nical assistance and M. Mazzoni, K. Yan, N. Artemyev, and H. Rarick for valuable discussions. REFERENCES. Bradford, M. M. (1976) Anal. Biochem. 72,248- ...
Vol. 269, No. 34, Issue of August 26, pp. 21519-21525, 1994 Printed in U.S.A.
THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.
Synthetic Peptides as Probes for G Protein Function CARBOXYL-TERMINAL Gas PEPTIDESMIMIC G, AND EVOKE HIGH AFFINITY AGONIST BINDING TO P-ADRENERGIC RECEPTORS* (Received for publication, April 11, 1994, and in revised form, June 15, 1994) Mark M. Rasenick*, Masayuki Watanabeg, Milenko B. Lazarevic, Shinichi Hatta§, and Heidi E. Hammfl From the Department of Ph.ysio1ogy and Bioph-ysics and the Committee on Neuroscience, University of Illinois College of Medicine at Chicago, Chicago, Illinois 60612-7342
The molecular interfaces between G, and the P-adrenergic receptor were investigated using synthetic peptides corresponding to various regions of its a subunit, a,. These experiments were carried out on saponin-permeable C6 glioma cellsin which the P-adrenergicreceptor appearstightly coupled to G,. Synthetic site-specific peptides from as (corresponding to amino acids 15-29, 354-372,and 384-394)and ai(8-22,315-324,and 345-455) were tested for their ability to interfere with coupling between the P-adrenergic receptor and G,. The two carboxyl-terminal peptides from a, blocked P-adrenergic stimulation of adenylyl cyclasein permeable cells.However, onlya,-354-372had this effect in C6 membranes. It is suggested that the partial uncoupling of G,, which occurs subsequent to cell disruption, may berelated to a change in the interaction of the a, carboxyl terminus with the P-adrenoreceptor. Two carboxyl-terminalpeptides, 354-372 and 384-394,could also mimic the effect of G, to increase agonist affinity for the P-adrenergic receptor. In combination, a,-354-372 and a,-384-394 increased the ability of isoproterenol to compete with ”‘Ipindolol binding in a partially additive manner. Synthetic peptides from aiand amino-terminalpeptides from a, had no effect on P-agonist binding, suggesting a high specificityof peptide effects. Two findings suggest that these peptides bind directly to the p-adrenergic receptor and stabilize its high agonist affinity conformation. First, GTP and hydrolysis-resistant GTP analogs did not alter thehigh affinitybinding in thepresence of high concentrations of the peptides. Second,in S49 lymphoma cyc- cells, whichlack G,, these peptides evoked the high affinity agonist binding state of the P-receptor. Neither peptide had an effect on antagonist binding affinity, as measured by propranolol displacement of ”‘Ipindolol. Thesedata suggest that atleast two regions on the a subunit of G, participate in high affinity G, binding to the P-adrenergic receptor. The fact that these small peptides could mimic the holo-G, effect onthe receptor is rather surprising, and the specificity of the effect suggests that theprimary and secondary structure of small regions of a, contain much of the information for specific interaction with 0-adrenergic receptors.
* This work was supported by the American Heart Association, National Science Foundation GrantIBN 91-21540, National Institutes of Health Grants MH39545 and EY06062. The costsof 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. $ Recipient of a Research Scientist Development Award from NIMH, National Institutes of Health (MH 00699). 5 Current address: Depts. of Neuropsychiatry and Pharmacology, Sapporo Medical College, Sapporo, 060 Japan. ll Recipient of a Glaxo Cardiovascular Discovery Award.
The adenylyl cyclase system is a membrane-associated complex of receptors, G proteins, and enzyme that is activated or inhibitedin response to hormones andneurotransmitters (Dohlman et al., 1991). The G proteins G, and Gi are able, respectively, to stimulate or inhibit adenylyl cyclase. These signal-transducing G proteins are heterotrimeric in structure, consisting of a, P, and y subunits. The activation of G proteins by receptors requires the presence of both a and Py subunits and involves a tight interaction of the heterotrimer with an agonist-activated receptor followed by receptor-catalyzed GDP dissociation, GTP binding, and altered interaction between a and Py subunits (Gilman,1987). It is clear that thea subunit plays a major role in both receptor and effector interaction and also contains GTP binding and GTPase activities (Conklin and Bourne, 1993; Noel et al., 1993). The functional domains on the a subunit involved in receptor interaction are a subject of current interest. Evidence from a variety of studies implicates the carboxyl terminus inreceptor interaction (West et al., 1985; Hamm et al., 1988; Simonds et al., 1989; Conklin et al., 1993). It is probable, however, that multiple interaction sites exist on G protein a subunits for receptors, since synthetic peptides from the amino terminus of the a subunit of transducin block its interactionwith rhodopsin (Hamm et al., 1988). Furthermore, and anantibody that binds to theamino and carboxyl termini of at is capable of interfering withrhodopsin-transducin interaction(Hamm et al., 1987; Hamm et al.,1988; Navon and Fung, 1988; Hingorani and Ho, 1990; Mazzoni et al., 1991) and P-adrenergic receptor activation of adenylyl cyclase (Hamm et al., 1989). The natural milieu of the adenylyl cyclase system is the plasma membrane, and it possible is that components that are present in intactcells, but lost following cell disruption, may be important for the normal regulation of the adenylyl cyclase system (Rasenick et al., 1981; Denker and Neer, 1991). Elements of the cytoskeleton are also likely to contribute t o an ordered association among components of the adenylyl cyclase cascade (Rasenick and Wang, 1988; Wang et al., 1990; Roychowdhury et al. ). A saponin-permeableC6 cell modelhas been developed for the study of receptor-effector coupling in a more native environment (Rasenick and Kaplan, 1986; Rasenick et al., 1993). In thisstudy, we have compared the ability of synthetic peptides corresponding to various regions of G, to mimic G, effects on agonist affinity of the P-adrenergic receptor and to disrupt the P-adrenergic activationof G,. It was observed that a peptide corresponding to the carboxyl terminus of a, blocked P-adrenergic activation of adenylyl cyclase only in permeable cells, while another peptide correspondingto an internal site blocked this activation in both permeable cells and cell membranes. Thus, it appears that within an intact membrane/cytoskeletal system, the molecular interfaces between the P-adrenergic receptor and GBare more elaborate than those seen following membrane isolation.
