Association of the Ras to Mitogen-activated Protein Kinase Signal Transduction Pathway with Microfilaments

Transcription

1 THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 274, No. 36, Issue of September 3, pp , by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. Association of the Ras to Mitogen-activated Protein Kinase Signal Transduction Pathway with Microfilaments EVIDENCE FOR A p185 neu -CONTAINING CELL SURFACE SIGNAL TRANSDUCTION PARTICLE LINKING THE MITOGENIC PATHWAY TO A MEMBRANE-MICROFILAMENT ASSOCIATION SITE* (Received for publication, January 11, 1999, and in revised form, April 29, 1999) Coralie A. Carothers Carraway, Maria E. Carvajal, and Kermit L. Carraway From the Departments of Biochemistry and Molecular Biology, and Cell Biology and Anatomy, University of Miami School of Medicine, Miami, Florida Microvilli of the aggressive ascites mammary adenocarcinoma contain a large, microfilament-associated signal transduction particle whose scaffolding is a stable glycoprotein complex (Li, Y., Hua, F., Carraway, K. L., and Carraway, C. A. C. (1999) J. Biol. Chem. 274, ) associated with the growth factor receptor p185 neu. The receptor is constitutively tyrosine-phosphorylated in the cells and microvilli, predicting that it should recruit mitogenic pathway components to this membrane-microfilament interaction site. Immunoprecipitation of cell lysates with anti-phosphotyrosine and immunoblotting showed phosphorylated forms of the mitogenic pathway proteins Shc and MAPK in addition to p185 neu, suggesting that the Ras to MAPK mitogenic pathway is activated. Immunoblotting of p185 neu -containing microvillar fractions revealed the presence in each of stably associated Shc, Grb-2, Sos, Ras, Raf, mitogen-activated protein kinase kinase, and mitogenactivated protein kinase/extracellular signal-regulated kinase, as well as the transcription factor-phosphorylating kinase Rsk. All of these pathway components coimmunoprecipitated with p185 neu from cleared lysates of microvilli solubilized under microfilament-depolymerizing conditions. The recruitment of constitutively phosphorylated p185 neu and the activated mitogenic pathway proteins to this membrane-microfilament interaction site provides a physical model for integrating the assembly of the mitogenic pathway with the transmission of growth factor signal to the cytoskeleton. This linkage is probably a requisite step in the global cytoskeleton remodeling accompanying mitogenesis. Growth factors trigger both mitogenesis and changes in cell morphology and microfilament organization (1). Since both of these effects result from the binding of the factors to a single type of receptor, their pathways must somehow be integrated. One suggestion for this integration is that the receptors and components of the signaling pathways are localized to sites where microfilaments interact with membranes (2). The EGF 1 * This work was supported in part by National Institutes of Health Grants GM and CA 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. To whom correspondence should be addressed: Dept. of Biochemistry and Molecular Biology (R-629), University of Miami School of Medicine, Miami, FL Tel.: ; Fax: ; ccarrawa@mednet.med.miami.edu. 1 The abbreviations used are: EGF, epidermal growth factor; ctmc, cytoskeletal transmembrane complex; MAP, mitogen-activated protein; MAPK, MAP kinase; MAPKK, MAPK kinase; PBS, phosphate-buffered This paper is available on line at receptor (3) and the EGF receptor family member p185 neu (ErbB2/HER2) (4) have been shown to be associated with microfilaments. EGF receptor binds directly to actin through a motif similar to the actin-binding motif of profilin (5). The mechanism for the association of p185 neu /ErbB2 with microfilaments is presently unclear. Interestingly, p185 neu /ErbB2, which has been implicated in tumor cell metastasis, is proposed to play a role in the regulation of cell structure and motility (6). The receptor is concentrated on microvilli in SK-BR-3 breast cancer cells. Activation of its phosphorylation leads to cell spreading, enlargement of microvilli, formation of pseudopodia, and aggregation of the p185 at the plasma membrane protrusions and pseudopodia (6). A specific role for p185 neu /ErbB2 in these microfilament-dependent processes known to enhance metastatic capability of tumor cells was shown in p185 neu - overexpressing transfectants of NCI-H460 lung carcinoma cells (7). Binding of a ligand to its receptor tyrosine kinase at the cell surface elicits numerous pleiotypic cellular changes culminating in cell division. This single activation event must integrate and orchestrate both temporally and spatially a multiplicity of pathways regulating metabolism, macromolecule biosynthesis, cytoskeletal organization, cell cycle, and cell division. An early consequence of ligand binding is a stimulation of the normally low level of tyrosine kinase activity of the receptor (8 10) and autophosphorylation of the receptor on its cytoplasmic domain, commonly on more than one tyrosine residue (11). Kinase activation triggers events at the plasma membrane cytoplasmic surface that initiate early steps in the transduction of growth signal to the cytoplasm and nucleus. These phosphorylated tyrosine residues create binding sites for other signal transduction components such as adaptor (linker) proteins, via sequence-specific SH2 domains (12 14). As a consequence, several signaling pathway proteins are recruited from the cytosol to the cytoplasmic surface of the plasma membrane, resulting in modulation of their functions. The autophosphorylation and recruitment permits at least transient associations of signal transduction proteins into complexes, called signal transduction particles (15). Studies of mitogenesis induced by binding of EGF or plateletderived growth factor to its receptor have described a putative common mitogenic pathway (16 19) that culminates in the phosphorylation of transcription factors. The key cellular elements are the activated receptor (8, 9, 11, 14, 15, 20, 21), the saline (Dulbecco s, no Ca 2 ); PIPES, piperazine-n,n -bis(2-ethanesulfonic acid); RIPA, radioimmunoprecipitation assay; PAGE, polyacrylamide gel electrophoresis; SH2, Src homology 2; STP, signal transduction particle; TMC, transmembrane complex; TMC-gp, transmembrane complex glycoprotein.

