Crystal Structure of the Transcriptional Regulator Rv0678 of Mycobacterium tuberculosis*

Background: The expression of the Mycobacterium tuberculosis MmpS5-MmpL5 transporter is controlled by the MarR-like transcriptional regulator Rv0678. Results: Rv0678 forms a dimeric two-domain molecule with the architecture similar to members of the MarR family of transcriptional regulators. Conclusion: Rv0678 is distinct in that its DNA-binding and dimerization domains cooperate to bind an inducing ligand. Significance: These findings suggest a mechanism for ligand and regulator derepression. Recent work demonstrates that the MmpL (mycobacterial membrane protein large) transporters are dedicated to the export of mycobacterial lipids for cell wall biosynthesis. An MmpL transporter frequently works with an accessory protein, belonging to the MmpS (mycobacterial membrane protein small) family, to transport these key virulence factors. One such efflux system in Mycobacterium tuberculosis is the MmpS5-MmpL5 transporter. The expression of MmpS5-MmpL5 is controlled by the MarR-like transcriptional regulator Rv0678, whose open reading frame is located downstream of the mmpS5-mmpL5 operon. To elucidate the structural basis of Rv0678 regulation, we have determined the crystal structure of this regulator, to 1.64 Å resolution, revealing a dimeric two-domain molecule with an architecture similar to members of the MarR family of transcriptional regulators. Rv0678 is distinct from other MarR regulators in that its DNA-binding and dimerization domains are clustered together. These two domains seemingly cooperate to bind an inducing ligand that we identified as 2-stearoylglycerol, which is a fatty acid glycerol ester. The structure also suggests that the conformational change leading to substrate-mediated derepression is primarily caused by a rigid body rotational motion of the entire DNA-binding domain of the regulator toward the dimerization domain. This movement results in a conformational state that is incompatible with DNA binding. We demonstrate using electrophoretic mobility shift assays that Rv0678 binds to the mmpS5-mmpL5, mmpS4-mmpL4, and the mmpS2-mmpL2 promoters. Binding by Rv0678 was reversed upon the addition of the ligand. These findings provide new insight into the mechanisms of gene regulation in the MarR family of regulators.

Recent work demonstrates that the MmpL (mycobacterial membrane protein large) transporters are dedicated to the export of mycobacterial lipids for cell wall biosynthesis. An MmpL transporter frequently works with an accessory protein, belonging to the MmpS (mycobacterial membrane protein small) family, to transport these key virulence factors. One such efflux system in Mycobacterium tuberculosis is the MmpS5-MmpL5 transporter. The expression of MmpS5-MmpL5 is controlled by the MarR-like transcriptional regulator Rv0678, whose open reading frame is located downstream of the mmpS5-mmpL5 operon. To elucidate the structural basis of Rv0678 regulation, we have determined the crystal structure of this regulator, to 1.64 Å resolution, revealing a dimeric two-domain molecule with an architecture similar to members of the MarR family of transcriptional regulators. Rv0678 is distinct from other MarR regulators in that its DNA-binding and dimerization domains are clustered together. These two domains seemingly cooperate to bind an inducing ligand that we identified as 2-stearoylglycerol, which is a fatty acid glycerol ester. The structure also suggests that the conformational change leading to substratemediated derepression is primarily caused by a rigid body rotational motion of the entire DNA-binding domain of the regulator toward the dimerization domain. This movement results in a conformational state that is incompatible with DNA binding. We demonstrate using electrophoretic mobility shift assays that Rv0678 binds to the mmpS5-mmpL5, mmpS4-mmpL4, and the mmpS2-mmpL2 promoters. Binding by Rv0678 was reversed upon the addition of the ligand. These findings provide new insight into the mechanisms of gene regulation in the MarR family of regulators. 3 is one of the oldest described diseases and remains a significant global problem with more than 8 million new cases reported annually (1). The World Health Organization estimates that one-third of the world's population is infected with Mycobacterium tuberculosis, and most of these individuals have latent TB (2). TB treatments are notoriously difficult and are compromised by the emergence of multiple drug-resistant, extensively drug-resistant, and totally drug-resistant bacterial strains (3)(4)(5)(6)(7). The development of drug-resistant M. tuberculosis strains is a major threat that challenges global prospects for TB control.

Tuberculosis (TB)
Although mycobacteria cluster phylogenetically with Grampositive prokaryotes, they are structurally more similar to Gram-negative bacteria. These mycobacteria are protected by an outer lipid bilayer made of mycolic acids and a cell envelope composed of non-covalently bound lipids and glycolipids. The unique structure and composition of the cell wall differentiates this highly pathogenic microorganism from other prokaryotes. The mycobacterial cell wall plays a crucial role in the hostpathogen interface on several levels (8). First, the thick, greasy cell wall acts as an effective layer of protection, providing intrinsic resistance to antibiotics and bactericidal components of the host immune response. Second, the surface-exposed polyketide and glycoconjugate lipids of the M. tuberculosis cell wall are associated with bacterial virulence (9 -12).
