GABA is the principal inhibitory neurotransmitter in the mammalian brain. It mediates fast synaptic inhibition by interaction with the GABA_A receptor. GABA_A receptors are ligand-gated ion channels that are modulated by a large number of clinically relevant drugs such as benzodiazepines (BZs), barbiturates, neurosteroids, and anesthetics (Sieghart, 1995). They are assembled from individual subunits forming a pentameric structure. Nineteen isoforms of mammalian GABA_A receptor subunits have been cloned: α_1–6, β_1–3, γ_1–3, δ, ϵ, p, ρ_1–3, and θ (Barnard et al., 1998; Simon et al., 2004). The major receptor subtype of the GABA_A receptor in adults consists of α₁, β₂, and γ₂ subunits, and the most likely stoichiometry is two α subunits, two β subunits, and one γ subunit (Sieghart and Sperk, 2002).

The subunit composition determines the GABA sensitivity and the pharmacological properties of the GABA_A receptor (Sieghart, 1995; Hevers and Luddens, 1998; Boileau et al., 2002). The subunit composition of the receptor also affects the time course of the GABA response (desensitization and deactivation of the chloride currents) (Bianchi et al., 2001; Boileau et al., 2003; Feng et al., 2004). Mutation of amino acid residues in α and γ₂ subunits modulate the BZ sensitivity of the receptor, suggesting that the BZ binding pocket is located at the interface between α and γ₂ (Sigel, 2002; Ernst et al., 2003). There is clear evidence that substitution of the γ₂ subunit by either γ₁ or γ₃ significantly alters the sensitivity for BZ (Hevers and Luddens, 1998).

In contrast to the γ₂ subunit, which is ubiquitously expressed in the central nervous system, the γ₁ subunit is expressed in only a few areas of the brain such as the amygdala (central and medial nuclei), the pallidum, the septum, the substantia nigra, and the thalamus (centrolateral and medial nuclei) (Pirker et al., 2000; Korpi et al., 2002). Compounds selectively interacting with receptors containing γ₁ subunits thus might have a substantial clinical potential.

Compared with receptors containing γ₂ subunits, little is known about the pharmacological profile of GABA_A channels composed of α₁, β₂, and γ₁ subunits. Ymer et al. (1990) observed a loss in affinity for the benzodiazepine antagonist Ro 15-1788 and the inverse agonist methyl-6,7-dimethoxy-4-ethyl-β-carboline-3-carboxylate (DMCM) when the γ₁ was substituted for γ₂ in α₁β₁γ₂ receptors. Negative modulatory effects of Ro 15-4513, β-CCM, and DMCM for GABA_A receptors composed of α_1/2/3β₁γ₂ subunits are changed to positive modulatory effects in α_1/2/3β₁γ₁ receptors (Puia et al., 1991; Wafford et al., 1993). Benke et al. (1996) observed a low affinity for clonazepam, zolpidem, and flunitrazepam and apparent insensitivity for flumazenil and Ro 15-4513 for γ₁-containing receptors. Wafford et al. (1993) demonstrated a reduced enhancement of chloride currents through α₃β₂γ₃ by diazepam, clonazepam, and bretazenil compared with α₃β₂γ₂ and a negative modulatory effect of zolpidem for α₂β₁γ₁ and alpidem for α₃β₁γ₁ receptors.

Overall, in 4 different studies, a total of 14 compounds from 5 different compound classes have been investigated so far for their ability to modulate GABA_A receptors containing γ₁ subunits. Unfortunately, most of these studies were carried out under different experimental conditions and with receptors containing different α and β subunits combined with γ₁. Thus, the relative efficacies of these compounds for αβγ₁ are not comparable (Hevers and Luddens, 1998).

In the present study, we analyzed the modulation of α₁β₂γ₁ receptors expressed in Xenopus laevis oocytes by 21 compounds comprising distinct chemical structures. Triazolam, clotiazepam, midazolam, and CGS 20625 exhibited a significant potency and efficiency, whereas the other compounds were either inactive or displayed only a low potency on γ₁-containing receptors.

