Disulfide Bonds in the Extracellular Calcium-Polyvalent Cation-sensing Receptor Correlate with Dimer Formation and Its Response to Divalent Cations in Vitro *

Extracellular calcium/polyvalent cation-sensing receptors (CaR) couple to G proteins and contain highly conserved extracellular cysteine residues. Immunoblotting of proteins from rat kidney inner medullary collecting duct endosomes with CaR-specific antibodies reveals alterations in the apparent molecular mass of CaR depending on protein denaturation conditions. When denatured by SDS under nonreducing conditions, CaR migrates as a putative dimeric species of 240–310 kDa. This is twice the predicted molecular mass of the CaR monomer observed after SDS denaturation in the presence of sulfhydryl-reducing agents. In sucrose density gradients, Triton X-100-solubilized CaR sediments as a 220-kDa complex, not explainable by binding of G proteins to CaR monomers. Treatment of Triton-soluble CaR with divalent (Ca2+, Mg2+) and trivalent (Gd3+) metal ion CaR agonists, but not monovalent ions (Na+), partially shifts the electrophoretic mobility of CaR under reducing conditions from a predominantly monomeric to this putative dimeric species on immunoblots in a manner similar to their rank order of functional potency for CaR activation (Gd3+≫ Ca2+ > Mg2+). This Ca2+effect is blocked by pretreatment withN-ethylmaleimide. We conclude that disulfide bonds present in CaRs mediate formation of dimers that are preserved in Triton X-100 solution. In addition, CaR exposure to Ca2+induces formation of additional disulfide bonds within the Triton-soluble CaR complex.

The binding of divalent (Ca 2ϩ , Mg 2ϩ ), trivalent (Gd 3ϩ ), and polyvalent (neomycin, protamine) cations to the extracellular domain of CaRs 1 initiates a variety of signal transduction cascades via G protein coupling (1)(2)(3). Following their expression in both oocytes (4,5) and cultured human embryonic kidney (HEK) cells (6 -8), CaRs have been characterized pharmacologically. Immunoblotting of these exogenously expressed CaRs present in membrane fractions using CaR-specific antisera re-veals multiple CaR-specific bands (7)(8)(9). To account for these diverse CaR species, it has been suggested that CaRs undergo post-translational modification including glycosylation (8). Moreover, studies using these same antibodies to probe endogenous CaRs present in rat kidney epithelial cells from both the thick ascending limb of Henle (10) as well as from IMCD (11) have revealed multiple CaR species exhibiting similar molecular masses to those reported for exogenously expressed CaRs. However, at present no detailed studies have examined the origin of these multiple CaR species that are present on immunoblots. In this report, we have utilized a combination of immunoblotting, SDS gel permeation chromatography, and sucrose density gradient centrifugation of Triton-solubilized CaR prepared from rat kidney IMCD to characterize CaR associations in both nonionic and ionic detergents. The CaR examined in this study is present in purified endosomes that are derived exclusively from the apical membrane of IMCD epithelial cells (11,12). Utilization of an endogenous CaR present in a defined intracellular compartment rather than a recombinant CaR species expressed in cultured cells at high levels and in multiple intracellular compartments precludes a potential complication of studying multiple CaR species that are actually present in different compartments within cells. The data reported here provide evidence that a major form of CaR is a putative dimeric species that is present following solubilization with both Triton X-100 and SDS detergents and perhaps in the cell membrane itself.
Previous studies have demonstrated that the non-glycosylated CaR polypeptide migrates as a band of approximately 120 kDa on SDS-PAGE immunoblots in close agreement with the putative molecular mass predicted from its corresponding cDNA (13). The presence of glycosylated CaRs possessing an estimated molecular mass of 120 -200 kDa has been demonstrated by conversion of this larger 120 -200-kDa CaR band to a 120-kDa band after digestion with glycosidases (8,13). However, a number of studies (7)(8)(9)(10)(11) have reported the presence of multiple CaR-reactive protein bands greater than 200 kDa on SDS-PAGE immunoblots that are not substantially effected by glycosidase digestion. These data suggest that CaRs may also associate with, or be bound covalently to, other protein(s) including a second CaR molecule. Support for the possibility of a dimeric CaR species is suggested by recent data showing that the structurally related metabotropic glutamate receptor, mGluR5, is a disulfide-linked homodimer in the plasma membrane of cells that exhibits a high molecular weight band on SDS-PAGE immunoblots (14). Putative mGluR5 dimer formation appears to be mediated by sulfhydryl (SH) linkages present in the N-terminal region of the mGluR5 extracellular domain that possesses at least 3 Cys residues that are identical in all CaRs as well as other mGluRs (13,(15)(16)(17)(18)(19)(20)(21).

