Mutational and Kinetic Evaluation of Conserved His-123 in Dual Specificity Protein-tyrosine Phosphatase Vaccinia H1-related Phosphatase PARTICIPATION OF TYR-78 AND THR-73 RESIDUES IN TUNING THE ORIENTATION OF HIS-123*

Active-site cysteine strategically positioned in the P-loop of protein-tyrosine phosphatases has been suggested to be further stabilized by hydrogen bonding arrays radiating out from the P-loop to neighboring residues. In this work, we investigated the structural role of histidine array in HC( X ) 5 RS motif of the vaccinia H1-related protein phosphatase (VHR), using site-di-rected mutagenesis in conjunction with an extensive kinetic analysis. Conserved His-123 was mutated along with neighboring residues Tyr-78 and Thr-73. The increased p K a values of active-site Cys-124 found in Y78F and T73A mutants (6.51 and 6.75, respectively) were comparable to those of H123A and H123F mutants. Kinetic evaluation of Y78F and T73A mutants further im-plicates that the mutations perturb the relative position of Cys-124 within the P-loop. These results imply that Tyr-78 and Thr-73 make up an essential part of the His-123 array and structurally tune the Cys-124 position. Tyr-78 of VHR turns out to be the invariant Tyr reported in several protein-tyrosine phosphatases by a structure-based sequence alignment. Therefore, orientation of the

The reaction mechanism of vaccinia H1-related phosphatase (VHR), which belongs to the dual-specific protein phosphatase (11)(12)(13), has been extensively studied (14,15). Recently, VHR has been identified as a tyrosine-specific extracellular signalregulated kinase phosphatase (16). Like other PTPs the activesite Cys-124 is stabilized by an extensive network of hydrogen bonding between backbone NH of the P-loop and the thiolate anion (4,17,18). The sulfhydryl group of Cys-124 has an unusually low pK a value and functions as a nucleophile to form a thiophosphate-enzyme intermediate (19). Asp-92 serves as a general acid catalyst for formation of the phosphoenzyme intermediate (20), and in the hydrolysis of thiophosphate intermediate the same Asp-92 acts as a general base, and the hydroxyl group of Ser-131 is transiently linked to the thiol (14) (Scheme 1). The phosphoenzyme intermediate of VHR has been verified by 31 P NMR (21), and the structure of the transition state forming the intermediate has been suggested to be a highly dissociative metaphosphate-like transition state (22). Additionally, by using a variety of nonpeptide substrates, the decomposition of phosphoenzyme intermediate has been shown to be the rate-limiting step for the VHR-catalyzed reaction (15).
The crystal structure of VHR further explains some catalytic properties of VHR (7). The shallow active-site pocket in VHR is likely to be related to the broad substrate specificity of VHR, which is known to be favorable toward sterically demanding nonpeptidic substrates (23). The absence of a water molecule in the cavity of VHR coincides with a nucleophilic reactivity of alcohols observed in VHR (24). The Glu-450 of Yersinia PTP ligates a water molecule and effectively prevents nucleophilic attack of alcohol from outside the cavity. The VHR structure also reveals that the position of conserved Asp-92 is different from the flexible WPD-loop found in Yersinia PTP or PTP1B (3). However, there are some structural features of VHR that have not been adequately recognized (7). One is that imidazole N-␦ of His-123 forms a hydrogen bond with the carbonyl oxygen of Cys-124. This imidazole ring of VHR is further connected to a water molecule held between Tyr-78 and Thr-73 by hydrogen bonds (see Fig. 1). The water molecule ligated at this position seems to be unique to VHR. His-123 of VHR is known to be conserved in most PTPs as HC(X) 5 R(S/T) except low molecular weight PTPs where His is replaced by Val (17). However, studies on the function of the conserved His are surprisingly sparse. * This work was supported in part by Korea Science and Engineering Foundation, Directed Basic Research Grant 95-0501-07-02-3, the Center for Molecular Catalysis at Seoul National University, by Brain Korea 21 Program, and by Korea Ministry of Science and Technology Grants 98-N1-06-01-A-01 and 98-NQ-08 -01-A-04 (to D. Y. S.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  One early report on the role of His-402 of Yersinia PTP shows that the His residue is critical for stability of the active-site Cys-403 (25). When His-402 was substituted with Asn or Ala, the pK a of the active-site Cys-403 increased dramatically. The authors, not knowing the structure of Yersinia PTP at that time, proposed a direct interaction between the imidazole ring of His-402 and thiol of Cys-403. This proposal is not compatible with the His position in the P-loop of Yersinia PTP determined by x-ray crystallography (4). Now all conserved His residues in VHR, PTP1B, and Cdc25A appear to interact with the carbonyl group of the active-site cysteine residue by a hydrogen bond.
