Structural Diversity and Differential Natural Pairings of MAT1-1-1 and MAT1-2-1 Proteins Essential for Sexual Reproduction in Ophiocordyceps sinensis Strains

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript by X.-Z. Li et al. is aimed at the elucidation of the mating strategy of fungus O. sinensis. This topic is of interest due to clinical importance and commercial value of this species and the necessity of its cultivation management. Genomes of studied strains simultaneously encode two mating proteins, MAT1-1-1 and MAT1-2-1. Based on this co-occurrence, earlier studies suggested that this fungus is self-fertilized. However, the authors of the present work perform a thorough analysis of the possible combinations of MAT proteins in each strain, and conclude that all these strains possess truncated versions of either one of the MAT proteins or both of them. The authors conclude that these strains could be self-sterile. This suggestion conforms with some previously published observations, e.g., inability of a monoculture of this species to reproduce.

While I appreciate the deep analysis of MAT proteins combinations, and the corresponding classification of strains which can be useful in future research, I still have several critical comments to this work and its presentation.

My general concern to the work: while the suggestion of self-sterility can be correct, the results reported in this manuscript do not indicate that. Even if the truncations affect the structure and properties of proteins, the truncated proteins can still be functional as transcriptional regulators. The authors themselves imply this (e.g., lines 642-647). Therefore, if both types of proteins are produced and functional, they both can contribute to self-fertilized manner of reproduction. Briefly, I cannot find any arguments against this strategy. So, I would suggest to make this conclusion more humble, or provide additional argumentation.

My comments to the manuscript:

1. The figures are not illustrative. They live separately of the text, while they have to provide the basis for text understanding. Examples:

• Figures 1 and 3: sequence alignments can be omitted here, since the following schematical illustrations on Figs. 2 and 4 are sufficient for understanding what parts of a chain are truncated. So, only reference sequences of full-length proteins without alignments can be provided.
• Figures 6 and 7 (and corresponding Figs. In Supplementary): they are meant to illustrate the differences between full-length protein and the truncated variants, however, this difference is not seen here. The plots shown in lower panels look the same for FL and truncated variants. The corresponding structures (meant for the pairwise comparison) are shown in different orientation, so it is difficult to compare them. The structure on Fig.6 is shown from such a point that the extra fragment of a chain (lost in truncated variant) is not seen. Structural figures are colored according to AlfaFold confidence, which is not related to the text or this work in general, however, the DNA-binding domain or three alfa-helices mentioned in the text several times (as important to the conclusions) are not highlighted.  Some elements of a structure are boxed in a circle and some text is provided (I think, their terminal sequences), but the legend do not indicate what it is. (again, I think it is the site of truncation, but it has to be explicitly indicated).
• If the authors want to illustrate the difference in structure, I would suggest showing the structure alignments between full-length and truncated proteins. This way, both structures will be in the same orientation, and the truncation site and any possible differences will be easily detected.
• If the authors want to illustrate the difference in properties, I would suggest showing the differential plots, i.e. not the absolute values of, say, hydrophobicity for two proteins but the result of their subtraction.
• The plots shown by the authors have little to do with functional properties, and anyway the distribution of these properties along the remaining protein chain would be almost unaffected by the truncation. I would suggest to illustrate some other properties that indeed can be affected by the truncation. First of all, the DNA-binding domains and other functionally important elements have to be highlighted (e.g., colored in the alternative colors). This would show where the truncation is positioned compared to these functional regions. Second, truncation creates the alternative surface, which is a primary site of recognition between DNA and protein, so it would be great to show the changes in the surface (instead of hydrophobicity of individual amino acids which remains almost the same). I would suggest one panel for structure alignment, (shown in secondary structure elements), and another pair of panels showing the close view of the surfaces of the two proteins (FL and truncated) in comparison.

2. Introduction provides only superficial information on the DNA-binding properties of studied proteins, without information on the functionally important structural elements or probably residues involved in DNA recognition which is key to this paper content. 
3. The Tables contain repetitive information which also is present on some figures and sometimes in the text. For instance, Table 2 contains GenBank accession numbers for the protein sequences which are also present in Table 1 and Figures 2 and 4. In some cases, it makes it easier to understand the numerous data, however, I would suggest to remove the unnecessary repetitions where it is possible.
4. As far as I understand, Table S3 in Supplementary does not contain the authors' results, it is the values that ProtScale uses for their calculations. If this is the standard values and not some special scale developed by the authors, the table has to be removed with the corresponding reference to the literature. 
5. Mutations (mentioned in Table 3 and in the text) should be given with the residue number, for example, Y143H, or Tyr143His, or suchlike. If case of different numbering (shifted as a result of truncations), the numbering can be given according to the reference protein, e.g. “Y143H (Full-length protein numbering)”.
6. The text is very long. I would suggest to make it more concise and remove repetitions. For example, Results section can be shortened if the authors generalize their findings without listing all the results for all the strains.
7. Several conclusions in Discussion are not exactly supported by the reported results. Examples:

• (lines 621-623) “These results provide protein-level evidence for evaluating how structural diversity… may influence DNA binding specificity”: functional properties were not studied in this paper, including DNA binding specificity, so the conclusion is not supported.
• (line 999) “This study provides protein structure evidence consistent with self-sterility…” – not supported.
• (line 22) "using the AlphaFold-based structural modeling" - as far as I understand, the authors did not perform modelling, they used the models deposited earlier to the models database. If I am wrong here, I apologize; however, in this case modeling protocol and validation metrics have to be included in the text.

To sum up, the paper cannot be accepted in the present form.

Author Response

1. Reviewer #1 commended: “the authors of the present work perform a thorough analysis of the possible combinations of MAT proteins in each strain, and conclude that all these strains possess truncated versions of either one of the MAT proteins or both of them. The authors conclude that these strains could be self-sterile. This suggestion conforms with some previously published observations, e.g., inability of a monoculture of this species to reproduce.”

Authors’ response:

We sincerely appreciate Reviewer #1 for this rigorous and constructive assessment and for acknowledging the significance of our structural analysis. Below are our point-by-point responses to all the concerns raised.

The present study focuses on characterizing the structural diversity of MAT1-1-1 and MAT1-2-1 proteins and, with a specific focus on their differential native pairing patterns across 20 “purportedly pure” O. sinensis strains. As outlined in the revised manuscript (Lines 108–115):

Li et al. [65−68] reported differential occurrence, alternative splicing, and differential transcription of mating-type and pheromone receptor genes in the genomes of H. sinensis strains, which are inconsistent with self-fertilization but consistent with self-sterility. Li et al. [65,69] described differential occurrence of the MAT1-1-1 and MAT1-2-1 proteins with heteromorphic tertiary structures of the DNA-binding domains in wild-type C. sinensis isolates, which may have originated from either GC-biased Genotype #1 or Genotype #3 of O. sinensis.

Collectively, these genomic, transcriptomic, and protein structural datasets provide irrefutable evidence for the self-sterility hypothesis for O. sinensis. This body of prior work motivated our current in silico study, which leverages publicly available datasets to analyze the differential native pairings of structurally variant MAT1-1-1 and MAT1-2-1 proteins simultaneously co-produced by the 20 O. sinensis strains.