2 25660 Microfilament-associated Signal Transduction Particle (proto)oncogene product Ras (22 28), and the protein kinases of the MAP kinase cascade (16 18, 29 32). Two of the critical events in this pathway entail recruitment of soluble cytoplasmic proteins to the plasma membrane. The first involves activation of Ras by Sos. In this proposed mechanism, the autophosphorylated receptor binds to an adaptor protein such as Grb2 via a phosphotyrosine linkage to an SH2 or a PTB domain (33 37). The adaptor protein interacts via its Src homology 3 domain with a proline-rich peptide on Sos (33), a nucleotide exchange factor that can activate Ras (33 39). This mechanism thus couples the phosphorylated growth factor receptor to the activation of the membrane-associated signal transduction switch Ras. Activation of the growth factor receptor results in localization of the cytosolic Grb2-Sos complex in juxtaposition with membrane-associated Ras via binding of the complex to a phosphorylated tyrosine residue on the activated receptor (24, 40 42). Grb2-Sos can also be recruited indirectly to activated receptors via the adaptor protein Shc (43, 44). The second event requiring membrane localization involves the recruitment of Raf and the downstream members of the MAPK pathway by activated Ras (45, 46). Activated Ras has been shown to bind Raf (47 51), a serine/threonine kinase that can initiate the MAPK cascade (16 18, 29 32). However, no direct activation of Raf by Ras has been observed (49). Instead, an important function of Ras appears to be the recruitment of Raf to the plasma membrane (45, 46), inducing its association with the microfilamentous cytoskeleton (46) by a protein(s) X (2) with concomitant activation. During this assembly, Raf becomes activated via an incompletely understood mechanism (52, 53) mediated in some cells by the protein family (54, 55), resulting in the activation of the MAPKs. In this cascade Raf can activate MAPKK, which activates MAPK. MAPK can then phosphorylate transcription factors as part of the mitogenic process (16 18, 56). These combined observations suggest the assembly at least transiently of a large, multimeric signaling complex or signal transduction particle (4, 15), which contains many if not all of the components of this pathway. The results cited above suggest that the formation of signaling complexes at membrane-microfilament interaction sites may play important roles in cellular responses to growth factors (57, 58). Partial pathway complexes have been demonstrated in some instances, although these complexes are often transient and difficult to characterize. Stable complete complexes more amenable to characterization would be expected in the case of cells that are constitutively activated, such as many types of tumor cells. The MAT-C1 subline of the rat mammary adenocarcinoma provides substantial advantages for the investigations of membrane-microfilament interactions and their regulation (59). Its cell surface is covered with abundant microvilli, which can be readily isolated in quantity by a simple shearing and differential centrifugation procedure that does not destroy the cells or disrupt the structural organization of the microvilli (60, 61). These microvilli provide a highly purified plasma membrane fraction with intact microfilamentmembrane interactions. Extraction and fractionation studies on the microvilli have indicated that microfilaments are linked to the membrane via a high molecular mass ( 10 6 Da) complex, designated the transmembrane complex (62). As isolated from microvillar membranes, this complex contains actin, at least five glycoproteins (63, 93), and a 58-kDa cytoplasmic protein, p58 gag. The growth factor receptor p185 neu (ErbB2/HER2), the protein product of the neu (proto)oncogene, is a component of the stably associated TMC glycoprotein complex that forms the core of the TMC (see model in preceding paper (93), Fig. 10). Thus, the TMC has properties that strongly suggest that it is a microfilament-associated signal transduction particle (4, 93). The STP-associated p58, which binds to both phospholipids and to microfilaments in the manner of a capping protein (64), has been implicated in stability of the cell surface receptors of ascites sublines (65, 66), which are xenotransplantable into mice (67). Complete cdna sequencing revealed that p58 is a truncated retroviral Gag-like protein lacking its nucleic acidbinding sequence (68). From lysates of untreated ascites cells, several tyrosine-phosphorylated proteins, including p58 gag, were immunoprecipitated with anti-phosphotyrosine antibody (69). Phosphoamino acid analysis indicated that p58 gag is phosphorylated on both tyrosine and serine. The membrane skeletal protein may be a substrate for the tyrosine kinases p60 src and/or p120 abl, which, along with the receptor kinase p185 neu, are components of the large signaling particle (69). p58 gag has PXXP-containing polyproline motifs, a general feature of retroviral Gag proteins, and was shown to bind to both p60 src (70) and p120abl. 2 The presence of the receptor, the cytoplasmic kinases, and in situ-phosphorylated substrate proteins, along with the aggressive properties of these highly metastatic tumor cells, suggest that p185 neu is constitutively phosphorylated and activated (69). This hypothesis also predicts that the activated growth factor receptor will recruit components of its signal transduction pathway(s). In the present study, we have used fractionation and co-immunoprecipitation approaches to investigate the presence of components of the Ras/MAPK pathway associated with the TMC. The results of these studies provide additional evidence for the presence of a microfilament-linked signal transduction particle at the cell surface in the microvilli of these cells. Moreover, they suggest that the STP/TMC is a site for association of the Ras-Raf-MAPK pathway components at the plasma membrane and provide a physical mechanism by which the cytoskeleton and mitogenic pathways can be coupled. To our knowledge, the isolation of a microfilament-associated, growth factor receptor-containing signaling complex containing all of the components of the mitogenic pathway has not been reported. EXPERIMENTAL PROCEDURES Materials Antibodies were obtained from the following sources: Oncogene Research Products (Cambridge), p185 neu /ErbB2 mouse monoclonal Ab3, pan-ras mouse monoclonal; Dako (Carpinteria), ErbB2 rabbit antihuman polyclonal antibody; Transduction Laboratories (Lexington, KY), Shc rabbit polyclonal antibody, Sos (1 and 2) rabbit polyclonal antibody, MAPKK (mitogen-activated protein kinase/extracellular signal-regulated kinase kinases 1 and 2) rabbit polyclonal antibody, Rsk mouse monoclonal antibody, and horseradish peroxidase-conjugated phosphotyrosine (PY20 recombinant); Upstate Biotechnology, Inc. (Lake Placid, NY), Grb2/Sem5 rabbit polyclonal antibody, human Raf-1 polyclonal rabbit antibody; Zymed Laboratories Inc. (South San Francisco, CA), mouse monoclonal anti-mapk (extracellular signal-regulated kinases 1 and 2) antibody; Promega Biotech (Madison, WI), mouse and rabbit IgG. All other chemicals were reagent grade and obtained from Sigma, unless specified otherwise. Methods Passaging of Ascites Tumor Cells and Preparation of Microvilli and Microvillar Fractions Ascites cells of the MAT-C1 subline of the rat mammary adenocarcinoma were passaged by weekly intraperitoneal injections into rats (65). Cells were collected and washed with PBS. Microvilli (60), a microfilament-containing membrane preparation, were prepared as previously reported (60, 61). The cell bodies, undisrupted by this procedure, were removed by centrifugation at 750 g, and the morphologically intact microvilli were harvested and washed with PBS at 45,000 g for 20 min prior to fractionation. Preparation of microvillar microfilament core, membrane, and TMC- 2 J. Huang, S. Price-Schiavi, R. Van Etten, B. J. Mayer, and C. A. C. Carraway, manuscript in preparation.

