Influence of Mutations on Physicochemical Properties of Spike Proteins from Prototypical SARS-CoV-2 Variants of Concern Detected in Amazonian Countries
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThe study thoroughly investigates the spread of various SARS-CoV-2 variants in Amazonian countries, utilizing sophisticated bioinformatics tools and statistical methods, but further research on the transmission mechanisms and public health implications of these variants is recommended for a more comprehensive understanding.
Comments for author File: 
Comments.pdf
The article' s quality of english language is good, demonstrating logical coherence and lacking obvious grammatical errors.
Author Response
Comment 1: The study thoroughly investigates the spread of various SARS-CoV-2 variants in Amazonian countries, utilizing sophisticated bioinformatics tools and statistical methods, but further research on the transmission mechanisms and public health implications of these variants is recommended for a more comprehensive understanding.
Response 1: Thank you for your positive feedback on our selection of bioinformatics tools and statistical methods. We appreciate your suggestion to delve deeper into the transmission mechanisms and public health implications of SARS-CoV-2 variants. In response to your recommendation, we have revised manuscript to add a new paragraph to the introduction section emphasizing the impact of these variants on transmission rates, severity of illness and vaccine effectiveness, particularly with regard to delta and omicron variants (lines 54–60).
Comment 2: The article mentions that volatile organic compounds (VOCs) have relatively high antigenic scores in the NTD region of the S protein, but it does not delve into the biological significance of these findings.
Response 2: Thank you for your careful review. In our article, VOCs refer to "variants of concern" of SARS-CoV-2, not “volatile organic compounds”, as defined upon its first use both in the abstract (line 12) and the introduction (line 49) sections to prevent misunderstandings. To incorporate your suggestion to delve into the biological significance of the high antigenic scores observed for specific S protein domains of SARS-CoV-2 VOCs, we have addressed how the increased antigenicity in both the NTD and RBD of these variants may influence immune recognition and thus vaccine efficacy in the discussion section (lines 306–309).
Comment 3: It is recommended that the authors further elaborate on how the structural changes in the S protein specifically affect the interaction between the virus and cellular receptors.
Response 3: Thank you for your valuable feedback, we appreciate the opportunity to clarify this important aspect. In response to your suggestion, we have revised the manuscript to provide a more detailed discussion in lines 310–322 on how the identified structural changes influence the S protein interaction with host cell receptors, considering every topic of our study (i.e., side-chain polarities, secondary structures, post-translational modifications and antigenic propensities). We believe that these additions will provide a comprehensive understanding of the functional implications of the structural changes in the S protein, thereby enhancing the clarity and depth of our manuscript.
Comment 4: It is recommended to add potential applications and suggestions of the research results in the discussion section.
Response 4: Thank you for your insightful feedback and suggestion to include potential applications of our research results, which enhance the relevance and impact of our findings. In response to your recommendation, we have expanded the discussion section to highlight several potential applications of our research results, especially toward the development of vaccines and drugs to better manage and mitigate the impact of COVID-19 (lines 323–335).
Reviewer 2 Report
Comments and Suggestions for AuthorsThe paper is written in clear, concise language. It is rather interesting and valuable as a comprehensive computational analysis of the mutations evoked in the different SARS-CoV-2 lineages. The authors aim to reveal the influence of these mutations on physicochemical properties of S proteins from prototypical SARS-CoV-2 VOCs detected in Amazonian countries.
However, I have a few concerns:
1. For non-bioinformatics auditorium is not easy to understand all the techniques used based on very short description and thus to believe the results and conclusions. I would suggest the authors to describe the methods applied in a bit more detail. E.g. to indicate the Consensus patterns of post-translational modifications searched is a must.
2. My main consern is N-myristoylation. As far as I know N-myristoylated is usually an N-terminal glycine. Which exaclty glycine might be actually N-myristoylated in the S-protein? Which one is predicted? In the SARS-CoV-2 S-protein's amino acid sequence there is no Gly2 (instead, Phe2 is following Met1). It also seems not feasible that the D614G substitution in S of Omicron strains might become N-terminal due to some proteolytical cleavage and that this Gly614 might get the myristoylation modification in a real protein. Usually, N-myristoylation is needed to anchor a protein or its part to lipid membrane. But according to S-protein's 3D-structure Gly614 is rather far away from the lipid membrane.
What I want to say is that the authors should be careful when interpreting their computational predictions since not every predicted post-translationally modifed site undergoes modification in a real protein.
