Enhancement of Guanine-Nucleotide Exchange Activity of C3G for Rap1 by the Expression of Crk, CrkL, and Grb2*

Crk is an adaptor protein that consists almost entirely of SH2 and SH3 domains. We have previously demonstrated, by using in vivo and in vitro systems, that C3G, which was identified as a Crk SH3 domain-binding guanine nucleotide exchange factor, specifically activates Rap1. C3G also binds to other adaptor proteins, including CrkL and Grb2. In the present study, we analyzed the effect of Crk, CrkL, and Grb2 on the C3G-Rap1 pathway. Expression of Crk, CrkL, and Grb2 with C3G in Cos1 cells significantly increased the ratio of GTP/GDP bound to Rap1. Both the SH2 and SH3 domains of Crk were required for this activity. However, Crk did not stimulate the guanine nucleotide exchange activity of C3G for Rap1in vitro, suggesting that Crk does not activate C3G by an allosteric mechanism. The requirement of the SH2 domain of Crk for the enhancement of guanine nucleotide exchange activity for Rap1 could be compensated for by the addition of a farnesylation signal to Crk, indicating that Crk enhanced the guanine nucleotide exchange activity of C3G by membrane recruitment of C3G. These results demonstrate that Crk, CrkL, and Grb2 positively modulate the C3G-Rap1 pathway primarily by recruiting C3G to the cell membrane.

main and two SH3 domains, whereas CrkI lacks the carboxylterminal SH3 domain. The SH2 domain of Crk binds epidermal growth factor receptor and Shc, which are phosphorylated by EGF 1 stimulation (5,11). Integrin stimulation also induces tyrosine phosphorylation of p130 cas and paxillin, resulting in binding to Crk (12,13). Cbl, which was originally identified as a cellular counterpart of the v-Cbl oncogene product (14), is tyrosine-phosphorylated upon various extracellular stimulations and binds to Crk (15,16). Therefore, Crk is implicated in the signal transduction pathway of both growth factor and cell adhesion.
Cellular targets of the SH3 domain of Crk include C3G, Sos, DOCK180, c-Abl, and EPS15 (17)(18)(19)(20)(21). C3G was isolated by screening of expression libraries with the amino-terminal SH3 domain of Crk as a probe. The carboxyl-terminal region of C3G has homology to the catalytic domain of CDC25 Mm , which is a guanine nucleotide exchange factor for Ras. Using both in vivo and in vitro systems, we demonstrated that the target GTPase of C3G is Rap1 (22). Rap1, also known as smg p21 or Krev-1, was originally identified as an anti-oncogenic protein which efficiently reverses the morphologic transformation of the v-Ki-ras-expressing NIH 3T3 cell line, DT (23)(24)(25)(26). Studies on chimeras of Rap1 and H-Ras have suggested that Rap1 antagonizes Ras through the competition of the effector proteins of Ras (27). This is also supported by the observation that the activation of the c-fos promoter/enhancer from Ras, but not from the activated Raf-1, which is an immediate downstream effector of Ras, is inhibited by Rap1 (28). Furthermore, constitutively activated Rap1 efficiently inhibits Ras-dependent activation of mitogen-activated protein kinase by lysophosphatidate or EGF (29). Rap1 is regulated mostly by two groups of proteins. The first one is the GTPase-activating protein, which stimulates the intrinsic GTPase activity of Rap1. Two proteins have been shown to possess GTPase-activating protein activity for Rap1 (30,31). The second are guanine nucleotide exchange factors for Rap1, which include smg GDS and C3G (22,32). smg GDS also stimulates the nucleotide dissociation of Ki-Ras, Rho, Rac, and mCdc42 (33). In contrast, C3G specifically stimulates the guanine nucleotide exchange reaction of Rap1 among various Ras-family G proteins (22).
Our aim in this study was to analyze the effect of adaptor proteins that bind to C3G on the C3G-Rap1 pathway. We found that Crk enhances the guanine nucleotide exchange reaction of * This work was supported in part by grants from the Ministry of Education, Science, and Culture, by grants from the Ministry of Health and Welfare, and by grants from the Human Science Foundation, Japan. 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.
