Structural basis of cell wall anchoring by SLH domains in Paenibacillus alvei

Self-assembling protein surface (S-) layers are common cell envelope structures of prokaryotes and have critical roles from structural maintenance to virulence. S-layers of Gram-positive bacteria are often attached through the interaction of S-layer homology (SLH) domain trimers with peptidoglycan-linked secondary cell wall polymers (SCWPs). Here we present an in-depth characterization of this interaction, with co-crystal structures of the three consecutive SLH domains from the Paenibacillus alvei S-layer protein SpaA with defined SCWP ligands. The most highly conserved SLH domain residue SLH-Gly29 is shown to enable a peptide backbone flip essential for SCWP binding in both biophysical and cellular experiments. Furthermore, we find that a significant domain movement mediates binding by two different sites in the SLH domain trimer, which may allow anchoring readjustment to relieve S-layer strain caused by cell growth and division.

The manuscript by Blackler and co-workers describes a careful crystallographic study performed with the SLH domains of the S-layer protein SpaA, in unbound form as well as associated to mimics of secondary cell wall polymers. Proteins involved in S-layer formation are interesting targets for the development of new, potential antibiotics, and work in this field is thus timely and of interest.
SpaA from the strain Paenibacillus alvei was thus selected for this work, and the objective was to employ X-ray crystallography and isothermal titration calorimetry to perform an extended investigation on the binding characteristics between SpaA and secondary cell wall polymers. Despite the fact that the structure of the SHL region of SpaA reported here is new, SLH domain proteins have been crystallized and their structures have been reported. Of note is the structure of Sap from B. anthracis, solved by Joachimiak and Schneewind and referred to in the manuscript, whose fold is highly reminiscent of the structure presented here.
Blackler and colleagues investigate the structure of SpaA in complex with a synthetic monosaccharide by solving the crystal structure of the complex in different space groups, and discuss the interactions observed in detail. In most cases, the ligand is bound in pocket G2, with the exception of the structure solved in space group C2, where they could observe some density in G1. Using ITC, they identify that the interaction between the monosaccharide and SpaA (in G2) occurs with nanomolar affinity, and relate binding in the G1 pocket to a potential artifact. They continue their work by solving the structure in the presence of a disaccharide, which binds in G2 similarly to the monoscaccharide. The most interesting observation involves a local glycine flip that occurs upon ligand binding within one of the important motifs. In order to characterize the flip, they created glycine-to-alanine single and double mutants, and characterized them by crystallography and ITC. The authors then discuss the structural contributions of different conserved residues to ligand binding.
The manuscript describes crystal structures that have been solved and refined carefully, and the local modifications engendered by ligand binding are certainly of interest for specialists in the field. However, in the absence of any microbiological or cellular data indicating how these results could be relevant for control of bacterial survival or novel inhibitor development, I believe that the paper would be more appropriate for a specialized structural biology journal. Authors should also ensure that labels are clearly visible, especially in figures that involve several superimpositions, such as Figs. 2, 3, and 4.

Reviewer #3 (Remarks to the Author):
The manuscript presented here by Blackler et. al. attempts to understand the role of SLH domains in the cell-wall of Gram-positive bacteria. Overall the manuscript is interesting, scientifically sound, and provides an extremely detailed look at how these domains interact with the cell-wall. An understanding of the S-layer and the cell-wall in general is not only academically fascinating but a critical step towards the creation of new antibiotics.
Overall based on scientific merit and importance this is a very nice study. The manuscript could benefit from some re-writing to better explain and put into context various experiments and hypothesis, for example a more in depth introduction and a more extended discussion. Currently, it is unclear what certain terms are or why certain experiments were performed unless you are familiar with the field (see comments).
Additionally, the authors have made some very interesting structural observations that may result directly to the biology of how SLH proteins interact with the S-layer. In particular, the study of the conserved glycine-motifs is very interesting but requires in vivo experimentation to prove.
The recommendation is to revise the manuscript before the final decision.

