The Interacting Cra and KdpE Regulators Are Involved in the Expression of Multiple Virulence Factors in Enterohemorrhagic Escherichia coli

Abstract

The human pathogen enterohemorrhagic Escherichia coli (EHEC) O157:H7 codes for two interacting DNA binding proteins, Cra and KdpE, that coregulate expression of the locus of enterocyte effacement (LEE) genes in a metabolite-dependent manner. Cra is a transcription factor that uses fluctuations in the concentration of carbon metabolism intermediates to positively regulate virulence of EHEC. KdpE is a response regulator that activates the transcription of homeostasis genes in response to salt-induced osmolarity and virulence genes in response to changes in metabolite concentrations. Here, we probed the transcriptional profiles of the Δcra, ΔkdpE, and Δcra ΔkdpE mutant strains and show that Cra and KdpE share several targets besides the LEE, but both Cra and KdpE also have independent targets. Several genes within O-islands (genomic islands present in EHEC but absent from E. coli K-12), such as Z0639, Z0640, Z3388, Z4267, and espFu (encoding an effector necessary for formation of attaching and effacing lesions on epithelial cells), were directly regulated by both Cra and KdpE, while Z2077 was only regulated by Cra. These studies identified and confirmed new direct targets for Cra and KdpE that included putative virulence factors as well as characterized virulence factors, such as EspFu and EspG. These results map out the role of the two interacting regulators, Cra and KdpE, in EHEC pathogenesis and global gene regulation.

INTRODUCTION

In order to adapt to changes in their environment, bacteria have developed signaling mechanisms that allow sensing of environmental cues to differentially regulate appropriate genes. These signals may be sensed by sensor kinases that can be membrane bound or cytoplasmic (1, 2). These signals are then transduced to response regulators (RR) that act on downstream genes, most often to regulate transcription. For example, QseB, an OmpR family member, is a response regulator that is phosphorylated by its cognate kinase, QseC, and, depending on its phosphorylation state, can either be an activator or inhibitor of gene expression (3, 4). The kinase QseC also has been shown to phosphorylate the noncognate response regulators QseF and KdpE (4). KdpE and its cognate kinase KdpD are important for K⁺ transport (5–7). The membrane-bound KdpD responds to K⁺ limitation or salt-induced high osmolarity to increase its phosphorylation state (8–10). The phosphorylated KdpD then transfers its phosphoryl group to KdpE, which, by binding to the promoter region of the kdpFABC operon, activates the transcription of these genes, consequently adjusting intracellular K⁺ levels to maintain homeostasis (7, 8, 11).

Differential gene transcription can also be controlled by transcriptional activators that bind to the promoter regions of their targets independently of phosphorylation. These regulatory proteins respond to intracellular cues known as inducers. The catabolite activator protein (CRP), for example, is a transcriptional factor that responds to changes in intracellular cyclic AMP (cAMP) levels (12, 13) to regulate its targets. Another protein, catabolite repressor protein (Cra; also known as FruR), has also been identified as a transcriptional factor that utilizes fluctuations in sugar concentrations to positively or negatively regulate target genes (14–17). Cra is a member of the LacI/GalR family, which has been shown to positively regulate the genes encoding gluconeogenic enzymes, such as fructose-1,6-diphosphatase, and negatively regulate genes encoding glycolytic enzymes, such as phosphofructokinase, and the energy-coupling proteins of the bacterial phosphotransferase system (18–21). Cra's function is inhibited by the presence of micromolar concentrations of fructose-1-phosphate (F1P) or millimolar amounts of fructose-1,6-bisphosphate (FBP) (14, 16). These metabolic intermediates bind to the inducer binding domain of Cra, decreasing its binding affinity for target promoters and consequently decreasing its regulatory function. Cra has been shown to regulate virulence in Salmonella enterica, with the cra mutant being avirulent in murine infections (22–24), and to be a facilitator of Shigella flexneri pathogenesis (25).

