Differential Sensitivity of Phosphatidylinositol 3-Kinase p110γ to Isoforms of G Protein βγ Dimers*

The ability of G protein α and βγ subunits to activate the p110γ isoform of phosphatidylinositol 3-kinase (PtdIns 3-kinase) was examined using pure, recombinant G proteins and the p101/p110γ form of PtdIns 3-kinase reconstituted into synthetic lipid vesicles. GTP-activated Gs, Gi, Gq, or Go α subunits were unable to activate PtdIns 3-kinase. Dimers containing Gβ1–4 complexed with γ2-stimulated PtdIns 3-kinase activity about 26-fold with EC50 values ranging from 4 to 7 nm. Gβ5γ2 was not able to stimulate PtdIns 3-kinase despite producing a 10-fold activation of avian phospholipase Cβ. A series of dimers with β subunits containing point mutations in the amino acids that undergo a conformational change upon interaction of βγ with phosducin (β1H311Aγ2, β1R314Aγ2, and β1W332Aγ2) was tested, and only β1W332Aγ2 inhibited the ability of the dimer to stimulate PtdIns 3-kinase. Dimers containing the β1 subunit complexed with a panel of different Gγ subunits displayed variation in their ability to stimulate PtdIns 3-kinase. The β1γ2, β1γ10, β1γ12, and β1γ13 dimers all activated PtdIns 3-kinase about 26-fold with 4–25 nm EC50 values. The β1γ11 dimer, which contains the farnesyl isoprenoid group and is highly expressed in tissues containing the p101/p110γ form of PtdIns 3-kinase, was ineffective. The role of the prenyl group on the γ subunit in determining the activation of PtdIns 3-kinase was examined using γ subunits with altered CAAX boxes directing the addition of farnesyl to the γ2 subunit and geranylgeranyl to the γ1 and γ11 subunits. Replacement of the geranylgeranyl group of the γ2 subunit with farnesyl inhibited the activity of β1γ2 on PtdIns 3-kinase. Conversely, replacement of the farnesyl group on the γ1 and γ11 subunit with geranylgeranyl restored almost full activity. These findings suggest that all β subunits, with the exception of β5, interact equally well with PtdIns 3-kinase. In contrast, the composition of the γ subunit and its prenyl group markedly affects the ability of the βγ dimer to stimulate PtdIns 3-kinase.

The generation of phosphatidylinositol 3,4,5-trisphosphate (PIP3) 1 in the inner leaflet of the plasma membrane is critical to the regulation of cell function (1)(2)(3). The phosphorylated inositol head group provides a docking site for proteins containing pleckstrin homology domains (PH domains) and leads to activation of many enzymes including the phosphoinositol-dependent protein kinase and protein kinase B (Akt). Activation of protein kinase B regulates multiple cellular functions including differentiation, regulation of metabolic events, cell survival, and motility (1)(2)(3). In keeping with this central role, the level of PIP3 is tightly regulated, it can be elevated by multiple classes of receptors (1,2), and there are specific phosphatidylinositol 5-phosphatases (SHIP and PTEN) that degrade the signal (4 -8).
A large family of phosphatidylinositol 4,5-bisphosphate 3-kinases (PtdIns 3-kinases) is responsible for generating PIP3 by phosphorylating the D3 position in the inositol ring of phosphatidylinositol 4,5-bisphosphate. These enzymes are divided into 3 classes, Class 1A and 1B, Class II, and Class III, based on their structure and substrate specificity (2). Class 1A and 1B enzymes represent the major receptor-regulated kinases and are heterodimers composed of a regulatory subunit and a catalytic subunit. The three major isoforms of the Class 1A catalytic subunits have molecular masses of about 110 kDa and are termed ␣, ␤, and ␦. Although there are multiple spliced isoforms of the p85 regulatory subunit for these p110 isoforms, the full-length molecule has two SH2 domains and a SH3 domain that interact with tyrosine-phosphorylated proteins (1,2). Thus, the Class 1A enzymes are markedly stimulated by growth factors and other stimuli that activate receptor tyrosine kinases (1,2).
The Class 1B catalytic subunit also has a molecular mass of 110 kDa; its catalytic subunit is termed p110␥. The regulatory subunit for this isoform has a molecular mass of 101 kDa, has no recognizable protein-protein interaction domains, and dimerizes with the p110␥ catalytic moiety (2,9,10). The p110␥ form of the enzyme is highly expressed in cells of hematopoietic origin and is markedly activated by interaction with the G␤␥ dimer released after activation of receptors (9 -12). Thus, ligands working through G protein-coupled receptors regulate the p110␥ form of PtdIns 3-kinase and, hence, affect cell survival, reorganization of the cytoskeleton, cell shape changes, and cell migration, which are of central importance to the biology of hematopoietic cells (13)(14)(15)(16). Based on experiments performed with mice in which the p110␥ isoform of PtdIns 3-kinase has been genetically ablated, this effector has multiple roles in the function of platelets, neutrophils, macrophages, mast cells, and monocytes (17)(18)(19). For example, neutrophils from p110␥ Ϫ/Ϫ mice show no increase in PIP3 levels upon stimulation of the fMet-Leu-Phe receptor, no activation of protein kinase B, and clear defects in the respiratory burst, cell shape changes, and migration (17)(18)(19).
The G protein-signaling pathway is surprisingly complex with large families of molecules comprising the receptors, G proteins, and effectors (20 -22). Most cell types express multiple isoforms of each category, raising questions regarding the signaling specificity inherent in the multiple isoforms of these proteins (20 -22). The isoforms of the G protein ␣ and ␤␥ subunits contribute extensively to the specificity of this signaling system. Receptors couple selectively to certain ␣ subunits, and most effectors interact selectively with the different isoforms of the ␣ and ␤␥ subunits (20,21,23,24). Thus, an important issue in cell signaling is identification of the protein-protein interactions underlying responses to G protein-coupled receptors.
Given the primary role of the G protein ␤␥ subunit in the activation of the p101/p110␥ form of PtdIns 3-kinase (9,11,12,25), we investigated the ability of a complete spectrum of G protein ␣ subunits and ␤␥ dimers to activate the p110␥ isoform of PtdIns 3-kinase. Using pure, recombinant proteins reconstituted into synthetic lipid vesicles, we find that the p110␥ isoform responds selectively to the different ␤␥ dimers and does not respond to activated ␣ subunits. Moreover, the specific form of the ␥ subunit in the ␤␥ dimer and the prenyl group on its C terminus play major roles in the activation of PtdIns 3-kinase.

