The G-protein G13 but Not G12 Mediates Signaling from Lysophosphatidic Acid Receptor via Epidermal Growth Factor Receptor to Rho
1998; Elsevier BV; Volume: 273; Issue: 8 Linguagem: Inglês
10.1074/jbc.273.8.4653
ISSN1083-351X
AutoresAntje Gohla, Rainer Harhammer, Günter Schultz,
Tópico(s)Sphingolipid Metabolism and Signaling
ResumoLysophosphatidic acid (LPA) utilizes a G-protein-coupled receptor to activate the small GTP-binding protein Rho and to induce rapid remodeling of the actin cytoskeleton. We studied the signal transduction from LPA receptors to Rho activation. Analysis of the G-protein-coupling pattern of LPA receptors by labeling activated G-proteins with [α-32P]GTP azidoanilide revealed interaction with proteins of the Gq, Gi, and G12 subfamilies. We could show that in COS-7 cells, expression of GTPase-deficient mutants of Gα12 and Gα13 triggered Rho activation as measured by increased Rho-GTP levels. In Swiss 3T3 cells, incubation with LPA or microinjection of constitutively active mutants of Gα12 and Gα13 induced formation of actin stress fibers and assembly of focal adhesions in a Rho-dependent manner. Interestingly, the LPA-dependent cytoskeletal reorganization was suppressed by microinjected antibodies directed against Gα13, whereas Gα12-specific antibodies showed no inhibition. The tyrosine kinase inhibitor tyrphostin A 25 and the epidermal growth factor (EGF) receptor-specific tyrphostin AG 1478 completely blocked actin stress fiber formation caused by LPA or activated Gα13 but not the effects of activated Gα12. Also, expression of the dominant negative EGF receptor mutant EGFR-CD533 markedly prevented the LPA- and Gα13-induced actin polymerization. Coexpression of EGFR-CD533 and activated Gα13 in COS-7 cells resulted in decreased Rho-GTP levels compared with expression of activated Gα13 alone. These data indicate that in Swiss 3T3 cells, G13 but not G12 is involved in the LPA-induced activation of Rho. Moreover, our results suggest an involvement of the EGF receptor in this pathway. Lysophosphatidic acid (LPA) utilizes a G-protein-coupled receptor to activate the small GTP-binding protein Rho and to induce rapid remodeling of the actin cytoskeleton. We studied the signal transduction from LPA receptors to Rho activation. Analysis of the G-protein-coupling pattern of LPA receptors by labeling activated G-proteins with [α-32P]GTP azidoanilide revealed interaction with proteins of the Gq, Gi, and G12 subfamilies. We could show that in COS-7 cells, expression of GTPase-deficient mutants of Gα12 and Gα13 triggered Rho activation as measured by increased Rho-GTP levels. In Swiss 3T3 cells, incubation with LPA or microinjection of constitutively active mutants of Gα12 and Gα13 induced formation of actin stress fibers and assembly of focal adhesions in a Rho-dependent manner. Interestingly, the LPA-dependent cytoskeletal reorganization was suppressed by microinjected antibodies directed against Gα13, whereas Gα12-specific antibodies showed no inhibition. The tyrosine kinase inhibitor tyrphostin A 25 and the epidermal growth factor (EGF) receptor-specific tyrphostin AG 1478 completely blocked actin stress fiber formation caused by LPA or activated Gα13 but not the effects of activated Gα12. Also, expression of the dominant negative EGF receptor mutant EGFR-CD533 markedly prevented the LPA- and Gα13-induced actin polymerization. Coexpression of EGFR-CD533 and activated Gα13 in COS-7 cells resulted in decreased Rho-GTP levels compared with expression of activated Gα13 alone. These data indicate that in Swiss 3T3 cells, G13 but not G12 is involved in the LPA-induced activation of Rho. Moreover, our results suggest an involvement of the EGF receptor in this pathway. Upon stimulation, a number of heterotrimeric G-protein-coupled receptors initiate cellular responses involving small GTP-binding proteins. For instance, the water-soluble phospholipid lysophosphatidic acid (LPA) 1The abbreviations used are: LPA, lysophosphatidic acid; G-protein, heterotrimeric guanine nucleotide-binding protein; PTX, pertussis toxin; EGF, epidermal growth factor; EGFR, EGF receptor; DMEM, Dulbecco's modified Eagle's medium; PAGE, polyacrylamide gel electrophoresis; FITC, fluorescein isothiocyanate. 1The abbreviations used are: LPA, lysophosphatidic acid; G-protein, heterotrimeric guanine nucleotide-binding protein; PTX, pertussis toxin; EGF, epidermal growth factor; EGFR, EGF receptor; DMEM, Dulbecco's modified Eagle's medium; PAGE, polyacrylamide gel electrophoresis; FITC, fluorescein isothiocyanate. binds to a G-protein-coupled heptahelical receptor (1Guo Z. Liliom K. Fischer D.J. Bathurst I.C. Tomei L.D. Kiefer M.C. Tigyi G. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 14367-14372Crossref PubMed Scopus (178) Google Scholar, 2Hecht J.H. Weiner J.A. Post S.R. Chun J. J. Cell Biol. 1996; 135: 1071-1083Crossref PubMed Scopus (656) Google Scholar) and stimulates the activation of Ras and Rho proteins via different pathways (3Howe L.R. Marshall C.J. J. Biol. Chem. 1993; 268: 20717-20720Abstract Full Text PDF PubMed Google Scholar, 4Ridley A.J. Hall A. Cell. 1992; 70: 389-399Abstract Full Text PDF PubMed Scopus (3785) Google Scholar). In fibroblasts, LPA lowers cAMP levels in a pertussis toxin (PTX)-sensitive manner, suggesting the coupling of the LPA receptor to G-proteins of the Gi subfamily (5Van Corven E.J. Groenink A. Jalink K. Eichholtz T. Moolenaar W.H. Cell. 1989; 59: 45-54Abstract Full Text PDF PubMed Scopus (674) Google Scholar). LPA-induced activation of the Ras/Raf/mitogen-activated protein kinase cascade also occurs in a PTX-sensitive fashion (3Howe L.R. Marshall C.J. J. Biol. Chem. 1993; 268: 20717-20720Abstract Full Text PDF PubMed Google Scholar, 6Van Corven E.J. Hordijk P.L. Medema R.H. Bos J.L. Moolenaar W.H. Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 1257-1261Crossref PubMed Scopus (334) Google Scholar), and Gβγ dimers of heterotrimeric G-proteins are thought to transduce this effect via intermediary protein-tyrosine kinases (7Crespo P. Xu N. Simonds W.F. Gutkind J.S. Nature. 1994; 369: 418-420Crossref PubMed Scopus (758) Google Scholar, 8Koch W.J. Hawes B.E. Allen L.F. Lefkowitz R.J. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 12706-12710Crossref PubMed Scopus (404) Google Scholar). In contrast, in a large variety of cells, the stimulation of phospholipase C induced by LPA was shown to be PTX-insensitive (9Moolenaar W.H. J. Biol. Chem. 1995; 270: 12949-12952Abstract Full Text Full Text PDF PubMed Scopus (566) Google Scholar), suggesting a coupling of the LPA receptor to members of the Gq subfamily. In addition to these pathways, LPA induces the Rho-dependent formation of focal adhesions and actin stress fibers in quiescent Swiss 3T3 fibroblasts (4Ridley A.J. Hall A. Cell. 1992; 70: 389-399Abstract Full Text PDF PubMed Scopus (3785) Google Scholar, 10Barry S.T. Critchley D.R. J. Cell Sci. 1994; 107: 2033-2045Crossref PubMed Google Scholar). A tyrosine kinase has been implicated in the LPA-induced Rho stimulation, since the tyrosine kinase inhibitor tyrphostin A 25 blocks the effects of LPA but not of microinjected Rho on actin polymerization (11Nobes C.D. Hawkins P. Stephens L. Hall A. J. Cell Sci. 1995; 108: 225-233Crossref PubMed Google Scholar). However, the G-proteins coupling the LPA receptor to Rho activation have not been specified. LPA-induced cytoskeletal reorganizations are PTX-insensitive and therefore presumably not transmitted via Gi subfamily members (12Ridley A.J. Hall A. EMBO J. 1994; 13: 2600-2610Crossref PubMed Scopus (439) Google Scholar). Gq proteins are apparently not involved in this pathway, because neither the activation of protein kinase C nor mobilization of intracellular calcium or expression of a GTPase-deficient mutant of Gαq induces the formation of actin stress fibers or focal adhesions (12Ridley A.J. Hall A. EMBO J. 1994; 13: 2600-2610Crossref PubMed Scopus (439) Google Scholar, 13Buhl A.M. Johnson N.L. Dhanasekaran N. Johnson G.L. J. Biol. Chem. 1995; 270: 24631-24634Abstract Full Text Full Text PDF PubMed Scopus (421) Google Scholar, 14Seufferlein T. Rozengurt E. J. Biol. Chem. 1994; 269: 9345-9351Abstract Full Text PDF PubMed Google Scholar). Interestingly, microinjection of constitutively active mutants of Gα12 and Gα13 into Swiss 3T3 cells triggers actin polymerization in a Rho-dependent manner (13Buhl A.M. Johnson N.L. Dhanasekaran N. Johnson G.L. J. Biol. Chem. 1995; 270: 24631-24634Abstract Full Text Full Text PDF PubMed Scopus (421) Google Scholar). In the present study, we determined the G-protein coupling pattern of the LPA receptor in membrane preparations and characterized the pathway leading from LPA receptor activation to stress fiber formation in intact cells. We report direct coupling of the LPA receptor to proteins of the G12, Gq, and Gi subfamilies. Furthermore, our data suggest that in intact cells G13 but not G12 mediates LPA-induced Rho activation involving the epidermal growth factor (EGF) receptor tyrosine kinase activity. Swiss 3T3 cells, kindly provided by Dr. Alan Hall (London), and COS-7 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum. To obtain quiescent and serum-starved Swiss 3T3 cells, cultures were rinsed three times in serum-free DMEM and incubated in DMEM supplemented with 25% Ham's F-12, 0.2% NaHCO3, 10 mm Na-Hepes, and 0.1% fetal bovine serum (modified DMEM) for 24 h followed by an incubation in modified DMEM without fetal bovine serum for 18 h. Membranes were prepared from subconfluent Swiss 3T3 cells. Cells were rinsed and scraped in a buffer containing 20 mm Tris/HCl, pH 8, 1 mm EDTA, 140 mm NaCl, 20 mm β-mercaptoethanol, and 10 mm phenylmethylsulfonylfluoride. The pelleted cells were resuspended in the buffer described above devoid of NaCl but containing 20 μg/ml leupeptin. Cells were disrupted by forcing the suspension 10 times through a 26-gauge needle. Undisrupted cells were pelleted (750 × g, 2 min), and the supernatant was centrifuged at 50,000 × g for 30 min at 4 °C. Membranes were resuspended and stored at −70 °C until use. Protein concentrations were determined according to Lowry et al. (15Lowry O.H. Rosebrough N.J. Farr A.L. Randall R.J. J. Biol. Chem. 1951; 193: 265-275Abstract Full Text PDF PubMed Google Scholar). For immunological detection of G-proteins, SDS-PAGE using 10% (w/v) polyacrylamide separating gels was employed. Proteins were transferred to nitrocellulose as described (16Harhammer R. Nürnberg B. Spicher K. Schultz G. Biochem. J. 1994; 303: 135-140Crossref PubMed Scopus (13) Google Scholar). The subtype specific antisera AS 232 and AS 233 (anti-α12); AS 343, AS 342, and AS 272 (anti-α13); AS 369 (anti-αq/11), and AS 266 (anti-αi) were characterized previously (17Spicher K. Kalkbrenner F. Zobel A. Harhammer R. Nürnberg B. Söling A. Schultz G. Biochem. Biophys. Res. Commun. 