Skip to PREreview

PREreview of Universally conserved isoasparylation sustains ribosome biogenesis and function

Published
DOI
10.5281/zenodo.23602488
License
CC BY 4.0

Overall assessment:

This study presents strong evidence that two proteins, YbeY (bacteria and organelles) and Fap7 (archaea, eukaryotes), are involved in the installation of isoaspartate in ribosomal protein uS11. Through bioinformatic analysis, biochemical assays, structural and mutational studies, this work demonstrates the generation and role of the universally-conserved isoaspartate residue in proper assembly and function of the ribosome. The main findings of the study are 1. The role of YbeY and Fap7 in modification of Asn or Asp residues to isoaspartate in the ribosomal protein uS11 across the tree of life. 2. The catalytic mechanisms of isoaspartylation that are novel and highlights mechanistic repurposing of protein folds. While YbeY shares structural and functional similarity with the metzincin metalloprotease in substrate engagement and α-peptide bond hydrolysis, the former keeps the C-terminal tail intact through the β-peptide bond of isoAsp. This mechanistic repurposing of a protein-fold is also accentuated by its eukaryotic counterpart Fap7 where an adenylate kinase fold, typically used for nucleotide phosphorylation, has been repurposed for phosphate transfer from ATP to the target Asp followed by backbone cyclization to form the aspartimide intermediate and subsequent hydrolysis to generate the isoAsp residue. 3. The importance of YbeY or Fap7-mediated modification in correct incorporation of the small ribosomal subunit. 4. The portability of the two isoaspartylation pathways between bacteria and archaea. 5. Providing molecular explanations for previous genetic data showing the importance of YbeY/Fap7 and their mutants.

The study is an important advancement in our understanding of this universal isoaspartylation in proper ribosome functioning. Earlier studies of universally conserved ribosome modifications focused on rRNA, and isoAsp was mainly known in other proteins as an accumulated ageing modification. Therefore, this work addresses a gap in the literature related to translation-related protein modifications.

Based on the experiments conducted to support the claims, I have the following concerns and suggestions which may improve the clarity of this study.

Major comments:

1. In Fig. 1B, the authors have conducted an extensive bioinformatic analysis of uS11 proteins of bacteria and organelle uS11 where species encoding for the YbeY protein predominantly contain an Asn residue at the C-terminal of uS11 which is ‘mostly’ absent in species that do not encode YbeY. However, Fig. 1B also shows a small proportion of an Asn residue at the modification site present in some species that also lack the ybeY gene. A short analysis of this subset would be helpful. i) Do the species that still contain the Asn in uS11 but do not code of YbeY also lack the modification? If so, it may be an evolutionary remnant. ii) If the residue is modified, is there a homologous protein that may carry out a similar function? Because the authors suggest that the isoaspartylation is universally conserved, this explanation would provide further clarity to this statement.

2. The authors show in their structural studies an extensive interaction between uS11 and the modifying enzyme. Yet, the successful transformation of the short peptide fragment (derived from the C-terminal tail of human uS11, Fig 3 and S7) by Fap7 suggests that such extensive interactions are not essential. This suggests that the modification may happen either on the C-terminal tail of the folded protein or on the nascent polypeptide chain. The authors may wish to discuss this.

3. In context of the information shown in Fig. S6, the Zn-free aspartimide could have formed during co-expression, sample preparation or crystallization. As I understand, hydroxylamine can weakly chelate divalent metal ions. At alkaline pH, hydroxylamine can cleave Asn-Gly bonds (as in this case; Fig S6F) which results in succinimide formation (reference 35). This product is the same C-terminal hydroxyamide that the authors detect in Fig S6F. While the zinc loss does not change the interpretation of the results by itself, the current presentation suggests the Zn is not involved in aspartimide generation and this conclusion may not be correct. The Zn-containing sample has already converted Asn119 to isoAsp. The comparison therefore cannot distinguish between a pre-existing aspartimide from one formed during treatment. Perhaps the authors would like to clarify this uncertainty.

