Tubulin stimulates adenylyl cyclase activity in C6 glioma cells by bypassing the beta-adrenergic receptor: a potential mechanism of G protein activation

While the cytoskeleton is known to play several roles in the biology of the cell, one role, which has been revealed only recently, is that of a participant in the signal transduction process. Tubulin binds specifically to the alpha subunits of Gs (stimulatory GTP-binding regulatory protein of adenylyl cyclase), Gi1 (inhibitory protein of adenylyl cyclase), and Gq and transactivates those molecules through direct transfer of GTP. The relevance of this transactivation process to G proteins which are normally activated by a neurotransmitter-occupied receptor is the subject of this study. C6 glioma cells, made permeable with saponin, retained tight coupling between Gs and the beta-adrenergic receptor. Although 5-guanylylimidodiphosphate (GppNHp) was incapable of activating Gs (and subsequently, adenylyl cyclase) in the absence of agonist, tubulin with GppNHp bound (tubulin-GppNHp) activated adenylyl cyclase with an EC(50) of 30 nM. Desensitization of beta-adrenergic receptors by isoproterenol exposure had no effect on the ability of tubulin-GppNHp to activate Gs and adenylyl cyclase. When the photoaffinity GTP analog, azidoanilido GTP (AAGTP; P3(4-azidoanilido)-P1-5'-GTP), was added to C6 membranes or permeable C6 cells, it was only weakly incorporated by G alpha s in the absence of isoproterenol. When the same concentration of dimeric tubulin with AAGTP bound was introduced, AAGTP was transferred from tubulin to G alpha s, activating the latter species. Similar 'preferential' activation of G alpha s by tubulin-AAGTP versus the free nucleotide was seen using purified components. Thus, membrane-associated tubulin may serve to activate G alpha s, independent of signals not normally coupled to that protein. Tubulin may act as an agent to link a variety of membrane-associated signalling systems.     

Identification by molecular cloning of two forms of the alpha-subunit of the human liver stimulatory (GS) regulatory component of adenylyl cyclase

Two DNA molecules complementary to human liver mRNA coding for the alpha-subunit of the stimulatory regulatory component Gs of adenylyl cyclase were cloned. One of the two forms is a full-length cDNA of 1614 nucleotides plus a poly(A) tail of 59 nucleotides. The deduced sequence of 394 amino acids encoded by its open reading frame is essentially identical to that of the alpha-subunits of Gs identified by molecular cloning from bovine adrenals, bovine brain and rat brain. Two independent clones of the other type of cDNA were isolated. Both were incomplete, beginning within the open reading frame coding for the alpha s polypeptide. One codes for amino acids 5 through 394 and the other for amino acids 48 through 394 of the above described cDNA of 1614 nucleotides, and both have the identical 3'-untranslated sequence. They differ from the first cDNA, however, in that they lack a stretch of 42 nucleotides (numbers 214 through 255) and have nucleotides 213 (G) and 256 (G) replaced with C and A, respectively. This results in a predicted amino acid composition of another alpha-subunit of Gs that is shorter by 14 amino acids and contains two substitutions (Asp for Glu and Ser for Gly) at the interface between the deletion and the unchanged sequence. We call the smaller subunit alpha s1 and the larger alpha s2. This is the first demonstration of a structural heterogeneity in alpha s subunits that is due to a difference in amino acid sequence.     

Purification and characterization of Go alpha and three types of Gi alpha after expression in Escherichia coli

