Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Activation of G protein-coupled receptors (GPCRs) commonly increases cytosolic free Ca²⁺ (Ca²⁺i). Both Gq/11- and Gi/Go-coupled receptors trigger the release of Ca²⁺ from intracellular stores upon activation of phospholipase C and formation of inositol phosphates. Opening of store-operated channels (SOCs; or capacitative Ca²⁺ entry), which allows extracellular Ca²⁺ (Ca²⁺e) to flow into the cell and replenish the stores (Clapham, 1995), follows this event. Moreover, GPCRs modulate ion channels in the plasma membrane via the release of soluble second messengers (Clapham, 1995). Last, receptors can activate ion channels directly by a membrane-delimited pathway (receptor operated Ca²⁺ channels) (Schaefer et al., 2000). Genes encoding the transient receptor potential family of cation channels seem to play a role in several of these processes (Schaefer et al., 2000; Zhang and Saffen, 2001). Each GPCR thus activates multiple pathways, both direct membrane delimited and indirect via second messengers.

Recent evidence supports the view that GPCRs exist in multiple conformational states that could trigger distinct signaling pathways. For example, octopamine and tyramine each stimulate a separate signaling pathway at their common receptor in Drosophila (Robb et al., 1994). Moreover, an activating mutation of the _1B-adrenergic receptor selectively stimulates only one of two _1B signaling pathways examined (Perez et al., 1996). Similarly, structurally distinct ligands differentially activate Gi and Go coupling of cannabinoid receptors (Houston and Howlett, 1998). Last, numerous ligand binding studies have revealed the existence of multiple receptor conformations (Wreggett and Wells, 1995; Brys et al., 2000). Because conformational receptor states vary with experimental conditions, it is difficult to link distinct conformational states with specific receptor functions induced by different ligands.

We have developed a rapid approach for measuring distinct Ca²⁺i pathways in response to GPCR activation, to determine whether different ligands might affect these pathways differentially at the same receptor. Herein, we have investigated the Ca²⁺i response to stimulation of the Gi/Go-coupled µ-opioid receptor (MOR), transfected into human embryonic kidney (HEK) 293 cells. Opioid receptors couple to PTX-sensitive G proteins, thereby activating inwardly rectifying K⁺ channels and inhibiting voltage-sensitive Ca²⁺ channels and adenylyl cyclases (North et al., 1987; Murthy and Makhlouf, 1996; Piros et al., 1996). However, Gi/Go-coupled opioid receptors also activate phospholipase C and phosphatidyl inositol turnover (Murthy and Makhlouf, 1996) and elevate intracellular Ca²⁺ levels in neuronal and non-neuronal tissues (Connor and Henderson, 1996; Hauser et al., 1996; Tang et al., 1996). This occurs either by stimulating influx of extracellular Ca²⁺e or release of intracellular Ca²⁺i stores via soluble second messengers (inositol trisphosphates), or both. Several GPCRs induce rapid Ca²⁺ influx independent of intracellular Ca²⁺ release (Felder et al., 1992; Montero et al., 1994). Smart et al. (1995) reported that Ca²⁺ influx preceded intracellular Ca²⁺ release upon MOR stimulation in human SH-SY5Y neuroblastoma cells shown to express MOR and -opioid receptors (Yu et al., 1986). However, it remains unclear to what extent receptor-operated Ca²⁺ channels contribute to the Ca²⁺i response of MOR. Whereas MOR agonists predominantly inhibit neuronal activity by modulating potassium channels and voltage-dependent Ca²⁺ channels, excitatory opioid effects have also been noted. For example, opioid enhancement of evoked enkephalin release in guinea pig myenteric plexus was found to involve elevation of intracellular Ca²⁺ levels (Xu and Gintzler, 1992). In the rat locus coeruleus, morphine did not simply decrease firing rates of LC neurons, but it also induced persistent oscillatory discharges (Zhu and Zhou, 2001). Therefore, opioid-dependent Ca²⁺ influx pathways could contribute these effects.

