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.

INTRODUCTION

Enterohemorrhagic Escherichia coli (EHEC) is an emerging zoonotic pathogen, particularly in industrialized countries (6). EHEC strains cause sporadic outbreaks of severe disease in humans, the most important being hemorrhagic colitis (HC) and hemolytic-uremic syndrome (HUS); the latter disease results in kidney damage and may lead to death (8, 28). Shiga toxins (Stx) produced by EHEC strains are the main factors responsible for these serious outcomes in humans. In contrast, Enteropathogenic E. coli (EPEC) is another pathogenic type of E. coli that can also cause severe intestinal disease in humans, although these infections are not usually associated with HC and HUS since these strains do not produce Shiga toxins. Unlike EHEC, there is no clear evidence that EPEC strains are zoonotic, although strains do circulate and cause diseases in animals (56).

Our understanding of EHEC pathogenesis is primarily based on studies of the EHEC O157 and EHEC O26 serogroups that are associated with most human EHEC infections in Europe, North America, and Japan (39, 40, 41). Both serogroups are considered to be present in ruminants, in particular cattle as the primary reservoir (5, 41, 59, 60). Although there are many EPEC serotypes, extensive research has been carried out on the sequenced human EPEC O127 strain E2348/69. EHEC and EPEC strains express a type III secretion system (T3SS) that is important for colonization of the human or animal host (33, 36, 56, 74). The T3SS injects effector proteins into host cells that manipulate cellular processes to promote the colonization and persistence of the bacterium in the gastrointestinal tract (16, 19, 20, 34, 52, 64). The primary phenotype associated with the T3SS is intimate attachment between the bacterial outer membrane protein intimin and the T3SS translocated intimin receptor (Tir) (42). In both EHEC and EPEC, the genes encoding this protein secretion system are expressed from the locus of enterocyte effacement (LEE) pathogenicity island (33, 36). Although several effector proteins are also expressed from the LEE, a number of additional secreted effector proteins have been identified that are expressed primarily from integrated phage elements scattered throughout the O157 chromosome (79). EHEC and EPEC strains have different combinations of effector proteins, potentially reflecting host adaptation and differences in pathogenesis.

EspF is a LEE-encoded effector protein that requires the CesF chaperone to be translocated by the T3SS into host cells (20). EspF has multiple proline-rich domains that act by binding to SH3 domains or enabled/VASP homology 1 (EVH1) domains of host cell signaling proteins (15, 55). For example, EPEC EspF_O127 binds to sorting nexin 9 (SNX9) via its SH3 amino-terminal region (1, 51). EspF is involved in disruption of tight junctions and increases monolayer permeability in part through the redistribution of occludins (54, 80). EspF sequences differ between EPEC and EHEC strains, and the EHEC O157 variant has a more modest impact on transepithelial electrical resistance (TER) (80). EspF in combination with other effectors inhibits the water transporter SGLT-1, highlighting the importance of effector interplay (16, 43). EPEC EspF_O127 is targeted to mitochondria with the N-terminal region of EspF functioning as an import signal. EPEC EspF_O127 causes an increase in mitochondrial membrane permeabilization in addition to the release of cytochrome c from mitochondria into the cytoplasm and subsequent caspase-9 and caspase-3 cleavage, leading to cell death (15, 58, 65, 66). More recent work has demonstrated that EspF can lead to loss of nucleolin from the nucleolus, an activity driven by EspF's activity on mitochondria (17). EPEC EspF_O127 also plays an important role in inhibition of bacterial uptake by macrophages (70), preventing macrophage phagocytosis via inhibition of the phosphatidylinositol 3-kinase (PI3K)-dependent pathway of bacterial uptake (11, 70).

Intestinal epithelium is composed of multiple cell types, including absorptive enterocytes, enteroendocrine, goblet, and Paneth cells. These cells derive through asymmetrical division migration and differentiation from pluripotent stem cells. An additional specialized epithelial cell type, termed M cells (i.e., “membranous” or “microfold” cells), are associated particularly with epithelium overlying gut-associated lymphoid tissue. This is referred to as follicle-associated epithelium (FAE) and is a site of active immunological function. In contrast to villous epithelium, FAE contains no or fewer goblet cells (67), defensin- and lysozyme-producing Paneth cells (26, 27) and expresses low amounts of membrane-associated hydrolases (68). The M cells generally lack the distinct microvilli and thick filamentous brush border glycocalyx (24) and instead have variable microfolds. Together, these features of M cells promote contact of antigens with gut epithelium and result in sampling of antigens from the intestinal lumen and transfer to antigen-presenting cells within an intraepithelial pocket (62) on its basolateral side (21, 22, 35). EPEC EspF_O127 has also been shown to inhibit EPEC translocation across antigen-transporting epithelial M cells in an in vitro model (53) based on a published coculture system (44).

In cattle, the terminal rectum is rich in lymphoid FAE containing M cells (49). Since this is the predominant site colonized by EHEC O157 in cattle (60), we hypothesized that EspF may have an important role in cattle colonization by inhibiting translocation of the organism by M cells and that the sequence differences of EspF in EHEC O157:H7 may reflect selection for its function in the bovine host. To test this, we compared the capacities of different espF alleles to inhibit phagocytosis and limit bacterial translocation through M cells derived in a coculture system to their capacities to inhibit bacterial uptake into and translocation through cells cultured from the bovine terminal rectum.

