Báo cáo khoa học: Allosteric functioning of dimeric class C G-protein-coupled receptors doc

MINIREVIEW Allosteric functioning of dimeric class C G-protein-coupled receptors J-P. Pin 1–5 , J. Kniazeff 1–5 ,J.Liu 1–5 , V. Binet 1–5 , C. Goudet 1–5 , P. Rondard 1–5 and L. Pre ´ zeau 1–5 1 Institut de Ge ´ nomique Fonctionnelle, Montpellier, France 2 CNRS, UMR5203, Montpellier, France 3 INSERM, Montpellier, France 4 Universite ´ Montpellier-I, France 5 Universite ´ Montpellier-II, France Most membrane receptors, including ligand-gated channels, tyrosine kinase receptors, cytokine receptors and guanylate cyclase receptors form oligomers. This was rapidly recognized as being crucial for the func- tioning of these receptors. In the case of ligand-gated channel receptors, association of 4–5 subunits is required to form an ion channel. In the case of recep- tors that have a single transmembrane domain, it was difficult to imagine how the signal could be transduced from the extracellular to the intracellular side of the membrane without subunit association. In that case, it was rapidly proposed that ligand binding in the extra- cellular domain induces receptor dimerization, allowing the associated intracellular enzymatic domains to inter- act and become activated. More recent data from the determination of the three-dimensional structure of the extracellular domains of such receptors with and with- out agonists, revealed that they can even be consti- tutive dimers, agonists stabilizing a specific active conformation of the dimer [1,2]. In contrast, all G-protein-coupled receptors (GPCRs) have a large membrane core domain com- posed of seven transmembrane-spanning helices, which is responsible, in most cases, for both ligand recogni- tion and activation of the intracellular effector, i.e. the heterotrimeric G-protein. This, plus other biophysical data, lead to the conclusion that GPCRs work as monomers that can oscillate between various confor- mations, the active conformations being stabilized by agonists, whereas the fully inactive conformations are stabilized by inverse agonists. However, it was difficult to explain some cooperativity phenomena observed in Keywords activation mechanism; allosteric modulators; dimerization; GPCR Correspondence J-P. Pin, Institut de Ge ´ nomique Fonctionnelle, 141 rue de la Cardonille, F-34094 Montpellier cedex 5, France Fax: +33 467 54 2432 Tel: +33 467 14 2988 E-mail: jppin@ccipe.cnrs.fr (Received 16 February 2005, accepted 6 April 2005) doi:10.1111/j.1742-4658.2005.04728.x Whereas most membrane receptors are oligomeric entities, G-protein- coupled receptors have long been thought to function as monomers. Within the last 15 years, accumulating data have indicated that G-protein-coupled receptors can form dimers or even higher ordered oligomers, but the gen- eral functional significance of this phenomena is not yet clear. Among the large G-protein-coupled receptor family, class C receptors represent a well- recognized example of constitutive dimers, both subunits being linked, in most cases, by a disulfide bridge. In this review article, we show that class C G-protein-coupled receptors are multidomain proteins and highlight the importance of their dimerization for activation. We illustrate several consequences of this in terms of specific functional properties and drug development. Abbreviations Acc, active-closed-closed conformation; Aco, active-closed-open conformation; CaS, receptor, calcium-sensing receptor; CRD, cystein-rich domain; ER, endoplasmic reticulum; HD, heptahelical domain; mGlu, receptor, metabotropic glutamate receptor; Roo, resting-open-open conformation; T1R: taste receptor type 1; VFT, Venus flytrap domain. FEBS Journal 272 (2005) 2947–2955 ª 2005 FEBS 2947 ligand binding. This led to the demonstration that most GPCRs can oligomerize as shown by both bio- chemical and energy transfer technologies [3]. In recent years, several publications have indicated that this phe- nomenon is involved in trafficking of the receptor to and from the plasma membrane, and in specific cross- talk between receptor subtypes [4]. However, the pre- cise role and importance of GPCR oligomerization in the activation process remains unknown. Five main classes of GPCRs can be defined in mam- mals based on sequence similarity [5–7]. Whereas the large number of rhodopsin-like receptors form class A, secretin-like and metabotropic glutamate (mGlu)-like receptors are members of classes B and C, respectively. Frizzled receptors and a subgroup of pheromone receptors form two additional classes. Class C GPCRs have been shown to be constitutive dimers and therefore represent a good model for studying the functional rele- vance of GPCR dimerization. These receptors include those for the main neurotransmitters, glutamate and GABA, as well as a receptor activated by extracellular Ca 2+ , some pheromone receptors and receptors for the sweet and umami taste compounds [8]. In this review article, we summarize our knowledge on the functioning of class C GPCRs and illustrate how allosteric inter- actions between