EHEC uses response regulators and transcriptional regulators in general to translate signals sensed by the bacterium into gene activation or inhibition. EHEC employs these regulators not only for the maintenance of bacterial homeostasis but also to differentially regulate virulence (26). One important characteristic of EHEC virulence is its ability to form attaching and effacing (AE) lesions on epithelial cells. AE lesions are characterized by the attachment of bacteria to the host epithelium, followed by the induction of extensive actin rearrangement within the epithelial cells, culminating in the formation of pedestal-like structures underneath the bacteria (27–30). Most of the genes necessary for AE lesion formation are contained within a pathogenicity island (PI) known as the locus of enterocyte effacement (LEE) (31, 32), and the majority of these genes are grouped into five operons (LEE1-5) (33). Hughes et al. also showed that KdpE regulates the transcription of ler, the first gene in the LEE1 operon and the master regulator of the LEE PI (4). In order to control LEE1 expression, KdpE acts in concert with Cra. Both transcription factors activate LEE1 expression under gluconeogenic conditions. These proteins bind to sites distant from one another and interact with each other to control LEE1 expression by promoting DNA bending (34). The LEE genes encode the structural components of a type three secretion system (T3SS), as well as some of its effectors (31, 35). LEE-encoded effectors, along with non-LEE-encoded effectors, like EspFu/TccP, mimic mammalian signaling proteins and hijack host cell signal transduction (36, 37). The non-LEE effector EspFu has been shown to be translocated through the T3SS into epithelial cells, where it contributes to the formation of AE lesions. The LEE PI and non-LEE effectors are encoded within blocks of sequences (O-islands) that are present in EHEC and are absent from the nonpathogenic Escherichia coli K-12 MG1655 genome. There are approximately 1,400 genes in O-islands (38), but fewer than 50% of them have assigned functions. However, because O-islands are most prevalent in pathogenic E. coli strains and some of the characterized genes within these islands have been linked to pathogenesis, it has been hypothesized that these islands are rich sources of virulence genes (39).

Previously, we described how KdpE and Cra interact to activate the LEE genes by directly binding to the ler promoter region (34). KdpE and Cra regulate the LEE PI in a glucose concentration-dependent manner. Here, we describe the transcriptional profiles of strains carrying nonpolar mutations of cra and kdpE. We show that Cra and KdpE share several virulence targets, including the genes encoding the effectors EspFu and EspG, as well as a number of genes within several O-island genes.

MATERIALS AND METHODS

Strains and plasmids.

All bacterial strains used in this study are listed in Table 1. Unless otherwise stated, strains were grown in Luria-Bertani (LB) medium or low-glucose Dulbecco's modified Eagle's medium (DMEM) at 37°C and 250 rpm. Media were supplemented, when necessary, with 50 μg ml⁻¹ streptomycin, 50 μg ml⁻¹ chloramphenicol, and 100 μg ml⁻¹ ampicillin. For protein expression, media were also supplemented with 0.5 mM isopropyl β-d-1-thiogalactopyranoside (IPTG). For growth curves, cultures grown overnight in LB were inoculated 1:100 (in triplicate) into either low-glucose (5.56 mM; 0.1%) DMEM or high-glucose (25 mM; 0.4%) DMEM, and the optical density at 600 nm (OD₆₀₀) was monitored over time.

Table 1.

E. coli strains and plasmids used in this study

Strain or plasmid	Genotype	Reference or source
Strain
8624	wt O157:H7 EDL933	64
DH11	ΔkdpE nonpolar mutant	4
JN01	Δcra	34
JN02	ΔkdpE Δcra	34
JN03	Δcra complemented with pCra in pBAD33	34
JN075	ΔkdpE complemented with pKdpE in pBAD33	34
Plasmid
pCG61	espFu regulatory region in pRS551	This study
pJN49	KpdE in pBAD33	34
pJN55	Cra in pET21	34
pJN57	Cra in pBAD33	This study
pKH4-28	KdpE in pET21	65
pRS551	lacZ reporter gene fusion vector	67
pET21	C-terminal His tag expression vector	EMBD Biosciences
pBAD33	Cloning vector	66

Construction of the strains used in this study has been previously described (Table 1) (4, 34). The transcriptional fusion reporter plasmid pCG50 was constructed by amplifying the espFu promoter region using primers EspFulacZF2 and EspFulacZR (Table 2) and cloning the PCR product into the BamHI and EcoRI/MfeI cloning sites of pRS551, a lacZ fusion vector (kanamycin and ampicillin resistant).

Table 2.

