Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Opioids produce their physiological effects by acting upon opioid receptors. Opioid receptors belong to the superfamily of G-protein-coupled receptors (GPCRs), and each of the three opioid receptor subtypes (µ, , and ) are coupled to heterotrimeric G-proteins (Standifer and Pasternak, 1997). Opioids find clinical utility as potent analgesic agents and are especially effective at treating nociceptive pain. Opioid analgesia was long thought to be mediated primarily through activation of µ-opioid receptors, as the classic opioid agonist, morphine, demonstrated high affinity for the µ-opioid receptor (Reisine and Pasternak, 1996). However, studies using agonists selective for the -opioid receptor also provided evidence of -opioid receptor-mediated analgesia. For example, one of the first selective -opioid agonists, [D-Pen²,D-Pen⁵]-enkephalin (DPDPE), produced analgesia following either intracerebroventricular or intrathecal administration in mice (Mosberg et al., 1983; Porreca et al., 1984). Other selective -opioid agonists, such as [D-Ser²,Leu⁵,Thr⁶]-enkephalin and [D-Thr²,Leu⁵,Thr⁶]-enkephalin, provided further evidence of -opioid receptor-mediated antinociception, and neither compound demonstrated cross-tolerance to morphine in murine analgesic assays (Porreca et al., 1987). Additional supporting evidence for -opioid analgesia came from the development of the highly selective -opioid receptor antagonist, ICI-174864 (N,N-diallyl-Tyr-Aib-Aib-Phe-Leu-OH, where Aib = -aminoisobutyric acid) (Cotton et al., 1984; Porreca et al., 1987). ICI-174864 blocked the analgesic effects of DPDPE but not those of the µ-opioid agonists morphine or [D-Ala²,N-MePhe⁴,Gly-ol⁵]-enkephalin, in the mouse tail-flick assay (Heyman et al., 1987). Thus, highly selective -opioid receptor ligands have proven to be invaluable aids in delineating the role of the -opioid receptor in opioid-mediated analgesia.

More recently, a new class of selective -opioid receptor antagonists was introduced termed the TIP(P) peptides. The prototype antagonist of this class was TIPP (H-Tyr-Tic-Phe-Phe-OH, where Tic = 1,2,3,4-tetrahydroisoquinoline), which demonstrated high affinity for -opioid receptors (i.e., subnanomolar) and was found to be a potent -opioid receptor antagonist in the mouse vas deferens bioassay (Schiller et al., 1992). TIPP displayed no agonist properties at either µ- or -opioid receptors when concentrations as high as 10 µM were tested (Schiller et al., 1992, 1999a). Although TIPP is classified as a selective -opioid antagonist, a recent study by our laboratory reported agonist activity for this compound (Martin et al., 2001). TIPP inhibited adenylyl cyclase activity in GH₃ cells stably expressing epitope-tagged -opioid receptors (GH₃DORT), as well as in N1E115 neuroblastoma cells, NG-108-15 neuroblastoma × glioma hybrid cells, Chinese hamster ovary-DOR, and human embryonic kidney-DOR cells. This novel agonist activity was shown to be -opioid receptor-mediated, concentration-dependent, and independent of receptor density. In addition, the agonist activity of TIPP was sensitive to blockade by pertussis toxin pretreatment, indirectly implying the involvement of G-proteins of the Gi/Go subtype. The efficacy of inhibition produced by TIPP was not significantly different from that produced by the full -opioid agonist, DPDPE. Importantly, these results are inconsistent with all previously reported studies confirming TIPP as a selective -opioid receptor antagonist. In particular, separate studies directly examining G-protein activation reported that TIPP failed to stimulate guanosine 5'-O-(3-[³⁵S]thio)triphosphate ([³⁵S]GTPS) binding or GTPase activity, key markers of G-protein activation (Mullaney et al., 1996; Szekeres and Traynor, 1997).

Since the finding of agonist activity by TIPP in various cellular models is inconsistent with over a decade of literature supporting TIPP as a purely selective -opioid receptor antagonist, it is conceivable that TIPP might represent a novel class of opioid agonists with unique properties. Therefore, the purpose of this study was to further investigate the agonist activity of TIPP in GH₃DORT cells. This was accomplished by examining the properties of TIPP at several points along the signal transduction pathway (i.e., receptor binding, G-protein activation, and effector regulation) and comparing these properties with those of the selective -opioid agonist, DPDPE. At all points examined in the signaling pathway, DPDPE demonstrated only the anticipated agonist properties. In contrast, TIPP, in the presence of a guanine nucleotide analog and sodium ions, demonstrated increased affinity for -opioid receptors, which is characteristic of an inverse agonist. In addition, TIPP failed to activate G-proteins, as measured by [³⁵S]GTPS binding, in GH₃DORT membrane preparations; yet, G-protein activation was detected in digitonin-permeabilized cells. Finally, TIPP maintained its ability to inhibit adenylyl cyclase activity in GH₃DORT membrane preparations and in whole cells, providing further evidence of its capability to act as an agonist. These results demonstrate that TIPP exhibits characteristics of an agonist, antagonist, or inverse agonist at -opioid receptors, depending upon the assay conditions and response being measured. Therefore, TIPP and structurally related compounds may possibly represent a new class of selective -opioid receptor agonists that inhibit adenylyl cyclase activity using very different mechanisms than other commonly used -opioid agonists.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. Penicillin/streptomycin (10,000 IU/ml and 10,000 µg/ml), fetal calf serum, and Dulbecco's modified Eagle's medium containing 4.5 g of glucose, L-glutamine, and pyruvate were purchased from Mediatech Cellgro (Herndon, VA). Hygromycin-B and digitonin were supplied by Calbiochem (San Diego, CA). DPDPE was obtained from Peninsula Laboratories (Belmont, CA), and TIPP was purchased from Phoenix Pharmaceuticals, Inc. (Belmont, CA). Naltriben, naloxone, and ICI-174864 were procured from Tocris Cookson, Inc. (Ballwin, MO). [³H]Diprenorphine (56 Ci/mmol) and [³⁵S]GTPS (1250 Ci/mmol) were purchased from PerkinElmer Life Sciences (Boston, MA). [8-³H]Adenine (26 Ci/mmol) was obtained from Amersham Biosciences (Piscataway, NJ). All other reagents were purchased from Fisher Scientific (Pittsburgh, PA).