Complementary DNAs for the G protein alpha subunits Gi alpha 1, Gi alpha 2, Gi alpha 3, and Go alpha were expressed in Escherichia coli, and the four proteins were purified to homogeneity. The recombinant proteins exchange and hydrolyze guanine nucleotide, are ADP-ribosylated by pertussis toxin, and interact with beta gamma subunits. The rates of dissociation of GDP from Gi alpha 1 and Gi alpha 3 (0.03 min-1) are an order of magnitude slower than that from rGo alpha; release of GDP from Gi alpha 2 is also relatively slow (0.07 min-1). However, the values of kcat for the hydrolysis of GTP by rGo alpha and the three rGi alpha proteins are approximately the same, about 2 min-1 at 20 degrees C. The recombinant proteins restore inhibition of Ca2+ currents in pertussis toxin-treated dorsal root ganglion neurons in response to neuropeptide Y and bradykinin, indicating that the proteins can interact functionally with all necessary components of at least one signal transduction system. The two different receptors function with different arrays of G proteins to mediate their responses, since all four G proteins restored responses to bradykinin, while Gi alpha 2 was inactive with neuropeptide Y. Despite these results, high concentrations of activated Gi alpha proteins are without effect on adenylyl cyclase activity, either in the presence or absence of forskolin or Gs alpha, the G protein that activates adenylyl cyclase. These results are consistent with the hypothesis that G protein beta gamma subunits are primarily responsible for inhibition of adenylyl cyclase activity.     

Genetic analysis of the Drosophila Gs(alpha) gene.

One of the best understood signal transduction pathways activated by receptors containing seven transmembrane domains involves activation of heterotrimeric G-protein complexes containing Gs(alpha), the subsequent stimulation of adenylyl cyclase, production of cAMP, activation of protein kinase A (PKA), and the phosphorylation of substrates that control a wide variety of cellular responses. Here, we report the identification of "loss-of-function" mutations in the Drosophila Gs(alpha) gene (dgs). Seven mutants have been identified that are either complemented by transgenes representing the wild-type dgs gene or contain nucleotide sequence changes resulting in the production of altered Gs(alpha) protein. Examination of mutant alleles representing loss-of-Gs(alpha) function indicates that the phenotypes generated do not mimic those created by mutational elimination of PKA. These results are consistent with the conclusion reached in previous studies that activation of PKA, at least in these developmental contexts, does not depend on receptor-mediated increases in intracellular cAMP, in contrast to the predictions of models developed primarily on the basis of studies in cultured cells.     

Activating mutations in the NH2- and COOH-terminal moieties of the Gs alpha subunit have dominant phenotypes and distinguishable kinetics of adenylyl cyclase stimulation.

The alpha subunit polypeptides of the G proteins Gs and Gi2 stimulate and inhibit adenylyl cyclase, respectively. The alpha s and alpha i2 subunits are 65% homologous in amino acid sequence but have highly conserved GDP/GTP binding domains. Previously, we mapped the functional adenylyl cyclase activation domain to a 122 amino acid region in the COOH-terminal moiety of the alpha s polypeptide (Osawa et al: Cell 63:697-706, 1990). The NH2-terminal half of the alpha s polypeptide encodes domains regulating beta gamma interactions and GDP dissociation. A series of chimeric cDNAs having different lengths of the NH2- or COOH-terminal coding sequence of alpha s substituted with the corresponding alpha i2 sequence were used to introduce multi-residue non-conserved mutations in different domains of the alpha s polypeptide. Mutation of either the amino- or carboxy-terminus results in an alpha s polypeptide which constitutively activates cAMP synthesis when expressed in Chinese hamster ovary cells. The activated alpha s polypeptides having mutations in either the NH2- or COOH-terminus demonstrate an enhanced rate of GTP gamma S activation of adenylyl cyclase. In membrane preparations from cells expressing the various alpha s mutants, COOH-terminal mutants, but not NH2-terminal alpha s mutants markedly enhance the maximal stimulation of adenylyl cyclase by GTP gamma S and fluoride ion. Neither mutation at the NH2- nor COOH-terminus had an effect on the GTPase activity of the alpha s polypeptides. Thus, mutation at NH2- and COOH-termini influence the rate of alpha s activation, but only the COOH-terminus appears to be involved in the regulation of the alpha s polypeptide activation domain that interacts with adenylyl cyclase.     

Existence of multiple novel Gs alpha splice variants in acute leukemia patients

The alpha subunit of the stimulatory G protein, Gs alpha, is involved in stimulation of the adenylate cyclase pathway of signal transduction. In this study, we investigated the status of the Gs alpha gene in 29 acute leukemia patients and identified three novel splice variants (designated Gs alpha L-1, Gs alpha L-2, and Gs alpha L-3), possibly derived from aberrant splicing. All of the splice variants have in-frame deletions, removing the functional domain responsible for GTPase activity of Gs alpha, and would encode truncated proteins of 160(Gs alpha L-1), 90(Gs alpha L-2) and 70(Gs alpha L-3) amino acids, respectively. The data suggest that these novel products may be implicated in an as-yet-unidentified signal transduction pathway in hematopoietic cells.     