Note that the authors indicate this N-myristoylation PTM as the most significant (showing reliable differece between the Omicron and non-Omicron strains).
3. Please indicate the post-translationsl modifications found and discussed at the S protein amino acid sequence.
4. I would suggest showing the boundaries of NTD and RBD regions discussed in Lines 119-120 in Figure S1. Actually, I guess that those regions should be marked on the sequence shown in Figure S1. Please refer to Figure S2 according to its meaning.
5. Line 244: I guess that in Ref. [22] another lipid modification, namely, S-palmitoylation (not N-myristoylation) is discussed as that important for membrane fusion and virus entry.
6. It would be also great if the authors give a scheme of S-protein primary structure illustrating its main domains / motives listed in Lines 41-44 as a Figure.
7. I did not find in the Results section the nucleotide mutations discussed in the Discussion Section (Line 217). Please add some factual data regarding nucleotide substitutions to the Results (or to Supplemental material), or indicate that you are not discussing your own data, but data obtained by other authors (with a reference).
Author Response
Comment 1: For non-bioinformatics auditorium is not easy to understand all the techniques used based on very short description and thus to believe the results and conclusions. I would suggest the authors to describe the methods applied in a bit more detail. E.g. to indicate the Consensus patterns of post-translational modifications searched is a must.
Response 1: Thank you for your valuable feedback. We appreciate your suggestion to provide a more detailed description of the methods used in our study, understanding that a comprehensive explanation is essential for a broader audience to fully grasp the techniques. In response, we have improved the materials and methods section by describing in more details all the applied computational tools in lines 84–90 (Clustal Omega), 96–106 (EMBOSS Pepstats), 111–116 (NPS@ PREDATOR), 121–130 (NPS@ PROSCAN) and 135–143 (NPS@ PCPROF) of the revised manuscript, including the indication of the consensus patterns of post-translational modifications searched as requested.
Comment 2: My main consern is N-myristoylation. As far as I know N-myristoylated is usually an N-terminal glycine. Which exaclty glycine might be actually N-myristoylated in the S-protein? Which one is predicted? In the SARS-CoV-2 S-protein's amino acid sequence there is no Gly2 (instead, Phe2 is following Met1). It also seems not feasible that the D614G substitution in S of Omicron strains might become N-terminal due to some proteolytical cleavage and that this Gly614 might get the myristoylation modification in a real protein. Usually, N-myristoylation is needed to anchor a protein or its part to lipid membrane. But according to S-protein's 3D-structure Gly614 is rather far away from the lipid membrane. What I want to say is that the authors should be careful when interpreting their computational predictions since not every predicted post-translationally modifed site undergoes modification in a real protein. Note that the authors indicate this N-myristoylation PTM as the most significant (showing reliable differece between the Omicron and non-Omicron strains).
Response 2: Thank you for your thorough review and for highlighting this important point regarding N-myristoylation. You are correct that N-myristoylation typically occurs on an N-terminal glycine residue, via an amide linkage. In the case of the SARS-CoV-2 S protein, there is indeed no Gly2 following the initial methionine (Met1), but rather a phenylalanine (Phe2). Upon reevaluating our analysis, we have realized that several consensus patterns for this post-translational modification in the S protein were inadvertently identified by the NPS@ PROSCAN program because its algorithm deliberately include as potential myristoylated glycine residues those which are internal to a sequence, as it could well be that the sequence under study represents a viral polyprotein precursor and that subsequent proteolytic processing could expose an internal glycine as the N-terminus of a mature protein, which, however, is not the case for the SARS-CoV-2 S protein. We apologize for this oversight. To fix this, we have now updated Figure 4 and its related text (lines 219–226, 231–236 and 289–294) to exclude any indication of N-myristoylation. Furthermore, we have improved the comprehension of Figure 4 by bringing together PKA/G, PKC, CK2 and TK phosphorylations into the same type of post-translational modification (i.e., phosphorylation). Regarding the interpretation of our computational predictions, we fully agree with your point that not every predicted post-translationally modified site undergoes modification in a real protein. To take this relevant observation into consideration, we have emphasized this limitation of our study in the discussion section and highlighted the necessity for further experimental validation to confirm these computational findings (lines 294–298).
Comment 3: Please indicate the post-translationsl modifications found and discussed at the S protein amino acid sequence.
