I. INTRODUCTION

Top Previous Next References

A great deal has occurred in the field of extracellular Ca²⁺ (Ca_o²⁺) sensing since 1991 when an earlier article in Physiological Reviews addressed this subject (52). At that time it was apparent that certain cells, such as the chief cells of the parathyroid gland, were capable of sensing (i.e., recognizing and responding to) small changes in the extracellular ionized calcium concentration. Moreover, indirect evidence suggested that Ca_o²⁺ sensing by parathyroid cells involved a process sharing certain properties with the mechanism through which G protein-coupled, cell surface receptors for a variety of extracellular messengers (e.g., peptides, catecholamines, prostaglandins) responded to their respective agonists (48, 95, 220, 315, 409). The cloning of a G protein-coupled Ca_o²⁺-sensing receptor (CaR) from bovine parathyroid gland in 1993 (58) proved that the calcium ion can, in fact, serve as an extracellular first messenger.¹

This review addresses the following areas in which progress has been particularly rapid over the past 5-10 years in elucidating the mechanisms underlying Ca_o²⁺ sensing. First, we briefly review the cloning of the CaR from various cells expressing it and discuss what is known about its homology with other members of the superfamily of G protein-coupled receptors (GPCRs), including some that also sense Ca_o²⁺. Next we describe what is known about the structure of the CaR gene and the factors regulating its expression and review the results of recent studies on the receptor's structure-function relationships and the various intracellular signaling pathways to which it couples. The discussion then covers the rapidly expanding range of cellular functions regulated by the CaR, including its physiological roles in the cells expressing it that are involved in as well as those that are uninvolved in systemic mineral ion homeostasis. Finally, we address the related area of Ca_o²⁺ signaling, that is, the mechanisms that underlie local and/or systemic changes in Ca_o²⁺, thereby producing signals that can modulate the receptor's activity both in tissues involved in systemic Ca_o²⁺ homeostasis as well as those that are not. We do not address disorders of Ca_o²⁺ sensing resulting from abnormalities in the CaR's structure and/or function (for review, see Ref. 54), except insofar as they elucidate the CaR's role of normal physiology.

	II. MOLECULAR CLONING OF PARATHYROID, RENAL, AND OTHER EXTRACELLULAR CALCIUM-SENSING RECEPTORS

Top Previous Next References

Expression cloning in Xenopus laevis oocytes enabled isolation of a single 5.3-kb clone (BoPCaR = bovine parathyroid Ca_o²⁺-sensing receptor) that exhibited pharmacological properties very similar to those of the Ca_o²⁺-sensing mechanism expressed endogenously in bovine parathyroid cells (58, 96, 368). Nucleic acid hybridization-based techniques then led to the cloning of full-length CaRs from several different tissues in various mammalian species, including human parathyroid (157); rat (381), human (3), and rabbit kidney (71); rat C cells (146, 158); and striatum of rat brain (392). All are highly homologous (>90% identical in their amino acid sequences to BoPCaR) and represent species and tissue homologs of the same ancestral gene.

A full-length CaR has also been cloned and characterized from chicken parathyroid (126), and a smaller segment of the CaR has been amplified and sequenced by RT-PCR from gastric mucosa of the mudpuppy, an amphibian (104). The chicken CaR and the portion of the mudpuppy CaR available for analysis exhibit slightly lower but still very substantial levels of homology to mammalian CaRs (84% identity at the amino acid level for the chicken CaR and 84% identity at the nucleotide level for the mudpuppy CaR), stressing the high degree of conservation of this gene among members of the mammals, birds, and amphibians examined to date.

Mammals, birds, amphibians, and reptiles, the so-called tetrapods, e.g., organisms having four extremities, all possess parathyroid glands and utilize Ca_o²⁺ homeostatic mechanisms similar in their overall design (335). Not surprisingly perhaps, given its wide tissue distribution, particularly in tissues apparently uninvolved in systemic mineral ion metabolism (see sect. XIII), the CaR did not originate when the parathyroid gland first appeared during evolution. A highly homologous CaR gene has been identified in fishes and in the dogfish shark (26). These species have levels of Ca_o²⁺ in their blood and extracellular fluids that are not dissimilar from those in humans and other mammals. They utilize hormones distinct from parathyroid hormone (PTH), such as stanniocalcin in fishes, for example (446, 447), to maintain Ca_o²⁺ homeostasis. Little work on the CaR's role in mineral ion homeostasis is available in these species. Further work is needed to determine whether, similar to its role in tetrapods, the CaR in these aquatic species controls Ca_o²⁺ both by regulating the secretion of calciotropic hormones, which then modulate the functions of target tissues (e.g., kidney, intestine, and gill), and by exerting direct, CaR-mediated actions on mineral ion transport by the latter. Given the crucial roles that both extra- and intracellular calcium play in essentially all organisms, it will be of great interest in future studies to understand the ontogeny and phylogeny of the CaR over a much broader evolutionary scale.

Surprisingly, given the diversity of structurally related GPCRs (see sect. III), an extensive search has yet to uncover definitive evidence for additional CaR isoforms arising from distinct genes, although there are several splice variants of the receptor that are expressed in various tissues (see sect. VC). The latter are currently of uncertain physiological relevance. Furthermore, there may be additional, physiologically relevant Ca_o²⁺ sensors, which are described in section IV.

The amino acid sequences of BoPCaR and the other CaRs cloned to date that are predicted from their nucleotide sequences reveal a common overall topology, which includes a very large (~600 amino acids), NH₂-terminal extracellular domain (ECD), a central core of some 250 amino acids with seven predicted transmembrane domains (TMDs) that are characteristic of the superfamily of GPCRs and a large intracellular COOH-terminal tail of ~200 amino acids (58) (Fig. 1). As described in more detail in section VID, data from studies on the CaR's structure-function relationships indicate that Ca_o²⁺ binds to its ECD. The receptor's ECD also contains multiple N-linked glycosylation sites (58, 134), whereas its intracellular domains [three intracellular loops (ICLs) and COOH tail] harbor several predicted consensus protein kinase C (PKC) and protein kinase A (PKA) phosphorylation sites (the PKA sites are present in all species studied to date except the bovine CaR). The PKC phosphorylation sites are known to modulate the receptor's activity, whereas the physiological relevance of the PKA sites, if any, is currently unknown (18, 58, 89) (see sect. VIF).

View larger version (43K): [in this window] [in a new window]   Fig. 1. Schematic representation of the principal predicted topological features of the extracellular Ca2+-sensing receptor (CaR) cloned from human parathyroid gland. SP, signal peptide; HS, hydrophobic segment; PKC, protein kinase C. Also delineated are missense and nonsense mutations causing either familial hypocalciuric hypercalcemia or autosomal dominant hypocalcemia. These are indicated using the three-letter amino acid code, with the normal amino acid indicated before and the mutated amino acid shown after the numbers of the relevant codons. [From Brown EM, Bai M, and Pollak M. Familial benign hypocalciuric hypercalcemia and other syndromes of altered responsiveness to extracellular calcium. In: Metabolic Bone Diseases (3rd ed.), edited by Krane SM and Avioli LV. San Diego, CA: Academic, 1997, p. 479-499.]

