Abstract
The recombinant human α2A-adrenoceptor (α2A-AR, RC 2.1.ADR.A2A) can be transformed into a constitutively activated form in CHO-K1 cells by coexpression with a rat Gαo protein. Constitutive activity could be enhanced more by both mutation of Thr373 of the α2A-AR to a Lys and Cys351 of the Gαo protein by an Ile. The basal [35S]GTPγS binding response displayed a constitutive α2A-AR activity that amounted to 21% of the maximal receptor activation as obtained with 10 μM (−)-adrenaline. UK 14304, BHT 920, d-medetomidine, oxymetazoline, and clonidine acted as efficacious agonists. The enhancement of basal activity was entirely blocked (−50 ± 3%) by ligands that thus appeared to act as inverse agonists (i.e., RX 811059 and its (+)-enantiomer, (+)-RX 821002, RS 15385, and yohimbine); the potencies of the ligands corresponded with their binding affinities for the α2A-AR. Fluparoxan and WB 4101 displayed partial inverse agonism. Atipamezole and dexefaroxan at 10 μM were virtually free of intrinsic activity and thus acted as neutral antagonists; idazoxan displayed potent partial agonist properties as observed with BRL 44408 and SKF 86466. The inverse agonist activity induced by (+)-RX 811059 could be reversed by atipamezole with a pKB value (8.73 ± 0.07) that was similar to that required for blockade of the UK 14304-mediated response. Constitutive α2A-AR activation was mainly observed with the Gαo Cys351Ile protein compared with the pertussis toxin-resistant mutants of the Gαi protein subtypes. The observed spectrum of intrinsic activities for the various ligands suggests that pure, neutral antagonists are rather uncommon in this specified α2A-AR system.
α2-Adrenoceptors (α2-AR) are implicated in the control of noradrenergic neurotransmission in the central nervous system and modulate several physiological processes peripherally (Timmermans et al., 1990; Szabadi and Bradshaw, 1996). There are now three characterized α2-AR subtypes: α2A, α2B, and α2C; these are G protein-coupled receptors that are predominantly coupled to the Gi/o signaling system, inhibiting and/or stimulating the activity of adenylate cyclase, inhibiting the opening of voltage-gated Ca2+ channels, and activating K+ channels (see Hein and Kobilka, 1997). The α2-AR may also couple to other intracellular pathways involving Na+/H+exchange and the activation of phospholipase A2and C (Limbird, 1988; Cotecchia et al., 1990; Kukkonen et al., 1998). The α2-AR subtypes are distributed differentially in cells and tissues, endowing these receptors with different physiological functions and pharmacological activity profiles. However, available ligands have only marginal subtype selectivity.
A widely accepted model used to describe agonist activation of G protein-coupled receptors is the ternary complex model, which accounts for the cooperative interactions among receptor, G protein, and agonist (De Lean et al., 1980). This model has recently been extended to accommodate the observation that many receptors can activate G proteins in the absence of agonist, and that mutations in different structural domains of the receptors can enhance the agonist-independent (constitutive) activity (Samama et al., 1993). The extended ternary complex model also accounts for the effects of different types of receptor ligands [full agonists, partial agonists, silent ligands (neutral antagonists), and inverse agonists (also defined as negative antagonists)] on receptor signaling (Gether and Kobilka, 1998). However, the pharmacological distinction between a neutral antagonist and an inverse agonist is often difficult to observe. One possible explanation is that the magnitude of inverse agonism is determined by constitutive receptor activation of specific G protein subtypes. Consequently, ligands may demonstrate distinct pharmacological effects (i.e., neutral antagonism or inverse agonism) depending on which receptor/G protein/effector pathway is involved. Perez et al. (1996)reported on a Cys128Phe mutation in the α1B-AR, resulting in G protein coupling in the absence of agonist and constitutive activation of the phospholipase C, but not of the phospholipase A2 pathway. A similar mutation (Cys116Phe) in the β2-AR causes selective constitutive activation of Na+/H+ exchange through a pathway not involving cAMP (Zuscik et al., 1998). These data suggest that multiple and distinct activation states exist for a receptor, and that the pharmacological profile of a single receptor subtype may be codetermined by the effector pathway that is being considered.
