Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The -aminobutyric acid type A (GABA_A) receptor has been shown to be a target of numerous depressant drugs, including benzodiazepines and general anesthetics (for review, see Carlson et al., 1997). At clinically relevant concentrations, all general anesthetics except ketamine enhance GABA_A receptor function in a reversible and stereospecific manner (Hales and Olsen, 1994). These findings suggest that the depressant behavioral effects of anesthetics are closely related to their actions on GABA_A receptors. The receptor domains pertinent for the actions of general anesthetics, however, have yet to be fully elucidated.

The GABA_A receptor is a ligand-gated Cl ion channel that belongs to a family of subunits (_1-6, _1-4, _1-3, , , and _1-3, in the mammalian central nervous system) that forms pentameric complexes (Amin and Weiss, 1996; Chang et al., 1996; Davies et al., 1997a; Tretter et al., 1997). Beside the -homomeric channels that are located primarily in the retina (Cutting et al., 1991), the proposed stoichiometry of native GABA_A receptors in the brain is believed to contain two  subunits, two  subunits, and one  or one , or one  subunit (Chang et al., 1996; Davies et al., 1997a; Tretter et al., 1997). The positive modulating actions of benzodiazepines on GABA_A receptors have been determined to depend on specific amino acids on both the  and the  subunits (Pritchett and Seeburg, 1991; Wieland et al., 1992; Buhr et al., 1996; Amin et al., 1997). Whereas the actions of general anesthetics appear not to depend on the presence of the  subunit (Jones et al., 1995), the  subunit has been shown to play an important role in the allosteric modulation of GABA_A receptors by i.v. anesthetics (Harris et al., 1995; Zezula et al., 1996). To date, only the  (in certain expression systems) and  subunits have been shown to confer insensitivity to i.v. anesthetics (Mihic and Harris, 1996; Davies et al., 1997a). With respect to the ₁ subunit, its anesthetic-distinct pharmacology has promoted the identification of specific residues in the transmembrane (TM) regions, specifically TM2 and TM3 of the glycine and GABA_A receptors, that harbor sites necessary for the positive allosteric modulation and direct activation induced by volatile and i.v. anesthetics (Belelli et al., 1997; Mihic et al., 1997; Moody et al., 1997; Amin, 1999).

To date, the body of evidence identifying structural determinants for anesthetic action on the GABA_A receptor has focused solely in the TM2 and TM3 regions. In this study, the emphasis was placed on amino acids from other than the TM2 and TM3 regions, one at the entrance of TM1 and the others in the extracellular region between TM2 and TM3 (Fig. 1). Except for the amino acids indicated, Fig. 1 illustrates that these two areas of interest reflect high homology between anesthetic-sensitive proteins and the anesthetic-insensitive ₁ subunit. Furthermore, the recent observation that lipid-water interfaces of membrane ion channels may be important sites of action for anesthetics supports the investigative interest in amino acids located near lipid-water interfaces as shown in Fig. 1 (Xu et al., 1998). With site-directed mutagenesis, a phenylalanine residue located at the entrance to TM1 (position 261, human ₁ subunit) has been transformed to a glycine, which is the homologous amino acid on _1-6, _1-3, _1-3, , and  GABA_A receptor subunits and on the ₁ subunit of the glycine receptor. The reciprocal point mutations were also made in the ₁ and ₂ subunits, G223F and G219F, respectively, to test the hypothesis that sensitivity to i.v. anesthetics would be diminished in GABA_A receptors on conversion of the conserved glycine to the ₁-residue phenylalanine. Anesthetic-induced enhancement of [³H]muscimol and [³H]flunitrazepam binding has shown that GABA_A receptors containing the ₂(G219F) mutation displayed a reduced efficacy in anesthetic modulation by all four of the i.v. anesthetics tested. Consistent with the binding data, functional analysis of these mutant receptors with whole-cell patch clamp demonstrated that enhancement of GABA currents by pentobarbital and propofol was also hindered in the presence of the ₂(G219F) point mutation. On the contrary, the four amino acids in the TM2/TM3 bridge were determined not to be essential for anesthetic modulation. This study identifies a new region of TM1 involved in channel gating and anesthetic modulation.

View larger version (30K): [in this window] [in a new window]   Fig. 1.   Amino acid sequence alignment of the pre-TM1/TM1 (A) and the inter-TM2/TM3 (B) domains for mammalian GABAA and 1 receptor subunits and the glycine receptor 1 subunit (gl1). A, this alignment shows that the pre-TM1 glycine (G) is conserved in all of the subunits aligned except the 1 (#), which has a phenylalanine (F). In this domain of 22 residues, there are five amino acids that are conserved in all the subunits listed including 1 (*) and seven amino acids that are conserved within subunit families (). B, this sequence of amino acids represents the extracellular region that bridges the TM2 and TM3 domains. Within a subunit family, there are six amino acids that are conserved. One amino acid, a proline (P), is conserved among all the subunits listed (*); # indicates the arginine (R), which is present in all of the subunits, except 1, which has an asparagine (N). Bolded residues are highly conserved within a subunit family and are the targeted four amino acids in the 1 subunit (bolded and underlined) that have been mutated to the GABAA receptor 1 subunit sequence (RNSL). In both A and B, bolded residues represent potentially relevant sites of action for i.v. anesthetics. Note that the glycine receptor 1 subunit is primarily insensitive to most i.v. anesthetics, except for propofol (Hales and Lambert, 1991). The source for all sequences was GenBank, apart from the glycine receptor 1 subunit (Grenningloh et al., 1990), GABAA receptor  subunit (Davies et al., 1997a), and 1 subunit (Cutting et al., 1991).

