2

Peer Review Information

Journal: Nature Microbiology Manuscript Title: Phase-variable capsular polysaccharides and lipoproteins modify bacteriophage susceptibility in Bacteroides thetaiotaomicron Corresponding author name(s): Eric Martens

Reviewer Comments & Decisions:

Dear Eric, Thank you for your patience over the past few weeks. Your manuscript entitled "Multiple phase-variable mechanisms, including capsular surface polysaccharides, modify bacteriophage susceptibility in Bacteroides thetaiotaomicron" has now been seen by 4 referees, whose comments are attached. Although they find your work of some potential interest, they have raised concerns which will need to be addressed before we can consider publication of the work in Nature Microbiology.

Should further experimental data allow you to address these criticisms, we would be happy to look at a revised manuscript.

Please note that referee #3 was delayed in submitting their review due to personal reasons, and so the review provided is incomplete, but I have attached it here for your information.

In particular, referee #1 asks for an analysis of other Bacteroidetes and gut phyla to determine how variable/conserved these cps loci are and referee #2 asks for clarification of the broader implications of the work. Referees #1 and #2 also have questions about some of the observed growth kinetics. Referee #4 asks for additional analysis of either the phage or bacterial host genomes to provide insight into the observed phenotypes. Referees #2 and #3 ask for the overall manuscript to be edited to ensure that it is succinct.

We are committed to providing a fair and constructive peer-review process. Please do not hesitate to contact us if there are specific requests from the reviewers that you believe are technically impossible or unlikely to yield a meaningful outcome.

If revising your manuscript:

Decision Letter, initial version:

3

* Include a “Response to referees” document detailing, point-by-point, how you addressed each referee comment. If no action was taken to address a point, you must provide a compelling argument. This response will be sent back to the referees along with the revised manuscript.

* If you have not done so already we suggest that you begin to revise your manuscript so that it conforms to our Article format instructions at http://www.nature.com/nmicrobiol/info/final-submission. Refer also to any guidelines provided in this letter.

* Include a revised version of any required reporting checklist. It will be available to referees (and, potentially, statisticians) to aid in their evaluation if the manuscript goes back for peer review. A revised checklist is essential for

2

re-review of the paper. Please use the link below to submit a revised paper:

[REDACTED]

<strong>Note:</strong> This url links to your confidential homepage and associated information about manuscripts you may have submitted or be reviewing for us. If you wish to forward this e-mail to co-authors, please delete this link to your homepage first.

Nature Microbiology is committed to improving transparency in authorship. As part of our efforts in this direction, we are now requesting that all authors identified as ‘corresponding author’ on published papers create and link their Open Researcher and Contributor Identifier (ORCID) with their account on the Manuscript Tracking System (MTS), prior to acceptance. This applies to primary research papers only. ORCID helps the scientific community achieve unambiguous attribution of all scholarly contributions. You can create and link your ORCID from the home page of the MTS by clicking on ‘Modify my Springer Nature account’. For more information please visit please visit <a href="http://www.springernature.com/orcid">www.springernature.com/orcid</a>.

If you wish to submit a suitably revised manuscript we would hope to receive it within 6 months. If you cannot send it within this time, please let us know. We will be happy to consider your revision, even if a similar study has been accepted for publication at Nature Microbiology or published elsewhere (up to a maximum of 6 months). Should your manuscript be substantially delayed without notifying us in advance and your article is eventually published, the received date would be that of the revised, not the original, version.

In the meantime we hope that you find our referees' comments helpful.

***************************************************** Reviewer Expertise:

Referee #1: Bacteroides, glycobiology, genomics Referee #2: glycobiology, bacteria-phage interactions, capsular polysaccharides Referee #3: bacteria-phage interactions, phage genomics Referee #4: Bacteroides, gut microbiome, metagenomics

3

Reviewer Comments:

Reviewer #1 (Remarks to the Author):

The paper by Porter et al describes an extensive and systematic study looking at the role of the multiple phase variable capsular polysaccharide loci (cps) encoded by the human gut symbiont Bacteroides thetaiotaomicron in susceptibility to phage infection (phage isolated from sewage samples so likely GI derived). The main novel finding seems to be that that there are distinct roles for each of the different CPS in mediating phage susceptibility (both inhibiting and assisting phage predation). In addition the paper identifies alternate phase variable mechanisms mediating phage susceptibility including an S-layer protein. The paper is very well written and presented making it clear and easy to read. The work is carefully and thoroughly carried out and the conclusions are justified based on the data. While the idea that CPS can modulate phage infection is not new, the scale of the study to systematically dissect the role of the different CPS in phage susceptibility is unprecedented and provides important insights into the role of phage in shaping the bacterial community of the human gut. I think the study described will be of significant interest to the wide range of researchers interested in the gut microbiota, as well more generally to virologists and bacteriologists working on microbes from other niches. As the authors point out their work will also have implications for the development of phage based therapies aimed at manipulating the gut microbiota. Comments: 1. A clear description of how the 8 cps loci were previously identified would be helpful. In the results (p5) it refers to Ref 25 (Hickey et al; line 92), but as far as I can see this paper doesn’t mention cps loci at all. An SI figure showing the 8 cps loci in B. thetaiotaomicron would also be helpful. 2. It would be good if the authors could provide some more insight into how conserved or variable the numbers and types of cps loci are in other Bacteroidetes from the gut (and indeed other environments) – the study mentions B. fragilis and B. intestinalis, but does not give any insights into how many cps loci these spp encode. Is Bacteroides thetaiotaomicron unique in having so many? And are they conserved in other spp? Do Bacteriodes spp have more cps loci than other Bacteroidetes? And how do they compare to the other major phyla of the gut (Firmicutes, Actinobacteria, Proteobacteria? Is anything known about this?). This doesn’t have to be exhaustive and could be an SI Fig, but it would be good to have some broader insights into cps loci distribution and conservation in the microbiota. 3. Fig 3. The cps7 strain seems to grow to lower OD to all the others with heat killed ARB25. Is this reproducible and can the authors speculate on why this might be? 4. Line 197 It states that strains expressing the single cps 4, 5, 7 and 8 show ‘….initial growth inhibition…’ but from Fig 3 it appears that initially all of these strains grow exactly the same as the WT (up to ~6h). Please rephrase or explain. Also in the case of cps 4 and 5 one of the live phage replicates does not display any defect at all. Why is this? And was the experiment only carried out once? 5. With regard to the explanation for the resumption of grow after initial inhibition (ie. Fig 3, mainly acapsular and cps1) I see how this is likely due to outgrowth of a resistant population, but how do you explain the other phenotype? ie. initial normal growth and then growth stagnation? (for cps4,5,7,8). Apologies if this is already described somewhere but I could easily see where. 6. Line 223. Confusing. Do you just mean there was a 3-4h extended lag before the culture grew as WT? 7. Fig 4B and elsewhere. When looking at cps locus expression by qPCR you use a single gene as a marker for expression of the whole locus as described in Methods and the primers given in Table S4. Can you explain how the gene for each cps was chosen? Is it simply the first in the locus? And how do you know this is representative of expression of the whole locus under all conditions?

4

8. Line 237-241. Did you do any stats to see if the difference in expression of cps 1, 3 and 4 between replicates 1-3 with heat killed is significant? It looks it but there are no error bars.

Minor points: Line 385. Some references for the potential mechanisms e.g. De Sordi et al. Cell Host and Microbe 25; 210-218 2019. Line 952 – spelling ‘tittered’. Fig S3. Y axis label on panel A missing. Line 956 - Missing ‘g’ from gene. Line 964 - CPS written twice. Line 971 – loci missing i. Ref 40 missing.

Reviewer #2 (Remarks to the Author):

The manuscript “Multiple phase-variable mechanisms, including capsular polysaccharides, modify bacteriophage susceptibility in Bacteroides thetaiotaomicron” from Porter et al. investigates an extensive collection of B. theta phages, isolated by the authors. The CPS of B. theta protects the bacteria from phages and is controlled by eight phase-variable genes. In this study, the panel of 71 phages was screened against various cps mutants and genotypes permissive of infection were identified. Examination of resistant bacteria not only identified changes in CPS, but also identified changes in surface lipoproteins and S-layer proteins in an acapsular strain that are under the control of independent phase-variable promoters. Overall, the results presented in this manuscript are important for demonstrating the role of CPS variability as a common defense against phage infections for both commensals and pathogens. More importantly, the results demonstrate that in addition to the ability to vary the expression of 8 different CPS structures, B. theta can alter at least 17 independent cell surface structures bringing to question the viability for phage therapy against bacteria in general.

1. The manuscript can be strengthened by providing brief rational upfront as to why this system was chosen and its broader importance including a more in-depth discussion on why phage-mediated modulation of the microbiota is of interest and why B. theta is a great model. 2. Please provide a comment as to why the dramatic loss of cps4 did not result in increased cps2 as one would expect based on the results in the deletion mutant shown in Figure 2G. 3. In line 225, the authors described variable growth kinetics, but this determination was made based on three colonies, where one appeared to show a difference in the corresponding Figure 4. Given this, the support does not appear to be significant for their claim. Was the expression done for the other two? 4. It is unclear why the SJCO1 phage was chosen for the S-layer protein experiment and the authors did discuss their thoughts as to why this showed an opposite phenotype when compared to the ARB25 phage the manuscript focused on.

Minor points: The writing could be more concise. I noted some wording that was better suited for the discussion that was in the results section. Similarly, some methodology could be removed from the figure legends. Also lab phrases such as strains were struck from freezer stocks; plaques were picked into buffer; etc should be fixed. The sentence on line 524 also appears to be missing a word.

