Abstract
We have identified four new μ-opiod receptor (MOR)-1 exons, indicating that the gene now contains at least nine exons spanning more than 200 kilobases. Replacement of exon 4 by combinations of the new exons yields three new receptors. When expressed in Chinese hamster ovary cells, all three variants displayed high affinity for μ-opioid ligands, but κ and δ drugs were inactive. However, there were subtle, but significant, differences in the binding profiles of the three variants among themselves and from MOR-1. Immunohistochemically, the major variant, MOR-1C, displayed a regional distribution quite distinct from that of MOR-1. Region-specific processing also was seen at the mRNA level. Antisense mapping revealed that the four new exons were all involved in morphine analgesia. Together with two other variants generated from alternative splicing of exon 4, there are now six distinct MOR-1 receptors.
The μ-opioid receptor (MOR) has a special place within the opioid receptor family. It mediates the actions of morphine and most clinical analgesic agents, as well as drugs of abuse such as heroin. In recent years, a number of μ receptor subtypes have been proposed. The first suggestion of μ1 and μ2receptor subtypes came from a combination of binding and pharmacological studies based on the antagonists naloxonazine and naloxazone (Wolozin and Pasternak, 1981; Pasternak, 1993; Reisine and Pasternak, 1996). More recently, pharmacological and molecular differences between morphine and morphine-6β-glucuronide (M6G) suggested yet another μ receptor subtype (Pasternak and Standifer, 1995; Rossi et al., 1995a,b, 1996).
Soon after a μ receptor, MOR-1, was cloned (Chen et al., 1993; Wang et al., 1993), antisense approaches confirmed its involvement with morphine analgesia (Rossi et al., 1994,1995a,b). Only a single μ receptor gene, MOR-1, has been identified (Min et al., 1994; Giros et al., 1995; Liang et al., 1995), yet antisense mapping studies against morphine and M6G analgesia revealed interesting patterns, with some exons implicated in the analgesic actions of one but not the other (Rossi et al., 1995a,b, 1997). Although the two analgesic agents acted through different receptors, the sensitivity of both agents to at least six different MOR-1 antisense probes implied that both receptors were closely associated with MOR-1, raising the possibility of pharmacologically relevant MOR-1 splice variants (Pasternak and Standifer, 1995; Rossi et al., 1995a,b). Alternative splicing has been observed with a number of G protein-coupled receptors, including somatostatin 2 (Vanetti et al., 1998), dopamine D2 (Guiramand et al., 1995), prostaglandin EP3 (Namba et al., 1993), serotonin receptor subtypes 5-hydroxytryptamine4and 5-hydroxytryptamine7 (Lucas and Hen, 1995), and MOR-1 (Bare et al., 1994; Zimprich et al., 1995). In view of the strong pharmacological evidence for distinct μ receptors, we have explored alternative splicing of the MOR-1 gene. We report the identification of four new exons for the MOR-1gene that combine to yield three novel MOR-1 splice variants.
Experimental Procedures
Materials.
Male Crl:CD-1(ICR)BR mice were obtained from Charles River Laboratories (Wilmington, MA). [3H][d-Ala2,N-MePhe4,Gly-ol5]enkephalin (DAMGO) was purchased from New England Nuclear Corp. (Boston, MA). Opiates and opioid peptides were the generous gift of the Research Technology Branch of the National Institute on Drug Abuse (Rockville, MD). All other materials were obtained from the sources listed.
Rapid Amplification of cDNA 3′ Ends (RACE) and Reverse Transcription-Polymerase Chain Reaction (RT-PCR).