In this study, to verify the structural role of the conserved His, we first carried out site-directed mutagenesis on His-123 of VHR, and we examined its effect on pK a of the active-site Cys-124. Then we examined the possible role of the hydrogen bonding array involving the conserved His-123, Tyr-78, and Thr-73 in orienting Cys-124 at the P-loop for optimal catalytic activity by mutating Tyr-78 and Thr-73. Extensive kinetic analysis was carried out with mutants of Y78F and T73A to demonstrate that the hydrogen bonding array of His-123 is crucial for maintaining the proper positioning of Cys-124 in the catalytic site.

EXPERIMENTAL PROCEDURES
Materials-The substrate materials para-nitrophenyl phosphate, phosphotyrosine, phosphothreonine, and phenophthalein monophosphate were purchased from Sigma. The sequencing grade endoprotease Lys-C and microcentrifuge filter with molecular cut-off of 10,000 were also obtained from Sigma. Malachite green oxalate was purchased from Junsei (Tokyo, Japan), and ammonium molybdate was from Shinyo (Osaka, Japan). pET-21a vector for bacterial expression was obtained from Novagen (Madison, WI), and isopropyl-thio-␤-D-galactoside was obtained from Roche Molecular Biochemicals. Hitrap Heparin and Superose 12 HPLC column was purchased from Amersham Pharmacia Biotech. [ 3 H]Iodoacetic acid was obtained from PerkinElmer Life Sciences.
Expression and Purification-The wild-type and mutant VHRs were cloned in pET-21a vector and transformed to Escherichia coli strain BL21(DE3) for protein expression. The overexpression and purification of VHR were performed according to Zhou et al. (19) with some modifications. The pre-chromatographic steps were the same as reported previously (19) except that Hitrap Heparin and Superose 12 HPLC chromatography were used instead of S-Sepharose and Sephadex G-75. Briefly, the 0.5% polyethyleneimine and 35-65% ammonium sulfate precipitate were dialyzed against 20 mM MES, pH 6.0, containing 1 mM EDTA and 1 mM dithiothreitol. The resulting solution was filtered through a 0.22-m pore filter and injected into a Hitrap Heparin HPLC column. After the column was washed extensively with 20 mM MES, pH 6.0, VHR was eluted at 0.15 M NaCl in a linear NaCl gradient. The active fractions collected were concentrated with a microcentrifuge filter and loaded onto a Superose 12 HPLC column that was preequilibrated with 20 mM MES, pH 6.0, containing 0.1 M NaCl. VHR was eluted at a retention time corresponding to 20 kDa. Mutant VHRs were purified by the same method, and the purity of isolated wild-type and mutant VHRs was checked on SDS-polyacrylamide gel electrophoresis.
VHR Phosphatase Assay-VHR assay was performed at 30°C as described by Denu and Dixon (14). The buffer system containing 0.1 M acetate, 0.05 M Tris, and 0.05 M Bis-Tris at a constant ionic strength of 0.1 M was used throughout the entire pH range. Nonpeptide phosphate esters, pNPP, pTyr, pThr, and PMP, were used as substrates. When pNPP was used, the phosphatase activity was measured by absorbance at 405 nm after quenching with 0.5 M NaOH. In the case of pTyr, pThr, and PMP, the inorganic phosphate released was measured by colorimetry at 660 nm using malachite green solution (27). For obtaining Michaelis-Menten parameters, k cat and k cat /K m , initial velocities were measured at eight different substrate concentrations between 0.1 and 5 K m . The initial velocities were fitted to the Michaelis-Menten equation using the nonlinear data analysis program (Axum, Mathsoft, Inc.). The enzyme concentration was determined from absorbance at 280 nm using an extinction coefficient of 11500 M Ϫ1 cm Ϫ1 (28).
