Polar localization of CheO under hypoxia promotes Campylobacter jejuni chemotactic behavior within host

Campylobacter jejuni is a food-borne zoonotic pathogen of worldwide concern and the leading cause of bacterial diarrheal disease. In contrast to other enteric pathogens, C. jejuni has strict growth and nutritional requirements but lacks many virulence factors that have evolved for pathogenesis or interactions with the host. It is unclear how this bacterium has adapted to an enteric lifestyle. Here, we discovered that the CheO protein (CJJ81176_1265) is required for C. jejuni colonization of mice gut through its role in chemotactic control of flagellar rotation in oxygen-limiting environments. CheO interacts with the chemotaxis signaling proteins CheA and CheZ, and also with the flagellar rotor components FliM and FliY. Under microaerobic conditions, CheO localizes at the cellular poles where the chemosensory array and flagellar machinery are located in C. jejuni and its polar localization depends on chemosensory array formation. Several chemoreceptors that mediate energy taxis coordinately determine the bipolar distribution of CheO. Suppressor screening for a ΔcheO mutant identified that a single residue variation in FliM can alleviate the phenotype caused by the absence of CheO, confirming its regulatory role in the flagellar rotor switch. CheO homologs are only found in species of the Campylobacterota phylum, mostly species of host-associated genera Campylobacter, Helicobacter and Wolinella. The CheO results provide insights into the complexity of chemotaxis signal transduction in C. jejuni and closely related species. Importantly, the recruitment of CheO into chemosensory array to promote chemotactic behavior under hypoxia represents a new adaptation strategy of C. jejuni to human and animal intestines.

Thank you again for your submission. We hope that our editorial process has been constructive so far, and we welcome your feedback at any time. Please don't hesitate to contact us if you have any questions or comments. Michael Malim Editor-in-Chief PLOS Pathogens orcid.org/0000-0002-7699-2064 *********************** Three experts in the field of bacterial motility and chemotaxis reviewed this work. Overall, the Reviewers concluded that this work is interesting as it identifies a new type of chemotaxis protein in Campylobacter and related species that has the potential to sense or be affected by oxygen tension or redox status. However, two Reviewers believe that there are significant experiments and details lacking in the work that leave too much ambiguity to develop strong conclusions for how CheO functions in chemotaxis in C. jejuni. The work was also reviewed by an editor and the editor largely concurs with all of the Major Criticisms listed by Reviewers 2 and 3. The authors should seriously consider all of these comments to improve the work. Furthermore, an editor notes that significant details are lacking from the Materials and Methods, such as detailed protocol for how the co-immunoprecipitation studies were performed (how were the proteins mixed together -purified proteins, proteins in cell lysates, mixing and washing conditions etc) and other methods that would limit the ability of others to perform similar analyses. The authors should examine this section and provide additional details for this method and potentially others.

