Expression and Purification of the Extracellular Ligand Binding Region of Metabotropic Glutamate Receptor Subtype 1*

Each metabotropic glutamate receptor possesses a large extracellular domain that consists of a sequence homologous to the bacterial periplasmic binding proteins and a cysteine-rich region. Previous experiments have proposed that the extracellular domain is responsible for ligand binding. However, it is currently unknown whether the extracellular ligand binding site can bind ligands without other domains of the receptor. We began by obtaining a sufficient amount of receptor protein on a baculovirus expression system. In addition to the transfer vector that encodes the entire coding region, transfer vectors that encode portions of the extracellular domain were designed. Here, we report a soluble metabotropic glutamate receptor that encodes only the extracellular domain and retains a ligand binding characteristic similar to that of the full-length receptor. The soluble receptor secreted into culture medium showed a dimerized form. Furthermore, we have succeeded in purifying it to homogeneity. Dose-response curves of agonists for the purified soluble receptor were examined. The effective concentration for half-maximal inhibition (IC50) of quisqualate for the soluble receptor was 3.8 × 10−8 m, which was comparable to that for the full-length receptor. The rank order of inhibition of the agonists was quisqualate ≫ ibotenate ≥l-glutamate ≈ (1S,3R)-1-aminocyclopentane-1,3-dicarboxylic acid. These data demonstrate that a ligand binding event in metabotropic glutamate receptors can be dissociated from the membrane domain.

subdivided into two groups: ␣-amino-3-hydroxy-5-methyl-4isoxazolepropionic acid receptors and kainate receptors. iGluRs are ligand-gated ion channels that transduce glutamate binding into cation influx. mGluRs that have been discovered most recently comprise eight subtypes, which are divided into three groups according to agonist selectivity, coupling to different effector systems, and sequence homology (3)(4)(5)(6). Group I includes mGluR1 and mGluR5, which are coupled to inositol phospholipid metabolism. Group II (mGluR2 and mGluR3) and group III (mGluR4, mGluR6, mGluR7, and mGluR8) are negatively coupled to adenylate cyclase activity. Functional analyses of these mGluRs are now avidly being performed. The evidence is accumulating that mGluRs modulate excitatory synaptic transmission (7) through various neural transduction pathways, such as regulation of neurotransmitter release (8), influences on ion channel activity (9), and modulation of synaptic plasticity (10).
mGluRs have a remarkably large extracellular domain that has no homology with the other G protein-coupled receptors (GPCRs) except Ca 2ϩ -sensing receptors (11). Previous experiments (12,13) have proposed that the ligand binding site resides mainly in the extracellular domain. Thus, the mode of ligand binding of mGluRs is different from that of other GPCRs for small molecule transmitters, such as adrenaline and acetylcholine. Thus, it is a subject of great interest how extracellular signals are transmitted into cells and what roles other parts of the receptor protein perform in mGluRs. Recently, various data of mGluRs on the brain function have been obtained, and the more specific and stronger antagonists and agonists have been urgently requested both in neurobiology and clinical fields (6,14). For designing the antagonists, the information on the three-dimensional structure of the receptor protein is very valuable. In order to determine a tertiary structure, a sufficient amount of pure receptor protein is needed. However, at present, it is extremely difficult to solubilize the receptor protein from membrane preparations. Therefore, we tried to produce a ligand binding region in a soluble form. If a ligand binding region can be produced independently of the membrane domain, it will be also a good tool for understanding activation mechanism of this intriguing receptor. In an ␣-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor, one of glutamate-gated ion channels, Kuusinen et al. (15) have shown that an agonist binding site can be reconstituted by making a fusion protein of two discontinuous segments: one is proximal to the first transmembrane segment, and the other is located between the third and fourth transmembrane segments.
In the current study, we have expressed a mGluR subtype 1␣ (described as mGluR1 from now on) in insect cells on a baculovirus system. We have isolated an extracellular ligand bind-ing region that has a biochemical characteristic comparable to that of the full-length receptor. Purification and characterization of the soluble mGluR are presented. This is the first demonstration among GPCRs for small molecule transmitters that a ligand binding region is produced in a soluble form.
