the of chemistry vol. 261, no. 4, of pp. 1656 … 1986 jbc.pdf · catecholamine-stimulated gtpase...

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1986 by The American Society of Biological Chemists, Inc

Vol. 261, No. 4, Issue of February 5 , pp. 1656-1664,1986 Printed in U.S.A.

Catecholamine-stimulated GTPase Cycle MULTIPLE SITES OF REGULATION BY @-ADRENERGIC RECEPTOR AND Mg2+ STUDIED IN RECONSTITUTED RECEPTOR-GB VESICLES*

(Received for publication, August 22, 1985)

Douglas R. Brandt and Elliott M. Ross$ From the Department of Phnrmacobgy, Southwestern Graduate School, University of Texas Health Science Center, Dallas, Texas 75235

The regulation of the intermediary steps of the catecholamine-stimulated GTPase cycle by &adrenergic agonists and Mg2+ was investigated using unilamellar phosphatidylethanolamine- phosphatidylserine vesicles that contained purified &adrenergic receptor and the stimulatory GTP-binding protein of the adenylate cyclase system, G.. The steady-state turnover number of the agonist-stimulated GTPase, normalized according to the receptor-responsive pool of G., was 0.8 min-' for untreated vesicles and 1.7 min" for vesicles that had been treated with dithiothreitol to activate the receptors. The binding and release of [~Y-"~P]GTP, ['HI GTP, and [y32P]GTP were used to measure the binding and hydrolysis of GTP and the release of GDP. Ago- nist-liganded receptor stimulated both the binding of GTP and the release of the GDP product, and GDP release per se did not appear to be the mechanism by which receptor stimulated the binding of GTP. Both processes displayed apparent first order rate constants of about 0.5 min" for untreated vesicles and both rates increased about 5-fold after dithiothreitol treatment. Both processes were formally catalytic with respect to receptor, in that several (up to 8) molecules of G. were stimulated per molecule of receptor. The hydrolysis of G.-GTP to G.-GDP was unaltered by agonist and occurred with a rate constant of about 4 min-l. The rates of these partial reactions were consistent with the overall rate of steady-state hydrolysis and with the ability of the agonist-liganded receptor to promote the formation of sufficient G.-GTP to fully stimulate adenylate cyclase in a native membrane.

The Mg2+ dependence of agonist-stimulated, steady- state GTPase activity appeared to consist of at least two, distinct Mg2+-requiring processes. Very low concentrations of Mg2+ (-20 nM) were required for hydrolysis of G.*GTP, and 10 MM Mg2+ was required to maximize the initial rate of agonist-stimulated [cY-'~P] GTP binding.

The generally accepted mechanism for the hormonal regulation of adenylate cyclase, postulated 10 years ago by Cassel and Selinger (l), holds that the binding of GTP stimulates the activity of the enzyme and that hydrolysis of the bound

* This work was supported by National Institutes of Health Grant GM30355 and R. A. Welch Foundation Grant 1-982. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "adver- tisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ Recipient of an Established Investigator Award from the Amer- ican Heart Association.

GTP to GDP causes its deactivation. A regulatory GTPase cycle is thus described in which the steady-state concentration of GTP-liganded enzyme determines the activity of the cyclase, and the rates of the formation and hydrolysis of this species is of central importance. Cassel and Selinger (2-4) further suggested that hormone promotes the binding of GTP (and consequent activation) by increasing the rate of release of GDP, and this view has been supported by evidence from several laboratories (see Refs. 5-7 e.g. and Ref. 8 for review). While this mechanism is probably correct in part, it remains unclear that the hormone-liganded receptor accelerates GTP binding solely by promoting the clearance of GDP (see Refs. 9-11, reviewed in, Ref. 12).

The site of stimulatory regulation of adenylate cyclase by guanine nucleotides has now been shown to be a distinct, GTP-binding regulatory protein, G,' (see Ref. 8 for review). We recently demonstrated that G, alone can catalyze the GTPase reaction associated with the control of adenylate cyclase and, when G, and ,&adrenergic receptor were co- reconstituted into phospholipid vesicles, the GTPase activity of G, was stimulated up to 15-fold by @-adrenergic agonists (13). Similar results have been reported by Cerione and co- workers (14). A highly homologous system exists in the retinal rod outer segment, where bleached rhodopsin regulates the GTPase activity of transducin, another GTP-binding protein that regulates a cyclic GMP phosphodiesterase (15, 16).

In an initial description of the agonist-stimulated GTPase activity of G,, we noted that the binding of guanine nucleotides to G, and the steady-state GTPase reaction were both slow processes and were similarly regulated, and we proposed that the rate-limiting step in hydrolysis might be the binding of the GTP substrate (13). This proposal was consistent with results of a subsequent study of the GTPase activity of isolated G, (17), which indicated that the binding of GTP and the observed dissociation of GDP were both relatively slow ( 4 - 1-5 min), but that the hydrolysis of G.-bound GTP to GDP was quite rapid (tH < 5 s). However, only a relatively small fraction of the GDP product accumulated as the tightly bound species, and up to 70% of the GDP was released too quickly to be measured. This suggested that the release of bound GDP was not the sole rate-limiting step in the GTPase cycle.

Because the control of the formation and turnover of the G,.GTP complex is central to the regulation of adenylate cyclase (and may serve as a prototype for other G protein-

'The abbreviations used are: G., the stimulatory, GTP-binding protein of the adenylate cyclase system; DTT, dithiothreitol; GTP-yS, guanosine-5'-0-(3-thiotriphosphate); Hepes, 4-(2-hydroxyethyl)-l- piperazineethanesulfonic acid; 1, liter; EGTA, ethylene glycol bis(p- aminoethyl ether)-N,N,N' ,N' -tetraacetic acid.

1656

Regulation of Catecholamine-stimulated GTPase 1657

mediated systems), we have begun to characterize the mechanism whereby purified P-adrenergic receptor controls the binding, hydrolysis, and release of guanine nucleotides by GB. In this report, we describe some basic kinetic and regulatory properties of this process and present data that argue for the independent control of both GTP binding and GDP release by the receptor.

EXPERIMENTAL PROCEDURES

Materials-Lubrol 12A9 was a gift from ICI, Ltd. Lubrol solutions

stored at 4 "C. [35S]GTPyS (800-1500 Ci/mmol), [8,5'-3H]GTP (30 (10%) were deionized by passage over Dowex Ag 501 (Bio-Rad) and

Ci/mmol), and [3H]dihydroalprenolo1 (90-130 Ci/mmol) were purchased from New England Nuclear. Nonradioactive GTPyS (Boeh- ringer Mannheim) was purified by elution from DEAE-Sephacel (Pharmacia) with a gradient of 0-0.5 M LiC1. [ c Y - ~ ' P ] G T P ~ ~ ~ [y-32P] GTP were prepared according to Johnson and Walseth (18). The purity of both labeled and unlabeled nucleotides was monitored by thin layer chromatography as described previously (19). ['251]Iodo~y- anopindolol was prepared according to Engel et al. (20). (*)-Cyano- pindolol was a gift from G. Engel, Sandoz Pharmaceuticals. Alpren- 0101 was a gift from Hassle Pharmaceuticals, (-)-isoproterenol was a gift from the Sterling-Winthrop Research Institute, and (-)-propranolol was a gift from Ayerst Laboratories. BA85 nitrocellulose filters were purchased from Schleicher & Schuell, Norit A charcoal from Fisher, GF/F filters from Whatman, and phospholipids from Avanti Polar Lipids. Digitonin was purchased from Sigma and was used without further purification. All other reagents were of the highest purity available.

