Introduction

Top Abstract Introduction Methods Results Discussion References

The GABA_A receptor is a transmembrane protein complex that forms an ion channel that is selectively permeable to chloride ions. The channel is directly gated by GABA, and the protein subunits that make up the receptor have specific binding sites for convulsants such as picrotoxin and TBPS, as well as barbiturates, benzodiazepines and the anesthetic steroids. Cloning of the GABA_A receptor subunit cDNAs has revealed considerable heterogeneity, with at least six alpha, four beta, three gamma, one delta and two rho subunits, which provides a molecular basis for multiple GABA_A receptor subtypes which distribute throughout the brain (McKernan and Whiting, 1996).

Organochlorine insecticides such as -HCH or lindane also interact with the GABA_A receptor, inducing convulsions of the Grand Mal type (Tussel et al., 1987; Portig and Schnorr, 1988) and inhibiting GABA-induced chloride flux in a wide range of mammalian brain preparations (Abalis et al., 1986; Bloomquist et al., 1986; Pomés et al., 1994) and in Xenopus oocytes expressing rat GABA_A receptors (Woodward et al., 1992). Studies showing that -HCH displaces [³⁵S]TBPS from its binding site in rat brain membrane homogenates (Lawrence and Casida, 1984; Llorens et al., 1990) suggest that this insecticide acts by interacting with the picrotoxin site of the GABA_A receptor. In contrast, the ,  and  isomers of -HCH have been reported to have only weak effects or act as central nervous system depressants (McNamara and Krup, 1948). The site of action of these isomers in the mammalian central nervous system has not been determined yet, although studies, which show them to potentiate GABA-activated chloride channels in Xenopus oocytes (Woodward et al., 1992) and in primary cultures of cortical (Pomés et al., 1994) and dorsal root ganglion neurons (Nagata and Narahashi, 1995), indicate that they interact with the GABA_A receptor. Here we report studies of [³⁵S]TBPS binding in membrane homogenates of a mammalian clonal cell line expressing the human 132S GABA_A receptor subtype and voltage-clamp recordings of human GABA_A receptors expressed in Xenopus oocytes, which suggests that -HCH is a partial inverse agonist of the picrotoxin site and that -HCH interacts with a site linked to the barbiturate locus of the GABA_A receptor.

	Methods

Top Abstract Introduction Methods Results Discussion References

Human GABA_A receptor subunit cDNAs. Complementary DNAs encoding human GABA_A 1, 6, 3 and 2S receptor subunits have been described elsewhere (Hadingham et al., 1993a,b, 1996).

Electrophysiology. To obtain oocytes, adult Xenopus frogs were anesthetized by immersion for 30 to 45 min in a solution of 0.4% ethyl m-aminobenzoate (Tricaine), and a small piece of ovary was removed through an incision in the abdominal wall. Stage V and VI oocytes were isolated and the theca and epithelial cell layer were removed mechanically with fine watchmaker forceps. Follicle cells were removed by an 8-min treatment in Sigma type IA collagenase (0.5 mg/ml) dissolved in MBS (88 mM NaCl, 1 mM KCl 10 mM HEPES, 0.82 mM MgSO₄, 0.33 mM Ca(NO₃)₂, 0.91 mM CaCl₂, 2.4 mM NaHCO₃, pH 7.5). Oocyte nuclei were directly injected with 10 to 20 nl of sterile buffer (88 mM NaCl, 1 mM KCl, 15 mM HEPES, pH 7.4) containing different combinations of human GABA_A subunit cDNAs engineered into the expression vectors pcDM8 or pcDNAI/Amp. Oocytes were incubated for 1 to 2 days in MBS supplemented with 2 mM sodium pyruvate, 2 ml/l penicillin streptomycin solution (containing 100 U/ml penicillin, 10 mg/ml streptomycin), and 50 mg/l gentamycin. For recording, oocytes were placed in a 50-µl bath and perfused with MBS at 4 to 6 ml/min. Cells were impaled with two 1 to 3 megohm agarose-cushion electrodes containing 2 M KCl and were voltage-clamped at 70 mV. Drugs were applied in the perfusate and GABA was applied until the peak of the response, which for most of the oocytes was 30 sec or less. To measure potentiation by allosteric modulators a submaximal concentration of 20% was chosen, which gave optimal potentiation while remaining on the linear portion of the curve. To measure inhibition of GABA responses an EC₅₀ concentration was chosen, which gave a robust signal without the run-down problems associated with continuous challenging with a maximum concentration. After maximal response to GABA (3 mM), constant responses to an EC₂₀ or EC₅₀ concentration were obtained. An EC₂₀/₅₀ concentration was the concentration of agonist that produced 20% or 50% of the maximal response for that agonist. At least 3 min were allowed between each drug application to prevent desensitization. Concentration-response curves were fitted by use of GraFit (Erithacus software, from Sigma-Aldrich Ltd., Poole, Dorset, U.K.) to the equation: f(x) = B_max/[1 + (EC₅₀/X)ⁿ], where B_max is the response at saturating concentration of ligand; EC₅₀, the concentration of ligand producing a half-maximal response; X, the concentration of ligand; and n, the Hill coefficient. Arithmetic mean values were calculated from data obtained from a number (n) of different cells. The statistical significance of differences between mean values were assessed by Student's two-tailed t tests, wherever appropriate.