21519
21520
P-Receptor-G Protein Interaction
TABLEI Radioligand binding studies on many receptors have shown Amino acid sequences of synthetic peptides multiple affinitystates for the bindingof agonist, and these are The sequenceswere taken from Jones and Reed (1987).Peptides were regulated by guanine nucleotides (Dohlman et al., 1991). For synthesized by the solid-phase Merrifield method withan Applied Prop-adrenergicreceptors, antagonists display a singlebinding tein Technologies synthesizer and purified by high-performanceliquid affinity. Several studies havesuggested that when a p-receptor chromatography. Identity of each peptide was monitored by amino acid is coupled to G,, the high affinity state for agonist binding analysis and fast neutron bombardment mass spectroscopy. results. The agonist-induced binding of GTP t o the G protein a,-38&394 QRMHLRQYELL causes uncoupling of that G protein from the receptor and a a8-acetyl-35&372-amide DGRHYCYPHSTCAVDTENI a,-15-29 EEKGQREANKKIEKQ subsequent shiftback to a lower affinity state for agonist bindEDKAAAERSKMIDKN a,-8-22 ing. Afunctionally analogous stabilization of the active receptor KDTKEIYTHF a,-315-324 conformation by G protein exists in the lightreceptor rhodopFLGCDKLNNKI a,-355-345 (nonsense) sin, inwhich the photoreceptor G protein transducin stabilizes metarhodopsin I1 (Hargrave et al., 1993). It hasbeen observed that syntheticpeptides corresponding t o carboxyl-terminal re- between 20 and 640 PM was usedto determine the dissociation constant insaturation exgions of the a subunit of transducin can mimic holo-G, and (K,)andthe density of P-adrenergic receptors stabilize active metarhodopsin I1 (Hamm et al., 1988). This periments. Agonist binding to P-adrenergic receptors was assayed by measurement of competition of isoproterenolfor IPIN binding sites. The suggests that similar regions from other G proteins might concentration of IPIN used (150PM for C6 cells and 200 PM for S49cycmimic or antagonize the interactionof a hormone receptorwith cells) was closeto the Kd value for the antagonist. Seven concentrations its cognate G protein. In thisstudy, peptidesfrom the carboxyl- of isoproterenolwere used ranging from 1n~ to 1m ~To. investigate the terminal regions of the a subunit ofG, have been shown to dose dependenceof peptide effects on agonist competition of IPIN bindincrease the agonist binding affinity of the p-adrenergic recep- ing, peptide concentrations of 1p~ to 1m~ were added in the presence tor. Thus, occupation of a single site on the receptor by a G of 100 n~ isoproterenol.The reaction mixture consisted of a final volume of 250 p1 containing Locke's solution, IPIN, and appropriate agents. protein peptide is sufficient to mimic the ability of heterotri- Assays were initiated by the addition of cells (120-140mg of total meric G, to stabilize the receptor in its active form. cellular protein) and were carried out for 50 min a t 30 "C. Assays were EXPERIMENTALPROCEDURES Preparation of Permeable Cells-Saponin treatment ofC6 glioma cells allows access for peptides or guanine nucleotides whilepreserving the receptor-G protein coupling seen in the intact cell (Rasenick and Kaplan, 1986; Rasenick etal., 1993). C6 glioma cells (passages 15-26) were grown to near confluency in Nuclon 175-cm2flasks in Dulbecco's modified Eagle's medium containing 4.5 giliter glucose and 10% ironsupplemented fetal bovine serum (Hyclone)in a humidified 37 "C incubator with 10% CO,. Just prior to confluence, cells werereleased from the flasks by brief treatment with 2 m~ EDTA in Dulbecco's modified Eagle's medium and gentle agitation. Cells werecentrifuged at 500 x g for 5 min, resuspended, and washed two times with Locke's solution (154 mM NaC1, 2.6 mM KCl,2.15 mM qHPO,, 0.85 KH,PO,, 10 m~ glucose, 2 mM CaCl,, and 5 m~ MgCl,, pH 7.4) at 500 x g for 5 min. Saponin solution (140 mM potassium glutamate, pH 6.8,2 m~ ATP, 100 pg/ml saponin) was added and mixed gently for 150s. The reaction was quenched by the addition of excess potassium glutamate buffer, and the cells werecentrifuged at 500 x g for 5 min and washed twice morewith potassium glutamate buffer. Permeable C6 cells wereresuspended with Locke's solution and assayed for receptor binding. All centrifugations were doneat room temperature. Preparation of permeable S49 cyc- cells was similar to C6 cells exceptthat horse serum was usedinstead of fetal bovine serum, and saponin exposure time was 20 s. Assay of Adenylyl Cyclase in Permeable C6 Cells-150 pl of Locke's solution containing [(Y-~~PIATP (2 x lo6 cpm), 0.5 mM ATP, 1mM MgCl,, and 0.5 mM isobutylmethylxanthine were added to each well and incubated for 3 min at 20 "C.Peptides, drugs, or other additions were added to each well, and the plates were incubated for 10 min at 37 "C. Reactions were stopped by adding 300 p1 of ice-cold 15 mM Hepes buffer pH 7.4 and