Abstract
In the current study, we investigated the neurochemical basis for the previously reported predominance of stimulatory μ-opioid signaling in guinea pig longitudinal muscle/myenteric plexus (LMMP) preparations after chronic in vivo morphine exposure. As expected, recombinant Gsα (rGsα) dose-dependently stimulated adenylyl cyclase (AC) activity in LMMP membranes obtained from opioid naive as well as tolerant LMMP tissue. However, the magnitude of the increase was significantly greater in the latter than in the former. The Gβγ blocking peptide QEHA (50 μm) essentially abolished stimulation by rGsα in LMMP membranes obtained from both opioid naive and tolerant animals. Interestingly, after partial blockade by lower QEHA concentrations, the incremental AC stimulation by rGsα in tolerant LMMP membranes was no longer observed, indicating augmented Gβγ stimulatory responsiveness. Concomitant changes in the content of AC isoform protein are consistent with these biochemical observations. After chronic systemic morphine, AC protein is augmented significantly (56%). This increment is most likely to be composed of AC isoforms that are stimulated by Gβγ. This is the first demonstration in a complex mammalian tissue that persistent activation of opioid receptors results in augmented Gβγ/Gsα AC stimulatory interactiveness. The relevance of such changes to the manifestation of opioid tolerance is discussed.
Despite the delineation of many of the physiological, biochemical, and molecular biological sequelae of persistent exposure to opioids, elucidation of the biochemical underpinnings of tolerance formation remains incomplete. AC superactivation (Sharma et al., 1975), diminution of spare opioid receptors (Chavkin and Goldstein, 1984), altered opioid receptor density (Chakrabarti et al., 1995), and altered G protein content (Ammer and Schulz, 1993) have long been explored as possible biochemical bases for the physiological sequelae of chronic morphine exposure. Additionally, we have focused attention on the contribution to opioid tolerant mechanisms of the recently discovered ability of opioids to regulate transmitter release and second messenger formation bimodally. These studies made use of the guinea pig ileum LMMP preparation. This tissue contains both enkephalin and dynorphin neurons, as well as a considerable density of the three predominant types of opioid receptor (μ, δ, and κ). It has long been used as a model system for acute and chronic effects of opioids.
LMMP opioid bimodality is concentration dependent (Gintzler and Xu, 1991; Wang and Gintzler, 1994). Low concentrations of sufentanil (nanomolar) facilitate evoked transmitter release (Gintzler and Xu, 1991) and cAMP accumulation (Wang and Gintzler, 1994), whereas higher concentrations (micromolar) are inhibitory. After chronicin vivo exposure to morphine, there is a qualitative in shift opioid responsiveness such that previously inhibitory concentrations of sufentanil facilitate neuronal function (Gintzleret al., 1987; Wang and Gintzler, 1995).
The shift from inhibitory to excitatory opioid modulation after chronicin vivo treatment with morphine could result from the loss of inhibitory responsiveness. This possibility, however, was eliminated by the observation that in these preparations, there seems to be a paradoxical augmentation of μ-opioid receptor coupling to the inhibitory (presumably Gi mediated) opioid pathway (Wang and Gintzler, 1997). Alternatively, the qualitative shift in opioid responsiveness could result from an augmentation of the opioid-facilitatory, Gs pathway. However, the inability of CTX to mitigate high dose (micromolar) sufentanil facilitation of AC activity in tolerant/dependent LMMP tissue (Wang and Gintzler, 1997) argues against the relevance of this mechanism. Such considerations fuel the hypothesis that chronic in vivoexposure to morphine results in the amplification or induction of an opioid receptor-coupled signal transduction pathway or pathways that are either poorly expressed in or absent from opioid naive LMMP tissue.
Consistent with this formulation, we recently demonstrated that chronic morphine augments LMMP levels of mRNA encoding AC in an isoform selective-fashion. Levels of mRNA encoding AC I do not change, but that encoding AC IV is increased significantly (Rivera and Gintzler, 1998). This has substantial qualitative implications regarding the consequences on AC activity of opioid receptor/Gisignal transduction because the former is inhibited by both Giα and Gβγ (Tang and Gilman, 1991a; Taussig et al., 1993a, 1993b), whereas the latter is activated conditionally by Gβγderived from Gi proteins (Gao and Gilman, 1991). Indeed, the relevance of AC isoform-specific Gβγ stimulatory signaling to the directionality (excitatory or inhibitory) of receptor-coupled regulation of AC activity has been demonstrated amply in several cell lines (Federman et al., 1992; Matsuoka et al., 1994; Avidor-Reiss et al., 1997).
