the of biol~cical vol. no. of pp. 0 for in u.s.a. 2’,3 ... · the joufwal of biol~cical chemistry...

THE JOUFWAL OF BIOL~CICAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 269, No. 50, Issue of December 16, pp. 31999-32007, 1994 Printed in U.S.A.

2’,3’-Dialdehyde GTP as an Irreversible G Protein Antagonist DISRUPTION AND RECONSTITUTION OF G PROTEIN-MEDIATED SIGNAL TRANSDUCTION IN CELLS AND CELL MEMBRANES”

(Received for publication, July 5 , 1994, and in revised form, September 26, 1994)

Christian Nanoff$, Stefan BoehmO, Martin HoheneggerS, Wolfgang SchutzS, and Michael Freissmuthh From the Institute of $Pharmacology and $Neuropharmacology, Vienna University, Wahringer Strasse 13a, -~ A-1090 Vienna, Austria

The 2’,3‘-dialdehyde analogue of GTP, oGTP, was de- vised as an irreversible antagonist of regulatory GTP- binding proteins (G proteins). Here, we show that oGTP uncouples transmembrane signaling mediated by a set of distinct G proteins both in isolated membranes and in whole cells. In human platelet membranes, pretreatment with oGTP suppressed receptor- and G protein- controlled regulation of adenylyl cyclase activity. In chick neuronal cells, inhibition of the voltage-sensitive Ca2+-current by various membrane receptors (a,-adrenergic, somatostatin, GABAJ was eliminated when oGTP was applied intracellularly in the whole cell patch-clamp configuration. Disruption of endogenous signaling pathways by oGTP occurred through specific blockage of the GTP-binding site of G protein a-subunits by the following criteria: (i) pretreatment of membranes with oGTP blocked direct G protein activation by guanine nucleotides as well as labeling of G,, and Gi, with the photoaffinity probe [a-S2P]GTP azidoanilide. (ii) The effect of oGTP was antagonized by the simultaneous in- troduction of guanosine 5‘-(3-O-thio)triphosphate into the patch-clamped cell. (iii) The time to onset of action was similar for oGTP and guanosine 5’-O-thio)diphosphate. (iv) Inactivation of G protein-dependent signaling was overcome by substituting G protein a-subunits. Addition of both the short and long form of recombinant Gs,(rGs,,.s and I-G~,.~) restored guanine nucleotide- dependent adenylyl cyclase activity to oGTP-treated platelet membranes with IG.,.~ being -3-10-fold more potent than rGs,.s. This apparent preference was due to the intrinsically different activation rates of rGsn-L and rG,,,. When reconstituted with exogenous rGs,, the 4- adenosine receptor did not discriminate among the two forms of rGa,. Thus, is the primary determinant of basal C A M P formation in platelets. In contrast, neither the addition of various recombinant subtypes of G, nor purified bovine brain py-dimers reconstituted adenylyl cyclase inhibition in oGTP-treated membranes. All subtypes of Gi, stimulated adenylyl cyclase. In the presence of rGsa, a conditional stimulation by fly-dimers was observed. This pattern of stimulation shows that platelet adenylyl cyclase is a type 11-like isoform. Either a differ- ently modified G protein or an ancillary GTP-binding

* This work was supported by Grants FWJ?-S6604 and P8875-MOB from the Austrian Science Foundation and by a grant from the Anton Dreher Gedachtnisstiftung. The costs of publication of this article were

therefore be hereby marked “aduertisement” in accordance with 18 defrayed in part by the payment of page charges. This article must

U.S.C. Section 1734 solely to indicate this fact. We dedicate this paper to Dr. Oleh Hornykiewicz. 7 To whom correspondence should be addressed: Institute of

Pharmacology, Vienna University, Wahringer Str. 13a, A-1090 Vienna, Austria. Tel.: 43-1-40-480-298; Fax: 43-1-402-48-33.

component is required for adenylyl cyclase inhibition in platelets. oGTP can be considered a useful tool for disruption and reconstitution of transmembrane signaling mediated by presumably all classes of heterotrimeric G proteins.

Heterotrimeric G proteins’ play a pivotal role in transducing information from receptor to effector. The family of heterotrimeric G proteins is characterized by a high degree of molecular diversity (1, 2). The biological significance of this diversity is not fully understood. The functional role of some members is not yet known, and new functions are assigned to even well characterized subunits. A whole set of experimental strategies has been employed to assess the specificity of interaction that governs coupling of the G protein to the receptor or to its cellular effector molecules (for review, see Refs. 2, 3). Originally, the bacterial exotoxins cholera toxin and pertussis toxin have been used to study the role of G, and GjG,, respectively. Addi- tional approaches include reconstitution experiments with purified components, subtype-specific antibodies, cotransfection into mammalian cells of cDNA coding for a particular G protein, receptor or effector molecule, as well as the knock-out of defined G protein species in the cell by antisense oligonucleo- tides. In addition, direct G protein ligands such as the wasp venom mastoparan, peptides derived from the primary se- quence of membrane receptors, and low molecular weight compounds may be of help for the analysis of specific signaling processes (4-6). These agents target the receptor interaction site where they act as G protein activators (receptor mimetics) or blockers. Their action is determined by their affinity for a particular G protein or by interference with a particular receptor-G protein tandem. Thus, the development of subtype-selec- tive low molecular weight compounds like GP-Ant2 (4 ) or D2N (6) may provide useful tools for the delineation of signaling pathways in membranes and intact cells.

In the present report, we chose an alternative approach and targeted the guanine nucleotide binding pocket of G protein a-subunits using the 2’,3’-dialdehyde analogue of GTP (oGTP). The underlying rationale is described in the accompanying paper (7). When tested on purified rGaa, oGTP binds in a quasi-

The abbreviations used are: G protein, regulatory GTP-binding protein; rGam.+ and rG,,.,, the short and long splice variant of the recombinant a-subunit of the G protein G, which stimulates adenylyl cyclase; G, and Go, G proteins mediating inhibition of adenylyl cyclase and regulation of neuronal calcium channels, respectively; rG,,, recombinant myristoylated G protein a-subunit of Gi, three subtypes of which (G,,.,, G,,,, Gin.J exist; rG,, recombinant myristoylated a-subunit of Go; NEXA, 5’-N-ethylcarboxamidoadenosine; GTPyS, guanosine 5‘43-0- thi0)triphosphate; GDPpS, guanosine 5’-(2-O-thio)diphosphate; PGI,, prostaglandin I, (prostacyelin); CHAPS, 3-[(3-cholamidopropyl)dimeth- ylammoniol-1-propanesulfonic acid; cpm, counts/min.

3 1999

32000 oGTP as Irreversible G Protein Antagonist

irreversible manner and traps the G protein cY-subunit in an inactive conformation thus acting as an irreversible G protein inhibitor (7). Here, we apply oGTP to a set of well defined signal transduction systems and demonstrate its effectiveness in un- coupling G8-, Gi-, and G,/G,-mediated pathways in cell membranes and whole cells. The oGTP-induced disruption is irreversible and can only be overcome by the exogenous readdition of G protein a-subunits.

