Eur. J. Biochem. 204, 725-731 (1992)
0FEBS 1992
Dissociation of guanosine 5’-[y-thio]triphosphate from guanine-nucleotide-binding regulatory proteins in native cardiac membranes Regulation by nucleotides and muscarinic acetylcholine receptors Gerhard HILF’, Christine KUPPRION I , Thomas WIELAND’ and Karl H. JAKOBS’
’ Pharmakologisches Institut der Universitat Heidelberg, Federal Republic of Germany ’ Institut fur Pharmakologie dcr Universitat GH Essen, Fcderal Republic of Germany (Received July 19/October 29, 1991) - EJB 91 1369
Binding of the poorly hydrolyzable GTP analog, guanosine 5’-[y-thioltriphosphate (GTP[S]), to purified guanine-nucleotide-bindingregulatory proteins (G proteins) has been shown to be nonreversible-in the presence of millimolar concentrations of Mg2 . In porcine atrial membranes, binding of [”S]GTP[S] to G proteins was stable in the presence of 1 mM Mgz+.However, either large dilution or, even more strongly, addition of unlabelled guanine nucleotides, in the potency order, GTP[S] > GTP > guanosine 5’-[b,y-iminoltriphosphate> GDP guanosine S-[b-thio]diphosphate > GMP, markedly enhanced the observed dissociation, with 20 - 30% of bound [35S]GTP[S]being released by unlabelled guanine nucleotide within 20 min at 25°C. Most interestingly, dissociation of [35Ss]GTP[S]was rapidly and markedly stimulated by agonist (carbachol) activation of cardiac muscarinic acetycholine receptors. Carbachol-stimulated release of [”S]GTP[S] was strictly dependent on the presence of Mg2+ and an unlabelled guanine nucleotide. Although having different potency and efficiency in releasing [3 5S]GTP[S] from the membranes by themselves, the guanine nucleoside triphosphates and diphosphates studied, at maximally effective concentrations, promoted the carbachol-induced dissociation to the same extent, while GMP and ATP were ineffective. GTP[S]binding-saturation experiments indicated that one agonist-activated muscarinic acetylcholine receptor can cause release of bound GTP[S] from three to four G proteins. The data presented indicate that binding of GTP[S] to G proteins in intact membranes, in contrast to purified G proteins, is reversible, and that agonist-activated receptors can even, either directly or indirectly, interact with GTP[S]bound G proteins, resulting in release of bound guanine nucleoside triphosphate. +
Heart muscarinic acetylcholine (mACh) receptors are coupled via guanine-nucleotide-bindingregulatory proteins (G proteins) to various effector systems [l - 91. Agonist-liganded receptors interacting with G proteins initiate signal transduction apparently by facilitating the release of GDP from inactive G proteins, which can then bind GTP. In the active, GTPbound state, the G proteins are dissociated from the receptors and interact with and regulate effector systems. The signal is switched off by the hydrolysis of bound GTP by the intrinsic GTPase activity of G proteins. The resulting GDP-bound G proteins can re-enter the cycle of activation and deactivation (for reviews see [lo, 111). This process of receptor-induced G-protein activation has been analyzed in cardiac and other membrane systems primarily by demonstrating receptor-stimulated GTPase activity of G proteins [I2 - 141 and receptor-induced binding of the poorly ___..
hydrolyzable GTP analog, guanosine 5‘-[y-thioltriphosphate (GTP[S]), to G proteins [15, 161. Only in some membrane systems, receptor-stimulated release of GDP from G proteins has been demonstrated [17 - 191. Interestingly, agonist-activated adrenoceptors have been reported to stimulate the release of guanosine 5’-[P,y-imino]triphosphate (Gpp[NHjp), another poorly hydrolyzable GTP analog, from G proteins in turkey erythrocyte and platelet membranes [20, 211. In contrast, binding of the GTP analog, GTP[S], to purified G proteins was shown to be non-reversible, provided MgZf was present at millimolar concentrations [22].In the present study, we investigated the binding of GTP[S] to G proteins in sarcolemma1 membrane preparations of porcine atria. We report here that, in contrast to purified G proteins, binding of GTP[S] to G proteins in intact membranes is, at least in part, reversible, and that the release of GTP[S] from G proteins in cardiac membranes is regulated by mACh receptors in a guaninenucleotide-dependent manner.
Correspondence l o K. H. Jakobs, Institut fur Pharmakologie der Universitat G H Essen, Universitatsklinikum, Hufelandstrasse 55, W4300 Essen, Federal Republic of Germany MATERIALS AND METHODS Abbreviations. mACh receptor, muscarinic acetylcholine receptor; G protein, guanine-nucleotide-bindingregulatory protein; GTP[S], Materials guanosine 5’-[y-thio]triphosphate; GDP[S], guanosine 5’-[/3-thio]ATP, adenosine S’-[P,y-imino]triphosphate(App[NH]p), diphosphate; Gpp[NH]p, guanosine 5’-[/3,y-imino]triphosphate; GTP, GDP, GMP, GTP[S], Gpp[NH]p and guanosine 5’-[pApp[NH]p, adenosinc, 5’-[P,y-imino]triphosphate.
726 I
thioldiphosphate (GDP[S]) were purchased from Boehringer Mannheim (Mannheim, FRG). Atropine sulfate and carbachol were from Sigma (Deisenhofen, FRG). [3'S]GTP[S] (1.2 - 3.4 Ci/pmol) and the radiolabelled mACh receptor antagonist, I -q~inuclinidyl[phenyl-4-~H]benzilate (45.7 Ci/ mmol) were obtained from DuPont-NEN (Bad Homburg, FRG). Poly(ethy1eneimine)-cellulose-F-coated TLC plastic sheets were from Merck (Darmstadt, FRG). Instagel was from Canberra Packard (Frankfurt, FRG) and Quicksafe A scintillalor from Zinsser Analytic (Frankfurt, FRG).
I
1
f 2&
Preparation of porcine atrial membranes Sarcolemmal membranes of porcine atria were prepared essentially as described [23] with some modifications [14] by sucrose-density-gradient centrifugation and were stored at - 70°C. The membranes contained about 1.4 pmol mACh receptor/mg protein, as determined by binding of the labelled mACh receptor antagonist. Protein concentration was measured according to Bradford [24], using bovine IgG as standard.
Labelling of cardiac membranes with [35S]GTP[S] Aliquots of the membrane preparations were thawed, washed with buffer A (25 mM Tris/HCl, pH 7.4,l mm EDTA, 2 mM MgC12, 1 mM dithiothreitol) and centrifuged for 30 min at 30000 g. The pellets were resuspended with a syringe needle in buffer A to a protein concentration of 4mg/ml. After addition of 0.2 mM App[NH]p, the membranes were incubated at 25°C for 3 min. Binding of [35S]GTP[S]to G proteins was performed in a total volume of 0.5 ml buffer A containing 1 mg membrane protein, 0.1 mM App[NH]p and 10 nM [35S]GTP[S]for 15 rnin at 25°C. The binding reaction was stopped by the addition of 0.75ml ice-cold buffer A containing, in addition, 10 pM GTP and centrifugation for 30 min at 30000 g. The supernatant was discarded and the pellet was washed four times with 1 ml buffer A to remove free ["S]GTP[S]. The membranes containing 2-3 pmol [35S]GTP[S] bound/mg protein were frozen in aliquots in liquid nitrogen and stored at -70°C without significant dissociation of [3'S]GTP[S].
0' 0
15
30
L5
60
I
Incubation time (minl
Fig. 1. Time course of [35S]CTP[S] release from G proteins in porcine atrial membranes. Porcine atrial membranes labelled with [3sS]GTP[S] (20000 cpm/80 pl) were incubated for up to 60 rnin at 25'C in the absence (A)or presence of 0.1 pM GTP[S]( 0 )or 0.1 pM GTP[S] plus 100 pM carbachol (m). [3sS]GTP[S]released from themembranes was determined as described in Materials and Methods.
in the filtrates. The concentration of free Mg2+was calculated as described [25].
Thin layer chromatography The released radioactive material was analyzed by TLC as described [26], using poly(ethy1eneimine)-cellulose-F-coated plastic sheets. 10-pl aliquots of the filtrates of the release assays were applied to the sheets and the plates were developed in 1.5 M potassium phosphate buffer, pH 3.4. The radioactive spots were visualized by autoradiography and quantified by scraping off the cellulose and measuring radioactivity in the scintillation fluid.
