|Home | About | Journals | Submit | Contact Us | Français|
γ-Aminobutyric acid type A (GABAA) receptors are an important target for general anesthetics in the central nervous system. Site-directed mutagenesis techniques have identified amino acid residues that are important for the positive modulation of GABAA receptors by general anesthetics. In the present study, we investigate the role of an amino acid residue in transmembrane (TM) domain 3 of the GABAA receptor β2 subunit for modulation by the general anesthetic 2,6-diisopropylphenol (propofol). Mutation of methionine 286 to tryptophan (M286W) in the β2 subunit abolished potentiation of GABA responses by propofol but did not affect direct receptor activation by propofol in the absence of GABA. In contrast, substitution of methionine 286 by alanine, cysteine, glutamate, lysine, phenylalanine, serine, or tyrosine was permissive for potentiation of GABA responses and direct activation by propofol. Using propofol analogs of varying molecular size, we show that the β2(M286W) mutation resulted in a decrease in the ‘cut-off’ volume for propofol analog molecules to enhance GABA responses at GABAAα1β2γ2s receptors. This suggests that mutation of M286 in the GABAA β2 subunit alters the dimensions of a ‘binding pocket’ for propofol and related alkylphenol general anesthetics.
2,6-Diisopropylphenol (propofol) has proven to be a useful intravenous general anesthetic since its introduction into clinical practice in the 1980s. There have been a number of efforts to understand the molecular mechanism of action of this clinically useful drug (Trapani et al., 2000). One hypothesis, supported by substantial experimental evidence, cites the ability of propofol to positively modulate the function of γ-aminobutyric acid type A (GABAA) receptors, a property shared with other general anesthetics (Franks and Lieb, 1994; Belelli et al., 1999; Krasowski and Harrison, 1999; Trapani et al., 2000). Propofol has been shown in electrophysiological assays to allosterically enhance (potentiate) the actions of GABA at the GABAA receptor (Hales and Lambert, 1991). In addition, propofol prolongs inhibitory postsynaptic currents mediated by GABAA receptors (Bai et al., 2001) and alters receptor deactivation and desensitization (Bai et al., 1999). Propofol can also open the GABAA receptor ion channel in the absence of GABA (termed ‘direct activation’), although this usually occurs at higher concentrations of propofol than necessary to potentiate responses to GABA (Hales and Lambert, 1991; Hara et al., 1993).
Molecular pharmacology studies have defined individual amino acid residues within GABAA receptor molecules that are critical for the allosteric effects of general anesthetics (Belelli et al., 1999; Krasowski and Harrison, 1999). Progress in this effort followed from the observation that receptors related to the GABAA receptor, such as the strychnine-sensitive glycine receptor (Mascia et al., 1996) and the GABAC ρ receptor (Mihic and Harris, 1996) differ in their sensitivity to general anesthetics. Amino acid residues in transmembrane (TM) domains 2 and 3 of GABAA receptor α and β subunits are particularly important for the modulatory actions of ether, alkane, and alcohol anesthetics (Mihic et al., 1997; Krasowski et al., 1998a,b; Koltchine et al., 1999; Ueno et al. 1999, 2000; Krasowski and Harrison, 2000; Jenkins et al., 2001) (Fig. 1), as well as for certain intravenous anesthetics such as propofol, etomidate, and the barbiturates (Amin, 1999; Belelli et al., 1999). In the case of propofol, mutation of methionine 286 to tryptophan (M286W) in TM3 of the GABAAβ1 receptor subunit abolishes potentiation of GABA responses (Krasowski et al., 1998b).