Solutions that were used in experiments involving the effects of low concentrations of Mf were passed over Chelex 100 resin (Bio- Rad) twice before use. This treatment also removes heavy metal ions. Contaminating M e in these solutions was measured by atomic absorption spectroscopy by Dr. Michael Nicar, Division of Mineral Metabolism, Department of Internal Medicine. Stock solutions of Mg2+ were standardized according to the calmagite method of Diehl (21).

Protein Purification-G. was purified from rabbit liver according to Sternweis et al. (22) and stored at -80 "C in 10 mM NaHepes (pH 8), 1 mM EDTA, 0.1% Lubrol 12A9.

8-Adrenergic receptors were purified using modifications of the method of Shorr et al. (23). All procedures were carried out at 0-4 "C except where noted. Plasma membranes from turkey erythrocytes, prepared as previously described (24), were suspended at 5 mg/ml in 20 mM NaHepes (pH 8), 0.1 M NaC1, 5 mM EDTA, 0.1 mM EGTA, 1.75% digitonin and stirred for 60 min. This and all other solutions contained 0.001% diisopropyl fluorophosphate, 0.1 mM phenylmeth- ylsulfonyl fluoride, 10 p~ benzamidine, 2.5 pg/ml bacitracin, gnd 2.8 pg/ml Trasylol except as noted. The mixture was centrifuged for 1 h at 35,000 rpm in a Beckman type 35 rotor. The supernatant (about 1.6 1) was filtered sequentially through Whatman No. 2 filter paper and 3 and 0.45 pm Millipore filters and was applied to a 160-ml column of alprenolol-Sepharose, prepared according to Caron et al. (25). Application took 12-20 h. The column was washed with 2 volumes of 20 mM NaHepes (pH 8), 5 mM EDTA, 0.1 mM EGTA, 1.0 M NaCl, 0.05% digitonin followed by 2 volumes of 20 mM NaHepes (pH 8) , 5 mM EDTA, 0.1 mM EGTA, 0.1 M NaC1, 0.05% digitonin ("low-salt buffer"). Trasylol, benzamidine, and bacitracin were omitted from this and subsequent buffers. The column was eluted at 30 "C with 10 p M (-)-alprenolo1 in low-salt buffer. The pooled, receptor- containing fractions (typically 80-150 ml) were concentrated 25-fold by ultrafiltration on an Amicon PM-30 filter and then diluted to the original volume with 20 mM NaHepes (pH 8), 0.5 mM EDTA, 0.1 mM EGTA, 0.05% digitonin. This material was applied to a 10-ml column of DEAE-Sephacel. The column was washed with 2 volumes of 20 mM NaHepes (pH 8), 0.5 mM EDTA, 0.1 mM EGTA, 0.2% digitonin and eluted with a linear gradient of 0-200 mM NaCl in the same buffer. Receptor was eluted at about 50 to 100 mM NaCI. The pooled, receptor-containing fractions were concentrated about 10-fold by ultrafiltration as above and stored at -80 "C. Typically, 1-1.5 nmol of receptor was obtained in each preparation.

This preparation typically contains 12-14 nmol of &adrenergic ligand binding sites/mg protein, and the highest value ever observed was 18 nmol/mg. This is significantly lower than the theoretical value of 23 nmol/mg that is based on a molecular weight of 43,000. Shorr

et al. (23) also reported a maximum of 18 nmol/mg. However, analysis of the receptor by sodium dodecyl sulfate-polyacrylamide gel electro- phoresis indicates that it is 90-97% pure, according to silver staining (shown in Refs. 26 and 27). The cause of this discrepancy is uncertain. The binding capacity is similar whether [3H]dihydroalprenolo1 (specific activity determined by the manufacturer) or ['251]iodocyanopin- dolo1 (radiochemically pure) was used for the assay, suggestlng that this value was not in error. Analysis by reverse phase high pressure liquid chromatography indicates that the major contaminant, globin, accounted for less than 5% of the protein. We therefore think it likely that this preparation of the receptor is substantially pure, but that it contains a variable amount of denatured material.

Reconstitution-Reconstitution of receptor and G. into phospholipid vesicles was performed essentially according to Brandt et al. (13). G. and receptor (molar ratio typically 3:1 to 101) were combined with a detergent/lipid mixture to yield a dispersion containing 0.3 mg/ml phosphatidylethanolamine, 0.2 mg/ml phosphatidylserine, 2 mg/ml deoxycholate, and 0.4 mg/ml cholate. Detergent was removed by gel filtration on Sephadex G-50 in 20 mM NaHepes (pH 8), 3 mM MgCl,, 1 mM EDTA, 100 mM NaC1. For some experiments, reconsti- tutions were performed in the absence of added MgCIZ. Vesicles were collected in an eluted volume 4-6 times that of the original mixture. The molar ratio of G. to receptor in the vesicles depended on recovery and varied from 7 to 40 in the experiments performed here. Recovery of receptor was monitored by assaying the binding of ['261]iodocyanopindolol or [3H]dihydroalprenolo1. Recovery of G. was assayed according to [35S]GTPyS binding after resolubilization by Lubrol 12A9 (see below). When necessary, vesicles were concentrated by centrifugation at 50,000 rpm for 2 h a t 2 "C in a Beckman 70.1 Ti rotor and resuspended in a minimal volume of 20 mM NaHepes (pH 8), 100 mM NaCI, 1 mM EDTA, 3 mM MgCl,. DTT Treatment of Reconstituted Vesicles-For some experiments,

receptors were activated by reduction of disulfides by incubating vesicles with 5 mM dithiothreitol for 1-3 h at 0 "C. Pedersen and Ross (28) have shown that such treatment enhances the ability of the receptor to promote the binding of GTPyS to G. and the steady-state hydrolysis of GTP. Subsequent to treatment, vesicles were diluted such that the final DTT concentration in the assay was always less than 0.5 mM. Because treatment with DTT results in activation of receptor even in the absence of agonist, the "basal" activity in vesicles was always determined using untreated vesicles in order to estimate the total stimulation caused by receptor. Thus, the incremental effect of isoproterenol on binding of nucleotides or GTPase was obtained by subtracting the basal activity for untreated vesicles from the agonist-stimulated activity of either DTT-treated or untreated vesicles.

Assays-The binding of [~I-~'P]GTP, [y-3ZP]GTP, [8,5'-3H]GTP, and [36S]GTPyS was measured essentially as described previously (17, 19). Vesicles were incubated in medium containing 20 mM NaHepes (pH 8) , 100 mM NaCl, 0.1 mM ascorbate, and the concentrations of EDTA, MgCL, nucleotide, and @-adrenergic agent shown in the text (0.05 ml total volume). The concentration of free M e in the assay medium was calculated using a K, for Mg-EDTA of 0.4 p~ at pH 8.0 (29) and ignoring the binding of M$' to lipid or nucleotide. Assays were initiated by the addition of vesicles. Unless indicated, assays were quenched with 2 volumes of ice-cold buffer containing 20 mM Tris-C1 (pH 8), 1 mM GTP, 10 mM MgC12, 100 mM NaCl, 0.1 mM (_+)-propranolol, 0.1% @-mercaptoethanol, and 0.1% Lubrol 12A9. Binding was stable in this solution for at least 3 h at 0 "C. Nucleotide bound to G. was separated by filtration on BA85 membrane filters and the filters were washed with 14 ml of 20 mM Tris-C1 (pH 8), 100 mM NaC1, 10 mM MgCl,. When the binding of [y3'P]GTP and [3H] GTP or [w3'P]GTP was assayed in parallel, reaction mixtures were quenched and filtered immediately. Nonspecific binding, determined in the presence of a 1000-fold excess of GTP, was always less than 10% of total binding at steady state with 0.2 p~ nucleotide and was accounted for by the binding of nucleotide to the filter. We report only specific binding, defined as the difference between total and nonspecific binding. The total amount of G. in the vesicles was determined by measuring the amount of [35S]GTPrS bound in the presence of 50 mM M e , 0.1% Lubrol 12A9, and 10 p~ GTPyS after 30 min of incubation at 30 "C (19). The identity of bound nucleotide was determined by thin layer chromatography as described (17).

