Enteropathogenic Escherichia coli dynamically regulates EGFR signaling in intestinal epithelial cells

Abstract

The diarrheagenic pathogen enteropathogenic Escherichia coli (EPEC) dynamically modulates the survival of infected host intestinal epithelial cells. In the initial stages of infection, several prosurvival signaling events are activated in host cells. These include the phosphorylation of epidermal growth factor receptor (EGFR) and the consequent activation of the phosphatidylinositol-3 kinase/Akt pathway. While studying this pathway in infected epithelial cells, we observed EGFR depletion at later stages of infection, followed subsequently by a decrease in phospho-EGFR. EGFR loss was not dependent on receptor phosphorylation, or on canonical proteasome- and lysosome-dependent processes. Although a type III secretion mutant (ΔescN) stimulated EGFR phosphorylation, it failed to induce receptor degradation. To identify the specific EPEC effector molecule(s) that influenced EGFR stability, epithelial cells infected with isogenic mutant EPEC strains were examined. An EPEC ΔespF strain failed to induce EGFR degradation, whereas EPEC ΔespZ accentuated receptor loss in infected cells. Given the known and contrasting effects of EspF and EspZ on caspase activation, and the known role of proteases in cleaving EGFR, we explored the effect of caspase inhibitors on infection-dependent EGFR loss. The pan-caspase inhibitor Q-VD-OPh blocked EPEC-induced EGFR cleavage in a dose-dependent manner. Taken together, our data suggest that EPEC EspF stimulates caspase-dependent EGFR cleavage and loss, whereas EspZ inhibits this process. Whereas EGFR phosphorylation contributes to the survival of host cells early in infection, EspF-driven caspase activation and consequent EGFR loss likely induce a precipitous increase in host cell death later in the infectious process.

intestinal pathogens, particularly those that are predominantly extracellular, have to contend with niche-specific parameters, including the relatively steady turnover of intestinal epithelial cells. The diarrheagenic pathogen enteropathogenic Escherichia coli (EPEC) subverts intestinal epithelial function by injecting specific proteins into host cells via a type III secretion system (T3SS). It is now increasingly evident that the pathogen dynamically controls and alters host cell physiology via these “effector” proteins (41). Among other effects, EPEC virulence factors modulate the survival of intestinal epithelial cells. It has been appreciated for some time that intestinal epithelial cells infected by EPEC, and related pathogens, show minimal cell death early in infection (7, 8, 14, 29) and a dramatic decrease in viability during the later stages of the infectious process.

Crane et al. (7) were the first to describe death in EPEC-infected cells as having features of both apoptosis and necrosis and first to make the suggestion that the pathogen regulates host cell survival via multiple pathways. Thus the EPEC secreted molecule EspF localizes to the mitochondria, causes membrane depolarization, and induces caspase-dependent cell death pathways (8, 17, 22–24). Some EPEC strains also translocate the inhibitory calmodulin Cif, which induces host cell cycle arrest at both G1/S and G2/M and triggers delayed apoptosis (30, 31, 36). Prolonged infection of epithelial cells results in caspase-independent death; however, the responsible EPEC molecules and corresponding host cell pathways have not been well defined.

On the other hand, EPEC also elaborates several effector molecules that inhibit host cell death. For instance, EspZ-dependent prosurvival signaling in host cells contributes to direct, and indirect, inhibition of caspase activation (4, 29, 32, 33). In addition, the secreted effectors NleH and NleD also contribute to cytoprotection. NleH interacts with Bax inhibitor-1 to prevent elevated cytoplasmic Ca²⁺ levels, nuclear condensation, caspase-3 activation, and membrane blebbing (15). NleD has a zinc metalloprotease activity that cleaves JNK and p38, thereby blocking JNK-mediated proapoptotic signaling (2).

We previously demonstrated the activation of the EGFR/phosphatidylinositol-3 kinase/Akt pathway in EPEC-infected intestinal epithelial cells and showed that it promotes the survival of these host cells (28). Secreted factor/s in EPEC culture supernatant are sufficient to induce EGFR phosphorylation (28). However, we also observed the loss of total EGFR at later time points postinfection. In the present study, we define the mechanisms regulating EGFR stability in infected cells and demonstrate that the secreted effector proteins EspZ and EspF modulate EGFR levels. Functional implications of EGFR regulation on host cell survival and EPEC colonization include disease progression and bacterial dissemination, and these are also discussed.