¶ C3G, and that the enhancement occurs mostly by promoting of the translocation of C3G to the membrane.
Cell Culture and Transfection-Cos1 cells and human embryonic kidney 293T cells were cultured in Dulbecco's modified Eagle's medium (Nissui, Tokyo) supplemented with 10% fetal calf serum. Cos1 cells were transfected by DEAE-dextran method as described previously (22). Human embryonic kidney 293T cells were transfected by calcium phosphate method.
Antibodies-Anti-Crk antibody was obtained from Transduction Laboratories. Production of the anti-C3G antibodies was described previously (35). Anti-C3G 1A/1B and anti-C3G-C sera were raised against the central region (amino acids 285-676) and the carboxyl terminus region (amino acids 1065-1077) of C3G, respectively. The epitope of anti-C3G monoclonal antibody SN21 was in the region between amino acids 393-676. Anti-GST rabbit antibody was prepared in our laboratory.
Analysis of Guanine Nucleotide Exchange Activity of C3G for Rap1 in Cos1 Cells-Guanine nucleotides bound to Rap1 were analyzed essentially as described previously (22). Briefly, 10 5 Cos1 cells plated on 35-mm dishes were transfected with 0.2 g of expression plasmids by the DEAE-dextran method. Forty-eight hours after transfection, cells were labeled for 2 h with 0.05 mCi of 32 P i in 0.5 ml of phosphate-free medium. GST-fused Rap1 was collected by glutathione-Sepharose 4B (Pharmacia). Guanine nucleotides bound to Rap1 were separated by thin-layer chromatography and quantitated with a PhosphorImager analyzer, BAS2000 (Fuji Film). For the analysis of protein expression, cells were similarly processed without isotopic labeling. The cleared lysates were separated by SDS-PAGE, transferred to a nylon membrane, probed with various antibodies, and detected by ECL chemiluminescence system (Amersham Corp.).
In Vitro Analysis of Guanine Nucleotide Exchange Activity of C3G for Rap1-The guanine nucleotide exchange activity of C3G for Rap1 was measured as described previously (22). Expression and purification of recombinant Rap1, C3G, and CrkI were described previously (22,34). C3G preincubated with GST or GST-CrkI was combined with [ 3 H]GDPloaded Rap1 and an excess of cold GTP. The radioactivity of [ 3 H]GDP remaining on Rap1 was measured after a 20-min incubation at 30°C.
Cell Fractionation-Cell fractionation was performed as described previously (36). Briefly, 10 5 Cos1 cells transfected with expression plasmids were washed twice with phosphate-buffered saline, disrupted by freeze-thawing in liquid nitrogen, suspended with 150 l of detergent-free W buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 20 mM MgCl 2 ). The soluble and the insoluble fraction were separated by centrifugation at 15,000 ϫ g for 10 min. Proteins in the insoluble fraction were solubilized with 37.5 l of lysis buffer (0.5% Triton X-100, 20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 5 mM MgCl 2 ). Extracts were centrifuged at 15,000 ϫ g for 10 min, and the supernatants were used as a particulate fraction. 2 l of soluble and 4 l of particulate fraction were subjected to SDS-PAGE and immunoblotting using anti-C3G 1A/1B sera.

Expression of Adaptor Proteins
Enhanced the Guanine Nucleotide Exchange Activity of C3G for Rap1-Cos1 cells were transfected with expression vectors encoding GST-tagged Rap1, C3G, and/or adaptor proteins. After 48 h, guanine nucleotides bound to GST-Rap1 were analyzed by thin-layer chromatography (Fig. 2, A and B). As reported previously (22), expression of C3G increased the ratio of GTP/(GDP ϩ GTP) on Rap1. Expression of CrkI or CrkII remarkably enhanced the C3G-dependent increase of GTP-bound Rap1, whereas CrkI or CrkII alone did not affect the ratio of GTP/(GDP ϩ GTP) on Rap1. Neither CrkI mutant with nonfunctional SH2, CrkI-R38V, nor that with nonfunctional SH3, CrkI-W169L, enhanced this C3G-dependent Rap1 activation. We confirmed that similar amounts of C3G and the wild-type and mutant Crk proteins were expressed in Cos1 cells (Fig. 2C). These results demonstrated that both the SH2 and SH3 domains of Crk were required for the enhancement of the guanine nucleotide exchange activity of C3G. We observed a similar enhancement of C3G activity by the other Crk-like protein, CrkL (Fig. 3, A and  B). Another adaptor protein, Grb2, also increased the ratio of GTP to GDP on Rap1. However, the increase in GTP-bound Rap1 was less than that by CrkI, CrkII, or CrkL. In these experiments, coexpression of the adaptor proteins did not alter the level of C3G expression (Fig. 3C).
