florio 1985

Embed Size (px)

Citation preview

  • 8/2/2019 florio 1985

    1/7

    THE OURNALF BIOLOGICAL CHEMISTRYQ 1985by The American Soeietyof Biological Chemists,Inc.Vol. 260. No. , ssue of March 25, pp. 3477-3483, 1985Printed in U.S.A.

    Reconstitution of Resolved Muscarinic Cholinergic Receptors ithPurified GTP-binding Proteins*(Received for publication, September 6, 1984)

    Vincent A. FlorioS and PaulC. SternweisFrom the Department of Pharmacology, University of Texas Health Science Center at Dallas, Dallas, Texas 75235

    The association of agonists with muscarinic recep-tors in membranes from bovine brainas affected onlyslightly by guanine nucleotides. However, solubiliza-tion of these membranes with deoxycholate and sub-sequent removal of detergent resulted in a preparationof receptors with increased affinity for agonists and alarge increase in responseo guanine nucleotides.Chromatography of deoxycholate extracts of mem-branes on DEAE-Sephacel resulted in the separationof receptors from 96%of the guanine nucleotide-bind-ing activity. Guanine nucleotides hadno effect on thebinding of agonists to these resolved receptors.The effect of guanine nucleotideswas restored afterthe addition of either of wo purified guanine nucleo-tide-binding proteins rom bovine brain. One of theseproteins, presumably brain I, is composed of subunitswith the same molecular eights (a, 1,000; B, 35,000;y, 11,000) and unctions as the nhibitoryguaninenucleotide-bindingprotein solated from liver. Theother protein, termedo, is a novel guanine nucleotide-binding protein that possesses a similar subunit com-position (a, 9,000; 8, 36,000; y, 11,000) but whosefunction is not yet known. Addition of either proteinto the resolved receptor preparation increased agonistaffinity by at least 10-20-fold, and low concentrationsof guanine nucleotidespecifically reversed this effect.Reconstitution of receptors with the resolved sub-units of Go demonstrates that the l subunit alone hadno effect on agonist binding,but that this subunit doesappear to enhance the ffects observed with the a sub-unit alone.

    The effect of guanine nucleotides in decreasing the affinitywith which agonists bind to receptors was first demonstratedfor hepatic glucagon receptors (1).A large body of subsequentwork demonstrated similar effects with a wide variety ofreceptors, particularly those linked to adenylate cyclase (e.g.Refs. 2 and 3).Several lines of evidence have suggested that this modula-tion of agonist affinity is mediated by regulatory proteins (G-proteins) that bind guanine nucleotides. Gs is the regulatory* This work was supported by Searle ScholarsAward 83-5-102 andMarch ofDimes Basil OConnor Grant -435.The costs of publicationof this article were defrayed in par tby the payment of page charges.This article must therefore be hereby marked advertisement inaccordance with 18U.S.C. Section 1734 solely o indicate this fact.$ Supported by National Institutes of Health Postdoctoral Train-ing Grant HD07190.The abbreviations used are: G-protein, any GTP-binding roteinwithin the family of homologous proteins composed minimally of G s,GI,Go, and transducin; Ga and GI, the identified regulatory compo-nents of adenylate cyclase that mediate stimulation and inhibition,respectively; Go, new GTP-binding protein from membranes ofbovine brain; GTPyS, guanosine 5-(3-O-thio)triphosphate;NB , 3-quinuclidinyl benilate; GppNHp, guanyl-5-yl imidodiphosphate;Hepes, 4-(2-hydroxyethyl)-l-piperazineethanesulfoniccid.

    protein tha t mediates the stimulation of adenylate cyclase byhormones (4). Membranes from the cyc- variant of S49 lym-phoma cells, which lack functional Gs, do not display 8-adrenergic stimulation of adenylate cyclase (5) ,and the affin-ity of @adrenergic receptors in these membranes for agonistsis not alteredby guanine nucleotides. The addition of a crudepreparation of G s restored the sensitivity of adenylate cyclaseto stimulatory hormones and caused an increase in the ffinityof 8-adrenergic receptors for agonists; this higher affinity wasreversed by the addition of guanine nucleotides (6).Hsia et al. (7 ) have recently provided evidence that theeffects of guanine nucleotides on enkephalin receptors thatmediate inhibition of adenylate cyclase are due to th 6nhibi-tory G-protein, GI (8). Treatment of NG108-15 cells withislet-activating protein, the toxin from Bordetella pertussisthat catalyzes ADP-ribosylation of GI and attenuates recep-tor-mediated inhibition of adenylate cyclase (9), caused adecrease in the affinity of enkephalin receptors for agonistsand abolished the effects of guanine nucleotides.Muscarinic receptors mediate inhibition of adenylate cy-clase in heart (lo ), brain (111, and other tissues (12). Theeffects of guanine nucleotides on the binding of muscarinicagonists in these tissues are well documented (Ref. 13 andreferences therein). We have chosen bovine brain as asourceof muscarinic receptors and G-proteins for use in elucidation

    of the mechanism of this modulation of agonist binding affin-ity by guanine nucleotides.We have recently reported the purification of two major G-proteins from membranes of bovine brain andhave presentedevidence pointing to the similarity of these proteins to thewell-characterized G-proteins, Gs, GI, and ransducin (14).One of these proteins was tentatively identified as GI andpossesses a , 8, and y subunits with molecular weights of41,000, 36,000and 11,000, respectively; this structure is iden-tical to that f GI purified from rabbit liver. The other protein,labeled Go, has identical p and y subunits but an a subunitof only 39,000 Da. While the functional and structural rela-tionship of these two G-proteins to each other and to otherG-proteins is not et clear, they exist in large quantities (about1-2% of membrane protein from brain) and can be easilypurified in the detergent sodium cholate. Furthermore, theGoa subunit (39,000 Da) can be obtained in an unligandedandstable form. Therefore, these proteinsare likely andconvenient candidates to use for study of the interaction ofG-proteins with the muscarinic receptor.We report here the solubilization of the muscarinic receptorfrom membranes of bovinebrain, its separationrom the bulkof GTP-binding activity in these membranes, and the subse-quent reconstitution of this receptor with purified GI and Gofrom the same source.MATERIALS ANDMETHOD S

