Desensitization of Ras activation by a feedback disassociation of the SOS-Grb2 complex.

Activation of Ras by the exchange of bound GDP for GTP is predominantly catalyzed by the guanylnucleotide exchange factor SOS. Receptor tyrosine kinases increase Ras-GTP loading by targeting SOS to the plasma membrane location of Ras through the small adaptor protein Grb2. However, despite the continuous stimulation of receptor tyrosine kinase activity, Ras activation is transient and, in the case of insulin, begins returning to the GDP-bound state within 5 min. We report here that the cascade of serine kinases activated directly by Ras results in a mitogen-activated protein kinase kinase (MEK)-dependent phosphorylation of SOS and subsequent disassociation of the Grb2-SOS complex, thereby interrupting the ability of SOS to catalyze nucleotide exchange on Ras. These data demonstrate a molecular feedback mechanism accounting for the desensitization of Ras-GTP loading following insulin stimulation.


Activation of Ras by the exchange of bound GDP for GTP is predominantly catalyzed by the guanylnucleotide exchange factor SOS. Receptor tyrosine kinases
increase Ras-GTP loading by targeting SOS to the plasma membrane location of Ras through the small adaptor protein Grb2. However, despite the continuous stimulation of receptor tyrosine kinase activity, Ras activation is transient and, in the case of insulin, begins returning to the GDP-bound state within 5 min. We report here that the cascade of serine kinases activated directly by Ras results in a mitogen-activated protein kinase kinase (MEK)-dependent phosphorylation of SOS and subsequent disassociation of the Grb2-SOS complex, thereby interrupting the ability of SOS to catalyze nucleotide exchange on Ras. These data demonstrate a molecular feedback mechanism accounting for the desensitization of Ras-GTP loading following insulin stimulation.
Previous studies have demonstrated that insulin stimulation of the insulin receptor tyrosine kinase results in Ras activation and subsequent downstream stimulation of the Raf/MEK/ ERK 1 pathway (1)(2)(3). The activation of Ras occurs predominantly through the tyrosine phosphorylation of Shc followed by the association with the Grb2-SOS complex (4,5). However, Ras activation is transient and rapidly returns to the inactive state despite continuous activation of the insulin receptor tyrosine kinase and prolonged Shc tyrosine phosphorylation (6,7). Since insulin does not affect Ras-GTPase activating protein activity and/or targeting (8,9), the mechanism responsible for the desensitization of Ras has remained obscure.
Recently it has been reported that stimulation of several cell types with growth factors and other mitogenic agents results in the serine/threonine phosphorylation of SOS (10,11). In addition, SOS phosphorylation precedes an insulin-dependent disassociation of the Grb2-SOS complex (7,12). The insulin time dependence of the SOS phosphorylation and uncoupling of Grb2 from SOS was consistent with the desensitization phase of Ras inactivation. To determine whether the ERK pathway is involved in this event, we used two independent approaches to inhibit MEK activity and, hence, ERK activation. In this study we demonstrate that prevention of insulin-stimulated SOS phosphorylation and subsequent disassociation of the Grb2-SOS complex results in a prolongation of Ras activation.

EXPERIMENTAL PROCEDURES
Cell Culture-Chinese hamster ovary cells expressing the human insulin receptor (CHO/IR) and 3T3L1 adipocytes were isolated and cultured as described previously (7). Cells were incubated for 16 h in serum-free media and then pretreated for 1 h with vehicle (0.5% dimethyl sulfoxide) or 100 M PD98059. The cells were then incubated with and without 100 nM insulin for various times, followed by lysis in 50 mM Hepes, pH 7.8, 1% Triton X-100, 2.5 mM EDTA, 100 mM sodium fluoride, 10 mM sodium pyrophosphate, 2 mM sodium vanadate, 2 M pepstatin, 0.5 trypsin inhibitory unit of aprotinin, 1 mM phenylmethylsulfonyl fluoride, and 10 M leupeptin.
