α Subunit Isoform Influences GABA_A Receptor Modulation by Propofol

Summary

We have investigated the role of the α subunit in the modulation of γ-aminobutyric acid type A (GABA_A) receptors by the general anesthetic propofol, using whole-cell patch clamp recordings made from distinct stable fibroblast cell lines which expressed only α₁β₃γ₂ or α₆β₃γ₂ GABA_A receptors. At clinically relevant anesthetic concentrations, propofol potentiated submaximal GABA currents in α₁β₃γ₂ receptors to a far greater degree than those in α₆β₃γ₂ receptors. The α subunit influenced the efficacy of propofol for modulation, but not its potency. In contrast, direct gating of the ion channel by propofol, in the absence of GABA, was significantly larger in the α₆ than the α₁ containing receptors. The potentiation of submaximal GABA by trichloroethanol, and the potentiation and direct gating by methohexital, was also studied, and showed the same relative trends as propofol.

Introduction

GABA_A (γ-aminobutyric acid type A) receptors are the major receptors for inhibitory neurotransmission in the mammalian brain. The GABA_A receptor is a pentameric complex formed by different glycoprotein subunits (α_1–6,β_1–4, γ_1–4,δ) which combine to form a chloride channel (reviewed by Burt and Kamatchi, 1991; Olsen and Tobin, 1990; Rabow et al., 1995). GABA_A receptor subunit expression in the central nervous system is heterogeneous (Laurie et al., 1992a,b; Wisden et al., 1991). For example, the GABA_A α₆ subunit isoform is localized solely in cerebellar granule cells, whereas α₁ is widely expressed throughout the brain (Luddens et al., 1990; Kato, 1990; Mertens et al., 1993; Wisden et al., 1992).

The GABA_A receptor complex is modulated allosterically by a wide range of compounds which act at discrete but unknown sites on the receptor (Macdonald and Olsen, 1994). One major group of GABA_A receptor modulators is the class of general anesthetics, many of which have been demonstrated to augment GABA_A receptor chloride currents at clinically relevant concentrations (e.g. Zimmerman et al., 1994), and are thought to elicit anesthesia by enhancing inhibitory synaptic transmission (Nicoll et al., 1975; Gage and Robertson, 1985). General anesthetics known to potentiate the actions of GABA (γ-aminobutyric acid) include propofol [2,6-diisopropylphenol (PRO); Hales and Lambert (1991); Hara et al. (1994)], steroid anesthetics (Harrison and Simmonds, 1984; Harrison et al., 1987), barbiturates (Study and Barker, 1981), chlormethiazole (Hales and Lambert, 1992), halogenated volatile anesthetics (Wakamori et al., 1991; Jones et al., 1992), and trichloroethanol (TCEt, the active metabolite of chloral hydrate; Lovinger et al., 1993; Peoples and Weight, 1994).

Different combinations of GABA_A receptor subunits show variable sensitivity to allosteric modulators (e.g. Horne et al., 1993). The best documented role of specific subunits is for benzodiazepine modulation, where the presence of a γ subunit is required (Pritchett et al., 1989), but benzodiazepine agonist sensitivity is also influenced critically by the type of α subunit isoform. Specifically, the exchange of a single amino acid confers benzodiazepine sensitivity on the normally benzodiazepine-insensitive α₆-containing GABA_A receptor (Kleingoor et al., 1993).

The role of GABA_A receptor subunits in general anesthetic modulation is less clear. For example, modulation of GABA_A receptors by PRO is qualitatively independent of the γ subunit (Jones et al., 1995) and can be observed in heteromeric αβ or even in homomeric β₁ receptors (Sanna et al., 1995a,b). It was recently shown that the α subunit influences modulation by pentobarbital (Thompson et al., 1996). This study was therefore designed to compare the effects of PRO on GABA-induced chloride currents in receptors containing two different α subunits against a common αβ background. The α₆ subunit was chosen as it shares the least homology [along with α₄, see Wafford et al. (1996)] with the other α subunit isoforms (Tyndale et al., 1995), and contributes to the differential pharmacology of benzodiazepines (Hadingham et al., 1993). We hypothesized that if differences in PRO modulation were to exist among α subunit isoforms, they might be most apparent in comparisons with α₆-containing receptors. Portions of this work were previously presented in abstract form (O’Shea et al., 1996).