Cell Culture. GH₃ cells stably transfected with cDNA encoding for -opioid receptors with a hemagglutinin epitope tag spliced at the N terminus (GH₃DORT) were used, as previously described (Martin et al., 2001). GH₃DORT cells were maintained in Dulbecco's modified Eagle's medium supplemented with NaHCO₃ (3.7 g/l), 10% (v/v) fetal calf serum, 100 units/ml penicillin, 100 µg/ml streptomycin, and 200 µg/ml Hygromycin-B. Cells were incubated at 37°C in a humidified atmosphere of 5% CO₂/95% air. Confluent cells were harvested with a 10 mM phosphate-buffered saline solution containing 1 mM EDTA, pH 7.4. Harvested cells were then centrifuged (1000 rpm, 4°C, 10 min), and the pellets were either used immediately or stored at 80°C for future use.

Membrane Preparation. GH₃DORT membranes were prepared as described previously (Martin et al., 2001) with slight modifications. Briefly, harvested cell pellets were thawed on ice and resuspended in ice-cold homogenization buffer consisting of 50 mM HEPES, pH 7.4, 3 mM MgCl₂, and 1 mM EGTA. Pellets were then homogenized using a glass Dounce homogenizer and pestle A (Wheaton, Philadelphia, PA). The cell homogenates were centrifuged at 40,000g for 10 min at 4°C, the supernatant was discarded, and the resultant pellet resuspended in the original volume of homogenization buffer. The procedure was repeated twice more, and the partially purified membrane pellet was resuspended at 10% of the original volume in 20 mM HEPES, pH 7.4, or 50 mM Tris, pH 7.4. The protein concentration of the partially purified membranes was determined, and aliquots were stored at 80°C.

Cell Permeabilization. Permeabilization of GH₃DORT cells was performed as previously described (Alt et al., 2001) with the following modifications. Harvested GH₃DORT cells were washed with ice-cold Krebs-Ringer-HEPES buffer, pH 7.4 (KRHB), containing 110 mM NaCl, 25 mM glucose, 55 mM sucrose, 10 mM HEPES, 5 mM KCl, 1 mM MgCl₂, and 1.8 mM CaCl₂. Cells were centrifuged at 1000 rpm for 10 min and then resuspended in KRHB containing 80 µM digitonin. The resuspended cells were incubated for 10 min at 0°C while gently shaking. Following the incubation period, cells were centrifuged, washed twice with ice-cold KRHB, and resuspended in assay buffer (i.e., 50 mM Tris, pH 7.4 or 20 mM HEPES, pH 7.4) at a final concentration of 0.5 or 1 × 10⁶ cells per assay.

Saturation and Competition Binding. Saturation and competition binding assays were performed in 50 mM Tris, pH 7.4, at room temperature using a 90-min incubation period, as described previously (Martin et al., 2001). For saturation binding studies, 0.05 to 10 nM [³H]diprenorphine was incubated with membranes (100 µg/assay) or permeabilized cells (1 × 10⁶ cells/assay). Nonspecific binding was determined in the presence of 10 µM naloxone. For competition binding assays, the ability of DPDPE, ICI-174864, or TIPP (10¹²-10⁵ M) to displace 1 nM [³H]diprenorphine from GH₃DORT membranes (200 µg of protein/assay) or permeabilized cells (5 × 10⁵ cells/assay) was assessed. Competition binding experiments were conducted in the presence or absence of 25 µM GppNHp and 100 mM NaCl. All binding reactions were terminated by rapid filtration with a Brandel 24-sample standard format harvester (Brandel Inc., Gaithersburg, MD) onto presoaked Whatman GF/B glass filters (Brandel Inc.). Filters were then placed in scintillation vials containing 4 ml of scintillation cocktail for liquid scintillation counting. The amount of radioactivity on the filters was determined the following day using a Packard Tri-Carb 2100TR liquid scintillation counter (Packard Bioscience, Meriden, CT).

[³⁵S]GTPS Binding Assay. G-protein activation by TIPP and DPDPE was examined using the method of [³⁵S]GTPS binding previously described (Liu and Prather, 2001). Briefly, [³⁵S]GTPS binding was performed in an assay buffer containing 20 mM HEPES, pH 7.4, 10 mM MgCl₂, 100 mM NaCl, and 0.1 nM [³⁵S]GTPS. To suppress basal G-protein activation, various concentrations (0.1 nM, 1 nM or 10 µM) of GDP were also included in the binding buffer. Binding was determined using either GH₃DORT membranes (50 µg), crude plasma membranes (50 µg), crude microsomal membranes (50 µg), or permeabilized cells (5 × 10⁵ cells/assay) in the presence of TIPP or DPDPE (10¹¹-10⁶ M). The entire mixture was incubated for 1 h at 30°C, with nonspecific binding determined by the inclusion of 10 µM GTPS. The [³⁵S]GTPS binding reaction was terminated with rapid filtration onto presoaked Whatman GF/B glass filters. Samples were washed twice with ice-cold 20 mM HEPES, pH 7.4. Filters were dried, and the radioactivity retained on the filters was determined by liquid scintillation counting as described above.