The carboxy-terminal portion of the third intracellular loop (ICL) has been suggested (Kjelsberg et al., 1992) to be involved in constraining G protein-coupled receptors in the inactive (G protein-uncoupled) conformation. Mutagenesis studies of the BBXXB motif (where B represents a basic residue andX a nonbasic residue) in the third ICL of α2A-AR demonstrated a constitutively active mutant (i.e., Thr373Lys; Ren et al., 1993) by analogy with mutations affecting the same region in α1B-AR and β2-AR, and more recently in 5-HT2A and 5-HT2C receptors (Egan et al., 1998;Herrick-Davis et al., 1997). The constitutively active mutant Thr373Lys α2A-AR demonstrates apparently disparate results (i.e., positive efficacy) for ligands that so far have been characterized as α2A-AR antagonists, by measuring the positive coupling of this mutant α2A-AR to the formation of inositol phosphates in the presence of a Gα15 protein (Wurch et al., 1999). This led us to suggest that the pharmacology of constitutively active receptors may be more complex than is commonly assumed. These data also suggest that pure neutral antagonists at α2A-AR may be relatively uncommon. It has yet to be determined whether constitutive activity of α2A-AR can be inhibited by inverse agonists, and whether this type of ligand activity can be reversed by neutral antagonists. Tian et al. (1994) demonstrated that some α2-AR antagonists display inverse agonist activity at precoupled wild-type (wt) rat α2D-AR in recombinant PC12 cells.
In this study, the intrinsic activity of α2-AR ligands was analyzed at the wt and mutant Thr373Lys α2A-AR on activation by a rat Gαo protein. ABordetella pertussis toxin (PTX)-resistant mutant Gαo Cys351Ile protein (Dupuis et al., 1999) was used to avoid potential coupling of the α2A-AR to endogenous Gαi/o proteins of CHO-K1 cells. The activation of α2A-AR by either endogenous G proteins or a recombinant Gαo protein was estimated by measuring agonist-independent and -dependent PTX-resistant binding of the stable GTP analog [35S]GTPγS. The process of constitutive activation of the mutant α2A-AR is considerably favored by coexpression of a Gαo Cys351Ile protein and can only be fully blocked by a minority of the putative α2-AR antagonists that were investigated.
Experimental Procedures
Cloning of wt and Mutant Human α2A-AR and Rat Gα Protein Genes.
Wild-type genes were cloned by polymerase chain reaction (PCR) using primers designed at the start and stop codons of the respective nucleotide sequences. Site-directed mutagenesis of the human α2A-AR gene (Fraser et al., 1989; Genbank: M 23533) was performed using a modified overlap extension technique based on PCR (Wurch et al., 1998) using appropriate complementary primers carrying the Thr373 (ACG codon) to Lys (AAA codon) mutation. The mutant rat GαoCys351Ile protein gene (Jones and Reed, 1987; Genebank: M 17526) was amplified in the same way with primers carrying the Cys (TGT codon) to Ile (ATT codon) mutation. The Cys351Ile mutants of either the Gαi1 (M 17527) and Gαi3 (M 20713) proteins, and the Cys352Ile mutant of the Gαi2 protein (M 17528) were constructed in a similar way. The respective PCR products were separated by agarose gel electrophoresis, purified using a Geneclean II kit, and cloned into 50 ng of a pCR3.1 vector. The recombinant genes are expressed under the transcriptional control of the human cytomegalovirus immediate-early gene promoter and enhancer sequence, which permit efficient and high expression (Boshart et al., 1985). Sequencing, performed automatically on an ABI Prism 310 Genetic Analyser using a dichlororhodamine terminator cycle sequencing kit, confirmed the respective nucleotide sequences.
Transient Expression of Human α2A-AR and Rat Gα Proteins in CHO-K1 Cells.
The CHO-K1 cell line (American Type Culture Collection, CCL 61) was cultured in Petri dishes (50 cm2) with nutrient mixture Ham's F12 supplemented with 10% heat-inactivated fetal calf serum. Cells grown to 60 to 80% confluence were used for transfection using a LipofectAMINE plus kit. Three micrograms of pCR3.1 plasmid containing either the wt or mutant Thr373Lys α2A-AR gene supplemented with 3 μg of empty pCR3.1 plasmid, or 3 μg of wt or mutant Thr373Lys α2A-AR gene and 3 μg (unless otherwise indicated) of either wt or mutant Gα protein gene were mixed with 10 μl of LipofectAMINE plus reagent in 0.2 ml of Opti-Mem and incubated at room temperature for 15 min. Subsequently, 20 μl of LipofectAMINE reagent diluted in 0.2 ml of Opti-Mem was added for 15 min and exposed with 5 ml of Opti-Mem to CHO-K1 cells for 3 h at 37°C. Thereafter, cells were incubated with 10 ml of complete growth medium and harvested 48 h after transfection. Treatment with PTX (20 ng/ml) was performed overnight before membranes were prepared.
