[Frontiers in Bioscience 2, d501-518, October 15, 1997]

	[Frontiers in Bioscience 2, d501-518, October 15, 1997] Reprints PubMed CAVEAT LECTOR
	ROLE OF c-Src TYROSINE KINASE IN EGF-INDUCED MITOGENESIS Allison P. Belsches, Michelle D. Haskell, and Sarah J. Parsons Department of Microbiology and Cancer Center, Box 441, Health Sciences Center, University of Virginia, Charlottesville, VA 22908 Received 10/7/97 Accepted 10/14/97 4. c-Src AND THE CYTOSKELETON Of the c-Src substrates or binding proteins that have been identified in the 10T model, all are conceptually connected to the actin cytoskeleton (figure 3). These include, p125FAK, p130CAS, p190RhoGAP, and cortactin. Within cells, P125FAK and P130CAS localize to focal adhesions, which serve rearrangement involving all three of these substrates, but the effect of tyrosine phosphorylation is uncertain at this time as anchoring sites for the internal actin cytoskeleton to the external substratum through integrins. P190RhoGAP regulates the small GTP binding protein, Rho, which in turn regulates actin stress fiber assembly and disassembly (108). Cortactin localizes to cortical actin in intact cells functions of c-Src in mitogenesis are mediated through the actin cytoskeleton. How c-Src signals via its substrates to contribute to the mitogenic response is a subject under intense investigation in our and other laboratories. A short summary of our current understanding of the functions and regulation of these proteins by c-Src follows (figure 3). Figure 3. c-Src phosphorylates proteins that are involved in cytoskeletal organization. c-Src phosphorylates cortactin, p190RhoGAP, and p130CAS. Cortactin is found associated with cortical actin at peripheral sites within the cell. p130CAS is found associated with focal adhesion kinase in focal adhesions. p190RhoGAP stimulates hydrolysis of Rho-GTP to Rho-GDP leading to actin stress fiber breakdown. Stimulation of C3H10T1/2 cells with EGF stimulates lipid metabolism and actin 4.1 Focal adhesion kinase P125FAK is a cytoplasmic tyrosine kinase that localizes to focal adhesions and contributes to the processes of integrin-mediated cell spreading and migration (reviewed in 109). Both of these processes involve continual remodeling of the actin cytoskeleton. Abundant evidence links c-Src to these processes, but its role (especially in cell spreading) may be largely non-catalytic, since defects in spreading of fibroblasts derived from c-Src null mice can be restored by the SH2 and SH3 domains but not by the catalytic domain of c-Src (110). Motility requires the coordinated activation of both growth factor and adhesion receptor signaling (111), in ways that are poorly understood. Since c-Src is required for the action of several growth factors that induce motility (such as EGF and PDGF) (109), c- Src/FAK interactions may be important in this process as well. How FAK participates in EGF-dependent mitogenic signaling is unexplored at this point, except perhaps to function as an adherence checkpoint. Currently, our data support only a minor role for c-Src/FAK interactions in EGF-induced growth. In unstimulated 10T½ cells the basal level of FAK tyrosyl phosphorylation is only slightly elevated in c-Src overexpressors as compared to controls (1-1.5 fold), and no increase in FAK tyrosyl phosphorylation is observed within 30 min of EGF stimulation. These findings suggest that FAK may be only a weak substrate for c-Src in 10T½ fibroblasts and is contraindicatory of a role for FAK in mitogenic potentiation by c-Src. However, if the functional interaction between FAK and c-Src is largely non-catalytic, one would not expect FAK tyrosyl phosphorylation to be affected much by c-Src overexpression, thereby leaving the questions unanswered as to whether FAK/c-Src interactions contribute to mitogenesis and further, whether adhesion/motility are coupled to mitogenesis. 4.2 p130CAS P130CAS was first identified as a highly tyrosyl-phosphorylated protein in cells transformed by a variety of oncogenes (112-114) and in normal cells following integrin-engagement (115-117) and stimulation with mitogenic neuropeptides, such as bombesin, vasopressin, and endothelin (118, 119). In the latter two processes, an intact actin cytoskeleton is required for phosphorylation of CAS. c-Src has also been implicated in integrin-mediated events and mitogenesis, and in fact, tyrosine phosphorylation of p130CAS in response to adhesion is largely dependent upon c-Src (117, 120). That p130CAS may be an integral part of the action of c-Src in EGF-induced mitogenesis is suggested by our finding that p130CAS co-precipitates with c-Src in 10T½ cells overexpressing c-Src and that its tyrosine phosphorylation is elevated in these cells as compared to control cells (N. Sevilir-Williams, J.