Identification of the target DNA sequence and characterization of DNA binding features of HlyU, and suggestion of a redox switch for hlyA expression in the human pathogen Vibrio cholerae from in silico studies

HlyU, a transcriptional regulator common in many Vibrio species, activates the hemolysin gene hlyA in Vibrio cholerae, the rtxA1 operon in Vibrio vulnificus and the genes of plp-vah1 and rtxACHBDE gene clusters in Vibrio anguillarum. The protein is also proposed to be a potential global virulence regulator for V. cholerae and V. vulnificus. Mechanisms of gene control by HlyU in V. vulnificus and V. anguillarum are reported. However, detailed elucidation of the interaction of HlyU in V. cholerae with its target DNA at the molecular level is not available. Here we report a 17-bp imperfect palindrome sequence, 5′-TAATTCAGACTAAATTA-3′, 173 bp upstream of hlyA promoter, as the binding site of HlyU. This winged helix-turn-helix protein binds necessarily as a dimer with the recognition helices contacting the major grooves and the β-sheet wings, the minor grooves. Such interactions enhance hlyA promoter activity in vivo. Mutations affecting dimerization as well as those in the DNA–protein interface hamper DNA binding and transcription regulation. Molecular dynamic simulations show hydrogen bonding patterns involving residues at the mutation sites and confirmed their importance in DNA binding. On binding to HlyU, DNA deviates by ∼68º from linearity. Dynamics also suggest a possible redox control in HlyU.


INTRODUCTION
The pathogenesis of cholera, caused by Vibrio cholerae, is a complex process involving a number of factors accounting for the establishment of the pathogen in the epithelium of the small intestine and production of the enterotoxin that disrupts ion transport by intestinal epithelial cells. The major features of the pathogenesis of cholera are well established, with toxin coregulated pilus (TCP), cholera toxin (CTX), accessory cholera enterotoxin (ACE), zona occludens toxin (ZOT) and hemolysin (HlyA) having been identified as the key virulence factors. HlyA, a cytotoxin as well as an enterotoxin (1)(2)(3)(4), has been implicated as a virulence determinant in the infant mouse cholera model (4). Numerous investigations proved that the hemolytic (Hly+) V. cholerae strains were of the El Tor biotype, whereas classical biotype strains were always found to be non-hemolytic (Hly−). In the recent past, the proportion of hemolytic V. cholerae isolates in Africa has increased dramatically (5), and the series of outbreaks of cholera along the U.S. Gulf Coast beginning in 1973 was due to a strongly hemolytic El Tor strain of V. cholerae (6,7). One of the possible pathogenicity factors of the V. cholerae strains that are not able to produce CTX but still cause gastrointestinal disorders (8,9) has been implicated to be the El Tor HlyA (1,10).
Besides the much discussed virulence regulatory systems, like ToxR, Fur, etc., in V. cholerae, the HlyU regulatory system has been found to be involved in hlyA gene expression and probably expression of other virulence factors that are required for the efficient colonization of the host small intestine by this organism (11,12). A hlyU mutant has been found to be deficient in the production of HlyA and Hcp (a 28 kDa secreted HlyA co-regulated protein, a part of the Type 6 secretion system) (13), as well as defective in colonization, with a 100-fold increase in the LD 50 in the infant mouse cholera model (12). HlyU of V. cholerae, (HlyU Vc) is known to belong to the SmtB/ArsR family of regulatory proteins including NolR of Rhizobium meliloti, SmtB of Synechococcus sp., ArsR of Staphylococcus aureus (14). The homology model of the protein and the very recently solved crystal structure (Protein Data Bank (PDB) ID: 4OOI, (15)) show that the protein is a homodimer with a winged helix-turn-helix (wHTH) motif (Supplementary Figure S1). Unlike the other metallorepressor members of its family, HlyU Vc is a transcription activator (16) lacking any metal binding domain (15). Though Williams and Manning demonstrated that a 710bp fragment spanning the predicted promoter in the 5 region of hlyA in conjunction with HlyU Vc increased chloramphenicol acetyl transferase activity and HlyA production in Escherichia coli (16), no further findings regarding the DNA-HlyU Vc interactions have been documented in the literature to date. Here we report studies pertaining to direct biochemical evidence of HlyU Vc binding to the region upstream of hlyA promoter region using gel shift assays and in vivo reporter gene assays. We have revealed the nucleotide bases interacting with HlyU Vc to specificity using DNase I footprinting studies. We have also endeavored to dissect the protein structure-function relation through mutational approaches and identify the actual residues of HlyU Vc that are essential for dimerization and DNAbinding activity. The involvement of both the major and the minor grooves of the double helix and the probable conformational changes of the nucleic acid in the DNA-HlyU Vc complex have also been reported. Additionally, we have carried out molecular dynamic (MD) simulations on a DNA-HlyU Vc model. The simulations validate the solution studies as well as suggest a possible redox role of the Cys residues in DNA binding by HlyU Vc.