RESULTS
In the presence of 1 mM Mg2+ and in the absence of unlabelled nucleotides, only about 3% of bound [35S]GTP[S] dissociated from labelled porcine atrial membranes during incubation for 60 min at 25°C (Fig. 1). After an initial peak, [35S]GTP[S]-releaseassay the amount of free radioactive material rapidly decreased Aliquots of labelled membranes were thawed and washed during the following 5 min of incubation and slowly increased with 1 ml buffer A by centrifugation for 30 min at 30000 g . thereafter up to 60 min of incubation. This pattern may be The pellets were resuspended in buffer A and diluted to 0.6 - caused by rapid binding of free [35S]GTP[S]still being present 1.2 cpm/nl, corresponding to 150- 300 pg protein/ml. If not from the labelling procedure, followed by a very slow release. otherwise indicated, the release assay was performed in tripli- Addition of unlabelled GTP[S] (0.1 pM) markedly increased cate for 15 min at 25'C in a total volume of 100 pl, containing the release of [35S]GTP[S],up to about 25% of total bound buffer A, 0.1 pM unlabelled GTP[S] and 20000-60000 cpm radioactivity after 45-60 min. The rapid phase of GTP[S]induced release of [35S]GTP[S]was followed by a slow dis[3 'S]GTP[S]-labelled membrane proteins. Thereafter, proteinbound and released free [35S]GTP[S]were separated by rapid sociation which was linear over 45 - 135 min of incubation filtration of 80 pl of the reaction mixture through nitrocellu- (Fig. 2). In the presence of 0.1 pM GTP[S], the cholinergic \ose filters (pore size 0.45 pm, Schleicher & Schuell, Dassel, agonist, carbachol (100 pM), stimulated the release of FRG). The filters were washed with 1.5 ml 50 mM Tris/HCl, [35S]GTP[S]up to twofold, without an apparent lag phase. pH 7.4, containing 5 mM MgC12. The filtrate and washing The carbachol-stimulated dissociation was maximal after solution were trapped in scintillation vials and radioactivity 5 min and remained constant up to 45 min of incubation. measured in 4.5 ml Instagel. Alternatively, the filters were After 45 min, when about 40% of the total radioactivity counted in Quicksafe A scintillator to determine [35S]GTP[S] was found in the filtrate, the release curve in the presence of bound to the membranes. Free [35S]GTP[S]remaining from carbachol plus unlabelled GTP[S] reached a plateau (Fig. 2). the labelling procedure (about 2% of total radioactivity The dissociation of [3'S]GTP[S] was not significantly enadded) was determined by filtration of samples without hanced by higher concentrations of unlabelled GTP[S]. Adfurther incubation and was subtracted from the radioactivity dition of carbachol for 15 min after treatment with GTP[S]
721
-
10 a
& 3.1 g
I
0
L
2.8 m -
IT m
2.2
"0
30
60
90
120
m
Incubation time lminl
Fig. 2. Time course of control and carbachol-stimulated release of [3sS]GI'PlS]. Release of ["S]lGTP[S] from labelled membranes (53000 cpm/80 pl) was determined in the presence of 0.1 pM GTP[S] (H) or 0.1 pM GTP[S] plus 100 pM carbachol(0) for up to 135 min at25'C. At theindicated timepoints(5-120min),carbachol(lOO pM final concentration) was added to reaction mixtures containing GTP[S] and the release of [3sS]GTP[S]was determined after 15 min of incubation (0). Standard deviations of triplicates are indicated by vertical bars.
7
6
5
L
3
-log([Carbachol]/M)
Fig. 4. mACh-receptor-stimulated release of [35SlCTPlS]from G proteins in cardiac membranes. Release of [3sS]GTP[S] from labelled mcmbranes (20000 cpm/80 pl) was determined in the presence of 0.1 pM unlabelled GTP[S] at increasing concentrations of carbachol in the absence (H)and presence of 1 pM atropinc ( 0 ) .
0
0 1 0 9 8 7 6 -log (GTP[S] c o n h a t i o n /M)
5
Fig. 5. Influence of GTPlS] on the release of ["S]GTP[S]. Release of [35S]GTP[S]from labelled membranes (17000 cpm/80 pl) was determined for 10 min at 25°C at increasing concentrations of unlabelled GTP[S] in the absence ( 0 )and presence of 100 1 M carbachol (H).
Fig. 3. Identification of the released radioactive material. Release of ["S]GTP[S] from labclled membranes was determined for 15 min at 25 C without (control) and with 0.1 pM unlabelled GTP[S] or 0.1 pM GTP[S] plus 100 pM carbachol. After filtration through nitrocellulose filters, the washing procedurc was omitted in order to avoid dilution of the samples. 10 pI of the filtrates were analyzed by TLC. No further spots were detectcd on the X-ray film. The radioactivity was quantified by scraping off and measuring the poly(ethy1eneimine)cellulose. (A) Autoradiogram of the GTP[S] rcgion of a TLC plate; (B) quantification of the radioactive spots. Standard, authentic [3sS]GTP[S]added directly to the platcs.
(0.1 pM) for 5 - 120 min decreased thc carbachol-induced release of [35S]GTP[S]in a time-dependent manner. The maximal release induced by unlabelled CTP[S] plus carbachol, however, was independent of the time of preincubation. These data indicate that only From a fixed portion of the G proteins in cardiac membranes [3sSs]GTP[S]can be released and that agonist-activated mACh receptors essentially accelerate this release reaction.
In order to investigate the molecular property of the released radioactive material, the filtrates wcre analyzed by TLC, using authentic [35S]GTP[S]as standard. The released radiolabelled compound comigrated with ["S]GTP\S] (Fig. 3). Additional spots were not detected, neither by exposure to X-ray films nor by scraping off and measuring radioactivity on the complete surface of the TLC plates. Quantification of the radioactive spots showed a twofold increase in ["S]GTP[S] following stimulation by carbachol, identical to that observed in the filter assays. Carbachol stimulated the dissociation of [3sSS]CTP[S]in a concentration-dependent manner. Half-maximal and maximal increase was observed at 0.3 pM and 10 pM carbachol, respcctively (Fig. 4). The mACh receptor antagonist, atropine (1 lM), competitively antagonized the carbachol effect. In the absence of carbachol, atropine by itself significantly reduced dissociation of bound [3'S]CTP[S] by 15 -20%. Stimulation of ["S]GTP[S] release by the agonist-activated mACh receptor was strictly dependent on the presence of an unlabelled guanine nucleotide. GTP[S] (Fig. 5) and other guanine nucleotides (Fig. 6) promoted the carbachol-stimu-
728
7 5 5 L L 3 3 [-log(C/M)] Control GTP[S] GTP Gpp[NH]P GDP GDP[S] GMP ATP
Fig. 6. Influence of various nucleotides on basal and carbachol-stimulated release of [35SlGTP[Sl. Incubation of cardiac membranes labelled with [35S]CiTP[S] (20000 cpm/80 pl) was performed for 15 min a1 25 C without (0) and with 100 KM carbachol (a) in the absence (control) and presence of maximally effective concentrations of various nucleotides. The standard deviation of triplicates is given by vertical bars.
lated dissociation of [3'S]GTP[S] from labelled sarcolemmal membranes in a concentration-dependent manner. Without guanine nucleotides in the reaction mixture, the release of [35S]GTP[S]was not stimulated by carbachol. Unlabelled GTP[S] promoted the carbachol-stimulated release of [3'S]GTP[S] at concentrations >, 1 nM, providing maximal carbachol stimulation at 100 nM. At maximally effective concentrations, GTP[S] (0.1 pM), GTP (10 pM), Gpp[NH]p (lOpM), GDP (100pM) and GDP[S] (100pM) promoted the carbachol stimulation of [35S]GTP[S]release to a similar extent. GMP and ATP were ineffective up to 1 mM (Fig. 6). In the absence of carbachol, the release of [35S]GTP[S]was differentially regulated by guanine nucleotides. GTP[S] was not only most potent but also most efficient, followed by GTP 2 Gpp[NH]p > GDP 2 GDP[S] > GMP 2 ATP. A possible role of the guanine nucleotides in causing release of [3 'S]GTP[S] is that these nucleotides block rebinding of dissociated [35S]GTP[S]to G proteins. Indeed, the concentration-response curves of unlabelled GTP[S] in inhibiting binding of [35S]GTP[S]to G proteins and leading to dissociation of bound [35S]CTP[S] from G proteins were superimposable (Fig. 7). However, GDP[S] and GMP, at concentrations inhibiting the binding of [35S]GTP[S]completely, caused only 30-50% of the dissociation of bound [j5S]GTP[S] observed with unlabelled GTP[S]. Dilution of the labelled membranes should also decrease the rebinding of dissociated ["S]GTP[S]. To investigate whether carbachol can stimulate the release of [35S]GTP[S]in the absence of guanine nucleotides, under conditions which decrease the rebinding, the membrane preparation was diluted 50-fold with the reaction mixture. Release of [35S]GTP[S]was measured by filtration of the complete reaction mixture and measuring the radioactivity on the nitrocellulose membranes. Apparently, due to the dilution, the release of [35S]GTP[S] observed in the absence of unlabelled GTP[S] was enhanced (Fig. 8). There was an initial rapid phase followed by a slow dissociation. After 30 min of incubation, about 20% bound [35S]GTP[S] was released, a number very similar to that observed upon addition of unlabelled guanine nucleotides in the non-diluted assay system (see e.g. Fig. 1). However, in contrast to this condition, carbachol (100 pM) did not
-log([Guanine nucleotide]/M)
Fig. 7. Inhibition of binding and stimulation of release of 13%]CTPlS] by various guanine nucleotides. Binding (0, A, 0) and release ( 0 , A , W ) of [3sS]GT'P[S]were carried out in the same reaction mixturc including the amount of [3'S]GTP[S] (free or bound) for 15 min at 25°C at increasing concentrations of GTP[S] (0, O ) , GDP[S] ( A , A) and GMP (0,W). The left ordinate gives percentage of maximal inhibition (at 10 pM GTP[S]) of binding of [35S]GTP[S].On the right ordinate, [3sS]GTP[S]released from labelled membranes is given as a percentage of total bound radioactivity added to the assay.
