Portal de Periódicos da CAPES

Loss of G Protein γ7 Alters Behavior and Reduces Striatal αolf Level and cAMP Production

2003; Elsevier BV; Volume: 278; Issue: 8 Linguagem: Inglês

10.1074/jbc.m211132200

1083-351X

William F. Schwindinger, Kelly S. Betz, Kathryn E. Giger, Angela Sabol, Sarah K. Bronson, Janet D. Robishaw,

Neuroscience and Neuropharmacology Research

The G protein βγ-dimer is required for receptor interaction and effector regulation. However, previous approaches have not identified the physiologic roles of individual subtypes in these processes. We used a gene knockout approach to demonstrate a unique role for the G protein γ7-subunit in mice. Notably, deletion ofGng7 caused behavioral changes that were associated with reductions in the αolf-subunit content and adenylyl cyclase activity of the striatum. These data demonstrate that an individual γ-subunit contributes to the specificity of a given signaling pathway and controls the formation or stability of a particular G protein heterotrimer. The G protein βγ-dimer is required for receptor interaction and effector regulation. However, previous approaches have not identified the physiologic roles of individual subtypes in these processes. We used a gene knockout approach to demonstrate a unique role for the G protein γ7-subunit in mice. Notably, deletion ofGng7 caused behavioral changes that were associated with reductions in the αolf-subunit content and adenylyl cyclase activity of the striatum. These data demonstrate that an individual γ-subunit contributes to the specificity of a given signaling pathway and controls the formation or stability of a particular G protein heterotrimer. The heterotrimeric G proteins control diverse biological processes by conveying signals from cell-surface receptors to intracellular effectors. Although function was originally ascribed to the GTP-bound α-subunit, it is now well established that the βγ-dimer plays active roles in the signaling process through upstream recognition of receptors and downstream regulation of effectors (1Clapham D.E. Neer E.J. Annu. Rev. Pharmacol. Toxicol. 1997; 37: 167-203Google Scholar). Molecular cloning has identified at least 5 β- and 12 γ-subunit genes in the mouse and human genomes. Structurally, γ-subunits are the most diverse, with four subgroups that show less than 50% identity to each other (2Balcueva E.A. Wang Q. Hughes H. Kunsch C. Yu Z. Robishaw J.D. Exp. Cell Res. 2000; 257: 310-319Google Scholar). Moreover, γ-subunits exhibit very different temporal (3Morishita R. Shinohara H. Ueda H. Kato K. Asano T. J. Neurochem. 1999; 73: 2369-2374Google Scholar, 4Schuller U. Lamp E.C. Schilling K. Histochem. Cell Biol. 2001; 116: 149-159Google Scholar) and spatial (5Betty M. Harnish S.W. Rhodes K.J. Cockett M.I. Neuroscience. 1998; 85: 475-486Google Scholar) patterns of expression. These characteristics suggest that γ-subunits have heterogeneous functions. However, comparison of their biochemical properties has revealed only modest differences (6Hou Y. Azpiazu I. Smrcka A. Gautam N. J. Biol. Chem. 2000; 275: 38961-38964Google Scholar, 7Lim W.K. Myung C.S. Garrison J.C. Neubig R.R. Biochemistry. 2001; 40: 10532-10541Google Scholar, 8Kuhn B. Christel C. Wieland T. Schultz G. Gudermann T. Naunyn-Schmiedeberg's Arch. Pharmacol. 2002; 365: 231-241Google Scholar), perhaps because of the inherent limitations of transfection and reconstitution approaches. Gene ablation in mice has proven to be a powerful approach to identifying the functional roles of several G protein α-subunits (9Offermanns S. Oncogene. 