Mechanism of Guanine Nucleotide Regulatory Protein-mediated Inhibition of Adenylate Cyclase STUDIES WITH ISOLATED SUBUNITS OF TRANSDUCIN IN A RECONSTITUTED SYSTEM*

The retinal nucleotide regulatory protein, transdu- cin, can substitute for the inhibitory guanine nucleo-tide-binding regulatory protein (Ni) in inhibiting ade- nylate cyclase activity in phospholipid vesicle systems. In the present work we have assessed the roles of the a (aT) and By (ByT) subunit components in mediating this inhibition. The inclusion of either a preactivated (YT.GTP~S (where GTPyS is guanosine 5’-0-(thiotri- phosphate)) complex, or the #?y complex, in phospholipid vesicles containing the pure human erythrocyte stimulatory guanine nucleotide-binding regulatory protein (N.) and the resolved catalytic moiety of bovine caudate adenylate cyclase (C) resulted in inhibition of the GppNHp-stimulated (where GppNHp is guanyl-5’-yl imidodiphosphate) activity (by -30-60 and 90%, respectively, at 2 mM MgC12). The inhibitions by both of these subunit species are specific for the N.-stimu- lated

The retinal nucleotide regulatory protein, transducin, can substitute for the inhibitory guanine nucleotide-binding regulatory protein (Ni) in inhibiting adenylate cyclase activity in phospholipid vesicle systems. In the present work we have assessed the roles of the a (aT) and By (ByT) subunit components in mediating this inhibition. The inclusion of either a preactivated (YT.GTP~S (where GTPyS is guanosine 5'-0-(thiotriphosphate)) complex, or the #?y complex, in phospholipid vesicles containing the pure human erythrocyte stimulatory guanine nucleotide-binding regulatory protein (N.) and the resolved catalytic moiety of bovine caudate adenylate cyclase (C) resulted in inhibition of the GppNHp-stimulated (where GppNHp is guanyl-5'yl imidodiphosphate) activity (by -30-60 and 90%, respectively, at 2 mM MgC12). The inhibitions by both of these subunit species are specific for the N.-stimulated activity with neither aT.GTPyS nor BYT having any direct effect on the intrinsic activity of the catalytic moiety. Increasing the MgClz concentration in the assay incubations significantly decreases the inhibitions by both aT.GTPyS and ByT. Similarly, when the pure hamster lung #?-adrenergic receptor is included in the lipid vesicles with N, and C, the levels of inhibition of the GppNHp-stimulated activity by both ~T * G T P~S and ByT are reduced compared to those obtained in vesicles containing just N. and C (but not stimulatory receptor). These inhibitions are reduced still further under conditions where the agonist stimulation of adenylate cyclase activity is maximal, Le. when stimulating with isoproterenol plus GTP. In these cases the aT.GTPyS inhibitory effects are completely eliminated and the inhibitions observed with holotransducin can be fully accounted for by the BYT complex. The ability of the #?-adrenergic receptor to relieve these inhibitions suggests that the receptor may remain coupled to N. (or as) during the activation of the regulatory protein and the stimulation of adenylate cyclase. These results also suggest that under physiological conditions the By subunit complex is primarily responsible for mediating the inhibition of adenylate cyclase activity.
* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertkement" in accordance with 18  Regulation of adenylate cyclase activity is accomplished through distinct stimulatory and inhibitory pathways. Both of these are initiated by agonist interactions with specific receptor proteins. These interactions enable the stimulatory receptors to promote the activation of a specific nucleotidebinding regulatory protein, designated as N, (1)(2)(3) or G, (4), and the inhibitory receptors to activate a distinct nucleotidebinding regulatory protein termed Ni ( 5 ) or Gi (6). These regulatory proteins are heterotrimers of composition olpy (7) and are structurally distinguishable by their a subunits (a, = 42,000 daltons and ai = 40,000 daltons) but not by their p (M, 35,000-36,000) (3, 8, 9) or y ( M , z 5,000-10,000) (9) subunits, which appear to be identical. An analogous type of system operates in vertebrate phototransduction. Here, the absorption of light by the photoreceptor rhodopsin allows it to interact with the heterotrimeric nucleotide regulatory protein transducin (M, aT = 39,000, p = 35,000, and YT s 9,000 (10, 11)). This interaction activates transducin and thus enables it to stimulate the activity of its biological effector, the cyclic GMP phosphodiesterase (10, 11).
