Functional characterization and crystal structure of the C215D mutant of protein-tyrosine phosphatase-1B.

We have characterized the C215D active-site mutant of protein-tyrosine phosphatase-1B (PTP-1B) and solved the crystal structure of the catalytic domain of the apoenzyme to a resolution of 1.6 A. The mutant enzyme displayed maximal catalytic activity at pH approximately 4.5, which is significantly lower than the pH optimum of 6 for wild-type PTP-1B. Although both forms of the enzyme exhibited identical Km values for hydrolysis of p-nitrophenyl phosphate at pH 4.5 and 6, the kcat values of C215D were approximately 70- and approximately 7000-fold lower than those of wild-type PTP-1B, respectively. Arrhenius plots revealed that the mutant and wild-type enzymes displayed activation energies of 61 +/- 1 and 18 +/- 2 kJ/mol, respectively, at their pH optima. Unlike wild-type PTP-1B, C215D-mediated p-nitrophenyl phosphate hydrolysis was inactivated by 1,2-epoxy-3-(p-nitrophenoxy)propane, suggesting a direct involvement of Asp215 in catalysis. Increasing solvent microviscosity with sucrose (up to 40% (w/v)) caused a significant decrease in kcat/Km of the wild-type enzyme, but did not alter the catalytic efficiency of the mutant protein. Structurally, the apoenzyme was identical to wild-type PTP-1B, aside from the flexible WPD loop region, which was in both "open" and "closed" conformations. At physiological pH, the C215D mutant of PTP-1B should be an effective substrate-trapping mutant that can be used to identify cellular substrates of PTP-1B. In addition, because of its insensitivity to oxidation, this mutant may be used for screening fermentation broth and other natural products to identify inhibitors of PTP-1B.

The protein-tyrosine phosphatases (PTPases) 1 compose a family of ϳ100 enzymes that play an important role in controlling several biological processes, including cell cycles and signal transduction pathways (1,2). One member of this family of phosphatases that is receiving increased attention because of its potential role in controlling the insulin signaling pathway is protein-tyrosine phosphatase-1B (PTP-1B). Studies have suggested that this ubiquitously expressed enzyme may play a role in regulating the function of the insulin receptor. Thus, PTP-1B is an attractive target for drug design in the treatment of Type II diabetes (for reviews, see Refs. 3 and 4).
Like other PTPases, PTP-1B contains a conserved 11-residue sequence motif (i.e. (I/V)HCXAGXXR(S/T)G) that harbors Cys 215 , which acts as the nucleophile and is essential for catalysis. The signature motif also forms the "P-loop" that is involved in substrate binding and catalysis. It is thought that PTP-1B-mediated catalysis occurs via a double-displacement mechanism in which the phosphoryl group of the substrate is first transferred to the active-site Cys residue (Cys 215 ) (5,6). The initial phosphoryl transfer is assisted by an invariant Asp residue (Asp 181 ) residing in a flexible loop region (WPD loop) that spans the conserved tripeptide Trp-Pro-Asp. It is generally believed that Asp 181 first acts as a general acid and protonates the leaving group in the phosphorylation step. Subsequently, Asp 181 functions as a general base, abstracting a proton from an attacking water molecule in the dephosphorylation step to enhance the rate of hydrolysis of the enzyme-thiophosphate intermediate (7)(8)(9)(10). Previous structural studies on PTP-1B have revealed interesting details regarding the conformations and structural organizations of the WPD loop and P-loop regions. Specifically, the WPD loop has been shown to adopt different conformations in the unliganded and liganded forms of the enzyme. In the unliganded structure, the WPD loop is in an open conformation, in which Asp 181 is ϳ10 Å away from the P-loop. Upon substrate binding, the WPD loop adopts a closed conformation and covers the active site like a "flap," thereby positioning Asp 181 closer to the leaving group (11,12).
