A Carboxyl-terminal Domain Controls the Cooperativity for Extracellular Ca2+ Activation of the Human Calcium Sensing Receptor

Calcium sensing receptors are part of a growing G protein-coupled receptor family, which includes metabotropic glutamate, γ-aminoisobutyric acid, and pheromone receptors. The distinctive structural features of this family include large extracellular domains that bind agonist and large intracellular, carboxyl-terminal domains of as yet undefined function(s). We have explored the contribution(s) of the carboxyl terminus of the human calcium sensing receptor (CaR) by assessing extracellular Ca2+-mediated changes in intracellular Ca2+ in individual HEK-293 cells transfected with CaR clones. In-frame fusion of EGFP to the carboxyl terminus of CaR had no effect on either the dose response for extracellular Ca2+ activation or CaR desensitization. Carboxyl-terminal truncations, fused in-frame with EGFP (CaRΔ1024-EGFP, CaRΔ908-EGFP, CaRΔ886-EGFP, and CaRΔ868-EGFP), were assessed for alterations in Ca2+-dependent activation or desensitization. Significant effects on the dose-response relation for extracellular Ca2+ were observed only for the CaRΔ868 truncation, which exhibited a decreased affinity for extracellular Ca2+ and a decrease in the apparent cooperativity for Ca2+-dependent activation. The alterations in extracellular Ca2+ affinity and cooperativity observed with CaRΔ868 were recapitulated by a point mutation, T876D, in the full-length CaR-EGFP background. All truncations with wild type dose-response relations exhibited desensitization time courses that were comparable to the full-length CaR, whereas the CaRΔ868 receptor desensitized completely after two exposures to 10 mmCa2+. Interestingly, the CaR point mutation T876D exhibited desensitization comparable to wild type CaR, suggesting that this mutation specifically modifies CaR cooperativity. In conclusion, these studies suggest that amino acid residues between 868 and 886 are critical to the apparent cooperativity of Ca2+-mediated activation of G proteins and to CaR desensitization.

Calcium sensing receptors (CaR) 1 have been identified in many tissues, including neurons (1,2), retinal epithelium (3), kidney (4), intestine (5,6), parathyroid (7), bone (8), keratinocytes (9), and fibroblasts (10). Its role in many of these tissues has yet to be fully defined, although a common functional consequence of extracellular Ca 2ϩ activation of CaR is a G qmediated increase in phospholipase C␤ activity leading to increases in phosphoinositide hydrolysis and a release of Ca 2ϩ from thapsigargin-sensitive intracellular stores (11). In some cell types, CaR activation also leads to rapid alterations in c-myc expression (5) or activation of the mitogen-activated protein kinase cascade (10), suggesting coupling to multiple signal transduction pathways in any given cell type/tissue is possible.
Calcium sensing receptors represent a novel member of a growing family of G protein-coupled receptors (family C), which include metabotropic glutamate receptors (mGluRs) (12,13), calcium sensing receptors (CaRs) (7,14), ␥-aminoisobutyric acid receptors (15), and a newly cloned multigene family of putative pheromone receptors identified in the rat vomeronasal organ (16 -18). Agonists for this family of receptors represent simple molecules present in the extracellular milieu, e.g. amino acids and ions, and these receptors most likely developed as a fusion between bacterial periplasmic binding proteins and the seven transmembrane/signaling domains common to G protein-coupled receptors (19,20). Agonist binding occurs within the large extracellular NH 2 -terminal domain (21), rather than in a pocket defined by the seven transmembrane helices as is characteristic of most G protein-coupled receptors (22,23). A cluster of CaR mutations identified in patients with hyper-and hypoparathyroidism, which affect CaR affinity for Ca 2ϩ , cluster within the extracellular domain (11), suggesting this as the location of the Ca 2ϩ binding site(s). Activation of CaR by extracellular Ca 2ϩ is highly cooperative (Hill coefficients for activation in the range of 2-4) both in vivo (24) and in vitro (25), and although the molecular mechanism is undefined, it has been suggested to result from multiple Ca 2ϩ binding sites (25,26).
