cDNA Cloning and Functional Expression of a Ca2+-sensing Receptor with Truncated C-terminal Tail from the Mozambique Tilapia (Oreochromis mossambicus)*

The complete cDNA sequence of the tilapia extracellular Ca2+-sensing receptor (CaR) was determined. The transcript length of tilapia CaR (tCaR) is 3.4 kbp and encodes a 940-amino acid, 7-transmembrane domain protein that is consistent in its structural features with known mammalian and piscine CaRs. The tCaR extracellular domain includes a characteristic hydrophobic segment, conserved cysteine residues that are implicated in receptor dimerization (Cys129 and Cys131) and in coupling to the transmembrane domain (nine conserved cysteine residues), and conserved serine residues (Ser147 and Ser169–171) that are linked to receptor binding of Ca2+ and l-amino acid-mediated potentiation of function. mRNA expression of tCaR was strong in kidney, brain, and gill. Weaker expression was observed in pituitary, stomach, intestine, urinary bladder, and heart. This distribution is consistent with possible physiological roles in endocrine cells, excitable tissues, and ion-transporting barrier epithelia. Expression of tCaR mRNA in kidney and intestine was salinity-dependent, suggesting a role for the receptor in iono-/osmoregulation in this euryhaline teleost species. Human embryonic kidney-293 cells transiently transfected with tCaR cDNA demonstrated dose-dependent phospholipase C activation in response to elevations in the extracellular Ca2+ concentration ([Ca2+]o). Functional activation of the mitogen-activated protein kinase cascade by high [Ca2+]o was also confirmed in these cells despite the naturally occurring truncation of the receptor's intracellular tail, which removes segments variably linked in mammalian CaRs to filamin-coupled activation of mitogen-activated protein kinase cascades. Sensitivity of phospholipase C activation to [Ca2+]o was dependent on the ionic strength of the bathing medium, supporting a role in salinity sensing.

The roles for ionic calcium (Ca 2ϩ ) in physiological functions are many and varied and essential for survival. Ca 2ϩ -dependent cellular processes include hormone secretion (stimulussecretion coupling), excitability and cell motility (excitationcontraction coupling), and neurotransmission (Ca 2ϩ -based action potentials). This critical involvement of Ca 2ϩ in these and other cellular functions is unique among ions. Ca 2ϩ is simultaneously a controlled variable in physiological systems, an extracellular messenger in multicellular organisms, and an intracellular mediator in a variety of effector cells.
Unicellular organisms acquire Ca 2ϩ from the external environment; they have little control over the composition of the surrounding medium and thus had to evolve plasma membrane-bound and other cellular mechanisms to realize cellular Ca 2ϩ homeostasis eventually permitting some degree of habitat survival and selection. In multicellular organisms, active regulation by integumentary tissues of the composition of an "extracellular fluid," which bathes all cells, assures a constant internal milieu and contributes to normal Ca 2ϩ -dependent cell functions. In terrestrial vertebrates, Ca 2ϩ homeostasis depends upon appropriate dietary intake, renal output, and osseous storage and mobilization of Ca 2ϩ . These processes are well known and are regulated by parathyroid hormone, calcitonin, and 1,25-dihydroxyvitamin D acting on kidney, intestine, and bone. Activation of the Ca 2ϩ -sensing receptor (CaR) 1 by its natural ligand, extracellular Ca 2ϩ , alters parathyroid hormone and calcitonin secretion and inhibits renal 1-hydroxylase to retard the synthesis of 1,25-dihydroxyvitamin D (1,2). Ca 2ϩ balance is achieved through regulated calciotropic hormone secretion in response to the extracellular Ca 2ϩ concentration ([Ca 2ϩ ] o ); the secreted hormones, in turn, control the activity of Ca 2ϩ -transporting tissues to bring about alterations in [Ca 2ϩ ] o .
Interestingly, in fishes, the calcium homeostatic process involves ion-transporting tissues not represented in other vertebrate classes (viz., the mitochondria-rich "chloride cells" of the gills and skin) and a different panel of hormones (prolactin and somatolactin from the pituitary, and stanniocalcin from the piscine corpuscles of Stannius (3)). Opportunities for cutaneous Ca 2ϩ exchange with the environment are limited in terrestrial vertebrates, but for aquatic, water-breathing fishes, in contrast, there is substantial environmental exposure and risk to survival that is typically met by the efficient vectorial transport of Ca 2ϩ across epithelial tissues contributing to the maintenance of a stable internal fluid composition (4,5). Moreover, the Ca 2ϩ regulatory mechanism is part of a greater osmoregulatory scheme in which successful salinity adaptation involves coordinated physiological responses to external Ca 2ϩ , Na ϩ and Cl Ϫ concentrations, all of which generally co-vary. Maintenance of a stable Ca 2ϩ concentration in the body fluids of fishes is complicated further in euryhaline species that may encounter daily, seasonal, or irregular changes in the composition of the external environment.
