Fig 1.
Growth of gut commensals in medium supplemented with exogenous p-cresol.
Growth of gut commensals in BHIS compared to media containing 0.01%, 0.5% and 0.1% (v/v) p-cresol. Each curve represents the mean growth rate of three independent replicates. Regression analysis was used to determine significant differences in growth rate compared to the BHIS control over the course of the experiment and marked *** p<0.01. Error bars represent the standard deviation.
Fig 2.
Co-culture assays in medium supplemented with exogenous p-cresol.
Relative fitness of C. difficile 630Δerm and hpdC::CT when grown in co-culture with gut commensals, with and without 0.05% (v/v) p-cresol. The relative fitness of 630Δerm and hpdC::CT grown in co-culture with a&b) E. coli, c&d) E. faecium, e&f) L. fermentum (e&f). Viable counts are represented by CFU/ml and displayed as a percentage of the total culture. Error bars are representative of three independent replicates. Regression analysis was used to determine significant differences in growth taking strain and media into consideration ***p<0.001.
Fig 3.
The effect of p-HPA supplementation on competitive co-culture of C. difficile with E. coli and detection of p-cresol production.
a) A comparison of the percentage survival of C. difficile strain 630Δerm and E. coli in competitive co-culture for 24 hours, performed in 0.1%, 0.2% and 0.3% p-HPA. Statistical difference in percentage survival was calculated using regression analysis taking strain and media into consideration ***p<0.001. b) HPLC quantification of p-cresol production in competitive co-culture of C. difficile strain 630Δerm with E. coli in media supplemented with 0.1%, 0.2% and 0.3% p-HPA. Statistical difference in p-cresol production was calculated using a two tailed t-test with Welch’s correction, comparing individual p-cresol concentrations to the level produced in 0.1% p-HPA ***p<0.001.
Fig 4.
Competition co-culture assays in the presence of endogenous p-cresol.
Relative fitness of C. difficile strains 630Δerm, hpdC::CT and the complement hpdC::CT::phpdC-A were grown in competitive co-culture for 24 hours with gut commensal species a) E.coli, b) K. oxytoca and c) B. thetaiotaomicron. The growth medium was supplemented with the intermediate in the p-cresol pathway, pHPA (at 0.2% v/v) and 50 ng/ml anhydrotetracycline to induce expression of hpdCA, encoded in trans. The relative representation of each strain was expressed as a percentage of the total CFU count. Error bars are representative of three independent replicates and show the variation across replicates and co-culture conditions. Regression analysis was used to determine significant differences in growth taking strain into consideration and marked ** p<0.01 and ***p<0.001. d) The concentration of p-cresol produced in the co-cultures of three independent replicates was quantified by HPLC and plotted in GraphPad Prism7. Statistical analysis was performed by Linear regression, significant differences in p-cresol production compared to 630Δerm are indicated *** p<0.001.
Fig 5.
Metabolic profiling of C. difficile in monoculture and competitive co-culture with intestinal commensal species.
a) Principle component analysis (PCA) score plot based on the 1H NMR spectra of the three C. difficile strains (630Δerm (black), hpdC::CT (red) and the complemented mutant (hpdC::CT::phpdC-A) in (blue), grown in monoculture (circle symbol) or in competitive co-culture with intestinal commensal species, E. coli (square), K. oxytoca (triangle) and B. thetaiotaomicron (star), in BHIS media supplemented with 0.2% p-HPA. b) Metabolite clustergram represented as a heatmap of Z-scores of the 1H NMR peak integrals, derived from the quantity of six main metabolites driving the separation between the samples (observed on the loading plots of the model) and a dendrogram showing the hierarchical clustering of the samples. The culture conditions, bacterial strains and p-HPA concentration are colour coded in a key on the heatmap.
Fig 6.
