Mechanism of beta-adrenergic receptor kinase activation by G proteins.

THEJOURNALOF BIOLOGICAL CHEMISTRY Q 1993 by The American Society for Biochemistryand Molecular Biology, Inc.

Vol. 268, No. 21, Issue of July 25, pp. 15412-15418,1993 Printed in U.S. A.

Mechanism of ,@-Adrenergic Receptor Kinase Activation by G Proteins* (Received for publication, December 30, 1992, and in revised form, March 30, 1993)

Chong M. Kim$, Stephane B. Dions, and Jeffrey L. Benovicll From the Department of Pharmacolozv, JeffersonCancer Institute, Thomas Jefferson University, Philadelphia, Pennsylvania 19107 I.

The t9-adrenergic receptor kinase (&ARK) specifically phosphorylates the activated form of various G protein-coupled receptors such as the &adrenergic receptor (Bz-AR).Recently, G protein j3r subunits have been demonstrated to activate B-ARK-mediatedreceptor phosphorylation. To further elucidate B-ARKIG protein interactions, we have developed a direct binding assay. The direct binding of [3BS]methionine-labeled @-ARKto either brain GJG, or @rsubunits was rapid and saturable with similar Kdvalues of -58 and -32 nM, respectively. Both heterotrimeric G proteins and fly subunits enhanced the initial rate of &AR and rhodopsin phosphorylation -10-fold. Kinetic studies demonstrate that @renhances @-ARK-mediated&-AR phosphorylation both by decreasing the K,,, for the j3,AR -4-fold and increasing the stoichiometry of phosphorylation from -4 to -11 mol/mol. An agonist- and ATP-dependent binding of &ARK to the reconstituted &-AR was also demonstrated. In addition, &ARK binding was enhanced in the presence of both the activated &AR and B-y subunits suggesting the formation of a transient ternary complex consisting of @-ARK, By, and &AR. Overall, these studies suggest that the specific association of @-ARK with heterotrimeric G proteins may play an important role in promoting receptorlkinase interaction and subsequent receptor phosphorylation.

The &-adrenergic receptor (P2-AR)’-coupledadenylylcyclase system has provided an important model for characterizing the molecular mechanisms of signal transduction. In this system, agonist binding to the&AR promotes activation of the G protein G, leading to increased adenylylcyclase activity and intracellular cAMP levels. However, prolonged agonist stimulation of the &AR results in an attenuation of cAMP generation (Benovic et al., 1989b; Hausdorff et al.,

* This research was supported in part by Grants GM44944 and HL45964 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 4 Supported by the M.D./Ph.D. Program, Temple University School of Medicine, Philadelphia, PA 19140. 3 Supported by a fellowship from the Heart andStroke Foundation of Canada. V To whom correspondence should be addressed. Tel.: 215-9554607; Fax: 215-923-1098. The abbreviations used are: &AR, &adrenergic receptor; App(NH)p, 5’-adenylyl-@y-imidodiphosphate;8-ARK, 8-adrenergic receptor kinase; PYB,By subunits from brain; PYT, @r subunits from retina; G protein, guanine nucleotide-binding protein; GTPrS, guanosine 5’-(3-0-thio)triphosphate;Gi,,,, G proteinsthat are abundantin brain and have the composition Rho’, light-activated rhodopsin; ROS, rod outer segments.

1990). This attenuation of second messenger generation is termed desensitization and is attributed to various molecular mechanisms. Rapid agonist-specific or homologous desensitization is due to a specific agonist-dependent uncoupling of the activated receptor from the G protein transducer. The padrenergic receptor kinase (&ARK) appears to play a major role in homologous desensitization of the &AR, likely due to its ability to specifically phosphorylate the agonist-occupied form of the receptor (Benovic et al., 1986). P-ARK-mediated receptor phosphorylation has been demonstrated to enhance the association of the proteinp-arrestin with the ,R2-AR, thereby uncoupling the receptor from G, (Lohse et al., 1990). A role for @ARK in regulating additional receptor/effector systems is suggested by its ability to phosphorylate the agonist-occupied platelet a,-adrenergic receptor (Benovic et al., 1987b) and chick heart muscarinic acetylcholine receptor (Kwatra et al., 1989). Although specific recognition of receptor subtypes provides an important determinant for P-ARK phosphorylation, multiple pathways may exist for conferring specificity. The increase in membrane-associated fi-ARK activity in response to various receptor agonists may play a role in conferring specificity (Strasser et al., 198% Mayor et al., 1987; Chuang et al., 1992). Recent reports suggest that G protein py subunits may be involved in the membrane localization and regulation of pARK. Using a porcine P-ARK-related kinase preparation, Haga and Haga (1989, 1992)demonstrated that G protein subunits stimulatedrhodopsin and m2 muscarinic cholinergic receptor phosphorylation -12-fold. No stimulation of receptor phosphorylation was achieved with the G, subunit alone. More recently, studies utilizing recombinant bovine /3-ARK demonstrated that By subunits enhanced P-ARK-mediated &-AR and rhodopsin phosphorylation -10-fold (Pitcher et al., 1992). Brain P-y subunits were -10-fold more potent than retinal By subunits in activating 8-ARK-mediated Pz-AR phosphorylation. The binding of Pr subunits to a fusion protein containing the carboxyl-terminal 222 amino acids of p-ARK was also demonstrated. These studiesstrongly support an integral role for G protein Py subunits in p-ARK-mediated desensitization of G protein-coupled receptors. The molecular mechanisms by which G protein @-y subunits enhance p-ARK-mediated receptor phosphorylation remain unclear. Unresolved questions regarding the kinetic mechanism of &ARK activation, the membrane localization of pARK, potential direct activation by By subunits, and &ARK interaction with heterotrimeric G proteins need to be further addressed. In thisstudy, we have used [35S]methionine-labeled @-ARKto investigate the direct interaction of 0-ARK with heterotrimeric G proteins, isolated /3r subunits, and purified P2-AR. The consequences of the various interactions were assessed by phosphorylation of receptors and nonreceptor substrates such as casein and synthetic peptides. The ability to quantitatively measure direct protein interactions provide