Membrane Preparation and Radioligand Binding Experiments.
Membrane preparation steps were performed at 4°C. Cells were washed twice with PBS and stored at −80°C. Cells were then scraped mechanically in 10 mM Tris-HCl supplemented with 0.1 mM EDTA (pH 7.5) and centrifuged for 10 min at 45,000g. The pellet was homogenized in the same buffer using a Polytron and recentrifuged. The final pellet was distributed in aliquots of 0.5 ml of Tris-EDTA buffer (0.5 to 1.5 mg/ml of protein), and stored at −80°C until used. Membrane preparations were diluted in 50 mM Tris-HCl (pH 7.7) containing 4mM CaCl2, 10 μM pargyline, and 0.1% ascorbic acid, and used for [3H]2-(2-methoxy-2,3-dihydro-benzo[1,4]dioxin-2-yl)-4,5-dihydro-1H-imidazole ([3H]RX 821002, 2 nM) binding experiments as described (Wurch et al., 1999). Ten micromolar phentolamine was used to determine nonspecific radioligand binding.
[35S]GTPγS Binding Responses.
Agonist-independent (basal) and -dependent [35S]GTPγS binding responses (Pauwels et al., 1997) were performed to the above described membrane preparation in 20 mM HEPES (pH 7.4) supplemented with 30 μM GDP, 100 mM NaCl, 3 mM MgCl2, and 0.2 mM ascorbic acid. [35S]GTPγS binding responses were systematically performed in the presence of 10 μM 2-(2-ethoxy-2,3-dihydro-benzo[1,4]dioxin-2-yl)-4,5-dihydro-1H-imidazole (RX 811059) to correct for enhanced basal [35S]GTPγS binding. Maximal stimulation of [35S]GTPγS binding was defined in the presence of 10 μM (−)-adrenaline and calculated versus basal [35S]GTPγS binding, unless otherwise indicated. pEC50 values were defined as the concentration of the ligand at which 50% of its own maximal stimulation of [35S]GTPγS binding was obtained. pIC50 values represent the concentration of the ligand that showed 50% inhibition of its own maximal inhibition of basal [35S]GTPγS binding. In antagonist experiments, atipamezole was coincubated with the indicated ligand. pKB values were calculated as KB = (B)/(A′/A) − 1, where B is the concentration of the antagonist, and A and A′ are the EC50 or IC50 values of ligand measured in either the absence or presence of antagonist, respectively. Saturation [35S]GTPγS binding responses were performed as described previously (Pauwels et al., 1998). Statistical analysis was performed onEmax values using one- way (either without or with repeated measures) ANOVA, followed by all pairwise multiple comparison procedures (Tukey's test).
Immunological Gα Protein Detection.
Membrane preparations of CHO-K1 cells transfected with the mutant Thr373Lys α2A-AR in either the absence or presence of Cys351Ile Gαo, Cys351Ile Gαi1, Cys352Ile Gαi2, and Cys351Ile Gαi3 protein were prepared as described above. Total proteins were separated by denaturing SDS-polyacrylamide gel electrophoresis (PAGE) (12.5%, w/v) as described (Laemmli, 1970). Thereafter, the proteins were blotted onto a nylon membrane by semidry electrotransfer (23 V, 45 min) in 192 mM glycine, methanol 20% (v/v), and 25 mM Tris-HCl buffer (pH 8.3). Proteins were probed using either a selective, monoclonal anti-Gαo antibody raised against a synthetic peptide corresponding to the amino acids 13 to 88 of the rat Gαo protein, or a nonselective polyclonal anti-Gα subunit antibody raised against a synthetic peptide corresponding to the amino acids 40 to 54 of the rat Gαz protein. The incubation was performed in PBS buffer containing 0.1% Tween 20 (w/v), 5% dried nonfat milk, and the indicated antibody at a dilution of 1:1000. Proteins were visualized with an anti-rabbit or an anti-mouse IgG antibody coupled to alkaline phosphatase using a colorimetric reaction (0.12 mM 4-nitroblue tetrazolium chloride monohydrate, 0.12 mM 5-bromo 4-chloro 3-indolylphosphate p-toluidine salt, 5 mM MgCl2 in 100 mM diethanolamine, pH 9.6). The computer-based image analysis system Imagena 2000 was used for quantification of the Gα protein signals.
