Ler Is a Negative Autoregulator of the LEE1 Operon in Enteropathogenic Escherichia coli

Abstract

Enteropathogenic Escherichia coli (EPEC) causes severe diarrhea in young children. Essential for colonization of the host intestine is the LEE pathogenicity island, which comprises a cluster of operons encoding a type III secretion system and related proteins. The LEE1 operon encodes Ler, which positively regulates many EPEC virulence genes in the LEE region and elsewhere in the chromosome. We found that Ler acts as a specific autorepressor of LEE1 transcription. We further show that Ler specifically binds upstream of the LEE1 operon in vivo and in vitro. A comparison of the Ler affinities to different DNA regions suggests that the autoregulation mechanism limits the steady-state level of Ler to concentrations that are just sufficient for activation of the LEE2 and LEE3 promoters and probably other LEE promoters. This mechanism may reflect the need of EPEC to balance maximizing the colonization efficiency by increasing the expression of the virulence genes and minimizing the immune response of the host by limiting their expression. In addition, we found that the autoregulation mechanism reduces the cell-to-cell variability in the levels of LEE1 expression. Our findings point to a new negative regulatory circuit that suppresses the noise and optimizes the expression levels of ler and other LEE1 genes.

Colonizing enteropathogens compete with the gut flora to gain a foothold in the host tissue by expressing powerful colonization factors. However, to reduce the immune response of the host, the pathogen should minimize the expression of the colonization factors. To resolve this dilemma, pathogens evolved regulatory mechanisms that optimize the expression levels and timing, thus maintaining expression of just enough colonization factors and only when needed. Another layer of complexity is added when the colonization is dependent on the assembly of organelles like the type III secretion systems (TTSS), which are composed of ∼30 different proteins of various relative amounts and encoded by several operons. In these cases, an orderly expression program is required for efficient assembly of the organelle.

Enteropathogenic Escherichia coli (EPEC) causes severe diarrhea in young children. It employs the TTSS as a molecular syringe to inject a battery of toxic or colonization proteins into the membrane and cytoplasm of infected host cells (4). The TTSS and some of the effectors are encoded by a 35.6-kbp pathogenicity island, termed the locus for enterocyte effacement (LEE). The LEE consists of 41 genes, organized in five major operons (LEE1 to LEE5) and several additional transcriptional units (10, 19). Ler, an H-NS paralog, encoded by the first gene of the LEE1 operon, is a key regulator of the LEE regulon, positively regulating expression of LEE2, LEE3, LEE4, LEE5, espG, and map (11, 19, 24, 30). The regulation of ler (LEE1 operon) is complex and involves many factors, including H-NS, integration host factor (IHF), Fis, PerC, BipA, GrlA, GrlR, GadX, and quorum sensing (2, 7, 11, 13, 14, 17, 19, 26, 28, 30, 32). Most of these factors appear to mediate the temporal regulation of Ler expression in response to the changing environment.

We investigated the mechanism that controls the level of Ler expression. We show that the Ler expression level is determined by autorepression. We also demonstrate that autoregulation reduces the cell-to-cell variability in the LEE1 expression levels. In addition, we show that Ler specifically binds to the LEE1 regulatory sequence with an affinity that allows expression of Ler levels that are relatively low but still sufficient for binding to the LEE2-LEE3 promoter region and activation of these promoters. Thus, autoregulation is required for balancing the expression of the Ler regulon to the optimal levels.

MATERIALS AND METHODS

Bacterial strains, culture conditions, and oligonucleotide primers.

The bacterial strains, plasmids, and primers used in this study are listed in Tables 1 and 2. Strains were grown overnight in Luria broth (LB) at 27°C, diluted 1:50 in buffered (20 mM HEPES, pH 7.4) Dulbecco modified Eagle medium (DMEM) or 1:10 in a modified Casamino-DMEM [0.25 μM Fe(NO₃)₃, 1.4 mM CaCl₂, 5.4 mM KCl, 0.8 mM MgSO₄, 110 mM NaCl, 1 mM Na₂HPO₄, 44 mM NaHCO₃, 0.45% glucose, 0.1 M HEPES, 0.1% Casamino Acids] or in LB. To achieve maximal repression of LEE1 expression (repressive conditions), overnight cultures grown in LB at 27°C were diluted 1:50 in LB containing 20 mM (NH₄)₂SO₄ and subsequently grown at 27°C with shaking. When needed, we used ampicillin (AMP) at 100 μg/ml, kanamycin (KAN) at 40 μg/ml, or chloramphenicol (CM) at 25 μg/ml.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Description	Reference or source
E. coli strains
E2348/69	EPEC wild type; Smr	J. Kaper
DF2	E2348/69 ler::kan	11
TU1403	E2348/69 carrying chromosomal ler-6his	This study
MC4100	F′ araD139 Δ(argF-lac) U169 deoC1 flb-5301 relA1 rpsL150 ptsF25 rbsR	3
DF1215	MC4100 lysogenized with λ PLEE1-lacZ	This study
W3110	F− λ−thyA36 deoC2 IN (rrnD-rrnE)1
Plasmids
pIR1	pKK177-3 derivative containing gfp mut3	11
pIR1Ler	pIR1 encoding PLEE1-ler-gfp	This study
pIR1LerL29R	pIR1 encoding PLEE1-lerL29R-gfp	This study
pSA10	pKK177-3 derivative containing lacIq	25
pTU12	pSA10 encoding Ptac-ler	This study
pTU14	pACYC184 encoding lacIq and Ptac-ler	This study
pSA8	pIR1 containing PoxyS-gfp	25
pSA11	pIR1 containing Ptac-gfp-lacIq	25
pHG86	A pRS415 derivative carrying a lacZ reporter gene	12
pDF9	pHG86 containing PLEE1-lacZ	This study
pSUB7	A template for six-His tagging of chromosomal genes	33
pKD46	λred recombinase expression plasmid	6
pTU609	pIR1 containing 25F-gfp	This study
pTU606	pIR1 containing 9F-gfp	This study
pTU608	pIR1 containing 23F-gfp	This study
pTU312	pIR1 containing PLEE1-gfp	This study

TABLE 2.