Missing labels on Figures 5 and 6

5

Refers to Figures 4 and S7A on line 318, but I believe this should be Figure 5C

Consider rewriting lines 253-256 for clarity

Missing axis label in Figure S3 Lines 976-980 in the methods refer to the incorrect table in the supplementary material

Throughout the manuscript “µl” should be corrected to “µL”

Spelling/grammar mistakes: Line 350 – “cold” to “could”, Line 841 – should be separated into two sentences, Line 952 – “tittered” to “titered”

Words cut out of BT1041 box in Table S3

Line 639 – remember to include project number once received

Are the structures for the 8 CPS known? And if so, can any correlations be made about phage infectivity based on structure?

Is there a straightforward method to confirm that the S-layer is really on and off with promoter changes?

The authors should be commended for replicating all the host range assays independently by two leading authors in Figure S2.

What was the rationale for using cultures grown to OD600=0.6-0.7 for RNAseq experiments? Do you know the approximate MOI at that point for each strain tested? MOI is only mentioned in a few places such as lines 234, 880, etc. And what are considered high and low MOIs in Figure S5?

What is the rationale for selecting cps7 expression as the reference? Is this because this CPS shows the lowest expression levels based on the inability to purify this CPS in an earlier paper?

Figure 6S legend - six loc should be six loci

Is it more appropriate to call the phage preparations throughout the manuscript as heat-killed and intact rather than heat-killed and live?

Reviewer #3 (Remarks to the Author):

In this intriguing paper the authors have looked at the interaction between bacteriophages and the human symbiont Bacteroides thetaiotaomicron. They have focused their studies on looking at the relationships between phase variation in the capsular polysaccharides of this bacterium and the ability of specific bacteriophage is to infect it. The manuscript is well written, the data interesting and I am confident that it will be of interest to a wide readership.

That said, I felt that the paper would benefit from being edited as it is very long.

The authors created a bank of Bacteroides thetaiotaomicron and used these to isolate phages, and test other phages

6

on. They reveal a vast amount of complexity within the capsular polysaccharides that results in different infection outcomes. They showed that the deletion of CPS caused several phages to be unable infect the bacteria. Furthermore, they showed that when they deleted the polysaccharide altogether, although this was fairly effective in preventing bacterial infection, the mutant bacteria that survived expressed high levels of a family of phase variable lipoproteins. When it is constitutively expressed, one of these lipoproteins then contributes to phage resistance. The first data reported by the authors is the isolation of bacteriophages on eight different versions of capsular mutants, and one version where the capsule was entirely deleted. This culminated in 71 bacteriophage is being isolated. They then tested each phage on the 10 variants - i.e. the mutants, the strain that lacked CPS, and the wild- type. Some of the data presented here is quite confusing, it isn't overly apparent which strains the different phages within the different branches were isolated. I think this could be written more clearly. The authors conclude that the data shows that CPS are important in dictating host range. I was not entirely clear about why figure 1 have lots of phage plaques on the right-hand side of the cluster analysis. It would seem odd if all of those phages beside the specific image had exactly that morphology? Interestingly, many phages do infect acapsular mutants but they also infect mutant variants that do have at least one of the capsule proteins. It could be therefore that some versions of the capsule are permissive and therefore allow infection whereas others cause it to be blocked. To try and work out what was happening the authors did further detailed studies on six bacteriophages. They made a series of capsule mutants whereby based on the host range data in the preceding section, they deleted the permissive combinations of CPS from strains and retested these 6 phages. The authors seemed surprised by some of their results in the section as often the double mutants did not display the phenotypes they expected. They conclude that phase variation may compensate for some of the mutations and either allow or not allow successful infection. They showed that the cps4 variant appears to be particularly important to phage infection. The authors also did experiments where they incubated the phages with CPS that had been removed from specific mutant strain lines to see if specific combinations of capsule might inhibit binding on the wild-type strain. This appeared to not be the case. The authors showed that much of the phage resistance seen in this set of experiments was transient and thus likely to have a genetic basis. The authors then have a section with a heading, Phage resistant wild-type B. thetaiotaomicron populations exhibit altered cps locus expression. Here, as you might expect, they showed that phages select for non-persmissive capsules.

Reviewer #4 (Remarks to the Author): The manuscript includes an impressive test of a large collection of phages against a panel of B. thetaioatomicron mutants, firmly establishing a link between capsular polysaccharides and phage infectivity. Further experiments are done with one phage especially, but the conclusions that can be drawn are limited, in part due to how complex the phase variable cps system in is. Overall, the manuscript would benefit from additional experiments to either broaden the scope or provide additional mechanistic insights into an individual phage.

Sequences of the phage genomes would be the most meaningful possible addition, as correlations between phage genome content and infectivity could lead to many predictions about the functions of phage genes. Though the authors acknowledge this as a future direction, the lack of this data makes the current manuscript shallow in its investigation of all but 2 of the phages. If inclusion of the phage genomes is not possible, expanding the investigation to additional strains of B. thetaiotaomicron (ideally those with already sequenced genomes) would be an alternative way to strengthen the paper. This would help assess how broadly reactive these phages are across strains and could also provide additional insight into the cps links made. For example, if phage infectivity can be predicted by the cps loci present on the genome of a given strain, this would be a powerful addition that validates the results of Figure 1.

7

As certain phage appear to require certain cps as receptors, this would be interesting to test.

The difference in results between plaque assay in Fig 1 and liquid growth in Fig 3 is interesting but not fully addressed. It would be helpful to repeat these experiments with at least the SJC01 phage which is later used in Fig 6 to see how it compares, since it had a similar profile in Fig 1.

The purified CPS having no effect on infectivity in Fig S3A is somewhat surprising. Did the authors try varying the relative amounts of bacteria, phage, and purified CPS? Perhaps the amount of CPS was simply not sufficient to block the amount of phage at the given MOI. If possible to purify, can the S-layer protein be used to modulate infectivity?

How does the S-layer affect phage infectivity in wildtype B. theta? Presumably in nature, B. thetaiotaomicron always expresses at least one cps, so the relevance of effects on the acapsular strain is uncertain.

minor:

Figure 1: suggest highlighting all the phage that are followed up on, in addition to ARB25, to aid the reader when referring back to this Figure.

line 350: cold should be could

Reviewer #1 (Remarks to the Author):

The paper by Porter et al describes an extensive and systematic study looking at the role of the multiple phase variable capsular polysaccharide loci (cps) encoded by the human gut symbiont Bacteroides thetaiotaomicron in susceptibility to phage infection (phage isolated from sewage samples so likely GI derived).

The main novel finding seems to be that that there are distinct roles for each of the different CPS in mediating phage susceptibility (both inhibiting and assisting phage predation). In addition, the paper identifies alternate phase variable mechanisms mediating phage susceptibility including an S-layer protein. The paper is very well written and presented making it clear and easy to read. The work is carefully and thoroughly carried out and the conclusions are justified based on the data.

While the idea that CPS can modulate phage infection is not new, the scale of the study to systematically dissect the role of the different CPS in phage susceptibility is unprecedented and provides important insights into the role of phage in shaping the bacterial community of the human gut.

I think the study described will be of significant interest to the wide range of researchers interested in the gut microbiota, as well as more generally to virologists and bacteriologists working on microbes from other niches. As the authors point out their work will also have implications for the development of phage-based therapies aimed at manipulating the gut microbiota.

Author response: We thank the reviewer for enthusiastic comments about our work.

Author Rebuttal to Initial comments

8

Comments: 1. A clear description of how the 8 cps loci were previously identified would be helpful. In the results (p5) it refers to Ref 25 (Hickey et al; line 92), but as far as I can see this paper doesn’t mention cps loci at all. An SI figure showing the 8 cps loci in B. thetaiotaomicron would also behelpful.

Author response: We agree that the Hickey et al. ref. causes some confusion in the way that it is cited here to reference construction of the single capsule strains. This study mostly focused on the role of B. theta in causing colitis and one monoclonal antibody that was used as a tool was found to recognize CPS3. Thus, we published this panel of single CPS expressing strains in that study for the first time, but the data was relegated to the supplement and not a focus of that study.

Historically, the 8 CPS loci were annotated in this strain’s genome (the second Bacteroides species to be fully sequenced) in a seminal paper from Jeff Gordon’s lab in 2003. We have now used the original Gordon reference after mention of the “…8 different, phase-variable CPS…”. In place of the Hickey et al. reference we also added Porter et al. 2017 in the following sentence, which utilized the single CPS expressing strains in much greater depth. As suggested, we have also added a detailed SI figure of the 8 loci to compare their conserved and variable functions (Figure S1B) and this also follows a new analysis (Figure S1A) to quantify the number of annotated CPS in 54 different human gut Bacteroidetes genomes. This latter addition was made in response to the next point below.

2. It would be good if the authors could provide some more insight into how conserved or variable the numbers and types of cps loci are in other Bacteroidetes from the gut (and indeed other environments) – the study mentions B. fragilis and B. intestinalis, but does not give any insights into how many cps loci these spp encode. Is Bacteroides thetaiotaomicron unique in having so many? And are they conserved in other spp? Do Bacteroides spp have more cps loci than other Bacteroidetes? And how do they compare to the other major phyla of the gut (Firmicutes, Actinobacteria, Proteobacteria? Is anything known about this?).

9

3 This doesn’t have to be exhaustive and could be an SI Fig, but it would be good to have some broader insights into cps loci distribution and conservation in the microbiota.

