Protein Kinase Cδ Induces Src Kinase Activity via Activation of the Protein Tyrosine Phosphatase PTPα*

Previously we have shown that protein kinase C (PKC)-mediated reorganization of the actin cytoskeleton in smooth muscle cells is transmitted by the non-receptor tyrosine kinase, Src. Several authors have described how 12-O-tetradecanoylphorbol-13-acetate (TPA) stimulation of cells results in an increase of Src activity, but the mechanism of the PKC-mediated Src activation is unknown. Using PKC isozymes purified from Spodoptera frugiperda insect cells, we show here that PKC is not able to activate Src directly. Our data reveal that the PKC-dependent Src activation occurs via the activation of the protein tyrosine phosphatase (PTP) PTPα. PTPα becomes activated in vivo after TPA stimulation. Further, we show that PKCδ phosphorylates and activates only PTPα in vitro but not any other of the TPA-responsive PKC isozymes that are expressed in A7r5 rat aortic smooth muscle cells. To further substantiate our data, we show that cells lacking PKCδ have a markedly reduced PTPα and Src activity after 12-O-tetradecanoylphorbol-13-acetate stimulation. These data support a model in which the main mechanism of 12-O-tetradecanoylphorbol-13-acetate-induced Src activation is the direct phosphorylation and activation of PTPα by PKCδ, which in turn dephosphorylates and activates Src.

Protein kinase C (PKC) 1 is a family of phospholipid-dependent serine/threonine kinases comprising 10 isozymes differing in their molecular domain organization of up to 4 variable and 3 constant regions and in their functions. These PKC isozymes are subdivided into three classes: (i)the "conventional" cPKCs, PKC-␣, -␤ (␤I and ␤II), and -␥, which can be activated by phosphatidylserine (PS), diacyl glycerol (DAG), or phorbol esters through binding to the C1 domain and Ca 2ϩ through binding to a Ca 2ϩ -binding site in their second constant region, C2; (ii) the "novel" nPKCs, PKC-␦, -⑀, -, -, which lack the C2 region and thus are Ca-independent but still DAG-, PS-, and phorbol ester-responsive; and (iii) the "atypical" aPKCs, PKC-/ and -, which also lack the C2 region and, in addition, are devoid of a functional DAG-binding site. Hence, the atypical PKCs are only responsive to PS but not to DAG or phorbol ester (reviewed in Refs. [1][2][3]. PKCs reside in the cytosol in an inactive conformation and translocate to the membrane (or other subcellular sites) upon activation, where they modify various cellular functions through phosphorylation of target substrates. Like other kinases, PKCs have been found to be involved in intracellular signal transduction pathways that regulate cell growth, differentiation, and apoptosis, and they have been implicated in the rearrangement of the cytoskeleton and migration (4).
Src-family tyrosine kinases also comprise a major group of cellular signal transducers. These tyrosine kinases can be activated by various extracellular signals and thus can modulate a variety of cellular functions, including proliferation, survival, adhesion, and migration (5). When phosphorylated at Tyr-527 Src is inactive. Activation is accomplished by dephosphorylation of this tyrosine residue, and the resulting conformational change facilitates an autophosphorylation at Tyr-416. After autophosphorylation Src is in an active state.
As described earlier, TPA activates Src in vivo (6 -9). Previously, we showed in A7r5 rat aortic smooth muscle cells that PKC-induced reorganization of the actin cytoskeleton involves the activation of Src tyrosine kinase activity (10). It is therefore very likely that PKC acts upstream of Src and mediates its activation, because PKC is known to be the main receptor of TPA in cells. However, neither a model for PKC-mediated Src activation nor even an explanation for the TPA stimulation of Src activity has been described so far.
The tyrosine phosphatase PTP␣, a 130-kDa transmembrane PTP, is known to be activated upon TPA stimulation and subsequent phosphorylation (11,12). Moreover, PTP␣ has been shown to be a physiological regulator of Src (13)(14)(15). Hence, we examined whether the TPA-induced activation of Src involves PTP␣. Here we present evidence of PKC-mediated Src activation via PTP␣.

