Flunitrazepam Has an Inverse Agonistic Effect on Recombinant α6β2γ2-GABAAReceptors via a Flunitrazepam-binding Site
1997; Elsevier BV; Volume: 272; Issue: 18 Linguagem: Inglês
10.1074/jbc.272.18.11723
ISSN1083-351X
AutoresCharlotte A. E. Hauser, Christian H. Wetzel, Barbara Berning, Franz Gerner, Rainer Rupprecht,
Tópico(s)Pharmacological Receptor Mechanisms and Effects
Resumoγ-Aminobutyric acid type A (GABAA) receptor subtypes containing the α6-subunit are generally thought to be insensitive to the action of benzodiazepine agonists. We describe the specific binding of the benzodiazepine agonist flunitrazepam to α6β2γ2-containing GABAA receptors, which has not been observed before and differs from previous reports. With the whole-cell voltage-clamp technique, we observed a functional discrimination between α1β2γ2- and α6β2γ2-receptors. Different benzodiazepines had different effects on GABA-evoked chloride currents. The agonist flunitrazepam had an inverse agonistic effect, whereas the antagonist flumazenil increased GABA-induced chloride currents. The action of flunitrazepam on the channel activity of α6β2γ2-receptors was opposite to its action on α1β2γ2-receptors. We conclude that flunitrazepam can act as either an agonist or an inverse agonist, depending on the GABAA receptor configuration. γ-Aminobutyric acid type A (GABAA) receptor subtypes containing the α6-subunit are generally thought to be insensitive to the action of benzodiazepine agonists. We describe the specific binding of the benzodiazepine agonist flunitrazepam to α6β2γ2-containing GABAA receptors, which has not been observed before and differs from previous reports. With the whole-cell voltage-clamp technique, we observed a functional discrimination between α1β2γ2- and α6β2γ2-receptors. Different benzodiazepines had different effects on GABA-evoked chloride currents. The agonist flunitrazepam had an inverse agonistic effect, whereas the antagonist flumazenil increased GABA-induced chloride currents. The action of flunitrazepam on the channel activity of α6β2γ2-receptors was opposite to its action on α1β2γ2-receptors. We conclude that flunitrazepam can act as either an agonist or an inverse agonist, depending on the GABAA receptor configuration. Benzodiazepines (BZs) 1The abbreviations used are: BZs, benzodiazepines; GABA, γ-aminobutyric acid; GABAA, γ-aminobutyric acid type A. 1The abbreviations used are: BZs, benzodiazepines; GABA, γ-aminobutyric acid; GABAA, γ-aminobutyric acid type A. act via the γ-aminobutyric acid type A (GABAA) receptor, thereby influencing the chloride influx into the cell (1Macdonald R.L. Olsen R.W. Annu. Rev. Neurosci. 1994; 17: 569-602Google Scholar, 2Lüddens H. Korpi E.R. J. Psychiatr. Res. 1995; 29: 77-94Google Scholar). Studies of a variety of GABAA receptor compositions have led to the conclusion that ternary receptors reconstituted from α-, β-, and γ-subunits have binding sites for BZs, with the high-affinity site being located on the α-subunit (3Pritchett D.B. Lüddens H. Seeburg P.H. Science. 1989; 245: 1389-1392Google Scholar, 4Ymer S. Draguhn A. Wisden W. Werner P. Keinänen K. Schofield P.R. Sprengel R. Pritchett D.B. Seeburg P.H. EMBO J. 1990; 9: 3261-3267Google Scholar). These subunits are differentially distributed throughout the central nervous system (5Wisden W. Laurie D.J. Monyer H. Seeburg P.H. J. Neurosci. 1992; 12: 1040-1062Google Scholar,6Laurie D.J. Seeburg P.H. Wisden W. J. Neurosci. 1992; 12: 1063-1076Google Scholar). It is generally thought that the subunit composition of the GABAA receptor complex organizes the receptor's pharmacology. The best known concept is that the α1βx γ2-receptors, called type I BZ receptors, preferentially bind full BZ agonists with high affinity (3Pritchett D.B. Lüddens H. Seeburg P.H. Science. 1989; 245: 1389-1392Google Scholar) and that the α2/3/5βx γ2-receptors, classified as type II BZ receptors, have a 10-fold lower affinity for BZ agonists (7Pritchett D.B. Seeburg P.H. J. Neurochem. 1990; 54: 1802-1804Google Scholar, 8Pritchett D.B. Seeburg P.H. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 1421-1425Google Scholar, 9Sieghart W. Schlerka W. Eur. J. Pharmacol. 1991; 197: 103-107Google Scholar, 10Gillard N.P. Quirk K. Ragan C.I. McKernan R.M. Eur. J. Pharmacol. 1991; 195: 407-409Google Scholar). The view is widely held that α6- and α4-containing receptors have virtually no affinity for full BZ agonists (11Lüddens H. Pritchett D.B. Köhler M. Killisch I. Keinänen K. Monyer H. Sprengel R. Seeburg P.H. Nature. 1990; 346: 648-651Google Scholar, 12Wisden W. Herb A. Wieland H. Keinänen K. Lüddens H. Seeburg P.H. FEBS Lett. 1991; 289: 227-230Google Scholar, 13Wieland H.A. Lüddens H. Seeburg P.H. J. Biol. Chem. 1992; 267: 1426-1429Google Scholar), although they retain a high affinity for the partial inverse BZ agonist Ro 15-4513 (14Suzdak P.D. Glowa J.R. Crawley J.N. Schwartz R. Skolnick P. Paul S.M.A. Science. 1986; 234: 1243-1247Google Scholar, 15Turner D.M. Sapp D.W. Olsen R.W. J. Pharmacol. Exp. Ther. 1991; 257: 1236-1242Google Scholar). In the mature mammalian brain, the α6-subunit is expressed exclusively in the cerebellum (6Laurie D.J. Seeburg P.H. Wisden W. J. Neurosci. 1992; 12: 1063-1076Google Scholar), where it is preferentially found in granular cells (16Nusser Z. Sieghart W. Stephenson F.A. Somogyi P. J. Neurosci. 