EutR promotes EA-dependent LEE expression in C. rodentium.

EutR is important for EHEC to sense EA and activate expression of the LEE (16, 17). The LEE includes five major operons (41,–43) (Fig. 2A). The T3SS filament protein is encoded by espA, which is carried in LEE4. Significantly, EA promoted C. rodentium virulence by enhancing expression of espA (Fig. 2B). Moreover, EutR was required for C. rodentium to sense EA and activate LEE expression, as the ΔeutR strain was unresponsive to the addition of EA to the culture medium (Fig. 2B). ler is the first gene within the LEE1 operon (Fig. 2A) and encodes the master regulator of the LEE (43,–46). Consistent with the case with EHEC, EutR was required for EA-dependent ler expression (Fig. 2C), and complementation of the ΔeutR strain restored LEE expression similar to WT levels (Fig. 2C).

Open in a separate window

FIG 2

EA sensing and EutR-dependent LEE expression in C. rodentium. (A) Schematic of the LEE pathogenicity island. (B) RT-qPCR of espA expression in WT and ΔeutR C. rodentium in the absence or presence of EA (n=3). (C) RT-qPCR of ler and espA expression in WT, ΔeutR, and complemented C. rodentium strains grown in the presence of EA (n=3). (D) EMSA of ler, regA, and bla with EutR::MBP. (E) Competition EMSA using biotin-labeled and/or unlabeled ler probes. Ratios compare amount of labeled probe to amount of unlabeled probe. (F) qPCR showing enrichment of ler and rpoA from in vivo ChIP of EutR::MBP compared to MBP (n=2). Error bars represent SD. *, P<0.05; **, P<0.01; ***, P<0.001.

To examine EutR interaction with the ler promoter, we purified EutR::MBP and performed electrophoretic mobility shift assays (EMSAs) using biotinylated ler, regA (previously shown to influence LEE expression in C. rodentium [47]), and bla (negative control) probes. These assays indicated that EutR directly binds the ler promoter to activate expression of the LEE in a specific manner, as EutR did not bind either of the control probes regA and bla (Fig. 2D). Furthermore, MBP alone did not shift the ler promoter (Fig. S3). To confirm specificity of EutR interaction with the ler promoter, we performed a competition EMSA. Addition of an equal concentration of unlabeled ler probe competed with the ler^biotin probe for EutR binding, as visualized by decreased band intensity in shifted DNA compared to that for labeled probe alone (Fig. 2E). Furthermore, when added at increasing concentrations, the unlabeled probe completely outcompeted the labeled ler probe for EutR binding (Fig. 2E). To substantiate the in vitro data, we performed in vivo ChIP. In these experiments, we measured an enrichment of ler in EutR::MBP samples compared to MBP alone as well as compared to the negative control rpoA (Fig. 2F). These data demonstrate that EutR-dependent regulation of LEE expression is conserved in EHEC and C. rodentium.

EutR promotes C. rodentium ler expression during infection.

The in vitro studies identified the LEE genes as EutR regulatory targets. We next tested the importance of EutR in controlling ler expression during intestinal infection. To do this, we constructed a transcriptional reporter in which the ler promoter was cloned upstream of the bacterial luciferase genes (Fig. S4A) and verified that this plasmid was retained during infection (Fig. S4B). At 10days postinfection (dpi), ler expression was significantly decreased in the ΔeutR strain compared to that in the WT (Fig. 3A and ​andB).B). To substantiate these data and control for potential differences in bacterial colonization, we infected mice with the WT or ΔeutR strain and harvested tissue at 10dpi for transcriptional analyses of ler expression normalized to 16S rRNA expression. These data were consistent with data shown in Fig. 2, in which the eutR deletion resulted in a statistically significant (~2-fold) decrease in ler expression (Fig. 3C). Altogether, these data indicate that EutR is important for sensing the host environment and promoting LEE expression in the GI tract. Moreover, the magnitude differences in ler gene expression between WT and ΔeutR measured in whole colons versus normalized to C. rodentium CFU suggest that EutR affects pathogen burden in the intestine.

Open in a separate window

FIG 3

EutR enhances C. rodentium virulence gene expression in vivo. (A) Representative bioluminescent image of ler expression in mice infected with WT or ΔeutR C. rodentium at 10dpi. (B) Quantification of bioluminescent data. (C) RT-qPCR of ler expression in WT and ΔeutR C. rodentium associated with colonic tissue. *, P<0.05; **, P<0.01.

EutR is required for maximal C. rodentium fitness in the intestinal tract.