Determination of pK a of Cys-124 -For determination of pK a of wildtype and mutant VHRs, a direct 3 H-carboxymethylation method was used (20,25). A portion of 2 g of enzyme in each pH value was incubated with 5 Ci of [ 3 H]iodoacetic acid (344.2 mCi/mmol) for 100 min at 30°C. The labeled protein was precipitated with 10% trichloroacetic acid on ice and washed three times with ice-chilled trichloroacetic acid, and the radioactivity bound on protein was counted. The counts were plotted as a function of pH and fitted to k inact ϭ k inact lim /(1 ϩ H/K a ). Here H and K a are the proton concentration and the acid dissociation constants of the Cys-124, respectively. In order to verify this method, peptide mapping was performed with T73A mutant as a representative. After carboxymethylation by [ 3 H]iodoacetic acid with 15 g of T73A at each pH, the labeled protein was divided into two portions as follows: one for measuring the total amount of bound radioactivity, and the other for digestion with 0.2 g of Lys-C endoprotease. The reversed phase chromatography of digested protein was done with Vydac C18 column (4.6 ϫ 250 mm, 300-Å pore size, 5-m particle size) at a flow rate of 0.5 ml/min. Solvent A was 0.1% trifluoroacetic acid in water, and solvent B was 80% acetonitrile in solvent A. The percentage of solvent B increased from 0 to 70% during the 40-min linear gradient. The radioactivity of each fraction was counted, and the composition of amino acid was determined. The Cys-124 was found on the major radioactive peak (fragment of NGRVLVHCREGYSRSPT) eluted at 24 min. The radioactivity counts from total protein and the counts from peptide fraction containing Cys-124 were simultaneously plotted as a function of pH and fitted to the above equation. To confirm the pK a values for Y78F the method of pseudo-first order inactivation rate constant was also used (20,25).
Rapid Reaction Kinetics-Pre-steady-state kinetic measurements of the hydrolysis of pNPP catalyzed by wild-type and mutant VHRs were performed at 30°C using a stopped-flow spectrophotometer (DX-17MV) from Applied Photophysics (Leatherhead, UK). The enzyme concentrations of wild-type and Y78F were 25 M each, and the concentration of pNPP was 20 mM in 0.1 M Tris, 0.05 M Bis-Tris, and 0.05 M acetate buffer system. After rapid mixing of pNPP and the enzyme, the increase in absorbance was monitored at 405 nm for 1, 5, and 10 s five times for each enzyme. The average of traces was fitted to the equation [pnitrophenolate] ϭ A e Ϫbt ϩ Bt ϩ C (15). Where A is the amplitude of the burst; B is the slope of the linear portion of the curve; C is the intercept of the line; b is the first order rate of burst, and t is time. The rate constants for formation (k 2 ) and breakdown (k 3 ) of phosphoenzyme intermediate were estimated from the equations of b ϭ k 2 ϩ k 3 and B ϭ k 2 k 3 /(k 2 ϩ k 3 ) (29).

RESULTS
Preparation of VHR Mutants-To evaluate the hydrogen bonding array present in His-123 of VHR ( Fig. 1), His-123 and residues linked to His-123 via a water molecule, namely Tyr-78 and Thr-73, were altered by site-directed mutagenesis. The His residue was changed to (a) alanine, in which case the imidazole ring was completely eliminated, to (b) phenylalanine, which has an aromatic ring but lacks hydrogen bonding capacity, and to (c) asparagine, which has the capability to form hydrogen bond without the imidazole ring. The Tyr-78 residue was changed to (d) phenylalanine to eliminate the hydrogen bonding capacity and the Thr-73 residue was replaced by (e) alanine, which lacks the hydroxyl group. We also simultaneously mutated Tyr-78 and Thr-73 to (f) phenylalanine and alanine, respectively. The six mutants prepared apparently had identical electrophoretic behavior to those of the wild-type VHR. To assess the structural integrity of the mutants, CD spectra of wild-type, Y78F, and T73A proteins were measured. The CD spectra over 190 -240 nm showed no significant differences among the wild-type and mutant VHRs (data not shown).
