Cys-140 Is Critical for Metabotropic Glutamate Receptor-1 Dimerization*

Metabotropic glutamate receptor 1 (mGluR1) expresses at the cell surface as disulfide-linked dimers and can be reduced to monomers with sulfhydryl reagents. To identify the dimerization domain, we transiently expressed in HEK-293 cells a truncated version of mGluR1 (RhodC-R1) devoid of the extracellular domain (ECD). RhodC-R1 was a monomer in the absence or presence of the reducing agents, suggesting that dimerization occurs via the ECD. To identify cysteine residues involved in dimerization within the ECD, cysteine to serine point mutations were made at three cysteines within the amino-terminal half of the ECD. A mutation at positions Cys-67, Cys-109, and Cys-140 all resulted in significant amounts of monomers in the absence of reducing agents. The monomeric C67S and C109S mutants were not properly glycosylated, failed to reach the cell surface, and showed no glutamate response, indicating that these mutant receptors were improperly folded and/or processed and thus retained intracellularly. In contrast, the monomeric C140S mutant was properly glycosylated, processed, and expressed at the cell surface. Phosphoinositide hydrolysis assay showed that the glutamate response of the C140S mutant receptor was similar to the wild type receptor. Substitution of a cysteine for Ser-129, Lys-134, Asp-143, and Thr-146 on the C140S mutant background restored receptor dimerization. Taken together, the results suggest that Cys-140 contributes to intermolecular disulfide-linked dimerization of mGluR1.

tors (T1Rs) (7), and GABA B receptors (8). The eight members of the mGluR family (mGluR1-8) of receptors have been divided into three subgroups based on sequence homology, signal transduction properties, and pharmacological profile (1). The mGluR1 and mGluR5 receptors and their splice variants make up the group 1 mGluRs, which are coupled to the stimulation of phosphoinositol turnover via the Gq subfamily of G-proteins.
The mGluRs and other family 3 GPCRs are characterized by a very large (approximately ϳ600 residues) extracellular amino-terminal domain (ECD). The mGluR1 ECD is the glutamate binding domain and believed to be structurally related to the bilobed "venus flytrap" structure of bacterial periplasmic binding proteins (9,10). Recently, it has been shown that both the mGluR1 (11) and the Ca 2ϩ receptor (12,13) are expressed at the cell surface as intermolecular disulfide-linked dimers. For mGluR1 (10), mGluR4 (14), and Ca 2ϩ receptor (15), the ECD of each receptor, purified as a secreted protein, exists as a disulfide-linked dimer, suggesting that one or more cysteines in the ECD is involved in receptor dimer formation. The rat mGluR subtype 1␣ (described as mGluR1 from now on) ECD contains a total of 19 cysteines (16) of which several are highly conserved in other mGluRs, Ca 2ϩ receptor, V2Rs, and T1Rs. Proteolysis of the mGluR5 receptor localized cysteine(s) critical for dimer formation to the first 17 kDa of the ECD (17). This region contains three cysteines conserved in all mGluRs, and we mutated these cysteines of the mGluR1 and investigated the role of these cysteine mutants in receptor dimerization and function. This study led to the identification of a conserved Cys-140 as critical for disulfide-linked dimerization of mGluR1, and functional study indicated that, like the wild type receptor, C140S mutant receptor is also capable of intracellular signaling via Gq-phospholipase C pathway.

Construction of a Rhodopsin Epitope-tagged Truncation Mutant of mGluR1-Using
Turbo Pfu DNA polymerase (Stratagene Inc.), a polymerase chain reaction (PCR) was performed to add 20 amino acid residues (MNGTEGPNFYVPFSNKTGVV) corresponding to the amino terminus of bovine rhodopsin to the mGluR1 before amino acid residue 584. The 1.8-kilobase PCR product was subcloned to the pCR3.1 expression vector (Invitrogen) as a HindIII-XhoI fragment. The entire nucleotide sequence of the PCR product was confirmed by using a dRhodamine terminator cycle sequencing reaction kit and ABI prizm-377 DNA sequencer (Applied Biosystems). The constructed truncation mutant, designated as RhodC-R1, was devoid of 1-583 amino-terminal ECD of the mGluR1 but included 20 amino acids of the amino terminus rhodopsin tag and the amino acid residues 584 -1199 of mGluR1.