EXPERIMENTAL PROCEDURES
Materials-Reagents used for Sf9 cell culture and purification of ␣ and ␤␥ dimers have been described (26,27). GDP, imidazole, and HEPES were from Sigma. Chaps and GTP␥S were from Roche Applied Science. Genapol C-100 was from Calbiochem. Ni 2ϩ -NTA Superflow resin was from Qiagen. [ 3 H]PIP2 was from PerkinElmer Life Sciences.
[␥-33 P]ATP was from ICN. Source TM 15Q anion exchange resin was from Amersham Biosciences. Centricon 30 concentrators were from Millipore, and Ultra 100 concentrators were from Amicon. Phosphatidylethanolamine (PE), phosphatidylinositol, PIP2, and PIP3 were purchased from Avanti Polar Lipids; TLC plates were from Whatman (60 Å silica gel, 250 m), Restriction enzymes were from New England Biolabs. Synthetic oligonucleotides were from Qiagen. All other materials were of the highest available purity.
Construction of Recombinant Baculoviruses-The methods for constructing the G protein baculoviruses used in this study have been published as follows: G s ␣, G i ␣, G o ␣ (28), G q ␣ (29), ␤ 1-4 (30,31), ␤ 5 (32), ␥ 1 , ␥ 2 (30), ␥ 10 -12 (33,34). The baculovirus for ␥ 13 was the kind gift of Dr. David Siderovski, University of North Carolina (35). The baculoviruses encoding the "conformational change" mutants (H311A, R314A, W332A) and the "prenyl pocket" mutants (T329K, S331A, W339A, V315A, F335A, K337A) in the ␤ 1 subunit were prepared as described (36). The cDNAs encoding the porcine PtdIns 3-kinase p110␥ subunit and the p101 subunit were kindly donated by Dr. Leonard R. Stephens, Cambridge University, Cambridge, UK (9), and used to construct new recombinant baculoviruses. The cDNA for the p101 subunit was used to prepare a p101 baculovirus with a FLAG epitope on the N terminus using complimentary synthetic oligonucleotides to create the FLAG sequence. The new cDNA was subcloned into the XhoI and EcoRI sites in the Sf9 transfer vector pAcSG2, sequenced, and the recombinant virus was prepared via our standard techniques (26,27). The cDNA for p110␥ was modified to contain a hexahistidine tag on the N terminus using complimentary synthetic oligonucleotides, subcloned into the EcoRI and NotI sites in the Sf9 transfer vector pAcSG2, and sequenced, and the recombinant virus prepared as described (26,27).
Culture and Infection of Sf9 Cells-Spodoptera frugiperda cells (Sf9 insect cells) were cultured and maintained at 27°C as described (26,27). Cell culture, viral amplification, and baculovirus infections for protein production were performed in the presence of 1.0% antibiotic/ antimycotic (10,000 units/ml penicillin G, 10,000 mg/ml streptomycin, and 25 mg/ml amphotericin B). Sf9 insect cells were infected at a multiplicity of infection of 3 with recombinant baculoviruses expressing G protein ␣ and/or ␤␥ subunits or with viruses for the p101 and the p110␥ subunits of PtdIns 3-kinase prepared as described above. The incubations were allowed to proceed for ϳ60 -80 h or until cell viability fell to 80%, and the cells were harvested by centrifugation at 100 ϫ g. For the purification of G protein ␣ and ␤␥ subunits, the cell pellet was washed three times in insect cell phosphate-buffered saline (7.3 mM NaH 2 PO 4 , pH 6.2, 58 mM KCl, 47 mM NaCl, 5.0 mM CaCl 2 ) and resuspended in an ice-cold buffer composed of 25 mM HEPES, pH 7.5, 1 mM MgCl 2 , 120 mM NaCl, 1 mM 2-mercaptoethanol supplemented with a mixture of freshly prepared protease inhibitors (aprotinin, leupeptin, and pepstatin (at 2 mg/ml), benzamidine at 20 mg/ml, and Pefabloc SC Plus at 100 mg/ml) and frozen. Cell pellets were stored in this buffer at Ϫ80°C before G protein purification (see below). Pellets that were to be used for purification of PtdIns 3-kinase were processed immediately as freezing and thawing cell pellets released proteases active on the p101 subunit (see below).
Purification of G Protein ␤␥ Subunits-The ␤␥ dimers were purified using a G i1 ␣ affinity chromatography procedure that uses a His 6 -tagged ␣ subunit to select properly modified G protein ␤␥ subunits (37). Sf9 insect cells were co-infected with the His 6 -G i1 ␣ baculovirus along with the specific ␤ and ␥ baculoviruses of interest, and the cell membranes were prepared as described and extracted with a buffer containing 10 mM Tris, pH 8.0, 25 mM NaCl, 5 M GDP, the above protease inhibitors, and 0.1% Genapol (31). Extracts were passed over a Ni 2ϩ -NTA column and washed with Ni 2ϩ -NTA base buffer (20 mM Tris, pH 8.0, 150 mM NaCl, 10 m GDP, 1 mM ␤-mercaptoethanol) containing the above mixture of protease inhibitors, 10 mM imidazole, 0.5% Genapol, and 500 mM NaCl. A subsequent wash with base Ni 2ϩ -NTA buffer containing 0.1% Chaps was performed before elution of the ␤␥ dimers with Ni 2ϩ -NTA base buffer containing 50 mM MgCl 2 , 10 mM NaF, and 30 M AlCl 3 . Protein was concentrated using an Amicon Ultra 30 concentrator and exchanged twice using Ni 2ϩ -NTA base buffer containing 0.1% Chaps. The ␤ 5 ␥ 2 HF dimer was purified as described (29). Verification of the proper post-translational processing of the ␥ subunit in the dimer was accomplished using matrix-assisted laser desorption ionization mass spectrometry as described (38). In keeping with our previous experience (34,38), the ␥ subunits in the dimers used in this study were fully and properly modified with the geranylgeranyl (␥ 2 , ␥ 10 , ␥ 12 , ␥ 13 ) or farnesyl (␥ 11 , ␥ 1 ) isoprenoid groups. Moreover, the ␥ subunits engineered to have altered CAAX sequences (e.g. ␥ 2 -L71S) were shown to have the expected change in the prenyl group at the C terminus of the ␥ subunit (i.e. farnesyl on ␥ 2 -L71S) (34).
