The regulatory component of adenylate cyclase. Purification and properties

The regulatory component (G/F) of adenylate cyclase, which has been purified previously, contains three putative subunits with molecular weights of 52,000, 45,000, and 35,000 (Northup, J. K., Sternweis, P. C., Smigel, M. D., Schleifer, L. S., Ross, E. M., and Gilman, A. G. (1980) Proc. Natl. Acad. Sci. U. S. A. 77, 6516-6520). The published procedure has been modified to reduce the time required for preparation and to increase the yield. Application of the improved procedure allows purification of .5 to 1.0 mg of purified G/F from 1.5 kg of frozen rabbit liver. Greater than 95% of the protein observed on sodium dodecyl sulfate polyacrylamide gels is found in the three bands mentioned above. Purified G/F has the following properties: 1. Hydrodynamic measurements in cholate indicate that purified hepatic G/F has a molecular weight of about 70,000. If G/F is activated with either fluoride or GTP analogs, its apparent molecular weight is reduced to 50,000. 2. The measurement of G/F by reconstitution with the catalytic moiety of adenylate cyclase is dependent on the concentrations of both G/F and catalytic moiety. This interaction is consistent with a model derived from a simple bimolecular binding equilibrium. 3. Purified G/F can be activated by fluoride and guanine nucleotide analogs in a Mg2+-dependent reaction. The rate of activation by guanine nucleotides is markedly stimulated by high concentrations of Mg2+, indicating a site of action of divalent metallic cations on G/F. 4. The 52,000- and 45,000-dalton polypeptides can be partially resolved by heptylamine-Sepharose chromatography. G/F fractions that are enriched in the 52,000-dalton protein reconstitute hormone-stimulated adenylate cyclase activity more efficiently and are activated by GTP analogs more rapidly than are fractions that are essentially free of this polypeptide. The 35,000-dalton protein is present in all cases.     

The regulatory GTPase cycle of turkey erythrocyte adenylate cyclase.

It has recently been suggested that adenylate cyclase activity is controlled by a regulatory cycle consisting of two reactions: a hormone induced formation of the active adenylate cyclase-GTP complex, and a subsequent turn-off reaction in which hydrolysis of the bound nucleotide reverts the system to the inactive state. To test this model each of the two reactions was measured separately and their rate constants were used to estimate the steady state adenylate cyclase and GTPase activities. The first order rate constants were kon = 3 min-1 for the activation reaction and koff = 15 min-1 for the turn-off reaction. Substitution of these rate constants in the steady state equation of the regulatory cycle gave values of hormone stimulated adenylate cyclase and GTPase activities similar to those determined by direct measurements. Treatment of the adenylate cyclase with cholera toxin caused a decrease of 96% in the rate constant of the turn-off reaction. In this case too the activities calculated from the steady state equation were in good agreement with those determined directly.     

Cholera-toxin-dependent ADP-ribosylation of the adenylate cyclase regulatory protein in turkey erythrocyte membranes

Incubation of turkey erythrocyte membranes with cholera toxin and [³²P]NAD caused toxin-dependent incorporation of ³²P into a 42,000 M_r peptide which could be distinguished from toxin-independent ³²P incorporation into other membrane proteins. The radiolabeled 42,000 M_r peptide could be extracted from the membranes using Lubrol PX. When toxin-treated membranes were incubated with isoproterenol and GMP before detergent solubilization, the 42,000 M_r labeled peptide was adsorbed by GTP-γ-agarose which, with the same conditions, adsorbed the adenylate cyclase guanine nucleotide regulatory protein. The labeled peptide and guanine nucleotide regulatory protein activity were coeluted from the affinity matrix by guanylyl-β,γ-imidodiphosphate, GDP, and GMP. Guanosine 5′-O-(2-thiodiphosphate), an analog of GDP which blocks guanine nucleotide- and fluoride-stimulated adenylate cyclase activity, caused elution of labeled peptide which exhibited no regulatory protein activity. Our data support the view that the 42,000 M_r peptide is part of the adenylate cyclase guanine nucleotide regulatory protein. The labeled peptide allows identification of both active and inactive regulatory protein and should be useful in monitoring the purification of the regulatory protein from turkey erythrocytes.     