Site-directed Mutagenesis of the mGluR1-mGluR1 cDNA was cloned in pCR3.1 expression vector (Invitrogen) as a BamHI-NotI fragment. Site-directed mutagenesis was performed using a commercial kit (QuikChange site-directed mutagenesis kit, Stratagene) as described by Ray et al. (18). Three cysteine point mutants, C67S, C107S, and C140S were created by changing cysteine at a given site to serine. Seven Cys-scanning mutants, S119C, S126C, S129C, K134C, D143C, T146C, and T152C were created in a second round of mutagenesis by using C140S mutant plasmid DNA as a template. The mutations were con-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
firmed by automated DNA sequencing using a dRhodamine terminator cycle sequencing reaction kit and the ABI prizm-377 DNA sequencer (Applied Biosystems).
Transient Transfection of Wild Type and Mutant Receptors in HEK-293 Cells-Receptor plasmid DNAs were prepared by a maxi-plasmid preparation kit (Qiagen) and were transiently transfection in HEK-293 cells using LipofectAMINE (Life Technologies) as described previously (19). Protein extraction for immunoblotting or biotinylation-immunoprecipitation experiments was performed 48 h after transfection.
Immunoblotting Analyses with Detergent-solubilized Crude or Whole Cell Extracts-Crude membrane extracts were prepared as described previously by Ray et al. (19). Briefly, confluent cells in 75-cm 2 flasks or 6-well plates were rinsed with ice-cold phosphate-buffered saline and scraped on ice in buffer A (5 mM Tris (pH 7.2), 2 mM EDTA), containing 10 mM iodoacetamide with freshly added Complete protease inhibitors mixture (Roche Molecular Biochemicals). The cells were forced through a 22-gauge needle five to eight times, and the lysate was spun in a TLA-45 centrifuge at 45,000 rpm for 30 min at 4°C to collect a crude membrane pellet. The pellet was resuspended in buffer B containing 20 mM Tris-HCl (pH 6.8), 150 mM NaCl, 10 mM EDTA, 1 mM EGTA, 1% Triton X-100 with 10 mM iodoacetamide and freshly added protease inhibitors mixture. Whole cell extracts were prepared by solubilizing cells directly in buffer B containing 1% Triton-100 as described earlier (13). The protein content of each sample was determined by the modified Bradford method (Bio-Rad) and 20 -30 g of protein per lane was separated on 5% gel by SDS-PAGE. Electrotransferred proteins to nitrocellulose membranes were incubated with monoclonal anti-mGluR1 antibody (Transduction laboratories, catalog # M72620) at a dilution of 1: 10,000 or anti-mGluR1 polyclonal antibody at a dilution of 1:500 (Chemicon, catalog # AB1553). Subsequently, the nitrocellulose membrane was incubated with a secondary goat anti-mouse or antirabbit antibody conjugated to horseradish peroxidase (Kierkegaard and Perry Laboratories) at a dilution of 1: 5000, respectively. The mGluR1 bands were detected with an Enhanced Chemiluminescence system (ECL) (Amersham Corp). Biotinylated protein bands were detected using peroxidase-conjugated streptavidin-POD followed by visualization of the biotinylated bands using a BM chemiluminescence kit (Roche Molecular Biochemicals).
Biotin-labeling of the Cell Surface mGluR1-48 h after transfection, cell surface proteins of the intact HEK-293 cells were labeled with membrane-impermeant Biotin-7-NHS using the cellular labeling kit (Roche Molecular Biochemicals) as described earlier (19). Briefly, intact cells were labeled with 50 g/ml Biotin-7-NHS in biotinylation buffer (50 mM sodium borate, 150 mM NaCl) to biotinylate cell surface proteins. The reaction was stopped by adding 50 mM NH 4 Cl for 15 min on ice. The cells were washed with phosphate-buffered saline and solubilized with lysis buffer B.