-H. Chang, and S. Parsons, unpublished). Tyrosine phosphorylation of c-Src-associated p130CAS is not altered by EGF, confirming that it is a preferred substrate of c-Src. How p130CAS participates in c-Src-mediated EGF signaling, however, is not clear. Recent evidence indicates that p130CAS functions as an adapter molecule, binding a number of signaling molecules that participate in cell adhesion, such as p125FAK (121) and PTP-PEST (122). Thus, the link between cell adhesion, the actin cytoskeleton, and mitogenesis is repeated, and the common involvement of c-Src in both processes suggests that c-Src may be a critical factor that links the two. 4.3 p190RhoGAP P190RhoGAP was first identified as a tyrosine phosphorylated protein that co-precipitated with p120RasGAP from v-Src transformed Rat-2 cells (90). The p190RhoGAP molecule has an amino-terminal GTPase domain that binds GTP (123, 124) and a carboxyl-terminal GTPase-activating domain (GAP) that is specific for small GTP-binding proteins of the Rho family (125). Rho in the GTP-bound form stimulates stress fiber formation while Rho in the GDP-bound form permits actin disassembly (108), a fact that functionally links p190RhoGAP to the actin cytoskeleton. The two functional domains of p190RhoGAP are separated by an 801 amino acid middle region that contains tyrosine residues 1087 and 1105. Both these residues are found in a motif (YXXPXD) thought to mediate in vitro binding to the SH2 domains of p120RasGAP (126). Results from transfection experiments suggest that p190 has tumor suppressor characteristics, in that it can inhibit Ras transformation of NIH3T3 cells. The GTP binding domain is thought to mediate this suppressive activity (127). Overexpression of c-Src in 10T cells results in an increase in the basal level of tyrosine phosphorylation of p190RhoGAP that is not further increased by EGF stimulation, which indicates that p190RhoGAP is a preferred substrate of c-Src and not the EGF receptor. Upon stimulation of normal murine fibroblasts by EGF, p190RhoGAP undergoes a rapid and transient redistribution from a diffuse cytoplasmic localization into concentric arcs that radiate away from the nucleus with a time course that mimics EGF-stimulated actin dissolution (128). Overexpression of wt c-Src expands the window of time in which EGF-induced p190RhoGAP arc formation and actin dissolution occurs, and overexpression of kinase defective c-Src contracts that window (128). These results implicate a role for tyrosine phosphorylated p190RhoGAP in regulating cytoskeletal reorganization, possibly by inactivating Rho. Several questions are raised by these experiments. Since the results are correlative, one question is whether tyrosine phosphorylation of p190 is required for the effects of c-Src overexpression on EGF-induced rearrangement of both p190 and actin. The second is related and asks whether or not the changes observed in the actin cytoskeleton that occur within 30 min after EGF stimulation are necessary for the two-five fold increase in EGF-induced DNA synthesis that is observed at 21 hours after addition of EGF in fibroblasts that overexpress c-Src (30). These questions are currently being addressed experimentally. Recently, our laboratory has characterized the complexity and relative levels of tyrosine phosphorylation on p190 from different c-Src overexpressors (Roof et al., 1997, in preparation). Phosphotryptic peptide analysis of endogenous p190 has uncovered only two tyrosine phosphorylated polypeptides in the molecule, one peptide containing Y1105 (located in the middle domain), and another as yet unidentified minor peptide. This pattern of phosphorylation holds true for p190 from control cells and from wt c-Src overexpressors. That c-Src is directly responsible for phosphorylating Y1105 in vivo is suggested by an increase in the level of phosphorylated Y1105 upon c-Src overexpression, a decrease in phosphorylation of Y1105 in cells overexpressing a kinase-deficient c-Src and phosphorylation of Y1105 by purified c-Src in an in vitro kinase assay (Roof et al. 1997, in preparation). One presumed function of tyrosine