Ribozyme Approach Identifies a Functional Association between the G Protein β1γ7 Subunits in the β-Adrenergic Receptor Signaling Pathway*

The complex role that the heterotrimeric G proteins play in signaling pathways has become increasingly apparent with the cloning of countless numbers of receptors, G proteins, and effectors. However, in most cases, the specific combinations of α and βγ subunits comprising the G proteins that participate in the most common signaling pathways, such as β-adrenergic regulation of adenylyl cyclase activity, are not known. The extent of this problem is evident in the fact that the identities of the βγ subunits that combine with the α subunit of Gs are only now being elucidated almost 20 years after its initial purification. In a previous study, we described the first use of a ribozyme strategy to suppress specifically the expression of the γ7 subunit of the G proteins, thereby identifying a specific role of this protein in coupling the β-adrenergic receptor to stimulation of adenylyl cyclase activity in HEK 293 cells. In the present study, we explored the potential utility of a ribozyme approach directed against the γ7 subunit to identify functional associations with a particular β and αs subunit of the G protein in this signaling pathway. Accordingly, HEK 293 cells were transfected with a ribozyme directed against the γ7 subunit, and the effects of this manipulation on levels of the β and αs subunits were determined by immunoblot analysis. Among the five β αs subunits detected in these cells, only the β1 subunit was coordinately reduced following treatment with the ribozyme directed against the γ7 subunit, thereby demonstrating a functional association between the β1 and γ7 subunits. The mechanism for coordinate suppression of the β1 subunit was due to a striking change in the half-life of the β1 monomerversus the β1 heterodimer complexed with the γ7 subunit. Neither the 52- nor 45-kDa subunits were suppressed following treatment with the ribozyme directed against the γ7 subunit, thereby providing insights into the assembly of the Gs heterotrimer. Taken together, these data show the utility of a ribozyme approach to identify the role of not only the γ subunits but also the β subunits of the G proteins in signaling pathways.

The complex role that the heterotrimeric G proteins play in signaling pathways has become increasingly apparent with the cloning of countless numbers of receptors, G proteins, and effectors. However, in most cases, the specific combinations of ␣ and ␤␥ subunits comprising the G proteins that participate in the most common signaling pathways, such as ␤-adrenergic regulation of adenylyl cyclase activity, are not known. The extent of this problem is evident in the fact that the identities of the ␤␥ subunits that combine with the ␣ subunit of G s are only now being elucidated almost 20 years after its initial purification. In a previous study, we described the first use of a ribozyme strategy to suppress specifically the expression of the ␥ 7 subunit of the G proteins, thereby identifying a specific role of this protein in coupling the ␤-adrenergic receptor to stimulation of adenylyl cyclase activity in HEK 293 cells. In the present study, we explored the potential utility of a ribozyme approach directed against the ␥ 7 subunit to identify functional associations with a particular ␤ and ␣ s subunit of the G protein in this signaling pathway. Accordingly, HEK 293 cells were transfected with a ribozyme directed against the ␥ 7 subunit, and the effects of this manipulation on levels of the ␤ and ␣ s subunits were determined by immunoblot analysis. Among the five ␤ ␣ s subunits detected in these cells, only the ␤ 1 subunit was coordinately reduced following treatment with the ribozyme directed against the ␥ 7 subunit, thereby demonstrating a functional association between the ␤ 1 and ␥ 7 subunits. The mechanism for coordinate suppression of the ␤ 1 subunit was due to a striking change in the halflife of the ␤ 1 monomer versus the ␤ 1 heterodimer complexed with the ␥ 7 subunit. Neither the 52-nor 45-kDa subunits were suppressed following treatment with the ribozyme directed against the ␥ 7 subunit, thereby providing insights into the assembly of the G s heterotrimer. Taken together, these data show the utility of a ribozyme approach to identify the role of not only the ␥ subunits but also the ␤ subunits of the G proteins in signaling pathways.
The most common type of signaling pathway requires the sequential interactions of a seven-transmembrane receptor, a G protein composed of ␣, ␤, and ␥ subunits, and an effector. How the specificity of this signaling pathway is encoded in the protein-protein interactions of ever increasing numbers of these signaling partners is not known. A simple way to encode the specificity would be for each receptor to interact with a specific G protein ␣␤␥ heterotrimer. In this regard, the recent identification of 23 ␣ subunits (1), 6 ␤ subunits (2,3), and 12 ␥ subunits (4 -6) predicts the potential existence of several hundred G protein ␣␤␥ heterotrimers that could serve as intermediaries between a similarly high number of receptors and a somewhat smaller number of effectors. Supporting this scenario, there is increasing evidence that the subunit composition of G protein ␣␤␥ heterotrimers may provide the level of selectivity that is needed to interact with the multitude of receptors and effectors that are now known to exist. Antisense (7,8) and ribozyme (9) strategies have proven to be most useful in identifying which of the potential G protein ␣␤␥ heterotrimers are physiologically relevant. These approaches have the advantage that they allow the selective suppression of individual G protein subunits and the subsequent identification of which signaling pathway(s) are impaired.