Major Comments
The authors have shown biochemically and structurally that the glycine residues in the conserved GIIxG motif are an important part of the ligand binding mechanism. However, they do not cite any literature where these mutations have been tested in bacteria. Is this experiment possible, or are there papers they can cite? If this is a completely new finding, the authors should test the glycine point mutations in their species (if possible) but at a minimum the equivalent mutations in another Gram-positive bacteria, such as B. subtilis, that is routinely studied genetically.
Page 11, Mutual-exclusive... I do not believe the conclusion here is justified. First, the authors use ITC to state that G1 binding is artificial, then they make mutants to show that binding in G1 is possible. They then conclude this is a novel form of intramolecular negative cooperativity. This section is best saved for the conclusions and discussed in a more hypothetical manner. For example, they could only observe a significant structural change by mutating a highly conserved residue. How do the authors show that it's not just an artifact of manipulating the system? Their conclusion/hypothesis here about the dual-binding in the context of the S-layer is intriguing and very possibly correct, but it hasn't been proven.

Introduction
Page 3. "there are over 54,000 specific hits…" What type of hits? The SLH domain? What was the search based upon exactly?
Page 4. "SLH domains and SCWP not possible previously." Switch possible and previously.
Page 4, last paragraph. What is mean by Inequality within other SLH domain repeats? Also, it is a better read if the last paragraph of the introduction highlights or at least previews one or two major points from the paper.
The introduction seems very short. Maybe explain the system better? For example, when it gets to the results its not explained why they used certain sugars for the co-crystal structures and biochemical characterization. It is eluded to, but when it comes up maybe state exactly why each ligand was used (eg. the Pyr is critical as mentioned, GlnNAc, ManNAc -why? -its nonobvious to the general reader).

Results
Page 5, first paragraph. Maybe compare to the B. antracis Sap? Is there anything biologically important from their similarities and differences? For example, the SLH domains and the G1-3 grooves, are they exactly the same as Sap?
Page 5, first paragraph. Maybe make it clear that Val125 of TVEE is in an equivalent position as R, but due to the residue change doesn't have the same contacts? Is that what the authors meant to imply?
Page 5, last paragraph."unliganded trigonal structures…" only a crystallographer is going to understand what you mean -maybe reference the table and PDB files here as well?
Page 6, first paragraph. Is this loop specifically not involved in any crystal packing contacts in any of the structures? What are the b-factors like in the 44-55 loop? Are they high relative to the mean of the structure?
Page 6, second paragraph. The authors need to place a starting sentence that explains why they performed ITC to address what problem. It seems just dropped in here without context. Additionally, the authors cannot say with 100% certainty that this is an artifact. They should put a qualifier "most likely, we believe, etc." Did the authors perform the control of buffer into protein? If so, it should be added to the extended data figure.
Page 7, last paragraph. The glycine flip is really interesting, however a table is not sufficient. The authors should show structural examples with the electron density of the region to show the flip.
Page 8, "To test this hypothesis, the SpaA/G109A" change "the" to "a". Additionally, the ligand still binds but is close to an order of magnitude worse in affinity, correct? 26 nM vs. 226 nM? The authors should state/say something about this.
Page 8, middle paragraph. "The unliganded crystal structure" The phrasing needs reworking.
Page 9, first paragraph. Please reference the extended data figure for the ITC.
"Interestingly, despite…" This is a long run-on sentence. Consider revising.
Page 9, Inequality within tandem. Please be sure that inequality is defined and its clear what this means, as per previous comment.
Page 12, conclusions. "These crystal structures…" Consider re-phrasing? "The crystal structures presented here…" I would like to express my gratitude to the Reviewers for their careful review. Below are the Reviewers" comments reproduced verbatim, with our responses interspersed.