The genome of M. tuberculosis H37Rv contains 15 genes that encode for the resistance-nodulation-cell division (RND) proteins designated MmpL transporters (13,14). Unlike the RNDtype efflux pumps of Gram-negative bacteria, MmpL proteins do not usually participate in antibiotic efflux. Instead, there is strong evidence that these MmpL proteins are responsible for exporting fatty acids and lipidic elements of the cell wall (8 -10, 12, 15, 16). Five mmpL genes are located adjacent to genes cod-ing for proteins involved in fatty acid or polyketide synthesis, suggesting that the MmpL membrane proteins transport these key virulence factors (9,10). Similar to RND proteins of Gramnegative bacteria, the MmpL transporters of M. tuberculosis are believed to work in conjunction with accessory proteins. Specifically, MmpL transporters form complexes with the MmpS family proteins in order to export cell wall lipid constituents (18). Five genes encoding MmpS proteins are adjacent to genes encoding MmpL proteins (8,13). Work in the model organism Mycobacterium smegmatis demonstrated that MmpS4 was required for bacterial sliding motility and biofilm formation (19). That the mmpS4 and mmpL4 mutants had similar phenotypes underscores a coordinated function for cognate MmpS-MmpL proteins.
Our efforts have focused on elucidating how M. tuberculosis transport systems are regulated. We previously crystallized the Rv3066 efflux regulator both in the absence and presence of bound substrate (20). Our data indicated that ligand binding triggers a rotational motion of the regulator, which in turn releases the cognate DNA and induces the expression of the Mmr efflux pump (20). We report here the crystal structure of the Rv0678 regulator, which has been proposed to control the transcriptional regulation of the MmpS5-MmpL5 transport system. Rv0678 belongs to the MarR family of regulators, which are found ubiquitously in bacteria and archaea and control various important biological processes, such as resistance to antimicrobials, sensing of oxidative stress agents, and regulation of virulence factors (21). Typically, the MarR family regulators are dimeric in form, and their protein sequences are poorly conserved. However, these proteins share a common fold, consisting of a helical dimerization domain and two winged helixturn-helix DNA-binding domains within the dimer (22). Our data suggest that fatty acid glycerol esters are the natural ligands of the Rv0678 regulator. An electrophoretic mobility shift assay indicates that Rv0678 binds promoters of the mmpL2, mmpL4, and mmpL5 operons. These results emphasize the importance of the Rv0678 regulator, which appears to regulate multiple MmpL transport systems.

EXPERIMENTAL PROCEDURES
Cloning of rv0678-The rv0678 ORF from genomic DNA of M. tuberculosis strain H37Rv was amplified by PCR using the primers 5Ј-CCATGGGCAGCGTCAACGACGGGGTC-3Ј and 5Ј-GGATCCTCAGTGATGATGATGATGATGGTCG-TCCTCTCCGGTTCG-3Ј to generate a product that encodes a Rv0678 recombinant protein with a His 6 tag at the C terminus. The corresponding PCR product was digested with NcoI and BamHI, extracted from the agarose gel, and inserted into pET-15b as described by the manufacturer (Merck). The recombinant plasmid (pET15b⍀rv0678) was transformed into DH5␣ cells, and the transformants were selected on LB agar plates containing 100 g/ml ampicillin. The presence of the correct rv0678 sequence in the plasmid construct was verified by DNA sequencing.
Expression and Purification of Rv0678-Briefly, the fulllength Rv0678 protein containing a His 6 tag at the C terminus was overproduced in Escherichia coli BL21(DE3) cells possessing pET15b⍀rv0678. Cells were grown in 6 liters of Luria broth (LB) medium with 100 g/ml ampicillin at 37°C. When the A 600 reached 0.5, the culture was treated with 0.2 mM isopropyl-␤-D-thiogalactopyranoside to induce Rv0678 expression, and cells were harvested within 3 h. The collected bacterial cells were suspended in 100 ml of ice-cold buffer containing 20 mM Na-HEPES (pH 7.2) and 200 mM NaCl, 10 mM MgCl 2 , and 0.2 mg of DNase I (Sigma-Aldrich). The cells were then lysed with a French pressure cell. Cell debris was removed by centrifugation for 45 min at 4°C and 20,000 rpm. The crude lysate was filtered through a 0.2-m membrane and was loaded onto a 5-ml Hi-Trap Ni 2ϩ -chelating column (GE Healthcare) preequilibrated with 20 mM Na-HEPES (pH 7.2) and 200 mM NaCl. To remove unbound proteins and impurities, the column was first washed with 6 column volumes of buffer containing 50 mM imidazole, 250 mM NaCl, and 20 mM Na-HEPES (pH 7.2). The Rv0678 protein was then eluted with 4 column volumes of buffer containing 300 mM imidazole, 250 mM NaCl, and 20 mM Na-HEPES (pH 7.2). The purity of the protein was judged using 12.5% SDS-PAGE stained with Coomassie Brilliant Blue. The purified protein was extensively dialyzed against buffer containing 100 mM imidazole, 250 mM NaCl, and 20 mM Na-HEPES (pH 7.5) and concentrated to 20 mg/ml.
