Proc. Natl. Acad. Sci. USA Vol. 88, pp. 2171-2175, March 1991
Physiology/Pharmacology
Reconstitution of high-affinity opioid agonist binding in brain membranes (ju opioid receptor/receptor-G protein coupling/membrane-membrane fusion/C6 glioma cells/brain synaptosomes)
ANN E. REMMERS AND FEDOR MEDZIHRADSKY* Departments of Biological Chemistry and Pharmacology, The University of Michigan Medical School, Ann Arbor, MI 48109
Communicated by Avram Goldstein, December 3, 1990 (received for review October 27, 1990)
ABSTRACT In synaptosomal membranes from rat brain cortex, the 1L selective agonist [3H]dihydromorphine in the absence of sodium, and the nonselective antagonist [3Hlnaltrexone in the presence of sodium, bound to two populations of opioid receptor sites with Kd values of 0.69 and 8.7 nM for dihydromorphine, and 0.34 and 5.5 nM for naltrexone. The addition of 5 ,uM guanosine 5'-[y-thioltriphosphate (GTP[yS]) strongly reduced high-affinity agonist but not antagonist binding. Exposure of the membranes to high pH reduced the number of GTP[y-rS] binding sites by 90% and low K., opioid-sensitive GTPase activity by 95%. In these membranes, high-affinity agonist binding was abolished and modulation of residual binding by GTP[yS] was diminished. High-affinity (Kd, 0.72 nM), guanine nucleotide-sensitive agonist binding was reconstituted by polyethylene glycol-induced fusion of the alkali-treated membranes with (opioid receptor devoid) C6 glioma cell membranes. Also restored was opioid agoniststimulated, naltrexone-inhibited GTPase activity. In contrast, antagonist binding in the fused membranes was unaltered. Alkali treatment of the glioma cell membranes prior to fusion inhibited most of the low Km GTPase activity and prevented the reconstitution of agonist binding. The results show that highaffmiity opioid agonist binding reflects the ligand-occupied receptor-guanine nucleotide binding protein complex.
In addition to stereospecificity, a functional characteristic of all opioid alkaloids (1), these compounds show biphasic receptor binding at equilibrium. In rodent brain membranes, the high- and low-affinity binding components of morphine typically exhibit affinities of approximately 0.3 and 3.0 nM, respectively (2-4). Similar binding properties were described for the opioid antagonists [3H]naloxone (2, 4) and [3H]naltrexone (NTX) (5). Initially, it was shown that such heterogeneity, in some instances, reflected either nonspecific tissue association of the opioid or was an artifact of the experimental conditions (5). An alternative explanation for the biphasic Scatchard plots obtained was the marginal selectivity of the ligand used for the multiple types of opioid receptors present in neural tissue (6-8). On the other hand, guanine nucleotides, in particular GTP and its nonhydrolyzable analogues in the presence of sodium, reduced radiolabeled dihydromorphine (DHM) but not naloxone binding at equilibrium (4), thereby abolishing the high-affinity agonist component (3). Guanine nucleotides also suppressed the binding of both ,and 8-selective opioid peptide agonists (9). A role for guanine nucleotide binding proteins (G proteins) in the interaction of opioid ligands with their receptors was further supported by the effects of pertussis toxin in membranes from NG108-15 neuroblastoma x glioma hybrid cells and rat brain: exposure to the toxin diminished the affinity and GTP sensitivity of opioid agonist binding (10, 11). The functional significance of The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. 2171
the modulation by GTP of ligand binding was demonstrated by the opioid agonist-induced stimulation of low Km GTPase, reflecting the mechanism of receptor-effector coupling (12, 13). Although the results outlined above have implicated modulationt of ligand-receptor interaction by G proteins, they did not establish the molecular identity of the high- and lowaffinity components of opioid binding. In this study, we have directly linked high-affinity opioid agonist binding and opioid-sensitive GTPase activity to membrane G protein. Inactivation of G protein at high pH abolished high-affinity agonist, but not antagonist, binding. Guanine nucleotidesensitive high-affinity opioid binding and opioid-sensitive GTPase activity were restored after fusion of the alkalitreated synaptosomal brain membranes with membranes from opioid receptor-deficient, but inhibitory G proteincontaining, C6 glioma cells. The results describe high-affinity opioid agonist binding as a ternary complex composed of ligand, receptor, and G protein. Guanine nucleotidedependent dissociation of the complex then yields the lowaffinity form of the ligand-occupied receptor. A preliminary report on these findings was presented recently (14).
MATERIALS AND METHODS Materials. [3H]NTX was provided by the National Institute on Drug Abuse. [3H]DHM and [Y-32PIGTP were from Amersham, guanosine 5'-[y-[355]thiojtriphosphate (GTP[y-35S]) was from New England Nuclear, GTP[yS] tetralithium salt was from Boehringer Mannheim, and PEG with an average molecular weight of 8000 was from Sigma. The unlabeled opioids were obtained through the Narcotic Drug and Opiate Peptide Basic Research Center at the University ofMichigan. Isolated Membranes and Cell Culture. The synaptosomal preparation was isolated from brain cortices of male SpragueDawley rats as described (15). Prior to their freezing at -80'C, the synaptosomes were disrupted in a Dounce homogenizer. To reach confluence, the C6 glioma cells were grown as a monolayer in a minimal essential medium supplemented with 10% fetal bovine serum and 1% pyruvate for 8-9 days at 370C in 95% air/5% CO2. After detachment from the flask wall by incubation at 370C in saline, the cells were disrupted in hypotonic buffer with a Dounce homogenizer. The pelleted membranes were resuspended in 50 mM Tris HCl (pH 7.4) to a protein concentration of 1 mg/ml and were stored at -800C. Protein was determined according to Bradford (16), using bovine serum albumin as standard. Alkaline Treatment and Membrane Fusion. For the inactivation of G proteins, the membrane suspension was diluted Abbreviations: NTX, naltrexone; DHM, dihydromorphine; G protein, guanine nucleotide binding protein; GTP[yS], guanosine 5'-[ythioltriphosphate; SWSD, sum of weighted squared deviations. *To whom reprint requests should be addressed at the Department of Biological Chemistry.
2172
Proc. Natl. Acad. Sci. USA 88 (1991)
Physiology/Pharmacology: Remmers and Medzihradsky A
c c .-CD
CY)
LLJ
E-
cr
CD m
CD
400 200 BOUND (fmol/mg protein) FIG. 1. Opioid agonist and antagonist binding in control and alkali-treated membranes. Scatchard plots of equilibrium binding of [3H]DHM (A) and [3H]NTX (B), carried out in control synaptosomal membranes (o and *), and pH 11.35-treated membranes (o and a) in the absence (o and o) and presence (o and *) of 5 gm GTP[yS]. (Inset) Dependence of the specific binding of 1.4 nM [3H]DHM (o) and 0.6 nM [3H]NTX (A) in the absence of GTP[yS] on the pH of membrane pretreatment. In all the experiments represented in this figure, binding of radiolabeled DHM and NTX was carried out in the absence and presence of 150 mM NaCl, respectively. Results from four experiments are shown. The mean values of the binding parameters and their statistical assessment are listed in Table 1.
