A network perspective on the role of c-di-GMP-associated protein complexes in biofilm formation
Curation statements for this article:-
Curated by eLife
eLife Assessment
This valuable study presents a comprehensive exploration of c-di-GMP-associated protein interaction networks in Pseudomonas fluorescens, with a particular focus on biofilm-related phenotypes. The evidence is convincing, supported by a genome-wide yeast two-hybrid screen, phenotypic analyses, and experimental validation. The work identifies multiple interaction hubs and provides a resource that will be of use to the biofilm and c-di-GMP communities, while additional mechanistic exploration would further enhance its impact by clarifying the biological roles of many of the identified interactions.
This article has been Reviewed by the following groups
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
- Evaluated articles (eLife)
Abstract
The secondary messenger cyclic di-GMP is a ubiquitous bacterial signal that regulates the switch from a free-swimming to a sessile biofilm-forming lifestyle. Many biofilm-forming Pseudomonas species possess numerous c-di-GMP-binding proteins (CDGs) which regulate gene expression, protein activity, and protein complexes. However, the mechanisms by which numerous CDG effectors form a coherent signaling network to coordinate lifestyle changes remain poorly understood. We addressed this knowledge gap by focusing on ten CDG proteins involved in biofilm development in P. fluorescens SBW25. We used an integrated approach combining a protein interaction network from genome-wide yeast two-hybrid (Y2H) screens with large-scale biofilm and motility phenotype analyses via CRISPR interference (CRISPRi). Our network associated c-di-GMP signaling with processes such as signal transduction, solute transport, secretion, virulence, transcriptional regulation, DNA repair, and cell division. We discovered unknown functions of two CDG proteins in DNA repair and cell division, supporting the significance of our network. Notably, the phosphodiesterase DipA interacts with numerous CDG proteins through GGDEF domains. Phenotypic analyses revealed that CDG partners were highly correlated or strongly anticorrelated with DipA. These findings suggest that DipA is a central hub for CDG interactions that integrates opposing modules. These findings support the hub-based model of c-di-GMP signaling, which is crucial for localized control and rapid adaptation to environmental changes.
Article activity feed
-
eLife Assessment
This valuable study presents a comprehensive exploration of c-di-GMP-associated protein interaction networks in Pseudomonas fluorescens, with a particular focus on biofilm-related phenotypes. The evidence is convincing, supported by a genome-wide yeast two-hybrid screen, phenotypic analyses, and experimental validation. The work identifies multiple interaction hubs and provides a resource that will be of use to the biofilm and c-di-GMP communities, while additional mechanistic exploration would further enhance its impact by clarifying the biological roles of many of the identified interactions.
-
Reviewer #1 (Public review):
Summary:
This manuscript by Noirot-Gros et. al. presents a herculean effort to map the protein-protein interactome of the c-di-GMP signaling network in Pseudomonas fluorescens (Pf). C-di-GMP, the key driver of biofilm formation in bacteria, is controlled by a highly complex network of synthesis, degradation and effector proteins. Pf is no exception as it encodes dozens of such proteins. The authors use a Yeast Two-Hybrid approach genome-wide screen with 10 diguanylate cyclase (DGC) enzymes as bait to assess protein-protein interactions in this network. The results identify over one hundred such interactions with several different hubs, including c-di-GMP signaling, other signaling systems, membrane proteins, etc. The authors then explore the original bait proteins as well as identify interactors on biofilm …
Reviewer #1 (Public review):
Summary:
This manuscript by Noirot-Gros et. al. presents a herculean effort to map the protein-protein interactome of the c-di-GMP signaling network in Pseudomonas fluorescens (Pf). C-di-GMP, the key driver of biofilm formation in bacteria, is controlled by a highly complex network of synthesis, degradation and effector proteins. Pf is no exception as it encodes dozens of such proteins. The authors use a Yeast Two-Hybrid approach genome-wide screen with 10 diguanylate cyclase (DGC) enzymes as bait to assess protein-protein interactions in this network. The results identify over one hundred such interactions with several different hubs, including c-di-GMP signaling, other signaling systems, membrane proteins, etc. The authors then explore the original bait proteins as well as identify interactors on biofilm formation-related phenotypes and swarming using a high-throughput CRISPRi expression knockdown approach. The amount of data generated is quite impressive. Much of the manuscript uses statistical-based network analysis to group different proteins based on their interactions or impact on phenotypes, which is a high-level analysis that can catalyze further study into this system. The authors chose three specific proteins to assess their impact on cell morphology, DNA repair, and protein localization. Overall, in my view, this is perhaps the best analysis of a c-di-GMP protein-protein interactome, and it provides a multitude of hypotheses to be tested. However, therein lies the weakness of the manuscript in that very few of these hypotheses are actually tested. But such is not the goal of this network analysis type of approach. Overall, I think the work will be highly impactful to those in the c-di-GMP field, and it provides a template for others attempting such analyses of protein-protein interactions.