GABA_A receptor cell line. Mouse fibroblasts stably expressing the human 132 GABA_A receptor were maintained as described previously (Hadingham et al., 1992). Semiconfluent cells containing the human 132s GABA_A receptor were induced with 2 µM dexamethasone for 8 days to gain maximum expression of the recombinant receptor.

[³⁵S]TBPS binding assay in membranes of induced mouse clonal cells. Eight-day induced cells were washed off culture flasks with phosphate-buffered saline (pH 7.4). Cells were washed twice for 5 min at 200 × g, resuspended in 5 ml of ice-cold 5 mM Tris citrate (pH 7.4) containing 1 mM ethylenediaminetetraacetic acid, 1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine hydrochloride, 1 mM leupeptin and 0.5 mM soybean trypsin, pH 7.4. Cells were then broken up by probe sonication on ice, for a total of 30 sec (three 10-sec bursts to prevent overheating). The homogenate was centrifuged for 5 min to remove nuclear material (200 × g; 4°C), and the supernatant was centrifuged for 1 hr at 90,000 × g, at 4°C. The resulting pellet was resuspended in 5 ml of 50 mM Tris citrate, pH 7.4, and quickly frozen and thawed. The thawed homogenate was then incubated for 30 min in 50 mM Tris citrate containing 0.01% Triton X-100 at 37°C. The homogenate was then washed twice in 50 mM Tris citrate buffer, pH 7.4, by centrifugation at 90,000 × g for 1 hr. The final pellet was resuspended in the required volume of 50 mM Tris citrate buffer. Protein content was determined using a commercially available Bio-Rad kit.

Drugs. Drugs used were: GABA (Sigma-Aldrich), picrotoxin (Sigma-Aldrich), - and -HCH (Sigma-Aldrich), loreclezole (a gift of Jannsens), 5-pregnan-3-ol-20-one (5,3-DHP) (Sigma-Aldrich), lanthanum (Sigma-Aldrich), pentobarbital (Sigma-Aldrich) and propofol (Aldrich Chemical Co., Gillingham, Dorset, U.K.). The - and -HCH isomers were a gift from Dr Stuart Dunbar, Zeneca Agrochemicals, U.K. Solutions of GABA were made in saline, while loreclezole, 5,3-DHP, lanthanum, picrotoxin, propofol and HCH isomers were prepared as 10¹ M or 10² M stocks in DMSO, while pentobarbital was supplied as a 60 mg/ml solution in ethanol. The highest concentration of DMSO or ethanol vehicle perfusing the oocyte was 1%, which had no effects on GABA-induced currents.

	Results

Top Abstract Introduction Methods Results Discussion References

Effects of -HCH and its ,  and  isomers on GABA responses. -HCH and its ,  and  isomers (fig. 1) were studied on recombinant human GABA_A 132S receptors to investigate their effect on GABA-mediated responses. The isomers (100 µM) were applied 30 sec before the coapplication of an EC₂₀ GABA concentration (or EC₅₀ for -HCH). As shown in figure 2, the - and -HCH isomers potentiated the GABA-mediated response by about 150% and 185%, respectively, as well as the  isomer producing a small amount of receptor activation in the absence of GABA. In contrast, -HCH inhibited the GABA EC₅₀ response by approximately 30%, and the  isomer had no significant effect. To elucidate the mechanism underlying the two opposite effects of the HCH isomers, we characterized the actions of -HCH and its  isomer in more detail.

