α-Thrombin-mediated Phosphatidylinositol 3-Kinase Activation through Release of Gβγ Dimers from Gαq and Gαi2*

Chinese hamster embryonic fibroblasts (IIC9 cells) express the Gα subunits Gαs, Gαi2, Gαi3, Gαo, Gαq/11, and Gα13. Consistent with reports in other cell types, α-thrombin stimulates a subset of the expressed G proteins in IIC9 cells, namely Gi2, G13, and Gq as measured by an in vitro membrane [35S]guanosine 5′-O-(3-thio)triphosphate binding assay. Using specific Gα peptides, which block coupling of G-protein receptors to selective G proteins, as well as dominant negative xanthine nucleotide-binding Gα mutants, we show that activation of the phosphatidylinositol 3-kinase/Akt pathway is dependent on Gq and Gi2. To examine the role of the two G proteins, we examined the events upstream of PI 3-kinase. The activation of the PI 3-kinase/Akt pathway by α-thrombin in IIC9 cells is blocked by the expression of dominant negative Ras and β-arrestin1 (Phillips-Mason, P. J., Raben, D. M., and Baldassare, J. J. (2000) J. Biol. Chem. 275, 18046–18053, and Goel, R., Phillips-Mason, P. J., Raben, D. M., and Baldassare, J. J. (2002) J. Biol. Chem. 277, 18640–18648), indicating a role for Ras and β-arrestin1. Interestingly, inhibition of Gi2 and Gq activation blocks Ras activation and β-arrestin1 membrane translocation, respectively. Furthermore, expression of the Gβγ sequestrant, α-transducin, inhibits both Ras activation and membrane translocation of β-arrestin1, suggesting that Gβγ dimers from Gαi2 and Gαq activate different effectors to coordinately regulate the PI 3-kinase/Akt pathway.

Upon activation, GPCRs catalyze the exchange of GTP for GDP on the G␣ subunit of specific G protein trimers, promoting the dissociation of G␣-GTP from the G␤␥ subunits, which remain tightly associated (3). Both the G␣-GTP and the G␤␥ dimers stimulate effectors including phospholipase C (4), Ras (5), and PI 3-kinases (6,7).
Twenty distinct G␣ subunits have been cloned that can be divided into four families based on sequence homology: G s , G i , G q , and G 12 (8). Members of the G i family can be identified by their sensitivity to ADP-ribosylation by pertussis toxin (PTX), which uncouples the G protein from the receptor (9). In addition to the 20 G␣ subunits, 5 G␤ subunits and 12 G␥ subunits have been identified (10). The downstream events associated with a particular GPCR depend on the heterotrimeric G proteins that associate with the receptor. Several G protein-coupled receptors, including ␣-thrombin, are known to activate multiple G proteins and these G proteins can activate different signaling pathways (11)(12)(13)(14)(15)(16)(17)(18)(19)(20). IIC9 cells express G s , G i2 , G i3 , G o , G q/11 , and G 13 (21). Pertussis toxin-sensitive G proteins are involved in ␣-thrombin-induced Ras activation (22), and G q family members mediate the activation of PLC-␤1 (4) and protein kinase C (22).
We have previously found that in IIC9 cells ␣-thrombin induces PI 3-kinase and Akt activities that are dependent on Ras and ␤-arrestin1 (23,24). At present the G␣ subunits that mediate activation of PI 3-kinase and Akt by ␣-thrombin are unknown. This article identifies the specific G␣ subunits by using transient expression of specific G␣ peptides, which block coupling of GPCRs with selective G proteins, as well as dominant negative xanthine nucleotide-binding G␣ mutants. We show that ␣-thrombin stimulates G q , G 13 , and G i2 in IIC9 cells. Interestingly, PI 3-kinase and Akt stimulation are sensitive to uncoupling of both G q and G i2 . Efforts to delineate the individual role of G q and G i2 in this common pathway show that Ras activation and the translocation of ␤-arrestin1 to membranes are mediated by G␤␥ dimers released from G␣ i2 and G␣ q , respectively. These data indicate that ␣-thrombin induces PI 3-kinase and Akt stimulation via the G␤␥ dimers from G␣ q acting cooperatively with the G␤␥ dimers from G␣ i2 , and that PI 3-kinase activation is a point of convergence in the action of the two G proteins.
Western Blot Analysis-Growth-arrested IIC9 cells were incubated in the absence or presence of the indicated mitogen for the times indicated in the figure legends. At the indicated times, cell lysates were prepared as previously described (23). Protein lysates (10 -25 g) were resolved by SDS-PAGE and transferred to a polyvinylidene difluoride membrane (Millipore, Boston, MA). Membranes were probed with polyclonal antibodies to Akt (Santa Cruz Biotechnology), phospho-AktSer-473 (Cell Signaling), phospho-ERK (Cell Signaling), ERK (Santa Cruz Biotechnology), ␤-arrestin1 (Transduction Laboratories), and PI 3-kinase p85 antibody (Upstate Biotechnology). Immunoreactive bands were visualized by enhanced chemiluminescence (ECL) (Amersham Biosciences) as recommended by the manufacturer.
Preparation of IIC9 Membranes and ␤-Arrestin1 Translocation Assay-Plasma membranes from IIC9 cells were prepared as previously described (23). The protein amounts of ␤-arrestin1 were determined by Western blot Analysis and quantified by using Amersham Biosciences PhosphorImager TM .