Purification of G Protein ␣ Subunits-The method for purification of G protein ␣ subunits is based on the affinity procedure used for the ␤␥ purification. The procedures used for purification of the ␣ subunits of G i , G q , G s , and G o were very similar; our methods for purification of G s and G q have been described in detail (29,31). Briefly, the desired G␣ subunit was overexpressed in Sf9 insect cells, with ␤ 1 and ␥ 2 subunits engineered to have hexahistidine and FLAG tags at their N termini (33). Sf9 cell membranes containing the G␣␤ 1HF ⅐␥ 2HF were prepared, the heterotrimer was extracted from the membranes with 0.2% Genapol, the extracts were applied to a Ni 2ϩ -NTA column and washed, and the ␣ subunits were purified by elution from the column with a buffer containing 50 mM MgCl 2 , 10 mM NaF, and 30 M AlCl 3 . These procedures are modifications of those described by Kozasa and Gilman (37) and Biddlecome et al. (39). The purified ␣ subunits were concentrated to a volume of 100 -200 l, separated into aliquots, and stored at Ϫ80°C. Each preparation of G protein ␣ subunit used in this study was functional based on its ability to couple to recombinant receptors and ␤␥ subunits when reconstituted into Sf9 cell membranes (29,31,36). SDS gels documenting the purity of the G␣ subunits prepared by this procedure are published elsewhere (29,31).
Purification of Phosphatidylinositol 3-Kinase-After infection with the recombinant baculoviruses for p101 and p110␥ as described above, Sf9 cells were harvested by centrifugation at 100 ϫ g. The cell pellet was not frozen; rather, it was washed three times in insect cell phosphate-buffered saline, then lysed by nitrogen cavitation at 0°C in a buffer containing 25 mM Hepes, pH 7.5, 120 mM NaCl, and the standard mixture of protease inhibitors. The extract was centrifuged at 1000 ϫ g for 10 min at 4°C to remove cell debris, then immediately centrifuged at 100,000 ϫ g for 45 min, and 100 ml of supernatant was loaded onto a 1.75-ml Ni 2ϩ -NTA column. The Ni 2ϩ -NTA column was washed with 17.5 ml (10 bed volumes) of a wash buffer containing 25 mM Hepes, pH 7.5, 120 mM NaCl with the above protease inhibitors and 1% (v/v) Triton X-100. To remove the Triton from the column it was washed with 17.5 ml of the wash buffer in which the Triton was replaced with 280 mM NaCl. Finally, the column was washed with 17.5 ml of the wash buffer containing 10 mM imidazole. PtdIns 3-kinase was eluted from the column with 10 ml of wash buffer containing 70 mM imidazole, 0.1% (v/v) Tween 20, and 1.0% (w/v) betaine. The protein eluted from the Ni 2ϩ -NTA column was immediately loaded onto a 3-ml anti-FLAG-agarose affinity gel column, and the column was washed with 30 ml (10 bed volumes) of 25 mM HEPES, pH 7.5, 3 mM MgCl 2 , 120 mM NaCl, 1 mM EGTA, 0.1% (v/v) Tween 20, 1.0% (v/v) betaine plus the above protease inhibitors. The column was washed with 30 ml of the above buffer containing 1% (v/v) Triton X-100 then 30 ml of the buffer with 280 mM NaCl. The column was eluted with 10 ml of the above buffer containing 0.5 mg/ml FLAG peptide, and 1-ml fractions were collected. To determine the yield and purity of each preparation, fractions collected from the FLAG column were resolved on an 8% SDS gel and stained with Coomassie Blue (Invitrogen, "Simply Blue"). Fractions containing Pt-dIns 3-kinase were pooled and concentrated to ϳ300 ng/l using an Amicon Ultra 100 concentrator. A typical preparation yielded about 100 g of protein. The pure protein was stored at Ϫ80°C in the elution buffer.
Assay of Phosphatidylinositol 4,5-Bisphosphate 3-Kinase-The activity of PtdIns 3-kinase was assayed using G protein ␣ subunits or ␤␥ dimers reconstituted into synthetic phospholipid vesicles. Phospholipid vesicles were prepared at a molar ratio of 4:1 PE to PIP 2 (1 mM PE, 250 M PIP 2 ) in a buffer containing 50 mM Hepes, pH 8.0, 80 mM KCl, 3 mM EGTA, and 1 mM dithiothreitol. The phospholipid mixture was dried under argon to evaporate the chloroform and methanol vehicle, resuspended in 0.5 ml of the above HEPES buffer, and vortexed gently. The suspension was then subjected to 10 freeze/thaw cycles to form large multilamellar vesicles. The large multilamellar vesicles were extruded through a polycarbonate filter (0.1-m-diameter pore size) to form a uniform preparation of large unilamellar vesicles (40). These vesicles improve the sensitivity and reproducibility of assays for enzymes such as PLC␤, which require lipid surfaces (40). The extruded large unilamellar vesicles were stored under argon at 4°C until use.
The activity of PtdIns 3-kinase was measured by the ability to produce PIP3 from PIP 2 using [␥-33 P]ATP to monitor the extent of the reaction. The purified G protein ␣ or ␤␥ subunits were diluted in 20 mM Hepes, pH 7.5, and reconstituted at final concentrations of 0.1-300 nM into the unilamellar vesicles on ice for 30 min in a buffer consisting of 20 mM Hepes, pH 7.5, 1 mM MgCl 2 , 150 mM NaCl, and 0.04% Chaps. The final Chaps concentration in the assay was kept below 0.01% because the activity of PtdIns 3-kinase is inhibited by concentrations of Chaps above 0.04%. The assay was started by the addition of 10 ng of PtdIns 3-kinase to a 50-l reaction mixture containing 20 mM Hepes, pH 7.5, 1 mM dithiothreitol, 50 M ATP, 3 mM MgCl 2 , 1 mg/ml bovine serum albumin, and 2.0 Ci of [␥-33 P]ATP and incubated for 15 min at 30°C. During the development of this assay it was found that bovine serum albumin concentrations ranging from 0 to 1.0 mg/ml did not affect the basal or G␤␥-stimulated activity of PtdIns 3-kinase. The rate of PIP3 production was linear for 30 min. The reaction was stopped by the addition of ice-cold 1 N HCl, and the lipids were extracted with CHCl 3 : MeOH (1:1 v/v). The samples were vortexed and centrifuged at 4°C at 1500 ϫ g for 2 min, the aqueous phase was aspirated off, and the organic phase was retained. The organic phase was washed with icecold 1 N HCl followed by a vortex/centrifugation cycle. Samples of the phospholipids retained in the organic phase were resolved by thin layer chromatography on a potassium oxalate-pretreated silica TLC plate (12). The mobile phase was 65 ml of 1-propanol, 4 ml of glacial acetic acid, and 31 ml of deionized water. After development for 2-3 h, the plate was dried and subjected to autoradiography with Amersham Biosciences Hyperfilm MP. The silica plates were scraped using the autoradiograph as a guide to identify the [ 33 P]PIP3, and the radioactivity was quantified by liquid scintillation counting.
