Activation of the focal adhesion kinase signaling pathway by ... - Nature

Oncogene (2001) 20, 951 ± 961 ã 2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00 www.nature.com/onc

Activation of the focal adhesion kinase signaling pathway by structural alterations in the carboxyl-terminal region of c-Crk II Agnes Zvara1,3,5, J Eduardo Fajardo1,5, Marcela Escalante1, Graham Cotton2, Tom Muir2, Kathrin H Kirsch1 and Raymond B Birge*,1,4 1

Laboratory of Molecular Oncology, The Rockefeller University, 1230 York Avenue, New York, NY 10021, USA; 2Laboratory of Synthetic Protein Chemistry, The Rockefeller University, 1230 York Avenue, New York, NY 10021, USA

The Crk II adaptor protein encodes an SH2/SH3-domain containing adaptor protein with an SH2 ± SH3 ± SH3 domain structure that transmits signals from tyrosine kinases. The two SH3 domains are separated by a 54 amino acid linker region, whose length is highly conserved in xenopus, chicken, and mamalian Crk II proteins. To gain a better understanding into the role of the C-terminal region of Crk, we generated a series of C-terminal SH3 domain and SH3 linker mutants and examined their role in tyrosine kinase pathways. Expression of point mutations in the C-terminal SH3 domain (W276K Crk), at the tyrosine phosphorylation site (Y222F Crk II), or truncation of the entire C-terminus (Crk I or Crk D242), all increased c-Abl binding to the N-terminal SH3 domain of Crk and, where relevant, increased Tyr222 phosphorylation. Deletion analysis of c-Crk II also revealed the presence of a C-terminal segment important for transactivation of FAK. Such mutants, Crk D255 or Crk D242 Extended Linker (Crk D242[EL]), characterized by a disruption in the SH3 linker/C-terminal SH3 boundary, induced robust hyperphosphorylation of focal adhesion kinase (FAK) on Tyr397, hyperphosphorylation of focal adhesion proteins p130cas and paxillin and increased focal adhesion formation in NIH3T3 cells. The eects of Crk D242[EL] could be abrogated by co-expression of dominant negative c-Src or the protein tyrosine phosphatase PTP ± PEST, but not by dominant negative Abl. Our results suggest that the C-terminal region of Crk contains negative regulatory elements important for both Abl and FAK dependent signal pathways, and oers a paradigm for an autoinhibitory region in the SH3 linker/C-terminal SH3 domain. Oncogene (2001) 20, 951 ± 961. Keywords: Crk; adaptor proteins; SH3 linker region; C-terminal SH3 domain; tyrosine phosphorylation; focal adhesion kinase

*Correspondence: RB Birge Current addresses: 3Biochip Laboratory, BRC, Hungary Academy of Sciences, Szeded POB 521, H6701 Hungary; 4UMDNJ-New Jersey Medical School, Department of Biochemistry and Molecular Biology MSB-E647, 185 South Orange Avenue, Newark, NJ 07103-2714, USA 5 These authors contributed equally to this work Received 24 August 2000; revised 8 December 2000; accepted 12 December 2000

Introduction The Crk family of adaptor proteins (c-Crk II, c-Crk I, CrkL) are Src Homology 2 (SH2) and SH3 domain containing proteins that have been implicated in many signaling events of proliferation, dierentiation, cell adhesion, and cytoskeletal reorganization (Birge et al., 1996; Feller et al., 1998). The role of Crk in the aforementioned signaling pathways is primarily mediated by the SH2 and the ®rst SH3 domain which form speci®c interactions with intracellular proteins. The SH2 domain of Crk and CrkL binds in the context of phospho Tyr-X-X-Pro, and primarily interact with tyrosine phosphorylated focal adhesion proteins p130cas and paxillin (Birge et al., 1993; Sakai et al., 1994; Songyang et al., 1994). The N-terminal SH3 domain of Crk and CrkL binds to proline-rich sequences in the context of Pro-X-X-Pro-X-Lys (Knudsen et al., 1994; Wu et al., 1995), and interacts with a limited number of cellular proteins including C3G, a guanine nucleotide exchange factor for Rap1 (Tanaka et al., 1994) and R-Ras (Mochizuki et al., 2000), DOCK180, a regulator of Rac1 (Hasegawa et al., 1996; Kiyokawa et al., 1998a,b), the hematapoietic progenitor kinase 1 (PRK1), an upstream activator of JNK (Ling et al., 1999), and the tyrosine kinase c-Abl (Feller et al., 1994). The involvement of Crk in intracellular signaling is initiated by many types of extracellular stimulation that include growth factors such as EGF, NGF, CSF1, IGF-1, mitogenic lipids, engagement of the T cell and B cell receptors, and adhesion of cells to extracellular matrix (reviewed in Feller et al., 1998; Matsuda and Kurata, 1996). Many of these induce the tyrosine phosphorylation of the focal adhesion proteins p130cas and paxillin, which recruit Crk and its SH3 binding partners to the cytoskeleton where the signals can be coupled. Previous studies have shown that binding of DOCK180 to tyrosine phosphorylated p130cas/c-Crk II complex results in GDP-GTP exchange activity on Rac1 (Kiyokawa et al., 1998a,b), which is critical for integrin-mediated cellular migration (Cheresh et al., 1999). Additionally, binding of C3G to the p130cas/c-Crk II complex has been implicated in integrin-mediated JNK activation (Dol® et al., 1998; Tanaka and Hanafusa, 1998) and G1 to S cell cycle

Crk and focal adhesion signaling A Zvara et al

952

progression (Oktay et al., 1999). Besides acting to transmit signals from tyrosine phosphorylated focal adhesion proteins to downstream eector pathways, there is also evidence that Crk can feedback and act upstream to activate FAK. For example, in cells transformed by v-Crk, the product of the avian sarcoma virus CT10, oncogenic Crk induces FAK activation (Altun-Gultekin et al., 1998; Polte and Hanks, 1995) and a speci®c elevation in the tyrosine phosphorylation of p130cas and paxillin (Mayer et al., 1988). v-Crk can also induce the formation of actin stress ®bers and de novo focal adhesion biogenesis in pheochromocytoma (PC12) cells (Altun-Gultekin et al., 1998), indicating that the recruitment of Crk to focal adhesions represents a point of convergence for both upstream and downstream signals to FAK. Despite the fact that several proteins have been identi®ed that bind to the N-terminal SH3 domain, no cellular molecules have been shown to interact with the Crk C-terminal SH3 domain, and the function of the C-terminus of c-Crk II is poorly understood. Mutations in the C-terminal SH3 domain have been shown to disrupt EGF-induced Ras activation in NRK cells (Kizaka-Kondoh et al., 1996), and the C-terminal SH3 domain negatively regulates the tyrosine phosphorylation of p130cas and the transforming ability of Crk (Ogawa et al., 1994). Alternative splicing of the c-crk gene yields two dierent translation products, c-Crk I (28 kD) and c-Crk II (42 kD) (Matsuda et al., 1994) which only dier in their C-terminal region. In cCrk II, the two SH3 domains are separated by a 54 amino acid proline-rich linker region whose sequence is highly conserved in xenopus, chicken and mammalian proteins (Reichman et al., 1992). This spacer region also contains a tyrosine motif Y222AQP that can be phosphorylated by c-Abl, creating an intramolecular binding site for the Crk SH2 domain that regulates the folding and ability of both SH2 and SH3 domain interactions with cellular proteins (Rosen et al., 1995). In this study, we have further investigated the role of the C-terminal region in Crk II by generating Cterminal Crk SH3 and linker mutants of dierent length and composition. Our results show that several C-terminal mutants, particularly in the C-terminal SH3 domain, enhance binding of Abl to Crk. Moreover, by generating a series of SH3 linker mutants, we uncovered an unusual mutant that strongly activates FAK and leads to elevation in the tyrosine phosphorylation of focal adhesion proteins. These results suggest the C-terminus of Crk contains distinct negative regulatory elements that control both Abl and FAKdependent signaling pathways.

