The interaction of the general anesthetic etomidate with the γ-aminobutyric acid type A receptor is influenced by a single amino acid

Abstract

The γ-aminobutyric acid type A (GABA_A) receptor is a transmitter-gated ion channel mediating the majority of fast inhibitory synaptic transmission within the brain. The receptor is a pentameric assembly of subunits drawn from multiple classes (α_1–6, β_1–3, γ_1–3, δ₁, and ɛ₁). Positive allosteric modulation of GABA_A receptor activity by general anesthetics represents one logical mechanism for central nervous system depression. The ability of the intravenous general anesthetic etomidate to modulate and activate GABA_A receptors is uniquely dependent upon the β subunit subtype present within the receptor. Receptors containing β₂- or β₃-, but not β₁ subunits, are highly sensitive to the agent. Here, chimeric β₁/β₂ subunits coexpressed in Xenopus laevis oocytes with human α₆ and γ₂ subunits identified a region distal to the extracellular N-terminal domain as a determinant of the selectivity of etomidate. The mutation of an amino acid (Asn-289) present within the channel domain of the β₃ subunit to Ser (the homologous residue in β₁), strongly suppressed the GABA-modulatory and GABA-mimetic effects of etomidate. The replacement of the β₁ subunit Ser-290 by Asn produced the converse effect. When applied intracellularly to mouse L(tk−) cells stably expressing the α₆β₃γ₂ subunit combination, etomidate was inert. Hence, the effects of a clinically utilized general anesthetic upon a physiologically relevant target protein are dramatically influenced by a single amino acid. Together with the lack of effect of intracellular etomidate, the data argue against a unitary, lipid-based theory of anesthesia.

The years 1996 and 1997 mark the sesquicentenary of the clinical use of ether and chloroform, respectively, as general anesthetics (1). However, the molecular mechanisms that underlie the remarkable behavioral actions of general anesthetics remain controversial (2). The long-established correlation between the anesthetic potency and lipid solubility of a wide range of structurally unrelated agents underpins the Meyer–Overton rule, which asserts that general anesthetics act at a hydrophobic site to produce anesthesia. Although the lipid bilayer is often equated with the molecular site of anesthetic action, the effects of clinically relevant concentrations of anesthetics upon the properties of the bilayer as a whole are modest and can be mimicked by very small changes in temperature (3). In addition, the loss of anesthetic activity that is observed for higher, yet still lipid soluble, members of homologous series of anesthetics such as the n-alcohols and n-alkanes [the cut-off phenomenon (3, 4)] poses unresolved problems for the Meyer–Overton rule (5). The validity of the latter has been further undermined by recent studies with certain polyhalogenated and perfluorinated compounds, which did not demonstrate the anesthetic activity predicted by the Meyer–Overton rule (6).

In view of the above, hydrophobic regions of membrane proteins, and/or discrete hydrophobic pockets at lipid-protein interfaces, are now widely regarded as more plausible molecular sites of general anesthetic action (2, 7). Neurotransmitter-gated ion channels are particularly sensitive to anesthetic agents (2), and within this superfamily a strong case can be made for the involvement of the inhibitory γ-aminobutyric acid type A (GABA_A) receptor in the action of structurally diverse general anesthetics (2, 7–9). At anesthetic or subanesthetic concentrations, many general anesthetics act to potentiate GABA_A receptor-mediated electrical responses. Higher concentrations of such agents can elicit direct activation of the receptor complex (2, 7–9).

A direct interaction between general anesthetics and the GABA_A receptor gains credence from studies that demonstrate the subunit composition of the pentameric complex (which may be drawn from α_1–6, β_1–3, γ_1–3, δ, or ɛ subtypes) to affect either the modulatory and/or agonist actions of certain agents (10–13). For the clinically utilized anesthetic etomidate [i.e., (R)-(+)-ethyl-1-(1-phenylethyl)-1H-imidazole-5-carboxylate], the α and β subunit isoforms present within the receptor complex influence these effects (12). In this communication, we report that the actions of etomidate at recombinant GABA_A receptors are strongly influenced by a single amino acid residue within the β subunit.

