For each of the transgenic fly lines, we PCR genotype DNA from the founder lines and test each of them with anti-Flag and anti-SMN antibodies prior to using them in experiments.

   1. Page 12: The authors stated "Both missense mutations display early onset SMA-like phenotypes". Was it age-dependent phenotype? Did adult animals show a more severe motor dysfunction?

   We changed the language here to be more precise. The mutations in question are Y208C and Y208A. The Y208C mutation causes SMA in humans and the fly model of it has been described (Spring et al. 2019). Y208A has not been identified in humans. Both alleles are considered Class 2 in flies (Figs. 5A and S4). These animals display locomotor phenotypes during larval stages and undergo developmental arrest during pupation. As shown in Figs 6C and 6D, Y208A is more severe than Y208C; we never observe Y208A adults, whereas a small fraction of Y208C animals can complete development and eclose.

   1. There are few statements that the authors should consider making clear. Here is an example "Presumably, the structural changes associated with Cys and Val substitutions do interfere with some aspect of SMN biology, leading to the intermediate and severe SMA phenotypes observed". What do you mean by some aspects? Oligomerization, stability or anything else?

   Thank you. This issue has been addressed throughout the text. With respect to the specific example noted above, that sentence was deleted.

   1. There are few typos throughout the manuscript that the authors should correct (western should be written as Western).

   It turns out that, in reference to various blotting techniques, the referee is incorrect: western and northern blotting are not capitalized. The only one that is capitalized is Southern blotting. This procedure is named after Edwin Southern, who first described it. When techniques for blotting RNAs and proteins were subsequently developed, those investigators cheekily named them in homage to Ed. In this context, western and northern are not compass directions, and so they are NOT generally capitalized. We are nevertheless happy to follow whatever editorial house style the journal uses. Also note that we have thoroughly re-checked the rest of manuscript for typographical errors.

   Reviewer #3:

   In "Assembly of higher-order SMN oligomers is essential for animal viability, requiring a motif exposed in TG zipper dimers," Gupta et al. present an impressive amount of data regarding the solution behavior of constructs of the protein SMN1 (or just SMN) from Homo sapiens, Drosophila melanogaster, and Schizosaccharomyces pombe. Defects in the Hs protein are known to cause the neuromuscular disease "Spinal Muscular Atrophy" (SMA). They also present experiments in genetically modified organisms (fission yeast and fruit flies) to test their hypotheses. Bioinformatics are used to generate and refine hypotheses. The potential power of these complementary methods is substantial, if employed well.

   The main finding of these researchers is that the oligomerization potential of SMN and its disease-causing variants (usually in complex with the protein Gemin 2 or "G2") mostly correlates with phenotype severity. In humans, this is correlated with the Type of SMA (I/0 for severe disease, ranging to IV for a milder form), and in fruit flies and yeast, it is correlated with viability and, in some cases, animal behavior. The results are extended through the creation of a model that purports to show how higher-order SMN oligomers can form.

   Strengths:

   The experiments appear to have been carried out competently. There is a virtual mountain of data presented in this paper, and, for the most part, they are summarized in a digestible fashion. The effort to correlate the biophysical solution data with observable phenotypes in human patients or genetically modified organisms is laudable, and it is done in a thoughtful fashion. The authors' structural intuition and savvy enables the generation of testable models that are explored in the paper. A plausible model for higher-order oligomers is presented.

   We thank the referee for their thorough reading and insightful summary of the approach and its strengths.

   Weaknesses:

   The most serious weakness of the paper is that the data cannot support the conclusion stated in the title, i.e. that multimerization of SMN is necessary for organismic viability. Instead, the data support an already-stated, decades-old conclusion (see their reference 21): that multimerization correlates with disease (viability). Even if the reader takes into account the new information about a multimerization interface that is separate from the dimerization one, the advance seems incremental.