Materials
Male Sprague-Dawley rats (200 -250 g) were purchased from Charles River Laboratories (Cambridge, MA). Various items were obtained from the following sources: EuCl 3 (Molecular Probes Inc., Eugene, OR); Triton X-100 (Bio-Rad); acrylamide (Amersham Pharmacia Biotech); polyvinylidene difluoride membrane (MSI, Westborough, MA); ECL reagents (Amersham Pharmacia Biotech); and autoradiography film (NEN Life Science Products). All other chemicals were purchased from Sigma. Anti-CaR mouse monoclonal antibody, raised to amino acids 214 -235 of the extracellular domain of the human parathyroid CaR (15), was obtained from NPS Pharmaceuticals (Salt Lake City, UT), and rabbit polyclonal anti-CaR antibody (A4641), raised to amino acids 215-237 of bovine parathyroid CaR, was a gift of Dr. Steven Hebert (Division of Nephrology, Vanderbilt University School of Medicine, Nashville, TN). Other antisera utilized included rabbit anti-Band III antibody, a gift of Dr. Samuel E. Lux (Harvard Medical School, Boston), rabbit anti-G␣ q /G␣ 11 , a gift of A. Tashjian (Harvard University, Boston, MA), and rabbit anti-G␣ i(1-2) and anti-G␣ i(3) antibodies (Upstate Biotechnology Inc., Lake Placid, NY).

Methods
Isolation and Triton X-100 Solubilization of Endosomes-After animals were sacrificed by cervical dislocation under anesthesia (100 mg/kg pentobarbital, intraperitoneally), endosomes were prepared as described previously (12) and utilized either immediately or stored at Ϫ80°C with identical results. Endosomal protein content was determined by the method of Bradford (22). Endosomal proteins (100 -200 g) were solubilized by incubation on ice for 30 min in solubilization buffer: 8.3 mM Tris (pH 7.4), 125 mM NaCl, 1.25 M pepstatin, 4 M leupeptin, 4.8 M phenylmethylsulfonyl fluoride, and 1.0% (v/v) Triton X-100 (final concentrations). The resulting mixture was then centrifuged at 100,000 ϫ g for 30 min to remove Triton X-100-insoluble material, and the supernatant containing Triton-soluble endosomal proteins was collected and aliquoted into equal volumes for the experiments detailed below. To prepare crude membranes, inner medullary papilla homogenate was spun at 2,500 ϫ g (5 min), and the resulting postnuclear supernatant was spun at 100,000 ϫ g (30 min) to give a crude membrane pellet.
Sucrose Density Gradient Ultracentrifugation-Triton X-100-soluble endosomal proteins were layered on top of a 5-20% (w/v) sucrose density gradient (8 ml) in 10 mM Tris (pH 7.4), 150 mM NaCl, and 0.05% (v/v) Triton X-100 and ultracentrifuged in an SW41 rotor at 39,000 rpm for 8 h at 4°C. Individual fractions were then collected from the bottom of the tube and processed for Western blotting. The mass of Triton X-100-soluble CaR was estimated by comparison of its mobility with those of standard proteins including bovine liver catalase, Band III, and hemoglobin as described by Martin and Ames (23) and Clarke (24).
SDS Gel Permeation Chromatography-Samples of SDS-denatured endosomal proteins were applied to a 1.3 ϫ 45-cm Sephacryl 4-B (Amersham Pharmacia Biotech) column equilibrated with 50 mM Tris (pH 8), 0.1% SDS buffer in the absence or presence of 10 mM DTT. One-ml fractions were collected for analyses of both protein amount (A 280 nm ) and content (Western blotting as described below). A single column was utilized for multiple chromatographic runs.