In the present study we further define the importance of this TM3 residue of the β subunit for the modulatory effects of propofol. We also test the effect of tryptophan substitution within TM2 of the β2 subunit and within TM2 and TM3 of the α1 subunit. The final objective was to utilize propofol analogs of different molecular size (Krasowski et al., 2001) to test whether amino acid substitutions for GABAA β2(M286) can alter the ‘cut-off’ for propofol analogs in potentiating GABA responses at GABAAα1β2γ2s receptors. The cut-off for a homologous series of molecules at a given receptor is believed to reflect a binding pocket of limited dimensions, such that molecules beyond a certain size cannot fit into the binding pocket (Franks and Lieb, 1994; Krasowski and Harrison, 1999; Dwyer and Bradley, 2000). The classic homologous series to show cut-off is the n-alcohols, which exhibit cut-offs specific for different ligand-gated ion channels (Li et al., 1994; Peoples and Weight, 1995; Mascia et al., 1996; Mihic and Harris, 1996). This has been taken to suggest that these receptors contain cavities of discrete size for the alcohols, which would be consistent with the finding that alcohol cut-off for modulation of glycine α1 and GABAC ρ receptors can be manipulated by mutagenesis within TM2 and TM3 (Wick et al., 1998). The idea in the present study was to investigate whether amino acid substitutions at GABAAβ2(M286) could produce a similar effect for analogs to propofol and thereby test the hypothesis thatβ2(M286) influences the dimensions of a propofol ‘binding pocket’.
The GABAAα1 (Schofield et al., 1989) and γ2s (Pritchett et al., 1989) receptor subunit cDNAs are of human origin. The GABAAβ2 receptor subunit cDNA is from the rat (Ymer et al., 1989). The human GABAAα1 and γ2s receptor subunit cDNAs were generously provided by the late Dr Dolan Pritchett (University of Pennsylvania, Philadelphia, PA, USA). The rat GABAA β2 subunit cDNA was provided by Dr Dennis Grayson (University of Illinois at Chicago, Chicago, IL, USA).
The β2(N265W), β2(M286F), and β2(M286Y) mutations were introduced by the unique site elimination method using the USE kit (Pharmacia Biotech, Piscataway, NJ, USA) as previously described (Krasowski et al., 1998a,b). For the α1(S270W,A291W) double mutation, the S270W and A291W mutations were initially introduced separately. The vector containing the S270W mutation was digested with Bsu36I and NotI (New England Biolabs) to create a 1 kb fragment, which then replaced the same fragment in the similarly digested vector containing the A291W mutation.
The β2(M286A), β2(M286C), β2(M286E), β2(M286K), β2(M286S), and β2(M286W) mutations in the GABAA β2 subunit were introduced with the QuikChange Site-Directed Mutagenesis kit (Stratagene, La Jolla, CA, USA) as previously described (Krasowski et al., 1998a,b). Positive clones were confirmed by double-stranded sequencing of the entire cDNA insert.
GABAA receptor cDNAs were expressed via the vector pCIS2 which contains one copy of the strong promoter from cytomegalovirus and a polyadenylation sequence from SV40. Human embryonic kidney (HEK) 293 cells (American Type Culture Collection, Rockville, MD, USA) were maintained in culture and passaged weekly by trypsin treatment for a maximum of 20 times before being discarded and replaced with early passage cells. HEK 293 cells were maintained in Eagle’s minimum essential medium (Sigma, St. Louis, MO, USA) supplemented with 10% fetal bovine serum (Hyclone, Logan, UT, USA), l-glutamine (0.292 μg/ml; Gibco BRL, Grand Island, NY, USA), penicillin G sulfate (100 units/ml; Gibco BRL), and streptomycin sulfate (100 μg/ml; Gibco BRL).
For electrophysiological experiments, cells were plated on glass coverslips coated with poly-D-lysine (Sigma). Each coverslip of cells was individually transfected with the calcium phosphate precipitation technique as previously described (Krasowski et al., 1998a). Each transfection used 1–5 μg of each cDNA; the cDNA was in contact with the cells for 24 h under an atmosphere containing 3% CO2 before being removed and replaced with fresh culture medium in an atmosphere of 5% CO2.
Stock solutions of GABA (Sigma) and the propofol analogs were diluted into extracellular solution daily before use. The propofol analogs were all prepared as stock solutions in dimethyl sulfoxide before being dissolved in the extracellular medium. The maximum final concentration of dimethyl sulfoxide was 0.2% (v/v), which was determined during carrier control experiments to have no significant effect on GABA-induced currents in the GABAA receptors analyzed in this study.