Oligonucleotides used in this study

Name	Sequence
EspFulacZF2	CATCAATTGCTGTCGGCTCTCTTCTAGAT
EspFulacZR	GTAGGATCCATATTGCGGTTGACGGTTGG
JespGrtF	ATGTCGAGGACTCGGCAATGCAAA
JespGrtR	TGCTATTTGCTCTGCATCATGGCG
Jz0639rtF	ATGAATGCGCTGACAACCGATGTG
Jz0639rtR	AACTGTTGGTGCGTTTGGGTTACG
Jz0640rtF	TGCCTCTGCCATGTCGCTGATTAT
Jz0640rtR	TTGCGTATACACCCACCCTTTCCA
Jz2077gsF	TGGGAGGGGAGAGAGTTAGAGTTTCTTATT
Jz2077gsR	GTTTTTTCTGTAATACAAGTCGATTGTTTGTGATTTCGC
Jz2077rtF	GCAACCTGGAACAGCAGATCAACA
Jz2077rtR	GGGCACTTAAGAAATTGTGTGTCGC
Jz3388gsF	TTGAATAATTCCCCTGATATTGCAAGGGCT
Jz3388gsR	GGCGCGTCTTACAAGGACGTTT
Jz3388rtF	CGGGGAACGCTTCAGCGATT
Jz3388rtR	CTAGATATTTTGTGTACTTGATTTGCAAACAGCTCCG
Jz39_40gsF	TCATTTTCTCTTGTTCAAAATAAGTCGTATTAATGTTTC
Jz39_40gsR	TTAAATTTTCCTGCCTGGCGTAAACC
Jz4267gsF	GCAAATCGTCCGGGGAAACCTTAC
Jz4267gsR	TGGAGTACTCCGAAACTCGGACG
Jz4267rtF	TGGTGAGCATCTTCATCTCTGCGT
Jz4267rtR	TCAAGGCTACCGATCACCAGTTCA
rpoA RTF	GCGCTCATCTTCTTCCGAAT
rpoA RTR	CGCGGTCGTGGTTATGTG
kan_EMSA_F1	CCGGAATTGCCAGCTGGGGCG
kan_EMSA_R1	TCTTGTTCAATCATGCGAAACGATCC

RNA extraction and qRT-PCR.

Cultures were grown in low-glucose DMEM to an OD₆₀₀ of 1.0. RNA from 3 biological replicates was extracted using a RiboPure bacterial isolation kit according to the manufacturer's protocols (Ambion). Quantitative reverse transcription-PCR (qRT-PCR) was performed as described previously (4). Briefly, diluted extracted RNA was mixed with validated primers (Table 2), RNase inhibitor, and reverse transcriptase (Applied Biosystems). The mix was used in a one-step reaction utilizing an ABI 7500 sequence detection system. Data were collected using ABI Sequence Detection 1.2 software, normalized to endogenous rpoA levels, and analyzed using the comparative critical threshold (C_T) method. Analyzed data were presented as fold changes over wild-type (wt) levels. The Student's unpaired t test was used to determine statistical significance. A P value of ≤0.05 was considered significant.

Microarray analysis.

Microarray global analysis was performed according to the manufacturer's instructions, as outlined in the Affymetrix gene expression technical manual (Affymetrix). Briefly, RNA extracted as described above was used as a template for reverse transcription to cDNA. The cDNA was then processed and hybridized to the E. coli genome GeneChip 2.0. The chips contain more than 10,000 probe sets directed toward the genes from four different strains of E. coli: the K-12 laboratory strain MG1655, the O157:H7 EHEC strain EDL933, the O157:H7 EHEC strain Sakai, and the uropathogenic strain CFT073.

To analyze the results, information from replicate scans of the chips was collected using GCOS v1.4 software according to the manufacturer's instructions. The data were then normalized using robust multiarray analysis (40, 41) and analyzed for differences in gene expression due to the deletion of cra and/or kdpE.

Protein purification and electrophoretic mobility shift assays (EMSAs).

The pET21-based plasmids encoding C-terminal His-tagged Cra and KdpE were transformed into the E. coli strain BL-21(DE3) (Invitrogen). Resulting transformants were grown to an OD₆₀₀ of 0.5 at 37°C in LB and then induced by adding IPTG to a final concentration of 0.5 mM and growing overnight at 25°C. The cells were harvested, suspended in lysis buffer (50 mM phosphate buffer, pH 8, 300 mM NaCl) with 20 mM imidazole, and lysed by homogenization. The lysed cells were centrifuged, and the lysate was loaded onto to a Ni²⁺-nitrilotriacetic acid-agarose gravity column (Qiagen). The column was washed with wash buffer (lysis buffer containing 50 mM imidazole), and protein was eluted with elution buffer (lysis buffer with 250 mM imidazole). Fractions containing purified protein (proteins were 95% pure) were confirmed by SDS-PAGE and Western blotting, while the protein concentration was quantified by Nanodrop and the Bio-Rad protein assay.

EMSAs were performed as previously described (42). Briefly, defined regions of the promoter (Table 2) were amplified by PCR, purified, quantified, and end labeled using radiolabeled [γ-³²P]ATP (PerkinElmer) and T4 polynucleotide kinase (NEB) according to the manufacturer's instructions. The radiolabeled probes were then repurified to remove unincorporated ATP. EMSAs were performed by adding the indicated amounts of purified recombinant protein to 2 ng labeled probe in binding buffer [60 nM HEPES, pH 7.5, 5 mM EDTA, 3 mM dithiothreitol (DTT), 300 mM KCl, 25 mM MgCl₂, 50 ng poly(dI · dC), 500 μg/ml bovine serum albumin (BSA); NEB] (3). When performing EMSAs using KdpE protein, 50 mM acetyl phosphate (lithium potassium salt; Sigma) was used as a phosphate donor when indicated. The reaction mixtures were incubated for 20 min at room temperature and then loaded on a 6% polyacrylamide gel after addition of a 5% Ficoll DNA loading buffer. The gel was run at 180 V for 6 h or 50 V overnight, dried, and exposed on a phosphorimager.

