Introduction

Summary Introduction Procedures Results Discussion References

Opioid agonists differ in their maximal ability to produce biological responses. As with other drugs that act by binding to specific receptors, agonists that produce maximal efficacy are termed full agonists, whereas those that produce less-than-maximal efficacy (at full receptor occupancy) are called partial agonists. For the superfamily of G protein-coupled receptors, of which opioid receptors are members (1-4), efficacy is primarily determined by the interaction between receptors and the G protein transducers (5). Opioid receptors produce biological responses by selectively activating G proteins of the pertussis toxin-sensitive G_i/G_o family (6), which couple to effectors, including inhibition of adenylyl cyclase (7, 8), stimulation of potassium channel conductance (9, 10), and inhibition of calcium channel conductance (11, 12). In the cycle of G protein activation, the receptor interacts with the G protein and decreases the affinity of the GDP-bound  subunit for GDP and increases its affinity for GTP, thus promoting guanine nucleotide exchange (13). The receptor catalytically activates G proteins, so that each receptor can activate multiple G proteins (14-16).

Recently, a method was developed to examine receptor activation of G proteins in isolated membranes by assaying agonist-stimulated binding of the hydrolysis-resistant GTP analog [³⁵S]GTPS in the presence of excess GDP (17-19). Traynor and Nahorski (19) have shown that different µ-selective opioid agonists have different efficacies in stimulating [³⁵S]GTPS binding to SH-SY5Y cell membranes, as measured by different maximal stimulation of [³⁵S]GTPS binding in agonist concentration-effect curves. Differences were also found among agonists that have been previously reported to produce "full" agonist responses in isolated organ preparations [e.g., the guinea pig ileum (20)] at different levels of receptor occupancy (21). These results correlate with the results of behavioral studies, which suggest that opioid agonists of different intrinsic efficacies have the ability to produce analgesic, discriminative stimulus, and reinforcing effects at different levels of receptor occupancy (22-26). Studies using G protein antisense oligonucleotides have indicated that analgesic efficacy differences may be related to the ability of these agonists to activate G proteins (27, 28). Thus, there is reasonable evidence to suggest that various opioid analgesics have different intrinsic efficacies and that these differences are likely due to the different abilities of these agonists to activate G proteins through µ-opioid receptors.

Recent studies in rat striatal membranes have determined that different types of receptors (i.e., µ-opioid, -opioid, and cannabinoid) exhibit different abilities to stimulate [³⁵S]GTPS binding under conditions of maximal receptor occupancy by full agonists (16). Comparison of the B_max value of receptor binding with the B_max value of agonist-stimulated [³⁵S]GTPS binding revealed that these differences were due to the overall catalytic amplification by receptors under these conditions (i.e., the number of G proteins activated by each receptor depended on the receptor type). On the other hand, the K_D value of agonist-stimulated [³⁵S]GTPS binding, a measure of the inherent ability of the agonist to change the conformation of the G protein  subunit to a high affinity GTP-preferring state, was not different among the different receptor types. Because these findings were determined for full agonists in three different receptor systems, the current study was designed to use the same technique to explore whether full and partial agonists of the same receptor system (in this case, µ-opioid receptors) produce similar differences in receptor-mediated G protein activation. These experiments were performed in rat thalamus, a brain region enriched in µ-opioid receptors (29, 30); human neuroblastoma SK-N-SH cells, which are also enriched in µ-opioid receptors (31-33); and CHO cells transfected with cDNA encoding the mouse µ-opioid receptor, which express high levels of the µ receptor (mMOR-CHO cells) (34a). These experiments reveal that differences in agonist efficacy may be the result of differential abilities of these agonists both to induce a receptor-mediated high affinity GTP-binding state in receptor-coupled G proteins and to stimulate different levels of receptor-mediated catalytic activation of G proteins. These studies help to elucidate the signal transduction mechanisms underlying agonist efficacy for µ-opioid receptors.

