The LysR-type regulator QseA regulates both characterized and putative virulence genes in enterohaemorrhagic Escherichia coli O157:H7

Authors

  • Melissa M. Kendall,

    1. 5323 Harry Hines Blvd., University of Texas Southwestern Medical Center at Dallas, Dallas, TX 75390-9048, USA.
    Search for more papers by this author
  • David A. Rasko,

    1. Institute for Genome Sciences, Department of Microbiology & Immunology, University of Maryland School of Medicine, BioPark Building II, 801 West Baltimore Street, Suite 619, Baltimore, MD 21201, USA.
    Search for more papers by this author
  • Vanessa Sperandio

    Corresponding author
    1. 5323 Harry Hines Blvd., University of Texas Southwestern Medical Center at Dallas, Dallas, TX 75390-9048, USA.
    Search for more papers by this author

E-mail vanessa.sperandio@utsouthwestern.edu; Tel. (+1) 214 648 1603; Fax (+1) 214 648 5905 (for express mail: 6000 Harry Hines Blvd. NA2.308, Dallas, TX 75235, USA).

Summary

Enterohaemorrhagic Escherichia coli (EHEC) colonizes the large intestine, causing attaching and effacing (AE) lesions. Most of the genes involved in AE lesion formation are encoded within a chromosomal pathogenicity island termed the locus of enterocyte effacement (LEE). The LysR-type transcriptional factor QseA regulates the LEE by binding to the regulatory region of ler. We performed transcriptome analyses comparing wild-type (WT) EHEC and the qseA mutant to elucidate QseA's role in gene regulation. During both growth phases, several genes carried in O-islands were activated by QseA, whereas genes involved in cell metabolism were repressed. During late-logarithmic growth, QseA activated expression of the LEE genes as well as non-LEE-encoded effector proteins. We also performed electrophoretic mobility shift assays, competition experiments and DNase I footprints. The results demonstrated that QseA directly binds both the ler proximal and distal promoters, its own promoter, as well as promoters of genes encoded in EHEC-specific O-islands. Additionally, we mapped the transcriptional start site of qseA, leading to the identification of two promoter sequences. Taken together, these results indicate that QseA acts as a global regulator in EHEC, co-ordinating expression of virulence genes.

Introduction

Enterohaemorrhagic Escherichia coli O157:H7 (EHEC) is a normal resident of cattle, but in humans EHEC causes severe disease characterized by bloody diarrhoea and haemolytic uraemic syndrome (Griffin et al., 1988). Infection commonly follows the ingestion of contaminated food or water (Riley et al., 1983; Honish, 1986; Ryan et al., 1986; Carter et al., 1987); however, infection may occur by direct contact with animals (Heuvelink et al., 2002) or via person-to-person transmission (Spika et al., 1986; Carter et al., 1987). EHEC survives passage through the acidic stomach and then enters the large intestine, where EHEC causes acute infection. There, EHEC senses the human hormones adrenaline and noradrenaline, as well as the bacterial hormone-like compound autoinducer 3 (AI-3) (Sperandio et al., 2003; Clarke et al., 2006). EHEC responds to these signals by activating expression of its flagella (Sperandio et al., 2002a) and of the locus of enterocyte effacement (LEE) (McDaniel et al., 1995). Upon reaching the epithelial cells, the LEE-encoded type three secretion system (TTSS) is expressed to translocate effector proteins directly into host epithelial cells (Jarvis et al., 1995). These effector proteins cause extensive cytoskeletal rearrangements and attaching and effacing (AE) lesion development (Jerse et al., 1990; Donnenberg and Kaper, 1991; Donnenberg et al., 1993; McKee et al., 1995; Kenny et al., 1997; Campellone et al., 2004; Deng et al., 2004; Garmendia et al., 2004; Mundy et al., 2004).

Most of the genes that encode for the TTSS and secreted proteins are located in the LEE (McDaniel et al., 1995). The LEE contains five operons (LEE1, LEE2, LEE3, LEE4 and LEE5) that are regulated directly or indirectly by more than 10 proteins (summarized in Fig. 1) (Friedberg et al., 1999; Mellies et al., 1999; Sperandio et al., 1999; 2000; Bustamante et al., 2001; Sánchez-SanMartín et al., 2001; Umanski et al., 2002; Haack et al., 2003; Deng et al., 2004; Iyoda and Watanabe, 2004; Zhang et al., 2004; Barba et al., 2005; Iyoda and Watanabe, 2005; Nakanishi et al., 2006). Regulation includes the global regulatory proteins integration host factor (IHF) that activates expression of ler (Friedberg et al., 1999) and H-NS that silences LEE transcription (Mellies et al., 1999; Sperandio et al., 2000; Bustamante et al., 2001; Sánchez-SanMartín et al., 2001; Umanski et al., 2002; Haack et al., 2003), as well as the environment-dependent regulator Hha that also acts to repress LEE transcription (Sharma and Zuerner, 2004). The sensor kinase QseC phosphorylates KdpE which subsequently activates expression of LEE1 (Hughes et al., 2009). The transcriptional regulators that are encoded by eivF and etrA in a second, non-functional TTSS are negative regulators of the LEE (Zhang et al., 2004). The ClpXP protease increases LEE transcription by inhibiting GrlR repression as well as by interacting with the stationary-phase sigma factor RpoS (Iyoda and Watanabe, 2005). RpoS positively regulates transcription of LEE3 (Sperandio et al., 1999; Iyoda and Watanabe, 2005), and the signalling molecule ppGpp can also increase transcription of the LEE (Nakanishi et al., 2006). The LEE-encoded regulator (Ler) that is encoded in LEE1 activates LEE2, LEE3 and LEE5 by binding to their promoters, displacing H-NS and allowing for transcription of these operons (Mellies et al., 1999; Sperandio et al., 2000; Sánchez-SanMartín et al., 2001; Haack et al., 2003; Russell et al., 2007). Also located in the LEE are grlR and grlA that encode for GrlR and GlrA respectively (Deng et al., 2004). GrlR represses LEE transcription, whereas GrlA activates LEE transcription (Deng et al., 2004; Barba et al., 2005; Russell et al., 2007). The RNA chaperone protein Hfq negatively controls levels of GrlA and GrlR through post-transcriptional regulation, and is thus an indirect regulator of LEE expression (Hansen and Kaper, 2009; Shakhnovich et al., 2009). The pch genes that are homologous to perC in enteropathogenic E. coli encode for positive regulators of the LEE genes (Iyoda and Watanabe, 2004).