Survival in fishes depends upon timely and appropriate physiological adaptation by epithelial tissues that separate inside from out; these include the branchial epithelium, the kidney, and the intestine. Direct effects of extracellular Ca 2ϩ on these osmoregulatory tissues are clearly indicated. Bieter (6) demonstrated many years ago that unilateral delivery of MgSO 4 solutions to the aglomerular toadfish kidney caused ipsilateral diuresis and elevated Mg 2ϩ excretion, consistent with a direct stimulatory action of Mg 2ϩ and similar to that shown for Ca 2ϩ in mammalian kidney (7,8). Mitochondria-rich cells in fish gills proliferate in response to reduced environmental [Ca 2ϩ ] (9).
The biochemical nature of the CaR was reported by Brown et al. (10), who isolated by expression cloning the full-length cDNA encoding the CaR from bovine parathyroid glands (bCaR). Subsequent studies on CaRs of other species demonstrate that CaR is a highly conserved seven-transmembrane domain receptor (1,7,10,11). CaR is a Group II member of Family C of the G proteincoupled receptor (GPCR) superfamily. Knowledge of the structure and physiological function(s) of CaR, admittedly an important and widespread molecule, is still based on its examination in only a few species representing a restricted phylogenetic perspective. Moreover, the CaR of several mammals, two teleost fishes, and a shark exhibit striking differences in the structures of their intracellular domains. Although the complete sequences of only three piscine CaRs have been reported (pufferfish (Fugu rubripes) CaR, fCaR (12); sea bream (Sparus aurata) CaR, sbCaR (13); and dogfish shark (Squalus acanthias) calcium polyvalent cation-sensing receptor, SKCaR (14)), recent studies by others offer both direct and indirect evidence for the presence and functional involvement of CaR in piscine Ca 2ϩ homeostasis and osmoregulation (14 -18).
In this study, we selected the euryhaline Mossambique tilapia (Oreochromis mossambicus) as our model system because aquatic vertebrates, in addition to the internal sensors of extracellular Ca 2ϩ , most likely possess external sensors of environmental Ca 2ϩ that are exposed to a broad, yet natural, range of ligand concentrations ([Ca 2ϩ ] in typical seawater: ϳ10 mM; in fresh water, Ͻ0.5 mM). We report here on the cDNA cloning and sequencing of the tilapia CaR (tCaR), its tissue distribution, and the signal transduction properties for the first time of a teleost CaR in a mammalian expression system.

EXPERIMENTAL PROCEDURES
Animals-Cultured Mossambique tilapia (O. mossambicus), weighing 50 -80 g, were maintained in freshwater aquaria at 25°C. Some of these tilapia were transferred to seawater aquaria, where they remained also at 25°C for at least 2 weeks. These freshwater-adapted (FW) and seawater-adapted (SW) tilapia were used for tissue harvesting as described below. The salinities and Ca 2ϩ concentrations of the fresh water and seawater were Ͻ10 mosM and 0.4 mM, and 1000 -1020 mosM and 10 mM, respectively. Fish were maintained and handled in accordance with the Guidelines for Care and Use of Animals approved by the University of Tokyo.
RNA Extraction-FW and SW tilapia were individually anesthetized in 0.1% (w/v) 3-aminobenzoic acid ethyl ester (Sigma), appropriately neutralized with sodium bicarbonate. After anesthesia, the brain, pituitary, gills (excluding the branchial arch cartilage), skin, heart, kidney, urinary bladder, liver, skeletal muscle, stomach, and intestine (that segment anterior to the easily identified sphincter marking the start of the posterior intestine) were isolated and immediately frozen in liquid nitrogen, and stored at Ϫ80°C. Total RNA was extracted from about 0.5 g of frozen tissue using ISOGEN (Nippon Gene, Toyama, Japan). RNA was quantified using measured absorbance at 260 nm (Model DU 640 spectrophotometer, Beckman Coulter, Fullerton, CA).