GTPase activity was determined as described previously (17). Vesicles were incubated at 30 "C in the same medium used for nucleotide binding assays, except that [y-32P]GTP was used as a substrate. The reaction (0.05 ml) was quenched with 0.5 ml of an ice-

1658 Regulation of Catechola

cold suspension of Norit A (5% (w/v) in 50 mM NaH2P04). The mixture was centrifuged (1500 rpm, 10 min) and an additional 0.5-ml aliquot of Norit A was added. After a second centrifugation, the 32P in 0.6 ml of the supernatant was determined by Cerenkov counting.

Binding of [1251]iodocyanopindolol or [3H]dihydroalprenolo1 was assayed by the method originally described for [1251]iodohydroxyben- zylpindolol (30). The binding of [3H]dihydroalprenolo1 to soluble receptors was assayed by the centrifugal gel filtration procedure (30). Protein was assayed by the method of Schaffner and Weissmann (31) using bovine serum albumin as standard.

RESULTS

Agonist-stimulated GTPase Activity in Receptor-G, Vesi- cles-When reconstituted with purified @-adrenergic receptors into phospholipid vesicles, G, displayed its characteristic GTPase activity? When the reaction was initiated by the addition of vesicles, there was no obseryed lag and the rate of hydrolysis was constant (Fig. 1, A and D). (-)-Isoproterenol stimulated the GTPase activity 15-fold over the basal activity and produced half-maximal stimulation at 0.1 to 0.2 ~ L M (not shown). The single K, for GTP was 0.1 PM (not shown). In contrast to our initial study in which only partially purified receptor was used (13), no high-K, component could be observed in vesicles composed of pure Teceptor, pure G., and two pure phospholipids.

Under identical conditions, agonist stimulated the rate of quasi-irreversible binding of [35S]GTPyS, a nonhydrolyzable analog of GTP, to receptor-(;, vesicles roughly 25-fold (Fig. 1, B and E). Binding reached plateau levels within 1 min under optimal conditions, and agonist-stimulated GTPyS binding typically represented 40 to 60% of the total G. present.

1 6 0 , , , , , , DTT TREATED UNTREATED

30

30 v

Time (min)

FIG. 1. Agonist-stimulated GTPase activity and nucleotide binding in receptor-G. vesicles. Assays were conducted at 30 “C for the times shown in medium containing 3 mM MgC12,l mM EDTA, and 0.2 ~ L M of either [y3’P]GTP (A and D), [35S]GTPrS ( B d E) , or [a-32P]GTP (C and F), Panels A, B, and C represent DTT-xeated vesicles that contained 1.6 fmol of receptor and 36.2 fmol of GJassay. Panels D, E, and F represent untreated vesicles from a different preparation that contained 3.3 fmol of receptors and 102 fmol of Gs/ assay. Data were obtained in the presence (closed symbols) or absence (open symbols) of 10 p~ (-)-isoproterenol. Duplicate binding and triplicate GTPase assays were performed.

G. is an active GTPase only in the absence of detergents (13,17), and inhibition by detergent may explain the reports of Sunyer et al. (32) and Northup et al. (33) that G, in Lubrol solution was negligibly active as a GTPase. Phospholipids stabiiize G,, but do not markedly enhance GTP hydrolysis.

:mine-stimulated GTPase

These data agree with those obtained by Asano et al. (19,34). As was observed for GTPase activity, 0.1 to 0.2 p~ isoproterenol caused half-maximal stimulation of the rate of GTPyS binding (not shown). This concentration is similar to the Kd for isoproterenol (19). In the experiment shown in Fig. lB, 8 molecules of G. were stimulated to bind nucleotide per receptor.

In our initial report on the reconstituted, agonist-stimulated GTPase activity of G,, we calculated a molar turnover number by normalizing the GTPase activity to the total amount of G, that could be assayed in each preparation of vesicles (13). In different preparations of vesicles, however, this turnover number has varied 4-5-fold (Table I). We believe that a more meaningful and appropriate index is the ratio of the agonist- stimulated GTPase activity to the number of receptor-accessible G. molecules, which is estimated according to the amount of GTPyS binding that is stimulated by agonist (Table I). For DTT-treated receptors, this turnover number varied only from 1.5 to 2.2. min-l in a number of different vesicle preparations (Table I). For untreated receptors, it varied from 0.5 to 1.1 min-l. The striking consistency of this value is a strong argument that the “pool” of G, that undergoes hormone- stimulated GTPase activity is the same as that which undergoes hormone-stimulated activation by binding GTPyS. It further argues that the true turnover number for the agonist- stimulated GTPase activity is either 0.8 or 1.7 min-l for untreated or DTT-treated receptors, respectively. These values should represent a maximum estimate of the transit time of a single G, molecule through a complete catalytic cycle. We do not know what determines the accessibility of G, to recep- tars in the vesicles, but the pool of G. that was not activated by GTPyS in response to agonist was also insensitive to Mgz+-promoted activation until detergent was added (19).

The binding of [cx-~~P]GTP or [3H]GTP can be used to measure the transient binding of both substrate and product to G. during steady-state hydrolysis. Using isolated G,, we previously found that the amount of isotope bound represented about 30-40% of the G, that could bind GTPyS. Of

, this amount, virtually all of the bound isotope was the GDP product (17). These studies have now been extended to the receptor-G. vesicles. As shown in Fig. 1, C and E: the initial rate of [LY-~’P]GTP binding was stimulated roughly 15-fold by agonist, as was the rate of binding of GTPyS and the overall GTPase rate. The steady-state level of agonist-stimulated binding of 32P-labeled nucleotide was typically 70-80% of steady-state GTPyS binding. The concentration of GTP at which half-maximal agonist-stimulated binding was observed was 0.1 p ~ , and the concentration of isoproterenol giving half-maximal stimulation of the rate of nucleotide binding ‘was 0.1 to 0.2 pM (data not shown). These concentration dependences are identical to those for the overall GTPase reaction.