List of primers^a

Primer	Sequence
5R	GCTCTAGATTCTCTGTTTTCTAATGTG (3924)
9F	GGAATTCGTTTATGCAATGAGATCTATC (3740)
25F	CGGGATCCGCGGTTACTTGTTCAGC (3470)
23F	CGGGATCCGTTGACATTTAATGATAATG (3879)
27R	GCTCTAGAGAGAGCGTCAGCGAAACG (4183)
36F	GGAATTCATGAATATGGAAACTAATTCAC (4088)
38R	AAACTGCAGCTTCCAGCTCAGTTATCG (4498)
46F	CCGCTCGAGTTTATGCAATGAGATCT (3740)
48R	CGGGATCCAATATTTTTCAGCGGTATTATTTCTTC (4453)
4F	GGTGGTTGTTTGATGAAATAG (3939)
10F	GGAATTCGATGGTTTTCTTCTTTATGATTG (3820)
126F	CCAATCATGATGGTTCATG (5067)
127R	CTGACCTGATTGACACCG (5125)
119F	GTGCCTGATGATGGACTC (4268)
123R	TCCTGATTCGCATTATCTC (4668)
Pler1R	GCTCTAGATGTTAAATATTTTTCAGCGG (4461)
Pler2F	GCAAATTGCAGTTCGACAGCAGGAAGCAAAGCG (4141)
Pler3R	CGCTTTGCTTCCTGCTGTCGAACTGCAATTTGC (4173)
30F	GCTCTAGACTGCGTACGCTCAGGAGC (14654)
31R	GCTCTAGACTGCGTACGCTCAGGAGC (15058)
EtkC-F	ATGGATCCGGCGCAATCCTCAATGGT
EtkC-R	ATAAGCTTTTACTCTTTCTCGGAGTAAC
lacZ2	CCGCTCGAGCCTTGTGGAGCGACATCC (2028)
lerhistagF	ATTTTCTTGTGAAGGACACTGAAGAAGAAATAATACCGCTGAAAAATATTCACCACCATCATCACCATTAG (4407)
lerhistagR	TATTATTTCGTCTTCCAGCTCAGTTATCGTTATCATTTAATTATTTTATGATGAATATCCTCCTTAGTTC (4501)

^aThe numbers in parentheses indicate the positions of the primers' 5′ ends, by use of the numbering of the LEE region and lacZ2 primer sequences submitted under accession numbers AF022236 and V00296, respectively.

Measurement of gene expression.

Cultures containing the gfp-expressing plasmids pIR1Ler, pIR1Ler(L29R), and pIR1 were grown overnight at 27°C with shaking in LB supplemented with 20 mM (NH₄)₂SO₄. The cultures were washed and diluted 1:10 in Casamino-DMEM and then grown in 96-well plates at 37°C in a microplate reader (SPECTRAFluor Plus; Tecan). The fluorescence intensity (filter set at 485-nm excitation wavelength and 535-nm emission wavelength) and optical densities at 590 nm (OD₅₉₀) were read every 5 min, and data was collected by XFluor4 software. For analysis of gfp expression by fluorescence microscopy, EPEC ler::kan containing pIR1Ler or pIR1Ler(L29R) was grown overnight at 27°C in LB, diluted 1:50 into DMEM, and grown at 27°C to an OD₆₀₀ of 0.3. The temperature was then shifted to 37°C to activate the LEE1 promoter, and after 30 min, the bacteria were treated with 3% formaldehyde, washed with phosphate-buffered saline, and attached by centrifugation to coverslips precoated with 0.01% poly-l-lysine. The slides were viewed with a fluorescence microscope (Zeiss Axiolab), and images were captured with a charge-coupled-device camera (DVC1310). The fluorescence intensity of individual bacteria was determined using Image-ProPlus version 4 (Media Cybernetics). In some cases, the level of gfp expression was monitored by flow cytometry as described previously (32). The β-galactosidase activity was measured as described before (20).

Construction of transcriptional fusions with gfp.

The plasmid pIR1 (11) containing the promoterless gfp reporter gene was used to construct the different transcriptional fusions. DNA fragments were amplified using specific oligonucleotides, digested with BamHI and XbaI, and cloned into the BamHI and XbaI sites of pIR1, yielding transcriptional fusions with gfp. Plasmid pIR1Ler was constructed with primers 9F and Pler1R. Plasmid pIR1Ler(L29R) was created by introducing the L29R point mutation in pIR1Ler by use of a QuickChange site-directed mutagenesis kit (Stratagene) and primers Pler2F and Pler3R.