Synthesis in Escherichia coli of GTPase-deficient mutants of Gs alpha

We have reduced the GTPase activity of the alpha subunit of Gs, the guanine nucleotide-binding regulatory protein that stimulates adenylyl cyclase, by introduction of point mutations analogous to those described in p21ras. Mutants G49V and Q227L differ from the wild type protein in the substitution of Val for Gly49 and Leu for Gln227, respectively (analogous to positions 12 and 61 in p21ras). Wild type and mutant proteins were synthesized in Escherichia coli, purified, and characterized. The rate constants for dissociation of GDP from G49V recombinant Gs alpha (rGs alpha) (0.47/min) and Q227L rGs alpha (0.23/min) differ by no more than 2-fold from that observed for the wild type protein (0.5/min). In marked contrast, the rate constants for hydrolysis of GTP by G49V rGs alpha (0.78/min) and Q227L rGs alpha (0.03-0.06/min) are 4-fold and roughly 100-fold slower than that for wild type rGs alpha (3.5/min). These reductions in the rate of hydrolysis of GTP result in significant fractional occupancy of these proteins by GTP in the presence of the nucleotide, 0.37 for G49V rGs alpha and 0.78 for Q227L rGs alpha, compared to 0.05 for wild type rGs alpha. When reconstituted with cyc- (Gs alpha-deficient) S49 cell membranes or purified adenylyl cyclase, both mutant proteins stimulate adenylyl cyclase activity in the presence of GTP to a much greater extent than does wild type Gs alpha; their maximal ability to activate the enzyme is largely unaltered. The fractional ability of a given Gs alpha polypeptide to active adenylyl cyclase in the presence of GTP correlates well with the fractinal occupancy of the protein by the nucleotide. The mutant subunits appear to interact normally with G protein beta gamma subunits, and their ability to activate adenylyl cyclase is enhanced by interaction with beta-adrenergic receptors. These results indicate that the structural analogy that has been inferred between the guanine nucleotide-binding domains of G proteins and the p21ras family is at least generally correct. They also provide confirmation of the kinetic model of G protein function and document mutations that permit the expression in vivo of constitutively activated G protein alpha subunits.     

Molecular cloning of five GTP-binding protein cDNA species from rat olfactory neuroepithelium

Biochemical studies in vertebrate olfactory tissue indicate that certain odorants stimulate adenylyl cyclase in a GTP-dependent manner. Additionally, immunochemical and toxin-labeling studies demonstrate the presence of several GTP-binding protein (G-protein) species in vertebrate olfactory epithelium. To identify the G-protein(s) responsible for olfactory signal transduction, we screened a rat olfactory cDNA library with an oligonucleotide probe and isolated 32 recombinant clones encoding five distinct types of G-protein alpha subunits. The majority of the clones encoded G alpha s, while the remaining clones encoded G alpha o, G alpha i1, G alpha i2, and a novel species, G alpha i3. Messenger RNA corresponding to each G alpha was detectable in all tissues examined; however, the levels for a given G alpha varied in a tissue-specific manner. In olfactory tissue, G alpha s was the most abundant of these messages and in combination with the biochemical studies suggests that G alpha s is the G-protein component of the olfactory signal transduction cascade.     

Mutated alpha subunit of the Gq protein induces malignant transformation in NIH 3T3 cells.

The discovery of mutated, GTPase-deficient alpha subunits of Gs or Gi2 in certain human endocrine tumors has suggested that heterotrimeric G proteins play a role in the oncogenic process. Expression of these altered forms of G alpha s or G alpha i2 proteins in rodent fibroblasts activates or inhibits endogenous adenylyl cyclase, respectively, and causes certain alterations in cell growth. However, it is not clear whether growth abnormalities result from altered cyclic AMP synthesis. In the present study, we asked whether a recently discovered family of G proteins, Gq, which does not affect adenylyl cyclase activity, but instead mediates the activation of phosphatidylinositol-specific phospholipase C harbors transforming potential. We mutated the cDNA for the alpha subunit of murine Gq in codons corresponding to a region involved in binding and hydrolysis of GTP. Similar mutations unmask the transforming potential of p21ras or activate the alpha subunits of Gs or Gi2. Our results show that when expressed in NIH 3T3 cells, activating mutations convert G alpha q into a dominant acting oncogene.     