In a recent study, we described the first use of the ribozyme strategy to suppress specifically the expression of G protein ␥ 7 subunit, thereby identifying a specific role of this subtype in coupling the ␤-adrenergic receptor to stimulation of adenylyl cyclase activity in HEK 293 cells (9). In the present study, we have explored the use of the same ribozyme approach to identify functional associations of the G protein ␥ 7 subunit with particular G protein ␤ or ␣ subunits in this signaling pathway. To this end, a ribozyme directed against the ␥ 7 subunit was transfected into HEK 293 cells to suppress specifically the expression of the ␥ 7 protein, and then the effects of this manipulation on the levels of the ␤ and ␣ s subunits of the G proteins were determined. Of the ␤ subunits, only the ␤ 1 subunit was coordinately reduced following treatment with the ribozyme directed against the ␥ 7 subunit, thereby demonstrating a functional association between the ␤ 1 and ␥ 7 subunits. By contrast, neither the 52-nor 45-kDa ␣ s subunits were suppressed following treatment with the ribozyme directed against the ␥ 7 subunit. Taken together, the results indicated that the ␥ 7 and ␤ 1 subunits play a specific role in the ␤-adrenergic receptor signaling pathway and that their role cannot be compensated for by other members of these large, multi-gene families.

EXPERIMENTAL PROCEDURES
Ribozyme Design and Delivery to HEK 293 Cells-A chimeric DNA-RNA hammerhead ribozyme targeted against the G protein ␥ 7 subunit mRNA was chemically synthesized and modified by adding two phosphorothioate linkages at the 3Ј-end, as described previously (9). Also, HEK 293 cells were grown in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) supplemented with 10% fetal bovine serum and plated into 100-mm dishes, as described previously (9). The cells were then transfected at approximately 60 -80% confluence with fresh serum-free medium containing premixed ␥ 7 ribozyme (2 M) and the cationic lipid, LipofectAMINE (20 g/ml, Life Technologies, Inc.). At 5 h post-transfection, fetal bovine serum was added to a final concentration of 6%, and sequential addition of 0.5 M ribozyme was supplemented to the total concentration of 4 M at 48 h post-transfection. The control cells were treated identically but without added ribozyme.
Immunoblot Analysis-To determine the effect of ribozyme treatment on G protein subunits, membranes from control and ␥ 7 ribozymetransfected cells were extracted with 1% sodium cholate overnight as described previously (10). The protein concentrations were determined by Amido Black assay, and equal amounts of proteins were resolved on 15% SDS-polyacrylamide gel and transferred to Nitro-plus nitrocellulose (0.45-m pore size, Micron Separations Inc.) using a high temperature procedure described previously (11). Following transfer to nitrocellulose, the blots were probed with anti-G protein subtype-specific antibodies (9,(11)(12)(13)(14). Briefly, after blocking, the nitrocellulose blots were incubated with the primary antibodies for 1 h in high detergent blotto at a dilution of 1:500 for B-69 (␤ 1 ); a dilution of 1:150 for D-17-6 (␤ 2 ); a dilution of 1:100 for B-24 (␤ 3 ); a concentration of 1 g/ml for C-16 (␤ 4 ) (Santa Cruz Biotechnology Inc.); a dilution of 1:2000 for ␤ 5 (Cy-toSignal Research Products); and a dilution of 1:500 for 584 (G␣ s ). After three successive washes, the blots were incubated for 1 h with 125 Ilabeled goat anti-rabbit F(abЈ) 2 fragment (1 ϫ 10 5 dpm/ml, NEN Life Science Products) in high detergent blotto. After washing, the blots were subjected to autoradiography by exposure to BioMax MS film (Eastman Kodak Co.), and the intensity of immunodetectable bands was quantified using the Molecular Dynamics PhosphorImager SI.
