Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Opiates induce opioid receptor adaptation, which may entail desensitization, internalization, and down-regulation (for reviews, see Roth et al., 1998; Connor and Christie, 1999). Recently, significant advances have been made in our understanding of the mechanisms of short-term receptor adaptation using recombinant opioid receptors expressed in cell lines. It has been discovered that, in addition to being receptor subtype-specific, receptor adaptation is differentially regulated by opioid agonists. For example, the µ-opioid receptor is a target of endogenous opioid peptides, synthetic compounds, and morphinan alkaloids (Matthes et al., 1996). Opioid agonists display remarkable differences in their ability to desensitize and induce the rapid internalization of the µ-opioid receptor both in transfected cells as well as in vivo (for review, see Whistler et al., 1999). Some opiates such as morphine are deficient in their ability to induce the desensitization and rapid endocytosis of receptors. Thus, it has been proposed by Whistler et al. (1999) that one can define ligands on the basis of the ratio of their relative activity versus endocytosis, and this value may correlate with their propensity to induce tolerance and dependence.

However, it is important to distinguish the short-term changes that result from a single (acute) exposure to opioids from the chronic effects of the drugs. The latter gradually develop over time in response to repeated drug treatments, resulting in tolerance/dependence and the addictive state persists for a long time after cessation of the exposure (Nestler and Aghajanian, 1997). These adaptive responses were accompanied by changes in the expression of certain genes after sustained opioid treatment (Nestler and Aghajanian, 1997). Chronic opioid agonists were shown to induce changes in levels of receptors as well as of various proteins of the signal transduction pathway including G proteins, adenylyl cyclase, and protein kinases (for reviews, see Cox, 1993; Gintzler and Chakrabarti, 2000; Law et al., 2000). However, the observed changes often were contradictory, and there is growing data to question prevailing explanations of tolerance. This includes the hypothesis that receptor desensitization and endocytosis are the underlying molecular mechanisms of physiological tolerance. Delineation of the molecular mechanisms involved in drug tolerance/dependence and their temporal interrelationships represent an important scientific challenge with substantial social impact.

It should be remembered that cell lines are models of more complex systems. Whole animal phenomena, such as analgesia or tolerance are mediated by multicellular networks, which have properties that transcend those of their individual components. The complex regulation that occurs in brain by interaction of different neuronal circuits and neurotransmitter systems has been well documented (Noble and Cox, 1996; Roth et al., 1998; Koob, 2000).

Here we studied changes in the subcellular distribution of µ-opioid receptors in morphine-tolerant/dependent rat brains. We prepared membrane fractions originating from the cell surface (SPM) and from "light vesicles" or microsomes (MI) that are enriched in endoplasmic reticulum, Golgi membranes, and endosomes. They have been previously subjected to a detailed characterization by marker enzymes, electron microscopy, and receptor binding experiments (Roth et al., 1981; Moudy et al., 1985; Szücs and Coscia, 1992). The availability of these membrane fractions allowed us to study the subcellular distribution of µ-opioid receptors and their cognate G proteins in morphine-tolerant rat brain. In addition, possible changes in the functional coupling of µ-opioid receptors to G proteins in the subcellular fractions have also been examined.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. Anti-G protein antibodies AS/7 (anti-G_i1,2; 1:500), RM/1 (anti-G_s; 1:500), and GC/2 (anti-G_o; 1:500) were purchased from PerkinElmer Life Sciences (Boston, MA). Guanosine-5'-O-(3-[³⁵S]thio)triphosphate ([³⁵S]GTPS) (37-42 Tbq/mmol) was obtained from Isotope Institute Ltd. (Budapest, Hungary); [-³²P]NAD; specific activity 800 Ci/mmol) was purchased from PerkinElmer Life Sciences. DAMGO and [³H]DAMGO were obtained from Multiple Peptide System (San Diego, CA) via the Drug Supply Program of National Institute on Drug Abuse (Rockville, MD). Urea was purchased form Merck (Darmstadt, Germany); ion-exchange resin for urea purification (AG-501 ×8) was from Bio-Rad (Richmond, CA); low-molecular weight markers were from Pharmacia (Peapack, NJ); and nitrocellulose (Hybond) and ECL were from Amersham Biosciences UK Ltd. (Little Chalfont, Buckinghamshire, UK). All other chemicals were purchased from Sigma-Aldrich (St. Louis, MO).