A second distinctive feature of many members of this receptor superfamily is the large intracellular carboxyl-terminal domain. Several mGluR genes are alternatively spliced to produce receptors having variations within the carboxyl terminus (27). Studies on heterologously expressed mGluR1 (a, b, or c) or mGluR5 (a or b) variants suggest that carboxyl-terminal differences affect receptor-mediated signaling (28,29) as well as the level of agonist-independent activity (30). Further, the carboxyl-terminal domain of mGluR1 has been shown to affect the sensitivity of the receptor to agonist in chimeras having identical NH 2 -terminal agonist-binding domains (28). Finally, the large carboxyl terminus may be involved in interactions with proteins that assist in transport of expressed receptors to the plasma membrane, as has been shown for Homer-mediated trafficking of mGluRs (31). Calcium sensing receptors also have a large carboxyl terminus (Ͼ210 amino acids). Truncations within the carboxyl terminus have recently been shown to severely limit phosphatidylinositol hydrolysis when the residual carboxyl terminus was limited to Ͻ15 amino acids, and a region between amino acids 874 and 888 has been identified as crucial to normal signal transduction (32).
In this report, we explore the functional consequences of carboxyl-terminal truncations in greater detail, utilizing an assay with enhanced sensitivity and time resolution, applied to individual cells. In order to resolve subtle differences in receptor function, we have characterized the truncations with respect to both the dose-response relation for Ca 2ϩ -dependent activation and CaR desensitization, since desensitization of G protein-coupled receptors often involves phosphorylation and/or conformation-dependent exposure of sequestration motifs within the carboxyl-terminal domain (33,34). Finally, all studies were performed with CaR constructs fused to EGFP at the carboxyl terminus, to allow identification of cells expressing receptors. We find that carboxyl-terminal fusion with EGFP does not affect the functional properties of CaR, and provides a useful marker for assessment of absolute expression levels of CaR within the cell. Modest truncations of the carboxyl terminus (CaR⌬1024, CaR⌬908, and CaR⌬886) had no effect on membrane localization, Ca 2ϩ dose response, or rates of desensitization, whereas truncation CaR⌬868 both altered the dose-response relation for Ca 2ϩ and enhanced desensitization. Finally, a point mutation T876D in the full-length CaR-EGFP background recapitulated the effects observed with CaR⌬868 with respect to the Ca 2ϩ dose-response relation, suggesting that this residue is critically involved in determining the cooperativity of G protein activation.

EXPERIMENTAL PROCEDURES
Materials-The full-length human CaR clone (in pBluescript) was provided by Dr. Jim Garrett of NPS Pharmaceuticals (Salt Lake City, UT); pEGFP was obtained from CLONTECH. Restriction enzymes were obtained from Life Technologies Inc. and New England Biolabs.
Plasmid Construction-PCR reactions were performed with PfuI polymerase (Stratagene) using 94, 53, and 72°C as denaturing, annealing, and elongation temperatures, respectively. All constructs were confirmed by restriction digests and automated sequencing. Full-length human CaR was subcloned into pCDNA3.1 (Invitrogen) after HindIII-NotI digestion (pCCaR). In order to facilitate identification of transfected cells, CaR and EGFP were sometimes expressed behind different promoters in the same mammalian expression vector (described as CaR construct/EGFP). The EGFP gene was excised from its original vector (pEGFP-N1; CLONTECH) by NheI-NotI digestion and inserted downstream of the SV40 promoter of pCCaR after digestion with BstBI-AvrII (replacing the neomycin resistance gene). To create a CaR-EGFP fusion protein, a two-step cloning procedure was used. First, a PCR product was generated using a 5Ј primer which anneals before the XmaI site of pBluescript/CaR (TTC CGC AAC