1994; 198: 906-914Crossref PubMed Scopus (62) Google Scholar). Antisera against Gα12 were incubated with the peptides (10 μg/ml) used for their generation, as published elsewhere (17Spicher K. Kalkbrenner F. Zobel A. Harhammer R. Nürnberg B. Söling A. Schultz G. Biochem. Biophys. Res. Commun. 1994; 198: 906-914Crossref PubMed Scopus (62) Google Scholar, 18Harhammer R. Nürnberg B. Harteneck C. Leopoldt D. Exner T. Schultz G. Biochem. J. 1996; 319: 165-171Crossref PubMed Scopus (15) Google Scholar). Detection of filter-bound proteins was carried out using the enhanced chemiluminescence (ECL) Western blotting system purchased from Amersham (19Harhammer R. Gohla A. Schultz G. FEBS Lett. 1996; 399: 211-214Crossref PubMed Scopus (28) Google Scholar). Synthesis of [α-32P]GTP azidoanilide from [α-32P]GTP (NEN Life Science Products) was performed according to published protocols (20Offermanns S. Schultz G. Rosenthal W. Methods Enzymol. 1991; 195: 286-301Crossref PubMed Scopus (59) Google Scholar). For photolabeling of Gα subunits, sedimented Swiss 3T3 membranes (200 μg of protein/tube) were preincubated in the absence or presence of 10 μm LPA at 30 °C, followed by the addition of [α-32P]GTP azidoanilide (3 μCi/tube; specific activity 3,000 Ci/mmol) at 30 °C for different times: 3 min for Gαi, 10 min for Gαq/11, and 30 min for Gα12 and Gα13 (18Harhammer R. Nürnberg B. Harteneck C. Leopoldt D. Exner T. Schultz G. Biochem. J. 1996; 319: 165-171Crossref PubMed Scopus (15) Google Scholar). Samples were irradiated for 15 s at 4 °C with a 254-nm UV lamp. Photolabeled membranes were pelleted and solubilized in 40 μl of 4% SDS (w/v) at room temperature, followed by the addition of precipitating buffer (280 μl) containing 10 mm Tris/HCl, pH 7.5, 1 mmEDTA, 150 mm NaCl, 1% deoxycholate (w/v), 1% Nonidet P-40 (w/v), 1 mm dithiothreitol, 0.2 μmphenylmethylsulfonylfluoride, and 10 μg/ml aprotinin. Samples were centrifuged (13,000 × g) for 5 min at 4 °C to remove insoluble material. The antisera (20 μl) AS 369 for Gαq/11, AS 233 for Gα12, AS 343 for Gα13, and AS 266 for Gαi were added to the supernatants followed by constant rotation of the samples for 2 h at 4 °C. Protein A-Sepharose beads (7 μg) were added, and samples were incubated at 4 °C overnight. The immunoprecipitates were washed three times, boiled in SDS-sample buffer, and subjected to SDS-PAGE. Gels were subsequently dried and analyzed using autoradiography. For transfection experiments, 1.2 × 105COS-7 cells were plated in 10-cm diameter cell culture dishes, grown overnight, and subsequently serum-starved for 18 h. Cells were then washed with phosphate-buffered saline, and 12 μg of DNA mixed with 60 μl of LipofectAMINE (Life Technologies, Inc.) in 6 ml of Opti-MEM (Life Technologies, Inc.) were added. In cotransfection experiments with two different plasmids, 6 μg of each plasmid were added. The total amount of DNA was kept constant by supplementing with DNA from a vector encoding β-galactosidase. Twelve h after transfection, COS-7 cell transfectants were labeled with [32P]orthophosphate (0.4 mCi/ml for 4 h) in phosphate-free Eagle's minimum essential medium. Cells were lysed on ice in 1 ml of 50 mm Hepes buffer, pH 7.4, containing 1% Triton X-100, 0.5% deoxycholate, 0.05% SDS, 150 mm NaCl, 5 mm MgCl2, 1 mm EGTA, 1 mg/ml of bovine serum albumin, 10 mm benzamidine, leupeptin-aprotinin-soybean trypsin inhibitor (10 μg/ml), and 1 mm phenylmethylsulfonylfluoride. Nuclei and cellular debris were sedimented, and NaCl was added to lysates at a final concentration of 500 mm. After preclearing for 30 min with 8 μg of protein A-Sepharose, samples were incubated with 4 μg of rabbit polyclonal anti-Rho A (Santa Cruz Biotechnology) for 60 min at 4 °C. Immunocomplexes were collected with 15 μg of protein A-Sepharose for 90 min. The beads were washed eight times in 50 mm Hepes buffer, pH 7.4, 500 mm NaCl, 0.1% Triton X-100, and 0.005% SDS. Nucleotides were eluted in 5 mm EDTA, 2 mm dithiothreitol, 0.2% SDS, 0.5 mm GTP, and 0.5 mm GDP for 20 min at 68 °C and separated by thin-layer chromatography on polyethyleneiminecellulose plates (Merck) run in 1 m KH2PO4, pH 3.4. The positions of unlabeled GDP and GTP standards on the plates were visualized under 254-nm UV light. Quantitations of radiolabeled nucleotides were performed with phosphoimaging (Fuji BAS 1500). For microinjection studies, quiescent, serum-starved Swiss 3T3 cells (approximately 103cells/mm2) plated on glass coverslips marked with squares to facilitate the localization of injected cells were used. The cDNAs of Gα12 (Q229L; Gα12QL), Gα13 (Q226L; Gα13QL), and Gαq(R183C; GαqRC), carried by the cytomegalovirus promotor-containing vector pCis, Gα11 (Q209L; Gα11QL) cloned in pZeo, and empty vector controls, were applied for microinjection. Plasmids were microinjected into the nucleus alone or with Texas red-coupled dextran (5 mg/ml, Molecular Probes) to visualize