EXPERIMENTAL PROCEDURES Preparation of oGTP and [CY-~~PIGTP azidoanilide-Preparation, pu-

rification, and quality control of oGTP was performed as described in the accompanying paper (7). For treatment of control membranes (and cells), mock oxidized GTP was used; periodate (0.11 mmol) was first dissolved in 10% glycerol. After 3 min, GTP (0.1 mmol) was dissolved in this solution and subsequently purified over a Sephadex G-10 column. GTP recovered from this treatment was indistinguishable from un- treated GTP. This was determined by binding to rG,, and by polyeth- ylene glycol-coated cellulose thin layer chromatography (data not shown). [a-32PlGTP azidoanilide was prepared according to Offermans et al. (8). Briefly, 1 mCi of [W~~PIGTP (specific activity diluted to 300Ci/ mmol by the addition of 1.44 nmol of carrier GTP) was incubated with an excess of 4-azidoaniline for 4 h at room temperature. [CY-~~PIGTP azidoanilide was purified by high performance liquid chromatography on a Spherisorb SAX ion-exchange column. The column was equili- brated with 2.8% ethanol in TEAC (triethylammoniumcarbonate), pH 8.5, and [a-32PlGTP azidoanilide was eluted with an ethanol gradient (2.848%). The entire reaction was carried out in the dark. About 75% of the radioactivity applied was incorporated into the reaction product. Quality control for [(Y-~~PIGTP azidoanilide was carried out by testing specific incorporation into recombinant G,. Upon W irradiation, incorporation efficiency amounted to 4% of the radioactivity bound.

Protein Purification-Myristoylated recombinant G protein a-subunits (rGia.l, rG,,.,, rGia.3, and rGoJ were purified as described (91, and recombinant G protein a,-subunits (rGsm) were obtained (10). The amount of G protein a-subunits used in individual experiments was determined by [36SIGTPyS binding in a medium containing 50 HEPES- NaOH mM, pH 8, 1 mM EDTA, 10 mM MgSO,, 0.1% Lubrol PX (=HEDL 0.1% 1, 1 p [35SlGTPyS (specific activity -20,000 cpdpmol). The incubation (2 h a t 30 "C) was terminated by filtration over nitrocellulose BA85 filters. The G protein py-dimer was chromatographically resolved from the Go,, oligomer (11). The concentration of the py-subunit was calculated from the protein concentration assuming a molecular mass of 45,000. Bovine brain adenylyl cyclase was purified over forskolin- agarose (12).

Immunoprecipitation and Immunoblotting of G Protein a-Subunits- Platelet membranes (=1 mg of membrane protein) pretreated with GTP or oGTP (see above) were incubated with [U-~~PIGTP azidoanilide (1 PM, specific activity 200 cpdfmol) in the presence of 10 mM MgCl, for 60 min at 25 "C. The membrane suspension (0.2 ml) was irradiated with W light (254 nm, 150 watts) for 15 s, and membranes were pelleted at 39,000 xg. The pellet was taken up in 0.25 ml of RIPA buffer containing 1% Nonidet P-40 and 50 mM Tris-HC1, pH 8.3, 5 mM EDTA, 150 mM NaCI, and was solubilized by sonication. After 30 min on ice, the solubilized suspension was centrifuged at 80,000 x g for 30 min. The supernatant was then incubated with non-immune serum for 1 h at 4 "C and cleared by the addition of Pansorbin (final concentration 1% w/v). After centrifugation, the resulting supernatant was incubated with CS1 antiserum (diluted 1: lO) for 6 h at 4 "C. G,, was precipitated by the addition of pansorbin. The pellet obtained by centrifugation was washed once and resuspended in Laemmli sample buffer to which 40 mM dithiothreitol had been added. The immunoprecipitate was resolved by SDS-polyacrylamide gel electrophoresis (5% stacking gel, 10% re- solving gel) and subjected to autoradiography.

Immunoblotting followed the procedure of Towbin et al. (13). Follow- ing transfer of proteins, the BA85 nitrocellulose membrane was blocked with 1% bovine serum albumin and incubated with CS1 or AS7 antiserum (dilution 1x500). The membrane was washed with 0.05% Tween, and immunostaining was carried out with '251-labeled anti-rabbit IgG. The antisera CSI (14) and AS7 (15) were kindly donated by Dr. G. Milligan (Glasgow University).

Preparation of Human Platelet Membranes and neatment of Platelet Membranes with oGTP-Platelet membranes were prepared according to the method outlined by Huttemann et al. (16). Briefly, platelets were washed twice in the presence of 20 mM EDTA, transferred to hypotonic lysis buffer (5 mM Tris-HCI, pH 7.5, and 5 mM EDTA) and subjected to

a freeze-thaw cycle. Lysed platelets were homogenized with a Teflon pestle, and a membrane pellet was obtained by centrifugation a t 39,000 x g. The pellet was washed, taken up in HME (20 mM HEPES, pH 8.0, 1 mM EDTA, 2 mM MgC1,) a t a concentration of 5 mg/ml membrane protein and immediately subjected to oGTP treatment. Fresh platelets purchased from a local blood donation service gave high adenylyl cyclase activity which declined once a batch had been kept for more than 3 days in the incubator. Platelet membranes were incubated (unless indicated otherwise) with 10 I.~M oGTP in the presence ofATP (0.5 mM), creatine phosphate (10 mM), and creatine kinase (1 mg/ml) for 30 min at 25 "C. Control membranes were treated in a similar manner except that mock-oxydized GTP instead of oGTP was used (see above). Membranes were kept on ice for 60 min, then diluted with ice-cold HME buffer and pelleted at 39,000 x g. The pellet was washed once, taken up in HME, and used immediately or was frozen in aliquots. The binding of oGTP was governed by G protein subtype-specific differences in the rate of GDP release (17). At lower ambient Mg2' concentrations (-1 mM), membrane treatment with oGTP (10 p~ for 60 min) predominantly affected G,. This reduced adenylyl cyclase inhibition while G,-mediated stimulation was less affected (Fig. 4). G, blockage resulting in deficient adenylyl cyclase stimulation, on the other hand, was more pronounced after oGTP pretreatment in the presence of high Mg2' concentrations (10 mM). This is in accordance with the observation that in membranes G, has a markedly lower GDP dissociation rate than Gi (10, 17).

Neuronal Cell Cultures-The procedures of dissociating and cultur- ing chick sympathetic neurons for electrophysiological experiments have been previously described in detail (19). Briefly, paravertebral sympathetic ganglia were dissected from 12-day-old chick embryos, trypsinized (0.1% for 30 min at 36 "C), subsequently triturated, and finally resuspended in Dulbecco's modified Eagle's medium (Life Tech- nologies, Inc., no. 041-01885M) containing 2.2 g/liter glucose, 10 mg/ liter insulin, 25000 IUAiter penicillin, and 25 mglliter streptomycin (Life Technologies, Inc., no. 043-05140D), 100 mg/liter gentamicin (Seromed no. A2712), 10 pglliter nerve growth factor (Life Technologies, Inc., no. 04360501, and 5% fetal calf serum (Life Technologies, Inc., no. 011-0620H). Dorsal root ganglia obtained from these embryos were dissociated and resuspended identically. All cells were plated onto poly- D-lysin-coated (Sigma no. 1149) tissue culture dishes (Nunc no. 153066). Sympathetic neurons were used 224 h after plating. This time interval was sufficient for the complete inactivation of G proteins involved in the modulation of Ca2+ currents if cells were treated by the inclusion of 100 ng/ml pertussis toxin (20).