Subcellular Fractionation-After transfection, cells were washed with ice-cold phosphate-buffered saline and then lysed with ice-cold lysis buffer which contains 2 mM MgCl 2 , 1 mM EDTA, 20 mM Hepes, 10 mM dithiothreitol, 1 mM aminoethylbenzenesulfonyl fluoride, 1 g/ml pepstatin A, and 1 mM benzamidine. The separation of the cytosolic and particulate fractions was accomplished by centrifugation at 250,000 ϫ g for 30 min. The particulate was extracted with 1% sodium cholate overnight at 4°C. Equal percentages of the cytosolic and solubilized particulate fractions from control and ribozyme-transfected cells were resolved by SDS-polyacrylamide gel electrophoresis and subjected to immunoblot analysis.
Construction of FLAG-tagged Human ␤ 1 cDNA and Hemagglutinintagged Human ␥ 7 cDNAs-In order to monitor the half-life of ␤ 1 subunit in the intact cell setting, epitope tagging was used in this study. The FLAG mammalian transient expression system was used for constructing FLAG-tagged human ␤ 1 subunit, and the influenza virus hemagglutinin (HA) 1 tag was used for constructing HA-tagged human ␥ 7 subunit cDNA. The FLAG system is designed for intracellular expression of amino-terminal Met-FLAG fusion protein which relies on the fusion of the FLAG peptide of eight amino acids (Asp-Tyr-Lys-Asp-Asp-Asp-Asp-Lys) to the coding sequence of a target gene in a pCMV5 expression vector. This FLAG epitope tag is recognized by an anti-FLAG M 2 monoclonal antibody (Sigma). The human ␤ 1 subunit cDNA coding region was generated by polymerase chain reaction and then cloned into the pFLAG-CMV-2 vector. The human G protein ␥ 7 subunit cDNA coding region was generated by polymerase chain reaction with 5Ј primer sequence coding for nine amino acids of HA tag epitope (Tyr-Pro-Tyr-Asp-Val-Pro-Asp-Tyr-Ala) (15) and then cloned into the pCI-neo Mammalian Expression Vector (Promega Corp., Madison, WI). A mutant form of the HA-tagged ␥ 7 m cDNA was generated in the same way by mutating Cys to Ser in the carboxyl-terminal CAAX box, thereby preventing its modification by isoprenylation (16). This HA epitope tag is recognized by anti-HA monoclonal antibody (Roche Molecular Biochemicals).
Metabolic Labeling and Immunoprecipitation-Following transfection, the FLAG-tagged ␤ 1 protein and HA-tagged ␥ 7 protein were detected by immunoprecipitation with anti-FLAG M 2 monoclonal antibody and anti-HA high affinity monoclonal antibody. Briefly, HEK 293 cells grown in 60-mm dishes were transfected with a plasmid encoding either FLAG-tagged hybrid protein ␤ 1 subunit or HA-tagged ␥ 7 protein by LipofectAMINE transfection method. At 24 h post-transfection, cells were incubated for 45 min in methionine-and cysteine-free Dulbecco's modified Eagle's medium and then pulse-labeled with 70 -100 Ci of [ 35 S]methionine (NEN Life Science Products) either for 2.5 min for monomer detection or 1 h for dimer detection, and then chased for the time points indicated in the figure legends. The incorporation of label was stopped by addition of complete medium supplemented with nonradioactive methionine at a final concentration of 1 mM. Cells were lysed in the lysis buffer (50 mM Tris-HCl, pH 7.6, 5 mM MgCl 2 , 0.5% Nonidet P-40, 0.1% Lubrol, 1 mM EDTA, 10 g/ml leupeptin, 10 mM benzamidine, 1 mM aminoethylbenzenesulfonyl fluoride, and 10 g/ml pepstatin A) and then passed through 25-gauge 5/8 needle and freezethawed once in Ϫ80°C freezer. The lysates were depleted of nuclei and cell debris by centrifugation for 10 min at 14,000 rpm in an Eppendorf centrifuge and then pre-cleared twice with 20 l of protein A/G plusagarose (Santa Cruz Biotechnology). Equal amounts of control and transfected cells were subjected to immunoprecipitation. For recovery of the ␤ 1 subunit in the monomer form, immunoprecipitation was performed by overnight incubation at 4°C on a rocker with 15 g/ml anti-FLAG M2 monoclonal antibody. For recovery of the ␤ 1 subunit in the dimer form, immunoprecipitation was carried out with 5 g/ml anti-HA high affinity monoclonal antibody. The immune complexes were precipitated by adsorption to protein A/G plus-agarose for an additional 3 h at 4°C followed by four washes in NET buffer (50 mM Tris-HCl, pH 7.6, 5 mM EDTA, 150 mM NaCl, 0.5% Nonidet P-40). Finally, the immune complexes were dissociated by heating for 10 min in SDS sample buffer. Protein A/G plus-agarose beads were pelleted by centrifugation, and the supernatants were resolved by SDS-polyacrylamide gel electrophoresis. The gel was fixed and soaked for 30 min in Amplify reagent (Amersham Pharmacia Biotech) and then dried and exposed to Kodak BioMax MS film with Kodak Biomax Transcreen-LE Intensifying Screen at Ϫ80°C. The intensity of the immunodetectable bands on the autoradiogram was quantified by densitometry analysis.