cDNA Cloning and Sequencing-For cloning and sequencing, poly(A)ϩ RNA was isolated from total RNA from the kidney of one FW tilapia using Oligotex-dT30 Super (Japan Synthetic Rubber, Tokyo, Japan). Double-stranded cDNA pools were prepared from 0.5 g of poly(A)ϩ RNA using a SMART TM cDNA Library Construction Kit (Clontech) with manufacturer-supplied primers. A partial cDNA fragment was amplified by PCR using a sense and antisense degenerate primer pair (omTM-1F and omCT-1R, respectively; Table I) based on the CaR cDNA sequences of cow (GenBank TM accession number S67307), human (GenBank TM accession number U20759), rat (Gen-Bank TM accession number U10354), mouse (GenBank TM accession number AF128842), pufferfish (GenBank TM accession number AB008857), and sea bream (GenBank TM accession number AJ289717). This primer pair was designed to recognize CaR sequences preferentially over structurally related pheromone receptor (PherR) sequences from the pufferfish (GenBank TM accession numbers AB008858 through AB008866). For this PCR amplification, the reaction conditions were: initial denaturation at 94°C, followed by 40 cycles of denaturation at 94°C for 50 s, annealing at 58°C for 30 s, and extension at 72°C for 1.5 min, and a final extension step at 72°C for 10 min. Each reaction mixture (25-l total volume) contained PCR buffer, 2.5 mM dNTPs, cDNA template, 20 pmol of each primer, and 0.25 unit of Ex Taq DNA polymerase (TaKaRa, Kyoto, Japan). PCR products were separated by electrophoresis on 1.2% agarose gels containing EtBr and were observed and recorded using a FAS-III Electronic UV Transilluminator (Toyobo, Osaka, Japan). Appropriate-sized bands were excised from the gels, and the PCR products were isolated using a GeneClean kit (Qbiogene, Carlsbad, CA). Following this, the PCR products were ligated and subcloned into pT7Blue T-vector (Novagen, Madison, WI). Sequencing was performed using an ABI Prism Model 310 or 3100 Genetic Analyzer (PerkinElmer Life Sciences/Hitachi Applied Biosystems, Foster City, CA). Subsequently, the remainder of the full-length tCaR cDNA was determined stepwise using the PCR, cloning, and sequencing protocol described above. The primers for each PCR step comprised both specific primers based on the determined sequence for the tilapia CaR cDNA and degenerate primers designed on the basis of other known CaRs (Table I). Complete sequencing of the tCaR cDNA required four sequencing steps from the original fragment in the 5Ј direction and a single sequencing reaction in the 3Ј direction. The final sequencing step in each direction utilized 5Ј-and 3Ј-rapid amplification of cDNA ends (RACE) with tCaR-specific primers (Table I) and the corresponding SMART TM 5Ј and 3Ј PCR primers, respectively.
Tissue Distribution-tCaR mRNA expression was initially examined by RT-PCR in the several tilapia tissues collected from two FW and two SW fish. Total RNA was prepared as described above and was treated with DNase (TaKaRa DNase I) at 37°C for 30 min to remove genomic DNA. Following DNase digestion, RNA was repurified, and its concentration was spectrophotometrically measured. Total RNA (1 g) from all eight tissues of each fish was reverse transcribed using the Superscript First Strand Synthesis System (Invitrogen) in a 20-l reaction volume according to the manufacturer's instructions. Negative control (without reverse transcriptase) PCR reactions were conducted for representative samples of all tissues to confirm the absence of contamination by genomic DNA. PCR was conducted in 25-l reaction volumes using 1 l of first strand cDNA template, 20 pmol of tCaR-specific primers omCaR-S1 and omCaR-A1, dNTPs, PCR buffer, and 0.25 unit of Ex Taq DNA polymerase, with the following program: initial denaturation at 94°C, followed by 30 or 35 cycles of denaturation at 94°C for 50 s, annealing at 60°C for 30 s, and extension at 72°C for 1.5 min, and a final extension step at 72°C for 10 min. The expression of tilapia ␤-actin mRNA was used as internal control. Specific primers for tilapia ␤-actin (om␤actin-L and om␤actin-R; Table I) were designed following confirmation of the cDNA sequence for tilapia ␤-actin (GenBank TM accession number AB037865). PCR for ␤-actin expression was conducted under the above conditions but for only 25 cycles of amplification. These amplification conditions were within the linear range of amplification (data not shown). PCR products were electrophoresed in 1.2% agarose gels containing EtBr. The gels were visualized, and the images were recorded using a Fuji FLA-2000 imaging analyzer with ImageReader and ImageGauge software (Fuji Photo Film Co., Ltd., Tokyo, Japan).
Northern Blotting-Northern blotting using 7.5 g of poly(A)ϩ RNA from the kidney of FW tilapia was performed according to standard procedures (19). A synthetic 32 P-labeled cDNA probe from the cloned tCaR (nucleotides 477-1375 in the extracellular domain (ECD)-coding region) was hybridized to membranes for 18 h. Membranes were washed under high stringency conditions (0.2 ϫ SSC with 0.1% SDS (v/v) at room temperature for 10 min and then twice for 1 h each at 60°C). Signal intensity was detected using the Fuji FLA-2000 imaging analyzer.