Lysogenization with lambda phage carrying PLEE1-lacZ.

A PCR DNA fragment extending from nucleotides −159 to +157, relative to the LEE1 transcriptional start site, was ligated into the BamHI and EcoRI sites of plasmid pHG86 (12), which was derived from pRS415 (27), yielding pDF9. This plasmid was used to construct E. coli K-12 strain MC4100 lysogenized with a lambda phage containing the PLEE1-lacZ transcriptional fusion, as described previously (3, 27). The new strain was designated DF1215.

Construction of an EPEC strain expressing native Ler with a six-His tag at the C terminus.

A DNA fragment containing a kanamycin resistance gene was amplified, using pSUB7 (33) as a template and lerhistagF and lerhistagR as primers. The ler gene was tagged as described previously (33). To confirm the construction of the tagged ler gene, a chromosomal DNA fragment containing ler-6his-kan-orf2 was amplified and sequenced using primers 119F and 123R. The expected sequence was obtained. The EPEC strain expressing the tagged Ler (TU1403) exhibited a phenotype similar to that of the wild type with respect to the expression of the LEE genes (data not shown). The tagged Ler bound efficiently to Ni- or Co-coated beads (Sigma), but for an unknown reason it did not react with anti-six-His antibody (Sigma).

Construction of plasmids expressing ler.

ler was amplified with primers 36F and 38R, digested with EcoRI and PstI, and cloned into the EcoRI-PstI sites of pSA10, yielding pTU12. This construction positioned the ler open reading frame downstream of the plasmid tac promoter and the ribosomal binding site. To construct pTU14, which is a pACYC184 derivative expressing ler, a BamHI-ScaI DNA fragment containing ler, the regulatory sequences, and the lacI^q gene was isolated from pTU12 and ligated into the BamHI-EcoRV sites of pACYC184.

Protein extraction and immunoblot analysis.

Bacteria were grown in DMEM to an OD₆₀₀ of 0.3, and cells were harvested by centrifugation and extracted by boiling for 5 min in 50 mM Tris-HCl (pH 7.0)-1% sodium dodecyl sulfate (SDS). The protein concentrations in the samples were determined using a bicinchoninic acid kit (Sigma), adjusted accordingly, and the samples were subjected to Western blot analysis (32).

Determination of Ler and Ler(L29R) stability.

An EPEC ler::kan mutant containing pIR1Ler or pIR1Ler(L29R) was grown in DMEM to an OD₆₀₀ of 0.33. CM at 100 μg/ml was added, bacteria were harvested at different time points, and the levels of Ler and a Ler protein containing the mutation L29R [Ler(L29R)] were determined by Western blot analysis using anti-Ler antibody.

ChIP assay.

The chromatin immunoprecipitation (ChIP) assays were based on previous reports (16, 18, 29, 31). Briefly, 50-ml cultures of EPEC and an EPEC ler::kan mutant (see Fig. 5A) or W3110/pIR1Ler and W3110/pIR1Ler(L29R) (see Fig. 5B) were grown to an OD₆₀₀ of 0.35 and treated with 1% formaldehyde for 10 min at room temperature followed by 30 min on ice. The bacteria were washed three times with 40 ml of PBS (pH 7.3) and frozen in liquid nitrogen. The frozen pellets were resuspended in 0.5 ml of solution A (10 mM Tris-HCl [pH 8.0], 20% sucrose, 50 mM NaCl, 10 mM EDTA) and 0.5 ml of 2× immunoprecipitation buffer (100 mM Tris-HCl [pH 7.0], 300 mM NaCl, 2% Triton X-100, 1 mM phenylmethylsulfonyl fluoride). The samples were sonicated repeatedly to shear the DNA to an average size of 300 to 600 bp. Insoluble cellular debris was removed by centrifugation, the supernatant was transferred to a fresh tube, and 75 μl was removed and stored for later analysis (total-DNA control). The Ler and Ler-DNA complexes were then immunoprecipitated from the remaining supernatant with agarose-protein A beads (Sigma) precoated with anti-Ler antiserum. The beads were incubated with the supernatants (4°C, overnight) to allow Ler binding and then washed 10 times with 1 ml of 1× immunoprecipitation buffer and twice with 1 ml of TE (10 mM Tris [pH 8.0], 0.1 mM EDTA), to remove unbound proteins and DNA. The slurry was resuspended in 0.05 ml of TE. The 75 μl of total-DNA control was treated with pronase E (final concentration, 0.1 mg/ml; Sigma) for 10 min at 37°C, and SDS was added (final concentration, 0.67%). The formaldehyde cross-links of both the total DNA and the immunoprecipitated DNA were reversed by incubation at 65°C overnight, and samples were used for PCR without further treatment. PCR was performed with Taq DNA polymerase, with serial dilutions of the immunoprecipitated DNA and the total-DNA control used as templates. The primers used to amplify the DNA sequence of the LEE1 regulatory region were 10F and 5R, those for the downstream LEE1 regulatory region were 126F and 127R, and those for the DNA region outside the LEE (etk) were EtkC-F and EtkC-R. The amplified products, 60 to 120 bp in length, were separated on 5% polyacrylamide gels and stained with ethidium bromide.

FIG. 5.