ACA CCC ATT GTC AAG G), and a 3Ј primer in which the CaR gene stop codon was replaced by a BclI site (GCT GAT CAA CTA CGT TTT CTG TAA CAG T). After PCR amplification, the product was digested with XmaI and subcloned back into the original template using XmaI and a T4 polymerase filled-in NotI site. The final chimeric gene was created by excising the mutated CaR gene from pBluescript/mutCaR using NotI-BclI and inserting it between the NotI-BamHI sites of pEGFP (CLONTECH). Two of the truncations were constructed by direct subcloning of portions of the CaR gene from pBluescript/CaR into pEGFP (CLONTECH). For CaR⌬886-EGFP, the CaR gene was cut and subcloned into pEGFP-N1 with HindIII-XmaI. The same was done for CaR⌬908-EGFP using HindIII-BamHI and pEGFP-N3. For two other truncations (CaR⌬1024-EGFP and CaR⌬868-EGFP), site-directed mutagenesis PCR products were generated using pBluescript/CaR as a template, a common 5Ј primer (GAA TTG TAA TAC GAC TCA CTA TAG GGC GA) and 3Ј primers containing an in-frame BamHI site (CGG TCG GAT CCA AGT CCG TTT CCC C for CaR⌬1024-EGFP and GGC GAT GGA TCC GGT GTT GCG GGA TGG C for CaR⌬868-EGFP). The PCR products were cut with HindIII-BamHI, purified by electrophoresis on 1% agarose gels, isolated, and subcloned into pEGFP. CaR⌬868 was also cloned into pCCaR/EGFP, replacing the full-length CaR gene. The PCR strategy described above was performed, using the same 5Ј primer and a 3Ј primer containing a stop codon immediately before a BamHI site (CAA CGC GGA TCC TAG ATG GTG TTG CGG GAT GG). CaR-T876D-EGFP was produced by PCR-based site-directed mutagenesis using the CaR-EGFP construct as template (5Ј primer, GCA CCT CCT CGA TGG TGT TGC G, 3Ј primer GTT GCA GCG ACG CAG CTC ACG C). Fig. 1 illustrates the topology of the human CaR, the location within the carboxyl-terminal domain of the various truncations, and the fusion proteins that were produced with EGFP.
Cell Culture/Transfections-HEK-293 cells were grown in high glucose Dulbecco's modified Eagle's medium (Life Technologies Inc.) supplemented with 10% bovine serum albumin (Sigma), 50 units/ml penicillin, and 50 g/ml streptomycin (37°C, 5% CO 2 ). Cells were transfected with 1 g of each construct using Superfect (Quiagen), plated on collagen-coated coverslips, and kept in a 5% CO 2 incubator at 37°C until use. Some constructs (CaR, CaR/EGFP, CaR-EGFP, CaR⌬868, and CaR⌬868-EGFP) were also stably transfected after linearization with SspI. The constructs CaR/EGFP and CaR⌬868/EGFP did not contain a neomycin resistance gene and were therefore cotransfected with linearized pCDNA3.1 in a 10:1 ratio. Cells were kept in medium containing 300 g/ml Geneticin (Life Technologies, Inc.). Colonies were selected after 3 weeks, and clones were confirmed by both cell fluorescence and CaR activity.
Confocal Microscopy-Coverslips containing transfected HEK-293 cells were fixed with acetone for 5 min at Ϫ20°C, mounted on glass slides with Vectashield (Vector Laboratories), and viewed with a Noran Oz confocal microscopy system. Selected regions of the field were excited at 340/380 nm (emission wavelength 510 nm) at 8-s intervals (8 frame averaging) on a Universal imaging system based on the MetaFluor software package. Background images at the same settings used during a particular experiment were obtained on regions of the coverslip devoid of cells. All solutions were osmolality-matched (290 -300 mOsm), measured on a Wescor 5500 vapor pressure osmometer (Salt Lake City, UT). Variations in extracellular Ca 2ϩ concentration were produced by isosmolar substitution for NaCl. All experiments were performed at room temperature (22-24°C). Multiple cells were analyzed from at least three independent transfections or cell passages (for stably transfected cell lines). Data were normalized/averaged as described in the text, and are presented as mean Ϯ standard error of the mean. Curves were fitted by least squares minimization using the Marquardt-Levenberg algorithm (NFIT, Island Products, Galveston, TX).