	Experimental Procedures

Summary Introduction Procedures Results Discussion References

Materials. [³⁵S]GTPS (1150-1300 Ci/mmol) was purchased from New England Nuclear (Boston, MA). SK-N-SH cells were obtained from American Type Culture Collection (Rockville, MD). mMOR-CHO cells were generously provided by Drs. Lawrence Toll (SRI International, Menlo Park, CA) and Christopher Evans (University of California, Los Angeles, CA). Ecolite scintillation fluid was obtained from ICN Biomedicals (Irvine, CA). DAMGO, naloxone, and penicillin/streptomycin were purchased from Sigma Chemical (St. Louis, MO). All nonpeptide opioid agonists were obtained from the National Institute for Drug Abuse drug supply program (Research Triangle Institute, Research Triangle Park, NC). FBS and geneticin (G-418) were purchased from GIBCO (Grand Island, NY). GTPS and guanosine-5-diphosphate were purchased from Boehringer-Mannheim Biochemicals (New York, NY). All other chemicals (reagent grade) were obtained from Sigma or Fisher Scientific (Fair Lawn, NJ).

Cell culture. Cells were cultured at 37° in a humidified atmosphere of 5% CO₂/95% air. SK-N-SH cells were cultured in Dulbecco's modified Eagle's medium containing 100 units/ml penicillin, 100 µg/ml streptomycin, and 10% FBS. mMOR-CHO cells were cultured in 50% Dulbecco's modified Eagle's medium and 50% Ham's F-12 Nutrient Mixture containing 100 units/ml penicillin, 100 µg/ml streptomycin, and 5% FBS. Cells were harvested by replacing the media with cold phosphate-buffered saline containing 0.04% EDTA for 5 min, followed by agitation. Cells were collected by centrifugation at 345 × g for 10 min.

Membrane preparation. Rats were killed by decapitation, and the thalamus was dissected on ice. Rat thalamic tissue, SK-N-SH cells, or mMOR-CHO cells were homogenized in 20 volumes of ice-cold 50 mM Tris·HCl, 3 mM MgCl₂, 1 mM EGTA, pH 7.4 (membrane buffer) with a Tissumizer (Tekmar, Cinncinnati, OH). The homogenate was centrifuged at 48,000 × g at 4° for 10 min, resuspended in membrane buffer, centrifuged again at 48,000 × g at 4° for 10 min, and finally resuspended in the assay buffer (50 mM Tris·HCl, 3 mM MgCl₂, 0.2 mM EGTA, 100 mM NaCl, pH 7.4). For SK-N-SH cell membranes, this procedure was preceded by a low-speed centrifugation at 500 × g, from which the supernatant was kept and the pellet was discarded. Membrane protein levels were determined according to the method of Bradford (35).

[³⁵S]GTPS binding assays. Agonist-stimulated [³⁵S]GTPS binding was examined by a modification of previously published methods (19, 30). Rat thalamic (10 µg of protein), SK-N-SH cell (50 µg of protein), or mMOR-CHO cell (25 µg of protein) membranes were incubated for 1 hr at 30°, with and without various drugs, in assay buffer containing 0.05 nM [³⁵S]GTPS and 10-30 µM GDP. Basal binding was assessed in the presence of GDP and absence of drug, whereas nonspecific binding was measured in the presence of 10 µM GTPS. In some experiments, 0.03-50 µM GDP was used. For Scatchard analysis, 0.1 nM [³⁵S]GTPS was incubated with 0.05-2000 nM GTPS in the presence of 50 µM (rat thalamus), 30 µM (SK-N-SH), or 10 µM GDP (mMOR-CHO) with and without various drugs for 1 hr (SK-N-SH and mMOR-CHO) or 2 hr (rat thalamus) at 30°. The incubation was terminated by rapid filtration under vacuum through Whatman GF/B glass fiber filters, followed by three washes with 3 ml of ice-cold 50 mM Tris·HCl, pH 7.4. Bound radioactivity was determined by liquid scintillation spectrophotometry at 95% efficiency after overnight extraction in 5 ml of Ecolite scintillation fluid.