Figure 1.

Model of LEE regulation. Adrenaline/noradrenaline and AI-3 are global regulators of virulence in EHEC. Black lines represent regulators whose direct interactions with the target promoter have been biochemically defined; grey lines represent interactions that occur indirectly or that have not been shown to bind biochemically to the target.

Additional regulation of the LEE also occurs via the LysR-type transcriptional regulator QseA. QseA was first identified as a regulator of the LEE genes as part of the cell-signalling cascade in the luxS system (Sperandio et al., 2002b). Initial studies demonstrated that an EHEC qseA mutant had a striking reduction in TTS activity, and secretion could be restored when qseA was complemented in trans (Sperandio et al., 2002b). Recent work has shown that QseA activates transcription of the LEE genes by directly binding to the ler promoter that encodes the master regulator of the LEE (Sharp and Sperandio, 2007). Moreover, QseA downregulates its own expression (Sircili et al., 2004) as well as expression of qseE that encodes a sensor kinase involved in sensing adrenaline (Reading et al., 2009) and in regulation of AE lesion formation (Reading et al., 2007). In this study, our objective was to explore the extent of QseA gene regulation in EHEC as well as to characterize the molecular mechanism involved in QseA regulation.

Results

QseA plays an extensive role in EHEC gene regulation

To elucidate whether QseA's role in regulating gene expression in EHEC was limited to ler and qseE, transcriptome analyses were undertaken, and these results were confirmed using real-time qPCR. We compared gene expression between wild-type (WT) EHEC strain 86-24 (Griffin et al., 1988) and the isogenic qseA mutant strain VS145 (Sperandio et al., 2002b) at both mid-exponential and late-exponential growth phases. The arrays contain approximately 10 000 probe sets for all 20 366 genes present in the genomes of the sequenced EHEC strains, EDL933 and Sakai, the uropathogenic E. coli strain CFT073, K-12 strain MG1655, and 700 probes covering intergenic regions that can encode non-annotated small ORFs, or small regulatory RNAs (sRNAs).

The microarray data [available at GSE18118 ( http://www.ncbi.nlm.nih.gov/geo/)] indicated that altered probe sets were both upregulated as well as downregulated indicating that QseA may act as an activator or repressor of its target genes (Table 1). Altered probe sets included genes derived from the E. coli K-12 strain MG1655, which contains a common E. coli backbone conserved among all E. coli pathovars (Rasko et al., 2008), as well as pathovar-specific genes (Table 2), suggesting a global role for QseA regulation.

Table 1.  Numbers of genes with altered expression in qseA mutant compared with WT.
 IncreaseDecreaseNo change
Mid-log22218747741
Late-log556058873
Table 2.  Comparison of K-12 and pathovar-specific genes with altered expression.
  IncreaseDecreaseNo change
MG1655-specificMid-log6033220
Late-log12153624
Pathovar-specificMid-log1121053409
Late-log341663907

Transcriptome comparisons between the qseA mutant and WT grown to mid-exponential phase revealed that QseA regulates genes involved in metabolism and virulence. We observed altered expression between the qseA mutant and WT for genes encoding for phosphonate metabolism and ethanolamine metabolism. Ethanolamine is abundant in the intestines (Sanatinia et al., 1995; Lawhon et al., 2003) and processed foods (Collier et al., 1991; Anderson, 1988) and may serve as a carbon, nitrogen and energy sources during aerobic or anaerobic growth (Blackwell et al., 1976; Kofoid et al., 1999; Buchrieser et al., 2003).

Microarray and qPCR data also revealed that qseA is most highly expressed at late-exponential growth (Fig. 2). Additionally, at late-exponential phase, the array data indicated that genes encoded in LEE1–5, including grlA, were significantly downregulated (Fig. 3A) in the qseA mutant strain compared with WT. We confirmed that QseA is a positive regulator of the LEE by performing RNA slot blot analysis and by qPCR of ler and escU (both encoded within LEE1), grlA and LEE1–5 (Fig. 3B and C, Table 3). Transcription of ler and escU could be restored to WT levels upon complementation with a functional copy of qseA (Fig. 3B and D). These data confirm that QseA promotes expression of the LEE by activating LEE1 and ler, thereby increasing expression of LEE2–5, as well as grlA.

Figure 2.

Transcriptional profile of qseA gene expression for WT EHEC grown to early-, mid-, late-exponential, and stationary growth phases as measured by qPCR and expressed as fold differences normalized to WT strain 86-24 at early-exponential growth phase. The error bars indicate the standard deviations of the ΔΔCT values (Anonymous, 1997). Significance is indicated as follows: one asterisk, P ≤ 0.05; two asterisks, P ≤ 0.005; and three asterisks, P ≤ 0.0005. The levels of rpoA transcript were used to normalize the CT values to account for variations in bacterial numbers.

Figure 3.

A. Heat map from microarray analysis representing LEE genes with altered expression in the qseA mutant compared with WT at mid-exponential and late-exponential growth phases.
B. RNA slot blot hybridized with the ler probe. RNA was purified from WT, the qseA mutant, and the qseA complemented strains grown in DMEM to late-exponential phase of growth. The graph corresponds to direct quantification of the concentration of target DNA from three slot blots. Error bars indicate standard deviations.
C. Transcriptional profile of LEE gene expression for WT EHEC and the isogenic qseA mutant as measured by qPCR and expressed as fold differences normalized to WT strain 86-24. The error bars indicate the standard deviations of the ΔΔCT values (Anonymous, 1997).
D. qRT-PCR measuring gene expression of escU (LEE1) of the WT, qseA mutant and qseA complemented strains. The error bars indicate the standard deviations of the ΔΔCT values (Anonymous, 1997). For the qRT-PCR analyses, the levels of rpoA transcript were used to normalize the CT values to account for variations in bacterial numbers.
In all graphs, significance is indicated as follows: one asterisk, P ≤ 0.05; two asterisks, P ≤ 0.005; and three asterisks, P ≤ 0.0005.