Comparison of the Functional Properties of Human CaR and
CaR-EGFP-To determine the localization and expression levels of the various CaR constructs, carboxyl-terminal fusions to EGFP were used (Fig. 1B). Preliminary studies were performed to confirm that EGFP did not affect the functional properties of Control experiments were performed in HEK-293 cells stably transfected with CaR-EGFP to determine whether EGFP interfered with measurements of intracellular Ca 2ϩ monitored by changes in fura-2 fluorescence. Cells expressing CaR-EGFP but not loaded with fura-2-AM were exposed to 0.5 and 10 mM extracellular Ca 2ϩ and the fluorescence measured at an emission wavelength of 510 nm (excitation wavelengths were 340 and 380 nm). The 340/380 fluorescence ratio was low (Ͻ0.5) and was unchanged by alterations in bath Ca 2ϩ . A second control determined the intracellular Ca 2ϩ responses to changes in extracellular Ca 2ϩ concentrations in the range from 1.5 to 30 mM in untransfected or vector (pCDNA3.1)-transfected HEK-293 cells. No significant changes in intracellular Ca 2ϩ were elicited in the absence of CaR.
A further means of comparing the functional properties of CaR and CaR-EGFP was to assess the rate of desensitization of the intracellular Ca 2ϩ response to repeated exposures to 10 mM extracellular Ca 2ϩ . Fig. 3 (A and B) illustrates the protocols used. For protocol 1 (illustrated in Fig. 3A), cells stably transfected with CaR were alternately exposed to 10 mM bath Ca 2ϩ for 2 min, followed by 0.5 mM Ca 2ϩ for 3 min. This protocol resulted in a modest rate of CaR desensitization over a 30-min period, as reflected by the peak response to successive 10 mM Ca 2ϩ applications. Protocol 2 (illustrated in Fig. 3B) utilized a 4-min exposure to 10 mM Ca 2ϩ , followed by a 2-min "recovery" period in 0.5 mM Ca 2ϩ . This protocol resulted in significant desensitization of CaR (Ϸ40 -50%) over a 30 min experiment. The responses to both protocols strongly suggest that the decline in response is due to CaR desensitization and not to slow depletion of intracellular Ca 2ϩ stores, since the extent of desensitization is proportional to the duration of receptor stimulation. Fig. 3C illustrates the time courses for desensitization of the peak intracellular Ca 2ϩ responses for CaR and CaR-EGFP for the two protocols, with all data normalized to the first peak response to 10 mM Ca 2ϩ of each cell. Carboxyl-terminal fusion of CaR to EGFP had no effect on the rate(s) of receptor desensitization.
Localization of CaR Carboxyl-terminal Truncations to the Plasma Membrane-Truncations of the carboxyl terminus of CaR were produced as fusion proteins with EGFP, as illustrated in Fig. 1B. Each construct was assessed for its cellular distribution prior to functional studies. Illustrated in Fig. 4 Table I. relation between EGFP fluorescence intensity (fluorescein isothiocyanate filter) and plasma membrane localization assessed by increases in intracellular Ca 2ϩ (fura-2 fluorescence changes) existed for all CaR-EGFP constructs.
Dose-Response Relations for Ca 2ϩ -mediated Activation of CaR Carboxyl-terminal Truncations-Experiments such as those illustrated in Fig. 2A were performed to assess the doseresponse relations for extracellular Ca 2ϩ -mediated activation of the CaR carboxyl-terminal truncations. Fig. 5 illustrates the results of fits to data averaged from at least three independent transfections for each construct, and Table I indicates the parameters of the fits. Only CaR⌬868-EGFP exhibited a difference in response to extracellular Ca 2ϩ , with both a decreased affinity for Ca 2ϩ (EC 50 of 10.9 mM) and a reduction in the Hill coefficient (from the range of 2.0 -2.9 down to 1.1). Since the ⌬868 truncation leaves a cytoplasmic tail of only 8 amino acids, we also determined the dose-response relation for an additional construct, CaR⌬868 (coexpressed with soluble EGFP) to eliminate the possibility that EGFP presented a physical barrier for access of G proteins to their sites of interaction with CaR. HEK-293 cell lines stably expressing CaR⌬868 were produced and the dose-response relationship for extracellular Ca 2ϩ determined. Results were similar for both the CaR⌬868-EGFP and CaR⌬868 constructs (data combined in Fig. 5 and Table I), implying that the decreased responsiveness to Ca 2ϩ is a result of the truncation and not of the carboxyl-terminal fusion to EGFP.
