Characterization of BNT2, an intrinsically curved DNA of Escherichia coli O157:H7

Abstract

The gene regulation by intrinsically curved DNA is one way for bacterial sensing of and response to environmental changes. Previously, we showed that the genetic element BNT2 upstream of the ecf (eae-positive conserved fragment) operon in the Escherichia coli O157:H7 virulence plasmid (pO157) has characteristics typical of intrinsically curved DNA, including the presence of multi-homopolymeric adenine:thymine tracts (AT tracts) and electrophoretic anomaly at 4°C. Here we report that a local intrinsic curvature induced by the two phased AT tracts within the unusual promoter sequence of BNT2 played a major role for its temperature-dependent promoter activity. The base substitution of the AT tract in the spacer DNA between the -35 and the unusual -10 regions of the BNT2 promoter with non-AT tract sequence reduced intrinsic curvature slightly at 4°C, but greatly affected its transcriptional activity. This implies that such a local intrinsic curvature within the unusual promoter of BNT2 is important for thermoregulation of the ecf operon.

Introduction

Escherichia coli O157:H7 is an important food-borne bacterial pathogen that causes hemorrhagic colitis and hemolytic uremic syndrome in humans and animals [1]. Interestingly, the bacterium is well known to be hardy in diverse in vivo and ex vivo environments. For example, E. coli O157:H7 can survive passage through the extremely acidic gastric stomach (pH<1.5) [2; 3], enhance virulence expression at the gastrointestinal milieu (i.e. induced expression of the genes in the LEE pathogenicity island) [4; 5], and persist in various environmental sources such as soil [6], raw manure [7], and farm water [8]. Therefore, it is likely that coordinated regulation of a set of E. coli O157:H7 genes might be required for this environmental flexibility although the underlying mechanisms are not clearly understood.

One of the environmental sensing and response mechanisms in cells is the gene regulation by intrinsically curved DNA, previously defined as DNA segments with a curved trajectory in its helix axis [9]. Intact nucleotide sequences such as short runs of homopolymeric adenine:thymine residues (AT tracts) is known to induce the intrinsic curvature of DNA especially when inserted in phase with the helical periodicity of 10 to 11-bp interval [9; 10]. Such changes in intrinsic DNA topology within or upstream of the promoter sequence may promote or inhibit binding of RNA polymerase (RNAP) and/or other DNA binding proteins (i.e. a nucleoid-associated protein H-NS, stationary sigma factor σ^s, or regulatory proteins HU, IHF, and FIS etc.) to the curved region of DNA [11; 12], which in turns modulates gene expression at the level of transcription. Interestingly, some environmental factors including temperature, pH, osmolarity, ionic strength such as [Mg²⁺], and anaerobiosis can affect gene expression by intrinsically curved DNA [13; 14; 15], suggesting that intrinsically curved DNA coordinates environmental signals with gene expression.

Some virulence-associated genes in bacterial pathogens are known to be controlled by intrinsic DNA curvature in response to certain environmental conditions. Such examples include transcriptional activation of type IV bundle forming pili in enteropathogenic E. coli [16; 17; 18] and VirF in Shigella flexneri and Yersinia enterocolitica [19; 20] during a temperature shift from 24 to 37°C. Moreover, a recent study reported that intrinsic curvature signals of DNA are highly conserved in putative regulatory regions of bacterial genomes, supporting their function as a global regulatory mechanism for prokaryotic genes [21].

Previously, we isolated BNT2 intrinsically curved DNA from the E. coli O157:H7 virulence plasmid (pO157) [22]. BNT2 has characteristics typical of intrinsically curved DNA such as electrophoretic anomaly at 4°C, six partially phased AT tracts, and the temperature-dependent transcription of the pO157-encoded ecf (eae-positive conserved fragments, ecf1 to 4) operon [22] whose gene products are involved in the structural modification of the outer membrane lipopolysaccharide (LPS) [22; 23; 24; 25]. For example, ecf1 and ecf2 encodes a putative polysaccharide deacetylase and a functional LPS α-1, 7-N-acetylglucosamine transferase, respectively and both are unique to pO157 [26]. ecf3 shows similarity to a putative outer membrane protein in E. coli K1 associated with bacterial invasion [27]. ecf4 encodes the second copy of a lipid A myristoyl transferase [22; 25]. The double mutant carrying deletions in the ecf4 and its chromosomal copy lpxM of E. coli O157:H7 had an altered lipid A structure based on growth temperature and membrane fatty acid composition, and showed decreased persistence in the bovine gastrointestinal tracts [22; 24; 26].