phosphorylation of p190RhoGAP is to enhance its interaction with p120RasGAP (84, 129, 130). Using deletion and site directed mutagenesis, Hu and Settleman (130) have proposed a model in which phosphorylation of both Y1087 and Y1105 are necessary for the interaction of p190RhoGAP with the two SH2 domains of p120RasGAP. In this model, the engagement of the two SH2 domains of p120RasGAP with Y1087 and Y1105 in p190RhoGAP induces a conformational change in the p120RasGAP molecule which exposes the RasGAP SH3 domain for interaction with its binding partners. This model represents a mechanism whereby tyrosine phosphorylation of p190 can regulate SH3 domain interactions of p120RasGAP. The data from our laboratory support a role for tyrosine phosphorylation of p190RhoGAP in enhancing its interaction with p120RasGAP but suggest that interactions other than the p120RasGAP SH2 domain/p190RhoGAP pTyr interaction may also be important. Phosphotryptic peptide mapping and phosphoamino acid analyses have been used to compare the level of tyrosine phosphorylation of p190 in control, wt c-Src overexpressors, and kinase deficient c-Src overexpressors at both Y1105 and the unidentified site. These comparisons have revealed that the level of phosphorylation of Y1105 in the wt c-Src overexpressors is increased approximately seven fold above that of control cells and decreased approximately six fold in the kinase deficient c-Src overexpressors relative to control cells. In contrast, the level of tyrosine phosphorylation at the unidentified site remains constant and below the level of Y1105 phosphorylation in all cell lines (approximately 25% of the level of Y1105 phosphorylation in control cells), irrespective of either c-Src overexpression or EGF addition (Roof et al. in preparation). The fact that we have been unable to identify two, equally phosphorylated tyrosines in p190 brings into question the dual pTyr/SH2 model for complex formation between p190 and p120RasGAP. Further questions are raised by the finding that the amount of p120RasGAP that is immunoprecipitated with p190RhoGAP from the cells overexpressing wt c-Src or kinase inactive c-Src is not consistent with the overall level of p190 tyrosine phosphorylation observed in these two cell lines. For instance, the amount of p120RasGAP that co-immunoprecipitates with p190RhoGAP from c-Src overexpressors is less than expected, while the amount of p120 in complex with p190 from kinase deficient c-Src overexpressors is greater than expected, given the striking differences in the extent of p190 tyrosine phosphorylation in the two cell lines. These data suggest that factors in addition to tyrosine phosphorylation may contribute to p190RhoGAP/p120RasGAP interactions and that tyrosine phosphorylation of p190RhoGAP may regulate another function of p190RhoGAP independently of its interaction with p120RasGAP. In fact, Koch et al. (131) have indicated that SH2 domains can interact with phosphoserine-containing peptides, which implies that there may be a role for phosphorylated serines in p190RhoGAP/p120RasGAP interactions. Another possibility is that tyrosine phosphorylation of p190RhoGAP regulates the two known functional activities ascribed to p190RhoGAP, namely, the GTP-binding and GAP activities. Preliminary data from our lab has shown that tyrosine phosphorylation of p190RhoGAP inhibits the in vitro binding of GTP to the N-terminal domain of p190 (Roof, R. and S. Parsons, unpublished observation). In conclusion, it appears that the mechanism of interaction between p190RhoGAP and p120RasGAP is more complicated than previously anticipated and that the consequences of the interaction await further experimentation before the exact function of the interaction between these two proteins is delineated. 4.4 Cortactin The v-Src and c-Src substrate p75/p80/85 cortactin is an actin-binding protein that consists of five tandem repeats in the N-terminus and a SH3 domain in the extreme C-terminus of the molecule (85-87). Both structural features of cortactin are found in a number of cytoskeletal proteins, and indeed the N-terminal repeats of cortactin mediate binding of the molecule to actin (87). Phosphoamino acid analysis of cortactin from normal chick embryo cells indicates that the protein is phosphorylated on serine and threonine residues, whereas in chick embryo cells transformed with activated c-Src (527F), cortactin is also phosphorylated on tyrosine (86). In normal murine 