Beta-galactosidase assays.

The beta-galactosidase assays were performed as previously described (43). Briefly, appropriate strains containing the espFu-lacZ transcription fusion-expressing plasmid (pCG50; ampicillin and kanamycin resistant) were grown overnight aerobically at 37°C in LB. Dilutions (1:100) were grown in triplicate in clear DMEM (low glucose, 0.1 M salt, 0.001 M pyruvate, no phenol red) and appropriate antibiotics. Cells were diluted in Z buffer and lysed with chloroform and 0.1% SDS. After addition of o-nitrophenyl-β-d-galactopyranoside (ONPG), the reaction was timed and stopped using 1 M Na₂CO₃. The OD₄₂₀ was measured and used to calculate the Miller units as previously described (43). The Student's unpaired t test was used to determine statistical significance. A P value of ≤0.05 was considered significant.

Microarray data accession numbers.

Array data have been deposited in the NCBI GEO database under accession numbers GSE15050 and GSE43659.

RESULTS

Global gene regulation by Cra and KdpE.

We have previously shown that Cra and KdpE regulate virulence in EHEC. These two transcriptional regulators directly bind to the promoter region of ler, activating the transcription of this gene in a metabolite-dependent manner (34). The regulation of the LEE genes by Cra and KdpE, which we showed to interact in vitro, culminates in significant induction of AE lesion formation (34). Since Cra and KdpE regulate genes integral to EHEC pathogenesis, we investigated the extent of their control of the EHEC transcriptome. We extracted RNA from wt, Δcra, ΔkdpE, and ΔkdpE Δcra strains grown in low-glucose DMEM to late exponential phase. Using Affymetrix E. coli 2.0 microarray chips, we compared the expression profiles of the mutants to that of the wt. These arrays contain more than 10,000 probe sets that cover the two EHEC strains EDL933 and Sakai, the uropathogenic E. coli strain CFT073, the K-12 strain MG1655 genome, and intergenic regions that may carry small regulatory RNAs (sRNAs) or nonannotated small ORFs.

The microarray analysis revealed that for the cra mutant compared to the wt, expression of 829 probe sets was increased, with 43% of these being pathogen specific (Table 3). Additionally, a total of 515 probe sets had decreased expression, with 49% of these being pathogen specific (not present in K-12 strains of E. coli). When the kdpE mutant was compared to the wt, 658 probe sets were upregulated while 636 probe sets were downregulated, with the pathogen-specific ratios being 34 and 60%, respectively. As the deletion of both cra and kdpE results in almost complete ablation of AE lesion formation (34), we also investigated global gene expression in the ΔkdpE Δcra strain. Compared to the wt, the ΔkdpE Δcra strain had 997 probe sets upregulated and 305 probe sets downregulated. The total number of probe sets differentially regulated in all three mutants was similar (1,344 probe sets in the Δcra strain, 1,294 probe sets in the ΔkdpE strain, and 1,302 sets in the ΔkdpE Δcra strain). This raised the possibility that there are more genes, besides the LEE, that are regulated by both Cra and KdpE. To test this hypothesis, we searched for genes within the transcriptome that were regulated by both Cra and KdpE. When we compared the ΔkdpE and Δcra strain transcriptomes, we found 57 genes commonly upregulated and 63 genes commonly downregulated in these mutants (Fig. 1A and B). The downregulated genes included the LEE PI and several non-LEE-encoded effectors (Fig. 2), while the upregulated ones included genes encoding sRNAs, such as micF and omrA. The double mutant (ΔkdpE Δcra) strain had more downregulated genes in common with the Δcra strain (188) than with the ΔkdpE strain (25). Also, the ΔkdpE Δcra strain had significantly more genes that were upregulated in common with the Δcra strain (490) than with the ΔkdpE strain (64). These data suggest that the double mutant has a phenotype that more closely resembles that of the Δcra strain.

Table 3.

Comparison of the effect of deletion of cra, kdpE, or both genes on global gene expression of EHEC O157

Strain comparison and comparison focus	No. (%) of probe sets experiencing:
	Increased expression	Decreased expression	No change	Total
wt versus Δcra mutant
MG1655 specific	473	266	3,331	4,070
Pathogen specific	356 (43)	249 (49)	5,338	5,943
Total	829	515	8,669	10,013
wt versus ΔkdpE mutant
MG1655 specific	434	253	3,383	4,070
Pathogen specific	224 (34)	383 (60)	5,336	5,943
Total	658	636	8,719	10,013
wt versus ΔkdpEC Δcra mutant
MG1655 specific	527	109	3,434	4,070
Pathogen specific	470	196	5,277	5,943
Total	997	305	8,711	10,013

Fig 1.