Protein Content.
Membrane protein levels were estimated with a dye-binding assay using a Bio-Rad kit; BSA was used as a standard (Bradford, 1976).
Materials.
The ABI Prism 310 Genetic Analyser and the dichlororhodamine terminator cycle sequencing kit were obtained from Perkin-Elmer (Foster City, CA). The pCR3.1 expression vector was purchased from Invitrogen (San Diego, CA). The Geneclean II kit was obtained from Bio 101 Inc. (La Jolla, CA). CHO-K1 cells were obtained from American Type Culture Collection (Rockville, MD). [3H]RX 821002 (50 Ci/mmol) and [35S]GTPγS (1035–1163 Ci/mmol) were obtained from Amersham (Les Ulis, France). The LipofectAMINE plus kit, cell culture media, fetal calf serum, culture plates, and PTX (50 μg/ml) were obtained from Gibco Biocult Laboratories (Paisley, UK). The Emulsifier-Safe was obtained from Packard (Warrenville, PA). The anti-Gαo, anti-Gα, and anti-rabbit IgG antibodies were purchased from Calbiochem Corp. (La Jolla, CA). The anti-mouse IgG antibody was obtained from NEN (Boston, MA). The Imagena 2000 quantification system was purchased from Biocom (Les Ulis, France). Clonidine, (−)-adrenaline, yohimbine, oxymetazoline, and substrates for the colorimetric immunodetection were obtained from Sigma (St. Louis, MO). (±)-2[(4,5-Dihydro-1H-imidazol-2-yl)methyl]-2,3-dihydro-1-methyl-1H-isoindole (BRL 44408) and 2-(2,6-dimethoxyphenoxyethyl)aminoethyl-1,4-benzodioxane hydrochloride (WB 4101) were obtained from Research Biochemicals, Inc. (Natick, MA).d-Medetomidine was purchased from SmithKline Beecham (Ploufragan, France). 5-Bromo-6-(2-imidazolin-2-ylamino)quinoxaline tartrate (UK 14304), dexefaroxan, atipamezole, (+)-(8aR,12aS,13aS)-3-methoxy-12-(methylsulphonyl)-5,8,8a,9,10,11,12,12a,13,13a-decahydro-6H-isoquino[2′,1-g] [1,6]naphthyridine (RS 15385), and racemic, (+)-, and (−)-RX 811059 were prepared intramuros. Idazoxan and RX 821002 were purchased from Reckitt and Colman (Kingston-upon-Hill, UK). 6-Allyl-5,6,7,8-tetrahydro-4H-thiazolo[4,5-d]azepin-2-ylamine (BHT 920) was a gift from Boehringer Ingelheim (Biberach an der Riss, Germany). 6-Chloro-2,3,4,5-tetrahydro-3-methyl-1H-3-benzazepine (SKF 86466) was obtained from SmithKline Beecham (Herts, UK). Fluparoxan was obtained from Glaxo (Hertfordshire, UK). Stock solutions of ligands were prepared at 10−3 M. Serial dilutions were made in the respective incubation buffers.