Crystallization of Rv0678-All crystals of the His 6 Rv0678 regulator were obtained using hanging drop vapor diffusion. The Rv0678 crystals were grown at room temperature in 24-well plates with the following procedures. A 2-l protein solution containing 20 mg/ml Rv0678 protein in 20 mM Na-HEPES (pH 7.5), 250 mM NaCl, and 100 mM imidazole was mixed with 2 l of reservoir solution containing 28% polyethylene glycol (PEG) 1000, 0.1 M sodium acetate (pH 4.0), 0.04 M NaCl, and 5% glycerol. The resultant mixture was equilibrated against 500 l of the reservoir solution. Crystals grew to a full size in the drops within 2 weeks. Typically, the dimensions of the crystals were 0.2 ϫ 0.05 ϫ 0.05 mm. Cryoprotection was achieved by raising the PEG 1000 concentration stepwise to 35% with a 3.5% increment in each step. Crystals of the tungsten derivative were prepared by incubating the crystals of Rv0678 in solution containing 28% PEG 1000, 0.1 M sodium acetate (pH 4.0), 0.04 M NaCl, 5% glycerol, and 1 mM (NH 4 ) 2 W 6 (-O) 6 (-Cl) 6 Cl 6 for 24 h at 25°C.
Data Collection, Structural Determination, and Refinement-All diffraction data were collected at 100 K at beamline 24ID-E located at the Advanced Photon Source, using an ADSC Quantum 315 CCD-based detector. Diffraction data were processed using DENZO and scaled using SCALEPACK (23). The crystals of Rv0678 belong to the space group P1 (Table 1). Based on the molecular mass of Rv0678 (18.34 kDa), the asymmetric unit is expected to contain four regulator molecules with a solvent content of 45.26%. Six tungsten cluster sites were identified using SHELXC and SHELXD (24), as implemented in the HKL2MAP package (25). Single isomorphous replacement with anomalous scattering was employed to obtain experimental phases using the program MLPHARE (26,27). The resulting phases were then subjected to density modification and NCS averaging using the program PARROT (28). The phases were of excellent quality and allowed for tracing of most of the molecule in PHENIX AutoBuild (29), which led to an initial model with over 90% amino acid residues containing side chains. The remaining part of the model was manually constructed using the program Coot (30). Then the model was refined using PHENIX (29), leaving 5% of reflections in the Free-R set. Iterations of refinement using PHENIX (29) and CNS (31) and model building in Coot (30) led to the current model, which consists of two dimers (587 residues in total in the asymmetric unit) with excellent geometrical characteristics ( Table 1).
Identification of Fortuitous Ligand-To identify the nature of the bound ligand in crystals of Rv0678, we used gas chromatography coupled with mass spectrometry (GC-MS). The Rv0678 crystals were extensively washed with the crystallization buffer and transferred into deionized water. The mixture was then incubated at 100°C for 5 min, and then chloroform was added into the mixture to a final concentration of 80% (v/v) to denature the protein and allow for the extraction of ligand. GC-MS analysis indicated that the bound ligand was octadecanoic acid, 2-hydroxyl-1-(hydroxymethyl)ethyl ester, also called 2-stearoylglycerol.
Virtual Ligand Screening Using AutoDock Vina-AutoDock Vina (32) was used for virtual ligand screening of a variety of compounds. The docking area was assigned visually to cover the internal cavity of the Rv0678 dimer. A grid of 35 ϫ 35 ϫ 35 Å with 0.375-Å spacing was calculated around the docking area for all atom types presented in the DrugBank (33) and ZINC (34) libraries using AutoGrid. The iterated local search global optimizer algorithm was employed to predict the binding free energies for these compounds.
Isothermal Titration Calorimetry for Ligand Binding-We used isothermal titration calorimetry to determine the binding affinity of 1-stearoyl-rac-glycerol (an isomer of 2-stearoylglycerol) to the purified Rv0678 regulator. Measurements were performed on a VP-Microcalorimeter (MicroCal, Northampton, MA) at 25°C. Before titration, the protein was thoroughly dialyzed against buffer containing 10 mM sodium phosphate, pH 7.2, 100 mM NaCl, and 0.001% n-dodecyl-␤-maltoside. The protein concentration was determined using the Bradford assay. The dimeric Rv0678 sample was then adjusted to a final concentration of 200 M and served as the titrant. The ligand solution contained 10 M 1-stearoyl-rac-glycerol, 10 mM sodium phosphate, pH 7.2, 100 mM NaCl, and 0.001% n-dodecyl-␤-maltoside. The protein and ligand samples were degassed before they were loaded into the cell and syringe. Binding experiments were carried out with the ligand solution (1.5 ml) in the cell and the protein solution as the injectant. Ten-microliter injections of the ligand solution were used for data collection.
