Single domain antibodies against enteric pathogen virulence factors are active as curli fiber fusions on probiotic E. coli Nissle 1917

Abstract

Enteric microbial pathogens, including Escherichia coli, Shigella and Cryptosporidium species, take a particularly heavy toll in low-income countries and are highly associated with infant mortality. We describe here a means to display anti-infective agents on the surface of a probiotic bacterium. Because of their stability and versatility, VHHs, the variable domains of camelid heavy-chain-only antibodies, have potential as components of novel agents to treat or prevent enteric infectious disease. We isolated and characterized VHHs targeting several enteropathogenic E. coli (EPEC) virulence factors: flagellin (Fla), which is required for bacterial motility and promotes colonization; both intimin and the translocated intimin receptor (Tir), which together play key roles in attachment to enterocytes; and E. coli secreted protein A (EspA), an essential component of the type III secretion system (T3SS) that is required for virulence. Several VHHs that recognize Fla, intimin, or Tir blocked function in vitro. The probiotic strain E. coli Nissle 1917 (EcN) produces on the bacterial surface curli fibers, which are the major proteinaceous component of E. coli biofilms. A subset of Fla-, intimin-, or Tir-binding VHHs, as well as VHHs that recognize either a T3SS of another important bacterial pathogen (Shigella flexneri), a soluble bacterial toxin (Shiga toxin or Clostridioides difficile toxin TcdA), or a major surface antigen of an important eukaryotic pathogen (Cryptosporidium parvum) were fused to CsgA, the major curli fiber subunit. Scanning electron micrographs indicated CsgA-VHH fusions were assembled into curli fibers on the EcN surface, and Congo Red binding indicated that these recombinant curli fibers were produced at high levels. Ectopic production of these VHHs conferred on EcN the cognate binding activity and, in the case of anti-Shiga toxin, was neutralizing. Taken together, these results demonstrate the potential of the curli-based pathogen sequestration strategy described herein and contribute to the development of novel VHH-based gut therapeutics.

Fig 1. Schematic overview of bacterial virulence factors used as VHH targets in this study.

Fig 2. Anti-Fla VHHs can inhibit REPEC motility.

(a) Representative images of REPEC growth on motility plates after incubation with varying VHH concentrations with spread diameter indicated (scale bar = 1 cm). (b) Relative motility as a function of αFla VHH concentration. Spread diameters were normalized to a “no VHH” control.

Fig 3. Anti-Tir and anti-intimin VHHs interfere with Tir-intimin binding and pedestal formation.

(a) Several anti-Tir VHHs blocked Tir-intimin binding. Data presented as mean ± SEM. P-values calculated by one-way ANOVA. ** P < 0.01; *** P < 0.001, **** P < 0.0001. (b-j) Anti-Tir and anti-intimin VHHs can inhibit EPEC-mediated pedestal formation (scale bar = 100 μm). HeLa cells were exposed to EPEC incubated with VHH, fixed, and stained with DAPI (blue) and Alexa Fluor-488 Phalloidin (green). Blinded pedestal scores were assigned by a separate researcher.

Fig 4. Formation of VHH-functionalized curli fibers by EcN-derived strain PBP8.

(a) Plasmid map of synthetic curli operon and domain map of CsgA-VHH fusion constructs. (b) Idealized schematic of curli fiber export and self-assembly mechanism. (c) Representative scanning electron micrograph demonstrating the formation of curli fibers by CsgA-VHH-expressing PBP8 (CsgA-αgp900-2, scale bar = 1 μm). (d) Congo Red (CR) binding assays of induced cultures of PBP8 bearing plasmids encoding various CsgA-VHH constructs. CR binding is indicative of the formation of curli protein fibers. Data corresponds to A₄₉₀ measurements normalized to the CsgA positive control, and is presented as mean ± SEM. One-way ANOVA (P < 0.0001) was performed to test presence of difference between conditions. All CsgA constructs exhibited P-values < 0.0001 compared to the “No CsgA” control, as calculated by Welch’s t-test.

Fig 5. CsgA-VHHs can bind soluble proteins.

(a) GFP pull-down assay. After incubation with 150 nM GFP, PBP8 CsgA-αGFP bound specifically to its target, as demonstrated by the depletion GFP in the cell supernatant. (b) Stx2 pull-down assay. PBP8 CsgA-αStx2 was used to selectively remove Stx2 upon incubation in a 10 ng/mL Stx2 solution. Supernatants were added to Vero cell monolayers and cell viability was measured. Data presented as mean ± SEM. One-way ANOVA (P < 0.0001) was performed to test presence of difference between conditions, P-values calculated by Welch’s t-test. ** P < 0.01; **** P < 0.0001.

Fig 6. CsgA-VHHs can neutralize S. flexneri contact-mediated hemolysis.

PBP8 expressing CsgA-αIpaD-2, but not CsgA-αIpaD-1, inhibited hemolysis of sheep red blood cells, significantly outperforming the off-target negative control (CsgA-αGFP). Data presented as mean ± SEM. Two-way ANOVA (P < 0.0001) was performed to test the presence of difference between conditions, P-values calculated by Welch’s t-test. * P < 0.05; ** P < 0.01.

Fig 7. CsgA-VHHs can bind to the eukaryotic pathogen Cryptosporidium parvum.

(a) ELISA demonstrating the ability of CsgA-αgp900-2 to bind gp900. PBP8 was adsorbed onto a well plate, followed by incubation with gp900-containing C. parvum lysate. In the positive control, no PBP8 was used, and the C. parvum lysate was allowed to directly adsorb onto the surface of the wells. Gp900, either bound to PBP8 or adsorbed to the surface, was detected using a non-competing anti-gp900 VHH, followed by an anti-Etag IgG-HRP conjugate. (b) Western blot analysis of gp900 binding to CsgA-αgp900-2. C. parvum lysate was incubated with PBP8 at different concentrations. After pelleting the PBP8, supernatant form was run on gel and subjected to Western blot to assess gp900 depletion. The gp900 band was detected by a specific non-competing VHH, followed by an anti-Etag IgG-HRP conjugate. Band intensity was normalized to a CsgA-αGFP negative control. (c) CsgA-αgp900-1 and -2 bind to C. parvum sporozoites. PBP8 was applied to immobilized and fixed sporozoites, followed by staining with anti-LPS Mab and anti-mouse IgG Alexa Fluor 568. Slides were inspected under TRITC filter and 5 images were taken under 200x magnification for each condition. Foci of PBP8 accumulation were quantified using ImageJ particle analyzer. Data presented as mean ± SEM. Two-way ANOVA (P < 0.0001) was performed to test presence of difference between conditions, P-values calculated by Welch’s t-test. * P < 0.05; ** P < 0.01; *** P < 0.001.