Reviewer #1 (Remarks to the Author):
This manuscript by Blackler et al describes the complete structure of the P. alvei SLH domain in complex with its secondary cell wall polysaccharide ligand. The SLH domain is the most widespread mechanism of S-layer anchoring in the Gram positive bacteria. This structure provides the first insight into the molecular basis of SLH function.
Overall I have no criticisms of the experimental design or execution. However there are some important omissions in background context and interpretation that I feel must be addressed for the sake of clarity in the scientific record.
COMMENT 1: The first description of the interaction between SLH domains and a pyruvylated SCWP was Mesnage et al (2000) EMBO J 19:4473-4484. Despite that study focusing on the S-layer of B. anthracis, it seems very strange that this paper has not been cited.
The structure of the B. anthracis SLH domain has also been previously described by the Schneewind lab. The only reference to this structure in the introduction is a passing comment on page 4. Although the structure is mentioned again at the beginning of the results, no detailed description is given. A more rounded comparison of this structure with that published previously should be added for completeness. I would suggest that a supplemental figure comparing the structures should be added at a minimum.  (2000) EMBO citation was an oversight. This paper is now referenced in the introduction.
We have also added a more detailed description of the published structure of B. anthracis Sap SLH to the introduction, and a new results section titled "Comparison of SpaA SLH to other SCWP-binding domains" that includes a thorough comparison (including a supplementary figure).  Although these structures are apparently distinct modules for S-layer anchoring, the structures appear to be strikingly similar.

RESPONSE 1.3:
The reviewer is correct that space constraints were the reason for the brief conclusions section, as the manuscript was initially prepared for Nature with a stricter word limit. In reformatting for Nature Communications, we have now appropriately revised and expanded the Conclusions (now Discussion). An additional results section "Comparison of SpaA SLH to other SCWP-binding domains" includes thorough comparisons to the recent structures of B. anthracis Sap SLH and C. difficile CWB2 domain-containing proteins Cwp6 and Cwp8.
Interestingly, the structure of Sap SLH in complex with a synthetic SCWP ligand published during our revisions (Sychantha et al. 2018, Biochemistry. DOI: 10.1021/acs.biochem.8b00060) revealed ligand bound only in G2, with G2 in a "closed" conformation and G1 and G3 in "open" conformations. In our study, wild-type Sap SLH was also observed with ligand bound in a "closed" G2 with G1 and G3 "open", while Sap SLH /G109A was observed instead with ligand bound in a "closed" G1 with G2 and G3 "open".
Both G1 and G3 of Sap SLH contain conserved residues involved in SCWP binding, which suggests Sap SLH may also utilize a switchable binding mechanism. However, it is important to note that Sychantha et al. do not hypothesize this binding mechanism, and that they have completely overlooked the function of the conserved SLH-Gly29 and have modelled it incorrectly in their deposited data. Therefore, the findings and discussion presented in our study remain completely novel.

Reviewer #2 (Remarks to the Author):
The manuscript by Blackler and co-workers describes a careful crystallographic study performed with the SLH domains of the S-layer protein SpaA, in unbound form as well as associated to mimics of secondary cell wall polymers. Proteins involved in S-layer formation are interesting targets for the development of new, potential antibiotics, and work in this field is thus timely and of interest.
SpaA from the strain Paenibacillus alvei was thus selected for this work, and the objective was to employ X-ray crystallography and isothermal titration calorimetry to perform an extended investigation on the binding characteristics between SpaA and secondary cell wall polymers. Despite the fact that the structure of the SHL region of SpaA reported here is new, SLH domain proteins have been crystallized and their structures have been reported. Of note is the structure of Sap from B. anthracis, solved by Joachimiak and Schneewind and referred to in the manuscript, whose fold is highly reminiscent of the structure presented here.
Blackler and colleagues investigate the structure of SpaA in complex with a synthetic monosaccharide by solving the crystal structure of the complex in different space groups, and discuss the interactions observed in detail. In most cases, the ligand is bound in pocket G2, with the exception of the structure solved in space group C2, where they could observe some density in G1. Using ITC, they identify that the interaction between the monosaccharide and SpaA (in G2) occurs with nanomolar affinity, and relate binding in the G1 pocket to a potential artifact. They continue their work by solving the structure in the presence of a disaccharide, which binds in G2 similarly to the monoscaccharide. The most interesting observation involves a local glycine flip that occurs upon ligand binding within one of the important motifs. In order to characterize the flip, they created glycine-to-alanine single and double mutants, and characterized them by crystallography and ITC. The authors then discuss the structural contributions of different conserved residues to ligand binding.
The manuscript describes crystal structures that have been solved and refined carefully, and the local modifications engendered by ligand binding are certainly of interest for specialists in the field.
COMMENT 2.1: However, in the absence of any microbiological or cellular data indicating how these results could be relevant for control of bacterial survival or novel inhibitor development, I believe that the paper would be more appropriate for a specialized structural biology journal.