The Amino-terminal Region of C3G Was Required for Enhancement by Crk-To understand the mechanism by which Crk enhances the guanine nucleotide exchange activity of C3G, we constructed a plasmid, C3G-CD, that encoded only the catalytic domain of C3G. To our surprise, C3G-CD activated Rap1 more efficiently than did the authentic C3G (Fig. 4, A and  B). We examined the expression of C3G-CD and the wild-type C3G by an antibody against the carboxyl-terminal region of C3G. As shown in Fig. 4C, the amount of C3G-CD was comparable to that of the wild-type C3G (Fig. 4C). Therefore, the amino-terminal region of C3G appears to regulate negatively the guanine nucleotide exchange activity for Rap1. Moreover, the catalytic activity of this C3G-CD mutant was enhanced no further by the co-expression of CrkI. This result strongly suggests that direct binding of Crk to the amino-terminal region of C3G is required for the up-regulation of C3G.
Crk Did Not Promote the Guanine Nucleotide Exchange Reaction of C3G for Rap1 in Vitro-To examine whether Crk activates C3G through an allosteric mechanism, we measured the guanine nucleotide exchange activity of C3G in vitro in the presence or absence of CrkI. As shown in Fig. 5A, addition of GST-fused CrkI did not enhance the catalytic activity of C3G for Rap1. The binding of the recombinant CrkI to C3G in vitro was confirmed by immunoblotting (Fig. 5B). We concluded that binding of Crk to C3G is not sufficient to enhance the catalytic activity of C3G in vitro.
Crk Recruited C3G to the Particulate Fraction-The SH2 domain of Crk binds to tyrosine-phosphorylated proteins such as EGF receptor (5,11), p130 cas (37), Shc (17), and paxillin (13). These proteins are localized to the plasma membrane either constitutively or transiently after various stimulations. It is also known that overexpression of v-Crk or CrkI induces tyrosine phosphorylation of these Crk SH2-binding proteins either by the activation of tyrosine kinases or by the competitive inhibition of protein-tyrosine phosphatases (9 -11). Thus, it has been postulated that at least one of the functions of the SH2 domain of Crk is to recruit the SH3-binding proteins to the plasma membrane. To test this possibility, we analyzed the subcellular distribution of C3G in the presence or absence of CrkI. As shown in Fig. 6, co-expression of CrkI increased the amount of C3G in the particulate fraction.
Membrane Targeting of CrkI Complemented the SH2 Mutation of CrkI-To examine whether membrane translocation of C3G can account for the activation of C3G by Crk, we constructed plasmids encoding CrkI-F and CrkI-R38V-F that have the Ki-Ras-derived CAAX box at the carboxyl termini of CrkI and CrkI-R38V, respectively. The CAAX box was shown to be farnesylated and required for the membrane-targeting of Ki-Ras protein. As shown in Fig. 7, A and B, CrkI-R38V-F en-  Fig. 2. B, the radioactivity of each spot was measured by PhosphorImager for calculation of the ratio of GTP to GTP ϩ GDP. Mean values obtained from three samples are shown with standard deviations. C, in parallel experiments, cells without 32 P i labeling were analyzed by SDS-PAGE followed by immunoblotting with anti-C3G 1A/1B sera.
hanced the activation of Rap1 by C3G to a level similar to that by CrkI-F. The expression level of C3G was not influenced by the expression of the wild-type and mutant CrkI proteins. This result clearly demonstrated that the SH2 domain of CrkI was not required when CrkI was localized to the plasma membrane by the farnesylation. Enhancement of the activation of C3G by CrkI-F was always less efficient than that by the wild-type CrkI, although CrkI-F was expressed more abundantly than the wild-type CrkI (Fig. 7C). Expression of these wild-type and mutant CrkI did not alter the amount of C3G expressed in Cos1 cells (Fig. 7C). We also confirmed that the addition of the farnesylation signal did not alter the amount of Crk bound to C3G in immunoprecipitation experiments (data not shown).