injected cells. Clostridium botulinumC3 toxin was co-microinjected with the cDNAs described above at a final concentration of 100 μg/ml. After microinjection, the cells were returned to the incubator for a further 90 min until fixation. As indicated, cells were treated with tyrphostin A 25 (150 μm), tyrphostin A 1 (150 μm), or genistein (30 μg/ml) for 60 min. Tyrphostin AG 1478 (1 μm) or tyrphostin AG 1296 (10 μm) were added 120 min before fixation. To study cytoskeletal effects of extracellular factors, cells were incubated with LPA (0.3 μm), thrombin (40 ng/ml), EGF (10 ng/ml), or bombesin (10 nm) for 10 min. The phosphatase inhibitor orthovanadate (100 μm) was added to serum-starved cells 25 min before fixation. Antibodies against Gα subunits were diluted and microinjected into the cytosol of cells followed by an incubation period of 60 min. The dominant negative EGF receptor mutant EGFR-CD533 carried by the pRK5 vector was expressed in Swiss 3T3 cells by nuclear microinjection. About 100 cells/field were injected in each case using a manual injection system (Eppendorf). Needles were pulled from capillaries with a horizontal micropipette puller (Sutter). Microinjected or growth factor-treated cells were fixed in 4% paraformaldehyde for 20 min and permeabilized in 0.2% Triton X-100 for 5 min. For localization of actin filaments, cells were stained with 0.5 μg/ml fluorescein isothiocyanate (FITC)-phalloidin (Sigma) for 40 min. Vinculin was detected after blocking with phosphate-buffered saline supplemented with 5% fetal bovine serum for 30 min, followed by incubation with a monoclonal anti-vinculin antibody (VIN-11–5; Sigma) for 60 min. Subsequently, cells were labeled with a FITC-conjugated goat anti-mouse antibody (Sigma) for 45 min. Antibody incubations were performed at room temperature. The coverslips were mounted on glass slides and examined on an inverted microscope (Zeiss Axiovert 100). To characterize the G-protein coupling pattern of the LPA receptor in Swiss 3T3 fibroblasts, we first determined the expression of different PTX-insensitive Gα subunits in these cells by performing immunoblot experiments with subtype-specific antibodies. Immunoblotting of membrane proteins with an antibody specific for Gαq/11(AS 369) identified the expression of the corresponding 42-kDa proteins (Fig. 1, lane 1). Antisera raised against a C-terminal peptide of Gα12 (AS 232, AS 233) recognized a protein with an apparent molecular mass of 43 kDa (Fig. 1, lanes 2 and 3). In addition, these antisera cross-reacted with several unknown proteins of higher and lower molecular masses that had not been detected in rodent and nonrodent brain membranes (17Spicher K. Kalkbrenner F. Zobel A. Harhammer R. Nürnberg B. Söling A. Schultz G. Biochem. Biophys. Res. Commun. 1994; 198: 906-914Crossref PubMed Scopus (62) Google Scholar). To assure the specificity of AS 232 and AS 233, which have been extensively characterized (17Spicher K. Kalkbrenner F. Zobel A. Harhammer R. Nürnberg B. Söling A. Schultz G. Biochem. Biophys. Res. Commun. 1994; 198: 906-914Crossref PubMed Scopus (62) Google Scholar, 18Harhammer R. Nürnberg B. Harteneck C. Leopoldt D. Exner T. Schultz G. Biochem. J. 1996; 319: 165-171Crossref PubMed Scopus (15) Google Scholar), we performed peptide competition experiments. Upon preincubation with the peptide used for their generation, the detection of Gα12was abolished (data not shown). Immunoblotting of membranes with an antiserum specific for Gα13 (AS 343) recognized a protein of approximately 43 kDa (Fig. 1, lane 4). These experiments indicate the expression of both Gα12 and Gα13 in Swiss 3T3 cell membranes. In previous studies, interactions of the LPA receptor with G-proteins of the Gi and Gq subfamilies were postulated mainly from experiments showing the existence of PTX-sensitive as well as PTX-insensitive cellular effects of LPA (9Moolenaar W.H. J. Biol. Chem. 1995; 270: 12949-12952Abstract Full Text Full Text PDF PubMed Scopus (566) Google Scholar). To directly show the LPA-mediated activation of G-proteins, membranes of subconfluent Swiss 3T3 cells were photolabeled with [α-32P]GTP azidoanilide in the presence or absence of LPA. G-protein Gα subunits were subsequently immunoprecipitated with anti-αq/11 (AS 369), anti-α12 (AS 233), anti-α13 (AS 343), or anti-αi (AS 266) antisera and visualized by autoradiography. Fig. 2 shows that stimulation of membranes with 10 μm LPA induced an incorporation of [α-32P]GTP azidoanilide into Gα subunits of Gq/11, G12, G13, and Gi proteins. For studying [α-32P]GTP azidoanilide incorporation into Gα12 and Gα13, the experiments were performed with prolonged incubation times of 30 min, whereas an agonist-induced activation of Gαi was only observed with short incubation times, reflecting the substantial basal guanine nucleotide turnover of Gi proteins. No increased incorporation of [α-32P]GTP azidoanilide into Gα12 or Gα13 was found using lower concentrations of LPA (e.g. 1 μm). Expression of constitutively active mutants of G12subfamily members has been reported to induce a Rho-dependent formation of actin stress fibers in Swiss 3T3 cells (13Buhl A.M. Johnson N.L. Dhanasekaran N. Johnson G.L. J. Biol. Chem. 1995; 270: 24631-24634Abstract Full Text Full