Collectively, these observations support the conceptual rubric that chronic morphine exposure induces a shift in the relative preponderance of opioid receptor/Giα inhibition to Gβγ (Gi-derived) stimulation of AC activity (Gsα dependent). This formulation presupposes enhanced AC Gβγstimulatory responsiveness in chronic morphine-treated tissue. Accordingly, the current study investigates whether such augmentation could be detected in LMMP tissue after chronic in vivomorphine exposure. Additionally, the occurrence of concomitant alterations in AC isoform protein as well as in G protein subunits, consistent with such changes in AC signaling was investigated.
Materials and Methods
Chronic in vivo morphine exposure and tissue preparation.
Male albino guinea pigs weighing 375 to 450 g were used. Chronic in vivo morphine exposure (6 days) was accomplished via the subcutaneous implantation of five morphine pellets (75 mg of morphine base each; generously supplied by National Institutes of Drug Abuse) into animals under light anesthesia [ketamine/xylazine (7:1)]. LMMP tissue were prepared, in parallel, from control and morphine-treated ilea as described previously (Wang and Gintzler, 1997).
Preparation of LMMP membranes for assessment of AC activity.
LMMP tissue was homogenized in Na-HEPES buffer (25 mm, pH 8.0) containing 10% sucrose, 1 mm EDTA, and the protease inhibitors bacitracin (100 mg/liter), leupeptin and soybean trypsin inhibitors (3.2 mg/liter each), aprotinin (0.5 mg/liter), and benzamidine (1 mm). Supernatant obtained from low speed centrifugation (1000 × g, 10 min, 4°) was subjected to higher speed centrifugation (30,000 × g, 30 min, 4°). The resulting membrane pellet was resuspended in the same buffer containing 8% sucrose, aliquoted, and stored at −70° for future use.
Gsα stimulation of AC activity.
Recombinant Gsα (rGsα) was expressed in and purified from Escherichia coli (Leeet al., 1994). rGsα (5 μg) was activated by incubation (60 min at 30°) with 100 μmGTPγS in 50 mm Na-HEPES buffer, pH 8.0, containing 1 mm EDTA, 1 mm DTT, 5 mmMgSO4, and BSA (1.25 mg/ml) as described previously (Tang et al., 1991). Separation from free GTPγS was achieved through gel filtration (Sephadex G-25). Recovery of activated rGsα (∼50%) was determined using [35S]GTPγS (New England Nuclear, Boston, MA). To determine whether any altered stimulatory responsiveness to Gsα requires the participation of Gβγ subunits, the effects of the Gβγ blocking peptide QEHA (Chen et al., 1995) on activated rGsα stimulation of AC also were investigated. Stimulatory responsiveness to rGsα was determined after treatment of the membranes with QEHA (50 μm; 20 min on ice) or its inactive control peptide, SKEE (50 μm).
Determination of AC activity in membrane preparations.
AC activity was determined by measuring the synthesis of [32P]cAMP from [α-32P]ATP (ICN, Costa Mesa, CA). Assays were initiated by the addition of the reaction mixture (50 mmHEPES buffer, pH 8.0, containing 10 mmMgCl2, 20 mm creatine phosphate, 10 units/sample creatine phosphokinase, 0.1 mm ATP, 10 μm GTP, 20 mm NaCl, 1 mm DTT, 50 mm EGTA, 0.125 mm rolipram, 0.1% BSA, and [α-32P]ATP; 1 μCi/sample) to cell membranes (10 μg) with or without prior incubation (30°, 10 min) with activated rGαs. Reactions (30°, 15 min) were terminated by the addition of 10 μl of 2.2 n HCl (4°). Thereafter, [32P]cAMP generated was separated by neutral alumina column chromatography as described previously (Alvarez and Daniels, 1990). [3H]cAMP (0.005 μCi), added to each sample before fractionation, was used as an internal standard to correct for column recovery of the [32P]cAMP. Radioisotopes were quantified using liquid scintillation spectroscopy (LKB 1209).
Production and characterization of AC IIC2 (6C6) monoclonal antibody.
The C2 domain of rat AC type II (IIC2 protein) was expressed in E. coli BL21(DE3) and purified by Ni-NTA affinity chromatography followed by Q-Sepharose column chromatography as described previously (Yan et al., 1996). The purified IIC2 protein (100 μg) was emulsified in complete Freund’s adjuvant and injected subcutaneously into 8-week-old Balb/c mice. Four additional immunization were administered subcutaneously in incomplete Freund’s adjuvant at 2-week intervals. Lymphocytes isolated from immunized mice were fused with murine myeloma cell line SP2/0 using PEG1500. Hybridoma supernatant were screened in enzyme-linked immunosorbent assay plates coated with 100 ng of purified IIC2. Hybridomas were cloned by limiting dilution into 96-well microculture plates containing Dulbecco’s modified Eagle’s medium, 15% fetal bovine serum, and hypoxanthine-thymidine. Of the five positive clones isolated, hybridoma 6C6 was the best antibody producer. Hybridoma supernatant was collected and purified using Q-Sepharose column chromatography. The isotype of heavy and light chain for monoclonal antibody 6C6 was determined to be IgG1 and κ, respectively (Id-sp Kit; Zymed Lab, San Francisco, CA).
