Substitution of Glutamate Residue by Lysine in the Dimerization Domain Affects DNA Binding Ability of HapR by Inducing Structural Deformity in the DNA Binding Domain

HapR has been given the status of a high cell density master regulatory protein in Vibrio cholerae. Though many facts are known regarding its structural and functional aspects, much still can be learnt from natural variants of the wild type protein. This work aims at investigating the nature of functional inertness of a HapR natural variant harboring a substitution of a conserved glutamate residue at position 117 which participates in forming a salt bridge by lysine (HapRV2G-E117K). Experimental evidence presented here reveals the inability of this variant to interact with various cognate promoters by in vitro gel shift assay. Furthermore, the elution profiles of HapRV2G-E117K protein along with the wild type functional HapRV2G in size-exclusion chromatography as well as circular dichroism spectra did not reflect any significant differences in its structure, thereby indicating the intactness of dimer in the variant protein. To gain further insight into the global shape of the proteins, small angle X-ray scattering analysis (SAXS) was performed. Intriguingly, increased radius of gyration of HapRV2G-E117K of 27.5 Å in comparison to the wild type protein from SAXS data analyses implied a significant alteration in the global shape of the dimeric HapRV2G-E117K protein. Structure reconstruction brought forth that the DNA binding domains were substantially “parted away” in this variant. Taken together, our data illustrates that substitution of the conserved glutamate residue by lysine in the dimerization domain induces separation of the two DNA binding domains from their native-like positioning without altering the dimeric status of HapR variant.


Introduction
Vibrio cholerae is endowed with a remarkable capacity to sense diverse environmental settings and accordingly modulate disparate cellular events to maintain its sustainability in a given niche. The process of sensing and further coordination of cellular activities is largely dependent on a microbial social networking program known as quorum sensing. In recent past, an explosive growth in studies on quorum sensing of Vibrio cholerae clearly illustrates the intricacy of such complex signal transduction at the molecular level and its influence on the lifestyle of this bacterium. The intricate architecture of sensory circuit has revealed the participation of myriad of factors, namely (i) two autoinducer systems (AI-2/LuxPQ and CAI-1/CqsS), (ii) one growth phase-regulated VarS/A-CsrA/BCD unit, (iii) a small nucleoid protein, Fis and (iv) a cascade of small RNAs, plays a critical role in balancing the periodic appearance and the performance of master regulatory proteins LuxO and HapR [1]. Being the high cell density master regulatory protein, HapR is under the spotlight as it governs a constellation of disparate physiological events, thereby shaping the virulence and survival strategy of Vibrio cholerae [2]. Collectively, it represses biofilm development and the production of primary virulence factors, while stimulates the production of HA/protease, promotes chitin induced competence, increases resistance to protozoan grazing, enhances the survival against oxidative stress and controls the expression of gene encoding Hcp [3]. Structurally, the protein comprises of two main domains namely DNA binding domain (DBD) and dimerization domain (DD). There are a total of nine a helices where the first three helices form a putative DNA binding domain. The HTH motif comprises of helices a2 and a3. The remaining six a-helices are distributed in the Cterminal dimerization domain [4]. In a continuing effort, Zhu and coworkers have elegantly characterized additional novel direct targets of HapR and illustrated two distinct binding motifs (motif 1 and motif 2) in various target promoters. Motif 1 sequence contains dyad symmetry and a consensus AATAR9 (where R represents A or G) whereas motif 2 sequence lacks any dyad symmetry and contains a highly conserved consensus 'TGT' [5]. Though many crucial facts about the structural and functional aspects of wild type HapR are available, additional insights into the functionality of this molecule can further be explored by studying natural variants of the same. For instance, a natural non functional HapR variant harboring aspartate in place of glycine residue at position 39 (G 39 D; named as HapR V2 ) clearly elucidates the contributory role of glycine 39 in DNA binding activity of HapR [6]. Further, structural analysis reveals the mutation driven shape change in the DNA binding domains, thereby explaining the loss in DNA binding ability of this HapR variant. In short, molecular analysis of HapR V2 (G 39 D) natural variant illustrates the importance of a conserved glycine residue in the performance of HapR [6] which was not evaluated in the previous studies [4].
A flurry of research articles evidences the preponderance of natural variants of quorum sensing regulatory proteins from various strains of Vibrio cholerae [6][7][8][9][10]. Of these, strains having non-functional HapR variants are more frequently obtained in nature. In a recent effort, Zhu and colleagues have reported the identification of several quorum sensing defective strains of Vibrio cholerae harboring either a single or multiple mutations in HapR protein [10]. Though, these natural variants of HapR have been categorized as non functional based on certain phenotypic traits, further molecular analysis is required to understand the cause of their functional inertness. In this work, we have chosen to investigate the molecular basis of the functional impairment of a HapR natural variant harboring substitution of a conserved glutamate residue (E 117 ) at position 117 with lysine in the dimerization domain. In literature, similar substitution where glutamate is replaced by lysine could result in mutant proteins either with a compromise in DNA binding ability or dimer stability. For example, glutamate to lysine substitution in many cases such as E 62 K and E 95 K severely affect the DNA binding ability of yeast transcription factor (TF) IIB and Oct-1 respectively [11,12]. Similarly, analogous substitution (E 99 K) destabilizes the dimer of Vibrio harveyi flavin reductase FRP [13]. It should be noted that glutamate at position 99 forms a salt bridge, thereby contributing to dimer stability of Vibrio harveyi flavin reductase FRP [13]. With this background, we wanted to investigate the