Western Blot Analysis of CaR Protein-Five-fold concentrated Laemmli buffer (0.32 M Tris (pH 6.8), 5% (w/v) sodium dodecyl sulfate, 25% (v/v) glycerol, 1% (w/v) bromphenol blue) was added in a 1:4 ratio to sample proteins that were incubated in the presence or absence of various SH group reducing agents at various temperatures and then fractionated using 5% SDS-polyacrylamide gels (25). Proteins were then transferred electrophoretically to polyvinylidene difluoride membrane in blotting buffer (25 mM Tris, 200 mM glycine, 15% (v/v) methanol) containing 0.5% (w/v) SDS so as to improve the transfer of the proteins Ͼ100 kDa (26,27). The membrane was then incubated in 1% (w/v) bovine serum albumin (20 min) to block nonspecific binding sites, followed by a 1-h incubation in either anti-CaR mouse monoclonal antibody (1:2500 dilution) or anti-CaR rabbit antiserum (1:1000 dilution). After washing to remove all nonspecifically bound anti-CaR antibodies, blots were then exposed to either horseradish peroxidaseconjugated anti-mouse or anti-rabbit secondary antibodies (1:5000) for 1 h. All incubations and intervening washes were performed in Tween/ TBS solution (15 mM Tris (pH 8), 150 mM NaCl, 0.1% (v/v) Tween 20) at room temperature, and blots were developed with the ECL kit as detailed in the manufacturer's instructions. Immunoreactivity was quantified by densitometry and is expressed as percent of total CaR immunoreactivity in the lane (Ϯ S.E.), and statistical significance was determined by paired t test. Fig. 1 shows how various experimental manipulations affect the distribution of CaR-reactive bands in immunoblots after fractionation of IMCD endosomal proteins by SDS-PAGE. As shown in panel A, a CaR-specific mouse monoclonal antibody identifies sequence determinants in a highly acidic region of the extracellular domain present in all CaRs reported to date (3). When purified endosomes are solubilized directly in SDS-PAGE Laemmli buffer in the absence of SH-reducing agents (Fig. 1, panel A, lane 1), greater than 90% of the CaR immunoreactivity is present in a broad band of approximate molecular mass 240 -310 kDa. In contrast, addition of ␤-mercaptoethanol (␤ME) to a final concentration of 143 mM results in both a diminution in the intensity of the 240 -310-kDa band and the appearance of additional CaR bands of 138 -169 and 121 kDa (Fig. 1, panel A, lane 2). These additional CaR-specific bands produced after exposure to ␤ME correspond to molecular masses of the glycosylated and nonglycosylated monomeric CaR proteins reported previously (8,13). Together, these apparent monomeric CaR species represent more than 40% of the total CaR signal within this purified subcellular fraction. Following solubilization in the nonionic detergent Triton X-100, integral membrane proteins such as Band III (28) and bacteriorhodopsin (29) maintain their intra-and intermolecular associations and native conformations (30,31). Triton X-100 solubilization of the CaR protein present in purified IMCD endosomes yielded a pellet of Triton X-100-insoluble proteins ( Fig. 1, panel A, lane 3) and a supernatant of Triton X-100-soluble proteins ( Fig. 1, panel A, lane 4) after centrifugation at 100,000 ϫ g for 30 min. Although both fractions contain a similar distribution of CaR-reactive bands, the Triton-insoluble fraction possesses a 240 -310-kDa band that is somewhat broader and constitutes 69 Ϯ 7% (n ϭ 3) of the total CaR immunoreactivity as compared with its counterpart in the Triton-soluble fraction where it represents only 42 Ϯ 5% (n ϭ 8) of total CaR immunoreactivity. The 121-kDa CaR band was not observed in any of the Triton-insoluble fractions indicating that it may have substantially greater solubility in Triton X-100 than the larger CaR species. To examine whether these alterations in CaR bands in immunoblots were present only after use of this mouse monoclonal antibody, a rabbit polyclonal antibody (panel C) was utilized as a second independent probe. Identical CaR species were obtained with the rabbit antiserum as compared with bands present in mouse monoclonal immunoblots.