Author response: The association between gut Bacteroidetes possessing large repertoires of CPS loci has indeed been known for some time, albeit the previous systematic analysis was done by the Comstock lab at Harvard before the currently large availability of sequenced reference strains. As a consequence of the limited genomic data ca. 2007, the previous study (PMID:17993536) did not analyze most gut Bacteroidetes. Nevertheless, with only a few strains, this study reported increased numbers of CPS loci in human gut Bacteroidetes compared to oral isolates. To address the reviewer’s question and update the field given the large number of new Bacteroidetes species that have been identified and/or sequenced in the past 10 years, we have performed a new analysis using the founding or type strains of 53 different sequenced species of Bacteroidetes isolated from human stool or other GI-derived samples. This analysis shows that there are between 0 and 13 CPS in these isolates, with only the Prevotella not possessing CPS based on the search criteria use (which are based on Bacteroides locus architecture and may miss more divergent variations that have different gene content). Among the remaining species that possess CPS, the range is from 2-13 (mean = 4), which is similar to the range and average reported by Coyne and Comstock for a smaller subset. Some lineages like the Alistipes appear to generally harbor fewer CPS-coding loci and could therefore rely more heavily on alternative phage-evasion strategies. While we considered a parallel systematic analysis of S-layer coding genes in these same genomes, which might begin to reveal each species’ reliance on these two strategies, we discovered that even the 8 S-layer proteins that are present in B. theta share relatively little homology to one another. This obfuscates searching by homology alone and will require development of an alternative approach to be used in a future mechanistic study of these functions that is already underway.

The newly generated data regarding human gut Bacteroidetes CPS loci is presented in the revised manuscript as Fig. S1A and presents the number of annotated CPS, using criteria based on functional gene content, in a phylogenetic context. We have added the following text in the first paragraph of results:

“The genomes of human gut Bacteroidetes frequently encode multiple CPS (Coyne et al, 2008), a phenomenon that we explored in a phylogenetic context in currently available genomes of 53 different human gut species. The type strains of all but the Prevotella species searched encoded between 2-13 CPS (mean = 4), suggesting that the ability to produce multiple CPS is typical and some lineages have undergone substantial expansion of these surface structures (Figure S1A).”

3. Fig 3. The cps7 strain seems to grow to lower OD to all the others with heat killed ARB25. Is this reproducible and can the authors speculate on why this might be?

Author response: There is indeed a consistent defect associated with this strain. We previously showed that this strain, which only possesses the CPS7 genes, does not express this locus as highly as other cps loci, raising the conclusion that the capsule is defective and the strain is therefore functionally similar to the acapsular strain, which also grows to about the same level (~0.1 OD600 units less than the others). This phenotype is likely to be attributable to these strains’ tendency to settle out of suspension without a

10

polysaccharide capsule coating their surface, a phenomenon that has also been reported for acapsular B. fragilis. This phenomenon may cause an absorbance reading artifact and therefore the slightly lower readings. Because we don’t explore any CP7-related phenotypes in additional detail, the small effect size of this growth phenomenon and the already tight space constraints in the text, we chose not to discuss this further in the revised manuscript.

4. Line 197 It states that strains expressing the single cps 4, 5, 7 and 8 show ‘….initial growth inhibition…’ but from Fig 3 it appears that initially all of these strains grow exactly the same as the WT (up to ~6h). Please rephrase or explain.

11

Author response: Thank you for pointing out this wording discrepancy. We have changed that text to more accurately reflect the growth phenomena that are observed and indicated by the reviewer. The revised text now reads:

“While susceptible strains initiated growth similarly to uninfected cultures and later showed loss of culture

density, they subsequently displayed either growth stagnation at an intermediate culture density (cps4, cps7, cps8) or resumption of growth (wild-type, acapsular, cps1, cps5) that approached the density achieved by uninfected controls.”

Also, in the case of cps 4 and 5 one of the live phage replicates does not display any defect at all. Why is this? And was the experiment only carried out once?

Author response: We have indeed replicated the CPS4 and CPS5 growths in 20 separate biological replicates (20 separate colonies) and have not observed any additional incidence of non-perturbed growth (see new Fig. S6B). Since this appears to either be a rare phenomenon (~1/23 or less) or a technical error in the original experiment, we have replaced the data for CPS4 and CPS5 strains with more representative n=3 growth response data. Although, given the many additional phase-variable evasion mechanisms uncovered in this work, it is still possible that 1/3 colonies in the original experiment were in a sufficiently resistant pre-existing state that they did not show growth inhibition by ARB25.

5. With regard to the explanation for the resumption of grow after initial inhibition (i.e. Fig 3, mainly acapsular and cps1) I see how this is likely due to outgrowth of a resistant population, but how do you explain the other phenotype? ie. initial normal growth and then growth stagnation? (for cps4,7,8). Apologies if this is already described somewhere but I could not easily seewhere.

Author response: We agree that this phenomenon is quite interesting, especially as it only reproducibly arose in certain CPS strain variants (i.e., 4,7,8). One possibility that needs to be explored with future work using new genetic variants of S-layer protein and CPS expressing strains is that there is cross-regulation between these mechanisms, such that the CPS4,7,8 strains phase-vary back to susceptible states more frequently and therefore fail to gain enough resistance in the population to increase density of viable cells. There is, for example, some evidence in the original B. theta BT1927 S-layer study that an unlinked, but nearby, surface protein (BT1954) involved in vitamin B12 acquisition (doi: 10.1016/j.chom.2013.12.007) negatively regulates BT1927 expression. There is also work from the Comstock lab (doi: 10.1073/pnas.1005039107) investigating B. fragilis CPS cross-regulation through negative regulation of anti-termination factors. Thus, it is possible that the 4 single CPS strains noted exhibit different global regulation of the various phase-variable resistance mechanisms described here. This is being actively explored in current work. Note that in follow experiments discussed below, we were able to discern that the age of a colony that is picked for liquid culture experiments influences the effectiveness of the BT1927 S- layer evasion mechanism (even though the colonies are grown the same way in liquid medium immediately before use). Even when treating cps4 strain colonies in a way that promotes the highest S-layer resistance (3 days on BPRM plate), we still see the above described intermediate phenotype with 20 separate colonies that only express CPS4, suggesting that colony age is not the only factor in mediating the ability of B. theta variants to partially or fully avoid ARB25 killing.

12

Because we initially failed to describe this phenomenon in the original submission, we have added a brief description of the points noted above, along with the need for future work to explore these potential regulatory phenomena in the context of all of the points noted above:

“The observation of growth stagnation after initial loss of bacterial density suggests that a more complex equilibrium is achieved between phage and resistant bacteria that prohibits either from becoming dominant. This behavior was reproducible with 20 separate cultures of the cps4 strain (Figure S6B).”

6. Line 223. Confusing. Do you just mean there was a 3-4h extended lag before the culture grew as WT?

13

5 Author response: There was a perceptible increase in growth before a subsequent decline and regrowth, so this is not extended lag. Since we described this growth phenomenon associated with WT and other strains after ARB25 infection, we have clarified the wording in this section to read as:

“Cultures treated with a high MOI (≈1 pfu/cfu) displayed similar growth kinetics as observed previously, with an apparently resistant population emerging after 3-4 hours (Figure 4A).”

7. Fig 4B and elsewhere. When looking at cps locus expression by qPCR you use a single gene as a marker for expression of the whole locus as described in Methods and the primers given in Table S4. Can you explain how the gene for each cps was chosen? Is it simply the first in the locus? And how do you know this is representative of expression of the whole locus under allconditions?

Author response: We previously examined expression of cps genes by both qPCR and global methods such as GeneChips and RNAseq (both of the latter helped established the length of the locus that is actually expressed and the strength of expression for each gene). Within a locus, there tends to be little variation in expression intensity, although some genes within a locus share homology with other loci. The primer sets used were extensively employed in a previous study (doi: 10.1016/j.chom.2017.08.020) and were validated for specificity using the other single CPS-expressing strains as controls. Moreover, the qPCR based data using these primer sets (Fig. 4) is concordant with RNAseq-based expression analysis, providing further validation of the selected amplicons. We have added a very brief description of the previously published testing/validation approach to the appropriate methods section:

“The primers used were selected to target a gene specific to each cps locus and were previously validated against the other strains that lack the target cps locus for specificity {Porter, 2017 #4}.”

8. Line 237-241. Did you do any stats to see if the difference in expression of cps 1, 3 and 4 between replicates 1-3 with heat killed is significant? It looks like it but there are no error bars.

Author response: The Drichlet regression analysis that we used requires multiple replicates in each treatment (live vs. heat-killed), but we cannot compare the 3 single replicates in the heat killed treatment without at least performing many more individual replicates to see if variant patterns emerge. Since each of the biological replicates is initiated with a single colony of the particular B. theta strain, it is quite likely that individual replicates start with a different set of phase variable states that may subsequently alter their response to phage infection. Indeed, the three individual replicates used for the live ARB25-infected wild- type experiment appeared to have different starting states, which is part of the data interpretation presented/discussed in this section. However, we did not run additional replicates to determine if the different patterns observed in the two different wild-type cultures are random expression patterns for these CPS, or consistent sub-patterns of CPS expression that would emerge more strongly with a large number of replicates.

Minor points: Line 385. Some references for the potential mechanisms e.g. De Sordi et al. Cell Host and Microbe 25; 210- 218 2019.

14

Author response: Thank you for this reference. We have added it as a citation at the end of the first sentence of this paragraph (previously line 385).

Line 952 – spelling ‘tittered’.

Author response: We have corrected this typo.

Fig S3. Y axis label on panel A missing.

Author response: We have corrected this error in the figure (now Fig. S5).

15

6

Line 956 - Missing ‘g’ from gene. Author response: We have corrected this typo.

Line 964 - CPS written twice.

Author response: We have corrected this typo.

Line 971 – loci missing i.

Author response: We have corrected this typo.

Ref 40 missing.

Author response: Thank you for catching this. We have now added the complete citation for ref. 40 (Gupta et al.) to the references list.