In wild-type PTP-1B, Cys 215 is present as a thiolate (13), and it is known that this active-site residue is absolutely necessary for PTP-1B-mediated catalysis. Mutation of this residue to a neutral Ser generates a "substrate-trapping" mutant, which is able to bind substrates with affinities similar to those of the wild-type enzyme, but does not display any measurable phosphatase activity (14,15). The crystal structure of the unliganded C215S PTP-1B mutant shows the P-loop in a conforma-tionally distinct orientation compared with that found in the wild-type protein. However, in the liganded form, the P-loop adopts the same conformation as the wild-type protein. In the C215S mutant, substitution of the negatively charged thiolate with a neutral (although polar) alcohol destabilizes the PTPase signature motif loop (P-loop) and the surrounding areas, favoring the extended conformation (16). The structural studies therefore suggest that the conformation of the P-loop region of the enzyme is inducible and may be dependent on the presence of the negative charge of the active-site nucleophile. The goal of this study was to explore the importance of the presence of a negatively charged residue other than Cys at position 215 in the conformation of the P-loop and in the catalytic activity of PTP-1B. We therefore substituted Cys 215 with Asp, as this residue is similar in charge and size density to the active-site thiolate. Here, we report the functional characterization and crystal structure of the C215D mutant enzyme and compare these properties with those of wild-type PTP-1B.

EXPERIMENTAL PROCEDURES
Materials-Bovine serum albumin and the SuperSignal TM West Pico immunoblotting kit were obtained from Pierce. SDS-polyacrylamide gels were purchased from Invitrogen, and Escherichia coli BL21 cells were obtained from Stratagene. Isopropyl-1-thio-␤-D-galactopyranoside and protease inhibitors were from Roche Applied Science. The Cibacron blue and Sepharose Q columns were bought from Amersham Biosciences. All other chemicals were obtained from Sigma.
General Methods-Protein concentrations were determined by a microplate adaptation of the Bradford assay using bovine serum albumin as a standard (47). For SDS-PAGE and immunoblotting of purified mutant and wild-type PTP-1B, proteins were boiled in SDS-PAGE loading buffer for 5-10 min and then separated on 10 -20% (w/v) polyacrylamide gradient gels containing SDS. The proteins were transferred onto nitrocellulose at a constant voltage of 100 V for 90 min. The blot was subsequently probed with an anti-FLAG monoclonal antibody and detected by chemiluminescence. The SuperSignal TM West Pico kit was used for immunoblotting of proteins following the manufacturer's suggested protocol.
Mutagenesis and Protein Expression-A plasmid containing the isolated catalytic domain of human PTP-1B (residues 1-298) in a pFLAG vector was used as a template for site-directed mutagenesis. The Cysto-Asp mutation was introduced using a PCR-based approach following established procedures (17). The resulting construct was verified by DNA sequencing on an ABI 373 DNA sequencer (Applied Biosystems), and the sequence data were analyzed using the software application package Sequencher Version 4.0.5 (Gene Code Corp.).
The resulting plasmid was transformed into E. coli BL21 cells for protein expression. The bacterial cultures were grown in LB broth at 37°C with shaking at 250 rpm to an absorbance of 0.7. The cultures were then induced by the addition of 1 mM isopropyl-1-thio-␤-D-galactopyranoside and grown for an additional 2 h. Cells were harvested by centrifugation, and cell lysates were analyzed by electrophoresis and immunoblotting using the protocol outlined above.
Purification of Mutant and Wild-type PTP-1B-Purification of mutant and wild-type PTP-1B was performed as outlined previously (18). Briefly, E. coli BL21 cells transformed with the expression vector were suspended in lysis buffer consisting of 20 mM Tris-HCl, 0.1 mM EDTA, 5 mM dithiothreitol, and two tablets of Complete TM protease inhibitors/50 ml of solution at a final pH of 7.5. Cell lysis was achieved by passing cells twice through an ice-chilled French pressure cell (SLM-AMINCO) at 16,000 p.s.i. The supernatant was retained following centrifugation of the cell lysate at 31,000 ϫ g for 30 min and then applied to a Cibacron blue affinity column pre-equilibrated with lysis buffer without protease inhibitors (binding buffer A). Following application of the supernatant, the affinity column was washed with 20 column volumes of binding buffer A or until the absorbance at 280 nm returned to base-line levels. The protein was then eluted with a linear gradient of NaCl (0 -500 mM) in binding buffer A over 10 column volumes. PTP-1B-rich fractions were determined by analyzing the fractions by SDS-PAGE and also by measuring PTPase activity. The appropriate fractions were pooled, and the ionic strength was adjusted to 100 mM NaCl. The partially purified preparation was subsequently applied to a Sepharose Q anion-exchange column pre-equilibrated with 20 mM Tris-HCl, 0.1 mM EDTA, 5 mM dithiothreitol, and 100 mM NaCl at pH 7.5 (binding buffer B). The column was then washed with binding buffer B to return the absorbance at 280 nm to base-line levels. Next, the protein was eluted with a linear gradient of increasing NaCl concentrations (100 -500 mM) over 8 column volumes. The eluted fractions were analyzed by SDS-PAGE and by measuring enzyme activity; the appropriate fractions were then pooled. The purified protein preparation was dialyzed into 20 mM Tris-HCl, 0.1 mM EDTA, 5 mM dithiothreitol, 150 mM NaCl, and 20% (v/v) glycerol at pH 7.5 prior to storage at Ϫ80°C. All steps in the purification scheme were carried out at 4°C or on ice with the aid of a fast protein liquid chromatography apparatus (Amersham Biosciences); a flow rate of 1 ml/min was used in all chromatographic steps.