Materials and Methods

Chemicals. Compounds were obtained from the following sources: flunitrazepam (7-nitro-1,3-dihydro-1-methyl-5-o-fluorophenyl-2H-1,4-benzodiazepin-2-one), diazepam (7-chloro-1,3-dihydro-1-methyl-5-phenyl-2H-1,4-benzodiazepin-2-one), flurazepam (7-chloro-1,3-dihydro-1-ethylaminodiethyl-5-o-fluorophenyl-2H-1,4-benzodiazepin-2-one), midazolam [8-chloro-6-(2-fluorophenyl)-1-methyl-4H-imidazo[1,5-a][1,4]benzodiazepine], Ro 15-1788, and bretazenil [t-butyl(s)-8-bromo-11,12,13,13a-tetrahydro-9-oxo-9H-imidazo[1,5-a]pyrrolo[2,1-c] [1,4]benodiazepine-1-carboxylate] were from Hoffmann La Roche (Basel, Switzerland); l-655,708 was purchased from Tocris Cookson Inc. (Bristol, UK); clotiazepam [5-(2-chlorophenyl)-7-ethyl-1,3-dihydro-1-methyl-2H-thieno[2,3-e][1,4]diazepin-2-one] was from Troponwerke (Köln, Germany); clobazam [7-chloro-1-methyl-5-phenyl-1H-1,5-benzodiazepine-2,4(3H,5H)-dione] was from sanofi-aventis (Bridgewater, NJ); triazolam [8-chloro-6-(2-chlorophenyl)-1-methyl-4H-1,2,4-triazolo[4,3-a][1,4]benzodiazepine] was from Sigma (Vienna, Austria); DMCM was from Ferrosan (Soeborg, Denmark); CGS 9896 and CGS 20625 were from Novartis (Basel, Switzerland); zolpidem [N,N,6-trimethyl-2-(4-methylphenyl)imidazo[1,2-a]-pyridine-3-acetamide] was from Synthelabo Recherche (Bagneux, France); Cl 218,872 was from American Cyanamide Comp. (Wayne, NJ); Ru 31719, Ru 32698, Ru 33203, and Ru 33356 were from Roussel Uclaf (Romainville, France); PK 8165 and PK 9084 were from Pharmuka Laboratories (Gennevilliers, France). For chemical structures, see Ogris et al. (2004).

Expression and Functional Characterization of GABA_A Receptors.X. laevis oocytes were prepared and injected as described previously (Grabner et al., 1996). Female X. laevis (Nasco, Fort Atkinson, WI) were anesthetized by exposing them for 15 min to a 0.2% MS-222 (methane sulfonate salt of 3-aminobenzoic acid ethyl ester; Novartis) solution before surgically removing parts of the ovaries. Follicle membranes from isolated oocytes were enzymatically digested with 2 mg/ml collagenase (type 1A, Sigma). One day after isolation, the oocytes were injected with approximately 10 to 50 nl of a solution of diethyl pyrocarbonate water containing the different cRNAs at a concentration of approximately 300 to 3000 pg/nl/subunit. The amount of cRNA was determined by means of a Nano-Drop ND-1000 (Kisker-Biotech, Steinfurt, Germany). To ensure expression of the γ subunit in the case of α₁β₂γ₁ and α₁β₂γ_2S receptors, cRNAs were mixed in a ratio of 1:1:10, and for receptors comprising only α₁ and β₂ subunits, they were mixed in a ratio of 1:1 (Boileau et al., 2002).

Oocytes were stored at 18°C in ND96 solution (Methfessel et al., 1986). Electrophysiological experiments were performed by the two-electrode voltage-clamp method making use of a TURBO TEC 01C amplifier (NPI Electronic GmbH, Tamm, Germany) at a holding potential of –70 mV. The bath solution contained 90 mM NaCl, 1 mM KCl, 1 mM MgCl₂, 1 mM CaCl₂, and 5 mM HEPES, pH 7.4.

Perfusion System. GABA was applied by means of a modified version of a fast perfusion system according to Hering (1998). A schematic drawing of the perfusion chamber and drug application device is shown in Fig. 1A. As described previously, the voltage-clamp experiments on X. laevis oocytes were performed in a small (≈15 μl) bath that was covered by a glass plate. Two angular inlet channels in the glass cover (diameter, <1 mm) enabled access of the two microelectrodes to the oocyte. A funnel for drug application surrounded both access channels for the microelectrodes compared with a funnel surrounding a single access channel in Hering (1998). This modification increased the stability of oocyte perfusion.