Immunoprecipitation of mGluR1-300 l (approximately 600 g of total protein) of the whole cell lysate of biotin-labeled cells was further diluted with 300 l of buffer B and incubated with 5 l of mouse monoclonal mGluR1-specific (made against the carboxyl-terminal peptide corresponding to amino acids 1042-1160; 0.1 mg/ml stock; Transduction laboratories, Lexington, KY) for 1-2 h at 4°C. Subsequently, 25 l of Protein A/G-agarose (Santa Cruz Biotechnologies) was added, and the incubation was continued for an additional 2 h. The Protein A/Gagarose was washed three times with buffer B containing 0.5% SDS, and the immunoreactive proteins were eluted in sample buffer containing either no ␤-mercaptoethanol or 300 mM ␤-mercaptoethanol. Samples were analyzed by SDS-PAGE, and immunoblotting was performed as described before.
Glycosidase Treatment of Detergent-solubilized Extracts-For cleavage with PNGase-F or Endo-H (Roche Molecular Biochemicals), whole cell extracts (30 l) were diluted in 20 l of 50 mM sodium acetate (pH 4.8). Samples were incubated with 0.5 milliunits of Endo-H or 1.0 unit of PNGase-F for 2 h at 37°C.
Phosphoinositide Hydrolysis Assay-Phosphoinositide (PhI) hydrolysis assay has been described previously (18). Briefly, 24 h after transfection, transfected cells from a confluent 75-cm 2 flask was replated in 24-well plates in medium containing 3.0 Ci/ml of [ 3 H]myoinositol (PerkinElmer Life Sciences) in complete Dulbecco's modified Eagle's medium containing no glutamine for another 24 h, followed by 1-h preincubation with 1ϫ PhI buffer (120 mM NaCl, 0.5 mM CaCl 2 , 5 mM KCl, 5.6 mM glucose, 0.4 mM MgCl 2 , 20 mM LiCl in 25 mM PIPES buffer, pH 7.2). After removal of PhI buffer, cells were incubated for an additional 30 min with different concentrations of L-glutamate in PhI buffer. The reactions were terminated by addition of 1 ml of acid-methanol (1:1000, v/v) per well. Total inositol phosphates were purified by chromatography on Dowex 1-X8 columns.

mGluR1 Is a Cell Surface-expressed Disulfide-linked
Dimer-To determine the domain(s) involved in mGluR1 dimerization, we sought to construct a truncated form of the mGluR1 lacking the ECD. The amino terminus of this truncated mutant (RhodC-R1) consisted of the first 20 amino acid residues of rhodopsin followed by the remainder of the mGluR1 beginning with residue 584. The rhodopsin N terminus tag was added, because it has been shown to enhance proper processing and cell surface expression of several GPCRs (20,21). Next, the wild type mGluR1 and RhodC-R1 cDNAs were transiently transfected in HEK-293 cells, and after 48 h, whole cell extracts were prepared. To prevent non-specific disulfide bond formation during protein extraction, 10 mM iodoacetamide was included in the lysis buffer. On an immunoblot, as shown in Fig.  1A (left), under non-reducing conditions, a mGluR1-specific monoclonal antibody made against a carboxyl-terminal epitope detected two major bands of the wild type mGluR1 around ϳ260to 270-kDa molecular mass range and two fainter 130and 135-kDa bands. The same four immunoreactive bands were also recognized by a mGluR1-specific polyclonal antibody ( Fig. 1B) and not detected in vector-transfected cells (data not shown). The intensity of the lower two 130-and 135-kDa bands varied from no immunoreactivity to faint immunoreactivity between different immunoblotting experiments. After reduction with ␤-mercaptoethanol, a majority of the wild type mGluR1 ϳ260to 270-kDa bands were reduced to a broad monomeric ϳ133-kDa band, and a small portion remained as SDS-resistant dimeric aggregates. The RhodC-R1 mutant expression pattern revealed a major ϳ90-kDa immunoreactive band, and the protein was expressed as a monomer, because the presence or absence of ␤-mercaptoethanol did not shift this band (Fig. 1A, right).