Activity Assays-Assays were carried out in a 96-well format at 22°C in buffer consisting of 100 mM Tris-HCl, 50 mM MES, 50 mM acetic acid, N,NЈ-dimethylbis(mercaptoacetyl)hydrazine, 2 mM EDTA, 5% (v/v) Me 2 SO, 2% (v/v) glycerol, and 0.01% (v/v) Triton X-100 at the appropriate pH. The use of this triple-component buffering system minimizes changes in ionic strength across a pH range of 3-9 (19). Enzyme activity was quantitated by monitoring C215D or wild-type PTP-1B-catalyzed hydrolysis of p-nitrophenyl phosphate (pNPP) to p-nitrophenol. Briefly, pNPP hydrolysis was measured by incubating C215D or wild-type PTP-1B with 0 -10 mM pNPP for 30 or 4 min, respectively, and then stopping the reaction by adding NaOH to a final concentration of 1 M. The absorbance at 405 nm was measured on a Cytofluor II TM plate reader, and the catalytic activity was calculated using the molar extinction coefficient of the p-nitrophenolate anion (18,800 M Ϫ1 cm Ϫ1 ). The observed rates of reactions were fitted to the Michaelis-Menten equation using the nonlinear curve fitting software program Grafit Version 4.0.10 (Erithacus Software Inc.) to determine kinetic constants. To extract pK app values from the pH-response studies, the data were also fitted using nonlinear regression analysis (SigmaPlot, Jandel Scientific).
For collection of Arrhenius plot data, C215D and wild-type PTP-1B activities were measured at their respective pH optima at 10 -37°C (Ϯ0.5°C) using 10 mM pNPP as a substrate and an assay time of 30 or 4 min, respectively. The data were transformed into Arrhenius plots, and the slopes of the lines were obtained using linear regression with the Marquardt-Levenberg algorithm (SigmaPlot).
For inactivation assays using 1,2-epoxy-3-(p-nitrophenoxy)propane (EPNP), the epoxide was dissolved in Me 2 SO and added to the reaction mixtures to a final concentration of 2.4 mM. Following incubation of the enzymes for 1 h at 4°C in the presence of EPNP, the catalytic activities were determined at their respective pH optima as described above.
Viscosity Studies-The effect of viscosity on C215D or wild-type PTP-1B activity was determined by measuring pNPP hydrolysis using the protocol outlined above in reaction mixtures containing 0 -40% (w/v) sucrose. Stock solutions of sucrose were prepared at twice the desired final concentration in assay buffer (pH 4.5 or 6) and added to the reaction mixtures to obtain a 2-fold dilution. The plates were continuously shaken during the incubation time to maintain homogeneity of the reaction mixtures. Relative solvent viscosities ( rel ϭ / o , where the superscript o denotes the reaction in buffer lacking sucrose) were calculated using the solution densities at 22°C (20,21). The calculated relative viscosities used in these experiments were 1.00, 1.37, 2.02, 3.32, and 6.39 for 0, 10, 20, 30, and 40% (w/v) sucrose solutions, respectively.