Bacterial strains, plasmids, radioisotopes and other chemicals
V. cholerae El Tor strain N16961 was a gift from Dr R. Nandy (National Institute of Cholera and Enteric Diseases, Kolkata). The specialized cloning vector pBend4 was provided by Dr S. Dasgupta (Bose Institute, Kolkata). All the bacterial strains, plasmids and clones used in the study are provided in Supplementary Table S1. [␥ 32 P]-ATP was obtained from Board of Radiation and Isotope Technology, India (1 milli Curie ml −1 ). X-ray films and film developing chemicals were purchased from Kodak. ortho-Nitrophenol galactoside (ONPG) is from Calbiochem. All enzymes are from Thermo Scientific. All other chemicals were from Merck and of molecular biology grade.

Designing of DNA fragments for electrophoretic mobility shift assay
About −582 to +246-bp DNA stretch spanning the hlyA transcription start site was considered for DNA binding studies. This region was initially divided into four overlapping DNA fragments (Table 1), and the corresponding forward and reverse primer sequences were designed according to the published genomic DNA sequence of V. cholerae N16961. Later, other fragments, DNA 5 and DNA 6 ( Table  1), were generated as stated above.

Purification of HlyU Vc
For over-expression of the proteins--wild-type HlyU Vc and the mutants--competent cells of E. coli strain BL21 (DE3) were transformed with the corresponding constructs. One colony from each was transferred from the plate into 50 ml Luria-Bertani (LB) broth containing 50 g ml −1 kanamycin for overnight growth at 37 • C. Over-expression of all the proteins was achieved by growing cells in LB broth at 37 • C to an OD 590 of ∼0.5 followed by induction with 0.5 mM isopropyl-␤-thiogalactopyranoside (IPTG) at 37 o C for 4 h. The cells were lysed under native conditions, and the protein was purified by using the nickel-nitrilotriacetic acid (Ni-NTA) resin (Qiagen) for affinity chromatography. The N-terminal 6X-His tag was removed using thrombin. The digestion was stopped by adding DNase I stop solution (50 mM Tris (pH 8.0), 50 mM EDTA, 2% (wt/vol) sodium dodecyl sulphate (SDS), 0.4 mg ml −1 proteinase K). DNA was then precipitated using ethanol and sodium acetate and then washed with ethanol and dried. Digested DNA fragments were resuspended in loading buffer (98% vol/vol) deionized formamide, 10 mM EDTA, 0.025% wt/vol xylene cyanol and 0.025% wt/vol bromophenol blue, boiled for 5 min, chilled rapidly and separated by gel electrophoresis on 8% (wt/vol) Urea-TBE polyacrylamide sequencing gel at 1300 V for 3.5-4 h. An A+G (using formic acid and piperidine) and a G ladder (using dimethyl sulphate and piperidine) were prepared with 10 nM of labeled DNA and analyzed along with digested DNA. The gels were imaged using a phosphor imager (Typhoon trio +).