Generation of PTX-resistant Mutants-PTX-resistant G␣ o , G␣ i2 , and G␣ i3 mutants were generated using PCR to amplify these inserts from their current pcDNA vectors with the addition of both 5Ј and 3Ј restriction sites and the COOH-terminal C 3 G mutation. For G␣ o a 5Ј HindIII site and a 3Ј XbaI site were introduced in addition to the C 3 G mutation using the following primers (5Ј 3 3 Ј): TATAAAGCTTGG-CCACCATGGGATGTACTC and TATATCTAGATCAGTACAAGCCTC-CGCCCCGGAGATTGTT. For G␣ i2 a 5Ј HindIII site and a 3Љ XhoI site were introduced in addition to the C 3 G mutation using the following primers (5Ј 3 3Ј): TATAAAGCTTGGCAGGATGGGCTGCA and TATA-CTCGAGTCAGAAGAGGCCTCCGTCCTTCAGGTTGTTCTTG. For G␣ i3 a 5Ј HindIII site and a 3Љ XhoI site were introduced in addition to the C 3 G mutation using the following primers (5Ј3 3Ј): TATAAAG-CTTGGCCGCCGTCATGGGCTGC and TATACTCGAGCCTCTCAGT-AAAGCCCACCTTCC. Restriction sites are underlined. PCR was carried out in the presence of 4 mM MgSO 4 using the Vent enzyme (Promega) and the following cycling conditions: 95°C, 5 min; 95°C, 1 min; 55°C, 1 min; 72°C, 1.5 min, 30 cycles. PCR products were restricted with the appropriate enzymes and subcloned into pcDNA3 (Invitrogen). Mutants were screened by restriction enzyme analysis and mutations were confirmed by sequencing using a PerkinElmer automated sequencer.
ADP Ribosylation Assay-IIC9 cells were pretreated with pertussis toxin at 100 ng/ml. IIC9 membranes were prepared as described above and resuspended in 30 l of 50 mM Tris-HCl, pH 7.5, 1 mM EDTA, and 1 mM MgCl 2 to each sample of 50 l of 2ϫ reaction buffer (2 mM ATP, 100 mM Tris-HCl, pH 7.5, 20 mM thymidine, 40 mM arginine, 0.4 mg/ml bovine serum albumin, 200 mM KPO 4 , pH 7.5, 10 mM ADP-ribose, 20 mM MgCl 2 , 2 mM EDTA, and 200 M GTP) and 4 l of activated pertussis toxin (pertussis toxin was activated by incubating 20 l of pertussis toxin (1 mg/ml) with 42 mM dithiothreitol, 20 mM Tris-HCl, pH 7.5, and incubated at 30°C for 30 min). Finally, 10 l of 5 M NAD containing 20 Ci of [ 32 P]NAD was added to each sample and incubated at 30°C for 30 min. The reaction was stopped by adding 100 l of NAD wash solution (5 mM NAD, 50 mM Tris-HCl, pH 7.5, 2.5 mM EDTA) to each sample. Membranes were pelleted by centrifugation at 14,000 ϫ g at 4°C for 5 min, washed twice with 200 l of NAD wash solution, resuspended in Laemmli sample buffer, and boiled for 5 min. Proteins were separated by SDS-PAGE gel containing 6 M urea and quantified using a Amersham Biosciences PhosphorImager TM .
Ras/Rho Activity Assays-Growth-arrested IIC9 cells or transfected cells were incubated in the presence or absence of 1 unit/ml ␣-thrombin for 5 min. After 5 min, the cells were washed two times at 4°C with PBS and harvested by scraping into 500 l of lysis buffer (50 mM Tris-HCl, pH 7.5, 20 mM MgCl 2 , 150 mM NaCl, 1% Triton X-100, 2 mM p-nitrophenyl phosphate, 10 g/ml pepstatin, 10 g/ml aprotinin, and 10 g/ml leupeptin), and the cell extracts were centrifuged for 2 min at 4°C. The cell extract was added to glutathione beads complexed with GST-RBD (fusion protein containing the Ras-binding domain of Raf-1 for Ras activity and fusion protein containing the RhoA binding domain of Rhotekin for Rho activity) for 1 h at 4°C. Samples were washed three times with 500 l of ice-cold lysis buffer, analyzed on 15% SDS-PAGE gel, and transferred to polyvinylidene difluoride. The membranes were probed with a 1/1000 dilution of a pan-Ras antibody (Santa Cruz) or RhoA monoclonal antibody (Santa Cruz). The amount of activated Ras or Rho (complexed with GTP) was visualized by ECL detection.
Assay of [ 35 S]GTP␥S Binding-Membranes (ϳ40 g of protein) from serum-arrested IIC9 cells were resuspended in 50 l of 50 mM Tris-HCl, pH 7.6, 2 mM EDTA, 100 mM NaCl, 1 mM MgCl 2 , 1 M GDP, 10 g/ml pepstatin, 10 g/ml aprotinin, 10 g/ml leupeptin, and 50 nM [ 35 S]GTP␥S (2000 Ci/mmol), and incubated in the presence or absence of 1 unit/ml ␣-thrombin at 37°C. After 5 min the reaction was terminated by the addition of 500 l of 50 mM Tris-HCl, pH 7.6, 150 mM NaCl, 20 mM MgCl 2 , 100 M GDP, 100 M GTP, 1% Nonidet P-40, 10 g/ml pepstatin, 10 g/ml aprotinin, and 10 g/ml leupeptin. The extracts were incubated with 3 l of preimmune serum and 200 l of pansorbin cells (Calbiochem) and centrifuged after 30 min to remove nonspecifically bound proteins. Extracts were incubated with antibody directed against a specific G␣ subunit for 1 h at 4°C. The immune complex was then incubated with 50 l of a 50% protein G-agarose suspension, the complexes were collected and washed three times in assay buffer plus 1% Nonidet P-40 and twice in assay buffer without detergent and the presence of G␣ subunits was analyzed by Western blot analysis using G␣-specific antibodies. [ 35 S]GTP␥S binding in the immunoprecipitates was quantified by scintillation counting.