	III. MOLECULAR SIMILARITY OF THE EXTRACELLULAR CALCIUM-SENSING RECEPTOR TO OTHER G PROTEIN-COUPLED RECEPTORS

Top Previous Next References

Based on the evolutionary tree predicted from the database for GPCRs [GCRD; http://www.uthscsa.edu (251)], the CaR belongs to the recently described family C within this large superfamily of genes. Family C GPCRs are defined as a group of receptors comprising at least three different subfamilies that share 20% amino acid identity over their seven membrane-spanning region (251) (Fig. 2). Group I includes the metabotropic glutamate receptors, mGluRs 1-8, which are receptors for the excitatory neurotransmitter glutamate and are widely expressed in the central nervous system (CNS) (311, 312). Unlike the ionotropic glutamate receptors (iGluRs) [i.e., the N-methyl-D-aspartate (NMDA) receptor], which are ion channels containing a binding site for their physiological agonist glutamate within the same channel molecule, the mGluRs are GPCRs.

View larger version (22K): [in this window] [in a new window]   Fig. 2. "Tree" diagram showing the degrees of homology and proposed evolutionary relationships among the various members of the family C G protein-coupled receptors (GPCRs) described to date. The farther to the left that a given receptor branches off, the less related it is to the other receptors. For details see text. [From Brown et al. (61).]

Group II contains at least two types of receptors: the CaR and a recently discovered, multigene subfamily of putative pheromone receptors, VRs (vomeronasal receptors) or GoVNs (193, 286, 397). The latter are found exclusively in neurons of the vomeronasal organ of the rat (VNO; hence the VRs) (a small sensory organ thought to be involved in regulating instinctual behavior through input from environmental pheromones) that express the guanine regulatory (G) protein, Ga_o (hence the GoVNs) (286). Additional receptors closely related to the CaR and/or VRs have recently been identified in mammals (203) and fishes (74, 310), which are taste and putative odorant receptors, respectively. These related receptors in fishes may represent evolutionary precursors of the pheromone receptors in terrestrial organisms (e.g., rats); they exhibit the topology characteristic of the family C GPCRs and are most closely related to the CaR among the members of this family.

Group III contains a subfamily of receptors, the GABA_B receptors, that bind and are activated by the inhibitory neurotransmitter GABA (238). As with the receptors for glutamate, there are both G protein-coupled and ionotropic (e.g., ligand-gated receptor channels) receptors for GABA (the latter are known as GABA_A receptors) (458). Interestingly, the formation of a functional GABA_B receptor capable of activating an inwardly rectifying K⁺ channel requires heterodimerization of the two different members of this subfamily of receptors identified to date [e.g., GABA_BR1 (of which there are two splice variants: GABA_BR1a and GABA_BR1b) and GABA_BR2], while homodimers of the individual GABA_B receptor subtypes do not activate this ion channel (218, 239, 257, 456). The GABA_BR2 form of this receptor, however, can inhibit adenylate cyclase when expressed by itself in heterologous mammalian expression systems (257). As described in more detail in sections VIA and VIB, the CaR resides on the cell surface (16) and in detergent extracts of at least some CaR-expressing cells (452) as a dimer [e.g., in bovine parathyroid chief cells, the epithelial cells of the inner medullary collecting duct (IMCD) of the rat kidney, and in CaR-transfected human embryonic kidney (HEK293) cells]. Furthermore, there appear to be functional interactions between the individual monomeric subunits within these dimers, as detailed in section VIE (17).

The extracellular, ligand-binding domains of the family C GPCRs are structurally related to those of the bacterial periplasmic binding proteins (PBPs) (110, 329). This hypothesis was initially based on molecular modeling (330, 428) using the known clawlike, three-dimensional structure of the PBPs (329). More recent studies have identified a significant degree of sequence homology when comparing the ECDs of the GABA_B receptor and the leucine-isoleucine-valine (LIV) bacterial nutrient-binding protein (238). This observation adds strong support to the hypothesis that there is an evolutionary link between the ECD of the various members of the family C GPCRs and the bacterial periplasmic, nutrient-binding proteins. Furthermore, recent data indicate that homologous regions of the ECDs of the GABA_B receptors, mGluRs, and CaR may participate in the binding of GABA, glutamate, and Ca_o²⁺ (41), respectively, while a nearby but spatially distinct region of the ECDs of some mGluRs participates in the modulation of these latter receptors' activities by Ca_o²⁺ (256) (see sect. IVA).

The bacterial PBPs include at least eight families that recognize a broad range of extracellular solutes, which are destined for cellular uptake and/or elicit chemotactic responses (407, 428). These solutes include organic nutrients as well as inorganic ions, e.g., phosphate and nickel (407, 428). Interestingly, one of the PBPs, PhoQ, a gene expressed by Salmonella, is an extracellular Mg²⁺ (Mg_o²⁺) sensor. PhoQ induces bacterial production of Mg²⁺ transport proteins in response to environmental Mg²⁺ deprivation (299).

In their capacities to act as cell surface receptors participating in chemoreception and sensory transduction or membrane transport, the PBPs interact with integral membrane proteins within the bacterial cell membrane after they bind specific chemosensory substances or nutrients to transmit the signal or transport the nutrient into the cell. Therefore, it seems likely that the family C GPCRs, including the CaR, evolved as "fusion proteins" comprising an NH₂-terminal ECD derived from an ancient family of solute-binding proteins and the seven membrane-spanning, "serpentine" motif that evolved separately to transmit extracellular signals to the interior of eukaryotic cells via the GPCRs. Interestingly, the CaR can also participate in the stimulation of chemotaxis by Ca_o²⁺ in monocytes (464) and, perhaps, osteoblasts and their precursors (462, 463). Thus there may be conservation of both the functional as well as the structural attributes of this domain across a very broad evolutionary time scale.

	IV. ARE THERE ADDITIONAL EXTRACELLULAR CALCIUM-SENSING RECEPTORS OR SENSORS?

Top Previous Next References

Recent work has revealed that some mGluRs can sense Ca_o²⁺ in addition to responding to glutamate as their principal agonist in vivo, although the physiological relevance of this Ca_o²⁺-sensing remains uncertain. Kubo et al. (256) demonstrated that mGluRs 1, 3, and 5 sense Ca_o²⁺ over a range of ~0.1-10 mM, while mGluR2 is considerably less responsive to changes in Ca_o²⁺. Construction of chimeric receptors, in which the ECDs of mGluR2 and -3 were fused to the TMDs and COOH tail of mGluR1a, proved that the capacities of the respective receptors to be activated by Ca_o²⁺ (or lack thereof) was conferred by their ECDs. All three of the mGluRs that sense Ca_o²⁺ have identical serines and threonines, respectively, at amino acid positions homologous to residues 165 and 188 in mGluR1a (41). These two residues are thought to play key roles in the binding of glutamate to the ECDs of the mGluRs (330). In contrast, while mGluRs 1a, 3, and 5 have a serine at a position equivalent to residue 166 in mGluR1a, mGluR2 has an aspartate in this position (256). Furthermore, changing the serines in mGluRs 1a, 3, and 5 to an aspartate considerably reduces their capacity to sense Ca_o²⁺, while replacing the aspartate in mGluR2 with a serine increases its apparent affinity for Ca_o²⁺, to a level similar to those of mGluRs 1a, 3, and 5 (256). Therefore, the serines at amino acid position 166 in mGluR1a and at the equivalent positions in mGluRs 3 and 5 apparently play key roles in their capacities to sense Ca_o²⁺, although the molecular mechanism underlying this action is not clear from these studies. It should be pointed out, however, that the Hill coefficients for the modulation of the activities of these mGluRs by Ca_o²⁺ (as well as by glutamate) are close to one (256), considerably lower than that for the CaR, which is ~3 (15, 52, 58, 152, 393). Thus there are apparently additional aspects of the binding of Ca_o²⁺ by the CaR (e.g., the presence of several binding sites) and/or subsequent steps in its activation that confer positive cooperativity on this overall process (see sect. VI). The latter is a key element contributing to the narrow range within which Ca_o²⁺ is maintained by the mineral ion homeostatic system.