Hydrodynamic properties of the regulatory component of adenylate cyclase

Hydrodynamic parameters of the regulatory component of adenylate cyclase, G/F, have been estimated by gel filtration and sucrose density gradient centrifugation. In solutions containing Lubrol 12A9, the protein has an apparent molecular weight of 130,000. G/F from various sources and resolved from the catalytic moiety of the enzyme by different techniques behaves similarly. Consistent with our previous proposal that this protein is the site of action of both guanine nucleotides and fluoride, treatment with these activating ligands causes a reduction in both the sedimentation coefficient and the Stokes radius of G/F. These changes suggest a loss of mass of approximately 40,000 daltons. Nevertheless, this alteration is fully reversible when ligands are removed, even if the liganded protein is first fractionated by gel filtration or sucrose density gradient centrifugation.     

The control of adenylate cyclase by calcium in turkey erythrocyte ghosts.

The adenylate cyclase of turkey erythrocytes is inhibited by low concentrations of calcium. Calcium binds to the enzyme system so tightly that the enzyme can compete with ethylene glycol bis(beta-aminoethyl ether)-N, N1-tetraacetic acid (EGTA) for the metal. The calcium binding site is shown to be distinct from the magnesium binding sites required for activity. Thus Ca2+ functions as a negative allosteric effector. Calcium decreases dramatically the V max of the catecholamine-stimulated activity without affecting the affinity for the hormone or for the substrate ATP. The cooperativity in the response toward Mg2+ dependence (Hill coefficient, nH equals 3) is also unaffected by Ca2+ where as the S0.5 (concentration yielding one-half V max) for Mg2+ is affected only slightly. The Ca2+ effect is cooperative (nH equals 2) and therefore brought about by a cluster of Ca2+ binding sites. Mn2+ can substitute for Mg2+ as the enzyme activator but the Mn2+-activated enzyme is no longer inhibited by Ca2+. The possible physiological significance of the Ca2+ effect is discussed.     

Adenosine receptor permanently coupled to turkey erythrocyte adenylate cyclase.

The mode of coupling of the adenosine receptor to adenylate cyclase in turkey erythrocyte membranes was probed by two independent approaches. The progressive inactivation of the adenosine receptor by an adenosine receptor affinity label resulted in the proportional reduction in the adenosine plus GppNHp dependent specific activity. In contrast, the intrinsic rate constant (k3), characterizing the process of adenylate cyclase activation by the adenosine-adenosine receptor complex, is independent of the extent of receptor inactivation. This behavior favors the precoupled mechanism, A + R.E: formula: (see text), where the receptor R and the enzyme E are permanently coupled to each other and the adenosine A binds to the receptor and induces the first-order process of cyclase activation to its active form ARE'. The finding that adenosine receptor is permanently coupled to the cyclase catalytic unit is corroborated by the observation that the progressive increase in membrane fluidity has no effect on the rate constant (k3) of adenylate cyclase activation by the adenosine-adenosine receptor complex and that the dose-response curve for adenosine is noncooperative.     

Characteristics of the guanine nucleotide regulatory component of adenylate cyclase in human erythrocyte membranes

This study probes the structure and mutual interactions of the components of adenylate cyclase. We use a complementation assay which involves the addition of an adenylate cyclase-related guanine nucleotide-binding protein component to a membrane lacking this component to measure guanine nucleotide-stimulated-adenylate cyclase. Instead of using detergent extracts we were able to achieve full complementation by mixing intact membrane preparations in the presence of the nucleotide component. Of particular interest was the human erythrocyte membrane which contains very low amounts of catalytic activity and no measurable beta-adrenergic receptor but has normal amounts of the nucleotide component. This component appears to be the same, by several criteria, as components found in pigeon and turkey erythrocytes and in rat liver plasma membrane. The component confers Gpp(NH)p, fluoride, and GTP stimulation of adenylate cyclase along a single reconstitution curve. It is labeled with NAD by cholera toxin, and has an apparent molecular weight of 39 000 upon sodium dodecyl sulfate gel electrophoresis. The presence of the nucleotide unit in the virtual absence of the active catalytic unit allowed us to determine those properties intrinsic to each unit and those conferred by the association of the units. The nucleotide component binds guanine nucleotides weakly in the human erythrocyte membrane, yet produces persistent activation of adenylate cyclase and tight binding (of Gpp(NH)p) upon combination with the catalytic unit. Treatment of the human erythrocyte membrane with N-ethylmaleimide causes a simultaneous diminution in both Gpp(NH)p and fluoride stimulation in reconstituted activities, suggesting that both activities are conferred by the same component.     