2001; 20: 1635-1642Google Scholar). We report the first use of a gene targeting strategy to identify a unique function for a member of the γ-subunit family. The G protein γ7-subunit (Gγ7) was originally cloned from bovine brain (10Cali J.J. Balcueva E.A. Rybalkin I. Robishaw J.D. J. Biol. Chem. 1992; 267: 24023-24027Google Scholar). In situhybridization of rat brain sections revealed that mRNA for Gγ7 is most highly expressed in the striatum (5Betty M. Harnish S.W. Rhodes K.J. Cockett M.I. Neuroscience. 1998; 85: 475-486Google Scholar), where it is found in 40–50% of medium sized neurons in the caudate putamen (11Watson J.B. Coulter P.M. Margulies J.E. de Lecea L. Danielson P.E. Erlander M.G. Sutcliffe J.G. J. Neurosci. Res. 1994; 39: 108-116Google Scholar). The regional expression of mRNA for Gγ7 in the brain mirrors that of the striatum-enriched D1 dopamine receptor (D1R), 1The abbreviations used are: D1R, dopamine D1 receptor; Gγ7, G protein γ7-subunit; MANOVA, multivariate analysis of variance; RT, reverse transcriptase; CLAMS, comprehensive laboratory animal monitoring system; CMV, cytomegalovirus; olf, olfactory; NMDA, N-methyl-d-aspartic acid Gαolf, and adenylyl cyclase Type V (12Zhuang X. Belluscio L. Hen R. J. Neurosci. 2000; 20: RC91Google Scholar), suggesting involvement of Gγ7 in the Gαolf-mediated stimulation of adenylyl cyclase by dopamine. Single cell RT-PCR analysis confirms that D1R and Gγ7 are expressed in the same subset of rat neurons (13Wang Q. Jolly J.P. Surmeier J.D. Mullah B.M. Lidow M.S. Bergson C.M. Robishaw J.D. J. Biol. Chem. 2001; 276: 39386-39393Google Scholar). Ribozyme suppression studies support a role for Gγ7 in the endogenous β-adrenergic receptor pathway (14Wang Q. Mullah B. Hansen C. Asundi J. Robishaw J.D. J. Biol. Chem. 1997; 272: 26040-26048Google Scholar) and the heterologously expressed D1R pathway in human embryonic kidney cells (13Wang Q. Jolly J.P. Surmeier J.D. Mullah B.M. Lidow M.S. Bergson C.M. Robishaw J.D. J. Biol. Chem. 2001; 276: 39386-39393Google Scholar). Mice were segregated by sex and group-housed in plastic microisolator cages in ventilated racks (Thoren Caging Systems, Inc., Hazelton, PA). Mice were given ad libitum access to water and Mouse Diet 9F (Purina Mills, St. Louis, MO). Environmental factors included temperature and humidity control and a 12-h light/dark cycle. The animal facility is maintained as virus antibody-free and parasite-free. Animal research protocols were approved by the Geisinger Clinic institutional animal care and use committee. Southern blot analysis was performed on genomic tail DNA cut with KpnI. The probe used for this analysis was an 0.8-kb fragment 5′ of the modified Gng7 allele. Alternatively, PCR analysis was performed using primers (Invitrogen) shown in Table I. Briefly, primers flanking the 3′ loxP site (JR385 and JR387) were used competitively to amplify the wild type and Gng7 fl alleles, and a third primer upstream of the 5′ loxP site (JR413) was included to amplify simultaneously the wild type and deleted Gng7 alleles. Amplification of the bacteriophage P1 Cre transgene was conducted with JR353 and JR354.Table IPrimers used in this studyNamea(a), (b), and (c) refer to arrows a, b, and c in Fig.1 A.DescriptionbMm., Mus musculus;Rn., Rattus norvegicus.SequenceJR385 (a)Mm. Gng7 exon 3 senseGTG CCA GCC TCT GAG AAT CCA TTC AAJR387 (b)Mm. Gng7 exon 3 antisenseTTG TGA CAC TGC ACC TGC ATG CTTJR413 (c)Mm. Gng7 intron 1 senseGTG ATT GTA TGA CTG GCT AAG GGAJR384Mm.Gng7 exon 2 senseTGT CAG GTA CTA ACA ACG TCG CCC AJR386Mm. Gng7 exon 3 antisenseATC TCA GTT AGG CCA GGC GAC AGT CAJR353P1 Cre senseGTT CGC AAG AAC CTG ATG GAC AJR354P1 Cre antisenseCTA GAG CCT GTT TTG CAC GTT CEFsRn. Eef1a2 senseGGA ATG GTG ACA ACA TGC TGEFasRn. Eef1a2 antisenseCGT TGA AGC CTA CAT TGT CCa (a), (b), and (c) refer to arrows a, b, and c in Fig.1 A.b Mm., Mus musculus;Rn., Rattus norvegicus. Open table in a