The GppNHp-activated adenylyl cyclase complex from turkey erythrocyte membranes can be isolated with its beta gamma subunits.

The adenylyl cyclase complex, derived from turkey erythrocyte membranes, was activated using guanosine 5'-[beta, gamma-imido]triphosphate (Gpp[NH]p) and separated under low-detergent and low-salt conditions using conventional molecular-sieve chromatography followed by high-pressure ion-exchange and molecular-sieve chromatography. Although the complex remains activated with Gpp[NH]p throughout the isolation, the beta gamma subunits copurify with the cyclase. The stoichiometry of the cyclase to the alpha subunit of the stimulatory guanosine-nucleotide-binding regulatory protein (alpha s) to the beta subunit is close to unity, demonstrating that the beta gamma subunits do not dissociate from the Gs.cyclase complex (Gs, guanosine-nucleotide-binding regulatory protein) upon activation of the enzyme. If the final purification step was performed at high-salt concentrations, the beta gamma subunits could be separated from the alpha s.cyclase complex. Previously reported results on bovine brain cyclase also showed that the Gs.cyclase complex remains intact subsequent to activation by hormone and Gpp[NH]p [Marbach, I., Bar-Sinai, A., Minich, M. and Levitzki, A. (1990) J. Biol. Chem. 265, 9999-10,004]. These results, using adenylyl cyclase from two different sources, support our previous kinetic experiments which first suggested that beta gamma subunits are not released from Gs upon cyclase activation. We, therefore, argue that the mode of adenylyl cyclase inhibition by the inhibitory guanosine-nucleotide-binding regulatory protein cannot be via shifting the alpha s to beta gamma equilibrium as is commonly believed, and an alternate hypothesis is proposed.     

The thyroliberin receptor interacts directly with a stimulatory guanine-nucleotide-binding protein in the activation of adenylyl cyclase in GH3 rat pituitary tumour cells. Evidence obtained by the use of antisense RNA inhibition and immunoblocking of the stimulatory guanine-nucleotide-binding protein.

The thyroliberin receptor in GH3 pituitary tumour cells is known to couple to phospholipase C via a guanine-nucleotide-binding protein (G protein). Thyroliberin is postulated also to activate adenylyl cyclase, via the stimulatory G protein (Gs). In order to study this coupling, we constructed an antisense RNA expression vector that contained part of the Gs alpha-subunit cDNA clone (Gs alpha) in an inverted orientation relative to the mouse metallothionein promoter. The cDNA fragment included part of the coding region and all of the 3' non-translated region. Transient expression of Gs alpha antisense RNA in GH3 cells resulted in the specific decrease of Gs alpha mRNA levels, followed by decreased Gs alpha protein levels. Thyroliberin-elicited adenylyl cyclase activation in membrane preparations showed a reduction of up to 85%, whereas phospholipase C stimulation remained unaffected. Activation of adenylyl cyclase by vasoactive intestinal peptide was reduced by 30-40%. Investigation of the effects of thyroliberin and vasoactive intestinal peptide on adenylyl cyclase in GH3 cell membranes pretreated with antisera against Gs alpha and Gi-1 alpha/Gi-2 alpha support the results obtained by the use of the antisense technique. We conclude that thyroliberin has a bifunctional effect on GH3 cells, in activating adenylyl cyclase via Gs or a Gs-like protein in addition to the coupling to phospholipase C.     

