A distinctive family of L,D-transpeptidases catalyzing L-Ala-mDAP crosslinks in Alpha- and Betaproteobacteria

The bacterial cell-wall peptidoglycan is made of glycan strands crosslinked by short peptide stems. Crosslinks are catalyzed by DD-transpeptidases (4,3-crosslinks) and LD-transpeptidases (3,3-crosslinks). However, recent research on non-model species has revealed novel crosslink types, suggesting the existence of uncharacterized enzymes. Here, we identify an LD-transpeptidase, LDTGo, that generates 1,3-crosslinks in the acetic-acid bacterium Gluconobacter oxydans. LDTGo-like proteins are found in Alpha- and Betaproteobacteria lacking LD3,3-transpeptidases. In contrast with the strict specificity of typical LD- and DD-transpeptidases, LDTGo can use non-terminal amino acid moieties for crosslinking. A high-resolution crystal structure of LDTGo reveals unique features when compared to LD3,3-transpeptidases, including a proline-rich region that appears to limit substrate access, and a cavity accommodating both glycan chain and peptide stem from donor muropeptides. Finally, we show that DD-crosslink turnover is involved in supplying the necessary substrate for LD1,3-transpeptidation. This phenomenon underscores the interplay between distinct crosslinking mechanisms in maintaining cell wall integrity in G. oxydans.

