Deciphering the Role of Histidine 252 in Mycobacterial Adenosine 5′-Phosphosulfate (APS) Reductase Catalysis*

Mycobacterium tuberculosis adenosine 5′-phosphosulfate reductase (APR) catalyzes the first committed step in sulfate reduction for the biosynthesis of cysteine and is essential for survival in the latent phase of tuberculosis infection. The reaction catalyzed by APR involves the nucleophilic attack by conserved Cys-249 on adenosine 5′-phosphosulfate, resulting in a covalent S-sulfocysteine intermediate that is reduced in subsequent steps by thioredoxin to yield the sulfite product. Cys-249 resides on a mobile active site lid at the C terminus, within a K(R/T)ECG(L/I)H motif. Owing to its strict conservation among sulfonucleotide reductases and its proximity to the active site cysteine, it has been suggested that His-252 plays a key role in APR catalysis, specifically as a general base to deprotonate Cys-249. Using site-directed mutagenesis, we have changed His-252 to an alanine residue and analyzed the effect of this mutation on the kinetic parameters, pH rate profile, and ionization of Cys-249 of APR. Interestingly, our data demonstrate that His-252 does not perturb the pKa of Cys-249 or play a direct role in rate-limiting chemical steps of the reaction. Rather, we show that His-252 enhances substrate affinity via interaction with the α-phosphate and the endocyclic ribose oxygen. These findings were further supported by isothermal titration calorimetry to provide a thermodynamic profile of ligand-protein interactions. From an applied standpoint, our study suggests that small-molecules targeting residues in the dynamic C-terminal segment, particularly His-252, may lead to inhibitors with improved binding affinity.

Tuberculosis remains a serious threat to public health, and new drugs are needed to simply and shorten treatment as well as fight multidrug-resistant tuberculosis. Toward this end, the inhibition of cysteine biosynthesis and, by extension, associated downstream metabolites represents fertile ground for the development of novel antibiotics (1,2). In mycobacteria, the enzyme adenosine 5Ј-phosphosulfate reductase (APR) 2 cata-lyzes the committed step in the biosynthesis of cysteine (Scheme 1) and is a validated target to develop new anti-tuberculosis agents, particularly for the treatment of latent infection (3,4). APR lacks a human homolog but is highly conserved across a wide range of bacterial species (5), raising the possibility that APR may also represent an attractive target for the discovery or rational design of broad spectrum antibiotics. APR is also present in plants and is recognized as a potential target for herbicide development (6 -8).
The importance of APR for microbial and plant survival has motivated investigations of its catalytic mechanism (9 -12). These studies provide support for the two-step mechanism shown in Scheme 2, which involves the nucleophilic attack by conserved Cys-249 3 on adenosine 5Ј-phosphosulfate (APS) leading to the formation of a covalent enzyme S-sulfocysteine intermediate, E-Cys-S␥-SO 3 Ϫ bound to AMP. The sulfite product is then released via thiol-disulfide exchange with free thioredoxin in bacterial APR or via the action of the C-terminal thioredoxin-like protein domain in plant APR. In addition, APR contains an [4Fe-4S] cluster 4 that is essential for catalytic activity (6,(13)(14)(15). However, it is not involved in redox chemistry, and its role remains an active area of investigation (16,17).
In 2006, the crystal structure of Pseudomonas aeruginosa APR (PaAPR) was solved in complex with APS, providing direct insight into substrate recognition (18). PaAPR and Mycobacterium tuberculosis APR (MtAPR) are homologous proteins sharing 27% identity and 41% similarity (supplemental Fig. S1), particularly in residues that line the active site (84% identity and 92% similarity). PaAPR and MtAPR have comparable reaction kinetics and ligand binding affinity (10,19). Likewise, the PaAPR structure has been successfully employed in virtual ligand screening to identify low micromolar chemical inhibitors of MtAPR (20).