Reviewer #2 (Remarks to the Author):

The manuscript “Multiple phase-variable mechanisms, including capsular polysaccharides, modify bacteriophage susceptibility in Bacteroides thetaiotaomicron” from Porter et al. investigates an extensive collection of B. theta phages, isolated by the authors. The CPS of B. theta protects the bacteria from phages and is controlled by eight phase-variable genes. In this study, the panel of 71 phages was screened against various cps mutants and genotypes permissive of infection were identified. Examination of resistant bacteria not only identified changes in CPS, but also identified changes in surface lipoproteins and S-layer proteins in an acapsular strain that are under the control of independent phase-variable promoters. Overall, the results presented in this manuscript are important for demonstrating the role of CPS variability as a common defense against phage infections for both commensals and pathogens. More importantly, the results demonstrate that in addition to the ability to vary the expression of 8 different CPS structures, B. theta can alter at least 17 independent cell surface structures bringing to question the viability for phage therapy against bacteria in general.

Author response: We thank the reviewer for enthusiastic comments about our work.

1. The manuscript can be strengthened by providing brief rational upfront as to why this system was chosen and its broader importance including a more in-depth discussion on why phage-mediated modulation of the microbiota is of interest and why B. theta is a great model.

Author response: We attempted to introduce the importance of this system in the first paragraph of the Introduction by underscoring the importance of community resilience in the face of environmental and

16

infectious perturbations. We’ve edited this first paragraph and the subsequent introductory paragraphs to better emphasize the utility of B. theta compared to other commensal Bacteroides as both a model symbiont and a representative member of the human gut microbiome. Given the original length of the manuscript and the fact that we added substantially more experiments, we did not introduce a lengthy discussion of why phage-mediated manipulation of the microbiome is of interest. However, we do conclude the Abstract, Introduction and Discussion with statements about the importance of understanding these bacterial host phage evasion mechanisms in order for phage therapy to be developed and work efficiently.

2. Please provide a comment as to why the dramatic loss of cps4 did not result in increased cps2 as one would expect based on the results in the deletion mutant shown in Figure 2G.

17

7 Author response: A line reference was not provided for this comment, but we’re assuming this comment is aimed at the difference between Fig. 2G (deletion of cps4) and Fig. 4 (selection against expression of cps4). Deletion of the cps4 locus clearly results in dominant expression of CPS2 (Fig. 2G). The response by wild-type B. theta after ARB25 infection in Fig. 4 is dominant expression of CPS3. However, these experiments use two different genetic backgrounds. When cps4 is deleted, it’s possible that loss of the associated upxY/Z regulators results in dominant CPS2 expression. In contrast, selection for the resistant sub-population from a wild-type culture that may be genetically programmed to express a different capsule repertoire could result in a different response. In Fig. 3B, the wild-type population expresses much more CPS3 than CPS2, which likely explains why short-term selection with ARB25 results in a population that dominantly expresses CPS3 instead of CPS2. This could be a result of the short culture time in the experiment shown in Fig. 4 and some stochastic selection for the non-permissive bacteria present when phage are introduced. Support for this comes from the RNAseq experiment shown right after this in the paper (Fig. 5A) in which wild-type B. theta that survives after ARB25 express similar levels of CPS2 and CPS3. As suggested by the reviewer, we’ve commented on this by adding the following text:

“While reduction in CPS4 expression did not correlate with increased CPS2 as observed with a ∆cps4 strain, the high abundance of CPS3 expressing bacteria in the initial culture may have enabled ARB25 to most rapidly select for this population in the hours post-infection.”

3. In line 225, the authors described variable growth kinetics, but this determination was made based on three colonies, where one appeared to show a difference in the corresponding Figure 4. Given this, the support does not appear to be significant for their claim. Was the expression done for the other two?

Author response: Yes, we measured cps gene expression for all three and show this data in Fig. S5. While there are only 3 replicates, there is a correlation between higher expression of non-permissive CPS3 in the single colony-derived culture that grew faster.

4. It is unclear why the SJCO1 phage was chosen for the S-layer protein experiment and the authors did not discuss their thoughts as to why this showed an opposite phenotype when compared to the ARB25 phage the manuscript focused on.

Author response: We originally chose this phage because it exhibits a very similar infection profile as ARB25 when tested against the single CPS-expressing strains in Fig. 1. We have re-tested ARB25 and SJC01 against the BT1927 locked on and off acapsular strains (in a different lab, too) and curiously found different results. Upon re-test in the UM lab, the BT1927 locked on strain was not inhibited by either ARB25 or SJC01, while the BT1927 locked off strain was more severely inhibited by both. Because these bacterial stocks were passaged before using in new experiments at the UM lab, they may have undergone global changes in other phase-variable factors that influence their infection sensitivity. We therefore carefully repeated these experiments a minimum of 12 separate times (all from separate colonies; 4 colonies on at least 3 separate days in both labs). Fortuitously, we discovered that the age of the colony (1, 2 or 3 days old) that is picked into liquid medium to initiate the growth experiment influences the amount of BT1927- mediated resistance, with 3-day old colonies being most resistant (now shown in Figure S13). Thus, we repeated the assays shown in Fig. 6 with 3-day old colonies of each locked strain and 4 different branch 1

18

or 2 phage, showing that locking BT1927 on promotes increased resistance to all 4 phage. The rationale for choosing these phages was to pick several phages from the two branches in which phage infect acapsular B. theta (Branch 3 mostly does not). The new results suggest that even a single S-layer BT1927 could be broadly protective against multiple phage, which changes the conclusions as previously written. The new results have been described in the section discussing Fig. 6 and the “colony age” phenomenon shown as Figure S13 to document this interesting observation.

Minor points: The writing could be more concise. I noted some wording that was better suited for the discussion that was in the results section. Similarly, some methodology could be removed from the figure legends. Also, lab

19

8 phrases such as strains were struck from freezer stocks; plaques were picked into buffer; etc should be fixed. The sentence on line 524 also appears to be missing a word.

Author response: We have conducted a thorough revision of the manuscript and eliminated all of the lab jargon we could identify and also made the writing more concise (although the former was somewhat offset by inclusion and description of new experimental data).

Missing labels on Figures 5 and 6

Author response: Thank you for noticing this. We have corrected these errors.

Refers to Figures 4 and S7A on line 318, but I believe this should be Figure 5C

Author response: We have revised to cite Fig. 5D and S10A (current MS numbering).

Consider rewriting lines 253-256 for clarity

Author response: We have revised these lines as suggested. They now read: “However, ARB25-infected acapsular B. thetaiotaomicron was still able to grow significantly after initial lysis by the phage and most bacteria isolated after phage infection had regained susceptibility (Figure 3, Table S2), suggesting the emergence of a transiently phage-resistant subpopulation in the absence of CPS.”

Missing axis label in Figure S3

Author response: We have added the appropriate axis label to what is now Fig. S4.

Lines 976-980 in the methods refer to the incorrect table in the supplementary material

Author response: We have updated to refer to Table S6 instead of S3.

Throughout the manuscript “µl” should be corrected to “µL” Author response: We have checked this throughout and changed as requested, along with “mL”.

Spelling/grammar mistakes: Line 350 – “cold” to “could”, Line 841 – should be separated into two sentences, Line 952 – “tittered” to “titered”

Author response: We have made these three edits as requested.

Words cut out of BT1041 box in Table S3

Author response: We have fixed this omission, which was caused by a formatting error in the Table cells.

20

Line 639 – remember to include project number once received Author response: Updated NCBI project numbers have been received from NCBI and provided. The project is set to be publicly available on May 1, 2020 or immediately upon paper publication.

Are the structures for the 8 CPS known? And if so, can any correlations be made about phage infectivity based on structure?

Author response: The glycosidic linkage structures of all 8 CPS are still unknown. We previously reported the sugar composition of these CPS (doi: 10.1016/j.chom.2017.08.020; as Figure S3 in that study).

21

9 However, aside from sugars like GlcNAc and mannose being common to the ARB25 non-permissive CPS2 and 3 there are no apparent correlations with resistance/susceptibility profiles and these sugars are also present in permissive CPS.

Is there a straightforward method to confirm that the S-layer is really on and off with promoter changes?

Author response: The same locked “on” strain was previously visualized by EM and shown to produce a different surface architecture. Although, this visualization was only in a wild-type (CPS+) background and not in the acapsular background that is new to this study. In new data that we have added, the phenotype of the BT1927 “on” strain is much more resistant to both ARB25, SJC01, ARB19 and SJC03, while the “off” strain is more susceptible to all 4. This result suggests that there is a functional outcome of the genetic events that lock the promoter on or off. Interestingly, we also describe a new phenomenon through which the age of a BT1927 locked on colony before being cultured into liquid media for phage infection assay influences BT1927-mediated resistance. We are actively exploring the basis of this, but show the initial data in Fig. S13.

The authors should be commended for replicating all the host range assays independently by two leading authors in Figure S2.

Author response: Thank you for this supportive comment.

What was the rationale for using cultures grown to OD600=0.6-0.7 for RNAseq experiments? Do you know the approximate MOI at that point for each strain tested? MOI is only mentioned in a few places such as lines 234, 880, etc. And what are considered high and low MOIs in Figure S5?

Author response: We harvested cells at a post phage-lysed OD for which we could be confident that the bacteria are still actively growing and not in stationary phase. The harvest point was based on previous growth curves with the same conditions that were allowed to proceed into stationary phase to establish growth maximum before stationary phase arrest occurred. Harvesting cells slightly later during active growth provides more biomass for RNA, but we would not anticipate that the number of phage present would change much after the initial burst(s) that reduces OD since the bacteria that grow after will be mostly resistant (note that we actually measured ARB25 titers in post-infected WT and acapsular B. theta to determine if the phage were still active and the expected results of no increase after initial burst were indeed observed; see Fig. S5B). Arguably, the most critical consideration in this RNAseq comparison is collecting cells in the live phage and heat-killed controls at similar growth ODs to ensure that other aspects of bacterial physiology are not influenced by comparing transcriptomes with large differences in growth phase. As such, in additional RNAseq, we have harvested cells at similar points in active growth.

