The deletion of 14 amino acids in the seventh transmembrane domain of a naturally occurring calcitonin receptor isoform alters ligand binding and selectively abolishes coupling to phospholipase C.

The cDNA that encodes the rabbit calcitonin receptor was cloned by screening a rabbit osteoclast library. Reverse transcription-polymerase chain reaction amplification of calcitonin receptor sequences from rabbit osteoclast RNA yielded cDNAs that encode two isoforms of the calcitonin receptor. One isoform is homologous to the C1a isoform previously identified in multiple cell types and species, while the second, designated CTRΔe13, is a previously unidentified isoform that is apparently generated by alternative splicing during mRNA processing that deletes exon 13, resulting in the absence of 14 amino acids in the predicted seventh transmembrane domain. Expression of mRNA transcripts encoding the two isoforms varies in a tissue-specific manner, with CTRΔe13 accounting for less than 15% of the total calcitonin receptor mRNA in osteoclasts, kidney, and brain, but comprising at least 50% of the transcripts in skeletal muscle and lung. The two isoforms were expressed, and the ligand binding and signal transduction properties were characterized. Deletion of the residues in the seventh transmembrane domain in CTRΔe13 reduced the binding affinity for salmon and human calcitonin by more than 10-fold and approximately 2-fold, respectively, resulting in a receptor that failed to discriminate between the two forms of calcitonin. Both isoforms activated adenylyl cyclase, with EC50 values consistent with the difference in ligand affinities. In contrast, only the C1a isoform, but not the CTRΔe13 isoform, activated phospholipase C. Thus, while the CTRΔe13 remains active despite the deletion of a significant portion of its seventh transmembrane domain, it has significantly altered ligand recognition and signal transduction properties.

acts to reduce serum calcium levels by inhibiting bone resorption and promoting renal calcium excretion. Besides this hypocalcemic effect, CT modulates the renal transport of water and several ions other than calcium, and acts on the central nervous system to induce analgesia, anorexia, and gastric secretion (1). The CT receptor (CTR) belongs to a family of G protein-coupled peptide receptors that includes receptors for parathyroid hormone/parathyroid hormone-related peptide, secretin, vasoactive intestinal peptide, growth hormone-releasing hormone, glucagon, glucagon-like peptide, pituitary adenylyl cyclase-activating peptide, corticotropin-releasing factor, and calcitonin gene-related peptide (2)(3)(4). The CTR, like other members of this family, is known to couple to multiple trimeric G proteins that activate several signaling molecules, thereby producing diverse biological responses in different cell types. For example, we and others have shown that some aspects of the osteoclast (OC) response to CT are mediated by the cAMP pathway, while others are mediated by protein kinase C (5,6), and studies in our laboratory, using synchronized LLC-PK1 kidney proximal tubule cells, have shown that the mode of receptor coupling and the biological effects of CT vary with the stage of the cell cycle (7,8). The coupling of a recombinant CTR to both the adenylyl cyclase and phospholipase C signaling pathways has been demonstrated (9,10). However, the mechanisms by which the coupling to individual G proteins might be differentially regulated are not defined. Cloning of the CTR gene and cDNAs has revealed the presence of several alternatively spliced cassettes, resulting in the expression of different isoforms of the receptor. Four isoforms have so far been identified. The most common isoform, C1a, was originally cloned from porcine kidney epithelial cells (11) and is also expressed in human, mouse, and rat cells (12)(13)(14)(15)(16). A second isoform, which contains a 16-amino acid cassette in the putative first intracellular loop, is expressed in both human and pig cells (12,17,18). Compared to the C1a, it has been reported to have a similar (19) or higher (20) binding affinity for salmon CT (sCT), but less potent ligand-dependent cAMP responses and no coupling to phospholipase C (19,21). A third isoform, C1b, contains a 37amino acid cassette in the putative first extracellular loop and is found in mouse brain and OC-like cells and rat brain (13,14,22). It shows lower ligand binding affinity than C1a, but no difference in the ability of the receptor to couple to signaling pathways. The fourth isoform, cloned from human breast carcinoma cells, lacks the first 47 amino acids of the Nterminal extracellular domain, but shows no change in ligand affinity to sCT or in ligand-dependent cAMP response relative to C1a (23). Taken together, these data suggest that the generation of functionally distinct isoforms of the CTR by alternative splicing of mRNA contributes to the modulation of cellular responses to CT.
