Portal de Periódicos da CAPES

Tyrosine 62 of the γ-Aminobutyric Acid Type A Receptor β2 Subunit Is an Important Determinant of High Affinity Agonist Binding

2000; Elsevier BV; Volume: 275; Issue: 19 Linguagem: Inglês

10.1074/jbc.275.19.14198

1083-351X

J. Glen Newell, Martin Davies, Alan N. Bateson, Susan M. J. Dunn,

Nicotinic Acetylcholine Receptors Study

The γ-aminobutyric acid type A receptor (GABAAR) carries both high (KD = 10–30 nm) and low (KD = 0.1–1.0 μm) affinity binding sites for agonists. We have used site-directed mutagenesis to identify a specific residue in the rat β2 subunit that is involved in high affinity agonist binding. Tyrosine residues at positions 62 and 74 were mutated to either phenylalanine or serine and the effects on ligand binding and ion channel activation were investigated after the expression of mutant subunits with wild-type α1 and γ2 subunits in tsA201 cells or in Xenopus oocytes. None of the mutations affected [3H]Ro15–4513 binding or impaired allosteric interactions between the low affinity GABA and benzodiazepine sites. Although mutations at position 74 had little effect on [3H]muscimol binding, the Y62F mutation decreased the affinity of the high affinity [3H]muscimol binding sites by ∼6-fold, and the Y62S mutation led to a loss of detectable high affinity binding sites. After expression in oocytes, the EC50 values for both muscimol and GABA-induced activation of Y62F and Y62S receptors were increased by 2- and 6-fold compared with the wild-type. We conclude that Tyr-62 of the β subunit is an important determinant for high affinity agonist binding to the GABAA receptor. The γ-aminobutyric acid type A receptor (GABAAR) carries both high (KD = 10–30 nm) and low (KD = 0.1–1.0 μm) affinity binding sites for agonists. We have used site-directed mutagenesis to identify a specific residue in the rat β2 subunit that is involved in high affinity agonist binding. Tyrosine residues at positions 62 and 74 were mutated to either phenylalanine or serine and the effects on ligand binding and ion channel activation were investigated after the expression of mutant subunits with wild-type α1 and γ2 subunits in tsA201 cells or in Xenopus oocytes. None of the mutations affected [3H]Ro15–4513 binding or impaired allosteric interactions between the low affinity GABA and benzodiazepine sites. Although mutations at position 74 had little effect on [3H]muscimol binding, the Y62F mutation decreased the affinity of the high affinity [3H]muscimol binding sites by ∼6-fold, and the Y62S mutation led to a loss of detectable high affinity binding sites. After expression in oocytes, the EC50 values for both muscimol and GABA-induced activation of Y62F and Y62S receptors were increased by 2- and 6-fold compared with the wild-type. We conclude that Tyr-62 of the β subunit is an important determinant for high affinity agonist binding to the GABAA receptor. γ-aminobutyric acid nicotinic acetylcholine receptor N,N-bis(2-hydroxyethyl)-2-aminoethanesulfonic acid flunitrazepam wild type analysis of variance The GABAAR1is a member of a superfamily of ligand-gated ion channels that includes the nicotinic acetylcholine receptor (nAChR), the glycine receptor, and the serotonin type 3 receptor (1.Dunn S.M.J. Bateson A.N. Martin I.L. Bradley R.C. Harris R.A. International Review of Neurobiology. 35. Academic Press, San Diego, CA1994: 51-96Google Scholar). The GABAAR carries binding sites for a number of therapeutic agents including the benzodiazepines, barbiturates, neurosteroids, some general anesthetics, and possibly also alcohol (1.Dunn S.M.J. Bateson A.N. Martin I.L. Bradley R.C. Harris R.A. International Review of Neurobiology. 