placing the entire plate on dry icefor 5 min. Plates were removed from dry ice and allowed to thaw, and cells were scraped into tubes. Tubes were boiled for8 min in a heatingblock and centrifuged at 15,000 x g for 8 min at 4 "C. The supernatant was tested for CAMP content as described (Hatta et al. (19861, modification of Salomon (1979)). C6 glioma membranes were prepared as described in Rasenick and Kaplan (1986)and stored under liquid N, until use. Membrane suspensions were assayed for adenylyl cyclase as described (Rasenick et al., 1984). Binding Assay for Permeable Cells-"251-pindolol (IPIN)' was used to assay antagonist binding to permeable cells. IPIN at concentrations The abbreviations used are: IPIN, 1251-pindolol;G,, the a subunit of G protein; GppNHp, 5'-guanylylimidodiphosphate; G,, stimulatory GTP-binding regulatory protein of adenylyl cyclase; Gi, inhibitory GTPbinding regulatory protein of adenylyl cyclase; Go, a Gprotein abundant in brainwith unknown functions; Gt, a Gprotein of rod photoreceptors; a,-384-394, a synthetic peptide corresponding to those residues of Ga,
terminated by the addition of 10 volumes of ice-cold 2 m~ Hepes buffer pH 7.4, containing 5 mM MgCl, and 2m~ EGTA. Bound and free ligand were separated by rapid vacuum filtration (Brandel Cell Harvester model M24R) onWhatman GFiB filters. The filters were washed with 3 ml of the same buffer three times, and bound radioactivity was quantitated in Beckman a 9000y counter. Specificbinding was defined as the difference between IPIN binding in the absence and presence of 10 p~ 1-propranolol. Nonspecific binding was less than 20% for C6 cells and less than 15%for S49cyc- cells in each assay. Bindingparameters were analyzed to obtain the values of Cimpby using the LUNDON 1program. Scatchard transformation of each saturation isotherm resulted in linear plots with correlation coefficientsgreater than 0.97 in C6 cells and 0.99 in S49 cyc- cells. IPIN binding data were best fit by a one-site model with Hill coefficientsof approximately 1. The values of Ci, were 82 PM and 11.1 fmoYmg of total cellular protein in C6 cells, and 164 PM and 22.6 fmol in S49 cyc- cells. The Cimpof IPIN in C6 membranes was 145 fmol/mg of membrane protein. Since the membrane protein is likely to represent about 10%or less of the total cellular protein, the Chp figures are actually similar. Protein content was determined by the Coomassie Brilliant Blue binding method with bovine serum albumin as the standard (Bradford, 1976). Site-specificSynthetic Peptides-Peptides listed in Table I were synthesized by the solid-phase Merrifield method with an AB1 peptide synthesizer and purified by reversed-phase HPLC using a C-18 column. Analytical HPLC monitoredat 220 nm showedless than 2% impurity in all samples. The identity of each peptide was monitored by amino acid analysis, and fast atom bombardment mass spectroscopy and, in some cases, by peptide sequencing on an Applied Protein Technologies sequenator. Peptide ai-355-345 isa control peptide with inverted sequence. MateriaZs-'251-pindolol (2200Ci/mmol;3200 KBqhl) was purchased from DuPont NEN. GppNHp was purchased from Boehringer Mannheim, and (-)isoproterenol was from Sigma. 1-Propranolol wasa gift from Ayerst.[CY-~~PIATP (800 Ci/mmol)was purchased from Dupont NEN. GTP and GppNHp were from Boehringer Mannheim. Forskolin, NaF, and isoproterenol were from Sigma. All other reagents were of analytical grade. RESULTS
Synthetic peptides from various regions of the (Y subunits of G, and Gi were used as probes of contact regions between the p-adrenergic receptor and G,, since peptides can serve as competitive inhibitors of receptor-G protein interaction (Hamm et al., 1988, 1989). Table I shows the amino acid sequences of the synthetic peptides used in this study. (any Gprotein a subunit followed by numbers refers to the corresponding peptide); HPLC, high-performanceliquid chromatography.
21521
p-Receptor-G Protein Interaction
b.......*"..".
"I .%
'"
'
"""'
'
'
"""'
'
" " "
.1
.Ol
1
Peptide (mM)
none
-9
-8
-7
-6
-5
-4
Log Isoproterenol [MI
14'[b
none .1
60,L//' . 0 10
'
"""'
.001
'
"""'
'
Peptide
(mM)
'
'
"""'
'
-9
-8
-7
-6
-5
-4
Log Isoproterenol [MI
' ' " " J
1
FIG.1. Inhibition of isoproterenol-stimulated adenylyl cyclase by site-specificsynthetic peptides inC6 glioma cells and membranes. a , permeable C6 cells. Peptides a,-354-372 (closed squares), a,-384-394 (closed circles),and nonsense peptide a,-355-345 (open triangles) were incubated with permeable C6 cells at the indicated con. C6 centrations in the presence of isoproterenol and GTP (10 p ~ ) b, membranes. Effects of the same peptides on isoproterenol-stimulated adenylyl cyclase in C6 membranes. Assay conditions for cellsand membranes were as described in the text. Values are mean & S.E. of triplicates from a representative experiment that was repeated three times.