Desensitization of CaR Carboxyl-terminal Truncations-The progressive desensitization of peak intracellular Ca 2ϩ in response to repeated exposures to 10 mM extracellular Ca 2ϩ was assessed for all of the CaR carboxyl-terminal truncations, according to the two protocols described, i.e. protocol 1, 2 min 10 mM Ca 2ϩ /3 min 0.5 mM Ca 2ϩ ; and protocol 2, 4 min 10 mM Ca 2ϩ /2 min 0.5 mM Ca 2ϩ . Illustrated in Fig. 6 are the combined data for all CaR carboxyl-terminal truncations. For CaR⌬1024-EGFP, CaR⌬908-EGFP, and CaR⌬886-EGFP, desensitization was unchanged from that observed for full-length CaR (lines drawn in the figure represent desensitization of full-length CaR in response to the two protocols, redrawn from Fig. 3C). Desensitization of CaR⌬868-EGFP, however, was significantly enhanced, with full desensitization occurring after two applications of agonist.
Comparison of CaR⌬868 with a Point Mutation, CaR-T876D-EGFP-Results with the truncations suggest that the region between amino acids 868 and 886 is critical for normal functioning of CaR by several criteria; 1) CaR⌬868 exhibits a decrease in the apparent affinity for extracellular Ca 2ϩ , 2) the degree of cooperativity as reflected in the Hill coefficient is decreased, and 3) desensitization of CaR⌬868 is significantly increased. The region between amino acids 874 and 888 has been previously identified as important in normal signal transduction (32). We sought to identify residues(s) critical to the observed alterations in Ca 2ϩ signaling, and have identified a point mutation, T876D (produced in the background of fulllength CaR-EGFP), which recapitulates a specific property observed with the truncation, CaR⌬868-EGFP, i.e. T876D exhibits a reduced affinity for Ca 2ϩ with respect to activation, and a reduction in the apparent cooperativity for Ca 2ϩ -dependent activation as reflected in the Hill coefficient. Fig. 7A illustrates the dose-response relations. Parameters of the fits are listed in Table I.
CaR⌬868-EGFP and T876D are different, however, with respect to desensitization. Fig. 7B compares the second response to 10 mM Ca 2ϩ (after a 3-min recovery in 0.5 mM extracellular Ca 2ϩ ) for three CaR constructs, full-length CaR, CaR⌬868-EGFP, and T876D. It is clear that T876D exhibits only a modest degree of desensitization, comparable to that of the full-length wild type CaR, whereas CaR⌬868-EGFP exhibits strong desensitization. DISCUSSION The series of carboxyl-terminal truncations in this study contributes to a mapping of functional domains within the carboxyl terminus of CaR. Significant truncations, i.e. removal of up to 192 of the 218 amino acids within the carboxyl terminus, had no effect on either extracellular Ca 2ϩ -dependent activation of the G q -mediated signaling pathway or desensitization of the response upon repeated exposures to 10 mM extracellular Ca 2ϩ . This implies that amino acid residues 886 -1078 are not critical to these particular acute responses. A recent study utilizing calcium-stimulated phosphoinositide hy- drolysis as a measure of CaR activity demonstrated that truncations at 865 or 874 were inactive despite adequate cell surface expression (32). Therefore, the most significant and unexpected result in the present study, most likely attributable to the increased sensitivity and time resolution obtained when utilizing intracellular Ca 2ϩ as a measure of CaR activity, was that the CaR⌬868-EGFP truncation was active. CaR⌬868-EGFP activity was, however, distinctly different from the other truncations or the full-length CaR, exhibiting both a decrease in the apparent affinity for extracellular Ca 2ϩ -mediated G q activation and a decrease in the apparent cooperativity. This result was recapitulated by a point mutation, T876D, suggesting that this residue within the proximal carboxyl-terminal domain is pivotal in determining the affinity and apparent cooperativity for Ca 2ϩ activation of CaR. Further support for this conclusion comes from additional CaR clones having point mutations at neighboring residues: R873A, C874A, or S875A; all of these mutated CaRs had wild type properties with respect to Ca 2ϩ -dependent activation (data not shown). Further explo-ration of the region between residues 868 and 886 by sitedirected mutagenesis is under way.