Construction of Transfer Vectors for Expression of mGluRs in Insect Cells-A SacII fragment that contains the full coding region of mGluR1 was excised from pmGR1 (3) and was ligated into SacII-digested pBlu-scriptII SK(ϩ) (Stratagene, La Jolla, CA). A NotI-XbaI fragment of the intermediate construct was ligated into XbaI-NotI-digested pVL1392 (PharMingen, San Diego, CA) to generate pmGluR102.
pmGluR104 was made as follows. pmGR1 was used as a template of polymerase chain reaction (PCR) to amplify a sequence (nucleotides 3276 -3597) according to the numbering of the sequence by Masu et al. (3). Primer TO3 (5Ј-CCAGGAGGAGTCCATC-3Ј) and primer TO4 (5Ј-CAGTGCTCTAGATCACTAGTGATGGTGATGGTGATGCAGGGTGG-AAGAGCTTTGC-3Ј) were used. TO4 contains six consecutive histidine codons and an XbaI site. A 310-base pair ClaI-XbaI fragment of the PCR product was ligated with a 3.3-kilobase pair NotI-ClaI fragment of pmGR1 and a 9.6-kilobase pair XbaI-NotI-digested pVL1392 vector.
pmGluR103 was made as follows. PCR was done using pmGR1 as a template to amplify a sequence (nucleotides 1239 -1506) with primers TO1 (5Ј-CATCAATGCCATCTATGCCATGGC-3Ј) and TO2 (5Ј-CAGTG-CTCTAGATCACTAGTGATGGTGATGGTGATGTTCATGCCAGGTCC-CCACG-3Ј). Six histidine codons and an XbaI site were included in TO2. The PCR product was digested with NcoI and XbaI and was cloned into the XbaI-NcoI-digested fragment of pmGluR102.
Generation of Recombinant Baculoviruses for mGluRs-Recombinant baculoviruses encoding full-length and truncated cDNAs were produced in Sf9 cells with the transfer vectors. Briefly, Sf9 cells (1 ϫ 10 5 ) were transfected with 0.1 g of BaculoGold linear DNA (Phar-Mingen) and 1.0 g of each plasmid using a liposome transfection kit (Life Technologies, Inc.). The viruses were harvested after 6 days. For the expression assay, 2.5 ϫ 10 6 Sf9 or High Five cells were infected with each baculovirus. 72 h after infection, cells were harvested, and membranes were prepared as below. In the case of the cells infected with baculoviruses for mGluR103, mGluR107, and mGluR108, culture medium was collected and centrifuged at 800 ϫ g for 10 min and filtered. Membranes were solubilized on ice for 30 min in lysis buffer (40 mM Hepes, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, and 1.0% Chaps) containing protease inhibitor mixtures (1 mM phenylmethylsulfonyl fluoride, and 10 g/ml each of leupeptin, aprotinin, pepstatin A, benzamidine, and trypsin inhibitor), and were centrifuged at 12,000 ϫ g for 20 min. Aliquots of the supernatants and medium were analyzed by immunoblotting.
The plaque-purified recombinant viruses containing the full-length cDNAs for mGluR were used to infect Sf9 cells (1 ϫ 10 9 ) at a multiplicity of infection of 1. The purified recombinant viruses containing truncated cDNAs for mGluR, mGluR103, mGluR107, and mGluR108 were used to infect High Five cells (1 ϫ 10 8 -2 ϫ 10 8 ). Cells were harvested at approximately 72 h post infection. Culture medium was prepared as described above for smaller cultures. Membranes were prepared as described below.
Membrane Preparation-Membranes were obtained by nitrogen cavitation. Briefly, CHO cells or Sf9 cells were suspended in Buffer A (20 mM Tris-HCl, pH 7.5, 1 mM EDTA, and protease inhibitor mixtures), placed in a pressure chamber (Parr bomb) and pressurized to 800 p.s.i. with nitrogen gas for 30 min on ice. The mixture was collected by pressure release and centrifuged at 800 ϫ g. The supernatant was then centrifuged at 100,000 ϫ g for 30 min to pellet down the membranes. The membranes were divided into aliquots, frozen in liquid nitrogen, and stored at Ϫ80°C until use.