Results Carboxyl terminal mutants of Crk increase the in vivo tyrosine phosphorylation of p130cas We investigated the function of the Crk C-terminal region by expressing wild-type Crk and various Crk Oncogene

mutants in the human kidney epithelial (HEK) cell line 293T (Figure 1). Co-expression of c-Crk with p130cas resulted in tyrosine phosphorylation of p130cas, which was dependent on Crk expression (Figure 2a, lanes 1 and 2). Co-expression of p130cas with carboxyl-terminal deletions of Crk up until the boundary of the second SH3 junction (Crk D242), or shorter truncation mutants Crk D232, Crk D222, Crk D212, or Crk I (terminates at amino acid 204), all stimulated p130cas phosphorylation to a similar extent, or slightly less than wild-type Crk II (Figure 2a, lanes 4 ± 6 and not shown). In contrast, expression of Crk D255, a truncation mutant containing the entire linker domain plus 13 amino acids into the C-terminal SH3 domain, strongly induced the tyrosine phosphorylation of p130cas relative to wild-type Crk II (lane 3). Moreover, Crk D255 and Crk D242 were also hyperphosphorylated on Tyr222 (Figure 2a, anti-pCrk blot); in contrast Crk D232 and Crk D222 mutants were not phosphorylated on the Tyr222 residue, even though Tyr222 was not mutated in these mutants. The above results might be explained if amino acids 242 ± 255 at the beginning of the SH3 domain bound to an eector of p130cas, or that an alteration in the SH3 linker/C-terminus SH3 domain boundary perturbed the conformation of the linker region, de-repressing a native inhibitory function. To investigate this, we generated an arti®cial Crk D255 by replacing amino acids 242 through 255 of Crk with 13 unrelated sequences from the pEBB vector, that are not homologous to any mammalian proteins (Figure 1). This mutant containing an arti®cal C-terminal extension will be referred to as Crk D242-Extended Linker (Crk D242[EL]) throughout the paper. Co-expression of p130cas with Crk D242[EL] was as eective as Crk D255 in inducing p130cas hyperphosphorylation (Figure 2b, lanes 2 ± 3). Co-immunoprecipitation experiments with anti-Crk revealed that Crk D242[EL] bound signi®cantly more p130cas than Crk II, indicating that increased tyrosine phopshorylation indeed represents increased association of Crk (Figure 2c). To test the eects of mutating the Abl binding site (Crk W170K D242[EL]) or the Abl phosphorylation site (Crk Y222F D242[EL]) in the context of Crk D242[EL], these mutants were coexpressed with p130cas (Figures 1 and 2D). Whereas both Crk mutants failed to become tyrosine phosphorylated (pCrk), only W170K Crk D242[EL] abrogated the p130cas hyper phosphorylation, suggesting that while the eects of Crk D242[EL] are mediated by carboxyl-terminal alterations, the biological eects of this mutant appear to be coupled to the N-terminal SH3 domain. Effect of C-terminal perturbations on Abl binding to Crk The fact that Crk D242[EL] was strongly phosphorylated on Tyr222 and required the N-terminal SH3 domain to induce p130cas hyperphosphorylation, suggested that augmented c-Abl binding to Crk might be an underlying factor for the eects of this mutant. To investigate this relationship, 293T cells were co-


953

Figure 1 Summary of Crk mutants used in this study. Amino acids 12 ± 125, 138 ± 189, and 243 ± 294 indicate the boundaries of the SH2, N-terminal SH3, and C-terminal SH3 domains, respectively in the Crk II adaptor protein. The SH3 linker region, indicated by amino acids 189 ± 243, is shown in red. The characteristics of individual mutants, as well as their eects on FAK activation (eects on FAK) or on the association of Abl (eects on Abl) are indicated in the right columns. In the lower panel, the putative regions in Crk responsible for FAK and Abl regulation are denoted by solid bars. The sequences in Crk D255 (a) and Crk D242EL (b) at the beginning of the C-terminal SH3 domain refers to R242 VIGKVPNAYDKTA and R242 GGRDSRVRVPTWDP respectively

transfected with c-Abl and various Crk SH2 and SH3 domain mutants as shown in Figure 3, and cellular extracts were immunoprecipitated with anti-Crk (antiRF51) that recognizes the Crk SH2 domain. Immune complexes were collected and analysed for Abl kinase activity using GST Crk 120 ± 225 as a substrate (Feller et al., 1994). Resolution of the amount of Crk protein indicated that the overall levels of Crk immunoprecipitated were similar in all transfected populations (Figure 3 insert, lower panel, and data not shown). As expected, mutation in the N-terminal SH3 domain (W170K Crk) virtually abrogated any Abl activity associated with Crk and mutation in the Crk SH2 (R38K Crk) slightly decreased this amount relative to wild-type Crk. In general, there was greater Abl activity associated with all the C-terminal Crk mutants, particularly Y222F Crk, W276K Crk (C-terminal SH3 domain mutant) and the splice variant c-Crk I, suggesting that the C-terminus of Crk directly in¯uences the binding of Abl, and where relevant, the

phosphorylation of Crk on Tyr222. Interestingly, the Abl activity associated with Crk D242[EL] was only slightly higher than c-Crk II and did not appear to correlate with the extent of Crk Tyr222 phosphorylation, suggesting that Crk D242[EL] eects on p130cas may be mediated in a c-Abl-independent manner. Crk D242[EL] expression results in specific phosphorylation on Tyr397 in FAK The enhanced tyrosine phosphorylation of p130cas observed in Crk D242[EL]-expressing cells might also have resulted from the activation of FAK, since p130cas phosphorylation is known to occur following integrinmediated cell adhesion (Nojima et al., 1995; Polte and Hanks, 1995; Vuori and Ruoslahti, 1995). To investigate a role of FAK in Crk signaling, we expressed wild-type FAK with Crk D242[EL] or various Crk SH2 and SH3 domain mutants (Figure 4). Consistent with the results in Figure 2, tyrosine phosphorylation FAK Oncogene