A steep Ca 2ϩ -response relation is a hallmark of CaR activation, observed both in freshly isolated bovine parathyroid cells (24) and in cell lines expressing recombinant CaR (25,26). Highly cooperative activation is unique to CaR, and is not observed even in the related receptor family, mGluR (23,35,36). There is little precedent, therefore, for the molecular origins of CaR cooperativity. Several explanations are possible, as initially suggested by Ruat et al. (26): 1) multiple Ca 2ϩ binding sites within the extracellular domain; 2) cooperative interactions between multiple CaR; or 3) highly cooperative, efficient activation of G proteins. There is as yet no information about the existence of multiple binding sites for Ca 2ϩ on CaR, since high affinity, selective agonists and/or antagonists suitable for binding studies have not yet been identified. A recent report supports the interaction among multiple CaR as a possible source of cooperativity, since CaR has been shown to be present as covalent disulfide-linked dimers in membranes (37). Further studies are needed, however, before the functional consequences of CaR dimerization are understood. The present results suggest that the third possibility, i.e. highly cooperative, efficient activation of G proteins, may be the explanation for the apparent cooperativity of Ca 2ϩ activation of CaR. A single  Fig. 2A were performed on HEK-293 cells transiently transfected with the CaR carboxyl-terminal truncations illustrated in Fig. 1B. Data from at least 40 cells from two to three independent transfections were averaged at each concentration. All data were fitted with the Hill equation, and the fit parameters are listed in Table I. Symbols are: CaR (closed circles), CaR-EGFP (closed triangles), CaR⌬1024-EGFP (closed squares), CaR⌬908-EGFP (closed diamonds), CaR⌬886-EGFP (closed inverted triangle), and CaR⌬868-EGFP (closed hexagons).

TABLE I Comparison of dose-response relations obtained for full-length human
CaR and carboxyl-terminal truncations, expressed in human embryonic kidney cells Data from at least 50 cells from at least three independent experiments (CaR and CaR-EGFP, CaR⌬868) or three independent transfections (CaR⌬1024-EGFP, CaR⌬908-EGFP, CaR⌬886-EGFP, and CaR⌬868-EGFP) were combined, normalized to the peak response at 30 mM Ca 2ϩ , and fitted with the following equation  (25,26). One possible explanation of variable cooperativity is that the various agonists differ in their ability to induce the highly cooperative conformation of the carboxyl-terminal domain. There is precedence for a differential ability of full and partial agonists to induce active conformations of G proteincoupled receptors (38 -40).
The amino acid residues between 868 and 886 contain no putative phosphorylation sites. In addition, there is no sequence homology with other G protein-coupled receptors, except a dipeptide, ST, which aligns with the same two residues within the carboxyl terminus of the two mGluR5 splice variants, mGluR5a and b. The threonine residue in mGluR5a is a protein kinase C phosphorylation site (41) and is responsible for the intracellular Ca 2ϩ oscillations observed upon activation of these receptors (41,42). The comparable amino acid residue in mGluR1, which does not exhibit intracellular Ca 2ϩ oscillations, is aspartic acid (41). This region, and in particular, these residues, are therefore implicated as part of the G protein interaction site (in combination with the i2 intracellular loop). We therefore mutated the threonine residue in CaR to aspartic acid. Although intracellular Ca 2ϩ oscillations were still observed in CaR-T876D (data not shown), the cooperativity with respect to activation was altered, strongly implicating this amino acid residue as critical to the G protein interaction site, specifically mediating highly cooperative G protein activation.
