Insulin-induced p21ras Activation Does Not Require Protein Kinase C, but a Protein Sensitive to Phenylarsine Oxide*

Insulin treatment of fibroblasts overexpressing the insulin receptor causes a rapid accumulation of the GTP-bound form of p2lras. We have studied the involvement of protein kinase C (PKC) in, and the effect of phenylarsine oxide (PAO), a putative inhibitor of tyrosine phosphatase activity on, this process. Activa- tion of p2lras was not observed when the cells were stimulated with the phorbol ester 12-0-tetradecanoyl-phorbol-13-acetate (TPA) and pretreatment with TPA for 16 h, sufficient to down-regulate PKC activity, did not abolish p2 lras activation by insulin. These results show that PKC is not involved in the insulin-induced activation of p2lras.

The ras genes encode closely related 21-kDa proteins, p2lras, that can bind GTP and GDP and may function as signal transducing molecules (1,2). ras proteins cycle between an active (GTP-bound) and an inactive (GDP-bound) conformation (3,4). This ras-cycle is thought to be controlled by at least two different activities. First, activation of p2lras is facilitated by nucleotide exchange factors that promote the exchange of GDP for GTP (5-7). Second, p2lras can be inactivated by GTPase activating proteins (GAP,' NF-1) that stimulate the intrinsic GTPase activity of p2lras (8)(9)(10)(11)(12). Oncogenic mutations in the ras gene, found in a large number of human tumors (13), give rise to a ras protein that is constitutively in the GTP-bound form, and therefore active Using a variety of experimental approaches it has been shown that p2lras may function in several signal transduction pathways directed by membrane-associated tyrosine kinase * This study was, in part, sponsored by a grant from the Dutch Cancer Society. 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. (14)(15)(16)(17).
4 To whom correspondence should be addressed.
In T lymphocytes, p2lras activation is suggested to occur via the activation of protein kinase C (PKC), resulting in the inactivation of GAP (or NF-1) activity (25). Furthermore, proper TCR-signaling seems to require the function of two tyrosine kinases, lck (27) and fyn (28), and a tyrosine phosphatase, CD45 (29). The mechanism by which insulin triggers the activation of p2lras, however, is still elusive. Similarities between TCR and IR signaling suggest that the mechanism of p2lras activation by insulin may resemble that of TCRinduced p2lras activation. For instance, signal transduction initiated by either one of these receptors requires tyrosine kinase activity (lck and fyn for TCR, IR-@-chain for IR), and in both pathways tyrosine phosphatases are involved that are sensitive to phenylarsine oxide (CD45 for TCR (30), HA1 and HA2 for IR (31)). Also, like in TCR-signaling, in insulinmediated signal transduction PKC may play an important role as well. First, insulin can stimulate the production of diacylglycerol from various sources, resulting in the subsequent activation of PKC (32,33). Secondly, PKC-activating agents, such as 12-0-tetradecanoylphorbol-13-acetate (TPA), mimic insulin action in some cases (34)(35)(36).
In this report we have investigated the involvement of PKC in, and the effect of PA0 on insulin-induced p2lras activation. We found that in contrast to TCR-induced p2lras activation, PKC is not involved in insulin-induced activation. However, insulin-induced p2lras activation is sensitive to PAO.
[3ZP]Orthophosphate Labeling of Cell Cultures-Cells were plated on 5-cm tissue culture dishes, at least 24 h prior to labeling, to exclude residual activation of the insulin receptor caused by trypsinization. After serum starvation for 16 h, cells were labeled for 3 h in phosphate-free/serum-free medium (GIBCO) supplemented with 200 pCi of [32P]orthophosphate per 5-cm dish (24).
Determination of GTPIGDP Ratio-'*P-Labeled cells were treated with different compounds as indicated, and lysates were made using a phase-split purification as described (37). Subsequently, p2lras was immunoprecipitated with monoclonal antibody Y13-259, and GTP/ GDP nucleotides were eluted and separated by thin-layer chromatography as described (25). After autoradiography, GTP/GDP ratios were determined by counting the separated nucleotides in a scintillation counter.
80-kDa Phosphorylation-During serum starvation cells were either left untreated or treated with 100 ng/ml TPA for 16 h, and cells were labeled for an additional 3 h with [32P]orthophosphate.
Both untreated and TPA pretreated cells were either left unstimulated or stimulated with 100 ng/ml TPA for 5 min. Total protein lysates were separated on a 7.5% SDS-polyacrylamide gel. Gels were fixed and dried prior to autoradiography.