Of interest, a recent study has documented that changes in Ca_o²⁺ also modulate the GABA_B receptors, although Ca_o²⁺ by itself has no effect on this class of receptors (459). Ca_o²⁺ potentiated the stimulatory action of GABA on GTP binding to this receptor and enhanced the coupling of the GABA_B receptor to activation of a K⁺ channel and inhibition of forskolin-stimulated cAMP accumulation. The actions of Ca_o²⁺ were not mimicked by other polyvalent cations. Therefore, given that not only Ca_o²⁺, but also amino acids (109; see also sect. IXE), modulate the function of the CaR, when taken in the context of the actions of amino acids (e.g., glutamate) or their derivatives (i.e., GABA) on the mGluRs and GABA_B receptors, respectively, further emphasizes the structural and functional relationships among these three types of receptors.

It is likely that there are Ca_o²⁺ receptors or sensors in addition to the CaR and mGluRs, which mediate some of the very substantial number of actions of Ca_o²⁺ on diverse cell types. The availability of the cloned CaR has made it feasible to begin to catalog the cell types that express this receptor, as described in more detail in sections X, XI, and XIII. For example, the inhibitory action of Ca_o²⁺ on PTH secretion, the stimulatory effect of Ca_o²⁺ on calcitonin (CT) release (146, 158, 288) and many of the actions of Ca_o²⁺ on the kidney (for review, see Ref. 180) are most likely CaR mediated. The presence of the CaR in a cell whose function is modulated by Ca_o²⁺ does not prove, however, that it mediates that particular action of Ca_o²⁺. The availability of mice with targeted disruption of the CaR gene (196) and the discovery of human diseases caused by CaR mutations (for review, see Ref. 53) have provided very useful tools for assessing this receptor's role in Ca_o²⁺-induced changes in various cellular functions in vivo and/or in vitro. The recent development of selective activators (320) and antagonists of the CaR (318) as well as the use of dominant negative CaR constructs (14, 15, 291) will likewise be of utility in determining whether the CaR mediates known actions of Ca_o²⁺ on specific, CaR-expressing cell types.

There are, however, cells whose functions are modulated by Ca_o²⁺ that do not express the CaR or have not yet been examined for its expression. In the former case, the actions of Ca_o²⁺ could potentially be mediated by one or more of the mGluRs that sense Ca_o²⁺, an hypothesis that has not yet been tested. Alternatively, this Ca_o²⁺-sensing capability may be conferred by one or more of the additional putative Ca_o²⁺ receptors/sensors discussed below. There may well be other Ca_o²⁺-receptors/sensing mechanisms as well, although a discussion of the evidence supporting the existence of these is beyond the scope of this review.

Monoclonal antibodies that are directed at a large protein, called megalin or gp330, that is present at high levels in parathyroid, proximal tubular, and placental cells (220) as well as in a variety of other cell types can modulate the Ca_o²⁺-sensing functions of these cells (223). For instance, such antibodies can interfere with the capacity of high Ca_o²⁺ to inhibit PTH secretion from human parathyroid cells (221). The level of expression of this protein is reduced substantially in pathological parathyroid glands from patients having various forms of hyperparathyroidism (HPT) (222). In these hyperparathyroid states, the abnormal cells are generally less sensitive than normal parathyroid cells to the suppressive effect of high Ca_o²⁺ on PTH release (51, 170, 326). Therefore, the reduced expression of the protein recognized by these antibodies could conceivably contribute to the defective Ca_o²⁺ sensing in HPT. Furthermore, the same protein could potentially participate in Ca_o²⁺ sensing by normal parathyroid cells. The level of expression of the CaR in pathological parathyroid cells, however, has also been found to be reduced in HPT in most (136, 162, 246) but not all studies (155), raising the possibility that the changes in megalin expression in hyperparathyroidism could be the consequence rather than the cause of the disease.

cDNAs coding for megalin/gp330 have been isolated from human (195, 273) and rat cDNA libraries (398). These cDNAs encode very large, ~500-kDa proteins that belong to the low-density lipoprotein receptor superfamily. Recent studies have provided strong evidence that megalin's principal role is to serve as an endocytic receptor (116) that binds to and mediates uptake of albumin (115), insulin (333), the transcobalamin-B12 complex (101), retinol and its binding protein (103), and thyroglobulin (279), as well as other proteins (298) and even drugs (137). Indeed, megalin "knockout" mice show defective proximal tubular uptake of the serum vitamin D and retinol binding proteins and their associated vitamin D metabolites and retinol, respectively, providing strong support for the role of this protein as an endocytic receptor (103). Although megalin does bind extracellular calcium ions (102), this Ca_o²⁺ binding probably does not participate directly in systemic mineral ion metabolism. It will be of interest to determine whether megalin interacts with the CaR in cells that coexpress both proteins and/or regulates the CaR's internalization or other aspects of its function, thereby participating indirectly in Ca_o²⁺ sensing.

Raising Ca_o²⁺ has several actions on cells of the osteoblastic lineage. Elevated levels of Ca_o²⁺ stimulate bone formation in explants of rodent bone (372). In addition, Ca_o²⁺ and other polycations [e.g., strontium (73) and aluminum (Al³⁺) (364)] stimulate the proliferation (161, 363, 421) and/or chemotaxis (161) of osteoblasts and their precursors, an effect that could be mediated, in part, by an associated increase in the release of insulin-like growth factor II (IGF-II) (201). High Ca_o²⁺ also modulates intracellular second messengers in the murine osteoblastic cell line, MC3T3-E1. Elevated levels of Ca_o²⁺ raise diacylglycerol (174) and cAMP levels (175) in these cells but do not promote the formation of inositol phosphates that would occur with activation of phosphoinositide (PI)-specific phospholipase (PL) C. Quarles et al. (362) were unable to detect CaR transcripts by RT-PCR and Northern analysis in MC3T3-E1 cells and suggested on the basis of this result as well as pharmacological differences from the CaR [including the latter's low affinity for Al_o³⁺ (417)] that a distinct Ca_o²⁺-sensing receptor mediated the actions of Ca_o²⁺ on this cell line. The same group has identified genomic clones for several CaR-related genes (194) that are ~60% similar and 40% identical to the CaR originally cloned from parathyroid (58) and kidney (381) within regions corresponding to their predicted TMDs. Transcripts for these putative receptors, however, are not expressed in bone cells at levels that can be detected by Northern analysis or RNase protection (194). It is possible, therefore, that they encode either pseudogenes or related receptors, viz., homologs of the putative pheromone receptors in the VNO of the rat (193, 286, 397) but are not involved in sensing Ca_o²⁺ and other polyvalent cations in osteoblasts. Moreover, as discussed in more detail in section XD, we (463) and others (230) have recently found that MC3T3-E1 cells express both CaR transcripts as assessed by RT-PCR and Northern analysis and receptor protein as detected by Western analysis and/or immunocytochemistry. Further studies of these and other osteoblastic cell lines are needed in which the CaR has been "knocked out" through the use of selective CaR activators (320) or antagonists (318) and/or dominant negative constructs of the CaR (291), to prove which, if any, of the effects of Ca_o²⁺ on osteoblastic cell lines are mediated by the CaR versus some other Ca_o²⁺-sensing mechanism(s). Furthermore, while MC3T3-E1 cells and other osteoblast-like cell lines represent useful models for investigating the control of osteoblastic function, they may or may not faithfully reproduce the phenotype of osteoblasts in vivo. It will be important, therefore, to determine whether bona fide osteoblasts and/or their precursor cells in intact bone express the CaR and/or other Ca_o²⁺-sensing mechanisms. Of interest in this regard, the studies of Pi et al. (349) have recently shown that primary osteoblasts derived from mice with targeted disruption of the CaR gene retain certain responses to Ca_o²⁺, consistent with the presence of another Ca_o²⁺-sensing mechanism (349). The presence of the latter and/or the CaR in osteoblasts could enable these cells to respond in physiologically relevant ways to local changes in Ca_o²⁺ within the bone/bone marrow microenvironment (see also sects. XD and XIV for further discussions of local Ca_o²⁺ sensing and Ca_o²⁺ signaling in bone, respectively).