METHODOLOGY

Cell lines.

The human intestinal epithelial cell lines Caco-2_BBE (C2_BBE), a brush border-expressing subclone of the Caco-2 cell line (25), and T84 were used in this study. C2_BBE cells were cultured in Dulbecco's modified Eagle's medium (DMEM) with 25 mM glucose, 10% fetal bovine serum, and 20 mM HEPES (Invitrogen, Grand Island, NY) at 37°C in the presence of 5% CO₂. Cells between passages 25 and 45 were used for experiments at days 7–10 postseeding. T84 cells were cultured in a 1:1 mixture of Ham's F12 medium and DMEM with 2.5 mM l-glutamine supplemented with 5% fetal bovine serum (Invitrogen) at 37°C in the presence of 5% CO₂.

Growth of bacteria and infection of cell lines.

Bacterial strains used in this study include the enteropathogenic Escherichia coli O127:H6 strain E2348/69 (EPEC) and its isogenic derivatives (Table 1). C2_BBE or T84 culture medium was changed to serum-free DMEM 14 h prior to bacterial infection. For infections, overnight bacterial cultures (in Luria-Bertani broth) were subcultured at 1:20 dilution in serum-free DMEM and grown to midlog growth phase (OD_{600 nm} = 0.4) at 37°C. Bacteria were added to the apical surface of C2_BBE or T84 cells corresponding to an initial multiplicity of infection (MOI) of 20 or 100 (C2_BBE), or 200 (T84; consistent with most previous studies from diverse laboratories, higher EPEC MOIs are required to observe comparable changes in this cell line) as required. Following incubation at 37°C in a 5% CO₂ water-jacketed incubator for 1 h, unattached bacteria were removed by washing with DMEM, fresh medium was added, and cell culture plates were incubated for additional time periods. Times indicated in the figures are from the point of initial addition of bacteria. For inhibitor studies, cells were pretreated with appropriate dilutions of the tyrosine kinase inhibitor tyrphostin AG1478 (EMD Millipore, Billerica, MA) or the pan-caspase inhibitor Q-VD-OPh (Sigma-Aldrich, St. Louis, MO), the proteasomal inhibitor MG-132 or the lysosomal inhibitor Bafilomycin A1 (Sigma-Aldrich), for 2 h prior to infection. Inhibitor levels were maintained throughout the course of subsequent infection.

Table 1.

List of strains

Strain Name Used in This Paper	Description/Relevant Genotype	Reference
EPEC	EPEC E2348/69 strain (serotype O127:H6); isolated from an infant during an outbreak in Taunton, United Kingdom in 1969; E2348/69 Nalr	37
ΔespZ	EPEC derivative MK41 with a nonpolar espZ deletion; E2348/69 Nalr Kmr::ΔespZ::aphA-3	16
ΔespF	EPEC derivative UMD874 with the nonpolar espF deletion; E2348/69 Nalr Kmr::ΔespF::aphA-3	16, 19
ΔescN	EPEC derivative CVD4524 with the nonpolar escN deletion; E2348/69 Nalr Kmr::ΔescN::aphA-3	6, 20
DH5α/vector	DH5α harboring the cosmid vector pCVD551	18
DH5α/pLEE	DH5α harboring pCVD462 (cloned LEE)	18

Immunoblot analysis.

Infected monolayers, and uninfected controls, were washed with sterile PBS, scraped, and total protein extracted in cell lysis buffer [20 mM Tris (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM β-glycerol phosphate, 1 mM Na₃VO₄, 1 μg/ml leupeptin, and 1 mM PMSF] (Sigma-Aldrich). Extracts were sonicated on ice three times to facilitate cell lysis. Protein concentrations were determined by use of the Pierce Coomassie (Bradford) Protein Assay Kit (Thermo Fisher Scientific, Rockford, IL). For immunoblot analyses, extracts (300 to 500 μg) were separated on 8% SDS-PAGE by use of the Protean II xi apparatus (Bio-Rad, Richmond, CA). Separated proteins were transferred to 0.2-μm nitrocellulose membranes (Transblot Cell Apparatus, Bio-Rad), blocked with 5% nonfat milk in Tris-buffered saline containing Tween 20 (TBST) and incubated with antibodies to pEGFR Tyr1068 (Epitomics, Burlingame, CA), EGFR (Santa Cruz Biotechnology, Dallas, TX), and actin (Sigma-Aldrich) at appropriate dilutions in blocking solution overnight at 4°C. Blots were incubated in horseradish peroxidase-conjugated goat anti-rabbit antibody (Sigma-Aldrich) for 1 h at room temperature. Membranes were washed three times for 5 min in TBST between each incubation step, and developed with SuperSignal West Femto Chemiluminescent Substrate (Thermo Fisher Scientific, Rockford, IL).