    The following reagents were purchased from Sigma: acetylcholine,ATP, atropine sulfate, bovine serum albumin, carbamylcholine, cho-3477

  • 8/2/2019 florio 1985

    2/7

    3478 Muscarinic Receptors: Interaction withTP-binding Proteinslesterol, dithiothreitol,-tubocurarine, egg phosphatidylcholine,GDP, GMP, GTP, Hepes, hexamethonium, ITP, methacholine, ni-cotine, oxotremorine, and physostigmine. Cholic acid was obtainedfrom Sigma and was purified on DEAE-cellulose as described by Rossand Schatz (15). Lubrol 12A9 was a gift from IC1 and was deionizedprior to use. Deoxycholic acid (Aldrich, gold seal 99+%) was usedwithout further purification. It should be noted that extraction effi-ciency, receptor stability, and the behavior of receptors on DEAE-Sephacel were all changed when deoxycholate from Sigma was used.We were unable to achieve satisfactory resolution of receptors fromG-proteins using the Sigma product. DEAE-Sephacel and SephadexG-50 were obtained from Pharmacia. [3H]QNBand [%]GTPyS werepurchased from New England Nuclear. GF/F filters and BA85 filterswere obtained from Whatman and Schleicher & Schuell, respectively.

    Prepar ation of Bovine Brain Mem branescrude membranes from bovine brain were prepared as described(14). Briefly, cerebra were dissected to remove brain stem and largeportions of white matter. The remaining tissue was homogenized ina blender with 4 volumes'of 10 mM Tris-C1, pH 7.5, 10% sucrose.After filtration through four layers of cheesecloth, membranes werecollected by centrifugation at 20,000 X g for 30 min. Pellets wereresuspended in the same buffer with a Potter-Elvehjemhomogenizerand collected by centrifugation at 20,000 X g for 60 min. This stepwas repeated and then pellets were resuspended to about 20 mg/mlin the same buffer and stored at -80 "C. This preparation was thestart ing material for purification of G-proteins and resolution of themuscarinic receptor.

    Purifica tion of G-proteins and Subunits from Bovine BrainThe major G-proteins, GI and GO,were purified as described (14).In summary, Go and GI were copurified through DEAE-Sephacel andUltrogel AcA and partially resolved during chromatography throughheptylamine-Sepharose, the final step for purification. The two pro-teins couldberesolved further by chromatography of Go- or GI-enriched pools through the same heptylamine-Sepharose columnunder identical conditions. In this fashion, fractions of pure Goa(39,000 Da) and fractions enriched in GI (a= 41,000 Da, p, and y)could be obtained. While no G~ol 41,000) could be detected in thepurified Goa, purified preparations highly enriched in GI still con-tained about 5-10% Goa (14). Th e purified proteins were pooled,concentrated by pressure filtration on an Amicon PM-30 membrane,

    and stored a t 0 or -80 'C. GOand GI werequantitated by their abilityto bind GTPyS; thus, 1 mol of binding sites indicated 1 mol of the asubunit or1mol of the multimeric protein.Purified p subunit was obtained from brain (14) by proceduresoriginally described by Northup and colleagues (16). The preparationcontained two forms of /3 subunits ( M r = 36,000 and M , = 35,000)and a potential y subunit ( M r= 11,000). The molar concentration ofpurified @ was calculated with an assumed M, f 35,000; the presenceof a y subunit suggests this overestimates the actual concentration.Solubilization of Receptors

    mixed with 95 ml of Tris-C1, pH 8, 1.5 mM NaEDTA, and 0.25 mMBrain membranes (70 ml, 1.4 g of protein) were rapidly thawed,atropine, and incubated at 30 "C for 15 min. The membranes werecollected by centrifugation a t 90,000 X g for 30 min at 4 "C andresuspended with a Dounce homogenizer in 47 mM Tris-C1, pH 8 , l . lmM NaEDTA, and 0.23 mM atropine to a volume of 88 ml. Sodiumdeoxycholate (20.3 ml of 4% (w/v)), pH 8, was slowlyadded to a finalconcentration of 0.75%, and themixture was incubated a t 0 "C for 30min with occasional shaking. Particulate material was then removedby centrifugation at 160,000 X g for 2 h. Atropine was included inthese and subsequent steps to stabilize receptors from marked lossesof binding activity during exposure to deoxycholate. Oxotremorinecan also prevent these losses, and one of these ligands was alwaysadded to preparations of receptors during treatments with detergents.

    Separation of Solubilized Muscarinic Receptors from G-protein sExtracts were supplemented with 20% glycerol, 80 mM NaCl, and0.2% sodium cholate (final concentrations).Theseadditions were

    dure. The mixture was applied to a column (3.2 X 5 cm) of DEAE-found to improve recovery of receptors through the following proce-Sephacel tha t had been equilibrated with 37.5 mM Tris-C1, pH 8, 1mM NaEDTA, 0.2 mM atropine, 80 mM NaCI, 0.65% sodium deoxy-cholate, 0.2% sodium cholate, and 20% glycerol. A short, wide column

    allowed a rapid flow rate without loss of resolution.) The gel was henwashed with 100 ml of the same solution; receptors were subsequentlyeluted by raising the NaCl concentration to 0.4 M. Fractions of 6 or3 ml (during the wash and elution, respectively) were collected intosiliclad tubes. Fractions near the receptor peak were collected into0.5 ml of 30 mg/ml eggphosphatidylcholine and 3 mg/ml cholesterolthat had been sonicated in the wash solution; this further improvedrecovery of receptors. The peak of QNB-binding activity (about 18ml)waspooled,mixed with 18 g of XAD-2 beads (MallinckrodtChemical Works), and incubated a t 0 "C for 1 h with shaking. Thisprocedure resulted in the removal of 95% of deoxycholate with aconcomitant loss of 25% of the QNB binding. The treated pool wasremoved from the beads and subjected to pressure filtration on anAmicon PM-30 membrane. After a 2-3-fold concentration, aliquotswere frozen at -80 'C. QNB-binding activity of this preparation wasunchanged after repeated freezing and thawing. For smaller prepa-rations, deoxycholate could be removed by gel filtration as describedin the legend to Fig. 3.