To obtain cell membranes with which to determine the specificity of the 6C6 antibody, Sf9 cells were infected with recombinant baculoviruses encoding one of AC types II–VI, 48 hr after which cell membranes were harvested and 20 μg was reduced and alkylated as described previously (Tang et al., 1991). Additionally, constructs for the expression of the C2 domain of AC I, II, VII, VIII, and IX were transformed into E. coli BL21(DE3) and used to extend the 6C6 specificity profile. The resulting clones were used for the induction and preparation of the corresponding C2 domain protein (obtained from the 200,000 × g supernatant ofE. coli lysate; Yan et al., 1996). Plasmid constructs for the expression of IC2 (bovine AC I: amino acids 806-1134) and IIC2 (rat AC II: amino acids 855-1090) were described previously (Yan et al., 1996). The coding sequence for VIIC2 (human AC VII: amino acids 864-1080) was amplified using polymerase chain reaction and cloned to pProExH6 vector, whereas that for VIIIC2 (mouse AC VIII: amino acids 946-1248) and IXC2 (mouse AC IX: amino acids 1011–1353) was amplified and cloned to pProEx-HAH6 vector.
AC isoform selectivity of 6C6 was determined using Western blot analysis. Sf9 cell membranes and E. coli lysates were loaded onto 11% and SDS-7.5% polyacrylamide gels, respectively, followed by electrophoreses and electrotransfer onto a nitrocellulose membrane. Recognition by 6C6 of AC isoforms or C2 domains thereof was visualized using ECL detection (Amersham, Arlington Heights, IL). To validate expression of constructs containing C2 domains, these nitrocellulose membranes also were blotted with antibody 12CA5 (Yanet al., 1996), which recognizes the hemagglutinin epitope.
Preparation of preadsorbed hybridoma supernatant.
Preadsorbed AC monoclonal antibody was prepared through incubation (60 min, 27°) of 5 μg of 6C6 antibody with excess peptide (AC IIC2; 50 μg) in 200 μl of buffer A (20 mm Tris·HCl, pH 8.0, 5 mmβ-mercaptoethanol, 100 mm NaCl) followed by Ni-NTA affinity chromatography (Yan et al., 1996). The column was sequentially eluted with 900 μl of buffer A (FT1A), twice with the same volume of buffer containing 20 mm imidazole (FT2 and FT2A), and once with buffer containing 150 mm imidazole (FT3). An aliquot (50 μl) of each column eluate was electrophoresed (SDS-7.5% polyacrylamide gel) and electrotransferred onto a nitrocellulose membrane. AC antibody protein was visualized using a secondary peroxidase-labeled anti-mouse antibody and ECL detection. The remainder of fractions FT1 and FT1A (from which 6C6 had been preadsorbed) was concentrated to a volume of ∼200 μl, pooled, and used as a negative control in Western blot analysis.
LMMP membrane preparation for Western blot analysis.
Membranes from LMMP tissue obtained from control and morphine-treated animals were prepared at 0–4o as described previously (Sternweis and Robishaw, 1984). Briefly, the LMMP tissue was homogenized in buffer A containing a previously described cocktail of protease inhibitors (Tang et al., 1991). Supernatants from a low speed centrifugation (8,000 × g for 30 min) were recentrifuged (20,000 × g; 60 min). Membranes were resuspended in 300–500 μl of the same buffer, and the protein concentration was determined by the Bradford Assay (BioRad, Hercules, CA).
Gel electrophoresis and Western blot analysis.
Membrane samples were heated (80° for 5 min) in Laemmli’s buffer (50 mm Tris·HCl, pH 6.8, containing 2% SDS and 0.2 mm DTT), after which they were treated with 50 mmN-ethylmaleimide (10 min at room temperature) and subjected to SDS-polyacrylamide gel electrophoresis. Gels were electroeluted onto nitrocellulose membranes, which were incubated with blocking solution (5% skim milk in Tris-buffered saline containing 0.1% Tween-20 at room temperature for 1 hr). Thereafter, selected lanes were incubated (room temperature for 2 hr) with either the AC monoclonal antibody 6C6 (0.87 μg/4 ml) or the polyclonal G protein subunit antibodies (Makhlouf et al., 1996) BC-1 (anti-Gβ, carboxyl terminus, 1:2000) or ASC-A (anti-Gsα, carboxyl terminus, 1:1500), each dissolved in blocking solution. The secondary antibody used was either peroxidase-labeled anti-mouse (for 6C6) or anti- rabbit (for G protein subunits). Antibody/substrate complex was visualized using ECL detection. Sample pairs, obtained from opioid naive and chronic morphine-treated LMMP tissues, were processed, electrophoresed, and blotted in parallel, after which they were exposed concomitantly to the same X-ray film (X-Omat film, Kodak). Signal intensity was quantified using NIH imaging software.