15412

G Protein Activation of P-ARK

15413

(w/w), 65 mM NaC1, 2 mM MgCl,, 10 mM Tris acetate buffer, pH 7.4, shaken vigorously, and centrifuged at 2,000 X g for 5 min. The of p-ARK. supernatant was diluted with 2 volumes of 10 mM Tris acetate buffer, pH 7.4, and centrifuged as above. The crude ROS pellets were EXPERIMENTALPROCEDURES resuspended in -30 ml of 0.77 M sucrose, 1 mM MgClz, 10 mM Tris Materials-Most chemicals were from sources previously described acetate buffer, pH 7.4, and purified further on a stepwise sucrose (Benovic et al., 1987a, 1989a). The chromatography resins S-Sepha- gradient. The interface between 0.84 and 1.00 M sucrose was collected (-25 ml), diluted with 25 ml of buffer, and centrifuged at 48,000 X g rose, heparin-Sepharose, and Mono S were from Pharmacia LKB Biotechnology Inc. Frozen bovine retinas were from George A. Hor- for 20 min. Rhodopsin kinase-free membranes were prepared as me1 and Co., whileendopeptidase Asp-N was from Boehringer Mann- described (Shichi and Somers, 1978). Briefly, purified ROS were heim. Partially dephosphorylated casein, (-)-isoproterenol, and soy- suspended in 50 mM Tris-HC1, pH 8.0,5 mM EDTA, 5 M urea (1ml/ bean phosphatidylcholine were from Sigma. [-y3'P]ATP, [35S]methi- retina), sonicated on ice (4 min, maximum power), diluted with 2 onine, and ['Z51]iodopindololwere from Du Pont-New England volumes of 50 mM Tris-HC1, pH 7.4, and centrifuged at 100,000 X g Nuclear. Tissue culture reagents were purchased from Life Technol- for 45 min. The pellet was washed three times with Tris buffer before ogies Inc. and Sigma. All other reagents were of the highest grade final resuspension in 50 mM Tris-HC1, pH 7.4, sonication on ice, and quick freezing. All operations were carried out in the dark or under commercially available. Purification of P-ARK-Bovine &ARK was overexpressed and dim red light. Urea-treated ROS showed negligibleendogenous kinase purified from infected Sf9 cells using the baculovirus expression activity and consisted of >90% rhodopsin as assessed by Coomassie system (Kim et al., 1993). Briefly, Sf9 cells were harvested 48 h post- Blue staining of polyacrylamide gels. The concentration of rhodopsin infection bylow speed centrifugation. The cells were washed and was determined in the presence of digitonin using a molar extinction then homogenized in 20 mM Hepes-HC1, pH 7.2,250 mM NaCl, 5 mM coefficient of 40,600 at 498 nm (Wald and Brown, 1953). Preparation of Truncated 329G-Rhodopsin-A truncated form of EDTA, 3 mM phenylmethylsulfonyl fluoride, 3 mM benzamidine (50 ml of buffer/liter of cells). A high speed supernatant fraction was rhodopsin with 19 amino acid residues proteolytically removed from then diluted with buffer A (20 mM Hepes-HC1, pH 7.2, 5 mM EDTA, the COOH terminus (329G-Rho)was made as previously described 0.02% Triton X-loo),loaded onto an S-Sepharose column, and then (Palczewski et al., 1991). Briefly, urea-treated ROS were incubated eluted with a linear gradient from 50 to 300 mM NaCl in buffer A. with endopeptidase Asp-N in 10 mM Tris-HC1, pH 7.5, in the dark Peak fractionswere pooled,diluted, loaded onto a heparin-Sepharose for 16 h at room temperature. The reaction was quenched by the column, and then eluted with a linear gradient from 100 to 600 mM addition of 1 mM dithiothreitol and 1 mM EDTA. The truncated rhodopsin was washed, centrifuged, and finally resuspended in 50 mM NaCl in buffer A. Radiolabeling of @-ARK with f5S]Methionine-At 24 h post-infec- Tris-HC1, pH 7.4. Peptide Synthesis and Phosphorylation-Peptides were synthesized tion, 1 liter of infected Sf9 cells were centrifuged and resuspended in 70 ml of methionine-free Graces media containing 1-2 mCi of [35S] on an Applied Biosystems 430A synthesizer using Fmoc (N-(g-flumethionine (1190Ci/mmol). Following an additional 20 h in a shaking oreny1)methoxycarbonyl) chemistry. Peptides were purified by high incubator at 27 "C, the cells were harvested and P-ARK was purified performance reverse-phase chromatography on a C-18 column using as described above. A finalchromatography step on a Mono S column a 0-50% acetonitrile gradient in 0.1% trifluoroacetic acid. A stock yielded -1 mgof radiolabeled @-ARKat >95% purity as judged by peptide solution was prepared and adjusted to pH7.4 with Tris base. SDS-polyacrylamide gel electrophoresis. The radiopurity was as- Varying concentrations of peptide (0-100 p M ) were incubated with 0sessed by autoradiography of the polyacrylamide gel and also judged ARK (20-100 nM), @-y subunits (0-200 nM), Gi/. (0-200 nM), &-AR (0-10 nM), 329G-Rho(0-2.5 p M ) , and (-)-isoproterenol (0-25 p M ) in to be >95%. The radiolabeled P-ARK had a specific activity of -2000 cpm/pmol as determined by protein analysis (Bradford, 1976) using a buffer containing 20 mM Tris-HC1, pH 7.5, 2 mM EDTA, 5 mM bovine