4. How do the authors envision Fap7 catalyzing the regioselective hydrolysis of the aspartimide? For YbeY, the Zn provides a clear avenue for regioselective hydrolysis (Fig. S6I), but in Fap7, this is less clear. PIMTs usually result in non-selective hydrolysis as also cited by the authors in reference 33. The authors may wish to discuss this, for example in light of the Lys106, Tyr102 and Pro9 interactions with the product (Fig. 3E). It appears that the Lys106 and Tyr102 anchor the α-carboxyl of isoAsp124 while the Pro9 stabilizes the methylene.

Minor comments

5. ‘YbeY and Fap7 evolutionarily co-occur with uS11 isoaspartylation signatures’ section: The main text states 5551 genomes were analysed (P3 line18), but Fig. 1B panel shows (5191 + 571) genomes in the plot. Perhaps the authors can address this discrepancy in the figure legend and the methods section. Were the organelle genomes counted separately from the bacterial genomes?

6. The authors have demonstrated that the C-terminal tail of uS11 interacts with the active site of YbeY and Fap7 and that the C-terminal peptide fragment alone can act as a substrate. However, the main text and legend of Fig. 3F are missing the sequence details of corresponding peptide fragments used for in vitro testing. The methods section (Table S9) mentions which peptide substrate was used but perhaps naming them and mentioning the sequence in the main text or the figure legends will provide better readability. The same is true for Fig S3B where capillary electrophoresis was used to distinguish Asp and isoAsp containing C-terminal peptide but the peptide used is not mentioned.

7. Fig S7B legend states that reaction with human uS11 peptide was performed for 1 hr at 37 C but the methods state the reaction was conducted for 50 min. The authors may want to address this minor discrepancy.

8. P 3 line 29: the authors may wish to clarify that the sequence dependence is for non-enzymatic isoAsp formation. Bacterial PIMT enzymes generate isoAsp/aspartimide in a variety of sequence contexts (e.g. Refs 50, 51 and Zhu and Link, 2026, Biochemistry, 65(6), 661–669, DOI: 10.1021/acs.biochem.6c00055. PMID: 41790433)

9. P 3 Line 42: The statement that PIMT ‘specifically recognizes and methylates isoAsp residues’ may be too strong, since recent studies have shown that PIMT can also methylate C-terminal Asp/Glu (a review: Zhu and Link, 2026, Biochemistry, 65(6), 661–669, DOI: 10.1021/acs.biochem.6c00055). This is also mentioned in the discussion (ref 52, 53). Also, the methods cite ref. 26 for the purification of PIMT, but ref. 26 does not contain the information.

10. The authors mention that Arg44 π-π stacks with Phe27 in two sections (P 4, line 27 and SI, Fig. S4 legend). It seems that cation-π interaction is the more appropriate interaction happening here since Arg lacks an aromatic π-system.

I reviewed this preprint as part of the Transparent and Accountable Peer Review training program. This report was written with the support of my research mentor.

Competing interests

The authors declare that they have no competing interests.

Use of Artificial Intelligence (AI)

The authors declare that they did not use generative AI to come up with new ideas for their review.

Comments

Write a comment
  1. Comment by Alexandre Smirnov

    Published
    License
    CC BY 4.0

    Overall assessment:

