Changes in the Association of G Protein Subunits with the Cloned Mouse Delta Opioid Receptor on Agonist Stimulation
- Department of Pharmacology, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania
Abstract
G proteins couple delta opioid receptors to multiple cellular effector systems and are critical components of thedelta opioid signal transduction cascade. To investigate the physical association of delta opioid receptors with G proteins, the cloned mouse delta opioid receptor was solubilized, and the G proteins associated with the receptor were identified through coimmunoprecipitation of the receptor/G protein complexes with antisera directed against different Gα and Gβ subunits. The delta receptor associates with Giα1, Giα3, Goα, Gβ1 and Gβ2 subtypes. On agonist binding to the receptor, a greater proportion of the receptor is associated with Giα than with Goα, Giα1dissociates from the receptor and Giα2 associates with the receptor, whereas Giα3 and the Gβsubunits remain coupled to the delta receptor. These findings reveal dynamic changes in the G proteins associated with the receptor after agonist binding that may be linked to the activation of the delta receptor. In addition to pertussis toxin-sensitive G proteins, the delta receptor physically interacts with the pertussis toxn-insensitive G proteins Gqα and Gzα. These interactions may be critical in linkingdelta receptors to phospholipase C. The diversity of G proteins associated with the delta opioid receptor may form the basis for the selective coupling of these receptors to multiple cellular effector systems.
An action of opioids is to inhibit the modulation of synaptic transmission in both the central and peripheral nervous systems. Opioids induce their biological actions through association with three classes of receptors: delta,kappa and mu. Each of these receptors has recently been cloned and is a member of the seven-transmembrane spanning superfamily of G protein-coupled receptors (Chen et al., 1993; Evans et al., 1992; Kieffer et al., 1992; Yasuda et al., 1993). The cloned opioid receptors are ∼65% identical in amino acid sequence and are most closely related to somatostatin receptors (Reisine and Bell, 1993). In fact, the mouse delta opioid receptor was isolated from a cDNA library using probes selective for the transmembrane spanning regions of the previously cloned somatostatin receptors (Yasudaet al., 1993).
At the cellular level, opioid receptors modulate various effector systems. All three subtypes of opioid receptors inhibit adenylyl cyclase (Attali et al., 1989; Chen et al., 1993;Frey and Kebabian, 1984; Sharma et al., 1975; Yasudaet al., 1993) and Ca++ conductance (Gross and MacDonald, 1987; Hescheler et al., 1987; Schroeder et al., 1991) while stimulating K+ conductance (Chen and Yu, 1994; North, 1993; Wimpey and Chavkin, 1991), and Na+/H+ exchange (Isom et al., 1987). Opioid receptors are coupled to these effector systems by G proteins. G proteins are heterotrimeric complexes consisting of alpha,beta and gamma subunits. Many subtypes ofalpha, beta and gamma subunits have been cloned, and the myriad of possible combinations may provide the basis for the divergent cellular actions of neurotransmitters and hormones (Simon et al., 1991).
Most of the cellular effects of delta opioid receptors are reported to be blocked by prior treatment with PTX. PTX ADP-ribosylates Giα/Goα (Giα1, Giα2, Giα3, Goα1 and Goα2) subtypes and effectively disrupts their activation by receptors. This result implies that any or all of the Giα/Goα subtypes might be mediating cellular actions of delta opioid receptors. The coupling of Gα subunits with delta opioid receptors has been examined by many methods with varying results. In a reconstituted system, delta opioid receptors were first shown to be linked by Goα to the inhibition of Ca++ conductance (Hescheler et al., 1987). More recently, in NG108-15 cells, an approach using PTX-insensitive mutants indicates that Goα1 couples the delta opioid receptor to the inhibition of Ca++ conductance (Taussig et al., 1992). In contrast, other studies in NG108-15 cells that measure cholera toxin-induced ADP-ribosylation conclude that Goα2associates with delta opioid receptors in these cells (Roerig et al., 1992). To study deltareceptor/Giα interactions, antisera have been used to uncouple the signaling pathway between delta opioid receptors and adenylyl cyclase. This approach in NG108-15 cells indicates that Giα2-directed antiserum disruptsdelta agonist-stimulated GTPase activity and inhibition of adenylyl cyclase activity (McKenzie and Milligan, 1990). The same technique in SH-SY5Y cells reveals that both Giα and Goα directed antisera uncouple delta opioid receptor signaling to adenylyl cyclase (Carter and Medzhradsky, 1993). Furthermore, labeling techniques including 32P-azidoanilide show many different combinations of Giα/Goαsubtypes coupling with delta opioid receptors (Laugwitzet al., 1993; Offermans et al., 1991; Roeriget al., 1992). These varying results may be due to the fact that different cell lines and techniques are used in the measurement ofdelta opioid receptor/Gα associations.
For somatostatin receptors, which are most closely related to opioid receptors, association with G proteins has been investigated by an immunoprecipitation approach (Law et al., 1991, 1993; Law and Reisine, 1992). These studies indicate that rat brain and AtT-20 cell somatostatin receptors associate with Giα1, Giα3 and Goα and the cloned somatostatin receptor SSTR2A associates with Giα3 and Goα. This correlates with the functional analyses, which have shown that Giα1 couples somatostatin receptors to adenylyl cyclase (Tallent and Reisine, 1992), Goα2couples the receptor to Ca++ channels (Kleuss et al., 1991; Taussig et al., 1992) and Giα3couples the receptor to K+ channels (Yatani et al., 1987). Therefore, the coupling of distinct G protein subunits with somatostatin receptors may determine which signaling pathways can be activated.
