The second intracellular loop of metabotropic glutamate receptor 1 cooperates with the other intracellular domains to control coupling to G-proteins.

Metabotropic glutamate receptors (mGluR) share no sequence homology with any other G-protein-coupled receptors (GPCRs). The characterization of their G-protein coupling domains will therefore help define the general rules for receptor-G-protein interaction. To this end, the intracellular domains of mGluR3 and mGluR1, receptors coupled negatively to adenylyl cyclase and positively to phospholipase C, respectively, were systematically exchanged. The ability of these chimeric receptors to induce Ca2+ signals were examined in Xenopus oocytes and HER 293 cells. The chimeric receptors that still possessed the second intracellular loop (i2) of these proteins were targeted correctly to the plasma membrane. Consistent Ca2+ signals could be recorded only with chimeric mGluR3 receptors that contains i2 and at least one other intracellular domains of mGluR3 have to be replaced by their mGluR1 equivalent to produce optimal coupling to G protein. These observations indicate that i2 of mGluR1 is a critical element in determining the transduction mechanism of this receptor. These results suggest that i2 of mGluRs may play a role similar to i3 of most other GPCRs in the specificity of coupling to the G-proteins. Moreover, as in many other GPCRs, our data revealed cooperation between the different mGluR intracellular domains to control efficient coupling to G-proteins.

tors activated by small ligands like catecholamines (1) and peptides (2,3). Their effector domains, the domains interacting with and activating the G-proteins, have been studied extensively (4,5). It has been proposed that their third intracellular loop (i3) plays a critical role for the activation of G-proteins (6 -12). However, the binding and selective recognition of the G-protein depends not only on the sequence of the i3 loop but is also influenced by other intracellular domains (13)(14)(15)(16)(17)(18)(19).
The cloning of 8 metabotropic glutamate receptors (mGluR1-8) and Ca 2ϩ -sensing receptors revealed the existence of a new GPCR family (20,21). Whereas mGluR1, mGluR5 (22)(23)(24), and the Ca 2ϩ -sensing receptors (25,26) activate phospholipase C, the other mGluRs inhibit adenylyl cyclase (27)(28)(29)(30)(31)(32). Although mGluRs also possess 7 putative TMD, they do not show sequence homology with the other GPCRs. Several lines of evidence suggest that mGluRs do not operate like the other GPCRs. Their large (about 600 residues) extracellular N-terminal domains are homologous to bacterial periplasmic binding proteins and constitute the agonist recognition domain (33,34). Their intracellular domains also show striking differences with the corresponding domains of the other GPCRs. For example, i3 is very short and highly conserved among mGluRs and Ca 2ϩsensing receptors (21), whereas it is the longest and more variable loop in all other GPCRs. The DRY (or ERW) tripeptide at the N-terminal end of i2 that is conserved in most GPCRs and plays a critical role in G-protein coupling (13,35) is absent in mGluRs. Nevertheless, and in spite of these differences, mGluRs and Ca 2ϩ -sensing receptors probably activate the same set of heterotrimeric G-proteins as all other GPCRs. The understanding of how mGluRs interact with and activate Gproteins will therefore help define the general rules for receptor-G-protein interaction.
In our preliminary analysis of rat mGluR domains involved in G-protein coupling selectivity, we only examined the role of i2 and the C-terminal tail of mGluR1 in the coupling to PLC (36). In contrast to what has been reported for many GPCRs in which only one short segment plays a critical role in G-protein coupling, we found that the simultaneous exchange of those two domains of the adenylyl cyclase-coupled mGluR3 with those of mGluR1 generated a chimeric receptor which coupled to PLC. However, the kinetics of the responses induced in Xenopus oocytes by this chimeric receptor differed from those of the wild type mGluR1, suggesting the involvement of other intracellular domains in G-protein coupling. The aim of the present work was therefore: 1) to examine whether only one, like in other GPCRs, of the two previously identified domains was critical for the PLC coupling and 2) to examine whether other intracellular domains (i1 and i3) could also influence the coupling of mGluR1 to PLC.