0
10 20 Incubation time lmin)
3c
Fig. 8. Influence of dilution on the dissociation of bound [35SlCTPlSl. Membranes labelled with [3sS]GTP[S] were diluted 50-fold (25000 cpm/5 ml) and incubated without ( A ) and with 100 pM carbachol ( O ) ,0.1 pM GTP[S] ( 0 )or 0.1 pM GTP[S] plus 100 pM carbachol ( W ) for up to 30 min at 25°C. 5 ml reaction mixture was filtered and membrane-bound radioactivity was determined by measuring the nitrocellulose filters.
enhance the release of bound [35S]GTP[S].When unlabelled GTP[S] (0.1 pM) was added, which increased the extent of [35S]GTP[S]release (about 30% after 30 min), carbachol caused a further substantial release of [35S]GTP[S],reaching about 50% after 30 min of incubation. Labelling of the membranes with increasing concentrations of GTP[S] increased the amount of GTP[S] bound to membrane G proteins (Fig. 9, control). After binding for 20 min, dissociation of bound GTP[S] was immediately initiated by the addition of 10 pM GTP[S] or 10 pM GTP[S] plus 100 pM carbachol and measured after 20 min of incubation. The amount of dissociated GTP[S] caused by addition of GTP[S] alone increased with increasing concentrations of GTP[S] in the labelling period. In relative terms (percentage of initial binding), dissociation of GTP[S] decreased from
729 2.5 20 I
f-
e
1.5
3 0
n
-
3 10 a I-
0
0.5
0
0
O 25 50 75 100 GTP[S] in preincubotion (nM1
Fig. 9. Influence of the GTPISj-binding site occupancy on the release of bound GTPISI. Sarcolemmal membranes (3 pg/tube) were labelled with [35S]GTP[S] (0.3 nM) and increasing concentrations of unlabelled GTP[S] (up to 100 nM) for 20 min. Thereafter, the release of bound GTP[S] was immediately initiated by the addition of 10 pM GTP[S] added without or with 100 pM carbachol. Membrane-bound GTP[S] was determined directly after the labelling period ( A ) or after the release reaction performed for 20 min in the absence (0) and presence of carbachol (0).
almost 50% dissociation at 0.3 nM GTP[S] to about 30% at 10 nM GTP[S] in the binding reaction, without further change up to 100 nM GTP[S]. The amount of GTP[S] being released under control of the mACh receptor was already maximal at about 1 nM GTP[S] in the labelling period and remained constant up to 100 nM GTP[S]. In relative terms, the carbachol effect decreased from almost 30% higher dissociation at 0.3 nM GTP[S] to only 8% at 100 nM GTP[S] in the binding reaction. These data indicate that only a fixed maximal portion of the GTP[S]-binding sites are regulated by mACh receptors. The reaction mixture contained about 40 pM mACh receptors, leading to release of GTP[S] from maximally about 150 pM binding sites upon addition of carbachol. In addition to guanine nucleotides, the carbachol-stimulated release of bound [3'S]GTP[S] from cardiac membranes required the presence of Mg". At concentrations below 10 pM free MgZt, dissociation of [35S]GTP[S]from G proteins was observed, even without addition of unlabelled guanine nucleotide and carbachol (Fig. 10). At 10- 1000 pM Mg2+,however, the dissociation remained constant at a low level. The increase in release of bound [3sSs]GTP[S]from the membranes caused by GTP[S] (0.1 pM) alone was not dependent on the presence and concentration of Mg2+.The MgZ' concentration response curves for [35S]GTP[S]release in the absence and presence of unlabelled GTP[S] alone were parallel. In contrast, the carbachol-stimulated release of [3'S]GTP[S] was strictly dependent on the presence of Mg2+. Half-maximal and maximal stimulation by carbachol occurred in the presence of about 0.5 pM and 10 pM Mg2+, respectively, with no decrease in carbachol effect being observed up to 1 mM Mg2+. DISCUSSION The measurement of receptor-stimulated binding of the poorly hydrolyzable GTP analog GTP[S] to G proteins is used to study the activation of G proteins in reconstituted systems [27, 281 and in membrane preparations [15, 161. For example,
-L c
O
8
7
6
5
L
3
-log([Free Mg2 ]/M)
Fig. 10. Influence of MgZ* on the release of [3sSlGTP[Slfrom cardiac membranes. In order to remove Mgz+, membranes labelled with [35S]GTP[S]were washed three times with Tris/HCl buffer (25 mM, pH 7.4) containing 1 mM dithiothreitol, and in the first washing 20 mM EDTA, in the second no EDTA, and in the third 2 mM EDTA. The reaction was performed in the presence of 2 mM EDTA, the calculated concentrations of MgCI2 and labelled membranes (19000 cpm/80 ~ 1for ) 10 min at 25°C. The release of [35S]GTP[S]was determined with no further additions (A)in the presence of 0.1 pM GTI'[S] ( 0 )or with 0.1 pM GTP[S] plus 100 pM carbachol (W).
in sarcolemmal membranes of porcine atria, the mACh receptor agonist, carbachol, caused a rapid and marked increase in binding of GTP[S] to G proteins [15]. It is generally assumed that this receptor action is due to the release of GDP from G proteins, to which GTP[S] then can bind [lo, 111. As shown herein, the agonist-liganded mACh receptor can also markedly and rapidly stimulate the release of GTP[S] from G proteins in cardiac membranes. The effect of carbachol was concentration dependent and specific for mACh receptors, since it was competitively antagonized by atropine. Interestingly, in the absence of carbachol, atropine by itself inhibited release of GTP[S] from G proteins by about 20%, which suggests that agonist-free, empty mACh receptors may have an effect on interacting G proteins, which can be prevented by antagonist binding. Two requirements were apparently necessary for the agonist-activated receptor action: first, the presence of Mg2+ and, second, addition of a guanine nucleotide. Low Mg2+ concentrations are known to destabilize the binding of GTP[S] to G proteins [22, 291 (Fig. 10). In order to obtain stable binding and, thus, minimal release of GTP[S] under control conditions, we routinely used 1 mM free Mg2+in the reaction mixture. However, even at this concentration of Mg", the carbachol effect was still maximal. Half-maximal mACh receptor-stimulated release of bound GTP[S] was observed at about 0.5 pM free MgZf,very similar t o the concentration of Mg2 required for carbachol stimulation of low-K, GTPase in these membranes [14], which apparently reflects the receptor-stimulated release of G-protein-bound GDP followed by GTP binding and hydrolysis [lo, 111. Thus, receptor-stimulated release of GDP as well as of GTP[S] required Mg2+ at very similar concentrations, pointing to a similar mechanism for these receptor actions. It is widely accepted that binding of GTP[S] t o G proteins is essentially irreversible in the presence of M g 2 + ,as shown for purified G proteins [22].However, in intact cardiac membranes there was a rapid and marked release of G-proteinbound [35S]GTP[S] upon addition of unlabelled guanine nucleotides or large dilution of the labelled membranes. One +
730 important difference between the studies using purified G proteins [22] and the data presented here on membrane G proteins is that the purified G proteins were studied in detergent-containing solutions, which may cause a trapping of the nucleotide in its binding site and, in addition, may facilitate dissociation of G proteins into f l y dimers and GTP[S] liganded r subunits. In fact, there was minimal if any release of [35S]GTP[S]from G proteins solubilized from cardiac membranes with the detergent, 3-[(3-~holamidopropyl)dimethylammonia]-1-propanesulfonate (data not shown). The most likely mechanism by which unlabelled guanine nucleotides increase the release of bound [35S]GTP[S]is by competition with released [35S]GTP[S] for rebinding. In fact, concentrations of GTP[S], GDP[S] and GMP, which blocked the binding of [35S]GTP[S]maximally, also lead to maximal dissociation of bound [35S]GTP[S] from labelled membranes. However, GDP[S] and GMP were much less effective than unlabelled GTP[S] in promoting the release of bound [35S]GTP[S].For example, GDP[S] was only about 50% as effective as GTP[S] in releasing [35S]GTP[S],while both guanine nucleotides were equally effective in inhibiting binding of [35S]GTP[S] to G proteins, although at different concentrations. Both the release and the binding assay were performed for the same period of time (15 min) and with the same amount of free or membrane-bound [35S]GTP[S]. Therefore, different binding kinetics of the unlabelled guanine nucleotides cannot be the reason for the reduced efficiencies of GDP[S] and GMP in causing release of bound [35S]GTP[S]. Thus, the effect of the guanine nucleotides in releasing G-protein-bound GTP[S] can only in part be explained by inhibition of rebinding of released GTP[S]. Additional mechanisms (see below) are apparently involved in this reaction. The most exciting question raised from the