Effects of Synthetic Peptides on P-adrenergic Stimulation of Adenylyl Cyclase in Membranes and Permeable Cells-When permeable C6 cells were incubated with the P-adrenergic agonist isoproterenol and GTP (10p~ each), two synthetic peptides from the carboxyl-terminal region of as,a,-384-394 and asacetyl-354-372-amide, were able to dose-dependently decrease isoproterenol stimulation of adenylyl cyclase (Fig. la). Peptides a,-384-394 and a8-acetyl-354-372-amide decreased stimulation maximally by 71 and 50%,with IC,, concentrations of 50 and 20 PM, respectively. Amino-terminal a, peptides, as well as aipeptides, had no effect on adenylyl cyclase activity under any condition tested. All synthetic peptides from a, or aiwere without effect on adenylyl cyclase in thepresence of 10 PM forskolin or 10 mM NaF. Basal enzyme activity was also unaltered by these peptides. In C6 membranes, thepresence of P-agonists is not required for stimulation of adenylyl cyclase by hydrolysis-resistant GTP analogs (Rasenick and Kaplan, 1986). The effect of synthetic peptides on this less tightly coupled adenylyl cyclase was determined. The synthetic peptide aS-acetyl-354-372-amide blocked isoproterenol-activated adenylyl cyclase by 50% (IC,, 6 p ~ ) while , a,-384-394 and other peptides had no effect on C6 membranes (Fig. lb). Use of Permeable Cells for Receptor Binding Studies-P-Adrenergic receptor binding was studied using a permeable cell suspension method. This method allows investigation of p-adrenergic receptor behavior in nearly intact cells under conditions where this receptor appears tightlycoupled to G, t o effect
none
-9
-8
-7
-6
-5
-4
Log Propranolol [MI FIG.2. Effect of site-specificsynthetic peptideson agonist or antagonist competition of 150 PM IPIN binding sites to permeable C6 cells. Each assay was performed at 30 "C for 50 min as described under "Experimental Procedures." Results are expressed as specific binding of IPIN in thepresence of indicated ligand concentrations. Data points are the means of triplicate determinations in one of three representative experiments. a , tissues were prepared and isoproterenol inhibition of IPIN binding was assayed in the absence (open circles) or presence (open squares)of 100 p~ GppNHp, 150 PM a,-315324 (open triangles), or 150 PM a,-384-394 (closed circles); b, isoproterenol inhibition of IPIN was assayed in the absence (open circles)or presence (closed squares)of 100 PM a,-354-372; c, antagonist competition of IPIN was determined in the absence (open circles) or presence (closed circles) of 100 p~ ai-315-324 or 100 p~ a,-384-394 (open triangles).The IC,, values of propranolol were 1.5 p~ for control, 1.6 p~ for a,-315-324, and 1.7 p~ for a,-384-394, respectively. dpm, disintigrationdminute.
activation of adenylyl cyclase. It is noteworthy that some difference exists inreceptor binding behavior between permeable cells and membrane preparations (Rasenick et al., 1993). For example (as can be seen in Fig. 2) in permeable cells, agonist competition curves are shallow and remainso in thepresence of GppNHp or peptides. The reasons for this difference are unclear. Effect of Peptides on ZPZN Binding-Inhibition of the padrenergic radioligand antagonist IPINby the full agonistisopro-
21522
P-Receptor-G Protein Interaction
TABLEI1 Effectsof peptide a,-384394and GppNHp on isoproterenol competition of iodopindolol binding in permeableC6 glioma cells Isoproterenol was used to displace iodopindolol (150 PM) as described in the text. Kd values were calculated using the LIGANDprogram. Where only one value is given, a single site model provided the best fit. Values are means f S.E. for three to five separate experiments, each done in triplicate. Kd high
Control + G ~ P N H P(10-4) + peptide as-384-394 + peptide as-384-394 and GppNHp
Kd low
M
M
1.91 x 10-9
3.08 x 10-7 1.80x 10-5
1.62 x 10”’ 2.70 x 10-9
3.83 x
terenol is shown in Fig. 2a.The well established effect of GppNHp in lowering agonist affinity is observed by a shift in the agonist displacement curveto the right.Two synthetic peptides from the carboxyl-terminal region of as,a,-384-394 (Fig. 2a) anda,-354-372 (Fig. 2b), by contrast, increased the affinity of the P-receptor for isoproterenol and shifted the curve to the left. A peptide corresponding to a similar sequence of ai,ai-315324, had no effect on agonist affinity (Fig. 2a). To determine whether a,-384-394 would also change the affinity of p-adrenergic antagonists, theability of the antagonistpropranolol to displace IPIN was tested. The dose-responsecurve for propranolol was not altered in thepresence of either a,-384-394, or the control peptide, ai-315-324 (Fig. 2c). The competition binding isotherms inFig. 2a were analyzed by LIGAND (Table 11). In the presence of GppNHp (without peptide a,-384-394) or in the presence of peptide a,-384-394 (without GppNHp) a single-site model was significant ( p = 0.003 and 0.032, respectively). This suggests that, in the presence of 100 PM GppNHp (without a,-384-394), 100% of the receptors are in thelow affinity state, while in thepresence of 100 PMa,-384-394 (without GppNHp), 100%high affinity binding is displayed. Data for control conditions and for conditions in which a,-384-394 and GppNHp are added (both at 100 PM) best fit a two-site model by LIGAND analysis. In control conditions, 72.4% of the receptors display high affinity binding, while 27.6%are low affinity. For a,-384-394 plus GppNHp, the values are 75.2 and 24.8%,respectively. Table I11 shows IC,, values of isoproterenol inhibitionof IPIN binding in the absence or presence of various peptides from a, or ai in C6 cells. Peptides from the amino-terminal region of either asor ai,as well as peptides from the carboxyl-terminal region of ai,had no significant effect on IC,, values, while the IC,, in thepresence of a,-384-394 was %fold lower than control values. In the presence