MATERIALS AND METHODS

Bacterial strains and media.

EPEC, EHEC, and other E. coli strains used in the present study are detailed in Table 1. Bacteria were cultured in Luria-Bertani (LB) broth, minimal essential medium (MEM)-HEPES, or Dulbecco modified Eagle medium (DMEM) with antibiotics included when required at the following concentrations: chloramphenicol, 20 μg/ml; kanamycin, 50 μg/ml; and ampicillin, 100 μg/ml.

Table 1.

Bacterial strains used in this study

Strain	Serotype	Source
E. coli ZAP1139	O26:H11	Cattle isolate, Stx negative; Chris Low, Scottish Agricultural College, Penicuik, Scotland
E. coli TUV93-0	O157:H7	Stx phage-negative derivative of the sequenced human strain EDL933; John Leong, Massachusetts Medical School, Boston, MA (10)
E. coli TUV93-0ΔespF	O157:H7	This study
S. enterica serovar (SL1344) Typhimurium		Mark Jepson (38)
E. coli EPEC E2348/69 (ΔespF)	O127:H6	Derivative of human isolate supplied by Brendan Kenny, University of Newcastle, Newcastle, United Kingdom (74)
E. coli AAEC 185	O148	Laboratory stock
E. coli DH5α	Rough	Laboratory stock

Cloning of espF alleles.

espF alleles were amplified from the different strains using the primers defined in Table 2 and the products cloned into pTS1 (Table 3). BamHI and HindIII sites were used to clone the alleles from EHEC O157:H7 and O26:H11 strains. For EPEC O127 espF, since this contains a natural HindIII site, an alternative cloning strategy was used: pMB102 (Table 3) was digested with HindIII, followed by Klenow polymerase (Roche) treatment to create a blunt end and then restriction with BamHI. The EPEC O127:H6 espF PCR product was digested with BamHI and EcoRV and then cloned into the restricted pTS1. Restriction digests were set up according to the manufacturer's instructions (New England Biolabs) with the recommended enzyme buffers. All clones were checked by sequencing, and matched published database sequences for the alleles giving the amino acid sequences are shown in Fig. 2. Gateway cloning of the espF alleles from the different E. coli strains was also carried out for the LUMIER assays. espF alleles were amplified by using a proofreading Taq polymerase with the primers defined in Table 2 and cloned into the Gateway vector pDONR-207 (Invitrogen) to form entry clones. The obtained clones were checked for the correct insert by DNA plasmid extraction and restriction digestion using BanII. Once in the Gateway system, clones were easily transferred to the required destination vectors, including pTREX-DEST30 (protein A fusions) and pRenilla (luciferase fusions).

Table 2.

Primers used in the study

Primer	Sequence (5′–3′)a	Use
EspF1F	CAGGATCCTCACAATGCAACTTAATATCG	Cloning of espFO26 into pTS1
EspF1R	CTAAGCTTGATATAAAAAGGCATGAATTATGC
EspF2F	CAGGATCCACCACACAATTTGATATCGGT	Cloning of espFO157 into pTS1
EspF2R	CTAAGCTTGATATAAAGAGGCATAAATTATGC
EspF3F	CAGGATCCACCACGCAATTTGATATCGGT	Cloning of espFO127 into pTS1
EspF3R	CTGATATCGATATAAAGAGGCATAAATTATGC
EspF4F	ATGCTTAATGGAATTAGTAACGC	Gateway cloning of espFO157 and espFO127
EspF4R	CTACCCTTTCTTCGATTGCTCATAG
EspF5F	ATGCTTAATGGAATTAGTCAAGC	Gateway cloning of espFO26
EspF5R	CTACACAAACCGCATAG

^aRestriction sites are underlined.

Table 3.

Plasmids used in this study

Plasmid	Description	Source or reference
pMB102	Low-copy cloning vector with inducible lac promoter	Tammari Schneiders, University of Belfast (69)
pAT1	pMB102 containing espFO127	This study
pAT2	pMB102 containing espFO157	This study
pAT3	pMB102 containing espFO26	This study
pUC-gfp	pUC18 derivative containing gfp	Laboratory stock
pAJR145	rpsM-GFP+ transcriptional fusion in pACYC184.	Laboratory stock
pAJR146	rpsM-RFP transcriptional fusion in pACYC184 (pDW16) based on the DsRed T3_S4T gene	71, 76
pDONR-207	Gateway entry cloning vector	Invitrogen
pTREX-DEST30-prA	Vector to generate amino-terminal protein A fusions for LUMIER binding assays	4
pRenilla	Vector to generate amino-terminal luciferase fusions for Lumier binding assays	4
pAT4-6	espFO157, espFO26, and espFO127 cloned into pTREX-DEST30-prA, respectively	This study
pAT7-9	espFO157, espFO26, and espFO127 cloned into pRenilla, respectively	This study
pTREX-DEST30-SNX9 pRenilla-SNX9	Vectors expressing SNX9-protein A or -luciferase fusion proteins	This study
pTREX-DEST30-N-WASP pRenilla-N-WASP	Vectors expressing N-WASP-protein A or-luciferase fusion proteins	This study

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).