Strengths:
The manuscript is impressive in the sheer scale of the protein-protein interactions identified, network analysis, and phenotypic analysis of specific proteins in the network. It is an impressive amount of work that could be very useful to the field. It is also statistically rigorous in its analysis of significant interactions or network nodes.
Weaknesses:
The weakness of the manuscript is that, with three exceptions, very few of the hypotheses are actually tested. For example, BifA is shown to be a network hub protein that interacts with many other diguanylate cyclases, and this is hypothesized to be through GGDEF heterodimerization. I appreciate that experimentally testing such a hypothesis is probably another entire manuscript, but some early forays into such ideas could be undertaken using AlphaFold structural modeling of protein-protein interactions compared with GGDEFs that don't form heterodimers. Also, an inherent weakness is that such detailed analyses of a c-di-GMP signaling network, in which each diguanylate cyclase and phosphodiesterase may respond to a unique cue, is that the network identified and the conclusions made are highly specific to the experimental conditions in which the work was done. Therefore, it is unclear how broadly these conclusions (i.e. BifA is the central regulator of c-di-GMP signaling) apply to other conditions. But it is impossible to get around such a limitation, and this work can lead to testing the robustness of the identified network in other environments.
-
Reviewer #2 (Public review):
Summary:
In this manuscript, Noirot-Gros and coworkers investigated the network of c-di-GMP associated protein complexes in Pseudomonas fluorescens. They did so by using a genome-wide yeast two-hybrid screen, and that was further probed by phenotypic screening that focused on biofilm and motility phenotypes. From this network map, they discovered that the phosphodiesterase DipA interacts with the GGDEF domains of many c-di-GMP-binding proteins.
Strengths:
(1) Broadness of screen led to identification of new interactions: The genome-wide yeast two-hybrid screening approach permitted broad investigation of c-di-GMP-associated protein-protein interactions. These interactions included some previously validated interactions as well as newly discovered interactions.
(2) Complementary experimental validation: The …
Reviewer #2 (Public review):
Summary:
In this manuscript, Noirot-Gros and coworkers investigated the network of c-di-GMP associated protein complexes in Pseudomonas fluorescens. They did so by using a genome-wide yeast two-hybrid screen, and that was further probed by phenotypic screening that focused on biofilm and motility phenotypes. From this network map, they discovered that the phosphodiesterase DipA interacts with the GGDEF domains of many c-di-GMP-binding proteins.
Strengths:
(1) Broadness of screen led to identification of new interactions: The genome-wide yeast two-hybrid screening approach permitted broad investigation of c-di-GMP-associated protein-protein interactions. These interactions included some previously validated interactions as well as newly discovered interactions.
(2) Complementary experimental validation: The proposed network was experimentally validated, including by using a CRISPRi-based approach in which the expression of genes encoding proteins identified in the network was systematically suppressed, and then the impact on the biofilm and motility phenotypes was assessed.
Weaknesses:
The findings would have been strengthened by further biochemical analysis, but this is likely beyond the scope of the paper.
-
Reviewer #3 (Public review):
Summary:
In this manuscript, Noirot-Gross et al take an open-ended approach to elucidate the c-diGMP-associated protein complexes in Pseudomonas fluorescens. Starting with 10 cyclic d-GMP putative proteins, they use a combination of genome-wide two-hybrid system followed by CRISPRi-mediated exploration of phenotypes to describe the cyclic di-GMP-associated regulation of biofilm formation, and how it relates to other functions. Overall, this work presents an excellent example of how genome annotations can be further confirmed with the use of integrated functional genomic approaches. Some areas of improvement can be applied to this manuscript to enhance readability and provide a clearer distinction between confirmatory results and new findings, which are provided below:
Strengths:
(1) The authors have explored …
Reviewer #3 (Public review):
Summary:
In this manuscript, Noirot-Gross et al take an open-ended approach to elucidate the c-diGMP-associated protein complexes in Pseudomonas fluorescens. Starting with 10 cyclic d-GMP putative proteins, they use a combination of genome-wide two-hybrid system followed by CRISPRi-mediated exploration of phenotypes to describe the cyclic di-GMP-associated regulation of biofilm formation, and how it relates to other functions. Overall, this work presents an excellent example of how genome annotations can be further confirmed with the use of integrated functional genomic approaches. Some areas of improvement can be applied to this manuscript to enhance readability and provide a clearer distinction between confirmatory results and new findings, which are provided below:
Strengths:
(1) The authors have explored their findings extensively and provide a comprehensive view of the topic.