Results
On transient expression in CHO-K1 cells, wt α2A-AR displayed a weak (−)-adrenaline-dependent [35S]GTPγS binding response, which could be strongly enhanced by coexpression with the Gαo protein (Fig.1). In the absence of recombinant Gα proteins, the basal [35S]GTPγS binding response was enhanced by 47 ± 15% with (−)-adrenaline (10 μM), whereas the basal response was not affected by the α2-AR antagonist RX 811059 (10 μM). PTX treatment (20 ng/ml) fully abolished the (−)-adrenaline-mediated [35S]- GTPγS binding response (Fig. 1B). Coexpression with a wt Gαo protein enhanced the (−)-adrenaline response by 367 ± 31%, whereas RX 811059 produced some inhibition (−15 ± 2%, P < .05) of the basal [35S]GTPγS response (Fig. 1C). Both effects were absent on treatment by PTX (20 ng/ml) (Fig. 1D). In contrast, coexpression with a PTX-resistant mutant Gαo Cys351Ile protein in the presence of PTX (20 ng/ml) yielded an (−)-adrenaline response (504 ± 78%) and RX 811059-mediated inhibition (−19 ± 3%,P < .05) of the basal [35S]GTPγS binding response (Fig. 1E). Nonetheless, the (−)-adrenaline-mediated [35S]GTPγS binding response was easily measurable in the transfected CHO-K1 cells, the basal [35S]GTPγS response was only poorly enhanced by coexpression with the Gαo protein. Therefore, a similar set of experiments was performed with a mutant Thr373Lys α2A-AR that previously has been shown to be constitutively active (Ren et al., 1993). Coexpression of the mutant Thr373Lys α2-AR with either a wt Gαo or mutant GαoCys351Ile protein clearly enhanced the basal [35S]GTPγS response. Both basal responses were attenuated by RX 811059 (10 μM). This effect was largest (−54 ± 3%) with the GαoCys351Ile protein and resistant to PTX treatment (20 ng/ml; Fig. 1J). The enhanced constitutive activity of mutant Thr373Lys α2A-AR was highly dependent on the amount of GαoCys351Ile plasmid expression in CHO-K1 cells as illustrated in Fig. 2. In contrast to transfection with empty plasmid, dose-dependent Gαo protein expression was observed in CHO-K1 cells on transfection with 0.3 to 3 μg of GαoCys351Ile plasmid (Fig. 2B). Under these conditions, constitutive Thr373Lys α2A-AR activity was enhanced by 46 to 114%. Figure 2C illustrates RX 811059 (10 μM)-mediated inhibition of the enhanced basal [35S]GTPγS binding response from a receptor amount of 100 fmol/mg protein onward.
A comparison between the α2-AR agonist's maximal [35S]GTPγS binding responses as mediated by wt and mutant Thr373Lys α2A-AR in either the absence or presence of Gαo protein is summarized in Fig.3. Besides UK 14304, none of these ligands attained at the wt α2A-AR protein a maximal [35S]GTPγS binding response similar to that of (−)-adrenaline. The maximal response of UK 14304 was not significantly different from that of (−)-adrenaline, whereas the agonists d-medetomidine, BHT 920, oxymetazoline, and clonidine displayed Emax values between 28 and 39%. SKF 86466 and BRL 44408 were virtually inactive at micromolar concentrations as agonists. The coexpression with a wt Gαo protein did not much affect this agonist's pattern of [35S]GTPγS binding responses (Fig.3B). However, after coexpression with a mutant Gαo Cys351Ile protein, the otherwise partial agonists BHT 920, d-medetomidine, and clonidine acted as agonists with an efficacy similar to that of (−)-adrenaline (Fig. 3C). Some intrinsic activity [13 and 8% versus (−)-adrenaline] was also obtained with SKF 86466 and BRL 44408. Both ligands displayed even more intrinsic activity at the mutant Thr373Lys α2A-AR, in particular, on coexpression with a GαoCys351Ile protein [56 and 51% versus (−)-adrenaline]. Otherwise, BHT 920, d-medetomidine, oxymetazoline, and clonidine did apparently display an intrinsic activity slightly (P < .05) exceeding that of (−)-adrenaline (Fig. 3F). Each of these [35S]GTPγS binding responses was dose-dependent; Emax and pEC50 values for the Thr373Lys α2A-AR with coexpression of a GαoCys351Ile protein are summarized in Table1.