In contrast to isolated G., GTP represented 20-30% of the nucleotide bound to reconstituted G. during agonist-stimulated hydrolysis. Relative amounts of bound GTP and GDP were determined routinely by comparing the binding of [a- 3ZP]GTP, which measures bound GDP plus GTP, with the binding of [y-32P]GTP, which measures bound GTP alone (Table 11). This measurement was confirmed by thin layer chromatography of the bound nucleotides (not shown). The increased steady-state concentration of G.. GTP observed here might be accounted for by the increased rate of binding of GTP, shown in Fig. 1, or by a decrease in its rate of hydrolysis. As shown in Fig. 2, the rate of hydrolysis of bound GTP was much slower for reconstituted G, than the rate previously observed with isolated G,. In this experiment, [y-

Regulation of Catecholamine-stimulated GTPase

TABLE I Agonist-stimulated GTPase turnover numbers relative to total G, and receptor-sensitive G,

1659

Turnover numbers for isoproterenol-stimulated GTPase activity were determined in several different preparations of recept0r-G. vesicles either before or after treatment with DTT. They are expressed relative either to the total amount of G, present in the vesicles or to the amount of G. that could be stimulated by agonist to bind GTP+. The fraction of GB that was sensitive to regulation by the receptor vaned from 9 to 54%. AB values are expressed either as ratios or as the activity in 1 p1 of vesicles. Avg., average.

Turnover number

DTT-treated vesicles Untreated vesicles Total G: Coupled C.6 Total G. Receptor'

Totald Coupled' Totald Coupled'

fmol fml mol. min" . mol G,-' ml. min" .mol G," 40.7 3.7 26.4 0.17 1.9 0.09 1.0 39.0 3.9 28.9 0.17 1.7 43.7 5.7 35.0 0.29 2.2 22.4 5.6 11.8 0.30 1.2 0.20 0.8 36.2 6.2 22.6 0.33 2.1 39.8 9.2 0.34 1.5 0.18 54.0

0.8 15.1 33.1 0.46 1.7

26.3 10.0 11.0 0.57 1.5 0.27 1.5

25.3 0.7

8.9 7.0 0.62 1.8 0.17 0.5 8.2 4.0 16.4 0.78 1.6 0.50

20.0 10.8 7.7 0.78 1.5 0.36 0.7 1.1

7.4 2.6 0.53

Avg. = 1.68 +. 0.09 (S.E.) Ave. = 0.80 f 0.08 (S.E.) ' Determinedby the binding of 10 p M [%?Y]GTP~S in the presence of 50 mM MgC12 and 0.1% Lubrol12A9 for

30 min at 30 "C (19). The agonist-stimulated binding of 0.2 pM [%]GTPrS was measured in the presence of 2 mM free M f l and

10 p~ (-)-isoproterenol after 10 min for untreated vesicles and after 3 min for DTT-treated vesicles. The increment in binding above basal levels indicates the amount of G. that was sensitive to regulation by the receptor ("coupled" Gd.

~ ~

Receptor was measured using ['261](-)-iodocyanopindolol, and the ratio of total G. to receptor is shown. GTPase activity was determined under the same conditions used to measure coupled G., except that [y"P]

The turnover number for agonist-stimulated GTPase was calculated by dividing by coupled G,. GTP was substituted for [36]GTPrS. The turnover number was calculated by dividing by total GB.

TABLE I1 Agonist-stimulated binding of [ c Y - ~ ~ P ] and [ T - ~ ~ P I G T P in the

presence and absence of Mg2+ Vesicles, either DTT-treated or untreated, were incubated for 3

min in assay medium at 30 "C with the indicated concentrations of MgC12, EDTA, and either 0.2 p~ [a-32P]GTP or [-y-32P]GTP. Binding was measured in the presence or absence of 10 pM (-)-isoproterenol by immediate filtration. The increment in binding of nucleotide caused by isoproterenol is shown. Each value, the average of triplicate determinations, represents 3.6 fmol of receptor and 81 fmol of G..

Nucleotide Nucleotide [y-a2plGTp

bound [o~-"P]GTP bound [EDTA1 IMgchl +DTT -DTT +DTT -DTT

Nucleotide bound [EDTA1 IMgchl bound [o~-"P]GTP

+DTT -DTT +DTT -DTT mM mM

[a-32P]GTP 1.0 - ~ T - ~ ~ P ~ G T P 1.0 18" 0.69 1.00 - 12.4 12.9 [a-32PiGTP 1.0 2.0 25.3 17.6 o.35 o.28 [ T - ~ ~ P ] G T P 1.0 2.0 8.8 4.9

The total concentration of M$+ was 1.8 p ~ , as determined by atomic absorption spectrophotometry. Given the EDTA concentration, the concentration of free M e was 0.7 nM.

32P]GTP was incubated with receptor-G, vesicles at about 1 nM Mg2+ until steady-state binding was reached. The reaction was quenched with unlabeled GTP, Mg2+ was added to 3 mM, and the disappearance of bound 32P was measured. At 30 "C, GTP was hydrolyzed at a rate of about 4 rnin", and this rate was not altered by the addition of antagonist. At 0 "C, no hydrolysis was observed in 15 min of incubation (not shown). These results are in marked contrast to those obtained with free G., where hydrolysis occurred at the rate of about 1 min" at 0 "C and was too fast to measure at 30 "C (17).

The experiments performed in Fig. 1, A-C utilized DTT-

treated receptor, which is significantly more active in stimu- lating G, (28). A similar pattern of data was obtained using untreated receptor (Fig. 1, D-F), although all rates were lower. The concentration dependence for agonist, antagonists, or nucleotide was unchanged by DTT treatment (data not shown). While the ratio of agonist-stimulated [cx-~'P]GTP binding to [35S]GTPyS binding at steady-state was slightly lower in the untreated vesicles shown here, 0.63 compared to 0.78 for DTT-treated vesicles, no consistent difference has been observed. A ratio of 0.7 to 0.8 has been the rule. For isolated G., this ratio was 0.3-0.4 (17). Relative to the total number of G. molecules, DTT treatment increased both the agonist-stimulated GTPase activity and the rate of nucleotide binding by 2- to 4-fold. In the following experiments, DTT- treated and untreated receptors were both used with similar qualitative results.

Agonist-stimulated Dissociation of Bound GDP-It has been reported by numerous investigators that agonists that activate adenylate cyclase can cause the release of GDP from plasma membranes that had been preincubated with GTP (see Ref. 12 for review). It was therefore of interest to determine if agonist-bound receptors could stimulate the dissociation of bound GDP in a purified, reconstituted system. In the experiment shown in Fig. 3, [cx-~'P]GTP was bound to the vesicles during agonist-stimulated hydrolysis, and the dissociation reaction was then monitored after the addition of a large excess of unlabeled GTP. Dissociation of bound nucleotide was rapid in the presence of isoproterenol, and the major, rapid component of dissociation was blocked by the @-adrenergic antagonist, propranolol. Its rate constant was approximately 0.5 rnin". Note that because the rate constant for hydrolysis was 4 rnin", the dissociating nucleotide was pre-

1660 Regulation of Catecholamine-stimulated GTPase

1 0 10 20 30 40

Time (s )

FIG. 2. Rate of hydrolysis of bound [r-S2P]GTP. Vesicles (DTT-treated) were incubated at 30 "C under GTPase assay conditions with 10 p M (-)-isoproterenol, 0.21 mM MgC12, 100 mM EDTA, and 0.2 p~ [y-32P]GTP. After 8 min, 0.01 volume of either 10 mM (-)-alprenolo1 (0) or Hz0 (0) was added. At 10 min, an aliquot from each incubation was assayed in duplicate for bound nucleotide by immediate filtration and the remainder was diluted 20-fold into assay buffer to yield final concentrations of 8 mM M P , 5 mM EDTA, 10 PM GTP, and either 10 p M (-)-isoproterenol (0) or 5 p M alprenolol plus 0.5 p~ isoproterenol (0). Subsequent to this dilution, equivalent amounts of vesicles were assayed in duplicate for bound nucleotide by immediate filtration. Each time point represents vesicles that contained 15 fmol of receptor and 160 fmol of G.. The amount of [y- 32P]GTP bound before dilution (100%) was 17 fmol. Error bars indicate the standard deviation for the 100% bound value ( n = 4) and the propagated error ( E ) for the other values, calculated as