3 Microfilament-associated Signal Transduction Particle enriched fractions Microfilament cores, microvillar membranes, and TMC-enriched microvillar fractions were prepared as previously reported (61, 63, 64, 71, 72). Microfilament cores were prepared by extraction of microvilli in an isotonic microfilament-stabilizing buffer (TPK buffer; 0.5% Triton, 100 mm KCl, 2 mm MgCl 2, 0.1 mm phenylmethylsulfonyl fluoride, and 20 mm PIPES, ph 6.8) for 15 min and either differential centrifugation (61) or phalloidin shift velocity sedimentation, a diagnostic procedure for analysis of microfilament-associated proteins (71). Microvillar membranes containing membrane-associated but not filamentous actin were prepared under microfilamentdepolymerizing conditions (low ionic strength, ph 9.5) according to the previously reported procedure (61, 63, 70). After centrifugation at 10,000 g to remove undissociated microfilaments, the membranes were harvested at 150,000 g for 1.5 h. This preparation contained few of the microfilament-associated proteins. A transmembrane complexenriched preparation was made by extraction of microvilli on ice for 30 min in a microfilament-depolymerizing buffer (S buffer (0.2% Triton X-100, 150 mm KCl, 2 mm MgCl 2, 0.2 mm ATP, 0.2 mm dithiothreitol, 0.1 mm phenylmethylsulfonyl fluoride, and 5 mm Tris, ph 7.6); Refs. 63 and 64) containing a protease inhibitor mixture. Low speed centrifugation removed undissociated filaments, and high speed centrifugation yielded a cytoskeletal TMC (ctmc) preparation containing small amounts of microfilament-associated proteins such as -actinin and ezrin (63, 64, 72). A more highly purified transmembrane complex was prepared by nonionic detergent solubilization of microvillar membranes, followed by fractionation by velocity sedimentation or gel filtration (62). Immunoprecipitation of Tyrosine-phosphorylated Proteins from Whole Cell Lysates Washed intact cells were solubilized in either radioimmunoprecipitation assay (RIPA) buffer (150 mm NaCl, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS, 50 mm Tris, ph 8.0; Ref. 74) or 2% SDS followed by the addition of the other components of RIPA buffer to give the same final concentrations of RIPA components. The extracts were centrifuged at 120 g for 4 min. To the supernatants was added 50 l of agarose-conjugated antibody to recombinant anti-phosphotyrosine py20, and the mixtures were incubated overnight at 4 C with rocking. The immune complexes were obtained by centrifugation of the beads at 120 g for 4 min and washing twice with PBS containing 0.02% Triton. The pellet was resuspended in 100 l of2 SDS electrophoresis buffer and subjected to electrophoresis and blot transfer. Distribution of Mitogenic Pathway Proteins between the Microvilli and the Cell Body The relative amounts of mitogenic pathway components in whole cells and in cell bodies remaining after shearing of microvilli and isolated microvilli were assessed by densitometry quantification from immunoblots. To estimate the amount of microvilli isolated from a given number of MAT-C1 cells, the cells were labeled in vivo with 20 Ci of [ 14 C]glucosamine (Amersham Pharmacia Biotech), according to our standard procedure for labeling glycosylated (predominantly cell surface) proteins (63, 75). Aliquots representing equal numbers of cells and cell bodies, obtained by counting, were prepared for SDS-PAGE. The amount of microvilli corresponding to that number of cells/cell bodies to be prepared for SDS-PAGE was taken to be the difference of the glucosamine counts between identical numbers of cell and cell bodies. After immunoblotting with antibodies to mitogenic pathway proteins (p185 neu, Shc, Grb2, Sos, Ras, Raf-1, MAPKK, MAPK, and Rsk; described below) and detection by chemiluminescence, the bands on the film were quantified by densitometry using a Hewlett- Packard Scanjet IIA Bioscan. Immunoprecipitations of Transmembrane Complex-containing Low Speed Supernatants with Antibodies to p185 neu /ErbB2, MAPK, and Raf l of microvilli were incubated on ice for 30 min in a microfilament-depolymerizing buffer containing a protease inhibitor mixture (S buffer; Refs. 62 and 73). The lysate was centrifuged at 15,000 g for 15 min at 4 C, and the supernatant enriched in a cytoskeletal transmembrane complex preparation (63, 64, 72) was used for immunoprecipitations with antibodies to c-neu, Raf, and MAPK. Mouse monoclonal anti-mapk (extracellular signal-regulated kinases 1 and 2), mouse monoclonal anti-c-neu, rabbit anti-human Raf-1, and appropriate nonimmune control antisera were incubated with protein A-agarose (Sigma). Approximately 800 l of lysate supernatant was added to each, and the mixtures were incubated overnight at 4 C. The pellets were obtained by centrifugation at 3000 rpm for 5 min and washed three times with PBS. Electrophoresis and Blotting Procedures SDS-PAGE was performed on 8% Laemmli minigels. For lectin binding analyses or immunoblotting, proteins were transferred to nitrocellulose membranes and processed as described previously (4, 63). Concanavalin A (ConA)/horseradish peroxidase (Sigma) was used at 50 g/ml and detected using the Renaissance TM chemiluminescence assay kit (NEN Life Science Products). Immunoblot Detection of Mitogenic Pathway Proteins -Immunoblotting was performed using the following primary antibodies (sources listed above) and dilutions (optimized for each antibody for detection by chemiluminescence): p185 neu mouse monoclonal antibody (1:4000); ErbB2 rabbit polyclonal antibody (1:500); pan-ras mouse monoclonal antibody (1:750); human Raf-1 polyclonal rabbit antibody (1:2500); MAPKK (mitogen-activated protein kinase/kinase) rabbit polyclonal antibody (1:1500); MAPK (extracellular signal-regulated kinases 1 and 2) (1:2500); Rsk mouse monoclonal antibody (1:500); and horseradish peroxidase-conjugated phosphotyrosine (PY20 recombinant) (1:1000). In all cases, the second antibody was either mouse (for polyclonal) or rabbit (for rabbit monoclonal) IgG (1:14,000). Immunoblotting was performed as described previously (63, 69). The replicas were incubated with primary antibody in Tween/Tris-buffered sodium chloride (500 mm NaCl, 50 mm Tris, ph 8.0) containing 1% bovine serum albumin for 1 h and with horseradish peroxidase-conjugated antibodies to rabbit or mouse immunoglobulin for 1 h. Blots were washed, and the reactive bands were detected using the NEN Life Science Products Renaissance TM chemiluminescence detection kit according to the manufacturer s instructions. RESULTS Constitutive Tyrosine Phosphorylation of p185 neu and Other MAT-C1 Proteins Observations from previous studies on the highly proliferative and metastatic MAT-C1 cells predict that the mitogenic pathway is constitutively activated. Several constitutively phosphorylated substrate proteins for both tyrosineand serine/threonine-specific kinases have been identified, including the membrane skeletal protein p58 gag (69). A high M r protein in the kDa range was tyrosine-phosphorylated, suggesting that p185 neu in the MAT-C1 cells may be constitutively phosphorylated and therefore capable of binding cytoplasmic signaling components having SH2 domains. This question was addressed by immunoprecipitation of