Effect of Ribozyme-mediated Suppression of the ␥ 7 Subunit on the Levels of ␤ Subunits in HEK 293
Cells-Under physiological conditions, the ␤ and ␥ subunits exist as a dimer that functions as a single entity (17)(18)(19)(20). Synthesis and assembly of the ␤␥ dimer begins in the cytosol, with subsequent translocation to the plasma membrane being dependent on post-translational lipid modifications of the ␥ subunit (21,22). Therefore, we reasoned that ribozyme-mediated loss of the ␥ 7 protein might limit the formation and translocation of a specific ␤␥ dimer to the plasma membrane. In this event, we expected that the content of one or more ␤ subunits might be reduced in the plasma membrane. Accordingly, the expression of the known ␤ subunits was examined following the treatment of HEK 293 cells with ribozyme specific for the ␥ 7 subunit.
To date, six ␤ subunits have been identified by molecular cloning (2,3). To determine which of these are expressed in HEK 293 cells, it was first necessary to generate (for ␤ 1 , ␤ 2 , and ␤ 3 ) or obtain commercially available (for ␤ 4 and ␤ 5 ) antibodies specific for each ␤ subtype. Then, the specificities of these antibodies were determined by immunoblot analysis of the recombinantly expressed ␤ proteins. As shown in Fig. 1, the sizes of the recombinantly expressed ␤ 1 , ␤ 2 , ␤ 3/6 , ␤ 4 , and ␤ 5 proteins ranged from 35 to 39 kDa and the antibodies reacted only with their corresponding ␤ proteins. To determine whether the ribozyme-mediated loss of the ␥ 7 protein would affect the expression of one or more of these ␤ subunits, membrane proteins from HEK 293 cells treated in the absence (C) or presence (RZ) of the ribozyme were immunoblotted with these ␤ subtype-specific antibodies. As shown in Fig. 2A and quantified for three separate experiments in Fig. 2B, the ␤ 1 , ␤ 2 , ␤ 4 , and ␤ 5 subunits were readily detected in membranes from HEK 293 cells treated in the absence (C) or presence (RZ) of the ribozyme. Under the same condition, the ␤ 3 subunit was not expressed at a detectable level, but whether this is due to the lack of expression or the relatively poor affinity of the ␤ 3 subtype-specific antibody is not known. When the relative amounts of the ␤ 1 , ␤ 2 , ␤ 4 , and ␤ 5 subunits were quantified in the ribozyme-treated membranes and then expressed as a percentage of their levels in control membranes, the levels of the ␤ 2 , ␤ 4 , and ␤ 5 subunits showed no reduction. However, the level of the ␤ 1 subunit showed a striking reduction in ribozyme-treated membranes to 30.2 Ϯ 6.9% of its level in control membranes. By way of comparison, the level of the ␥ 7 subunit showed a similarly striking loss in ribozyme-treated membranes to 21 Ϯ 8.4% of its level in control membranes (Fig. 2, A and B). These data showing coordinate suppression of the ␤ 1 subunit along with the ␥ 7 subunit provide strong evidence for their functional association in the intact cell setting.
Mechanism of Loss of the ␤ 1 Protein-The resulting decrease in the expression of the ␤ 1 protein following ribozyme treatment indicated that the ␥ 7 protein is required for the appearance of the ␤ 1 protein in the membrane. This requirement could reflect the need for the ␥ 7 protein to allow the translocation of the ␤ 1 protein from its place of synthesis in the cytosol to the site of its function in the plasma membrane. To examine this possibility, subcellular fractionation experiments were performed. HEK 293 cells treated in the absence or presence of ribozyme directed against the ␥ 7 subunit were fractionated and the resulting soluble and particulate fractions were subjected to immunoblot analysis with the ␤ 1 subtype-specific antibody (B-69). As shown in Fig. 3A and quantified for three separate experiments in Fig. 3B, the amount of ␤ 1 protein in the particulate fraction was reduced to 38 Ϯ 2.98% (n ϭ 3) in the ribozyme-treated cells (RZ) compared with the control cells (C). However, this loss in the particulate fraction did not result in detectable accumulation of the ␤ 1 protein in the cytosolic fraction in the ribozyme-treated cells (RZ, S) compared with the control cells (C, S). These data support the notion that the ␤ 1 protein, without benefit of association with ␥ 7 protein, may be rapidly degraded.