To analyze type III secretion profiles and EspF secretion from the EPEC O127:H6 ΔespF and complemented strains, previous procedures were used (50, 61, 71). EspD was detected by using a monoclonal antibody kindly supplied by T. Chakraborty (Giessen, Germany). EspF was detected by using polyclonal antibodies supplied from C. Sasakawa (Tokyo, Japan) (58), and G. Hecht (University of Illinois at Chicago) (80).

Phagocytosis assays.

The murine macrophage cell line (RAW 264.7) was cultured in Dulbecco modified Eagle medium (DMEM; Sigma) supplemented with 10% heat-inactivated fetal bovine serum (FBS; Sigma), 1 U of penicillin, 1 μg of streptomycin/ml, and 2 mM l-glutamine (final concentrations). Cells were grown at 37°C in 5% CO₂ and moisture. The bacteria, transformed with pUC-gfp (Table 3), were inoculated from LB broth overnight cultures into DMEM to an optical density at 600 nm (OD₆₀₀) of 0.3. When analyzed, the espF alleles were induced with 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside). At an OD₆₀₀ of 0.7, the bacterial cultures were diluted 1:5 into prewarmed DMEM, and 300 μl was added to the macrophages (initially seeded at 10⁵ cells/well the day before). The cells were then incubated at 37°C in 5% CO₂ in a moist box for the desired incubation time. The cells were washed three times with sterile phosphate-buffered saline (PBS) and fixed using 4% paraformaldehyde (PFA). After three washes with PBS, the samples were incubated for 90 min with the relevant anti-lipopolysaccharide (LPS) antibody (MAST Group, Ltd.) at 1:4,000 and, after several washes, were incubated with Alexa Fluor 594-conjugated goat anti-rabbit immunoglobulins antibody (1/1,000; Molecular Probes) for an hour. The slides were then washed three times with PBS and mounted with Hydromount (National Diagnostic), and coverslips were applied. All bacteria, internal and external, express green fluorescent protein (GFP), whereas only the bacteria external to the macrophages were stained with the anti-O157 LPS antibody. This differential staining allows the proportion of internalized bacteria to be quantified.

Microscopy.

Confocal data were acquired using a 1,024 by 1,024 pixel image size, a Zeiss Plan Apochromat 1.4 NA ×63 oil immersion lens, and a multitrack (sequential scan) experimental setup on a Zeiss LSM510. Image data, acquired at Nyquist sampling rates, were deconvolved using Huygens software (Scientific Volume Imaging, Netherlands), the resulting three-dimensional models were analyzed, and orthogonal views were created using NIH ImageJ software; the final figures were assembled in Adobe Photoshop.

LUMIER assays.

Proteins were transiently expressed in HEK293 cells as hybrid proteins with the Staphylococcus aureus protein A tag or Renilla reniformis luciferase fused to their amino termini. Portions (20 ng) of each expression construct were transfected into HEK293 cells using 0.05 μl of Lipofectamine 2000 (Invitrogen) in 96-well plates. After 40 h, the medium was removed, and the cells were lysed on ice in 10 μl of ice-cold lysis buffer (20 mM Tris [pH 7.5], 250 mM NaCl, 1% Triton X-100, 10 mM EDTA, 10 mM dithiothreitol [DTT], protease inhibitor cocktail and phosphatase inhibitor cocktail [both from Roche], and Benzonase [Novagen]). Sheep anti-rabbit IgG-coated magnetic beads were also added (Dynabeads M280 [Invitrogen]; 2 mg/ml, final concentration), followed by incubation on ice for 15 min. Then, 100 μl of washing buffer (PBS, 1 mM DTT) was added per well, and 10% of the diluted lysate was removed to determine the luciferase activity present in each sample before washing. The rest of the sample was washed six times in washing buffer. Luciferase activity was measured in the lysate as well as in washed beads. Negative controls (nc) were wells transfected with the plasmid expressing the luciferase fusion protein and a vector expressing a dimer of protein A. For each sample, four values were measured: the luciferase present in 10% of the sample before washing (“input”), the luciferase activity present on the beads after washing (“bound”), and the same two values for the negative controls (i.e., “input nc” and “bound nc”). Normalized signal-to-noise ratios were calculated as published previously (4, 9): (bound/input)/(bound nc/input nc). Z-scores were then calculated using these normalized values by subtracting the population mean from an individual raw score and then dividing the difference by the population standard deviation. In normal distributions, results with a z-score of >1.96 lie within the 95% prediction interval, and z-scores of >2.58 lie within the 99% prediction interval.

Bovine primary rectal epithelial cell cultures.

Bovine primary epithelial cells were derived from lymphoid dense terminal rectum 0 to 5 cm proximal to rectal-anal junction, as described previously (7, 30). Briefly, the mucosal scrapings from terminal rectum of adult cattle from a local abattoir were digested in DMEM (1% [vol/vol] fetal calf serum [FCS], 100 U of penicillin ml⁻¹, 30 μg of streptomycin ml⁻¹, and 25 μg of gentamicin ml⁻¹) containing 20 μg of dispase and 75 U of collagenase (Roche) ml⁻¹, with gentle shaking at 37°C for 90 min to release intact crypts. Crypts were enriched from contaminating gut microflora and single cells, including fibroblasts using a series of differential centrifugation steps with DMEM containing 2% sorbitol (30). Approximately 500 crypts were seeded onto collagen (Vitrogen collagen; Nutacon, Netherlands)-coated four-well chamber slides (Costar; Corning). The cells were grown to a stage of confluence (approximately 3 × 10⁵ cells per well, typically at 5 to 7 days after initial primary epithelial cell culture). To inhibit the growth of fibroblasts, the cells were cultured in MEM-d-valine (Servichem GmbH) containing 2.5% (vol/vol) FCS, 0.25 U of insulin, 10 ng of epidermal growth factor ml⁻¹, and 25 mg of gentamicin ml⁻¹ (23, 30, 46).