Western Blot Analyses-Cell membranes and culture medium were separated by SDS-polyacrylamide gel electrophoresis (PAGE) and electroblotted onto a nitrocellulose membrane (Schleicher & Schuell). The membrane was then blocked for 1 h in TBST (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, and 0.05% Tween 20) with 1% bovine serum albumin and incubated overnight at room temperature with a 1:5000 dilution of anti-mGluR1 monoclonal antibody (MAb) mG1Na-1 and a 1:5000 dilution of polyclonal antibody A52 in TBST. After being washed, the membrane was incubated with a goat anti-mouse IgG or a goat antirabbit IgG conjugated with alkaline phosphatase. Color development was done by a commercial detection kit (Promega). A MAb mG1Na-1 was made by Dr. Akio Neki as follows. The spleen cells from Balb/c mice immunized with glutathione S-transferase (GST)-fused mGluR1 peptide (amino acid residues 104 -154) were fused with X63-Ag8.653 cells as described (17). The character of MAb mG1Na-1 was verified using CHO cells, which produce mGluR1, in comparison with polyclonal antibodies G18 and A12 (18). Polyclonal antibody A52 was raised against an mGluR1␣ C-terminal peptide (amino acid residues 859 -1199) (19).
Ligand Binding Assay-Membrane fractions (20 -100 g) of baculoviruses-infected cells were incubated with [ 3 H]quisqualate (20 nM) in a total volume of 200 l of Buffer B (40 mM Hepes, pH 7.5, and 2.5 mM CaCl 2 ) for 1 h at room temperature. The reaction mixture was aspirated onto GF/C filter (Millipore Corp.). After being washed with Buffer B, the filters were briefly dried and counted by the scintillation counter. Nonspecific binding was determined in the presence of 1 mM L-glutamate. For soluble mGluRs secreted into culture medium, a new binding assay was established using nickel conjugated beads as described below. The concentrated medium (100 -200 l) or purified soluble receptor mGluR108 (1 g) was mixed with Ni 2ϩ -NTA conjugated beads (Qiagen Inc., Chatsworth, CA) at 4°C for 15 h. After several washings with Buffer B, [ 3 H]quisqualate was added with and without ligands. After being shaken for 1 h at room temperature, the reaction mixture was washed with Buffer B several times and filtered through GF/C filters. The remaining radioactivity was counted. Because aspirates through GF/C filters did not show any significant signal by immunoblotting, the entrapment of His-tagged receptor protein by Ni 2ϩ -beads seemed to be complete.
Purification of the Soluble mGluR-A conditioned medium of soluble mGluR1-producing High Five cells was concentrated to 10 -50-fold by XM50. After being dialyzed against Buffer C (10 mM Hepes, pH 7.5, and 200 mM NaCl), the dialyzate was centrifuged at 100,000 ϫ g for 30 min and loaded onto an immunoaffinity column, which was conjugated with MAb mG1Na-1 as described above. The column was washed with 30 ml of Buffer C, and the bound material was eluted with 500 mM NaCl containing 500 mM sodium thiocyanate and 10 mM Hepes, pH 7.5. Aliquots of the eluate were analyzed by SDS-PAGE followed by immunoblotting. Positively stained fractions were pooled and analyzed by ligand binding assay. Then the pooled fraction was dialyzed against 10 mM Hepes, pH 7.5, containing 200 mM NaCl and 10% glycerol, and the dialyzate was loaded onto a HiLoad 16/60 Superdex 200 (Amersham Pharmacia Biotech) by high performance liquid chromatography (Waters 626 LC). The flow rate was 0.5 ml/min, and fractions of 2.5 ml each were collected. Fig. 1 shows a schematic drawing of the transfer vectors used in the baculovirus infection experiments. pmGluR102 and pm-GluR104 encode the membrane-bound forms of mGluR1. pm-GluR103, pmGluR107, and pmGluR108 are designed to produce the truncated forms of mGluR1, which are expected to be secreted into the culture medium.

