Interspecies interactions induce exploratory motility in Pseudomonas aeruginosa

Essential revisions:

The authors should do three things to improve the paper:

1) Better contextualize their study in light of the Oliveira paper (this can be done in the Introduction and Discussion, and also experimentally – see point 3).

In response to both points 1 and 3, we further investigated a role for Chp chemotaxis genes in the context of the studies performed by Oliveira paper. This study described a role for the CheY-like response regulator PilG, which regulates the ATPase necessary for TFP extension (PilB), to play a role in P. aeruginosa response to a chemotactic gradient. We therefore examined a role for PilG in P. aeruginosa response to S. aureus. Similar to the results by Oliveira, we found that a ΔpilG mutant was diminished in its capacity to respond to S. aureus (Figure 6). This was particularly striking when we compared the motility of ΔpilG to ΔpilJ, the predicted TFP chemoreceptor. Despite exhibiting low motility in traditional macroscopic assays, the ΔpilJ mutant was capable of responding to S. aureus, while the response of the pilG mutant was significantly diminished. These data suggest that PilG participates in modulating directional TFP-mediated motility independent of the predicted chemoreceptor, PilJ.

These data are discussed in detail in the following areas of the edited manuscript:

Introduction; Results: subsection “The CheY-like response regulator, PilG, modulates P. aeruginosa response to S. aureus”; Figure 6 (including Figure 6—figure supplements 1-3) and Discussion.

2) Expand understanding of the specificity of this interspecies interaction: test whether Pseudomonas responds similarly to other non-Staph species and/or

In response to this helpful suggestion, we examined P. aeruginosa response to other non-Staph species, including clinical isolates from CF patients and also a range of model organisms, in order to determine the specificity of P. aeruginosa interactions with other bacterial species. We examined Haemophilus influenzae, Burkholderia cepacia, and Achromobacter xylosoxidans isolates from CF patients and Bacillus subtilis, Escherichia coli, and Salmonella Typhimurium laboratory strains. While the magnitude of the effect varied among organisms and strains, we found that B. subtilis, E. coli, S. typhimurium, and B. cepacia were capable of inducing motility in P. aeruginosa, suggesting induction of P. aeruginosa motility is not specific to S. aureus or CF pathogens. Investigation into the differences between the signals generated by these organisms is underway and is the focus of a follow-up manuscript.

These data are discussed in detail in the following areas of the edited manuscript:

Figure 7; Results: subsection “P. aeruginosa responds to a broad range of model organisms and CF pathogens” and Discussion.

3) Determine whether the interactions observed here are operating through the Chp chemotaxis genes shown by Oliveira to be important in P. aeruginosa in exploratory motility.

Please see response for point #1, above.

Reviewer #1:

In this study, Limoli et al. used single-cell imaging to discover a novel behavioral interaction between Pseudomonas aeruginosa and Staphylococcus aureus. By imaging cells in co-culture, they found that P. aeruginosa alters its motility in response to a chemical "signal" made by S. aureus. P. aeruginosa "senses" that signal and activates a form of motility that the authors term "exploratory motility." The authors go on to show that exploratory motility is driven, at least in part, by type IV pili. They also show that the signal generated by S. aureus is regulated by the Agr quorum sensing system. Overall, this work represents the beginning of a potentially exciting story with important implications to research on biofilms, bacterial interactions, and motility. However, in its current state the paper falls in between two incomplete stories. If the primary goal is to characterize a new behavioral transition of P. aeruginosa from collective to solitary movement, the authors need to better understand the mechanism underlying the different motility modes and the transition between them. And if the primary emphasis is the interspecies signaling, the authors need to better understand the nature of the signaling involved. In my view, answering one of these two questions would elevate the paper to the standards of eLife.

1) Is there really a "signal" sensed by P. aeruginosa? While it seems clear that S. aureus can affect P. aeruginosa motility and that its ability to do so depends on Agr, I am concerned that this could be a passive effect of a secreted factor (such as a surfactant or a protease) rather than an active signaling process. The argument for active changes in cAMP levels (Figure 6B) shows a modest decrease in cAMP levels when P. aeruginosa is exposed to S. aureus, but this effect is only seen at late time-points which do not appear to be consistent with the kinetics of the response (Figure 2A). To really confirm this point the authors would ideally be able to elicit the response with the purified signal. For example, if the signaling is direct, they could try simply adding AIP to P. aeruginosa. If not, the authors could better characterize the spent media activity they describe, for example by showing whether it is heat-stable, protease-resistant, organically-soluble, etc and then pursue purification appropriately.