the subunits play a fundamental role in their activation. Of interest, we see that this com- plex functioning of class C receptors offers a number of possibilities to regulate their activity with synthetic ligands acting at sites different from the natural ligand- binding site, the so-called allosteric modulators. The multiple domains of class C GPCRs In contrast to most class A rhodopsin-like GPCRs, class C receptors are composed of three main struc- tural domains, not including the C-terminal tail which can be very long (up to 376 residues for mGlu5b) and where a multitude of intracellular scaffolding and sig- nalling molecules bind. These domains are the Venus flytrap domain (VFT), which contains the agonist- binding site, the cysteine-rich domain (CRD) and the heptahelical domain (HD) involved in G-protein acti- vation (Fig. 1). The VFT module is a bilobate domain that shares structural similarity with bacterial periplasmic amino acid-binding proteins. The structure of the mGlu1 VFT has been solved by X-ray crystallography in the absence and presence of either agonist or antagonist [9,10]. These studies revealed that both types of ligand bind in the cleft that separates both lobes. As already shown for bacterial proteins, these studies also revealed that the VFT of class C GPCRs can adopt either an open or a closed conformation (Fig. 1). Inter- estingly, both conformations have been seen in the absence of ligand, as well as in the presence of agon- ists. In contrast, only the open conformation was observed with bound antagonist. It was therefore pro- posed that the VFT can naturally oscillate between these two states, the closed state being stabilized by agonists, whereas antagonists prevent the closure. Further studies performed on full-length receptors confirmed this functioning of the VFT. For example, by removing steric or ionic hindrance that prevents mGlu8 VFT closing upon antagonist binding, two antagonists were converted into full agonists [11]. Moreover, the introduction of two cysteine residues that are expected, based on modelling studies, to cross-link both lobes of the GABA B1 receptor and lock it in a closed state, generates a fully constitutively active receptor [12]. The CRD links the VFT to the HD in most class C GPCRs. The structure of this CRD is not known although a three-dimensional model has been proposed recently [13] (Fig. 1). Although the CRD is absent in the GABA B receptor subunits, it appears necessary for the activation of either mGlu or calcium-sensing (CaS) receptors [14], but its specific mode of action is not yet known. Like any other GPCRs, class C receptors possess a HD that shares very low sequence similarity with rho- dopsin-like receptors (Fig. 1). Indeed, few residues are conserved in these two groups of receptors and model- ling studies suggest that both types of HD share a similar structure [8]. As in class A receptors, the intra- cellular loops of class C GPCRs as well as the C-terminal tail are involved in G-protein coupling. For various class C GPCRs, including the mGlu5, GABA B2 and CaS receptors, the HD can fold correctly and be trafficked to the cell surface when expressed alone after deletion of both the large extracellular domain and the long C-terminal tail [15–17]. More- over, these isolated HDs retain their ability to activate G-proteins as illustrated by their constitutive activity, an activity that can either be inhibited by inverse agon- ists known to bind in the HD, or further stimulated by other molecules known as positive allosteric modula- tors. Accordingly, the HD of class C GPCRs appears to behave like rhodopsin, oscillating between various states each being possibly stabilized by specific com- pounds (Fig. 1). In summary, class C GPCRs are multimodule pro- teins and both major modules (the agonist-binding VFT and the G-protein-activating HD) retain their specific functional properties when isolated. As expec- ted for allosteric proteins, these modules can oscillate between various states, each being stabilized by specific Class C G-protein-coupled receptors J-P. Pin et al. 2948 FEBS Journal 272 (2005) 2947–2955 ª 2005 FEBS molecules. However, how can the ligand-binding domain control the activity of the HD? In other words, how is the signal transduced from one domain to the other? Class C GPCRs are constitutive dimers An important piece of information to understand the activation process of class C GPCRs came with the discovery that these receptors are constitutive dimers. The first observation came from the mGlu5 receptor, which was shown in western blot and immunopreci- pitation experiments to be a homodimer in both transfected cells and native tissue [18]. Only upon di- thiothreitol treatment was the monomeric form detec- ted. Soon after, the CaS receptor was also shown to form dimers stabilized by a disulfide bridge via Cys129 located in the VFT [19], and this was confirmed in both mGlu1 and mGlu5 receptors [20]. Because this residue is conserved in all mGlu receptors, as well as in the taste and pheromone receptors, these are also expected