RESULTS
Recombinant baculoviruses for the expression of full-length mGluR1 were prepared and used to infect Sf9 insect cells in monolayer cultures. Immunoblotting analysis of expressed mGluR1 is shown in Fig. 2A. Both the polyclonal antibody A52, which was raised against a C-terminal peptide of mGluR1, and the MAb mG1Na-1 yielded a 155-kDa band in the membrane preparation from Sf9 cells infected with recombinant viruses but not in the control membrane preparation from Sf9 cells infected with wild-type virus, Autographa californica nuclear polyhedrosis virus (AcNPV). mGluR1 expressed in CHO cells showed a more slowly migrating band. This size difference may reflect less extensive glycosylation in the insect cells. A ligand binding assay was performed using membrane fractions of Sf9 cells infected with their respective recombinant viruses. 3 H-Labeled quisqualate specifically bound the membranes, and the binding was displaced by unlabeled glutamate (Fig. 2B). The control membranes prepared in parallel from wild-type virus (AcNPV)-infected cells did not show any specific binding. Both levels of expression of the receptor protein and the ligand binding ability seemed not to show a significant difference between the cells infected with recombinant viruses for mGluR102 and for mGluR104. The inhibition of [ 3 H]quisqualate binding with several agonists was examined as shown in Fig. 2C. The effective concentration for half-maximal inhibition (IC 50 ) of quisqualate was 3.0 ϫ 10 Ϫ8 M, which is comparable to the effective concentration for half-maximal response (EC 50 ) in inducing the stimulation of phosphatidylinositol hydrolysis in CHO cells permanently expressing mGluR1 (16). The rank order of inhibition was quisqualate Ͼ Ͼ L-glutamate Ϸ ibotenate Ն (1S,3R)-ACPD. Ohashi et al. (20) have obtained comparable data using the membrane fractions of Sf9 cells infected with recombinant baculovirus for the human full-length mGluR1. They reported that 3 H-labeled quisqualate binds the receptor in a saturable manner, with a K d value of 5.26 ϫ 10 Ϫ8 M. Their IC 50 value, 2 ϫ 10 Ϫ8 M, of quisqualate for the human full-length mGluR1 agrees with ours.
Next, we tried to dissect the ligand binding region of mGluR1. We infected High Five cells with recombinant viruses for mGluR103, mGluR107, and mGluR108, which were expected to produce the extracellular portions of mGluR1. Fig. 3 shows the Western blotting analyses of culture medium of cells infected with the recombinant baculoviruses. mGluR103 and mGluR107 produced a 66-kDa band, and mGluR108 produced a 74-kDa band under the reducing conditions. The level of expression of mGluR108 was much higher than those of mGluR103 and mGluR107. Under the nonreducing conditions, both bands shifted to the bands of approximately twice the size of the bands observed under the reducing conditions, suggesting a dimer formation. A minor faster migrating band observed in mGluR108 under the nonreducing conditions seemed to be a monomer. The 74-kDa band under the reducing conditions looked broad and may reflect heterogeneities of the glycosylation pattern. mGluR107, which differs from mGluR103 only in the flag tag sequence added to the N-terminal end, did not show any obvious band under the nonreducing conditions. Although the cause of the anomalous electrophoresis of the mGluR107 protein was not clear, it suggested an aggregated or an unfolded protein formation, which may be brought about by the disruption of the tertiary structure due to the peptide tag.
To study ligand binding properties of soluble receptors, we developed a new binding assay using Ni 2ϩ -conjugated beads, which can trap the histidine tag added to the soluble mGluRs. Concentrated culture medium (100 -200 l) was incubated with Ni 2ϩ -conjugated beads. Although mGluR103 and mGluR107 showed no appreciable binding, mGluR108 did reveal a significant binding ability. This binding was inhibited in the presence of 1 mM L-glutamate. Neither concentrated control culture medium nor separately prepared His-tagged soluble receptor (soluble low density lipoprotein receptor; data not shown) showed any significant binding. We examined the agonist selectivity of mGluR108 as shown in Fig. 4. Inhibition of labeled quisqualate binding by several agonists was analyzed. The rank order of inhibition was quisqualate Ͼ Ͼ L-glutamate Ϸ (1S,3R)-ACPD Ϸ ibotenate. (2S,1ЈR,2ЈR,3ЈR)-2-(2,3-Dicarboxycyclopropyl)-glycine and (1S,3S)-ACPD inhibited the quisqualate binding by 37 and 56% at 1 ϫ 10 Ϫ4 M, respectively; however, L-2-amino-4-phosphonobutyrate, which is a group III-specific agonist, hardly inhibited it (data not shown). The iGluR ago- nists N-methyl-D-aspartate, ␣-amino-3-hydroxy-5-methyl-4isoxazolepropionic acid, and kainate did not show any significant inhibition at 1 ϫ 10 Ϫ4 M (data not shown). These binding characteristics clearly indicate that the expressed soluble receptor, mGluR108, is capable of binding ligands in the same manner as the native group I mGluR.