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Site-Directed Mutagenesis and Generation of Recombinant Baculoviruses. The mutations were introduced into the cDNAs of the GABA_A receptors subunits with the Altered Sites II in vitro mutagenesis systems (Promega, Madison, WI). Briefly, the entire coding region of the human ₁ subunit was subcloned into pAlter plasmid (the same was performed for the rat ₁ and ₂ subunits), and a mutagenic oligonucleotide was used to incorporate the desired mutation according to the manufacturer's suggestions. The oligonucleotides used were: ₁ NASM(311-314)-RNSL, 5'-GTCCACCATCA- TCACGGGCGTGAGAAACTCCCTGCCGCGCGTCTCCTACATC-3'; ₁(F261G), 5'-TTGCGTCGCCACATCGGCTTCTTCTTGCTCCAA-3'; ₁(G223F), 5'-TTGAATAACAAAGTAGAAGATTTTTCTCTTCAA-3'; and ₂(G219F), 5'-CAGGATGAAGTAGAAAATGTTTCTTTTCAG-3'.

Cell Culture and Baculovirus Infection. Spodoptera frugiperda 9 (Sf9) cells (Life Technologies, Paisley, Scotland) were grown in serum-free medium (Sf900 II medium; Life Technologies) as a shaking culture (140 rpm) at 27°C. At a density of 2 × 10⁶ cells/ml, Sf9 cells were infected with a single or various combination(s) of the following recombinant Autographa californica nuclear polyhedrosis viruses (AcNPV), coding for wild-type and point-mutant GABA_A receptors: AcNPV-₁, AcNPV-₁(G223F), AcNPV-₂, AcNPV-₂(G219F), AcNPV-₂, AcNPV-₁, AcNPV-₁(F261G), and AcNPV-₁(NASM-RNSL). The amount of recombinant baculovirus added for infection was determined by maintaining a cumulative multiplicity of infection 15. Titers of recombinant baculovirus ranged from 8 × 10⁷ to 8 × 10⁸ plaque-forming units/ml. Virus titer was determined with a plaque assay according to protocol from the Life Technologies Instruction Manual, "Guide to Baculovirus Expression Vector Systems (BEVS) and Insect Cell Culture Techniques." The wild-type viruses AcNPV₁, AcNPV₂, and AcNPV₂ were gifts from Dr. D. Gallager (Neurogen, Branford, CT).

Sf9 Cell Membrane Preparation. At 65 h after infection, Sf9 cells were harvested by centrifugation at 1500g for 5 min. The pelleted cells were subsequently resuspended in 20 mM KH₂PO₄/K₂HPO₄ and 50 mM KCl (pH 7.4) buffer and pelleted by centrifugation at 1500g for 5 min. The buffer was aspirated, and the pellet was homogenized in 10 mM KH₂PO₄/K₂HPO₄, 100 mM KCl (pH 7.4) binding buffer, using an Ultra-Turrax T-25 homogenizer (Janke & Kunkel, Staufen, Germany) at 12,000 rpm for 20 s. The homogenate was centrifuged for 20 min at 20,000g at 4°C. This washing/centrifugation procedure was repeated twice. Sf9 membrane pellets were stored at 80°C until use. Before binding, the pellet was resuspended in 10 mM KH₂PO₄/K₂HPO₄, 100 mM KCl binding buffer, using an Ultra-Turrax T-25 homogenizer (Janke & Kunkel) at 12,000 rpm for 30 s.

Binding Assays. Binding of [³H]muscimol (19.1-20.0 Ci/mmol; New England Nuclear, Boston, MA) or [³H]flunitrazepam (84.5 Ci/mmol; New England Nuclear) was determined in a total volume of 0.25 ml consisting of 0.15 ml of Sf9 membrane-bound proteins (300-800 mg/ml), 0.075 ml of binding buffer, and 0.025 ml of radioactive ligand. Nonspecific binding was measured by adding GABA (final concentration, 100 µM) or diazepam (final concentration, 10 µM) in the presence of radioactive ligand. Anesthetic-induced enhancement of radioligand binding was determined by adding each anesthetic solution in the presence of radioactive ligand. Anesthetic concentration-response assays were performed with either 3 nM [³H]muscimol for -containing Sf9 membranes or 40 nM [³H]muscimol for -containing membranes. For assessing anesthetic-modulated benzodiazepine binding in -containing Sf9 membranes, a final concentration of 1 nM [³H]flunitrazepam was used. All concentrations used to study anesthetic modulation were below the K_D values calculated from competition assays.

Experimental Design for Binding Assays. Ten concentrations of an anesthetic or cold ligand were used to construct the concentration-response curves for determining EC₅₀ and E_max values or competition curves for determining K_I values, respectively. For each curve, total radioligand binding in the absence of anesthetic or cold ligand (control), total radioligand binding in the presence of the anesthetic (concentration-response curves) or cold ligand (competition curves), and nonspecific binding were measured in triplicate. Control, anesthetic-treated (or cold ligand-treated), and nonspecific binding groups were assayed in the same experiment, for a total of 36 tubes per receptor combination. Each anesthetic assay of 36 tubes was repeated three or four times for each of the 11 different receptor combinations. Competition assays were repeated at least twice for each of the receptor combinations.