to be disulfide-linked dimers. Mutation of this Cys residue does not prevent dimer formation [21]. Indeed, the VFT, even when produced as a soluble protein, forms stable dimers via a hydrophobic surface area located on one side of lobe-I, as clearly revealed in the crystal structure of the dimers of mGlu1 VFTs [9,10] (Fig. 2A). Mutation of the Cys residue involved in the covalent linkage of the subunit also does not affect functioning of the receptor [22]. Although the role of this disulfide bridge remains elusive, it certainly prevents any possible dissociation of the subunits under normal conditions, making these receptors con- stitutive dimers. To date, no heterodimeric mGlu receptors have been described. Only mGlu1–CaS heterodimers have been open closed VFT CRD HD HD* HD HDg Fig. 1. The main domains of class C GPCRs and their various conformational states. Class C GPCRs are composed of three main structural domains, the Venus flytrap domain (FVT) where agonists and competit- ive antagonists bind, the cysteine-rich domain (CRD) that interconnects the VFT to the heptahelical domain (HD), and HD, which if similar to rhodopsin-like GPCRs. Each structural domain is shown in a ribbon view. Both the VFT and HD are coloured according to the succession of secondary structure elements from dark blue (N-termi- nus) to red (C-terminus). Both the open unliganded and agonist-bound closed confor- mation of the VFT are shown. The three expected conformational states for the HD are indicated, as also proposed for the rhodopsin-like GPCRs: HDg, ground totally inactive state; HD, basal state; HD*, fully active state. The ribbon views were gener- ated using the coordinates of the mGlu1 VFT (protein data bank Accession nos 1EWT:A and 1EWK:A, respectively), the pro- posed model of the CRD, and the coordi- nates of rhodopsin (protein data bank Accession no. 1F88). J-P. Pin et al. Class C G-protein-coupled receptors FEBS Journal 272 (2005) 2947–2955 ª 2005 FEBS 2949 observed [23], but more work is required to validate their functional and physiological relevance. However, the related taste receptors need to heterodimerize to form functional receptors. The association of taste receptor type 1 (T1R1) and taste receptor type 3 (T1R3) results in the formation of umami receptors [24], whereas taste receptor type 2 (T1R2) and T1R3 constitute the sweet receptors [25]. Although not observed in heterologous expression systems, T1R3 may also be able to form a functional low-affinity sweet receptor in the absence of T1R1 and T1R2 [26]. In contrast to the other class C GPCRs, the GABA B receptor is not a disulfide-linked dimer. However, this receptor was the first GPCR identified as an obligatory heterodimer composed of two distinct subunits, GABA B1 and GABA B2 [27]. During evolution, a sys- tem has been selected to ensure that only the func- tional heterodimer reaches the cell surface. Indeed, the GABA B1 subunit contains an endoplasmic reticulum (ER) retention signal in its intracellular tail, preventing it from reaching the surface alone [28]. Only when associated with GABA B2 can this subunit reach the cell surface and be functional. Although no covalent linkage between the subunits has been observed, these dimers are likely very stable due to a coiled coil inter- action at the level of their intracellular tail, as well as by direct interaction of their VFTs and also likely their HDs [29]. These observations revealed that class C GPCRs are complex multidomain molecules and raised an import- resting active Lobe-I VFTs Lobe-II HDs A B Fig. 2. General structure of dimeric class C GPCRs. (A) Ribbon view of the crystal struc- ture of the resting Roo (left, pdb Accession no. 1EWT) and fully active Acc (right, pdb Accession no. 1ISR) state of the mGlu1 VFT dimer, and apposition of two rhodopsin structures. The yellow subunit is in the front, whereas the blue subunit is in the back. Note the difference in the relative ori- entation of the two VFTs probably leading to a different mode of association of the two HDs within the dimer. (B) Scheme illustra- ting that agonist binding in one VFT can activate the HD of the same subunit (cis-activation) and ⁄ or the HD of the other subunit (trans-activation). In the wild-type heterodimeric GABA B receptor only trans- activation occurs (agonist binding in the GABA B1 VFT leads to the activation of the GABA B2 HD), but both cis- and trans-activa- tion occur in the homodimeric mGlu recep- tors. Class C G-protein-coupled receptors J-P. Pin et al. 2950 FEBS Journal 272 (2005) 2947–2955 ª 2005 FEBS ant issue: the interplay between the various states of each domain in the dimer, and how this can be con- trolled by agonists. Activation mechanism of class C GPCRs involves allosteric interaction between the VFTs As described above, the mGlu1 VFT can reach a closed state stabilized by agonists, and form dimers via a hydrophobic area on one side of its lobe-I [9,21]. This contact between the VFTs is likely required for receptor activation, because a point mutation in that area results in a loss of function of the