of a,-354-372, the IC,, value was similar t o that of a,-384-394 (data not shown). Thus, eitherof two peptides from the carboxyl-terminal region of a, could mimic the holo-G, protein t o promote an increased agonist affinityfor P-receptors. The above studies were conducted in the presence of a fixed amount of each synthetic peptide (100 p).To examine the peptide dose dependence of the increase in P-agonist affinity, one concentration of isoproterenol (100 n ~was ) chosen, which half-maximally inhibited IPIN binding under control conditions. The dose-response curve for the effect of peptide a,-384394 on agonist binding is shown in Fig. 3. Under these conditions, the peptide a,-384-394 increasedagonist-displaced binding from approximately 15 to 65%. Thus, the peptide increased the potency of isoproterenol to maximally displace M. IPIN binding from to Effect of GppNHp on Peptide-induced Increases in P-agonist Affnity-The peptide-induced increase in agonist affinity is not expected t o be altered by guanine nucleotides, since pep-
tides would not be influenced by these compounds. Data inFig. 4a indicate that even in the presence of 100 p~ a,-384-394, GppNHp can stilldecrease agonist affinity. This suggests that, at this peptide concentration, the G protein G, and the peptide compete for binding to the P-receptor. This possibility was further studied asa function of peptide concentration in thepresence of 100 nM isoproterenol (Fig. 4b). At peptide concentrations higher than 200 PM, there was a negligible difference in agonist-displaceable binding in the presence and absence of GppNHp. This suggests that the peptide has two effects, a direct binding t o P-adrenergic receptorsand a competition with presence of endogenous G, for binding t o receptors. Thus, in the peptide concentrations that uncouple all endogenous G, from 0-receptors, there is no longer any effect of GppNHp. Multiple Sites of Interaction with Receptor-Previous studies with the photoreceptor system indicated that synthetic peptides corresponding to homologous regions of at were synergistic in their effect (Hamm et al., 1988). At low concentrations, a,-354-372 and a,-384-394 have additive, but not synergistic, effects (Fig. 5). At higher concentrations, however, only partial additivity is seen. Peptide Effects on Agonist Affinity in S49 cyc- Cells”S49 cyclymphoma cells lack G, but retain p-adrenergic receptors. If G, peptides exert theireffect by binding to P-adrenergic receptors rather than affecting G, directly, a similar result would be predicted for C6 and cyc- cells. The potency of isoproterenol in displacing IPIN was enhancedby 100 p a,-384-394 in permeable S49 cyc- cells. Calculated IC,, values for isoproterenol were 546 n~for control and 123 nM in the presence of peptide 384-394. DISCUSSION
This report demonstrates that site-specific synthetic peptides can disrupt thecoupling between the P-adrenergic receptor and G, and mimic G, in eliciting high affinity agonist binding by the receptor. Evidence from many laboratories supports the notion that thehigh affinity component of agonist binding occurs within a ternary complex of agonist, receptor, and G,, while the low affinity component represents agonist binding to free receptor (Dohlman et al., 1991). Guanine nucleotides perturb the ternary complex and cause a decrease in agonist affinity. In this work, we have found that two small synthetic peptides from the carboxyl-terminal a, sequence increase agonist affinity compared with control membranes with a normal complement of P-receptors and G,. Several lines of evidence suggest that these peptides bind directly t o the P-adrenergic receptor and stabilize the agonistbound conformation. First, GppNHp did not decrease agonist affinity in the presence of high concentrations of peptide a,384-394, suggesting that the peptide competes with G, for a common binding site on the receptor. Second, in S49 cyc- lymphoma cells, which lack G, but retain p-adrenergic receptors, peptide a,-384-394 increased the affinity for agonist. Third, there was no effect of these peptides on antagonist affinity. Thus, it appears likely that thesepeptides bind to siteson the P-receptor that normally interact withan,Analysis of the data shows that the effect of the peptides is to convert biphasic competition binding isotherms into high affinity monophasic isotherms (Table 11).This would suggest that thepeptides can mimic G, effects; when receptors are not occupied by G,, the peptides can convert free receptors with low agonist affinity into the highaffinity state. This peptide-induced high affinity agonist binding is analogous to theability of Gat-derived carboxyl-terminal peptidesto stabilize the active metarhodopsinI1 conformation of rhodopsin (Hamm et al., 1988). This suggests an important role of the carboxyl terminus of a subunits in mediating G protein-in-
21523
p-Receptor-G Protein Interaction TABLEI11
IC,, of isoproterenol competition for 1251-pindololbinding sites in theabsence or presence of 0.1 mM peptide in permeableC6 cells Agonist binding to /3-adrenergic receptors was assayed by measurement of competition of isoproterenol for IPIN binding sites. SevenconcenM 1mM. The concentrationof added IPIN was 150 PM. The values represent means 2 standard trations of isoproterenol were used ranging from 1 I ~ to errors (MI). Values determined by Student's t-test.
275 393 360
No.
None
1 2 3
288 336 262 22" 295 f342
a,-a22
192 317 363 f
35'
a,-315-324
319 377 366 354 2 18'
a,-15-29
a,-38P394
291 f 51d
108 73 88 90 f 10"
.apNot significantly different from 6, c, and d. Not significantly different froma, c, and d. e Not significantly different from a , b, and d. Not significantly different from a , b, and d. e Significantly different from a ( p < 0.001),b (p < 0.003),c ( p < 0.001),and d (p < 0.02).
'
7-
1000L11 none - 9 I
,
I
-8
,
I
-7
,
I
-6
,
I
-5
.
