Phosphorylation of the G Protein γ12 Subunit Regulates Effector Specificity*

Although the G protein βγ dimer is an important mediator in cell signaling, the mechanisms regulating its activity have not been widely investigated. The γ12subunit is a known substrate for protein kinase C, suggesting phosphorylation as a potential regulatory mechanism. Therefore, recombinant β1γ12 dimers were overexpressed using the baculovirus/Sf9 insect cell system, purified, and phosphorylated stoichiometrically with protein kinase C α. Their ability to support coupling of the Gi1 α subunit to the A1 adenosine receptor and to activate type II adenylyl cyclase or phospholipase C-β was examined. Phosphorylation of the β1γ12 dimer increased its potency in the receptor coupling assay from 6.4 to 1 nm, changed theK act for stimulation of type II adenylyl cyclase from 14 to 37 nm, and decreased its maximal efficacy by 50%. In contrast, phosphorylation of the dimer had no effect on its ability to activate phospholipase C-β. The native β1γ10 dimer, which has 4 similar amino acids in the phosphorylation site at the N terminus, was not phosphorylated by protein kinase C α. Creation of a phosphorylation site in the N terminus of the protein (Gly4 → Lys) resulted in a β1γ10G4K dimer which could be phosphorylated. The activities of this βγ dimer were similar to those of the phosphorylated β1γ12 dimer. Thus, phosphorylation of the β1γ12 dimer on the γ subunit with protein kinase C α regulates its activity in an effector-specific fashion. Because the γ12 subunit is widely expressed, phosphorylation may be an important mechanism for integration of the multiple signals generated by receptor activation.

pCNTR with BamHI and XbaI and ligated into these sites in the pVL1393 transfer vector. The N terminus of the ␥ 10 protein was modified to have a protein kinase C phosphorylation site by mutagenesis of the ␥ 10 cDNA using PCR. The primers used were: (sense primer: 5Ј-GGATCCATGTCCTCCAAGGCTAGC-3Ј; antisense primer: 5Ј-CACTT-TGTGCTTGAAGGAATTCC-3Ј). This modification introduced a Gly 4 3 Lys mutation (G4K) into the protein. The products of the PCR reaction were subcloned into pCNTR, digested with BamHI and EcoRI, and ligated into these sites in the pVL1393 transfer vector. To add the hexahistidine-FLAG affinity tags to the 5Ј end of the ␤ 1 subunit, the polymerase chain reaction was used to add XbaI and BamHI restriction sites to the 5Ј and 3Ј ends of the ␤ 1 coding region, respectively. The primers used were: (sense primer: 5Ј-TCTAGAATGAGTGAGCTTGAC-CAGTT-3Ј; antisense primer: 5Ј-GGATCCTTAGTTCCAGATCTTGA-GGA-3Ј). The products of the reaction were digested with XbaI and BamHI and ligated into the pDouble Trouble (pDT) vector, which adds the nucleotide sequences for the hexahistidine and FLAG affinity tags to the 5Ј end of the ␤ 1 coding region (17). The ␤ 1HF coding region was excised from pDT with HindIII and BamHI and subcloned into the pCNTR shuttle vector. The ␤ 1HF coding sequence was excised from pCNTR with BamHI and ligated into the BamHI site of pVL1393. The pVL1393 transfer vectors containing these four constructs were sequenced to verify the fidelity of the ␤ and ␥ sequences. Recombinant baculoviruses were constructed by co-transfecting each transfer vector with linearized BaculoGold® viral DNA into Sf9 cells using the PharMingen BaculoGold® kit as described (18). The recombinant baculoviruses were purified by one round of plaque purification using standard techniques (19). The construction of the recombinant baculoviruses coding for the G i1 and G s ␣ subunits and the A1 adenosine receptor have been described (20,21).
Expression and Purification of Recombinant G Protein ␣ and ␤␥ Subunits-G protein ␣ and ␤␥ subunits were overexpressed by infecting suspension cultures of Sf9 insect cells with recombinant baculoviruses (22,23). The G i1 ␣ subunit was purified to homogeneity as described (22). The recombinant ␤␥ dimers were extracted from Sf9 cells and purified using DEAE chromatography and an ␣ subunit affinity column (23).