MATERIALS AND METHODS

GABA_A Receptor Subunit cDNA Clones.

The cDNAs encoding human α₆, β₁, β₂, β₃, γ_2S, and γ_2L have been described previously (14–17). The β subunit chimeras used here, denoted β_1–2 and β_2–1, correspond to the β2Δ1.1 and β2Δ1.2 constructs, respectively, of Wingrove et al (18). In the β_1–2 chimera, all residues including and carboxyl terminal to Lys-238 of the β₁ subunit, are replaced by the corresponding β₂ sequence. In the β_2–1 chimera, the β₁ sequence substitutes that of β₂ over the same region. The construction of the single point mutants β₁ Ser-290 → Asn, β₃ Asn-289 → Ser and β₃ Asn-289 → Met has been reported (18, 19).

Expression of GABA_A Receptor Subunits.

cDNAs encoding wild-type, chimeric, and point mutant human GABA_A receptor subunits in the pCDM8 vector were linearized by the appropriate restriction enzymes and cRNA transcripts prepared as described (20). Transcripts were injected (50 nl of 1 mg ml⁻¹ cRNA per subunit) into Xenopus laevis oocytes (stage V–VI) that had been defolliculated by treatment with collagenase (2 mg ml⁻¹ in Ca²⁺-free Barth’s saline) for 2–3 hr at 18–23°C (20). Oocytes were maintained in Barth’s saline [composition 88 mM NaCl/1 mM KCl/2.4 mM NaHCO₃/1 mM MgSO₄/0.5 mM CaCl₂/0.5 mM Ca(NO₃)₂/15 mM Hepes, pH 7.5] supplemented with gentamicin (0.1 mg ml⁻¹). The transiently expressed receptors were studied 2–12 days following injection. Human recombinant GABA_A receptors with the subunit composition α₆β₃γ_2S were additionally studied in mouse L(tk−) cells stably transfected with individual subunit DNAs in the dexamethasone-inducible eukaryotic expression vector pMSGneo as previously described (17). Cells were maintained in growth medium containing minimal essential medium (MEM) supplemented with 10% fetal calf serum (FCS; vol/vol), 2 mM glutamine, 100 units ml⁻¹ penicillin, and 100 μg ml⁻¹ streptomycin at 37°C in a humid atmosphere of 95% air/5% CO₂. Receptor expression was induced by the addition of 1 μM dexamethasone at least 24 hr prior to recordings.

Preparation and Culture of Cerebellar Granule Cells.

Primary cultures of cerebellar granule neurons were prepared from neonatal Wistar rat (day 8) cerebella essentially as described by Courtney et al. (21). In brief, cerebella were cut into blocks, incubated with trypsin (0.25 mg ml⁻¹) in PBS for 20 min at 37°C, and subsequently dispersed into a single cell suspension by trituration. Neuronal perikarya were sedimented by centrifugation (100 × g for 5 min) through a cushion of 4% bovine serum albumin (wt/vol) in Earle’s balanced salt solution and the isolated cells transferred to a culture medium comprising MEM supplemented with 10% FCS (vol/vol), 20 mM KCl, 30 mM glucose, 2 mM glutamine, 50 units ml⁻¹ penicillin, and 50 μg ml⁻¹ streptomycin. Cells were seeded onto poly-l-lysine-coated glass coverslips and maintained as described above for mouse L(tk−) cells. 5-Fluor-2′-deoxyuridine (80 μM) was added to the medium 48 hr after plating to suppress the replication of nonneuronal cells.

Electrophysiological Recordings.