Functional Expression-For assessing the function of tCaR, cDNA from kidney was synthesized using the SMART TM cDNA Library Construction kit as described above, but amplification was performed by primer extension at 95°C for 2 min, 72°C for 10 min, and 95°C for 1 min, followed by 3 cycles at 95°C for 15 s and 68°C for 8 min. Using this cDNA pool as template, a 3181-kbp full-length tCaR cDNA was amplified by PCR using the specific primers omCaR-5Full and omCaR-3Full (Table I) with the following thermocycler program: denaturation at 94°C, followed by 40 cycles of denaturation at 94°C for 1 min, annealing at 60°C for 30 s, and extension at 72°C for 5 min, and a final extension step at 72°C for 10 min. Because the full-length cDNA had no appropriate cloning sites for direct transfer to the pcDNA3.1/Hygro (Invitrogen) expression vector, the PCR product was ligated first into pT7Blue T-vector and subcloned. Clones were sequenced to confirm the identity and orientation of the cDNA insert. Using restriction digestion with SalI and BamHI (all restriction enzymes from TaKaRa), the fulllength tCaR cDNA was shuttled from the pT7Blue vector into the pcDNA3.1/Hygro vector, restriction digested with BamHI and XhoI, again subcloned, and sequenced.
Signal transduction by tCaR was evaluated by measuring high [Ca 2ϩ ] o -induced activation of phospholipase C (PLC) in transiently transfected human embryonic kidney (HEK)-293 cells (20). HEK-293 cells were grown and transfected with tCaR or bCaR cDNA as previously described except for the use of transfection reagent, which was 40 l/ml Lipofectamine TM 2000 (Invitrogen) in these experiments (20). At 24 h after transfection, cells were washed, plated into 6-well culture plates, and allowed to attach and grow for 24 - Subsequently, the ability of tCaR to activate the mitogen-activated protein (MAP) kinase signaling cascade was assessed by examining phosphorylation of extracellular signal-regulated kinases 1 and 2 (ERK1/2). Cells transfected with tCaR or bCaR cDNAs for 48 h were switched to a serum-deprived basal medium (Ca 2ϩ -and Mg 2ϩ -free Dulbecco's modified Eagle's medium supplemented with bovine serum albumin (0.2% w/v), MgSO 4 (0.5 mM), and CaCl 2 (0.5 mM)) for 18 h, and the medium was renewed 2 h prior to experiments. For time-course experiments, cells were switched to Basal Medium containing various CaCl 2 concentrations (0.5, 3, and 5 mM) for 2-20 min. For Ca 2ϩ doseresponse experiments, cells were treated with different CaCl 2 concentrations (0.5-10 mM) in Basal Medium for 5 min. Whole cell lysates were prepared by treating cell monolayers with rapid immunoprecipitation assay lysis solution (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Nonidet P-40, 0.1% SDS, and 0.5% sodium deoxycholate plus Complete TM Protease Inhibitor tablets (1 tablet/50 ml; Roche Applied Science), 1 mM sodium vanadate, and 25 mM sodium fluoride). Total protein in the lysates was determined by a commercial kit (BCA Protein Assay Kit, Pierce). Cell proteins (20 -35 g) were separated by SDS-PAGE and analyzed for expression of phosphorylated and total ERK1/2 by immunoblotting as detailed below, using specific antibodies (Cell Signaling Technology, Beverly, MA) and horseradish peroxidase-conjugated goat anti-rabbit IgG antibodies (Amersham Biosciences) for detection. Signals were developed with a Signal West Dura Substrate (Pierce), detected by BioMax x-ray films (Eastman Kodak Co., Rochester, NY), and quantified by densitometry with a Microtek Scan Maker i900 scanner and Kodak 1D image analysis software (Kodak). The effects of different [Ca 2ϩ ] o on MAP kinase activation were quantified by first calculating the signal ratio of phosphorylated ERK1/2 to total ERK1/2. These ratios were then normalized to the ratio of phosphorylated ERK1/2 to total ERK1/2 in tCaR-expressing cells at 0.5 mM Ca 2ϩ .
The effects of different NaCl concentrations on PLC activation were tested in HEK-293 cells 3 days after transfections with tCaR or bCaR cDNAs. Cells loaded with [ 3 H]myoinositol were incubated with HEPESbuffered saline solution (5 mM KCl, 1 mM Na 2 HPO 4 , 1 g/liter dextrose, 25 mM HEPES (pH 7.4)) containing 0.5 mM MgSO 4 , 0.1% bovine serum albumin, 0.1 mM CaCl 2 , and 10 mM LiCl (added to block phosphatase activity) in the presence of either 50 mM NaCl and 300 mM mannitol, or 200 mM NaCl and 0 mM mannitol. Cells in the low (50 mM) and high (200 mM) NaCl treatment groups were exposed to varying [Ca 2ϩ ] o (0.1-30 mM) for 60 min at 37°C. Total [ 3 H]InsP accumulation was determined, and data were normalized and expressed as a percentage of the accumulation measured at 30 mM Ca 2ϩ .
Statistics-Experimental data on tissue mRNA, [ 3 H]InsPs, and MAP kinase activation are reported as the mean Ϯ 1 S.E. Statistical comparisons were conducted using paired or unpaired Student's t tests, as appropriate.