Site-directed mutagenesis
All the desired mutations were carried out using QuikChange II Site-directed mutagenesis kit (Stratagene). The plasmid clone for hlyU Vc gene in pET28a or pET23a vector (pET28a-HlyU Vc, pET23a-HlyU Vc; Supplementary Table S1) was used as the template for site-directed mutagenesis PCR using the corresponding primers. The wild-type templates in the PCR product mix were digested away with DpnI and the final product was transformed into E. coli XLIB cells. The plasmids for the mutant proteins (pET28a-XnY, pET23a-XnY; Supplementary Table S1) were extracted from the screened colonies and sequenced for the correct mutations.

Size exclusion chromatography
Analytical gel filtration experiments were carried out in an HPLC system AKTA prime Plus. 500 g protein was injected at a time. The column was pre-equilibrated with 50 mM sodium phosphate and 200 mM sodium chloride buffer, pH 6.8. Bovine serum albumin (66.5 kDa), RNase A (13.7 kDa), chymotrypsin (24.8 kDa) and ovalbumin (44.5 kDa) were used as molecular weight markers. The void volume was calculated by running blue Dextran. The molecular weight of the sample proteins was calculated from the standard plot of R f versus molecular weight, M w , generated from the marker proteins.
␤-galactosidase assay E. coli BL21 (DE3) cells were transformed with the two plasmids: one bearing the hlyU Vc gene under the control of T7 promoter (pET23a-HlyU Vc) with Amp R and the other, pDA1a containing DNA 7 ( Table 1) Table S1). The double antibiotic resistant colonies were cultured overnight at 37 • C in LB broth. Similar transformations were done with pDA1a and each of the pET23a-XnY plasmids (Supplementary Table S1) encoding L25D, L98D, Q63A and Y85A mutant proteins. A control experiment was also carried out with E. coli cells carrying the empty vector pET23a and pDA1a. Secondary cultures were grown for 5 h from the overnight cultures and were induced for 1 h with 0.5 mM IPTG. The ␤-galactosidase (␤-gal) activity of aliquots from the culture was determined by the method of Miller (17). The OD 590 of the cultures were noted. 100 l cultures were removed in Z-buffer and lysed with SDS and chloroform. 200 l ONPG was added and reactions were allowed to continue till development of yellow color. Reactions were stopped with 300 mM sodium carbonate and the color of released ortho-nitrophenol (ONP) and scattering were noted at 420 and 550 nm, respectively.
For statistical analysis, all data reported are the arithmetic mean (±SD) from five independent experiments performed in triplicate. The results were analyzed by one-way analysis of variance (ANOVA) followed by post hoc Tukey's Test accepting P < 0.001 as a level of significance. Data analyses were performed using the Prism software (Graph-Pad, San Diego, CA, USA).

Circular Permutation Assay
DNA 6 was cloned at the XbaI site of the specialized vector pBend4 to get the clone pBendDA (Supplementary Table S1) and six DNA fragments for the circular permutation assay were isolated from pBendDA by digestion with BglII, XhoI, EcoRV, NcoI, SmaI and NruI, respectively. These fragments were end labeled using T4 polynucleotide kinase in the presence of [␥ 32 P]-ATP using standard procedures. Gel shift assays were performed as described above; the DNA-protein complexes were separated on 8% (wt/vol) polyacrylamide gels run at 200 V for 6 h at 4 • C. The magnitudes of the apparent bending were calculated from the variations in the mobilities of the DNA-protein complexes in the circular permutation analysis using the formula where R L (relative mobility) is defined as the ratio of the minimally (r min ) and maximally (r max ) retarded bands. 'r' is the ratio of μ bound and μ free . Mobility (μ) is in turn defined as the distance migrated from the origin. α is the angle by which the DNA is bent from linearity (18).