Immunoprecipitation and Immunoblotting-Grb2 was immunoprecipitated from the whole cell lysates by incubation with a Grb2 polyclonal antibody (Santa Cruz Biotechnology) for 2 h at 4°C. The resultant immune complexes were precipitated by incubation with protein A-agarose for 1 h at 4°C. The pellets were washed three times with Tris-buffered saline (20 mM Tris, pH 7.6, 150 mM NaCl), resuspended in SDS-sample buffer (125 mM Tris-HCl, pH 6.8, 20% (v/v) glycerol, 4% (w/v) SDS, 100 mM dithiothreitol, 0.1% (w/v) bromphenol blue) and heated at 100°C for 5 min. Whole cell lysates or immunoprecipitates were separated on reducing 5-10% SDS-polyacrylamide gradient gels and transferred to polyvinylidene difluoride membranes using 1 A for 2 h at 4°C. Immunoblotting of the whole cell lysates or Grb2 immunoprecipitates was performed using an ERK polyclonal antibody (Zymed), a pp90 rsk polyclonal antibody, or a Raf polyclonal antibody (Santa Cruz Biotechnology) and a SOS polyclonal antibody (Upstate Biotechnology Inc).
Quantitative Transient Transfection by Electroporation-We have recently demonstrated that electroporation can be used to express various cDNAs in CHO/IR with 85-100% transfection efficiency (13). Briefly, CHO/IR cells were electroporated with a total of 40 g of the dominant-interfering MEK mutant (MEK/K97R) or the empty parent vector (CMV5) at 340 V and 960 microfarads in 500 l of phosphatebuffered saline. Following electroporation, the cells were plated in ␣-minimal essential medium containing 10% serum. Cell debris was removed by replacing media with fresh media 12 h later.
Determination of GTP-bound Ras-CHO/IR cells were incubated in serum-and phosphate-free media for 2 h, followed by the addition of 0.2 mCi/ml carrier-free 32 P for 3 h. The cells were pretreated with either vehicle (0.5% dimethyl sulfoxide) or 100 M PD98059 during the last hour of the labeling period. Cells were then either left untreated or stimulated with 100 nM insulin for the indicated times. Cells were solubilized in 50 mM Hepes, 1 mM sodium phosphate, pH 7.4, 1% Triton X-100, 100 mM NaCl, 20 mM MgCl 2 , 1 mg/ml bovine serum albumin, 0.1 mM GTP, 0.1 mM GDP, 1 mM ATP, 0.4 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 10 g/ml leupeptin, 10 g/ml soybean trypsin inhibitor, and 10 mM benzamide. The extract was immunoprecipitated with the Ras antibody (Y13-259, Oncogene Science) for 60 min, and the immune complexes were washed 5 times with lysis buffer and 5 times with wash buffer (50 mM Hepes, pH 7.4, 20 mM MgCl 2 , 150 mM NaCl, and 0.005% SDS). Ras-associated guanylnucleotides were eluted in 20 l of 2 mM EDTA, pH 8.0, 2 mM dithiothreitol, 0.2% SDS, 0.5 mM GTP, and 0.5 mM GDP) at 65°C for 20 min. Eluted GDP and GTP were * This study was supported in part by grants from the National Institutes of Health (to K.-L. G., G. A. K., and J. E. P.). 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. § These two authors contributed equally to this study. ** Established Investigator of the American Heart Association. ‡ ‡ To whom correspondence should be addressed: Dept. of Physiology & Biophysics, The University of Iowa, Iowa City, IA 52242-1109. 1 The abbreviations used are: MEK, mitogen-activated and extracellular signal-regulated kinase kinase; MEK/K97R, dominant-interfering MEK mutant in which lysine 97 was replaced with arginine; ERK, extracellular signal-regulated kinase; SOS, Son-of-Sevenless; CHO/IR, Chinese hamster ovary cells expressing the human insulin receptor. separated on polyethyleneimine cellulose plates (Baker) by thin layer chromatography using 1 M KH 2 PO 4 (pH 3.4) as the solvent. Labeled nucleotides were visualized by autoradiography and counted using an AMBIS ␤ detector.