Binding of Ler and Ler(L29R) to the LEE1 regulatory region in vivo. (A) EPEC wild-type (wt) and EPEC ler::kan mutant bacteria were grown in DMEM at 37°C to an OD₆₀₀ of 0.35. The cultures were then used for ChIP analysis with anti-Ler antibody. Serial 2-fold dilutions of immunoprecipitated DNA (lanes 1 to 5) or 10-fold dilutions of the total input DNA (lanes 6 to 9) were used as templates for PCR. The reactions utilized primers specific to sequences located upstream of the LEE1 promoter (LEE1 regulatory region), primers specific to a region downstream of the LEE1 regulatory region, and primers specific to sequences located outside the LEE (see Materials and Methods for details). The PCR products were separated on 5% polyacrylamide gels. (B). E. coli K-12 W3110 bacteria containing pIR1Ler or pIR1Ler(L29R) were grown and subjected to ChIP analysis, as described above. The 1:5 (lane 1), 1:25 (lane 2), and 1:125 (lane 3) dilutions of immunoprecipitated DNA were subjected to PCR using primers specific to the LEE1 regulatory region. The PCR products were separated on 5% polyacrylamide gels.

Preparation of Ler.

Ler was purified as described previously (T. Berdichevski et al., unpublished data). Briefly, E. coli W3110 containing pDF5, coexpressing Ler and 6His-Ler, was grown in 400 ml of LB at 37°C, with shaking, to an OD₆₀₀ of 0.6. The temperature was then shifted to 20°C. After 20 min at 20°C, IPTG (isopropyl-β-d-thiogalactopyranoside; 1 mM) was then added, and the culture was incubated for 18 h at 20°C. The bacteria were then harvested by centrifugation, resuspended in sonication buffer (50 mM Tris-HCl [pH 8.0], 600 mM NaCl, 100 μg of phenylmethylsulfonyl fluoride/μl), sonicated on ice, subsequently supplemented with 0.5% Triton X-100, and cleared by repeated centrifugation. Complexes of 6His-Ler bound to Ler were purified from the cleared supernatant under native conditions by metal affinity chromatography (Talon; Clontech), according to the manufacturer's instructions. Ler was specifically eluted from the complex by use of urea elution buffer (8 M urea, 50 mM Tris-HCl [pH 8.0], 0.05% Tween 20). To remove residual 6His-Ler, the eluted Ler was adsorbed several times onto fresh agarose-Co beads, preequilibrated with urea elution buffer. The purified Ler was dialyzed with renaturation buffer (100 mM Tris-HCl [pH 7.2], 100 mM NaCl) containing decreasing concentrations of urea (4 M, 2 M, 1 M, 0.5 M, and no urea). Dialysis was carried out at 4°C for 1 h against buffer containing urea at each of the decreasing concentrations and for 18 h against the urea-free buffer. The protein concentration of the purified Ler was determined using the bicinchoninic acid kit (Sigma). The Ler purity was verified by SDS-polyacrylamide gel electrophoresis and Coomassie blue staining. The purified Ler was stored at −20°C in Ler storage buffer (50 mM Tris-HCl [pH 7.2], 50 mM NaCl, 30% glycerol).

DNA mobility shift assay.

Three DNA fragments were amplified by use of the primers indicated: a 186-bp fragment corresponding to nucleotides −173 to +11, relative to the LEE1 transcriptional start site, with primers 9F and 5R (LEE1 fragment −173+11); a 244-bp fragment corresponding to nucleotides +26 to +270, relative to the LEE1 transcriptional start site, with primers 4F and 27R (LEE1 fragment +26+270); and a 395-bp PCR fragment corresponding to nucleotides +117 to −288, relative to the LEE2 transcriptional start site, with primers 30F and 31R (LEE2-LEE3 fragment +117−288). These fragments were used for gel shift analysis with the Ler protein as described previously (11, 32).

DNase I footprinting.

For DNase I footprinting, the 186-bp LEE1 fragment (−173 +11) was end labeled at the XbaI site, as described previously (11, 32). For DNA sequencing, the 5R primer was labeled at the 3′ end with [γ-³²P]dATP (3,000 Ci/mmol; Amersham) by using polynucleotide kinase (MBI Fermentas). Ler or IHF proteins were bound to DNA in the binding buffer as described previously (11, 32). After an incubation period of 20 min, 1 μl of 100 mM MgCl₂ (10 mM final concentration) and 1 μl of a 1:2.5 dilution of DNase I (RQ1 DNase; 1 U/μl; Promega) were added for 2 min at 25°C. The reaction was stopped by adding 4 μl of stop solution containing 95% formamide, 20 mM EDTA, 0.05% bromophenol blue, and 0.05% xylene cyanol (USB-Amersham). The samples were heated at 80°C for 3 min and loaded onto a 6% sequencing gel along with the corresponding sequencing reaction. The gel was dried and exposed to autoradiography.

RESULTS

Identification of a minimal regulatory region of the LEE1 operon.

To define the LEE1 regulatory region, we constructed plasmids containing transcriptional fusions between gfp and serially deleted DNA sequences upstream of the LEE1 promoter. Plasmids pTU608, pTU606, and pTU609, containing fusions 23F-gfp, 9F-gfp, and 25F-gfp, respectively (Fig. 1), were introduced into EPEC bacteria, and the levels of gfp expression were compared. The data show that 9F-gfp contains a region essential for LEE1 transcription. This region is 123 bp in size, between nucleotides −159 and −36 relative to the transcription start site (Fig. 1), and contains an IHF-binding site that is required for LEE1 expression (11, 34). However, 9F-gfp also contains the putative ler ribosome binding site and a part of the ler coding region, which could interfere with the translation of gfp. We therefore constructed a plasmid (pTU312) expressing gfp under the same regulatory region but lacking the ler coding region and the putative ler ribosome binding site (Fig. 1). This plasmid indeed expressed elevated GFP levels (Fig. 1). This fusion of gfp to the LEE1 promoter (PLEE1-gfp) on pTU312 was used for our further investigations.