Response 3: Thank you for your feedback. In response to this suggestion, we have updated Figures S1 and S2 to clearly indicate the predicted glycosylation (highlighted in yellow) and phosphorylation (highlighted in green) sites in the amino acid sequences of both subunits of the S protein for all variants under study (figure callouts in lines 225–226 and figure legends in lines 349–353).
Comment 4: I would suggest showing the boundaries of NTD and RBD regions discussed in Lines 119-120 in Figure S1. Actually, I guess that those regions should be marked on the sequence shown in Figure S1. Please refer to Figure S2 according to its meaning.
Response 4: Thank you for your insightful feedback. In response, we have updated Figure S1 to clearly mark with underlines the boundaries of NTD (residues 13–305) and RBD (residues 319–541) on the amino acid sequence of the S1 subunit. In addition, we have also updated Figure S2 to clearly indicate in the same way the major domains of the S2 subunit – i.e., FP (residues 788–806), CH (residues 987–1035) and TM (residues 1213–1237). Finally, we have revised the manuscript to refer to Figures S1 and S2 appropriately in the results section, ensuring that it aligns with the data on their respective domains (lines 164–168). We believe these changes enhanced the clarity and utility of both supplementary figures for readers.
Comment 5: Line 244: I guess that in Ref. [22] another lipid modification, namely, S-palmitoylation (not N-myristoylation) is discussed as that important for membrane fusion and virus entry.
Response 5: Thank you for pointing out this important detail in your careful review. You are correct that this reference discusses S-palmitoylation, not N-myristoylation, as being significant for membrane fusion and virus entry. We apologize for any confusion caused by this oversight. Although N-myristoylation has been removed from our dataset to address one of your previous concerns, we have retained Ref. 22 (now Ref. 25) because it also explores N-glycosylation, which is still pertinent to our study, adjusting the related text in the discussion section (lines 289–294). Thank you once again for your valuable input, which has helped us improve the accuracy and clarity of our manuscript.
Comment 6: It would be also great if the authors give a scheme of S-protein primary structure illustrating its main domains / motives listed in Lines 41-44 as a Figure.
Response 6: Thank you for your valuable suggestion. Although we have not included a new figure, we have addressed your suggestion by identifying the main domains/motifs of S protein in the primary structures of both its subunits shown in Figures S1 and S2 of the revised manuscript. Together, these figures now indicate the extension of N-terminal domain (NTD), receptor-binding domain (RBD), fusion peptide (FP), central helix (CH) and transmembrane domain (TM), in addition to the precise position of predicted glycosylation and phosphorylation sites. Also, the exact region that each domain spans in the primary structure of S protein has been described in lines 41–46 of the introduction section.
Comment 7: I did not find in the Results section the nucleotide mutations discussed in the Discussion Section (Line 217). Please add some factual data regarding nucleotide substitutions to the Results (or to Supplemental material), or indicate that you are not discussing your own data, but data obtained by other authors (with a reference).
Response 7: Thank you for your valuable feedback, we appreciate your careful reading of our manuscript. To address your pertinent concern about the nucleotide substitutions mentioned in the discussion section, we have revised the manuscript to explicitly state that these mutations are based on findings from previous studies, using the expression “as demonstrated earlier” in the beginning of the corresponding sentence (lines 265–266), whose reference (Ref. 20) is at the end of the contiguous sentence reflecting on the possible impacts of these mutations for the virus (line 271). Thank you once again for your insightful comments, which have undoubtedly enhanced the clarity and accuracy of our manuscript.
Reviewer 3 Report
Comments and Suggestions for AuthorsThe paper "Influence of Mutations on Physicochemical Properties of Spike Proteins from Prototypical SARS-CoV-2 Variants of Concern Detected in Amazonian Countries" by Silva and Carvalho collects evidence from several computational tools in order to identify the biochemical/structural impact of Variants of Concer (VOCs) of the SARS-CoV-2 Spike protein. The paper is well written and the experimental design is sound, but I have a few points to raise.
- Line 63: please explicitly state which are the "Amazonian countries" analyzed as the definition can be subjective or too vague. For example, I see the authors have not included samples from Bolivia, Ecuador, Suriname, Venezuela or French Guyana. The authors can keep the term "Amazonian countries" throughout the paper, but it should be defined here in the Introduction, and not later in the Discussion (line 207).
- Line 73: why did the authors use Clustal Omega and not, e.g., MUSCLE or Probcons, which are faster? I think it's because of the higher accuracy of Clustal Omega, but the authors should justify their choice (since it's not necessarily the standard tool for the task).