35. Academic Press, San Diego, CA1994: 51-96Google Scholar). In brain membranes, there are at least two classes of binding sites for the endogenous neurotransmitter, which differ by more than an order of magnitude in their affinity for GABA or its structural analogues (2.Olsen R.W. Bergman M.O. Van Ness P.C. Lummis S.C. Watkins A.E. Napias C. Greenlee D.V. Mol. Pharmacol. 1981; 19: 217-227PubMed Google Scholar, 3.Olsen R.W. Yang J. King R.G. Dilber A. Stauber G. Ransom R.W. Life Sci. 1986; 39: 1969-1976Crossref PubMed Scopus (75) Google Scholar, 4.Agey M.W. Dunn S.M.J. Biochemistry. 1989; 28: 4200-4208Crossref PubMed Scopus (22) Google Scholar). This heterogeneity in binding was originally thought to reflect the diversity of GABAAR subtypes in brain tissue. However, the presence of both classes of sites in a stable cell line expressing a specific subtype (5.Davies M. Steele J. Hadingham K.L. Whiting P.J. Dunn S.M.J. Can. J. Physiol. Pharmacol. 1994; 72: 337Google Scholar) suggests that both exist in a single receptor molecule. On the basis of biochemical studies, the reasonable correlation between the concentration of agonist required to elicit ion flux and to potentiate the binding of benzodiazepine ligands suggested that the low affinity sites are important for channel gating (see Ref. 6.Sieghart W. Pharmacol. Rev. 1995; 47: 181-233PubMed Google Scholar). However, the role(s) of the high affinity binding sites in receptor function remains unclear.All members of this receptor family are believed to be pentameric complexes formed by homologous subunits assembled to form a central ion channel (7.Nayeem N. Green T.P. Martin I.L. Barnard E.A. J. Neurochem. 1994; 62: 815-818Crossref PubMed Scopus (376) Google Scholar). Recent models (see Ref. 8.Changeux J.-P. Edelstein S.J. Neuron. 1998; 21: 959-980Abstract Full Text Full Text PDF PubMed Scopus (374) Google Scholar) predict that ligand binding sites occur at subunit-subunit interfaces. This was first demonstrated in the nAChR in which the α-γ and α-δ interfaces were implicated in forming nonequivalent binding sites for d-tubocurarine (9.Pedersen S.E. Cohen J.B. Proc. Natl. Acad. Sci. U. S. A. 1990; 87: 2785-2789Crossref PubMed Scopus (223) Google Scholar). In the GABAAR, low affinity GABA sites (i.e. those that have been implicated in channel activation) are thought to be located at the interfaces between the β and α subunits (10.Smith G.B. Olsen R.W. J. Biol. Chem. 1994; 269: 20380-20387Abstract Full Text PDF PubMed Google Scholar, 11.Amin J. Weiss D.S. Nature. 1993; 366: 565-569Crossref PubMed Scopus (372) Google Scholar, 12.Sigel E. Baur R. Kellenberger S. Malherbe P. EMBO J. 1992; 11: 2017-2023Crossref PubMed Scopus (169) Google Scholar, 13.Boileau A.J. Evers A.R. Davis A.F. Czajkowski C. J. Neurosci. 1999; 19: 4847-4854Crossref PubMed Google Scholar), whereas the benzodiazepine binding site is predicted to occur at the homologous α-γ interface (14.Duncalfe L.L. Carpenter M.R. Smillie L.B. Martin I.L. Dunn S.M.J. J. Biol. Chem. 1996; 271: 9209-9214Abstract Full Text Full Text PDF PubMed Scopus (103) Google Scholar, 15.Wieland H.A. Luddens H. Seeburg P.H. J. Biol. Chem. 1992; 267: 1426-1429Abstract Full Text PDF PubMed Google Scholar, 16.Davies M. Martin I.L. Bateson A.N. Hadingham K.L. Whiting P.J. Dunn S.M.J. Neuropharmacology. 1996; 35: 1199-1208Crossref PubMed Scopus (19) Google Scholar, 17.Buhr A. Baur R. Malherbe P. Sigel E. Mol. Pharmacol. 1996; 49: 1080-1084PubMed Google Scholar, 18.Buhr A. Schaerer M.T. Sigel E. Mol. Pharmacol. 1997; 52: 676-686Crossref PubMed Scopus (87) Google Scholar, 19.Amin J. Brooks-Kayal A. Weiss D.S. Mol. Pharmacol. 