    Reconstitution of Receptors with G-proteinsThe procedures described below utilized resolved receptors fromwhich detergent had been removed by either gel filtration or XAD-2bead treatment. Receptors that had been prepared by the atterprocedure were gel-filtered into 20 mM NaHepes, pH 8 , lmM EDTA,160 mM NaCl before reconstitution with G-proteins; this step re-

    moved atropine and residual deoxycholate. Three methods were de-veloped for reconstitution of the resolved receptor. The resultantpreparations could be assayed directly for binding of QNB.Method 1: Gel Filtration (Used in Fig. 3)"Resolved receptors (1-2pmol and 2-8 mg of protein in 0.5 ml) were mixed with 0.5-1.0 nmolof G-protein (either 0.5 ml of crude G-protein from early fractions ofthe DEAE column or 0.02-0.15 ml of purified G-protein in 20 mMTris-CI, pH 8, 1 mM NaEDTA, 1 mM dithiothreitol, 1% sodiumcholate, 100 mM NaCI; this solution was also added t o control tubes).Deoxycholate (0.5% final) and oxotremorine(1mM final) were added,and themixture was allowed o sta nd n ice for 1h. The mixture wasthen diluted with 4 ml of 20 mM NaHepes, pH 8, 1mM NaEDTA, 1mM dithiothreitol, 160 mM NaCl and applied to a 50-ml G-50 columnwhich was developed in the same solution. Leading turbid fractionswere pooled (turbidity corresponded well to QNB-binding capacity).Unfortunately, the recovery of receptor binding activity through thisprocedure was variable. Therefore, alternative procedures for recon-stitution were also developed.Method 2: Sedimen tation through Sucrose Gradients (Used in Fig.4)"Receptors and G-proteins were mixed at 0 "C in a volume of 0.3-1.2 ml. Atropine (10 p~ final) and either deoxycholate (0.7% final)or cholate (0.4% final) were added. After 1h a t 0 "C, the mixture wasapplied to the op of a step gradient of sucrose (30 ml of 5% (w/v), 5ml of 40% (w/v)) in 20 mM NaHepes, pH 8 , l mM NaEDTA, 160 mMNaCI. The gradients were centrifuged for 3 hat 27,000 rpm in an SW27 rotor. Receptors were harvested from the 5-40% interface andwere resuspended with a syringe and needle and diluted in the solutionabove without sucrose. Recovery of receptors from thi s procedure was

    Method 3: Dilution (Used in Table I I and Figs.5 and 6)"G-proteinswere mixed with resolved receptors at 0 "C in a total volume of 0.5-1.0 ml; MgCI, was added to a concentration of 20 mM. Concentrationsof cholate in the initial mixtures did not exceed 0.3%. After 1 h at0 "C, the mixture was diluted slowly (over 5 min) with 5-10 ml (10-fold dilution) of 20 mM Hepes, pH 8, 1mM NaEDTA, 160 mM NaC1.The omission of deoxycholate and muscarinic ligands from thisprocedure eliminates the need for a separation step and does notappear to reduce the efficiency of reconstitution. Recovery of recep-tors w a s approximately 100%.

    50-80%.

    Incorporation of Small Aliquots of Receptor into Phospholipid forAssaySince binding of QNB could not be detected directly in samplescontaining deoxycholate (i.e. extracts and DEAE fractions), thismethod was devised to prepare small aliquots for assay by removal ofdetergent and concomitant incorporation of protein into phospho-lipid. Aliquots (50 or 100 p l ) of samples were mixed with 500 p1of alipid mixture containing egg phosphatidylcholine (3 mg/ml), choles-terol (0.3 mg/ml), and sodium cholate (3 mg/ml) that had beensonicated under nitrogen in Solution A (25 mM NaHepes, pH 8, 2mM MgC12,1mM NaEDTA, and 100 mM NaC1). The mixtures were

  • 8/2/2019 florio 1985

    3/7

    Mus carin ic Receptors :nteraction with GTP -b ind ingrote ins 3479applied to 5-ml columns of Sephadex G-50 (medium) tha t had beenequilibrated with Solution A. The columns were washed with 1.1 mlof Solution A, and the receptors were eluted with another 1.0 ml ofthe same solution. Eluted receptors could be assayed for QNB bindingby vacuum filtration a s described below. Recovery of receptors afterextraction and detergent removal was 70-80% relative to originalmembranes, suggesting tha t thi sprocedure does not result in signifi-cant loss of receptors. This treatment also provided for efficientremoval of unbound ligands present in receptor preparations. Anycarryover of ligand bound to receptors would not be sufficient tocompete significantly with QNB during the binding assay.

    Assay of Q N B BindingBinding of QNB to membranes and reconstituted mixtures wasmeasured by vacuum filtration, essentially as described by Yamamuraand Snyder (17). Samples were incubated at 30 'C in 1 ml of 25 mMpotassium phosphate, pH 7.5, 08mM NaEDTA, 10mMMgC12, 230mM NaCl, 0.6 mg/ml bovine serum albumin, 4 mM NaHepes, pH8, and the indicated concentrations of nucleotides and cholinergicligands. ['HIQNB was present at a concentration of 0.19 nM forcompetition experiments and 3 nM for measurement of total receptorbinding sites. After 1 h, samples were diluted with 3 ml of ice-cold25mM potassium phosphate, pH 7.5, 5 mM MgC12,1 mM EDTA, andfiltered through Whatman GF/F filters, followed by three washeswith 5 ml of the same buffer. The filters were dried and counted intoluene containing 0.5% 2,5-diphenyloxazole with an efficiency of55%.The nonspecific binding for 13H]QNBwas less than 5% of totalbinding for all preparations of receptors utilized.

