Insulin Activates the Raf-1 Protein Kinase*

Several growth factors and mitogens have been shown to activate the proto-oncogene product Raf-1 protein kinase in murine fibroblasts, apparently through a direct agonist-stimulated tyrosine phosphorylation of the Raf-1 protein. We investigated the possibility that insulin could also activate the Raf-1 kinase, since its receptor also contains an intrinsic insulin-activated protein tyrosine kinase activity. In several cell lines expressing relatively large numbers of insulin receptors, insulin rapidly stimulated the phosphorylation of immunoreactive Raf-1 protein. In H35 cells, a line of well differentiated rat hepatoma cells, the effect of insulin was maximal by 6 min and at 7 nM insulin and occurred normally in cells virtually completely depleted of protein kinase C activity. The insulin-stimulated increase in Raf-1 protein phosphorylation occurred concurrently with a 3-fold increase in Raf-1 protein kinase activity. However, phosphoamino acid analysis showed that only phosphoserine and a trace of phosphothreonine were present in the Raf-1 protein after insulin stimulation of the cells. This was true even when investigated at shorter times (4 min) after insulin stimulation and despite the use of phosphotyrosine phosphatase inhibitors. We conclude that insulin can rapidly activate the Raf-1 kinase in some insulin-sensitive cell types but that this activation probably occurs through a mechanism distinct from direct phosphorylation of the Raf-1 protein by the insulin receptor protein tyrosine kinase.

acid analysis showed that only phosphoserine and a trace of phosphothreonine were present in the Raf-1 protein after insulin stimulation of the cells. This was true even when investigated at shorter times (4 min) after insulin stimulation and despite the use of phosphotyrosine phosphatase inhibitors.
We conclude that insulin can rapidly activate the Raf-1 kinase in some insulin-sensitive cell types but that this activation probably occurs through a mechanism distinct from direct phosphorylation of the Raf-1 protein by the insulin receptor protein tyrosine kinase. is the Ruf-1 proto-oncogene product. The Raf-1 proto-oncogene is the cellular homologue of v-ruf, the transforming oncogene of murine sarcoma virus 3611 (1,2). Morrison et al. (3,4) reported recently that the activity of the Raf-1 protein kinase could be stimulated by exposure of cells to plateletderived growth factor (PDGF),' acidic fibroblast growth factor, epidermal growth factor, the phorbol ester phorbol 12myristate 13-acetate, and by several types of oncogenic transformation of murine 3T3 cells. Insulin had no effect in these studies. Moreover, they reported that the activation of Raf-1 that occurred in response to PDGF was accompanied by phosphorylation of the protein on tyrosine residues. Finally, they demonstrated direct physical association of the Raf-1 protein with the PDGF receptor in intact cells and that tyrosine phosphorylation of the Raf-1 kinase by the PDGF receptor kinase in vitro led to a marked increase in Raf-1 kinase activity (4). Thus, Raf-1 was proposed as a protein kinase whose activation in cells was a direct result of phosphorylation on tyrosine residues by the PDGF receptor kinase (3-5).
Although Morrison et al. (3) found no activation or phosphorylation of Raf-1 in response to insulin, we evaluated this possibility in cells that express many more insulin receptors than the 3T3 cells used in their study. This approach had led to the discovery of several insulin effects that are not found in 3T3 fibroblasts (6)(7)(8). The studies described here demonstrate that insulin can rapidly activate the Raf-1 kinase in various insulin-responsive cell types; however, we could find no evidence that this activation was accompanied by tyrosine phosphorylation of the Raf-1 protein.

MATERIALS AND METHODS
Cells-The cell lines used in this study were: NIH EC,, an NIH-3T3 cell derivative that overexpresses Raf-1 (9); HIRC-B, a rat-l cell derivative that overexpresses the normal human insulin receptor (10); H35, a well differentiated rat hepatoma cell line (11, 12); and HIR 3.5, an NIH-3T3 cell derivative that overexpresses normal human insulin receptors (13). Cells were grown to confluence, changed to serum-free medium for 16 h, and incubated with 0.1 mCi/ml "P for 2 h as described (8,14 Phosphoamino Acid Analysis-Serum-deprived cells were labeled for 3 h with approximately 1 mCi/ml :r'P and then exposed to control conditions or 70 nM insulin for 10 min. The cells were then processed for immunoprecipitation, electrophoresis, and autoradiography as described above. The radiolabeled Raf-I protein was eluted from the gel and processed for phosphoamino acid analysis as described (18).