RESULTS

EPEC induces EGFR loss in infected intestinal epithelial cells.

In our previous study on EGFR phosphorylation in EPEC-infected intestinal epithelial cells (28), we observed the depletion of pEGFR at late time points postinfection. Similarly, in this work, and in EPEC-infected C2_BBE cells (MOI = 20), peak EGFR phosphorylation was evident at 4 h postinfection, with a subsequent reduction in pEGFR at 6 h postinfection (Fig. 1A). Additionally, there was a reduction in total EGFR levels at 6 h postinfection. EGFR depletion was EPEC dose dependent, with a more rapid decline in receptor levels in cells infected at a higher MOI (100; Fig. 1B).

Fig. 1.

Epidermal growth factor receptor (EGFR) degradation in infected cells is enteropathogenic Escherichia coli (EPEC) dose dependent. C2_BBE cells were treated with medium alone or infected with EPEC at multiplicity of infection (MOI) = 20 (A) or MOI = 100 (B). Total protein extracts were prepared, separated, and analyzed for pEGFR Tyr1068, EGFR, and actin abundance by Western hybridization. Blots shown are representative of 3 similar experiments. Densitometric analysis was performed to determine EGFR abundance relative to actin and normalized to levels in control uninfected cells. Significant changes were assessed by ANOVA with Bonferroni post hoc test. *P < 0.01 and **P < 0.01 for specific conditions compared with uninfected control. Error bars represent standard error.

EPEC-induced EGFR degradation is phosphorylation independent.

In the canonical pathway, downregulation of EGFR signaling involves internalization via clathrin-mediated endocytosis, ubiquitination of the phosphorylated receptor, and subsequent proteasomal degradation (1, 10, 21). Therefore, the EGFR kinase inhibitor tyrphostin AG1478 was used to assess the contribution of EGFR phosphorylation and activation to degradation in infected cells. Although AG1478 effectively abrogated EGFR phosphorylation in EPEC-infected cells, it did not significantly alter the kinetics of EGFR loss (Fig. 2A). Following activation and endocytosis, EGFR is typically degraded via lysosome- or proteasome-dependent processes (35). However, infection-induced EGFR loss was evident even in the presence of proteasomal (MG132) and lysosomal (bafilomycin A) inhibitors, respectively, suggesting alternate mechanisms for this process (Fig. 2B).

Fig. 2.

A: EPEC induces phosphorylation-independent degradation of EGFR. C2_BBE cells were pretreated with the tyrosine kinase inhibitor AG1478 (1 μM) or vehicle control (DMSO) for 2 h prior to infection with EPEC at MOI = 100. Inhibitor was maintained throughout the course of infection. Total protein extracts isolated at specified time points were immunoblotted for pEGFR Tyr1068, EGFR, and actin. Blots shown are representative of 3 similar experiments. Densitometric analysis was performed to determine EGFR abundance relative to actin and normalized to levels in control uninfected cells. Significant changes were assessed by ANOVA with Bonferroni post hoc test. *P < 0.01 for specific conditions compared with uninfected control. Error bars represent standard error. B: EPEC-induced EGFR degradation is not mediated by proteosomal and lysosomal processes. C2_BBE cells were pretreated with the proteosomal inhibitor MG132 (10 μM), lysosomal inhibitor bafilomycin A (100 nM; Baf A), or vehicle control (DMSO) for 2 h prior to infection with EPEC at MOI = 100. Inhibitor was maintained throughout the course of infection. Total protein extracts isolated at specified time points were immunoblotted for EGFR and actin. Blots shown are representative of 3 similar experiments. Densitometric analysis was performed to determine EGFR abundance relative to actin and normalized to levels in control uninfected cells. P values for specific conditions compared were assessed by ANOVA with Bonferroni post hoc test. EGFR loss at 4 h postinfection was not significantly affected by MG132 or bafilomycin A. Error bars represent standard error.