Confocal Microscopy-Growth-arrested or transfected IIC9 cells grown on chamber slides (Nalgeen® Labware, Rochester, NY) were incubated in the absence or presence of 1 unit/ml ␣-thrombin for the indicated time points after preincubation in the absence or presence of 100 ng/ml PTX for 6 h. Subsequent to activation, the cells were fixed in a 3.7% formalin (Sigma) solution for 10 min at room temperature followed by 6 min incubation in ice-cold methanol at Ϫ20°C. The cells were washed in PBS and then blocked in 1 ml of blocking buffer (0.8 g of fatty acid-free bovine serum albumin (Sigma) in 100 l of PBS) for 2 h at room temperature. ␤-Arrestin1 monoclonal antibody was added at a 1:75 dilution (antibody:blocking buffer) and incubated at room temperature for 2 h. The cells were washed three times with PBS. The secondary antibody (Jackson ImmunoResearch Inc.) was added (1:5000 dilution in blocking buffer) for 45 min at room temperature. Again the cells were washed three times with PBS and then mounted using gel mount (Biomedia Corp, Foster City, CA) and microscope coverslips (Fisher Scientific, Pittsburgh, PA). Z series images were obtained using a Bio-Rad MRC 1024 confocal microscope. The acquired images were assembled using Adobe Photoshop and MS PowerPoint.
Co-immunoprecipitations-Growth-arrested or transfected IIC9 cells were incubated in the presence or absence of 1 unit/ml ␣-thrombin. Cell lysates were prepared and protein concentrations were determined as mentioned above. Protein lysates (100 g) were incubated with 5 g of ␤-arrestin1 antibody or PI 3-kinase p85 antibody at 4°C with gentle rocking for 2 h. The immune complexes were then immunoprecipitated as described previously (26), resolved by SDS-polyacrylamide gel electrophoresis, and transferred to a polyvinylidene difluoride. Membranes were then probed with antibodies to ␤-arrestin1, PI 3-kinase, or G proteins (Calbiochem). Immunoreactive bands were visualized by ECL detection.

Rapid PI 3-Kinase Activity Is Pertussis Toxin-sensitive and
Mediated by G␤␥ Dimers-Previous data from our laboratory (23) show that in IIC9 cells, ␣-thrombin activates rapid ␤-ar-restin1-, and Ras-dependent PI 3-kinase activity. Because changes in Akt phosphorylation and activity are dependent on PI 3-kinase activity (23), we quantified differences in Akt phosphorylation as a measure of changes in PI 3-kinase activity. Similar to our reported results (23, 24), ␣-thrombin induces a 10-fold increase in Akt activity as determined by Western blot analysis using antibodies directed against phosphorylated Akt. To determine whether the increase in PI 3-kinase is dependent on a member of PTX-sensitive G proteins, we treated IIC9 cells with 100 ng/ml pertussis toxin for 6 h prior to stimulation. Pretreatment with pertussis toxin completely blocks the ␣-thrombin-induced increase in Akt phosphorylation (Fig. 1A) and PI 3-kinase activity (Fig. 1C), indicating that the ␣-thrombin-induced PI 3-kinase/Akt pathway is dependent on a member of the G i/o subfamily of G proteins. Whereas pertussis toxin blocks ␣-thrombin-stimulated Akt phosphorylation, it is ineffectual on Akt phosphorylation by platelet-derived growth factor (Fig. 1A), which acts through a receptor tyrosine kinase and is insensitive to PTX. To determine whether this activation is mediated by the G␣ subunit or the G␤␥ dimers, we examined the effect of G␤␥ sequestrants on the ␣-thrombin-induced activation of PI 3-kinase and Akt. Transient expression of both G␣ t (Fig. 1B) and a membrane anchored G protein receptor kinase 3 (GRK3) carboxyl-terminal polypeptide, MAS-GRK3ct (data not shown), block ␣-thrombin-stimulated Akt phosphorylation by over 65%, whereas expression of the vector alone is ineffectual ( Fig. 1B). Similar results are seen for PI 3-kinase activity (Fig. 1C). Consistent with the known mechanism of signaling through tyrosine kinase receptors, G␣ t does not affect epidermal growth factor-induced PI 3-kinase activity (Fig. 1C). To ensure that the G␤␥ sequestrants do not affect a G␣ subunitmediated effect, we examined the result of expression of G␣ t on PIP 2 hydrolysis, which is known to be mediated by G␣ q (22). As previously found (13), addition of ␣-thrombin induces a 3-4fold increase in inositol phosphate release (Fig. 3B). Whereas expression of G␣ t blocks Akt phosphorylation (Fig. 1B), no change in the amount of thrombin-induced inositol phosphate release is observed (Fig. 3B). Taken together, these data indicate that ␣-thrombin stimulates rapid PI 3-kinase activity and Akt phosphorylation by the G␤␥ dimers released from a PTXsensitive G protein.