Electrophysiology. Recordings were made from Sf9 cells that had been incubated with virus for 27 to 29 h. A Petri dish containing the cells was transferred to the recording chamber on the stage of a Zeiss (Oberkochen, Germany) Axiovert-10 inverted phase-contrast microscope, where the individual cells were viewed at 200× magnification. The recording chamber contained 2 to 3 ml of artificial balanced salt solution (ABSS) that was renewed by constant perfusion at 0.5 ml/min¹ at room temperature (20-22°C). The composition of ABSS was 162.5 mM NaCl, 3.5 mM KCl, 1.25 mM Na₂HPO₄, 2 mM MgSO₄, 2 mM CaCl₂, 10 mM glucose, and 10 mM HEPES, pH 7.35, at 22°C. Standard patch clamp techniques (Hamill et al., 1981) were used to record from the infected cells in the whole-cell configuration with an EPC-9 amplifier (HEKA Electronik, Lambrecht, Germany). The patch electrodes were manufactured from 1.5-mm o.d. glass (World Precision Instruments, Sarasota, FL). The patch electrodes were pulled just before use with a BB-CH-PC microelectrode puller (Mecanex, Nyon, Switzerland) and were filled with a solution containing 160 mM KCl, 1 mM MgCl₂, 1 mM CaCl₂, 10 mM EGTA, 2 mM Mg-ATP, and 10 mM HEPES, pH 7.3, at 20°C. Electrodes had resistances of 2 to 5 M. A holding potential of 40 mV was used. Series resistance was 60% compensated. Whole-cell currents were plotted on a low-fidelity chart recorder during the experiment. The signals were stored on computer and also recorded on a video recorder with a VR-100 digital data recorder (Instrutech Corporation, Elmont, NY). Results were analyzed with Pulse (HEKA) and Igor Pro (Wavemetrics, Lake Oswego, OR) software.

Data Analysis. Curve fitting via nonlinear regression analyses of binding data was used to determine EC₅₀, E_max, and K_I values (GraphPad Software). Statistical analyses of EC₅₀, E_max, and K_I values from the binding data were performed using one-way ANOVA, evaluated at a criterion of P < .01. Pairwise multiple comparisons, using the Tukey test, were calculated from the mean and S.E. values generated from one-way ANOVA and evaluated at P < .05. Additional pairwise comparisons were determined with Student's t test. Analysis of the electrophysiological data was performed with either Kruskal-Wallis ANOVA or Mann-Whitney test, with follow-up comparisons with Dunn's test. All statistical analyses were conducted with Jandel Statistical Software (SigmaStat version 2; San Rafael, CA).

Drugs. Stock solutions of GABA (10 mM; Riedel-deHaen, Seelze, Germany) and diazepam (1 mM; Sigma, St. Louis, MO) were diluted into binding buffer daily before use. The i.v. anesthetics used in this study were sodium pentobarbital (Sigma), alphaxalone (5-pregnan-3-ol-11,20-dione; Sigma), etomidate (org 24242; gift from Dr. Niall Hamilton, Akzo Nobel, Organon Labs Ltd., Lanarckshire, UK), and propofol (diprivan, 1% 2,6-diisopropylphenol in soy and animal lecithin as an aqueous emulsion, Stuart Pharmaceuticals, Inc., Wilmington, DE, or 2,6-diisopropylphenol, Tocris, Bristol, UK). Pentobarbital stock solutions were prepared in 50 mM Tris base, 120 mM NaCl, 5 mM KCl, pH 10. Alphaxalone and etomidate stock solutions were prepared in DMSO (Riedel-deHaen), and the maximum final concentration of DMSO was 0.1% (v/v), which was determined not to interfere with [³H]muscimol or [³H]flunitrazepam binding. All i.v. anesthetics were dissolved daily in binding buffer immediately before the experiment. With regard to the intralipid version of propofol in this study, the potentiating effects of propofol in this formulation has been shown not to differ from propofol made from an ethanol stock solution (Hales and Lambert, 1991).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Characterization of Wild-Type and Mutant GABA_A Receptors Expressed in Sf9 Cells. For the purposes of this study, the ₁₂₂ receptor is also identified by the IUPHAR nomenclature name A1a2 (Barnard et al., 1998). GABA_A receptors containing the ₂(G219F) subunit demonstrated a slightly higher affinity for [³H]muscimol binding. The rank order for [³H]muscimol affinity was ₁₂(G219F), ₁(G223F)₂(G219F), ₁₂, and ₁(G223F)₂ (K_I, 5.9 ± 1.83, 6.6 ± 2.26, 8.1 ± 0.71, and 8.8 ± 2.47 nM ± S.D., respectively). For receptors, the rank order for [³H]muscimol affinity was different from that for the receptors: ₁(G223F)₂(G219F)₂, ₁₂₂, ₁₂(G219F)₂, and ₁(G223F)₂₂ (K_I, 10.4 ± 0.85, 13.1 ± 2.68, 21.0 ± 1.4, and 29.3 ± 0.92 nM ± S.D., respectively). For the ₁ receptors, wild type, ₁(F261G), and ₁(NASM-RNSL) showed similar K_I values in muscimol competition assays (78.0 ± 4.2, 70.0 ± 7.1, and 71.0 ± 12.7 nM ± S.D., respectively). Flunitrazepam affinity was assessed by nonradioactive competition assays, and these assays showed comparable K_I values between the wild-type and mutant receptor combinations [K_I (nM ± S.D.) = ₁₂₂, 2.2 ± 0.14; ₁(G223F)₂₂, 2.25 ± 0.21; ₁₂(G219F)₂, 2.5 ± 0.0 nM; and ₁(G223F)₂(G219F)₂, 2.3 ± 0.14 nM]. One-way ANOVA on ranks determined that all the K_I values within each group tested, except the K_I values for [³H]muscimol between ₁(G223F)₂(G219F)₂ and ₁(G223F)₂₂, were not significantly statistically different from each other (P > .05).

