Reversible Inactivation of Protein-tyrosine Phosphatase 1B in A431 Cells Stimulated with Epidermal Growth Factor*

Stimulation of various cells with growth factors results in a transient increase in the intracellular concentration of H 2 O 2 that is required for growth factor-in- duced protein tyrosine phosphorylation. The effect of H 2 O 2 produced in response to epidermal growth factor (EGF) on the activity of protein-tyrosine phosphatase 1B (PTP1B) was investigated in A431 human epidermoid carcinoma cells. H 2 O 2 inactivated recombinant PTP1B in vitro by oxidizing its catalytic site cysteine, most likely to sulfenic acid. The oxidized enzyme was reacti-vated more effectively by thioredoxin than by glutaredoxin or glutathione at their physiological concentrations. Oxidation by H 2 O 2 prevented modification of the catalytic cysteine of PTP1B by iodoacetic acid, suggesting that it should be possible to monitor the oxidation state of PTP1B in cells by measuring the incorporation of radioactivity into the enzyme after lysis of the cells in the presence of radiolabeled iodoacetic acid. The amount of such radioactivity associated with PTP1B immunoprecipitated from A431 cells that had been stimulated with EGF for 10 min was 27% less than that associated with PTP1B from unstimulated cells. The amount of iodoacetic acid-derived radioactivity associated with PTP1B reached a minimum 10 min after stimulation of cells with EGF and returned to base line values by 40 min, suggesting that the oxidation of PTP1B is reversible in cells. polymerase chain polymerase I pET-14b PTP1B thrombin-sensitive Escherichia coli strain BL21 (DE3) standard procedures. histidine-tagged PTP1B fu- sion purified Escherichia coli extract with the use of an immobilized nickel resin (Novagen). The of the PTP1B prep- aration as assessed by SDS-polyacrylamide gel electrophoresis (PAGE) PTP1B cleave tag and purified histidine-tagged and thrombin-treated incu- bated DTT and dialyzed in an anaerobic chamber at °C against a deoxygenated solution containing 20 m M Mes-NaOH (pH 6.5) and 0.1 m M EDTA. Portions of the dialyzed enzyme were stored in an anaerobic chamber at 4 °C. Unless otherwise specified, histidine-tagged PTP1B was used throughout this and formance liquid chromatography (HPLC) DEAE-5PW ion exchange col-umns. purified from described Determination of Protein Concentration— The concentrations of recombinant PTP1B and Trx were determined spectrophotometrically, and the A 280 values of 0.1% solutions were 1.231 and 0.738, respectively. The concentrations of other proteins were determined with the BCA protein assay reagent (Pierce), with bovine serum albumin as a standard. Assay of PTP1B Activity— Two different methods were used to assay PTP1B activity. The activity of recombinant PTP1B was measured spectrophotometrically with p -nitrophenyl phosphate (pNPP) as a sub- strate in a reaction mixture containing 40 m M Bis-Tris-HCl (pH 7.0), 2 m M EDTA, 50 m M NaCl, and 10 m M pNPP (13). The initial velocity of p -nitrophenol formation was measured by monitoring the change in A 405 . An assay of PTP1B activity in immune complexes was performed with 33 P-phosphorylated Raytide as a substrate (23). Assay mixtures (50 m l) containing the immune complex, 20 m M Tris-HCl (pH 7.5), bovine serum albumin (0.1 mg/ml), 1 m M EDTA, 10 m M DTT, and 90 n M 33 P-phosphorylated were incubated at 30 °C for 30 min, after which the reaction was terminated by the addition of 10 m l of glacial acetic acid, and the radioactivity associated with the was meas- ured as (24). anaerobic of centrifuged for super- Protein-containing pooled, pooled precleared incubating for 4 at 4 with of anti-mouse immunoglobulin-coated immunobeads (25% slurry) and centri-fuging

Ligation of a variety of cell surface receptors, including those for growth factors and cytokines, induces a transient increase in the intracellular concentration of H 2 O 2 in mammalian cells (1)(2)(3). Inhibition of this effect blocks receptor-mediated signal transduction. For example, inhibition of the accumulation of H 2 O 2 by introducing catalase into NIH 3T3 or A431 cells pre-vented the induction of tyrosine phosphorylation by plateletderived growth factor or epidermal growth factor (EGF) 1 (2,3). Direct exposure of cells to H 2 O 2 also increases protein tyrosine phosphorylation and activates signal transduction pathways (2)(3)(4)(5). Because the extent of protein tyrosine phosphorylation in a cell reflects an equilibrium between the actions of protein-tyrosine kinases (PTKs) and protein-tyrosine phosphatases (PTPs), either stimulation of PTKs or inhibition of PTPs would be expected to shift the equilibrium toward phosphorylation. PTP activity in crude cell extracts can be inactivated by various oxidants, including H 2 O 2 , and this inactivation can be reversed by incubation with thiol compounds such as dithiothreitol (DTT) and GSH (6,7). These observations suggest that PTPs might undergo H 2 O 2 -dependent inactivation in cells, resulting in a shift in the equilibrium with PTKs toward phosphorylation.
