A novel role for phosphatidylinositol 3-kinase beta in signaling from G protein-coupled receptors to Akt.

The protein kinase Akt plays a central role in a number of key biological functions including protein synthesis, glucose homeostasis, and the regulation of cell survival or death. The mechanism by which tyrosine kinase growth factor receptors stimulate Akt has been recently defined. In contrast, the mechanism of activation of Akt by other cell surface receptors is much less understood. For G protein-coupled receptors (GPCRs), conflicting data suggest that these receptors stimulate Akt in a cell type-specific manner by a yet to be fully elucidated mechanism. Here, we took advantage of the availability of cells, where Akt activity could not be enhanced by agonists acting on this large family of cell surface receptors, such as NIH 3T3 cells, to investigate the pathway linking GPCRs to Akt. We present evidence that expression of phosphatidylinositol 3-kinase (PI3K) beta is necessary and sufficient to transmit signals from G proteins to Akt in these murine fibroblasts and that the activation of PI3Kbeta may represent the most likely mechanism whereby GPCRs stimulate Akt, as the vast majority of cells do not express PI3Kgamma, a known G protein-sensitive PI3K isoform. Furthermore, available evidence indicates that GPCRs activate Akt by a pathway distinct from that utilized by growth factor receptors, as it involves the tyrosine phosphorylation-independent activation of PI3Kbeta by G protein betagamma dimers.

The protein kinase Akt plays a central role in a number of key biological functions including protein synthesis, glucose homeostasis, and the regulation of cell survival or death. The mechanism by which tyrosine kinase growth factor receptors stimulate Akt has been recently defined. In contrast, the mechanism of activation of Akt by other cell surface receptors is much less understood. For G protein-coupled receptors (GPCRs), conflicting data suggest that these receptors stimulate Akt in a cell type-specific manner by a yet to be fully elucidated mechanism. Here, we took advantage of the availability of cells, where Akt activity could not be enhanced by agonists acting on this large family of cell surface receptors, such as NIH 3T3 cells, to investigate the pathway linking GPCRs to Akt. We present evidence that expression of phosphatidylinositol 3-kinase (PI3K) ␤ is necessary and sufficient to transmit signals from G proteins to Akt in these murine fibroblasts and that the activation of PI3K␤ may represent the most likely mechanism whereby GPCRs stimulate Akt, as the vast majority of cells do not express PI3K␥, a known G protein-sensitive PI3K isoform. Furthermore, available evidence indicates that GPCRs activate Akt by a pathway distinct from that utilized by growth factor receptors, as it involves the tyrosine phosphorylation-independent activation of PI3K␤ by G protein ␤␥ dimers.
Many important intracellular downstream targets for phosphatidylinositol 3-kinases (PI3Ks) 1 have been recently identified (1,2). Among them, the serine-threonine kinase, Aktprotein kinase B, has received special attention because of its central role in the regulation of cell survival or death in a large number of cellular systems (3). This kinase can be potently activated by growth factor receptors of the tyrosine kinase class, a process that involves the tyrosine phosphorylation-dependent activation PI3Ks (1,4,5). However, how the large family of cell surface receptors that transmit signals through heterotrimeric G proteins regulates the activity of Akt is still poorly understood. In this regard, seemingly conflicting data suggest that the ability of G protein-coupled receptors (GPCRs) to stimulate Akt is cell type-specific. For example, initial reports indicated that lysophosphatidic acid (LPA), which stimulates G i -coupled receptors, does not enhance the activity of Akt in NIH 3T3 cells (4,6), but recent studies have provided evidence that Akt can be stimulated by a variety of agonists acting on GPCRs in other cell types (7)(8)(9). The latter includes COS-7 cells, in which ectopically expressed G q and G i -coupled receptors, m1 and m2, muscarinic acetylcholine receptors (mAChRs), respectively, were shown to activate Akt effectively when stimulated by carbachol, a cholinergic agonist (9).