Opioid ligands elicit distinct effects at MOR. Different MOR agonists vary dramatically in their ability to induce receptor internalization (Arden et al., 1995; Keith et al., 1996). DAMGO and etorphine, but not morphine, were shown to cause receptor internalization, even though all three strongly stimulate G protein coupling. This distinguishes receptor forms active in coupling and internalization. Furthermore, various agonists couple the opioid receptor to a different spectrum of G proteins (Chakrabarti et al., 1995; Allouche et al., 1999), but our results failed to show significant differences in the activation of individual G proteins (Burford et al., 1998). Last, some opioid agonists, such as etorphine and dihydroetorphine, cause less dependence/withdrawal than morphine (Tokuyama et al., 1994). The molecular mechanism underlying these distinct effects remains unknown.

To examine Ca²⁺ signaling pathways of GPCRs, whole cell fluorescence measurements were performed by measuring rapid Ca²⁺i responses using a plate reader format (Lin et al., 1999). HEK293 stably cells transfected with recombinant MOR (HEK-µ) were used as a model for testing possible links between GPCR conformation and signaling pathways. We demonstrate that MOR activation leads to a multiphasic Ca²⁺ response that is differentially affected by the presence or absence of Ca²⁺e, by temperature, and by different classes of opioid ligands.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Materials. [³⁵S]Guanosine-5'-O-(3-thio)triphosphate was obtained from PerkinElmer Life Sciences (Boston, MA) and Calcium Green 1-AM from Molecular Probes (Eugene, OR). All other reagents were from Fisher Scientific (Fair Lawn, NJ), unless noted otherwise. Morphine, etorphine, naloxone, diprenorphine, and the opioid peptide DAMGO were obtained from the National Institute on Drug Abuse (Bethesda, MD).

Tissue Culture. HEK293 cells, untransfected or transfected with cDNA encoding recombinant rat MOR (HEK-µ) (Arden et al., 1995; Lin et al., 1999), expressing approximately 2 pmol of [³H]diprenorphine sites per milligram of protein, were grown in T-175-cm³ tissue culture flasks containing 30 ml of HEK293 medium (a 1:1 mixture of Dulbecco's modified Eagle's medium and Ham's F-12 medium, supplemented with 10% fetal bovine serum and 200 µg/ml G418). To test the requirement for Gi/Go in the Ca²⁺ response, we pretreated the cells with 300 ng of PTX for 3 h at 37°C. This protocol blocks the coupling of MOR to Gi/Go, determined with the [³⁵S]guanosine-5'-O-(3-thio)triphosphate binding assay described below.

Ca²⁺ Measurements. Intracellular free Ca²⁺i levels were measured according to the procedure described by Lin et al. (1999). Briefly, HEK293 or HEK-µ cells were grown to 80% confluence, harvested with trypsin, and pelleted at 300g for 2 min. The pelleted cells were then resuspended in fresh HEK293 medium and allowed to recover for 1 to 2 h in a tissue culture incubator at 37°C under 5% CO₂ and humidified atmosphere. Cells were then removed from the incubator and prepared for Ca²⁺ measurements by pelleting and resuspending twice in Krebs-HEPES buffer, pH 7.4, containing 118 mM NaCl, 4.8 mM KCl, 1.2 mM MgSO₄, 1.2 mM KH₂PO₄, 1.3 mM CaCl₂, 11.7 mM D-glucose, and 10 mM HEPES-free acid. After incubation in the same buffer with 3 mM Calcium Green 1-AM at room temperature for 30 min, the cells were rinsed three times with the above-mentioned buffer containing 0.5% (w/v) bovine serum albumin (Sigma-Aldrich, St. Louis, MO), diluted to approximately 200,000 cells/ml, and plated (~30,000 cells/well) into opaque white 96-well plates (Costar, Cambridge, MA). To test for any requirement of Ca²⁺e, similar sets of experiments were conducted with the same buffer but containing the indicated concentrations of CaCl₂ and 10 µM BAPTA. The addition of BAPTA to the Ca²⁺-free buffer had no effect per se on indicator dye fluorescence.

cAMP Assays. cAMP accumulation was stimulated with 10 µM forskolin for 5 min (Arden et al., 1995). Assays were performed either in cell suspensions using the same buffer as for the Ca²⁺ assay, or by using attached monolayers as described previously (Yu et al., 1986). Opioid agonists were added in varying concentrations along with forskolin, to establish dose-response curves for inhibition of cAMP accumulation.