MATERIALS AND METHODS
Nomenclature of Chimeric Receptors-The chimeric receptor names are RX/Y-(A,B, . . . ), where X is the mGluR type in which small portions have been exchanged with those of mGluRY. The fragments exchanged are indicated in the parentheses and described in Fig. 2. For example R3/1-(i2) is a chimeric mGluR3 receptor containing the second intracellular loop of mGluR1. When necessary, the subtype of alternatively spliced mGluR is specified. For example, since the C-terminal region of mGluR1a, -b, and -c are different, R3/1a-(C2) means that the entire C-terminal end of mGluR3 has been replaced by that of mGluR1a.
The sequence of the domains exchanged between mGluR1 and mGluR3 is the following: i1 of mGluR1 611 (36). If necessary, making use of the redundancy of the genetic code, primers for some chimeras were designed based on known rat cDNA sequences to introduce diagnostic restriction enzyme cleavage sites. This allowed rapid screening for the mutant genotype. After construction, the segments of the chimeric receptor DNA generated by PCR were sequenced on both strands using 17-25 mere primers using the dideoxynucleotide method using Sequenase (U. S. Biochemical Corp.).
Expression into Xenopus Oocytes-The preparation of oocytes and the in vitro synthesis of RNA transcripts from the cloned cDNA were performed as described previously (37). Recordings were performed in Barth's medium using the two-electrode voltage-clamp technique (Axoclamp-2A) 3 days after injection. Data were recorded on a PC computer and analyzed using pclamp software.
Culture and Transfection of HEK 293 Cells-Culture and transfection of the human embryonic kidney cells HEK 293 were conducted essentially as described previously (38). Briefly, cells were cultured in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) supplemented with 10% fetal calf serum (complete medium). The cDNAs encoding the different chimeras were subcloned into an eucaryotic vector with a cytomegalovirus immediate-early promoter. Cells were transiently transfected by using 10 g of DNA as a Ca 2 PO 4 precipitate in 1.2 ml of 0.82% (w/v) NaCl, 0.6% (w/v) Hepes, pH 7.1, 0.02% (w/v) Na 2 HPO 4 , and 0.125 M CaCl 2 added to 10 ml of complete medium per 100-mm plate. Cells were exposed to DNA for 3 h in a 10% CO 2 incubator in order to form a proper precipitate. After glycerol shock by exposure to 15% (v/v) glycerol for 30 s, the cells were washed with PBS and incubated in complete medium for 1 h. Cells were plated on coverslips coated with poly-L-lysine (Sigma) and used after 24 -48 h for immunofluorescence assays or calcium measurements.
Immunofluorescence of Transfected Cells-The preparation, purification, and characterization of antibodies directed against synthetic peptides corresponding to the C-terminal ends of mGluR1a, mGluR1c, and mGluR3, respectively, have already been described (38,39).
24 h after transfection, HEK 293 cells grown on coverslips were washed three times in PBS, fixed for 15 min at room temperature in 3% paraformaldehyde in PBS, and washed three times in PBS. The fixed cells were then permeabilized with ice-cold methanol for 2 min. After washing four times with PBS, the cells were incubated for 1 h at room temperature in PBS containing 3% bovine serum albumin and rabbit anti-mGluR1a (1:500), anti-mGluR1c (1:1000), or anti-mGluR3 (1:2000) antibody. Cells were washed in PBS containing 3% bovine serum albumin, and bound primary antibodies were detected with a secondary antibody (1/200 dilution of fluorescein-labeled goat anti-rabbit IgG) for 30 min at room temperature. Cells were washed thoroughly, and the coverslips were mounted in 1,4-Diazabicyclo[2.2.2]octane (100 mg/ml) in PBS:glycerol (1:1). Fluorescence microphotographs were prepared on a Zeiss Axioskop microscope.