serum albumin as standard andmeasurement of radioactivity MgC12, 0.1 mM [-y3*P]ATP(0.5 cpm/fmol) in a final volume of 50 pl at 30 "C. Reactions were stopped by the addition of30 pl of 30% with scintillant. Purification and Reconstitution of @-AdrenergicReceptor-The hu- trichloroacetic acid. The quenched reactions were centrifuged at man &adrenergic receptor was also expressed using the baculovirus 48,000 X g for 10 min and theresulting supernatants were transferred expression system. The human &-AR cDNA was subcloned into the to a 2 X 2-cm square of P-81 paper followed by fivewashes in 75 mM vector pBluebac and then cotransfected into Sf9 cells with wild-type phosphoric acid as previously described (Cook et al., 1982). @-ARK AcNPV DNA to produce the recombinant baculovirus. For routine activity was defined as the difference in phosphate incorporation in receptor purification, 2 liters of Sf9 cells were infected with recom- the presence and absence of peptide. Assay of P-ARK-mediated Receptor Phosphorylation-Phosphorylbinant virus a t a multiplicity of infection of 2:l and cells were harvested 48 h post-infection. The cells were lysed by Dounce ho- ation reactions contained, in a total volume of15-100 pl, either mogenization and the membrane fraction pelleted by centrifugation reconstituted Pz-AR (10-200nM) or urea-treated ROS (40-2000 nM), a t 33,000 X g. The membranes were solubilized with a 5:l ratio of @-ARK(4-50 nM), P-y subunits (0-150 nM), Gi,, (0-150 nM), and (-)digitonin to protein and the P,-AR was purified by affinity chroma- isoproterenol (0-50 p M ) in 20 mM Tris-HC1, pH 7.5, 2 mM EDTA, 5 tography on an alprenolol-Sepharose column (Benovic et al., 1984). mM MgCl,, and 0.1 mM [-y3'P]ATP (0.1-2 cpm/fmol). Incubations The receptor concentration was determined by radioligand binding were exposed to fluorescent laboratory lighting at 30 "C for 1-3 min using [1251]iodopindolol. The purified receptor was reconstituted into for rhodopsin or 5-100 min for the &AR. Reactions were stopped soybean phosphatidylcholine vesicles as previously described (Cer- with 10-20 pl of SDS sample buffer (8% SDS, 25 mM Tris-HC1, pH ione et al., 1983).Vesicles containing the reconstituted receptors were 6.5,10% glycerol, 5% mercaptoethanol, and 0.003% bromphenol blue) resuspended in 20 mM Tris-HC1, pH 7.2, 2 mM EDTA and used as a and electrophoresed on 10%SDS polyacrylamide gels (Laemmli, substrate for purified @-ARK. 1970). After autoradiography, the 32P-labeled receptor bands were Purification of G Proteins-The GTP-binding proteins Go and Gi excised and counted to determine the pmol of phosphate transferred. were purified from bovine brain by the method of Sternweis and In some rhodopsin phosphorylation studies, the reactions were Robishaw (1984). Briefly, bovine brain membranes were solubilized stopped by the addition of 30% trichloroacetic acid. The quenched with 1%cholate, centrifuged, and theresulting supernatant was then reactions were filtered throughWhatman GF/B filter paper and successively chromatographed on DEAE-Sephacel, Sephacryl S200, washed with 10 mM phosphate buffer, pH 7.0, using a Brandel Cell and heptylamine-Sepharose. The G protein preparation, consisting Harvester. The filters were then directly counted to determine the primarily of Gi and Go,was further purified by chromatography on a pmol of phosphate transferred. Mono Q column and then stored in 20 mM Tris-HC1, pH 8.0, 1 mM @-ARKBinding to G Proteins and Receptors-Direct binding was EDTA, 1mM dithiothreitol, 150 mM NaCl, and 0.05% Lubrol (buffer assessed in reaction volumes of 100-300 pl with 35S-labeled@-ARK B) at -80 "C. P-y Subunits were isolated by chromatography of the (5-100 nM), @rsubunits (0-44 nM), Gi/, (0-20 nM), and GTP+ (0purified Gproteins on heptylamine-Sepharose in the presence of 25 p M ) in a buffer containing 100 pg/ml sonicated soybean phosphaAMF (30 p M AlCl3, 6 mMMgC12, and 10 mM NaF) (Katada et al., tidylcholine, 20 mM Tris-HC1, pH 7.5,2 mM EDTA, and 5 mM MgCl,. 1984). 0-y Subunits free of AMF and cholate were obtained by final The Gi,, or @rsubunits were directly added to the binding reaction anion exchange chromatography on a Mono Q column. The purified already containing the phospholipid vesicles or reconstitutedreceptor. P-y was stored in buffer B a t -80 "C. The binding reactions were incubated at 30 "C for 30 min followed Preparation of Urea-treated Rod Outer Segments-Rod outer seg- by centrifugation at 70,000 X g for 10 min. The phospholipid/protein ments (ROS) were prepared essentially as described (Wilden and pellet was solubilized with 1% SDS and counted with scintillation Kuhn, 1982). Fifty retinas were suspended in 50 ml of 34% sucrose fluid. Specific binding was the difference between phosphatidylchosignificant insight into the mechanism of G protein activation

