The Rcs-Regulated Colanic Acid Capsule Maintains Membrane Potential in Salmonella enterica serovar Typhimurium

Abstract

The Rcs phosphorelay and Psp (phage shock protein) systems are envelope stress responses that are highly conserved in gammaproteobacteria. The Rcs regulon was found to be strongly induced during metal deprivation of Salmonella enterica serovar Typhimurium lacking the Psp response. Nineteen genes activated by the RcsA-RcsB response regulator make up an operon responsible for the production of colanic acid capsular polysaccharide, which promotes biofilm development. Despite more than half a century of research, the physiological function of colanic acid has remained elusive. Here we show that Rcs-dependent colanic acid production maintains the transmembrane electrical potential and proton motive force in cooperation with the Psp response. Production of negatively charged exopolysaccharide covalently bound to the outer membrane may enhance the surface potential by increasing the local proton concentration. This provides a unifying mechanism to account for diverse Rcs/colanic acid-related phenotypes, including susceptibility to membrane-damaging agents and biofilm formation.IMPORTANCE Colanic acid is a negatively charged polysaccharide capsule produced by Escherichia coli, Salmonella, and other gammaproteobacteria. Research conducted over the 50 years since the discovery of colanic acid suggests that this exopolysaccharide plays an important role for bacteria living in biofilms. However, a precise physiological role for colanic acid has not been defined. In this study, we provide evidence that colanic acid maintains the transmembrane potential and proton motive force during envelope stress. This work provides a new and fundamental insight into bacterial physiology.

Keywords: Salmonella; biofilms; colanic acid; exopolysaccharide; extracytoplasmic stress; proton motive force.

FIG 1

RcsA and EPS genes are induced in SFZP mutant Salmonella during metal restriction. (A) Results of qPCR assays performed with cDNA obtained from cultures grown in chelated LB for 2 h. Absolute qPCR values were normalized to the bacterial housekeeping gene rpoD and are expressed as the fold change over the WT. Mean qPCR values from three biological replicates ± the standard deviation are shown. (B) Growth curves of strains in LB with 550 μM dipyridyl at 37°C.

FIG 2

S. Typhimurium sitA feoB zupT metal transport mutants lacking the phage shock response lose cell membrane integrity. Salmonella sitA feoB zupT (SFZ) (A to D) or sitA feoB zupT pspA (SFZP) (E to H) mutants were diluted 1,000-fold in LB with 625 μM dipyridyl and incubated for 3 h (A, B, E, F) or 4 h (C, D, G, H). Representative DIC (A, C, E, G) and TEM (B, F, D, H) images are shown. DIC, ×100, oil immersion. B and F, TEM at ×20,000; D and H, TEM at ×30,000.

FIG 3

The Psp and Rcs responses maintain Δψ in metal transport-deficient mutants during growth in metal-limited medium. Δψ was measured by flow cytometry of aliquots from cultures incubated for 2 h in LB with 625 μM dipyridyl. Flow cytometry was performed with live bacterial cells following 15 min of incubation with the Δψ-sensitive dye DiOC₂(3), which exhibits green fluorescence that shifts toward red fluorescence following Δψ-dependent intracellular aggregation. Data are expressed as the mean red-to-green emission ratio of a population of 2 × 10⁴ cells; a representative plot of three biological replicates is shown. Statistical significance was determined with an unpaired t test (*, P < 0.05; **, P < 0.01). Abbreviations: S, sitA; F, feoB; Z, zupT; P, pspA; R, rcsA.

FIG 4

Mutations in the wza colanic acid and yjbEFGH operons eliminate EPS production. Expression of the colanic acid and yjbEFGH regulons is dependent on the RcsCDB phosphorelay. EPS production was induced by RcsA expressed in trans on a pBR322 replicon. (A) Purified EPS was subjected to a spectrophotometric assay for fucose and uronic acid. Values represent the mean of three biological replicates ± the standard deviation, and significance was determined with an unpaired t test (*, P < 0.05; n.s., not significant; n.d., not detected). (B) Representative images of colonies on LB agar plates formed by the WT strain or EPS-deficient mutants. Each strain contains either the pRcsA expression plasmid or the vector control. Colonies were allowed to grow for 3 days at 25°C. Scale bar, 5 mm.

FIG 5

S. Typhimurium pspA mutants deficient in EPS production have increased sensitivity to antimicrobial peptides. (A) Sensitivity to the cationic BPI-P2 peptide (8 µg ml⁻¹) was measured by enumerating CFU after 45 min of treatment at 37°C. Input CFU counts were calculated at time zero by plating untreated samples. Percent survival was calculated by dividing the CFU count at 1 h by the input CFU count and normalizing the result to the WT count. (B) Sensitivity to PMB (1 µg ml⁻¹) was determined by enumerating CFU after 1 h of treatment at 37°C. Percent survival was calculated as described for P2. The mean percent survival ± the standard deviation of a minimum of four biological replicates is shown. Significance was determined by paired t test (*, P < 0.05; n.d., not detected).

FIG 6

Rcs-regulated colanic acid capsule production maintains Δψ in stationary phase. Salmonella cultures were grown to stationary phase, and Δψ was measured by flow cytometry with DiOC₂(3) as described in the legend to Fig. 3. Histograms show the red-to-green emission ratio as a measurement of Δψ distribution in a population of 2 × 10⁴ cells. The MFI of each histogram was determined from four biological replicates. WT cells depolarized by CCCP were included as a control. (A) Representative histograms of WT and mutant cells. (B) Replicate MFIs of the strains represented in panel A. (C) Representative histogram showing mutants complemented with pRcsB. (D) Replicate MFIs of strains represented in panel C. Significance was determined by paired t test (*, P < 0.05; **, P < 0.01).

FIG 7

Biofilms formed by WT and mutant Salmonella strains. Strains were added to 96-well PVC microtiter plates containing LB and incubated without agitation at 25°C for 48 h. CV binding to biofilms was quantified by measuring absorbance at 595 nm. A csgD mutant lacking curli and cellulose was included as a biofilm negative control. Biofilm formation in five biological replicates was measured, and the mean values ± the standard deviation are shown. Statistical significance was determined by paired t test (*, P < 0.05; **, P < 0.01).

FIG 8

S. Typhimurium colanic acid capsule mutants show increased sensitivity to ampicillin. Strains were grown for 3 h to logarithmic phase, at which point 200 μg ml⁻¹ ampicillin was added to culture tubes. Cultures were allowed to continue growth in the presence of antibiotic, samples were taken and plated at the time points indicated, and CFU were enumerated after 24 h. Susceptibility was determined by dividing the CFU count at the posttreatment times indicated by the CFU count of cultures immediately before antibiotic addition. The average survival ± the standard deviation of four biological replicates are shown. Significance was determined with a paired t test (*, P < 0.05; **, P < 0.01).