Another example of a cell that appears to possess a Ca_o²⁺-sensing mechanism distinct from the CaR is the osteoclast, based largely on indirect, pharmacological evidence. Several groups first reported in 1989 that elevating Ca_o²⁺ had direct actions on isolated osteoclasts in vitro, inhibiting bone resorption and producing elevations in the cytosolic calcium concentration (Ca_i²⁺), which were reminiscent of those elicited in parathyroid cells by raising Ca_o²⁺ (277, 476). Although it remains to be determined whether this mechanism functions in a physiologically relevant manner in vivo (e.g., by creating mice with targeted disruption of the relevant gene), it could represent a Ca_o²⁺-sensing system through which the osteoclast regulates its own resorptive activity; that is, when Ca_o²⁺ rises above a certain level, owing to osteoclast-mediated bone resorption, activation of the putative Ca_o²⁺-sensing receptor in this cell type would feed back to inhibit further bone breakdown. Subsequent studies, principally by Zaidi et al. (475), have elucidated several features of the process of Ca_o²⁺ sensing by osteoclasts (see below), although characterization of the sensor/receptor at a molecular level has not yet been accomplished.

Elevating Ca_o²⁺ in vitro produces marked retraction of osteoclasts, decreased expression of podosomes (the structures that anchor resorbing osteoclasts to the underlying bone), inhibition of the release of hydrolytic enzymes, and a reduction in bone resorption (277, 476). The observed Ca_o²⁺-induced increases in Ca_i²⁺ are probably an important mediator of the associated alterations in cellular function, because the calcium ionophore ionomycin causes similar effects. Not all osteoclasts possess this Ca_o²⁺-sensing mechanism. Those freshly isolated from medullary bone of the Japanese quail, for example, do not exhibit these responses to elevated levels of Ca_o²⁺ (25). After being cultured for 5-8 days, however, these cells develop the capacity to sense Ca_o²⁺ in a manner similar to that of osteoclasts freshly isolated from chick or rat bone (25). These cultured quail osteoclasts could, therefore, represent an appropriate source of mRNA encoding the putative sensor that could be utilized to isolate the relevant gene using an expression cloning strategy.

A variety of polyvalent cations mimic the actions of Ca_o²⁺ on the osteoclast, but they generally exhibit a pharmacological profile that differs distinctly from that exhibited by parathyroid cells and other cells expressing the CaR (405) [although more recent studies have provided examples of pharmacological profile more similar to that of the CaR, including effects of extracellular Gd³⁺ and neomycin resembling those of high Ca_o²⁺ (474); see also sect. XD]. In general, activation of the CaR in parathyroid cells by Ca_o²⁺, Mg_o²⁺, or extracellular Ba²⁺ takes place at concentrations of these divalent cations that are severalfold lower than those modulating the function of osteoclasts (57, 409, 477). The lower affinity of the Ca_o²⁺-sensing mechanism in osteoclasts for Ca_o²⁺ may be physiologically appropriate, because Ca_o²⁺ measured directly beneath osteoclasts that are actively resorbing bone can be as high as 8-40 mM (412). Other polyvalent cations that activate the osteoclast's Ca_o²⁺-sensing mechanism include extracellular Ni²⁺, extracellular Cd²⁺ (which do not stimulate the CaR) (405), and extracellular La³⁺ (which does activate the CaR) (404).

The putative Ca_o²⁺-sensing receptor in the osteoclast may be related to the ryanodine receptor (479). Agents [e.g., ryanodine (478) or caffeine (406)] that interact with and modulate the activity of the ryanodine receptor (which mediates high Ca_i²⁺-induced release of Ca²⁺ from intracellular stores in skeletal muscle and other cell types) modify osteoclastic Ca_o²⁺ sensing. Moreover, osteoclasts bind [³H]ryanodine, and this binding is displaced by Ca_o²⁺ and by the ryanodine receptor antagonist ruthenium red (479). Finally, an antibody that recognizes an epitope within the ryanodine receptor's channel-forming domain potentiates the effects of extracellular Ni²⁺ on osteoclasts and labels the plasma membrane of nonpermeabilized osteoclasts. Conversely, an antibody that interacts with an intracellular epitope does not exert either of these actions. Taken together, these results suggest the presence of a ryanodine receptor-like molecule on the osteoclast plasma membrane (479) (in contrast to other cell types in which the ryanodine receptor is located intracellularly) that functions as a Ca_o²⁺ sensor or in close association with some other Ca_o²⁺-sensing mechanism. It should be pointed out, however (as described in more detail in sect. XD), that recent studies have suggested that the CaR is also expressed in osteoclasts and/or their precursors. It remains to be determined whether there are actually two distinct Ca_o²⁺-sensing mechanisms in this cell type.

The identification of inherited diseases of Ca_o²⁺ sensing has not only provided strong genetic evidence for the central role of the CaR in systemic Ca_o²⁺ homeostasis but has also raised the possibility that there may be additional Ca_o²⁺ sensors/receptors. Familial hypocalciuric hypercalcemia (FHH) is a generally benign, inherited condition (indeed it is sometimes called FBH, familial benign hypercalcemia) in which there is autosomal dominant inheritance of hypercalcemia accompanied in most cases by relative hypocalciuria (i.e., lower rates of urinary calcium excretion than would have been expected in the setting of hypercalcemia) (for review, see Ref. 49). The great majority of families with this condition (at least 90%) show genetic linkage to the long arm of chromosome 3 in the region where the CaR gene is known to reside (100, 179, 353, 354, 434). Of the families exhibiting this linkage to chromosome 3, about two-thirds have heterozygous inactivating mutations within the coding region of the CaR gene (49). Most of these mutations are point mutations that reduce the receptor's activity by decreasing its cell surface expression and/or reducing its intrinsic biological activity. Some mutations exert an additional dominant negative action on the wild-type CaR (15, 178, 342, 353).

