Two Vicinal Cysteines Confer a Peculiar Redox Regulation to Low Molecular Weight Protein Tyrosine Phosphatase in Response to Platelet-derived Growth Factor Receptor Stimulation*

Low molecular weight protein tyrosine phosphatase (LMW-PTP) is an enzyme involved in platelet-derived growth factor (PDGF)-induced mitogenesis and cytoskeleton rearrangement because it is able to bind and dephosphorylate the activated receptor. LMW-PTP pre-sents two cysteines in positions 12 and 17, both belong-ing to the catalytic pocket; this is a unique feature of LMW-PTP among all protein tyrosine phosphatases. Our previous results demonstrated that in vitro LMW-PTP is oxidized by either H 2 O 2 or nitric oxide with the forma- tion of a disulfide bond between Cys-12 and Cys-17. This oxidation leads to reversible enzyme inactivation because treatment with reductants permits catalytic activity rescue. In the present study we investigated the in vivo inactivation of LMW-PTP by either extracellularly or intracellularly generated H 2 O 2 , evaluating its action directly on its natural substrate, PDGF receptor. LMW-PTP is oxidized and inactivated by exogenous oxidative stress and recovers its activity after oxidant removal. LMW-PTP is oxidized also during PDGF signaling, very likely upon PDGF-induced H 2 O 2 production, and recov- ers its activity within 40 min. Our results strongly suggest

Protein tyrosine phosphorylation plays a key role in the regulation of many cellular processes in eukaryotes such as cellular metabolism, proliferation, differentiation, and onco-genic transformation (1). Accumulating evidence indicates that the contribution of phosphotyrosine protein phosphatases (PTPs) 1 to the control of the cell phosphorylation state is as relevant as that of phosphotyrosine protein kinases. The PTP superfamily is composed of over 70 enzymes that, despite very limited sequence similarity, share a common CX 5 R active site motif and an identical catalytic mechanism. On the basis of their function, structure, and sequence, PTPs can be classified in four main families: 1) tyrosine-specific phosphatases, 2) VH1-like dual specificity PTPs, 3) the cdc25 phosphatases, and 4) the low molecular weight phosphatase (2).
The low molecular weight protein tyrosine phosphatase (LMW-PTP) is an 18-kDa enzyme that is expressed in many mammalian tissues (3). Our previous studies on the molecular biology of LMW-PTP in NIH3T3 cells demonstrated a well defined role of this enzyme in platelet-derived growth factor (PDGF)-induced mitogenesis. The most relevant phenotypic effect of LMW-PTP overexpression is a strong reduction of the cell growth rate in response to PDGF stimulation. We have previously shown that activated PDGF-R is a LMW-PTP substrate (4) and that LMW-PTP is involved in the control of specific pathways triggered by PDGF-R activation. In particular, LMW-PTP is able to modulate both myc expression, interfering with the Src pathway, and fos expression through a mitogen-activated protein kinase-independent pathway mediated by the signal transducers and activators of transcription (STAT) proteins (5). More recently, we have found that in NIH3T3 cells LMW-PTP is constitutively localized in both cytoplasmic and cytoskeleton-associated fractions. These two LMW-PTP pools are differentially regulated because only the cytoskeleton-associated LMW-PTP fraction is specifically phosphorylated by c-Src after PDGF stimulation (6). As a consequence of its phosphorylation, LMW-PTP shows an average 20-fold increase in its in vitro catalytic activity (7,8,9). Cytoskeleton-associated LMW-PTP influences cell adhesion, spreading, and migration, controlling the phosphorylation level of p190Rho-GAP, a protein that is able to regulate Rho activity and, consequently, cytoskeleton rearrangement in response to PDGF stimulation. Hence, LMW-PTP is able to perform multiple roles in PDGF-induced mitogenesis: while cytosolic LMW-PTP binds and dephosphorylates PDGF-R (4), thus modulating part of its signaling cascade, cytoskeleton-associated LMW-PTP acts on phosphorylated p190Rho-GAP, consequently play-ing a role in PDGF-mediated cytoskeleton rearrangement (10).
Caselli et al. (11,12) have demonstrated that in vitro LMW-PTP is oxidized and inactivated by NO or H 2 O 2 and that P i is able to protect the enzyme from oxidation. H 2 O 2 causes the oxidation of Cys-12 and Cys-17, the two vicinal cysteines in the catalytic pocket, as demonstrated by mass spectroscopy and radiolabeled tryptic fingerprint analysis. Furthermore, Caselli et al. (11,12) have demonstrated that H 2 O 2 oxidizes Cys-12 and Cys-17 to form a disulfide bond. Treatment of LMW-PTP with dithiothreitol restores its enzymatic activity.