954

Figure 2 Expression of Crk II in 293T cells is sucient to induce tyrosine phosphorylation of p130cas. (a) 293T cells were cotransfected with expression plasmids for GST-p130cas and the Crk mutants indicated. After 48 h, detergent lysates were collected, normalized for protein, and 15 mg total protein was immunoblotted with anti-pTyr antibodies, and then stripped and reprobed with anti-p130cas, anti-phospho Crk (pCrk) or an anti-Crk SH2 domain speci®c Ab RF51. The molecular weights of known standards (6103) are indicated on the right. (b) Lysates (15 mg protein) from 293T cells transiently transfected with the indicated Crk Cterminal deletion mutants were resolved by SDS ± PAGE and analysed by immunoblotting with anti-pTyr, anti-p130cas, or antipCrk. The localization of Tyr222 phosphorylated Crk is indicated by the bracket. The Crk D255 and Crk D242EL constructs are described in Figure 1. (c) 293T cells were transfected with the indicated plasmids and detergent lysates (500 mg) were immunoprecipitated with anti-Crk, followed by Western blotting with anti-p130cas or anti-Crk. (d) 293T cells were transfected with wild-type Crk II or mutants and p130cas as indicated. The lysates were immunoblotted as in a. Similar levels of Crk proteins were observed in all lanes (not shown)

Figure 3 C-terminal mutants of Crk increase the binding of cAbl to Crk. (a) 293T cells were co-transfected with WT c-Abl and one of the c-Crk expression plasmids indicated. Cellular lysates (500 mg protein) were immunoprecipitated with anti-Crk RF51 antibodies, dividied into equal aliquots, and analysed for Abl kinase activity (Abl kinase assay) using GST Crk 120 ± 225 as a substrate (histogram), or immunoblotted with anti-c-Abl antibody 8E9 or anti-Crk RF51 to determine the amount of protein (inset blots). For the kinase assays, the gels were quanti®ed with a Molecular Dynamics phosphoimagerTM and the data are expressed as the average+standard error of three independent experiments Oncogene

was strictly dependent upon c-Crk II expression (compare lanes 1 versus 7). The tyrosine phosphorylation of FAK is thought to re¯ect catalytic activity of FAK itself since FAK proteins bearing a kinaseinactive mutation (M454 FAK) are not phosphorylated (Richardson et al., 1997, and data not shown). Crkinducible FAK tyrosine phosphorylation was completely abrogated by co-expression with R38K Crk (lane 2), and decreased by co-expression with W170K Crk (lane 3), indicating that the SH2 domain of c-Crk is absolutely essential for this eect. In contrast, mutants in the C-terminus of Crk that include Y222F c-Crk, which abrogates the c-Abl phosphorylation site (Feller et al., 1994), and W276K Crk, which mutates the Cterminal SH3 domain, modestly increased FAK activation compared to wild-type c-Crk II. However, Crk D242[EL] clearly exhibited the strongest eect on the in vivo FAK tyrosine phosphorylation of all the mutants tested under these conditions (anti-pFAK, lane 5). A hallmark of FAK activation is the recruitment of c-Src, via its SH2 domain, to Tyr397 and the subsequent tyrosine phosphorylation of p130cas and paxillin (Hildebrand et al., 1995; Schlaepfer et al., 1994,


955

Figure 4 Eects of Crk C-terminal mutations on FAK activation, and the tyrosine phosphorylation of focal adhesion proteins p130cas and paxillin. (a) 293T cells were cotransfected with FAK and either WT Crk II or one of the c-Crk mutants indicated. The lysates (15 mg protein) were immunoblotted with a generic anti-phospho FAK (pFAK) antibody or with phosphospeci®c anti-FAK antibodies phosphoTyr397 or phosphoTyr925 (Biosource). After stripping, the blots were reprobed with anti-FAK to verify expression. In the lower panels, replicate samples were immunoblotted with anti-pCrk or anti-Crk RF51. (b) 293T cells were cotransfected with GST-epitope tagged p130cas (upper panel) or Flag-epitope tagged paxillin (lower panel) and either WT Crk II (lane 2), or one of the mutants indicated. Lysates (15 mg protein) were immunoblotted with anti-pTyr 4G10, anti-GST, anti-Flag, antipCrk, or anti-RF51 in respective panels. (c) Kinase-dead K295M c-Src inhibits Crk D242EL-mediated paxillin phosphorylation. 293T cells were co-transfected with paxillin, Crk D242EL and either kinase dead K290M c-Abl or K295M c-Src as indicated. Lysates were prepared and proteins (15 mg) immunoblotted with anti-pTyr, anti-Crk, anti-Flag, anti-Src, or anti-Abl to verify expression

1997). Conversely, Tyr925 in FAK creates a binding site for the SH2 domain of Grb2, resulting in Rasdependent ERK activation (Schlaepfer and Hunter, 1997). Using phospho-speci®c anti-FAK antibodies (Figure 4a), Crk-inducible FAK phosphorylation occurred exclusively on Tyr397, with the greatest eects induced by Crk D242[EL], although both Y222F Crk and W276K Crk had a modest potentiating eect. In fact, the pattern of p130cas (Figure 4b, upper panel) and paxillin (Figure 4b, lower panel) phosphorylation induced by the dierent c-Crk mutants closely paralleled the phosphorylation of FAK shown in Figure 2 whereby R38K c-Crk abolished Crk-induced phosphorylation and the C-terminal mutants, particularly Crk D242[EL], strongly stimulated phosphorylation. Consistent with a role for FAK/Src, overexpression of kinase-de®cient K295M c-Src inhibited Crk D242[EL]-mediated phosphorylation of paxillin, although it had no eect on the Crk Tyr222 phosphorylation (Figure 4c, compare anti-pTyr panel with anti-pCrk panel). In contrast, overexpression of kinase-de®cient c-Abl blocked Tyr222 phosphorylation, but not paxillin or p130cas tyrosine phosphorylation (Figure 4c, lane 4). Taken together, these data suggest that the C-terminal region of Crk contains distinct regulatory elements for controlling Abl and FAK signaling pathways.