Preparation of total RNA extract.
LMMP tissue was homogenized in 5 volumes of guanidinium thiocyanate (5.7 m) containing sodium citrate (5 mm, pH 7.0) and β-mercaptoethanol (1 mm). Extraction of total RNA via CsCl centrifugation and ethanol precipitation was performed as described previously (Rivera and Gintzler, 1998). LMMP tissues from control and morphine-treated animals were processed in parallel. All RNA preparations yielded ratio of absorption (260:280 nm) of ∼2.0.
Preparation of RNA probes.
Full-length cDNA for rat AC type II (4008 bp; Feinstein et al., 1991) was provided as a plasmid in pBluescript SK− (courtesy of Dr. Randall Reed). The excised cDNA/HindIII/SacI fragment was subcloned into the pSP64 and pSP65 vectors. Restriction enzyme mapping indicated that the AC II/pSP64 and AC II/pSP65 plasmids that were obtained were of the sense and antisense orientation, respectively. Both plasmids were linearized with SacI andHindIII to generate the respective templates. An SP6 transcription system (Promega, Madison, WI) and [α-32P]GTP (New England Nuclear, Boston, MA) were used to form radiolabeled AC II (∼8.2 × 108 cpm/μg) riboprobe antisense. CF11 column chromatography was used to purify the RNA transcripts (Franklin, 1966).
Solution hybridization assay for mRNA encoding AC II.
The solution hybridization/ribonuclease treatment/TCA precipitation protocol used to quantify RNA levels was a modification of that described previously (Franklin et al., 1991) as reported earlier (Rivera and Gintzler, 1998). Briefly, duplicate aliquots of total RNA extract (40 μg) were incubated (75° overnight) with AC II32P-labeled riboprobe (200,000 cpm) contained in 2× TESS (20 mm TES, pH 7.4, 20 mm EDTA, 0.6m NaCl, 1% SDS). Thereafter, samples were incubated (30° for 1 hr) with a high salt buffer (10 mm Tris·HCl, pH 7.5, 0.3 m NaCl, 5 mm EDTA) containing 40 μg/ml RNase A (Worthington Biochemicals, Freehold, NJ) and 2 μg/ml RNase T1 (Calbiochem, San Diego, CA). TCA precipitates, obtained via the addition (1 ml) of 5% TCA containing 0.75% sodium pyrophosphate and one drop of 0.5% BSA, were collected onto glass microfiber filters. Radioactivity was quantified by liquid scintillation spectroscopy. Extract from a chronic morphine-treated LMMP tissue always was assayed in parallel with its paired opioid naive tissue. AC II mRNA levels in LMMP tissue extract were quantified from an AC II mRNA standard curve that was generated using the “sense transcript” (5, 10, 20, 40, 60, 80, 160, 320 pg) containing 10 μg ofE. coli tRNA as a carrier as described previously (Rivera and Gintzler, 1998). Standard AC II RNA was synthesized using unlabeled ribonucleotides and the SP6 RNA polymerase. Regression analysis of the standards was carried out to determine the least-squares best fit, from which values for assigning concentrations to the tissue samples were obtained by interpolation.
Northern blot analysis.
Total RNA from LMMP tissue as well as AC I and AC IV “sense transcript” was subjected to 1% agarose/2.2 m formaldehyde gel electrophoresis and transferred onto nitrocellulose membranes as described previously (Rivera and Gintzler, 1998). Membranes were hybridized with an α-32P-GTP-labeled cRNA probe specific for mRNA encoding AC II (106 cpm/ml; 60° overnight in sodium phosphate buffer, 0.4 m, pH 7.0, containing 3% SDS, 1 mm EDTA, 43% formamide, 0.025% acetylated BSA). Subsequently, membranes were washed once for 5 min followed by a 20-min wash (60°) and a 30-min wash (75°) with 0.5× standard saline citrate buffer (15 mm sodium citrate, 150 mmsodium chloride, pH 7.2) containing 1% SDS and 1 mm EDTA. Hybridized signal was visualized by autoradiography using storage PhosphorImaging techniques (Molecular Dynamics, Sunnyvale, CA).
Results
Effect of chronic morphine on Gsα stimulation of AC activity.