receptor, even though agonist binding can still be measured [30]. Comparison of the crystal structure of the VFT dimer in the absence or presence of glutamate also revealed a major change in the relative orientation of the two VFTs [9]. In a first orientation, lobe-IIs are far apart in the absence of agonist or in the presence of antag- onist. This orientation is, therefore, called ‘resting’. A second orientation is observed in the presence of agon- ist and is therefore considered active. In that case, lobe-IIs are in close contact and one VFT is closed, whereas the other remains open. More recently, a structure has been solved in the presence of both agon- ist and Gd 3+ [10]. In that case, the same active orien- tation is observed, but both VFTs are in a closed state (Fig. 2A). These data illustrate that the dimer of mGlu VFTs can have at least three conformations: the resting-open-open (Roo, resting orientation with both VFTs in an open state), the asymmetric active-closed- open (Aco) and the symmetric active-closed-closed (Acc) conformations. How can agonist binding affect the relative orienta- tion of the VFTs? Much can be deduced from analysis of the interface between the subunits at the level of lobe-II when both VFTs are maintained in the active orientation. This interface revealed major charge repul- sion if both VFTs are open, consistent with the great instability of this form of the dimer (note this is deduced from modelling studies, because this form of the receptor has never been observed) [10]. In contrast, in the Aco state, the interface consists of a number of ionic interactions between the two subunits. Finally, when both VFTs are closed (Acc state), four acidic side chains are facing each other, creating a cation- binding site that likely needs to be occupied for this state to be stable [10]. We recently examined whether both Aco and Acc conformations lead to similar properties of the dimeric mGlu receptor [31]. To that aim, we used the quality- control system of the GABA B receptor to generate mGlu receptor dimers composed of two distinct bind- ing sites, either from two distinct mGlu receptors or from a wild-type and a mutated VFT. This allowed us to show that a single ligand per dimer stabilized the Aco conformation, leading to partial activation of the receptor (Fig. 3A). Only upon binding of two agonists per dimer was the Acc state reached, leading to full activity [31]. Of interest, this fully active state is further stabilized by cations such as Ca 2+ or Gd 3+ . Although two glutamates bind in a dimeric mGlu receptor, no strong cooperativity could be measured by analysing the Hill coefficient. However, functional analysis suggests a positive cooperativity between both sites. Indeed, agonist potency is 3–5 times lower in a receptor dimer that possesses a single wild-type site. Moreover, our data also revealed that when one VFT A B Fig. 3. Activation mechanism of homodimeric class C GPCRs and its regulation by allosteric modulators. (A) In the absence of agon- ist, the receptor is in a resting state (Roo-HD), and switches to a partially active state upon binding of a first agonist [Aco-HD (*) ], and to a fully active state upon binding of a second agonist (Acc-HD*). Binding of an inverse agonist in the HD stabilizes the fully inactive ground state of the receptor, whereas binding of a positive allo- steric modulator further stabilizes the fully active state of the agon- ist-bound dimer. (B) Schematic representation of the functioning of the class C GPCR after deletion of the large extracellular domain, and illustrating the main three states: the basal state HD that can generate basal activity of the receptor, the ground inactive state HDg stabilized by inverse agonists, and the fully active state stabil- ized by positive allosteric modulators. J-P. Pin et al. Class C G-protein-coupled receptors FEBS Journal 272 (2005) 2947–2955 ª 2005 FEBS 2951 is in the closed state, it stabilizes the associated VFT in the closed state. Such observations are in contrast to the negative allosteric interaction reported between the mGlu1-binding sites using binding experiments on purified and soluble VFTs [32]. However, this may be explained by the absence of the other part of the receptor (the CRD and the HD), as well as by the absence of cations that stabilize the Acc state. Although two agonists per dimer are required for full activation of homodimeric class C GPCRs, a single agonist is sufficient to fully activate the heterodimeric receptors. This has been demonstrated in the case of the GABA B receptor in which GABA binds in the GABA B1 VFT only [33]. Surprisingly, although the GABA B2 subunit also possesses a VFT, no natural ligand probably binds in this domain, as illustrated by the absence of selective conservation of residues in the putative binding pocket during evolution. Even though the GABA B2 VFT does not bind GABA, it is necessary for GABA B receptor activation. Indeed, among the various combinations of GABA B1 –GABA B2 subunit chimera generated, only those possessing both the GABA B1 and GABA B2 VFTs