Biochemical and crystallographic studies on Mycobacterium tuberculosis 3-hydroxyisobutyric acid dehydrogenase (MtHIBADH), a member of the 3-hydroxyacid dehydrogenase superfamily, have been carried out. Gel filtration and blue native PAGE of MtHIBADH show that the enzyme is a dimer. The enzyme preferentially uses NAD+ as the cofactor and is specific to S-hydroxyisobutyric acid (HIBA). It can also use R-HIBA, l-serine and 3-hydroxypropanoic acid (3-HP) as substrates, but with much less efficiency. The pH optimum for activity is ∼11. Structures of the native enzyme, the holoenzyme, binary complexes with NAD+, S-HIBA, R-HIBA, l-serine and 3-HP and ternary complexes involving the substrates and NAD+ have been determined. None of the already known structures of HIBADH contain a substrate molecule at the binding site. The structures reported here provide for the first time, among other things, a clear indication of the location and interactions of the substrates at the active site. They also define the entrance of the substrates to the active site region. The structures provide information on the role of specific residues at the active site and the entrance. The results obtained from crystal structures are consistent with solution studies including mutational analysis. They lead to the proposal of a plausible mechanism of the action of the enzyme.

Introduction

3-Hydroxyacid dehydrogenases constitute a large structurally conserved superfamily of enzymes, showing substrate specificity for ligands which have 3-OH and 1-COOH groups, in which the hydroxyl group acts as a donor and a cofactor (NAD+ or NADP+) functions as an acceptor in the oxidoreductase pair [1]. This superfamily has evolved from short-chain alcohol dehydrogenase. The members of the superfamily characterized so far include 3-hydroxyisobutyrate dehydrogenase (HIBADH), 6-phosphogluconate dehydrogenase (6-PGDH), d-phenylserine dehydrogenase, tartronate semialdehyde reductase, 2-(hydroxymethyl)glutarate dehydrogenase and serine dehydrogenase. These enzymes are crucial in amino acid catabolism [2], hepatic gluconeogenesis [3], pentose phosphate pathway [4], glycerate biosynthesis [5], nicotinate fermentation [6] and l-serine metabolism [7].

HIBADH (EC 1.1.1.31) catalyzes the sixth step in the valine catabolism pathway and converts the central metabolite S-hydroxyisobutyrate (3-hydroxy-2-methyl propanoate) reversibly to S-methylmalonate semialdehyde (3-oxo-2-methyl propanoate) in the presence of the cofactor NAD+ or NADP+ in a metal-independent manner [8]. Human HIBADH is expressed in the mitochondria of the liver, kidney and muscle [9]. The end product of valine catabolism, propionyl CoA, enters the tricarboxylic acid (TCA) cycle and generates glucose through gluconeogenesis [10]. The involvement of the enzyme has also been predicted in generation of energy via valine intake in cultured neuronal cells [11]. Its expression has been shown to be up-regulated in the mid-piece of sperm, which has crucial role in motility [12]. Functional compromise of human HIBADH results in elevated level of the substrate HIBA in serum and urine, resulting in hydroxyisobutyric aciduria, methylmalonic acidemia and several brain disorders [1315]. Although HIBADH, in general, has been known for its participation in valine catabolism, in Pseudomonas aeruginosa, it has been reported to use l-serine as its substrate, thus participating in l-serine catabolism [7].

HIBADH enzymes from different organisms characterized till date follow a similar fold. The structural fold consists of an N-terminal (N-) domain with a Rossmann fold and a C-terminal (C-) domain with all α-helices, connected by a long and rigid interconnecting helix [7,16]. HIBADH enzyme consists of four well-defined functional motifs that have broadly defined roles in cofactor binding, substrate binding and catalysis. HIBADHs from different sources exhibit a similar fold, yet show remarkable differences in their substrate selectivity and specificity [7,16]. HIBADH purified from pig kidney and human liver were shown to be highly specific for S-hydroxyisobutyric acid (S-HIBA); however, the enzyme purified from Candida rugosa and rabbit liver utilizes both R and S isoforms of HIBA as substrates [8,10,16]. HIBADH from P. aeruginosa shows selectivity for l-serine over HIBA and that from P. putida shows selectivity for 2-methyl-dl-serine [17]. HIBADH from Thermus thermophilus can utilize l-serine, d-serine and HIBA (R- and S-HIBA) as substrates, irrespective of their chirality [18]. Structural studies carried out so far on the apo or the cofactor bound forms of HIBADHs from Homo sapiens (PDB ID: 2GF2 and 2I9P), P. aeruginosa [7], T. thermophilus [18], Salmonella typhimurium (PDB ID: 3G0O) and Bacillus cereus [19] are available. In each of the above structures, the presumed substrate-binding pocket is occupied by one of the components of the crystallization medium. Cofactor bound crystal structures indicate the sequential ordered mechanism where substrate appears to follow cofactor binding. While the order of substrate and the cofactor binding for HIBADH from other bacterial sources are not well characterized, an ordered Bi-Bi kinetic mechanism has been shown biochemically for human HIBADH enzyme [10].

HIBADH, annotated as mmsB gene in Mycobacterium tuberculosis (Rv0751c), is present in a single operon along with probable fadE9 (Rv0752c) and mmsA (Rv0753c) genes and is regulated under the influence of SigG factor under stress condition [20,21]. At pH 5.7, growth of Mycobacterium species ceases, resulting in the shift of metabolic flux. The individual genes (Rv0751c, Rv0752c and Rv0753c) are nearly two-fold up-regulated at pH 7 than at pH 5.7 [22]. Though the valine metabolism is well studied, knowledge of valine catabolism as a source of energy during stress condition remains unexplored [23,24].

Here, we report the biochemical and structural characterization of HIBADH from M. tuberculosis and its interactions with the cofactor and substrates/substrate analogs such as S-HIBA, R-HIBA, l-serine and 3-HP. We show that the enzyme is dimeric and the preferred cofactor is NAD+. The enzyme follows a sequential ordered Bi-Bi kinetic mechanism where cofactor binds first. The structures reported here are the only ones in which substrate molecules have been located. The crystal structures provide a detailed picture of the active site and the path to it. The interactions observed in the structures provide a structural rationale for the activity of the enzyme against different substrates/substrate analogs. The rationale is also supported by mutational analysis. Thus, the results presented here provide a comprehensive description of the structure and action of MtHIBADH.

Experimental procedures

Construction of MtHIBADH clone and site-directed mutagenesis

Gene-specific primers were designed (Supplementary Table S1) based on the gene sequence of Rv0751c (MtHIBADH) open reading frame in Tuberculist [25]. Full-length Rv0751c was amplified by PCR using genomic DNA of M. tuberculosis H37Rv obtained from BEI Resources. PCR amplicons and pET22b(+) vector (Novagen) were digested with NdeI and XhoI restriction enzymes (NEB), followed by ligation and screening to obtain the MtHIBADH clone. The gene sequence was further confirmed by sequencing.

Point mutations were generated by the single primer method [26]. PCR was performed using the MtHIBADH construct as a template and a mutant single primer for extension (Supplementary Table S1). PCR was followed by digestion with the DpnI restriction enzyme (NEB) to cleave the parental plasmid DNA. The digested mixture was transformed in Escherichia coli DH5α strain (Promega) and the mutant colonies were screened by plasmid isolation, followed by digestion. The mutations were confirmed by sequencing.

Protein expression and purification

The protein was purified following the standard protocol for Ni-NTA affinity chromatography using E. coli BL21 (λDE3) (Promega) as host cells. In brief, competent cells were transformed with the recombinant clone and transformed colonies were selected on LB agar plate containing 100 µg/ml ampicillin (Sigma) as an antibiotic. A single colony was inoculated as primary culture, followed by large-scale culture in 1 l of LB media. The culture was induced at A600 of 0.6 with 0.4 mM IPTG and was incubated further for 16 h at 22°C. The cells were harvested by centrifugation at 6000 rpm and resuspended in lysis buffer [20 mM Tris–HCl (pH 8), 500 mM NaCl, 5% glycerol and 1 mM PMSF]. The cells were lysed by sonication at 4°C and the cell debris was removed by centrifugation at 14 000 g. The collected supernatant was allowed to bind to the pre-equilibrated Ni-NTA resin (GE Healthcare) followed by washing with buffer [20 mM Tris–HCl (pH 8), 150 mM NaCl, 5% glycerol, 1 mM PMSF and 10 mM imidazole]. The protein fractions were eluted with a gradient (20–200 mM) of imidazole. The purified fractions were pooled and the purity of the protein was checked on 12% SDS–PAGE. The pooled fractions were dialyzed against buffer [20 mM Tris–HCl (pH 8), 150 mM NaCl and 5% glycerol]. The same protocol was followed for purification of mutants.