In the present study, we took advantage of the failure of GPCRs to stimulate Akt in NIH 3T3 cell lines to investigate the nature of the signaling molecules linking GPCRs to Akt. Here, we show that PI3K␤ is necessary and sufficient to stimulate Akt by the large family of GPCRs in the vast majority of cells, which do not normally express the G protein-regulated PI3K␥ isoform. Furthermore, we present evidence that GPCRs stimulate PI3K␤ by a pathway distinct from that utilized by tyrosine kinase growth factor receptors, as it does not require the phosphorylation of the p85 subunit in tyrosine residues or its recruitment to phosphotyrosine-containing complexes, but instead involves the activation of this PI3K isoform by the G ␤␥ subunits of heterotrimeric G proteins.
Akt and MAPK Assay-MAPK and Akt activity in cells transfected with plasmids for an epitope-tagged MAPK (HA-ERK2) or Akt (HA-Akt) were determined as described (9,14) using myelin basic protein (Sigma) or histone 2B (Roche Molecular Biochemicals) as substrates, respectively. Autoradiograms were scanned using Nikon Outlook software and subsequently quantified using NIH Image, Version 1.61 software.
* 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.
Fluorescence Microscopy-NIH 3T3 cells were grown in 24-well plates on coverslips and transfected with a plasmid containing the AH of Akt fused to EGFP (pCEFL-EGFP-AH-AKT), engineered as described by Watton and Downward (16), together with additional plasmids using LipofectAMINE PLUS™ following the manufacturer's instructions. 24 h later, cells were serum-starved for 12 h and subjected to stimulation by PDGF (30 ng/ml), carbachol (1 mM), or LPA (10 M), rapidly washed with 1ϫ phosphate-buffered saline, and fixed and permeabilized with cold methanol. Cells were then mounted and photographed under uv light in an Axioplan2 microscope (Zeiss).

RESULTS AND DISCUSSION
To begin exploring the molecular basis for GPCR activation of Akt in each cellular setting, we first re-examined the effect of LPA and carbachol on the enzymatic activity of Akt in COS-7 and NIH 3T3 cells expressing endogenous and transfected LPA and m1 receptors, respectively. As shown in Fig. 1A, these treatments stimulated Akt in COS-7 cells but not in NIH 3T3 cells, although these agonists enhanced the activity of MAPK in both cell types, which served as a control. Interestingly, serum stimulated Akt in both COS-7 and NIH 3T3 cells, thus suggesting that NIH 3T3 cells may be defective in one or more key molecular components involved in the transmission of signals from GPCRs to Akt.
As PI3Ks are central elements of intracellular signaling pathways linking cell surface receptors to Akt (4, 5, 17), we next examined which PI3K isoforms were expressed in each cell type. Two of the PI3K catalytic subunits, p110␣ and p110␤, are believed to be ubiquitously expressed (1,18). These PI3K subunits form heterodimers with non-catalytic subunits of the p85 family and are stimulated by receptor tyrosine kinases upon their recruitment to the plasma membrane by the assembly of phosphotyrosine-containing multimolecular complexes (1,18). Another PI3K catalytic subunit, p110␥, lacks the binding region for p85 proteins (19) but instead associates with a G ␤␥ -sensitive regulatory subunit, p101 (10,20,21), and thus can be regulated by heterotrimeric G proteins. This PI3K isoform exhibits a more restricted tissue distribution (22). To explore which PI3K isoforms are expressed in NIH 3T3 and COS-7 cells, cellular lysates were subjected to immunoprecipitation with antibodies specific for each p110 PI3K subunit, and their kinase activity was assessed using an in vitro lipid-kinase reaction. As shown in Fig. 1B, endogenous PI3K␣ activity was readily detectable both in NIH 3T3 and COS-7 cells. In contrast, PI3K␥ activity was either absent or at levels below its limit of detection in both cell types. However, this PI3K was recovered from PI3K␥-transfected NIH 3T3 cells (Fig. 1B, second panel) even if its