Introduction

Top Abstract Introduction Materials Results Discussion References

Opioid receptors play important roles in many physiological functions. The presence of multiple types of opioid receptors, at least mu, delta, kappa and epsilon, in the nervous system has been established by pharmacological and binding studies as well as differential anatomical localization in the brain (for a review, Chang, 1984). Activation of kappa opioid receptors produces many effects including analgesia (Von Voigtlander et al., 1983; Dykstra et al., 1987), dysphoria (Pfeiffer et al., 1986) and water diuresis (Leander, 1983a; Dykstra et al., 1987). Dynorphins are thought to be endogenous ligands for kappa receptors (Chavkin and Goldstein, 1981). ()-U50,488H is the prototype of selective kappa agonists (Von Voigtlander et al., 1983), whereas norbinaltorphimine is a selective kappa antagonist (Portoghese et al., 1987).

There is ample evidence supporting the notion that the kappa opioid receptor belongs to the superfamily of GPCRs. Activation of kappa opioid receptors leads to inhibition of adenylate cyclase (Attali et al., 1989; Konkoy and Childers, 1989, 1993; Lawrence and Bidlack, 1993; Lawrence et al., 1995; Prather et al., 1995), activation of low-K_m GTPase (Clark et al., 1986; Clark and Medzihradsky, 1987; Lawrence et al., 1995), enhancement of incorporation of [³²P]azidoanilido-GTP into G subunits (Prather et al., 1995), increase in inward-rectifying K⁺ conductance (Ma et al., 1995; Henry et al., 1995; Grudt and Williams, 1993) and decrease in Ca⁺⁺ conductance through N-type channels (Gross and MacDonald, 1987). Kappa agonist-induced inhibition of adenylate cyclase and increase in K⁺ conductance were shown to be mediated through pertussis toxin-sensitive G proteins (Lawrence and Bidlack, 1993; Prather et al., 1995; Ma et al., 1995). After the cloning of the mouse delta opioid receptor (Kieffer et al., 1992; Evans et al., 1992), we (Zhu et al., 1995) and Mansson et al. (1994) reported cloning of the human kappa opioid receptor. Hydropathy analysis of the amino acid sequence shows the presence of seven putative transmembrane domains, a structural motif common to all G protein-coupled receptors.

G proteins undergo activation/inactivation cycle (for reviews, Gilman, 1987; Birnbaumer et al., 1990). At resting state, the guanine nucleotide binding site of the alpha subunits of G proteins is occupied by GDP. When the GPCRs are activated by agonists, GDP bound to G is displaced by GTP. Binding by GTP causes dissociation of G proteins into alpha and beta/gamma subunits. GTP-bound alpha subunit and beta/gamma subunits activate downstream effectors. Hydrolysis of GTP bound to alpha subunits by a GTPase intrinsic to G to GDP terminates agonist effects. G-GDP is then reassociated with beta/gamma subunits to re-enter the cycle. When the poorly hydrolyzed GTP analog GTPS is used instead of GTP, the half-lives of GTPS-bound alpha subunits are prolonged and  subunits are persistently activated.

Agonist-stimulated binding of [³⁵S]GTPS to G proteins has been used as a sensitive assay of activation of many GPCRs. It was first used in a reconstitution system of purified G protein and purified beta adrenergic receptors (Asano et al., 1984). Since then, this method has been applied to many GPCRs in cell membranes, including muscarinic (Hilf et al., 1989; Lazareno et al., 1993), A1 adenosine (Lorenzen et al., 1993), alpha-2D adrenergic (Tian et al., 1994) and mu opioid (Traynor and Nahorski, 1995) receptors. It was recently used to localize G proteins activated by agonists for mu opioid, cannabinoid and -aminobutyric acid_B receptors in rat brain sections by in vitro autoradiography (Sim et al., 1995).

In this study, we characterized effects of activation of the hkor stably expressed in CHO cells (CHO-hkor cells) on [³⁵S]GTPS binding to membranes. In addition, we determined whether this assay could be used to determine efficacy and potency of various opioid ligands on the kappa receptor.

	Methods and Materials

Top Abstract Introduction Materials Results Discussion References

Materials. [³H]Diprenorphine (35 Ci/mmol) and [³⁵S]GTPS (1000-1300 Ci/mmol) were obtained from Amersham Corp. (Arlington Heights, IL) and NEN Life Sciences Co. (Boston, MA), respectively. Naloxone was a gift from DuPont/Merck (Wilmington, DE), and ()-U50,488H was provided by Upjohn (Kalamazoo, MI). Dynorphin, DPDPE and DAMGO were purchased from Peninsula Laboratories (Belmont, CA). GDP and GTPS were purchased from Sigma Chemical Co. (St. Louis, MO). (±)-Ethylketocyclazocine, tifluadom, nalorphine, nalbuphine, pentazocine, -funaltrexamine, norbinaltorphimine and buprenorphine were provided by the National Institute on Drug Abuse.