Bacterial strains and media
The bacterial strains and plasmids used in this study are listed in Table 1. Vibrio cholerae strains were derived from a non-O1, non-O139 strain V2, serogroup O37. Strains were maintained at 270uC in Luria-Bertani (LB) medium containing 20% glycerol. Escherichia coli BL21 (DE3) (Novagen) was used for the overexpression of proteins. All strains were propagated at 37uC in liquid with agitation or on solid (1.5% agar) in Luria Broth unless mentioned otherwise. For protease assay, V. cholerae strains were grown with aeration at 37uC in tryptic soya broth without dextrose (TSB-D). When appropriate, the growth medium was supplemented with ampicillin (100 mg ml 21 ) and chloramphenicol (17 mg ml 21 ). All antibiotics were purchased from Sigma-Aldrich and media ingredients from Himedia and Difco.

Protease assay
Protease activity was measured using an azocasein assay as described earlier [7]. Briefly, the recombinant derivatives of V. cholerae strain V2s (Table 1) were grown in tryptic soya broth without dextrose (TSB-D), containing chloramphenicol (17 mg ml -1 ) and ampicillin (100 mg ml -1 ) accordingly, with agitation to stationary phase at 37uC. 100 ml of stationary phase culture supernatant was incubated with 100 ml azocasein (5 mg ml -1 in 100 mM Tris, pH 8.0) for 1 h at 37uC. The reaction was stopped by the addition of 400 ml of 10% tricholoroacetic acid. After centrifugation, supernatant was transferred to 700 ml of 525 mM NaOH, and the optical density was determined at 442 nm. One azocasein unit was defined as the amount of enzyme producing an increase of 0.01 OD units per hr.

Site-specific mutagenesis
Glutamate 117 to lysine (E 117 K) mutation in the functional hapR ORF was generated on plasmid pSV2G (Table 1) using quick change mutagenesis kit from Stratagene as per manufacturer's guidelines. The primers are listed in Table 2. Positive clones were analyzed by sequencing in their entirety to confirm the desired mutation. One of such mutant clone designated as pSV2G-E117K was further transformed into V2 S and the recombinant strain was designated as V2 S -R V2G -E 117 K.
Protein purification and Electrophoretic gel mobility shift assay with promoter regions of aphA, hapA and vc0900 HapR V2G and HapR V2G -E 117 K proteins were purified by Ni 2+ -Nitrilotriacetic acid chromatography as described previously [6]. Both wild type and variant HapR were cloned into NdeI-BamHI site of the pET15b vector (Novagen) to generate N-terminal 6X His-HapR fusion proteins. All the clones were confirmed by sequencing and transformed into E. coli BL21 (DE3). After induction with 0.4 mM isopropyl 1-thio-b-D-galactopyranoside (IPTG), HapR proteins were purified through Qiagen Ni 2+nitrilotriacetic acid columns and subjected to overnight dialysis in a solution of buffer A containing 10 mM Tris-HCl, pH 7.9, 300 mM KCl, 0.1 mM EDTA. Gel mobility shift assay was done essentially as described earlier [6]. Briefly, three fragments of 399, 665 and 467 bp corresponding to promoter regions of aphA, hapA and vc0900 respectively, were amplified with primer pairs as listed in Table 2. The fragments were gel-purified and end labeled with [c 32 -P] dATP using T4 polynucleotide kinase (New England Biolabs). The binding reaction was carried out with 4 ng of labeled fragment in 10 mM Tris-HCl, pH 7.9, 1 mM EDTA, 1 mM DTT, 60 mM KCl, 10% glycerol, 5 mg BSA and 1 mg poly (dI- dC) in a 20 ml reaction volume for 20 min at 26uC. The reaction mixture was applied to a 5.5% polyacrylamide gel and subjected to electrophoresis in 1X Tris-Acetate-EDTA, pH 8.5, at 4uC. The gel was dried and autoradiographed to examine the shift of the band.