Ligand Binding. [³H]Naloxone binding to MOR in HEK-µ cells was determined under two experimental conditions, i.e., competition binding by incubation over 30 min, or measuring initial tracer binding rates over 10 s. Equilibrium binding analysis was performed at room temperature either with intact suspended cells (using similar conditions described for the Ca²⁺ assay) or with cell membranes prepared and incubated in Tris-HCl, pH 7.4, buffer as described previously (Yu et al., 1986; Burford et al., 1998), at 25 or 4°C. The tracer was incubated together with competing ligands for 30 min.

Data Analysis. Concentration-response curves and standard errors were analyzed by fitting experimental observations to a logistic equation, defined by DeLean et al. (1978), using SigmaPlot curve-fitting software (SPSS Inc., Chicago, IL).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Ca²⁺ Response to Morphine in HEK-µ Cells. Injection of 1 µM morphine into a suspension of HEK-µ cells evoked a fluorescence peak at 5 to 7 s (peak 1) followed by a sustained elevated plateau, indicating an increase in intracellular free cytosolic Ca²⁺i capable of complexing with Calcium Green 1-AM (Fig. 1A) (Lin et al., 1999). The response to control buffer injections was subtracted as reported previously (Lin et al., 1999). No response to morphine was seen in nontransfected (wild-type) HEK293 cells, or cells transfected with empty plasmid. Pretreatment of the cells with PTX eliminated the response to morphine, indicating that the Ca²⁺ response depended on activation of PTX-sensitive G proteins (Gi and Go) (Fig. 1A).

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Ca2+ fluorescence response in HEK-µ cells. A, morphine (1 µM) was injected at time 0 into suspensions (with or without 1.3 mM Ca2+) of HEK-µ cells labeled with Calcium Green 1-AM, and fluorescence was read each second. Control readings (no drug) were subtracted from sample readings at each time point. Each point represents the mean ± S.D. of four measurements per experiment. Each experiment was replicated at least twice. Solid circles, medium containing 1.3 mM Ca2+; open circles, Ca2+-free medium; and solid triangles, PTX-pretreated cells in medium with 1.3 mM Ca2+. B, fluorescence response to injections of 1 µM morphine into wells containing HEK-µ cells suspended in varying concentrations of Ca2+. The morphine solution used for injection (8% of total volume) contained 1 µM Ca2+. C, effect of temperature on the Ca2+ response to 1 µM morphine injections. All experiments were replicated at least twice, with similar results.

Opioid Agonist Concentration-Response Curves for Two Distinct Phases of Ca²⁺ Response in HEK-µ Cells. To obtain separate measures of peak 1 and peak 2, relative fluorescence intensities were determined by integration of the signal at 5 to 7 and 20 to 22 s, respectively, after addition of morphine, etorphine (Fig. 2, A-D), and DAMGO to HEK-µ cells, in the presence and absence of Ca²⁺e. The calculated EC₅₀ values are provided in Table 1. In the presence of Ca²⁺e, morphine had similar potency in eliciting both phases of the Ca²⁺ response (EC₅₀ of ~40 nM), whereas etorphine was 20-fold more potent in stimulating peak 2 (EC₅₀ of 0.14 nM; 20-22 s) than peak 1 (EC₅₀ of 2.9 nM; 5-7 s; p < 0.001). DAMGO was 3-fold more potent in stimulating peak 2 than peak 1 (p < 0.001) (Table 1).

View larger version (51K): [in this window] [in a new window]   Fig. 2.   Ca2+-fluorescence response to varying doses of morphine (A and C) and etorphine (B and D) in HEK-µ cells, in the presence (A and B) or absence (C and D) of 1.3 mM Ca2+. Each agonist was injected to yield concentrations of 1, 10, and 100 pM, and 1, 10, and 100 nM and 1 µM, as indicated. Arbitrary fluorescence units observed at 5 to 7 s and 20 to 22 s were averaged, and the resultant EC50 values are listed in Table 1.