    This study presents strong evidence that two proteins, YbeY (bacteria and organelles) and Fap7 (archaea, eukaryotes), are involved in the installation of isoaspartate in ribosomal protein uS11. Through bioinformatic analysis, biochemical assays, structural and mutational studies, this work demonstrates the generation and role of the universally-conserved isoaspartate residue in proper assembly and function of the ribosome. The main findings of the study are 1. The role of YbeY and Fap7 in modification of Asn or Asp residues to isoaspartate in the ribosomal protein uS11 across the tree of life. 2. The catalytic mechanisms of isoaspartylation that are novel and highlights mechanistic repurposing of protein folds. While YbeY shares structural and functional similarity with the metzincin metalloprotease in substrate engagement and α-peptide bond hydrolysis, the former keeps the C-terminal tail intact through the β-peptide bond of isoAsp. This mechanistic repurposing of a protein-fold is also accentuated by its eukaryotic counterpart Fap7 where an adenylate kinase fold, typically used for nucleotide phosphorylation, has been repurposed for phosphate transfer from ATP to the target Asp followed by backbone cyclization to form the aspartimide intermediate and subsequent hydrolysis to generate the isoAsp residue. 3. The importance of YbeY or Fap7-mediated modification in correct incorporation of the small ribosomal subunit. 4. The portability of the two isoaspartylation pathways between bacteria and archaea. 5. Providing molecular explanations for previous genetic data showing the importance of YbeY/Fap7 and their mutants.

    The study is an important advancement in our understanding of this universal isoaspartylation in proper ribosome functioning. Earlier studies of universally conserved ribosome modifications focused on rRNA, and isoAsp was mainly known in other proteins as an accumulated ageing modification. Therefore, this work addresses a gap in the literature related to translation-related protein modifications.

    Based on the experiments conducted to support the claims, I have the following concerns and suggestions which may improve the clarity of this study.

    Response: We thank the reviewer for such a thorough and detailed analysis of our manuscript and their enthusiasm regarding our work. The reviewer’s remarks and suggestions helped us to clarify several important notions and experimental details and remove inconsistencies.

    Major comments:

    1. In Fig. 1B, the authors have conducted an extensive bioinformatic analysis of uS11 proteins of bacteria and organelle uS11 where species encoding for the YbeY protein predominantly contain an Asn residue at the C-terminal of uS11 which is ‘mostly’ absent in species that do not encode YbeY. However, Fig. 1B also shows a small proportion of an Asn residue at the modification site present in some species that also lack the ybeY gene. A short analysis of this subset would be helpful. i) Do the species that still contain the Asn in uS11 but do not code of YbeY also lack the modification? If so, it may be an evolutionary remnant. ii) If the residue is modified, is there a homologous protein that may carry out a similar function? Because the authors suggest that the isoaspartylation is universally conserved, this explanation would provide further clarity to this statement.

    Response: There are indeed cases of an asparagine in this position even in the absence of YbeY. By far the most prominent group of such occurrences (~170 species) are mitochondrial uS11m proteins of insects. In these proteins, this asparagine is always followed by a proline, the only downstream residue that is incompatible with isoaspartylation, since the aspartimide cycle cannot be closed on an imino acid. Thus, the isoaspartylation of such proteins is chemically impossible. We will mention this interesting exception in a revised version of the manuscript.

    A rare, but curious type of uS11 proteins with an isoaspartylatable asparagine (followed by a glycine) has been observed in two bacterial species lacking a ybeY gene: Candidatus Carsonella ruddii and Candidatus Atelocyanobacterium thalassa. Both are obligate endosymbionts of eukaryotic hosts with reduced genomes (182 and 1227 genes, respectively). We hypothesize that the ‘missing’ YbeY protein may be encoded in the host’s genome and imported into the symbionts, along with a host of other proteins, much like this happens with mitochondria and plastids.

    Finally, the scenario of the ‘evolutionary remnant’ proposed by the reviewer appears to be realized in the Mycoplasma genus. Most Mycoplasma species have YbeY and an isoaspartylatable asparagine (followed by a glycine) in their uS11 proteins. In M. ovis and M. wenyonii, YbeY has been lost, and, accordingly, their uS11 proteins contain a glycine instead of the asparagine. Beside these ‘canonical’ cases, there are M. parvum and M. suis, which have no YbeY, but still preserve an asparagine in uS11, suggesting indeed that they are at an intermediate evolutionary stage on the way of losing the isoaspartylation pathway.