Investigations into the coupling of delta opioid receptors with G proteins have yielded many contradictory results. With the cloning of the delta opioid receptor, the ability to study its association with various G protein subunits in isolation from other opioid receptors may allow insight into the signaling process fromdelta opioid receptors to effector systems. To determine which G protein subunits associate with the cloned deltaopioid receptor, a similar approach was used as previously described for somatostatin receptor/G protein coupling (Law et al., 1991, 1993; Law and Reisine, 1992). Our results show thatdelta opioid receptors associate with Giα1, Giα3, Goα, Gβ1 and Gβ2 subunits. On binding of agonist to the receptor, the G protein association changes so that a greater proportion of the receptor associates with Giα than with Goα, much less Giα1 is coupled to the receptor and the receptor forms a new association with Giα2 and remains coupled to Giα3, Gβ1 and Gβ2. Furthermore, a component of the high-affinity agonist binding to thedelta opioid receptor was found to be PTX insensitive but GTPγS sensitive. It was then discovered that the deltaopioid receptor associates with Gzα and Gqαin both the absence and presence of agonist. Our findings for the first time reveal alterations in delta receptor/G protein coupling that occur as a result of agonist binding and, most importantly, show interactions of the receptor with PTX-insensitive G proteins. These dynamic changes in delta receptor/G protein association may be the underlying basis of activation for the delta opioid receptor signal transduction pathway.
Materials and Methods
Stable cell line.
The mouse delta opioid receptor has been stably expressed in CHO-DG44 variant cell line (Yasudaet al., 1993). The pharmacological characteristics of the cloned delta opioid receptor have been studied in detail (Raynor et al., 1994) and are similar to thedelta receptor endogenously expressed in NG108 cells, a cell line from which the receptor was cloned. The density ofdelta receptors in CHO cells is 1251 fmol/mg of protein (Raynor et al., 1994).
Western blotting and antisera.
Conditions for gel electrophoresis and the transfer of proteins to nitrocellulose membranes to analyze G proteins have been previously described (Carlsonet al., 1989). The primary antibodies bound to proteins on the nitrocellulose membranes were detected with the Protoblot alkaline phosphatase kit (Promega, Madison, WI). All antisera were used at a dilution of 1:250 for Western blotting.
The G protein-directed antisera that were used were all obtained after injection of synthetic peptides corresponding to specific regions of the Gα or Gβ subunits (table1). Antiserum 8730 is directed against a carboxyl-terminal region of Giα and selectively detects and immunoprecipitates Giα1, Giα2 and, to a lesser extent, Giα3 (Carlson et al., 1989). Antiserum 9072 (anti-Go) is directed against the same region of Giα as 8730, which differs in sequence between Goα and Giα (Law et al., 1993;Law and Reisine, 1992). It selectively detects the recombinant forms of Goα by Western blotting1. Antiserum 2918 is directed against an internal region of Goα and selectively interacts with Goα (Carlson et al., 1989; Lawet al., 1991; Williams et al., 1990). Antisera 3646 (anti-Giα1) and 1521 (anti-Giα2) are made to the same internal region of Giα that is divergent in sequence for Giα1 and Giα2. The antisera are selective on Western blots and in detecting recombinant Giα subtypes (Carlson et al., 1989; Williamset al., 1990). Antiserum 1518 specifically interacts with Giα3, as shown by its selectivity for recombinant Giα3 (Williams et al., 1990), and is made to a peptide corresponding to a different region of Giα3 than antisera 3646 or 1521. These antisera have been previously used to selectively immunoprecipitate or uncouple somatostatin receptor/G protein complexes (Law et al., 1991, 1993; Law and Reisine, 1992).
Peptide-directed antisera against the different G protein subunits
Antiserum 2919 selectively detects Gzα by immunoblotting (Lounsbury et al., 1993). Similar to antisera 8730 and 9072, antiserum 946 is carboxyl-terminally directed and detects Gqα (Carlson et al., 1989).
Antiserum 8136 is generated against a peptide with an amino acid sequence common to Gβ1 and Gβ2. Antiserum 8132 is directed against a unique sequence of Gβ1 and corresponds to U49 (Gao et al., 1987). Antiserum 8129 is directed against an amino acid sequence unique to Gβ2 and corresponds to K-523 (Gao et al., 1987).
Solubilization of delta opioid receptors.