cell lysates with anti-tyr(p) and analysis of the immunoprecipitate by immunoblotting. Our previous studies on MAT-C1 microvilli had indicated that preparing cell lysates with nonionic detergent is completely inadequate for solubilizing the large, stable complexes at the membrane-microfilament interface (4, 62 64). Therefore, denaturing conditions were required for the solubilization of TMC-containing fractions. MAT-C1 cell lysates were prepared in two ways. The first employed RIPA buffer (74), which was designed to dissociate complexes prior to immunoprecipitation. RIPA did not completely dissociate the TMC (data not shown), and for this reason SDS was routinely used to dissociate completely all protein-protein interactions in the complex. The latter procedure also denatures and destroys the activities of cellular enzymes, such as phosphatases and proteases, which could alter the phosphorylated proteins during the analyses. When diluted appropriately, the SDS-solubilized material was amenable to immunoprecipitation and other types of binding experiments (see preceding paper (93), in which concanavalinagarose is used to fractionate SDS-solubilized cell fractions). To determine whether p185 neu is the high M r protein, the anti-tyr(p) immunoprecipitate from the whole cells lysed with SDS was assayed by immunoblot with anti-p185 neu. Fig. 1 shows the presence of p185 neu in the immunoprecipitates. Exhaustive immunoprecipitation of cell lysates with anti-phosphotyrosine-agarose, followed by immunoblotting of the final supernatant with either the monoclonal or polyclonal anti- ErbB2, revealed that 95% of the p185 neu in the intact, untreated cells is phosphorylated (data not shown). The phosphorylated receptor would be expected to recruit SH2 domaincontaining proteins to the plasma membrane to initiate the assembly of a signaling complex containing components of the mitogenic pathway. One mechanism for the recruitment involves the adaptor protein Shc (43, 44), which can be phos-

4 25662 Microfilament-associated Signal Transduction Particle FIG. 1.Anti-Tyr(P) immunoprecipitation of solubilized whole cells. Washed cells were lysed in RIPA buffer and subjected to immunoprecipitation with anti-tyr(p)-agarose, as described under Methods. The pellets were analyzed by SDS-PAGE and electrotransfer to nitrocellulose for immunoblotting with anti-tyr(p) and anti-erbb2, -Shc, and -MAPK antibodies. MF, microfilament; IP, immunoprecipitation. phorylated on tyrosine. Analysis of the anti-phosphotyrosine immunoprecipitates by immunoblotting with anti-shc demonstrated that it is phosphorylated in the ascites cells (Fig. 1B). A major indicator of mitogenic activation is the dual phosphorylation of MAPK on both tyrosine and threonine (16 18, 29 32). Tyrosine phosphorylation of MAPK was demonstrated by the presence of the MAPK in the anti-phosphotyrosine immunoprecipitates (Fig. 1B). The presence of activated MAPK was also shown by immunoblotting with antibody to activated MAPK. 3 Recruitment of Signaling Components to the Plasma Membrane In previous studies we showed that the receptor tyrosine kinase p185 neu is associated with a high M r glycoprotein complex (63) linked to microfilaments in microvilli from ascites rat mammary adenocarcinoma cells (Ref. 4 and preceding paper (Ref. 93)). We have proposed that the TMC-p185 complex can act as a signal transduction particle that can recruit signal transduction components from the cytoplasm to the membrane-microfilament interface as part of the cell activation process. These results led us to predict that the glycoprotein complex forms the core of a signal transduction particle which can organize and integrate signaling pathways at the membrane-microfilament interface (2, 58). As demonstrated above, these highly proliferative cells have constitutively phosphorylated p185 neu /ErbB2. These findings would predict that the tyrosine-phosphorylated receptor can recruit signal transduction components containing SH2 domains from the cytoplasm to the membrane. To analyze association of mitogenic pathway components with the plasma membrane, we isolated microvilli (60, 61), which comprise over 80% of the plasma membrane of these cells, by shearing them from the ascites cells, leaving intact cell bodies (61). A direct quantitative comparison is not feasible because of losses in yield during preparation. In order to compare microvilli and cell body preparations and correct for losses of microvilli during their preparation, the ascites cells were metabolically labeled with glucosamine. To obtain an estimate of the amount of purified microvilli obtained from a given number of cells, we used as starting material ascites cells labeled metabolically with glucosamine (63, 75). In previous studies, we have shown that about 90% of the glucosamine is incorporated into an abundant cell surface glycoprotein complex, the sialomucin complex (75), serving as a quantifiable marker for the plasma membrane. 3 M. Carvajal, unpublished observation. FIG. 2. Distribution of pathway components between tumor cell bodies and microvilli. MAT-C1 cells were metabolically labeled in vivo with glucosamine and sheared to remove microvilli. Cell bodies and microvilli were isolated by differential centrifugation. Intact cells and cell bodies were counted and analyzed for glucosamine label by scintillation counting. For SDS-PAGE, equal numbers of cells and cell bodies were solubilized and loaded onto the gels. The sample of microvilli loaded was equivalent to the difference in radioactivity between the cells and cell bodies. The SDS-PAGE gels were analyzed by immunoblotting with the appropriate antibodies. Since this complex is distributed throughout the cell surface, on both the microvilli and nonvillous areas of the plasma membrane, glucosamine labeling provides a minimal estimate of the plasma membrane content of the cells. For these experiments, cells were labeled with glucosamine in vivo (63) and then sheared to remove microvilli. Cell bodies were isolated by centrifugation. For quantitation, we analyzed by SDS-PAGE equal numbers of intact cells and cell bodies, determined by cell counting. To determine the amount of microvilli for analysis, aliquots of cells and cell bodies were assayed for radioactivity by scintillation counting. The difference in label between the cells and cell bodies represents the amount of label in the microvilli. By loading that amount of microvillar fraction radioactivity onto SDS-PAGE, we can analyze equivalent amounts of cells, cell bodies, and microvilli and eliminate problems from loss of microvilli during the procedure. This method provides a minimal estimate for microvilli, since the membrane of the cell bodies retain a portion of the microvilli (61). According to our glucosamine measurements, the amount of membrane on the sheared cells is approximately 20% of the total. The SDS-PAGE gels were then assayed by immunoblotting with antibodies to selected components of the Ras/MAPK pathway to evaluate their recruitment to the plasma membrane from the cytoplasm. Results from these immunoblots are shown in Fig. 2. Qualitatively, it is clear that there are similar amounts of each of the signaling components in microvilli and in the cell bodies, which retain some plasma membrane. This observation was verified quantitatively by densitometric scanning of the immunoblot bands, as shown in Table I for a typical experiment. Since the microvilli contain little cytoplasmic material (61) and the cell bodies contain a fraction of the membrane, these results suggest that the signaling proteins are to a significant extent associated with the microvilli (pure plasma membrane) in these cells. Regardless of the exact quantification, these results clearly demonstrate that substantial amounts of the components of the Ras/MAPK signaling pathway are associated with the plasma membrane, consistent with our hypothesis that they are recruited to the p185 neu -containing signaling particle at the membrane-microfilament interface. As shown in Table I, minimally half of each of the Ras to MAPK central pathway components are associated with the microvilli. This is expected for Ras, which is targeted to the