␣-Thrombin Stimulates Multiple Heterotrimeric G Proteins G q , G i2 , and G 13 -Several G protein-coupled receptors stimulate cellular responses by activating multiple G proteins that couple to different signaling pathways. Because in many cell types, ␣-thrombin activates several G proteins, including G q/11 , G 13/12 , and members of the pertussis toxin-sensitive G i subfamily, we next investigated the G proteins activated by ␣-thrombin in IIC9 cells (Fig. 2). Previously we found that IIC9 cells express G␣ s , G␣ i2 , G␣ i3 , G␣ o , G␣ q/11 , and G␣ 13 (21). We next examined ␣-thrombin-induced binding of [ 35 S]GTP␥S to specific G␣ subunits after immunoprecipitation of the G proteins with G␣-specific antibodies against G␣ s , G␣ i2 , G␣ i3 , G␣ o , G␣ q , and G␣ 13 (27,28) . Although significant binding of [ 35 S]GTP␥S occurs in unstimulated cells, ␣-thrombin induces a 4-fold increase in [ 35 S]GTP␥S binding to G␣ q , a 3-fold increase to G␣ 13 , and an approximate 5-fold increase to G␣ i2 ( Fig. 2A). ␣-Thrombin does not mediate significant binding to G␣ s , G␣ i3 , or G␣ o ( Fig. 2A). Because lysophosphatidic acid-mediated increases in Akt phosphorylation are blocked by expression of G␣ subunits COOH-terminal peptides to G␣ o2 and not G␣ i2 , 2 we reasoned that lysophosphatidic acid should couple to G o and not G i2 . Consistent with the Akt phosphorylation data, lysophosphatidic acid stimulates a significant increase in [ 35 S]GTP␥S binding to G␣ o but not G␣ i2 (Fig. 2B). Finally, treatment of isolated IIC9 membranes with the active peptide of cholera toxin induces detectable binding of [ 35 S]GTP␥S to only G s (Fig. 2C). The thrombin-induced increases in [ 35 S]GTP␥S binding to G␣ i2 , G␣ q , and G␣ 13 indicate that ␣-thrombin couples to, and therefore can signal through, these G proteins. Because some of the G proteins may be expressed at levels that are difficult to detect, and we have not identified ligands that stimulate all of FIG. 1. PI 3-kinase pathway is pertussis toxin-sensitive and mediated by G␤␥ dimers. Cells were untreated or transfected with pcDNA (Invitrogen) or pcDNA containing G␣ t . Growth-arrested IIC9 cells were untreated or incubated in the presence of 100 ng/ml PTX 6 h prior to stimulation with the indicated mitogen for 5 min (1 unit/ml ␣-thrombin, 10 ng/ml platelet-derived growth factor (PDGF), or 10 ng/ml epidermal growth factor (EGF)). A and B, lysate proteins (20 g) were separated by SDS-PAGE, immunoblotted with either anti-Akt or anti-phospho-Akt polyclonal antibody, and the blots were quantified using a Amersham Biosciences densitometer. Relative activation of Akt (p-Akt/Akt) is shown, with activity of serum-arrested IIC9 as 1.0. Results for both A and B are representative blots of four independent experiments. C, PI 3-kinase complexes were immunoprecipitated from lysates containing equal proteins using an anti-p85 polyclonal antibody and assayed for the ability to phosphorylate PI in vitro. PI 3-kinase activity was quantified as previously described (1). The data shown are the mean Ϯ S.D. for triplicates in one experiment and are representative of four independent experiments. The Western blot is a representative blot. the G proteins expressed, it is not possible to determine the G proteins activated by ␣-thrombin with certainty. However, these studies suggest that ␣-thrombin couples primarily to G i2 , G q , and G 13 . Furthermore, we find no detectable increases in [ 35 S]GTP␥S binding to other members of the pertussis toxin class as observed in other cell types (19).
G␣ q and G␣ i2 Are Critical for ␣-Thrombin-stimulated Akt Phosphorylation-Because pertussis toxin blocks the ␣-thrombin-induced increase in Akt phosphorylation (Fig. 1A), and ␣-thrombin induces a significant increase in GTP␥S binding to G i2 (Fig. 2), we reasoned that G i2 mediates PI 3-kinase stimulation. To further demonstrate the involvement of G i2 , we expressed peptides containing the COOH terminus of G␣ subunits (Fig. 3A). Expression of the carboxyl terminus G␣ peptides of specific G␣ subunits inhibits binding and activation of the targeted G␣ subunit (14,29), indicating that these peptides effectively inhibit specific GPCR-mediated responses. Consistent with the data in Figs. 1 and 2, expression of the COOH-terminal peptide of G␣ i1/2 inhibits ␣-thrombin-induced Akt phosphorylation (Fig. 3A). Furthermore, expression of G␣ 13 (Fig. 3A), G␣ 01 , and G␣ 02 (data not shown) is ineffective. Surprisingly, rapid Akt phosphorylation is inhibited by expression of the COOH-terminal peptide of G␣ q (Fig. 3A), suggesting that ␣-thrombin-induced PI 3-kinase activation is dependent on G␣ q . We obtained similar results for PI 3-kinase activity (data not shown). The unexpected activity of the COOH-terminal peptide of G␣ q prompted us to characterize these peptides further. We reasoned that PIP 2 hydrolysis and Rho activity would be dependent on G q and G 13 (14), respectively. Whereas expression of the G␣ q COOH-terminal peptide blocks PIP 2 hydrolysis, expression of the G␣ i1/2 COOH-terminal peptide or PTX treatment does not (Fig. 3B). Furthermore, only expression of the COOH-terminal peptide of G␣ 13 inhibits Rho activation, whereas G␣ i1/2 and G␣ q are ineffective (Fig. 3C). We also measured the ability of the COOH-terminal peptides to inhibit sustained ERK phosphorylation, which is dependent on PTX-sensitive G proteins (G i2 ). We find that expression of the COOH-terminal peptide of G␣ i1/2 inhibits sustained ERK activation, whereas G␣ 13 and G␣ q do not inhibit it (Fig. 3C). Taken together these data indicate that the COOH-terminal peptides effectively inhibit targeted G␣ subunits and that Akt phosphorylation is dependent on both G q and G i2 .