As expected, rGsα stimulated AC activity in LMMP membranes obtained from both opioid naive and tolerant LMMP tissue in a dose-dependent manner (p < 0.05 for all concentrations of rGsα; Fig.1A). Of particular relevance to the current study, the magnitude of the increase in the latter group was significantly greater than that observed in the former compared with all rGsα concentrations tested (1, 2.5, 5, 10 nm). However, the increment in Gsαstimulation between groups did not vary across concentrations tested. A two-way analysis of variance revealed a group and dose effect (F1,6 = 10.25 and F3,18 = 119.28; p < 0.02 for both) but no group × dose interaction (F3,18 = 2.77; p > 0.05).
Effect of the Gβγ blocking peptide QEHA on Gsα stimulation of AC.
To investigate the role of Gβγ in the altered Gsα stimulation of AC after chronic morphine, the effect of the Gβγ blocking peptide QEHA (a peptide corresponding to residues 956–982 of AC II; Chen et al., 1995) on Gsα-stimulated AC activity was determined in opioid naive and tolerant membranes (Fig. 1B). A two-way analysis of variance revealed a dose effect of QEHA (F1,10 = 186.39, p < 0.001). QEHA (50 μm) essentially abolished stimulation by Gsα in membranes obtained from opioid naive and morphine-treated LMMP membranes. Lower QEHA concentrations (10 and 20 μm) incompletely attenuated Gsαstimulation of AC activity. Interestingly, after partial blockade by QEHA, the difference in the magnitude of AC stimulation by Gsα between naive and tolerant LMMP membranes was no longer observed (F1,4 = 1.83,p > 0.2; Fig. 1A, inset). This effect did not vary across concentrations of QEHA (F1 10 = 1.64; p > 0.2). Moreover, in LMMP membranes obtained from four animals (two opioid naive and two chronic morphine treated), coincubation with QEHA and purified Gβγ (50 nm) reversed the QEHA block by 69–99%. In these experiments, the concentration of Lubrol PX present in the buffer containing Gβγ was maintained at 0.01% in all assays.
Interestingly, QEHA failed to lower the basal AC activity in either control or morphine-treated preparations despite the presence of 10 μm GTP, a concentration sufficient to generate activated Gsα and Gβγ. However, it should be noted that higher concentrations of GTP (up to 100 μm) failed to produce any further increase in basal AC activity in both preparations (three experiments), which remained significantly lower than that observed after the addition of 5 nm activated Gsα. This suggests that although GTP (10 μm) results in the generation of activated Gsα, the concentration achieved is <5 nm and below that required for AC stimulation by Gβγ. In the absence of Gβγ, stimulation of AC, QEHA should not be expected to alter AC activity.
Analogous to Gsα stimulatory responsiveness, forskolin (10 μm) stimulation of AC activity also was augmented after chronic morphine (∼17%, p < 0.05, Fig. 1B). However, in contrast to Gsαstimulation of AC activity, QEHA (50 μm) did not alter the forskolin activation of this enzyme (Fig. 1B). To explore further the specificity of the effect of QEHA, the effect of a related peptide, SKEE, was determined. This peptide corresponds to the region of AC III (not stimulated by Gβγ) that is analogous to the Gβγ binding region on AC II (stimulated by Gβγ), the amino acid sequence of which composes QEHA (Chen et al., 1995). A concentration of SKEE equal to that used for QEHA (50 μm) was devoid of any effect on rGsα AC stimulation (Fig. 1B).
Specificity of monoclonal antibody 6C6 raised against AC IIC2 domain.
Antibody 6C6 was able to detect recombinant types II and IV ACs but not types III, V, VI, or rutabaga (Fig.2A) in membranes obtained from Sf9 cells transfected with these isoforms. Western blot analysis of the C2 protein domain of various adenylyl cyclases expressed in E.coli indicated that 6C6 recognized IC2, IIC2, and VIIC2 but not VIIIC2 and IXC2 (Fig. 2B). Expression of VIIIC2 and IXC2 was confirmed by demonstrating that lysate obtained from E.coli infected with either gave a positive signal, at the expected molecular mass, when blotted with antibody 12CA5 (Fig. 2C; Yanet al., 1996). Thus, antibody 6C6 recognizes the type II AC enzyme family (type II, IV, and VII) and AC I but not AC types III, V, VI, VIII, IX, and rutabaga.
Specificity of AC Western blot analysis.
Specificity of the signal visualized using 6C6 antibody in Western blot analysis was determined using preadsorbed hybridoma cell extract. Fig.3A illustrates the elution profile of the adsorbed AC monoclonal antibody from a Ni-NTA column that had been incubated with AC antibody/IIC2 peptide complex. The absence of any signal after incubation with goat anti-mouse immunoglobulin indicates that column eluate obtained with imidazole-free buffer (FT1 and FT1A) was devoid of all immunoglobulin. In contrast, AC antibody was eluted with buffer containing imidazole (fractions FT2 and FT3); therefore, fractions FT1 and FT1A were pooled and used for adsorption control immunoblotting.