PTPs constitute a diverse family of enzymes that can be divided into several subgroups, including receptor PTPs and nonreceptor PTPs (8 -11). All PTPs contain an essential cysteine residue in the signature active site motif, HCXXGXXR(S/ T). The PTP active site cysteines exhibit low pK a values (5.4 for mammalian PTP1 (12), 5.6 for human dual specific PTP (13), and 4.7 for Yersinia PTP (14)) and are readily ionized at neutral pH, whereas the pK a of a typical cysteine residue is 8.5. The ionized essential sulfhydryl group (thiolate anion) contributes to the formation of a thiol-phosphate intermediate in the catalytic mechanism of PTPs (15). In addition, the essential cysteine is the target of specific modification by various sulfhydryl-alkylating reagents (12-14, 16, 17).
We now demonstrate that H 2 O 2 , either added extracellularly or generated intracellularly in response to EGF, can cause reversible inactivation of PTPs in cells, and we identify the most plausible electron donor responsible for the reactivation of such inactivated PTPs. PTP1B, the widely expressed cytosolic enzyme originally purified from human placenta (18), was chosen as the target enzyme and was studied in A431 human epidermoid carcinoma cells.

EXPERIMENTAL PROCEDURES
Materials-Dulbecco's modified Eagle's medium, fetal bovine serum, penicillin, and streptomycin were from Life Technologies, Inc. Rabbit polyclonal antibodies to PTP1B were kindly provided by B. G. Neel (Harvard University). A monoclonal antibody to PTP1B was obtained from Oncogene Science. Horseradish peroxidase-conjugated antibodies to mouse or to rabbit immunoglobulin G were from Amersham Pharmacia Biotech. Yeast glutathione reductase (GR) and bovine catalase were from Boehringer Mannheim. The synthetic peptide Raytide and the kinase p43 v-abl were from Oncogene Science. [␥-33 P]ATP was from Amersham Pharmacia Biotech, and 14 C-or 3 H-labeled iodoacetic acid was from NEN Life Science Products.
Recombinant PTP1B-Complementary DNA corresponding to the 37-kDa form (NH 2 -terminal 321 residues) of PTP1B was obtained by the polymerase chain reaction, placed downstream of the phage T7 RNA polymerase promoter at the NdeI site of pET-14b (Novagen) (thus providing a histidine tag attached to the NH 2 terminus of PTP1B by a thrombin-sensitive sequence), and expressed in Escherichia coli strain BL21 (DE3) by standard procedures. The histidine-tagged PTP1B fusion protein was purified from Escherichia coli extract with the use of an immobilized nickel resin (Novagen). The purity of the PTP1B preparation as assessed by SDS-polyacrylamide gel electrophoresis (PAGE) was Ͼ95%. A portion of the purified PTP1B was treated with thrombin to cleave the histidine tag and was repurified on a Mono S column. Both the purified histidine-tagged and thrombin-treated enzymes were incubated in the presence of 10 mM DTT for 1 h and then dialyzed in an anaerobic chamber at 4°C against a deoxygenated solution containing 20 mM Mes-NaOH (pH 6.5) and 0.1 mM EDTA. Portions of the dialyzed enzyme were stored in an anaerobic chamber at 4°C. Unless otherwise specified, histidine-tagged PTP1B was used throughout this study.