FIG. 1.

Schematic representation of the LEE1 regulatory region and of the gfp transcriptional fusions used to determine the minimal regulatory region of LEE1. The transcriptional start site of the LEE1 operon is marked +1, and the translational start site of Ler is marked +174. The plasmids containing the gfp fusions are indicated. The values to the right of each fusion indicate the fluorescence intensity, emitted by EPEC bacteria containing the corresponding fusion. These bacteria were grown at 37°C in DMEM to an OD₆₀₀ of 0.3. Fluorescence intensity was measured from 10,000 bacteria by flow cytometry, and a mean result for three independent experiments is presented with the standard deviation shown in parentheses.

Ler overexpression specifically represses the LEE1 promoter.

In E. coli, many central regulators are negatively autoregulated. This mechanism enables rapid promoter activation upon stimulation while maintaining a relatively low steady-state level of the regulator (22). In addition, an autorepression mechanism reduces the cell-to-cell variability in the levels of promoter activity (1). To determine whether Ler is a negative autoregulator, we introduced plasmid pTU14, expressing ler from the tac promoter, into an EPEC strain containing pTU312, which carries the PLEE1-gfp transcriptional fusion. Overexpression of Ler resulted in a 7.5-fold reduction in gfp expression by the LEE1 promoter (Fig. 2A), suggesting that Ler negatively regulates PLEE1. In contrast, Ler overexpression did not repress the activity of tac and oxyS promoters, which were used as negative controls (Fig. 2A). Similar results were obtained when instead of EPEC we used E. coli K-12 (data not shown). Cumulatively, these results indicate that Ler specifically represses the LEE1 promoter.

FIG. 2.

Specific repression of the LEE1 promoter by Ler overexpression. (A) Plasmids pTU312, pSA11, and pSA8, carrying gfp under the regulation of the LEE1 promoter (PLEE1), the tac promoter (Ptac), or the oxyS promoter (PoxyS), respectively, were transformed into EPEC bacteria containing (+) or lacking (−) a compatible plasmid (pTU14), expressing Ler from Ptac. Cultures were grown to an OD₆₀₀ of 0.13, and IPTG (1 mM) was added for about 2.5 h to induce Ptac. The PoxyS promoter was activated by adding hydrogen peroxide (1 mM) to cultures at an OD₆₀₀ of 0.27 for 15 min. The level of gfp expression from 10,000 bacteria was determined by flow cytometry, and mean results for three independent experiments are shown. Vertical lines on the bars indicate standard errors. (B) Effect of ectopically expressed Ler on the expression of a chromosomal PLEE1-lacZ transcriptional fusion in E. coli K-12 MC4100. Ptac-ler was induced by IPTG as described for panel A. The expression levels are the mean values for three experiments, with vertical lines indicating the standard errors. (C) EPEC bacteria encoding chromosomal Ler-6His (EPEC ler-6his) containing (+) or lacking (−) pTU14 (expressing ler) were grown to an OD₆₀₀ of 0.35. The bacteria were then harvested and lysed. A portion of the crude extract was subjected to SDS-polyacrylamide gel electrophoresis and Western analysis using anti-Ler, anti-DnaK, anti-Tir, or anti-EscJ antibodies. 6His-Ler was precipitated from the remaining crude extract by use of Talon cobalt beads, eluted, and analyzed by Western blotting, using anti-Ler antibody. The antibodies, used for the development of the blots, are indicated above each blot, and arrows indicate the corresponding proteins. Molecular weights (in thousands) are indicated. The specificity of the anti-Ler antibody was verified. The antibody did not react with a band corresponding to Ler in an extract of the EPEC ler::kan mutant, and preimmune antiserum did not react with a band corresponding to Ler in an extract of wild-type EPEC (data not shown).

To substantiate the results, we investigated the effect of Ler on the expression of the EPEC native chromosomal ler gene. The chromosomal ler gene was tagged at its 3′ end with a sequence encoding six histidine residues and a stop codon. This strain, EPEC ler-6xhis (designated TU1403), was transformed with pTU14, expressing ler from the tac promoter. TU1403 and TU1403 containing pTU14 were grown under conditions repressive to the LEE1 and tac promoters to an OD of 0.13 (see Materials and Methods). These promoters were activated by placing the cultures into inducing conditions (DMEM supplemented with 0.5 mM IPTG at 37°C), and the cultures were grown to an OD₆₀₀ of 0.35 without shaking. Then, the bacteria were harvested and lysed. The 6His-Ler protein was precipitated by using six-His-binding beads and then eluted and analyzed by Western blotting, using anti-Ler antibody. As a control, the lysate was subjected to Western analysis using antibodies to Ler, DnaK, Tir, and EscJ. We found that Ler expressed in trans caused a >5-fold repression in the production of the chromosomally encoded 6His-Ler (Fig. 2C). In contrast, the DnaK and EscJ levels were not significantly affected by Ler overexpression, and those of Tir were increased upon Ler overexpression. As expected, increased levels of Ler were detected in the crude extract of a strain containing pTU14 (Fig. 2C). Together these results indicated that Ler, expressed in trans, specifically represses the expression of the chromosomally encoded Ler in EPEC.