Phosphorylation of Purified ␤␥ Subunits by PKC-The purified ␤ 1HF ␥ 12 subunit was incubated for 30 min at 30°C in 50 mM Tris, pH 7.5, 1 mM ␤-mercaptoethanol, 10 mM MgCl 2 , 0.4 mM CaCl 2 , 40 -100 M ATP, recombinant PKC ␣ or ␤, 40 g/ml phosphatidylserine, and 0.8 g/ml diolein. Usually, 25 g of ␤␥ dimer was incubated with 0.74 unit of PKC ␣ to achieve stoichiometric phosphorylation of the ␥ subunit. Control reactions contained deionized water in the place of PKC. The stoichiometry of the phosphorylation reaction was measured by including 40 M [ 32 P]ATP (500 -2000 cpm/pmol) in the reaction mix and subjecting the phosphorylated ␤␥ subunits to Tricine/SDS-PAGE (24). The resolved ␥ subunit was cut from the dried gel and the amount of radioactive phosphate incorporated estimated by scintillation counting. After the phosphorylation reaction and before use in the assays, the ␤␥ subunit was repurified from the reaction mixture by loading it onto a 0.25 ml Ni 2ϩ -NTA-agarose column (Qiagen) and washing with 15 ml of 20 mM Hepes, pH 8.0, 150 mM NaCl, 1 mM MgCl 2 , 1 mM ␤-mercaptoethanol, 0.6% (w/v) CHAPS, and 5 mM imidazole to remove PKC. The ␤␥ dimer was eluted with 1 ml of the wash buffer containing 200 mM imidazole. To ensure that no PKC activity was carried into the assays with the ␤␥ dimer, its activity was monitored in the elution fractions using the kinase reaction buffer described above with 25 g/ml histone 3S as substrate (25). As expected, the kinase activity eluted in the void volume of the column and not with the ␤␥ dimer. Controls were also performed to determine whether the 30°C incubation with PKC and subsequent re-purification reduced the activity of the ␤␥ dimer in the functional assays. In these experiments, a mock-phosphorylation reaction was performed in the absence of PKC, the ␤␥ dimer re-purified and its activity compared with that of dimers subjected only to the ␣-agarose column (see "Results").
Measurement of the High Affinity Ligand Binding Conformation of the A1 Adenosine Receptor-Sf9 insect cell membranes overexpressing recombinant A1 adenosine receptors were prepared as described (21) and reconstituted with G protein ␣ and ␤␥ subunits on ice for 30 min (26). The high affinity, agonist binding conformation of the receptor was measured using the agonist ligand 125 I-N 6 -(aminobenzyl)adenosine as described (26). Each reaction tube contained 20 fmol of receptor, 6 nM G i1 ␣ subunit, 0 -10 nM ␤␥ dimer, 50 nM GDP, and 0.3 nM 125 I-N 6 -(aminobenzyl)adenosine. Because this assay was incubated for 3 h before filtration, 100 nM microcystin was included to inhibit protein phosphatases in the Sf9 cell membrane preparation (27).
Measurement of Phospholipase C-␤ Activity-Large unilamellar phospholipid vesicles were prepared by extrusion into a buffer containing 50 mM Hepes, pH 8.0, 3 mM EGTA, 80 mM KCl, and 1 mM dithiothreitol with a Avanti Polar Lipids mini-extruder (28 (29). The reaction was begun by addition of 10 ng of recombinant, turkey erythrocyte PLC-␤ and 3 M free Ca 2ϩ to each assay tube. The mixture was incubated for 15 min at 30°C and stopped by the addition of ice-cold 10% trichloroacetic acid followed by the addition of 10 mg/ml bovine serum albumin. Assay tubes were centrifuged at 4,000 ϫ g and the [ 3 H]inositol 1,4,5-trisphosphate released measured by liquid scintillation counting (30). Measurement of Adenylyl Cyclase Activity-Sf9 insect cell membranes overexpressing recombinant, rat type II adenylyl cyclase (31) were prepared as described (18). The G s ␣ subunit was extracted from an Sf9 cell preparation with 0.1% (w/v) CHAPS as described (18). Cyclase containing membranes (5 g of protein/assay tube) were reconstituted with GTP␥S-activated G s ␣ subunit (32) and varying concentrations of ␤␥ dimer on ice for 30 min. The reaction buffer (25 mM Hepes, pH 8.0, 10 mM phosphocreatine, 10 units/ml creatine phosphokinase, 0.4 mM 3-isobutyl-1-methylxanthine, 10 mM MgSO 4 , 0.5 mM ATP, and 0.1 mg/ml bovine serum albumin) was preincubated at 30°C for 20 min. Production of cyclic AMP was initiated by addition of the reconstituted membranes to the reaction buffer and the incubation continued for 10 min at 30°C. Reactions were stopped by the addition of 0.1 N HCl and cyclic AMP measured using an automated radioimmunoassay (33).
Electrophoresis-Tricine/SDS-polyacrylamide gels were run according to the procedure of Schagger and von Jagow (24). The separating gel contained 16.5% total acrylamide, 0.4% bisacrylamide, and 10% (v/v) glycerol. The stacking gel contained 4% total acrylamide and 0.1% bisacrylamide. Gels were run at constant voltage (ϳ100 volts) at 10°C for 4 -5 h. Resolved proteins were stained with silver by the method of Morrissey (34), with the modification that the dithiothreitol incubation was reduced to 15 min.