Preparation of Thioredoxin (Trx), Glutaredoxin (Grx), and Trx Reductase (TR)-Rat Trx cDNA (19) was obtained by the polymerase chain reaction, cloned into the pET-17b expression vector, and expressed in E. coli by standard procedures. Recombinant Trx was purified to homogeneity from the cytosolic fraction of E. coli by heat treatment at 65°C followed by sequential chromatography on Sephacryl S-100 HR gel filtration (Amersham Pharmacia Biotech) and high performance liquid chromatography (HPLC) DEAE-5PW ion exchange columns. TR and Grx were purified from rat liver as described (20 -22).
Determination of Protein Concentration-The concentrations of recombinant PTP1B and Trx were determined spectrophotometrically, and the A 280 values of 0.1% solutions were 1.231 and 0.738, respectively. The concentrations of other proteins were determined with the BCA protein assay reagent (Pierce), with bovine serum albumin as a standard.
Assay of PTP1B Activity-Two different methods were used to assay PTP1B activity. The activity of recombinant PTP1B was measured spectrophotometrically with p-nitrophenyl phosphate (pNPP) as a substrate in a reaction mixture containing 40 mM Bis-Tris-HCl (pH 7.0), 2 mM EDTA, 50 mM NaCl, and 10 mM pNPP (13). The initial velocity of p-nitrophenol formation was measured by monitoring the change in A 405 . An assay of PTP1B activity in immune complexes was performed with 33 P-phosphorylated Raytide as a substrate (23). Assay mixtures (50 l) containing the immune complex, 20 mM Tris-HCl (pH 7.5), bovine serum albumin (0.1 mg/ml), 1 mM EDTA, 10 mM DTT, and 90 nM 33 P-phosphorylated Raytide were incubated at 30°C for 30 min, after which the reaction was terminated by the addition of 10 l of glacial acetic acid, and the radioactivity associated with the peptide was measured as described (24).
Determination of Free SH Groups-PTP1B (100 g) that had been treated with DTT and then dialyzed against a DTT-free buffer under anaerobic conditions was incubated at 25°C with 100 M H 2 O 2 in a total volume of 100 l containing 40 mM Bis-Tris-HCl (pH 7.0), 150 mM NaCl, and 1 mM EDTA. After 20 min, the oxidation reaction was stopped by adding 1 g of catalase. The resulting oxidized and control unoxidized enzymes (100 g in 100 l) were separately mixed with 300 l of a solution containing 266 M 5,5Ј-dithiobis-(2-nitrobenzoic acid) (DTNB), 6 M guanidine hydrochloride, and 50 mM Tris-HCl (pH 8.0). The concentration of thionitrobenzoic acid released was determined spectrophotometrically with a molar extinction coefficient of 13,700 at 412 nm (25).
Identification of the H 2 O 2 -sensitive Residue in PTP1B-PTP1B (64 g) that had been incubated with 100 M H 2 O 2 for 10 min and then treated with catalase as described above was incubated with 2 mM [ 3 H]iodoacetic acid in a total volume of 100 l of 40 mM Bis-Tris-HCl (pH 6.5) containing 0.1 mM EDTA. As a control, unoxidized PTP1B (64 g) was likewise treated with [ 3 H]iodoacetic acid. After 10 min at room temperature, the reaction was stopped by adding 100 mM 2-mercaptoethanol and the reaction mixtures were subjected to gel filtration chromatography on a Sephadex G-25 column to remove unreacted iodoacetic acid. The PTP1B-containing fractions were pooled and incubated overnight at room temperature with 2.5 g of endoproteinase Lys-C in a total volume of 400 l. The resulting digestion products were analyzed by HPLC on a C 18 column with a linear gradient (0 -60%, v/v) of acetonitrile in 0.1% trifluoroacetic acid at a flow rate of 1 ml/min over 60 min. Fractions corresponding to each peptide peak were collected manually, and a portion (10%) of each fraction was analyzed for 3 H radioactivity.