Assay of Phospholipase C␤-The activity of PLC was measured via the ability to produce inositol trisphosphate from PIP 2 incorporated into synthetic lipid vesicles in the presence of pure G protein ␣ or ␤␥ subunits (34). The large unilamellar vesicles were prepared as described above (40), but [inositol-2-3 H]PIP 2 was included in the vesicles to provide a radioactive substrate for PLC␤. The G protein ␣ or ␤␥ subunits were reconstituted into the vesicles on ice for 30 min in 20 mM Hepes, pH 7.5, 1 mM MgCl 2 , 150 mM NaCl, and 0.04% Chaps. Before reconstitution, they were diluted in the reconstitution buffer so that their final concentration in the vesicles was 0.1-300 nM. As with the PtdIns 3-kinase assay, the final Chaps concentration was kept below 0.01% in the assay. The assay of PLC␤ activity was carried out using 5 ng of pure, recombinant turkey PLC␤ as described (34). The [ 3 H]inositol trisphosphate released into the aqueous solution was quantified by liquid scintillation spectrometry.
Activation of G Protein ␣ Subunits-Activation of the G␣ subunits before reconstitution into the phospholipid vesicles used in the PLC␤ and PtdIns 3-kinase assays was achieved by diluting the various ␣ subunits to a final concentration of 1000 nM in a buffer containing 50 mM Hepes, pH 8.0, 1 mM EGTA, 3 mM EDTA, 5 mM MgCl 2 , 100 mM NaCl, 1% sodium cholate, 1 mM dithiothreitol, and 1 mM GTP␥S. The ␣ subunits were incubated in this solution at 30°C for 30 min before the addition to the unilamellar phospholipid vesicles. The final concentration of the ␣ subunits reconstituted into the vesicles was 0.1-300 nM. The protocols for PtdIns 3-kinase and PLC␤ assays as described above were followed with the exception that a final concentration of 10 mM NaF and 20 M AlCl 3 was added to each incubation at the time the PtdIns 3-kinase or avian PLC␤ was added to the vesicles. Gel Electrophoresis and Immunoblotting-The identity and purity of protein samples were confirmed by gel electrophoresis followed by silver staining and/or by Western blot analysis. Preparations were resolved on 8 or 12% SDS gels, and the gels were stained with Coomassie Blue or silver or transferred to nitrocellulose membranes. Visualization of proteins on Western blots with specific antisera was performed using the ECL chemiluminescence system (Amersham Biosciences).
Calculations, Statistics, and Expression of Results-The experiments presented under "Results" are representative of four or more experiments. The EC 50 values presented in Table I were obtained by fitting the points to sigmoidal curves using the routines provided in the Graph Pad Prism software. The statistical significance between the fits of different data sets was determined on normalized data using the F statistic (41). Data were normalized because the V max of PtdIns 3-kinase varied. Differences in the maximal effect of ␤␥ dimers on PtdIns 3-kinase activity were determined by the paired t test.

RESULTS
Pure, recombinant G proteins, and PtdIns 3-kinase were reconstituted into unilamellar synthetic lipid vesicles to investigate the ability of various G protein ␣ subunits and ␤␥ isoforms to stimulate PtdIns 3-kinase activity. All proteins were expressed in Sf9 insect cells infected with recombinant baculoviruses and purified as described under "Experimental Procedures." The data in Fig. 1 demonstrate the purity of these proteins and the ability of the ␤␥ dimer to markedly activate PtdIns 3-kinase. Fig. 1A presents a silver-stained polyacrylamide gel resolving typical preparations of G␤ 1 ␥ 2 and G␤ 1 ␥ 11 , and Fig. 1B demonstrates the purity of the p101/p110␥ form of PtdIns 3-kinase. Fig. 1C presents a section of an autoradiograph demonstrating the resolution of the 33 P-labeled PIP3 from other components in the vesicle assay used to monitor PtdIns 3-kinase activity. Fig. 1D shows the low background activity in the assay and shows that ␤ 1 ␥ 2 can activate PtdIns 3-kinase more than 20-fold with an EC 50 of ϳ5 nM. This result agrees with the literature (9,12).
It has been recognized that the ␤␥ dimer markedly activates the p110␥ form of PtdIns 3-kinase (9,11,12); however, there are differing reports about the effect of G protein ␣ subunits on the activity of this enzyme. The original report describing the cloning of the p110␥ isoform reported that the G i ␣ and G t ␣ subunits purified from bovine brain and retina, respectively, caused a stimulation of PtdIns 3-kinase (42); however, subsequent work has suggested that G i ␣ subunits do not activate the enzyme (9 -11). To examine this issue with pure, recombinant proteins, we prepared four different ␣ subunits from the G s , Gi , and G q families by purifying them from a ␤␥ affinity column and examined their ability to activate the enzyme. The ␣ subunits were activated using a 30-min incubation with GTP␥S and reconstituted into synthetic lipid vesicles, and the activity of PtdIns 3-kinase activity was measured in the presence of AlF Ϫ . Fig. 2A compares the ability of pure, recombinant G q , G i1 , G o , and G s ␣ subunits to activate PtdIns 3-kinase with that of ␤ 1 ␥ 2 . None of the ␣ subunits tested activated PtdIns 3-kinase, whereas ␤ 1 ␥ 2 stimulated the enzyme more than 20-fold. To ensure that the protocols used to activate the ␣ subunits were effective, we examined the ability of an aliquot of the activated G q to stimulate its effector, PLC␤, in a lipid vesicle assay using avian PLC␤ (34). This assay was performed at the same time as the PtdIns 3-kinase assay. Fig. 2B indicates that the activated Gq ␣ subunit can stimulate avian PLC␤ activity about 10-fold and equally with the ␤ 1 ␥ 2 dimer. Activation of G q ␣ provides a stringent control, as the purified form of this ␣ subunit slowly exchanges GTP and is difficult to activate (43). As expected, the G s , G i , and G o ␣ subunits did not activate PLC␤ (44). Each of the preparations of G ␣ subunit used in these experiments was tested for its ability to couple to the appropriate recombinant receptors expressed in Sf9 cell membranes (31,45). All of the G␣ subunits were active proteins in this assay. The G s ␣ subunit was also able to stimulate type I and type II adenyl cyclase (31).
