Alexa 488- and 546-conjugated phalloidin, were obtained from Invitrogen. Primary antibodies used were mouse anti-ezrin (Zymed # 357300), rabbit anti-O157 (Gibco), and goat anti-O157 (Fitzgerald Industries #70-XG13). Secondary antibodies are species-specific Alexa 350-, 488-, and 546-labeled anti-IgG (Invitrogen).

CaCo-2a enterocytes, stably transfected with the calpastatin high over-expression plasmid pRC/CMV-3ΔCSN (HOX) or the pRC/CMV empty vector (CON), have been previously described (Potter et al., 2003). Cells were grown in Dulbecco’s Modified Eagle’s Medium supplemented with 10% fetal bovine serum, 1% each penicillin–streptomycin–fungizone (PSF) and l-glutamine. Sub-confluent cultures were maintained in T175 flasks. For experiments, cultures were washed twice with HBSS, released with trypsin–EDTA, and seeded at confluence (approximately 1.5 × 10⁵ cells/cm²) in tissue cultures vessels precoated with collagen (BD #354236). CaCo-2a cells were then allowed to differentiate for 7–14 days, and fed every 2 or 3 days. For electron microscopy, cells were seeded onto standard plastic six-well, tissue culture plates. For immunofluorescence staining, cells were seeded into 24-well glass bottom tissue culture plates (Mattek #P24G-1.5-13F). For calpain activity, cells were seeded into plastic 96-well black walled tissue culture plates with optically clear bottoms (Potter et al., 2003).

Table 1 lists bacterial strains used in this study. The wild type, Stx-deficient EHEC strain of serotype O157:H7, strain TUV-93, a derivative of the prototypic E. coli O157:H7 strain EDL933, has been previously described. The lack of Shigatoxin production by TUV-93-0 permits the prolonged infection of a wide variety of cell lines without the induction of Stx-mediated cytotoxicity. TUV-93ΔescN was generated by lambda Red recombineering as previously described (Murphy and Campellone, 2003).

In preparation for infection, individual colonies from freshly streaked plates were grown in 1 mL of LB with appropriate antibiotics for up to 8 h. Ten microliter of this day culture was transferred to DMEM/100 mM HEPES/antibiotics and incubated overnight at 37°C with 5% CO₂ to induce T3SS expression. For infection, bacteria were resuspended in DMEM/2% FBS/20 mM HEPES/2 mM glutamine to the following MOI’s: EHEC at 500:1, EPEC at 100:1. Higher MOI’s were necessary for EHEC infections due to less efficient binding of EHEC to cultured cells; this is a commonly observed phenotype that was noted at least two decades ago (Cantey and Moseley, 1991). Strain TUV-93 also does not adhere to in vitro cultured cells as well as EHEC harvested from infected piglets (Brady et al., 2011). The latter observation suggests that the poor cell attachment by in vitro-grown EHEC, as were analyzed here (and in virtually all published cell binding studies), reflect the relative lack of a putative adhesin (or adhesins) that is expressed at higher levels when EHEC is growing in the mammalian host.

Calpain activity was assayed by either one of two fluorogenic calpain substrate kits. For assays using AnaSpec’s Calpain Activity Kit (#72150), CaCo-2a cell monolayers are seeded in 96-well TC treated plates with clear bottoms and black walls (Ibidi #89626) 10–14 days prior to use. Cells were infected as described above. Infected monolayers were washed twice with PBS. Fifty microliter of assay buffer was added to each well and incubated at room temperature (RT) for 5–10 min. Cells were scraped with a pipette tip and 50 mL of fluorogenic substrate in assay buffer was added to each well. Plates were spun at 1000 rpm on a tabletop centrifuge (700 × g) for 5 min to reduce bubbles and incubated at RT in the dark for up to 60 min. For assays using Calbiochem’s InnoZyme™ Calpain Activity Assay Kit, (#CBA054), CaCo-2a cell monolayers are seeded in six-well TC treated plates 10–14 days prior to use. Cells were infected as described above. Infected monolayers were washed twice with PBS, and lysed with CytoBuster^TM (EMD #71009) on ice for 5 min. Cell lysates were scraped, transferred into 1.5 mL tubes and spun at 12,000 g at 4°C for 15 min. The supernatants were collected and used immediately for measuring calpain activity per kit instructions or stored at −80°C. Fluorescence was measured using a Molecular Devices SpectraMax Gemini XS at 354 nm excitation and 442 nm emission for the Anaspec kit and 320 nm excitation and 480 nm emission for the Calbiochem kit, and corrected for background fluorescence (i.e., that of wells with substrate only).

Calpastat, a cell-penetrating calpastatin peptide (Croce et al., 1999) was synthesized at the Tufts Peptide Core Facility using solid phase Fmoc chemistry. Peptides were purified by high performance liquid chromatography; molecular mass and purity were confirmed by mass spectrometry. A 25 mM peptide stock was prepared in 100 mM HEPES, pH 7.4 and stored at −20°C. For treatments, cell monolayers were washed once with HBSS and cell culture media containing Calpastat at the specified concentrations was added. Cells were incubated at 37°C in 5% CO₂ for 1 h.