Calcium Measurements-Cells were incubated at room temperature for 30 min in a phosphate-buffered Krebs-Henseleit solution (Sigma) containing a 10 g/ml concentration of the calcium indicator fura-2/AM (Molecular Probes Inc.), 10 l/ml dimethyl sulfoxide, and 25% fetal calf serum. Glass coverslips carrying the dye-loaded cells were mounted into a perfused cuvette in a fluorescence spectrophotometer (F-4500, Hitachi). Changes in [Ca 2ϩ ] i were monitored by measuring the ratio of fura-2 fluorescence (510 nm) excited alternatively (1.6 Hz) at 340 and 380 nm. Valves operated by a digital timer were used to switch the perfusate from control solution to 1 mM glutamate (Glu)-containing solution.
Chemicals-Quisqualate and L-CCG-I were from Tocris Cookson (Essex, United Kingdom); glutamic acid (Glu) was from Sigma (France, L'Isle d'Abeau). 1S,3R-ACPD was synthesized by K. Curry (Vancouver, Canada). All other compounds were of the highest possible grade from commercial sources.

RESULTS
In our preliminary analysis of mGluR domains involved in G-protein coupling, we examined only the role of i2 and the C terminus of mGluR1c in determining the coupling to PLC (36). However, sequence alignment of the mGluR intracellular domains also revealed that several residues in i1 and i3 are conserved in adenylyl cyclase-coupled mGluRs, but different in PLC-coupled mGluRs. For example, Arg-618, Ser-625, and Ser-627 (numbers correspond to the mGluR1 sequence) in the first intracellular loop of mGluR1 and mGluR5, are replaced by Asn, Ala, and Gly, respectively, in the adenylyl cyclase-coupled mGluRs (Fig. 1). Moreover, the neutral residue Ala-779 in the middle of i3 in PLC-coupled mGluRs is replaced by a glutamic acid residue in all adenylyl cyclase-coupled mGluRs (Fig. 1). Finally, the C-terminal intracellular domains show little homology between mGluRs (21). A role for the C terminus in G-protein coupling is strengthened by the observation that mGluR1 splice variants, mGluR1a, mGluR1b, and mGluR1c, which have different C-terminal intracellular domains, are all coupled to PLC but have slightly different functional properties (37, 40) (see Fig. 2a). In mGluR1b and mGluR1c, the C-terminal 318 residues of mGluR1a are replaced by 20 and 11 residues, respectively (32,37). These observations suggested that additional intracellular domains may somehow be involved in the control of the interaction between mGluRs and the G-protein.
In order to analyze the exact role of the different intracellular domains of mGluR1 in its coupling to PLC, we exchanged the equivalent domains of the adenylyl cyclase-coupled mGluR3 with those of mGluR1a or mGluR1c. The coupling to PLC of the resulting chimeric receptors was first analyzed after transient expression in Xenopus oocytes. In these cells, it is well established that activation of PLC results in the activation of a Ca 2ϩ -activated chloride current due to the IP3-induced release of intracellular Ca 2ϩ (41). Not only the amplitude of the current was measured, but also the time needed to reach the maximal current after the response began (time-to-peak value). This later parameter has previously been shown to be a functional characteristic of a given receptor, not directly related to the level of expression of this receptor (7,11,37,42). This parameter can therefore be considered as a more accurate measure of the efficiency of coupling of the receptor to this transduction pathway in oocytes (7,11,37,(42)(43)(44).
As shown in Fig. 2b, none of the chimeric receptors that still possess the second intracellular loop of mGluR3 elicited consistent responses when expressed in Xenopus oocytes. Although responses were recorded from some oocytes expressing the chimeric receptor R3/1a-(i1,i3,C2) (see "Materials and Methods" and Fig. 2 for the nomenclature of chimeric receptors), these were very small and observed in only 8 out of 55 recorded oocytes. In contrast, many chimeric receptors that possess i2 of mGluR1 activated chloride current upon stimulation with 1 mM Glu (Fig. 2, c and d). Taken together, these results suggested that, in contrast to what has been reported for most GPCRs, i2 rather than i3 of mGluR1a plays a critical role in determining the PLC coupling of this receptor. To further support this conclusion, we constructed chimeric mGluR1a receptors with the second or third intracellular loops of mGluR3. The exchange of i3 in mGluR1a did not prevent the receptor from activating the chloride current when expressed in oocytes (Figs. 2e and 5), whereas no responses could be measured with the mGluR1a chimeric receptor with i2 of mGluR3.
