The RpoE Stress Response Pathway Mediates Reduction of the Virulence of Enteropathogenic Escherichia coli by Zinc

The bacterial strains and plasmids used in this study are listed in Table 1. Liquid cultures were grown in lysogeny broth (LB) at 37°C with aeration and with ampicillin (100 μg/ml) or kanamycin (50 μg/ml). Cultures were also grown on LB agar plates with antibiotic selection at 37°C. Dulbecco's modified Eagle's medium (DMEM) was prepared from DMEM-D5030 (Sigma-Aldrich, St. Louis, MO) and supplemented with 1 g/liter glucose, 17.9 mM sodium bicarbonate, 25 mM HEPES, 0.543 mM adenine sulfate, 4.00 mM l-glutamine, and 1.00 mM sodium pyruvate, pH 7.4. For routine maintenance of tissue culture, DMEM was prepared by supplementing F-12 medium (D5523; Sigma-Aldrich) with 10% fetal bovine serum (FBS). Unless otherwise noted, DMEM in the rest of study refers to DMEM-D5030. A working stock of zinc acetate was dissolved in 1% glacial acetic acid solution to 100× or 1,000× treatment concentrations.

Bacteria grown overnight in LB were diluted 1:100 in DMEM for subsequent assays. Expression of pLC245 or pTrunc RpoE was induced by 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG); expression of pRseA was induced by 0.2% arabinose.

PCR was conducted to amplify a truncated rpoE insert from pLC245 template (Table 1) with EcoRI_rpoE_trunc_fwd and XbaI_rpoE_trunc_rev primers (Table 2) using Q5 High-Fidelity DNA polymerase (New England BioLabs, Ipswich, MA) according to standard procedures. The insert was digested with EcoRI and XbaI and ligated into pTrc99a backbone vector. The construct was sequenced and confirmed to be correct by the core sequencing service at the Vollum Institute (Oregon Health and Science University) using pTrc99a sequencing primers (Table 2).

Turbidity was measured by an Infinite 200 Pro microplate reader (Tecan, Männedorf, Switzerland) at 600-nm absorbance. Bacteria grown in DMEM or in DMEM supplemented with inducer were incubated at 37°C with shaking for the amount of time indicated on the figures. IPTG (1 mM) was used to induce expression of RpoE from plasmids pLC245 and pTrunc RpoE, while 0.2% l-arabinose was added to induce RseA expression from plasmid pRseA. For zinc susceptibility assays, bacteria were grown in medium with or without inducer at 37°C with shaking for 90 min, and then zinc acetate or vehicle (1:1,000 dilution of 10% glacial acetic acid) solutions were added into the cultures directly. Cultures were further incubated and measured for 600-nm absorbance for 160 min. Resulting cultures were serially diluted and plated on nonselective or ampicillin-selective LB agar plates in triplicates and incubated at 37°C overnight. Viability was assayed by counting the CFU per milliliter of culture.

Bacteria grown in DMEM supplemented with 0.3 mM zinc acetate, 0.5 mM zinc acetate, or 1% glacial acetic acid (vehicle) were incubated at 37°C with aeration for 4 h. Turbidity of each sample as the absorbance at 600 nm was measured, and each sample was equalized to an optical density at 600 nm (OD₆₀₀) of 0.15. Bacterial samples were centrifuged at 8,000 × g for 10 min to collect cell pellets that were then resuspended in an equal volume of phosphate-buffered saline (PBS; pH 7.4). A group of vehicle-treated sample pellets were resuspended in −20°C 70% ethanol (EtOH) for 15 min and centrifuged to collect the pellets, which were resuspended in PBS as a positive control. SYTOX blue (Invitrogen, Carlsbad) was added to the mixture to a final concentration of 5 μM, and the mixture was incubated at room temperature for 10 min. Cell pellets were collected by centrifugation again at 8,000 × g for 10 min and then resuspended in 20 mM MgSO₄. Cells were immobilized on 2% agar pads. Samples were then immediately imaged on a Nikon H550L fluorescence microscope at ×400 magnification using a 40× water immersion objective. At least seven fields of view were captured for each sample. For each field of view, dark-field and 4′,6′-diamidino-2-phenylindole (DAPI) filter images were obtained. Cells in each image were counted by CellProfiler, version 2.0 (Broad Institute, Cambridge, MA), and the ratio of fluorescent cells to total cells was computed. Statistically significant differences in ratios between treatments were determined by a two-tailed, unequal variance Student's t test, and a P value of <0.05 was considered significant.