MATERIALS AND METHODS

Tissue culture

An Ltk⁻ mouse fibroblast cell line from ATCC (Rockville, MD, U.S.A.) was stably transfected with dexamethasone-inducible GABA_A receptor α₁β₃γ₂ or α₆β₃γ₂ cDNA, as previously described in detail (Hadingham et al., 1992). The resultant cell lines were cultured in supplemented Eagle’s minimum essential medium (Sigma, St. Louis, MO, U.S.A.) as previously described.

Electrophysiology

Electrophysiological recordings were performed at room temperature using the whole-cell patch clamp technique. The electrode solution contained (in mM): 147, N-methyl-D-glucamine hydrochloride; 5, CsCl; 5, K₂ATP; 5, 4-(2-hydroxyethyl)-l-piperazine ethane sulfonic acid (HEPES); 1, MgCl₂; 0.1, CaCl₂; and 1.1, ethylene glycol bis(b-aminoethyl ether)-N,N,N′,N′-tetraacetic acid (EGTA), pH 7.2, osmolarity 315 mosmol. Pipette-to-bath resistance was 4–7 MΩ. During an experiment, cells were constantly perfused with extracellular solution containing (in mM) 145 NaCl, 3 KCl, 1.5 CaC1₂, 1 MgCl₂, 5.5 D-glucose and 10 HEPES, pH 7.4, osmolarity 320–330 mosmol. Cells were voltage-clamped at −60 mV. Since the intracellular and extracellular solutions contained symmetrical chloride concentrations, the chloride reversal potential was ca 0 mV.

Electrophysiology-picospritzer

GABA was applied to the cell under study by brief pressure ejection (2–6 p.s.i., 10–100 msec) from low-resistance micropipettes filled with 20 μM GABA for cells containing α₁β₃γ₂ receptors and 2 μM GABA for cells containing α₆β₃γ₂ receptors. This produced transient inward currents which were standardized by varying the duration of the pressure pipette pulse. For both cell lines, a maximal response was elicited by a 1 sec pulse of GABA from the pressure pipette, which was then decreased to 25–50 msec to achieve transient chloride currents that were ca 20% of the maximum current obtainable by 20 or 2 μM (test response; 19.3 ± 0.89% and 19.2 ± 0.94% for the α₁β₃γ₂ and α₆β₃γ₂ cells, respectively). GABA was applied every 20 sec. Anesthetic agents were applied to the bath until a stable, maximal potentiation of the GABA response was achieved. Drugs were then washed out until the pre-drug current was regained. Using this method, anesthetic equilibrium and recovery typically took 5–10 min. Current responses were low-pass filtered at 5 kHz (−3 dB, Bessel filter 902; Frequency Devices, Inc, MA, U.S.A.), digitized (TLl-125 interface; Axon Instruments, Foster City, CA, U.S.A.), and stored for off-line analysis (AXOBASIC, Axon Instruments).

Electrophysiology-rapid solution changer

GABA and/or anesthetics were rapidly applied to the cell by local perfusion [as described in Koltchine et al. (1996)] using a motor-driven solution exchange device (Bio Logic Rapid Solution Changer RSC-100; Molecular Kinetics, Pullman, WA, U.S.A.). An approximate EC₁₀ GABA test dose was used as the control value (α₁β₃γ₂/PRO: 9.0 ± 2.2% of maximal current; α₁β₃γ₂/methohexital (MTX): 8.4 ± 0.7%; α₆β₃γ₂/PRO: 9.2 ± 0.2%; α₆β₃γ₂/MTX: 10.8 ± 1.4%) for the potentiation experiments. The solution changer was controlled by protocols in the acquisition programs AXOBASIC or pCLAMP5 (Axon Instruments). Laminar flow was maintained by applying all solutions at identical flow rates via a multi-channel infusion pump (Smelting, Wood Dale, IL, U.S.A.). Prior to recording, a blue dye (FD and C Blue No. 1 and FD and C Red No. 40, McCormick, Baltimore, MD, U.S.A.) was used to check the alignment of the solution streams from the rapid solution changer.

Data analysis-electrophysiology

Drug-induced potentiation of a GABA-induced current was defined as the percentage increase of the control GABA response (defined as the average of the pre-drug and post-drug GABA induced currents). Concentration-response data were fitted (KaleidaGraph; Reading, PA, U.S.A.) with the logistic equation: I/I_max = 100*[drug]^N/([drug]^N+(EC₅₀)^N), where I/I_max is the percentage of the maximum obtainable GABA response, EC₅₀ is the concentration producing a half-maximal response, and N is the Hill coefficient. Pooled data are presented as mean ± SEM. Statistical significance was determined by Student’s two-tailed, unpaired t-test.

