Introduction

Top Summary Introduction Procedures Results Discussion References

Several families of GPCRs can be defined based on their sequence homologies. Although they all possess seven transmembrane domains, members from different families share no sequence homology, even within these transmembrane domains. Family 1 is constituted by the large number of receptors homologous to rhodopsin and to adrenergic and glycoprotein receptors (Savarese and Fraser, 1992). Family 2 is composed of the receptors homologous to the glucagon and pituitary adenylyl cyclase-activating polypeptide receptors (Spengler et al., 1993), whereas family 4 corresponds to a subset of pheromone receptors (Dulac and Axel, 1995). The eight mGlu receptors (Conn and Pin, 1997; Nakanishi, 1992), the Ca²⁺-sensing receptor (Brown et al., 1993), the GABA_B receptor (Kaupmann et al., 1997), and a second subset of 100 pheromone receptors define the family three GPCRs (Herrada and Dulac, 1997; Matsunami and Buck, 1997).

In contrast to the other GPCRs, family 3 receptors possess a long amino-terminal domain (550-600 residues) that shares some sequence homology with bacterial PBPs (O'Hara et al., 1993). Based on the known three-dimensional structure of some PBPs and subsequent site-directed mutagenesis, it has been proposed that the amino terminus of mGlu receptors is the glutamate recognition domain (O'Hara et al., 1993). In agreement with this proposal, swapping the extracellular domain of the PLC-coupled mGlu1 receptor with that of either the mGlu2 or mGlu4 receptor that inhibits adenylyl cyclase generates chimeric receptors that activate PLC and possess agonist pharmacological characteristics that resemble those of mGlu2 and mGlu4 receptors, respectively (Takahashi et al., 1993; Tones et al., 1995).

As in the other GPCRs, most intracellular loops and the carboxyl-terminal tail of family 3 receptors are responsible for the G protein coupling (Pin et al., 1994; Gomeza et al., 1996a; Mary et al., 1998). Although some short segments in the intracellular loops of 7TM receptors play an important role in G protein-coupling selectivity, the conformational state of the 7TM region also may play a role. Indeed, the current hypothesis regarding 7TM receptors is that they can adopt various conformations stabilized by different ligands. These different conformations of a receptor can have distinct G protein-coupling selectivities or desensitization properties (Spengler et al., 1993; Eason et al., 1994; Robb et al., 1994; Gether et al., 1995; Perez et al., 1996).

The aim of the current study was to examine whether the ligand binding extracellular domain of these receptors has any influence on the G protein-coupling selectivity of family 3 receptors. For that purpose, we used mGlu receptors as representatives of this receptor family. Three groups of mGlu receptors can be distinguished based on their sequence homology and pharmacology. The group I includes mGlu1 and mGlu5 receptors, which activate PLC. The second group is composed of mGlu2 and mGlu3 receptors. The recently cloned Drosophila melanogaster DmGluA receptor belongs to this group of mammalian receptors (Parmentier et al., 1996). The third group includes mGlu4, mGlu6, mGlu7, and mGlu8 receptors. Both group II and III mGlu receptors inhibit cAMP formation (for a review, see Conn and Pin, 1997).

In the current study, we first show that a G protein-coupling profile of receptors can be defined using exogenous and chimeric G proteins (Conklin et al., 1993; Offermanns and Simon, 1995) in the same way that synthetic ligands are commonly used to define an agonist or antagonist pharmacological profile. To analyze the influence of the extracellular domain on the G protein-coupling profiles of mGlu receptors, we first examined whether this profile may be agonist dependent. Then, the extracellular domains of the most distantly related mGlu receptors, the rat mGlu1 and the D. melanogaster DmGluA receptors, were swapped. Our data demonstrate that the G protein-coupling and ligand-recognition selectivities are exclusively determined by the 7TM and extracellular regions of mGlu receptors, respectively.

	Experimental Procedures

Top Summary Introduction Procedures Results Discussion References

Materials. Ibotenate, L-quisqualate, L-CCGI, DHPG, DCG IV, DL-AP4, and L-SOP were obtained from Tocris Cookson (Essex, UK). The two isomers (1S,3S)-ACPD and (1S,3R)-ACPD were a generous gift of Dr. K. Curry (Vancouver, Canada). Glutamate was from Sigma-Aldrich (L'isle d'Abeau, France). LY354740 and (2R,4R)-APDC were a generous gift of Dr. D. D. Schoepp (Eli Lilly, Indianapolis, IN). ACPT-I was a gift from Dr. F. Acher.