For the transcytosis studies, primary bovine terminal rectal epithelial cells were allowed to establish monolayers on 3-μm-pore-size polycarbonate filters of 35-mm transwell chambers (Corning Inc.) by growing them for 2 to 3 weeks until they reached confluence. Before seeding primary bovine terminal rectal epithelial cells, the transwell plates were coated with collagen (Nutagen) since it is important to establish primary cell polarization. The collagen did not block the pore, and this was checked in comparison to non-collagen-coated plates. The integrity of the cell monolayer, polarization, and formation of tight junctions was tested by measuring the TER using an epithelial voltmeter (WPI). Only filters of cell monolayers that displayed the required TER (100 Ω/cm²) were used for the bacterial transcytosis assay. Bacteria were prepared as described for the phagocytosis assays, and 100 μl of bacterial suspension was added to each upper chamber. Bacterial counts from the lower chamber were determined at 60 min relative to the levels inoculated.

Characterization of bovine primary rectal cultured cells.

The epithelial origin of the cells was confirmed by immunostaining for cytokeratin intermediate filaments. The cells at 5 days of culture were fixed with 2% PFA, permeabilized with cold acetone for 5 min, washed with PBS, and stained with a pan-cytokeratin and or anti-vimentin antibodies (Sigma, 1:300) for 3 h at 25°C. The cells were washed with PBS, and the specific monoclonal antibodies were detected using fluorescein isothiocyanate (FITC)- or TRITC (tetramethyl rhodamine isothiocyanate)-labeled secondary anti-rabbit or anti-mouse monoclonal antibody (Sigma) at 1:80. The cell nuclei were stained with TO-PRO iodide (Molecular Probes). The stained cells were mounted in fluorescence mounting medium Fluoromount (Dako) and examined by using a Leica DMLB epifluorescence microscope. To ascertain whether lymphoid cells were present in primary cell cultures from the crypts isolated from lymphoid rich mucosal tissue, immunostaining was carried out with a panel of seven monoclonal antibodies (Table 4) specific for different immune cell types of cattle.

Table 4.

Monoclonal antibodies used in immunocytochemical screening of bovine rectal primary epithelial cell cultures

Monoclonal antibody	Specificity	Cellular expression	Sourcea	Reference
CC21	CD21	Follicular dendritic cells, mature B cells	IAH	57
CC20	Bovine CD1b	Dendritic cells	IAH	31
CC15	Bovine WC1	γd T cell	IAH	32
ILA-12	CD4	T helper cells	ILRAD	78
ILA-51	CD8	Cytotoxic T cells	ILRAD	78
ILA-156	CD40	B cells, antigen-presenting cells	ILRAD	63
ILA-111	CD25	Activated T and B cells, macrophages	ILRAD	12
ILA-43	CD2	αβ T cells, natural killer cells	ILRAD	78
Anti-hPH	β-Subunit of hPH	Fibroblasts	Acris GmbH	3

^aILRAD, International Laboratory for Research on Animal Diseases, Nairobi, Kenya; IAH, Institute for Animal Health, Compton, United Kingdom.

The percentages of cells expressing vimentin in the bovine terminal rectal primary epithelial cell cultures were determined using flow cytometry. Briefly, the trypsinized cells were fixed and permeabilized in 500 μl of cell permeabilization solution (BD Pharmingen) for 10 min at room temperature. The cells were washed in PBS with 0.5% bovine serum albumin and 0.1% sodium azide. The cells were incubated with primary anti-vimentin and the respective species-specific isotype control antibodies, diluted in flow cytometry (FC) medium (DMEM, 1% FBS, 0.1% sodium azide) at 4°C for 30 min, centrifuged (2,500 rpm for 3 min), and washed twice in FC medium. The cells were then labeled with FITC-labeled secondary antibodies at 4°C for 30 min, centrifuged again, and washed twice in FC medium. The cell pellets were suspended in 200 μl of FC medium, and vimentin-expressing cells were quantified by using Calibur flow cytometer (Becton Dickinson).

Uptake of inert microparticles and Salmonella enterica serovar Typhimurium has been used as in vitro functional assay to ascertain M cells in cultures (14, 44). Similar technique was adapted to identify M-cell subsets in bovine rectal epithelial cell culture. Briefly, FITC-conjugated latex beads with a 0.2-μm pore size (Polysciences, Inc., Germany) were diluted (1:1,000) in DMEM containing 2% FBS. Aliquots (100 μl) of diluted beads were pipetted evenly onto 6-day-old cultures, followed by incubation at 37°C for 45 min; the cells were then washed three times in PBS and further incubated with Salmonella Typhimurium (pUC18GFP-labeled SL1344 strain) (multiplicity of infection [MOI] of 1:100) for 10 min. The cells were washed three times in PBS, fixed, and permeabilized with 2% (wt/vol) PFA-0.25% (vol/vol) Triton X-100 at room temperature for 20 min. Staining of F-actin was carried out with Phalloidin-647 (diluted 1:40 in PBS; Molecular Probes) for 45 min at room temperature. For colocalization studies with vimentin, the cells were incubated overnight at 4°C with mouse anti-vimentin monoclonal antibody (1:100; Sigma). This primary antibody was detected with Alexa Fluor 594-tagged rabbit anti-mouse or goat anti-mouse polyclonal antibodies according to the manufacturer's instructions (Invitrogen). The cell nuclei were stained with either TO-PRO iodide or DAPI (4′,6′-diamidino-2-phenylindole; Merck) nuclear stains. The mounted slides were examined by confocal microscopy. To examine the position of the beads, 0.4-μm optical sections were acquired and processed using the Imaris Surpuss module (Bitplane) computer software program.