[³⁵S]GTPS binding autoradiography. [³⁵S]GTPS autoradiography was performed in vitro as previously described (30). Briefly, coronal sections from male Sprague-Dawley rats were cut at the level of the thalamus on a cryostat at 20° and thaw-mounted onto gelatin-coated slides. Sections were rinsed in assay buffer at 25° and then preincubated with 2 mM GDP in assay buffer at 25° for 15 min. Sections were incubated with and without various drugs in the presence of 0.04 nM [³⁵S]GTPS and 2 mM GDP in assay buffer at 25° for 2 hr. Basal binding was assessed in the presence of GDP and absence of drug. Slides were then rinsed twice in ice-cold 50 mM Tris·HCl, pH 7.0, at 25° and rinsed once briefly in deionized water. Slides were dried overnight and exposed to Reflections film (New England Nuclear) for 6 days. Films were digitized with a Sony XC-77 video camera and analyzed using the NIH Image program for Macintosh computers. Images were quantified by comparison with ¹⁴C standards, and values were adjusted to ³⁵S as previously described (36).

Data analysis. Unless otherwise indicated, data are reported as mean ± standard error of at least three separate experiments that were each performed in triplicate. Net stimulated [³⁵S]GTPS binding is defined as stimulated binding minus basal binding. Percent stimulation is defined as (net stimulated binding/basal binding) × 100%. Percent maximal stimulation is defined as (net stimulated binding by agonist/by 10 µM DAMGO) × 100%. This parameter was determined within each individual experiment by inclusion of 10 µM DAMGO in each assay, and the normalized data were subjected to nonlinear regression analysis to determine ED₅₀ and efficacy (E_max) values. Statistical significance of the data was determined by analysis of variance followed by the nonpaired two-tailed Student's t test, using JMP (SAS Institute, Cary, NY). Nonlinear regression analysis of concentration-effect curves was also performed with JMP using an iterative model. Scatchard analyses were conducted using EBDA and LIGAND.

	Results

Summary Introduction Procedures Results Discussion References

Concentration-effect relationship of opioid stimulation of [³⁵S]GTPS binding in membranes. To determine the E_max and ED₅₀ values of opioid agonists for stimulation of [³⁵S]GTPS binding in membranes, concentration-effect curves were constructed using opioid agonists that have been reported to have different efficacies for G protein activation (19): DAMGO, fentanyl, morphine, buprenorphine, and levallorphan. Because agonist stimulation of [³⁵S]GTPS binding in membranes requires the presence of excess GDP, a GDP concentration (30 µM) was chosen that has previously been found to allow maximal stimulation of [³⁵S]GTPS binding by DAMGO in rat thalamic membranes (30). The concentration-effect curves for opioid agonist stimulation of [³⁵S]GTPS binding to rat thalamic membranes are shown in Fig. 1, A and B. Because DAMGO was the most efficacious agonist examined, producing an average stimulation of >165% above base-line, all agonists are presented as a percentage of the maximal stimulation produced by DAMGO (Fig. 1B). This calculation of the data was based on the inclusion of an internal "full efficacy" standard of 10 µM DAMGO in each individual experiment to control for interassy variability in the absolute stimulation. The results of these experiments showed that morphine and fentanyl were approximately equally efficacious and produced ~55-60% of the maximal stimulation observed with DAMGO. Buprenorphine was much less efficacious and produced only ~10% of the maximal stimulation obtained with DAMGO. Levallorphan acted as a pure antagonist in this system and produced no significant effect. The ED₅₀ and E_max values of the opioid agonists are summarized in Table 1. There was no correlation between efficacy and potency: the order of decreasing efficacy was DAMGO > fentanyl = morphine  buprenorphine, whereas the order of decreasing potency (increasing ED₅₀) was buprenorphine  fentanyl > DAMGO > morphine.