Specific Anti-CaR Monoclonal and Polyclonal Antibodies Both Identify Multiple CaR Bands in Immunoblots from the Inner Medulla of Rat Kidney-
Exposure of either native or Triton X-100-solubilized endosomes to 100°C after solubilization in SDS-containing Laemmli buffer with 143 mM ␤ME causes aggregation of CaR as assayed by SDS-PAGE (Fig. 2, panel A). Therefore, CaR samples were routinely immunoblotted without being preheated. As shown in Fig. 2, panel B, SDS denaturation of Triton-soluble endosomal proteins in the absence (lane 3) or presence (lane 4) of 143 mM ␤ME, respectively, yielded CaR immunoreactivity that was mostly present as either a single 240 -310-kDa band (lanes 3 and 7) or as multiple CaR-reactive bands (lanes 4 -6) identical to those shown in Fig. 1, panel A. The binding of the anti-CaR monoclonal antibody to the CaR bands was specific since it was ablated by preincubation of antibody with excess immunizing peptide (Fig. 2, panel B, lanes  1 and 2). Substitution of ␤ME with other SH-reducing agents, including 10 mM dithiothreitol (DTT, lane 5) and 109 mM ␤-mercaptopropanol (lane 6), yielded similar results. In each case, there was only partial conversion of the 240 -310-kDa CaR-reactive band to the smaller CaR-reactive bands. To determine whether oxidation of SH groups to form S-S bonds during SDS denaturation contributes to formation of the 240 -310-kDa CaR band, Triton X-100-soluble endosomal proteins were denatured in SDS-Laemmli buffer containing 1 mM Nethylmaleimide (NEM) that modifies free SH groups but will not reduce S-S bonds (Fig. 2, panel B, lane 7). As shown in panel C, native endosomes were also preincubated on ice for 5 min with 1 mM NEM prior to their denaturation in SDS in the absence of reducing agents (lane 3). Subsequent immunoblotting of both samples revealed that CaR immunoreactivity was again present as a 240 -310-kDa band (Fig. 2, panel C, lane 2) identical to that exhibited by respective controls that were not exposed to any SH-reducing agents (Fig. 2, panel B, lane 3, and Fig. 1, panel A, lane 1, respectively). In addition, exposure of endosomes to chelating agents (1 mM EGTA, 1 mM EDTA on ice for 5 min) prior to SDS solubilization also had no effect on the 240 -310-kDa CaR band. These data suggest the involvement of pre-existing disulfide bonds in the formation of the 240 -310-kDa immunoreactive CaR band and not oxidation of SH groups to form S-S bonds during exposure of CaR to either Triton X-100 or SDS detergents.
To assay whether the apparent CaR dimeric complex may form artifactually from oxidation of SH groups during our tissue processing and endosome purification, we perfused the kidneys of anesthetized rats with 10 mM NEM in situ then removed the kidneys and rapidly prepared a crude membrane fraction from the inner medulla and papilla using homogenization buffer supplemented with 1 mM NEM. These membranes were then solubilized in Laemmli buffer in the absence (Fig. 2, panel D, lanes 1 and 3) or presence (lanes 2 and 4) of 143 mM ␤ME and compared with controls using anti-CaR Western blotting. As shown in panel D of  1 and 3) and Triton X-100-solubilized (TxSol, lanes 2 and 4) endosomes subjected to denaturation in Laemmli buffer containing 143 mM ␤ME where the protein mixture was applied to the gel after a 3-min incubation at either 20°C (lanes 1 and 2) or 100°C (lanes 3 and 4). Panel B, immunoblot of Triton X-100-solubilized endosomal proteins denatured in Laemmli buffer in either the absence of a reducing agent (lanes 1 and 3) or containing either 143 mM ␤ME (lanes 2 and 4) S-S bonds responsible for the 240 -310-kDa CaR complex preexist prior to tissue processing.