EXPERIMENTAL PROCEDURES
Reagents and Antibodies-The monoclonal Src-antibody (Ab-1) was obtained from Oncogene, the HA-antibody was from Roche Applied Science, and the pY(416) Src antibody to measure Src activity (16 -18) was obtained from BIOSOURCE. The PTP␣-specific antibody was a kind gift from Jeroen den Hertog (Hubrecht Laboratories, Utrecht, Netherlands) (19). All secondary antibodies were from Dianova. TPA, GF109203X, and phosphatidylserine were obtained from Calbiochem. Thrombin, polyvinylidene difluoride and nitrocellulose membranes, the ECL-kit and [␥-32 P]ATP were from Amersham Biosciences. All other * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. chemicals were obtained from Sigma. All cell culture reagents were from Invitrogen.
Plasmids-To express GST-tagged PKCs in Spodoptera frugiperda (Sf9) insect cells, the complete coding regions of the relevant murine PKC cDNAs (20) were cloned in-frame into pAcGHLT-C or a modified pAcGHLT-C vector (BD Biosciences), where the thrombin cleavage site was exchanged for a TEV cleavage site. The accuracy of the inserted PKC sequences was proved by DNA sequencing.
To generate the recombinant baculovirus, PAcGHLT-C-PKC plasmids were cotransfected with Baculo-Gold (BD PharMingen, Dianova) DNA in Sf9 cells according to the manufacturers' recommendations.
Five plaques of each baculovirus were picked and amplified in Sf9 cells. Protein expression in infected cells was examined using immunoblot analysis. The activity of the recombinant protein was examined in an in vitro kinase assay, and the respective baculoviruses were used to infect Sf9 cells as described elsewhere (21). The baculovirus for c-Src was a kind gift from Margaret Frame (Beatson Laboratories, Glasgow, UK). The tetracyclin-inducible expression vector for HA-tagged PTP␣ was a kind gift from David Shalloway (Cornell University, Ithaca, NY) (14).
Cell Culture and Transfections-A7r5 rat vascular smooth muscle cells were grown in Dulbecco's modified Eagle's medium without phenol red containing 10% fetal calf serum and 2 mM glutamine. The medium was changed every 3 days. For transfection, A7r5 cells were seeded freshly and transfections were performed after ϳ24 h of culture or when cells were grown to 50 -70% confluence. The expression vector for HA-PTP␣ and transactivator-plasmid pTet-tTAk (Invitrogen) were resuspended in the above medium without serum at a ratio of 1:2 (10 g/10-cm plate). Superfect transfections reagent (Qiagen, Hilden, Germany) was added according to the manufacturer's recommendations. Transfected cells were examined 2 days after transfection.
Primary skin fibroblasts were isolated from newborn PKC␦Ϫ/Ϫ and PKC␦ϩ/ϩ mice from the same litter as described earlier (22) and cultured in the same way as A7r5 cells. Experiments were performed between passage 3-5. Sf9 insect cells were cultured at 27°C without CO 2 in spinner flasks, using Grace's insect medium containing 10% fetal bovine serum, 3.3 g/liter lactalbumin hydrolysate, 3.3 g/liter yeastolate, 50 g/ml gentamicin, and 2.5 g/ml fungizone.
PKC Kinase Assays-Recombinant PKC isozymes were prepared from Sf9 cells as described (21), bound to GSSH-Sepharose, and cleaved by thrombin. The activity of each isozyme was standardized in an in vitro kinase assay using myelin basic protein (MBP). One g of MBP was incubated with different amounts (5-20 ng) of the PKC isozymes in the presence of 20 mM Tris, pH 7.5, 5 mM MgCl 2 , 0.1% Nonidet P-40, 0.5 M TPA, 1.5 M phosphatidylserine, 10 M ATP, 5 Ci of [␥-32 P]ATP (3000Ci/mM, Amersham Biosciences), 1 g/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride (kinase buffer) in a final volume of 50 l for 15 min at 30°C. The reaction was stopped by adding 10 l of 6ϫ SDS sample buffer and heating to 95°C for 5 min. The samples were analyzed by SDS-PAGE (12.5%). The gel was stained with Coomassie Brilliant Blue before it was exposed to an x-ray film.
PTP␣ Activity and Src Activity Assays-A7r5 cells were transfected with HA-tagged PTP␣. 48 h after transfection the cells were serumstarved for 6 h, stimulated with 1 M TPA for 15 min, washed with phosphate-buffered saline, and lysed on ice in RIPA buffer containing 50 mM Tris, pH 7.5, 150 mM NaCl, 10 mM MgCl 2, 1% Triton, 0.1% SDS, 0.5% deoxycholate, 1 mM EDTA, 1 mM NaF, 1 g/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride. Lysates were cleared by centrifugation for 10 min at 14,000 rpm. Cleared lysates were incubated overnight at 4°C with 30 l of protein-G-Sepharose beads and 0.5 g of HAantibody or 20 l of the polyclonal PTP␣ antibody. Precipitated PTP␣ immunocomplexes were washed once with RIPA buffer (now containing 500 mM NaCl), twice with PPase assay buffer (0.1 mM succinic acid, pH 6.0, 150 mM NaCl, 1 mM EDTA) and subsequently incubated for 30 min in 200 l of PPase assay buffer containing 10 M p-nitrophenyl phosphate (pNPP). PTP␣ immunocomplexes were sedimented by centrifugation, an appropriate volume of 2ϫ sample buffer was added, and the amount of PTP␣ was estimated by immunoblotting using the PTP␣specific antibody. The supernatant was mixed with 200 l of 2 M NaOH, and the absorption at 415 nM was measured.
To determine the phosphatase activity toward Src, PTP␣ was immunoprecipitated from TPA-stimulated or unstimulated cells. Immunocomplexes were washed once with RIPA buffer (500 mM NaCl), twice with PPase assay buffer, and resuspended in 200 l of PPase-assay buffer. Recombinant Src was isolated from Sf9 cells via glutathione-GST interaction. The GST tag was then cleaved with thrombin and removed by the GSSH-Sepharose chromatography. 1 g of purified Src was added to each sample, and the sample was incubated at 30°C for 1 h with agitation.
The samples were washed twice with 600 l of buffer containing 50 mM Tris, pH 7.5, 150 mM NaCl, 5 mM MgCl, 5% glycerol, 1% Triton-X-100, 2 mM Na 3 VO 4, 2 mM ␤-glycerophosphate, 2 mM NaF, 2 mM pyrophosphate, 1 g/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride, and the supernatant was placed in a new tube. Src was immunoprecipitated from the sample using 1 g of the Src-specific antibody AB-1 coupled to protein-G-Sepharose for 2-4 h at 4°C. Src-containing beads were washed twice with kinase buffer (20 mM Tris, pH 7.5, 5 mM MgCl 2 , 0.1%Nonidet P-40, 2 mM ␤-glycerophosphate, 2 mM NaF, 2 mM pyrophosphate, 1 g/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride) and was permitted to undergo autophosphorylation in 50 l of kinase buffer containing 5 M ATP for 15 min at 37°C. The reaction was stopped by adding 10 l of 6ϫ SDS-sample buffer and heated to 95°C for 5 min. Equal amounts of the sample were separated in parallel by SDS-PAGE (10%) and subjected to immunoblot analysis, using the anti-Src antibody Ab-1 to quantitate the levels of c-Src and the Src pY(416) antibody to determine the activation state of Src (16 -18). To prove that equal amounts of PTP␣ were analyzed in each sample, an appropriate volume of 2ϫ sample buffer was added to PTP␣ immunocomplexes, and the amount of PTP␣ was estimated by immunoblotting using the PTP␣ -specific antibody.
To analyze the activity of PTP␣ after direct phosphorylation by PKC isozymes, HA-PTP␣ was immunoprecipitated from serum-starved unstimulated A7r5 cells. PTP␣ immunocomplexes were washed once with RIPA buffer (500 mM NaCl) and twice with kinase buffer and were subjected to the kinase assay with ϳ20 ng of recombinant PKC isozymes in the presence of 0.5 M TPA, 1.5 M phosphatidylserine, and 10 M ATP for 30 min at 30°C. For direct analysis of the phosphatase activity on pNPP, the phosphorylated phosphatase was washed twice with PPase assay buffer and incubated for 30 min in 200 l of PPase assay buffer containing 10 M pNPP. PTP␣ immunocomplexes were sedimented by centrifugation, an appropriate volume of 2ϫ sample buffer was added, and the amount of PTP␣ was again estimated by immunoblotting using the PTP␣-specific antibody. The supernatant was mixed with 200 l of 2 M NaOH, and the absorption at 415 nM was measured.
To measure the activity of PTP␣ on Src after direct PKC phosphorylation, PTP␣ immunocomplexes were washed twice with PPase assay puffer. The immunocomplexes were resuspended in 200 l of PPase assay buffer containing 1 g of purified Src, and the sample was incubated at 30°C for 1 h with agitation. The analysis of the Src activity was performed as described above. The Src kinase assays shown in Figs. 1A and 4B were done as described earlier (10) RESULTS Src Is Not Directly Activated by PKC-As shown previously (10), TPA stimulation of A7r5 cells leads to an increase of Src kinase activity in vivo. Src was immunoprecipitated from serum-starved A7r5 cells before and after TPA stimulation for 15 min, and the activity of Src was determined in immunoblot analysis using the pY(416)-specific antibody for Src. A robust increased immunoreactivity of Src from TPA-stimulated cells with the pY(416)-specific antibody was observed (Fig. 1A).
To ensure that the measurement of Src kinase activity with the Src pY(416)-specific antibody is a valid method to determine the activity of Src, we analyzed the Src activity in parallel directly in an in vitro kinase assay. Src was immunoprecipitated from A7r5 cells before and after TPA stimulation, and the kinase activity of Src was determined using acid-denatured enolase as an exogenous substrate. As shown in Fig. 1A, the kinase activity of Src measured on enolase correlates directly with the immunoreactivity of the Src pY(416)-specific antibody. A 2-3-fold increase of Src kinase activity could be observed after TPA stimulation.
Then we asked whether PKC isozymes could directly activate the kinase activity of Src, because Src has been described as a PKC substrate (23). To this end, Src was phosphorylated in vitro with the five TPA-responsive PKC isozymes that are present in wild-type A7r5 cells (␣, ␤I, ␦, ⑀, ), (24). Subsequently, autophosphorylation of Src at Tyr-416 was examined as a marker for Src activity utilizing a pY(416)-specific anti-body (see above). As is evident from Fig. 1B, phosphorylation of Src with the different PKC isozymes did not result in any significant differences in the phosphorylation status of Tyr-416. As a control for equal activity of the PKC isozymes that were used in this assay, the activity of PKC isozymes was determined in an in vitro kinase assay using MBP as a substrate in the presence of radiolabeled phosphate (Fig. 1B, lower  panel). Incubation of MBP with the different PKC isozymes resulted in an equal incorporation of radioactively labeled phosphate, indicating a comparable activity of the five PKC isozymes (lower panel).
These results show, as expected, that the kinase activity of Src is not directly regulated by PKC-mediated phosphorylation. Hence, an indirect mechanism must be responsible for the observed activation of Src in vivo after TPA stimulation.
TPA Increases the Activity of PTP␣ against pNPP and Src-The main regulatory mechanism regulating Src activity is phosphorylation of Tyr-527. Dephosphorylation of this residue is essential for Src kinase activation. Hence, phosphatases represent plausible candidates for signal transmitters from PKC to Src. One such candidate is PTP␣, which was reported to be positively regulated by PKC and which also is capable of activating Src (11, 12, 14). To test this scenario, we examined whether the activity of PTP␣ could be up-regulated in a TPAdependent manner in A7r5 cells.
To determine whether TPA treatment enhances the phosphatase activity of PTP␣, A7r5 cells were transfected with an HA-tagged PTP␣ expression vector. The transfected cells were stimulated with TPA for 15 min, and HA-PTP␣ was isolated by immunoprecipitation with an anti-HA antibody and tested for phosphatase activity with pNPP. As is evident in Fig. 2A, PKC activation did result in activation of PTP␣. The isolated HA-PTP␣ showed a 2-3-fold increased activity compared with the enzyme that was isolated from cells that were not stimulated with TPA. The inset immunoblot shows that equal amounts of PTP␣ were analyzed in both samples.
We next tested whether PTP␣ from TPA-stimulated cells is also capable of dephosphorylating and activating Src. To this end, we immunoprecipitated HA-tagged PTP␣ from A7r5 cells before and after stimulation with TPA for 15 min. Recombinant Src from Sf9 insect cells was incubated with the isolated PTP␣ in vitro. Subsequently, Src was isolated from the reaction mixture by immunoprecipitation. Based on the fact that autophos-

FIG. 1. Src is not activated directly by PKC phosphorylation.
A, Src was immunoprecipitated from serum-starved A7r5 cells before (Ϫ) and after (ϩ) TPA stimulation (15 min). The activity of Src was determined using the pY(416)-specific antibody for Src or directly in an in vitro kinase assay using acid-denatured enolase to ensure that the immunoreactivity of the pY(416)-specific antibody correlates directly with the kinase activity of Src. In both cases a robust increase of Src kinase activity could be observed after TPA stimulation. As a control, the amount of Src was analyzed using the Src-specific antibody. B, Src was immunoprecipitated from serum-starved, unstimulated A7r5 cells and subjected to an in vitro kinase assay with the five TPA-responsive PKC isozymes expressed in A7r5 cells (purified recombinant proteins from Sf9 cells) or left untreated as a control. The activity of Src was analyzed using the Src p(Y416)-specific antibody. As a control, the amount of Src was analyzed using the Src-specific antibody. No significant activation of Src could be observed in this assay. The intensity of the signal obtained from the phospho-specific antibody directly correlated with the signal from the Src-specific antibody. As a control for equal activity of the PKC isozymes that were used in this assay, MBP was used as a substrate in an in vitro kinase assay. Incubation of MBP with the different PKC isozymes resulted in an equal incorporation of radioactively labeled phosphate, indicating a comparable activity of the five PKC isozymes (lower panel). Data shown here are representative of three experiments.

FIG. 2. PTP␣ is activated after TPA stimulation.
A7r5 cells were transfected with an HA-PTP␣-encoding construct, and PTP␣ was immunoprecipitated with an HA-specific antibody from either TPA-stimulated (15 min) or unstimulated cells. A, the phosphatase activity was assayed using pNPP. A 2-3-fold increase of PTPase activity was detected in immunocomplexes that were derived from TPA-stimulated cells in comparison to unstimulated cells. The inserted upper panel shows a Western blot stained for PTP␣, indicating that equal amounts of PTP␣ were assayed. Data shown are mean Ϯ S.E. (n ϭ 3). B, the phosphatase activity of PTP␣ was analyzed with Src as a physiological substrate, and the activation state of Src was determined using the pY(416)-specific antibody. Incubation of recombinant Src with PTP␣ immunoprecipitated from TPA-stimulated cells resulted in an increase of Src autophosphorylation at Tyr-416 in comparison to Src that was incubated with PTP␣ from unstimulated cells. As a control that equal amounts of PTP␣ and Src were analyzed in each sample, the amount of PTP␣ and Src detected by immunoblot analyses is shown. Data shown here are representative of three experiments. phorylation, and hence activation, of Src depends on the previous dephosphorylation of pY(527), we used the autophosphorylation at Tyr-416 as a measure for the activity of Src. As is evident in Fig. 2B, the activity of Src is markedly increased after treatment with PTP␣ from TPA-stimulated cells, as compared with PTP␣ from cells that had not been stimulated with TPA.
PKC␦ Phosphorylates PTP␣ in Vitro-Although our data prove that PKC activation in vivo results in activation of PTP␣, we cannot exclude that another PKC-controlled kinase is responsible for the phosphorylation of PTP␣. Hence we examined whether PKC directly phosphorylated the phosphatase in vitro. A fusion protein of the cytoplasmic domain of PTP␣ and GST (GST⅐PTP␣cyt) was used as PKC substrate in an in vitro kinase assay. Only the PKC isozymes that are present in wild-type A7r5 cells were tested (␣, ␤I, ␦, ⑀, ) (24). The assay was performed with the same conditions used in Fig. 1B.
Equal amounts of GST⅐PTP␣cyt were present in the individual assays as judged by Coomassie staining. Although MBP was phosphorylated equally by all five isozymes (see Fig. 1), GST⅐PTP␣cyt was significantly phosphorylated only by PKC␦, as shown in Fig. 3A. Hence, the cytoplasmic domain of PTP␣ serves as a specific substrate for PKC␦ only.
To further substantiate these observations, we examined the PKC-phosphorylated PTP␣ utilizing two-dimensional phosphopeptide maps. Phosphorylation of PTP␣ with PKC␦ in vitro resulted in three major spots on the map (not shown). This resembles the data presented by Tracy et al. (11), with Ser-180 being responsible for two spots and Ser-204 for one.
In Vitro Phosphorylation of PTP␣ by PKC␦ Increases the Phosphatase Activity Significantly-To examine whether PKC␦ was able to activate the PTP␣ phosphatase activity, HA-tagged PTP␣ was isolated from serum-starved cells by immunoprecipitation. The isolated protein was phosphorylated in vitro with each of the five PKC isozymes expressed in A7r5. After incubation with PKC, the phosphatase activity was determined using pNPP as a substrate for the phosphatase. As clearly evident from Fig. 3B, only previous treatment with PKC␦, but not with any other PKC isozyme, increased the phosphatase activity significantly (up to 10-fold).
Next we examined whether the PTP␣ phosphatase activity was also increased when Src was used as a substrate. To this end, purified Src was subjected to autophosphorylation after PTP␣ treatment. The activity of Src was then determined in a Western blot utilizing the pY(416)-specific antibody. As is evident from Fig. 3C, PTP␣ that had been treated with PKC␦ resulted in a significantly higher activity/autophosphorylation of Src than PTP␣ that had been treated with other PKC isozymes. These data strongly suggest that the TPA-induced activation of Src is mediated by phosphorylation/activation of PTP␣ by PKC␦.
TPA-induced PTP␣ and Src Activation Is Decreased in Fibroblasts from PKC␦ Knockout Mice-To further substantiate our findings, we set out to examine PTP␣ and Src activation after TPA treatment of cells that lack PKC␦ (22). Fibroblasts from PKC␦ knockout mice were prepared and compared with fibroblasts from wild-type mice. These cells do not express PKC␦, as shown by Leitges et al. (22) and confirmed by our Western blot analysis (data not shown). Equal numbers of cells were stimulated with TPA for 15 min (or left untreated as a control). PTP␣ was immunoprecipitated from the cell lysates and analyzed for activity using pNPP as a substrate. As shown in Fig. 4A, TPA stimulation of wild-type fibroblasts resulted in activation of PTP␣. The isolated PTP␣ from TPA-stimulated cells showed roughly 2-fold increased activity when compared with PTP␣ from unstimulated cells. In contrast, we found no induction of PTP␣ activity in PKC␦ Ϫ/Ϫ fibroblasts after TPA stimulation. The inset immunoblot shows that equal amounts of PTP␣ were analyzed in both samples.
To test whether the diminished activation of PTP␣ in PKC␦Ϫ/Ϫ fibroblasts also resulted in a diminished activation of Src after TPA stimulation, we analyzed the activity of Src before and after TPA stimulation. Src was immunoprecipitated from PKC␦ ϩ/ϩ and Ϫ/Ϫ before and after TPA stimulation for 15 min, and the activity of Src was analyzed by Western blot using the Src pY(416)-specific antibody (Fig. 4B, upper panel) and directly in an in vitro kinase assay using acid-denatured enolase as a substrate (lower panel). As shown in Fig. 4B, TPA stimulation of wild-type fibroblasts resulted in a roughly 3-fold Src activation as compared with the non-stimulated cells. As expected from our previous results, cells from PKC␦Ϫ/Ϫ mice FIG. 3. PKC␦ phosphorylates and activates PTP␣ in vitro. A, 1 g of a GST⅐PTP␣cyt fusion protein was subjected to an in vitro kinase assay with the five TPA-responsive isozymes under the same conditions as in Fig. 1B, fractionated by gel electrophoresis, and stained with Coomassie prior to autoradiography. Only the incubation with PKC␦ resulted in a strong incorporation of radiolabeled phosphate in PTP␣, whereas the other PKC isozymes did not. Data shown here are representative of three experiments. B, A7r5 cells were transfected with an HA-PTP␣-construct, and PTP␣ was immunoprecipitated from unstimulated serum-starved cells using the HA-specific antibody. PTP␣ immunocomplexes were subjected to an in vitro kinase assay with the five purified recombinant PKC isozymes tested previously (see Fig. 1B) or left untreated as a control, and the phosphatase activity was assayed using pNPP as substrate. Only the incubation with PKC␦ resulted in a manifest activation of the phosphatase as indicated by the strong absorbance at 415 nm, whereas the incubation with the other PKC isozymes did not result in any increased activity compared with the basal activity. In the inserted upper panel the amount of PTP␣ protein in each reaction, as determined by immunoblot analysis, is indicated. Data shown are mean Ϯ S.E. (n ϭ 3). C, Src is a substrate for PTP␣. PTP␣ was subjected to in vitro kinase reaction with the different PKC isozymes. Subsequently, 1 g of purified recombinant Src was incubated with the PKC-phosphorylated PTP␣. Src was immunoprecipitated from the sample, and the capacity of Src to undergo autophosphorylation was determined using the pY(416)-specific antibody. As shown, only phosphorylation of PTP␣ with PKC␦ leads to increased Src activity. Results shown are representative of three experiments.
showed only a marginal increase of Src activity after TPA stimulation. These data reveal that the lack of PKC␦ results in an impaired ability of the cells to activate PTP␣, and thereby Src, upon TPA stimulation in vivo.

DISCUSSION
Several laboratories, including ours, have shown previously that TPA treatment of cells leads to the activation of Src. The molecular mechanism, however, has long been elusive. Because our initial observations showed that TPA increases Src activity, we first evaluated the effect of TPA on Src. Using specific PKC inhibitors we could easily demonstrate that this effect was transmitted via PKC, while other phorbol ester receptors like ␤-chimaerin and Ras-GRP have been excluded from being responsible for the observed effects (Ref. 10 and data not shown).
The activation of Src has been described before in detail (14). Src activation is dependent on dephosphorylation at Tyr-527.
Subsequently, Src autophosphorylates at Tyr-416 to obtain full tyrosine kinase activity (Fig. 5). These data led to the assumption that an immediate activator of Src is a phosphatase, rather than a kinase like PKC. As we show here, phosphorylation of Src with purified PKC isozymes does not result in the activation of Src. Hence, PKC must be upstream of a phosphatase and essential for the TPA-induced Src activation.
We have shown evidence that PTP␣ directly activates Src kinase after previous activation of PTP␣ only by PKC␦ but not by any other PKC isozyme tested. These experiments were performed in vitro and in vivo to substantiate our conclusions.
PTP␣ becomes activated after TPA treatment as judged by pNPP dephosphorylation. This is consistent with data published by den Hertog et al. (12), who also showed an increase of phosphatase activity of PTP␣ after TPA treatment of epidermal growth factor receptor-phosphorylated MBP. However, from these data it remained unclear whether the PKC-dependent phosphorylation activates PTP␣ to recognize and dephosphorylate a physiological substrate like Src. Here we show that PTP␣, immunoprecipitated from TPA-stimulated cells, is able to dephosphorylate and activate Src.
The suggestion that PTP␣ might be a target for PKC has also been made by Tracy et al. (11). In that report the authors showed that PTP␣ becomes phosphorylated on two serine residues (Ser-184/204) within the cytoplasmic domain of the molecule after TPA treatment of NIH3T3 cells. The same results were obtained by the use of purified PKC from rat brain and a GST fusion protein of PTP␣ in an in vitro kinase assay. In addition, it has been shown that changing these serine residues to alanine diminished the ability of PTP␣ to activate Src after mitotic stimulation in vivo (27). Therefore, these phosphorylation sites must be, at least in part, necessary for the regulation and activation of PTP␣.
Because the in vitro kinase assays by Tracy et al. (11) were carried out with only partially purified PKC, it was still unclear whether PKC is directly involved in the phosphorylation of PTP␣ or whether another kinase that is regulated by PKC is responsible for this phosphorylation. In addition, it was unclear which of the PKC isozymes might be responsible for PTP␣ phosphorylation.