Change in Half-life of ␤ 1 Subunit in Monomeric Versus Heterodimeric State-To test if the stability of the ␤ 1 subunit is different in the monomeric state versus heterodimeric state in complex with the ␥ 7 subunit, pulse-chase labeling studies were performed for each condition. For this purpose, FLAG-and HA-tagged versions of the human ␤ 1 and ␥ 7 subunits were constructed, respectively. Previous studies have shown that the use of these epitope tags does not interfere with assembly of the ␤␥ dimer (23,24). Moreover, these tagged proteins can be readily monitored with well characterized monoclonal antibodies whose immunoprecipitation capabilities and spectrums of cross-reactivity with non-tagged proteins are already known.
To determine the half-life of the ␤ 1 monomer, HEK 293 cells expressing the FLAG-tagged ␤ 1 subunit alone were labeled with [ 35 S]methionine for 2.5 min and then chased for various time points. Subsequently, cells were lysed and immunoprecipitated with the anti-FLAG M2 antibody in the presence of 0.5% SDS for the recovery of ␤ 1 monomer. The immune complexes were precipitated, denatured in SDS sample buffer, and resolved on 15% SDS-polyacrylamide gels. As shown in Fig. 4A, the 35 S-labeled ␤ 1 monomer was readily recovered at 0 min of chase. However, essentially no labeled ␤ 1 monomer was detected by 3 h of chase. To determine the half-life of the ␤ 1 monomer more accurately, shorter chase periods were employed. As shown in Fig. 4B and quantified for three separate experiments in Fig. 4C, there was progressive loss of the 35 Slabeled ␤ 1 monomer from 0 to 3 h of chase. Based on densitometric and curve fitting analysis, the loss of the labeled ␤ 1 monomer was best fit to a single phase exponential decay, with an estimated half-life of only 20.8 min.
To determine the half-life of ␤ 1 protein in a heterodimeric FIG. 1. Specificity of G protein ␤ subtype-specific antibodies. The recombinantly expressed G protein ␤ 1 , ␤ 2 , ␤ 3 , and ␤ 5 subunits were resolved on a 15% SDS-polyacrylamide gels, transferred to nitrocellulose, and then immunoblotted with ␤ subtype-specific antibodies, as described previously (11,12). Since recombinantly expressed G protein ␤ 4 subunit is not available yet, a bovine brain membrane preparation containing this protein was used as positive control. The sizes of the various ␤ subtypes ranged from 35 to 39 kDa. Br, brain.

FIG. 2. Ribozyme suppression of G protein ␤ subtypes in HEK 293 cells at the protein level.
Membrane proteins from control (C) and ␥ 7 ribozyme-transfected (RZ) cells were extracted, resolved on 15% SDS-polyacrylamide gels, transferred to nitrocellulose, and immunoblotted as described under "Experimental Procedures." Following transfer, the nitrocellulose was cut along the 30-kDa marker; the higher molecular blot was probed with one of the ␤ subtype-specific antibodies, and the lower molecular blot was probed with the ␥ 7 subtype-specific antibody (A-67). A, shows representative immunoblots demonstrating the selective loss of the ␤ 1 and ␥ 7 subunits. B, quantifies the loss of the ␤ 1 and ␥ 7 subunits in the ␥ 7 ribozyme-treated cells. The intensities of the bands from three separate experimental sets of immunoblots were determined by PhosphorImager analysis. The relative amounts of proteins in the ␥ 7 ribozyme-treated cells were expressed as a percentage of their levels in the control cells. The data shown are mean Ϯ S.E. complex with the ␥ 7 protein, HEK 293 cells expressing the ␤ 1 subunit alone or in combination with the HA-tagged ␥ 7 subunit were labeled with [ 35 S]methionine for 1 h. Then cells were lysed and immunoprecipitated with the anti-HA antibody in the absence or presence of 0.5% SDS. Subsequently, immune complexes were resolved on 15% SDS-polyacrylamide gels. As shown in Fig. 5, recovery of the 35 S-labeled ␥ 7 subunit was the same when immunoprecipitated with the anti-HA antibody in the presence (lane 1) or absence (lane 2) of SDS. By contrast, recovery of the labeled ␤ 1 subunit was very different between these two conditions. No labeled ␤ 1 subunit was detected when the ␥ 7 subunit was immunoprecipitated in the presence of SDS (lane 1), but it was readily observed when the ␥ 7 subunit was immunoprecipitated in the absence of SDS (lane 2). These results indicated that the ␤ 1 subunit is present as a complex with the ␥ 7 subunit in the absence of SDS and, as a result, can be brought down with the anti-HA antibody directed against the ␥ 7 subunit. Without transfection of ␥ 7 subunit (lane 3) or without addition of anti-HA antibody (lane 4), neither the ␤ 1 subunit nor the ␥ 7 subunit were immunoprecipitated, showing that the appearance of the ␤ 1 subunit is dependent on the presence of the ␥ 7 subunit or the anti-HA antibody, respectively.