RESULTS AND DISCUSSION
To investigate the role of the ERK pathway mediating SOS phosphorylation and disassociation of the Grb2-SOS complex, we initially took advantage of the recently identified specific inhibitor of MEK activity, PD98059 (14,15). This reagent is noncompetitive for ATP and does not affect the enzyme activities of over 30 tyrosine and serine/threonine kinases examined, including the highly related Jun kinase kinase, JNKK. Since MEK is the immediate upstream kinase responsible for ERK phosphorylation, we examined the effect of PD98059 on insulin-stimulated ERK phosphorylation (Fig. 1a). Insulin treatment of Chinese hamster ovary cells expressing high levels of the insulin receptor (CHO/IR) resulted in a timedependent phosphorylation of ERK1 and ERK2 characterized by decreased electrophoretic mobility (1). The phosphorylation of ERK was transient, with a maximal decrease in mobility following 5 min of insulin stimulation and a return to the basal state by 30 min (Fig. 1a, left, lanes 1-6). Pretreatment of cells with PD98059 prior to insulin stimulation markedly reduced the extent of ERK phosphorylation, as indicated by the nearcomplete absence of the slower migrating isoforms of ERK (Fig.  1a, right, lanes 1-6).
MEK is activated by serine phosphorylation catalyzed by members of the Raf family of protein kinases (16 -18). The Raf kinase lies in a feedback pathway, in which Raf undergoes phosphorylation secondary to ERK activation (19,20). Thus, insulin also stimulated a characteristic time-dependent gel shift of Raf which was initially detected at 5 min and was maximal by 10 min (Fig. 1b, left, lanes 1-6). In contrast to ERK, the decrease in Raf electrophoretic mobility persisted over the 30-min time period examined. However, pretreatment with the MEK inhibitor prevented the insulin-stimulated Raf gel shift (Fig. 1b, right, lanes 1-6) but did not block Raf kinase activity (data not shown). These data further demonstrate that Raf phosphorylation occurs by a MEK-dependent feedback pathway.
Previous studies have demonstrated that SOS is phosphorylated on serine/threonine residues following growth factor activation of the Ras/Raf/MEK/ERK pathway (11). Consistent with this, insulin stimulated a time-dependent reduction in the electrophoretic mobility of SOS (Fig. 1c). The decrease in SOS mobility was detected following 5 min of insulin treatment with a maximal effect reached by 10 min (Fig. 1c, left, lanes 1-6). Pretreatment of the cells with PD98059 also prevented the insulin-stimulated SOS gel shift (Fig. 1c, right, lanes 1-6), consistent with a role of MEK in the cascade leading to the phosphorylation of SOS.
In addition to the stimulation of SOS phosphorylation, various agents which activate the ERK pathway have also been observed to induce the disassociation of the Grb2-SOS complex (7,12). Immunoprecipitation of Grb2 from unstimulated cells resulted in the co-immunoprecipitation of SOS (Fig. 1d, left,  lane 1). A similar amount of co-immunoprecipitated SOS was detected in the Grb2 immunoprecipitates from cells treated with insulin for 3 min (Fig. 1d, left, lane 2). However, following 10 or 30 min of insulin stimulation, there was a marked decrease in the ability of the Grb2 antibody to co-immunoprecipitate SOS (Fig. 1d, left, lanes 3 and 4). Pretreatment with PD98059 had no significant effect on the extent of Grb2-SOS association from unstimulated cells or from cells treated with insulin for 3 min (Fig. 1d, right, lanes 1 and 2). However, the MEK inhibitor completely blocked the disassociation of the Grb2-SOS complex following 10 and 30 min of insulin stimulation (Fig. 1d, right, lanes 3 and 4). Neither insulin nor PD98059 had any effect on the amount of Grb2 protein that was immunoprecipitated by the Grb2 antibody (data not shown).