1996; 16: 103-114Google Scholar).The studies we report here focus on the BZ pharmacology of the α6β2γ2-receptor. We show that α6β2γ2-GABAAreceptors have high affinity for the BZ agonist flunitrazepam, although the histidine at position 100 in the ligand-binding domain, which is thought to be crucial for BZ binding, is missing (13Wieland H.A. Lüddens H. Seeburg P.H. J. Biol. Chem. 1992; 267: 1426-1429Google Scholar). Furthermore, using the whole-cell voltage-clamp technique on α6β2γ2-transfected cells, we demonstrate that flunitrazepam has an inverse agonistic effect on GABA-induced chloride currents, whereas the BZ antagonist Ro 15-1788 (flumazenil) increases GABA-mediated chloride currents.RESULTS AND DISCUSSIONClassical BZ agonists are generally thought to bind to all ternary GABAA receptor complexes reconstituted from α-, β-, and γ-subunits, except when the α6- or α4-subunit is involved (13Wieland H.A. Lüddens H. Seeburg P.H. J. Biol. Chem. 1992; 267: 1426-1429Google Scholar). This conclusion was drawn indirectly from competition studies with classical BZ agonists as competitors for the inverse agonist-binding site of [3H]Ro 15-4513. A single histidine was found to be essential for prototypic BZ agonist binding (13Wieland H.A. Lüddens H. Seeburg P.H. J. Biol. Chem. 1992; 267: 1426-1429Google Scholar). We analyzed the direct binding of [3H]flunitrazepam to recombinant α1β2γ2- and α6β2γ2-GABAAreceptors. The results for the recombinant α6β2γ2-receptor were unexpected. Scatchard analysis with [3H]flunitrazepam in concentrations between 1.2 and 40 nm revealed a single category of saturable high-affinity binding sites, with aK d value of 8.7 nm for α6β2γ2-receptors compared with a K d value of 3.3 nm for α1β2γ2-receptors (Fig.1, A and B). With this finding for α6β2γ2-receptors, we postulate a binding site for flunitrazepam on this receptor subtype, which has not been detected previously. Whereas α1β2γ2-receptors are found throughout the central nervous system, α6β2γ2-receptors are localized predominantly on cerebellar granule cells in the neighborhood of a variety of excitatory ion channels (12Wisden W. Herb A. Wieland H. Keinänen K. Lüddens H. Seeburg P.H. FEBS Lett. 1991; 289: 227-230Google Scholar, 16Nusser Z. Sieghart W. Stephenson F.A. Somogyi P. J. Neurosci. 1996; 16: 103-114Google Scholar).In competition binding studies, the α1β2γ2-receptors, which are the "classical" GABAA receptors, exhibited the expected displacement of flunitrazepam binding by other BZ agonists (11Lüddens H. Pritchett D.B. Köhler M. Killisch I. Keinänen K. Monyer H. Sprengel R. Seeburg P.H. Nature. 1990; 346: 648-651Google Scholar), whereas for the α6β2γ2-receptors, we obtainedK i values 150–1000-fold higher than the observedK d (Table I). The weak displacement of flunitrazepam binding by classical BZ agonists suggested that we had found a binding site on the α6β2γ2-receptor with a preference for flunitrazepam. We then looked at whether the peripheral BZ receptor ligand 4′-chlorodiazepam (Ro 5-4864), which interacts with a unique low-affinity site on the GABAA receptor, and a peripheral BZ receptor antagonist, the isoquinoline carboxamide derivative PK 11195 (22Basile A.S. Bolger G.T. Lüddens H.W. Skolnick P. J. Pharmacol. Exp. Ther. 1989; 248: 463-469Google Scholar), bound to the flunitrazepam-binding site. We did not detect any flunitrazepam displacement (Table I). Finally, we examined whether other BZ agonists, such as alprazolam and chlordiazepoxide, or BZ inverse agonists, such as methyl-6,7-dimethoxy-4-ethyl-β-carboline 3-carboxylate,n-butyl-β-carboline 3-carboxylate, and desmethyldiazepam, were able to compete for the flunitrazepam-binding site on the α6-subunit, but found only weak displacement or none at all (Table I).Table IBinding affinity of various benzodiazepinesα1β2γ2α6β2γ2K d (nm)aValues represent the mean K d or mean K i ± S.E. from at least three independent experiments done in triplicate.[3H]Flunitrazepam3.3 ± 1.98.7 ± 3.0K i (nm)Diazepam5.5 ± 2.91190 ± 160Clonazepam2.0 ± 0.55000 ± 1900Bromazepam190 ± 708300 ± 2400Lorazepam5.6 ± 1.22900 ± 900Ro 5–4864ND1-bND, not determined; DMCM, methyl-6,7-dimethoxy-4-ethyl-β-carboline 3-carboxylate; β-CCB,n-butyl-β-carboline 3-carboxylate.>10,000PK 11195ND>10,000DMCM8.6 ± 1.8>10,000β-CCB28.8 ± 3.01580 ± 480Chlordiazepoxide2080 ± 750>10,000Alprazolam23.1 ± 2.5>10,000Desmethyldiazepam159 ± 742130 ± 7001-a Values represent the mean K d or mean K i ± S.E. from at least three independent experiments done in triplicate.1-b ND, not determined; DMCM, methyl-6,7-dimethoxy-4-ethyl-β-carboline 3-carboxylate; β-CCB,n-butyl-β-carboline 3-carboxylate. Open table in a new tab We also wanted to test whether the Ro 15-4513 binding site on α6β2γ2 receptors had any influence on the newly described flunitrazepam binding site. Analysis of α1(Arg-101)β2γ2 and α6(His-100)β2γ2 receptors along with the wild-type receptors showed that the sensitivity of [3H]Ro 15-4513 binding to flunitrazepam as well as to diazepam coincided with the presence of the crucial amino acid histidine (Table II). The newly proposed high affinity binding site of flunitrazepam on α6β2γ2 receptors did not compete with the Ro 15-4513 binding site. The BZ antagonist flumazenil exhibited a 100–300-fold lower affinity for mutant receptors and α6β2γ2 receptors than for α1β2γ2 receptors which is in agreement with data previously published (13Wieland H.A. Lüddens H. Seeburg P.H. J. Biol. Chem. 1992; 267: 1426-1429Google Scholar).Table IIBinding affinity of [3H]Ro 15–4513 and displacement by various benzodiazepinesReceptor typeK d, [3H]Ro 15–45132-aValues represent the mean K d or mean K i ± S.E. from at least three independent experiments done in triplicate.K iFlunitrazepamDiazepamFlumazenilnmnmα119.9 ± 2.00.2 ± 0.0523.4 ± 7.410.4 ± 0.5α1(Arg-101)6.6 ± 0.4>10,000>10,000273 ± 15α66.2 ± 1.9>10,000>10,000149 ± 59α6(His-100)5.1 ± 1.91.2 ± 0.01153 ± 9096 ± 442-a Values represent the mean K d or mean K i ± S.E. from at least three independent experiments done in triplicate. Open table in a new tab To investigate whether an exchange of the crucial amino acid histidine, responsible for high-affinity BZ agonist binding, would change the flunitrazepam affinity for α6β2γ2-receptors, we used two mutants, already described elsewhere (13Wieland H.A. Lüddens H. Seeburg P.H. J. Biol. Chem. 1992; 267: 1426-1429Google Scholar), for [3H]flunitrazepam binding studies. These mutant receptors, α1(Arg-101)β2γ2and α6(His-100)β2γ2, showed a single class of binding sites, with mean K d values of 19.7 and 5.4 nm, respectively (Fig. 1, A andB). This gives evidence that histidine 100 is not the exclusive determinant of flunitrazepam binding as postulated earlier (13Wieland H.A. Lüddens H. Seeburg P.H. J. Biol. Chem. 1992; 267: 1426-1429Google Scholar).The response to GABA characterized by the electrophysiological investigation of recombinant GABAA receptors is known to be enhanced consistently by flunitrazepam (23Malherbe P. Draguhn A. Multhaup G. Beyreuther K. Möhler H. Mol. Brain Res. 1990; 8: 199-208Google Scholar, 24Wafford K.A. Whiting P.J. Kemp J.A. Mol. Pharmacol. 