Differential Effects of Naloxone and Diprenorphine on Opioid Agonist-Mediated Ca²⁺ Responses. We next determined the ability of naloxone (an oxymorphinan chemically similar to morphine) and diprenorphine (an oripavine chemically similar to etorphine) to block the agonist response. Both naloxone and diprenorphine (1-10 µM) added alone failed to cause a significant Ca²⁺ response, as expected for an antagonist (data not shown). Antagonists were either coinjected with the agonist or were preinjected 30 min before the agonist to permit equilibration with the receptor. The selected agonist concentrations corresponded to 2- to 4-fold the respective EC₅₀ values at 5 to 7 s (100 nM morphine, 10 nM etorphine, and 100 nM DAMGO; Table 1). Note, however, that 10 nM etorphine exceed its EC₅₀ at 20 to 22 s substantially. As with the agonist dose-response curves, effects of increasing concentrations of the antagonists were measured at 5 to 7 s and 20 to 22 s (representing peaks 1 and 2) (Fig. 3; Table 2). The resultant IC₅₀ values were compared between peaks 1 and 2 of the same run, and ratios were calculated for IC₅₀ peak 1/peak 2. This also permitted the comparison between the antagonists acting against each of the three agonists tested. Table 2 further contains statistical evaluations of relevant comparisons. The following results were calculated from one series of experiments (n = 4/data point) to permit comparison between different conditions in the same batch of cells. Replicate experiments gave similar results as indicated in the text and legends.

View larger version (49K): [in this window] [in a new window]   Fig. 3.   Representative antagonist dose-response curves for inhibiting the Ca2+ response to 100 nM morphine in HEK-µ cells. Antagonists (1 pM to 10 µM, in increments of 10) were either coinjected (A and B) or the antagonist was preincubated for 30 min (C). EC50 values are listed in Table 2. D, morphine was coinjected with 1 µM diprenorphine or naloxone in Ca2+-free medium. These experiments were repeated at least once with similar results.

View this table: [in this window] [in a new window]   TABLE 2 Antagonist dose-response curves for the inhibition of the Ca2+ response to 100 nM morphine, 10 nM etorphine, and 100 nM DAMGO The antagonists naloxone and diprenorphine were added either together with (co-injection) or 30 min before the agonist (30 min). The IC50 values (±S.D.) are provided for fluorescence measured at 5 to 7 s or 20 to 22 s after agonists injection (see Fig. 3 for fluorescence response profiles). The data in this table are derived from one series of experiments with the same batch of cells (n = 4 per data point). At least one repeat experiment was done for all conditions (except for DAMGO), yielding similar results.

Coinjection of Antagonist with Agonist. Shown in Fig. 3A, naloxone given together with morphine (100 nM) suppressed the Ca²⁺ response with IC₅₀ of 86 nM (5-7 s) and 32 nM (20-22 s) (Table 2). Naloxone displayed different potencies at peaks 1 and 2, against each agonists, with decreasing antagonist potency in the rank order DAMGO > morphine > etorphine (Table 2). In all repeat experiments, naloxone (1 µM or above) fully eliminated the Ca²⁺ response of the three agonists (100 nM morphine, 10 nM etorphine, and 100 nM DAMGO).

Preincubation of Antagonists. In a second series of experiments, we allowed the antagonists to equilibrate with the receptor before injecting the agonist 30 min later. This resulted in dramatically reduced IC₅₀ values calculated from concentration-response curves with naloxone and diprenorphine against morphine (100 nM) and etorphine (10 nM) (Table 2). In contrast to coinjection (Fig. 3B), diprenorphine was quite potent against morphine upon preincubation even at 5 to 7 s (0.48 nM; Fig. 3C; Table 2). However, unexpected differences persisted in relative potencies between naloxone and diprenorphine. Under the preincubation conditions, naloxone was paradoxically twice as potent as diprenorphine in blocking rapid Ca²⁺e influx stimulated by etorphine (peak 1; IC₅₀ of 0.16 nM versus 0.30 nM; Table 2) even though it has ~5-fold lower binding affinity for MOR than diprenorphine (Sadée et al., 1982). Moreover, both antagonists were significantly more potent inhibitors against etorphine at 5 to 7 s than at 20 to 22 s (P < 0.05, P < 0.01). In contrast, naloxone showed no preference for peaks 1 and 2 against morphine, whereas diprenorphine is less potent at peak 1 (opposite to what is observed against etorphine). These differences are statistically significant as indicated in Table 2.