Measurement of Adenylyl Cyclase Activity in GH₃DORT Membranes and in Whole Cells. The measurement of cAMP production and inhibition of adenylyl cyclase activity in GH₃DORT whole cells was performed as detailed elsewhere (Martin et al., 2001). In GH₃DORT membranes, the assay was performed as previously described (Law et al., 1982, 1983) with slight modifications. Briefly, membranes (50 µg/assay) were incubated in the presence or absence of drug in an assay mixture. The assay mixture contained 50 mM Tris, pH 7.4, 3 mM MgCl₂, 0.2 mM EDTA, 0.1 mM ATP, 0.2 mM dithiothreitol, 0.1 mM GTP, 0.1 mM cAMP, 120 mM NaCl, 0.25 mM R020-1724 [4-(3-butoxy-4-methoxybenzyl)-2-imidazolidinone]], 0.5 mM 3-isobutyl-1-methylxanthine, 20 mM phosphocreatine, 10 units of creatine kinase, 5 µM forskolin, and 1 µCi of [³²P]ATP. Membranes were combined with the assay mixture in a final total volume of 200 µl. Samples were incubated for 15 min at 30°C, and the reaction was terminated with the addition of 20 µl of 2.2 N HCl. Samples were boiled for 4 min, cooled on ice for 4 min, and then loaded onto chromatography columns containing 1.3 g of neutral alumina. [³²P]cAMP was then collected from the columns by elution with 4 ml of 0.1 M ammonium acetate, pH 7.0, into vials containing 10 ml of scintillation fluid. Radioactivity was counted using a liquid scintillation counter (noted above).

Data Analysis and Statistics. For saturation binding experiments, determination of receptor affinity (K_d) and receptor density (B_max) was performed using the nonlinear regression analysis of GraphPad Prism v2.0b for Macintosh (GraphPad Software, San Diego, CA). GraphPad Prism was also used to determine IC₅₀ values from the competition binding experiments. The IC₅₀ values obtained from full concentration-effect curves were converted to K_i values using the Cheng-Prusoff equation (Cheng and Prusoff, 1973). Data are expressed as mean ± S.E.M. and represented by a minimum of three experiments performed in triplicate or duplicate, unless otherwise stated. For statistical comparisons involving three or more groups, differences between means were determined by a one-way ANOVA followed by post hoc comparisons using either Dunnett's or Tukey's test. When only two groups were compared, differences between means were determined by the nonpaired Student's t test.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

The properties of TIPP at specific points in the signal transduction pathway (i.e., receptor binding, G-protein activation, and effector regulation) were studied in GH₃DORT cells. To examine the ligand/receptor interaction, receptor binding was employed. Guanine nucleotides and sodium ions uncouple the receptor from associated G-proteins and thus shift the equilibrium from the active (R*) state toward the inactive or uncoupled (R) state of the receptor (see Discussion). This results in a decreased affinity for the receptor by agonists, an increased affinity by inverse agonists, and no change in affinity by antagonists (Childers and Snyder, 1980). To examine the manner in which TIPP bound to -opioid receptors in GH₃DORT membranes, the affinity of the established agonist DPDPE, inverse agonist ICI-174864, or TIPP was compared in the absence or presence of the guanine nucleotide analog, GppNHp, and NaCl (Table 1). This was accomplished by examining competition binding between these ligands and the nonselective opioid antagonist [³H]diprenorphine (Fig. 1). Using this method, the IC₅₀ values derived from the competition binding curves are converted to a measure of receptor affinity (K_i) by the Cheng-Prusoff equation. This calculation requires knowledge of the K_d of the radioactive ligand employed and assumes that displacement of the radiolabeled compound by the competing ligand is competitive. Therefore, it was first demonstrated by saturation binding that [³H]diprenorphine bound to -opioid receptors expressed in GH₃DORT membranes with a B_max of 2.28 ± 0.11 pmol/mg and a K_d of 0.71 ± 0.04 nM. Consistent with an antagonist profile, neither the K_d (1.06 ± 0.14 nM) nor the B_max (2.57 ± 0.18 pmol/mg) determined for [³H]diprenorphine by saturation binding was significantly altered in the presence of GppNHp and NaCl. Finally, it was determined that DPDPE and TIPP competitively inhibited [³H]diprenorphine binding because saturation binding experiments conducted in the presence of a fixed concentration of DPDPE or TIPP (equivalent to their K_i), increased the K_d of diprenorphine by approximately 2-fold, without having a significant effect on the B_max (data not shown).

View this table: [in this window] [in a new window]   TABLE 1 Receptor binding properties of TIPP, DPDPE, and ICI-174864 in the absence or presence of GppNHp/NaCl in GH3DORT membranes or in digitonin-permeabilized cells Competitive inhibition of [3H]diprenorphine binding (1.0 nM) in GH3DORT membranes or in permeabilized cells by TIPP, DPDPE, or ICI-174864 (1012-105 M) in the presence or absence of GppNHp (25 µM) and NaCl (100 mM) was performed as described under Experimental Procedures. Affinity (Ki) values are expressed as mean ± S.E.M. and were calculated from IC50 values derived from complete concentration-effect curves using the Cheng-Prusoff equation (see Figs. 1 and 5). The Kd for diprenorphine used to calculate the Ki values in membranes was 0.71 ± 0.04 nM (GppNHp/NaCl) and 1.06 ± 0.14 nM (+GppNHp/NaCl). The Kd for diprenorphine in permeabilized cells was 1.29 ± 0.11 nM (GppNHp/NaCl) and 0.93 ± 0.10 nM (+GppNHp/NaCl). Data for each value presented represent three separate experiments performed in duplicate.