data presented herein is how agonist-activated mACh receptors can stimulate the release of bound GTP[S] from G proteins. It is generally assumed that receptors are uncoupled from G proteins with bound guanine nucleoside triphosphate, particularly with Gpp[NH]p and GTP[S] [lo, 11, 30, 311. Similar data as for many other receptors have been reported for cardiac mACh receptors [32]. On the other hand, in intact turkey erythrocyte membranes and in systems reconstituted of partially purified G proteins and receptors, P-adrenoceptors have been shown to interact with Gpp[NH]p-liganded G proteins [33- 351, even with higher affinity than with GDP-liganded G proteins [35]. In contrast, the receptors did not apparently interact with and release GTP[S] from GTP[S]-liganded G proteins [35, 361, whereas Gpp[NH]p was rapidly released [20, 361. Thus, although a variety of data suggest that guanine-nucleosidetriphosphate-bound G proteins are uncoupled or even physically dissociated from receptors, receptors can apparently efficiently interact with Gpp[NH]-liganded G proteins. The data presented herein indicate that receptors can even interact with GTP[S]-bound G proteins. The interaction of agonist-liganded receptors with GTPIS]-bound G proteins is either direct or indirect. If receptors interact directly with GTP[S]-liganded G proteins and, thereby, cause release of bound guanosine nucleoside triphosphate, an additional consequence is that these interacting GTP[S]-liganded G-protein c1 subunits are associated with Gprotein By dimers, known to be required for receptor-Gprotein interaction [lo, 11, 37, 381. Thus, it may be speculated that binding of an agonist to its receptor causes dissociation of any guanine nucleotide from the interacting trimeric G protein, irrespective of whether GDP or GTP[S] is bound and,
thereby, initiates replacement of the bound by a free guaninc nucleotide. On the other hand, it is also feasible that oligomeric G proteins [39,40] are the basis for the observed receptor-stimulated release of GTP[S] and that receptors interact indirectly with the GTP[S]-liganded G proteins. Agonist-liganded receptors may primarily interact with GTP[S]-free G proteins. In a second step, this receptor G-protein interaction somehow induces GTP[S]-bound G proteins to release GTP[S] in a guanine-nucleotide-dependent manner. Such an indirect mechanism of action seems to be corroborated by the finding that addition of a guanosine nucleoside diphosphate or triphosphate was absolutely necessary for agonist-stimulated release of bound GTP[S]. Although large dilution caused release of GTP[S] to a similar extent as observed with, for example, GDP or GDP[S], carbachol did not increase the observed release. This finding suggests that the effect of guanine nucleotides in promoting the carbachol-stimulated release of GTP[S] is not only due to inhibition of rebinding of released GTP[S] but involves an active process. Analysis of GTP[S]binding-saturation experiments indicated that cardiac membranes apparently contain two binding sites for GTP[S], one with high affinity and low capacity and another one with lower affinity and high capacity, and that effects of mACh receptor agonists on GTP[S] binding (increase or release) are preferentially observed with the high-affinity binding component (G. Hilf and K. H. Jakobs, unpublished observations). Binding of guanine nucleotides to the low-affinity component may modify the binding affinity of GTP[S] to the high-affinity sites, resulting in release of bound GTP[S]. Thus, the observed receptor-stimulated release of G-protein-bound GTP[S] may be due to a direct or indirect interaction of receptors with GTP[S]-liganded G proteins or even a combination of these two reactions. Considering the number of mACh receptors present in the membrane preparation used and the amount of G proteins from which GTP[S] can be released under stimulation by carbachol, it has to be concluded that one agonist-activated mACh receptor can cause rapid release of GTP[S] from three to four G proteins. The time course of receptor-stimulated GTP[S] binding and release, and the maximal amount of G proteins being controlled by mACh receptors to bind or release GTP[S], were very similar [15]. These findings suggest that receptor-stimulated release of G-protein-bound GDP, followed by binding of GTP[S], may follow similar reactions, as discussed above for receptor-stimulated release of G-proteinbound GTP[S]. Taken together, evidence is presented that, in contrast to a widely accepted view, binding of the poorly hydrolyzable GTP analog, GTP[S], to G proteins is reversible in intact membranes and that agonist-activated receptors can, either directly or indirectly, interact with GTP[S]-bound G proteins, thereby accelerating the dissociation of the bound guanine nucleoside triphosphate. Although the mechanism(s) by which agonist-liganded mACh receptors stimulate the dissociation of GTP[S] from G proteins in cardiac membranes remains incompletely understood at the moment, the measurement of agonist-stimulated release of G-protein-bound GTP[S] may provide a further sensitive tool to study receptor - Gprotein interactions also in other membrane and receptor systems. This work was supported by the Deutsche Forsc,hungsgemeinschaft.
731
REFERENCES 1. Watanabe, A. M., McConndughey, M. M., Strawbridgc, R. A., Fleming, M. W., Jones, L. & Besch, R. A. (1978) J. Biol. Chem. 353,4833-4836. 2. Jakobs, K . H., Aktories, K. & Schultz, G. (1979) NaunynSchmitdeherg's Arch Pharmacol. 310, 113 - 119. 3. Brown, S. L. & Brown, J. H. (1983) Mol. Pharmucol. 24, 351 356. 4. Brown, J. H. & Brown, S. L. (1984) J . B i d . Chem. 259. 37773781. 5. Pfaffinger, P. J., Martin, J. M., Hunter, D. D., Nathdnson, N . M. & tlille, B. (1985) Nuturr317, 536-538. 6. Breitwieser, G. E. & Szabo, G. (1985) Nature 317, 538-540. 7. Cohen-Armon, M. & Sokolovsky, M. (1986) J . B i d . Chem. 261, 12498-12505. 8. Cohen-Armon, M., Garly, H. & Sokolovsky, M. (1988) Riochemistry 27, 368 - 374. 9. Danilcnko, M. P., Panchenko, M. P., Yesirev, 0. V. & Tkachuk, V. A. (1990) Eur. J . Biochem. 194, 155-160. 10. Gilman, A. G. (1987) Annu. Rev. Biochem. 56,615-649. 11. Birnbaumer, L., Abramowitz, J. & Brown, A. M. (1990) Biochim. Biophys. Acta 1031, 163--224. 12. Cassel, D. & Selinger, Z. (1976) Biochim. Biophys. Acfu452,538 551. 13. Aktories, K., Schultz, G. & Jakobs, K. H. (1982) Mol. Pharmacol. 21,336- 342. 14. Hilf, G. & Jakobs, K. H. (1989) Eur. J. Pharmacol. Mol. Pharmocol. 172, 155- 163. 15. Hilf, G., Gierschik, P. & Jakobs, K. H. (1989) Eur. J . Biochetn. 186,725-731. 16. Gierschik, P., Moghtader. R., StrdUb, C., Dieterich, K. &Jakobs, K. H. (1991) Eur. J . Biochem. 197,725-732. 17. Cassel, D. & Selinger, Z. (1978) Proc. Nut1 Acud. Sci. USA 75, 41 55 -41 59. 18. Pike, J. L. & Lefkowitz, R. J. (1981) J . Biol. Cjiem. 256, 22072212. 19. Muraydma, T. & Ui, M. (1984) J . B i d . Chern. 259,761 -769.
20. Casscl, D. & Selinger, Z.(1977) J . Cyclic Nucleotide Res, 3, 1 I 22. 21. Michel, T. & Lefkowitz, R. J. (1982) .I. Biol. C'hem.257, 1355713563. 22. Higashijima, T., Ferguson, K. M., Sternweis, P. C., Smigel, M. D. & Gilman, A. G. (1987) J . Biol. Chem. 262, 762-766. 23. Petersen, G. L. & Schimerlik, M. I. (1984) Prep. Biochem. 14, 33 - 74. 24. Bradford, M. M. (1976) Anal. Biochem. 72,248-254. 25. Bartfai, T. (1979) Adv. Cyclic Nucleotide Res. 10, 219-242. 26. Seifert, R., Roscnthal, W., Schultz, G., Wieland, T., Gierschik. P. & Jakobs, K. H. (1988) Eur. J . Biochem. 175, 51 -55. 27. Asdno, T., Pedersen, S. E., Scott, C. W. & Ross, E. M. (1984) Biochemistry 23, 5460- 5467. 28. May, D. C. & Ross, E. M. (1988) Biochemistry 27,4888-4893. 29. Fcrguson, K. M., Higashijima, T., Smigel, M. D. & Gilman, A . G. (1986) J. Biol. Chem. 261,7393-7399. 30. Maguire, M. E., Van Arsdale, P. M. & Gilman, A. Cr. (1976) Mol. Pharmacol. 12, 335-339. 31. Limbird, L. E., Gill, D. M. & Lefkowitz, R. J. (1980) Yruc. Natl Acad. Sci. USA 77, 775 - 779. 32. Kosenberger, L. B., Yamamura, H. I. & Roeske, W. R. (1980) J . Biol. Chem. 255,820 - 823. 33. Sevilla, N. & Levitzki, A. (1977) FEES Lett. 76, 129- 134. 34. Arad, H., Rimon, G. & Levitzki, A. (1981) J . B i d . Chem. 256, 1593- 1597. 35. Citri, Y. & Schramm, M. (1982) J. B i d . Chem. 257, 1325713262. 36. Cassel, D. & Selinger, Z. (1977) Biochem. Biophys. Res. C O ~ V ~ M ~ . 77,868-873. 37. Fung, B. K. K . (1983) J . Biol. Chem. 258,10495-10502. 38. Florio, V. A. & Sternweis, P. C. (1985) J . B i d . Chem. 260,34773483. 39. Vaillancourt, R. R., Dhdnasekardn, N., Johnson, G . L. & Ruoho, A. E. (1990) Proc. Natl Acud. Sci.U S A 87, 3645-3649. 40. Nakamura, S.-I. & Rodbell, M. (1990) Proc. Nut1 A c d . Sci. USA 87,6413-6417.