Text PDF PubMed Scopus (421) Google Scholar). To further characterize the functional role of G12 and G13 in LPA-induced and Rho-mediated regulation of the actin cytoskeleton, we first expressed constitutively active mutants of Gα12 (Gα12 Q229L; Gα12QL) and Gα13 (Gα13 Q226L; Gα13QL) in Swiss 3T3 fibroblasts by microinjecting expression plasmids into quiescent cells. Both Gα subunits stimulated the formation of actin stress fibers, whereas GTPase-deficient forms of Gαq (Gαq R183C; Gαq RC) and Gα11 (Gα11 Q209L; Gα11 QL) were unable to induce stress fiber formation (Fig. 3). The expression of Gα12QL and Gα13QL was also accompanied by the formation of vinculin-containing spots with a characteristically elongated, arrowhead shape typical of Rho-induced assembly of focal adhesions (22Sekine A. Fujiwara M. Narumiya S. J. Biol. Chem. 1989; 264: 8602-8605Abstract Full Text PDF PubMed Google Scholar) that were not observed upon microinjection of plasmids encoding GαqRC and Gα11QL (Fig. 3). The dependence of Gα12QL- and Gα13QL-mediated cytoskeletal effects on stimulation of Rho activity could be demonstrated employing purified recombinant C3 exoenzyme from C. botulinum (13Buhl A.M. Johnson N.L. Dhanasekaran N. Johnson G.L. J. Biol. Chem. 1995; 270: 24631-24634Abstract Full Text Full Text PDF PubMed Scopus (421) Google Scholar). C3 exoenzyme has been shown to inhibit the isoforms of Rho but no other members of the Rho family (21Aktories K. Braun U. Rosener S. Just I. Hall A. Biochem. Biophys. Res. Commun. 1989; 158: 209-213Crossref PubMed Scopus (213) Google Scholar, 22Sekine A. Fujiwara M. Narumiya S. J. Biol. Chem. 1989; 264: 8602-8605Abstract Full Text PDF PubMed Google Scholar). Co-microinjection of C3 exoenzyme (100 μg/ml) abolished the actin stress fiber formation induced by Gα12QL and Gα13QL (Fig. 4). Likewise, focal adhesion assembly caused by both activated Gα subunits was inhibited by C3 exoenzyme (data not shown). These results confirm that G12 and G13 are able to trigger Rho-dependent effects in Swiss 3T3 cells, whereas cytoskeletal remodeling typical of an activation of Cdc42 or Rac,i.e. formation of filopodia or membrane ruffles (23Nobes C.D. Hall A. Cell. 1995; 81: 53-62Abstract Full Text PDF PubMed Scopus (3678) Google Scholar,24Ridley A.J. Paterson H.F. Johnston C.L. Diekmann D. Hall A. Cell. 1992; 70: 401-410Abstract Full Text PDF PubMed Scopus (3039) Google Scholar), was not observed. Interestingly, the phosphotyrosine phosphatase inhibitor orthovanadate (100 μm), which has previously been shown to stimulate Rho in Swiss 3T3 fibroblasts (11Nobes C.D. Hawkins P. Stephens L. Hall A. J. Cell Sci. 1995; 108: 225-233Crossref PubMed Google Scholar), also induced a selective formation of actin stress fibers without membrane ruffling or formation of filopodia (not shown). To test G12subfamily-induced Rho activation biochemically, we transiently overexpressed Gα12QL and Gα13QL in COS-7 cells and analyzed the incorporation of radioactive nucleotides in Rho A. Expression of activated Gα12 and Gα13, but not of the β-galactosidase control, caused an accumulation of Rho-bound radioactive GTP. The ratio of bound GTP to total nucleotides bound to Rho [(GTP/GTP + GDP) × 100] was 22 ± 0.57% for the lacZ control, 28.75 ± 0.92% for Gα12QL, and 35.7 ± 4.1% for Gα13QL (Fig. 5). These data demonstrate that expression of Gα12QL and Gα13QL leads to an activation of Rho.Figure 4Botulinum C3 exoenzyme abolishes the cytoskeletal effects induced by constitutively active Gα12, Gα13, and LPA. Serum-starved Swiss 3T3 cells were stimulated with LPA (200 ng/ml) or microinjected with expression plasmids (100 ng/μl) encoding Gα12(Q229L) or Gα13 (Q226L). Cells in the right panel were co-microinjected with botulinum C3 exoenzyme (100 μg/ml). Actin stress fibers were visualized by indirect immunofluorescence as described under "Experimental Procedures." Cells from one of three independent experiments are shown.View Large Image Figure ViewerDownload (PPT)Figure 5Rho-activation due to overexpression of constitutively active Gα12 and Gα13.COS-7 cells were transiently transfected with cDNAs encoding activated Gα12 and Gα13. After metabolic labeling with [32P]orthophosphate, cells were lysed and RhoA was immunoprecipitated. A, radioactive nucleotides bound to Rho were eluted and resolved by TLC. The position of GDP and GTP standards is indicated. B, ratio of GTP to total labeled nucleotides complexed to Rho [(GTP/GTP + GDP) × 100]. Data represent means ± S.D. (n = 2). Error bars not shown are masked by the symbols.View Large Image Figure ViewerDownload (PPT) To characterize the function of individual G12 subfamily members for LPA-induced activation of Rho in intact cells, we microinjected antibodies directed against the carboxyl termini of Gα12 and Gα13. Actin stress fiber formation caused by LPA at a concentration of 0.3 μm was suppressed by application of Gα13-specific antiserum AS 343 (dilution of 1:50), whereas microinjection of the corresponding preimmune serum failed to influence LPA-induced cytoskeletal effects. Inhibition of LPA-induced stress fiber formation was also obtained using other Gα13-specific antisera, i.e. AS 342 and AS 272 (not shown). Surprisingly, the Gα12-specific antisera (AS 232, AS 233) did not