Activation G Protein

15414

of @-ARK

line vesicles with or withoutG proteins. Nonspecific binding was 1020% of the total binding. Direct binding to the &-AR (0-14 nM) was obtained under similar conditions with the addition of (-)-isoproterenol (0-100 p ~and ) App(NH)p(0-100 p ~ )Nonspecific . binding was determined using phospholipid vesicles taken through the reconstitution procedure in the absence of receptors.

to be 32 f 5 nM (Fig. 2 A ) and to heterotrimericG proteins to be 58 k 14 nM (Fig. 2B). Further evidence that G, does not significantly affect @-ARKbinding to Py subunits was provided by including GTPyS in the binding assay. Dissociation of heterotrimeric G proteins into the respective a and @y subunits has been demonstrated to be greatly enhanced by RESULTS et al., nonhydrolyzable G T P analogs such as GTPyS (Katada 1984). Thus, if G, competes with &ARK for binding to By, Previous studies have suggested a role for G protein @y dissociationtriggered by GTPyS should subunits in the activation and possible membrane association then the subunit of @-ARK (Haga and Haga, 1992; Pitcher et al., 1992). To result in an increase in total @-ARK binding. However, no further elucidate the mechanism of P-ARK activation by G appreciable difference in (3-ARK binding toGi,, was observed in the absence or presenceof GTPyS (Fig. 2B). These direct protein @ysubunits, we have utilized the inherent property binding studies demonstrate that @-ARK interacts with both of brain @y subunits to tightly associate with phospholipid vesicles (Sternweis, 1986). This enabled us to develop a bind- heterotrimeric G proteins andBy subunits with a comparable ing assay that involved incubating [35S]methionine-labeled@- high affinity. Since equivalent binding of P-ARK to heterotrimeric G ARK with G protein @ysubunits in the presenceof phosphatidylcholine lipid vesicles. The phospholipidvesicles, contain- proteins and @ysubunits was observed, we also compared the ing Py and @ y @-ARK . complexes,were then rapidly pelleted by centrifugation. Most (>85%) of the Py subunits in the binding reaction pelleted with the phospholipid vesicles as assessed by polyacrylamide gel electrophoresis while free PARK remained in the supernatant. The amount of P-ARK that was associated with the@ysubunits was then determined by measuring the radioactivity in the pellets. In this assay system 35S-labeled @-ARKserved as a specific radioligand for probing @-ARKinteraction with various membrane proteins such asG proteins and receptors. @-ARK binding to either heterotrimeric G proteins (Gila) or t o purified @y subunits was rapid and saturable, with halfmaximal binding achieved in less than 1 min (Fig. 1).The binding was equivalent whether heterotrimeric G proteins or assay. This isolated @y dimers were used inthebinding equivalent binding was unexpected since earlier reports suggested that G, subunits competed with @-ARK for binding to @ysubunits (Pitcher et al., 1992). Additional binding studies confirmed that the affinities of P-ARK for brain By subunits andheterotrimeric G proteinsare very similar.Scatchard Kd for P-ARK binding toBy subunits analysis determined the 2.0

-

1.0

.

0.8

.

Y

a

@z

El

-

0.6

E

n

v

PARK (nn) I

0.0

I

I

FIG. 2. A , direct binding of 35S-labeled 0-ARK to brain 0-y subunits. Direct binding was assessed in a total reaction volume of 200 11, containing 44 nM (3-y subunits, 100 pg/ml phosphatidylcholine,5 FIG. 1. Time course of direct binding of 36S-labeled&ARK mM MgC12,20mM Tris-HC1, pH 7.5, 2 mM EDTA, and 10-100 nM to & subunits and Gi,.. Directbindingwasassessedin a total &ARK (300 cpm/pmol). Specific (3-ARK binding was determined as reaction volume of 3 ml containing 20 nM brain 0-y (0)or Gila (a), described in the legend to Fig. 1. Data points arethe mean -+ S.E. of 20 nM 0-ARK (1000 cpm/pmol), 5 mM MgC12,100 pg/ml phosphati- three separate experiments. The Scatchard analysis is shown in the 0-ARK t o Gil, in the absence or dylcholine, 20mM Tris-HC1, pH 7.5, and 2 mM EDTA. The reaction inset. B , direct binding of 35S-labeled mixtures were incubated at 30 "C and at the times indicated, 240-pl presence of GTPrS. Direct binding was assessed in a total reaction aliquots were removed and centrifuged at 70,000 X g for 10 min. The volume of 200 pl, containing 17 nM G;lo,100 pg/ml pbosphatidylchop~ GTP-yS, 5 mM MgC12,20 mM Tris-HC1, pH pellets were solubilized with 1%SDS and counted with scintillant. line, 0 ( 0 )or 25 (0) Specific bindingwas the difference between phosphatidylcholine ves- 7.5, 2 mM EDTA, and 5-100 nM &ARK (2000 cpm/pmol). Specific icles with or without G proteins. Data points are the mean k range 0-ARK binding was determined as described in Fig. 1. Data points are the mean ? range of two separate experiments done in duplicate. of two separate experiments done in duplicate. 0

10

20

30

t ~ m e( m l n )

40

50

G Protein Activation of P-ARK ability of G proteins and @ysubunits to directly activate @ARK-mediated receptor phosphorylation. When testedat 1:l molar ratios (G protein:@-ARK) both Gi/, and @y subunits stimulated the initial rateof &AR phosphorylation -10-fold (Fig. 3 A ) . In addition, both Gi/, and @yincreased the stoichiometry of &AR phosphorylation from -4 mol of phosphate/ mol of receptor to-11 mol/mol (Fig. 3 A ) .To further compare the ability of heterotrimeric G proteins and @y dimers t o activate @-ARK-mediated receptor phosphorylation we also used rhodopsin as a substrate. At low concentrations of pr subunitsor Gila, equivalentactivation of rhodopsinphosphorylation was observed (Fig. 3 B ) . However, significant differences arose as higher concentrations of Gi/, were used. While increasing concentrations of @y subunits resulted in