The structure of PaAPR shows that APS binds in a deep active site cavity with the phosphosulfate extending toward the protein surface (see Fig. 1A). Conserved and semiconserved residues participate in four main chain hydrogen bonds with adenine and the ribose O 2 Ј hydroxyl. Interaction between the phosphosulfate and APR occurs via strictly conserved residues Lys-145, Arg-237, and Arg-240. The phosphosulfate is also positioned opposite the [4Fe-4S] cluster. 5 The C-terminal 18 residues, carrying the catalytically essential Cys-249, were disordered in the structure of PaAPR. The lack of electron density information, coupled with limited proteolysis studies, led to the proposal that Cys-249 resides on a flexible "lid peptide" that closes over the active site pocket upon ligand binding (18).
This conformational change hypothesis was later confirmed when Fisher and co-workers (21) reported the crystal structure of the related enzyme, 3Ј-phosphoadenosine-5Ј-phosphosulfate (PAPS) reductase from Saccharomyces cerevisiae (ScPAPR) in complex with adenosine 3Ј,5Ј-diphosphate (PAP). Although APS and PAPS differ by a 3Ј-phosphate and PAPR lacks the [4Fe-4S] cluster, 6 structural and functional studies show that the two-step mechanism for sulfite production in Scheme 2 is conserved among this family of enzymes, known collectively as sulfonucleotide reductases (10,18,22,23). Sulfonucleotide reductases share ϳ25% overall amino acid identity, including two highly conserved domains, the sulfonucleotide-binding pocket, and C-terminal segment containing the K(R/T)ECG(L/ I)H catalytic motif (supplemental Fig. S1; see also Ref. 18 for an alignment of 38 APR and 34 PAPR amino acid sequences). Likewise, sulfonucleotide reductases have virtually identical threedimensional structures (superposition C␣ backbone atoms from PaAPR and ScPAPR yields an root mean square deviation of 0.98 Å; see Fig. 1B). The crystal structure of ScPAPR is especially significant as it shows the flexible C-terminal segment folded over the active site pocket. In this conformation, a strictly conserved histidine residue His-252 within the K(R/ T)ECG(L/I)H motif is proximal to the active site ligand (ϳ3 Å) and Cys-249 (ϳ4 -5 Å) 7 (see Fig. 1B). These three-dimensional relationships are recapitulated well in the homology model of MtAPR, generated on the basis of sequence alignment and the ScPAPR template structure (root mean square deviation of 0.1 Å for C␣ backbone atoms; Fig. 1C).
On the basis of conservation and juxtaposition to the catalytic cysteine, it was recently proposed that His-252 acts as a general base in sulfonucleotide reductases to deprotonate the Cys-249 nucleophile (21). However, this hypothesis has not yet been directly tested, and thus, the precise function of this active site residue remains unknown. Herein, we have used site-directed mutagenesis to change His-252 in MtAPR to an alanine residue and analyzed the effect of this mutation on the kinetic parameters, pH rate profile, and ionization of Cys-249 of APR. In addition, isothermal titration calorimetry (ITC) was performed to provide a thermodynamic profile of ligand-protein interactions. Collectively, our data indicate that His-252 does not perturb the pK a of Cys-249 or play a direct role during chemical steps that lead to S-sulfocysteine formation. Instead, we show that interactions with His-252 increase substrate affinity, which might be used in further inhibitor design to trap the enzyme in a closed, inactive conformation.

EXPERIMENTAL PROCEDURES
Materials-All chemicals, unless stated otherwise, were purchased from the Sigma and were of the highest purity available. The C-terminal peptide (AKTECGLHASW) was synthesized by solid-phase peptide synthesis using Fmoc-based chemistry and HPLC-purified to Ͼ98%. The molecular mass of the peptide was confirmed by mass spectrometry (1202.4 Da). Aristeromycin was synthesized from dimethyl-3-cyclopentene-1, 1-dicarboxylate as described previously (24). 5Ј-Phosphoaristeromycin was prepared by chemical phosphorylation of aristeromycin using established methods (25). The physical and spectral data for 5Ј-phosphoaristeromycin were consistent with values reported in the literature for this nucleotide (25).
Mutagenesis and Protein Expression-The construction of the expression vector encoding wild-type MtAPR cloned into the vector pET24b (Novagen) has been described previously (10). The H252A mutant plasmid was prepared using QuikChange site-directed mutagenesis (Stratagene). Wild-type and mutant MtAPR were overexpressed and purified to homogeneity according to published procedures using nickel affinity and gel filtration column chromatography (17).