1993; 43: 240-244Google Scholar), and this effect can be blocked completely by flumazenil (3Pritchett D.B. Lüddens H. Seeburg P.H. Science. 1989; 245: 1389-1392Google Scholar, 24Wafford K.A. Whiting P.J. Kemp J.A. Mol. Pharmacol. 1993; 43: 240-244Google Scholar). In the case of the rat α1β2γ2-receptor, 1 μm flunitrazepam has been shown to enhance 1 μm GABA-evoked currents by 64% (25Kleingoor C. Ewert M. von Blankenfeld G. Seeburg P.H. Kettenmann H. Neurosci. Lett. 1991; 130: 169-172Google Scholar). We observed that α1β2γ2-receptors reacted as expected to the coapplication of GABA and flunitrazepam. Flunitrazepam (10 μm) markedly affected the amplitude of the peak current at both GABA concentrations used and in a dose-dependent manner. At 1 μm GABA, we detected a 117 ± 46% increase in the amplitude, whereas coapplication with a higher concentration of GABA (10 μm) elicited a smaller potentiation of 46 ± 6% (Fig. 2 A).Figure 2Effects of flunitrazepam on recombinant GABAA receptors. Inward currents of HEK 293 cells activated by 1 or 10 μm GABA and measured at −50 mV are shown for α1β2γ2-receptors (A) and α6β2γ2-receptors (B) as a representative experiment of 6–12 independent experiments. Transfected cells were first challenged with GABA for 5 s and again challenged with GABA plus 10 μmflunitrazepam. Black curves (b) are control GABA responses for the situation using either 1 μm GABA (left panels) or 10 μm GABA (right panels) without flunitrazepam. Gray curves(a) are GABA responses in the presence of 10 μm flunitrazepam. Also shown (C) is the effect of 10 μm flunitrazepam on α6β2γ2-receptors at different concentrations of GABA expressed as mean ± S.E. Data are expressed in terms of percent inhibition, which was calculated by the following formula: (1 −I GABA+10 μm flunitrazepam/I GABA) × 100. The asterisks indicate a significant inhibitory effect (t test for paired samples; p < 0.05).View Large Image Figure ViewerDownload (PPT)For α6β2γ2-receptors, however, 10 μm flunitrazepam and 10 μm GABA resulted in a significant reduction of control currents to 76 ± 12%, whereas coapplication of 1 μm GABA and 10 μm flunitrazepam did not produce a significant change in amplitude (Fig. 2, B and C). This effect of flunitrazepam was shown to be dependent on the concentration of GABA and became significant at a concentration of 6 μm GABA (Fig. 2 C). Our data thus indicate that flunitrazepam acts on α6β2γ2-receptors in an inverse agonistic manner.Our investigation of the effects of flumazenil on GABA-evoked chloride currents showed opposite effects for the two receptors studied. For cells expressing α1β2γ2-receptors, 10 μm flumazenil reduced the amplitude of chloride currents induced by application of 10 μm GABA to 89 ± 3% of control, whereas for cells expressing α6β2γ2-receptors, there was an increase to 118 ± 3% of control induced by application of 1 μm GABA and 10 μm flumazenil (Figs.3 (A and B) and 4A). We have used different GABA concentrations depending on the receptor subtype because we found in an earlier investigation (17Hauser C.A.E. Wetzel C.H.R. Rupprecht R. Holsboer F. Biochem. Biophys. Res. Commun. 1996; 219: 531-536Google Scholar) that α6β2γ2-receptors effectively responded to GABA at lower concentrations than α1β2γ2-receptors. The effect of flumazenil on α6β2γ2-receptor-mediated responses is low compared with previous reports on Sf-9 cells and HEK 293 cells (26Im W.B. Im H.K. Pregenzer J.F. Hamilton B.J. Carter D.B. Jacobsen E.J. TenBrink R.E. VonVoigtlander P.F. Br. J. Pharmacol. 1993; 110: 677-680Google Scholar). In an earlier investigation, an apparent agonistic activity of flumazenil was reported for a binary GABAAreceptor using the rat α1- and β2-subunits expressed in Xenopus oocytes (23Malherbe P. Draguhn A. Multhaup G. Beyreuther K. Möhler H. Mol. Brain Res. 1990; 8: 199-208Google Scholar). Here, 10 μmflumazenil potentiated the response induced by 2–10 μmGABA to 175% of control.Figure 3Effects of flumazenil or Ro 15-4513 on recombinant GABAA receptors. Inward currents of HEK 293 cells activated by 1 μm (B) or 10 μm (A) GABA and measured at −50 mV are shown for α1β2γ2-receptors (A) and α6β2γ2-receptors (B) as a representative experiment of 6–12 independent experiments. Transfected cells were similarly treated and recorded as described in the legend of Fig. 2. Black curves(b) are control GABA responses for either 1 μm(B) or 10 μm (A) GABA without BZ.Gray curves (a) are GABA responses in the presence of either flumazenil (left panels) or Ro 15-4513 (right panels). For flumazenil (10 μm) application, in cells expressing α6β2γ2-receptors, application of 1 μm GABA induced chloride currents with a mean amplitude of 639 ± 213 pA, whereas in cells expressing α1β2γ2-receptors, application of 10 μm GABA induced chloride currents with a mean amplitude of 2585 ± 435 pA. For Ro 15-4513 (1 μm) application, in cells expressing α6β2γ2-receptors, application of 1 μm GABA induced chloride currents with a mean amplitude of 984 ± 230 pA, whereas in cells expressing α1β2γ2-receptors, application of 10 μm GABA induced chloride currents with a mean amplitude of 2835 ± 655 pA.View Large Image Figure ViewerDownload (PPT)Coapplication of flunitrazepam and flumazenil to recombinant α1β2γ2- and α6β2γ2-receptors (Fig.4 B) showed that in the case of the α6β2γ2-receptors, the inverse agonistic properties of flunitrazepam were reduced by the agonistic action of flumazenil. As expected, for the α1β2γ2-receptors, the agonistic properties