The specific mechanisms by which receptors promote the activation of these regulatory proteins in a lipid milieu, as well as the mechanisms by which the activated regulatory proteins go on to stimulate or inhibit effector enzyme activity, have been the subject of much investigation. A particularly controversial question concerns the mechanism(s) of nucleotide regulatory protein-mediated inhibition of adenylate cyclase activity. Two quite different schemes have been suggested. One of these, based on hydrodynamic studies (7, 12-15), indicates that the activation of the regulatory proteins, in detergent solutions, results in their dissociation into a . guanine nucleotide and intact (3-y complexes. It can then be predicted that by increasing the levels of p-y in a cellular membrane, as a result of the activation of Ni, there would be a decrease in hormonal and guanine nucleotide-stimulated cyclic AMP production due to the increased deactivation of N. ( i e . through the reassociation of a. with Py). Support for this scheme has come from studies in platelet membranes and S49 lymphoma wild type membranes, where the addition of fly effects a dose-dependent inhibition of Ne-stimulated adenylate cyclase activity (16,17).
However, various studies using cyc-membranes suggest that the ai subunit is primarily responsible for Ni-mediated inhibition of adenylate cyclase. It has been well documented that inhibitory agonists like somatostatin (la), or even guanine nucleotides alone (2, 16,[19][20][21], can cause significant inhibition of the adenylate cyclase activity in these mutant membranes, which contain neither a functional N, (22,23) nor RNA sequences coding for it (24). Furthermore, the addition of activated resolved ai subunit complexed with GTPyS,' to both cyc-and wild type S49 cells, also leads to attenuation of activity (17). An examination of the inhibition by Ni of as. GTPyS-stimulated activity in cyc-membranes indicated clearly that this inhibition is noncompetitive in nature (21). These results cannot be explained by a By deactivation mechanism but rather suggest some type of direct interaction between Ni and the catalytic moiety of adenylate cyclase.
Recent achievements in purifying the individual stimulatory and inhibitory components of the adenylate cyclase system (3, 6, 25-29) have enabled us to study the mechanisms by which this enzyme is regulated using reconstituted systems. The advantage of such an approach is that it permits the study of interactions between each of these components in a lipid milieu under well defined conditions. Recently, we succeeded in reconstituting a guanine nucleotide-dependent inhibition of adenylate cyclase activity in phospholipid vesicles (30). An important conclusion of these studies was that the retinal nucleotide regulatory protein, transducin, can effectively substitute for Ni in mediating this inhibition. This finding offers an advantage for studying the inhibitory pathway of adenylate cyclase since both the a and Py subunits of the retinal regulatory protein can be readily isolated in high yield and in the complete absence of detergent (31). In the work described here, we have exploited this finding to study the roles of the isolated a and Pr subunits of transducin in mediating inhibition of adenylate cyclase activity in different types of phospholipid vesicle systems.
The catalytic moiety (C) of adenylate cyclase was solubilized from bovine caudate with sodium cholate and isolated from the other components of the system by Sepharose 6B chromatography, essentially using the procedures of Strittmatter and Neer (36). The specific activity of the resolved C preparations for the forskolin-stimulated activity was 20-40 nmol of cAMP/mg of protein/min (37). In these preparations C appeared to be effectively separated from the other functional components of the system N., Ni, and the P subunit since all of these activities were below detectable levels in the various assays performed. N. was assessed by reconstitution of adenylate cyclase stimulation in cyc-S49 membranes. N, (Le. ai) was detected by ADP-ribosylation using pertussis toxin in the presence of excess /3y subunit. The content of By subunit in these preparations was assessed by immunoblotting with an antibody (RV6) capable of recognizing the P subunit of N proteins (38). These adenylate cyclase preparations were stored (-90 "c) in 50 mM Tris-HC1, 200 mM sucrose, 1 mM DTT, 15 mM MgC12, 3.5 mg/ml crude soybean phosphatidylcholine, and 0.6% sodium cholate (pH 7.6) (buffer F).