Drug or control solutions were applied to the funnel by means of a Miniprep 60 (Tecan, Durham, NC) that was controlled by a DigiData 1322A (Clampex version 9.2; Molecular Devices, Sunnyvale, CA), permitting automation of the experiments [see also Supplemental Videos S1 and S2 (files methods_1.avi and methods_2.avi) for animation of the solution exchange].

To estimate the rate of solution exchange independent of the ligand-receptor interaction, we expressed K_v1.1 channels in X. laevis oocytes and analyzed the time course of current decay during a rapid increase of the extracellular potassium concentration from 1 to 10 mM (sodium was reduced to 80 mM, respectively). Figure 1B illustrates a typical K_v1.1 current during a voltage-clamp step from –80 to +20 mV. The current decrease upon fast perfusion with 10 mM potassium at a speed of 1 ml/s is shown on the right in Fig. 1B. A mean current decline time t_10–90% of 140.8 ± 17.5 ms (n = 7) was estimated.

To elicit GABA-induced chloride currents (I_GABA), the chamber was perfused with 120 μl of GABA-containing solution at the same volume rate (1 ml/s). The rise time of I_GABA ranged usually between 100 and 250 ms (Fig. 1C), which is comparable with the rate of solution exchange estimated in Fig. 1B.

After the initial fast perfusion step for rapid agonist application, the chamber was continuously perfused at a rate of 1 μl/s for a total of 18 s. Before rapid washout of agonist and/or drug, the funnel was emptied by a suction pulse applied to the two funnel outlets (Fig. 1A). The time between this suction pulse and the application of new solution to the funnel was <1 s to avoid evaporation of the bath surrounding the oocyte.

Duration of washout periods was extended from 3 to 30 min, with increasing concentrations of applied GABA to account for slow recovery from increasing levels of apparent desensitization. Oocytes with maximal current amplitudes >3 μA were discarded to minimize voltage-clamp errors.

Analyzing Concentration-Response Curves. Enhancement of chloride currents by modulators of the GABA_A receptor was measured at a GABA concentration eliciting between 5 and 10% of the maximal current amplitude (EC_5–10). The EC_5–10 (usually ranging between 3 and 8 μM) was determined at the beginning of each experiment.

Enhancement of the chloride current (I_GABA) was defined as (I_{(GABA + Comp)}/I_GABA) – 1, where I_{(GABA + Comp)} is the current response in the presence of a given compound, and I_GABA is the control GABA current. To measure the sensitivity of the GABA_A receptor for a given compound, it was applied for an equilibration period of 1 min before concomitant application of GABA (EC_5–10) and increasing concentrations of the compound. None of the compounds investigated was able to induce chloride flux in the absence of GABA. Concentration-response curves were generated, and the data were fitted by nonlinear regression analysis using ORIGIN software (OriginLab Corp, Northampton, MA). Data were fitted to the equation 1/(1 + (EC₅₀/[Comp])^nH), where EC₅₀ is the concentration of the compound that increases the amplitude of the GABA-evoked current by 50%, and n_H is the Hill coefficient. Data are given as mean ± S.E. from at least four oocytes and ≥2 oocyte batches. Statistical significance was calculated using unpaired Student's t test with a confidence interval of p < 0.05.

Results

Modulation of α₁β₂γ₁ Receptors. The aim of the present study was to investigate the pharmacological properties of compounds interacting with α₁β₂γ₁ GABA_A receptors. We have therefore analyzed the modulation of this GABA_A receptor subtype by 21 compounds from 12 different structural classes comprising 1,4-benzodiazepines (flunitrazepam, diazepam, flurazepam, midazolam, and triazolam), 1,4-thienodiazepines (clotiazepam), 1,5-benzodiazepines (clobazam), imidazobenzodiazepines (flumazenil, bretazenil, and l-655,708), β-carbolines (DMCM), pyrazoloquinolines (CGS 9896), pyrazolopyridines (CGS 20625), imidazopyridines (zolpidem), triazolopyridazines (Cl 218,872), imidazoquinolines (Ru 31719), imidazopyrimidines (Ru 32698, Ru 33203, and Ru 33356), and quinolines (PK 8165 and PK 9084).