Mutation of TM1 Glycine on the ₂ Subunit Alters Pentobarbital-Induced Modulation of Ligand Binding. GABA_A receptors containing the ₂(G219F) subunit displayed a decreased maximal effect in pentobarbital-induced potentiation of [³H]flunitrazepam and [³H]muscimol binding (Fig. 2; Table 1). The double-mutant receptors (i.e., ₁(G223F)₂(G219F)₂) demonstrated an intermediate pentobarbital-induced E_max compared with the GABA_A receptors containing the single point mutations. In addition, pentobarbital-induced enhancement of [³H]flunitrazepam binding was statistically greater with the ₁(G223F)₂₂ than with the wild-type ₁₂₂ (P < .05; Fig. 2). Although there was no difference in the potency of pentobarbital among the receptors (Fig. 2), pentobarbital was significantly more potent in the mutant receptors than in the wild-type ₁₂ receptors (P < .05; Table 1).

View larger version (24K): [in this window] [in a new window]   Fig. 2.   Concentration-response curves for pentobarbital-induced modulation of ligand binding in wild-type and mutant GABAA receptors expressed in Sf9 cells. All curves represent the percentage change of specific [3H]flunitrazepam binding (1 nM) in the presence of increasing concentrations of pentobarbital (n = 3 experiments performed in triplicate for each receptor combination). One-way ANOVA (P < .001) determined that there were significant differences in the Emax values between the receptors. Both 12(G219F)2 (, 24.7 ± 3.93) and 1(G223F)2(G219F)2 (, 28.7 ± 0.67) Emax values were statistically less than 122 (, 62.7 ± 3.18) and 1(G223F)22 (, 79.0 ± 4.93) Emax values (Tukey test, P < .05). The EC50 values for pentobarbital in the receptors were not statistically different from each other (one-way ANOVA, P = .054; 122, , 120.3 ± 8.65 µM; 1(G223F)22, , 104.3 ± 5.18 µM; 12(G219F)2, , 100.9 ± 6.09 µM; 1(G223F)2(G219F)2, , 152.3 ± 20.51 µM). All values are mean ± S.E.

The Effect of the ₂ Mutant on Alphaxalone-Induced Modulation of Ligand Binding Is Altered in the Presence of the ₂ Subunit. Compared with ₁₂₂, GABA_A receptors containing the ₂(G219F) mutant subunit did not demonstrate any significant changes in the maximal effect induced by alphaxalone (P > .05; Fig. 3). Furthermore, alphaxalone was equipotent for all receptor combinations (Fig. 3). In the combinations, however, alphaxalone induced negative modulation of [³H]muscimol binding in the ₁₂(G219F) receptor, and alphaxalone was more potent in the mutant receptors than in the wild-type ₁₂ receptors (P < .05; Table 1).

View larger version (26K): [in this window] [in a new window]   Fig. 3.   Concentration-response curves for alphaxalone-induced modulation of ligand binding in wild-type and mutant GABAA receptors expressed in Sf9 cells. All curves represent the percentage change of specific [3H]flunitrazepam binding (1 nM) in the presence of increasing concentrations of alphaxalone (n = 3 experiments performed in triplicate for each receptor combination). The rank order for Emax values of combinations is as follows: 1(G223F)22, 37.7 ± 4.41 (); 122, 34.3 ± 1.20 (); 1(G223F)2(G219F)2, 30.0 ± 3.22 (); and 12(G219F)2, 22.7 ± 1.85 (). One-way ANOVA and follow-up Tukey test determined that there was a significant difference (P = .033 and P < .05, respectively) between the mean Emax value of 12(G219F)2 and the mean Emax value calculated from the 1(G223F)22 curve. The EC50 values for alphaxalone in the receptors were not statistically different from each other (one-way ANOVA, P = .80; 122, , 0.97 ± 0.37 µM; 1(G223F)22, , 0.67 ± 0.19 µM; 12(G219F)2, , 0.76 ± 0.14 µM; 1(G223F)2(G219F)2, , 0.83 ± 0.03 µM). All values are mean ± S.E.

Mutation of TM1 Glycine on the ₁ and ₂ Subunits Alters Etomidate-Induced Modulation of Ligand Binding. GABA_A receptors containing the ₁(G223F) and/or the ₂(G219F) subunit(s) displayed a significant decrease in the maximal effect of etomidate-induced potentiation of [³H]flunitrazepam and [³H]muscimol binding (Fig. 4; Table 1). The efficacies (i.e., E_max) of etomidate on ₁(G223F)₂₂, ₁₂(G219F)₂, and ₁(G223F)₂(G219F)₂ receptors were all approximately half of that measured with wild-type ₁₂₂ receptors (Fig. 4). For the receptors, etomidate was the most potent with the ₁(G223F)₂(G219F)₂ combination and the least potent with the ₁₂(G219F)₂ receptor (Fig. 4). For the receptors, etomidate was significantly more potent at receptors containing the ₁(G223F) subunit than the ₁₂ or ₁₂(G219F) receptors (P < .05; Table 1). The double-mutant receptor ₁(G223F)₂(G219F) demonstrated an intermediate etomidate-induced enhancement of [³H]muscimol binding compared with the ₁(G223F)₂ and ₁₂(G219F) receptor combinations (Table 1).