In Vivo Opioid Treatments. Wistar rats weighing 250 to 350 g were studied using protocols approved by the Animal Care Committee of Szeged University and Biological Research Center. All experiments were performed in freely moving animals during the same period of the day (8:00-13:00 h) to exclude diurnal variations in pharmacological effects. The animals were randomly assigned to treatment groups (n = 6-16 per group), and the observer was blind to the treatment administered. Animals were made dependent on morphine by a series of subcutaneous injections of morphine hydrochloride administered twice daily at 8:00 AM and 6:00 PM for 3 (M3), 5 (M5), and 10 (M10) days. The initial dose, 10 mg/kg, was gradually increased as shown in the paradigm in Table 1. Control animals were handled simultaneously by saline injections.

Hot-Plate and Tail-Flick Tests. Nociceptive sensitivity after different treatments (3, 5, or 10 days morphine administration) was assessed by using hot-plate and tail-flick techniques. The latency of licking one of the hind paws or jumping was measured on the hot-plate (52.5°C, cut-off time, 60 s). The reaction time in the tail-flick test was determined by immersing the lower 5-cm portion of the tail in the hot water until typical tail-withdrawal response was observed (51.5°C water, cut-off time, 20 s). Baseline latencies were obtained immediately before and 30, 60, 90, and 120 min after the drug injection (saline or 10 mg/kg morphine). Analgesic latencies in acute pain tests were converted to percentage of maximum possible effect (%MPE) by using the formula: %MPE = [(observed latency  baseline latency)/(cut-off time  baseline latency)] × 100.

Membrane Preparation. Subcellular fractionations of rat brains were performed as described by Roth et al. (1981) and Szcs and Coscia (1992) with some modifications. In brief, fresh forebrains were gently homogenized, and the homogenates were centrifuged at 1000g as described except that a second 1000g centrifugation was included, and combined supernatants were centrifuged at 12,000g for 20 min. Pellets were suspended in 10% buffered sucrose and subjected to consecutive centrifugations at 20,000g for 25 min and 14,000g for 20 min twice. The resulting pellets were lysed, followed by fractionation on a 10, 28.5, and 34% sucrose density step gradient that was spun at 128,000g for 2 h. Highly enriched SPMs were obtained from the 28/34% interface. Crude MI were obtained from the 12,000g supernatant by consecutive centrifugations at 20,000g for 25 min and 128,000g for 1 h. MI were purified on a 10 and 28.5% sucrose step gradient centrifuged at 128,000g for 2 h and collected at 10/28.5% interface. SPM and MI fractions were diluted 3-fold with Tris-HCl, pH 7.4, pelleted at 128,000g for 1 h, suspended in 50 mM Tris-HCl, pH 7.4, and either freshly used for receptor binding or frozen and stored at 80°C for up to several months for G protein studies.

Receptor Binding. Binding assays were performed with 1 nM [³H]DAMGO, 11 concentrations (10¹⁰-10⁵ M) of unlabeled DAMGO, and the membrane suspensions (200-300 µg of protein) in 50 mM Tris-HCl, pH 7.4, buffer in a final volume of 1 ml. Incubation (25°C, 1 h) was terminated by vacuum filtration through GF/C glass fiber filters (Whatman, Clifton, NJ). Filters were rapidly washed twice with 10 ml of ice-cold 50 mM Tris-HCl (pH 7.4) buffer, air-dried, and counted. Untransformed binding data from homologous displacement curves were analyzed by means of the nonlinear least-squares regression computer program, LIGAND, to obtain K_D (dissociation constant) and B_MAX (densities of binding sites) values (Munson and Rodbard, 1980). To compare changes in receptor number due to agonist treatments, the significance was determined by a t test in B_MAX from matched samples of treated versus control.

ADP Ribosylation. ADP ribosylation of membrane proteins was performed as described (Rottmann et al., 1998). Each sample of approximately 50 µg of membrane protein was subjected to ADP ribosylation with 1 × 10⁷ dpm [-³²P]NAD and 60 µM unlabeled NAD.

Gel Electrophoresis and Immunoblotting. Gel electrophoresis and immunoblotting were performed as described (Fábián et al., 1998). Films were scanned, and data files were evaluated with ImageQuant computer software (version 4.1; Molecular Dynamics, Sunnyvale, CA). The density of immunolabeling was linear with protein levels in the weight range (7.5-60 µg) applied onto the gel (not shown). For radiolabeled samples, gels were dried after electrophoresis and exposed to X-ray films (X-OMAT AR; Eastman Kodak, Rochester, NY). Films were analyzed by laser densitometry [Ultroscan XL enhanced laser densitometer (LKB, Uppsala, Sweden) and GelScan XL laser densitometer program computer software].

