Insulin and Epidermal Growth Factor Stimulate a Conformational Change in Rap1 and Dissociation of the CrkII-C3G Complex*

Insulin and epidermal growth factor (EGF) stimulation of Chinese hamster ovary cells expressing the human insulin and EGF receptors resulted in a time-dependent decrease in the ability of a Rap1 antibody (amino acid epitope 121–136) to immunoprecipitate Rap1 from whole cell detergent extracts. This was due to an apparent masking of Rap1 as heat denaturation of the whole cell detergent extracts (5 min at 100 °C) resulted in equal immunoprecipitation of Rap1 with this epitope-specific antibody. The time-dependent change in Rap1 immunoreactivity was paralleled with an insulin-stimulated dissociation of the CrkII-C3G complex. Similarly, EGF treatment also resulted in a time-dependent dissociation of the CrkII-C3G complex that occurred concomitant with the masking of the 121–136 Rap1 epitope. Furthermore, pretreatment of the cells with the tyrosine kinase inhibitor, genistein, decreased both the basal and insulin-stimulated tyrosine phosphorylation of CrkII that directly correlated with the amount of CrkII that was immunoprecipitated with C3G. Together, these data suggest that insulin and EGF stimulation result in the dissociation of the CrkII-C3G complex, thereby inducing an apparent conformation change in Rap1.

Ras is a low molecular weight (M r 21,000) GTP-binding protein that plays an essential role in cell proliferation and differentiation (1,2). Mutations that result in constitutive activation of Ras are associated with several types of neoplastic tissue in mammals, and expression of these alleles into cultured fibroblasts results in cellular transformation (3). The best characterized proximal downstream Ras target is the Raf serine/ threonine kinase, which following growth factor stimulation, becomes activated upon association with GTP-bound Ras (4 -8). In turn, the Raf kinase phosphorylates and activates the dual functional protein kinase, termed MEK 1 (9 -13). This kinase phosphorylates the ERK family of mitogen-activated pro-tein kinases on both threonine and tyrosine residues in a characteristic TEY motif. The subsequent activation of ERK provides an important bifurcation point for the stimulation of numerous signaling pathways, including metabolic, transcriptional, and mitogenic events (14 -17).
Opposing the Ras activation pathway, the low molecular GTP-binding protein Rap was originally observed to revert or suppress the transformed phenotype in Ki-Ras-transformed fibroblasts (18,19). This apparent antagonism between Ras and Rap function may reflect the ability of Rap and Ras to interact with the same downstream effectors, since these proteins share identical sequences within their respective effector domains (18 -20). For example, several studies have demonstrated that both Rap and Ras can bind the same regulator (p120RasGAP) and effectors (RalGDS, Raf1, and B-Raf) in a GTP-dependent manner (21)(22)(23)(24)(25). In particular, association of Raf1 with GTP-bound Ras results in the activation of Raf1 protein kinase activity, whereas the association with GTPbound Rap results in an inhibition of Raf1 protein kinase activity (25).
Although the precise molecular details remained to be established, there is also substantial similarity in the upstream signaling mechanisms that regulate both Ras and Rap activation. Ras GTP binding is stimulated upon the targeting of the Ras guanylnucleotide exchange factor SOS to the plasma membrane location of Ras (26,27). The carboxyl-terminal domain of SOS contains a proline-rich regions that directs its association with the SH3 domains of the small adapter protein, Grb2 (28 -33). In analogy, the formation of active GTP-bound Rap results from the specific interaction with the Rap guanylnucleotide exchange factor C3G (34,35). Similarly, the proline-rich regions of C3G are responsible for the association with the central SH3 domain of the small adapter protein, CrkII (35)(36)(37)(38). Since Rap can function as a suppresser of Ras downstream signaling, we hypothesized that in order for growth factors to activate the Raf/MEK/ERK pathway, there must be a mechanisms to rapidly inhibit Rap function. In this manuscript, we demonstrate that insulin and EGF stimulation result in a rapid dissociation of the CrkII-C3G complex that correlates with an apparent conformational change and/or masking of Rap1 immunoreactivity.