RESULTS
Tilapia CaR cDNA Sequence and Deduced Protein Structure-Using a combination of RT-PCR and 5Ј-and 3Ј-(RACE), a 3394-bp length CaR cDNA was isolated from the kidney of freshwater tilapia (GenBank TM accession number AY541693 ). An open reading frame marked by a Met start codon associated with a partial Kozak sequence began from nucleotide 149 of the transcript sequence. This open reading frame encoded a 940-aa protein (Fig. 1). The 3Ј-untranslated region (UTR) comprised 423 bp (excluding the stop codon) and included a single AATAAA polyadenylation signal.
An N-terminal signal peptide of 18-aa length was predicted using SignalP version 2.0 (21). From statistical analysis of the TMbase data base of naturally occurring transmembrane proteins using the TMpred program (22), the deduced tCaR protein structure includes seven predicted helical transmembrane segments (TM) that comprise a transmembrane domain (TMD) separating a 599-aa residue N-terminal ECD and a 96-aa Cterminal intracellular domain (ICD). Matching predictions of transmembrane segments were achieved using a Hidden Markov Model analysis (TMHMM 2.0 (23)).
The deduced tCaR protein is a member of Family C/Group II of the GPCR superfamily. tCaR exhibits substantial sequence similarity to both piscine and mammalian (human, hCaR; rat, rCaR; bovine) CaRs, and more modest similarity to PherRs (Fig. 2). At the nucleotide level, tCaR displays higher sequence identity to the CaRs of two teleosts (sea bream, 88%; and pufferfish, 84%) compared with those of dogfish shark (64%; GenBank TM accession number AF406649) or mammals (63-65%). The same relationship is seen in amino acid sequence identity (sea bream, 94%; pufferfish, 91%; dogfish shark, 68%; and mammals, 66 -67%). Compared with mammalian CaRs, the teleost CaRs have a C-terminal tail ICD that is abbreviated by over 100 aa residues due to both deletion of 38 residues at aa position 924 (relative to bCaR) and truncation by 90 residues at the C terminus. Restricting the comparisons to the ECD and TMD only, amino acid sequence identities between these domains in tCaR and the corresponding domains in the mammalian receptors are somewhat higher (71-73% and 87-88%, respectively). Lower nucleotide and amino acid sequence identities are observed when tCaR is compared with pufferfish pheromone receptors (48 -51% and 34 -39%, respectively), which are related members of Family C/Group II of the GPCR superfamily possessing even more truncated ICDs. Overall, this analysis indicates that "tCaR" is indeed a CaR, rather than a pheromone receptor, and reaffirms that the truncated Cterminal tail is a feature typical of teleostean (but not all piscine; cf. SKCaR) CaRs known to date.
The tCaR protein is rich in cysteine residues, possessing 18 in the ECD alone, that are conserved compared with piscine and mammalian CaRs (Figs. 1 and 2). The ECD also contains 11 predicted N-glycosylation sequons (Asn-Xaa-Ser/Thr) in the ECD, at least several of which are predicted to be highly likely sites for glycosylation (24).
Tissue Distribution of Tilapia CaR mRNA-Using RT-PCR, tCaR mRNA expression was demonstrated in several tissues of FW tilapia (Fig. 3). The strongest expression was observed in the kidney, brain, and gill, with weaker expression noted in the pituitary, stomach, intestine, urinary bladder, and heart; negligible expression was seen in liver and skeletal muscle. This expression pattern was confirmed in additional trials on tissues from FW and SW fish (Figs. 4 and 5). In some tissues, differential mRNA expression in FW and SW fish was suggested based on visual examination. Expanded analysis on selected tissues from 6 -7 each FW and SW fishes revealed that tilapia kidney and intestine exhibited opposite patterns of tCaR mRNA expression as a function of acclimation salinity (Fig. 5, A and B). Expression in kidney of FW tilapia was 3-fold greater than in the kidney of SW fish (p Ͻ 0.001). Conversely, tCaR mRNA expression in SW fish intestine exceeded that in FW fish intestine by about 3-fold (p Ͻ 0.01). Although the calculated mean for mRNA expression in gills of FW tilapia was greater than the calculated value for SW fish gills, there was no statistically significant difference (p ϭ 0.12; Fig. 5C).
Northern Blotting-Northern blot analysis of kidney poly(A)ϩ RNA under high stringency conditions (Fig. 6) revealed a strong band at 3.4 kb, closely matching in size the full-length cloned tCaR cDNA. An additional, weaker band was observed at 4.6 kb, the identity of which is currently unknown. Multiple bands of CaR transcripts on Northern analysis have also been detected in the rat and shark kidney (14,25,26). Based on the results of several independent nucleotide-sequencing reactions of tCaR cDNA, there was no indication of alternatively spliced transcripts; extended UTRs remain a possibility in explaining the second, larger transcript.