Effect of Mutation on the pK a of Cys-124 -Since the reactivity of thiol of active-site cysteine is due to its low pK a , the pK a values of mutant VHRs offer crucial information for assessing the catalytic capability of the mutants. Therefore, we examined the p K a values of wild-type and mutant VHRs using radioactive iodoacetate ( Table I). The carboxymethylation of cysteine by iodoacetate has been successfully tested in Yersinia PTP (25) and VHR (19). The pK a values were estimated from the radioactivity bound to proteins at different pH values (Fig. 2). The pK a of 5.52 (5.44 -5.62) estimated for wild-type VHR agrees well with the reported value of 5.6 Ϯ 0.1 (20). The pK a values of His-123 mutants, H123A and H123F, were 6.88 (6.75-7.07) and 6.93 (6.82-7.08), respectively. These values are at least 1.3 pH units higher than that of the wild-type VHR. The shift of pK a value of the active-site cysteine by eliminating the hydrogen bonding capacity of His-123 could be compared with the pK a shift reported on His-402 mutants of Yersinia PTP (25). However, the H123N mutant, which still forms hydrogen bonding with neighboring residues, showed a similar pK a value of 5.47 (5.37-5.58) compared with that of the wild-type enzyme. The second series of mutants, Y78F and T73A, were designed to investigate the role of amino acids around His-123. The pK a values of the active-site cysteine in mutants, Y78F and T73A, were found to be 6.51 (6.38 -6.69) and 6.75 (6.68 -6.90), respectively. These results are intriguing, because these mutants disturb only the hydrogen bond connected to the N-⑀ side of imidazole ring of His-123 via a water molecule, and it is not apparent how the distant hydrogen bonds would affect the active-site cysteine (Fig. 1). To confirm further the pK a value obtained, we carried out peptide mapping for 3 H-carboxymethylated cysteine of T73A mutant (Fig. 2B) and measured pseudo-first order inactivation rate constant by iodoacetate for Y78F mutant (Fig. 2C). The pK a values determined by the peptide mapping or the inactivation kinetics are in good agreement with the corresponding values estimated by direct counting method of total protein. When the effect of double mutation of T73A/Y78F on the pK a of cysteine was examined, the highest pK a of 7.04 (6.94 -7.15) was observed.
Steady-state Kinetics of Wild-Type and Mutant VHRs-The steady-state kinetic parameters of wild-type and mutant VHRs were obtained using pTyr as a substrate (Table II). The k cat and K m values of wild-type VHR were determined to be 3.08 Ϯ 0.34 s Ϫ1 and 3.69 Ϯ 0.42 mM, respectively. These values are reasonably close to those previously reported (15) under similar conditions (5.87 Ϯ 0.37 s Ϫ1 for k cat and 2.20 Ϯ 0.38 mM for K m ). The His-123 mutants, particularly H123A and H123F, exhibited k cat values that were ϳ10 and 4% that of the wild-type VHR. This is to be expected since these mutants having increased pK a of the cysteine means reduced reactivity of the active-site cysteine. However, the k cat of H123N, which did not significantly alter the pK a of the cysteine, was only minimally affected, indicating that asparagine still could manage to form a hydrogen bonding network. In the same token, mutations of Y78F and T73A, having increased pK a of the cysteine, showed substantial reductions on their kinetic parameters with ϳ4and 14-fold decreases in k cat and ϳ11-fold and 36-fold decreases in k cat /K m , respectively. For the double mutant T73A/ Y78F, similar reduction in k cat was observed. However, a more dramatic reduction of 480-fold in k cat /K m was observed due to the unusually high value of K m . To substantiate further our kinetic analysis, we also carried out kinetic study of D92N mutant, which has been studied extensively by Denu et al. (20). The kinetic parameters of D92N obtained with pTyr were comparable to the reported values.
Effect of pH on Kinetic Parameters of Y78F and T73A Mutants-Observation of the marked alterations in kinetic parameters obtained with Y78F and T73A mutants led us to investigate the pH dependence of the kinetic parameters using pTyr as a substrate (Fig. 3). Determination of kinetic parameters, k cat and k cat /K m , as a function of pH could establish the residues that function as acid or base catalysts (30). The k cat /K m parameter is the apparent second order rate constant for the reaction of substrate and free enzyme. Therefore, in the pHk cat /K m profile, ionization residues that are important for binding of substrate and/or catalysis through the first irreversible step would be displayed, and the pH-k cat profile would display the ionization state of residues involved in the rate-limiting step in the overall reaction. The pH-k cat /K m profiles for wildtype, Y78F, and T73A displaying bell shapes, with an ascending slope of approximately ϩ2 and a descending slope of approximately Ϫ1, were observed (Fig. 3A). These are mostly consistent with the known participations of one unprotonated residue together with the dianion of phosphate monoester for the ascending curve and a protonated residue for the descending curve during the VHR-catalyzed formation of phosphoenzyme intermediate (20). However, there was a small but significant shift in optimal pH of the mutants toward alkali compared with that of the wild type. When k cat was plotted against pH, the mutants showed a more drastic difference from  that of the wild-type enzyme (Fig. 3B). The descending part (alkaline wing) of the normal bell shape disappeared in the mutants. Since the alkaline wing reflects the participation of protonated residue in the rate-limiting step, these k cat alterations of mutants are likely due to shift of the ionization state of protonated residue(s) involved in the breakdown of phosphoenzyme intermediate (Scheme 1).