The 20 O. sinensis strains examined herein were originally reported to be homogeneous pure cultures [14,36]. However, some phylogenic datasets simultaneously released by these authors may not be fully consistent with their claim of clonal homogeneity. Notably, the authors reported that the 20 O. sinensis strains simultaneously expressed both MAT1-1-1 and MAT1-2-1 proteins without any omissions. It remains unclear whether structurally modified MAT1-1-1 and MAT1-2-1 proteins might carry attenuated, altered, or even augmented biological functions.

 

2. Reviewer #1 commended: “My general concern to the work: while the suggestion of self-sterility can be correct, the results reported in this manuscript do not indicate that. Even if the truncations affect the structure and properties of proteins, the truncated proteins can still be functional as transcriptional regulators. The authors themselves imply this (e.g., lines 642-647). Therefore, if both types of proteins are produced and functional, they both can contribute to self-fertilized manner of reproduction. Briefly, I cannot find any arguments against this strategy. So, I would suggest to make this conclusion more humble, or provide additional argumentation.”

Authors’ response:

We appreciate Reviewer #1 for this thoughtful and constructive critique. We fully agree that truncations or amino acid substitutions within a protein domain do not necessarily abolish biological function.

Before addressing the functional implications of the structural variants, we wish to clarify that the self-sterility hypothesis was not grounded solely in the protein structural analyses presented in this study. Instead, it arises from the integration of genomic, transcriptomic, and protein structural evidence (Lines 108–115), including the differential occurrence, alternative splicing, and differential transcription of mating-type and pheromone receptor genes in the genomes of H. sinensis strains and the differential natural occurrence of MAT1-1-1 and MAT1-2-1 proteins in wild-type, heterogeneous C. sinensis isolates with heteromorphic tertiary structures of the DNA-binding domains.

The present in silico study further characterizes the differential natural pairing patterns of structurally diverse MAT1-1-1 and MAT1-2-1 proteins across the 20 O. sinensis strains. As described in Section 2.6, we identified a critical discrepancy: genome assemblies from 6 pure H. sinensis strains exhibit variable presence/absence profiles for mating-type genes, whereas all 20 O. sinensis strains examined in this study uniformly co-expressed naturally paired MAT1-1-1 and MAT1-2-1 proteins, with no exceptions.

The revised manuscript states the following (Lines 610–614):

Table 3 summarizes the reported co-occurrence and differential occurrence patterns of MAT1-1-1 and MAT1-2-1 proteins across C. sinensis insect‒fungal complexes, wild-type C. sinensis isolates, and O. sinensis strains. The MAT1-1-1 and MAT1-2-1 proteins individually occurred in 52.7% and 22.3% of the 184 samples, respectively, but co-occurred in only 25.0% of the samples.

Approximately three quarters of all surveyed samples lack one of the 2 mating proteins. These uneven distribution patterns are difficult to reconcile with a strictly homothallic reproductive mode and instead are consistent with the previously proposed self-sterility hypothesis.

We further note the following in Lines 624–627:

This differential presence/absence pattern observed at the genomic level in H. sinensis strains, however, was not replicated at the protein level in the dataset analyzed herein. All 20 O. sinensis strains simultaneously produced naturally paired MAT1-1-1 and MAT1-2-1 proteins without omissions (Tables 1−2; Figure 3).

The reviewer's interpretation that co-expressed mating proteins could enable homothallic reproduction rests on the assumption that the paired MAT1-1-1 and MAT1-2-1 proteins originate from a single genetically homogeneous fungal strain. Our analysis of the available datasets leads us to a different interpretation: the observed paired MAT1-1-1 and MAT1-2-1 proteins detected within each strain may be derived from multiple co-occurring fungal taxa rather than from a single fungal genome. We interpret the collected datasets in Lines 641–653:

The above observations indicate that the 20 O. sinensis strains simultaneously produced naturally paired MAT1-1-1 and MAT1-2-1 protein variants, which exhibited N- and/or C-terminal truncations together with diverse amino acid substitutions at distinct sites. These findings underscore the extensive genetic variation within the MAT loci of distinct fungal genomes residing within the impure O. sinensis cultures, irrespective of their designation as pure strains previously reported by Li et al. [36]. Unlike the variable presence/absence patterns detected across the genome assemblies of the 6 H. sinensis strains, all 20 O. sinensis strains analyzed in this study consistently co-expressed MAT1-1-1 and MAT1-2-1 proteins as natural pairs with divergent domain architectures. Given that no repetitive copies of the MAT1-1-1 or MAT1-2-1 genes exist in the H. sinensis genome [69], the differential pairing patterns of MAT1-1-1 and MAT1-2-1 proteins observed herein are compatible with contributions from multiple co-occurring fungal taxa to the observation of differential natural pairings of the MAT1-1-1 and MAT1-2-1 proteins within these strains.

To clarify our interpretation, we revised the concluding sentence of this paragraph as follows:

the contrast between genomic and protein-level occurrence patterns suggests multiple fungal contributors to the mating proteins within these impure O. sinensis strains.

to the revised version (Lines 627–630):

the contrasts between the genomic and protein-level occurrence patterns suggest the possibility that multiple co-occurring fungi within these impure O. sinensis strains contributed to the naturally paired mating proteins, including both authentic and variant forms.

With respect to protein functionality, we have already addressed this point in our response to Comment #2 from the Academic Editor. Throughout the revised manuscript, we have adopted far more cautious language to clearly separate objective experimental observations from interpretive hypotheses. We explicitly acknowledge that dedicated functional assays will ultimately be needed to determine whether the structurally variant MAT1-1-1 and MAT1-2-1 proteins retain, attenuate, or even enhance their biological activities.

Importantly, our core interpretation does not hinge entirely on the functionality of the structurally variant proteins. Rather, it is supported by integrated evidence from previous genomic and transcriptomic studies together with the differential natural pairing patterns identified in the present protein-level analyses. Collectively, these multilayered observations are consistent with, but do not by themselves prove, the hypothesis that compatible mating partners originating from distinct co-occurring fungi mediate either heterothallic mating (involving 2 genomically distinct, self-sterile O. sinensis) or interspecies hybrid reproduction rather than strictly homothallic selfing.

We have revised the manuscript accordingly to emphasize that this represents a hypothesis supported by multiple complementary findings and that targeted experimental validation is needed to discriminate between alternative reproductive mechanisms.