   We understand why the referee might take the position that the overall advance seems incremental. We chose this manuscript title because the decades-old ‘conclusion’ to which the referee refers is based on a false assumption. The false assumption is that the self-oligomerization activity of the YG box is required for SMN function. Note that Ref. 21 (Lefebvre et al. 1995) is the positional cloning paper for the SMN locus, not the publication that first showed the correlation betwen SMN self-binding and SMA severity (Lorson et al. 1998). To date, no one has shown that oligomerization of SMN (n > 2) is actually required for its function.

   Thus, the whole point of our going through a pain-staking mutational analysis was to uncover a separation-of-function allele. Namely, one that maintains the dimerization activity of SMN but interferes with its higher-order oligomerization activity. Y208A (Y277A in human) is one such allele. We show that this allele fails to form higher-order oligomers in vitro and that it causes locomotor dysfunction and early lethality in vivo. The referee does not explain; what is the basis for discounting our conclusion? Is it because we did not specifically demonstrate a lack of tetramer formation for this mutant in vivo (a nearly impossible experiment, btw)?

   Given the overall slow rate of progress on SMN ultrastructure in the literature, we felt that the identification of this second tetramer-forming interface is an important finding. We understand if the editors feel that we have not reached the rather subjective threshold necessary for the manuscript to be published in eLife. But we disagree with the referee on the idea that our data fail to support the conclusion.

   The large amount of data leads to numerous difficulties for the reader in the text:

   As the referee mentions above, there is a veritable mountain of data in this manuscript and presenting it clearly and succintly has been a challenge. Finding the right balance between too much detail for the casual reader and too little detail for the cognoscenti is a difficult task. Add to that, the breadth of techniques employed and it is nearly impossible to please everybody. We thank the referee for pointing out some of the areas they found to be problematic.

   1. Complex biophysical measurements, due to space, are usually summarized by one or two words in tabular format.

   1. When these measurements are shown, there is no visual context for the reader to assess the pre-digested conclusions that are included in the figures. For example, all SEC-MALS data show a conclusion ("Tetramer-Octamer"), but there is no visual cue for the reader to know what the theoretical masses for these species are (so that the reader may draw an independent conclusion).

   There are more than eighty different SEC-MALS experiments presented in this manuscript. We simply cannot display a visual representation for all of them. Most of these constructs contain point substitutions and so they have equivalent theoretical masses. We showed traces for the wild-type constructs for each of the four species analyzed (human, fly, nematode and yeast), as well as for the mutations that are highlighted in Figs 5 and 6.

   To provide a better visual cue, we re-formatted the Y-axes of all of the SEC-MALS traces in the paper. As you can now see, the tick marks for the molar masses are expressed as units of a single SMN•Gem2 heterodimer. In this way, the oligomerization of yeast, fly, human and nematode heterodimers can all be compared by simply counting up the rungs on the right hand side of the Y-axis.

   In some cases, the conclusions reached in the paper are not clearly supported by the data or are self-contradictory. An example is the discussion of the residue H273 (human numbering). In Fig. 4B, the mutation H273R is said to have a wild-type "Oligomer Status". But in Fig. 5B, it is "Dimer-Tetramer+". The text says that H273R is "only partially impaired" in forming oligomers; the authors apparently mean the data presented in Fig. 5B but refer to the contradictory result in Fig. 4B.

   The disconnect between Fig 4B and 5B was a simple cut/paste error that has been corrected. First of all, the primary data are the same. Those data are shown visually in Fig. 5B and in tabular form in Fig. 4B. The biophysical parameters of this particular mutant (H273R) are, however, unusual and hard to describe. Although high MW species are detected (some of which appear to be very high MW aggregates that are on the leading edge of the peak), the mutation also clearly pushes the equilibrium toward dimers (and in the case of the fly ortholog, monomers). We have revised the text and the display elements to so indicate.

   Another example centers on the discussion of the putative "dominant-negative" effect of some human missense mutations. But they do not point to any human data that support this contention (SMA-associated missense mutations are usually discovered in mixed heterozygotes have a deletion in the other copy of the Smn gene), but they cite data that suggest a more nuanced position regarding negative dominance would be appropriate.