[³H]Muscimol binding

Binding was performed on homogenized membranes prepared from cells cultured as above. Cells were harvested in ice-cold binding buffer containing (in mM): 20, Tris; 2, ethylenediamine tetraacetic acid (EDTA); 150, KCl, pH 7.4, then centrifuged at 2700g for 10 min at 4°C. The cell pellet was hypotonically lysed in deionized water and centrifuged at 48 000g for 20 min at 4°C. Finally, the pellet was rinsed with binding buffer and stored at −80°C for up to 2 months. Prior to experiments, membranes were thawed on ice and resuspended in assay buffer containing (in mM): 124, NaCl; 1.3, MgS0₄; 25, HEPES; 2.9, KCl; 1.2, KH₂P0₄; D-glucose, 5.2, pH 7.4. The pellet was then homogenized with a Brinkman polytron (final protein concentration 0.3–0.5 mg/ml by Bradford assay, BSA standard). Membrane homogenates, along with eight concentrations of [³H]muscimol (saturation curve) or GABA plus a fixed concentration of [³H]muscimol (inhibition curve), were incubated in triplicate. After equilibration at room temperature, the assay mixtures were vacuum filtered through GF/B filter paper (Whatman, Clifton, NJ, U.S.A.) using a cell harvester (Brandel model MB-48, Gaithersburg, MD, U.S.A.), and quickly washed 10 times with 0.25 ml assay buffer. Specific binding was determined by subtracting the radioactive counts in the presence of 10 mM GABA from the radioactive counts in buffer alone. Similar results were obtained with a centrifugation assay using binding buffer (data not shown).

Data analysis-binding

Data points are taken from specific binding in individual experiments, and represent the mean of triplicate assays±SEM. Saturation binding data were fit (KaleidaGraph; Reading, PA, U.S.A.) with the equation: B = (B_max * [L])/([[L] + K_d), where B is the specific binding, [L] is [muscimol], and K_d is the dissociation constant for muscimol. Inhibition data were fit with the equation:

$$B=B_{\mathit{max}}-[\text{drug}]/({\text{IC}}_{50}+[\text{drug}])\ast ({\text{B}}_{max}-{\text{B}}_{min}),$$

where B_min is the lowest specific binding, [drug] is the final concentration of GABA, IC₅₀ is the concentration producing a half-maximal displacement of [³H]muscimol. The K_i value for GABA was calculated from: K_i = IC₅₀/(1 + ([L]/K_d).

Drugs used

Stock solutions of GABA, bicuculline methiodide, TCEt (Sigma), PRO (Aldrich, Milwaukee, WI, U.S.A), MTX sodium (Brevital®, Eli Lilly, Indianapolis, IN, U.S.A.), and midazolam hydrochloride (Versed®, intravenous/intramuscular solution preparation; Roche Pharmaceuticals, Manati, PR, U.S.A.) were diluted into extracellular solution daily before use. PRO was first dissolved into dimethyl sulfoxide (DMSO; Sigma) to form a stock solution of 10 mM PRO which was then dissolved into the extracellular solution to form the final PRO solutions (maximum final concentration of DMSO was 0.05% for a 50 μM PRO solution). Carrier controls were performed with 0.05% DMSO in extracellular medium. No significant effects of this DMSO solution were observed on GABA-induced currents in cells expressing either α₁β₃γ₂ or α₆β₃γ₂ receptors. [³H]muscimol (sp. act. 14.9 and 19.1 Ci/mmol) was obtained from NEN (Wilmington, DE, U.S.A.).

RESULTS

Receptor characteristics

We initially characterized the electrophysiological responses to GABA of the α₁β₃γ₂ and α₆β₃γ₂ combinations. Application of GABA to either cell line by the rapid solution changer elicited concentration-dependent inward currents [Fig. 1(A, B)]. Analysis of the concentration-response data [Fig. 1(C) and Table 1] shows that the α₆β₃γ₂ receptor has a higher apparent affinity for GABA (EC₅₀ 0.8 ± 0.1 μM) and a lower Hill coefficient (N= 0.9 ± 0.2) than the α₁β₃γ₂ line (EC₅₀ 2.2 ± 0.2 μM, N = 1.9 ± 0.2, respectively; p < 0.05 for both comparisons). These results are in accord with other investigations that compared GABA responses in α₁ and α₆-containing GABA_A receptor combinations (Ducic et al., 1995; Thompson et al., 1996). Also, the α₁β₃γ₂ line produced larger maximal GABA currents than the α₆β₃γ₂ line (2690 ± 400 pA vs 1186 ± 448 PA, p < 0.05). Midazolam (1 μM) potentiated GABA currents in cells containing the α₁ subunit, but not in cells containing the α₆ subunit, and GABA currents in both receptor subtypes were blocked by 20 μM bicuculline (data not shown).