Injections occurred at intervals of 300 s, and the duration time of each injection was 20 s. Heat transfer (cal/s) was measured as a function of elapsed time (s). The mean enthalpies measured from injection of the ligand in the buffer were subtracted from raw titration data before data analysis with ORIGIN software (MicroCal). Titration curves were fitted by a nonlinear least squares method to a function for the binding of a ligand to a macromolecule. Nonlinear regression fitting to the binding isotherm provided the equilibrium binding constant (K a ϭ 1/K D ) and enthalpy of binding (⌬H). Based on the values of K a , the change in free energy (⌬G) and entropy (⌬S) were calculated with the equation, ⌬G ϭ ϪRT lnK a ϭ ⌬H Ϫ T⌬S, where T is 273 K and R is 1.9872 cal/K/mol. Calorimetry trials were also carried out in the absence of Rv0678 in the same experimental conditions. No change in heat was observed in the injections throughout the experiment.
Electrophoretic Mobility Shift Assay-Probes were amplified from the H37Rv genome using the primers listed in Table 2. All probes were labeled with digoxigenin using the Roche DIG Gel Shift kit. For EMSA analysis, 12 nM DIG-labeled probe and the indicated micromolar concentrations of protein were incubated for 45 min at room temperature in the Roche binding buffer modified by the addition of 0.25 mg/ml herring sperm DNA, and 0.75 mg/ml poly(dI-dC). For ligand competition studies, 1-stearoyl-rac-glycerol (an isomer of 2-stearoylglycerol) (Sigma-Aldrich) was resuspended in hot acetone and added to EMSA reactions at 1 M final concentration. Competition reactions were performed at 37°C. All reactions were resolved on a 6% native polyacrylamide gel in TBE buffer and transferred to nylon membrane, and DIG-labeled DNA-protein complexes were detected following the manufacturer's recommendations. Chemiluminescent signals were acquired using an ImageQuant LAS 4000 imager (GE Healthcare).  Dye Primer-based DNase I Footprint Assay-DNase I footprinting was performed as described by Zianni et al. (35). The 296-bp Rv0678-mmpS5 probe was generated by PCR using the primers 6FAM-Rv0678-F and HEX-Rv0678-R. Gel-purified, fluorescently labeled probe (0.6 pmol) was incubated with either 1 M Rv0678 or BSA for 30 min at room temperature in standard EMSA binding buffer. After incubation, 10 mM MgCl 2 and 5 mM CaCl 2 were added to the reaction mixture in a final volume of 50 l. Then 0.0025 units of DNase I (Thermo) was added and incubated for 5 min at room temperature. Digested DNA fragments were purified with QIAquick PCR purification columns (Qiagen) and eluted in 20 l of water. Digested DNA samples were analyzed at the Center for Genome Research and Biocomputing at Oregon State University. Purified DNA (2 ml) was mixed with HiDi formamide and GeneScan-500 LIZ size standards (Applied Biosystems) and analyzed using an Applied Biosystems 3730 DNA analyzer.
The 296-bp fragment was sequenced with the primers 6FAM-Rv0678-F and HEX-Rv0678-R, respectively, using the Thermo Sequenase dye primer manual cycle sequencing kit according to the manufacturer's instructions. Each reaction was diluted 5-fold in water, and 4 l was added to 5.98 l of HiDi formamide and 0.02 l of GeneScan-500 LIZ size standard. Samples were analyzed using the 3730 DNA analyzer, and electropherograms were aligned using the GENEMAPPER software (version 5.0, Applied Biosystems).
Site-directed Mutagenesis-Site-directed point mutations on residues Asp-90 and Arg-92, which are expected to be critical for DNA binding, were performed to generate the single point mutants D90A and R92A. The primers used for these mutations are listed in Table 3. All oligonucleotides were purchased from (Integrated DNA Technologies, Inc., Coralville, IA) in a salt-free grade.