The pK a values of ionizable group involved in the pH dependence data of pTyr substrate were analyzed, and the results are listed in Table III. For wild-type VHR, the analysis of k cat /K m shows pK a values of 5.61 (pK 1 ) and 5.43 (pK 2 ) as the two unprotonated species, whereas a pK a value of 5.71 (pK 3 ) seems to reflect the protonated residue. Here the pK 1 of 5.61 should be the unprotonated dianion of substrate pTyr (pK a2 of pTyr). The pK 2 of 5.43, another unprotonated group calculated from pH-k cat /K m profile, seems to correspond to the pK a of Cys-124. The value of pK 2 is in good agreement with the pK a of Cys-124 determined by iodoacetate titration (Table I). The most probable candidate for protonated residue (pK 3 ) could be the Asp-92, since Asp-92 acts as a general acid catalyst for formation of the phosphoenzyme intermediate. These pK a (pK 1 , pK 2 , and pK 3 ) values obtained with pTyr substrate are basically similar to those derived from pNPP hydrolysis except the pK 1 (pK a2 of pTyr) (20). For mutants k cat /K m , the analysis shows that both pK 2 and pK 3 apparently increased by 1-1.7 and 0.3-0.7 pK units, respectively, indicating that the mutations affect unprotonated and protonated states of residues involved in substrate binding and/or catalytic step (k 2 ) for formation of phosphoenzyme intermediate. The analysis of pHdependent k cat values of wild-type enzyme indicates that one residue with a pK a value of 5.52 (pK 2 ) should be unprotonated and the other with a pK a value of 7.02 (pK 3 ) should be protonated in the rate-limiting step, likely the breakdown of phosphoenzyme intermediate in the case of substrate pTyr. The pK 3 components of k cat of mutants have not been observed as evidenced by the absence of a descending curve in the alkaline pH studied here. When the kinetics were carried out with PMP as a substrate, a substantially diminished effect of pH on k cat /K m was observed (Fig. 3C). For mutants, the pH dependence of k cat /K m values were further reduced.
Substrate Specificity of T78F and T73A Mutants-A series of studies on substrate specificity of VHR has established that VHR catalyzes hydrolysis of aromatic phosphate as well as alkyl phosphate, although the former is preferred (15,23,28). Since kinetic parameters of Y78F and T73A mutants determined with pTyr substrate reveal substantial changes from those of wild-type VHR, we extended our kinetic studies to different types of nonpeptide substrates (Table IV). Effectively similar k cat values for each wild-type or mutant VHRs were obtained with different substrates at pH 6.0 as expected by the fact that the breakdown of phosphoenzyme intermediate is the rate-limiting step for VHR-catalyzed hydrolysis of phosphate monoesters (15). However, there was a rather big discrepancy between k cat values for pNPP and pThr at pH 7.0. Like k cat /K m for Y78F and T73A measured using pTyr as substrate at pH 6.0 (Table II), the k cat /K m of mutants obtained with pNPP and pThr showed also a substantial reduction (8 -13-fold for Y78F and 16 -54-fold for T73A) compared with that of wild-type VHR. In contrast, the k cat /K m values of mutants observed with PMP as substrate show a rather small decrease (3-5-fold). The effects of substrates on k cat /K m are more evident when the ratios of k cat /K m values obtained with different substrates are compared with each other among the mutants. The 11-fold difference of k cat /K m values observed between PMP and pNPP in wild-type VHR jumps to a 120-fold difference in the T73A mutant at pH 6.0. The K m values obtained for each substrate are consistent with the results of substrate effect on k cat /K m values. All substrates examined except PMP showed increased K m values upon mutation. The unaffected K m values for PMP in mutants, together with the smallest K m value observed with PMP, suggest that the active site of VHR tends to be easily occupied by a large hydrophobic substrate.
