The heptahelical domain of GABA(B2) is activated directly by CGP7930, a positive allosteric modulator of the GABA(B) receptor

HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. The heptahelical domain of GABA(B2) is activated directly by CGP7930, a positive allosteric modulator of the GABA(B) receptor. Virginie Binet, Carole Brajon, Laurent Le Corre, Francine Acher, Jean-Philippe Pin, Laurent Prézeau

The GABA B receptor is a G protein-coupled receptor (GPCR) activated by the most abundant inhibitory neurotransmitter of the central nervous system, γ-amino butyric acid (GABA). This receptor is involved in numerous physiological processes via the regulation of both GABAergic and glutamatergic synapses at either the pre-or post-synaptic level (1).
Accordingly, GABA B receptors are involved in various types of epilepsy, in nociception and drug addiction, and in spasticity associated with multiple sclerosis (2). Although it has been pharmacologically described for twenty years, only in 1998 was the first GABA B receptor (GABA B1 ) cloned (3). It belonged to the class-III of the GPCR super family, together with the metabotropic glutamate (mGlu), the calcium sensing (CaS), and some pheromone and taste receptors (4). In addition to the typical GPCR heptahelical domain (HD), GABA B1 possesses a large extracellular domain (ECD), like most other class-III GPCRs (4). In contrast to the rhodopsin-like receptors (class-I GPCRs), the ligand binding site of class-III GPCRs is located within their large ECD, in the so-called Venus Fly-Trap Module (VFTM). Indeed, agonists bind within a cleft that separates the two lobes of the VFTM and stabilize a closed active conformation. This has been recently illustrated by the crystal structures of the mGlu1 ECD that have been solved both in the absence and presence of agonist (5), and confirmed in the case of GABA B1 by multiple mutagenesis studies (6)(7)(8).
However, to form a receptor able to efficiently activate G-proteins, GABA B1 need to be associated with a homologous protein called GABA B2 (9)(10)(11)(12)(13). GABA B receptor was then the first described obligatory heterodimeric receptor. Several studies unraveled some specific roles dedicated to each subunit. First, GABA B2 takes GABA B1 to the cell surface probably by masking a retention signal located in GABA B1 C-terminal tail (14)(15)(16). Secondly, GABA B1 VFTM, but not that of GABA B2 , binds all known GABA B agonists and antagonists, whereas GABA B2 HD is critical for G protein activation (17)(18)(19)(20). Third, there are complex allosteric interactions between the ECD and the HD of both subunits (20)(21)(22) leading to optimal agonist affinity and coupling efficacy.
Although the co-expression of both GABA B1 and GABA B2 appears to be required for an efficient activation of G-proteins, some studies report a possible functioning of one GABA B subunit independently of the other (9). (23). In support of a functional role of GABA B2 in the absence of GABA B1 , homodimeric GABA B2 receptors have been observed at the surface of heterologous cells (24), and homodimeric GABA B2 HDs are capable of activating G-proteins (20,21). Moreover, localization studies revealed that some neurons in the brain express much more mRNA of one subunit than of the other consistent with a possible role of homomeric GABA B receptors (25). Finally, GABA B1 may be able to activate intracellular pathways independently of G-proteins (23, [26][27][28]. The identification of selective compounds acting on GABA B2 will probably help unravel this issue. Within the last few years, allosteric modulators of class-III GPCRs have been identified for CaS and mGlu receptors (29-31). Such compounds act either as non-competitive antagonists (32)(33)(34), or as positive allosteric modulators (35)(36)(37)(38)(39)(40). In each case, these compounds have been shown to bind within the HD of their targeted receptor (34,(41)(42)(43).
They are also highly selective for one receptor subtype, in contrast to most of the ligand acting at the orthosteric site. Urwyler and coll. recently described the first GABA B specific positive allosteric modulators, CGP7930 and CGP13501 and more recently the compound GS39783 (44,45). Their site of action was not identified, but according to what was known for the mGlu specific positive allosteric modulators and to the fact that GABA B2 coupled to G protein, we hypothesized that these compounds act in the GABA B2 HD (46).
In the present work, we not only demonstrate that CGP7930 indeed modulates the GABA B receptor by directly acting in the GABA B2 HD, but also that it activates the homomeric GABA B2 receptor, indicating that GABA B2 could be functional by itself.
Moreover, CGP7930 also activates a truncated version of GABA B2 deleted of the ECD, demonstrating that this HD can behave like a rhodopsin-like receptor. These data bring much information on the mechanism of action of this GABA B positive modulator and reveal that GABA B2 selective drugs can be identified. Such drugs will be useful to better dissect the specific role of GABA B1 and GABA B2 in the brain.