expression levels were not high enough to be detectable by standard immunoblotting techniques (data not shown). Using this enzymatic assay as a sensitive approach to explore the expression of PI3Ks, we also did not observe any endogenous PI3K␥ activity in COS-7 cells. This suggested that the presence of this G protein-regulated PI3K isoform cannot explain the ability of GPCRs to stimulate Akt in COS-7 cells and raised the possibility that another PI3K catalytic subunit may transduce the G protein-initiated signals in this cell type. In this regard, very high levels of endogenous PI3K␤ activity were found in COS-7 cells (Fig. 1B). Interestingly, no PI3K␤  1 mM), or 10% calf serum (serum) for 15 min for Akt assays (Akt) or 5 min for MAPK assays (MAPK). Kinase activity (Kinase) was determined in anti-HA immunoprecipitates, and expression levels of epitopetagged kinases were determined by Western blot (WB) analysis. NIH 3T3 cells stably expressing m1 receptors were processed identically but without transfecting m1 receptors. The autoradiograms shown are from a representative experiment that was repeated four times with nearly identical results. Values in the bar graph correspond to the mean Ϯ S.E. of four independent experiments and are expressed as -fold increase with respect to their corresponding control. The migration of HA-Akt, HA-ERK, and their corresponding 32 P-labeled substrates is indicated. MBP, myelin basic protein. B, lysates (400 g of total cellular proteins) from NIH 3T3 cells, COS-7 cells, or NIH 3T3 cells transfected or not with expression constructs for p110␣ (NIH 3T3 ϩ PI3K␣), p110␤ (NIH 3T3 ϩ PI3K␤), or p110␥ and p101 (NIH 3T3 ϩ PI3K␥) (4 g per 10-cm dish) were immunoprecipitated (IP) using specific antibodies (Santa Cruz Biotechnology) for each PI3K isoform or a control serum as indicated, and PI3K activity was assayed. The chromatographic mobility of 32 P-labeled phosphatidylinositol-3-phosphate (PI-3P) is indicated. Autoradiograms shown are representative of three or more independent experiments. activity could be detected in NIH 3T3 cells, although it was readily demonstrable upon exogenous expression of p110␤.
These findings prompted us to explore whether the expression of PI3K␤ in NIH 3T3 might confer the ability to activate Akt in response to agonists acting on GPCRs. As shown in Fig.  2A, whereas overexpression of p110␣ did not change the pattern of Akt activation observed in control-transfected cells, as expected, the expression of a G protein-regulated isoform, PI3K␥, was sufficient to support the activation of Akt by the stimulation of m1 and LPA receptors ( Fig. 2A). Remarkably, expression of the p110␤ subunit of PI3K was also sufficient to enable the activation of Akt by both GPCR agonists. A detailed time course analysis of the Akt stimulation by carbachol and LPA in PI3K␤-transfected NIH 3T3 cells revealed that the expression levels of PI3K␤ achieved (Fig. 2B) were sufficient to support the rapid and potent activation of Akt by these GPCR agonists, which was demonstrable as soon as 5 min after ligand addition and remained above control levels for more than 1 h (Fig. 2C). Furthermore, expression of this PI3K isoform also enabled the subcellular redistribution of Akt in response to GPCR agonists in NIH 3T3 cells. As shown in Fig. 2D, an Akt-EGFP chimera containing the membrane-targeting domain of Akt (AH) fused to EGFP (16) displays a cytoplasmic localization in control cells, either untreated or when stimulated with carbachol or LPA, but readily translocates to the plasma membrane upon PDGF treatment (Fig. 2D, upper panel). In contrast, carbachol and LPA promoted the accumulation of EGFP-Akt at the level of the plasma membrane when cells coexpressed PI3K␤ (Fig. 2D, lower panel). Taken together, these results indicate that PI3K␤ can mediate the activation of Akt by GPCRs and that the expression of this PI3K isoform is necessary and sufficient to transmit signals from G proteins to Akt in these murine fibroblasts.