In a typical experiment, ∼10 million CHO cells stably expressing thedelta opioid receptor grown to confluency would yield sufficient solubilized receptor for one immunoprecipitation sample (e.g., either the nonimmune serum or one G protein-directed antiserum immunoprecipitation). The delta receptor/G protein complexes from CHO cells were solubilized by removing media, washing the plated cells with 50 mM Tris · HCl (pH 7.8) and then scraping the cells from the flasks. The cells were centrifuged at 24,000 ×g for 7 min at 4°C, and the supernatant was discarded. To the pelleted cells, 2 ml of buffer was added containing 50 mM Tris · HCl (pH 7.8), 1 mM EGTA, 5 mM MgCl2, 10 μg of leupeptin, 2 μg of pepstatin and 200 μg of bacitracin (buffer A). For these studies, buffer A also contained 4 μl of 50 mM phenylmethylsulfonyl fluoride in ethanol. The sample was homogenized with a Brinkmann Polytron and centrifuged at 120 × gfor 10 min at 4°C. The resulting supernatant was removed and centrifuged at 45,000 × g for 15 min at 4°C. The supernatant was subsequently discarded, and 1 ml of buffer A was added to the membrane pellet, which was then homogenized as before. Next, the cell membranes were incubated with the deltareceptor-selective agonist 1 μM DPDPE or buffer A for 30 min at 30°C and then chilled for 10 min at 4°C. Then, 250 μl of solubilization buffer (50 mM CHAPS and 50% glycerol) was added to the samples, and the samples were placed on ice with constant stirring for 30 min. The sample was then centrifuged at 100,000 × gfor 60 min at 4°C. The supernatant was diluted 1:3 in buffer A with 7.5% glycerol and 0.5 μg/ml aprotinin and concentrated to a final volume of 1 ml. The 1-ml sample was loaded onto a Sephadex G-50 column (0.7 × 15 cm, BioRad, Melville, NY) and run in the following buffer: 50 mM Tris (pH 7.8), 1 mM EGTA, 5 mM MgCl2, 5% glycerol and 2 mM CHAPS. Eluted fractions were assayed for specific binding to the opioid receptor agonist 125I-β-endorphin. Fractions containing such activity were pooled, concentrated with Centricon 30 (Amicon) ultrafiltration devices and placed in the immunoprecipitation assay. The recovery of delta receptors after solubilization was ∼10% compared with the amount of receptor present in the starting membrane preparation.
Immunoprecipitation of delta opioid receptor/G protein complexes.
Solubilized delta opioid receptors were incubated with G protein-specific antisera (table 1) at a final dilution of 1:20. The amount of antisera used was similar to that used in previous studies of somatostatin and alpha-2 adrenergic receptors (Law et al., 1991, 1993; Law and Reisine, 1992;Okuma and Reisine, 1992) and was optimal for the immunoprecipitation ofdelta opioid receptor/G protein complexes. Higher concentrations of antisera had no further effect. A 1:12 dilution of 50% (w/v) protein A-Sepharose beads (CL-4B, Sigma Chemical, St. Louis, MO), washed three times and diluted in buffer A, was also added. The samples were then placed on a rotator at 4°C and incubated overnight. Following this, the samples were centrifuged at 10,000 rpm for 4 minutes in an Ependorf microcentrifuge. The supernatant was removed, and the presence of solubilized delta opioid receptors was detected in this portion of the sample using the125I-β-endorphin binding assay. The immunoprecipitate was washed in buffer A and centrifuged. The supernatant was then discarded, the immunoprecipitate was resuspended in buffer A and the presence ofdelta opioid receptors detected using the125I-β-endorphin binding assay.
125I-β-Endorphin binding assay.
The presence of solubilized delta opioid receptors was detected with the high-affinity agonist 125I-β-endorphin (specific activity, 2200 Ci/mmol; Amersham, Arlington Heights, IL). Solubilizeddelta opioid receptors were incubated with 25 pM125I-β-endorphin in a total volume of 0.3 ml of buffer A. Nonspecific binding of 125I-β-endorphin was determined as the amount of binding remaining in the presence of 1 μM β-endorphin or DSLET and accounted for <20% of the total125I-β-endorphin binding. The binding reaction was carried out at 25°C for 60 min. Under these conditions, the binding reaction reached equilibrium. The binding reaction was terminated by the addition of 9 ml (three consecutive additions of 3 ml) of 50 mM Tris · HCl (pH 7.8) at 4°C. The samples were vacuum filtered over Whatman (GF/F) glass-fiber filters that had been presoaked in 0.5% polyethylenimine at 4°C. The filters were dried, and radioactivity was measured in a gamma counter (80% efficiency). A similar procedure was used to detect immunoprecipitated delta opioid receptors. In these studies, after immunoprecipitation ofdelta opioid receptor/G protein complexes with antisera directed against different G protein subunits, the immunoprecipitate was resuspended in an appropriate volume with buffer A, and the presence of delta opioid receptors was detected using the125I-β-endorphin binding assay.
PTX treatment.
CHO cells stably expressing thedelta opioid receptor were treated overnight with media containing 100 ng/ml PTX (List Biologicals, Campbell, CA). Membranes were then prepared and the membrane binding assay was performed as previously described (Raynor et al., 1994). The CHO cell membranes were incubated with 25 pM 125I-β-endorphin in a total volume of 0.3 ml of buffer A. Nonspecific binding of125I-β-endorphin was determined as the amount of binding remaining in the presence of 1 μM β-endorphin or DSLET and accounted for <20% of the total 125I-β-endorphin binding. The binding reaction was carried out at 25°C for 60 min, when equilibrium was accomplished. The binding assay was terminated by rapid vacuum filtration, and the filters were washed with 12 ml of a Tris · HCl buffer (pH 7.8) and counted in a gamma counter (80% efficiency).
Data analysis programs.
Inhibition curves were analyzed using the National Institutes of Health computer-based PROPHET program as previously described (Raynor et al., 1994). Immunoprecipitation data were analyzed by the statistical program Number Cruncher Statistical Systems, Version 501 (Kaysville, UT).
Results
The cloned mouse delta opioid receptor expressed in CHO (DG44 variant) cells has pharmacological properties similar to those of the delta-2 opioid receptor endogenously expressed in brain and NG108 cells (Raynor et al., 1994). After solubilization of the receptor from CHO cells with the nonionic detergent CHAPS, the receptor was labeled with 125I-β-endorphin. Specific binding was inhibited by the opioid agonists β-endorphin and thedelta receptor-selective agonist DPDPE as well as the antagonists diprenorphine and naloxone (fig. 1). The specific binding was not affected by dextrorphan.