Recent findings indicate that treatment with growth factors or H 2 O 2 induces an elevation of tyrosine-phosphorylated proteins. This elevation can be achieved by the activation of protein tyrosine kinases and/or inactivation of PTPs. It has been shown that reactive oxygen species (ROS), such as O 2 . or H 2 O 2 , are transiently generated intracellularly when cells are stimulated with cytokines or growth factors (13). In addition, an exogenous oxidative stress could be produced in various physiological conditions, such as the lymphocytic and macrophagic oxidative burst, or in pathological conditions, such as reperfusion, etc. (13). Although there is not convincing evidence that protein tyrosine kinases are activated by ROS, PTPs have been shown to be regulated by a redox mechanism (14 -16 (17,18). We now demonstrate that H 2 O 2 , either added extracellularly or generated intracellularly in response to PDGF stimulation, can cause inhibition of the activity of LMW-PTP on both its known phosphorylated substrates PDGF-R and p190Rho-GAP. In addition, we show that glutathione is most probably the electron donor responsible for the reactivation of LMW-PTP after the removal of the oxidative stress. We demonstrate that the two vicinal cysteines in the catalytic pocket of the phosphatase confer a peculiar and reversible redox regulation to LMW-PTP uncommon among PTP family members.

EXPERIMENTAL PROCEDURES
Materials-Unless specified all reagents were obtained from Sigma. NIH3T3 and C2C12 cells were purchased from ATCC, human recombinant PDGF-BB was from Peprotech, the Enhanced Chemi-Luminescence kit was from Amersham Pharmacia Biotech, all antibodies were from Santa Cruz Biochemicals, and BCA protein assay reagent was from Pierce. 5Ј-F-IAA and affinity-purified rabbit antibodies to fluorescein were obtained from Molecular Probes.
Site-specific Mutagenesis and Cloning of LMW-PTP Mutants in Eukaryotic Expression Vector-Oligonucleotide-directed mutagenesis was performed using the Unique Restriction Elimination Site kit from Amersham Pharmacia Biotech. The 26-base-long target mutagenesis primer contained an ACA codon (alanine) substituted for the original TGC codon (cysteine). The mutated LMW-PTP coding sequence was completely sequenced by the Sanger method and subcloned into the HindIII and ApaI restriction sites of the pRcCMV eukaryotic expression vector harboring the neomycin resistance gene.
Cell Culture and Transfections-NIH3T3 cells were routinely cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum in a 5% CO 2 -humidified atmosphere. 10 g of pRc-CMV-wtLMW-PTP or pRcCMV-C17A-LMW-PTP were transfected in NIH3T3 cells using the calcium phosphate method. Stable transfected clonal cell lines were isolated by selection with G418 (400 g/ml). Control cell lines were obtained by transfecting 2 g of pRcCMVneo alone. The clonal lines were screened for expression of the transfected genes by a) Northern blot analysis and b) enzyme-linked immunosorbent assay using polyclonal anti-LMW-PTP rabbit antibodies, which do not cross-react with murine endogenous LMW-PTP.
Immunoprecipitation and Western Blot Analysis-1 ϫ 10 6 cells were seeded in 10-cm plates in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. Cells were serum-starved for 24 h before receiving 30 ng/ml PDGF-BB. Cells were then lysed for 20 min on ice in 500 l of RIPA lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% Nonidet P-40, 2 mM EGTA, 1 mM sodium orthovanadate, 1 mM phenylmethanesulfonyl fluoride, 10 g/ml aprotinin, 10 g/ml leupeptin). Lysates were clarified by centrifugation and immunoprecipitated for 4 h at 4°C with 0.1 g of the specific antibodies. Immune complexes were collected on protein A-Sepharose (Amersham Pharmacia Biotech), separated by SDS-polyacrylamide gel electrophoresis, and transferred onto nitrocellulose (Sartorius). Immunoblots were incubated in 3% bovine serum albumin, 10 mM Tris-HCl, pH 7.5, 1 mM EDTA, and 0.1% Tween 20 for 1 h at room temperature, probed first with specific antibodies, then probed with secondary antibodies conjugated with horseradish peroxidase, washed, and developed with the Enhanced Chemi-Luminescence kit.
Cell Lysate Fractionations-Cell lysate fractions were obtained as already described (6). Briefly, PDGF-stimulated NIH3T3 cells were lysed in RIPA buffer, and the lysates were clarified by centrifugation for 30 min at 20,000 ϫ g. Pellets were washed twice with 1 ml of RIPA and then resuspended in complete RIPA buffer (cRIPA), which is RIPA buffer plus 0.5% sodium deoxycholate and 0.1% sodium dodecylsulfate, by shaking for 1 h at room temperature and clarifying by centrifugation at 20,000 ϫ g for 30 min. RIPA or cRIPA fractions were then used for immunoprecipitation analysis.
PTP Activity Assay-The PTP activity was measured as previously reported (7). Briefly, 1.5 ϫ 10 6 cells were collected in 300 l of 0.1 M sodium acetate, pH 5.5, 10 mM EDTA, 1 mM ␤-mercaptoethanol and sonicated for 10 s. The lysates were clarified by centrifugation, and 50 l were used in the PTP activity assay with 50 l of 10 mM paranitrophenylphosphate, at 37°C for 30 min. The production of p-nitrophenol was measured colorimetrically at 410 nm. The results were normalized on the basis of total protein content. Furthermore, PTP activity was measured in anti-LMW-PTP immunoprecipitates in 300 l of 0.1 M sodium acetate, pH 5.5, 10 mM EDTA, 1 mM ␤-mercaptoethanol at 37°C for 15 and 45 min for wtLMW-PTP or C2C12 myoblasts, respectively.