0FEBS 1992
Dissociation of guanosine 5’-[y-thio]triphosphate from guanine-nucleotide-binding regulatory proteins in native cardiac membranes Regulation by nucleotides and muscarinic acetylcholine receptors Gerhard HILF’, Christine KUPPRION I , Thomas WIELAND’ and Karl H. JAKOBS’
’ Pharmakologisches Institut der Universitat Heidelberg, Federal Republic of Germany ’ Institut fur Pharmakologie dcr Universitat GH Essen, Fcderal Republic of Germany (Received July 19/October 29, 1991) - EJB 91 1369
Binding of the poorly hydrolyzable GTP analog, guanosine 5’-[y-thioltriphosphate (GTP[S]), to purified guanine-nucleotide-bindingregulatory proteins (G proteins) has been shown to be nonreversible-in the presence of millimolar concentrations of Mg2 . In porcine atrial membranes, binding of [”S]GTP[S] to G proteins was stable in the presence of 1 mM Mgz+.However, either large dilution or, even more strongly, addition of unlabelled guanine nucleotides, in the potency order, GTP[S] > GTP > guanosine 5’-[b,y-iminoltriphosphate> GDP guanosine S-[b-thio]diphosphate > GMP, markedly enhanced the observed dissociation, with 20 - 30% of bound [35S]GTP[S]being released by unlabelled guanine nucleotide within 20 min at 25°C. Most interestingly, dissociation of [35Ss]GTP[S]was rapidly and markedly stimulated by agonist (carbachol) activation of cardiac muscarinic acetycholine receptors. Carbachol-stimulated release of [”S]GTP[S] was strictly dependent on the presence of Mg2+ and an unlabelled guanine nucleotide. Although having different potency and efficiency in releasing [3 5S]GTP[S] from the membranes by themselves, the guanine nucleoside triphosphates and diphosphates studied, at maximally effective concentrations, promoted the carbachol-induced dissociation to the same extent, while GMP and ATP were ineffective. GTP[S]binding-saturation experiments indicated that one agonist-activated muscarinic acetylcholine receptor can cause release of bound GTP[S] from three to four G proteins. The data presented indicate that binding of GTP[S] to G proteins in intact membranes, in contrast to purified G proteins, is reversible, and that agonist-activated receptors can even, either directly or indirectly, interact with GTP[S]bound G proteins, resulting in release of bound guanine nucleoside triphosphate. +
Heart muscarinic acetylcholine (mACh) receptors are coupled via guanine-nucleotide-bindingregulatory proteins (G proteins) to various effector systems [l - 91. Agonist-liganded receptors interacting with G proteins initiate signal transduction apparently by facilitating the release of GDP from inactive G proteins, which can then bind GTP. In the active, GTPbound state, the G proteins are dissociated from the receptors and interact with and regulate effector systems. The signal is switched off by the hydrolysis of bound GTP by the intrinsic GTPase activity of G proteins. The resulting GDP-bound G proteins can re-enter the cycle of activation and deactivation (for reviews see [lo, 111). This process of receptor-induced G-protein activation has been analyzed in cardiac and other membrane systems primarily by demonstrating receptor-stimulated GTPase activity of G proteins [I2 - 141 and receptor-induced binding of the poorly ___..
hydrolyzable GTP analog, guanosine 5‘-[y-thioltriphosphate (GTP[S]), to G proteins [15, 161. Only in some membrane systems, receptor-stimulated release of GDP from G proteins has been demonstrated [17 - 191. Interestingly, agonist-activated adrenoceptors have been reported to stimulate the release of guanosine 5’-[P,y-imino]triphosphate (Gpp[NHjp), another poorly hydrolyzable GTP analog, from G proteins in turkey erythrocyte and platelet membranes [20, 211. In contrast, binding of the GTP analog, GTP[S], to purified G proteins was shown to be non-reversible, provided MgZf was present at millimolar concentrations [22].In the present study, we investigated the binding of GTP[S] to G proteins in sarcolemma1 membrane preparations of porcine atria. We report here that, in contrast to purified G proteins, binding of GTP[S] to G proteins in intact membranes is, at least in part, reversible, and that the release of GTP[S] from G proteins in cardiac membranes is regulated by mACh receptors in a guaninenucleotide-dependent manner.
Correspondence l o K. H. Jakobs, Institut fur Pharmakologie der Universitat G H Essen, Universitatsklinikum, Hufelandstrasse 55, W4300 Essen, Federal Republic of Germany MATERIALS AND METHODS Abbreviations. mACh receptor, muscarinic acetylcholine receptor; G protein, guanine-nucleotide-bindingregulatory protein; GTP[S], Materials guanosine 5’-[y-thio]triphosphate; GDP[S], guanosine 5’-[/3-thio]ATP, adenosine S’-[P,y-imino]triphosphate(App[NH]p), diphosphate; Gpp[NH]p, guanosine 5’-[/3,y-imino]triphosphate; GTP, GDP, GMP, GTP[S], Gpp[NH]p and guanosine 5’-[pApp[NH]p, adenosinc, 5’-[P,y-imino]triphosphate.
726 I
thioldiphosphate (GDP[S]) were purchased from Boehringer Mannheim (Mannheim, FRG). Atropine sulfate and carbachol were from Sigma (Deisenhofen, FRG). [3'S]GTP[S] (1.2 - 3.4 Ci/pmol) and the radiolabelled mACh receptor antagonist, I -q~inuclinidyl[phenyl-4-~H]benzilate (45.7 Ci/ mmol) were obtained from DuPont-NEN (Bad Homburg, FRG). Poly(ethy1eneimine)-cellulose-F-coated TLC plastic sheets were from Merck (Darmstadt, FRG). Instagel was from Canberra Packard (Frankfurt, FRG) and Quicksafe A scintillalor from Zinsser Analytic (Frankfurt, FRG).
I
1
f 2&
Preparation of porcine atrial membranes Sarcolemmal membranes of porcine atria were prepared essentially as described [23] with some modifications [14] by sucrose-density-gradient centrifugation and were stored at - 70°C. The membranes contained about 1.4 pmol mACh receptor/mg protein, as determined by binding of the labelled mACh receptor antagonist. Protein concentration was measured according to Bradford [24], using bovine IgG as standard.
Labelling of cardiac membranes with [35S]GTP[S] Aliquots of the membrane preparations were thawed, washed with buffer A (25 mM Tris/HCl, pH 7.4,l mm EDTA, 2 mM MgC12, 1 mM dithiothreitol) and centrifuged for 30 min at 30000 g. The pellets were resuspended with a syringe needle in buffer A to a protein concentration of 4mg/ml. After addition of 0.2 mM App[NH]p, the membranes were incubated at 25°C for 3 min. Binding of [35S]GTP[S]to G proteins was performed in a total volume of 0.5 ml buffer A containing 1 mg membrane protein, 0.1 mM App[NH]p and 10 nM [35S]GTP[S]for 15 rnin at 25°C. The binding reaction was stopped by the addition of 0.75ml ice-cold buffer A containing, in addition, 10 pM GTP and centrifugation for 30 min at 30000 g. The supernatant was discarded and the pellet was washed four times with 1 ml buffer A to remove free ["S]GTP[S]. The membranes containing 2-3 pmol [35S]GTP[S] bound/mg protein were frozen in aliquots in liquid nitrogen and stored at -70°C without significant dissociation of [3'S]GTP[S].
0' 0
15
30
L5
60
I
Incubation time (minl
Fig. 1. Time course of [35S]CTP[S] release from G proteins in porcine atrial membranes. Porcine atrial membranes labelled with [3sS]GTP[S] (20000 cpm/80 pl) were incubated for up to 60 rnin at 25'C in the absence (A)or presence of 0.1 pM GTP[S]( 0 )or 0.1 pM GTP[S] plus 100 pM carbachol (m). [3sS]GTP[S]released from themembranes was determined as described in Materials and Methods.
in the filtrates. The concentration of free Mg2+was calculated as described [25].
Thin layer chromatography The released radioactive material was analyzed by TLC as described [26], using poly(ethy1eneimine)-cellulose-F-coated plastic sheets. 10-pl aliquots of the filtrates of the release assays were applied to the sheets and the plates were developed in 1.5 M potassium phosphate buffer, pH 3.4. The radioactive spots were visualized by autoradiography and quantified by scraping off the cellulose and measuring radioactivity in the scintillation fluid.