To explore the question of whether the many isoforms of the G␤␥ dimer differentially activate PtdIns 3-kinase, we prepared a panel of 10 recombinant ␤␥ dimers. The 7 known ␤ subunits fall into two subfamilies, one comprised of ␤ 1-4 and one con-taining ␤ 5 and ␤ 5L . The ␥ subunits separate into 5 subfamilies, one with ␥ 1,8,11 , one containing ␥ 2,3,4,9, one comprised of ␥ 7,12 , and another containing ␥ 5,10 (22,46). G␥ 13 is quite divergent and forms its own subfamily (22,46) but seems to have functional activities comparable with ␥ 2 (35). To minimize the complexity of exploring all 84 possible combinations of the existing ␤ and ␥ subunits, we chose subunits known to be effective activators of PtdIns 3-kinase such as ␤ 1 and ␥ 2 and selected ␤ or ␥ isoforms from each of the subfamilies to compare with these benchmarks. Thus, dimers were prepared containing the ␤ 1-5 subunits, each complexed with ␥ 2 as the partner, and a panel containing different ␥ subunits was prepared with the ␤1 subunit.
The effects of these 10 different ␤␥ dimers on PtdIns 3-kinase activity are presented beginning in Fig. 3. In agreement with previous experiments (47), Fig. 3 illustrates that dimers containing the ␤ 1-3 subunits complexed with the ␥ 2 subunit all activate PtdIns 3-kinase with EC 50 values ranging from 4 to 7 nM. Interestingly, the ␤ 4 ␥ 2 dimer activates PtdIns 3-kinase activity to a greater extent than the ␤ 3 ␥ 2 dimer; this difference has also been seen with type II adenyl cyclase (31). As expected (47), the ␤ 5 ␥ 2 HF dimer did not activate PtdIns 3-kinase at concentrations up to 100 nM. In experiments not shown, 300 -400 nM ␤ 5 ␥ 2 HF did not activate the enzyme. Importantly, this particular preparation of ␤ 5 ␥ 2 HF was an active molecule, as indicated by its ability to activate PLC␤ ϳ10-fold (not shown), an activation equivalent to that seen with ␤ 1 ␥ 2 in Fig. 2B. The EC 50 values of the five ␤␥ dimers for activation of PtdIns 3-kinase are presented in Table I. The values range from 4 to 7 nM and are not statistically different from each other.
To investigate the regions of the ␤ subunit that might be important for activating PtdIns 3-kinase, we made two types of mutations in the ␤ 1 subunit. These mutations have been termed the conformational change and prenyl pocket mutations (36). Based on the two x-ray structures of ␤ 1 ␥ 1 complexed with phosducin, three amino acids (His-311, Arg-314, and Trp-332) in the ␤ subunit undergo significant conformational changes when the dimer binds to phosducin (48,49). Two G␤ 1 subunits containing point mutations in which these amino acids were changed to alanine (H311A and W332A) bind phosducin and phosducin-like-protein less well than the native ␤ 1 ␥ 2

FIG. 1. Purification of PtdIns 3-kinase and G␤␥ from Sf9 cells.
A, silver-stained 12% SDS-polyacrylamide gel demonstrating the purity of two representative ␤␥ dimers, ␤ 1 ␥ 2 and ␤ 1 ␥ 11 . All dimers used in this study had similar purity. The proteins were purified as described under "Experimental Procedures." Molecular mass calibration in kDa (MW) is shown at the right. B, silver-stained 8% SDS-polyacrylamide gel demonstrating the purity of the p101/p110␥ form of PtdIns 3-kinase (PI3K). The enzyme was purified from the extracts of Sf9 cells as described under "Experimental Procedures." Molecular mass calibration in kDa (MW) is shown at the right. Both the p101 and p110␥ subunits migrate slightly higher than their indicated molecular masses because of the N-terminal epitope tags. C, section of autoradiograph made with Amersham Biosciences Hyperfilm MP from the potassium oxalate-pretreated silica TLC plate resolving [ 33 P]PIP3 from other components in the PtdIns 3-kinase assay. The enzyme was stimulated with increasing concentrations of the ␤ 1 ␥ 2 dimer. The migration position of [ 33 P]PIP3 is shown by the band, and the migration positions of PE, phosphatidylinositol (PI), and PIP2 standards are indicated by arrows. The TLC plate was loaded and developed as described under "Experimental Procedures." D, activation of PtdIns 3-kinase by G␤␥ reconstituted into synthetic lipid vesicles. Large unilamellar phospholipid vesicles were prepared at a molar ratio of 4:1 PE to PIP 2 (1 mM PE, 250 M PIP 2 ) as described under "Experimental Procedures." The ␤ 1 ␥ 2 dimer (open circles) was reconstituted into the vesicles at the indicated concentrations, and the activity of PtdIns 3-kinase was measured as the ability to produce PIP3 from PIP 2 using [ 33 P]ATP to monitor production of radioactive PIP3 . FIG. 2. Effect of G protein ␣ and ␤␥ subunits on PtdIns 3-kinase and PLC␤ activity. A, G protein ␣ subunits (multiple symbols) or the ␤ 1 ␥ 2 dimer (closed circles) were purified as described under "Experimental Procedures" and reconstituted into phospholipid vesicles. The G protein ␣ subunits were activated with GTP␥S before reconstitution and with AlF Ϫ during the assay. The activity of PtdIns 3-kinase was measured as described under "Experimental Procedures." B, the ability of activated G q ␣ (open squares) to stimulate avian PLC␤ was measured as the ability to produce inositol trisphosphate from PIP 2 incorporated into synthetic lipid vesicles in the presence of pure G proteins (␣ subunits or the ␤ 1 ␥ 2 dimer). The reaction mixture for each assay of PLC␤ activity used ϳ5 ng of PLC␤ and the indicated concentrations of ␣ or ␤ 1 ␥ 2 subunits (closed circles). The activity of PLC␤ was measured as described under "Experimental Procedures." or ␤ 1 ␥ 1 dimers (50). In addition, dimers with G␤ subunits containing mutations in any of these three residues are 25-100-fold less potent than native ␤ 1 ␥ 2 dimers at activating type II adenyl cyclase or PLC␤ (36). In contrast, dimers containing point mutations at any of these three residues are equally effective as the ␤ 1 ␥ 2 dimer in supporting coupling of the adenosine A1 receptor to the G i ␣ subunit (36). This result and the fact that the dimers are purified from a G␣ subunit column suggest that the ␤ subunits with these point mutations are properly folded.