In Vitro Labeling with 5Ј-F-IAA-Affinity-purified wtLMW-PTP was obtained using a procedure reported elsewhere (19). 1 g of the purified enzyme was incubated in a final volume of 20 l at 37°C in 50 mM buffers of the indicated pH (sodium acetate for pH 5.5, MES for pH 6.5, and Tris-HCl for pH 7.5). H 2 O 2 was added at the concentration of 0.2 or 2 mM, and the mixture was incubated for an additional 10 min at 37°C. After the addition of 20 M 5Ј-F-IAA from freshly prepared stock in dimethylformamide, the mixture was incubated for an additional 10 min. The labeling reaction was stopped by the addition of 2ϫ SDS sample buffer.
In Vivo 5Ј-F-IAA Labeling-Cells were lysed directly in RIPA buffer at pH 7.5, and 5Ј-F-IAA was added from freshly prepared stock to a final concentration of 5 M. The lysates were maintained for 1 h at 4°C for the labeling step and then were treated for immunoprecipitation with anti-fluorescein antibodies.
Measurement of Intracellular ROS-2 ϫ 10 5 cells were plated in 6-cm plates in standard culture medium, serum-starved for 24 h, and then stimulated with 30 ng/ml PDGF-BB. 20 g/ml 2Ј,7Ј-dichlorodihydrofluorescein diacetate, an oxidant-sensitive fluorescent dye, were added 5 min before analysis. Cells were then rapidly detached from the substrate by trypsinization and analyzed immediately by flow cytometry using a Becton Dickinson FACScan flow cytometer equipped with an argon laser lamp (FL-1; emission, 480 nm; band pass filter, 530 nm).

Regulation of LMW-PTP Activity on PDGF Receptor during
Extracellularly Generated Oxidative Stress-As already reported LMW-PTP is oxidized in vitro by H 2 O 2 and NO (11,12) in its catalytic site cysteines 12 and 17. To study if the generation of an oxidative stress in vivo could lead to LMW-PTP oxidation, thus influencing its catalytic activity, we have used the NIH3T3 fibroblast cell line overexpressing wtLMW-PTP. We have quantitated the LMW-PTP activity directly on activated PDGF-R because we previously reported that the tyrosine-phosphorylated receptor is an in vivo natural substrate for LMW-PTP (4). We immunoprecipitated the PDGF-R from cells and evaluated its tyrosine phosphorylation level by antiphosphotyrosine immunoblotting as an indication of the LMW-PTP in vivo activity. We exposed NIH3T3 cells, which were either expressing wtLMW-PTP or mock-transfected, to an exogenous oxidative stress generated by adding 100 milliunits/ml glucose oxidase (G/O) to the culture medium. It has been reported that after 10 min of treatment H 2 O 2 concentration reaches a steady state level of 30 M (20). Fig. 1A shows the PDGF-R phosphorylation level in mock-transfected and wtLMW-PTP-expressing cells during G/O treatment. Dephosphorylation of agonist-activated PDGF-R by ectopically expressed wtLMW-PTP was almost completely blocked by the oxidative stress. The same results were obtained using 0.1 mM sodium pervanadate as the oxidant (Fig. 1B). It has been reported that although sodium orthovanadate is a competitive inhibitor for phosphotyrosine phosphatase, its oxidized form pervanadate is a strong oxidant of the PTPs active site cysteine (21).
We have previously demonstrated that LMW-PTP exists in two different cellular pools: a cytosolic pool, which is recruited to the membrane upon PDGF stimulation acting directly on PDGF-R, and a second pool anchored to the cytoskeleton that acts on a different substrate, p190Rho-GAP (10). LMW-PTP is thus able to influence cell growth through PDGF-R dephosphorylation and cytoskeleton rearrangement through Rho regulation. The results shown in Fig. 1C demonstrate that H 2 O 2 can down-regulate LMW-PTP activity on p190Rho-GAP as well as on PDGF-R, suggesting a general and strong action of oxidative stress on both cytosolic and cytoskeleton-associated LMW-PTP pools.
Furthermore we were interested in directly evaluating the influence of an exogenous oxidative stress on LMW-PTP activity, using a synthetic substrate, para-nitrophenylphosphate (PNPP), in an ex vivo assay. Fig. 1D shows the results on G/O-stressed cells. The PTP activity of wtLMW-PTP-expressing cells is strongly inhibited by H 2 O 2 treatment, confirming the results obtained with phosphorylated PDGF-R and p190Rho-GAP. Caselli et al. (11,12) have demonstrated that LMW-PTP is oxidized in vitro by both H 2 O 2 and NO and is able to rescue its catalytic activity after treatment with reducing agents. It is interesting to note that further treatment with catalase for 10 min (to remove H 2 O 2 ) led to a complete recovery of LMW-PTP activity (Fig. 1D), indicating that in the cell LMW-PTP is efficiently reduced again after removal of the oxidant.