Reconstitution of Adenylate Cyche Activity in Phospholipid Vesicles-The reconstitution of adenylate cyclase in the various phospholipid vesicle systems was performed as follows. Soybean phosphatidylcholine (100 p1 of a 17 mg/ml sonicated solution) was first incubated with 0.05-0.1 ml of the adenylate cyclase preparation (in buffer F and typically yielding 5-10 nmol of forskolin-stimulated CAMP production/30 min at 30 "C using previously described assay conditions (30)), 0.19-0.28 ml of 100 mM NaCl, 10 mM Tris-HC1 (pH 7.4), and 25 pl of octyl glucoside (17%) for 20-30 min on ice. At this point, when appropriate, N. (0.8 pg in 10 p1 of buffer A), rhodopsin (6.5 pg in 15 p1 of buffer C), and PAR (10-15 pmol in 40 p1 of buffer E) were added to the lipid solutions (incubated -1-2 min at 4 "C). The PAR (40 pl) was preincubated with 5 p~ alprenolol (10 pl) and 500 pg of bovine serum albumin (10 pl) prior to its addition to the above mixture. Rhodopsin was included in the vesicle preparations in order to ensure maximum activation of holotransducin (32) (thus, for comparative purposes, rhodopsin was present in all cases to be described). In all cases the final volume of the reconstitution incubation was 0.5 ml (i.e. when PAR was not present, an equivalent volume of buffer was added to the incubation).
Upon mixing the lipid solutions with N. and the receptor, the mixtures were applied to Extracti-Gel columns (1 ml of gel) at 4 "C which were pretreated with 4 volumes of 100 mM NaCl, 10 mM Tris-HCl (pH 7.4), 2 mg/ml BSA and then equilibrated with 100 mM NaCl, 10 mM Tris-HC1 (pH 7.4), 1 mM dithiothreitol and in some cases 15 mM MgCl,. The eluates from the Extracti-Gel (2 ml in 100 mM NaCl, 10 mM Tris-HC1 (pH 7.4), 1 mM dithiothreitol) were incubated with polyethylene glycol (6000-8000) for 5 min at room temperature and then isolated by ultracentrifugation at 4 "C as previously described (30,32). The resultant protein-lipid pellets were resuspended in 75 mM Tris-HC1 (pH 7.8), 1 mM DTT (final volume 0.6-1.3 ml) and either directly assayed for adenylate cyclase activity or treated with holotransducin or its subunit components prior to assaying for activity, as described below. The efficiency of reconstitution of PAR (as determined by ['251]iodocyanopindolol binding) typically ranged from 5 to 10%; the efficiency of reconstitution of N. ranged from 20 to 60% (as assessed by [35S]GTPyS binding), and the efficiency of reconstitution of adenylate cyclase ranged from 15 to 40% as assessed by forskolin-stimulated activity (37).

Examination of the Effects of (YT, P~T ,
and Hobtransduein on the Reconstituted Adenylute Cyclase Activity-In order to determine the effects of (YT, PYT, or holotransducin on the reconstituted adenylate cyclase activity, the resuspended lipid vesicles containing the different components, described in the preceding section, were divided into 0.18-0.35-m1 aliquots. Varying amounts of the pure subunit components (0.03-3 pg in 2-20 pl of buffer D), pure holotransducin (0.05-2 pg in 2-15 pl of buffer B), or just buffer were directly added to the vesicle aliquots prior to assaying for adenylate cyclase activity (for 30 min at 30 "C as previously described (30)). The fact that none of these components are stored in detergent allows them to be added directly to the lipid vesicles without disrupting the vesicle structure (as indicated by the maintenance of receptor-nucleotide regulatory protein coupling in these vesicles). The rhodopsin-promoted binding of [%3]GTPyS to the holotransducin or to (YT in the presence of substoichiometric amounts of Py suggests that 30-75% (typically -50%) of the protein added properly associates with the lipid vesicles.