Circular Dichroism (CD) measurement
Both HapR V2G and HapR V2G -E 117 K were examined by CD using a Jasco J-810 spectropolarimeter. Measurements in the far ultraviolet region (250-200 nm) were performed on protein solutions (0.2 mg/ml) employing a cell with path length of

Molecular weight determination
HapR V2G , HapR V2 and HapR V2G -E 117 K were dialyzed overnight in buffer A. All proteins were subjected to molecular sieve chromatography using a Superdex 200 5/150 GL analytical column coupled to an ''AKTA-Purifier'' chromatography system (GE-Healthcare). The molecular weight (Mr) of all the proteins were evaluated according to a calibration curve generated with the gel filtration of molecular mass markers (GE-Healthcare) including blue dextran (200 kDa), conalbumin (75 kDa), ovalbumin (44 kDa), carbonic anhydrase (29 kDa), ribonulcease A (13.7 kDa) and aprotinin (6.5 kDa) ( Figure 1A). These standards were loaded independently at the concentrations recommended by GE-Healthcare. The elution volume of blue dextran was used to determine the void volume (Vo = 1.215 ml) and the total volume (Vt = 3 ml) was provided by the product instruction manual. The peak elution volumes (Ve) were calculated from the chromatogram. Kav were calculated using the following equation: The calibration curve was determined by plotting the Kav of the protein standards against the Mr of the standards. Elution volumes were determined by monitoring the absorbance at 280 nm. All proteins were eluted at the same volume. Kav values for all the proteins were calculated based on elution volumes and fitted to a calibration curve to estimate the size of the proteins.