Coculture of Caco-2 cells with Raji-B cells and measurement of bacterial translocation.

Caco-2 cells were grown for 14 days in DMEM supplemented with 10% FBS, 15 mM l-glutamine, and 1% penicillin-streptomycin (Sigma) on transwell polycarbonate inserts (3-μm pore size). This time allows the differentiation and development of microvilli. These cells were then cocultured with 0.5 × 10⁶ Raji-B cells in the basal compartment for 6 days (44). The TER was measured every day before adding the Raji-B cells until it become consistent at 200 to 300 Ω/cm²; however, it was not changed by coculture with Raji-B cells. Bacteria were prepared as described for the phagocytosis assays, and 100 μl of bacterial suspension was added to each upper chamber. Bacterial counts from the lower chamber were determined at 60 min relative to the levels inoculated.

Kanamycin protection assay to detect intracellular E. coli.

Bacteria were prepared as described for the phagocytosis assays. The bovine primary epithelial cells were washed twice with MEM-HEPES without antibiotic for 1 h before the bacterial culture reached an OD₆₀₀ of 0.4, and the wells of confluent epithelial monolayer were infected by using 100 μl of bacterial suspension for each well. The infected cells were incubated at 37°C with 5% CO₂ and moisture for 1.5 h. The bacterial suspension was removed, and the wells were washed four times with PBS to remove the nonattached bacteria. A total of 500 μl of MEM-HEPES containing 750 μg of kanamycin/ml was added to each well of the infected cells. The infected cells were incubated for 3 h and then washed three times with PBS. Then, 300 μl of 0.1% Triton X-100 (in PBS) was added to each well to lyse the cells. The bacteria were collected after scraping of the cells, serially diluted in PBS, and triplicate plated onto LB plates with appropriate antibiotics. The plates were incubated at 37°C for overnight, and the colonies were counted the next day.

Statistical design and analyses.

All statistical analyses were carried out in R (v2.10.1; R Foundation for Statistical Computing). All binding, uptake, and translocation assays were repeated on at least separate occasions. For all analyses, a similar methodology was adopted. Overall differences between strains were first assessed, and if there were statistically significant differences, then post hoc Tukey pairwise comparisons were carried out.

Three types of statistical models were used. Differences between strains in the percentages of intracellular bacteria at either set time intervals, as determined by fluorescence microscopy or as a measure of phagocytosis, were examined by general linear models with binomial errors to account for the percentage nature of the data. To ensure that any differences between experiments were accounted for in the time interval analysis, which of the three experiments the data came from was also entered as a covariate. When we considered differences in the number of bacteria bound per microscopic slides between the three strains, general linear models with Poisson errors were used to account for the integer nature of the data.

For all other statistical analyses, analysis of covariance of the log₁₀-transformed number of bacteria posttranslocation was carried out to assess the differences between strains. To adjust for differences in pretranslocation levels and experiments, the number of bacteria pretranslocation and which of the three experiments the translocation carried out were entered as covariates. Log transformation was undertaken to normalize the residuals. Statistical significance was taken when P was <0.05.

RESULTS

Strain-specific susceptibility to phagocytosis by cultured macrophages.

Previous research has demonstrated that EspF is able to inhibit bacterial uptake into macrophages, although the majority of the previous research examining this inhibition has focused on the EspF_O127 variant from EPEC O127 E2348/69. Initially, EHEC O157, EHEC O26, and EPEC O127 were compared for their capacity to inhibit their nonopsonized phagocytosis into cultured macrophages. EPEC O127 bound in significantly higher numbers to the macrophages (Fig. 1A, P < 0.001), and the majority remained external (Fig. 1B and C). EHEC O26 exhibited an adherence level similar to that of EHEC O157 but significantly less than that of EPEC O127 (Fig. 1A, P < 0.001); however, it resisted uptake by macrophages in a way comparable to that of the EPEC strain (Fig. 1B and C). In contrast, EHEC O157 adhered at lower levels, and a significantly higher proportion was phagocytosed (Fig. 1B and C). For example, at 90 min after addition 63% ± 3.3% of EHEC O157 bacteria were internalized compared to only 33% ± 3.5% of EPEC O127 or 31% ± 6.3% of EHEC O26 (Fig. 1B). By 6 h after addition to the cultures, the macrophages lost their structural integrity in the presence of EPEC O127. In contrast, macrophages exposed to EHEC were still healthy at this time point, with relatively very few bacteria evident (data not shown).

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.

Strain-specific susceptibility to phagocytosis is associated with espF allele expression.