Desensitization of turkey erythrocyte adenylate cyclase. Beta-adrenergic receptor phosphorylation is correlated with attenuation of adenylate cyclase activity

Preincubation of turkey erythrocytes with beta-adrenergic agonists leads to an attenuation of the responsiveness of adenylate cyclase to subsequent hormonal stimulation. Recently, our laboratory has shown (Stadel, J. M., Nambi, P., Shorr, R. G. L., Sawyer, D. D., Caron, M. G., and Lefkowitz, R. J. (1983) Proc. Natl. Acad. Sci. U. S. A. 80, 3173-3177) using 32Pi incorporation that phosphorylation of the beta-adrenergic receptor accompanies this desensitization process. We now report that, as determined from intracellular [gamma-32P] ATP specific activity measurements, this phosphorylation reaction occurs in a stoichiometric fashion. Under basal conditions there exists 0.75 +/- 0.1 mol of phosphate per mol of receptor whereas under maximally desensitized conditions this ratio increases to 2.34 +/- 0.13 mol/mol. This phosphorylation of the receptor is dose-dependent with respect to isoproterenol and exhibits a dose-response curve coincidental with that for isoproterenol-induced desensitization of adenylate cyclase. The time courses for receptor phosphorylation and adenylate cyclase desensitization are identical. In addition, the rate of resensitization of adenylate cyclase activity is comparable to the rate of return of the phosphate/receptor stoichiometries to control levels. Both the phosphorylation and desensitization reactions are pharmacologically specific as indicated by the high degree of stereoselectivity, rank order of catecholamines, and blockade by the specific beta-adrenergic antagonist, propranolol. Incubation of turkey erythrocytes with cAMP and cAMP analogs maximally activates cAMP-dependent protein kinase but only partially mimics isoproterenol in promoting phosphorylation of the receptor in concordance with their partial effects in inducing desensitization. Conversely, activators or inhibitors of Ca2+/calmodulin kinase or protein kinase C do not affect the isoproterenol-induced desensitization. These results indicate that desensitization of turkey erythrocyte adenylate cyclase is highly correlated with phosphorylation of the beta-adrenergic receptor and that these events are mediated, at least partially, by cAMP.     

The inhibitory guanine nucleotide-binding regulatory component of adenylate cyclase. Properties and function of the purified protein

Treatment of membranes with islet activating protein (IAP), a toxin from Bordetella pertussis, results in abolition of GTP-dependent, receptor-mediated inhibition of adenylate cyclase. This appears to result from IAP-catalyzed ADP-ribosylation of a 41,000-Da membrane-bound protein. A protein with 41,000- and 35,000-Da subunits has been purified from rabbit liver membranes as the predominant substrate for IAP. This protein has now been shown to be capable of regulating membrane-bound adenylate cyclase activity of human platelets under various conditions. The characteristics of the actions of the IAP substrate are as follows. 1) Purified 41,000/35,000-Da dimer is capable of restoring the inhibitory effects of guanine nucleotides and the alpha 2-adrenergic agonist, epinephrine, on the adenylate cyclase activity of IAP-treated membranes. 2) The subunits of the dimer dissociate in the presence of guanine nucleotide analogs or A1(3+), Mg2+, and F-. The 41,000-Da subunit has a high affinity binding site for guanine nucleotides. 3) The resolved 35,000-Da subunit of the dimer mimics guanine nucleotide- and epinephrine-induced inhibition of adenylate cyclase. 4) The resolved (unliganded) 41,000-Da subunit stimulates adenylate cyclase activity and relieves guanine nucleotide- +/- epinephrine-induced inhibition of the enzyme. In contrast, the GTP gamma S-bound form of the 41,000-Da subunit inhibits adenylate cyclase activity, although with lower apparent affinity than does the 35,000-Da subunit. 5) The 35,000-Da subunit increases the rate of deactivation of Gs, the stimulatory regulatory protein of adenylate cyclase. In contrast, the 41,000-Da subunit can interact with Gs and inhibit its deactivation. These data strongly suggest that the IAP substrate is another dimeric, guanine nucleotide-binding regulatory protein and that it is responsible for inhibitory modulation of adenylate cyclase activity.     