S111N mutation in the helical domain of human Gs(alpha) reduces its GDP/GTP exchange rate

G-protein alpha subunits consist of two domains: a Ras-like domain also called GTPase domain (GTPaseD), structurally homologous to monomeric G-proteins, and a more divergent domain, unique to heterotrimeric G-proteins, called helical domain (HD). G-protein activation, requires the exchange of bound GDP for GTP, and since the guanine nucleotide is buried in a deep cleft between both domains, it has been postulated that activation may involve a conformational change that will allow the opening of this cleft. Therefore, it has been proposed, that interdomain interactions are playing an important role in regulating the nucleotide exchange rate of the alpha subunit. While constructing different Gs(alpha) quimeras, we identified a Gs(alpha) random mutant, which was very inefficient in stimulating adenylyl cyclase activity. The introduced mutation corresponded to the substitution of Ser(111) for Asn (S111N), located in the carboxi terminal end of helix A of the HD, a region neither involved in AC interaction nor in the interdomain interface. In order to characterize this mutant, we expressed it in bacteria, purified it by niquel-agarose chromatography, and studied its nucleotide exchange properties. We demonstrated that the recombinant S111N Gs(alpha) was functional since it was able to undergo the characteristic conformational change upon GTP binding, detected by the acquisition of a trypsin-resistant conformation. When the biochemical properties were determined, the mutant protein exhibited a reduced GDP dissociation kinetics and as a consequence a slower GTPgammaS binding rate that was responsible for a diminished adenylyl cyclase activation when GTPgammaS was used as activator. These data provide new evidence that involves the HD as a regulator of Gs(alpha) function, in this case the alphaA helix, which is not directly involved with the nucleotide binding site nor the interdomain interface.     

The NH2-terminal alpha subunit attenuator domain confers regulation of G protein activation by beta gamma complexes.

Gs and Gi, respectively, activate and inhibit the enzyme adenylyl cyclase. Regulation of adenylyl cyclase by the heterotrimeric Gs and Gi proteins requires the dissociation of GDP and binding of GTP to the alpha s or alpha i subunit. The beta gamma subunit complex of Gs and Gi functions, in part, to inhibit GDP dissociation and alpha subunit activation by GTP. Multiple beta and gamma polypeptides are expressed in different cell types, but the functional significance for this heterogeneity is unclear. The beta gamma complex from retinal rod outer segments (beta gamma t) has been shown to discriminate between alpha i and alpha s subunits (Helman et al: Eur J Biochem 169:431-439, 1987). beta gamma t efficiently interacts with alpha i-like G protein subunits, but poorly recognizes the alpha s subunit. beta gamma t was, therefore, used to define regions of the alpha i subunit polypeptide that conferred selective regulation compared to the alpha s polypeptide. A series of alpha subunit chimeras having NH2-terminal alpha i and COOH-terminal alpha s sequences were characterized for their regulation by beta gamma t, measured by the kinetics of GTP gamma S activation of adenylyl cyclase. A 122 amino acid NH2-terminal region of the alpha i polypeptide encoded within an alpha i/alpha s chimera was sufficient for beta gamma t to discriminate the chimera from alpha s. A shorter 54 amino acid alpha i sequence substituted for the corresponding NH2-terminal region of alpha s was insufficient to support the alpha i-like interaction with beta gamma t. The findings are consistent with our previous observation (Osawa et al: Cell 63:697-706, 1990) that a region in the NH2-terminal moiety functions as an attenuator domain controlling GDP dissociation and GTP activation of the alpha subunit polypeptide and that the attenuator domain is involved in functional recognition and regulation by beta gamma complexes.     

The alpha subunit of the stimulatory guanine-nucleotide-binding regulatory protein forms high-molecular-mass aggregates, concomitant with iloprost-induced desensitization of human platelet adenylyl cyclase.