Bacteria grown in DMEM or in DMEM supplemented with inducer were incubated at 37°C with aeration for 2 h. For zinc treatment, zinc acetate or vehicle (1% glacial acetic acid) solutions were added to the cultures in the beginning of incubation or 10 min before RNA extraction. One milliliter of each culture was extracted with TRI reagent (Sigma-Aldrich, St. Louis, MO). RNA samples were reverse transcribed with a SuperScript III first-strand synthesis kit (Life Technologies, Carlsbad, CA) using random hexamer primers provided with the kit. Reverse-transcribed cDNA samples were subsequently diluted 1:10 with diethyl pyrocarbonate (DEPC)-treated H₂O to reduce inhibitory effects of reverse transcription reagents on quantitative PCR (qPCR) efficiency.

Transcripts were assayed with ImmoMix (Bioline Reagents, London) using primers listed in Table 2. The delta-delta comparative method (22) was used to analyze transcriptional changes in target genes using rrsB as the reference gene. A standard curve was generated by serially diluting cDNA samples 1:4, 1:16, 1:64, and 1:256. PCR efficiency was determined by analyzing standard curves of each primer set; qPCR efficiencies were determined to be similar (<10% difference) between experimental and control groups. Data were analyzed with the REST 2009 relative expression software tool (Roche, Penzberg, Germany) with analysis of error propagation. Statistical significance was determined by a two-tailed, unequal variance Student's t test, and a P value of <0.05 was considered significant.

Bacteria were grown in DMEM at 37°C with aeration in the presence of 1 mM IPTG for 5 h. Absorbance at 600 nm was measured for each culture. A total of 100 μl of each culture equalized to 1.50 absorbance was mixed with 100 μl of 6× SDS-PAGE loading buffer and incubated on a 95°C heat block for 10 min. Twenty microliters of dyed lysates was loaded onto a 4 to 20% Mini-Protean TGX Gel (Bio-Rad, Hercules) and separated at 100 V for 80 min. Proteins were transferred to a polyvinylidene difluoride (PVDF) membrane at 100 V for 60 min. After the membrane was blocked with 5% nonfat milk, it was probed with anti-EspC, anti-Tir, and anti-DnaK primary antibodies (Enzo Life Sciences, Farmingdale, NY) and rabbit anti-mouse horseradish peroxidase conjugate as the secondary antibody (Life Technologies, Carlsbad, CA) and developed using the Western Lightning enhanced chemiluminescence method (Thermo Fischer Scientific, Waltham).

HEp-2 (human larynx carcinoma) cells were seeded at a density of 6 × 10⁴ cells per well in a 24-well plate containing round coverslips in DMEM (4.5 mg/ml glucose) supplemented with 10% FBS and 25 μg/ml gentamicin. HEp-2 cells were grown for 60 h in a 4.5% CO₂ atmosphere at 37°C. Overnight bacterial inoculants were diluted 1:100 in DMEM (1 mg/ml glucose) supplemented with 2% FBS, 1% d-mannose, and 1 mM IPTG and grown at 37°C with aeration for 2 h. Each well containing HEp-2 cells was washed twice with PBS, and 1 ml of diluted bacteria was added directly to each well. The cells were then coincubated at 37°C for 4 h. Each well was washed twice with PBS and then fixed in 4% paraformaldehyde for 10 min; cells were then washed once with PBS, treated with 0.1% Triton X-100 for 10 min, and washed again with PBS. Samples were then incubated in 5 μg/ml of fluorescein isothiocyanate (FITC)-labeled phalloidin (Sigma-Aldrich, St. Louis) for 45 min at 37°C. Each well was washed once with PBS and then incubated in 300 nM DAPI stain for 5 min. After being washed again with PBS and mounted onto slides, samples were randomly indexed. They were then visualized and photographed with a confocal microscope (Nikon Eclipse) using a 60×, 1.4 numerical aperture (NA) oil immersion objective lens. Samples were analyzed blindly. At least three field images were taken for each slide, with at least 400 cells examined for each treatment. Actin-rich pedestals beneath DAPI-stained bacteria were counted and quantified. Statistical significance was determined by a two-tailed, unequal variance Student's t test, and a P value of <0.05 was considered significant.

All plasmid constructs were confirmed to be correct by DNA sequencing analysis performed at the Vollum Institute at the Oregon Health and Science University.

Previous studies have demonstrated that micromolar concentrations of zinc acetate suppressed virulence gene expression and protein secretion in EPEC, with only a modest decrease in cell density after overnight growth in DMEM (6, 14, 15). To determine whether these results were due to an altered growth rate or viability of EPEC in DMEM with low glucose (0.1%), which is optimal for virulence expression, we monitored EPEC growth in this medium in the presence of zinc acetate. The doubling times for EPEC during balanced growth in the presence of vehicle treatment and of 0.3 mM, 0.5 mM, and 1.0 mM zinc acetate treatments were 39.4 min, 46.4 min, 73.3 min, and 117 min, respectively (Fig. 1A). We found that while the bacterial growth rate was significantly decreased due to the addition of zinc (comparison of linear fit slopes, P < 0.05), bacterial viability was not significantly different between zinc acetate and vehicle treatment groups when the number of CFU per milliliter was determined at the 4-h time point, which is after cultures had reached the stationary phase of growth (Fig. 1B). This result was consistent with the finding that cell densities were similar after overnight growth in DMEM with and without 0.4 mM zinc acetate (14), suggesting that zinc in the 400 μM range did not exert bactericidal effects on EPEC. These results show that growth is slowed in the presence of micromolar concentrations of zinc, that the time to reach stationary phase is increased, but that growth yield is not altered. Thus, inhibition of EPEC virulence by micromolar concentrations of zinc was likely not caused by a reduction in bacterial viability but, rather, by changes in bacterial physiology and/or by regulatory phenomena.