10T fibroblasts, cortactin has a low level of tyrosine phosphorylation that is increased upon overexpression of c-Src. (85). In addition, EGF stimulates a time-dependent increase in the level of tyrosine phosphorylation of cortactin that is further enhanced by overexpression of c-Src (66, 85). Thus, cortactin may be a substrate of the EGF receptor as well as of c-Src, or c-Src may mediate EGF-induced phosphorylation of cortactin. The observation that increased tyrosine phosphorylation of cortactin is seen in Csk-deficient cells favors the notion that c-Src and/or its family members are responsible for phosphorylating cortactin (132, 133). Interestingly, two phases of EGF-induced cortactin tyrosine phosphorylation can be observed in 10T cells, one occuring within 2-10 min following stimulation and another occurring later in G1, with the maximum level being reached approximately nine hours post-treatment. The level of cortactin tyrosine phosphorylation in both phases is increased by c-Src overexpression. Indirect immunofluorescence microscopy of 10T cells reveals that cortactin is localized within the cytoplasm to punctate sites that are concentrated around the nucleus (85). A significant amount of cortactin is also co-localized with actin in the plasma membrane and at peripheral adhesion sites (85). This pattern of subcellular distribution is not altered upon EGF treatment or c-Src overexpression. However, in cells transformed with activated c-Src (527F), cortactin collapses into rosettes or podosomes, which are sites of membrane/substratum interactions (86). These studies are suggestive of a role for peripheral cortactin in maintaining cortical actin integrity and membrane association, but the role of intracellular cortactin is unknown. Furthermore, the role that tyrosine phosphorylation plays in regulating cortactin function is not clear. Recent studies have shown that in vitro tyrosine phosphorylation of cortactin results in a dramatic decrease in F actin cross-linking activity which indicates a role for cortactin in regulating actin cytoskeletal reorganization (134). 4.5 Sam68 Although Sam68 is not a known cytoskeleton-associated protein, it is included in this section, because it is a well-documented substrate of c-Src. The activity of c-Src increases when cells enter mitosis, suggesting the existence of mitosis-specific substrates (135). Sam68 (Src-associated in mitosis) is a 68 kDa RNA-binding protein that is both a c-Src binding protein and a c-Src substrate (figure 1). Sam68 was discovered simultaneously by two laboratories that were comparing pTyr-containing proteins from cells arrested in mitosis with those found in asynchronous cells (136, 137). These two groups found that in mitotic cells but not in asynchronous cells a heavily tyrosine phosphorylated protein of 68 kDa can be co-immunoprecipitated with c-Src. The apparently high affinity interaction of Sam68 with c-Src requires both the SH2 and SH3 domains of Src, although it appears that the association is mediated mainly through the Src SH2 domain (136, 137). The functional consequence of phosphorylation of Sam68 is currently unknown, although it has been reported that phosphorylation of Sam68 decreases its RNA binding ability (138). Recent work by Barlat et al. (139) has shown that a naturally occurring isoform of Sam68 exists in NIH3T3 cells. This isoform has a portion of the RNA-binding domain deleted (amino acid residues 170-208) and is called Sam68deltaKH. Sam68deltaKH is expressed at growth arrest in confluent cells but is absent from cells that have been transformed by activated c-Src (Y527F) (139). Interestingly, transfection of NIH3T3 cells with a myc-tagged Sam68deltaKH leads to a 50% decrease in BrdU incorporation in response to serum stimulation, while transfection of either myc-tagged Sam68 or vector alone had no effect on BrdU incorporation (139). Furthermore, co-expression of myc-tagged Sam68 with myc-tagged Sam68deltaKH reversed the myc-tagged, Sam68deltaKH-induced inhibition of BrdU incorporation. These results indicate that the RNA-binding domain of Sam68 is important for progression through the cell cycle. However, it is unclear what effect tyrosine phosphorylation of Sam68 has on cell cycle progression. Whether the deletion of the KH domain affects the ability of SAM68 to become tyrosine phosphorylated is not known, although the portion of the KH domain that is deleted does not contain any tyrosine residues.