    Assay of GTPyS BindingBinding of [%]GTPyS was assayed as described by Northup etal.(18). amples to be assayed were diluted in 5 or more volumes of 20mM NaHepes, pH 8, 1 mM NaEDTA, 1 mM dithiothreitol, and 0.1%Lubrol 12A9. The diluted samples (50 l) were added to 50pl of 50mM NaHepes, pH 8, 100 mM MgCL, 1 mM EDTA, 200 mM NaCl, 4

    p~ GTP rS and "SIGTPyS (1-2 X lo6 cpm) in siliclad tubes. After1 h at 30 'C, samples were diluted and filtered as described (18).Filters were dried, dissolved, and counted in Liquiscint (NationalDiagnostics).Other Assays

    Protein concentration was determined by staining with AmidoBlack as described by Schaffner and Weissman (19) sing bovineserum albumin as a standard. Phosphate concentration was deter-mined by the method of Ames (20) fter digestion of samples withperchloric acid.RESULTS

    Effects ofGuanineNucleotide in Membranes and Reconsti-tuted Extracts-Membranes from bovine brain bound [3H]QNB with high affinity. The KO or [3H]QNB was 250 p~an d was unaffected by the presence of 0.1 mM GTP. Bindingin the presence of 50 WM atropine (nonspecific binding) wasless tha n 5% of total binding. T he muscarinic specificity ofthese sites is shown in Table I; thus, theorder of potency forTABLEEffectsof GTP on the binding of cholinergic ligands to muscarinicreceptors in membranes f rom bovine brainMembranes from bovine brain were prepared and assayed as de-scribed (see "Materials and Methods"). Assays contained 80 pgofmembrane protein with 79 fmol of binding sites for QNB, 0.19 p M['HIQNB, and 0.1 mM GT P where indicated.

    ICm-GTPGTPLigand

    OxotremorineCarbacholAtropineHexamethoniumNicotined-Tubocurarine

    P M2.80.0055.0055650 19004500 55001100100

    15050

    7.0

    competition with QNB was atropine > oxotremorine >>nicotine.Table I also shows th at G TP could decrease th e affinity ofthese binding sites for agonists, but not antagonists. T heeffect of GT P in this prep aratio n (!&"-fold increase in theIC, of agonists) was smaller tha n effects observed with mus-carinic receptors from rat brain (21) and myocardium (22).Fig. ZA demon strates thi s small effect of GT P on membranepreparation from bovine brain. T he low app arent affinity ofthese receptors for agonists and the small effects of GT Psuggest th at themajority of these receptors are not ubject toregulation by G-proteins. However, subsequent data demon-stra te that regulation of these receptors by G TP is possible.If receptors were solubilized with deoxycholate and hedetergent was then removed by gel filtration, the apparen taffinity of receptors for agonists was increased m arkedly (Fig.1B). hus, the IC, for oxotremorine in the preparation tha thas been treated with detergent was 0.03PM, hile th e K O fo r[3H]QNB was unchanged (not shown). Th e addition of 0.1mM G T P resulted in the reduc tion of the a ppar ent affinity ofoxotremorine to an C,, of 10 ~ C I M ith no effect on the K D for[3H]QNB.Without t he addition of detergent, no treatme nt ofth e membranes has been able to induce this result. Therefore,the process of solubilization and rem oval of detergent appe arsto promote the interaction of muscarinic receptors with acomponent sensitive to guanine nucleotide. Moreover, the

    Log [~xotremorine]FIG. 1. Effect ofGTP on agonist binding to membranes and

    reconstituted receptors. Brain membranes (60mg of protein) wereincubated for 20min at 30 "C in 6 ml of 20mM NaHepes, pH 7.5,5mM MgC12, 1 mM EDTA, 1 mM dithiothreitol, and 100 p M oxotre-morine. After collection by centrifugation, 30 mg of membranes werewashed twice at 4 "C in the absence of MgC12 and oxotremorine andsuspended for assay ( A ) .The other 30 mg of membrane protein weresuspended at 0 C in 2.6 ml of the solution containing oxotremorinebut without MgC12. Deoxycholate (0.4 ml of5% (w/v)) was added,and extraction wasallowed to proceed for 20 min on ice.Aftercentrifugation, 2ml of the extract were mixedwith 200pl of sonicatedegg phosphatidylcholine (15 mg/ml) and applied to a 50-ml columnof Sephadex G-50 for removal of detergent and oxotremorine. Afterelution with 20 mM NaHepes, pH 7.5, 1 mM EDTA, and 160 mMNaCl, cloudy fractions containing vesicles were pooled and assayeddirectly for binding of QNB by vacuum filtration ( B ) .The yield ofbinding sites for QNB in the pool of vesicles was 50% of the sitesmeasured in the equivalent amount of original membrane. Assayscontained 0.18 nM [3H]QNB.GTP (0 )was included at 100 p ~ .

  • 8/2/2019 florio 1985

    4/7

    3480 M u s c a r i n i c Receptors: I n t e r a c t i o n with GTP-binding Proteinsdata suggest th at a large proportion of the treated receptorswere capable of this interactio n.It is of interest that the me mb ranes used in Fig. 1 wereincubated with ox otremorine pr ior to and during extraction.If the an tagonist , a tropine,was substituted for oxotremorineduring the treatme nt, the ecovery of binding si tes for QNBwas increased slightly, but the expression of high affinity foroxotremorine was much less (ICs0 of 1 a n d 10 p~ i n heabsence and presence of GTP, respectively). The differencebetween oxotremo rine an d atropine islways observed n thi scrude experiment; however, the difference isnot observedwhen either ligand is utilized during reconstitution f resolvedreceptors.

    Resolution of Receptors from G-proteins and Reconstitutionof Guanine NucleotideEffects-Fig. 2 demon strates the sep a-ration of solubilized muscarinic receptors from G-proteins.Application of a deoxycholate extract of brain membranes toDEAE-Sephacel resulted in retention f the binding si tes forQ N Ba nd hema jorit y of th e phospholipid; most of theGTPy S-binding activity an d oughly half of the protein waswashed through. Receptors andhospholipid were then elutedby inclusion of 0.4 M NaCl in thewash solution. T he yield ofreceptors in these fractions was 35% of the original extract.Fur the r washes with higher conc entratio ns of NaCl releasedsmall amou nts of protein and phospholipid but very l i t t leQNB-binding activity. The peakof eluted receptorsalso con-tained about 5% of the or iginal GTPyS -binding activity butessential ly no Gs activity (not shown ). Since I an d Go elutefrom DEA E slightly prior to Gs (14), mo st of the ac tivity inthe receptorpeak is probably due to other protein s th at b indguanine nucleotides.Th e peak of QNB-binding activitywas pooled, treated withXAD-2 beads, concentrated, and stored at -80 C (see M a-teria ls and M etho ds ). Th e overall yield of receptors fromthi s procedure was 15-20% relative to he original mem -branes; no significant purificationwas achieved. This prepa-ration of resolved recep tors was used in the recon stitutio nswith G-pro teins from bovine brain described below.T h e KO of 0.2 nM for bindingof [3H]QNB to thes e esolvedreceptors was simila r to alues obtained with untreated mem-branes . This prepara tion also isplayed the same muscarinicspecificity as theoriginal membranes (Table 11); he order ofpotency forcholinergic age nts to isplace [3H]QNB was iden-