In a first step, GABA_A receptors were activated by GABA concentrations corresponding to EC_5–10, and drug effects were screened at a single concentration of 1 μM. As illustrated in Fig. 2 (□), only the benzodiazepines triazolam, midazolam, and diazepam, the thienodiazepine clotiazepam, the pyrazolopyridine CGS 20625, and the pyrazoloquinoline CGS 9896 induced an enhancement of >20% at this concentration. The other compounds induced either a very small but statistically significant enhancement (zolpidem and bretazenil), no statistically significant effect (DMCM, Cl 218,872, clobazam, flumazenil, Ru 33203, PK 9084, flurazepam, l-655,708, Ru 31719, Ru 33356, and Ru 32698), or even an inhibition of GABA-induced chloride currents (flunitrazepam and PK 8165), and details are given in Table 1.

To establish compounds displaying low potency but high efficiency, all compounds were subsequently tested at higher concentrations (10 and 100 μM) (Fig. 2A). Modulation of GABA receptors at low GABA concentrations (EC_5–10 close to “tonic concentrations”) can substantially differ from modulation at high concentrations (i.e., millimolar “synaptic concentrations”). We have, therefore, analyzed the effects of all 21 compounds at 1 mM GABA (Fig. 3A). None of the compounds, however, substantially enhanced or inhibited I_GABA. Representative chloride currents induced by 1 mM in the absence or presence of 1 μM triazolam or CGS 20625 are shown in Fig. 3B.

Contribution of α₁β₂ Receptors. Previous studies have clearly shown that the extent of the incorporation of γ subunits into heterologously expressed GABA_A receptors may vary between oocyte batches and decrease with time (Boileau et al., 2002). To clarify whether the effects observed were caused by α₁β₂γ₁ or by α₁β₂-comprising receptors, we analyzed the effect of these compounds on GABA_A channels composed of α₁β₂ subunits. With the exception of CGS 20625 (+20 ± 5%), CGS 9896 (+16 ± 7%), flumazenil (+13 ± 5%), and flurazepam (+26 ± 4%), all compounds (1 μM) were either inefficient in enhancing I_GABA or even induced significant inhibition (triazolam, clotiazepam, midazolam, Cl 218,872, and PK 8165; Table 1).

For the most potent stimulators of the α₁β₂γ₁ receptors triazolam, clotiazepam, and midazolam, we also analyzed the inhibition of I_GABA in oocytes expressing only α₁β₂ subunits at higher (10 μM) concentrations. Triazolam inhibited the GABA-induced chloride flux in α₁β₂ receptors by –33 ± 4% (n = 12), clotiazepam by –30 ± 8% (n = 7), and midazolam by –31 ± 9% (n = 5) (experiments not shown).

Comparing the Effects of Benzodiazepine Site Ligands on α₁β₂γ₁ and α₁β₂γ_2s Receptors. Triazolam, clotiazepam, midazolam, and CGS 20625 were subsequently analyzed in more detail by comparing their effects on α₁β₂γ₁ and α₁β₂γ_2S receptors. Figure 4 illustrates the concentration-dependence of the enhancement of the currents (EC_5–10) by triazolam, clotiazepam, and midazolam. The EC₅₀ value was determined by fitting the concentration-effect data to the Hill equation. Triazolam enhanced the maximum chloride current of α₁β₂γ₁ receptors by 85% while displaying the highest potency (EC₅₀ ≈ 90 nM) of all tested benzodiazepines. Clotiazepam elicited an enhancement of the GABA response of approximately 170% but had a 19-fold lower potency (EC₅₀ ≈ 1.7 μM) than triazolam. Midazolam was more potent than clotiazepam (EC₅₀ ≈ 1.2 μM) but was 13 times less potent than triazolam, with a maximum enhancement (92%) comparable with triazolam.

A comparison with the concentration-response data obtained on GABA_A channels containing γ_2S subunits reveals a 3-fold higher efficiency of triazolam, a 1.5-fold higher efficiency of clotiazepam, and a 3.7-fold higher efficiency of midazolam on α₁β₂γ_2S receptors. The ratio of the EC₅₀ values for triazolam, clotiazepam, and midazolam (EC_50/γ1/EC_50/γ2S) reflect 4-, 9-, and 8-fold lower potencies of these compounds for α₁β₂γ₁ receptors, respectively (Table 2). The apparent EC₅₀ values, maximum enhancement, and the corresponding ratios for the benzodiazepines tested are given in Table 2.