Previous biochemical characterization of BNT2 identified a functional promoter sequence with an unusual -10 (AAAAAT) element. Interestingly, two AT tracts are positioned in phase between this unusual -10 element and the spacer DNA in the promoter sequence of BNT2, suggesting the possibility of an AT tract-induced intrinsic curvature at this defined region [22]. In this study, therefore, we hypothesized that the AT tract-induced intrinsic curvature between the spacer DNA and the unusual -10 element of BNT2 contributes to the promoter activity of BNT2 that depends on temperature. To test this hypothesis, the AT tract in the spacer DNA was replaced with non-AT tract sequence by site-directed mutagenesis and its biological significance was examined by measuring both intrinsic DNA curvature and the promoter activity at the different temperatures.

Materials and Methods

Bacterial strains, plasmids, and BNT2 nucleotide sequence

Bacterial strains and plasmids used in this study are listed in Table 1. All bacterial strains were grown and maintained in Luria-Bertani (LB) media (MP Biomedicals) with or without 1.5% (w/v) agar. If necessary, the antibiotics were used at the following concentrations: ampicillin (Ap), 100 μg/ml and kanamycin (Km), 50 μg/ml.

Table 1. Bacterial strains and plasmids in this study.

Strain / plasmid	Description	Reference
E. coli strains:
DH5a	A general cloning host	RBCa
ATCC 43894	E. coli O157:H7 (a human clinical isolate, stx1+/stx2+)	ATCCb
YH2021	ATCC 43894 with a single copy BNT2∷lacZ	This study
Plasmids:
pSKBNT2	1,041-bp BNT2 fragments in pBluescript SK+	[22]
pRBC/BNT2	BNT2 fragments in RBC TA cloning vector	This study
pRBC/AT24M	AT24M in RBC TA cloning vector	This study
pRBC/BNT2-134	a 134-bp region of BNT2 in RBC TA cloning vector	This study
pRBC/AT24M-134	a 134-bp region of AT24M in RBC TA cloning vector	This study
pRS551	a lacZYA' operon fusion vector (pBR322 origin, ApR KmR)	[31]
pRSBNT2	BNT2 in pRS551 (ApR KmR)	[22]
pSP417	A lacZYA' operon fusion vector (pBR322 origin, ApR)	[34]
pSPBNT2	BNT2 in pSP417 (ApR)	[22]
pSPAT24M	AT24M in pSP417 (ApR)	This study

^aReal Biotech Corporation, Taiwan

^bAmerican Type Culture Collection, Manassas, VA

The DNA nucleotide sequence of BNT2 was obtained from the GenBank database (http://www.ncbi.nlm.nih.gov/sites/entrez?db=nucleotide) available at National Center for Biotechnology Information (NCBI) with the accession number AY330225.

General procedures

General molecular techniques including polymerase chain reaction (PCR), gene cloning, plasmid purification, and transformation were carried out as previously described [28]. All the enzyme reactions were followed by the manufacturers' recommendation and DNA sequencing analyses were done by Macrogen Inc. (Seoul, Korea), which was commercially available.

Site-directed mutagenesis by inverse PCR

A 1,041-bp BNT2 fragment from pSKBNT2 [22] was cloned into the RBC TA cloning vector (Real Biotech Corporation). The resultant plasmid was confirmed by DNA nucleotide sequencing and designated pRBC/BNT2 (Table 1).

To disrupt a local intrinsic curvature within the unusual promoter sequence of BNT2, the AT tract (AAAAAT) in the spacer DNA of BNT2 (near +960-bp position) was substituted with non-AT tract sequences (TCG) by introducing the ClaI restriction site, yielding AT24M. Briefly, the ClaI site-incorporated primers, J11 (5′-TGATATCGATATTAAAAAATCCCTGTTGTCGG-3′) and J12 (5′-ATATCGATATCAGTTAATCAACCAGAACT-3′), were used for inverse PCR with the template plasmid pRBC/BNT2 (Table 1). The amplified PCR products were purified from the agarose gel using HiYield Gel/PCR DNA extraction kit (Real Biotech Corporation), digested with ClaI (New England Biolabs), and self-ligated with T4 ligase (New England Biolabs). The resultant plasmid was confirmed by DNA nucleotide sequencing and designated pRBC/AT24M (Table 1).