DNA-protein docking and MD simulations
The PDB coordinates of the B-DNA with the sequence as found by DNase I footprinting were generated using 3D-DART web server (19). The DNA models were analyzed using Curves+ (20). The DNA and the crystal structure of HlyU Vc (PDB ID: 4OOI, (15)) docking was done using Haddock server: easy interface (21). The experimentally determined active residues for both the protein and the DNA were mentioned in the submitted job. The best docked model of DNA-HlyU Vc complex with maximum Haddock score was chosen for MD studies. All simulations and equilibration were performed using the sander module of AMBER 10.0 software package at 300 K (22). The structure was solvated in a cuboid periodic box of explicit water with water molecules extending 6Å outside the protein on all sides. The water molecules were described by the transferable intermolecular potential three point (TIP3P) model (23). The molecules were minimized using steepest descent for 500 cycles, followed by conjugate gradient method for 20 000 cycles prior to equilibration and dynamics run. The system was heated to 300 K within 40 ps and equilibrated following minimization. Bonds involving hydrogen were constrained with the help of the SHAKE algorithm. The production run was carried out for 20 ns, employing constant pressure periodic boundary conditions (24). A non-bonded cutoff distance was set to 12Å and 2 fs integration time step was used. The coordinates were saved after each 2 ps. The analyses of the trajectories were performed using the ptraj module of AMBER. The figures and movies were generated using Pymol (http://www. pymol.org/) and VMD (http://www.ks.uiuc.edu/Research/ vmd/), respectively. The bending of helices was computed using HELANAL-PLUS, an online server that is useful in characterizing the overall geometry of helices from its C␣ atom coordinates (25).

HlyU Vc binds to an imperfect palindrome about 164 bp upstream of hlyA transcription start site
As Williams and Manning showed a 710-bp DNA sequence upstream of hlyA gene in conjunction with HlyU Vc increases HlyA production (16), we scanned the region upstream of the hlyA gene for the precise delineation of the HlyU Vc binding site. Four DNA fragments (Table 1) from the region (−582 to +246 bp) spanning the hlyA promoter were amplified by PCR from the V. cholerae N16961 genomic DNA such that there is ∼50-bp overlap between consecutive fragments. EMSA revealed that HlyU Vc binds to both DNA 2 and DNA 3 ( Figure 1A). The results indicated that HlyU Vc binding can possibly occur by either of the two ways as modeled in Figure 1B. According to model I, both the DNA fragments can show shift if there are separate binding sites for HlyU Vc. To test this possibility, EMSA was performed with a 127-bp DNA fragment (DNA 5, Table 1); the fragment is a part of DNA 3 excluding ∼58 bp at the 5 end that has a sequence overlap with the 3 end of DNA 2. Since there was no complex formation with this fragment ( Figure 1C), we assessed the other possibility that HlyU Vc binds to the overlapping regions of DNA 2 and DNA 3 (model II). A 205-bp fragment (DNA 6, Table 1) with the overlap between DNA 2 and 3 residing at the center of the nucleotide string was considered. On performing gel shift assays, HlyU Vc protein was found to bind DNA 6 ( Figure 1C).
Subjecting the HlyU Vc binding stretch (DNA 6) to DNase I footprinting, a protection spread over 35 residues (−164 to −198 bp) was found (Figure 2A). An imperfect palindrome (5 -TAATTCAGACTAAATTA-3 ), −173 to −189 bp upstream of the hlyA transcription start site, on which the recognition helices of HlyU Vc sits, was identified--over the stretch of 17 bases between the complementary strands, there is only one mismatch on either side of the central position ( Figure 2B). This internal 17 residue palindrome can be extended by seven residues (shown in yellow color in Figure 2B) on either side--however, it introduces three more mismatches. This additional region may be important for interaction of the wing of wHTH motif. The binding site sequences found for HlyU Vv and HlyU Vc were aligned ( Figure 2C), which shows ∼71% base identity with common stretches of T and A. Next we performed EMSA with oligonucleotide sequences O1, O2 and O3 (Table 2). O1 is the 35-bp palindrome that binds HlyU Vc. The 5 end of O2 is the sequence from DNA 6 immediate upstream of O1; the sequence of O2 ends with the first half of O1--thus it retains one half of the binding stretch. O3 is an altered sequence that retains the second half of O1 while the first half is replaced by a G trail. HlyU Vc was found to bind only to O1, while the others showed no shift ( Figure 2D). Thus the individual half sites are incapable of binding and the entire palindrome is crucial in binding HlyU Vc.
We also checked the effect on DNA binding of a few nucleotides in the palindrome. dT (-190), dA(-187), dC(-180) and dT(-175) were mutated, and were found to have no affect (data not shown). Comparing the binding sites for HlyU in different Vibrio species, it is interesting to note that there is no strict consensus binding site of HlyU for the transcription of different genes (26,27). Considering HlyU as a global regulator (12), it can be expected that there would be flexibility in the binding sequence and, therefore, single mutations are less expected to affect binding.   (Table 2), O1 being the original palindrome, and O2 and O3 having each of the half sites of O1.

