Transcriptional regulation of the esp genes of enterohemorrhagic Escherichia coli

Abstract

We have determined that the genes encoding the secreted proteins EspA, EspD, and EspB of enterohemorrhagic Escherichia coli (EHEC) are organized in a single operon. The esp operon is controlled by a promoter located 94 bp upstream from the ATG start codon of the espA gene. The promoter is activated in the early logarithmic growth phase, upon bacterial contact with eukaryotic cells and in response to Ca2+, Mn2+, and HEPES. Transcription of the esp operon seems to be switched off in tightly attached bacteria. The activation process is regulated by osmolarity (induction at high osmolarities), modulated by temperature, and influenced by the degree of DNA supercoiling. Transcription is sigmaS dependent, and the H-NS protein contributes to its fine tuning. Identification of the factors involved in activation of the esp operon and the signals responsible for modulation may facilitate understanding of the underlying molecular events leading to sequential expression of virulence factors during natural infections caused by EHEC.

FIG. 1

Infection of HeLa cells with the EHEC strain EDL933. Cells were infected with EDL933 (a and c) or its ΔespA derivative (b and d) for 3 h. Then, monolayers were fixed, bacteria were labelled with TRITC-conjugated antibodies (a and b) and F-actin was stained with FITC-labelled phalloidin (c and d), and coverslips were examined by immunofluorescence microscopy. While the wild-type strain forms microcolonies with consistent actin accumulation (a and c), the ΔespA mutant has lost the ability to attach to HeLa cells (b) and to induce actin accumulation (d). Scales are in micrometers.

FIG. 2

(a) Northern blot analysis of the mRNA transcript encompassing the espA, espD, and espB genes. Total RNA extracted from EDL933 grown on DMEM-HEPES was fractionated on a 1% agarose gel, transferred to Byodine B membranes, and hybridized with probes specific for espA, espD, and espB. As a control, a probe that hybridizes within regions located upstream of espA (sepL) was used. The main RNA transcript is indicated by an arrow (approximately 2.8 kb). (b) Identification of the transcriptional start site from the esp operon by primer extension analysis. Total RNA was extracted from EDL933(pUJ3) grown exponentially at 37°C in medium supplemented with either 10 (lane 1) or 430 (lane 2) mM NaCl. A 24-bp oligonucleotide (FAB56), which hybridizes with positions +3 to −21 of the espA region, was used to perform primer extension and to generate a sequence ladder. The position of the first base in the main RNA message relative to the adenosine (base +1) of the ATG start codon is indicated.

FIG. 3

(a) Sequence of the esp promoter region. The start of transcription (+1), the putative −10 and −35 consensus sequences (underlined), the Shine-Dalgarno sequence (SD), the consensus binding sequence for H-NS (5′-TNTNAN-3′ [in boldface italic type]), and several inverted and direct repeats (underlined) are indicated. (b) Schematic representation of the constructs employed to study transcriptional regulation of the esp operon. Abbreviations: EV, EcoRV; lacZ, β-galactosidase-encoding gene.

FIG. 4

Expression of β-galactosidase by EDL933(pUJ3) after infection of HeLa cells. At different time intervals after infection, enzymatic activity was determined in bacteria present in supernatants (■) or attached to HeLa cells (●) and compared to that produced by EDL933(pUJ3) grown in DMEM (▴). The basal values of β-galactosidase obtained from EDL933 containing the promoterless plasmid under matching conditions were at least 10-fold lower than the basal levels of the tested clones and were subtracted from each sample. β-Galactosidase activities are expressed as relative light units (rlu) per 10⁵ bacteria and are means of three independent experiments; standard deviations were lower than 5%. Open symbols indicate numbers of CFU at each time point.

FIG. 5

β-Galactosidase induction in response to different media and micronutrients. EDL933(pUJ3) was grown in DMEM (▴) or DMEM supplemented with 100 mM HEPES (pH 7) (⧫) (a) or in M9-glucose medium supplemented with either CaCl₂ (⧫, 0 mM; ●, 0.01 mM; ■, 0.1 mM; ▴, 1 mM) (b) or MnSO₄ (⧫, 0 mM; ●, 0.0033 mM; ■, 0.33 mM; ▴, 3.3 mM) (c), and β-galactosidase activities were determined at different time intervals. Growth rate is indicated by open symbols (OD₆₀₀). Results are expressed as relative light units (rlu) per 10⁵ bacteria and are means of three independent experiments; standard deviations were lower than 5%. The background values for EDL933 containing the promoterless plasmid under matching conditions were at least 10-fold lower than the basal values at the tested conditions and were subtracted from each sample.

FIG. 6

Activation of the esp promoter in response to changes in temperature and osmolarity. EDL933 containing either pUJ3 (●) or its deletion derivative, pUJ3-285 (■), was grown in minimal medium (pH 7) supplemented with 10 mM (dotted lines) or 430 mM (solid lines) NaCl at 25°C (a), 37°C (b), or 42°C (c), and β-galactosidase activities were determined at different time intervals. Results are expressed as relative light units (rlu) per 10⁵ bacteria and are means of three independent experiments; standard deviations were lower than 5%. The background values for EDL933 containing the promoterless plasmid were at least 10-fold lower than the values at the tested conditions and were subtracted from each sample.

FIG. 7

Transcription of the esp operon is dependent on ς^S and H-NS. (a) Plasmid pUJ3 was transformed into E. coli MC4100 (●) and its ς^S-deficient derivative, RH90 (■), and β-galactosidase activities were determined under inducing (430 mM NaCl [solid lines]) and noninducing (10 mM NaCl [dotted lines]) conditions. (b and c) To investigate the role of the H-NS protein, plasmids pUJ3 (●) and pUJ3-285 (■) were transformed into E. coli GM37 (b) and its hns derivative GM230 (c). Strains were grown in M9-glucose medium supplemented with 10 mM (dotted lines) or 430 mM (solid lines) NaCl, and production of β-galactosidase was determined after different time intervals. Results are expressed as relative light units (rlu) per 10⁵ bacteria and are means of three independent experiments; standard deviations were lower than 5%. The background values for strains containing the promoterless plasmid were at least 10-fold lower than the basal values at the tested conditions and were subtracted from each sample. No differences in growth were observed.

FIG. 8

Influence of the degree of supercoiling on transcription of the esp promoter. EDL933(pUJ3) was grown in M9-glucose medium supplemented with 10 mM (open symbols) or 430 mM (solid symbols) NaCl in the presence of 0 (triangles), 5 (circles), 20 (squares), or 50 (diamonds) μg of the gyrase inhibitor novobiocin per ml, and β-galactosidase activities were determined at different time intervals. Results are expressed as relative light units (rlu) per 10⁵ bacteria and are means of three independent experiments; standard deviations were lower than 5%. The background values for EDL933 containing the promoterless plasmid were at least 10-fold lower than the basal values at the tested conditions and were subtracted from each sample. No differences in growth were observed at the novobiocin concentrations tested (not shown).