The regulation of adenylate cyclase activity in turkey erythrocyte membranes by forskolin

The mechanisms by which forskolin stimulates adenylate cyclase activity in turkey erythrocyte membranes and is influenced by manganese and Gpp(NH)p were studied. Forskolin-dependent adenylate cyclase activity in particulate turkey erythrocyte membranes is enhanced following preincubation of membranes with isoproterenol and GMP (cleared membranes). In contrast, solubilization of turkey erythrocyte membranes, previously cleared, renders them relatively refractory to forskolin but not to Gpp(NH)p. Whereas adenylate cyclase activity due to the simultaneous presence of forskolin and Mn2+ in particulate turkey erythrocyte membranes is additive, their copresence becomes synergistic after solubilization. The apparent Kact for forskolin activation of adenylate cyclase is not influenced by clearance or by the presence of Mn2+ in particulate turkey erythrocyte membranes. Following solubilization, the Vmax for forskolin-dependent adenylate cyclase activation determined in the presence of Mn2+ is also independent of clearance. Forskolin activation of turkey erythrocyte adenylate cyclase appears to be influenced at sites in addition to the catalytic unit.     

A reconstitution assay for the guanine nucleotide regulatory protein of the adenylate cyclase system using turkey erythrocyte membranes as the acceptor preparation

The turkey erythrocyte beta-adrenergic receptor-adenylate cyclase system has the unusual property that neither GTP nor Gpp(NH)p are effective in activating adenylate cyclase unless a beta-agonist is present simultaneously. This property results in essentially no basal activity and the inability of GTP or Gpp(NH)p alone to activate the catalytic moiety. In this study, we have exploited these characteristics to utilize turkey erythrocyte membranes as the acceptor preparation in a reconstitution assay. Rat reticulocyte or turkey erythrocyte membranes that have been activated with isoproterenol and Gpp(NH)p followed by solubilization with sodium cholate serve as the donor source of the guanine nucleotide regulatory protein (N). By reconstituting this Gpp(NH)p-activated N protein, it has been found that: (1) exogenous Gpp(NH)p-associated N could activate the catalytic unit of adenylate cyclase in turkey erythrocyte membranes; (2) this system can be used to assay N protein activity; (3) the endogenous pathway for activation of turkey erythrocyte membrane adenylate cyclase by hormones and fluoride remains qualitatively functional; and (4) the effects of combined activation via the endogenous and exogenous pathways are additive and saturable.     

Purification of a protein cofactor required for ADP-ribosylation of the stimulatory regulatory component of adenylate cyclase by cholera toxin

A factor (ARF) that is required for the cholera toxin-dependent ADP-ribosylation of the stimulatory, GTP-binding regulatory component (Gs) of adenylate cyclase has been purified about 2000-fold from cholate extracts of rabbit liver membranes. ARF is an intrinsic membrane protein with Mr = 21,000. The final product can be resolved into two polypeptides with very similar molecular weights; each of these has ARF activity. The ADP-ribosylation of Gs can now be studied with defined components. GTP and ARF are both necessary cofactors. The data imply that the substrates for the activated toxin are NAD and a GTP X Gs X ARF complex, and the reaction proceeds in a lipid environment. The apparent ability of ARF to bind to the alpha subunit of Gs suggests that it may play another, unknown role in the regulation of adenylate cyclase activity.     