A Marathon cDNA Amplification Kit (Clontech, Palo Alto, CA) was used in 3′-RACE. A sense primer (sense primer A, 5′-CCCAACTTCCTCCACAATCGAA-3′), which is located at the 3′-end of the exon 3 and at position 1338–1359 of the mouse μ receptor; GenBank accession no. U26915) and an antisense primer, AP1 included in the kit, were used in PCR with a mouse brain marathon cDNA template. Multiple bands with different abundance were amplified. Each band was excised from agarose gel and amplified with the second set of nested primers: sense primer B (5′-GGGAACACCCCTCCACGGC-3′), which is at position 1394 to 1412 of the receptor, and an antisense primer, AP2. The PCR fragments were then subcloned into Bluescript plasmids and sequenced. Although most sequences were aligned with that of MOR-1, the sequence of one clone, 110222, of approximately 500 base pairs (bp) in length did not. It contained a partial sequence of the 3′-end of MOR-1exon 3 sequence, followed by a sequence that was totally different from that of MOR-1 exon 4. The new sequence predicted seven amino acids beyond exon 3 (RNEEPSS), followed by a termination codon.
To obtain full-length cDNA clones of the variant, a sense primer based on the 5′ untranslated region of MOR-1 at position 217 to 240 (5′-GGAACCCGAACACTCTTGAGTGCT-3′) and an antisense primer located at the 3′ untranslated region of the new sequence, antisense primer A (5′-CCACACTGCTCACCAGCTCATCCC-3′), were used in PCR with the first-strand cDNA reverse-transcribed from mouse brain RNA as template. Three fragments of approximately 1.3, 1.4, and 1.5 kilobases (kb) in length, respectively, were obtained; subcloned into pCRII-TOPO plasmid (InVitrogen, Carlsbad, CA); and sequenced in both directions with appropriate primers. Sequence analysis of the fragments revealed that all three clones contained coding exons 1, 2, and 3 fromMOR-1 but different sequences downstream from exon 3. The three clones were named MOR-1C, MOR-1D, andMOR-1E. MOR-1D has the same sequence seen with the 3′-RACE from clone 110222. MOR-1C contained a 89-bp insertion between exon 3 and the new sequence from clone 110222. Although MOR-1C contains the same new sequence found in MOR-1D, the 89-bp insertion results in a reading-frame shift. As a result, MOR-1C predicted 52 amino acids that did not include the amino acid sequence from MOR-1D.MOR-1E had a 209-bp insertion between exon 3 and the new sequence found in clone 110222, making it the longest novel sequence. The last 89 bp in this insertion were identical to those inMOR-1C; however, the MOR-1E sequence predicted only 15 amino acids.
Isolation and Characterization of Genomic Clones.
Genomic clones containing the MOR-1 exons and the new sequences were obtained from screening two mouse genomic bacterial artificial chromosome (BAC) libraries (GenomeSystems, Inc., St. Louis, MO, and Research Genetics, Huntsville, AL) and a mouse genomic P1 library (GenomeSystems, Inc.) by either PCR or hybridization methods. All the clones obtained were analyzed by restriction enzyme digestion, long PCR, Southern blotting, and sequencing with appropriate primers and fragments. Initially, BAC clone A, approximately 75 kb in length, was obtained from the GenomeSystems BAC library using MOR-1 exon 4 primers in PCR. BAC clone A contained the MOR-1 exons 1, 2, 3, and 4 but none of the new sequences. Because no positive clones were obtained by screening the BAC library with the new sequence in clone 110222, we screened the P1 library and obtained one P1 clone (P1 clone A) of approximately 100 kb in length, which contained the new sequence. However, it contained neither exon 4 nor the additional insertions seen in MOR-1C or MOR-E. To identify a clone containing these insertions and to fill the gap between the BAC clone A and the P1 clone A, another mouse BAC library (Research Genetics, Inc.) was screened by hybridization with the insertional sequence. Five positive clones of different lengths were identified. One of these, BAC clone B (∼120 kb) contained exon 4 and the insertions present inMOR-1C and MOR-1E. The three BAC and P1 clones overlapped each other, predicting an MOR-1 gene of approximately 230 kb, which is consistent with distance between BAC clone A and the P1 clone measured by fluorescence in situ hybridization (FISH) in interphase nuclei (see below).
Chromosomal Localization by FISH.