Dimerization: a prerequisite for DNA binding by HlyU Vc
In 2-fold symmetry-related protein subunits, a residue close to the 2-fold axis may interact with the same residue from the other subunit thus making up a pair of 'self-contacting' residues which may be important in defining the homodimeric interface (28). Analysis of the dimeric interface (using ProFace (29)) of the crystal structure of HlyU Vc revealed the presence of four self-contact residues: Leu25, Ala29, Met95 and Leu98 (Supplementary Figures S2 and  S3) (16). These hydrophobic residues were mutated to a charged residue, Asp, to induce charge-charge repulsion, consequently preventing protein dimerization. The ultimate target was to emphasize the importance of protein dimerization on DNA binding. Size exclusion chromatography of the proteins indicated that while the wild-type protein eluted entirely as dimer ( Figure 3A), L25D, A29D and M95D showed predominant monomer population ( Figure  3B-D). Interestingly, L98D mutation affected dimerization to a lesser extent with both the dimer and monomer being present in nearly equal amounts ( Figure 3E). In accordance to the oligomerization behavior, L25D, A29D and M95D did not bind DNA ( Figure 3B-D), whereas in the case of L98D, there was considerable DNA binding ( Figure 3E). At even higher concentrations of the proteins, L25D, A29D and M95D failed to bind DNA ( Supplementary Figure S4A). However, L98D showed a similar DNA binding profile as wild-type HlyU Vc in the same high range of protein concentrations. Among the four self-contact residues, Leu98 is the only one that lies at the rim of the dimeric interface (15); the others lie in the core region and therefore influence the dimeric association more strongly (30). Moreover, Leu98 makes fewer contacts than the others (Supplementary Figure S2), lying close to the protein exterior (Supplementary Figure S3). Therefore, a mutation in this residue is less likely to alter the oligomeric status of HlyU as compared to other self-contact mutants.

Amino acids essential for DNA-HlyU Vc binding
Based on the crystal structure of HlyU Vc (PDB ID: 4OOI, (15)), in silico model of HlyU Vc -DNA complex (14) and DNA binding studies on HlyU from Vibrio vulnificus (HlyU Vv) (31), we predicted some amino acids which may be essential for HlyU Vc's biological function. The amino acids chosen were Lys26 from the ␣1 helix, Asn30 from the loop between ␣1 and ␣2 and Arg32, and Arg33 from the ␣2, Ser62, Gln63, His64, Leu65, Ala66, Trp67, Arg69 and Arg70 from ␣4, Lys78 from ␤2, Gln81 from the turn, and Tyr85 from ␤3. In addition, we selected some residues away from the DNA binding sites as negative controls, Glu49 from the ␣3 and Thr90 from ␣5. Ala was converted to Ser and all others were mutated to Ala to understand their functional significance. K26A, N30A, R32A, R33A, S62A, Q63A, H64A, L65A, W67A, R69A, R70A, K78A and Y85A mutants did not show any DNA binding at the protein concentrations tested ( Figure 4A). The control mutations E49A and T90A did not affect DNA binding. A66S showed very slight DNA binding and unexpectedly, Q81A exhibited considerable shift. All the mutations mentioned above did not impair the dimerization of the protein ( Figure  4B). The mutants were neither partially degraded nor did they behave differently from the wild-type protein on SDSpolyacrylamide gel electrophoresis (PAGE) ( Figure 4C). Although the mutants were shown to be incapable of DNA binding ( Figure 4A), we wanted to see their behavior at higher protein concentrations (Supplementary Figure S4B). While N30A, R32A, R33A, Q63A, R69A and R70A mutants were absolutely defective in DNA binding, K26A, W67A and Y85A produced some DNA-protein complexes. Like HlyU Vc at higher concentrations, the other mutants (S62A, H64A, K78A, Q81A) produced supershifted bands, as also the control mutant T90A.