15415

greater stimulation of rhodopsin phosphorylation, increasing concentrations of Gila resultedininhibition of rhodopsin phosphorylation. This inhibitionwould be predicted if the G protein competes with @-ARK for binding to the receptor. Since dissociation ofG, from the activated receptor is proposed to occur upon G T P binding (Hepler and Gilman, 1992), the inhibitionof @-ARK-mediated rhodopsin phosphorylation by Gi/, should be relieved when GTPyS is present in the reaction. This is exactly what is observed in the presence of GTPyS ( i e . the inhibition athigh Gi,o was partially relieved). It appears that although heterotrimeric G proteins and pr dimers can associateequally well with p-ARK, full activation of P-ARK-mediated receptor phosphorylation is only achieved by the dissociated py dimer. The kinetic mechanism by which py activates p-ARK was studied by measuringreceptorphosphorylation at varying concentrations of &-AR in the absence or presence of @y subunits. Double-reciprocal plots of this data demonstrate that @y enhances the initial rate of P2-AR phosphorylation primarily by decreasing theK,,, for the substrate (from 194 f 19 to 49 f 2 nM) with little change in the V,,, (Fig. 4). To assess whether the interaction between by subunits and @ARK directly activates the kinase, we assessed the effect of p r on P-ARK phosphorylation of non-receptorsubstrates such as casein and the synthetic peptide RRREEEEESAAA. In these studies Pr subunits promoted a 2-4-fold increase in the rate of P-ARK-mediatedphosphorylation of the nonreceptorsubstrates(datanotshown).Thisactivation was found tobe lipid dependent, aswould be predictedconsidering the very hydrophobic nature of brain Py subunits. However, the modest and variable direct activation of P-ARK by By subunits suggests that additional interactions involving the activated receptor are necessary for py activation of @-ARKmediated receptor phosphorylation. We next directly probed the interaction of 35S-labeled pARK with thePZ-AR. In the absenceof any ligand a low level of specific binding of P-ARK to the &AR was detected (Fig. 5). This binding was increased -1.7-fold when the agonist

70

-

Pv

6050 -

0

2

4

G p r o t e l n PARK

6

40

-

30

-

8

(molar ratlo)

FIG. 3. A , (3-ARK-mediated&-AR phosphorylation. Phosphorylation reactions contained, in a total reaction volume of 100 pl, 20 nM reconstituted &AR, 50 nM (3-ARK, 20 mM Tris-HC1, pH 7.5, 2 mM EDTA, 5 mM MgCl,, 0.1 mM [y3’P]ATP (2 cpm/fmol), 50 p M (-)isoproterenol, and no G protein (0), 50 nM Gi/, (O),or 50 nM (37(V). At the times indicated, 15-pl aliquots were removed, mixed with 10 p1 of SDS sample buffer and then electrophoresed on 10% SDSpolyacrylamide gels. Following autoradiography the receptor bands were excised and counted. Data points are the mean & range from two separate experiments. B , (3-ARK-mediated rhodopsin phosphorylation. Phosphorylation reactions contained, in atotal reaction volume of 200 pl, 40 nM rhodopsin, 4 nM p-ARK, 0.1 mM [y3’P]ATP (0.1 cpm/fmol), 5 mM MgCl,, 20 mM Tris-HC1, pH 7.5, 2 mM EDTA, and 0-32 nM brain (37 (V) or Gi/, in the absence (0)or presence (0) of 20 p M GTPyS. The reactions were incubated for 3 min and quenched by the addition of 100 pl of 30% trichloroacetic acid as described under “Experimental Procedures.” Datapointsare the mean + range of two separate experiments done in duplicate.

20

1 / p,AR

(nM)

FIG. 4. Kinetics of &-AR phosphorylation by &ARK in the presence or absence of subunits. Phosphorylation reactions contained, in a total reaction volume of 15 ,ul, 0-200 nM p2-AR, 20 nM @-ARK,0 (0)or 100 (0) nM brain (37, 0.1 mM [y3’P]ATP (2 cpm/ fmol), 20 mM Tris-HC1, pH 7.5, 5 mM MgCl,, 5 mM NaP04, and50 p~ (-)-isoproterenol. The reactions were incubated for 10 min and quenched by the addition of SDS sample buffer as described under “Experimental Procedures.” Following autoradiography the receptor bands were excised and counted. Data points arethe means from two separate experiments. Kinetic analysis was performed using the BASIC translation of the SEQUEN program (Cleland, 1979).

15416 0.5

0.4

Activation G Protein

,

1

I

" "

" "

Ia0 " "

I a0 AIlP-PNP

FIG.5. Direct binding of "S-labeled &ARK to &AR. Direct binding was assessed in a total reaction volume of 100 pl, containing 14 nM &AR, 100 nM @-ARK (2000 cpm/pmol), 0 or 100 p~ (-)isoproterenol,0 or 100 p M App(NH)p,5 mM MgClz, 20 mM Tris-HC1, pH 7.5, and 2 mM EDTA. Specific @-ARK binding was determined as described in the legend to Fig. 1. Data points arethe mean f range of two separate experimentsdone in duplicate.

-

I

Q

/

Pr

U

P2m

3

0

m

0.3

0

20

40

& I

80

100

PARK (nM)

of P-ARK

phorylation by P-ARK underconditions where P-ARK is normally inhibited. Previous studies have demonstrated that bovine @-ARKis potently inhibited by ionic strength with0.1 M NaCl giving 90% inhibition of &AR phosphorylation (Benovic et d., 1987a). Purified P-ARK is also strongly inhibited by 0.1 M NaCl (Fig. 7). However, both brain P-y subunits and Gi/, are able to prevent the NaCl inhibition of @-ARK, presumably by contiguous association with P-ARK during the catalytic transfer of the phosphate moiety to the activated receptor. The ability to prevent P-ARK inhibition was also specific for brain PT since retinal Pr subunits only partially prevented the inhibition (Fig. 7). This finding is consistent with the decreased potency of retinal Py to activate P-ARKmediated receptor phosphorylation (Pitcheret al., 1992). T o further characterize the potential formation of a ternary complex we tested the ability of the P2-AR and PT subunits to modulate the phosphorylationof the peptide RRREEEEESAAA by P-ARK. While the individual addition of either fir subunits or P,-AR didnotsignificantlyalterthe level of peptide phosphorylation, the P-ARK-mediated peptide phosphorylation was increased -5-fold when both &AR and PT were added (Fig. 8A). In addition, this synergistic activation of peptidephosphorylation was totallydependentonthe presence of the P-agonist isoproterenol. To fully mimic the formation of a ternary complex consisting of P-ARK, G protein (or p~ subunits), and activated receptor, we used truncated rhodopsin (329G-Rho)which is devoid of its COOHterminal phosphorylation domain (Palczewski et al., 1991). 329G-Rhois not phosphorylatedby &ARK either in the presence or absenceof P-y subunits so its affinityfor P-ARK and/ or @T is not altered by its phosphorylation state (data not shown). In these studies 329G-Rho*alone stimulated peptide phosphorylation -5-fold consistent with our previous results (Chen et al., 1993). However, the combination of 329G-Rho and either PT or Gi/, synergistically enhanced P-ARK-mediated peptide phosphorylation -100-fold (Fig. 8B). These data further support the formation of aphysical complex consisting of P-ARK, G protein, and activatedreceptor. r