Coinjection of Antagonists in Absence of Ca²⁺e. Because coinjection of diprenorphine failed to block peak 1 of the Ca²⁺ response to morphine (Fig. 3B), we tested whether the residual peak 1 seen with morphine + diprenorphine consists of Ca²⁺ influx or intracellular Ca²⁺ release, or both. In nominally Ca²⁺-free medium (Ca²⁺ concentration contributed as trace from all other reagents estimated at 1 µM), coinjection of diprenorphine (1 µM) fully blocked the Ca²⁺ response to morphine (100 nM) (Fig. 3D), in contrast to the results obtained in the presence of Ca²⁺e (Fig. 3B). As expected, naloxone fully blocked the morphine response regardless of Ca²⁺ levels in the medium (Fig. 3, A and D). These results indicate that the Ca²⁺ response to morphine + diprenorphine (Fig. 3B) entirely depends upon the presence of Ca²⁺e (peak 1), and therefore, seems to represent Ca²⁺e influx.

Agonist-Induced Inhibition of cAMP Accumulation. For comparison to the Ca²⁺ response, we also measured the potency of morphine and etorphine to inhibit forskolin-stimulated cAMP production under identical conditions in cell suspension (Table 1B). Morphine's EC₅₀ value (24 nM) did not differ significantly from that observed in the Ca²⁺ assay. Etorphine had similar potency in the cAMP assay (EC₅₀ of 1.6 nM) and in stimulating peak 1 Ca²⁺ response (2.9 nM) but was more potent in stimulating Ca²⁺i release (peak 2; EC₅₀ of 0.14 nM). We also measured the cAMP response in attached HEK-µ monolayers to determine whether cell suspension as used for the Ca²⁺ assay affected the results. EC₅₀ values were similar, indicating that cell suspension had marginal effects on the cAMP response (data not shown).

Ligand Binding Studies. For comparison to functional responses, we first used intact suspended cells incubated under similar conditions as for the Ca²⁺ assays. In three sets of experiments, the tracer [³H]naloxone and competing unlabeled ligands were incubated with the cells either for 10 s or for 30 min. Over 10 s, the tracer occupied approximately 15% of the MOR binding sites. When the competing ligand was added for 30 min, the shape of tracer displacement curves was identical regardless of whether the tracer was added also for 30 min or only for 10 s at the end of the incubation (Fig. 4). This result indicates that the short tracer incubation, designed to measure initial binding rates over 10 s, correctly assessed the receptor population remaining unoccupied after 30-min equilibration with unlabeled ligand. The IC₅₀ values for the unlabeled ligands added 30 min before the tracer (10 s) are provided in Table 3A.

View larger version (28K): [in this window] [in a new window]   Fig. 4.   [3H]Naloxone binding curves in the presence of unlabeled ligands in intact HEK-µ cell suspensions. The tracer was either incubated for 10 s together with the unlabeled ligand (10-s coincubation) or the unlabeled ligand was added 30 min before the 10-s tracer incubation (30-min preincubation). In a third experiment, both tracer and unlabeled ligand were incubated together for 30 min (30-min coincubation). All experiments were performed at room temperature and repeated twice. The calculated IC50 values and slope factors are given in Table 3A.