[Frontiers in Bioscience 2, d501-518, October 15, 1997]
Reprints
PubMed
CAVEAT LECTOR

ROLE OF c-Src TYROSINE KINASE IN EGF-INDUCED MITOGENESIS

Allison P. Belsches, Michelle D. Haskell, and Sarah J. Parsons

Department of Microbiology and Cancer Center, Box 441, Health Sciences Center, University of Virginia, Charlottesville, VA 22908

Received 10/7/97 Accepted 10/14/97

4. c-Src AND THE CYTOSKELETON

Of the c-Src substrates or binding proteins that have been identified in the 10T model, all are conceptually connected to the actin cytoskeleton (figure 3). These include, p125FAK, p130CAS, p190RhoGAP, and cortactin. Within cells, P125FAK and P130CAS localize to focal adhesions, which serve rearrangement involving all three of these substrates, but the effect of tyrosine phosphorylation is uncertain at this time as anchoring sites for the internal actin cytoskeleton to the external substratum through integrins. P190RhoGAP regulates the small GTP binding protein, Rho, which in turn regulates actin stress fiber assembly and disassembly (108). Cortactin localizes to cortical actin in intact cells functions of c-Src in mitogenesis are mediated through the actin cytoskeleton. How c-Src signals via its substrates to contribute to the mitogenic response is a subject under intense investigation in our and other laboratories. A short summary of our current understanding of the functions and regulation of these proteins by c-Src follows (figure 3).

Figure 3. c-Src phosphorylates proteins that are involved in cytoskeletal organization. c-Src phosphorylates cortactin, p190RhoGAP, and p130CAS. Cortactin is found associated with cortical actin at peripheral sites within the cell. p130CAS is found associated with focal adhesion kinase in focal adhesions. p190RhoGAP stimulates hydrolysis of Rho-GTP to Rho-GDP leading to actin stress fiber breakdown. Stimulation of C3H10T1/2 cells with EGF stimulates lipid metabolism and actin

4.1 Focal adhesion kinase

P125FAK is a cytoplasmic tyrosine kinase that localizes to focal adhesions and contributes to the processes of integrin-mediated cell spreading and migration (reviewed in 109). Both of these processes involve continual remodeling of the actin cytoskeleton.

Abundant evidence links c-Src to these processes, but its role (especially in cell spreading) may be largely non-catalytic, since defects in spreading of fibroblasts derived from c-Src null mice can be restored by the SH2 and SH3 domains but not by the catalytic domain of c-Src (110). Motility requires the coordinated activation of both growth factor and adhesion receptor signaling (111), in ways that are poorly understood. Since c-Src is required for the action of several growth factors that induce motility (such as EGF and PDGF) (109), c- Src/FAK interactions may be important in this process as well. How FAK participates in EGF-dependent mitogenic signaling is unexplored at this point, except perhaps to function as an adherence checkpoint. Currently, our data support only a minor role for c-Src/FAK interactions in EGF-induced growth.

In unstimulated 10T½ cells the basal level of FAK tyrosyl phosphorylation is only slightly elevated in c-Src overexpressors as compared to controls (1-1.5 fold), and no increase in FAK tyrosyl phosphorylation is observed within 30 min of EGF stimulation. These findings suggest that FAK may be only a weak substrate for c-Src in 10T½ fibroblasts and is contraindicatory of a role for FAK in mitogenic potentiation by c-Src. However, if the functional interaction between FAK and c-Src is largely non-catalytic, one would not expect FAK tyrosyl phosphorylation to be affected much by c-Src overexpression, thereby leaving the questions unanswered as to whether FAK/c-Src interactions contribute to mitogenesis and further, whether adhesion/motility are coupled to mitogenesis.