AB B,.ABo E = - + - Bo Bo2

where Bo and B, are the amount of GTP bound at zero time and at each time point, AB0 is the standard deviation of Bo, and A B , is 0.5 times the difference between duplicate measurements of each B,.

lo u 0 I O 20 30 Time (mid

FIG. 3. Agonist-stimulated dissociation of bound guanine nucleotide. Vesicles were incubated at 30 "C in the presence of 3 mM MgCI2, 1 mM EDTA, 10 pM (-)-isoproterenol, and 0.2 p M [a- 32P]GTP. After 5 min, 0.01 volume of 10 mM (-)-propranolol (0) or an equal volume of water (0) was added. After 2 min more, bound nucleotide was assayed and a 0.01 volume of 0.1 M GTP was added. Bound nucleotide was assayed at the indicated times in duplicate. Each time point represents vesicles that contained 3.6 fmol of receptor and 158 fmol of G.. The amount of nucleotide bound at zero time was 30 fmol.

dominantly GDP. In this experiment, about 6 molecules of G, were stimulated to release bound GDP per receptor. Thus, the agonist-bound receptor could catalytically cycle among multiple G, molecules to promote dissociation, analogous to its catalysis of nucleotide binding.

The receptor-catalyzed phase of dissociation accounted for over 50% of the assayable G,.GDP complex. A slower component, about 25% of the total, was not blocked by propranolol and displayed a rate constant of about 0.015 rnin". This was the principal kinetic component of basal dissociation. How- ever, neither agonist-stimulated nor basal dissociation could be fit by these two components alone; a very rapid early component was also evident in the presence or absence of propranolol. The nature of this component is not clear. It might represent GDP bound to the free a subunit of G, (see Ref. 17), GDP bound to a preexistent GB-receptor complex, or

Propranolol blocked the agonist-stimulated dissociation of bound GDP in a dose-dependent manner (Fig. 4). Blockade of the stimulatory effect of agonist appeared to be complete, because the fractional dissociation of nucleotide in the absence of agonist was identical to that observed in the presence of isoproterenol plus saturating propranolol. The concentration of isoproterenol that will half-maximally promote GDP dissociation can be estimated from such an experiment by assuming simple competition between agonist and antagonist. Given the Kd for propranolol of 40 nM (determined by competition for binding with ['251]iodocyan~pindolol (19)), its half-maximally effective concentration in the experiment (about 3 KM), and the concentration of isoproterenol used, we calculated that approximately 0.13 KM isoproterenol could stimulate the rate of GDP dissociation half-maximally. This estimate is in good agreement with the half-maximally effec-

G,.GTP.

50 - - 0

40 - .3 c

m 0' 30 al

20 0 " s 10

-

0 t 1

FIG. 4. Concentration dependence on propranolol for inhibition of agonist-stimulated nucleotide dissociation. Receptor- G. vesicles were incubated at 30 "C with 10 PM (-)-isoproterenol, 3 mM MgC12, 1 mM EDTA, and 0.2 pM [a-32P]GTP. At 8 min, 0.04 volume of (-)-propranolol was added to each incubation to yield the final concentrations indicated on the abscbsa. Two min later, aliquots were assayed for bound nucleotide (0) and 0.04 volume of 25 mM GTP was added to yield a final concentration of 1 mM. Two min later, the remaining bound nucleotide was assayed again (0). The solid line that connects the open circles was drawn assuming simple competition and using a Kd for propranolol of 40 nM, an ECm for isoproterenol of 130 nM, and lower and upper limits on the amount of GDP bound after 2 min of dissociation of 18.9 and 37.5 fmol. As a control, a separate volume of vesicles was incubated as above except that both isoproterenol and propranolol were omitted in order to measure basal levels of binding (A) at 10 min and dissociation after 2 min with excess GTP (A). Note that 20% of GDP dissociated in 2 min either in the presence of isoproterenol plus M propranolol or in the absence of both &adrenergic ligands. Each point represents vesicles that contained 14.7 fmol of receptor and 150 fmol of G,.


u \ O O 10 x) 30;'; ' ; ' A

Time (min)

FIG. 5 . Agonist-stimulated GTPrS binding prior to and subsequent to the dissociation of bound GDP. All assays were conducted at 30 "C for the times shown in medium containing 3 mM MgCL,, 1 mM EDTA, and the concentrations of agonist, antagonist, and nucleotide indicated below. Each time point in each section of this experiment represents vesicles that contained a total of 15 fmol of receptor and 260 fmol of G,. A, agonist-stimulated nucleotide dissociation from vesicles. Vesicles (0.65 ml) were preincubated at 30 "C in the presence of 1 PM (-)-isoproterenol and 0.2 NM 13H]GTP for 8 min. This is 4 times as long as required to attain steady-state agonist-stimulated binding (see Fig. 1). Propranolol was added to yield a 10 GM final concentration for an additional 2 min. At this point, binding of 3H-nucleotide was 120 fmol, about 46% of total G,. The vesicles were then immediately chilled, diluted with 4 ml of ice- cold 20 mM NaHepes (pH 8.0), 100 mM NaCI, 3 mM MgCl2, I mM EDTA ("buffer") and centrifuged for 2 h at 50,000 rpm in a Beckman 70.1 Ti rotor. The surface of the tight pellet was washed gently 3 times with 0.5 ml of ice-cold buffer, and the pellet was resuspended in 0.325 ml of buffer. At zero time, the vesicles were diluted 10-fold into assay medium at 30 "C that contained either 0.2 pM GTPrS (0 and 0) or no added nucleotide (e and 0). In either case, the dissociation of nucleotide was followed in the presence of either 10 PM (-)- isoproterenol (0 and +) or 0.5 pM (-)-propranolol (0 and 0). Data points represent duplicate binding assays. B, agonist-stimulated GTPyS binding to vesicles. Vesicles (0.65 ml) were preincubated to allow agonist-stimulated nucleotide binding and then washed exactly as described above, except that unlabeled GTP was used. At zero time, two aliquots of vesicles were diluted 10-fold into assay medium that contained 0.2 NM [35S]GTPyS (A and A) and either 10 p~ (-)- isoproterenol (A) or 0.5 NM (-)-propranolol (A). The binding of f3%]

GTPyS was followed over 30 min at 30 "C. Two other aliquots were diluted into assay medium at 30 "C that contained isoproterenol but no added nucleotide (. and 0) to initiate agonist-stimulated dissociation of any bound GDP. At 27 min, either 0.01 volume of (-)- alprenolol (0, 200 +M final concentration) or an equal volume of water (W) was added to these samples and the binding of [35SS]GTPrS was initiated at 30 min by the addition of 0.01 volume of [36SS]GTPrS (0.2 PM final concentration). Data points represent duplicate (A and A) or triplicate (. and 0) binding assays. C , first order replots of data on the dissociation of GDP and the association of GTPrS. Data on GDP dissociation (panel A ) are replotted as the differences in the amount of bound i3H]GDP with and without isoproterenol (0). The agonist-promoted decreases in bound nucleotide at 10 and 30 min were averaged in order to give BT, which was used to normalize the agonist-promoted dissociation data at shorter times. Agonist-stimulated dissociation was complete by 10 min. Data were fit by un- weighted least squares analysis of In{(& - B)/BT] uersus t, where B equals the decrement in nucleotide binding due to isoproterenol a t any time t. Data from panel B are replotted as the increment in nucleotide binding due to isoproterenol (A, 0 to 30 min; H, 30 to 60

tive concentrations for stimulation of GTPase activity and the rate of nucleotide binding.