    2. The authors show in their structural studies an extensive interaction between uS11 and the modifying enzyme. Yet, the successful transformation of the short peptide fragment (derived from the C-terminal tail of human uS11, Fig 3 and S7) by Fap7 suggests that such extensive interactions are not essential. This suggests that the modification may happen either on the C-terminal tail of the folded protein or on the nascent polypeptide chain. The authors may wish to discuss this.

    Response: Considering that the C-terminal tail emerges from the ribosome after the N-terminal domain, the isoaspartylation can only happen once the entire protein has been translated and released.

    3. In context of the information shown in Fig. S6, the Zn-free aspartimide could have formed during co-expression, sample preparation or crystallization. As I understand, hydroxylamine can weakly chelate divalent metal ions. At alkaline pH, hydroxylamine can cleave Asn-Gly bonds (as in this case; Fig S6F) which results in succinimide formation (reference 35). This product is the same C-terminal hydroxyamide that the authors detect in Fig S6F. While the zinc loss does not change the interpretation of the results by itself, the current presentation suggests the Zn is not involved in aspartimide generation and this conclusion may not be correct. The Zn-containing sample has already converted Asn119 to isoAsp. The comparison therefore cannot distinguish between a pre-existing aspartimide from one formed during treatment. Perhaps the authors would like to clarify this uncertainty.

    Response: We thank the reviewer for bringing up these points. Indeed, it is not clear at this stage whether the catalytic zinc participates directly in the first part of the reaction (cyclisation). Therefore, we avoided all speculation about this in the manuscript, limiting our conclusions to the second part of the reaction (aspartimide hydrolysis), which we consider reliably established. This said, given the steric constraints of the active center, it is conceivable, if not compelling, that the zinc may be involved in substrate anchoring from the very early steps of the reaction.

    Regarding distinguishing between aspartimides formed prior and during hydroxylamine treatment, several considerations argue against the latter possibility. The cleavage of Asn-Gly bonds with hydroxylamine in the absence of a strong denaturant (which is the case here) is a very inefficient reaction (ref. 35) and could hardly account for the massive uS11 truncation observed in fig. S6E. Most importantly, the protein samples treated in fig. S6E,F are the same as used for crystallization, and fig. S6C shows unambiguously an aspartimide in this uS11-YbeY complex. This indicates that the aspartimide was present there before the hydroxylamine treatment. Since this important detail is somewhat buried in Materials and methods, we will add this information directly to the main text to avoid any further confusion.

    4. How do the authors envision Fap7 catalyzing the regioselective hydrolysis of the aspartimide? For YbeY, the Zn provides a clear avenue for regioselective hydrolysis (Fig. S6I), but in Fap7, this is less clear. PIMTs usually result in non-selective hydrolysis as also cited by the authors in reference 33. The authors may wish to discuss this, for example in light of the Lys106, Tyr102 and Pro9 interactions with the product (Fig. 3E). It appears that the Lys106 and Tyr102 anchor the α-carboxyl of isoAsp124 while the Pro9 stabilizes the methylene.

    Response: This is an excellent point—and a major direction of our ongoing work. At this stage, it is clear that no ‘canonical’ proteinase-like mechanism, as it is the case with YbeY, could account for the second step of the Fap7-catalysed isoaspartylation. Given the need for two proton transfers during such a hydrolysis reaction, residues like His73 and Lys106 appear to be strong candidates for the catalytic acid-base pair positioned right next to the substrate aspartimide.

    Minor comments

    5. ‘YbeY and Fap7 evolutionarily co-occur with uS11 isoaspartylation signatures’ section: The main text states 5551 genomes were analysed (P3 line18), but Fig. 1B panel shows (5191 + 571) genomes in the plot. Perhaps the authors can address this discrepancy in the figure legend and the methods section. Were the organelle genomes counted separately from the bacterial genomes?