5 TABLE I Distribution of mitogenic pathway proteins between microvilli and cell bodies of MAT-C1 ascites cells Densitometry quantification Whole cells Cell bodies Microvilli Percentage in microvilli arbitrary units % Grb2 20,542 8,542 12, Ras 17,245 6,853 9, Raf 21,590 11,092 10, MAPKK 15,425 5,323 9, MAPK (p42 p44) 29,596 15,579 12, p90 rsk 19,458 9,867 10, Microfilament-associated Signal Transduction Particle membrane component by prenylation of its carboxyl terminus (76). On the other hand, Grb2-Sos complex, Raf, MAPKK, and MAPK are cytoplasmic or nuclear until recruited to the membrane by an activation event (overviewed in Ref. 2). The presence of all of these normally intracellular proteins at the membrane-microfilament interface provides further evidence for the constitutive activation of the ascites tumor cells. Interestingly, p90 rsk, a MAP kinase substrate found in the cytoplasm and the nucleus in both activated and unactivated cells (77), was present at the membrane in similar proportions to the other pathway proteins. Association of Signal Transduction Pathway Components with STP/TMC-containing Microvillar Particulate Fractions Our hypothesis that components of the Ras/MAPK pathway are associated with the membrane-microfilament interface was first tested by fractionating the microvilli to determine whether the signaling components co-fractionate with the p185 neu -containing STP. Extraction under mild conditions (isotonic buffer, ph 6.8) and differential centrifugation yields a relatively intact microfilament core that sediments at low speed (61, 72). Centrifugation of the supernatant from this sedimentation at 100 kg yields the TMC containing actin but no microfilaments (61, 72). A microfilament-depleted high speed pellet called the cytoskeletal TMC, which is enriched in membrane skeletal proteins (63, 64, 72), was made by Triton extraction of microvilli at ph 9.5 and low ionic strength to depolymerize microfilaments. Immunoblot analyses of the particulate microvillar fractions revealed that Ras, Grb2, Sos, Raf, MAPKK, and MAPK were associated with the p185 neu -containing TMC fractions, including the ctmc (Fig. 3). Interestingly, there is some variation in the fraction of the different components that remain associated with the microfilaments. As shown previously, p185 neu is largely associated with the microfilament core when extraction is performed under mild conditions (4). Moreover, Grb2 is stably associated with the core, as expected for linkage via an SH2-tyrosine phosphate interaction. Interestingly, MAPK associated preferentially with the microfilament-containing low speed pellet, suggesting a more immediate association of the terminal Ras-to-MAPK pathway component with a microfilament-associated protein. Most of the other components are less stably associated. The stringency of the extraction conditions for generating the ctmc suggests a stable complex of these components with the glycoprotein complex that forms the core of the TMC, although the interactions involved in this complex are undefined. A noteworthy observation is that the MAPK substrate p90 rsk, found in the previous experiment to be recruited in significant amount to the microvilli at the membrane, is also stably associated with the p185 neu -containing complex in the particulate microvillar fractions. The interaction of Rsk with the membrane skeletal complex appears stable, since this kinase is abundant despite the stringent conditions for ctmc preparation. While the large, stable signaling complex contains all of the Raf to MAPK to Rsk mitogenic pathway components, a number FIG. 3. Association of mitogenic pathway components with STP-containing microvillar microfilament core and TMC fractions. Microvilli and fractions prepared as described previously (63) were analyzed by SDS-PAGE and electrotransfer onto nitrocellulose for immunoblotting using antibodies against mitogenic pathway components. MV, microvilli; MF, microfilament. of signal transduction proteins, both receptor and nonreceptor types, were not associated. Table II shows that Ras GTPaseactivating protein (Ras GAP), the insulin receptor, and transforming growth factor receptor types I and II, although present in the microvilli, were found largely in the complexdepleted soluble fraction; EGF receptor is not expressed in the ascites cells (4). Velocity Sedimentation Analysis of the Association of Signal Transduction Components with Microfilament Core To circumvent potential co-sedimentation artifacts, which may erroneously imply the existence of proteins in a specific microfilament-associated complex, we have developed the phalloidin shift analyses of microvilli (71). Phalloidin binds specifically to microfilaments and stabilizes them to dissociation. A comparison of the migration on velocity sedimentation of the microfilament fractions of phalloidin-treated and -untreated microvillar extracts shows a shift of the stabilized filaments and any proteins in that complex farther into the gradient. Fig. 4 shows the distribution of the mitogenic pathway proteins in the gradient fractions, with both the large STP/TMC at the bottom and in soluble fractions in the upper portion. As expected, the amount of the components associated with microfilaments is less on the gradients than in the sedimented pellets because of the time required for analysis by gradient centrifugation. Interestingly, the presence of Raf in middle fractions suggests that it is being released during centrifugation, suggesting a more direct association with microfilaments than other components. The fact that microfilament-unassociated forms of other components also do not migrate as suggested by their molecular sizes (for example, see Rsk) suggests that they may be present in smaller complexes released from the microfilaments. Analyses of these putative complexes may help to resolve questions about molecular interactions in associations of the signaling proteins with the microfilaments. Co-immunoprecipitation of p185 neu and Mitogenic Pathway Proteins from STP/TMC-enriched Microvillar Extract The major problem in characterizing interactions between specific proteins in the microfilament-associated STP has been the generation of small enough complexes. Extensive cross-linking and stabilization of microfilaments with proteins such as -actinin (60) prevent complete depolymerization under traditional actin-depolymerizing conditions (62). Further, the large size of the stably associated receptor-containing TMC/STP-glycoprotein complex (93) with which the signaling proteins associate precludes simple immunoprecipitation of microvillar lysates. A smaller complex was generated by lysing microvilli with S buffer, the depolymerization buffer used in the preparation of the TMC (62), and clearing the lysate of undepolymerized mi-