4.2 p130CAS

P130CAS was first identified as a highly tyrosyl-phosphorylated protein in cells transformed by a variety of oncogenes (112-114) and in normal cells following integrin-engagement (115-117) and stimulation with mitogenic neuropeptides, such as bombesin, vasopressin, and endothelin (118, 119). In the latter two processes, an intact actin cytoskeleton is required for phosphorylation of CAS. c-Src has also been implicated in integrin-mediated events and mitogenesis, and in fact, tyrosine phosphorylation of p130CAS in response to adhesion is largely dependent upon c-Src (117, 120). That p130CAS may be an integral part of the action of c-Src in EGF-induced mitogenesis is suggested by our finding that p130CAS co-precipitates with c-Src in 10T½ cells overexpressing c-Src and that its tyrosine phosphorylation is elevated in these cells as compared to control cells (N. Sevilir-Williams, J.-H. Chang, and S. Parsons, unpublished). Tyrosine phosphorylation of c-Src-associated p130CAS is not altered by EGF, confirming that it is a preferred substrate of c-Src. How p130CAS participates in c-Src-mediated EGF signaling, however, is not clear. Recent evidence indicates that p130CAS functions as an adapter molecule, binding a number of signaling molecules that participate in cell adhesion, such as p125FAK (121) and PTP-PEST (122). Thus, the link between cell adhesion, the actin cytoskeleton, and mitogenesis is repeated, and the common involvement of c-Src in both processes suggests that c-Src may be a critical factor that links the two.

4.3 p190RhoGAP

P190RhoGAP was first identified as a tyrosine phosphorylated protein that co-precipitated with p120RasGAP from v-Src transformed Rat-2 cells (90). The p190RhoGAP molecule has an amino-terminal GTPase domain that binds GTP (123, 124) and a carboxyl-terminal GTPase-activating domain (GAP) that is specific for small GTP-binding proteins of the Rho family (125). Rho in the GTP-bound form stimulates stress fiber formation while Rho in the GDP-bound form permits actin disassembly (108), a fact that functionally links p190RhoGAP to the actin cytoskeleton. The two functional domains of p190RhoGAP are separated by an 801 amino acid middle region that contains tyrosine residues 1087 and 1105. Both these residues are found in a motif (YXXPXD) thought to mediate in vitro binding to the SH2 domains of p120RasGAP (126). Results from transfection experiments suggest that p190 has tumor suppressor characteristics, in that it can inhibit Ras transformation of NIH3T3 cells. The GTP binding domain is thought to mediate this suppressive activity (127).

Overexpression of c-Src in 10T cells results in an increase in the basal level of tyrosine phosphorylation of p190RhoGAP that is not further increased by EGF stimulation, which indicates that p190RhoGAP is a preferred substrate of c-Src and not the EGF receptor. Upon stimulation of normal murine fibroblasts by EGF, p190RhoGAP undergoes a rapid and transient redistribution from a diffuse cytoplasmic localization into concentric arcs that radiate away from the nucleus with a time course that mimics EGF-stimulated actin dissolution (128). Overexpression of wt c-Src expands the window of time in which EGF-induced p190RhoGAP arc formation and actin dissolution occurs, and overexpression of kinase defective c-Src contracts that window (128). These results implicate a role for tyrosine phosphorylated p190RhoGAP in regulating cytoskeletal reorganization, possibly by inactivating Rho. Several questions are raised by these experiments. Since the results are correlative, one question is whether tyrosine phosphorylation of p190 is required for the effects of c-Src overexpression on EGF-induced rearrangement of both p190 and actin. The second is related and asks whether or not the changes observed in the actin cytoskeleton that occur within 30 min after EGF stimulation are necessary for the two-five fold increase in EGF-induced DNA synthesis that is observed at 21 hours after addition of EGF in fibroblasts that overexpress c-Src (30). These questions are currently being addressed experimentally.

Recently, our laboratory has characterized the complexity and relative levels of tyrosine phosphorylation on p190 from different c-Src overexpressors (Roof et al., 1997, in preparation). Phosphotryptic peptide analysis of endogenous p190 has uncovered only two tyrosine phosphorylated polypeptides in the molecule, one peptide containing Y1105 (located in the middle domain), and another as yet unidentified minor peptide. This pattern of phosphorylation holds true for p190 from control cells and from wt c-Src overexpressors. That c-Src is directly responsible for phosphorylating Y1105 in vivo is suggested by an increase in the level of phosphorylated Y1105 upon c-Src overexpression, a decrease in phosphorylation of Y1105 in cells overexpressing a kinase-deficient c-Src and phosphorylation of Y1105 by purified c-Src in an in vitro kinase assay (Roof et al. 1997, in preparation).