View larger version (23K): [in this window] [in a new window]   Fig. 4.   Concentration-response curves for etomidate-induced modulation of ligand binding in wild-type and mutant GABAA receptors expressed in Sf9 cells. All curves represent the percentage change of [3H]flunitrazepam binding (1 nM) in presence of increasing concentrations of etomidate (n = 3 experiments performed in triplicate for each receptor combination). One-way ANOVA and the follow-up Tukey test (P < .001 and P < .05, respectively) determined that there were significant differences between the Emax values of 1(G223F)22 (, 22.7 ± 1.20), 12(G219F)2 (, 22.3 ± 1.20), and 1(G223F)2(G219F)2 (, 27.3 ± 2.91) and the Emax value of 122 (, 55.0 ± 1.53). One-way ANOVA (P < .001) determined that there were significant differences in the EC50 values between the receptors. Both 12(G219F)2 (, 4.63 ± 0.37 µM) and 1(G223F)2(G219F)2 (, 0.48 ± 0.22 µM) EC50 values were statistically different from each other as well as from 122 (, 2.69 ± 0.37 µM) and 1(G223F)22 (, 2.00 ± 0.20 µM) EC50 values (Tukey test, P < .05). All values are mean ± S.E.

Mutation of TM1 Glycine on the ₁ and ₂ Subunits Alters Propofol-Induced Modulation of Ligand Binding. Similarly to etomidate, GABA_A receptors containing either or both point mutations, i.e., ₁(G223F) and/or ₂(G219F), demonstrated a diminished maximal effect of propofol-induced enhancement of [³H]flunitrazepam and [³H]muscimol binding (Fig. 5; Table 1). For the receptors, propofol was significantly less efficacious for single- and double-mutant GABA_A receptors containing the ₂(G219F) subunit compared with the ₁₂₂ and ₁(G223F)₂₂ receptor combinations (P < .05; Fig. 5). In addition, the presence of either or both point mutations in an receptor significantly increased the potency of propofol compared with the wild-type complex (P < .05; Fig. 5). With receptors, propofol was the least efficacious and least potent at the double-mutant receptor (Table 1). Although the ₁₂(G219F) receptor displayed a 4- to 5-fold decrease in the efficacy of propofol modulation, propofol had a higher potency for this receptor than the wild-type receptor ₁₂ (P < .05; Table 1).

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Concentration-response curves for propofol-induced modulation of ligand binding in wild-type and mutant GABAA receptors expressed in Sf9 cells. All curves represent the percentage change of specific [3H]flunitrazepam binding (1 nM) in the presence of increasing concentrations of propofol (n = 3 experiments performed in triplicate for each receptor combination). One-way ANOVA and the follow-up Tukey test (P < .001 and P < .05, respectively) determined that there were significant differences between the Emax values of 1(G223F)22 (, 51.0 ± 5.20), 12(G219F)2 (, 21.0 ± 3.79), and 1(G223F)2(G219F)2 (, 22.0 ± 3.61) and the Emax value of 122 (, 78.7 ± 8.17). In addition, the Emax value for 1(G223F)22 was evaluated to be significantly different from the Emax values of 12(G219F)2 and 1(G223F)2(G219F)2 (P < .05). One-way ANOVA and the follow-up Tukey test (P < .001 and P < .05, respectively) determined that there were significant differences in the EC50 values of 1(G223F)22 (, 14.3 ± 2.38 µM), 12(G219F)2 (, 13.0 ± 1.95 µM), and 1(G223F)2(G219F)2 (, 6.8 ± 1.40 µM) compared with the EC50 value of 122 (, 26.9 ± 2.13 µM). All values are mean ± S.E.

Point Mutations on ₁ receptors Alter Anesthetic Insensitivity. The TM1 mutation in ₁(F261G) receptors manifested a pentobarbital-induced and alphaxalone-induced inhibition of [³H]muscimol binding (Table 1). This inhibition was found to be significantly different from wild-type GABA_A receptors (i.e., ₁, ₁₂, and _122; P < .05). Pentobarbital was significantly less potent in modulating ₁(F261G) receptors than wild-type receptors ₁₂ and ₁₂₂ (P < .05; Table 1); however, alphaxalone was equally potent in modulating ₁(F261G) receptors as seen with ₁₂ and ₁₂₂ receptors. In addition to the ₁(F261G) mutant, ₁(NASM-RNSL) homomers demonstrated a significant alpaxalone-induced inhibition of [³H]muscimol binding compared with ₁, ₁₂, and ₁₂₂ receptor combinations (P < .05; Table 1). The potency of alphaxalone in modulating ₁(NASM-RNSL) receptors also was not statistically different from ₁₂ and ₁₂₂ receptors (P > .05). In the presence of alphaxalone, wild-type ₁ receptors did not show any statistically significant changes in specific [³H]muscimol binding (P > .05; Table 1).

Functional Characterization of Wild-Type and Mutant Heteromeric and Homomeric GABA_A Receptors. From the GABA concentration-response curves (Fig. 6B),^, mutation of the ₁ subunit (G223F) did not significantly affect the concentration-response relation for GABA-induced peak currents. The corresponding mutation in the ₂ subunit (G219F), however, significantly decreased the EC₅₀ of the receptor for GABA (P < .001), suggesting an increase in affinity for GABA. The ₁₂(G219F)₂ combination had an EC₅₀ comparable with the homomeric ₁ receptor (not significantly different). The Hill coefficients determined for the four subunit combinations were not significantly different. To confirm the presence of the ₂ subunit in the receptor combinations tested, the effect of 1 µM diazepam on a GABA-induced current (20 µM GABA) was tested. The diazepam modulated peak current was 165 ± 27% greater than without applied benzodiazepine (n = 3, ₁₂₂). The effect of diazepam disappeared within 1 to 2 min. The ₁₂(G219F)₂ combination was also observed to be positively modulated by benzodiazepines (data not shown).