AND DISCUSSION
We first evaluated the ability of insulin to stimulate the phosphorylation of the Raf-1 protein in several insulin-sensitive cell lines. Because of a relatively high level of apparent basal phosphorylation of the Raf-1 protein, stimulated phosphorylation has been demonstrated in previous studies by a characteristic decrease in Raf-1 migration in SDS-polyacrylamide gels, accompanied by a modest increase in net phosphorylation.
In three different insulin-sensitive cell lines, insulin (70 nM for 10 min) stimulated the phosphorylation of Raf-1 (Fig. 1). It also appeared to stimulate the phosphorylation of several other proteins in the immunoprecipitate, some of which have previously been noted and described as potential substrates for the Raf-1 kinase (19). Because of the apparent relative abundance of "'P-labeled Raf-1 in the H35 rat hepatoma cells, this cell line (11,12) was chosen for further study.
Insulin-stimulated phosphorylation of the Raf-1 protein in H35 cells was evident as early as 4 min after addition of insulin to the cells and reached a maximum by 6-10 min (data not shown). This time course was somewhat slower than that observed for insulin-stimulated tyrosine phosphorylation of several proteins in a different insulin-sensitive cell line (20). Maximal activation of the response occurred at 7 nM insulin (not shown). This dose-response relationship parallels that for insulin induction of c-myc and insulin-stimulated thymidine incorporation in H35 cells (11, 12) and insulin receptor occupancy in a closely related cell line (21). This concentration of insulin does not result in significant binding to insulinlike growth factor II receptors in these cells (22) receptors (22). Thus, the insulin effect appears to be the consequence of insulin binding to insulin receptors in these cells. Although activation of protein kinase C can lead to phosphorylation of the Raf-1 protein (3, 4), insulin appeared to stimulate Raf-1 phosphorylation normally in cells preincubated for 16 h with 16 pM phorbol 12-myristate 13-acetate (data not shown). This treatment has been shown to deplete these cells of more than 93% of cytosolic and 96% of particulate protein kinase C activity (8).
An assessed by measuring Raf-1 kinase activity in immunoprecipitates prepared from control and insulin-treated cells, insulin treatment (70 nM for 10 min) resulted in an approximately 3-fold increase in Raf-1 kinase activity directed toward a synthetic Raf-1 peptide as an exogenous substrate (Fig. 2). We conclude from these studies that insulin can activate the Raf-1 kinase in these well differentiated rat hepatoma cells and in other insulin-sensitive cell lines. We next attempted to determine whether the insulin effect occurred as the result of tyrosine phosphorylation of the Raf-1 protein. Phosphoamino acid analysis of the Raf-1 protein from cells exposed to either control conditions or insulin (70 nM for 15 min) showed that phosphoserine was by far the predominant phosphoamino acid present, with a trace of phosphothreonine being detected after long exposure of the autoradiograph (Fig. 3). No phosphotyrosine was detected, even after much longer autoradiographic exposure. In other experiments, we pretreated the cells with phenylarsine oxide (PAO), an inhibitor of protein tyrosine phosphatases that has been used to uncover cryptic insulin-stimulated tyrosine phosphorylation in cells (20,23 Serum-deprived H35 cells were labeled for 3 h with approximately 1 mCi/ml 'rrP and then exposed to control conditions or 70 nM insulin for 10 min. The cells were then processed for immunoprecipitation, electrophoresis, and autoradiography (A) as described in the legend for Fig. 1 were detected (not shown). We conclude from these studies that insulin activates the Raf-1 kinase in these cells by a mechanism that results in an increase in serine phosphorylation without a significant increase in tyrosine phosphorylation.
There are several potential mechanisms for this effect. One possibility is that insulin activates an intermediate protein serine/threonine Raf-1 protein kinase kinase. Known candidates for this kinase include the insulin-activated S6 (24-26) and MAP2 (27-29) kinases. However, it is unlikely that this putative intermediate kinase is protein kinase C, since insulin-stimulated phosphorylation of the Raf-1 protein occurred normally in cells virtually completely depleted of protein kinase C. Another potential mechanism would involve an insulin-generated second messenger that could bind to and activate the Raf-1 kinase in a manner analogous to diacylglycerol or cyclic AMP. Numerous studies have proposed the existence of such an insulin "mediator" (30-32), but details of its structure are not available.
In support of such a mechanism is the fact that the Raf-1 kinase and protein kinase C are both protein kinases with hydrophobic "zinc finger"-like domains (9). This sequence motif appears to be responsible for binding phorbol esters and diacylglycerols in the case of protein kinase C (33); the analogous domain in Raf-1 could conceivably function in the same manner to bind a still undiscovered ligand. This type of allosteric activation would result in an increase in Raf-1 serine phosphorylation upon activation that is presumably the result of autophosphorylation.
Whatever its mechanism, activation of the phosphotransferase activity of Raf-1 by insulin could have important implications for the mechanism of insulin action. Activated Raf-1 could phosphorylate and alter the activities of enzymes and proteins involved in insulin-modulated processes. For example, insulin induces the rapid transcription of the c-fos protooncogene by utilizing the c-fos promoter serum response element (6,7); activated Raf-1 also appears to stimulate c-fos transcription through the serum response element (34), thus making Raf-1 a potential intermediate in this insulin-stimulated process. Another possibility is that the Raf-1 kinase is responsible for insulin-stimulated serine phosphorylation of .the insulin receptor, perhaps altering the behavior of the receptor in some way. Finally, it could be involved in certain insulin-resistant states in which the abnormality is thought to reside at a "post-receptor" locus.