Table 1. Parameters of opioid ligand binding Experimental conditions Membrane treatment Ligand KdI Control 0.69 [3H]DHM ND +GTP[yS] ND pH 11.35 ND +GTP[yS] ND pH 11.75 0.72 +fusion* 0.34 Control [3H]NTX 0.40 +GTP[yS] 0.57 11.35 pH
1:20 with 50 mM sodium phosphate adjusted to pH values ranging from 7.4 to 12 (17). After incubation at 4°C for 1 hr, the membranes were sedimented at 20,000 x g and resuspended in 50 mM Tris HCl (pH 7.4). Fusion of G proteindeficient synaptosomal membranes with untreated or alkalitreated C6 glioma cell membranes in the presence of PEG was performed in principle as described (18, 19). However, it was observed that exposure ofthe membranes to PEG prior to the alkali treatment enhanced the inactivation of G proteins. Therefore, brain membranes were initially treated with both PEG and high pH buffer and then mixed with C6 glioma cell membranes at an approximate 1:1 ratio (based on membrane protein). After centrifugation at 20,000 x g for 15 min, the pellet was resuspended at 7 mg of protein per ml in a buffer medium (pH 7.4) containing 50%o (wt/wt) PEG, 5 mM glucose, 2 mM CaC12, 2 mM MgCl2, 1 mM ATP, 0.1 mM EDTA, 135 mM NaCl, 5 mM KCl, and 20 mM Tris-HCl. The suspension was incubated for 5 min at 37°C, then 20 vol of 50 mM Tris buffer (pH 7.4) was layered on top and the tubes were centrifuged at 145,000 x g for 30 min. The pellet was resuspended in 20 vol of the 50 mM Tris buffer, centrifuged at 20,000 x g for 15 min, and the resulting pellet was suspended in the buffer at appropriate protein concentrations. Opioid Receptor Binding. Equilibrium ligand binding was carried out at 250C as described (5). Specific binding was defined as the difference between ligand bound in the absence and presence of an excess of the ju& selective agonist levorphanol (binding of [3H]DHM) or NTX (binding of [3H]NTX). Using levorphanol and dextrorphan, respectively, it was ascertained that specific and stereospecific ligand binding were identical. The incubation times and concentrations of unlabeled ligands were 65 min and 1 AM levorphanol for [3H]DHM, and 30 min and 1 A&M NTX for [3H]NTX. GTPvyS] Binding. In principle, a previously published procedure was followed (20, 21). In addition to isolated membranes, the assay medium contained 25 mM Tris-HCl (pH 7.4), 1 mM dithiothreitol, and GTP[y35S] to yield a final concentration of 0.01-10 uM. Nonspecific binding was determined with 500 ,M GTP[yS]. Nucleotide binding was linear between 10 and 100 ,ug of membrane protein. After incubation for 50 min at 25°C to reach binding equilibrium, the samples were filtered through nitrocellulose disks and
Kd2 8.71 3.64 7.60 13.60 10.53 ND 5.49 6.06 2.00
Binding parameters Bmax1 BmaX2 151 ND ND ND ND 58 202 270 245
205 225 230 264 117 ND 160 172 161
SWSD 38.38 20.17 41.08 33.40 120.63 13.48 20.73 23.% 45.50
n
28 25 28 25 27 39 54 39 42
+GTP[yS] Equilibrium binding of [3H]DHM (in the absence of 150 mM NaCl) and [3H]NTX (in the presence of 150 mM NaCl) in the absence and presence of 5 AM GTP[yS], respectively, was carried out in synaptosomal brain membranes. The experimental conditions are described in the legends to Figs. 1 and 4. The data from three or four experiments were pooled and subjected to nonlinear least-squares analysis. Kd,, Bm.1 and Kd2, B 11,2 reflect high- and low-affinity binding, respectively. SWSD is a
criterion of the variance ratio (or F test). n, Number of observations. ND, not detectable. Kd values are expressed as nM; Bn values are expressed as fmol per mg of protein. *The results in fusates of synaptosomal and C6 glioma cell membranes are expressed per mg of brain membrane protein. Because of high nonspecific binding, the concentration of [3H]DHM in the experiments with fused membranes was limited to 3 nM.
Physiology/Pharmacology: Remmers and Medzihradsky subjected to liquid scintillation counting. Standards of GTP[y35S] were dried on filters and included along with the samples. GTPase Activity. The assay of low Km GTPase activity in brain membranes has been described (22) and is based on the hydrolysis of [y-32P]GTP in the presence of Mg2' and an ATP-regenerating system (23). The following controls were routinely included in each assay: basal GTPase activity in the absence of opioids, GTP hydrolysis in the absence of membranes, and GTP hydrolysis in the assay medium prior to sample incubation. Maximal stimulation of basal enzyme activity and the ligand concentration producing half-maximal stimulation were expressed as S,,. and Ks, respectively. Data Analysis. The results on ligand binding were evaluated by using the computer program NONLIN, a weighted nonlinear least-squares regression analysis (5, 24). The data were weighted according to the inverse of the observed experimental values and were fitted to one- or two-site models. Initial estimates for two-site binding parameters were obtained from the respective double reciprocal plots, while one-site parameter estimates were obtained from nonlinear least-squares analysis using the computer program GRAPH PAD. The statistical variability of all the data was assessed either by SEM or by the sum of weighted squared deviations (SWSD). The latter is a parameter of the nonlinear regression analysis and, as a criterion of the F test, provides a statistical measure of variance. The relationship between SEM and SWSD is given by the equation SEM = SWSD/n, where n is the number of observations.
Proc. Natl. Acad. Sci. USA 88 (1991)
2173
To reconstitute high-affinity agonist binding, the alkalitreated synaptosomal membranes were fused with membranes from C6 glioma cells as a source of G proteins. The latter cells do not express opioid receptors (26), but they contain the inhibitory G proteins, Gi and Go (27). In the course of the fusion experiments, we have observed that PEG added to alkali-treated brain membranes in the absence of C6 glioma membranes restored some of the guanine nucleotidesensitive high-affinity [3H]DHM binding (Fig. 3). This effect apparently reflected mobilization by PEG of G protein from membrane compartments not susceptible to high pH. Thus, to increase the efficiency of fusion, PEG was added to membranes before alkali treatment. In the fused membranes, high-affinity [3H]DHM binding was restored, whereas [3H]NTX binding remained unaffected (Fig. 3). Equilibrium ligand binding revealed a population of high-affinity [3H]DHM binding sites with a Kd of 0.72 nM
A -q
x
z A: -r-
z
v
0 04 0
RESULTS The specific binding of both [3H]DHM and [3H]NTX at equilibrium revealed two saturable binding components. While abolishing high-affinity binding of the agonist, GTP[yS] did not significantly alter the affinity of [3H]NTX binding (Fig. 1 and Table 1). This effect of GTP[yS] suggested that agonist binds to the opioid receptor-G protein complex with higher affinity than to the uncoupled form of the receptor (e.g., see refs. 3 and 4). To directly assess the role of G protein in the expression of the two binding components, its inactivation in membranes by alkali treatment was carried out. Incubation of synaptosomal membranes at a pH up to 11.0 for 1 hr at 40C did not affect the specific binding of radiolabeled DHM or NTX. Above pH 11.0, high-affinity DHM binding decreased, while NTX binding was unaltered until approximately pH 11.4 (Fig. 1 Inset). Therefore, alkali treatment of the membranes at pH 11.35 was initially selected to resolve agonist and antagonist binding. However, in subsequent experiments, the membranes were exposed to pH 11.75 for a more complete inhibition of functional G proteins, necessary for the efficient reconstitution of high-affinity agonist binding by membrane-membrane fusion. In control synaptosomal membranes GTP[y_35S] bound to two populations of sites with Kd values of 0.03 and 1.00 ,uM, and Bma values of 233 and 478 pmol per mg of protein, respectively. In these membranes, the opioid agonist etorphine stimulated low Km GTPase activity with a K, of 40 nM. As reported earlier (25), the agonist-stimulated GTPase was inhibited by NTX (Fig. 2B). After alkali treatment of the membranes, the number of nucleotide binding sites (Fig. 2A) as well as basal (Fig. 2 Inset) and opioid-stimulated low Km GTPase activity (Fig. 2B) decreased by -95%. The effect of the treatment on GTP[y-35S] binding was to lower the Bmax at constant Kd values of the two binding components. Nonlinear least-squares analysis of ligand binding in the alkali-treated membranes revealed the absence of the high-affinity [3H]DHM component, whereas [3H]NTX affinity was unaltered. Moreover, residual [3H]DHM binding had lost its sensitivity to modulation by GTP[yS] (Table 1).