Electrophysiology-Whole cell Ca2+ currents were recorded at room temperature (20-24 "C) employing the whole cell configuration of the patch-clamp technique (21) with a List LI"EPC7 amplifier (List Medical, Darmstadt, Germany) and the pClamp Hard- and Software (Axon TL-1 DMA Interface, Axon Instruments, Burlingame, CAI. Patch electrodes were made of glass capillaries (inside diameter 0.86 mm, Science Products, Frankfurt, Germany) with a vertical pipette puller (no. 720, David Kopf Instruments, Tujunga, CAI yielding a tip resist- ance of 3-5 Ma. We used choline chloride instead of NaCl and tetrodotoxin in the bathing solution to block Na' currents and N-methyl+- glucamine and tetraethylammonium chloride in the internal solution (for filling the pipettes) to block K' currents. This internal solution contained 115 mM N-methyl+-ghcamine, 20 mM tetraethylammonium chloride, 1.6 mM CaCI,, 2 mM Mg-ATP, 10 mM EGTA, 10 mM glucose, 20 nm HEPES, adjusted to pH 7.3 with HCI. For control recordings the internal solution contained 2 mM Li-GTP. The inclusion of GTP in the internal solution is not essential for the determination of G protein- mediated modulation of Ca2+ currents in chick sympathetic (19) and sensory (22) neurons. GTP in combination with ATP, however, has been shown to preserve the Ca2+ current inhibition by noradrenaline in frog sympathetic neurons for up to 60 min (23). oGTP (0.01 mM), GTP$ (0.001 to 0.1 mM), or GDPpS (0.1 mM) replaced Li-GTP in the internal solution as indicated. The external bathing solution consisted of 120 mM choline chloride, 5 mM CaCI,, 20 mM glucose, 10 mM HEPES, and 0.001 mM tetrodotoxin, adjusted to pH 7.3 with KOH.

Ca2+ currents of sympathetic neurons which extend numerous processes during a 24-h cultivation period were recorded from the cell bodies (19). Sensory neurons were used for electrophysiological experiments within 5 h after plating. During this time, processes were formed only in a limited number of neurons, and these cells were not used for Ca2+ current recordings. In order to enable complete equilibration of the cytoplasm with the internal solution of the pipette (241, current recordings were started more than 10 min after the establishment ofthe whole cell configuration, unless indicated otherwise. Ca2+ currents were elicited by depolarization from the holding potential of -80 to 0 mV. In order to account for the time-dependent run down of Ca2' currents (25;

oGTP as Irreversible G Protein Antagonist 32001

see also Fig. 51, effects of externally applied drugs were quantified by measuring current amplitudes in the presence of test drugs (B) and by comparing them to control currents recorded before (A) and at least 20 s after (washout, C) the application of the drugs (191, according to the equation: % inhibition of Ca'+current = [l - 2B/(A + C)] x 100.

Determination of Adenylyl Cyclase Actiuify-Adenylyl cyclase activity in human platelet membranes was assayed in a reaction mixture containing 50 mM HEPES-NaOH, pH 8.0, 0.05 mM [a-"'PIATP (specific activity -200 cpm/pmol), 10 mM MgCI,, 0.1 mM Rolipram (RO 20172). 10 mM creatine phosphate, 1 mg/ml creatine kinase, 5 pg/ml adenosine deaminase, 1% bovine serum albumin, and 15-40 pg of membrane protein. Inhibitory regulation of adenylyl cyclase was assayed in the presence of 20 p~ forskolin and 2 mM MgCI,. Reconstitution of oGTP- treated platelet membranes with recombinant a-subunits was performed as described previously (10). Purified recombinant a-subunits were diluted in HEDL 0.1% and a volume of 10 pl was added to 40 pl of platelet membranes in HME (15-20 pg of membrane protein). After preincubation (30 min on ice), the adenylyl cyclase reaction was initiated by adding the reaction mixture to a final volume of 0.1 ml. The incubation was terminated after 20 min a t 25 "C, and I"'P1cAMP was separated by double column chromatography (26). The final Lubrol concentration in the assay was adjusted to 0.01%. When the adenylyl cyclase activity was determined upon addition of exogenous G, or py, purified proteins were diluted in HED with 25 mM CHAPS, and the final CHAPS concentration in the assay was 2.5 mM. Interaction of the components involved is not impaired by the presence of residual amounts of detergent (see also 10, 27). Under similar conditions, addition of rG,,, did not amplify adenylyl cyclase stimulation in GTP-treated control membranes (data not shown).

The assay conditions for determining the activity of purified bovine brain adenylyl cyclase were as described previously (28); py-dimer preparations (0.05-2.5 pg) were preincubated with 50-100 ng of purified adenylyl cyclase for 30 min on ice in 20 pl of HEDL 0.1%; thereafter the reaction was started by the addition of 80 1.11 of prewarmed substrate solution containing 0.6 mM [a-32PlATP (specific activity 20 cpdpmol) and 10 mM MgCl, and carried out for 5 min a t 20 "C.

RESULTS

Irreversible Binding of oGTP to G, and G,, in Human Platelet Membranes-In the accompanying paper, we showed that oGTP acts as a quasi-irreversible antagonist on purified Gsm. Quasi- irreversible binding of oGTP to G protein a-subunits can also be demonstrated in cell membranes. Following pretreatment of platelet membranes with oGTP or GTP, guanine nucleotide-binding proteins were visualized by photoaffinity labeling with [a-"PIGTP azidoanilide (Fig. 1). In GTP-treated membranes, a band migrating at 40 kDa was visualized. The labeling of this protein met the criteria for the identification of G protein a-subunits. The extent of incorporation was enhanced in the presence of millimolar concentrations of magnesium and was suppressed by excess GTPyS (Fig. 1, lanes a-c); this band comigrated with Gi,, and was recognized in immunoblots by antiserum AS7 (not shown). Gi,.2 is known to be the predominant G, subtype in human platelets (29). In platelet membranes preincubated with 10

oGTP, labeling of the Gi,.z band as well as other non-identified bands, which incorporated [a-"P]GTP azidoanilide, was dramatically reduced (Fig. 1, lane d).

Since in platelet membranes G,, is present a t much lower levels than Gi,.2, the detection of the photoaffinity-labeled G,, required protein enrichment by means of immunoprecipitation. Antiserum CS1 precipitated both the long and the short splice variant of purified rG,,, with similar efficiency (-40%, not shown). In human platelet membranes, immunoblots with the same antiserum revealed that the long form is predominantly expressed compared to the short form (Fig. 24 ). Accordingly, in GTP-pretreated platelet membranes, a soluble extract prepared after [a-"PIGTP azidoanilide labeling contained more Gl,.,, than G,,., as visualized by immunoprecipitation. Again, labeling of immunoprecipitated G,, was greatly reduced after pretreatment with oGTP. The reduction appeared less complete than that observed for Gi,.* (see Fig. 1). This is probably due to the lower GDPIGTP exchange rate of oligomeric G,, resulting in

a b c d

Mg - + + + GTPJS - - + -

with [CY-~~PIGTP azidoanilide. Membranes were pretreated either FIG. 1. Photoaffinity labeling of human platelet membranes

with GTP (a+) or oGTP ( d l in the presence of 10 mM MgCI,, washed, and subsequently incubated with 1 p~ [a-"PIGTP azidoanilide (specific activity 200 cpdfmol) for 30 min a t 20 "C. This second incubation medium included MgCI, a t a free concentration of 0.5 mM (a) or 10 mM (b-d or in the presence of 10 PM GTPyS (c). Samples were photolyzed by W irradiation (15 s ) , centrifuged, and taken up in Laemmli sample buffer. Equal amounts of membranes were subjected to SDS-polyacrylamide gel electrophoresis (-20 pg of membrane proteidane). The gel was dried and exposed to x-ray film for 16 h. This experiment was repeated twice.