Expression of Crk D242[EL] increases the number of focal adhesions in NIH3T3 cells Previous studies have shown that FAK activation and focal adhesion biogenesis are closely related events (Giancotti and Ruoslahti, 1999). To ascertain whether expression of Crk D242[E] increased the number of focal adhesions in cells expressing this mutant, a biscistronic retroviral vector encoding c-Crk II or Crk D242[EL], preceded by an internal ribosome entry site (IRES) and the GFP gene, was transiently introduced into NIH3T3 cells by retrovirus infection. Sorting of GFP-positive cells by FACSortTM, followed by Western blotting with anti-GFP and Crk con®rmed that GFPpositive cells co-expressed Crk (Figure 5i). After infection, cells harboring GFP, wild-type c-Crk II, or Crk D242[EL] were plated on ®bronectin-coated coverslips, and stained with either anti-paxillin (to examine the numbers of focal adhesions) or with anti-pTyr antibodies to assess the pTyr content of focal adhesions in GFP-positive cells. Interestingly, the surface area of Crk D242[EL]-expressing cells was approximately two-times that of vector- or wild-type Crk II transfected cells (Figure 5i), which accompanied an increased staining of focal adhesions particularly noted on the ventral surface of the cell (Figure 5e). These focal adhesions exhibited robust anti-pTyr Oncogene


956

staining (Figure 5f), suggesting that Crk D242[EL] increased the stability of mature focal adhesion structures in adherent cells. In contrast, the cells expressing wild-type c-Crk II were more spindle shaped with notable ®lopodia, but had fewer focal adhesions compared to vector (GFP), suggesting that the intact C-terminal SH3 domain may negatively aect the stability of focal adhesions in vivo. Moreover, staining of cells with anti-Crk RF51 antisera showed greater Crk D242[EL] co-localized with focal adhesions relative to wild-type Crk (Figure 5c and e, inserts), suggesting that the eects of Crk D242[EL] may be mediated, in part, by increased binding to focal adhesions. Crk D242[EL]-induced tyrosine phosphorylation of FAK and focal adhesion proteins is counterbalanced by expression of the protein tyrosine phosphatase PEST (PTP ± PEST) The net stability of focal adhesions in cells is balanced by inductive versus destabilizing forces acting on the focal adhesions. Interestingly, in ®broblasts derived from PTP ± PEST-de®cient mice, FAK, paxillin, and p130cas are found in a hyperphosphorylated state, and

PTP ± PEST (7/7) cells display increased numbers of focal adhesions (Angers-Loustau et al., 1999). Moreover, both p130cas and paxillin associate with PTPPEST via speci®c interactions (Cot et al., 1999; Garton and Tonks, 1999), and p130cas has also recently been shown to be a preferred substrate for PTP ± PEST in vivo (Cote et al., 1998; Garton and Tonks, 1999). To investigate whether Crk-mediated p130cas and paxillin hyperphosphorylation were linked by a common pathway to PTP ± PEST, 293T cells were cotransfected with c-Crk II or Crk D242[EL] and PTP ± PEST, and the individual focal adhesion proteins shown in Figure 6. Overexpression of PTP ± PEST completely counteracted both c-Crk II and Crk D242[EL]-induced p130cas (Figure 6a, compare lanes 3 and 4 with 1 and 2 in antipTyr panel) and paxillin phosphorylation (not shown). In contrast, overexpression of PTP ± PEST had no eect on the extent of Crk Tyr222 phosphorylation in either c-Crk II or Crk D242[EL]-expressing cells (Figure 6a, anti-pCrk panel). Since PTP ± PEST binding to focal adhesion proteins has been linked to focal adhesion disassembly and turnover (Garton et al., 1997), we investigated whether Crk D242[EL] might indirectly eect such interactions.

Figure 5 Crk D242EL increases the number of focal adhesions in NIH3T3 cells. Murine NIH3T3 cells were infected with recombinant pCX retrovirus expressing GFP alone (a and b), WT c-Crk II and GFP (c and d), or Crk D242EL and GFP (e and f). Infection eciency was over 80% as judged by green ¯uorescence. Forty-eight hours after infection, the cells were triturated and replated onto ®bronectin-coated coverslips for 6 h, after which they were ®xed, permeabilized, and stained with either anti-paxillin mAb (a,c,e) or anti-pTyr 4G10 (b,d,f) followed by rhodamine-conjugated secondary antisera. Original magni®cation6630 Bar=3 mm. (g) (anti-paxillin) and (h) (GFP ¯uoresence) show Crk D242EL plus and minus expressing cells in the same ®eld for comparison (see arrowhead for GFP non-expressing cell). Note that in GFP-negative cells, the staining for focal adhesion is much weaker. In the (c) and (e) inserts, cells were stained with anti-Crk RF51 Ab to compare Crk expression. Note increased localization to focal adhesions in Crk D242EL expressing cells (e). In (i), the contours of the GFP-positive cells was traced and expressed as the surface area (in pixels)+standard error. At least 10 cells was traced for each measurement. In the i insert, GFP-positive cells were FACSTM-sorted and collected. Detergent lysates were prepared and 20 mg protein immunoblotted with either anti-GFP or anti-Crk Abs to show co-expression of Crk and GFP from the bicistronic vector Oncogene


However, as shown in Figure 6b, expression of Crk D242[EL] did not eect the association between p130cas or paxillin and the endogenous PTP ± PEST (Figure 6b, anti-pTyr panel). No association was observed between FAK and PTP ± PEST (Figure 6b, lanes 6 and 7). Interestingly, when anti-PEST immune complexes were analysed for co-precipitating c-Crk II or Crk D242[EL], only the former was detected, particularly in cells cotransfected with p130cas (Figure 6b, lane 4). Although we found no direct interaction between PTP ± PEST and FAK (Figure 6b, lanes 6 and 7), co-expression of PTP ± PEST completely abrogated Crk D242[EL]-induced FAK hyperphosphorylation (Figure 6c, top panel). These eects appear to be speci®c for PTP ± PEST, since expression of PTEN, which can also dephosphorylate FAK under certain conditions (Tamura et al., 1998, 1999), had no eect on the Crkinducible pathway (Figure 6d). Discussion The Crk II adaptor protein is an SH2/SH3 domain containing protein with the structure SH2/SH3/SH3. While much has been learned about the function of the SH2 and N-terminal SH3 domains via the identi®cation of speci®c interacting proteins, the functions of the C-terminal SH3 domain and the SH3 linker regions of Crk are not well understood. In this study, we present evidence that the C-terminal region of c-Crk II contains negative regulatory elements important for

mediating distinct functions associated with the protein. On the one hand, mutants in the C-terminal SH3 domain (W276K Crk) or deletions in the entire Cterminal SH3 domain, including the splice variant cCrk I, enhance Abl binding to Crk and increase Tyr222 phosphorylation. On the other hand, by analysing a series of Crk C-terminal mutants, we describe an unusual mutant, Crk D242[EL], characterized by a disruption in the SH3 linker/C-terminal SH3 domain boundary that results in FAK activation, the tyrosine phosphorylation of focal adhesion proteins p130cas and paxillin, and increased numbers of focal adhesions in ®broblasts. These results are consistent with the proposed negative regulatory role of the Crk Cterminus (Ogawa et al., 1994), and suggest that the C-terminal region of Crk may contribute to the regulation of distinct tyrosine kinase pathways involving Abl and FAK. How does Crk D242[EL] increase FAK397 hyperphosphorylation? Presently, the answer is unclear, since no biological eectors have been identi®ed to interact with the C-terminal Crk SH3 domain or the SH3 linker region. It is interesting that Y222F Crk and W276K Crk also increased FAK397 phosphorylation, albeit more weakly that Crk D242[EL] possibly suggesting that such activation might result from de-repression of an auto-inhibitory interaction between linker sequences and the C-terminal SH3 domain. In a separate study, Kizaka ± Kondoh and colleages (1996) found that speci®c point mutations, Crk D245R and Crk K254E, also at the beginning of the C-terminal SH3