-
(pmol/mg)
GTP-T-S BOUND
sC
z -H
-H 0, 04
C,
af)
rz
E-
A
B
C
D
E
F
FIG. 2. GTP[yS] binding and GTPase activity in control and alkali-treated membranes. Equilibrium binding of GTP[y-35S] (A) in control (o) and alkali-treated (a) brain membranes was determined. Results of a representative experiment repeated at least four times are shown. (Inset) Basal, low Km GTPase activity in control and alkali-treated brain membranes (bars A and B) and C6 glioma cell membranes (bars C and D). Opioid agonist-stimulated (10 /AM etorphine) low Km GTPase activity (B) was determined in untreated brain membranes in the absence (bar A) and presence (bar B) of 10 tIM NTX, in brain membranes exposed to pH 11.35 (bar C), pH 11.75 (bar D), pH 11.75 followed by PEG-induced fusion with membranes from C6 glioma cells (bar E), and inhibition of the enzyme activity depicted in bar E by 10 ,uM NTX (bar F). Means + SEM of six assays, each run in triplicate, are shown.
2174
Physiology/Pharmacology: Remmers and Medzihradsky -E
X 200 -o
0
E
100
C] C: CD
p v
F_
(.I
-vfl]
--
A
B
C
A
B C
D
A B C
+PEG
-PEG
A B C D
+PEG
-PEG
FIG. 3. Opioid ligand binding in brain membranes and in C6 glioma cell membranes before and after exposure to high pH and PEG. The initial treatment of brain membranes at pH 11.75 was carried out. Specific binding of 1.4 nM [3H]DHM in the absence of 150 mM NaCl (Left) and 0.6 nM [3H]NTX in the presence of 150 mM NaCl (Right) was determined in C6 glioma cell membranes (bars A), pH 11.75 pretreated brain membranes (bars B), a mixture of both (bars C), and in a mixture of untreated brain membranes and pH 11.75-treated C6 glioma cell membranes (bars D). Ligand binding was carried out after incubation in either the absence (-PEG) or presence (+PEG) of 50%o PEG. The results obtained with membrane mixtures or fusates (bars C and D) are expressed on the basis of brain membrane protein. Means + SEM of a representative experiment performed in triplicate and repeated two to four times are shown.
(Fig. 4), a value remarkably similar to the Kd in unaltered synaptosomal membranes (Table 1). The reconstituted agonist binding was strongly modulated by GTP[yS] (Fig. 4). The fusion also restored opioid-stimulated low Km GTPase activity (Sma, 21%; Ks, 10 nM) that was inhibited by NTX (Fig. 2B). Furthermore, high pH treatment of C6 glioma cell membranes virtually eliminated low Km GTPase activity (Fig. 2 Inset), and the fusion of these membranes with untreated brain membranes failed to restore high-affinity [3H]DHM binding (Fig. 3) and opioid-sensitive GTPase activity.
DISCUSSION were characterized by biphasic [3H]DHM binding, opioid-stimulated low Km GTPase activity, and well-expressed binding of GTP[ty35S]. In view
The untreated brain membranes
R
c i c
CL)
A
6 io-A
0
0M
zk
0,
E
10
4
0
AA AA
E--4
LU
AA
A
2)o
A A
LLJ
a: CD cn
0
AsA IoI J
v v-
0o
20 40 BOUND (frmol/mg protein)
60
FIG. 4. [3H]DHM binding in alkali-treated brain membranes before and after their fusion with C6 glioma cell membranes. Equilibrium binding of [3H]DHM in pH 11.75-treated synaptosomal membranes before (n) and after (A) PEG-induced fusion with C6 glioma cell membranes was carried out. Also shown is the effect of 5 AM GTP[yS] on ligand binding in the pH 11.75-treated membranes before (m) and after (A) fusion. The results in the fusates are expressed on the basis of brain membrane protein. Results of a representative experiment repeated five times with similar results are shown. The corresponding binding parameters and data variability are listed in Table 1.
Proc. Natl. Acad. Sci. USA 88 (1991) of previous studies, it is of interest that we have observed two components of nucleotide binding. Although biphasic dissociation of [3H]GTP has been described in brain membranes (28), and the existence of two binding sites for GTP[yS] on the regulating component of adenylate cyclase from liver was implied (20), biphasic equilibrium binding of guanine nucleotides was not demonstrated. The affinity of the single binding compound (Kd 0.7 tzM) described in the studies cited above correlates well with that of the low-affinity site (Kd 1 ,uM) observed in our experiments. It is plausible to assume that in brain membranes the heterogeneity of guanine nucleotide binding reflects its interaction with multiple G proteins. The GTPase activity that was stimulated by the opioid agonist etorphine was inhibited by NTX. In brain membranes exposed to high pH, the strong inhibition of low Km GTPase and GTP[t-35S] binding was accompanied by the loss of the high-affinity [3H]DHM binding component. The profile of agonist binding after alkaline treatment was remarkably similar to that obtained in control membranes in the presence of GTP[yS], indicating receptor-G protein uncoupling as the underlying common mechanism. Since in both cases the loss of high-affinity binding was not fully accounted for by the increase in low-affinity sites, it is possible that we have failed to detect a very-lowaffinity form of the uncoupled receptor. High nonspecific binding of [3H]DHM limited its use to concentrations below 20 nM. The optimal pH for G protein inactivation was selected on the basis of differential sensitivity toward alkali treatment of the opioid receptor in its agonist and antagonist conformation. Despite elimination of 90%o of GTP[y_-35S] binding sites at pH 11.35, the membranes retained some agoniststimulated GTPase and GTP-sensitive ligand binding. Therefore, to increase the efficiency of G protein inactivation, and subsequent fusion, the pH step was also carried out at pH 11.75, albeit with some loss in receptor sites. Fusion of the alkali-treated synaptosomal membranes with C6 glioma cell membranes restored the high-affinity, GTPsensitive binding component of [3HJDHM. Alkali treatment of the glioma cell membranes before fusion prevented the reconstitution of agonist binding, thus accentuating the role of G protein in this process. The virtually identical ligand binding affinity in control and fused membranes suggests either that the ,u opioid receptor is capable of unperturbed function in the lipid environment of C6 membranes or that it maintains the composition of its original lipid boundary layer. The latter assumption is supported by the unaltered characteristics of [3H]NTX binding in the fused membranes. We have recently described the strong influence of membrane lipids on ligand binding to opioid receptors (29) and receptor coupling to low Km GTPase (30). Earlier it was shown that, after inactivation of opioid-sensitive adenylate cyclase in NG108-15 cells by N-ethylmaleimide, enzyme activity was restored by fusion of the hybrid cells and C6 glioma cells. Ligand binding, however, was not characterized (31). It is significant that the reconstitution of opioid binding described in this study also involved the opioid-sensitive GTPase activity, a functional consequence of receptor occupancy (22, 25). Low Km GTPase, intrinsic to inhibitory G protein (32), was virtually absent after alkali treatment of either synaptosomal or glioma cell membranes. In analogy to its characteristics in untreated brain membranes, the agoniststimulated enzyme activity restored in the fusates was inhibited by the antagonist NTX. Similar to opioid binding, the reconstitution of opioid-sensitive GTPase was unsuccessful in fusates of brain membranes and alkali-treated glioma cell membranes.