The C. rodentium infection cycle in C57BL/6 mice is characterized by initial low levels of C. rodentium, followed by proliferation, and steady state, before clearance (Fig. 4A) (29). To examine how EutR affects C. rodentium colonization, we determined numbers of WT and ΔeutR strain CFU shed in stool throughout the infection cycle. Similar levels of ΔeutR and WT C. rodentium CFU were measured during the initial stages of infection. However, starting at 8dpi and continuing until the end of the experiment, the ΔeutR strain was recovered at lower numbers than the WT (Fig. 4B). These findings reveal an important role for EutR in promoting colonization and persistence in the mammalian GI tract.

Open in a separate window

FIG 4

EutR is required for robust C. rodentium colonization independent of changes in EA availability. (A) CFU of WT C. rodentium shed in stool during infection. Data are presented as medians ± interquartile ranges. (B) CFU of WT and ΔeutR C. rodentium shed in stool. The dotted line represents the limit of detection. (C) EA levels in colons of mock-infected or WT C. rodentium-infected mice. *, P<0.05.

C. rodentium infection might affect EA levels in the intestine, which could be a contributing factor in the differences in colonization between the WT and ΔeutR strains measured during later stages of infection. To examine this possibility, we harvested whole colons from mock- or C. rodentium-infected mice at 6 and 14dpi and measured EA levels in the tissue homogenates using liquid chromatography-mass spectrometry (LC-MS). At each time point, similar EA concentrations were measured from the colons of mice infected with 10⁹ CFU of C. rodentium and mock-infected mice (Fig. 4C). These data suggest that C. rodentium infection does not alter EA concentrations in the intestine and that differences in WT and ΔeutR strain colonization are not due to changes in EA availability in the colon.

EutR is important for transmission efficiency.

Passage through the GI tract results in a hypertransmissible state in which significantly fewer C. rodentium CFU are sufficient to cause infection than for bacteria cultivated in vitro (48). In C57BL/6 mice, ingestion of 10⁷ to 10⁸ CFU of freshly shed bacteria results in infection (48), compared to 10⁹ CFU of C. rodentium grown in broth. At 10dpi, the median CFU of the ΔeutR strain shed in stool was ~10-fold lower than the reported threshold that results in efficient transmission between animals. Therefore, we next investigated how EutR affects transmission. Following 24 h of cohousing with 6-dpi donor mice, equal numbers of WT and ΔeutR C. rodentium organisms were recovered in stool from recipient mice (Fig. 5A). In contrast, following cohousing with 10-dpi donor mice, the ΔeutR strain was recovered at significantly lower levels in stool of (formerly) naive mice (Fig. 5B and Fig. S5), indicating that EutR is important for pathogen persistence not only during primary infection but also within a host population.

Open in a separate window

FIG 5

EutR promotes C. rodentium transmission. (A) CFU of WT and ΔeutR C. rodentium shed in stool of recipient mice following cohousing with 6-dpi donor mice. (B) CFU of WT and ΔeutR C. rodentium shed in stool of recipient mice following cohousing with 10-dpi donor mice.

Culture media.

Bacteria were grown in Luria-Bertani medium (LB), Dulbecco’s modified Eagle medium (DMEM; Invitrogen), or M9 minimal medium (55) that was supplemented with 10mM EA hydrochloride (Sigma) as the sole nitrogen source. Cyanocobalamin (150nM; Sigma) was added to the medium whenever EA was added. In general, bacteria were grown overnight (OVN) with shaking in LB at 37°C and then diluted 1:100 into fresh medium for experimental analysis.

EutR amino acid alignment.

EutR sequences were obtained using the annotated EHEC EDL933 (56) and C. rodentium (57) genome sequences. The EMBOSS Needle program (58) was used to generate the alignment.

Transcriptional fusions.

C. rodentium gDNA including the ler promoter, including 400 upstream and 43 nucleotides downstream of the ATG start site, was amplified by PCR. Resulting PCR products were cloned into the pGEN-luxCDABE vector (54) (Addgene MTA).

Purification of EutR under native conditions.