View larger version (37K): [in this window] [in a new window]   Fig. 1.   Concentration-effect relationship of opioid stimulation of [35S]GTPS binding in rat thalamic and mMOR-CHO cell membranes. Membranes prepared from (A and B) rat thalamus or (C and D) mMOR-CHO cells were incubated with 0.05 nM [35S]GTPS, 30 µM (rat thalamus), or 10 µM (mMOR-CHO) GDP and various concentrations of DAMGO, fentanyl, morphine, or buprenorphine. Data are mean ± standard error of (A and C) net agonist-stimulated [35S]GTPS binding or (B and D) percent of stimulation produced by 10 µM DAMGO. Basal [35S]GTPS binding values and ED50 and Emax values from curve-fitting of these data are shown in Table 1.

View this table: [in this window] [in a new window]   TABLE 1 Efficacy and potency of opioid agonists for stimulation of [35S]GTPS binding to rat thalamus, SK-N-SH, and mMOR-CHO membranes Membranes were incubated with various concentrations of opioid agonists in the presence of 0.05 nM [35S]GTPS and 30 µM (rat thalamus and SK-N-SH) or 10 µM (mMOR-CHO) GDP. Data are mean ± standard error of ED50 (nM) and Emax values (percent maximal stimulation obtained with 10 µM DAMGO). Maximal stimulation over basal by DAMGO was 167 ± 8.4%, 222 ± 18%, and 414 ± 18% in rat thalamic, SK-N-SH, and mMOR-CHO cell membranes, respectively. Basal [35S]GTPS binding was 87.6 ± 12.0, 7.8 ± 0.7, and 21.3 ± 1.3 fmol/mg of protein in rat thalamic, SK-N-SH, and mMOR-CHO cell membranes, respectively.

Autoradiographic examination of opioid agonist-stimulated [³⁵S]GTPS binding in rat brain sections. To confirm that membranes prepared from dissected rat thalamus actually contained the µ receptor-enriched region, and to examine the efficacy relationships among the opioid agonists in a more intact system, in vitro autoradiography of opioid agonist-stimulated [³⁵S]GTPS binding was examined in rat brain sections at the level of the thalamus, using maximally effective concentrations of agonists as determined from the concentration-effect curves. The results (Fig. 2) showed that all four agonists stimulated [³⁵S]GTPS binding with a distribution consistent with that of µ receptors (29, 36). This is best shown by DAMGO, which produced high levels of stimulation throughout the medial thalamus, medial hypothalamus, and amygdala, and produced low levels of stimulation in hippocampus and cortex. Comparison of labeling stimulated by all four agonists revealed similar efficacy relationships as observed in thalamic membranes: DAMGO > fentanyl = morphine  buprenorphine. When agonist-stimulated binding in the medial thalamus was quantified by densitometry (Table 2), some significant quantitative differences emerged between the results obtained in membranes and brain sections. Although buprenorphine produced approximately the same efficacy (13%) relative to DAMGO as that observed in thalamus membranes, the relative efficacies of fentanyl and morphine (37%) were much less than expected based on the results obtained in membranes.

View larger version (52K): [in this window] [in a new window]   Fig. 2.   In vitro autoradiography of basal and opioid agonist-stimulated [35S]GTPS binding in rat brain sections at the level of the thalamus. Sections were incubated with 0.04 nM [35S]GTPS, 2 mM GDP, and 5 µM DAMGO, 5 µM fentanyl, 10 µM morphine, or 0.03 µM buprenorphine. Images shown are from a typical experiment that was performed on triplicate (adjacent) sections and replicated three times. Values from densitometric analysis of this data are given in Table 2.

View this table: [in this window] [in a new window]   TABLE 2 Effect of opioid agonists on [35S]GTPS binding: in vitro autoradiographic analysis in rat thalamus Sections were incubated with opioid agonists (concentrations shown in legend to Fig. 2) in the presence of 0.04 nM [35S]GTPS and 2 mM GDP. Data are mean ± standard error of absolute [35S]GTPS binding, percent stimulation of [35S]GTPS binding over basal, and percent of maximal stimulation produced by 5 µM DAMGO.