Pharmacological characterization of the solubilizeddelta opioid receptor. The specific binding of125I-β-endorphin to the solubilized deltaopioid receptor was inhibited by varying concentrations of β-endorphin (▴), DPDPE (□), naloxone (•), diprenorphine (○) or dextrorphan (▵). Data are mean results of three different experiments.
The solubilized delta opioid receptor sample was examined by Western blot analysis to determine its G protein complement. All of the Giα and Gβ subtypes as well as Goα, Gqα and Gzα were present (fig. 2). Thus, any or all of these subunits could potentially couple the delta opioid receptor to various cellular effector systems.
G proteins present in CHO (DG44 variant) cells expressing the cloned delta opioid receptor. Western blotting was preformed on CHO cells stably expressing the cloneddelta opioid receptor to detect the Gαsubunits Giα (antiserum 8730), Goα(antiserum 9072), Gqα (antiserum 946) and Gzα (antiserum 2919) (A); the Giα subtypes Giα1 (antiserum 3646), Giα2 (antiserum 1521) and Giα3 (antiserum 1518) (B); and the Gβ subunits Gβ1 (antiserum 8132), Gβ2 (antiserum 8129) and Gβ1/2 (antiserum 8136 detects both Gβ1 and Gβ2).
Delta receptor/Gα interactions.
Peptide directed antisera against Goα and Giα were used to investigate deltareceptor/Gα associations. Antiserum 8730, which is directed against the carboxyl terminus of Giα and selectively detects Giα1, Giα2 and Giα3, does not immunoprecipitate or uncoupledelta receptor/Giα complexes (fig.3, A and B). However, 8730 is very effective in immunoprecipitating Giα (Law and Reisine, 1992). This indicates that either Giα is not associated with thedelta opioid receptor or the 8730 epitope on Giα is inaccessible. In contrast, antiserum 9072, which is directed against the carboxyl terminus of Goα, uncouples delta receptor/Goα complexes (fig.3, A and B). This is demonstrated by a loss of binding sites in the supernatant but no immunoprecipitation of these sites. A similar result was seen previously as antiserum 9072 uncoupled the cloned somatostatin receptor SSTR2 from Goα (Law et al., 1993). The differences seen with the two carboxyl-terminally directed antisera, the uncoupling by antiserum 9072 and the inability of antiserum 8730 to have an effect may indicate that the deltareceptor contact sites are different for Goα vs. Giα. Antiserum 2918, which is directed against an internal sequence of Goα, immunoprecipitatesdelta receptor/Goα complexes (fig. 3, A and B). This is indicated by the appearance of specific125I-β-endorphin binding sites in the 2918 immunoprecipitate and the corresponding loss of deltareceptor binding sites in the supernatant. The data obtained with the Goα-directed antisera show that the cloneddelta opioid receptor associates with Goα.
The association of Giα and Goα with the cloned delta opioid receptor. Solubilized delta opioid receptor/G protein complexes were immunoprecipitated with various G protein antisera. Deltaopioid receptors in the immunoprecipitate or supernatant were detected with 125I-β-endorphin, and the resulting specific binding was compared with nonimmune serum. Data are the mean ± S.E.M. specific cpm bound in seven different experiments. A, Solubilizeddelta opioid receptor/G protein complexes were immunoprecipitated with Giα (8730) and Goα(9072 and 2918)-directed antiserum or nonimmune serum (NI). Open bars, specific binding of 125I-β-endorphin in the immunoprecipitates. B, Hatched bars, delta receptor/G protein complexes detected by 125I-β-endorphin in the supernatants from A. C, Solubilized delta opioid receptor/G protein complexes were immunoprecipitated with antisera against Giα1 (3646), Giα2 (1521) or Giα3 (1518) and nonimmune serum (NI). Open bars, specific binding of 125I-β-endorphin in the immunoprecipitates. D, Hatched bars, delta receptor/G protein complexes detected by125I-β-endorphin in the supernatants from C.
To further investigate whether Giα associates with thedelta receptor, selective peptide-directed antisera against internal regions of the Giα subtypes were used. Antiserum 3646 selectively recognizes Giα1 and coimmunoprecipitatesdelta opioid receptor/Giα1 complexes (fig. 3, C and D). Antiserum 1521, directed against Giα2, did not coimmunoprecipitate the delta receptor or cause uncoupling of the receptor from Giα2. It is unlikely that the antiserum 1521 epitope in Giα2 is hidden because antisera 3646 and 1521 are made to the same region of Giα, which differs in sequence, and antiserum 3646 is able to immunoprecipitatedelta receptor/Gα complexes. This indicates that Giα2 does not associate with the deltareceptor under these conditions. Antiserum 1518, directed against Giα3, did not immunoprecipitate the deltareceptor. However, it did uncouple the receptor from Giα3, as reflected by the loss of high-affinity agonist binding sites in the supernatant (fig. 3, C and D). Antiserum 1518 similarly uncouples somatostatin receptors from Giα3 (Lawet al., 1991), possibly indicating that this antiserum interacts with domains of Giα important for receptor/Giα3 coupling. These findings indicate that thedelta receptor associates with Giα1 and Giα3 and that the inability of antiserum 8730 to immunoprecipitate or uncouple receptor/Giα complexes is likely to be due to epitope inaccessibility.