   Due to the variable copy number of SMN2 genes and the extremely low frequency of SMA patients bearing SMN1 point mutations, the human data on this topic are a complete minefield. So the fact that we do not point to any human data on this is irrelevant. Less complex model systems often point the way to uncover nuances like this that are later identified in humans.

   We were the first to show a dominant phenotypic effect for certain SMN point mutants in vivo (Praveen et al. 2014). Specifically, we showed that two YG box missense mutants (dmY203C/hsY272C and dmM194R/hsM263R) have a more severe phenotype than the SmnX7 micro-deletion null mutation. That is, animals that contain no zygotic SMN protein actually live longer than the ones that express these missense mutations. The fact that the dmG026S/hsG275S mutation fails to bind to the wild-type protein is consistent with the slightly milder phenotype we observe for G206S vs Y203C and M194R (Spring et al. 2019). All three of these mutant proteins (human or fly) fail to self-interact (many previous papers), but only G206S fails to interact with wild-type SMN (Fig 4C).

   We included the data here because they help illustrate the conservation between human and fly systems. We also note that, in their detailed comments (see detailed subpoint 9c, below), this same referee called this finding “ground breaking.” We apologize for the lack of clarity in this section and have revised the text.

   Finally, the paper suffers throughout from a lack of precision of language that undercuts its conclusions at numerous points. They continually rely on qualitative statements rather than hard, statistically rigorous facts, e.g. "more intimate," "a bit of a sequence outlier," "very modest."

   This has been addressed.

   Read the original source
3. 

   eLife

   May 25, 2021

   Reviewer #3 (Public Review):

   In "Assembly of higher-order SMN oligomers is essential for animal viability, requiring a motif exposed in TG zipper dimers," Gupta et al. present an impressive amount of data regarding the solution behavior of constructs of the protein SMN1 (or just SMN) from Homo sapiens, Drosophila melanogaster, and Schizosaccharomyces pombe. Defects in the Hs protein are known to cause the neuromuscular disease "Spinal Muscular Atrophy" (SMA). They also present experiments in genetically modified organisms (fission yeast and fruit flies) to test their hypotheses. Bioinformatics are used to generate and refine hypotheses. The potential power of these complementary methods is substantial, if employed well.

   The main finding of these researchers is that the oligomerization potential of SMN and its disease-causing variants …

   Reviewer #3 (Public Review):

   In "Assembly of higher-order SMN oligomers is essential for animal viability, requiring a motif exposed in TG zipper dimers," Gupta et al. present an impressive amount of data regarding the solution behavior of constructs of the protein SMN1 (or just SMN) from Homo sapiens, Drosophila melanogaster, and Schizosaccharomyces pombe. Defects in the Hs protein are known to cause the neuromuscular disease "Spinal Muscular Atrophy" (SMA). They also present experiments in genetically modified organisms (fission yeast and fruit flies) to test their hypotheses. Bioinformatics are used to generate and refine hypotheses. The potential power of these complementary methods is substantial, if employed well.

   The main finding of these researchers is that the oligomerization potential of SMN and its disease-causing variants (usually in complex with the protein Gemin 2 or "G2") mostly correlates with phenotype severity. In humans, this is correlated with the Type of SMA (I/0 for severe disease, ranging to IV for a milder form), and in fruit flies and yeast, it is correlated with viability and, in some cases, animal behavior. The results are extended through the creation of a model that purports to show how higher-order SMN oligomers can form.

   Strengths:

   The experiments appear to have been carried out competently. There is a virtual mountain of data presented in this paper, and, for the most part, they are summarized in a digestible fashion. The effort to correlate the biophysical solution data with observable phenotypes in human patients or genetically modified organisms is laudable, and it is done in a thoughtful fashion. The authors' structural intuition and savvy enables the generation of testable models that are explored in the paper. A plausible model for higher-order oligomers is presented.