Immunoprecipitation and Immunoblotting-Immunoprecipitations were performed from whole cell lysates by incubation with 2.0 g of Rap1 121-136 or C3G polyclonal antibodies (Santa Cruz) for 2 h at 4°C. The resulting immune complexes were precipitated by incubation with protein A-Sepharose for 1 h at 4°C and washed as described above. The pellets were then resuspended in SDS sample buffer (0.188 mM Tris-HCl, pH 6.8, 30% (v/v) glycerol, 15% (w/v) SDS, 15% 2-mercaptoethanol, 0.01% bromphenol blue) and heated at 100°C for 5 min. Whole cell lysates or immunoprecipitates were separated on 10% SDS-polyacrylamide gels (30:0.4 acrylamide to bisacrylamide) and transferred to polyvinylidene difluoride membranes at 4°C. Immunoblotting was performed with monoclonal antibodies (Transduction Laboratories) directed against Rap1, CrkII, and phosphotyrosine (PY20) or with a polyclonal antibody (Santa Cruz) directed against C3G.

RESULTS AND DISCUSSION
Insulin and EGF Stimulate a Masking of Rap1 Immunoreactivity-Insulin and EGF have been well documented to stimulate Ras activation through the regulated exchange of GDP for GTP by the guanylnucleotide exchange factor SOS (42, 43). Since we have been unable to develop a GTP binding assay for the endogenous Rap1 protein, we examined the ability of antibodies to detect potential conformational changes in Rap1. Insulin stimulation had no significant effect on the total amount of Rap1 protein that was detergent-solubilized in the whole cell extracts as detected by immunoblotting with a Rap1specific monoclonal antibody (Fig. 1, lanes 1-5). In contrast, we observed that insulin stimulation resulted in a marked reduction in the amount of Rap1 that was immunoprecipitated with a polyclonal antibody prepared against amino acids 121-136 of Rap1 (Fig. 1, lanes 6 -10). This apparent time-dependent masking of the Rap1 epitope was detected following 1 min of insulin stimulation and was maximal between 5 and 15 min.
To further demonstrate that the Rap1 121-136 antibody was detecting a masking of the Rap1 epitope, we compared the ability of this antibody to immunoprecipitate detergent-solubilized and heat-denatured Rap1 ( Fig. 2A). Insulin stimulation for 15 min had no effect on the total amount of detergent extracted Rap1 protein as detected by immunoblotting of the whole cell lysates ( Fig. 2A, lanes 1 and 2). In contrast, the ability of the Rap1 121-136 antibody to immunoprecipitate Rap1 was markedly decreased following 15 min of insulin stimulation ( Fig. 2A, lanes 3 and 4). However, heat denaturation of the same extracts restored the ability of the Rap1 121-136 antibody to immunoprecipitate identical amounts of Rap1 protein from both the unstimulated and insulin-stimulated cells ( Fig. 2A,  lanes 5 and 6).
Similar to insulin, EGF stimulation for 15 min had no effect on the total amount of detergent extracted Rap1 protein (Fig.  2B, lanes 1 and 2). Immunoprecipitation of the detergent ex-tracts with the Rap1 121-136 antibody demonstrated a marked reduction in the amount of Rap1 protein following EGF stimulation (Fig. 2B, lanes 3 and 4). The decreased immunoreactivity of Rap1 was completely reversed following heat denaturation of the detergent cell extracts (Fig. 2B, lanes 5 and 6). Together, these data demonstrate that growth factor stimulation (insulin and EGF) induced the masking of the Rap1 121-136 amino acid epitope. Although we have not yet been able to determine whether this reflects a direct conformational change in Rap1 or is due to the association of Rap1 with another molecule that blocks this epitope, it is likely to represent the conversion of active GTP-bound Rap1 to the inactive GDPbound state.