General APS Reductase Assay-APR assays were performed as described previously (17,19). All assays were conducted at 30°C. Unless otherwise indicated, the reaction conditions included 100 mM Bis-Tris propane (pH 6.5), supplemented with 5 mM DTT, and 10 M Escherichia coli thioredoxin. Production of 35 SO 3 2Ϫ from 35 S-labeled APS was monitored using charcoalbased separation and scintillation counting as reported previously (19). For each time point, the fraction product was calculated according to Equation 1, where F is the fraction converted to product, P is product, and S is intact substrate. Reactions progress curves were analyzed using Kalediagraph (Synergy Software) as described below.
Single-Turnover Kinetics-Single-turnover APR assays were performed in the standard reaction buffer as described above. To ensure single-turnover reactions, the concentration enzyme was kept in excess over the concentration of [ 35 S]APS (typically 2.5 nM). Reactions were followed to completion (Ն5 half-lives) except for very slow reactions. The reaction progress curve was plotted as the fraction of product versus time and was fit by a single exponential using Kaleidagraph (Equation 2), where F is the fraction product, A is the fraction of substrate converted to product at completion, k obs is the observed rate constant, and t is the time.
Under single-turnover conditions, the concentration dependence of the enzyme is hyperbolic (Equation 3). The maximal observed rate constant (k max ) corresponds to the rate of reaction at a saturating enzyme concentration, and the K1 ⁄ 2 value indicates the concentration at which half of the substrate is bound. For K1 ⁄ 2 determinations, the APR concentration was varied over a wide range, and reactions were carried out in the absence of thioredoxin, as described previously (19). Although we refer to the K1 ⁄ 2 value for maximal activity as K m , the value could differ from the K m value for multiple turnover because the latter can be affected by product release. At least two or more enzyme concentrations were averaged to obtain the . Under these conditions, the observed rate constant is linearly dependent upon enzyme concentration, and independent of substrate across a concentration range of at least 4-fold, which demonstrated that substrate was not saturating. The reported k cat /K m values are for single-turnover conditions, but are equivalent to steady-state k cat /K m (17).
The single-turnover rate constant (k max ) was determined at saturating concentration of APR, and this was confirmed by the observation of the same rate constant at two different concentrations of APR. Under these conditions, the observed rate constant is equal to the maximal single-turnover rate constant (k obs ϭ k max ) and monitors steps after binding up to and including the chemical step (Equation 3). The inhibition constant (K i ) was measured for various ligands by inhibiting the APR reac-tion under k cat /K m conditions with varying concentration of inhibitor (I). The data were fit to a simple model for competitive inhibition (Equation 4) and, with subsaturating APR, the K i is equal to the equilibrium dissociation constant (K d ) of the inhibitor.
pH Dependence for k cat /K m -The following buffers were used for the indicated pH range: sodium 4-morpholineethanessulfonic acid (6.0 -7.0), Bis-Tris propane (6.5-7.5), Tris-HCl (7.5-9.0), and sodium N-cyclohexyl-3-aminopropanesulfonic acid (9.0 -9.5). Reactions were carried out with 100 mM buffer. The rate constants obtained at each pH value for multiple reactions were averaged, and the S.D. were Յ25% of the average. The data were fit to a model for a single rate-controlling ionization as described by Equation 5.
pH Dependence of Inhibitor Binding-The following buffers were used for the indicated pH range: sodium 4-morpholineethanessulfonic acid (6.0 -7.0), Bis-Tris propane (6.5-7.5), Tris-HCl (7.5-9.0), and sodium N-cyclohexyl-3-aminopropanesulfonic acid (9.0 -9.5). Reactions were carried out with 100 mM buffer. The conditions described above were used to monitor k cat /K m for reduction of [ 35 S]APS in the presence and absence of inhibitor. The rate constants at each pH value for multiple reactions were averaged, and S.D. were Յ25% of the average. K a values were determined using Equation 6, derived from a model where the binding of the ligand depends on a single ionizable group.