Crystallization and Data Collection-Apo-C215D crystals were obtained by vapor diffusion in sitting drops at 4°C by mixing 2 l of protein (10 mg/ml in 20 mM HEPES, 50 NaCl, 1 mM EDTA, and 5 mM N,NЈ-dimethylbis(mercaptoacetyl)hydrazine, pH 7.0) and 2 l of precipitant solution (13-16% polyethylene glycol 3350, 100 mM HEPES, and 200 mM MgCl 2 , pH 7.0). X-ray diffraction data were collected on an ADSC Q210 detector from a single crystal (ϳ0.3 ϫ 0.2 ϫ 0.1 mm in size) using synchrotron radiation. Data to 1.6 Å were collected at beamline 17-ID in the facilities of the Industrial Macromolecular Crystallography Association Collaborative Access Team at the Advanced Photon Source. Data processing, scaling, and merging were done with the software DPS/MOSFLM (22,23). The crystal was trigonal, with space group P3 1 2 and unit cell parameters a ϭ b ϭ 88.45 Å, c ϭ 104.36 Å, ␣ ϭ ␤ ϭ 90.0°, and ␥ ϭ 120°. See ''Results'' for the statistics for the data collected.
Structure Solution and Refinement-The crystal was isomorphous to previously reported PTP-1B crystals (i.e. Protein Data Bank 1PTY) (24), and the three-dimensional structure of the C215D mutant was solved by difference Fourier analysis using, as the initial model, the 1.8-Å structure of the mutant enzyme in complex with phosphotyrosine (Protein Data Bank code 1PTY). Bound ligand, solvent molecules, and protein residues 175-184 (WPD loop) and 213-223 (P-loop) were deleted from the coordinate file. The initial electron density maps to 2.4 Å calculated from this model showed that the WPD loop was in a closed conformation and that the peptide containing the catalytic site mutation assumed the same conformation observed in the wild-type enzyme (see Fig. 4). Both loops were built into the available density using the graphic software O (25). Refinement of the model was carried out by alternating cycles of manual rebuilding of the model in O and computerbased refinement using CNX, slowly including all available data to 1.6-Å resolution. Typically, two cycles of torsion angle dynamics and positional and temperature factor refinement were run in each cycle. Bulk solvent correction was applied throughout the entire refinement, and the refinement was performed using the cross-validated maximal likelihood approach (26,27). When high resolution data were included, it became evident in the electron density maps that the WPD loop (Thr 177 -Ser 188 ) was present in both "open" and "closed" conformations. Several other residues were also modeled as having dual conformation for their side chains; the occupancy for atoms in dual conformations was initially set to 50% for each conformer and then manually adjusted to reflect the temperature factors. At the end of the refinement, 30 additional cycles of occupancy refinement were carried out for the atoms modeled in the alternate configurations. See ''Results'' for the statistics for the refined model.

RESULTS
To evaluate the importance of a negatively charged residue at position 215 of PTP-1B, we substituted Cys 215 with Asp. The isosteric nature of this amino acid replacement minimizes the introduction of structural perturbations in the conformation of the P-loop. Following expression and purification of the C215D mutant, the functional and structural properties of the enzyme were compared with those of wild-type PTP-1B.
Effect of pH on C215D Activity-We investigated the pH dependence of C215D and wild-type PTP-1B-catalyzed hydrolysis of pNPP at pH 3.5-9 (Fig. 1). Although the pH profiles were generally bell-shaped for both enzymes, suggesting the existence of two ionizable amino acid side chains involved in catalysis, significant differences in the profiles were noted. As expected, wild-type PTP-1B displayed maximal catalytic activity at pH 6.0; however, when the active-site Cys was replaced with Asp, the pH optimum of the enzyme was shifted to 4.5-4.7. Interestingly, the pH profiles showed very little overlap, and significant differences in the relative levels of catalytic activity were observed at the pH optima of the two enzymes. For example, at pH 4.5, where the C215D mutant displayed maximal activity, the catalytic activity of wild-type PTP-1B was Ͻ5% of its maximum. Similarly, whereas wild-type PTP-1B exhibited maximal catalytic activity at pH 6, only 15% of the maximal catalytic activity of the C215D mutant was present. From the pH profiles, the first and second apparent ionization constants of the enzyme-substrate complexes (i.e. pK 1(app) and pK 2(app) ) for wild-type PTP-1B were estimated to be ϳ5.5 and 6.8, respectively. Due to the steep slope of the acid limb of the pH profile for the C215D mutant, the pK 1(app) could not be determined accurately, and it could only be estimated that the value was between 4.3 and 4.5. According to the pH profile, the pK 2(app) of the mutant enzyme (ϳ5.5) was also significantly lower than that of wild-type PTP-1B. These results clearly show that the two forms of the enzyme displayed distinct pH dependences with respect to substrate hydrolysis.