To ensure that the effect of MEK inhibition on Grb2-SOS interactions was not cell type-specific, we also determined the effect of PD98059 on 3T3L1 adipocytes. As observed in the CHO/IR cells, insulin stimulation for 5 min resulted in the characteristic decrease in electrophoretic mobility characteristic of ERK phosphorylation (Fig. 2a, lanes 1 and 2). Pretreatment of the 3T3L1 adipocytes with PD98059 completely prevented the insulin stimulation of ERK phosphorylation (Fig.  2a, lanes 3 and 4). Similarly, PD98059 also inhibited the phosphorylation of SOS following 15 min of insulin stimulation (Fig.  2b) and the subsequent disassociation of the Grb2-SOS complex (Fig. 2c). Thus, the insulin-stimulated pathways leading to SOS phosphorylation and disassociation of the Grb2-SOS complex converged at the level of MEK in both CHO/IR cells and 3T3L1 adipocytes.
Since Ras desensitization and the disassociation of the Grb2-SOS complex were temporally related, we examined the effect of the MEK inhibitor on insulin-stimulated Ras-GTP loading ( Fig. 3). In the absence of insulin, a small fraction of Ras was in the GTP-bound state (Fig. 3a, lane 1). Insulin stimulated a rapid increase in the amount of GTP-bound Ras (Fig. 3a, lanes  2 and 3) which subsequently declined by 5 min but did not completely return to the basal state by 30 min (Fig. 3a, lanes  4 -6). Pretreatment of cells with the MEK inhibitor had no significant effect on the initial (1-3 min) insulin-stimulated increase in GTP-bound Ras (Fig. 3B, lanes 1-3), but significantly prolonged the amount of Ras that remained in the GTPbound state between 5 and 30 min (Fig. 3B, lanes 4 -6). These data demonstrate that in the continuous presence of insulin the transient nature of GTP-bound Ras was dependent on the disassociation of the Grb2-SOS complex.
Together, these data provide evidence that the insulin-stimulated phosphorylation and disassociation of the Grb2-SOS complex occurs in a MEK-dependent manner. Furthermore, the persistent association of the Grb2-SOS complex resulted in a prolonged activation of Ras suggesting that the uncoupling of Grb2 from SOS is at least one mechanism accounting for the inactivation of Ras. Since persistent Ras activation results in profound changes in cell growth and development, the molecular mechanism(s) regulating the desensitization of GTP loading is a critical feature of normal Ras function. Although our data have demonstrated that phosphorylation of SOS is MEKdependent, several studies have observed that ERK can phos-phorylate SOS both in vivo (20) and in vitro (7,11). However, it remains possible that MEK itself or other downstream protein kinases may be involved. Further studies are necessary to determine the specific kinases and sites of SOS phosphoryla-  2) or presence of the MEK inhibitor PD98059 (lanes 3 and 4), followed by a second incubation in the absence (lanes 1 and 3) or the presence (lanes 2 and 4) of 100 nM insulin. Whole cell lysates were prepared and subjected to Western blotting using an ERK antibody (a) or a SOS antibody (b) as described under "Experimental Procedures." The whole cell lysates were also immunoprecipitated with a Grb2 antibody, and the resultant immunoprecipitates were then subjected to Western blotting using a SOS antibody (c). IB, immunoblot; IP, immunoprecipitate.  1-3). The cells were then incubated in the absence (lanes 1 and 4) or presence of 100 nM insulin for 5 (lanes 2 and 5) and 20 (lanes 3 and 6) min. Whole cell lysates were prepared and subjected to Western blotting using an ERK antibody (a), a Raf antibody (b), a Rsk antibody (c), or a SOS antibody (d) as described under "Experimental Procedures." The whole cell lysates were immunoprecipitated with a Grb2 antibody, and the resultant immunoprecipitates were then subjected to Western blotting using a SOS antibody (e). The whole cell lysates were also immunoprecipitated with a Ras antibody, and the resultant immunoprecipitates were then subjected to thin layer chromatography to separate the bound guanylnucleotides (f). IB, immunoblot; IP, immunoprecipitate. tion that are responsible for the disassociation of the Grb2-SOS complex leading to Ras inactivation.