EHEC and EPEC serotypes express different variants of EspF (Fig. 2). EHEC O157 contains four polyproline repeat regions (PRRs) compared to three in EPEC O127 and EHEC O26. In addition, the three sequences differ within the polyproline repeat regions and in the amino terminus of the protein shown to be important for organelle targeting (1, 17). To determine the relative contributions of the different espF alleles in inhibiting phagocytosis, these were amplified from E. coli O157, O127, and O26 and cloned into pMB102 (Table 2) under the control of a pTAC promoter that is inducible with IPTG. The sequences were confirmed to be identical to those published in the National Center for Biotechnology Information (NCBI) database for these serotypes, resulting in the predicted amino acid sequences shown in Fig. 2. The clones were transformed into EPEC E2349/69ΔespF (Table 1), and the secretion profiles were examined. All strains secreted comparable levels of the translocon protein EspD; EspF secretion was detected from all three complemented strains by Western blotting (Fig. 3B) but with potentially lower levels detectable for the O26 variant. The complemented strains were then compared in a macrophage phagocytosis assay at 90 min postinfection (Fig. 3A). All three alleles were able to complement the espF knockout (Fig. 3, P < 0.001), although the espF gene from the E. coli serotype O157 (espF_O157) allele was significantly less effective at inhibiting uptake compared to the espF_O127 and espF_O26 alleles (Fig. 3A, P < 0.001). There were no significant differences in the adherence levels for the complemented strains (data not shown). These results indicate that the capacity of the different strains to inhibit phagocytosis correlates with the activity of the respective espF allele.

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.

Strain-specific variation in M-cell translocation is associated with espF variation.

Coculturing of human intestinal epithelial Caco-2 cells and Raji-B cells in vitro leads to the differentiation of a subset of epithelial cells into antigen-sampling cells (M cells) (44, 47). This in vitro “M-cell” culture system can then be used to analyze bacterial translocation and the capacity of bacteria to inhibit this trafficking (53). This trafficking is presumed to occur by transcytosis, but this and other research does not rule out the possibility of paracellular transport. Evidence for M-cell translocation in the assay is provided by comparison of translocation through a standard epithelial Caco-2 monolayer which occurs at significantly lower levels (Fig. 4A to C). To determine the contribution of espF to the translocation of EHEC O157 through this M-cell culture system, a defined deletion of espF was constructed in E. coli O157:H7 TUV93-0 (Table 1). Translocation of the EHEC O157 ΔespF strain was significantly higher than that of the wild-type strain across both the Caco-2/rajiB coculture and the Caco-2 monolayer. The levels of translocation were restored to those of the wild type by complementation with espF_O157 in trans (Fig. 4A).

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.

The EHEC O157, EHEC O26, and EPEC O127 strains were then compared in the same coculture translocation assay. EHEC O157 demonstrated significantly higher levels of translocation compared to EPEC O127 and EHEC O26 (Fig. 4B, P < 0.001). To examine whether the espF alleles have an effect on the level of translocation inhibition, EPEC E2349/69ΔespF transformed with the three amplified and cloned espF alleles was analyzed in the same assay. All three complemented the espF mutation in the EPECΔespF strain to a significant level (Fig. 4C, P < 0.001). As with the phagocytosis assay, the espF_O157 allele was the least effective of the three, being significantly less effective than the espF_O127 allele at inhibiting translocation. This was the case for translocation across both the Caco-2/Raji-B coculture and the Caco-2 monolayers, although translocation through the coculture system always occurred at significantly higher levels (P < 0.001). Given the correlation with strain origin, these differences again indicate variation in EspF activity between the different espF alleles.

The significance of espF alleles on the interaction of E. coli with primary cultures of bovine rectal epithelial cells.

Previous work has established that EHEC O157:H7 colonizes the terminal rectum of cattle, the main reservoir host (60). The tissue at this site contains a high number of lymphoid follicles and M-like cells that are present in the FAE (49). Primary cells were cultured from a mixed population of crypts isolated from this bovine terminal rectum (see Materials and Methods). The cells in the monolayers expressed cytokeratins (Fig. 5A) indicative of epithelial cells. Screening with a panel of antibodies (Table 4) provided no indication of contaminating cells, such as fibroblasts (data not shown). It was apparent that a small proportion of cells expressed vimentin, an intermediate filament protein indicative of M cells (Fig. 5A and B). A subset of vimentin-expressing cells endocytosed latex fluorescent microparticles (Fig. 5C and C1). S. enterica serovar Typhimurium preferentially targets M cells in the gut (14) and therefore was used to test whether the bacteria and the latex particles were internalized by vimentin-expressing cells in culture. Indeed, during early stages of infection, Salmonella Typhimurium interacted primarily with vimentin-expressing cells in culture (Fig. 5D) that also had internalized the latex beads (Fig. 5E), a further indication that these cells can be considered M cells based on previous M-cell characterization studies (44). Taken together, these data indicate that primary cells cultured from crypts isolated from bovine rectal FAE do contain a subset of cells with characteristics of M cells and are capable of taking up particles, including bacteria.

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.