Throughout these studies the amounts of N., transducin (YT, &T, subunit has migrated off the gels in lunes 2 and 3. The arrows to the left of these lanes show the relative mobility of known molecular weight standards. The a subunit shown in lune 1 was prepared by eluting the rod outer segment membranes with GTPyS. Identical results are obtained using a subunits which were prepared by eluting the membranes with G T P (results not shown). or rhodopsin are expressed in terms of total protein as determined by the fluorescamine method (3) or by the methods of Lowry (39) or Bradford (40). The molar concentrations of these proteins were determined assuming an M, of 95,000 for the holonucleotide regulatory protein (3), 39,000 for a~ (ll), 46,000 for BYT (ll), and 37,000 for rhodopsin (41). The molar concentrations of the 8-adrenergic receptor throughout the text are expressed in terms of ['251]iodocyanopindolol binding activity (pmol).

RESULTS
Effects of (YT and @-yr on the Adenylate Cyclase Activities in NJC Vesicles-Previously we reported that the retinal nucleotide regulatory protein, transducin, can substitute for Ni in inhibiting the GppNHp-stimulated adenylate cyclase activity in lipid vesicles containing the pure human erythrocyte N. and a resolved bovine caudate C preparation (30). In order to probe the mechanisms of inhibition of adenylate cyclase we have resolved the (Y (aT) and @y ( @ -y~) subunit components of transducin by Blue Sepharose chromatography as previously described (31). Fig. 1 shows the SDS gel electrophoresis profiles of these subunit preparations.
Both the ( Y~ subunit, which has been preactivated with GTP+ ((YT-GTP~S), and the @-yT complex inhibit the GppNHp-stimulated adenylate cyclase activity following their addition to N,/C vesicles as shown in Fig. 2 is -30% in the experiment shown, although it is frequently as high as 50-6076 (49 f 17% S.E., n = 6, also see Fig. 5 below). At 50 mM MgC1, this inhibition is markedly reduced (<5%). Similarly, the inhibition by the @YT complex (where [@YT] z 100-200 X [N.]) is much greater at low [MgCl,], being 90% (&2% S.E., n = 5) a t 2 mM MgC1, compared to about 15% a t 50 mM MgC12. In all cases the @YT complex is a much more potent inhibitor than the active (YT species. The inhibitory effects of both subunit components are specific for the Ne-stimulated adenylate cyclase activity. Fig.  3 shows that the direct addition of either the (YT.GTP~S or ByT species to phospholipid vesicles containing the resolved caudate adenylate cyclase preparation, alone, has essentially no effect on basal (Mg2"stimulated) or forskolin-stimulated adenylate cyclase activity. These results are completely consistent with our earlier studies where it was observed that neither pure holo-Ni nor pure holotransducin had any effect on the intrinsic activity of the bovine caudate catalytic unit (30).
The inhibitory effects of the ( Y~ subunit are also specific for FIG. 3. Effects of aT*GTPyS and &T on basal and forskolinstimulated adenylate cyclase activity. Resolved C preparation (100 pl) was added to a reconstitution incubation as described under "Materials and Methods." The isolated phospholipid vesicles were resuspended in 0.6 ml of 75 mM Tris-HCI (pH 7.8), 1 mM DTT, and then aliquots (0.18 ml) of these vesicles were incubated with 20 pl of buffer D, ~T . G T P~S (3 pg, ~8 0 pmol), or /3yT (3 pg, =SO pmol) prior to assaying their basal and forskolin-stimulated adenylate cyclase activity. Forskolin, 0.1 mM; [MgCI,] = 2 mM. B, the basal activity measured in the presence of MgC12 alone. The actual basal activities were 2.9 pmol of cAMP for C alone, 2.6 pmol of cAMP for C + a~' GTPyS, and 2.7 pmol of cAMP for C + &T. Each data point represents the mean of triplicate determinations from a single experiment which was repeated three times with comparable results. The error bars represent the standard deviation for the triplicate determinations. ence of GppNHp. These results most likely reflect the ability of aT.GDP to associate with the By complex of N, and thus promote the dissociation (and activation) of N. at low [Mg2+] (see Discussion).

