Phosphatidylinositol 3-Kinase Is an Early Intermediate in the Gβγ-mediated Mitogen-activated Protein Kinase Signaling Pathway

The βγ-subunit of Gi mediates mitogen-activated protein (MAP) kinase activation through a signaling pathway involving Shc tyrosine phosphorylation, subsequent formation of a multiprotein complex including Shc, Grb2, and Sos, and sequential activation of Ras, Raf, and MEK. The mechanism by which Gβγ mediates tyrosine phosphorylation of Shc, however, is unclear. This study assesses the role of phosphatidylinositol 3-kinase (PI-3K) in Gβγ-mediated MAP kinase activation. We show that Gi-coupled receptor- and Gβγ-stimulated MAP kinase activation is attenuated by the PI-3K inhibitors wortmannin and LY294002 or by overexpression of a dominant negative mutant of the p85 subunit of PI-3K. Wortmannin and LY294002 also inhibit Gi-coupled receptor-stimulated Ras activation. The PI-3K inhibitors do not affect MAP kinase activation stimulated by overexpression of Sos, a constitutively active mutant of Ras, or a constitutively active mutant of MEK. These results demonstrate that PI-3K activity is required in the Gβγ-mediated MAP kinase signaling pathway at a point upstream of Sos and Ras activation.

The cellular signaling pathways leading to receptor-tyrosine kinase-(RTK) 1 and G protein-coupled receptor-(GPCR) stim-ulated mitogen-activated protein (MAP) kinase activation have recently been the subject of intense investigation (1)(2)(3)(4)(5)(6). The signaling pathway of RTK-mediated MAP kinase activation is the most clearly understood. Epidermal growth factor (EGF) stimulation, for example, produces activation and autophosphorylation of the EGF receptor leading to the formation of a multiprotein complex containing the phosphorylated receptor, the phosphoprotein Shc, the adaptor protein Grb2, and the Ras-GTP exchange factor Sos (7)(8)(9). Sos catalyzes exchange of GTP for GDP on the small guanine nucleotide-binding protein, Ras, thereby stimulating Ras activation (10). Ras-GTP activates a kinase cascade involving Raf, MEK, and MAP kinase (11)(12)(13). Activated MAP kinases phosphorylate and activate transcription factors involved in cell growth and proliferation (1).
The signaling pathways utilized by G i -, G s -, G o -, and G qcoupled receptors to stimulate MAP kinase activation have also been assessed and compared (14 -18). In many cell types, G icoupled receptors mediate MAP kinase activation via the ␤␥dependent activation of Ras (14 -16). Several of the intermediate steps in the G␤␥-stimulated MAP kinase pathway are identical with the RTK-stimulated signaling cascade including Shc phosphorylation, Shc/Grb2 association, and Sos activation (19). Inhibitors of Src family tyrosine kinase activity abrogate G i -coupled receptor-and G␤␥-mediated Shc phosphorylation and MAP kinase activation in COS-7 cells (3,17,20,21) suggesting that a Src family tyrosine kinase may be involved in the G␤␥-mediated MAP kinase activation pathway at a point upstream of Ras activation. The mechanism by which G␤␥ subunits mediate activation of a tyrosine kinase resulting in increased Shc phosphorylation, however, is unclear.

EXPERIMENTAL PROCEDURES
Materials-COS-7 and CHO cells were from the American Type Culture Collection. Culture media and LipofectAMINE were from Life Technologies, Inc. Fetal bovine serum (FBS) and gentamicin were from Life Technologies Inc. LY294002 was from Bio-Mol. Monoclonal antibody 12CA5 was from Boehringer Mannheim, and anti-ERK2 polyclonal antibody was from Santa Cruz Biotechnology. Protein A-agarose was from Pharmacia Biotech Inc. [␥-32 P]ATP was from DuPont NEN. Myelin basic protein (MBP) and wortmannin were from Sigma. UK-14304 was from Pfizer. The cDNAs for the human ␣2-C10 AR and ␤ARK1 were cloned in our laboratory (31); cDNAs encoding G␤1 and G␥2 were provided by M. Simon; cDNA encoding hemagglutinin (HA)tagged p44 mapk (ERK1) was from J. Pouysségur; cDNA encoding ⌬p85 was from M. Sakaue. Constitutively active MEK was from R. Erickson.
Cell Culture and Transfection-COS-7 and CHO-K1 cells were maintained in Dulbecco's Modified Eagle's Medium (DMEM) and F-12 medium, respectively, supplemented with 10% FBS and 50 g/ml gentamicin. Cells were transiently transfected using LipofectAMINE as described previously (17). Assays were performed 48 h after transfection. In MAP kinase activation assays, transfected cells were serumstarved in DMEM containing 0.5% serum for 16 to 20 h prior to simulation.
Measurement of MAP Kinase Activity-Activity of epitope-tagged * This work was supported in part by National Institutes of Health Grant HL 16037 (to R. J. L.). 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. p44 HA-mapk (HA-ERK1), or endogenous ERK2, was determined following immunoprecipitation, using MBP as substrate (32) with modifications as described (33). Quantitation of labeled MBP was performed using a Molecular Dynamics PhosphorImager.