Dimers containing ␤ 1 subunits harboring these three mutations were purified in a complex with the ␥ 2 subunit and examined for their ability to activate PtdIns 3-kinase. Fig. 4 illustrates that ␤ subunits containing the H311A and R314A mutations have similar activity as native ␤ 1 ␥ 2 and that ␤ subunits containing the Trp-332 mutation are less potent. The EC 50 value for the Trp-332 mutation is shifted to the right about 4-fold, to 12 nM (Table I). Interestingly, dimers with ␤ subunits harboring the Trp-332 mutation are uniformly less active in assays of multiple effectors including PLC␤, type II adenyl cyclase, K ϩ channels, Ca 2ϩ channels, and binding to phosducin (50 -53). However, even though this residue is also important for binding of the ␣ subunit to the ␤ subunit (54), the W332A mutation does not appear to inhibit the A1 adenosine receptor G i ␣-␤␥ interaction to a significant extent (36).
Representative ␤␥ dimers containing members of the 5 families of ␥ subunits had much larger differences in their activity on PtdIns 3-kinase. The data in Fig. 5 indicate that dimers containing the ␥ 2 , ␥ 12 , and ␥ 13 subunits were about equally effective in activating PtdIns 3-kinase with EC 50 values ranging from 4 to 7 nM. Dimers containing the ␥ 10 subunit were statistically less effective, and dimers containing ␥ 11 did not effectively activate the enzyme. The EC 50 values for all combinations tested are shown in Table I. Note that the weak activation of PtdIns 3-kinase caused by ␤ 1 ␥ 11 had an EC 50 value of about 12 nM, suggesting that it bound to PtdIns 3-kinase with reasonable affinity but could not cause the conformational change needed for increased activity.
In addition to differences in amino acid sequence, a major difference between ␥ 11 and the other ␥ subunits used in these experiments is that ␥ 11 contains a C15 farnesyl group on its C terminus (34). Multiple studies indicate that the composition of the prenyl group on the C terminus of the ␥ subunit is very important for the interaction of the ␤␥ dimer with receptors and effectors (34,(55)(56)(57)(58). For this reason we tested the role of the prenyl group in the activation of PtdIns 3-kinase by using ␥ subunits that contained altered CAAX boxes to direct the FIG. 3. Effect of G protein ␤␥ dimers containing different ␤ subunits on PtdIns 3-kinase activity. Increasing concentrations of ␤␥ dimers containing the ␤ 1 , ␤ 2 , ␤ 3 , ␤ 4 , or ␤ 5 subunits (all dimerized with the ␥ 2 subunit) were reconstituted into unilamellar synthetic lipid vesicles, and the activity of PtdIns 3-kinase measured as described under "Experimental Procedures." Ten ng of pure PtdIns 3-kinase was added to each assay. Open circles, ␤ 1 ␥ 2 ; closed squares, ␤ 2 ␥ 2 ; closed triangles, ␤ 3 ␥ 2 ; open diamonds, ␤ 4 ␥ 2 ; open inverted triangles, ␤ 5 ␥ 2 .

TABLE I EC 50 values for activation of PtdIns 3-kinase by isoforms of G␤␥
The various forms of recombinant G␤␥ were purified from baculovirus-infected Sf9 cells as described under "Experimental Procedures." PtdIns 3-kinase activity was measured by reconstituting pure G␤␥ isoforms into unilamellar, synthetic lipid vesicles containing PIP 2 and ␥-33 P-labeled ATP, adding 10 ng of pure PtdIns 3-kinase, and generating [ 33 P]PIP 3 as described under "Experimental Procedures." The EC 50 values were calculated with the Graph Pad Prism curve-fitting program as described under "Experimental Procedures." The EC 50 value for each different ␤␥ dimer was compared against that of ␤ 1 ␥ 2 . Statistical differences between the fitted curves were determined using the F-statistic. NA, not applicable.  4. Ability of three point mutations in the ␤ subunit to activate PtdIns 3-kinase. Three conformational change mutants were tested for their ability to activate PtdIns 3-kinase. These three amino acids in the G␤ 1 subunit (His-311, Arg-314,and Trp-332) undergo conformational changes when the ␤␥ dimer binds to phosducin. The ability of dimers containing these point mutations to activate pure PtdIns 3-kinase in unilamellar, synthetic lipid vesicles was measured as described under "Experimental Procedures." Open circles, ␤ 1 ␥ 2 ; closed triangles, ␤1H311A␥ 2 ; open squares, ␤1R314A␥ 2 ; closed inverted triangles, ␤1W332A␥ 2 .
addition of a C15 farnesyl group to ␥ 2 (␥ 2 -L71S) or a C20 geranylgeranyl group to ␥ 11 (␥ 11 -S74L) or ␥ 1 (␥ 1 -S74L). Thus, the amino acid sequence of the ␥ subunit is not altered. Dimers containing ␥ subunits with these mutations have been tested for their activity on PLC␤ and type II adenyl cyclase with the result that dimers containing the C15 farnesyl group are less active than those containing the geranylgeranyl group (34).
The data presented in Fig. 6A indicate that switching the prenyl group on ␥ 2 from geranylgeranyl to farnesyl shifts the activation of PtdIns 3-kinase about 1 log to the right (EC 50 from 5.6 to 35 nM, see Table I). As expected (47), Fig. 6B indicates that the ␤ 1 ␥ 1 dimer, like the similar ␤ 1 ␥ 11 dimer, poorly activates PtdIns 3-kinase. More importantly, switching the farnesyl group on ␥ 11 (or ␥ 1 ) to the geranylgeranyl group produces dimers that are nearly as active as the native ␤ 1 ␥ 2 . Note from Fig. 6B that ␤ 1 ␥ 11 -S74L and ␤ 1 ␥ 1 -S74L can activate PtdIns 3-kinase to about the same extent as ␤ 1 ␥ 2 . However, they are still less potent at activating the enzyme (EC 50 values of about 53 nM). The requirement of the geranylgeranyl group for proper activation of PtdIns 3-kinase is very similar to the near absolute requirement for the geranylgeranyl group in the stimulation of type II adenyl cyclase by ␤␥ dimers (34). Importantly, previous studies show that dimers containing ␥ subunits modified with either the farnesyl or geranylgeranyl groups intercalate into phospholipid vesicles equally well (40), suggesting that the type of prenyl group on the ␥ subunit is very important for the interaction with and activation of PtdIns 3-kinase.