However, exchanging i2 of mGluR3 with that of mGluR1 was not sufficient to generate a receptor able to activate a chloride current in Xenopus oocytes. Glu-induced responses could be recorded if, in addition to i2, at least another intracellular loop (i1 and/or i3) was simultaneously exchanged (Figs. 2c and 3c).
Chimeric receptors that contain i2 and the end of the C-terminal domain (C1 domain) of mGluR1a did not generate responses upon stimulation with Glu (Fig. 2c). However, the exchange of both i2 and the entire C-terminal intracellular domain of mGluR1a generated a chimeric receptor activating PLC in oocytes (Fig. 2c). In contrast to the responses induced by the wild type mGluR1a that were fast and transient (Figs. 2a  and 3a), responses generated by this chimeric receptor R3/1a-(i2,C2) had time-to-peak values often greater than 10 s (Figs. 2c and 3d). As shown in Fig. 2c and Fig. 3, e and f, the additional exchange of either i1 or i3 in R3/1a-(i2,C2) generated receptors that induced fast responses in oocytes, similar to those induced by the wild type mGluR1a, regardless of the maximal amplitude of the response (Fig. 3, e and f). This suggests that most intracellular domains of mGluR1a have to be present to recover the functional coupling to the G-protein of the wild type mGluR1a. Similarly, among the chimeric receptors with the C terminus of mGluR1c, only the R3/1c-(i1,i2,i3,C2) receptor that contains all intracellular domains of mGluR1c, induced responses similar to those generated by mGluR1c (Fig. 2d).
The pharmacological profile of R3/1c-(i2,C2) previously has been studied extensively, and full agonist dose-response curves were constructed (36). This revealed that the chimeric receptors had a pharmacology very similar to that of mGluR3. In the present study, we verified that the new R3/1 chimeric receptors constructed had a pharmacology similar to that of mGluR3, but distinct from that of mGluR1. As shown in Fig. 4, the potent agonist of mGluR1, quisqualate, did activate R1a/3-(i3) chimeric receptors, but not R3/1 receptors. In contrast, L-CCG-I which is a potent agonist at mGluR3, activated R3/1 chimeric receptors, but not R1a/3-(i3). Finally, 1S,3R-ACPD that activates both mGluR1 and mGluR3 activated all chimeric receptors tested.
The expression and coupling to Ca 2ϩ signaling of chimeric receptors was also examined in mammalian cells. After transient expression in HEK 293 cells, all chimeric and wild type  3. Analysis of the kinetics of  the responses induced by mGluR1a  (a), mGluR1c (b), and the chimeric  receptors R3/1-(i1,i2,i3) (c), R3/1a-(i2,C2) (d), R3/1a-(i1,i2,C2) (e), and R3/  1a-(i2,i3,C2)  receptors analyzed were found to be expressed at the plasma membrane level as illustrated by immunohistochemistry using antibodies directed against the C terminus of either mGluR1a (Fig. 5), mGluR3, or mGluR1c (data not shown). None of these antibodies gave immunostaining in mock-transfected cells or cells expressing a receptor with a different C terminus (Fig. 5  and data not shown). The chimeric receptors that did not activate chloride current in Xenopus oocytes did not induce Ca 2ϩ signal in HEK 293 cells when activated with 1 mM Glu (Fig. 6). In contrast, Glu induced clear Ca 2ϩ signals in HEK 293 cells expressing chimeric receptors that were functional in oocytes (Fig. 6). DISCUSSION The present study was aimed at examining the role of all intracellular domains of mGluR1a and mGluR1c in determin- ing their coupling to PLC-activating G-proteins. For that purpose, the different intracellular domains of the adenylyl cyclase-coupled mGluR3 were systematically exchanged by their mGluR1a or mGluR1c equivalent. Our results revealed that all chimeric receptors able to strongly activate the chloride current in Xenopus oocytes, or to induce Ca 2ϩ signals in transfected HEK 293 cells, possess the second intracellular loop of mGluR1. The importance of i2 in determining the transduction mechanism of mGluR1 was strengthened by showing that the exchange of i2 of mGluR1a with that of mGluR3 abolishes its coupling to PLC. Although the absence of a radioactive ligand with high enough affinity prevented us from estimating the receptor density in the transfected cells, immunofluorescence studies indicated that all chimeric receptors including those that did not show a Glu-induced Ca 2ϩ signal were properly targeted to the plasma membrane of HEK 293 cells. Whether some of these receptors retained their ability to inhibit adenylyl cyclase has not been examined yet.