Iodoacetic Acid Labeling and Immunoprecipitation of PTP1B-A431 cells, maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, penicillin (100 units/ml), and streptomycin (100 units/ml), were allowed to reach 80 -90% confluence in 150-mm dishes. The cells were then deprived of serum for 16 h and subsequently stimulated with EGF (200 ng/ml) or H 2 O 2 (1 or 3 mM). The cells were rinsed and then exposed in an anaerobic chamber to 1 ml of O 2 -free lysis buffer (50 mM Bis-Tris-HCl (pH 6.5), 0.5% Triton X-100, 0.5% deoxycholate, 0.1% SDS, 150 mM NaCl, 1 mM EDTA, leupeptin (0.5 g/ml), aprotinin (0.5 g/ml), and 0.1 mM 4-(2-aminoethyl)-benzenesulfonyl fluoride hydrochloride) containing 2 mM [ 14 C]iodoacetic acid and briefly sonicated. As a control, serum-deprived cells that were not stimulated with EGF or H 2 O 2 were likewise lysed and labeled. After 30 min at room temperature in the dark, the labeling reaction was stopped by adding 0.2 ml of 200 mM cold iodoacetic acid in 0.8 M Tris-HCl (pH 7.5) and followed by the addition of 0.1 ml of 1 M DTT. The reaction mixtures were then centrifuged at 10,000 ϫ g for 20 min. The resulting supernatants were subjected to a G-25 gel filtration chromatography to remove excess radioactivity. Protein-containing fractions were pooled, and protein concentrations were measured. The pooled samples were precleared by incubating for 4 h at 4°C with 40 l of goat anti-mouse immunoglobulin-coated immunobeads (25% slurry) (Sigma) and centrifuging at 3000 ϫ g for 30 s. Afterward, 40 l of goat anti-mouse immunoglobulin immunobeads (25% slurry) that had been absorbed with an excess of monoclonal antibody to PTP1B were added to the resulting supernatants, and incubation was continued for an additional 4 h. The beads were separated; extensively washed once in ice-cold O 2 -free lysis buffer, twice in O 2 -free 50 mM Tris-HCl buffer(pH 7.5), and twice in O 2 -free phosphate-buffered saline (pH 7.4); and subjected either to SDS-PAGE in order to measure the amount of radioactivity incorporated into PTP1B or to assay the amount of PTP activity regenerated by treatment with DTT.

RESULTS
PTP1B was originally isolated from human placenta as a soluble 37-kDa protein (18). It was subsequently shown that the full-length (50-kDa) PTP1B protein contains 435 amino acids and that the 37-kDa protein corresponds to the NH 2terminal 321 residues (26). The COOH-terminal 114 residues of the full-length protein contain a sequence responsible for localization of PTP1B in the endoplasmic reticulum (23). The 37-kDa PTP1B contains six cysteine residues, which do not appear to form disulfide bonds on the basis of the crystal structure of the protein (27). Both Cys 121 and Cys 215 of PTP1B are conserved among all members of the PTP family (28). Cys 215 , which has a pK a value of 5.4, is the essential cysteine residue located at the active site. However, mutation of Cys 121 in PTP1, the rat homolog of PTP1B, markedly reduces PTP activity (15). The active site cysteines of various PTPs, including Cys 215 of PTP1B, are specifically targeted by the sulfhydryl-modifying reagent iodoacetic acid (12-14, 16, 17). The same active site cysteine residues have also been implicated as the site of oxidation by various oxidants (2,3,6,7,29), in which case one should be able to monitor the extent of H 2 O 2 -induced inactivation of PTPs in cells by measuring the amount of radioactivity incorporated into the enzyme after cell lysis in the presence of radiolabeled iodoacetic acid. With this aim, we expressed the 37-kDa PTP1B in E. coli, purified the recombinant protein, and subjected it to modification by H 2 O 2 and iodoacetic acid.
Identification of the H 2 O 2 -sensitive Cysteine in PTP1B-Incubation with H 2 O 2 resulted in the inactivation of the purified recombinant 37-kDa PTP1B in a manner dependent on time and H 2 O 2 concentration (Fig. 1, A and B). When the enzyme samples from the experiment shown in Fig. 1B (which had been inactivated to various extents by incubation with different concentrations of H 2 O 2 ) were subjected to labeling with [ 14 C] iodoacetic acid at pH 6.5, the extent of labeling decreased in proportion to the extent of inactivation (Fig. 1C), indicating that the site of oxidation by H 2 O 2 is the same as the site of labeling by iodoacetic acid. The data shown in Fig. 1 were obtained with histidine-tagged enzyme, but similar results were obtained with PTP1B lacking the histidine tag.