In a few families, consanguineous marriages of individuals with FHH (yielding infants homozygous for CaR inactivation) (213, 353, 354) or union of persons with FHH harboring different CaR mutations (producing a compound heterozygous infant) (249) produces a much more severe form of hypercalcemia, termed neonatal severe hyperparathyroidism (NSHPT) (for review, see Ref. 49). The discovery that FHH and NSHPT can represent, respectively, the equivalent of the heterozygous and homozygous forms of complete or partial knockout of the CaR gene has 1) established the central, nonredundant role of the CaR in mineral ion metabolism, 2) proved that expression of the CaR is required for normal regulation of PTH secretion and probably parathyroid cellular proliferation by Ca_o²⁺ (see also sect. XA), and 3) documented that the CaR plays a key role in regulating the renal tubular handling of divalent cations (see sect. XC for more details).

Of great interest, clinical conditions similar in many of their features to FHH can be caused by genetic defects at chromosomal loci other than that harboring the CaR gene. The first such condition was assigned to a locus on chromosome 19p13.3 by Heath et al. (179) in a family with clinical characteristics indistinguishable from those present in the form of FHH caused by mutations in the CaR. Subsequently, Trump et al. (434) have shown that another family with FHH exhibiting certain atypical features (e.g., osteomalacia and progressive elevations in serum PTH with increasing age in certain family members) exhibits linkage to a different locus on chromosome 19 (19q13) (269), further documenting the genetic heterogeneity of this clinical syndrome. It is possible, therefore, that these two genetic loci contain genes encoding Ca_o²⁺ sensors other than the CaR. Alternatively, these genes might represent additional, presumably downstream elements along the Ca_o²⁺-sensing pathway(s) regulated by the CaR (or some other Ca_o²⁺ sensor) that, when mutated, interfere with the ability of parathyroid and kidney to respond normally to the Ca_o²⁺ signal. It is also conceivable that these genes encode transcription factors or other proteins necessary for expression of the CaR gene in parathyroid and kidney. In the latter case, loss of the relevant transcription factor might reduce the expression of the CaR, analogous to certain forms of diabetes that result from mutations in transcription factors participating in expression of the insulin gene (165).

	V. THE EXTRACELLULAR CALCIUM-SENSING RECEPTOR GENE AND REGULATION OF EXTRACELLULAR CALCIUM-SENSING RECEPTOR EXPRESSION

Top Previous Next References

Very little work has been carried out directed at characterizing the CaR gene. The human gene is located on the long arm of chromosome 3 (3q21-q24) as assessed by linkage analysis (100) and at band 3q13.3-21 as determined by fluorescent in situ hybridization (214). In the rat and mouse, the gene resides on chromosomes 11 and 16 (214), respectively, while in the bovine species it is present on chromosome 1 (314). The human CaR gene contains at least seven exons (343). Six encode the receptor's large ECD and/or its upstream untranslated regions, while a single exon codes for the receptor's TMDs and COOH terminus (343, 353). The regulatory regions of the gene have not yet been characterized but will be of substantial interest, since expression of the CaR can change under several circumstances in vivo and/or in vitro as described below.

Three apparently benign polymorphisms have been identified in the predicted COOH tail of the CaR: A986S, G990R, Q1011E, which were present in 30, 15, and 10% of more than 100 persons in the United States who were apparently unaffected with any disturbance in calcium homeostasis (178). In a subsequent study (106), the A986S polymorphism was found to be present in 16% of 163 Canadian individuals and was associated with a slight increase in serum total calcium concentration when corrected for albumin and in fasting calcium concentration (106). It is possible that this polymorphism could contribute to a genetic predisposition to certain bone and/or mineral disorders.

Several splice variants of the CaR gene have been described. One cDNA clone of the human parathyroid CaR contained a 30-nucleotide insertion within the region of the gene encoding the receptor's predicted ECD (157). This 10-amino acid insertion had no apparent effect on the function of the CaR as assessed by expression in Xenopus laevis oocytes (157). The same splice variant has subsequently been identified in human breast cancer tissue (97). Its functional significance, if any, when studied in mammalian expression systems requires further investigation. Another alternatively spliced CaR transcript that is expressed in human cytotrophoblasts and parathyroid lacks exon 3 and encodes a truncated, presumably inactive receptor (40). Whether this receptor could interfere in some way with the normal CaR's function or exhibits any other functional attribute(s) remains to be determined. Oda et al. (327) have recently described an alternatively spliced form of the CaR in keratinocytes that lacks exon 5, producing an in-frame deletion of 77 amino acids and resulting in an expressed protein that is smaller and exhibits an altered glycosylation pattern compared with the full-length CaR. The truncated CaR was inactive when transfected into HEK293 cells or keratinocytes as assessed by high Ca_o²⁺-evoked increases in inositol phosphates, and it interfered with the function of the coexpressed full-length CaR (327). This latter observation may explain the reduced responsiveness of differentiated keratinocytes to Ca_o²⁺-induced elevations of Ca_i²⁺, as the alternatively spliced form of the receptor is expressed at greater levels in the differentiated cells and could exert, therefore, a dominant negative action on the full-length CaR expressed within the same cells (327).

Finally, there can be alternative splicing within the 5'-untranslated region (UTR) of the CaR gene. For instance, transcripts in human parathyroid vary within their 5'-UTRs, consistent with alternative splicing of noncoding exons within the 5'-upstream region of the gene, without, however, altering the coding region (157). Such alternative splicing within the gene's putative upstream regulatory regions could clearly participate in tissue-specific expression and/or regulation of the CaR gene, but further studies are needed to define further its importance in this regard. Chikatsu et al. (99) have recently cloned a portion of the upstream region of the human CaR gene and identified two promoters (present within, respectively, exons 1A and 1B). The more upstream of the two promoters has TATA and CAAT boxes, while the downstream promoter is GC rich.

D.  Regulation of CaR Expression

1.  CaR expression in cultured parathyroid cells

Recent studies have documented that the expression of the CaR mRNA and/or protein can change in a variety of circumstances, although the mechanisms underlying these alterations in gene expression are, as yet, poorly understood. Calf parathyroid cells show rapid (within hours and 1-2 days, respectively) and marked (up to 80-85%) reductions in CaR mRNA and protein after they are put in culture (47, 297). This reduction in CaR expression probably contributes to a major extent to the accompanying reduction in high Ca_o²⁺-elicited inhibition of PTH release (47, 259, 260, 297). Of interest, the expression of the receptor and the associated suppression of PTH secretion by high Ca_o²⁺ are maintained to a substantially greater extent in long-term cultures of human parathyroid cells, for unclear reasons (391).

The level of expression of the CaR also decreases in the kidney in chronic renal insufficiency induced in rats by subtotal nephrectomy (283, 285). This reduction in CaR expression may contribute to the associated reduction in urinary Ca²⁺ excretion occurring in this setting, based on the inverse relationship between CaR activity and/or expression levels and concomitant renal excretion of Ca²⁺ that is present in persons with inactivating mutations in the CaR (see sect. XC) (181). Since, as described later, 1,25-dihydroxyvitamin D [1,25(OH)₂D] can increase the renal expression of the CaR (46), the decrease in CaR expression in the kidney with impaired renal function could result, at least in part, from the associated reduction in the level of 1,25(OH)₂D that occurs during the development of renal insufficiency (419). Alternatively, the rise in circulating PTH levels in the setting of chronic renal failure (419) may also contribute to the reduction in CaR gene expression in the kidney (285). Further studies are needed to distinguish between these possibilities.