View larger version (22K): [in this window] [in a new window]   Fig. 1.   Competitive inhibition of [3H]diprenorphine (1 nM) binding in GH3DORT membranes by DPDPE (panel A), ICI-174864 (panel B), or TIPP (panel C) in the absence (closed symbols) or presence (open symbols) of GppNHp (25 µM) and NaCl (100 mM). Data are presented as the percentage of [3H]diprenorphine binding in the presence of the indicated drug concentration compared with binding in the absence of any competing ligand (i.e., % of control). Control values for DPDPE competition were 2.57 ± 0.22 pmol/mg (GppNHp/NaCl) and 2.23 ± 0.19 pmol/mg (+GppNHp/NaCl). Control values for ICI-174864 competition were 1.96 ± 0.54 pmol/mg (GppNHp/NaCl) and 1.63 ± 0.42 pmol/mg (+GppNHp/NaCl). Control values for TIPP competition were 2.0 ± 0.13 pmol/mg (GppNHp/NaCl) and 1.94 ± 0.16 pmol/mg (+GppNHp/NaCl). Data represent the mean ± S.E.M. for three to five experiments performed in duplicate. The affinity (Ki) values calculated for each binding condition are presented in Table 1.

When competition binding studies were performed with [³H]diprenorphine, the receptor affinity of the well characterized -opioid agonist DPDPE was reduced over 20-fold in the presence of GppNHp/NaCl (K_i = 14.5 ± 2.22 nM versus 418 ± 51.2 nM) (Fig. 1A; Table 1). In contrast, these conditions increased the affinity of the accepted inverse agonist ICI-174864 by over 16-fold, from 304 ± 73.2 nM to 18.8 ± 3.8 nM (Fig. 1B; Table 1). Surprisingly, TIPP demonstrated a slight, but significant (P < 0.05), increase in affinity in the presence of GppNHp/NaCl (K_i = 3.02 ± 0.16 nM versus 1.16 ± 0.11 nM) (Fig. 1C; Table 1). These results indicated that unlike the agonist DPDPE, but similar to the inverse agonist ICI-174864, TIPP bound preferentially to the inactive or uncoupled (R) state of the -opioid receptor.

It has previously been demonstrated that TIPP acts as an agonist to produce inhibition of adenylyl cyclase activity in whole cells (Martin et al., 2001). Furthermore, overnight pretreatment with pertussis toxin blocks the agonist activity of TIPP, indirectly indicating that Gi/Go G-proteins are involved in this response. Therefore, in the present study the question of whether TIPP could activate G-proteins was examined directly by measuring the increase in [³⁵S]GTPS binding to GH₃DORT membranes in the presence of increasing concentrations of TIPP or DPDPE (Fig. 2; Table 2). [³⁵S]GTPS is a radiolabeled nonhydrolyzable GTP analog that binds irreversibly to G-proteins as a result of their activation produced by the exchange of GTP for GDP on the G subunit (Clark et al., 1997; Szekeres and Traynor, 1997). Characteristic of agonists, DPDPE increased [³⁵S]GTPS binding to GH₃DORT membranes in a concentration-dependent manner by 138.4 ± 34.4% with an EC₅₀ of 12.0 ± 2.8 nM. The amount of DPDPE required to produce half-maximal G-protein activation was similar to the affinity (K_i) of DPDPE for -opioid receptors expressed in this cell line (Table 1). Quite unexpectedly, no increase in [³⁵S]GTPS binding was observed with TIPP, indicating that under these conditions TIPP did not activate G-proteins, even at concentrations as high as 1 µM (Fig. 2). A relatively high concentration of GDP (10 µM) was included in the binding buffer to suppress basal G-protein activation in order to observe optimal agonist stimulation. However, such high concentrations of GDP have been demonstrated to impair the activation of G-proteins by partial agonists in this assay (Breivogel et al., 1998). Therefore, to determine whether TIPP might possess partial agonist properties, [³⁵S]GTPS binding experiments were repeated, employing substantially decreased GDP concentrations (0.1 and 1 nM) and TIPP (10⁹-10⁵ M). Even under these conditions, TIPP showed no increase in [³⁵S]GTPS binding to GH₃DORT membranes (data not shown). Additionally, G-protein activation by DPDPE (100 nM) was completely reversed when coadministered with TIPP (1 µM). Thus, direct measurement using [³⁵S]GTPS binding to GH₃DORT membranes failed to detect G-protein activation by TIPP, and TIPP was able to block the agonist action produced by DPDPE. These observations indicated that TIPP possessed characteristics similar to those of an antagonist.

View larger version (20K): [in this window] [in a new window]   Fig. 2.   [35S]GTPS binding in response to increasing concentrations of TIPP (filled circles) or DPDPE (filled squares) in GH3DORT membranes. Nonspecific binding was defined by the inclusion of 10 µM GTPS. Data are presented as the percentage of [35S]GTPS binding in the presence of the indicated drug concentration compared with basal binding in the absence of any test ligand (i.e., % over control). Control values (i.e., basal [35S]GTPS binding) were 53.9 ± 2.6 fmol/mg for DPDPE and 68.6 ± 7.5 fmol/mg for TIPP. Values represent the mean ± S.E.M. of three separate experiments performed in triplicate. The EC50 and Emax values determined for each drug are presented in Table 2.