We next tested whether autorepression of the chromosomally carried LEE1 promoter also takes place in E. coli K-12. We constructed a transcriptional fusion, PLEE1-lacZ, between lacZ and a LEE1 regulatory region identical to that contained within pTU312. E. coli K-12 (strain MC4100 ΔlacZ) was subsequently lysogenized with lambda phage containing a PLEE1-lacZ transcriptional fusion, generating strain DF1215. We then compared PLEE1-mediated expression levels of lacZ in DF1215 and DF1215 containing pTU14, which expresses ler. Ectopic expression of ler resulted in a 12-fold reduction in lacZ expression (Fig. 2B). This finding further indicated that Ler is an autorepressor and that EPEC-specific factors are not needed for this activity.

The steady-state level of Ler is controlled by a negative autoregulation mechanism.

In the previous experiments, a plasmid-borne ler gene was overexpressed from a tac promoter, and thus the 1:1 ratio between the copy number of ler and its promoter (PLEE1) was not maintained. To test for Ler-negative autoregulation in a setting that better reflects the physiological conditions, we constructed two bicistronic operons (Fig. 3A). Both contained ler and gfp under the control of the native LEE1 regulatory region. In these constructs, the 1:1 ratio between ler and its native promoter is maintained, but they differ in one base that introduced a missense mutation in ler to replace leucine 29 with arginine [Ler(L29R)]. This mutation inactivates Ler by disrupting a coiled-coil domain, which is required for its activity (30). We predicted that if Ler is an autorepressor, the L29R mutation should result in increased activity of the LEE1 promoter and thus an increase in gfp expression.

FIG. 3.

(A) Schematic representation of the two bicistronic operons in pIR1Ler and pIR1Ler(L29R). (B and C) Expression of gfp by the EPEC ler::kan mutant (B) or E. coli K-12 W3110 (C), containing pIR1 (dotted line), pIR1Ler (thin line), or pIR1Ler(L29R) (thick line). The experiments shown in panels B and C were repeated several times, and the results of typical experiments are shown. In the inset in panel B, the levels of Ler and Ler(L29R) in the EPEC ler::kan mutant containing pIR1Ler and pIR1Ler(L29R) are shown by use of a Western blot developed with anti-Ler antibody. (D) Stability of Ler and Ler(L29R). EPEC ler::kan mutant bacteria containing either pIR1Ler or pIR1Ler(L29R) were grown in DMEM, at 37°C, to an OD₆₀₀ of 0.3. Chloramphenicol was then added to the culture (time zero), and levels of Ler and Ler(L29R) at different time points posttreatment were determined by Western blotting using anti-Ler antibody. To achieve similar initial amounts of Ler and Ler(L29R), extracts containing Ler(L29R) were diluted fivefold before the gel was loaded.

We introduced these constructs into an EPEC ler::kan mutant, and cultures were grown under conditions repressive for the LEE gene expression [27°C with shaking in LB broth supplemented with 20 mM (NH₄)₂SO₄] (23, 24, 32). To activate the LEE1 promoter, the cultures were diluted into modified DMEM medium and grown in 96-well plates at 37°C. Growth (OD₅₉₀) and fluorescence levels were determined at 5-min intervals, and the specific GFP activity was plotted (Fig. 3B). We found a ∼10-fold increase in gfp expression when the L29R mutation was introduced into Ler (Fig. 3B). This result indicates that the wild-type Ler negatively autoregulates the LEE1 promoter, while Ler(L29R) is deficient in autorepression. The Ler and Ler(L29R) levels produced by the two constructs were analyzed by Western blotting with anti-Ler antibody (Fig. 3B, inset). In agreement with the determined GFP levels, the level of Ler(L29R) is about fivefold higher than that of wild-type Ler. To test the possibility that Ler(L29R) (Fig. 3B, inset) is accumulating due to increased stability, we compared the stabilities of Ler and Ler(L29R). We found that in comparison to Ler, Ler(L29R) is somewhat less stable (Fig. 3D). These results indicate that the increased levels of Ler(L29R) are not due to its increased stability.

Ler autorepression was also observed when we used the E. coli K-12 strain W3110, which contains neither ler nor other LEE genes, instead of the EPEC ler::kan strain (Fig. 3C). These results indicate that the Ler autorepression is not dependent on other EPEC-specific genes and may suggest that Ler interacts directly with the LEE1 regulatory region to mediate the repression.

Ler interacts with the LEE1 regulatory region in vitro.