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Structures of the -, -, - and -HCH stereoisomers.

View larger version (14K): [in this window] [in a new window]   Fig. 2.   Effects of HCH isomers on GABA responses in oocytes expressing human 132S GABAA receptors. Oocytes were clamped at 70 mV, and EC20 or EC50 GABA responses were elicited in the presence or absence of 100 µM HCH isomer. To study potentiation of the GABAA receptor, a submaximal EC20 concentration was chosen in (a), (b) and (d). To measure inhibition with -HCH (c), an EC50 concentration was used (see "Methods"). -HCH inhibited EC50 GABA responses, whereas the  and  isomers clearly potentiated EC20 GABA responses, with -HCH having an additional direct action on GABAA receptors. -HCH had no effect on EC20 GABA-mediated currents.

- and -HCH concentration-response curves. To determine the potency and efficacy of the two isomers on GABA responses in oocytes expressing human 132S GABA_A receptors, concentration-response curves were constructed with an EC₂₀ concentration of GABA for the potentiation by the  isomer and EC₅₀ for the inhibition by the  isomer determined on individual oocytes. Figure 3A shows that the effect of -HCH was concentration-dependent, giving a maximum inhibition of 33.8 ± 3.6% (n = 7), and gave no further inhibition at concentrations greater than 10 µM. The EC₅₀ for -HCH was 1.0 ± 0.3 µM (n = 7). Similar results have also been reported for Xenopus oocytes expressing RNA from rat cerebral cortex (Woodward et al., 1992). Concentration-response curves for -HCH potentiation of GABA responses demonstrated a potentiation of the GABA EC₂₀ with a maximum of 216 ± 31% (n = 7) (fig. 3B). The EC₅₀ value for -HCH potentiation was 15.3 ± 3.7 µM (n = 7).

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Concentration-response curves for the effects of - and -HCH on GABA responses on oocytes expressing human 132S (filled circles) or 632S (open circles) GABAA receptors. (a) The inhibitory action of increasing concentrations of -HCH on EC50 GABA responses was studied. (b) The effect of the  isomer was analyzed on EC20 GABA responses (see "Methods").

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Concentration-response curve for the direct effect of -HCH on oocytes expressing 132S (filled circles) or 632S (open circles) GABAA receptors. Each point represents the arithmetic mean ± S.E.M. calculated as a percentage of the response obtained with a maximum concentration of GABA (3 mM).

Effects of - and -HCH on picrotoxin inhibition of GABA responses of 132S GABA_A receptors. Because -HCH gave only partial inhibition of the GABA EC₅₀ response, we determined whether either of the isomers could interact with the picrotoxin binding site. We studied the effect of these compounds on picrotoxin inhibition of GABA responses on the 132S subunit combination. As shown in figure 5A, both -HCH and the  isomer when coapplied with picrotoxin could reverse the inhibitory effect of picrotoxin on GABA responses. Figure 5B shows that the concentration-response curve for picrotoxin inhibition of GABA responses was shifted to the right by coapplication of -HCH in a dose-dependent manner, with EC₅₀ values of 1.3 ± 0.3 µM (n = 4), 7.0 ± 2.1 µM (n = 4) in 10 µM -HCH, and 116.2 ± 30.9 µM (n = 5) in 100 µM -HCH, which suggested that -HCH is a competitive inhibitor at the picrotoxin site. The  isomer also produced a similar rightward shift of the picrotoxin dose-response curve (fig. 5B), which suggested that this isomer too displaces picrotoxin from its binding site.

View larger version (21K): [in this window] [in a new window]   Fig. 5.   The effect of - and -HCH on picrotoxin-mediated inhibition of GABA responses on oocytes expressing human 132S GABAA receptors. (a) Both - and -HCH reversed the inhibitory action of picrotoxin on EC50 GABA responses. Cells were voltage-clamped at 70 mV and exposed to GABA, picrotoxin and HCH isomers as indicated by the bars above each trace. (b) The effect of 10 µM (open triangles) and 100 µM (open squares) -HCH, and 100 µM -HCH (filled triangles) on concentration-response curves for picrotoxin-mediated inhibition of EC50 GABA responses (filled circles). Each point represents the arithmetic mean ± S.E.M. of four or five experiments.