View this table: [in this window] [in a new window]   TABLE 3 [3H]Naloxone (0.5 nM) binding to HEK-µ intact cells (A) and cell membranes (B) A. [3H]Naloxone was incubated for 10 s with the intact cells suspended in the same buffer used for the Ca2+ and cAMP assays. Competing unlabeled drugs were added either together with the tracer for the 10-s incubation or they were added 30 min before the tracer to permit equilibration of unlabeled ligand to occur. IC50 values (±S.D.) and slope factors are calculated from the displacement curves shown in Fig. 4. Calculated values with 30-min incubations of both tracer and displacer together (not included here; see Fig. 4) were indistinguishable from the results obtained with the 30-min displacer/10-s tracer incubations. All experiments were performed at room temperature. B. Using cell membrane incubations, prepared as described (Burford et al., 1998), EC50 values and slope factors n were calculated from competition binding curves of [3H]naloxone incubated together with the unlabeled ligand for 30 min at 4°C or room temperature (RT) (see Figure 6). In all experiments, ligand concentrations ranged from 10 pM to 10 µM (in 3- to 3.3-fold intervals, duplicate values at each point, n = 3).

View larger version (9K): [in this window] [in a new window]   Fig. 5.   [3H]Naloxone binding to intact HEK-µ cells at room temperature in the absence (solid circles) or the presence of unlabeled ligands added either together with the tracer (time 10 s) or before the tracer at the indicated times (open circles). Unlabeled ligands were 30 nM morphine (A), 5 nM naloxone (B), and 3 nM diprenorphine (C). The experiment was replicated twice with similar results.

View larger version (12K): [in this window] [in a new window]   Fig. 6.   Equilibrium binding curves measuring the competition between [3H]naloxone and unlabeled naloxone (A) or diprenorphine (B) in HEK-µ membranes. Incubations were done for 30 min at 4 or 25°C. The calculated IC50 values and slope factors are provided in Table 3B. Each experiment was replicated twice..

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Morphine stimulated a multiphasic Ca²⁺ response in HEK-µ cells. This heterologous cell culture served as a test model for generic GPCR-Ca²⁺ signaling, because we have observed similar multiphasic Ca²⁺ responses with muscarinic receptors endogenous to HEK293 cells, and with HEK293 cells transfected with the dopamine D2 receptor (unpublished data). Different cell types, and specifically neuronal cells, may display different Ca²⁺ signaling patterns. Previous studies with a variety of cells had shown that both Gi/Go- and Gq/11-coupled receptors stimulate phospholipase C, resulting in the release of Ca²⁺ from intracellular stores and subsequent capacitative Ca²⁺ entry via SOCs. Herein, we demonstrate the presence of an early phase of the Ca²⁺ response, which seems to represent rapid influx of extracellular Ca²⁺ (peak 1). Ca²⁺e influx before, or independent of, Ca²⁺i release had been shown to occur upon activation of muscarinic receptors and other GPCRs (Felder et al., 1992; Montero et al., 1994), including MOR in SH-SY5Y cells (Smart et al., 1995). However, a direct experimental demonstration of rapid induction of Ca²⁺ influx by MOR was lacking.

Several findings support the view that fluorescence peak 1 represents rapid Ca²⁺ entry and peak 2 intracellular Ca²⁺ release as distinct responses. First, deletion of Ca²⁺ from the medium abolished the rapid Ca²⁺ response (peak 1) while leaving intracellular Ca²⁺ release unchanged (peak 2). Second, coinjection of morphine and diprenorphine seemed to activate exclusively Ca²⁺ entry, yielding peak 1 only and enabling one to study this signaling pathway separately. Third, peaks 1 and 2 were differentially sensitive to temperature: peak 1 disappeared at temperatures below 30°C, whereas peak 2 declined only when the temperature dropped below 20°C. These results support the hypothesis that peaks 1 and 2 are two distinct response pathways, but the molecular basis of the Ca²⁺ currents remains to be resolved.