View this table: [in this window] [in a new window]   TABLE 2 [35S]GTPS binding in response to TIPP or DPDPE to GH3DORT membranes or to digitonin-permeabilized cells Increasing concentrations of TIPP or DPDPE (1011-106 M) to GH3DORT membranes or to permeabilized cells was assessed (see Figs. 2, 3, and 4). Data were converted to the percentage of [35S]GTPS binding in the presence of increasing drug concentrations, compared with basal binding in the absence of any drug (i.e., percentage over control). The EC50 and Emax values for each drug were determined by nonlinear regression analysis. Control values (i.e., basal [35S]GTPS binding) for DPDPE experiments were 53.9 ± 2.6 fmol/mg (membranes) or 157 ± 15.6 fmol/mg (permeabilized cells). Control values for TIPP experiments were 68.6 ± 7.5 fmol/mg (membranes) or 163 ± 3.89 fmol/mg (permeabilized cells). Values represent the mean ± S.E.M. of three separate experiments performed in triplicate.

Szekeres and Traynor (1997) reported a similar lack of stimulation of [³⁵S]GTPS binding to NG108-15 membranes by TIPP. This is in agreement with the well established classification of TIPP as an antagonist. However, the observation that TIPP acts as an agonist at -opioid receptors to produce a pertussis toxin-reversible inhibition of adenylyl cyclase activity (Martin et al., 2001) strongly implies that G-protein activation (specifically of the Gi/Go subtype) must occur. Importantly, experiments that have demonstrated agonist activity of TIPP were assessed by assays employing whole cells (Martin et al., 2001). However, past studies (Mullaney et al., 1996; Szekeres and Traynor, 1997) and present findings indicative of antagonist characteristics of TIPP evaluated G-protein activation in membrane preparations. Therefore, a possible intriguing explanation for this discrepancy is that G-protein activation by TIPP may require more physiological conditions (i.e., whole cells and/or an intact cytoskeleton) that are lost or disrupted upon membrane preparation. To examine this question, G-protein activation was re-examined using GH₃DORT whole cells that were semi-permeabilized with the detergent digitonin to allow [³⁵S]GTPS to penetrate into the cell (Alt et al., 2001). Results demonstrated that the -opioid agonist DPDPE concentration dependently increased [³⁵S]GTPS binding in cells permeabilized with digitonin, and neither the potency nor the efficacy of DPDPE significantly differed from results observed employing membrane preparations (Fig. 3; Table 2). In contrast to the results obtained using GH₃DORT membranes, TIPP also stimulated [³⁵S]GTPS binding in a concentration-dependent manner (EC₅₀ = 5.8 ± 1.5 nM) to levels 40.8 ± 1.83% above control in digitonin-permeabilized cells (Fig. 4; Table 2). Importantly, the EC₅₀ for G-protein activation by TIPP in this assay was similar to its affinity (K_i) for -opioid receptors (Table 1). Last, when DPDPE (100 nM) and TIPP (1 µM) were coadministered, G-protein activation was reduced to a level that was slightly but significantly (P < 0.05) greater than when TIPP was tested alone (i.e., 59.0 ± 4.9%). These results indicate that TIPP is able to activate G-proteins in digitonin-permeabilized cells, although less efficaciously than the -opioid agonist DPDPE. Therefore, under these conditions, TIPP acts as a partial -opioid agonist.

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Comparison of the stimulation of [35S]GTPS binding with increasing concentrations of DPDPE employing either GH3DORT membranes (filled squares) or digitonin-permeabilized cells (open squares). Nonspecific binding was defined by the inclusion of 10 µM GTPS. Data are presented as the percentage of [35S]GTPS binding in the presence of the indicated drug concentration compared with basal binding in the absence of DPDPE (i.e., % over control). Control values (i.e., basal [35S]GTPS binding) were 53.9 ± 2.6 fmol/mg for membranes and 157 ± 15.6 fmol/mg for permeabilized cells. Values represent the mean ± S.E.M. of three separate experiments performed in triplicate. The EC50 and Emax values determined for each binding condition are presented in Table 2.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Comparison of the stimulation of [35S]GTPS binding with increasing concentrations of TIPP employing either GH3DORT membranes (filled circles) or digitonin-permeabilized cells (open circles). Nonspecific binding was defined by the inclusion of 10 µM GTPS. Data are presented as the percentage of [35S]GTPS binding in the presence of the indicated drug concentration compared with basal binding in the absence of TIPP (i.e., % over control). Control values (i.e., basal [35S]GTPS binding) were 68.6 ± 7.5 fmol/mg for membranes and 163 ± 3.89 fmol/mg for permeabilized cells. Values represent the mean ± S.E.M. of three separate experiments performed in triplicate. The EC50 and Emax values determined for each binding condition are presented in Table 2.

To assure that the observed agonist effect of TIPP using digitonin-permeabilized cells was specifically due to activation of -opioid receptors, [³⁵S]GTPS binding experiments were conducted using a maximally effective concentration of TIPP or DPDPE (i.e., 100 nM) in the absence or presence of a saturating concentration of the selective -opioid antagonist, naltriben (1 µM). In digitonin-permeabilized cells, the stimulation of [³⁵S]GTPS binding by TIPP was decreased 60% in the presence of naltriben (47.7 ± 1.2% versus 19.2 ± 4.4%, n = 6). In comparison, naltriben also decreased DPDPE-induced stimulation of [³⁵S]GTPS binding by 89% (219.9 ± 15.0% versus 24.3 ± 6.9%, n = 4). Because the presence of the -opioid antagonist naltriben attenuated the observed increase in [³⁵S]GTPS binding stimulated by TIPP or DPDPE in digitonin-permeabilized cells, the effect was concluded to be -opioid receptor-mediated. Finally, the stimulation of G-proteins by TIPP or DPDPE (1 µM) was not observed in cells that had been pretreated overnight with pertussis toxin (100 ng/ml) (data not shown).