DNA-HlyU Vc interaction enhances promoter activity
The ability of HlyU Vc to activate transcription, in vivo, was monitored by ␤-gal assay. A plasmid containing lacZ gene under the control of the hlyA promoter along with the HlyU Vc binding site and a clone of hlyU Vc gene were cotransformed in E. coli BL21 (DE3). Wild-type HlyU Vc significantly boosted transcription of the reporter gene as compared to the negative control (pET23a) ( Figure 5). Y85A, Q63A and L25D mutants failed to physically interact with DNA, and when compared to wild-type HlyU Vc showed much less activity, which was statistically significant at P < 0.001. The relative activity found for the positive control in the reporter gene assay, thus, demonstrates the in vivo binding of HlyU Vc to the identified DNA partner and the positive control exerted by HlyU Vc on the hlyA promoter. The significantly different activity levels for the mutation in the dimerization interface (L25D), recognition helix (Q63A) and in the wing (Y85A) as compared to the wildtype activity emphasize the importance of these regions in the biological functioning of HlyU Vc. ␤-gal activity comparable to wild-type HlyU Vc was obtained for L98D (the difference being statistically non-significant at P < 0.001). This is in agreement with the finding that this mutant retains the dimeric structure to a considerable extent and binds DNA in vitro ( Figure 3E).

DNA-HlyU Vc interactions induce bending of DNA
DNA-binding proteins often use combination of base readout and shape readout mechanisms to bind to specific sites. Shape readouts may be results of intrinsic conformation of the nucleic acid or the structural changes brought upon DNA-protein interaction. The circular permutation assay, used to decipher any structural change of the double helix on DNA-HlyU Vc complex formation, indicated that the angle of divergence of DNA 6 from linearity is 68 ± 2 • (Figure 6A-C).

MD simulations emphasize the solution studies
Models of the 35-bp partner DNA and HlyU Vc were generated with different bent angles of the DNA. However, the best results were obtained for the complex with 68 • bend and were used for MD simulations for 10 ns. The complex retains a bend angle of ∼55 • after the simulation (Figure 6D), which is consistent with the finding that HlyU Vc binding indeed induces considerable curvature in the DNA (Supplementary Movie SM1).
DNA binding also induces some conformational changes in the protein structure. Noticeable change occurs in ␣1 whose geometry changes from being kinked (unbound state) to curved (bound state) with a significant decrease in the value of bend angle (Supplementary Figure S5). In ␣4, there is a 2-fold increase in helix bending on DNA binding; the opposite being the case with ␣5.
A model showing the putative interactions of the functionally important amino acids with the DNA is shown in Figure 7. Among the residues that were experimentally found to be important in DNA binding, the side chains of Lys26, Asn30, Arg32, Arg33, Ser62, Gln63, His64, Arg69, Arg70, Lys78 and Tyr85 point toward the DNA bases and simulations indicate the existence of hydrogen bonding (Supplementary Table S2 and Figure 7A) or close van der Waal's contact involving Trp67 ( Figure 7B). Two non-polar residues, Leu65 and Ala66, whose mutations resulted in altered DNA binding, may have an indirect effect by changing the stability of the recognition helix in which they are located.
The percentage occupancy of the hydrogen bonds between the amino acids and the corresponding bases during MD simulations (Supplementary Table S2) may explain the dependence of mobility shift on protein concentration ( Figure 4A and Supplementary Figure S4B). If we classify the strength of DNA binding into three groups, 'strong' (hydrogen bond occupancy of >15%), 'medium' (10-15%) and 'weak' (<10%), mutants of "strong" residues (Asn30, Arg32, Arg33, Gln63, Arg69 and Arg70) would be absolutely non-DNA binding irrespective of protein concentration. The mutants of 'medium' residues (Lys26, Trp67 and Tyr85) showed DNA binding only at high protein concentrations. For the last category (Ser62, His64, Lys78 and Gln81), though there may be some binding at low concentrations, all show supershift at high concentrations.
Another interesting observation in the MD simulation trajectory is the relation of the distance between the two Cys residues in the structure (at 38 and 104) and the contacts important in DNA binding. During the course of simulation there was not much visible change in the interaction between the major groove and the recognition helix. However, the interaction of the wing with the DNA seemed to be quite dynamic. Therefore, two very important wing residues, Lys78 and Tyr85, were chosen to highlight their mode of DNA contact. It was found that these residues moved away from the partner bases when the distance between Cys38 and Cys104 was small and vice versa ( Figure  8).
Simulation of the unbound HlyU Vc exhibited a reduction in the Cys38-Cys104 distance. The structure from the trajectory with the shortest Cys38-Cys104 distance (∼3Å) indicated a greater distance between the wing and the recognition helix (∼17Å). In contrast, in the DNA bound form, the distance between Cys38-Cys104 showed an increase in the value during simulation (Supplementary Movies SM2 and SM3). Also the wing is closer to the recognition helix (by ∼3Å) (Supplementary Figure S6), allowing generous contacts with both the grooves of the DNA.