Functional Expression-Expression of tCaR and bCaR cDNA in HEK-293 cells produced dose-dependent increases in [ 3 H]InsP accumulation up to 30 mM Ca 2ϩ (Fig. 7, lower panel) Immunoblotting revealed the expected pattern for bCaR expression in HEK-293 cells (20). The strong and faint bands at 140 and 160 kDa, respectively, likely represent glycosylated variants of the receptor. Bands larger than the 214 kDa marker are thought to be receptor dimers and oligomers (Fig. 7, upper  panel). For tCaR cDNA-transfected cells, multiple bands were also visualized. The two bands with apparent molecular masses of about 95 and 98 kDa may represent the non-glycosylated receptor protein core. The next four larger bands at ϳ114, 118, 144, and 156 kDa are likely glycosylated monomers, and the two largest and strongest bands Ͼ214 kDa are probably dimers and oligomers. With the exception of some residual immunostaining of the tCaR 118-kDa band, all of the bands noted above in these blots were absent when blotting was done with antiserum that was preabsorbed with bCaR peptide.
In both tCaR-and bCaR-expressing HEK-293 cells, raising [Ca 2ϩ ] o from 0.5 to 3 and 5 mM increased expression of phos- The upper panel shows membrane protein immunoblotting results using an antiserum to a bCaR ECD peptide. Positions of immunoreactive bands are indicated at the right-hand edge of each gel image. For tCaR, immunoreactive bands were observed at 95 and 98 kDa (non-glycosylated core proteins), at ϳ114, 118, 144, and 156 kDa (glycosylated monomers), and Ͼ214 kDa (dimers and oligomers); for bCaR, bands were observed at 140 and 160 kDa (glycosylated monomers) and Ͼ214 kDa (dimers and oligomers). Preabsorption of the antiserum with bCaR peptide prevented almost all immunoreactivity. See text for details. with ED 50 values increasing 2-to 3-fold (Fig. 10). As with earlier trials conducted in 140 mM NaCl-containing medium (Fig. 7), transfected cells again displayed a biphasic dose-response relationship. DISCUSSION We cloned and sequenced a cDNA encoding an extracellular CaR from tilapia kidney. The tCaR cDNA sequence, deduced protein structure, and functional activity confirm its identity as an authentic CaR, rather than a closely related fish PherR or other 7-TM domain protein of the GPCR superfamily. Several structural features support the assignment of this molecule as a CaR (1). In the ECD, the tCaR contains a predicted signal sequence at the N terminus (2). tCaR possesses the two conserved cysteine residues, Cys 129 and Cys 131 , that are thought to be responsible for intermolecular disulfide bond formation in receptor dimerization of human CaR (27). These two cysteines are positioned in a protruding region of the bilobed venus flytrap domain model for the CaR ECD monomer, and they are suggested to function as part of the dimer interface (27)(28)(29). Preliminary three-dimensional structure prediction with the protein homology-modeling servers CPHmodels (30) and SWISS-MODEL (31), using the crystal structure of the rat metabotropic glutamate receptor subtype 1 (mGluR1)-free form I (PDB code: 1EWT) as a template, confirmed a similar orientation for these two cysteines in tCaR (3). A hydrophobic segment, comprising residues 135-174 of tCaR, mirrors that typically found in other CaRs and in mGluRs (4,10,32). Conserved serine residues in the ECD of mammalian CaRs are suggested to be involved in activation of the receptor by Ca 2ϩ (Ser 147 and Ser 170 (33,34)) and for L-amino acid-mediated potentiation of receptor function (Ser 169 -171 (35)); these serine residues are conserved in tCaR at the identical positions 147 and 169 -171 (5). In a 70-aa ECD sequence just preceding the first membrane-spanning segment, tCaR possesses the nine fully conserved cysteine residues found among vertebrate CaRs; this cysteine-rich region putatively contributes to functional linkage of the venus flytrap domain of the ECD with the TMD (6,36). The eleven predicted N-glycosylation sequons in tCaR are located at corresponding positions to those in mammalian CaRs (29,34). In summary, it is not surprising that there is great similarity among CaRs across species, and even between CaRs and PherRs, especially in their TMDs as opposed to the corresponding ECDs. These ECDs in CaRs contain Ca 2ϩ -binding domains rather than sequences responsive to pheromones.
Other structural features of tCaR are noteworthy. The presence of a substantial C-terminal ICD distinguishes tCaR from PherRs, which have very short (ϳ15-20 aa) ICDs (Fig. 2). tCaR does, however, contain an ICD that is still substantially shorter (by about 100 residues) than that of mammalian CaRs. This abbreviated ICD is sufficient to support activation of PLC in mammalian HEK-293 cells transfected with tCaR cDNA.