Calculations and Expression of Results-Experiments presented under "Results" are representative of three or more similar experiments. Data expressed as dose-response curves were fit to rectangular hyperbolas using the fitting routines in the GraphPad Prizm® software. Statistical differences between the curves were determined using all the individual data points from multiple experiments to calculate the F statistic as described (35).
Materials-All reagents used in the culture of Sf9 cells and for the expression and purification of G protein ␣ and ␤␥ subunits have been described in detail (20,23). The baculovirus transfer vector, pVL1393, was purchased from Invitrogen; the BaculoGold® kit from PharMingen; 10% GENAPOL® C-100, CHAPS, microcystin, and the ␣ and ␤ 1 isoforms of PKC from Calbiochem; Ni 2ϩ -NTA-agarose from Qiagen; [ 3 H]phosphatidylinositol bisphosphate from NEN Life Science Products; PMA from Sigma; the pCNTR shuttle vector from 5 Prime 3 3 Prime, Inc. (Boulder, CO). All other reagents were of the highest purity available.

Stoichiometry of Phosphorylation of the ␥ 12 and the ␥ 10G4K
Subunits-Recent experiments have demonstrated that the bovine ␥ 12 subunit is a substrate for protein kinase C (12), but the functional significance of this phosphorylation event has not been extensively studied. As this newly discovered ␥ 12 subunit is widely expressed (12)(13)(14)(15), its phosphorylation may have important consequences. To examine the effects of ␥ 12 subunit phosphorylation on its activity, we purified recombinant ␤ 1 ␥ 12 dimers from Sf9 insect cells, phosphorylated them with PKC ␣ and ␤, and tested their activity in assays of ␤␥ function. The human ␥ 12 subunit was rapidly phosphorylated by PKC ␣ to a stoichiometry of about 1 mol/mol as shown in Fig. 1, A and B. Protein kinase C ␤ 1 was less effective than PKC ␣; the stoichiometry only reached 0.5 mol/mol after 60 min of incubation. Addition of a hexahistidine-FLAG affinity tag to the ␤ 1 subunit (␤ 1HF ) facilitated removal of PKC from the reaction mixture prior to assessing ␤␥ function. Thus, the phosphorylated ␤ 1HF ␥ 12 dimer in the reaction mix can be applied to a Ni 2ϩ -NTA-agarose column and pure ␤␥ dimer eluted with imidazole. The left side of Fig. 1C shows that stained PKC protein was removed by this procedure and assay of kinase activity in the column elution fractions determined that the ␤␥ dimer was free of residual kinase activity (see "Experimental Procedures"). The right side of the figure shows that only the ␥ 12 subunit in the dimer is phosphorylated. Addition of the hexahistidine-FLAG affinity tag to the N terminus of the ␤ 1 subunit did not affect the association of the ␤ and ␥ subunits or purification of the dimer on an ␣-subunit affinity column (data not shown). Ser 1 in the N terminus of the bovine ␥ 12 subunit is the site phosphorylated by protein kinase C (12,13). Thus, the consensus phosphorylation sequence S*XK is the motif in the ␥ 12 subunit recognized by the kinase (36). None of the other known ␥ subunits have this motif, but the ␥ 10 subunit has a similar pair of serine residues at its N terminus (37). To further evaluate the effect of phosphorylation on the activity of ␥ subunits, we mutated residue 4 in the ␥ 10 subunit from glycine to lysine (Gly 4 3 Lys) to introduce the SSK motif ( Fig. 2A). The ability of the wild type and the mutated ␥ 10 subunit to be phosphorylated by protein kinase C ␣ was examined. As expected, the wild type ␤ 1HF ␥ 10 dimer was not phosphorylated, but the dimer containing the ␥ 10G4K subunit was rapidly phosphorylated by PKC ␤ (autoradiograph in Fig. 2B) with a time course similar to that seen with ␥ 12 (data not shown). The silver-stained gel in Fig. 2C compares the mobilities of the wild type and mutated ␥ 10G4K subunits with those of the ␥ 1 and ␥ 12 subunits. Clearly, the SSK motif created in the ␥ 10G4K subunit can be phosphorylated effectively. The stoichiometry of phosphorylation for the ␥ 10G4K protein was about 0.5 mol/mol using PKC ␣ and about 0.25 mol/mol using PKC ␤ 1 (data not shown). Thus, the differences in phosphorylation rates observed using ␥ 12 as a substrate were also seen using the ␥ 10G4K subunit.