    4.2 -1.0 -

    0.6 -0.4 -

    4 8 I2 I6 20 24 28 32 36 40Fraction Number

    FIG. 2. Resolution of muscarinic receptors from GTPyS-binding activity by DEAE-Sephacel chromatography. Proce-dures and assays are described under Materials and Methods.Fraction volumes of either 6 or 3 ml were collected in fractions 1-17and 18-38, respectively. Application of 0.4 M NaCl was begun duringfraction 18. Fractions 21-38were supplemented with lipid as de-scribed; determinations of inorganic phosphate were corrected forthese additions. Individual fractions of the initial flow-through werenot collected; collection of fraction 1 coincides with the applicationof wash solution.

    tical in the two prepa rations (compare Tables I an d 11), an dth e ICs0 values w ere similar.The appa rent affinity f resolved recepto rs for the gonist,oxotremorine, was low (ICs0 - 10 p M ) and was not affectedby guanin e nucleotides Fig. 3) . If, however, some of theGTPy S-binding activity rom the DEAE fractionswas addedTABLE1

    Effects of purified G-protein and GT P on the binding of cholinergicligands to resolved muscarinic receptors from bovine brainResolved receptors were prepared by DEAE chromatography andtreatment with XAD-2 beads, reconstituted with G-protein byMethod 3 (dilution), and assayed for binding of QNB as described(see Materials and Methods). G-protein was added at a 500-1000-fold molar excess over receptors. Assays contained about 40 fmol ofbinding sites, 0.19 nM [3H]QNB, and 0.1 mM GT P when indicated.The da ta were compiled from several experiments. Each ligand wasexamined at least twice, except for hexamethonium and d-tubocurar-ine; a ligand that showed a shift in affinity was included in eachexperiment. The qualitative results were always the same, althoughthe degree of reconstitution was variable. The purified G-proteinutilized was enriched in Gr as described under Materials and Meth-ods.

    IC,Ligand No G-proteinlu s-protein

    -GTPGTPGTPGTPP M

    Oxotremorine 10 15 0.25 20Carbachol 1000 1000 30 800Methacholine 450 450 120 700Acetylcholine 30 30 4.5 45Atropine 0.010 0.010 0.010 0.010Hexamethonium 6500 6500 5500 5500Nicotine 1900 1900 1400 1800d-Tubocurarine 60 60 60 120

    - 2.0-: .0-0c o12.0z 10.00 8.O

    4 oO1: ,G , ,/~028 1-9 -8 -7 -6 -5 -4 -3

    RECONSTITUTED 2.0 RECEPTOR \. s0 LOP [ ~ x o t r e m o r i n e ]FIG. 3. Reconstitution of resolved muscarinic receptorswith resolved G-protein from DEAE chromatography. Recep-tors were resolved from G-protein by DEAE chromatography. Frac-tions containing eceptor (about 5 ml) were applied to a 25-ml columnof Sephadex G-50 and eluted with 20 mM NaHepes, pH 8, 1 mMEDTA, 160 mM NaCl, and 0.1 mM oxotremorine. Turbid fractionswere pooledand utilized for reconstitution. Lower, the resolved recep-tors (0.5 ml) were mixed with 0.5 ml of a crude fraction of G-proteinfrom the DEAE flow-through and reconstituted by Method 1 (gelfiltration) as described under Materials and Methods; upper, theresolved receptors were subjected to he same treatmentsas forreconstitution with the exception of the addition of G-protein. Assayscontained 0.19 nM [3H]QNB and100p M GTP (0) .

  • 8/2/2019 florio 1985

    5/7

    Muscarinic Receptors: Interaction with GTP-binding Proteins 3481back to the receptors, the apparent affinity for oxotremorinewas increased by about 100-fold in the absence of GTP (IC60of 0.1 p ~ ) . ddition of GTP resulted in the complete reversalof this effect (IC6,, of10p ~ ) .hus, chromatography on DEAEseparated receptors from a component(s) necessary for theexpression of high affinity binding of agonists. This compo-nent(s)co-fractionated with GTPyS-binding activity andcould be used to restore the high affinity for agonists toresolved receptors.Reconstitution with Purified G-proteins-Two proteins withGTPyS-binding activity have been purified from bovine brain(14); these proteins have been designated GI and GO. Theaddition of either of theseproteins to resolved receptorsresulted in an increase in the apparent affinity of receptorsfor agonists. Guanine nucleotides reversed this effect com-pletely. Fig. 4 documents the displacement of t3H]QNB fromisolated receptors by oxotremorine in the absence or presenceof a GI-enriched preparation of purified G-protein from brain.In the presence of the GI, significant displacement was ob-served at low concentrations of oxotremorine (0.1 W M ) andthe IC, was decreased 15-fold. The IC, in the presence of GIplus GTP was the same as in the bsence of GI (10p ~ ) .