The 134-bp promoter regions at the 3′-end of BNT2 or AT24M (corresponding to the nucleotide positions from +907 to +1,041-bp) were obtained by PCR using the template plasmids, pRBC/BNT2 or pRBC/AT24M, with the primers J86 (5′-GGGGAATTCGAGCATATGCGTTAAGGGAA-3′) and M13R (5′-GGAAACAGCTATGACCATG-3′) and subcloned into the RBC TA cloning vector. The resultant constructs were confirmed by DNA sequencing and designated pRBC/BNT2-134 or pRBC/AT24M-134, respectively (Table 1).

In silico prediction of intrinsic DNA curvature

The bend.it server (http://hydra.icgeb.trieste.it/dna/bend_it.html) was used to calculate the intrinsic curvature of BNT2 or AT24M DNA molecules as predicted from their DNA nucleotide sequences. This prediction was based on values tabulated for di- or tri-nucleotides and the intrinsic curvature was calculated using standard algorisms [29] with a window size of 31-bp. For visualizing the 3-dimensional structures of BNT2 and AT24M, the model.it server (http://hydra.icgeb.trieste.it/dna/model_it.html) was used as followed by the server's direction.

Experimental detection of intrinsic DNA curvature

To analyze the electrophoretic gel retardation of intrinsically curved DNA at 4°C, the gel-purified DNA fragments were subject to one-dimensional (1-D) polyacrylamide gel electrohoresis (PAGE) at 60 or 4°C on 5% non-denatured polyacrylamide gels. Based on the molecular marker, the expected sizes of migrated DNA fragments at both temperatures were calculated and compared using Microsoft Excel program (version 2003, Microsoft Corporation).

For more sensitive detection of intrinsic DNA curvature, two-dimensional (2-D) PAGE was used as previously described [20]. Briefly, one microgram of pRBC/BNT2-134 or pRBC/AT24M-134 was digested completely with HindIII at 37°C, mixed with 1 ug of 100-bp DNA ladder (Invitrogen), and separated on 10 or 15% non-denaturing polyacrylamide gels. After the first dimensional electrophoresis at 60°C, the glass-gel sandwich was rotated and transferred to pre-chilled 1×TBE buffer (45 mM Tris-borate, 1mM EDTA, pH8.4) and subject to the second dimensional electrophoresis at 4°C. The gel was stained with ethidium bromide (EtBr) and visualized using the Core-Bio iMAX image analysis system (Core-Bio Inc.).

Determination of transcriptional activity

To avoid possible biases from heterogeneous strains or plasmid copy-number effects, a single-copy BNT2∷lacZ transcriptional fusion system was constructed on the chromosome of E. coli O157:H7 ATCC 43894 by allelic exchange as previously described [30]. Briefly, BNT2 was ligated into EcoRI and BamHI sites of pRS551, a multi-copy lacZ transcriptional fusion plasmid [31], resulting in pRSBNT2. The 6.0-kb ScaI-SacI fragment of pRSBNT2 was ligated with 8.7-kb EcoRV-SacI fragment of the conjugative suicide vector, pJRlacZins [32]. After transformation into E. coli S17-1 λpir, the resulting construct was transferred to the wildtype E. coli O157:H7 strain by conjugation and recombined into intact lacZ gene on the E. coli chromosome. The E. coli O157-positive, Ap-resistant (Ap^R) and Km^R exconjugants were counter-selected under presence of 6% sucrose and finally, the Km^R/Ap^S colonies were confirmed by PCR.

β-galactosidase activity

Transcriptional fusion strains were grown in LB containing appropriate antibiotics with or without 0.4% glucose at various temperatures such as 10, 15, 24, 37, and 45°C with shaking at 150 rpm to an optical density at 600 nm (OD₆₀₀) of 0.5. The β-galactosidase activity was measured with 0.2% sodium dodecyl sulfate/chloroform permeabilized cells as previously described [33]. At least, three independent experiments were performed and the data were analyzed statistically by the student t-test using Microcal Origin™ (version 8.0, Microcal Software, Inc.).

Construction of a multi-copy AT24M∷lacZ transcriptional fusion

The transcriptional fusion vector pSP417 carrying a promoterless lacZ gene [34] was used to analyze the transcriptional activity of BNT2 or AT24M. The 1,041-bp BNT2 or AT24M fragments were restricted by both EcoRI and BamHI and ligated into the EcoRI and BamHI sites of pSP417, resulting in pSPBNT2 and pSPAT24M (Table 1). The resultant plasmids were transformed into E. coli strains and screened for appropriate antibiotic resistance.