One presumed function of tyrosine phosphorylation of p190RhoGAP is to enhance its interaction with p120RasGAP (84, 129, 130). Using deletion and site directed mutagenesis, Hu and Settleman (130) have proposed a model in which phosphorylation of both Y1087 and Y1105 are necessary for the interaction of p190RhoGAP with the two SH2 domains of p120RasGAP. In this model, the engagement of the two SH2 domains of p120RasGAP with Y1087 and Y1105 in p190RhoGAP induces a conformational change in the p120RasGAP molecule which exposes the RasGAP SH3 domain for interaction with its binding partners. This model represents a mechanism whereby tyrosine phosphorylation of p190 can regulate SH3 domain interactions of p120RasGAP.

The data from our laboratory support a role for tyrosine phosphorylation of p190RhoGAP in enhancing its interaction with p120RasGAP but suggest that interactions other than the p120RasGAP SH2 domain/p190RhoGAP pTyr interaction may also be important. Phosphotryptic peptide mapping and phosphoamino acid analyses have been used to compare the level of tyrosine phosphorylation of p190 in control, wt c-Src overexpressors, and kinase deficient c-Src overexpressors at both Y1105 and the unidentified site. These comparisons have revealed that the level of phosphorylation of Y1105 in the wt c-Src overexpressors is increased approximately seven fold above that of control cells and decreased approximately six fold in the kinase deficient c-Src overexpressors relative to control cells. In contrast, the level of tyrosine phosphorylation at the unidentified site remains constant and below the level of Y1105 phosphorylation in all cell lines (approximately 25% of the level of Y1105 phosphorylation in control cells), irrespective of either c-Src overexpression or EGF addition (Roof et al. in preparation).

The fact that we have been unable to identify two, equally phosphorylated tyrosines in p190 brings into question the dual pTyr/SH2 model for complex formation between p190 and p120RasGAP. Further questions are raised by the finding that the amount of p120RasGAP that is immunoprecipitated with p190RhoGAP from the cells overexpressing wt c-Src or kinase inactive c-Src is not consistent with the overall level of p190 tyrosine phosphorylation observed in these two cell lines. For instance, the amount of p120RasGAP that co-immunoprecipitates with p190RhoGAP from c-Src overexpressors is less than expected, while the amount of p120 in complex with p190 from kinase deficient c-Src overexpressors is greater than expected, given the striking differences in the extent of p190 tyrosine phosphorylation in the two cell lines. These data suggest that factors in addition to tyrosine phosphorylation may contribute to p190RhoGAP/p120RasGAP interactions and that tyrosine phosphorylation of p190RhoGAP may regulate another function of p190RhoGAP independently of its interaction with p120RasGAP. In fact, Koch et al. (131) have indicated that SH2 domains can interact with phosphoserine-containing peptides, which implies that there may be a role for phosphorylated serines in p190RhoGAP/p120RasGAP interactions. Another possibility is that tyrosine phosphorylation of p190RhoGAP regulates the two known functional activities ascribed to p190RhoGAP, namely, the GTP-binding and GAP activities. Preliminary data from our lab has shown that tyrosine phosphorylation of p190RhoGAP inhibits the in vitro binding of GTP to the N-terminal domain of p190 (Roof, R. and S. Parsons, unpublished observation).

In conclusion, it appears that the mechanism of interaction between p190RhoGAP and p120RasGAP is more complicated than previously anticipated and that the consequences of the interaction await further experimentation before the exact function of the interaction between these two proteins is delineated.