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and western blot analysis
Western blot analysis was done to check the expression of HapR. Briefly, FLAG-tagged recombinant derivatives of V. cholerae strain V2s were grown overnight in TSB-D medium with agitation. Protein samples were prepared from equal amount of bacterial cells and separated on 13.5% gel by SDS-PAGE. For immunoblotting, the proteins were transferred to Immobilon-P membrane (Millipore) at 75 mA for 1 h. Membrane was subsequently blocked in PBS-T with 5% skim milk at 37uC for 2 h and then shifted to 4uC for overnight with shaking. Blot was then washed in PBS-T five times for 10 min each, incubated in monoclonal HRP-conjugated Anti-flag (Sigma Aldrich) at a dilution of 1:5000 in PBS-T with 2% skim milk for 1 h and again washed in PBS-T five times for 10 min each. Proteins were  Small Angle X-ray Scattering Data Acquisition, Analysis and Structure Reconstruction All SAXS intensity profiles were acquired at the X9 beam line (National Synchrotron Light Source, Brookhaven National Laboratory). Scattered X-rays were recorded as images on Pilatus detectors, corrected for slit smearing, circularly averaged about the beam centre, and the contribution from buffer components were subtracted using Python script based programs written by Dr. Lin Yang (X9 Beam line, NSLS). For each experiment, 120 ml of protein solutions (HapR V2G and HapR V2G -E 117 K) and their matched buffers from FPLC were exposed to X-rays in a quartz capillary flow cell. Using programmable liquid handling robotics at X9 beam line, the collection of entire SAXS data described in this study were carried out in triplicate and averaged during processing. To calibrate the beam intensity at zero angles and estimate the actual concentration of protein samples, the SAXS datasets were collected under identical conditions on a series of hen egg white lysozyme (in NaOAc buffer containing 150 mM NaCl, pH 3.8) and recombinant gelsolin (in Tris-EGTA, buffer pH 8) with their predetermined concentrations. Image processing provided the scattering intensity (I) as a function of momentum transfer vector, Q (Q = [4psinh]/l), where l and h are the wavelength of X-ray and the scattering angle, respectively. Kratky plots (I (Q) Q 2 vs. Q) were generated to interpret globular scattering nature of the wild type and variant protein in solution. The Guinier approximations of the low Q region were carried out using PRIMUS software [14] presuming globular and rod-like scattering shape which provided the radius of gyration (R G ) and radius of cross-section (R C ) of the predominant shape of protein molecules. Using the relationship between R G and R C [Eq. 2], we estimated the linear dimension of the molecules with following equation: Further, indirect Fourier transformation of the SAXS data (Q range: 0.008-0.25 Å -1 ) was performed using GNOM45 program [15] to obtain probability of finding various pairwise vectors arising from the scattering shape of the protein molecules. During P(r) curve estimation, the probability of finding a pairwise vector of length equal to 0 Å and equal to the maximum linear dimension (D max ) of the molecule was considered to be zero. This analysis also provided an estimate of the R G and scattering intensity at zero scattering angles, I 0 . To visualize the three dimensional shapes of wild type and the variant HapR protein, ten independent models of each were generated using the DAMMINIQ program and the respective SAXS I(Q) profile as reference [16]. No symmetry or shape bias was employed during each shape reconstruction. For each protein, the independent structural models were averaged using DAMAVER suite of programs [17] which resulted in a structural model capable of best-representing the solution shape of the proteins in solution. The SAXS data based dummy model and the available crystal structure of HapR (PDB ID: 2PBX) were overlaid by aligning inertial axes of the models using SUP-COMB20 program [18]. The open source VMD version 1.9.1 and Chimera 1.6.1 were used for graphical analysis and figure generation.

Results and Discussion
Molecular weight and secondary structure determination of non-functional HapR V2G -E 117 K Like other members of TetR family proteins, HapR also acts as a dimer [4]. Analysis of the dimerization domain and interface reveals the significance of critical residues in the formation and stabilization of dimer. One such residue glutamate at position 117 (E 117 ) contributes to the stability of the HapR dimer by forming a salt bridge with arginine residue at position 123 (R 123 ) of the other monomer [4]. There is a preponderance of data elucidating the crucial role of salt bridge in stabilizing protein structure. For example, substitution of a glutamate residue at position 99 with lysine (E 99 K) in case of Vibrio harveyi flavin reductase FRP completely abolishes the dimer formation. It should be noted that E 99 makes salt bridge with R 133 and R 225 of the other monomer in FRP [13]. In our study, we were driven by a desire to understand the molecular basis of the non-functionality of previously identified HapR variant harboring substitution of a conserved glutamate residue at position 117 (E 117 ) with lysine. As described in the preceding section, E 117 of a functional wild type HapR participates in the formation of salt bridge, it is therefore conceivable that substitution of this residue by lysine in non-functional HapR   Figure 1A). In addition to the wild type functional protein as control, we also kept a previously identified non functional HapR variant (HapR V2 ) that is defective in DNA binding but forms a stable dimer [6] as a second control in gel filtration chromatography experiment ( Figure 1B). Intriguingly, no significant difference was observed in the gel filtration elution profiles of the purified His-tagged HapR functional (HapR V2G ) protein, DNA binding defective variant (HapR V2 ) and dimerization domain variant (HapR V2G -E 117 K). We also carried out CD data analysis to examine any alteration in the secondary structure ( Figure 2). K2D2 analysis of the CD data (in the range of 250-  Evaluate DNA binding ability of variant HapR V2G -E 117 K HapR justifies its role as a master regulator by interacting with a range of cognate promoter sequences. To assess the DNA binding ability of HapR V2G -E 117 K, a gel shift assay was employed with hapA, aphA and vc0900 promoter regions. It should be noted that the promoter region of vc0900 contains motif 1 binding site while the promoter regions of hapA and aphA contain motif 2 binding site. Unlike its wild type functional counterpart HapR V2G , HapR V2G -E 117 K fail to bind to any of these promoter regions, thereby indicating a compromise in DNA binding ability (Figure 3). Combining gel filtration and gel retardation results, we conclude that substitution of a salt bridge glutamate residue with lysine in the dimerization domain abrogates DNA binding activity without affecting dimeric status of HapR variant.