PKC Is Not Involved in Insulin-induced Activation of
p21 ras-We have previously shown a rapid accumulation of the GTP form of p2lras induced by insulin in NIH3T3 fibroblasts overexpressing the human insulin receptor (A14 cells) (24). To analyze an upstream function for PKC activation in this process, we measured the effect of TPA treatment on the activation state of p2lras. As shown in Fig. 1, a 5-min FIG. 1. The role of protein kinase C in p2lras activation. Cells were serum-starved for 16 h on 0.5% FCS, in the absence (untreated), or presence of 100 ng/ml TPA (TPA-treated). The cells were labeled with ["P]orthophosphate for 3 h, and stimulated with growth factors for 5 min. Cells were lysed and p2lras was collected by immunoprecipitation with monoclonal Y13-259. Bound nucleotides were eluted and separated by thin-layer chromatography. A14 cells: unstimulated (-), stimulated insulin for 5 min (ins, 10 pg/ml), or stimulated with TPA for 5 min (TPA, 100 ng/ml).  Fig. 1. After pretreatment with PAO, or DMSO for 5 min, cells were stimulated with insulin (5 min, 10 pg/ml). Cells were lysed and the insulin-receptor was precipitated using monoclonal antibody RPN538. The precipitates were analyzed on a 7.5% SDS-polyacrylamide gel. A14 cells: unstimulated (-) or stimulated with insulin (ins); without pretreatment, pretreated with PA0 (PAO, 25 p~) , or pretreated with DMSO. The arrow indicates the position of the insulin receptor &chain. , or with addition of PA0 for 10, 5, or 2 min prior to lysis. B, A14 cells: pretreated with PA0 for 5 min (25 PM), and stimulated with insulin (10 pg/ml) in the absence (-) or presence of DMP for 5 min prior to lysis.

GDP
TPA stimulation of A14 cells did not result in a shift of the p2lras GTP/GDP ratio. Next, we pretreated the A14 cells with 100 ng/ml TPA for a period of 16 h, to deplete the cells of PKC activity (38), and measured the p2lras GTP/GDP ratio after insulin stimulation. This TPA pretreatment did not influence the insulin-induced shift in the GTP/GDP ratio on p2lras. As a control for PKC down-modulation, we analyzed the phosphorylation of the 80-kDa MARCKS protein, a major substrate of PKC often used as a marker for PKC activity in intact cells (38)(39)(40). A 5-min stimulation with TPA was sufficient to stimulate 80-kDa phosphorylation (Fig. 2).

Insulin-induced p2lras Activation Inhibited by P A 0
Induction of 80-kDa phosphorylation by a short stimulation with TPA was abolished after pretreatment with 100 ng/ml TPA for 16 h, indicating that cells are depleted of TPAsensitive PKC activity using this protocol. From these results we conclude that PKC is not involved in the activation of p2lras by insulin. P A 0 Inhibits the Activation of p2lras by Insulin-Phenylarsine oxide has been found to interfere with signal transduction from the insulin receptor (41,42). To test its effect on the shift in the p2lras GTP/GDP ratio induced by insulin, we added PA0 5 min prior to insulin stimulation. PA0 completely blocked activation of p2lras at a concentration of 25 PM (Fig. 3). Addition of solvent alone (DMSO) had no effect on the insulin-induced activation of p2lras. To ascertain that addition of PA0 does not influence insulin-induced phosphorylation of the @-chain of the insulin receptor, thereby inhibiting p2lras activation, we immunoprecipitated the insulin receptor from 32P-labeled cell lysates. consistent with previous findings (42), PA0 treatment resulted in a small increase in phosphorylation rather than an inhibition of insulin-induced @-chain phosphorylation (Fig. 4). This suggests that PA0 inhibition does not affect the insulin receptor itself, but interacts with a protein that functions somewhere between the activated receptor and p2lras. Reversion of p2lras Activation and P A 0 Inhibition-To gain further insight in the process affected by PAO, we investigated the kinetics of inhibition. For this, we stimulated the cells 15 min prior to lysis with insulin. Full activation of p2lras (-70% GTP-bound) is achieved within 2 min, and remains at this level for at least 15 min (24). PA0 was added at different time points during a 15-min insulin stimulation, starting at 10 min prior to lysis. We did no longer observe an insulin-induced shift in the GTP/GDP ratio on p2lras when PA0 was added for 10 min, but activation was hardly affected when PA0 treatment lasted only 5 min (Fig. 5A).
Inhibition of insulin-induced responses by PA0 can be prevented by addition of DMP, a compound containing vicinal sulfhydryls, thus capable of competing for PA0 binding. We observed complete reversion of PA0 inhibition when 50 p~ DMP was added simultaneously with insulin (Fig. 5B), similar to findings with insulin-induced glucose uptake (41).