Determination of Substrate Affinity-The K d for [ 35 S]APS from wild-type and H252A MtAPR-ligand complexes was measured using an ultrafiltration binding assay reported by Hernick and Fierke (26). In brief, the concentration of substrate was kept low (i.e. below the K d ) and constant, and the concen- Hydrogen bond interactions are depicted as yellow dashes. C, no empirical three-dimensional structure information is currently available for MtAPR; however, the amino acid sequence of sulfonucleotide reductases is highly conserved, particularly among residues that define the active site (supplemental Fig. S1). In view of this, a homology model was built using the structure of ScPAPR (Protein Data Bank code 2OQ2) as a template and the Swiss-Model server. The model predicts several interactions between APS and His-252, analogous to those observed in the crystal structure of ScPAPR bound to PAP. Hydrogen bond interactions are depicted as yellow dashes. The magenta dashes indicate the predicted distance between the two nearest atoms of His-252 and Cys-249 (ϳ5 Å). AUGUST 12, 2011 • VOLUME 286 • NUMBER 32 JOURNAL OF BIOLOGICAL CHEMISTRY 28569 tration of the enzyme was varied from 0 to 50 M (wild-type) or 0 to 400 M (H252A). The enzyme was added to reaction buffer containing 100 mM Bis-Tris propane (pH 6.5) with 5 nM APS at 30°C and then transferred into ultrafiltration devices (Microcon, 30-kDa cut-off, Millipore), and the free and bound ligand separated by centrifuging the samples at 3000 rpm for 2.5 min. Equal volumes of the filtrate and retentate were removed and quantified using scintillation counting. The ratio of EL/L total was determined as a function of [E] total , and the K d value was obtained by fitting Equation 7 to these data.

Role of His-252 in M. tuberculosis APS Reductase
Spectrophotometric pK a Determination of Cys-249-Buffer exchange of APR was performed using a PD-10 column (GE Healthcare) that had been pre-equilibrated with degassed water. Ionization of Cys-249 was monitored by absorption of the thiolate anion at 240 nm (23) using a Cary 50 UV-visible spectrometer (Varian) and a 1-cm path length quartz cuvette. APR was diluted to a final concentration of 20 M in 10 mM MES buffers of various pH (5.0 -8.0), and the absorption of the sample was monitored at 240 and 280 nm after correction for the absorption of the MES buffer alone. The extinction coefficient at 240 nm (⑀ 240 ) was calculated using the ratio of absorbance at 280 and 240 nm (Equation 8).
A 240 /A 280 is the ratio of the absorbance of the protein at 240 and 280 nm, ⑀ 280 is the known extinction coefficient of APR at 280 nm (36,815 M Ϫ1 cm Ϫ1 ), and ⑀ 240 is the extinction coefficient at 240 nm (23). The data were plotted as a function of pH, and the pK a was determined by fitting a version of the Henderson-Hasselbalch equation to the data (Equation 9).
where q is the directly measured amount of heat released during each injection, v is the volume of the reaction, and L i is the ligand concentration at the ith injection. the K d was calculated as the inverse of the K a .

RESULTS AND DISCUSSION
To advance our understanding of the molecular recognition and catalytic mechanism of MtAPR, we used site-directed mutagenesis to change His-252 to an alanine residue and characterized single turnover kinetic parameters for wild-type and H252A. Mutation of His-252 to Ala reduced k cat /K m by 230fold (Table 1), indicating that this residue contributes to catalytic efficiency by enhancing substrate affinity and/or stabilizing the catalytic transition state. To gain further insight into the role of the conserved active site histidine residue, we compared the saturating single-turnover rate constant (k max ), the K m , and the substrate dissociation constant (K d ) for wild-type and H252A MtAPR (Table 1; see also supplemental Figs. S2-S5 for representative data). The results show that alanine substitution of His-252 decreased the value of k max by only 2-fold, whereas K m and K d were both weakened by more than two orders of magnitude. Control experiments showed that there was no difference in iron incorporation or [4Fe-4S] cluster stability between the wild-type and variant MtAPR (supplemental Fig.  S6), consistent with the long-range distance (ϳ10 Å) that is predicted between His-252 and the metallocenter.