of flunitrazepam were almost completely suppressed by flumazenil.Figure 4Relative change in the amplitude of GABA-induced chloride currents of recombinant α1β2γ2- and α6β2γ2-receptors. A, application of different BZ receptor ligands to α1β2γ2- and α6β2γ2-receptors;B, coapplication of flunitrazepam and flumazenil. HEK 293 cells expressing α1β2γ2- or α6β2γ2-receptors were first challenged with GABA for 5 s. Cells were again challenged with GABA plus 10 μm flunitrazepam. A washout period of 5 min followed. Cells were further challenged with GABA plus 10 μm flunitrazepam and 10 μm flumazenil. After a new 5-min washout period, the last application of GABA plus 10 μm flumazenil followed. Transfected cells were treated similarly as described in the legend of Fig. 3 using 1 μmGABA for α6β2γ2-receptors and 10 μm GABA for α1β2γ2-receptors.View Large Image Figure ViewerDownload (PPT)The partial inverse agonist Ro 15-4513 had the opposite effect on GABA-induced currents in α1β2γ2- and α6β2γ2-receptors. In cells expressing α1β2γ2-receptors, 1 μm Ro 15-4513 decreased the amplitude of 10 μm GABA-induced chloride currents. For α6β2γ2-receptors with 1 μm GABA, there was a comparable change in current amplitude, but in the opposite direction (Figs. 3 (A andB) and 4A).In this study, we demonstrate that the two receptor subtypes α1β2γ2 and α6β2γ2 react with the BZ receptor ligands flunitrazepam, flumazenil, and Ro 15-4513 in a functionally opposite manner. The exact mechanism underlying the distinct BZ pharmacology of the α6β2γ2-receptor is not clear at present. Our findings, together with those of earlier investigations (13Wieland H.A. Lüddens H. Seeburg P.H. J. Biol. Chem. 1992; 267: 1426-1429Google Scholar, 26Im W.B. Im H.K. Pregenzer J.F. Hamilton B.J. Carter D.B. Jacobsen E.J. TenBrink R.E. VonVoigtlander P.F. Br. J. Pharmacol. 1993; 110: 677-680Google Scholar), lead us to conclude that distinct amino acids of the BZ-binding pocket determine the BZ agonistic or inverse agonistic activity of the GABAA receptor subtype. For the wild-type α6β2γ2-receptor, we postulate that after binding of flunitrazepam, the bulky amino acid arginine induces a structural change that modifies the channel in a way that decreases chloride ion flux. If the smaller amino acid histidine (wild-type α1β2γ2-receptor) is present, such a hindrance of influx may be absent. Studies involving site-directed mutagenesis should be conducted to elucidate the molecular characteristics of the flunitrazepam-binding site on the α6β2γ2-receptor. Benzodiazepines (BZs) 1The abbreviations used are: BZs, benzodiazepines; GABA, γ-aminobutyric acid; GABAA, γ-aminobutyric acid type A. 1The abbreviations used are: BZs, benzodiazepines; GABA, γ-aminobutyric acid; GABAA, γ-aminobutyric acid type A. act via the γ-aminobutyric acid type A (GABAA) receptor, thereby influencing the chloride influx into the cell (1Macdonald R.L. Olsen R.W. Annu. Rev. Neurosci. 1994; 17: 569-602Google Scholar, 2Lüddens H. Korpi E.R. J. Psychiatr. Res. 1995; 29: 77-94Google Scholar). Studies of a variety of GABAA receptor compositions have led to the conclusion that ternary receptors reconstituted from α-, β-, and γ-subunits have binding sites for BZs, with the high-affinity site being located on the α-subunit (3Pritchett D.B. Lüddens H. Seeburg P.H. Science. 1989; 245: 1389-1392Google Scholar, 4Ymer S. Draguhn A. Wisden W. Werner P. Keinänen K. Schofield P.R. Sprengel R. Pritchett D.B. Seeburg P.H. EMBO J. 1990; 9: 3261-3267Google Scholar). These subunits are differentially distributed throughout the central nervous system (5Wisden W. Laurie D.J. Monyer H. Seeburg P.H. J. Neurosci. 1992; 12: 1040-1062Google Scholar,6Laurie D.J. Seeburg P.H. Wisden W. J. Neurosci. 1992; 12: 1063-1076Google Scholar). It is generally thought that the subunit composition of the GABAA receptor complex organizes the receptor's pharmacology. The best known concept is that the α1βx γ2-receptors, called type I BZ receptors, preferentially bind full BZ agonists with high affinity (3Pritchett D.B. Lüddens H. Seeburg P.H. Science. 1989; 245: 1389-1392Google Scholar) and that the α2/3/5βx γ2-receptors, classified as type II BZ receptors, have a 10-fold lower affinity for BZ agonists (7Pritchett D.B. Seeburg P.H. J. Neurochem. 1990; 54: 1802-1804Google Scholar, 8Pritchett D.B. Seeburg P.H. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 1421-1425Google Scholar, 9Sieghart W. Schlerka W. Eur. J. Pharmacol. 1991; 197: 103-107Google Scholar, 10Gillard N.P. Quirk K. Ragan C.I. McKernan R.M. Eur. J. Pharmacol. 1991; 195: 407-409Google Scholar). The view is widely held that α6- and α4-containing receptors have virtually no affinity for full BZ agonists (11Lüddens H. Pritchett D.B. Köhler M. Killisch I. Keinänen K. Monyer H. Sprengel R. Seeburg P.H. Nature. 1990; 346: 648-651Google Scholar, 12Wisden W. Herb A. Wieland H. Keinänen K. Lüddens H. Seeburg P.H. FEBS Lett. 1991; 289: 227-230Google Scholar, 13Wieland H.A. Lüddens H. Seeburg P.H. J. Biol. Chem. 1992; 267: 1426-1429Google Scholar), although they retain a high affinity for the partial inverse BZ agonist Ro 15-4513 (14Suzdak P.D. Glowa J.R. Crawley J.N. Schwartz R. Skolnick P. Paul S.M.A. Science. 1986; 234: 1243-1247Google Scholar, 15Turner D.M. Sapp D.W. Olsen R.W. J. Pharmacol. Exp. Ther. 1991; 257: 1236-1242Google Scholar). In the mature mammalian brain, the α6-subunit is expressed exclusively in the cerebellum (6Laurie D.J. Seeburg P.H. Wisden W. J. Neurosci. 1992; 12: 1063-1076Google Scholar), where it is preferentially found in granular cells (16Nusser Z. Sieghart W. Stephenson F.A. Somogyi P. J. Neurosci. 