Insulin and EGF Stimulate the Dissociation of the CrkII-C3G Complex-We and others have also recently observed that the inactivation of Ras back to the GDP-bound state occurs concomitant with the dissociation of the Grb2-SOS complex (41, 44 -46). We therefore speculated that the apparent conformational change and/or masking of Rap1 immunoreactivity might have resulted from the inactivation of Rap1 due to an uncoupling of the CrkII-C3G complex. To examine the association state of the CrkII-C3G complex, we next determined the effect of insulin and EGF on the co-immunoprecipitation of these complexes (Fig. 3). As expected, in unstimulated cells immunoprecipitation of C3G resulted in the co-immunoprecipitation of CrkII (Fig. 3A, lane 1). However, following insulin stimulation, there was a time-dependent decrease in the amount of CrkII that could be co-immunoprecipitated with C3G (Fig. 3A, lanes  1-6). The dissociation of CrkII from C3G was detectable as early as 1 min and was maximal between 3 and 5 min (Fig. 3A,  lanes 3 and 4). The insulin-stimulated decrease of C3G immunoprecipitated CrkII protein was not due to differences in C3G immunoprecipitation as assessed by C3G immunoblotting of the C3G immunoprecipitates (Fig. 3A, lanes 7-12). Furthermore, the time dependence of CrkII-C3G dissociation was similar to or slightly preceded the insulin-stimulated decrease in Rap1 immunoreactivity (Fig. 1).
In comparison, EGF stimulation also resulted in a rapid dissociation of the CrkII-C3G complex (Fig. 3B, lanes 1-6). The time-dependent uncoupling of CrkII from C3G was similar to insulin with detectable dissociation following 1 min of EGF treatment (Fig. 3B, lane 2). The maximal EGF-stimulated dissociation of the CrkII-C3G complex occurred between 3 and 5 min and was persistent for up to 30 min (Fig. 3B, lanes 3-6). As  (lanes 1, 3, and 5) or incubated with 100 nM insulin (lanes 2, 4, and 6) for 15 min at 37°C. B, CHO/IR/ER cells were either untreated (lanes 1, 3, and 5) or incubated with 20 nM EGF (lanes 2, 4, and 6) for 15 min at 37°C. In each case, whole cell detergent extracts (Lysate) were immunoblotted for Rap1 (lanes 1 and 2). The extracts were either then incubated with the antibody directed against the amino acid 121-136 Rap1 epitope prior to (Native) or following heat denaturation (Denatured) by incubation for 5 min at 100°C. The resulting immunoprecipitates were then immunoblotted with a Rap1 monoclonal antibody as described under "Experimental Procedures." a control for immunoprecipitation, immunoblotting with the C3G antibody demonstrated equivalent amounts of immunoprecipitated C3G protein under each condition (Fig. 3B, lanes  7-12). These data are consistent with the CrkII-C3G complex functioning to maintain Rap1 in the active GTP-bound state, which is recognized by the Rap1 121-136 epitope-specific antibody. However, the dissociation of the CrkII from C3G complex terminates this activation signal, thus allowing Rap1 to convert to the inactive GDP-bound state.
The Dissociation of the CrkII-C3G Complex Requires Activation of Protein Tyrosine Kinase Activity-Recently it has been observed that CrkII becomes tyrosine-phosphorylated following growth factor stimulation (47). Based upon these previous findings and the rapid insulin-and EGF-stimulated dissociation of the CrkII-C3G complex, we next examined the potential role for the tyrosine phosphorylation of CrkII. This was assessed by use of the selective tyrosine kinase inhibitor, genistein (Fig. 4). Phosphotyrosine immunoblots of whole cell detergent extracts demonstrated the insulin stimulation of IRS1/2 and insulin receptor ␤ subunit tyrosine phosphorylation (Fig.  4A, lanes 1 and 2). Genistein pretreatment of the cells reduced both the basal level of tyrosine-phosphorylated proteins as well as decreased the extent of insulin-stimulated IRS1/2 and insulin receptor ␤ subunit tyrosine phosphorylation (Fig. 4A, lanes  3 and 4). It should be noted that genistein only partially inhibited insulin-stimulated tyrosine phosphorylation, probably due to the high levels of insulin receptors expressed in this cell line. As expected, insulin stimulation resulted in an increased tyrosine phosphorylation of CrkII (Fig. 4A, lanes 5 and 6). Interestingly, a tyrosine-phosphorylated protein in the 120 -130-kDa range was also found to co-immunoprecipitate with CrkII, which decreased following insulin treatment (Fig. 4A, lanes 5  and 6). In any case, genistein pretreatment inhibited both the basal and insulin-stimulated tyrosine phosphorylation of CrkII as well as the tyrosine dephosphorylation of the 120 -130-kDa band (Fig. 4A, lanes 7 and 8).