View larger version (20K): [in this window] [in a new window]   Fig. 6.   Concentration-response curves for GABA-induced peak currents. Top, representative current traces from the maximum GABA-induced response for each combination tested. The line above each current trace represents duration of GABA application. Desensitization in the presence of GABA is prominent in all currents, except for the 1 receptor. Bottom, results are shown as mean ± S.E. (n = 4-14 Sf9 cells tested/combination). The currents for each cell were normalized relative to the GABA-induced peak current of the highest concentration used for each subunit combination. The concentration-response relation for all four combinations shown were estimated by nonlinear regression with a logistic equation of the form: E = Emax × [GABA]n/(EC50n+ [GABA]n), where n is the Hill coefficient. Each curve was normalized to make Emax = 100%. The following values (EC50, µM; Hill coefficient, n) represent the nonlinear regression analyses of GABA-induced peak currents for wild-type and mutant GABAA receptors (122: 39 µM (29-49), 1.19 (0.98-1.40); 1(G223F)22: 30 µM (23-37), 1.27 (0.91-1.62); 12(G219F)2: 5.4 µM (4.3-6.5)*, 0.93 (0.78-2.01); 1: 8.8 µM (5.9-11.6), 1.14 (0.76-1.53). Numbers in brackets represent 95% CIs. *P < .001, significant difference between the EC50 of 12(G219F)2 compared with 122 and 1(G223F)22,.

7

Mutation of TM1 Glycine on ₂ Subunit Eliminates Pentobarbital-Induced Enhancement of GABA Currents. The effect of pentobarbital was tested both with and without a 10-s pretreatment of pentobarbital (Fig. 7). There was no statistically significant effect of pretreatment compared with no pretreatment (Fig. 7, filled columns). For the wild-type ₁₂₂ and ₁(G223F)₂₂ combinations, pentobarbital concentrations of 10 and 50 µM resulted in statistically significant concentration-dependent increases in the GABA-induced peak current, compared with control (10 µM: P < .05 with or without pretreatment for both receptor combinations; 50 µM: ₁₂₂, P < .01 with and P < .001 without pretreatment; ₁(G223F)₂₂, P < .01 with or without pretreatment; Kruskal-Wallis ANOVA and Dunn's tests). No significant differences were found between the wild-type receptor ₁₂₂ and the mutant ₁(G223F)₂₂ at either 10 or 50 µM pentobarbital. For the ₁₂(G219F)₂ combination, only a low concentration of 5 µM was tested, because higher concentrations of pentobarbital gave rise to relatively large currents (refer to direct activation curves, Fig. 8), making it impossible to determine the modulating effect on GABA currents. In the presence of the ₂(G219F) mutant, 5 µM pentobarbital did not enhance GABA currents. In addition, the wild-type ₁ receptor was not modulated by 50 µM pentobarbital.

View larger version (28K): [in this window] [in a new window]   Fig. 7.   Modulating effect of pentobarbital on GABA-induced peak currents. Top, representative current traces from each combination depicting the modulating effect of pentobarbital on currents induced by GABA (EC20). The increased current variation in the beginning of the pentobarbital-modulated traces is caused by transient voltage pulses used to monitor cell membrane conductance and capacitance. These transient pulses are suspended 5 s before agonist application. Bottom, peak currents were normalized for each cell to the response of GABA at approximately the EC20 for each receptor combination (122, 10 µM GABA; 1(G223F)22, 10 µM GABA; 12(G219F)2, 1 µM GABA; 1, 4 µM GABA) and are shown as a mean ± S.E. (n = 3-9 Sf9 cells tested/combination). *P < .05; **P < .01, significant increase in pretreated responses compared with control.

View larger version (20K): [in this window] [in a new window]   Fig. 8.   Modulating effect of propofol on GABA-induced peak currents. Top, representative current traces from the 122 and 12(G219F)2 combinations depicting the modulating effect of propofol on currents induced by GABA (EC20: 122, 10 µM GABA; 12(G219F)2, 1 µM GABA). Bottom, peak currents were normalized for each cell to the response of GABA at approximately the EC20 for each combination and are shown as mean ± S.E. (n = 6-10 Sf9 cells/combination). In all cases, the mean modulation was smaller for the 12(G219F)2 combination. *P < .05; **P < .01, significant concentration-dependent increases in pretreated responses compared with control.

Mutation of TM1 Glycine on ₂ Subunit Diminishes Propofol-Induced Enhancement of GABA Currents. The effect of propofol was tested both with and without a 10-s pretreatment of propofol (Fig. 8). There was no statistically significant effect of pretreatment compared with no pretreatment (Fig. 8, solid columns). For the wild-type ₁₂₂ and ₁₂(G219F)₂ combinations, propofol concentrations of 1 and 5 µM resulted in statistically significant concentration-dependent increases in the GABA-induced peak current compared with control (1 µM: P < .05 with pretreatment for both receptor combinations; 5 µM: P < .01 with pretreatment for both receptor combinations). In all cases, however, the mean modulation was smaller for the ₁₂(G219F)₂ combination, and the modulating effect of 5 µM propofol (without pretreatment) was statistically significantly smaller for the ₁₂(G219F)₂ combination compared with the ₁₂₂ combination (Mann-Whitney, P = .030). Note that concentrations of 1 and 5 µM propofol did not elicit any direct activation of GABA_A chloride currents in either receptor combination.