By having validated this method for detection of labeled ␤ 1 subunit in the heterodimeric state, we sought to determine the half-life of the ␤ 1 subunit when complexed with the ␥ 7 subunit. For this purpose, HEK 293 cells expressing the ␤ 1 subunit and the HA-tagged ␥ 7 subunit together were labeled with [ 35 S]methionine for 1 h, chased for the various time points, and then immunoprecipitated with the anti-HA antibody in the absence of SDS. As shown in Fig. 6A and quantified for three separate experiments in Fig. 6B, there was progressive loss of the 35 Slabeled ␤ 1 protein from 0 to 54 h of chase. Based on densitometric and curve fitting analysis, the loss of the majority of the labeled ␤ 1 subunit when complexed with the wild type HAtagged ␥ 7 subunit was best fit to a single phase exponential decay, with an estimated half-life of 14.2 h. However, a small portion of the labeled ␤ 1 protein in the heterodimeric state was stable for at least 54 h. Taken together, these data indicated that the ␤ 1 ␥ 7 dimer is turned over Ͼ40-fold more slowly than the ␤ 1 monomer in the intact cell setting.
To determine whether association with the ␥ 7 subunit is sufficient to stabilize the half-life of the ␤ 1 subunit or whether FIG. 3. Subcellular fractionation of ␤ 1 subunit. After transfection, control (C) and ␥ 7 ribozyme-treated cells (RZ) cells were lysed, separated into soluble (S) and particulate (P) fractions by centrifugation at 250,000 ϫ g, and then equal percentages of these fractions were resolved on 15% SDS-polyacrylamide gels, transferred to nitrocellulose, and immunoblotted with the ␤ 1 -specific antibody (B-69). A shows representative immunoblots demonstrating the loss of the ␤ 1 subunit from the particulate fraction of the ␥ 7 ribozyme-treated cells (RZ, P) compared with the control (C, P) but no detectable accumulation of ␤ 1 subunit in the soluble fraction (RZ, S) compared with control (C, S). B quantifies the change in the ␤ 1 subunit. The intensities of the bands from three separate experimental sets of immunoblots were determined by densitometric analysis. The relative amounts of proteins in the particulate fraction from the ␥ 7 ribozyme-treated cells (RZ) were expressed as a percentage of their levels in the control (C) cells. The data shown are mean Ϯ S.E.  35 S-labeled ␤ 1 subunit was immunoprecipitated with anti-FLAG M2 monoclonal antibody for recovery of ␤ 1 monomer and then resolved on 15% SDS-polyacrylamide gels. Gels were fixed, dried, and exposed to Kodak BioMax MS film with a Kodak Biomax Transcreen-LE Intensifying Screen at Ϫ80°C. The images were quantified by densitometric analysis. A and B show representative autoradiograms of the 35 S-labeled ␤ 1 monomer over longer and shorter time points, respectively. C quantifies the loss of the 35 S-labeled ␤ 1 monomer over the shorter time points (B). Curve fitting analysis of the results from three separate experimental sets revealed the loss of the 35 Slabeled ␤ 1 monomer was best fit to a single phase exponential decay, with an estimated half-life of only 20.8 min.
translocation of the ␤ 1 ␥ 7 complex to the membrane is required for this stabilization, we generated a mutant form of the HAtagged ␥ 7 subunit by replacing the Cys residue with a Ser residue in the carboxyl-terminal "CAAX" motif. As a result of this substitution, the mutant HA-tagged ␥ 7 subunit is not able to undergo prenylation or translocation to the membrane (16). HEK 293 cells expressing the ␤ 1 subunit and the mutant HAtagged ␥ 7 subunit together were labeled with [ 35 S]methionine for 1 h, chased for the various time points, and then immunoprecipitated with the anti-HA antibody in the absence of SDS. As shown in Fig. 7A and quantified for two separate experiments in Fig. 7B, there was a progressive and similar loss of the 35 S-labeled ␤ 1 protein whether complexed with mutant or wild type HA-tagged ␥ 7 subunit (compare with Fig. 6A). Based on densitometric and curve fitting analysis, the loss of the majority of the labeled ␤ 1 subunit when complexed with the mutant HA-tagged ␥ 7 subunit was best fit to a single phase exponential decay, with an estimated half-life of 16.3 h. These data demonstrated that association with the ␥ 7 subunit is sufficient for stabilization of the ␤ 1 subunit.