The relative tyrosine phosphorylation state of CrkII also correlated with the extent of association between CrkII and C3G (Fig. 4B). As observed previously, immunoprecipitation of C3G resulted in the co-immunoprecipitation of CrkII, which was decreased following insulin stimulation (Fig. 4B, lanes 1  and 2). In the unstimulated cells, although pretreatment with genistein decreased the amount of tyrosine-phosphorylated CrkII, there was an increase in the relative extent of CrkII that was co-immunoprecipitated with C3G (Fig. 4B, lane 3). Furthermore, genistein pretreatment also prevented the insulinstimulated increase CrkII tyrosine phosphorylation and concomitantly inhibited the insulin-stimulated dissociation of the CrkII-C3G complex (Fig. 4B, lane 4). As controls for immunoprecipitation, the amount of C3G immunoprecipitated under these conditions remained unchanged (Fig. 4B, lanes 5-8). Thus, these data indicated that the extent of CrkII tyrosine phosphorylation was inversely related to the association state of the CrkII-C3G complex.
In summary, we have observed that growth factor stimulation results in a rapid dissociation of the CrkII-C3G complex, which parallels or slightly precedes an apparent conformational change and/or masking of the Rap1 121-136 amino acid epitope. Since the association state of CrkII with C3G appears to correlate with the extent of CrkII tyrosine phosphorylation, we speculate that receptor tyrosine kinase phosphorylation of CrkII regulates this interaction. Further studies will be necessary to determine whether CrkII is a direct substrate of receptor tyrosine kinases, the site(s) of CrkII phosphorylation, and the functional role of this tyrosine phosphorylation in modulating the interaction between CrkII and C3G. In any case, based upon the ability of Rap1 to suppress Ras signaling, the ability of growth factors to activate the Raf/MEK/ERK cascade requires a mechanism for the inactivation of Rap1 function. Our data suggest that the growth factor-stimulated uncoupling of the CrkII-C3G complex may be one such mechanism.  1 and 7) or in the presence of 20 nM EGF for 1 (lanes 2 and 8), 3 (lanes 3 and 9), 5 (lanes 4 and 10), 15 (lanes 5 and 11), and 30 (lanes 6 and 12) min at 37°C. In both cases, whole cell detergent extracts were prepared and were immunoprecipitated with a C3G antibody. The C3G immunoprecipitates were immunoblotted with a CrkII antibody (lanes 1-6) or a C3G antibody (lanes 7-12) as described under "Experimental Procedures." FIG. 4. The amount of CrkII co-immunoprecipitated with C3G is inversely proportional to the extent of CrkII tyrosine phosphorylation. CHO/IR/ER cells were preincubated in the absence or presence of 300 M genistein for 1 h at 37°C. The cells were then either unstimulated or incubated in the presence of 100 nM insulin for an additional 15 min at 37°C. A, whole cell detergent extracts (Lysates) were prepared and directly immunoblotted with the PY20 phosphotyrosine antibody (lanes 1-4) or immunoprecipitated with a CrkII antibody and immunoblotted with the PY20 phosphotyrosine antibody (lanes 5-8) as described under "Experimental Procedures". B, the whole cell detergent extracts were immunoprecipitated with a C3G antibody and immunoblotted with a CrkII antibody (lanes 1-4) or with a C3G antibody (lanes 5-8).
Growth Factor-stimulated Dissociation of CrkII from C3G 28181