 

3. Reviewer #1 commended: “1. The figures are not illustrative. They live separately of the text, while they have to provide the basis for text understanding. Examples: ”

• Figures 1 and 3: sequence alignments can be omitted here, since the following schematical illustrations on Figs. 2 and 4 are sufficient for understanding what parts of a chain are truncated. So, only reference sequences of full-length proteins without alignments can be provided.
• Figures 6 and 7 (and corresponding Figs. In Supplementary): they are meant to illustrate the differences between full-length protein and the truncated variants, however, this difference is not seen here. The plots shown in lower panels look the same for FL and truncated variants. The corresponding structures (meant for the pairwise comparison) are shown in different orientation, so it is difficult to compare them. The structure on Fig.6 is shown from such a point that the extra fragment of a chain (lost in truncated variant) is not seen. Structural figures are colored according to AlfaFold confidence, which is not related to the text or this work in general, however, the DNA-binding domain or three alfa-helices mentioned in the text several times (as important to the conclusions) are not highlighted. Some elements of a structure are boxed in a circle and some text is provided (I think, their terminal sequences), but the legend do not indicate what it is. (again, I think it is the site of truncation, but it has to be explicitly indicated).
• If the authors want to illustrate the difference in structure, I would suggest showing the structure alignments between full-length and truncated proteins. This way, both structures will be in the same orientation, and the truncation site and any possible differences will be easily detected.
• If the authors want to illustrate the difference in properties, I would suggest showing the differential plots, i.e. not the absolute values of, say, hydrophobicity for two proteins but the result of their subtraction.
• The plots shown by the authors have little to do with functional properties, and anyway the distribution of these properties along the remaining protein chain would be almost unaffected by the truncation. I would suggest to illustrate some other properties that indeed can be affected by the truncation. First of all, the DNA-binding domains and other functionally important elements have to be highlighted (e.g., colored in the alternative colors). This would show where the truncation is positioned compared to these functional regions. Second, truncation creates the alternative surface, which is a primary site of recognition between DNA and protein, so it would be great to show the changes in the surface (instead of hydrophobicity of individual amino acids which remains almost the same). I would suggest one panel for structure alignment, (shown in secondary structure elements), and another pair of panels showing the close view of the surfaces of the two proteins (FL and truncated) in comparison.

Authors’ response:

We thank Reviewer #1 for these detailed and constructive suggestions to improve figure clarity and interpretability. We agree that the original figures could be improved to facilitate an intuitive visual comparison between the reference and variant proteins. We have therefore comprehensively revised both the figures and the accompanying legends accordingly, as detailed below.

(1) Sequence alignment figures

Following the recommendations of both the Academic Editor and Reviewers, we have relocated the original main-text sequence alignment in Figures 1 and 3 to the Supplementary Materials (now Figures S1 and S2). These supplementary figures preserve complete sequence alignment information on both the N- and/or C-terminal truncations and the various amino acid substitutions at different sites for readers requiring full primary-sequence detail.

(2) New pairwise structural superposition panels

Following the editorial and reviewer’s guidance, we performed pairwise structural superposition to facilitate direct visual comparison between reference authentic proteins and mating protein variants. Revised main-text Figures 4–5 and Supplementary Figures S3–S9 in the revision (originally Figures 6–7 and Figures S1–S7) each contain 2 complementary side-by-side superposition views:

• overall structural superpositions (on the left in each panel) generated using the Template-Model align (TM-align) algorithm, which compares the overall folding architectures of reference and variant proteins. The 3 critical α-helices within the DNA binding domain are green-outlined highlighted.
• locally magnified structural superpositions computed using the Smith–Waterman 3D algorithm to maximize local structural similarity/dissimilarity and optimized to resolve subtle conformational divergence within the 3 α-helices that form the hydrophobic core of the DNA-binding domains.

These dual complementary structural superposition views facilitate direct comparisons of both the overall protein folds and the local subtle structural shifts within the functionally essential hydrophobic core of the DNA-binding domains.

(3) Technical constraints on differential ProtScale subtraction plots

We thoroughly evaluated the reviewer's suggestion to generate differential subtraction plots (variant minus reference) rather than the raw independent ProtScale profiles. Unfortunately, this approach is technically unfeasible for ExPASy ProtScale output. ProtScale metrics are derived using sliding-window algorithms, where each plotted value is calculated sequentially from multiple neighboring residues rather than corresponding to a single amino acid. Our analyses adopted a default 9-residue sliding window with sequentially normalized weighting values for all physicochemical scale calculations.

Consequently, directly subtracting two independent ProtScale traces would not accurately represent the underlying physicochemical changes and could even be misleading. For this reason, we retained the original ProtScale profiles. However, to improve direct visual comparability and readability, we added multiple dashed auxiliary lines, arrows, and annotated frames to the plots to highlight the shifts in the peak height/amplitude, trough depth, peak shape and overall waveform topology. As described in the figure legend for Figure 5, for instance (Lines 367–372):

ExPASy ProtScale analyses are shown in Panels (B)-(F), illustrating changes in hydrophobicity and predicted secondary-structure characteristics, including hydropathy, α-helices, β-sheets, β-turns, and coils, respectively. The open blue rectangles highlight the C-terminal truncation regions in the ExPASy plots, whereas the open green rectangles or circles with red dashed auxiliary lines or arrows highlight changes in the topological configuration and waveform patterns due to amino acid substitutions.

The modified ExPASy ProtScale plots improve the comparability and readability of the ExPASy ProtScale structural profiles.

(4) Follow-up plan for molecular surface functional analyses

We greatly appreciate the reviewer’s recommendation to compare molecular surfaces involved in DNA recognition.

We agree that analyses of electrostatic surface profiling, DNA-binding interfaces, protein–DNA docking, and protein–protein interaction modeling would yield valuable complementary functional insights. However, such functional simulations extend beyond the defined scope of the present study, which focuses exclusively on characterizing differential natural pairing patterns of structurally diverse MAT1-1-1 and MAT1-2-1 proteins using publicly available sequence and precomputed AlphaFold structural datasets. We have therefore included this direction among the future independent studies proposed in the Discussion section. We anticipate that full-length cDNA sequencing from pure H. sinensis strains and wild-type impure C. sinensis isolates, together with the corresponding full-length protein structural characterization, will provide a more appropriate experimental foundation for subsequent functional surface studies and experimental validation.

We believe that in accordance with these suggestions, this revision, in which structural superposition panels are added and ProtScale plots are modified with auxiliary marking, substantially improves the visual comparison of the reference and variant proteins and directly addresses the reviewer's concern regarding orientation and structural comparability while maintaining the primary objective of the present work.

 

4. Reviewer #1 commended: “ Introduction provides only superficial information on the DNA-binding properties of studied proteins, without information on the functionally important structural elements or probably residues involved in DNA recognition which is key to this paper content.”

Authors’ response:

We thank Reviewer #1 for this helpful suggestion. We agree that a brief description of the structural features underlying DNA binding provides a clearer conceptual foundation for understanding the subsequent structural comparisons presented in the paper.

In the original manuscript, we described the DNA-binding domains of the 2 mating proteins as follows (Lines 103−108):

The MAT1-1-1 protein harbors a mating-type alpha high mobility group box (MATα_HMGbox) domain, whereas the MAT1-2-1 protein contains a high mobility group box ROX1-like (HMG-box_ROX1-like) domain [14,58,63−65]. These DNA-binding domains regulate the transcription of genes associated with sexual reproduction.

In the revised Introduction, we have supplemented key structural details by introducing the core structural feature central to our analyses (Lines 106−107):

…… Each DNA-binding domain contains a hydrophobic core formed by 3 asymmetrically oriented α-helices.