We determined the number of sulfhydryl groups in PTP1B To identify the H 2 O 2 -sensitive cysteine, we exposed PTP1B oxidized with H 2 O 2 (100 M for 10 min to achieve 80% inactivation) as well as unoxidized control enzyme to [ 3 H]iodoacetic acid. The labeled enzymes were then cleaved with endoproteinase Lys-C, the resulting peptides were separated by HPLC on a C 18 column to yield 25 identifiable peaks, and the amount of 3 H radioactivity associated with each peak was measured (data not shown). The peptides from oxidized and unoxidized PTP1B enzymes produced virtually identical elution profiles. For unoxidized PTP1B, ϳ80% of total radioactivity was associated with a major peak that eluted at 36.4 min, whereas the remaining 20% of radioactivity was distributed among several other peptides. However, for the H 2 O 2 -treated enzyme, although the peak that eluted at 36.4 min contained the most radioactivity, this amount was only 23% of that associated with the corresponding peak for the unoxidized enzyme. Edman degradation of the major radioactive peptide derived from the unoxidized enzyme yielded a complete sequence of VRESGSLSPEHGPV-VVHXSAGIGRSGTFXLADTXLLLMDK, which matches exactly the sequence of amino acids 198 -237 of PTP1B with the exception that the residue corresponding to position 215 was identified as carboxymethylated cysteine and the residues corresponding to positions 226 and 231 were unknown. The PTP1B sequence between residues 198 and 237 contains three cysteines at positions 215, 226, and 231. When the Edman degradation products after each cycle were collected and measured for radioactivity, most of the 3 H was detected at the 18th cycle (Cys 215 ), with smaller amounts also present for the next several cycles. No radioactivity was detected at the cycles corresponding to Cys 226 and Cys 231 . Together, these results suggest that the active site Cys 215 is the major site of alkylation by iodoacetic acid as well as the site of oxidation by H 2 O 2 . Thus, it should be possible to monitor the extent of PTP1B inactivation by H 2 O 2 by measuring the amount of radioactive iodoacetic acid incorporated into the enzyme at pH 6.5.
Oxidative Inactivation of PTP1B in A431 Cells-A431 cells were incubated for various times with EGF and then lysed in an anaerobic chamber by exposure to a pH 6.5 buffer containing Triton X-100 and [ 14 C]iodoacetic acid. PTP1B was then immunoprecipitated from the cell lysate with a highly specific monoclonal antibody, and the radioactivity associated with the precipitated 50-kDa PTP1B protein was measured after SDS-PAGE. The extent of incorporation of [ 14 C]iodoacetic acid into PTP1B decreased with time of incubation of the cells with EGF, reaching a minimum at 10 min, and returned to the basal value by 40 min (Fig. 2).
Data from six independent experiments revealed that the extent of the decrease in [ 14 C]iodoacetic acid incorporation into the PTP1B immunoprecipitate measured after stimulation of cells for 10 min with EGF was 26.8 Ϯ 6.4% (Fig. 3A). Stimulation of A431 cells for 10 min with H 2 O 2 at concentrations of 1 or 3 mM reduced the extent of radioactivity incorporated into the PTP1B immunoprecipitate by 33.9 Ϯ 7.4 and 42.3 Ϯ 6.1%, respectively (Fig. 3A). These results suggest that PTP1B is oxidized at the active site cysteine in response to stimulation of cells with EGF or H 2 O 2 .