1997; 51: 833-841Crossref PubMed Scopus (106) Google Scholar). More detailed analyses of the properties of these sites have led to a “loop model” of ligand binding sites (see Ref. 20.Smith G.B. Olsen R.W. Trends Pharmacol. Sci. 1995; 16: 162-168Abstract Full Text PDF PubMed Scopus (451) Google Scholar) in which amino acid residues from at least three discontinuous regions (denoted “loops” A-C) of one subunit together with residues from at least one region of the adjacent subunit (“loop” D) form the binding pocket (see Fig. 1 A).In the GABAAR, evidence for the location of the high affinity agonist site(s) is derived from a number of experimental approaches. Photoaffinity labeling studies first suggested that the β subunit is a major determinant of high affinity binding, because this was the principle site of photoincorporation of [3H]muscimol (21.Deng L. Ransom R.W. Olsen R.W. Biochem. Biophys. Res. Comm. 1986; 138: 1308-1314Crossref PubMed Scopus (69) Google Scholar, 22.Casalotti S.O. Stephenson F.A. Barnard E.A. J. Biol. Chem. 1986; 261: 15013-15016Abstract Full Text PDF PubMed Google Scholar, 23.Bureau M. Olsen R.W. Biochem. Biophys. Res. Comm. 1988; 153: 1006-1011Crossref PubMed Scopus (36) Google Scholar), although another report has given some indication that the α subunit can also be labeled (10.Smith G.B. Olsen R.W. J. Biol. Chem. 1994; 269: 20380-20387Abstract Full Text PDF PubMed Google Scholar). Heterologous expression of different GABAAR subunit combinations indicates that coexpression of α and β subunits is required for high affinity binding, and the α1β3, α1β3γ2, α1β2, and α1β2γ2 combinations have all been shown to form high affinity binding sites for [3H]muscimol (24.Zezula J. Slany A. Sieghart W. Eur. J. Pharmacol. 1996; 301: 207-214Crossref PubMed Scopus (80) Google Scholar,25.Pregenzer J.F. Im W.B. Carter D.B. Thomsen D.R. Mol. Pharmacol. 1993; 43: 801-806PubMed Google Scholar). Furthermore, expression of a tandem construct in which the C terminus of α6 was covalently linked to the N terminus of the β2 subunit produced high affinity binding sites (26.Im W.B. Pregenzer J.F. Binder J.A. Dillon G.H. Alberts G.L. J. Biol. Chem. 1995; 270: 26063-26066Abstract Full Text Full Text PDF PubMed Scopus (81) Google Scholar), although the receptors were nonfunctional.Based on the above observations and homology considerations, we speculated that a high affinity agonist site in the GABAAR may be located at the α-β subunit interface, in which the β subunit would contribute residues in loop D according to the model described above (see Fig. 1 A). Candidate tyrosine residues (at positions 62 and 74) of the β2 subunit were identified by amino acid sequence alignment (Fig.1 B) based on previous work that residues in the homologous positions to Tyr-62 in the α and γ subunits have been implicated in (low affinity) GABA and benzodiazepine binding, respectively (10.Smith G.B. Olsen R.W. J. Biol. Chem. 1994; 269: 20380-20387Abstract Full Text PDF PubMed Google Scholar, 12.Sigel E. Baur R. Kellenberger S. Malherbe P. EMBO J. 1992; 11: 2017-2023Crossref PubMed Scopus (169) Google Scholar, 18.Buhr A. Schaerer M.T. Sigel E. Mol. Pharmacol. 1997; 52: 676-686Crossref PubMed Scopus (87) Google Scholar). Both tyrosine residues were mutated to phenylalanine and to serine to evaluate the relative contributions of the aromatic rings and hydroxyl groups of these tyrosine residues to high affinity muscimol binding.In this report, we demonstrate that Tyr-62 of the β subunit is an important determinant of high affinity muscimol binding. Substitution of phenylalanine at this position decreased the affinity for both mucimol and GABA, whereas substitution by serine led to a loss of detectable high affinity binding sites. In functional assays, both mutations increased the EC50 for channel activation. These results suggest that Tyr-62 of the β subunit is an important determinant for high affinity agonist binding and that although this residue may play some