Effect of Ribozyme-mediated Suppression of the ␥ 7 Subunit on the Levels of 45 and 52 kDa ␣ s Subunits in HEK 293
Cells-To determine whether a ribozyme directed against the ␥ 7 subunit would also affect the expression of one or more of the ␣ s subunits, membrane proteins from HEK 293 cells treated in the absence or presence of the ribozyme against ␥ 7 subunit were immunoblotted with the G protein ␣ s subtype-specific antibody (antibody 584). As shown in Fig. 8A, both the 45-and 52-kDa ␣ s subunits, which are derived from alternative splicing (25), were detected in control and ribozyme-treated cells. The relative amounts of the two forms of ␣ s subunits were quantified in control and ␥ 7 ribozyme-treated cells by densitometry and then expressed as percentages of their levels in control cells for three separate experiments in Fig. 8B. There were no differences in the intensities of the 45-and 52-kDa ␣ s subunits between the control and ribozyme-treated cells even though the intensities of the ␤ 1 and ␥ 7 subunits were markedly and coordinately suppressed in the ribozyme-treated cells. DISCUSSION The G protein ␤␥ subunits exist as a tightly associated complex that plays a prominent role in transducing information from a receptor to an appropriate effector in a specific fashion. Thus, the ␤␥ subunit complex binds directly to receptors (26 -28), where it may act as a conformational switch to direct receptor-G protein coupling (29). Also, the ␤␥ subunit complex interacts directly with a variety of effectors to regulate their activities (for reviews see Refs. 4 and 30). Finally, the ␤␥ subunit complex regulates kinases involved in desensitization of receptors (31). Recently, in vitro (32)(33)(34)(35)(36) and in vivo (7)(8)(9) studies raise the strong possibility that the specific composition of the ␤␥ subunit complex contributes to the recognition of these signaling components.
With the identification of 6 ␤ subtypes (2, 3) and 12 ␥ subtypes (4 -6), it has become increasingly important to decipher which ␤␥ subunit complexes actually exist in intact cells and to identify their roles in particular signaling pathways. In vitro strategies have revealed that certain combinations are not possible, but most combinations are physically able to form ␤␥ dimers (18,37,38). While providing valuable information on structure-function relationships, these reconstitution approaches fall critically short of establishing the roles of specific ␤␥ subunits in particular signaling pathways in the intact cell setting. Increasingly, reverse genetics strategies are being employed to fill this gap (7)(8)(9). In a previous study, we used the ribozyme approach to identify a specific role of the ␥ 7 subunit in the ␤-adrenergic receptor signaling pathway (9). HEK 293 cells transfected with a ribozyme directed against the ␥ 7 subunit showed a specific reduction of the ␥ 7 protein that was associ- ated with a significant attenuation of isoproterenol-stimulated adenylyl cyclase activity. In the present study, we asked whether loss of the ␥ 7 protein would have any effect on the expression and/or membrane localization of the associated ␤ and ␣ s proteins that comprise the G protein heterotrimer in the ␤-adrenergic receptor signaling pathway. Our results show the first successful use of a ribozyme approach directed against a specific ␥ subunit to identify a functional association with a particular ␤ subunit.