It is possible that permeabilization of cells allows TIPP to activate G-proteins through activation of -opioid receptors located at distinct intracellular sites rather than those located in the plasma membrane. For example, it is known that -opioid receptors are present not only in plasma membranes, but also as immature receptors in microsomal membranes likely localized to the endoplasmic reticulum (Roth et al., 1981). Since the membrane preparation used in experiments thus far contained a mixture of both plasma and microsomal membranes, we next sought to examine the potential for G-protein activation by TIPP in these different isolated membrane preparations. DPDPE produced a 138.4 ± 34.4% increase in [³⁵S]GTPS binding in GH₃DORT membranes (Table 2) and a 212.0 ± 42.0% (n = 3) stimulation of G-proteins in isolated plasma membranes. However, DPDPE did not elevate [³⁵S]GTPS binding in isolated microsomal membranes. In contrast, TIPP failed to significantly increase [³⁵S]GTPS binding in any membrane preparation tested (data not shown). These data indicate that when isolated membrane preparations are employed, TIPP is unable to produce G-protein activation by acting at plasma membrane or microsomal -opioid receptors.

The ability of TIPP to act as an agonist to stimulate G-proteins could be observed only in digitonin-permeabilized cells (Fig. 4) but not in membranes (Fig. 2). Therefore, it is possible that the receptor binding properties of TIPP might also resemble those of an agonist if digitonin-permeabilized cells were employed in the assay. To test this hypothesis, the ability of the agonist DPDPE, inverse agonist ICI-174864, or TIPP to competitively inhibit [³H]diprenorphine binding in the absence or presence of GppNHp/NaCl in digitonin-permeabilized cells was compared (Fig. 5). Saturation binding revealed that [³H]diprenorphine bound to -opioid receptors in permeabilized GH₃DORT cells with a B_max of 1.24 ± 0.04 pmol/mg and a K_d of 1.29 ± 0.11 nM. The presence of GppNHp and NaCl did not significantly alter the K_d (0.93 ± 0.1 nM) or the B_max (1.32 ± 0.03 pmol/mg) determined for [³H]diprenorphine by saturation binding. Similar to the results obtained in membranes, DPDPE demonstrated a significantly decreased affinity in the presence of GppNHp/NaCl (K_i = 13.2 ± 5.60 nM versus 379 ± 61.1 nM) (Fig. 5A; Table 1). In contrast, the affinity of the inverse agonist ICI-174864 for -opioid receptors was enhanced by these conditions (K_i = 680 ± 46.1 nM versus 83.8 ± 5.46 nM) (Fig. 5B; Table 1). Surprisingly, but also similar to the results obtained in membranes, TIPP demonstrated a significant 10-fold increase in affinity for the uncoupled form of the -opioid receptor (K_i = 4.77 ± 1.39 nM versus 0.42 ± 0.09 nM) (Fig. 5C; Table 1). These results indicate that the distinct binding characteristics of DPDPE, ICI-174864, and TIPP, in the presence of guanine nucleotides and sodium ions, are similar whether examined using membrane preparations or in digitonin-permeabilized cells. In either preparation, DPDPE binds to -opioid receptors like an agonist (i.e., having a higher affinity for the coupled form of the receptor), whereas TIPP binds like the inverse agonist ICI-174864 (i.e., having a higher affinity for the uncoupled form of the receptor).

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Competitive inhibition of [3H]diprenorphine (1 nM) binding in digitonin-permeabilized GH3DORT cells by DPDPE (panel A), ICI-174864 (panel B), or TIPP (panel C) in the absence (closed symbols) or presence (open symbols) of GppNHp (25 µM) and NaCl (100 mM). Data are presented as the percentage of [3H]diprenorphine binding in the presence of the indicated drug concentration compared with binding in the absence of any competing ligand (i.e., % of control). Control values for DPDPE competition were 1.81 ± 0.30 pmol/mg (GppNHp/NaCl) and 1.83 ± 0.07 pmol/mg (+GppNHp/NaCl). Control values for ICI-174864 competition were 1.44 ± 0.03 pmol/mg (GppNHp/NaCl) and 1.55 ± 0.02 pmol/mg (+GppNHp/NaCl). Control values for TIPP competition were 0.897 ± 0.044 pmol/mg (GppNHp/NaCl) and 0.956 ± 0.060 pmol/mg (+GppNHp/NaCl). Data represent the mean ± S.E.M. for three to four experiments performed in duplicate. The affinity (Ki) values calculated for each binding condition are presented in Table 1.

Because such distinct results for G-protein activation by TIPP were obtained in membranes relative to digitonin-permeabilized cells, it was next determined whether these differences would also be observed in the ability of TIPP to regulate the activity of the effector adenylyl cyclase. Therefore, the hypothesis that TIPP would be able to regulate adenylyl cyclase activity in whole cells, but not in membrane preparations, was tested. Similar to previously reported findings (Martin et al., 2001), maximally effective concentrations of TIPP or DPDPE (1 µM) reduced intracellular cAMP levels to similar amounts in GH₃DORT whole cells (61.8 ± 4.3% and 70.1 ± 2.5%, respectively) (Fig. 6A). As expected, the -opioid agonist DPDPE (1 µM) similarly inhibited adenylyl cyclase activity in membranes prepared from GH₃DORT cells by 31.6% (Fig. 6B). The inhibition produced by DPDPE was reversed by coadministration with the opioid antagonist naloxone (10 µM) and by overnight pertussis toxin pretreatment (100 ng/ml). Surprisingly, TIPP (1 µM) also significantly reduced the production of cAMP by 15.8% in GH₃DORT membranes (P < 0.05). Similar to the action of DPDPE, the agonist activity of TIPP was blocked by an opioid antagonist and by overnight pertussis toxin pretreatment. These results indicate that, as far as regulation of the effector adenylyl cyclase is concerned, TIPP possesses agonist activity similar to that of DPDPE.