- NaC I

FIG.6. Direct binding of %-labeled &ARK to &-AR in the absence or presence of By subunits. Direct binding was assessed in a total reaction volume of 100 pl containing 5-100 nM @-ARK (2000cpm/pmol), 0 or 14 nM reconstituted @,-AR, 0 or 10 nM brain 8-y subunits, 100 p M (-)-isoproterenol, 100 p M App(NH)p, 5 mM MgCl,, 20 mM Tris-HC1, pH 7.5, and 2 mM EDTA. Specific @-ARK binding was determined as described in the legend to Fig. 1. Data points are the mean of two separate experiments done in duplicate. isoproterenol was included in the reaction. Interestingly, an -%fold increase in @-ARK binding to theP2-AR was observed in the presence of both isoproterenol and the ATP analog App(NH)p (Fig. 5). Under these conditions,which mimic the phosphorylation reaction, @-ARKbinding was linear up to a concentration of 100 nM suggesting that the Kd for @-ARK binding to the &AR is >lo0 nM (Fig. 6). We also tested whether fir subunits had anyeffect on @ARK binding to the P2-AR. At concentrations of @ARK near or below its K d for 6-y (520 nM), @-ARKbinding to theactivated P2-AR plus Py was -50% higher than to the additive value of P-ARK binding to P-y and P2-AR separately (Fig. 6). These results suggest potential formation of a ternary complex consisting of PARK, By, and Pz-AR. The formation of a ternary complex is also suggested by the ability of PT subunits and Gi/, to stabilize receptor phos-

FIG.7. Effect of By subunits and Gvoon NaCl inhibition of 8-ARK-mediated rhodopsin phosphorylation. Phosphorylation reactions containedin a total reaction volume of 40 pl, 12 nM 8-ARK, 2 p M rhodopsin, 0 or 100 mM NaCl, 0.1 mM [y3'P]ATP (0.1 cpm/ fmol), 5 mM MgCl,, 20 mM Tris-HC1, pH 7.5, 2 mM EDTA, and no G protein, 120 nM brain @r(@-y~), 120 nM retinal 0-y (PYT),or 120 nM Gi,,,. The reactions were incubated for 3 min and quenchedby the addition of 50 pl of 10% trichloroacetic acid as described under "Experimental Procedures." Datapoints arethe mean f range of two separate experiments done in duplicate. nd,not detectable.

G Protein Activation of P-ARK

ZOO

a

1

'"t " "

--- -

BY - ---

--- -

----

BY

Rho*

Rho*

GI, Rho*

FIG. 8. A, synergistic activation of P-ARK-mediatedpeptide phosphorylation by P-y and &AR. Phosphorylation reactions contained, in a total reaction volume of 50 p l , 100p~ RRREEEEESAAA peptide, 20 nM 8-ARK, 0 or 10 nM &AR, 0 or 40 nM P-y subunits, 0 or 25 p M (-)-isoproterenol ( h ) , 0.1 mM [-y3'P]ATP (0.5 cpm/fmol), 5 mM MgCL, 20 mM Tris-HC1, pH 7.5, and 2 mM EDTA. The reactions were incubated for 20 min and quenched by the addition of 30 p1 of 30% trichloroacetic acid as described under "Experimental Procedures.'' Data points are the mean range from two separate experiments done in duplicate. B, synergistic activation of P-ARK-mediated peptide phosphorylation by G proteins and 3z9G-Rho'.Phosphorylation reactions contained, in a total reaction volume of 50 p1, 100 pM RRREEEEESAAA peptide, 100 nM p-ARK, 0 or 2.5 p~ 329G-Rho' (Rho'), 0 or 200 nM By, 0 or 200 nM Gila, 0.1 mM [y3*P]ATP (0.5 cpm/fmol), 5 mM MgC12, 20 mM Tris-HC1, pH 7.5, and 2 mM EDTA. The reactions were incubated for 3 min and quenched by the addition of 30 p1 of 30% trichloroacetic acid as described under "Experimental Procedures." Data points are the mean f range from two separate experiments done in duplicate.

*

DISCUSSION

The role of G protein Pr subunits in modulating effectors independent of the G, subunit has grown significantly in recent years. Regulation of adenylylcyclase (Tang and Gilman, 1991), phospholipase C (Blank et al., 1992; Camps et d., 1992; Katz et al., 1992), and various ion channels (Okabe et al., 1990) by Gprotein dimers have all recently been reported. The demonstration that 8-y subunits specifically