role in receptor activation, high affinity binding per se is not a requirement for channel gating.DISCUSSIONThe elucidation of the mechanisms underlying GABAAR function is important for the understanding of inhibitory synaptic transmission in the central nervous system. The aim of the present study was to identify residues within a specific domain of the β2 subunit of the GABAAR that contribute to high affinity muscimol binding and to define, in part, potential roles of this site in receptor function. Although the subunit stoichiometry of native receptors is unknown, most recent evidence indicates that the recombinant α1β2γ2 GABAAR contains 2α, 2β, and 1γ subunits (39.Farrar S.J. Whiting P.J. Bonnert T.P. McKernan R.M. J. Biol. Chem. 1999; 274: 10100-10104Abstract Full Text Full Text PDF PubMed Scopus (231) Google Scholar, 40.Tretter V. Ehya N. Fuchs K. Sieghart W. J. Neurosci. 1997; 17: 2728-2737Crossref PubMed Google Scholar, 41.Chang Y. Wang R. Barot S. Weiss D.S. J. Neurosci. 1996; 16: 5415-5424Crossref PubMed Google Scholar). One likely arrangement of these subunits within the pentamer has been suggested to be α-β-α-γ-β (40.Tretter V. Ehya N. Fuchs K. Sieghart W. J. Neurosci. 1997; 17: 2728-2737Crossref PubMed Google Scholar,42.Sigel E. Buhr A. Trends Pharmacol. Sci. 1997; 18: 425-429Abstract Full Text PDF PubMed Scopus (345) Google Scholar). This arrangement (see Fig. 1 A) provides one α-γ interface where the benzodiazepine site is thought to be located and two β-α interfaces, each of which may carry a low affinity agonist site (11.Amin J. Weiss D.S. Nature. 1993; 366: 565-569Crossref PubMed Scopus (372) Google Scholar, 12.Sigel E. Baur R. Kellenberger S. Malherbe P. EMBO J. 1992; 11: 2017-2023Crossref PubMed Scopus (169) Google Scholar). The presence of two low affinity sites would be consistent with a Hill coefficient of ∼2 for channel activation (43.Macdonald R.L. Olsen R.W. Annu. Rev. Neurosci. 1994; 17: 569-602Crossref PubMed Scopus (1780) Google Scholar). This leaves two additional interfaces (α-β and γ-β), which could potentially form high affinity muscimol/GABA sites. According to the popular loop model (see Introduction), loop D contributing to these sites would be found in the N-terminal domain of the β2 subunit. Previous work has shown that residues within this domain of the γ and α subunits are determinants of benzodiazepine (17.Buhr A. Baur R. Malherbe P. Sigel E. Mol. Pharmacol. 1996; 49: 1080-1084PubMed Google Scholar) and low affinity GABA (10.Smith G.B. Olsen R.W. J. Biol. Chem. 1994; 269: 20380-20387Abstract Full Text PDF PubMed Google Scholar, 12.Sigel E. Baur R. Kellenberger S. Malherbe P. EMBO J. 1992; 11: 2017-2023Crossref PubMed Scopus (169) Google Scholar) binding, respectively. This provided the rationale for targeting homologous residues in the β2 subunit (Tyr-62 and Tyr-74) to investigate their role in high affinity muscimol binding.The major finding reported here is that Tyr-62 of the β2 subunit is a determinant of high affinity agonist binding. Its substitution by phenylalanine reduced the affinity for both muscimol and GABA (6-fold), whereas its substitution by serine resulted in a dramatic reduction in affinity (>30-fold) such that no high affinity binding was measurable in this mutant. However, receptors containing the Y62S mutation were still functional, albeit with an increased EC50 for channel activation by about 6-fold. Thus high affinity agonist binding does not appear to be obligatory for receptor activation.In a previous study, Sigel et al. (12.Sigel E. Baur R. Kellenberger S. Malherbe P. EMBO J. 1992; 11: 2017-2023Crossref PubMed Scopus (169) Google Scholar) also mutated residue Tyr-62 of the β2 subunit. Although these authors did not investigate receptor binding properties, they found that the Y62L mutation reduced the maximum current elicited by GABA by ∼5-fold, leading to the conclusion that the mutation had disrupted receptor assembly. In the present study, we did not observe any reduction of the maximum current as a result of either phenylalanine or serine substitution at this position, suggesting that there were no major effects on receptor synthesis and expression.There is no general consensus as to the number of agonist binding sites on a single GABAAR (44.Hevers W. Lüddens H. Mol. Neurobiol. 