What is the rationale for selecting cps7 expression as the reference? Is this because this CPS shows the lowest expression levels based on the inability to purify this CPS in an earlier paper?

Author response: Yes, cps7 shows low levels of expression, but this is from previous qPCR-based transcriptional analysis of cps7 genes (plus, an inability to recover much surface material which suggests a

22

defect). The Dirichlet regression requires one condition to be held as a reference and returns no significance value for this reference, so we chose to use CPS7 for this reference. We have clarified this in the corresponding methods section.

Figure 6S legend - six loc should be six loci

Author response: We have made the indicated revision.

Is it more appropriate to call the phage preparations throughout the manuscript as heat-killed and intact rather than heat-killed and live?

23

10 Author response: While we agree the phage are not actually “alive”, we think the adjective “live” will be the most intuitive for the general reader to indicate that the phage are “active” and “infectious”. The latter two terms are appropriate alternatives, as is the suggested “intact”, although all come with some potential confusion of their own, so we chose to retain “live” throughout.

Reviewer #3 (Remarks to the Author):

In this intriguing paper the authors have looked at the interaction between bacteriophages and the human symbiont Bacteroides thetaiotaomicron. They have focused their studies on looking at the relationships between phase variation in the capsular polysaccharides of this bacterium and the ability of specific bacteriophage is to infect it. The manuscript is well written, the data interesting and I am confident that it will be of interest to a wide readership.

That said, I felt that the paper would benefit from being edited as it is very long.

Author response: We thank the reviewer for the supportive comments and have thoroughly edited the manuscript to reduce length and increase clarity of the previously existing sections. It is important to note that we have also added substantial new data, which requires additional text for description. Nevertheless, we have tried to describe all of the new data additions in a clear and concise way.

The authors created a bank of Bacteroides thetaiotaomicron and used these to isolate phages, and test other phages on. They reveal a vast amount of complexity within the capsular polysaccharides that results in different infection outcomes.They showed that the deletion of CPS caused several phages to be unable infect the bacteria. Furthermore, they showed that when they deleted the polysaccharide altogether, although this was fairly effective in preventing bacterial infection, the mutant bacteria that survived expressed high levels of a family of phase variable lipoproteins. When it is constitutively expressed, one of these lipoproteins then contributes to phage resistance.

The first data reported by the authors is the isolation of bacteriophages on eight different versions of capsular mutants, and one version where the capsule was entirely deleted. This culminated in 71 bacteriophage is being isolated. They then tested each phage on the 10 variants - i.e. the mutants, the strain that lacked CPS, and the wild-type. Some of the data presented here is quite confusing, it isn't overly apparent which strains the different phages within the different branches were isolated. I think this could be written more clearly. The authors conclude that the data shows that CPS are important in dictating host range.

I was not entirely clear about why figure 1 have lots of phage plaques on the right-hand side of the cluster analysis. It would seem odd if all of those phages beside the specific image had exactly that morphology?

Author response: To clarify which strain is the preferred host for each phage, we have now added this information as a parenthetical next to each phage name in Figure 1. We agree that it would be unexpected for the various phage shown to have the same plaque morphology. This was indeed the intended point of

24

showing several phage from each group that show similar bacterial infection profiles, but vary in morphology. This is obviously not a precise way to determine that the phages with similar infection profiles are not simply identical, but our recently submitted study on phage genomes (biorxiv: doi: https://doi.org/10.1101/2020.03.04.977157) provides additional evidence of genetic diversity for a subset of phages, and we feel that the plaque visualization provides some extra evidence of difference for the remaining phage that were not sequenced. In the revised figure, we took care to juxtapose pictures of phage that had similar infection profiles and were photographed on the same host bacterium for more direct comparison.

Interestingly, many phages do infect acapsular mutants but they also infect mutant variants that do have at least one of the capsule proteins. It could be therefore that some versions of the capsule are permissive and

25

therefore allow infection whereas others cause it to be blocked. To try and work out what was happening the authors did further detailed studies on six bacteriophages. They made a series of capsule mutants

whereby based on the host range data in the preceding section, they deleted the permissive combinations of CPS from strains and retested these 6 phages. The authors seemed surprised by some of their results in the section as often the double mutants did not display the phenotypes they expected. They conclude that phase variation may compensate for some of the mutations and either allow or not allow successful infection. They showed that the cps4 variant appears to be particularly important to phage infection. The authors also did experiments where they incubated the phages with CPS that had been removed from specific mutant strain lines to see if specific combinations of capsule might inhibit binding on the wild-type strain. This appeared to not be the case. The authors showed that much of the phage resistance seen in this set of experiments was transient and thus likely to have a genetic basis.

The authors then have a section with a heading, Phage resistant wild-type B. thetaiotaomicron populations exhibit altered cps locus expression. Here, as you might expect, they showed that phages select for non- permissive capsules.

Author response: We thank the reviewer for comments on the paper and understand that the review could not be completed, so please note that there were no additional questions associated with this section.

Reviewer #4 (Remarks to the Author):

The manuscript includes an impressive test of a large collection of phages against a panel of B. thetaioatomicron mutants, firmly establishing a link between capsular polysaccharides and phage infectivity. Further experiments are done with one phage especially, but the conclusions that can be drawn are limited, in part due to how complex the phase variable cps system in is. Overall, the manuscript would benefit from additional experiments to either broaden the scope or provide additional mechanistic insights into an individual phage.

Author response: We thank the reviewer for the positive comments and have added numerous new experiments (some described above, but others detailed below) to broaden what is shown in this manuscript. Due to the large amount of data that would be added, we did not describe genome sequences of any phage in this manuscript, but have recently submitted a paper describing the genomes of 27 phage from this study and other isolates, revealing substantial diversity within three main phage families. This is currently available on biorxiv: doi: https://doi.org/10.1101/2020.03.04.977157.

Sequences of the phage genomes would be the most meaningful possible addition, as correlations between phage genome content and infectivity could lead to many predictions about the functions of phage genes. Though the authors acknowledge this as a future direction, the lack of this data makes the current manuscript shallow in its investigation of all but 2 of the phages. If inclusion of the phage genomes is not possible, expanding the investigation to additional strains of B. thetaiotaomicron (ideally those with already sequenced genomes) would be an alternative way to strengthen the paper. This would help assess how broadly reactive these phages are across strains and could also provide additional insight into the cps links

26

made. For example, if phage infectivity can be predicted by the cps loci present on the genome of a given strain, this would be a powerful addition that validates the results of Figure 1. As certain phage appear to require certain cps as receptors, this would be interesting to test.

Author response: To address the concerns about the work not being broad enough with respect to detailed investigations of all but two phage, we have added an in vivo experiment demonstrating that wild- type and acapsular B. thetaiotaomicron strains can co-exist in relative equilibrium for over 2 months with ARB25, which targets both strains, and the corresponding bacterial responses involved in this extended survival within the gut. We have added experiments using another genetically manipulable, sequenced B. theta strain (7330) that has a partially overlapping repertoire of cps loci. We have also added new RNAseq

27

12 for ARB25 infected CPS1-expressing B. theta WT, which shows that cells that only express a single, permissive capsule can still evade this phage and do so by upregulating S-layers. This experiment importantly also shows that S-layers can be turned on in a strain with forced expression of at least one CPS. In addition, we have added RNAseq of WT and acapsular B. theta after infection with another phage (SJC01), which shows that, similarly to ARB25, WT B. theta evades infection with a shift toward non- permissive CPS. In parallel, the acapsular strain evades infection by upregulating S-layer proteins and ARB25 and SJC01 each select for downregulation of different sets of PULs involved in nutrient acquisition. Importantly, in the latter experiment, we identified a new phase-variable restriction enzyme system that is activated in the SJC01 infected population. We have also performed additional tests with acapsular B. theta strains in which the BT1927 S-layer gene in the locked “on” and “off” positions, revealing that this protein increases resistance to 4 different phage (all that were tested). In tandem with the new in vivo co-existence data, which reveal another phase-variable lipoprotein (see Figure S12 for both restriction and lipoprotein systems), we feel that this latter result brings the narrative of the study to a more complete conclusion.

The difference in results between plaque assay in Fig 1 and liquid growth in Fig 3 is interesting but not fully addressed. It would be helpful to repeat these experiments with at least the SJC01 phage which is later used in Fig 6 to see how it compares, since it had a similar profile in Fig 1.

Author response: As suggested, we have repeated this with SJC01 and also see that CPS (like CPS4) that are non-permissive in solid medium assays (e.g., cps4) show susceptibility in liquid culture assays.

The purified CPS having no effect on infectivity in Fig S3A is somewhat surprising. Did the authors try varying the relative amounts of bacteria, phage, and purified CPS? Perhaps the amount of CPS was simply not sufficient to block the amount of phage at the given MOI. If possible to purify, can the S-layer protein be used to modulate infectivity?