One x-ray structure of the ␤ 1 ␥ 1 dimer bound to phosducin indicates that the prenyl group may fold into a pocket in the seventh blade on the surface of the ␤ 1 subunit (48). Mutations of the residues in this area designed to destabilize the binding of the prenyl group in its pocket (F335A, K337A, V315A, S331A, W339A, and T329K in G␤ 1 ) reduce the ability of dimers to activate type II adenyl cyclase and PLC␤ (36). Point mutations at Val-315, Thr-329, and Trp-339 in G␤ 1 also reduce the affinity of the interaction between ␤ 1 ␥ 1 or ␤ 1 ␥ 2 and phosducin (50). Accordingly, ␤ 1 ␥ 2 dimers containing mutations in this area of the ␤ subunit were tested for their ability to activate PtdIns 3-kinase. The data in Fig. 7 indicate that the effects of these prenyl pocket mutations fall into two groups. Dimers containing ␤ 1 T329K␥ 2 and ␤ 1 S331A␥ 2 and ␤ 1 W339A␥ 2 activate PtdIns 3-kinase with EC 50 values equal to that observed with the ␤ 1 ␥ 2 dimer. Dimers containing the F335A, K337A, and V315A point mutations shift the EC 50 for activation of PtdIns 3-kinase about 4 -5-fold to the right (see Table I). A slightly larger rightward shift in EC 50 was observed in the ability of dimers harboring these mutations to activate type II adenyl cyclase and PLC␤ (36). The observation that they cause a right shift in the ability of the dimer to activate PtdIns 3-kinase suggests that the prenyl group also plays an important role in the interaction of the ␥ subunit with this enzyme.

DISCUSSION
The family of heterotrimeric G proteins is composed of ␣, ␤, and ␥ subunits with many isoforms of each subunit expressed in most cells. The G␣ subunits are expressed from 17 different genes and, through the splice variants that exist in cells, form 22 different ␣ subunits. These subunits separate into four major families (21,22,46). The multiple ␣ subunits couple to the superfamily of seven transmembrane domain receptors, which initiate signaling from light, odorants, hormones, neurotransmitters, and autocoids (21,23). The ␤ and ␥ subunits form a dimer that is an integral and important part of the G protein-signaling mechanism. The ␤␥ dimer is required for the receptor to form a high affinity complex with agonists (59 -61), The ability of two representative prenyl pocket mutants to activate PtdIns 3-kinase was compared with that of ␤ 1 ␥ 2 (open circles) as described under "Experimental Procedures." The dimer containing ␤1S331A␥ 2 is representative of three dimers that activated PtdIns 3-kinase with EC 50 values equal to those of ␤ 1 ␥ 2 . The dimer containing ␤1V315A␥ 2 is representative of three dimers with rightward shifted EC 50 values (see Table I). The ability of all dimers to activate 10 ng of pure PtdIns 3-kinase in unilamellar, synthetic lipid vesicles was measured as described under "Experimental Procedures." for the receptor to efficiently catalyze the exchange of GDP for GTP on the ␣ subunit, and as the mediator of signaling to multiple downstream targets (20,21). The ␤ subunits are expressed from 5 genes (with 2 splice variants, ␤ 3S and ␤ 5L ), and the ␥ subunits are expressed from 12 genes. The ␤ subunits separate into 2 families, one containing ␤ [1][2][3][4] , which are about 90% identical, and one containing ␤ 5 and ␤ 5L , which are only 50% similar to ␤ 1 (22,46). The ␥ subunits can be divided into 5 subfamilies, one subfamily consists of ␥ subunits that are modified with farnesyl (C15) at their C terminus (␥ subunits 1,8, and 11) and three other subfamilies containing geranylgeranyl (C20) at their C terminus. The second subfamily is comprised of ␥ subunits 2,3,4,9, whereas the third and fourth subfamilies only consist of two members each, ␥ subunits 7 and 12, and ␥ subunits 5 and 10, respectively. G␥ 13 is the most divergent ␥ and forms its own subfamily (22,46) but has signaling properties similar to ␥ 2 (35).
The ␤␥ dimer has been demonstrated to regulate more than 20 effectors including ion channels, adenyl cyclase, phospholipase C, tyrosine kinases, and PtdIns 3-kinase (20,21). The large number of targets for the ␤␥ dimer, and the possibility that the many different isoforms of the dimer represent selective signals make it important to determine the signaling specificity inherent in these molecules. The generation of PIP3 in the membrane of hematopoietic cells such as neutrophils, mast cells, and macrophages is markedly increased by G i -linked receptors such as the fMet-Leu-Phe or C5a receptors. Activation of these receptors releases the ␤␥ dimer (17)(18)(19) and regulates the respiratory burst, cell shape changes via reorganization of the cytoskeleton, and motility (17)(18)(19). In vitro, the p101/p110␥ form of PtdIns 3-kinase can be activated 20 -60fold by the ␤␥ dimer (9,11,12,25). Thus, the primary goal of this study was to understand the activation of the p101/p110␥ form of PtdIns 3-kinase by the multiple different isoforms of the ␤␥ dimer. The experiments were performed with pure, recombinant proteins to ensure that direct effects of the ␤␥ dimer were measured.