To better understand the molecular basis of the function of the CTR in OCs, we have investigated the expression of CTR isoforms in isolated rabbit OCs. In addition to the widely expressed C1a isoform, we found a novel isoform, CTR⌬e13, that is generated by deletion of the cassette that corresponds to exon 13 of the porcine CTR gene (17) and encodes 14 amino acids in the putative seventh transmembrane domain. The CTR⌬e13 mRNA is present in all rabbit tissues that express the C1a isoform. Despite the deletion of more than half of the putative seventh transmembrane domain, which could have been expected to have a profound negative effect on the CTR expression or activity, characterization of the CTR⌬e13 isoform demonstrated that the protein is a functional CTR, albeit with unique differences in ligand binding and signal transduction properties relative to those of C1a.

EXPERIMENTAL PROCEDURES
Osteoclast Isolation-Osteoclasts were isolated as described previously (24), with minor modifications. Long bones were isolated from 1-week-old rabbits (body weight, 90 -120 g). After removal of muscle and cartilage, the bones were minced in minimum essential medium-␣ modification (␣-MEM) (Sigma) (pH ϭ 6.9) containing 5% fetal bovine serum, 1% penicillin-streptomycin, 26 mM sodium bicarbonate, and 10 mM HEPES. Cells were dissociated from bone fragments by gentle vortexing, then bone fragments were allowed to settle under normal gravity. The supernatant was removed and saved, and the mincing and sedimentation were repeated three more times. The supernatants were pooled and centrifuged for 5 min at 60 ϫ g. The pelleted cells obtained from the bones of one rabbit were resuspended in 40 ml of ␣-MEM and plated into four 10-cm culture dishes. After 18 h of culture, the adherent cells were washed three times with phosphate-buffered saline and then treated with phosphate-buffered saline containing 0.02% EDTA and 0.001% Pronase E (Sigma) for 10 min at 37°C to remove contaminating cells. The highly enriched osteoclasts (Ͼ90%) were washed three times in phosphate-buffered saline and cultured in ␣-MEM for 18 h prior to isolation of RNA.
Cloning of the Receptor cDNA-Standard molecular biology techniques were performed as described elsewhere (25) unless otherwise noted. RNA was isolated from purified osteoclasts using the guanidine isothiocyanate method (Stratagene, Menasha, WI). A cDNA library was generated from rabbit OC mRNA in the EXlox vector. Oligonucleotide primers were designed based on sequences that are highly conserved among CTR sequences from other species (11,12,14,15). For RT-PCR, first strand cDNA was synthesized by reverse transcription using a gene-specific primer at the C-terminal end of the coding region, and the CTR cDNA was amplified by PCR with various pairs of primers chosen to correspond to sense and antisense sequences in the N-terminal extracellular domain (nucleotides 458 -484 of human C1a cDNA; 5Ј-TATTGCAACCGCACCTGGGATGGATGG-3Ј) (primer 1), the fifth transmembrane domain (primer 2, antisense complementary to nucleotides 1187-1213 of human C1a cDNA, 5Ј-CATGACAGGTCCATGGAT-GATGTAAAG-3Ј; primer 3, sense corresponding to nucleotides 1139 -1165 of human C1a cDNA, 5Ј-GCAATCTACTTCAATGACAACT-GCTGG-3Ј), and the C-terminal tail (antisense complementary to nucleotides 1616 -1639 of human C1a cDNA; 5Ј-CCTCGGTTCCTGGTGG-CAGATGTA-3Ј) (primer 4). The rabbit OC EXlox cDNA library was screened by the plaque hybridization method with a 32 P-labeled DNA fragment probe obtained by RT-PCR. Hybridization was performed at 42°C in 40% formamide and the filters were washed in 0.25 ϫ SSC at 60°C. Isolated cDNAs were subcloned from the EXlox phage vector into the EXlox plasmid vector (Novagen, Madison, WI) for sequencing. Double-stranded template sequencing was performed using the dideoxynucleotide chain termination method (Sequenase kit, U. S. Biochemical Corp.).