Comparison of the relative levels of delta receptor coimmunoprecipitated or uncoupled by the Goα and Giα antisera indicates that the delta receptor may predominantly associate with Goα.
Effect of agonist treatment on deltareceptor/Gα coupling.
The effect of agonist ondelta receptor/Gα interactions was examined by incubating the receptor before solubilization with thedelta-specific agonist DPDPE for 30 min. It should be noted that the levels of 125I-β-endorphin binding todelta receptors in the nonimmune controls varied little between the untreated and agonist-treated samples. Thus, similar levels of delta receptor binding were added in the immunoprecipitation procedure. Furthermore, residual free agonist was removed through column chromatography and was not present in the sample during the binding assay because fractions eluted from the gel filtration column were detected by their ability to bind the agonist125I-β-endorphin, which could not have labeled the solubilized receptor if the receptor were still bound to DPDPE. In the presence of agonist, antiserum 8730 is able to coimmunoprecipitate thedelta opioid receptor (fig. 4, A and B), which is in contrast to its inability to immunoprecipitate the agonist free receptor (fig. 3, A and B). Antisera 2918 and 9072 are able to coimmunoprecipitate and uncouple, respectively, the agonist-bounddelta receptor from Goα (fig. 4, A and B), indicating that Goα remains coupled to thedelta receptor after the binding of DPDPE. However, the relative levels of delta receptor able to associate with Goα and Giα changed after agonist binding to the receptor so that a similar proportion of the receptor associated with Giα and Goα (compare figs. 3 and 4). Agonist binding to the delta receptor may induce conformational changes in the delta receptor and/or Giα so that the carboxyl terminus of Giα is no longer occluded and is freely accessible to antiserum 8730, which can then coimmunoprecipitate deltareceptor/Giα complexes and may also facilitate receptor/Giα interactions.
Interaction of the agonist DPDPE with thedelta receptor changes deltareceptor/Giα associations. Delta opioid receptor/G protein complexes treated with the delta receptor agonist DPDPE and then solubilized were immunoprecipitated with various G protein antisera. Delta opioid receptors in the immunoprecipitate or supernatant were detected with125I-β-endorphin, and the resulting specific binding was compared with nonimmune serum. Data are the mean ± S.E.M. specific cpm bound in at least four different experiments. A,Delta receptor/G protein complexes were immunoprecipitated with antisera against Giα (8730), Goα (9072 or 2918) and the nonimmune serum (NI). Open bars, specific binding in the immunoprecipitate. B, Hatched bars, specific binding in the supernatants from A. C, Delta receptor/G protein complexes were immunoprecipitated with antisera against Giα1(3646), Giα2 (1521) or Giα3 (1518) and nonimmune serum (NI). D, Hatched bars, specific binding in the supernatants from C.
After the binding of DPDPE to the delta receptor, it was observed that different Gα subunits associate with the receptor. Antiserum 3646 was much less able to immunoprecipitate or uncouple the delta receptor/Giα1 complexes (fig. 4, C and D). The levels of specific binding of125I-β-endorphin in the immunoprecipitate of the agonist naive and pretreated receptors were 568 and 191 cpm, respectively. In contrast, antiserum 1521 coimmunoprecipitates and predominantly uncouples the receptor/Giα2 complexes and antiserum 1518-coimmunoprecipitated delta receptor/Giα3complexes (fig. 4, C and D). This finding indicates that Giα2 associates with the agonist-treated deltareceptor, which is in contrast to experiments done in the absence of agonist (fig. 3, C and D). The interaction of Giα3 with the delta receptor is slightly altered in the presence of agonist so that antiserum 1518 is able to immunoprecipitatedelta receptor/Giα3 complexes (fig. 4, C and D), whereas it previously uncoupled such complexes (fig. 3, C and D). Therefore, agonist binding to the delta receptor causes the dissociation of Giα1 from the receptor, causes a new association of Giα2 with the receptor and promotes the coupling of Giα3 with the receptor.
Both antisera 1521 and 1518 independently disrupted most of thedelta receptor/Giα coupling (fig. 4D). It is not clear why the almost complete uncoupling of the receptor from Giα by each antisera occurs. This may have occurred if the antisera cross-reacted with the same Giα subunits. However, the antisera are generated against different peptides and do not cross-react by either immunoblotting or immunoprecipitating the different alpha subunits (Carlson et al., 1989;Law et al., 1991, 1993; Law and Reisine, 1992; Williamset al., 1990). Furthermore, figure 3 shows that 1518 uncouples delta receptor/Giα complexes, whereas 1521 does not, again indicating that the antisera do not cross-react under the conditions used in these studies.
Delta receptor/Gβ interactions.
In addition to Giα1, Giα3 and Goα, Gβ1 and Gβ2 form stable complexes with the delta receptor. Antisera 8132, 8129 and 8136, which are directed against Gβ1, Gβ2and Gβcommon, respectively, coimmunoprecipitate thedelta receptor (fig. 5, A and B). Immunoprecipitation of delta receptor/Gβ1complexes by antiserum 8132 could be blocked by the peptide to which the antiserum was generated with no effect seen by the peptide alone. However, analysis of the supernatant binding results show that antiserum 8132 potentiates binding above the nonimmune control level (fig. 5, A and B). The potentiation by antiserum 8132 was blocked to the level of the nonimmune control when the antiserum was incubated with 8132 peptide. This probably represents a full blockade because the 8132 peptide completely blocks the ability of antiserum 8132 to immunoprecipitate delta receptor/Gβ complexes. The potentiation by 8132 may be due to the ability of the antiserum to alter Gβ conformation, thereby promoting thedelta receptor high-affinity state. The 8132 peptide itself could significantly decrease delta receptor binding in the supernatant (fig. 5, A and B). It is possible that this peptide contains the receptor/Gα recognition domain and competes with other Gβ subunits for these sites, thus causing a decrease in specific binding.