Chromosomal localization of the P1 clone was carried out using FISH methods by GenomeSystems, Inc. In brief, the P1 clone A was labeled with digoxigenin dUTP and hybridized to metaphase chromosomes derived from a mouse embryo fibroblast cell line. Specific hybridization signals were detected by incubating the hybridized slides in fluoresceinated antidigoxigenin antibodies, followed by counterstaining with 4′,6-diamidine-2′-phenylindole HCl. The initial experiment resulted in specific labeling of the proximal portion of a medium-sized chromosome, believed to be chromosome 10 on the basis of 4′,6-diamidine-2′-phenylindole HCl staining. Cohybridization of a specific probe for the telomeric region of chromosome 10 with the P1 clone A demonstrated that the P1 clone A was located immediately adjacent to the heterochromatic euchromatic boundary of chromosome 10, an area corresponding to band 10A2. A total of 80 metaphase cells were analyzed, with 68 exhibiting specific labeling.
Chromosomal Physical Distance Measurement by FISH.
An interphase FISH analysis (van den Engh et al., 1992) was used to estimate the physical distance between the BAC clone A and the P1 clone A by GenomeSystems, Inc. In brief, the BAC clone A and P1 clone, labeled with biotin dATP or digoxigenin dUTP, respectively, were hybridized as differentially labeled pairs to interphase nuclei derived from mouse embryo fibroblasts. Specific hybridization signals were detected by fluorescein conjugate anti-digoxigenin antibodies and Texas red avidin. The mean distance between the two clones was calculated from measurements made from photographs of interphase cells exhibiting paired red and green signals and converted to the actual distance between the two clones in kilobase pairs. The estimated distance between the BAC clone A and the P1 clone A was approximately 250 kb, with a possible error of approximately 30%.
Northern Blot Analysis.
To investigate the lengths of the transcripts encoding the new variants, Northern blot analysis was performed as described previously (Pan et al., 1994). In brief, total RNA was isolated from mouse brain by the guanidinium thiocyanate phenol-chloroform extract method. Then, 50 μg of total brain RNA/lane was loaded, separated on a 0.8% formaldehyde agarose gel, and transferred to GenePlus membrane. The membrane was hybridized with32P-labeled fragments of the new sequences generated by PCR with appropriated primers.
Expression of MOR-1C, MOR-1D, andMOR-1E.
The cDNA fragments containing the full-lengthMOR-1 or the MOR-1 variants in pCRII-TOPO were subcloned into pcDNA3.1 (InVitrogen), a mammalian expression vector. The resulting plasmids, MOR-1/pcDNA3,MOR-1C/pcDNA3, MOR-1D/pcDNA3, andMOR-1E/pcDNA3, respectively, were used to transfect Chinese hamster ovary (CHO) cells by LipofectAMINE reagent (GIBCO, Gaithersburg, MD). Stable transformants were obtained 2 weeks after selection with G418 and screened with [3H]DAMGO binding assay.
In Vitro Translation.
MOR-1/pcDNA3,MOR-1C/pcDNA3, MOR-1D/pcDNA3, andMOR-1E/pcDNA3 plasmids were transcribed and translated in vitro with a TNT-coupled reticulocyte lysate kit (Promega, Madison, WI). Briefly, the plasmids were incubated with T7 RNA polymerase and reticulocyte lysate in the presence of 0.04 mCi of [35S]methionine (>1000 Ci/mmol; DuPont-NEN, Boston, MA) at 30°C for 1 h. The translation products were separated by a 12.5% SDS-polyacrylamide gel, and the gel was treated with Amplify (Amersham Life Science), dried, and exposed to Kodak BioMax MR film.
Regional Expression of MOR-1C, MOR-1D, and MOR-1E mRNA.