S t , memb. GTP OGTP

FIG. 2. Immunoblot analysis (panel A) and immunoprecipitation of G,, splice variants in human platelet membranes following photoaffinity labeling (panel B ) . Panel A, human platelet membranes (20 pg) were dissolved in Laemmli sample buffer (lane memb. ), separated by SDS-polyacrylamide gel electrophoresis and transferred to a BA85 nitrocellulose membrane; lane st., purified rGw.L (100 ng) and rG8m.s (50 ng). The membrane was subjected to immunoblotting using the CS1 antiserum a t a dilution of 1500; immunoreactive bands were visualized by incubating with a "'1-anti-rabbit antibody and autoradiography. Panel B , human platelet membranes (-1 mg of membrane protein) were labeled with [a-"2P1GTPazidoanilide (1 p ~ , specific activity - 200 cpdfmol) and solubilized in RIPA buffer as outlined under "Ex- perimental Procedures." Immunoprecipitation was carried out by se- quential incubations with non-immune serum and antiserum CSl (dilution 1 : l O ) followed by adsorption to Pansorbin. The recovery of radioactivity in the immunoprecipitate amounted to -0.5% of the membrane bound radioactivity. Shown is the autoradiography of immuno- precipitates obtained from GTP- and oGTP-treated platelet membranes.

less oGTP binding to G,, than to Gi,.z (see below). If human platelet membranes were incubated with [a-"PloGTP, the bound radioactivity was not removed by repeated washing but was released in the presence of 1 M HC1. These observations show that oGTP is irreversibly bound to G protein a-subunits as well as to additional guanine nucleotide-binding proteins in the membrane.

Disruption of G Protein-dependent Regulation of Adenylyl Cyclase by oGTP-The functional consequences of irreversible


0 7 6 5 4

G T Q S (-log M)

I T I - 100 h

E" - 75 .9

E L-

\

-50 o E,

- 2 5 a E

v

2 - 0

FIG. 3. Stimulation of adenylyl cyclase activity in human platelet membranes. Panel A, adenylyl cyclase activity was determined in platelet membranes (40 pg of membrane protein) pretreated with 10 p~ oGTP (0) or GTP (0) in the presence of increasing Concentrations of GTPyS as outlined under "Experimental Procedures." Inset, data were replotted after normalization by setting the respective activity maximum 100% to illustrate similar EC,, values. Data are from a representative experiment performed in duplicate and repeated three times with similar results. Panel B, adenylyl cyclase activity assessed in the presence of GTPyS (1 PM), GTPyS + NECA (10 p), or GTPyS + PGI, (1 p~). Bars represent mean values ( n = 4) obtained in control (hatched bars) and oGTP-treated membranes (cross-hatched bars).

oGTP binding were assessed by measuring the bidirectional, G protein-dependent regulation of adenylyl cyclase. In platelet membranes pretreated with 10 PM oGTP, the ability of GTPyS to stimulate adenylyl cyclase was dramatically reduced with respect to the maximal effect (Fig. 3A). However, the concentration dependence of the GTPyS-effect was unaltered and comparable in oGTP-treated and in control membranes (Fig. 3 A , inset 1. This provides functional evidence that the majority of G,, coupling to adenylyl cyclase is blocked by oGTP and that the inhibition of G,, cannot be reversed even through high concentrations of GTPyS. oGTP also diminished basal adenylyl cyclase activity and receptor agonist-induced cyclase stimulation (Fig. 3B). In spite of a large difference in the amount of CAMP formation, the fold stimulation of adenylyl cyclase by guanine nucleotides or receptor activation (by NECA or PGI,) was only modestly reduced after oGTP treatment. This is in- dicative of unimpaired signal transfer from receptor to adenylyl cyclase by the remaining population of non-oGTP-liganded G,. The effects elicited by membrane pretreatment with oGTP was concentration-dependent. Starting in the submicromolar concentration range, oGTP inhibited GTPyS-stimulated cyclase activity. Up to a concentration of 10 p ~ , oGTP neither inhibited purified bovine brain adenylyl cyclase nor forskolin-stimulated adenylyl cyclase activity in G,,-deficient S49 cyc- membranes; significant inhibition of catalytic activity occurred at oGTP concentrations 2100 (data not shown).

We have also verified that pretreatment with oGTP caused irreversible inhibition of the G,-dependent regulation of adenylyl cyclase. Experiments were carried out in the presence of forskolin which gave an approximately 40-fold increase in the level of adenylyl cyclase activity. In control membranes, GTPyS produced a concentration-dependent inhibition of adenylyl cyclase activity up to -50% (Fig. 4A). The inhibition by low concentrations of GTPyS (10 nM) was enhanced by activation of the a,-adrenergic receptor with epinephrine (Fig. 4B). Pre- treatment of membranes with oGTP led to a moderate decrease in forskolin-induced cyclase activity (292 2 58 versus 254 z 44 pmol/min/mg in GTP- and oGTP-treated membranes, respectively; n = 5) while the maximal extent of Gi-mediated inhibition by GTPyS (Fig. 4A) or following a,-adrenergic receptor activation by epinephrine was greatly reduced (Fig. 4B).

Disruption of G Protein-dependent Regulation of Neuronal

0 'b/J 1 'I L 0 1 10 100 1000

GTP$ (nM)

, , m 2 s & E s 9 s ?

FIG. 4. Inhibition of adenylyl cyclase activity in human platelet membranes. Panel A, enzyme activity was determined in the presence of 20 p forskolin and 2 mM MgCl,. The assay was carried out at 25 "C for 20 min with 30 pg of membrane protein of either GTP-treated control (0) or oGTP-treated membranes (0) in the presence of increasing concentrations of GTPyS. Comparable IC,, values (-10 nM) were estimated for GTPyS in control and oGTP-treated membranes. Deter- minations were performed in duplicate and the data shown are representative of three independent experiments. Panel B , attenuation of platelet adenylyl cyclase activity by epinephrine (30 p ~ ; EPI) in the presence of GTPyS (10 MI). Mean values f S.E. (n = 4) are given for the inhibition of adenylyl cyclase activity in the presence of forskolin (= 100%). Forskolin-stimulated adenylyl cyclase activity was 286 -c 47 and 262 f 51 pmol/min/mg in GTP-treated control and oGTP-treated membranes, respectively.

Calcium Channels by oGTP-The data presented so far show that oGTP effectively disrupts G protein-dependent signaling in isolated cell membranes. Additionally, the effect of oGTP was investigated in whole cells by applying oGTP to the interior of sensory and sympathetic neurons of chick embryos. In these cells, activation of pertussis toxin-sensitive G proteins causes inhibition of voltage-activated Ca2+-channels (20, 22). Agonists at somatostatin, la1 ,-adrenergic, and GABA, receptors markedly reduced Ca2+ currents evoked by depolarizations from -80 to 0 mV Inclusion of oGTP into the recording pipette invariably abolished current inhibition regardless of the type of cell or type of receptor tested (Fig. 5). oGTP also eliminated the ox- otremorine (= muscarinic) and bromoxidine (= a,-adrenergic) induced inhibition of Ca2+ currents in rat superior cervical gan- glion neurons (not shown, see also Ref. 30). The effect of oGTP was identical to that of a pretreatment with pertussis toxin (Fig. 5 A , open squares) and comparable to the effect of GDPPS, which blocks G protein mediated signaling in a reversible manner. Time course experiments revealed similar time intervals for the onset of the effect of GDPPS and oGTP indicating that association with the target site (i.e. a GTP-binding site on the a-subunit) constitutes the rate-limiting step (Fig. 6).