(2) The combination of genome-wide explorations of protein-protein interactions with the more focused phenotypic exploration of the interactions found provides a solid framework for the work presented.
Weaknesses:
(1) Overall goal of the work:
While articles that describe open-ended approaches can be comprehensive and descriptive in nature, the authors should have a main overall goal, which can guide the reader through the main and most compelling findings at the end. As written, the overall goal is not clear. The network perspective is interesting, and the focus on biofilm formation appears in the title. Why P. fluorescens? How is cyclic di-GMP-mediated regulation of biofilm formation in P. fluorescens different from P. aeruginosa? Why would it be studied?
(Positive or negative regulation of biofilm formation?)
(2) Abstract:
The abstract is very well written and guides the reader to the DipA as a hub protein in the network. From further reading, the article could clarify whether this finding is confirmatory or novel (does DipA play a similar role in P. aeruginosa?) It would be appropriate to mention the role of DipA in other Pseudomonas species from the beginning, and not only in the discussion session.
(3) Introduction:
The introduction is nicely written. An area of improvement could be giving more attention to protein interactions as relevant to c-di-GMP. The authors could consider an independent paragraph starting with line 84-85 "Protein-protein interactions involving DGCs, PDEs, and target effectors are crucial in establishing localized signalling through the generation of local pools of c-di-GMP", expanding on this particular aspect with an example of localized signal, after explaining that localization could help decipher specific function within the network of DGCs and PDEs. Then go into connecting biofilms with c-di-GMP and protein-protein interactions, using the example of GcbC and LapD.
(4) The rationale of choosing 10 PDEs could be clarified. The nice diagrams shown in the supplementary table could be used as part of Figure 1, so the reader understands why these proteins were used, and what is known about them (for example, add them as Figure 1a).
(5) Figures 1b and 2 convey the same information as in Figure 1a. They could be removed without affecting the understanding of the article.
(6) CRISPRi and Figure 3. Figure 3 shows the methodology of CRICPR phenotypic screening. A diagram showing the CRISPRi system in P. fluorescens could help the non-expert reader. While the choice of 23 proteins related to the emerging hub DipA is clear, the choice of the other 33 genes could be better explained. Are these proteins already related to biofilm formation? Where are they part of the network detected? How about the other 14 SBW25 genes? The authors could clarify the rationale of the choices. Figure 4 could be combined with Figure 3 or moved to the supplementary material.
(7) Figures 5, 6 and 7 represent solid network analysis of the findings. Still, they could be improved in clarity on the main findings. The authors conclude at the end of section 3.2.3 that there are networks that exert a "positive role" and a "negative role". The authors could show that in the figures, explaining what those roles are: more biofilm structural coding genes? positive or negative regulation of biofilm formation?)