EGFR depletion is mediated by EPEC translocated effector(s).

To define the mechanism of EPEC-induced EGFR loss, C2_BBE cells were treated with wild-type (WT) EPEC, filtered EPEC culture supernatant, or a type III secretion-defective mutant (ΔescN). Whereas EGFR phosphorylation was evident with all three treatments (Fig. 3), receptor loss was observed only in WT EPEC-infected cells. This further supports the contention that receptor loss is likely not a consequence of phosphorylation-dependent events, and that it is dependent on one or more type III-secreted effector proteins.

Fig. 3.

A secreted EPEC component promotes EGFR phosphorylation, but not its degradation. C2_BBE cells were treated with medium alone or filter-sterilized EPEC supernatants (EPEC S/n), or infected with wild-type EPEC or an isogenic secretion-defective strain (ΔescN) at MOI = 20. Total protein extracts collected at specific time points were immunoblotted for pEGFR Tyr1068, EGFR, and actin. Blots shown are representative of 3 similar experiments; for each blot, the spacing after lane 7 denotes the cropping of irrelevant lanes. Densitometric analysis was performed to determine EGFR abundance relative to actin and normalized to levels in control uninfected cells. Significant changes were assessed by ANOVA with Bonferroni post hoc test. *P < 0.01 for specific conditions compared with uninfected control. Error bars represent standard error.

To determine whether the effector/s responsible for EGFR degradation are encoded within the 35-kb EPEC pathogenicity island (locus of enterocyte effacement, LEE), C2_BBE cells were infected with a nonpathogenic E. coli strain (DH5α) harboring plasmid-encoded LEE (DH5α/pLEE). This strain was previously demonstrated to support type III secretion-dependent translocation of LEE-encoded effectors into host cells (18). EGFR loss was observed in DH5α/pLEE-infected C2_BBE cells, but not in DH5α/vector-treated cells, implicating one or more LEE-encoded effectors in this process (Fig. 4). Analyses were restricted to 2- and 4-h infected cells since the focus was solely on determining whether the relevant effectors were LEE-encoded, and since the strain background (DH5α) precluded direct comparison to EPEC-infected host cells.

Fig. 4.

The EPEC locus of enterocyte effacement (LEE) is necessary and sufficient to promote EGFR degradation. C2_BBE cells were treated with medium alone or infected with EPEC, DH5α harboring the cosmid vector pCVD551 (DH5α/vector) or DH5α harboring the LEE-encoding pCVD462 (DH5α/pLEE) at MOI = 100. Total protein extracts isolated at specific time points were immunoblotted for EGFR and actin. Blots shown are representative of 3 similar experiments. Densitometric analysis was performed to determine EGFR abundance relative to actin and normalized to levels in control uninfected cells. P values for specific conditions compared were assessed by ANOVA with Bonferroni post hoc test; *P < 0.01. Error bars represent standard error.

EPEC EspF stimulates the EGFR loss in infected intestinal epithelial cells.

To identify effector(s) influencing EGFR loss, we evaluated phospho- and total EGFR levels in C2_BBE cells infected with isogenic mutants lacking specific virulence proteins. As with WT EPEC infection, EGFR depletion was observed in C2_BBE cells infected with EPEC Δtir, Δmap, ΔespG, ΔespH, ΔnleA, ΔnleH1/2, ΔespG2, and ΔfliC (data not shown), ruling out a role for the corresponding proteins in this process. In the case of ΔespF- and ΔespZ-infected cells, however, the kinetics of EGFR loss was clearly different from that of WT EPEC-infected cells (Fig. 5). In ΔespF-infected cells, EGFR persisted at 6 h postinfection, in contrast to WT-infected C2_BBE cells (Fig. 5A), suggesting that EspF contributes to EGFR loss. To verify whether EspF- and EspZ-dependent EGFR alterations were cell line specific, we performed similar studies in T84 cells. Consistent with the observations in C2_BBE cells, EPEC induced EGFR loss in T84 cells, and this phenotype was dependent on EspF (Fig. 6). Compared with C2_BBE cells, however, EGFR loss in infected T84 cells was more modest at 6 h postinfection.