To determine the significance espF has on the interaction of E. coli O157 with these primary cells, translocation across these primary cell cultures was then assessed for the wild-type strains, the EHEC O157 ΔespF deletion, and EPEC ΔespF strain complemented with the three different alleles. In agreement with the coculture translocation assay, espF significantly contributed to inhibition of translocation through these cultured cells (Fig. 6A); however, in clear contrast to the “human” coculture system, EHEC O157 was significantly better in the bovine assay at inhibiting translocation compared to EPEC O127 and EHEC O26 (P < 0.001) (Fig. 6B). In agreement with this, the espF_O157 allele showed the highest activity in restricting transcytosis in the EPEC ΔespF background, although this was not significant (Fig. 6C) compared to the two other alleles. As an alternative assessment of the function of these alleles on the bovine primary cells, an intracellular kanamycin protection assay was developed that determined the intracellular bacterial numbers at 90 min postinfection. The relative percentage of bacteria taken up into cells is low but increased significantly upon deletion of espF (Fig. 7A). EHEC O157 was taken up into cells at significantly lower levels compared to EPEC O127 and EHEC O26 (Fig. 7B, P < 0.001) in agreement with the transcytosis results. This assay was then carried out with the three espF alleles in the EPEC E2348/69ΔespF strain. All three espF alleles reduced uptake significantly (Fig. 7C, P < 0.001) with the espF_O157 variant showing significantly higher activity than the espF_O127 (P < 0.05) and espF_O26 (P < 0.001) alleles (Fig. 6E).

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.

The fact that the relative activities of the espF clones are reversed in the bovine assays indicates that the differences measured between the variants are not due to expression or secretion levels. This is supported by the correlations with the relative activity levels of the parental strains in the different assays. Taken together, it is likely that the relative activities are due to different functional capacities of the EspF variants and, in turn, these are dependent on the host cell type in which the variants are acting.

Molecular basis to the comparative activity of EHEC and EPEC espF alleles.

Previous work has established that EspF has a number of interacting partners in eukaryotic cells including interactions with both sorting nexin 9 (SNX9) and N-WASP. Since SNX9 is important for endocytosis dynamics and N-WASP is central to actin polymerization, we investigated by LUMIER binding assay whether differences in the interactions of the EspF variants with these proteins could account for the differences in functional levels measured in the present study. To quantify the binding activity, the three EspF variants, as well as their binding partners SNX9 (human) and N-WASP (human), were cloned into plasmids that express either protein A-tagged proteins or Renilla luciferase fusion proteins as described in Materials and Methods. Cell lysates containing both sets of tagged proteins were generated, and then the protein A-tagged complexes were removed by using antibody-coated beads. Total fluorescence and captured fluorescence were measured and normalized, and z-scores for each interaction were determined as outlined in Materials and Methods and as previously published (4). The results confirmed the interaction of all of the EspF variants with both SNX9 and N-WASP (Fig. 8). The EspF_O157 had the lowest interaction score of the three variants with luciferase-linked SNX9 and also had the lowest interaction score of the three with respect to binding to N-WASP.

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.

DISCUSSION

EPEC is known to inhibit phagocytosis via inhibition of PI3K activity, and this has been shown to be due to the activity of the type III-secreted effector EspF (11, 70). Although more recent research has established that a number of secreted proteins can also function to inhibit uptake into cells, including EspB (34), EspJ (52), and EspH (18). The inhibition of EPEC translocation through M cells has been shown to be dependent on espF, presumably in a manner analogous to its activity on macrophages. Comparison of EspF protein sequences from EPEC O127, EHEC O26, and EHEC O157 shows a number of differences in both the localization domain and in the number and sequence of proline-rich repeats (PRRs). The aim of the present study was to investigate whether these differences affected the capacity of the EspF variants to (i) inhibit bacterial phagocytosis into macrophages, (ii) inhibit M-cell translocation in a human-derived Caco-2/Raji-B coculture system, and (iii) inhibit uptake into bovine primary epithelium, containing M-like cells, cultured from the terminal rectum of cattle, the predominant colonization site of EHEC O157:H7 in cattle.

Strain comparisons demonstrated that both EPEC O127 E2348/69 and E. coli O26:H11 had a much greater capacity to block nonopsonized phagocytosis into cultured murine macrophages compared to EHEC O157:H7. Both the EHEC O157 and the EHEC O26 strains adhered to the macrophages at lower levels than the EPEC O127 strain (Fig. 1). To determine whether this strain difference could be accounted for by variation in the EspF effector protein, the genes from the three serogroups, EPEC O127, EHEC O157, and EHEC O26 were cloned and expressed in EPEC O127 from which espF had been deleted. Although the espF alleles from both O127 and O26 completely complemented the mutant, the espF_O157 allele showed a significantly reduced ability to block phagocytosis. This indicates that the initial strain observations may be accounted for by differences in the espF alleles. This result does not rule out important antiphagocytic functions for the other effector proteins but does indicate synergistic roles with EspF and a significant dependence on EspF. The different alleles did not alter the level of attachment of the complemented EPEC strain to Caco-2 cells, a finding in line with previous research indicating that EspF has no direct role in attachment and attaching/effacing lesion formation (55, 61, 74).