To analyze whether the oxidized wtLMW-PTP is able to recover its catalytic activity in vivo, we evaluated the wtLMW-PTP activity on PDGF-R after the removal of the oxidative stress. After 10 min of G/O and PDGF treatment the produced H 2 O 2 was removed by adding 1 g/ml catalase to the medium. The activity of wtLMW-PTP was quantitated by anti-phosphotyrosine immunoblotting of anti-PDGF-R immunoprecipitates as above. Band intensities were quantitated by analytical software, and the results are shown in Fig. 2A. Our data show that 20 min after H 2 O 2 removal the level of PDGF-R phosphorylation was comparable with that of the untreated control, suggesting that wtLMW-PTP is able to recover its catalytic activity on PDGF-R. Hence, in vivo the oxidation of LMW-PTP active site cysteines is followed by a reduction process, which permits the recovery of the activity of the phosphatase on PDGF-R. To study the reactivation mechanism of oxidized LMW-PTP we analyzed the role of the redox cellular system based on reduced glutathione, which has been shown to be involved in the safety mechanism of various cell molecules (22). We reduced the availability of cellular reduced glutathione by using a strong depleting agent, diethyl maleate (DEM). NIH3T3 cells overexpressing wtLMW-PTP were pretreated for 1 h with 0.2 mM DEM. This treatment led to a 90% decrease of the level of free reduced glutathione (data not shown). Fig. 2B shows the tyrosine phosphorylation level of the activated PDGF-R from oxidized and non-oxidized wtLMW-PTP-expressing cells pretreated with DEM. The depletion of reduced glutathione severely impaired the rescue of oxidized wtLMW-PTP activity after H 2 O 2 removal, suggesting a central role of this reducing agent in the reduction/reactivation of LMW-PTP.
Oxidative Inactivation of LMW-PTP during PDGF Signaling-Stimulation of a variety of cell surface receptors, including those of growth factors such as epidermal growth factor and PDGF, induces a transient increase in the intracellular concen-tration of H 2 O 2 (23,24). In fact, the removal of cellular H 2 O 2 by catalase treatment blocks PDGF-mediated signal transduction. In this context it is possible that LMW-PTP is oxidized during PDGF signaling and that this oxidation transiently influences its catalytic action on PDGF-R.
LMW-PTP contains eight cysteine residues, which do not appear to form disulfide bonds. In fact, the reduced state of all the sulfhydryl groups of LMW-PTP has been demonstrated by our group and others (25,26). Both cysteines in positions 12 and 17 are conserved among the LMW-PTP subfamily, thus suggesting an important role for both residues. We have already demonstrated that Cys-12 is the essential cysteine residue performing in the cysteinyl-phosphate intermediate during catalysis, whereas Cys-17 has a role in phosphate binding in cooperation with Arg-18 (27). Van Etten et al. (26) have demonstrated that LMW-PTP has two reactive cysteines in the catalytic site with a low pK a (at pH 7.52 and 9.05 for iodoacetamide). Caselli et al. (11,12) have demonstrated that H 2 O 2 and NO lead to the specific oxidation of Cys-12 and Cys-17 in the catalytic pocket of LMW-PTP. In addition the active site cysteines of LMW-PTP, as in other PTPs, are specifically targeted by the sulfhydryl-modifying reagent iodoacetic acid (25,26).
Recently an iodoacetamide-fluorescein (5Ј-F-IAA) labeling method of proteins containing low pK a cysteine residues has been reported (28), and we applied this method to study the oxidation state of LMW-PTP in vivo. We verified that in vitro the reaction of purified wtLMW-PTP with 5Ј-F-IAA was almost completely blocked by 0.2 mM H 2 O 2 (Fig. 3A) at pH 6.5-7.5, suggesting a specificity of 5Ј-F-IAA for LMW-PTP active site reduced cysteines in these conditions.
To verify whether during PDGF signaling LMW-PTP becomes oxidized at the two catalytic cysteines, wtLMW-PTPexpressing cells were incubated for various times with PDGF and then lysed in RIPA buffer, pH 7.0, containing 5 M 5Ј-F-IAA according to Zhang et al. (26). LMW-PTP was then immunoprecipitated from lysates, and anti-fluorescein immunoblotting was performed. The result is shown in Fig. 3B. LMW-PTP was strongly oxidized during PDGF signaling, reaching a maximum 10 min after stimulation with PDGF. The oxidation was followed by a new reduction, which reached the level of unstimulated cells 40 min after stimulation, suggesting that the oxidation/inactivation of LMW-PTP is transient and the enzyme is reactivated after the removal of the oxidants by endogenous mechanisms. The blot was reprobed with anti-LMW-PTP antibodies, and band intensities of both anti-fluorescein and anti-LMW-PTP immunoblots were quantitated. The ratio between these two values is reported in Fig. 3C as percentage of LMW-PTP reduction, allowing us to demonstrate that about 80% of LMW-PTP is already oxidized 2 min after the stimulation with PDGF and that oxidation becomes almost complete after 10 min. Moreover, within 40 min about 70% of LMW-PTP recovered the reduced state. Moreover, in a parallel experiment, we quantitated the LMW-PTP activity of immunoprecipitated LMW-PTP in PDGF-treated cells by using an enzymatic assay with PNPP as substrate. The results, reported in Fig. 3D, demonstrate that the oxidation of the phosphatase temporally coincided with enzyme inactivation while the rereduction was concomitant with the rescue of catalytic activity. Interestingly, oxidation/inactivation of LMW-PTP reached the maximum after 10 min together with the ROS production peak, which was measured by 2Ј,7Ј-dichlorodihydrofluorescein diacetate labeling (data not shown).