957

Figure 6 PTP-PEST attenuates Crk-mediated hyperphosphorylation of focal adhesion proteins. (a) 293T cells were cotransfected with GST-tagged p130cas and WT Crk II or Crk D242EL in the presence or absence of PTP ± PEST as indicated. The lysates (15 mg protein) were immunoblotted with the indicated antibodies. (b) Crk D242EL does not eect the interaction between PTP ± PEST and focal adhesion proteins. 293T cells were co-transfected with WT c-Crk II or Crk D242EL and one of the focal adhesion proteins indicated. The lysates (500 mg protein) were immunoprecipitated with anti-PTP ± PEST antibody, followed by immunoblot analysis with a mixture of monoclonal antibodies towards, anti-p130cas, anti-Flag (to detect paxillin), and anti-FAK (top panel) or with antiCrk (lower panel). The amount of Crk proteins in the lysates and the amount of immunoprecipitated PTP ± PEST is indicated in the loading controls. The expression of focal adhesion proteins in each transfection was veri®ed by reprobing the blots with respective antibodies (data not shown). (c,d) PTP-PEST, but not PTEN, attenuates Crk-induced FAK hyperphosphorylation. 293T cells were cotransfected with the combinations of plasmids as in a. Lysates (15 mg) were immunoblotted with the indicated antibodies Oncogene


958

Oncogene

domain, impaired c-Crk II signaling in NRK cells by disrupting EGF signaling to Ras, further illuminating the importance of this region. It is also noteworthy that the sequence homology in the linker is highly conserved in the xenopus, chicken, mouse, and human crk II genes (Reichman et al., 1992). Nine of the 53 amino acids in the linker are Pro, some of which have homology to putative SH3 domain binding proteins, and hence the linker region might represent an independent binding domain in Crk. Although there are no Pro-X-X-Pro-X-Lys motifs in the linker region that conform to known Crk SH3 binding motifs, there is evidence that intramolecular interactions between SH3 domains and a single prolyl residue can act as autoinhibitory modules in the Abl and Src family tyrosine kinases (Barila and Superti-Furga, 1998; Sicheri et al., 1997). Furthermore, six of the nine Pro residues in the c-Crk II linker are conserved in the mouse and human CrkL linker region, even though the sequence and length of the linker in CrkL is quite dierent from Crk II (ten Hoeve et al., 1993). While detailed structural and biochemical studies are warranted to analyse the conformation of the linker with respect to the Crk SH2 and SH3 domains, it will be interesting to observe whether speci®c motifs in the linker region bind directly to either of the SH3 domains of Crk, and whether speci®c Pro substitutions in the Crk II linker eect FAK activation, or enhance binding of Abl to Crk. Such analysis should help elucidate whether the linker region participates in an intramolecular interaction with the SH3 domains and whether this de®nes an independent autoinhibitory module in the Crk protein. Using a variety of carboxyl terminal mutants, we show that virtually all C-terminal mutants tested, including SH3 truncation mutants, W276K Crk, and the splice variant Crk I mimick Y222F Crk and enhance the binding of Abl to the N-terminal Crk SH3 domain. Likewise, several of these mutants which retain Tyr222, including W276K Crk, Crk D255, and Crk D242[EL], also resulted in enhanced phosphorylation on this site, suggesting that the C-terminal SH3 domain/linker region may fold back and impose a negative constraint on the N-terminal SH3 domain. In native Crk II, biochemical and biophysical studies have shown that Crk Tyr222 phosphorylation induces conformational changes in the N-terminal region of Crk, inhibiting both SH2 and N-terminal SH3 domainmediated binding to cellular targets (Escalante et al., 2000; Rosen et al., 1995), although whether this involves a physical interaction between the C-terminus is not known. In the Drosophila (Galletta et al., 1999) and C. elegans (Reddien and Horvitz, 2000) Crk II proteins, there is no conserved Y222 AQP Abl consensus phosphorylation site, and the linker region is conspicuously absent in Pro residues and considerably shorter in length than the mammalian Crk proteins. Thus, we may speculate that Crk Tyr222 phosphorylation could regulate Abl/Crk binding by modulating an interaction between the C-terminal SH3 domain and Pro residues in the linker. However, from

the studies with the Crk D242[EL] mutant, our results imply that the positive regulation of FAK signaling by C-terminal perturbations appear to be independent from the eects of c-Abl and Tyr222 phosphorylation. Evidence to support this includes the fact that: (i) kinase-dead K295M Abl or mutatagenesis of Tyr222 in Crk D242[EL] had no eect on Crk D242[E]-induced FAK or p130cas hyperphosphorylation; (ii) overexpression of PTP-PEST blocked Crk D242[EL] eects on FAK, but not on Crk Tyr222 phosphorylation; and (iii) the amount of Abl associated with dierent Crk mutants did not correlate with the extent of FAK activation. Based upon the series of Crk mutants used in this study, our data suggest that Crk utilizes two distinct regulatory mechanisms to bidirectionally regulate focal adhesion signaling. On the one hand, Crk can transmit focal adhesion signals by conventional SH2/SH3 domain coupling, in which the Crk SH2 domain binds tyrosine phosphorylated focal adhesion proteins (e.g, p130cas and paxillin) and propagates signals through proteins such as C3G, DOCK180, and Abl, that bind to the Crk N-terminal SH3 domain (Kiyokawa et al., 1998a,b). Additionally, as shown here, c-Crk may also feedback and turn o signals that activate FAK by a mechanism under negative regulation by the C-terminal SH3 domain and the SH3 linker region. It is interesting that R38K Crk (the SH2 mutant) abrogated Crk's ability to activate FAK (and p130cas and paxillin) suggesting that direct binding of Crk to focal adhesions may trigger a feedback loop that aects signals upstream to FAK. In this capacity, Crk binding to p130cas or paxillin could act as a thermostat to ®ne-tune the strength of signals transmitted through focal adhesions, the exact setting maintained by cellular factors that either promote or destabilize the Cterminal region. As alluded to above, phosphorylation of Tyr222 may participate in this event, since Y222F Crk mutants (or Y207F CrkL which corresponds to the Y222F mutant in Crk II) increase the in vivo tyrosine phosphorylation of p130cas and paxillin (Senechal et al., 1998). Given that current models predict that integrinmediated adhesion and migration along the extracellular matrix is a dynamic process, requiring reversible tyrosine phosphorylation and dephosphorylation for continuous assembly and disassembly of focal contacts (Giancotti and Ruoslahti, 1999), we posit that molecules like c-Crk II that can regulate upstream and downstream signals in a dynamic equilibrium would be well suited to permit cyclical events required for focal adhesion turnover and plasticity. The capacity of c-Crk D242[EL] to induce hyperphosphorylation of focal adhesion proteins and stabilize focal adhesion structures is remarkably reminescent of the phenotype observed in PTP ± PEST (7/7) ®broblasts in which p130cas, paxillin, and FAK are also found in a hyperphosphorylated state, and focal adhesions in these cells are more numerous throughout the ventral surface of the cell (Angers-Loustau et al., 1999). Functionally, this suggests that the C-terminal region of Crk may function to induce turnover of focal