Total RNA was extracted from different mouse brain regions as described and reverse-transcribed with SuperScript II Reverse Transcriptase (GIBCO) in the presence of random hexamers. For the new variants, the first-strand cDNAs were amplified with a nested PCR strategy. The first-round PCR using the sense primer A (see above), which was designed from exon 3, and an antisense primer (5′-GAAAGGCATCTTCCCTCTCGCTGT-3′), which was derived from exon 9 did not yield visible bands on agarose gel. We then used the first PCR products in the second-round PCR with a pair of nested primers (sense primer B and antisense primer A, see above). Three major bands were amplified. RNA loading was estimated by from a parallel PCR with β2-microglobulin primers (Clontech). The agarose gel was stained with ethidium bromide and photographed with Kodak DC120 Digital Camera and Imagine System. The predicted sizes of the PCR products for MOR-1C, MOR-1D, andMOR-1E are 246, 157, and 366 bp, respectively. Each band was extracted from agarose gel, subcloned into pCRII-TOPO plasmid, and sequenced. The sequences showed that they all correspond to respective variants. For MOR-1, PCR was performed using a sense primer (5′-GCATCCCAACTTCCTCCACAATCG-3′) and an antisense primer (5′-CCAGGAAACCAGAGCCTCCCACAA-3′).
Binding Assays.
Membranes were prepared from stable transfectants with the pcDNA3.1 constructs as previously described (Pan et al., 1994, 1996). [3H]DAMGO binding was performed at 25°C for 60 min in 50 mM potassium phosphate buffer, pH 7.4, containing 5 mM magnesium sulfate. Specific binding was defined as the difference between total binding and nonspecific binding, defined by levallorphan (1 μM). K D andK i values were calculated by nonlinear regression analysis (Prism; GraphPAD Software, San Diego, CA). Protein concentration were determined as described by Lowry et al. (1951) using BSA as the standard.
Immunohistochemistry.
After an injection of sodium pentobarbital (100 mg/kg i.p.), mice received an intracardiac perfusion of PBS 0.1 M (50 and 20 ml, respectively) followed by 4% formaldehyde in 0.1 M phosphate buffer (300 or 50 ml, respectively). After the perfusion, the brain and spinal cord were removed, postfixed for 4 h in the same fixative, and then cryoprotected overnight in 30% sucrose in 0.1 M PBS. Immunostaining was performed on 40-μm sections cut in the coronal plane on a freezing microtome. Immunostaining was performed according to the avidin-biotin peroxidase method of Hsu et al. (1981). Sections were incubated with a solution of 0.1 M PBS with 3% normal goat serum and 0.3% Triton-X. The blocking solution was removed from the tissue, and the sections were incubated overnight at room temperature in the primary antiserum. The sections were washed and then incubated in biotinylated goat anti-rabbit IgG and avidin-biotin-peroxidase complex (Vector Labs). To localize the horseradish peroxidase immunoreaction product, we used a nickel-intensified diaminobenzidine protocol with glucose oxidase adapted from Llewellyn-Smith and Minson (1992). Finally, the sections were washed in phosphate buffer, mounted on gelatin-coated slides, dried, and coverslipped with DPX (Aldrich, Milwaukee, WI). For immunohistochemical controls, the primary antibody was either omitted, replaced by preimmune sera, or adsorbed with several concentrations of the synthesized peptide.
Antisense Mapping.
Groups of mice (n ≥ 20) received the indicated oligodeoxynucleotide (10 μg i.c.v.) daily for 5 days, after which analgesia was assessed in the radiant heat tail-flick assay (Rossi et al., 1995a,b, 1996). In brief, the tails of mice were exposed to a light source and a baseline latency determined, which was typically between 2 and 3 s. Analgesia was defined as a doubling or greater of the baseline latency. Significance between groups was assessed using Fisher’s exact test.
Results
Cloning New Splice Variants of MOR-1 Gene and Its Structure.