It has been reported that the H 2 O 2 transient increase observed after growth factor stimulation is due to the activity of a nonphagocytic NADPH oxidase (29). To assess whether the oxidation of LMW-PTP during PDGF signaling is really due to endogenously produced H 2 O 2, we used a strong inhibitor of the NADPH oxidase complex, diphenyl iodide (DPI). wtLMW-PTPexpressing cells were pretreated or not for 30 min with 10 M DPI, and the oxidation of LMW-PTP was assayed by 5Ј-F-IAA labeling and anti-fluorescein immunoblotting as described above (Fig. 3E). In addition, DPI pretreatment caused a decrease of the PDGF receptor phosphorylation level in agreement with the absence of phosphatase inhibition (data not shown). The result demonstrates that the oxidation of LMW-PTP after PDGF treatment is impaired by the inhibition of NADPH oxidase by DPI, suggesting a central role of endogenously produced H 2 O 2 in LMW-PTP oxidation.
To confirm the redox regulation of LMW-PTP during PDGF signaling, we assayed the redox state of endogenous phosphatase in NIH3T3 cells and in C2C12 myoblasts. Unfortunately the level of LMW-PTP in growing NIH3T3 cells is almost undetectable but rapidly increases when cells reach confluence (3). 2 On the other hand, C2C12 mouse myoblasts naturally express a level of LMW-PTP higher than that in NIH3T3 cells.
For these reasons, we have analyzed the redox state of endogenous LMW-PTP in PDGF-stimulated confluent NIH3T3 cells and in C2C12 cells. The oxidation of LMW-PTP was assayed by 5Ј-F-IAA labeling. LMW-PTP was then immunoprecipitated from lysates, and anti-fluorescein immunoblotting was performed. The result is shown in Fig. 4A. Endogenous LMW-PTP was strongly oxidized after PDGF treatment. Moreover, the oxidation was followed by a new reduction thereafter. These data confirm that the oxidation/inactivation of LMW-PTP (either endogenous or overexpressed) is transient, and the enzyme is reduced/reactivated after the removal of the oxidants.
In addition, in a parallel experiment we quantitated PTP activity in anti-LMW-PTP immunoprecipitates after PDGF treatment. Endogenous LMW-PTP from C2C12 cells was immunoprecipitated from PDGF-stimulated cells, and enzyme activity is shown in the plot of Fig. 4B. The data confirm that the oxidation of LMW-PTP during PDGF stimulation is followed by enzyme inactivation, while the re-reduction thereafter is accompanied by enzyme reactivation.
Furthermore, we studied the role of the reactivation of LMW-PTP during PDGF signaling. For this purpose, we blocked the 2 P. Chiarugi, unpublished observations. glutathione-dependent cellular system of oxidized protein reduction by using an inhibitor of the ␥-glutamylcysteine synthetase, buthionine sulfoxide (BSO) (30). The cellular glutathione concentration was decreased to 90% by a 24-h pretreatment of NIH3T3 cells (data not shown). Mock-transfected and wtLMW-PTP-expressing NIH3T3 cells were pretreated or not for 16 h with 25 mM BSO, and the tyrosine phosphorylation level of PDGF-R was assayed by anti-phosphotyrosine immunoblotting. The results are shown in Fig. 5. In wtLMW-PTP-expressing cells the BSO treatment almost completely blocked the action of wtLMW-PTP on PDGF-R, whereas in mock-transfected cells BSO was ineffective on the level of phosphorylation of PDGF-R. These findings suggest that the GSH-dependent reduction process on oxidized LMW-PTP after PDGF treatment is a key control event of the phosphatase activity on PDGF-R.