Plasmids and site-directed mutagenesis
The plasmids encoding the wild-type and chimeric GABA B1a and GABA B2 subunits epitope tagged at their N-terminal ends (pRK-GABA B1a -HA, pRK-GABA B1a/2 -HA, pRK-GABA B2/1 -HA and pRK-GABA B2 -HA or -cMyc), under the control of a CMV promoter, were previously described (16,20). PRK-HD2 was generated by deletion of the sequence coding for the ECD in the pRK-GABA B2 -HA plasmid, with the use of the MluI restriction site located just after the sequence coding for the HA tag and a MluI site created just after the proline residues at position 463 in GABA B2 .

Cell culture and transfection
Human embryonic kidney (HEK) 293 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% FCS and transfected by electroporation as described elsewhere (47,48). Unless stated otherwise, 10.10 6 cells were transfected with plasmid DNA containing the coding sequence of the receptor subunits, and completed to total amount of 10 µg plasmid DNA with pRK 6 . For determination of inositol phosphate accumulation, the cells were also tranfected with the chimeric Gαqi9 G-protein which allows the coupling of the recombinant heteromeric GABA B receptor to PLC (47).

Measurement of inositol phosphate production
Determination of inositol phosphate (IP) accumulation in transfected cells was performed in 96-wells plates (0.2.10 6 cells/well) after over-night labeling with 3 H-myoinositols (0.5µCi/well) as already described (49). The stimulation was conducted for 30 minutes in a medium containing 10mM LiCl and the indicated concentration of agonist or antagonist. The reaction was stopped with a 0.1M formic acid solution. Supernatants were recovered and IP were purified by ion exchange chromatography using DOWEX AG1-X8 resin (Biorad, Marnes-la-Coquette, France) in 96 well filter plates (ref: MAHVN4550 Millipore, Bedford, MA). Total radioactivity remaining in the membrane fractions was counted after treatment of cells with a solution containing 10% triton X-100 and 0.1N NaOH.
Radioactivity was quantified using Wallac 1450 MicroBeta liquid scintillation counter. Data were expressed as IP/Membrane, meaning the amount of total IPs produced over the amount of radioactivity remaining in the membranes, multiplied by 100. Unless stated otherwise, all data are means ± sem of at least 3 independent experiments. The dose-response curves were fitted using the Kaleidagraph program and the following equation "y=[(y maxy min )/(1+(x/EC 50 ) nH )) +y min " where the EC 50 is the concentration of the compound necessary to obtain 50% of the maximal effect and nH is the Hill coefficient.

Anti-HA ELISA for quantification of cell surface expression
Twenty-four hours after transfection (10.10 6 cells, HA-tagged GABA B1 (2µg) and cMyc-tagged GABA B2 (2µg) subunits), cells were fixed with 4% paraformaldehyde and then blocked with PBS + 5% FBS. After 30 minutes reaction with primary antibody (monoclonal anti-HA clone 3F10 (Roche, Basel, Switzerland) at 0.5µg/mL) in the same buffer, the goat Anti-Rat antibody coupled to horseradish peroxidase (Jackson Immunoresearch, West Grove, PA) was applied for 30 minutes at 1µg/mL. After intense washes with PBS, secondary antibody was detected and quantified instantaneously by chemiluminescence using Supersignal® ELISA femto maximum sensitivity substrate (Pierce, Rockford, IL) and a Wallac Victor 2 luminescence counter.