[³⁵S]GTPS Functional Assay. Rat brain membrane fractions (10 µg of protein) were incubated in Tris-EGTA buffer (50 mM Tris-HCl, 1 mM EGTA, and 3 mM MgCl₂, pH 7.4) containing 0.05 nM [³⁵S]GTPS and increasing concentrations (10⁸-10⁴ M) of DAMGO in the presence of 100 µM GDP in a total volume of 1 ml for 60 min at 30°C as described by Sim et al. (1996) and Traynor and Nahorski (1995) with slight modifications. Nonspecific binding was determined with 10 µM GTPS and subtracted. Bound and free [³⁵S]GTPS were separated by vacuum filtration through Whatman GF/F filters with a Millipore manifold (Millipore Corporation, Bedford, MA). Filters were washed with 3 × 5 ml of ice-cold buffer, and radioactivity of the dried filters was detected in a toluene-based scintillation cocktail in a Wallac 1409 scintillation counter (Wallac, Turku, Finland). Data were fitted with the Prism 2.01 program (GraphPad Software Inc., San Diego, CA). EC₅₀ values were defined as the concentration of DAMGO producing 50% maximal response.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Antinociceptive Responses. As shown in Fig. 1, a single acute injection of 10 mg/kg morphine produced profound analgesia in both tail-flick and hot-plate tests at each time point tested in morphine naive rats. Chronic administration of increasing doses of morphine for 3, 5, or 10 consecutive days significantly decreased the antinociceptive effect of morphine. Treatment with morphine for 3 days caused significantly lower level of tolerance than the 5- and 10-day treatments. There were no significant differences in the degree of tolerance between the 5- and 10-day treatment groups.

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Hot-plate (A) and tail-flick (B) analgesia tests of saline-treated (control, ), acute morphine (, 10 mg/kg), and 3 (), 5 (), or 10 () days of chronic morphine-treated rats. Shown are the percentage of maximal possible effects (%MPE) of a challenge dose (10 mg/kg) of morphine at different times after the injection. Two-way analysis of variance of data for repeated measures was used to test for overall effects. Statistical comparison of the saline- and morphine-treated groups at individual times was conducted using the Newman-Keuls test. , p < 0.05 statistically different from control values; , p < 0.05 statistically different from 3, 5, and 10 days treatments.

Subcellular Localization of µ-Opioid Receptors. Saline-treated control (C) and morphine-treated (M5 and M10) brain homogenates were simultaneously assessed in every experiment. Homogenates were subjected to subcellular frac-tionation yielding enriched SPMs and a light vesicle or MI fraction. Agonist-induced changes in ligand-binding parameters of µ-opioid receptors were determined by [³H]DAMGO binding (Fig. 2). Receptor affinity (K_D) did not change due to chronic morphine treatment in either SPM or MI fractions (data not shown). The densities of µ-binding sites were 203 ± 26 and 199 ± 15 fmol/mg of protein in opioid naive SPM and MI fractions, respectively. Upon 5-day morphine treatment (M5), µ-opioid receptor B_MAX values did not change in SPM. In contrast, a 68% increase in B_MAX values of [³H]DAMGO binding sites was measured in 5-day morphine-treated MI over controls. Chronic morphine treatment for 10 days (M10) resulted in elevated levels of densities of µ-binding sites; the resulting B_MAX values of [³H]DAMGO binding were 320 ± 71 fmol/mg and 263 ± 3 fmol/mg in SPM and MI, respectively (Fig. 2).

View larger version (42K): [in this window] [in a new window]   Fig. 2.   Changes in the BMAX of [3H]DAMGO binding due to opioid exposure. Animals were treated with saline (C) or increasing doses of morphine for 5 (M5) or 10 (M10) days. Subcellular fractionations of brain homogenates were performed to prepare synaptic plasma membrane (SPM) and microsomal (MI) membranes as described under Experimental Procedures. Results shown are expressed as femtomoles per milligram of protein in each fraction. Mean ± S.E.M.; n = 3 to 8; significance was determined by t test; *, p < 0.05;**, p < 0.01.