Role of Cys-17 in Rescue of LMW-PTP Activity on PDGF Receptor after Removal of Oxidative Stress-It has been reported both in vivo and in vitro that the oxidation of a sulfhydryl group to sulfenic acid (SOH) rapidly and spontaneously evolved into further oxidation products such as sulfinic acid and sulfonic (SO 2 and SO 3 ) (22). The milder oxidation to sulfenic acid of the catalytic site cysteine of a phosphatase is sufficient to inactivate the enzyme, although the reversibly of the reaction (i.e. the reduction to SH) is yet possible. The further oxidation states (SO 2 and SO 3 ) are probably terminal products that do not permit cellular reduction systems to op-erate on them, thus leading to inactive modified proteins (see Scheme 1, where cat. is catalytically). In this light, only a milder oxidation could be intended as a possible redox functional regulation. It has been reported very recently that in vitro PTP1B is able to recover its catalytic activity after the removal of the oxidative stress by the thioredoxin system (14) or by the GSH system (15). The authors (14,15) suggest that the cellular reduction system could play a central role in maintaining a low oxidation grade in PTP1B through the formation of a mixed disulfide between PTP1B and GSH. All members of the LMW-PTP subfamily have two vicinal cysteines in positions 12 and 17, which are both in the catalytic pocket and totally conserved among the members of the subfamily. These two cysteines in vitro are able to form an intramolecular S-S bridge (12). We supposed that the redox functional regulation of this enzyme in vivo could be simply achieved by the formation of an intramolecular disulfide bond between Cys-12 and Cys-17. Mutations at the two cysteines had already been generated. Although the mutation in the Cys-12 totally inactivates the enzyme in vivo and in vitro (4), the C17A mutant retains in vitro about 70% of its activity, suggesting a minor involvement of this residue in the catalytic mechanism (27,28). To evaluate the in vivo activity of the C17A-LMW-PTP mutant, we transfected this mutant in NIH3T3 cells. First we demonstrated that C17A-LMW-PTP was active on phosphorylated PDGF-R. The results are reported in Fig. 6A. C17A-LMW-PTP retained almost 70% of the activity on the receptor with respect to wtLMW-PTP. This finding is in agreement with the activity of the C17A-LMW-PTP mutant measured in vitro on PNPP (32). 3

FIG. 4. Endogenous LMW-PTP redox regulation in response to
PDGF. Panel A, confluent NIH3T3 cells or C2C12 myoblasts were serum-starved for 24 h and then stimulated with 30 ng/ml PDGF-BB for the indicated times (Ј represents minutes). Lysates were treated with 5Ј-F-IAA as indicated under "Experimental Procedures." LMW-PTP was immunoprecipitated, and anti-fluorescein immunoblotting was performed. Loading equalization was performed by immunoblotting the stripped blots with anti-LMW-PTP. Panel B, 1 ϫ 10 6 C2C12 cells were serum-starved for 24 h and then stimulated with 30 ng/ml PDGF-BB for the indicated times (in minutes). LMW-PTP was immunoprecipitated at the indicated times, and a PTP activity assay was performed using PNPP as substrate. LMW-PTP activity is shown in units/mg. To study the role of the cysteine in position 17 on the oxidation/inactivation and reduction/reactivation processes of LMW-PTP, we treated NIH3T3 cells expressing wtLMW-PTP or C17A-LMW-PTP with PDGF and G/O for 10 min (Fig. 6B), and then we removed the oxidative stress by catalase treatment for an additional 30 min (Fig. 6C). The level of the PDGF-R tyrosine phosphorylation was quantitated by anti-phosphotyrosine immunoblotting, and band intensities were revealed by soft-ware analysis. The results show that both the wild type enzyme and the mutant C17A-LMW-PTP were inactivated by H 2 O 2 (Fig. 6B). On the contrary, although wtLMW-PTP recovered its activity 30 min after the removal of the oxidative stress, the rescue of the C17A-LMW-PTP H 2 O 2 -inactivated mutant was severely impaired upon catalase treatment (Fig. 6C). These findings are in agreement with the results of Caselli et al. (11,12) who have shown that both LMW-PTP catalytic site cys-FIG. 6. Role of Cys-17 in the redox regulation of LMW-PTP during extracellularly generated oxidative stress. Panel A, 1 ϫ 10 6 NIH3T3 cells overexpressing wtLMW-PTP or C17A-LMW-PTP were serum-starved for 24 h and then stimulated or not with 30 ng/ml PDGF-BB for 10 min (10Ј). PDGF-R was immunoprecipitated. The anti-phosphotyrosine immunoblot is shown. Panel B, 1 ϫ 10 6 NIH3T3 cells overexpressing wtLMW-PTP (wt) or C17A-LMW-PTP (C17A) were serum-starved for 24 h and then stimulated with 30 ng/ml PDGF-BB and 100 milliunits/ml G/O for 10 min. Panel C, cells were treated first as in panel B, the medium was changed, and then 1 g/ml of catalase was added for an additional 30 min. PDGF-R was immunoprecipitated, and anti-phosphotyrosine immunoblotting was performed. The histogram shows the data from the densitometric analysis of the blots. Panel D, cells were treated as in panels B and C, but fresh lysates were used for an in vitro PTP assay using PNPP as substrate. PTP-specific activity is shown in units/mg. C, control; G/OX, G/O; cat, catalase. teines are oxidized by H 2 O 2 and NO forming a disulfide bond and suggest that the specific role of Cys-17 is in the reactivation/reduction of LMW-PTP and not in the inactivation/oxidation. To better assess these findings we repeated the same experiment as above but directly evaluated LMW-PTP-specific activity on PNPP (Fig. 6D). The results again demonstrate that C17A-LMW-PTP was inactivated by oxidation, as was wtLMW-PTP (although to a lower extent), but was completely unable to recover its activity on PNPP upon removal of the oxidative stress.