Modulation of I_GABA by Triazolam, Clotiazepam, and Midazolam at Different GABA Concentrations. To gain insight into the mechanism of I_GABA enhancement, we studied the GABA dose-effect curves in the absence and presence of the modulators. The results are shown in Fig. 5, A to C. In control, the mean EC₅₀ value for GABA was 39 ± 3 μM in α₁β₂γ₁ receptors and 50 ± 3 μM in α₁β₂γ_2S receptors. The three benzodiazepine receptor ligands shifted the dose-effect curves to the left without affecting the maximal response (Fig. 3A). It is noteworthy that the drug-induced shift was more pronounced for α₁β₂γ_2S than for α₁β₂γ₁ receptors, reflecting the higher efficiency of these ligands on α₁β₂γ_2S.

Effect of the Pyrazolopyridine CGS 20625 on α₁β₂, α₁β₂γ₁ and α₁β₂γ_2S Receptors. CGS 20625 elicited maximum enhancement of chloride currents through α₁β₂γ₁ receptors of approximately 645%, with a half-maximal enhancement occurring at approximately 20 μM (Fig. 6B). CGS 20625 thus represents a low potency but highly efficient positive modulator of α₁β₂γ₁ receptors. This compound enhanced the GABA response of α₁β₂γ_2S receptors with comparable efficiency (I_max/γ2/I_max/γ1≈ 1.12) and was less efficient on α₁β₂ receptors (Fig. 6B). At higher CGS 20625 concentrations (≥300 μM), we observed weaker enhancement of the GABA-induced chloride flux than at 100 μM for all subunit compositions (Fig. 6B).

Effect of Flumazenil on α₁β₂γ₁ Receptors. Flumazenil is a ligand of the BZ binding site of α₁β₂γ₂ GABA_A receptors and competitively inhibits the enhancement of GABA-induced chloride currents by benzodiazepine agonists (Wafford et al., 1993). It was therefore interesting whether this compound would also inhibit the effects of triazolam and clotiazepam on α₁β₂γ₁ receptors.

Figure 7, A and B (left), illustrates the inhibition by flumazenil of triazolam- or clotiazepam-induced I_GABA enhancement in α₁β₂γ_2S receptors. As shown on the right, the effects of triazolam or clotiazepam on α₁β₂γ₁ receptors were not inhibited by 1 μM flumazenil. Moreover, we were unable to study possible antagonistic effects at higher concentrations, because flumazenil induced significant enhancement of I_GABA in oocytes expressing α₁β₂γ₁ subunits at 10 (22 ± 2%, n = 4) and 100 μM (64 ± 8%, n = 4; Fig. 7C).

Discussion

In the present study, we made use of a fast and automated perfusion technique (Fig. 1) to test a selection of 21 modulators on GABA_A receptors composed of α₁β₂γ₁ subunits. Solution exchange occurred between 100 and 250 ms (see Materials and Methods and Fig. 1C), which reduced the effects of desensitization on peak current detection compared with conventional bath perfusion.

I_GABA of α₁β₂γ₁ Subunit Receptors Is Enhanced by Some Benzodiazepines and the Pyrazolopyridine CGS 20625. The 21 compounds tested comprised benzodiazepines and representatives of other structural classes of ligands of the BZ binding site of GABA_A receptors. To determine the most potent modulators of α₁β₂γ₁ subunit receptors, we first tested the compounds at a concentration of 1 μM. Of the 21 compounds, six induced an enhancement of the GABA response (EC_5–10) by more than 20% with the following order of potency: triazolam > clotiazepam > midazolam > CGS 20625 > CGS 9896 > diazepam. Zolpidem and bretazenil induced a small enhancement (≈13 and 10%, respectively). The effects of DMCM, Cl 218,872, clobazam, flumazenil, Ru 33203, PK 9084, flurazepam, l-655,708, Ru 31719, Ru 33356, and Ru 32698 were not significantly different from control. In contrast, flunitrazepam and PK 8165 significantly inhibited the GABA-induced chloride flux. Application of higher concentrations (10 and 100 μM) revealed a moderate enhancement at high concentrations (low-potency modulation) by bretazenil, DMCM, flumazenil, Ru 33203, Ru 32698, and flunitrazepam. All compounds were tested in the absence of GABA (data not shown). None of them induced measurable currents, even at high concentrations (up to 100 μM). Thus, the 1,4-benzodiazepines triazolam and midazolam, the 1,4-thienodiazepine clotiazepam, and the pyrazolopyridine CGS 20625 seemed to be the most promising candidates for further detailed analysis.

