Association of Vibrio cholerae 569B outer membrane vesicles with host cells occurs in a GM1-independent manner

Abstract

The primary virulence factor of Vibrio cholerae, cholera toxin (CT), initiates a pathway in epithelial cells that leads to the severe diarrhoea characteristic of cholera. Secreted CT binds to GM1 on the surface of host cells to facilitate internalisation. Many bacterial toxins, including CT, have been shown to be additionally delivered via outer membrane vesicles (OMVs). A fraction of the closely related heat labile toxin produced by enterotoxigenic Escherichia coli has been demonstrated to reside on the surface of OMVs, where it binds GM1 to facilitate OMV internalisation by host cells. In this work, we investigated whether OMV-associated CT is likewise trafficked to host cells in a GM1-dependent mechanism. We demonstrated that a majority of CT is secreted in its OMV-associated form and is located exclusively inside the vesicle. Therefore, the toxin is unable to bind GM1 on the host cell surface, and the OMVs are trafficked to the host cells in a GM1-independent mechanism. These findings point to a secondary, noncompeting mechanism for secretion and delivery of CT, beyond its well-studied secretion via a Type II secretion system and underscore the importance of focusing future studies on understanding this GM1-independent delivery mechanism to fully understand Vibrio cholerae pathogenesis.

FIGURE 1. Characterization of V. cholerae OMVs

(A) Scanning electron micrographs of V. cholerae 569B OMVs. (B) Western blot analysis of V. cholerae 569B OMVs. Two strong anti-CT polyclonal antibody-reactive bands were detected, demonstrating the presence of CT in the OMVs. (C) Relative CT concentration in 569B OMVs and OMV-free supernatant. The mass of CT present in purified OMVs or the OMV-free supernatant was determined by comparing the intensity of an antibody-reactive spot to that of CT standards with known mass.

FIGURE 2. V. cholerae 569B OMV-mediated cytotoxicity

(A) Representative phase contrast image of OMV-treated FHs 74 intestinal epithelial cells. (i) Untreated cells and (ii) OMV-treated cells (0.4 ng/μL of CT, 4 hr). Enlarged view of the regions in small boxes are shown in the insets. (B) Quantitative analysis of the OMV-treated cell morphology. Cells were treated with a serial dilution of 569B OMVs (5 to 40 ng CT/well), and the resulting morphology scored on a scale of 1 (spindly) to 4 (rounded), based on the percentage of cell rounding. Data are expressed as mean ± SD (N=3). One-way ANOVA followed by Bonferroni’s post hoc test was used to compare differences between treated and untreated samples. * p < 0.05, ** p < 0.01, and *** p < 0.001. (C) Quantification of 569B OMV- and OMV-free supernatant-mediated cytotoxicity. Cells were treated with purified OMVs or OMV-free supernatant. Control cells were treated with HBSS. OMV-mediated alterations in cell morphology were scored on a scale of 1 (spindly) to 4 (rounded). Data are expressed as mean ± SD (N=3). The level of significance was determined between cells treated with OMVs and OMV-free supernatant using an unpaired two-tailed t-test. *** p < 0.001.

FIGURE 3. Characterization of soluble CT association with OMV lipids

(A) Lipid dot blot analysis of CT affinity for (i) POPE, POPG, TOCL, 569B LPS and GM1, and (ii) POPE, POPG, TOCL, and 569B LPS (with no GM1). In the presence of GM1, free CT was found to have strong affinity for GM1, and in absence of GM1, it had some affinity for POPE. In both cases no affinity for 569B LPS was seen. (B) SDS-PAGE analysis of purified LPS from V. cholerae 569B. Lane 1 is the LPS standard and lane 2 is 569B LPS.

FIGURE 4. Localization of OMV-associated CT

(A) Proteinase K digestion of 569B OMVs. Lane 1: untreated OMVs. Lane 2: Proteinase K-treated OMVs. Lane 3: SDS- and proteinase K-treated OMVs. (B) Indirect noncompetitive ELISA of intact (squares) and EDTA-treated (triangles) 569B OMVs. EDTA-disrupted OMVs bound to the anti-CT polyclonal antibody-coated plate, but intact OMVs did not. Data are shown as mean + SD (N=3). The level of significance was determined using one-way ANOVA followed by Bonferroni’s multiple comparison test showing a statistically significant difference between intact and EDTA-treated 569B OMVs. ** p < 0.01 and *** p < 0.001.

FIGURE 5. 569B OMV association with GM1

(A) GM1 ELISA to measure soluble CT binding to GM1. Soluble CT (open circles) binds to coated GM1; CT pretreated with GM1 (filled circles) does not bind GM1. Data are shown as mean ± SD (N=3). The level of significance was determined using one-way ANOVA followed by Bonferroni’s multiple comparison test. (B) GM1 ELISA to measure OMV binding to GM1. Neither OMVs (open squares) nor GM1-treated OMVs (filled squares) bind GM1. Data are expressed as mean ± SD (N=3). One-way ANOVA analysis indicated no significant difference between untreated and GM1-treated OMVs. (C) Phase contrast image of FHs 74 intestinal epithelial cells treated with OMVs. (i) Untreated OMVs and (ii) GM1-pretreated OMVs (0.4 ng/μL of CT, 4 hr). Both untreated and GM1-treated OMVs induce cell rounding. (D) Quantitative analysis of the morphology of cells treated with untreated OMVs or OMVs pre-treated with GM1. Data are expressed as mean ± SD (N=3). No significant difference was observed between two samples as determined by an unpaired two-tailed t-test.

FIGURE 6. GM1-dependence of cellular association of OMVs

FHs 74 intestinal epithelial cells were labeled with AF555-WGA (red) and 569B OMVs were labeled with FITC (green). Both untreated (top) and GM1-treated OMVs (bottom) are associated with host cells within two hrs. The scale bar in all the figures is 5 μm.

FIGURE 7. Schematic of proposed mechanism of CT association with 569B OMVs

Our results demonstrate that CT has no affinity for LPS, which comprises the outer leaflet of the OM, and resides entirely within the lumen of the OMVs. We therefore hypothesize that during T2S, some CT remains in the periplasm, allowing it to become incorporated in the OMV lumen. Unlike soluble CT, OMV-associated CT does not bind to the GM1 receptor in the lipid rafts of the host cell membrane because of the toxin’s location within the OMV. We propose that CT secretion via OMVs and through the T2S system are therefore complementary mechanisms for CT delivery.