The following experimental design was chosen to demonstrate an antagonism between GTPyS and oGTP which both act as quasi-irreversible G protein ligands with opposing ac- tivities. oGTP and GTPyS were added to the internal solution simultaneously a t varying molar ratios. In chick sympathetic neurons, intracellular GTP@ mimics the inhibition of Ca2+ currents by receptor agonists (31,32). Either upon intracellular application of GTPyS or in the presence of receptor agonists, the inhibition of Ca2+ currents may be relieved by large depolarizing prepulses preceding the test depolarization (23, 31, 32). As illustrated in Fig. 7 (upperpanel), a prepulse enhanced the peak current amplitude obtained with intracellular GTPyS while it had no effect under control conditions in the presence of GTP. The current amplitude recorded after the prepulse was normalized to the control amplitude preceding the prepulse (Z&,, Fig. 7 B ) . This ratio is an index for G protein-mediated inhibition of the Ca2+ current. As indicated by an enhanced prepulse effect, increasing concentrations of GTPyS in the pres-


FIG. 5. oGTP-induced disruption of agonist dependent modulation of N- type calcium currents in chick sympathetic and sensory neurons. Panels A and B, reduction of Ca2+ currents by

nergic agonist bromoxidine (= UK14,304; saturating concentrations of the a,-adre-

10 p ~ ) in chick sympathetic neurons (panel A) and of the GABAB agonist baclofen (100 p ~ ) in chick sensory neurons (panel B). The cells were dialyzed with either 2 mM GTP (upper traces, O), 0.01 m~ oGTP (center truces, O), or 0.1 mM GDPPS (panel B, lower trace, A). Alterna- tively, sympathetic neurons had been pretreated with pertussis toxin (100 ng/ml for 24 h) and were dialyzed with 2 m~ GTP (panel A, lower trace, 0). Recordings were started 10 min after breaking the cell membrane. Each data point repre- sents the peak current amplitude aver- aged from three recording periods performed subsequently in a single cell. Amplitudes have been normalized to the first current amplitude recorded in each

recordings (calibration: 0.5 nA, 30 ms). series. Insets show representative original

Panels C and D, quantification of the inhibitory effects of bromoxidine (10 PM) and somatostatin (0.1 PM) on Ca2+ currents in chick sympathetic (panel C ) and of baclofen (100 p ~ ) and bromoxidine (10 p ~ ) in chick sensory neurons (panel Dl. Cells were dialyzed with the guanine nucleotides as indicated; chick sympathetic neurons had also been pretreated with pertussis toxin and were dialyzed with GTP (see above). Inhibitory effects of externally applied drugs were evaluated as described under "Experimental Procedures."

2mM GTP 2minm ......._..__........._.

5min m,

O . O l m M oGTP u ........................

1: ...................... .

A: s y m p a t h e t i c neurons B: senso ry neu rons

bromoxidine 10 M -6

baclofen lO"M

0.0 J I I L 0.0 I I ' I ' ( 1 I I 1 0 80 120 180 0 80 120 180

t ime (5) t ime ( 5 )

,"=5 somatos ta t in I [D n = 5 baclofen 7 5

0 bromoxidine 0 bromoxidine F 50 9

c 25 .E

w

G T P P T X o G T P G T P G D P P S o G T P 0

O . l r n M GDPBS

L I L I L

FIG. 6. Time-dependent onset of the oGTP- and GDPPS- induced block of the Ca2+ current inhibition by baclofen. Ca2+ currents were elicited by a 140-ms depolarization every 30 s starting 90 s after establishing the whole cell configuration in chick sensory neurons. Baclofen (100 p ~ ) was applied 2,3,4, and 5 min after breaking the cell membrane (always for 25 s before recording currents). Currents obtained in the presence of agonist are indicated by asterisks; they are shown together with control currents recorded before application of receptor agonist or after washout. The internal solution contained guanine nucleotides as indicated (calibration: 1 nA, 30 ms).

ence of oGTP gradually inhibited the Ca2+ current (Fig. 7, lower panel). Under these conditions receptor agonists failed to re- duce Ca2+ currents (data not shown). If GTP+ was present at a 10-fold molar excess, Ca2+ current inhibition was close to the effect of GTPyS found in the absence of oGTP (Fig. 7 B ) thus showing that the guanine nucleotides competed for a common

site of action. In addition, a direct effect of oGTP on Ca2+ channels is rather unlikely since the presence of oGTP did not significantly alter Ca" currents (mean current amplitudes: 1009 2 185 and 752 t 152 nA with intracellular GTP or oGTP, respectively; n = 7). Moreover, Diverse-Perluissi and Dunlap (22) have demonstrated that the receptors tested in sensory neurons use distinct signaling pathways converging to Ca2+ current inhibition. While the a,-adrenergic effect requires protein kinase C activity, the GABA, receptor action does not. Both effects are, however, initiated through activation of pertussis toxin-sensitive G proteins and were, as shown in the present work, identically suppressed by oGTP. This further supports the conclu- sion that G proteins are the cellular site of action of oGTP.

Reconstitution of Adenylyl Cyclase Regulation to oGTP- treated Membranes by Exogenous Addition of G Protein Subunits-Since oGTP acts as an irreversible blocker of endogenous G proteins in membranes, it should be possible to recon- stitute the signaling pathway after removal of oGTP. We have therefore examined whether addition of purified rG,, reconsti- tutes the receptor and guanine nucleotide-dependent stimulation of adenylyl cyclase to oGTP-treated platelet membranes. The addition of purified rG,, enhanced CAMP formation by both guanine nucleotide-induced as well as receptor-mediated G protein stimulation (Fig. 8). The maximum levels achieved in this reconstitution experiment approximated the level of cyclase activity observed in control membranes thus providing further evidence that the catalytic activity of the adenylyl cyclase enzyme was not affected by oGTP. The level of enzyme activity increased with the amount of a-subunit added, with both rGsa.L (0, W) and rGs,.B (A, A). rGBO.L was about 3-10-fold more potent than rGsa.a when cyclase stimulation was tested in the presence of GTPyS (0.1 PM) or NECA (10 VM) plus GTPyS.


I (c t1) I (PP)

OmV 50mV OmV -80mV

2mM G T W "."""_ "//. 'Os vmv "."."..".

GTP (mM) 2.0 " "

GTP -, S (mM) -- 0.001 0.01 0.1 0.1 oGTP (mM) -- 0.01 0.01 0.01 "

-- - _

FIG. 7. Competition of GTPyS and oGTP for interaction with the G protein mediating Ca2+ current inhibition. Panel A, effect of

neurons dialyzed with 2 mM GTP (upper trace) or 0.1 m~ GTPyS (lower a large depolarizing prepulse on the Ca2+ current in chick sympathetic

trace). A depolarizing prepulse (50 mV, 40 ms) followed by a 5-ms repo- larization precedes the test depolarization. Control currents are recorded 10 s before application of the prepulse protocol. Panel B, the ratio of current amplitudes following (I(pp)) and preceding a depolarizing prepulse (I(ctl)) measures the extent of Ca2+ current inhibition. As con- trols, cells were dialyzed with 2 mM GTP (no current inhibition) or 0.1 m~ GTPyS (maximal current inhibition). I(pp)/I(ctl) ratios are shown for increasing concentrations of GTPyS in the presence of 10 p~ oGTP;

tively) from the value in the presence of 2 mM GTP; #, ##, significantly *, **, indicates a significant difference ( p < 0.05 and p < 0.01, respec-

different ( p < 0.05 and p <0.01, respectively) from the value in the presence of 0.1 mM GTPyS (unpaired Student's t test; n = 4).

In Fig. SB, the NECA-induced increment in adenylyl cyclase activity was replotted versus the amount of a-subunit added depicting the concentration requirements for the interaction with the &-adenosine receptor. This illustrates that the 4- adenosine receptor did not discriminate between the two forms of G,, and that low concentrations of rGaa sufficed to saturate the interaction with the &-adenosine receptor (EC,, -1 nmol/ liter). Results were comparable when GTP was used instead of GTPyS although the NECA-induced increments in adenylyl cyclase stimulation were smaller in the presence of GTP (10 VM). The effect of NECA on adenylyl cyclase stimulation in the presence of exogenous rGso was completely antagonized by the adenosine receptor antagonist xanthine amine congener (10 p~). Likewise, PG1,-mediated regulation of adenylyl cyclase was also reconstituted by addition of rGsu.L or rGS,.% (not shown).