Electrophysiological recordings were performed at room temperature using the whole-cell patch clamp technique as previously described (Krasowski et al., 1998a,b). The coverslips were transferred 48–96 h after removal of the cDNA to a large chamber that was continuously perfused (2–3 ml/min) with extracellular medium containing (in mM): 145 NaCl, 3 KCl, 1.5 CaCl2, 1 MgCl2, 5.5 D-glucose, and 10 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), pH 7.4, osmolarity 320–330 mOsmol. The electrode solution contained (in mM): 145 N-methyl-D-glucamine hydrochloride, 5 K2ATP, 5 HEPES/KOH, 2 MgCl2, 0.1 CaCl2, and 1.1 ethylene glycol bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid (EGTA), pH 7.2, osmolarity 315 mOsmol. Pipette-to-bath resistance was 4–6 MΩ. Cells were voltage-clamped at _60 mV. Since the intracellular and extracellular solutions contained symmetrical chloride concentrations, the chloride equilibrium potential was approximately 0 mV.
All drugs were applied to the cell by local perfusion (Krasowski et al., 1998a,b) using a motor-driven solution exchange device (Bio Logic Rapid Solution Changer RSC-100; Molecular Kinetics, Pullman, WA, USA). Laminar flow was maintained by applying all solutions at identical flow rates via a multi-channel infusion pump (Stoelting, Wood Dale, IL, USA). The solution changer was driven by protocols in the acquisition program pCLAMP5 (Axon Instruments, Foster City, CA, USA). Responses were low-pass-filtered at 5 kHz and digitized (TL-1-125 interface; Axon Instruments) using pCLAMP5 and stored for off-line analysis.
Potentiation of GABA responses by propofol and other drugs was always assessed by co-application of the drug with an EC20 GABA concentration (i.e. the concentration of GABA that produces 20% of the maximal response to GABA). Anesthetics were always preapplied prior to co-application with GABA to ensure that the anesthetic had reached equilibrium with the receptor. At the end of each potentiation experiment, a maximal GABA response was elicited by an appropriately high concentration of GABA to verify that the average control GABA response fell between EC10 and EC30. This experimental design permits comparison not only of anesthetic EC50 between various receptor combinations but also of the maximal magnitude (relative efficacy) of agonist potentiation between anesthetics (Krasowski and Harrison, 2000). Anesthetic direct activation currents were expressed as a fraction of the maximal current response to GABA.
Drug-induced potentiation of a GABA-induced current was defined as the percentage increase of the control GABA response (defined as the average of the pre-drug and post-drug GABA-induced currents). Concentration response data were fitted (KaleidaGraph; Reading, PA, USA) with the equation: I/Imax=100[drug]nH/([drug]nH+(EC50)nH), where I/Imax is the percentage of the maximum obtainable response, EC50 the concentration producing a half-maximal response, and nH is the Hill coefficient. Pooled data are presented throughout as mean±SE. Statistical significance was determined by two-way analysis of variance with Dunnet’s post-hoc test, unless otherwise specified.
Molecular volumes for the propofol analogs were calculated using the cerius2 qsar software version 3.0 (Molecular Simulations, San Diego, CA, USA). The molecular volumes of the propofol analogs are listed in Fig. 7. The molecular volumes of the amino acid residue side chains considered in this study are (Harpaz et al., 1994): alanine (A; 0.026 nm3), asparagine (N; 0.064 nm3), cysteine (C; 0.04 nm3), glutamic acid (E; 0.077 nm3), glycine (G; 0 nm3), lysine (K; 0.106 nm3), methionine (M; 0.104 nm3), phenylalanine (F; 0.13 nm3), serine (S; 0.03 nm3), tryptophan (W; 0.168 nm3), tyrosine (Y; 0.133 nm3), and valine (V; 0.075 nm3).