To elucidate further the effect of zinc on EPEC, we quantified the number of bacterial cells permeable by SYTOX blue, a nucleic acid stain that penetrates only cells with compromised membranes, after treatment with 0.3 and 0.5 mM zinc acetate. Treatments with 70% EtOH and a dilution of 1% glacial acetic acid solution were used as positive and negative vehicle controls, respectively. We observed a dosage-dependent increase in cell permeability in response to the different concentrations of zinc acetate (Fig. 2). The percentage of permeable cells was significantly higher when samples were treated with 0.5 mM zinc acetate (20.9%) than with 0.3 mM zinc acetate (7.8%). Combined with results from Fig. 1, this result demonstrated that zinc acetate compromised membrane integrity at micromolar concentrations without significantly reducing bacterial viability, likely contributing to the increased EPEC doubling time compared to that of the vehicle control in DMEM containing 0.1% glucose (Fig. 1).

Given that zinc increased membrane permeability of EPEC without decreasing its viability, we predicted the envelope stress regulon to be activated in response to zinc in order to promote bacterial survival. Zinc was previously reported to upregulate the transcription of rpoE and, as part of its regulon, degP, in a K-12-derived strain of E. coli (23, 24). We previously demonstrated that transcription of rpoE in a K-12-derived strain was modestly increased and that LEE gene expression was downregulated in EPEC strains in response to zinc using multicopy, plasmid-derived reporter gene fusions (15). In this study, we used qRT-PCR to quantify the effect of 0.3 mM zinc acetate on the transcriptional level of envelope stress markers degP, rpoE, and dsbA as well as LEE virulence genes, ler (LEE1) and espA (LEE4), in EPEC. Transcription of dsbA was measured to assess activation of the Cpx pathway in response to zinc as a control as dsbA is part of the Cpx regulon.

Upon 10 min of exposure to zinc acetate, envelope stress markers were all significantly upregulated, indicative of an activated envelope stress response (Fig. 3A). We observed the activating effect of zinc on the RpoE-dependent stress response as both rpoE and degP were significantly upregulated. In contrast to previous findings (24), we saw a 2-fold upregulation of dsbA, suggesting that the Cpx two-component response was also modestly activated by zinc. Upon 120 min of exposure to zinc acetate, rpoE was no longer transcriptionally increased (Fig. 3A). RseA, an inhibitor of RpoE, is encoded as part of a four-gene operon along with rpoE, which might explain why the prolonged treatment with zinc exerted negative feedback on the transcription of rpoE via the expression of RseA. The LEE virulence genes ler and espA were downregulated 2-fold when bacteria were treated with 0.3 mM zinc acetate for 120 min (Fig. 3B). Thus, activation of the RpoE regulon, which occurs mostly by posttranslational mechanisms, was associated with downregulation of LEE virulence genes in EPEC.

RpoE activity is posttranslationally controlled by RseA, the envelope-bound, sequestering inhibitor of RpoE. Stress signals and outer membrane protein dynamics can trigger cleavage of RseA and subsequent release of RpoE for downstream activation of the envelope stress response pathway. We therefore took advantage of an inducible RseA construct carried on a plasmid vector to investigate whether the activation of an RpoE-dependent response was necessary for bacterial survival in the presence of 0.3 mM zinc acetate. We measured the doubling times of strain E2348/69 carrying pRseA grown in the presence and absence of the 0.2% arabinose inducer with and without the addition of 0.3 mM zinc acetate (Fig. 4).

As a control, the doubling times comparing the absence and presence of arabinose, inducing RseA, were 38.8 and 43.3 min, respectively, which represent only a modest decrease in growth rate (Fig. 4A). Induction of RseA in the absence of zinc addition led to a 10-fold reduction in EPEC viability (Fig. 4B), suggesting that basal activity of RpoE is necessary for bacterial survival. With the addition of zinc, the doubling times in the absence and presence of RseA overexpression were 51.2 and 68.5 min, respectively, indicating a 34% decrease in growth rate (Fig. 4C; comparison of linear fit, P > 0.05). This demonstrated that induction of RseA was able to further reduce the bacterial growth rate when it was combined with 0.3 mM zinc acetate. As previously demonstrated, mutation in rpoE sensitized EPEC to a variety of divalent metal ions, including zinc (23). Deletion of rpoE is lethal in E. coli without uncharacterized, compensatory mutations (25, 26). Our viability results demonstrated support for these two findings.