1
-4
Log Isoproterenol [MI FIG.3.Dose dependenceof synthetic peptide effects. The effect of varying concentrations of synthetic peptides ai-315-324(open triangles) or a,-384-394 (closed circles) was examined onthe competition of 0.1 isoproterenol for 150 PM IPIN binding to permeable C6 cells. Data points are means of triplicate determinations in one of three representative experiments. dpm, disintigrationdminute.
duced stabilization of the active state of receptors. That a G protein simply stabilizes an active conformational state of agonist-bound receptor, which can be attained even in theabsence of the G protein,has long been clear in the case of rhodopsin but has recently been highlighted by the finding that constitutively activated mutant /3-adrenergic receptors have higher agonist affinity evenin theabsence of G proteins (Samamaet al., 1993). The important role of the carboxyl-terminal of G protein a subunits for receptor interaction has been established in vari2000 L//--... . , , .... . . . . .- b ous ways. Pertussis toxin ADP ribosylation of a cysteine resio .001 .01 .I 1 due 4 residues upstream from the carboxyl terminus of ai,a,, Peptide Gs 384 (mM) and at blocks interaction of these proteins withreceptors (Van FIG.4. a , competition between a,-384-394 and GppNHp on agonist Dop et al., 1984; Kuroseet al., 1991; Cote et al., 1984). Synthetic (open circles) or presence (open peptides from the carboxyl-terminal region of at both inhibit affinity was studied in the absence rhodopsin-G, coupling and mimic G, in stabilizing the active squares) of 100 PM GppNHp, in the presence of 100 p~ a,-384-394 (closed circles), or in the presence of 100 PM GppNHp plus 100 PM metarhodopsin I1 conformation (Hamm et al., 19881, and a, a,-384-394 (closed squares). While GppNHp decreased the potency of carboxyl-terminal peptides affect P-receptor-G, coupling (Palm isoproterenol, the peptide elicited a n increased agonist affinity. In the et al., 1990). Anti-peptide antibodies produced against a car- presence of peptide, the GppNHp-induced decrease in agonist affinity boxyl-terminaldecapeptide of ai attenuate the receptor-de- was diminished. Data represent the meanof triplicate determinations in a typical experiment which was repeated four times. b, effect of pendent inhibition of adenylyl cyclase (Simonds et al., 1989). increasing concentrations of a,-384-394 on 100 V M isoproterenol comMastoparan, a receptor mimetic tetradecapeptide from wasp petition of IPIN binding in the absence (open circles) or presence (closed venom (Higashijima et al., 19901, was shown to couple to the circles) of 100 PM GppNHp in permeableC6 cells. At the higherconcencarboxyl-terminal region of ai,and its action was blocked by trations of cu,-38&394 used here, GppNHp hadonly a negligible effect. dpm, disintigrations/minute. antibodies against the aicarboxyl-terminal (Weingarten et al., 1990). The effect of carboxyl-terminal peptides from a, on stabilization of P-agonist affinity shows that this region is not of a, may be involved in preceptor coupling. The homologous simply involved in contact with the receptor, but it canmimic region of at is also involved in rhodopsin coupling, and in both even the functional coupling of the agonist binding site with thesystems, combinations of peptides have additive or synergistic G protein. effects (Hamm et al., 1988). This study provides the first eviThe finding that another region, 30-34 aminoacidsupdence that this region is important in other receptor systems. stream from the a, carboxyl terminus, is also involved in re- The region, a,-354-372, is quite highly conserved in all heteromultiple regions trimeric G proteins (Pupilloet dl., 19891, and a small portion of ceptor coupling is the first demonstration that 1
I-
,.....I
21524
P-Receptor-G Protein Interaction 50001
tor-G, interface, since stabilizationof the active conformation of both receptors is similarly achieved by carboxyl-terminal peptides. It is noteworthy that a conformational change around Lys345occurs upon binding of the carboxyl-terminal peptide from at (340-350) to metarhodopsin I1 (Dratz et al., 1993). The unc mutation of 549 lymphoma cells, in which receptor and G, are uncoupled, changes the homologous AI-$*^ to Pro, which would not have the sameconformational flexibility. Thus, conformational change of the G protein carboxyl terminus may be important for productive G protein-receptor interaction and subsequent receptor-mediatedGDP releaseand Gprotein N“ 2000 . .... , . . . . . . ...., . activation. c 0 .001 .01 .1 1 We have shown that two small peptides from the carboxylPeptide (mM) terminal region of a,can each mimic G, effects on P-adrenergic FIG.5. Partial additivity of a,-364-372 and a,-384-394 on in- receptors. Although either peptide can elicit a nearly maximal crease of agonist binding affinity to P-adrenergic receptors in permeable C6 cells. Indicated concentrations of a,-354-372 (closed effect, combinations of the two peptides are partially additive squares), a,-38&394 (closed circles), or indicated concentrations of aB- (Fig. 51, leading to a multisite model of receptor-(=.protein in384-394 + 100 p~ a,-35&372 (closed triangles) were incubated with teraction. These data areconsistent with severalpoint-to-point 100 V M isoproterenol and 150 PM IPIN. This experiment indicatesthat interactions between receptors and G proteins at regions that these peptides may exhibit a partially additive effect on increasing might be some distance from each other on the primary seagonist affinity. Dataare expressed as lZ5I disintigrationdminute (dprn) of specific IPIN binding