Furthermore, throughout the revised figures, we visually highlight these core 3 α-helices with green-outlines in the structural alignments of the AlphaFold 3D structural models, especially when the SW 3D algorithm is used to maximize the resolution of the local folding similarity/divergences within the conserved hydrophobic cores, enabling straightforward identification of structural variations across the mating protein variants.

 

5. Reviewer #1 commended: “ The Tables contain repetitive information which also is present on some figures and sometimes in the text. For instance, Table 2 contains GenBank accession numbers for the protein sequences which are also present in Table 1 and Figures 2 and 4. In some cases, it makes it easier to understand the numerous data, however, I would suggest to remove the unnecessary repetitions where it is possible.”

Authors’ response:

We thank Reviewer #1 for this practical suggestion. We agree that unnecessary repetition should be avoided where possible. Accordingly, we carefully reviewed the tables and figures to minimize redundant content while preserving overall data clarity.

Although Tables 1 and 2 share some overlapping labels (including strain IDs and ITS sequences and protein GenBank accession numbers), they are organized to report different aspects of our dataset and thus fulfill complementary analytical roles. Our analysis covered 20 O. sinensis strains, each of which simultaneously co-produced one matched pair of MAT1-1-1 and MAT1-2-1 proteins. Thus, careful multiangle data presentation is essential to unpack this complex dataset. The unique design logic of each table and revised figure is outlined below:

• Table 1 catalogs the 20 tested sinensis strains by integrating strain identifiers with their phylogenetic ITS information, mating-protein accession codes, and links to their respective AlphaFold structural models. Critically, the 20 O. sinensis strains are sorted according to their published phylogenetic classification (genetically homogeneous, heterogeneous, or uncharacterized);
• Table 2 focuses on the structural characteristics of the naturally paired MAT1-1-1 and MAT1-2-1 proteins simultaneously produced by each of the 20 sinensis strains. Here, the 20 strains are reclassified purely on the basis of the structural features of the paired mating proteins, with subgroups defined by distinct AlphaFold structural morphotypes (marked with unique suffix symbols). This grouping scheme streamlines comparisons of differential natural pairing patterns across samples:

Group I: N- and C-terminally truncated MAT1-1-1 proteins naturally paired with C-terminally truncated MAT1-2-1 proteins; and

Group II: N- and C-terminally truncated MAT1-1-1 proteins naturally paired with full-length MAT1-2-1 proteins.

• Figures 1−2 in the revised manuscript (previously Figures 2 and 4) display schematic comparisons of the sequence distributions and sequence ranges of the DNA binding domains of the mating proteins rather than the complete protein sequences. To enable direct cross-matching between each schematic representation and the corresponding strain and protein, minimal identifying information (strain numbers, protein accession IDs, domain coordinate ranges, and associated 3D structure models) has been retained.

Therefore, although some identifiers necessarily appear in more than one table or figure, each occurrence sits within a separate organizational framework and serves a separate analytical purpose. Removing these cross-referencing labels would require readers to repeatedly toggle between separate tables and figures to match strains, proteins, and their structural data, thereby hindering rather than improving readability.

 

6. Reviewer #1 commended: “ As far as I understand, Table S3 in Supplementary does not contain the authors' results, it is the values that ProtScale uses for their calculations. If this is the standard values and not some special scale developed by the authors, the table has to be removed with the corresponding reference to the literature.”

Authors’ response:

We appreciate Reviewer #1 for raising this critical point. It is correct that Table S3 does not present the results generated in the present study. Accordingly, Table S3 is included in the Materials and Methods section (rather than the Results section) because it fully documents the amino acid scales used by the selected ExPASy ProtScale algorithms for computational reproducibility.

We included Table S3 to document the computational parameters used in our analyses, thereby facilitate reproducibility and readability. Changes in hydropathy index values that are listed in Table S3 are frequently discussed in our manuscript in relation to amino acid substitutions, as these changes indicate altered hydrophobicity. As shown on the ProtScale server (https://web.expasy.org/protscale/), the platform hosts a total of 57 literature-derived predefined amino acid scales covering diverse physicochemical and secondary structural features. Multiple independent scales are available for a single property; for example, the platform includes 23 distinct hydropathy scales and 3 separate sets of propensity values for α-helix, β-sheet, β-turn les, and random coil. To ensure consistent comparative analysis across all protein variants in the present study, we consistently selected the Deleage & Roux propensity set [73] (which covers α-helix, β-sheet, β-turn and coil metrics) and the Kyte & Doolittle hydropath propensity scale [74].

To eliminate potential confusion for readers, we have revised both the legend of Table S3 and the Materials and Methods section to explicitly state that the listed values are published reference scales implemented in ExPASy ProtScale rather than values generated in the present study. We have cited the appropriate reference papers for the original scale developers and for the ExPASy ProtScale resource. The updated explanatory note in Table S3 reads as follows:

Note: An amino acid scale is defined at https://web.expasy.org/protscale/ by a numerical value assigned to each type of amino acid. The most frequently used scales are the hydrophobicity or hydrophilicity scales and the secondary structure conformational parameter scales, but many other scales exist, which are based on the different chemical and physical properties of the amino acids. The ExPASy ProtScale program provides 57 predefined scales based on the literature [74]. *, Hydropathy index [73]; the larger the value, the stronger the hydrophobicity; negative values indicate hydrophilicity.

 

7. Reviewer #1 commended: “ Mutations (mentioned in Table 3 and in the text) should be given with the residue number, for example, Y143H, or Tyr143His, or such like. If case of different numbering (shifted as a result of truncations), the numbering can be given according to the reference protein, e.g. “Y143H (Full-length protein numbering)”.”

Authors’ response:

We appreciate this constructive suggestion from Reviewer #1. We agree that using standard mutation nomenclature greatly improves the clarity and consistency of the sequence comparisons across our dataset.

Owing to the extensive truncations and internal amino acid deletions across the protein variants, residue numbering cannot be consistently referenced to a unified full-length protein sequence. Alternatively, we have updated the main text, tables, and figures to report amino acid substitutions using the arrow format (e.g., Y¹⁴⁴→H¹⁴⁴, CDRA^65-68→SSFT^2-5).

 

8. Reviewer #1 commended: “ The text is very long. I would suggest to make it more concise and remove repetitions. For example, Results section can be shortened if the authors generalize their findings without listing all the results for all the strains.”

Authors’ response:

We appreciate this valuable suggestion from Reviewer #1. We agree that the original manuscript was longer than desirable. We have made substantial efforts to improve its conciseness while preserving the completeness of the comparative analyses.

Following the recommendations of the Academic Editor and Reviewers, we relocated the original protein sequence alignment figures (original Figures 1 and 3; now Figures S1–S2) to the Supplementary Materials. Combined with other text edits, the main manuscript length has been reduced from 45 to 38 pages.