-
Author response:
Public Reviews:
Reviewer #1 (Public review):
Summary:
This manuscript by Noirot-Gros et. al. presents a herculean effort to map the protein-protein interactome of the c-di-GMP signaling network in Pseudomonas fluorescens (Pf). C-di-GMP, the key driver of biofilm formation in bacteria, is controlled by a highly complex network of synthesis, degradation and effector proteins. Pf is no exception as it encodes dozens of such proteins. The authors use a Yeast Two-Hybrid approach genome-wide screen with 10 diguanylate cyclase (DGC) enzymes as bait to assess protein-protein interactions in this network. The results identify over one hundred such interactions with several different hubs, including c-di-GMP signaling, other signaling systems, membrane proteins, etc. The authors then explore the original bait proteins as well as …
Author response:
Public Reviews:
Reviewer #1 (Public review):
Summary:
This manuscript by Noirot-Gros et. al. presents a herculean effort to map the protein-protein interactome of the c-di-GMP signaling network in Pseudomonas fluorescens (Pf). C-di-GMP, the key driver of biofilm formation in bacteria, is controlled by a highly complex network of synthesis, degradation and effector proteins. Pf is no exception as it encodes dozens of such proteins. The authors use a Yeast Two-Hybrid approach genome-wide screen with 10 diguanylate cyclase (DGC) enzymes as bait to assess protein-protein interactions in this network. The results identify over one hundred such interactions with several different hubs, including c-di-GMP signaling, other signaling systems, membrane proteins, etc. The authors then explore the original bait proteins as well as identify interactors on biofilm formation-related phenotypes and swarming using a high-throughput CRISPRi expression knockdown approach. The amount of data generated is quite impressive. Much of the manuscript uses statistical-based network analysis to group different proteins based on their interactions or impact on phenotypes, which is a high-level analysis that can catalyze further study into this system. The authors chose three specific proteins to assess their impact on cell morphology, DNA repair, and protein localization. Overall, in my view, this is perhaps the best analysis of a c-di-GMP protein-protein interactome, and it provides a multitude of hypotheses to be tested. However, therein lies the weakness of the manuscript in that very few of these hypotheses are actually tested. But such is not the goal of this network analysis type of approach. Overall, I think the work will be highly impactful to those in the c-di-GMP field, and it provides a template for others attempting such analyses of protein-protein interactions.
Strengths:
The manuscript is impressive in the sheer scale of the protein-protein interactions identified, network analysis, and phenotypic analysis of specific proteins in the network. It is an impressive amount of work that could be very useful to the field. It is also statistically rigorous in its analysis of significant interactions or network nodes.
Weaknesses:
The weakness of the manuscript is that, with three exceptions, very few of the hypotheses are actually tested. For example, BifA is shown to be a network hub protein that interacts with many other diguanylate cyclases, and this is hypothesized to be through GGDEF heterodimerization. I appreciate that experimentally testing such a hypothesis is probably another entire manuscript, but some early forays into such ideas could be undertaken using AlphaFold structural modeling of protein-protein interactions compared with GGDEFs that don't form heterodimers. Also, an inherent weakness is that such detailed analyses of a c-di-GMP signaling network, in which each diguanylate cyclase and phosphodiesterase may respond to a unique cue, is that the network identified and the conclusions made are highly specific to the experimental conditions in which the work was done. Therefore, it is unclear how broadly these conclusions (i.e. BifA is the central regulator of c-di-GMP signaling) apply to other conditions. But it is impossible to get around such a limitation, and this work can lead to testing the robustness of the identified network in other environments.
We would like to thank the reviewer sincerely for their positive comments on our manuscript and for their constructive feedback. We recognize the limitations arising from the lack of extensive knowledge regarding the environmental cues that trigger the regulation of all CDG activities in P. fluorescens. We hypothesize that DipA acts as a central local hub that positively or negatively regulates the activity of its interacting CDG partners throughout the cell life cycle, lifestyle transitions and environmental signals. Testing this hypothesis would indeed require extensive biochemical and omics approaches. However, strengthening the significance of DipA complexes in silico using AlphaFold is a very appealing proposition and we are currently considering including this analysis in the revised version of the manuscript.
Reviewer #2 (Public review):
Summary:
In this manuscript, Noirot-Gros and coworkers investigated the network of c-di-GMP associated protein complexes in Pseudomonas fluorescens. They did so by using a genome-wide yeast two-hybrid screen, and that was further probed by phenotypic screening that focused on biofilm and motility phenotypes. From this network map, they discovered that the phosphodiesterase DipA interacts with the GGDEF domains of many c-di-GMP-binding proteins.
Strengths:
(1) Broadness of screen led to identification of new interactions: The genome-wide yeast two-hybrid screening approach permitted broad investigation of c-di-GMP-associated protein-protein interactions. These interactions included some previously validated interactions as well as newly discovered interactions.
(2) Complementary experimental validation: The proposed network was experimentally validated, including by using a CRISPRi-based approach in which the expression of genes encoding proteins identified in the network was systematically suppressed, and then the impact on the biofilm and motility phenotypes was assessed.
Weaknesses:
The findings would have been strengthened by further biochemical analysis, but this is likely beyond the scope of the paper.
We would like to express our gratitude to the reviewer for their positive evaluation assessment, and for taking into account the limitations of the study's scope.