DNA Constructs-The full-length of CTR⌬e13 cDNA was generated by deletion of exon 13 from C1a cDNA using the oligonucleotide-directed mutagenesis PCR method. Epitope-tagged CTR were generated as described previously (26). The hemagglutinin (HA) peptide (YPYD-VPDYA) was introduced into the site between Thr 30 and His 31 after the deduced signal peptide cleavage site in the N-terminal coding region of CTR. The sequences of DNA constructs were verified by dideoxynucleotide sequencing and subcloned into the eukaryotic expression vector pBK-CMV (Stratagene).
RNA Analysis-For PCR amplification, first strand cDNA was gen-erated as described above. The region of CTR exon 13 was amplified using a pair of sense (nucleotides 1381-1403 of rabbit C1a; 5Ј-GGAA-GATCTATGGTTACCTCATG-3Ј) and antisense primers (nucleotides 1601-1622 of rabbit C1a; 5Ј-CTACCACCAGGAGCCCAGGAAC-3Ј). Thirty-five cycles of amplification were performed using AmpliTaq DNA polymerase (Perkin-Elmer) with cycles of 1 min at 94°C, 1 min at 60°C, 30 s at 72°C and a final extension for one cycle at 72°C for 1 min. The amplified PCR products were run on an 4% agarose gel, transferred to a nitrocellulose membrane, and probed with a 32 P-end-labeled oligonucleotide that complexed to sequence common to both isoforms (nucleotides 1486 -1508 of rabbit C1a). RNase protection assay was performed using the RPA II kit (Ambion, Austin, TX) according to the manufacturer's instructions. Total RNA (20 g from OC and 100 g from the other tissues) was mixed with the 32 P-labeled (100,000 cpm) antisense RNA probe in 20 l of hybridization buffer and incubated at 42°C overnight. The following morning, 200 l of 1:100 diluted RNase solution was added, and the mixture was incubated for 30 min at 37°C. The reaction was stopped by the addition of 300 l of RNase inactivation/precipitation mixture, and the RNA precipitated at Ϫ20°C for 30 min. Protected RNA fragments were separated on a 8% denaturing polyacrylamide gel and quantified using a PhosphorImager (Molecular Dynamics, Sunnyvale, CA) overnight.
Cell Culture and Transfections-The COS-7 cells were maintained in Dulbecco's minimum essential medium (DMEM) (Life Technologies, Inc.) containing 4500 mg/liter glucose, 5% heat-inactivated fetal bovine serum and 1% penicillin-streptomycin. The HEK-293 cells were maintained in the same medium except that 10% heat-inactivated fetal bovine serum was used. Cells grown to 50% confluence in 10-cm tissue culture dishes were transfected with 8 g of plasmid DNA and 60 g of Lipofectin reagent (Life Technologies, Inc.) in 5 ml of DMEM (27) and incubated at 37°C in a 5% CO 2 atmosphere for 6 h, after which 5 ml of DMEM containing 10% fetal bovine serum were added into transfection medium. Cells were cultured for an additional 18 h, at which time the transfection medium was replaced with regular culture medium. 24 h later, cells were replated into 24-well plates at a density of 5 ϫ 10 4 cells/well and cultured for an additional 48 h before assaying the receptor function. Stable transfectants were generated by culturing transfected HEK-293 cells with 500 g/ml G418 (Life Technologies, Inc.). Clones were isolated using cloning rings and were maintained in the presence of G418 (500 g/ml).
Ligand Binding Assay-Both saturation and competition binding assays were performed as described previously (19). Cells were rinsed with binding medium (DMEM containing 1 mg/ml bovine serum albumin) and incubated in this medium for 15 min at room temperature. In saturation binding assays, the medium was then removed and replaced with binding medium containing increasing concentrations of 125 I-labeled calcitonin (Amersham Corp.) with or without 1000-fold excess of unlabeled calcitonin (Peninsula, Belmont, CA). In competition binding assays, 20,000 -25,000 cpm/well of 125 I-labeled human calcitonin (1 nM for C1a and 3 nM for CTR⌬e13-transfected cells) were added to the wells in the presence of increasing concentrations of unlabeled ligands. The cells were incubated for 1 h at room temperature, washed three times with phosphate-buffered saline, and solubilized with 0.5 N NaOH. The samples were collected from each well and counted in a gamma counter.
cAMP Measurements-cAMP was measured as described elsewhere (10). Cells were incubated with varying concentrations of ligand in DMEM containing 10 mM isobutylmethylxanthine (Sigma) for 10 min. The reaction was stopped by the addition of 95% ethanol containing 3 mM HCl. The cAMP was quantitated using a scintillation proximity assay (Amersham Corp.) following the manufacturer's instructions.