Part I -Summary
Please use this section to discuss strengths/weaknesses of study, novelty/significance, general execution and scholarship.
Reviewer #1: This is a very comprehensive, elegant and detailed study of a new chemotaxis regulator in campylobacters. The manuscript is well written, and the study will greatly contribute the chemotaxis field in general. There are a few thoughts for the authors to consider and some minor language corrections: Response: We thank the reviewer for positive comments.
In your discussion: Consider postulating that CheO is likely to associate with CheA within the chemosensory array in a manner similar to CheY and that is the reason that once the array formation is destabilised, it does not localise to the poles.
Response: Thanks. We have added this postulation in the discussion, please see Lines 372-375: "Furthermore, CheO is likely to associate with CheA within the chemosensory array in a manner similar to CheY but our in vitro phosphorylation assays did not yield conclusive results. Thus, the detailed mechanism of CheO in chemotaxis signal transduction is unclear." Line 361-364: Consider that vigorous motility does not really allow for adaptation, but rather allows the cell to sense and relocate to more favourable or preferred oxygen environment/level.
Response: Thanks for the good point. We have changed the sentence as "Moreover, because CheO is more necessary in microaerobic than in aerobic conditions, this protein likely promotes more vigorous chemotactic motility to help C. jejuni sense and relocate to niches with preferred oxygen level and other favorable conditions." (Lines 392-394).
There are a few suggestions for the discussions within the minor comments below: Reviewer #2: The manuscript describes a chemotaxis protein responsive to oxygen, named CheO. CheO homologs are conserved in Campylobacter, Helicobacter and Wolinella. The authors show that a functional CheOsgfp localizes at cell poles in a CheAVW-dependent manner, suggesting its localization depends on chemotaxis signaling array formation. CheOsgfp polar localization is observed when cells are grown under microaerobic conditions but is lost under aerobic conditions. The contribution of CheO to chemotaxis is also increased under microarobic conditions compared to aerobic conditions. CheO appears to affect the swimming reversal, but not speed, and the authors provide evidence that it physically interacts with CheA, CheZ, FliM and FliY. The authors also show that CheO polar localization depends on the presence of energy taxis receptors (tlp6, tlp9aer1 and aer2). The findings are potentially significant and are certainly novel. However, some of the experiments are incomplete and some conclusions are preliminary with the data presented. The discussion is also highly speculative and lacks citation of relevant references. this reviewer suggests a few additional experiments and a more focused discussion and data presentation.
Response: We thank the reviewer for constructive comments. We have performed additional experiments and revised the manuscript accordingly, please see detailed answers in Part II & III.
Reviewer #3: Mo and colleagues report the study of the newly identified CheO protein of C. jejuni. They present a solid characterization showing that cheO is important for chemotaxis particularly under microaerobic conditions. The protein localizes to the pole more under microaerobic conditions, possibly due to interactions with a subset of chemoreceptors. This part of the manuscript is interesting, and could be developed more. The authors show that CheO interacts with multiple chemotaxis and flagella proteins. These findings are well supported, but confusing because it seems surprising for a protein to have such wide interactions, and given that all are present in aerobic and microaerobic, not clear why CheO would lose the interactions and move off the pole. Overall, this work is interesting but would be more powerful if further developed.
Response: We appreciate the reviewer's comments and briefly clarify the confusing part here: (1) The loss of polar localization of CheO in quadruple mutant Δtlp6Δtlp9Δaer1Δaer2 suggests that these energy taxis receptors can affect CheO localization but does not suggest that they directly interact with CheO. We performed new BTH experiment and the results showed that CheO does not directly interact with any chemoreceptors (new Fig. S7). We only confirmed that CheO interacts with CheA, CheZ, FliM and FliY.
(2) New experiment suggest that FliMY does not affect the CheO localization (new (3) Δtlp6Δtlp9Δaer1Δaer2 retains chemotaxis ability suggesting that the chemosensory array are still formed (new Fig. S6). Based on current results, we suspect that upon sensing oxygen level change, a subset of chemoreceptors can affect the chemosensory array conformation and particularly the CheA kinase, thus indirectly affect the CheO localization and function.
We carefully answered all the questions in Part II & III.

Part II -Major Issues: Key Experiments Required for Acceptance
Please use this section to detail the key new experiments or modifications of existing experiments that should be absolutely required to validate study conclusions.
Generally, there should be no more than 3 such required experiments or major modifications for a "Major Revision" recommendation. If more than 3 experiments are necessary to validate the study conclusions, then you are encouraged to recommend "Reject".
Reviewer #1: no major issues Reviewer #2: 1. CheO is proposed to physically interact with CheA, CheZ, FliM and FliY and to localize to the cell poles in a CheAVW dependent manner, as expected as well as in presence of tlp6, tlp9aer1aer2. It is unclear how these multiple interactio0ns as well as an "energy taxis" receptors dependent localization and major role under microaerobic conditions may take place but a few experiments could clarify the findings. Does CheO interact with the energy taxis receptors?
Response: We examined the interaction between CheO and energy taxis receptors Tlp6, Tlp9, Aer1, Aer2 and additional five Tlps (Tlp1/4/7/8/10) by BTH method (Note: Tlp 3 and Tlp5 were not included because they are pseudogenes in C. jejuni 81-176; Tlp2 was not cloned here because it has an identical cytoplasmic MA domain as Tlp4). Does CheO-sGFP localization at the cel poles in aerobic versus microaerobic conditions depends on CheAVW as well as these energy taxis receptors?