4.4 Cortactin

The v-Src and c-Src substrate p75/p80/85 cortactin is an actin-binding protein that consists of five tandem repeats in the N-terminus and a SH3 domain in the extreme C-terminus of the molecule (85-87). Both structural features of cortactin are found in a number of cytoskeletal proteins, and indeed the N-terminal repeats of cortactin mediate binding of the molecule to actin (87). Phosphoamino acid analysis of cortactin from normal chick embryo cells indicates that the protein is phosphorylated on serine and threonine residues, whereas in chick embryo cells transformed with activated c-Src (527F), cortactin is also phosphorylated on tyrosine (86). In normal murine 10T fibroblasts, cortactin has a low level of tyrosine phosphorylation that is increased upon overexpression of c-Src. (85). In addition, EGF stimulates a time-dependent increase in the level of tyrosine phosphorylation of cortactin that is further enhanced by overexpression of c-Src (66, 85). Thus, cortactin may be a substrate of the EGF receptor as well as of c-Src, or c-Src may mediate EGF-induced phosphorylation of cortactin. The observation that increased tyrosine phosphorylation of cortactin is seen in Csk-deficient cells favors the notion that c-Src and/or its family members are responsible for phosphorylating cortactin (132, 133). Interestingly, two phases of EGF-induced cortactin tyrosine phosphorylation can be observed in 10T cells, one occuring within 2-10 min following stimulation and another occurring later in G1, with the maximum level being reached approximately nine hours post-treatment. The level of cortactin tyrosine phosphorylation in both phases is increased by c-Src overexpression.

Indirect immunofluorescence microscopy of 10T cells reveals that cortactin is localized within the cytoplasm to punctate sites that are concentrated around the nucleus (85). A significant amount of cortactin is also co-localized with actin in the plasma membrane and at peripheral adhesion sites (85). This pattern of subcellular distribution is not altered upon EGF treatment or c-Src overexpression. However, in cells transformed with activated c-Src (527F), cortactin collapses into rosettes or podosomes, which are sites of membrane/substratum interactions (86). These studies are suggestive of a role for peripheral cortactin in maintaining cortical actin integrity and membrane association, but the role of intracellular cortactin is unknown. Furthermore, the role that tyrosine phosphorylation plays in regulating cortactin function is not clear. Recent studies have shown that in vitro tyrosine phosphorylation of cortactin results in a dramatic decrease in F actin cross-linking activity which indicates a role for cortactin in regulating actin cytoskeletal reorganization (134).

4.5 Sam68

Although Sam68 is not a known cytoskeleton-associated protein, it is included in this section, because it is a well-documented substrate of c-Src. The activity of c-Src increases when cells enter mitosis, suggesting the existence of mitosis-specific substrates (135). Sam68 (Src-associated in mitosis) is a 68 kDa RNA-binding protein that is both a c-Src binding protein and a c-Src substrate (figure 1). Sam68 was discovered simultaneously by two laboratories that were comparing pTyr-containing proteins from cells arrested in mitosis with those found in asynchronous cells (136, 137). These two groups found that in mitotic cells but not in asynchronous cells a heavily tyrosine phosphorylated protein of 68 kDa can be co-immunoprecipitated with c-Src. The apparently high affinity interaction of Sam68 with c-Src requires both the SH2 and SH3 domains of Src, although it appears that the association is mediated mainly through the Src SH2 domain (136, 137). The functional consequence of phosphorylation of Sam68 is currently unknown, although it has been reported that phosphorylation of Sam68 decreases its RNA binding ability (138).

Recent work by Barlat et al. (139) has shown that a naturally occurring isoform of Sam68 exists in NIH3T3 cells. This isoform has a portion of the RNA-binding domain deleted (amino acid residues 170-208) and is called Sam68deltaKH. Sam68deltaKH is expressed at growth arrest in confluent cells but is absent from cells that have been transformed by activated c-Src (Y527F) (139). Interestingly, transfection of NIH3T3 cells with a myc-tagged Sam68deltaKH leads to a 50% decrease in BrdU incorporation in response to serum stimulation, while transfection of either myc-tagged Sam68 or vector alone had no effect on BrdU incorporation (139). Furthermore, co-expression of myc-tagged Sam68 with myc-tagged Sam68deltaKH reversed the myc-tagged, Sam68deltaKH-induced inhibition of BrdU incorporation. These results indicate that the RNA-binding domain of Sam68 is important for progression through the cell cycle. However, it is unclear what effect tyrosine phosphorylation of Sam68 has on cell cycle progression. Whether the deletion of the KH domain affects the ability of SAM68 to become tyrosine phosphorylated is not known, although the portion of the KH domain that is deleted does not contain any tyrosine residues.