Fig. 5.

EspF promotes EGFR degradation, whereas EspZ protects infected cells from EGFR loss. C2_BBE cells were treated with medium alone or infected with the wild-type EPEC, ΔespF mutant (A), or ΔespZ mutant (B) at MOI = 100. Total protein extracts collected were analyzed for the abundance of EGFR and actin by Western hybridization. Blots shown are representative of 4 similar experiments. Densitometric analysis was performed to determine EGFR abundance relative to actin and normalized to levels in control uninfected cells. Significant changes were assessed by ANOVA with Bonferroni post hoc test. *P < 0.01 for specific conditions compared. Error bars represent standard error.

Fig. 6.

Validation of EPEC-mediated EGFR degradation in T84 cells. T84 cells were treated with medium alone or infected with EPEC, ΔespF mutant, or ΔespZ strain at MOI = 200. Total protein extracts isolated at specific time points were analyzed for the abundance of EGFR and actin by Western hybridization. Blots shown are representative of 3 similar experiments. Densitometric analysis was performed to determine EGFR abundance relative to actin and normalized to levels in control uninfected cells. Significant changes were assessed by ANOVA with Bonferroni post hoc test. *P < 0.01 and **P < 0.05 for specific conditions compared. Error bars represent standard error.

EPEC EspZ protects against EGFR loss from infected intestinal epithelial cells.

In contrast to ΔespF, ΔespZ infection of both C2_BBE and T84 cells resulted in increased loss of EGFR relative to WT EPEC-infected cells (Figs. 5B and 6). Thus, although EGFR is substantially preserved at 4 h postinfection in WT EPEC-infected C2_BBE cells, it is almost completely depleted in ΔespZ-infected cells at this time point. The time course of EGFR depletion paralleled the caspase activation in ΔespZ-infected C2_BBE cells that we previously reported (29). Since ΔespZ-infected host cells undergo extensive rounding, detachment, and death by 6 h postinfection (29), EGFR levels were not assessed for this time point.

EGFR persistence in infected epithelial cells is dictated by EspF- and EspZ-dependent regulation of caspase activity.

Both EspF and EspZ have been implicated in the regulation of the intrinsic apoptotic pathway caspases (3, 7 and 9) and the survival of EPEC-infected intestinal epithelial cells. EspF localizes to the mitochondria and induces a loss in mitochondrial potential, and consequent activation of caspases 9 and 3 (8, 22–24), whereas EspZ inhibits caspase 3/7/9 activation (29). Caspases, or cysteine-dependent aspartate-directed proteases, have previously been implicated in site-specific EGFR cleavage (DVVD1012 and DNPD1171) (13). Therefore, the pan-caspase inhibitor Q-VD-OPh was used to probe caspase dependence of EGFR loss in EPEC-infected C2_BBE cells (Fig. 7). Consistent with our hypothesis, a Q-VD-OPh dose-dependent restoration of total EGFR was observed in EPEC-infected epithelial cells. Taken together, our results suggest that activated caspases degrade EGFR in EPEC-infected cells. EspF stimulates EGFR cleavage by activating caspases, whereas EspZ delays EGFR loss by inhibiting caspase activation.

Fig. 7.

EGFR loss during EPEC infection is caspase dependent. C2_BBE cells were pretreated with increasing concentrations of the pan-caspase inhibitor Q-VD-OPh (0, 10, 20, or 40 μM) for 2 h. Cells were then treated with medium alone or infected with EPEC at MOI = 100. At 5 h postinfection, total protein extracts were isolated and immunoblotted for EGFR and actin. Densitometric analysis was performed to determine EGFR abundance relative to actin and normalized to levels in control uninfected cells (0 μM Q-VD-OPh). Significant changes were assessed by ANOVA with Bonferroni post hoc test; *P < 0.01, n = 3. Error bars represent standard error.