Functional Assays of Opioid Receptors. Functional coupling of µ-opioid receptors to G proteins after 5-day morphine treatment was examined by measuring the ability of DAMGO to stimulate [³⁵S]GTPS binding of membrane fractions. This µ-agonist resulted in a concentration-dependent stimulation over basal values of [³⁵S]GTPS binding in control SPM and to a lesser extent in control MI fractions (Fig. 3). Upon chronic administration of morphine, there was a trend toward lower DAMGO efficacy in SPM that, however, is not statistically significant (Fig. 3A). An opposite phenomenon, namely a leftward shift of the dose-response curve of DAMGO indicating increased coupling to G proteins, was noted in the MI fraction of morphine-tolerant animals (Fig. 3B). Accordingly, the EC₅₀ value of DAMGO significantly decreased (p < 0.05) from 73 ± 3 to 11 ± 1 nM in the MI fraction. In parallel, the extent of maximal stimulation by DAMGO significantly (p < 0.05) increased from 49 ± 1 to 67 ± 3% in MI upon chronic morphine exposure.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Stimulation of [35S]GTPS binding by varying concentrations of DAMGO in control (open symbols) and 5-day morphine-treated (filled symbols) rat brain SPM and MI fractions. Data are mean ± S.E.M. of six experiments each performed in triplicate.

Pertussis Toxin Labeling of G Proteins. PTX, which catalyzed ADP ribosylation of the -subunits of G_i/G_o proteins was used first to assess the G protein distribution in subcellular fractions of rat brain. The proteins incubated with [³²P]NAD were analyzed by SDS-PAGE and autoradiography. PTX catalyzed the labeling of three proteins in the molecular mass range of 39 to 42 kDa that were present in all fractions (Fig. 4). The most intense band at 42 kDa represented about 70 to 80% of the total labeling, the other two bands were faint. The density of the main band increased by about 20 and 90% in the SPM and MI, respectively, compared with their saline-treated counterparts after 5-day chronic morphine treatment (Fig. 4).

View larger version (31K): [in this window] [in a new window]   Fig. 4.   Pertussis toxin-dependent labeling of G proteins in rat brain subcellular fractions. Shown are the percentage of density changes in the 5-day morphine-treated rat brain SPM and MI fractions over their saline-treated (control) counterpart. Mean ± S.E.M.; n = 3; significance was determined by t test; *, p < 0.05. Inset, autoradiogram of one representative experiment. Lane 1, morphine-treated MI; lane 2, saline-treated MI; lane 3, morphine-treated SPM; and lane 4, saline-treated SPM. The position of the molecular weight marker is shown on the right side.

Immunoblotting of G Proteins. In preliminary experiments, specificity of the G protein antibodies was assessed in opioid naive subcellular fractions. All three antibodies displayed similar labeling patterns in SPM and MI fractions (Fig. 5). The intensity of labeling was proportional to the amount of membrane proteins loaded on the gel (7.5, 15, or 30 µg) for all antibodies tested. Antisera anti-G_o labeled a faint band at 41 and a strong one at 39 kDa, as expected (Goldsmith et al., 1987; Morris et al., 1990). Labeling of the 41-kDa protein (presumably G_o1) was not detectable at lower protein amounts, thus only the 39-kDa band (presumably G_o2) was further assessed. Anti-G_i1,2 antisera labeled two bands at 40 and 41 kDa, which correspond to G_i2 and G_i1, respectively (Fábián et al., 1998). The anti-G_s antisera identified two poorly separated G protein -subunits at 45 to 46 kDa. Specificity of the antisera and identity of the labeled proteins was further supported by the similar mobility of immunolabeled and PTX-labeled bands (Figs. 4 and 5), of photoaffinity labeled bands (not shown), and by the identification of G protein subtypes with a set of antibodies, including those used here, in cultured cerebral endothelial cells (Fábián et al., 1998).

View larger version (48K): [in this window] [in a new window]   Fig. 5.   Western blotting of various G protein -subunits by subtype-specific antibodies. Varying amounts of sample proteins (7.5, 15, and 30 µg) of opioid naive rat brain SPM (A) and MI (B) fractions were run on SDS-PAGE, blotted onto nitrocellulose, and incubated with the appropriate antibody GC/2 (anti-Go; 1:500), AS/7 (anti-Gi1,2; 1:500), and RM/1 (anti-Gs; 1:500). Immunolabeled proteins were visualized by ECL chemiluminescent detection system on Kodak X-OMAT AR film. The position of the molecular weight marker is shown on both sides.