Agonist-evoked currents were recorded from RNA-injected oocytes with a two-electrode voltage clamp and a holding potential of −60 mV (20). The voltage sensing and current passing electrodes were filled with 3 M KCl and had resistances of 0.5–1.5 MΩ. The oocytes were held in a 0.5 ml chamber and were continually superfused (7–10 ml min⁻¹) with a solution containing 120 mM NaCl, 2 mM KCl, 1.8 mM CaCl₂, and 5 mM Hepes⋅NaOH (pH 7.4). All drugs were applied via the superfusate. GABA-modulatory activity was examined against the GABA EC₁₀ determined for each oocyte. The GABA-modulatory or GABA-mimetic activities of the drugs tested were expressed as a percentage of the response to a maximally effective concentration of GABA (12). Concentration response relationships were fitted with the logistic equation [E = E_max/1 + (EC₅₀/D)ⁿ] to derive the maximal drug effect (E_max), the concentration of drug producing a half-maximal effect (EC₅₀) and interaction coefficient (n) where D is the concentration of the drug and E is the response in the presence of the drug at concentration D. Agonist-evoked currents were recorded from rat cerebellar granule cells and mouse L(tk−) cells using the whole-cell recording mode of the patch-clamp technique. In all experiments, the cells were superfused (3–5 ml min⁻¹) with an extracellular solution comprising 140 mM NaCl, 2.8 mM KCl, 1 mM CaCl₂, 2 mM MgCl₂, 10 mM glucose, 20 mM sucrose, and 10 mM Hepes (pH 7.2). GABA (3 μM) or etomidate (10–30 μM) was applied locally by pressure ejection (1.4 × 10⁵ Pa; 20–60 msec) from a modified patch pipette, and modulatory agents were delivered by the superfusate. Patch electrodes filled with a solution containing 140 mM CsCl (or KCl), 1 mM CaCl₂, 2 mM MgCl₂, 11 mM EGTA, and 10 mM Hepes (pH 7.2) had resistances in the range 3–5 MΩ and series resistance compensation of up to 50% was applied when appropriate. In experiments examining the site of action of etomidate, the drug was included within the pipette solution at a concentration of 10 μM. Under the conditions employed, the time constant of diffusional exchange between the pipette and cell interiors (22) was estimated to be approximately 96 sec (range = 63–178 sec). All data are reported as the mean ± SEM.

RESULTS

Acting upon oocytes expressing recombinant GABA_A receptors with the subunit composition α₆β₃γ₂, etomidate produced a potent (EC₅₀ = 0.7 ± 0.06 μM) and large enhancement (E_max = 135 ± 7%; n = 4) of the control response evoked by GABA. The potency and maximal effect of etomidate were similar to the values previously determined under identical conditions for the α₆β₂γ₂ subunit combination, but were distinct from those found for receptors composed of α₆β₁γ₂ subunits (12) (Fig. 1, Table 1). In the absence of GABA, etomidate, acting at α₆β₃γ₂ receptors, evoked a concentration-dependent (EC₅₀ = 23 ± 2.4 μM) inward current response with a maximal amplitude amounting to 96 ± 24% of that elicited by a saturating concentration of GABA (n = 4). Once more, these values are similar to those found for α₆β₂γ₂ receptors, but contrast with results obtained with α₆β₁γ₂ receptors, at which etomidate is ineffective as an agonist (12) (Table 1). Hence, when compared with the β₁ subunit, the coexpression of either the β₂ or β₃ subunits with α₆ and γ₂ subunits greatly favors both the potentiating and agonist actions of etomidate. This selectivity is shared by loreclezole (23). However, for the subunit combination α₆β₃γ₂, both the GABA-enhancing (EC₅₀ = 2.5 ± 0.3 μM, E_max = 59 ± 2%; n = 3) and the GABA-mimetic (11 ± 4%, n = 3 for 10⁻⁴ M loreclezole) effects of this anticonvulsant are less than those of etomidate (Fig. 1).

Figure 1.

Comparison of the actions of etomidate and loreclezole on GABA_A (α₆β₃γ_2L) receptors expressed in X. laevis oocytes. (A) Graph showing the concentration-effect relationship for the GABA-modulatory (solid symbols) and GABA-mimetic (open symbols) of etomidate (triangles) and loreclezole (circles) expressed as a percentage of the peak amplitude of the current produced by a maximally effective concentration of GABA (3 mM). Each point represents the mean ± SEM determined from 3–5 oocytes. The curves, fitted by eye, have no theoretical significance. (B and C) Exemplar traces illustrating the effects of a maximally effective concentration (10 μM) of etomidate (B) or loreclezole (C) on the inward current induced by GABA applied at EC₁₀. Note the direct inward current elicited by the pre-application of these agents prior to the addition of GABA.