Effect of Phosphorylation of the ␥ Subunit on Receptor Coupling-Having established the ability of protein kinase C to phosphorylate the two recombinant ␥ subunits, we examined the effects of phosphorylation on the function of the dimer. The ␤␥ subunits play several important roles in the signaling mechanisms used by receptors to activate effectors. In combination with the ␣ subunit, they participate in forming the high affinity agonist binding conformation of the receptor (21,38,39); they stabilize the basal state of the system by increasing the affinity of the ␣ subunit for GDP (7,8); and when released from the ␣ subunit, they activate effectors such as type II adenylyl cyclase and phospholipase C-␤ (6, 7). We first examined the effect of ␥ subunit phosphorylation on the ability of the ␤␥ dimer to support establishment of the high affinity, agonist binding conformation of a G protein-coupled receptor. In membranes from Sf9 cells overexpressing recombinant A1 adenosine receptors, about 90% of the receptors are in a low affinity conformation. Reconstitution of pure G i ␣ and ␤␥ subunits into these membranes establishes high affinity agonist binding and provides a sensitive assay for receptor-␣␤␥ interactions (21). Fig. 3 shows that the dimers used in this study, ␤ 1 ␥ 12 , ␤ 1 ␥ 10 , ␤ 1 ␥ 10G4K , and ␤ 1HF ␥ 12 , were able to re-establish the high affinity, agonist binding conformation of the receptor with a potency and efficacy equal to the well studied and highly effective ␤ 1 ␥ 2 dimer (26). All dimers tested support coupling with a K act of 0.5-1.0 nM (see Fig. 3 legend). Thus, the newly discovered ␥ 10 and ␥ 12 subunits are able to couple very effectively to the G i1 ␣ subunit and the A1 adenosine receptor when dimerized with the ␤ 1 subunit. Importantly, neither the hexahistidine-FLAG tag added to the N terminus of the ␤ 1 subunit nor the G4K mutation made in the N terminus of the ␥ 10 subunit affect the ability of these recombinant ␤␥ dimers to induce the high affinity conformation of the A1 adenosine receptor.
The data in Fig. 4 illustrate the ability of phosphorylated and unphosphorylated forms of the ␤ 1HF ␥ 12 dimer to support reconstitution of the high affinity binding state of the A1 adenosine receptor. Phosphorylation of the ␤ 1HF ␥ 12 subunit (open circles) increased the potency of the dimer in this assay from about 6.4 to 1 nM (Fig. 4A). This difference was significant (p Ͻ 0.001). Interestingly, phosphorylation of the ␥ 12 subunit has been reported to increase the affinity of the ␤␥ dimer for the ␣ subunit (12), a result consistent with the increased potency seen in Fig.  4A. A similar result was obtained when the phosphorylated and unphosphorylated forms of the ␤ 1HF ␥ 10G4K dimer were tested (Fig. 4B). This small difference in potency was also significant (p Ͻ 0.05). The observation that the phosphorylated ␤ 1HF ␥ 10G4K dimer does not shift the curve as greatly as the phosphorylated ␤ 1HF ␥ 12 dimer may be due to the fact that the stoichiometry of phosphorylation is only about 0.5 mol/mol (see Fig. 2 and text). It is important to note that these differences were observed only when microcystin was included in the binding assay buffer, suggesting that the Sf9 cell membranes contain a protein phosphatase able to dephosphorylate ␥ 12 . To examine this possibility, we incubated 32 PO 4 -labeled ␤ 1HF ␥ 12 with the Sf9 cell membranes at 30°C in the presence or absence of 100 nM microcystin, removed aliquots from the incubation medium over a 60-min time period, and resolved the proteins on a Tricine/SDS gel. In the absence of microcystin, the amount of radioactivity in the ␥ 12 subunit decreased more than 80% over the 60-min incubation. Little dephosphorylation occurred if microcystin was included in the 30°C incubation or if the mixture was held at 0°C without microcystin (data not shown). This result confirms the existence of an effective phosphatase for the ␥ 12 subunit. The insect cell phosphatase must be analogous to mammalian protein phosphatases 1 and 2, which are very sensitive to microcystin (27). Finally, controls were performed to determine whether the steps used to phosphorylate and re-purify the ␤␥ dimers decreased their ability to couple to the adenosine receptor (see "Experimental Procedures"). The activity of a pure ␤ 1HF ␥ 12 dimer used directly in the binding assay was about 10% higher than dimers subjected to a mockphosphorylation incubation and re-purified on the Ni 2ϩ -NTA column prior to assay (compare maximal binding in Figs. 3 and 4). Fig. 5A show the ability of the native and modified ␤␥ dimers used in this study to activate PLC-␤ in an in vitro assay using [ 3 H]PIP 2 incorporated into phospholipid vesicles as substrate. Note that ␤ 1 ␥ 12 , ␤ 1 ␥ 10 , ␤ 1HF ␥ 10G4K , and ␤ 1HF ␥ 12 are equally as effective as the ␤ 1 ␥ 2 dimer. All four forms of the protein stimulated the release of [ 3 H]inositol 1,4,5-trisphosphate with a K act of 6 -8 nM and were equally effective (ϳ8-fold increase in activity). The data in Fig. 5B demonstrate that there is no difference in the ability of either the phosphorylated or unphosphorylated ␤ 1HF ␥ 12 dimers to activate PLC-␤. The phosphorylated or unphosphorylated forms of the ␤ 1HF ␥ 10G4K dimers were also tested and no differences were observed (data not shown). As can be seen by comparison of the maximal activities shown in Fig. 5, A and B, the phosphorylation protocol caused about a 40% decrement in ␤␥ activity in the PLC assay.