Table 3.  Relative change in expression in the qseA mutant strain VS 145 compared with WT strain 86-24.
GeneFold changeP-value
escU (LEE1)40.0047
grlA90.0005
escC (LEE 2)30.0005
escV (LEE 3)20.004
espA (LEE 4)20.0005
eaeA (LEE 5)20.0005
Z521240.0021
Z521350.0005
Z521460.0027
Z296783 0000.0003
Z29681 7940.0008
Z29691 2240.0008
Z297110 7110.0001
Z29721 6830.0001
Z297365 5660.0001
Z29743 5770.0001
Z29752 5920.0189
Z29761 3100.0087
Z29774870.0059
Z418130.0005
Z41852 5920.0189
Z4189220.0001
yqeI60.002
ygeK360.0006
Z032620.0009

Moreover, the microarray and real-time qPCR analyses indicated that QseA regulated numerous genes located in O-islands, regions of the EHEC genome not found in K-12 (Hayashi et al., 2001; Perna et al., 2001), during late-logarithmic growth. Specifically, the genes encoded in O-island 153 (Fig. 4A) were downregulated in qseA mutant compared with WT, and we confirmed these data by qPCR (Fig. 4B, Table 3). The genes in O-island 153 encode for the recently identified effector proteins EspX3′ (Z5212 and Z5213) and EspY5′ (Z5214) (Tobe et al., 2006). Reverse transcription PCR (RT-PCR) revealed that the Z5212 and Z5213 genes contained in O-island 153 are members of the same operon, and thus are expressed co-ordinately (Fig. 4C). Tobe et al. also reported that the prophage-like element Sp17 encodes secreted effector proteins (Tobe et al., 2006). Sp17 contains a cluster of genes from Ecs3512 to Ecs3508 that seem to comprise an operon (Abe et al., 2008). The microarray analyses revealed that these genes were downregulated in the qseA mutant compared with WT.

Figure 4.

A. Schematic representation of O-island 153.
B. Transcriptional profiles of Z5212, Z5213 and Z5214 gene expression for WT EHEC and the qseA mutant.
C. RT-PCR analysis showing that Z5212 and Z5213 are co-transcribed. Primers flanking Z5212 and Z5213 show product when either the genomic DNA control (lane 1) or cDNA is used (lane 2), but show no product when no RT is added (lane 3).
D. Schematic representation showing the arrangement of O-island 76.
E. Transcriptional profile of O-island 76 genes for WT EHEC and the qseA mutant.
F. RT-PCR showing that Z2966 through Z2969 are co-transcribed and that Z2970 (coxT) through Z2973 are co-transcribed. Controls were set up as described in (C).
G. Schematic representation showing the arrangement of O-island 115.
H. Transcriptional profiles of Z4181, Z4185 and Z4189 for WT EHEC and the qseA mutant.
I. Transcriptional profiles of yqeI, ygek and Z0326 for WT EHEC and the qseA mutant.
In (B), (E), (H) and (I), gene expression was measured by qPCR and expressed as fold differences normalized to WT strain 86-24. The error bars indicate the standard deviations of the ΔΔCT values (Anonymous, 1997). Significance is indicated as follows: one asterisk, P ≤ 0.05; two asterisks, P ≤ 0.005; and three asterisks, P ≤ 0.0005. The levels of rpoA transcript were used to normalize the CT values to account for variations in bacterial numbers.

Additionally, the microarray data indicated that QseA is involved in regulation of the genes located in the phage-encoded O-islands 76 (Fig. 4D) and 115 (Fig. 4G). The O-island 76 contains 12 genes that are downregulated in the qseA mutant strain compared with WT (Fig. 4E, Table 3). RT-PCR confirmed that the first four genes in the operon (Z2966–Z2969) were co-transcribed and that the divergent genes (Z2970–Z2973) were also co-transcribed (Fig. 4F). The O-island 115 contains 10 genes, and we analysed the effect of QseA on expression of three genes contained within this O-island. The real-time qPCR revealed that these genes are downregulated in the qseA mutant compared with WT (Fig. 4H, Table 3). Many of these genes have not been functionally characterized; however, they are similar to genes that encode secreted effector proteins PrgK, PrgH and SpaQ in Salmonella, YscS in Yersinia and Spa9 in Shigella (Venkatesan et al., 1992; Li et al., 1995; Callazo and Galán, 1996, Klein et al., 2000). We also performed qPCR on yqeI that encodes a predicted transcriptional regulator (Blattner et al., 1997), the gene ygeK that encodes for a response regulator protein (Yamamoto et al., 2005), as well as the prophage-encoded gene Z0326, and the data showed that these genes are downregulated in the qseA mutant compared with WT (Fig. 4I, Table 3). Together, these analyses further support that QseA is an important regulator of virulence gene expression in EHEC.

In addition to regulating genes that encode for proteins, the microarray data indicated that QseA regulates expression of small RNAs (sRNAs). More specifically, the sRNA downstream region sRNA (dsrA) was significantly decreased in the qseA mutant strain compared with WT (P ≤ 0.036) in RNA slot blot assays (Fig. 5). Moreover, the microarray data revealed that QseA regulates expression of a dicF and a dicF homologue – a sRNA involved in cell division (Faubladier et al., 1990; Bomchil et al., 2003) – as well as additional intergenic regions (i.e. sRNAs) that have not yet been characterized.

Figure 5.

RNA slot blot hybridized with the dsrA probe. RNA was purified from the WT and the qseA mutant strains grown in DMEM to late-exponential phase of growth. The graph corresponds to direct quantification of the concentration of target DNA from three slot blots. Error bars indicate standard deviations.

The qseA promoter contains two transcription start sites

To characterize the mechanism of QseA regulation, we first performed primer extension analysis to determine the qseA transcriptional start site. A primer was designed approximately 100 base pairs downstream from the qseA translational start codon. Then, primer extension analysis was performed using cDNA synthesized from RNA that was purified from WT EHEC, transformed with pMK40 and grown to late-exponential phase in DMEM.