Comparison of K m , k cat , and E act of Mutant and Wild-type PTP-1B-Next, we compared the kinetic parameters of the C215D mutant with those of wild-type PTP-1B. Table I shows a comparison of the kinetic parameters of C215D and wild-type PTP-1B-mediated pNPP hydrolysis at the determined pH optima. Although both enzymes displayed similar K m values of 1.1-1.4 mM, the k cat values of the C215D mutant were ϳ70and ϳ7000-fold lower than those of wild-type PTP-1B at pH 4.5 and 6, respectively.
To determine whether the observed differences in k cat between the mutant and wild-type enzymes could be at least partially explained by potential differences in the thermodynamics of catalysis, we compared the activation energies (E act ) of C215D and wild-type PTP-1B-catalyzed pNPP hydrolysis by measuring the catalytic activities as a function of temperature. The Arrhenius plots and the corresponding calculated E act values for C215D and wild-type PTP-1B are shown in Fig. 2 and Table II, respectively. The Arrhenius plot for wild-type PTP-1B-mediated catalysis was continuous over the temperature range of 10 -35°C, with E act ϭ 18 Ϯ 2 kJ/mol. At temperatures above 35°C, significant denaturation of the enzyme was observed (data not shown). The C215D mutant was more thermally labile, and a continuous Arrhenius plot was obtained only up to 32°C. Above this temperature, the Arrhenius plot also began to level off due to thermal denaturation of the enzyme. Interestingly, the E act for the C215D mutant (i.e. 61 Ϯ 1 kJ/mol) was Ͼ3-fold higher than that for wild-type PTP-1B.
Inactivation of C215D-mediated Catalysis by EPNP-EPNP is a potent inactivator of enzymes that possess an active-site carboxyl residue (for example, see Refs. 28 -30). Irreversible inactivation is presumably due to alkylation of the active-site carboxyl residue by the epoxide moiety of EPNP. Because the C215D mutant contains a carboxyl residue in its active site, we wished to determine the effect of EPNP on the catalytic activity of this mutant derivative of PTP-1B. We therefore monitored the amount of C215D and wild-type PTP-1B catalytic activities that remained following a 1-h incubation of the enzymes at 4°C in the presence of 2.4 mM EPNP. The activity of the mutant protein was reduced by Ͼ80% following incubation of the enzyme in the presence of the inactivator (Fig. 3). However, the activity of wild-type PTP-1B was not significantly affected under similar conditions. As the only amino acid difference between the mutant and wild-type enzymes was the substitution  Crystal Structure of the Apo-C215D Mutant-To evaluate the effect of the Cys-to-Asp substitution on the tertiary structure of PTP-1B, we solved the x-ray crystal structure of the mutant enzyme and refined it against 1.6-Å data. The final model had a crystallographic R-factor of 18.6% (R free is 20.4%) for 59,409 reflections between 15.0 and 1.6 Å (5% flagged for R free calculation) and maintained good geometry (root mean square deviations for bond lengths and bond angles of 0.01 Å and 1.5°, respectively (Table III)). The backbone conformation of 91.6% of the residues was within the most favored regions of the Ramachandran plot, with none in disallowed regions, as defined using PROCHECK (31). The side chains of Met 3 , Asp 48 , Pro 87 , Arg 105 , Ser 118 , Leu 119 , Gln 157 , Glu 159 , Ser 190 , Ser 216 , Arg 221 , Cys 226 , Ile 246 , Met 253 , and Ser 285 were modeled as having dual conformations. Two peptides (His 60 -Asp 63 and Thr 177 -Ser 188 , the WPD loop) and five solvent molecules were also modeled as having alternate conformations. The P-loop is in the wild-type conformation (Fig. 4), and extensive hydrogen bond interactions are present between the Asp 215 carboxylate and the main chain and side chain nitrogens of Ser 216 , Ala 217 , Gly 218 , and Ser 222 , and the side chain oxygen of Ser 222 as depicted schematically in Fig. 4A. Asp 215 is also hydrogenbonded to one of several ordered water molecules located in the binding site. This water is located at the position normally occupied by one of the phosphate oxygens of Tyr(P) in the Protein Data Bank 1PTY structure (Fig. 4B) and makes similar hydrogen bond interactions with the main chain nitrogens of Ile 219 and Gly 220 . A second