adhesions, and our data are consistent with a Crk?PTP-PEST?FAK cascade that regulates focal adhesion phosphorylation, although the nature of the crosstalk between Crk and PTP ± PEST is not clear. Mapping studies have shown that PTP ± PEST binding to paxillin and p130cas involves the LIM 3 domain of paxillin (Shen et al., 1998), and the SH3 domain of p130cas (Garton et al., 1996) both of which are also believed to target paxillin and p130cas (via FAK) to focal adhesions. Accordingly, Cot et al. (1999) have proposed a model in which p130cas and paxillin binding to PTP ± PEST mediates relocation of these complexes to the cytoplasm, where they favor dephosphorylation and turnover of focal adhesions. The substrate speci®city for PTP ± PEST has been analysed in vitro and in vivo, and shown to prefer pYDXP-containing sequences contained in the focal adhesion protein p130cas that confer Crk SH2 binding (Cote et al., 1998). Interestingly, the pY222 AQP motif in Crk does not appear to be a relevant substrate for PTP ± PEST, and the nature of the Crk phosphatase is still an open question. Thus, it is likely that other factors contribute to the action of Crk on PTP ± PEST, and future studies should address whether the C-terminal region of c-Crk participates in some way to the regulation of PTP ± PEST activity. To eventually de®ne the functions of Crk in focal adhesion signaling, it will be important to de®ne proteins to which Crk interacts as well as how the domains interact to modulate these interactions. In the present study, we have uncovered a potentially novel regulatory mechanism for Crk II that involves the SH3 linker region/ C-terminal SH3 domain. The results of this study and previous studies support the notion that c-Crk can have both positive and negative roles during focal adhesion signaling by regulating both the binding to, as well as aecting the tyrosine phosphorylation status of FAK, paxillin, and p130cas. Elucidation of the molecular mechanisms and potential activators of focal adhesion formation should facilitate our understanding of the regulation of Crk signaling.

Materials and methods Antibodies An anti-SH2 domain rabbit antibody (RF51), raised against a GST-Crk SH2 polypeptide (Knudsen et al., 1994), was used to detect c-Crk II and C-terminal deletion mutants. AntiPTP ± PEST polyclonal antibody (CSH8 pAb) and antipeptide polyclonal antibodies that recognizes Tyr222 phosphorylated c-Crk II were generous gifts from Nick Tonks (Cold Spring Harbor Laboratories) and Michiyuki Matsuda (National Institutes of Health, Tokyo, Japan) respectively (Garton et al., 1997; Hashimoto et al., 1998). Mouse monoclonal anti-FAK and anti-p130cas antibodies (Transduction Laboratories), mouse anti-phosphotyrosine (pTyr) 4G10 (Upstate Biotechnologies) mouse anti-Flag M2 monoclonal antibody (Kodak), mouse anti-GFP (NE Biolabs), and mouse anti-paxillin monoclonal antibody (Zymed) were purchased from their respective vendors. The phosphospeci®c antibody

rabbit anti-FAK antiserum sampler package was obtained from BioSource.

959

Cell culture The human embryonic kidney (HEK) 293T line and the Bosc23 replication incompetent ecotropic packaging line (RetromaxTM, Imgen) were maintained in Dulbecco's modi®ed Eagle's medium (Gibco ± BRL) containing 10% fetal calf serum (FCS) and 100 U/ml penicillin and streptomycin. NIH3T3 ®broblasts were maintained in DMEM containing 10% calf serum and antibiotics. All cells were incubated in a humidi®ed atmosphere at 378C under 5% CO2. Plasmid constructions, DNA transfections and Crk-expressing virus production cDNAs encoding the tyrosine-speci®c phosphatases PTEN and PTP ± PEST were cloned into the mammalian expression vector pFLAG-CMV-2 (Kodak) and pcDNA respectively, and kindly provided by Dr Maria Georgescu (The Rockefeller University) and Dr M Trembley (McGill University), respectively. Expression plasmids for Flag-tagged paxillin, GST-p130cas, kinase-de®cient (K290M) Abl, and kinasede®cient (K295M) Src have been described previously (Escalante et al., 2000), as have the mammalian pEBB expression plasmids driving expression of c-Crk II, R38K cCrk, W170K c-Crk, and W276K c-Crk (Mayer et al., 1995). The deletion mutants Crk D255, Crk D242, Crk D232, Crk D222 and Crk D212 were generated by polymerase chain reaction (PCR) using synthetic oligonucleotide primers (the forward primer 5'-CCGGGGATCCATGGC CGGGCAGTTCGACTCCGAGG and the respective reverse primers containing termination codons at the indicated positions using the wild-type c-crk II plasmid as a template. Crk D242[E] containing the sequence GGRDSRVVPTWDP or GGRDSRVVPTWDA (generated by site directed mutagenesis) were generated by ligating Crk D242 into pEBB and allowing readthrough until the vector termination codon. The ampli®ed fragments were digested with NotI and BamHI and ligated into the NotI/BamHI-digested pEBB vector. The Crk Y222F and Crk D242 Y222F mutants were generated by sitedirected mutagenesis using PCR and a single base-pair mismatch primers (Stratagene, La Jolla, CA, USA). All mutations were con®rmed by DNA sequencing. For the construction of pCX-c-crk-GFP and pCX c-crk D242[EL] retroviral constructs, pEGB c-Crk or pEGB c-Crk D242[EL] were digested with BamHI/NotI, ®lled with Klenow polymerase, and ligated into EcoRI-digested Klenow-treated pCX vector (gift from Tsuyoshi Agati, Osaka Bioscience Institute). The c-Crk expression in pCX is driven by the viral long terminal repeat (LTR) and a CMV promoter; the green ¯uoresence protein (GFP) is expressed from the same transcript via an internal ribosomal entry site cloned downstream of the c-crk gene. To generate Crk-expressing virus particles, pCX-c-crk-GFP and pCX c-crk D242[EL] plasmid DNA was co-transfected with pCL-Eco plasmid (Imgenex) containing retroviral gag, pol, and Ecotrophic gp70 at a ratio of 1 : 1 into a 50% con¯uent 10 cm tissue culture dish of Bosc23 cells by the Lipofectamine method in serum-free Optim-MEM (Gibco ± BRL). After 4 h, 10 ml of DMEM containing 10% FCS was added, the cells were incubated for 48 h. The virus-containing supernatant was collected, ®ltered through 0.45 mm syringeloaded ®lter (Millipore), and added to 40% con¯uent NIH3T3 ®broblast cells in the presence of 5 mg/ml Polybrene (Sigma). For transient transfections of 293T cells, 0.56106 Oncogene