As shown in new
Response: Under aerobic condition, CheO-sfGFP lost polar localization in the presence of CheVAW as well as all chemoreceptors as shown in Fig. 3A.
These results suggest that the polar localization of CheO-sfGFP depends on CheVAW and coordinated function of several energy taxis receptors under microaerobic conditions; the loss of polar localization under aerobic condition is likely related to energy taxis receptors that can affect the chemosensory array conformation and particularly CheA kinase upon sensing oxygen and redox status.
What is the expression patterns of the energy taxis receptors in aerobic versus microaerobic conditions?
Response: To answer this question, we refer to a paper from David J Kelly's group which is a comprehensive study of C.jejuni transcriptome and proteome under different oxygen availability (PMID: 28892295). We checked the expression patterns of all chemotaxis genes in this genome-wide study and found that only Tlp6 is significantly up-regulated at both mRNA and protein levels at low oxygen availability, while Tlp9 is down-regulated at mRNA level from high to low oxygen. All chemotaxis genes with differential expression at transcriptional level were summarized in the table below, while their proteomics data are not shown due to no significant changes or low abundance except Tlp6.
Our results showed that the polar localization of CheO is not altered in Δtlp6 or Δtlp9Δaer1Δaer2 but is lost in quadruple mutant Δtlp6Δtlp9Δaer1Δaer2, suggesting that the polar localization of CheO depends on concerted effects of more than one Tlps including Tlp6 and Tlp9 (Aer1 and Aer2 are PAS domain alone proteins that interact with Tlp9). But these two Tlps showed distinct expression pattern from high to low oxygen level according to the David J Kelly's study. We understand that the reviewer is interested in how energy taxis receptors can affect CheO localization and function. So far, we can only demonstrate that CheO does not directly interact with Tlps. In addition, the answer below showed the absence of FliMY does not affect CheO localization at the pole. Altogether, we speculate that the energy taxis receptors might affect the chemosensory array conformation to change the interaction between CheA/Z and CheO.
2. The authors provide strong evidence that CheO interacts with FliM (ppi assays, suppressor analysis and role of cheO on swimming reversals). However, the discussion and most conclusions revolve around the link between cheO, microaerobiosis and energy taxis receptors. How does CheO-sgfp localize in mutants lacking FliY, FliM or relative to these proteins? Are all of these proteins expressed equally in aerobic versus microaerobic conditions?
Response: We introduced cheO-sfGFP into ΔfliMY mutant and the absence of FliM and FliY does not affect the polar localization of CheO (new Fig. S8). We also added this new result in the Results: "Besides, since the above results suggested CheO directly interacts with two flagellar rotor components FliM and FliY, we investigated whether these flagellar proteins affect CheO localization. A copy of cheO-sfGFP was introduced into ΔfliMY mutant to replace the native cheO, and fluorescence imaging showed bipolar distribution of CheO-sfGFP in C. jejuni cells in the absence of FliM and FliY (Fig. S8)." (Lines 289-294).
Both FliM and FliY genes did not show significantly differential expression under different oxygen levels in David J Kelly's study (PMID: 28892295).
3. Discussion : the last two paragraphs lack relevant references that include the published role of Tlp6 CZB in Helicobacter, the subcellular localization of TlpD in H. pylori that is relevant here and the aer1/aer2 and PAS domains of Tlp9 as well as what is known about energy taxis in these different species. Given the multiple interactions of CheO suggested here, additional information on the flagellar motor and chemotaxis in this species should be included. As is, the discussion is speculative and too removed from the existing literature.
Response: The lack of related references in Discussion is an unintentional mistake. We now added all relevant references regarding Tlp6/9 and flagellar rotor and also moderately extended our discussion (see the last three paragraphs in Discussion).
Reviewer #3: Major comments 1. The expression of cheO under aerobic and microaerobic conditions. Fig 3E shows that the expression level of cheO gene remains unchanged in aerobic and microaerobic condition. What about the protein, which could be done by anti-GFP? This is important because if we look at the middle picture of panel A (Fig. 3), the total fluorescence intensity of cheO-sfGFP along the cell body is very different between aerobic and microaerobic conditions. The fluorescence intensity of CheO-sfGFP in microaerobic condition is much higher than that under aerobic condition. The expression level of gfp could affect GFP intracellular distribution.