Electrospray ionization mass spectrometry

The protein was subjected to liquid chromatography (Agilent HP1100 Series), wherein the sample was passed through a reverse-phase column (Agilent Poroshell 120, SB C8, 4.6 mm × 150 mm, 2.7 µm) attached to an electrospray ionization ion trap mass spectrometer. A linear gradient elution was performed with H2O/ACN/0.1% formic acid for 60 min at a flow rate of 0.2 ml/min. The eluted protein was analyzed on the mass spectrometer, HCT Ultra PTM Discovery (Bruker Daltonics), which houses a classic ion trap (Paul type) with following parameters: nebulizer pressure of 30 psi, dry gas at a flow rate of 10 l/min and dry temperature at 330°C.

Size exclusion chromatography

Oligomeric state of the protein was determined by size exclusion chromatography using a Superdex 200, 10/300 column (GE Healthcare) connected to AKTA avant 25 FPLC (GE Healthcare). The column was first equilibrated using buffer [20 mM Tris–HCl (pH 8), 150 mM NaCl and 5% glycerol], followed by calibration with molecular mass standards (Bio-Rad): thyroglobulin (bovine), 670 kDa; γ-globulin (bovine), 158 kDa; ovalbumin (chicken), 44 kDa; myoglobin (horse), 17 kDa and vitamin B12, 1.35 kDa. The void volume of the column was determined by the elution volume of thyroglobulin and calibrated by plotting log molecular weight (Mw) of proteins versus their elution volumes. For calibration, 100 µl of protein aliquot (4 mg/ml) was loaded onto the column in a separate run.

Blue native PAGE

Bis–tris-based blue native PAGE was prepared and electrophoresis was performed using the modified protocol [27]. A 5× sample buffer [250 mM Tricin, 250 mM Bis–tris (pH 7.7), 40% glycerol and 0.1% Coomassie brilliant blue G250] was added to protein samples and loaded onto 10% BN gel and allowed to be electrophoresed at 4°C for 5 h at 70 V. Cathode buffer was composed of 15 mM Bis–tris, 50 mM Tricine and 0.02% Coomassie blue G250. A total of 50 mM Bis–tris was used as anode buffer. BSA (monomer, 66 kDa; dimer, 132 kDa) was used as a marker.

Dehydrogenase activity assay

Activities of the native and mutant enzymes with S-HIBA and NAD+ were determined at 30°C in reaction buffer consisting of 100 mM Tris–HCl (pH 8.5) in the presence of 1–10 µM of enzyme, by measuring A340 of NADH produced by reduction of NAD+ at Varioskan™ Flash Multimode Reader (Thermo Fisher). Steady-state kinetic parameters of the cofactors were determined at saturating concentration of S-HIBA (8 mM) and with concentration of NAD+/NADP+ cofactor varying between 0.2 and 6 mM. Likewise, the concentration of NAD+ was kept constant at 8 mM and that of S-HIBA was varied between 0.2 and 5 mM, to determine the parameters for the substrate. In both the sets of experiments, the enzyme concentration used was 1 µM. For other putative substrates, the saturating substrate concentrations were determined in separate activity assays and the kinetic assays were performed till concentrations of 10 mM for R-HIBA; 400 mM for l-serine; 3 M for d-serine; 20 mM for 3-HP; 600 mM for S-HIBA methyl ester and 250 mM for R-HIBA methyl esters. The rate of product formation was expressed as NADH produced (mM/min/mg or µmol/min/mg) as calculated using the extinction coefficient for NADH at 340 nm (6220 M−1 cm−1). Steady-state kinetic parameters (Km, kcat and Vmax) were determined by fitting the data to the Michaelis–Menten equation with nonlinear least-square regression with GraphPad Prism 6 software.

The enzyme assay at different pH values was performed in a buffer mixture of succinic acid, NaH2PO4 and glycine (2 : 7 : 7) [28], and specific activity was calculated at varying pH using the above standard reaction mixture with NAD+ as a cofactor and S-HIBA as a substrate.

Dynamic scanning fluorimetry

The melting temperature of MtHIBADH at different pHs was determined by the dynamic scanning fluorimetry (DSF) method [29] using a fluorescence microplate reader (iQ5, Bio-Rad iCycler Multicolor real-time PCR detection system). A sample volume of 25 µl of solution (a buffer mixture of succinic acid, NaH2PO4 and glycine; 2 : 7 : 7) at different pHs containing 2.5 µM protein was mixed with 5× Sypro Orange dye. Fluorescence intensities were measured at intervals of 0.5°C over a temperature range of 25–90°C, with a dwell time of 1 min. Each experiment was repeated three times. Fluorescence intensities were plotted as a function of temperature and the plots were normalized to lie between 0 and 1. Melting temperature of the protein was determined by the Boltzmann sigmoidal dose response function using GraphPad Prism 6. For WT and W211A mutant proteins, a similar procedure was followed with 2.5 µM protein concentration in a buffer of 20 mM Tris–HCl (pH 8), 150 mM NaCl and 5% glycerol.

Preparation of crystals

The purified enzyme was dialyzed against a buffer consisting of 20 mM Tris–HCl (pH 7.5), 100 mM NaCl and 5% glycerol and was concentrated to 12 mg/ml for crystallization trials. Initial screening for crystallization was carried out with Crystal Screen, Crystal Screen 2 and Index Screen (Hampton Research, Aliso Viejo, CA, U.S.A.) using the microbatch-under-oil method. Crystals grew in 3–5 days when index screen reagent no. 59 [0.02 M MgCl2·6H2O, 0.1 M HEPES (pH 7.5), 22% (w/v) polyacrylic acid sodium salt 5100] was used as the precipitant. Binary complexes with S-HIBA, R-HIBA, l-serine and 3-HP were prepared by soaking the crystals in mother liquor containing 10 mM of the appropriate ligands. Ternary complexes were likewise obtained by soaking the crystals of the holoenzyme in the appropriate substrate solution for 3–5 min. Longer soaking resulted in the disintegration of the crystal. Therefore, crystals were frozen after 3–5 min of soaking.

Data collection and processing

Crystals were soaked for 10–20 s in a cryoprotectant (the crystallization solution with 10% glycerol) and flash-cooled at 100 K in a stream of cold nitrogen. X-ray diffraction data from all crystals, except one (S-HIBA + NAD+), were collected at a home source, on a Rigaku ultraX 18 CuKα rotating anode generator using an MAR345 image plate. The crystal to detector distance was kept at 130–150 mm with an oscillation angle of 1° per frame. The diffraction data from the ternary complex involving NAD+ and S-HIBA were collected at the BM-14 beamline of ESRF, Grenoble, with a crystal to detector distance of 187.4 mm and an oscillation angle of 0.5° per frame. The images were processed using MOSFLM [30] and scaled using SCALA [31] from the CCP4 program suite [32].

Structure determination and refinement

The structure of a subunit of l-serine dehydrogenase from P. aeruginosa (PDB ID 3OBB) was used as the search model to solve the structure of the apoenzyme employing the molecular replacement method using Phaser [33,34] from the CCP4 program suite. Phaser gave a single solution for two subunits in the asymmetric unit. The structure of the apoenzyme was used to analyze the remaining nine isomorphous crystals. Matthews's coefficients of the crystals ranged between 2.0 and 2.2 [35]. The solutions obtained from Phaser were then built into the electron density manually using COOT [36] before refinement. An initial rigid-body refinement followed by positional refinement was performed on each structure using REFMAC5 [37] from the CCP4 program suite. Refinement and model building were continued till R and Rfree converged. A few amino acid side chains having alternative conformations were assigned fractional occupancies. Some of the bound NAD+ molecules were also assigned fractional occupancies, so that their displacement parameters are comparable to those of molecules with full occupancies. Ligands were fitted using difference maps. The presence of ligands was further confirmed by simulated-annealing Fo− Fc OMIT maps calculated using CNS v.1.3 [38]. Water molecules were built into the model based on peaks greater than 1σ in 2Fo− Fc and 3σ in Fo − Fc maps. The refinement parameters, along with data collection statistics, are given in Table 2.