Fluorescence Polarization Assay for DNA Binding-Fluorescence polarization assays were used to determine the affinity for DNA binding by Rv0678 and its mutants. Both the 26-bp oligodeoxynucleotide and fluorescein-labeled oligodeoxynucleotide were purchased from Integrated DNA Technologies, Inc. (Coralville, IA). These oligodeoxynucleotides contain the consensus 18-bp putative promoter DNA sequence (TTT-CAGAGTACAGTGAAA) for Rv0678. The sequences of the oligodeoxynucleotides were 5Ј-CAGATTTCAGAGTACAGT-GAAACTTG-3Ј and 5Ј-F-CAAGTTTCACTGTACTCTGAA-ATCTG-3Ј, where F denotes the fluorescein that was covalently attached to the 5Ј-end of the oligodeoxynucleotide by a hexamethylene linker. The 26-bp fluoresceinated dsDNA was prepared by annealing these two oligodeoxynucleotides together. The fluorescence polarization experiment was done using a DNA binding solution containing 10 mM sodium phosphate (pH 7.2), 100 mM NaCl, 5 nM fluoresceinated DNA, and 1 g of poly(dI-dC) as nonspecific DNA. The protein solution containing 2,500 nM dimeric Rv0678 or Rv0678 mutant and 5 nM fluoresceinated DNA was titrated into the DNA binding solution until the millipolarization became unchanged. All measurements were performed at 25°C using a PerkinElmer LS55 spectrofluorometer equipped with a Hamamatsu R928 photomultiplier. The excitation wavelength was 490 nm, and the fluorescence polarization signal (in ⌬P) was measured at 525 nm. Each titration point recorded was an average of 15 mea-

Structure of the Transcriptional Regulator Rv0678
surements. Data were analyzed using the equation, P ϭ ((P bound Ϫ P free )[protein]/(K D ϩ [protein])) ϩ P free , where P is the polarization measured at a given total protein concentration, P free is the initial polarization of free fluorescein-labeled DNA, P bound is the maximum polarization of specifically bound DNA, and [protein] is the protein concentration. The titration experiments were repeated three times to obtain the average K D value. Curve fitting was accomplished using the program ORIGIN (OriginLab Corp., Northampton, MA).

RESULTS AND DISCUSSION
Overall Structure of Rv0678-M. tuberculosis Rv0678 belongs to the MarR family of regulators. It possesses 165 amino acids, sharing 14 and 15% protein sequence identity with MarR (22) and OhrR (36) (Fig. 1). The crystal structure of Rv0678 was determined to a resolution of 1.64 Å using single isomorphous replacement with anomalous scattering (Table 1). Four molecules of Rv0678 are found in the asymmetric unit, which assemble as two independent dimers (Fig. 2). Superim-position of these two dimers gives a root mean square deviation of 0.8 Å over 271 C ␣ atoms, indicating that their conformations are nearly identical to each other.
The structure of Rv0678 (Fig. 3) is quite distinct in comparison with the known structures of the MarR family regulators (22, 36 -39). Each subunit of Rv0678 is composed of six ␣-helices and two ␤-strands: ␣1 (residues 17-31), ␣2 (residues 36 -47), ␣3 (residues 55-62), ␣4 (residues 66 -79), ␤1 (residues 82-85), ␤2 (residues 94 -97), ␣5 (residues 101-127), and ␣6 (residues 132-160) (Fig. 1). The monomer is L-shaped, with the shorter side forming a DNA-binding domain. However, the longer side contributes to an extended long arm, creating a dimerization domain for the regulator. Residues 34 -99, which include ␣2, ␣3, ␣4, ␤1, and ␤2, are responsible for constructing the DNA-binding domain. The dimerization domain of Rv0678 is generated by residues 16 -32 and 101-160, which cover ␣1, ␣5, and ␣6 of the protomer. Each protomer of Rv0678 is ϳ55 Å tall, 35 Å wide, and 35 Å thick. As a member of the MarR family of regulators, the DNAbinding domain of Rv0678 features a typical winged helix-turnhelix binding motif. The two anti-parallel ␤1 and ␤2 strands are found to generate a ␤-hairpin structure, which also forms the wing of the DNA-binding domain. The crystal structure of the OhrR-DNA complex (36) showed that this ␤-hairpin directly participates to contact the double-stranded DNA and is critical for repressor-operator interactions. Another important component of the winged helix-turn-helix motif for DNA recognition is helix ␣4. In the OhrR-DNA complex (36), the corresponding ␣-helix is found to bind within the deep major groove of the B-DNA. Protein sequence alignment suggests that Rv0678 contains three conserved amino acids common among members of the MarR family. These three residues, Arg-84, FIGURE 3. Structure of the M. tuberculosis Rv0678 regulator. a, ribbon diagram of a protomer of Rv0678. The molecule is colored using a rainbow gradient from the N terminus (blue) to the C terminus (red). b, ribbon diagram of the Rv0678 dimer. Each subunit of Rv0678 is labeled with a different color (yellow and orange). The bound 2-stearoylglycerol within the dimer is shown in sphere form (gray, carbon; red, oxygen). The figure was prepared using PyMOL.   Fig. 1b). The bound 2-stearoylglycerol is shown as a stick model (green, carbon; red, oxygen). The simulated annealing 2F o Ϫ F c electron density map is contoured at 1.2 (blue mesh). The left and right subunits of Rv0678 are shown as orange and yellow ribbons. b, the 2-stearoylglycerol binding site. Amino acid residues within 3.9 Å of the bound 2-stearoylglycerol (green, carbon; red, oxygen) are shown with one-letter codes. The side chains of selected residues from the right subunit of Rv0678 in Fig. 1b are shown as yellow sticks (yellow, carbon; blue, nitrogen; red, oxygen). Residues from the next subunit of Rv0678 are shown as orange sticks (orange, carbon; blue, nitrogen; red, oxygen). c, schematic representation of the Rv0678 and 2-stearoylglycerol interactions. Amino acid residues within 4.5 Å from the bound 2-stearoylglycerol are shown with one-letter codes. Dotted lines, hydrogen bonds. The hydrogen-bonded distances are also indicated. Asp-90, and Arg-92, are located within the DNA-binding domain of the regulator (Fig. 1) and are probably important for protein-DNA interactions. Among them, Asp-90 and Arg-92 are positioned within the ␤-hairpin of the wing. The corresponding amino acids located at the winged loop region of the ST1710 regulator play a major role in regulator-promoter interactions (39).