water molecule occupies the position of another phosphate oxygen, but appears to be quite mobile; and it has been modeled as having two distinct positions. This water molecule interacts with the main chain and ⑀-nitrogen of Arg 222 , the side chain oxygen of Asp 215 , and three water molecules. The other four solvent molecules, at least one of which appears to be quite delocalized, have been located in the binding site; together, these molecules form an extensive hydrogen bond network connecting Asp 215 to Asp 181 in the WPD loop (Fig. 4, A and B). The WPD loop assumes both conformations: the catalytically active closed conformation (preferred, with an average occupancy of 73.1% and an average temperature factor of 17.6 Å 2 ) and the inactive open conformation typical of the unliganded enzyme. This alternate position refined to a much lower occupancy (26.9% on average, with an   3. Effect of EPNP on C215D and wild-type PTP-1B-catalyzed pNPP hydrolysis. The catalytic activities of C215D and wildtype (WT) PTP-1B were measured following a 1-h incubation at 4°C in the absence (black bars) or presence (gray bars) of 2.4 mM EPNP using 10 mM pNPP as a substrate. The data represent the means of triplicate determinations, and activity is expressed as percent maximal activity of C215D and wild-type PTP-1B at pH 4.5 and 6, respectively. average temperature factor of 22.8 Å 2 ), but electron density was clearly available in the difference Fourier maps that could be justified only by assuming a partially open loop. Following the two different positions of the WPD loop, the side chain of Arg 221 also assumes two distinct conformations (with similar occupancies) that indeed correspond to the conformations previously observed in the apo-and liganded PTP-1B (Protein Data Bank codes 2HNP and 1PTY, respectively).
Effect of Viscosity on C215D and Wild-type PTP-1B-mediated Substrate Hydrolysis-As the major structural difference between the wild-type enzyme and the C215D derivative was observed in the conformation of the WPD loop, we wished to investigate this further. Therefore, we evaluated the contribution of loop motion to C215D and wild-type PTP-1B-mediated catalysis by monitoring the effect of solvent microviscosity. We measured the K m and k cat of enzyme-catalyzed hydrolysis of pNPP in the presence of increasing concentrations of sucrose. As shown in Fig. 5A, the k cat of the wild-type enzyme was slightly higher in the presence of 40% (w/v) sucrose. However, there was a 30% decrease in k cat /K m of wild-type PTP-1B (Fig.  5B), suggesting that the catalytic efficiency of wild-type PTP-1B was significantly hindered in the presence of sucrose.
In contrast, there was no change in either the K m or k cat of C215D-mediated catalysis under similar conditions, resulting in a lack of deviation of (k cat /K m ) o /(k cat /K m ) as a function of relative solvent microviscosity (Fig. 5, C and D). DISCUSSION Previously, we reported that substitution of Cys 215 with Ser results in a conformationally distinct P-loop in the unliganded mutant enzyme, and we suggested that a negative charge in the active site is required to maintain the P-loop conformation observed in the wild-type enzyme (16). In this study, we generated and characterized the C215D mutant of PTP-1B to obtain further insights into the structural basis of PTP-1B catalysis. We substituted Cys 215 with Asp to preserve both the charge and size of this active-site residue.
Substitution of the active-site Cys with Asp resulted in a dramatic change in the pH profile for PTP-1B-catalyzed hydrolysis of pNPP. As expected, wild-type PTP-1B displayed a pH optimum of ϳ6. This value agrees favorably with the optimum values of 5.5-6.5 determined previously for the enzyme (32,33). In contrast, the pH optimum for the C215D mutant was shifted to a significantly lower value of 4.5-4.7 and likely reflects the fact that the active-site Cys, which displays a pK a of ϳ5.5 in wild-type PTP-1B (13), was replaced with the strongly acidic Asp residue, with a pK a of ϳ4 in model systems. The lower pH optimum of the C215D derivative is similar to the pH optima of some enzymes that contain active-site Asp residues such as pepsin (28) and the simian immunodeficiency virus protease (34).