Fig. 1.

GABA has higher apparent affinity for α₆β₃γ₂ than α₁β₃γ₂. GABA was applied with the rapid solution changer. (A and B) GABA responses from a cell expressing GABA_A α₁β₃γ₂ (A) or α₆β₃γ₂ (B) receptors. Bars over current traces indicate the duration of rapid GABA application, with the concentration of applied GABA in μM. (C) Concentration-response curves for GABA_A α₁β₃γ₂ or α₁β₃γ₂ receptors. Data points are shown as the normalized means of multiple experiments (5 ≤ n ≤ 7), error bars indicate SEM.

Table 1.

A summary of electrophysiological data using the rapid solution changer (values shown are means ± SEM with number of experiments in parentheses)

Cell line	EC50 (μM)	Hill coefficient	Maximum (pA)
α1β3γ2	2.2 ± 0.2 (6)	1.9 ± 0.2 (6)	2690 ± 400 (18)
α6β3γ2	0.8 ± 0.1 (7)	0.9 ± 0.2 (7)	1186 ± 448 (8)

As shown in Fig. 2(A) and Table 2, the α₆β₃γ₂ receptor shows a higher affinity for [³H]muscimol than α₁β₃γ₂ (α₆β₃γ₂ K_d 28 ± 5 nM versus α₁β₃γ₂ K_d 90 ± 23, p <0.05). A Scatchard plot of the data (Fig. 2(B)) illustrates that binding at both receptors was appropriately fit by a one-site model. In displacement assays (Fig. 2(C)), GABA binding affinity to α₆ was also higher than at α₁-containing receptors, as shown by its lower IC₅₀ (Table 2: α₆β₃γ₂: 45 ± 14 nM vs α₁β₃γ₂: 370 ± 78, p <0.05), and lower K_i value (α₆β₃γ₂: 34 ± 10 versus α₁β₃γ₂: 337 ± 71, p < 0.05).

Fig. 2.

Binding affinity is higher for muscimol and GABA at α₆β₃γ₂ than α₁β₃γ₂. (A) Saturation binding isotherms for α₁β₃γ₂ and α₆β₃γ₂ receptors obtained by using increasing concentrations of [³H]muscimol. Data points show the means ± SEM from a single experiment, with each concentration performed in triplicate. This experiment was performed five times, with similar results. (B) Shows a Scatchard plot from the same experiment from (A). Data points are the means of triplicate measurements. (C) An inhibition curve showing displacement of 10 nM [³H]muscimol with increasing concentrations of GABA. This figure represents the means ± SEM from a single experiment, with each point performed in triplicate. This experiment was performed three times with similar results.

Table 2.

A summary of radioligand binding data using [³H]muscimol (values represent the pooled means±SEM from curve fits of individual experiments with number of experiments in parentheses)

Cell line	Muscimol Kd (nM)	GABA IC50 (nM)	GABA Ki (nM)
α1β3γ2	90 ± 23 (5)	370 ± 78 (3)	337 ± 71 (3)
α6β3γ2	28 ± 5 (5)	45 ± 14 (3)	10 ± 10 (3)

Propofol potentiation of GABA responses

GABA potentiation by PRO was measured electrophysiologically using two separate methods: the picospritzer and rapid solution changer. The picospritzer was used from the initial set of experiments (Fig. 3, Table 3), since this method allowed for a wide range of modulator concentrations to be tested rapidly. However, once the appropriate PRO concentration range was established, later experiments with the rapid solution changer (Fig. 4) provided the means to apply known concentrations of GABA.

Fig. 3.

α subunit affects the efficacy, but not potency of PRO. Submaximal GABA was applied by picospritzer as detailed in Materials and Methods; anesthetics were applied to the bath. (A) PRO (0.05–50 μM) potentiates GABA-induced currents in both α₁β₃γ₂ and α₆β₃γ₂ GABA_A receptors. Data points are means ± SEM from pooled experiments (4 ≤ n ≤ 13). (B) TCEt (0.02–10 mM) potentiates GABA-induced currents in both receptor combinations (5 ≤ n ≤ 15; pooled data are shown).