We proposed that the same Cys-17-mediated redox regulation of inactivation/reactivation of LMW-PTP is achieved during the PDGF signaling-dependent oxidative stress mediated by H 2 O 2 production. To test this hypothesis, we analyzed the redox state of LMW-PTP by in vivo 5Ј-F-IAA labeling and anti-fluorescein immunoblotting after PDGF treatment of wtand C17A-LMW-PTP-expressing cells (Fig. 7). The results show again that the position 17 was irrelevant in the oxidation of LMW-PTP after PDGF stimulation. On the contrary, Cys-17 was essential in the reduction process of LMW-PTP, which leads to enzyme reactivation.
On the basis of these data we propose that LMW-PTP is able to rescue its enzymatic activity upon GSH-dependent reduction of the disulfide bridge between Cys-12 and Cys-17 that is formed upon oxidation. This disulfide bond can be considered as a protection system of LMW-PTP against further oxidation. In fact, protection of the catalytic Cys-12 from oxidation states higher than sulfenic acid of the catalytic Cys-12 could be achieved either by GSH mixed disulfide or by an intramolecular Cys-12-Cys-17 disulfide. The impairment of oxidized C17A-LMW-PTP reduction (which still possess a functional Cys-12) excludes the possibility of a Cys-12-GSH mixed disulfide as a protection mechanism and supports our hypothesis of the formation of an intramolecular Cys-12-Cys-17 bridge. DISCUSSION On the basis of the growing evidence of the requirement of endogenously produced oxidants during growth factor signal transduction (13,27,28), we investigated in vivo the possible redox regulation of LMW-PTP by H 2 O 2 during PDGF stimulation in NIH3T3 cells. It has been reported that the H 2 O 2 produced upon growth factor stimulation is really a cellular second messenger directly influencing many signal transducers such as the STAT proteins, p70 s6k, protein kinase D, and PTPs (33)(34)(35)(36). We have already demonstrated that LMW-PTP is a key regulator of PDGF-R phosphorylation upon recruitment to the membrane when cells are stimulated with the growth factor (5,6,10).
In this article we first show that LMW-PTP was oxidized in vivo by an exogenous oxidative stress such as H 2 O 2 produced by G/O or sodium pervanadate. We have already demonstrated that in NIH3T3 cells LMW-PTP is active on phosphorylated PDGF-R and on p190Rho-GAP, thus influencing both mitosis rate and cytoskeleton rearrangement. Herein, we quantitated the in vivo activity of LMW-PTP directly on its natural phosphorylated substrates and demonstrate that the oxidation of LMW-PTP led to the inactivation of the enzyme preventing dephosphorylation of both PDGF-R (Fig. 2, A and B) and p190Rho-GAP (Fig. 2B). We have previously demonstrated that LMW-PTP is distributed in the cell in two different pools, a cytosolic pool, which acts on PDGF-R, and a cytoskeleton-associated pool, which is active on p190Rho-GAP (10). In agreement with the extremely rapid diffusion of small oxidants such as H 2 O 2 among cellular compartments, we observed a strong inhibition of LMW-PTP (via oxidation) on both activated PDGF-R and phosphorylated p190Rho-GAP.
The inactivation of LMW-PTP by oxidation is reversible in vitro upon reductant treatment with such agents as dithiothreitol (11,12). Herein we show that LMW-PTP was able to rescue its catalytic activity in vivo on its substrate PDGF-R after the removal of the oxidative stress. The reversibility of oxidation and the consequent recovery of enzymatic activity are generally recognized as key points in the redox regulation of a protein. In fact, it has been widely discussed that protein oxidation should have two different meanings in cell behavior. Oxidation of the protein backbone or the simple oxidation of methionine and/or cysteine residues could lead to direct protein fragmentation or to irreversibly oxidized "dead" products (37). On the other hand, regulated oxidation of amino acid side chains and in particular of cysteine residues should be considered a reversible functional regulation of proteins because the oxidation might be reversed by the redox cellular systems (thioredoxin and GSH/glutaredoxin). Herein, we show that the reduction process that permits the recovery of LMW-PTP activity was dependent on the cellular reduced glutathione content as indicated by the data obtained using BSO or DEM to deplete glutathione (Figs. 2 and 5). Thioredoxin/thioredoxin reductase/ NADPH and glutaredoxin/glutathione/glutathione reductase/ NADPH are the two major redox cellular systems (38). Our findings indicate that in vivo the LMW-PTP reactivation is under the control of the glutaredoxin/glutathione/glutathione reductase/NADPH system. Moreover, we have shown that the redox regulation of LMW-PTP in vivo activity i) was active upon PDGF stimulation, ii) was performed by the endogenously produced H 2 O 2 , and iii) was required for LMW-PTP dephosphorylation of PDGF-R itself. In fact, we demonstrate that upon PDGF treatment almost 80% of LMW-PTP was oxidized and inhibited after 10 min and that after 45 min almost 70% of the phosphatase was reduced and had recovered its catalytic activity. Also in these conditions the reduction process was dependent on the availability of reduced glutathione because the reduction process was impaired by BSO treatment (Fig. 7). In addition, we have shown that this redox regulation was active also on the endogenous FIG. 7. Role of Cys-17 in the redox regulation of LMW-PTP during PDGF signaling. 1 ϫ 10 6 NIH3T3 cells overexpressing wtLMW-PTP or C17A-LMW-PTP were pretreated or not with BSO for 16 h, serum-starved for 24 h, and then stimulated with 30 ng/ml PDGF-BB for 10 min. The medium was then changed, and 1 g/ml catalase (cat) was added when indicated for an additional 20 min. LMW-PTP was immunoprecipitated, and anti-fluorescein immunoblotting was performed. enzyme and hence that was not a particular feature of overexpressed LMW-PTP.