Effect of ␥ Subunit Phosphorylation on Effector Activity-The data in
The ␤␥ dimer causes a synergistic activation of type II adenylyl cyclase in the presence of GTP␥S-activated G s ␣ subunit (40). Therefore, we examined the effect of phosphorylation on the ability of the ␤ 1HF ␥ 12 and ␤ 1HF ␥ 10G4K dimers to stimulate recombinant, type II adenylyl cyclase in Sf9 cell membranes. Fig. 6A shows that the unphosphorylated forms of all the ␤␥ dimers used in this study effectively stimulate type II adenylyl cyclase. As before, the ␤ 1 ␥ 12 , ␤ 1HF ␥ 12 , and the ␤ 1HF ␥ 10G4K dimers are all as effective as the ␤ 1 ␥ 2 dimer. The ␤ 1 ␥ 10 dimer had activity equal to the ␤ 1HF ␥ 10G4K dimer in this assay (data not shown). The K act for all the dimers ranges from 3 to 14 nM. Each of these dimers was purified with the ␣ subunit affinity column and assayed directly. Thus, the newly discovered ␥ 12 and ␥ 10 subunits are able to effectively stimulate type II adenylyl cyclase when combined with the ␤ 1 subunit. Moreover, neither the hexahistidine-FLAG tag on the ␤ 1 subunit nor the G4K mutation in the ␥ 10 subunit greatly affected the ability of these subunits to stimulate adenylyl cyclase.
Interestingly, the phosphorylated forms of the ␤ 1HF ␥ 12 and ␤ 1HF ␥ 10 dimers were significantly less potent and effective in their ability to activate adenylyl cyclase (Fig. 6B). Note that the phosphorylated forms of the ␤␥ dimers are less active than the unphosphorylated forms over the concentration range of 1-40 nM. Fitting the data to rectangular hyperbolas estimates that phosphorylation of the dimers decreases their activity about 50% and shifts their K act from 14 to 37 nM (see legend). As before, controls were performed to determine whether the steps needed to phosphorylate and re-purify the ␤␥ dimers decreased their activity. The protocol causes about a 15% decrement in activity (compare the maximal activities in Fig. 6, A and B). To verify the effect of ␥ 12 subunit phosphorylation on cyclase activity, we prepared dimers with different stoichiometries of phosphorylation and assayed their activity. As shown in Fig. 7, varying the stoichiometry of ␥ 12 subunit phosphorylation between 0 -1 mol/mol resulted in a gradual reduction in the dimer's ability to stimulate type II adenylyl cyclase from about 17 nmol/min/mg of protein to 10 nmol/min/mg of protein. Each of the four curves in Fig. 7 is significantly different from the other (see legend). This result clearly demonstrates that phosphorylation of the serine residue at the N terminus of the ␥ subunit can reduce its ability to stimulate type II adenylyl cyclase.
Previous work has demonstrated that activation of PKC with phorbol 12-myristate 13-acetate (PMA) in Sf9 cells expressing type II adenylyl cyclase results in phosphorylation of the enzyme and an increase in both basal and forskolin-stimulated cyclase activity (31). Moreover, direct phosphorylation of the type II adenylyl cyclase expressed in Sf9 cell membranes with PKC alters its responsiveness to both ␣ and ␤␥ subunits (41). Thus, our finding that phosphorylation of ␥ 12 by the same kinase reduces the ability of the ␤␥ dimer to stimulate type II FIG. 5. Comparison of the ability of phosphorylated and unphosphorylated ␤␥ dimers to activate phospholipase C-␤. A, the ability of native and modified ␤ 1 ␥ 12 and ␤ 1 ␥ 10 dimers to activate phospholipase C-␤ as compared with the effect of ␤ 1 ␥ 2 . The indicated concentrations of ␤␥ dimers were reconstituted with recombinant, turkey phospholipase C-␤ in phospholipid vesicles containing [ 3 H]phosphatidylinositol bisphosphate and phospholipase activity measured as described under "Experimental Procedures." B, comparison of the ability of unphosphorylated (closed circles) and phosphorylated (open triangles) ␤ 1HF ␥ 12 dimers to activate phospholipase C-␤. Data are representative of six independent experiments, each performed in duplicate. Rectangular hyperbolas were fit to the data and statistical differences between the curves determined as described under "Experimental Procedures." There was no significant difference in the curves.