Triton X-100-solubilized CaR Exists as a Complex of Approximately 220 kDa-Triton X-100-soluble extracts of endosomal proteins were subjected to ultracentrifugation in sucrose density gradients containing Triton X-100 in the absence of sulfhydryl reducing agents and the sedimentation of CaR compared with that of marker proteins of known native molecular weights. As shown in Fig. 3, panel A, i, CaR immunoreactivity was observed in a single discrete region of the sucrose gradient in close proximity to the catalase (230 kDa) marker, where the peak CaR immunoreactivity corresponded to an estimated molecular mass of approximately 220 kDa (Figs. 1 and 2). As shown in panel A, ii and iii, of Fig. 3, SDS-PAGE analysis of each respective sucrose gradient fraction revealed that each displayed a series of CaR bands including 121, 138 -169, or 240 -310 kDa identical in appearance to those shown in Figs. 1 and 2. Note that the dissociated monomeric CaR species are not equally represented in each gradient fraction (Fig. 3, panel A,  iii). For example, the 121-kDa CaR protein is localized only in fractions 12-14, whereas the 138 -169-kDa CaR protein is enriched in fractions 10 -13.
To determine whether the molecular mass of the detergentsolubilized CaR complex could be accounted for by the presence of a bound G protein as described for the galanin (32), kainate (33), melatonin (34), opioid (35), and pancreastatin (36) receptors, a second identical sample was centrifuged in the presence of GTP␥S in order to dissociate any complex-bound G proteins. As shown in Fig. 3, panel A, i, GTP␥S had no effect on the apparent mass of the Triton-soluble CaR. Consistent with these data were additional results showing that no G␣ q or G␣ i(1-3) immunoreactivity was detectable in the CaR-containing fractions. Instead, G␣ q and G␣ i(1-3) immunoreactivity was present in other fractions nearer the top of the gradient suggesting that they migrate independently of CaRs (data not shown). Fig. 3, panel B, shows Triton X-100-solubilized CaR after density gradient ultracentrifugation in the presence of 10 mM DTT. DTT failed to alter the apparent molecular mass of Triton-soluble CaR as compared with ultracentrifugation under nonreducing conditions. Figs. 1-3, SDS gel permeation chromatography was utilized to demonstrate that DTT alters the apparent size of SDS-denatured CaR. As shown in Fig. 4, fractionation of SDS-solubilized endosomes by gel permeation chromatography shows that the K av of SDS-solubilized CaR without DTT (0.41) is smaller as compared with a paired identical sample denatured and chromatographed in the presence of 10 mM DTT (K av 0.5). These data are in agreement with SDS-PAGE immunoblotting analyses and suggest that SH-reducing agents decrease the molecular mass of SDS-denatured CaR from a value consistent with a CaR dimeric species to that consistent with a monomeric species.

SDS-solubilized CaR Is Decreased in Size upon Reduction-To validate further the data shown in
The simplest interpretation of data displayed in Figs. 1-4 is that CaR protein in rat IMCD endosomes is present as a dimeric species that can only be partially dissociated by SDS denaturation and SH-reducing agents under standard conditions utilized to isolate most membrane preparations. Derivatization of CaR SH groups with NEM to prevent SH oxidation and the formation of disulfide bonds prior to or during SDS denaturation fail to replicate the dissociation of the CaR complex achieved by SH-reducing agents such as ␤-mercaptoethanol or DTT. These data suggest that reduction of pre-existing S-S bonds are primarily responsible for the formation of CaR monomers on immunoblots. However, a minimum of 40% of the CaR immunoreactive protein contained in purified apical membrane endosomes is successfully reduced to CaR monomers, whereas the remainder is resistant to treatment with multiple SH-reducing agents. PAGE-Although the binding of divalent, trivalent, and polyvalent cations to CaRs are known to activate downstream signaling cascades (13,37), only limited data exist investigating the structural alterations that occur in CaRs after ion binding. To examine whether the electrophoretic mobility of CaR is altered by agonist binding, Triton X-100-solubilized CaR was first incubated for 20 min in various concentrations of Ca 2ϩ , Mg 2ϩ , or Na ϩ at 37°C and then analyzed by SDS-PAGE immunoblotting as shown in Figs. 1-3. As shown in Fig. 5, Ca 2ϩ exposure of Triton X-100-solubilized CaR induced a concentration-dependent decrease in the intensity of the monomeric (121 and 138 -169 kDa) CaR bands accompanied by an increase in intensity of the 240 -310-kDa species (20 Ϯ 6% increase in 240 -310-kDa signal induced by 20 mM Ca 2ϩ , as percent of total CaR signal, p Ͻ 0.05, n ϭ 4). A similar effect was also observed following Mg 2ϩ treatment (ϩ16 Ϯ 3% by 20 mM Mg 2ϩ , p Ͻ 0.01, n ϭ 4), but higher (20 -40 mM) Mg 2ϩ concentrations were required to prevent formation of CaR monomers as compared with 10 -20 mM Ca 2ϩ (Fig. 5). These effects appeared specific since exposure to 0 -80 mM Na ϩ produced no significant change in CaR monomer formation (ϩ7 Ϯ 6% by 40 mM Na ϩ , n ϭ 3). These data shown in Fig. 5 are similar to the relative potencies of Ca 2ϩ and Mg 2ϩ obtained from analyses of expressed recombinant CaRs. In a similar manner, exposure to trivalent lanthanides that also activate CaRs, including Gd 3ϩ (200 M) and Eu 3ϩ (200 M), also prevented formation of monomeric CaR species on SDS-PAGE immunoblots (Fig. 5). It should be noted, however, that the appearance of the high molecular weight complex induced by lanthanides is different from that produced by Ca 2ϩ and Mg 2ϩ , since the trailing edge of the lanthanide-induced complex is more diffuse and extends almost into the stacking gel. This is consistent with the possibility that Ca 2ϩ and Gd 3ϩ may actually activate the CaR at different binding sites (8,38).