To examine the role of His-252 in greater detail and provide additional insight into the overall catalytic mechanism of APR, we measured the pH dependence of k cat /K m for the wild-type enzyme and the H252A mutant. Fig. 2 illustrates that the acidic Procedures"). d Dissociation constants were measured using ultrafiltration at 30°C (100 mM Bis-Tris propane, pH 6.5) as described under "Experimental Procedures." e From Ref. 26. f In Bis-Tris propane at pH 7.5. Due to technical limitations of the TLC-based assay only a lower limit could be obtained. limb for reaction of APS with wild-type or H252A has a firstorder dependence on the proton concentration, consistent with a single inactivating protonation at acidic pH. For k cat /K m , these kinetic pK a values could represent ionization of either free enzyme or substrate. The data described below support the model with ionization of the Cys-249 nucleophile.
The acidic limb of the pH dependence for the APR-catalyzed reduction of APS is best fit by a pK a of 6.1 Ϯ 0.1 and 6.3 Ϯ 0.1 for wild-type and H252A MtAPR, respectively (Fig. 2). The most likely candidate for this ionization is the enzyme, specifically of catalytic cysteine, because the substrate pK a falls significantly below this region. To test this proposal, we determined the pK a of Cys-249 by measuring the change in absorbance of UV light at 240 nm resulting from formation of the thiolate anion, as described previously (23,27,28). For these studies, we utilized C59A MtAPR, which has identical kinetic properties to the native enzyme (10, 15) but eliminates a nonconserved cysteine that could confound the analysis.
The pH dependence of the molar extinction coefficient of C59A MtAPR at 240 (⑀ 240 ) displays a transition with a pK a of 6.2 Ϯ 0.1 (Fig. 3A). The change in ⑀ 240 is most likely due to ionization of Cys-249, as indicated by the absence of a pH-dependent transition for C59A/C249A MtAPR (Fig. 3A). The pH dependence of the molar extinction coefficient of C59A/ H252A MtAPR at 240 (⑀ 240 ) shows a transition with a pK a of 6.0 Ϯ 0.1 (Fig. 3B). For comparison, we evaluated the pK a of Cys-249 within a synthetic peptide derived from the last 10 C-terminal residues of MtAPR (supplemental Fig. S4). The pK a of the thiol in the peptide segment was determined as 8.3 Ϯ 0.1, consistent with the pK a value of free cysteine solution (29). Interestingly, our experiments indicate that thiolate formation at Cys-249 correlates with decrease in signal at ⑀ 240 , as opposed to the increase that is normally observed. Therefore, the ionization constant of Cys-249 was verified by an independent method using the thiol-specific reagent, monobromobimane (24). In this assay, the pK a value of Cys-249 for C59A MtAPR was determined to be 6.0 Ϯ 0.1 (supplemental Fig. S4), which is within error of the UV-based method. The similarity of the kinetic pK a and the pK a value for Cys-249 deprotonation strongly suggest that the observed inflection in k cat /K m corresponds to the ionization of the active site cysteine to form the thiolate anion.
To further investigate the molecular recognition properties for wild-type and H252A MtAPR, we compared the pH dependence for binding of the nucleotide product, AMP. As shown in Fig. 4, the logarithm of the association constant (K a ϭ 1/K d ) shows a first-order dependence on the proton concentration. The acidic limb for wild-type MtAPR binding to AMP has a pK a of 8.1 Ϯ 0.1, as reported previously (19). The pK a values observed in product affinity could reflect ionizations in either or both the ligand and the enzyme, analogous to the pH dependence for k cat /K m discussed above. A likely explanation  for the weaker binding of AMP below pH 8 is that the dianion binds more tightly than the monoanion. However, the apparent pK a differs from the pK a of AMP in solution (ϳ6.8) by more than one unit. The discrepancy between the experimental data and this model is most likely due to concurrent ionization of the enzyme that affects ligand binding, leading to shift in the apparent pK a of AMP. One model that could account for this upward deviation is that an enzymatic group with a pK a of ϳ6 contributes slightly (ϳ5-fold) to AMP binding when protonated. Given its proximity to the ␣-phosphate, the most likely residue to exert such an effect on ligand binding is His-252. Consistent with this proposal, the acidic limb for H252A MtAPR binding to AMP displays a pK a of 6.4 Ϯ 0.1 (Fig. 4). An additional observation from these data is that binding of the nucleotide product to H252A is weaker at physiological pH and above, as compared with wild-type MtAPR. For example, at pH 7.5 wild-type and H252A MtAPR bind to AMP with respective K d values of 5.4 Ϯ 0.2 M and 50.5 Ϯ 3 M.