Effect of Mutation on Turnover Rate-To ascertain the ratelimiting step of wild-type and Y78F mutant, pre-steady-state  kinetic experiments were carried out with 20 mM pNPP at 30°C (Fig. 4). The burst kinetic traces observed with both VHRs apparently revealed the rate-limiting step of Y78F was also the breakdown of the phosphoenzyme intermediate. From the burst kinetic data of wild type, k 2 (rate of leaving group release) and k 3 (rate of phosphoenzyme break) were estimated to be 33.5 and 1.93 s Ϫ1 , respectively (Scheme 1). These values are in reasonably good agreement with data reported previously (15). The value of k 3 is also reasonably similar to the k cat value of 3.27 s Ϫ1 obtained with pNPP by steady-state kinetics at pH 6.0 (Table IV). For Y78F, k 2 values was 27.2 s Ϫ1 , and k 3 was 0.61 s Ϫ1 . The estimated k 3 value of Y78F mutant is again similar to the k cat values obtained by steady-state kinetics with pNPP substrate at pH 7.0 (Table IV). As expected in burst kinetics the values of k 3 were consistently smaller than the values of k 2 . Although the k 3 value diminished significantly by mutation, the k 2 values were virtually unchanged. This result indicates that the effect of mutation mainly affects the k 3 , the turnover rate (k cat ). For further examination of the k 3 slowdown effect by mutation, we examined the effect of ethylene

Analysis of kinetic parameters for wild-type and mutant VHR proteins with pTyr as a substrate
Data were fit to Equations 1-3 for each parameter using the nonlinear data analysis program Axum.
, k cat ϭ k cat lim /(1 ϩ H/K 2 ). In equations, H is the proton concentration, K 1 , K 2 , and K 3 are the ionization constants of the groups involved in the reaction. (k cat /K m ) lim was pH-independent second-order rate constant, and k cat lim was pH-independent turnover number. Error ranges were expressed as Ϫlog (K 1,2,3 Ϯ S.E. k cat (s Ϫ1 ) 3 1.14 6.80 (6.68-6.95) a The second acid dissociation constant of phosphotyrosine (pK a2 ) was estimated from titration curve obtained with an autotitroprocessor at 30°C (ionic strength 0.1 M with NaCl).

TABLE IV Kinetic parameters of wild-type and mutant VHRs obtained with various non-peptide substrates
The phosphatase activity of each mutant was measured in indicated pH in 0.1 M Tris, 0.05 M Bis-Tris, 0.05 M acetate buffer system at 30°C. Errors were calculated by nonlinear least square fitting of the kinetic data to the Michaelis-Menten equation and expressed as Ϯ S.D. Two duplicated sets of experiments were conducted.  3. Effects of pH on kinetic parameters of wild-type, Y78F, and T73A VHRs. A, pH dependence of the k cat /K m value with pTyr as substrate. B, pH dependence of the k cat value with pTyr as substrate. C, pH dependence of the k cat/ K m value with PMP as a substrate. The VHR activity was measured at 30°C in 0.1 M Tris, 0.05 M Bis-Tris, and 0.05 acetate buffer system. E, wild; q, Y78F; f, T73A. glycol (Fig. 5). In the presence of alcohol, the phosphoryl group of phosphoenzyme intermediate of VHR can be transferred to alcohol in addition to water. This phosphotransfer activity hence accelerated the turnover rate (k cat ) of VHR (24). The k cat of wild-type VHR was elevated by addition of ethylene glycol as reported previously (24). However, the k cat of Y78F showed little dependence on ethylene glycol. The second order rate constants for ethylene glycol determined with wild-type and Y78F were 0.23 and 0.037 M Ϫ1 s Ϫ1 , respectively.

DISCUSSION
Low pK a values of active-site cysteines observed in various PTPs support the notion that pK a is a key determinant for thiol reactivity of the P-loop at physiological pH (31). The low pK a value (5.52 (5.44 Ϫ 5.62)) of VHR cysteine obtained here is in the same range as PTP1 (5.57 Ϯ 0.12) (32). For some PTPs like Yersinia PTP and bovine low molecular PTP, the pK a values are even lower than 5 (17,25). Therefore, the increased pK a values found in His-123 mutants of VHR (H123A and H123F), which are at least 1.3 pH units higher than that of wild-type VHR (Table I), implicate an impairment of the normal thiol reactivity of Cys-124 by perturbing its configuration in the P-loop. The elimination of the hydrogen bond between the carbonyl oxygen of active-site Cys-124 and the imidazole ring of conserved His-123 apparently affects the cysteine at the center of the active site (Fig. 1). It is remarkable that the increases in pK a values of Y78F and T73A, the mutants with respect to neighboring residues, are comparable with those of His-123 mutants ( Table I). The Tyr-78 and Thr-73 residues are not part of the VHR P-loop itself but are located outside of the P-loop. One possible explanation is that the hydrogen bonding array of the carbonyl group from the active-site Cys-124 is extended to Tyr-78 and Thr-73 through two nitrogen atoms of the His-123 imidazole ring, and this array holds the imidazole ring in a critical orientation, so that the ring acts as a lever for positioning the cysteine in the P-loop (Fig. 1A). Therefore, any disruption of this hydrogen bonding array perturbs the position of Cys-124 at the center of the P-loop and thereby affects its pK a . This is quite reasonable since the unusually low pK a value of PTP active-site cysteine is stemmed from the strategic positioning of the cysteine at the hub of hydrogen bonding network including hydrogen bonding arrays radiating out from the Ploop to neighboring residues (4,6,7,18). In contrast to the sensitivity of the pK a values to mutations in Tyr-78 and Thr-73, mutations in Asp-92 and Ser-131, the other key residues critical for catalytic action of VHR (Fig. 1B), such as D92N and S131A, were reported not to significantly affect the pK a value of the Cys-124 (14,20).