To test the above-described hypothesis, we examined whether purified Ler binds to the LEE1 regulatory region, using the mobility shift assay. We found that Ler binds to the LEE1 fragment −173+11 to form a low-mobility Ler-DNA complex (Fig. 4A). In contrast, under the same conditions, Ler did not bind to the LEE1 fragment +26+270 (Fig. 4A). We also tested the binding of Ler to a DNA fragment that contains a Ler binding site which is required for activation of the LEE2 and LEE3 promoters (the LEE2-LEE3 fragment +117−288) (30). The purified Ler interacted with this fragment, forming a low-mobility complex (Fig. 4A). To calculate the K_d, the gels were quantified by phosphoimaging analysis. Ler affinity to the LEE2-LEE3 fragment +117−288 (K_d = 48 nM) was found to be about 1.4-fold higher than its affinity to the LEE1 fragment −173+11 (K_d = 68 nM). In both cases, the Ler-DNA interaction results displayed an ultrasensitive sigmoidal curve, indicating a cooperative mode of binding (Fig. 4B). At a concentration close to the K_d, a Ler concentration increase of about twofold was sufficient to shift the majority of the DNA from the free to the Ler-bound form (Fig. 4B). Thus, small changes in Ler concentration strongly affect the Ler-DNA interaction. It appears that the difference in binding affinities of Ler to the LEE1 and the LEE2-LEE3 regulatory regions reflects the need to express just enough, but not too much, of ler and consequently of the other LEE genes.

FIG. 4.

In vitro binding of Ler to the promoter region of LEE1 and LEE2-LEE3. (A) A gel mobility shift assay was carried out using purified Ler and three different DNA fragments: LEE1 fragment −173+11, containing the LEE1 regulatory region; LEE1 fragment +26+270, a negative control DNA; and LEE2-LEE3 fragment +117−288 containing the regulatory region of the LEE2-LEE3 promoters (see Materials and Methods for details). The DNAs were mixed with twofold-increasing concentrations of Ler. The numbers above the lanes indicate nanomolar Ler concentrations. (B) The DNA-protein interaction with LEE1 fragment −173+11 and LEE2-LEE3 fragment +117−288 was quantified by phosphorimaging, and the results were plotted. The corresponding fragment is indicated above each graph. (C) DNase I footprinting analysis of LEE1 fragment −173+11. DNA was mixed and incubated with the indicated amounts of Ler or IHF prior to the DNase I treatment. The vertical line on the left side indicates the DNA region protected by Ler. Vertical lines on the right side indicate the DNA regions protected by IHF and the −35 sequence. The corresponding nucleotide sequence is shown in the lanes labeled G and A.

To further define the binding of Ler to the LEE1 regulatory region, we performed a DNase I footprint analysis using the LEE1 fragment −173+11. Ler protected an extended DNA sequence of the LEE1 regulatory region that includes the IHF binding site and the promoter, but no specific boundaries were found (Fig. 4C). For a positive control, we also tested the IHF footprint (Fig. 4C). In this case, only a relatively small and specific region was protected (11).

Ler binds to the LEE1 regulatory region in vivo.

To examine whether the in vitro interaction of Ler with the LEE1 regulatory region mirrors the in vivo situation, we used ChIP. Cultures of the EPEC wild type and the EPEC ler::kan mutant, which was used as a negative control, were treated with formaldehyde to cross-link DNA and DNA-associated proteins. Ler and Ler-DNA complexes were then immunoprecipitated from the bacterial extracts, using anti-Ler antibody. The DNA was extracted from the complexes and used as a template in PCR, using specific primer pairs. A product was obtained with primers specific to the LEE1 regulatory region when the template was the DNA precipitated from wild-type EPEC but not when the template originated from the EPEC ler::kan mutant (Fig. 5A). In contrast, we did not obtain any products with the same templates when primers specific to the region downstream of the LEE1 regulatory region or of the etk gene, which is located outside the LEE, were used (Fig. 5A). As a control for amplification efficiency with the different primers, we verified that a similar amount of PCR products was obtained when total DNA as a template and either pair of primers were used (Fig. 5A). Together with the findings of the in vitro analysis, these results support the hypothesis that Ler specifically binds, in vivo, to the LEE1 regulatory region.

Reduced binding efficiency of Ler(L29R) to the LEE1 regulatory region.

ChIP was used to compare the binding efficiencies of Ler and Ler(L29R) to the LEE1 regulatory region in vivo. We used anti-Ler antibody to precipitate protein-DNA complexes from extracts of strains W3110/pIR1Ler and W3110/pIR1Ler(L29R). DNA coprecipitated with Ler or Ler(L29R) was amplified using primers specific to the LEE1 regulatory region (Fig. 5B). The results indicated that introducing the L29R mutation into Ler resulted in an ∼10-fold reduction in binding to the LEE1 regulatory region (Fig. 5B). However, the level of Ler(L29R) in the expressing bacteria is ca. fivefold higher than that of Ler (Fig. 3B). Therefore, it appears that the Ler(L29R) binding efficiency to the LEE1 regulatory region is reduced more than 10-fold.

Ler autorepression enhances the expression stability of the LEE1 promoter.

Using the tetR promoter, Becskei and Serrano (1) demonstrated that an important consequence of autorepression is an increase in the expression stability of the autoregulated promoter. The term “stability” refers to the degree of cell-to-cell variability in the expression levels of a given promoter, and it can be quantified by the coefficient of variation (standard deviation divided by the mean). To determine if the LEE1 promoter is also stabilized by the autoregulation, we tested the effect of the L29R mutation on the expression stability. To this end, we used the EPEC ler::kan mutant containing plasmid pIR1Ler or pIR1Ler(L29R) (Fig. 3A). The two strains were grown as described in Materials and Methods, fixed, and observed with a fluorescence microscope, and representative images were captured (Fig. 6). In addition, the gfp expression levels of individual bacteria (n = 600 for each strain) were determined, and the results were plotted (Fig. 6). The coefficient of variation of GFP levels in cells producing Ler(L29R) was 2.4-fold higher than that of cells producing Ler (Fig. 6). These differences in gfp expression are also clearly visible in the image shown in Fig. 6. A similar reduction in stability upon introducing the L29R mutation in Ler was observed when we used E. coli K-12 strain W3110 instead of the EPEC ler::kan mutant (data not shown). Altogether these results indicate that negative autoregulation enhances the expression stability of the LEE1 promoter.