Effect of GDP on opioid agonist efficacy for stimulation of [³⁵S]GTPS binding. The quantitative differences observed in the relative efficacies of the opioid agonists between rat thalamic membranes and sections suggested that a difference in assay conditions may produce different relative efficacies. The main difference in assay conditions was the GDP concentration: 30 µM GDP was used in rat thalamic membranes and 2 mM GDP was used in the autoradiography. Furthermore, similar quantitative differences were observed between relative agonist efficacies in SK-N-SH cell membranes in the current study and those previously reported in SH-SY5Y cell membranes, where 3 µM GDP was used in the assay (19). Therefore, the effect of GDP concentration on relative opioid agonist efficacy was investigated with maximal stimulatory concentrations of agonists (determined by the concentration-effect curves). The results in rat thalamic membranes showed that [³⁵S]GTPS binding was inhibited by GDP in both the absence and presence of opioid agonists (Fig. 3A), with 50 µM GDP inhibiting >90% of total [³⁵S]GTPS binding. However, the potency of GDP in inhibiting agonist-stimulated [³⁵S]GTPS binding was inversely proportional to the efficacy of the agonist. Net buprenorphine-stimulated [³⁵S]GTPS binding (not shown) was decreased half-maximally by ~3 µM GDP, whereas DAMGO-, fentanyl-, and morphine-stimulated binding was reduced half-maximally by ~30 µM GDP, and naloxone did not significantly stimulate [³⁵S]GTPS binding at any concentration of GDP. In Fig. 3B, the the data are expressed as percent stimulation of [³⁵S]GTPS binding above basal. These results shows that for DAMGO, fentanyl, and morphine, the percent stimulation of binding increased with increasing GDP concentration. However, the percent stimulation by buprenorphine was maximal at 0.2 µM GDP. Perhaps most importantly, the relative differences in efficacy among the four agonists apparently increased with increasing GDP concentration. Similar results were obtained in SK-N-SH cell membranes (not shown), in which half-maximal inhibition of net buprenorphine-stimulated [³⁵S]GTPS binding was observed at 4 µM GDP and half-maximal inhibition of net-stimulated [³⁵S]GTPS binding with the higher efficacy agonists was obtained at ~20 µM GDP. Again, relative efficacy differences among the agonists were magnified by increasing GDP concentrations.

View larger version (36K): [in this window] [in a new window]   Fig. 3.   Effect of GDP on basal and agonist-stimulated [35S]GTPS binding in rat thalamic and mMOR-CHO cell membranes. Membranes prepared from (A and B) rat thalamus or (C and D) mMOR-CHO cells were incubated with 0.05 nM [35S]GTPS and various concentrations of GDP in the presence and absence of opioid agonists. Agonist concentrations were 5 µM DAMGO, 5 µM fentanyl, 10 µM morphine, and 0.03 µM buprenorphine (rat thalamus) and 1 µM DAMGO, 1 µM fentanyl, 5 µM morphine, 0.03 µM buprenorphine, and 0.1 µM levallorphan (mMOR-CHO). Data shown are mean ± standard error of (A and C) percent total [35S]GTPS binding in the absence of GDP or agonist and (B and D) percent stimulation by agonist over basal binding measured at each GDP concentration. Total [35S]GTPS binding in the absence of GDP was 2174 ± 56 and 460 ± 17 fmol/mg of protein in rat thalamic and mMOR-CHO cell membranes, respectively.