The crystal structure of scPAPR (21) and the model of MtAPR shown in Fig. 1 indicate that the side chain of His-252 is positioned within hydrogen bonding distance of the ␣-phosphate and the endocyclic ribose oxygen of the active site ligand. Previous studies have demonstrated the relative importance of the ␣-phosphate group for AMP binding to MtAPR (ϳ3 kcal/ mol) (19); however, the contribution of O-4 in the ribose sugar has not been investigated. To examine the importance of the hydrogen bond contact between His-252 and the endocyclic ribose oxygen, we synthesized 5Ј-phosphoaristeromycin, which replaces O-4 in AMP with a methylene unit. Binding studies indicate that at pH 7.5, this analog binds to MtAPR with a K d value of 25 M Ϯ 2.5 M (Fig. 5) or 5-fold more weakly than AMP. These data indicate that the interaction of His-252 with the ribose O-4 makes a modest contribution to ligand recognition (ϳ1 kcal/mol).
To substantiate the role of His-252 in substrate binding, we performed additional biophysical experiments. In initial experiments, we attempted to monitor spectral perturba-tion of noncatalytic 2Ј(3Ј)-O-(N-methylanthraniloyl) and N6-etheno substrate analogs. However, the affinity of these ligands for wild-type MtAPR was extremely weak (K d Ͼ 1 mM), and the associated signal changes were unreliably small (not shown). As an alternative approach, we employed ITC to measure affinities for wild-type and H252A MtAPR for substrate, APS, and product, AMP (supplemental Fig. S7). ITC offers a direct and complete characterization of the thermodynamic interaction whereby the ligand is titrated into the protein (30,31). This analysis indicates that APS binds to wild-type MtAPR with a K d of 0.6 Ϯ 0.3 M as compared with 42 Ϯ 6.2 M for H252A. Furthermore, AMP binds to wild-type MtAPR with a K d of 7.5 Ϯ 1.4 M compared with 67 Ϯ 8.4 M for H252A. These data are in excellent agreement with the other kinetic and thermodynamic values obtained from our radiolabeled biochemical assay (i.e. Table 1 and Fig. 4).
Collectively, the functional data presented herein provide strong support for a direct interaction between His-252 in the C terminus with ligands, including APS and the nucleotide product AMP. The flexible C-terminal segment must fold over the active site upon substrate binding to bring Cys-249 in proximity to the ␤-sulfate group. In this context, our studies do not support a role for His-252 as a general base that deprotonates catalytic Cys-249 because (i) wild-type and H252A exhibit similar pK a values for both k cat /K m and Cys-249 deprotonation, and (ii) alanine substitution of His-252 has an extremely modest affect on k max . Rather, our data show that His-252 plays an important role in ligand binding and likely facilitates docking of the C-terminal residues. These studies also reveal that the pK a value of the Cys-249 nucleophile is perturbed downward by more than two units (i.e. 6.2) relative to the value that we obtained for this residue in the context of the free peptide (i.e. 8.3). The low pK a value of Cys-249 in MtAPR is consistent with the essential catalytic function of this residue. Positively charged amino acids in the active site, including Lys-145, Arg-237, and Arg-240, are likely candidates for stabilization of the thiolate.
A critical motivating factor for these studies is that APR is essential for mycobacterial survival during persistent infection  (4). This key discovery has led to the proposal that small molecule inhibitors of APR might be a source for new drugs to treat latent tuberculosis infection (1,3,20). The increasing number of antibiotic-resistant strains suggests that the availability of such compounds could play an important role in treating the disease and minimizing the impact on human health. Defining active site residues that are essential for molecular recognition in MtAPR sets the stage for the development of such drugs. Toward this end, results from the present study suggest that targeting dynamic elements within the active site, particularly Cys-249 and His-252, may increase the potency of APR inhibitors.