Global analysis of effects of Cra and KdpE on EHEC O157 gene transcription. Venn diagrams show the number of overlapping downregulated genes (A) and upregulated genes (B) between the Δcra, ΔkdpE, and ΔkdpE Δcra mutant strains compared to the wt. (C) Venn diagram indicating genes that are decreased in the Δcra and increased in the ΔkdpE mutant strain. (D) Venn diagram indicating genes that are increased in Δcra and decreased in ΔkdpE strains. Strains for the microarrays were grown to an OD₆₀₀ of 1.0 in low-glucose DMEM.

Fig 2.

Effect of deletion of the transcriptional regulators Cra and KdpE on EHEC putative and characterized virulence genes. (A) A heat map representing differential regulation of the LEE genes. (B) A heat map showing the differential expression of non-LEE genes.

We next investigated oppositely regulated genes in the ΔkdpE and Δcra strain arrays. In the single mutant arrays, we identified a total of 290 probe sets conversely regulated (164 in the “decreased in the Δcra strain, increased in the ΔkdpE strain” batch and 126 in the “increased in the Δcra strain, decreased in the ΔkdpE strain” batch) (Fig. 1C and D). Altogether, these data indicated that there is both convergent regulation and converse regulation of some genes by Cra and KdpE.

Deletion of Cra and KdpE increases growth rate in a metabolite-independent manner.

Due to the fact that Cra and KdpE not only regulate virulence factors (22, 23, 34) but also are involved in carbon metabolism regulation (15–17, 44), we investigated whether these regulators affected EHEC growth rates. In particular, we wanted to examine differences during growth in low-glucose DMEM, conditions that have been shown to be optimal for LEE gene expression and AE lesion formation in vitro (45, 46). Strains were grown overnight in LB and then inoculated at a dilution of 1:100 in low-glucose DMEM. The Δcra and ΔkdpE mutant strains had significantly shorter doubling rates than the wt and ΔkdpE Δcra strains (P < 0.05 by two-way analysis of variance [ANOVA]) and reached a higher density, which they maintained for several hours (Fig. 3A). A possible explanation for this is that Cra and KdpE directly or indirectly regulate systems that delay exponential growth, such that when these two genes are deleted, the mutant strains grow faster. Another possibility is that the mutant strains have some genes induced that are not induced in the wt strain until after the strain is exposed to fresh media. Interestingly, when both genes are deleted, the growth rate of the ΔkdpE Δcra strain is comparable to that of the wt (no significance using two-way ANOVA), suggesting that in the double mutant the systems that control delayed growth nullify each other. The doubling rates of the ΔkdpE, Δcra, and ΔkdpE Δcra strains were 42, 37, and 41 min, respectively, compared to 48 min for the wt.

Fig 3.

Δcra and ΔkdpE strain growth curves. (A) Overnight samples of wt and mutant strains were diluted 1:100 in low-glucose (0.1%) DMEM, and their OD₆₀₀s were monitored over time. Their doubling rates were determined to be the following: wt, 48 min; Δcra mutant, 37 min; ΔkdpE mutant, 42 min; and ΔkdpE Δcra mutant, 41 min. The strain's ability to grow in high-glucose (0.4%) versus low-glucose medium was evaluated by monitoring the OD₆₀₀ for wt (B), ΔkdpE (C), Δcra (D), and ΔkdpE Δcra (E) strains. All strains were grown in triplicate.

Since ler transcription in high-glucose (0.4%) DMEM is downregulated compared to its transcription in low-glucose (0.1%) DMEM (34), we next examined the effect that increase in glucose had on growth. When the wt was grown in high glucose, its doubling time was comparable to that of growth in low glucose (Fig. 3B). However, as expected, the wt growing in high glucose had a longer log phase, which is probably due to the fact that the high-glucose media provides more nutrients, thus sustaining exponential growth for a longer period of time. When low/high-glucose growth curve experiments were performed for the mutants (Fig. 3C to E), a similar pattern of extended exponential growth was observed. The switch from low to high glucose did not significantly alter exponential growth in the wt or mutant strains (no significance using two-way ANOVA).

In silico analysis of characterized and putative virulence factors potentially regulated by Cra.