Additional evidence for the oxidation of the active site cysteine of PTP1B was provided by reactivation experiments. Cys 215 of PTP1B is likely oxidized to sulfenic acid (Cys-SOH) by After 10 min, the reaction was stopped by the addition of 1 g of catalase (65 units/g). One-half of the reaction mixture was assayed for PTP activity with pNPP as a substrate. Data are from a representative experiment. C, to the remaining half of the reaction mixtures from B, [ 14 C]iodoacetic acid was added to a final concentration of 1 mM. After 10 min at room temperature, the reaction was stopped by the addition of 2-mercaptoethanol to a final concentration of 100 mM, and the samples were subjected to SDS-PAGE on a 10% gel. The separated proteins were transferred to a nitrocellulose membrane, which was then dried and subjected to autoradiography (top). Equal application of protein among gel lanes was confirmed by immunoblot analysis of the same nitrocellulose membrane with antibodies to PTP1B (bottom). The positions of PTP1B are indicated in both top and bottom. H 2 O 2 (see "Discussion"). Because sulfenic acid is readily reduced by a thiol, incubation of the immunoprecipitates from the experiments shown in Fig. 3A with DTT would be expected to reactivate the PTP1B molecules oxidized by H 2 O 2 but not those modified by iodoacetic acid. To test this prediction, we subjected the iodoacetic acid-labeled immunoprecipitates from the EGF-treated and H 2 O 2 -treated cells to DTT treatment and subsequent assay for PTP activity. As expected, the PTP activity of immunoprecipitates from EGF-treated (0.56 Ϯ 0.09 pmol/30 min) or H 2 O 2 -treated (0.60 Ϯ 0.13 and 0.69 Ϯ 0.12 pmol/30 min for 1 and 3 mM H 2 O 2 , respectively) cells was greater than that for immunoprecipitates from unstimulated cells (0.29 Ϯ 0.08 pmol per 30 min). These results thus support the notion that PTP1B is oxidized by H 2 O 2 and that the modified enzyme can be reduced to its original state by DTT.
To estimate the total activity of unlabeled PTP1B, immunoprecipitates were prepared from A431 cell lysates and subjected to DTT treatment and subsequent assay of PTP activity as described in the legend of Fig. 3, with the exception that iodoacetic acid was omitted from the lysis buffer. The total PTP activity thus assessed was 4.5 Ϯ 0.09 pmol/30 min and was not affected by stimulation of cells with EGF or H 2 O 2 , suggesting that the total amount of PTP1B was not affected by EGF or H 2 O 2 . The PTP activities of DTT-treated immunoprecipitates derived from iodoacetic acid-labeled lysates were then expressed as a percentage of the total PTP activity (Fig. 3B). The observation that such immunoprecipitates from unstimulated cells exhibited substantial PTP activity (6.4% of total activity) suggests that PTP1B was not fully carboxymethylated by iodo-acetic acid or that a small fraction of PTP1B in unstimulated cells exists in an oxidized state. The percentages of the total activity regenerated by DTT for the stimulated cells (12.4% for EGF-treated cells; 13.3 and 15.3% for cells treated with 1 or 3 mM H 2 O 2 ) (Fig. 3B) were substantially lower than the corresponding percentage decreases in iodoacetic acid labeling (26.8, 33.9, and 42.3%, respectively) (Fig. 3A). The reason for this discrepancy is not clear. It is possible that exposure of immunoprecipitates to air during handling might have resulted in further oxidation of sulfenic acid to sulfinic acid (Cys-SOOH), which cannot be reduced by DTT.

FIG. 2. Time course of the effect of EGF stimulation on [ 14 C]
iodoacetic acid labeling of PTP1B from A431 cells. A431 cells were stimulated with EGF (200 ng/ml) for the indicated times and then lysed in a buffer containing 2 mM [ 14 C]iodoacetic acid. PTP1B was precipitated from the cell lysates with a specific monoclonal antibody and subjected to SDS-PAGE on a 10% gel. A, the dried gel was exposed to x-ray film (Eastman Kodak Co.) for 4 days to yield an autoradiogram (top), after which the separated proteins were probed by immunoblot analysis with rabbit antibodies to PTP1B and immune complexes were visualized with horseradish peroxidase-conjugated antibodies to rabbit immunoglobulin G and ECL reagents (Amersham Pharmacia Biotech) (bottom). B, the intensity of the PTP1B bands on autoradiograms (A) was quantitated with a PhosphorImager (Molecular Dynamics) equipped with a scanner (Umax). Data are expressed as a percentage of the value for immunoprecipitates derived from unstimulated A431 cells and are means Ϯ S.E. from three independent experiments.