    Fig. 5 shows the effect of the a subunit of Go on the ffinityof receptors for oxotremorine. All preparations were supple-mented with equal concentrations of the purified /3 subunit;this subunit, by itself, did not affect agonist binding (seebelow;Fig. 6B). n he absence of added a subunit, thedisplacement of QNB was similar in the absence or presenceof GTP. The addition of a subunit in a 100-fold molar excessover receptors resulted in an increase in apparent agonistaffinity, with significant displacement at 0.1 p M oxotremorineand a decrease of 4-fold in the ICw. A 10-fold higher concen-tration of 1y subunit resulted in a slightly greater increase in

    15tl o t

    RECEPTOR 0 +GTPISOLATED 0 -GTP \%.,

    lot51 RECEPTORRECONSTITUTED L 4FIG. 4. Reconstitution of resolved receptors with purifiedG-protein from brain. Receptors were prepared by DEAE chro-matography and treatment with XAD-2 beads as described under"Materials and Methods." Resolved receptors (1.8 pmol, 5.5 mg ofprotein) in 0.95 ml were mixed with the purified preparation of G -protein enriched inGI (1.6 nmol in 0.12 m l) or 0.12 ml of I%cholateas a detergent control. Atropine, MgC12, and cholate were added tofinal concentrations of 10FM, 0 mM, and 0.476, respectively. Incor-poration of proteins into phospholipid was carried out by sedimen-tation through sucrose (Method 2, reconstitution, see "Materials andMethods"). Binding assays contained 0.19 nM ['HIQNB and 100 P MGTP (0).

    25

    I I I I I-7 -6 - 5 -4 -3Log [Oxotremorine]

    FIG. 5. Reconstitution of resolved receptors with purifiedGO.Receptors were prepared by DEAE chromatography and treat -ment with XAD-2 beads. The receptors (1.5 pmol) were mixed with1 nmol of purified /3 subunit and the indicated amount of purifiedGoa subunit (39,000Da) in a total volume of 0.5 ml. Reconstitutionwas carried ou t by Method 3 (dilution) as described under "Materialsand Methods." Binding assays contained 0.19 nM [3H]QNB and 100pM GTP (0) .agonist affinity with a 'I-fold decrease in IC5,,. In each case,100 ~ L M TP reversed the effect. The addition of a 100-foldexcess of Go (i.e. 100-fold excess of binding sites for GTPyS)is about the same as the number of binding sites for GTPySremaining in the preparation of receptor. The proteins re-sponsible for this residual binding activity either do notinteract with the muscarinic receptor or are much less potentthan Go for effecting the expression of high affinity foragonists.

    The increase in apparent binding affinity induced by GIand Go was specific for agonists of the muscarinic receptor(Table 11).GI and Go did not change the affinities observedfor muscarinic (atropine) and nicotinic (&tubocurarine) an-tagonists or for nicotinic agonists (nicotine). The differencesobserved in the extent of shifts observed for different mus-carinic agonists are due at least in part o variability inreconstitution obtained in separate experiments; no attempthas been made to quantitate effects of several agonists withinthe same experiment.The increased affinity for muscarinic agonists induced byGo and GI was specifically abolished by guanine nucleotides.The order of potency among various nucleotides for this effectwas the same as reported for effects on muscarinic agonistbinding to intact membranes (23). Thus, GTPyS (K,,, = IO-M ) was the most potent of all the nucleotides tested, followedby GppNHp MI, GTP M ), GDP (3 x M), ndITP (10"j M ) . Neither GMP nor ATP had any effect atM (data not shown).

    Reconstitution with Resolved Components of Go-The puri-fication procedure for the G-proteins from bovine brain yieldsthe free homogeneousa subunit of Go. This a subunit (39,000-Da polypeptide) binds guanine nucleotides, is stable at 0 "Cin cholate, and interacts with the /3 subunit, which decreasesits affinity for GTPyS (14). The effect of the resolved LY andfi subunits on agonist binding to isolated muscarinic receptorsis shown in Fig. 6. Fig. 6B shows tha t the 3 subunit alone didnotreconstitute guanine nucleotide modulation of agonistbinding; no change in apparent agonist affinity has been foundwith addition of up to a 3000-fold molar excess of /3 subunitover receptors. Addition of the a39 subunit alone did yield anincrease in agonist affinity (Fig. 6C); a 3-fold decrease in theIC6@f oxotremorine was obtained and was reversed by GTP.

  • 8/2/2019 florio 1985

    6/7

    3482 Muscarinic Receptors: Interaction w ithTP-bindingroteins

    Log [Oxotremorine ]FIG.6. Reconstitution of resolved receptors with purifiedsubunits. Receptors were prepared by DEAE chromatography andtreatment with XAD-2 beads. The receptors (1.6 pmol in 475 pl) weremixed with 90 pl of 1% cholate ( A ) ,70 pl of B subunit (2 nmol) and20 pl 1% cholate ( B ) ,20 pl of Goa subunit (0.34 nmol) and 70 p1 1%cholate (C), and 70 p1 of subunit (2 nmol) and 20 pl of Goa subunit(0.34 nmol) ( D ) . Reconstitution was accomplished by Method 3(dilution) as described under "Materials and Methods." Bindingassays contained 0.19 nM [3H]QNBan d 100p~ G TP (0) .

    This effect is smaller tha n those shown in Figs. 4 an d 5,whereboth a and @ subunits were added. If the same amounts ofpurified a and #3 subunits were added together (Fig. 6D), thedecrease in IC60 (9-fold) of oxotremorine was significantlygreater tha n with th e addition of the a subunit alone. In foursuch experiments, the decrease in IC, for oxotremorine was2.8 k 0.5-fold in the absence of added j3 subunit and 7.1 f0.5-fold when th e j3 subunit was included. Thus, th e j3 subunitappears to enhan ce the effects of the a subunit.