The receptor-independent component of adenylyl cyclase activation (Fig. &4, open symbols) suggests that the adenylyl cyclase enzyme preferentially couples to G,,,. This difference in potency, however, was eliminated if rG8a.L and rG8,.s were activated prior to addition to the membrane. Preincubation with GTPyS abolished the lag phase in activation which results from the dissociation of prebound GDP. This gave superimpos- able concentration response curves for rGsa.L and rG,,, (Fig. 9). Hence, the apparent ability of the adenylyl cyclase enzyme to discriminate between rGsol.L and rGsa.B was not due to a difference in absolute affinity but arose from the intrinsically different dissociation rates of prebound GDP, a distinct feature of

C

GTP

0 1 10 100 1000

rGs, (ng/assay)

FIG. 8. Reconstitution of adenylyl cyclase to oGTP-treated platelet membranes by addition of rGBa. Panel A, purified rGsm.L (0, .) or r G S m . @ (V, V) was diluted in HEDL 0.1%; the amounts indicated were added to oGTP-treated platelet membranes (-20 pg of membrane protein), and the mixture was preincubated on ice for 30 min. The reaction was initiated by the addition of substrate ATP and the following agents: GTPyS (0.1 p ~ ; 0, V) or GTPyS + 10 J ~ M NECA (., V). The assay was carried out in a final volume of 0.1 ml (20 min at 25 "C) containing MgC1, (10 m ~ ) and 0.01% Lubrol. This experiment was performed in duplicate and repeated twice with similar results. Panel E, the NECA-induced increment in cAMP formation, i.e. cAMP generated in the presence of NECA + GTPyS minus CAMP generated in response to GTPyS alone, was replotted versus the amount of rGea.L (.) or rGao.B (V) to illustrate the receptor-mediated activation Panel C , control adenylyl cyclase activity in oGTP and GTP-treated platelet membranes in the absence of added rGsa.

each splice variant of G,, (10). Furthermore, following preactivation, the apparent affinity of rG,,.L and rGaa.s was increased resulting in EC,, values which correspond to those required for half-maximal saturation of the receptor-mediated effect (see Fig. 8B).

While reconstitution of adenylyl cyclase stimulation to oGTP-treated platelet membranes was achieved by straightfor- ward addition of rG,,, attempts to restore inhibition with recombinant Gia not only failed but produced an unexpected result. Addition of preactivated myristoylated recombinant Gi,- subtypes increased adenylyl cyclase activity in oGTP-treated platelet membranes (Fig. 1OA). At a concentration of 100 nmou liter, G,,.l and Gi,.l enhanced the activity of forskolin stimulated adenylyl cyclase by 1.8-fold; G,,, was slightly less active and Go, barely had an effect. Similar results were obtained under the following conditions: (i) the final concentration of G,, was raised up to 0.4 p ~ . (ii) Non-preactivated G,, was added to the membranes in the presence of 1 p~ GTPyS or 10 p~ GTP. (iii) Experiments were also carried out at a final GTPyS concentration of 0.1 p~ (not shown). Under none of these conditions was inhibition restored. Since the Gi, subtypes were ap- parently not capable of acting as inhibitors of platelet adenylyl cyclase, we have also tested the effect of purified bovine brain py-dimers. Upon receptor activation of oligomeric G proteins, free py complexes are generated which are capable of inhibiting


0 "0 200 400 600 800 1000 0 25 50 7 5 100

rGsa (ng/assay) V (CAMP pmol/mg/min)

tivated rG,, in oGTP-treated platelet membranes. Preactivation of FIG. 9. Reconstitution of adenylyl cyclase activity with preac-

rGsm.L (.) or rG8m.s (V) was performed at 25 "C for 45 min in the presence of 20 p~ GTPyS, 10 m~ MgSO,, and 0.1% Lubrol. Amounts of rGaa indicated were withdrawn and were preincubated with oGTP-treated platelet membranes (-20 pg of membrane protein) for 30 min on ice in 60 pl. The adenylyl cyclase reaction was initiated by adding substrate

m~ Me, and 0.01% Lubrol for 20 min at 25 "C. Data were fitted to a and was carried out in 100 pl at final concentrations of 2 p~ GTPyS, 10

rectangular hyperbola (panel A) or were linearized by Scatchard trans- formation (panel B ) . The EC,, value for preactivated rGsm.L and rGB(I.B in

was performed twice with similar results. stimulating adenylyl cyclase amounts to -2 nmol/liter. This experiment

some but not all forms of adenylyl cyclase (33). Addition of By (500 nmoyliter) to oGTP-treated platelet membranes had only a minor effect in the presence of forskolin (Fig. 1OA). However, in the presence of GTPyS-liganded rG,,.,, By produced an additional stimulation (Fig. 1OB). We have verified that the P y preparations employed were in principle capable of inhibiting an adenylyl cyclase isoform. The equivalent amount of By, at a maximally active concentration, inhibited purified bovine brain adenylyl cyclase by 30% (Fig. 1OC). This preparation contains predominantly the type I isozyme (34).

DISCUSSION

In the present work, the action of the 2',3'-dialdehyde analogue of GTP, oGTP, on signaling pathways has been investigated in isolated membranes and in whole cells. We demonstrate that oGTP disrupts receptor-effector coupling in a set of signal transduction systems whose activation is known to in- volve several G proteins, i.e. G,, Gi, and GJG,. oGTP persis- tently antagonizes effector regulation induced by guanine nucleotides or via receptor activation, including three distinct effector mechanisms (stimulation and inhibition of adenylyl cyclase, regulation of the voltage-sensitive Ca2+-channel) and several types of G protein-coupled receptors (&-adenosine, PGI,, a,-adrenergic, GABA,, and somatostatin). Inhibition by oGTP is irreversible in that it cannot be relieved through washing of the membranes through subsequent addition of excess concentrations of GTPyS or through receptor activation which maximally enhances guanine nucleotide exchange in native G proteins. The oGTP-induced disruption of transmembrane signaling results from the specific interaction with the guanine nucleotide binding pocket of G protein a-subunits. This conclu- sion is based on the following findings. (i) Labeling of G protein a-subunits by [a-32P]GTP azidoanilide is suppressed upon pretreatment of platelet membranes with oGTP. (ii) When added concomitantly, GTPyS prevents the effect of oGTP in whole cells. (iii) Time course experiments reveal that in the whole cell patch-clamp configuration the interval to onset of action is comparable for GDPpS and for oGTP, i.e. G protein inactivation occurs gradually with time, and the inactivation rate is determined primarily by association with the guanine nucleotide binding site. (iv) Once established, the block by oGTP cannot be

overcome by guanine nucleotides, and reconstitution of the signaling pathway requires addition of G protein a-subunits. All these observations in membranes and whole cells are consist- ent with the kinetics of oGTP binding to purified rGBa described in the accompanying paper (7).