Interestingly, both forms of PTP-1B displayed identical K m values at the pH optimum of the C215D mutant. The K m values of 5.4 and 5.8 mM for pNPP for both mutant and wild-type PTP-1B, respectively, were 4-fold higher than the K m values determined at pH 6. That both PTPases displayed the same shift in K m at a lower pH suggests that the increase in K m was not a result of substitution of the active-site Cys with Asp. Rather, the increase in K m may be due to the influence of pH on the ionization state of the substrate or may be a result of alterations in the ionization states of other important residues in the vicinity of the active-site regions of the proteins. At pH 4.5, C215D-mediated pNPP hydrolysis still remained Ͼ70-fold lower than wild-type enzyme-mediated pNPP hydrolysis at the same pH. Thus, substitution of the strongly nucleophilic activesite Cys with negatively charged Asp caused a significant reduction in the rate of substrate turnover that could not be reversed to that of the wild-type enzyme at the pH optimum of the mutant protein.
The Arrhenius plots for C215D and wild-type PTP-1B revealed that the activation energy of C215D-mediated pNPP hydrolysis was Ͼ3-fold higher than that of the reaction catalyzed by wild-type PTP-1B. The plots were linear over temperature ranges of 10 -32°C for the C215D mutant and 10 -35°C for the wild-type enzyme and did not reveal any discontinuities over the temperature ranges. Thus, the large drop in k cat observed for the mutant protein with respect to pNPP hydrolysis may be partially attributable to the 3-fold higher activation energy of C215D-mediated catalysis. The crystal structure of the apo-C215D mutant shows several water molecules sequestered at the active site. These water molecules would have to be excluded from the active site to enable substrate binding and catalysis. Thus, we hypothesize that the energy penalty that would result from the desolvation of the active site would translate to the higher activation energy observed for the mutant protein.
EPNP is a potent inactivator of enzymes that possess an active-site carboxyl residue and has been used to study the kinetic mechanisms of various aspartyl proteases such as pepsin (28) and the simian and human immunodeficiency virus proteases (for example, see Refs. 30 and 35). Inactivation by this uncharged molecule is presumably due to alkylation of the active-site carboxyl residue by the epoxide moiety of EPNP. Inactivation of the C215D derivative by EPNP and the lack of an effect on wild-type PTP-1B suggest a direct involvement of Asp 215 in the catalytic mechanism of the mutant protein. Previously, Zhang et al. (36) showed that EPNP also acts as an irreversible inactivator of the low molecular weight PTPase from bovine heart. In the case of this PTPase, however, two cysteine residues were proposed to be the target of the epoxide. As EPNP did not significantly inhibit or inactivate wild-type PTP-1B, it seems reasonable to suggest that the irreversible inactivation of the C215D mutant was a result of a chemical modification of the active-site Asp by the epoxide as observed for aspartyl proteases. These results corroborate the pH studies and provide strong evidence that Asp 215 is crucial to the reaction mechanism of the mutant protein.