The association of Gβ subunits with solubilized delta opioid receptors. Delta opioid receptor/G protein complexes treated with or without thedelta receptor agonist DPDPE, solubilized and then immunoprecipitated with various G protein antisera. Deltaopioid receptors in the immunoprecipitate or supernatant were detected with 125I-β-endorphin, and the resulting specific binding was compared with nonimmune serum controls. Data are the mean ± S.E.M. specific cpm bound in five different experiments. A, In the absence of agonist, delta receptor/G protein complexes were immunoprecipitated with antisera against Gβ1 (8132), Gβ2 (8129) and an antiserum against both Gβsubunits (8136) or nonimmune serum (NI). Open bars, specific binding in the immunoprecipitate. B, Hatched bars, specific binding in the supernatants from A. C, In the presence of the delta opioid agonist DPDPE, delta receptor/G protein complexes were immunoprecipitated with the antisera listed in A. Open bars, specific binding in the immunoprecipitate. D, Hatched bars, specific binding in the supernatants from C. 8132p is CEGNVRSRELAGHTGY.
Agonist treatment of the delta receptor does not dramatically alter the association of Gβ subunits with the receptor. All three antisera against Gβ1 and Gβ2 immunoprecipitate deltareceptor/Gβ complexes (fig. 5, C and D). These results are similar to those obtained in the absence of agonist (fig. 5, A and B). The only difference is that antiserum 8132 directed against Gβ1 no longer potentiates binding in the supernatant. The presence of agonist then may induce in Gβ1 a conformational change that stabilizes the receptor/G protein complex so that antiserum 8132 can no longer influence the affinity state of the receptor.
Delta receptor association with PTX-insensitive G proteins.
The possibility that PTX-insensitive G proteins associate with delta opioid receptors was investigated with PTX and the nonhydrolyzable GTP analog GTPγS. As figure6 indicates, there is a portion of specific high-affinity agonist binding to delta opioid receptors that is unaffected by PTX treatment but is sensitive to GTPγS. Higher concentrations of PTX did not further reduce125I-β-endorphin binding, and a similar concentration of PTX abolished agonist binding to the cloned somatostatin receptor SSTR2. PTX reduced 125I-β-endorphin binding to approximately half that reduced by GTPγS. These results suggest thatdelta opioid receptors may associate with PTX-insensitive G proteins. To examine this possibility, the immunoprecipitation approach was used to determine whether two of the PTX-insensitive G proteins, Gzα and Gqα, couple with deltaopioid receptors. In the absence of agonist, antiserum 2919 directed against Gzα immunoprecipitates deltareceptor/G protein complexes, whereas antiserum 946 directed against Gqα uncouples these complexes (fig. 7, A and B). Therefore, both of the PTX-insensitive G proteins tested appear to associate with delta opioid receptors. In the presence of the opioid agonist DPDPE, the interaction between the deltareceptor and Gzα remains unchanged, whereas the association of Gqα is altered so that antiserum 946 is able to immunoprecipitate instead of uncouple the deltareceptor/G protein complexes (fig. 7, C and D). This is the first report that delta opioid receptors physically associate with PTX-insensitive G proteins and that there are differences in these associations depending on the presence of agonist.
The effect of GTPγS and PTX pretreatment on the specific binding of 125I-β-endorphin to the cloneddelta opioid receptor. CHO cells (DG44 variant) expressing either the cloned delta opioid receptor (open bars) or the cloned somatostatin receptor SSTR2 (hatched bars) (S) were either pretreated overnight with PTX or with medium alone. Deltareceptors were detected by the binding of the agonist125I-β-endorphin. SSTR2 was detected with the somatostatin agonist 125I-MK678 (specific binding defined with 1 μM SRIF). Binding to SSTR2 was used as a control to establish the effectiveness of the PTX treatment that abolished specific 125I MK678 binding. Specific binding of the radioligands to membranes from control cells or cells pretreated with PTX was measured in either the presence or absence of 10 μM GTPγS. Data are the mean ± S.E.M. of three different experiments.
The association of Gzα and Gqα with the cloned delta opioid receptor.Delta opioid receptor/G protein complexes treated with or without the delta receptor agonist DPDPE and then solubilized were immunoprecipitated with various G protein antisera.Delta opioid receptors in the immunoprecipitate or supernatant were detected with 125I-β-endorphin, and the resulting specific binding was compared with nonimmune serum. Data are the mean ± S.E.M. specific cpm bound in four different experiments. A, In the absence of agonist, delta receptor/G protein complexes were immunoprecipitated with antisera against Gzα (2919), Gqα (946) or nonimmune serum (NI). Open bars, specific binding in the immunoprecipitate. B, Hatched bars, specific binding in the supernatants from A. C, In the presence of the delta opioid agonist DPDPE, deltareceptor/G protein complexes were immunoprecipitated with the antisera listed in A. Open bars, specific binding in the immunoprecipitate. D, Hatched bars, specific binding in the supernatants from C.