RESULTS
In the presence of 1 mM Mg2+ and in the absence of unlabelled nucleotides, only about 3% of bound [35S]GTP[S] dissociated from labelled porcine atrial membranes during incubation for 60 min at 25°C (Fig. 1). After an initial peak, [35S]GTP[S]-releaseassay the amount of free radioactive material rapidly decreased Aliquots of labelled membranes were thawed and washed during the following 5 min of incubation and slowly increased with 1 ml buffer A by centrifugation for 30 min at 30000 g . thereafter up to 60 min of incubation. This pattern may be The pellets were resuspended in buffer A and diluted to 0.6 - caused by rapid binding of free [35S]GTP[S]still being present 1.2 cpm/nl, corresponding to 150- 300 pg protein/ml. If not from the labelling procedure, followed by a very slow release. otherwise indicated, the release assay was performed in tripli- Addition of unlabelled GTP[S] (0.1 pM) markedly increased cate for 15 min at 25'C in a total volume of 100 pl, containing the release of [35S]GTP[S],up to about 25% of total bound buffer A, 0.1 pM unlabelled GTP[S] and 20000-60000 cpm radioactivity after 45-60 min. The rapid phase of GTP[S]induced release of [35S]GTP[S]was followed by a slow dis[3 'S]GTP[S]-labelled membrane proteins. Thereafter, proteinbound and released free [35S]GTP[S]were separated by rapid sociation which was linear over 45 - 135 min of incubation filtration of 80 pl of the reaction mixture through nitrocellu- (Fig. 2). In the presence of 0.1 pM GTP[S], the cholinergic \ose filters (pore size 0.45 pm, Schleicher & Schuell, Dassel, agonist, carbachol (100 pM), stimulated the release of FRG). The filters were washed with 1.5 ml 50 mM Tris/HCl, [35S]GTP[S]up to twofold, without an apparent lag phase. pH 7.4, containing 5 mM MgC12. The filtrate and washing The carbachol-stimulated dissociation was maximal after solution were trapped in scintillation vials and radioactivity 5 min and remained constant up to 45 min of incubation. measured in 4.5 ml Instagel. Alternatively, the filters were After 45 min, when about 40% of the total radioactivity counted in Quicksafe A scintillator to determine [35S]GTP[S] was found in the filtrate, the release curve in the presence of bound to the membranes. Free [35S]GTP[S]remaining from carbachol plus unlabelled GTP[S] reached a plateau (Fig. 2). the labelling procedure (about 2% of total radioactivity The dissociation of [3'S]GTP[S] was not significantly enadded) was determined by filtration of samples without hanced by higher concentrations of unlabelled GTP[S]. Adfurther incubation and was subtracted from the radioactivity dition of carbachol for 15 min after treatment with GTP[S]
721
-
10 a
& 3.1 g
I
0
L
2.8 m -
IT m
2.2
"0
30
60
90
120
m
Incubation time lminl
Fig. 2. Time course of control and carbachol-stimulated release of [3sS]GI'PlS]. Release of ["S]lGTP[S] from labelled membranes (53000 cpm/80 pl) was determined in the presence of 0.1 pM GTP[S] (H) or 0.1 pM GTP[S] plus 100 pM carbachol(0) for up to 135 min at25'C. At theindicated timepoints(5-120min),carbachol(lOO pM final concentration) was added to reaction mixtures containing GTP[S] and the release of [3sS]GTP[S]was determined after 15 min of incubation (0). Standard deviations of triplicates are indicated by vertical bars.
7
6
5
L
3
-log([Carbachol]/M)
Fig. 4. mACh-receptor-stimulated release of [35SlCTPlS]from G proteins in cardiac membranes. Release of [3sS]GTP[S] from labelled mcmbranes (20000 cpm/80 pl) was determined in the presence of 0.1 pM unlabelled GTP[S] at increasing concentrations of carbachol in the absence (H)and presence of 1 pM atropinc ( 0 ) .
0
0 1 0 9 8 7 6 -log (GTP[S] c o n h a t i o n /M)
5
Fig. 5. Influence of GTPlS] on the release of ["S]GTP[S]. Release of [35S]GTP[S]from labelled membranes (17000 cpm/80 pl) was determined for 10 min at 25°C at increasing concentrations of unlabelled GTP[S] in the absence ( 0 )and presence of 100 1 M carbachol (H).
Fig. 3. Identification of the released radioactive material. Release of ["S]GTP[S] from labclled membranes was determined for 15 min at 25 C without (control) and with 0.1 pM unlabelled GTP[S] or 0.1 pM GTP[S] plus 100 pM carbachol. After filtration through nitrocellulose filters, the washing procedurc was omitted in order to avoid dilution of the samples. 10 pI of the filtrates were analyzed by TLC. No further spots were detectcd on the X-ray film. The radioactivity was quantified by scraping off and measuring the poly(ethy1eneimine)cellulose. (A) Autoradiogram of the GTP[S] rcgion of a TLC plate; (B) quantification of the radioactive spots. Standard, authentic [3sS]GTP[S]added directly to the platcs.
(0.1 pM) for 5 - 120 min decreased thc carbachol-induced release of [35S]GTP[S]in a time-dependent manner. The maximal release induced by unlabelled CTP[S] plus carbachol, however, was independent of the time of preincubation. These data indicate that only From a fixed portion of the G proteins in cardiac membranes [3sSs]GTP[S]can be released and that agonist-activated mACh receptors essentially accelerate this release reaction.
In order to investigate the molecular property of the released radioactive material, the filtrates wcre analyzed by TLC, using authentic [35S]GTP[S]as standard. The released radiolabelled compound comigrated with ["S]GTP\S] (Fig. 3). Additional spots were not detected, neither by exposure to X-ray films nor by scraping off and measuring radioactivity on the complete surface of the TLC plates. Quantification of the radioactive spots showed a twofold increase in ["S]GTP[S] following stimulation by carbachol, identical to that observed in the filter assays. Carbachol stimulated the dissociation of [3sSS]CTP[S]in a concentration-dependent manner. Half-maximal and maximal increase was observed at 0.3 pM and 10 pM carbachol, respcctively (Fig. 4). The mACh receptor antagonist, atropine (1 lM), competitively antagonized the carbachol effect. In the absence of carbachol, atropine by itself significantly reduced dissociation of bound [3'S]CTP[S] by 15 -20%. Stimulation of ["S]GTP[S] release by the agonist-activated mACh receptor was strictly dependent on the presence of an unlabelled guanine nucleotide. GTP[S] (Fig. 5) and other guanine nucleotides (Fig. 6) promoted the carbachol-stimu-
728
7 5 5 L L 3 3 [-log(C/M)] Control GTP[S] GTP Gpp[NH]P GDP GDP[S] GMP ATP
Fig. 6. Influence of various nucleotides on basal and carbachol-stimulated release of [35SlGTP[Sl. Incubation of cardiac membranes labelled with [35S]CiTP[S] (20000 cpm/80 pl) was performed for 15 min a1 25 C without (0) and with 100 KM carbachol (a) in the absence (control) and presence of maximally effective concentrations of various nucleotides. The standard deviation of triplicates is given by vertical bars.
lated dissociation of [3'S]GTP[S] from labelled sarcolemmal membranes in a concentration-dependent manner. Without guanine nucleotides in the reaction mixture, the release of [35S]GTP[S]was not stimulated by carbachol. Unlabelled GTP[S] promoted the carbachol-stimulated release of [3'S]GTP[S] at concentrations >, 1 nM, providing maximal carbachol stimulation at 100 nM. At maximally effective concentrations, GTP[S] (0.1 pM), GTP (10 pM), Gpp[NH]p (lOpM), GDP (100pM) and GDP[S] (100pM) promoted the carbachol stimulation of [35S]GTP[S]release to a similar extent. GMP and ATP were ineffective up to 1 mM (Fig. 6). In the absence of carbachol, the release of [35S]GTP[S]was differentially regulated by guanine nucleotides. GTP[S] was not only most potent but also most efficient, followed by GTP 2 Gpp[NH]p > GDP 2 GDP[S] > GMP 2 ATP. A possible role of the guanine nucleotides in causing release of [3 'S]GTP[S] is that these nucleotides block rebinding of dissociated [35S]GTP[S]to G proteins. Indeed, the concentration-response curves of unlabelled GTP[S] in inhibiting binding of [35S]GTP[S]to G proteins and leading to dissociation of bound [35S]CTP[S] from G proteins were superimposable (Fig. 7). However, GDP[S] and GMP, at concentrations inhibiting the binding of [35S]GTP[S]completely, caused only 30-50% of the dissociation of bound [j5S]GTP[S] observed with unlabelled GTP[S]. Dilution of the labelled membranes should also decrease the rebinding of dissociated ["S]GTP[S]. To investigate whether carbachol can stimulate the release of [35S]GTP[S]in the absence of guanine nucleotides, under conditions which decrease the rebinding, the membrane preparation was diluted 50-fold with the reaction mixture. Release of [35S]GTP[S]was measured by filtration of the complete reaction mixture and measuring the radioactivity on the nitrocellulose membranes. Apparently, due to the dilution, the release of [35S]GTP[S] observed in the absence of unlabelled GTP[S] was enhanced (Fig. 8). There was an initial rapid phase followed by a slow dissociation. After 30 min of incubation, about 20% bound [35S]GTP[S] was released, a number very similar to that observed upon addition of unlabelled guanine nucleotides in the non-diluted assay system (see e.g. Fig. 1). However, in contrast to this condition, carbachol (100 pM) did not
-log([Guanine nucleotide]/M)
Fig. 7. Inhibition of binding and stimulation of release of 13%]CTPlS] by various guanine nucleotides. Binding (0, A, 0) and release ( 0 , A , W ) of [3sS]GT'P[S]were carried out in the same reaction mixturc including the amount of [3'S]GTP[S] (free or bound) for 15 min at 25°C at increasing concentrations of GTP[S] (0, O ) , GDP[S] ( A , A) and GMP (0,W). The left ordinate gives percentage of maximal inhibition (at 10 pM GTP[S]) of binding of [35S]GTP[S].On the right ordinate, [3sS]GTP[S]released from labelled membranes is given as a percentage of total bound radioactivity added to the assay.