Mutation of TM1 Glycine on ₂ Subunit Alters Pentobarbital- and Propofol-Induced Direct Activation in the Absence of GABA. The ₁₂(G219F)₂ combination demonstrated a biphasic concentration-response curve and was significantly more sensitive than the ₁₂₂ combination to the direct effect of pentobarbital in the lower concentration range of 20 to 50 µM. No difference was detected between the direct effect of pentobarbital on the ₁₂₂ and the ₁(G223F)₂₂ combinations (Fig. 9). The concentration-response relation for the ₁(G223F)₂₂ combination could be fitted with a logistic equation of the form: E = E_max × [pentobarbital]ⁿ/[EC₅₀ⁿ + (pentobarbital)ⁿ], where n is the Hill coefficient (Fig. 9). The results for pentobarbital were as follows (95% CI in parentheses): E_max = 26% (20-32%), EC₅₀ = 47 µM (22-73 µM), n = 1.68 (0.39-2.97). For the ₁₂₂ combination, a Hill coefficient could not be estimated because of the lack of points in the middle (increasing) range of the curve. Instead, using a Hill coefficient of the same value as estimated for the ₁(G223F)₂₂ combination (n = 1.68), the corresponding E_max and EC₅₀ values were estimated as follows (95% CI in parentheses): E_max = 38% (27-50%) and EC₅₀ = 84 µM (24-145 µM). These values were not significantly different from the corresponding values for the ₁(G223F)₂₂ combination. For the ₁₂(G219F)₂ combination, the concentration-response curve was clearly biphasic, and at 500 µM pentobarbital, the peak current increased significantly relative to 200 µM (P < .001). Because of the lack of an asymptote for the second phase of the curve, the data were impossible to fit into the logistic equation. Rather, for display purposes, a splined curve is shown in Fig. 9.

View larger version (17K): [in this window] [in a new window]   Fig. 9.   Concentration-response curves for pentobarbital-induced direct activation. Peak currents were normalized to the peak current of a maximum GABA response in each Sf9 cell and are shown as the mean ± S.E. (n = 5-15 Sf9 cells tested/combination). The concentrations used to elicit maximum GABA responses were: 122, 2 mM; 1(G223F)22, 2 mM; and 12(G219F)2, 200 µM. No significant differences were found between the 122 combination () and the 1(G223F)22 combination () at any of the concentrations of pentobarbital tested, whereas the peak currents induced by pentobarbital differed significantly between the 12(G219F)2 combination () and the 122 combination at some concentrations. *P < .05; **P < .01; ***P < .001).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Mutation of TM1 glycine diminishes anesthetic efficacy in GABA_A receptor binding. Our data suggest that G219 on the rat ₂ subunit may be an important component for allosteric changes induced by i.v. anesthetics. Finding an essential residue on the ₂ subunit is consistent with other studies implicating the  subunit as an important subunit for anesthetic modulation. For example, -homomeric channels have been shown to be insensitive to GABA but can be directly activated by propofol and pentobarbital (Cestari et al., 1996; Davies et al., 1997b). More specifically, key amino acids in the TM2 domain, i.e., ₁(S265), ₂(N289), and ₃(N290), and in the TM3 domain, i.e., M286, of the ₁ subunit have been shown to be essential for the positive potentiating effects of volatile and i.v. anesthetics (Belelli et al., 1997; Mihic et al., 1997; Moody et al., 1997; Amin, 1999). However, the same point mutations, which were critical for the positive modulation of GABA_A and glycine receptors by volatile anesthetics (Mihic et al., 1997), appeared not to affect the potentiating actions of several i.v. anesthetics, such as methohexital (a barbiturate), alphaxalone, etomidate, and propofol (Krasowski et al., 1998). These findings indicate that the structural determinants for volatile and i.v. anesthetics are not necessarily the same.

Compared with wild-type receptors (i.e., ₁₂ and ₁₂₂), our binding data demonstrated that inclusion of the ₂(G219F) subunit in a recombinant or GABA_A receptor hindered positive modulation by all four anesthetics tested. However, as determined by the K_I values, binding affinities for [³H]muscimol (on receptors) or [³H]flunitrazepam (on receptors) were not altered. This result indicates that this point mutation most likely did not modify the binding sites for GABA/muscimol or benzodiazepines. Furthermore, on studying the effect of the ₂ mutant on anesthetic modulation of ligand binding between and receptors, it was concluded that the presence of ₂ subunit did not change the modulation induced by all the anesthetics tested, except for alphaxalone. In the presence of the steroid anesthetic, a reduced efficacy was not apparent for ₁₂(G219F)₂ receptors; however, alphaxalone was the only anesthetic to induce negative modulation of ligand binding in ₁₂(G219F) recombinant receptors. Perhaps the presence of the ₂ subunit in an ₁₂(G219F)₂ complex may provide additional structural determinants that are sufficient for preserving alphaxalone-induced enhancement of [³H]flunitrazepam binding, therefore masking the deleterious effect of the point mutation on the ₂ subunit, as seen in the combination. Sanna et al. (1997) have shown that receptors can be directly activated by alphaxalone but not etomidate, underlining the importance of the ₂ subunit for the allosteric changes induced by alphaxalone on GABA_A receptors.

With regard to the TM1 point mutation on the ₁ subunit, ₁(G223F), the potentiating effects of pentobarbital and alphaxalone on ligand binding were not altered by this mutation. However, the efficacy of etomidate- and propofol-induced enhancement of both [³H]muscimol and [³H]flunitrazepam binding was significantly reduced in receptors containing the ₁(G223F) subunit. This finding, first, indicates that structural criteria for allosteric modulation is different among these i.v. anesthetics, and second, in addition to the TM1 glycine on the ₂ subunit, this residue on the ₁ subunit is also essential for the allosteric effects of etomidate and propofol. This observation is consistent with a functional study demonstrating the importance of the ₁ subunit for the potentiating effects of both etomidate and propofol (Uchida et al., 1997). Furthermore, it has been shown that amino acids from the TM2 and TM3 domains of the  subunit, which were homologous to essential residues on the  subunit, were critical for positive modulation by volatile anesthetics (Mihic et al., 1997). Taken together, the  subunit can be important for the allosteric modulation induced by both i.v. and volatile anesthetics. Note there was not an additive effect in reducing the efficacy of etomidate and propofol in receptors containing the double mutant (i.e., point mutations in both the ₁ and ₂ subunits). This result may indicate that one of the point mutations is sufficient to disrupt the conformational changes needed for anesthetic-induced modulation of ligand binding.