To further determine whether the dimeric mGluR1 and monomeric RhodC-R1 forms were expressed at the cell surface, cell surface proteins were labeled with membrane impermeant Biotin-7-NHS prior to lysing the cells. The wild type mGluR1 and the RhodC-R1 were then immunoprecipitated with anti-mGluR1 monoclonal antibody and eluted with gel loading sample buffer containing no ␤-mercaptoethanol. Immunoprecipitates were analyzed by SDS-PAGE, and immunoblots were stained either with streptavidin-POD to detect biotinylated cell surface proteins or with a mGluR1 polyclonal antibody to detect all the immunoreactive bands. As shown in Fig. 1B, under non-reducing conditions, streptavidin-POD detected the upper ϳ260to 270-kDa dimeric bands but no visible 130-or 135-kDa monomeric bands of the mGluR1. Similarly, streptavidin-POD detected the 90-kDa band of the RhodC-R1 mutant. A duplicate blot of these samples with anti-mGluR1 polyclonal antibody detected all the immunoreactive bands of the wild type mGluR1 and RhodC-R1. mGluR1 ECD contains four potential asparagine-linked glycosylation sites (Asn-Xaa-Ser/Thr) and is shown to undergo glycosylation (11), and the RhodC-R1 mutant contains two potential asparagine-linked glycosylation sites within the rhodopsin tag sequence. Therefore, to determine further the biochemical identity of the ϳ260to 270-kDa dimeric bands of the wild type and the 90-kDa RhodC-R1 mutant receptors, we conducted deglycosylation experiments with two glycosidase enzymes, PNGase-F and Endo-H. The PNGase-F enzyme cleaves all asparagine-linked carbohydrates (both intermediate high mannose forms and fully processed complex carbohydrate forms) from glycoproteins (22). Sensitivity to Endo-H digestion distinguishes between the fully processed mGluR1 forms that are modified with complex carbohydrates (Endo-H-resistant) and intermediate high mannose modified forms (Endo-H-sen-sitive), which have not trafficked from the endoplasmic reticulum to the Golgi (13,18,19). As shown in Fig. 1C, under non-reducing conditions, whole cell extracts prepared from mGluR1-and RhodC-R1-transfected cells digested with PN-Gase-F showed a decrease in the size of both the ϳ260to 270-kDa bands of the wild mGluR1 and the 90-kDa band of the RhodC-R1 mutant receptors. However, these bands remained mostly resistant to Endo-H digestion. These data suggest that the dimeric wild type mGluR1 and monomeric RhodC-R1 receptors are expressed at the cell surface as asparagine-linked glycosylated and mostly fully processed receptor forms.
Screening of Single Cysteine Mutants for Their Ability to Form Homodimers-We generated single C 3 S mutants of three cysteines at positions 67, 109, and 140 in the mGluR1 ECD. The C 3 S mutants were further analyzed by determining their dimerization patterns on immunoblots ran under non-reducing conditions. Transiently transfected cells expressing these mutant receptors were treated with iodoacetamide to prevent aggregates forming secondary to non-specific disulfide bond formation prior to crude membrane preparation. The membrane extracts were run on immunoblots, and immunoreactive bands were detected with an anti-mGluR1 monoclonal antibody. As seen in Fig. 2A, the wild type mGluR1 receptor expressed as two dimeric bands with little or no monomeric forms visible on immunoblot. C67S and C109S mutant receptors expressed predominantly as two 130-and 135-kDa monomeric forms, and one or two fainter upper dimeric band corresponding to the upper ϳ260to 270-kDa dimeric bands of the wild type mGluR1 were sometimes visible. These fainter dimeric bands were more visible when samples ran on gels by SDS-PAGE were enriched as crude membrane extracts or as immunoprecipitated samples but not routinely seen in whole cell extracts prepared from C67S-and C109S-transfected cells.
Interestingly, C140S mutant receptor expressed as a broad ϳ140-kDa monomeric band that ran at slightly higher molecular mass than the 130-and 135-kDa monomeric bands of C67S and C109S mutants and also generated a ϳ300-kDa dimeric band. This dimeric band of the C140S mutant ran higher than the dimeric bands of the wild type mGluR1 or C67S or C109S mutant receptors. The intensities of the immunore- active dimeric bands relative to the monomeric bands for each mutant receptor and wild type receptor were then measured by densitometric scanning (Table I). The data showed that C67S, C109S, and C140S mutant receptors expressed mostly as monomeric forms and generated a small amount of dimeric forms. In contrast, the wild type receptor expressed predominantly as dimers and generated a small amount of monomers. After reduction (Fig. 2B), a majority of the dimeric ϳ260to 270-kDa bands of the wild type receptor were reduced to a broad monomeric ϳ133-kDa band and a small portion remained as SDSresistant dimeric aggregates, whereas the dimeric bands generated by C67S, C109S, and C140S mutant receptors remained mostly as SDS-resistant dimeric aggregates. Also, the 140-kDa monomeric band of the C140S mutant showed a small shift of molecular mass to a 133-kDa band like the wild type receptor. Taken together, the results suggest that substituting serine for cysteines 67, 109, and 140 directly or indirectly blocks dimerization.