960

cells were plated in 6 cm tissue culture dishes and transfected with indicated amounts of plasmid DNA using lipofectamine. Forty-eight hours after transfection, cells were washed in icecold phosphate buered saline (PBS) and lysed. For cell sorting, NIH3T3 cells infected with pCX-crk-GFP expression vector and sorted based on green ¯uoresence using FACStarPlus (Benton Dickinson, Inc, San Jose, CA, USA). Pools of GFP positive and GFP negative cells were collected, and analysed by Western blotting with anti-Crk and anti-GFP antibodies to demonstrate co-expression. Cell lysis, immunoprecipitations and Western blotting Cells were lysed in ice-cold HNTG buer (20 mM HEPES, 150 mM NaCl, 10% glycerol, 1% Triton-X-100 pH 7.5) containing a mix of protease and phosphatase inhibitors; 1 mM Na2VO4, 0.1 mM Na2MoO4 1% Aprotonin, 1 mM phenyl methyl sulfonyl ¯uride (PMSF) and 5 mg/ml Leupeptin. Cells were incubated on ice for 15 min, and detergent lysates were clari®ed by centrifugation in a microcentrifuge at 13 000 r.p.m. for 5 min. Protein concentrations were normalized by Bio-Rad protein assay and equivalent amounts of protein lysate were boiled for 5 min in sodium dodecyl sulfate (SDS) containing sample buer, loaded on an 8 ± 10% SDSpolyacylamide gel electrophoresis (PAGE), and electrophoresed by standard protocols. For co-immunoprecipitation assays, equivalent amounts of cellular protein (500 mg) was incubated with primary antibody for 3 h at 48C, followed by incubation with Protein A Sepharose or rabbit anti-mouse passi®ed Protein A Sepharose for 1 h (Pharmacia). The beads were washed three times in low detergent (0.1% Triton X100) HNTG lysis buer prior to electrophoresis. For Western blotting, cell lysates or immunocomplexes were separated on SDS ± PAGE gels and transferred to PVDF membranes (Millipore). The membranes were blocked with Tris-buered saline (TBS) containing 1% BSA and then incubated for 3 h at room temperature with one of the following antibodies: anti-Crk RF51, anti-PTP-PEST, antiphosphotyrosine 4G10 (0.1 mg/ml), anti-Flag (0.1 mg/ml), anti-P130cas (0.25 mg/ml) or anti-paxillin (0.1 mg/ml) all at 1 : 1000 ®nal dilutions. Membranes were washed three times with TBS containing 0.02% Tween 20 and then incubated for 1 h with horseradish peroxidase (HRP)-conjugated secondary antibodies (Jackson Laboratories, Bar Harbor, ME, USA) at 1 : 5000 dilution (0.02 mg/ml). Immunoreactive protein bands were detected by enhanced chemiluminescence (ECL) reagents (Amersham, Arlington Heights, IL, USA) diluted 1 : 10 with distilled water. Abl kinase assays Immune complex kinase assays for c-Abl were performed using [g-32P]ATP and GST-Crk 120 ± 225 as an exogenous

substrate as described previously (Escalante et al., 2000). After 20 min at room temperature, kinase reactions were terminated by boiling in SDS sample buer prior to SDS ± PAGE. Gels ®xed in 50% methanol/10% acetic acid were subjected to autoradiography. Quanti®cation of kinase activity was performed using a Molecular Dynamics Phospho-Imager. Fluorescent labeling of cells Glass coverslips were coated with 50 mg/ml human ®bronectin (Becton Dickinson) in PBS for 60 min at room temperature; excess matrix protein was then removed by rinsing with PBS. Forty-eight hours after infection of NIH3T3 cells, cells were triturated in PBS-containing 0.05% Trypsin and 0.5 mM EDTA, and replated onto 5 mg/ cm2 ®bronectin coated glass coverslips in 12-well tissue culture dish for 12 h in DMEM 10% calf serum. Before labeling, cells were ®xed in PBS-containing 3% paraformaldehyde for 30 min, washed, and permeabilized with 0.2% Triton X-100 for 5 min to permit antibody penetration. After permeabilization, sites were blocked by incubation with PBScontaining 50 mM Glycine (10 min) followed by incubation in PBS-containing 0.25% Gelatin (Merck) in PBS for an additional 10 min at room temperature. Primary antibodies (anti-paxillin (Zymed), anti-phospho-Tyr 4G10, or anti-Crk RF51) were diluted 1 : 200 in PBS/gelatin and 50 ml droplets pipetted on para®lm. Coverslips were incubated inverted with primary antibodies for 45 min, extensively washed, and further incubated with 1/200-diluted rhodamine-conjugated goat anti-mouse secondary antibody for an additional 30 min. After four washes, cells were mounted on glass slides with ProLong antifade reagent (Molecular Probes Inc. Eugene, OR, USA). Immuno¯uorescent staining was analysed using Nikon E800 Eclipse ¯uorescence microscope attached to a digital camera and ®les were collected and saved as JPEG ®les. The surface area of the cells was calculated by tracing the contour of the cells, and integrating the area using Adobe Photoshop software.

Acknowledgments We would like to thank Tomoyuki Shishido and Hidesaburo Hanafusa for discussing unpublished results, and Jong-Il Kim for helpful discussions. We also thank Paul Kaloudis for his assistance in the digital imaging of the immuno¯uoresence studies and Jason Ptacek for help in the preparation of ®gures. This work was supported by Public Health Service Awards to RB Birge (RO1 GM55760) and from a Muscular Dystrophy Association (MDA) grant.

References Altun-Gultekin ZF, Chandriani S, Bougeret C, Ishizaki T, Narumiya S, de Graaf P, Van Bergen en Henegouwen P, Hanafusa H, Wagner JA and Birge RB. (1998). Mol. Cell. Biol., 18, 3044 ± 3058. Angers-Loustau A, Cote JF, Charest A, Dowbenko D, Spencer S, Lasky LA and Tremblay ML. (1999). J. Cell. Biol., 144, 1019 ± 1031. Barila D and Superti-Furga G. (1998). Nat. Genet., 18, 280 ± 282. Oncogene

Birge RB, Fajardo JE, Reichman C, Shoelson SE, Songyang Z, Cantley LC and Hanafusa H. (1993). Mol. Cell. Biol., 13, 4648 ± 4656. Birge RB, Knudsen BS, Besser D and Hanafusa H. (1996). Genes Cells, 1, 595 ± 613. Cheresh DA, Leng J and Klemke RL. (1999). J. Cell. Biol., 146, 1107 ± 1116. Cot JF, Turner CE and Tremblay ML. (1999). J. Biol. Chem., 274, 20550 ± 20560.