DISCUSSION
To have a complete picture of a gene expression system, it is of primary importance to elucidate the regulatory sequences that include the promoter region together with enhancer elements to which activator proteins bind. HlyU is an important virulence activator, responsible for the activation of the rtxA1 operon in V. vulnificus, the plp-vah1 gene cluster and the rtxACHBDE gene cluster in Vibrio anguillarum, and hlyA gene and probably other virulence factors in V. cholerae (15). Recent investigations by several groups have thrown light on regulation of virulence in V. vulnificus and V. anguillarum by HlyU (27,32). As far as the V. cholerae system is concerned, the promoter sequence for hlyA gene had been determined by Williams and Man-ning in the 1990s (16). They also identified HlyU Vc to play some role in hlyA gene activation. However, at this time there is a lack of information on specific interaction between HlyU Vc and the enhancer sequence in the hlyA expression system.
In this article, HlyU Vc has been demonstrated to specifically bind to a 17-bp imperfect palindrome 5 -TAATTCAGACTAAATTA-3 173 bp upstream of the hlyA transcription start site (Figure 2). Studies with the mutations at the self-contacting interface residues, Leu25, Ala29, Met95 and Leu98, indicated that homodimerization of the protein is an essential criterion for DNA binding (Figure 3). From the crystal structure of HlyU Vc (PDB ID: 4OOI, (15); Supplementary Figure S1), the residues constituting the wHTH motif were clearly deciphered. Lys26, Asn30, Arg32, Arg33, Ser62, Gln63, His64, Leu65, Trp67, Arg69, Arg70, Lys78 and Tyr85 residues were found to be indispensable for DNA binding in vitro (Figure 4), as also suggested by the structure. The reporter gene assay strongly  suggests that DNA-HlyU Vc interaction augments the promoter activity while the mutated proteins (Y85A, Q63A and L25D) fail to do so ( Figure 5). The protein binding induced the target DNA to deviate 68 ± 2 • from linearity ( Figure 6). This is in conformity with the protein structure which indicates a distance less than 34Å between the two recognition helices in the dimer (15), necessitating a bent DNA partner. On docking, the protein fits perfectly with the ␣4 helices and the wings in the major and minor grooves, respectively, of a significantly bent DNA ( Figure  6D and Supplementary Movie SM1). wHTH proteins are   Figure 2B. known to interact with DNA with the recognition helices sitting on the major grooves and the wings drooping over the minor grooves (33). In our case, the involvement of both the types of grooves was also ascertained by competition with groove-specific dyes, the details of which are available as Supplementary Text S1. MD simulation studies validate the contacts of the DNA phosphates/bases with the experimentally determined important amino acids of HlyU Vc (Supplementary Table S2).
Our current findings open up many more questions about the molecular mechanism of regulation by HlyU Vc. Prokaryotic transcription control may involve direct contact of regulatory proteins (bound to DNA adjacent to the target promoter) with RNA polymerase, or alternatively, the regulatory proteins bound at more distally located enhancer sites may activate the promoter by DNA-looping, or with the help of mediator proteins, or by anti-repression (34). HlyU from V. vulnificus and V. anguillarum has been shown to act as an anti-repressor relieving the repression by H-NS (26,35). Liu and Crosa have proposed a model where the regulatory DNA is looped and involved in a bridge with H-NS that represses the rtx gene expression (36). Our results indicate the HlyU Vc binding site to be quite distant from the promoter which compels us to reflect on the possible mode of DNA looping with involvement of accessory mediator or repressor proteins. It is likely to observe such H-NS-mediated anti-repression in HlyU Vc too (Pal,A. et al., unpublished data).