    Response: For eukaryotes possessing both mitochondria and plastids, uS11m and uS11c proteins were counted separately. Therefore, the number of species (= complete genomes) is smaller than that of uS11 proteins analyzed. We highlighted this in Materials and methods. The confusion may stem from Fig. 1B,C, where n actually refers to the number of distinct proteins, not species. This will be corrected.

    6. The authors have demonstrated that the C-terminal tail of uS11 interacts with the active site of YbeY and Fap7 and that the C-terminal peptide fragment alone can act as a substrate. However, the main text and legend of Fig. 3F are missing the sequence details of corresponding peptide fragments used for in vitro testing. The methods section (Table S9) mentions which peptide substrate was used but perhaps naming them and mentioning the sequence in the main text or the figure legends will provide better readability. The same is true for Fig S3B where capillary electrophoresis was used to distinguish Asp and isoAsp containing C-terminal peptide but the peptide used is not mentioned.

    Response: Thank you for spotting this. The peptide details will be added to the corresponding Figure legends.

    7. Fig S7B legend states that reaction with human uS11 peptide was performed for 1 hr at 37 C but the methods state the reaction was conducted for 50 min. The authors may want to address this minor discrepancy.

    Response: Thank you for spotting this discrepancy. It is 1 h indeed.

    8. P 3 line 29: the authors may wish to clarify that the sequence dependence is for non-enzymatic isoAsp formation. Bacterial PIMT enzymes generate isoAsp/aspartimide in a variety of sequence contexts (e.g. Refs 50, 51 and Zhu and Link, 2026, Biochemistry, 65(6), 661–669, DOI: 10.1021/acs.biochem.6c00055. PMID: 41790433)

    Response: The reviewer is right. We will add this detail to the text.

    9. P 3 Line 42: The statement that PIMT ‘specifically recognizes and methylates isoAsp residues’ may be too strong, since recent studies have shown that PIMT can also methylate C-terminal Asp/Glu (a review: Zhu and Link, 2026, Biochemistry, 65(6), 661–669, DOI: 10.1021/acs.biochem.6c00055). This is also mentioned in the discussion (ref 52, 53). Also, the methods cite ref. 26 for the purification of PIMT, but ref. 26 does not contain the information.

    Response: This is correct. Although the main and best-documented function of PIMT, directly leveraged in our study, is methylating isoaspartates, it can also methylate certain C-terminal residues, such as Gln, although this ability seems to depend strongly on the sequence and structure context (ref. 53). Proteins with such C-terminal residues are relatively rare and at any rate irrelevant in the context of the present study.

    The reference 26 does not describe specifically PIMT purification; it provides the detailed general protocol of protein purification that we adapted in the present study to purify uS11, Fap7, YbeY and PIMT proteins.

    10. The authors mention that Arg44 π-π stacks with Phe27 in two sections (P 4, line 27 and SI, Fig. S4 legend). It seems that cation-π interaction is the more appropriate interaction happening here since Arg lacks an aromatic π-system.

    Response: This is a very good point. Arg possesses a very stable 6-electron π-system capable of significant stacking interactions with aromatic systems. Physicochemical analyses indicate that the interactions occurring between Arg and aromatic residues are neither purely cation-π (i.e. electrostatics-dominated), nor π–π (i.e. dispersion-dominated) (Singh et al., J Chem Theory Comput, 2009, doi.org/10.1021/ct800471b). It is something in between, with electrostatic and dispersion forces contributing almost equally to the binding energy. We will follow the nomenclature proposed in the cited study and designate this interaction as π+–π.

    I reviewed this preprint as part of the Transparent and Accountable Peer Review training program. This report was written with the support of my research mentor.

    We would like to thank again the reviewer for having chosen our paper and their helpful and constructive comments.

    Competing interests

    The author of this comment declares that they have no competing interests.