High Levels of Cyclic Diguanylate Interfere with Beneficial Bacterial Colonization

Abstract

During colonization of the Hawaiian bobtail squid (Euprymna scolopes), Vibrio fischeri bacteria undergo a lifestyle transition from a planktonic motile state in the environment to a biofilm state in host mucus. Cyclic diguanylate (c-di-GMP) is a cytoplasmic signaling molecule that is important for regulating motility-biofilm transitions in many bacterial species. V. fischeri encodes 50 proteins predicted to synthesize and/or degrade c-di-GMP, but a role for c-di-GMP regulation during host colonization has not been investigated. We examined strains exhibiting either low or high levels of c-di-GMP during squid colonization and found that while a low-c-di-GMP strain had no colonization defect, a high c-di-GMP strain was severely impaired. Expression of a heterologous c-di-GMP phosphodiesterase restored colonization, demonstrating that the effect is due to high c-di-GMP levels. In the constitutive high-c-di-GMP state, colonizing V. fischeri exhibited reduced motility, altered biofilm aggregate morphology, and a regulatory interaction where transcription of one polysaccharide locus is inhibited by the presence of the other polysaccharide. Our results highlight the importance of proper c-di-GMP regulation during beneficial animal colonization, illustrate multiple pathways regulated by c-di-GMP in the host, and uncover an interplay of multiple exopolysaccharide systems in host-associated aggregates. IMPORTANCE There is substantial interest in studying cyclic diguanylate (c-di-GMP) in pathogenic and environmental bacteria, which has led to an accepted paradigm in which high c-di-GMP levels promote biofilm formation and reduce motility. However, considerably less focus has been placed on understanding how this compound contributes to beneficial colonization. Using the Vibrio fischeri-Hawaiian bobtail squid study system, we took advantage of recent genetic advances in the bacterium to modulate c-di-GMP levels and measure colonization and track c-di-GMP phenotypes in a symbiotic interaction. Studies in the animal host revealed a c-di-GMP-dependent genetic interaction between two distinct biofilm polysaccharides, Syp and cellulose, that was not evident in culture-based studies: elevated c-di-GMP altered the composition and abundance of the in vivo biofilm by decreasing syp transcription due to increased cellulose synthesis. This study reveals important parallels between pathogenic and beneficial colonization and additionally identifies c-di-GMP-dependent regulation that occurs specifically in the squid host.

Keywords: Euprymna scolopes; Hawaiian bobtail squid; Vibrio fischeri; bacterial colonization; biofilm; c-di-GMP; cyclic diguanylate; flagellar motility; microbiome; motility; squid; symbiosis.

FIG 1

High c-di-GMP levels inhibit swimming motility, promote Congo red binding, and promote cellulose gene transcription. (A) Quantification of c-di-GMP concentration for indicated V. fischeri strains in LBS and FSIO using the pFY4535 c-di-GMP reporter plasmid. Constitutive AmCyan was used to normalize GFP to cell density. For each strain, n = 3 biological and n = 3 technical replicates per biological replicate, each point represents the mean of technical replicates, average bars represent the means of biological replicates, and error bars represent standard errors of the mean. (B) Quantification of migration through soft (0.3%) agar for indicated V. fischeri strains. For each strain, n = 8 to 9 biological replicates, average bars represent means, and error bars represent standard errors of the mean. Motility plate images are representative. (C) Quantification of Congo red binding for indicated V. fischeri strains. For each strain, n = 3 technical replicates, average bars represent means, and error bars represent standard errors of the mean. Congo red spot images are representative. (D) Quantification of in vitro bcs transcription by indicated V. fischeri strains using the pRYI063 bcsQ′-gfp⁺ transcriptional reporter plasmid. For each strain, n = 3 biological replicates. Points represent each biological replicate, average bars represent the means, and error bars represent standard errors of the mean. Constitutive mCherry was used to normalize GFP to cell density. For panels A to D, one-way analysis of variance (ANOVA) was used for statistical analysis, and asterisks represent significance relative to the WT strain (ns, not significant; *, P < 0.05; **, P < 0.006; ****, P < 0.0001).

FIG 2

Elevated c-di-GMP levels inhibit host colonization. (A) Quantification of squid colonization levels at 18 hpi by indicated V. fischeri strains and aposymbiotic (Apo) control. Sample sizes from left to right are 8, 14, 25, and 13 squid. (B) Quantification of squid colonization levels at 48 hpi by indicated V. fischeri strains and aposymbiotic (Apo) control. Sample sizes from left to right are 23, 68, 75, and 65 squid. For panels A and B, box-and-whisker plots represent the minimum, 25th percentile, median, 75th percentile, and maximum. A Kruskal-Wallis test was performed for statistical analysis for squid that were introduced to bacteria (ns, not significant; **, P < 0.002; ****, P < 0.0001).