display agonist-induced activity [34]. This is consistent with the proposal that a change in the relative orientation of the VFTs in the dimer is associated with receptor activation. As shown for the mGlu receptors, isolated GABA B1 and GABA B2 VFTs form dimers (heterodimers in that case), and this increases affinity for agonists but not for antagonists [29]. This effect likely results from a stabilization of the closed state of the agonist-bound GABA B1 VFT by the GABA B2 VFT, a proposal that is reminiscent to the positive allosteric coupling between the VFTs of mGlu receptors described above. Although closure of the GABA B1 VFT is sufficient to fully activate the recep- tor, whether the associated GABA B2 VFT also has to reach a closed empty form remains unknown. As observed in the GABA B receptor heterodimer, a single agonist is also likely to be sufficient to activate the sweet and umami taste receptors, the sweeteners aspartame and neotame interacting in the T1R2 VFT of the sweet taste T1R2 : T1R3 heterodimer, whereas glutamate binds in the T1R1 VFT in the umami taste T1R1 : T1R3 heteromer [35]. However, in contrast to the GABA B2 subunit, the T1R3 VFT-binding site is very well conserved during evolution, suggesting that natural ligands bind in this subunit also. Such ligand remains to be identified, but may likely act in synergy with the oligosaccharides and glutamate. In summary, interaction between VFTs is crucial for class C GPCR activation. Although agonist binding stabilizes the closed state of the bound VFT, this does not correspond to the major difference in the resting and active conformation of the VFT dimer. Indeed, whether one or two ligands interact in this dimeric unit, the main consequence is the stabilization of a new relative orientation of the VFTs. But how is this transmitted to the HDs within the dimer? Allosteric coupling between the extracellular and heptahelical domains within the dimer Whether agonist binding interacting in one VFT of the dimer activates the HD of the same subunit and ⁄ or that of the associated subunit has been carefully exam- ined in both heterodimeric GABA B and homodimeric mGlu receptors (Fig. 2B). In the case of the GABA B receptor, it was soon observed that the GABA B1 subunit could not activate the G-protein even when its ER retention signal was mutated [28,36]. As such it was soon proposed that the GABA B2 subunit was responsible for G-protein activa- tion. This was firmly demonstrated in several ways. First, mutations into either the i2 or i3 loop of GABA B2 suppressed G-protein activation by the het- erodimer, whereas the equivalent mutation in GABA B1 had a minor effect [37,38]. Second, a receptor combi- nation composed of the VFTs of both GABA B1 and GABA B2 , but of two HDs from GABA B2 , can activate G-proteins upon agonist application, although with a much lower efficacy than the heterodimer, demonstra- ting that the HD of GABA B2 possesses enough of the molecular determinants required for G-protein coup- ling [34]. Finally, it has recently been shown that this GABA B2 HD expressed alone can be activated by CGP7930 [17], a positive allosteric modulator of the GABA B receptor. It was therefore concluded that trans-activation occurs in the GABA B receptor, GABA binding in the GABA B1 VFT leading to activation of the GABA B2 HD. Although GABA B1 VFT binds the agonist and the GABA B2 HD couples to G-protein, a chimeric con- struct composed of these two domains cannot be acti- vated by agonists when expressed alone [34]. Normal functioning can be restored when such a chimeric con- struct is coexpressed with the reverse chimera bearing the GABA B2 VFT and the GABA B1 HD, demonstra- ting the importance of dimer formation for function. Of interest, note that in the case of this combination of chimeric subunits, cis-activation occurs, because the agonist binding domain and the G-protein coupling domain are part of the same subunit. Coupling between ligand binding and HD activation has also been recently examined in the homodimeric mGlu receptors. As described above, by manipulating Class C G-protein-coupled receptors J-P. Pin et al. 2952 FEBS Journal 272 (2005) 2947–2955 ª 2005 FEBS each subunit in a receptor dimer, it was shown that the monoliganded dimer of VFTs in the Aco confor- mation led to partial activity, whereas the Acc confor- mation with two bound agonists led to a full activity of the receptor [31]. By examining the effect of a point mutation known to prevent G-protein activation in the i3 loop of either HD, it was shown that both the Aco and Acc conformations of the VFT dimer activate either one or the other HD [31]. This demonstrates that both cis- and trans-activation occur in homo- dimeric mGlu receptors (Fig. 2B). Taken together, these data highlight the need for dimer formation for the signal transmission from the VFT to the HD, and also show that in homodimeric receptors, the signal from one VFT can be transmitted to either HD. These observations fit nicely with the proposal that the stabilization of a specific relative orientation of the VFTs by agonists, also stabilizes a specific association of the HDs leading to their activa- tion (Fig. 3A). Such a proposal is supported by recent data obtained using a FRET approach and showing a specific change in the general conformation of the HD dimer upon receptor activation [39]. Allosteric functioning of the HD of class C GPCRs As observed for class A GPCRs, some class C recep- tors display constitutive, agonist-independent activity. As described above, because the VFTs have the ability to close in the absence of agonist, spontaneous closure may well be at the origin of constitutive activity in some of these receptors, as observed for the GABA B receptor [40]. Indeed, in that case, competitive antago- nists act as inverse agonists by preventing the sponta- neous closure of GABA B1 VFT. However, in the case of the mGlu1 and mGlu5 receptors, their constitutive activity was not inhibited by competitive antagonists, demonstrating that their HD can reach an active state even when the VFTs stay open. This was further dem- onstrated in two ways. First, noncompetitive mGlu1 and mGlu5 antagonists known to bind directly in the HD of these receptors were found to have inverse agonist properties [41,42]. Second, the HD of mGlu5 expressed alone (mGlu5 receptor deleted of its large extracellular domain) was found to display the same constitutive activity as the full-length receptor, an activity that can be inhibited by inverse agonists bind- ing in this domain [16]. In addition to the noncompetitive antagonists, posit- ive allosteric modulators of class C GPCRs also bind in the HD [43]. In most cases, these compounds do not have agonist activity, but potentiate both the effic- acy and the potency of agonists. However, when the large extracellular domain was deleted from the recep- tor these compounds act as full agonists [16], and are therefore able to stabilize a new fully active conforma- tion of the HD (Fig. 3B). As such, as observed with rhodopsin, the HD of class C GPCRs can exist in at least three major states: an HDg (ground) state, which corresponds to the totally inactive state stabilized by inverse agonists; an HD state, which is able to activate G-proteins although with a low efficacy (this state being responsible for the constitutive activity of some receptors); and an HD* state, which corresponds to the active state of the receptor stabilized by positive allosteric modulators (Fig. 3B). Why are the mGlu5 positive modulators unable to activate the full-length receptor although they can fully activate an isolated HD? This indicates that the HD* state cannot be reached if the VFT dimer is not in the active orientation. This suggests that the HD* state is likely associated with a specific orientation of the HDs in the dimer that can be reached when the extracellular parts of the subunits are deleted (Fig. 3A). Taken together, these observations show that the HD of class C GPCRs can oscillate between various conformational states, each being stabilized either by synthetic ligand directly interacting in this domain, or by specific conformations of the VFT dimer. Conclusion Although class C GPCRs appeared to be more com- plex proteins than the class A receptors, because of their multiple domains and their association into con- stitutive dimers, much information on their activation process has been gained in recent years. These findings illustrate the importance of allosteric transition between various conformations of each domain. These transitions can be summarized as follow. The extracel- lular binding domains (the VFTs) can oscillate between an open and a closed conformation, the latter being stabilized by agonists. The relative orientation of the VFTs also oscillate between at least two positions, the resting ‘R’ orientation, and the active ‘A’ orientation, the latter being stabilized when at least one VFT is in a closed conformation, and further stabilized if both VFTs are closed. The HDs can also exist in at least three states, the HD state responsible for the constitu- tive activity of some receptors, the fully inactive state HDg stabilized by inverse agonist, and the fully active state HD* stabilized by the active form of the dimer of VFTs (the Acc conformation). Such complex functioning of these receptors offers a number of possibilities for allosterically regulating J-P. Pin et al. Class C G-protein-coupled receptors FEBS Journal 272 (2005) 2947–2955 ª 2005 FEBS 2953 their activity using compounds acting at various sites of the receptor. One such possibility is to further sta- bilize the closed state of the VFT after agonist binding. Such a possibility has been proposed for the positive allosteric effect of Ca 2+ on the GABA B receptor [44]. Another possibility is to stabilize the Acc conforma- tion of the dimer of VFTs, as seen with Gd 3+ in the mGlu receptors [10]. As already reported for many class C GPCRs, compounds directly interacting with the central pocket of the HD also stabilize a specific conformation of this domain and affect functioning of the receptor (acting as inverse agonists or positive modulators), but other possibilities exist, such as mole- cules acting at the contact interface between the HDs. Eventually, although the specific role of the CRD in the activation process is not known, compounds acting at this level may also influence functioning of the receptor. In support of this idea, large sweet proteins such as brazzein appear to contact the CRD of the T1R3 receptor subunit [45]. Accordingly, class C GPCRs represent good targets for drug development not only because of their important physiological roles, but also because the large number of possibilities for regulating their activity. References 1 He X-L, Chow D-C, Martick MM & Garcia KC (2001) Allosteric activation of a spring-loaded natriuretic pep- tide receptor dimer by hormone. Science 293, 1657– 1662. 2 Livnah O, Stura EA, Middleton SA, Johnson DL, Jolliffe LK & Wilson IA (1999) Crystallographic evidence for preformed dimers of erythropoietin recep- tor before ligand activation. Science 283, 987–990. 3 Bouvier M (2001) Oligomerization of G-protein-coupled transmitter receptors. Nat Rev Neurosci 2, 274–286. 4 Terrillon S & Bouvier M (2004) Roles of G-protein- coupled receptor dimerization. EMBO Rep 5, 30–34. 5 Kolakowski LF (1994) GCRDb: a G-protein-coupled receptor database. Receptors Channels 2, 1–7. 6 Fredriksson R, Lagerstrom MC, Lundin LG & Schioth HB (2003) The G-protein-coupled receptors in the human genome form five main families. Phylogenetic analysis, paralogon groups, and fingerprints. Mol Phar- macol 63, 1256–1272. 7 Bockaert J & Pin J-P (1999) Molecular tinkering of G-protein coupled receptors: an evolutionary success. EMBO J 18, 1723–1729. 8 Pin J-P, Galvez T & Prezeau L (2003) Evolution, struc- ture and activation mechanism of family 3 ⁄ C G-protein coupled receptors. Pharmacol Ther 98 , 325–354. 9 Kunishima N, Shimada Y, Tsuji Y, Sato T, Yamamoto M, Kumasaka T, Nakanishi S, Jingami H & Morikawa K (2000) Structural basis of glutamate recognition by a dimeric metabotropic glutamate receptor. Nature 407, 971–977. 10 Tsuchiya D, Kunishima N, Kamiya N, Jingami H & Morikawa K (2002) Structural views of the ligand-bind- ing cores of a metabotropic glutamate receptor com- plexed with an antagonist and both glutamate and Gd 3+ . Proc Natl Acad Sci USA 99, 2660–2665. 11 Bessis A-S, Rondard P, Gaven F, Brabet I, Triballeau N, Pre ´ zeau L, Acher F & Pin J-P (2002) Closure of the Venus flytrap module of mGlu8 receptor and the activa- tion process: insights from mutations converting antago- nists into agonists. Proc Natl Acad Sci USA 99, 11097– 11102. 12 Kniazeff J, Saintot P-P, Goudet C, Liu J, Charnet A, Guillon G & Pin J-P (2004) Locking the dimeric GABA B G-protein coupled receptor in its active state. J Neurosci 24, 370–377. 13 Yu L, Liang S, Liu X, He Q, Studholme DJ & Wu Q (2004) NCD3G: a novel nine-cysteine domain in family 3 GPCRs. Trends Biochem Sci 29, 458–461. 14 Hu J, Hauache O & Spiegel AM (2000) Human Ca 2+ receptor cysteine-rich domain. Analysis of function of mutant and chimeric receptors. J Biol Chem 275, 16382–16389. 15 Ray K & Northup J (2002) Evidence for distinct cation and calcimimetic compound (NPS 568) recognition domains in the transmembrane regions of the human Ca 2+ receptor. J Biol Chem 277, 18908–18913. 16 Goudet C, Gaven F, Kniazeff JC, Liu J, Cohen- Gonsaud M, Acher F, Prezeau L & Pin JP (2004). Heptahelical domain of metabotropic glutamate receptor 5 behaves like rhodopsin-like receptors. Proc Natl Acad Sci USA 101, 378–383. 17 Binet V, Brajon C, Le Corre L, Acher F, Pin JP & Pre ´ zeau L (2004) The heptahelical domain of GABA B2 is directly activated by CGP7930, a positive allosteric modulator of the GABA B receptor. J Biol Chem 279, 29085–29091. 18 Romano C, Yang W-L & O’Malley KL (1996) Meta- botropic glutamate receptor 5 is a disulfide-linked dimer. J Biol Chem 271, 28612–28616. 19 Ray K, Hauschild BC, Steinbach PJ, Goldsmith PK, Hauache O & Spiegel AM (1999) Identification of the cysteine residues in the amino-terminal extra- cellular domain of the human Ca 2+ receptor critical for dimerization. Implications for function of mono- meric Ca 2+ receptor. J Biol Chem 274, 27642–27650. 20 Ray K & Hauschild BC (2000) Cys-140 is critical for metabotropic glutamate receptor-1 (mGluR-1) dimeriza- tion. J Biol Chem 275, 34245–34251. 21 Tsuji Y, Shimada Y, Takeshita T, Kajimura N, Nomura S, Sekiyama N, Otomo J, Usukura J, Nakanishi S & Jingami H (2000) Cryptic dimer interface and domain organization of the extracellular region of metabotropic Class C G-protein-coupled receptors J-P. Pin et al. 2954 FEBS Journal 272 (2005) 2947–2955 ª 2005 FEBS glutamate receptor subtype 1. J Biol Chem 275, 28144– 28151. 22 Romano C, Miller JK, Hyrc K, Dikranian S, Menner- ick S, Takeuchi Y, Goldberg MP & O’Malley KL (2001) Covalent and noncovalent interactions mediate metabotropic glutamate receptor mGlu5 dimerization. Mol Pharmacol 59, 46–53. 23 Gama L, Wilt SG & Breitwieser GE (2001) Heterodi- merization of calcium sensing receptors with metabotro- pic glutamate receptors in neurons. J Biol Chem 276, 39053–39059. 24 Nelson G, Chandrashekar J, Hoon MA, Feng L, Zhao G, Ryba NJ & Zuker CS (2002) An amino-acid taste receptor. Nature 416, 199–202. 25 Nelson G, Hoon MA, Chandrashekar J, Zhang Y, Ryba NJP & Zuker CS (2001) Mammalian sweet taste receptors. Cell 106, 381–390. 26 Zhao GQ, Zhang Y, Hoon MA, Chandrashekar J, Erlenbach I, Ryba NJP & Zuker CS (2003) The recep- tors for mammalian sweet and umami taste. Cell 115, 255–266. 27 Marshall FH, Jones KA, Kaupmann K & Bettler B (1999) GABA B receptors – the first 7TM heterodimers. Trends Pharmacol Sci 20, 396–399. 28 Margeta-Mitrovic M, Jan YN & Jan LY (2000) A traf- ficking checkpoint controls GABA B receptor hetero- dimerization. Neuron 27, 97–106. 29 Liu JF, Maurel D, Etzol S, Brabet I, Ansanay H, Pin JP & Rondard P (2004) Molecular determinants of the allo- steric control of agonist affinity in GABA B receptor by the GABA B2 subunit. J Biol Chem 279, 15824–15830. 30 Sato T, Shimada Y, Nagasawa N, Nakanishi S & Jingami H (2003) Amino acid mutagenesis of the ligand binding interface of the metabotropic glutamate recep- tor crucial residues for setting the activated state. J Biol Chem 278, 4314–4321. 31 Kniazeff J, Bessis A-S, Maurel D, Ansanay H, Prezeau L & Pin J-P (2004) Closed state of both binding domains of homodimeric mGlu receptors is required for full activity. Nat Struct Mol Biol 11, 706–713. 32 Suzuki Y, Moriyoshi E, Tsuchiya D & Jingami H (2004) Negative cooperativity of glutamate binding in the dimeric metabotropic glutamate receptor subtype 1. J Biol Chem 279, 35526–35534. 33 Kniazeff J, Galvez T, Labesse G & Pin J-P (2002) No ligand binding in the GB2 subunit of the GABA B receptor is required for activation and allosteric interaction between the subunits. J Neurosci 22, 7352– 7361. 34 Galvez T, Duthey B, Kniazeff J, Blahos J, Rovelli G, Bettler B, Pre ´ zeau L & Pin J-P (2001) Allosteric interac- tions between GB1 and GB2 subunits are required for optimal GABA B receptor function. EMBO J 20, 2152– 2159. 35 Xu H, Staszewski L, Tang H, Adler E, Zoller M & Li X (2004) Different functional roles of T1R subunits in the heteromeric taste receptors. Proc Natl Acad Sci USA 101, 14258–14263. 36 Pagano A, Rovelli G, Mosbacher J, Lohmann T, Duthey B, Stauffer D, Ristig D, Schuler V, Meigel I, Lampert C et al. (2001) C-Terminal interaction is essential for surface trafficking but not for heteromeric assembly of GABA B receptors. J Neurosci 21, 1189–1202. 37 Robbins MJ, Calver AR, Filippov AK, Hirst WD, Russell RB, Wood MD, Nasir S, Couve A, Brown DA, Moss SJ et al. (2001) GABA B2 is essential for G-protein coupling of the GABA B receptor heterodimer. J Neuro- sci 21, 8043–8052. 38 Duthey B, Caudron S, Perroy J, Bettler B, Fagni L, Pin J-P & Pre ´ zeau L (2002) A single subunit (GB2) is required for G-protein activation by the heterodimeric GABA B receptor. J Biol Chem 277, 3236–3241. 39 Tateyama M, Abe H, Nakata H, Saito O & Kubo Y (2004) Ligand-induced rearrangement of the dimeric metabotropic glutamate receptor 1alpha. Nat Struct Mol Biol 11, 637–642. 40 Gru ¨ newald S, Schupp BJ, Ikeda SR, Kuner R, Steigerw- ald F, Kornau H-C & Ko ¨ hr G (2002) Importance of the c-aminobutyric acid B receptor C-termini for G-pro- tein coupling. Mol Pharmacol 61, 1070–1080. 41 Carroll FY, Stolle A, Beart PM, Voerste A, Brabet I, Mauler F, Joly C, Antonicek H, Bockaert JM, Mu ¨ ller T et al. (2001) BAY36-7620: a potent non-competitive mGlu1 receptor antagonist with inverse agonist activity. Mol Pharmacol 59, 965–973. 42 Pagano A, Regg D, Litschig S, Stoehr N, Stierlin C, Heinrich M, Floersheim P, Pre ´ zeau L, Carroll F, Pin J-P et al. (2000) The non-competitive antagonists 2- methyl-6-(phenylethynyl) pyridine and 7-hydroxyimino- cyclopropan[b]chromen-1a-carboxylic acid ethyl ester interact with overlapping binding pockets in the trans- membrane region of group I metabotropic glutamate receptors. J Biol Chem 275, 33750–33758. 43 Knoflach F, Mutel V, Jolidon S, Kew JN, Malherbe P, Vieira E, Wichmann J & Kemp JA (2001) Positive allo- steric modulators of metabotropic glutamate 1 receptor: characterization, mechanism of action, and binding site. Proc Natl Acad Sci USA 98, 13402–13407. 44 Galvez T, Urwyler S, Pre ´ zeau L, Mosbacher J, Joly C, Malitschek B, Heid J, Brabet I, Froestl W, Bettler B et al. (2000) Ca 2+ -requirement for high affinity c-amino- butyric acid (GABA) binding at GABA B receptors: involvement of serine 269 of the GABA B R1 subunit. Mol Pharmacol 57, 419–426. 45 Jiang P, Ji Q, Liu Z, Snyder LA, Benard LMJ, Mar- golskee RF & Max M (2004) The cysteine-rich region of T1R3 determines responses to intensely sweet pro- teins. J Biol Chem 279, 45068–45075. J-P. Pin et al. Class C G-protein-coupled receptors FEBS Journal 272 (2005) 2947–2955 ª 2005 FEBS 2955 . so-called allosteric modulators. The multiple domains of class C GPCRs In contrast to most class A rhodopsin-like GPCRs, class C receptors are composed of. drug development. Abbreviations Acc, active-closed-closed conformation; Aco, active-closed-open conformation; CaS, receptor, calcium-sensing receptor; CRD, cystein-rich domain;