View larger version (27K): [in this window] [in a new window]   Fig. 6.   Concentration-effect curves for the inhibition of [35S]TBPS-specific binding by HCH isomers in a mouse fibroblast cell line expressing human 132S GABAA receptors. HCH isomers plotted are -HCH (open hexagon); -HCH (open triangle); -HCH (filled square); and -HCH (filled circle). The data for each HCH isomer represent four independent experiments. HCH isomers (1 nM to 100 µM) were incubated with membrane homogenates and 5 nM [35S]TBPS.

The effect of -HCH on -HCH-dependent potentiation of GABA responses of 132S GABA_A receptors. To test whether -HCH had any effect on the positive potentiating action of -HCH, we examined the effect of 10 µM and 100 µM -HCH on the concentration-response curve of -HCH potentiation in oocytes expressing 132S GABA_A receptors. As shown in figure 7, 10 µM -HCH reduced the maximum potentiation of the GABA EC₂₀ induced by the  isomer from 216 ± 31% to 139 ± 13% without causing a shift in the EC₅₀ value of the -HCH-dependent potentiation. When the experiment was repeated with 100 µM -HCH, the potentiation was completely blocked (fig. 7). These studies suggest that -HCH can allosterically affect the potentiation of GABA responses by the  isomer.

View larger version (20K): [in this window] [in a new window]   Fig. 7.   The effect of -HCH on -HCH-mediated potentiation of GABA responses on oocytes expressing the human 132S GABAA receptors. The concentration-response curve for -HCH-induced potentiation of EC20 GABA responses was obtained in the absence (open circles) and presence of 10 µM (open square) or 100 µM (filled circle) -HCH. Each point represents the arithmetic mean ± S.E.M. of four or seven experiments.

Effects of -HCH on other positive modulators of the GABA_A receptor. To characterize in more detail the effect of the  isomer, we tested the effect of 100 µM -HCH on the positive allosteric action of several modulators of the GABA_A receptor. As shown in figure 8 saturating concentrations of loreclezole (10 µM), 5-pregnan-3-ol-20-one (300 nM), lanthanum (1 mM), pentobarbital (100 µM) and propofol (10 µM) caused a potentiation of the GABA response. Coapplication of 100 µM -HCH had an additive effect with the potentiation mediated by loreclezole, 5-pregnan-3-ol-20-one and lanthanum, which indicated that -HCH potentiation is not mediated through any of these binding sites. In contrast, -HCH either gave no further potentiation, or reduced the potentiation mediated by either pentobarbital or propofol (fig. 8, D and E, respectively), which suggested that that -HCH may act through a site common to both propofol and pentobarbital. To determine whether pentobarbital and propofol interacted with the same site, we carried out additive experiments with these two compounds. As shown in figure 8F, the effects of propofol and pentobarbital show some additivity when combined. In addition, -HCH inhibited the direct activation of the receptor by pentobarbital (data not shown).

View larger version (12K): [in this window] [in a new window]   Fig. 8.   Effects of -HCH on the potentiation of GABA responses induced by positive allosteric modulators of the GABAA receptor. Oocytes expressing human 132S GABAA receptors were voltage-clamped at 70 mV and exposed to an EC20 concentration of GABA (G), followed by the allosteric modulator alone, followed by this in combination with 100 µM -HCH (-HCH), as indicated by the bars above each trace. Traces show (a) 10 µM loreclezole (Lor), (b) 300 nM 5-pregnan-3-ol-20-one (5), (c) 1 mM lanthanum (Lan), (d) 100 µM pentobarbital (Pen) and (e) 10 µM propofol (Pro). Trace (f) shows the combination of pentobarbital (100 µM) and propofol (10 µM). Traces are typical of at least four experiments in each case.

Effects of -HCH on pentobarbital potentiation of the GABA response. In light of the previous finding that -HCH could allosterically inhibit -HCH potentiation, we investigated whether -HCH had any effect on the potentiation by the barbiturate pentobarbital. As shown in figure 9, 100 µM -HCH inhibited the dose-dependent potentiation of the EC₂₀ GABA responses of 132S GABA_A receptors by pentobarbital, which suggested that, in addition to its effect on the picrotoxin binding site, -HCH has an allosteric effect inhibiting barbiturate potentiation.