inhibit the LPA-induced actin polymerization using dilutions of 1:1 to 1:50 (Fig. 6). The application of antibodies against Gαq/11 (AS 369) showed no effects on cytoskeletal reorganisation caused by LPA. These results suggest that in intact Swiss 3T3 cells, LPA-induced Rho stimulation may be mediated by G13 but not by G12. Since the tyrosine kinase inhibitor tyrphostin A 25 has been shown to block stress fiber formation evoked by LPA, but not by microinjected Rho protein, an involvement of a tyrosine kinase acting upstream of Rho in the pathway of LPA-induced Rho activation appears to be likely (11Nobes C.D. Hawkins P. Stephens L. Hall A. J. Cell Sci. 1995; 108: 225-233Crossref PubMed Google Scholar). To further study the mechanism whereby G13 mediates the LPA-induced activation of Rho, we investigated effects of tyrosine kinase inhibitors on cytoskeletal changes induced by different stimuli. As demonstrated in Fig. 7 A, tyrphostin A 25 (150 μm) completely inhibited stress fiber formation induced by LPA and Gα13QL (100 ng/μl of cDNA) but not the cytoskeletal effects of Gα12QL (100 ng/μl of cDNA), further confirming the notion that LPA signals via G13 to Rho. Focal adhesion assembly caused by LPA or Gα13QL was also blocked by incubation with tyrphostin A 25, whereas focal adhesions induced by Gα12QL were not influenced by this tyrosine kinase inhibitor (Fig. 7 B). To exclude that these differences in tyrphostin sensitivity were results of different expression levels of Gα12QL and Gα13QL, we studied cytoskeletal effects caused by various concentrations of expression plasmids. Cytoskeletal effects of Gα13QL induced by injection of up to 1 μg/μl of cDNA were abolished by tyrphostin A 25. However, Gα12QL-induced submaximal stress fiber formation after microinjection of 10 ng/μl of cDNA was tyrphostin A 25-insensitive. Unlike tyrphostin A 25, the tyrosine kinase inhibitor genistein (30 μg/ml), which has previously been shown to act downstream of Rho (12Ridley A.J. Hall A. EMBO J. 1994; 13: 2600-2610Crossref PubMed Scopus (439) Google Scholar), blocked the formation of actin stress fibers induced by LPA, Gα13QL, and Gα12QL (Fig. 7 A). The biologically inactive tyrphostin A 1 (150 μm) showed no effects on cytoskeletal changes caused by LPA or the activated Gα subunits (data not shown). Using more specific inhibitors of tyrosine kinases, we found that 1 μm EGF receptor-specific tyrosine kinase inhibitor tyrphostin AG 1478 (25Levitzki A. Gazit A. Science. 1995; 267: 1782-1788Crossref PubMed Scopus (1608) Google Scholar) completely blocked both the LPA- and Gα13QL-induced formation of actin stress fibers (Fig. 8). Lower concentrations of tyrphostin AG 1478 (e.g. 250 or 500 nm) partially suppressed actin polymerization caused by Gα13QL. Cytoskeletal effects caused by Gα12QL were not influenced by this tyrphostin at concentrations of up to 20 μm. Tyrphostin AG 1296 (10 μm), a selective platelet-derived growth factor receptor tyrosine kinase inhibitor, failed to affect stress fiber formation induced by LPA and Gα13QL or by Gα12QL (Fig. 8). Actin stress fiber formation caused by orthovanadate was also inhibited by tyrphostin A 25 and tyrphostin AG 1478 but not by tyrphostin AG 1296 (data not shown).Figure 8Inhibition of G13-mediated Rho activation by tyrphostin AG 1478. Actin stress fiber formation is shown in Swiss 3T3 cells treated with LPA (0.3 μm) or injected with plasmids encoding constitutively active Gα12 or Gα13. As indicated, cells were incubated with tyrphostin AG 1478 (1 μm), tyrphostin AG 1296 (10 μm), or buffer (control). Actin cytoskeleton was visualized by staining of fixed cells with FITC-phalloidin. Cells from one of three independent experiments are shown.View Large Image Figure ViewerDownload (PPT) The inhibitory effects observed with tyrphostins suggested a participation of the EGF receptor in Gα13QL-induced Rho activation. Therefore, we expressed an EGF receptor mutant lacking 533 C-terminal amino acids (EGFR-CD533) in Swiss 3T3 cells by microinjection. It has been shown that this mutant exerts a dominant negative function on EGF receptor signaling by formation of signaling-incompetent heterodimers with the wild type receptor (26Redemann N. Holzmann B. von Ruden T. Wagner E.F. Schlessinger J. Ullrich A. Mol. Cell. Biol. 1992; 12: 491-498Crossref PubMed Scopus (119) Google Scholar). Fig. 9 shows that expression of the truncated receptor markedly suppressed actin stress fiber formation caused by LPA or by Gα13QL. In agreement with the tyrphostin data shown above, Gα12QL-induced actin polymerization was not inhibited by expression of EGFR-CD533. Furthermore, expression of the dominant negative EGF receptor mutant also abolished the cytoskeletal effects caused by orthovanadate (Fig. 9). To assess a role for the EGF receptor in Gα13-mediated Rho activation, we transiently coexpressed Gα13QL and EGFR-CD533 in COS-7 cells and determined the activation state of Rho by analyzing Rho-bound nucleotides. Fig. 10 demonstrates that the Gα13QL-induced Rho-GTP accumulation was attenuated by co-expression of the mutated EGF receptor. The ratio of bound GTP to total nucleotides bound to Rho [(GTP/GTP + GDP) × 100] was 17.8 ± 1.9% for the lacZ control, 25.8 ± 0.2% for the coexpression of Gα13QL