Distinct Ligand Effects at MOR. We tested whether chemically distinct ligands differentially affect different phases of the Ca²⁺ and cAMP responses. Significant differences were observed for peaks 1 and peak 2 of the Ca²⁺ response between the agonists morphine, etorphine, and DAMGO. Moreover, the antagonists naloxone and diprenorphine also displayed substantial differences in blocking these Ca²⁺ signals. cAMP measurements were taken over longer time periods (5 min) than Ca²⁺ signals (<1 min), allowing for more equilibration time, which could have affected apparent ligand potency by several mechanisms. This impedes a direct comparison of relative potencies in different signaling pathways. Nevertheless, morphine displayed similar potencies for the Ca²⁺ and cAMP responses, whereas etorphine had similar potency for cAMP (observed over 5 min) and the peak 1 Ca²⁺ response (5-7 s), but was much more potent for the peak 2 response (15-17 s), arguing against time factors accounting for these differences. Yet, different levels of receptor activation may have been required for Ca²⁺ and cAMP responses. To avoid these confounding factors, we focused on comparing peak 1 (thought to represent rapid Ca²⁺e influx) and peak 2 (Ca²⁺i release) over relatively short time periods in direct comparison with ligand binding. However, ligand-receptor binding is not in equilibrium, which must be considered in the interpretation. This approach enables one to test whether different receptor conformations are associated with different response pathways, using observations over relatively short time periods. Our result suggests the following hypotheses: 1) effective receptor on-rates differ greatly among the ligand tested, or 2) MOR exists in several conformations that activate separate signaling pathways.

Different On-Rates for Oxymorphinans and Oripavines. The early Ca²⁺ response (peak 1) rapidly desensitized (Fig. 1B); therefore, varying on-rates could strongly affect the Ca²⁺ response. Hence, the observed 3-fold lower potency of DAMGO and 30-fold lower potency of etorphine in stimulating peak 1 compared with peak 2 could have resulted from slower on-rates for these ligands compared with morphine. Similarly, inability of coinjected diprenorphine to block peak 1 stimulated by morphine might be related to a slower receptor on-rate compared with morphine and naloxone. Consistent with this hypothesis, diprenorphine's potency against morphine increased dramatically upon preincubation, indicating slow access to or equilibration with MOR sites. Recent evidence suggests that essential molecules of GPCR signaling pathways are held in proximity of each other in microdomains such as caveoli and do not depend upon random collision to interact (Ostrom et al., 2000). Therefore, access of ligands to receptor microdomains may differ between oxymorphinans and the more lipophilic oripavines. However, the polar peptide DAMGO had properties intermediate to those of morphine and the more lipophilic etorphine, suggesting that access to microdomains may not play a role.

Multiple MOR Conformations. Distinct receptor conformations/complexes could trigger different signaling pathways. Target size analysis of GPCRs in the plasma membrane has revealed large GPCR complexes exceeding 1 × 10⁶ Da, which partially break up upon agonist stimulation (Rodbell, 1992). Receptor aggregation as a main organizing principle could lead to oligomeric receptors (George et al., 2000) shown to affect receptor functions (Jones et al., 1998; Li et al., 2002). Specifically, GPCRs exist in physical contact with ion channels, as shown for complexes between dopamine D5 receptors and -aminobutyric acid-A (Liu et al., 2000), and 2 receptors and Ca_v1.2 Ca²⁺ channels (Davare et al., 2001). Therefore, different ligands could induce distinct signaling pathways at the same receptor.

Pharmacological Significance of Distinct Ligand Effects at MOR. Although the present study was performed with MOR transfected into a non-neuronal cell line, our results demonstrate the principle that different MOR agonists can cause significantly different Ca²⁺ response patterns. Ca²⁺ appears to play a major role in the development of tolerance and dependence (Sanghvi and Gershon, 1976), involving voltage-sensitive Ca²⁺ channels (Tokuyama et al., 1995), N-methyl-D-aspartate (Trujillo and Akil, 1991), and -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors (Carlezon et al., 1997), and calmodulin (Mestek et al., 1995), which interacts directly with MOR and serves as a regulator of G protein coupling and as a second messenger per se (Wang et al., 1999, 2000). Hence, the observed differences in Ca²⁺ signaling between morphine and etorphine could contribute to differences in their pharmacological effects. Indeed, etorphine produces less dependence in animal studies (Tokuyama et al., 1994). The finding that N-methyl-D-aspartate antagonists selectively affect antinociceptive tolerance to morphine, but not to other opioid agonists tested (Bilsky et al., 1996), suggests different tolerance mechanisms may exist for different agonists. Distinct signaling pathways triggered by different opioid ligands could contribute to their varying pharmacological properties.