Fig. 3B illustrates the Western blots obtained using either preadsorbed hybridoma cell extracts (FT1 and FT1A) or unabsorbed 6C6 monoclonal antibody that was equivalently diluted. The predominant signal visualized by Western blot analysis using unadsorbed 6C6 had a molecular mass of ∼150 kDa. In addition, a faster moving band was detected at ∼140 kDa (lane 1). Neither of these bands was detected when LMMP extract was blotted with preadsorbed hybridoma extract (Fig. 3B, lane 2). The molecular weight difference between LMMP ACs recognized by 6C6 versus that expressed in Sf9 cells (Fig. 2A) most likely is due to variations in the degree of glycosylation, sites for which are present in ACs. Indeed, it has been shown that the molecular masses of ACs vary, depending on the source (Mollner and Pfeuffer, 1988)
Effect of chronic morphine treatment on AC protein.
The signal intensity of the 150-kDa band was augmented significantly in LMMP extract obtained from guinea pigs chronically treated with morphine (56%; p < 0.05, four experiments; Fig. 3C). The content of the 140-kDa species was quite variable among experiments. In fact, it was detectable in only two of four experiments. Therefore, a comparison of signal intensity between samples obtained from opioid naive and chronic morphine-treated tissues was restricted to the larger molecular mass species.
Effect of chronic morphine treatment on G protein subunits.
Western blot analyses of Gsα and Gβ (Fig. 4A) revealed two bands for the former (45 and 48 kDa) and one for the latter (36 kDa). This is consistent with results from previous studies using these as well as other antibodies (Makhlouf et al., 1996; Ammer and Schulz, 1997). Neither the intensity of the two immunoreactive bands corresponding to Gsα nor that corresponding to Gβ differed among cholate extracts of membranes obtained from opioid naive versus chronic morphine-treated LMMP tissue (lane 1 versus lane 2 for each blot).
Specificity of the AC II riboprobe.
Northern blot analysis was used to determine the size of AC II mRNA in guinea pig LMMP, as well as the specificity of the radiolabeled riboprobe used in these experiments (Fig. 4B). For AC II, Northern blots of total RNA obtained from guinea pig LMMP tissue revealed an abundant hybridized transcript of ∼4.0 kb (Fig. 4B, lane 4). This is consistent with the 4.1-kb size reported previously for the type II isoform isolated from rat brain and olfactory bulb (Feinstein et al., 1991), as well as the size of the sense AC II RNA transcribed in this study (Fig. 4B, lane 3). Additionally, a hybridized transcript estimated to be 1.9 kb was observed. Because it survived high stringency washes (1% SDS, 0.5× standard saline citrate buffer, 1 mm EDTA at 70° for 60 min), it most likely does not result from nonspecific hybridization. This band could represent splice variants of the original transcript or alternative transcription initiation sites. An analogous pattern of the size distribution for AC II mRNA has been reported for rat tissues (Feinstein et al., 1991).
Specificity of the riboprobe for AC II was evaluated further by assessing the extent to which AC II antisense cross-hybridizes with AC I or AC IV sense. Northern blot analyses (Fig. 4B, lanes 1 and 2) and solution hybridization (data not shown) revealed that the radiolabeled riboprobe for AC II does not cross-hybridize with sense RNA encoding AC I or IV.
Quantification of AC II mRNA in opioid naive and chronic morphine-treated tissue.
Solution hybridization assays revealed that levels of mRNA encoding AC II in opioid naive LMMP tissue are 0.93 ± 0.16 pg/μg total RNA. After persistent in vivo activation of opioid receptors, levels of AC II mRNA were increased slightly to 1.0 ± 0.10 pg/μg total RNA, but this numerical difference was not statistically significant (p = 0.6, five experiments; see Table1 for a comparison with mRNA encoding AC I and IV). In these assays, the standard curves for AC II sense RNA had a correlation coefficient of ≥0.98. It should be noted that the riboprobe for AC II contains complementary sequences encoding bovine AC II (Feinstein et al., 1991). Because these sequences are, undoubtedly, not perfectly complementary to guinea pig AC II mRNA, their quantification in the current study probably is an underestimation. It is not possible to determine the precise correction factor to be used, so assay values are expressed as “pg equivalents” (Franklin et al., 1991).
Discussion
It has been demonstrated that ACs from murine S49 cells and platelets, insensitive to Gβγ, are stimulated by Gsα (Tang and Gilman, 1992). These observations are consistent with the more recent finding that rGsα stimulated purified AC in the absence of Gβγ (Yan et al., 1997). Therefore, the abolishment by QEHA of rGsα and stimulation of AC in membranes prepared from opioid naive and tolerant tissue indicate that in both preparations, it is mediated via Gβγ (Gsα dependent); that is, stimulation of LMMP AC by exogenous rGsα (1–10 nm) results from its ability to augment stimulation by endogenous Gβγ (Tang and Gilman, 1991b). Thus, chronic morphine-induced augmented Gβγ stimulation (Gsα dependent) of AC is most likely reflected by the increased rGsα stimulatory responsiveness of tolerant preparations. This inference is supported by the ability of submaximal concentrations of QEHA to obliterate the incremental rGsα stimulatory response that occurs after chronic morphine because a similar degree of Gβγ neutralization by QEHA in both preparations would have a greater effect on rGsα stimulation in the latter.