Calcium-induced Alteration of CaR Electrophoretic Mobility Is Mediated by Disulfide Bonds within the CaR Complex-To
determine whether exposure of CaR to Ca 2ϩ produces a structural alteration in CaR protein that prevents CaR monomer formation upon subsequent SDS denaturation, Triton X-100solubilized CaR was incubated at either 4°C (Fig. 6, lane 1) or 37°C (Fig. 6, lane 3) prior to denaturation in SDS-Laemmli buffer containing 143 mM ␤ME at 20°C. As compared with 4°C, CaR incubation at 37°C consistently resulted in increased CaR immunoreactivity present in the 240 -310-kDa band with a corresponding decrease in monomer immunoreactivity of CaR. NEM pretreatment of the CaR samples exposed to 37°C blocked the shift in CaR electrophoretic mobility to the larger 240 -310-kDa band (Fig. 6, lane 3; p Ͻ 0.01, n ϭ 6) but did not affect the distribution of CaR bands after incubation at 4°C (Fig. 6, lane 2). These data suggest that a 10-min exposure of Triton X-100-solubilized CaR to 37°C results in the formation of disulfide bonds that are not susceptible to reduction during subsequent SDS denaturation in the presence of SHreducing agents. In a similar manner, exposure of Triton-soluble CaR to 20 mM Ca 2ϩ at 4°C produced a shift in CaR immunoreactivity identical to that produced by incubation at 37°C that was also significantly inhibited by pre-exposure to 1 mM NEM (Fig. 6, lane 5; p Ͻ 0.05, n ϭ 3). The shifts produced by either calcium or exposure to 37°C are not additive, and their combination was also significantly attenuated by NEM pretreatment (Fig. 6, lane 8; p Ͻ 0.001, n ϭ 6). These data suggest that exposure to Ca 2ϩ induces an alteration in Triton X-100solubilized CaR at 4°C that is similar to that produced by exposure of CaR alone to 37°C. Both treatments alter the ability of the combination of SDS and SH-reducing agents to fully denature CaRs to monomeric species. DISCUSSION These data reported here confirm and greatly extend earlier reports (8,13) showing that the electrophoretic profile of CaR immunoreactive bands is more complex than simply nonglycosylated and glycosylated monomeric CaRs present on immunoblots prepared from CaR-containing protein mixtures. Whereas the presence of a high molecular weight CaR-immunoreactive band on immunoblots has been previously reported (2,(7)(8)(9)(10), it has been unclear as to whether its origin represented a real biochemical association or was simply an artifact of denaturation in SDS-containing detergent solutions. Data presented in Figs. 1 and 2 show that SDS denaturation of rat kidney IMCD CaR in the absence of SH-reducing agents results in the formation of a single 240 -310-kDa CaR band, as determined using both monoclonal and polyclonal anti-CaR antibodies. The specificity of this polyclonal anti-CaR antiserum has been demonstrated previously where no bands were observed in immunoblots with membranes prepared from mock-transfected HEK cells, whereas membranes from HEK cells transfected with human CaR and bovine parathyroid tissue possessed CaR bands identical to those reported here (8).