It was then investigated whether the different espF alleles varied in their capacity to inhibit E. coli translocation through M cells. The rationale for investigating this function is the finding that EHEC O157:H7 predominantly colonizes the terminal rectum of its main reservoir host, cattle, and that this site is rich in lymphoid follicles (60). Epithelium associated with lymphoid follicles is subject to different signals from the high abundance of B and T cells in these follicles leading to production of M cells (44) that sample lumen particles from the gut, delivering them to antigen-presenting cells that initiate appropriate adaptive immune responses at these sites. M cells have been postulated to be an important cell for the initial uptake or colonization by different enteric bacterial pathogens such as Salmonella, Shigella, and E. coli (37, 45). For EHEC O157:H7, it is a cell type that the bacteria will encounter when colonizing the terminal rectum (49, 60) and, consequently, inhibiting translocation may promote colonization, for example, by providing an initial attachment site and/or limiting bacterial presentation to the host's immune system. The origin of M cells is unclear since studies either support their derivation from enterocytes via extrinsic stimuli (luminal antigens and/or lymphoid-follicle derived signals) or conclude that they originate from lineage-specific precursor cells (25, 72, 73, 75). An M-cell coculture system was first defined in 1997 (44) and makes use of the capacity of Raji-B cells to signal differentiation of human colon-derived Caco-2 cells. Initial experiments demonstrated that espF from EHEC O157 is required to limit EHEC O157 translocation through M cells but that the EHEC O157 espF allele was again reduced in its capacity to limit translocation of E. coli compared to the O127 and O26 alleles. This finding mirrored the relative capacity of the specific strains to inhibit their translocation, again demonstrating the significance of the espF allele for the strain phenotype.

The strains and espF alleles were then compared on primary cells cultured from crypts isolated from the bovine terminal rectum. These cultures were further characterized in the present study. A subset of the primary cells expressed the intermediate filament protein, vimentin, that is found in rabbit intestinal M cells (13). M cells in the terminal rectum of cattle have been demonstrated to express vimentin as the predominant intermediate filament protein (49), and the same staining was apparent in the cultured primary cells. A subset of vimentin-expressing cells was also able to take up S. Typhimurium (Fig. 5D) alone and S. Typhimurium and beads together (Fig. 5E). Therefore, the bovine rectal primary cultures contain a subset of cells that express vimentin and have characteristics of M cells. There was a significant role for espF in inhibiting bacterial uptake into cells, and translocation through these cells and the activity of the strains correlated with that of the espF alleles. An important result was that, in contrast to the previous experiments on mouse macrophages and the human-derived M-cell coculture system, EHEC O157 was the most effective strain at inhibiting uptake and translocation and this correlated with the relative activity of the espF alleles (Fig. 6). Taken together, these results indicate that EHEC O157 and the espF_O157 allele are more effective at inhibiting uptake into bovine M cells than the O26 and O127 strains and alleles, a reversal of the situation observed for the interactions with murine macrophages and the human coculture system.

There are multiple phenotypes associated with EspF in addition to inhibition of phagocytosis, including inhibition of water transport and disruption of tight junctions mitochondria and the nucleolus (17, 29, 58, 80). Uptake functions are likely to be related to EspF interactions with N-WASP and SNX9 (1). N-WASP stimulates actin filament assembly by direct activation of the Arp2/3 complex, and SNX9 is essential for clathrin-coated pits (CCPs) at the late stages of vesicle formation. SNX9 binds to β2 appendages of adaptor protein complex 2 (AP-2) and interacts with clathrin and dynamin-2, two other important molecules in the endocytic process (48, 77). SNX9 has two lipid interaction domains; a phospholipid-binding region termed the phox (PX) domain, followed by a putative Bin/amphiphysin/Rvs (BAR) lipid-binding domain. The BAR domain is a banana-shaped helical dimer that senses membrane curvature and can reconfigure lipid vesicles or sheets into membrane tubules (1). SNX9 also possesses an N-terminal Src homology-3 (SH3) protein interaction region that was recently shown to bind WASP (2) and to functionally activate dynamin at CCPs. Therefore, SNX9 is an important factor in remodeling the membrane and cytoskeletal during endocytosis (1, 77). The SH3 domain of SNX9 binds to the PPRs of EspF (1, 17). Despite both direct and indirect (via SNX9) activation of N-WASP, EspF is not considered to have a role in A/E lesion formation (54, 55). To determine whether the different EspF variants have different affinities for SNX9 and N-WASP, the interactions were assayed inside transfected HEK293 cells using a LUMIER assay (4; described in Materials and Methods). This in-cell assay confirmed that all three EspF variants bound to both N-WASP and SNX9, although EspF_O157 showed a lower interaction score with both SNX9 and N-WASP compared to the other two variants. Whether this is due to a reduced affinity of EspF_O157 for these two proteins remains to be determined. A specific region in the PRRs of EspF (shaded gray in Fig. 2) interacts directly with SNX9 through its SH3 domain-binding motif. There are only minor differences in this region between the serotypes examined here, and we cannot determine whether these or their presentation, due to sequence changes in flanking regions, account for the differences observed.

The published Bos taurus sequence for SNX9 (NCBI) contains a number of changes over the human variant used in the LUMIER assays; however, these differences are mainly present in the amino terminus of the predicted bovine SNX9 and lie outside the SH3 domain. There are also very few sequence differences between the predicted Bos taurus N-WASP protein sequence (NCBI) and the human variant. Taken together, our results indicate that there are likely to be other protein interactions involving EspF that could also contribute to the host specificity demonstrated here. The greatest region of diversity between the variants lies in the motifs preceding the second and third PRRs (amino acids 110 to 118 and amino acids 157 to 165 in Fig. 2), with no established function for these regions. Future work will seek to identify other interacting partners for EspF and address how important M-cell interactions are for colonization of cattle at the terminal rectum or whether other factors may explain the tropism of the EHEC O157:H7 for this gastrointestinal site.