The ₂(G219F) Point Mutation Alters Functional Aspects of GABA-, Pentobarbital-, and Propofol-Induced GABA_A Chloride Channel Gating. In this study, pentobarbital was not able to enhance GABA currents in ₁₂(G219F)₂ receptors at concentrations that were not confounded by the direct activation effects of pentobarbital. Note, however, that at higher concentrations of pentobarbital, the potentiating effect could possibly be present, although it would be impossible to quantitate in the presence of the direct activation effect. Furthermore, propofol-induced GABA currents were less efficacious in the ₁₂(G219F)₂ receptors at the same anesthetic concentrations tested in ₁₂₂ receptors. These data, first, provide functional correlates for the receptor binding data, demonstrating reduced efficacies in anesthetic-induced enhancement of ligand binding with ₂(G219F) mutant receptors; and second, strongly support the hypothesis that this glycine residue at the entrance to TM1 is important for the allosteric actions of anesthetics. Considering that this TM1 point mutation had minimal or no effect on anesthetic potency in binding experiments or on high-affinity ligand binding, the phenylalanine on the ₂(G219F) subunit may not impair the actual anesthetic or GABA binding site but perhaps alters conformational changes involved in channel gating that are allosterically regulated by GABA and anesthetics. Our findings that the sensitivity of GABA, pentobarbital, and propofol in direct channel-gating function were increased support this conclusion. Note that essential residues in the TM2 domain of the ₂ subunit also appear to be critical for the conformational changes induced by GABA as well as pentobarbital (Birnir et al., 1997; Tierney et al., 1998). Thus, it appears that anesthetics and GABA allosterically induce a potentially similar transduction mechanism for direct channel gating. Because the TM1 point mutation on the  subunit altered the agonistic and modulatory actions of pentobarbital and propofol in a diametrically opposed fashion, this observation supports the working hypothesis that there are distinct structural requirements for this duality of anesthetic action (Jones et al., 1995).

TM1 Point Mutation on the ₁ Subunit Inhibits the Expression of Functional Channels. As discussed, the mutation of the polar glycine residue to the hydrophobic phenylalanine residue appeared to alter conformational flexibility of the receptors. Because glycine is known to confer conformational freedom to peptide chains (Renard et al., 1999), it was surprising to find that the ₁(F261G) receptors did not produce any functional channels. This finding is difficult to resolve with the binding data, which show that the ₁(F261G) subunit produced a small change in anesthetic sensitivity for modulating [³H]muscimol binding as well as demonstrated comparable K_I values for [³H]muscimol, as seen in wild-type ₁ receptors. It is possible that the binding data for the ₁(F261G) receptors could reflect intracellular homomeric receptors that have yet to be expressed on the cell surface. A recent study has shown that mutagenesis of certain extracellular residues in GABA_A receptor ₁ subunit, expressed in Sf9 cells, resulted in muscimol binding activity without cell surface expression (Srinivasan et al., 1999). On the other hand, because all receptor combinations demonstrating binding also conferred channels that were gated by GABA, including the ₁(NASM-RNSL) homomer (data not shown), the mutant, ₁(F261G), homomeric receptors may have been expressed on the cell surface yet were not capable of being gated by GABA (or pentobarbital). Our findings may indicate that this point mutation, which is expressed on every subunit of the ₁(F261G) homomer, obstructed the proper subunit-subunit interactions needed for channel gating. Because the functional wild-type and bridge mutant ₁(NASM-RNSL) receptors have three consecutive phenylalanines at the entrance to the TM1 domain on each subunit, it would be of interest to pursue the importance of the adjacent phenylalanine (F262) in ₁ with regard to channel gating and allosteric events induced by anesthetics.

Concluding Remarks. There is strong evidence in support of anesthetics binding directly to protein targets as opposed to indirectly affecting protein targets via disrupting the lipid bilayer for their mechanism of action (Franks and Lieb, 1991; Eckenhoff, 1998). However, until specific radiolabeled anesthetics are developed, it will remain difficult to determine whether the residues studied here, as well as others (Belelli et al., 1997; Mihic et al., 1997; Moody et al., 1997; Amin, 1999), are involved specifically in forming the binding sites for anesthetics. From our data, first, anesthetic modulation of binding appears to be a predictable indicator of its functional correlate. This conclusion is not novel, and the allosteric modulation of GABA, benzodiazepine, and picrotoxin sites by positive and negative modulators of GABA_A receptors has been found to correlate extremely well for a series of compounds, including stereoisomers, with modulation of GABA_A currents in cultured cells, and with animal behavior (reviewed in Olsen et al., 1991; Carlson et al., 1997). Second, it is concluded that the TM1 glycine residue, which is located at the membrane interface of the protein with the extracellular fluid, is more likely involved in the conformational or allosteric control of channel gating by anesthetics (or GABA) rather than a specific anesthetic binding site. Consistent with this idea are studies on interactions of anesthetics with model membrane ion channels indicating the importance of amphiphilic channel residues at the lipid-water interface (Xu et al.,1998). Furthermore, in muscle-type acetylcholine ligand-gated ion channels and GABA_A receptors, it has been demonstrated that the N-terminal region of TM1 domain may work in concert with the TM2 domain for channel gating (Akabas and Karlin, 1995; Thompson et al., 1999). Thus, the TM1 domain is potentially a link in the chain of conformational events elicited by GABA and anesthetics.