Addition of SH-reducing agents to SDS-denatured CaR reduces its apparent size as measured by gel permeation chromatography (Fig. 4) as well as diminishes the intensity of the 240 -310-kDa band with a concomitant appearance of CaRreactive bands of 121 and 138 -169 kDa that correspond to non-glycosylated and glycosylated monomeric CaRs, respectively ( Figs. 1 and 2). Perfusion of intact kidney tissue with NEM to prevent any SH oxidation followed by immediate processing of membranes in NEM fails to prevent isolation of a CaR dimeric species. However, denaturation of CaR in SDS in the presence of SH-reducing agents is capable of only partial conversion of the total CaR immunoreactivity to CaR monomeric species. As shown in Figs. 1 and 2, a significant proportion (0 -60%) of immunoreactive CaR protein remains as a discrete 240 -310-kDa band. At present, the exact nature of this large SDS-resistant CaR is unknown. However, data shown in Fig. 2 suggest that this SDS-resistant CaR complex may possess S-S linkages that resist the combination of SDS and reducing agents, and its formation is increased by both exposure to elevated temperature (Fig. 2, panel A; Fig. 5) and prolonged exposure of membrane preparations to buffers that do not contain SH-reducing agents (Fig. 2, panel D). These data do not permit us to distinguish whether intramolecular or intermolecular S-S bonds contribute to formation of this SDS-resistant CaR complex.
To study the interactions of CaR with itself or other proteins, CaR was solubilized in the nonionic detergent Triton X-100 and analyzed by a combination of sucrose gradient sedimentation and SDS-PAGE immunoblotting. As shown in panel A of Fig. 1, 58 Ϯ 5% of Triton X-100-solubilized CaR was converted to CaR monomers upon denaturation in SDS-reducing agents, whereas the Triton X-100-insoluble CaR fraction was correspondingly enriched for the larger 240 -310-kDa CaR species. As shown in Fig. 3, Triton X-100-soluble CaR sediments as a complex of approximately 220 kDa in the presence or absence of SH-reducing agents as determined in sucrose gradients. Previous analyses of Triton X-100-solubilized membrane proteins such as Band 3, the major anion exchanger of the human red cell, using identical sucrose gradient sedimentation techniques have demonstrated that Triton X-100 solubilization does not disrupt intermolecular protein-protein associations between Band 3 dimers or interfere with the binding of other cytoskeletal and cytoplasmic proteins to Band 3 (24). These data dis-TABLE I Conservation of extracellular cysteine residues between CaRs, mGluRs, and VRs Conservation of extracellular cysteine residues, including those in the extracellular loops of the transmembrane domain, present in CaRs from bovine parathyroid (BoP), rat kidney (RaK), human parathyroid (HuP), rat brain (RaB), and human kidney (HuK), and in mGluRs 1-8 and the mouse vomeronasal odorant receptors (mV2Rs). Conservation (ϩ) or non-conservation (o) is determined by comparison to BoPCaR, the position of whose cysteine residues are given, and additional cysteine residues present in each molecule but not present in BoPCaR are listed in the right-hand column (13,(15)(16)(17)(18)(19)(20)(21)39). played in Fig. 3 show the presence of a 220-kDa high molecular CaR complex after Triton X-100 solubilization suggesting that CaR complex formation does not arise via artifactual denaturation by SDS. Furthermore, it seems unlikely that the larger molecular mass of the CaR complex is the result of associations with heterotrimeric G proteins as has been reported for 6 other G protein-coupled receptors (33)(34)(35)(36)(37)41) since its sedimentation is not altered by preincubation with GTP␥S (Fig. 3, panel A). Instead, the IMCD CaR appears to correspond more closely to the muscarinic acetylcholine (42), neurotensin (43), and gastrin-releasing peptide (44) receptors that do not possess associated G proteins after nonionic detergent solubilization.