Plasmids. The cDNAs encoding mGlu1a, mGlu2, mGlu4a, and DmGluA receptors were in the pRK vector under the control of a CMV promoter, as described previously (Joly et al., 1995; Gomeza et al., 1996b; Parmentier et al., 1996). The cDNAs of mGlu7a and mGlu8a originally in the pBluescript vector (gifts from Dr. J. Saugstad) were subcloned into the pRK vector. The cDNAs of the wild-type G_q and the chimeric G_qi (G_q with the carboxyl-terminal nine amino acids of G_i2), G_qo, and G_qz (G_q with the carboxyl-terminal five amino acids of G_o and G_z, respectively) were in the pcDNA-I expression vector (gift from Drs. B. Conklin and H. Bourne) (Conklin et al., 1993). The cDNAs of the mouse G₁₅ and the human G₁₆ were in the pCIS vector (gift from Drs. S. Offermanns and M. Simon) (Offermanns and Simon, 1995).

Culture and transfection of HEK 293 cells. HEK 293 cells were cultured in DMEM (GIBCO BRL, Gaithersburg, MD) supplemented with 10% fetal calf serum and transfected by electroporation as described previously (Gomeza et al., 1996b). Electroporation of 5 × 10⁶ cells was performed in a total volume of 150 µl with 6 µg of carrier DNA (500 µF, 260 V). Plasmid DNA expressing mGlu receptors (100 ng of mGlu1a, 1 µg of mGlu2, 2.5 µg of DmGluA (Parmentier et al., 1996); 5 µg of mGlu4, mGlu7a, or mGlu8a; or the chimeric mGlu1a/D receptors) was transfected alone or with 3 µg of plasmid DNA expressing G_qi, G_qo, G_qz, G_q (Conklin et al., 1993), G₁₅, or G₁₆ (Offermanns and Simon, 1995). Cells were resuspended and split into half of a 12-well cluster. For the cyclase experiments, 10⁷ cells were electroporated in a total volume of 300 µl with 12.5 µg of carrier DNA, 5 µg of plasmid DNA expressing DmGluA receptor, and 600 ng of LH receptor-expressing plasmid (950 µF, 280 V). Cells were resuspended and split into 12-well clusters.

Determination of IP accumulation. Determination of IP accumulation in transfected cells was performed as described previously after labeling the cells overnight with myo-[³H]inositol (23.4 Ci/mol; New England Nuclear, Le Blanc Mesnil, France) in DMEM Glutamax I (GIBCO-BRL, Life Technologies, Cergy Pontoise, France) (Gomeza et al., 1996b). The stimulation was conducted for 30 min in a medium containing 10 mM LiCl and the agonists. The basal IP formation was determined after a 30-min incubation in the presence of 10 mM LiCl and of the glutamate-degrading enzyme glutamate pyruvate transaminase (1 unit/ml plus 2 mM pyruvate). Results are expressed as the amount of IP produced divided by the radioactivity present in the membranes. The concentration-response curves were fitted using the equation y = [(y_max  y_min)/1 + (x/EC₅₀)ⁿ] + y_min and the Kaleidagraph program, where n is the Hill coefficient, x is the concentration of agonist, and y_max and y_min are the maximal and minimal responses, respectively.

Determination of cAMP accumulation. The cellular cAMP production was measured using the prelabeling technique as described previously (Parmentier et al., 1996). Four hours after being electroporated, cells were washed and incubated for 14 hr in DMEM Glutamax I containing 1 µCi/ml [³H]adenine (27 Ci/mmol). Stimulation of the cells was performed in HEPES buffer saline (146 mM NaCl, 4.2 mM KCl, 0.5 mM MgCl₂, 1 mM CaCl₂, 0.1% glucose, 20 mM HEPES, pH 7.4). cAMP formation is expressed as percentage conversion of [³H]ATP to [³H]cAMP: ([³H]cAMP × 100)/([³H]cAMP + [³H]ATP).