The primer extension results showed that qseA had two transcriptional start sites (TSSs) (Fig. 6A). We mapped the P2 (proximal) promoter TSS to 30 base pairs upstream of the qseA translation start site. For P2, the −10 sequence, TATATT, had one mismatch from the σ70 consensus TATAAT, and the −35 sequence, TTAGCG, had three mismatches from the consensus TTGACA (Fig. 6B). The −10 and −35 promoter sequences were separated by a 19-base-pair spacer (Fig. 6B).

Figure 6.

A. The promoter of qseA was mapped using primer extension. The promoter region of qseA contains two promoters, a distal promoter P1 and a proximal promoter P2 responsive to QseA. The proximal (P2) and the distal (P1) promoters, sequences and conserved −10 and −35 regions are shown.
B. The +1, −10 and −35 promoter sites are indicated by bold typeface for the P1 and P2 promoters respectively.

The second promoter identified from primer extension was the P1 (distal) promoter located −125 upstream of the translational start site. In the P1 promoter, the distance between the −10 and −35 sequences was 18 base pairs (Fig. 6B). Four out of six base pairs matched the consensus sequence at the −10 promoter sequence, and three out of six matched with the consensus sequence at the −35 site (Fig. 6B).

QseA directly interacts with both the P1 and P2 ler promoters to activate transcription of ler

LysR-type transcriptional regulators may activate or repress transcription of its target genes by either direct or indirect action (Schell, 1993). Because microarray and real-time qPCR analyses cannot determine whether QseA regulation is direct or indirect, we performed electrophoretic mobility shift assays (EMSAs) with the promoter regions of several characterized and putative virulence genes identified in the transcriptome analyses. The EHEC LEE1 operon contains two promoters (Fig. 7A), the P1 distal promoter that is found in EPEC and Citrobacter rodentium and the P2 proximal promoter that is unique to EHEC (Mellies et al., 1999; Sperandio et al., 2002b; Deng et al., 2004; Sharp and Sperandio, 2007). Previously, it was shown that QseA binds to the P1 promoter to activate ler expression (Sharp and Sperandio, 2007), and our current work is consistent with those data (Figs 7B and 8B). Additionally, our study revealed that QseA bound to the P2 promoter (Figs 7B and 8C). To confirm specificity of binding, competition assays were performed (Fig. 7C). We showed that upon addition of unlabelled probe, QseA binding was competed out by the cold ler P2 promoter probe at a ratio of 1:4 (cold probe: labelled probe, volume/volume). The addition of unlabelled kan probe, as a negative control reaction, shows no competition, indicating that QseA specifically binds its target promoters.

Figure 7.

A. Schematic representation of the LEE1/ler promoter region, indicating the P1 and P2 promoter regions.
B. EMSA of the P1 promoter (−273 to −69 bp), the P2 promoter (−205 to −6) and the kan negative control promoter regions.
C. Competition assays with QseA. The assay was performed using 60 nM QseA in increasing amount of unlabelled ler promoter (−273 to −69 bp) probe. A competition assay was also performed using the kan promoter as a negative control.
D. Primer extension of the ler P2 promoter. RNA was harvested from WT EHEC and the qseA mutant strains grown to mid- and late-exponential growth phases. The qseA-dependent promoter is indicated by the arrow.
E. EMSAs of QseA with target promoters, including the −175 bp to +125 bp region of the qseA promoter, the −300 bp to +1 bp region (based on the translation start site) of the Z5212 promoter of O-153, the −300 bp to +1 bp region (based on the translation start site) of Z4189 of O-island 76, and the divergently transcribed Z2969 and coxT promoters from O-island 115 (−300 to +1 bp region based on the translation start site).

Figure 8.

DNase I footprints.
A. DNase I footprint was performed using the qseA promoter (−127 to +125 bp based on the proximal promoter start site) and increasing amounts of QseA.
B. DNase I footprint was performed using the ler distal promoter (−382 to −89 bp based on the proximal promoter).
C. DNase I footprint was performed using the proximal promoter (−205 to −6 bp) and increasing amounts of QseA.
D. DNA sequences of the qseA and ler promoters regions including the +1 transcriptional start sites and the −10 and −35 recognition sequences. Bold lettering indicates the QseA binding sites, and asterisks indicate the putative consensus sequences T-N11-A.

Previous work indicated that QseA activated ler through the P1 promoter (Sperandio et al., 2002b); however, QseA interaction with the P2 promoter was not examined. Thus, we further characterized QseA regulation of ler by performing primer extension analysis of the P2 promoter (Fig. 7D). RNA was isolated from WT EHEC and the qseA mutant strains grown to both mid-logarithmic and late-logarithmic growth phases. In this analysis, a weaker band mapped to the P2 transcriptional start site in the qseA mutant RNA isolated from late-exponential growth phase compared with the qseA mutant RNA isolated from mid-exponential growth phase as well as RNA isolated WT EHEC at mid- or late-exponential growth.

QseA regulates transcription of qseA and secreted effectors through direct interaction

LysR regulators generally bind to and repress their own expression (Schell, 1993). To test if QseA binds to its own promoter, we designed primers from −300 to +1 base pairs upstream of the qseA translational start site. Accordingly, QseA binds to its own promoter within 300 nucleotides of the qseA ATG translational start site (Fig. 7E).

QseA was able to bind and shift radiolabelled DNA probes comprising the Z5212 (O-island 153), Z4189 (O-island 115), coxT and Z2969 (O-island 76) promoter regions (Fig. 7E). Thus, these results suggest that QseA activates transcription of Z5212 (O-island 153), Z4189 (O-island 115), coxT and Z2969 (O-island 76) by directly binding to the first gene in the respective operon to increase expression of the downstream genes.

QseA binds upstream of the −35 promoter element when activating ler and binds directly to the qseA−35 promoter element when repressing its own transcription

In order to confirm the results of the EMSA analyses and potentially define a QseA-recognition consensus sequence, we undertook DNase I footprinting experiments of both ler promoters as well as the qseA promoter region. The footprinting analyses indicated that QseA directly bound its −35 P2 promoter element (Fig. 8A and D). QseA bound nucleotides −23 to −52 bp with respect to its P2 transcription start site. QseA bound both ler promoters immediately upstream of their respective −35 promoter elements (Fig. 8B–D). More specifically, QseA protected the promoter region of the ler P2 promoter located at −44 to −66 bp with respect to the P2 transcription start site. In the P1 promoter region, QseA protected the nucleotides located at −203 to −186 bp with respect to the P2 transcription start site.