The current study was unable to directly demonstrate augmented Gβγ stimulation of AC activity after chronic morphine treatment. In fact, stimulation of AC activity by exogenous Gβγ in either opioid naive or tolerant LMMP membranes was not detectable. This most likely results from the fact that the LMMP tissue contains multiple AC isoforms, some of which are inhibited by Gβγ. For example, in addition to AC II and IV (which are stimulated by Gβγ), the LMMP tissue also contains AC I (Rivera and Gintzler, 1998). All would be stimulated by activated exogenous Gsα. However, Gβγ inhibition of AC I (Tang and Gilman, 1991b) could mask any augmented Gβγstimulation of ACs of the type II family. In this regard, it is relevant to note that to date, virtually all studies demonstrating direct Gβγ stimulation of AC activity have been conducted with cultured cells overexpressing a single AC isoform.
The effect of chronic morphine on LMMP levels of AC isoform-specific mRNA and AC isoform protein is consistent with the augmented Gβγ stimulatory AC signaling demonstrated in the current study. Furthermore, these data suggest a mechanism by which such changes could occur. The antibody used to quantify AC protein recognizes ACs of the type II family (AC II, IV, and VII), all of which are stimulated by Gβγ, and AC I. It is unlikely that cross-reactivity with AC I confounds interpretation of the Western blot analyses because levels of mRNA encoding the latter do not change after chronic morphine (Rivera and Gintzler, 1998). This supposition is supported by a previous report (Hellevuo et al., 1996) that in several brain areas, the mRNA levels of six AC isoforms were predictive of the regulatory characteristics of AC activity in corresponding membrane preparations. Similarly, it also is unlikely that AC II contributes to the chronic morphine-induced up-regulation of AC protein because its corresponding mRNA fails to increase after this treatment.
In contrast, the LMMP content of AC IV mRNA is increased (∼37%) after chronic systemic morphine treatment (Rivera and Gintzler, 1998). Interestingly, the magnitude of the increase in AC protein observed in the current study (56%) is of comparable magnitude. Thus, it seems plausible to infer that the morphine-dependent increase in AC protein observed in the current study reflects elevated levels of AC IV. The effect of chronic morphine on levels of mRNA encoding AC VII has not been determined. Consequently, its contribution to the morphine-induced up-regulation of AC protein remains unknown. This notwithstanding, distinction between these isoforms is not relevant to the current study because an increase in either would have the same Gβγ regulatory implications (Gao and Gilman, 1991; Yoshimura et al., 1996). Thus, the increased Gβγ stimulatory signaling observed in LMMP tissue obtained from chronic morphine-treated animals most likely results from the elevated content of the Gβγ-stimulated AC isoforms AC IV, VII, or both.
Paradoxically, basal AC activity did not differ between LMMP tissue obtained from opioid naive versus morphine-treated animals, despite the increase in AC protein that occurred in the latter group. However, it should be noted that after agonist removal from chronic morphine-treated COS-7 cells, activity of transfected AC II, IV, or VII not only failed to manifest superactivation but actually decreased (AvidorReiss et al., 1997). This presumably also would occur as a result of the removal of morphine from chronic morphine-treated LMMP tissue during the preparation of LMMP membranes. Thus, any expected increase in basal AC activity resulting from chronic morphine-induced elevated levels of AC IV and/or VII protein could have been masked by a concomitant decrement in AC basal specific activity.
As a consequence of increased levels of AC isoforms of the type II family, a shift from predominantly opioid receptor/Gi inhibitory coupling to opioid receptor Gβγ (Giderived) stimulatory signaling could ensue. Such a shift would be consistent with and could underlie both the CTX-insensitive excitatory effects of previously inhibitory concentrations of sufentanil and the paradoxical increase in μ-opioid receptor coupling to Gi that has been reported in these preparations (Wang and Gintzler, 1997). This mechanism also would explain enhanced Gβγ stimulatory AC responsiveness in the absence of any increase in the LMMP content of this G protein subunit (Fig. 4A).
In a previous study (Lane-Ladd et al., 1997), the effects of chronic morphine on AC isoform content was investigated in one brain region, the locus ceruleus. Chronic morphine was found to up-regulate AC I and AC VIII protein, the magnitude of which was comparable to that observed in the current study. In contrast, levels of AC III, IV, or V were not altered. These findings in combination with the current report suggest that up-regulation of AC protein by chronic morphine may be a generalizable consequence of persistent opioid receptor activation. However, the nature of the specific isoform or isoforms so affected and, consequently, the physiological significance of such changes vary across tissues and, presumably, brain regions.