Agonist-stimulated Nucleotide Binding to Receptor-G, Ves- icles from Which Endogenous GDP Has Dissociated-Cassel and Selinger ( 2 ) and others (4-8) originally proposed that agonists promote binding of GTP by clearing bound GDP from G,. Because G. was assumed to be unliganded as purified (33), we assumed in our studies of GTPyS binding that the agonist-liganded receptor promoted the binding reaction itself (19). Recently, however, it has been shown that rabbit hepatic G,, purified by the method of Sternweis et al. (22) , and similar preparations of other G proteins, contain nearly stoichiometric amounts of bound GDP.3 Thus, the rate of nucleotide binding to G, and steady-state hydrolysis of GTP might in fact be limited simply by the rate of dissociation of bound GDP, since significant amounts of GDP remain bound to G, through the reconstitution process (Fig. 5 and other data not shown). To test this possibility more rigorously, receptor-(;, vesicles were allowed to exchange endogenous GDP for i3H] GDP by incubating them with isoproterenol and [3H]GTP. The vesicles were washed and the release of t3H]GDP (Fig. 5A) was then compared with the binding of [35S]GTPyS, which was added either immediately or after the [3H]GDP had largely dissociated (Fig. 5B). Several aspects of this experiment are significant. First, the rate of agonist-stimulated [35S]GTPyS binding subsequent to the agonist-stimulated release of prebound [3H]GDP was no faster than that observed when both processes occurred concurrently. There was no burst of basal binding after 30 min of dissociation, as might have been expected if clearing GDP were rate-limiting prior to diffusion-controlled, receptor-independent binding of GTPyS. Furthermore, the rate of agonist-stimulated dissociation of GDP was about twice as fast as that of the binding of [3'S]GTPyS, whether the GTP+ was added immediately or after GDP had dissociated (Fig. 5C). While some residual [3H] GDP remained bound after the initial 30-min incubation in the absence of added nucleotide, it was less than the amount of [35SJGTP$3 that bound under the influence o f agonist. From these data, it seems clear that agonist stimulated both nucleotide association and dissociation, further suggesting that nucleotide dissociation per se is not necessarily the rate- limiting step in agonist-stimulated GTPase activity. It should also be noted that the rates of substrate binding and product release determined here are consistent with the turnover numbers for steady-state hydrolysis shown in Table I for non- DTT-treated vesicles.

Multiple Effects of Mg' on the Agonist-stimulated GTPase Reuction-Isoproterenol-stimulated GTPase activity required Mg"+ or other divalent cations. The effects of M e were observed over a wide range of concentrations, with stimulation noticeable at about 0.1 FM and continuing until about 5 mM (Fig. 6). At least three actions of Mg2+ were detected. To investigate the effects at low concentrations of M e , we prepared vesicles as described under Experimental Procedures

K. M. Ferguson, personal communication.

rnin). Plateau levels of bound [35S]GTPrS were averaged to give BT, which was used to normalize binding at shorter times. The data were fit as described above. In addition to the data shown in panel B, agonist-stimulated binding of [35S]GTPyS to vesicles that were neither preincubated nor centrifuged is also shown (V). BT equaled 100 ( W , 60 (A), 53 (U), and 24 fmol (0). The calculated rate constants for GTPrS binding were 0.56 (V), 0.52 (A), and 0.49 min" (m); the rate constant for f3H]GDP dissociation was 1.3 rnin".

1662 Regulation of Catecholamine-stimulated GTPase

100 r I

= 20

I Mg”1 (log M 1

FIG. 6. Influence of M 3 + concentration on agonist-stimulated GTPase activity. Each time point represents DTT-treated vesicles containing 8.7 fmol of receptor and 175 fmol of G., assayed in triplicate at 30 ‘C for 5 min. Assays were performed in medium containing 0.2 FM [T-~*PJGTP, from 0.3 to 30 mM Me, and from 0.1 to 20 mM EDTA. The free M$+ concentration was calculated using an apparent dissociation constant of 0.4 pb! for the EDTA.Mg complex (pH 8.0; see Ref. 29). The binding of M$+ to nucleotide (Kd,epp, - 30 FM) and to lipid (&pp - 1 mM) was assumed to be negllglble. Data were obtained in the absence (0) or presence of 10 PM (-)-isoproterenol. Shown are the basal rate (0) and the agonist- promoted increase above basal (@).

0 2 4 6 8 10 Time (min)

FIG. 7. Influence of M e on agonist-stimulated [LY-’~P]GTP binding. Vesicles (DTT-treated) were incubated in duplicate in the presence of 0.2 PM [ce3*P]GTP and either 3 mM Me and 1 mM EDTA (0) or 0.3 mM M e and 10 mM EDTA ( 0 ; free Me concen- tation - 20 nM). Under each condition, either 10 g M or no (-)- isoproterenol was included in order to obtain the incremental binding of nucleotide due to agonist. Each assay point represents 4 fmol of receptor and 60 fmol of G,.

except that Mg2’ was omitted from the buffers. Analysis for M$+ of the buffers that were used for these experiments indicated the presence, of 1.8 p~ total M$’ in the final assay volumes. Thus, in the.presence of 1 mM EDTA, the calculated free concentration was 0.7 nM (assuming a Kd for the complex of 0.4 p~ (29)). Under these conditions, GTPase activity was virtually zero, and 60-100% of the agonist-stimulated nucleotide binding at steady-state was accounted for as GTP (Table 11), indicating that G.. GTP was not hydrolyzed to G, GDP in the absence of divalent cation. This result confirms data previously obtained using isolated G, (17). While it has been difficult to determine the concentration dependence of hydrolysis precisely, we found that the agonist-stimulated binding of [y-3ZP]GTP was half-maximally inhibited at 2-10 nM free Mg2+ (data not shown), and this is probably a good estimate of the concentration of Mg2t requirement for hydrolysis.

[Mg2*] (log MI

FIG. 8. Influence of M P concentration on the initial rate of agonist-stimulated fa-SZP]GTP binding. Vesicles ( D m - treated) were assayed in duplicate at 30 “C for 10 s. Assays were performed in medium containing 0.2 PM [a-32P]GTP, from 0.3 to 10 mM M&L, and from 0.1 to 20 mM EDTA. The free Mg2+ concentration was calculated as described in the legend to Fig. 6. Data were obtained in the absence (0) or presence of 10 p~ (-)-isoproterenol. Shown is the incremental binding of nucleotide due to agonist (e). Each assay point represents 8.7 fmol of receptor and 175 fmol of G..