Functional Association of ␤ 1 and ␥ 7 Subunits-The ␤ subunit is synthesized in the cytosol and then translocated to the membrane upon association with the appropriately modified ␥ subunit (21,22). This suggested the possibility that ribozymemediated loss of the ␥ 7 protein might be a useful approach to obtain information on its functional association with a particular ␤ subtype. Following treatment of HEK 293 cells with ribozyme specific for the ␥ 7 subunit, only the ␤ 1 protein showed a coordinate reduction with the ␥ 7 protein in the membranes of ribozyme-treated cells compared with control cells (Fig. 2). Next, the mechanism underlying the coordinate suppression of the ␤ 1 protein was explored. Subcellular fractionation studies revealed that loss of the ␤ 1 protein in the membrane did not lead to any detectable increase in the ␤ 1 protein in the cytosol of ribozyme-treated cells. However, pulse-chase labeling studies showed a dramatic difference in the half-life of the ␤ 1 monomer (20.8 min, Fig. 4) compared with the ␤ 1 ␥ 7 dimer (14.2 h, Fig. 6). These data indicate that the ␤ 1 protein is rapidly and specifically degraded when sufficient ␥ 7 protein is not available for dimerization. The ␥ 7 protein undergoes post-translational processing, including prenylation and carboxyl methylation (16,20,21). To explore the influence of processing, a mutant ␥ 7 protein was produced by replacing the Cys residue four residues from the carboxyl terminus with a Ser residue. The effect of this substitution is to prevent prenylation and carboxyl methylation of the mutant ␥ 7 protein. Pulse-chase labeling studies revealed no significant difference in the half-life of the ␤ 1 protein when complexed with the wild type versus the mutant ␥ 7 protein (compare Figs. 6 and 7). This result indicates that dimerization rather than post-translational processing and translocation to the membrane is sufficient to increase the half-life of the ␤ 1 protein. This contrasts with a previous study showing prenylation and carboxyl methylation of RhoA leads to an increase in the half-life of this protein (39). While coordinate suppression of the ␤ 1 and ␥ 7 subunits provides strong evidence for their functional association in the intact cell setting, these results also raise a number of questions. One question is whether the pairing of the ␤ 1 ␥ 7 subunits is specific to HEK 293 The immune complexes were resolved on 15% SDS-polyacrylamide gels, and the gels were fixed, dried, and exposed to Kodak BioMax MS film with a Kodak Biomax Transcreen-LE Intensifying Screen at Ϫ80°C. Membrane proteins from control (C) and ␥ 7 ribozyme-transfected (RZ) cells were extracted, resolved on 15% SDS-polyacrylamide gels, transferred to nitrocellulose, and immunoblotted as described under "Experimental Procedures." A shows representative immunoblots demonstrating no loss of the ␣ s subtypes even though selective loss of the ␤ 1 and ␥ 7 subunits occurred. B quantifies the results in control (C) versus ␥ 7 ribozyme-treated (RZ) cells. The intensities of the bands from three separate experimental sets of immunoblots were determined by PhosphorImager analysis. The relative amounts of proteins in the ␥ 7 ribozyme-treated cells were expressed as a percentage of their levels in the control cells. The data shown are mean Ϯ S.E. cells or is common to various types of cells. A second question revolves around what factors govern the preferential assembly of the ␤ 1 ␥ 7 dimer in cells expressing multiple ␤ and ␥ subtypes. One possibility is a structural feature that favors certain ␤␥ subunit combinations. However, in vitro studies show that the ␤ 1 and ␥ 7 subunits have the potential to interact with a wide variety of subtypes (37,38). Another possibility is a spatial factor that directs selective assembly of the ␤ 1 ␥ 7 dimer within a particular subcellular compartment. In this regard, there is evidence that certain mRNAs are localized within cells, resulting in proteins being synthesized within discrete subcellular compartments (40,41). That the ␥ subtypes are localized within different subcellular domains is increasingly clear (42,43).
Lack of Functional Association of ␣ s and ␥ 7 Subunits-Ribozyme-mediated suppression of the ␥ 7 protein is associated with significant attenuation of ␤-adrenergic receptor-stimulated adenylyl cyclase activity (9). Since loss of the ␥ 7 protein results in corresponding reduction of the ␤ 1 protein, the results of the present study provide strong evidence that the ␥ 7 and ␤ 1 subunits, together with the ␣ s subunit, comprise the G s heterotrimer that couples the ␤-adrenergic receptor to adenylyl cyclase (44). The ␣ s subunit comes in multiple forms that are generated by alternative splicing of a single gene (45,46). Interestingly, ribozyme-mediated suppression of the ␥ 7 subunit has no effect on either the 52-or 45-kDa ␣ s proteins (Fig. 8). This result indicates that the ␣ s subunits associate with the membrane and are stable in the absence of the ␤ 1 ␥ 7 subunit complex. This is consistent with studies showing that the ␣ s subunits contain their own membrane targeting signals (47,48).
Assembly of G Protein Heterotrimers-These and other studies (49) have begun to answer very basic questions regarding the synthesis and assembly of G protein heterotrimer. The G protein ␣, ␤, and ␥ subunits are synthesized on ribosomes in the cytosolic compartment, and they are directed to the plasma membrane by the way of post-translational modification of both ␣ (48, 50) and ␥ (21,22) subunits. Increasing evidence suggests that there is a specific order of addition proceeding from the individual monomers through the ␤␥ dimer to the ␣␤␥ trimer (20,22,47,51). Importantly, this order of assembly has not been examined for a specific combination of G protein ␣␤␥ subunits that is known to exist in the intact cell setting. The results of the present study showing a functional association of the G protein ␣ s ␤ 1 ␥ 7 subunits should allow examination of this question.