The plasmid expressing EutR::MBP was constructed by amplifying the eutR gene from C. rodentium strain DBS100 using primers EutRMBP_F1 and EutRMBP_R1 and cloning the resulting PCR product into the NcoI/BamHI cloning sites of vector pMAL-c5X (New England BioLabs [NEB]). In order to purify MBP and EutR::MBP, E. coli strain BL-21(DE3) containing MBP or EutR::MBP was grown at 37°C in LB with glucose (0.2% [final concentration]) and ampicillin (100μg/ml) to an optical density at 600nm (OD₆₀₀) of 0.5. Then, isopropyl-β-d-thiogalactopyranoside (IPTG) was added to a final concentration of 0.3mM, and protein expression was induced OVN at 18°C. Cells were harvested by centrifugation at 4,000×g for 20min, then resuspended in 25ml of column buffer (20mM Tris-HCl, 200mM NaCl, 1mM EDTA), and lysed by homogenization. The lysed cells were centrifuged, and the lysate was loaded onto a gravity column (Qiagen) with amylose resin. The column was washed with column buffer and then eluted with column buffer containing 10mM maltose. Fractions containing purified proteins were confirmed by SDS-PAGE and Western blot analysis. Protein concentration was determined using a NanoDrop spectrophotometer.

ChIP-qPCR.

Chromatin immunoprecipitation (ChIP) was performed using the C. rodentium ΔeutR strain transformed with pEutR::MBP or pMBP. Strains were grown in DMEM supplemented with EA, cyanocobalamin, and 2.5μM IPTG until cells reached an OD₆₀₀ of ~0.8. Cross-linking and ChIP were performed based on established methods (5, 59). After growth, formaldehyde was added (1% [final concentration]), and cells were incubated for 15min at room temperature. Reactions were quenched with 0.5 M glycine, and then samples were pelleted, resuspended in phosphate-buffered saline (PBS), and washed. Cells were lysed with 2mg/ml of lysozyme at 37°C for 30min. Subsequently, samples were placed on ice and sonicated. Insoluble cell debris was removed by centrifugation, and supernatants were collected. RNase A was added to the samples, which were then incubated at 37°C for 1 h. One hundred units of Benzonase nuclease (EMD Millipore) was added to digest DNA into smaller fragments (60). Immunoprecipitation was performed using amylose beads (NEB) for 2 h at 4°C with gentle mixing. Beads were pelleted and washed. Then samples were incubated for 10min at 65°C in elution buffer with occasional gentle mixing. Samples were centrifuged and supernatants were collected. To reverse the cross-link, samples were boiled for 10min and DNA was purified using the Qiagen PCR purification kit. For ChIP-qPCR, an aliquot of each reaction was taken prior to immunoprecipitation, de-cross-linked by boiling for 10min, and purified for use as the input control. The fold enrichment of each promoter represents the value of the immunoprecipitated DNA divided by the input unprecipitated DNA. These values were normalized to the values obtained for each promoter precipitated using MBP empty vector in order to account for nonspecific enrichment.

EMSAs.

Electrophoretic mobility shift assays (EMSAs) were performed basically as described previously (5). We modified the protocol to use 5′-biotinylated primers to generate the labeled DNA probes. EMSAs were performed by adding increasing amounts of purified EutR protein to end-labeled probe (30ng) in binding buffer (10mM Tris-HCl [pH 8.0], 1mM Na-EDTA, 80mM NaCl, 10mM β-mercaptoethanol, and 4% glycerol) (61), and reactions were incubated for 20min at 37°C. For competition EMSAs, 25ng of labeled probe was used with 0ng, 25ng, 125ng, and 250ng of unlabeled probe added. Immediately before the samples were loaded, Ficoll solution was added to the reaction mixtures. The samples were electrophoresed for approximately 4 to 5 h at 175 V on a 6% polyacrylamide gel, transferred to Zeta-Probe membranes, and visualized using the chemiluminescent nucleic acid detection module kit (Thermo Scientific).

Animal infections.

All experiments were approved by the Institutional Animal Care and Use Committee at the University of Virginia School of Medicine. Female C57BL/6J (6- to 10-week-old) mice were inoculated by oral gavage of 1×10⁹ CFU of C. rodentium. Mice were monitored daily for weight loss and other signs of distress. At the end of each experiment, mice were euthanized by CO₂ asphyxiation followed by cervical dislocation.

Enumeration of CFU from stool.

For enumeration of CFU from stool, fresh fecal pellets were collected at the desired time points, weighed, and homogenized in 1ml of PBS. Serial dilutions were plated on LB plates containing appropriate antibiotic selectivity to evaluate CFU. Enumerated CFU per gram of stool were log transformed prior to statistical analysis, as previously described (62).

RNA extraction and RT-qPCR.