1998; 18: 35-86Crossref PubMed Scopus (410) Google Scholar). There is, however, abundant evidence for the presence of high affinity sites in addition to one or more classes of sites having lower affinity (see Refs. 1.Dunn S.M.J. Bateson A.N. Martin I.L. Bradley R.C. Harris R.A. International Review of Neurobiology. 35. Academic Press, San Diego, CA1994: 51-96Google Scholar and 6.Sieghart W. Pharmacol. Rev. 1995; 47: 181-233PubMed Google Scholar). Previous studies have demonstrated that there are approximately twice as many high affinity sites for muscimol as for flunitrazepam (45.Sigel E. Barnard E.A. J. Biol. Chem. 1984; 259: 7219-7223Abstract Full Text PDF PubMed Google Scholar), suggesting that there are two high affinity sites/receptor. As described above, we predict that these sites are located at the α-β and γ-β interfaces, which by their nature are nonequivalent. Although we have detected no heterogeneity in high affinity [3H]muscimol binding, the bicuculline displacement experiments (Table II) suggest that in the wild type receptor, this antagonist may discriminate between the two putative high affinity agonist sites. Although the Y62F mutation caused a significant decrease in affinity for the agonist, bicuculline binding was apparently unaltered. This result is in agreement with the previous observation that the Y62L mutation did not affect the IC50 values for functional antagonism of GABA-mediated chloride conductance by bicuculline (12.Sigel E. Baur R. Kellenberger S. Malherbe P. EMBO J. 1992; 11: 2017-2023Crossref PubMed Scopus (169) Google Scholar). Conversely, neither of the Tyr-74 mutations reported here affected agonist binding, but they did alter the characteristics of [3H]muscimol displacement by bicuculline. These observations suggest that although muscimol and bicuculline compete for the same binding sites, different subsets of amino acids may be involved in the recognition of the different ligands. Alternatively, the Tyr-74 mutations may have produced changes in the conformation of the receptor, which indirectly affect the binding of bicuculline. Further complexity arises from the apparent preference of bicuculline for binding to the low affinity agonist sites (46.Möhler H Okada T. Mol. Pharmacol. 1978; 14: 256-265PubMed Google Scholar, 47.Enna S.J. Snyder S.H. Mol. Pharmacol. 1977; 13: 442-453PubMed Google Scholar, 48.Olsen R.W. Snowman A.M. J. Neurochem. 1983; 41: 1653-1663Crossref PubMed Scopus (85) Google Scholar, 49.Maksay G. Mol. Pharmacol. 1994; 46: 386-390PubMed Google Scholar). Further studies to explore this novel observation will be required to identify the specific residues with which bicuculline interacts.The presence of multiple agonist binding sites in the GABAA receptor raises the question of their roles in receptor function. Discrepancies between the concentrations of agonists that are required to activate the receptor and agonist affinities that are measured in equilibrium binding assays are generally thought to reflect differences in receptor conformation (i.e. between the activated and desensitized states). In this and many other studies (11.Amin J. Weiss D.S. Nature. 