We were surprised that Akt phosphorylation is sensitive to uncoupling of G␣ q from ␣-thrombin. To ensure that these results are not selective for COOH-terminal peptides, we decided to use another approach to inhibit G protein activation, utilizing xanthine nucleotide-binding G␣ mutants (G␣Xs). Expression of G␣X inhibits coupling of GPCR, including the thrombin receptor, to specific G proteins indicating that these G␣Xs act as dominant negative inhibitors (30). We transiently transfected IIC9 cells with G␣ i2 X, G␣ q X, and G␣ 13 X, and quantified Akt phosphorylation (Fig. 3A). Similar to the results with the COOH-terminal peptides, G␣ i2 X and G␣ q X block Akt phosphorylation, whereas G␣ 13 X does not. Consistent with the ability of these G␣X constructs to inhibit the activation of specific G proteins by ␣-thrombin, expression of G␣ q X blocks and G␣ i2 X does not inhibit PIP 2 hydrolysis (data not shown). Taken together, these data show that ␣-thrombin-induced PI 3-kinase activity and Akt phosphorylation is dependent on both G q and G i2 . Membranes were prepared from serum-arrested IIC9 cells and the G protein immune complexes containing specific G proteins were immunoprecipitated as described under "Materials and Methods" after treatment with: A, 1 unit/ml thrombin; B, 20 M lysophosphatidic acid; or C, 1 g/assay of cholera toxin. The presence of G␣ subunits was analyzed by Western blot analysis using the indicated G␣-specific antibodies. In control samples, membranes proteins were immunoprecipitated with anti-G␣ subunit antibody (Calbiochem, CA).

␣-Thrombin-stimulated Akt Phosphorylation in the Presence of
Pertussis Toxin-Because our data show that ␣-thrombin-induced Akt phosphorylation is dependent on both G q and G i2 , we reasoned that expression of a pertussis toxin-resistant mutant of G␣ i2 should rescue pertussis toxin inhibition of ␣-thrombininduced Akt phosphorylation. Pertussis toxin-resistant mu-tants were generated by converting the COOH-terminal of G␣ from a cysteine residue to a glycine (G␣ o C351G, G␣ i2 C352G, and G␣ i3 C351G). To verify that the C 3 G mutation renders the expressed G␣ subunits resistant to pertussis toxin treatment, IIC9 cells were transiently transfected with vector alone or vector containing G␣ o C351G, G␣ i2 C352G, or G␣ i3 C351G,

FIG. 3. Effect of COOH-terminal peptides of G␣ subunits on ␣-throm-
bin-stimulated effectors. IIC9 cells were untreated or transfected with the COOH-terminal peptides of G␣ subunits and then serum arrested. A, inhibition of G␣ q or G␣ i2 activation blocks ␣-thrombinstimulated Akt phosphorylation. Cells were transfected with the G␣Xs mutants or the COOH-terminal peptides as indicated in figure. The serum-arrested cells were untreated or treated with 100 ng/ml PTX for 6 h prior to incubation in the absence or presence of 1 unit/ml ␣-thrombin for 5 min. Cell lysates (20 g) were separated by SDS-PAGE, and immunoblotted with either anti-Akt or anti-phospho-Akt polyclonal antibody. Blots were quantified using a Amersham Biosciences densitometer. The data shown are the mean Ϯ S.D. for triplicates in one experiment and are representative of three independent experiments. The Western blot is a representative blot. B, effect of COOH-terminal peptides on PIP 2 hydrolysis. Serum-arrested cells were labeled with myo-[ 3 H]inositol (1 Ci/ml) for 24 h, and then incubated in serum-free medium supplemented with 20 mM LiCl in the absence or presence of 1 unit/ml ␣-thrombin for 20 min at 37°C, and inositol trisphosphate amounts were quantified. The data shown are the mean Ϯ S.D. for duplicates in one experiment and are representative of three independent experiments. C, effect of COOH-terminal peptides on Rho activity and sustained ERK phosphorylation. Serum-arrested cells were incubated in the absence or presence of 1 unit/ml ␣-thrombin for 5 min and 4 h for Rho activity and sustained ERK phosphorylation, respectively. For Rho activity, cells lysates were prepared, incubated with recombinant GST-RBD for 1 h, and the immunocomplexes were subjected to SDS-PAGE and Western blot analysis with anti-RhoA antibody. For ERK activity the cell lysates were subjected to SDS-PAGE and Western blot analysis with anti-phospho-ERK polyclonal antibody. Relative activation of RhoA (complexed with GTP) and ERK are shown in the same graph. We calculated these activities in the following manner. We subtracted the basal activities (for Rho and phospho-ERK) from the ␣-thrombin-induced activities in the untransfected cells and set these values equal to 100%. The transfected samples are calculated by subtracting the basal from the experimental and dividing these values by the untransfected values times 100. The data shown are representative of three independent experiments. treated with 100 ng/ml pertussis toxin for 10 h, and then cell lysates were analyzed by SDS-PAGE in the presence of 6 M urea (Fig. 4A). Previous data in IIC9 cells show that pretreatment with 100 ng/ml pertussis toxin for 10 h results in the ADP-ribosylation of more than 95% of the pertussis toxinsensitive G proteins (22). For each construct, in the absence of pertussis toxin Western blot analysis with an antibody that recognizes all G i family members shows a single band with apparent molecular weight of 43,000 (Fig. 4A). Treatment with pertussis toxin results in a single band shifted to higher molecular weight (Fig 4B). In the presence of pertussis toxin, Western blot analysis of lysate proteins from cells expressing each construct shows a band at 43 kDa and a faster migrating band (Fig. 4B). Presumably, the lower band is the unmodified G␣ subunit and is absent in lysates from cells transfected with empty vector. Furthermore, transient expression of the three G␣ C 3 G constructs results in similar protein levels (Fig. 4B) ensuring that the ability to rescue is independent of expression levels. These data clearly demonstrate that treatment with pertussis toxin for 10 h results in ADP-ribosylation of the endogenous G i family members and the G␣ C 3 G mutants are resistant to pertussis toxin. In agreement with our data indicating a role for G i2 in ␣-thrombin-stimulated Akt phosphorylation, expression of G␣ i2 C352G but not G␣ o C351G or G␣ i3 C351G rescues ␣-thrombin-induced rapid Akt phosphorylation in IIC9 cells treated with pertussis toxin (Fig. 4C).