Analysis of the structures

The refined models were evaluated using PROCHECK [39]. Structural superpositions were carried out using ALIGN [40] and figures were generated using PyMOL and CHIMERA [41,42]. Interatomic distances were calculated using CONTACT from the CCP4 program suite. Hydrogen bonds were assigned using HBPLUS [43] based on a distance less than or equal to 3.6 Å between the donor (D) and the acceptor (A) atoms and a D-H---A angle greater than 90°. Energy minimization of the protein–ligand (model) complex was performed using the GROMACS v.5.0.7 package with Gromos96 force field [44]. The energy minimization was carried out for 500 steps of the steepest descent method.

Statistical analysis

All the kinetic experiments were performed with three different protein preparations. Statistical analyses were conducted using Graphpad Prism 6 software. Each datum represents the mean ± SD of a minimum three measurements in each experiment.

Results and discussion

Biochemical characterization

SDS–PAGE demonstrates that the expressed recombinant protein is pure (Supplementary Figure S1A) and mass spectroscopic analysis gives the molecular mass as 30.875 kDa, which agrees well with the molecular mass calculated from the sequence (Supplementary Figure S1B). Size exclusion chromatography indicates the protein to be a dimer with Mr 60 000 (Figure 1A). Blue native PAGE also indicates the protein to be a dimer (Figure 1B). Steady-state kinetic parameters obtained from the Michaelis–Menten equation using NAD+ and NADP+ as cofactors and S-HIBA as the substrate yield four-fold higher kcat/Km values with NAD+ compared with that of NADP+, demonstrating the preference of the enzyme for NAD+ (Table 1). Subsequent experiments were therefore carried out with NAD+ as the cofactor. The measured activities of the holoenzyme against different possible substrates (Michaelis–Menten kinetic curves shown in Supplementary Figure S1C) illustrate that the enzyme exhibits 40-fold higher activity with S-HIBA than it does against R-HIBA, indicating substrate specificity for S-chirality (Figure 1C and Table 1). l-serine, which has a conformation similar to that of S-HIBA, but in which the methyl group is replaced by an amino group (Figure 1D), is only as good a substrate as R-HIBA. The activity of the enzyme against 3-HP, which is achiral due to the removal of the methyl group at the asymmetric carbon, is much lower than that against S-HIBA but higher than those against R-HIBA and l-serine (Figure 1C and Table 1). The enzyme did not show any detectable activity against the other possible substrates examined. Only compounds against which the enzyme showed detectable activity, namely S-HIBA, R-HIBA, l-serine and 3-HP, were used in subsequent crystallographic investigations.

Biochemical characterization of MtHIBADH.

Figure 1.
Biochemical characterization of MtHIBADH.

(A) Size exclusion chromatography. (B) 10% Blue native PAGE with varying MtHIBADH amounts (2.5–20 µg). (C) Enzymatic activity against different substrates with NAD+ as the cofactor. (D) Chemical structures of substrates and corresponding substrate analogs.

Figure 1.
Biochemical characterization of MtHIBADH.

(A) Size exclusion chromatography. (B) 10% Blue native PAGE with varying MtHIBADH amounts (2.5–20 µg). (C) Enzymatic activity against different substrates with NAD+ as the cofactor. (D) Chemical structures of substrates and corresponding substrate analogs.

Table 1
Kinetic parameters of cofactors and substrates
Ligands Km (mM) kcat (s−1kcat/Km (M−1 s−1
S-HIBA (8 mM) as a substrate 
 NAD+ 0.563 ± 0.058 1.182 ± 0.045 2.1 × 103 
 NADP+ 0.582 ± 0.18 0.29 ± 0.041 0.498 × 103 
NAD+ (8 mM) as a cofactor 
S-HIBA 0.574 ± 0.097 1.018 ± 0.037 1.773 × 103 
R-HIBA 4.56 ± 1.02 0.205 ± 0.027 0.0448 × 103 
l-serine 21.33 ± 1.3 0.94 ± 0.019 0.044 × 103 
d-serine 579.1 ± 120 0.323 ± 0.120 0.558 
 3-HP 2.53 ± 0.217 0.262 ± 0.063 0.1036 × 103 
S-HIBA methyl ester 31.63 ± 5.8 0.020 ± 0.001 0.632 
R-HIBA methyl ester 11.87 ± 3.7 0.002 ± 0.0002 0.168 
Ligands Km (mM) kcat (s−1kcat/Km (M−1 s−1
S-HIBA (8 mM) as a substrate 
 NAD+ 0.563 ± 0.058 1.182 ± 0.045 2.1 × 103 
 NADP+ 0.582 ± 0.18 0.29 ± 0.041 0.498 × 103 
NAD+ (8 mM) as a cofactor 
S-HIBA 0.574 ± 0.097 1.018 ± 0.037 1.773 × 103 
R-HIBA 4.56 ± 1.02 0.205 ± 0.027 0.0448 × 103 
l-serine 21.33 ± 1.3 0.94 ± 0.019 0.044 × 103 
d-serine 579.1 ± 120 0.323 ± 0.120 0.558 
 3-HP 2.53 ± 0.217 0.262 ± 0.063 0.1036 × 103 
S-HIBA methyl ester 31.63 ± 5.8 0.020 ± 0.001 0.632 
R-HIBA methyl ester 11.87 ± 3.7 0.002 ± 0.0002 0.168 
Table 2
Summary of data collection and structure refinement

All crystals contained a dimeric molecule in the asymmetric unit. The space group of all the crystals is P65. Values in parentheses are for the highest resolution shell.