View larger version (34K): [in this window] [in a new window]   Fig. 6.   Comparison of the inhibition of adenylyl cyclase activity by TIPP and DPDPE in GH3DORT whole cells and in membranes. The ability of maximally effective concentrations of TIPP (1 µM) or DPDPE (1 µM) to reduce adenylyl cyclase activity in GH3DORT whole cells (panel A) and membranes (panel B) was examined. In panel B, the effect of overnight incubation with pertussis toxin (100 ng/ml) or coadministration with the opioid antagonist naloxone (10 µM) on the inhibition of adenylyl cyclase activity produced by DPDPE and TIPP was also examined. Data are presented as the percentage of cAMP levels in the presence of the indicated drug compared with that observed in the absence of any drug (i.e., % of control). Control values for whole cell adenylyl cyclase assays were 212 ± 5.98 fmol/mg/min and 236 ± 17.3 fmol/mg/min for assays conducted with membranes. Data represents the mean ± S.E.M. for three to eight experiments performed in triplicate. , significantly different from control (P < 0.05, Dunnett's post hoc test); #, significantly different from inhibition produced by DPDPE (P < 0.05, Tukey's post hoc test).

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

TIPP was identified as a selective and potent -opioid receptor antagonist (Schiller et al., 1992, 1999a,b) but has since been reported to possess agonist activity comparable with that of the -opioid agonist DPDPE (Martin et al., 2001). Therefore, it is possible that TIPP, and structurally related compounds, might represent a new class of opioid agonists exhibiting unique properties. The current study examined the characteristics of TIPP in a cellular model at several points along the signal transduction pathway (i.e., receptor binding, G-protein activation, and effector regulation) and compared them with that of the well characterized -opioid agonist DPDPE. The most significant finding of this study was that whereas DPDPE displayed agonist properties in all assays, TIPP demonstrated characteristics of an agonist, antagonist, or inverse agonist, depending on the step in the signal transduction cascade examined and the specific assay conditions employed (Table 3). These results suggest that although both DPDPE and TIPP act similarly as agonists to regulate the intracellular effector adenylyl cyclase, they apparently demonstrate significant differences in the signal transduction cascade preceding this final point of convergence.

View this table: [in this window] [in a new window]   TABLE 3 Comparison of the characteristics of TIPP and DPDPE at selected points in the signal transduction pathway of -opioid receptors in GH3DORT membranes or in digitonin-permeabilized cells TIPP displays characteristics of an agonist, antagonist, or inverse agonist depending on the step in the signal transduction cascade examined and the assay conditions employed. In contrast, the well characterized -opioid receptor agonist DPDPE displays only properties anticipated for an agonist.

Many GPCRs exhibit constitutive activity, producing spontaneous regulation of effectors in the absence of activation by agonists (Lefkowitz et al., 1993; Merkouris et al., 1997). A two-state receptor model has been proposed to account for constitutive activity in which receptors exist in an equilibrium between inactive (R) and active (R*) states (Costa et al., 1992). The active (R*) state effectively couples to, and activates G-proteins, whereas the inactive (R) state does not. Upon binding, agonists stabilize or enrich the proportion of receptors in the active (R*) state, inverse agonists stabilize the inactive (R) state and antagonists have equal preferences for both states (Kenakin, 2001). Therefore, agonists, antagonists, and inverse agonists bind to opioid receptors in distinct manners in model systems containing constitutively active receptors (Childers and Snyder, 1980; Childers et al., 1993; Knapp et al., 1996). One approach to examine the unique binding characteristics of agonists, antagonists, and inverse agonists for GPCRs is to compare the affinity of these ligands for receptors under conditions that promote coupling, versus uncoupling, to G-proteins. For example, guanine nucleotides (i.e., GppNHp) and sodium ions uncouple GPCRs from G-proteins, shifting the receptor equilibrium to the inactive (R) state. Thus, in the presence of GppNHp/NaCl, an agonist demonstrates a decrease in affinity for the receptor, the affinity of an inverse agonist is increased, and the affinity of an antagonist remains unaffected. Since TIPP appeared to potentially possess both antagonist and agonist properties, experiments were conducted to determine the manner in which TIPP bound to -opioid receptors. Competitive inhibition of [³H]diprenorphine binding by TIPP in the presence of GppNHp/NaCl demonstrated a significant increase in affinity for the -opioid receptor in both GH₃DORT membranes and in digitonin-semi-permeabilized cells. This increase in affinity suggests that TIPP preferentially binds to the inactive/uncoupled (i.e., R) state of the -opioid receptor, which was similarly demonstrated in this study for the well characterized inverse agonist ICI-174864 (Merkouris et al., 1997; Neilan et al., 1999). This was quite surprising given that TIPP has only been demonstrated to possess either agonist or antagonist (but not inverse agonist) characteristics.

Although the increase in the affinity of TIPP for -opioid receptors in the presence of GppNHp/NaCl is highly unusual, given its previously observed agonist activity, not all opioid agonists exhibit the anticipated decreased affinity under these conditions. For example, the nonpeptide -opioid agonists BW373U86 and one of its enantiomers, SNC80, retain their high affinity in the presence of guanine nucleotides and sodium ions (Childers et al., 1993; Knapp et al., 1996). Similarly, the affinity of etorphine, a nonselective nonpeptide opioid agonist, is also unaffected by the presence of guanine nucleotides and sodium ions (Childers and Snyder, 1980). BW373U86 is a potent inhibitor of adenylyl cyclase in both brain membranes and NG108-15 cells, and it has been argued that the lack of sensitivity of this compound to guanine nucleotides and sodium ions is the reason it retains its potency in both brain tissue and cultured cells (Childers et al., 1993). SNC80 also retains full agonist activity despite its insensitivity to GppNHp and sodium ions. These studies suggest that BW373U86 and SNC80 bind to the -opioid receptor very differently from other -opioid agonists, and this hypothesis is supported by studies demonstrating that various -opioid ligands interact distinctly with the -opioid receptor (Befort et al., 1996; Valiquette et al., 1996).