␣-Thrombin-induced Translocation of ␤-Arrestin1 to the Plasma Membrane Is Mediated by G␤␥ Released from G␣ q -Because ␤-arrestin1 is required for thrombin-induced PI 3-ki-nase and Akt activities in IIC9 cells (23), we addressed the possibility that ␤-arrestin1 translocation is dependent on a single G-protein. We have previously established that treatment of IIC9 cells with ␣-thrombin results in the translocation of ␤-arrestin1 to the plasma membrane (23). We have also shown that expression of a membrane-anchored GRK3 carboxyl-terminal polypeptide (MAS-GRK3ct) blocks the translocation, indicating the involvement of a G␤␥ (23). We next examined the effect of expression of COOH-terminal peptides on ␣-thrombin-induced translocation of ␤-arrestin1 (Fig. 5) by both biochemical fractionation (Fig. 5A) and confocal microscopy (Fig. 5, B and C). We found it necessary to transiently express ␤-arrestin1 to examine the translocation. ␣-Thrombin induces a significant increase in the translocation of ␤-arres-tin1 to the membrane fraction within 5 min (Fig. 5A) and in the translocation of a substantial fraction of the cellular fluorescence to the plasma membrane within 10 -30 min (Fig. 5B, 2  and 3). After 30 min, ␤-arrestin1 is found mainly in the cytoplasm (Fig. 5B, 4). Whereas expression of the COOH-terminal peptide of G␣ q inhibits translocation (Fig. 5, A and C, 4), treatment with PTX or expression of the COOH-terminal peptide of G␣ i1/2 does not block the amount of ␤-arrestin1 translocation (Fig. 5, A and C, 3 and 6). However, transient co-expression of G␣ t blocks the increase (Fig. 5A) and also results in an evenly distributed cytoplasmic fluorescence (Fig. 5C, 5). These data indicate that ␤-arrestin1 translocation to the plasma membrane in IIC9 cells is mediated by the G␤␥ released from G␣ q and suggest that G i2 affects PI 3-kinase activation by another mechanism .   FIG. 4. C 3 G G i mutants are not ADP-ribosylated. Cells were transiently transfected with either vector (pcDNA), or pcDNA containing G o ␣C351G, G i2 ␣-C352G, and G i3 ␣C351G. The transfected cells were growth-arrested, and cell lysates prepared from: A, cells treated with 100 ng/ml PTX for 6 h prior to the preparation of lysates; or B, untreated cells. Lysate proteins (20 g) were separated by SDS-PAGE containing 6 M urea and immunoblotted with a polyclonal antibody to an internal sequence recognized by G␣ subunits. C, expression of PTX-insensitive G␣ i2 rescues ␣-thrombin-stimulated Akt phosphorylation in IIC9 cells treated with PTX. The transfected cells were incubated in the presence or absence of 100 ng/ml pertussis toxin 6 h prior to stimulation with 1 unit/ml ␣-thrombin for 5 min and cell lysates were prepared. Lysate proteins (20 g) were immunoblotted with anti-phospho-Akt or anti-Akt polyclonal antibody. Akt phosphorylation was quantified using a Amersham Biosciences PhosphorImager TM and reported as percent maximal stimulation. Data in A and B are representative of three independent experiments. The data shown in C are the mean Ϯ S.D. for triplicates in one experiment and are representative of four independent experiments. The Western blot is a representative blot.

FIG. 5. Expression of G␤␥ sequestrant or COOH-terminal G␣ q blocks
␣-thrombin-induced translocation of ␤-arrestin1. IIC9 cells were transfected with ␤-arrestin1 alone or with G␣ t or the COOH-terminal peptides of G␣ q or G␣ i1/2 . The cells were then stimulated with 1 unit/ml ␣-thrombin. A, membranes were prepared after stimulation for 15 min. Membrane protein lysates (50 g) were separated by SDS-PAGE, immunoblotted, and the amount of ␤-arrestin1 was quantified with a Amersham Biosciences densitometer. The data shown are the mean Ϯ S.D. for triplicates in one experiment and are representative of three independent experiments. The Western blot is a representative blot. B and C, IIC9 cells were grown on chamber slides and transfected as described above. The cells were fixed, permeabilized, and visualized by confocal microscopy as described under "Materials and Methods." B, ␤-arrestin1 was visualized after 0 min, 10 min, 30 min, and 1 h post-stimulation with ␣-thrombin. C, the cells were serum arrested, and were untreated or treated for 4 h with 100 ng of PTX. After treatment the cells were stimulated with 1 unit/ml ␣-thrombin for 30 min. Data in B and C are representative of three independent experiments.