The sources of the propofol analogs and other anesthetics were as follows: 2,6-dimethylphenol, 2-isopropylphenol (Aldrich Chemical Co., Milwaukee, WI, USA), 2,6-diethylphenyl isothiocyanate (Lancaster Synthesis, Windham, NH, USA), and methohexital sodium (Brevital sodium; Eli Lilly and Co., Indianapolis, IN, USA). Each analog was of the highest purity grade available commercially. Propofol, 2,6-diethylphenol, and 2,6-di-tert-butylphenol were generously provided by Drs. J.B. Glen and Roger James of Zeneca Pharmaceuticals (Macclesfield, Cheshire, UK; see James and Glen, 1980). Research-grade etomidate was a generous gift from Professor Alfred Doenicke (Institute of Anesthesiology, Ludwig Maximillians University of Munchen, Germany).
The wild-type GABAA receptor and all of the receptors containing point mutations in the α1 or β2 subunits responded to GABA in a concentration-dependent manner. Fig. 2(A) and (B) shows responses to GABA for the wild-type GABAAα1β2γ2s receptor and for the GABAAα1β2(M286W)γ2s receptor. Fig. 3 displays GABA concentration–response curves for the wild-type GABAAα1β2γ2s receptor and for the series of eightα1β2(M286X)γ2s mutated receptors.
The EC50 values for GABA for the α1β2(M286X)γ2s series of receptors varied over approximately one order of magnitude, with the greatest absolute difference found in the α1β2(M286C)γ2s receptor, which had a 4.5-fold higher EC50 to GABA than the α1β2γ2s receptor. GABA EC50, Hill slope, and maximal response data for all receptors are summarized in Table 1. In contrast to the relatively modest effects on GABA sensitivity produced by the series of β2(M286X) mutant subunits, the GABAA α1(S270W,A291W)β2γ2s receptor had an EC50 value for GABA 25 times lower than that for the wild-type GABAAα1β2γ2s receptor. This is consistent with the previously described effects of theα2(S270W) andα2(A291W) mutations, which shift the GABA concentration–response relationship for the GABAAα2β1γ2s receptor to the left by 7.1- and 3.4-fold, respectively (Krasowski et al., 1998b; Koltchine et al., 1999).
Propofol potentiated submaximal (EC20) GABA responses at the wild-type GABAA α1β2γ2s receptor (Figs. 2(C) and and4)4) but failed to enhance submaximal GABA responses at the GABAA α1β2(M286W)γ2s receptor (Figs. 2(D) and 4(A); Table 1). In contrast to the lack of potentiation of GABA responses, propofol directly activated the GABAA α1β2(M286W)γ2s receptor, although with a 2.3-fold higher EC50 value than at the wild-type GABAA α1β2γ2s receptor (Figs. 2(D) and 5(A); Table 1). Complete concentration–response curves for potentiation of GABA responses and direct activation by propofol for the wild-type GABAA α1β2γ2s receptor and for the series of α1β2(M286X)γ2s mutated receptors are illustrated in Figs. 4 and and5,5, with summary of data for all receptors presented in Table 1.
To test the importance of the α subunit for the effects of propofol, potentiation of GABA responses by propofol was studied at the GABAA α1(S270W,A291W)β2γ2s receptor. This receptor showed potentiation of GABA responses by propofol similar to that at the wild-type GABAA α1β2γ2s receptor (Table 1).
We next investigated the relationship between the physical properties of the amino acid residue at position 286 of the GABAAβ2 subunit and the responses to GABA and propofol. The amino acid properties chosen were molecular volume (Harpaz et al., 1994), polarity (Zimmerman et al., 1968), hydropathy (Kyte and Doolittle, 1982), and hydrophilicity (Hopp and Woods, 1981). There was no significant correlation between any of these amino acid properties and the GABA EC50, GABA maximal current, EC50 for potentiation of GABA responses by propofol, potentiation of control GABA response by 2, 5, 10, or 20 μM propofol, EC50 for propofol direct activation, or maximal direct activation by propofol (Pearson’s correlation r2=0.002–0.56, P<0.05 for each). In addition, the GABA EC50 value did not correlate with the EC50 values for direct activation or potentiation of GABA responses by propofol (r2=0.07 and 0.13, respectively, P<0.05 for each); similarly, the EC50 value for potentiation of GABA responses by propofol also did not correlate with the EC50 value for propofol direct activation (r2=0.01, P<0.05).