Table 3.

A comparison of electrophysiological data from α₁β₃γ₂ and α₆β₃γ₂ using bath-applied PRO or TCEt, picospritzer-applied GABA (values are pooled means ± SEM, 4 ≤ n ≤ 15)

Anesthetic	Cell line	EC50	Hill coefficient	Maximum (% enhancement)
PRO	α1β3γ2	2.6 ± 0.5 μM	1.3 ± 0.2	564 ± 53
	α6β 3γ2	4.6 ± 1.9 μM	1.3 ± 0.3	136 ± 17
TCEt	α1β3γ2	0.8 ± 0.2 mM	1.0 ± 0.6	258 ± 21
	α6β 3γ2	0.6 ± 0.2 mM	1.3 ± 0.3	60 ± 3

Fig. 4.

PRO causes greater direct gating in α₆ than α₁, but greater potentiation in α₁ than α₆-containing receptors. GABA and anesthetics were applied with the rapid solution changer. (A and B) PRO enhances EC₁₀ GABA responses in either (A) α₁β₃γ₂ or (B) α₆β₃γ₂ GABA_A receptors. The first trace shows a pre-anesthetic control response to GABA; the last trace shows the response to a maximal concentration of GABA. Bars over current traces indicate GABA and PRO applications, respectively, with the concentration of each drug applied given in μM. In both receptor combinations, pre-application of PRO often elicited a direct inward current in the absence of GABA. (C) PRO directly elicits an inward current in the absence of applied GABA, displayed here as a percentage of the maximal GABA response. Bars show the percentage of direct gating ± SEM from pooled experiments, with the number of experiments given in parentheses. This direct current is significantly greater at 10 μM PRO for the α₆β₃γ₂ receptors (p < 0.05). (D) Enhancement of an EC₁₀ GABA by 1 and 10 μM PRO is significantly greater in the α₁β₃γ₂-containing receptors. The ordinate shows the percentage enhancement of the GABA dose (± SEM) in the presence of PRO compared to the GABA dose alone. (Number of experiments shown in parentheses.)

Significant potentiation of submaximal GABA-induced chloride currents by PRO was first observed at 0.2 μM and 0.5 μM PRO for the α₁β₃γ₂ and α₆β₃γ₂ GABA_A receptors, respectively (Fig. 3(A), p < 0.05 for each). Table 3 shows that the magnitude of the maximal potentiation of GABA-induced currents by PRO in these picospritzer experiments was significantly greater in cells expressing the α₁ isoform relative to the α₆ isoform (maximum: 4. l-fold greater efficacy; p < 0.05). Statistically significant differences between the efficacy for PRO modulation of α₁ and α₆ receptors were observed at all concentrations > 0.2 μM PRO.

The rapid solution changer was then used to investigate GABA_A receptor modulation in greater detail. As shown in Fig. 4, and Table 4, the efficacy of PRO potentiation applied at 10 μM was higher in α₁ than at α₆-ontaining receptors (1024 ± 203% vs 120 ± 39%; percent enhancement of an EC₁₀ GABA test concentration by α₁β₃γ₂ and α₆β₃γ₂-containing receptors, respectively; p < 0.05).

Table 4.

A comparison of α₁β₃γ₂ and α₆β₃γ₂ using the rapid solution changer to apply GABA or PRO or MTX (values represent means ± SEM pooled from multiple experiments with number of experiments in parentheses)

Cell line	Concentration (μM)	PRO (% Enhancement)	MTX (% Enhancement)
α1β3γ2	1	178 ± 21 (5)	125 ± 16 (6)
	10	1024 ± 203 (5)	816 ± 63 (6)
α6β3γ2	1	70 ± 18 (6)	124 ± 43 (7)
	10	120 ± 39 (5)	309 ± 86 (6)

Direct activation of GABA_A receptors

As Fig. 4 illustrates, some of the single cell responses to PRO show an inward current during pre-equilibration of the anesthetic, prior to co-application with GABA. This direct channel gating by 10 μM PRO was much more pronounced in α₆β₃γ₂ than α₁β₃γ₂-containing receptor (Fig. 4(C): 24.0 ± 6.4% versus 5.3 ± 1.9% of maximal current, respectively; p< 0.05).