The generally accepted LTTR (LysR-type transcriptional regulator) box consists of the sequence T-N11-A (Goethals et al., 1992; Schell, 1993; Parsek et al., 1994b; Deghmane and Taha, 2003; Maddocks and Oyston, 2008). We identified this sequence both in the ler promoters and in the qseA promoter sequences (Fig. 8D), and current studies using site-directed mutagenesis are being undertaken in order to confirm that these nucleotides are essential for QseA recognition of its target genes.

Discussion

QseA belongs to the family of LysR-type transcriptional regulators. LysR-type transcriptional regulators are the most abundant class of bacterial DNA-binding proteins (Schell, 1993). Early studies suggested that LysR-type transcriptional regulators were involved only in activation expression of a divergently transcribed gene and autorepressing their own transcription (Lindquist et al., 1989; Schell, 1993; Parsek et al., 1994a; Maddocks and Oyston, 2008). However, characterization of additional LysR-type regulators suggested that these proteins were global regulators, activating or repressing single or operonic genes that may be divergently transcribed or located throughout the chromosome (Deghmane et al., 2000; 2002; Heroven and Dersch, 2006; Hernández-Lucas et al., 2008; Maddocks and Oyston, 2008). Our study revealed that QseA regulates genes involved in metabolism, regulation, cell-to-cell signalling and virulence (summarized in Fig. 9), and thus supports the paradigm that LysR-type proteins act as global regulators. QseA regulation of both K-12 and pathogen-specific genes may allow for integration and maintenance of horizontally acquired virulence genes without disrupting central metabolic processes, as has been suggested for Pch (Abe et al., 2008).

Figure 9.

Summary of QseA gene regulation in EHEC at late-exponential growth phase. Lines with arrows indicate that QseA promotes expression, and lines with bars indicate that QseA downregulates expression of the respective genes encoding for the phenotypes listed.

Previous studies demonstrated that QseA regulates expression of ler, the master regulator of the LEE (Sperandio et al., 2002b; Sharp and Sperandio, 2007). This study has revealed that QseA regulation of ler affects expression of all the genes in the LEE pathogenicity island, including grlA, another positive regulator of ler (Deng et al., 2004) (Fig. 3A–D). We further characterized the interaction between QseA and ler using EMSA, primer extension and footprinting analyses. Our results confirmed QseA interaction with the P1 promoter (Sperandio et al., 2002b; Sharp and Sperandio, 2007) (Figs 7B and 8B). We also showed that QseA interacts directly with the P2 promoter (Figs 7B and D and 8C). LysR-type transcriptional regulators promote gene expression by typically binding to target sequences near the −35 RNA polymerase binding site (Schell, 1993). QseA is a positive activator of ler and binds to both the P1 and P2 promoter sites immediately upstream of their respective −35 promoter sequences (Figs 8B and C and 10A). QseA activation of ler at both the P2 and P1 promoter sites occurs mainly at late-exponential growth phase (Fig. 7D and Sperandio et al., 2002b). In EHEC, the LEE genes are most highly expressed at late-exponential phase (Walters and Sperandio, 2006), thus QseA plays important role in co-ordinating the timing of LEE gene expression.

Figure 10.

Model of QseA transcriptional regulation in EHEC.
A. When QseA act as a positive regulator of gene expression, QseA binds immediately upstream of the σ70−35 recognition sequence, as occurs on both the ler P2 and P1 promoters.
B. QseA represses its own transcription by binding directly to the σ70−35 promoter sequence, thereby occluding the RNA polymerase from binding to and transcribing the DNA.

EHEC encodes characterized and putative virulence factors outside the LEE, including fimbriae that may mediate attachment and effectors (Nles) that may enhance virulence (Hacker and Kaper, 1999; Kaper et al., 1999; Hacker and Carniel, 2001; Karmali et al., 2003; Deng et al., 2004; Garmendia et al., 2004; Shen et al., 2005; Tobe et al., 2006; Coburn et al., 2007; Coombes et al., 2008; ). Our study revealed that QseA is involved in promoting expression of genes encoding some Nles (Fig. 4A–I). QseA activates expression of Nles encoded for in O-islands 153, 76 and 115 by direct interaction with their promoter sequences (Fig. 7E). The O-island 153 encodes secreted proteins (Tobe et al., 2006), and the O-island 115 encodes putative virulence factors similar to the TTSS and secreted proteins of other enteric pathogens (Perna et al., 2001). The precise roles that these genes play in EHEC pathogenesis have not yet been elucidated.

Moreover, our study indicates that QseA regulates expression of sRNAs. sRNAs are increasingly being recognized as important in post-transcriptional gene regulation. QseA significantly downregulated expression of the sRNA dsrA during mid-logarithmic growth and upregulates this gene during late-exponential growth (Fig. 5). In non-pathogenic E. coli, DsrA was initially identified as a regulator of capsule synthesis, and subsequent studies revealed that DsrA binds with Hfq to regulate the translation of two global regulatory proteins, RpoS and H-NS (Sledjeski and Gottesman, 1995). In EHEC, H-NS acts to silence transcription of the LEE, thus if QseA promotes transcription of dsrA, H-NS repression of the LEE might be over-ridden. Indeed, a recent study showed that when overexpressed, DsrA promoted transcription of ler (and thus the LEE) (Laaberki et al., 2006).

Finally, we characterized the molecular mechanism of QseA autoregulation. LysR-type transcriptional regulators generally negatively regulate their own expression (Schell, 1993). A previous study showed that QseA represses its own transcription, and in the current, we undertook EMSA analyses and confirmed that QseA binds to its promoter region to autorepress transcription (Fig. 7E). Using primer extension analyses, we determined that the qseA promoter sequence contains two promoter sites (Fig. 6). We determined that QseA binds to the P2 promoter site directly overlapping the −35 promoter sequence (Fig. 8A and D). It is likely that QseA represses its own transcription by prohibiting RNA polymerase binding to the −35 promoter element (Fig. 10B).