We next determined the effects of methohexital (a barbiturate) and etomidate on submaximal GABA responses at the GABAA α1β2γ2s and α1β2(M286W)γ2s receptors, to test whether this mutant receptor was capable of being modulated by anesthetics from different chemical classes than propofol. Methohexital (10 μM) potentiated EC20 GABA responses at both receptors: 199±39% potentiation at the GABAA α1β2γ2s receptor (n=5) and 118±18% at the GABAA α1β2(M286W)γ2s receptor (n=6) (P<0.05 compared with response at the wild-type receptor). Etomidate (10 μM) also potentiated EC20 GABA responses at both receptors: 237±45% potentiation at the GABAA α1β2γ2s receptor (n=6) and 162±11% at the GABAA α1β2(M286W)γ2s receptor (n=5) (P<0.05 compared with response at the wild-type receptor).
Two approaches were adopted to test the possibility that substitutions at the β2(M286) position might alter the size of propofol analog that could potentiate GABA responses. One approach was to mutate β2(M286) to a smaller amino acid residue to see if this would allow an analog bulkier than propofol to gain access to the propofol binding site by creating a larger binding cavity (an increase in cut-off). The analog chosen was 2,6-di-tertbutylphenol, shown in a separate study to be completely inactive at wild- type GABAA α1β2γ2s receptors (Krasowski et al., 2001). The second approach investigated whether submaximal GABA responses at the GABAA α1β2(M286W)γ2s receptor could still be potentiated by analogs smaller than propofol (a decrease in cut-off), even though propofol itself is inactive at this mutant receptor.
In the first set of experiments, the effects of 2,6-ditert-butylphenol were tested at GABAA α1β2(M286A)γ2s and GABAA α1β2(M286S)γ2s receptors. The M286A and M286S mutations result in 0.078 and 0.074 nm3 decreases in amino acid molecular volume at the 286 position, respectively, while 2,6-di-tert-butylphenol is 0.034 nm3 greater in molecular volume than propofol. 2,6-Di-tert-butylphenol did not, however, potentiate GABA responses at the GABAA α1β2(M286A)γ2s and GABAA α1β2(M286S)γ2s receptors (Table 2). 2,6-Ditert-butylphenol also produced no direct activation at either receptor at concentrations up to 500 μM (n=5). In contrast, four analogs smaller than propofol (molecular volume=0.193 nm3), 2,6-dimethylphenol (0.126 nm3), 2-isopropylphenol (0.143 nm3), 2,6-diethylphenol (0.160 nm3), and 2,6-diethylphenyl isothiocyanate (0.187 nm3), each potentiated GABA responses at the GABAA α1β2(M286A)γ2s and α1β2(M286S)γ2s receptors (Table 2). Thus, the GABAA α1β2(M286A)γ2s and GABAA α1β2(M286S)γ2s receptors have a cut-off for propofol analogs similar to that for the wild-type GABAA α1β2γ2s receptor, at least for the six propofol analogs tested (Fig. 7).
In the second set of experiments, the actions of propofol and five analogs were tested at the GABAA α1β2(M286W)γ2s receptor. The M286W mutation results in a 0.064 nm3 increase in molecular volume at position 286. The two analogs with the smallest molecular volumes, 2,6-dimethylphenol (0.126 nm3) and 2-isopropylphenol (0.143 nm3), enhanced submaximal GABA currents at the GABAA α1β2(M286W)γ2s receptor (Figs. 6 and and7;7; Table 2). In contrast, 2,6-diethylphenol (0.16 nm3), 2,6- diethylphenyl isothiocyanate (0.187 nm3), propofol (0.193 nm3), and 2,6-di-tertbutylphenol (0.227 nm3) all failed to potentiate submaximal GABA currents at the GABAA α1β2(M286W)γ2s receptor (Figs. 2(D), ,66 and and7;7; Table 2). Thus, the GABAA β2(M286W) mutation reduces the propofol analog cut-off for potentiation of GABA responses from 0.193 nm3 (propofol) to below 0.16 nm3 (2,6-diethylphenol) (Fig. 7).
