Switching Shiga Toxin (Stx) Type from Stx2d to Stx2a but Not Stx2c Alters Virulence of Stx-Producing Escherichia coli (STEC) Strain B2F1 in Streptomycin (Str)-Treated Mice

Abstract

Shiga toxin (Stx)-producing Escherichia coli (STEC) strain B2F1 produces Stx type 2d, a toxin that becomes more toxic towards Vero cells in the presence of intestinal mucus. STEC that make Stx2d are more pathogenic to streptomycin (Str)-treated mice than most STEC that produce Stx2a or Stx2c. However, purified Stx2d is only 2- or 7-fold more toxic by the intraperitoneal route than Stx2a or Stx2c, respectively. We hypothesized, therefore, that the toxicity differences among Stx2a, Stx2c, and Stx2d occur at the level of delivery from the intestine. To evaluate that hypothesis, we altered the toxin type produced by stx_2d+ mouse virulent O91:H21 clinical isolate B2F1 to Stx2a or Stx2c. Because B2F1 encodes two copies of stx_2d, we did these studies in a derivative of B2F1 in which stx_2d1 was deleted. Although the strains were equivalently virulent to the Str-treated mice at the 10¹⁰ dose, the B2F1 strain that produced Stx2a was attenuated relative to the ones that produced Stx2d or Stx2c when administered at 10³ CFU/mouse. We next compared the oral toxicities of purified Stx2a, Stx2c, and Stx2d. We found that purified Stx2d is more toxic than Stx2a or Stx2c upon oral administration at 4 µg/mouse. Taken together, these studies suggest that Stx2 toxins are most potent when delivered directly from the bacterium. Furthermore, because Stx2d and Stx2c have the identical amino acid composition in the toxin B subunit, our results indicate that the virulence difference between Stx2a and Stx2d and Stx2c resides in the B or binding subunit of the toxins.

Keywords: B2F1; STEC; Shiga toxin (Stx); Shiga toxin type 2c (Stx2c); Shiga toxin type 2d (Stx2d).

Figure 1

Toxicity and activation. The mean cytotoxicity of supernatants from the strains on Vero cells (A) and fold-activation (B) is shown. Fold-activation is the ratio of cytotoxicity of the supernatant after incubation with mouse intestinal mucus or the buffer control. The error bars represent standard deviation. Each point represents an individual replicate. The p values were determined by one-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test and are relative to the B2F1Stx2a (A) or to the B2F1Stx2d2 values (B).

Figure 2

Survival curves in Str-treated BALB/c mice for B2F1Stx2d2, B2F1Stx2a, and B2F1Stx2c. (A) Survival after an inoculum of 10⁹ CFU/mouse. (B) Survival after an inoculum of 10³ CFU/mouse. The survival curves were not statistically different at the 10⁹ dose. At the 10³ dose, the B2F1Stx2a survival curve was different than both the B2F1Stx2d2 and B2F1Stx2c survival curves, p = 0.005 or p = 0.0008, respectively. n = 5 (B2F1Stx2d2 and B2F1Stx2c) or 10 (B2F1Stx2a) mice.

Figure 3

The CFU (A) and CD₅₀/g feces (B) from mice inoculated with 10³ CFU of the indicated strains as measured on days 1 and 3 post-infection. The CD₅₀/g feces were higher on day 3 as compared to day 1 for both B2F1Stx2d2 and B2F1Stx2c (p = 0.001), though the CFU/g feces did not change. Each symbol represents an individual mouse. n = 5 (B2F1Stx2d2 and B2F1Stx2c) or 10 (B2F1Stx2a) mice. Data were compared by two-way ANOVA on the log-transformed values.