FIG. 3. Effect of stimulation of A431 cells with EGF or H 2 O 2 on
the labeling of PTP1B with iodoacetic acid and on its catalytic activity. A431 cells were incubated in the absence (control) or presence of EGF (200 ng/ml) or H 2 O 2 (1 or 3 mM) for 10 min and then lysed in a buffer containing 2 mM [ 14 C]iodoacetic acid. PTP1B was immunoprecipitated from the lysates as described in Fig. 2, and the precipitates were divided into two equal portions. A, the amount of radioactivity associated with one-half of the immunoprecipitates was measured from the autoradiograms as described in Fig. 2. Data are means Ϯ S.E. from six independent experiments. B, the other half of the immunoprecipitates was incubated for 5 min at room temperature in a solution containing 20 mM Tris-HCl (pH 7.5), 1 mM EDTA, and 10 mM DTT and then assayed for PTP activity for 30 min with 33 P-phosphorylated Raytide as a substrate. Data are expressed as a percentage of the value (4.5 pmol of 33 P-Raytide hydrolyzed during 30 min) for immunoprecipitates from A431 cell extracts that had not been incubated with iodoacetic acid and are means Ϯ S.E. from three independent experiments.
Reactivation of PTP1B-Recombinant PTP1B that had been inactivated by exposure to H 2 O 2 was incubated in the presence of various electron donors and reactivation was monitored (Fig.  4). In addition to DTT and GSH, the Trx system (consisting of Trx, TR, and NADPH) and the Grx system (consisting of Grx, GSH, GR, and NADPH) were used as electron donors. In the Trx system, oxidized Trx is reduced by TR with the use of electrons supplied by NADPH. In the Grx system, oxidized Grx is reduced by GSH, which in turn receives electrons from NADPH through the action of GR. The most rapid reactivation was achieved by DTT (4 mM), whereas GSH (4 mM) was the least efficient electron donor. The Trx system was slightly less efficient than DTT, and the Grx system was substantially less effective than the Trx system. The functional efficacy of the Grx and GR preparations was demonstrated by measuring GSHdisulfide transhydrogenase (30) and glutathione peroxidase (31) activities, respectively. Each component of the Trx and Grx systems was used at a saturating concentration; i.e. increasing their concentrations did not increase the rate of reactivation (data not shown).
The reduction of oxidized PTP1B by the Trx system and the Grx system was also measured spectrophotometrically by monitoring the decrease in A 340 attributable to the oxidation of NADPH. PTP1B reduction by GSH was also coupled to NADPH oxidation by including GR and NADPH in the reaction mixture. In agreement with the data shown in Fig. 4, the rank order for the rate of NADPH oxidation was Trx system Ͼ Grx system Ͼ GSH system (Fig. 5). With the PTB1B concentration of 7 M used in Fig. 5, the maximal change in A 340 is expected to be 0.042 absorbance units.
Reactivation of oxidized PTP1B by the Trx system required all three components (Trx, TR, and NADPH); reactivation in the absence of one of the three components was negligible (data not shown). The half-maximal concentration of Trx required for reactivation was 0.15 M.

DISCUSSION
Given the recent observation that H 2 O 2 is required for the growth factor-induced tyrosine phosphorylation of cellular proteins, we investigated whether H 2 O 2 produced in response to EGF is capable of inactivating PTP1B in A431 cells.
First, with the use of the recombinant 37-kDa form of PTP1B, we demonstrated that the essential residue Cys 215 is the site of oxidation by H 2 O 2 . The oxidized products of cysteine include sulfenic acid, disulfide, sulfinic acid, and sulfonic acid (Cys-SO 3 H). The disulfide intermediate can be excluded as the H 2 O 2 -modified form of PTP1B on the basis of our observation that only one out of six DTNB-sensitive residues was lost after H 2 O 2 oxidation, and the sulfinic and sulfonic acid intermediates can be excluded on the basis of the observation that the oxidized PTP1B can be reduced back to its original state by DTT. Nevertheless, PTP1B was shown to form sulfinic and sulfonic acid intermediates when oxidized in the presence of osteoporosis drug alendronate (32) and pervanadate (33), respectively. Cysteine sulfenic acid is highly unstable and readily undergoes condensation with a thiol. However, the sulfenic acid intermediate of PTP1B is probably stabilized by the fact that, according to the x-ray structure of the 37-kDa form of PTP1B (27), no cysteine residues are located near Cys 215 . Furthermore, the sulfenate anion (Cys-SO Ϫ ) is also probably stabilized by a salt bridge to Arg 221 , which was shown to stabilize the thiolate anion of Cys 215 and consequently to reduce its pK a .