We explain these experimental data by proposing that mutation of M286 in the GABAA receptor β2 subunit alters the dimensions of a binding cavity for the general anesthetic propofol. We will first discuss our results in relation to those of other studies employing site-directed mutagenesis to probe sites of anesthetic action on GABAA receptor subunits and then turn to discussion of the putative propofol binding cavity.
The effects of the GABAA β2 mutations described in this study contrast with previously described effects of mutations within the GABAA α2 subunit or the glycineα1 subunits. The GABAA β2(M286W) mutation, which abolishes potentiation of GABA responses by propofol, has only small effects on the potency of GABA or for direct activation by propofol at the GABAA α1β2γ2s receptor (1.2 to 2.6-fold shifts of the respective concentration–response curves; see Table 1). In contrast, site-directed substitutions of GABAA α2(S270) or α2(A291), or of glycine α1(S267) or α1(A288) (see Fig. 1), markedly affect both the ability of general anesthetics to potentiate receptor function and receptor gating by agonists or anesthetics (Ye et al., 1998; Krasowski et al., 1998a,b; Koltchine et al., 1999; Ueno et al. 1999, 2000; Yamakura et al., 1999).
More specifically, the molecular volume of the amino acid side chain at the GABAA α2(S270), glycine α1(S267), or glycine α1(A288) positions correlates negatively with agonist EC50 and with anesthetic potentiation of agonist responses (Ye et al., 1998; Koltchine et al., 1999; Yamakura et al., 1999). In contrast, the present study and others show that site-directed substitutions of the 265 or 286 positions of GABAA β1 or β2 subunits do not produce any consistent alterations in agonist EC50 (Krasowski et al., 1998a,b; Amin, 1999; Ueno et al., 1999). This may relate to the dominant role of GABAA α subunits in determining agonist potency and efficacy (Levitan et al., 1988; Sigel et al., 1990; Ebert et al., 1994; O’Shea and Harrison, 2000).
There are a number of ways in which amino acid substitutions at β2(M286) could potentially disrupt the modulatory effects of propofol. First, the mutations could result in receptors that are already ‘maximally sensitive’ to GABA, such that propofol cannot shift the concentration–response relationship to GABA further to the left. This mechanism, which may apply to GABAA receptors containing the β2(L259S) mutation (Thompson et al., 1999), is improbable in the GABAA α1β2(M286W)γ2s receptor due to the minimal alterations in EC50 for GABA and the observation that etomidate and methohexital potentiate submaximal responses at this receptor (albeit with responses significantly smaller than at the wild-type GABAA α1β2γ2s receptor).
A second possibility is that the mutated receptor is more easily gated by propofol, which might obscure the measurement of potentiation of GABA-evoked currents. This problem has been encountered in other studies of anesthetic effects on mutated GABAA α2 and β2 receptor subunits (Krasowski et al., 1998b; Thompson et al., 1999). However, the GABAA α1β2(M286W)γ2s receptor analyzed in this study actually shows less direct activation by propofol concentrations below 20 μM than the wild-type GABAA α1β2γ2s receptor, discounting the possibility that direct activation obscures potentiation of GABA responses by propofol at this mutated receptor.
The remaining possibilities to explain the effects of the β2 mutations are that the mutations disrupt the binding site for propofol or, alternatively, disrupt conformational change(s) of the GABAA receptor that occur following the binding of propofol to the receptor. Distinguishing between these two possibilities is difficult because fine-resolution, three-dimensional structural data for the ligand-gated ion channels are not yet available. We therefore pursued an alternate approach, namely that of considering whether site-directed mutagenesis ofβ2(M286) altered the characteristics of a putative propofol binding cavity.
A number of investigators have suggested that general anesthetic molecules may exert their effects by binding fortuitously to cavities naturally present on proteins, thereby stabilizing certain conformational states such as the open and closed states of the GABAA receptor ion channel (LaBella, 1981; Franks and Lieb, 1985; Williams and Akabas, 1999). Although GABAA receptors and other ligand-gated ion channels have so far resisted attempts to determine high-resolution structures, there have been a number of studies which demonstrate general anesthetic binding to cavities in soluble proteins that serve as models for true anesthetic targets. The high resolution structures of firefly luciferase complexed with bromoform (Franks et al., 1998) and human serum albumin complexed with halothane and propofol (Bhattacharya et al., 2000) demonstrate striking examples of anesthetic binding to pre-existing amphiphilic protein cavities.