- Figure 2A: the figure is unclear, the ruler (0 10 20 30...) should have units of measurement stated either on the figure itself or in the legend.
- Figure 2B: it is really hard to appreaciate the differences between Omicron and Non-Omicron. The fact that Neutral and Basic differences are super significant (p<0.0001 !) is based on a few pixels. the authors should move this picture to supplementary, and substitute this with a four panel figure, with each class of polarities having their own y-axis. In this case, the differences can be appreciated, while leaving the grouped plot in the supplementary will provide a comparable (although less informative) figure.
- The same considerations for Figure 2A and 2B are valid for Figures 3A and 3B.
- Figure 5. This is a great figure but it somehow lacks on the comparative side. Would it be possible to highlight which regions have significantly different antigenicity? E.g. by calculating, in each 10aa window, the standard deviation in antigenicity across the 8 sequences? This would add a 9th line plot to the figure, which will show where (in terms of Spike aa coordinates) the VOCs are more or less varying.
Author Response
Comment 1: Line 63: please explicitly state which are the "Amazonian countries" analyzed as the definition can be subjective or too vague. For example, I see the authors have not included samples from Bolivia, Ecuador, Suriname, Venezuela or French Guyana. The authors can keep the term "Amazonian countries" throughout the paper, but it should be defined here in the Introduction, and not later in the Discussion (line 207).
Response 1: Thank you for your suggestion. In response, we have revised the manuscript to clearly define the Amazonian countries in the introduction section (lines 61–62) by listing the nine countries that extend across the Amazon basin (i.e., Brazil, Peru, Colombia, Bolivia, Venezuela, Guyana, Suriname, Ecuador and French Guiana, in order of Amazon tree cover), all initially considered for retrieval of S protein sequences. Despite this inclusion criterion informed in lines 72–74 of the materials and methods section, Brazil, Peru and Colombia were the only countries that contained the prototypical SARS-CoV-2 VOCs of the Amazon region according to the NCBI Virus database, as shown in Table 1. However, we believe that this issue not undermines our representativeness of the region of interest, since Brazil, Peru and Colombia are the largest Amazonian countries by size and together hold 80% of the Amazon tree cover (60, 12 and 8%, respectively), as stated in the discussion section (lines 256–259). We hope this clarification addresses your concern and improves the understanding of our study design.
Comment 2: Line 73: why did the authors use Clustal Omega and not, e.g., MUSCLE or Probcons, which are faster? I think it's because of the higher accuracy of Clustal Omega, but the authors should justify their choice (since it's not necessarily the standard tool for the task).
Response 2: Thank you for your insightful comment. While MUSCLE and ProbCons are indeed faster, Clustal Omega is known for its higher accuracy in multiple sequence alignments, as well as being specifically optimized to handle long sequences. Our study prioritized accuracy over speed to ensure the reliability of the results, especially given the complexity of the analyzed sequences, making Clustal Omega more suitable than MUSCLE and ProbCons for the task. We have highlighted this improved accuracy and how it is achieved by Clustal Omega in the materials and methods section (lines 84–90) to clarify our choice of sequence alignment tool. We hope this addresses your concern and provides a clear rationale for our decision.
Comment 3: Figure 2A: the figure is unclear, the ruler (0 10 20 30...) should have units of measurement stated either on the figure itself or in the legend.
Response 3: Thank you for your valuable feedback regarding the clarity of Figure 2A. We understand the importance of including units of measurement for the ruler to ensure the data is easily interpretable. Since adding a title for the ruler in a radar chart could clutter the visualization, making it harder to read and interpret, we have revised the figure legend to clearly state the units of measurement for the ruler (lines 199–200).
Comment 4: Figure 2B: it is really hard to appreaciate the differences between Omicron and Non-Omicron. The fact that Neutral and Basic differences are super significant (p<0.0001 !) is based on a few pixels. the authors should move this picture to supplementary, and substitute this with a four panel figure, with each class of polarities having their own y-axis. In this case, the differences can be appreciated, while leaving the grouped plot in the supplementary will provide a comparable (although less informative) figure.
Response 4: Thank you for your feedback regarding Figure 2B. To address your concerns about the difficulty in appreciating the differences between omicron and non-omicron variants, we have enhanced the clarity of the existing figure by introducing a discontinuity (identified by the symbol “≈”) on its y-axis to decompress lower values and make the differences more apparent, as the grouped plot is essential for providing a holistic view and direct comparison across all classes of polarities. Moving the figure to the supplementary materials and splitting it into separate panels could fragment the data, potentially making it harder to perceive the relative proportion of the different classes. We believe this approach will help readers better understand the data without losing the comprehensive perspective offered by the grouped plot. Finally, to also address your concerns about statistical significance, as the focus of our hypothesis testing was to compare only two specific groups (non-omicron vs. omicron) rather than examining interactions between factors, ordinary two-way analysis of variance (ANOVA) was replaced by unpaired t test with Bonferroni correction to mitigate the risk of type I errors (lines 145–149), providing a more appropriate analysis.
Comment 5: The same considerations for Figure 2A and 2B are valid for Figures 3A and 3B.
Response 5: Thank you for your keen observation. In response, we have made changes to Figures 3A and 3B similar to those applied to Figures 2A and 2B (i.e., clarification of the units of measurement for the ruler in line 216 of the legend of Figure 3A, introduction of a discontinuity on the y-axis of Figure 3B to make the differences between the groups more apparent and use of a more appropriate statistical hypothesis test). We believe these adjustments will address your concerns and enhance the clarity and informativeness of Figures 3A and 3B.
Comment 6: Figure 5. This is a great figure but it somehow lacks on the comparative side. Would it be possible to highlight which regions have significantly different antigenicity? E.g. by calculating, in each 10aa window, the standard deviation in antigenicity across the 8 sequences? This would add a 9th line plot to the figure, which will show where (in terms of Spike aa coordinates) the VOCs are more or less varying.
Response 6: Thank you for your positive feedback on Figure 5 and for your insightful suggestion to enhance its comparative aspect. In response, we have revised this figure to include arrows (leftwards for the NTD and righwards for the RBD) indicating regions with relatively pronounced antigenic propensity in the S proteins of SARS-CoV-2 VOCs. As requested, we have also calculated the standard deviation (SD) of Parker antigenicity score for each aligned amino acid position across the eight protein sequences, clearly indicating where the VOCs exhibit more or less variability in terms of antigenic propensity. Instead of adding this new plot as a ninth line in Figure 5, we have included it as a supplementary figure (Figure S3) to avoid crowding the main figure (figure callout in lines 242–243 and figure legend in lines 353–355). We believe this addition will provide a more comprehensive and comparative view of the antigenicity variations across the sequences, addressing your suggestion and improving the overall informative value of the figure. Thank you once again for your valuable input, which has significantly enhanced the quality of our manuscript.
Round 2
Reviewer 2 Report
Comments and Suggestions for AuthorsAll my concerns were corrected. The paper warrants publication.
Author Response
Comment: All my concerns were corrected. The paper warrants publication.
Response: Thank you for your positive feedback and for confirming that all your concerns have been addressed. We are pleased to hear that you consider our paper warrants publication.
Reviewer 3 Report
Comments and Suggestions for AuthorsI am positively impressed by the authors for putting real effort in revising their manuscript. I like their smart solution to the issues in Figure 2. I also think that calculating SD for figure 5 was a nice addition. However, I really think that the value of the SD profile (now shown as Figure S3) is when compared with the other profiles in Figure 5, which is indeed a two-page figure, but it is already so without the extra panel, and so overcrowdedness won't get worse.
Despite this small point, I think the manuscript is informative and complete enough for publication.
Author Response
Comment: I am positively impressed by the authors for putting real effort in revising their manuscript. I like their smart solution to the issues in Figure 2. I also think that calculating SD for figure 5 was a nice addition. However, I really think that the value of the SD profile (now shown as Figure S3) is when compared with the other profiles in Figure 5, which is indeed a two-page figure, but it is already so without the extra panel, and so overcrowdedness won't get worse. Despite this small point, I think the manuscript is informative and complete enough for publication.
Response: Thank you for your encouraging feedback and for acknowledging our efforts in revising the manuscript. We appreciate your positive comments on the solutions implemented in Figure 2 and the addition of standard deviation calculations for Figure 5. Regarding your suggestion to include the SD profile from Figure S3 in Figure 5, we understand your point about the potential benefits of this comparison and we have integrated the SD profile into Figure 5 to enhance the comparison with the other profiles as recommended. We are grateful for your thorough review and constructive feedback, which have significantly improved our manuscript.
