Multiple mu opioid receptors were proposed in many years ago, mainly based upon pharmacological studies [3, 43–45]. However, to date only a single mu opioid receptor gene has been identified, raising the questions of how a single OPRM1 gene could explain the complex pharmacology of mu opioids in animals and humans. Our early antisense mapping studies suggested different exon combinations for the analgesic actions of the two mu agonists morphine and M6G, raising the possibility of alternative splicing in the OPRM1 gene [17, 46]. Since then, much effort has been devoted to identifying these OPRM1variants. To date, over 28 splice variants of the mouse OPRM1 gene have been isolated [22–24, 27, 35], some of which had been previously identified in humans  and rats .
The majority of variants were C-terminal variants, differing only at the C-terminal tip. These variants revealed marked differences in their regional distribution at both mRNA and protein level [22–24, 27, 47–52] and agonist-induced G protein activation and internalization [25, 36, 37]. We then identified a second set of variants associated with exon 11, a previously unknown exon located 30 kb upstream of exon 1 , and established their functional significance in an exon 11 KO mouse model . Disrupting exon 11 diminished M6G and heroin analgesia without affecting morphine or methadone actions, suggesting that exon 11 and its associated variants played an important role in the actions of a subset of mu opioids that include M6G and heroin. A number of C-terminal variants have been isolated from the rat [20, 25] and human OPRM1 genes [21, 26, 37]. We recently isolated a homolog exon 11 and three its associated variants in the human OPRM1 gene . The current studies have now extended a similar splicing pattern to the rat with the identification of a homologous exon 11 and seven associated variants in the rat OPRM1 gene. Additionally, the exon 11 sequence has been predicted from the OPRM1 genomic locus of six other mammalian species through NBCI and Ensembl databases, including chimpanzees, monkeys, guinea pigs, bats, cows and armadillos , but not in lower vertebrate species such as fish and amphibians that contain OPRM1 gene.
The nucleotide sequence and genomic location of the rat exon 11 were similar to those in the mouse and human. However, some differences exist between the rat and mouse exons 11. Whereas the rat contains an alternative splice site within exon 11 which splits it into exon 11a and exon 11b, a situation similar to the human exon 11, the mouse does not. Alternative usage of these two splice sites within exon 11, together with a choice of downstream exons, created a number of different variants. The rat exon 11b has a predicted stop codon when translated from its first AUG in exon 11a, leading to only seven amino acids in rMOR-1G1, rMOR-1H1 and rMOR-1i1. However, initiating translation from the first AUG of exon 2 in rMOR-1G1 predicts a 6 TM protein. Using the AUG in exon 1a of rMOR-1H1 and rMOR-1i1, rMOR-1i2, rMOR-1i3 leads to the same protein as the original rMOR-1.
The variants that skip exon 11b can translate through from the AUG in exon 11a, but differed in amino acid sequence depending upon their downstream exons. In rMOR-1i2 and rMOR-1i3, translation using the first AUG in exon 11a still predicted small proteins due to early termination of translation in exons 1b and exon 1c, respectively. However, both rMOR-1i2 and rMOR-1i3 can initiate translation from the AUG of exon 1a to generate the same protein as the original rMOR-1. Thus, together with rMOR-1H1 and rMOR-1i1, a total of four exon 11-associated transcripts can produce the identical rMOR-1 protein, a similar situation seen in the mouse exon 11-aasociated variants. This raises questions regarding why four different splice variants are needed to generate the same protein. It is interesting to speculate, that these differences may differentially regulate their cellular location and their ability to express the protein, but there is no evidence to date to support this possibility.
On the other hand, translation from the AUG of exon 11a also generated the 6 TM protein, rMOR-1G2. Skipping exon 11b maintained the reading frame from AUG of exon 11a through exons 2/3/4. Similarly, skipping exon 11b also enabled rMOR-1H2 to read through, yielding a novel receptor with extra 50 amino acids extended at the N-terminus of rMOR-1, a prediction that was supported by in vitro transcription coupled with translation. Thus, rMOR-1H2 is the first full length (i.e. 7 TM) rat splice variant isolated with a different protein sequence at the N-terminus.
The exon 11 mRNAs are relatively abundant in the brain, as illustrated by Northern blot analysis that displayed a major ~ 12 kb band with intensity comparable to that observed with the exons 2/3 probe. The expression of the exon 11-associated variant mRNAs differed markedly among brain regions, contrasting with the relatively homogenous expression levels of rMOR-1. This suggested that, like the mouse, there is region- and/or cell-specific RNA processing of the variant pre-mRNAs and/or varying levels of upstream promoter activity. Differential expression of the variant mRNAs among brain regions also raised questions regarding their functions. Recently, we observed high correlations between mRNA expression levels, including exon 11-associated variants, in selected brain regions with the degree of morphine and heroin dependence and tolerance among four inbred strains of mice (J Xu, B Kest and YX Pan, unpublished observations). These results suggest a possible contribution of alternative splicing of the OPRM1 gene in mu opioid tolerance and addiction in mice, although the relevance of these correlations needs to be further validated. It will be interesting to see if these correlations also exist in rat.
The genomic location of the rat exon 11 approximately 21 kb upstream of exon 1 suggested the existence of an upstream promoter controlling the expression of the exon 11-associated variants. Preliminary studies indicate that the 5' flanking region of the rat exon 11 has promoter activity, particularly in the neuroblastoma cell lines NIE115 and Be(2)C cells, assessed using a secreted alkaline phosphotase (SEAP) reporter assay (J Xu and XY Pan, unpublished observation).
rMOR-1H2 encoded a full length 7 TM mu opioid receptor with a unique, extended N-terminus. Its similar binding profile is consistent with the other variants was expected since it is believed that the binding pocket is contained within the transmembrane regions, which are identical among all the full length variants. However, the additional N-terminal 50 amino acids in rMOR-1H2 did influence agonist-induced G protein activation, a similar scenario seen in the human N-terminal variant, hMOR-1i . While similar results were observed with the C-terminal variants, but this was more easily understood because of the presumed ability of the C-terminus to influence coupling to transduction proteins. How the additional N-terminal sequence influences receptor activation is as yet unknown.
In the mouse, the exon 11-associated variants mMOR-1G, mMOR-1M and mMOR-1N predict 6 TM variant due to skipping of exon 1, which encodes the first TM. rMOR-1G2 predicted a similar 6 TM protein with translation of exon 11 that resembles mMOR-1G with exon 4 as the last coding exon. While rMOR-1G1 also predicts a 6 TM variant with a terminal exon 4, it requires using the AUG within exon 2 to initiate translation. Despite our efforts, we were unable to isolate rat homologs of the mouse mMOR-1M and mMOR-1N. While it is possible that they do not exist in rats, it also is possible that these homologs are localized to very specific brain regions with a low overall abundance. The functional relevance of the 6 TM mouse variants has been suggested by a range of studies. First, although they do not bind radiolabeled mu agonists with high affinity, the mouse 6 TM variants displayed a moderate binding affinity towards [3H]-diprenorphine (KD approximately 10 nM; J Xu, GW Pasternak and YX Pan, unpublished observation). Second, the 6 TM variants can physically associate with the regular 7 TM MOR-1 and modulate the expression of the 7 TM receptors on cell surface membrane (J Xu, GW Pasternak and YX Pan, unpublished observation). More importantly, disrupting exon 11 diminished M6G and heroin analgesia without affecting morphine and methadone, suggesting selective roles of the 6 TM exon 11-associated variants in the actions of M6G and heroin. Finally, the conservation of the exon 11 and exon 11-associated variants across species further supports their role.