6 25664 Microfilament-associated Signal Transduction Particle TABLE II Fractionation of signal transduction components from MAT-C1 ascites rat mammary tumor cell microvilli Signaling protein Microvilli Microfilament core ctmc High speed supernatant EGF receptor ErbB2 ( / ) Insulin receptor TGF R-I a ( / ) ( ) TGF R-II b ( / ) Ras GAP c ( / ) ( ) Rsk a Transforming growth factor receptor I. b Transforming growth factor receptor II. c GTPase-activating protein. FIG. 5.Co-immunoprecipitation of p185 neu and signaling proteins by anti-p185 immunoprecipitation of TMC-enriched soluble fraction of microvilli extracted in actin-depolymerizing buffer containing Triton. Microvilli were extracted in the microfilament-dissociating buffer used and subjected to velocity sedimentation analysis, as previously reported for the original isolation of the TMC (62). SDS-PAGE was performed on the fractions, and immunoblots were analyzed with antibodies to mitogenic pathway proteins. FIG. 4.Distribution of mitogenic pathway proteins by velocity sedimentation analysis of microvilli lysed in Triton-containing microfilament-dissociating buffer. Microvilli extracted in low ionic strength, high ph buffer containing Triton were subjected to velocity sedimentation analysis (62). The gradient fractions were analyzed by immunoblotting with antibodies to mitogenic pathway components and to TMC-gp65 and TMC-gp55. crofilaments and associated proteins by centrifugation at low speed. The supernatant, which retains a portion of the STP released from microfilaments, was subjected to immunoprecipitation with antibodies to p185 neu /ErbB2. As a negative control, immunoprecipitations were performed with antibodies to EGF receptor, which is not expressed in these cells (4). All of the proteins of the mitogenic pathway were co-immunoprecipitated with p185 neu /ErbB2, as shown in Fig. 5. Immunoprecipitation with antibodies to Raf and MAPK, as well as any of the other components, gave essentially identical results (data not shown). The observation that the components immunoprecipitated were the same regardless of the antibody used indicates that different complexes containing one component were not being co-immunoprecipitated with its specific antibody. DISCUSSION Mitogenesis entails a complex series of pleiotypic events, requiring not only the transmission of signals to the nucleus, but also massive remodeling of the cell cytoskeleton in preparation for and during cell division. Although considerable progress has been made in understanding the pathways involved in transmitting growth factor signals to the nucleus to activate transcription factors, much less is known about the regulation of the organization and reorganization of the cytoskeleton. In particular, little is known about how the pathways integrating the signal to the nucleus and to the cytoskeleton are physically organized. A developing paradigm for eliciting pleiotypic cellular responses to diverse ligands for plasma membrane receptors is that transmission of the signal to several pathways involves recruitment of the components of the pathways to sites of interaction of the plasma membrane with the cytoskeleton (discussed at length in Ref. 58). These membrane-microfilament interaction sites in the plasma membrane are often involved in both establishing and stabilizing cell morphology (2, 58) and have been shown to be enriched in tyrosine kinases and their substrates, which have been implicated in mitogenesis. Microvilli of rat mammary adenocarcinoma cells provide a useful model for investigating membrane-microfilament interactions. Fractionation studies on the microvilli have shown that a p185 neu /ErbB2-containing, high M r glycoprotein complex, the TMC-gp complex, is associated with microfilaments (4, 63). The glycoprotein complex has been isolated as a transmembrane complex with actin and cytoplasmic p58 gag (62) or, under more stringent conditions (1 M KCl, ph 9.5), as a complex of glycoproteins (TMC-gp complex; gp55, gp65, gp80, gp110, and gp120 (63)). Under denaturing conditions it could be purified free of cytoplasmic proteins and reconstituted into the large glycoprotein complex (93). In that study, overlay analyses with actin showed that at least one of the glycoproteins interacts directly with actin. Furthermore, fractionation (4) and reconstitution studies (93) showed that the tyrosine receptor kinase p185 neu /ErbB2 associates directly with the TMC-gp. The presence of the growth factor receptor kinase at a site involved in regulating microfilament and cell surface dynamics suggests that, like integrin-containing adhesion plaques (79) and cadherin-containing adherens junctions (80), the glycoprotein complex is a membrane-microfilament interaction site, which transduces signal to microfilaments to regulate cell morphology. Hence, this site provides a physical scaffolding for the integration of growth factor signals to the mitogenic pathway and to the cytoskeleton. An additional important feature of these highly proliferative tumor cells is that their mitogenesis pathway appears to be constitutively stimulated. MAT-C1 microvilli show a high level of tyrosine kinase activity and several specific tyrosine-phosphorylated proteins (69). At least three tyrosine kinases copurify with the STP, p185 neu /ErbB2 (4), p60 src, and p120 abl (69). Of these, at least one, p185 neu /ErbB2, is activated, based upon its immunoprecipitation from lysates of the MAT-C1 cells with anti-phosphotyrosine (Fig. 1). Further, MAP kinase is activated, as shown in two types of experiments: 1) abolition of its activity by dephosphorylation of immunoprecipitated MAP