1993; 366: 565-569Crossref PubMed Scopus (372) Google Scholar, 12.Sigel E. Baur R. Kellenberger S. Malherbe P. EMBO J. 1992; 11: 2017-2023Crossref PubMed Scopus (169) Google Scholar, 13.Boileau A.J. Evers A.R. Davis A.F. Czajkowski C. J. Neurosci. 1999; 19: 4847-4854Crossref PubMed Google Scholar) it has been found that micromolar concentrations of GABA and muscimol are required to open the ion channel (Fig. 6), suggesting that the sites involved in channel activation are of intrinsically low affinity, indeed lower than can be measured in direct equilibrium binding studies. In recent functional studies, we have found that the concentrations of GABA and muscimol that induce receptor desensitization are in good agreement with the lower affinity binding component measured directly. 2J. G. Newell and S. M. J. Dunn, unpublished observations. The role of the high affinity sites, however, is less clear.The Tyr-62 mutations disrupted high affinity agonist binding and also increased the EC50 values for channel activation. It is likely, therefore, that the high affinity binding sites may play a role in the efficiency of channel activation. The EC50 value is a macroscopic constant that depends on several microscopic processes, including ligand binding and channel gating (50.Colquhoun D. Br. J. Pharmacol. 1998; 125: 924-947Crossref PubMed Scopus (777) Google Scholar). It is, therefore, difficult to discriminate among the various contributing factors on the basis of concentration-response curves alone. This is particularly true when complications arising from multiple classes of agonist sites are introduced. One possibility is that the high affinity sites are allosterically coupled to other domains intimately involved in channel activation and that their occupancy at low concentrations of agonists increases the affinity of the latter sites to enhance the efficiency of synaptic transmission. It has been theorized that two nonequivalent sites, in nAChR, provide an ideal kinetic mechanism to enhance and potentially accelerate receptor activation, which may satisfy physiological requirements for rapid activation and termination of response (51.Jackson M.B. Proc. Natl. Acad. Sci. U. S. A. 1989; 86: 2199-2203Crossref PubMed Scopus (108) Google Scholar).In the present study, substitution of the tyrosine residue at position 62 by serine had a more dramatic effect than the phenylalanine substitution. Although we have not made multiple amino acid substitutions at this position, the aromaticity of the residue in this position appears to be particularly important in agonist binding. As has been previously reported for agonist binding to the nicotinic acetylcholine receptor (52.Dougherty D.A. Stauffer D.A. Science. 1990; 250: 1558-1560Crossref PubMed Scopus (573) Google Scholar) and for benzodiazepine binding to the GABAAR (32.Davies M. Bateson A.N. Dunn S.M.J. J. Neurochem. 1998; 70: 2188-2194Crossref PubMed Scopus (54) Google Scholar), aromatic residues may be involved in a π-π stacking interaction with the ligand.Detailed analyses of structure-function relationships without knowledge of the crystal structure of the protein should be interpreted with caution (53.Ward W.H.J. Timms D. Fresht A.R. Trends Pharmacol. Sci. 1990; 11: 280-284Abstract Full Text PDF PubMed Scopus (44) Google Scholar). As with all site-directed mutagenesis studies, the major limitation of the present study is that we cannot state with any degree of certainty that implicated residues are directly or indirectly involved in ligand binding. However, the mutations do appear to be specific for the high affinity agonist site and this study provides the first evidence for the structural basis of high affinity binding that has been noted for more than 20 years.In conclusion, we have identified residue Tyr-62 of the β2 subunit as a determinant of high affinity [3H]muscimol binding in the recombinant α1β2γ2 GABAAR. Further, we have shown that the reduction in affinity of high affinity binding site(s) does not have a large effect on receptor activation (54.Im W.B. Pregenzer J.P. Binder J.A. Alberts G.L. Im H.K. Br. J. Pharmacol. 