Results and Discussion

Temperature dependency of the transcriptional activity of BNT2 intrinsically curved DNA

A typical characteristic of intrinsically curved DNA is electrophoretic gel retardation at low temperatures, which is due to temperature-dependent changes in DNA topology. Consistent with our previous observation [22], BNT2 intrinsically curved DNA showed slowed migration at 4°C compared to 60°C (Fig. 1A). Indeed, the 1.0-kb sized fragments of BNT2 at 60°C migrated like the 1.6-kb sized fragments at 4°C (Fig. 1A). Since BNT2 is known to have a functional promoter whose activity relies on growth temperatures [22], temperature dependency of the BNT2-driven transcriptional activity was investigated at various temperatures ranging from 10 to 45°C (Fig. 1B). Our results demonstrated that BNT2-driven transcription was significantly high at low temperatures such as 10, 15, and 24°C with a peak activity at 24°C, compared to at high temperatures 37 and 45 °C (p<0.05, Fig. 1B). Indeed, approximately 3.1-fold higher activity in transcription was observed at low temperatures (10, 15, and 24°C) compared to at high temperatures (37 and 45 °C), regardless of the presence of glucose (Fig. 1B). These results imply that DNA relaxation at high temperatures above 37°C may affect the BNT2-driven transcriptional activity which depends on temperature. Supporting this notion, a previous study demonstrated that the intrinsic DNA curvature in the Y. enterocolitica virulence plasmid (pYV) was relaxed at between 30 and 37°C [20].

Fig. 1. Electrophoretic anomaly of BNT2 at 4°C.

(A) One-dimensional polyacrylamide gel electrophoresis of BNT2. Each purified DNA was separated on 5 % polyacrylamide gel at 60°C or 4°C to examine intrinsic curvature. (B) Temperature dependency of the BNT2-driven transcription. The strain (YH2021) carrying a single-copy chromosomal BNT2∷lacZ transcriptional fusion construct in ATCC 43894 was grown in LB with (filled bar) or without 0.4% glucose (open bar) at various temperatures indicated and β-galactosidase activity was measured. The enzyme activities are represented as the means ± standard deviation from at least three independent experiments.

Site-directed mutagenesis of the AT-tract in the spacer DNA of the unusual promoter sequence of BNT2

The previous genetic mapping of BNT2 has identified the six partially phased AT tracts located upstream of the transcriptional start site (a guanine residue in Fig. 2) of the E. coli O157:H7 pO157 ecf operon (Fig. 2). Among those six AT tracts, two AT tracts were positioned in phase with a 10-bp periodicity and localized within the functional, but unusual promoter sequence of BNT2 (Fig. 2). To examine if a possible intrinsic curvature induced by those two phased AT tracts within the unusual promoter sequence of BNT2 contributes to the BNT2-driven promoter activity responding to temperature, we targeted and replaced the AT tract in the spacer DNA with non-AT tract sequence by site-directed mutagenesis (Fig. 2). Defined mutation was confirmed by DNA nucleotide sequencing (data not shown). The resultant DNA fragment, AT24M, had intact -35 and -10 elements of BNT2, but contained non-AT tract sequence in the spacer DNA region unlike in BNT2 (Fig. 2).

Fig. 2. Site-directed mutagenesis of the AT tract in the spacer DNA of the unusual promoter sequence of BNT2.

The schematic diagram of the BNT2 containing region in pO157 is shown. The nucleotide base position in pO157 is indicated at the top left and right. The black boxes indicate the AT tracts and the arrow shows the transcriptional start site (+1 and the asterisk). The capitalized letter P in a circle represents the functional, but unusual promoter sequence of BNT2. The AT tract in the spacer DNA region between the -35 element and the unusual -10 element of BNT2 was replaced with non-AT tract sequences by PCR mutagenesis (see Materials and Methods). The resultant mutation was directly confirmed by DNA nucleotide sequencing.

In silico modeling of BNT2 and AT24M

To investigate whether or not the base substitution of the AT tract in the spacer DNA with non-AT tract (Fig. 2) could disrupt a possible local intrinsic curvature, the nucleotide sequences from BNT2 or AT24M were predicted for the presence of intrinsic curvature (see Materials and Methods). As expected, the results revealed that a predicted intrinsic curvature in the 134-bp promoter region at the 3′-end of BNT2 became reduced at the corresponding region at AT24M where the defined base substitution was introduced (Fig. 3A). Consistently, in silico modeling of both BNT2 and AT24M indicated that the 134-bp promoter region of AT24M was relaxed by approximately 16 degrees compared to the corresponding region of BNT2 (Fig. 3B).