Physiology/Pharmacology: Remmers and Medzihradsky An essential finding of the reconstitution described here of high-affinity agonist binding was the absence of corresponding changes in antagonist binding during either the alkaline treatment or the subsequent fusion. In agreement with lacking regulation by GTP (4), the binding of [3H]NTX was not influenced by the elimination or readdition of G protein. Therefore, in contrast to DHM, the two binding components of NTX cannot be identified as different association states of receptor and G protein. Indeed, recent experiments on the inhibition of [3H]NTX binding by u,- and 6-selective opioids have shown that the biphasic binding of this antagonist reflects interactions with both u (high-affinity component) and 6 (low-affinity component) opioid receptors (33). These interpretations are in line with the limited binding selectivity of NTX in brain (33, 34). Whereas previous studies on the effect of guanine nucleotides on opioid receptor binding have suggested a role for G protein in ligand-receptor interaction (e.g., see ref. 35), the results of this study provide direct evidence for the ternary complex of ligand, receptor, and G protein as the molecular basis for high-affinity opioid agonist binding. The key support for this conclusion was provided by the strict dependence of this binding component and of functional receptor coupling (GTPase) on the membrane content of G protein. The methodological approaches described in this paper, in conjunction with specific membrane modifications described earlier (36), should prove useful to study the actual process of collision coupling that leads to the formation of receptor-effector complexes in the membrane environment. In addition, these techniques should be valuable to assess the characteristics of purified opioid receptor in neural cell membranes after its incorporation from liposomes of specific composition by membrane-membrane fusion. The transfer of G protein (and presumably of opioid receptor) with an intact lipid boundary layer has the advantage of retaining high-affinity opioid binding in the subnanomolar range (Table 1), unlike the reconstitution of purified receptor and G protein in lipid vesicles (37). We are grateful to Dr. G. L. Nordby for his advice in the statistical treatment of the data, and to Dr. W. R. Mancini for the supply of C6 glioma cells. We thank Ms. R. McLaughlin for her excellent secretarial assistance in preparing the manuscript. This work was supported in part by Grant DA 04087 from the United States Public Health Service. 1. Goldstein, A. L., Lowrey, L. I. & Pal, B. K. (1971) Proc. Nati. Acad. Sci. USA 68, 1742-1747. 2. Pasternak, G. W. & Snyder, S. H. (1975) Nature (London) 253, 563-565. 3. Blume, A. J. (1978) Proc. Nati. Acad. Sci. USA 75, 1713-1717. 4. Childers, S. R. & Snyder, S. H. (1980) J. Neurochem. 34, 583-593. 5. Fischel, S. V. & Medzihradsky, F. (1981) Mol. Pharmacol. 20, 269-279. 6. James, I. F. & Goldstein, A. (1984) Mol. Pharmacol. 25, 337-342.
Proc. Natl. Acad. Sci. USA 88 (1991)
2175
7. Goldstein, A. & James, I. F. (1984) Mol. Pharmacol. 25, 343-348. 8. Clark, M. J., Carter, B. D. & Medzihradsky, F. (1988) Eur. J. Pharmacol. 148, 343-351. 9. Chang, K.-J., Blanchard, S. G. & Cuatrecasas, P. (1983) Proc. Natl. Acad. Sci. USA 80, 940-944. 10. Hsia, J. A., Moss, J., Hewlett, E. L. & Vaughan, M. (1984) J. Biol. Chem. 259, 1086-1090. 11. Abood, M. E., Lee, N. M. & Loh, H. H. (1987) Brain Res. 417, 70-74. 12. Koski, G. & Klee, W. A. (1981) Proc. Natl. Acad. Sci. USA 78, 4185-4189. 13. Barchfeld, C. C. & Medzihradsky, F. (1984) Biochem. Biophys. Res. Commun. 121, 641-648. 14. Remmers, A. E. & Medzihradsky, F. -(1990) Trans. Am. Soc. Neurochem. 21, 270 (abstr.). 15. Cahill, A. L. & Medzihradsky, F. (1976) Biochem. Pharmacol. 25, 2257-2264. 16. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254. 17. Kim, M. H. & Neubig, R. R. (1985) FEBS Lett. 192, 321-325. 18. Schramm, M. (1979) Proc. Natl. Acad. Sci. USA 76, 11741178. 19. Kim, M. H. & Neubig, R. R. (1987) Biochemistry 26, 36643672. 20. Northrup, J. K., Smigel, M. D. & Gilman, A. G. (1982) J. Biol. Chem. 257, 11416-11423. 21. Sternweis, P. C. & Robishaw, J. D. (1984) J. Biol. Chem. 259, 13806-13813. 22. Clark, M. J., Nordby, G. L. & Medzihradsky, F. (1989) J. Neurochem. 52, 1162-1169. 23. Cassel, J. & Selinger, Z. (1976) Biochim. Biophys. Acta 452, 538-551. 24. Fischel, S. V. & Medzihradsky, F. (1986) Neuropharmacology 25, 351-359. 25. Clark, M. J. & Medzihradsky, F. (1987) Neuropharmacology 26, 1763-1770. 26. Klee, W. A. & Nirenberg, M. (1974) Proc. Natl. Acad. Sci. USA 71, 3474-3477. 27. Itoh, H., Kozasa, T., Nagata, S., Nakamura, S., Katada, T., Ui, M., Iwai, S., Ohtsuka, E., Kawasaki, H., Suzuki, K. & Kaziro, Y. (1986) Proc. Natl. Acad. Sci. USA 83, 3776-3780. 28. Childers, S. R. & Snyder, S. H. (1980) J. Neurochem. 35, 183-192. 29. Remmers, A. E., Nordby, G. L. & Medzihradsky, F. (1990) J. Neurochem. 55, 1993-2000. 30. Lazar, D. & Medzihradsky, F. (1990) Prog. Clin. Biol. Res. 328, 113-116. 31. Tocque, B., Pfeiffer, A. & Klee, A. (1987) FEBS Lett. 222, 327-331. 32. Milligan, G. & Klee, W. A. (1985) J. Biol. Chem. 260, 20572063. 33. Remmers, A. E. & Medzihradsky, F. (1991) J. Neurochem., in press. 34. Gillan, M. G. C., Kosterlitz, H. W. & Paterson, S. J. (1980) Br. J. Pharmacol. 70, 481-490. 35. Werling, L. L., Puttfarcken, P. S. & Cox, B. M. (1988) Mol. Pharmacol. 33, 423-431. 36. Medzihradsky, F. (1989) Adv. Biosci. 75, 41-44. 37. Ueda, H., Harada, H., Nozaki, M., Katada, T., Ui, M., Satoh, M. & Takagi, H. (1988) Proc. Natl. Acad. Sci. USA 85, 7013-7017.