Stable expression of hkor in CHO cells. CHO cells were transfected with the hkor cDNA in the vector pBK-CMV (Zhu et al., 1995) and clonal cell lines stably expressing hkor (CHO-hkor) were established with Geneticin selection as described (Sambrook et al., 1989).

[³⁵S]GTPS binding. Determination of [³⁵S]GTPS binding to G proteins was carried out with a modified procedure of Traynor and Nahorski (1995).

11

5

Opioid receptor binding. Membranes were prepared from CHO-hkor cells as described previously (Zhu et al., 1996). Opioid receptor binding was conducted with [³H]diprenorphine according to our published procedure (Zhu et al., 1995). Binding was carried out in 50 mM Tris-HCl buffer containing 1 mM EGTA and 10 µM leupeptin (pH 7.4 at room temperature) (TEL buffer) or in [³⁵S]GTPS binding buffer at 25°C for 1 h in duplicate in a volume of 1 ml with 30 to 60 µg protein. Binding data were analyzed with the EBDA program (McPherson, 1983). K_i values were determined by use of the equation of Cheng and Prusoff (1973). Comparison of K_i values of each drug in the two buffers were performed with Student's t test, with P <.05 being considered significantly different.

Determination of protein content. Protein contents of membranes were determined by the BCA method of Smith et al. (1985) with bovine serum albumin as the standard.

	Results

Top Abstract Introduction Materials Results Discussion References

Stable expression of the human kappa opioid receptor cDNA in CHO cells. The human kappa opioid receptor cDNA was expressed stably in CHO cells. Saturation binding of [³H]diprenorphine to membranes of CHO-hkor cells was carried out in 50 mM TEL buffer. K_d and B_max were determined to be 0.12 ± 0.02 nM and 1292 ± 52 fmol/mg protein (mean ± S.E.M., n = 3), respectively. [³H]Diprenorphine binding to the membranes was potently inhibited by kappa selective ligands, but not by mu or delta ligands (see table 1).

View this table: [in this window] [in a new window]   TABLE 1 EC50 values and maximal effects of opioid ligands in stimulating [35S]GTPS binding to CHO-hkor cell membranes and their Ki values in inhibiting [3H]diprenorphine binding to the human kappa receptors expressed in CHO cells [35S]GTPS binding was performed in the presence of 3 µM GDP, 5 mM Mg++ and 100 mM Na+ with ~10 µg membrane proteins at 30°C for 60 min with various concentrations of each drug. The 100% maximal effect was designated as the level of binding in response to 10 µM U50,488H. Competitive inhibition of 0.3~0.5 nM [3H]diprenorphine binding by opioid ligands was performed in [35S]GTPS binding buffer as well as in TEL buffer at 25°C for 60 min as described under "Methods and Materials." Each value represents mean ± S.E.M. of three independent experiments in duplicate.

Effects of GDP on [³⁵S]GTPS binding. Experiments were performed with 5 mM Mg⁺⁺, 100 mM Na⁺ and various concentrations of GDP in the presence and absence of 3 µM ()-U50,488H (fig. 1). Without GDP, no increase in [³⁵S]GTPS binding by the kappa receptor agonist ()-U50,488H was observed. GDP decreased [³⁵S]GTPS binding in a dose-dependent manner both in the presence and absence of ()-U50,488H (fig. 1). With 0.3 µM GDP, the magnitude of reduction in [³⁵S]GTPS binding was larger in the absence than the presence of ()-U50,488H and, thus, the increase in [³⁵S]GTPS binding caused by the agonist could be observed. The presence of 3 µM GDP produced the maximal stimulatory response. At concentrations of GDP  100 µM, effects of ()-U50,488H were greatly diminished or obliterated. In this study, 3 µM GDP was used in all other experiments.

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Effect of GDP on [35S]GTPS binding to CHO-hkor membranes in the absence (open triangle) or presence (dark triangle) of 3 µM ()-U50,488H. [35S]GTPS binding was performed in the absence or presence of various concentrations of GDP, 5 mM Mg++ and 100 mM Na+ with ~10 µg membrane proteins at 30°C for 60 min as described under "Materials and Methods." Each value represents the mean ± S.E.M. of three independent experiments in duplicate. Control represents [35S]GTPS binding without GDP in the absence of 3 µM ()-U50,488H, which yielded 4 to 5.5 fmol/10 µg protein.