Crk Recruited C3G to EGF Receptor upon EGF Stimulation-It has been reported that Crk binds to EGF receptor upon EGF stimulation (11). We examined whether the EGF stimulation also recruits C3G to the EGF receptor in 293T cells. As shown in Fig. 8, C3G bound to EGF receptor upon EGF stimulation in a manner dependent on the expression of CrkI. This result strongly suggests that Crk recruits C3G to the plasma membrane upon physiological stimulation. DISCUSSION In the present study, we have demonstrated that the expression of Crk enhances the guanine nucleotide exchange activity of C3G to Rap1 in vivo. Crk did not enhance the guanine nucleotide exchange activity of C3G in vitro, arguing against the allosteric activation of C3G by Crk. Analysis by use of SH2   FIG. 4. Crk-binding sites of C3G are required for enhancement of guanine nucleotide exchange activity by Crk. A, Cos1 cells were transfected with the expression plasmids encoding the proteins indicated on the bottom and analyzed by thin-layer chromatography as described in Fig. 2. C3G-CD consists only of the catalytic region of C3G. B, the radioactivity of each spot was measured by PhosphorImager for calculation of the ratio of GTP to GTP ϩ GDP. Mean values obtained from three samples are shown with S.D. C, in parallel experiments, cells without 32 P i labeling were analyzed by immunoblotting with the anti-C3G-C serum, which was raised against the carboxyl-terminal region of the C3G (35).

FIG. 5. Crk does not enhance the catalytic activity of C3G in vitro.
A, Rap1 was loaded with [ 3 H]GDP as described in the text. The guanine nucleotide exchange reaction was started by the addition of C3G preincubated with GST or GST-CrkI. After 20-min incubation at 30°C, Rap1 was absorbed on a nitrocellulose membrane, and the radioactivity was measured. B, C3G was incubated with either GST or GST-CrkI on ice for 10 min. C3G was immunoprecipitated anti-C3G 1A/1B sera and protein A-Sepharose. GST or GST-CrkI was collected by glutathione-Sepharose 4B. The proteins collected on the beads were analyzed by SDS-PAGE and immunoblotting by either anti-C3G 1A/1B sera or anti-GST antibody.
FIG. 6. Crk recruits C3G to the particulate fraction. Cos1 cells transfected with either pCAGGS-C3G alone or pCAGGS-C3G and pCAGGS-CrkI were subjected to cell fractionation as described in the text. The soluble (sup.) and the particulate (ppt.) fractions were analyzed by SDS-PAGE and immunoblotting using anti-C3G 1A/1B sera or anti-Crk antibody. and SH3 mutants of Crk demonstrated that both the SH2 and SH3 domains of Crk are required for the enhancement. In contrast to CrkI-R38V, a CrkI mutant with a non-functional SH2 domain, the mutant that has a membrane localization signal on its carboxyl terminus, CrkI-R38V-F, enhanced the catalytic activity of C3G to a level similar to that of CrkI-F, the wild-type CrkI with a membrane localization signal. The result indicates that the function of the SH2 domain is essentially to recruit Crk to the plasma membrane.
However, we should point out the difference between the membrane recruitment of C3G by CrkI and that by the farnesylation. The enhancement of C3G activity by CrkI-F was less efficient than that by the wild-type CrkI (Fig. 7). Thus, CrkI-F may not recruit C3G to the place where the optimum activation of C3G occurs. In the cells overexpressing v-Crk or CrkI, we and others observed the increase of several phosphotyrosinecontaining proteins including paxillin and p130 cas , to which v-Crk and CrkI bind (9,10). C3G may be activated most efficiently in the close vicinity of these Crk-binding proteins, which is known to localize at the focal adhesions (8).