Xenopus laevis oocytes assay. The preparation of oocytes and the in vitro synthesis of RNA transcripts from the cloned cDNA were performed as described previously (Pin et al., 1994). Recordings were performed in Barth's medium using the two-electrode voltage-clamp technique (Axoclamp-2A; Axon Instruments, Burlingame, CA) 3-4 days after injection. Data were recorded on a PC and analyzed using the pCLAMP software.

List and sequence of the oligonucleotides used to generate chimeras. The sequence of the oligonucleotide primers used to generate the chimeric receptors are presented in the 5'-to-3' direction. The following nucleotides have their 5'-end sequences corresponding to mGlu1a receptor, with the junction between the mGlu1 and DmGluA receptor sequences indicated by a hyphen: 1aD/1S, ACA AAA GCG GAA TGG TAC GA-T CGG CCT GTT CAC TAC CAT G; 1aD/2S, GGA AAG GAG AAG TGA GCT GC-T GCT GGA TAT GCG ACA GCT G; 1aD/3S, AGT TCA CCT GCA GAG CCT GT-G GTC CTG GAC TTT GGC CCT A; 1aD/4S, TCC GTT ATC TTG AGT GGA GT-T CGT TGT TTG CCC TTA TTC C; D1a/1AS, CAA GGC TCA CTGCAC ACA GA-A GTA GGT TGT TCG GTT TCC T; D1a/2AS, CAG GCC GTG CAG ATC CAG CA-A CAG GTA TCC CCC TGT TGT T; D1a/3AS, TTG GGC CAC CAC CCC AGG TC-G CAG TCT TTA CAC GTA AAC T; and D1a/4AS, GCT ATG ATA GAT TCT ATG TC-G TTC CAT TTC ATA TAC TGG A. The following oligonucleotides have their 5' sequence corresponding to DmGluA, with the junction between the DmGluA and mGlu1a receptor sequences indicated by a hyphen: D1a/1S, AGG AAA CCG AAC AAC CTA CT-T CTG TGT GCA GTG AGC CTT G; D1a/2S, AAC AAC AGG GGG ATA CCT GT-T GCT GGA TCT GCA CGG CCT G; D1a/3S, AGT TTA CGT GTA AAG ACT GC-G ACC TGG GGT GGT GGC CCA A; D1a/4S, TCC AGT ATA TGA AAT GGA AC-G ACA TAG AAT CTA TCA TAG C; 1aD/1AS, GTA CCA TCA CTT GTC CGG CT-T CGT ACC ATT CCG CTT TTG T; 1aD/2AS, CAG CTG TCG CAT ATC CAG CA-G CAG CTC ACT TCT CCT TTC C; 1aD/3AS, TAG GGC CAA AGT CCA GGA CC-A CAG GCT CTG CAG GTG AAC T; and 1aD/4AS, GGA ATA AGG GCA AAC AAC GA-A CTC CAC TCA AGA TAA CGG A. The following oligonucleotides also were used: MGR15, Kos3, Dchim1, MGR42: MGR 15 (CGC TTC CAG TGT CGC CTA CC), Kos3 (TCT AGA ACT TAT TTA CTA TAT GAC AGA T), and Dchim1 (GGC CGG ATC CTT GGT TTG CTG AAT AC), MGR 42 (AGT GTA CAT GGT GAA GGC G).

Construction of chimeric receptors. Chimeras were constructed using the PCR overlap extension method as described previously (Pin et al., 1994). To construct 1a/D-1 chimera that contains the extracellular domain of the mGlu1a receptor and the 7TM and carboxyl-terminal regions of DmGluA receptor, a PCR was conducted with pKosDMGRAs as template and Kos3 and 1a/D-1S as primers to amplify the DmGluA receptor part of the chimera. A parallel PCR was conducted using pRKG1aS (Prézeau et al., 1996) as template DNA and MGR15 and 1a/D-1AS as primers to amplify the mGlu1a receptor part of the chimera. The products of both reactions were mixed and used as template in a third PCR using MGR15 and Kos3 as primers. The resulting chimeric PCR product was digested with PmlI and XbaI and subcloned into pmGR1a [mGlu1a receptor cDNA inserted into pBluescript-SK⁽⁾ (Pin et al., 1994)] cut with the same enzymes. After sequence verification, the resulting plasmid was digested with BamHI and EspI, and the smaller fragment was inserted into pKosDMGRAs previously cut with the same enzymes. The same strategy was used for the construction of 1a/D-2, 1a/D-3, and 1a/D-4 with the corresponding chimeric sense and antisense oligonucleotides.