7 Microfilament-associated Signal Transduction Particle FIG. 6. Model of the TMC/STP as a microfilament-associated mitogenic pathway complex, a site for cross-talk between the Ras 3 MAPK pathway and the cytoskeleton. kinase) and 2) immunoblotting with anti-activated MAPK. 4 The phosphorylation of p185 neu /ErbB2 should permit it to recruit SH2-containing factors from the cytoplasm, initiating the formation of a signaling particle. Using both co-sedimentation and co-immunoprecipitation, we demonstrated the association of Shc, Grb-2, and Sos with the TMC. This was not surprising, since they are required to initiate the mitogenic pathway by activating membrane-associated Ras. However, blotting for the other members of the MAP kinase cascade (Raf, MAP kinase kinase, and MAP kinase) showed that the complete pathway, as well as p90 rsk, is stably assembled in the STP/TMC (Fig. 6). Although the complex is large ( Da; Refs. 62 and 63), the assembly of the complete mitogenic pathway complex was not necessarily expected. This point emphasizes a very important caveat in analyzing signaling complexes; multiplicity of components is recognized only if appropriate assays are used. Size analyses of complexes, as reported for the TMC (62), the TMC-gp complex (63, 93), and a complex of Raf with the molecular chaperones hsp90 and p50 (81), provide valuable information on the nature of the complexes. The size of the TMC indicated that individual components were present in multiple copies and/or that multiple components were present. The presence of the signaling components in STP-enriched fractions prepared under relatively harsh conditions (e.g. high ph/low ionic strength used in the preparation of ctmc (63, 64, 72)) suggests stable interactions among components in the complex. Indeed, the fact that RIPA buffer, developed for dissociation of proteins prior to immunoprecipitation (74), is ineffective in completely solubilizing the various forms of the STP/ TMC indicates an unusual degree of stability of protein-protein interactions in the complex. These observations necessitate a second caveat; incomplete dissociation can lead to an erroneous inference of direct interaction if more than two proteins are present in the complex. This problem becomes particularly difficult when dealing with large, stable complexes, such as those frequently found at membrane-microfilament interaction sites. The isolation of components of the mitogenic pathway with the STP/TMC was facilitated by the fact that the p185 neu / ErbB2 in the highly proliferative ascites cells is constitutively activated. Essentially all of the MAT-C1 cell p185 neu /ErbB2 was shown to be tyrosine-phosphorylated, providing binding sites for SH2 domain-containing proteins. One possible mechanism for the activation is through mutation of specific components of the pathway, such as Ras or Raf; we have no evidence for such a mutation(s), whose activation mechanisms, on any account, would bypass the growth factor receptor. An alternative for activation of the receptor and the pathway is via autocrine growth factors, since many tumors are known to produce these factors. Of particular interest is the fact that the ascites cells contain a specific activator of the p185 neu /ErbB2 (82). This modulator ASGP-2 is an integral membrane glycoprotein having two EGF-like domains (83), which associates with p185 neu /ErbB2 specifically, stimulating its phosphorylation and tyrosine kinase activity (82). Further, ASGP-2 potentiates the level of NDF/heregulin-activated phosphorylation of ErbB2 and ErbB3 (82). Thus, we propose that ASGP-2, which binds to ErbB2 in the absence of the other ErbB receptors, can activate ErbB2 as well as modulating signaling through the ErbB2/ErbB3 heterodimer. In previous studies, the purported ligand was characterized as the transmembrane subunit ASGP-2 (83) of the ascites cell sialomucin complex (75). ASGP-2 appears to be an epithelial specific protein found in lung (84), uterus (85), colon (86), and lactating but not virgin mammary gland in the rat (87). A fraction of ASGP-2 associates stably with the TMC (88), presumably through an interaction with p185 neu /ErbB2, since the purported ligand can be coimmunoprecipitated with p185 neu /ErbB2 from solubilized ascites cell membranes under conditions (ph 9.5, low ionic strength) in which a fraction of the p185 neu /ErbB2-containing complex becomes dissociated from the microfilament-associated TMC (82). Regardless of the mechanism of constitutive activation, it is clear that the STP/TMC is a major site for cross-talk in the ascites tumor cells. Minimally, mitogenic signaling through p185 neu /ErbB2 is transduced to the microfilamentbased cytoskeleton in the cell surface microvilli, which in turn must be linked to the intracellular cytoskeleton in the intact cell. It is likely that this receptor complex mediates cross-talk with yet other pathways, since it also contains p60 src, p120 abl (69), the p85 subunit of PI-3-kinase, phospholipase C, and modulators of the mitogenic pathway, including some isoforms of protein kinase C and specific phosphatases. 5 Interestingly, the insulin receptor, transforming growth factor receptors I and II, and Ras GAP, although present in the microvilli, are excluded from the complex. A plethora of questions remain on the organization and functions of the signaling complex. The size and stability of the multimeric complex precludes a simple answer to the question of molecular interactions among complex components. The first step is presumably recruitment of signaling proteins to the phosphorylated receptor via SH2 or PTB domains. It is not completely clear at present which signaling proteins bind directly to the activated ErbB2 receptor. The Grb2 SH2 domain has been reported to have a consensus binding sequence for p185 neu /ErbB2 (89), and this adaptor protein binds in situ to ErbB2 (90). In addition, Shc can bind in vitro to ErbB2 via its PTB domain (91). Either or both of these adaptors are candidates for binding to phosphorylated ErbB2 and recruitment of Ras to the STP/TMC. From that point, a plausible assembly pathway would involve association of Ras with Raf, of Raf with MAPKK, and of MAPKK to MAPK, interactions that have been shown to occur in mitogenic complexes. The somewhat surprising observation was the stable associ- 4 D. Lorenzo and M. Carvajal, unpublished observations. 5 M. Carvajal and I. Lovinescu, unpublished observations.

8 25666 Microfilament-associated Signal Transduction Particle ation of p90 rsk with membrane-microfilament interaction site. However, Rsk has been shown to bind to MAPK and has been reported to localize to the cell periphery (92), although the mechanism or cellular significance is not clear. Our data suggest a potential mechanism involving linkage of a signal transduction particle to microfilaments and implicate p90 rsk at least indirectly in linking the mitogenic pathway to morphological alterations after a signaling event. Current knowledge is minimal in the important area of mechanisms for the temporal and spatial regulation of the localization and relocalization of mitogenic pathway components in signaling (discussed in Chapter 8 of Ref. 58). The significance of the nuclear transport of cytoplasmic pathway proteins on activation is intuitive. Not so easily understood is the significance of the presence in both the cytoplasm and the nucleus of pathway components such as MAPK, with the kinase at both locations becoming activated. A second poorly understood area is the significance, regulation, and mechanisms for cytoplasmic sequestration and for relocalization of nuclear export motif-containing proteins such as mitogen-activated protein kinase/extracellular signal-regulated kinase into and then back out of the nucleus after activation. Finally, the appearance of nuclearly localized signaling proteins at the plasma membrane has from time to time been reported, yet the mechanisms for membrane localization have not been described nor have insights into the cellular function of this localization been proffered. These important regulatory questions are a major challenge and will predictably involve, in many cases, the cytoskeleton and its associations with the plasma