View larger version (19K): [in this window] [in a new window]   Fig. 9.   Concentration-response curves for pentobarbital, with (filled circle) and without (open circle) 100 µM -HCH, on oocytes expressing human 132S GABAA receptors. Each point represents the arithmetic mean ± S.E.M. of four experiments and calculated as percentage of EC20 potentiation.

	Discussion

Top Abstract Introduction Methods Results Discussion References

In the present study, we have shown that -HCH and its  and  isomers interact with human 132S or 632S GABA_A receptors expressed in oocytes, which supports the view that these substances act primarily at GABA_A receptors. In contrast, we found that the  isomer is essentially inactive on either GABA responses or [³⁵S]TBPS binding, confirming previous suggestions that this isomer is essentially a non-GABAergic HCH isomer (Woodward et al., 1992; Nagata and Narahashi, 1995). However, a depressant action for the  isomer has been well substantiated (Vohland et al., 1981; Stark et al., 1986), and Pomés et al. (1994) has demonstrated a biphasic action (positive modulation at concentrations lower than 1 µM and inhibition at higher concentrations) of -HCH on GABA-dependent ³⁶Cl uptake into primary cultures of rat neocortical neurons. It is not easy to solve these discrepancies, but it is possible that differences in intracellular signaling mechanisms in different preparations and the ability of HCH isomers to interact with intracellular targets such as ryanodine-sensitive calcium channels (Pessah et al., 1992) may account for them. Thus, the presence of HCH-sensitive intracellular receptors in a preparation may result in the activation of kinase enzymes, which are known to modulate the activity of GABA_A receptors (Leidenheimer et al., 1992).

-HCH suppresses GABA-induced currents in dorsal root ganglion neurons (Nagata and Narahashi, 1995), insect neurons (Wafford et al., 1988; Bermudez et al., 1991) and oocytes expressing mammalian GABA_A (Woodward et al., 1992) or Drosophila GABA (Zhang et al., 1994; Belelli et al., 1995) receptors. -HCH also inhibited GABA currents on oocytes expressing human 132S or 632S GABA_A receptors with an EC₅₀ value of about 1 µM, which is in agreement with values reported for other mammalian preparations (Woodward et al., 1992; Nagata and Narahashi, 1995). Our studies showed -HCH to be a partial inhibitor of GABA_A receptors, suppressing EC₅₀ GABA responses by a maximum of 34%. Partial inhibition of GABA-induced currents by -HCH has also been demonstrated by Woodward et al. (1992) in Xenopus oocytes expressing rat or bovine GABA_A receptors. However, full -HCH-induced inhibition has been reported for dorsal root ganglion neurons (Nagata and Narahashi, 1995) and Xenopus oocytes expressing a Drosophila GABA receptor subunit (Zhang et al., 1994). This suggests that the degree of efficacy may depend on the type of GABA receptor subunits present; we have shown here that it is unlikely to be caused by different alpha subunits.

-HCH acts in a similar fashion to picrotoxin. For example, there are several reports showing -HCH-induced inhibition of [³⁵S]TBPS binding to neuronal membranes (Lawrence and Casida, 1984; Llorens et al., 1990) and GABA-mediated ³⁶Cl fluxes (Abalis et al., 1986; Pomés et al., 1994). The binding site of picrotoxin is likely to be in the M2 region of the Cl channel as suggested by site-directed mutagenesis studies of the putative M2 domain of rho receptors (Wang et al., 1995), -glycine receptor subunits (Pribilla et al., 1992) and a Drosophila GABA subunit (ffrench-Constant et al., 1993; Belelli et al., 1995). Interestingly, mutations of the picrotoxin site (alanine 302) in a Drosophila GABA subunit decreases the potency of both picrotoxin and -HCH inhibition (Zhang et al., 1994; Belelli et al., 1995), which suggests that these two compounds interact with identical residues in the M2 region of GABA_A receptors.

It is not clear as yet whether picrotoxin, and hence -HCH, inhibits GABA responses by merely obstructing the Cl channel or by allosteric blockade. Studies showing that picrotoxin binds preferentially to the desensitized conformation of GABA_A receptors in sympathetic rat neurons (Newland and Cull-Candy, 1992), and that -HCH accelerates the desensitization of GABA_A receptors in dorsal root ganglion neurons (Nagata and Narahashi, 1995), suggest that picrotoxin and -HCH are allosteric inhibitors of the GABA_A receptor. Thus, by analogy to the benzodiazepine site, we propose that picrotoxin is a full inverse agonist of the picrotoxin modulatory site. In contrast, -HCH, which both partially blocks GABA responses mediated by GABA_A receptors expressed in Xenopus oocytes (this study; Woodward et al., 1992) and relieves picrotoxin-induced inhibition of human GABA_A receptors (this study), behaves as a partial inverse agonist of the picrotoxin site. This action is similar to that of fluorinated methylbutyrolactone, which has also been proposed to be a partial inverse agonist of the picrotoxin site (Yoon et al., 1990). In other preparations such as dorsal root ganglion neurons, -HCH may act as a full inverse agonist, presumably reflecting species or GABA receptor subtype differences.