We thank the reviewer for these helpful comments. These questions are of significant interest to us as well and are the focus of current studies. We have evidence that suggests there are multiple signals, which can independently function to promote motility or function as a chemoattractant. Our preliminary studies indicate that the addition of S. aureus AIP is not sufficient to promote P. aeruginosa motility or chemotaxis. Thus, as the reviewer suggested, we have analyzed the spent supernatant and determined the factors necessary to increase motility are heat stable, protease sensitive, hydrophilic protein(s). We have identified candidate secreted Staphylococcal products we predict promote motility and/or chemotaxis and are working to define their specific roles therein. These experiments are the focus of a follow-up manuscript.

The reviewer brings up a second salient point regarding the kinetics of cAMP signaling in response to S. aureus. Experiments are currently underway to interrogate cAMP levels at the single cell level at earlier time points with live-imaging. We predict these experiments will also inform the important questions the reviewer raises in point 2, as we are able to more specifically visualize the kinetics of cAMP levels and the initiation of single-cell motility. We have therefore chosen to remove the current analysis and focused our discussion on the role of the PilG in light of the Oliveira paper, as recommended.

2) How does the "signal" affect P. aeruginosa motility in general and type IV pili and flagella in particular? Figure 6 argues that cAMP drives exploratory motility, but I feel that this is simply correlation. In addition, it is unclear to me how changes in cAMP affect pilus activity and how pilus activity leads to "collective" or "single-cell" motility. At a minimum the authors could use single-cell imaging, perhaps with sparse labeling of fluorescent cells, to better understand the role of pili and flagella in collective motility. Additional cAMP mutants could help show the real necessity for this pathway. And finally, the role of PilJ needs to be untangled, as the signs of the signaling are confusing (how does inhibition of PilJ function stimulate solitary motility?).

We agree with the reviewer that it is important to understand how changes in cAMP and pilus activity influence the transition between collective and single-cell motility. Studies are currently underway to measure pili biogenesis and activity via immunoblot and live-imaging of fluorescently labeled pili in the presence and absence of both S. aureus cells and secreted factors identified to be necessary as discussed above.

Please see response to point #1 above, and response to essential revisions #1.

Reviewer #2:

[…] I have only two major comments:

1) The authors propose that "exploratory behavior" may have implications for co-infections in cystic fibrosis and may represent a novel therapeutic target. This is highly speculative as the conditions in the CF lung are substantially different than the in vitro conditions used in the study. While P. aeruginosa/S. aureus co-infection is common in CF, P. aeruginosaa is an environmental bacterium that otherwise rarely coexists with S. aureus. It seems likely that this behavior evolved in an environmental niche, where it may provide a competitive advantage. It would be interesting to known if other species (in particular other environmental Gram-positive bacteria) can evoke the same phenotype in P. aeruginosa. This would provide information about the generality of the behavior in response to other potential bacterial competitors.

We thank the reviewer for these insightful comments. We absolutely agree that these interspecies behaviors likely evolved in an environmental niche, not the CF airway. Accordingly, we examined additional bacterial species, both isolates from CF patients and model organisms not common to the CF airway. As the reviewer suspected, P. aeruginosa was able to respond to a range of organisms, although some specificity was observed (please see response to essential revision #2 for more detail).

2) The signal and response mechanisms are only superficially explored. More information on either or both would increase the impact of the paper; however, I concede that an exhaustive characterization is unlikely to be achieved in a reasonable timeframe. Minimally, it would be useful to comment on other known factors implicated in P. aeruginosa Tfp function, including ChpA, PilG/PilH and PilY1. Also, a generic characterization of the chemotactic signal would be informative and support the findings; for example, is it a protein or heat labile.

These questions are of significant interest to us as well and are the focus of current studies. We have evidence that suggests there are multiple signals, which can independently function to promote motility or function as a chemoattractant. Thus, as the reviewer suggested, we have analyzed the spent supernatant and determined the factors necessary to increase motility are heat stable, protease sensitive, hydrophilic protein(s). We have identified candidate secreted Staphylococcal products we predict promote motility and/or chemotaxis and are working to define their specific roles therein. These experiments are the focus of a follow-up manuscript.

