Structure and conformational dynamics of Clostridioides difficile toxin A

This study presents a complete structural model of TcdA holotoxin and sheds new lights into the conformational dynamics of TcdA and its roles in TcdA intoxication.

Here, the authors use protein crystallography to determine a neutral pH structure of a TcdA fragment that lacks two enzyme domains at the N-terminus (residues 843 to 2481) but is slightly more extended into the C-terminal repetitive "CROP"-domainscontaining region than previously published structures. The structure provides detailed insight into the lasso-loop-like interaction between the CROPs-tail and the delivery and receptor-binding domain (DRBD) of TcdA. This seems to keep the toxin in a closed stated required to e.g. protect it from protease degradation, and it seems to open up at lower pH when the toxin resides in endosomes and initiates membrane translocation of the glycosylating domain that is at the heart of TcdA's and TcdB's toxicity. Such a higher resolution closed-form structure of TcdA or TcdB was not available until very recently, and the work presented here provides important insight into the structural principles of keeping these toxins in the closed state.
Being a crystallographer myself, I understand that the crystal structure determination reported here required very high skills and was extremely laborious to achieve. I am therefore certain that a recent publication of a similar cryoEM structure of TcdA by a group from Denmark (Aminzadeh et al., 2022; by coincidence I was a referee for this paper as well) must have come as a shock to Chen and colleagues. The present manuscript, however, takes our insight into the structural dynamics of TcdA (and likely TcdB as well) further by employing additional methods such as protease sensitivity and single-molecule FRET experiments to corroborate pH-induced movements within TcdA. Finally, the authors show that toxicity of TcdA is strongly reduced when the toxin is blocked in the closed state.
Overall, the work by Chen et al. is an important contribution for the field that I enjoyed reading, and I have only very few remarks that I suggest changing to improve the manuscript: -In general, it would have been helpful if the manuscript would have come with line and page numbers -The reference to the recently published paper by Aminzadeh needs to be updated -In their discussion of the structure, the authors should more openly indicate if their analysis comes to the same conclusions as the Aminzadeh paper and where it deviates -In Fig. 6A, the small cyan dots are easily overlooked -please consider changing it to something more visible -The final sentence in the paragraph "The interplay between the DRBD and the CROPs of TcdA" reads a little strange, and the reference to Krissinel & Henrick, 2007 sounds as if these authors have done an analysis of TcdA -consider rewording -I could not really make out if "CROPs" is a plural or singular, I would guess it is plural. Consequently, the "undergoes" in the chapter "pH-dependent dynamics of TcdA" should be changed to "undergo" -I may be wrong here. -Data availability: the authors indicate that they have not yet deposited the structure of their TcdA fragment. I think that this is not acceptable, and that a PDB entry code must be provided.
Reviewer #2 (Comments to the Authors (Required)): The authors report on the crystal structure of a fragment of C. difficile toxin A (TcdA) at 3.18 Å resolution, covering residues 843 through 2481, including the receptor binding and translocation domain DRBD and the CROPs domain parts I-V. In addition, they present data on limited proteolysis studies, which were performed at neutral (pH 7.4) and acidic pH, and data of single molecule FRET analyses, which support dynamic changes of the toxin structure at low pH. This is an interesting paper, which is in line with recent publications. Especially, the recent publication on the cryo-EM structure of TcdA (Aminzadeh et al. EMBO reports 2022) has addressed many of the questions studied here. Therefore, novelty and impact of the data are limited. Specific comments: 1. In general, the recent manuscript by Aminzadeh et al. should be discussed in comparison to data presented here more extensively. 2. Sometimes, it is not clear what is new and what is from former studies. Moreover, if data are from modeling this should clearly be stated. For example, dynamic changes of toxin structures are not directly observed in this study, therefore a statement (page 9) like "We found that the SGS domain in the DRBD...." is not really supported by experimental data. 3. Abstract: "Here, we report ..., which fills the knowledge gap...". Actually I could not really see which "knowledge gap" e.g. in comparison to the paper by Aminzadeh et al. is filled. Please change the wording. 4. The citation of Aminzadeh et al. should include the publication year. 5. The first paragraph of the results part can be omitted, because the information is not important. It might be partly included into the Introduction. 6. Please, give the C. difficile strain from which the toxin comes also under Material and Methods. 7. What is known about the impact of the flexible region covering residues 1661-1667. 8. Statement (page 4): "......resulting the first complete structure of TcdA holotoxin...." This is a modeling and has been done by others, too. 9. The presentation and discussion of the structural data is very similar to those of the paper by Aminzadeh et al.. Therefore, these data should be compared with the present data. 10. Page 6: The interactions at interface II are partly different as compared to the paper by Aminzadeh et al.. this should be stated and discussed. 11. Discussion (page 10): The paragraph starting with "While the ....." needs additional references and, again, a discussion about the results obtained by Aminzadeh et al.