Physiology/Pharmacology
Reconstitution of high-affinity opioid agonist binding in brain membranes (ju opioid receptor/receptor-G protein coupling/membrane-membrane fusion/C6 glioma cells/brain synaptosomes)
ANN E. REMMERS AND FEDOR MEDZIHRADSKY* Departments of Biological Chemistry and Pharmacology, The University of Michigan Medical School, Ann Arbor, MI 48109
Communicated by Avram Goldstein, December 3, 1990 (received for review October 27, 1990)
ABSTRACT In synaptosomal membranes from rat brain cortex, the 1L selective agonist [3H]dihydromorphine in the absence of sodium, and the nonselective antagonist [3Hlnaltrexone in the presence of sodium, bound to two populations of opioid receptor sites with Kd values of 0.69 and 8.7 nM for dihydromorphine, and 0.34 and 5.5 nM for naltrexone. The addition of 5 ,uM guanosine 5'-[y-thioltriphosphate (GTP[yS]) strongly reduced high-affinity agonist but not antagonist binding. Exposure of the membranes to high pH reduced the number of GTP[y-rS] binding sites by 90% and low K., opioid-sensitive GTPase activity by 95%. In these membranes, high-affinity agonist binding was abolished and modulation of residual binding by GTP[yS] was diminished. High-affinity (Kd, 0.72 nM), guanine nucleotide-sensitive agonist binding was reconstituted by polyethylene glycol-induced fusion of the alkali-treated membranes with (opioid receptor devoid) C6 glioma cell membranes. Also restored was opioid agoniststimulated, naltrexone-inhibited GTPase activity. In contrast, antagonist binding in the fused membranes was unaltered. Alkali treatment of the glioma cell membranes prior to fusion inhibited most of the low Km GTPase activity and prevented the reconstitution of agonist binding. The results show that highaffmiity opioid agonist binding reflects the ligand-occupied receptor-guanine nucleotide binding protein complex.
In addition to stereospecificity, a functional characteristic of all opioid alkaloids (1), these compounds show biphasic receptor binding at equilibrium. In rodent brain membranes, the high- and low-affinity binding components of morphine typically exhibit affinities of approximately 0.3 and 3.0 nM, respectively (2-4). Similar binding properties were described for the opioid antagonists [3H]naloxone (2, 4) and [3H]naltrexone (NTX) (5). Initially, it was shown that such heterogeneity, in some instances, reflected either nonspecific tissue association of the opioid or was an artifact of the experimental conditions (5). An alternative explanation for the biphasic Scatchard plots obtained was the marginal selectivity of the ligand used for the multiple types of opioid receptors present in neural tissue (6-8). On the other hand, guanine nucleotides, in particular GTP and its nonhydrolyzable analogues in the presence of sodium, reduced radiolabeled dihydromorphine (DHM) but not naloxone binding at equilibrium (4), thereby abolishing the high-affinity agonist component (3). Guanine nucleotides also suppressed the binding of both ,and 8-selective opioid peptide agonists (9). A role for guanine nucleotide binding proteins (G proteins) in the interaction of opioid ligands with their receptors was further supported by the effects of pertussis toxin in membranes from NG108-15 neuroblastoma x glioma hybrid cells and rat brain: exposure to the toxin diminished the affinity and GTP sensitivity of opioid agonist binding (10, 11). The functional significance of The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. 2171
the modulation by GTP of ligand binding was demonstrated by the opioid agonist-induced stimulation of low Km GTPase, reflecting the mechanism of receptor-effector coupling (12, 13). Although the results outlined above have implicated modulationt of ligand-receptor interaction by G proteins, they did not establish the molecular identity of the high- and lowaffinity components of opioid binding. In this study, we have directly linked high-affinity opioid agonist binding and opioid-sensitive GTPase activity to membrane G protein. Inactivation of G protein at high pH abolished high-affinity agonist, but not antagonist, binding. Guanine nucleotidesensitive high-affinity opioid binding and opioid-sensitive GTPase activity were restored after fusion of the alkalitreated synaptosomal brain membranes with membranes from opioid receptor-deficient, but inhibitory G proteincontaining, C6 glioma cells. The results describe high-affinity opioid agonist binding as a ternary complex composed of ligand, receptor, and G protein. Guanine nucleotidedependent dissociation of the complex then yields the lowaffinity form of the ligand-occupied receptor. A preliminary report on these findings was presented recently (14).
MATERIALS AND METHODS Materials. [3H]NTX was provided by the National Institute on Drug Abuse. [3H]DHM and [Y-32PIGTP were from Amersham, guanosine 5'-[y-[355]thiojtriphosphate (GTP[y-35S]) was from New England Nuclear, GTP[yS] tetralithium salt was from Boehringer Mannheim, and PEG with an average molecular weight of 8000 was from Sigma. The unlabeled opioids were obtained through the Narcotic Drug and Opiate Peptide Basic Research Center at the University ofMichigan. Isolated Membranes and Cell Culture. The synaptosomal preparation was isolated from brain cortices of male SpragueDawley rats as described (15). Prior to their freezing at -80'C, the synaptosomes were disrupted in a Dounce homogenizer. To reach confluence, the C6 glioma cells were grown as a monolayer in a minimal essential medium supplemented with 10% fetal bovine serum and 1% pyruvate for 8-9 days at 370C in 95% air/5% CO2. After detachment from the flask wall by incubation at 370C in saline, the cells were disrupted in hypotonic buffer with a Dounce homogenizer. The pelleted membranes were resuspended in 50 mM Tris HCl (pH 7.4) to a protein concentration of 1 mg/ml and were stored at -800C. Protein was determined according to Bradford (16), using bovine serum albumin as standard. Alkaline Treatment and Membrane Fusion. For the inactivation of G proteins, the membrane suspension was diluted Abbreviations: NTX, naltrexone; DHM, dihydromorphine; G protein, guanine nucleotide binding protein; GTP[yS], guanosine 5'-[ythioltriphosphate; SWSD, sum of weighted squared deviations. *To whom reprint requests should be addressed at the Department of Biological Chemistry.
2172
Proc. Natl. Acad. Sci. USA 88 (1991)
Physiology/Pharmacology: Remmers and Medzihradsky A
c c .-CD
CY)
LLJ
E-
cr
CD m
CD
400 200 BOUND (fmol/mg protein) FIG. 1. Opioid agonist and antagonist binding in control and alkali-treated membranes. Scatchard plots of equilibrium binding of [3H]DHM (A) and [3H]NTX (B), carried out in control synaptosomal membranes (o and *), and pH 11.35-treated membranes (o and a) in the absence (o and o) and presence (o and *) of 5 gm GTP[yS]. (Inset) Dependence of the specific binding of 1.4 nM [3H]DHM (o) and 0.6 nM [3H]NTX (A) in the absence of GTP[yS] on the pH of membrane pretreatment. In all the experiments represented in this figure, binding of radiolabeled DHM and NTX was carried out in the absence and presence of 150 mM NaCl, respectively. Results from four experiments are shown. The mean values of the binding parameters and their statistical assessment are listed in Table 1.