Effects of Mg⁺⁺ concentration on [³⁵S]GTPS binding. The importance of Mg⁺⁺ in [³⁵S]GTPS binding was determined in the absence and presence of 3 µM ()-U50,488H with 3 µM GDP and 100 mM Na⁺ (fig. 2). At concentrations between 5 µM and 0.5 mM, there was no significant difference in [³⁵S]GTPS binding between with and without ()-U50,488H. At 5 or 15 mM, the basal level of [³⁵S]GTPS binding was doubled. From 1.5 to 15 mM, binding in the presence of ()-U50,488H was greatly increased and a plateau was reached at 5 and 15 mM. At the plateau, agonist-induced binding was 2.5- to 3.5-fold of basal binding. At 50 mM, both basal level and agonist-induced binding decreased. Based on these results, 5 mM Mg⁺⁺ in the form of MgCl₂ was used in all other experiments.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Effect of Mg++ on [35S]GTPS binding to CHO-hkor membranes in the absence (open triangle) or presence (dark triangle) of 3 µM ()-U50,488H. [35S]GTPS binding was performed in the presence of various concentrations of Mg++, 3 µM GDP and 100 mM Na+ with ~7.5 µg membrane proteins at 30°C for 60 min as described under "Materials and Methods." Each value represents the mean ± S.E.M. of three independent experiments in duplicate.

Effects of Na⁺ on [³⁵S]GTPS binding. Effects of Na⁺ on [³⁵S]GTPS binding were investigated in the presence of 3 µM GDP and 5 mM Mg⁺⁺ with or without 3 µM ()-U50,488H (fig. 3). Without Na⁺, ()-U50,488H-stimulated increase in [³⁵S]GTPS binding was evident, although the magnitude of the increase was relatively small (not shown, but similar to 0.1 mM NaCl of fig. 3). At 1 mM, Na⁺ did not affect basal or agonist-induced [³⁵S]GTPS binding. Basal [³⁵S]GTPS binding was decreased by [Na⁺]  10 mM, whereas binding in the presence of ()-U50,488H was reduced at [Na⁺]  100 mM, both in a concentration-dependent manner. Binding in the presence of 100 mM Na⁺ provides the best signal-to-background difference. At this concentration, the basal binding amounted to approximately 28% of the maximal binding achieved by ()-U50,488H. In all other experiments, binding was performed in the presence of 100 mM NaCl.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Effect of Na+ on [35S]GTPS binding to CHO-hkor membranes in the absence (open triangle) or presence (dark triangle) of 3 µM ()-U50,488H. [35S]GTPS binding was performed in the presence of various concentrations of Na+, 3 µM GDP and 5 mM Mg++ with ~ 7.5 µg membrane proteins at 30°C for 60 min as described under "Materials and Methods." Each value represents the mean ± S.E.M. of three independent experiments in duplicate.

Time course of basal and ()-U50,488H-stimulated [³⁵S]GTPS binding. Time courses of [³⁵S]GTPS binding were conducted in the presence and absence of 3 µM ()-U50,488H at 30°C. Basal [³⁵S]GTPS binding increased with time, whereas ()-U50,488H-stimulated [³⁵S]GTPS binding (difference between with and without ()-U50,488H) reached a plateau at ~90 min.

Relationship between [³⁵S]GTPS binding and the amount of CHO-hkor membrane. [³⁵S]GTPS binding increased linearly with or without 3 µM ()-U50,488H up to 40 µg of membrane proteins in each assay tube. Routinely, 7.5 to 10 µg of membrane proteins were used, which gave 3,500 to 5,500 dpm and 1,000 to 1,500 dpm for stimulated and basal [³⁵S]GTPS binding, respectively.

Effect of ()-U50,488H on [³⁵S]GTPS binding to CHO-hkor membranes. [³⁵S]GTPS binding to CHO-hkor membranes with various concentrations of ()-U50,488H was examined in the presence of 3 µM GDP, 5 mM Mg⁺⁺ and 100 mM Na⁺ (fig. 4). Binding of [³⁵S]GTPS was increased by ()-U50,488H in a concentration-dependent manner with an EC₅₀ of 3.1 nM. The maximal response, which represented ~3.5-fold of the basal level, was reached at 0.3 µM ()-U50,488H (fig. 4). Ten micromolar ()-U50,488H failed to stimulate [³⁵S]GTPS binding to membranes of untransfected CHO cells (not shown). (+)-U50,488H up to 1 µM did not increase [³⁵S]GTPS binding, compared with the basal level (not shown), which indicated stereospecificity of receptor activation. Naloxone (1 µM) or norbinaltorphimine (10 nM) shifted the concentration-effect curve of ()-U50,488H to the right by about 100-fold (fig. 4). K_e values of naloxone and norbinaltorphimine calculated from these data were 10 nM and 0.1 nM, respectively, indicating that their potencies in the CHO-hkor system are similar to those in other kappa opioid receptor assays (Portoghese et al., 1987). The mu agonist DAMGO and the delta agonist DPDPE had no effect (fig. 4). Taken together, these results indicate that ()-U50,488H-induced increase in [³⁵S]GTPS binding is mediated by specific activation of the kappa opioid receptor.