new tab RNA was isolated from olfactory bulb, frontal cortex, striatum, hypothalamus, midbrain, cerebellum, pons, or whole brain using TRIzol reagent (Invitrogen). First-strand cDNA, prepared from 2 μg of RNA using Moloney murine leukemia virus reverse transcriptase (M-MLV RT, Promega Corp., Madison, WI), was used as a template to amplify the Gγ7 transcript with primers JR384 and JR386 or the elongation factor transcript with primers EFs and EFas, respectively (Table I). Powdered brain tissues were homogenized in HME with proteinase inhibitors (20 mm Hepes, pH 8.0, 2 mm MgCl2, 1 mm EDTA, 1 mm benzamidine, 0.1 mm4-(2-aminoethyl)benzenesulfonyl fluoride, 20 μmleupeptin, 1.4 μm pepstatin, 27 μm1-chloro-3-tosylamido-7-amino-2-heptanone, 28 μml-1-tosylamido-2-phenylethyl chloromethyl ketone). Membranes were obtained by centrifugation onto a sucrose cushion (see below), and membrane-associated proteins were extracted with 1% cholate at 4 °C overnight. Protein concentrations were determined using an Amido Black assay (15Schaffner W. Weissmann C. Anal. Biochem. 1973; 56: 502-514Google Scholar). Equal amounts of proteins were loaded onto 12% Nu-Page gels (Invitrogen) and transferred to NitroPure nitrocellulose (Osmonics, Inc., Westborough, MA) using a high temperature transfer procedure (16Robishaw J.D. Balcueva E.A. Methods Enzymol. 1994; 237: 498-509Google Scholar). Immunoblotting was performed as described previously (14Wang Q. Mullah B. Hansen C. Asundi J. Robishaw J.D. J. Biol. Chem. 1997; 272: 26040-26048Google Scholar) with antisera specific for Gγ7(17Robishaw J.D. Kalman V.K. Moomaw C.R. Slaughter C.A. J. Biol. Chem. 1989; 264: 15758-15761Google Scholar) at a 1:200 dilution, rat Na+/K+-ATPase β-subunit (Research Diagnostics, Inc., Flanders, NJ) at a 1:1000 dilution, Gαs (a generous gift of Dr. Catherine Berlot) used at a 1:500 dilution, Gαolf (a generous gift of Dr. Denis Hervé) used at a 1:1000 dilution, Gαo (16Robishaw J.D. Balcueva E.A. Methods Enzymol. 1994; 237: 498-509Google Scholar) used at 1:500, Gα13 (a generous gift of Dr. N. Dhanasekaran) used at 1:1000, Gαq/11 (16Robishaw J.D. Balcueva E.A. Methods Enzymol. 1994; 237: 498-509Google Scholar) used at 1:200, and Ras (BD Biosciences) used at a 1:2000 dilution. Immunoblots were imaged with a PhosphorImager and analyzed with ImageQuant software (Amersham Biosciences). Brain tissues were homogenized in Buffer A (10 mm Tris, pH 7.4, 1 mm EDTA, 1 mm dithiothreitol, 0.3 mm4-(2-aminoethyl)benzenesulfonyl fluoride, 30 μmleupeptin, 1 μm pepstatin A) with 10% sucrose using a Brinkmann homogenizer (Brinkmann Instruments). Membranes were then isolated by centrifugation (65 min at 100,000 × g) onto a cushion of Buffer A with 44.5% (w/v) sucrose. The membranes at the interface were transferred to a new tube, washed twice with Buffer A, and collected by centrifugation (30 min at 100,000 ×g). Protein concentrations were determined with Coomassie Plus (Pierce). Adenylyl cyclase activity (18Johnson R.A. Alvarez R. Salomon Y. Methods Enzymol. 1994; 238: 31-56Google Scholar) was determined by incubating membrane protein (20 μg) at 30 °C for 10 min in 0.1 ml of buffer containing 50 mm Hepes, pH 7.4, 0.2 mm EGTA, 1 mm MgCl2, 1 mm dithiothreitol, 0.5 mm ATP, 1 mmisobutylmethylxanthine, 5 mm creatine phosphate, 50 units/ml creatine phosphokinase, and various agonists as indicated in Fig. 3. For stimulation of striatal membranes with dopamine agonists, more consistent results were obtained using buffer containing 10 mm imidazole, pH 7.3, 0.2 mm EGTA, 0.5 mm MgCl2, 0.5 mm dithiothreitol, 0.1 mm ATP, and 0.5 mm isobutylmethylxanthine (19Friedman E. Jin L.Q. Cai G.P. Hollon T.R. Drago J. Sibley D.R. Wang H.Y. Mol. Pharmacol. 