The results demonstrate four major points. First, GTP-activated G s , G i , G q , or G o ␣ subunits were not able to stimulate the p101/p110␥ form of PtdIns 3-kinase. These experiments clarify earlier conflicting results on this point in the literature (9,10,42). As each ␣ subunit was purified from a column of immobilized ␤␥ dimers after activation with AlCl 3 and NaF, which induces a conformational change, the proteins are properly folded, active molecules. Moreover, each preparation of ␣ subunit coupled to the appropriate recombinant receptors (29,31,36); the G q ␣ preparation was able to markedly activate PLC␤ (Fig. 2B), and the preparations of G s ␣ have been demonstrated to activate adenyl cyclase (31). These experiments offer clear evidence that the p101/p110␥ form of PtdIns 3-kinase falls into the group of G protein effectors that are regulated exclusively by the ␤␥ dimers. Ion channels and the Rac guanine nucleotide exchanger, P-Rex1, are other examples (20,62). This finding is also consistent with observations that the ␤␥ dimer is a major regulatory molecule released by activation of G i -coupled receptors such as the fMet-Leu-Phe or C5a receptors in cells expressing the p101/p110␥ form of PtdIns 3-kinase such as neutrophils, platelets, mast cells, and macrophages (3, 17-19, 63, 64). Second, all dimers containing the G␤ 1-4 subunits appear to activate PtdIns 3-kinase equally. As expected (47), the ␤ 5 ␥ 2 HF dimer did not activate the enzyme. However, the observations that the ␤ 5 subunit is highly expressed in nerve tissue (65,66), which does not express high levels of the p110␥ isoform of PtdIns 3-kinase (42) and may dimerize with certain isoforms of RGS proteins rather than ␥ subunits (67), make it difficult to speculate about the meaning of this finding. Third, the isoform of the ␥ subunit in the dimer had a major effect on its ability to activate PtdIns 3-kinase. Clearly, dimers containing representatives of the ␥ 2 , ␥ 10 , ␥ 12 , and ␥ 13 families all activate PtdIns 3-kinase well. However, dimers containing ␥ 11 are not effective. Finally, the prenyl group on the ␥ subunit was found to be very important for the activation of PtdIns 3-kinase. G␥ subunits modified with the farnesyl group do not activate the enzyme well. Previous studies have shown that dimers containing ␥ subunits modified with the farnesyl group are ineffective in activating type II adenyl cyclase (34) and regulating the T-type Ca 2ϩ channel (68). This is an intriguing finding in that the ␥ 11 subunit is found in most tissues (69), is highly expressed in hematopoietic cells (69), and couples very well to G i -linked receptors in reconstitution assays (70).
The most interesting implications of these findings are those surrounding the ␥ 11 subunit. As noted above, dimers containing the ␥ 11 subunit couple very well to both G i (70)-and G qlinked receptors, 2 and the ␥ 11 isoform is abundant in hematopoietic cells (69). However, the present results demonstrate that it is not an effective activator of a major ␤␥ target in these cells, PtdIns 3-kinase. Moreover, the G i ␣ subunit is not able to activate the enzyme (Fig. 2B), yet G i -linked receptors such as the fMet-Leu-Phe and C5a receptors markedly activate the p101/p110␥ form of PtdIns 3-kinase in mast cells, neutrophils, and macrophages (3, 17-19, 63, 64). Thus, it seems highly unlikely that dimers containing the ␥ 11 subunit are released after activation of fMet-Leu-Phe or C5a receptors. Activation of these receptors must release a dimer such as ␤ 1 ␥ 2 , which is able to activate PtdIns 3-kinase 20 -40-fold. Interestingly, dimers containing ␥ 11 are not very effective at activating type II adenyl cyclase (34) or regulating ion channels (68). Because this ␥ subunit is widely expressed (69), these observations suggest that cells containing ␥ 11 could use dimers containing ␥ 11 to couple ␣ subunits to receptors, allowing receptors to activate ␣ subunits without inducing effects on downstream ␤␥ targets. Taken further, this argues that receptor activation may release specific ␤␥ dimers, a concept that gains experimental support from biochemical experiments (31,(71)(72)(73), antisense mRNA experiments (74,75), small interfering RNA experiments (76), and data obtained with knockout mice (77).
Interestingly, activation of receptors coupled to the G s ␣ subunit are able to inhibit the activation of neutrophils, mast cells, eosinophils, and platelets (78 -81). The inhibition is due in part to a reduction in the PIP3 signal (82,83). Thus, in hematopoietic cells, receptors coupled to G i provide stimulatory inputs, and receptors coupled to G s provide negative inputs into a major biological process. However, activation of G s must also release some isoform of the ␤␥ dimer, posing a potential question, Why would an inhibitory pathway release a signaling molecule (the ␤␥ dimer) that can markedly activate the very pathway to be inhibited? Thus, it seems logical that the ␤␥ dimer released by activation of G s -coupled receptors would not be able to stimulate PtdIns 3-kinase.
There are many possible reasons that ␤␥ dimers released from G s may not be able to stimulate PtdIns 3-kinase. The dimer might be compartmentalized (84), it could be released in too low a concentration to activate the effectors, or it might be part of a scaffolding complex that restricts its activity (85). Another interesting possibility is that the dimer which couples to G s ␣ in these cells may contain ␤ or ␥ subunits that do not interact well with the effectors. A significant amount of data suggests that distinct ␤␥ isoforms can selectively regulate effectors (22,86). For example, the ␤ 2 ␥ 2 dimer is able to inhibit the T-type Ca 2ϩ channel, whereas the ␤ 1 ␥ 2 dimer is ineffective (68), the ␤ 5 ␥ 2 dimer inhibits the GIRK 1,4 K ϩ channel, whereas dimers containing the other four ␤ subunits activate the channel (87), and ␤ 2 ␥ x dimers appear to be important for stimulating PLC␤ and producing Ca 2ϩ transients in mouse macrophages whereas dimers containing the ␤ 1 subunit are not effective (76).
Another body of literature indicates that specific ␤␥ dimers may be released after activation of specific receptors. Examples include the original antisense experiments performed in GH3 cells, suggesting that specific ␤␥ isoforms couple to muscarinic and somatostatin receptors (74,75,88), results from mice in which the ␥ 7 gene has been ablated, showing that loss of ␥ 7 causes loss of the Golf ␣ subunit, resulting in specific changes in cyclase activity in the striatum (77), reconstitution studies showing a preference of certain receptors for specific ␤␥ dimers (31, 70 -73), and results using small interfering RNA to knock down the ␤ 2 subunit in the J774.A1 macrophage line, showing loss of this subunit blunts the ability of the C5a receptor to stimulate a rise in intracellular Ca 2ϩ (76). Our own data show that dimers containing ␥ 11 support coupling of ␣ subunits to receptors very well (70), yet these dimers are very poor at regulating effectors such as PtdIns 3-kinase (Figs. 5 and 6), type II adenyl cyclase (34), PLC␤ (34), and T-type Ca 2ϩ channels (68). Taken together these results suggest the possibility that activation of specific receptors may release specific ␤␥ dimers. Based on our results, if the dimer released by receptor activation is ␤ 1 ␥ 11 , it is not likely to activate effectors. This hypothesis suggests that receptor activation may not always release a ␤␥ dimer that acts on downstream signaling pathways. This possibility would restrict receptor activation of effectors to those able to interact with that ␣ subunit coupled to the receptor and provides interesting avenues for future experiments in intact cells.