Stepwise C-terminal truncation of the bCaR receptor had progressive effects in reducing receptor-mediated PLC activation in previous studies from our laboratory, but with a C-terminal tail of 66 residues still enabling full signaling via PLC (37). Relative to these findings with bCaR, the 96-aa residue ICD of tCaR comprises sufficient length (and, probably, sequence identity or similarity) to confer competence to activate PLC. In the Cterminal tail, there are two amino acids in bCaR (His 880 and Phe 882 ) that appear to be important for trafficking of the receptor to the cell surface or maintaining the receptor there (37). Phospho-ERK1/2 and total ERK1/2 levels in experimental groups exposed to 0.5-10 mM Ca 2ϩ for 5 min were assessed by immunoblotting (panel A) using antisera recognizing either phospho-or total ERK1/2, respectively. Ratios of phospho-ERK1/2 expression to total ERK1/2 expression (panel B) were determined as described under "Experimental Procedures" and normalized to the value obtained for tCaR-expressing cells treated with 0.5 mM Ca 2ϩ (n ϭ 3 immunoblots from two separate experiments). These residues are fully conserved in tCaR (as His 863 and Phe 865 ) as are the surrounding residues in this segment.
These studies demonstrate that tCaR senses changes in [Ca 2ϩ ] o and activates PLC and MAP kinase as evidenced by ERK1/2 phosphorylation. tCaR, however, has a reduced sensitivity to [Ca 2ϩ ] o and lower transduction capacity compared with bCaR. We consistently saw statistically lower InsP accumulation and MAP kinase activation at all levels of [Ca 2ϩ ] o tested in HEK-293 cells, transfected with the same amounts of tCaR cDNA compared with bCaR cDNA. Possible explanations for the observed differences between these two CaRs include: 1) lower transfection efficiency for tCaR compared with bCaR cDNA in HEK-293 cells and consequently reduced cell-surface receptor number; 2) reduced stability of membrane tCaRs compared with bCaRs, perhaps due to the truncated C-terminal tail of tCaR; 3) altered ligand binding and, therefore, lower signal generation by the tCaR, owing to the stretch of missing amino acids in the tCaR ECD corresponding to residues 366 -384 in bCaR; 4) less effective coupling of tCaR to mammalian G proteins and PLC (compared with bCaR) in this expression system; and 5) less productive receptor dimerization by tCaR compared with bCaR.
We attempted to assess receptor protein levels by immunoblotting but those studies have inherent limitations. The epitope against which the polyclonal antiserum was raised is a peptide from bCaR that differs by 5 amino acids (out of 22) from the same stretch in tCaR (residues 214 -235). This antiserum would be expected to detect tCaR protein less efficiently than bCaR, accounting for the need to resolve larger amounts of tCaR protein (85 g versus 5 g for bCaR protein) to visualize tCaR bands by immunoblotting. This technical issue makes comparing the immunoblots in a quantitative manner difficult.
tCaR clearly supports phosphorylation of ERK1/2, which might not be expected based on current knowledge of CaR structure-function relationships. Three groups recently reported that amino acid sequences in the C-terminal segment of the ICD bind filamin (38 -40). These authors suggested that CaR may engage the MAP kinase or other signaling pathways through this scaffolding protein and lead to subsequent activation of Rho GTPases. Hjälm et al. (39) proposed that a 91-aa segment (907-997) of the hCaR C-terminal ICD includes the filamin-binding domain. This segment of the hCaR corresponds to the C-terminal-most regions of the teleost CaRs, beginning presumably at residue 891 in tCaR. However, interestingly, compared with hCaR, the three described teleost CaRs (fCaR, sbCaR, and tCaR) lack a 32-aa sequence at the center of the proposed filamin-binding domain in the mammalian CaR (which would be located between positions 906 and 907 in the teleost CaRs). The teleost CaRs also terminate 9 aa before the end of the proposed filamin-binding domain in the hCaR. In another study, Pi et al. (40) co-immunoprecipitated rCaR, filamin, and Rho, upon which CaR-activated G␣ q is dependent for stimulation of serum response element transcription. Furthermore, they demonstrated inhibition of rCaR-and G␣ q -stimulated serum response element activity by an rCaR minigene comprising a peptide from residues 906 -980 (corresponding to the tCaR sequence beginning from position 890). Finally, Awata et al. (38) assigned the filamin-binding site to within a shorter segment (60 aa in length; positions 972-1031 in hCaR) that is located further downstream along the C-terminal tail of the mammalian receptor. In teleost CaR, this proposed binding domain would begin at position 924 and include only the Cterminal-most 17-aa residues. In the dogfish shark calcium polyvalent cation-sensing receptor (SKCaR), the missing sequence is 35 aa in length, and although the receptor, at 1027 aa in length, is substantially longer than the teleost CaR, the extended ICD bears little sequence similarity to the corresponding region in mammalian receptors. Together, these sequence differences might have suggested that piscine CaRs might not effectively activate MAP kinase phosphorylation in mammalian cell systems. To the contrary, our findings of high [Ca 2ϩ ] o -stimulated, tCaR-mediated ERK 1/2 phosphorylation indicate substantial coupling to MAP kinase activation in this heterologous expression system, despite the absence of ICD sequences in tCaR that have been implicated in this transduction pathway. The absolute minimum sequence requirements for ERK1/2 phosphorylation, with any of these CaRs, has not been finely mapped.