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Dose-response relationship of [35S]GTPS binding to CHO-hkor membranes induced by ()-U50,488H, ()-U50,488H + 1 µM naloxone, ()-U50,488H + 10 nM norbinaltorphimine, DAMGO and DPDPE. [35S]GTPS binding was performed in the presence of 3 µM GDP, 5 mM Mg++ and 100 mM Na+ with ~7.5 µg membrane proteins at 30°C for 60 min as described under "Materials and Methods." Basal [35S]GTPS binding was 0.5 to 0.7 fmol/10 µg protein, whereas stimulated levels were 1.8 to 2.7 fmol/10 µg protein. Basal binding was subtracted from each datum. Data were expressed as % of maximal binding produced by ()-U50,488H. Each value represents the mean ± S.E.M. of three independent experiments in duplicate.

Effects of pretreatment with pertussis toxin or cholera toxin on [³⁵S]GTPS binding. To assess the G proteins that were involved in the action of ()-U50,488H, CHO-hkor cells were treated with pertussis toxin (100 ng/ml) or cholera toxin (20 µg/ml) for 24 h before membrane preparation. Pertussis toxin pretreatment completely abolished ()-U50,488H-induced increase in [³⁵S]GTPS binding and reduced basal binding of [³⁵S]GTPS by 60% (fig. 5A). In contrast, cholera toxin pretreatment did not affect basal or ()-U50,488H-induced [³⁵S]GTPS binding (fig. 5B). Thus, ()-U50,488H-induced [³⁵S]GTPS binding was most likely caused by binding to pertussis toxin-sensitive G proteins, i.e., G_i and/or G_o.

View larger version (27K): [in this window] [in a new window]   Fig. 5.   Effect of pretreatment with pertussis toxin (A) or cholera toxin (B) on ()-U50,488H-induced [35S]GTPS binding. CHO-hkor cells were treated with pertussis toxin (PTX; 100 ng/ml) or cholera toxin (CTX; 20 µg/ml) for 24 h and [35S]GTPS binding to membranes was performed with or without ()-U50,488H. Binding to membranes of untreated cells served as control. [35S]GTPS binding was performed in the presence of 3 µM GDP, 5 mM Mg++ and 100 mM Na+ with ~7.5 µg membrane proteins at 30°C for 60 min as described under "Materials and Methods." Each value represents the mean ± S.E.M. of three independent experiments in duplicate. Data are expressed as % of the control basal level. All pertussis toxin-treated binding values were significantly lower than the control basal level (P <.05, Student's t test).

Determination of K_d and B_max of ()-U50,488H-induced [³⁵S]GTPS binding. Displacement of [³⁵S]GTPS binding with unlabeled GTPS was performed in the absence and presence of 10 µM ()-U50,488H for 180 min (fig. 6) to determine K_d and B_max of [³⁵S]GTPS binding that could be maximally stimulated by (-),U50,488H. Scatchard analysis of the difference between the two curves revealed a K_d of 0.34 ± 0.08 nM and a B_max of 431 ± 29 fmol/mg protein for [³⁵S]GTPS binding (mean ± S.E.M., n = 3).

View larger version (20K): [in this window] [in a new window]   Fig. 6.   Displacement curves of [35S]GTPS binding by GTPS in the presence (open circle) and absence (dark circle) of 3 µM ()-U50,488H. [35S]GTPS binding was performed in the presence of 3 µM GDP, 5 mM Mg++ and 100 mM Na+ with ~7.5 µg membrane proteins at 30°C for 60 min as described under "Materials and Methods." Each value represents the mean ± S.E.M. of three independent experiments in duplicate. Differences between two curves were analyzed with Scatchard analysis (inset), which yielded Kd and Bmax values of 0.34 ± 0.08 nM and 431 ± 29 fmol/mg protein.

Determination of potencies and efficacies of opioid ligands on [³⁵S]GTPS binding. Several opioid receptor ligands were examined for their potencies and efficacies in stimulating [³⁵S]GTPS binding to membranes. EC₅₀ values and maximal responses were determined and compared with those of ()-U50,488H (fig. 7, table 1). Dynorphin A 1-17, (±)-ethylketocyclazocine, tifluadom and -funaltrexamine produced maximal responses similar to ()-U50,488H and were full agonists. Dynorphin A 1-17 and (±)-ethylketocyclazocine had higher potencies than ()-U50,488H with EC₅₀ values of 0.18 and 0.57 nM, respectively, whereas -funaltrexamine and tifluadom were potent similar to ()-U50,488H with EC₅₀ values of 1.7 nM and 3.9 nM, respectively. Nalorphine and pentazocine acted as partial agonists with the maximal stimulation of [³⁵S]GTPS binding at 68% and 23% of that of ()-U50,488H, respectively. EC₅₀ values of nalorphine and pentazocine, determined as the dose that elicited 50% of the maximal response of each drug, were 17.9 nM and 103 nM, respectively. Nalbuphine and buprenorphine had some stimulatory effects at 100 nM, but no plateau in maximal response was detected at up to 30 µM. The rank order of potencies of ligands tested for stimulation of [³⁵S]GTPS binding was dynorphin A 1-17 > (±)-ethylketocyclazocine > -funaltrexamine, tifluadom, ()-U50,488H > nalorphine > pentazocine, nalbuphine > buprenorphine. Although norbinaltorphimine and naloxone had high affinity for the kappa receptor, they did not increase [³⁵S]GTPS binding. In addition, naloxone and norbinaltorphimine shifted the dose-response curve of ()-U50,488H to the right (fig. 4), indicating that they are antagonists.