The consequence of mutations that affect the pK a value of the cysteine would alter the catalytic capability of the enzyme. As suspected, substantial reductions of steady-state kinetic parameters were observed for mutants with pTyr as a substrate (Table II). Particularly, the 4-fold reduction in k cat and 11-fold reduction in k cat /K m for Y78F and 14-fold reduction in k cat and 36-fold reduction in k cat /K m for T73A are comparable with those of H123A and H123F mutants, suggesting that mutations in neighboring residues affect the Cys-124 position similarly to those in His-123. This similarity of kinetic parameters supports again that Tyr-78 and Thr-73 constitute an essential part of the histidine hydrogen bonding array and that these residues function by structurally tuning the position of His-123 in the histidine array.
A dislocation of the active-site cysteine in the P-loop means alteration of the relative position of other key residues, mainly Asp-92 and Ser-131, within the P-loop, besides the increased pK a of the cysteine itself (Fig. 1B). Therefore, mutational effects on the catalytic patterns such as the pH dependence of kinetic parameters and the effect of ethylene glycol on turnover rate (k cat ) were investigated. As evidenced by the burst kinetic traces of p-nitrophenol derived from pNPP substrate (Fig. 4), the breakdown of phosphoenzyme intermediate (k 3 step) is the rate-limiting step for both wild-type and Y78F mutant VHRs. This observation suggests that the overall catalytic mechanism of VHR is not significantly affected by the mutation, although the mutation markedly diminished the k 3 value. The k 2 values that were virtually unchanged by the mutation implicate that the thiol nucleophilicity of the active-site cysteine is not significantly altered by the mutation in Tyr-78. A small but significant shift in pH optima of k cat /K m versus pH curves observed in mutants (Fig. 3) appears to be a manifestation of subtle alteration of the Cys-124 position in the core of the P-loop. In the view of catalytic mechanism depicted in Scheme 1, pK 1 and pK 2 , the ionization of constants observed from k cat /K m -pH profile for unprotonated species, have to be assigned to the reactive dianion of pTyr substrate (pK a2 of pTyr) and to the nucleophilic thiolate anion (pK a of Cys-124), respectively (Table III). The residue responsible for pK 3 , which must be protonated for activity, should be the Asp-92. The protonated aspartic acid has been known to act as a general acid in the release of leaving group (k 2 step). When Asp-92 was completely eliminated as in the D92N mutant, the alkaline wing (descending part) of the k cat /K m -pH profile was totally absent (20).