Functional examination and stability of variant HapR V2G -E 117 K under in vivo condition
From in vitro gel shift assay, inability of the HapR variant (HapR V2G -E 117 K) to shift promoter region of gene encoding HA/ protease (hapA) is evident. It is therefore expected that this variant will remain non functional to stimulate protease production in a protease negative strain of Vibrio cholerae. To ascertain its in-vivo functionality, recombinant constructs of hapR V2G (pSV2G, wild type functional HapR control) and hapR-E 117 K (pSV2G-E117K) were transformed into a protease negative Vibrio cholerae strain V2 S (hapR disrupted strain of V2, Table 1) to generate strains V2 S -R V2G and V2 S -R V2G -E 117 K. In contrast to V2 S -R V2G , V2 S -R V2G -E 117 K remains protease negative ( Figure 4A). The protease data recapitulates in vitro gel shift data with promoter hapA where protease production is directly proportional to the functional level of HapR. In addition to the protease production, we have also checked the rugosity pattern of V2 S -strains harboring HapR proteins. It has been observed that strain having non-functional HapR develops rugosity ( Figure 5). This data is in agreement with that of Wang et al, who demonstrated rugosity in the strain harboring such mutant HapR protein [10].
It has been documented that substitution of an amino acid often results in an increased susceptibility of mutants to proteases, if critical for global stability of the folded protein [19]. It is thus possible that substitution of glutamate by lysine may affect the in vivo stability of HapR V2G -E 117 K variant. To investigate the stability, FLAG epitope was inserted in the C-terminal domain of wild type HapR V2G and variant protein ( Table 1). The recombinant constructs were transformed into a protease negative Vibrio cholerae strains V2 S (hapR disrupted strain of V2, Table 1) to generate strains V2 S -R V2G -FLAG and V2 S -R V2G -E 117 K-FLAG. The recombinant strains were then subjected to western blot analysis to examine the stability of each protein. Our western blot analysis clearly indicated the intactness of lysine (K 117 ) variant along with wild type HapR protein ( Figure 6A). The protease assay with FLAG containing recombinant constructs further confirmed no functional alteration upon FLAG insertion ( Figure 4B). Until now results support that substitution of a salt bridge contact residue glutamate by lysine abrogates DNA binding ability without affecting stability and dimeric status of HapR V2G -E 117 K variant.