C 3 S Mutation of Cys-140 Generates a Monomeric Form of mGluR1 Expressed at the Cell Surface-Because point mutations of three cysteine residues in mGluR1 generated monomeric receptors, we wanted to distinguish between properly processed cell surface versus intracellularly trapped monomeric forms. Cell surface proteins of the HEK-293 cells transiently transfected with wild type, C67S, C109S, and C140S mutant receptors were labeled with membrane impermeant Biotin-7-NHS. The whole cell extracts were immunoprecipitated with anti-mGluR1 monoclonal antibody, ran under nonreduced condition, and analyzed on immunoblots stained either with streptavidin-POD to detect biotinylated cell surface proteins or with anti-mGluR1 polyclonal antibody to detect total mGluR1 immunoreactive species. As seen in Fig. 3A, streptavidin-POD detected only the dimeric ϳ260to 270-kDa bands of the wild type mGluR1 and a fainter dimeric band of C67S and C109S mutants but no monomeric 130-to 135-kDa bands of these receptors. In contrast, streptavidin detected both the dimeric 300-kDa and monomeric 140-kDa bands of the C140S mutant receptor. A duplicate blot of the samples with anti-mGluR1 polyclonal antibody detected both the cell surface and the intracellular forms of the wild type and mutant receptors (Fig. 3A, anti-mGluR1 blot). To further determine the biochemical identity of the expressed bands of the C67S, C109S, and C140S mutant receptors, we tested for sensitivity to Endo-H digestion. Whole cell extracts prepared from HEK-293 cells transiently transfected with wild type, C67S, C109S, and C140S cDNAs were digested with Endo-H and analyzed by SDS-PAGE under non-reducing conditions. As seen in Fig. 3B, digestion with Endo-H caused no decrease in the sizes of the dimeric ϳ260to 270-kDa bands of the wild type mGluR1, which remained mostly resistant to Endo-H digestion. Whereas, both the lower 130-and 135-kDa bands of the wild type, C67S, and C109S mutant receptors showed sensitivity to Endo-H digestion with downward mobility shifts, indicating that these are intracellularly trapped, high mannose receptor forms. For the C140S mutant, however, both the 300-and 140-kDa bands were largely resistant to Endo-H digestion, and only a small fraction of both bands showed mobility shift after Endo-H digestion as shown by arrows in Fig. 3B (right side). These results suggest that, of the different monomeric forms generated by C67S, C109S, or C140S mutants, only the 140-kDa monomeric form of C140S mutant receptor is expressed at the cell surface.
Mapping a Dimeric Interface Region Encompassing Cys-140 -The ability of the monomeric forms of the C140S mutant to express at the cell surface suggested that the Cys-140 residue is sufficient to form an intermolecular disulfide bond critical for dimerization. To further confirm this notion, we introduced ectopic cysteines at different positions on C140S mutant  background to determine whether enforced apposition of these cysteine residues would rescue any dimeric receptor form from the C140S monomers. Band patterns of seven Cys-scanning mutants, namely S119C, S126C, S129C, K134C, D143C, T146K, and T152C generated on the C140S mutant background were analyzed under non-reducing conditions and compared with the wild type and C140S mutant receptors. As shown in Fig. 4, four mutants S129C, K134C, D143C, and T146C, expressed mostly as two dimeric bands like the wild type receptor and did not show the140-kDa monomeric band of the C140S mutant; S126C mutant expressed as two dimeric bands but also showed some residual expression of the 140-kDa band; whereas, the expression pattern of the S119C mutant was similar to the C140S mutant and showed both 300-and 140-kDa bands. The T152C mutant expression pattern showed a single dimeric band like the C140S mutant, however, the 140-kDa band was largely missing. For all the mutants, some immunoreactive bands below the 140-kDa band were detectable, probably corresponding to the 130-to 135-kDa unprocessed, intracellular forms of mGluR1.