Cra has been reported to be an activator of virulence in both Salmonella (22) and EHEC (34). We previously reported on Cra's ability to activate Ler, the master regulator of the LEE genes, and consequently the transcription and translation of genes in this PI (34). Since the microarray analyses suggested that numerous virulence factors were regulated by Cra, and since Cra has a very well-defined DNA binding consensus sequence, we performed an in silico analysis of the EHEC genome using Virtual Footprint software, version 3.0 (http://prodoric.tu-bs.de/vfp/vfp_regulon.php) (47), to search for potential Cra binding sites. This software uses the consensus binding site sequence of transcriptional regulators to identify putative targets that may be directly bound by the transcription factor. Virtual Footprint allowed us to both narrow the list of potential direct Cra hits and to identify targets that the microarray may have missed. We used Cra's consensus binding sequence RSTGAAWCSNTHHW (48) to scan the genomes of two EHEC strains, EDL933 and Sakai, for potential targets in O-islands, regions of the EHEC genome not found in K-12 (38, 49). Since KdpE has no clear consensus binding sequence, we did not use this transcriptional factor in the bioinformatics analysis. We identified about a dozen potential O-island targets (Table 4), including espG, which is located next to ler and is divergently transcribed. Of the in silico predicted targets, only ler and espG have been characterized. EspG is a secreted effector that has been reported to disrupt host cell activity by perturbing the function of microtubules, Golgi membrane, and tight junction barriers (50–53). Several Virtual Footprint hits, including Z2077 and Z4267, were also differentially regulated in the microarrays. The fact that the hits reported by this software were so few, and that one of the positive hits was already characterized to be regulated by Cra, gave us confidence in this bioinformatics analysis method.

Table 4.

O-island genes predicted to be putative Cra targets by Virtual Footprint

Namea	Predicted Cra binding site	O-island	Function
Z0402	TGAATGGATTC	15	Putative beta-barrel outer membrane protein, 55% identity to putative ATP-binding component of a transport system and an adhesin protein in Escherichia coli, aidA-like
Z0639*	TGAAGCGGTTC	29	23% identity to putative adhesion/invasion gene in Neisseria meningitidis
Z0640*	TGAAGCGGTTC	29	Unknown
Z1163***	TGAATCGATC	43	36% identity to gene of unknown function in Sinorhizobium meliloti
Z1602***	TGAATCGATC	48	36% identity to gene of unknown function in Sinorhizobium meliloti
Z2077	TGAATGGATTA	57	Encoded by prophage CP-933O; nleG7′; secreted
Z3388	TGAATCGCTTAT	94	Unknown
Z3934	TGAATGGTTTAT	108	92% identity to NinG protein (bacteriophage 21)
Z4267	TGAAGCGTTTCA	119	32% identity to gene of unknown function in Methanobacterium thermoautotrophicum
espG**	TGAACCGTTTC	148	Binds p21-activated kinase, regulates endomembrane trafficking
Z5890	TGAATGGCTTA	172	74% identity to prophage P4 integrase in Escherichia coli
Consensus	TGAAT/G/CG/CGA/GT

^aAn asterisk indicates that it shares an intergenic region with ler; **, shares an intergenic region and Cra binding site with ler; ***, duplicate gene.

Cra and KdpE differentially regulate the transcription of characterized and putative virulence factors.

To confirm the in silico analyses, we performed qRT-PCR. Cra has previously been shown to interact at least in vitro with another transcriptional regulator KdpE, which together with Cra directly regulates ler (34). This previous observation suggested that Cra and KdpE could have more targets in common. We therefore investigated whether KdpE also regulated the Cra targets identified bioinformatically and by the microarray. We extracted RNA from wt, Δcra and ΔkdpE strains grown in low-glucose DMEM and analyzed the changes in transcription of the top five genes predicted to be regulated by Cra in the in silico analysis, namely, espG, Z0639, Z0640, Z3388, Z2077, and Z4267, that were also identified in the microarray.

Figure 4A shows a cartoon representation of the potential target genes. The espG gene shares a 1,200-bp regulatory region with ler. Previous work has shown that Cra binds with high affinity 350 bp upstream of ler (34), and this binding site is approximately 800 bp from espG. Scanning of the espG/ler promoter region closer to the espG transcription start site did not produce another potential Cra binding site closer than this previously described 800 bp. The confirmed binding site on the espG/ler promoter, as well as the putative sites identified by the Virtual Footprint software, are indicated as vertical solid arrows in Fig. 4. The KdpE binding site has also been identified approximately 100 bp from the ler translation start site and about 1,100 bp from the espG start site (34). When espG's transcription levels were examined in the mutants, the mRNA levels were decreased 2.5-fold in the Δcra strain compared to that of the wt, but in the ΔkdpE strain there was only a slight increase that was not significant (Fig. 4B). The lack of a significant effect on espG transcription in the ΔkdpE strain suggests that KdpE does not influence expression of this gene. Comparison of the mRNA levels of genes encoding the putative virulence factors in the Δcra strain to that in the wt depicted 2-fold reductions for Z0639, Z0640, and Z2077, a 3-fold reduction for Z4267, and a 4-fold increase in Z3388 transcription. The increase in Z3388 mRNA levels may be explained by the fact that Cra has been shown to be both an activator and repressor of transcription (14, 16). In the ΔkdpE strain, the transcription of Z0639, Z0640, Z3388, and Z4267 was decreased 2.5-, 3-, 4-, and 2-fold, respectively, but Z2077 mRNA levels remained unchanged (Fig. 4B). These results indicated that Cra and KdpE have several targets in common (ler, Z0639, Z0640, Z3388, and Z4267), as well as targets whose regulation is Cra dependent but KdpE independent (espG and Z2077).