Scatchard analysis of agonist-stimulated [³⁵S]GTPS binding. The primary mechanism of G protein activation is the receptor-induced shift in the affinity of the G protein  subunit for guanine nucleotides: the GDP affinity is decreased, and the GTP affinity is increased. Thus, it is possible to determine whether differences in agonist efficacy are due to differences in the degree to which the agonist-occupied receptor can induce a high affinity GTP binding site on the G protein (measured as the K_D value of [³⁵S]GTPS binding) and/or the number of G proteins activated by the agonist-occupied receptor (measured as the B_max value of [³⁵S]GTPS binding). To examine these parameters, Scatchard analysis of basal and agonist-stimulated [³⁵S]GTPS binding to membranes was performed in the presence of GDP, as previously described (15).¹ These experiments revealed little high affinity [³⁵S]GTPS binding in the absence of agonist, as demonstrated by both the homologous displacement of [³⁵S]GTPS binding by increasing concentrations of GTPS (Fig. 4A) and Scatchard analysis of these data (Fig. 4B). The addition of the full agonist DAMGO produced significant stimulation of [³⁵S]GTPS binding above basal levels (Fig. 4A). This stimulation resulted in a biphasic Scatchard plot (Fig. 4B) with a large increase in high affinity binding and no significant change in low affinity binding (Table 3), as previously observed for agonist-stimulated [³⁵S]GTPS binding in both SK-N-SH¹ and HL-60 (15) cell membranes. Scatchard analysis of net receptor-stimulated [³⁵S]GTPS binding was obtained by subtraction of basal [³⁵S]GTPS binding from agonist-stimulated binding, according to previously published calculations (15, 16, 19).¹ The resulting Scatchard plot of net-stimulated binding consisted of a single class of high affinity [³⁵S]GTPS binding sites (Fig. 4B, inset, and Table 3). This type of analysis is not only simpler than biphasic Scatchard modeling of the data but also statistically valid because there was no significant difference between K_D and B_max values of agonist-stimulated high affinity [³⁵S]GTPS binding obtained from this monophasic Scatchard analysis of net agonist-stimulated [³⁵S]GTPS binding and those obtained by biphasic Scatchard analysis of absolute agonist-stimulated [³⁵S]GTPS binding (Table 3).¹

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Homologous displacement and Scatchard analysis of [35S]GTPS binding in mMOR-CHO cell membranes. Membranes were incubated with 0.1 nM [35S]GTPS, 10 µM GDP, and 0.5-2000 nM unlabeled GTPS in the presence and absence of 5 µM DAMGO. Data shown are (A) mean percent of control [35S]GTPS binding (binding measured in the absence of agonist or unlabeled GTPS) ± standard error and (B) Scatchard analysis of basal and agonist-stimulated [35S]GTPS binding calculated from a typical experiment that was performed in triplicate and replicated four times. Inset, Scatchard analysis of net DAMGO-stimulated [35S]GTPS (stimulated minus basal binding measured at each concentration of GTPS) calculated from a typical experiment that was performed in triplicate and replicated four times. Control [35S]GTPS binding was 61.1 ± 1.4 fmol/mg of protein. KD and Bmax values from Scatchard analyses are given in Table 3.

View this table: [in this window] [in a new window]   TABLE 3 Scatchard analysis of basal and DAMGO-stimulated [35S]GTPS binding to mMOR-CHO cell membranes Membranes were incubated with 0.01 nM [35S]GTPS and 10 µM GDP in the presence and absence of 1 µM DAMGO. Data are mean ± standard error of KD and Bmax values from biphasic Scatchard analysis of absolute [35S]GTPS binding and from monophasic Scatchard analysis of net DAMGO-stimulated [35S]GTPS binding.

View larger version (32K): [in this window] [in a new window]   Fig. 5.   Homologous displacement of [35S]GTPS binding in rat thalamic and mMOR-CHO cell membranes. Membranes prepared from (A) rat thalamus or (B) mMOR-CHO cells were incubated with 0.1 nM [35S]GTPS and 50 µM (rat thalamus) or 10 µM (mMOR-CHO) GDP and 0.1-50 nM unlabeled GTPS in the presence and absence of opioid agonists. Agonist concentrations were 5 µM DAMGO, 5 µM fentanyl, 10 µM morphine, and 0.03 µM buprenorphine (rat thalamus) and 1 µM DAMGO, 1 µM fentanyl, 5 µM morphine, 0.03 µM buprenorphine, and 0.1 µM levallorphan (mMOR-CHO). Data are mean percent of control [35S]GTPS binding (binding measured in the absence of agonist or unlabeled GTPS) ± standard error. Control [35S]GTPS binding was 313 ± 30 and 68.9 ± 3.8 fmol/mg of protein in rat thalamic and mMOR-CHO cell membranes, respectively. KD and Bmax values from Scatchard analyses of these data are shown in Table 4.