The level of expression of the CaR in parathyroid cells is diminished in pathological parathyroid glands resected from patients with primary hyperparathyroidism or with the severe hyperparathyroidism that can develop during chronic hemodialysis in patients with renal failure due to end-stage renal disease (162, 246). This reduction in CaR mRNA and protein expression has been observed in most (45, 135, 136, 232) but not all studies (155). The study that did not show a decrease in receptor expression employed semi-quantitative RT-PCR to compare the levels of expression of CaR transcripts in normal and pathological parathyroid glands. One potential problem in the interpretation of the results of this study (155) is that normal but not hyperparathyroid parathyroid glands contain substantial numbers of fat cells that can account for ~50% of the volume of the normal parathyroid gland. Studies using immunocytochemistry or in situ hybridization have not detected expression of the CaR in the fat cells of normal parathyroid glands (162, 246). Extraction of RNA from normal parathyroid glands (155), therefore, may have in effect "diluted" CaR transcripts from parathyroid chief cells with RNA not containing these transcripts from fats cells, thereby reducing apparent CaR mRNA expression in normal parathyroid glands to levels that are comparable to those in pathological parathyroid cells. A recent study has shown a selective reduction in parathyroid adenomas of the CaR transcript arising from the further upstream of the two promoters regulating this gene's expression (99). The mechanism underlying this change in the pattern of expression of the CaR gene, however, remains to be determined.

In view of the reduced sensitivity to Ca_o²⁺ of parathyroid glands from patients with inactivating mutations of the CaR gene, the observed reduction in CaR expression in pathological parathyroid tissue could contribute to the elevated set point for Ca_o²⁺-regulated PTH secretion that is often observed not only in primary but also in severe, uremic secondary HPT (51, 62). The relationship of the reduced CaR expression to the associated excessive cellular proliferation in these pathological parathyroid glands remains to be determined. It should be noted, however, that the parathyroid cellular proliferation observed in mice with knockout of the CaR gene (196) as well as in patients homozygous for inactivating mutations of the human CaR gene (for review, see Ref. 49) strongly support the CaR's involvement in tonically suppressing parathyroid cellular proliferation. There is further discussion of the CaR's roles in controlling various aspects of parathyroid function in section XA.

There are interactions between the CaR (or at least the effects of high Ca_o²⁺) and vitamin D receptor (VDR) that are likely physiologically relevant in that the two receptors regulate their own levels of expression and/or the expression of the other receptor. Vitamin D, specifically its active form 1,25(OH)₂D, upregulates its own receptor at the transcriptional level (313). Vitamin D also increases the expression of the CaR in parathyroid and kidney in vivo in the rat (46), although another study, carried out using a slightly different experimental approach, failed to observe any vitamin D-induced changes in the CaR's expression in these two tissues (387). If confirmed in future studies, vitamin D-elicited upregulation of the CaR in the parathyroid could be physiologically appropriate, since it would tend to facilitate inhibition of parathyroid function by high Ca_o²⁺, e.g., suppression of PTH secretion and parathyroid cellular proliferation (see sect. XA).

High Ca_o²⁺ raises the levels of expression of the CaR in the pituitary-derived, ACTH-secreting murine AtT-20 cell line (131) and of the VDR in rat parathyroid glands in vivo (394). Although these actions of Ca_o²⁺ have not yet been proven to be CaR mediated, taken together with the known effects of vitamin D on the expression of the CaR and VDR, they would afford the opportunity for synergistic interactions between Ca_o²⁺ and vitamin D in regulating their target tissues. Such interactions could contribute, for instance, to the known synergistic actions of vitamin D and Ca_o²⁺ in promoting the differentiation of the human colon cancer cell line Caco-2 (113) and in enhancing the expression of calbindin D_28K in the kidney (105). Furthermore, the combination of high Ca_o²⁺ and vitamin D could produce a synergistic inhibition of the expression of the preproPTH gene (413, 467), at least in part, by upregulating the receptors involved in mediating this action as well as through the direct effects of each agent on its own receptor. Further studies are needed, however, aimed at understanding the mechanism(s) (e.g., transcriptional or posttranscriptional) by which high Ca_o²⁺ and vitamin D regulate the expression of the CaR gene. Additional discussion of the regulation of the expression of various genes by high Ca_o²⁺ and the CaR can be found in section VIIID.

Suzuki et al. (425) have recently identified thyroid transcription factor 1 (TTF-1) as a potentially important element in the mechanism(s) through which Ca_o²⁺ may induce changes in gene expression in CaR-expressing cells. TTF-1 is a transcription factor that is a key mediator of thyroid-specific gene expression. It is also expressed in thyroid C cells and in parathyroid chief cells and interacts with elements within the 5'-flanking regions of the CaR, calmodulin, and calcitonin genes (425). Increases or decreases in Ca_i²⁺ enhance or reduce, respectively, the activity of this promoter, its RNA levels, and the binding of TTF-1 to these genes. In CaR-expressing cells that also express TTF-1, in which activation of the receptor is linked to elevations in Ca_i²⁺, therefore, TTF-1 may mediate the regulation of Ca_o²⁺-dependent genes (425).

Whether phosphorus intake modulates CaR expression is controversial. Brown et al. (45) showed that high phosphorus intake was associated with reduced CaR mRNA and protein expression in the parathyroid glands of rats with secondary hyperparathyroidism owing to subtotal nephrectomy. In another study using a similar model, in contrast, there was no change in CaR mRNA expression during intake of a high phosphorus diet (192). In the first of these studies, the reduced CaR immunostaining was limited to regions of active chief cell proliferation, suggesting that the reduction in CaR expression might be secondary to enhanced proliferation rather than the change in phosphorus intake per se (45). It is possible that if high dietary phosphorus induces only focal changes in CaR expression, the latter might be missed by techniques that measure total (i.e., integrated) tissue levels of the receptor (e.g., Northern analysis) (192) as opposed to those that assess localized differences in expression (i.e., in situ hybridization or immunocytochemistry).

There are substantial developmental increases in the expression of the CaR in both kidney (84, 88) and hippocampus of the rat (87). The upregulation of the CaR in the kidney occurs in the immediate peri- and postnatal period, and the ensuing higher level of expression of the receptor persists through adulthood (84, 88). The increase in CaR expression in brain, in contrast, occurs about a week postnatally. Furthermore, it is transient, decreasing severalfold ~2 wk later to a lower level that remains stable into adulthood (87). The mechanisms underlying these alterations in expression of the CaR gene, including the relative importance of alterations in gene transcription versus posttranscriptional mechanisms, require further investigation.

Interleukin (IL)-1 modestly raises the level of CaR mRNA in bovine parathyroid gland fragments in association with a reduction in PTH secretion (324). In preliminary studies, we have also demonstrated that sheep that have undergone experimental burn injury show increased CaR mRNA and protein expression in parathyroid but not in kidney (E. D. Murphey, N. Chattopadhyay, M. Bai, O. Kifor, D. Harper, D. L. Traber, E. M. Brown, and G. L. Klein, unpublished data). Because the levels of inflammatory cytokines are elevated in patients who have suffered a burn, including IL-1 levels, the latter could potentially account, at least in part, for the accompanying rise in CaR expression in the sheep model of burn injury. Furthermore, the increase in CaR protein expression presumably contributes to the associated reduction in PTH secretion and to the failure of PTH levels to rise normally despite hypocalcemia that can occur after burn injury (248).