Table 1. Parameters of opioid ligand binding Experimental conditions Membrane treatment Ligand KdI Control 0.69 [3H]DHM ND +GTP[yS] ND pH 11.35 ND +GTP[yS] ND pH 11.75 0.72 +fusion* 0.34 Control [3H]NTX 0.40 +GTP[yS] 0.57 11.35 pH
1:20 with 50 mM sodium phosphate adjusted to pH values ranging from 7.4 to 12 (17). After incubation at 4°C for 1 hr, the membranes were sedimented at 20,000 x g and resuspended in 50 mM Tris HCl (pH 7.4). Fusion of G proteindeficient synaptosomal membranes with untreated or alkalitreated C6 glioma cell membranes in the presence of PEG was performed in principle as described (18, 19). However, it was observed that exposure ofthe membranes to PEG prior to the alkali treatment enhanced the inactivation of G proteins. Therefore, brain membranes were initially treated with both PEG and high pH buffer and then mixed with C6 glioma cell membranes at an approximate 1:1 ratio (based on membrane protein). After centrifugation at 20,000 x g for 15 min, the pellet was resuspended at 7 mg of protein per ml in a buffer medium (pH 7.4) containing 50%o (wt/wt) PEG, 5 mM glucose, 2 mM CaC12, 2 mM MgCl2, 1 mM ATP, 0.1 mM EDTA, 135 mM NaCl, 5 mM KCl, and 20 mM Tris-HCl. The suspension was incubated for 5 min at 37°C, then 20 vol of 50 mM Tris buffer (pH 7.4) was layered on top and the tubes were centrifuged at 145,000 x g for 30 min. The pellet was resuspended in 20 vol of the 50 mM Tris buffer, centrifuged at 20,000 x g for 15 min, and the resulting pellet was suspended in the buffer at appropriate protein concentrations. Opioid Receptor Binding. Equilibrium ligand binding was carried out at 250C as described (5). Specific binding was defined as the difference between ligand bound in the absence and presence of an excess of the ju& selective agonist levorphanol (binding of [3H]DHM) or NTX (binding of [3H]NTX). Using levorphanol and dextrorphan, respectively, it was ascertained that specific and stereospecific ligand binding were identical. The incubation times and concentrations of unlabeled ligands were 65 min and 1 AM levorphanol for [3H]DHM, and 30 min and 1 A&M NTX for [3H]NTX. GTPvyS] Binding. In principle, a previously published procedure was followed (20, 21). In addition to isolated membranes, the assay medium contained 25 mM Tris-HCl (pH 7.4), 1 mM dithiothreitol, and GTP[y35S] to yield a final concentration of 0.01-10 uM. Nonspecific binding was determined with 500 ,M GTP[yS]. Nucleotide binding was linear between 10 and 100 ,ug of membrane protein. After incubation for 50 min at 25°C to reach binding equilibrium, the samples were filtered through nitrocellulose disks and
Kd2 8.71 3.64 7.60 13.60 10.53 ND 5.49 6.06 2.00
Binding parameters Bmax1 BmaX2 151 ND ND ND ND 58 202 270 245
205 225 230 264 117 ND 160 172 161
SWSD 38.38 20.17 41.08 33.40 120.63 13.48 20.73 23.% 45.50
n
28 25 28 25 27 39 54 39 42
+GTP[yS] Equilibrium binding of [3H]DHM (in the absence of 150 mM NaCl) and [3H]NTX (in the presence of 150 mM NaCl) in the absence and presence of 5 AM GTP[yS], respectively, was carried out in synaptosomal brain membranes. The experimental conditions are described in the legends to Figs. 1 and 4. The data from three or four experiments were pooled and subjected to nonlinear least-squares analysis. Kd,, Bm.1 and Kd2, B 11,2 reflect high- and low-affinity binding, respectively. SWSD is a
criterion of the variance ratio (or F test). n, Number of observations. ND, not detectable. Kd values are expressed as nM; Bn values are expressed as fmol per mg of protein. *The results in fusates of synaptosomal and C6 glioma cell membranes are expressed per mg of brain membrane protein. Because of high nonspecific binding, the concentration of [3H]DHM in the experiments with fused membranes was limited to 3 nM.
Physiology/Pharmacology: Remmers and Medzihradsky subjected to liquid scintillation counting. Standards of GTP[y35S] were dried on filters and included along with the samples. GTPase Activity. The assay of low Km GTPase activity in brain membranes has been described (22) and is based on the hydrolysis of [y-32P]GTP in the presence of Mg2' and an ATP-regenerating system (23). The following controls were routinely included in each assay: basal GTPase activity in the absence of opioids, GTP hydrolysis in the absence of membranes, and GTP hydrolysis in the assay medium prior to sample incubation. Maximal stimulation of basal enzyme activity and the ligand concentration producing half-maximal stimulation were expressed as S,,. and Ks, respectively. Data Analysis. The results on ligand binding were evaluated by using the computer program NONLIN, a weighted nonlinear least-squares regression analysis (5, 24). The data were weighted according to the inverse of the observed experimental values and were fitted to one- or two-site models. Initial estimates for two-site binding parameters were obtained from the respective double reciprocal plots, while one-site parameter estimates were obtained from nonlinear least-squares analysis using the computer program GRAPH PAD. The statistical variability of all the data was assessed either by SEM or by the sum of weighted squared deviations (SWSD). The latter is a parameter of the nonlinear regression analysis and, as a criterion of the F test, provides a statistical measure of variance. The relationship between SEM and SWSD is given by the equation SEM = SWSD/n, where n is the number of observations.
Proc. Natl. Acad. Sci. USA 88 (1991)
2173
To reconstitute high-affinity agonist binding, the alkalitreated synaptosomal membranes were fused with membranes from C6 glioma cells as a source of G proteins. The latter cells do not express opioid receptors (26), but they contain the inhibitory G proteins, Gi and Go (27). In the course of the fusion experiments, we have observed that PEG added to alkali-treated brain membranes in the absence of C6 glioma membranes restored some of the guanine nucleotidesensitive high-affinity [3H]DHM binding (Fig. 3). This effect apparently reflected mobilization by PEG of G protein from membrane compartments not susceptible to high pH. Thus, to increase the efficiency of fusion, PEG was added to membranes before alkali treatment. In the fused membranes, high-affinity [3H]DHM binding was restored, whereas [3H]NTX binding remained unaffected (Fig. 3). Equilibrium ligand binding revealed a population of high-affinity [3H]DHM binding sites with a Kd of 0.72 nM
A -q
x
z A: -r-
z
v
0 04 0
RESULTS The specific binding of both [3H]DHM and [3H]NTX at equilibrium revealed two saturable binding components. While abolishing high-affinity binding of the agonist, GTP[yS] did not significantly alter the affinity of [3H]NTX binding (Fig. 1 and Table 1). This effect of GTP[yS] suggested that agonist binds to the opioid receptor-G protein complex with higher affinity than to the uncoupled form of the receptor (e.g., see refs. 3 and 4). To directly assess the role of G protein in the expression of the two binding components, its inactivation in membranes by alkali treatment was carried out. Incubation of synaptosomal membranes at a pH up to 11.0 for 1 hr at 40C did not affect the specific binding of radiolabeled DHM or NTX. Above pH 11.0, high-affinity DHM binding decreased, while NTX binding was unaltered until approximately pH 11.4 (Fig. 1 Inset). Therefore, alkali treatment of the membranes at pH 11.35 was initially selected to resolve agonist and antagonist binding. However, in subsequent experiments, the membranes were exposed to pH 11.75 for a more complete inhibition of functional G proteins, necessary for the efficient reconstitution of high-affinity agonist binding by membrane-membrane fusion. In control synaptosomal membranes GTP[y_35S] bound to two populations of sites with Kd values of 0.03 and 1.00 ,uM, and Bma values of 233 and 478 pmol per mg of protein, respectively. In these membranes, the opioid agonist etorphine stimulated low Km GTPase activity with a K, of 40 nM. As reported earlier (25), the agonist-stimulated GTPase was inhibited by NTX (Fig. 2B). After alkali treatment of the membranes, the number of nucleotide binding sites (Fig. 2A) as well as basal (Fig. 2 Inset) and opioid-stimulated low Km GTPase activity (Fig. 2B) decreased by -95%. The effect of the treatment on GTP[y-35S] binding was to lower the Bmax at constant Kd values of the two binding components. Nonlinear least-squares analysis of ligand binding in the alkali-treated membranes revealed the absence of the high-affinity [3H]DHM component, whereas [3H]NTX affinity was unaltered. Moreover, residual [3H]DHM binding had lost its sensitivity to modulation by GTP[yS] (Table 1).