The manuscript by Espaillat et al. presents an in-depth characterization of an L,D-transpeptidase involved in unconventional 1,3 peptidoglycan crosslinks.The existence of such an enzyme was foreseen in a previous study by the same authors that revealed (1-3) L-Ala-mDAP cross-links in the peptidoglycan of several Proteobacteria.A very recent study (J. Biol. Chem., in press, doi.org/10.1016/j.jbc.2023.105494)identified a similar L,D transpeptidase and described a first preliminary characterization.The present study by Espaillat et al. provides a comprehensive study of this unconventional LD transpeptidase with additional and meaningful data, relative to its substrates, its 3D structure and its role in maintaining cell wall integrity.A thorough investigation of the enzymatic activity is provided using purified recombinant enzyme and purified sacculi as substrates.They provide convincing results showing that LD1,3-TPases use muropeptides with tri or tetrapeptide chains as acyl donors, thus engaging non-terminal amino acids in the cross-linking process.This property makes these enzymes unique among characterized L,D-transpeptidases.They also obtained crystallographic structural data showing distinctive features of LD1,3transpeptidases compared to LD3,3-transpeptidases, allowing to support their original substrate specificity.They further identified a D,D-endopeptidase (of the PBP family) involved in regulating substrate availability.Furthermore, they provide first insights into the role of the 1,3 LD crosslinks to counteract cell envelope stresses.The presented experimental data are sound and the manuscript is well written.Therefore, this study should be of great interest to the readers of Nature Communications.I have a few specific comments: L181-182: "without a corresponding impact on M2".Since M2 is present in very low amount in V. cholerae sacculi (as indicated L177), it seems difficult to conclude on the role of M2 as a donor substrate with this experiment.Please modify the text.Fig. 3A: M2 is not visible on the chromatograms.Was it detected only by LC-MS?Indicate its position on the chromatograms in Fig. 2A.Fig. 3B (legend) : how the "molar abundance" of each muropeptide form was quantified?L180-186: the data presented to conclude that M4 is a substrate of LDTBcn are quite convincing.However, the clarity of this paragraph could be improved to better explain the double activity of LDTbcn (LD-TPase and LD-endopeptidase).The decrease in M4 may be associated with either TPase or EPase activity and this is the same for D44 decrease.It may be mentioned that like for classical L,D-transpeptidase, this is in agreement with the formation of an acyl-enzyme intermediate (covalent intermediate between the catalytic Cys residue and disaccharide-L-Ala (M1)) requiring cleavage of the L-Ala1-Glu2 bond)).At least in vitro, this acyl intermediate could be further hydrolyzed leading to LD-EPase activity or reacts with an acyl acceptor M4 or D44 leading to LD-TPase activity.L210-211: the authors indicate that they tested LDTgo activity on muramidase-digested peptidoglycan or purified muropeptides without success.Which experiments were performed?Was the activity tested with non-reduced muropeptides?Reduction of muropeptides with sodium borohydride could inhibit enzymatic activity.
Reviewer #2 (Remarks to the Author): Traditionally, the peptidoglycan (PG) studies were done in model bacteria such as E. coli, Vibrio cholerae or Bacillus subtilis and these yielded a wealth of information on the basic fundamental processes involved in bacterial cell wall biogenesis.PG is a polymer consisting of linear glycan strands attached to a short-stem peptides.In Gram-negative bacteria, the predominant components of the stem peptide include L-alanine (L-Ala1), D-glutamate (D-Glu2), mesodiaminopimelic acid (mDAP3), D-alanine (D-Ala4), and D-Ala5, with cross-links occurring between D-Ala4 and mDAP3 (4-3) or between two mDAP3 residues (3-3).While 4-3 cross-links are made by penicillin-binding proteins, 3-3 cross-links are the products of L,D-Transpeptidases.Recent studies on the PG of non-model culturable bacteria have uncovered novel modifications in the cell wall, providing intriguing insights into their peptidoglycan biology.In the present study, Espaillat et al. describe how the non-canonical 1-3 cross-links in Glucanobacter oxydans, a member of alpha-proteobacteria are synthesized and regulated.Through extensive bioinformatics and biochemical approaches, the authors discovered and characterized a new class of enzymes involved in the formation of these 1,3 cross-links, providing evidence of how it differs from its similar counterparts, LDTs (3-3).Additionally, they elucidated the protein structure and offered mechanistic insights into its catalytic activity.This study is truly fascinating presenting clear data and a complete picture.Manuscript is written well with good flow and logic.I thoroughly enjoyed reading the manuscript.This study provides a major breakthrough in the field of peptidoglycan biology and will be of interest to a large class of bacterial research community.I have no major concerns, but authors should consider the following: 1. Fig. 1A.Representation of T144 needs to be rechecked.Will the central mDAP make 4 bonds which are theoretically impossible?I think the D-ala of central monomer may make a crosslink with mDAP of right monomer. 2. Recommend adding a panel to Fig. 1 illustrating all the structures/ denotations of muropeptides including M1, M2 and others (which are already here).3. Fig. 3A.A representation of M2 in the chromatogram is needed for the reader to compare the chromatogram for level of M2.Also label the M2 peak in the chromatogram.4. In the results section 6 (322-324), it is suggested that three amino acids, Tyr 180, 181, and 182, play crucial role in conferring the characteristic 1-3 cross-linking properties to these enzymes.Have the authors attempted to mutate these residues and assess whether the enzyme still functions as a 1-3 cross-linker? 5.In Fig. 1, I would suggest adding a western blot showing the expression of catalytic site mutants.6.The figures need to be enlarged and font sizes increased for ease in reading.7. Discussion, lines 420, this paragraph discusses the regulation and homeostasis of different crosslinks.It is possible that 1,3-crosslink hydrolase may exist in these organisms and it may control the level of 1,3 crosslinks.Only the synthases need not be controlled; hydrolases may also contribute to the crosslinkage frequency.8. May be a sentence on how these crosslinked muropeptides are recycled would help in the Discussion.9. Line 54-comma to be included after pentapeptide.10.Line 87-'type' after novel can be removed.

Reviewer #3 (Remarks to the Author):
This is an interesting mechanistic study regarding a novel L,D transpeptidase that catalyzes 1-3 stem peptide cross-links.Authors employ extensive bacterial genetics techniques, paired with structural biology and molecular dynamics, to extensively characterize not only the novel transpeptidase itself but also the biochemical and mechanistic environment that lead to its activation.The text is elegantly written and the figures are quite clear.There are a few points that merit clarification, highlighted below.
Line 20, and elsewhere: authors should include a few sentences to describe the generalized term 'acetic acid bacteria', i.e., specific growth conditions, unusual environments, or metabolism.
Lines 116-117: what were the consequences of the 1-3 crosslinks on the overall bacterial phenotype?Did authors observe modifications in growth, shape, or survival capacity?It could be of interest to mention this here.
Lines 142-145: can authors mention a few prototypical strains in these phyla, and are there specific characteristics (such as typical growth environment, pathogenicity potential, etc) that could explain the need for such a transpeptidase in these strains?Line 179: kindly clarify the need for Cu+2 in this experiment.
Lines 212-213: can authors comment on the preference of LDT-Bcn for PG chains rather than individual muropeptides?Lines 246-247: it would be of interest to the reader to find out at this point a few more details regarding how the crystal structure was solved.This reviewer had to look for this info in the Materials & Methods section.The structure was solved by using an AlphaFold model, but no details of the experiment were provided, neither in the Results section nor in M&M itself.Can authors comment on differences between the initial model and the final structure