The Rv0678 crystal structure reveals that helices ␣1, ␣5, and ␣6 are involved in the formation of the dimer. Specifically, helices ␣5, ␣6, ␣5Ј, and ␣6Ј (where the prime denotes the next subunit) form intertwined helical bundles and constitute the dimerization domain. Helices ␣6 and ␣6Ј are oriented in an antiparallel fashion and form the scaffold of the dimer (Fig. 3). Extensive hydrophobic interactions are observed at the interface between the two subunits of the regulator. In addition, Tyr-147 and Tyr-159Ј and their identical counter pair perform aromatic stacking interactions, securing the dimeric organization. Additional salt bridges between Arg-32 and Glu-115Ј and between Glu-106 and Arg-109Ј (as well as their counter pairs) stabilize the binding.  Perhaps the most striking difference between the structures of Rv0678 and other MarR family members is the relative orientation of the DNA-binding and dimerization domains. The structures of MarR (22), OhrR (36), and MexR (37,38) suggest that helices ␣4 and ␣4Ј orient approximately perpendicular to the pseudo 2-fold axis of the dimeric regulators. However, our crystal structure of Rv0678 depicts these two helices more or less in parallel with the dimer's pseudo-2-fold axis. Similar orientation of helix ␣4 has also been found in the structure of the Vibrio cholerae AphA transcriptional activator (40). This conformation is not compatible and does not allow the regulator to interact with the B-form DNA. To bind its cognate DNA, the Rv0678 regulator must undergo a large conformational movement that reorients the DNA-binding domain such that the positions of helices ␣4 and ␣4Ј can be matched with the two consecutive major grooves of the promoter DNA. Based on the OhrR-DNA (36) and ST1710-DNA (39) crystal complexes, we predict that the entire DNA-binding domain of Rv0678, including ␣2, ␣3, ␣4, ␤1, and ␤2, has to rotate downward by ϳ70°with respect to ␣5 of the dimerization domain before DNA binding (Fig. 4). If this is the case, then the loop region between ␤2 and ␣5 forms the hinge for this rotational motion.
Rv0678 Was Liganded-Unexpectedly, a large extra electron density was found at the interface between the DNA-binding and dimerization domains of Rv0678, indicating the existence of a fortuitous bound ligand co-purified and co-crystallized with the regulator (Fig. 5). Thus, this region is also a substratebinding site. To identify the unknown bound ligand, GC-MS was applied to investigate the Rv0678 crystals (Fig. 6). The result suggests that the fortuitous ligand is 2-stearoylglycerol, also called octadecanoic acid, 2-hydroxyl-1-(hydroxymethyl)ethyl ester, which contains 21 carbons with the molecular formula C 21 H 42 O 4 . That this fatty acid glycerol ester is co-purified with the Rv0678 regulator suggests that fatty acid glycerol esters may be the natural substrates for this protein.