Virulence genes are usually conditionally expressed under specific stimuli and are likely to be downregulated by some other. V. cholerae is able to respond to host environmental signals by activating transcriptional regulatory cascades. However, the mode of functioning switch in HlyU family of proteins for regulating the virulence determinants is yet unexplored. While several SmtB/ArsR family proteins use metal coordination for transformation into DNA non-binding form, many wHTH proteins of related structures, especially BigR (29% sequence identity to HlyU Vc (www.ebi.ac.uk/Tools/msa/clustalw2/)), while 174 residues out of 187 (in dimer) structurally aligned ((www.ccp4.ac. uk/); Supplementary Figure S8A and B) have been found to undergo DNA-binding/release using a redox switch. In the DNA binding, reduced thiol form of BigR, the two Cys residues are wide apart (>9Å); however, a C-terminal Tyr and an N-terminal Met residue from the two different subunits come closer as if to disrupt the formation of the disulphide bond (Supplementary Figure S8C). The oxidized BigR structure indicates increased distance between the Tyr and the Met residues (Supplementary Figure S8D) (37). The MD simulations suggest the possibility of a similar redox switch in HlyU. Simulation of the unbound HlyU Vc indicated a negative correlation between two distances--one between Cys38 and Cys104 and the other involving Met14 and His100, all of these residues occupying the equivalent positions of the redox switch protagonists in BigR. The trajectory of the DNA bound HlyU Vc revealed a steady concomitant decrease in the distance between Met14 and His100 with increasing stabilization of the DNA-protein complex and increasing distance between the Cys residues ( Figure 9 and Supplementary Movie SM2). The closing of the wings onto the minor groove with the parallel distanc- ing of the Cys residues, thus, leading to a better complex interaction, points toward a plausible analogy to the BigR redox model. Therefore, it would be interesting to know whether the cysteine(s) of the non-metal binding HlyU Vc play any such redox role. It is worth mentioning here that under oxygen-limiting conditions, an environment similar to the host intestines, V. cholerae virulence genes are highly expressed and the pathogen may use a thiol-based switch mechanism to sense oxygen-rich aquatic environments and oxygen-limited human host signals (38)(39)(40). Anaerobiosis is indeed reported to increase HlyA production in E. coli and Bacillus cereus (41,42). It may also be added that conservation of the residues proposed to be important in the probable redox switch of HlyU Vc is noted in several other species of Vibrio which share considerable sequence identity; HlyU from yet other species having divergent sequence show absence of Cys and the accessory residues (Supplementary Figure S9). With our ongoing research, we look forward to answering the yet unresolved questions, dissect the mechanistic details of regulation of hlyA gene by HlyU Vc and elucidate any possible conservation in the functioning of this family of protein.

SUPPLEMENTARY DATA
Supplementary Data are available at NAR Online.