Table 1.

A comparison of the GABA-modulatory and GABA-mimetic activities of etomidate across wild-type, chimeric, and point-mutant GABA_A receptors expressed in X. laevis oocytes

Subunit combination	Modulatory EC50, μM	Modulatory Emax, %	Agonist EC50, μM	Agonist Emax, %
α6β1γ2*	7.4 ± 0.6	28 ± 2	ND	5 ± 2 (1 mM)
α6β2γ2*	0.6 ± 0.04	169 ± 14	22 ± 1	51 ± 15
α6β3γ2	0.7 ± 0.06	135 ± 7	23 ± 2.4	96 ± 24
α6β1-2γ2	1.3 ± 0.1	226 ± 47	34 (n = 2)	38 (n = 2)
α6β2-1γ2	7.7 ± 1.0	23 ± 2	ND	4.4 ± 1.2 (0.6 mM)
α6β1 S290Nγ2	1.6 ± 0.3	150 ± 13	79 ± 6	45 ± 13
α6β3 N289Sγ2	5.7 ± 2.6	36 ± 1	55 ± 6	14 ± 3
α6β3 N289Mγ2	ND	11.8 ± 0.6	ND	3.5 ± 0.4 (0.6 mM)

Numbers in parentheses refer to the maximum concentration of etomidate tested as an agonist for receptors containing the wild-type β₁ subunit, β_2-1 chimera, or β3_N289M point mutant. The E_max is expressed relative to the maximal response to GABA. All data, except where indicated, are mean ±SEM values obtained from three to five oocytes. ND, not determined due to limited potentiation or direct agonist action. 

^*Data from ref. 12 included for the purpose of comparison. 

The GABA-mimetic action of etomidate was further investigated by using whole-cell clamp techniques on L(tk−) cells stably transfected with α₆β₃γ_2S human GABA_A receptor subunits (17). Here, the brief local application of etomidate (30 μM, 1.4 × 10⁵ Pa, 40 msec) produced an inward current (Vh = −40 mV) on all cells tested (Fig. 2A). These currents reversed in sign at approximately 0 mV and similar to GABA (not shown), exhibited outward rectification (Fig. 2B). The responses induced by etomidate were inhibited to 54 ± 8% of control (n = 3) by the GABA_A receptor antagonist bicuculline (30 μM; Fig. 2A), and enhanced to 177 ± 20% of control (n = 3) by the neurosteroid 5α-pregnan-3α-ol-20-one (300 nM, not shown). Collectively, these observations are consistent with etomidate directly activating the GABA_A receptor. In contrast to the neurosteroid, 3 μM of loreclezole reduced the etomidate-induced response to 74 ± 1% of control (n = 7), supporting the contention that these agents may act through a common site. In view of the β subunit selective actions of etomidate upon recombinant receptors, we additionally examined the effect of this agent upon the GABA_A receptors native to cerebellar granule neurons, which from studies examining the regional distribution of GABA_A receptor subunit mRNAs are likely to incorporate β₂/β₃ subunits. Consistent with the data obtained with the heterologously expressed receptor, all granule cells challenged with locally applied etomidate responded with an inward current, which was susceptible to blockade by picrotoxin (30 μM; Fig. 2C).

Figure 2.

The GABA-mimetic effects of etomidate. (A) Traces illustrating inward currents (holding potential = −40 mV) evoked by the brief local application of etomidate (30 μM) to an L(tk−) cell stably transfected with human α₆β₃γ_2S receptor subunits. Currents evoked by etomidate are antagonized by 30 μM bath-applied bicuculline. (B) Representative current-voltage plot depicting the relationship between holding potential and the peak amplitude of responses induced by 30 μM etomidate recorded from an L(tk−) cell expressing the α₆β₃γ_2S receptor subunits. (Inset) Family of currents from which the graph was derived. (C) Brief application of 30 μM etomidate to a cerebellar granule cell (holding potential = −60 mV) induces an inward current that is antagonized by 30 μM picrotoxin.

