Identification of the DNA binding sites of PerA, the transcriptional activator of the bfp and per operons in enteropathogenic Escherichia coli

Abstract

The bundle-forming pilus (BFP) is an important virulence factor for enteropathogenic Escherichia coli (EPEC). Genes involved in its biogenesis and regulation are tightly regulated by PerA (BfpT), a member of the AraC/XylS family of transcriptional regulators. The aim of this work was to purify PerA and determine its association with bfpA and perA (bfpT) regulatory regions by electrophoretic mobility shift and DNase I footprinting assays. PerA was purified as a maltose-binding protein (MBP) fusion, which was capable of complementing bfpA expression and which was able to restore the localized adherence phenotype of an EPEC perA mutant strain. Upstream of bfpA and perA, MBP-PerA recognized with similar affinity asymmetric nucleotide sequences in which a 29-bp-long AT-rich consensus motif was identified. These DNA motifs share 66% identity and were previously shown, by deletion analysis, to be involved in the PerA-dependent expression of both genes. Interestingly, in perA, this motif spans the sequence between positions -75 and -47, approximately one helix turn upstream of the -35 promoter sequence, while in bfpA, it spans the sequence between positions -83 and -55, approximately two helix turns upstream from the promoter. An additional PerA binding site was identified at the 5' end of the bfpA structural gene, which was not required for its activation. Experiments with LexA-PerA fusions suggested that PerA acts as a monomer to activate the transcription of both perA and bfpA, in contrast to what has been documented for other members of this family of transcriptional regulators.

FIG. 1.

Purification and characterization of the MBP-PerA fusion. (A) PAGE of samples corresponding to different steps of the purification process. Lanes: 1, IPTG-induced cells; 2, crude extract after sonication and centrifugation; and 3, amylose column-purified protein. (B) Detection of MBP-PerA fusion by Western blotting with anti-PerA polyclonal antibodies. Lanes: 1, noninduced cells; 2, crude extract of IPTG-induced cells after sonication and centrifugation; 3, amylose column-purified protein. (C) Complementation of an EPEC E2348/69 perA::Km mutant with plasmid pMALT2 encoding MBP-PerA. Shown are Western blotting results with anti-BfpA antibodies. Lanes: 1, EPEC E2348/69; 2, JPN15; 3, perA::Km; 4, perA::Km/pMALT2. (D) LA assay on HeLa cells.

FIG. 2.

Binding of MBP-PerA to bfpA and perA. EMSAs with a bfpA fragment from positions −214 to −15 (A) and a perA fragment from positions −155 to +21 (B) were carried out with 0, 0.11, 0.21, 0.43, 0.87, 2.16, and 4.3 μg (0, 74 nM, 148 nM, 298 nM, 596 nM, 1.5 μM, and 3 μM, respectively) of purified MBP-PerA (lanes 1 to 7). Samples were resolved in 2.5% agarose gels, stained with ethidium bromide, and photographed with an Alpha-Imager.

FIG. 3.

Mapping of the PerA binding site on the bfpA regulatory region by EMSA. (A) Overlapping 200-bp-long PCR fragments that span the bfpA upstream and downstream promoter region from positions −374 to +262 were incubated with (+) or without (−) 2.16 μg (1.5 μM) of purified MBP-PerA. Samples were resolved in 6% native polyacrylamide gels and visualized by ethidium bromide staining. The asterisk denotes the second complex formed with fragment 7. (B) Schematic representation of the PCR fragments spanning the bfpA regulatory and structural regions. The fragment number is indicated to the left, and the sequence range covered for each fragment is indicated to the right. The approximate locations of the putative PerA binding sites, upstream and downstream of the bfpA promoter, are indicated by black and gray boxes, respectively. PBS, previously proposed PerA binding sequence

FIG. 4.