Function of Cysteine Mutants in PhI Hydrolysis
Assay-Because the monomeric form of the C140S mutant reached the cell surface, we tested whether this mutant receptor was capable of signal transduction using the intact cell L-glutamatestimulated PhI hydrolysis assay. PhI hydrolysis was measured by incubating cells with different micromolar doses of L-glutamate for 30 min, and formation of total inositol phosphates were monitored by counting the radioactivity incorporated. The C67S and C109S mutant receptors showed no significant response in dose-dependent L-glutamate response (data not shown) compared with the wild type mGluR1, but the C140S mutant showed a similar dose-response curve as the wild type mGluR1 (Fig. 5). DISCUSSION To confirm observations derived from study of mGluRs (10,11) that suggested that the dimerization domain is localized to the ECD, we expressed a truncation mutant receptor (RhodC-R1) that contained the seven transmembrane domains and carboxyl-terminal tail of the mGluR1 but lacked the ECD. As expected, the RhodC-R1 mutant receptor expressed as properly processed, cell surface monomers in the presence or absence of reducing agents. This identified the ECD as the locus for cova-lent-linked dimerization of mGluR1. Within this ECD, the suspected covalent-linked dimerization domain was localized by proteolysis of mGluR5 ECD within the amino-terminal 17 kDa of the ECD (17). Thus, we began to analyze three cysteine mutations within this region of mGluR1. Mutations of the cysteine residues at positions 67, 109, and 140 generated a substantial amount of monomers, making it difficult to confirm whether all these cysteines are directly involved in disulfidelinked dimerization. Further studies showed that monomers generated by mutations of the cysteine residues at positions 67 and 109 lead to primarily intracellularly trapped, incompletely processed receptors, indicating a potential problem with folding, processing, or trafficking of these receptors to the cell surface. In contrary, mutation at Cys-140 generated monomers that were processed normally and expressed at the cell surface. Thus, it is conceivable that mutation at Cys-67 and Cys-109 may affect the conformation of the ECD in a manner that prevents dimerization, whereas it is highly likely that the cell surface-expressed, properly processed (thus properly folded) monomeric receptor forms generated by mutation of Cys-140 is a direct consequence of disruption of an intermolecular disulfide linkage. Surprisingly, all three cysteine mutations also generated some residual dimeric receptor forms along with the monomeric forms. It is unclear how these dimeric forms are generated and expressed at the cell surface, but some of these processes could represent "artifacts" of mutagenesis and/or an overexpression problem in the heterologous cell system. Reduction with ␤-mercaptoethanol showed that much of these high molecular weight bands do not get reduced like the wild type dimers and mostly remain as SDS-resistant receptor forms, which may have resulted from aggregation of overexpressing proteins. However, we cannot conclude from this study whether mutations at these cysteines fail to completely block dimer formation, either because another cysteine ordinarily uninvolved in intermolecular disulfide linkage forms an "illegiti- mate" intermolecular disulfide bond or there is involvement of another unidentified cysteine in this dimerization process.
Our results of the Cys-scanning mutagenesis study, although not conclusive, show a simple approach for localizing the extent of a putative dimer interface domain and further confirm that the Cys-140 is indeed critical for mGluR1 dimerization. By introducing a cysteine at different amino acid locations in the context of the C140S mutant background, we can show rescue of dimeric forms from the monomeric C140S mutant forms. This allowed us to tentatively determine the extent of this putative dimer interface of mGluR1 on both sides of the Cys-140 position. We observed that introduction of an ectopic cysteine at positions Ser-129, Lys-134, Asp-143, and Thr-146 can fully rescue a dimeric form from the monomeric 140-kDa form of C140S, whereas cysteine at positions Ser-119, Ser-126, and Thr-152 can either partially or completely fail to rescue this dimeric form. This result thus favors the existence of a specific interreceptor contact face at least encompassing Ser-129 to Thr-146 and implies that this juxtamembrane domain of the monomeric receptors are closely apposed in the receptor dimer. Similar introduction of ectopic cysteines has been used to identify dimer interfaces in signal transduction by bacterial aspartate receptors, erythropoietin receptors, and receptor tyrosine kinases (23)(24)(25). Interestingly, this putative Cys-140 dimer interface in all mGluRs, Ca 2ϩ receptor, V2Rs, and T1Rs shows substantial sequence diversity. It is possible that this dimer interface in different mGluRs confers specificity for homodimerization of different mGluR subtypes; because the mGluR5 was shown to homodimerize but not to heterodimerize with mGluR1 (17), it reflects the importance of specific recognition sequences in addition to the conserved cysteine necessary for dimerization.