In response to comments by both reviewers, we have further investigated the role for PilG in light of the Oliveira paper. Please see response above to essential revision #1.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

While we appreciate that you successfully responded to our previous set of requests, and think this paper is on-track for acceptance in eLife, reviewer 1 raises an important point about the nature of the signal (and how it is discussed) that requires clarification. While we sympathize that revealing the identity of the signal might be something you are saving for another paper, inclusion of its identify here would significantly elevate the work. How easy would it be/how willing are you to include the identity of the unspecific "signal" in this manuscript, and/or test the hypothesis that it is surfactin (as suggested by reviewer 1)? If there is a compelling reason why taking it this one step further is impossible or undesirable, please explain.

In response to the suggestion by reviewer #1 and recommendations from the editors, regarding the investigation of a role for surfactants in the induction of P. aeruginosa motility by S. aureus, we have included data describing the biochemical analysis of S. aureus supernatant we performed in an effort to identify the factor(s) made by S. aureus that promote motility in P. aeruginosa. These data reveal that the factors are heat stable, hydrophilic, and protease sensitive.

Since we previously demonstrated that Agr was necessary for the secretion of these products, we were able to identify S. aureus phenol soluble modulins (PSMs) as a putative factor necessary for induction of motility. We have included the genetic data showing that PSMs indeed play a role in the induction of P. aeruginosa motility, but additional unidentified factors are also likely required based on the data we present here. PSMs are multifunctional peptides: they can function as surfactants, lyse eukaryotic and some prokaryotic cells, and are chemoattractant for neutrophils. We speculate that the ability of PSMs to reduce interfacial tension increases P. aeruginosa TFP-mediated motility. As reviewer 1 notes, the observation that the undomesticated B. subtilis strain, known to secrete high levels of surfactant, also promotes P. aeruginosa motility supports this hypothesis. As we previously mentioned, experiments to understand how bacterial surfactants (including PSMs, surfactin, rhamnolipids, and serrawettin) promote T4P motility per se and influence the directionality of P. aeruginosa motility are underway. We have included a discussion of potential models in the manuscripts, but we feel that additional experimental investigation is outside the scope of this paper.

Below, please find the direct critique from reviewer 1 that provides constructive feedback. Please address why you can or cannot comply with this request.

Reviewer #1:

In the original submission, Limoli et al. described an increase in motility of P. aeruginosa in the presence of S. aureus, which they argued could explain some of the interesting interactions known to occur between these species in clinically-important setting such as cystic fibrosis. Taking the two reviewers' comments into consideration, the editor requested that they address three major points: the species-specificity of the interaction, its genetic requirements, and its comparison to a previous potentially-related study by Oliveira et al.

The authors should be commended for indeed addressing all three points. Specifically, they now add additional data showing that the interactions are not species-specific and that the genetic requirements for their phenotype are similar to those previously reported by Oliveira et al.

On one hand, the authors successfully did what was asked of them. On the other hand, unfortunately, some of the new results dampened my enthusiasm for the work. For example, I am excited that clinical CF isolates of P. aeruginosa and S. aureus exhibit the same interspecies behavior as the model strains tested previously. However, I am worried that the new data showing that other bacterial species also induce P. aeruginosa motility indicates that the "signal" the authors are interested in is actually a more general secreted factor that passively induces motility. Specifically, I am intrigued that B. subtilis 3610 induces the response while B. subtilis PY79 (note: there is a typo in Figure 7D that calls it PV79) does not, as those two strains differ in their ability to make surfactin. Based on these new data, I feel that it is appropriate to directly test the hypothesis that surfactin is involved by testing a mutant of 3610 lacking the ability to make surfactin.

Please see above for general response to this comment. Data describing these results are included in Figure 8 and in the Results. Further discussion is included in the Discussion section.

The typo in Figure 7D has been corrected.

I would also encourage the authors to edit the tone of the Abstract and Introduction to better reflect their new findings. For example, I would feel more comfortable if the claim of "signaling" is decreased throughout the paper, as I am not convinced that active signaling is occurring. Along similar lines, the revised Abstract still makes it seem like the findings on the P. aeruginosa-S. aureus interactions are specific.

The use of the word signaling has been decreased throughout the paper, including the title and Abstract.