As described under "Results," careful analyses of the monomeric CaR species shown in Fig. 3, panel A, iii, reveal that the Triton X-100-soluble CaR complex present in individual fractions of the sucrose gradient contains a nonrandom mixture of glycosylated and non-glycosylated CaR monomers. The simplest interpretation of these data is that the Triton X-100solubilized CaR complex consists of CaR dimers composed of various combinations of glycosylated and glycosylated-nonglycosylated pairs. These various types of CaR dimers observed here are only partially dissociated by subsequent denaturation in SDS and exposure to SH-reducing agents. However, we cannot eliminate entirely the possibility that other proteins might also be part of the CaR complex present after solubilization in either Triton X-100 or SDS.
Romano et al. (14) have used similar techniques to suggest that the metabotropic glutamate receptors (mGluRs), which are structurally related to CaRs, exist in the membrane as disulfide-linked dimers and that the cysteine residues responsible for the homodimerization of mGluR5 reside within 17 kDa of the N terminus of the extracellular domain. Of note, this mGluR5 region contains 4 Cys residues (19), 3 of which are conserved between all of the mGluRs (19 -21) and all CaRs so far reported (13,(15)(16)(17)(18). Indeed, Table I shows that the conservation of extracellular cysteine residues within the extracellular domains of the various known members of this receptor family is very high despite varying degrees of overall homology. A previous report (45) using immunoblotting analyses of both cells transfected with various mGluR species as well as specific regions of rat brain shows the formation of specific bands of approximately twice the apparent molecular mass of monomeric mGluRs after SDS denaturation in the presence of SHreducing agents. These data suggest that CaRs form dimers or complexes via their extracellular domains. A truncated CaR lacking all of the intracellular domain and 4 of the 7 transmembrane domains still exhibits dimer-like high molecular weight immunoreactivity on CaR immunoblots (9). Taken together, these data reported here and previous reports (14) suggest that an ability to form dimeric CaR species may be a characteristic common to all members of this G protein-coupled receptor subfamily.
The data shown here suggest that both the IMCD CaR and mGluR5 (14) are distinct from the receptor tyrosine kinases that undergo dimerization principally upon agonist binding. However, addition of Ca 2ϩ to Triton X-100-solubilized CaR significantly reduced the ability of SDS and SH-reducing agents to generate monomeric CaR species, raising the possibility that Triton-soluble CaR may undergo structural modification in response to metal ion agonist treatment. The effect of Ca 2ϩ was mimicked by other metal ion CaR agonists in a rank order of potency equivalent to that determined in functional assays (Gd 3ϩ Ͼ Ͼ Ca 2ϩ Ͼ Mg 2ϩ ). These data, shown in Figs. 5 and 6, suggest that binding of divalent and trivalent cations to the CaR molecule induces a conformational change in the CaR complex that promotes oxidation of free sulfhydryl groups re-sulting in a Ca 2ϩ -mediated reduction in CaR monomer formation upon subsequent exposure to SDS and SH-reducing agents. Although metal ions such as Ca 2ϩ and Zn 2ϩ can themselves promote oxidation of free sulfhydryl residues, the data shown in Figs. 5 and 6 suggest these effects are not simply nonspecific, since Ca 2ϩ and Mg 2ϩ exposure causes a shift in CaR electrophoretic mobility between discrete, monomeric, and putative dimeric bands, as opposed to producing a broad smear of CaR immunoreactivity along the lane which could indicate random aggregation. Thus, the current data are most consistent with the concept that putative dimeric CaR species become stabilized upon addition of Ca 2ϩ in vitro. Further work is required to determine what proportion of CaRs actually exist as either monomers or dimers in the membranes of intact cells and whether an equilibrium between the two states has a functional significance similar to that recently established for the G protein-coupled ␦ opioid receptor (46).
The importance of extracellular domain disulfide bonds on agonist-induced G protein-coupled receptor activation has been investigated recently using site-directed mutagenesis studies of both the thyrotropin-releasing hormone receptor (47,48) and gonadotropin-releasing hormone receptor (49). To date, most studies of the CaR have investigated the expression of various normal and mutant recombinant CaRs (obtained from naturally occurring mutations and by site-directed mutagenesis) in oocytes (4,5,13) and HEK cells (6 -9). The studies reported here have suggested an important role for sulfhydryl groups in formation of a putative CaR dimeric species and provide a series of testable hypotheses that should be the object of future studies of CaR using site-directed mutagenesis.