-
(pmol/mg)
GTP-T-S BOUND
sC
z -H
-H 0, 04
C,
af)
rz
E-
A
B
C
D
E
F
FIG. 2. GTP[yS] binding and GTPase activity in control and alkali-treated membranes. Equilibrium binding of GTP[y-35S] (A) in control (o) and alkali-treated (a) brain membranes was determined. Results of a representative experiment repeated at least four times are shown. (Inset) Basal, low Km GTPase activity in control and alkali-treated brain membranes (bars A and B) and C6 glioma cell membranes (bars C and D). Opioid agonist-stimulated (10 /AM etorphine) low Km GTPase activity (B) was determined in untreated brain membranes in the absence (bar A) and presence (bar B) of 10 tIM NTX, in brain membranes exposed to pH 11.35 (bar C), pH 11.75 (bar D), pH 11.75 followed by PEG-induced fusion with membranes from C6 glioma cells (bar E), and inhibition of the enzyme activity depicted in bar E by 10 ,uM NTX (bar F). Means + SEM of six assays, each run in triplicate, are shown.
2174
Physiology/Pharmacology: Remmers and Medzihradsky -E
X 200 -o
0
E
100
C] C: CD
p v
F_
(.I
-vfl]
--
A
B
C
A
B C
D
A B C
+PEG
-PEG
A B C D
+PEG
-PEG
FIG. 3. Opioid ligand binding in brain membranes and in C6 glioma cell membranes before and after exposure to high pH and PEG. The initial treatment of brain membranes at pH 11.75 was carried out. Specific binding of 1.4 nM [3H]DHM in the absence of 150 mM NaCl (Left) and 0.6 nM [3H]NTX in the presence of 150 mM NaCl (Right) was determined in C6 glioma cell membranes (bars A), pH 11.75 pretreated brain membranes (bars B), a mixture of both (bars C), and in a mixture of untreated brain membranes and pH 11.75-treated C6 glioma cell membranes (bars D). Ligand binding was carried out after incubation in either the absence (-PEG) or presence (+PEG) of 50%o PEG. The results obtained with membrane mixtures or fusates (bars C and D) are expressed on the basis of brain membrane protein. Means + SEM of a representative experiment performed in triplicate and repeated two to four times are shown.
(Fig. 4), a value remarkably similar to the Kd in unaltered synaptosomal membranes (Table 1). The reconstituted agonist binding was strongly modulated by GTP[yS] (Fig. 4). The fusion also restored opioid-stimulated low Km GTPase activity (Sma, 21%; Ks, 10 nM) that was inhibited by NTX (Fig. 2B). Furthermore, high pH treatment of C6 glioma cell membranes virtually eliminated low Km GTPase activity (Fig. 2 Inset), and the fusion of these membranes with untreated brain membranes failed to restore high-affinity [3H]DHM binding (Fig. 3) and opioid-sensitive GTPase activity.
DISCUSSION were characterized by biphasic [3H]DHM binding, opioid-stimulated low Km GTPase activity, and well-expressed binding of GTP[ty35S]. In view
The untreated brain membranes
R
c i c
CL)
A
6 io-A
0
0M
zk
0,
E
10
4
0
AA AA
E--4
LU
AA
A
2)o
A A
LLJ
a: CD cn
0
AsA IoI J
v v-
0o
20 40 BOUND (frmol/mg protein)
60
FIG. 4. [3H]DHM binding in alkali-treated brain membranes before and after their fusion with C6 glioma cell membranes. Equilibrium binding of [3H]DHM in pH 11.75-treated synaptosomal membranes before (n) and after (A) PEG-induced fusion with C6 glioma cell membranes was carried out. Also shown is the effect of 5 AM GTP[yS] on ligand binding in the pH 11.75-treated membranes before (m) and after (A) fusion. The results in the fusates are expressed on the basis of brain membrane protein. Results of a representative experiment repeated five times with similar results are shown. The corresponding binding parameters and data variability are listed in Table 1.
Proc. Natl. Acad. Sci. USA 88 (1991) of previous studies, it is of interest that we have observed two components of nucleotide binding. Although biphasic dissociation of [3H]GTP has been described in brain membranes (28), and the existence of two binding sites for GTP[yS] on the regulating component of adenylate cyclase from liver was implied (20), biphasic equilibrium binding of guanine nucleotides was not demonstrated. The affinity of the single binding compound (Kd 0.7 tzM) described in the studies cited above correlates well with that of the low-affinity site (Kd 1 ,uM) observed in our experiments. It is plausible to assume that in brain membranes the heterogeneity of guanine nucleotide binding reflects its interaction with multiple G proteins. The GTPase activity that was stimulated by the opioid agonist etorphine was inhibited by NTX. In brain membranes exposed to high pH, the strong inhibition of low Km GTPase and GTP[t-35S] binding was accompanied by the loss of the high-affinity [3H]DHM binding component. The profile of agonist binding after alkaline treatment was remarkably similar to that obtained in control membranes in the presence of GTP[yS], indicating receptor-G protein uncoupling as the underlying common mechanism. Since in both cases the loss of high-affinity binding was not fully accounted for by the increase in low-affinity sites, it is possible that we have failed to detect a very-lowaffinity form of the uncoupled receptor. High nonspecific binding of [3H]DHM limited its use to concentrations below 20 nM. The optimal pH for G protein inactivation was selected on the basis of differential sensitivity toward alkali treatment of the opioid receptor in its agonist and antagonist conformation. Despite elimination of 90%o of GTP[y_-35S] binding sites at pH 11.35, the membranes retained some agoniststimulated GTPase and GTP-sensitive ligand binding. Therefore, to increase the efficiency of G protein inactivation, and subsequent fusion, the pH step was also carried out at pH 11.75, albeit with some loss in receptor sites. Fusion of the alkali-treated synaptosomal membranes with C6 glioma cell membranes restored the high-affinity, GTPsensitive binding component of [3HJDHM. Alkali treatment of the glioma cell membranes before fusion prevented the reconstitution of agonist binding, thus accentuating the role of G protein in this process. The virtually identical ligand binding affinity in control and fused membranes suggests either that the ,u opioid receptor is capable of unperturbed function in the lipid environment of C6 membranes or that it maintains the composition of its original lipid boundary layer. The latter assumption is supported by the unaltered characteristics of [3H]NTX binding in the fused membranes. We have recently described the strong influence of membrane lipids on ligand binding to opioid receptors (29) and receptor coupling to low Km GTPase (30). Earlier it was shown that, after inactivation of opioid-sensitive adenylate cyclase in NG108-15 cells by N-ethylmaleimide, enzyme activity was restored by fusion of the hybrid cells and C6 glioma cells. Ligand binding, however, was not characterized (31). It is significant that the reconstitution of opioid binding described in this study also involved the opioid-sensitive GTPase activity, a functional consequence of receptor occupancy (22, 25). Low Km GTPase, intrinsic to inhibitory G protein (32), was virtually absent after alkali treatment of either synaptosomal or glioma cell membranes. In analogy to its characteristics in untreated brain membranes, the agoniststimulated enzyme activity restored in the fusates was inhibited by the antagonist NTX. Similar to opioid binding, the reconstitution of opioid-sensitive GTPase was unsuccessful in fusates of brain membranes and alkali-treated glioma cell membranes.