0
10 20 Incubation time lmin)
3c
Fig. 8. Influence of dilution on the dissociation of bound [35SlCTPlSl. Membranes labelled with [3sS]GTP[S] were diluted 50-fold (25000 cpm/5 ml) and incubated without ( A ) and with 100 pM carbachol ( O ) ,0.1 pM GTP[S] ( 0 )or 0.1 pM GTP[S] plus 100 pM carbachol ( W ) for up to 30 min at 25°C. 5 ml reaction mixture was filtered and membrane-bound radioactivity was determined by measuring the nitrocellulose filters.
enhance the release of bound [35S]GTP[S].When unlabelled GTP[S] (0.1 pM) was added, which increased the extent of [35S]GTP[S]release (about 30% after 30 min), carbachol caused a further substantial release of [35S]GTP[S],reaching about 50% after 30 min of incubation. Labelling of the membranes with increasing concentrations of GTP[S] increased the amount of GTP[S] bound to membrane G proteins (Fig. 9, control). After binding for 20 min, dissociation of bound GTP[S] was immediately initiated by the addition of 10 pM GTP[S] or 10 pM GTP[S] plus 100 pM carbachol and measured after 20 min of incubation. The amount of dissociated GTP[S] caused by addition of GTP[S] alone increased with increasing concentrations of GTP[S] in the labelling period. In relative terms (percentage of initial binding), dissociation of GTP[S] decreased from
729 2.5 20 I
f-
e
1.5
3 0
n
-
3 10 a I-
0
0.5
0
0
O 25 50 75 100 GTP[S] in preincubotion (nM1
Fig. 9. Influence of the GTPISj-binding site occupancy on the release of bound GTPISI. Sarcolemmal membranes (3 pg/tube) were labelled with [35S]GTP[S] (0.3 nM) and increasing concentrations of unlabelled GTP[S] (up to 100 nM) for 20 min. Thereafter, the release of bound GTP[S] was immediately initiated by the addition of 10 pM GTP[S] added without or with 100 pM carbachol. Membrane-bound GTP[S] was determined directly after the labelling period ( A ) or after the release reaction performed for 20 min in the absence (0) and presence of carbachol (0).
almost 50% dissociation at 0.3 nM GTP[S] to about 30% at 10 nM GTP[S] in the binding reaction, without further change up to 100 nM GTP[S]. The amount of GTP[S] being released under control of the mACh receptor was already maximal at about 1 nM GTP[S] in the labelling period and remained constant up to 100 nM GTP[S]. In relative terms, the carbachol effect decreased from almost 30% higher dissociation at 0.3 nM GTP[S] to only 8% at 100 nM GTP[S] in the binding reaction. These data indicate that only a fixed maximal portion of the GTP[S]-binding sites are regulated by mACh receptors. The reaction mixture contained about 40 pM mACh receptors, leading to release of GTP[S] from maximally about 150 pM binding sites upon addition of carbachol. In addition to guanine nucleotides, the carbachol-stimulated release of bound [3'S]GTP[S] from cardiac membranes required the presence of Mg". At concentrations below 10 pM free MgZt, dissociation of [35S]GTP[S]from G proteins was observed, even without addition of unlabelled guanine nucleotide and carbachol (Fig. 10). At 10- 1000 pM Mg2+,however, the dissociation remained constant at a low level. The increase in release of bound [3sSs]GTP[S]from the membranes caused by GTP[S] (0.1 pM) alone was not dependent on the presence and concentration of Mg2+.The MgZ' concentration response curves for [35S]GTP[S]release in the absence and presence of unlabelled GTP[S] alone were parallel. In contrast, the carbachol-stimulated release of [3'S]GTP[S] was strictly dependent on the presence of Mg2+. Half-maximal and maximal stimulation by carbachol occurred in the presence of about 0.5 pM and 10 pM Mg2+, respectively, with no decrease in carbachol effect being observed up to 1 mM Mg2+. DISCUSSION The measurement of receptor-stimulated binding of the poorly hydrolyzable GTP analog GTP[S] to G proteins is used to study the activation of G proteins in reconstituted systems [27, 281 and in membrane preparations [15, 161. For example,
-L c
O
8
7
6
5
L
3
-log([Free Mg2 ]/M)
Fig. 10. Influence of MgZ* on the release of [3sSlGTP[Slfrom cardiac membranes. In order to remove Mgz+, membranes labelled with [35S]GTP[S]were washed three times with Tris/HCl buffer (25 mM, pH 7.4) containing 1 mM dithiothreitol, and in the first washing 20 mM EDTA, in the second no EDTA, and in the third 2 mM EDTA. The reaction was performed in the presence of 2 mM EDTA, the calculated concentrations of MgCI2 and labelled membranes (19000 cpm/80 ~ 1for ) 10 min at 25°C. The release of [35S]GTP[S]was determined with no further additions (A)in the presence of 0.1 pM GTI'[S] ( 0 )or with 0.1 pM GTP[S] plus 100 pM carbachol (W).
in sarcolemmal membranes of porcine atria, the mACh receptor agonist, carbachol, caused a rapid and marked increase in binding of GTP[S] to G proteins [15]. It is generally assumed that this receptor action is due to the release of GDP from G proteins, to which GTP[S] then can bind [lo, 111. As shown herein, the agonist-liganded mACh receptor can also markedly and rapidly stimulate the release of GTP[S] from G proteins in cardiac membranes. The effect of carbachol was concentration dependent and specific for mACh receptors, since it was competitively antagonized by atropine. Interestingly, in the absence of carbachol, atropine by itself inhibited release of GTP[S] from G proteins by about 20%, which suggests that agonist-free, empty mACh receptors may have an effect on interacting G proteins, which can be prevented by antagonist binding. Two requirements were apparently necessary for the agonist-activated receptor action: first, the presence of Mg2+ and, second, addition of a guanine nucleotide. Low Mg2+ concentrations are known to destabilize the binding of GTP[S] to G proteins [22, 291 (Fig. 10). In order to obtain stable binding and, thus, minimal release of GTP[S] under control conditions, we routinely used 1 mM free Mg2+in the reaction mixture. However, even at this concentration of Mg", the carbachol effect was still maximal. Half-maximal mACh receptor-stimulated release of bound GTP[S] was observed at about 0.5 pM free MgZf,very similar t o the concentration of Mg2 required for carbachol stimulation of low-K, GTPase in these membranes [14], which apparently reflects the receptor-stimulated release of G-protein-bound GDP followed by GTP binding and hydrolysis [lo, 111. Thus, receptor-stimulated release of GDP as well as of GTP[S] required Mg2+ at very similar concentrations, pointing to a similar mechanism for these receptor actions. It is widely accepted that binding of GTP[S] t o G proteins is essentially irreversible in the presence of M g 2 + ,as shown for purified G proteins [22].However, in intact cardiac membranes there was a rapid and marked release of G-proteinbound [35S]GTP[S] upon addition of unlabelled guanine nucleotides or large dilution of the labelled membranes. One +
730 important difference between the studies using purified G proteins [22] and the data presented here on membrane G proteins is that the purified G proteins were studied in detergent-containing solutions, which may cause a trapping of the nucleotide in its binding site and, in addition, may facilitate dissociation of G proteins into f l y dimers and GTP[S] liganded r subunits. In fact, there was minimal if any release of [35S]GTP[S]from G proteins solubilized from cardiac membranes with the detergent, 3-[(3-~holamidopropyl)dimethylammonia]-1-propanesulfonate (data not shown). The most likely mechanism by which unlabelled guanine nucleotides increase the release of bound [35S]GTP[S]is by competition with released [35S]GTP[S] for rebinding. In fact, concentrations of GTP[S], GDP[S] and GMP, which blocked the binding of [35S]GTP[S]maximally, also lead to maximal dissociation of bound [35S]GTP[S] from labelled membranes. However, GDP[S] and GMP were much less effective than unlabelled GTP[S] in promoting the release of bound [35S]GTP[S].For example, GDP[S] was only about 50% as effective as GTP[S] in releasing [35S]GTP[S],while both guanine nucleotides were equally effective in inhibiting binding of [35S]GTP[S] to G proteins, although at different concentrations. Both the release and the binding assay were performed for the same period of time (15 min) and with the same amount of free or membrane-bound [35S]GTP[S]. Therefore, different binding kinetics of the unlabelled guanine nucleotides cannot be the reason for the reduced efficiencies of GDP[S] and GMP in causing release of bound [35S]GTP[S]. Thus, the effect of the guanine nucleotides in releasing G-protein-bound GTP[S] can only in part be explained by inhibition of rebinding of released GTP[S]. Additional mechanisms (see below) are apparently involved in this reaction. The most exciting question raised from the