Inositol Phosphate Measurements-Inositol phosphates were measured as described previously (10). 24 h before assay, the culture medium was replaced by the medium containing 2 Ci/ml myo-[ 3 H]inositol (Amersham Corp.) and 5% dialyzed serum. Labeled cells were rinsed at room temperature with the assay medium (DMEM containing 10 mM LiCl and 20 mM HEPES). After washing, the cells were incubated in assay medium containing various concentrations of salmon CT at 37°C for 30 min. The reaction was stopped by the addition of 0.75 ml of ice-cold 20 mM formic acid. Inositol phosphates were separated from myo-inositol by ion exchange chromatography using Dowex columns (Bio-Rad AG 1-X8, 100 -200 mesh, formate form). Data are expressed as the percentage of the total recovered [ 3 H]inositol that is found in the inositol phosphate (IP) fraction.

Molecular
Cloning of the CTR Isoforms in Rabbit Osteoclasts-Other laboratories have identified several isoforms of the CTR that differ in their ligand binding and signaling properties in a variety of tissues and cells from human, pig, rat, and mouse (11-14, 16, 22, 23, 28, 29). We therefore sought to identify the CTR isoforms expressed in rabbit OC. An 0.8-kb fragment was amplified from total rabbit OC RNA by RT-PCR using primers 1 and 2, and used to screen a rabbit OC cDNA library. Four cDNA clones, approximately 3.5 kb in length, were obtained. All four clones contained the complete coding region for the C1a isoform, as well as 5Ј-and 3Ј-untranslated sequences (Fig. 1). The rabbit C1a CTR is 71% identical at the nucleic acid level and 79% identical at the amino acid level to the porcine C1a. The putative translation initiation site was assigned at position 276 based on its similarity to predicted porcine, human, and rodent CTR initiation sites. In addition, it has adenosine residues at the ϩ4 and the Ϫ3 positions, consistent with the requirements for translation initiation (30). The N-terminal domain encoded by the sequence on the 3Ј side of the assigned initiation site included a hydrophobic domain flanked by polar regions, consistent with the general outline of a signal peptide. The rabbit OC CTR displays the cardinal features of the CTR family, including a large N-terminal extracellular domain with six conserved cysteine residues and three N-linked glycosylation sites (Asn-X-Ser/Thr) and the seven transmembrane domain topography shared by all G proteincoupled receptors. It also contains a motif in the third intracellular loop (R-X 11 -K-A-V-K 343 ) postulated to be involved in coupling to G s␣ (31) and a potential protein kinase C phosphorylation site in the C-terminal tail (Thr 401 ).
In parallel with screening the cDNA library, RT-PCR was used to determine if alternatively spliced CTR mRNAs were present in rabbit OCs. The expression of CTR isoforms that contain the inserts in the first intracellular loop and the first extracellular loop was examined using primers 1 and 2 which amplified the region from near the N terminus to the fifth transmembrane domain (756 bp in the human C1a sequence). A single product was observed in this reaction and sequence analysis indicated that it corresponded to the C1a isoform. None of the other isoforms previously reported in pig, human, and rodent were observed. However, amplification of the regions from the fifth transmembrane domain to near the C terminus (primers 3 and 4) and from near the N terminus to near the C terminus (primers 1 and 4) each produced two fragments, one with the size predicted from the C1a cDNA sequence (501 and 1182 bp, respectively) and a second slightly shorter fragment. For both amplification reactions, restriction digestion of the two PCR products with NcoI and PstI restriction enzymes generated a fragment of 382 bp and a shorter fragment of 340 bp, indicating a deletion of a DNA sequence between the NcoI and PstI sites (data not shown). Sequence analysis of these two PCR products indicated that the deletion corresponded to the 42 bp that encode 14 amino acids (Gly 384 -Glu 397 ) in the putative seventh transmembrane domain. This 42-bp cassette corresponds exactly to exon 13 of the porcine CTR gene (17), suggesting that this novel isoform of CTR, "CTR⌬e13," is generated by alternative splicing of exon 13.