and represent the meansof triplicate determi- quence of the proteins. Thus, itis likely that three-dimensional nations from one of three experiments. folding of these proteins is critical for high affinity molecular recognition. This may helpto explain why “consensussethis region (Thr-Cys-Ala-Thr) is found also in a subclass of quences” of receptors and G proteins have not been found, small G proteins related to the ADP-ribosylation factor (Sewell despite the wealth of information on primary amino acid seand Kahn, 1988). This region of a,,Thr320-Thr323, has recently quences of both protein families. The elegant studies of the been shownto be involved in bindingt o the guanine ring of the mechanism of action of mastoparan (Higashijima et al., 1988, guanine nucleotide (Noel et al., 1993). The possibility that this 1991) demonstrate that thetwo- and three-dimensional conforregion also interacts with receptors suggests a putative mech- mation of interacting sitesis critical for the interaction to take anism for receptor-mediated GDP release, since reorientation place. Our studies demonstrate thatsome small peptides have of the guaninebinding side chains would cause a diminution of the ability to fold into the appropriate conformation in the GDP affinity. Mutagenesis of a, has confirmed this find- context of a recognition site on the receptor. It is noteworthy that either a,-354-372 or a,-384-394 was ing (Thomas et al., 1993). Furthermore, comparative studies of able toacheive a maximal effect on increasing P-receptor affininteraction of various classes of GTP binding proteins with their cognate guanine nucleotide exchange factor proteins sug- ity (Fig. 5). This suggests that, although multiple sites on G gest that the exchange factor interacts with the TCAT region, proteins may normally contactthe P-receptor, occupation of one which is directly involved in binding the guanine ring (Hwangof those sites is sufficient to mimic the binding of a G protein. Increasing concentrations of peptide eliminate the ability of et al., 1992, 1993). It is clear that other regions of the G protein maybe involved GTP to shift receptor into the low affinity state for agonist in receptor interaction. The amino-terminal region of at was binding. This resultis consistent with the notion of a continushown to be involved in rhodopsin interaction (Hamm et al., ous competition between a,and peptide for a bindingsite on the 19881, and this may be the case with otherG proteins aswell, receptor, which can be overcome by great excess of peptide since Higashijima et al. (1990) have shown that the receptor relative t o a, (Fig. 4b). Such a mechanism might also be inmimetic tetradecapeptide mastoparan can be cross-linked t o volved in the ability of the a, peptides t o shift the @receptor the amino terminus of ao.The crystal structureof the a subunit into a higher affinity state for agonist. Peptides bind to the of G, shows that theamino and carboxyl termini are in spatial appropriate sites to mimic high agonist affinity without disproximity (Noel et al., 1993). In this study,amino-terminal playing the comprehensive characteristics of the receptor-G peptides from a, had no effect on P-adrenergic receptor-G, cou- protein complex. Even in the absence of agonist, this complex pling or agonist affinity. Thus, there arekey differences in the may have some inherent function. The observation that reconexperimental findings regarding regions on different G protein stituted receptor-(; protein complexes (without agonist) have GTPaseactivity that is not seen with the G protein alone a subunits of interaction with their cognate receptors. These findings provide a rationale for the study of multiple receptor (Kurose et al., 1991) is consistent with thisidea. Thus, receptor systems toelucidate a general mechanism and todiscern varia- plus peptide representsthe“pristine” highaffinityagonist binding state. tions on the common theme. It is of interest that the ability of a given peptide to block the The fact that homologous peptides from the sequence of ai had no effect on agonist affinity (Table 111)suggests that short coupling between the P-adrenergic receptor and G, (Fig. 1)is peptide sequences carry a surprising degree of specificity. An dependent upon the environment surrounding theseproteins. average of40%of the amino acids of the carboxyl-terminal Thus, while the a, peptide 354-372 blocks p-adrenergic stimuregion are identical between a, and ai;thus, the specificity lation of adenylyl cyclase in both permeable cells and memdeterminants must reside in the unique residues of the se- branes, the carboxyl-terminal peptide, 384-394, is capable of quence. Peptide proton NMR and peptide substitution studies blocking a, activation by the activated P-receptor only when in have shown that a P-turn at the carboxyl terminus of at involv- a near native environment. Furthermore, the carboxyl-termiing Gly348,Leu349,and Phe350is critical for the peptide- nal peptide appears to be disrupting only the “tightlycoupled induced stabilization of active metarhodopsin I1 (Dratz et al., state” of the P receptor-G,-adenylyl cyclase complex. Two possible reasons for this discrepancy exist. First, it is 1993). It may be possible t o extrapolate from these structural studies on rhodopsin-G, interaction to the P-adrenergic recep- possible that disruptionof the membrane andcytoskeletal en-
-
L,/. I
.....I