1997; 51: 6-11Google Scholar). Reactions were terminated by the addition of 1 ml of 0.1n HCl, 1 mm EDTA. The cAMP concentrations were determined by automated radioimmunoassay using a Gamma-Flo instrument (Atto Instruments, Inc., Potomac, MD) as described previously (20Krupinski J. Lehman T.C. Frankenfield C.D. Zwaagstra J.C. Watson P.A. J. Biol. Chem. 1992; 267: 24858-24862Google Scholar). A functional observational battery was used to assess the behavior of mice outside of their home cages. On 5 successive days, mice were placed in the center of a translucent polypropylene box (45 × 30 × 60 cm). The behavior of each mouse was observed for 2 min by a researcher who was unaware of the genotypes of the mice. The latency to first step in seconds was recorded. The number of times the mouse reared was counted. The fraction of the box floor explored by the mouse was estimated. The startle response of the mouse to the sound of a latex glove being snapped was graded. After the mouse was removed from the box, the fecal boli and urine pools left in the box were counted, and the box was wiped clean with a contact disinfectant. Locomotor activities were quantified in CLAMS cages (Columbus Instruments, Columbus, OH). The cages were clear plastic boxes (20 × 10 × 12.5 cm) fitted with three rows of eight photoelectric sensors (x, y and z directions). The mice were placed in the CLAMS cages at 11 a.m. and remained in the cages for 3 h. During this time the mice had ad libitum access to water. Every minute the number of total and consecutive photobeam breaks for each of the three sensor arrays and the number of contacts with the sipper tube were recorded. Acoustic startle was measured using an SR-Lab startle reflex system (San Diego Instruments, San Diego, CA). A speaker delivered a continuous background noise of 65 db. During each trial a 40-ms pulse of broadband noise at 110 db was delivered. In prepulse trials, 20 ms of broadband noise at 70, 80, or 90 db was administered at 100 ms before the pulse. Each mouse received six consecutive pulse trials followed by six of each prepulse, pulse, or no stimulation trial, in random order. The interval between trials varied from 12 to 30 s. Data were collected at 1-ms intervals starting 50 ms before the first stimulus. The startle magnitude was calculated by summing the voltage from the startle chamber (minus the base-line voltage) over the 100 ms after the onset of the pulse. A single chamber was used for all mice. Data are presented as the mean ± S.E. Data were analyzed by t test, chi square, or repeated measures MANOVA using JMP (SAS Institute Inc., Cary, NC). To investigate the physiologic function of Gγ7, we targeted Gng7, the murine gene encoding Gγ7, using a Cre/Loxstrategy to provide the potential for conditional inactivation of the gene in future studies (Fig. 1 A). The complete coding region of Gng7 was contained within an ∼10-kb mouse genomic clone (21Wattler S. Kelly M. Nehls M. BioTechniques. 1999; 26: 1150-1160Google Scholar). This included the two coding exons split by an ∼1.1-kb intron. A targeting vector was designed to flank the coding region of Gng7 with loxPsites. Targeted embryonic stem cells, chimeric mice, and F1heterozygotes with the floxed Gng7 allele (Gng7 fl/+) were produced at Lexicon Genetics, Inc., The Woodlands, TX. Gng7 fl/+ mice were mated with mice carrying the Cre recombinase driven by the CMV promoter, BALB/c-TgN(CMV-Cre)#Cgn (Jackson Laboratories, Bar Harbor, Maine). This resulted in deletion of the coding region of Gng7 following Crerecombinase-mediated excision (Fig. 1 B). Heterozygous mice with the deleted Gng7 allele (Gng7 +/−) were back-crossed with C57BL/6J mice (Jackson Laboratories) for up to five generations, to eliminate the CMV-Cre transgene and to obtain a more