Bacterial cells grown in vitro were resuspended in TRIzol (Life Technologies). Colonic tissue was harvested, washed with PBS to remove luminal contents, and then homogenized in TRIzol. For all samples, RNA was extracted using a PureLink RNA minikit (Invitrogen). Reverse transcription-qPCR (RT-qPCR; for in vitro and colonic samples) was performed as previously described (63) in a one-step reaction using an ABI 7500-FAST sequence detection system and software (Applied Biosystems). For each 10-μl reaction mixture, 5μl of 2× SYBR master mix (Ambion), 0.05μl of Multi-Scribe reverse transcriptase (Invitrogen), and 0.05μl of RNase inhibitor (Invitrogen) were added. For stool samples, cDNA was synthesized using SuperScript II reverse transcriptase (Invitrogen) with random primers prior to qPCR to overcome low RNA yields. Reactions are identical to those described for RT-qPCR omitting reverse transcriptase and adjustment for a final 10-μl volume. Primers were designed using Primer BLAST (NCBI) to ensure no cross-reactivity to other genes in the C. rodentium chromosome. Amplicon length was approximately 100bp. Amplification efficiency of each primer pair was verified using standard curves of known DNA concentrations. Melting-curve analysis was used to ensure template specificity by heating products to 95°C for 15 s, followed by cooling to 60°C and heating to 95°C while monitoring fluorescence. After the amplification efficiency and template specificity were determined for each primer pair, relative quantification analysis was used to analyze the samples using the following conditions for cDNA generation and amplification: 1cycle at 48°C for 30min, 1cycle at 95°C for 10min, and 40 cycles at 95°C for 15 s and 60°C for 1min. Two technical replicates of each biological replicate were included for each gene target. Data were normalized to the reference controls rpoA (in vitro samples) or 16S rRNA specific to C. rodentium (tissue samples) and analyzed using the comparative critical threshold (C_T) method (64). The expression levels of the target genes were compared using the relative quantification method (64).

EA measurements.

EA levels were measured as previously described (13), with minor modifications. Colons were harvested, weighed, and homogenized in 1.5ml of ultrapure water. Homogenized tissue was incubated with 150μl of 40% sulfosalicylic acid (Sigma) for 15min and then centrifuged at 11,000×g for 15min at room temperature. Then 1ml of the supernatant was mixed with 300μl of 0.5 M NaHCO₃ (Sigma), 2ml of 20-mg/ml dansyl chloride solution (Sigma) in acetone, and 200μl of 1 M NaOH (Fisher). Following incubation in the dark for 20min at room temperature, 200μl of 25% NH₄OH (Sigma) was added. The volume was adjusted to 5ml with acetonitrite (Sigma), and 1ml was centrifuged at 11,000×g for 1min. The supernatant was taken for analysis. The LC-MS system consists of a ThermoElectron Orbitrap ID-X mass spectrometer with a Heated Electrospray Ionization source interfaced to a Thermo Accucore Vanquish C₁₈ 1.5-μm, 2.1-by100-mm column. Two microliters of the extract was injected and the compounds were eluted from the column by a methanol-0.1% formic acid gradient at a flow rate of 250μl/min (total time, 15 min). The nanospray ion source was operated at 3.2kV. The sample was analyzed by MS and tandem MS (MS/MS). The dansyl-EA was detected as a peak at ~4.36min and a mass of 295.111+. A 1 M in vitro sample of EA hydrochloride (Sigma) taken through the dansylation process was used to generate a standard curve.

In vivo imaging system (IVIS) analysis.

For in vivo bioluminescence on living mice, an IVIS Spectrum (Caliper Lifesciences, Alameda, CA) was used. Luminescence (radiance) was quantified using the software program Living Image (Xenogen). Radiance is reported as photons per second.

Transmission studies.

Donor mice were infected with WT or ΔeutR C. rodentium for 6 or 10days. At these time points, each donor mouse was placed in a fresh cage with 3 naive mice for 24 h. For data shown in Fig. 5, at 24 h after cohousing, all mice were separated and placed in fresh, individual cages. For data shown in Fig. S5, respective recipient mice were cohoused after exposure to donor mice. At 2 or 4days after cohousing with donor mice, stool was collected from each recipient mouse, homogenized in PBS, and plated on appropriate selective media to determine CFU of transmitted C. rodentium.

We thank members of the Kendall lab for thoughtful discussions on this work and Hervé Agaisse for feedback on the manuscript. We thank Sarah Ewald for advice on the IVIS experiments and James Nataro for the Citrobacter rodentium DBS100 strain. We thank the Biomolecular Analysis Facility at the University of Virginia for LC-MS and analysis services.

This work was supported by National Institutes of Health NIAID grants AI118732, AI130439, and AI146888 to M.M.K. C.A.R. received support from NIH training grant 5T32AI007046 and a UVA Wagner Fellowship. A.B.S. received support from NIH training grant 5T32AI055432.

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.