Hfq affects the expression of the LEE pathogenicity island in enterohaemorrhagic Escherichia coli

Abstract

Colonization of the intestinal epithelium by enterohaemorrhagic Escherichia coli (EHEC) is characterized by an attaching and effacing (A/E) histopathology. The locus of enterocyte effacement (LEE) pathogenicity island encodes many genes required for the A/E phenotype including the global regulator of EHEC virulence gene expression, Ler. The LEE is subject to a complex regulatory network primarily targeting ler transcription. The RNA chaperone Hfq, implicated in post-transcriptional regulation, is an important virulence factor in many bacterial pathogens. Although post-transcriptional regulation of EHEC virulence genes is known to occur, a regulatory role of Hfq in EHEC virulence gene expression has yet to be defined. Here, we show that an hfq mutant expresses increased levels of LEE-encoded proteins prematurely, leading to earlier A/E lesion formation relative to wild type. Hfq indirectly affects LEE expression in exponential phase independent of Ler by negatively controlling levels of the regulators GrlA and GrlR through post-transcriptional regulation of the grlRA messenger. Moreover, Hfq negatively affects LEE expression in stationary phase independent of GrlA and GrlR. Altogether, Hfq plays an important role in co-ordinating the temporal expression of the LEE by controlling grlRA expression at the post-transcriptional level.

Figure 1

Hfq affects the abundance of proteins in culture supernatants and LEE expression in EHEC. (A) Proteins present in culture supernatants of EHEC TUV93-0 wild type, hfq mutant and the mutant strain complemented with hfq from pAMH100 grown in DMEM at 37°C to an optical density at 600 nm (OD₆₀₀) ~ 0.2 (lanes 1–3) and 1.0 (lanes 4–6). Proteins corresponding to 300 µl of culture supernatant were prepared as described in Experimental procedures, resolved in a 12% SDS-PAGE and visualized by silver staining. (B and C) The temporal expression profiles of LEE-encoded proteins in wild type TUV93-0, a hfq mutant and a complemented hfq mutant were determined by Western blot analyses as described in Experimental procedures. Equal amounts of total protein including both whole cell and secreted proteins, was prepared from cultures grown at 37°C in DMEM (B) and LB (C) to the cell densities indicated, resolved by 4–20% SDS-PAGE and blotted onto nylon membranes. Levels of Tir, EspA, EspB and GroEL were detected by Western blot analyses using polyclonal antisera against the respective proteins. GroEL served as an internal control for the total protein amount.

Figure 2

LEE transcript levels are increased in a hfq mutant. Quantitative real-time PCR measurements of LEE transcripts in total RNA isolated from TUV93-0 wild type and hfq mutant strains grown at 37°C in LB to OD₆₀₀ ~ 0.6 and 3.0 were performed as described in Experimental procedures. The relative fold expression representing the change in transcript levels of ler, sepZ, escV, espB and tir in a hfq mutant relative to wild type grown to OD₆₀₀ ~ 0.6 (A) and 3.0 (B) was determined. Detection of the rpoB transcript served as a control for a gene which expression is not significantly stimulated in a hfq mutant. Results represent means and standard deviations from triplicate experiments. Error bars show the standard deviations of ΔΔC_T values. Levels of 16S rRNA were used to normalize the C_T values of target genes to variations in bacterial numbers.

Figure 3

Hfq indirectly affects Ler levels. (A) Schematic outline of constructs used to determine ler-his expression from ler distal and proximal promoters (pAMH103), a ler promoter with 120 nt of the 160 nt leader sequence including the proximal promoter region deleted (pAMH105) and kan-his expressed from a native ler promoter (pAMH113). P_{ler d} and P_{ler p} designate the distal and proximal ler promoter, respectively. (B) The ler leader region is not required for Hfq-mediated regulation of Ler levels. Western blot analysis of Ler-His expressed from pAMH103 (lanes 1–3) and pAMH105 (lanes 4–6) in TUV93-0 ler, TUV93-0 ler hfq and EDL933 ΔLEE strains grown in LB at 37°C to OD₆₀₀ ~ 0.4. (C) Expression of Kan-His from the ler promoter is negatively affected by hfq. Western blot analysis of Kan-His expressed from pAMH113 in TUV93-0 ler and ler hfq strains grown in LB at 37°C to OD600 ~ 0.4 (lanes 1–2) and 3.0 (lanes 3–4). Ler-His, Kan-His and GroEL were detected using antibodies against the His-tag and GroEL, respectively. GroEL served as an internal loading control.

Figure 4

Hfq affects LEE expression in EPEC E2348/69. Wild type and hfq mutant strains grown in LB at 37°C to OD₆₀₀ ~ 0.6 (lanes 1–2) and 3.0 (lanes 3–4) were assayed for Ler-His expressed from pAMH104 (A) and EspB (B) by Western blot analyses using His and EspB specific antibodies, respectively. (C) Hfq affects LEE expression independent of per. Western blot analyses to measure EspB levels in E2348/69, E2348/69 hfq, JPN15 (per) and JPN15 hfq strains, grown in LB at 37°C to OD₆₀₀ ~ 0.6 (lanes 1–4) and OD₆₀₀ ~ 3.0 (lanes 5–8), were performed using antibodies specific to EspB and GroEL.