View larger version (26K): [in this window] [in a new window]   Fig. 7.   Dose-response curves of opioid ligands on [35S]GTPS binding to CHO-hkor membranes. [35S]GTPS binding was performed in the presence 3 µM GDP, 5 mM Mg++ and 100 mM Na+ with ~7.5 µg membrane proteins at 30°C for 60 min as described under "Materials and Methods." A dose-response curve of ()-U50,488H was included in each experiment and data are expressed as % of the maximal response induced by ()-U50,488H. Basal [35S]GTPS binding was 0.5 to 0.7 fmol/10 µg protein, whereas stimulated levels were 1.8 to 2.5 fmol/10 µg protein. Basal binding was subtracted from each datum. Each value represents the mean ± S.E.M. of three independent experiments in duplicate.

Binding affinities of opioid ligands to the human kappa opioid receptor. Competitive inhibition of [³H]diprenorphine binding by opioid ligands to CHO-hkor membranes was conducted to determine binding affinities of opioid ligands tested in [³⁵S]GTPS binding. Binding was performed in [³⁵S]GTPS binding buffer as well as in TEL buffer, because Tris-HCl buffer is the most commonly used buffer for binding studies. K_i values of these ligands were listed in table 1. There was no significant difference in K_i values of each drug in these two buffers. These values determined in membranes of CHO-hkor cells were similar to those determined in membranes of COS-1 cells transiently expressing the kappa receptor (Zhu et al., 1995). Ligands selective for kappa receptor (()-U50,488H, norbinaltorphimine, dynorphin A 1-17) had much higher affinity than the mu selective ligand DAMGO or the delta selective ligand DPDPE. For the five full agonists (dynorphin A 1-17, (±)-ethylketocyclazocine, tifluadom, -funaltrexamine and ()-U50,488H), the K_i values were very similar to the EC₅₀ values for stimulation of [³⁵S]GTPS binding (table 1).

	Discussion

Top Abstract Introduction Materials Results Discussion References

Opioid receptors mediate functional effects of agonists via activation of G proteins. In the present study, we demonstrated the ability of kappa agonists to enhance binding of [³⁵S]GTPS to membranes of CHO-hkor cells. The presence of GDP and Mg⁺⁺ was essential for agonist-stimulated [³⁵S]GTPS binding. Na⁺ was necessary for maximal signal-to-background ratios. The extent of ()-U50,488H-induced increase in [³⁵S]GTPS binding was dependent on agonist concentration, incubation time and tissue concentration. The effect of ()-U50,488H was stereospecific and reversed by naloxone or norbinaltorphimine, which indicates that this effect is mediated by the kappa receptor. Pertussis toxin-, but not cholera toxin-sensitive G proteins were involved. In this system, [³⁵S]GTPS binding assay allowed classification of high-affinity kappa opioid ligands into full agonists, partial agonists and antagonists. Although full agonists elicit maximal effects, partial agonists and antagonists produce submaximal and no effects, respectively. This biochemical assay thus permits determination of efficacies of kappa opioid ligands. In addition, the CHO-hkor cell, being devoid of other receptors, is an excellent system for this purpose. EC₅₀ values of five full agonists tested in stimulating [³⁵S]GTPS binding were in the nanomolar range, very similar to their K_i values in inhibiting [³H]diprenorphine binding, which strongly suggests that there is no or little spare receptors in this system.

In addition to its relatively large stimulated signals and its allowing determination of ligand efficacy, [³⁵S]GTPS binding assay has some practical advantages. Membranes can be prepared and frozen for convenience. The assay itself is easy, quick and reproducible. Recently, Befort et al. (1996) reported use of [³⁵S]GTPS binding to evaluate activity of delta opioid receptors transiently expressed COS cells. The utility of this assay for transiently expressed receptors will facilitate examination of effects of mutations on receptor functional activity.