Analysis of a series putative α2-AR antagonists at the constitutively active Thr373Lys α2A-AR on coexpression with a rat GαoCys351Ile protein by concentration [35S]GTPγS binding response curves is shown in Fig. 4. The correspondingEmax, pIC50, or pEC50 values are summarized in Table2 and compared with their respective pKi values. RX 811059 potently inhibited basal [35S]GTPγS binding by −54%; the activity resided in the (+)-enantiomer, which was 2 times more potent, whereas the (−)-enantiomer was almost inactive (Fig. 4A). (+)-RX 821002, yohimbine, and RS 15385 yielded maximal inhibition of the basal [35S]GTPγS binding with a magnitude almost similar to that of (+)-RX 811059. Fluparoxan and WB 4101 displayed partial inhibition of basal [35S]GTPγS binding, whereas dexefaroxan and atipamezole were almost inactive at 10 micromolar on basal [35S]GTPγS binding. In contrast, idazoxan yielded stimulation of [35S]GTPγS binding, like SKF 86466, with a potency in agreement with its pKivalue for the mutant Thr373Lys α2A-AR (Fig. 4C). The inhibition of basal [35S]GTPγS binding as mediated by (+)-RX 811059 could be blocked by atipamezole (1 μM) in a competitive manner. The antagonist potency (pKB: 8.73 ± 0.07) of atipamezole fitted with that observed for the antagonism of the UK 14304-mediated [35S]GTPγS binding response (pKB: 8.55 ± 0.04; Fig.5).
Another set of experiments with (+)-RX 811059 was performed by coexpression of the mutant Thr373Lys α2A-AR with the PTX-resistant Cys351Ile mutants of the Gαi1 and Gαi3 proteins, and Cys352Ile mutant of the Gαi2 protein instead of a Gαo Cys351Ile protein. Membrane preparations for [35S]GTPγS binding responses as mediated by these various Gαproteins were selected on the basis of a similar amount of Gα protein expression as shown in Fig.6. Analysis of (−)-adrenaline-specific saturation [35S]GTPγS binding indicated a single class of high-affinity [35S]GTPγS binding sites for each of the investigated Gαproteins, with a slightly higher affinity for the Gαo Cys351Ile protein. The maximal adrenaline-mediated [35S]GTPγS binding capacity for each of these Gαproteins was in the same range; it varied between 0.67 and 1.73 pmol/mg protein (Fig. 7). Whereas basal [35S]GTPγS binding in the presence of each of the Gαi proteins was virtually not affected by (+)-RX 811059, it attenuated basal [35S]GTPγS binding in case of a GαoCys351Ile protein. Consequently, the maximal [35S]GTPγS binding capacity of (−)-adrenaline to the GαoCys351Ile protein was enhanced (Fig. 7D). Table3 summarizes similar (+)-RX 811059 data as mediated by these Gα proteins for four independent experiments.
Discussion
The present study demonstrates that the α2A-AR can be transformed into a constitutively activated form by coexpression with a rat Gαoprotein. This process could be enhanced by a single amino acid mutation (Thr373Lys) in the distal part of the third ICL of the α2A-AR. By measuring the agonist-independent activation of a PTX-resistant Gαo Cys351Ile protein by this mutant α2A-AR, the constitutive activity represented about 21% of the maximal receptor activation as obtained with (−)-adrenaline. This mutant receptor has previously been shown to be constitutively active (Ren et al., 1993; Wurch et al., 1999). However, no evidence was provided for blockade of the constitutive activity by inverse agonists, and whether this type of ligand activity can be reversed by neutral antagonists. The present study illustrates that the enhanced basal activity is yet present at a low receptor amount (≥100 fmol/mg protein) and can be fully blocked by several ligands previously characterized as presumably antagonists. The ligands RX 811059, RX 821002, RS 15385, and yohimbine acted as efficacious inverse agonists with potencies that corresponded with their binding affinities for the α2A-AR. This type of ligand activity could be competitively reversed by a neutral antagonist, such as atipamezole, with properties similar to its reversal of agonist-mediated (positive) responses. Moreover, the described model system demonstrates a spectrum of ligand-mediated intrinsic activities that allow a clear distinction between various degrees of either inverse agonism or positive agonism, and silent neutral antagonists.