FIG 3

Reduced swimming motility, increased Congo red binding, and diminished squid colonization of the High cdG strain are dependent on high c-di-GMP levels. (A) Quantification of c-di-GMP concentration for indicated V. fischeri strains using the pFY4535 c-di-GMP reporter plasmid. For each strain, n = 3 biological and n = 3 technical replicates, each dot represents the mean of technical replicates, average bars represent the means of biological replicates, and error bars represent standard errors of the mean. (B) Quantification of migration through TBS soft (0.3%) agar for indicated V. fischeri strains. For each strain, n = 5 biological replicates, average bars represent means, and error bars represent standard errors of the mean. (C) Quantification of Congo red binding for indicated V. fischeri strains. For each strain, n = 3 technical replicates, average bars represent the means, and error bars represent standard errors of the mean. Congo red spot images are representative. For panels A to C, one-way ANOVA was used for statistical analysis (ns, not significant; *, P = 0.017; ****, P < 0.0001). (D) Quantification of squid colonization levels at 48 hpi by indicated V. fischeri strains and aposymbiotic (Apo) control. Box-and-whisker plots represent the minimum, 25th percentile, median, 75th percentile, and maximum. Sample sizes from left to right are 30, 26, 34, 37, and 11 squid. A Kruskal-Wallis test was performed for statistical analysis for squid that were introduced to bacteria (ns, not significant; ****, P < 0.0001).

FIG 4

High c-di-GMP levels do not eliminate motility during host colonization. Quantification of squid competitive colonization index at 48 hpi by indicated V. fischeri strains. The competitive index represents the log₁₀[(test strain/High cdG)_output/(test strain/High cdG)_input]. Box-and-whisker plots represent the minimum, 25th percentile, median, 75th percentile, and maximum. Sample sizes from left to right are 43 and 75 squid. A Kruskal-Wallis test was performed for statistical analysis (****, P < 0.0001).

FIG 5

Elevated c-di-GMP levels lead to a greater number of V. fischeri aggregates in the squid host. (A) Representative fluorescence microscopy images of aggregates made by indicated V. fischeri strains carrying a constitutive GFP on the pVSV102 plasmid within the host mucus. Arrows indicate location of aggregates. (B) Quantification of aggregate area per squid for indicated V. fischeri strains carrying a constitutive GFP on the pVSV102 plasmid. Zero values are displayed at the bottom of the plot. Box-and-whisker plots represent the minimum, 25th percentile, median, 75th percentile, and maximum. Sample sizes from left to right are 73, 131, 101, 38, and 12 squid. A Kruskal-Wallis test was used for statistical analysis (ns, not significant; *, P < 0.04; **, P < 0.02; ****, P < 0.0001). (C) Quantification of number of aggregates formed per squid by indicated V. fischeri strains. Sample sizes from left to right are 73, 131, and 101 squid.

FIG 6

c-di-GMP downregulates syp via BcsA in host mucus but not in culture. (A) Quantification of syp transcription for indicated V. fischeri strains using the pM1422 sypA′-gfp⁺ transcriptional reporter plasmid in aggregates within the host mucus. Sample sizes from left to right are 47, 40, and 22 aggregates. (B) Quantification of bcs transcription for indicated V. fischeri strains using the pRYI063 bcsQ′-gfp⁺ transcriptional reporter plasmid in aggregates within the host mucus. Sample sizes from left to right are 19, 24, and 28 aggregates. For panels A and B, box-and-whisker plots represent the minimum, 25th percentile, median, 75th percentile, and maximum. A Kruskal-Wallis test was used for statistical analysis (ns, not significant; ****, P < 0.0001). (C) Quantification of in vitro syp transcription in LBS and TBS-Ca²⁺ for the indicated V. fischeri strains carrying the pM1422 sypA′-gfp⁺ transcriptional reporter plasmid. For each strain, n = 3 biological replicates. Points represent each replicate, average bars represent the means of replicates, and error bars represent the standard errors of the mean. One-way ANOVA was used for statistical analysis (ns, not significant relative to the WT strain in the same media). For panels A to C, constitutive mCherry was used to normalize GFP to cell density.