␣-Thrombin-mediated Formation of Complexes Containing PI 3-Kinase and ␤-Arrestin1-Because the ability of ␤-arres-tin1 to associate with members of mitogen-activated protein kinase pathways is known to affect activation of these pathways (31), we reasoned that ␤-arrestin1 serves as an adapter molecule, associating with and recruiting PI 3-kinase to the membrane. To determine whether ␣-thrombin induces association of ␤-arrestin1 with PI 3-kinase, we immunoprecipitated ␤-arrestin1 pre and post ␣-thrombin stimulation and examined the immunoprecipites for the presence of PI 3-kinase (Fig. 6). In the absence of ␣-thrombin, low amounts of PI 3-kinase are found in the immunoprecipitates (Fig. 6). ␣-Thrombin induces an increase of PI 3-kinase in ␤-arrestin1 immunoprecipitates within 15 min that decreases to amounts similar to those found in unstimulated cells after 30 min (data not shown). Similar results are found when PI 3-kinase immunoprecipitates are assayed for the presence of ␤-arrestin1 (data not shown). Expression of the G␣ q COOH-terminal peptide inhibits complex formation, whereas the G␣ i1/2 COOH-terminal peptide does not (Fig. 6). Taken together, these data indicate that ␣-thrombin stimulates membrane translocation of ␤-arrestin1 and its association with PI 3-kinase suggesting that, as is found in mitogen-activated protein kinase pathway activation (32), ␤-arrestin1 functions as an adaptor molecule recruiting PI 3-kinase.
G i2 Mediates PI 3-Kinase Activation by ␣-Thrombin via Ras-Whereas our results clearly demonstrate a role for G q in the activation of PI 3-kinase, the data do not provide a mechanistic role for G i2 . Previously we found that Ras is required for ␣-thrombin-induced PI 3-kinase activation (24). Furthermore, Ras stimulation by ␣-thrombin is independent of ␤-arrestin1 (23), suggesting that Ras activation is independent of G q . These data suggest that G i2 may affect PI 3-kinase via Ras. To test this hypothesis, we next examined the role of G proteins in the activation of ␣-thrombin-induced Ras activation using the RBD fragment of the Ras effector Raf-1 (33). Treatment of IIC9 cells with ␣-thrombin results in an approximate 12-fold increase of GTP-bound Ras (Fig. 7). The ␣-thrombin-induced increase is blocked by pretreatment of IIC9 cells with pertussis toxin or expression of the G␣ i1/2 COOH-terminal peptide (Fig. 7A). In contrast to the ability of the G␣ i1/2 COOH-terminal peptide to block the Ras activation, expression of the G␣ q COOH-terminal peptide does not affect the increase (Fig. 7A). Also we find that expression of pertussis toxin-insensitive G␣ i2 C352G but not G␣ o C351G nor G␣ i3 C351G rescues ␣-thrombin-induced Ras activation in IIC9 cells treated with PTX (Fig. 7B), confirming the ability of the COOH-terminal peptide of G␣ i1/2 to block Ras activation.
Because GPCRs that are sensitive to pertussis toxin often stimulate Ras through G␤␥ subunits (5), we reasoned that expression of G␤␥ sequestrants would block Ras activation by ␣-thrombin. Expression of G␣ t completely abolishes ␣-thrombin-mediated GTP binding to Ras (Fig. 7A). Taken together, these data demonstrate the involvement of G␤␥ dimers from G␣ i2 in ␣-thrombin-stimulated Ras and thus a role in PI 3-kinase activation. DISCUSSION ␣-Thrombin stimulates multiple effectors in fibroblasts (1,22,34), including PI-PLC, ERK, and PI 3-kinase activities. In this paper, we identify the G proteins that mediate the activation of PI 3-kinase and Akt by ␣-thrombin. We find that treatment of IIC9 cells with PTX markedly inhibits PI 3-kinase and Akt activation in response to ␣-thrombin (Fig. 1). Because PTX catalyzes the ADP-ribosylation and inactivation of members of the G i/o family (9), these data indicate the involvement of G i/o . Consistent with these data, ␣-thrombin activates the PTXsensitive G-protein, G i2 (Fig. 2). We then reasoned that G i2 mediates PI 3-kinase and Akt activation in response to ␣-thrombin. Consistent with this notion, expression of either G␣ i1/2 COOH-terminal blocking peptide (29) or the dominant negative G␣ i2 X mutant (30) blocks Akt phosphorylation and PI 3-kinase activity (Fig. 3). Surprisingly, expression of a G␣ q COOH-terminal peptide, which blocks the activation of G q , or the dominant negative G␣ q X mutant, inhibits Akt phosphorylation (Fig. 3A), indicating an important role for G q in PI 3-kinase activation by ␣-thrombin.
Previously, we reported that ␤-arrestin1 is required for ␣-thrombin-induced PI 3-kinase and Akt activities in IIC9 cells (23,35). However, we had not identified the G proteins that mediate activation of the PI 3-kinase/Akt pathway. Expression of a dominant negative G␣ q X mutant and a G␣ q COOH-terminal peptide blocks ␤-arrestin1 translocation, whereas expression of dominant negative G␣ i2 X mutant or the G␣ i1/2 COOHterminal peptide is ineffective (Fig. 5). Furthermore, ␤-arrestin1 translocation is dependent on the G␤␥ subunits of G␣ q , indicating an essential role for G␤␥ in ␤-arrestin1 translocation.
In addition to G q , there is also a role for G i2 in ␣-thrombininduced PI 3-kinase and Akt stimulation. Our results demonstrate that Ras is upstream of PI 3-kinase activation and the dependence of PI 3-kinase on Ras is unrelated to the dependence of PI 3-kinase on ␤-arrestin1. First, inhibition of G␣ i2 activation abrogates Ras activation (Fig. 7A). Second, expression of only PTX-resistant G␣ i2 rescues ␣-thrombin-induced Ras activation in PTX-treated IIC9 cells (Fig. 7B). Furthermore, ␣-thrombin induces Ras via the G␤␥ dimers released from G␣ i2 (Fig. 7A). Taken together, these results indicate that ␣-thrombin stimulates Ras via the G␤␥ from G i2 .