Several features were apparent by coexpression of the α2A-AR with a Gαoprotein. The magnitude of the (−)-adrenaline [35S]GTPγS binding response was enhanced about 10-fold by coexpression with a wt Gαoprotein, and this magnitude was not differently affected by a mutant Gαo Cys351Ile protein and/or Thr373Lys α2A-AR. Besides UK 14304, maximal responses of the agonists investigated at wt α2A-AR were different from that of (−)-adrenaline. Similar data have been reported at wt α2A-AR in recombinant HEK 293 and CHO cells using [35S]GTPγS and Ca2+ responses (Jasper et al., 1998; Kukkonen et al., 1998). Less differentiation between the agonists' maximal responses was found by measuring α2-AR agonist modulation of [35S]GTPγS binding to G proteins in human platelet membranes (Gessi et al., 1999). In the present study, the partial agonists BHT 920, d-medetomidine, and clonidine were converted at the wt α2A-AR into highly efficacious agonists by the presence of a mutant Gαo Cys351Ile protein. This observation underlines the role of the amino acid at position 351 in the Gαo protein. This amino acid position has been shown to be involved in the magnitude of agonist-mediated responses in addition to its key role in PTX resistance (Dupuis et al., 1999). The hydrophobic amino acid isoleucine instead of a cysteine at the position 351 of the Gαo protein increases the magnitude of responses mediated by partial agonists. Similar data have also been reported for the Gαi1 protein and porcine α2A-AR in Cos-7 cells (Bahia et al., 1998; Jackson et al., 1999). An increase in the magnitude of the intrinsic activity of partial agonists was also apparent at the mutant Thr373Lys α2A-AR; this effect was enhanced by coexpression with a Gαoprotein and more importantly with a mutant GαoCys351Ile protein. Wurch et al. (1999) also observed an increased efficacy for partial agonists by following the mutant Thr373Lys α2A-AR by measuring the stimulation of inositol phosphates in the presence of a Gα15 protein. The wt as well as the mutant Thr373Lys α2A-AR exhibited (in our experimental conditions in the absence of recombinant G proteins) no measurable constitutive activity although they were responsive to (−)-adrenaline stimulation. This contrasts with data obtained in transfected HEK 293 cells (Ren et al., 1993); their basal activity represented 8 and 82% of UK 14304-dependent inhibition of stimulated cAMP formation for the wt and Thr373Lys α2A-AR, respectively. The basal activity of the wt α2-AR in CHO-K1 cells was slightly increased to 3% of maximal agonist-dependent receptor activation by coexpression with either a wt Gαo or mutant Gαo Cys351Ile protein. Furthermore, the increase of the basal activity at the mutant Thr373Lys α2A-AR was highly dependent on the coexpression with Gαoproteins; 6 and 21% of maximal receptor activation was observed with a wt Gαo and mutant GαoCys351Ile protein, respectively. Constitutive α2A-AR activation in our study was observed mainly with a GαoCys351Ile protein. It is not clear why the mutant Cys352Ile and Cys351Ile forms of, respectively, the Gαi2 and Gαi3 protein did not yield constitutive α2A-AR activity. The possibility that coupling of the α2A-AR to different Gα proteins in different cells would be causing distinct pharmacological properties (i.e., silent antagonism versus inverse agonism) for a given ligand cannot be excluded.
In contrast to assessment of α2A-AR antagonist efficacy with a GTPase assay at wt α2A-AR, where none of the investigated α2 AR antagonists acted as an inverse agonist (Virolainen et al., 1997), we here did observe inverse agonist activity at α2A-AR. The magnitude of this activity was more pronounced at the mutant Thr373Lys α2A-AR. RX 811059, RX 821002, RS 15385, and yohimbine displayed maximal inhibition of constitutive activity at potencies relevant to their binding affinities. Fluparoxan and WB 4101 acted as partial inverse agonists. Tian et al. (1994) also reported on α2-AR antagonists that reduced basal G protein activation by the recombinant α2D-AR in PC12 cells with the following rank order of maximal effectiveness: yohimbine = phentolamine > idazoxan = rauwolscine > WB 4101. Rauwolscine and atipamezole have also been shown to display inverse agonist activity at endogenous α2A-AR in human erythroleukemia cells (HEL 92.1.7) by following both Ca2+ elevation and cAMP production (Jansson et al., 1998). This effect was less marked for idazoxan (Jansson et al., 1998). In the present study, idazoxan and atipamezole did not display inverse agonist activity. Atipamezole as well as dexefaroxan acted as essentially neutral antagonists, whereas idazoxan, BRL 44408, and SKF 86466 displayed partial agonist properties. The stereoselective interaction of the (+)-enantiomer of RX 811059 compared with its inactive (−)-enantiomer emphasizes the specificity of the blockade of constitutive activity by α2-AR ligands in our study. The reversal of this effect by atipamezole in a competitive manner, and at a potency similar to that observed for its blockade of the agonist UK 14304, confirms that the interaction of the observed inverse agonists with the mutant Thr373Lys α2A-AR is indeed transduced by α2-AR. The data on idazoxan, BRL 44408, and SKF 86466 are consistent with the finding (Wurch et al., 1999) that these ligands also acted as partial agonists at the mutant Thr373Lys α2A-AR in the presence of a Gα15 protein. However, dexefaroxan and atipamezole, which appeared also partial agonists at the mutant α2A-AR in the presence of a Gα15 protein (Wurch et al., 1999), seem to be neutral antagonists by coexpression with a GαoCys351Ile protein. It is possible, therefore, that certain pharmacological differences in intrinsic activity for some of these ligands may be due to α2A-AR interactions with selective Gα protein subunits. The ligands RS 15385, RX 811059, and WB 4101 characterized as inverse agonists at the Thr373Lys α2A-AR showed less binding affinity compared with the wt α2A-AR (Wurch et al., 1999) in line with what would be expected for inverse agonists (Westphal and Sanders-Bush, 1994). Although these binding shifts were significant, they were weak. The resolution of the herein described [35S]GTPγS binding responses is higher and therefore more suitable to monitor intrinsic activity of α2-adrenergic ligands.