with lacZ, and 22.5 ± 0.1% for the coexpression of Gα13QL with EGFR-CD533. This demonstrates that the dominant negative EGF receptor (EGFR-CD533) was able to inhibit Gα13-induced activation of Rho. The expression of EGFR-CD533 with lacZ did not affect basal Rho-GTP levels compared with lacZ (not shown). These data clearly suggest the EGF receptor as a critical component in the pathway leading from Gα13 to Rho activation.Figure 10Reduction of Gα13-induced Rho activation after expression of the dominant negative EGF receptor mutant EGFR-CD533. COS-7 cells were transfected with either lacZ alone or Gα13QL/lacZ or Gα13QL/EGFR-CD533. Cells were labeled with [32P]orthophosphate for 4 h, and RhoA was immunoprecipitated from the lysates. A, migration of Rho-associated radioactive nucleotides after separation by TLC in comparison to cold GDP/GTP standards. B, ratio of GTP to total labeled nucleotides complexed to Rho [(GTP/GTP + GDP) × 100]. Data represent means ± S.D. (n = 2). Error bars not shown are contained within the symbols.View Large Image Figure ViewerDownload (PPT) The phospholipid LPA has been characterized as a multifunctional messenger acting on adenylyl cyclases, phospholipases, mitogen-activated protein kinases, and the actin cytoskeleton via different pathways (9Moolenaar W.H. J. Biol. Chem. 1995; 270: 12949-12952Abstract Full Text Full Text PDF PubMed Scopus (566) Google Scholar). Previous studies demonstrated the existence of high affinity, guanine nucleotide-sensitive binding sites for LPA in membranes prepared from Swiss 3T3 cells and rat brains (27Thomson F.J. Perkins L. Ahern D. Clark M. Mol. Pharmacol. 1994; 45: 718-723PubMed Google Scholar). Recently, the cloning of LPA receptors from neuronal cells and Xenopus oocytes has been described (1Guo Z. Liliom K. Fischer D.J. Bathurst I.C. Tomei L.D. Kiefer M.C. Tigyi G. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 14367-14372Crossref PubMed Scopus (178) Google Scholar, 2Hecht J.H. Weiner J.A. Post S.R. Chun J. J. Cell Biol. 1996; 135: 1071-1083Crossref PubMed Scopus (656) Google Scholar). It has been suggested from experiments using PTX that some cellular effects of LPA are mediated by G-proteins belonging to Gi or Gq subfamilies (9Moolenaar W.H. J. Biol. Chem. 1995; 270: 12949-12952Abstract Full Text Full Text PDF PubMed Scopus (566) Google Scholar). The signaling cascade coupling the LPA receptor to an activation of Rho and subsequent remodeling of the actin cytoskeleton leading to stress fiber formation has, however, remained unclear. Results from previous studies indicated that LPA-induced Rho signaling is PTX-insensitive and that constitutively activated Gαq subunits are not able to stimulate Rho, excluding Gi and Gq proteins from this pathway (9Moolenaar W.H. J. Biol. Chem. 1995; 270: 12949-12952Abstract Full Text Full Text PDF PubMed Scopus (566) Google Scholar, 13Buhl A.M. Johnson N.L. Dhanasekaran N. Johnson G.L. J. Biol. Chem. 1995; 270: 24631-24634Abstract Full Text Full Text PDF PubMed Scopus (421) Google Scholar). On the other hand, expression of constitutively active mutants of Gα12 and Gα13 has been shown to trigger Rho-mediated stress fiber formation in Swiss 3T3 cells (13Buhl A.M. Johnson N.L. Dhanasekaran N. Johnson G.L. J. Biol. Chem. 1995; 270: 24631-24634Abstract Full Text Full Text PDF PubMed Scopus (421) Google Scholar). Therefore, we studied the involvement of G12 subfamily proteins in the pathway from LPA receptor to stimulation of Rho/stress fiber formation. Activated G12 and G13 induced Rho-mediated stress fiber formation (see Fig. 3; Ref. 13Buhl A.M. Johnson N.L. Dhanasekaran N. Johnson G.L. J. Biol. Chem. 1995; 270: 24631-24634Abstract Full Text Full Text PDF PubMed Scopus (421) Google Scholar). Furthermore, we were able to demonstrate activation of Rho by constitutively active mutants of Gα12 and Gα13 (see Fig. 5). In photolabeling experiments, both members of the G12subfamily appeared to interact with the LPA receptor; however, in intact cells, only G13, but not G12, mediated signaling from LPA receptor to Rho. This was suggested by discriminating effects of microinjected subtype-specific antibodies, by differences in tyrphostin sensitivity, and by the different requirement of EGF receptor function in LPA and Gα13QL-versus Gα12QL-induced Rho activation (see below). These data led us to identify G13 as a mediator from LPA receptors to Rho. A likely explanation for the observed discrepancy between the photoaffinity assay and microinjection/Rho activation studies is that photolabeling studies had to be performed employing a high concentration of LPA (10 μm), whereas a significantly lower concentration of 0.3 μm LPA was sufficient to induce maximum Rho activation in live Swiss 3T3 cells. Moreover, our data suggest that selectivity in receptor signaling is often found to be higher in intact complex systems than in in vitro assays. Similar conclusions may be deducible from the specificity of Gβγ dimers in signaling pathways, since studies with antisense oligonucleotides performed in intact cells revealed a higher selectivity than biochemical assays using purified components (28Dippel E. Kalkbrenner F. Wittig B. Schultz G. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 1391-1396Crossref PubMed Scopus (70) Google Scholar, 29Kalkbrenner F. Degtiar V. Schenker M. Brendel S. Zobel A. Hescheler J. Wittig B. Schultz G. EMBO J. 1995; 14: 4728-4737Crossref PubMed Scopus (69) Google Scholar, 30Ueda N. Iniguez-Lluhi J.A. Lee E. Smrcka A.V. Robishaw J.D. Gilman A.G. J. Biol. Chem. 1991; 269: 4388-4395Google Scholar, 31Wickman K.D. Iniguez-Lluhi J.A. Davenport P.A. Taussig R. Krapivinsky G.B. Linder M.E. Gilman A.G. Clapham D.E. Nature. 1994; 368: 255-257Crossref PubMed Scopus (377) Google Scholar). The identification of G13 as signal transducer from LPA receptor to Rho may suggest that other agonists of G-protein-coupled receptors that induce effects on the actin cytoskeleton may also act via G13 and/or G12. Previously, Nobes et al. (11Nobes C.D. Hawkins P. Stephens L. Hall A. J. Cell Sci. 1995; 108: 225-233Crossref PubMed Google Scholar) have indicated the existence of a protein-tyrosine kinase acting in the LPA pathway upstream of Rho. One major finding of the present study is that LPA- as well as Gα13QL-induced effects on the actin cytoskeleton can be blocked with tyrphostins, suggesting that a tyrosine kinase activity is required for the signaling from G13 to Rho. Although an inhibition of GTPase activity of transducin by tyrphostin A 25 has been reported (32Wolbring G. Hollenberg M.D. Schnetkamp P.P.M. J. Biol. Chem. 1994; 269: 22470-22472Abstract Full Text PDF PubMed Google Scholar), it appears unlikely that the suppression of Gα13QL-induced Rho activation by this inhibitor might be a result of nonspecific interactions between tyrphostin A 25 and the G-protein, since GTPase-deficient mutants of Gα subunits were used. Furthermore, also the specific EGF receptor inhibitor tyrphostin AG 1478 blocked the effects. The implication of EGF receptor kinase activity in the LPA/Gα13QL-induced, Rho-dependent stress fiber formation was substantiated by the expression of a dominant negative EGF receptor mutant in Swiss 3T3 cells. In addition, Gα13-induced Rho-GTP accumulation was at least partially reversed after overexpression of the truncated EGF receptor mutant. Recently, several G-protein-coupled receptors have been reported to induce a ligand-independent activation of the EGF or platelet-derived growth factor receptor with subsequent activation of mitogen-activated protein kinase cascades, and an involvement of Gi-type G-proteins in this transactivation pathway has been suggested (33Linseman D.A. Benjamin C.W. Jones D.A. J. Biol. Chem. 1995; 270: 12563-12568Abstract Full Text Full Text PDF PubMed Scopus (273) Google Scholar, 34Daub H. Weiss F.U. Wallasch C. Ullrich A. Nature. 1996; 379: 557-560Crossref PubMed Scopus (1312) Google Scholar, 35Luttrell L.M. Della Rocca G.J. van Biesen T. Luttrell D.K. Lefkowitz R.J. J. Biol. Chem. 1997; 272: 4637-4644Abstract Full Text Full Text PDF PubMed Scopus (425) Google Scholar). However, neither active Gαi nor Gβγ dimers lead to a reorganization of the actin cytoskeleton (13Buhl A.M. Johnson N.L. Dhanasekaran N. Johnson G.L. J. Biol. Chem. 1995; 270: 24631-24634Abstract Full Text Full Text PDF PubMed Scopus (421) Google Scholar). Our data suggest an involvement of the receptor tyrosine kinase EGF receptor in the pathway of G13-mediated actin stress fiber formation. The presumably complex mechanism whereby the EGF receptor influences signaling mediated by activated Gα13 has yet to be resolved. Using serum-starved Swiss 3T3 cells stimulated with LPA, we found neither a coimmunoprecipitation of Gα13with the EGF receptor or with a nonreceptor tyrosine kinase of the Src family nor a tyrosine phosphorylation of Gα13, arguing against the possibility of a direct interaction between Gα13 and a tyrosine kinase. 2A. Gohla and R. Harhammer, unpublished results. On the other hand, strong similarities were observed between cytoskeletal events induced by LPA or Gα13QL and those caused by orthovanadate, pointing to the possibility that activated Gα13 may cause a ligand-independent stimulation of the EGF receptor by interacting with a phosphotyrosine phosphatase. The identification of a putative receptor tyrosine kinase/phosphatase cycle controlling the communication between activated G13 and Rho is under current investigation. Taken together, our data indicate that G13 couples the LPA receptor to Rho activation in Swiss 3T3 cells. In addition, we provide evidence for an involvement of the EGF receptor in the pathway leading from G13 to Rho activation/stress fiber formation. We thank Nadine Albrecht for expert technical assistance. We are grateful to Dr. Klaus Aktories (Freiburg) for providing botulinum C3 exoenzyme, Dr. Alan Hall (London) for the donation of Swiss 3T3 fibroblasts, Dr. Melvin I. Simon (Pasadena) for providing cDNAs of wild type and constitutively active forms of Gαq, Gα11, Gα12, and Gα13, Dr. Christoph Sachsenmaier (München/Seattle) and Andreas Herrlich (Berlin/Karlsruhe) for the construct encoding the EGF receptor mutant EGFR-CD533, and Dr. Karsten Spicher (Berlin/Los Angeles) for providing antisera against Gα subunits. Furthermore, we thank Dr. Thomas Gudermann and Dr. Christian Harteneck for helpful suggestions and critical reading of the manuscript and Dr. Frank Kalkbrenner for help with the microinjection technique.
Referência(s)