Induction of RseA combined with 0.3 mM zinc acetate treatment reduced EPEC viability of cultures grown in the presence of arabinose by ∼100-fold relative to growth in the absence arabinose (Fig. 4D). Our data demonstrate that the absence of a functional RpoE regulon compromises EPEC viability in the presence of only 0.3 mM zinc acetate.

To address whether the expression of RpoE alone was sufficient to suppress virulence gene expression, we transformed EPEC with an inducible RpoE-carrying vector (pLC245) and an empty vector (pTrc99a) control. EPEC was grown in the presence of IPTG, and the transcriptional levels of envelope stress markers and LEE virulence genes were measured via qRT-PCR (Fig. 5). As previously reported in E. coli (27), induction of RpoE led to over 3-fold upregulation of degP with no upregulation of dsbA, which is part of the Cpx regulon, showing that the inducible vector is functional in EPEC. We observed an approximate 10-fold downregulation of the ler (LEE1) and espA (LEE4) genes when RpoE was induced, indicating that the induction of RpoE either directly inhibited the transcription of LEE virulence genes or altered other regulatory elements in order to affect expression (Fig. 5A). As a control for the effect of IPTG, we measured no significant difference in transcriptional levels of the genes of the EPEC strain carrying the empty vector in the presence and absence of IPTG (data not shown). We concluded that zinc induced the RpoE regulon and that overexpression of RpoE resulted in downregulation of LEE virulence genes in EPEC.

The alternate sigma factor RpoE attaches to RNA polymerase via its N terminus and directs transcription by binding to consensus promoter sequences via a helix-turn-helix motif (HTH) on its C terminus (28–30). We determined whether a mutated variant of RpoE, carried on a multicopy plasmid, would alter target gene expression in the presence of the chromosomally located wild-type (wt) rpoE gene. To do this we generated a truncated variant of RpoE lacking the −35 promoter sequence-binding motif in its C terminus (residues 167 to 190) (see Fig. S1 in the supplemental material) (30). As a control, we showed that EPEC doubling time and viability were not affected by overexpression of either the wt or truncated RpoE (data not shown). Expression of the truncated RpoE variant failed to upregulate degP, as expected (Fig. 5B). Also as predicted, ler was no longer downregulated when the truncated RpoE was overexpressed. However, induction of truncated RpoE still led to significant downregulation of espA (Fig. 5B). This suggested that espA and ler are differentially regulated by the induction of RpoE and that only the regulation of ler is dependent upon the DNA-binding capacity of RpoE. As the overexpression of the truncated RpoE no longer regulated ler, we concluded that the RNA polymerase holoenzyme containing σ^E regulated LEE1 either directly or indirectly through the expression of another gene product under the control of an RpoE-dependent promoter. It was unclear how RpoE controlled espA expression.

To monitor protein expression, we performed an immunoblot assay using whole-cell proteins and found that both Tir (LEE5) and EspC expression levels were significantly decreased by induction of both RpoE and truncated-RpoE (Fig. 5C). These data were consistent with our observations by qRT-PCR showing LEE1 and LEE4 downregulation upon RpoE overexpression (Fig. 5A and B).

To evaluate the effect of RpoE and truncated RpoE overexpression on virulence, we employed a fluorescent actin stain (FAS) assay to assess the formation of attaching and effacing (A/E) lesions, a hallmark of EPEC infection on HEp-2 host cells. Images of HEp-2 cells alone or with coincubated EPEC strains were taken by confocal scanning laser microscopy (Fig. 6A to E). Total host cells and host cells with A/E lesions were counted blindly for each treatment. Our results showed that induction of RpoE in EPEC significantly reduced formation of A/E lesions by 39% compared to the control level (Fig. 6F). Intriguingly, induction of truncated RpoE reduced A/E lesion formation even further than induction of wild-type RpoE, with a 62% reduction in A/E lesions compared to the level in the control (Fig. 6D and F). Given that the secreted protein Tir is essential for intimin-mediated binding and the formation of pedestal structures by EPEC, this finding is consistent with our observation previously that Tir protein expression was most robustly suppressed by the induction of truncated RpoE (Fig. 5C), whereas induction of wild-type RpoE led only to a modest reduction of Tir expression. Our results clearly show that induction of both wild-type RpoE and truncated RpoE led to diminished LEE gene expression and attenuated A/E lesion formation. The results demonstrate that RpoE modulates EPEC virulence gene expression in response to zinc by multiple mechanisms, one of which requires the DNA-binding motif of RpoE.