Stimulation of [³⁵S]GTPS binding by agonists allows one to examine G protein activation by ligand-occupied receptors regardless of the types of G proteins and effector systems involved. In this regard, it is similar to kappa agonist-induced enhancement of GTPase activity of G proteins (Clark et al., 1986; Clark and Medzihradsky, 1987; Lawrence et al., 1995) and increase in labeling of G subunits by [³²P]azidoanilido-GTP (Prather et al., 1995). ()-U50,488H and dynorphin peptides stimulated low-K_m GTPase activity by 10 to 20% with EC₅₀ values of 3 to 23 µM in guinea pig cerebellum membranes (Clark et al., 1986) and in rat and monkey brain membranes pretreated with -funaltrexamine and cis-(±)-3-methylfentanyl isothiocyanate to alkylate the mu and delta receptors, respectively (Clark and Medzihradsky, 1987). These EC₅₀ values are more than 100-fold greater than their K_i values of binding to the kappa receptor. In three mouse thymoma cell lines, stimulation of low-K_m GTPase activity by ()-U50,488H was observed at  10 nM and had a 21 to 53% increase over the basal level at 30 µM ()-U50,488H (Lawrence et al., 1995). Activation of a cloned rat kappa opioid receptor stably expressed in CHO cells increased the incorporation of [³²P]azidoanilido-GTP into four G subunits with maximal increases of 20 to 44%, and EC₅₀ values of ()-U50,488H and dynorphin A 1-17 were in the nanomolar range, similar to their K_i values in binding to the kappa receptor (Prather et al., 1995). In these studies, ()-U50,488H and dynorphin peptides, but no partial agonists, were examined.

In addition to activation of GTPase, inhibition of forskolin-stimulated adenylate cyclase activity has been used as a biochemical measure of activation of the kappa opioid receptor (Attali et al., 1989; Konkoy and Childers, 1989, 1993; Lawrence and Bidlack, 1993; Lawrence et al., 1995; Prather et al., 1995). In CHO cells stably transfected with a cloned rat kappa receptor, dynorphin A 1-17 and ()-U50,488H were very potent in inhibiting forskolin-stimulated adenylate cyclase with EC₅₀ values in the nanomolar range (Prather et al., 1995). In contrast, in many other systems, EC₅₀ values of ()-U50,488H, U69,593 and dynorphin A 1-17 for the inhibition of forskolin-stimulated adenylate cyclase activity were in the micromolar range, 100- to 1000-fold of their K_i values of binding to the kappa opioid receptor. These included the rat spinal cord and spinal cord-dorsal root ganglion culture (Attali et al., 1989), guinea pig cerebellar membranes (Konkoy and Childers, 1989), guinea pig brain membranes (Konkoy and Childers, 1993) and mouse thymoma cells (Lawrence et al., 1995; Lawrence and Bidlack, 1993). Maximal inhibition that could be achieved by ()-U50,488H or dynorphin A 1-17 was 40% in guinea pig brain membranes (Konkoy and Childers, 1993), 37 to 66% in mouse thymoma cell lines (Lawrence et al., 1995; Lawrence and Bidlack, 1993) and 56 to 75% in CHO cells expressing a rat kappa receptor (Prather et al., 1995). Only ()-U50,488H, U69,593 and dynorphin peptides were examined in the studies on inhibition of adenylate cyclase. Whether partial agonists were active in these systems was not determined.

The presence of GDP is essential for ()-U50,488H-induced increase in [³⁵S]GTPS binding. Optimal signals were obtained with 1 to 10 µM GDP. The requirement of GDP for the agonist-induced increase in binding of [³⁵S]GTPS has also been demonstrated for many other receptors including muscarinic, A1 adenosine and mu opioid receptors (Hilf et al., 1989; Lorenzen et al., 1993; Traynor and Nahorski, 1995). For the alpha-2D adrenergic receptor, GDP is not required for agonist-induced increase in [³⁵S]GTPS binding, but it increased the magnitude of stimulation caused by an agonist (Tian et al., 1994). The mechanism of action of GDP, however, is not fully understood. From the data in figure 1, we propose the following hypothesis. In the absence of an agonist, GDP inhibits [³⁵S]GTPS binding in a dose-dependent manner with an IC₅₀ value of ~0.8 µM. With no or a low concentration of GDP, [³⁵S]GTPS competes favorably over GDP for binding to guanine nucleotide binding sites of G subunits. As a result, the agonist-induced increase in [³⁵S]GTPS binding is obscured. Addition of GDP in the micromolar range results in the binding of GDP to most guanine nucleotide binding sites of G subunits. Binding of an agonist to the receptor increases dissociation of GDP from the interacting G proteins and their association of [³⁵S]GTPS (Gilman, 1987) and, therefore, allows the increase in [³⁵S]GTPS binding to be observed.

Kappa agonist-stimulated binding of [³⁵S]GTPS depends on the presence of Mg⁺⁺, and the maximal signal-to-background ratio was observed between 5 and 15 mM. Mg⁺⁺ has multiple effects on signal transduction of G protein-coupled receptors (for reviews, Gilman, 1987; Birnbaumer et al., 1990). The effects that require Mg⁺⁺ are as follows: 1) formation of agonist-receptor-G protein ternary complex; 2) activation of G proteins by agonist-occupied receptors; 3) stimulation of GTP binding to G proteins and GDP dissociation by an agonist-receptor complex; 4) reduction of dissociation of GTPS from G proteins to near zero; 5) -stimulated GDP dissociation and GTPS-induced subunit dissociation; 6) GTPase activity. The net results are that Mg⁺⁺ promotes dissociation of oligomeric G proteins and the formation of an "activated" state of G. Effects of Mg⁺⁺ on [³⁵S]GTPS binding are likely caused by a combination of effects 1 through 5.