	VI. STRUCTURE-FUNCTION RELATIONSHIPS OF THE EXTRACELLULAR CALCIUM-SENSING RECEPTOR

Top Previous Next References

CaR proteins extracted from HEK293 cells transiently transfected with the human CaR (15, 134) have similar expression patterns to those isolated from bovine parathyroid cells (15). Western blot analysis using anti-CaR antisera reveals a doublet of protein bands at molecular masses corresponding to ~130-140 and 150-160 kDa that are present in considerably greater amounts than another band at ~115-120 kDa (15, 134). The latter is close to, or slightly smaller than, the expected size of the full-length, nonglycosylated receptor protein predicted from the CaR's cDNA. Direct sequence analysis of the receptor's ECD, in fact, has revealed that the putative signal peptide at the CaR's NH₂ terminus predicted from its nucleotide sequence has been cleaved off (163). Thus the first residue encountered is the tyrosine predicted at amino acid position 20 (163) of the human CaR cDNA (157).

Biochemical analysis of CaR-transfected HEK293 cells using appropriate endoglycosidases has shown that the immunoreactive band at 130-140 kDa corresponds to an immature form(s) of the CaR glycosylated with carbohydrate(s) high in mannose content, while the band at 150-160 kDa is the mature form of the receptor glycosylated with complex carbohydrates (15, 134). The receptor present on the cell surface of these cells is the mature form of the receptor, although the latter probably only represents a relatively small fraction of the total cellular immunoreactivity of the CaR as assessed by Western blot analysis (16). Therefore, much of the CaR is located intracellularly. Indeed, immunocytochemistry using anti-CaR antisera performed on a variety of CaR-expressing cells, when performed following detergent permeabilization (e.g., Triton X-100), demonstrates substantial CaR immunoreactivity over the cytoplasm, often with a prominent perinuclear component (for review, see Ref. 88). It is not currently known whether intracellular forms of the CaR simply represent nascent receptor protein in passage through the biosynthetic pathway or whether these intracellular receptors have distinct biological functions. For instance, the concentration of Ca_o²⁺ within its intracellular stores in the endoplasmic reticulum (ER) (300) approaches the millimolar range and could potentially be sensed by the CaR's NH₂-terminal "extracellular" domain, which would face the lumen of the ER.

In addition to these monomeric forms of the CaR, there are variable amounts of a doublet of immunoreactivity on Western analysis performed using reducing agents at the expected molecular weights of high mannose-containing and fully glycosylated CaR dimers and, to a lesser extent, higher oligomers of the receptor (15, 452). These are not simply artifacts resulting from aggregation of the receptor occurring during its extraction from cells and subsequent PAGE in a denaturing buffer, since these dimers are also present when the receptor is extracted using nondenaturing buffers, and its size is estimated using gel permeation chromatography (452). Furthermore, Bai et al. (16) have shown using a nonpermeant cross-linking reagent combined with cell surface biotinylation that most of the CaR on the cell surface of transiently transfected HEK293 cells is present as a dimer. These CaR dimers appear to be disulfide linked, as the inclusion of reducing agents, such as dithiothreitol or mercaptoethanol, is needed for their conversion to monomers (16, 452). Even following reduction, however, a substantial fraction of the receptor protein can still run on denaturing PAGE as a dimer (16). Therefore, there may be additional intermolecular interactions contributing to dimerization. Interestingly, the CaR has within its fifth TMD a putative hydrophobic dimerization motif present in one of the ₂-adrenergic receptor's TMDs (185). Moreover, the CaR is far from unique among the GPCRs in forming dimers. Recent studies have emphasized that a number of GPCRs (184, 219, 480), including the mGluRs (389), may exist as dimers and that dimerization may potentially play important roles in the function of these receptors.

As noted above, the CaR's mature, cell surface form has a carbohydrate content of ~35-40 kDa/receptor monomer as assessed on Western blots (15, 134). These carbohydrate residues could potentially contribute to the binding of Ca_o²⁺ or participate in other aspects of the receptor's structure and/or function. Indeed, treatment of CaR-transfected HEK293 cells with tunicamycin, which blocks N-linked glycosylation, markedly reduced the response of the cells to raised levels of Ca_o²⁺ in association with reduced cell surface expression of the receptor (134). The use of site-directed mutagenesis subsequently revealed that of the nine predicted N-liked glycosylation sites within the human CaR's ECD, eight are efficiently glycosylated (375). Removal of any four or five of these produced substantial (50-90%) reductions in cell surface expression and biological activity, and at least three intact glycosylation sites were required for efficient cell surface expression. Glycosylation per se, however, did not appear to be critical for the CaR's biological activity as assessed by high Ca_o²⁺-induced increases in inositol phosphate accumulation (375).

The first direct evidence that Ca_o²⁺ binds to the CaR's ECD was provided by studies utilizing chimeric receptors in which the ECD of either the CaR or an mGluR was fused to the TMDs and COOH tail of the other receptor (317). A chimeric receptor containing the CaR's ECD and the TMDs and COOH tail of mGluR1a was activated by high Ca_o²⁺ but not by mGluR agonists. Conversely, a chimeric receptor comprising the mGluR's ECD and the CaR's TMDs and COOH tail was activated by glutamate but not by high Ca_o²⁺. [In these studies, in which the receptors were expressed in X. laevis oocytes, activation of the mGluR by high Ca_o²⁺ was not observed (317); subsequent studies have shown, however, that certain mGluRs can sense Ca_o²⁺, as noted above (256)]. Thus the ECD confers ligand specificity upon the CaR, the mGluRs, and presumably the other members of the family C GPCRs.

Subsequent studies (41) have confirmed that Ca_o²⁺ acts on the CaR by binding to its ECD, taking a similar approach that utilized chimeric receptors and showing that a chimeric receptor comprising the CaR's ECD and mGluR1a's TMDs and COOH tail was activated by Ca_o²⁺, Mg_o²⁺ and Ba_o²⁺ with EC₅₀ values very similar to those of the wild-type CaR. These workers also extended this analysis to define specific residues (e.g., Ser-147 and Ser-170) within the CaR's ECD that may be involved in determining the receptor's apparent affinity for Ca_o²⁺ (41). As described above, these residues are in positions homologous to those of Ser-165 and Thr-188 in mGluR1a. These two serine residues in the CaR and the equivalent residues in mGluR1a and the other mGluRs are thought to play key roles in the binding of these receptors' respective ligands to their ECDs. Moreover, amino acid residues in similar positions within the GABA_B receptors have recently been suggested to play important roles in the binding of GABA. Brauner-Osborne et al. (41) found that mutating Ser-147 to alanine produced a fourfold reduction in the EC₅₀ of the CaR for Ca_o²⁺, whereas a CaR in which Ser-170 was changed to alanine showed no activation by 50 mM Ca_o²⁺. Therefore, it is possible that the binding of Ca_o²⁺ by the CaR involves residues within the ECD that are equivalent to those within the mGluRs and GABA_B receptors that bind glutamate and GABA, respectively.