15417

activate 0-ARK-mediated receptor phosphorylation adds to the growing appreciation for the complex nature of heterotrimerit G proteins. Previous studies demonstrated an -1O-fold activation of p-ARK-mediated receptor phosphorylation by brain subunits (Haga and Haga, 1992; Pitcher et al., 1992). However, attempts to dissect the mechanisms of activation were limited by the inability to assess direct protein interaction between native P-ARK and G protein Pr subunits. In the present study we provide the first demonstration of direct protein interaction between isolated G protein Pr subunits and a fully functional effector such as p-ARK. The specific binding of @-ARKto purified brain Pr subunits or Gila was observed only in the presence of phospholipid vesicles. There appeared to be no direct interaction of P-ARK andsubunits in detergent or detergent-free buffersas assessed by gelfiltration.' This lipid dependence suggests that specific @ARK binding is due to interaction at sites on the Pr subunit that are accessible only when localized on membrane surfaces. P-ARK rapidly bound to both heterotrimeric G proteins and dimers with similar K d values of -58 and -32 nM, respectively. This rapid association between P-ARK andbrainsubunits or Gi/, is consistent with the rapid activation of &ARK mediated receptor phosphorylation. The similar Kd values of &ARK binding, as well as the inability of GTPrS to alter P-ARK binding to heterotrimeric G proteins, suggests that &ARK and G protein a subunit binding to Pr are not mutually exclusive. The ability of heterotrimeric G proteinsto directly associate with p-ARK was also evident by the activation of receptor phosphorylation. Equivalent activation was observed only at sublow concentrations of heterotrimeric Gproteinsand units. Similar observations by Haga and Haga (1990, 1992) were made using a crude porcine P-ARK-related kinase preparation. They showed that low concentrations of heterotrimeric G proteinsactivated m2 muscarinic cholinergic receptor and rhodopsin phosphorylation. They also concluded that higher concentrations of heterotrimeric G proteins inhibited receptor phosphorylation due to competitive binding ofGi/, and kinase to the activated receptor. These studies, in addition to the direct binding data, demonstrate that P-ARK can associate equally well to the heterotrimeric G protein or to the /3r dimer alone. This equivalent affinity is important because it suggests that binding to P-y does not require dissociation of the CY subunit andthat theheterotrimeric G protein may be the molecular entity which promotes @-ARKrecognition of activated receptors. Presumably, the inherent ability of the G, subunit to couple and then dissociate from the activated receptor would temporarily compete for &ARK interaction with the activated receptor. This would explain the inhibition of receptor phosphorylation observed at higher concentrations of Gi/,,. In addition, the high affinity interaction of &ARK with heterotrimeric G proteins may also be involved in the cellular localization of @ARK in vivo. If (3ARK is associated with heterotrimeric G proteins prior to agonist stimulation, then the coupling of G proteins to agonist-occupied receptors will serve to target P-ARK to thesame activated receptors. In this proposed scheme, &ARK recognition of activated receptors and activation of P-ARK by Py subunits would satisfy the temporal and spatial sequence of events proposed to occur after agonist activation of the receptor and G protein coupling (Hepler and Gilman, 1992). Kinetic analysis of &AR phosphorylation by P-ARK suggests that 07 subunits primarily enhance receptor phosphorylfor the P,-AR from -194 to -49 ation by decreasing the K,,, nM, with little alteration of the V,,,. These results are in C. Kim, personal observation.

15418

G Protein Activation of @-ARK

contrast to previous studies reporting that Pr subunits primarily increase the Vmaxof receptor phosphorylation with little effect on the K , (Haga and Haga, 1992). We have found that brain subunitscan preventinhibition of @-ARKmediated receptor phosphorylation, presumably by stabilizing the kinase/receptor interaction. This ability of Pr to prevent kinase inhibitioncould be interpreted as aneffect on the V,,, of the phosphorylation reaction. This may explain some of the differences between our results and those of Haga and Haga, since their studies utilized crude kinase preparations and assay conditions that would substantially inhibit P-ARK. In addition to the decrease in the K,,, for the P2-AR, our studies reveal an increase in stoichiometry from -4 mol of phosphate/mol of P2-AR to -11 mol/mol in the presence of Py subunits. The potentialfunctional significance of this increased stoichiometry is currently being addressed. Pr Subunits were also able to directly activate P-ARK-mediated phosphorylation of nonreceptor substrates. This direct activation was phospholipid dependent and varied from 2-4-fold depending on the substrate. This direct activation in itself, however, does not fully explain the -10-fold stimulation of receptor phosphorylation. Additional interactions involving the activated receptor appear to be required for Py stimulation of P-ARK activity. Binding studieswith the P2-AR detected an -1.7-fold stimulation of &ARK binding in response to isoproterenol treatment. Interestingly, a %fold increase over basal P-ARK binding was observed in response to both isoproterenol and App(NH)p. Thisagonist and nucleotide dependence for high affinity binding to the receptor suggests a multisite interaction between the receptor and kinase that is dynamically controlled by conformational changes induced in both the PzAR and P-ARK. Previous studies have shown that P-ARKmediated receptor phosphorylation follows a two-substrate sequentialkinetic mechanism (Kim et al., 1993). Since App(NH)p increases p-ARK binding to the P2-AR this suggests that ATP binding to P-ARK likely precedes binding to the activated receptor. This is also supported by the ability of @-ARK to undergo autophosphorylationin the absence of receptor. The -%fold increase in P-ARK binding observed in response to isoproterenol and App(NH)p as well as theability of subunits toactivate P-ARK may account for the previous reports of increased membrane-associated @-ARKactivity in response to various agonists (Strasser et al., 1986; Mayor et al., 1987; Chuang et al., 1992). The independent interactions between @-ARKand either fir subunits or the activated receptor in a distinct temporal sequence does not appear to fully explain the activation of PARK-mediated receptor phosphorylation. In fact, several lines of evidence suggest the formation of a ternary complex composed of G protein, receptor, and P-ARK as an initiating event. First, both brain Pr subunits and heterotrimeric Gi/, repress the ability of ionic strength to inhibit p-ARK-mediated receptor phosphorylation. This ability of G proteins to inhibit the effects of ionic strength on @-ARKmay well have a physiological role given the ionic environment inside the cell. In addition, direct binding studies also suggest the formation of a ternary complex. At concentrations of P-ARK near or below its Kd for Pr, P-ARK binding to the activated &AR plus was -50% higher than to the additive value of @-ARKbinding to Pr and P2-AR separately. Finally, ternary complex formation is also suggested by the ability of Gi/, (or Py subunits) and receptor to synergistically activate @-ARK phosphorylation of a synthetic peptide. Moreover, this synergistic activation is agonist-dependent, demonstrating that