Effects of aT and PyT on the Adenylate Cyclase Activities in PAR/N,/C
Vesicles-In order to examine whether the presence of the stimulatory PAR affects the inhibition of adenylate cyclase activity by aT and ByT, phospholipid vesicle systems containing the pure guinea pig lung PAR, together with N. and C, were constructed. Fig. 5A compares the GppNHpstimulated adenylate cyclase activities (in the presence and absence of the @-agonist isoproterenol) in PAR/N,/C vesicles alone and in these same vesicles following the addition of aT.
GTPyS or byT. Fig. 5B shows the analogous sets of comparisons, performed in the same experiment, using N,/C vesicles. In the case of the PAR/N,/C vesicles, there is only a slight stimulation of the GppNHp-stimulated activity (4.1-fold) by isoproterenol. This is due to the fact that when GppNHp is the activating guanine nucleotide, the agonist advantage is lost, i.e. many of the N, molecules can be activated in the absence of any agonist-receptor-N, interaction (see Discus- the active subunit species. An inactive aT subunit can be prepared by initially eluting transducin from rod outer segments with GTP. The hydrolysis of GTP to GDP results in an aT. GDP complex which can be resolved and purified. The &T complex has a much higher affinity for aT. GDP than it does for aT. GTPyS based on the relative abilities of PyT to promote the ADP-ribosylation by pertussis toxin (42) of these CYT complexes.' Fig. 4A shows the results of the addition of (YT. GDP to N,/C vesicles for conditions where the total [MgCl'] = 2 mM in the assay incubation. Unlike the case for the aT. GTPyS complex which inhibits the GppNHp-stimulated activity under these conditions, the aT-GDP complex is actually stimulatory. This stimulatory effect can be greatly PAR. Specifically, in direct comparative studies the aT. GTP-yS complex effected about a 50% inhibition (52 k 2%, n = 2 ) of the GppNHp-stimulated activities in the N./C vesicles while in the presence of PAR the extents of inhibition of the GppNHp-stimulated activities were reduced to 22% (&3%, n = 2 ) in the absence of isoproterenol and 16% (+1%, n = 2 ) in the presence of agonist. Likewise the @yT complex was a more effective inhibitor of the adenylate cyclase activities in N./C vesicles (94 & 1%, n = 2 ) compared to ,t?AR/N,/C vesicles where the inhibitions of GppNHp-stimulated activities were 77% (-+2%, n = 2) and 70% ( f 2 % , n = 2 ) in the absence and presence of agonist, respectively.
These inhibitions can be further reduced under conditions more akin to the true physiological situation where a greater percentage of the N, molecules have been activated via isoproterenol. PAR complexes, i.e. when assaying the isoproterenol plus GTP-stimulated adenylate cyclase activity. Fig. 6 presents the results of such an experiment where the activities are compared in the presence and absence of the P-yT subunit.
In this case the isoproterenol stimulation of the adenylate cyclase activity in PAR/N,/C vesicles, relative to the activity stimulated by GTP alone, is ~2.5-fold. Thus the percentage of N. molecules activated via agonist-receptor complexes is significantly increased relative to the conditions shown in Fig.  5A. Under such conditions the basal (GTP alone)-stimulated adenylate cyclase activity is inhibited -72% upon addition of the P-yT complex, similar to the extents of inhibition by P-yT of the GppNHp-stimulated activities in these vesicles (Fig.  5A). However, the isoproterenol (plus GTP)-stimulated activity is only inhibited by -45%. The differential inhibition of basal and agonist-stimulated activity in turn results in about a %fold increase (from 2.5-to 5.5-fold) in the -fold stimulation by isoproterenol for PAR/N,/C vesicles containing PYT. This in effect tightens the coupling between stimulatory agonists and adenylate cyclase (43). The nucleotide regulatory protein-mediated inhibition of isoproterenol (plus GTP)-stimulated adenylate cyclase activity can be completely accounted for by the @yT subunit. This is illustrated by the results presented in Fig. 7 which show that the inhibition by holotransducin can be fit rather well to the dose-response curve for the effects of @-yT on this activity. Under these conditions the effective activation of N, molecules by isoproterenol-PAR complexes is such that the addition of (YT. GTP-yS or N T . GDP has no effect on the agoniststimulated activity.