Using 3′-RACE, we identified a novel sequence downstream from exon 3 in MOR-1 that replaced the sequence of exon 4, subsequently identified as exons 8 and 9. Reverse transcription (PT)-PCR with an upstream primer in the 5′-untranslated region of MOR-1 along with a downstream primer from the new sequence yielded three different full-length cDNA clones:MOR-1C, MOR-1D, and MOR-1E (Figs.1 and 2). All three contained exons 1, 2, and 3 as originally described inMOR-1, and the nucleotide sequence first identified using 3′-RACE. MOR-1C had an additional 89-bp insertion, whereas the insertion in MOR-1E was longer (209 bp).
We next established the gene structure of MOR-1 using genomic clones (Fig. 1). BAC clone A (∼75 kb) contained all four original MOR-1 exons as well as the mouse homolog of the rat exon 5, which encodes the splice variant MOR-1B (Zimprich et al., 1995) (Figs. 1 and 2). However, BAC clone A did not contain the new sequence we identified using 3′-RACE (exons 8 and 9). We then isolated P1 clone A (∼100 kb), which contained the sequence found with the original 3′-RACE (exons 8 and 9), but this clone did not overlap with BAC clone A. Furthermore, it also did not contain the additional sequences found in either MOR-1C or MOR-1E. The use of FISH localized the P1 clone to band 10A2 of chromosome 10, the same region to which MOR-1 was localized (Giros et al., 1995; Kaufman et al., 1995). Interphase FISH analysis revealed a distance of ∼250 kb between BAC clone A, which contained the original MOR-1gene, and the P1 clone. This distance was confirmed after we finally obtained BAC clone B, which overlapped with the originalMOR-1 exon 4 and the P1 clone (Fig. 1).
Using these overlapping clones, we then mapped the new sequences. The three new variants were generated from combinations of four separate exons located downstream from the original exon 4 (Fig. 1). Exon 6 was 120 bp, whereas exon 7 was only 89 bp. Exon 8 was the shortest, only 66 bp, whereas exon 9 was the longest, 388 bp. The cDNA sequences obtained in the original cloning were identical with the gene sequences. The original sequence obtained in the 3′-RACE corresponded to exons 8 and 9. The 89-base insertion in MOR-1C was exon 7, whereas the insertion in MOR-1E was exons 6 and 7. All four new exons had flanking sequences consistent with consensus splice junctions (Fig.1B). Thus, the MOR-1 gene consists of nine exons spanning at least 200 kb.
The predicted amino acid sequences for these new exons differed fromMOR-1 and from each other. Exon 4 of MOR-1 has 12 predicted amino acids. MOR-1C had the longest predicted sequence, 52 amino acids, whereas MOR-1D had only 7. Although exon 8 is translated in both MOR-1C andMOR-1D, the presence of exon 7 with its 89 bp inMOR-1C produced a reading frame shift in exon 8. The termination codon in MOR-1E, which has 15 amino acids, is in exon 6, so that exons 7, 8, and 9 are not translated. These sequences differed from MOR-1 in other respects as well. For example,MOR-1E is quite basic and contains a protein kinase C phosphorylation site not present in MOR-1. MOR-1Ccontains two putative casein kinase II phosphorylation sites.
Analysis of MOR-1 mRNA.
We next examined the expression of the mRNA encoding the variants. A probe containing exons 7, 8, and 9 hybridized to a diffuse band, ranging in size from ∼6 to 9 kb. This was easily distinguished from an exon 4 probe from the original MOR-1, which revealed a single transcript of ∼12 kb (Fig. 3). An exon 7 probe, which would detect only MOR-1C and MOR-1E, revealed a weaker band of similar size as the combined exon 7/8/9 probe (data not shown). We were unable to identify MOR-1E in this assay using an exon 6 probe.