Fig 4.

Confirmation of Cra and KdpE targets. (A) Cartoon representation of potential target genes identified using Virtual Footprint. The solid vertical arrows indicate the predicted Cra binding site, while the dotted vertical arrow indicates the KdpE binding site. The black horizontal arrows indicate genes evaluated via qRT-PCR, while the gray horizontal arrows indicate neighboring genes. (B) qRT-PCR analysis examining the expression of the indicated genes in wt, Δcra, and ΔkdpE strains grown to an OD₆₀₀ of 1.0 in low-glucose DMEM. The genes' transcript levels were quantified as fold differences normalized to wt gene transcription levels. The samples' rpoA transcript levels were used as internal controls to normalize the output C_T values. The data are from at least three independently grown replicates.

Transcriptional regulation of the O-island putative and characterized virulence genes by Cra and KdpE is direct.

Having confirmed that several genes identified by bioinformatics and/or by the microarrays are indeed differentially targeted by Cra, we next investigated whether their regulation by Cra, and for some, their regulation by KdpE, was direct. We designed 500-bp probes encompassing the proposed binding sites of Z0639/Z0640, Z3388, Z2077, and Z4267 and then performed electrophoretic mobility shift assays (EMSAs). These probes were radiolabeled, and approximately 150 pM of the probes was mixed with increasing amounts of either Cra or KdpE recombinant proteins. These mixtures were then run on a polyacrylamide gel (Fig. 5A and B). As a negative control, we used the radiolabeled nonspecific probe kan, which we previously showed to not be bound by either Cra or KdpE (34). We observed that increasing amounts of Cra were able to shift all of the probes except for the negative-control kan. When increasing concentrations of KdpE protein were incubated with the probes, all probes except the negative-control kan and the Z2077 probe were shifted. The fact that KdpE did not shift the Z2077 probe correlates with the qRT-PCR data (Fig. 5B), which showed that the mRNA levels of Z2077 remain unchanged in the ΔkdpE strain. These EMSA results provided evidence that the differential regulation of the putative virulence factors Z0639, Z0640, Z3388, and Z4267 by Cra and KdpE is due to direct binding. We also confirmed that although there was direct regulation of Z2077 by Cra, there was no binding of KdpE to the promoter region of this gene, which would explain the lack of regulation by this transcriptional factor. It is apparent from the qRT-PCR and the EMSAs that although Cra and KdpE share a number of targets, not all targets regulated by Cra are also controlled by KdpE.

Fig 5.

Cra and KdpE regulate O-island genes by direct interaction with their promoter regions. EMSAs of the regulatory regions of several genes within O-islands regulated by Cra and/or KdpE with these proteins are shown. Increasing amounts of His-purified recombinant Cra and KdpE was used to shift 2 ng of the indicated radiolabeled DNA probe. A radiolabeled kan DNA probe was used as a negative control.

Cra and KdpE directly regulate espFu expression.

We have shown that the two transcription regulators Cra and KdpE, by interacting with each other, can directly regulate transcription of ler, the master regulator of the LEE pathogenicity island (34). We have also shown that Cra and KdpE share O-island targets. The gene espFu codes for an effector, which has been shown to be important for pedestal formation, an important step in the progression of the disease caused by EHEC (37, 54). However, there are no probes for espFu in the arrays; hence, we scanned the espFu promoter region in silico and identified a putative Cra binding site (Fig. 6A). Interestingly, this binding site is downstream from the promoter, yet Cra activates espFu transcription. These data suggest that Cra works in vivo in context with other transcription factors; indeed, there is a precedence for a transcriptional activator to bind downstream of a LEE promoter that it activates (55). The target espFu was not one of the Virtual Footprint hits; this was probably due to the stricter parameters we set for the software. We transformed an espFu::lacZ transcriptional fusion into the single and double mutants and performed beta-galactosidase assays. We observed a drastic reduction in espFu transcription in all three mutants (Δcra, ΔkdpE, and ΔkdpE Δcra strains) (Fig. 6B), which could be complemented by expression of these genes in trans. We next wanted to investigate whether this regulation was direct. Using a probe encompassing the first 500 bp of the espFu promoter, we performed EMSAs with Cra and KdpE. Both proteins were able to shift the labeled espFu probe (Fig. 6C and D), indicating that espFu is indeed a direct target of both Cra and KdpE. It is interesting that unphosphorylated KdpE (Fig. 6D, left) showed higher binding affinity to the espFu promoter than KdpE phosphorylated with 50 mM lithium potassium acetyl phosphate (Fig. 6D, right). This observation was similar to what was observed with the ler promoter (34) and opposite of what was reported for the kdpFABC promoter region (44). The target-specific differential effect of phosphorylation on KdpE's ability to bind DNA indicates that KdpE has an added level of control that allows for optimal differential expression of virulence genes (LEE genes, espFu, and other O-island genes) and nonvirulence genes (kdpFABC).