GTP--[ 35 S] binding measurements
Cells were transfected using PolyFect transfection reagent (Qiagen, Hilden, Germany) under optimized conditions. Complex were formed using total amount of 8µg plasmid DNA with 60µL of polyfect in 300µL of serum free antibiotic free DMEM for 10 minutes and then added to cells at 40-60% confluence. According expression results, the amount of DNA is GABA B1 2µg, GABA B2 1µg, Gαo1c 2µg and pRK 6 3µg for wild-type receptor. with antagonist (15 minutes) and after with agonist (15 minutes). 60µL of incubation buffer (50mM Tris, 1mM EDTA, 10µM GDP, 5mM MgCl 2 , 0.01mg/mL leupeptine, 100mM NaCl), and 20µL of H 2 O per well were added, and then, the plate was incubated one hour at 30°C.
After vacuum filtration and plate filter drying, the radioactivity was measured using a Wallac 1450 MicroBeta liquid scintillation counter. The dose-response curves were fitted using the Kaleidagraph program and the following equation "y=[(y max -y min )/(1+(x/EC 50 ) nH )) +y min " where the EC 50 is the concentration of the compound necessary to obtain 50% of the maximal effect and nH is the Hill coefficient.

CGP7930 is a positive allosteric modulator of the GABA B receptor
As previously described by Urwyler  Therefore, the effect of CGP7930 was further analyzed using an IP production assay which is supposed to give a higher signal to noise ratio. Indeed the GABA B receptor can efficiently activate PLC when co-expressed with the chimeric G-protein Gqi9 (47). As previously noticed, the IP assay is much more sensitive than the GTP-γ[ 35  However, in this assay, CGP7930 also clearly stimulated IP production even in the absence of added GABA, further suggesting that the CGP7930 could be a GABA B partial agonist ( Fig.2 and 3).

CGP7930 is a partial agonist of the GABA B receptor
Additional experiments were performed in order to demonstrate that CGP7930 directly activates the GABA B receptor. First, CGP7930 alone did not induce IP formation in pRK6 and Gqi9 transfected control cells, nor in cells co-expressing mGlu5, demonstrating that IP production did not result from a direct action on either the transfected G-protein or PLC (data not shown). Second, the observed stimulation of IP production by CGP7930 (Fig.3A), is dose-dependent with an EC 50 similar to that observed for the potentiating effect (32.5 ± 7.2 µM; Fig.4).
As we could not exclude that the observed effect of CGP7930 was du to overexpression of the receptors, we decided to examine the effect of CGP7930 at different receptor expression levels. To that aim the effect of saturating concentrations of GABA or CGP7930 (1mM) were measured in cells transfected with increasing amount of GABA B1 -and GABA B2 -expressing plasmids (Fig.5). The expression levels of heteromeric GABA B receptor were determined using ELISA performed on intact cells with an anti-HA antibody labeling the N-terminal HA-tagged GABA B1 subunit. As shown in Fig.5, the maximal agonist activity of CGP7930 was always lower than the maximal GABA activity, and of course than the maximal activity induced by CGP7930 together with GABA, indicating that CGP7930 was only a partial agonist.