agonist binding is acritical step in the formation of this ternary complex. Previous reports have demonstrated that rhodopsin kinase and @-ARKcan bind to discrete domains of the activated receptor as assessed by activation of peptide phosphorylation (Palczewski et al., 1991; Chen et al., 1993). This activation is enhanced when the carboxyl-terminal tail of rhodopsin, which contains the phosphorylation sites, is proteolytically removed. The synergistic levels of peptide phosphorylation also indicate that &ARK and G proteins may bind to discrete domains of the receptor which differ from the sites of phosphorylation. Our data suggest that the formation of this transient G protein. receptor. @-ARKcomplex may play an important role in promoting p-ARK-mediated receptor phosphorylation in vivo. The present study was undertaken in an attempt to unambiguously dissect the various protein interactions involved in @r subunit stimulation of @-ARKactivity. This was greatly aided by the use of fully functional radiolabeled P-ARK. Quantitative binding and kinetic data demonstrates a high affinityinteraction between P-ARK and heterotrimeric G proteins via the Pr subunit. This interaction may well play an important role in the cellular localization of P-ARK as well as serving as a mechanism to specifically target @-ARK to activated receptors. REFERENCES Benovic, J. L., Shorr, R. G. L., Caron,M. G., and Lefkowitz, R. J. (1984) Biochemistry 23,4510-4518 Benovic, J. L., Strasser, R. H., Caron, M. G., and Lefkowitz, R.J. (1986) Proc. Natl. Acad. Sci. U, S. A. 8 3 , 2797-2801 Benovic. J. L.. Mayor. F.. Jr., Staniszewski, C., Lefkowitz, R. J., and Caron, M. G. (1987a) 3. Bibl. Ck'm. 262,9026-9032 Benovic, J. L., Regan, J. W., Matsui, H., Mayor, F., Jr., Cotecchia, S., LeebLundberg, L.M. F., Caron, M. G., and Lefkowitz, R. J. (1987b) J. Biol. Chem. 2 6 2 , 17251-17253 Benovic, J. L., DeBlasi, A,, Stone, W. C., Caron, M. G., and Lefkowitz, R. J. (1989a) Science 2 4 6 , 235-240 Benovic. J. L.. Bouvier. M.. Caron. M. G.. and Lefkowitz. R. J. (1989h) Annu. Rev. Cell Bid. 4 , 405-428 Blank. J. L.. Brattain. K. A,. and Exton. J. H. (1992) J. Biol. Chem. 267. 23069-23075 Bradford, M. M. (1976) Anal. Biochem. 72,248-254 Camus. M.. Carozzi. A,. Schnabel, P., Scheer, A,, Parker, P. J., and Gierschik, p.'(i992j Nature 360,684-686 Cerione, R. A,, Strulovici, B., Benovic, J. L., Lefkowitz, R. J., and Caron, M. G. (1983) Nature 306,562-566 Chen, C.-Y., Dion, S. B., Kim, C. M., and Benovic, J. L. (1993) J. Bid. Chem. 268,7825-7831 Chuang, T. T., Sallese, M., Ambrosini, G., Parruti, G., and De Blasi, A. (1992) J. Biol. Chem. 267.6886-6892 Cleland, W. W. (1979) Methods Enzymol. 6 3 , 103-138 Conk. Neville. M. E.. Vrana. K. E.. Hart]. F. T.. and Roskoski, R., Jr. ".", P. F.. . . ~~~. (1982) Biochemistj 21,5794-5799 ' Haga, T., and Haga, K. (1989) FEES Lett. 268,43-47 Haea. T..and Haea. K. (1992) J. Biol. Chem. 2 6 7 , 2222-2227 He&, J. R., and-Gilman, A.'G. (1992) Trend5: Biochem. Sci. 17,383-387 Hausdorff, W.P., Caron,M. G., and Lefkowitz, R.J. (1990) FASEB J. 4,2881~~

~~~

I

~

~

~~

I

~~

~

~~

~

911114

KL&za, T., Bokoch, G. M., Smigel, M. D., Ui, M., and Gilman, A. G. (1984) J . Biol. Chem. 2 5 9 , 3586-3595 Katz, A,, Wu, D., and Simon, M. I. (1992) Nature 3 6 0 , 686-689 Kim, C. M., Dion, S. B., Onorato, J. J., and Benovic, J. L. (1993) Receptor, in press Kwatra, M. M., Benovic, J. L., Caron, M. G., Lefkowitz, R. J., and Hosey, M. M. (1989) Biochemlstry 28,4543-4547 Laemmli, U. K. (1970) Nature 227,680-685 Lohse. M. J.. Benovic. J. L.. Codina,. J... Caron, M. G., and Lefkowitz, R. J. (1990) Scilice 248,'1547-i550 Mayor, F.,Jr., Benovic, J. L., Caron, M. G., and Lefkowitz,R. J. (1987) J. Biol. Chem. 262,6468-6471 Okabe, K., Yatani, A,, Evans, T., Ho, Y-K., Codina, J., Birnbaumer, L., and Brown, A. M. (1990) J. Biol. Chem. 265,12854-12858 Palczewski, K., Buczylko, J., Kaplan, M. W., Polans, A. S., and Crabb, J. W. (1991) J. Biol. Chem. 2 6 6 , 12949-12955 Pitcher, J. A,, Inglese, J., Higgins, J. B., Arriza, J. L., Casey, P. J., Kim, C., Benovic. J. L.. Kwatra. M. M.. Caron, M. G., and Lefkowitz, R. J. (1992) Science 5 5 7 , i264-1267 Shichi, H., and Somers, R. L. (1978) J. Biol. Chem. 253, 7040-7046 Sternweis, P. C. (1986) J. Biol. C h m . 261,631-637 Sternweis, P., and Robishaw, J. D. (1984) J. Biol. Chem. 259,13806-13813 Strasser, R. H., Benovic, J. L., Caron, M. G., and Lefkowitz, R. J. (1986) Proc. Natl. Acad Sei. U. S. A. 83,6362-6366 Tang, W-J., and G h a n , A. G. (1991) Science 2p4,1500-1503 Wald, G., and Brown, P. K. (1953) J. Gen. Physcol. 37, 189-200 Wilden, U., and Kuhn, H. (1982) Biochemistry 21,3014-3022