    DISCUSSIONWe have presented evidence for the functional interactionof two major guanine nucleotide-binding prote ins from bovinebrain with muscarinic receptors from the same source. Thecriteria we have used to measure this functional interactionwere the ability of the G -proteins to increase th e affinity ofmuscarinic receptors for agonists a nd the specific reversal ofthis effect by low concentrations of guanine nucleotides.Chromatography of extracts from brain membranes onDEAE-Sephacel in deoxycholate and cholate resulted in theseparation of receptors from 95% of the high affinity bindingsites for guanine nucleotides. These resolved receptors bound

    agonists with a low affinity tha t was no t affected by guaninenucleotides (Figs. 3 upper, and 4, upper). The addition ofeither GI or Go resulted in an increase in theaffinity of thesereceptors for agonists (Figs. 4, lower, an d 5) with no changein the affinity for 13H]QNBor other antagonists (Table 11).T he increase in affinities for agonists could be reversed withguanine nucleotides.While it is clear that a portion of the m uscarinic receptorsin these preparations hasbeen recon stituted with G-proteins,it is not clear th at all of the receptors were responding. Th eshallow competition curves obtained with agonists in theabsence of G T P suggest the presence of a t least two popula-tions of receptor, even when apparent saturating concentra-tions of G-protein were present. I t is not known whether th esites with low affinity for agonists after reconstitution are ueto defective receptors that no longer interact with G-proteins

    but still bind m uscarinic ligands, to experimental constraintssuch as vesicle composition that do not allow all of thereceptors to exist in a high affinity state, or to the presenceof two independent populations of muscarinic receptor. In th elatte r case , only one class of receptors w ould be capable ofregulation by G-proteins. While we have attempted to utilize[3H]pirenzepine o examine potential subclasses of m uscarinicreceptor, high nonspecific binding in reconstituted vesicleshas not allowed reliable evaluation of this possibility.Furthermore, it is not known whether GI and Go interactwith the same or different populations of muscarinic recep-tors. Experiments inwhich GI and GOwere added together orseparately to receptors revealed no conclusive differences inthe exte nt of Feconstitution achieved (data not shown). Th eobservation that GI and Go can achieve similar extents ofreconstitution within the same experiment suggests th at th etwo proteins may interact w ith the same population of recep-tor.The availability of the purified G-proteins enabled us toevaluate their role in more detail. T he action of the G -proteinson receptors was reversed by guanine nucleotides with thesame order of potency tha t was reported for effects on agonistbinding in membranes (23). Measurement of nucleotide bind-ing to the a subunit of GO also yielded the same order ofspecificity an d potency (14). This indicates th at th e ubunitsof GI and Go are the ites of action of guanine nucleotides onthe regulation of affinity of the muscarinic receptors foragonists.T he ability to prepare th e isolated j3 subunit of G-proteins(16) as well as the isolated a subu nit of Go (14) has allowedevaluation of the effects of these individual subunits on thebinding of agonists. The j3 subunit alone does not appear tomodulate agonist binding (Fig. 6B). The a39subunit doesinduce an increase in affinity for agonists (Fig. 6C), and thesimultaneous addition of a and j3 produc es a larger effect (Fig.6D). Th e effect of j3 subunit in enhancing th e effect of a wasprobably not due to stabilization of (Y against denaturation,since the recoveryof GTP yS-binding activity from thesereconstitutions was not alteredappreciably by the addition of@ subunit (not shown). The significance of the reconstitutionby th e a! subunit alone is not totally clear. It is likely th at th eresolved receptor preparation contained significant amountsof the j3 subunit? Therefore, it is possible th at both LY an d j3subunits are required for modulation of agonist binding andtha t the effect of (Y alone was due to cooperative action withendogenous j3. The assays for the j3 subunit in these crudeextracts have not proven sensitive enough to eliminate thispossibility. While th e j3 subunit certainly improves the inter -action of Goa with receptors, the determ ination of w hether @is a total requirement awaits further purification of th e recep-tors.

    We have not observed any major differences between GIand Go in this reconstitutive assay. A large excess of eitherprotein w as required before any clear change in agonist affin-ity was observed. T he requirement for large amounts of G-protein existed regardless of th e method employed for recon-stitution. Th is requirement did not seem to reflect inactiva-tion of th e G-proteins during th e procedures for reconstitutionsince the recovery of GTPyS-binding activity is 50-100%.T he maximal effects achieved with G O-or GI-enriched prep-aration s (still contain ing abo ut 5-10% Goa) are similar andoccurred at approximately a 1000-fold excess of G-proteinover receptors. Therefore, it appe ars unlikely tha t th e effect2T he behavior of G-proteins on DEAE during purification incholate results in variable quantities of the B subunits eluting withhigher salt than required for elution of the GTPyS-binding activity.

  • 8/2/2019 florio 1985

    7/7

    Fduscariniceceptors: Interaction with GTP -b ind ingrote ins 3403of either G-protein was attributable t o contamination by theother. An alternative explanation would be th at the activitywas due to th e presence of a minor contam inant in bothpurified Go and GI. This, however, seems unlikely as an ctivecontaminant would have to be present to the same extent inboth preparations of G-proteins and be a guanine nucleotide-binding protein with th e sam e specificity as GO.T he mechanism of reconstitution is not clear. Initial exper-iments with crude extrac ts suggested th at anactivated recep-to r (ie. n the presence of agonist, oxotremorine) enhancedthe interaction of G-proteins with the receptor during deter-gent treatmen ts. T his property was not observed after reso-lution of th e receptors from th e G-proteins; therefore, eitheroxotremorine or atropine could be used to stabilize the recep-tor with equivalent results. The difference in these experi-ments may indicate that there is another factor (lost duringresolution of receptors) involved in efficient interaction of theG-proteins with receptor. Alternatively, th e receptor h as beenaltered during esolution such tha t th e fficiency of its recon-stitution is no longer affected by th e two different ligands.Th ere are everal reasons to expect a large requirement forG-protein during these econstitutions. The first is the phys-ical state of ourpreparation. Although no morphologicalcharacterization of our recon stituted m ixtures ha s been per-formed, the ratio of lipid to protein (2-32 by weight) isfavorable for the fo rmation of vesicles upon r emov al of deter-gent by dilution or gel filtration. If such vesicles were 500-1000A in diameter, a typical size reported after the slowremov al of ionic deterge nts (24), there would be an average ofonly one recep tor for every 20-80 vesicles. If G-proteinspar tition equally amon g vesicles, a 20-80-fold excess of G-proteins over receptors would be necessary to approach con-ditions where each receptor had access to ateast one moleculeof G-protein.It is also possible that each receptor requires more tha n onemolecule of G-protein for the expression of a high affinitybinding site for agonists; this could be due to a stoichiometryof G-protein to receptor of greater than 1 or to kinetic con-sta nt s tha t equire high concentrations of G-protein to effecta high proportion of complex formation. Evaluation of thesepossibilities will be contingent upon purification of the mus-carinic receptor and reconstitution with highly concentratedreceptor preparations. A final explanation for the high re-quiremen t for G-protein is th e loss of reconstitutive activityduring purification. In this case, both the Go and GI prepa-rations would have to be equivalently susceptible to this lossof activity. While we have attempted to explain why thereconstitution may require large amounts of G-protein, wenote tha t this might be predicted, as th e native membranesfrom brain also contain 300-600 times as much G-protein asmuscarinic receptor.