The heptahelical domain of GB2 is required for the action of CGP7930
In order to identify the site of action of CGP7930, we first examined whether the stimulatory effect was inhibited by the competitive antagonist CGP54626. As shown in GABA B heterodimers can be considered as the association of 4 distinct domains (ECD1, ECD2, HD1 and HD2) that correspond to the ECD and HD of GABA B1 and GABA B2 subunits, respectively. To identify which of these domains is required for the effect of CGP7930 its action on various combinations of chimeric and mutated subunits was examined.
The chimeric subunits used were GABA B1/2 and GABA B2/1 , in which the entire ECD have been swapped between GABA B1 and GABA B2 (20).
The GABA B1/2 was expressed with GABA B1 to form a receptor that does not contain the ECD2, and vice-versa, GABA B2/1 was expressed with GABA B2 to form a receptor devoid of ECD1. Both combinations have already been shown to be expressed at the cell surface and to form heteromeric complexes (20). Although they are not sensitive to GABA, both activate Gqi9, as illustrated by the high constitutive IP formation measured in cells expressing these subunit combinations (20). As shown in Fig.6, CGP7930 stimulated IP production in cells expressing either combination (Fig.6, compare lanes 3 and 6). Accordingly, none of the ECD was required for the effect of CGP7930. Then, although we could not rule out that CGP7930 could act similarly on both ECD1 and ECD2 and that only one ECD would be enough for the effect of CGP7930, the more likely possibility was that CGP7930 acts in the HD of either GABA B1 or GABA B2 .
In order to determine which HD could be the site of action of CGP7930, the effect of CGP7930 was analyzed on the combinations GABA B1 + GABA B2/1 and GABA B2 + GABA B1/2 . The first combination possessed only HD1 and not HD2, whereas the second possessed HD2 only and not HD1 (Fig.7). In order to allow the correct expression of both GABA B1 and GABA B2/1 at the cell surface, the ER retention signal RSR located in the Cterminal tail of these subunits was mutated into ASA (20). Although the combination GABA B1 + GABA B2/1 was not activated by GABA, whereas the combination GABA B2 + GABA B1/2 was, both combinations were similarly expressed at the cell surface ((20) and data not shown). As shown in Fig. 7, CGP7930 stimulated the combination containing only HD2 (4 fold increase of the IP production, from 4.7 to 13.8 normalized cpm in the absence and presence of 100µM CGP7930, respectively), but was devoid of activity on that possessing HD1 only. Taken together, these data illustrated the requirement of HD2 for the partial agonist activity of CGP7930 in the GABA B heteromer. However, because GABA B1 cannot activate the G protein by itself, one can not exclude that CGP7930 bound to GABA B1 but failed at allowing it to stimulate G proteins.

CGP7930 activated GABA B2 in the absence of GABA B1
As mentioned above, GABA B2 possesses sufficient molecular determinants for Gprotein activation (17)(18)(19)(20). Even when transfected alone, GABA B2 was highly expressed at the HEK293 cell surface (Fig.8A), allowing us to examine whether CGP7930 could activate this subunit expressed alone. Indeed, whereas GABA is devoid of activity on GABA B2 , CGP7930 increased IP production 3 fold (Fig.8B) with an EC 50 of 57.1 ± 3.8 µM (Fig. 9). A similar effect of CGP7930 was observed with the chimeric subunit GABA B1/2 , even though GABA was inactive (Fig.8A and B). This confirms that the HD of GABA B2 is crucial for the CGP7930 effect. However, it is still possible that CGP7930 requires the presence of an ECD, either that of GABA B1 or GABA B2 , to turn on the GABA B2 HD.

CGP7930 activated GABA B2 HD (HD2) expressed alone
According to the data described above, it appeared that CGP7930 could activate GABA B2 by a direct action in its HD. Thus, we looked for the action of CGP7930 on the HD of GABA B2 alone. Indeed, we have recently shown that the HD of mGlu5 receptor could be expressed alone and could be directly activated by a positive allosteric modulator of this receptor (49). We therefore generated a truncated version of GABA B2 lacking the large ECD ( Fig. 8). Thanks to the presence of a signal peptide and of a HA tag inserted at the Nterminus, HD2 was found at the cell surface, although at a lower level than the wild-type subunit (30% of the wild type Fig.8A)). On cells expressing HD2, CGP7930 increased IP production more than two folds, with an EC 50 of 64.8 ± 38.7 µM (Fig.9). These data showed that CGP7930 directly stabilizes an active conformation of the GABA B2 HD, and can be considered as a first GABA B2 ligand.

DISCUSSION
In the present study, we explored the mechanism of action of CGP7930, one of the first described positive allosteric modulators of the GABA B receptor (44). Using both GTPγ[ 35 S] binding and IP production assays, that reflect the activation of Gαo and Gαqi9 respectively, we confirmed the positive allosteric action of CGP7930. However, using the more sensitive IP assay -i.e. activation of PLC via Gαqi9 -CGP7930 was also found to directly activate the GABA B receptor though with a low efficacy. The action of CGP7930 on various combinations of wild type and chimeric subunits, led us to propose that CGP7930 activated HD2 within the heteromer. This proposal was then directly demonstrated, as we showed that CGP7930 acted as an agonist of HD2 expressed alone, demonstrating this domain of the GABA B receptor behaved like a rhodopsin-like receptor. It is noteworthy that even if CGP7930 could bind to the HD of GABA B1 , it would not induced G protein activation, as we never observed any activation of the recombinant G-proteins by GABA B receptors lacking HD2 (17,19,20).

