Comparative analysis of EspF variants in inhibition of Escherichia coli phagocytosis by macrophages and inhibition of E. coli translocation through human- and bovine-derived M cells

Abstract

The EspF protein is secreted by the type III secretion system of enteropathogenic and enterohemorrhagic Escherichia coli (EPEC and EHEC, respectively). EspF sequences differ between EHEC O157:H7, EHEC O26:H11, and EPEC O127:H6 in terms of the number of SH3-binding polyproline-rich repeats and specific residues in these regions, as well as residues in the amino domain involved in cellular localization. EspF(O127) is important for the inhibition of phagocytosis by EPEC and also limits EPEC translocation through antigen-sampling cells (M cells). EspF(O127) has been shown to have effects on cellular organelle function and interacts with several host proteins, including N-WASP and sorting nexin 9 (SNX9). In this study, we compared the capacities of different espF alleles to inhibit (i) bacterial phagocytosis by macrophages, (ii) translocation through an M-cell coculture system, and (iii) uptake by and translocation through cultured bovine epithelial cells. The espF gene from E. coli serotype O157 (espF(O157)) allele was significantly less effective at inhibiting phagocytosis and also had reduced capacity to inhibit E. coli translocation through a human-derived in vitro M-cell coculture system in comparison to espF(O127) and espF(O26). In contrast, espF(O157) was the most effective allele at restricting bacterial uptake into and translocation through primary epithelial cells cultured from the bovine terminal rectum, the predominant colonization site of EHEC O157 in cattle and a site containing M-like cells. Although LUMIER binding assays demonstrated differences in the interactions of the EspF variants with SNX9 and N-WASP, we propose that other, as-yet-uncharacterized interactions contribute to the host-based variation in EspF activity demonstrated here.

Fig. 1.

Comparison of EHEC O157, EHEC O26, and EPEC O127 interactions with RAW 264.7 macrophages. Confluent monolayers of the mouse macrophages were infected with EPEC O127:H6 (E2348/69), EHEC O26:H11 (ZAP1139), and EHEC O157:H7 (ZAP1163) strains, and the number of bacteria inside (green) or outside (red) of the macrophages was determined by fluorescence microscopy as detailed in Materials and Methods. (A) Mean (± the 95% confidence intervals) numbers of adherent bacteria per macrophage at 30 min after addition of the bacteria at an MOI of 100. (B) Percentage (± the 95% confidence intervals) of intracellular bacteria at the time points shown as determined by fluorescence microscopy for EHEC O157 (•), EPEC O127 (♦), and EPEC O26 (▴). Significant differences of <0.001 for O157 versus EPEC (***), O157 versus O26 (+++), and EPEC (xxx) versus EHEC O26 are indicated. (C) Confocal images were acquired using a Zeiss Plan Apochromat 1.4 NA ×63 oil immersion lens and a multitrack (sequential scan) experimental set up on a Zeiss LSM510. Scale bar, 10 μm.

Fig. 2.

Alignment of EspF amino acid sequences from EHEC O157:H7 (EDL933); EPEC O127:H6 (E2348/69), and EHEC O26:H11 (ZAP1139). The proline-rich repeats (PRRs) are boxed with EHEC O157 containing an additional fourth repeat. Amino acid differences from the EHEC O157 sequence are indicated in boldface. A binding site for SNX9 is highlighted in gray within the PRRs (1). The putative N-WASP binding region is within the middle of the PRRs (1), although the most significant sequence diversity between these variants is at the ends of each PRR. The leucine at position 16 (arrow) has been shown to be essential for EspF translocation into mitochondria, and this is changed to the similar aliphatic amino acid isoleucine in EHEC O26 (58). Analysis of NCBI E. coli O157:H7 sequences showed no significant variation in the predicted EspF amino acid sequence. The predicted EspF from an E. coli O26:H2 strain was identical to that shown for E. coli O26:H11 (data not shown).

Fig. 3.

Inhibition of phagocytosis (mean ± the 95% confidence intervals) by different espF alleles. (A) A confluent monolayer of RAW 264.7 macrophages was infected with a panel of GFP-labeled ΔespF EPEC strains transformed with espF cloned from EPEC O127:H6, EHEC O157:H7, and EHEC O26:H11. The proportions of extracellular bacteria were determined at 90 min. The infected cells were fixed in 4% PFA. Bacteria were stained with O-antigen-specific antibody detected with Alexa Fluor 594-conjugated secondary antibody. The intracellular versus total bacteria were imaged and quantified as described in Materials and Methods. ***, P < 0.001. (B) Levels of EspD and EspF secretion by EPEC O127 ΔespF and then complemented with the respective three espF alleles as indicated. Supernatants were prepared and separated in a standard sodium dodecyl sulfate denaturing gel that was then stained with colloidal blue. The bottom panel shows the staining for the main EspD/B band. The middle panel shows the same samples with detection for EspD. The top panel shows the detection of EspF. The supernatants were prepared from equal volumes of bacteria (50 ml) cultured to an OD₆₀₀ of 0.8. Experimental details are given in Materials and Methods.