Because the p110␤ PI3K catalytic subunit forms complexes with regulatory subunits of the p85 family (23) and the latter participates in the activation of PI3Ks by tyrosine kinase receptors (24,25), we next investigated whether tyrosine phosphorylation of p85 or its association to phosphotyrosine-containing proteins is also involved in the activation of PI3K by GPCRs in NIH 3T3 cells. As an approach, we performed PI3K assays on anti-phosphotyrosine immunoprecipitates from p110␤-transfected NIH 3T3 cells upon treatment with serum, LPA, and carbachol. As shown in Fig. 3A, treatment with serum resulted in the recovery of PI3K activity in anti-phosphotyrosine immunoprecipitates. In contrast, LPA and carbachol treatments did not result in any demonstrable effect. Similarly, the amount of endogenous p85 protein recovered in anti-phosphotyrosine immunoprecipitates was only enhanced by serum addition (Fig. 3A). In agreement with these findings, pretreatment of p110␤-transfected cells with tyrosine kinase inhibitors such as genistein (100 M for 30 min) diminished the activation of Akt by serum and PDGF but not when elicited by LPA and carbachol (data not shown). In addition, a mutant p85 protein lacking the p110 interacting domain, ⌬p85, which acts as a dominant negative interfering molecule for p85-mediated pathways (13), caused a significant reduction in the activation of Akt by PDGF but did not affect the activation of Akt by LPA and carbachol (Fig. 3B). Taken together, these results strongly suggest that the activation of PI3K␤ by GPCR stimulation does not require tyrosine phosphorylation-dependent events.

FIG. 2. Expression of PI3K␤ in NIH 3T3 cells is sufficient to
stimulate Akt kinase activity and to promote its localization to the plasma membrane in response to GPCR agonists. A, NIH 3T3 cells expressing m1 receptors were transfected with HA-Akt and plasmids (DNA) for GFP or the catalytic subunits of PI3K␣, PI3K␤, or p101 and PI3K␥, serum-starved, and left untreated (c) or treated with LPA (10 M), carbachol (cch) (1 mM), or 10% calf serum (serum) for 15 min. Akt kinase assay (Kinase) and Western blots (WB) were performed as for Fig. 1. Autoradiograms are representative of three independent experiments. Arrows depict the migration of the phosphorylated H2B and HA-Akt. Values in the bar graph correspond to the mean Ϯ S.E. of three independent experiments and are expressed as -fold increase with respect to their corresponding control. B, COS-7 and NIH 3T3 cells transfected with expression plasmids for GFP (control) or p110␤ (PI3K␤) (4 g per 10-cm dish) were subjected to immunoprecipitation using an anti-PI3K␤-specific antibody (S-19; Santa Cruz Biotechnology) and processed for Western blot (WB) analysis with the same antibody. An arrow indicates the migration of the catalytic subunit of PI3K␤. A non-specific band of 105 kDa was observed in both control-and PI3K␤transfected cells, which served as a protein-loading control. C, NIH 3T3 cells expressing m1 receptors were transfected as in A with expression plasmids for HA-Akt and p110␤, serum-starved, and stimulated for the indicated times with LPA (10 M) or carbachol (cch) (1 mM), lysed, and subjected to Akt kinase assays (Kinase) and anti-HA Western blots (WB) as above. Arrows indicate the migration of labeled H2B and the epitope-tagged Akt. D, NIH 3T3 cells expressing m1 receptors were grown in 24-well plates on coverslips and transfected with a plasmid expressing the AH of Akt fused with EGFP (pCEFL-EGFP-AH-Akt, 250 ng per well) together with an empty vector (control) or an expression plasmid for the catalytic subunit of PI3K␤ (PI3K␤) (100 ng per well). As indicated, after serum starvation cells were stimulated with PDGF (10 ng/ml), carbachol (cch) (1 mM), or