 Native NAD+ S-HIBA R-HIBA l-serine 3-HP S-HIBA +  NAD+ R-HIBA +  NAD+ l-serine +  NAD+ 3-HP +  NAD+ 
PDB code 5Y8G 5Y8H 5Y8I 5Y8J 5Y8K 5Y8P 5Y8L 5Y8M 5Y8N 5Y8O 
Data collection statistics 
 Unit cell parameters (Å) 
  a (Å) 128.79 129.43 128.58 128.53 129.40 128.92 128.79 129.02 129.16 129.06 
  b (Å) 128.79 129.43 128.58 128.53 129.40 128.92 128.79 129.02 129.16 129.06 
  c (Å) 70.35 70.49 70.23 70.20 70.57 70.19 70.22 70.30 70.43 70.13 
 Resolution range (Å) 111.5–2.01 (2.12–2.01) 112.1–2.1 (2.21–2.1) 111.3–2.04 (2.15–2.04) 111.3–1.86 (1.96–1.86) 112.0–2.04 (2.15–2.04) 111.6–2.15 (2.27–2.15) 111.5–1.85 (1.95–1.85) 111.73–2.04 (2.15–2.04) 111.8–2.68 (2.82–2.68) 111.7–2.05 (2.16–2.0) 
 Solvent content (%) 56.0 56.5 55.9 55.7 56.5 55.9 55.5 56.1 56.2 56.0 
 No. of observed reflection 363 277 (47 301) 344 993 (47 204) 326 373 (36 024) 527 566 (63 441) 405 525 (49 135) 279 645 (30 640) 300 869 (43 419) 406 508 (56 705) 184 111 (17 528) 150 836 (19 478) 
 No. of unique reflections 44 342 (6417) 38 867 (5503) 41 507 (5917) 52 774 (5917) 42 645 (6105) 33 570 (4407) 56 709 (8247) 42 572 (6189) 18 305 (2553) 41 484 (6092) 
 Multiplicity 8.2 (7.4) 8.9 (8.6) 7.9 (6.1) 10.0 (8.7) 9.5 (8.0) 8.3 (7.0) 5.3 (5.3) 9.5 (9.2) 10.1 (6.8) 3.6 (3.2) 
 Completeness (%) 100 (99.9) 98.6 (96.5) 98.2 (96.6) 94.9 (89.8) 99.2 (97.8) 92.5 (83.3) 100 (100) 99.9 (99.7) 96.4 (92.9) 99.0 (99.6) 
I/(σI10.1 (2.7) 9.5 (2.5) 8.1 (2.1) 14.3 (2.3) 9.5 (2.2) 7.4 (2.1) 7.6 (1.8) 11.2 (2.5) 7.0 (2.4) 6.0 (1.2) 
Rmerge1 (%) 14.4 (75.7) 17.4 (91.7) 17.7 (97.6) 12.7 (99.0) 17.6 (92.0) 22.3 (96.2) 16.1 (92.5) 18.0 (93.5) 27.6 (94.9) 16.0 (94.2) 
Refinement and model statistics 
R-factor (%) 14.4 15.9 16.6 14.5 16.7 16.7 14.2 15.2 18.1 17.6 
Rfree (%) 18.5 21.1 20.7 18.4 20.6 21.6 18.3 19.8 24.1 22.5 
R.m.s. deviation from ideal values 
 Bond length (Å) 0.019 0.016 0.008 0.021 0.008 0.016 0.019 0.018 0.010 0.014 
 Bond angle (°) 1.8 1.8 1.3 1.9 1.4 1.7 1.9 1.9 1.3 1.6 
Residues (%) in Ramachandran plot 
 Favored region (%) 96.7 95.6 96.7 96.8 96.3 96.1 97.2 96.3 95.6 96.8 
 Allowed region (%) 2.9 3.9 3.3 2.9 2.9 3.4 2.2 3.4 3.9 2.9 
 Disallowed region (%) 0.3 0.3 0.0 0.1 0.6 0.3 0.5 0.1 0.3 0.1 
 Native NAD+ S-HIBA R-HIBA l-serine 3-HP S-HIBA +  NAD+ R-HIBA +  NAD+ l-serine +  NAD+ 3-HP +  NAD+ 
PDB code 5Y8G 5Y8H 5Y8I 5Y8J 5Y8K 5Y8P 5Y8L 5Y8M 5Y8N 5Y8O 
Data collection statistics 
 Unit cell parameters (Å) 
  a (Å) 128.79 129.43 128.58 128.53 129.40 128.92 128.79 129.02 129.16 129.06 
  b (Å) 128.79 129.43 128.58 128.53 129.40 128.92 128.79 129.02 129.16 129.06 
  c (Å) 70.35 70.49 70.23 70.20 70.57 70.19 70.22 70.30 70.43 70.13 
 Resolution range (Å) 111.5–2.01 (2.12–2.01) 112.1–2.1 (2.21–2.1) 111.3–2.04 (2.15–2.04) 111.3–1.86 (1.96–1.86) 112.0–2.04 (2.15–2.04) 111.6–2.15 (2.27–2.15) 111.5–1.85 (1.95–1.85) 111.73–2.04 (2.15–2.04) 111.8–2.68 (2.82–2.68) 111.7–2.05 (2.16–2.0) 
 Solvent content (%) 56.0 56.5 55.9 55.7 56.5 55.9 55.5 56.1 56.2 56.0 
 No. of observed reflection 363 277 (47 301) 344 993 (47 204) 326 373 (36 024) 527 566 (63 441) 405 525 (49 135) 279 645 (30 640) 300 869 (43 419) 406 508 (56 705) 184 111 (17 528) 150 836 (19 478) 
 No. of unique reflections 44 342 (6417) 38 867 (5503) 41 507 (5917) 52 774 (5917) 42 645 (6105) 33 570 (4407) 56 709 (8247) 42 572 (6189) 18 305 (2553) 41 484 (6092) 
 Multiplicity 8.2 (7.4) 8.9 (8.6) 7.9 (6.1) 10.0 (8.7) 9.5 (8.0) 8.3 (7.0) 5.3 (5.3) 9.5 (9.2) 10.1 (6.8) 3.6 (3.2) 
 Completeness (%) 100 (99.9) 98.6 (96.5) 98.2 (96.6) 94.9 (89.8) 99.2 (97.8) 92.5 (83.3) 100 (100) 99.9 (99.7) 96.4 (92.9) 99.0 (99.6) 
I/(σI10.1 (2.7) 9.5 (2.5) 8.1 (2.1) 14.3 (2.3) 9.5 (2.2) 7.4 (2.1) 7.6 (1.8) 11.2 (2.5) 7.0 (2.4) 6.0 (1.2) 
Rmerge1 (%) 14.4 (75.7) 17.4 (91.7) 17.7 (97.6) 12.7 (99.0) 17.6 (92.0) 22.3 (96.2) 16.1 (92.5) 18.0 (93.5) 27.6 (94.9) 16.0 (94.2) 
Refinement and model statistics 
R-factor (%) 14.4 15.9 16.6 14.5 16.7 16.7 14.2 15.2 18.1 17.6 
Rfree (%) 18.5 21.1 20.7 18.4 20.6 21.6 18.3 19.8 24.1 22.5 
R.m.s. deviation from ideal values 
 Bond length (Å) 0.019 0.016 0.008 0.021 0.008 0.016 0.019 0.018 0.010 0.014 
 Bond angle (°) 1.8 1.8 1.3 1.9 1.4 1.7 1.9 1.9 1.3 1.6 
Residues (%) in Ramachandran plot 
 Favored region (%) 96.7 95.6 96.7 96.8 96.3 96.1 97.2 96.3 95.6 96.8 
 Allowed region (%) 2.9 3.9 3.3 2.9 2.9 3.4 2.2 3.4 3.9 2.9 
 Disallowed region (%) 0.3 0.3 0.0 0.1 0.6 0.3 0.5 0.1 0.3 0.1 
1

, where Ii(hkl) is the ith observation of reflection hkl and 〈I(hkl)〉 is the weighted average intensity for all i observations of reflection hkl.

Assays carried out at different pHs indicate that the pH optimum for activity is ∼11 (Supplementary Figure S2A), suggesting the involvement of a highly basic residue in catalysis. The stability of the protein as a function of pH was also explored using DSF. The protein remains stable up to pH 9 (Supplementary Figure S2B). The protein begins to exhibit instability when the pH of the medium is increased further. At pH 11, the melting temperature is as low as 34.5°C. A further increase in the pH of the medium results in the precipitation of the protein. Thus, while activity increases with pH, action of the protein becomes unviable on account of its instability at pH ∼11.

The reaction mechanism of MtHIBADH was analyzed based on reciprocal plots of reaction velocities with varying concentrations of S-HIBA (0.2–2 mM) which followed intersecting lines with Vmax of the reaction kinetics being same with varying NAD+ concentrations (0.2–2 mM) (Supplementary Figure S3A). This indicates a sequential reaction mechanism. The enzyme showed competitive inhibition in the presence of NADH when NAD+ was varied and showed noncompetitive inhibition with NADH when S-HIBA was varied (Supplementary Figure S3B,C). The data profile indicates an ordered Bi–Bi reaction mechanism with NAD+ binding first followed by S-HIBA.

Overall structural features

Crystals of apo MtHIBADH were obtained first. Crystals of binary complexes of the enzyme with NAD+, S-HIBA, R-HIBA, l-serine and 3-HP were obtained by soaking the native crystals in solutions containing the respective ligands. Ternary complexes involving NAD+ and each substrate were also prepared by soaking. The structure of the apoenzyme, determined by molecular replacement, was used in the structure analysis of the complexes. The structures of 9 out of the 10 crystals were refined using data at resolutions ranging from 1.85 to 2.15 Å (Table 2). Crystals of the ternary complex involving NAD+ and l-serine diffracted only to a resolution of 2.68 Å (Table 2). The protein molecule is well defined in all the structures with interpretable electron density for at least 290 of the 293 residues in the protein. The three C-terminal residues are disordered to different extents in the crystals. The N-terminal residue is ill defined in most of them.

The 10 crystal structures are isomorphous to one another. Each contains an MtHIBADH dimer in the crystallographic asymmetric unit. Among the five HIBADH molecules of known three-dimensional structure, four exist as tetramers (dimer of dimers), while that from B. cereus exists as a dimer, as the MtHIBADH molecule does. As in the five, each of the subunits (designated A and B) is made up of two domains, an N-domain and a C- domain (Figure 2). The N-domain (residues 1–158) is an elaboration of the Rossman fold with a nine-stranded β-sheet decorated with helices. The C-domain (168–285) is almost wholly helical. A few N-terminal residues of α8, most of which belong to the C-domain, form part of the linker region (Figure 2).