Infection of mice with 630Δerm and hpdC::CT.
a) A schematic of the C. difficile relapse infection in C57BL/6 mice. b) CFU counts enumerated from the faeces of individually caged animals (n = 5) infected with either wild-type C. difficile strain (630Δerm) or the corresponding p-cresol null mutant (hpdC::CT) over the duration of the primary infection (days 1–21) and relapse (days 1R-7R). c) CFU enumerated from the faeces day 4 post-infection and d) day 4 relapse. Mann Whitney U test was used to calculate significant differences between infection with wild-type C. difficile and the hpdC mutant marked *p<0.05. Graphs were produced in GraphPad Prism 6.0.
Fig 7.
The diversity of the intestinal microbiota was assessed using 16S rRNA sequencing to determine the bacterial operational taxonomic units present during a relapse model of C. difficile infection. a) Composition of the intestinal microbiota, at the Family level, between wild type C. difficile 630Δerm and p-cresol mutant (hpdC::CT) infected mice and the naïve control at day 7 post-infection, day 0, day 2 and day 4 post relapse. b) Box and whisker plots displaying the total number of operational taxonomic units (OTUs) at the Family level. Statistically significant differences between the diversity of OTU’s are labelled * p<0.05 (ANOVA). c) Principle component analysis of the microbial diversity at the Family level. Each dot represents a single mouse at a given time point, colour-coded for day 7 post-infection, and days 0, day 2 and day 4 post-relapse. d) The relative prevalence of bacterial Class in wild type C. difficile and hpdC::CT infected mice at day 7 post-infection, day 0, day 2 and day 4 post relapse. Statistical analysis was performed in Stata15 by regression analysis taking strain and bacterial class into consideration * p<0.05.
Fig 8.
Metabolic profiling of the in vivo mouse relapse model of CDI.
1H NMR spectroscopy was used to investigate metabolic differences between the individually caged mice infected with C. difficile strain 630Δerm, hpdC::CT or uninfected naïve mice. a) Principle component analysis of the metabolite diversity. Each dot represents a single sample at a given time point. A symbol code was used for the strains 630Δerm (circle), hpdC::CT (square) and naïve mice (star), and a colour code was used to represent the different time points, Day 2 post-infection (dark blue), Day 4 post-infection (light blue), Day 7 post-infection (green), Day 0 relapse (yellow) and Day 4 relapse (orange). b) Loading plots of the PCA model, representing the metabolites driving the distribution of the samples along principal component 1 (PC 1) and principal component 2 (PC 2).
Fig 9.
Percentage survival of bacterial phyla from healthy human faecal slurry treated with p-cresol.
Healthy human donor stool samples were emulsified and treated with exogenously added p-cresol (0.1% or 0.3% v/v) for 90 minutes and the CFU counts were expressed as percentage survival relative to the CFU of the PBS control at t = 0. Faecal emulsions were plated onto differentially selective agar to determine the survival of endogenous intestinal bacteria to p-cresol stress. Error bars are representative of triplicate experiments. Statistical analysis was performed by linear regression to determine the correlation between survival in 0.1% and 0.3% (v/v) p-cresol relative to the PBS control. Significant differences are indicated with ***p<0.001. Error bars are Standard error of the Mean (SEM).
Fig 10.
Phosphate release in Gram-positive and Gram-negative common gut commensal species.
The release of intracellular phosphate was determined spectroscopically (OD650) using malachite green and ammonium molybdate from: a) C. difficile 630Δerm, hpdC::CT and E. coli stationary phase cultures in various concentrations of p-cresol. Significant differences in phosphate release compared to 630Δerm were calculated taking concentration into account using linear regression and are marked *** p<0.001. b) phosphate relapse was measured from Gram-negative common gut commensals and c) phosphate relapse was measured from Gram-positive common gut commensals following incubation with 0.3% (v/v) p-cresol over a 90 minute time period. Red line indicates maximum phosphate release determined by boiling a comparable cell suspension. Curves and error bars are representative of three independent replicates. Significant differences in phosphate release from the maximum were calculated using regression analysis taking time points into consideration, significant differences are marked **p<0.01, ***p<0.001. Error bars are SEM.