Na⁺ is not required for kappa agonist stimulation, because ()-U50,488H can increase [³⁵S]GTPS binding even without Na⁺. However, better stimulated-to-basal difference was obtained in the presence of >30 mM [Na⁺]. Similar observations have also been reported in studies of G protein activation by muscarinic (Hilf et al., 1989), formyl peptide (Gierschik et al., 1989), A1 adenosine (Lorenzen et al., 1993) and alpha-2D adrenergic (Tian et al., 1994) receptors. It appears that Na⁺ prevents activation of G proteins by unoccupied receptors (Gierschik et al., 1989). This effect of Na⁺ is in accord with the findings that Na⁺ was required for kappa agonist-induced increase in low-K_m GTPase activity (Clark et al., 1986; Clark and Medzihradsky, 1987; Lawrence et al., 1995), inhibition of forskolin-stimulated adenylate cyclase (Attali et al., 1989; Konkoy and Childers, 1989, 1993; Lawrence and Bidlack, 1993; Lawrence et al., 1995) and increase in [³²P]azidoanilido-GTP into G subunits (Prather et al., 1995).

Enhancement of [³⁵S]GTPS binding by the kappa opioid agonist ()-U50,488H was completely blocked by pretreatment of the cells with pertussis toxin, but not with cholera toxin, which confirms that the event is mediated entirely through pertussis toxin-sensitive G proteins. This result agrees with the findings that actions of kappa opioid receptors are mediated through pertussis toxin-sensitive G_i and/or G_o proteins (Lawrence and Bidlack, 1993; Prather et al., 1995; Ma et al., 1995). Inhibition of adenylate cyclase by kappa receptor activation was blocked by pertussis toxin pretreatment (Lawrence and Bidlack, 1993). Activation of a cloned rat kappa opioid receptor increased the incorporation of [³²P]azidoanilido-GTP into four G subunits, three of which were identified as G_i3, G_i2 and G_o2 (Prather et al., 1995). Pertussis toxin treatment abolished kappa agonist-induced increase in K⁺ conductance through an inward rectifying K⁺ channel (Ma et al., 1995). Basal [³⁵S]GTPS binding was also lowered by pertussis toxin treatment (see fig. 5A). A similar reduction of basal [³⁵S]GTPS binding by pertussis toxin was observed in SHSY-5Y cells (Traynor and Nahorski, 1995). This may be caused by active coupling of unoccupied receptors to pertussis toxin-sensitive G proteins and, thus, stimulation of [³⁵S]GTPS binding to G proteins at resting condition, similar to alpha-2D adrenergic receptor (Tian et al., 1994). In this CHO-hkor system, kappa opioid receptors are not coupled to cholera toxin-sensitive G proteins. Coupling of the kappa opioid receptor to cholera toxin-sensitive G proteins was reported in the dorsal root ganglion-spinal dorsal horn culture (Crain and Shen, 1990) and in myenteric plexuses (Gintzler and Xu, 1991).

Potencies and efficacies of opioid ligands in stimulating mu opioid receptor-mediated [³⁵S]GTPS binding to SHSY-5Y cell membranes correlated very well with those in other functional assays, such as an in vivo antinociceptive test and inhibition of contraction in guinea pig ileum in vitro (Traynor and Nahorski, 1995). In the present study, opioid agonists had different potencies in kappa receptor-mediated enhancement of [³⁵S]GTPS binding and produced varying maximal responses, indicative of their efficacies. The rabbit vas deferens contained kappa opioid receptor, but not mu and delta receptors (Oka et al., 1981). Inhibition of field-stimulated contraction of the rabbit vas deferens was used to determine efficacies and potencies of ligands on kappa opioid receptors (Hayes and Kelly, 1985; Miller et al., 1986; Verlinde and De Ranter, 1988). Our finding that ethylketocyclazocine, tifluadom and ()-U50,488H were full agonists in stimulating [³⁵S]GTPS binding is consistent with the observation that these three compounds were full agonists in the rabbit vas deferens (Romer et al., 1982; Hayes and Kelly, 1985; Verlinde and De Ranter, 1988). ()-U50,488H and tifluadom were equally potent in stimulating [³⁵S]GTPS binding, whereas tifluadom was 7 to 400 times more potent than ()-U50,488H in the rabbit vas deferens (Hayes and Kelly, 1985; Miller et al., 1986; Verlinde and De Ranter, 1988). Although ethylketocyclazocine was about 10 times more potent than ()-U50,488H, in stimulating [³⁵S]GTPS binding it is 6 or 60 times more potent than ()-U50,488H in the rabbit vas deferens (Hayes and Kelly, 1985; Miller et al., 1986). The observation that -funaltrexamine enhances [³⁵S]GTPS binding in CHO-hkor cells is consistent with the original findings that it is a kappa agonist, in addition to having irreversible mu antagonist activities (Portoghese et al., 1980).