data presented herein is how agonist-activated mACh receptors can stimulate the release of bound GTP[S] from G proteins. It is generally assumed that receptors are uncoupled from G proteins with bound guanine nucleoside triphosphate, particularly with Gpp[NH]p and GTP[S] [lo, 11, 30, 311. Similar data as for many other receptors have been reported for cardiac mACh receptors [32]. On the other hand, in intact turkey erythrocyte membranes and in systems reconstituted of partially purified G proteins and receptors, P-adrenoceptors have been shown to interact with Gpp[NH]p-liganded G proteins [33- 351, even with higher affinity than with GDP-liganded G proteins [35]. In contrast, the receptors did not apparently interact with and release GTP[S] from GTP[S]-liganded G proteins [35, 361, whereas Gpp[NH]p was rapidly released [20, 361. Thus, although a variety of data suggest that guanine-nucleosidetriphosphate-bound G proteins are uncoupled or even physically dissociated from receptors, receptors can apparently efficiently interact with Gpp[NH]-liganded G proteins. The data presented herein indicate that receptors can even interact with GTP[S]-bound G proteins. The interaction of agonist-liganded receptors with GTPIS]-bound G proteins is either direct or indirect. If receptors interact directly with GTP[S]-liganded G proteins and, thereby, cause release of bound guanosine nucleoside triphosphate, an additional consequence is that these interacting GTP[S]-liganded G-protein c1 subunits are associated with Gprotein By dimers, known to be required for receptor-Gprotein interaction [lo, 11, 37, 381. Thus, it may be speculated that binding of an agonist to its receptor causes dissociation of any guanine nucleotide from the interacting trimeric G protein, irrespective of whether GDP or GTP[S] is bound and,
thereby, initiates replacement of the bound by a free guaninc nucleotide. On the other hand, it is also feasible that oligomeric G proteins [39,40] are the basis for the observed receptor-stimulated release of GTP[S] and that receptors interact indirectly with the GTP[S]-liganded G proteins. Agonist-liganded receptors may primarily interact with GTP[S]-free G proteins. In a second step, this receptor G-protein interaction somehow induces GTP[S]-bound G proteins to release GTP[S] in a guanine-nucleotide-dependent manner. Such an indirect mechanism of action seems to be corroborated by the finding that addition of a guanosine nucleoside diphosphate or triphosphate was absolutely necessary for agonist-stimulated release of bound GTP[S]. Although large dilution caused release of GTP[S] to a similar extent as observed with, for example, GDP or GDP[S], carbachol did not increase the observed release. This finding suggests that the effect of guanine nucleotides in promoting the carbachol-stimulated release of GTP[S] is not only due to inhibition of rebinding of released GTP[S] but involves an active process. Analysis of GTP[S]binding-saturation experiments indicated that cardiac membranes apparently contain two binding sites for GTP[S], one with high affinity and low capacity and another one with lower affinity and high capacity, and that effects of mACh receptor agonists on GTP[S] binding (increase or release) are preferentially observed with the high-affinity binding component (G. Hilf and K. H. Jakobs, unpublished observations). Binding of guanine nucleotides to the low-affinity component may modify the binding affinity of GTP[S] to the high-affinity sites, resulting in release of bound GTP[S]. Thus, the observed receptor-stimulated release of G-protein-bound GTP[S] may be due to a direct or indirect interaction of receptors with GTP[S]-liganded G proteins or even a combination of these two reactions. Considering the number of mACh receptors present in the membrane preparation used and the amount of G proteins from which GTP[S] can be released under stimulation by carbachol, it has to be concluded that one agonist-activated mACh receptor can cause rapid release of GTP[S] from three to four G proteins. The time course of receptor-stimulated GTP[S] binding and release, and the maximal amount of G proteins being controlled by mACh receptors to bind or release GTP[S], were very similar [15]. These findings suggest that receptor-stimulated release of G-protein-bound GDP, followed by binding of GTP[S], may follow similar reactions, as discussed above for receptor-stimulated release of G-proteinbound GTP[S]. Taken together, evidence is presented that, in contrast to a widely accepted view, binding of the poorly hydrolyzable GTP analog, GTP[S], to G proteins is reversible in intact membranes and that agonist-activated receptors can, either directly or indirectly, interact with GTP[S]-bound G proteins, thereby accelerating the dissociation of the bound guanine nucleoside triphosphate. Although the mechanism(s) by which agonist-liganded mACh receptors stimulate the dissociation of GTP[S] from G proteins in cardiac membranes remains incompletely understood at the moment, the measurement of agonist-stimulated release of G-protein-bound GTP[S] may provide a further sensitive tool to study receptor - Gprotein interactions also in other membrane and receptor systems. This work was supported by the Deutsche Forsc,hungsgemeinschaft.
731
REFERENCES 1. Watanabe, A. M., McConndughey, M. M., Strawbridgc, R. A., Fleming, M. W., Jones, L. & Besch, R. A. (1978) J. Biol. Chem. 353,4833-4836. 2. Jakobs, K . H., Aktories, K. & Schultz, G. (1979) NaunynSchmitdeherg's Arch Pharmacol. 310, 113 - 119. 3. Brown, S. L. & Brown, J. H. (1983) Mol. Pharmucol. 24, 351 356. 4. Brown, J. H. & Brown, S. L. (1984) J . B i d . Chem. 259. 37773781. 5. Pfaffinger, P. J., Martin, J. M., Hunter, D. D., Nathdnson, N . M. & tlille, B. (1985) Nuturr317, 536-538. 6. Breitwieser, G. E. & Szabo, G. (1985) Nature 317, 538-540. 7. Cohen-Armon, M. & Sokolovsky, M. (1986) J . B i d . Chem. 261, 12498-12505. 8. Cohen-Armon, M., Garly, H. & Sokolovsky, M. (1988) Riochemistry 27, 368 - 374. 9. Danilcnko, M. P., Panchenko, M. P., Yesirev, 0. V. & Tkachuk, V. A. (1990) Eur. J . Biochem. 194, 155-160. 10. Gilman, A. G. (1987) Annu. Rev. Biochem. 56,615-649. 11. Birnbaumer, L., Abramowitz, J. & Brown, A. M. (1990) Biochim. Biophys. Acta 1031, 163--224. 12. Cassel, D. & Selinger, Z. (1976) Biochim. Biophys. Acfu452,538 551. 13. Aktories, K., Schultz, G. & Jakobs, K. H. (1982) Mol. Pharmacol. 21,336- 342. 14. Hilf, G. & Jakobs, K. H. (1989) Eur. J. Pharmacol. Mol. Pharmocol. 172, 155- 163. 15. Hilf, G., Gierschik, P. & Jakobs, K. H. (1989) Eur. J . Biochetn. 186,725-731. 16. Gierschik, P., Moghtader. R., StrdUb, C., Dieterich, K. &Jakobs, K. H. (1991) Eur. J . Biochem. 197,725-732. 17. Cassel, D. & Selinger, Z. (1978) Proc. Nut1 Acud. Sci. USA 75, 41 55 -41 59. 18. Pike, J. L. & Lefkowitz, R. J. (1981) J . Biol. Cjiem. 256, 22072212. 19. Muraydma, T. & Ui, M. (1984) J . B i d . Chern. 259,761 -769.
20. Casscl, D. & Selinger, Z.(1977) J . Cyclic Nucleotide Res, 3, 1 I 22. 21. Michel, T. & Lefkowitz, R. J. (1982) .I. Biol. C'hem.257, 1355713563. 22. Higashijima, T., Ferguson, K. M., Sternweis, P. C., Smigel, M. D. & Gilman, A. G. (1987) J . Biol. Chem. 262, 762-766. 23. Petersen, G. L. & Schimerlik, M. I. (1984) Prep. Biochem. 14, 33 - 74. 24. Bradford, M. M. (1976) Anal. Biochem. 72,248-254. 25. Bartfai, T. (1979) Adv. Cyclic Nucleotide Res. 10, 219-242. 26. Seifert, R., Roscnthal, W., Schultz, G., Wieland, T., Gierschik. P. & Jakobs, K. H. (1988) Eur. J . Biochem. 175, 51 -55. 27. Asdno, T., Pedersen, S. E., Scott, C. W. & Ross, E. M. (1984) Biochemistry 23, 5460- 5467. 28. May, D. C. & Ross, E. M. (1988) Biochemistry 27,4888-4893. 29. Fcrguson, K. M., Higashijima, T., Smigel, M. D. & Gilman, A . G. (1986) J. Biol. Chem. 261,7393-7399. 30. Maguire, M. E., Van Arsdale, P. M. & Gilman, A. Cr. (1976) Mol. Pharmacol. 12, 335-339. 31. Limbird, L. E., Gill, D. M. & Lefkowitz, R. J. (1980) Yruc. Natl Acad. Sci. USA 77, 775 - 779. 32. Kosenberger, L. B., Yamamura, H. I. & Roeske, W. R. (1980) J . Biol. Chem. 255,820 - 823. 33. Sevilla, N. & Levitzki, A. (1977) FEES Lett. 76, 129- 134. 34. Arad, H., Rimon, G. & Levitzki, A. (1981) J . B i d . Chem. 256, 1593- 1597. 35. Citri, Y. & Schramm, M. (1982) J. B i d . Chem. 257, 1325713262. 36. Cassel, D. & Selinger, Z. (1977) Biochem. Biophys. Res. C O ~ V ~ M ~ . 77,868-873. 37. Fung, B. K. K . (1983) J . Biol. Chem. 258,10495-10502. 38. Florio, V. A. & Sternweis, P. C. (1985) J . B i d . Chem. 260,34773483. 39. Vaillancourt, R. R., Dhdnasekardn, N., Johnson, G . L. & Ruoho, A. E. (1990) Proc. Natl Acud. Sci.U S A 87, 3645-3649. 40. Nakamura, S.-I. & Rodbell, M. (1990) Proc. Nut1 A c d . Sci. USA 87,6413-6417.