LPA (10 M) for 5 min, washed, fixed, mounted, and photographed under uv light using an Axioplan2 microscope (Zeiss) (ϫ 63). Depicted areas were magnified (ϫ 4) using Adobe Photoshop software. We next explored the nature of the G protein(s) implicated in the activation of Akt through PI3K␤, using the expression of GTPase-deficient forms of representative members of each G protein ␣ subunit family as an approach. As shown in Fig. 4A, coexpression of the epitope-tagged Akt together with constitutively active mutants of G␣ s , G␣ i2 , G␣ q , G␣ 12 , and G␣ 13 (12,26,27) did not result in the activation of Akt in wild-type NIH 3T3 or in PI3K␤-transfected cells, although Ras stimulated Akt potently in both cases. Thus, because of the failure of activated G␣ subunits to stimulate Akt, we asked whether ␤␥ dimers could induce the PI3K␤-dependent activation of Akt. As shown in Fig. 4A, overexpression in control NIH 3T3 cells of ␤ 1 ␥ 2 subunits either together or individually did not cause any significant change in the Akt activity. In contrast, overexpression of ␤ 1 ␥ 2 dimers induced a remarkable increase in the activity of Akt in PI3K␤-expressing cells. This activation was not observed when ␤ 1 or ␥ 2 were expressed alone or when HA-Akt was coexpressed with ␤ 1 and ␥ 2* , a mutant ␥ 2 that lacks the ␥-isoprenylation signal and therefore fails to associate with the plasma membrane (26,28). These results indicated that ␤␥ but not G␣ subunits promote the PI3K␤-dependent activation of Akt and that functional, membrane-bound ␤␥ dimers are required for this effect.
To investigate whether ␤␥ dimers participate in signaling from GPCRs to Akt, we expressed a chimeric molecule containing the extracellular and transmembrane domains of CD8 fused to the carboxyl-terminal domain of ␤ARK, which includes the high affinity ␤␥ binding region of this kinase and thus acts as a G ␤␥ scavenger. As shown in Fig. 4B, transfection of this chimeric molecule in NIH 3T3 cells expressing PI3K␤ dramatically reduced the activation of Akt in response to carbachol and LPA, whereas the stimulation of Akt by tyrosine kinase receptors, such as PDGF, was not affected. Taken together, these findings strongly suggest that ␤␥ subunits of heterotrimeric G proteins play a key role in signaling from GPCRs to Akt by acting on the p110␤ isoform of PI3K.
Accumulated evidence indicates that PI3K␥ can link GPCRs to a number of signaling pathways (1, 29 -31). However, recent studies have suggested that the expression of this PI3K is restricted to few cell types (1,22). Indeed, we did not observe detectable levels of PI3K␥ protein or its lipid kinase activity in NIH 3T3 and COS-7 cells. Instead, this study and recently available evidence provide support for a novel role for PI3K␤ in signaling by GPCRs in the vast majority of cells, which do not express PI3K␥. For example, it has been recently shown that this PI3K isozyme can be stimulated in vitro by purified G protein ␤␥ subunits in synergism with p-Tyr peptides (32,33) and that a synergistic effect of receptor tyrosine kinases and GPCRs could be observed in PI3K␤ activation (34,35). Furthermore, antibody-blocking experiments implicated PI3K␤ in the mitogenic pathway initiated by LPA (36). Thus, these results and the present findings that PI3K␤ expression is necessary and sufficient to activate PI3K-depedent pathways by GPCRs in cells lacking PI3K␥, such as fibroblasts, demonstrate a central role for the PI3K␤ isoform in signaling by G protein -FIG. 3. GPCRs do not promote the recruitment of PI3K␤ to phosphotyrosine-containing complexes. A, NIH 3T3 cells expressing m1 receptors were transfected with an expression plasmid for p110␤ as above, serum starved, and stimulated with 10% calf serum (serum), LPA (10 