Contribution of the γ₁ Subunit to the Enhancement of I_GABA. On injection of X. laevis oocytes with α, β, and γ subunits, not only are receptors containing all three subunits formed but possibly also receptors composed of α and β subunits only (Boileau et al., 2002). To investigate whether the observed drug effects were caused by effects on α₁β₂γ₁ receptors or could also be explained by effects on α₁β₂ receptors, the effects of drugs on the latter receptors were also investigated. A comparison of drug effects on α₁β₂γ₁ and α₁β₂ receptors revealed that the γ₁ subunit was essential for enhancement of I_GABA by triazolam, clotiazepam, and midazolam (Table 1 and Fig. 2), because these compounds at 1 μM concentration induced a significant inhibition by –21 ± 4% (n = 12, p < 0.05), (–14 ± 2%, n = 6), and (–20 ± 5%, n = 6), respectively, of I_GABA through α₁β₂ subunit receptors. In addition, the effects of several other drugs were clearly different in α₁β₂γ₁ and α₁β₂ receptors. Thus, CGS 20625, CGS 9896, diazepam, zolpidem, and bretazenil exhibited an enhancement of I_GABA that was higher in α₁β₂γ₁ than in α₁β₂ receptors, supporting the importance of the γ₁ subunit for the effects observed. Finally, Cl 218,872 and flurazepam exhibited effects in opposite direction in α₁β₂γ₁ and α₁β₂ receptors, again supporting the conclusion that the drug effects observed on injection of oocytes with α₁, β₂, and γ₁ subunits were predominantly caused by α₁β₂γ₁ receptors (Table 1).

Our finding that triazolam, clotiazepam, and midazolam significantly inhibit I_GABA through α₁β₂ receptors suggests that we might have underestimated the enhancement of I_GABA in α₁β₂γ₁ receptors by these drugs because of simultaneously occurring inhibition of α₁β₂ subunit receptors. The differential effects of various BZ binding site ligands on receptors composed of α₁β₂ subunits are highly interesting by themselves and significantly extend previous evidence for the existence of a low-affinity benzodiazepine binding site at αβ receptors (Wafford et al., 1993; Thomet et al., 1999; Walters et al., 2000).

To determine the potency and efficiency (maximum ability to enhance the GABA EC_5–10 response of the three most potent BZ; Table 2), we studied their concentration effect for receptors composed of α₁β₂γ₁ subunits (Fig. 3). The 1,4-benzodiazepine triazolam displayed the highest potency (EC₅₀ ≈ 90 nM), followed by midazolam (EC₅₀ ≈ 1.2 μM) and the 1,4-thienodiazepine clotiazepam (EC₅₀ ≈ 1.7 μM, Table 2). A comparison of the concentration-effect curves for these benzodiazepine-type ligands with receptors composed of α₁β₂γ_2S receptors revealed a significantly lower efficiency and potency for γ₁-containing receptors (Fig. 3 and Table 2).

The Pyrazolopyridine CGS 20625 Displays the Highest Efficiency in Enhancing GABA-Induced Chloride Flux of α₁β₂γ₁ Subunit Receptors. CGS 20625 was identified as the most efficient compound in terms of maximum enhancement of the GABA-induced chloride currents through α₁β₂γ₁ subunit receptors (Fig. 6). This compound induced a maximum enhancement of 645 ± 55% greater than control, which is approximately 3.75-fold the enhancement achieved with clotiazepam (172 ± 24% greater than control), and more than 7 times the enhancement induced by midazolam (92 ± 8% greater than control) or by triazolam (85 ± 7% above control), respectively (Fig. 3). CGS 20625, however, had a potency for α₁β₂γ₁ receptors (EC₅₀ ≈ 20 μM) approximately 200 times lower than that of triazolam and 10 to 20 times lower than that of the other benzodiazepines (Table 2).