The discriminatory influence of the β subunit isoform was further investigated in experiments performed on the chimeric β₁/β₂ subunits described in Materials and Methods. For oocytes expressing receptors assembled from α₆ and γ₂ subunits and a β subunit bearing the N-terminal domain of β₁ (i.e., β_1–2), etomidate produced a potent (EC₅₀ = 1.3 ± 0.1 μM, n = 3) and large potentiation (E_max = 226 ± 47%, n = 3) of the current evoked by GABA (Fig. 3A and Table 1). In addition, etomidate acted as an agonist, with an EC₅₀ of 34 μM and a maximal effect amounting to 38% of the GABA maximum (n = 2). When the aforementioned values are compared with those obtained for receptors containing wild-type β subunits (Table 1), it is clear that they approximate most closely those observed for β₂ or β₃ subunit containing receptors.

Figure 3.

The potency and efficacy of the modulatory and agonist actions of etomidate varies across GABA_A receptors incorporating chimeric and point-mutant β subunits. (A) Graph depicting the modulatory (solid symbols) and agonist (open symbols) actions of etomidate at the β_1–2 (triangles) and β_2–1 (circles) subunit chimeras coexpressed with the wild-type α₆ and γ₂ subunits in X. laevis oocytes. For clarity, data for the extremely weak agonist activity of etomidate at the chimeric β_2–1 subunit-containing receptor are omitted. (B) Graph showing the concentration-effect relationship for the GABA-modulatory (solid symbols) and GABA-mimetic (open symbols) of etomidate at the β subunit point mutants β_{1 S290N} (triangles) or β_{3 N289S} (circles) coexpressed with the wild-type α₆ and γ₂ subunits. Data are expressed relative to the maximal response elicited by a saturating concentration of GABA. Each point represents the mean data determined from 2 to 5 oocytes and the associated SEM where appropriate and exceeding the size of the symbol. The curves, fitted by eye, have no theoretical significance.

In comparison to GABA receptors containing the β_1–2 chimera, those presenting the N-terminal domain of the β₂ subunit (i.e., β_2–1) demonstrated reductions in both the potency (EC₅₀ = 7.7 ± 1 μM) and maximal effect (E_max = 23 ± 2%, n = 3) of the GABA-enhancing effects of etomidate (Fig. 3A). These values most closely approximate those found for β₁ subunit containing receptors (Table 1). Similarly, the parameters associated with the very weak agonist action of etomidate (Table 1) resemble those obtained when the β₁ subunit is present within the hetero-oligomeric complex. Collectively, these observations suggest that a key region that confers the enhanced activity of etomidate is located distal to the N-terminal extracellular portion of the receptor protein. For loreclezole, a single amino acid located within the M2 region of the β subunit (an asparagine for β₂ and β₃ and a serine for β₁) dictates the differential interaction of this anticonvulsant with the GABA_A receptor (18). Similarly, for oocytes expressing α₆, γ_2L and a β₁ subunit in which this serine was changed to an asparagine residue (β_{1 S290N}), both the potency (EC₅₀ = 1.6 ± 0.3 μM) and the maximal effect (E_max = 150 ± 13%, n = 5) of etomidate was increased relative to receptors (α₆ β₁ γ_2L) expressing the wild-type β₁ subunit (Fig. 3B, Table 1). Furthermore, in contrast to receptors containing the wild-type β₁ subunit, α₆β_{1 S290N} γ_2L receptors exhibited a clear (EC₅₀ = 79 ± 6 μM, E_max = 45 ± 13%, n = 5) agonist effect (Fig. 3B). By contrast, the reciprocal mutation of the β₃ subunit (β_{3 N289S}) reduced both the GABA modulatory (EC₅₀ = 5.7 ± 2.6 μM; E_max = 36 ± 1%; n = 4) and GABA-mimetic (EC₅₀ = 55 ± 6 μM; E_max = 14 ± 3%; n = 4) effects of this anesthetic (Fig. 3B, Table 1). A GABA receptor isolated from Drosophila, which is etomidate-insensitive, possesses a methionine residue at a position homologous to the β₂/β₃ subunit asparagine (24). The coexpression of α₆ and γ₂ subunits with a β₃ subunit in which Asn-289 is mutated to methionine results in a receptor that is essentially insensitive to both the GABA-modulatory and GABA-mimetic actions of etomidate (Fig. 4 A and B). By contrast, the GABA-enhancing action of the neurosteroid 5α-pregnan-3α-ol-20-one (1 μM) is not impaired by this mutation (Fig. 4C).