RESPONSE 2.1:
We have performed additional in vivo studies that prove the conserved SLH-Gly29 is required for SCWP binding and cell-surface protein anchoring in a biological context. Unfortunately, direct in vivo experiments with SpaA are impossible because genetic manipulation of the spaA gene in P. alvei results in a lethal phenotype. Therefore, to address the reviewers" concerns, we probed a different SLH domain-containing cell surface protein, SlhA. It was shown previously that deletion of the P. alvei slhA gene produces changes in colony morphology and impedes biofilm formation and swarming motility of P. alvei cells. We now show that two single-point SLH-Gly29Ala mutations in the slhA gene produces the same phenotype as gene deletion, thus indicating a loss of cell-surface anchoring and proving the significance of this conserved residue in a biological context. S-layers are critical for the survival and virulence of diverse microorganisms, and we believe that our detailed elucidation of a conserved anchoring mechanism, now supported by in vivo data using a second P. alvei SLH domain-carrying surface protein, will be of broad interest and significance. The manuscript presented here by Blackler et. al. attempts to understand the role of SLH domains in the cell-wall of Gram-positive bacteria. Overall the manuscript is interesting, scientifically sound, and provides an extremely detailed look at how these domains interact with the cell-wall. An understanding of the S-layer and the cell-wall in general is not only academically fascinating but a critical step towards the creation of new antibiotics.
Overall based on scientific merit and importance this is a very nice study. The manuscript could benefit from some re-writing to better explain and put into context various experiments and hypothesis, for example a more in depth introduction and a more extended discussion. Currently, it is unclear what certain terms are or why certain experiments were performed unless you are familiar with the field (see comments).
Additionally, the authors have made some very interesting structural observations that may result directly to the biology of how SLH proteins interact with the S-layer. In particular, the study of the conserved glycine-motifs is very interesting but requires in vivo experimentation to prove.
The recommendation is to revise the manuscript before the final decision.

Major Comments
COMMENT 3.1: The authors have shown biochemically and structurally that the glycine residues in the conserved GIIxG motif are an important part of the ligand binding mechanism. However, they do not cite any literature where these mutations have been tested in bacteria. Is this experiment possible, or are there papers they can cite? If this is a completely new finding, the authors should test the glycine point mutations in their species (if possible) but at a minimum the equivalent mutations in another Gram-positive bacteria, such as B. subtilis, that is routinely studied genetically.