1997; 120: 559-564Crossref PubMed Scopus (14) Google Scholar). It has previously been suggested that the nicotinic acetylcholine receptor carries sites of low and high affinity, and although the former are involved in channel activation, the latter may be important in mediating receptor desensitization (55.Dunn S.M.J. Conti-Tronconi B.M. Raftery M.A. Biochemistry. 1983; 22: 2512-2518Crossref PubMed Scopus (40) Google Scholar). By analogy to the nAChR, the high affinity site(s) of GABAAR may fulfill the same role. Other investigators have likewise suggested that two molecules of GABA are required for activation, and two independent molecules of neurotransmitter are required for desensitization (56.Cash D.J. Subbarao K. Biochemistry. 1987; 26 (7270): 7526Google Scholar). Experiments to examine the consequences of the above mutations on the desensitization of GABAAR are currently in progress. The GABAAR1is a member of a superfamily of ligand-gated ion channels that includes the nicotinic acetylcholine receptor (nAChR), the glycine receptor, and the serotonin type 3 receptor (1.Dunn S.M.J. Bateson A.N. Martin I.L. Bradley R.C. Harris R.A. International Review of Neurobiology. 35. Academic Press, San Diego, CA1994: 51-96Google Scholar). The GABAAR carries binding sites for a number of therapeutic agents including the benzodiazepines, barbiturates, neurosteroids, some general anesthetics, and possibly also alcohol (1.Dunn S.M.J. Bateson A.N. Martin I.L. Bradley R.C. Harris R.A. International Review of Neurobiology. 35. Academic Press, San Diego, CA1994: 51-96Google Scholar). In brain membranes, there are at least two classes of binding sites for the endogenous neurotransmitter, which differ by more than an order of magnitude in their affinity for GABA or its structural analogues (2.Olsen R.W. Bergman M.O. Van Ness P.C. Lummis S.C. Watkins A.E. Napias C. Greenlee D.V. Mol. Pharmacol. 1981; 19: 217-227PubMed Google Scholar, 3.Olsen R.W. Yang J. King R.G. Dilber A. Stauber G. Ransom R.W. Life Sci. 1986; 39: 1969-1976Crossref PubMed Scopus (75) Google Scholar, 4.Agey M.W. Dunn S.M.J. Biochemistry. 1989; 28: 4200-4208Crossref PubMed Scopus (22) Google Scholar). This heterogeneity in binding was originally thought to reflect the diversity of GABAAR subtypes in brain tissue. However, the presence of both classes of sites in a stable cell line expressing a specific subtype (5.Davies M. Steele J. Hadingham K.L. Whiting P.J. Dunn S.M.J. Can. J. Physiol. Pharmacol. 1994; 72: 337Google Scholar) suggests that both exist in a single receptor molecule. On the basis of biochemical studies, the reasonable correlation between the concentration of agonist required to elicit ion flux and to potentiate the binding of benzodiazepine ligands suggested that the low affinity sites are important for channel gating (see Ref. 6.Sieghart W. Pharmacol. Rev. 1995; 47: 181-233PubMed Google Scholar). However, the role(s) of the high affinity binding sites in receptor function remains unclear. All members of this receptor family are believed to be pentameric complexes formed by homologous subunits assembled to form a central ion channel (7.Nayeem N. Green T.P. Martin I.L. Barnard E.A. J. Neurochem. 1994; 62: 815-818Crossref PubMed Scopus (376) Google Scholar). Recent models (see Ref. 8.Changeux J.-P. Edelstein S.J. Neuron. 