LysR-type transcriptional regulators are the most common type of regulator found within the genomes of bacteria (Schell, 1993), thus better understanding of their contribution to virulence gene expression could lead to insights into disease progression, therapeutics and vaccine development (Maddocks and Oyston, 2008). This study has elucidated the extensive role that QseA plays in EHEC gene expression and contributes to the understanding of LysR-type transcriptional regulators' importance in bacterial pathogenesis.

Experimental procedures

Strains and growth media

All bacterial strains used in the study are listed in Table 4. Strains were grown at 37°C in Dulbecco's modified Eagle's medium (DMEM) (Invitrogen) or in LB (Invitrogen). Antibiotics were added at the following concentrations: ampicillin, 100 µg ml−1; streptomycin, 50 µg ml−1; and kanamycin, 50 µg ml−1.

Table 4.  Bacterial strains and plasmids used in this study.
Strain/plasmidGenotype/descriptionReference
Strain  
 86-24Wild-type EHEC strain (serotype O157:H7)Griffin et al. (1988)
 BL-21(DE3)F− ompT hsdSB(rB−, mB−) gal dcm (DE3)Invitrogen
 VS14586-24 qseA mutantSperandio et al. (2002b)
 VS151VS145 with plasmid pVS150 (qseA complement)Sperandio et al. (2002b)
 MK11EHEC strain 86-24 with plasmid pMK40This study
 MK14EHEC strain 86-24 with plasmid pVS23This study
 MK15VS145 with plasmid pVS23This study
Plasmid  
 pET28Cloning/expression vectorNovagen
 pMK08EHEC 86-24 qseA in pET28This study
 pVS23Regulatory region of LEE1 in pBluescriptSperandio et al. (2002b)
 pRS551lacZ reporter gene fusion vectorSimons et al. (1987)
 pVS150EHEC 86-24 qseA in pACYC177Sperandio et al. (2002b)
 pMK40QseA promoter sequence in pRS551This study

Recombinant DNA techniques

Standard methods were used to perform plasmid purification, PCR, ligation, restriction digests, transformations and gel electrophoresis. All oligonucleotide primers are listed in Table 5.

Table 5.  Primers used in this study.
Primers for qPCR and RNA slot blots
escU_RT_F1TCCACTTTGTATCTCGGAATGAAG
escU_RT_R1CAAGGATACTGATGGTAACCCTGAA
escC_RT_F1GCGTAAACTGGTCCGGTACGT
escC_RT_R1TGCGGGTAGAGCTTTAAAGGCAAT
escV_RT_F1TCGCCCCGTCCATTGA
escV_RT_R1CGCTCCCGAGTGCAAAA
grlA_RT_F1CCGGTTGTTCCAGGACTTTC
grlA_RT_R1TAAGCGCCTTGAGATTTTCATTT
espA_RT_F1TCAGAATCGCAGCCTGAAAA
espA_RT_R1CGAAGGATGAGGTGGTTAAGCT
eae_RT_F1GCTGGCCCTTGGTTTGATCA
eae_RT_R1GCGGAGATGACTTCAGCACTT
espX3′ (Z5212)_RT_F1AACCACGCAGTTCCCCATAA
espX3′ (Z5212)_RT_R1GTTAGACAATTTAGAAAAACGATTGAGATG
espX3′ (Z5213)_RT_F1GGGACAAATTTTAGCGGTTCTACA
espX3′ (Z5213)_RT_R1CGTCCACTTTGTTGGTGTCTTAAT
espY5′ (Z5214)_RT_F1CGTCGTACTAAAGCGCCATTT
espY5′ (Z5214)_RT_R1ACTGAGGACAAAGTTAAGAGATTTGAGA
ygeK_RT_F1TTTTCCATACGCATCCTTTCG
ygeK_RT_R1TTTGTAGCAAAGATGCCGTATATTG
yqe1_RT_F1GGCCTGAGCACCGAAAATTA
yqe1_RT_R1CTGGCTGGCGTTGTCAGA
Z0326_RT_F1GCAGGCTGTTCTGACACCAA
Z0326_RT_R1CCAAATGATGGCATATTCCTTTTT
Z2967_RT_F1CAAGACACTGTAATTGAACCTCATCA
Z2967_RT_R1CCGTCGCAATGCCTTTTAAA
Z2968_RT_F1CGGATTAATTCTGAAGCACGAA
Z2968_RT_R1TCGGCATCAAGCCACATAGA
Z2969_RT_F1CCTTTTCACTGATCGCACCTTT
Z2969_RT_R1TCGAGCAGCTTCGTTAGATGTG
Z2971_RT_F1GCCCACGCAATCCGTTT
Z2971_RT_R1GATTGTCGCCGCAGTTCAA
Z2972_RT_F1TTGATTTTCGCGGTGAACTG
Z2972_RT_R1TTTTTTCGGATGGAACAATTCC
Z2973_RT_F1GTGCAGCCCGTGACTCATT
Z2973_RT_R1TCAGCGTAAAAAGCAACGTCAT
Z2974_RT_F1CAGGGCAGCCCAGAAAGATA
Z2974_RT_R1GTGGCTATTGCGATGGTCAGT
Z2975_RT_F1CCTGGCGAACCGGAAAC
Z2975_RT_R1AACGTTATCGATGCGTTGCA
Z2976_RT_F1CGTCGTCACATCCAGACTGAA
Z2976_RT_R1TGCGCGACGCGGAT
Z2977_RT_F1CCTCTGCAATCGTCGATGTG
Z2977_RT_R1TCGGCATGGCGATGTGT
Z4181_RT_F1GAGCGGGTCAACGTCTCAA
Z4181_RT_R1GAGGTTTCATTGGTGGGAACA
Z4185_RT_F1TGGCAATCATCAAACGGCTAT
Z4185_RT_R1GAGAGCGCCATCAGGAAATTC
Z4189_RT_F1ACCAACTCCGGTTCAGAATATTTT
Z4189_RT_R1CGTCGGTTGTTAATTTTGTTGAGA
qseA _RT_F1CGCAAGGGAGATTTGTGACTAAT
qseA_RT_R1GGCACCCGCCGTTAGC
rpoA_RT_F1GCGCTCATCTTCTTCCGAAT
rpoA_RT_R1CGCGGTCGTGGTTATCTG
ler_dotblot_F1GAGATTATTTATTATGAATA
ler_dotblot_R1TCTTCTTCAGTGTCCTTCAC
dsrA_F2GCACAATAAAAAAATCCCGA
dsrA_R2ATTGCGGATAAGGTGATGAAC
Primers for RT-PCR
Z2966-69_F1TTAATTGCCCATCAGGAAGC
Z2966-69_R1TTCCTGATGCTGCGTATTTG
Z2970-73_F1TCAATGTGAATGCCCCCTAT
Z2970-73_ R1ATGCGATATTCCCCAGCTTA
Z5212-Z5213_F1GGCACACCACCAAACCACGC
Z5212-Z5213_R1GCTCTTGCCTTACACTATA
Primers for EMSAs/footprinting
QseA_pETF1CATATGGAACGACTAAAACGCATGTCGGTG
QseA_pET28R1AAGCTTTTACTTCTCTTTCCCGCGCCCGTG
LerProximal_F1
Ler promoter −173 FCGGGATCCCGATGATTTTCTTCTATATCATTG
LerProximal_R1
Ler promoter −42 RCGGAATTCCGCGACCTTATCAGGAAGGACC
LerDistal_F2CGGAATTCCTGGGGATTCACTCGCTTGC
LerDistal_R1TCTAATGTGTAAAATACAT
QseA_EMSA_F1TAGATACCTGTTGGCACAAGTACCCGGCACAC
QseA_EMSA_R1CATTATTCACTCTGACTTAAAAGTGATTTAGA
EspX3′ (Z5212)_EMSA_F1TCCTCCGGCGTGTAATATTG
EspX3′ (Z5212)_EMSA_R1TATCCATAAATCACCTTGAT
Z4189_EMSA_F1GCATCATTACAAACATTGAC
Z4189_EMSA_R1TGAAACAGATGACTTTGATA
Z2969_EMSA_F1GTCGCTCATGGCAACTTGAC
Z2969_EMSA_R1AGCTGTTAGCCAAAGCGTCT
coxT_EMSA_F1CAATCTTTCCCTTTCCTCAC
coxT_EMSA_R1TGATGGTTATACCTGACATC
kan_EMSA_F1CCGGAATTGCCAGCTGGGGCG
kan_EMSA_R1TCTTGTTCAATCATGCGAAACGATCC
amp_EMSA_F1GGAATTCGAAAGGGCCTCGTGATACGC
amp_EMSA_R1CGGGATCCGGTGAGCAAAAACAGGAAGG
Primers for primer extension
qseA_PE_F1CACGATAACGGGAAACAGACTCATGTTGACCT
qseA_PE_R1TGACACCGTCTGACTGATGGACGA
K983GGACTTGTTGTATGTG