Structurally, wild-type PTP-1B and the C215D derivative are identical, with the exception of the so-called WPD loop, which appears in both open and closed conformations in the mutant protein. The closed conformation is clearly favored over the open conformation (relative occupancies of 73.1 and 26.9%, respectively) and is probably induced by the extensive hydrogen bond network, involving both solvent and protein atoms, identified in the binding site (Fig. 4). These differences in the flexible WPD loop prompted us to investigate whether solvent microviscosity could affect the catalytic properties of these two enzymes. If rapid movement of the WPD loop is crucial to the catalytic mechanism of PTP-1B, then it is possible that increasing solvent microviscosity could have a detrimental effect on PTP-1B-mediated catalysis by imposing a physical energy barrier to the movement of this region. The catalytic domain (residues 1-298) of wild-type PTP-1B was influenced by solvent microviscosity, resulting in a 30% decrease in k cat /K m in the presence of 40% (w/v) sucrose, primarily stemming from a higher K m value. In addition, we have performed similar viscosity studies on the highly homologous T-cell PTPase and have found that the catalytic efficiency of the corresponding region of this enzyme (i.e. amino acids 1-296) responded to changes in solvent microviscosity in a manner similar to PTP-1B (data not shown). Specifically, k cat /K m of T-cell PTPase-catalyzed pNPP hydrolysis was ϳ1.5-fold lower in the presence of 40% (w/v) sucrose. The decrease in catalytic efficiency in the presence of sucrose is consistent with the hypothesis that movement of the WPD loop is crucial to the catalytic mechanism of PTP-1B. However, increasing the relative viscosity to the same extent did not influence either the K m or k cat of the C215D mutant and hence resulted in no change in the second-order rate constant of C215D-mediated pNPP hydrolysis. A comparison of the crystal structures of apo-C215D and wild-type PTP-1B gives an insight into the effect of increasing solvent viscosity on catalytic efficiency. The closed conformation necessary for catalysis is observed in the apo-C215D structure, but not in the wildtype apo-PTP-1B structure. This suggests that there is a higher propensity for the loop to assume the catalytically competent conformation in the mutant derivative even in absence of substrate than in the wild-type enzyme. It seems possible that this structural change in the WPD loop conformation in the C215D mutant, which results in a preformed active site, influences the lack of sensitivity to the increasing solvent microviscosity. This induced fit mechanism may explain the higher K m value for the wild-type enzyme in the presence of the viscogen.
An important strategy in the identification of potential substrates of PTPases is the use of substrate-trapping mutants that are structurally similar to the wild-type enzyme, but display either a lower dissociation constant between the enzyme and the substrate or a slower substrate turnover. To date, four examples of substrate-trapping mutants have been used to characterize PTP-1B. In the first case, the active-site Cys is replaced with Ser (37)(38)(39). This mutant still retains the ability to bind substrates, but displays no measurable catalytic activity. However, differences in the thermodynamic parameters for ligand binding between the human form of the C215S mutant and wild-type PTP-1B have been noted (40) and could be partially explained by the observed altered conformation of the P-loop in the mutant enzyme (16). In the second type of substrate-trapping mutant, the general acid Asp (i.e. Asp 181 ) is replaced with Ala (41,42). Like the C215S mutant, this enzyme also binds substrate, but its catalytic activity is drastically reduced (42,43). A third type of substrate-trapping mutant is the Q262A derivative, which has been used to obtain a crystal structure of the phosphoryl-enzyme intermediate (44). Recently, a double mutant (D181A/Q262A) of PTP-1B has been generated. This substrate-trapping mutant exhibits higher affinity than both the independent D181A and C215S mutants for the epidermal growth factor receptor and displays 3000and 11,000-fold lower k cat values for pNPP and epidermal growth factor receptor, respectively (45). It is currently unknown, however, how this mutant compares structurally with wild-type PTP-1B.
In wild-type PTP-1B, Cys 215 functions as a strong nucleophile, forming a cysteinyl-phosphate intermediate in the reaction mechanism. We hypothesized that substitution of this residue with negatively charged Asp would decrease the catalytic activity of the enzyme. We observed that the turnover number of the C215D mutant was indeed significantly lower (Ͼ7000-fold) than that of the wild-type protein at pH 6, suggesting that this PTP-1B derivative may also serve as a highly efficient substrate-trapping derivative of the enzyme.
In this study, we have characterized some of the functional properties of the C215D derivative of PTP-1B. Although the results of this study suggest that Asp 215 is involved in catalysis, it is currently unknown whether the catalytic mechanism of this derivative is identical to that of the wild-type enzyme or is somewhat altered. Experiments are underway in our laboratory to address this question. Nonetheless, the C215D enzyme is a novel substrate-trapping mutant whose structure is nearly identical to, but displays significantly lower catalytic activity than, wild-type PTP-1B. Thus, C215D could be used to isolate and identify physiological substrates of PTP-1B. Introduction of a carboxylate residue in place of a thiolate has rendered the enzyme resistant to oxidation; and therefore, this mutant may also be used for screening of fermentation broth and natural products to identify inhibitors of PTP-1B. Preliminary studies in our laboratory indicate that the profiles of inhibitors screened with the mutant enzyme are equivalent to those obtained with the wild-type enzyme.