   Weaknesses:

   The most serious weakness of the paper is that the data cannot support the conclusion stated in the title, i.e. that multimerization of SMN is necessary for organismic viability. Instead, the data support an already-stated, decades-old conclusion (see their reference 21): that multimerization correlates with disease (viability). Even if the reader takes into account the new information about a multimerization interface that is separate from the dimerization one, the advance seems incremental.

   The large amount of data leads to numerous difficulties for the reader in the text:

   1. Complex biophysical measurements, due to space, are usually summarized by one or two words in tabular format.

   2. When these measurements are shown, there is no visual context for the reader to assess the pre-digested conclusions that are included in the figures. For example, all SEC-MALS data show a conclusion ("Tetramer-Octamer"), but there is no visual cue for the reader to know what the theoretical masses for these species are (so that the reader may draw an independent conclusion).

   In some cases, the conclusions reached in the paper are not clearly supported by the data or are self-contradictory. An example is the discussion of the residue H273 (human numbering). In Fig. 4B, the mutation H273R is said to have a wild-type "Oligomer Status". But in Fig. 5B, it is "Dimer-Tetramer+". The text says that H273R is "only partially impaired" in forming oligomers; the authors apparently mean the data presented in Fig. 5B but refer to the contradictory result in Fig. 4B. Another example centers on the discussion of the putative "dominant-negative" effect of some human missense mutations. But they do not point to any human data that support this contention (SMA-associated missense mutations are usually discovered in mixed heterozygotes have a deletion in the other copy of the Smn gene), but they cite data that suggest a more nuanced position regarding negative dominance would be appropriate.

   Finally, the paper suffers throughout from a lack of precision of language that undercuts its conclusions at numerous points. They continually rely on qualitative statements rather than hard, statistically rigorous facts, e.g. "more intimate," "a bit of a sequence outlier," "very modest."

   Read the original source
4. 

   eLife

   May 25, 2021

   Reviewer #2 (Public Review):

   The current study nicely demonstrates that high-order assembly of SMN protein oligomerization is necessary for animal survival and is dependent on a motif exposed to YG zipper dimers. Mutations in the human SMN1 gene have been shown to cause a neurodegenerative disease named Spinal Muscular Atrophy (SMA). About 50% of the SMA-causing mutations are located in the YG zipper domain. The authors used multi-disciplinary approaches such as biophysical, bioinformatic, computational and genetic approaches to demonstrate that a set of YG box amino acids in SMN protein are not involved in dimerization process and formation high-order oligomers is dependent on these residues. Importantly, mutating key residues within this new structural domain impairs SMN dimerization and causes motor dysfunction as well as viability …

   Reviewer #2 (Public Review):

   The current study nicely demonstrates that high-order assembly of SMN protein oligomerization is necessary for animal survival and is dependent on a motif exposed to YG zipper dimers. Mutations in the human SMN1 gene have been shown to cause a neurodegenerative disease named Spinal Muscular Atrophy (SMA). About 50% of the SMA-causing mutations are located in the YG zipper domain. The authors used multi-disciplinary approaches such as biophysical, bioinformatic, computational and genetic approaches to demonstrate that a set of YG box amino acids in SMN protein are not involved in dimerization process and formation high-order oligomers is dependent on these residues. Importantly, mutating key residues within this new structural domain impairs SMN dimerization and causes motor dysfunction as well as viability defects in Drosophila. Overall, this is a well-written paper that offers new insights into the structural and functional aspects of SMN protein. The authors should consider addressing the following issues:

   1. The authors should discuss the impact of the YG zipper domain mutations on snRNP biogenesis. SMN protein is a master regulator of snRNP biogenesis. It is a little surprising that the authors did not mention snRNP biogenesis in the whole manuscript.

   2. The authors should provide evidence that their transgenic lines express the desired transgene. A WB or qPCR would be great (even as supplementary data).

   3. Page 12: The authors stated "Both missense mutations display early onset SMA-like phenotypes". Was it age-dependent phenotype? Did adult animals show a more severe motor dysfunction?