1996; 16: 103-114Google Scholar). The studies we report here focus on the BZ pharmacology of the α6β2γ2-receptor. We show that α6β2γ2-GABAAreceptors have high affinity for the BZ agonist flunitrazepam, although the histidine at position 100 in the ligand-binding domain, which is thought to be crucial for BZ binding, is missing (13Wieland H.A. Lüddens H. Seeburg P.H. J. Biol. Chem. 1992; 267: 1426-1429Google Scholar). Furthermore, using the whole-cell voltage-clamp technique on α6β2γ2-transfected cells, we demonstrate that flunitrazepam has an inverse agonistic effect on GABA-induced chloride currents, whereas the BZ antagonist Ro 15-1788 (flumazenil) increases GABA-mediated chloride currents. RESULTS AND DISCUSSIONClassical BZ agonists are generally thought to bind to all ternary GABAA receptor complexes reconstituted from α-, β-, and γ-subunits, except when the α6- or α4-subunit is involved (13Wieland H.A. Lüddens H. Seeburg P.H. J. Biol. Chem. 1992; 267: 1426-1429Google Scholar). This conclusion was drawn indirectly from competition studies with classical BZ agonists as competitors for the inverse agonist-binding site of [3H]Ro 15-4513. A single histidine was found to be essential for prototypic BZ agonist binding (13Wieland H.A. Lüddens H. Seeburg P.H. J. Biol. Chem. 1992; 267: 1426-1429Google Scholar). We analyzed the direct binding of [3H]flunitrazepam to recombinant α1β2γ2- and α6β2γ2-GABAAreceptors. The results for the recombinant α6β2γ2-receptor were unexpected. Scatchard analysis with [3H]flunitrazepam in concentrations between 1.2 and 40 nm revealed a single category of saturable high-affinity binding sites, with aK d value of 8.7 nm for α6β2γ2-receptors compared with a K d value of 3.3 nm for α1β2γ2-receptors (Fig.1, A and B). With this finding for α6β2γ2-receptors, we postulate a binding site for flunitrazepam on this receptor subtype, which has not been detected previously. Whereas α1β2γ2-receptors are found throughout the central nervous system, α6β2γ2-receptors are localized predominantly on cerebellar granule cells in the neighborhood of a variety of excitatory ion channels (12Wisden W. Herb A. Wieland H. Keinänen K. Lüddens H. Seeburg P.H. FEBS Lett. 1991; 289: 227-230Google Scholar, 16Nusser Z. Sieghart W. Stephenson F.A. Somogyi P. J. Neurosci. 1996; 16: 103-114Google Scholar).In competition binding studies, the α1β2γ2-receptors, which are the "classical" GABAA receptors, exhibited the expected displacement of flunitrazepam binding by other BZ agonists (11Lüddens H. Pritchett D.B. Köhler M. Killisch I. Keinänen K. Monyer H. Sprengel R. Seeburg P.H. Nature. 1990; 346: 648-651Google Scholar), whereas for the α6β2γ2-receptors, we obtainedK i values 150–1000-fold higher than the observedK d (Table I). The weak displacement of flunitrazepam binding by classical BZ agonists suggested that we had found a binding site on the α6β2γ2-receptor with a preference for flunitrazepam. We then looked at whether the peripheral BZ receptor ligand 4′-chlorodiazepam (Ro 5-4864), which interacts with a unique low-affinity site on the GABAA receptor, and a peripheral BZ receptor antagonist, the isoquinoline carboxamide derivative PK 11195 (22Basile A.S. Bolger G.T. Lüddens H.W. Skolnick P. J. Pharmacol. Exp. Ther. 1989; 248: 463-469Google Scholar), bound to the flunitrazepam-binding site. We did not detect any flunitrazepam displacement (Table I). Finally, we examined whether other BZ agonists, such as alprazolam and chlordiazepoxide, or BZ inverse agonists, such as methyl-6,7-dimethoxy-4-ethyl-β-carboline 3-carboxylate,n-butyl-β-carboline 3-carboxylate, and desmethyldiazepam, were able to compete for the flunitrazepam-binding site on the α6-subunit, but found only weak displacement or none at all (Table I).Table IBinding affinity of various benzodiazepinesα1β2γ2α6β2γ2K d (nm)aValues represent the mean K d or mean K i ± S.E. from at least three independent experiments done in triplicate.[3H]Flunitrazepam3.3 ± 1.98.7 ± 3.0K i (nm)Diazepam5.5 ± 2.91190 ± 160Clonazepam2.0 ± 0.55000 ± 1900Bromazepam190 ± 708300 ± 2400Lorazepam5.6 ± 1.22900 ± 900Ro 5–4864ND1-bND, not determined; DMCM, methyl-6,7-dimethoxy-4-ethyl-β-carboline 3-carboxylate; β-CCB,n-butyl-β-carboline 3-carboxylate.>10,000PK 11195ND>10,000DMCM8.6 ± 1.8>10,000β-CCB28.8 ± 3.01580 ± 480Chlordiazepoxide2080 ± 750>10,000Alprazolam23.1 ± 2.5>10,000Desmethyldiazepam159 ± 742130 ± 7001-a Values represent the mean K d or mean K i ± S.E. from at least three independent experiments done in triplicate.1-b ND, not determined; DMCM, methyl-6,7-dimethoxy-4-ethyl-β-carboline 3-carboxylate; β-CCB,n-butyl-β-carboline 3-carboxylate. Open table in a new tab We also wanted to test whether the Ro 15-4513 binding site on α6β2γ2 receptors had any influence on the newly described flunitrazepam binding site. Analysis of α1(Arg-101)β2γ2 and α6(His-100)β2γ2 receptors along with the wild-type receptors showed that the sensitivity of [3H]Ro 15-4513 binding to flunitrazepam as well as to diazepam coincided with the presence of the crucial amino acid histidine (Table II). The newly proposed high affinity binding site of flunitrazepam on α6β2γ2 receptors did not compete with the Ro 15-4513 binding site. The BZ antagonist flumazenil exhibited a 100–300-fold lower affinity for mutant receptors and α6β2γ2 receptors than for α1β2γ2 receptors which is in agreement with data previously published (13Wieland H.A. Lüddens H. Seeburg P.H. J. Biol. Chem. 1992; 267: 1426-1429Google Scholar).Table IIBinding affinity of [3H]Ro 15–4513 and displacement by various benzodiazepinesReceptor typeK d, [3H]Ro 15–45132-aValues represent the mean K d or mean K i ± S.E. from at least three independent experiments done in triplicate.K iFlunitrazepamDiazepamFlumazenilnmnmα119.9 ± 2.00.2 ± 0.0523.4 ± 7.410.4 ± 0.5α1(Arg-101)6.6 ± 0.4>10,000>10,000273 ± 15α66.2 ± 1.9>10,000>10,000149 ± 59α6(His-100)5.1 ± 1.91.2 ± 0.01153 ± 9096 ± 442-a Values represent the mean K d or mean K i ± S.E. from at least three independent experiments done in triplicate. Open table in a new tab To investigate whether an exchange of the crucial amino acid histidine, responsible for high-affinity BZ agonist binding, would change the flunitrazepam affinity for α6β2γ2-receptors, we used two mutants, already described elsewhere (13Wieland H.A. Lüddens H. Seeburg P.H. J. Biol. Chem. 1992; 267: 1426-1429Google Scholar), for [3H]flunitrazepam binding studies. These mutant receptors, α1(Arg-101)β2γ2and α6(His-100)β2γ2, showed a single class of binding sites, with mean K d values of 19.7 and 5.4 nm, respectively (Fig. 1, A andB). This gives evidence that histidine 100 is not the exclusive determinant of flunitrazepam binding as postulated earlier (13Wieland H.A. Lüddens H. Seeburg P.H. J. Biol. Chem. 1992; 267: 1426-1429Google Scholar).The response to GABA characterized by the electrophysiological investigation of recombinant GABAA receptors is known to be enhanced consistently by flunitrazepam (23Malherbe P. Draguhn A. Multhaup G. Beyreuther K. Möhler H. Mol. Brain Res. 1990; 8: 199-208Google Scholar, 24Wafford K.A. Whiting P.J. Kemp J.A. Mol. Pharmacol. 