M), or carbachol (cch) (1 mM) for the indicated times. Cells were lysed and immunoprecipitated (IP) with anti-phosphotyrosine antibodies (anti-p-Tyr) and assayed for PI3K activity as in Fig. 1. The chromatographic mobility of the resulting 32 P-labeled PI-3P is indicated. The amount of PI3K p85 subunit recovered in anti-p-Tyr immunoprecipitates was assessed in anti-p85 (New England Biolabs) Western blots (WB). Autoradiograms correspond to representative experiments that were repeated two additional times. B, NIH 3T3 cells were transfected as described for Fig. 1 with HA-Akt, p110␤, and either pSR␣-⌬p85 (⌬p85, 0.5 g) (13) or the same amount of a GFP expression plasmid (GFP) as a control. After serum starvation, cells were stimulated with PDGF (10 ng/ml), LPA (10 M), or carbachol (cch) (1 mM) for 15 min, and Akt assays were performed in the corresponding anti-HA immunoprecipitates as described in Fig. 1. Values in the graph represent the mean Ϯ S.E. of three independent experiments and are expressed as a percentage of stimulation with respect to that observed in response to the same treatments in GFP-transfected cells, which was taken as 100%.

FIG. 4. G protein ␤␥ subunits mediate the activation of Akt by
GPCRs in PI3K p110␤-transfected NIH 3T3 cells. A, NIH 3T3 cells were transfected with expression plasmids for HA-Akt and GFP (NIH 3T3) or the catalytic subunit of PI3K␤ (NIH 3T3 ϩ PI3K␤) as in Fig. 1, together with expression vectors for GFP (c), the activated form of Ras (rasV12), the GTPase-deficient (QL) forms of the G protein ␣ subunits ␣ s , ␣ i , ␣ q , ␣ 12 , and ␣ 13 , or ␤ 1 , ␥ 2 , and ␥ 2* G protein subunits (0.5 g per dish in each case) as indicated. Akt kinase reactions were performed in anti-HA immunoprecipitates from the corresponding lysates. Values in the graphs represent the mean Ϯ S.E. of three independent experiments and are expressed as -fold induction with respect to controltransfected cells. B, Akt activity was examined in cells transfected with plasmids for p110␤ and HA-Akt together with a plasmid expressing the carboxyl terminus of ␤ARK fused to the CD8 receptor (CD8-␤ARK, 1 g per 6-cm dish) or with a vector expressing GFP (GFP) as a control. Data are expressed as a percentage of the response observed for each treatment in cells expressing ␤ARK relative to that of GFP-transfected cells, which was taken as 100%. Results correspond to the mean Ϯ S.E. of three independent experiments. linked receptors.
Interestingly, although there is extensive sequence homology between the p110␣ and p110␤ PI3K catalytic subunits (37), recent observations in knockout mice models suggest that these PI3K isoforms may perform non-complementary functions (1,38). In this regard, our current findings suggest that one such distinct feature is the ability of PI3K␤ to act downstream from heterotrimeric G proteins. The structural elements responsible for this distinct coupling specificity are still unknown and are under current investigation.
The emerging picture is that GPCRs stimulate PI3K␤ in most tissues by a mechanism distinct from those utilized by tyrosine kinase receptors, as it does not require the functional activity of the p85 non-catalytic subunit, but instead involves the activation of p110␤ by G ␤␥ subunits. We can also conclude that the distinct tissue distribution of each PI3K isoform may play an unexpected role in controlling the specificity in signal transmission, as it may govern the ability of GPCRs to stimulate a variety of intracellular signaling pathways that require PI3K function. In turn, the availability and level of expression of PI3K␤ and PI3K␥ may determine the nature of the biological responses elicited by the large family of G protein-linked cell surface receptors in each cell type.