Author response: We agree that experiments to investigate the trans-acting nature of CPS could be done in more depth (additional phage, additional capsules) and would be especially interesting if purified S-layer proteins were used (we have not attempted this yet, but it may require extra effort given the larger size, >1,000aa, of these proteins and their apparent nature to aggregate or form a patterned lattice). Nevertheless, the experiment that was conducted in the original version of the study used a large amount of extracted CPS2 mixed with ARB25 prior to infecting B. theta in a soft agar assay in which target bacterial cells are plentiful. While we don’t know the size of individual CPS2 polymers, a liberal estimate of 1,000 sugars per molecule would provide a molecular weight of 180,000 Da, which at 1mg/ml would be 9x1013

CPS glycan molecules. Even if the prep were only 10% pure (9x1012 glycan molecules), this was incubated with 103 ARB25 based on pfu/ml titer, prior to adding to bacteria. So, the CPS2 molecules were very likely to be many orders of magnitude more abundant than individual phage even if the phage have multiple receptors per phage. In this scenario, if there was strong trans inhibition of non-permissive CPS2 on ARB25 infection, we expect that we would have observed it. We have briefly mentioned this excess of CPS to phage and also added the above numerical calculation in the corresponding methods section. Note that to reduce length, we have only briefly mentioned this in the main text and moved the discussion to the Fig. S5 legend.

2

How does the S-layer affect phage infectivity in wildtype B. theta? Presumably in nature, B. thetaiotaomicron always expresses at least one cps, so the relevance of effects on the acapsular strain is uncertain.

Author response: This is a great point. Given the antagonistic regulation that occurs between Bacteroides CPS (documented in B. fragilis and implied here in our cps4 knockouts) it is certainly plausible that expression of a capsule could interfere with expression of S-layer or that these two surface features are mutually exclusive of one another. While testing all of the individual CPS x S-layer combinations would be beyond the scope of the current study, we performed an additional RNAseq experiment using the ARB25- permissive CPS1-expressing strain infected with ARB25. While this strain is forced to express a capsule that is not protective, it is still able to survive infection (Figure 3). We observed that the dominant

13 transcriptional response after ARB25 survival was upregulation of a subset of the previously identified S- layer proteins, suggesting that S-layers can be expressed in a strain forced to express just CPS1. Moreover, we used germfree mice to perform new in vivo co-colonization experiments with B. theta WT and acapsular strains in which the respective bacterial strain was added first and then ARB25 introduced 7 days later. In both cases, ARB25 and its respective host bacterium co-exist for up to 72 days together in relative equilibrium (new Figure 6E). While RNAseq of the acapsular strain shows a predominant profile characterized by increased expression of S- layer proteins (and two additional functions that we validate to indeed be phase-variable: a restriction endonuclease and another predicted surface lipoprotein!), the RNAseq of WT B. theta does indeed indicate that the cell population can express both CPS and S-layer. However, WT B. theta in vivo expresses mostly permissive CPS (5, 6, 7, 8) with weak upregulation of non- permissive CPS2. This implies that other aspects of the in vivo environment may be dominant in selecting for these permissive CPS and the cell must compensate by adapting to express an alternative mechanisms of phage resistance such as S-layers. It is still not possible to decipher if individual cells are expressing both types of resistance mechanisms or if there is exclusion at the single cell level. Resolving this will require future experiments. Our overall interpretation of these data is that CPS variation in WT B. theta is usually a high-frequency adaptive mechanism, which may be why it’s observed first in WT cells grown for a short time in vitro. For example, ~40% of the population before ARB25 infection expresses CPS3, which is likely why it is selected for so quickly in post- infected cells. In contrast, the BT1927 S-layer was originally estimated to be expressed by ~1:1000 cells, so selection for cells expressing this mechanism may require more time to emerge than the relatively short in vitro batch culture experiments that we performed (even if it is a more effective evasion strategy). In contrast, long-term coexistence in vivo allows populations to emerge that express both non-permissive CPS and S-layers because there is more time for the low-frequency of phase- variation S-layer responses to emerge. We provide brief comments on the cps1 result from in vitro culture and the above interpretation of the wild-type strain in vivo in the corresponding new sections.

minor:

2

Figure 1: suggest highlighting all the phage that are followed up on, in addition to ARB25, to aid the reader when referring back to this Figure.

Author response: As suggested, we have highlighted the phage in Figure 1 that are followed up on in later experiments.

line 350: cold should be could

Author response: We have made the indicated correction

Dear Eric,

Thank you for your patience while your manuscript "Multiple phase-variable mechanisms, including capsular surface polysaccharides, modify bacteriophage susceptibility in Bacteroides thetaiotaomicron" was under peer review at Nature Microbiology. It has now been seen by our referees, and in the light of their advice I am delighted to say that we can in principle offer to publish it. First, however, we would like you to revise your paper to address the points made by the reviewers, and to ensure that it is in Nature Microbiology format.

The referees’ have no remaining comments. Editorially, we will need you to make some changes so that the paper complies with our Guide to Authors at http://www.nature.com/nmicrobiol/info/gta.

Nature Microbiology offers a transparent peer review option for new original research manuscripts submitted from 1st December 2019. We encourage increased transparency in peer review by publishing the reviewer comments, author rebuttal letters and editorial decision letters if the authors agree. Such peer review material is made available as a supplementary peer review file. <b>Please state in the cover letter ‘I wish to participate in transparent peer review’ if you want to opt in, or ‘I do not wish to participate in transparent peer review’ if you don’t.</b> Failure to state your preference will result in delays in accepting your manuscript for publication. Please note: we allow redactions to authors’ rebuttal and reviewer comments in the interest of confidentiality. If you are concerned about the release of confidential data, please let us know specifically what information you would like to have removed. Please note that we cannot incorporate redactions for any other reasons. Reviewer names will be published in the peer review files if the reviewer signed the comments to authors, or if reviewers explicitly agree to release their name. For more information, please refer to our <a href="https://www.nature.com/documents/nr-transparent- peer-review.pdf" target="new">FAQ page</a>.

Specific points: In particular, while checking through the manuscript and associated files, we noticed the following specific points which we will need you to address:

Decision Letter, first revision:

3

1) Extended data. Per journal guidelines, we use "Extended Data". Please see below for additional information on how to format and refer to Extended Data. Please convert 10 of your supplementary figures into Extended Data figures. These should be the most pertinent figures. The remaining figures can be provided in a Supplementary Information pdf document, or if possible, can be merged into other figures.

2) Supplementary tables. Please provide Supplementary tables in a single Excel file with one table per tab and a title at the top of each.

3) Priority claims. Per journal guidelines, we recommend that you avoid the use of terms like ‘new’, ‘novel’ and other priority claims throughout the text in order to avoid any perception of grandstanding and so that the reader can focus on the significance, rather than the novelty, of the findings. Therefore, please revise lines 294, 300,439,1176,1178,1190,1228,1229,1233 and 1243(‘new’ or 'newly'), and any other relevant sections accordingly.

4) ORCID. As mentioned previously, we ask all corresponding authors to provide their unique ORCID identifiers at the time of final submission. Please see below for additional information on how to sign up to ORCID and link your account to the one in our manuscript tracking system.

5) Title. Unfortunately we cannot use punctuation in our titles due to journal style. We would therefore suggest the following: "Multiple phase-variable mechanisms modify bacteriophage susceptibility in Bacteroides thetaiotaomicron" OR "Phase-variable capsular polysaccharides and lipoproteins modify bacteriophage susceptibility in Bacteroides thetaiotaomicron". However, these are just suggestions so please feel free to provide an alternative.

6) Competing interests. Per journal guidelines, please add a competing interest statement to the text.

7) Author Contributions. Please provide a more detailed and specific author contributions statement. A good example can be found at the end of the following article http://www.nature.com/nature/journal/v532/n7599/full/nature17433.html

8) Scale bars. Per style guidelines, all microscopy figures should include a scale bars, which should be defined in the legend. Please add scale bars to figure S4c and define the length of the bar in the legend.

9) Figures. While carefully checking the figures, we noted a few things that need to be revised so that the findings are clear to the reader:

Figure 1 - please explain the tree and branches on the left hand side of the figure in the legend. Figure2a-f - please clarify in the legend what the dashed black line represents. Figure 5a-b - please clarify what the dashed lines represent in the legend. Figure 5d - please provide additional labelling or description in the legend of the colours used, what the green shapes and numbers represent, what the black and red dashed lines represent and all arrows. Figure S4a - please provide a key for the colours used. Figure S6 - please clarify what the green line represents in the key. Figure S10b - please clarify the colour coding used in the legend e.g. what do the salmon pink and yellow represent.

4

Figure S12 - please clarify in the legend what the black dashed lines represent and what the grey/black circles represent.

10) Source data. Please provide source data for all in vitro and in vivo experiments. Please see below for more information on how to provide this information.

11) Full length gels. Per journal guidelines, we require that full length versions of all gels or blots are included in SI. These should be raw, uncropped, unmanipulated versions of all gels, ideally showing the original molecular weight markers and using text boxes to indicate which sections of the full gels were cropped to generate the figures shown. Therefore, please add a new SI figure with full length blots for the gels shown in figures 5e and S12e.

12) Animal experiments. Please add a statement in the methods section describing, for animal experiments, how was sample size chosen, and whether/how randomization and blinding were performed. Please also note the sex and age of the animals used, and whether the work was IRB approved.

13) Data availability. Please include a data availability section at the end of the methods - see below for additional details on how to format this section. Note that this section should include all accession codes for all sequencing and transcriptomics data deposited to databases. Please note that all accession codes must be live by the time of publication of the piece.

14) Code Availability. Please provide a Code Availability statement and deposit any custom code to GitHub.

15) Reporting checklist. Note that a final version of the reporting checklist will be published with your manuscript. Therefore, please revise this document according to the instructions found in the annotated PDF attached to this message.

Below are points arising from the reporting checklist that will need to be addressed in the manuscript:

Reporting Summary – Extended Comments

General comments: • Please state in the legends how many times each experiment was repeated independently with similar results. This is needed for all experiments but is particularly important wherever representative experiments are shown. If space in the legends is limiting, this information can be included in a section titled “Statistics and Reproducibility”.