DISCUSSION
We have studied the role of PKC in the insulin-induced activation of p2lras, since PKC mediates p2lras activation in T lymphocytes (25). Depletion of PKC by prolonged TPA treatment in A14 cells did not abolish the ability of insulin to induce a shift in the p2lras GTP/GDP ratio. Such a pretreatment did result in the loss of TPA-sensitive PKC activity as judged by the ability of TPA to induce phosphorylation of the 80-kDa substrate (MARCKS protein) of PKC. Likewise, PKC activation by a short incubation with TPA did not mimic the effect of insulin on the activation state of p2lras. Although not all PKC isozymes are sensitive to TPA (43)(44)(45), the PKC isozyme that mediates activation of p2lras in T cells is activated by TPA (25). Furthermore, in NIH3T3 cells the PKC-LY isozyme seems to be the major form that is expressed, and this PKC isozyme is sensitive to TPA down-modulation (46)(47)(48). We conclude that (a TPA-sensitive) PKC is not involved in insulin-induced activation of p2lras, and that the mechanism of p2lras activation by insulin in A14 cells differs from the activation of p2lras by TCR stimulation.
The finding that PA0 can inhibit both insulin-induced, as well as TCR-induced signal transduction, prompted us to evaluate the effect of P A 0 on insulin-induced p2lras activation. A pretreatment of 5 min with P A 0 resulted in a complete inhibition of p2lras activation, although the overall insulin receptor phosphorylation remained unaffected. Addition of 2,3-dimercaptopropanol to cells preincubated with PA0 resulted in restoration of the insulin-induction of p2lras activation. Thus, PA0 seems to exhibit its inhibitory effect on p2lras activation through binding of vicinal sulfhydryls (41).
The kinetics and characteristics of inhibition by PA0 are similar to that found for the inhibition of insulin-induced glucose uptake by PA0 (42). This may indicate that the same PAO-sensitive protein is involved in the generation of both effects. Inhibition of insulin-induced glucose uptake by PA0 is thought to occur through inactivation of a tyrosine phosphatase, since several proteins, phosphorylated in response to insulin, were shown to accumulate in the presence of PA0 (49). PA0 had no effect on proteins phosphorylated in response to serum or platelet-derived growth factor, indicating that there is some specificity in the inhibiting effect of PAO.
One of the proteins that could be detected after insulin stimulation in the presence of PA0 is a 15-kDa phosphoprotein (50). This protein was shown to be the fatty acid-binding protein 422(aP2) that can be phosphorylated by the insulin receptor in vitro, in a fatty acid-dependent manner (51).
Recently, it was shown that the dephosphorylation of this protein can be mediated by two different tyrosine phosphatases, HA1 and HA2, both associated with the cell membrane (31). These two phosphatases were purified from 3T3 adipocytes, and shown to be inactive in the presence of PAO. In T lymphocytes PA0 inhibits CD45 phosphatase activity, having no effect on the kinase activity of both lck and fyn, again indicating that PA0 may be a specific tyrosine phosphatase inhibitor (30). However, PA0 binds to vicinal sulfhydryl groups (52) and may interact with other proteins as well, so clearly a more detailed analysis is needed before a definite conclusion concerning the function of the PAO-sensitive component in this mechanism can be drawn. Accumulation of the GTP-bound form of p2lras in A14 cells is maximal within 2 min after insulin addition, and remains at this elevated level (-70% GTP bound) for at least 15 min (24). Thus we were able to study whether PA0 could not only prevent, but also reverse the activation of p2lras induced by insulin. When cells were stimulated with insulin for 15 min and PA0 was present during the last 10 min, we observed complete reversion of insulin-induced p2lras activation, whereas addition of PA0 5 min prior to lysis had almost no effect. The implications of these results are severalfold. First, the fact that we indeed observe reversion points to a mechanism in which constitutive activation of upstream elements is necessary for sustained p2lras activation. Secondly, while a 5-min treatment with P A 0 is still without effect on the insulin stimulation, 10 min after addition of PA0 all p2lras is converted to the GDP form. Thus, complete inactivation takes place in 5 min. The fact that we find no effect of PA0 during the first 5 min of treatment is probably due to the time P A 0 needs to cross the cellular membrane (53). Since hydrolysis of GTP bound to p2lras in vitro is rather slow (54,55), this implies that the activity of a GTPase activating protein (GAP or NF-1) is considerable in these cells, at least in the absence of an insulin stimulus.
Thus far, the nature of the PAO-sensitive protein is still unknown. It could be one of the tyrosine phosphatases as discussed above. This would imply that tyrosine dephosphorylation is an essential step in insulin induction of p2lras activation, alternatively, PA0 might inhibit another protein activity. In both cases, PA0 can be used as a tool for deciphering the pathway between insulin stimulation and p2lras activation.