Measurement of Inositol Phosphate Production-Transfected cells were labeled overnight with [ 3 H]inositol
(1-2 Ci/ml) in DMEM containing 10% FBS, washed once with phosphate-buffered saline, and incubated in phosphate-buffered saline containing 1.0 mM CaCl 2 , 10 mM LiCl, and the indicated agonist for 45 min. Cells were lysed in 0.4 M perchloric acid (1 ml/well), and 0.8-ml aliquots were neutralized with 0.4 ml of 0.72 M KOH, 0.6 M KHCO 3 . Total inositol phosphate accumulation was quantitated using Dowex anion exchange chromatography as described previously (34).
Measurement of Ras Activation-Ras activation was determined as described previously (14,35). Ras-bound GDP and GTP were quantitated using a Molecular Dynamic PhosphorImager. Ras activation is expressed as the amount of GTP bound to Ras as a percent of total guanine nucleotide bound to Ras.

RESULTS AND DISCUSSION
The ␤␥ subunit of G i mediates Ras-dependent MAP kinase activation produced by stimulation of both the lysophosphatidic acid (LPA) receptor and ␣ 2A adrenergic receptor (␣ 2 AR) (14). In order to determine whether PI-3K activity is required in the G␤␥-mediated mitogenic signaling pathway, we assessed the effect of two chemical PI-3K inhibitors (wortmannin and LY294002) and the effect of overexpression of a dominant negative mutant of the p85 subunit of PI-3K (⌬p85) on G i -coupled receptor-and G␤␥-mediated MAP kinase activation. As shown in Fig. 1A, pretreatment of COS-7 cells with wortmannin or LY294002 markedly inhibits MAP kinase activation produced by stimulation of the endogenously expressed LPA receptor. This inhibition of the LPA signal was also detectable at the level of endogenous MAP kinase (ERK2) (Fig. 1A, right panel). Similar inhibition by wortmannin and LY294002 is observed on MAP kinase activation provoked by stimulation of transiently overexpressed ␣ 2 AR and direct transfection of G␤␥ (Fig.  1B). Wortmannin and Ly294002 pretreatment also inhibit G i -PCR-and G␤␥-mediated MAP kinase activation in CHO-K1 cells (data not shown). In contrast, wortmannin and LY294002 have a lesser effect on MAP kinase activation stimulated by phorbol 12-myristate 13-acetate (PMA), or epidermal growth factor (EGF) (Fig. 1B). The inhibition by wortmannin and LY294002 is limited to the MAP kinase signaling pathway, in that the PI-3K inhibitors do not affect ␣ 2 AR-or G␤␥-mediated phosphoinositide hydrolysis (Fig. 1C). Both wortmannin and LY294002 inhibit LPA and G␤␥-stimulated MAP kinase activation in a concentration-dependent manner (Fig. 1, D and E), with IC 50 values of 100 nM and 1.0 M for wortmannin and LY294002, respectively.
A recent study (36) reported that expression of a dominant negative mutant of p85␣ (⌬p85), which lacks the p110 binding site, inhibits insulin-stimulated PI-3K activity and PIP 3 production. The effect of ⌬p85 expression on G i -coupled receptorand G␤␥-stimulated MAP kinase activation is shown in Fig. 2. ative mutant of PI-3K produces effects similar to wortmannin and LY294002. These results strongly suggest that PI-3K activity is an essential component of the G␤␥-mediated MAP kinase signaling pathway.
In PC12 cells, GTP-dependent association of Ras with the catalytic subunit of PI-3K has been described (23), suggesting that activation of the p85/p110 PI-3K occurs subsequent to Ras activation. In contrast, a constitutively active mutant of PI-3K stimulates Ras-dependent Xenopus oocyte maturation and fos transcription (37) suggesting PI-3K activation precedes Ras activation in this system. To determine whether PI-3K activation in the G i -coupled receptor/G␤␥-mediated MAP kinase signaling pathway precedes or follows Ras activation, the effect of PI-3K inhibitors on G i -coupled receptor-mediated Ras activation was determined. As shown in Fig. 3, LPA-stimulated Ras activation is abolished by wortmannin and LY294002 pretreatment. In contrast, EGF-stimulated Ras activation is not significantly affected. The striking sensitivity of the LPA-stimulated signal to inhibition of PI-3K activity suggests a crucial role for PI-3K activity early in the G i -coupled receptor/G␤␥-mediated MAP kinase signaling pathway.
Further evidence that PI-3K activation is an early event in the G␤␥-mediated MAP kinase signaling pathway is provided by the results in Fig. 4. The effect of PI-3K inhibition on MAP kinase activation provoked by overexpression of Sos, constitutively active Ras (T24Ras), and constitutively active MEK (MEKϩ) in COS-7 cells was assessed. Pretreatment with wortmannin or LY294002 (Fig. 4A) or overexpression of ⌬p85 (data not shown) has no effect on the increase in MAP kinase activation stimulated by Sos, T24Ras, or MEKϩ (Fig. 4A), suggesting PI-3K activity is upstream of these intermediates in the G␤␥-mediated MAP kinase activation pathway.