Cell monolayers were washed twice with HBSS, then fixed by immersion in 2.5% glutaraldehyde in 100 mM sodium phosphate buffer (pH 7.2) for a minimum of 2 h at RT. The fixed samples were then washed three times in the same buffer. Following the third wash, monolayers were dehydrated through a graded series of ethanol to 100% and then critical point dried in liquid CO₂. The bottoms of the dishes were cut off and using silver conductive paste, the plastic disks with the cells attached on the surface were affixed to aluminum scanning electron microscopy (SEM) stubs and sputter coated with Au/Pd (80/20). The specimens were then examined using an FEI Quanta 200 FEG MK II scanning electron microscope at 10 kV accelerating voltage. Each specimen was systematically observed at 1000×, 2500×, and 5000× at five separate areas spaced throughout the disk. Images were assessed by four individuals, one of whom was blinded and another of whom was not involved in the study. Assessments were unanimous concerning the presence or absence of effacement. Representative images at 10,000× were taken with disks tilted about 30°.

For indirect immunofluorescence imaging studies, CaCo-2a cells were seeded at confluence and allowed to differentiate in 24-well glass bottom culture plates (Mattek #P24G-1.5-13F). At the indicated time points following infection, cells were washed with warm DMEM, then fixed in 4% formaldehyde/DMEM for 5 min at RT. Cells were permeabilized with 0.1% Triton buffer for 90 s at RT, washed three times with PBS, and then incubated with the specified primary antibodies for 1 h at RT. Labeled secondary antibodies and/or Alexa 488- or 544-conjugated phalloidin (Invitrogen) were applied for 45 min at RT. Experiments were repeated three times. For each experiment, greater than 5 fields were examined per condition; roughly 20–50 cells per field were viewed/scored. Images were acquired on a Zeiss 200 M inverted microscope with a Hamamatsu cooled-CCD digital camera (ER) and MetaMorph 7.0 imaging software (Molecular Devices Corp, PA, USA).

Infected and uninfected cells were washed with warm TBS to remove media and unattached bacteria. Extraction buffer containing 40 mM HEPES pH 7.2, 50 mM PIPES, 75 mM NaCl, 1 mM MgCl₂, 0.5 mM EGTA, Protease Inhibitor Cocktail (1:1000, Sigma), 1 mM sodium orthovanadate, and 0.1% Triton detergent was applied to the cultures, 0.5 ml per well of a six-well plate, and plates were placed on an orbital shaker at 50 rpm for 10 min at RT. The extraction buffer was then removed to tubes held on ice. The remaining cellular residue was collected with 200 μL boiling hot Laemmli sample buffer (Sambrook and Russell, 2001). All extracts were dialyzed against 4 L distilled water at 4°C for 4 h with one water change at 2 h, using SnakeSkin dialysis tubing, 10 K MWCO (Pierce). Dialyzed samples were snap frozen in liquid nitrogen, lyophilized overnight, and brought up in 200 μL each Laemmli sample buffer. All samples were boiled for 3 min and loaded onto 10% acrylamide gels for SDS-PAGE, followed by transfer to nitrocellulose membranes. Membranes were stained for ponceau to confirm equal protein loading, then probed for ezrin using monoclonal mouse anti-ezrin (Zymed #357300).

To investigate whether calpain activity is affected by EHEC, we used two different commercially available kits that utilized calpain-specific fluorogenic substrates. Compared to uninfected cell monolayers, calpain activity increased after 3 h of infection by EHEC. As reported (Dean et al., 2010), EPEC also induced an increase in calpain activity, while the T3SS deficient EHEC mutant Δescn did not. Figure 1 shows the results of one representative experiment. Pretreatment of polarized cell monolayers with a cell-penetrating calpastatin inhibitor for 1 h prior to EHEC infection reduced EHEC-induced calpain activation to near uninfected levels (Data not shown).

To assess the role of calpain in EHEC-induced effacement, we infected polarized CaCo-2a CON cells with or without pretreatment with Calpastat, a cell-penetrating version of the endogenous calpain inhibitor, calpastatin (Croce et al., 1999; Carragher, 2006). SEM of uninfected control cells revealed distinct microvilli and cellular borders, although some variation in the density and arrangement of microvilli was observed (Figure 2A, and data not shown).

When CON cells were infected with EHEC and assessed by SEM at low (1000×) magnification (Figure 2) or TEM (data not shown), CON cells suffered severe “blebbing” and rounding, indicating cellular damage (Figure 2B). This apparent toxicity was not observed after infection with EPEC (Figure 2C), in which cell monolayers looked nearly identical to uninfected monolayers (Figure 2A). Pretreatment with 2.5 mM Calpastat for 1 h prior to EHEC infection diminished these effects (Figure 2D), suggesting that calpain activity is required for this cellular damage and raising the possibility that calpain may be critical for EHEC-induced cellular pathogenesis.