It is important to note that TIPP is the first ligand, that we are aware of, to demonstrate an inverse agonist property of preferential binding to the inactive state of the -opioid receptor, and yet possess an agonist characteristic of producing adenylyl cyclase inhibition. This observation has significant implications. First, it suggests that although TIPP may preferentially bind to the inactive form of -opioid receptors, this interaction does not result in an enrichment or stabilization of this receptor state. If this were the case, since -opioid receptors are constitutively active in this cell line (see the following), it would be expected that TIPP would also act functionally as an inverse agonist. For example, the inverse agonist ICI-174864 examined in this study also produces an increase in cAMP levels (Martin et al., 2001) and a reduction in [³⁵S]GTPS binding (unpublished data) in GH₃DORT cells. In contrast, as has been demonstrated, TIPP produces an inhibition of adenylyl cyclase activity and G-protein activation. Second, there is convincing evidence that peptide and alkaloid ligands bind to -opioid receptors differently (Meng et al., 2000). However, the strikingly distinct binding characteristics of the two peptide agonists examined in the present study suggest that individual peptides may also differ in their recognition of -opioid receptors.

Another important observation from the present study is the finding that TIPP is only able to activate G-proteins in GH₃DORT cells made semi-permeable with digitonin, whereas DPDPE can activate G-proteins in either membrane preparations or in digitonin-permeabilized cells. Furthermore, TIPP even failed to activate G-proteins in isolated plasma or microsomal membrane preparations. Therefore, in membranes, TIPP appears to act as an antagonist whereas in digitonin-permeabilized cells, TIPP exhibits characteristics of an agonist. Such "conditional" activation of G-proteins is intriguing and could explain the lack of G-protein activation previously reported for TIPP when using membrane preparations (Mullaney et al., 1996; Szekeres and Traynor, 1997). However, at first glance it appears difficult to reconcile these observations with other findings presented in the present study. For example, despite an apparent inability to activate G-proteins in membrane preparations, TIPP is nevertheless able to produce inhibition of membrane adenylyl cyclase activity. Potential explanations for this inconsistency require a more detailed examination of the data. First, it is clear that Gi/Go proteins are involved in the inhibition of adenylyl cyclase activity produced by TIPP in membrane preparations because this effect is completely reversed by pretreatment of cells with pertussis toxin. Therefore, it is likely that TIPP does indeed stimulate G-proteins in membrane preparations but below the level of detection by the method and/or assay conditions used in the present study. Second, support for this hypothesis is provided by the observation that the efficacy of TIPP, relative to DPDPE, to inhibit adenylyl cyclase activity is also significantly reduced (although not eliminated) in membrane but not in whole cell assays. Although it is unclear why the efficacy for G-protein activation and regulation of adenylyl cyclase activity produced by TIPP are dramatically reduced or absent in membrane preparations, it is likely that preparation of membranes from whole cells results in a disruption, loss, or a significant reduction in an essential component required to convey the agonist activity of TIPP.

There are several examples of potential elements, crucially involved in the signal transduction cascade initiated by TIPP, that could be disrupted upon membrane preparation. First, G-proteins are not integral membrane proteins but are rather attached to the membrane by post-translational lipid modifications such as myristoylation (Jones et al., 1990) and palmitoylation (Linder et al., 1993). Homogenization and subsequent high-speed centrifugation employed to prepare membranes might result in a reduction in the amount of pertussis toxin-sensitive G-protein(s) recovered. Indeed, this has been reported to occur for the Gi/Go protein transducin (Chabre and Deterre, 1990), and a more recent study demonstrated that this resulted in a loss of detectable activation of this G-protein by -opioid receptors in membrane preparations (Varga et al., 2000). Therefore, if TIPP preferentially activated G-protein(s) that are lost during membrane preparation, this would obviously result in a decrease in measurable G-protein activation (i.e., GTPS binding) and in the subsequent efficacy of effector regulation (i.e., adenylyl cyclase inhibition) produced by TIPP. Second, there is an increasing amount of evidence suggesting that GPCR signaling can occur in cellular microdomains containing receptors, G-proteins, and effectors (Simons and Toomre, 2000). If the agonist effects of TIPP required that participating signaling components be contained within such microdomains, disruption of these specialized cellular compartments by membrane preparation would obviously interfere with the normal signal transduction cascade. Last, it is also possible that an unidentified accessory protein present in the cytosol, removed by membrane preparation, might be required to achieve the activation of G-proteins by TIPP. In any case, since DPDPE is able to activate G-proteins and regulate adenylyl cyclase similarly in both membrane preparations and in whole cells, the most important implication of these data is that DPDPE and TIPP most likely require and/or utilize different signaling mechanisms to ultimately activate G-proteins and regulate the same intracellular effector adenylyl cyclase.

In summary, the previously established selective -opioid receptor antagonist TIPP demonstrates characteristics of an agonist, antagonist, or inverse agonist, depending on the step in the signal transduction cascade examined and the assay conditions employed. Therefore, TIPP appears to bind to and activate the -opioid receptor in a manner distinct from that of other agonists. These unique properties may possibly characterize a new class of selective -opioid receptor agonists, derived from the TIP(P) peptides. The mechanism(s) underlying this novel activity remain to be elucidated.