Physiology/Pharmacology: Remmers and Medzihradsky An essential finding of the reconstitution described here of high-affinity agonist binding was the absence of corresponding changes in antagonist binding during either the alkaline treatment or the subsequent fusion. In agreement with lacking regulation by GTP (4), the binding of [3H]NTX was not influenced by the elimination or readdition of G protein. Therefore, in contrast to DHM, the two binding components of NTX cannot be identified as different association states of receptor and G protein. Indeed, recent experiments on the inhibition of [3H]NTX binding by u,- and 6-selective opioids have shown that the biphasic binding of this antagonist reflects interactions with both u (high-affinity component) and 6 (low-affinity component) opioid receptors (33). These interpretations are in line with the limited binding selectivity of NTX in brain (33, 34). Whereas previous studies on the effect of guanine nucleotides on opioid receptor binding have suggested a role for G protein in ligand-receptor interaction (e.g., see ref. 35), the results of this study provide direct evidence for the ternary complex of ligand, receptor, and G protein as the molecular basis for high-affinity opioid agonist binding. The key support for this conclusion was provided by the strict dependence of this binding component and of functional receptor coupling (GTPase) on the membrane content of G protein. The methodological approaches described in this paper, in conjunction with specific membrane modifications described earlier (36), should prove useful to study the actual process of collision coupling that leads to the formation of receptor-effector complexes in the membrane environment. In addition, these techniques should be valuable to assess the characteristics of purified opioid receptor in neural cell membranes after its incorporation from liposomes of specific composition by membrane-membrane fusion. The transfer of G protein (and presumably of opioid receptor) with an intact lipid boundary layer has the advantage of retaining high-affinity opioid binding in the subnanomolar range (Table 1), unlike the reconstitution of purified receptor and G protein in lipid vesicles (37). We are grateful to Dr. G. L. Nordby for his advice in the statistical treatment of the data, and to Dr. W. R. Mancini for the supply of C6 glioma cells. We thank Ms. R. McLaughlin for her excellent secretarial assistance in preparing the manuscript. This work was supported in part by Grant DA 04087 from the United States Public Health Service. 1. Goldstein, A. L., Lowrey, L. I. & Pal, B. K. (1971) Proc. Nati. Acad. Sci. USA 68, 1742-1747. 2. Pasternak, G. W. & Snyder, S. H. (1975) Nature (London) 253, 563-565. 3. Blume, A. J. (1978) Proc. Nati. Acad. Sci. USA 75, 1713-1717. 4. Childers, S. R. & Snyder, S. H. (1980) J. Neurochem. 34, 583-593. 5. Fischel, S. V. & Medzihradsky, F. (1981) Mol. Pharmacol. 20, 269-279. 6. James, I. F. & Goldstein, A. (1984) Mol. Pharmacol. 25, 337-342.
Proc. Natl. Acad. Sci. USA 88 (1991)
2175
7. Goldstein, A. & James, I. F. (1984) Mol. Pharmacol. 25, 343-348. 8. Clark, M. J., Carter, B. D. & Medzihradsky, F. (1988) Eur. J. Pharmacol. 148, 343-351. 9. Chang, K.-J., Blanchard, S. G. & Cuatrecasas, P. (1983) Proc. Natl. Acad. Sci. USA 80, 940-944. 10. Hsia, J. A., Moss, J., Hewlett, E. L. & Vaughan, M. (1984) J. Biol. Chem. 259, 1086-1090. 11. Abood, M. E., Lee, N. M. & Loh, H. H. (1987) Brain Res. 417, 70-74. 12. Koski, G. & Klee, W. A. (1981) Proc. Natl. Acad. Sci. USA 78, 4185-4189. 13. Barchfeld, C. C. & Medzihradsky, F. (1984) Biochem. Biophys. Res. Commun. 121, 641-648. 14. Remmers, A. E. & Medzihradsky, F. -(1990) Trans. Am. Soc. Neurochem. 21, 270 (abstr.). 15. Cahill, A. L. & Medzihradsky, F. (1976) Biochem. Pharmacol. 25, 2257-2264. 16. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254. 17. Kim, M. H. & Neubig, R. R. (1985) FEBS Lett. 192, 321-325. 18. Schramm, M. (1979) Proc. Natl. Acad. Sci. USA 76, 11741178. 19. Kim, M. H. & Neubig, R. R. (1987) Biochemistry 26, 36643672. 20. Northrup, J. K., Smigel, M. D. & Gilman, A. G. (1982) J. Biol. Chem. 257, 11416-11423. 21. Sternweis, P. C. & Robishaw, J. D. (1984) J. Biol. Chem. 259, 13806-13813. 22. Clark, M. J., Nordby, G. L. & Medzihradsky, F. (1989) J. Neurochem. 52, 1162-1169. 23. Cassel, J. & Selinger, Z. (1976) Biochim. Biophys. Acta 452, 538-551. 24. Fischel, S. V. & Medzihradsky, F. (1986) Neuropharmacology 25, 351-359. 25. Clark, M. J. & Medzihradsky, F. (1987) Neuropharmacology 26, 1763-1770. 26. Klee, W. A. & Nirenberg, M. (1974) Proc. Natl. Acad. Sci. USA 71, 3474-3477. 27. Itoh, H., Kozasa, T., Nagata, S., Nakamura, S., Katada, T., Ui, M., Iwai, S., Ohtsuka, E., Kawasaki, H., Suzuki, K. & Kaziro, Y. (1986) Proc. Natl. Acad. Sci. USA 83, 3776-3780. 28. Childers, S. R. & Snyder, S. H. (1980) J. Neurochem. 35, 183-192. 29. Remmers, A. E., Nordby, G. L. & Medzihradsky, F. (1990) J. Neurochem. 55, 1993-2000. 30. Lazar, D. & Medzihradsky, F. (1990) Prog. Clin. Biol. Res. 328, 113-116. 31. Tocque, B., Pfeiffer, A. & Klee, A. (1987) FEBS Lett. 222, 327-331. 32. Milligan, G. & Klee, W. A. (1985) J. Biol. Chem. 260, 20572063. 33. Remmers, A. E. & Medzihradsky, F. (1991) J. Neurochem., in press. 34. Gillan, M. G. C., Kosterlitz, H. W. & Paterson, S. J. (1980) Br. J. Pharmacol. 70, 481-490. 35. Werling, L. L., Puttfarcken, P. S. & Cox, B. M. (1988) Mol. Pharmacol. 33, 423-431. 36. Medzihradsky, F. (1989) Adv. Biosci. 75, 41-44. 37. Ueda, H., Harada, H., Nozaki, M., Katada, T., Ui, M., Satoh, M. & Takagi, H. (1988) Proc. Natl. Acad. Sci. USA 85, 7013-7017.
