Loss of very-long O-antigen chains optimizes capsule-mediated immune evasion by Salmonella enterica serovar Typhi

Abstract

Expression of capsular polysaccharides is a variable trait often associated with more-virulent forms of a bacterial species. For example, typhoid fever is caused by the capsulated Salmonella enterica serovar Typhi, while nontyphoidal Salmonella serovars associated with gastroenteritis are noncapsulated. Here we show that optimization of the immune evasive properties conferred by the virulence-associated (Vi) capsular polysaccharide involved an additional alteration to the cell envelope of S. Typhi, namely inactivation of the fepE gene, encoding the regulator of very-long O-antigen chains. Introduction of the capsule-encoding viaB locus into the nontyphoidal S. enterica serovar Typhimurium reduced complement deposition in vitro and intestinal inflammation in a mouse colitis model. However, both phenotypes were markedly enhanced when the viaB locus was introduced into an S. Typhimurium fepE mutant, which lacks very-long O-antigen chains. Collectively, these data suggest that during the evolution of the S. Typhi lineage, loss of very-long O-antigen chains by pseudogene formation was an adaptation to maximize the anti-inflammatory properties of the Vi capsular polysaccharide.

Importance: Genomic comparison illustrates that acquisition of virulence factors by horizontal gene transfer is an important contributor to the evolution of enteric pathogens. Acquisition of complex virulence traits commonly involves horizontal transfer of a large gene cluster, and integration of the gene cluster into the host genome results in the formation of a pathogenicity island. Acquisition of the virulence-associated (Vi) capsular polysaccharide encoded by SPI7 (Salmonella pathogenicity island 7) was accompanied in the human-adapted Salmonella enterica serovar Typhi by inactivation of the fepE gene, encoding the regulator of very-long O-antigen chains. We show that the resulting loss of very-long O-antigen chains was an important mechanism for maximizing immune evasion mediated by the Vi capsular polysaccharide. These data suggest that successful incorporation of a capsular polysaccharide requires changes in the cell envelope of the hosting pathogen.

FIG 1

(A to C) Fixation of C3 after incubation of the indicated S. Typhi and S. Typhimurium strains (wild type and mutants) in 10% human serum was detected by flow cytometry using an anti-human C3 (α-human C3) FITC conjugate. The experiments shown in panels A and C were repeated 3 times independently with similar outcomes, and a representative example is shown. The average maximum fluorescence intensity (MFI) values ± standard errors (error bars) determined for these three independent experiments are shown in panel B. (D) Silver-stained SDS-PAGE of LPS preparations from the indicated S. Typhimurium and S. Typhi strains. The positions of short, long, and very-long O-antigen chains are indicated to the left of the gel. A magnification of the region showing long and very-long O-antigen chains is shown on the right, and the presence of very-long O-antigen chains in S. Typhimurium strains is indicated by black arrows.

FIG 2

(A) Silver-stained SDS-PAGE of LPS preparations from the indicated S. Typhimurium and S. Typhi strains (wild type and mutants). Plasmid pRC37 carries the cloned S. Typhimurium fepE gene. The positions of very-long O-antigen chains are indicated by black arrows. (B) Vi capsular polysaccharide expression was detected by flow cytometry. Cells of the indicated S. Typhi and S. Typhimurium strains were labeled with rabbit anti-Vi serum/goat anti-rabbit FITC conjugate (α-Vi on the y axis), and fluorescence intensities were determined for 10,000 particles. Each experiment was repeated 3 times independently with similar outcomes, and a representative example is shown. (C) Fixation of C3 after incubation of the indicated S. Typhi and S. Typhimurium strains in 10% human serum was detected by flow cytometry using an anti-human C3 FITC conjugate. The experiment was repeated 3 times independently with similar outcomes, and a representative example is shown.

FIG 3

Streptomycin-pretreated mice were infected with the indicated S. Typhimurium strains, and the cecum and colon contents were collected 72 h after infection. (A) Recovery of S. Typhimurium from colon contents. Values are geometric means of CFU per gram colon contents ± standard errors (error bars). (B to F) Transcript levels of Ifng (B), Tnfa (C), Il22 (D), Kc (E), and Mip2 (F) in the cecal mucosa were determined by quantitative real-time PCR. Values are geometric means ± standard errors of fold increases over mRNA levels in mock-infected animals. Values that are statistically significant (P < 0.05) are indicated by a bar and asterisk. Values that are not statistically significant (ns) are indicated.

FIG 4

Streptomycin-pretreated mice were either mock infected or infected with the indicated S. Typhimurium strains, and the cecum was collected 72 h after infection. (A) Combined histopathology score of pathological changes observed in sections from the cecum. Each symbol represents the combined histopathology score for an individual animal. The average for each group of mice is indicated by a short line. (B) Representative images of histopathological changes.