Figure 5

Hfq negatively regulates levels of LEE-encoded proteins through grlRA. The effect of grlRA on LEE expression was measured by Western blot analyses of TUV93-0 wild type, a hfq mutant, a grlRA mutant, a hfq grlRA double mutant and a hfq grlRA double mutant strain complemented with grlRA (pAMH116) and hfq (pAMH100) grown in LB to exponential (OD₆₀₀ ~ 0.6, lanes 1–6) and stationary phase (OD₆₀₀ ~ 3.0, lanes 7–12). Levels of Tir, EspA, EspB and GroEL were detected using polyclonal antibodies against the respective proteins. GroEL served as an internal control for total protein levels.

Figure 6

Hfq affects grlRA expression independent of Ler. The expression of GrlA (A and C) and GrlR (B and D) was determined in TUV93-0 wild type and hfq backgrounds in the presence (A and B) and in the absence (C and D) of ler by Western analyses using a His specific antibody. TUV93-0 wild type, hfq and ler mutant strains, harboring plasmids encoding grlRA-his (pAMH117) and grlR-his (pAMH118), were grown in LB at 37°C to exponential (A–D lanes 1–2) and stationary phases (A–D lanes 3–4). His-tagged GrlR and GrlA proteins were detected by Western blot analyses using a His specific antibody. Detection of GroEL served as an internal control. (E) Hfq affects grlA mRNA stability. Wild type TUV93-0 and hfq mutant strains were grown in LB at 37°C to exponential phase (OD₆₀₀ ~ 0.6), transcription was blocked by rifampicin (0 min) and aliquots were collected at the time indicated, and total RNA was extracted as described in Experimental procedures. Analysis of grlA mRNA stability was carried out using qRT-PCR to measure the amount of grlA transcript normalized to levels of 16S rRNA. Data represent the mean ± SD percentage of grlA transcript remaining in three independent RNA samples isolated from wild type (circles) and the hfq mutant (triangles) at the time indicated after rifampicin addition relative to time 0. The percentage of transcript remaining for each time point was significantly different in RNA isolated from wild type and hfq mutant strains as determined by the t-test (P < 0.01). Half-lives of grlA mRNA in wild type and hfq mutant strains were calculated from the lin-log graph to 1.5 min and 3 min, respectively.

Figure 7

The Hfq-mediated regulation of grlRA requires a factor that is located outside the LEE. Western blot analyses were used to detect levels of GrlA (A and C) and GrlR (B and D) in ΔLEE and ΔLEE hfq mutant derivates of EHEC EDL933 (A and B) and wild type and hfq mutant backgrounds of E. coli K-12 MG1655 (C and D) grown in LB at 37°C to OD₆₀₀ ~ 0.6 (lanes 1–2) and 3.0 (lanes 3–4). His-tagged GrlR and GrlA were expressed from pAMH117/pAMH118 and pAMH125/pAMH126 in K-12 and EHEC backgrounds, respectively. His-tagged proteins were detected using a His specific antibody as described in Experimental procedures. GroEL served as an internal control for total protein levels.

Figure 8

Hfq affects the A/E phenotype through grlRA. FAS test using HeLa cells infected for 3 h (A–F) and 6 h (G–L) with TUV93-0, TUV93-0 hfq, TUV93-0 hfq pAMH100, TUV93-0 grlRA, TUV93-0 grlRA hfq and TUV93-0 grlRA hfq pAMH116. The actin cytoskeleton of infected HeLa cells was stained with FITC-phalloidin. Representative phase-contrast (left panels) and fluorescence actin (right panels) images are shown. Black and white arrowheads indicate bacteria and A/E lesions, respectively.

Figure 9

Model depicting the regulation of LEE by Hfq, GrlA, GrlR and Ler. Black arrows and T-lines represent positive and negative regulation, respectively. Gray solid and broken T-lines indicate Hfq-mediated regulation in exponential and stationary phase, respectively. Ler is a positive regulator of LEE2-5 and grlRA expression, whereas GrlA and GrlR activate and repress LEE1 transcription, respectively, as described in the introduction. Moreover, GrlA and GrlR positively affect transcription of LEE2 and LEE4 independent of Ler (Russell et al., 2007). Hfq negatively affects grlRA expression by modulating transcript stability to prevent expression of GrlA and GrlR, and thereby the initiation of the Ler-GrlA positive regulatory loop and subsequent LEE expression during exponential growth. In addition, Hfq negatively affects LEE4 and LEE5 expression independent of grlRA in stationary phase. To emphasis the impact of Hfq on LEE expression, the regulation of LEE by additional regulators is excluded from this model (for a detailed model of LEE regulation see Mellies et al., 2007).