As shown in Figs. 7 and 8 and Table 11, a requirement for Mg2+ could not be detected for the agonist-stimulated, high- affinity binding of GTP to G, (or, by inference, for the release of bound GDP). GTP bound to roughly the same extent with or without added Mg”’, although the rate of binding was about 3-fold slower in its absence. In the absence of Mg“, binding that was observed in the absence of agonist was quite low, so the agonist-stimulated increment shown was close to the total amount. The influence of MgZ+ on the rate of [LX-~’P]GTP binding to receptor-(;. vesicles is shown in Fig. 8. The agonist- stimulated rate of GTP binding was increased half-maximally at 2-3 p~ Mg+, roughly 1000-fold above the concentration required for the rapid hydrolysis of GTP to GDP. This discrepancy suggests that there are a t least two distinct sites and/or mechanisms through which Mg2+ contributes to the overall agonist-stimulated GTPase and probably accounts for the overall effect of M$+ shown in Fig. 6. The data of Fig. 8 are also consistent with the concentration dependence observed by Asano et al. (19, 34) for the agonist-stimulated binding of [36S]GTPrS.

The stimulatory effect of high concentrations of M$+ (10- 200 mM) on the receptor-independent GTPase activity of G, (see Ref. 17) is not shown well in Fig. 6, but has been observed in other experiments using receptor-(=, vesicles.

DISCUSSION

This study provides an initial characterization of the ways in which the &adrenergic receptor regulates the levels of G,- nucleotide intermediates during agonist-stimulated, steady- state hydrolysis of GTP. Because the concentration of the activated species G,. GTP is crucial for the stimulation of adenylate cyclase, control of the overall GTPase rate is less important than the rates of the partial reactions involved. For the present study, we have used reconstituted vesicles composed of purified receptors and G. and two phospholipids to identify basic mechanisms whereby the receptor regulates these reactions. While the system is artificial in several re- spects, it has allowed certain detailed experiments that have not been feasible in native membranes, which generally contain several G proteins, receptors, and associated proteins at variable concentrations.

The agonist-stimulated GTPase cycle described here is outlined in Fig. 9. Both the high-affinity binding of GTP and the dissociation of tightly bound GDP are promoted by the agonist-liganded receptor in events that appear to be mecha-


/f "h J

GDP R-HCATALYZED

BY 4 FIG. 9. Schematic representation of intermediate steps in

the receptor-stimulated CTPase cycle. Initial low affinity binding of GTP to form [G..GTP] (see Ref. 34) is followed by a receptor- catalyzed isomerization to the high-affinity, activated species, G.*. GTP. The activated species is deactivated by hydrolysis of the bound GTP to form G.. GDP, which dissociates to yield free G. in a receptor- catalyzed process. The rate constant for hydrolysis (kh) is approximately 4 min" and the rate constant for agonist-stimulated, high- affinity GTP binding (kJ and release of GDP (&) are approximately 1 min" without pretreatment with DTT.

nistically very similar. Agonist-stimulated GTP binding and GDP release display similar absolute rates and similar ECso values for agonists, both require a low concentration of Mp", and in both reactions a single receptor can service multiple molecules of G, in a formally catalytic fashion. We therefore find it most convenient to consider the &adrenergic receptor as promoting an "open" conformation for G, rather than favoring a liganded versus unliganded, or active versus inac- tive, state. Such an open state would allow relatively free binding or release of guanine nucleotides. This concept helps to rationalize the fact that the receptor can catalyze the release of GDP, the binding of GTP or GTPyS and concom- itant activation of G,, or the release of GTPyS4 and concom- itant deactivation. It is a slightly more general kinetic concept than the idea that receptor promotes nucleotide exchange, as suggested by Fung and Stryer for transducin (35), because it allows for the dissociation of receptor from unliganded G,, as might be observed at low concentrations of agonist. It is also consistent with the ability of agonist to promote the dissociation of nucleotide in the absence of added nucleotide (Fig. 5A).

Our findings, particularly those of Fig. 5, argue strongly that receptor does not facilitate binding solely by promoting the release of previously bound nucleotide to free the binding site, as was initially proposed (2, 8). Several previous experiments argued against this idea, most notably the simple observation made by Birnbaumer and co-workers (9) that different guanine nucleotides cause the activation of adenylate cyclase to occur with different rates. This finding would be reconciIabIe with a purely dissociation-based mechanism only by extreme conjecture. In Fig. 5, we showed that the rate and extent of binding of GTPyS were not altered if more than half of the total bound GDP, representing essentially all of the receptor-sensitive pool of G., was allowed to dissociate before the binding reaction was initiated. Furthermore, the amount of G. that bound GTPyS in the rapid, agonist-

T. Asano, D. R. Brandt, and E. M. Ross, unpublished observa- tions.

stimulated binding reaction was larger than the amount of G,.GDP that was present at the initiation of the binding reaction (compare Fig. 5A with 5B, right side).5 Last, the agonist-stimulated rate of dissociation of GDP was about twice as fast as the rate of GTPrS binding, suggesting that release is not rate-limiting (Fig. X).

Because the rate of agonist-stimulated GDP dissociation and GTP binding differed by only a factor of 2 in this reconstituted system, we cannot say which might predominate in uiuo. However, a single mechanism of receptor action, the opening of G, to nucleotide binding or release, appears capable of explaining both events and both favor the increase in the active G.. GTP species.

The release of GDP during the agonist-stimulated GTPase reaction follows a different pattern than that previously observed for isolated G, (17). In that study, only about one-third of the GDP product remained bound to G, in a form stable enough to be measured. This fraction was estimated both from the fraction of G, measurable as G, + GDP and from the ratio of the initial rate of G.. GDP formation to the GTPase rate (see Ref. 17). Because the /3y subunits of G. stabilize the binding of GDP to the a subunit, we proposed that the assayable GB. GDP represents the intact a& trimer and that the two-thirds that dissociated rapidly was probably bound to the free a subunit. Because 70-80% of the receptor-sensitive G. was found as G,. GDP in the vesicles, it may be that dissociation of a and Py occurs to a lesser extent under these circumstances. Direct measurement of this dissociation equi- librium will be difficult but should be quite rewarding.

The actual hydrolysis of G,-bound GTP to GDP was not reguIated by receptor (Fig. 21, but the rate was much slower when G, was reconstituted into phospholipid vesicles than what we had previously observed using isolated G. (17). Be- cause of the slower rate of hydrolysis and the stimulation of GTP binding by agonist, up to 35% of the receptor-accessible G, was bound to GTP under conditions of rapid, steady-state hydrolysis. Because G, seems to be in stoichiometric excess over adenylate cyclase itself (8), this fraction is adequate to account for complete activation of the cyclase by agonist plus GTP. We do not know the reason for this effect of reconstitution on the rate of hydrolysis. The rate that is observed in the receptor-G, vesicles is similar to that previously described for transducin, the G protein of the retinal rod outer segment (36).

The hydrolysis of G.-bound GTP did not occur in the presence of EDTA and absence of added M e . We do not know if this requirement for M$+ is for the actual chemical hydrolytic reaction or for a change in the conformation of G. that precedes rapid, metal-independent hydrolysis. M e , or a similar cation, is typically required for nucleoside triphos- phatase reactions, but usually in the micromolar range and usually as a ligand of the nucleotide. The EC50 for Mg2+,

Interpretation of this experiment is predicated to some extent on the assumption that there exists a fixed fraction of reconstituted G. molecules that is sensitive to receptor and a fraction that is not. This assumption is supported by the data of Table I and Figs. 3 and 4 and by the previous work of Asano et al. (19). However, if one were to assume that all the G, in the vesicles used for the experiment of Fig. 5 was initially bound to GDP and that none of the fraction that did not exchange with [3H]GDP dissociated either during the washing of the vesicles or during the initial 30-min incubation with agonist, then there would remain enough G..GDP to account for the amount of agonist-stimulated GTPyS binding shown in Fig. 5B. We feel that these assumptions are implausible, since they require that the receptors suddenly "recruit" a population of G.. GDP that had been insensitive to receptor during the initial L3H]GTP binding reaction and for the first 30-min incubation.