1993; 43: 240-244Google Scholar), and this effect can be blocked completely by flumazenil (3Pritchett D.B. Lüddens H. Seeburg P.H. Science. 1989; 245: 1389-1392Google Scholar, 24Wafford K.A. Whiting P.J. Kemp J.A. Mol. Pharmacol. 1993; 43: 240-244Google Scholar). In the case of the rat α1β2γ2-receptor, 1 μm flunitrazepam has been shown to enhance 1 μm GABA-evoked currents by 64% (25Kleingoor C. Ewert M. von Blankenfeld G. Seeburg P.H. Kettenmann H. Neurosci. Lett. 1991; 130: 169-172Google Scholar). We observed that α1β2γ2-receptors reacted as expected to the coapplication of GABA and flunitrazepam. Flunitrazepam (10 μm) markedly affected the amplitude of the peak current at both GABA concentrations used and in a dose-dependent manner. At 1 μm GABA, we detected a 117 ± 46% increase in the amplitude, whereas coapplication with a higher concentration of GABA (10 μm) elicited a smaller potentiation of 46 ± 6% (Fig. 2 A).For α6β2γ2-receptors, however, 10 μm flunitrazepam and 10 μm GABA resulted in a significant reduction of control currents to 76 ± 12%, whereas coapplication of 1 μm GABA and 10 μm flunitrazepam did not produce a significant change in amplitude (Fig. 2, B and C). This effect of flunitrazepam was shown to be dependent on the concentration of GABA and became significant at a concentration of 6 μm GABA (Fig. 2 C). Our data thus indicate that flunitrazepam acts on α6β2γ2-receptors in an inverse agonistic manner.Our investigation of the effects of flumazenil on GABA-evoked chloride currents showed opposite effects for the two receptors studied. For cells expressing α1β2γ2-receptors, 10 μm flumazenil reduced the amplitude of chloride currents induced by application of 10 μm GABA to 89 ± 3% of control, whereas for cells expressing α6β2γ2-receptors, there was an increase to 118 ± 3% of control induced by application of 1 μm GABA and 10 μm flumazenil (Figs.3 (A and B) and 4A). We have used different GABA concentrations depending on the receptor subtype because we found in an earlier investigation (17Hauser C.A.E. Wetzel C.H.R. Rupprecht R. Holsboer F. Biochem. Biophys. Res. Commun. 1996; 219: 531-536Google Scholar) that α6β2γ2-receptors effectively responded to GABA at lower concentrations than α1β2γ2-receptors. The effect of flumazenil on α6β2γ2-receptor-mediated responses is low compared with previous reports on Sf-9 cells and HEK 293 cells (26Im W.B. Im H.K. Pregenzer J.F. Hamilton B.J. Carter D.B. Jacobsen E.J. TenBrink R.E. VonVoigtlander P.F. Br. J. Pharmacol. 1993; 110: 677-680Google Scholar). In an earlier investigation, an apparent agonistic activity of flumazenil was reported for a binary GABAAreceptor using the rat α1- and β2-subunits expressed in Xenopus oocytes (23Malherbe P. Draguhn A. Multhaup G. Beyreuther K. Möhler H. Mol. Brain Res. 1990; 8: 199-208Google Scholar). Here, 10 μmflumazenil potentiated the response induced by 2–10 μmGABA to 175% of control.Figure 3Effects of flumazenil or Ro 15-4513 on recombinant GABAA receptors. Inward currents of HEK 293 cells activated by 1 μm (B) or 10 μm (A) GABA and measured at −50 mV are shown for α1β2γ2-receptors (A) and α6β2γ2-receptors (B) as a representative experiment of 6–12 independent experiments. Transfected cells were similarly treated and recorded as described in the legend of Fig. 2. Black curves(b) are control GABA responses for either 1 μm(B) or 10 μm (A) GABA without BZ.Gray curves (a) are GABA responses in the presence of either flumazenil (left panels) or Ro 15-4513 (right panels). For flumazenil (10 μm) application, in cells expressing α6β2γ2-receptors, application of 1 μm GABA induced chloride currents with a mean amplitude of 639 ± 213 pA, whereas in cells expressing α1β2γ2-receptors, application of 10 μm GABA induced chloride currents with a mean amplitude of 2585 ± 435 pA. For Ro 15-4513 (1 μm) application, in cells expressing α6β2γ2-receptors, application of 1 μm GABA induced chloride currents with a mean amplitude of 984 ± 230 pA, whereas in cells expressing α1β2γ2-receptors, application of 10 μm GABA induced chloride currents with a mean amplitude of 2835 ± 655 pA.View Large Image Figure ViewerDownload (PPT)Coapplication of flunitrazepam and flumazenil to recombinant α1β2γ2- and α6β2γ2-receptors (Fig.4 B) showed that in the case of the α6β2γ2-receptors, the inverse agonistic properties of flunitrazepam were reduced by the agonistic action of flumazenil. As expected, for the α1β2γ2-receptors, the agonistic properties of flunitrazepam were almost completely suppressed by flumazenil.Figure 4Relative change in the amplitude of GABA-induced chloride currents of recombinant α1β2γ2- and α6β2γ2-receptors. A, application of different BZ receptor ligands to α1β2γ2- and α6β2γ2-receptors;B, coapplication of flunitrazepam and flumazenil. HEK 293 cells expressing α1β2γ2- or α6β2γ2-receptors were first challenged with GABA for 5 s. Cells were again challenged with GABA plus 10 μm flunitrazepam. A washout period of 5 min followed. Cells were further challenged with GABA plus 10 μm flunitrazepam and 10 μm flumazenil. After a new 5-min washout period, the last application of GABA plus 10 μm flumazenil followed. Transfected cells were treated similarly as