FIG. 6.

Effect of autorepression on the stability of the LEE1 promoter. Images of gfp expression via the PLEE1 by EPEC ler::kan mutant bacteria producing Ler or Ler(L29R) are shown. The level of gfp expression within individual bacteria (n = 600) was determined. The distribution of gfp expression levels is plotted below each image for EPEC ler::kan/pIR1Ler (black bars) and EPEC ler::kan/pIR1Ler(L29R) (gray bars). The calculated coefficient of variation (CV; standard deviation divided by the mean) is indicated.

DISCUSSION

The LEE is common to enteropathogens that cause attaching and effacing histopathology (21). It encodes the TTSS, which is essential for colonization of the host intestine. The LEE is composed of 41 genes organized in about 11 transcriptional units. These include the LEE1, LEE2, LEE3, LEE4, and LEE5 operons and the espG and map transcriptional units (8, 24). Concerted regulation of these operons is essential for successful colonization of the host. The LEE1 operon encodes nine genes: ler, orf2, cesAB, orf4, orf5, escR, escS, escT, and escU. Ler, the product of the first gene, positively regulates the transcription of LEE2, LEE3, LEE4, LEE5, espG, and map. The other genes in the LEE1 operon encode proteins that are required for the assembly of functional TTSS (5, 7). Thus, synchronized and well-timed activity of the LEE1 promoter is important to determine the expression of other LEE operons and is essential for efficient assembly of the TTSS and host colonization.

The temporal regulation of the LEE1 operon involves a complex interplay between many factors. Expression is positively regulated by IHF, Fis, PerC, BipA, GrlA, and a quorum-sensing factor and is negatively regulated by H-NS, GrlR, and GadX (7, 11, 13, 14, 17, 19, 26, 28, 32). GadX regulates LEE gene expression indirectly by affecting PerC expression (26). The IHF and H-NS regulation of the LEE1 promoter correlates with direct interaction of these proteins with the LEE1 regulatory region (11, 32, 34). We predict that at least some of the other LEE regulators, including PerC, GrlA, GrlR, Fis, BipA, or quorum-sensing factor, directly regulate the LEE1 promoter by binding to the LEE1 regulatory region. Thus, complex protein-protein and protein-DNA interactions are required for LEE1 regulation. This complexity hinders the setting up of an in vitro transcription system that will soundly mimic the in vivo situation.

In the present study, we did not investigate the temporal regulation of LEE1 but did investigate the mechanism that determines the steady-state level of Ler. Our results suggest that Ler autorepression is required to limit the Ler concentration to a level that is just sufficient for activation of the other LEE genes. Negative autoregulation is common in E. coli and has been proposed to allow a rapid response to the changing environment while maintaining relatively low steady-state levels of the regulator (22). Negative autoregulation is also required to enhance expression stability (1). Indeed, we demonstrated that Ler autoregulation significantly reduces the cell-to-cell variability in the PLEE1 activity.

In the absence of Ler, the LEE genes are repressed by H-NS (2, 15, 24, 32). Upon its expression, Ler negates the H-NS repression to allow expression of other LEE operons (2, 32). In the case of the LEE2-LEE3 and LEE5 promoters, the mechanism of promoter activation by Ler involves its specific binding to the DNA region of >100 bp, which contains defined boundaries. The Ler binding regions are located about 70 to 100 bp upstream (LEE2 and LEE5) or downstream (LEE3) of the promoters (15, 30). We show that autorepression by Ler involves its interaction with the LEE1 promoter region. Using a DNase protection assay, we found that Ler protects an extended region of >100 bp that overlaps with the IHF binding site and the promoter region. We could not, however, identify defined boundaries to this region. It is possible that Ler represses the LEE1 promoter by interfering with the RNA polymerase-promoter interaction and/or with the IHF binding, which is required for LEE1 expression.

Binding of Ler to the LEE1 and the LEE2-LEE3 regulatory regions exhibits an ultrasensitive sigmoidal curve, indicating a cooperative mode of binding. Thus, when the cellular [Ler] is close to the K_d, small fluctuations in concentration can strongly affect the Ler-DNA interaction and, thus, Ler-mediated repression and activation of the promoters. We found that the affinity of Ler to the LEE2-LEE3 region is about 1.4-fold higher than its affinity to the LEE1 region. Thus, we think that this higher affinity reflects the need to maintain Ler concentrations that are low but sufficient for activation of the LEE2, LEE3, and other LEE operons.

Ler was reported to directly or indirectly repress the production of non LEE-encoded fimbriae (9), and we also recently identified another non-LEE operon that is negatively regulated by Ler. Given our results with negative autoregulation, it is conceivable that Ler acts directly to repress several operons in the EPEC genome.

In conclusion, Ler's negative autoregulation prevents the overshooting of Ler expression upon its activation and stabilizes its expression in the population. These results suggest that during infection of the human intestine, the bacteria uniformly express the optimal levels of TTSS to allow efficient colonization.