Because newer ligands have recently been used in physiological and pharmacological experiments on mGluRs, we examined their reactivities with the soluble mGluR108. Fig. 5A shows the doseresponse curves of the newer ligands. (S)-3,5-Dihydroxyphenylglycine, which is a group I-specific agonist (21)(22)(23) and (R,S)-1-aminoindan-1,5-dicarboxylic acid, which is a group I-specific antagonist Next, we started the purification of the soluble receptor mGluR108 from the culture medium. Approximately 2.5 liters of the culture medium was concentrated, dialyzed, and loaded on an immunoaffinity column as described under "Experimental Procedures." After being washed with the binding buffer, the bound material was eluted. Aliquots of the flow-through and the eluates were analyzed by silver staining (Fig. 6A) and immunoblotting (Fig. 6B). Fractions that were positively stained by the immunoblot completely agreed with the ligand binding activity, and these fractions were pooled and purified by size exclusion column chromatography (Superdex 200) as shown in Fig. 7. A single peak was obtained. Peak fractions were collected and analyzed by SDS-PAGE under reducing and nonreducing conditions, followed by silver staining. On the basis of the result of an overloaded gel electrophoresis of the purified soluble receptor protein, we estimated that the purity of our final material is 99% (data not shown). Calibration with the molecular size markers showed the eluting position of the soluble mGluR108 to be around 190 kDa. The purified material is definitely larger than a monomer and seems to be a dimer or a larger oligomer. Fig. 8 shows dose-response curves of group I mGluR agonists for the purified mGluR108. IC 50 value of quisqualate, 3.8 ϫ 10 Ϫ8 M, was close to that for the full-length receptor, mGluR104, as described in Fig. 2C. The rank order of inhibition of the agonists was quisqualate Ͼ Ͼ ibotenate Ն L-glutamate Ϸ (1S,3R)-ACPD. The IC 50 values of L-glutamate and (1S,3R)-ACPD for the purified receptor seems to be 2-5 times as large as those for the full-length receptor. The rank order of the inhibition is reversed between L-glutamate and ibotenate. Although the significance of this small difference is unknown, it may reflect the difference between the crude membrane preparation and the purified receptor protein. DISCUSSION mGluRs have occupied a unique place in the sense that they have a very large extracellular domain and no amino acid sequence homology in the transmembrane helix with the other conventional GPCRs. Takahashi et al. (12) have made chimeric receptors between mGluR1 and mGluR2 and have shown that the large extracellular domain plays a crucial role in determining agonist selectivity. O'Hara et al. (13) have discovered the homology between the N-terminal extracellular region (residues 1-496) (which we tentatively call region I) and the bac- terial amino acid-binding proteins. They have constructed the molecular model of the glutamate binding site. The model has led them to point out the two critical amino acid residues, substitutions of which abolished [ 3 H]glutamate binding. In the extracellular domain of mGluR, there is also a short stretch of the cysteine-rich sequence (region II) just proximal to the membrane-spanning domain. However, the ligand binding site has not yet been isolated as a soluble molecule. In this investigation, we have succeeded in producing a soluble metabotropic glutamate receptor without the membrane-anchored domain and have provided compelling evidence that the soluble receptor consisting of the extracellular regions I and II is sufficient for conferring the affinity and selectivity of ligand binding characteristic of the mGluR.
Extensive studies of structure-function relationships have been done in GPCRs, and different modes of agonist binding are known (30,31). Monoamines and other small ligands, such as catecholamine and acetylcholine, are bound to the pockets formed within the transmembrane segments. In the receptors for neuropeptides or chemokines, a major binding site is located in the N-terminal segments of the receptors, with some contribution of the extracellular loops or the outer portions around the transmembrane segments. In receptors for large glycoprotein hormones, such as luteinizing hormone and thyrotropin, a large extracellular domain is responsible for high affinity ligand binding (32)(33)(34)(35), but a low affinity site is located in the extracellular loops in luteinizing hormone receptor (36,37). Although the role(s) of the extracellular loops and transmembrane segments of mGluR remains to be determined in the present study, it should be pointed that the soluble form of mGluR1 has a ligand affinity and specificity comparable to those of the full-size mGluR1. Thus, our results indicate that the ligand binding event is really dissociable from transmem-brane signaling in this receptor system. Pharmacological analyses with the conventional agonists and the newer ligands suggest that the soluble mGluR would be a reliable method for binding studies of future ligands that are expected to be developed.