The regional mRNA levels of the different new variants was particularly intriguing. Using RT-PCR, we examined the mRNA expression of the different variants in a number of brain regions (Fig.4). MOR-1 was expressed in all regions (Fig. 4B). Of the new variants, MOR-1C was the predominant isoform in all the regions, but the relative expression of the other variants varied widely (Fig. 4A). This was particularly evident in the thalamus, a region noted for high levels of μ-opioid receptor binding (Pert et al., 1976; Atweh and Kuhar, 1977a,b,c;Goodman and Pasternak, 1985), where there was little evidence for either MOR-1D or MOR-1E expression. The spinal cord also contained predominantly MOR-1C mRNA, with low levels of MOR-1E and no observable MOR-1D. In contrast, the periaqueductal gray and striatum displayed all three variants, with highest levels of MOR-1C, followed byMOR-1E and then MOR-1D. In contrast, the cortex had higher levels of expression of MOR-1D thanMOR-1E, as did the cerebellum and brainstem.
Characterization of Expressed Variants.
In vitro translation of the three full-length cDNA clones revealed that MOR-1Dand MOR-1E had molecular weights similar to that ofMOR-1, whereas the size of MOR-1C was larger than the others, as expected based on the predicted sequence (Fig.5). We then stably transfected CHO cells with the three clones and examined opioid binding. In saturation studies, [3H]DAMGO displayed high affinity for all the variants (Table 1). Indeed, the new variants bound [3H]DAMGO with higher affinities than MOR-1, but the differences did not achieve statistical significance (Table 1).
Competition studies established that the variants all belonged within the μ-opiod receptor family (Table 2). The μ ligands such as morphine, DAMGO, and M6G and the endomorphins all competed binding quite potently, whereas the κ1-opioid U50,488H and the δ-opiod ligand [d-Pen2,d-Pen5]enkephalin were ineffective. However, the binding selectivity profiles among the variants were significantly different. For example, morphine competed binding to the MOR-1D variant over 3-fold more potently than against MOR-1 itself (p < .05). Similarly, the opioid peptide [d-Ser2,Leu5]enkephalin-Thr was twice as potent against binding to the MOR-1D variant than MOR-1 (p < .05). However, the most dramatic differences in potency were seen with the endogenous opioids dynorphin A (p < .0001) and β-endorphin (p < .0003). The MOR-1D variant had the highest affinity for both dynorphin A and β-endorphin.MOR-1E also had a significantly higher affinity for β-endorphin than MOR-1. Dynorphin A had significantly higher affinity for MOR-1C and MOR-1D than eitherMOR-1 or MOR-1E.
Regional Expression of MOR-1C.
We explored the regional distribution of MOR-1C, the most abundant of the three variants, using a polyclonal antibody generated against a unique amino acid sequence in this variant (Abbadie et al., in press). The polyclonal antibody recognized MOR-1C but notMOR-1 in transfected cells in Western blots (data not shown)
Sections through the striatum (Fig. 6, A and B) demonstrated marked differences between MOR-1 andMOR-1C. The labeling of MOR-1 corresponded closely to regions previously reported to contain high levels of μ binding autoradiographically. MOR-1 immunolabeling was observed in patches in the striatum, as well is in the subcallosal streak. Dense areas of labeling also were seen in the nucleus accumbens. In contrast, the MOR-1C antiserum failed to label these areas. There also was MOR-1C immunoreactivity in regions of the lateral septum that had minimal staining with theMOR-1 antiserum. The hypothalamus also revealed significant differences between the two antisera (Fig. 6, C and D). Although there was some MOR-1 staining, the intensity of theMOR-1C immunoreactivity was far more intense in the arcuate nucleus and median eminence. Additional studies documented intenseMOR-1C immunoreactivity in the trigeminal tract and the dorsal horn of the spinal cord, as well as in the periaqueductal gray (data not shown). Overall, the mouse distributions were very similar to those observed in the rat (Abbadie et al., in press).
Antisense Mapping Exons 6, 7, 8, and 9.