Response: We performed additional experiment to examine the protein expression level of CheO-sfGFP under both aerobic and microaerobic conditions. For the western blots with anti-GFP antibody and anti-RNA polymerase beta antibody, we controlled the cell lysates and loading volume at the same number for each sample. As shown below, the amount of CheO-sfGFP in equal amount of cell lysates did not show significant differences under microaerobic and aerobic conditions. However, the anti-RNA polymerase beta antibody (for E. coli) does not work on C. jejuni to serve as a control, so we did not add this result in the manuscript.
To further clarify this question, CheO is not a transmembrane protein. It is a cytoplasmic protein and the fusion gene copy cheO-sfGFP was introduced to C. jejuni wild type strain via recombination to replace the native cheO gene. So the expression of cheO-sfGFP is under native promoter and there are no overexpression of this protein.
2. The result of CheO relocalization is very interesting--but having the GFP fusion could be introducing artifacts. Could authors confirm this the localization of CheO under different conditions using immonofluorescence with untagged CheO, or at least differently tagged CheO?
Response: Thanks for the suggestion. We do not have an antibody for CheO but we have tried immunofluorescence experiments with Anti-FLAG antibody for CheO-3xFLAG fusion protein. The quality of the immunofluorescence images was not as good as the in situ CheO-sfGFP expression, and we did not collect enough immunofluorescence pictures for statistics. For the reviewer's concern regarding the CheO-sfGFP fusion, we have two points to clarify: (1) CheO is a cytoplasmic protein with no transmembrane domain, and the fusion gene copy of cheO-sfGFP was introduced into the C. jejuni chromosome to replace the native cheO copy, thus the expression of cheO-sfGFP is under native promoter to avoid over-expression; (2) We performed soft agar assay with cheO-sfGFP mutant, which showed similar level of swarming motility as the wild type (Fig. S3), suggesting that the function of CheO is not altered due to the GFP fusion.
3. The authors show that CheO is dispersed in the quadruple tlp9, aer1, aer2, tlp6 mutant, but not in any single chemoreceptor mutants. This work needs follow up. First, are CheA, CheW, CheV dispersed in these quadruple mutants, e.g. is it specific? Line 241-245 and Fig. S5, not all the single tlp/aer mutants are shown, e.g. no single tlp9, aer1 or aer2 deletion. Also, the mutant labeled by deltatlp9aer1aer2 is a bit unclear--should it be labeled as deltatlp9 deltaaer1 deltaaer2.
Response: First, we want to clarify that Aer1 and Aer2 in C. jejuni are only composed of PAS domain alone, different from the Aer receptor in E. coli (Please see Fig. S1 for domain organization of all chemoreceptors in C. jejuni). Previous studies suggested that Aer1 (CetB) and Aer2 (CetC) interact with Tlp9 (CetA) to form complex for signal transduction (references 39, 40, 41). Specifically, Tlp9 and Aer1 are co-transcribed and is a bipartite energy taxis system in C. jejuni (ref 39, 40). The aer2 is at the upstream of tlp9/aer1 genes and aer2 (cetC) functionally complements a Δaer1 (cetB) mutant, required for energy taxis in concert with Aer1 (CetB) (ref 41). Thus, we decided to make a triple Δtlp9Δaer1Δaer2 mutant to test whether the complex affects CheO localization. In Δtlp9Δaer1Δaer2 mutant, CheO still localizes at the cell poles, so we did not go further to make individual mutant for tlp9, aer1, and aer2. Other than these 3 genes and two pseudogenes (tlp3 and tlp5), we made single gene mutant for all the other tlps in C. jejuni to test their effect on CheO localization (Please see Fig. S5).
We have renamed Δtlp6Δtlp9/aer1/aer2 as Δtlp6Δtlp9Δaer1Δaer2 in the text and Figures.
Our additional experiments described below in answers to question 5 is an indirect evidence to show that CheV, CheA, CheW proteins still form array at the cell poles.
4. It would be great to include double and triple tlp6, tlp9, aer1, aer2 mutants to really understand which receptor(s) are driving the localization.
Response: The answers to the above question #3 clarified why we only generated Δtlp6, Δtlp9Δaer1Δaer2, and Δtlp6Δtlp9Δaer1Δaer2, but not more double or triple mutants on these four genes. We understand that the reviewer is interested in how each or more Tlps affect the CheO localization, and we have added an additional experiment in the revised manuscript to check whether CheO interact with any Tlps. As shown in Fig. S7, CheO does not interact with any receptors tested here. We added this new result in the Results: "Since the polar localization of CheO is lost in quadruple knockout mutant Δtlp9Δaer1Δaer2Δtlp6, we further examined whether CheO interacts with any of the chemoreceptors through BTH assay. As shown in Fig.  S7, CheO does not interact with any of the energy taxis receptor Tlp6, Tlp9, Aer1, Aer2, and other Tlps (Tlp1/4/7/8/10)" (Lines 286-289).
Based on our current results, more than one Tlps concertedly affect the polar localization of CheO. For future thorough examination on this question, more tlp combinational mutants will be needed, not limited to tlp6, tlp9, aer1 and aer2.