Figure 4.

Suppression of the modulatory and agonist actions of etomidate by a site-directed mutation of the β₃ subunit. (A) Traces illustrating the potentiation by a maximally effective concentration of etomidate (10 μM) of the inward current response evoked by GABA at EC₁₀ at receptors composed of wild-type α₆ β₃ and γ₂ subunits expressed in X. laevis oocytes. Note the prominent inward current evoked by the pre-application of etomidate. (B) Traces illustrating the virtual absence of effect of a high concentration of etomidate (100 μM) of the inward current response evoked by GABA at EC₁₀ at receptors composed of wild-type α₆, γ₂, and mutant β_{3 N289M} subunits. (C) 5α-Pregnan-3α-ol-20-one (1 μM) retains GABA-modulatory activity at receptors composed of wild-type α₆, γ₂, and mutant β_{3 N289M} subunits.

Given the intramembrane locus of the asparagine residue, it was of interest to determine whether intracellular etomidate influences the GABA_A receptor. We assessed this possibility by using L(tk−) cells (expressing α₆β₃γ_2S receptors) and patch pipettes that contained a concentration (10 μM) of etomidate that would be expected to maximally potentiate responses to GABA should, indeed, the modulatory site be accessible from the cell interior. However, responses to 3 μM pressure-applied GABA recorded within 20 sec of establishing a whole-cell recording remained constant over 10 min. During this time, the intracellular concentration of etomidate would be expected to rise with a calculated average time constant of 96 sec to 10 μM. In addition, the holding current did not change during dialysis, suggesting that intracellularly applied etomidate cannot exert a GABA-mimetic action. Moreover, when acting upon control or etomidate-loaded cells, the enhancement of GABA-evoked responses produced by 1 μM of extracellularly applied etomidate was very similar, amounting to 211 ± 29% (n = 5) and 184 ± 14% (n = 6) of control, respectively. A clear GABA-mimetic activity of 1 μM extracellular etomidate persisted in the cells loaded with the compound at a 10-fold molar excess. Collectively these observations suggest that intracellularly applied etomidate cannot access the site(s) mediating the modulatory and agonist actions of this agent (Fig. 5).

Figure 5.

Intracellularly applied etomidate does not affect the GABA_A receptor. Traces depicting inward current responses elicited by the repetitive local application of 3 μM GABA to L(tk−) cells stably expressing the α₆β₃γ_2S subunit combination. Responses were recorded under whole-cell voltage clamp at holding potential of −40 mV from cells dialyzed with the standard pipette solution (A) or pipette solution supplemented with 10 μM etomidate (B). Approximately 10 min after establishing the whole-cell recording, the cells were challenged with 1 μM of bath-applied etomidate. In either instance, etomidate caused a clear potentiation of the GABA-evoked current and directly elicited an inward current response.

DISCUSSION

Synaptically located transmitter-gated ion channels are now considered prime targets of general anesthetic action (2). Some general anesthetics are highly selective for particular members of this receptor superfamily. For example, anesthetic steroids such as alphaxalone enhance the function of the GABA_A receptor but, at clinically relevant concentrations, are inert at glycine, nicotinic, and glutamate receptors (25). By contrast, pentobarbitone enhances GABA_A receptor function at concentrations that also inhibit glutamate (non-N-methyl-d-aspartate) receptors and it is conceivable that for this anesthetic these effects act in concert to produce central depression (25). Our preliminary studies suggest that etomidate is highly selective for the GABA_A receptor (12), having little or no effect upon neuronal nicotinic (α₄β₂), 5-HT₃ (h5-HT₃R-A_s), glycine (α₁β), or kainate receptors (rat brain mRNA) expressed in Xenopus oocytes (ref. 26 and unpublished observations).