Response to reviewers
Reviewer #1 (Remarks to the Author): The manuscript by Espaillat et al. presents an in-depth characterization of an L,Dtranspeptidase involved in unconventional 1,3 peptidoglycan crosslinks.The existence of such an enzyme was foreseen in a previous study by the same authors that revealed (1-3) L-Ala-mDAP cross-links in the peptidoglycan of several Proteobacteria.A very recent study (J.Biol.Chem., in press, doi.org/10.1016/j.jbc.2023.105494)identified a similar L,D transpeptidase and described a first preliminary characterization.The present study by Espaillat et al. provides a comprehensive study of this unconventional LD transpeptidase with additional and meaningful data, relative to its substrates, its 3D structure and its role in maintaining cell wall integrity.A thorough investigation of the enzymatic activity is provided using purified recombinant enzyme and purified sacculi as substrates.They provide convincing results showing that LD1,3-TPases use muropeptides with tri or tetrapeptide chains as acyl donors, thus engaging nonterminal amino acids in the cross-linking process.This property makes these enzymes unique among characterized L,D-transpeptidases.They also obtained crystallographic structural data showing distinctive features of LD1,3-transpeptidases compared to LD3,3-transpeptidases, allowing to support their original substrate specificity.They further identified a D,D-endopeptidase (of the PBP family) involved in regulating substrate availability.Furthermore, they provide first insights into the role of the 1,3 LD crosslinks to counteract cell envelope stresses.The presented experimental data are sound and the manuscript is well written.Therefore, this study should be of great interest to the readers of Nature Communications.
We thank the reviewer for their detailed assessment of our manuscript and their enthusiastic comments.
I have a few specific comments: L181-182: "without a corresponding impact on M2".Since M2 is present in very low amount in V. cholerae sacculi (as indicated L177), it seems difficult to conclude on the role of M2 as a donor substrate with this experiment.Please modify the text.
We agree with the reviewer's concern, and we have removed the sentence (line 181).The M2 was detected only by LC-MS analysis in Vibrio cholerae samples.Following the reviewer's suggestion, we have indicated its position in Fig. 3A.Fig. 3B (legend) : how the "molar abundance" of each muropeptide form was quantified?
The relative molar abundance is the percentage of the peak area of a muropeptide, divided by its molecular weight, compared to the sum of peak areas in the chromatogram.We have clarified this in the figure legend (lines 1011-1013) and included it in the Methods section (lines 596-598).L180-186: the data presented to conclude that M4 is a substrate of LDTBcn are quite convincing.However, the clarity of this paragraph could be improved to better explain the double activity of LDTbcn (LD-TPase and LD-endopeptidase).The decrease in M4 may be associated with either TPase or EPase activity and this is the same for D44 decrease.It may be mentioned that like for classical L,D-transpeptidase, this is in agreement with the formation of an acyl-enzyme intermediate (covalent intermediate between the catalytic Cys residue and disaccharide-L-Ala (M1)) requiring cleavage of the L-Ala1-Glu2 bond)).At least in vitro, this acyl intermediate could be further hydrolyzed leading to LD-EPase activity or reacts with an acyl acceptor M4 or D44 leading to LD-TPase activity.
We have included a sentence in line with the reviewer's recommendation (lines 184-188).L210-211: the authors indicate that they tested LDTgo activity on muramidasedigested peptidoglycan or purified muropeptides without success.Which experiments were performed?Was the activity tested with non-reduced muropeptides?Reduction of muropeptides with sodium borohydride could inhibit enzymatic activity.
The reviewer raises a valid point.We tested both LDTGo and LDTBcn activities in vitro under different conditions and using different purified substrates such as M2, M4, M5 and their combinations.To rule out a possible effect of reduction of the muropeptides on the enzymatic activity, we also used non-reduced mutanolysin digests of different sacculi such as that of Gluconobacter oxydans, Escherichia coli, Vibrio cholerae.These in vitro assays were done using different reaction buffer conditions, including a range of pH (pH 6-8), and supplementation with different salts and metal ions (NaCl,CaCl2,ZnCl2,MgCl2).Unfortunately, all our efforts to demonstrate the activity on purified muropeptides were unsuccessful and have led us to conclude the enzyme is preferentially active on intact sacculi.We agree with the reviewer that the incorporation of Gln is independent of LDTGo; however, we prefer to keep this information in the figure as it includes the identification of non-canonical M2 muropeptides.Nonetheless, we have edited the text to emphasize M2 Gln is not a product of LDTGo (lines 236-238).L549: change to .. "N-acetyl-muramic acid residues were reduced to muramicitol".
Modified accordingly in the revised version of the manuscript (line 570).
L563: does "MS-MS/MS" mean "MS and MS/MS"?Yes, the reviewer is right, so we have changed it in the revised manuscript (line 585).