DISCUSSION

Multiple studies have revealed that EPEC manipulates the survival of host intestinal epithelial cells by eliciting both cytoprotective and cytolethal responses, likely in a temporally regulated manner (7, 8, 15, 28, 29, 33). Although we previously reported on the protective effects of EPEC-induced EGFR phosphorylation, this study demonstrates a type III secretion-dependent downregulation of EGFR signaling later in infection. The effector protein EspF stimulates caspase-dependent EGFR cleavage, whereas the early secreted EPEC molecule EspZ curtails these effects. It is interesting to note that phospho-EGFR levels initially increase even as total EGFR levels begin to decrease in infected cells, possibly suggesting that the phosphorylated receptor is relatively refractory to caspase-dependent cleavage. Formally, however, we cannot rule out the possibility of an alternate role for EGFR loss in infected cells, in addition to, or independent of, the canonical phosphorylation-related signaling pathways.

Caspase-mediated EGFR cleavage has been reported in apoptotic human keratinocyte HaCaT and human cervical epithelial HeLa cell lines (20), as well as during UVA damage of HaCaT cells (26). Curiously, the ubiquitin ligase c-Cbl was also degraded in EPEC-infected cells in a type III secretion-dependent manner (data not shown); it is unclear whether EGFR cleavage precipitates c-Cbl loss or whether c-Cbl is independently cleaved by caspases (11, 40). Although caspase-mediated cleavage of c-Cbl has not been reported, in silico analysis of the protein using the caspase cleavage prediction server CASVM version 1.0 (39) revealed the presence of four putative DxxD caspase cleavage sites (DPFD-435, DDDD-459, DDDD-460, and DGYD-775).

In an earlier study, we demonstrated EGFR-dependent phosphorylation of Akt Ser473 (pAkt) in infected epithelial cells (28). In macrophages, WT EPEC induced a transient increase in pAkt, but, interestingly, Akt phosphorylation was sustained in cells infected with a T3SS-deficient strain (5, 26). EPEC-induced dephosphorylation of host proteins, including pAKT Ser473, contributes to bacterial evasion of phagocytosis by macrophages (12, 26). Although EGFR signaling in EPEC-infected macrophages has not been studied, caspase-dependent downregulation of EGFR could be one mechanism contributing to the eventual loss of pAkt.

EGFR signaling has been implicated in the pathogenesis of several bacterial infections. EGFR activation by Neisseria gonorrhoeae and Salmonella typhimurium enhances bacterial invasion of host epithelial cells (9, 34). EGFR activation in enterohemorrhagic Escherichia coli (O157:H7)- and Helicobacter pylori-infected epithelial cells, respectively, leads to enhanced interleukin-8 production (38). Importantly, pathogenic bacteria have mechanisms to influence the duration of EGFR signaling, typically to sustain it for long periods. Mechanisms include the manipulation of endocytic and/or proteasomal/lysosomal degradation steps in the ligand-induced EGFR downregulation pathway. Helicobacter pylori CagA, for instance, prevents receptor endocytosis, and the prolonged surface exposure of EGFR likely contributes to sustained increase in EGFR signaling and to the neoplastic transformation of host cells (3). Similarly, the Shigella flexneri type III-secreted effector protein IpgD prolongs EGFR signaling by preventing endocytosis and impairing lysosomal degradation (27).

Our observations are consistent with a precipitous loss of viability of infected cells during late stages of EPEC infection. Although host cell survival during the initial phase of infection, promoted by EPEC effector molecules such as EspZ, may be critical for colonization and pathogenesis, it is likely that accelerated cell death during later stages may promote symptomatic disease and facilitate bacterial dispersion.

GRANTS

This research project was supported by the National Institutes of Health (V. K. Viswanathan; NIAID1R01AI081742). Work in the Viswanathan laboratory is also supported by the USDA CSREES Hatch Program (ARZT-5704100-A02-140).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

J.L.R., G.V., and V.V. conception and design of research; J.L.R. and K.A.R. performed experiments; J.L.R., G.V., and V.V. analyzed data; J.L.R., G.V., and V.V. interpreted results of experiments; J.L.R. and K.A.R. prepared figures; J.L.R. drafted manuscript; J.L.R., K.A.R., G.V., and V.V. edited and revised manuscript; G.V. and V.V. approved final version of manuscript.