The interaction of etomidate with the GABA_A receptor is dependent upon subunit composition. In particular, both the GABA-modulatory and GABA-mimetic effects of this anesthetic are favored by receptors containing either a β₂ or a β₃ subunit versus a β₁ subunit (12, 13). Sanna and colleagues (13) have shown that receptors assembled only from α₁ and γ₂ subunits are not activated by etomidate. However, somewhat surprisingly, GABA-evoked currents recorded from this construct could be potentiated by the anesthetic. Nonetheless, the β_2/3 preference of etomidate is not shared by the neurosteroids (13, 15), barbiturates (12), or the general anesthetic propofol (12). Hence, in vivo, it is likely this anesthetic will distinguish between GABA_A receptor isoforms. Given the distinctive distribution of GABA_A receptor subunits within the central nervous system (27), it is probable that the actions of this anesthetic may be region or, indeed, neuron selective. A precedent is given by the anticonvulsant loreclezole, a closely related analogue of etomidate. Loreclezole has been shown to enhance the GABA_A receptor-mediated currents in only 50% of rat hippocampal dentate granule cells (28). Given the β isoform selectivity of loreclezole (23), it might be postulated that β₁ subunits are present upon loreclezole-insensitive neurons, whereas sensitive cells may express β₂ or β₃ subunits. Indeed, mRNAs for the three β isoforms are evident in dentate granule cells (27). Clearly, it would be of interest to determine the actions of etomidate on these cells, together with the technique of single-cell reverse transcription–PCR (29) to confirm the β subunit “fingerprint” of the neuron tested. In this study, all cerebellar granule cells tested gave a large direct inward current to locally applied etomidate (30 μM) suggesting the expression of GABA_A receptors containing β₂ or β₃ subunits. These subunit mRNAs are well represented in cerebellar granule cells, whereas the β₁ subunit mRNA is limited (27).

Whether or not etomidate and loreclezole mediate their effects through a common binding site or transduction mechanism, it is clear from Fig. 1 that the former has a greater apparent efficacy, both as a modulator and mimetic of GABA. Recently, the positive allosteric effects of the β-carboline methyl-6,7-dimethoxy-4-ethyl-carboline have also been shown to be dependent upon the β subunit isoform (19). Furthermore, it has been suggested that a number of benzodiazepines and β-carbolines may influence GABA_A receptor-mediated responses through this mechanism (19, 30). This action is in addition to the well-established effect of these drugs at the benzodiazepine recognition site, which is highly dependent upon the nature of the α- and γ subunits (31). Hence, this domain of the β subunit may be an important locus of drug action, and drugs that act through this site may differ both in their potency and efficacy.

The molecular determinant(s) of the β subunit selectivity of etomidate was initially investigated using subunit chimeras. The results obtained with the chimeric β_1–2 and β_2–1 subunits demonstrate that structural elements that reside within the transmembrane domains, connecting loops, or carboxyl-terminal region of the β subunit are important for both the GABA-modulatory and GABA-mimetic effects of etomidate. The GABA-modulatory effects of loreclezole exhibit a similar profile (18). For this anticonvulsant, site-directed mutagenesis studies have highlighted the importance of an asparagine residue (β₂ and β₃) versus a serine residue (β₁) located toward the carboxyl-terminal end of the second transmembrane domain, the putative pore-forming region of the chloride ion channel (18). This amino acid also plays a crucial role in the actions of etomidate. Hence, exchange of the equivalent serine residue of the β₁ subunit enhanced both the GABA-modulatory and GABA-mimetic effects of the anesthetic, whereas the reciprocal exchange of the asparagine-to-serine residue of the β₃ subunit reduced these actions. We have previously shown that the selective interaction of etomidate with β₂- or β₃-containing receptors occurs irrespective of the α subtype present (α_1,2,3,6) (12). Therefore, although not investigated here, it is possible that this β subunit amino acid similarly influences the interaction of etomidate with GABA_A receptor isoforms containing other α (α_1,2,3) subunits.