Similarly, the alterations of the k cat values versus pH curves SCHEME 1. Kinetic mechanism for the hydrolysis of phosphomonoester by VHR. The kinetic mechanism of VHR is usually defined by two chemical steps as follows: one for formation of a thiophosphateenzyme intermediate (k 2 ), and the other for breakdown of the phosphoenzyme intermediate (k 3 ). Transition states corresponding to these steps are depicted. The Asp-92 acted as general acid catalyst in the transition state for the leaving group release (k 2 step), whereas the same Asp acted as general base catalyst in the phosphoenzyme break (k 3 step). The hydroxyl group of Ser-131 is transiently linked to the thiolate to expel phosphate anion. observed in the mutants imply a perturbation of the relative position of both Asp-92 and Ser-131 residues against the Cys-124. The increased pK 2 in k cat appears to correspond to the ascending slope of unprotonated Asp-92 (Table III). It should be kept in mind, however, that the catalytic contribution of Asp-92 in k cat is different from that in k cat /K m , since k cat represents the breakdown of phosphoenzyme intermediate (k 3 step), and the aspartate acts as a general base (Scheme 1). The protonated residue responsible for pK 3 derived from k cat -pH profile ought to be the Ser-131 residue. This assignment appears unrealistic because of the low pK 3 (7.02) observed in wild-type VHR. However, when considering the transient nature of the hydrogen bond rendered by Ser-131 to thiolate to expel phosphate anion during the breakdown of phosphoenzyme intermediate and the insensitivity of the descending part of mutants to pH (absence of the pK 3 component), the protonated state of Ser-131 appears to be very sensitive on the neighboring hydrogen bonding network surrounding this residue. It is most likely that the hydroxyl group of Ser-131 involved in the dephosphorylation step (k 3 ) requires an optimal distance as well as a precise alignment to the leaving thiolate (Scheme 1 and Fig. 1B). This interpretation is consistent with the marked decrease of turnover rate (k cat ) observed upon mutating Ser-131 to alanine in the catalytic mechanism of VHR (14). However, the exceptionally small k cat values observed for pThr at pH 7.0 might implicate that the pH-dependent change of rate-limiting step occurred. In this case, the k cat -pH profile for pThr would reflect the ionization of groups involved in phosphorus-oxygen bond cleavage (k 2 step), which would mean that the unprotonated group is Cys-124 and the protonated group is Asp-92.
The reduced dependence of k cat on ethylene glycol nucleophile observed with Y78F (Fig. 5) again implies that the mutation affects the relative position of Asp-92 and Ser-131 to Cys-124. Since Asp-92 acts as a general base to promote nucleophilic attack of a water molecule in the breakdown of the phosphoenzyme intermediate (k 3 step) (Scheme 1), distortion of the Asp-92 position by mutation definitely reduces the transfer reactivity of ethylene glycol. Similarly, the conserved Ser-18 and Asp-128 residues in Stp1 have been previously recognized as critical residues involved in the transferase activity of the enzyme (33).
One characteristic of the substrate specificity of VHR is that VHR prefers aromatic phosphate to alkyl phosphate (15). The kinetic data obtained with different types of nonpeptide substrates showed substantial changes from wild-type VHR (Table  IV). These kinetic data confirm the key feature of substrate specificity of VHR, namely that PMP is a better substrate than pNPP, and pThr is the poorest among those examined. However, unlike other substrates, the kinetic parameters of mutants for PMP showed rather small decreases, particularly in k cat /K m , implying that the PMP fills up the active site relatively well even in the mutants. However, the small K m values for substrate PMP observed in mutants could not be assessed in terms of the binding affinity for PMP, because K m values are not generally representing the true equilibrium binding constants. Nonetheless, it appears that the alteration of the P-loop geometry brought about by the mutations is not enough to affect the K m values for PMP. Since the pK a value of 9.4 for PMP is comparable to those of 7.14 for pNPP and 10.07 for pTyr (15,23), the possibility that PMP might be expelled without being protonated could be ruled out.
The results of mutational and kinetic experiments support the pivotal role of conserved His-123 as a positioning lever for the active-site Cys-124. Furthermore, positioning of cysteine requires not only the conserved His-123 but also Tyr-78 and Thr-73 to orient the imidazole ring. In a structure-based sequence alignment, Tyr-78 of VHR is one of 10 invariant residues recognized with other PTPs (7). Like Tyr-78 and Thr-73 in VHR, the crystal structure of Yersinia PTP reveals Tyr-301 and His-270 residues hydrogen-bonded to orient the imidazole ring of His-402 (4). Indeed, a common structural feature of hydrogen bonding between histidine and neighboring residue(s) is found to be well preserved in a large number of PTPs. Examples of the invariant tyrosine residues playing a structural role are Tyr-124 of PTP1B (6), Tyr-366 of SHP-1 (9), and Tyr-729 of CD45 (34). Mutational study of CD45 showed that only Tyr-729 in domain I was specifically required for activity (34). Therefore, we conclude that the invariant tyrosine residue is important for the catalytic activity of PTP through hydrogen bonding to the conserved histidine. In other words, without proper orientation of the imidazole ring of histidine by the invariant tyrosine residue, the precise positioning of cysteine is impaired and the catalytic activity disrupted accordingly. Thus our mutational and kinetic approaches unveiled the possible structural role of the N-⑀ side of the imidazole ring in the highly coordinated hydrogen bonding array of the conserved histidine in modulating the catalytic activity of PTP.