described in the legend of Fig. 3 using 1 μmGABA for α6β2γ2-receptors and 10 μm GABA for α1β2γ2-receptors.View Large Image Figure ViewerDownload (PPT)The partial inverse agonist Ro 15-4513 had the opposite effect on GABA-induced currents in α1β2γ2- and α6β2γ2-receptors. In cells expressing α1β2γ2-receptors, 1 μm Ro 15-4513 decreased the amplitude of 10 μm GABA-induced chloride currents. For α6β2γ2-receptors with 1 μm GABA, there was a comparable change in current amplitude, but in the opposite direction (Figs. 3 (A andB) and 4A).In this study, we demonstrate that the two receptor subtypes α1β2γ2 and α6β2γ2 react with the BZ receptor ligands flunitrazepam, flumazenil, and Ro 15-4513 in a functionally opposite manner. The exact mechanism underlying the distinct BZ pharmacology of the α6β2γ2-receptor is not clear at present. Our findings, together with those of earlier investigations (13Wieland H.A. Lüddens H. Seeburg P.H. J. Biol. Chem. 1992; 267: 1426-1429Google Scholar, 26Im W.B. Im H.K. Pregenzer J.F. Hamilton B.J. Carter D.B. Jacobsen E.J. TenBrink R.E. VonVoigtlander P.F. Br. J. Pharmacol. 1993; 110: 677-680Google Scholar), lead us to conclude that distinct amino acids of the BZ-binding pocket determine the BZ agonistic or inverse agonistic activity of the GABAA receptor subtype. For the wild-type α6β2γ2-receptor, we postulate that after binding of flunitrazepam, the bulky amino acid arginine induces a structural change that modifies the channel in a way that decreases chloride ion flux. If the smaller amino acid histidine (wild-type α1β2γ2-receptor) is present, such a hindrance of influx may be absent. Studies involving site-directed mutagenesis should be conducted to elucidate the molecular characteristics of the flunitrazepam-binding site on the α6β2γ2-receptor. Classical BZ agonists are generally thought to bind to all ternary GABAA receptor complexes reconstituted from α-, β-, and γ-subunits, except when the α6- or α4-subunit is involved (13Wieland H.A. Lüddens H. Seeburg P.H. J. Biol. Chem. 1992; 267: 1426-1429Google Scholar). This conclusion was drawn indirectly from competition studies with classical BZ agonists as competitors for the inverse agonist-binding site of [3H]Ro 15-4513. A single histidine was found to be essential for prototypic BZ agonist binding (13Wieland H.A. Lüddens H. Seeburg P.H. J. Biol. Chem. 1992; 267: 1426-1429Google Scholar). We analyzed the direct binding of [3H]flunitrazepam to recombinant α1β2γ2- and α6β2γ2-GABAAreceptors. The results for the recombinant α6β2γ2-receptor were unexpected. Scatchard analysis with [3H]flunitrazepam in concentrations between 1.2 and 40 nm revealed a single category of saturable high-affinity binding sites, with aK d value of 8.7 nm for α6β2γ2-receptors compared with a K d value of 3.3 nm for α1β2γ2-receptors (Fig.1, A and B). With this finding for α6β2γ2-receptors, we postulate a binding site for flunitrazepam on this receptor subtype, which has not been detected previously. Whereas α1β2γ2-receptors are found throughout the central nervous system, α6β2γ2-receptors are localized predominantly on cerebellar granule cells in the neighborhood of a variety of excitatory ion channels (12Wisden W. Herb A. Wieland H. Keinänen K. Lüddens H. Seeburg P.H. FEBS Lett. 1991; 289: 227-230Google Scholar, 16Nusser Z. Sieghart W. Stephenson F.A. Somogyi P. J. Neurosci. 1996; 16: 103-114Google Scholar). In competition binding studies, the α1β2γ2-receptors, which are the "classical" GABAA receptors, exhibited the expected displacement of flunitrazepam binding by other BZ agonists (11Lüddens H. Pritchett D.B. Köhler M. Killisch I. Keinänen K. Monyer H. Sprengel R. Seeburg P.H. Nature. 1990; 346: 648-651Google Scholar), whereas for the α6β2γ2-receptors, we obtainedK i values 150–1000-fold higher than the observedK d (Table I). The weak displacement of flunitrazepam binding by classical BZ agonists suggested that we had found a binding site on the α6β2γ2-receptor with a preference for flunitrazepam. We then looked at whether the peripheral BZ receptor ligand 4′-chlorodiazepam (Ro 5-4864), which interacts with a unique low-affinity site on the GABAA receptor, and a peripheral BZ receptor antagonist, the isoquinoline carboxamide derivative PK 11195 (22Basile A.S. Bolger G.T. Lüddens H.W. Skolnick P. J. Pharmacol. Exp. Ther. 1989; 248: 463-469Google Scholar), bound to the flunitrazepam-binding site. We did not detect any flunitrazepam displacement (Table I). Finally, we examined whether other BZ agonists, such as alprazolam and chlordiazepoxide, or BZ inverse agonists, such as methyl-6,7-dimethoxy-4-ethyl-β-carboline 3-carboxylate,n-butyl-β-carboline 3-carboxylate, and desmethyldiazepam, were able to compete for the flunitrazepam-binding site on the α6-subunit, but found only weak displacement or none at all (Table I). We also wanted to test whether the Ro 15-4513 binding site on α6β2γ2 receptors had any influence on the newly described flunitrazepam binding site. Analysis of α1(Arg-101)β2γ2 and α6(His-100)β2γ2 receptors along with the wild-type receptors showed that the sensitivity of [3H]Ro 15-4513 binding to flunitrazepam as well as to diazepam coincided with the presence of the crucial amino acid histidine (Table II). The newly proposed high affinity binding site of flunitrazepam on α6β2γ2 receptors did not compete with the Ro 15-4513 binding site. The BZ antagonist flumazenil exhibited a 100–300-fold lower affinity for mutant receptors and α6β2γ2 receptors than for α1β2γ2 receptors which is in agreement with data previously published (13Wieland H.A. Lüddens H. Seeburg P.H. J. Biol. Chem. 1992; 267: 1426-1429Google Scholar). To investigate whether an exchange of the crucial amino acid histidine, responsible for high-affinity BZ agonist binding, would change the flunitrazepam affinity for α6β2γ2-receptors, we used two mutants, already described elsewhere (13Wieland H.A. Lüddens H. Seeburg P.H. J. Biol. Chem. 1992; 267: 1426-1429Google Scholar), for [3H]flunitrazepam binding studies. These mutant receptors, α1(Arg-101)β2γ2and α6(His-100)β2γ2, showed a single class of binding sites, with mean K d values of 19.7 and 5.4 nm, respectively (Fig. 1, A andB). This gives evidence that histidine 100 is not the exclusive determinant of flunitrazepam binding as postulated earlier (13Wieland H.A. Lüddens H. Seeburg P.H. J. Biol. Chem. 1992; 267: 1426-1429Google Scholar). The response to GABA characterized by the electrophysiological investigation of recombinant GABAA receptors is known to be enhanced consistently by flunitrazepam (23Malherbe P. Draguhn A. Multhaup G. Beyreuther K. Möhler H. Mol. Brain Res. 1990; 8: 199-208Google Scholar, 24Wafford K.A. Whiting P.J. Kemp J.A. Mol. Pharmacol. 