What is the soft agar migration of the quadruple receptor mutant?
Response: We performed additional experiments to answer this question, please see the results in new Fig. S6. Basically, the motility ring of Δtlp6Δtlp9Δaer1Δaer2 mutant was slightly reduced in the soft agar plate, but its swimming velocity and reversal frequency were not significantly different from the wild-type in the single cell tracking analysis. These results suggested that the quadruple mutant retains chemosensory array structure, i. e. the CheV, CheA, CheW proteins still form array at the cell poles. A copy of fliM L99F was also introduced into C. jejuni ΔcheVA mutant, which did not restore its motility, suggesting that the FliM L99F is not a suppressor of the motility phenotype of ΔcheVA mutant.

Part III -Minor Issues: Editorial and Data Presentation Modifications
Please use this section for editorial suggestions as well as relatively minor modifications of existing data that would enhance clarity.
Response: Please see lines 152-156: "Sequence analyses of CheO did not identify any known domain or motif that provided a clue for its cellular function. In addition, this protein does not contain any obvious signal peptide or transmembrane region, most likely to be a cytoplasmic protein. In the genome of C. jejuni 81-176, cheO is a stand-alone gene with 86bp intergenic region to its upstream guaA (CJJ81176_1264) and 240bp to its downstream purD (CJJ81176_1266)." 10. Line 166-170, single-cell tracking was used to test the role of CheO in chemotaxis (Figs. 1F and G). Is the swimming behavior tracked without adding chemotaxis ligands? The conditions and cognate ligands used in this assay should be indicated.
Response: The swimming behavior was measured in liquid BHI medium without adding additional ligands. We have added detailed information in the Methods. Figure 1C lacks the statistical analysis. The caption about Fig 1F and G is incomplete.