CGP7930, a partial agonist of the GABA B receptor
Not only could CGP7930 potentiate the effect of GABA, as previously reported by others (44), but it could also directly activate the wild-type receptor. This effect occured in a similar range of concentration as those observed for the potentiating effect, and various arguments excluded the possibility of a potentiation of the effect of a possible endogenous agonist present in the assay medium. Indeed, the effect of CGP7930 was not fully inhibited by a competitive antagonist. Moreover, the effect of CGP7930 could still be observed on various mutated GABA B receptors not sensitive to GABA.
Since CGP7930 activated GABA B2 in the absence of GABA B1 , then the agonist effect observed could well be the consequence of some GABA B2 subunits not associated with GABA B1 . Although this possibility can not be firmly excluded, we think it is unlikely since we and others observed that in heterologous systems GABA B2 was less expressed than GABA B1 (8,50). Accordingly, it is very unlikely that there was enough isolated GABA B2 subunits (either in a monomeric or homodimeric form) in cells transfected with both GABA B1 and GABA B2 to generate a CGP7930-induced response higher than that observed in cells expressing GABA B2 only.
Of interest, CGP7930 activated not only the heteromeric GABA B receptor and the GABA B2 subunit expressed alone, but also a GABA B2 subunit deleted of its ECD. Moreover, all these effects were observed in the same range of concentration of CGP7930 (with very similar EC 50 values). This observation clearly indicates that neither GABA B2 ECD nor the GABA B1 are required for the agonist activity of CGP7930. Also, this further demonstrates that agonist binding is not required for CGP7930 interaction with HD2.

On a possible allosteric control of CGP7930 effect
The affinity of GABA on the ECD of GABA B1 ass allosterically regulated by the other domains of the heteromeric GABA B receptor, like the ECD of GABA B2 . We recently hypothesized that this effect was probably due to the relief by the GABA B2 ECD of an inhibitory action of the GABA B1 HD on the GABA B1 ECD (22). Accordingly, one would expect that the affinity of CGP7930 in the HD of GABA B2 would also be under the allosteric control of the other domains of the heteromeric complex. The EC 50 of CGP7930 on the wild type receptor, on GABA B2 or on HD2 were quite similar, whether the agonist effect or the potentiating effect of CGP7930 was measured. This suggests that CGP7930 affinity was not as dependent as GABA affinity on the specific state of the other domains. Alternatively, within this range of concentration tested, CGP7930 may only bind and exert its effect on receptors in a specific state. The identification of radioactive compounds interacting at the same site than CGP7930 would be required to further study this point.
However, the antagonist CGP54626 that likely stabilizes the inactive open state of GABA B1 ECD, partly inhibited the agonist effect of CGP7930 on the wild-type receptor. Such an effect was not observed on the GABA B2 subunit expressed alone or on HD2. Then this demonstrated that the action of CGP7930 was dependent of the specific state of the GABA B1 ECD. To our actual knowledge on the mechanism of activation of the GABA B receptor, GABA binding in the VFT of GABA B1 leads to the closure of the VFT, and to a possible reorientation of the dimer of VFTs. As a consequence, this stabilizes the active conformation of the dimer of HDs. The competitive antagonist like CGP54626 is expected to prevent the closure of the GABA B1 VFT, and thus to prevent the stabilization of the active state of the HD dimer by agonists. According to this model, any drug directly stabilizing the dimer of HDs will also stabilize the active conformation of the dimer of VFTs. In agreement with this proposal, CGP7930 increased agonist affinity. Due to allosteric coupling between the dimer of VFTs and the dimer of HD, locking the dimer of VFTs in the inactive state by a competitive antagonist, are expected to make more difficult the change in conformation of the dimer of HDs required for its activation. This is expected to decrease the effect of CGP7930, as observed here. .