In the present study, nalorphine and pentazocine were partial agonists at the kappa receptor in stimulating [³⁵S]GTPS binding and nalbuphine had low level of agonist activity. Our findings are consistent with those of Miller et al. (1986) that pentazocine and nalorphine inhibited electrically induced contraction of guinea pig ileum by acting on the kappa opioid receptor with a K_e value of naloxone of ~20 nM. In contrast, pentazocine, nalbuphine and nalorphine were inactive in vas deferens preparations of the rat, mouse and rabbit, but they could antagonize the action of ethylketocyclazocine in these preparations (Hayes and Kelly, 1985; Miller et al., 1986; Verlinde and De Ranter, 1988). The variation in potencies of these compounds in these in vitro preparations was hypothesized to be caused by the difference in the number of spare receptors in these models (Hayes and Kelly, 1985; Miller et al., 1986).

Our observations that ()-U50,488H, tifluadom and ethylketocyclazocine are full agonists and nalorphine and nalbuphine are partial agonists are in accord with results from in vivo pharmacological studies. ()-U50,488H, tifluadom and ethylketocyclazocine were found to be highly efficacious kappa agonists in kappa receptor-mediated antinociception (Von Voigtlander et al., 1983; Piercey et al., 1982; France et al., 1994; Dykstra et al., 1987) and diuresis (Von Voigtlander et al., 1983; Leander, 1983a; Dykstra et al., 1987; Takemori et al., 1988). In drug discrimination procedures, these drugs substitute completely for kappa agonists ethylketocyclazocine, bremazocine and spiradoline (Holtzman et al., 1991; France et al., 1994; Dykstra et al., 1987; Picker, 1994b; Smith and Picker, 1995). On the contrary, maximal effects produced by nalorphine were less than those of full agonists in kappa receptor-mediated diuresis (Leander, 1983a, b). Nalorphine and nalbuphine substituted partially or failed to substitute for bremazocine in drug discrimination tests (Smith and Picker, 1995; Picker, 1994a). In addition, nalorphine and nalbuphine antagonized the stimulus effect of bremazocine (Picker, 1994a; Smith and Picker, 1995). Nalorphine antagonized the diuretic effect of bremazocine (Leander, 1983b) as well as the antinociceptive effect of U50,488H (Dykstra, 1990).

Naloxone and norbinaltorphimine shifted the dose-response curve of ()-U50,488H to the right, but neither had any effect on the basal [³⁵S]GTPS binding. These results indicate that both compounds are pure antagonists without positive or negative intrinsic activity in this system.

For five full agonists, the EC₅₀ values in stimulating [³⁵S]GTPS binding were similar to their K_i values in inhibiting [³H]diprenorphine binding. This finding indicates that there are few or no spare receptors in this CHO-hkor system.

The binding affinities of ligands examined, including full agonists, partial agonists and antagonists, for the kappa receptor were similar in [³⁵S]GTPS binding buffer and in commonly used Tris-based receptor binding buffer. [³⁵S]GTPS binding buffer contains 100 mM Na⁺, 5 mM Mg⁺⁺ and 3 µM GDP, whereas the Tris-based receptor binding buffer does not. We found that the addition of 100 mM Na⁺, 5 mM Mg⁺⁺, 3 µM GDP or all three to Tris-based receptor binding buffer did not change the apparent K_i value of ()-U50,488H (Li, J.-G. and Liu-Chen, L.-Y., unpublished observation). Based on results in table 1, it is likely that Na⁺, Mg⁺⁺, GDP or all three at the concentrations used did not affect binding affinities of other ligands examined. Contrary to our observations, Lawrence and Bidlack (1992) showed that in R1.1 thymoma cell membranes, 100 mM Na⁺ and 10 mM Mg⁺⁺ inhibited 50 pM ()-[³H]bremazocine binding by 45% and 37%, respectively. Although 100 µM GDP did not have an effect on ()-[³H]bremazocine binding, a combination of 100 µM GDP and 30 mM Na⁺ inhibited binding by 54%. Thus, effects of Na⁺, Mg⁺⁺ and GDP on kappa agonist binding may also depend on the ligand examined. This issue needs further investigation.

In conclusion, stimulation of [³⁵S]GTPS binding to membranes of CHO cells stably expressing the human kappa opioid receptor by kappa agonists provides a useful functional measure for interaction between kappa opioid receptors and pertussis toxin-sensitive G proteins. In addition, this assay distinguishes ligands of full agonists, partial agonists and antagonists, which, in general, agrees with results obtained from in vitro tissue preparations and in vivo pharmacology.