11.
Response: The statistical analysis between the ΔcheO mutant and the WT strain was added in Fig. 1C. The caption of Fig. 1F and 1G was modified as "(F) Single-cell tracking of C. jejuni wild-type and mutant strains in BHI medium with the microscope slides and coverslips sealed in microaerobic atmosphere. Three individual experiments were performed for each strain and 20-30 cells were tracked in each experiment. The images shown here are representatives for each strain to compare their swimming behavior. (G) Quantification of reversal rates of C. jejuni wild-type and mutant strains. The reversal rates were calculated as the number of directional switches per minute per cell and data are shown as mean ± SEM. Differences between the mutant and the wild-type were statistically analyzed by Student's t-test".
12. Line 187-189 ( Fig. 2A). Is the fluorescence observation done under microaerobic conditions? The description is not clear.
Response: The fluorescence observation was done under microaerobic condition. The sentence were modified as "Fluorescence microscopy showed that CheO-sfGFP localizes exclusively to the poles of C. jejuni under microaerobic condition ( Fig. 2A)."  13. Figure 3A. The backgrounds of fluorescence pictures is abnormal. Did authors adjust the image contrast or lightness a lot to make the backgrounds vague? Please make this clear in the figure legend and methods.
Response: The microscope slide appeared contaminated with background fluorescence and we replaced a clearer microscope image for Fig. 3A.
14. Figure 3C and D need statistical analysis.
Response: references were added, please see line 252 now.
16. Figure 4 shows that CheO interacts with both the HPT and linker region of CheA. Do authors have any idea about how does that work? For this and the other interactions, it might be useful to determine the cognate interacting region of CheO with other proteins/domains? Response: Thanks for the suggestion. CheO does not contain any identifiable domain or motif to guide reasonable truncation design to study its interaction region. Future structural information may provide details to guide this study.
17. Fig. 4 might work better moved to come before Fig. 3.
Response: Thanks for the suggestion. With the newly added experiments suggested by reviewer #2, we think it is better to remain the order of Fig 3 & 4 and corresponding text.
18. Line 292: The authors should make it more clear that the G297T refers to nucleotides (not AA, as I initially thought).
Response: We have rephrased this sentence as "We replaced the wild-type fliM gene with fliM G297T (nucleotide mutation) in the chromosome of ΔcheO mutant to test whether FliM L99F (the corresponding amino acid mutation) can alleviate the spreading defect in the absence of CheO." 19. The discussion part is pretty simplified. It looks like that CheO can interact with many proteins including CheA, CheZ, FliM, and FliY. What do the authors believe is the reason that CheO can interact so many proteins with different functions and domains?
Response: The chemosensory system (F3 class) in C. jejuni contains several auxiliary components that are not found in the E. coli paradigm, such as CheV, ChePep, CheX, ChePQ. In addition, core components also display distinct domain organizations, such as fusion of REC domain in CheA and lack of REC domain in CheB. The identified CheO here is another element that is not found in other model organisms used for chemotaxis studies. We did not identify any known domain or motif in this protein, so it is not a classical component in known chemotaxis signaling pathways. Besides, the CheO homologs in species of Campylobacterota are generally not conserved (Fig. S4) and have a higher pI value (~8.5). We have collaborated with a structural lab to crystalize CheO but not successful yet since this protein is not a compact structure in solution. We suspect that the sequence flexibility (low complexity region) might enable more interactions of this protein with the others.
We have moderately expanded our discussion, please see the last three paragraphs in Discussion. 20. Methods: There are multiple plate names that are not clearly defined, e.g. karmali agar, and sometimes use of TSA, Brucella Broth, Blood (not clearly defined). Please check the media used and make sure each is clearly described.
Response: In the revised Methods, we have clarified the media used in each experiment. Figure S3 caption about 'strains expressing cheO-sfGfp or sfGFP-cheY' is unclear, wildtype strain or cognate mutants?

21.
Response: We have modified the Fig. S3 caption as "Soft agar motility assay of C. jejuni wild-type, C. jejuni strains expressing cheO-sfGFP or sfGFP-cheY at the native cheO or cheY loci without other mutations, with ΔmotA mutant as a negative control."

Figure Files:
While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org.

Data Requirements:
Please note that, as a condition of publication, PLOS' data policy requires that you make available all data used to draw the conclusions outlined in your manuscript. Data must be deposited in an appropriate repository, included within the body of the manuscript, or uploaded as supporting information. This includes all numerical values that were used to generate graphs, histograms etc.. For an example see here on PLOS Biology: http://www.plosbiology.org/article/info%3Adoi%2F10.1371%2Fjournal.pbio.1001908#s5.

Reproducibility:
To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols