12-O-Tetradecanoylphorbol-13acetate Activates the Ras/ Extracellular Signal-regulated Kinase (ERK) Signaling Pathway Upstream of SOS Involving Serine Phosphorylation of Shc in NIH3T3 Cells*

We investigated the activation of the Ras/ERK signaling pathway by 12-O-tetradecanoylphorbol-13-acetate (TPA) in NIH3T3 fibroblasts. Interestingly, the activation was suppressed not only by dominant negative Raf-1 but also by dominant negative Ras and SOS. Further analysis revealed that TPA treatment induced, dependently on protein kinase C, the mobility shift of p66 shc in SDS-polyacrylamide gel electrophoresis, which could be prevented by treatment of the Shc immunoprecipitate with serine/threonine-specific protein phosphatase 1 (PP1) or 2A (PP2A). Phosphoamino acid analysis of Shc showed that unlike growth factor-induced Shc phosphorylation, where Shc is mainly phosphorylated at tyrosine residues, TPA-induced phosphorylation was only at serine residues. Like growth factor-induced Shc phosphorylation, which leads to the association of Shc with Grb2, TPA also induced this association, but, correspondingly to the above results, the TPA-induced association was disrupted by in vitro treatment of the Shc immunoprecipitate with PP1. Taken together, these results suggest that the TPA signal was fed at or upstream of Shc to activate the Ras/ERK signaling pathway involving serine phosphorylation of Shc.

Ligand-activated growth factor receptors induce translocation of the Grb2-SOS complex to the plasma membrane where p21 ras is localized by two alternative mechanisms: directly by recruiting Grb2 via its SH2 domain (1) or indirectly by recruiting Shc via its SH2 or PTB domain (2), the subsequent tyrosine phosphorylation of which leads to its association with Grb2 via its SH2 domain (3). The high affinity binding of Grb2 to Shc proteins requires phosphorylation of Shc at Tyr 317 , which lies within the high affinity binding motif for the Grb2 SH2 domain (4). 12-O-Tetradecanoylphorbol-13-acetate (TPA) 1 is one of the compounds widely used to probe various cellular activities. Its only known function in the cell is to activate protein kinase C (PKC), mimicking the physiological lipid metabolite diacylglycerol (5). Numerous studies have indicated that TPA activates the Ras/ERK signaling pathway (6 -8), but it is still controversial as for the entry site of the TPA signal in the pathway. In NIH3T3 fibroblasts protein kinase C phosphorylates Raf-1 and stimulates its kinase activity (9); however Raf-1 phosphorylation by PKC in vitro does not lead to MEK activation (10). It has also been reported that the phosphorylation of Raf-1 is irrelevant to signal transduction (11). We analyzed TPA-induced Ras/ERK signaling in the context of ERK-mediated gene regulation in NIH3T3 cells and found that the TPA signal is fed upstream of SOS and involves serine phosphorylation of Shc.

MATERIALS AND METHODS
Plasmids-The reporter plasmid pGL2 muPA-8.2 was constructed by inserting the murine urokinase-type plasminogen activator (uPA) gene promoter (from Ϫ8, 200 to ϩ398 with respect to the transcription initiation site) upstream of the luciferase-coding region of the promoterless plasmid pGL2-basic (Promega). Expression vectors for various signaling molecules used in this work were previously described (12,13) except that for constitutively active pPKC␦ (14).
Transient Transfection and Analysis of Reporter Gene Expression-NIH3T3 cells (0.1 ϫ 10 6 /well) were plated in 6-well (35-mm) tissue culture plates with 2 ml DMEM containing 10% CS and transfected 20 h later by the calcium phosphate precipitation method (Pharmacia Biotech Inc.). Luciferase expression was determined as described (13).
Transient Transfection of p44 mapk -Tag and Determination of ERK-1 Kinase Activity-NIH3T3 cells (1 ϫ 10 6 ) were plated in 10-cm dishes and transfected 20 h later using 60 l of LipofectAMINE TM (Life Technologies, Inc.) with 15 g of pcDNA-p44-tag encoding HA-tagged ERK-1 together with 15 g of the coexpressed plasmid for 5 h. The cells were incubated in fresh DMEM with 10% CS for 5 h and in DMEM with 0.1% CS for 12 h. The cells were then treated with FGF-2 or TPA for 10 min and washed with phosphate-buffered saline. Whole cell extracts were immunoprecipitated with anti-HA-tag mouse monoclonal antibody (12CA5), and ERK activity was determined as described (15).
Immunoprecipitation and Western Blot Analysis-NIH3T3 cells stimulated for 10 min by TPA, FGF-2, or PDGF were immunoprecipitated as described (16) using a polyclonal anti-Shc antibody (Transduction Laboratories). The immunoprecipitates were analyzed by Western blots using a monoclonal anti-Shc antibody (1:250; Transduction Laboratories) or a monoclonal anti-Grb2 antibody (1:500; Transduction Laboratories). An enhanced chemiluminescence detection method (Amersham) was employed, and the membrane was exposed to Kodak X-Omat AR film.
Phosphoamino Acid Analysis-NIH3T3 cells were grown to confluency on a 15-cm dish, starved for 16 h in phosphate-free DMEM containing 0.1% dialyzed calf serum, incubated for 4 h in the same medium with the addition of 2 mCi of [ 32 P]orthophosphate, and then induced with 100 ng/ml TPA or 10 ng/ml FGF-2 for 10 min. Cell extracts were immunoprecipitated using polyclonal anti-Shc antibody, and the precipitates were fractionated by SDS-polyacrylamide gel electrophoresis. Each Shc isoform was recovered separately from the gel after autoradiography and subjected to phosphoamino acid analysis as described (17). * This work was supported in part by the Gottlieb Daimler-and Karl Benz-Stiftung Fellowship 2.91.07 (to D. B.). 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. ‡ The first two authors contributed to the work equally. § Present address: The Rockefeller University, Dept. of Molecular Oncology, 1230 York Ave., New York, NY 10021.

RESULTS AND DISCUSSION
Transcriptional Activation by TPA but Not by PKC␦ Is Dependent on SOS-In NIH3T3 cells, the uPA gene is activated by TPA and FGF-2 via the Ras/ERK signaling pathway (12). To know where in the pathway the TPA signal is fed, we first examined the effects of various signaling molecules on uPA gene induction in transient transfection assays. The induction of the uPA promoter by TPA and FGF-2 was suppressed by dominant negative mutants of Ras (Ras17N) or Raf-1 (N⌬Raf-1) or by mitogen-activated protein kinase-specific protein phosphatase MKP-1 (Fig. 1A), confirming the involvement of the Ras/ERK pathway in uPA gene activation by TPA and FGF-2. Surprisingly, the induction by TPA was also suppressed by dominant negative mutants of SOS (⌬mSOS1, Fig. 1B), showing that TPA activates the Ras/ERK signaling pathway at or upstream of SOS. Wild-type SOS (mSOS1) enhanced both the basal and induced activities of the uPA promoter (Fig. 1B). SOS is a GTP-GDP exchange factor which is able to activate Ras (18). Accordingly, the induction of the uPA gene by constitutively active Ras or Raf was not suppressed by coexpression of ⌬mSOS1 (data not shown), indicating that inhibitory effects of dominant SOS are not general on gene expression or protein synthesis.
A constitutively active mutant of PKC␦ (PKC␦ RA) strongly activated the uPA promoter in the transient transfection assay, but the activation was not blocked by ⌬mSOS1 (Fig. 1C).
N⌬Raf-1 and wild-type MKP-1 were able to block this induction. These results show that PKC␦ RA activates the pathway at a step downstream of SOS.
It has been shown that PKC␣ phosphorylates Raf-1 at the same sites in a cell-free system which are phosphorylated in vivo (9). Accordingly, overexpressed PKC␦ may activate Raf-1 directly in the cytoplasm without Ras-dependent translocation of Raf-1 to the membrane (19), possibly because of an increase in the number of PKC␦ molecules in the transfection assays described here. In the case of TPA induction, which activates several PKC isoforms including ␣ and ␦ (20), the absolute number of PKC molecules is not elevated and, therefore, Rasdependent translocation of Raf-1 to the membrane may still be needed for the induction.
Dominant Negative SOS and Ras and a Ras Inhibitor Block ERK-1 Kinase Activation by TPA and FGF-2-Both ⌬mSOS1 and Ras17N strongly suppressed not only the uPA promoter activity but also the ERK-1 activity induced by TPA and FGF-2 ( Fig. 2A). mSOS1 showed no inhibitory effect on ERK-1 activity, and constitutively active Ras (Ras61L) alone activated ERK-1 as strongly as TPA or FGF-2. The levels of tagged-ERK-1 expression were similar in different transfections (Fig.  2B). Pretreatment of the cells for 18 h with an inhibitor of farnesylation, an obligatory step in Ras processing (21), blocked TPA and FGF-2 stimulation of ERK-1 activity about 60% in both inductions (Fig. 2C), proving again that the activation of Ras is necessary for activation of the ERK pathway by TPA and FGF-2. membrane is induced either by Shc or by a receptor tyrosine kinase. Shc activation involves its phosphorylation at a tyrosine residue and subsequently interaction with Grb2 (22). Accordingly, both FGF-2 and PDGF, but not TPA, induced tyrosine phosphorylation of all three Shc isoforms (Fig. 3A). Interestingly, all three stimuli, TPA, FGF-2, and PDGF, led to a mobility shift of the p66 shc isoform (Fig. 3B), suggesting an alternative modification on the Shc protein other than tyrosine phosphorylation after induction with these agents.
TPA and Growth Factors Lead to Serine Phosphorylation of the p66 shc Isoform-To analyze the mobility shift of the p66 shc isoform, Shc was immunoprecipitated from cell extracts after stimulation of the cells with FGF-2 or TPA. The immunoprecipitates were treated with either calf intestine alkaline phosphatase (CIP), tyrosine-specific phosphatase LAR (23), or serine/threonine-specific protein phosphatase 2A (24) (PP2A). Treatment with either CIP or PP2A, but not LAR, prevented the mobility shift of p66 shc (Fig. 3C), suggesting that the shift is due to serine/threonine phosphorylation. TPA-induced mobility shift of p66 shc was also sensitive to serine/threoninespecific protein phosphatase 1 (PP1) (data not shown). The results in Fig. 3D show that the mobility shift of p66 shc induced by TPA is dependent on protein kinase C; treatment of cells with the protein kinase C inhibitor bisindolylmaleimide (25) or TPA for 24 h down-regulating PKC prior to TPA treatment ablated the shift. The shift induced by FGF-2 was partially suppressed by both treatments, suggesting that serine/threonine phosphorylation of Shc after FGF-2 treatment also involves PKC isoforms, at least partially. Phosphoamino acid analysis revealed that TPA and FGF-2 induced phosphorylation of serine residues in p52 shc and p66 shc (Fig. 3E). Appar-

FIG. 3. Modification of Shc.
A, tyrosine phosphorylation of Shc. Cells were induced by treatment for 10 min with 100 ng/ml TPA, 10 ng/ml FGF-2, or 30 ng/ml PDGF and immunoprecipitated with the Shc antibody. Immunoblotting was performed using a mouse monoclonal antiphosphotyrosine antibody. B, retardation of the p66 shc isoform in SDS-gel electrophoresis. The same blot as in A was washed and reprobed using a mouse monoclonal anti-Shc antibody. C, effects of various phosphatases on the mobility shift of Shc. Shc immunoprecipitates were treated with either calf intestine alkaline phosphatase (Boehringer Mannheim) (10 units/sample) at 37°C or specific tyrosine phosphatase LAR (New England Biolabs) (10 units/sample) or PP2A (provided by Brian Hemmings) (10 nM) at 30°C for 2 h, then immunoblotted with a mouse monoclonal anti-Shc antibody. D, effects of PKC inactivation on the TPA-and FGF-2-induced p66 shc mobility shift. Cells were pretreated with 500 nM PKC inhibitor (bisindolylmaleimide; Calbiochem) for 1 h or 100 ng/ml TPA for 24 h to down-regulate PKC before stimulation with TPA or FGF-2 for 10 min. Shc immunoprecipitates were analyzed by Western blotting using a mouse monoclonal anti-Shc antibody. E, phosphoamino acid analysis. Cells were induced with TPA or FGF-2 in the presence of radioactive orthophosphate. Shc immunoprecipitates were fractionated, and each isoform was separately analyzed as described under "Materials and Methods." Serine phosphorylation of p52 shc was stimulated 5-and 1.5-fold, while that of p66 shc was stimulated 3-and 2-fold by TPA and FGF-2, respectively.
FIG. 4. TPA-induced Shc association with Grb2. A, cells were induced with 100 ng/ml TPA or 10 ng/ml FGF-2 for the different times indicated, and cell extracts were immunoprecipitated using a polyclonal anti-Shc antibody and analyzed by Western blotting using a rabbit polyclonal anti-Shc antibody. The same blot was reprobed with either anti-phosphotyrosine antibody (4G10) or mouse monoclonal anti-Grb2 antibody. B, Shc immunoprecipitates from control (Ϫ), TPA-induced, or FGF-2-induced cells were treated at 30°C for 30 min with PP1 (Calbiochem; 5 units), PP1 together with calyculin A (50 nM), or calyculin A alone and analyzed by Western blotting using mouse monoclonal anti-Grb2 antibody. ently, serine phosphorylation of p52 shc does not affect its mobility in SDS-polyacrylamide gel electrophoresis under the conditions used. In accordance with the result shown in Fig.  3A, phosphorylation of tyrosine residues was induced by FGF-2 but not by TPA.
TPA Leads to Increased Association of Shc with Grb2-We were interested in whether the TPA-induced serine phosphorylation of Shc also promotes its association with Grb2. After immunoprecipitation of Shc from either TPA-or FGF-2-treated or untreated cells we could detect an increase in Grb2 association with Shc after 2 min of either treatment, which remained constant at least up to 30 min. The association between Shc and Grb2 shows a similar duration as the modification of the p66 shc . In accordance with the results described in Fig. 3, tyrosine phosphorylation of Shc was increased only after FGF-2 treatment (Fig. 4A). When Shc immunoprecipitates were treated with PP1, Grb2 association induced by TPA or FGF-2 was suppressed or strongly reduced, respectively (Fig. 4B). The PP1-specific inhibitor calyculin A suppressed the inhibitory effect of PP1.
It has been well established that the major modification of Shc is the phosphorylation of Tyr 317 , which is induced by many receptor tyrosine kinases and cytoplasmic tyrosine kinases (Ref. 26 and references cited therein). Tyrosine-phosphorylated Shc recruits the Grb2-SOS complex to the membrane via the SH2 domain of Grb2, thus allowing SOS to activate Ras. The present results suggest that Shc can also recruit Grb2 through serine phosphorylation. An increase in serine phosphorylation of Shc as an G␤␥-induced event (27) or in response to epidermal growth factor stimulation (28) has been observed, but its biological significance has not been addressed. It has been shown recently that serine phosphorylation of Raf-1 is involved its interaction with 14-3-3 protein (29) and that serine phosphorylation of Raf-1 is also involved in the recruitment of the Fyn SH2 domain (30), although it is not known whether this transduces signaling. Taken together, our results suggest that TPA can activate Ras/ERK signaling by inducing serine phosphorylation of Shc. It remains to be seen whether PKC phosphorylates Shc directly or indirectly and how the serine phosphorylation of Shc is able to promote the interaction of Shc with Grb2.