Finally, we explored the functional significance of these new variants. Antisense mapping (Standifer et al., 1994; Pasternak and Standifer, 1995) has been extensively used to correlate opioid pharmacology with theMOR-1 receptor (Rossi et al., 1994, 1995a,b; Kolesnikov et al., 1996). We examined the activity of antisense probes designed to target each of the four new exons against both morphine and M6G analgesia (Fig. 7), two μ drugs whose actions have been distinguished using antisense approaches (Rossi et al., 1994, 1995a,b). Control studies have documented that intrathecal administration of the antisense targeting exon 8 down-regulates MOR-1C immunohistochemistry in the dorsal horn by 40 to 50% (Abbadie et al., in press). All four antisense probes significantly lowered morphine analgesia (Fig. 7). A mismatched control based on the antisense targeting exon 7 was inactive, confirming the specificity of the response. The activity of antisense probes against all four exons implied that each was present in the mRNA encoding a receptor or receptors involved with morphine analgesia. The activity of the probe against exon 6 clearly implied the involvement of MOR-1E because it is the only variant containing this exon. At this point, it is not possible to determine whether MOR-1C or MOR-1D was involved because they both share exons 7, 8, and 9 with MOR-1E. Thus, the response may involve MOR-1E alone or a combination of MOR-1Cand MOR-1D. In contrast to their significant blockade of morphine analgesia, none of the antisense probes significantly lowered M6G analgesia; thus, these exons were not a component of the postulated M6G receptor.
Discussion
Correlation of the actions of morphine and other μ opioids with cloned receptors has long been a goal in the opioid field. The cloning of the μ-opioid receptor quickly led to its association with many morphine actions using both antisense (Rossi et al., 1994; Pasternak and Standifer, 1995) and gene disruption techniques (Matthes et al., 1996; Sora et al., 1997; Loh et al., 1998; Schuller et al., 1999). Pharmacological studies have long suggested subtypes of μ receptors (Wolozin and Pasternak, 1981; Pasternak, 1993; Reisine and Pasternak, 1996), and recent studies have raised the possibility that some of these may reflect splice variants of the MOR-1 gene (Pasternak and Standifer, 1995), the only identified gene encoding a μ receptor. Two MOR-1 variants were identified shortly after the initial cloning of MOR-1 (Bare et al., 1994;Zimprich et al., 1995). Our current results identified an additional three MOR-1 splice variants that result from combinations of four new exons.
The identification of the new four exons indicated that theMOR-1 gene contained at least nine exons spanning over 200 kb. There was extensive splicing, with six variants differing only at the intracellular carboxyl terminus. In MOR-1A, exon 4 was missing, leaving an extended exon 3 that encodes four additional amino acids (Bare et al., 1994). MOR-1A was first detected in a human cell line, but a similar murine variant also has been isolated (Y.-X.P., J.X., and G.W.P., unpublished observations). MOR-1B, isolated from the rat, contains an alternatively spliced exon 5 instead of the original exon 4. Exon 5, which is located between exons 3 and 4 (Fig. 2), now has been identified in the murine gene (Y.-X.P., J.X., and G.W.P., unpublished observations).
Unlike the other variants, the new ones consisted of two to four exons in place of the original exon 4. These exons were quite short, ranging from only 66 to 388 bases, and they were widely separated from the original exons comprising MOR-1 and from each other. The amino acid sequences of all the new variants were different. Although exon 8 contained coding sequences in both MOR-1C andMOR-1D, the 89 bases from exon 7 produced a reading frame shift in exon 8 of MOR-1C compared with MOR-1D. Exon 8 was not translated in MOR-1E because the termination codon was in exon 6. The sequences encoded by MOR-1D andMOR-1E were both relatively short: 7 and 15 amino acids, respectively. MOR-1C had an extended sequence, 52 amino acids, which was far longer than MOR-1 itself.