FIG. 6. Comparison of the ability of phosphorylated and unphosphorylated ␤␥ dimers to stimulate type II adenylyl cyclase.
A, Sf9 cells were infected with a recombinant baculovirus encoding the type II adenylyl cyclase, membranes prepared, and the cyclase reaction performed with the indicated concentrations of four recombinant ␤␥ dimers as described under "Experimental Procedures." B, the effect of phosphorylation on the ability of ␤ 1HF ␥ 12 and ␤ 1HF ␥ 10G4K to stimulate type II adenylyl cyclase. The ␤␥ dimers were phosphorylated by protein kinase C ␣ and purified as described under "Experimental Procedures" and the legend to Fig. 1. The cyclase assay was performed with unphosphorylated (closed circles) and phosphorylated (open circles) ␤␥ dimers. Whereas the data does not define a complete curve, fitting the data to rectangular hyperbolas estimates phosphorylation to decrease the V max from 15 nmol/min/mg of protein to about 10 nmol/min/mg of protein and the K act from 14 to 37 nM. The differences were significant (p Ͻ 0.001). The results are representative of 10 similar experiments, each performed in duplicate. Rectangular hyperbolas were fit to the data and statistical differences between the curves determined as described under "Experimental Procedures." adenylyl cyclase frames an interesting problem. To determine the effect of a phosphorylated ␤ 1 ␥ 12 dimer on the activity of type II adenylyl cyclase in PMA-treated cells, we infected Sf9 cells with a recombinant baculovirus encoding type II adenylyl cyclase for 48 h, added 1 M PMA to the medium for 30 min before harvest, and prepared membranes as described under "Experimental Procedures." The type II adenylyl cyclase activity in the control and PMAtreated membranes was compared following treatment with vehicle, 100 nM forskolin, activated G s ␣, and G s plus 20 nM ␤␥. The activities of the control membranes treated with these agents were 0.2, 1.0, 1.2, and 6.25 nmol of cAMP/min/mg of protein, respectively. In keeping with previous results (31), pretreatment of cyclase infected Sf9 cells with PMA did result in a 10 -20% increase in the rates of basal, 100 nM forskolin, G s ␣, and G s plus 20 nM ␤␥-stimulated cyclic AMP synthesis relative to control membranes. Next, dose-response curves were performed with both membrane preparations using phosphorylated and unphosphorylated ␤ 1HF ␥ 12 dimers in the presence of activated G s ␣. Interestingly, in the membranes from PMAtreated cells, the phosphorylated dimers were still markedly less effective than unphosphorylated dimers at stimulating type II adenylyl cyclase. In membranes from control cells, phosphorylation of the dimer reduced the stimulation of cyclase by the following percentages: at 10 nM ␤␥, by 31%; at 20 nM ␤␥, by 40%; and at 40 nM ␤␥, by 45% (n ϭ 5). In membranes from PMA treated Sf9 cells, the reductions were very similar: at 10 nM ␤␥, by 33%; at 20 nM ␤␥, by 40%; and at 40 nM by ␤␥, 55% (n ϭ 5). Thus, in a cell expressing the ␥ 12 subunit, its phosphorylation by PKC would still inhibit the ability of a ␤␥ dimer to stimulate type II adenylyl cyclase activity, even though basal or G sstimulated cyclase activity itself might be slightly elevated by kinase activation. DISCUSSION Identification of the diversity in the family of G protein ␥ subunits has prompted studies to determine whether the differences in these proteins translate to specificity in transmembrane signaling (42). In this regard, one major finding of this study is that, when combined with the ␤ 1 subunit, the newly discovered ␥ 12 and ␥ 10 subunits are equal in potency and efficacy to the well studied ␤ 1 ␥ 2 dimer. Both the ␤ 1 ␥ 12 and ␤ 1 ␥ 10 dimers were fully effective in the receptor coupling assay using the G i1 ␣ subunit and the A1 adenosine receptor and able to maximally activate type II adenylyl cyclase and PLC-␤. As these ␥ subunits are widely expressed in brain and peripheral tissues (12,14,15,37), they are likely to play important roles in signaling by a large number of G protein coupled receptors. A second important finding is that phosphorylation of the ␤ 1 ␥ 12 dimer with protein kinase C has distinct effects on the activity of the molecule, increasing its potency in the receptor coupling assay and inhibiting its ability to stimulate type II adenylyl cyclase. Previous studies have determined that the phosphorylation site in ␥ 12 is Ser 1 at the N terminus of the molecule (12,13). This finding is consistent with our observations that the phosphorylation site created in the ␥ 10G4K subunit makes it a substrate for protein kinase C and that phosphorylation regulates its activity. Although not fully explored, the finding that the dephosphorylation of ␥ 12 is blocked by microcystin suggests that the protein phosphatases that dephosphorylate the protein in the intact cell are most likely protein phosphatase 1 and/or 2A. Overall, the protein kinases and phosphatases regulating the phosphorylation state of the ␥ 12 subunit are those known to participate in responses to receptors generating diacylglycerol and Ca 2ϩ (27,(43)(44)(45), suggesting an important role for this event in cell signaling.