Overall structure of the molecule.

Figure 2.
Overall structure of the molecule.

(A) Ribbon representation of the monomer. (B) Surface representation of the monomer shown in a different orientation. (C) The dimer shown in the same orientation as in (A). The N-domain is in green, the C-domain in orchid and the linker region in blue. Secondary structural elements are shown in (A). Bound substrates at the entry site and reaction center are indicated in ball and stick representation.

Figure 2.
Overall structure of the molecule.

(A) Ribbon representation of the monomer. (B) Surface representation of the monomer shown in a different orientation. (C) The dimer shown in the same orientation as in (A). The N-domain is in green, the C-domain in orchid and the linker region in blue. Secondary structural elements are shown in (A). Bound substrates at the entry site and reaction center are indicated in ball and stick representation.

The two molecules in the crystallographic asymmetric unit form a dimer with a near-perfect two-fold symmetry in all the crystals. The N-domain has the same structure in the 10 crystals, irrespective of the presence or absence of bound ligands, with r.m.s.d. ranging from 0.26 to 0.32 Å on pairwise superposition of α-carbon atoms of residues 1–158; the same is true about the C-domain and the range of r.m.s.d. being 0.15–0.17 Å for residues 168–285. The orientation between the two domains is nearly the same in subunit A in all the crystals. Again, the orientation in subunit B in the crystals also is nearly the same, but is slightly different from that in subunit A by ∼3°. Among the HIBADHs of known structures, the closure involving the two domains is observed for the coenzyme binding in the case of the HIBADH from T. thermophilus. In this case, the coenzyme is NADP+. No such closure upon coenzyme binding is apparent in the enzymes from H. sapiens and P. aeruginosa, irrespective of whether the coenzyme-bound enzyme is prepared by soaking or co-crystallization. In both the cases, the coenzyme is NAD+. For the enzyme from B. cereus, only the structure of the holoenzyme is available while in the case of S. typhimurium, only that of the apoenzyme is available. In MtHIBADH, the mutual orientation of the two domains is unrelated to ligand binding. Within each type of subunit, the interdomain orientation is nearly the same in the apoenzyme, holoenzyme and the substrate-bound apo or holoenzyme.

Protein–ligand interactions

NAD+ has well-defined electron density in both the subunits of the holoenzyme, although the occupancies are in the range of 0.7–0.8 (Figure 3). Density for the cofactor exists in both the subunits in the ternary complexes involving S-HIBA and l-serine while it exists only in subunit B in the remaining two ternary complexes (R-HIBA and 3-HP). The molecules bound to subunit A of the complex involving l-serine and that in subunit B in the complex involving 3-HP again have fractional occupancies. Protein–NAD+ interactions, which almost entirely involve the N-domain, are nearly the same in the two subunits in all the structures. They are in turn similar to those in HIBADHs of known structure from other sources. Among these enzymes, only that from T. thermophilus uses NADP+ as the cofactor. In other HIBADHs, including MtHIBADH, Asp-31 would have an unfavorable interaction with the phosphate group attached to the 2′O of the adenosine moiety of the NADP+ cofactor. This residue is replaced by arginine in T. thermophilus HIBADH (TtHIBADH). Small local conformational changes in the structure facilitate favorable stabilizing interactions of the side chain guanidyl group with the free phosphate group of NADP+, which presumably leads to the preference of NADP+ over NAD+ in TtHIBADH.

NAD+-binding site.

Figure 3.
NAD+-binding site.

Residues involved in NAD+ binding are shown in sticks. Simulated annealing Fo − Fc omit map for NAD+ in subunit A of the binary complex is also shown. Contours are at 3.0 σ levels.

Figure 3.
NAD+-binding site.

Residues involved in NAD+ binding are shown in sticks. Simulated annealing Fo − Fc omit map for NAD+ in subunit A of the binary complex is also shown. Contours are at 3.0 σ levels.

Those reported here are the only HIBADH structures in which substrates/substrate analogs have been located. Substrates/substrate analogs located in the different crystal structures reported here are associated with well-defined electron densities, although the densities are not as sharp as those for the polypeptide chains and NAD+ (Supplementary Figure S4). The average B values of most substrate molecules are in the region of 60 Å2, values which are high, but not unacceptable. The high B values could be a reflection of weak binding of the ligands to the protein. It is also possible that the sites are only partially occupied. However, refinement with partial occupancies resulted in positive densities at the sites in the difference Fourier map. Therefore, full occupancies for the ligand were retained. Furthermore, the repeated occurrence of the appropriate density at the same location in different crystal structures added confidence to the interpretation. In view of these factors and the small size of the substrate molecules, extreme care was taken in interpreting the appropriate electron densities and in the refinement of the relevant atomic positions. In any case, the located positions of the substrate molecules made eminent chemical sense.

The two subunits in the molecule behave very differently in relation to substrate binding. In subunit B, the substrate is located away from the catalytic site at what appears to be the entrance to the path toward the catalytic residues, situated between the two domains (Figure 4). At this site (Site I), made up of residues Ser-119, Gly-120, Gly-121, Trp-211, Phe-236 (all from subunit B) and Thr-207 (from subunit A), the protein forms several water bridges with the substrate. Significantly, the hydroxyl group of the substrate forms a strong hydrogen bond with the side chain of Asp-275 from a symmetry-related neighboring molecule. Bound substrate molecules in subunit A are found only in the complexes involving l-serine. The situation in the ternary complex is particularly interesting in relation to the activity of the enzyme. Two serine molecules are bound to the enzyme in the ternary complex (Figure 5). One of them is at a location (Site I′) similar to Site I in subunit B, but with a different orientation. Unlike at Site I, the molecule at Site I′ is not involved in any intermolecular interaction. The second molecule is at the catalytic center at a site (Site II) made up of residues Val-118, Ser-119, Gly-120, Gly-121, Lys-168, Asn-172, Asp-244 and Phe-236. The molecule is firmly embedded at the center with hydrogen bonds with several residues belonging to both the domains. It also interacts with the cofactor. It is difficult to distinguish between the substrate and the product in a medium-resolution X-ray analysis, as the difference between the two is in the presence of a hydrogen atom and the hybridization state of one atom. Thus, the molecule at Site II could be the substrate poised for the reaction or the product poised for leaving the catalytic center. Interestingly, the molecules at Sites I′ and II are in contact with each other. The one at Site I′ can move into the catalytic site (Site II) only when the incumbent moves out.

The entry site of the substrate.

Figure 4.
The entry site of the substrate.

Location of S-HIBA and NAD+ in subunit B of the ternary complex. Simulated annealing Fo − Fc omit map for S-HIBA is shown at contour levels of 3.0 σ. Asp-275 belongs to a symmetry-related molecule which is shown in gray.

Figure 4.
The entry site of the substrate.

Location of S-HIBA and NAD+ in subunit B of the ternary complex. Simulated annealing Fo − Fc omit map for S-HIBA is shown at contour levels of 3.0 σ. Asp-275 belongs to a symmetry-related molecule which is shown in gray.

Ternary complex of subunit A.

Figure 5.
Ternary complex of subunit A.

Location of NAD+ and l-serine molecules in subunit A of the ternary complex involving NAD+ and l-serine. Simulated annealing Fo − Fc omit maps of l-serine molecules are shown at contours levels of 3.0 σ.

Figure 5.
Ternary complex of subunit A.

Location of NAD+ and l-serine molecules in subunit A of the ternary complex involving NAD+ and l-serine. Simulated annealing Fo − Fc omit maps of l-serine molecules are shown at contours levels of 3.0 σ.