   4. There are few statements that the authors should consider making clear. Here is an example "Presumably, the structural changes associated with Cys and Val substitutions do interfere with some aspect of SMN biology, leading to the intermediate and severe SMA phenotypes observed". What do you mean by some aspects? Oligomerization, stability or anything else?

   5. There are few typos throughout the manuscript that the authors should correct (western should be written as Western).

   Read the original source
5. 

   eLife

   May 25, 2021

   Reviewer #1 (Public Review):

   Gupta et al. provide a very detailed and in depth analysis of the dimerization / oligomerization behavior of the protein Survival Motor Neuron (SMN). The protein is able to use a modified glycine zipper motif to form tightly packed dimers and additional hydrophobic amino acids for higher oligomeric states. Mutations in SMN cause Spinal Muscular Atrophy and the authors show that mutations leading to this disease affect the oligomerization state of the protein. Overall, this is a very detailed study using several biophysical techniques and extensive mutagenesis. The data are of high importance for researchers working in the field of SMN proteins.

   A mechanistic link of how these differences in oligomeric states changes the cellular behavior leading to Spinal Muscular Atrophy is unfortunately missing. The …

   Reviewer #1 (Public Review):

   Gupta et al. provide a very detailed and in depth analysis of the dimerization / oligomerization behavior of the protein Survival Motor Neuron (SMN). The protein is able to use a modified glycine zipper motif to form tightly packed dimers and additional hydrophobic amino acids for higher oligomeric states. Mutations in SMN cause Spinal Muscular Atrophy and the authors show that mutations leading to this disease affect the oligomerization state of the protein. Overall, this is a very detailed study using several biophysical techniques and extensive mutagenesis. The data are of high importance for researchers working in the field of SMN proteins.

   A mechanistic link of how these differences in oligomeric states changes the cellular behavior leading to Spinal Muscular Atrophy is unfortunately missing. The authors stress several times that SMN is part of membraneless organelles. Multivalent interactions are characteristic of such organelles, although they are typically based on "fuzzy" interactions involving low complexity regions (and not all dimerization / oligomerization events can be classified as liquid-liquid phase separation). This limits the impact of this detailed analysis.

   While this very detailed analysis is an excellent source for researchers working in this field the interest beyond SMN proteins will be limited. The paper could also be written in a less dense manner, which would make its reading easier. The main weakness is a missing mechanistic model that can explain how differences in the oligomerization behavior relates to the function of the protein and causes Spinal Muscular Atrophy. The impact of oligomerization on the formation of membraneless organelles would be important.

   Read the original source
6. 

   eLife

   May 25, 2021

   Summary:

   The manuscript describes a very detailed mutagenesis analysis of the dimerization / oligomerization behavior of the protein Survival Motor Neuron. Mutations in this protein cause Spinal Muscular Atrophy. Analysis of disease causing mutations show a correlation with their impact on oligomerization. A structural model that includes different domains of the protein involved in oligomerization is built from these analyses.

   This analysis is an excellent source for researchers working in the field of SMN proteins. A mechanistic interpretation of how changes in the oligomerization lead to the disease or impact the formation of membraneless organelles, is however missing. Thus, the manuscript provides an enormous amount of important mutational analysis data but does not lead to a significant advancement in our understanding of the …

   Summary:

   The manuscript describes a very detailed mutagenesis analysis of the dimerization / oligomerization behavior of the protein Survival Motor Neuron. Mutations in this protein cause Spinal Muscular Atrophy. Analysis of disease causing mutations show a correlation with their impact on oligomerization. A structural model that includes different domains of the protein involved in oligomerization is built from these analyses.

   This analysis is an excellent source for researchers working in the field of SMN proteins. A mechanistic interpretation of how changes in the oligomerization lead to the disease or impact the formation of membraneless organelles, is however missing. Thus, the manuscript provides an enormous amount of important mutational analysis data but does not lead to a significant advancement in our understanding of the disease mechanism.

   (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 agreed to share their name with the authors.)

   Read the original source
7. 

   Version published to 10.1101/2020.11.07.336172v1 on bioRxiv

   Nov 8, 2020