The finding that phosphorylation of the ␥ subunit can reduce the ability of the ␤␥ dimer to stimulate one effector without changing its activity on other effectors adds complexity to the regulation of this signal. Previously, the known mechanisms for regulating ␤␥ activity only involved sequestration by either G protein ␣ subunits or phosducin (3,6,7). Because the GDP bound form of the ␣ subunit has a higher affinity for the ␤␥ subunit, an important mechanism for regulating the activity of the ␤␥ subunit is the return of the active, GTP-bound ␣ subunit to its basal state (1,3,4,7). Indeed, overexpression of ␣ subunits is an effective means of decreasing the activity of ␤␥ subunits in cultured cells (46,47). Phosducin also has a nanomolar affinity for the ␤␥ subunit (48) and, whereas its role in cell function is still emerging, it appears to inhibit the ␤␥ dimer by sequestration of the protein following dissociation of the ␣:␤␥ heterotrimer (49). Because the ␤␥ dimer is needed for coupling the ␣ subunit to receptors (21,38,39), the continued activation of ␣ subunits is decreased. Accordingly, overexpression of phosducin in cultured cells can inhibit the ability of released ␤␥ to activate PLC-␤ or type II adenylyl cyclase (50). Interestingly, phosphorylation of Ser 73 in phosducin via the cyclic AMP-dependent protein kinase decreases its affinity for the ␤␥ dimer (49,51), and treatment of cells overexpressing phosducin with dibutyryl cyclic AMP can relieve its inhibitory effects (50). In this context, the finding that phosphorylation of the ␥ 12 subunit with protein kinase C can inhibit the activity of the ␤␥ dimer toward certain effectors offers new paradigms for understanding the regulation of this important signal.
The observation that phosphorylation of the ␥ 12 subunit on Ser 1 inhibits the ability of the dimer to stimulate type II adenylyl cyclase suggests that the N-terminal region of the ␥ subunit is important for interaction with this effector. In support of this concept, pilot experiments demonstrated that a peptide mimicking the N-terminal 21 amino acids of the ␥ 2 FIG. 7. Stimulation of type II adenylyl cyclase by ␤ 1HF ␥ 12 dimers with different stoichiometries of phosphorylation. Sf9 cell membranes infected with the recombinant baculovirus encoding type II adenylyl cyclase were used to monitor the effect of the ␤␥ dimers on cyclic AMP production as described under "Experimental Procedures." The ␤ 1HF ␥ 12 dimers phosphorylated to different stoichiometries were prepared as follows: 0 mol of PO 4  protein can inhibit the ability of ␤ 1 ␥ 2 to stimulate cyclase. 2 Thus, the negatively charged phosphate group at the N terminus of the protein may inhibit the interaction of the dimer with the cyclase molecule. Two other functional domains of the ␥ subunit have been intensively studied using both biochemical assays and site directed mutagenesis. The C-terminal 15 amino acids and the prenyl group are important for interaction with the plasma membrane, the ␣ subunit, and the receptor (4,5,26,52,53), and the central region of the molecule is important for specific interaction with the ␤ subunits (54,55). The importance of these interaction sites is clearly supported by the x-ray structures of the ␣␤␥ heterotrimer, which indicate a stretch of 15 amino acids beginning at Arg 30 of ␥ 1 that interact with the ␤ subunit and likely contacts between the C-terminal domain of the ␥ subunit and the membrane-␣ subunit interface (56,57). Overall, these findings are interesting because they indicate that three different domains of this small protein are used for interactions with other proteins. Comparison of the N-terminal sequences of the 11 ␥ subunits through the first helical region (Val 26 in ␥ 1 (57)) shows that amino acid identity varies from 15 to 85%. Thus, the diversity of this domain may be important for specificity in effector signaling.