Plasmid pMK08 was constructed by amplifying the qseA gene from the EHEC strain 86-24 using AccuTaq polymerase (Sigma) using primers QseA_pETF1 and QseA_pET28R1 and cloning the resulting PCR product into the NdeI/HindIII cloning site of vector pET28.

Plasmid pMK40 was constructed by amplifying −200 bp to +100 bp in relation to the qseA translational start site using primers QseA_PE_F1 and QseA_PE_R1. The resulting product was cloned into the BamHI/EcoRI cloning site of pRS551. EHEC strain 86-24 was then transformed with pMK40 to create strain MK11.

RNA extraction

Cultures of strains 86-24 and VS145 were grown aerobically in LB medium at 37°C overnight, and then were diluted 1:100 in DMEM and grown at 37°C. RNA from three biological replicate cultures of each strain/condition was extracted at the early-exponential (OD600 of 0.2), mid-exponential growth phase (OD600 of 0.5), late-exponential growth phase (OD600 of 1.0) or stationary growth phase (OD600 of 1.5) using the RiboPure Bacteria RNA isolation kit (Ambion).

Reverse transcription PCR

The SuperScriptTM First-Strand Synthesis System for RT-PCR was used to synthesize cDNA. The resulting cDNA was utilized for regular PCR with gene-specific primers that originated in Z2966 and Z2969, Z2970 and Z2973, Z5212 and Z5213, and Z5212 and Z5214 (listed in Table 5). A positive control with genomic DNA and a negative control of RNA without reverse-transcriptase-added were also used.

Microarray preparation and analyses

Affymetrix 2.0 E. coli gene arrays were used to compare gene expression in strain 86-24 with that in strain VS145 (qseA mutant) at both mid- and late-exponential growth phases grown aerobically in DMEM at 37°C. The GeneChip E. coli genome 2.0 array includes approximately 10 000 probe sets for all 20 366 genes present in the following four strains of E. coli: K-12 lab strain MG1655, uropathogenic strain CFT073, O157:H7 enterohaemorrhagic strain EDL933 and O157:H7 enterohaemorrhagic strain Sakai ( http://www.affymetrix.com). The RNA processing, labelling, hybridization and slide-scanning procedures were performed as described in the Affymetrix Gene Expression Technical Manual (http://www.affymetrix.com/support/technical/manual/expressionmanual.affx).

The array data analyses were performed as described previously (Kendall et al., 2007). The output from scanning a single replicate of the Affymetrix GeneChip E. coli Genome 2.0 array for each of the biological conditions was obtained using GCOS v 1.4 according to the manufacturer's instructions. Data were normalized using Robust Multiarray analysis (Irizarry et al., 2003) at the RMAExpress website (http://rmaexpress.bmbolstad.com/); the resulting data were compared with determined features whose expression was increased or decreased in response to the qseA deletion. Custom analysis scripts were written in Perl to complete multiple array analyses. The results of the array analyses were further confirmed using qPCR as described below. We note that the isolate used in these studies has not been sequenced and thus is not fully contained on the array and that differences in genome content are evident. Expression data can be accessed using Accession No. GSE18118 at the NCBI GEO database.

Real-time qPCR

RNA was extracted as described above from three biological replicates each of strain 86-24 and strain VS145. The primers used in the real-time qPCR assays were designed using Primer Express v1.5 (Applied Biosystems) (Table 5). Amplification efficiency and template specificity of each of the primer pairs were validated and reaction mixtures were prepared as described previously (Walters and Sperandio, 2006). Real-time RT-PCR was performed in a one-step reaction using an ABI 7500 sequence detection system (Applied Biosystems).