We also carefully reviewed the text throughout the manuscript to remove repetitive statements wherever possible. Unlike typical structure studies that analyze a small set of representative proteins, our work investigates the differential natural pairing patterns of 40 structurally divergent MAT1-1-1 and MAT1-2-1 proteins across 20 distinct O. sinensis strains. To minimize redundancy, in particular, we consolidated similar observations for proteins sharing the same AlphaFold-predicted structural morph and described them collectively rather than repeatedly discussing each strain individually where feasible. We believe that this grouped presentation substantially improves readability while fully retaining the complete dataset evidence supporting our core conclusions.

We also revised several statements in the Discussion and Conclusions to clarify our observations and interpretations and to ensure that the conclusions remain appropriately supported by the available evidence.

 

9. Reviewer #1 commended: “ Several conclusions in Discussion are not exactly supported by the reported results. Examples:

• (lines 621-623) “These results provide protein-level evidence for evaluating how structural diversity… may influence DNA binding specificity”: functional properties were not studied in this paper, including DNA binding specificity, so the conclusion is not supported.
• (line 999) “This study provides protein structure evidence consistent with self-sterility…” – not supported.
• (line 22) "using the AlphaFold-based structural modeling" - as far as I understand, the authors did not perform modelling, they used the models deposited earlier to the models database. If I am wrong here, I apologize; however, in this case modeling protocol and validation metrics have to be included in the text.”

Authors’ response:

We thank Reviewer #1 for pointing out these critical mismatches between our original interpretations and the available experimental evidence. We fully acknowledge that several statements in the original manuscript were phrased too strongly relative to the evidence presented. Accordingly, we have comprehensively reworded the manuscript to distinguish more clearly between our structural observations, interpretations, and hypotheses, which require follow-up experimental validation.

(1) Functional implications of structural variation

The original statement (lines 621-623):

These results provide protein-level evidence for evaluating how structural diversity among differentially naturally paired mating proteins may influence DNA binding specificity and downstream regulation of mating-type transcriptional programs in O. sinensis.

could be interpreted as implying functional conclusions that were not directly investigated in the present study.

We have therefore rephrased this passage to adopt a tentative, hypothesis-driven tone (revised Lines 669–671):

The identified variations are associated with changes in hydrophobicity and secondary/tertiary structures, including the altered 3D conformations of the hydrophobic cores formed by 3 α-helices of each functional domain.

This revised wording clearly decouples the observed structural differences from their possible functional consequences.

(2) Conclusion regarding self-sterility

We agree with the original statement:

This study provides protein structure evidence consistent with self-sterility...

was insufficiently justified by the dataset presented in the current study. This sentence has therefore been completely deleted from the Conclusions section.

We have toned down all the relevant commentary throughout the Discussion and Conclusions to draw a clearer boundary between our empirical observations and interpretations. Specifically, our results are merely compatible with the self-sterility hypothesis but cannot independently verify it. The self-sterility hypothesis for O. sinensis was proposed on the basis of two lines of evidence (Lines 107–115):

1. Previous genomic and transcriptomic studies revealed differential occurrence, alternative splicing, and differential transcription of mating-type and pheromone receptor genes across 7 sinensis genome and transcriptome assemblies, and these studies concluded the self-sterility for O. sinensis [65−68];
2. More than 180 wild-type sinensis isolates display differential occurrence of the MAT1-1-1 and MAT1-2-1 proteins, whose DNA-binding domains adopt heteromorphic tertiary structures [65,69].

(3) Clarification of AlphaFold structural model usage

The reviewer’s assessment is accurate: we did not perform de novo AlphaFold structure modeling in this work. All 3D structures analyzed in our study were precomputed models downloaded from the public AlphaFold Protein Structure Database.

Accordingly, we have revised the Abstract as follows:

...using the AlphaFold-predicted structural modeling.

To (Lines 22–23)

... based on AlphaFold-predicted 3D structural models and pairwise structural superposition analyses.

We have clearly described this in the Materials and Methods section, stating that the deposited AlphaFold models were retrieved from the public database rather than generated in the present study (Lines 1023–1026):

To examine heteromorphic stereostructures in this study, the 3D structures of the MAT1-1-1 and MAT1-2-1 proteins from O. sinensis strains were predicted computationally from their amino acid sequences using AlphaFold, an artificial intelligence (AI)-driven machine learning platform [134−144].

 

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

In this study, we systematically evaluated the natural pairing patterns of co-expressed MAT1-1-1 and MAT1-2-1 proteins across 20 supposedly "homogeneously purified" Ophiocordyceps sinensis strains, utilizing AlphaFold structural modeling along with ExPASy ProtScale hydrophobicity and secondary structure analyses. Our findings reveal that the DNA-binding domains (MATA_HMGbox and HMG-box_ROX1-like) of these mating-type proteins exhibit distinct N-or C-terminal truncations in their primary sequences, accompanied by 1–4 amino acid mutations at critical sites. These variations significantly alter their hydrophobicity, secondary structures, and tertiary spatial conformations. Based on their pairing characteristics, the 20 strains were classified into two groups: Group I (5 strains) co-expressed truncated forms of both MAT1-1-1 and MAT1-2-1, whereas Group II (15 strains) expressed truncated MAT1-1-1 alongside full-length MAT1-2-1. Given that a single H. sinensis genome does not harbor duplicate copies of mating-type genes, the simultaneous detection of diverse, heterogeneous MAT protein pairings within an individual strain indicates that these strains are not mycologically pure, but rather harbor multiple co-existing fungal taxa (characteristic of a multi-fungal complex). This structural diversity provides evidence against a strict homothallic (selfing) model, strongly suggesting that O. sinensis favors a life-cycle strategy of self-sterility, heterothallic mating, or hybrid reproduction. However, the persuasiveness of the manuscript's conclusions is currently heavily constrained due to an over-reliance on purely in silico predictions, a lack of supporting biochemical activity assays, low confidence scores in certain sequencing data, and substandard graphical presentation.