	Results

Top Summary Introduction Procedures Results Discussion References

Exogenous G proteins can be used to define a G protein-coupling profile of mGlu receptors. To study the G protein-coupling selectivity of mGlu receptors, we needed tools that define a G protein-coupling profile of receptors. For that purpose, we used a series of exogenous wild-type or modified G protein  subunits in the same way that synthetic ligands are commonly used to define the agonist and antagonist pharmacological profiles.

15

16

q

i2

qi

o

z

qo

qz

16

qi

qo

15

qz

View larger version (41K): [in this window] [in a new window]   Fig. 1.   PLC activation by mGlu2, DmGluA, mGlu4a, mGlu7a, or mGlu8a receptors coexpressed with the Gq subunit, the promiscuous G16 and G15 subunits, or the chimeric Gq subunits. Basal and agonist-stimulated IP formation in cells expressing (a) the G proteins only, (b) the G proteins plus mGlu2 or (c) DmGluA, (d) mGlu4a, (e) mGlu7a, or (f) mGlu8a receptors. The agonist used is glutamate (1 mM) except for the mGlu7a receptor, for which the more potent DL-AP4 agonist was used at 3 mM. Values are mean ± standard error of at least two experiments performed in triplicate.

q

15

qi

qo

qi

qo

15

qi

qo

16

qz

qz

16

15

qi

qo

16

qz

15

16

q

q

qi

qo

qz

16

15

q

qi

qo

qz

16

15

View larger version (46K): [in this window] [in a new window]   Fig. 2.   PLC activation by the mGlu1a receptor expressed alone or with the wild-type or chimeric Gq subunits or the promiscuous G15 or G16 receptor. Basal and 1 µM glutamate-stimulated IP formation was determined in cells expressing the mGlu1a receptor alone () or with the indicated G proteins. Basal IP formation also was determined in cells expressing the indicated G proteins alone. Values presented are obtained by subtracting this latter value corresponding to the effect of the G protein alone from the basal and glutamate-induced IP formation measured in cells coexpressing the mGlu1a receptor and the indicated G protein. *, Significantly different (p < 0.05 by analysis of variance) from the basal (unbroken line) or glutamate-induced (dotted line) activities in cells expressing the mGlu1a receptor alone. Values are mean ± standard error of one typical experiment performed in triplicate.

The G protein-coupling profile of mGlu receptors seems to not depend on the agonist used. Previous reports indicated that the G protein-coupling specificity of a receptor can vary depending on the agonist used (see, for example, Spengler et al., 1993). To test whether this also may be the case for mGlu receptors, the coupling of mGlu2 and mGlu4 receptors to the exogenous G proteins described above was examined after stimulation with various agonists. As shown on Fig. 3a, the three specific group II agonists LY354740 (5 µM), (2R,4R)-APDC (500 µM), and DCG IV (50 µM) (Conn and Pin, 1997) stimulated IP production in cells expressing mGlu2 receptor and G_qi, G_qo, G₁₅, or G₁₆ but not in cells coexpressing mGlu2 receptors with G_q or G_qz, as observed with glutamate. Similarly, the three specific group III agonists DL-AP4 (50 µM), L-SOP (500 µM), and ACPT-I (500 µM) (Acher et al., 1997; Conn and Pin, 1997) stimulated IP formation in cells expressing the mGlu4a receptor and G_qi, G_qo, G_qz, or G₁₅ but not in cells coexpressing mGlu4a receptors with G_q or G₁₆ (Fig. 3b). These results revealed no difference in the G protein-coupling specificity of mGlu2 and mGlu4a receptors on stimulation with various agonists.

View larger version (45K): [in this window] [in a new window]   Fig. 3.   The G protein-coupling profile of mGlu2 and mGlu4a receptors is the same whatever the agonist used. a, HEK293 cells coexpressing mGlu2 receptors with the wild-type Gq; the chimeric Gqi, Gqo, or Gqz; or the promiscuous G15 and G16 were incubated for 30 min with control medium (basal) or medium containing LY354740 (5 µM), (2R,4R)-APDC (500 µM), or DCG IV (50 µM), and the IP formation measured. b, Same as in a with cells expressing the mGlu4a receptor and stimulated with DL-AP4 (50 µM), L-SOP (500 µM), or ACPT-I (500 µM). Values are mean ± standard error of triplicate determinations. The experiment were repeated twice with the same results.