Reviewer #3 (Comments to the Authors (Required)):
This manuscript by Chen, et al. presents structural and biochemical data probing the conformational dynamics of TcdA at both neutral and acidic pH. The X-ray crystal structure of a TcdA fragment was determined, covering the majority of the delivery and receptor binding domain (DRBD), including an N-terminal region the authors label the small globular subdomain (SGS), along with CROPs repeats I-V. With this new structural data, the authors assemble a complete model of the TcdA holotoxin. Importantly, the authors contrast their structure, determined at neutral pH, with existing TcdA structures (Aminzadeh, et al. 2021;Pruitt, et al. 2010), and with the TcdB holotoxin structure, determined at low pH. Although previously established that TcdA and other large clostridial family toxins undergo drastic pH-dependent conformational rearrangements, the authors identify specific structural states (open and closed) that the toxin may sample under endosomal pH. These were confirmed by limited trypsin digestion at pH 7.4 and pH 5.4, identification of trypsin cleavage products by mass spectrometry, structure-based site-directed mutagenesis to generate TcdA mutants that either disrupt and strengthen interactions between the DRBD and CROPs, singlemolecule FRET to establish TcdA's pH-dependent dynamics, and lastly, cell-rounding assays to test the cytotoxicity of TcdA mutants that influence the flexibility of the CROPs. The authors conclude that the SGS of the DRBD undergoes coordinated conformation changes with the CROPs to protect the pore-forming region from proteolysis at neutral pH, and allow for exposure of the pore-forming region at endosomal pH.
The manuscript is well-written and easy to follow. The experiments performed were thorough and considerate of outstanding questions concerning TcdA dynamics and how the information generated by these studies may inform future efforts to neutralize TcdA through limiting its structural plasticity. All claims are well supported by the data.
Major Concerns: None.

Structure and conformational dynamics of Clostridioides difficile toxin A
We thank the editors and reviewers for their careful reading of the manuscript and their constructive suggestions that have guided our revision. We have revised the manuscript to address reviewers' concerns as outlined below. We would like to point out that our studies presented here were carried out independently, using different experiments and analyses, from the beautiful cryoEM studies recently published by Jørgensen and colleagues (Aminzadeh et al. EMBO reports 2022). Our crystal structure was determined back in 2019 but we decided to continue with more mechanistic studies, which is our ultimate goal when resolving a protein structure. Unfortunately, these studied were delayed by COVID restrictions. As Reviewer #1 pointed out, the recent structure paper did come as a shock to us when we were preparing our manuscript. We are glad that our two groups arrived at similar conclusions regarding the structure of TcdA and in fact corroborate each other's findings. We cite Aminzadeh et al. throughout our paper, and have added more structural comparisons in the revised manuscript. However, we hope the reviewers appreciated that we have presented our data as an independent work, with a neutral point of view, instead of focusing on comparison with the recent cryoEM structure.

Reviewer #1
The manuscript, Structure and conformational dynamics of Clostridioides difficile toxin A, by Chen and colleagues describes crystallographic, biochemical, biophysical and cell-biology work with TcdA, one of the large toxins of the pathogenic bacterium C. difficile. With over 2700 amino acids, TcdA is the "little" brother of TcdB, and high-resolution structure determination of more or less complete TcdA/B has only recently been achieved in a handful of landmark papers. This breakthrough was also fueled by the better availability of cryoEM methods.
Here, the authors use protein crystallography to determine a neutral pH structure of a TcdA fragment that lacks two enzyme domains at the N-terminus (residues 843 to 2481) but is slightly more extended into the C-terminal repetitive "CROP"-domains-containing region than previously published structures. The structure provides detailed insight into the lasso-loop-like interaction between the CROPs-tail and the delivery and receptor-binding domain (DRBD) of TcdA. This seems to keep the toxin in a closed stated required to e.g. protect it from protease degradation, and it seems to open up at lower pH when the toxin resides in endosomes and initiates membrane translocation of the glycosylating domain that is at the heart of TcdA's and TcdB's toxicity. Such a higher resolution closed-form structure of TcdA or TcdB was not available until very recently, and the work presented here provides important insight into the structural principles of keeping these toxins in the closed state.
Being a crystallographer myself, I understand that the crystal structure determination reported here required very high skills and was extremely laborious to achieve. I am therefore certain that a recent publication of a similar cryoEM structure of TcdA by a group from Denmark (Aminzadeh et al., 2022;  -The final sentence in the paragraph "The interplay between the DRBD and the CROPs of TcdA" reads a little strange, and the reference to Krissinel & Henrick, 2007

sounds as if these authors have done an analysis of TcdA -consider rewording
Response: Sorry for the confusion. This reference was cited here for the PISA program we used. We have deleted this reference from this sentence and moved it to the Method: "Calculation of the buried molecular surface area was carried out using PISA (Proteins, Interfaces, Structures and Assemblies) program (Krissinel & Henrick, 2007)." (Lines 512-514, Page 12) -I could not really make out if "CROPs" is a plural or singular, I would guess it is plural. Consequently, the "undergoes" in the chapter "pH-dependent dynamics of TcdA" should be changed to "undergo" -I may be wrong here.
Response: It has been changed to "undergo".
-Data availability: the authors indicate that they have not yet deposited the structure of their TcdA fragment. I think that this is not acceptable, and that a PDB entry code must be provided.

Response:
The structure has been deposited to PDB under access code 7U1Z.

Reviewer #2
The authors report on the crystal structure of a fragment of C. difficile toxin A (TcdA) at 3.18 Å resolution, covering residues 843 through 2481, including the receptor binding and translocation domain DRBD and the CROPs domain parts I-V. In addition, they present data on limited proteolysis studies, which were performed at neutral (pH 7.4) and acidic pH, and data of single molecule FRET analyses, which support dynamic changes of the toxin structure at low pH. This is an interesting paper, which is in line with recent publications. Especially, the recent publication on the cryo-EM structure of TcdA