L565: define MSE
According to the reviewer's suggestion, we have defined what MS E is in the text (lines 587-588).L578: how were the sacculi prepared?
We have done the requested replacement (line 566) and throughout the whole manuscript.

Reviewer #2 (Remarks to the Author):
Traditionally, the peptidoglycan (PG) studies were done in model bacteria such as E. coli, Vibrio cholerae or Bacillus subtilis and these yielded a wealth of information on the basic fundamental processes involved in bacterial cell wall biogenesis.PG is a polymer consisting of linear glycan strands attached to a short-stem peptides.In Gram-negative bacteria, the predominant components of the stem peptide include Lalanine (L-Ala1), D-glutamate (D-Glu2), meso-diaminopimelic acid (mDAP3), Dalanine (D-Ala4), and D-Ala5, with cross-links occurring between D-Ala4 and mDAP3 (4-3) or between two mDAP3 residues (3-3).While 4-3 cross-links are made by penicillin-binding proteins, 3-3 cross-links are the products of L,D-Transpeptidases.Recent studies on the PG of non-model culturable bacteria have uncovered novel modifications in the cell wall, providing intriguing insights into their peptidoglycan biology.In the present study, Espaillat et al. describe how the non-canonical 1-3 crosslinks in Glucanobacter oxydans, a member of alpha-proteobacteria are synthesized and regulated.Through extensive bioinformatics and biochemical approaches, the authors discovered and characterized a new class of enzymes involved in the formation of these 1,3 cross-links, providing evidence of how it differs from its similar counterparts, LDTs (3-3).Additionally, they elucidated the protein structure and offered mechanistic insights into its catalytic activity.This study is truly fascinating presenting clear data and a complete picture.Manuscript is written well with good flow and logic.I thoroughly enjoyed reading the manuscript.This study provides a major breakthrough in the field of peptidoglycan biology and will be of interest to a large class of bacterial research community.
We thank the reviewer for the careful read of our manuscript and their positive comments.
I have no major concerns, but authors should consider the following: 1. Fig. 1A.Representation of T144 needs to be rechecked.Will the central mDAP make 4 bonds which are theoretically impossible?I think the D-ala of central monomer may make a crosslink with mDAP of right monomer.
We thank the reviewer for pointing this out and have fixed the schematics of the T144 in Fig. 1A.
2. Recommend adding a panel to Fig. 1 illustrating all the structures/ denotations of muropeptides including M1, M2 and others (which are already here).
As suggested by the reviewer, we have expanded Fig. 1A to include all the muropeptides indicated in the chromatogram.
3. Fig. 3A.A representation of M2 in the chromatogram is needed for the reader to compare the chromatogram for level of M2.Also label the M2 peak in the chromatogram.
Thanks for pointing this out.We have labelled the M2 peak in the chromatogram in Fig. 3A, but we believe including this muropeptide in the schematics could be misleading as these only illustrate the species that we observe participate in the LD1,3-TPase reaction.
4. In the results section 6 (322-324), it is suggested that three amino acids, Tyr 180, 181, and 182, play crucial role in conferring the characteristic 1-3 cross-linking properties to these enzymes.Have the authors attempted to mutate these residues and assess whether the enzyme still functions as a 1-3 cross-linker?
We are currently working on a follow-up mechanistic study which includes biochemical characterization of LDTGo mutant variants in key residues defined by the structure such as those in the active site or shaping the belt. 5.In Fig. 1, I would suggest adding a western blot showing the expression of catalytic site mutants.
We appreciate the reviewer's suggestion.Unfortunately, the constructs used for complementation in G. oxydans lack a tag that would allow immunodetection of the protein.However, all these constructs were done on the same plasmid background and also induction conditions were the same.A Coomassie-stained SDS-PAGE of total protein extract of G. oxydans samples shows a band corresponding to LDTGo MW in the complemented samples (Fig. R2).Given that the Pm promoter of the pSEVA238 is leaky in Gluconobacter, this band is visible both in non-induced and induced conditions with 1 mM 3-methylbenzoic acid, however it does increase upon addition of the inducer.Further, the intensity of this band is similar for the WT and the catalytic site mutant variants.
[redacted] Additionally, the role of the catalytic site mutants in its homologue LDTBcn is further demonstrated in the heterologous expression in E. coli assays shown in Fig. 2C and in the in vitro results presented in Fig. 3C and Supplementary Fig. 4 where protein and substrate amounts were the same in all the assays performed (Fig. R3). 6.The figures need to be enlarged and font sizes increased for ease in reading.
We have increased font sizes and enlarged the figures whenever possible.7. Discussion, lines 420, this paragraph discusses the regulation and homeostasis of different crosslinks.It is possible that 1,3-crosslink hydrolase may exist in these organisms and it may control the level of 1,3 crosslinks.Only the synthases need not be controlled; hydrolases may also contribute to the crosslinkage frequency.
We appreciate the reviewer's comment.We have added a sentence to the discussion as suggested (lines 445-448).8. May be a sentence on how these crosslinked muropeptides are recycled would help in the Discussion.
We agree with the reviewer that this is an interesting topic.However, G. oxydans has no homologues to AmpG or other known muropeptide transport systems described so far (Gilmore and Cava, Nat Commun 2022, PMID: 36566216).Therefore, we prefer not to speculate and focus on the description and characterization of the novel LD1,3-TPase.9. Line 54-comma to be included after pentapeptide.
Modified accordingly in the revised version of the manuscript (line 47).10.Line 87-'type' after novel can be removed.
Modified accordingly in the revised version of the manuscript (line 80).