Structural basis for catalytic activity

The molecules located at Sites I, I′ and II provide a plausible picture of the movement of the substrate in the binding region. The fortuitous occurrence of an intermolecular interaction helps define the entrance to the binding region. The hydrogen bond between the hydroxyl group of the substrate bound to subunit B and the carboxylate group of Asp-275 of a neighboring molecule is presumably strong enough to hold the substrate at the entrance, thus preventing it from reaching the active site. No such constraint exists in subunit A and substrate molecules can readily access the active site. In most cases, presumably, reaction takes place and the product leaves the enzyme. l-serine appears to be an exception. Among the substrates used in the present crystallographic studies, the only one which carries an amino group is l-serine. This amino group forms a hydrogen bond with the side chain of Ser-119 that appears to prevent the substrate from leaving the active site. The molecule immobilized at Site II presumably causes the next incoming molecule to remain at Site I′. In a dynamic situation, the hydrogen bond referred to earlier may not exist in all molecules of the complex. It is possible that only those molecules in which it exists crystallized. Alternatively, it is possible that the site is only partially occupied. If that is the situation, only a portion of the molecules with the N–H—O hydrogen bond gets trapped in an unproductive state. In the case of the rest of molecules, catalytic reaction takes place and the product leaves the site.

The most favored substrate of the enzyme is S-HIBA which differs from l-serine only in the replacement of the amino group in the latter by a methyl group. This substitution abolishes the crucial hydrogen bond between the amino group and Ser-119 Oγ in the complex involving l-serine, if l-serine is replaced by S-HIBA. Furthermore, the larger size of the methyl group and the increased length of the CA—C bond compared with the CA—N bond results in unacceptable steric clash between the methyl group and Ser-119 Oγ. This clash can be relaxed by a rotation involving χ1 which changes the position of OG. The absence of a hydrogen bond with Ser-119 Oγ found in the l-serine complex makes it possible to make small changes in the location and conformation of S-HIBA at Site II. Making use of such changes, a plausible model of the ternary complex involving S-HIBA could be readily constructed manually. The model was then energy-minimized. In this model, χ1 of Ser-119 is close to 180° corresponding to a staggered conformation (Figure 6A). The carboxylate group of the substrate now interacts with Ser-119 Oγ, Gly-120 N and Gly-121 N. The hydroxyl group has strong interactions with Lys-168 Nζ, Asn-172 Nδ2 and Asp-244 Oδ2. In addition, the methyl group nestles against the side chain of Phe-236 (Figure 6A). Incidentally, the orientation of the substrate at Site II in the model is the same as that at Site I′ in the crystal structure. Thus, the movement from I′ to II would involve only a simple translation.

Interaction of S-HIBA in the model structure.

Figure 6.
Interaction of S-HIBA in the model structure.

(A) Modeled location of S-HIBA at the active site (see the text for details). (B) Plausible reaction mechanism where Lys-168 participates as a base in the general base catalysis.

Figure 6.
Interaction of S-HIBA in the model structure.

(A) Modeled location of S-HIBA at the active site (see the text for details). (B) Plausible reaction mechanism where Lys-168 participates as a base in the general base catalysis.

S-HIBA with the location and orientation specified in the model is poised for catalysis following the mechanism outlined (Figure 6B). The proposed mechanism is consistent with the mechanism proposed for 6-PGDH [45]. The hybridization state of CB in the product is sp2 while it was sp3 in the substrate. Furthermore, CB is now linked to OG by a shorter double bond. Consequently, OG shifts substantially from its original position resulting in the disruption of interactions with the protein. The product, thus destabilized, leaves the active site.

The model of the ternary complex involving S-HIBA provides a framework for examining the activities against the other substrates. Interactions of S-HIBA with the enzyme involves hydrogen bonds of the hydroxyl group with a patch made up of residues Lys-168, Asn-172 and Asp-244 (patch A), those of the carboxylate group with the VSGG motif (residues 119–121, patch B) and hydrophobic interactions of the methyl group with the side chain of Phe-236 (patch C). The mutual disposition of the three groups is governed by the configuration at CA. The configuration is different in R-HIBA and in no orientation it can simultaneously be involved in all the three sets of interactions. In the absence of the methyl group, HP can form only the two sets of hydrogen bonds. This could be the reason for the reduced activity of R-HIBA and 3-HP. In the case of l-serine, as indicated earlier, part of the molecules could be immobilized at the active site. The remaining molecules could assume an orientation similar to that of S-HIBA. The amino group could now interact with the Phe-236 side chain. Reaction could take place at this location in this orientation. These remaining molecules could be responsible for the residual activity of the enzyme against l-serine.

The location of l-serine at Site II and that of S-HIBA in the model also lead to plausible explanations for the absence of activity against compounds, the interactions of which have not been crystallographically explored. d-serine can occupy Site II in an orientation slightly different from that of l-serine, but with similar interactions with protein, taking advantage of the possibility of the rotation of the side chain hydroxyl group of Ser-119 by varying χ1. Therefore, d-serine can get trapped at Site II. As in the case of R-HIBA, the interaction of d-serine with the active site would be weak when placed in an orientation corresponding to that in the model. Considering that only a fraction of molecules, which are not trapped at Site II, would be available to productively interact with the enzyme in this orientation, that too only weakly, the activity is likely to be undetectable. In the case of l-threonine and d-threonine, the additional methyl group would be involved in steric clashes with the protein at Site II as well as at modeled location of the ligand. In the case of S-HIBA methyl ester and R-HIBA methyl ester, obviously the important interactions involving the carboxylate group are abolished. Furthermore, the additional methyl group would clash with the protein. Not surprisingly, the enzyme is inactive against them.

Mutational analysis

As indicated earlier, Lys-168, Asn-172 and Asp-244 interact with the hydroxyl group of S-HIBA while Ser-119 interacts with the carboxyl group at the active site. Phe-236 interacts with the methyl group. The entrance of the substrate is made up of the main chain atoms of Gly-120, Gly-121 and side chains of Thr-207, Trp-211 and Phe-236. The ligands at the entry point are also stabilized by Asp-275 of a neighboring molecule. Comparison with the HEPES-bound structure of l-serine-specific HIBADH of P. aeruginosa also revealed the identical residues at the corresponding sites (Supplementary Figure S5). Therefore, the side chains of all the above-mentioned residues, except of course the glycines, were mutated to a methyl group, as in alanine, one by one and the activities of the point mutants were examined with S-HIBA as a substrate (Figure 7). For comparison, the specific activities of mutants and WT proteins against S-HIBA and l-serine are represented as normalized specific activity with respect to the WT activity with respective substrates (namely, S-HIBA and l-serine). Among the point mutants, a marginal loss of activity is exhibited by D275A, indicating that the interaction involving Asp-275 is a crystallographic artifact. Nevertheless, its participation in the interaction with all the substrates at Sites I and I′ has led to their transient trapping at the entrance of the catalytic site. Hence, observation of substrates bound there provides an inkling of the path of the entry of the substrate there. The activity is wholly abolished when residues involved in interactions with the hydroxyl or carboxylate groups of S-HIBA are mutated (Figure 7). Among the residues located at the entrance of the pathway to the active site, Trp-211 appears to be the most important. It was verified that the dramatic reduction in the activity of W211A is not due to any structural instability by comparing the thermal denaturation temperature of WT (Tm 47°C) and the mutant enzymes (Tm 45°C) (Supplementary Figure S6). Mutation of Thr-207 and Phe-236 results in partial loss of activity. The activity profile of point mutants with l-serine also followed the similar pattern (Figure 7), indicating that the substrate specificity is not regulated through the residues of the active site region, strengthening the conformational selectivity of ligands as a reason for substrate preference. Thus, mutational analysis corroborates structural data.

Mutational analysis.

Figure 7.
Mutational analysis.

Enzymatic activities of MtHIBADH mutants involving the active site and entry site residues for S-HIBA and l-serine.

Figure 7.
Mutational analysis.

Enzymatic activities of MtHIBADH mutants involving the active site and entry site residues for S-HIBA and l-serine.