1. The entire manuscript relies exclusively on AlphaFold predictions and in silico simulations, lacking any empirical biochemical or biophysical validation (such as X-ray crystallography, NMR, or Cryo-EM) to confirm the actual spatial conformations of these truncated and mutant proteins.
2. Although the authors speculate that the truncation of the core α-helix impairs its binding to AT-rich promoter elements, empirical experimental evidence is lacking. A truncation of 60 amino acids could directly trigger domain unfolding or lead to a complete loss of function, essentially rendering it a nonfunctional pseudogene product.
3. Figures 6 and 7 show that the average pLDDT scores for these MAT proteins range only between 67.31 and 69.25, falling into the low-to-medium confidence interval (pLDDT < 70). Consequently, AlphaFold predictions regarding locally flexible regions or severely truncated variants may harbor structural artifacts.
4. On one hand, the manuscript notes that prior studies identified alternative splicing in the mating-type genes; on the other hand, it insists that the strains are "mycologically impure" based solely on the detection of truncations. These two possibilities—alternative splicing of a single gene within the same sample versus multi-genomic expression from a mixed culture—have not been clearly decoupled.
5. All variant truncated proteins were amplified using the exact same primer pairs (m1F3/m1R3 and Mat1-2F/Mat1-2R). In complex wild multi-fungal samples, non-specific amplification, PCR bias, or sequencing assembly errors can easily cause incomplete pseudogenes to be misidentified as authentically expressed transcripts.
6. The text-based alignment format in Figure 1 is highly cluttered, with vertical alignment lines (|) and numeric layouts confusingly intermixed. It lacks the standard color-coding and canonical visualization typically provided by modern bioinformatics alignment software (such as BoxShade or Jalview), making it extremely difficult for readers to intuitively identify specific mutation sites and gaps.
7. In Table 2 and Section 2.2, a "Lysine-to-unidentified residue (K-to-X)" mutation is mentioned for AGW27553. Under modern sequencing technologies, the appearance of an unidentified amino acid (X) typically implies poor sequencing quality or the presence of ambiguous bases (Ns).
8. The ITS sequence, serving as one of the pivotal lines of evidence for the "co-existence of multiple fungi," is completely missing (indicated as "-") in GenBank for 5 out of the 20 strains. Furthermore, several strains (such as many in Group II) lack readily available AlphaFold 3D models. Consequently, the pairing chain of evidence remains incomplete both statistically and phylogenetically.
9. Although the manuscript points out that structural diversity supports self-sterility and heterothallism, it fails to elaborate on the evolutionary advantages that this complex "self-sterility and multi-fungal allelic mating" system confers to Ophiocordyceps sinensis in maintaining population genetic diversity and host adaptation within the harsh, alpine extreme environments of the Qinghai-Tibet Plateau.

Author Response

1. Reviewer #2 commended: “ The entire manuscript relies exclusively on AlphaFold predictions and in silico simulations, lacking any empirical biochemical or biophysical validation (such as X-ray crystallography, NMR, or Cryo-EM) to confirm the actual spatial conformations of these truncated and mutant proteins.”

Authors’ response:

We thank Reviewer #2 for this important comment. We fully agree that the present study is entirely based on publicly available sequence datasets, AlphaFold-predicted structural models, and comparative in silico analyses. Accordingly, the findings presented herein should be interpreted as comparative structural observations rather than direct experimental proof of protein conformation or biological function.

The objective of the present study was not to determine the experimentally verified 3D structures of the MAT1-1-1 and MAT1-2-1 proteins. Instead, we systematically characterized and compared the naturally occurring mating protein variants across the 20 O. sinensis strains with accessible data deposited in the GenBank and AlphaFold databases. We further evaluated whether their structural diversity and their native pairings align with previous genomic and transcriptomic observations (including differential occurrence, alternative splicing, and differential transcription of the mating-type genes) regarding the previously proposed self-sterility hypothesis for heterothallism or hybrid reproductive biology of O. sinensis and the microecosystem of the C. sinensis insect-fungal complex.

Our research specifically addresses four core research questions:

1. whether the pairing patterns of the MAT1-1-1 and MAT1-2-1 proteins across the 20 sinensis strains match or diverge from the presence/absence patterns of their corresponding MAT1-1-1 and MAT1-2-1 genes and transcripts identified in the genome and transcriptome assemblies of H. sinensis strains;
2. To determine whether the coexpressed MAT1-1-1 and MAT1-2-1 proteins from each of the 20 strains exhibit structural polymorphisms, including N- and/or C-terminal truncations and amino acid substitutions, relative to the authentic reference proteins, as also presented by pairwise structure superposition comparisons in the revision;
3. Whether structurally divergent, differentially naturally paired MAT1-1-1 and MAT1-2-1 protein variants within individual sinensis strains are compatible with contributions from multiple co-occurring fungal taxa within the impure O. sinensis strains; and
4. whether our integrated structural data, when integrated with prior genomic and transcriptomic evidence, are consistent with the previously established self-sterility hypothesis for sinensis, while stressing that our computational results alone are not sufficient to validate this hypothesis.

We agree that comprehensive experimental structural and functional validation is indispensable for determining the native conformations and functional activities of these polymorphic mating proteins. Such experimental approaches include X-ray crystallography, cryo-EM, NMR spectroscopy, protein biochemistry, protein‒DNA and protein‒protein interaction assays, and in vivo reproductive physiology phenotyping. We have therefore revised the manuscript to explicitly acknowledge these limitations and to emphasize that our conclusions are based on comparative structural analyses and should be regarded as hypothesis supporting rather than conclusive experimental proof.

 

2. Reviewer #2 commended: “ Although the authors speculate that the truncation of the core α-helix impairs its binding to AT-rich promoter elements, empirical experimental evidence is lacking. A truncation of 60 amino acids could directly trigger domain unfolding or lead to a complete loss of function, essentially rendering it a nonfunctional pseudogene product.”

Authors’ response:

We thank Reviewer #2 for this insightful comment. We fully acknowledge that our work depends entirely on AlphaFold-predicted structures and their superposition comparisons and comparative in silico analyses rather than direct structural resolution via X-ray crystallography, NMR spectroscopy, cryo-EM, or complementary biochemical and biophysical approaches.

The objective of this study was not to experimentally determine the 3D structures of mating proteins but instead to systematically analyze all currently available MAT1-1-1 and MAT1-2-1 sequence data from 20 purportedly pure O. sinensis strains and their corresponding AlphaFold-predicted structural models deposited in the public GenBank and AlphaFold databases. Pairwise structure superposition algorithms (TM-align and SW 3D) were applied to compare the 3D architectures of the reference authentic proteins and naturally occurring variants. Using these publicly available datasets, we characterized the structural diversity of naturally paired MAT1-1-1 and MAT1-2-1 proteins across the 20 documented O. sinensis strains and interpreted these structural observations alongside previously reported genomic and transcriptomic evidence (including differential occurrence, alternative splicing, and differential transcription of mating-type loci).

Accordingly, the conclusions of the revised manuscript have been phrased more cautiously throughout. We avoid definitive deductions that the observed structural variations necessarily lead to altered biological functions. Instead, we conclude that only the identified amino acid substitutions and N- and/or C-terminal truncations correlate with changes in secondary and tertiary structures. Multiple lines of evidence suggest that these variant mating proteins may not necessarily be derived from H. sinensis (GC-biased Genotype #1 of O. sinensis); alternatively, they may be simultaneously coproduced by co-occurring heterogeneous fungal taxa, including AT-biased, genomically independent O. sinensis genotypes. These heterogeneous origins of variant mating proteins are consistent with the previously proposed self-sterility hypothesis for H. sinensis on the basis of genomic and transcriptomic observations. This hypothesis requires compatible mating partners to accomplish heterothallic or hybrid reproduction. Importantly, the functional outcomes of these structural variations remain to be experimentally established.

We have explicitly highlighted this requirement for experimental validation in the revised manuscript. As now noted in the Discussion (Lines 715–719):

Experimental validation through protein biochemistry and reproductive physiology studies remains necessary to evaluate this hypothesis, clarify species-specific reproductive mechanisms, and determine how structural variations of the mating proteins influence complementary interactions and mating function in O. sinensis [65–69].

The present in silico study offers a comparative structural framework built upon currently available public datasets. We anticipate that these findings will facilitate future experimental investigations using full-length mating proteins and appropriate fungal materials.

 

3. Reviewer #2 commended: “ Figures 6 and 7 show that the average pLDDT scores for these MAT proteins range only between 67.31 and 69.25, falling into the low-to-medium confidence interval (pLDDT < 70). Consequently, AlphaFold predictions regarding locally flexible regions or severely truncated variants may harbor structural artifacts.”

Authors’ response:

We thank Reviewer #2 for this crucial observation and fully acknowledge this inherent limitation of AlphaFold-based structural prediction.

We agree that regions with relatively low pLDDT scores should be interpreted cautiously, particularly for intrinsically locally flexible regions, ligand-binding pockets, and heavily truncated protein variants, where predicted conformations may not fully reflect the native structures.

In the revised manuscript, we have therefore moderated our interpretation throughout the text to stress that the AlphaFold models provide computational structural predictions rather than experimentally validated structures. Accordingly, the reported structural disparities should therefore be treated as hypotheses pending future biochemical and structural–functional validation.

Nevertheless, the objective of the present study was not to determine the absolute native conformations of individual MAT proteins. Instead, we aimed to perform systematic comparative analyses of all publicly reported MAT1-1-1 and MAT1-2-1 protein variants across the 20 strains using one unified structural prediction algorithm. All the predicted 3D structures of the proteins were generated and subsequently superimposed using identical structural alignment workflows. Under such uniform computational settings, the analyses permit consistent rational comparisons of the relative structural and folding differences among naturally paired mating protein variants.

In response to suggestions from the Academic Editor and reviewers, we have improved the structural analyses presented in the figures. We implemented pairwise structural superposition using TM-align to evaluate overall folding similarity, alongside Smith–Waterman 3D to capture local similarity and dissimilarity within the hydrophobic α-helix cores within the DNA-binding domains. These structural overlays facilitate direct visual comparison between the reference and variant proteins, while we explicitly note that all underlying models remain in silico predictions.

We further reinforce in the Discussion and Conclusions that a definitive biological interpretation of these structural differences requires experimental validation. Required experiments include high-resolution structure determination of full-length proteins, biochemical characterization, protein–DNA interaction assays, and reproductive physiology phenotyping.

 

4. Reviewer #2 commended: “ On one hand, the manuscript notes that prior studies identified alternative splicing in the mating-type genes; on the other hand, it insists that the strains are "mycologically impure" based solely on the detection of truncations. These two possibilities—alternative splicing of a single gene within the same sample versus multi-genomic expression from a mixed culture—have not been clearly decoupled.”

Authors’ response:

We thank Reviewer #2 for raising this important point. We agree that these 2 biologically distinct scenarios need to be clearly distinguished: (i) alternative splicing of mating-type genes within a single fungal genome and (ii) simultaneous expression of mating proteins derived from multiple co-occurring fungal entities in a mixed sample.

Our inference of mycological impurities does not rely solely on the detection of truncated mating protein variants. Instead, this interpretation arises from integrated analysis of multiple independent observations, which reveal distinct protein expression profiles and unique natural pairing patterns of MAT1-1-1 and MAT1-2-1 proteins across the 20 O. sinensis strains.

Alternative splicing of the MAT1-2-1 gene has indeed been documented previously [14]. Unlike very large human proteins, fungal MAT1-1-1 and MAT1-2-1 proteins are encoded by 2 mating-type genes, each of which contain 3 exons and 2 introns (cf. Figure S2 of [14] and Figures S1−S2 of [68]). As illustrated in Figure S2 of [14], Intron II of the MAT1-2-1 gene is spliced, while Intron I remains retained. This retained Intron I carries stop codons and causes a reading frame shift and premature translation termination (cf. Figure S2 of [68]). As a result, Exons II and III, which encode the HMG-box_ROX1-like domain of MAT1-2-1, are not translated. Consequently, this differential intron splicing pattern is expected to translate only a short N-terminal peptide fragment encoded by Exon I, which would likely be subject to normal intracellular quality control machinery.

Crucially, this reported event does not produce a stable, alternative truncated MAT1-2-1 isoform, given that no repetitive copies of the MAT1-1-1 or MAT1-2-1 genes exist in the H. sinensis genome [69]. Accordingly, the reported alternative splicing event cannot explain the C-terminally truncated MAT1-2-1 protein isoforms in the present study (Table 2; Figures 2−3 and S2), which retain full or partial HMG-box_ROX1-like domains. To date, Ref. [14] provides the only fully documented example of mating-type gene alternative splicing in this system. It remains unclear whether analogous intracellular splicing processes contribute to the frequent absence of detectable MAT1-1-1 or MAT1-2-1 proteins observed in three-quarters of the >180 biological samples summarized in Table 3.

Our interpretation of mixed, mycologically impure samples is built upon the integration of multiple complementary lines of evidence, rather than merely the observation of truncated proteins:

• Genomic assemblies from 6 pure sinensis strains reveal differential occurrence of MAT1-1-1 and MAT1-2-1 genes, four of which lack one of the 2 mating-type genes [64,80–84].
• Differential occurrences of the MAT1-1-1 and MAT1-2-1 genes have been reported across 237 wild-type sinensis isolates and H. sinensis strains (cf. Table S1 of [68]).
• Transcription surveys have revealed alternative splicing and differential transcription of the MAT1-1-1 and MAT1-2-1

Collectively, Evidence #1–3 account for cases in which only one mating protein is detected (unpaired profiles), providing key observational support for the self-sterility hypothesis of O. sinensis [65–68]. These molecular outcomes are fundamentally incompatible with our own dataset (Evidence 4):

• All 20 sinensis strains examined in this work consistently yield both MAT1-1-1 and MAT1-2-1 proteins, which form natural pairs without loss of either paralog (Figures 3; Tables 1−2).
• Five of the 20 sinensis strains harbor at least 2 genomically distinct O. sinensis genotypes [36]: GC-biased Genotype #1 and AT-biased Genotype #17 (Tables 1 and S4−S5). These strains are unequivocally mycologically impure. An additional 5 strains lack deposited ITS sequences in GenBank, leaving their heterogeneity/homogeneity unresolved. Even those remaining strains previously considered homogeneous (which carry only GC-biased Genotype #1 ITS) may harbor additional fungal taxa, as discussed in Section 3.3 (Lines 779−855). Notably, the culture-dependent approach adopted in the study [36] likely underestimates the full capture of co-occurring fungi (Sections 2.3 and 3.2).
• The naturally paired mating proteins exhibit differential combinations of N- and/or C-terminal truncations, divergent amino acid substitutions, and distinct tertiary structural morphologies, resulting in the formation of reproducible pairing combinations across all 20 samples. We propose that these structurally variable protein variants originate from multiple coinfecting fungal entities within mycologically impure sinensis strains, maintained by natural selection favoring heterothallic or hybrid reproductive strategies.