Effect of other anesthetics

To determine whether these findings applied to other GABA modulators, the barbiturate MTX was also studied (Fig. 5 and Table 4). As with PRO, the efficacy of MTX enhancement of an EC₁₀ GABA test concentration was higher in α₁ than α₆-containing receptors (Table 4; 816 ± 63% vs 309 ± 86%, respectively; p < 0.05). Figure 5 shows that, in agreement with previously published data studying pentobarbital modulation of GABA_A receptors expressed in Xenopus oocytes (Thompson et al., 1996), 10 μM MTX also elicits significantly larger direct current in α₆β₃γ₂ than α₁β₃γ₂ containing receptors (16.7 ± 3.3% versus 2.4 ± 0.8% of maximal current, respectively; p < 0.05). Trichloroethanol was also studied, using picospritzer applied GABA (Fig. 3(B), Table 3). This drug also potentiated submaximal GABA currents, although it was less efficacious than PRO (2.2-fold less efficacious at both receptor subtypes). In a similar fashion to PRO, the efficacy of TCEt potentiation was significantly greater in α₁β₃γ₂ than α₆β₃γ₂-containing receptors p < 0.05).

Fig. 5.

MTX causes greater direct gating in α₆ than α₁, but greater potentiation in α₁ than α₆-containing receptors. Experimental design and figure representations are analogous to Fig. 4. (A and B) enhancement of submaximal GABA responses by MTX in cells expressing either (A) α₁β₃γ₂ or (B) α₆β₃γ₂ GABA_A receptors. As with PRO, preapplication of MTX often elicited an inward current directly. (C) MTX directly activates an inward current in the absence of applied GABA, shown as a percentage of the maximal GABA response. This direct current is significantly greater at both 1 and 10 μM MTX for the α₆β₃γ₂ receptors (p < 0.05). (D) Enhancement of an EC₁₀ GABA current response by 1 and 10 μM MTX is significantly greater in the α₁β₃γ₂ at 10 μM.

DISCUSSION

The concentrations of PRO, TCEt, and MTX that caused enhancement of GABA-induced currents in this study correlate with clinically relevant anesthetic concentration ranges determined in vivo. PRO induces general anesthesia in rats and dogs with an estimated EC₅₀ of 0.4 μM free PRO (Franks and Lieb, 1994), and TCEt anesthetizes canines and humans at concentrations in the range of 0.2–2 mM (Breimer, 1977; Garrett and Lambert, 1973). Threshold anesthetic concentrations for MTX in humans are estimated to range from 12 to 37 μM (Lauven et al., 1987).

The results of this study underline the importance of subunit composition in the modulatory effects of PRO. As summarized in Tables 3 and 4, replacing an α₆ with an α₁ subunit in GABA_A receptors with an identical β₃γ₂ subunit background markedly increased the modulatory action of PRO. Although the larger maximal current (as shown in Table 1) and higher maximal binding [B_max; Fig. 2(A and B)] suggest that more α₁-containing receptors are being expressed, normalization of the enhancement against parallel GABA test concentrations demonstrates that the differential enhancement cannot be explained by differences in receptor expression levels.

Our data with MTX contrast slightly with other investigations of the role of the α subunit in modulation by barbiturates. Our finding that direct activation by MTX is greater in α₆ than α₁-containing receptors is in agreement with two studies assessing pentobarbital modulation in GABA_A receptors expressed in Xenopus oocytes (Wafford et al., 1996; Thompson et al., 1996). However, in contrast to our findings, those two studies found that pentobarbital potentiated submaximal GABA currents to a slightly greater degree in α₆ than α₁-icontaining receptors. These differences may be accounted for by the use of different expression systems, barbiturates, and βγ backgrounds between those studies and ours. Use of the rapid solution changer allows for rapid equilibration of the applied GABA concentration, a condition which may be more difficult to achieve in whole oocytes. Nevertheless, from these studies, as well as our own, it is clear that the α subunit plays a significant role in both PRO and barbiturate modulation.

Inspection of Table 3 reveals a difference in the maximal potentiation (efficacy), but not the EC₅₀, for GABA potentiation by PRO between α₁β₃γ₂ or α₆β₃γ₂ GABA_A receptors. Thus, differences between the structures of the α₁ and α₆ subunit isoforms may be important in the extent of allosteric modulation by PRO, but perhaps not direct binding per se, of the drug to the GABA_A receptor. Previous work has already shown that modulation by PRO does not require the γ subunit (Jones et al., 1995), and in fact, is even seen in β₁ homomers expressed in Xenopus oocytes (Sanna et al., 1995b). Our finding that the α subunit isoform type does not affect the apparent affinity of PRO is consistent with the primary determinants of PRO binding being located on the β subunit, or on components of the α subunit which are highly conserved.