Salmonella Extracellular Matrix Components Influence Biofilm Formation and Gallbladder Colonization

Abstract

Salmonella enterica serovar Typhi, the causative agent of typhoid fever in humans, forms biofilms encapsulated by an extracellular matrix (ECM). Biofilms facilitate colonization and persistent infection in gallbladders of humans and mouse models of chronic carriage. Individual roles of matrix components have not been completely elucidated in vitro or in vivo To examine individual functions, strains of Salmonella enterica serovar Typhimurium, the murine model of S Typhi, in which various ECM genes were deleted or added, were created to examine biofilm formation, colonization, and persistence in the gallbladder. Studies show that curli contributes most significantly to biofilm formation. Expression of Vi antigen decreased biofilm formation in vitro and virulence and bacterial survival in vivo without altering the examined gallbladder pro- or anti-inflammatory cytokines. Oppositely, loss of all ECM components (ΔwcaM ΔcsgA ΔyihO ΔbcsE) increased virulence and bacterial survival in vivo and reduced gallbladder interleukin-10 (IL-10) levels. Colanic acid and curli mutants had the largest defects in biofilm-forming ability and contributed most significantly to the virulence increase of the ΔwcaM ΔcsgA ΔyihO ΔbcsE mutant strain. While the ΔwcaM ΔcsgA ΔyihO ΔbcsE mutant was not altered in resistance to complement or growth in macrophages, it attached and invaded macrophages better than the wild-type (WT) strain. These data suggest that ECM components have various levels of importance in biofilm formation and gallbladder colonization and that the ECM diminishes disseminated disease in our model, perhaps by reducing cell attachment/invasion and dampening inflammation by maintaining/inducing IL-10 production. Understanding how ECM components aid acute disease and persistence could lead to improvements in therapeutic treatment of typhoid fever patients.

FIG 1

Mutations in components of the ECM showed variations in biofilm formation compared to WT S. Typhimurium exposed to 0.1% bile on cholesterol-coated surfaces in vitro. Biofilm screening was performed in untreated microtiter plates coated with 5 mg/ml cholesterol at 24 h post-bacterial inoculation. The crystal violet staining method was used to estimate biofilm production. Experiments were performed in triplicate and repeated 3 times. A one-way analysis of variance (ANOVA) with Dunnett's multiple comparison test correction was used to determine significant differences between mutant strains and the WT. *, P < 0.05; **, P < 0.01. The error bars indicate standard errors of the means.

FIG 2

Survival plot of mice given intraperitoneal injections containing 10⁶ WT (●) or WT+ViAg (▼) cells. No further mice died after day 7, and the surviving animals were sacrificed at day 13 postinfection (P.I.). n = 8 mice per group.

FIG 3

Bacterial gallbladder CFU enumeration of Salmonella at 7 days postinfection in the presence of gallstones. (A) Mice were given intraperitoneal injections of WT (circles) or WT+ViAg (triangles) cells. (B) In the competition experiment, mice were given intraperitoneal injections containing a mixture of WT cells marked with streptomycin resistance and WT+ViAg cells marked with kanamycin resistance. In both cases, the resulting final dose was 10⁴, and the limit of detection was 100 CFU/ml. The dotted line represents a completely neutral competitive index. An unpaired t test with Welch's correction was used to determine significant differences between mutant strains and the WT in each organ. **, P < 0.01. Bars indicate means. The WT Cam^r versus WT Strep^r comparison (ctrl) showed a neutral competitive index (CI) with a median value of 0.2059.

FIG 4

IL-10 and TNF-α production of mouse gallbladder samples infected with either the WT (gray bars) or WT+ViAg (white bars). Each bar consists of the combined data from 10 separate samples. Error bars indicate standard errors. ns, not significant.

FIG 5

(A) Survival plot of mice given intraperitoneal injections containing 10⁵ WT (●) or ΔwcaM ΔcsgA ΔyihO ΔbcsE mutant (■) cells. All mice were sacrificed at day 8 postinfection. n = 8 mice/group. (B) CFU enumeration in gallbladder or bile or on gallstones from mice given intraperitoneal injections containing 10⁴ WT (●) or ΔwcaM ΔcsgA ΔyihO ΔbcsE mutant (■) cells at 8 days postinfection in the presence of gallstones. The limit of detection was 100 CFU/ml. An unpaired t test with Welch's correction was used to determine significant differences between the mutant strain and WT (**, P < 0.01). Bars indicate means. (C) CFU enumeration in gallbladder or bile or on gallstones from mice given intraperitoneal injections containing 10⁴ WT (●) or ΔwcaM ΔcsgA (◆), ΔwcaM ΔcsgA ΔyihO (▲), or ΔwcaM ΔcsgA ΔyihO ΔbcsE (■) mutant cells at 8 days postinfection in the presence of gallstones. The limit of detection was 100 CFU/ml. One-way ANOVA with Dunnett's multiple comparison test was used to determine significant differences between mutant strains and the WT (*, P < 0.05). Bars indicate means. (D) Bacterial CFU enumeration of Salmonella at 8 days postinfection in the presence of gallstones. Mice were challenged with intraperitoneal injections containing a 10⁴-CFU mixture of WT cells marked with streptomycin resistance and ΔwcaM ΔcsgA ΔyihO ΔbcsE mutant cells marked with chloramphenicol. The limit of detection was 100 CFU/ml. The dotted line represents a completely neutral competitive index. Wilcoxon's signed rank test was used to determine significant differences between mutant strains and the WT (*, P < 0.05; **, P < 0.01). Bars indicate means. Positive values represent the mutant outcompeting the WT.

FIG 6

(A) IL-10 and (B) TNF-α levels in gallbladder tissue or bile from mice infected with either WT (gray bars) or ΔwcaM ΔcsgA ΔyihO ΔbcsE mutant (white bars) cells. The data from gallbladder tissue or bile are from 10 separate samples. An unpaired t test was used to determine significant differences between the mutant strain and WT (*, P < 0.05). Error bars indicate standard errors.

FIG 7

(A) Macrophage cell line J774.1 was infected with either WT (●) or ΔwcaM ΔcsgA ΔyihO ΔbcsE mutant (■) cells. Numbers of bacteria per cell (both inside and membrane associated) were recovered at 1 h postinfection and after 45 min of gentamicin treatment. An unpaired t test was used to determine significant differences between mutant strains and the WT in each organ (ns, not significant). Bars indicate means. (B) Human serum sensitivity assays were performed on WT (●) and ΔwcaM ΔcsgA ΔyihO ΔbcsE mutant (■) cells. E. coli (◆) was used as a positive control to ensure complement killing, and E. coli with heat-inactivated serum (♢) served as a negative control. The limit of detection was 100 CFU/ml. Experiments were repeated 3 times. An unpaired t test with Welch's correction was used to determine significant differences between mutant strain and WT (ns, no significance).

FIG 8

Bacterial CFU enumeration of Salmonella in competition experiments at 8 days postinfection in the presence of gallstones. Mice were given intraperitoneal injections containing 10⁴ CFU that represent an equal mixture of WT cells marked with streptomycin resistance and (A) ΔwcaM::Kan, (B) ΔcsgA::Kan, (C) ΔyihO::Cam, or (D) ΔbcsE::Cam mutant cells. The limit of detection was 100 CFU/ml. The dotted line represents a completely neutral competitive index. Wilcoxon's signed rank test was used to determine significant differences between mutant strains and the WT (*, P < 0.05; **, P < 0.01; ***, P < 0.001). The bars indicate the means of the groups. Positive values represent the mutant outcompeting the WT.