It is instructive to consider the present results in the context of a model of the three-dimensional structure of the mGluR1 ECD as originally proposed by O'Hara et al. (9). Sequence comparison of the ECD of the family 3 receptors, e.g. mGluRs, Ca 2ϩ receptor, and GABA B receptor with the known threedimensional crystal structure of the bacterial periplasmic binding proteins (LIVP, LIVPA, LEUBP, etc.) supports a bilobed structure of these receptor ECDs (8,9,13). It is proposed that, like bacterial periplasmic binding proteins, the two lobes of the ECD domain of family 3 receptors also close upon ligand binding like a "venus flytrap" and that this change in conformation is transduced to the transmembrane domain to activate Gprotein signaling. For mGluR1, mGluR4 and GABA B ligand binding domains were mapped on lobe-1 of the ECD by modeling and mutagenesis studies (8 -10, 26). Interestingly, mutations of all three cysteines, Cys-67, Cys-109, and Cys-140 that block dimer formation are located within the lobe-1 of this bilobed mGluR1 ECD model (9,13). Based on our mutagenesis experiments, we hypothesize that a lobe-1 versus lobe-1 intermolecular disulfide linkage between Cys-140 of two monomeric mGluR1 forms generates a disulfide-linked mGluR1 dimer. We also suspect that Cys-67 and Cys-109 of the mGluR1 may be intramolecular disulfide-linked and mutation of either residue disrupting this linkage may lead to misfolding of lobe-1 of the monomers. This misfolding can prevent formation of the Cys-140 dimer interface region that normally protrudes from the lobe-1, thus leading to improperly processed, intracellularly trapped monomeric receptor forms.
Disulfide-linked dimerization is unique for the GPCR family and was first identified in mGluRs and Ca 2ϩ receptor. Unlike the receptor tyrosine kinases, which are known to dimerize upon ligand binding (24,25), the functional significance of GPCR dimerization is relatively unknown. The GABA B receptor forms a functional heterodimer by a coiled-coil interaction involving a region within the intracellular carboxyl terminus of two GABA B receptor subtypes (27). Structural and biochemical studies show that the rhodopsin family of GPCRs is mostly monomeric and that some homodimer may form, mostly by non-covalent interactions within seven transmembrane domains, but the functional consequences of such dimer formation is not well understood (28). Recently, it has been shown that two fully functional opioid receptors, and ␦, heterodimerize and synergistically bind highly selective agonists and potentiate signal transduction (29). In the present study, we observed that the overall pattern of glutamate-induced phosphoinositol dose-response profile of the C140S mutant receptor was similar to the wild type mGluR1. HEK-293 cells overexpressing C140S mutant receptor also showed a higher basal phosphoinositide activity, but this constitutive higher basal activity was not unique for cells expressing the C140S mutant receptor and was also seen in cells expressing the wild type mGluR1, as reported by other investigators (30). This result implies that the monomeric C140S mutant receptors are functional, but this function must be qualified by the caveat that the glutamate response of the C140S mutant may depend on the formation of the residual dimers at the cell surface. In accordance, reduction with the reducing agent dithiothreitol has recently been shown to block the function of group1 mGluRs in transfected HEK-293 cells, cerebellar granule cells, and hippocampal slices (31). However, it is possible that dithiothreitol may block functional response of receptors by disrupting several putative intramolecular disulfide linkages thus changing the overall conformation of the receptor binding/activation domains. Interestingly, we have recently found that monomeric mutant form of another family 3 receptor, the Ca 2ϩ receptor, is functional (13); likewise, it is possible that monomeric mGluR1 receptor form may also be functional.
In conclusion, this study has provided new insights into the structural requirements for dimer formation of mGluR1 and provided important clues to further understand the functional impact of this dimerization process.