DISCUSSION
The specific mechanism(s) by which inhibitory receptors and the inhibitory guanine nucleotide-binding regulatory protein, Ni, mediate the inhibition of adenylate cyclase activity has been the subject of some controversy over the past few years. In order to examine this mechanism in a well defined system, we developed procedures for reconstituting the guanine nucleotide-dependent inhibition of adenylate cyclase activity in phospholipid vesicles containing pure N, and a resolved C preparation (30) and in lipid vesicles containing the pure @-adrenergic receptor together with N. and C (43). Our previous studies using such systems indicated that transducin, like Ni, could inhibit the adenylate cyclase activity. Thus, in the present work we have taken advantage of the ready availability of the isolated subunits of transducin to examine the roles of the N and P-y components of the regulatory protein in the inhibitory process.
As clearly illustrated by the data presented in Figs with no Ns). These results are fully consistent with earlier studies using holo-Ni or holo-transducin (30). The inhibition of the N,-stimulated adenylate cyclase activity by P-yT also supports prior work in various membrane systems where free P-y from Ni was observed to be a potent inhibitor of guanine nucleotide-stimulated activity (16, 17). As outlined earlier, such inhibition is consistent with a scheme where nucleotide regulatory protein activation reflects the dissociation of the heterotrimer into its component subunits. In this case, the addition of free P~T , by shifting the association-dissociation equilibrium of N. to the inactive heterotrimeric state, would deactivate the system. This is depicted in Scheme l where GN represents an activating guanine nucleotide such as GTP or GppNHp. For simplicity only one equilibrium is depicted for the dissociation of N,, the binding of M$+, and the binding of guanine nucleotide. The fact that Mg2+ overcomes inhibition is compatible with this scheme since increasing the divalent metal concentration would shift the equilibrium (in Scheme 1) to the right.
It should be noted that other mechanisms could account for the inhibition of N,-stimulated activity by PYT. For example, activation of N, might reflect a conformational change (but not dissociation) to an intact N.*-guanine nucleotide complex. If both the a, and Py subunits were involved in the binding of N,* to C during stimulation, then it would be expected that free PYT could act as a competitive inhibitor of this interaction. However, the stimulatory effects observed when aT. GDP complexes are added to N./C vesicles (Fig. 4) are not readily compatible with such a mechanism but rather suggest that N. activation does directly result in subunit dissociation, as depicted in Scheme 1, at least for cases where GN = GppNHp. Specifically, the (YT. GDP complex, by virtue of its high affinity for By, would act to pull the activation equilibrium to the right (in Scheme l), thereby increasing the levels of active as* and thus cause a net stimulatory response when added to N./C vesicles. The fact that this stimulatory effect is greatest under conditions where the levels of N, activation (and dissociation) are quite low, i.e. when total [MgCIP] < 2 mM, is further compatible with the mechanism presented in Scheme 1. Still, it has yet to be documented that stimulatory receptors, in the presence of hormones and GTP, promote the dissociation of N. into its subunit components. Thus, at the present time, it is not possible to rule out mechanisms of inhibition by By (of hormone plus GTPstimulated adenylate cyclase activity) which involve a direct interaction between this complex and the catalytic moiety. The inhibition of the N,. GppNHp-stimulated adenylate cyclase activity, by the preactivated aT.GTPyS species, is clearly weaker than that effected by PYT. Specifically, under conditions where PyT causes greater than 90% inhibition of the activity, the maximum inhibition obtained with aT. GTPyS typically ranges from 30 to 60%. However, the fact that some inhibition is observed with the preactivated aT species indicates that this subunit must be capable of directly interacting with C . This interaction could either occur at a site which overlaps the N, (or as) binding site on C or at a distinct site from which the (YT subunit effects an allosteric inhibition of N. (a,) binding to the catalytic moiety. The latter possibility would be compatible with recent results from the cyc-variant of S49 lymphoma cells where kinetic evidence indicates that aB and ai bind at two distinct sites on adenylate cyclase (21). However, in the case of the cyc-adenylate cyclase, binding by ai can apparently exert allosteric effects which inhibit the intrinsic activity of the catalytic moiety. This differs from the results of previous studies using bovine caudate enzyme, as well as those presented here, which indicate that the inhibitory effects of holo-Ni, holotransducin, and aT are all confined to the N.-stimulated activity (30). These differences may reflect subtle differences between the catalytic properties of the bovine caudate enzyme and that of the mutant cyc-cell.