Fig. 3. Prediction of intrinsic DNA curvature (A) and in silico modeling (B) of BNT2 and AT24M.

Individual nucleotide sequence was analyzed for intrinsic curvature using the bend.it server. The 134-bp promoter regions containing the two phased AT tracts from BNT2 and AT24M were reconstituted by the model.it server. The arrow indicates the defined spacer DNA regions between the -35 element and the unusual -10 element of BNT2 or AT24M (panel A). In the panel B, changes in intrinsic DNA curvature were compared by measuring the angles produced by the two artificial lines (dotted lines).

Biological significance of the AT tract-induced intrinsic curvature within the unusual promoter sequence of BNT2

The predicted intrinsic DNA curvature present in BNT2 and AT24M, especially within their promoter regions, was confirmed by 2-D PAGE (Fig. 4A). The HindIII-restricted DNA fragments containing the 134-bp promoter regions from the 3′-end of BNT2 or AT24M were purified from pRBC/BNT2-134 or pRBC/AT24M-134, respectively, and analyzed on both 10 and 15% non-denatured polyacrylamide gels (Fig. 4A). Consistent with our in silico modeling, the restricted fragments of BNT2 showed a slightly retarded migration at 4°C compared to those of AT24M on both 10 and 15% non-denatured gels (Fig. 4A).

Fig. 4.

Involvement of a local intrinsic curvature within the unusual promoter sequence of BNT2 in its temperature-dependent transcription. (A) The HindIII-restricted fragments containing the 134-bp BNT2 or AT24M promoter regions were subjected to 2-D PAGE as indicated with 1.0 ug of the 100-bp ladder DNA marker (Gibco BRL). The 2-D gels were visualized with UV light after EtBr staining. The arrow indicates the DNA fragments containing the 134-bp BNT2 or AT24M promoter regions. The dash lines showed the migration status of the 200-bp DNA marker band on the 2-D gels and used to compare to the migration status of the DNA fragments containing the 134-bp BNT2 or AT24M promoter regions. (B) Temperature-dependent transcriptional activity of BNT2 and AT24M. The multi-copy transcriptional fusion plasmids of BNT2 (pSPBNT2) or AT24M (pSPAT24M) were constructed on a promoterless operon fusion vector (pSP471) and transformed into E. coli DH5α. The lacZ expression was measured as β-galactosidase activity after grown in LB at 37 or 24°C. The enzyme activities are represented as the means ± standard deviation from at least three independent experiments.

Can such a subtle change in DNA topology of BNT2 affect the BNT2-driven promoter activity responding to temperature? To address this question, the transcriptional activity of BNT2 and AT24M was examined at 37 and 24°C. As shown in Fig. 4B, surprisingly, the significant changes in transcription were observed between BNT2 and AT24M at both temperatures (p<0.05, Fig. 4B). At 24°C, the BNT2-driven transcription was about 4.6 folds higher than that at 37°C, while the AT24M-driven transcription was about 2.1 folds higher than that at 37°C (Fig. 4B). Although a simple base substitution of the AT tract in the spacer DNA did not fully turned off the promoter activity of BNT2, it significantly affected the BNT2-driven transcription as well as its temperature dependency. Taken together, theses results imply that a local intrinsic curvature induced by the two phased AT-tracts within the unusual promoter sequence of BNT2 are important for the temperature-dependent transcription of the pO157-encoded ecf operon.

In conclusion, bacterial coordination of environmental signals with gene expression is most important for their survival and persistence in various environmental conditions as well as contributes to an established infection in vivo. Not surprisingly, a recent study has reported that DNA curvature signals are conserved in regulatory regions of a cluster of prokaryotic genes such as proteins HU and IHF, DNA gyrase subunits A and B, FIS protein-coding genes, transposases, cell division-related genes, and flagellum genes [21]. Our data suggest that the gene regulation through intrinsic DNA curvature especially within or near the promoter sequences seems to be very sensitive to environmental changes because only a subtle change in intrinsic DNA topology within the unusual promoter sequence of BNT2 greatly reduced its transcription. It would be interested to elucidate the mechanisms behind this regulation as well as to identify other in vivo or ex vivo environmental signals which can modulate the BNT2-driven promoter activity.