Although the physiological implications of a constitutively active mutant α2A-AR remain unclear, Neubig et al. (1988) concluded that approximately 30% of platelet α2-AR exist in a precoupled state. The levels of G protein expression are regulated in vivo (Burstein et al., 1997) and can possibly operate to affect α2A-AR functioning. Up-regulation of G protein levels may provide increased sensitivity to signaling and/or enhance alternative signaling pathways (see Burstein et al., 1997). In the present study, enhanced sensitivity to [35S]GTPγS binding responses was observed by coexpression of both the wt α2A-AR and the activating mutant Thr373Lys α2A-AR with Gαoproteins. The observed spectrum of ligands' intrinsic activities in this specified α2A-AR system suggests that several common antagonists behave as either partial agonists or (partial) inverse agonists. Ligands may probably demonstrate distinct pharmacological effects, depending on which G proteins and effector pathways are involved. Therefore, it cannot be excluded that a ligand that behaves as an inverse agonist at a mutant receptor with a specific G protein may as well behave as a neutral antagonist at native receptors. Future studies are needed to establish whether the constitutively active mutant Thr373Lys α2A-AR mimics a transient state of the native mechanism of α2A receptor activation.
Acknowledgments
We thank Dr. Isabelle Rauly and Delphine Dupuis for their assistance with the transfection protocol, and Stéphanie Cecco for secretarial assistance.
Footnotes
-
Send reprint requests to: Dr. Peter J. Pauwels, Department of Cellular and Molecular Biology, Centre de Recherche Pierre Fabre, 17 avenue Jean Moulin, 81106 CASTRES Cédex, France. E-mail:peter.pauwels{at}pierre-fabre.com
- Abbreviations:
- AR
- adrenoceptor
- BHT 920
- 6-allyl-5,6,7,8-tetrahydro-4H-thiazolo[4,5-d]azepin-2-ylamine
- BRL 44408
- (±)-2[(4,5-dihydro-1H-imidazol-2-yl)methyl]-2,3-dihydro-1-methyl-1H-isoindole
- PTX
- Bordetella pertussis toxin
- SKF 86466
- 6-chloro-2,3,4,5-tetrahydro-3-methyl-1H-3-benzazepine
- RS 15385
- (+)-(8aR,12aS,13aS)-3-methoxy-12-(methylsulphonyl)-5,8,8a,9,10,11,12,12a, 13,13a-decahydro-6H-isoquino[2′,1-g][1,6]naphthyridine
- RX 811059
- 2-(2-ethoxy-2,3-dihydro-benzo[1,4]dioxin-2-yl)-4,5-dihydro-1H-imidazole
- RX 821002
- 2-(2-methoxy-2,3-dihydro-benzo[1,4]dioxin-2-yl)-4,5-dihydro-1H-imidazole
- UK 14304
- 5-bromo-6-(2-imidazolin-2-ylamino)quinoxaline tartrate
- WB 4101
- 2-(2,6-dimethoxyphenoxyethyl)aminoethyl-1,4-benzodioxane hydrochloride
- ICL
- intracellular loop
- PCR
- polymerase chain reaction
- wt
- wild-type
- PAGE
- polyacrylamide gel electrophoresis
- Received August 17, 1999.
- Accepted October 25, 1999.
- The American Society for Pharmacology and Experimental Therapeutics