View this table: [in this window] [in a new window]   TABLE 4 Scatchard analysis of net agonist-stimulated [35S]GTPS binding to rat thalamus, SK-N-SH, and mMOR-CHO membranes Membranes were incubated with 0.1 nM [35S]GTPS, 50 µM (rat thalamus), 30 µM (SK-N-SH), or 10 µM (mMOR-CHO) GDP and 0.01-50 nM GTPS in the presence and absence of opioid agonists (concentrations shown in legend to Fig. 5). Data are mean ± standard error of KD and Bmax values from Scatchard analysis of net agonist-stimulated [35S]GTPS binding.

	Discussion

Summary Introduction Procedures Results Discussion References

The results of the current study confirm those of Traynor and Nahorski (19), demonstrating that opioid agonists of various reported intrinsic efficacies have different abilities to maximally stimulate [³⁵S]GTPS binding to µ receptor-containing cell membranes. The results of the current study also show that similar efficacy differences for stimulation of [³⁵S]GTPS binding are observed both in membranes prepared from rat thalamus and (by [³⁵S]GTPS autoradiography) in rat brain sections at the level of the thalamus. The finding that qualitatively similar results have been obtained in three different tissues, SH-SY5Y (19) and SK-N-SH neuroblastoma cells, as well as rat thalamus, indicates that these efficacy differences are a property of the agonist/receptor interaction and not of the system under investigation. However, the quantitative differences observed among the different systems indicate that the tissue and/or conditions under which the experiment is conducted will determine how these efficacy differences are expressed in terms of a functional response. For example, the effect of GDP concentration on the magnitude of the differences in agonist efficacy that were observed in the [³⁵S]GTPS binding assays are quite profound. The finding that efficacy differences were magnified by increasing GDP concentrations was also an indication that there was a guanine nucleotide affinity component involved in the determination of agonist efficacy. These results demonstrated that agonists of higher efficacy were better able to overcome the driving force of excess GDP to "turn off," or prevent the "turn on," of receptor-activated G proteins. The catalytic activation component of agonist efficacy may also be affected by GDP concentration: Because GDP stabilizes G proteins in the inactive, holotrimeric state (37, 38), it is possible that the catalytic rate of receptor/G protein activation is slowed by excess GDP. Agonists of higher efficacy may be less affected by this stabilizing effect of GDP in their ability to catalytically activate G proteins. It is also clear that the receptor density ratio is an important factor in the determination of both absolute and relative agonist efficacy, as demonstrated by the results obtained in mMOR-CHO cell membranes, which contain an overabundance of µ-opioid receptors compared with neural tissues (34).

The mechanisms of µ-opioid agonist efficacy were explored in this study using Scatchard analysis of agonist-stimulated [³⁵S]GTPS binding. However, there are both advantages and disadvantages to this type of analysis. It is important to note that Scatchard analysis of [³⁵S]GTPS binding actually measures the competition of a radiolabeled ligand ([³⁵S]GTPS) for a nonlabeled ligand (GDP) under nonequilibrium conditions. These conditions are necessary to measure agonist-stimulated binding; at equilibrium, the [³⁵S]GTPS would displace as much of the GDP as it is capable of displacing, and no agonist-stimulated binding could then be observed. Therefore, this type of analysis is not quantitatively accurate in the sense that a given B_max represents the exact maximal number of G proteins that are capable of being activated by the agonist-occupied receptor. However, this type of analysis is useful for relative comparisons. Previous studies have demonstrated the usefulness of Scatchard analysis of [³⁵S]GTPS binding to examine catalytic activation of G proteins (15), differences in the ability of different receptor types to catalytically activate G proteins (16), and changes in the catalytic activation of G proteins during agonist-induced receptor desensitization.¹ Thus, Scatchard analysis of [³⁵S]GTPS binding is useful for examining relative measurements of agonist activity, such as with agonists of different efficacies, as in the current study. One problem with using this type of analysis to compare agonists of different efficacies is that the Scatchard analysis was performed at high GDP concentrations, which maximizes efficacy differences among the different agonists but considerably reduces [³⁵S]GTPS binding and potentially decreases the accuracy of B_max determinations. Nevertheless, significant differences between B_max values were obtained with DAMGO versus morphine or fentanyl in either rat thalamus or SK-N-SH cell membranes, which validates the conclusion that differences in catalytic activation of G proteins contribute to the determination of agonist efficacy (see below).