The tissue expression pattern of tCaR mRNA and the differences in expression with salinity adaptation suggest an iono-/ osmoregulatory role for CaRs in kidney and intestine. Supporting these patterns of expression in tilapia are several lines of evidence that point to the involvement of CaRs in ion sensing generally, and in the regulation of Ca 2ϩ metabolism specifically, in vertebrates and in fishes. Several techniques (RT-PCR, in situ hybridization, and immunohistochemistry) confirm abundant CaR expression in mammalian kidney, intestine, and brain (41)(42)(43). Within both the kidney and alimentary tract, CaRs appear to be expressed along the fulllength of the renal tubule and intestinal tract and variably on both apical (lumen-facing) and basolateral cell membranes (44 -50). These general patterns of distribution are mirrored specifically, and further extended, in fishes. CaR is detected in the kidney (including the endocrine corpuscles of Stannius) and urinary bladder, alimentary tract, and gills of several species (sea bream; winter flounder, Pseudopleruonectes americanus; Atlantic salmon, Salmo salar; rainbow trout, Oncorhynchus mykiss; and dogfish shark), and in the secretory shark rectal gland (13,14,17,18). In the brain of fishes, CaR is localized in several regions, including the pituitary gland and olfactory tissues (13,14), and in the European flounder (Platichthys flesus), "CaR protein" has been localized to Dahlgren cells of the caudal neurosecretory system (16). This correlative evidence points to the engagement of CaR in fish iono-/osmoregulation not only at the level of ion-transporting epithelial tissues but also at the level of relevant hormone-secreting endocrine cells.
Physiological evidence also supports the involvement of CaRs in piscine osmoregulation. The role of CaRs in mediating directly (i.e. without the involvement of an endocrine messenger) the effect of Ca 2ϩ on ion transport is proposed for the shark rectal gland. In that tissue, high [Ca 2ϩ ] o triggers an elevation in the intracellular [Ca 2ϩ ] and thereby both promotes rectal artery vasoconstriction and diminishes cAMP-dependent active NaCl secretion (15). It is known that the Ca 2ϩ -sensing function of the CaR is sensitive to ionic strength. Elevations in ionic strength are thought to reduce the receptor's sensitivity to Ca 2ϩ literally by shielding the ligand-binding site. In that manner, it is proposed, the CaR actually functions as a "salinity sensor" in fishes (14,15). Our findings confirm a dependence of Ca 2ϩ sensing on ionic strength for tCaR and bCaR. Hebert et al. (8) proposed for mammalian kidney a local mechanism for Ca 2ϩ -stimulated divalent ion excretion in which CaR activation inhibits NaCl reabsorption in the thick ascending limb and through a consequent reduction in transepithelial voltage, decreases Ca 2ϩ and Mg 2ϩ reabsorption. The indirect control of Ca 2ϩ and ion regulation through CaR-sensitive release of endocrine messengers is suggested by the demonstration of CaR protein in the Dahlgren cells of flounder (16). These cells are part of the caudal neurosecretory system that is a source of urotensins, with both vasoactive and osmoregulatory functions (51), and of parathyroid hormone-related peptide, which is hypercalcemic in the sea bream (16). Renal and urinary blad-der effects of caudal neurosecretory system-derived hormones may be expected in fishes, because the caudal vein, which receives these hormonal secretions, delivers blood to these iontransporting tissues (52). Although these data from Dahlgren cells are circumstantial, there is more direct evidence from the rainbow trout where both extracellular Ca 2ϩ and the calcimimetic NPS R-467 stimulated the release of the teleost hypocalcemic hormone stanniocalcin from the corpuscles of Stannius, and reduced radiocalcium uptake across the gills (17,53).
Electrophysiological recordings from the olfactory nerve of the sea bream demonstrated sensitivity to environmental [Ca 2ϩ ] over a naturally occurring range (0 -10 mM), with reductions in environmental [Ca 2ϩ ] leading to increased nerve firing rate (54). Taken together with molecular and immunohistochemical data on CaR expression in the gills presented here and elsewhere, it appears that in fishes the CaRs are deployed to detect the external as well as the internal (i.e. extracellular) [Ca 2ϩ ].
In conclusion, tilapia expresses a CaR that couples to signaling pathways in mammalian cells. The tissue distribution of tCaR is consistent with possible physiological roles in barrier epithelia and excitable tissues, including perhaps Ca 2ϩ sensing in endocrine cells of the kidney (corpuscles of Stannius) or the brain and pituitary gland (prolactin-and somatolactin-secreting cells). Future immunohistochemical studies to localize tCaR can confirm these findings and add valuable information on subcellular distribution (apical versus basolateral membranes in polarized ion-transporting epithelial cells, for example) that will be useful in relating receptor expression to specific functions such as regulation of divalent cation homeostasis in fishes.