Reviewer #3 (Public review):
Summary:
In this manuscript, Noirot-Gross et al take an open-ended approach to elucidate the c-diGMP-associated protein complexes in Pseudomonas fluorescens. Starting with 10 cyclic d-GMP putative proteins, they use a combination of genome-wide two-hybrid system followed by CRISPRi-mediated exploration of phenotypes to describe the cyclic di-GMP-associated regulation of biofilm formation, and how it relates to other functions. Overall, this work presents an excellent example of how genome annotations can be further confirmed with the use of integrated functional genomic approaches. Some areas of improvement can be applied to this manuscript to enhance readability and provide a clearer distinction between confirmatory results and new findings, which are provided below:
Strengths:
(1) The authors have explored their findings extensively and provide a comprehensive view of the topic.
(2) The combination of genome-wide explorations of protein-protein interactions with the more focused phenotypic exploration of the interactions found provides a solid framework for the work presented.
Weaknesses:
(1) Overall goal of the work:
While articles that describe open-ended approaches can be comprehensive and descriptive in nature, the authors should have a main overall goal, which can guide the reader through the main and most compelling findings at the end. As written, the overall goal is not clear. The network perspective is interesting, and the focus on biofilm formation appears in the title. Why P. fluorescens? How is cyclic di-GMP-mediated regulation of biofilm formation in P. fluorescens different from P. aeruginosa? Why would it be studied? (Positive or negative regulation of biofilm formation?)
We would like to express our appreciation to the reviewer for their thorough evaluation of our manuscript and for the constructive feedback they provided. The overall goal of this study will be further refined, and outlined in the introduction in the revised version of the manuscript.
(2) Abstract:
The abstract is very well written and guides the reader to the DipA as a hub protein in the network. From further reading, the article could clarify whether this finding is confirmatory or novel (does DipA play a similar role in P. aeruginosa?) It would be appropriate to mention the role of DipA in other Pseudomonas species from the beginning, and not only in the discussion session.
(3) Introduction:
The introduction is nicely written. An area of improvement could be giving more attention to protein interactions as relevant to c-di-GMP. The authors could consider an independent paragraph starting with line 84-85 "Protein-protein interactions involving DGCs, PDEs, and target effectors are crucial in establishing localized signalling through the generation of local pools of c-di-GMP", expanding on this particular aspect with an example of localized signal, after explaining that localization could help decipher specific function within the network of DGCs and PDEs. Then go into connecting biofilms with c-di-GMP and protein-protein interactions, using the example of GcbC and LapD.
We propose highlighting the example to the local signalling cascade formed by the tripartite system YdaM, YciR and MlrA. This will be addressed in the revised version of the manuscript.
(4) The rationale of choosing 10 PDEs could be clarified. The nice diagrams shown in the supplementary table could be used as part of Figure 1, so the reader understands why these proteins were used, and what is known about them (for example, add them as Figure 1a).
We propose to include a specific section in the supplementary file to explain the whole rationale behind choosing these CDGs. These proteins were selected based on their involvement in different steps of biofilm formation in Pseudomonas, as well as their role in the ability of P. fluorescens strains to colonize plant roots.
(5) Figures 1b and 2 convey the same information as in Figure 1a. They could be removed without affecting the understanding of the article.
Figure 2 will be transferred in Supplementary as part of the Figure S1
(6) CRISPRi and Figure 3. Figure 3 shows the methodology of CRICPR phenotypic screening. A diagram showing the CRISPRi system in P. fluorescens could help the non-expert reader. While the choice of 23 proteins related to the emerging hub DipA is clear, the choice of the other 33 genes could be better explained. Are these proteins already related to biofilm formation? Where are they part of the network detected? How about the other 14 SBW25 genes? The authors could clarify the rationale of the choices. Figure 4 could be combined with Figure 3 or moved to the supplementary material.
A better description of the rationale behind the choice of tested interacting protein partners will be provided. We also agree to combine Figure 4 with Figure 3.
(7) Figures 5, 6 and 7 represent solid network analysis of the findings. Still, they could be improved in clarity on the main findings. The authors conclude at the end of section 3.2.3 that there are networks that exert a "positive role" and a "negative role". The authors could show that in the figures, explaining what those roles are: more biofilm structural coding genes? positive or negative regulation of biofilm formation?)
-
-
-