Construction of chimeric mGlu1a/DmGluA receptors. Because the ligand PBP-like recognition domain of mGlu receptors seems to be distant from the 7TM region, it may be proposed that when occupied by an agonist, the domain of the family 3 receptors acts as the activator (as an agonist) of the 7TM region. The PBP-like domain of family 3 receptors may therefore, by stabilizing a specific conformation of the 7TM region, have some influence on the G protein-coupling specificity of the receptor. To test this possibility, the extracellular domains of the most distantly related mGlu receptors (i.e., the rat mGlu1a and the D. melanogaster DmGluA receptors) were swapped. These two receptors were chosen because they have distinct pharmacological profiles (Parmentier et al., 1996) and G protein-coupling specificities (Table 1). Four chimeras containing the PBP-like domain of the mGlu1a receptor and the transmembrane and intracellular domains of the DmGluA receptor (1a/D-1-1a/D-4 chimeras) were constructed. They differ by the position of the junction within the two sequences (Fig. 4). The converse chimeras also were constructed and called D/1a-1-D/1a-4.

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Positions of the junction points between mGlu1a and DmGluA receptor sequences in the different chimeras. Boxed, residues conserved in all mGlu receptors. Shaded, beginning of the first transmembrane domain.

G protein-coupling profiles of chimeric mGlu receptors. The ability of the four D/1a chimeric mGlu receptors to activate PLC was examined by injecting their corresponding in vitro synthesized transcripts into X. laevis oocytes. It is established that among G protein-coupled receptors, only those coupled to PLC activate Ca²⁺-dependent Cl channels by releasing Ca²⁺ from internal stores. Each of the D/1a chimeric mGlu receptors elicited an inward current when activated with 300 µM glutamate, whereas DmGluA receptors did not (data not shown). The D/1a-1 chimeric protein elicited the highest inward current, which was similar to that obtained with the wild-type mGlu1a receptor (data not shown). These results are in agreement with previous studies showing that mGlu1a chimeric receptors with the extracellular domain of the group II mGlu2 or the group III mGlu4 receptors activate PLC (Takahashi et al., 1993; Tones et al., 1995).

qi

15

16

qi

qo

qz

q

View larger version (33K): [in this window] [in a new window]   Fig. 5.   Analysis of the G protein coupling of mGlu1a/D-4 receptor chimera. a, Cotransfection of DmGluA or mGlu1a/D4 receptors with the LH receptor. Percentage of conversion of ATP in cAMP is estimated without any agonist and in the presence of either 1 mM glutamate or 100 ng/ml LH or in the presence of both agonists. Values are mean ± standard error of triplicate determinations from one typical experiment. b, Basal and 1 mM glutamate-stimulated IP formation in cells expressing the mGlu1a/D-4 receptor alone or with the Gq, G16, G15, Gqi, Gqo, or Gqz subunits. Values are mean ± standard error of three experiments performed in triplicate.

The chimeric mGlu1a/D-4 receptor displays the same agonist recognition profile as the mGlu1a receptor. To analyze whether the 7TM region of the DmGluA receptor influences the agonist selectivity of the receptor, we systematically compared the agonist pharmacological profiles of DmGluA, mGlu1a, and 1a/D-4 receptors using seven agonists (Table 2, Fig. 6, a and b). The rank order of potency of these agonists was found to be identical between the mGlu1a receptor and the 1a/D-4 chimeric receptor, with the PBP-like domain of the mGlu1a receptor (correlation coefficient r = 0.97, p < 0.005) whereas it was not between DmGluA and 1a/D-4 receptors (r = 0.115) (Fig. 6c). This indicates that the 7TM region of mGlu receptors does not have a major influence on the agonist pharmacological profile of these receptors.