RESPONSE 3.1:
The involvement of SLH-Gly29 and the GIIxG motif in SCWP binding is a novel finding of our study, and mutations of this motif have not been assessed previously to the best of our knowledge. As discussed in Response 2.1, we performed additional in vivo studies that prove the conserved SLH-Gly29 is required for SCWP binding in a biological context, and these results are included in the revised manuscript. With regard to testing SLH-Gly29 mutations in other species; assessing SLH-Gly29 mutations in B. subtilis is not possible as there are no SLH domain-containing proteins identified in that organism. Switching to an additional new system, where we do not know the structure of the cell wall ligand, let alone have available pure, synthesized partial structures of the ligand would, we feel, be well beyond the scope of this study.
COMMENT 3.2: Page 11, Mutual-exclusive... I do not believe the conclusion here is justified. First, the authors use ITC to state that G1 binding is artificial, then they make mutants to show that binding in G1 is possible. They then conclude this is a novel form of intramolecular negative cooperativity. This section is best saved for the conclusions and discussed in a more hypothetical manner. For example, they could only observe a significant structural change by mutating a highly conserved residue. How do the authors show that it"s not just an artifact of manipulating the system? Their conclusion/hypothesis here about the dual-binding in the context of the S-layer is intriguing and very possibly correct, but it hasn"t been proven. RESPONSE 3.2: Based on 1:1 binding with 29 nM K D measured by ITC for wild-type SpaA SLH binding to the SCWP monosaccharide 4,6-Pyr--D-ManNAcOMe, and the consistent observation of ligand bound in G2 of crystal structures, we did conclude that the single observation of ligand bound with fragmented electron density in G1 of wild type SpaA SLH was an artifact of crystallization. However, when binding in G2 was disrupted by SLH-Gly29Ala mutation, we observed a significant structural change that allowed binding in G1 with 226 nM K D . Unfortunately, it is impossible for us to prove or disprove that G1 is functional in wild-type SpaA SLH because G2 has higher affinity for ligand and binding in G2 precludes the structural change required for binding in G1. However, we believe that the maintenance of highly conserved residues in G1 and its displayed 226 nM KD suggest that it does have biological significance. Nevertheless, we acknowledge that this section is largely hypothetical and have moved it to the discussion as recommended by the reviewer.
COMMENT 3.3: This paper definitely requires more discussion. A single paragraph seems insufficient.
COMMENT 3.10: The introduction seems very short. Maybe explain the system better? For example, when it gets to the results its not explained why they used certain sugars for the cocrystal structures and biochemical characterization. It is eluded to, but when it comes up maybe state exactly why each ligand was used (eg. the Pyr is critical as mentioned, GlnNAc, ManNAcwhy?its non-obvious to the general reader). RESPONSE 3.10: Introduction was expanded to explain the system better, including descriptions of SCWPs and SLH domain structures. A new figure (Fig. 2) was added that shows schematics of selected SCWPs and the corresponding ligands used in the study.

Results
COMMENT 3.11: Page 5, first paragraph. Maybe compare to the B. antracis Sap? Is there anything biologically important from their similarities and differences? For example, the SLH domains and the G1-3 grooves, are they exactly the same as Sap? COMMENT 3.12: Page 5, first paragraph. Maybe make it clear that Val125 of TVEE is in an equivalent position as R, but due to the residue change doesn"t have the same contacts? Is that what the authors meant to imply? RESPONSE 3.12: That is correct. We revised this sentence for clarity: "In the case of SLH2, Val125 of the TVEE motif corresponds in position to the conserved SLH-Arg43 of the TRAE and TRAQ motifs but does not protrude into the neighboring G3." COMMENT 3.13: Page 5, last paragraph "unliganded trigonal structures…" only a crystallographer is going to understand what you meanmaybe reference the table and PDB files here as well? RESPONSE 3.13: All structures are now referred to by PDB ID.
COMMENT 3.14: Page 6, first paragraph. Is this loop specifically not involved in any crystal packing contacts in any of the structures? What are the b-factors like in the 44-55 loop? Are they high relative to the mean of the structure? RESPONSE 3.14: we have added a sentence about crystal contacts and B-factors of this loop, and added Supplementary Fig. 1 showing multiple conformations colored by B-factor. COMMENT 3.15: Page 6, second paragraph. The authors need to place a starting sentence that explains why they performed ITC to address what problem. It seems just dropped in here without context. Additionally, the authors cannot say with 100% certainty that this is an artifact. They should put a qualifier "most likely, we believe, etc." RESPONSE 3.15: Paragraph was revised to clarify the incentive for ITC and qualify the hypothesis.