    The reconstitution of high affinity agonist binding by theaddition of G-proteins to m uscarinic receptors dem onstratesthe existence of an interaction between these two proteins.Why is this interaction not readily observed in membranesfrom bovine brain andwhy does solubilization and subsequentremoval of detergent result in greatly increased interaction?Although th e 2-3-fold change in th e ICeofor agonists tha t weobserve in membranes is smaller than effects observed byothers in t he sam e tissue from other species, we have not beenable to increase this effect of gu anine nucleotides by treatme ntwith agonists for the receptor or with GM P.A large number of possible explanations for the poor inte r-action of receptors and G-proteins in membranes exist. Theprocedure for membrane preparation may resuIt in the phys-ical separation of G-proteins from receptors; therefore, a

    receptor in a membrane particle may have access to fewermolecules of G-protein than itdoes in an intac t ell. Solubil-ization of these rece ptors in the presence of agonist (F ig. 1)may then favor association with soluble G-proteins. It is alsopossible th at th e disruption of m embrane structure by solu-bilization improves the interaction between receptors and G-proteins.It is possible th at receptors and G-proteins n brain are, forthe most part, present in different cell types and, therefore,interact only after solubilization. However, the high quantityof GI nd Go in brain makes this seem unlikely. Gs and GIalso appear to be rather ubiquitous in that they are presentin all cell types tested so far. An intracellular separation ofreceptors and G-proteins, however, remains a possibility withpotentially interesting possibiIities for regulation. The evalu-ation of these possibilities awaits the cellular and subcellularlocalization of th e proteins.Th e results presented here demonstrate that a substantialproportion of the muscarinic receptors in bovine brain caninteract with regulatory G-proteins, in specific, two purifiedproteins, GI and Go, th at exist in high quan tity in this tissue(14). This suggests th at at eas t pa rt of the effects of musca-rinic agonists on cellular function a re directed through theseG-proteins. Th e procedures developed for isolation and recon-stitution of the muscarinic receptor will allow more detailedanalysis of its specificity and mechanism of interaction withvarious purified G-proteins.

    Acknowledgments-Superb techn ical assistan ce during exper imen-tation wasprovidedbyHowardCummings,Marcia Timko, andBarbara Creighton. We thank Dr. Murray Smigel for very usefuldiscussion of the manuscript and Nancy B ryant for skillful technicalassistance in its preparation.REFERENCES

    1. Rodbell, M., Krans, H. M. J., Pohl, S. L., and Birnbaumer, L. (1971) J.2. Ross, E.M., Maguire, M. E., Sturgill, T. W., Biltonen, R. L., nd Gilman,3. Tsai, B., and Lefiowitz, R. J. (1979) Mol. P h u rm o l . 1 6 , 6 1 4 84. Sternweis, P. C.. NorthuD, J. K., Sm iael.M. .. and Gilman. A. G. (1981)

    Blol. Chem. 246,1872-1876A. G. (1977)J. ioi. Chem. 262,5761-5775J.BioL Chem.~266 , l i517-11526 - 'Bwl. Chem. 263,6401-6412

    . .5. Rosa, E.M., Howlett, A. C.,Ferguaon, K. M., and Gilman, A. G. (1978) J.6. Sternweis, P. C., and Gilman, A. G. (1979)J.Bid. Chem. 264,3333-33407. Hsia, J. A., Moss, ., Hewlett, E. L., and Vaughan, M. 1984) . ioL Chem.269,1086-10908. Bokoch, G.M., atada, T., Northup, J. K., Ui, M., nd Gilman, A. G.9. Katada, T.,nd Ui, M. (1982) J.Biol. Chem. 267,7210-7216(1984)J.Biol. Chem. 269,3560-356710. Watanabe, A. M., McConnaughey, M. M., Strawbridge, R. A,, Fleming, J.W., Jones, L. R., and Besch, H. R. , Jr. (1978) J.Bid . Chem. 263 ,4833-483611. Onafi, P.. Olianas. M. C.. Neff.N. H.. and Costa.E. (1983) Mol. P h u r m o l .24,380-386 '

    6.367-375. . . . .

    12. Lkhtstein, D., oone. G., and Blume, A. (1979) J.Cyclic Nucleotide Res.13. Sokolo vsky, M. , Gurwitz, D., and Kloog.T. 1983)Adu. Enzymol. 6 6 , 1 3 7 -14. Sternweis, P. C., and Robishaw, J. D. 1984) J . Biol. Chem. 269, 13806-16. Northup, J. K.,Smiget, M. D. , Sternweis, P. C., and Gilman, A. G. 1983)15. Ross, E. M., and Schatz, G. (1978) Methods Enzymol. 63,222-22917. Yamamura, H. ., and Snyder, S. H. 1974) Proc. Natl. Acod. Sci. U. . A.18. Northup, . X., miael, M. D., and Gilman. A. G. (1982) J.Biol. Chem.

    . , ~ . .

    19613813J.Biol. Chem. 268,11369-1137671 , 1725-1729267,-11416-11423- . .19. Schaffn er, W., and Weissman, C. (1973) Anal. Biochem. S6 , 502-50420. Ames, B. N. (1966) Methcds Enzymol. 8,115-11821. Birdsall, N. J. M.. Bureen.A. S. V.. and Hulme.E.C. 1978)Mol.Phrmocol.- .14,723-760Biochem. Biophys.Res. Commun. 8 7 , 1000-1005Commun.94,487-492

    . I ~~~~22. Berrie, C. P., Blrdsall, N. J. M., Burgen, A. S. V. , and Hulme, E.C. (1979)23. Sokolov sky, M., Gurwitz,D., and Galron, R. (1980) Biochern. Biophys. Res.24. Szoka, F., and Papahadjopoulos, D. (1980) Annu. Reu. Biophys. Bioeng. 9,

    467-508