A closer inspection of the subunit composition specificity of CGS 20625 action revealed that this drug potentiates α₁β₂γ₁ and α₁β₂γ_2S subunit receptors to an almost similar extent (Fig. 6B). However, a significantly lower efficiency on α₁β₂ receptors revealed an essential role of a γ subunit.

CGS 20625 at 300 μM caused less enhancement than at 100 μM in α₁β₂, α₁β₂γ₁, and α₁β₂γ_2S subunit receptors (Fig. 6B), suggesting that this compound might inhibit chloride currents at high concentrations. Similar behavior was shown previously for the action of another pyrazolopyridine (tracazolate) on GABA_A channels (Thompson et al., 2002).

CGS 20625 thus represents a low-potency (EC₅₀ ≈ 20 μM) but high-efficiency modulator of α₁β₂γ₁ and α₁β₂γ_2S subunit-containing receptors (Fig. 6B). This compound was almost not selective for either γ₁ or γ₂ subunits. α₁β₂ Subunit receptors were, however, stimulated to a significantly lesser extent (≈160%) compared with α₁β₂γ_1/2S subunit receptors (640–730%) (Fig. 6B).

The competitive antagonist flumazenil (1 μM) inhibited I_GABA enhancement of α₁β₂γ_2S receptors but failed to affect the enhancement of I_GABA through α₁β₂γ₁ receptors by triazolam and clotiazepam (Fig. 7, right). These data suggest that flumazenil exhibits either no or a very low affinity for the BZ binding site of α₁β₂γ₁ receptors, or that flumazenil interacts with a binding site different from that for triazolam and clotiazepam at these receptors. The first conclusion is consistent with the observation that the affinity of flumazenil for its binding site was reduced approximately 1000-fold in GABA_A receptors in which phenylalanine 77 of the γ₂ subunit was mutated to the corresponding residue (isoleucine) of the γ₁ subunit (γ₂F77I; Buhr et al., 1997; Wingrove et al., 2002; Ogris et al., 2004). At higher concentrations, flumazenil displayed properties of a low-affinity agonist on α₁β₂γ₁-receptors (10 μM potentiated I_GABA by 22 ± 2% and 100 μM by 64 ± 9%).

In addition to flumazenil, the affinities of bretazenil, l-655,708, DMCM, zolpidem, Cl 218,872, and PK 8165 were drastically reduced in receptors containing the γ₂F77I point mutation (Ogris et al., 2004), as measured by [³H]flunitrazepam binding studies. A low affinity of these compounds for α₁β₂γ₁ receptors could have contributed to their small effects on these receptors observed in the present study. In contrast, replacement of phenylalanine (γ₂F77) by the corresponding isoleucine of the γ₁ subunit only weakly (2- to 7-fold) reduced the affinity of the classic 1,4-benzodiazepines, the 1,4-thienodiazepine clotiazepam, the 1,5-benzodiazepine clobazam, or the pyrazoloquinoline CGS 9896 (Ogris et al., 2004). The small (4- to 9-fold) reduction in potency of triazolam, midazolam, and clotiazepam for enhancing α₁β₂γ₁ compared with α₁β₂γ₂ receptors could be explained by a reduced apparent affinity of these compounds for α₁β₂γ₁ receptors underlining the importance of Phe77 for high-affinity BZ binding (Buhr et al., 1997). For other compounds such as flunitrazepam (1 μM), we observed significant inhibition of I_GABA in α₁β₂γ₁ receptors, indicating the importance of additional amino acids for drug binding and gating.

The different efficiency of triazolam, midazolam, clotiazepam, and CGS 20625 are explained by the different amounts of shifts of the GABA concentration-effect curves (Fig. 5, A–C). Larger shifts induced on α₁β₂γ_2S receptors reflect the higher apparent efficiency.

In summary we systematically investigated 21 ligands of the BZ binding site from chemically distinct classes to obtain insight in the pharmacological profile of GABA_A receptors comprising a γ₁ subunit. Triazolam was identified as a high-potency and CGS 20625 as a high-efficiency modulator of this receptor subtype. Different potencies of triazolam, midazolam, clotiazepam, and CGS 20625 can be explained by different shifts of the GABA dose-effect curve, reflecting different apparent affinities of these compounds.