• Wherever statistics have been derived (e.g., error bars, box plots, statistical significance), the legend needs to provide and define the n number (i.e., the sample size used to derive statistics) as a precise value (not a range), using the wording “n=X biologically independent samples/animals/independent experiments,” etc. as applicable.

• All error bars need to be defined in the legends (e.g., SD, SEM) together with a measure of centre (e.g., mean, median), and should be accompanied by their precise n number defined as noted above.

5

• For all bar graphs, the corresponding dot plot must be overlaid.

• The figure legends must indicate the statistical test used and if applicable, whether the test was one- or two-sided. A description of any assumptions or corrections such as tests of normality and adjustment for multiple comparisons must also be included.

• Test results (e.g., p-values, q-values) should be given as exact values whenever possible and appropriate, and confidence intervals noted.

• For null hypothesis testing, please indicate the test statistic (e.g., F, t, r) with confidence intervals, effect sizes, degrees of freedom and P values noted.

Specific comments:

Figure 2 a-f • The p value has been provided as a range; please provide the precise value if possible and appropriate. Figure 2 g • Please state the statistical test used to test the significance of the regression. If applicable, indicate whether the test was one- or two-sided. • The p value has been provided as a range; please provide the precise value if possible and appropriate. • Please overlay each data point as dot plots to indicate the distribution of the data. Figure 2 h • Please indicate in the figure legend how many times these experiments were repeated independently with similar results. Figure 3 • Please indicate in the figure legend how many times these experiments were repeated independently with similar results. Figure 4 a • Please indicate in the figure legend how many times these experiments were repeated independently with similar results. Figure 4 b • Please state the statistical test used to test the significance of the regression. If applicable, indicate whether the test was one- or two-sided. • The p value has been provided as a range; please provide the precise value if possible and appropriate. • Please overlay each data point as dot plots to indicate the distribution of the data. Figure 5 a • Please state the statistical test used. If applicable, indicate whether the test was one- or two-sided and describe any adjustments for multiple comparisons. Figure 5 b • Please clarify the sample size (n) used to derive the statistics. Figure 5 d • Please provide the p-value cutoff used to define significance. Figure 5 e • Please indicate in the figure legend how many times these experiments were repeated independently

6

with similar results. Figure 6 a-d • Please include the sample size as an exact value, not a range. • Please define the centre values and error bars in the figure legend. Figure 6 e • Please provide and define the sample size (n) in the figure legend as described in the General comments above. • Please define the centre values and error bars in the figure legend. • Please indicate in the figure legends whether the statistical test was one-sided or two-sided. • The p value has been provided as a range; please provide the precise value if possible and appropriate. Figure S2 a, b • Please indicate in the figure legend how many times these experiments were repeated independently with similar results. Figure S3 • Please indicate in the figure legend how many times these experiments were repeated independently with similar results, as an exact number and not a range. Figure S4 b • Please provide and define the sample size (n) in the figure legend as described in the General comments above. • Please define the measure of centre (e.g., median or mean) in the figure legend. • Please indicate in the figure legends whether the statistical test was one-sided or two-sided. • The p value has been provided as a range; please provide the precise value if possible and appropriate. • Please provide the precise p value where “ns” is noted. Figure S4 c • Please indicate in the figure legend how many times these experiments were repeated independently with similar results. Figure S4 d • Please indicate in the figure legend how many times these experiments were repeated independently with similar results, as an exact value. Figure S5 a • Please indicate in the figure legends whether the statistical test was one-sided or two-sided. • Please provide the precise p values if possible and appropriate. • Please overlay each data point as dot plots to indicate the distribution of the data. Figure S6 a, b • Please indicate in the figure legend how many times these experiments were repeated independently with similar results. Figure S7 b • Please overlay each data point as dot plots to indicate the distribution of the data. Figure S11 • Please indicate in the figure legends whether the statistical test was one-sided or two-sided. • Please define the measure of centre (e.g., median or mean) in the figure legend. Figure S12 a-c • Please provide and define the sample size (n) in the figure legend as described in the General comments above. • Please state the statistical test used. If applicable, indicate whether the test was one- or two-sided and describe any adjustments for multiple comparisons.

7

Figure S12 e • Please indicate in the figure legend how many times these experiments were repeated independently with similar results. Figure 12 f, g • Please provide and define the sample size (n) in the figure legend as described in the General comments above. • Please state the statistical test used. If applicable, indicate whether the test was one- or two-sided and describe any adjustments for multiple comparisons. Figure S13 a, b • Please clarify the sample size (n) used to derive the statistics. • Please define the centre values and error bars in the figure legend. Table S3 a-g • Please provide and define the sample size (n) in the figure legend as described in the General comments above. • Please state the statistical test used. If applicable, indicate whether the test was one- or two-sided and describe any adjustments for multiple comparisons.

General points:

Please read carefully through all of the following general formatting points when preparing the final version of your manuscript, as submitting the manuscript files in the required format will greatly speed the process to final acceptance of you work.

Our normal length limit for Articles is about 3,000 words. We have some flexibility, and can allow a revised manuscript at 3,500 words, but please consider this a firm upper limit. You could achieve some shortening by moving some details to the Methods section that should follow the main text (the length of the Methods section is unlimited and does not count towards the main text length).

Titles should give an idea of the main finding of the paper and ideally not exceed 90 characters (including spaces). We discourage the use of active verbs and do not allow punctuation.

The paper's summary paragraph (about 150-200 words; no references) should serve both as a general introduction to the topic, and as a brief, non-technical summary of your main results and their implications. It should start by outlining the background to your work (why the topic is important) and the main question you have addressed (the specific problem that initiated your research), before going on to describe your new observations, main conclusions and their general implications. Because we hope that scientists across the wider microbiology community will be interested in your work, the first paragraph should be as accessible as possible, explaining essential but specialised terms concisely. We suggest you show your summary paragraph to colleagues in other fields to uncover any problematic concepts.

We strongly support public availability of data. Please place the data used in your paper into a public data repository, if one exists, or alternatively, present the data as Source Data or Supplementary Information. If data can only be shared on request, please explain why in your Data Availability Statement, and also in the correspondence with your editor. For some data types, deposition in a public repository is mandatory - more information on our data deposition policies and available repositories can be found at https://www.nature.com/nature-research/editorial-policies/reporting-

8

standards#availability-of-data.

Please include a data availability statement as a separate section after Methods but before references, under the heading "Data Availability”. This section should inform readers about the availability of the data used to support the conclusions of your study. This information includes accession codes to public repositories (data banks for protein, DNA or RNA sequences, microarray, proteomics data etc…), references to source data published alongside the paper, unique identifiers such as URLs to data repository entries, or data set DOIs, and any other statement about data availability. At a minimum, you should include the following statement: “The data that support the findings of this study are available from the corresponding author upon request”, mentioning any restrictions on availability. If DOIs are provided, we also strongly encourage including these in the Reference list (authors, title, publisher (repository name), identifier, year). For more guidance on how to write this section please see: http://www.nature.com/authors/policies/data/data-availability-statements-data-citations.pdf

Please supply the figures as vector files - EPS, PDF, AI or postscript (PS) file formats (not raster or bitmap files), preferably generated with vector-graphics software (Adobe Illustrator for example). Try to ensure that all figures are non-flattened and fully editable. All images should be at least 300 dpi resolution (when figures are scaled to approximately the size that they are to be printed at) and in RGB colour format. Please do not submit Jpeg or flattened TIFF files. Please see also 'Guidelines for Electronic Submission of Figures' at the end of this letter for further detail. Please view http://www.nature.com/authors/editorial_policies/image.html for more detailed guidelines.

We will edit your figures/tables electronically so they conform to Nature Microbiology style. If necessary, we will re-size figures to fit single or double column width. If your figures contain several parts, the parts should be labelled lower case a, b, and so on, and form a neat rectangle when assembled.

Please check the PDF of the whole paper and figures (on our manuscript tracking system) VERY CAREFULLY when you submit the revised manuscript. This will be used as the 'reference copy' to make sure no details (such as Greek letters or symbols) have gone missing during file-transfer/conversion and re-drawing.

All Supplementary Information must be submitted in accordance with the instructions in the attached Inventory of Supporting Information, and should fit into one of three categories:

1. EXTENDED DATA: Extended Data are an integral part of the paper and only data that directly contribute to the main message should be presented. These figures will be integrated into the full-text HTML version of your paper and will be appended to the online PDF. There is a limit of 10 Extended Data figures, and each must be referred to in the main text. Each Extended Data figure should be of the same quality as the main figures, and should be supplied at a size that will allow both the figure and legend to be presented on a single legal-sized page. Each figure should be submitted as an individual .jpg, .tif or .eps file with a maximum size of 10 MB each. All Extended Data figure legends must be provided in the attached Inventory of Accessory Information, not in the figure files themselves.

2. SUPPLEMENTARY INFORMATION: Supplementary Information is material that is essential

9

background to the study but which is not practical to include in the printed version of the paper (for example, video files, large data sets and calculations). Each item must be referred to in the main manuscript and detailed in the attached Inventory of Accessory Information. Tables containing large data sets should be in Excel format, with the table number and title included within the body of the table. All textual information and any additional Supplementary Figures (which should be presented with the legends directly below each figure) should be provided as a single, combined PDF. Please note that we cannot accept resupplies of Supplementary Information after the paper has been formally accepted unless there has been a critical scientific error.

All Extended Data must be called you in your manuscript and cited as Extended Data 1, Extended Data 2, etc. Additional Supplementary Figures (if permitted) and other items are not required to be called out in your manuscript text, but should be numerically numbered, starting at one, as Supplementary Figure 1, not SI1, etc.