1998; 21: 959-980Abstract Full Text Full Text PDF PubMed Scopus (374) Google Scholar) predict that ligand binding sites occur at subunit-subunit interfaces. This was first demonstrated in the nAChR in which the α-γ and α-δ interfaces were implicated in forming nonequivalent binding sites for d-tubocurarine (9.Pedersen S.E. Cohen J.B. Proc. Natl. Acad. Sci. U. S. A. 1990; 87: 2785-2789Crossref PubMed Scopus (223) Google Scholar). In the GABAAR, low affinity GABA sites (i.e. those that have been implicated in channel activation) are thought to be located at the interfaces between the β and α subunits (10.Smith G.B. Olsen R.W. J. Biol. Chem. 1994; 269: 20380-20387Abstract Full Text PDF PubMed Google Scholar, 11.Amin J. Weiss D.S. Nature. 1993; 366: 565-569Crossref PubMed Scopus (372) Google Scholar, 12.Sigel E. Baur R. Kellenberger S. Malherbe P. EMBO J. 1992; 11: 2017-2023Crossref PubMed Scopus (169) Google Scholar, 13.Boileau A.J. Evers A.R. Davis A.F. Czajkowski C. J. Neurosci. 1999; 19: 4847-4854Crossref PubMed Google Scholar), whereas the benzodiazepine binding site is predicted to occur at the homologous α-γ interface (14.Duncalfe L.L. Carpenter M.R. Smillie L.B. Martin I.L. Dunn S.M.J. J. Biol. Chem. 1996; 271: 9209-9214Abstract Full Text Full Text PDF PubMed Scopus (103) Google Scholar, 15.Wieland H.A. Luddens H. Seeburg P.H. J. Biol. Chem. 1992; 267: 1426-1429Abstract Full Text PDF PubMed Google Scholar, 16.Davies M. Martin I.L. Bateson A.N. Hadingham K.L. Whiting P.J. Dunn S.M.J. Neuropharmacology. 1996; 35: 1199-1208Crossref PubMed Scopus (19) Google Scholar, 17.Buhr A. Baur R. Malherbe P. Sigel E. Mol. Pharmacol. 1996; 49: 1080-1084PubMed Google Scholar, 18.Buhr A. Schaerer M.T. Sigel E. Mol. Pharmacol. 1997; 52: 676-686Crossref PubMed Scopus (87) Google Scholar, 19.Amin J. Brooks-Kayal A. Weiss D.S. Mol. Pharmacol. 1997; 51: 833-841Crossref PubMed Scopus (106) Google Scholar). More detailed analyses of the properties of these sites have led to a “loop model” of ligand binding sites (see Ref. 20.Smith G.B. Olsen R.W. Trends Pharmacol. Sci. 1995; 16: 162-168Abstract Full Text PDF PubMed Scopus (451) Google Scholar) in which amino acid residues from at least three discontinuous regions (denoted “loops” A-C) of one subunit together with residues from at least one region of the adjacent subunit (“loop” D) form the binding pocket (see Fig. 1 A). In the GABAAR, evidence for the location of the high affinity agonist site(s) is derived from a number of experimental approaches. Photoaffinity labeling studies first suggested that the β subunit is a major determinant of high affinity binding, because this was the principle site of photoincorporation of [3H]muscimol (21.Deng L. Ransom R.W. Olsen R.W. Biochem. Biophys. Res. Comm. 1986; 138: 1308-1314Crossref PubMed Scopus (69) Google Scholar, 22.Casalotti S.O. Stephenson F.A. Barnard E.A. J. Biol. Chem. 1986; 261: 15013-15016Abstract Full Text PDF PubMed Google Scholar, 23.Bureau M. Olsen R.W. Biochem. Biophys. Res. Comm. 1988; 153: 1006-1011Crossref PubMed Scopus (36) Google Scholar), although another report has given some indication that the α subunit can also be labeled (10.Smith G.B. Olsen R.W. J. Biol. Chem. 1994; 269: 20380-20387Abstract Full Text PDF PubMed Google Scholar). Heterologous expression of different GABAAR subunit combinations indicates that coexpression of α and β subunits is required for high affinity binding, and the α1β3, α1β3γ2, α1β2, and α1β2γ2 combinations have all been shown to form high affinity binding sites for [3H]muscimol (24.Zezula J. Slany A. Sieghart W. Eur. J. Pharmacol. 1996; 301: 207-214Crossref PubMed Scopus (80) Google Scholar,25.Pregenzer J.F. Im W.B. Carter D.B. Thomsen D.R. Mol. Pharmacol. 1993; 43: 801-806PubMed Google Scholar). Furthermore, expression of a tandem construct in which the C terminus of α6 was covalently linked to the N terminus of the β2 subunit produced high affinity binding sites (26.Im W.B.