The propanetriol of the bound 2-stearoylglycerol is completely buried within the dimer interface, leaving the tail portion of its elongated octadecanoate hydrophobic carbon chain oriented at the entry point of this binding site. This orientation facilitates the contribution of Arg-32 and Glu-106Ј to form two hydrogen bonds with the glycerol headgroup of the fatty acid. The backbone oxygen of Phe-79Ј also participates to create the third hydrogen bond with this glycerol headgroup. In addition, the carbonyl oxygen of the octadecanoate group contributes to make another hydrogen bond with Arg-109, securing the binding. Interestingly, Rv0678 further anchors the bound fatty acid molecule through hydrophobic interactions with residues Phe-79, Phe-79Ј, and Phe-81Ј. Therefore, the binding of 2-stearoylglycerol in Rv0678 is extensive; within 4.5 Å of the bound fatty acid glycerol ester, 20 amino acids contact this molecule (Table  4). It should be noted that residues Phe-79, Phe-79Ј, and Phe-81Ј belong to helices ␣4 and ␣4Ј. In the OhrR-DNA structure (36), the corresponding ␣4 and ␣4Ј helices were buried within the two consecutive major grooves, directly contacting the promoter DNA. Thus, we suspect that helices ␣4 and ␣4Ј have dual

Structure of the Transcriptional Regulator Rv0678
responsibilities in the Rv0678 regulator. They form the DNAbinding site for operator DNA as well as the substrate-binding site for inducing ligands.
In the second Rv0678 dimer of the asymmetric unit, it is also found that a 2-stearoylglycerol molecule is bound within the corresponding substrate-binding site. Residues contributed to form this binding site are nearly identical but with a slightly different subset of amino acids in comparison with those of the first Rv0678 dimer described above (Table 4).
Virtual Ligand Library Screening-Virtual ligand screening was then performed to elucidate the nature of protein-ligand interactions in the Rv0678 regulator. The 2-stearoylglycerol binding site was chosen as a substrate binding cavity for this docking study. AutoDock Vina (32) was used to screen small molecules listed in the DrugBank (33) and ZINC (34) libraries. Vina utilizes the iterated local search global optimizer algorithm, which results in predicted binding free energies for these compounds ranging from Ϫ13.8 to ϩ20 kcal/mol. Of the 70,000 screened compounds, it is predicted that the best substrate for Rv0678 is the heterocyclic compound diethyl-[(5E)-5-(6,8,9,10tetrahydro-5H-benzo[c]xanthen-11-ylmethylene)-7,8-dihydro-6H-xanthen-3-yli. Table 5 lists the top three substrates, which have the lowest predicted binding free energies, for the Rv0678 regulator.
Because the crystal structure of Rv0678 shows that a fatty acid glycerol ester is bound within the substrate binding site of this regulator, Vina (32) was also used to examine whether these fatty acids are able to interact with Rv0678. As a positive control, the molecule 2-stearoylglycerol was docked into the substrate-binding site of this regulator, resulting in a predicted binding free energy of Ϫ7.6 kcal/mol. Vina was then used to screen for 2,500 different fatty acids. Based on the lowest predicted binding free energies, the top three compounds in this class was selected and listed in Table 6, where 18-[8-chloro-1-  mmpS2-mmpL2, mmpS4-mmpL4, mmpS5, and rv0991-2c. a, schematic depicting the DNA probes used in EMSAs to examine the promoter and intragenic regions of the mmpS2-mmpL2, mmpL3, mmpS4-mmpL4, mmpS5-mmpL5, and rv0991-2c genes. b, EMSAs were performed using 12 nM DIG-labeled probe and the indicated micromolar concentrations of protein. An arrow denotes the shifted probes. c, to demonstrate specificity, EMSAs were performed in the presence of non-labeled ("cold") probe. Reactions were performed with 6 nM DIG-labeled probe, the indicated micromolar concentrations of protein, and 0.6 M cold probe. *, accumulation of free DIG-labeled probe. d, EMSAs were performed using 12 M DIG-labeled probe and 6 M Rv0678 in the presence or absence of 1 M 1-stearoyl-rac-glycerol, as indicated above the blot. e, the sequence of the probes bound by Rv0678 in b and c were compared using the motif-based sequence analysis tool MEME, yielding a putative Rv0678 binding motif.
Rv0678-Ligand Interaction-The binding affinity of 1-stearoyl-rac-glycerol for the Rv0678 regulator was then determined using isothermal titration calorimetry, which obtained a binding affinity constant, K a , of 4.9 Ϯ 0.4 ϫ 10 5 M Ϫ1 . The titration is characterized by a negative enthalpic contribution, which gives rise to a hyperbolic binding curve (Fig. 7). The thermodynamic parameters of binding of 1-stearoyl-rac-glycerol to Rv0678 display enthalpic (⌬H) and entropic (⌬S) contributions of Ϫ1.0 Ϯ 0.1 kcal/mol and 22.5 cal⅐mol⅐degrees Ϫ1 , respectively. Interestingly, the molar ratio for this binding reaction based on isothermal titration calorimetry is one Rv0678 dimer/ligand. This ligand-binding experiment confirms that Rv0678 is capable of recognizing fatty acid glycerol esters.