3. SOURCE DATA: We strongly encourage you to provide source data for your figures whenever possible. Full-length, unprocessed gels and blots must be provided as source data for any relevant figures, and should be provided as individual PDF files for each figure containing all supporting blots and/or gels with the linked figure noted directly in the file. Numerical source data that underlie graphs are required for in vivo experiments and strongly encouraged generally. They should be provided in Excel format, one file for each relevant figure, with the linked figure noted directly in the file. They should be clearly labelled such that individual experiments and/or animals are labelled (for example, across a time course if applicable). For imaging source data, we encourage deposition to a relevant repository, such as figshare (https://figshare.com/) or the Image Data Resource (https://idr.openmicroscopy.org).

Please include any references for the Methods at the end of the reference list. Any citations in the Supplemental Information will need inclusion in a separate SI reference list.

It is a condition of publication that you include a statement before the acknowledgements naming the author to whom correspondence and requests for materials should be addressed.

Finally, we require authors to include a statement of their individual contributions to the paper -- such as experimental work, project planning, data analysis, etc. -- immediately after the acknowledgements. The statement should be short, and refer to authors by their initials. For details please see the Authorship section of our joint Editorial policies at http://www.nature.com/authors/editorial_policies/authorship.html

We will not send your revised paper for further review if, in the editors' judgement, the referees' comments on the present version have been addressed. If the revised paper is in Nature Microbiology format, in accessible style and of appropriate length, we shall accept it for publication immediately.

Please resubmit electronically

* the final version of the text (not including the figures) in either Word or Latex.

* publication-quality figures. For more details, please refer to our Figure Guidelines, which is available here: https://mts-nmicrobiol.nature.com/letters/Figure_guidelines.pdf

10

* Extended Data & Supplementary Information, as instructed

* a point-by-point response to any issues raised by our referees and to any editorial suggestions.

* any suggestions for cover illustrations, which should be provided at high resolution as electronic files. Please note that such pictures should be selected more for their aesthetic appeal than for their scientific content. I am sure you will understand that we cannot make any promise as to whether any of your suggestions might be selected for the cover of Nature Microbiology.

Please use the following link to access your home page:

{REDACTED}

* This url links to your confidential homepage and associated information about manuscripts you may have submitted or be reviewing for us. If you wish to forward this e-mail to co-authors, please delete this link to your homepage first.

Please also send the following forms as a PDF by email to microbiology@nature.com.

* Please sign and return the <a href="http://www.nature.com/documents/snl-ltp.docx" target="_blank">Licence to Publish form</a> .

* Or, if the corresponding author is either a Crown government employee (including Great Britain and Northern Ireland, Canada and Australia), or a US Government employee, please sign and return the <a href="http://www.nature.com/documents/snl-ltp-crown.docx"_blank"> Licence to Publish form for Crown government employees</a>, or a <a href="http://www.nature.com/documents/snl-ltp- govus.docx" target="_blank"> Licence to Publish form for US government employees</a>.

* Should your Article contain any items (figures, tables, images, videos or text boxes) that are the same as (or are adaptations of) items that have previously been published elsewhere and/or are owned by a third party, please note that it is your responsibility to obtain the right to use such items and to give proper attribution to the copyright holder. This includes pictures taken by professional photographers and images downloaded from the internet. If you do not hold the copyright for any such item (in whole or part) that is included in your paper, please complete and return this <a href="http://www.nature.com/documents/thirdpartyrights-origres.doc">Third Party Rights Table</a>, and attach any grant of rights that you have collected.

For more information on our licence policy, please consult http://npg.nature.com/authors.

AUTHORSHIP

CONSORTIA -- For papers containing one or more consortia, all members of the consortium who contributed to the paper must be listed in the paper (i.e., print/online PDF). If necessary, individual authors can be listed in both the main author list and as a member of a consortium listed at the end of the paper. When submitting your revised manuscript via the online submission system, the consortium name should be entered as an author, together with the contact details of a nominated consortium representative. See https://www.nature.com/authors/policies/authorship.html for our authorship

11

policy and https://www.nature.com/documents/nr-consortia-formatting.pdf for further consortia formatting guidelines, which should be adhered to prior to acceptance.

<b>ORCID</b>

Nature Microbiology is committed to improving transparency in authorship. As part of our efforts in this direction, we are now requesting that all authors identified as ‘corresponding author’ create and link their Open Researcher and Contributor Identifier (ORCID) with their account on the Manuscript Tracking System (MTS) prior to acceptance. ORCID helps the scientific community achieve unambiguous attribution of all scholarly contributions. For more information please visit http://www.springernature.com/orcid

For all corresponding authors listed on the manuscript, please follow the instructions in the link below to link your ORCID to your account on our MTS before submitting the final version of the manuscript. If you do not yet have an ORCID you will be able to create one in minutes. https://www.springernature.com/gp/researchers/orcid/orcid-for-nature-research

IMPORTANT: All authors identified as ‘corresponding author’ on the manuscript must follow these instructions. Non-corresponding authors do not have to link their ORCIDs but are encouraged to do so. Please note that it will not be possible to add/modify ORCIDs at proof. Thus, if they wish to have their ORCID added to the paper they must also follow the above procedure prior to acceptance.

To support ORCID's aims, we only allow a single ORCID identifier to be attached to one account. If you have any issues attaching an ORCID identifier to your MTS account, please contact the <a href="http://platformsupport.nature.com/">Platform Support Helpdesk</a>.

Nature Research journals <a href="https://www.nature.com/nature-research/editorial- policies/reporting-standards#protocols" target="new">encourage authors to share their step-by-step experimental protocols</a> on a protocol sharing platform of their choice. Nature Research's Protocol Exchange is a free-to-use and open resource for protocols; protocols deposited in Protocol Exchange are citable and can be linked from the published article. More details can found at <a href="https://www.nature.com/protocolexchange/about" target="new">www.nature.com/protocolexchange/about</a>.

We hope to hear from you within two weeks; please let us know if the revision process is likely to take longer.

***************************************************** Reviewer Expertise:

Reviewer Comments:

12

Reviewer #1 (Remarks to the Author):

The authors have satisfactorily addressed all of my comments and that of the other reviewers (as far as I can see) and I see no further issues to be raised.

Reviewer #2 (Remarks to the Author):

The authors have made a substantial number of changes to the manuscript, both experimental and written, that not only satisfy all of my concerns, but also, I believe, those raised by the other reviewers. This is a significant study that is of broad interest.

Reviewer #4 (Remarks to the Author):

The authors have adequately addressed my critiques, making an earnest effort to perform additional experiments wherever feasible. The manuscript is well written and well presented, the experiments are fairly rigorous, the data support the conclusions, and overall the work will be impactful to field. I have no further comments.

Dear Eric, I am pleased to accept your Article "Phase-variable capsular polysaccharides and lipoproteins modify bacteriophage susceptibility in Bacteroides thetaiotaomicron" for publication in Nature Microbiology. Thank you for having chosen to submit your work to us and many congratulations. Before your manuscript is typeset, we will edit the text to ensure it is intelligible to our wide readership and conforms to house style. We look particularly carefully at the titles of all papers to ensure that they are relatively brief and understandable. The subeditor may send you the edited text for your approval. Once your manuscript is typeset you will receive a link to your electronic proof via email within 20 working days, with a request to make any corrections within 48 hours. If you have queries at any point during the production process then please contact the production team at rjsproduction@springernature.com. Once your paper has been scheduled for online publication, the Nature press office will be in touch to confirm the details. Acceptance of your manuscript is conditional on all authors' agreement with our publication policies (see www.nature.com/nmicrobiolate/authors/gta/content-type/index.html). In particular your manuscript must not be published elsewhere and there must be no announcement of the work to any media outlet until the publication date (the day on which it

Final Decision Letter:

13

is uploaded onto our website). The Author's Accepted Manuscript (the accepted version of the manuscript as submitted by the author) may only be posted 6 months after the paper is published, consistent with our <a href="http://www.nature.com/authors/policies/license.html">self-archiving embargo</a>. Please note that the Author’s Accepted Manuscript may not be released under a Creative Commons license. For Nature Research Terms of Reuse of archived manuscripts please see: <a href="http://www.nature.com/authors/policies/license.html#terms">http://www.nature.com/authors/policies/license.html#terms</a> If you have posted a preprint on any preprint server, please ensure that the preprint details are updated with a publication reference, including the DOI and a URL to the published version of the article on the journal website. An online order form for reprints of your paper is available at <a href="https://www.nature.com/reprints/author-reprints.html">https://www.nature.com/reprints/author-reprints.html</a>. All co-authors, authors' institutions and authors' funding agencies can order reprints using the form appropriate to their geographical region. We welcome the submission of potential cover material (including a short caption of around 40 words) related to your manuscript; suggestions should be sent to Nature Microbiology as electronic files (the image should be 300 dpi at 210 x 297 mm in either TIFF or JPEG format). Please note that such pictures should be selected more for their aesthetic appeal than for their scientific content, and that colour images work better than black and white or grayscale images. Please do not try to design a cover with the Nature Microbiology logo etc., and please do not submit composites of images related to your work. I am sure you will understand that we cannot make any promise as to whether any of your suggestions might be selected for the cover of the journal. You can now use a single sign-on for all your accounts, view the status of all your manuscript submissions and reviews, access usage statistics for your published articles and download a record of your refereeing activity for the Nature journals. To assist our authors in disseminating their research to the broader community, our SharedIt initiative provides you with a unique shareable link that will allow anyone (with or without a subscription) to read the published article. Recipients of the link with a subscription will also be able to download and print the PDF. As soon as your article is published, you will receive an automated email with your shareable link.

14