Structural deformity of HapR V2G E 117 K variant as evidenced by SAXS data analysis and structure restoration
In the preceding section, our characterization of hydrodynamic, secondary structural and stability analysis of the proteins ruled out a grossly misfolded shape of HapR dimerization domain HapR V2G -E 117 K variant, but lacked in explaining the loss of DNA binding function. In this scenario, high resolution structural analysis may shed further light on the global architecture, thereby enabling us to understand the loss of DNA binding activity of these dimerization domain variant proteins. To further delve into the structural details, SAXS analysis was performed as described previously [6]. SAXS scattering data lacking an upturned profile in low Q region assured lack of aggregation in samples during data collection ( Figure 7A). Considering globular scattering nature of the protein, Guinier analysis indicated that both wild type and variant HapR adopted a monodisperse profile in solution. Steeper slope of the linear fit for the HapR V2G -E 117 K variant protein (compared to wild type) indicated a bigger R G for the variant (Figure 7A inset). In fact, the R G values were 23.4 Å and 27.1 Å for the HapR V2G (functional) and HapR V2G -E 117 K protein, respectively (Table 3). Presuming rod-like scattering shape for the protein molecules provided the R C values for the two proteins to be 10.1 Å and 8.3 Å . Using the Eq. 2, the L values for wild type and HapR variant were estimated to be 73 Å and 89 Å , respectively. Interestingly, peak-like Kratky plot of the wild type protein confirmed a globular scattering nature of HapR V2G molecules in solution, a wider peak with I(Q)*Q 2 not dampening to zero as a function of increasing Q for the mutant suggested relatively higher inherent disorder in the HapR V2G -E 117 K protein molecules ( Figure 7B). Increased dimensions of the variant protein were also seen from the indirect Fourier transformation of the SAXS datasets ( Figure 7C). The wild type functional protein HapR V2G showed P(r) curve with D max and R G about 78 Å and 23.9 Å , respectively [6]. At the same time, the P(r) analysis for HapR V2G -E 117 K protein provided D max and R G close to 93 Å and 27.5 Å , respectively. The computed P(r) profile for the HapR V2G -E 117 K variant suggested an extended curve seen for shapes with tail-like features.
To gain insight into the shape of the two proteins, ab initio modeling approaches were followed using dummy residues. The model solved for the predominant solution shape of HapR V2G agreed well with the crystal structure of the HapR protein (PDB ID: 2PBX), as published earlier ( Figure 8A). In contrast, the model solved for HapR V2G -E 117 K variant protein revealed a much more open structure, particularly the DNA binding domains (comparing with the HapR V2G ). Overlay of the two models confirmed that while the shape profile of the dimerization domain of the variant remained similar to wild type protein, the DNA binding domains in mutant somewhat ''fell-apart''. Moreover, this deviation in the global shape of the mutant occurred only in dimension (please see the side-views in Figure 8B). To realize how this single point mutation is causing such a deleterious effect, we analyzed the polar interactions surrounding E 117 in the crystal structure of the HapR V2G (PDB ID: 2PBX) ( Figure 8C). It brought forth that alongside interacting with R 123 , HapR V2G -E 117 K indirectly positions other residues in vicinity viz. K 117 and some water molecules. A mutation to lysine, not only reverses the local electrostatic surface potential which would alter the intricate network of interactions resolved for the native protein, but also the extended length of lysine side-chain compared to that of glutamate would find it difficult to accommodate in the groove formed between the upper portion of the dimeric interface. Replacement of two glutamates in the dimer (in the zone highlighted in Figure 8C) by two lysines would cause a substantial remodeling of the local topography. Very likely, the adjustments are done by parting away of the upper portion (since lower part of molecule is engaged in dimerization). A culmination of local changes eventually results in eventual separation of the upper domains, a geometry which has evolved for binding different double stranded DNA ( Figure 8D).

Conclusions
There is a growing body of evidence that elucidates how substitution of a single amino acid could affect the functionality of a DNA binding protein by altering its DNA binding activity, dimeric status and overall stability by increasing susceptibility to proteases [6,11,13,[19][20][21][22]. Keeping this view in mind, we wanted to examine the underlying cause of non-functionality of a natural HapR variant designated as HapR V2G -E 117 K where a salt bridge glutamate residue is replaced by lysine moiety. It has been documented that similar substitution where glutamate is replaced by lysine could result in mutant proteins either with a compromise in the DNA binding ability or dimer stability [11][12][13]. A combined in vivo and in vitro analysis helped us to establish that HapR V2G -E 117 K variant protein forms a stable dimer with a compromise in its DNA-binding activity. As evidences suggest salt bridges not only maintain the proper orientation as well as stability of dimer interface, but also controls the wedge angle between DNA-binding domain and the dimerization domain, thereby mediating interaction between the protein molecules and their cognate DNA sequences [23,24,25]. In our case, replacement of a salt bridge contact residue glutamate by lysine neither affects the dimer stability nor overall protein stability, but it introduces a deformity in the DNA binding domain as revealed by SAXS analysis, thereby offering a plausible explanation of lack of DNA binding activity and overall cellular performance of HapR V2G -E 117 K variant. Based on our previous work and present observation, it is seemingly apparent that HapR adopts a particular ''Y'' shape that facilitates its interaction with cognate promoters ([6]; Figure 8). Any deviations from its typical ''Y'' shape either too close [6] or too widen (Figure 8) substantially compromises its functionality. Interestingly, glutamate (E 117 ) residue is highly conserved in the HapR homologues of other Vibrio species (Figure 9). It would be interesting to examine the effect of such mutation on the functionality of HapR homologues of different Vibrio species. Studies being pursued in our laboratory will possibly shed more insight into the encoded balance between the sequence content, global fold and functionality of HapR homologues, and our current work will be interesting to community involved in shapefunction studies of proteins, particularly Vibrio species.