Reconstitution of adenylyl cyclase stimulation with exogenous G,, indicates that receptor and effector are free to interact physically with newly inserted GB,. This interaction fully re- stores the regulation of adenylyl cyclase as CAMP generation approximates the levels observed in control membranes. Ad- enylyl cyclase associates tightly with activated G,, resulting in copurification of the complex (35,361. Thus, oGTP destabilizes the G,,-adenylyl cyclase complex, presumably by favoring formation of the inactive G protein heterotrimer. Using oGTP- pretreated membranes, we are in the position to test the rel- evance of the differential expression of splice variants for receptor- and guanine nucleotide-dependent activation of membrane-bound adenylyl cyclase. As predicted from experiments with the purified brain enzyme and S49 cyc- membranes (lo), the human platelet adenylyl cyclase does not display preferen- tial interaction with either splice variant of G8,. With non- preactivated G,,, however, the apparent affinity is higher for the long than for the short from. This difference is determined by the intrinsic activation rates of the two forms of GSe. We therefore conclude that basal adenylyl cyclase activity in platelet membranes primarily depends on Gsu.L. Our observations also further support the hypothesis that, irrespective of the isoform of adenylyl cyclase involved, the regulated expression of individual G,, splice variants may serve to fine tune the basal level of CAMP turnover in different tissues.

The &-adenosine receptor, however, does not differentiate between G,,., and Gaa.L. A similar lack in discrimination has been observed previously for the human P-adrenergic receptors when expressed in Escherichia coli and reconstituted with the two forms of G,, (37). Likewise, the estimates for the affinity between receptor and G,, (-1-3 m) obtained in both studies are comparable (37). Reconstitution of native receptors with recombinant purified a-subunits has provided an experimental approach to test G protein subtype specificity in receptor coupling. Several types of receptors coupling to G proteins of the Guo class have thus displayed a clear-cut selectivity, i.e. the bovine brain A,-adenosine receptor and the human 5HT,,- receptor for Gia.3 (38, 39) and the pituitary D,-dopamine receptor for Gi,.2 (40). The &-adenosine receptor has long been known to activate adenylyl cyclase. However, activation rates are incompatible with the collision-coupling model, which ad- equately describes normal receptor-G, interaction (41). Fur- thermore, guanine nucleotides fail to modulate the binding of agonists to the &-adenosine receptor raising the possibility that a G protein distinct from G, mediates the action of the receptor (4243). Using the present approach of disruption and reconstitution, we, nonetheless, provide direct evidence that the &-adenosine receptor interacts with Gsw. Mammalian cD- NAs coding for &-adenosine receptors have been cloned (44, 45). Evaluation of the peculiarities in the interaction between &-adenosine receptor and G, will be feasible once the receptor protein is expressed in an appropriate system.

In contrast, neither addition of a-subunits of the G, class nor of By-dimers reconstituted inhibition of adenylyl cyclase, but produced a pattern of stimulation which has recently been observed as a characteristic feature of type I1 adenylyl cyclase (46). Hence, the mechanism, which causes inhibition of the type I1 isoform and type 11-like platelet adenylyl cyclase, remains elusive. Several lines of evidence point to Gi, as the principal mediator of cyclase inhibition in platelet membranes. (i) Recep- tor-induced attenuation of adenylyl cyclase activity involves a


T A

L

600 - d

'2 450 2

E L E 300 0

fi

150 a E

v

3 0

FIG. 10. Effect of recombinant Gi, subtypes and of P-y on adenylyl cyclase inhibition in oGTP-treated platelet membranes. Panel A, platelet membranes were pretreated with oGTP at low Mg2t concentrations (1 mM) to disrupt endogenous adenylyl cyclase inhibition as outlined under "Experimental Procedures." rGim.l, rG,,,, rGiu.3, and rG,, were preactivated with 10 w GTPyS in HED, 25 m CHAPS, 10 m MgSO, at 25 "C for 90 min. 10 pmol of a-subunit were subsequently added to oGTP-treated platelet membranes (20 pg). Alternatively, 2 pg of py was added in the same buffer without GTPyS. The mixture was preincubated on ice for 30 min, and the assay was carried out for 20 min at 25 "C in a final volume of 0.1 ml. The respective reference 100% value was determined in the presence of the appropriate buffer to control for carryover of detergent and magnesium with and without GTPyS. In the presence of forskolin (20 w) and forskolin (20 w) + GTPyS (1 w), adenylyl cyclase activity amounted to 298 2 52 and 315 2 49 pmol/mg/min CAMP (means f S.E.; n = 3), respectively. Since interassay variation ranged from 170 to 370 pmol/min/mg, data are given in percent of the respective base-line value. Panel B, effect of py in the presence of rG,, on adenylyl cyclase in oGTP-treated platelet membranes. Platelet membranes were used that had been pretreated with oGTP at high M e concentrations (see above). rG,,, was preactivated in the presence of GTPyS (10 m) and MgSO, (10 mM) for 30 min at 25 "C; 0.3 pg of rG80.L was added to oGTP-treated membranes (-20 pg of membrane protein) in the absence or presence of Py (2 pg) as indicated, and the assay was carried out as described in the legend to Fig. 9. Results are shown from a single experiment out of two reconstitution experiments. Panel C , py directly inhibits the catalytic activity of purified bovine brain adenylyl cyclase. Purified preparations of bovine brain py-dimer (1.5 pg) used in panels A and B were added to -100 ng of purified bovine brain adenylyl cyclase in the presence of 0.02% final Lubrol; the assay was carried out for 5 min at 30 "C in a final volume of 0.1 ml. GTPyS (10 m) did not alter basal activity levels. The means 2 S.E. from three independent experiments are shown.

pertussis toxin-sensitive mechanism. (ii) a,-Adrenergic inhibition of adenylyl cyclase is reduced in the presence of antibodies directed against the carboxyl terminus of Gia., (29,47). (iii) The a,-adrenergic receptor displays an overwhelming coupling preference for Gi over G, and G, (48). Inhibition of type I1 adenylyl cyclase by G,,, has been previously demonstrated in whole cells co-transfected with the catalyst and a mutant-activated form of the a-subunit (49). Two hypothetical explanations are avail- able. It has recently been reported that Go, and Gi,s are palmitoylated (51, 52). Myristoylation of Gi,s is essential for inhibition of several isoforms adenylyl cyclase (52). An additional lipid modification such as palmitoylation may be required for inhibition of the type I1 isozymes. Obtaining palmitoylated preparations of G protein a-subunits may be difficult due to the lability of the thioester linkage of the palmitate. Alternatively, the inhibitory pathway in platelets may require an ancillary factor, beside the G protein, to attenuate catalyst activity. Sup- port for this assumption is based on the finding that an intrinsic GTP-dependent nucleoside diphosphate kinase, which is blocked by UDP and supported by phosphorylated fly-dimers, amplifies adenylyl cyclase inhibition in platelet membranes (53, 54). oGTP is likely to inhibit the catalytic activity of this nucleoside diphosphate kinase.

Our observations suggest that oGTP may act as a universal G protein antagonist. oGTP provides an alternative approach of disruption (and reconstitution) of G protein-mediated signaling, which is analogous to the use of pertussis toxin for GjG,, in native systems (2, 3). With oGTP, this strategy is presumably applicable to the entire class of G proteins. It may thus be useful to delineate the role of newly identified G protein a- subunits as well as other GTP-binding proteins in signaling pathways.

ous gift of antisera CS1 and AS7 and to Dr. M. E. Linder for providing Acknowledgments-We are grateful to Dr. G. Milligan for the gener-

purified recombinant Gi, and Go, and for critical reading of the manu- script. We thank G. Koth, A. Motejlek, and K. Schwarz for excellent technical assistance in the preparation of embryonic neuronal cells for

electrophysiological procedures. We also thank E. Tuisl and E. Weis for skillful artwork.