It has been proposed that the transient increase of ROS observed in response to growth factor administration is due to the induced activation of a membrane NADPH oxidase complex (39). This event is dependent on the administration of PDGF and is probably under the control of the signal transduction pathway of phosphatidylinositol 3-kinase (40). We have shown that the oxidation of LMW-PTP was performed by an endogenously produced oxidative stress as indicated by the use of DPI, an inhibitor of the NADPH oxidase. This treatment clearly prevented the oxidation of LMW-PTP in response to PDGF (Fig. 3D).
Furthermore, our data suggest that the redox regulation of LMW-PTP during PDGF stimulation was required for PDGF-R dephosphorylation. In fact, the inhibition of the recovery of the LMW-PTP catalytic activity by reduction with glutathionedepleting agents such as BSO severely impaired the action of LMW-PTP on phosphorylated receptor. These data, together with the time-dependent oxidation/inactivation and reduction/ reactivation of LMW-PTP upon PDGF stimulation, indicate a specific role of the inhibition of LMW-PTP during PDGF-R signal transduction. In fact, the phosphatase was oxidized/ inactivated (up to 80%) at the very beginning of agonist stimulation and reduced/reactivated later when switching off of the receptor signal is required. Lee et al. (14) have proposed a different explanation for a similar behavior of PTP1B during epidermal growth factor stimulation. They propose that the growth factor stimulation alone may not be sufficient to increase the steady state level of protein tyrosine phosphorylation in the cell and that concurrent inhibition of protein tyrosine phosphatases by H 2 O 2 might also be required (14). This second hypothesis is supported by data about the PDGF-R activation by UV rays through PTP inactivation/oxidation (41). PTP inhibition permits the transduction of the signal through the maintenance of a high receptor phosphorylation level. Hence, PTPs could be seen either as regulators of the turning off of the growth factor signal or as concurrent activators of the receptor phosphorylation. Our data on BSO impairment of LMW-PTP action on PDGF-R dephosphorylation suggest that the main action of the phosphatase is after the redox reactions upon PDGF stimulation and hence in the late phase of the receptor signaling.
The function of glutathione in the redox regulation of PTP1B is to protect the sulfenic derivative from further oxidation because it can easily be oxidized to form the irreversible sulfinic and sulfonic products. In fact, cysteine sulfenic acids are highly unstable and readily undergo condensation with a thiol. In the presence of reduced glutathione, the sulfenic derivative can be converted to a more stable S-thiolated product (15). This reaction prevents a further oxidation to sulfinic or sulfonic acid. This is a general mechanism that should operate on every PTP that possesses the catalytic cysteine. The different behavior of glutathione in the protection of PTP1B or LMW-PTP from further oxidation could be because of differences in the structures of the catalytic sites of the two phosphatases. In fact, the PTP1B active site cleft is surrounded by positively charged lysine residues, which could interact with the negatively charged glutathione. On the other hand, the LMW-PTP active site crevice is rich in aromatic residues, a structure that should not facilitate the entry of GSH.
We propose that LMW-PTP is able to undergo a different mechanism because of the presence of an additional cysteine in the catalytic site, which can form an intramolecular S-S bridge. This intramolecular S-S bond between Cys-12 and Cys-17 has only been observed in vitro in mass spectroscopy-analyzed of NO-oxidized LMW-PTP (11). In fact, the C17A-LMW-PTP mutant still possesses the catalytic cysteine in position 12, which could be glutathionated and thus preserved from further oxidation. On the contrary, we observed a dramatic decrease in the catalytic activity rescue of the C17A-LMW-PTP mutant upon reduction, suggesting a key role of the intramolecular S-S disulfide in the protection of LMW-PTP from excessive oxidation. We suggest that the presence of this additional cysteine in the catalytic site, which allows intramolecular S-S bond formation, confers a peculiar rapidity and efficiency in the redox regulation of the phosphatase. The formation of an intramolecular disulfide bond in Yab1 transcription factor has already been proposed to be responsible for the rapid response to the stress condition in H 2 O 2 sensing (31).
In conclusion, we propose that LMW-PTP has a peculiar safety mechanism during its redox regulation, based on the formation of an intramolecular disulfide bond between the two catalytic cysteines, that could allow a rapid and efficient recovery of catalytic activity during both extracellularly and intracellularly generated oxidative stress, thus leading to a fine tuning of PDGF-R activation.