Discussion
Delta opioid receptors mediate the physiological actions of endogenous opioid neurotransmitters by coupling to various effector systems, such as adenylyl cyclase (Chen et al., 1993; Evans et al., 1992; Kieffer et al., 1992;Sharma et al., 1975; Yasuda et al., 1993), Ca++ and K+ channels (Gross and MacDonald, 1987; Hescheler et al., 1987; North, 1993; Wimpey and Chavkin, 1991) and the Na+/H+ exchanger (Isomet al., 1987). G proteins link delta opioid receptors to these effector systems, and G proteins may determine which effector systems are activated.
In this study, the physical interactions between deltaopioid receptors and G proteins were investigated. Both Giα and Goα associate with deltareceptors, and changes in receptor/G protein association occur after agonist binding to the receptor. These changes may be involved in the activation of the delta receptor signal transduction pathway. An important new finding is that PTX-insensitive G proteins couple to the delta receptor. These interactions may be involved in mediating the effects of delta agonists on phospholipase C activation and changes in intracellular phosphoinositol turnover. In fact, Tsu et al. (1995) recently reported that Gzα can couple the cloned delta receptor to adenylyl cyclase and phospholipase C.
In our studies, we used a CHO cell line transfected to express the cloned mouse delta receptor. The levels of receptor in these cells are higher than those found naturally in the body. However, it is unlikely that the high levels of receptor resulted in a distorted G protein association. Law et al. (1994a) reported that the cloned delta receptor expressed in CHO cells at 1.4 (comparable to our levels of 1.25 pmol), 0.7 and 0.27 pmol/mg of protein gave similar levels of DPDPE-induced inhibition of cAMP accumulation (83–61%), indicating that varying the densities ofdelta receptor did not significantly affect the interaction of the receptor with the G proteins involved in coupling the receptor to adenylyl cyclase. Furthermore, the agonist affinities for the cloneddelta receptor in our CHO cells is comparable to those reported by Evans et al. (1992), Kieffer et al.(1992), Yasuda et al. (1993), Tsu et al. (1995)and Blake et al. (1995), who performed their studies in different cells and at different receptor densities. Because agonist affinities are dependent on G protein association, the similar agonist affinities in our cells and those of others suggest that the receptor/G protein associations we identified are not unusual because of the high receptor densities.
A carboxyl-terminally directed antiserum against Giα, 8730, did not immunoprecipitate deltareceptor/Giα complexes in the absence of agonist. The inability of antiserum 8730 to immunoprecipitate receptor/Giα complexes in the absence of agonist was not due to a lack of Giα association with the receptor because antisera directed against internal regions of Giα1 and Giα3 were able to immunoprecipitate or uncouple the deltareceptor/Giα complex. This suggests that the inability of antiserum 8730 to immunoprecipitate or uncouple the agonist free receptor/Giα complex is likely to be due to the lack of epitope accessibility. This is in contrast to the Goαantiserum 9072, which is directed against the same epitope as 8730 that is a domain of Ga that differs in amino acid sequence between Giα and Goα and uncoupled the agonist free delta receptor from Goα. These findings imply that differences exist in the coupling of thedelta receptor with Giα and Goα.
The binding of agonist to the delta receptor permitted antiserum 8730 to coimmunoprecipitate the delta receptor. Agonist binding to the delta receptor may therefore increase the accessibility of the 8730 epitope in Giα. This alteration could be due to conformational changes in the carboxyl terminus of Giα, which result from agonist association with the delta opioid receptor. The carboxyl terminus of Giα subunits is thought to be important for association with receptors, and recent results with somatostatin receptors have shown the carboxyl terminus is essential for mediating somatostatin inhibition of adenylyl cyclase activity (Law et al., 1994b).
The physical association of Goα with the deltaopioid receptor is consistent with results of Roerig et al.(1992), who showed in NG108 cells that delta agonists increase the ADP-ribosylation of Goα. Our results are also consistent with functional studies showing that thisalpha subunit selectively couples delta receptors to Ca++ channels (Hescheler et al., 1987;Taussig et al., 1992). Goα was the predominantalpha subunit associated with the agonist-freedelta receptor. Agonist binding to the receptor reduced the proportion of delta receptor coupled to Goαcompared with Giα. This shift in association of the receptor with Goα to Giα may further reflect the dynamic changes that result from receptor activation.
In the absence of agonist, Giα1 was found to associate with the delta receptor. However, in the presence of agonist, much less association of Giα1 with thedelta receptor was detected. These findings suggest two possibilities: either agonist binding to the receptor may promote the dissociation of Giα1 from the receptor, or the epitope for the Giα1-directed antiserum becomes inaccessible on agonist binding. The former possibility is more plausible given that the Giα1- and Giα2-directed antisera were generated against the same region of Gα, which differs in sequence, and that the Giα2-directed antiserum can immunoprecipitate receptor/G protein complexes in the presence of agonist. Thus, both Giα1 and Giα2 associate with delta opioid receptors. However, agonist binding to the receptors may cause Giα1 to dissociate and Giα2 to associate with the receptor. Both Giα1 and Giα2 have been proposed to coupledelta receptors to adenylyl cyclase. The possible dissociation of Giα1 from the receptor after agonist binding might be expected to increase accessibility of Giα1 to effector systems. The coupling of Giα2 with the receptor might link the receptor to membrane-associated forms of adenylyl cyclase. Our findings are consistent with the results of previous studies showing a role for both Giα1 and Giα2 in delta receptor signaling and provide the first evidence for dynamic changes indelta receptor/Giα interaction induced by agonists.