REFERENCES 1. Simon, M. I, Strathmann, M. P., and Gautam, N.(1991) Science 262,802-808 2. Spiegel, A. M., Shenker, A,, and Weinstein, L. S. (1992) Endocrine Reu. 13,

3. Milligan G. (1993) fiends Pharmacol. Sci. 14,239-244 4. Mukai, H., Munekata, E., and Higashijima, T. (1992) J. Biol. Chem. 267,

5. Oppi, C. Wagner, T., Crisari, A., Camerini, B., and Tocchini Valentini, G. P.

6. Voss, T., Wallner, E., Csernilofsky, A. P., and Freissmuth, M. (1993) J. Biol.

7. Hohenegger, M., Nanoff, C., Ahorn, H., and Freissmuth, M. (1994) J. Biol.

8. Offermans, S., Schultz, G., and Rosenthal, W. (1991) Methods Enzymol. 195,

9. Linder, M. E., and Mumby, S. M. (1994) Methods Emymol. 237,254-268

536565

16237-16243

(1992) Proc. Natl. Acad. Sci. U. S. A. 89, 8268-8273

Chem. 268,46374642

Chem. 269,3200%32015

286301

10. Graziano, M. P., Freissmuth, M., and Gilman, A. G. (1989) J. Biol. Chem. 264,

11. Casey, P. J., Graziano, M. P., and Gilman, A. G. (1989) Biochemistry 28,611-

12. Pfeuffer, E., Mollner, S., and PfeufTer, T. (1985) EMBO J. 4,3675-3679 13. Towbin, H., Staehelin, T., and Gordon J. (1979) Proc. Nutl. Acad. Sci. U. S. A.

14. Milligan, G., Unson, C. G., and Wakelam, M. J. 0. (1989) Biochem. J. 262,

15. Goldsmith, P., Gierschik, P., Milligan G., Unson, C. G., Vinitsky, R., Malech, H.

16. Hiittemann, E., Ukena, D., Lenschow, I?, and Schwabe, U. (1984) Naunyn

17. Gilman A. G. (1987) Annu. Reu. Biochem. 56,615449 18. Offermanns, S., Schultz G., and Rosenthal, W. (1991) J. Biol. Chem. 266,

19. Boehm, S., and Huck, S. (1991)Naunyn-Schmiedeberg's Arch. Pharmacol. 344,

20. Boehm, S., Huck, S., Drobny, H., and Singer, E. A. (1992) Naunyn-

21. Hamill, 0. P., Marty, A., Neher, E., Sakman, B., and Sigworth, F. J. (1981)

409-418

616

76,4350-4354

643449

L., and Spiegel, A. M. (1987) J. Biol. Chem. 262, 14683-14688

Schmiedeberg's Arch. Pharmacol. 326,226233

3365-3368

382-385

Schmiedeberg's Arch. Pharmacol. 345,606-609

Pfliigers Arch. 391,85-100 22. Diverse-Perluissi, M., and Dunlap K. (1993) Neuron 10, 753-760 23. Elmslie, K. S., Werz, M. A,, Overholt, J. L., and Jones, S. W. (1993) Pfliigers

Arch. 423. 472479 24. Pusch, M., and Neher, E. (1988) Pfliiger's Arch. 411,204-211 25. Carbone, E., and Swandulla, D. (1989) Prog. Biophys. Mol. Biol. 54, 31-58 26. Johnson, R. A., and Salomon, Y. (1991) Methods Enzymol. 195,3241

28. Freissmuth, M., and Gilman, A. G. (1989) J. Biol. Chem. 264, 21907-21914 27. Freissmuth, M., Selzer, E., and Schiitz, W. (1991) Biochem. J. 275, 651-656

29. Simonds, W. F., Goldsmith, P. K., Codina, J., Unson, C. G., and Spiegel, A. M.


30. Bwhm, S., Nanoff, C., Hohenegger, M., Freissmuth, M., and Huck, S. (1993)

31. Golard, A,, and Siegelbaum, S. A. (1993) J. Neurosci. 13, 3884-3894 32. Boehm, S., Huck, S., Koth, G., Drobny, H., Agneter, E., and Singer, E. A. (1994)

33. Hepler, J., and Gilman, A. G. (1992) Dends Biochem. Sci. 17,383-387 34. Krupinski, W. J., Coussen, F., Bakalyar, H. A., Tang, W. J., Feinstein, P. G.,

35. Pfeuffer, E., Dreher, R. M., Metzger, H., and Pfeuffer, T. (1985) Proc. Natl.

36. Morris, D., McHugh-Sutkowski, E., Moos, M., Simonds, W. F., Spiegel, A. M.,

37. Freissmuth, M., Selzer, E., Marullo, S., Schiitz, W., and Strosberg, A. D. (1991)

(1989) Proc. Natl. Acad. Sei. U. S. A. 86,7809-7813

Soc. Neurosci. 19, 1263 (abstr.)

J. Neurochem. 63, 146154

Orth, K., Reed, R. R., and Gilman, A. G. (1989) Science 244, 1558-1564

Acad. Sci. U. S . A. 82,30863090

and Seamon, K B. (1990) Biochemistry 29,9079-9084

Proc. Natl. Acad. Sci. U. S. A. 88, 8548-8552

17783 38. Freissmuth, M., Schutz, W., and Linder, M. (1991) J. Biol. Chem. 266, 17776

39. Bertin, B., Freissmuth, M., Breyer, R., Schiitz, W., Strosberg, A. D., and

40. Senogles, S. E., Spiege1,A. M., Padrell, F., Iyengar, R., and Caron, M. G. (1990)

41. Braun, S., and Levitzky, A. (1979) Biochemistry 18,2134-2138 42. Nanoff, C. , Jacobson, K. A., and Stiles, G. L. (1991) Mol. Phurmacol. 39,

Mamllo, S. (1992) J. Biol. Chem. 267, 8200-8206

J. Biol. Chem. 266,4507-4514

43. Nanoff, C. , and Stiles, G. L. (1993) J. Rec. Res. 13,961-973 44. Maenhaut, C., van Sande, J., Libert, F., Abramowicz, M., Parmentier, M.,

Vanderhaegen J. J., Dumont, J. E., Vassart, G., and Schiffmann, S. (1990)

45. Pierce, K. D., Furlong, T. J., Selbie, L.A., and Shine, J. (1992) Biochem. Biochem. Biophys. Res. Commun. 173,1169-1178

130-135

46. Taussig, R., Tang, W. J., Hepler, J. R., and Gilman, A. G. (1994) J. Biol. Chern. Biophys. Res. Comiun. 187,8693

47. McClue, S., Selzer, E., Freissmuth, M., and Milligan, G. (1992) Biochem. J.

48. Chabre, O., Conklin, B. R., Brandon, S., Bourne, H. R., and Limbird, L. E.

49. Chen, J., and Iyengar, R. (1993) J. Biol. Chem. 268, 12253-12256 50. Linder, M. E., Middleton, P., Hepler, J., Taussig, R., and Gilman, A. G. (1993)

51. Parenti, M., Vigano, M. A., Newman, M., Milligan G., and Magee, A. I . (1993)

52. Taussig, R., Iiiiguez-Lluhi, J. A,, and Gilman, A. G. (1993) Science 261,218-

53. Jakobs, K. H., and Wieland, T. (1989) EUK J. Biochem. 183,115-121 54. Wieland, T., Ronzani, M., and Jakobs, K. H. (1992)J. Biol. Chem. 267,20791-

269,6093-6100

284,565-568

(1994) J. Biol. Chem. 269,5730-5734

Proc. Natl. Acad. Sci. U. S . A. 90, 3675-3679

Biochem. J. 291,349-353

221

20797

the of biol~cical vol. no. of pp. 0 for in u.s.a. 2’,3 ... · the joufwal of biol~cical chemistry...

Documents