Scatchard analysis of basal and agonist-stimulated [³⁵S]GTPS binding confirmed that agonists of higher efficacy produced a higher affinity GTP-binding state, presumably in the guanine nucleotide binding site of µ receptor-coupled G protein  subunits, than agonists of lower efficacy. This was most evident when comparing the agonist of lowest efficacy, buprenorphine, with the agonist of highest efficacy, DAMGO, for which significant differences in the agonist-stimulated [³⁵S]GTPS binding affinity were observed in both SK-N-SH and rat thalamus membranes. Among the agonists of higher efficacy, DAMGO, fentanyl, and morphine, differences in the agonist-induced guanine nucleotide affinity shifts were less evident. However, differences in catalytic activation of G proteins, as measured by the agonist-stimulated B_max of [³⁵S]GTPS binding, were apparent among the higher efficacy agonists. Although the differences between B_max values obtained with fentanyl and morphine versus those obtained with DAMGO were relatively small (15-30% lower than DAMGO) compared with the differences between DAMGO and buprenorphine (50-60% lower), statistical significance was achieved with at least one of the two agonists in both SK-N-SH and rat thalamus membranes. Thus, opioid agonist efficacy seems to be determined by at least two signal transduction components: differences in the ability to promote receptor-stimulated guanine nucleotide exchange on the G protein  subunits (K_D) and differences in the number of G proteins activated per occupied receptor (B_max). These data suggest that there are actually three classes of agonists: 1) full agonists (e.g., DAMGO), which induce a maximally high affinity GTP binding state and also catalytically activate a maximal number of G proteins; 2) classic partial agonists (e.g., buprenorphine and levallorphan), which induce a lower affinity GTP binding state and (possibly by virtue of the lower probability that a given G protein will bind GTP and thus become activated) also catalytically activate a lower number of G proteins; and 3) mixed full/partial agonists (high efficacy partial agonists; e.g., fentanyl and morphine), which induce the same high affinity GTP binding state as full agonists but may catalytically activate a lower number of G proteins. This hypothesis predicts that classic partial agonists, such as buprenorphine and levallorphan, would be partial agonists in any cell type but that the efficacy of the mixed full/partial agonist, such as morphine, would depend on additional tissue-specific factors, such as the ratio of receptors to G proteins. In systems where the receptor number is high relative to the pool of G proteins available for activation by the receptor, such agonists may seem to be full agonists.

This hypothesis was confirmed by the experiments performed in mMOR-CHO cell membranes. These cells have been reported to contain µ receptors at a density of 7.5 pmol/mg of membrane protein (34) compared with 0.15 pmol/mg in SK-N-SH cell membranes¹ and ~0.1 pmol/mg in rat thalamic membranes (39). In membranes from mMOR-CHO cells, DAMGO stimulated [³⁵S]GTPS binding by >400% above basal, as did morphine and fentanyl. The apparent lack of an efficacy difference for G protein activation by the high efficacy agonists in this cell line is likely due to the high receptor density compared with the density of G proteins available for activation because the B_max of agonist-stimulated [³⁵S]GTPS binding was in the 3-5 pmol/mg range for the "full" agonists. Only those agonists that did not fully shift the affinity of the G proteins into a GTP-preferring state (i.e., buprenorphine and levallorphan) showed decreased efficacy in this system, although they did show increased relative efficacy compared with rat thalamus or SK-N-SH cells. These results suggest that relative agonist efficacy depends on the receptor/G protein ratio as well as the activity state of the G proteins and the intrinsic ability of the agonist to activate the receptor. The findings of the present study may help to explain the differences in µ-opioid agonist intrinsic efficacy observed at the level of effectors (40, 41) and in behavioral and analgesic studies (22-26).