We used five PKC isozymes, purified from Sf9 cells, to analyze whether PKC isozymes are directly involved in the phosphorylation and activation of PTP␣ and, if so, which PKC isozyme is responsible for the phosphorylation of PTP␣. Because we could show earlier that TPA stimulation of A7r5 cells resulted in a strong activation of Src (10), only PKC isozymes FIG. 4. Cells from PKC␦ knockout mice show decreased TPAstimulated PTP␣ and Src activity. Dermal fibroblasts from PKC␦ knockout (Ϫ/Ϫ) or matching wild-type (ϩ/ϩ) mice were serum-starved overnight and then stimulated with TPA for 15 min or left untreated, and the activity of PTP␣ and Src was determined. A, PTP␣ was immunoprecipitated from these cell lysates and analyzed for activity using pNPP as a substrate. PKC␦ ϩ/ϩ cells showed an increase of PTP␣ activity after TPA stimulation, whereas the activity of PTP␣ from Ϫ/Ϫ cells was unaffected by the TPA treatment. The inserted upper panel shows a Western blot stained for PTP␣, indicating that equal amounts of PTP␣ were assayed. Data shown are mean Ϯ S.E. (n ϭ 3). B, Src was immunoprecipitated from PKC ϩ/ϩ and PKC␦ Ϫ/Ϫ cells before and after TPA stimulation for 15 min, and the activity was determined using the Src pY(416)-specific antibody (upper panel) or in an in vitro kinase assay using acid-denatured enolase (lower panel). To demonstrate that equal amounts of Src were present, the level of Src protein was estimated by immunoblotting using the Src-specific antibody. Results shown are representative of three experiments. that are expressed in A7r5 rat smooth muscle cells (24) were used. In these assays, only PKC␦ was able to phosphorylate a GST fusion protein containing the entire cytoplasmic domain of PTP␣ (amino acids 176 -802). Similar results were obtained using PTP␣ immunoprecipitated from A7r5 cells (data not shown).
Analysis of PTP␣ protein phosphorylated by PKC␦ using a two-dimensional phosphotryptic map showed a similar pattern of phosphopeptides that had been shown before by Tracy et al. (11). From these data we conclude that the in vivo phosphorylation sites within PTP␣ that have been described before are phosphorylated by PKC␦.
We next set out to determine whether the observed phosphorylation of PTP␣ by PKC␦ resulted in increased activity of the phosphatase. As shown in Fig. 4, PKC␦ induced an increase in PTP␣ activity as measured using pNPP dephosphorylation. It also led to an increase in Src activity, as reflected in pY(416) levels of baculovirus-synthesized Src. Thus, while other kinases may contribute to the activation of PTP␣ (28), our data show that phosphorylation by PKC␦ was sufficient to increase its phosphatase activity.
Utilizing cells derived from PKC␦ knockout mice, we were able to further substantiate our findings. As we show here, PKC␦ Ϫ/Ϫ cells showed no induction of PTP␣ activity after TPA stimulation. In accordance with these data, we found only a marginal induction of Src activity after TPA stimulation in these cells, indicating that the PKC␦-mediated phosphorylation of PTP␣ is a critical moment for the activation of Src after TPA stimulation.
However, a low but yet significant induction of Src activity can consistently be observed in PKC␦ Ϫ/Ϫ cells after TPA stimulation. This finding indicates that additional, yet less prominent, mechanisms of TPA-induced Src activation exist, the nature of which remains to be identified. However, the reduced basal and induced levels of Src activity might explain why these cells have a severe defect in migration and cell cycle progression). 2 Because both the colocalization of PKC␦ and Src (9,29) and the colocalization of PTP␣ and PKC␦ (30) have been described, our data suggest a model according to which these molecules (most likely together with additional associated proteins like RACK) reside in a signaling complex, as has been described for the mitogen-activated protein kinases (reviewed in Ref. 26). According to our data, we propose a model in which the main mechanism of TPA-induced Src activation is the direct phos-phorylation and activation of PTP␣ by PKC␦, which in turn dephosphorylates and activates Src.