homogeneous genetic background, and then were intercrossed to obtain the mice used in the following experiments (Fig. 1 C). To confirm the effectiveness of the gene knockout, expression of Gng7 was examined in brain, where the gene is most highly expressed (10Cali J.J. Balcueva E.A. Rybalkin I. Robishaw J.D. J. Biol. Chem. 1992; 267: 24023-24027Google Scholar). RT-PCR analysis demonstrated that the Gγ7 mRNA was absent in brains from Gng7−/− mice (Fig. 1 D). Likewise, immunoblot analysis with Gγ7 antisera (17Robishaw J.D. Kalman V.K. Moomaw C.R. Slaughter C.A. J. Biol. Chem. 1989; 264: 15758-15761Google Scholar) showed that the Gγ7 protein was reduced by 50 ± 6% in cholate-solubilized membrane extracts of brains fromGng7 +/− mice and was not detectable in brains from Gng7 −/− mice (Fig. 1 E). Genotype analysis of offspring of heterozygous (Gng7 +/−) intercrosses revealed the expected numbers of wild type, heterozygous, and homozygous mice, indicating that disruption of Gng7 did not affect survival to weaning. There were no significant differences in the weights ofGng7−/− mice and their wild type littermates. Moreover, no increased mortality was seen inGng7−/− mice over a 6-month median observation time. Finally, in two homozygous crosses,Gng7−/− mice were fertile and weaned litters of apparently normal size. These observations are in stark contrast to mice with a deficiency of other components of the D1R signal transduction pathway. Mice with a homozygous deletion ofGnas, the gene encoding Gαs, die in utero before implantation (22Yu S. Yu D. Lee E. Eckhaus M. Lee R. Corria Z. Accili D. Westphal H. Weinstein L.S. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 8715-8720Google Scholar). Most mice with a homozygous deletion of Gnal (23Belluscio L. Gold G.H. Nemes A. Axel R. Neuron. 1998; 20: 69-81Google Scholar), the gene encoding Gαolf, fail to thrive and die in the neonatal period. Moreover, although Gnal −/− mice are fertile, dams lack fostering skills and none of their litters survive. Similar failure to thrive is observed in mice with a homozygous deletion ofDrd1a (24Drago J. Gerfen C.R. Lachowicz J.E. Steiner H. Hollon T.R. Love P.E. Ooi G.T. Grinberg A. Lee E.J. Huang S.P. Bartlett P.F. Jose P.A. Sibley D.R. Westphal H. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 12564-12568Google Scholar, 25Xu M. Moratalla R. Gold L.H. Hiroi N. Koob G.F. Graybiel A.M. Tonegawa S. Cell. 1994; 79: 729-742Google Scholar), the gene encoding the D1R. These feeding and fostering deficits may be related to loss of the sense of smell inGnal −/− mice (23Belluscio L. Gold G.H. Nemes A. Axel R. Neuron. 1998; 20: 69-81Google Scholar) and impaired motivated behavior in Drd1a −/− mice (26Smith D.R. Striplin C.D. Geller A.M. Mailman R.B. Drago J. Lawler C.P. Gallagher M. Neuroscience. 1998; 86: 135-146Google Scholar), respectively. Thus, Gng7−/− mice display a more circumscribed phenotype than mice with a deficiency of Gnas,Gnal, or Drd1a; and offer a more robust animal model for behavioral studies. Behavior reflects the underlying function of the brain, making it a sensitive indicator of alterations induced by genetic manipulation (27Tarantino L.M. Bucan M. Hum. Mol. Genet. 2000; 9: 953-965Google Scholar). Anecdotal reports of increased handling reactivity from the animal care technicians provided the first evidence for a behavioral phenotype. More systematic screens of their behavior were carried out using a functional observational battery (28Moser V.C. Neurotoxicol. Teratol. 