In CHO-K1 cells, coexpression of G␤␥ with Sos results in a synergistic increase in MAP kinase activation (19). Expression of G␤␥ or Sos alone stimulates a 2-3-fold and 5-fold increase, respectively, in MAP kinase activation (Fig. 4B). Coexpression of G␤␥ and Sos results in a 15-20-fold increase in MAP kinase activation. As in COS-7 cells, G␤␥-mediated MAP kinase activation in CHO-K1 cells is abolished by wortmannin or LY294002 pretreatment, and Sos-stimulated MAP kinase activation is unaffected. The synergistic increase in MAP kinase activation produced by G␤␥ and Sos coexpression is not observed in cells pretreated with wortmannin or LY294002 (Fig.  4B). Thus, the ability of G␤␥ subunits to synergize with Sos is dependent on PI-3K activity, suggesting that PI-3K activity is required downstream of G␤␥, but upstream of Sos in the G imediated MAP kinase activation signaling pathway.
A role for PI-3K in mitogenic signaling has been suggested by previous studies showing that PI-3K can associate with activated RTKs and Src family kinases (38 -40). PI-3K association with the Grb2-Sos complex has been demonstrated following monocyte colony-stimulating factor stimulation of human peripheral blood monocytes (41) and increased PI-3K/Shc association in cells transformed by BCR/abl oncoprotein has been reported (42). Thus, PI-3K is capable of interacting with many mitogenic signaling intermediates.
Several studies have suggested a role for PI-3K in GPCRmediated signaling. Activation of neutrophils by formylated-Met-Leu-Phe involves pertussis toxin-sensitive increases in PI-3K activity and PIP 3 production (25)(26)(27)(28)(29)43). Increased PI-3K activity is observed in anti-phosphotyrosine immunoprecipitates following activation of G protein-mediated systems (27). Wortmannin attenuates platelet-activating factor-stimulated MAP kinase activation in guinea pig neutrophils (24) and pertussis toxin-sensitive somatostatin receptor-stimulated MAP kinase activation in CHO-K1 cells (22). The results of the present study demonstrate that PI-3K activity is required in the G␤␥-mediated MAP kinase signaling pathway and that the site of PI-3K activity in the pathway is upstream of Sos.
We have previously observed that G␤␥-mediated Shc phosphorylation is sensitive to tyrosine kinase inhibitors and wortmannin (20). Wortmannin-sensitive G␤␥-stimulated PI-3K activity has been described in platelets and neutrophils (44 -46), and a G␤␥-sensitive PI-3K (designated p110␥ or PI-3K␥) has been cloned (47). It is thus attractive to speculate that PI-3K␥ activity may be required for recruitment and activation of the tyrosine kinase(s) responsible for mediating Ras and MAP kinase activation.
It has recently been reported that a product of PI-3K activity, phosphatidylinositol 3,4,5-trisphosphate (PIP 3 ), is capable of binding with high affinity to the SH2 domains of proteins such as Src and the p85 subunit of PI-3K (48). Further, PIP 3 can compete with tyrosine-phosphorylated proteins for binding to these sites. The PIP 3 /SH2 interaction may suggest a novel mechanism for regulating signaling pathways in PI-3Kdependent systems. PIP 3 may block phosphoprotein binding to SH2 domain-containing proteins or even supplant phosphoproteins bound to an SH2 domain. It is therefore possible that PIP 3 may serve as an intermediate in the G␤␥-mediated MAP kinase signaling pathway.
The ability of ⌬p85 to inhibit G␤␥-mediated MAP kinase activation may indicate a requirement for the p85␣-p110 complex in the pathway, possibly as part of a complex containing Shc and a Src family tyrosine kinase (49). Alternatively, ⌬p85 may inhibit the G␤␥ signal by binding directly to PIP 3 . Binding of ⌬p85 to PIP 3 may disrupt PI-3K-dependent signaling by preventing PIP 3 from competing with phosphoproteins for binding to an SH2 domain. Therefore, it is possible that both p85-p110 PI-3K-and PI-3K␥-dependent signaling can be inhibited by expression of ⌬p85. Further investigation is required to determine which PI-3K isotype is utilized in the G␤␥-mediated MAP kinase signaling pathway, the mechanism by which PI-3K activity provokes tyrosine kinase activation resulting in Shc phosphorylation and the identity of the tyrosine kinase utilized in this pathway.