A qualitative shift in opioid receptor signaling from inhibitory (mediated via Giα) to stimulatory (mediated via Gβγ) would compensate for the blunted stimulated formation of cAMP in the absence of opioid agonist that occurs in chronic morphine-treated LMMP tissue (Wang and Gintzler, 1995), as well as in COS-7 cells transfected with AC isoforms II, IV, and VII (Avidor-Reiss et al., 1997). This would result in the restoration of cAMP levels that is characteristic of opioid naive tissue, despite the continued presence of inhibitory concentrations of opioid (i.e., tolerance) (Wang and Gintzler, 1995).
The importance of the relative abundance of Gβγ-stimulated AC isoform or isoforms to the directionality of opioid responsiveness has been demonstrated in several cell lines. For example, dopamine activation of AC types V and VII can be inhibited and stimulated, respectively, by morphine in a dose-dependent manner (Yoshimura et al., 1996). The stimulatory effects of morphine on AC VII are blocked by pertussis toxin, as well as the α subunit of transducin, suggesting the importance of Gβγ derived from Gi. Similarly, in COS-7 cells, acute opioid treatment inhibits or stimulates transfected AC activity in an isoform-specific fashion (Avidor-Reiss et al., 1997), the pattern of which correlates with that of βγ stimulatory responsiveness.
The current demonstration of enhanced rGsαstimulation of AC activity after chronic morphine is consonant with results obtained using A431 carcinoma cells stably transfected with μ-opioid receptors (Ammer and Schulz, 1997). This study reported increased Gsα stimulation of AC activity after chronic morphine treatment, albeit via a different mechanism than that reported here. In this case, augmented Gsαstimulation of AC activity was attributed to increased Gsα depalmitoylation. The current results do not preclude the occurrence of this mechanism in the LMMP tissue because the consequences of any chemical modification of endogenous Gsα would not be reflected by assessing stimulatory responses to exogenous rGsα, as was done in the current study.
Chronic morphine treatment does not alter the content of Gsα in LMMP tissue (Fig. 4A), as has been observed in A431 carcinoma cells (Ammer and Schulz, 1997). Thus, the previously reported (Wang and Gintzler, 1997) increased low dose (Gsα mediated) sufentanil stimulation of AC activity in the LMMP tissue after chronic morphine most likely results from increased AC stimulatory responsiveness to Gsα. This could result from augmented Gsα-dependent Gβγstimulation of AC, increased Gsαdepalmitoylation, or a combination.
Heretofore, attempts to elucidate neurochemical mechanisms that underlie narcotic tolerance have focused predominantly on altered G proteins and opioid receptor content, as well as changes in their coupling. The current report points to the relevance of augmented AC Gβγ stimulatory signaling to mechanisms that underlie opioid tolerance. Such changes reflect that chronic morphine qualitatively as well as quantitatively alters opioid receptor-coupled signaling.
Acknowledgments
We thank Dr. Ravi Iyengar (Mount Sinai Medical School, New York, NY) for QEHA and SKEE peptides and Dr. John Hildebrandt (Medical University of South Carolina, Charleston, SC) for antibodies against Gsα and Gβ.
Footnotes
-
Send reprint requests to: Dr. Alan Gintzler, Box 8, Department of Biochemistry, SUNY HSCB, 450 Clarkson Avenue, Brooklyn, NY 11203.
-
↵1 Current affiliation: United States Department of Agriculture, ARS, NPA, Roman L. Hruska U.S. Meat Animal Research Center, Clay Center, Nebraska 68933.
-
This work was supported in part by Aaron Diamond Foundation Grant HRI 817–533 (M.R., A.R.G.) and National Institutes of Health Grant GM53459 (W.J.T). M.R. was an Aaron Diamond Foundation Postdoctoral Fellow. S.C., M.R., and S.-Z.Y contributed equally to this article.
- Abbreviations:
- AC
- adenylyl cyclase
- LMMP
- longitudinal muscle/myenteric plexus
- EGTA
- ethylene glycol bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid
- HEPES
- 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid
- SDS
- sodium dodecyl sulfate
- TCA
- trichloroacetic acid
- GTPγS
- guanosine-5′-O-(3-thio)triphosphate
- DTT
- dithiothreitol
- BSA
- bovine serum albumin
- SSC
- standard saline citrate
- TES
- N-Tris[hydroxymethyl] methyl-2-aminoethane sulfonic acid
- rGsα
- recombinant Gsα
- Received April 17, 1998.
- Accepted July 13, 1998.
- The American Society for Pharmacology and Experimental Therapeutics