We were surprised to find that the translocation of ␤-arres-tin1 and the activation of Ras were mediated by the G␤␥ subunits from G q and G i2 , respectively. These results show that the G␤␥s associated with G q and G i2 can activate different effectors. There are several possible explanations to account for the specificity of G␤␥ signaling of G q and G i2 . It is possible that FIG. 6. ␣-Thrombin-mediated formation of complexes containing PI 3-kinase and ␤-arrestin1. IIC9 cells were transiently cotransfected with 1 g/ml ␤-arrestin1 and COOH-terminal peptides of G␣ i1/2 or G␣ q . The cells were serum arrested, and treated with or without 1 unit/ml ␣-thrombin for 15 min. Protein lysates (100 g) were incubated with antibodies directed against ␤-arrestin1 or PI 3-kinase p85 for 2 h. The immune complexes were immunoprecipitated as described under "Materials and Methods." Samples were subjected to SDS-PAGE and probed with the indicated antibodies directed against ␤-arrestin1 or PI 3-kinase p85 and the amount of ␤-arrestin1 or PI 3-kinase p85 was quantified with a Amersham Biosciences densitometer. This is a representative blot of five independent experiments FIG. 7. G i2 mediates Ras activation. Cells were transfected with: A, G␣ t or the COOH-terminal peptides G␣ i1/2 or G␣ q ; or B, pcDNA containing G␣ o C351G, G␣ i2 C352G, and G␣ i3 C351G. The cells were growth arrested. Growth-arrested cells were incubated in the presence or absence of 100 ng/ml PTX 6 h prior to stimulation and then incubated for 15 min in the absence or presence of 1 unit/ml ␣-thrombin. Cell lysates were prepared and incubated with recombinant GST-RBD for 1 h. Samples were subjected to SDS-PAGE, probed with anti-pan-Ras antibody, and the amounts of Ras quantified. The data shown in A are the mean Ϯ S.D. for triplicates in one experiment and are representative of three independent experiments. The Western blot is a representative blot. Data in B is representative of two independent experiments. G␣ q and G␣ i2 associate with different subsets of G␤␥s, and the distinct subsets activate distinct effectors. Several studies lend support to the notion that select G␤␥ subunits activate distinct effectors (36). Co-transfection experiments in COS-7 cells found that expression of only specific G␤ or G␥ subunits significantly activate PI-PLC␤ 2 (37). Whereas expression of G␤ 1 ␥1, G␤ 1 ␥5, or G␤ 2 ␥5 stimulates PI-PLC␤ 2 activity, expression of G␤ 2 ␥1 has no effect. Similarly, expression of G␤ 5 ␥2 stimulates PI-PLB␤ 2 but not ERK, whereas expression of G␤ 1 ␥2 activates both (38). Other studies (39) found that expression of only specific G␤ subunits activate GRK2. Of the G␤ subunits examined, G␤ 1 and G␤ 2 but not G␤ 3 activates GRK2. Recent data using immobilized G␤ 1 ␥ 2 to screen phage-displayed random peptide libraries (40) found a peptide, which contains a conserved sequence found in PI-PLC␤. Interestingly, expression of this peptide blocks activation of PI-PLC␤, but not G␤␥-dependent voltage-gated channels or G␤␥-mediated inhibition of type I adenylate cyclase suggesting that effectors bind to specific G␤␥ peptide sequences (40).
For different G-proteins to activate distinct effectors via release of G␤␥ suggests that the G␣s are associated with distinct subsets of G␤␥s. Whereas there is strong evidence for the activation of effectors by specific G␤␥ subunits, there is little data on the identity of the G␤␥ associated with specific G␣s. Consistent with this requirement, analysis of the G␤ subunits associated with G␣ i2 and G␣ q in IIC9 cells 3 find different subsets of G␤s associated with G␣ i2 and G␣ q .
A second explanation for our results is that activation of Ras and translocation of ␤-arrestin1 occurs in different membrane domains, for example, in raft and non-raft membrane fractions (41)(42)(43)(44). These membrane domains are enriched in proteins important in intracellular signaling including GPCRs, suggesting that these domains may play a role in signaling from GPCRs to their effectors. Huang et al. (45) have reported enrichment of G i , G s , G o , and G␤␥s in detergent-resistant membrane domains, lipid rafts. Furthermore, H-Ras has been shown to be localized to rafts, suggesting that Ras activation could occur in these domains (46).
At present our data cannot distinguish between these mechanisms. However, any mechanism must take into account our results showing that the activation of PI 3-kinase by ␣-thrombin (Fig. 8) is mediated via G q and G i2. G q plays a crucial role in mediating ␣-thrombin-induced PI 3-kinase activation through ␤-arrestin1 by stimulating ␤-arrestin1 association with PI 3-kinase, suggesting that ␤-arrestin1 functions as an adaptor molecule recruiting PI 3-kinase. PI 3-kinase activation also requires Ras activation, which is downstream of G i2 . Both effectors are blocked by G␤␥ sequestrants, indicating that they are G␤␥-regulated. Because Ras activation and ␤-arrestin1 translocation are crucial for ␣-thrombin-mediated PI 3-kinase activity and, therefore, likely occur in the same domain, we think that the association of different G␤␥s with specific G␣s is likely essential. However, it is also important to keep in mind that these mechanisms are not mutually exclusive. Thus, both could play a role. The possible importance of these mechanisms in G␤␥ signaling remains to be elucidated.