1993; 43: 240-244Google Scholar), and this effect can be blocked completely by flumazenil (3Pritchett D.B. Lüddens H. Seeburg P.H. Science. 1989; 245: 1389-1392Google Scholar, 24Wafford K.A. Whiting P.J. Kemp J.A. Mol. Pharmacol. 1993; 43: 240-244Google Scholar). In the case of the rat α1β2γ2-receptor, 1 μm flunitrazepam has been shown to enhance 1 μm GABA-evoked currents by 64% (25Kleingoor C. Ewert M. von Blankenfeld G. Seeburg P.H. Kettenmann H. Neurosci. Lett. 1991; 130: 169-172Google Scholar). We observed that α1β2γ2-receptors reacted as expected to the coapplication of GABA and flunitrazepam. Flunitrazepam (10 μm) markedly affected the amplitude of the peak current at both GABA concentrations used and in a dose-dependent manner. At 1 μm GABA, we detected a 117 ± 46% increase in the amplitude, whereas coapplication with a higher concentration of GABA (10 μm) elicited a smaller potentiation of 46 ± 6% (Fig. 2 A). For α6β2γ2-receptors, however, 10 μm flunitrazepam and 10 μm GABA resulted in a significant reduction of control currents to 76 ± 12%, whereas coapplication of 1 μm GABA and 10 μm flunitrazepam did not produce a significant change in amplitude (Fig. 2, B and C). This effect of flunitrazepam was shown to be dependent on the concentration of GABA and became significant at a concentration of 6 μm GABA (Fig. 2 C). Our data thus indicate that flunitrazepam acts on α6β2γ2-receptors in an inverse agonistic manner. Our investigation of the effects of flumazenil on GABA-evoked chloride currents showed opposite effects for the two receptors studied. For cells expressing α1β2γ2-receptors, 10 μm flumazenil reduced the amplitude of chloride currents induced by application of 10 μm GABA to 89 ± 3% of control, whereas for cells expressing α6β2γ2-receptors, there was an increase to 118 ± 3% of control induced by application of 1 μm GABA and 10 μm flumazenil (Figs.3 (A and B) and 4A). We have used different GABA concentrations depending on the receptor subtype because we found in an earlier investigation (17Hauser C.A.E. Wetzel C.H.R. Rupprecht R. Holsboer F. Biochem. Biophys. Res. Commun. 1996; 219: 531-536Google Scholar) that α6β2γ2-receptors effectively responded to GABA at lower concentrations than α1β2γ2-receptors. The effect of flumazenil on α6β2γ2-receptor-mediated responses is low compared with previous reports on Sf-9 cells and HEK 293 cells (26Im W.B. Im H.K. Pregenzer J.F. Hamilton B.J. Carter D.B. Jacobsen E.J. TenBrink R.E. VonVoigtlander P.F. Br. J. Pharmacol. 1993; 110: 677-680Google Scholar). In an earlier investigation, an apparent agonistic activity of flumazenil was reported for a binary GABAAreceptor using the rat α1- and β2-subunits expressed in Xenopus oocytes (23Malherbe P. Draguhn A. Multhaup G. Beyreuther K. Möhler H. Mol. Brain Res. 1990; 8: 199-208Google Scholar). Here, 10 μmflumazenil potentiated the response induced by 2–10 μmGABA to 175% of control. Coapplication of flunitrazepam and flumazenil to recombinant α1β2γ2- and α6β2γ2-receptors (Fig.4 B) showed that in the case of the α6β2γ2-receptors, the inverse agonistic properties of flunitrazepam were reduced by the agonistic action of flumazenil. As expected, for the α1β2γ2-receptors, the agonistic properties of flunitrazepam were almost completely suppressed by flumazenil. The partial inverse agonist Ro 15-4513 had the opposite effect on GABA-induced currents in α1β2γ2- and α6β2γ2-receptors. In cells expressing α1β2γ2-receptors, 1 μm Ro 15-4513 decreased the amplitude of 10 μm GABA-induced chloride currents. For α6β2γ2-receptors with 1 μm GABA, there was a comparable change in current amplitude, but in the opposite direction (Figs. 3 (A andB) and 4A). In this study, we demonstrate that the two receptor subtypes α1β2γ2 and α6β2γ2 react with the BZ receptor ligands flunitrazepam, flumazenil, and Ro 15-4513 in a functionally opposite manner. The exact mechanism underlying the distinct BZ pharmacology of the α6β2γ2-receptor is not clear at present. Our findings, together with those of earlier investigations (13Wieland H.A. Lüddens H. Seeburg P.H. J. Biol. Chem. 1992; 267: 1426-1429Google Scholar, 26Im W.B. Im H.K. Pregenzer J.F. Hamilton B.J. Carter D.B. Jacobsen E.J. TenBrink R.E. VonVoigtlander P.F. Br. J. Pharmacol. 1993; 110: 677-680Google Scholar), lead us to conclude that distinct amino acids of the BZ-binding pocket determine the BZ agonistic or inverse agonistic activity of the GABAA receptor subtype. For the wild-type α6β2γ2-receptor, we postulate that after binding of flunitrazepam, the bulky amino acid arginine induces a structural change that modifies the channel in a way that decreases chloride ion flux. If the smaller amino acid histidine (wild-type α1β2γ2-receptor) is present, such a hindrance of influx may be absent. Studies involving site-directed mutagenesis should be conducted to elucidate the molecular characteristics of the flunitrazepam-binding site on the α6β2γ2-receptor. We are grateful to Peter Seeburg and Heike Wieland for providing the GABAA receptor cDNAs, Hartmut Lüddens for supplying several BZs, and Thorsten Trapp for critical comments on the manuscript. We are also grateful to Thomas Kuckuk and Bettina Burkart-Lauer for excellent assistance with the artwork.
Referência(s)