Fig 6.

Cra and KdpE both directly regulate the espFu promoter region. (A) DNA sequence of the espFu promoter region showing the −35 and −10 positions and the transcription start site. The position of the putative Cra binding site is indicated in boldface. Below this is an alignment of the Cra binding site sequence from the espFu promoter with the consensus binding site sequence of Cra. (B) An espFu-lacZ transcriptional fusion was transformed into wt, ΔkdpE, Δcra, and ΔkdpE Δcra strains. The transcription fusion was also transformed into ΔkdpE and Δcra strains complemented with a low-copy chloramphenicol-resistant plasmid expressing KdpE and Cra, respectively. The beta-galactosidase assay was performed at an OD₆₀₀ of 0.5. (C and D) To investigate direct binding, EMSAs were performed using the espFu probe and increasing concentrations of recombinant Cra (C) or KdpE (D). For KdpE, EMSAs were done in the absence (D, left) or in the presence (D, right) of acetyl phosphate (lithium potassium salt).

DISCUSSION

Transcriptional regulators are involved in the control of gene expression by bacteria, allowing these microorganisms to efficiently and effectively adapt to their colonization niche. Pathogens, in particular, use transcriptional factors to express their virulence genes in an energy- and spatiotemporally efficient manner. KdpE, an OmpR/PhoB family member, positively regulates E. coli's kdpFABC operon in response to osmotic stress (56) and is phosphorylated by QseC to activate LEE expression (4, 34). KdpE has also been recently identified as an important virulence factor in a number of pathogens. It has been shown to be important in the intracellular survival of pathogens such as Yersinia pestis, Mycobacterium bovis, and Photorhabdus asymbiotica (57–59). We had previously shown that these two transcriptional factors, Cra and KdpE, directly bind to the promoter region of ler, the master regulator of the LEE genes, and consequently activate this island, leading to the formation of AE lesions (34). We had also presented evidence that these two transcription factors interacted (34). Here, we used microarrays and Virtual Footprint to identify more targets, including espFu, espG, and several O-island genes. We confirmed that a subset of these were not only real targets of Cra and/or KdpE but also that their regulation was due to the direct interaction of these transcription factors with their promoter regions. This regulation results in the activation of EHEC virulence, including the formation of the hallmark AE lesions (Fig. 7).

Fig 7.

Simple model for Cra and KdpE regulation of EHEC known and putative virulence factors. In response to metabolites, Cra and KdpE positively regulate the transcription of the ler, espFu, espG, and O-island genes and, consequently, EHEC virulence.

Our data (Fig. 1 and 4) suggest that although Cra and KdpE clearly have individual targets, they also share either convergent or divergent regulation of subsets of genes. These studies suggest that sets of transcription factors have intertwined regulons. Cra and KdpE converge in the regulation of the LEE T3SS (34) and several non-LEE-encoded effectors (Fig. 4 to 6), which agrees with the fact that expression of these genes should be coupled, even though they are found in different chromosomal regions.

One of the environmental signals that bacteria respond to is carbon nutrition. EHEC's ability to initiate growth and maintain colonization in vivo depends on whether the carbon source is glycolytic or gluconeogenic (60). Glucose polymers have been shown to be important sources of carbon nutrition (61–63). In vitro studies have shown that metabolites can regulate the expression of both metabolism- and non-metabolism-related genes. We previously showed that raising glucose concentrations in media from 0.1 to 0.4% inhibited not only the transcription but also the translation of the LEE genes (34). Here, we show that this difference in glucose levels also prolongs the log phase but does not significantly alter the overall growth of EHEC.

Altogether, we showed that Cra and KdpE are global regulators of gene expression and, in particular, the expression of virulence genes. These regulators respond to cues that include changes in glucose levels. As both of these transcription factors are found in many pathogens, understanding the Cra and KdpE regulatory cascade may provide useful information into virulence regulation in other pathogens.