The differences between the new variants and MOR-1 itself were restricted to the terminal portion of the intracellular tail of the receptor. The transmembrane regions of G protein-coupled receptors are critical in the binding selectivity of the receptors. Because all three new variants shared the exons encoding all seven transmembrane regions, it was not surprising that they all selectively bound μ opiods and had poor affinity for δ and κ drugs, yet the variants did have subtle binding profile differences, particularly for the endogenous opioids dynorphin A and β-endorphin. It is not yet possible to establish the reasons underlying these differences, which might reflect a generalized structural change in the receptor or possibly coupling to a different G protein. This question needs further examination. The presence of additional phosphorylation sites inMOR-1C and MOR-1E also raises questions about the regulation of their function. Finally, recent reports on the γ-aminobutyric acidB receptor found that the intracellular tail was important in forming receptor heterodimers necessary for a functional receptor (Jones et al., 1998; Kaupmann et al., 1998; White et al., 1998; Kuner et al., 1999). Opioid receptor dimerization also has been proposed (Cvejic and Devi, 1997). The potential role of these differences in the intracellular tail of the receptor in these interactions also needs to be explored.
The most dramatic differences between MOR-1 and the variants were seen in their regional expressions. Although MOR-1 andMOR-1C are derived from the same gene, their markedly different immunohistochemical distributions implicated region-specific processing. This concept was supported at the mRNA level as well. Among the variants, MOR-1C was the most abundant in all regions examined. However, the relative expression of MOR-1D andMOR-1E to MOR-1C varied from region to region. For example, the expression of MOR-1E was greater than that of MOR-1D in the hypothalamus, whereas the reverse was true in cortex. These regional differences in expression further support the possibility that these variants encode pharmacologically relevant receptors.
Anatomic studies in both mice and rats (Abbadie et al., in press) have demonstrated the presence of MOR-1Cimmunoreactivity in regions important in pain processing, including the dorsal horn of the spinal cord, the trigeminal nucleus, and the periaqueductal gray. The presence of MOR-1C receptors in this region was consistent with the antisense studies. Antisense probes targeting all four of the new exons lowered morphine analgesia. The activity of the exon 6 antisense probe clearly implicatedMOR-1E in this action because this is the only variant containing this exon. The activity of the antisense probes targeting exons 7, 8, and 9 might be due to their actions against onlyMOR-1E, which contain these exons as well. Because there are no unique antisense probes for MOR-1C and MOR-1D, their involvement in morphine analgesia remains unclear. Pharmacological studies have suggested that morphine analgesia can involve more than one subtype of μ receptor (Paul and Pasternak, 1988; Pick et al., 1992). The blockade of morphine analgesia by antisense probes targeting the new variants, as well as probes based on exon 4, which is present in MOR-1 (Rossi et al., 1994,1995a,b), provide evidence at the molecular level for the involvement of at least two different μ receptor subtypes.
Our initial impetus into the search for additional splice variants was prompted, in large part, by the pharmacological differences between morphine and M6G (Rossi et al., 1995a,b, 1996, 1997). The antisense mapping results suggest that these new exons are not a component of the receptor responsible for M6G analgesia. However, the possibility remains for alternative splicing at other exons as well, leading to an even greater diversity of MOR-1 variants. It will be interesting to see whether other variants will be uncovered.
Footnotes
- Received February 3, 1999.
- Accepted April 13, 1999.
-
Send reprint requests to: Dr. Gavril W. Pasternak, Department of Neurology, Memorial Sloan-Kettering Cancer Center, 1275 York Ave., New York, NY 10021. E-mail:pasterng{at}mskmail.mskcc.org
-
This work was supported in part by the National Institute on Drug Abuse Grants DA02615, DA06241, and DA07242; Senior Scientist Award DA00220 (to G.W.P.) and Research Scientist Development Award DA00296 (to Y.-X.P.); and Core Grant CA08748 (to the Memorial Sloan-Kettering Cancer Center).
Abbreviations
- MOR
- μ-opioid receptor
- DAMGO
- [d-Ala2,N-MePhe4,Gly-ol5]enkephalin
- M6G
- morphine-6β-glucuronide
- RT
- reverse transcription
- PCR
- polymerase chain reaction
- FISH
- fluorescence in situ hybridization
- BAC
- bacterial artificial chromosome
- The American Society for Pharmacology and Experimental Therapeutics