1990; 12: 483-488Google Scholar). For this purpose, mice were observed outside of their home cages; measurements of neuromuscular function (gait abnormalities), sensorimotor defects (auditory startle), autonomic responses (fecal boli and urine pools), and activity levels (latency to first step, rears, and exploration) were scored on predefined rating scales by a trained observer who was unaware of the genotypes of the animals. Using this test battery, the most striking observation was that the Gng7 −/−mice exhibited an increased startle response as compared with their wild type littermates. To quantify this effect, a Startle Reflex System (San Diego Instruments) was used to measure the startle reactivity ofGng7 −/− mice and their wild type littermates. Notably, the Gng7 −/− mice displayed a greater startle amplitude than their wild type littermates for each stimulus tested (Fig. 2 A). Prepulse inhibition refers to the reduction in startle response that occurs when the startling stimulus is preceded by a stimulus of lower intensity and is often used as a measure of sensorimotor gating. Despite the enhanced startle response, the Gng7 −/− mice showed a similar degree of prepulse inhibition compared with their wild type littermates for each stimulus tested (Fig. 2 A,inset). The finding of increased startle response but normal prepulse inhibition of the startle response is a phenotype that has been observed previously in mice with mutations in the glycine binding site of the N1-subunit of the NMDA receptor,Grin1 D481N/D481N (29Kew J.N. Koester A. Moreau J.L. Jenck F. Ouagazzal A.M. Mutel V. Richards J.G. Trube G. Fischer G. Montkowski A. Hundt W. Reinscheid R.K. Pauly-Evers M. Kemp J.A. Bluethmann H. J. Neurosci. 2000; 20: 4037-4049Google Scholar) andGrin1 D481N/K483Q (30Ballard T.M. Pauly-Evers M. Higgins G.A. Ouagazzal A.M. Mutel V. Borroni E. Kemp J.A. Bluethmann H. Kew J.N. J. Neurosci. 2002; 22: 6713-6723Google Scholar). This is intriguing because earlier studies had shown that dopamine acting through cAMP to stimulate protein kinase A, and through DARPP-32 (32-kDa dopamine- and cAMP-responsive phosphoprotein) to inhibit protein phosphatase 1, increases the phosphorylation state of the N1-subunit of the N-methyl-d-aspartic acid (NMDA) receptor and potentiates NMDA responses (31Snyder G.L. Fienberg A.A. Huganir R.L. Greengard P. J. Neurosci. 1998; 18: 10297-10303Google Scholar). This suggests that Gγ7may be involved in the signal transduction pathway that regulates the response of striatal γ-aminobutyric acid-producing (GABAergic) neurons to glutamate, a pathway that has been implicated in the pathogenesis of schizophrenia (32Mohn A.R. Gainetdinov R.R. Caron M.G. Koller B.H. Cell. 1999; 98: 427-436Google Scholar). Alternate mechanisms by which a deficiency of Gγ7 may increase the startle response are suggested by other mouse models. Mice with a deficiency of adrenergic α2c-receptors demonstrate an increased startle response; however, these mice have diminished prepulse inhibition (33Sallinen J. Haapalinna A. Viitamaa T. Kobilka B.K. Scheinin M. J. Neurosci. 1998; 18: 3035-3042Google Scholar). Transgenic mice expressing a dominant mutant of the human inhibitory glycine receptor α1-subunit, TgN(GLRA1R271Q), as well as mice with the recessive mutation spasmodic(Glra spd/spd), display a complex neuromuscular phenotype that includes increased startle response and mimics the human disease of hyperekplexia (34Becker L. von Wegerer J. Schenkel J. Zeilhofer H.U. Swandulla D. Weiher H. J. Neurosci. 