CRISPR Screen Reveals that EHEC's T3SS and Shiga Toxin Rely on Shared Host Factors for Infection

Abstract

Enterohemorrhagic Escherichia coli (EHEC) has two critical virulence factors-a type III secretion system (T3SS) and Shiga toxins (Stxs)-that are required for the pathogen to colonize the intestine and cause diarrheal disease. Here, we carried out a genome-wide CRISPR/Cas9 (clustered regularly interspaced short palindromic repeats with Cas9) loss-of-function screen to identify host loci that facilitate EHEC infection of intestinal epithelial cells. Many of the guide RNAs identified targeted loci known to be associated with sphingolipid biosynthesis, particularly for production of globotriaosylceramide (Gb3), the Stx receptor. Two loci (TM9SF2 and LAPTM4A) with largely unknown functions were also targeted. Mutations in these loci not only rescued cells from Stx-mediated cell death, but also prevented cytotoxicity associated with the EHEC T3SS. These mutations interfered with early events associated with T3SS and Stx pathogenicity, markedly reducing entry of T3SS effectors into host cells and binding of Stx. The convergence of Stx and T3SS onto overlapping host targets provides guidance for design of new host-directed therapeutic agents to counter EHEC infection.IMPORTANCE Enterohemorrhagic Escherichia coli (EHEC) has two critical virulence factors-a type III secretion system (T3SS) and Shiga toxins (Stxs)-that are required for colonizing the intestine and causing diarrheal disease. We screened a genome-wide collection of CRISPR mutants derived from intestinal epithelial cells and identified mutants with enhanced survival following EHEC infection. Many had mutations that disrupted synthesis of a subset of lipids (sphingolipids) that includes the Stx receptor globotriaosylceramide (Gb3) and hence protect against Stx intoxication. Unexpectedly, we found that sphingolipids also mediate early events associated with T3SS pathogenicity. Since antibiotics are contraindicated for the treatment of EHEC, therapeutics targeting sphingolipid biosynthesis are a promising alternative, as they could provide protection against both of the pathogen's key virulence factors.

Keywords: CRISPR screen; EHEC; EPEC; LAPTM4A; Shiga toxin; T3SS; TM9SF2; host susceptibility; sphingolipid synthesis.

FIG 1

Design of a CRISPR/Cas9 screen to identify host factors underlying susceptibility to EHEC infection. (A) Schematic of the infection and outgrowth process for an HT-29 CRISPR/Cas9 library undergoing multiple rounds of infection with ΔespZ EHEC, which has an active T3SS and secretes Stx1 and Stx2. (B and C) Abundance of HT29 cells infected with the indicated strain relative to the abundance of mock-infected cells at day 1 (B) and day 5 (C) postinfection. Graphs display the mean and standard deviation (SD) from 3 independent experiments. **, P < 0.01; ****, P < 0.0001.

FIG 2

Mutations that disrupt sphingolipid biosynthesis and poorly characterized genes are enriched in the HT-29 CRISPR/Cas9 library following repeated infection with espZ EHEC. (A) Scatterplot of the statistical significance in each library (A and B) associated with the genes ranked in the top 5% by the STARS algorithm. Genes with a P value of ≤0.001 in both libraries (upper right quadrant) are named; genes within the ellipse all have P values of <2.0e−06. (B) Products of genes shown in panel A with P ≤ 0.001 in both libraries and schematic representation depicting the subcellular localization of enzymes (black) that contribute to sphingolipid biosynthesis. A subset of substrates/products is depicted in red. (C) Abundance of HT29 control and mutant cells infected with ΔespZ EHEC relative to the abundance of mock-infected cells at day 5 postinfection. Graphs display the mean and SD from 3 independent experiments compared to HT-29 Cas9 (leftmost bar). *, P < 0.05; **, P < 0.005; ***, P < 0.001; ****, P < 0.0001.

FIG 3

Disruption of host sphingolipid biosynthesis genes and poorly characterized genes reduces the activity and cytotoxicity of EHEC’s T3SS. (A) Abundance of control and mutant HT29 Cas9 cells infected with ΔespZ Δstx1 Δstx2 EHEC relative to the abundance of mock-infected cells at day 1 postinfection. Graphs display the mean and SD from 3 independent experiments. P values were obtained from one-way ANOVA with Dunnett’s postcorrection (*, P < 0.02; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001). (B) Relative translocation of Tir-CyA from wt EHEC into HT-29 Cas9 control cells and the indicated HT-29 mutants based on cAMP levels. Translocation into HT-29 Cas9 control cells was set as 100%. Data reflect the mean and SD from 3 independent experiments. P values (****, P < 0.0001) are based on one-way ANOVA with Dunnett’s postcorrection. (C) Confocal microscopy of control and mutant HT-29 Cas9 cells infected for 6 h with GFP-EHEC and then stained for F-actin with Alexa 647-phalloidin (pink) and DAPI (blue [labels nuclei]). Merged images are shown. Focal colocalization of bacteria and actin reflects formation of actin pedestals. White boxes show enlarged images to highlight pedestals. (D) Percentage of the indicated host cells with actin pedestals 6 h after infection. Two hundred fifty cells were assessed for each host genotype. (E) Number of pedestals per host cell. Box plots show the range (minimum to maximum) of pedestal numbers. One hundred cells with AE lesions were counted per genotype. ****, P < 0.0001.

FIG 4

TM9SF2 and LAPTM4A promote sensitivity to Stx. (A) Abundance of wt and mutant HT29 Cas9 cells infected with T3SS-deficient EHEC (ΔespZ ΔescN) relative to the abundance of mock-infected cells at day 5 postinfection. P values are based on one-way ANOVA with Dunnett’s posttest correction (****, P < 0.0001). (B) Flow cytometry analysis of Stx2-Alexa 647 binding to wt and mutant HT-29 Cas9 cells. Histograms show the distribution of fluorescence intensity in the total cell population in the presence and absence of toxin. (C and D) Confocal microscopy of Stx2-Alexa 488 (green) binding to nonpermeabilized (C) and permeabilized (D) control and mutant HeLa Cas9 cells. Cells were also stained with DAPI and Alexa 568-phalloidin.

FIG 5

Subcellular localization of TM9SF2, LAPTM4A, and A4GALT in wt and mutant HeLa cells. (A) Confocal immunofluorescence microscopy of HeLa cells stained with anti-TM9SF2 (green), anti-GM130 to label the Golgi complex (pink) and DAPI. For LAPTM4A localization, HeLa cells were transfected with GFP-tagged LAPTM4A, which was imaged directly after counterstaining as described above. (B) Confocal immunofluorescence microscopy of control and mutant HeLa Cas9 cells labeled with anti-A4GALT antibody (pink), anti-58K to label the Golgi complex (pink) and DAPI. GM130 and 58K stain similar populations of Golgi complex membranes and were used interchangeably to accommodate the primary antibodies of interest.