Fig. 4.

Translocation of EPEC and EHEC strains across a Caco-2 and lympho-epithelial M-cell coculture system (black columns [A, C, and E]) and Caco-2 cells only (white columns [B, D, and F]). (A and B) espF is required to inhibit EHEC O157 TUV93-0 translocation through M cells. The translocation of EHEC O157 TUV93-0 was compared to an isogenic espF deletion mutant and complemented with espF_O157 (pAT1, Table 2). (C and D) Comparative translocation of EPEC O127:H6 strain E2348/69, EHEC O157:H7 ZAP198, and EHEC O26:H11 across the coculture system. (E and F) Comparative translocation of an EPEC O127ΔespF mutant complemented with the three defined espF alleles. *, **, and ***, statistical significances of P < 0.05, P < 0.01, and P < 0.001, respectively.

Fig. 5.

Characterization of bovine primary rectal epithelial cells. (A) Heterogeneous population of epithelial cells in a primary cell culture from the bovine terminal rectum. A 5-day-old culture was prepared and immunolabeled to detect vimentin (green), pan-cytokeratins (red), and nuclei (blue) as described in Materials and Methods. A subset of the cells expressed the intermediate filament protein, vimentin (green), which is indicative of M cells. (B) Vimentin-expressing cells were quantified by using flow cytometry from four independent primary cultures (standard deviation, ±0.3). (C) Microparticle colocalization with vimentin-expressing cells. A subset of vimentin expressing (red) cells interacted with fluorescent microparticles (green). Panel C1 shows the inset image in panel C digitally magnified by a factor of 4. The primary rectal epithelial cells were incubated with latex particles (green, 0.2 μm in size) at 37°C and 5% CO₂ for 45 min, fixed, permeabilized, and labeled with anti-vimentin (red) and TO-PRO nuclear stain (blue). (D) Orthogonal section demonstrating S. enterica serovar Typhimurium uptake by vimentin-expressing cells during the early stages of interaction. The cells were infected with mid-log-phase S. Typhimurium (pUC18GFP-labeled SL1344 strain) at an MOI of 1:100 at 37°C and 5% CO₂ for 10 min. (E) Orthogonal section demonstrating the combined uptake of S. Typhimurium (white arrow, DAPI stained) and microparticles (yellow arrow, green) by vimentin-positive cells (red) in a bovine rectal primary culture. The cells were incubated with latex particles (green) for 45 min, washed (3× PBS), and further infected with mid-log-phase S. Typhimurium at an MOI of 1:100 for 10 min. The infected cells were fixed and labeled with anti-vimentin (red) and DAPI nuclear stain. Confocal images were acquired by using a Zeiss LSM510 with a ×63 objective lens). Scale bar, 10 μm.

Fig. 6.

Interaction of EHEC and EPEC strain with cultured epithelial cells from the bovine terminal rectum. (A) espF limits EHEC O157 TUV93-0 uptake into rectal primary cells. The transcytosis levels (%) of EHEC O157 TUV93-0 were compared to an isogenic espF deletion mutant and complemented with espFO157 (pAT1, Table 2).(B) Comparative transcytosis levels (%) of wild-type strains EPEC O127:H6 E2348/69, EHEC O157:H7 TUV93-0, and EHEC O26:H11 on interaction with bovine rectal primary cells. (C) Comparative transcytosis levels (%) of an EPEC O127ΔespF mutant complemented with the three defined espF alleles. * and ***, statistical significances of P < 0.05 and P < 0.001, respectively.

Fig. 7.

Interaction of EHEC and EPEC strain with cultured epithelial cells from the bovine terminal rectum. (A) espF limits EHEC O157 TUV93-0 uptake into rectal primary cells. The intracellular levels of EHEC O157 TUV93-0 were compared to an isogenic espF deletion mutant and complemented with espF_O157 (pAT1, Table 2). (B) Comparative intracellular levels of wild-type strains EPEC O127:H6 E2348/69, EHEC O157:H7 TUV93-0, and EHEC O26:H11 on interaction with bovine rectal primary cells. (C) Comparative intracellular levels of an EPEC O127ΔespF mutant complemented with the three defined espF alleles. * and ***, statistical significances of P < 0.05 and P < 0.001, respectively.

Fig. 8.

Comparative binding analysis between EspF variants and human SNX9 or N-WASP. In the LUMIER binding assay, the espF alleles were expressed with a protein A tag and then immobilized on immunoglobulin beads. The interaction between proto-oncogenes, c-Fos and c-Jun, was used as a positive control. SNX9 and N-WASP were expressed with Renilla luciferase tags and incubated with the immobilized EspF variants. The luminescence signals were determined and normalized against negative controls, and z-scores were calculated as described previously (4, 9) and in Materials and Methods.