View larger version (20K): [in this window] [in a new window]   Fig. 6.   Dose-response curves of agonist-induced IP stimulation (expressed in percentage of maximal glutamate stimulation) in mGlu1a receptor-expressing cells (a) and in cells expressing the mGlu1a/D-4 receptor and Gqi (b). In a and b, dose-response curves were obtained with quisqualate (), glutamate (), (1S,3R)-ACPD (), and (1S,3S)-ACPD (). c, For each of these drugs [glutamate, L-quisqualate, ibotenate, L-CCGI, DHPG, (1S,3R)-ACPD, and (1S,3S)-ACPD], the logarithm of the EC50 value obtained with the mGlu1a () or DmGluA () receptors was plotted versus the logarithm of the EC50 value obtained with the chimeric mGlu1a/D4 receptor. Only the significative linear regression obtained between the mGlu1 and mGlu1a/D-4 receptor values is shown (as well as the corresponding equation). Value are mean ± standard error () or mean only () of at least two independent determinations.

	Discussion

Top Summary Introduction Procedures Results Discussion References

Our results indicate that exogenous G protein  subunits can be used to define the G protein-coupling selectivity profile of mGlu receptors, in the way synthetic ligands are commonly used to define their pharmacological profile. By using G_q; the chimeric G_qi, G_qo, and G_qz; and the promiscuous G₁₅ and G₁₆ and measuring the ability of most mGlu receptors to activate PLC, we show here that group I, II, and III mGlu receptors have distinct G protein-coupling profiles. This constitutes the first demonstration that group II and III mGlu receptors can be distinguished not only by their pharmacological profiles and sequence homology (Nakanishi, 1992; Conn and Pin, 1997) but also by their G protein-coupling selectivity.

Our experiments were not aimed at identifying the endogenous G proteins activated by mGlu receptors in vivo. However, because the carboxyl-terminal end of the G protein  subunits plays a critical role in their specific interaction with GPCRs (Hamm et al., 1988; Wong et al., 1992; Conklin et al., 1993, 1996; Conklin and Bourne, 1993), our data may help identify the endogenous G proteins that could be activated by these receptors. The pertussis toxin-sensitive G_o and G_i and the pertussis toxin-insensitive G_z are expressed in many brain areas (Hinton et al., 1990; Wong et al., 1992), as are mGlu receptors. Our analysis of the coupling of group II and III mGlu receptors to G_qi, G_qo, and G_qz (Table 1) suggests that group III receptors couple to G_i, G_o, and G_z, whereas group II receptors couple to G_i and G_o but may not couple to G_z. The possible endogenous coupling of group II and III mGlu receptors to G_i and G_o is supported by the pertussis toxin-sensitive inhibition of adenylyl cyclase and voltage-sensitive Ca²⁺ channels in neurons induced by selective group II and III agonists (for a review, see Conn and Pin, 1997). However, the endogenous coupling of group III mGlu receptors to G_z remains to be demonstrated. Because G₁₅ and G₁₆ are found only in mouse and human hematopoietic cells, respectively (Amatruda et al., 1991; Wilkie et al., 1991), they cannot be involved in the natural transduction of mGlu receptors in the brain.

Our data reveal that the carboxyl-terminal end of the  subunits plays a critical role in their specific coupling not only to the large family of GPCRs (Conklin and Bourne, 1993; Dratz et al., 1993; Liu et al., 1995) but also to the mGlu receptor family, even though these two families of receptors do not share any sequence homology. Moreover, they show that a difference of only three residues (between G_qo and G_qz; see Fig. 7) in this short domain can discriminate between cyclase-inhibiting receptors. However, the carboxyl-terminal end of the  subunit is not the only region involved in the specific recognition of mGlu receptors. Although G₁₅ and G₁₆ possess an identical carboxyl terminus (Amatruda et al., 1991; Wilkie et al., 1991), G₁₅ couples to group II and III mGlu receptors, whereas G₁₆ couples to group II mGlu receptors only. Other regions with low sequence homology between these  subunits therefore may be involved in the interaction with these receptors. Among these regions is the sequence around the loop L9 of the  subunit that has been proposed to contact the receptor (Hamm et al., 1988; Conklin and Bourne, 1993; Lee et al., 1995; Lichtarge et al., 1996).

View larger version (19K): [in this window] [in a new window]   Fig. 7.   Alignment of the five carboxyl-terminal amino acids of G15 (which are the same for G16), Gi2, Go, Gz, and Gq. Boxed, amino acids conserved in the five sequences. Shaded, amino acids that differ between Gz and Gi/o.