We have made and analyzed several truncated forms of soluble mGluR1, which encode parts of the extracellular domain of mGluR1. The IC 50 value of quisqualate for mGluR108 is remarkably similar to those of the full-length membraneanchored form of the receptor. The agonist selectivity examined also constitutes a typical feature of mGluR1. These results indicate that the secreted extracellular domain forms a correct conformation for ligand binding. The soluble form of mGluR1 shows a dimerized form. A similar dimerization has been reported for several members of the conventional GPCRs, and this dimerization is ascribed to the molecular association in the transmembrane domains of these receptors (38,39). In addi- FIG. 6. Immunoaffinity chromatography of the soluble mGluR108. Culture medium (2.5 liters) of High Five cells infected with recombinant virus for mGluR108 was concentrated 50-fold and dialyzed against Buffer C. The dialyzate was loaded on a HiTrap Affinity column (1 ml) that had been coupled with mGluR1-specific MAb mG1Na-1 and equilibrated with Buffer C. The column was washed with 30 ml of Buffer C, and the bound material was eluted with 500 mM NaCl containing 500 mM sodium thiocyanate and 10 mM Hepes, pH 7.5. A, indicated aliquots of the eluate were subjected to 7.5% SDS-PAGE and silver stained. B, identical samples were analyzed by Western blot with MAb mG1Na-1 as described in the legend to Fig. 2A.   FIG. 7. Gel filtration chromatography of the soluble mGluR108. Soluble mGluR108 (64 g) from the immunoaffinity column ( Fig. 6) was loaded onto a HiLoad 16/60 Superdex 200 gel filtration column (Amersham Pharmacia Biotech). The column was equilibrated and eluted with 10 mM Hepes, pH 7.5, containing 200 mM NaCl and 10% glycerol. 2.5 ml of each fraction was collected. The molecular masses of the marker proteins were blue dextran (Ͼ2000 kDa), bovine thyroglobulin (669 kDa), horse spleen apoferritin (443 kDa), sweet potato ␤-amylase (200 kDa), yeast alcohol dehydrogenase (150 kDa), bovine serum albumin (66 kDa), and bovine erythrocytes carbonic anhydrase (29 kDa). Aliquots (20 l) of the indicated gel filtration fractions were subjected to 7.5% SDS-PAGE under the reducing conditions with 2% ␤-mercaptoethanol and under the nonreducing conditions, followed by silver staining. tion, Romano et al. (40) have reported that the truncated form of mGluR5 containing the first transmembrane segment forms a dimer. They demonstrated that an N-terminal 17-kDa region is required for dimerization. Our observation that mGluR103 dimerized is consistent with theirs. However, the truncated forms of the receptor, mGluR103 and mGluR107, which contain only region I, do not express well, and they show no significant ligand binding. Furthermore, mGluR108, which contains both region I and region II, expresses well and binds the ligand. This cysteine-rich region II of mGluR may thus impose a structural constraint on the receptor protein. A less likely possibility is that the ligand binding site might be composed of the two regions. In this context, the GABA B receptor (41), although belonging to the mGluR family with a large extracellular domain and a low but significant sequence homology to region I of mGluR, does not possess a region homologous to the cysteine-rich region II of mGluR. Furthermore, such a cysteine-rich region is absent in bacterial periplasmic amino acid-binding proteins. Thus, the real role of the cysteine-rich region II of mGluR remains an open question.
The bacterial periplasmic binding proteins serve as initial receptors of active transport for a variety of amino acids, sugars, peptides, oxyanions, and other nutrients. Although the binding proteins have different sizes (20 -60 kDa) and share little sequence homology, they all fold a similar two-lobed tertiary structure. Atomic structures of the binding proteins have been analyzed by Quiocho and co-workers (42) and other groups (43) using crystallography. On the basis of the comparison between ligand-bound and ligand-free forms, the "Venus flytrap" model has been proposed (44): the ligand binds preferentially to one lobe of the open, ligand-unloaded form. A bending motion at the hinge region between the two lobes in turn causes the other lobe to participate in binding and completely entrap the ligand. The closed, ligand-loaded form of the binding protein then interacts with membrane-bound components, thereby initiating nutrient translocation or flagella motion. Therefore, it should give us insights into the receptor activation mechanism of mGluR to reveal the three-dimensional structure of the mammalian glutamate receptor and determine whether a domain movement similar to that of the bacterial binding protein is produced upon ligand binding in mGluR. We speculate that glutamate induces a conformational change of the preexisting dimer or oligomer of glutamate receptor and triggers the signal transmission to the cytoplasmic signaling domain through the cysteine-rich region and the seven-transmembrane domain. The soluble form of mGluR1 produced in a sufficient amount will make it possible to conduct the biophysical analysis of mGluR1.