Recently, we reported that the @-adrenergic receptor can overcome the inhibition of N,-stimulated adenylate cyclase activity by Ni (43). The same holds true 'for the inhibitory effects of transducin. Specifically, the inhibitions by aT. GTPyS and ByT of the GppNHp-stimulated activity are significantly reduced in PAR/N,/C vesicles compared to NJC systems. The fact that the receptor relieves this inhibition, even in the absence of isoproterenol, can be attributed to the "basal" activation of N. molecules by receptor alone (32,44). It is interesting that the inhibitions by both subunits are more severely depressed when GTP is used, together with isoproterenol, to stimulate activity compared to when GppNHp is used as the activating guanine nucleotide. This can be explained with the aid of Scheme 2 where PAR= R, and H = agonist. For simplicity the activation of N. is assumed to directly result in subunit dissociation. In addition, the receptor-promoted activation is depicted as occurring exclusively via agonist-receptor complexes, although some basal activation by receptor alone would in fact also occur (32,44). Based on the reductions in the -fold stimulation of adenylate cyclase activity by isoproterenol, when GppNHp is the activating guanine nucleotide (Fig. 5A) compared to GTP (Fig. 6), it is likely that in the former case the N, molecules can be activated either via pathway 1 or pathway 3, while in the latter case pathway 3 predominates. (It should be noted that in these phospholipid vesicle systems as many as 50% of the total N. molecules which are accessible to guanine nucleotides are not accessible to HR complexes but nevertheless can be spontaneously activated by high concentrations of nonhydrolyzable GTP analogs and Mg2+ (32).) As illustrated by the data presented in Figs. 5 and 6, N. molecules activated via pathway 1 are more susceptible to the inhibitory effects of both aT. GTPyS and ByT than are N, molecules activated via PAR (ie. pathway 3). Thus, it would be expected that a greater percentage of the total adenylate cyclase activity is susceptible to inhibition when GppNHp is the activating guanine nucleotide compared to when GTP is used.
An important outcome of these results is that in PARIN./ C vesicles the inhibitions by (YT. GTPyS or PYT are not strictly dependent on the amount of active N, present (relative to these subunits) but rather on the amount of active N, which has been formed via the agonist-receptor pathway. This, then, suggests that the HRa, complex (Scheme 2) must remain intact for a finite period during both the activation process and the stimulation of adenylate cyclase activity. Thus the inhibition by PyT would be reduced if the a. subunit, when complexed to HR, had a much weaker affinity for ,ByT compared to a. alone. Or, in the event that subunit dissociation is not required for activation, HRN,* complexes might interact more tightly with C (and thus more effectively compete with free &T for a binding site on the catalytic moiety) compared to N.* alone. Similarly, if HRa,* or (HRN,*) complexes had an increased affinity for C they would better overcome the inhibitory effects of CYT-GTPyS. For every case examined, we found the ByT subunit to be a more potent inhibitor of the N,-stimulated adenylate cyclase activity than the CYT. GTPyS complex. In fact, under conditions where agonist stimulation of adenylate cyclase activity is maximal (i.e. when stimulating with isoproterenol plus GTP) the CY^. GTPyS inhibitory effects are completely eliminated, and the inhibition observed with holo-transducin can be fully accounted for by the byT complex. These results strongly suggest that under physiological conditions, By is primarily responsible for both mediating the inhibitory response as well as effecting a tighter coupling between stimulatory agonists and the effector enzyme by severely depressing basal (GTP alone)-stimulated activity.