__I
Interaction ProteinP-Receptor-G vironment alter the receptor such that the site available for binding 384-394 is no longer accesible to the aqueousenvironment. Alternately, disruption of the membrane may increase the strength of the receptor-G, interface so that the carboxylterminal peptide no longer competes with as.It is known that disruption of the cell frees Ga, from the constraint normally imposed by the P-receptor. This is evidenced by the ability of adenylyl cyclase in theabsence GTP or GTP analogs to activate of P-agonist. A mechanistic explanation for this might be that the change in the association between receptor and Ga following cell disruption fosters a tighter interaction between the receptor. This carboxyl-terminal region of the CY subunit and the would constitute a primary recognition event, which could prime a secondary interaction between the receptor and the thus proguanine nucleotide binding region within CY,-354-372, moting guanine nucleotide release. According to thismodel, in the undisrupted cellular milieu, G proteins in their GDP-bound state have a low affinity for receptors. G protein activation occurs as a two-stage process. First, receptor recognition and primary activation due to contact between the activated receptor and the carboxyl-terminal of the CY subunit, followed by a secondary interaction of the receptor with the TCAT region leading to guanine nucleotide release. Upon cellular disruption, especially in certain cell and tissue types, a changed environment (including, perhaps, low GDP concentration) promotes the high affinity mode of receptor-G protein interaction preventing access of the ligand to its binding pocket. Thus, this study may suggesta molecular rationale for the well documented "uncoupling" between receptor and G protein seen following cell disruption. Through the use of site-specific synthetic peptides, the coupling between receptor, G protein, and intracellular effector molecules can be explored and some of the intracacies of in situ G protein-mediated signal transduction might be revealed. Acknowledgments-We thank F. Belga and M. Talluri for expert technical assistance and M. Mazzoni, K. Yan, N. Artemyev, andH. Rarick for valuable discussions. REFERENCES Bradford, M. M. (1976) Anal. Biochem. 72,248-254 Conklin, B. R., and Bourne, H. R. (1993) Cell 73, 631-1 Conklin, B. R., Farfel, Z., Lustig, K D., Julius, D., and Bourne, H.R. (1993)Nature 363,274-276
21525 Cote, T. E., Frey, E. A,, and Sekura R. D. (1984) J. Biol. Chem. 269,86934698 Denker, B. M., and Neer, E. J. (1991) FEES Lett. 279,98-100 Dohlman, H. G., Thorner, J., Caron, M. G., Lefkowitz, R. J. (1991) Annu. Reu. Biochem. 6 0 , 6 5 3 4 8 Dratz, E. A,, Furstenau, J. E., Lambert, C. G., Thireault, D. L., Rarick, H. R., E. (1993)Nature 363,276-281, Schepers, T., Pakhlevaniants, S., and Hamm, H. 1993 Gilman, A. G . (1987) Annu. Reu. Biochem. 66,615-649 Hamm, H. E., Deretic, D., Hofmann, K. P., Schleicher, A,, and Kohl, B. (1987) J. Biol. Chem. 262, 10831-10838 Hamm, H. E., Deretic, D., Arendt, A,, Hargrave, P. A,, Koenig, B., Hofmann K. P. (1988) Science 241, 832-835 Hamm, H. E., Deretic, D., Mazzoni, M.R., Moore, C. A., Takahashi, J. S., and Rasenick, M. M. (1989) J. Biol. Chem. 264, 11475-11482 Hargrave, P. A,, Hamm, H. E., and Hofmann, K. P. (1993) Bioessuys 16, 1-8 Hatta, S., Marcus, M. M., and Rasenick, M. M. (1986)Proc.Natl. Acad. Sci.U. S. A. 83,5439-5443 Higashijima, T., Uzu, S., Nakajima, T., and Ross, E. M. (1988) J. Biol. Chem. 263, 6491-6494 Higashijima, T., Burrier, J., and Ross, E. M. (1990) J. Biol. Chem. 266, 14176-86 Hingorani, V. N., and Ho, Y.-K. (1990) J. Biol. Chem. 266, 19923-19927 Hwang,Y.-W., Carter, M., andMiller, D. L.(1992) J. Biol. Chem. 267,22198-22205 Hwang, Y.-W., Zhong, J.-M., Poullet,P., and Parmeggiani,A. (1993) J. Biol. Chem. 268,24692-24698 Jones, D. T., and Reed, R. R. (1987) J. Biol. Chem. 262, 14241-14249 Kurose, H., Regan, J. W., Caron, M. G., and Lefkowitz, R. J. (1991) Biochemistry 30,3335-3341 Mazzoni, M.R., Malinski, J. A,, and Hamm, H. E. (1991) J. Biol. Chem. 266, 14072-14081 Navon, S. E., and Fung, B. K.-K. (1988) J. Biol. Chem. 262, 15746-15751 Noel, J. P., Hamm, H. E., and Sigler, P. B. (1993) Nature 366, 654-663 Palm, D., Munch, G., Malek, D., Dees, C., and Hekman,M. (1990) FEES Lett. 261, 294-298 Pupillo, M., Kumagai, A., Pitt, G . S., Firtel, R. A,, and Devreotes, P. N. (1989) Proc. Natl. Acad. Sci. U.S. A. 86, 4892-4896 Rasenick, M. M., and Kaplan, R. S. (1986) FEBS Lett. 207, 296-301 Rasenick, M. M., and Wang, N. (1988) J. Neurochem. 61,300-311 Rasenick, M. M., Stein, P. J., and Bitensky, M. W. (1981) Nature 294,560-562 Rasenick M. M., Wheeler, G., Bitensky, M. W., Kosack, L., Malina, R., and Stein,P. J. (1984) J. Neurochm. 43, 1447-1454 Rasenick, M. M., Lazarevic, M., Watanabe M., and Hamm, H. E. (1993) Methods (Orlando) 6,252-257 Roychowdhury, S., Wang, N., and Rasenick, M. M. (1993) Biochemistry 32,49554961 Salomon, Y. (1979)Adu. Cyclic Nucleotide Res. 10, 35-54 Samama, P., Cotecchia, S., Costa T., and Lefkowitz, R.J. (1993)J . Biol. Chem. 268, 4625-4636 Sewell, J. L., and Kahn, R. A. (1988) Proc. Natl. Acad. Sci. U. S. A. 86, 4620-4624 Simonds, W. F., Goldsmith, P. K., Codina, J., Unson, C . G., and Spiege1,A. M. (1989) Proc. Natl. Acad. Sci. U. S. A. 86, 7809-7813 Thomas, T. C., Schmidt, C. J., andNeer, E. J. (1993)Proc. Natl. Acad.Sci. U. S. A. 90, 10295-10299 Van Dop, C., Yamanaka, G., Steinberg, F., Sekura, R. D., Manclark, C. R., Stryer, L., and Bourne, H. R. (1984) J. Biol. Chem. 269,23-26 Wang, N., Yan, K., and Rasenick, M. M. (1990) J. B i d . Chem. 266, 1239-1242 Weingarten, R., Rasnas, L., Mueller, H., Sklar, L. A,, and Bokoch, G . M. (1990) J. Biol. Chem. 266, 11044-11049 West, R. E., Moss, J., Vaughan, M., Liu, T., and Liu, T.Y. (1985) J. €501. Chem. 260, 14428-14430