Purification and properties of the inhibitory guanine nucleotide regulatory unit of brain adenylate cyclase

Hormonal inhibition of adenylate cyclase is mediated by a guanine nucleotide regulatory protein (Ni) which is different from the one which mediates hormonal stimulation. There is substantial evidence that the active component of Ni (termed alpha i can be ADP-ribosylated by a toxin from Bordetella pertussis. We have found that in bovine cerebral cortex there are three proteins of similar molecular weight (39,000-41,000) which are modified by pertussis toxin. We have purified these proteins and have resolved the 41,000-dalton protein from the 40,000/39,000-dalton doublet. All three forms of pertussis toxin substrate can be isolated in free form or together with a 36,000 beta component. We have also purified this beta component. ADP-ribosylation of the three pertussis toxin substrates is greatly enhanced by the addition of the purified beta component. This makes possible an assay of beta subunit activity based on its interaction with alpha i. The three forms of pertussis toxin substrate which we have purified differ in two functions: susceptibility to ADP-ribosylation and GTPase activity. The 41,000-dalton protein is more readily ADP-ribosylated by pertussis toxin than the smaller forms. The 39,000-dalton protein has GTPase activity with a low Km (0.3 microM) for GTP. The GTPase activity can be doubled by phospholipids. The GTPase activity of the 41,000-dalton protein is almost undetectable. It is not yet known what the relationship of the forms is to each other. The smaller forms may be derived from the larger by proteolysis or it may be intrinsically different. It remains to be shown whether one of the forms represents a different type of regulatory protein which transmits a hormonal signal to effectors other than adenylate cyclase.     

The guanine nucleotide-binding regulatory component of adenylate cyclase in human erythrocytes

The guanine nucleotide-binding regulatory component of adenylate cyclase (G/F) has been purified from human erythrocyte membranes. It is composed of two major polypeptides with molecular weights of 35,000 and 45,000. When cyc- S49 lymphoma cell plasma membranes are reconstituted with purified human erythrocyte G/F, stimulation of adenylate cyclase by beta-adrenergic agonists, guanine nucleotides, and fluoride is restored. Binding of GTP gamma S to human erythrocyte G/F and GTP gamma S-mediated activation of the protein are closely correlated. The agreement between the apparent dissociation constants for these two reactions suggests that the measured binding site is identical to the site responsible for activation. A 41,000-dalton protein has been identified as a contaminant of preparations of G/F that have been purified by four successive chromatographic steps. This protein serves as a specific substrate for ADP-ribosylation and labeling by islet activating protein (IAP) and [32P]NAD, and it appears to contribute an additional high-affinity guanine nucleotide binding site to such preparations.     

Evidence that forskolin activates turkey erythrocyte adenylate cyclase through a noncatalytic site

The diterpene forskolin has been reported to activate adenylate cyclase in a manner consistent with an interaction at the catalytic unit. However, some of its actions are more consistent with an interaction at the coupling unit that links the hormone receptor to the adenylate cyclase activity. This report adds support to the latter possibility. Under conditions that lead to stimulation of adenylate cyclase in turkey erythrocyte membranes by GTP, forskolin also becomes more active. Additional evidence to support an influence of forskolin upon adenylate cyclase via the GTP-coupling protein N includes the following: (i) forskolin, at submaximal concentrations, leads to enhanced sensitivity and responsiveness of isoproterenol-dependent adenylate cyclase activity in turkey erythrocyte membranes; (ii) under specified conditions, the nucleotide GDP, an inhibitor of the stimulating nucleotide GTP and its analog, guanyl imidodiphosphate (Gpp(NH)p), also markedly inhibits the action of forskolin; (iii) both Gpp(NH)p and forskolin are associated with a decrease in agonist affinity for the beta-adrenergic receptor. However, actions of forskolin in the turkey erythrocyte are not identical to those of GTP: (i) forskolin is never as potent as Gpp(NH)p in activating adenylate cyclase; (ii) the magnitude of synergism between isoproterenol and forskolin is not equal to that observed with isoproterenol and Gpp(NH)p; (iii) at high concentrations, forskolin inhibits antagonist binding to the beta-receptor. Forskolin appears to have several sites of action in the turkey erythrocyte membrane, including an influence upon the adenylate cyclase regulatory protein N.     