A second feature of CaR function, which is localized to the region of the proximal carboxyl terminus between amino acids 868 and 886, is acute desensitization of the peak CaR-mediated response. The CaR⌬868-EGFP truncation desensitizes very rapidly compared with less severe truncations or wild type CaR. All of the carboxyl-terminal domain's putative protein kinase A and C sites (43) are eliminated in the CaR⌬886-EGFP truncation, which desensitizes normally. Therefore, these protein kinase sites do not participate in acute desensitization of CaR. Additional potential protein kinase C sites are localized to intracellular loops; in particular, there is one in the i3 loop that contributes, in many G protein-coupled receptors, to G protein activation. CaR⌬886-EGFP may desensitize normally because the amino acid residues between 868 and 886 retard access of the proteins involved in desensitization to the relevant phosphorylation site(s). Direct interactions between G protein-coupled receptor kinases and possibly arrestin with muscarinic receptors have been functionally demonstrated (44), although the consensus binding site(s) on the receptor have not been identified. Deletion of the amino acid residues of CaR between 868 and 886 may expose potential regulatory protein interaction site(s), leading to more rapid and complete phosphorylation and desensitization of CaR. Further studies with sitedirected mutants that eliminate all potential phosphorylation sites on CaR will be required to confirm this possibility.
Although the distal carboxyl terminus (from amino acids 886 -1078) is not involved in modulating acute CaR activation (reflected in the dose-response relation and desensitization rate), there are other potential role(s) for this domain. Several mGluR1 genes are expressed as slice variants, which differ primarily in the length of the carboxyl terminus (27). The longer splice variant of mGluR1 exhibits significant agonistindependent activity (30), and the carboxyl-terminal domain of  Table I. B, comparison of second response to application of 10 mM Ca 2ϩ for three constructs: full-length CaR, CaR⌬868-EGFP, and T876D, as a percentage of the first response. Experiments were performed as in Fig. 3A, i.e. there was a 3-min interval in 0.5 mM Ca 2ϩ between the first and second 2-min applications of 10 mM Ca 2ϩ . this variant (plus the i2 loop) confers enhanced sensitivity to agonist with respect to G protein activation (28). Although only a single CaR gene has been studied to date, functional differences have been observed among CaR cloned from different species (45), and several additional partial clones that localize to different mouse chromosomes have been identified (14). A family of CaR genes may therefore soon be available for more detailed sequence/function comparisons within the carboxyl terminus as well as throughout the protein.
Another potential role for the distal carboxyl terminus of CaR may be in trafficking and/or targeting of the receptor within particular cell membrane domains. CaR is highly expressed in various regions of the brain (46), and a subcellular localization within pre-and/or post-synaptic domains (1,47), in analogy with mGluRs, may require interaction with targeting proteins such as Homer (31), which has been recently shown to participate in mGluR transport to the plasma membrane in an activity-dependent manner. Significant concentrations of both mGluRs and CaR have been observed within intracellular membranes in vivo (31,47), and thus a similar mechanism may operate with respect to CaR targeting. Finally, CaR is highly expressed in epithelia within the kidney (48) and intestine (6), and therefore the potential for interactions with the cytoskeleton or other proteins within the target membrane to anchor and/or retain CaR may occur via this domain. The available of a series of carboxyl-terminal truncations of CaR which retain activity but have variable deletions of potential protein-protein interaction domains will make hunting for regulatory protein partners straightforward.
In this study, we have begun to dissect the molecular roles for the carboxyl terminus of CaR. A proximal portion, within 20 amino acid residues of the membrane, contributes specifically to regulation of the cooperativity of G protein activation, as well as to acute receptor desensitization. The availability of functional EGFP-tagged receptor mutants and truncations makes further exploration of the role(s) for the carboxyl-terminal intracellular domain of CaR feasible, and may generate paradigms generalizable to the superfamily of G protein-coupled receptors including CaR, mGluRs, ␥-aminoisobutyric acid receptors, and pheromone receptors.