We have previously demonstrated that the function of a GABA-gated recombinant receptor isolated from Drosophila melanogaster (Rdl) is enhanced by both pentobarbitone and propofol, but is insensitive to etomidate (20, 24). Alignment of the M2 region of this invertebrate receptor with the mammalian β subunits reveals a methionine residue at the equivalent position to the serine (β₁) and asparagine (β₂ and β₃). Intriguingly, when the β₃ subunit asparagine residue was mutated to a methionine (β_{3 N289M}) and coexpressed with α₆ and γ_2L subunits, the resultant receptor was, like the invertebrate receptor, insensitive to both the modulatory and direct effects of etomidate. We are presently investigating whether mutation of the methionine residue of the Rdl receptor to an asparagine may impart etomidate sensitivity to this invertebrate receptor.

The well-documented GABA-modulatory and GABA-mimetic actions of a number of intravenous anesthetics are probably mediated through distinct binding sites. This premise is based primarily on the distinct concentration dependence of these effects, together with the observation that for certain recombinant GABA receptors, the anesthetics demonstrate only the modulatory actions (20, 24) or preferentially exhibit the mimetic actions (10). Here, a single point mutation disrupts both actions of etomidate. A speculative explanation for this observation postulates that occupation of a high-affinity modulatory site facilitates the binding of the anesthetic to a lower affinity “GABA-mimetic” site. Alternatively, the asparagine residue may permit the anesthetic-induced allosteric modification of the GABA_A receptor to be transduced, a suggestion compatible with the ion channel location of this amino acid.

Etomidate is optically active. The interaction of this anesthetic with the GABA_A receptor (as assessed by radioligand binding and ³⁶Cl⁻ efflux from rodent brain slices) exhibits a clear preference for the R-(+) isomer (32). This stereoselectivity, together with the observation that a single amino acid can dramatically influence the interaction of etomidate with the GABA_A receptor, would appear to preclude a membrane locus for this anesthetic. However, given the location of this asparagine residue within the transmembrane ion channel pore sequence (M2), this assumption may be unsound. By mutating amino acids in the M2 region to cysteine residues and determining their subsequent susceptibility to covalent blockade by charged sulfhydryl reagents, inferences have been made concerning the amino acids that face the ion channel lumen (33). This analysis suggests that for the GABA α₁ subunit the serine residue that occupies the equivalent position to the asparagine residue for β₂ and β₃ subunits does not project into the channel lumen, but faces either the interior of the protein or the lipid bilayer. Should the orientation of the β subunit M2 region be similar, etomidate could feasibly access this asparagine residue via the membrane. However, relatively high concentrations of etomidate when applied intracellularly to L cells expressing α₆ β₃ γ_2S GABA_A receptors did not influence the GABA-modulatory or GABA-mimetic actions of etomidate. Therefore, either intracellular etomidate cannot access this site or the asparagine residue is necessary for the anesthetic to modify the transduction process. Interestingly, for the β₁ subunit, the exchange of an M2 located threonine to a glutamine residue (situated three amino acids amino terminal to the asparagine, i.e., approximately one helical turn distant) abolishes the GABA-modulatory and GABA-mimetic effects of pentobarbitone (34). By contrast, this amino acid exchange for the α₁ subunit is inert in this respect (34). Hence, clearly this region of the β subunit ion channel plays a crucial role in the interaction of certain (i.e., pentobarbitone and etomidate) intravenous general anesthetics with the GABA_A receptor.

Over 150 years since the first clinical use of general anesthetics, their molecular mechanism remains to be elucidated. During the last decade, the focus of anesthetic research has shifted from the membrane lipid to synaptic receptor proteins, and the GABA_A receptor in particular. For etomidate this interaction is greatly influenced by a single amino acid. In the future this should permit the generation of mice carrying this specific mutation and allow the role of the GABA_A receptor in the behavioral actions of this clinically utilized general anesthetic to be determined unequivocally.

ABBREVIATION

GABA_A

γ-aminobutyric acid type A

Note Added in Proof

An amino acid residue homologous to that investigated here has been implicated in the positive allosteric modulation of both GABA_A and strychnine-sensitive glycine receptors by ethanol and enflurane (35).