It is unknown why membrane recruitment of C3G by CrkI enhances guanine nucleotide exchange activity of C3G. It is possible that the membrane targeting of C3G simply facilitates its interaction with the substrate, Rap1. Alternatively, C3G may be modified at the membrane and activated enzymatically. To address this question, we have to measure the specific activity of C3G from the cells with or without the expression of CrkI.
We previously reported that farnesylated C3G activated Rap1 in Cos1 cells less efficiently than did the wild-type C3G (22). This observation appears to argue against our proposal that CrkI enhances guanine nucleotide exchange activity of C3G by membrane recruitment. We have found that the amount of the farnesylated C3G expressed in Cos1 cells is significantly less than that of the wild-type; therefore, we cannot conclude that the farnesylated C3G was not as active as the wild-type C3G. Moreover, only the farnesylated C3G but not the wild-type C3G could revert the transformation of NIH 3T3 cells expressing Ki-Ras (22). This finding supports our proposal that membrane-targeted C3G is more active than the wild-type C3G.
The significance of the translocation of the guanine nucleotide exchange factor to the membrane has been suggested for Sos, a guanine nucleotide exchange protein of Ras (38 -41). In this case, another adaptor protein, Grb2, recruits Sos to the membrane upon various stimulations, as we have observed for CrkI and C3G. It has also been reported that the binding of Grb2 to Sos does not activate Sos in vitro (42).
Rap1 is phosphorylated by protein kinase A at Ser 169 , located at the carboxyl-terminal basic region that is critical for the interaction with smg GDS (43). This phosphorylation stimulates the guanine nucleotide dissociation of Rap1 by smg GDS (44). However, we could not observe any difference in the phosphorylation level of Rap1 in the presence or absence of CrkI (data not shown). Therefore, it is unlikely that Crk enhances the guanine nucleotide exchange reaction of C3G through the phosphorylation of Rap1.
Crk and CrkL enhance the guanine nucleotide exchange FIG. 7. Requirement of the SH2 domain of Crk for enhancement of C3G activity is complemented by the addition of a farnesylation signal to Crk. A, Cos1 cells were transfected with the expression plasmids encoding the proteins indicated on the bottom and analyzed by thin-layer chromatography as described in Fig. 2. CrkI-F and CrkI-R38V-F have a farnesylation signal on their carboxyl terminus. B, the radioactivity of each spot was measured by PhosphorImager for calculation of the ratio of GTP to GTP ϩ GDP. Mean values obtained from three samples are shown with standard deviations. C, in parallel experiments, cells without 32 P i labeling were analyzed by immunoblotting with either anti-C3G 1A/1B sera or anti-Crk antibody.
FIG. 8. C3G bound to EGF receptor upon EGF stimulation in a manner dependent on CrkI. 293T cells were transfected with expression plasmids and stimulated with EGF for 3 min. Cells were lysed and immunoprecipitated with anti-EGF receptor (EGFR) polyclonal antibody. The immune complex was analyzed by SDS-PAGE and immunoblotting using either anti-EGF receptor monoclonal antibody or anti-C3G 1A/1B sera. activity of C3G more efficiently than does Grb2. It is possible that Grb2 cannot alter C3G localization efficiently because Grb2 does not bind to C3G as strongly as do Crk and CrkL (45). Alternatively, because the SH2 domains of Crk and CrkL bind similar sets of proteins (45,46), both may recruit C3G to a position closer to Rap1 than does Grb2. Experiments on SH2/ SH3 chimeras of Crk/CrkL and Grb2 would be useful.
C3G-CD, which lacks the amino-terminal region of C3G, catalyzed the guanine nucleotide exchange reaction for Rap1 more efficiently than did the authentic C3G. Although the in vitro study neglected the allosteric activation of C3G by Crk, it is still possible that expression of Crk can counteract the negative regulation of C3G by the amino-terminal region in vivo. A similar phenomenon is known for Sos, a guanine nucleotide exchange factor for Ras. The deletion of the carboxyl-terminal Grb2-binding region sites enhances the guanine nucleotide exchange activity and the transforming activity of the farnesylated Sos protein (41). Although the exact role of the aminoterminal region of C3G has yet to be elucidated, it is possible that it is involved in the down-regulation of the C3G-Rap1 signaling pathway.