membrane. Our results indicate that the TMC contains all of the elements necessary for integrating the regulation of microfilament organization and the transmission of the mitogenic signal to MAP kinase and Rsk and thus to the nucleus. These elements include sites of interaction with microfilaments, the receptor tyrosine kinase p185 neu, and the MAP kinase cascade, possibly through binding of Raf to the TMC. Since MAP kinase in the microvilli is phosphorylated and active, the TMC may also contain a component that activates Raf as the first step in stimulating the MAP kinase cascade. Studies are under way to define these specific components of the TMC. Integration of proliferation and cytoskeletal signals suggests a common site. We have proposed that the dynamics of microfilaments in the MAT-C1 cells is regulated at the site of the microfilament-membrane interaction, because p58 gag at that site appears to stabilize the microvilli, presumably by stabilizing the TMC-gp interactions. Thus, the STP includes both the scaffold for the mitogenic pathway components and a site for regulating microfilament dynamics, as required for integration. The study of the regulation of microfilament dynamics at this site is difficult in these cells, because of the presence of the stabilizing factor p58 gag. However, other MAT sublines lacking p58 (62) and having dynamic microvilli (65, 66) may be amenable to such studies. REFERENCES 1. Bretscher, A. (1991) Annu. Rev. Cell Biol. 7, Carraway, K. L., and Carraway, C. A. C. (1995) BioEssays 17, Wiegant, F. A. C., Blok, F. J., Defize, L. H. K., Linnemans, W. A. M., Verkleij, A. J., and Boonstra, J. (1986) J. Cell Biol. 103, Carraway, C. A. C., Carvajal, M. E., Li, Y., and Carraway, K. L. (1993) J. Biol. Chem. 268, den Hartigh, J. C., van Bergen en Henegouwen, P. M. P., Verkleij, A. J., and Boonstra, J. (1992) J. Cell Biol. 119, De Potter, C. R., and Quatacker, J. (1993) J. Clin. Exp. Metastasis 11, Yu, D., Wang, S. S., Dulski, K. M., Tsai, C. M., Nicolson, G. L., and Hung, M. C. (1994) Cancer Res. 54, Fantl, W. J., Johnson, D. E., and Williams, L. T. (1993) Annu. Rev. Biochem. 62, Lemmon, M. A., and Schlessinger, J. (1994) Trends Biochem. Sci. 19, van der Geer, P., and Hunter, T. (1994) Annu. Rev. Cell Biol. 10, Panayotou, G., and Waterfield, M. D. (1993) BioEssays 15, Pawson, T., and Gish, G. D. (1992) Cell 71, Birge, R. B., and Hanafusa, H. (1993) Science 262, Feller, S. M., Ren, R., Hanafusa, H., and Baltimore, D. (1994) Trends Biochem. Sci. 19, Ullrich, A., and Schlessinger, J. (1990) Cell 61, Blenis, J. (1993) Proc. Natl. Acad. Sci. U. S. A. 90, Davis, R. J. (1993) J. Biol. Chem. 268, Ruderman, J. V. (1993) Curr. Opin. Cell Biol. 5, Maruta, H., and Burgess, A. W. (1994) BioEssays 16, Yarden, Y., and Ullrich, A. (1988) Annu. Rev. Biochem. 57, Cadena, D. L., and Gill, G. N. (1992) FASEB J. 6, Bollag, G., and McCormick, F. (1991) Annu. Rev. Cell Biol. 7, Haubruck, H., and McCormick, F. (1991) Biochim. Biophys. Acta 1072, Downward, J. (1992) BioEssays 14, Satoh, T., Nakafuku, M., and Kaziro, Y. (1992) J. Biol. Chem. 267, Marshall, M. S. (1993) Trends Biochem. Sci. 18, Lowy, D. R., and Willumsen, B. M. (1993) Annu. Rev. Biochem. 62, Polakis, P., and McCormick, F. (1993) J. Biol. Chem. 268, Pelech, S. L., and Sanghera, J. S. (1992) Science 257, Thomas, G. (1992) Cell 68, Leevers, S. J., and Marshall, C. J. (1992) Trends Cell Biol. 2, Nishida, E., and Gotoh, Y. (1993) Trends Biochem. Sci. 18, Egan, S. E., Giddings, B. W., Brooks, M. W., Buday, L., Sizeland, A. M., and Weinberg, R. A. (1993) Nature 363, Rozakis-Adcock, M., Fernley, R., Wade, J., Pawson, T., and Bowtell, D. (1993) Nature 363, Li, N., Batzer, A., Daly, R., Yajnik, V., Skolnik, E., Chardin, P., Bar-Sagi, D., Margolis, B., and Schlessinger, J. (1993) Nature 363, Kavanaugh, W. M., and Williams, L. T. (1994) Science 266, Buday, L., and Downward, J. (1993) Cell 73, Simon, M. A., Bowtell, D. D. L., Dodson, G. S., Laverty, T. R., and Rubin, G. M. (1991) Cell 67, Simon, M. A., Dodson, G. S., and Rubin, G. M. (1993) Cell 73, Olivier, J. P., Raabe, T., Henkemeyer, M., Dickson, B., Mbamalu, G., Margolis, B., Schlessinger, J., and Hafen, E. (1993) Cell 73, Chardin, P., Camonis, J. H., Gale, N. W., van Aelst, L., Schlessinger, J., Wigler, M. H., and Bar-Sagi, D. (1993) Science 260, Downward, J. (1996) Cancer Surv. 27, Sasaoka, T., Langlois, W. J., Leitner, J. W., Draznin, B., and Olefsky, J. M. (1994) J. Biol. Chem. 269, Yamauchi, K., and Pessin, J. E. (1994) J. Biol. Chem. 269, Leevers, S. J., Paterson, H. F., and Marshall, C. J. (1994) Nature 369, Stokoe, D., Macdonald, S. G., Cadwallader, K., Symons, M., and Hancock, J. F. (1994) Science 264, Moodie, S. A., Willumsen, B. M., Weber, M. J., and Wolfman, A. (1993) Science 260, van Aelst, L., Barr, M., Marcus, S., Polverino, A., and Wigler, M. (1993) Proc. Natl. Acad. Sci. U. S. A. 90, Zhang, X.-F., Settleman, J., Kyriakis, J. M., Takeuchi-Suzuki, E., Elledge, S. J., Marshall, M. S., Bruder, J. T., Rapp, U. R., and Avruch, J. (1993) Nature 364, Votjek, A. B., Hollenberg, S. M., and Cooper, J. A. (1993) Cell 74, Warne, P. H., Rodriguez-Viciana, P., and Downward, J. (1993) Nature 364, Morrison, D. K. (1995) Mol. Reprod. Dev. 42, Suen, K. L., Bustelo, X. R., and Barbacid, M. (1995) Oncogene 11, Freed, E., Symons, M., Macdonald, S. G., McCormick, F., and Ruggieri, R. (1994) Science 265, Fantl, W. J., Muslin, A. J., Kikuchi, A., Martin, J. A., MacNicol, A. M., Gross, R. W., and Williams, L. T. (1994) Nature 371, Hunter, T., and Karin, M. (1992) Cell 70, Carraway, C. A. C., and Carraway, K. L. (1996) in Treatise on the Cytoskeleton, Vol. 2 (Hesketh, H. E., and Pryme, I. F., eds) pp , JAI Press Inc., Greenwich, CT 58. Carraway, K. L., Carraway, C. A. C., and Carraway, K. L., III (1998) Signal Transduction and the Cytoskeleton, Springer-Verlag, Heidelberg 59. Carraway, K. L., and Carraway, C. A. C. (1989) Biochim. Biophys. Acta Biomembr. Rev. 988, Carraway, K. L., Huggins, J. W., Cerra, R. F., Yeltman, D. R., and Carraway, C. A. C. (1980) Nature 285, Carraway, C. A. C., Cerra, R. F., Bell, P. B., and Carraway, K. L. (1982) Biochim. Biophys. Acta 719, Carraway, C. A. C., Jung, G., and Carraway, K. L. (1983) Proc. Natl. Acad. Sci. U. S. A. 80, Carraway, C. A. C., Hua, F., Ye, X., Juang, S.-H., Liu, Y., Carvajal, M., and Carraway, K. L. (1991) J. Biol. Chem. 266, Liu, Y., Carraway, K. L., and Carraway, C. A. C. (1989) J. Biol. Chem. 264, Carraway, K. L., Doss, R. C., Huggins, J. W., Chesnut, R. W., and Carraway, C. A. C. (1979) J. Cell Biol. 83, Huggins, J. W., Trenbeath, T. P., Yeltman, D. R., and Carraway, K. L. (1980) Biochim. Biophys. Acta 127, Howard, S. C., Hull, S. R., Huggins, J. W., Carraway, C. A. C., and Carraway, K. L. (1982) J. Natl. Cancer Inst. 69, Juang, S.-H., Huang, J., Li, Y., Salas, P., Fregien, N., Carraway, C. A. C., and Carraway, K. L. (1994) J. Biol. Chem. 269, Juang, S.-H., Carvajal, M. E., Whitney, M., Liu, Y., and Carraway, C. A. C. (1996) Oncogene 12, Huang, J., Li, Y., Juang, S.-H., Mayer, B. J., Carraway, K. L., and Carraway, C. A. C. (1999) Oncogene 18, Carraway, C. A. C., and Weiss, M. (1985) Exp. Cell Res. 161, Carraway, C. A. C., Jung, G., Hinkley, R. E., and Carraway, K. L. (1985) Exp.