Prolonged treatment of human platelets with the prostacyclin analog iloprost led to desensitization of the response to various prostaglandin derivatives. However, basal adenylyl cyclase activity and stimulation by agents acting directly via Gs, the stimulatory guanine-nucleotide-binding regulatory protein of adenylyl cyclase, were likewise decreased. Reconstitution of desensitized membranes with purified Gs from turkey erythrocytes indicated no alteration in the catalyst itself. However, the function of Gs (in cholate extracts) appeared to be severely impaired when reconstituted with adenylyl cyclase catalyst. Modification of Gs was also indicated by its altered sedimentation in sucrose density gradients. From Western blots, the alpha subunit of Gs, alpha s, from control platelets sedimented as a 5.6S species, while that from desensitized cells appeared at higher S values (in a polydisperse distribution). Activation by guanosine 5'-[gamma-thio]triphosphate of Gs from control platelets shifted alpha s to 3.5-3.7S, while activation of Gs from desensitized platelets induced such shift only for a minor portion of alpha s. This small fraction alone appeared to be susceptible to ADP-ribosylation by cholera toxin/[32P]NAD. Furthermore, an antibody directed against the C-terminal hexadecapeptide of alpha s precipitated much less alpha s from cholate extracts derived from desensitized platelets. Modification of alpha s during desensitization was also suggested from cross-linking experiments using the homobifunctional agent bismaleimidohexane: alpha s from desensitized platelets formed a single product of 80 kDa, while that from untreated platelets yielded a doublet (100 kDa and 110 kDa).     

Expression of Gs alpha in Escherichia coli. Purification and properties of two forms of the protein

Cloning of complementary DNAs that encode either of two forms of the alpha subunit of the guanine nucleotide-binding regulatory protein (Gs) that stimulates adenylyl cyclase into appropriate plasmid vectors has allowed these proteins to be synthesized in Escherichia coli (Graziano, M.P., Casey, P.J., and Gilman, A.G. (1987) J. Biol. Chem. 262, 11375-11381). A rapid procedure for purification of milligram quantities of these proteins is described. As expressed in E. coli, both forms of Gs alpha (apparent molecular weights of 45,000 and 52,000) bind guanosine 5'-(3-O-thio)triphosphate stoichiometrically. The proteins also hydrolyze GTP, although at different rates (i.e. 0.13.min-1 and 0.34.min-1 at 20 degrees C for the 45- and the 52-kDa forms, respectively). These rates reflect differences in the rate of dissociation of GDP from the two proteins. Both forms of recombinant Gs alpha have essentially the same kcat for GTP hydrolysis, approximately 4.min-1. Recombinant Gs alpha interacts functionally with G protein beta gamma subunits and with beta-adrenergic receptors. The proteins can also be ADP-ribosylated stoichiometrically by cholera toxin. This reaction requires the addition of beta gamma subunits. Both forms of recombinant Gs alpha can reconstitute GTP-, isoproterenol + GTP-, guanosine 5'-(3-O-thio)triphosphate-, and fluoride-stimulated adenylyl cyclase activity in S49 cyc- membranes to maximal levels, although their specific activities for this reaction are lower than that observed for Gs purified from rabbit liver. Experiments with purified bovine brain adenylyl cyclase indicate that the affinity of recombinant Gs alpha for adenylyl cyclase is 5-10 times lower than that of liver Gs under these assay conditions; however, the intrinsic capacity of the recombinant protein to activate adenylyl cyclase is normal. These findings suggest that Gs alpha, when synthesized in E. coli, may fail to undergo a posttranslational modification that is crucial for high affinity interaction of the G protein with adenylyl cyclase.     

The carboxy-terminal domain of Gs alpha is necessary for anchorage of the activated form in the plasma membrane

GTP-binding proteins which participate in signal transduction share a common heterotrimeric structure of the alpha beta gamma-type. In the activated state, the alpha subunit dissociates from the beta gamma complex but remains anchored in the membrane. The alpha subunits of several GTP-binding proteins, such as Go and Gi, are myristoylated at the amino terminus (Buss, J. E., S. M. Mumby, P. J. Casey, A. G. Gilman, and B. M. Sefton. 1987. Proc. Natl. Acad. Sci. USA. 84:7493-7497). This hydrophobic modification is crucial for their membrane attachment. The absence of fatty acid on the alpha subunit of Gs (Gs alpha), the protein involved in adenylate cyclase activation, suggests a different mode of anchorage. To characterize the anchoring domain of Gs alpha, we used a reconstitution model in which posttranslational addition of in vitro-translated Gs alpha to cyc- membranes (obtained from a mutant of S49 cell line which does not express Gs alpha) restores the coupling between the beta-adrenergic receptor and adenylate cyclase. The consequence of deletions generated by proteolytic removal of amino acid sequences or introduced by genetic removal of coding sequences was determined by analyzing membrane association of the proteolyzed or mutated alpha chains. Proteolytic removal of a 9-kD amino-terminal domain or genetic deletion of 28 amino-terminal amino acids did not modify the anchorage of Gs alpha whereas proteolytic removal of a 1-kD carboxyterminal domain abolished membrane interaction. Thus, in contrast to the myristoylated alpha subunits which are tethered through their amino terminus, the carboxy-terminal residues of Gs alpha are required for association of this protein with the membrane.     