View larger version (28K): [in this window] [in a new window]   Fig. 6.   Chronic morphine-induced changes in the level of G protein -subunits (, Gs; , Gi1; , Gi2; , Go). Samples from saline- and chronic morphine-treated rat brain synaptic SPM and MI fractions were run on SDS-PAGE as in Fig. 5. Films were scanned, and data were evaluated by ImageQuant software (Molecular Dynamics). Results are presented as percent densities of saline-treated control values in appropriate fractions, which were defined as 100%. A, 3-day morphine (M3); B, 5-day morphine (M5); C, 10-day morphine treatment (M10). Mean ± S.E.M.; n = 3; significance was determined by t test; , p < 0.05; , p < 0.01.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

Here, we describe chronic morphine-induced changes in levels of µ-opioid receptors and G proteins in subcellular fractions of rat brains. New findings include a significant up-regulation of µ-sites and G_i/G_o proteins in the light membrane fraction of morphine-tolerant/dependent animals. These effects have gone undetected in the past with methods that yield crude membrane fractions. Comparable B_MAX values were measured in control SPM and MI membranes (Fig. 2) that are consistent with earlier reports (Moudy et al., 1985). µ-Sites were up-regulated by 68% in MI upon 5 days of morphine exposure. When morphine treatment was performed with a higher dose for 10 days, the density of both surface and intracellular µ-sites was elevated, consistent with previous findings (Brady et al., 1989; Rothman et al., 1991).

The leftward shift of DAMGO-stimulation of [³⁵S]GTPS binding in MI suggested increased coupling of intracellular µ-sites to G proteins after 5 days of morphine treatment compared with control MI (Fig. 3B). The increased efficacy, i.e., maximal stimulation by DAMGO in MI after chronic morphine, may reflect an increased number of µ-binding sites and/or G proteins, which is supported by the results of ligand binding experiments and G protein measurements (Figs. 2 and 5). In contrast, µ-opioid receptors in SPM did not display such changes. Neither the maximal effect nor the EC₅₀ of DAMGO in stimulating [³⁵S]GTPS binding changed significantly in SPM upon 5 days of morphine-treatment (Fig. 3A). These results are reminiscent of data on the same parameters by autoradiography of rat brain regions (Sim et al., 1996) and whole brain membranes (Heyliger et al., 2000). In contrast, chronic morphine-induced increases in G protein coupling in MI proved significantly different from that observed by these two groups and represent a novel finding.

However these results were somewhat unexpected on the basis of the model of tolerance invoking receptor desensitizaton/internalization. The possibility that morphine or its metabolite behaved like a partial antagonist cannot be ruled out a priori. Sternini et al. (1996) showed that morphine partially inhibited the etorphine-induced µ-opioid receptor rapid endocytosis in neurons. It is well documented that chronic administration of opioid receptor antagonists produce up-regulation of surface µ-sites and their increased G protein coupling (Moudy et al., 1985; Belcheva et al., 1991). However, it should be noted that in our experiments, only the intracellular sites were affected by shorter morphine exposure but not those in the SPM (Figs. 2 and 3).

We suggest that multiple cellular adaptations are elicited by chronic exposure to opioids in vivo. To explain the chronic morphine-induced shift in the subcellular distribution of µ-sites, we propose that it may induce the activation of cryptic or latent receptors and/or may enhance new receptor synthesis. Previously, we showed that newly synthesized µ-opioid receptors that are highly enriched in MI of neonatal brains display enhanced coupling to G proteins compared with their adult counterparts (Szücs and Coscia, 1992). These data were also consistent with autoradiographic studies showing that, although newly synthesized receptors in transit from the soma toward the nerve terminal are GTP-sensitive, internalized opioid receptors undergoing axoplasmic flow are GTP-insensitive (Zarbin et al., 1990). New receptor synthesis is not responsible for all receptor up-regulation. Although sodium butyrate-elicited up-regulation of opioid receptors is abolished by cycloheximide in cultured cells, opioid antagonist-induced up-regulation of opioid receptors is only partially blocked by this protein synthesis inhibitor (Tempel et al., 1986; Belcheva et al., 1991). Although new protein synthesis is not necessarily preceded by increases in new message, there have been no reports of alterations of mRNA levels of brain µ-opioid receptors after either chronic morphine or naltrexone treatments in rats (Brodsky et al., 1995; Buzas et al., 1996). Up-regulation of intracellular nicotinic acetylcholine receptors was reported to occur after chronic nicotine (an agonist) treatment in primary cultures of fetal rat brain. Nicotine-stimulated conversion of the low-affinity reserve pool of receptor into a high-affinity conformer was entailed (Bencherif et al., 1995). Emerging data in the literature indicate that receptor ligands can act as pharmacological chaperones (for review, see Morello et al., 2000). It has been shown that early events in the folding of the human -opioid receptor are probably rate-limiting and that receptor-folding intermediates are retained in the endoplasmic reticulum until they can adopt the correct conformation (Petäjä-Repo et al., 2000). Thus, one can speculate that the morphine-induced increase in the number of µ-binding sites we detected in the MI fraction may be due to such chaperoning effect of the ligand.