These data indicate that, within the intracellular domains of mGluR1, i2 plays a critical role in the activation of PLC-coupled G-protein. Considering this and our previous studies (36), our data indicate that, within i2, the 16 C-terminal residues are necessary for optimum coupling to PLC. That such a short segment plays a critical role in G-protein coupling in mGluRs is supported by our recent observation that only few residues in the ␣ subunit of G-proteins play a critical role in their selective interaction with mGluRs. As observed with other G i -coupled GPCRs (45), G i -coupled mGluRs can activate PLC when coexpressed with a chimeric G␣ q subunit that has its 5 C-terminal residues replaced by those of G␣ i2 . 2 Finally, it is interesting to note that the C-terminal portion of i2 is likely to fold into an amphiphilic ␣ helix (36). Such a secondary structure has been proposed to be important for the receptor domains involved in G-protein activation (46 -51), although this may not be a general rule (52).
As previously reported, the exchange of the second intracellular loop of mGluR3 with that of mGluR1 is not sufficient to obtain a chimeric receptor able to activate the chloride current in Xenopus oocytes. The additional exchange of another intracellular domain is necessary. We previously reported that the additional exchange of the C terminus of mGluR1c was sufficient to allow the chimeric receptor to activate PLC in oocytes (36). Similar results were obtained with the chimeric receptor with i2 and the entire C-terminal intracellular domain of mGluR1a. Our data also revealed that the simultaneous exchange of i2 and either i3 or i1 plus i3 also generated chimeric receptors able to activate chloride currents in oocytes, even though the C-terminal domain was not exchanged. Therefore, all intracellular domains of mGluRs play a role in G-protein activation by facilitating the action of i2. However, none of the specific residues found in i1, i3, or the C-terminal intracellular domain of PLC-coupled mGluRs are absolutely required for the activation of PLC. In agreement with this conclusion, the sequences of i1 and i3 of the Ca 2ϩ -sensing receptors (that activate PLC) are almost identical with those of adenylyl cyclase-coupled mGluRs (Fig. 1).
As reported for other receptors expressed in Xenopus oocytes (7,11,37,(42)(43)(44), the responses elicited by the R3/1 chimeric receptors can be either fast (small time to peak values) or more slowly generated (larger time to peak values). This difference is observed whatever the maximal amplitude of the responses and is therefore unrelated to the level of expression of the receptors. The kinetic of the response appears therefore as a good measure of the efficacy of coupling of a given receptor to this transduction cascade (7,11,44). Interestingly, most of the intracellular domains have to be present in the R3/1 chimeras to recover the time to peak value of the corresponding wild type receptor (mGluR1a or mGluR1c). Taken together, these results suggest that the coupling efficacy of mGluRs is determined by the combination of all their intracellular domains. Accordingly, a cooperation between several intracellular domains in controlling either the specificity or the efficacy of coupling to Gproteins has already been reported for other GPCRs (8,14,16,19).
In conclusion, our results indicate that i2 of mGluR1 plays a critical role in the G-protein coupling selectivity of this receptor and may therefore be regarded as the equivalent of i3 of most other GPCRs. We show, moreover, that the coupling and activation of the G-protein involve a cooperation between all intracellular domains of mGluR1. Therefore, although the mGluR share no sequence homology with the other GPCRs, similar mechanisms have been selected during evolution for their coupling to G-proteins.