Tissue Distribution of the C1a and CTR⌬e13 Isoforms-The tissue distribution of the C1a and CTR⌬e13 transcripts was investigated by amplification of RNA from several tissues, using a pair of primers flanking the region of exon 13 that produce a 242-bp product for C1a and a 200-bp product for CTR⌬e13. Fig. 2A shows that both the C1a and CTR⌬e13 were expressed in OC, kidney, brain, cerebellum, heart, and muscle. Only the CTR⌬e13 was weakly detectable in lung, and neither isoform was detectable in liver.
The relative expression of C1a and CTR⌬e13 was then quantified by RNase protection assay using a radiolabeled antisense RNA probe that spans the region of exon 13. The predicted 253-base protected fragment for C1a and the predicted 170base protected fragment for CTR⌬e13 were both detected in all tissues except liver, lung, and heart (Fig. 2B). The relative levels of the mRNAs for the two isoforms, i.e. CTR⌬e13/C1a, were 0.14 Ϯ 0.04, 0.12 Ϯ 0.03, 0.09 Ϯ 0.02, 0.10 Ϯ 0.02, and 1.93 Ϯ 0.14 in OC, kidney, brain, cerebellum, and muscle, respectively (n ϭ 2). Only the CTR⌬e13 transcript was detectable in lung, and neither transcript was detectable in liver and heart. These data are consistent with the results of targeted PCR, again demonstrating that the CTR⌬e13 was expressed in various tissues of rabbit, and indicating that CTR⌬e13 represented a much higher proportion of total CTR transcripts in lung and muscle than in the other tissues.

Ligand Binding Characteristics of the C1a and CTR⌬e13
Isoforms-The ligand binding properties of the two isoforms were determined using a transient expression system. COS-7 cells transfected with cDNA constructs that encoded C1a or CTR⌬e13 receptors with and without N-terminal HA epitope tags were used for saturation binding assays with either 125 Ilabeled salmon or human CT as described under "Experimental Procedures." Preliminary studies demonstrated that the presence of the N-terminal HA epitope tag had no effect on the binding characteristics of either isoform, and the epitopetagged receptors were used for subsequent studies. Saturable binding of 125 I-sCT (Fig. 3A) and 125 I-hCT (Fig. 3B) was observed with both the C1a-and CTR⌬e13-transfected cells. No specific binding was detected in cells transfected with the empty pBK-CMV vector. Scatchard analysis of the binding data (Table I) showed that the affinity of the C1a isoform for sCT was more than 10-fold greater than the affinity of the same isoform for hCT. Deletion of the exon 13 cassette reduced the binding affinity of the CTR for both sCT and hCT, but with a substantially greater effect on sCT binding, resulting in a receptor that no longer distinguished between sCT and hCT.
Similar results were obtained in competitive binding assays analyzing the displacement of 125 I-hCT by unlabeled sCT, hCT, or calcitonin gene-related peptide (CGRP) (Fig. 4 and Table I). sCT displaced 125 I-hCT from the C1a isoform nearly 10 times more efficiently that hCT (K i ϭ 0.21 nM and 1.82 nM, respec- tively), but was only slightly more potent than hCT in displacing the labeled ligand from the CTR⌬e13 isoform (K i ϭ 1.23 nM and 1.65 nM for sCT and hCT, respectively). CGRP did not displace 125 I-hCT from either of the isoforms at concentrations up to 10 M.
Coupling of the C1a and CTR⌬e13 Isoforms to Adenylyl Cyclase-CT induces increased levels of intracellular cAMP, pre-sumably by coupling to G s , which activates adenylyl cyclase (2). The CT-dependent accumulation of cAMP was measured in both transiently transfected COS-7 cells and HEK-293 cells lines expressing the HA-tagged C1a and CTR⌬e13 isoforms, where the ligand binding characteristics of the two isoforms were similar (data not shown). The relative potencies of CT for inducing adenylyl cyclase activity were consistent with the   relative affinities obtained from binding data although the EC 50 values were lower than the K d and K i values, suggesting that full receptor occupancy is not necessary to obtain a maximal cAMP response. As shown in Fig. 5A and Table II, only a 2-fold difference between the relative potencies of salmon and human CT was observed in CTR⌬e13-transfected cells. In contrast, sCT was approximately 10-fold more potent than hCT in C1a-transfected cells. sCT failed to elevate cAMP levels in cells transfected with the vector alone.