Conclusion

Biochemical studies of MtHIBADH show that the dimeric enzyme preferentially utilizes NAD+ as the cofactor. The preferred substrate of the enzyme is S-HIBA with a pH optimum of 11 for activity. The enzyme also acts, much less efficiently, on R-HIBA, l-serine and 3-HP. The crystal structures of the native enzyme and those of binary and ternary complexes involving NAD+, S-HIBA, R-HIBA, 3-HP and l-serine confirm that the molecule is a dimer. The reaction kinetics follow an ordered Bi–Bi mechanism where NAD+ binds first followed by S-HIBA binding. Although binary complexes with cofactors are available, none of the other known crystal structures of HIBADH contain a bound substrate. The crystal structures reported here along with simple modeling provide, for the first time, direct plausible information on the location of the substrate and its interactions with the enzyme. The interactions thus defined are similar to those observed in the crystal structure of the enzyme–substrate complex involving E. coli 6-phosphoglyceraldehyde dehydrogenase, another member of the superfamily in which the structural details of the relevant interactions have been completely defined [45]. They are also consistent with the results of mutational analysis. The specificity of the enzyme for S-HIBA is generated by the interactions of the hydroxyl group of the substrate with a hydrophilic patch in the enzyme made up of residues Lys-168, Asn-172 and Asp-244, those of the carboxylate group with the VSGG motif and hydrophobic interactions of the methyl group with the side chain of a phenylalanine residue. One or the other of these interactions is disturbed in the substrate analogs against which the enzyme has reduced activity. Together with mutant activity assay with l-serine substrate and the ternary complex of nonproductive conformation of bound l-serine, a logical explanation for substrate selectivity involving the conformational selection of ligands but not the residues from the active site region have been inferred. With advantage of an artifact involving intermolecular interactions, the entrance of the substrate to the active site region could be delineated. The crystal structures along with appropriate models also lead to the plausible mechanism of action of the enzyme.

PDB references

Mycobacterium tuberculosis HIBADH Native, 5Y8G; complex with NAD+, 5Y8H; complex with S-HIBA, 5Y8I; complex with R-HIBA, 5Y8J; complex with l-serine, 5Y8K; complex with 3-HP, 5Y8P; complex with S-HIBA and NAD+, 5Y8L; complex with R-HIBA and NAD+, 5Y8M; complex with l-serine and NAD+, 5Y8N; complex with 3-HP and NAD+, 5Y8O.

Abbreviations

     
  • 3-HP

    3-hydroxypropanoic acid

  •  
  • 6-PGDH

    6-phosphogluconate dehydrogenase

  •  
  • DSF

    differential scanning fluorimetry

  •  
  • ESI

    electrospray ionization

  •  
  • MtHIBADH

    Mycobacterium tuberculosis 3-hydroxyisobutyric acid dehydrogenase

  •  
  • PDB

    Protein Data Bank

  •  
  • S-HIBA

    S-3-hydroxyisobutyric acid

  •  
  • TtHIBADH

    T. thermophilus HIBADH

Author Contribution

Idea is conceived by A.S. and M.V. R.S. has carried out structural studies and Am.S. has carried out biochemical studies. Data have been analyzed by Am.S., A.S., R.S. and M.V. Am.S., R.S., A.S. and M.V. have written manuscript. A.S. and M.V. have provided framework to work.

Funding

This work has been supported by DBT, the Ministry of Science and Technology, Govt. of India and partially funded by CSIR, Govt. of India.

Acknowledgments

Intensity data were collected at the X-ray Facility for Protein Crystal Structure Determination and Protein Design at the Indian Institute of Science, supported by the Department of Science and Technology and at ESRF, Grenoble through an arrangement made by the Department of Biotechnology (DBT), Govt. of India. We acknowledge BEI resources for providing M. tuberculosis genomic DNA and Kaushik Saha for support in cloning. A.S. is a Bhatnagar Fellow of the Council of Scientific and Industrial Research and M.V. is a NASI Platinum Jubilee Senior Scientist.

Competing Interests

The Authors declare that there are no competing interests associated with the manuscript.