2002; 22: 2505-2512Google Scholar). The abundant expression of Gng7 in the striatum suggests a possible role in control of locomotor activity. As a quantitative test of this activity, Gng7−/− mice and their wild type littermates were placed in CLAMS cages (Columbus Instruments), which are equipped with photobeams to measure movement in thex, y, and z directions. Both wild type and Gng7−/− mice exhibited elevated horizontal and vertical locomotor activity when introduced into this new environment. Gng7 −/− mice showed less horizontal (Fig. 2 B) and vertical activity (not shown) than their wild type littermates at all time points, but these differences were not statistically significant. Habituation refers to the tendency for the increased locomotor activity to decline upon repeated or sustained exposure to a new environment, which is often used as a measure of a learned response. Importantly, both groups of mice displayed comparable levels of habituation between and within sessions. On the basis of pharmacologic studies showing D1 dopamine agonists produce a strong stimulatory effect on locomotor activity (35Cabib S. Castellano C. Cestari V. Filibeck U. Puglisi-Allegra S. Psychopharmacology (Berl.). 1991; 105: 335-339Google Scholar), one might have expected a more marked decrease in locomotor activity in Gng7 −/− mice. In this regard, however, other gene knockout studies have not provided strong support for the pharmacologic studies. For example,Drd1a −/− mice have been variously reported as being either hypoactive (24Drago J. Gerfen C.R. Lachowicz J.E. Steiner H. Hollon T.R. Love P.E. Ooi G.T. Grinberg A. Lee E.J. Huang S.P. Bartlett P.F. Jose P.A. Sibley D.R. Westphal H. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 12564-12568Google Scholar, 26Smith D.R. Striplin C.D. Geller A.M. Mailman R.B. Drago J. Lawler C.P. Gallagher M. Neuroscience. 1998; 86: 135-146Google Scholar), hyperactive (25Xu M. Moratalla R. Gold L.H. Hiroi N. Koob G.F. Graybiel A.M. Tonegawa S. Cell. 1994; 79: 729-742Google Scholar), or neither but showing an altered pattern of activity (36Clifford J.J. Tighe O. Croke D.T. Sibley D.R. Drago J. Waddington J.L. Neuropharmacology. 1998; 37: 1595-1602Google Scholar). Moreover,Gnal −/− mice have increased locomotor activity in an open field (12Zhuang X. Belluscio L. Hen R. J. Neurosci. 2000; 20: RC91Google Scholar), which is the opposite of the result expected based on pharmacologic studies. This variability may reflect the complex nature of locomotor activity (37Waddington J.L. Clifford J.J. McNamara F.N. Tomiyama K. Koshikawa N. Croke D.T. Prog. Neuropsychopharmacol. Biol. Psychiatry. 2001; 25: 925-964Google Scholar), the confounding effects of olfactory deficits in Gnal −/− mice (23Belluscio L. Gold G.H. Nemes A. Axel R. Neuron. 1998; 20: 69-81Google Scholar), or unidentified compensatory changes that may occur in knockout mice (38Sibley D.R. Annu. Rev. Pharmacol. Toxicol. 1999; 39: 313-341Google Scholar). Further study will be needed to clarify this issue, including more a restricted disruption of Gng7 within specific regions of the brain that is made possible by our floxed mouse model. Finally, the expression of Gng7 along withDrd1a in a subset of neurons within the striatum suggests a possible role in regulation of adenylyl cyclase activity (13Wang Q. Jolly J.P. Surmeier J.D. Mullah B.M. Lidow M.S. Bergson C.M. Robishaw J.D. J. Biol. Chem. 2001; 276: 39386-39393Google Scholar). To evaluate this possibility, adenylyl cyclase activity was compared between those regions of the brain that normally expressGng7, such as striatum, and other regions that do not normally express Gng7, such as cerebellum (Fig.3 A). Intriguingly, dopamine and the D1-specific agonist 6-chloro-PB had stimulatory effects on adenylyl cyclase activity in the striatum from wild type mice, but the responses were virtually abolished in the striatum from Gng7 −/− mice (Fig.3 B). By contrast, dopamine had no effect on adenylyl cyclase activity in the cerebellum, but the response to isoproterenol was comparable in both wild type and Gng7 −/− mice (Fig. 3 C). These results establish a functional link between the expression of Gγ7 and adenylyl