The D. melanogaster mGlu receptor DmGluA belongs to group II mGlu receptors. Indeed, this G_i-coupled receptor share 47% sequence similarity with group II mammalian mGlu receptors, whereas it share 42% and 35% similarity with group I and III receptors, respectively. Moreover, it displays a pharmacological profile closer to that of group II than to that of the other groups of mGlu receptors (Parmentier et al., 1996). The use of the chimeric and promiscuous G proteins also reveals that despite sequence differences between the 7TM region of the DmGluA receptor and those of mammalian group II mGlu receptors, the DmGluA receptor has the same G protein-coupling specificity as the mGlu2 receptor. This confirms that DmGluA receptor is a D. melanogaster homologue of group II mammalian mGlu receptors, indicating that group II mGlu receptors were already defined in the common ancestor of vertebrates and arthropods.

It has been proposed that the large amino-terminal extracellular domain of mGlu receptors constitutes their ligand recognition domain. Homology modeling and site-directed mutagenesis of the mGlu1a receptor allowed the identification of two residues within this domain involved in glutamate binding (O'Hara et al., 1993). Moreover, the exchange of the extracellular domain of the group I mGlu1a receptor with that of either mGlu2 or mGlu4 receptors generates chimeric receptors able to activate PLC. These receptors also were shown to have pharmacological profiles similar to those of mGlu2 and mGlu4 receptors, respectively, using two or three different agonists (Takahashi et al., 1993; Tones et al., 1995). In the current study, we analyzed in greater detail the pharmacological profile of a converse chimeric receptor with the extracellular domain of mGlu1 and the 7TM region of the cyclase-inhibiting DmGluA receptor. Our results show that the agonist pharmacological profile of this chimeric receptor is identical to that of mGlu1 receptors. They further indicate that the extracellular domain of mGlu receptors is responsible for the ligand recognition in this receptor family. Moreover, they demonstrate that the 7TM region has no influence on the agonist pharmacological profile of these receptors. According to these data, the extracellular domain of mGlu receptors may be considered the activator of the 7TM region.

It recently was been reported that different agonists, probably by stabilizing different conformational states of the 7TM domains, can lead to the activation of different G proteins (Spengler et al., 1993; Eason et al., 1994; Robb et al., 1994; Gether et al., 1995; Perez et al., 1996). This means that the G protein-coupling selectivity of a 7TM receptor depends not only on the sequence of its intracellular loops but also on its conformational state. Our results revealed, however, no difference in the G protein-coupling profiles of either mGlu2 or mGlu4a receptor on activation with four different agonists. However, because the extracellular domain of mGlu receptors, when occupied by glutamate, can be considered the activator (or agonist) of the 7TM region, we also tested whether it could influence the G protein-coupling specificity by stabilizing a specific conformation of the 7TM region. By using exogenous G proteins, we show here that the G protein-coupling profile of the 7TM region of the DmGluA receptor is the same regardless of whether the extracellular domain is that of DmGluA or mGlu1 receptors, although these two domains share only 35% of sequence homology. This supports the conclusion that the extracellular domain of mGlu receptors may have no influence on their G protein-coupling profiles.

Taken together, these results demonstrate that mGlu receptors are composed of two domains, with the PBP-like domain entirely responsible for the ligand recognition and the 7TM region responsible for the G protein-coupling specificity. Because the agonist pharmacological and G protein-coupling profiles are not influenced by the 7TM and PBP-like domains, respectively, one may expect these two domains to evolve independently. Surprisingly, the only example of a mGlu receptor from a distant species, the DmGluA receptor, does not fit this hypothesis. Indeed, although the DmGluA receptor shares only 45% homology with the mGlu2 receptor, both its agonist pharmacological (Parmentier et al., 1996) and G protein-coupling profiles have been conserved during evolution. The functional analysis of other distant mGlu receptors would help us to understand this discrepancy between the apparent independence of the agonists and G protein-coupling specificities of mGlu receptors and their parallel conservation during evolution.

Within the 7TM domain, it has been proposed that the sequence of the intracellular loops plays an important role in G protein-coupling selectivity. We previously reported that the second intracellular loop was critical in determining the coupling of the mGlu1 receptor to PLC (Pin et al., 1994; Gomeza et al., 1996a). Further experiments would reveal whether this loop is indeed responsible for the different G protein-coupling profiles observed here between group II and III mGlu receptors.