References

References
1
Hawes
,
J.W.
,
Harper
,
E.T.
,
Crabb
,
D.W.
and
Harris
,
R.A.
(
1996
)
Structural and mechanistic similarities of 6-phosphogluconate and 3-hydroxyisobutyrate dehydrogenases reveal a new enzyme family, the 3-hydroxyacid dehydrogenases
.
FEBS Lett.
389
,
263
267
2
Steele
,
M.I.
,
Lorenz
,
D.
,
Hatter
,
K.
,
Park
,
A.
and
Sokatch
,
J.R.
(
1992
)
Characterization of the mmsAB operon of Pseudomonas aeruginosa PAO encoding methylmalonate-semialdehyde dehydrogenase and 3-hydroxyisobutyrate dehydrogenase
.
J. Biol. Chem.
267
,
13585
13592
PMID:
[PubMed]
3
Letto
,
J.
,
Brosnan
,
M.E.
and
Brosnan
,
J.T.
(
1986
)
Valine metabolism. Gluconeogenesis from 3-hydroxyisobutyrate
.
Biochem. J.
240
,
909
912
4
Njau
,
R.K.
,
Herndon
,
C.A.
and
Hawes
,
J.W.
(
2001
)
New developments in our understanding of the β-hydroxyacid dehydrogenases
.
Chem. Biol. Interact.
130–132
,
785
791
5
Liu
,
Y.
,
Koh
,
C.M.J.
,
Sun
,
L.
and
Ji
,
L.
(
2011
)
Tartronate semialdehyde reductase defines a novel rate-limiting step in assimilation and bioconversion of glycerol in Ustilago maydis
.
PLoS ONE
6
,
e16438
6
Reitz
,
S.
,
Alhapel
,
A.
,
Essen
,
L.-O.
and
Pierik
,
A.J.
(
2008
)
Structural and kinetic properties of a β-hydroxyacid dehydrogenase involved in nicotinate fermentation
.
J. Mol. Biol.
382
,
802
811
7
Tchigvintsev
,
A.
,
Singer
,
A.
,
Brown
,
G.
,
Flick
,
R.
,
Evdokimova
,
E.
,
Tan
,
K.
et al. 
(
2012
)
Biochemical and structural studies of uncharacterized protein PA0743 from Pseudomonas aeruginosa revealed NAD+-dependent l-serine dehydrogenase
.
J. Biol. Chem.
287
,
1874
1883
8
Robinson
,
W.G.
and
Coon
,
M.J.
(
1957
)
The purification and properties of beta-hydroxyisobutyric dehydrogenase
.
J. Biol. Chem.
225
,
511
521
PMID:
[PubMed]
9
Rougraff
,
P.M.
,
Zhang
,
B.
,
Kuntz
,
M.J.
,
Harris
,
R.A.
and
Crabb
,
D.W.
(
1989
)
Cloning and sequence analysis of a cDNA for 3-hydroxyisobutyrate dehydrogenase. Evidence for its evolutionary relationship to other pyridine nucleotide-dependent dehydrogenases
.
J. Biol. Chem.
264
,
5899
5903
PMID:
[PubMed]
10
Rougraff
,
P.M.
,
Paxton
,
R.
,
Kuntz
,
M.J.
,
Crabb
,
D.W.
and
Harris
,
R.A.
(
1988
)
Purification and characterization of 3-hydroxyisobutyrate dehydrogenase from rabbit liver
.
J. Biol. Chem.
263
,
327
331
PMID:
[PubMed]
11
Murín
,
R.
,
Schaer
,
A.
,
Kowtharapu
,
B.S.
,
Verleysdonk
,
S.
and
Hamprecht
,
B.
(
2008
)
Expression of 3-hydroxyisobutyrate dehydrogenase in cultured neural cells
.
J. Neurochem.
105
,
1176
1186
12
Tasi
,
Y.-C.
,
Chao
,
H.-C.A.
,
Chung
,
C.-L.
,
Liu
,
X.-Y.
,
Lin
,
Y.-M.
,
Liao
,
P.-C.
et al. 
(
2013
)
Characterization of 3-hydroxyisobutyrate dehydrogenase, HIBADH, as a sperm-motility marker
.
J. Assist. Reprod. Genet.
30
,
505
512
13
Congdon
,
P.J.
,
Haigh
,
D.
,
Smith
,
R.
,
Green
,
A.
and
Pollitt
,
R.J.
(
1981
)
Hypermethioninaemia and 3-hydroxyisobutyric aciduria in an apparently healthy baby
.
J. Inherit Metab. Dis.
4
,
79
80
14
Landaas
,
S.
(
1975
)
Accumulation of 3-hydroxyisobutyric acid, 2-methyl-3-hydroxybutyric acid and 3-hydroxyisovaleric acid in ketoacidosis
.
Clin. Chim. Acta
64
,
143
154
15
Barshop
,
B.A.
,
Yoshida
,
I.
,
Ajami
,
A.
,
Sweetman
,
L.
,
Wolff
,
J.A.
,
Sweetman
,
F.R.
et al. 
(
1991
)
Metabolism of 1–13C-propionate in vivo in patients with disorders of propionate metabolism
.
Pediatr. Res.
30
,
15
22
16
Hasegawa
,
J.
(
1981
)
Purification, crystallization and some properties of β-hydroxyisobutyrate dehydrogenase from Candida rugosa IFO 0750
.
Agric. Biol. Chem.
45
,
2805
2814
17
Chowdhury
,
E.K.
,
Nagata
,
S.
and
Misono
,
H.
(
1996
)
3-Hydroxyisobutyrate dehydrogenase from Pseudomonas putida E23: purification and characterization
.
Biosci. Biotechnol. Biochem.
60
,
2043
2047
18
Lokanath
,
N.K.
,
Ohshima
,
N.
,
Takio
,
K.
,
Shiromizu
,
I.
,
Kuroishi
,
C.
,
Okazaki
,
N.
et al. 
(
2005
)
Crystal structure of novel NADP-dependent 3-hydroxyisobutyrate dehydrogenase from Thermus thermophilus HB8
.
J. Mol. Biol.
352
,
905
917
19
Park
,
S.C.
,
Kim
,
P.-H.
,
Lee
,
G.-S.
,
Kang
,
S.G.
,
Ko
,
H.-J.
and
Yoon
,
S.i.
(
2016
)
Structural and biochemical characterization of the Bacillus cereus 3-hydroxyisobutyrate dehydrogenase
.
Biochem. Biophys. Res. Commun.
474
,
522
527
20
Cole
,
S.T.
and
Barrell
,
B.G.
(
1998
)
Analysis of the genome of Mycobacterium tuberculosis H37Rv
.
Novartis Found. Symp.
217
,
160
177
;
discussion 172–177
21
Lee
,
J.-H.
,
Geiman
,
D.E.
and
Bishai
,
W.R.
(
2008
)
Role of stress response sigma factor SigG in Mycobacterium tuberculosis
.
J. Bacteriol.
190
,
1128
1133
22
Baker
,
J.J.
,
Johnson
,
B.K.
and
Abramovitch
,
R.B.
(
2014
)
Slow growth of Mycobacterium tuberculosis at acidic pH is regulated by phoPR and host-associated carbon sources
.
Mol. Microbiol.
94
,
56
69
23
Tripathy
,
S.
,
Lytle
,
K.A.
,
Stevens
,
R.D.
,
Bain
,
J.R.
,
Newgard
,
C.B.
,
Greenberg
,
A.S.
et al. 
(
2014
)
Fatty acid elongase-5 (Elovl5) regulates hepatic triglyceride catabolism in obese C57BL/6J mice
.
J. Lipid Res.
55
,
1448
1464
24
Crown
,
S.B.
,
Marze
,
N.
and
Antoniewicz
,
M.R.
(
2015
)
Catabolism of branched chain amino acids contributes significantly to synthesis of odd-chain and even-chain fatty acids in 3T3-L1 adipocytes
.
PLoS ONE
10
,
e0145850
25
Lew
,
J.M.
,
Kapopoulou
,
A.
,
Jones
,
L.M.
and
Cole
,
S.T.
(
2011
)
TubercuList — 10 years after
.
Tuberculosis
91
,
1
7
26
Edelheit
,
O.
,
Hanukoglu
,
A.
and
Hanukoglu
,
I.
(
2009
)
Simple and efficient site-directed mutagenesis using two single-primer reactions in parallel to generate mutants for protein structure-function studies
.
BMC Biotechnol.
9
,
61
27
Fiala
,
G.J.
,
Schamel
,
W.W.A.
and
Blumenthal
,
B.
(
2011
)
Blue native polyacrylamide gel electrophoresis (BN-PAGE) for analysis of multiprotein complexes from cellular lysates
.
J. Vis. Exp.
48
,
2164
28
Newman
,
J.
(
2004
)
Novel buffer systems for macromolecular crystallization
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
60
,
610
612
29
Vivoli
,
M.
,
Novak
,
H.R.
,
Littlechild
,
J.A.
and
Harmer
,
N.J.
(
2014
)
Determination of protein-ligand interactions using differential scanning fluorimetry
.
J. Vis. Exp.
91
,
51809
30
Leslie
,
A.G.W.
,
Powell
,
H.R.
,
Winter
,
G.
,
Svensson
,
O.
,
Spruce
,
D.
,
McSweeney
,
S.
et al. 
(
2002
)
Automation of the collection and processing of X-ray diffraction data — a generic approach
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
58
,
1924
1928
31
Evans
,
P.
(
2006
)
Scaling and assessment of data quality
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
62
,
72
82
32
Winn
,
M.D.
,
Ballard
,
C.C.
,
Cowtan
,
K.D.
,
Dodson
,
E.J.
,
Emsley
,
P.
,
Evans
,
P.R.
et al. 
(
2011
)
Overview of the CCP4 suite and current developments
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
67
,
235
242
33
McCoy
,
A.J.
(
2007
)
Solving structures of protein complexes by molecular replacement with Phaser
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
63
,
32
41
34
McCoy
,
A.J.
,
Grosse-Kunstleve
,
R.W.
,
Adams
,
P.D.
,
Winn
,
M.D.
,
Storoni
,
L.C.
and
Read
,
R.J.
(
2007
)
Phaser crystallographic software
.
J. Appl. Crystallogr.
40
,
658
674
35
Matthews
,
B.W.
(
1968
)
Some crystal forms of bovine chymotrypsinogen B and chymotrypsinogen A
.
J. Mol. Biol.
33
,
499
501
36
Emsley
,
P.
,
Lohkamp
,
B.
,
Scott
,
W.G.
and
Cowtan
,
K.
(
2010
)
Features and development of Coot
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
66
,
486
501
37
Murshudov
,
G.N.
,
Skubak
,
P.
,
Lebedev
,
A.A.
,
Pannu
,
N.S.
,
Steiner
,
R.A.
,
Nicholls
,
R.A.
et al. 
(
2011
)
REFMAC5 for the refinement of macromolecular crystal structures
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
67
,
355
367
38
Brunger
,
K.
and
Henry
,
A.
(
1998
)
Processing load and the extended optional infinitive stage
.
Int. J. Lang. Commun. Disord.
33
(
suppl.
),
422
427
39
Laskowski
,
R.A.
,
Moss
,
D.S.
and
Thornton
,
J.M.
(
1993
)
Main-chain bond lengths and bond angles in protein structures
.
J. Mol. Biol.
231
,
1049
1067
40
Cohen
,
G.H.
(
1997
)
ALIGN: a program to superimpose protein coordinates, accounting for insertions and deletions
.
J. Appl. Crystallogr.
30
,
1160
1161
41
DeLano
,
W.L.
(
2002
)
Pymol: An open-source molecular graphics tool
.
CCP4 Newsletter On Protein Crystallography
40
,
82
92
42
Pettersen
,
E.F.
,
Goddard
,
T.D.
,
Huang
,
C.C.
,
Couch
,
G.S.
,
Greenblatt
,
D.M.
,
Meng
,
E.C.
et al. 
(
2004
)
UCSF Chimera—a visualization system for exploratory research and analysis
.
J. Comput. Chem.
25
,
1605
1612
43
McDonald
,
I.K.
and
Thornton
,
J.M.
(
1994
)
Satisfying hydrogen bonding potential in proteins
.
J. Mol. Biol.
238
,
777
793
44
Patra
,
M.
and
Karttunen
,
M.
(
2004
)
Systematic comparison of force fields for microscopic simulations of NaCl in aqueous solutions: diffusion, free energy of hydration, and structural properties
.
J. Comput. Chem.
25
,
678
689
45
Chen
,
Y.-Y.
,
Ko
,
T.-P.
,
Chen
,
W.-H.
,
Lo
,
L.-P.
,
Lin
,
C.H.
and
Wang
,
A.H.-J.
(
2010
)
Conformational changes associated with cofactor/substrate binding of 6-phosphogluconate dehydrogenase from Escherichia coli and Klebsiella pneumoniae: implications for enzyme mechanism
.
J. Struct. Biol.
169
,
25
35

Author notes

*

These authors contributed equally to this work.