Developmental expression of G proteins that differentially modulate adenylyl cyclase activity in mouse brain.

Changes in the relative abundance of the G protein alpha subunits were observed during early mouse development Gs alpha was almost exclusively present as a large form (Gs-1) in prenatal brain. Postnatally with a substantial increase in Gpp[NH]p stimulated adenylyl cyclase activity, the small form (Gs.s) increased in amount while Gs-1 decreased. These results suggest that the Gs-s may be the more effective cyclase activator and that changes in alternative splicing are developmentally regulated. Gi1 and Go appeared before birth whereas Gi2 developed postnatally. Opiate stimulation of GTPase and inhibition of adenylyl cyclase were fully expressed prenatally.     

Tubulin binds specifically to the signal-transducing proteins, Gs alpha and Gi alpha 1

Participation of cytoskeletal elements in regulation of hormonal response and responsiveness has been suggested by several laboratories. Addition of dimeric tubulin to rat cerebral cortex synaptic membranes causes stable inhibition of adenylyl cyclase, and the molecular basis for this effect appears to require a direct interaction between tubulin and G proteins. To test whether such tubulin-G protein interaction occurred, several purified G proteins were bound to nitrocellulose, and 125I-tubulin overlay studies were performed. 125I-Tubulin bound to the alpha subunits of Gs and Gil with high specificity and an apparent Kd of approximately 130 nM. Other G protein alpha subunits (alpha i2, alpha i3, alpha 0, and transducin) displayed a much lower affinity for tubulin, despite the much closer relationship of those proteins to alpha il than to alpha s. Association of beta gamma subunits with alpha il or alpha s did not alter the binding of tubulin to these G protein heterotrimers, and the binding of a hydrolysis-resistant GTP analog to the alpha subunits was similarly without effect. These results suggest that tubulin forms complexes with specific G proteins and these complexes might provide a locus for the interaction of cytoskeletal components and signal transduction cascades. These results also provide evidence of a functional distinction among the closely related alpha i subtypes.     

Beta subunit copurifies with GppNHp-activated adenylyl cyclase

Previous kinetic studies (Tolkovsky, A.M., Braun, S., and Levitzki, A. (1982) Proc. Natl. Acad. Sci. U. S.A. 79, 213-222) and biochemical studies (Arad, H., Rosenbusch, J., and Levitzki, A. (1984) Proc. Natl. Acad. Sci. U.S.A. 81, 6579-6583) from our laboratory suggest that Gs or alpha s remain associated with the catalytic subunit of adenylyl cyclase (C) throughout the activation cycle of adenylyl cyclase by hormone receptors. In this study we have purified GppNHp-activated bovine brain adenylyl cyclase over 3000-fold under mild solution conditions. We demonstrate that although the enzyme is permanently activated it retains the beta subunit when bound to a forskolin-agarose affinity column as long as it is not exposed to high salt concentrations. The stoichiometry of alpha s to beta to C is close to unity, suggesting that beta gamma subunits do not dissociate from Gs upon its activation. The complex gamma beta alpha s (GppNHp). C dissociates partially when migrating on a Superose 12 fast protein liquid chromatography molecular-seiving column. This partial dissociation probably results from the relatively diluted state of the enzyme at a high degree of purity. Prolonged ultracentrifugation of the complex also causes partial dissociation of the beta gamma subunits from alpha s (GppNHp). C. The apparent contradiction between the results reported here and the observation that beta gamma subunits inhibit cyclase activity when added to platelet membranes (Katada, T., Bokoch, G. M., Northrup, J. K., Ui, M., and Gilman, A. G. (1984a) J. Biol. Chem. 259, 3568-3577) is discussed. We suggest an alternative model to account for this inhibitory effect of added beta gamma subunits.