Electrophoretic Mobility Shift Assay-To demonstrate direct transcriptional regulation, we performed EMSAs using a probe corresponding to the intergenic region between mmpS5 and rv0678 (Fig. 8a). This probe shifted in a concentration-dependent manner (Fig. 8b). This result is consistent with previous reports of altered mmpS5/mmpL5 gene expression in Mycobacterium bovis BCG spontaneous rv0678 mutants (13). Preliminary CHIPSeq data from the TB Systems Biology Consortium suggests that Rv0678 regulates the expression of additional genes (41). We designed additional probes to experimentally demonstrate binding of Rv0678 to the promoter regions of mmpS2-mmpL2, mmpS4-mmpL4, and rv0991-0992. Probes are depicted schematically in Fig. 8a. We also saw concentration-dependent binding of Rv0678 to these two probes (Fig. 8b). As a control, EMSAs were performed in the presence of non-labeled probes. Release of DIG-labeled probe was observed consistent with specific binding of Rv0678 to the rv0678-mmpS5, rv0505-mmpS2, and mmpL4 probes (Fig. 8c). Using the sequence of the six probes that shifted, we identified a putative consensus binding sequence for Rv0678 using the MEME algorithm (17) (Fig. 8e). Rv0678 co-crystallized with a ligand whose binding renders the protein unable to bind DNA. The addition of 1-stearoyl-rac-glycerol (an isomer of 2stearoylglycerol) to the EMSA reaction buffer reduced Rv0678 binding to a target promoter probe (Fig. 8c).
Dye Primer-based DNase I Footprint Assay-To further refine the binding site of Rv0678 in the rv0678-mmpS5 intergenic region, a DNase I footprint assay was performed on the Rv0678-mmpS5 probe using established methods (35). Electropherograms in Fig. 9 show the DNA sequence bound by Rv0678. The control protein BSA did not result in DNA protection at the same concentration. Interestingly, the region bound by Rv0678 includes the start codon of the rv0678 gene (underlined nucleotides in Fig. 9b). The bound sequence contains a potential inverted repeat motif (GAACGTCACA-GATTTCA . . . N 8 . . . TGAAACTTGTGAGCGTCAAC).
Rv0678-DNA Interaction-A fluorescence polarizationbased assay was carried out to study the interaction between Rv0678 and the 26-bp DNA containing the 18-bp putative promoter DNA sequence (TTTCAGAGTACAGTGAAA). Our footprint assay has suggested that this promoter DNA sequence was protected by the Rv0678 regulator. Fig. 10a illustrates the binding isotherm of Rv0678 in the presence of 5 nM fluoresceinated DNA. The titration experiment indicated that this regulator binds the 26-bp promoter DNA with a dissociation constant, K D , of 19.6 Ϯ 3.0 nM. The binding data also indicate that Rv0678 binds its cognate DNA with a stoichiometry of one Rv0678 dimer per dsDNA.
In addition, fluorescence polarization was used to determine the binding affinities of this 26-bp DNA by the Rv0678 mutants D90A and R92A. These two residues are located within the ␤-hairpin of the winged helix-turn-helix motif of the N-terminal DNA-binding domain. In ST1710, the corresponding two residues are critical for regulator-promoter interactions. Interestingly, our measurements indicate that the K D values of the D90A-DNA and R92A-DNA complexes are 113.3 Ϯ 16.8 and 86.0 Ϯ 7.4 nM (Fig. 10, b and c), revealing that the DNA binding affinities for these mutants are significantly weaker than that of the native Rv0678 regulator. Like ST1710, our experimental results suggest that residues Asp-90 and Arg-92 are important for DNA recognition.
With the rising incidence of drug resistant strains of M. tuberculosis, it is increasingly important to understand the molecular mechanisms underlying virulence and drug resist- ance of this pathogen. This knowledge will inform the development of new strategies to combat TB. In this report, we describe the crystal structure the Rv0678 transcriptional regulator, which controls the expression level of the MmpS5-MmpL5, MmpS4-MmpL4, and MmpS2-MmpL2 transport systems. MmpS4 and MmpS5 contribute to siderophore export, but the substrate of MmpL2 is not known (15). Fortuitously, the structure of Rv0678 was resolved in complex with a 2-stearoylglycerol molecule, suggesting that fatty acid glycerol esters are the natural substrates for the Rv0678 transcriptional regulator. Further work is required to demonstrate whether this ligand is structurally related to the substrate of either efflux system or how its availability changes in different environments and mycobacterial growth phases. The crystal structure of the 2-stearoylglycerol-Rv0678 complex probably provides a snapshot of the ligand-binding state of this regulator, whereby both the DNA-binding and dimerization domains are recruited to participate in ligand binding. In this case, the DNA-binding domain must bend upward and shift toward the dimerization domain to accommodate the bound ligand. As crystallized, the regulator is incompatible with the operator DNA. When the inducing ligand is removed from the ligand-binding site, freeing helices ␣4 and ␣4Ј to rotate downward and shift away from the dimerization domain, this conformational state should be compatible with the B-DNA and allow for DNA binding.