Inhibition of hormone-stimulated adenylate cyclase activity after altering turkey erythrocyte phospholipid composition with a nonspecific lipid transfer protein. Phosphatidylinositol uncouples catecholamine binding from adenylate cyclase activation

The nonspecific lipid transfer protein from beef liver was used to modify the phospholipid composition of intact turkey erythrocytes in order to study the dependence of isoproterenol-stimulated adenylate cyclase activity on membrane phospholipid composition. Incorporation of phosphatidylinositol into turkey erythrocytes inhibited isoproterenol-stimulated cyclic AMP accumulation in a linear, concentration-dependent manner. Inhibition was relatively specific for phosphatidylinositol; phosphatidylcholine, phosphatidylethanolamine, phosphatidylserine, phosphatidylglycerol and phosphatidic acid were from 3 to 7 times less effective as inhibitors of hormone-stimulated cyclase activity. Inhibition by phosphatidylinositol was not reversible when up to 90% of the incorporated phosphatidylinositol was removed, either by incubation with phosphatidylinositol-specific phospholipase C or a second incubation with transfer protein; possibly adenylate cyclase activity depends on a small pool of phosphatidylinositol that is inaccessible to either phospholipase C hydrolysis or removal by lipid transfer protein. Phosphatidylinositol incorporation inhibits adenylate cyclase activity by uncoupling beta-adrenergic receptors from the remainder of the cyclase complex. Phosphatidylinositol incorporation had no effect on stimulation of cAMP accumulation by either cholera toxin or forskolin, indicating that inhibition occurs only at the level of receptor. Phosphodiesterase activity was not altered in phosphatidylinositol-modified cells. Inhibition of cAMP accumulation was not the result of changes in either membrane fluidity or in cAMP transport out of modified turkey erythrocytes. Phosphatidylinositol inhibition of isoproterenol-stimulated cyclase activity may serve as a useful model system for hormone-induced desensitization.     

Cell-free desensitization of catecholamine-sensitive adenylate cyclase. Agonist- and cAMP-promoted alterations in turkey erythrocyte beta-adrenergic receptors

Conditions have been developed for desensitizing the beta-adrenergic receptor-coupled adenylate cyclase of turkey erythrocytes in a cell-free system. Desensitization is observed when cell lysates are incubated with isoproterenol or cAMP analogs for 30 min at 37 degrees C. Maximally effective concentrations of isoproterenol produce a 41.0 +/- 1.55% loss of iosproterenol-stimulated and a 15.0 +/- 2.35% loss of fluoride-stimulated enzyme activity. cAMP causes a 26.5 +/- 1.5% fall in isoproterenol-stimulated and a 21.5 +/- 4.4% fall in fluoride-sensitive activity. Desensitization by isoproterenol is dose-dependent, stereospecific, and blocked by the beta-adrenergic antagonist propranolol. Cell-free desensitization required ATP, Mg2+, and factor(s) present in the soluble fraction of the cell. Nonphosphorylating analogs of ATP did not support desensitization. Desensitization by agonist or cAMP in the cell-free system caused structural alterations in the beta-adrenergic receptor peptides apparent as an altered mobility of the photoaffinity labeled receptor peptides on sodium dodecyl sulfate-polyacrylamide gel electrophoresis. As with the desensitization reaction, supernatant factors and ATP were also required for the agonist or cAMP-promoted receptor alterations. These data indicate that beta-adrenergic agonists promote a cAMP-mediated process which leads to receptor alterations and desensitization. The reactions involved in this process require ATP and soluble cellular factors. Additional processes must also occur to account for decreases in fluoride-sensitive enzyme activity. The availability of this cell-free system should facilitate elucidation of the molecular mechanisms involved in these processes.