CGP7930 as an activator of the GABA B2 HD
As recently reported for mGlu5 (49), the HD of GABA B2 could be expressed as a membrane protein at the cell surface in HEK293 cells. Although the mGlu5 truncated receptor displayed constitutive activity, no such activity could be detected with GABA B2 . This was in agreement with the higher constitutive activity measured with mGlu5 compared to GABA B receptor (51,52). However, in both cases, these HDs were activated by positive allosteric regulators, CGP7930 and DFB for GABA B2 and mGlu5 HDs, respectively. This shows that, even though class-III and class-I rhodopsin-like GPCRs diverged early during evolution, their HDs still possess common structural properties and likely share similar activation mechanism.
Whether GPCRs function as monomer or dimers is a matter of intense debate in the field (53)(54)(55). However, in the case of class-III GPCRs it is well accepted that the dimeric nature of these receptors is crucial for the intra-molecular transduction -i.e. transfer of information from the agonist binding domain to the heptahelical G-protein activating domain (4). However, whether a dimeric nature of the HD of these receptors is necessary for the agonist effect of positive allosteric modulators is not known. Obviously, the demonstration that HDs of class-III GPCRs can function like class-I GPCRs will help unravel this important issue.

A model for the action of CGP7930 on the GABA B receptor
We recently proposed a model for the functioning of class-III GPCRs (56). This model integrated our common view of the functioning of Venus Fly-Trap modules, meaning the binding of the ligand in the cleft between the two lobes of the ECD, and the stabilization of a closed state by agonists, or prevention of the closure of such a domain upon antagonist binding. Moreover, our model takes into account the putative mechanism of activation of the HD, with the existence of at least two states, active and inactive, the equilibrium between these two states being under the control of the specific conformation of the ECD. As discussed in this paper, this model fits very nicely with a series of specific properties of class-III GPCRs. For example, this model provides a reasonable explanation for the lack of inverse agonist activity of competitive antagonists of mGluRs, whereas non-competitive antagonists interacting in the HD were found to be inverse agonists (56).
In this model, we proposed two options to explain the effect of positive allosteric modulators acting in the HD of class-III GPCRs. A first possibility is that such compounds act by stabilizing the active state of the HD, and as a consequence stabilize the active state of the binding domain therefore increasing agonist affinity. According to this proposal, such positive allosteric modulators were expected to also activate with a low efficacy the full-length receptor. This nicely fits with our observation that CGP7930 acted both as an activator and as a positive allosteric modulator of the full-length GABA B receptor. The second possibility was that the positive modulators acted by increasing the allosteric coupling between the active ECD and the HD, rather than by directly activating the HD. According to the second possibility, the positive modulator directly activated neither the full-length receptor, nor the HD. Our recent observation that the mGlu5 positive modulator, DFB (3,3'-Difluorobenzaldazine), did not activate the full-length receptor, but acted as a full agonist on the receptor deleted of its ECD, did not fit with any of these two possibilities. Accordingly, a more complex model involving 3 states of the mGlu5 HD was proposed (49). This was based on the recognized three states of rhodopsin (57,58) Such an observation provides evidence for a different activation mechanism of GABA B receptors and the other Class-III GPCRs, such as mGluRs. Indeed, although these two types of receptors share sequence similarities, the GABA B receptor subunits lack the cystein-rich domain that interconnects the ECD to the HD in the mGlu-like receptors. Further studies will be necessary to better clarify the specific properties of both types of class-III GPCRs.

Conclusion
The main observation of this study is that it is possible to identify compounds acting on the GABA B2 subunit. As discussed above, such compounds will be useful to elucidate the activation mechanism of such a complex GABA B heteromeric receptor. In addition, as pointed out in our introduction, it is well recognized that the vast majority of GABA B1 and GABA B2 subunits associate with each other to form a functional GABA B receptor in the brain. However, some observations suggest that either GABA B1 or GABA B2 could be active on their own, or in association with another type of subunit. Our data show that it should be possible to identify compounds acting on GABA B2 specifically. Such compounds will help unravel the possible function of this subunit in the brain.