Characterization of CT-induced Inositol Phosphate Production-Previous studies have shown that, in addition to activating adenylyl cyclase, the CTR also activates phosphatidyl inositol-specific phospholipase C, generating IPs and diacylglycerol (9, 10, 32). We therefore examined the relative abilities of the C1a and CTR⌬e13 isoforms to mediate a CT-dependent increase in IP production in HEK-293 cell lines expressing HA-tagged C1a and CTR⌬e13. sCT induced a saturable and dose-dependent accumulation of IP with the EC 50 ϭ 0.48 Ϯ 0.12 nM in HEK293 cells expressing the C1a isoform (Fig. 5B). In contrast, in CTR⌬e13-transfected cells and cells transfected with the empty vector, no significant increase of IP production was observed up to 10 M CT. DISCUSSION In the present study, we have cloned and characterized the cDNA encoding the CTR from highly purified rabbit OC. Our results demonstrate that rabbit OCs express at least two CTR isoforms. One, which accounts for the majority of CTR mRNA in OC, corresponds to the C1a isoform previously described in porcine, human, mouse, and rat (11-14, 16, 22, 23, 28, 29) that lacks inserts in the first intracellular loop and the first extracellular loop. Analysis of the structural features of the rabbit C1a CTR (Fig. 1) revealed that it is highly homologous (71-87% identical at the amino acid level) to the same isoform in other species and has the conserved structural motifs of the CTR receptor family. The second, CTR⌬e13, is apparently generated by alternative splicing that results in the deletion of exon 13 and hence a 14-amino acid deletion in the C-terminal two thirds of the putative seventh transmembrane domain (Fig. 6).
Deletion of 14 amino acids from a transmembrane domain of the seven-transmembrane receptor might be expected to markedly reduce or even abolish expression or function of the resulting protein. We found, however, that the CTR⌬e13 protein is a functional receptor, although the properties of the new isoform differ in certain ways from those of the C1a isoform. Thus, while the binding of both sCT and hCT to CTR⌬e13 is somewhat weaker than to C1a (Figs. 3 and 4, and Table I), the reduction in affinity is much greater for sCT than for hCT, resulting in a receptor with the same affinity for the two ligands. Furthermore, unlike C1a, which couples to both adenylyl cyclase and phospholipase C, CTR⌬e13 couples to adenylyl cyclase with an EC 50 slightly higher than that of C1a, consistent with the different ligand affinities, but fails to couple to phospholipase C (Fig. 5).
The stronger binding affinity and hypocalcemic activity that normally characterize sCT relative to mammalian CTs (1) correlates with the strong propensity of the central region of sCT to form an amphipathic helix when placed in a relatively hydrophobic environment (33), a property that is not shared with mammalian CTs such as hCT (34). The relatively greater decrease in affinity for CTR⌬e13 that is seen with sCT and the resulting failure of CTR⌬e13 to discriminate between sCT and hCT suggest that the seventh transmembrane domain of the C1a isoform interacts either with the helix-forming residues of sCT or with other residues of CT that may be positioned adjacent to the seventh transmembrane domain only when the helix is formed. It is well established that residues in the   (35). The seventh transmembrane helix is the site of the covalent attachment of 11-cis-retinal to rhodopsin (36) and contributes to the binding of antagonists to several receptors for biogenic amines and to the binding of both neurokinin peptides and some nonpeptide antagonists to neurokinin receptors (37,38). Our results suggest that, as in the case of these other G proteincoupled receptors, the seventh transmembrane helix of the CTR also contributes to the receptor's interaction with ligands. The deletion of the amino acids encoded by exon 13 changes not only the ligand binding properties but also the signal transduction potential of the receptor. While the C1a activates cAMPdependent pathways at low ligand concentrations and diacylglycerol-and Ca 2ϩ -dependent pathways at higher ligand concentrations, CTR⌬e13 activates only the cAMP-dependent pathways. The structural determinants of CTR coupling to the adenylyl cyclase and phospholipase C signaling pathways have not been completely defined. Previous studies of CTR isoforms revealed that an inserted sequence in the first intracelllar loop of human C1a abolished stimulation of the IP signal transduction pathway while allowing stimulation of the cAMP pathway (19 -21). This is interpreted as indicating that the additional amino acids in the first intracellular loop interfere with coupling to G q but not G s (21). Our results indicate that an intact seventh transmembrane helix is required for coupling to phospholipase C, although the basis for this requirement is not clear. The deletion of 14 amino acids in the seventh transmembrane helix could draw the proximal residues of the C-terminal tail into the lipid bilayer, thereby possibly shielding them from access by trimeric G proteins. Although perhaps less likely, it is also possible that the deletion of exon 13 coding region reduces the hydrophobic character of this part of the protein sufficiently that it is no longer able to anchor the remaining sequence in the membrane, resulting in the absence of any seventh transmembrane domain and therefore an extracellular C terminus. Further experiments will be required to address this issue. In either case, the results suggest that sequences in the C-terminal domain that are proximal to the membrane may be important for coupling to phospholipase C specifically. Findlay et al. (39) have reported that truncation of the C-terminal tail of porcine CTR inhibits both the cAMP response and the increase in intracellular Ca 2ϩ . The loss of the cAMP response in this study may indicate that more distal residues in the C-terminal tail are required for coupling of the CTR to adenylyl cyclase.