Taken together, our data represent a departure from the classical hypothesis of opioid tolerance being the result of receptor desensitization and/or internalization, and are consistent with new concepts of the mechanism of tolerance (for review, see Kieffer and Evans, 2002). A recent work has demonstrated that receptor internalization could reduce morphine tolerance in vivo (He et al., 2002). The first ultrastructural evidence confirmed that µ-opioid receptors are internalized by agonists such as etorphine, but not the partial agonist morphine, in the locus coeruleus (Van Bockstaela and Commons, 2001)

Growing evidence suggests that a major contributor to opioid tolerance is the emergence of new opioid-coupled signaling strategies. Increased phosphorylation and altered association of signaling proteins have been demonstrated in the myenteric plexus after chronic morphine (Chakrabarti et al., 2001). A switch from predominantly inhibitory to stimulatory opioid signaling was proposed in opioid tolerance (Gintzler and Chakrabarti, 2000). Thus, chronic morphine exposure was associated with both the predominance of sufentanil excitatory effects and with increased coupling of µ-opioid receptors to G_i (Wang and Gintzler, 1997). The observed increased coupling of MI µ-sites to G proteins after chronic morphine treatment in our experiments is reminiscent of that data. Increased efficacy, rather than tolerance, of opioid agonists at µ-receptors on GABAergic nerve terminals in periaqueductal gray area was demonstrated (Ingram et al., 1998). Chronic morphine induced a sustained increase in extracellular signal-regulated kinase phosphorylation and activity, which may initiate increases in tyrosine hydroxylase expression and other adaptations in the ventral tegmental area (Berhow et al., 1996). We showed that the adaptor protein dynamin translocated from the intracellular pool to SPM upon chronic morphine (Noble et al., 2000). This recruitment could be critically involved in long-lasting changes such as alterations of axonal transport observed in opioid dependence.

Sustained agonist treatment of cells can elicit alterations in both cellular distribution and levels of G proteins activated by their cognate G protein-coupled receptor (Milligan, 1993; Nestler and Aghajanian, 1997). Up-regulation of G_i/G_o proteins was noted by PTX-dependent ADP ribosylation and by immunoblotting in MI of morphine-tolerant animals (Fig. 4 and 6). A transient decrease of G_i1 in SPM with an apparent translocation of the same protein in MI also occurred after the 3-day morphine treatment. Nevertheless, putative translocation cannot be the single source of G protein increase in MI since in M5 membranes an increase was detected for G_i1, G_i2, and G_o without a concomitant decrease of the same proteins in SPM. Also, after the 10-day morphine treatment, the increase in the MI was greater than the decrease in SPM. Reduced amounts of G_o were detected in M10 SPM, whereas elevated levels of the same proteins in MI were also noted. The stimulatory G protein, G_s, was shown to play a specific role in morphine tolerance in the peripheral nervous system (Wang and Gintzler, 1997). Here we found that the total amount (surface + intracellular) of G_s tended to be less in brains of 3- and 5-day morphine-treated animals than that in saline-treated brains but unchanged in M10 membranes (Fig. 6).

The characteristic pattern of changes of G protein subtypes that were detected after 3, 5, and 10 days of morphine administration (Fig. 6) may represent different stages of cellular adaptation to the continuous presence of drug and may reflect different roles of the G protein subtypes in this progression. Our data fit into the scheme of drug regulation of neuronal gene expression suggested by Nestler and Aghajanian (1997), where one main group of genes targeted by the drug effect is the group encoding G proteins. Altered gene expression of several components of the cell signaling system, resulting in tolerance and addiction, may be part of the adaptation processes to compensate for the impact of agonist exposure (Cox, 1993).

To our knowledge, this is the first report of molecular changes in opioid receptors and G proteins elicited by chronic morphine at subcellular levels in rat brain. This approach provided detailed analysis of intracellular changes in addition to those on the cell surface and provided heretofore unrecognized findings that are consistent with a mechanism of adaptation entailing activation of latent ("cryptic") receptors and some new synthesis. It is suggested that, whereas surface opioid receptors and their cognate G proteins mediate the acute effects of morphine, intracellular events may play a crucial role in the long-term changes elicited by chronic drug exposure.