Reviewer #3 (Remarks to the Author):
This is an interesting mechanistic study regarding a novel L,D transpeptidase that catalyzes 1-3 stem peptide cross-links.Authors employ extensive bacterial genetics techniques, paired with structural biology and molecular dynamics, to extensively characterize not only the novel transpeptidase itself but also the biochemical and mechanistic environment that lead to its activation.The text is elegantly written and the figures are quite clear.There are a few points that merit clarification, highlighted below.
Fig. S9: Incorporation of Gln in M2 by exchange with D-Glu seems to be independent of LDTGo, since it is present in the deletion mutant.Consequently, the extraction ion chromatogram shown for M2Gln shown in Fig. 4D could be removed because it does not support the D,L-aminoacid exchange activity of LDTGo.Fig. S8: Similar suggestion for the MS spectra of M2Gln shown in Fig. S8.
, especially in what relates to the belt region and other loops?This is shown in Fig 5 but not explored in the text.Minor comments -line 289, should read ' we subsequently leveraged …' -line 607, should read 'Coot was used …'-there are a few units missing in the crystallography table, notably in what relates to wavelength, resolution, B-factors, RMS … in addition, the number of digits to the right of the decimal point could be homogenized.

Fig. 3A :
Fig. 3A: M2 is not visible on the chromatograms.Was it detected only by LC-MS?Indicate its position on the chromatograms in Fig. 2A.

Fig. S9 :
Fig. S9: Incorporation of Gln in M2 by exchange with D-Glu seems to be independent of LDTGo, since it is present in the deletion mutant.Consequently, the extraction ion chromatogram shown for M2Gln shown in Fig. 4D could be removed because it does not support the D,L-aminoacid exchange activity of LDTGo.

Fig. S8 :
Fig. S8: Similar suggestion for the MS spectra of M2Gln shown in Fig. S8.

Figure R2 .
Figure R2.Coomassie-stained gel of total protein extract of the G. oxydans samples shown in Fig. 1C.

Figure
Figure R3.Coomassie-stained gel of purified LDTBcn proteins used in in vitro assays shown in Fig. 3C and Supplementary Fig.4.