In addition, alkylphenol binding cavities have been demonstrated on a variety of proteins, including the enzymes phenol hydroxylase (Neujahr, 1990; Enroth et al., 1998), vanillyl-alcohol oxidase (Mattevi et al., 1997; van den Heuvel et al., 1998), and p-cresol methylhydroxylase (McIntire et al., 1985; Cunane et al., 2000). In these three enzymes, structural determination by X-ray crystallography has revealed hydrophobic cavities that selectively bind alkylphenols of specific shapes and sizes. For instance, the 0.2 nm3 catalytic cavity of vanillyl-alcohol oxidase is completely filled by the competitive inhibitor p-(-1-heptenyl)phenol, which explains why alkylphenols bearing aliphatic substituents longer than seven carbon atoms at the para position do not bind to the enzyme (Mattevi et al., 1997; van den Heuvel et al., 1998). The substrate binding cavity of phenol hydroxylase has a shape and size that can accommodate 4-tert-butylphenol but not 2-isopropylphenol (Enroth et al., 1998; Neujahr, 1990).
Much like the structure–affinity relationships for alkylphenols at the three enzymes described above, the structure–activity relationships for propofol analogs for modulating GABAA receptors or producing loss of righting reflex in Xenopus laevis tadpoles exhibit distinct ligand requirements, for which steric factors seem to play a major role (Krasowski et al., 2001). For example, propofol (molecular volume=0.193 nm3) and 2,6-diethylphenyl isothiocyanate (0.187 nm3) are ‘active’, whereas 2,6-dicyclopentylphenol (0.24 nm3) and 2,6-diisopropylphenyl isocyanate (0.211 nm3) are completely inactive. These structure–activity relationships are consistent with the hypothesis that propofol analogs interact with a binding cavity of limited dimensions on the GABAA receptor, similar perhaps to the alkylphenol binding cavities on phenol hydroxylase, vanillyl-alcohol oxidase, or p-cresol methylhydroxylase.
If β2(M286) is indeed part of a binding cavity for propofol, then mutagenesis of this residue might alter the specificity of propofol analogs that can modulate GABAA receptor function. This is precisely what happened, with the propofol analog cut-off for potentiation of GABA responses reduced from 0.193 nm3 (propofol) in the wild-type GABAA α1β2γ2s receptor to below 0.16 nm3 (2,6-diethylphenol) in the GABAA α1β2(M286W)γ2s receptor (Fig. 6). In this context, it is of interest to consider studies of the catalytic pockets of alcohol dehydrogenase (Weinhold and Benner, 1995) and L-hydroxyisocaproate dehydrogenase (Feil et al., 1997). In both of these enzymes, it is possible to alter substrate specificity by site-directed mutagenesis of amino acid residues lining the substrate binding pocket.
Future studies will continue to elucidate the role of the amino acid residues involved in the actions of anesthetics. Even more exciting, perhaps, is the use of targeted gene manipulations in mice to test the relevance of certain ligand-gated ion channels in mediating the behavioral actions of general anesthetics (Belelli et al., 1999; Krasowski and Harrison, 1999). The GABAA receptor β2(M286W) mutation described in this study may be an ideal candidate for the creation of knock-in mice, in which the endogenous β2 subunit is replaced by the mutated, anesthetic-insensitive subunit.
Funding for this study was provided by the C.V. Starr Foundation (New York City, NY, USA) and the Rice Foundation (Chicago, IL, USA) to N.L.H., by National Institutes of Health grants P01-GM62195 and GM56850 to N.L.H., and by National Institute of Mental Health training fellowship MH11504 to M.D.K. K.N. was supported by a research fellowship from the Uehara Memorial Foundation (Tokyo, Japan). The authors are grateful to Amiinah Kung, Irene Paraskevakis, and Steve Lopez for invaluable technical support. The authors also thank Drs J.B. Glen and Roger James of Zeneca Pharma ceuticals for providing propofol analogs and Professor Alfred Doenicke for supplying etomidate.