DNase I protection of the bfpA (A) and perA (B) regulatory regions by MBP-PerA. Increasing amounts of MBP-PerA were mixed with 100,000 cpm of a ³²P-end-labeled DNA fragment corresponding to positions −201 to +26 of bfpA and −155 to +81 of perA and treated with 0.003 U of DNase I. Samples were subjected to electrophoresis on an 8% polyacrylamide sequencing gel. The −35 promoter sequence and other upstream positions are indicated on the left; they were determined by running in parallel sequencing reactions with the same fragments (data not shown). The protected regions are indicated by vertical black bars. The black triangles above the gels represent increasing amounts of MBP-PerA. (A) Lanes: 1, DNA alone; 6 to 2, 159 nM, 316 nM, 634 nM, 1.26 μM, and 2.2 μM, respectively. (B) Lanes: 1, DNA alone; 7 to 2, 37 nM, 74 nM, 148 nM, 295 nM, 753 nM, and 1.5 μM, respectively.

FIG. 5.

PerA binds to double-stranded oligonucleotides containing the putative PerA binding site. EMSAs were performed by incubating ³²P-labeled double-stranded oligonucleotides of different lengths containing bfpA (A) or perA (B) sequences (as described in Tables 2 and 3, respectively), with 1.06 and 2.1 μM MBP-PerA. Control reactions with DNA alone were included as controls. A nonspecific 47-mer oligonucleotide containing an unrelated sequence was included as a negative control. In addition, control reactions with 50- and 100-fold molar excesses of the unlabeled 45-mer (PBSA45) were included to demonstrate the specificity of the interaction (data not shown). Samples were subjected to electrophoresis on a 6% native polyacrylamide gel.

FIG. 6.

PerA binding consensus sequence. (A) Nucleotide sequence alignment of the bfpA (upper line) and perA (lower line) regulatory regions. The boxed sequence denotes the 29-bp-long region that shares 66% identity between bfpA and perA, is required for PerA binding (this study), and has been previously shown to be required for the PerA-dependent activation of both promoters (5, 23). The proposed consensus PerA binding sequence is shown below the alignment. The −35 promoter sequence of both bfpA and perA is underlined. (B). Comparison of the second PerA binding sequence identified at the bfpA structural gene. Positions relative to the transcriptional start site are indicated to the left and right of each sequence.

FIG. 7.

Characterization of the second binding site on bfpA. A) The DNase I protection assay of a fragment containing the sequence from positions −54 to +166 was carried out as described in the legend to Fig. 4. The black triangle above the lanes indicates decreasing amounts of MBP-PerA (0, 2.6 μM, 1.7 μM, 880 nM, 528 nM, 176 nM, and 88 nM). (B) EMSA of a 56-mer double-stranded oligonucleotide containing the bfpA sequence between positions +38 and +93 (PBSA2-56) (Table 2). The triangle above the lanes indicates increasing amounts of MBP-PerA (0, 133 nM, 238 nM, 475 nM, 958 nM, 1.85 μM, 5.04 μM, 6.6 μM, and 8.4 μM). Probe PBSA56 was included as a control in the absence or presence of 1.85 μM MBP-PerA. (C) The second PerA binding sequence is not required for bfpA expression. EPEC strains E2348/69 (wild type) and JPN15 (pEAF cured) were transformed with either pCAT201 or pCAT + 27, which carry transcriptional fusions between the reporter cat gene and bfpA regulatory fragments from positions −201 to +76 and −201 to +27, respectively. The CAT specific activity was determined from samples collected from bacterial cultures grown on DMEM at an OD₆₀₀ of 1.0. The results are the average of three different experiments.

FIG. 8.

A LexA-PerA fusion does not form dimers. The ability of the full-length PerA protein fused to the LexA DBD to dimerize was evaluated by measuring the activity of a sulA::lacZ fusion in E. coli SU101. The β-galactosidase activity was determined from samples obtained from bacterial cultures grown on LB medium supplemented with 1 mM IPTG at an OD₆₀₀ of 0.8. The LexA-AraC, LexA-H-NS, and LexA-CAT fusions were used as positive controls. The results are the average of three different experiments.