cyclase activity. Forskolin is a potent activator of adenylyl cyclase activity. Remarkably, forskolin had a potent stimulatory effect on adenylyl cyclase activity in the striatum from wild type mice, but the response was reduced by 50% in the striatum fromGng7 −/− mice (Fig. 3 D). By contrast, response to forskolin was comparable in the cerebellum from both groups of mice (Fig. 3 E). Taken together, these data demonstrate that Gγ7 has an important but regionalized role in the regulation of adenylyl cyclase activity in the brain. Because the striatal adenylyl cyclase Type V isoform (39Glatt C.E. Snyder S.H. Nature. 1993; 361: 536-538Google Scholar) is synergistically activated by a combination of forskolin and Gαs (40Zimmermann G. Zhou D. Taussig R. J. Biol. Chem. 1998; 273: 19650-19655Google Scholar), one mechanism that could account for defects in both receptor-mediated and forskolin-stimulated adenylyl cyclase activity is a reduced level of the stimulatory G protein α-subunit in the striatum of Gng7 −/− mice. To test this possibility, the levels of Gαs and Gαolf, the two known activators of adenylyl cyclase activity, were determined. Notably, the levels of Gαs were comparable in the striatum from both groups of mice level, but the amount of Gαolf was reduced by 82 ± 3% in the striatum (n = 8) from Gng7 −/− mice (Fig. 4 A). To provide additional evidence for a specific reduction in Gαolf, we examined the levels of representative members of other G protein α-subunit subunit families. Notably, the levels of Gαo, Gα13, and Gαq/11 were not significantly reduced in the striatum of Gng7 −/− mice (Fig.4 B). These results provide the first evidence that loss of a γ-subunit can result in loss of its α-subunit partner. In summary, this paper provides several novel insights. First, Gγ7 plays a unique role in regulation of adenylyl cyclase signaling in certain regions of the brain. This is demonstrated by the finding that loss of Gγ7 produces both a behavioral and a biochemical phenotype, indicating that other types of γ-subunit are not able to substitute for this function. Thus, members of the γ-subunit family are not functionally interchangeable in the context of the whole animal. Second, Gγ7 plays a role in the stabilization or formation of a G protein heterotrimer (αolfβγ7) that is required for stimulation of adenylyl cyclase activity in the striatum. This is substantiated by the finding that loss of Gγ7 reduces the level of Gαolf in the striatum in a specific and coordinate fashion. Further studies are needed to address the underlying mechanisms and to determine whether this process is applicable to other γ-subunits. If widespread, these results will reveal an important new signaling paradigm, namely, the level of a specific γ-subunit controls the stability or assembly of a particular G protein heterotrimer. This, in turn, provides a probable basis for the selectivity of the multitude of G protein-coupled receptor signaling pathways that are now known to exist. Finally, the production of mice lacking Gγ7 provides a unique mouse model for the study of numerous diseases in which dysfunction of the adenylyl cyclase signaling pathway in the striatum has been implicated, such as Parkinson's disease, Huntington's chorea, Tourette's syndrome, and schizophrenia. We thank the outstanding technicians in our animal care facility, Cynthia J. Rhone, Gail L. Gregory, and Shannon Wescott. We are grateful to Dr. Denis Hervé for providing the αolf antibody, to Dr. Catherine Berlot for supplying the αs antibody and providing helpful advice, and to Dr. N. Dhanasekaran for supplying the α13 antibody. We are indebted to Drs. Michael C. Nehls and Jean-Pierre Revelli (Lexicon Genetics Inc.) for producing the Gng7 +/flmice.