The genomic structure of many members of the CTR family has been characterized (17, 18, 40 -46). The genes of the CTR family are characterized by the presence of numerous introns within the coding region, and there are already several examples of functionally distinct CTR isoforms, some with tissuespecific patterns of expression, that are generated by alternative splicing (12-14, 17, 18, 22, 23). Subtypes of other G protein-coupled receptors generated by alternative splicing that are expressed in a tissue-specific fashion and have different pharmacological properties have also been reported (47,48). The protein sequence encoded by exon 13 of the porcine CTR gene is one of the most conserved regions in the CTR family, to the point where it is used as a signature sequence in data bases. Furthermore, in all the genes of the CTR family that have been characterized to date, with the exception of the rat glucagon receptor (43), the boundaries of the small exon that corresponds to the porcine CTR exon 13 are also highly conserved, suggesting that this sequence and possibly the ability to selectively exclude it from the final gene product are under a high degree of evolutionary constraint. It might, therefore, be expected that the presence or absence of this portion of the CTR or other receptors in the CTR family would be of considerable physiological importance. To our knowledge, however, this is the only description of an isoform of any member of the CTR family that is generated by alternative splicing of the exon that is homologous to exon 13 of the CTR, other than a preliminary report of a similar alternatively spliced PTH/ PTHrP receptor mRNA identified in immortalized human renal tubular cells (49).
The CTR⌬e13 mRNA is expressed in several tissues. Analysis of the tissue distribution of the C1a and CTR⌬e13 transcripts by RT-PCR ( Fig. 2A) and quantitative RNase protection assay (Fig. 2B) revealed that the C1a isoform accounts for 85-90% of the transcripts in OC, kidney, brain, and cerebellum. In contrast, the CTR⌬e13 transcripts comprise more than half of the CTR mRNA in lung and skeletal muscle, indicating that the expression of the two CTR isoforms is regulated in a tissue-specific manner. The significantly higher levels of expression of the CTR⌬e13 mRNA in skeletal muscle, and to a lesser extent in lung, is intriguing. Although these tissues are not classically thought of as targets for CT, sCT binding and stimulation of adenylyl cyclase activity have been reported in membranes derived from muscle (50), and sCT inhibits insulinstimulated glucose incorporation into glycogen in rat soleus muscle (51). The failure of the CTR⌬e13 to activate phospholipase C suggests that the pattern of second messengers and the resulting changes in downstream responses that are evoked by CT in muscle will be different to some degree from those elicited in OC, kidney, and brain.
In summary, we have identified a new isoform of the CTR that is apparently generated by alternative splicing of mRNA. The absence of 14 amino acids in the seventh transmembrane domain results in ligand binding and signal transduction characteristics that differ from those of the more commonly found CTR isoform. The levels of CTR⌬e13 expression relative to those of the C1a isoform, particularly in muscle and lung, suggest that this isoform may be of significant physiological importance. The differences in ligand binding and signal transduction properties of the two CTR isoforms, together with their different tissue distributions, provide further evidence that the generation of different CTR isoforms by alternative splicing of the CTR mRNA may contribute to the diverse biological responses induced by CT.