It should be noted that there are currently no assays of the binding to the CaR of its physiological ligands. Therefore, studies on the determinants within the ECD that are potentially involved in binding these ligands (41, 317) have so far relied on indirect measures of binding, namely, high Ca_o²⁺- and other agonist-evoked increases in Ca_i²⁺ or PLC activity (e.g., as assessed by accumulation of inositol phosphates). Clearly mutating a residue within a GPCR need not modify that receptor's function solely through a direct action (e.g., by interfering with the binding of a ligand to a specific amino acid residue). By altering the receptor's conformation, mutating such a residue could also modulate the protein's function indirectly by secondarily perturbing agonist binding and/or subsequent steps involved in activating intracellular signal transduction [e.g., coupling of the receptor to its respective G protein(s)]. Thus further direct structural studies (e.g., using X-ray crystallography) will be necessary to establish definitively whether the serines at amino acid positions 147 and 170 within the CaR's ECD participate directly or indirectly in Ca_o²⁺ sensing. Moreover, the CaR exhibits a Hill coefficient for its activation by high Ca_o²⁺ of ~3 (15), while the Hill coefficients for activation of the mGluRs by glutamate or by Ca_o²⁺ are ~1 (256). Therefore, it is likely that there are several (probably at least 3 as estimated from the Hill coefficient) binding sites for Ca_o²⁺ within the CaR's ECD and/or elsewhere on the receptor. As discussed in section VID, the fact that the CaR exists on the cell surface principally as a disulfide-linked dimer may contribute to this receptor's apparent positive cooperativity in its binding of Ca_o²⁺.

It is noteworthy that the agonists of the family C receptors, namely calcium ions, glutamate, and GABA, are small molecules (or ions) that bind predominantly, if not entirely, to these receptors' large ECDs. The other major class of GPCRs that bind their agonists to a substantial extent within their ECDs are the receptors for the glycoprotein hormones: thyroid-stimulating hormone, luteinizing hormone, follicle-stimulating hormone, and human chorionic gonadotropin, all of which, in contrast to the agonists for the family C GPCRs, are relatively large heterodimeric glycoproteins (440). In the great majority of the other GPCRs, even those whose agonists are small molecules, such as the biogenic amines, the receptors' agonists have binding sites that likely involve amino acid residues near the extracellular ends of the TMDs, deeper within the TMDs, or in the extracellular loops (454).

The CaR shares with the mGluRs the same relative positions of a total of 20 cysteines: 17 within the ECD, 1 each in the first and second predicted extracellular loops, and 1 in the fifth TMD (58, 157). Clearly these cysteines could be involved in intra- and/or intermolecular disulfide bonds that are important in stabilizing the CaR's tertiary and quaternary structures, e.g., by participating in receptor dimerization. It is of interest in this regard that mGluR5 is a dimer held together by intermolecular disulfide bond(s) present within the first 17 kDa of its NH₂ terminus (389). In this region, the CaR harbors six cysteines, although two of them are within the predicted signal peptide and are likely removed during biosynthesis. Recent studies using site-directed mutagenesis have documented that Cys-129 and Cys-131 are necessary for dimer formation (377), since the CaR migrates on PAGE principally as a monomer when these two cysteines are replaced by serines. Interestingly, the resultant CaR is substantially more sensitive to Ca_o²⁺ than the wild-type receptor, suggesting that these intermolecular disulfide bonds could participate in constraining the receptor in its inactive conformation(s). Although Ray et al. (377) interpreted these findings as showing that the CaR lacking Cys-129 and Cys-131 resides on the cell surface as monomers, we have recently found using immunoprecipitation and cell surface cross-linking that noncovalently bound dimers represent the major form of the cell surface CaR in HEK293 cells transiently transfected HEK293 with a receptor lacking Cys-129 and Cys-131 (M. Bai and E. M. Brown, unpublished data).

As noted above, the cell surface form of the CaR, at least in transiently transfected HEK293 cells, is principally a disulfide-linked dimer. Recent studies have addressed the question of whether dimerization of the CaR is of functional significance; that is, do the two monomeric partners within these dimers bind calcium and/or other CaR ligands and subsequently activate G proteins and intracellular signaling pathways in a largely independent manner or do the monomers interact functionally in some way? Bai et al. (17) showed that individually inactive CaRs containing either inactivating point mutations within their ECDs or sufficient truncation of their COOH tails to abrogate biological activity showed substantial reconstitution of biological activity when cotransfected in HEK293 cells (Fig. 3) (17). Much less reconstitution of activity, in contrast, occurred when two inactive or partially active CaRs were cotransfected that both harbored ECD mutations, or both had defects (e.g., either truncation of the COOH tail or an inactivating point mutation within the third cytoplasmic loop) within domains of the receptor likely involved in intracellular signaling (17). This result suggested that the CaR has at least two functional domains, i.e., the ECD and the remainder of the receptor. These two domains can complement one another in cotransfections involving mutant receptors having defects in different domains. It is also likely that intermolecular interactions between CaR monomers within the dimeric receptor may contribute to the dominant negative action of some CaR mutations identified in FHH, particularly those exhibiting relatively robust levels of cell surface expression and forming substantial quantities of heterodimers of the mutant and wild-type receptors (14, 15). Such dominant negative CaRs have provided a useful tool for determining the CaR's mediatory role in Ca_o²⁺-regulated cellular processes (see sect. VIII) (291).

View larger version (21K): [in this window] [in a new window]   Fig. 3. Cotransfection of inactive mutant CaRs reconstitutes CaR-mediated, extracellular Ca2+ (Cao2+)-elicited cytosolic Ca2+ (Cai2+) signaling in HEK cells. Responses are normalized to the maximal cumulative Cai2+ responses observed with cells transfected with normal (wt) receptor alone for both A and B. A: HEK cells were transfected with either wt or one of the two mutant CaRs, G143E or E297K, either of which had very little activity by itself. B: cells were transfected with the truncation mutant A877Stop or were cotransfected with A877Stop and the full-length wt (wt/A877Stop) or a mutant CaR, either G143E (G143E/A877Stop) or E297K (E297K/A877Stop). Points are the mean values ± SE (n = 3-9). ECD, extracellular domain of a GPCR; TMs, transmembrane domains. [From Bai et al. (17). Copyright 1999 National Academy of Sciences, USA.]

Activators of PKC, such as phorbol 12-myristate 13-acetate (PMA), substantially reduce Ca_o²⁺-evoked increases in inositol phosphates and Ca_i²⁺ in bovine parathyroid cells (243, 293, 326, 369, 408). The presence of predicted PKC phosphorylation sites in the CaR's intracellular domains suggests that PKC could participate in modulating the receptor's function by phosphorylating one or more of these sites (88). Bai et al. (18) recently examined the functional importance of the predicted PKC phosphorylation sites within the human CaR's intracellular domains. The human receptor contains a total of five predicted PKC sites, one each within the second and third ICLs and three within the COOH tail (18, 157). Deleting the two PKC sites within the ICLs had little or no effect on PMA-induced modulation of high Ca_o²⁺-elicited increases in Ca_i²⁺ in HEK cells transiently transfected with the mutant CaRs. Deletion of the PKC site at residue 888 within the CaR's COOH tail, in contrast, substantially reduced the effect of PMA. Individually removing the two other PKC sites within the tail had relatively little impact on the effect of PMA on the CaR's function, but when all three PKC sites within the COOH tail were deleted, there was a modest further reduction in the effect of PKC activators on high Ca_o²⁺-evoked Ca_i²⁺ responses (18). Thus the phosphorylation of the CaR's PKC sites, particularly that present at residue 888 in its COOH tail, can account for much of the inhibitory effect of PKC activators on CaR-mediated signaling through the PLC-inositol trisphosphate pathway. The small (~30%) residual effect of PMA that remains after deletion of all of the CaR's PKC si