Cote JF, Charest A, Wagner J and Tremblay ML. (1998). Biochemistry, 37, 13128 ± 13137. Dol® F, Garcia-Guzman M, Ojaniemi M, Nakamura H, Matsuda M and Vuori K. (1998). Proc. Natl. Acad. Sci. USA, 95, 15394 ± 15399. Escalante M, Courtney J, Chin WG, Teng KK, Kim JI, Fajardo JE, Mayer BJ, Hempstead BL and Birge RB. (2000). J. Biol. Chem., 275, 24787 ± 24797. Feller SM, Knudsen B and Hanafusa H. (1994). EMBO J., 13, 2341 ± 2351. Feller SM, Posern G, Voss J, Kardinal C, Sakkab D, Zheng J and Knudsen BS. (1998). J. Cell. Physiol., 177, 535 ± 552. Galletta BJ, Niu XP, Erickson MR and Abmayr SM. (1999). Gene, 228, 243 ± 252. Garton AJ, Burnham MR, Bouton AH and Tonks NK. (1997). Oncogene, 15, 877 ± 885. Garton AJ, Flint AJ and Tonks NK. (1996). Mol. Cell. Biol., 16, 6408 ± 6418. Garton AJ and Tonks NK. (1999). J. Biol. Chem., 274, 3811 ± 3818. Giancotti FG and Ruoslahti E. (1999). Science, 285, 1028 ± 1032. Hasegawa H, Kiyokawa E, Tanaka S, Nagashima K, Gotoh N, Shibuya M, Kurata T and Matsuda M. (1996). Mol. Cell. Biol., 16, 1770 ± 1776. Hashimoto Y, Katayama H, Kiyokawa E, Ota S, Kurata T, Gotoh N, Otsuka N, Shibata M and Matsuda M. (1998). J. Biol. Chem., 273, 17186 ± 17191. Hildebrand JD, Schaller MD and Parsons JT. (1995). Mol. Biol. Cell., 6, 637 ± 647. Kiyokawa E, Hashimoto Y, Kobayashi S, Sugimura H, Kurata T and Matsuda M. (1998a). Genes Dev., 12, 3331 ± 3336. Kiyokawa E, Hashimoto Y, Kurata T, Sugimura H and Matsuda M. (1998b). J. Biol. Chem., 273, 24479 ± 24484. Kizaka-Kondoh S, Matsuda M and Okayama H. (1996). Proc. Natl. Acad. Sci. USA, 93, 12177 ± 12182. Knudsen BS, Feller SM and Hanafusa H. (1994). J. Biol. Chem., 269, 32781 ± 32787. Ling P, Yao Z, Meyer CF, Wang XS, Oehrl W, Feller SM and Tan TH. (1999). Mol. Cell. Biol., 19, 1359 ± 1368. Matsuda M, Hashimoto Y, Muroya K, Hasegawa H, Kurata T, Tanaka S, Nakamura S and Hattori S. (1994). Mol. Cell. Biol., 14, 5495 ± 5500. Matsuda M and Kurata T. (1996). Cell Signal, 8, 335 ± 340. Mayer BJ, Hamaguchi M and Hanafusa H. (1988). Cold Spring Harb. Symp. Quant. Biol., 53, 907 ± 914. Mayer BJ, Hirai H and Sakai R. (1995). Curr. Biol., 5, 296 ± 305. Mochizuki N, Ohba Y, Kobayashi S, Otsuka N, Graybiel AM, Tanaka S and Matsuda M. (2000). J. Biol. Chem., 275, 12667 ± 12671.

Nojima Y, Morino N, Mimura T, Hamasaki K, Furuya H, Sakai R, Sato T, Tachibana K, Morimoto C, Yazaki Y, et al. (1995). J. Biol. Chem., 270, 15398 ± 15402. Ogawa S, Toyoshima H, Kozutsumi H, Hagiwara K, Sakai R, Tanaka T, Hirano N, Mano H, Yazaki Y and Hirai H. (1994). Oncogene, 9, 1669 ± 1678. Oktay M, Wary KK, Dans M, Birge RB and Giancotti FG. (1999). J. Cell. Biol., 145, 1461 ± 1469. Polte TR and Hanks SK. (1995). Proc. Natl. Acad. Sci. USA, 92, 10678 ± 10682. Reddien PW and Horvitz HR. (2000). Nat. Cell. Biol., 2, 131 ± 136. Reichman CT, Mayer BJ, Keshav S and Hanafusa H. (1992). Cell Growth Dier., 3, 451 ± 460. Richardson A, Malik RK, Hildebrand JD and Parsons JT. (1997). Mol. Cell. Biol., 17, 6906 ± 6914. Rosen MK, Yamazaki T, Gish GD, Kay CM, Pawson T and Kay LE. (1995). Nature, 374, 477 ± 479. Sakai R, Iwamatsu A, Hirano N, Ogawa S, Tanaka T, Mano H, Yazaki Y and Hirai H. (1994). EMBO J., 13, 3748 ± 3756. Schlaepfer DD, Broome MA and Hunter T. (1997). Mol. Cell. Biol., 17, 1702 ± 1713. Schlaepfer DD, Hanks SK, Hunter T and van der Geer P. (1994). Nature, 372, 786 ± 791. Schlaepfer DD and Hunter T. (1997). J. Biol. Chem., 272, 13189 ± 13195. Senechal K, Heaney C, Druker B and Sawyers CL. (1998). Mol. Cell. Biol., 18, 5082 ± 5090. Shen Y, Schneider G, Cloutier JF, Veillette A and Schaller MD. (1998). J. Biol. Chem., 273, 6474 ± 6481. Sicheri F, Moare® I and Kuriyan J. (1997). Nature, 385, 602 ± 609. Songyang Z, Shoelson SE, McGlade J, Olivier P, Pawson T, Bustelo XR, Barbacid M, Sabe H, Hanafusa H, Yi T, et al. (1994). Mol. Cell. Biol., 14, 2777 ± 2785. Tamura M, Gu J, Danen EH, Takino T, Miyamoto S and Yamada KM. (1999). J. Biol. Chem., 274, 20693 ± 20703. Tamura M, Gu J, Matsumoto K, Aota S, Parsons R and Yamada KM. (1998). Science, 280, 1614 ± 1617. Tanaka S and Hanafusa H. (1998). J. Biol. Chem., 273, 1281 ± 1284. Tanaka S, Morishita T, Hashimoto Y, Hattori S, Nakamura S, Shibuya M, Matuoka K, Takenawa T, Kurata T, Nagashima K, et al. (1994). Proc. Natl. Acad. Sci. USA, 91, 3443 ± 3447. ten Hoeve J, Morris C, Heisterkamp N and Groen J. (1993). Oncogene, 8, 2469 ± 2474. Vuori K and Ruoslahti E. (1995). J. Biol. Chem., 270, 22259 ± 22262. Wu X, Knudsen B, Feller SM, Zheng J, Sali A, Cowburn D, Hanafusa H and Kuriyan J. (1995). Structure, 3, 215 ± 226.

961

Oncogene