1664 Regulation of Catecholamine-stimulated CTPase

estimated a t 2-10 nM, seems low for such a role. We have no data on the specificity of this reaction among other divalent cations.

M$+ is also involved in enhancing the agonist-stimulated rate of GTP binding to G.. While we are aware of the difficulty in calculating these very low concentrations of free Mg2+ given the available data, it seems clear that a higher concentration of M e is needed to accelerate binding (Fig. 8) than is needed for the release of Pi from bound GTP. These two sites of action for M$+ seem to account for the overall dependence on M e of the steady-state GTPase reaction that is shown in Fig. 6. We cannot directly relate either of these M$+-dependent events to the Me-dependent GTPase activity of Lubrol-solubilized Gi observed by Sunyer et al. (32). The absolute rates described by those authors were quite low, the effective concentration of Mg2+ was claimed to be substantially below nanomolar, and multiple turnovers of the GTPase were not observed in most cases.

The overall agonist-stimulated GTPase reaction is clearly complex, and we have barely approached the specific roles of the phospholipid bilayer and the G protein subunits in its control. Regardless, the present description seems internally consistent, delineates two specific actions of the receptor, and accounts for the maintenance of an activated G.. GTP complex in sufficient amounts to reconcile GTPase activity with the activation of adenylate cyclase by agonists plus GTP.

Acknowledgments-We are grateful to Dr. Michael Nicar, of the Department of Internal Medicine, for performing the atomic absorption spectroscopic determinations of magnesium in our reagents. We thank Kenneth M. Ferguson, of this department, for assaying GDP bound to several preparations of G. and for critical discussion of this work. We also thank C. Eric Cordero and Beth Strifler for their excellent technical assistance, and Jennifer Cordero for preparation of the manuscript.

REFERENCES

1. Cassel, D., and Selinger, Z. (1976) Biochim. Biophys. Acta 452,

2. Cassel, D., and Selinger, Z. (1977) J. Cyclic Nucleotide Res. 3,

3. Cassel, D., and Selinger, Z. (1977) J. Cyclic Nucleotide Res. 3,

4. Cassel, D., and Selinger, Z. (1978) Proc. Natl. Acad. Sci. U. S. A.

5. Pike, L. J., and Lefkowitz, R. J. (1981) J. Biol. Chem. 256,2207-

6. Murayama, T., and Ui, M. (1984) J . Biol. Chem. 259,761-769 7. Blume, A. J., and Foster, C . J. (1976) J. Biol. Chem. 251, 3399-

8. Ross, E. M., and Gilman, A. G. (1980) Annu. Reu. Biochem. 49,

9. Birnbaumer, L., Swartz, T. L., Abramowitz, J., Mintz, P. W., and

538-551

393-406

11-22

75,4155-4159

2212

3404

533--564

Iyengar, R. (1980) J. Biol. Chem. 255, 3542-3551 10. Levitzki, A. (1980) FEBS Lett. 115,9-10 11. Neer, E. J., and Salter, R. S. (1981) J. Biol. Chem. 256, 12102-

12107 12. Smigel, M. D., ROSS, E. M., and Gilman, A. G. (1984) in Cell

Membranes, Methods and Reuiews (Elson, E. L., Frazier, W. A., and Glaser, L., eds) Vol. 2, pp. 247-294, Plenum Press, New York

13. Brandt, D. R., Asano, T., Pedersen, S. E., and Ross, E. M. (1983) Biochemistry 22,4357-4362

14. Cerione, R. A., Codina, J., Benovic, J. L., Lefkowitz, R. J., Birnbaumer, L., and Caron, M. G. (1984) Biochemktry 23, 4.51 9-4525

15. Wheeler, G. L., and Bitensky, M. W. (1977) Proc. Natl. Acud. Sci. . - - - - - - -

U. S. A. 74,4238-4242 16. Fung, B. K.-K., Hurley, J. B., and Stryer, L. (1981) Proc. Natl.

Acad. Sci. U. S. A. 78, 152-156 17. Brandt, D. R., and Ross, E. M. (1985) J. Biol. Chem. 260, 266-

272 18. Johnson, R. A., and Walseth, T. F. (1979) Adu. Cyclic Nucleotide

Res. 10, 135-168 19. Asano, T., Pederson, S. E., Scott, C. W., and Ross, E. M. (1984)

Biochemistry 23,5460-5467 20. Engel, G., Hoyer, D., Berthold, R., and Wagner, H. (1981) Nuu-

nyn-Schmiedeberg’s Arch. Pharmacol. 317, 277-285 21. Diehl, H. (1964) Cukein, Calmagite, and o,o’-Dihydroxyazoben-

zene. Titrimetric, Colorimetric and Fluorometric Reagents for Calcium and Magnesium, G. Frederick Smith Chemical Co., Columbus, OH

22. Sternweis, P. C., Northup, J . K., Smigel, M. D., and Gilman, A. G. (1981) J. Biol. Chem. 256, 11517-11526

23. Shorr, R. G. L., Strohsacker, M. W., Lavin, T. N., Lefkowitz, R. J., and Caron, M. G. (1982) J. Biol. Chem. 257,12341-12350

24. Pedersen, S. E., and Ross, E. M. (1982) Proc. Natl. Acad. Sci. U.

25. Caron, M. G., Srinivasan, Y., Pitha, J., Kociolek, K., and Lefkow-

26. Asano, T.. Brandt. D. R.. Pedersen. S. E. and Ross. E. M. (1984)

S. A. 79, 7228-7232

itz, R. J. (1979) J . Biol. Chem. 254, 2923-2927

Ad”.’ Cyclic Nucleotide Res. 19, i7-56 , ,

27. May, D. C.. Ross. E. M., Gilman. A. G.. and Smicrel. M. D, (1985) Biol. Chem. 260, 15829-15833 ’

- I . ,

28. Pedersen, S. E., and Ross, E. M. (1985) J. Biol. Chem. 260,

29. O’Sullivan, W. J. (1969) in Datu for Biochemical Research (Daw- son, R. M. C., Elliott, D. C., and Jones, K. M., eds) pp. 426- 427, Oxford University Press, London

30. Fleming, J. W., and Ross, E. M. (1980) J. Cyclic Nucleotide Res.

31. Schaffner, W., and Weissmann, C. (1973) Anal. Biochem. 56,

32. Sunyer, T., Codina, J., and Birnbaumer, L. (1984) J . Biol. Chem.

33. Northup, J. K., Smigel, M. D., and Gilman, A. G. (1982) J. Biol.

34. Asano, T., and Ross, E. M. (1984) Biochemistry 23, 5467-5471

14150-14157

6,407-419

502-514

259,15447-15451

Chem. 257, 11416-11423

35. Fung, B. K.-K., and Stryer, L. (1980) Proc. Natl. Acud. Sci. U. S. A. 77, 2500-2504

36. Navon, S. E., and Fung, B. K.-K. (1984) J. Biol. Chem. 259, 6686-6693

the of chemistry vol. 261, no. 4, of pp. 1656 … 1986 jbc.pdf · catecholamine-stimulated gtpase...

Documents