Data were collected using the ABI Sequence Detection 1.2 software (Applied Biosystems). All data were normalized to levels of rpoA and analysed using the comparative critical threshold (CT) method (Anonymous, 1997). The expression levels of the target genes under the various conditions were compared using the relative quantification method (Anonymous, 1997). Real-time data are expressed as the changes in expression levels compared with the WT levels. Statistical significance was determined by Student's t-test, and a P-value of ≤ 0.05 was considered significant.

RNA slot blot

RNA was isolated as described above from strains MK14 and MK15 grown at 37°C to mid- (OD600, 0.5) and late-exponential (OD600, 1.0) phases in DMEM. RNA slot blotting was performed with 2 µg of total RNA in triplicate. The RNA was dissolved in 10 µl of RNase-free water and then denatured 30 µl of RNA denaturation solution [660 µl of formamide, 210 µl of 37% (w/v) formaldehyde, and 130 µl of 10× MOPS (morpholinepropanesulfonic acid, pH 6.8)] at 65°C for 5 min. Then, the RNA was applied to a nylon membrane under vacuum by using a Bio-Rad dot blot apparatus. The well was then washed two times with 1 ml of 10× SSC (1× SSC is 0.15 M NaCl and 0.015 M sodium citrate). The membranes were cross-linked, hybridized with either the ler or dsrA probe by using UltraHyb from Ambion at 42°C, then washed first with 1× SSC, 0.1% SDS (sodium dodecyl sulfate) and then three times with 0.5× SSC, 0.1% SDS. Bound probes were detected using the Amersham PhosphorImager. Liquid scintillation counting was used to quantify the concentration of target DNA.

DNA probes were generated with ler_dotblot_F1 and ler_dotblot_R1 or dsrA_F2 and dsrA_R2 primer sets and radio-labelled using the Ready-To-Go DNA Labelling Beads (GE Healthcare).

Purification of QseA in native conditions

In order to purify the His-tagged QseA protein, the E. coli strain BL-21 (DE3) (Invitrogen) containing pMK08 was grown at 37°C in LB to an OD600 of 0.5, at which point IPTG was added to a final concentration of 0.4 mM and allowed to induce for 3 h. Purification was then performed using nickel columns (Qiagen).

Electrophoretic mobility shift assays (EMSAs)

To determine the direct binding of QseA to target promoters, EMSAs were performed using the purified QseA-His and PCR amplified DNA probes (Table 5). DNA probes were then end-labelled with [γ-32P]-ATP (Perkin-Elmer) using T4 polynucleotide kinase (NEB) using standard procedures (Sambrook et al., 1989). End-labelled fragments were run on a 6% polyacrylamide gel, excised and purified using the Qiagen PCR purification kit.

EMSAs were performed by adding increasing amounts of purified QseA protein (0–15 µg) to end-labelled probe (10 ng) in binding buffer [500 µg ml−1 BSA (NEB), 50 ng of poly-dIdC, 60 mM HEPES pH 7.5, 5 mM EDTA, 3 mM dithiothreitol (DTT), 300 mM KCl and 25 mM MgCl2] (Clarke and Sperandio, 2005) and incubated for 20 min at room temperature. A 5% ficol solution was added to the reactions immediately before loading the samples on the gel. The reactions were electrophoresed for approximately 8 h at 150 V on a 6% polyacrylamide gel, dried and exposed to KODAK X-OMAT film.

DNase I footprinting

DNase I footprints were performed as described previously (Sperandio et al., 2000). Briefly primers LerDistal R2, Ler −42R (Sharp and Sperandio, 2007), and QseA R1 (Table 5) were end labelled by standard procedures using [γ-32P]-ATP (Perkin Elmer). The resulting labelled primers were used in a PCR reaction with unlabelled primers Ler F4, Ler 173F (Sharp and Sperandio, 2007) or QseA F3 respectively. The resulting single-end-labelled PCR products were utilized in binding reactions (described in the EMSAs) with purified QseA for 20 min at 25°C. At this time, a 1:100 dilution of DNase I (Invitrogen) and the manufacturer-supplied buffer were added and allowed to digest at room temperature for 4 min. The digestion was stopped by adding 100 µl of stop solution (200 mM NaCl, 2 mM EDTA and 1% SDS). All protein was then extracted via phenol-chloroform, and the DNA was precipitated using 3 M potassium acetate, 100% ethanol and 1 µl of glycogen. The DNase reactions were run in a 6% polyacrylamide gel next to a sequencing reaction (Epicentre). Amplified DNA from strain 86-24 was used to generate the sequencing ladder using primers LerProximal_F1 and LerProximal_R1, LerDistal_F2 and LerDistal_R1, and QseA_EMSA_F1 and QseA_EMSA_R1.

Primer extension

Primer extension analysis was performed as described previously (Mellies et al., 1999). Briefly, primer QseA_PE_R1 (Table 5), located 100 bp upstream of the ATG translational start site, was end-labelled using [γ-32P]-ATP. A total of 40 µg of RNA was isolated as described above from strain MK11 and was used with the Primer Extension System – AMV Reverse Transcriptase kit (Promega). The resultant cDNA was precipitated, run on a 6% polyacrylamide-urea gel and visualized by autoradiography. A sequencing ladder (Epicentre) was run adjacent to the primer extension reaction. The sequencing ladder was generated from EHEC 86-24 genomic DNA amplified with primers QseA_PE_F1 and QseA_PE_R1.

Primer extension of the ler P2 promoter was performed as described above using RNA isolated from strains MK14 and MK15 grown to mid- and late-exponential phases. The sequencing ladder was generated from EHEC 86-24 genomic DNA amplified with primers ler distal F2 and K983.

Acknowledgements

This work was supported by National Institutes of Health Grant AI053067, the Ellison Medical Foundation, the Burroughs Wellcome Fund, and a NIH Ruth L. Kirschstein Fellowship F32AI80115 to M.M.K. Its contents are solely the responsibility of the authors and do not represent the official views of the NIH NIAID.

Ancillary