Preliminary studies (Blake et al., 1995) have shown that the cloned delta receptor expressed in HEK 293 cells couples to adenylyl cyclase and mediates agonist inhibition of cAMP formation. These cells have been shown previously to express Giα1and Giα3 immunoreactivity but not Giα2 or Goα immunoreactivity (Law et al., 1993), suggesting that Giα2 is not required for deltareceptors to couple to adenylyl cyclase and that either Giα1 and/or Giα3 can couple the receptor to this enzyme. These studies do not exclude a role for Giα2in coupling delta receptors to adenylyl cyclase. Behavioral studies (Sanchez-Blazquez and Garcon, 1995) have suggested that Giα2 is critical for mediating the analgesic effects ofdelta-selective agonists because a “knockdown” of Giα2 in mice using antisense blocked deltaagonist-induced analgesia.
Giα3 was the only Giα subtype consistantly associated with both the agonist-free and -bound receptor. Agonist binding did increase the apparent stability of the deltareceptor/Giα3 complexes because they were only coimmunoprecipitated in the presence of agonist. The ability of Giα3 to associate with the receptor and the lack of coupling of Giα1 with the agonist-bound deltareceptor are of interest because these proteins are 94% identical in amino acid sequence, indicating that only a few residues are responsible for the major differences in physical association with thedelta receptor. In contrast, the cloned somatostatin receptor SSTR3 was coupled to adenylyl cyclase by Giα1but not Giα2 or Giα3 (Law et al., 1994b). Therefore, receptor-specific coupling to various Gα subunits may be due to the recognition by receptor contacts of divergent regions of these subunits.
Our results showing that delta receptors associate with Giα2 and Giα3 are consistent with those reported by Roerig et al. (1992) in NG108 cells, which showed that agonists can increase ADP-ribosylation of Giα2 and Giα3. Furthermore, McKenzie and Milligan (1990) reported that delta receptors in NG108 cells interact with Giα2 to transduce agonist inhibition of adenylyl cyclase activity. Unique to our findings is the association of the delta receptor with Giα1, a G protein that was not found to be expressed in NG108 cells by McKenzie and Milligan (1990). Because brain does express Giα1, our studies reveal for the first time the potential importance of this G protein indelta receptor signaling.
Gβ subunits are important for the modulation of particular effector systems such as adenylyl cyclase (Tang and Gilman, 1991; Taussig et al., 1993). Thus, the determination of specific Gβ coupling with receptors may yield clues regarding which of the increasing number of Gβ-mediated effector functions might be activated by a particular receptor. Both Gβ1 and Gβ2 subunits formed stable complexes with the delta receptor. This multiplicity of Gβ subunit association with the delta receptor differs from the findings with the somatostatin and alpha-2 adrenergic receptors, which primarily associate with Gβ1(Law et al., 1991; Okuma and Reisine, 1992). The additional coupling of Gβ2 with the delta receptor may allow it to couple with a different set of effector systems not activated by somatostatin or alpha-2 adrenergic receptors. The potentiation of specific high-affinity delta receptor binding with the addition of the Gβ1-directed antiserum may indicate that this antiserum modulates various deltareceptor/G protein contacts in the absence of agonist. However, on agonist binding, a more stable complex may be formed between the receptor and Gα/β subunits that does not allow Gβ1-directed antiserum to modify the complex. Further study of this region involving site-directed mutagenesis would be useful because there is only one amino acid change in the Gβ1 vs. Gβ2 epitope domain.
The study of opioid signal transduction pathways has primarily focused on PTX-sensitive signaling mechanisms. However, we provide data indicating that delta opioid receptors physically associate with PTX-insensitive G proteins. Delta receptors in both the presence and absence of agonist associate with two of the PTX-insensitive G proteins, Gzα and Gqα. These two proteins in turn may lead to the activation of signaling systems such as phospholipase C. In fact, recent studies have shown that the cloned delta opioid receptor expressed inXenopus oocytes can mediate agonist activation of phospholipase C (Miyamae et al., 1993). Furthermore, thedelta receptor-selective agonist DPDPE has been reported to increase intracellular Ca++ levels in NG108 cells, presumably via a phospholipase C mechanism (Connor et al., 1994). The cloned delta receptor coexpressed with Gzα in 293 cells was reported to couple to adenylyl cyclase in a PTX-insensitive manner (Tsu et al., 1995). Furthermore, Gzα reconstituted delta receptor coupling to phospholipase C in Ltk− cells (Tsu et al., 1995). Therefore, Gzα and possibility Gqα may serve critical roles in delta receptor signaling.
We have shown that there are dynamic changes in G protein association with the delta opioid receptor after agonist binding and that delta receptors associate with PTX-insensitive as well as -sensitive G proteins. Thus, the specific association of particular G proteins with the delta opioid receptor may provide the diversity and direction for the various opioid-attributed signaling events.
Acknowledgments
This work was supported by National Institutes of Health Grant DA-08951.
Footnotes
-
Send reprint requests to: Dr. Terry Reisine, Department of Pharmacology, University of Pennsylvania School of Medicine, 36th and Hamilton Walk, Philadelphia, PA 19104.
-
↵1 D. R. Manning, personal communication.
- Abbreviations:
- PTX
- pertussis toxin
- CHO
- Chinese hamster ovary
- DPDPE
- cyclic [d-Pen2, d-Pen5]enkephalin, DSLET, [d-ser2, d-leu5]enkephalin-Thr6
-
- Received March 6, 1996.
- Accepted January 31, 1997.
- The American Society for Pharmacology and Experimental Therapeutics