As highlighted above, the two mechanisms raised by Reviewer #2 represent distinct biological processes and must be evaluated separately. The documented alternative splicing of the MAT1-2-1 gene arises from Intron I retention, abolishes synthesis of the HMG-box domain, and generates no stable functional protein isoform matching those analyzed protein datasets. This splicing event tends to eliminate one mating protein entirely, producing unpaired expression patterns consistent with the self-sterility hypothesis.

In stark contrast, all 20 strains analyzed in our study displayed complete but differential pairing of structurally diverse MAT1-1-1 and MAT1-2-1 proteins (Evidence #5−6). The natural pairing pattern of mating proteins cannot be readily explained by single-genome alternative splicing, differential occurrence and transcription of mating-type genes but is parsimoniously reconciled by contributions from multiple coexisting fungal genotypes or heterospecific fungal taxa within impure samples. These findings suggest that an alternative pathway under natural selection pressure is also compatible with the self-sterility hypothesis involving heterothallic or hybrid reproduction.

In summary, alternative splicing predicts the absence of one mating protein; our mating protein dataset consistently shows complete pairing of both mating proteins.

 

5. Reviewer #2 commended: “ All variant truncated proteins were amplified using the exact same primer pairs (m1F3/m1R3 and Mat1-2F/Mat1-2R). In complex wild multi-fungal samples, non-specific amplification, PCR bias, or sequencing assembly errors can easily cause incomplete pseudogenes to be misidentified as authentically expressed transcripts.”

Authors’ response:

We thank Reviewer #2 for this technical comment. We recognize that, when working with complex wild samples harboring multiple fungal taxa, technical factors, including nonspecific amplification, PCR bias, incomplete transcript recovery, or sequencing errors, should always be considered when interpreting transcript-derived protein sequences.

We wish to clarify that the present study, however, is a secondary in silico comparative analysis based exclusively on publicly available nucleotide and protein sequences deposited in GenBank, as well as their corresponding AlphaFold structural models. No MAT1-1-1 or MAT1-2-1 nucleotide or protein sequences analyzed here were generated experimentally in our laboratory. Instead, we performed a systematic in silico analysis of the completely publicly available datasets originally produced and submitted to GenBank by Bushley et al. (2013, Fungal Biology) and Li et al. (2013, Molecular Phylogenetics and Evolution) [14,36].

As documented by Bushley et al. (2013) [14], all MAT1-1-1 cDNAs were amplified using the same primer pair (m1F3/m1R3), and all MAT1-2-1 cDNAs were amplified using the uniform primer pair (Mat1-2F/Mat1-2R) under identical experimental conditions described in the original publication. Accordingly, any systematic technical biases introduced during the original PCR amplification experiments under the unified technical setting would exert comparable effects across all samples within the margin of systematic error. Such a consistent experimental framework enables valid comparative analysis of the reported protein variants contained within that dataset, although the uniformity of methodology alone cannot confirm that each recovered transcript corresponds to the native biological transcript.

In light of the above, we interpret the structural differences through comparative analysis rather than treating each retrieved sequence as definitive evidence of the native protein architecture. Independent experimental validation, encompassing the re-sequencing of full-length MAT1-1-1 and MAT1-2-1 transcripts, and proteomic identification of corresponding proteins derived from well-characterized pure H. sinensis strains and wild-type C. sinensis isolates will be necessary to formally verify the biological authenticity of these sequence variants.

 

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

I appreciate the work made on the revision of the manuscript. I agree with the current interpretation, that the observed results are consistent with the possibility of the self-sterility, rather than prove it to be a single mechanism. I have no serious critique now, so in the current form the manuscript may be accepted for publication. My only minor comment is to the revised versions of structural figures (Figure 4, G and H, and Figure 5, G and H). In the structural biology papers, it is a common practice, when superposing the complex structures, for instance, those of multidomain proteins, to superpose them using only one domain/part of a chain. The authors obviously in both cases used full-length superpositions. Considering that the proteins have complex structures (involve several structural domains as well as elements with unresolved structure), it would be better to superpose those α-helical domains that are discussed in the text and are highlighted with green outline in the reference proteins. This would give the idea of how these particular domains are compared to each other, e.g. in respect of helix orientations, while in the current variant we can compare only the position of this domain in the overall structure which is not fixed anyway (since the domains are linked with the flexible parts of a chain).

 

Author Response

Comments from Reviewer #1 for Authors

I appreciate the work made on the revision of the manuscript. I agree with the current interpretation, that the observed results are consistent with the possibility of the self-sterility, rather than prove it to be a single mechanism. I have no serious critique now, so in the current form the manuscript may be accepted for publication. My only minor comment is to the revised versions of structural figures (Figure 4, G and H, and Figure 5, G and H). In the structural biology papers, it is a common practice, when superposing the complex structures, for instance, those of multidomain proteins, to superpose them using only one domain/part of a chain. The authors obviously in both cases used full-length superpositions. Considering that the proteins have complex structures (involve several structural domains as well as elements with unresolved structure), it would be better to superpose those α-helical domains that are discussed in the text and are highlighted with green outline in the reference proteins. This would give the idea of how these particular domains are compared to each other, e.g. in respect of helix orientations, while in the current variant we can compare only the position of this domain in the overall structure which is not fixed anyway (since the domains are linked with the flexible parts of a chain).

 

Authors response to the comment:

We thank Reviewer #1 for this helpful clarification. We fully agree that, for multidomain proteins containing flexible linkers or structurally unresolved regions, domain‑restricted superposition yields high‑resolution comparisons of local structural traits, which complements full‑length alignment.

We would like to clarify that our revised figures already incorporate both types of structural superposition recommended by the Reviewer. In Figures 4 and 5 (as well as Figures S3–S10), Panels (G) display the full‑length protein superposition using the TM-align algorithm. The MATα_HMGbox or HMG-box_ROX1-like DNA-binding domain highlighted by pale‑green outlines. This comparison illustrates the overall structural similarity between the reference and variant proteins.

Importantly, Panels (H) provide a separate, domain-focused structural superposition. In Panels (H), we exclusively superposed the DNA-binding domains, focusing specifically on the region containing the 3 α-helices forming the hydrophobic core region within the DNA-binding domain. Local‑domain superpositions were carried out with the Smith–Waterman 3D (SW 3D) algorithm to optimize the local structural alignment and thereby facilitate direct comparison of helix orientations and local conformational differences.

To summarize, Panels (G) address the overall fold of the full-length proteins, whereas Panels (H) specifically address the structural dissimilarity of the DNA-binding domains highlighted in the text. Both alignment modes are preserved because they deliver complementary information: the full-length superposition illustrates the overall structural framework, while the domain-focused superposition allows direct assessment of the local conformational variation within the functionally essential α-helical DNA-binding region.

We appreciate the Reviewer’s valuable advice. We have recognized the necessity to explicitly distinction the two alignment strategies (in Panels (G) and (H)) within main‑text descriptions and figure captions.

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have satisfactorily responded to all of my comments and revised the manuscript accordingly. I recommend that the manuscript be accepted for publication.

Author Response

We greatly appreciate your kind recommendation to the journal editors for the acceptance of our manuscript.