-HCH potentiates GABA responses in Xenopus oocytes expressing rat and bovine GABA_A (Woodward et al., 1992), Drosophila GABA (Belelli et al., 1996) receptors, dorsal root ganglion (Nagata and Narahashi, 1995) and cortical (Pomés et al., 1994) neurons. We also observed -HCH-dependent potentiation of human 132S and 632S GABA_A receptors. -HCH potentiation was not influenced by alpha subunit composition, which is in contrast to other positive allosteric modulators of the GABA_A receptor such as barbiturates (Thompson et al., 1996) and benzodiazepines (Wafford et al., 1993)

Several lines of evidence argue against the idea of an identical binding locus for the  and  isomers. First, a A302S mutant of a Drosophila GABA subunit receptor shows reduced susceptibility to inhibition by picrotoxin and -HCH without changes in its affinity toward the positive allosteric modulator -HCH (Belelli et al., 1995). Second, rho receptors are endowed with a picrotoxin binding locus but lack a site for positive allosteric modulation by the  isomer (Woodward et al., 1992; Wang et al., 1995). However, the two sites may be close, adjacent or even overlapping, which could explain the inhibition of -HCH-dependent effects by the  isomer, as well as the inhibition of picrotoxin antagonism by the  isomer. Alternatively the -HCH activity at the picrotoxin site may be masked by its potentiating effect, and only when pentobarbital is used as an agonist is the inhibition observed. The inhibition of -HCH potentiation by -HCH is clearly not a competitive interaction because the inhibition cannot be overcome at high concentrations, and the -HCH EC₅₀ is not affected. This would suggest an allosteric interaction between the two sites.

Clues as to where the -HCH site might be located come from our observation that the  isomer reduces the positive allosteric action of both pentobarbital and propofol. Although neither the barbiturate or propofol sites have been defined as yet, there is neurochemical evidence to suggest that they are separate entities (Concas et al., 1991), with the alpha and beta subunits probably contributing to the barbiturate site (Amin and Weiss, 1993; Thompson et al., 1996). The effect of -HCH on the enhancement action of pentobarbital and propofol may be explained then by postulating that the -HCH locus overlaps with the barbiturate and propofol sites, which are close or adjacent to each other. The observation that rho receptors, which are insensitive to positive allosteric modulation by -HCH are insensitive to barbiturates and propofol (Bormann and Feigenspan, 1995) further support the idea of three separate though overlapping loci. It is unlikely that the -HCH site is linked to other sites such as the benzodiazepine site, because flumazenil, an antagonist of the benzodiazepine site, has no effect on -HCH-mediated potentiation of GABA responses (Woodward et al., 1992).

We demonstrated a direct action of -HCH on GABA_A receptors, which has not been reported previously. This action resembled the direct activation of GABA_A receptors by barbiturates (Belelli et al., 1996; Thompson et al., 1996) and propofol (Orser et al., 1994; Belelli et al., 1996). Direct activation of GABA_A receptors by barbiturates (Mathers and Barker, 1980; Thompson et al., 1996) and propofol (Orser et al., 1994) is not mediated through the GABA binding site (Thompson et al., 1996; Ueno et al., 1997), which suggests that the site for direct activation by the  isomer is a separate locus from the agonist site. Recent evidence suggests that homomeric beta subunits form receptors that can be gated by barbiturates and propofol (Sanna et al., 1995; Krishek et al., 1996), but only weakly by GABA, which suggests a role for the beta subunit in this activation. Because the -HCH isomer seems to be acting via a site common to these compounds, it may also be activating the receptor in a similar manner, although to a lesser extent than pentobarbital or propofol. A single channel approach as well as the use of pharmacological tools such as competitive inhibitors of the GABA binding site might help to define more precisely this site.