Second, we measured the amount of radiolabeled iodoacetic acid incorporated into PTP1B as a means of monitoring changes in the oxidation state of the protein in A431 cells stimulated with EGF. This approach was based on our observations that iodoacetic acid reacts almost exclusively with Cys 215 -SH of PTP1B at pH 6.5 and that the oxidation of this cysteine residue by H 2 O 2 prevents its reaction with iodoacetic acid. Exposure of A431 cells to EGF resulted in a decrease in the extent of iodoacetic acid labeling of PTP1B, with the maximal (27%) decrease apparent 10 min after stimulation and the labeling returning to base-line values by 40 min. This reduced labeling is probably due to the oxidation of PTP1B by H 2 O 2 ,   (7).
Third, we compared the abilities of Trx, Grx, and GSH to reactivate oxidized recombinant PTP1B in vitro; Trx at 3.8 M was markedly more efficient than 3.8 M Grx or 4 mM GSH. The cellular concentrations of Trx and GSH are approximately 2-14 M and 1-10 mM, respectively (34,35). Thus, Trx is predicted to be a major electron donor for PTP1B reduction in cells. Trx was previously shown to be the preferred electron donor for the reduction of glyceraldehyde-3-phosphate dehydrogenase containing an active site cysteine sulfenic acid, whereas Grx was preferred for the reduction of the same enzyme containing a disulfide (36 -38).
Although we have demonstrated the sensitivity of PTP1B to H 2 O 2 formed in response to treatment of cells with EGF, other PTP enzymes are also probably susceptible to such inactivation. However, it is possible that the concentration of H 2 O 2 is sufficiently high to inactivate PTPs only in limited regions of the cell in which the H 2 O 2 -producing components are recruited and that H 2 O 2 molecules that diffuse away from such regions are readily eliminated by various peroxidases. Such a localized inactivation of PTP1B is one possible explanation for our observation that only 27% of this enzyme was oxidized in EGFactivated A431 cells. Neither the mechanism of H 2 O 2 generation, the site of generation, nor the concentration of H 2 O 2 generated in response to growth factors and cytokines is known.
On the basis of the previous observation that growth factor-induced protein tyrosine phosphorylation requires H 2 O 2 production and our current observation that growth factor-induced generation of H 2 O 2 is sufficient to cause inactivation of PTP1B, we propose that the activation of a receptor PTK by interaction with a growth factor may not be sufficient to increase the steady state level of protein tyrosine phosphorylation in a cell; rather, concurrent inhibition of PTPs by H 2 O 2 may also be required for this effect. The extent of tyrosine phosphorylation of receptor PTKs and their substrates would then return to basal values after degradation of H 2 O 2 and the subsequent reactivation of PTPs by electron donors. Our in vitro data suggest that Trx might be a physiological electron donor for PTP1B. It remains to be determined whether other PTPs also form a sulfenic acid intermediate on oxidation with H 2 O 2 and whether Trx reduces such oxidized intermediates. The low molecular weight PTP, which shows no apparent sequence similarity to other PTPs but which shares several common features in active site architecture (9,39), forms a disulfide on oxidation with nitric oxide (40). A scheme depicting the proposed roles of H 2 O 2 and Trx in growth factor-induced protein tyrosine phosphorylation is shown in Fig. 6. This scheme is consistent with the following observations: 1) production of H 2 O 2 via NADPH oxidase results in inhibition of the PTP activity of CD45 in neutrophils, 2) blocking the H 2 O 2 production by treatment with N-acetylcysteine or diphenylene iodonium, an inhibitor of NADPH oxidase, restores its PTP activity (41,42), and 3) removal of intracellular oxidants by pyrrolidine dithiocarbamate diminishes protein tyrosine phosphorylation (43). The scheme is also consistent with the suggestion that the ligand-independent basal activity of receptor PTKs might be sufficient to increase the extent of protein tyrosine phosphorylation in cells treated with thiolalkylating agents, such as iodoacetic acid and iodoacetamide, or oxidants, such as ultraviolet light, that cause the inactivation of PTPs (29).
Previously proposed mechanisms for the regulation of nonreceptor PTP activity include phosphorylation at serine and tyrosine residues (23, 44 -48), anchoring via SRC homology 2 domains (49,50), and proteolysis (51,52). In contrast to these positive regulation mechanisms of PTP activity, oxidation by H 2 O 2 provides a means for negative regulation of PTP activity. Negative modulation mediated by ligand-induced dimerization has also been proposed for the receptor PTP CD45 (53).