The domains in the PLC-␤ molecule that interact with the ␤␥ dimer have been examined using a number methods. Multiple lines of evidence indicate that the ␤␥ dimer binds near the Y domain in the catalytic core of PLC-␤ (58). Refinement of this location using overexpression of glutathione S-transferase fusion proteins containing small regions of the molecule suggests that the amino acids between Leu 580 and Val 641 are involved in binding the ␤␥ dimer (59). Experiments using small peptides indicate a potential ␤␥ binding domain in the 10 amino acids between Glu 574 and Lys 583 (60). The domains in the ␤ subunit that interact with effectors appear to be similar to those responsible for binding the ␣ subunit (7). However, the domains in the ␥ subunits responsible for interaction with PLC-␤ have not been clearly identified. The finding that the ability of the ␤␥ subunit to stimulate PLC-␤ is not affected by phosphorylation suggests that the central or C-terminal regions in the protein are more likely to interact with PLC-␤ than the N-terminal domain. Alternatively, the negative charge introduced by phosphorylation of the protein may not inhibit binding of the dimer to PLC-␤.
It is especially interesting that phosphorylation of the ␥ 12 subunit increases the affinity of the receptor-␣␤␥ interaction, since the structure of the heterotrimer shows no interaction between the N terminus of the ␥ subunit and the ␣ subunit (56,57). In addition, the N-terminal region of the ␥ subunit is predicted to be some distance from the receptor and the membrane (57). A similar situation occurs with phosducin, where phosphorylation changes its affinity for the ␤␥ dimer (49,51), yet the phosphorylation site is not directly in contact with either the ␤ or ␥ subunit (61). One possible explanation of this result is that phosphorylation may cause an indirect effect on receptor-heterotrimer interactions by an induced conformational change in the ␤␥ dimer. However, a definitive answer should emerge from direct structural analysis of the receptor-␣␤␥ interaction.
The finding that phosphorylation of ␤ 1 ␥ 12 markedly decreases its ability to stimulate type II adenylyl cyclase and focuses the ␤␥ signal toward other effectors is likely to be a broadly important regulatory mechanism. The ␥ 12 subunit is widely expressed and has been demonstrated to be phosphorylated by PKC in intact cells following receptor activation (12,13). The type II adenylyl cyclase is expressed at high levels in the brain and the type IV adenylyl cyclase, with nearly identical regulatory properties, is widely expressed in peripheral tissues such as lung, heart, kidney, and liver (62). Thus, the phosphorylation of the ␥ 12 subunit has the potential to affect the interplay of the Ca 2ϩ and cyclic AMP signaling networks in most cells. As one example, in vascular smooth muscle cells where phosphorylation of the ␥ 12 subunit occurs following application of the contractile agonists vasopressin or angiotensin II (12), this mechanism may augment the ability of Ca 2ϩ to cause contraction by blunting a rise in cyclic AMP, which relaxes smooth muscle (63). A similar mechanism could be used in neural networks to amplify the effects of a Ca 2ϩ signal and blunt those of cyclic AMP.
The fact that only the ␥ 12 subunit has a protein kinase C phosphorylation site in its N terminus is intriguing. Whether other ␥ subunits can be phosphorylated is an important issue. However, the ␥ 12 subunit is highly expressed in all regions of the brain (14) and in most peripheral tissues (12). Thus, ␥ 12 may be the ␥ subunit in the ␤␥ dimers used in many important signaling systems. Our findings that the ␤ 1 ␥ 12 dimer is equal in activity to the better studied ␤ 1 ␥ 2 dimer support this conclusion. However, the ␥ 12 subunit may also be used by cells for undiscovered or specialized signaling roles where phosphorylation is critical to its activity. In this regard, in Swiss 3T3 and C6 cells, dimers containing ␥ 12 appear to be localized to actin stress fibers, whereas those containing the ␥ 5 subunit are found in focal adhesions. Moreover, the ␤␥ 12 dimers appear to bind much more tightly to purified actin filaments than do dimers containing the ␥ 5 subunit (15). These findings suggest that the cytoskeleton may be an important site for ␥ 12 function and may lead to discovery of new roles for the ␤␥ dimer in cell signaling.
The observation that phosphorylation can change the effect of the ␤␥ dimer on certain effectors may have consequences for many signaling systems not studied in this report. The ␤␥ dimer is emerging as an important regulatory signal for a large number of effectors including: K ϩ and Ca 2ϩ channels (64), the ␤ adrenergic receptor kinase (65), phosphatidylinositol 3-kinase (66 -68), mitogen-activated protein kinase (16), and novel kinases such as p21-activated protein kinase (69). It will be important to determine whether phosphorylation alters the activity of the ␤ 1 ␥ 12 dimer toward any of these signaling molecules.