Higher magnification assessment of monolayers at 5000× to 10,000× revealed that a 2-h infection by EHEC resulted in the predicted microvillar effacement (Figure 3B) as compared to uninfected controls (Figure 3A). When CON cells were subjected to a 1-h pretreatment with 2.5 mM of Calpastat prior to EHEC infection and assessed by 10,000× magnification SEM, most EHEC-bound cells retained largely intact microvilli (Figure 3C), similar in quality to uninfected CaCo-2a CON monolayers (Figure 3A). As expected, the control infection by EPEC in the absence of Calpastat treatment also revealed effacement (Figure 3D).

We previously showed that the calpastatin-overexpressing CaCo-2a cell line, HOX, for which the pRC/CMV vector-transfected CaCo-2a CON line serves as an appropriate “wildtype” control, resists effacement by EPEC (Potter et al., 2003). Like the parent CaCo-2a and vector control CON line, these cells polarize and appear to differentiate, in that they develop trans-epithelial resistance and, by SEM, generate microvilli (albeit morphologically distinct from those on CON cells; Potter et al., 2003). Like CON cells, some cell-to-cell variability with respect to microvillar length and density (data not shown) were observed. HOX microvilli were, intact and in general, shorter than those observed in CON cells, (Figure 4A). As reported; these phenotypic differences compared to wild type CaCo-2 cells are likely due to the altered cytoskeletal architecture resulting from disrupted calpain-driven cytoskeletal remodeling (Potter et al., 2003).

When stained for F-actin using fluorescent phalloidin and imaged near the apical surface of the monolayers, the microvillar F-actin core bundles of polarized CON and HOX cells appeared as a punctate fluorescent pattern on the apical surface (Figures 4C,E) when viewed “end-on,” while the circumferential “belt” of F-actin encircling the apical cytoskeletal–membrane domain could also be readily distinguished. This apical F-actin staining pattern was consistent with the pattern of microvilli observed by SEM (Figures 3D and 4A). In addition, infection of polarized CON cells by EHEC induced the loss of this punctate fluorescent staining (Figure 4D), consistent with the loss of microvilli observed during EHEC-induced effacement as revealed by high magnification SEM (Figure 3A). (Note that the EHEC strain used in this study requires 4–6 h of infection to form actin pedestals, and were not detected in the timeframe of the current study.)

In stark contrast to the effacement of CON cells by EHEC, infected HOX cells retained their microvillar-associated punctate F-actin fluorescence staining (Figure 4F). The retention of microvillar integrity by HOX cells after 2 h (Figure 4F) or 3 h (data not shown) was confirmed by SEM (Figure 4B and data not shown). These results indicate that ectopic expression of calpastatin and the concomitant inhibition of calpain renders HOX cells resistant to effacement by EHEC infection.

We have previously shown that calpain-mediated cleavage of ezrin occurs during cell spreading and migration, a process that, like effacement, involves extensive actin remodeling (Shuster and Herman, 1995; Potter et al., 1998, 2003). Given that EHEC induces calpain activity upon infection of mammalian cells, we tested for ezrin cleavage upon infection by EHEC. Immunoblotting of subcellular fractions of EHEC-infected CaCo-2a CON cells revealed ezrin cleavage (Figure 5A). The full-length 80 kDa species was detected in both soluble and insoluble fractions. The cleaved 55 kDa ezrin fragment, which is not associated with the plasma membrane, was detected at very slight levels in the soluble fraction of uninfected CON cells. Endogenous calpain activity is present in all cells, supporting various functions in situ such as cellular differentiation, membrane polarization, or cell cycle progression. Therefore a small amount of cleaved ezrin should exist, even in uninfected cells. Infection with EHEC increased this pool of cleaved ezrin. A significant amount of full-length ezrin was still detected in EHEC-infected CON monolayers likely because, as can be seen in Figures 4D,F, EHEC does not infect cells uniformly, but rather heavily infects a fraction of cells. Thus, a mixture of infected and uninfected cells contributes to the lysate collected for western blotting. The 55 kDa ezrin cleavage product was not observed in uninfected or EHEC-infected HOX cells, indicating that calpain is required for its generation (Figure 5A).

Immunofluorescence staining of ezrin in polarized CaCo-2a CON cells revealed a punctate pattern at the apical surface almost identical to that of microvilli staining by phalloidin (Figure 5B). Images taken at the apical surface showed a staining pattern that is consistent with its role in linking the plasma membrane to core microvillar actin bundles (Fehon et al., 2010). To determine if EHEC-mediated cleavage of ezrin correlated with a change in the cellular distribution of ezrin, we stained CON cells for ezrin after EHEC infection. Interestingly, the apical punctate ezrin pattern disappeared throughout the entire infected cell (Figure 5C). Cleavage of ezrin would result from its release from the apical surface, thus appearing as a loss of staining in the focal plane imaged. This change in localization was dependent on calpain activity, because calpastatin-overexpressing HOX cells retained their punctate ezrin staining pattern during infection (Figure 5D).