Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Since its discovery in 1980, propofol (2,6-diisopropylphenol) has proven to be a clinically useful general anesthetic. The advantages of propofol include a rapid onset and offset of action and relatively low toxicity, which has led to the use of propofol in many surgical and critical care settings (Langley and Heel, 1988). 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 -aminobutyric acid type A (GABA_A) receptors, a property common to many other general anesthetics (Franks and Lieb, 1994; 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 GABA_A receptor (Hales and Lambert, 1991) and also to prolong inhibitory postsynaptic currents mediated by GABA_A receptors (Orser et al., 1994). Propofol can also open the GABA_A receptor ion channel in the absence of GABA (termed "direct activation") although this usually occurs at higher concentrations of propofol than necessary to potentiate submaximal receptor responses to GABA (Hales and Lambert, 1991; Hara et al., 1993; Jones et al., 1995).

The general anesthetic properties of propofol were initially discovered during a screen of 97 alkylphenols in mice and rabbits, following up on the initial observation that 2,6-diethylphenol possessed potent anesthetic effects (James and Glen, 1980). There have been few studies using structure-activity relationship (SAR) analysis for propofol (Trapani et al., 1998; Sanna et al., 1999; Lingamaneni et al., 2001), even though the relatively simple molecular structure of propofol would seem to favor SAR analysis.

The studies in this manuscript report concentration-response relationships for three actions of 27 propofol analogs (Fig. 1): loss of righting reflex in Xenopus laevis tadpoles, potentiation of submaximal GABA responses at the GABA_{A 122s} receptor, and direct activation of the GABA_{A 122s} receptor. X. laevis tadpoles were chosen for the determination of in vivo anesthetic potency because propofol and other phenols have very complicated pharmacokinetics in mammals, including extensive binding to plasma proteins (Langley and Heel, 1988; Trapani et al., 2000). The loss of righting reflex assay in tadpoles, which has a long history with respect to the study of general anesthetics, was therefore chosen to generate a self-consistent set of potencies for the immobilizing properties of the propofol analogs (Downes and Courogen, 1996). The GABA_{A 122s} receptor was selected for the electrophysiological studies because this represents the most common subunit combination in the mammalian central nervous system (McKernan and Whiting, 1996).

View larger version (23K): [in this window] [in a new window]   Fig. 1.   Chemical structures of the 27 propofol analogs analyzed in this study. The top structure illustrates propofol, depicting the numbering of the carbon atoms on the aromatic ring.

Additional studies characterized the pharmacological properties of GABA_A receptors in dissociated spinal neurons from X. laevis tadpoles. Spinal neurons were chosen because spinal cord circuitry is involved in mediating the immobilization produced by general anesthetics (for review, see Collins et al., 1995), although it is not yet clear if the neuronal circuitry underlying nocifensive movements in mammals and righting reflex in tadpoles is the same. The results of these experiments further justified the use of tadpoles as a model system for studying general anesthesia and the role of GABA_A receptors in the effects of propofol and its analogs.

We also compared the potencies of the propofol analogs with their calculated octanol/water partition coefficients, a measure of lipid solubility. The relationship between general anesthetic potency and the octanol/water partition coefficient, the "Meyer-Overton correlation", has often been invoked to suggest that general anesthetics act at lipid, as opposed to protein, targets. Traditional theories of general anesthesia, however, have proved increasingly untenable in the face of experimental evidence (Franks and Lieb, 1994; Krasowski and Harrison, 1999), including the discovery of compounds known as "nonimmobilizers", which possess high lipid solubility and yet lack anesthetic activity, despite being structurally related to ether and alkane general anesthetics (Koblin et al., 1994). The propofol analogs selected for investigation here included compounds such as 2,6-di-tert-butylphenol known to lack anesthetic activity in mice (James and Glen, 1980). The experiments thus tested what molecular effect best accounts for both the anesthetic activity and inactivity of a series of propofol analogs.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Determination of Anesthetic Potencies for Loss of Righting Reflex in X. laevis Tadpoles. General anesthetic potencies were determined as previously described (Krasowski and Harrison, 2000) for X. laevis tadpoles (Xenopus 1; Ann Arbor, MI) in the prelimb-bud stage of development, corresponding to stages 43 to 50 of the standard nomenclature for X. laevis development (Nieuwkoop and Faber, 1956). Tadpoles were maintained in dechlorinated tap water in an aerated aquarium at room temperature.

1

Dissociation and Culturing of Tadpole Spinal Neurons. The method used to dissociate X. laevis tadpole spinal neurons was adapted from a procedure previously described (Dale, 1991). The following saline solutions were used: HEPES saline, 115 mM NaCl, 3 mM KCl, 1 mM Ca(NO₃)₂, 1 mM CaCl₂, 1 mM MgCl₂, 2.4 mM NaHCO₃, 10 mM glucose, 10 mM HEPES, adjusted to pH 7.6 with NaOH; PIPES saline, 115 mM NaCl, 3 mM KCl, 0.1 mM CaCl₂, 0.1 mM MgCl₂, 25 mM glucose, 10 mM piperazine-N,N'-bis[2-ethanesulfonic acid] (PIPES), adjusted to pH 7.0 with HCl; and dissociation saline, 115 mM NaCl, 3 mM KCl, 2 mM EDTA, 25 mM glucose, 10 mM PIPES, adjusted to pH 7.0 with HCl. All saline solutions were saturated with 100% O₂ before use.

Cell Culture and Transfection of Receptor cDNAs. The GABA_{A 1} (Schofield et al., 1989) and _2s (Pritchett et al., 1989) receptor subunits cDNAs are of human origin. The GABA_{A 2} receptor subunit cDNA is from the rat (Ymer et al., 1989). The human GABA_{A 1} and _2s receptor subunit cDNAs were generously provided by the late Dr. Dolan Pritchett (University of Pennsylvania, Philadelphia, PA). The rat GABA_{A 2} subunit cDNA was provided by Dr. Dennis Grayson (University of Illinois at Chicago, Chicago, IL).

Electrophysiology and Design of Pharmacology Experiments. Electrophysiological recordings were performed at room temperature using the whole-cell patch-clamp technique as previously described (Krasowski et al., 1998). The coverslips were transferred 48 to 96 h after removal of the cDNA to a large chamber that was continuously perfused (2-3 ml/min) with extracellular medium containing 145 mM NaCl, 3 mM KCl, 1.5 mM CaCl₂, 1 mM MgCl₂, 5.5 mM D-glucose, and 10 HEPES, pH 7.4, osmolarity 320 to 330 mosmol. The electrode solution contained 145 mM N-methyl-D-glucamine hydrochloride, 5 mM K₂ATP, 5 mM HEPES/KOH, 2 mM MgCl₂, 0.1 mM CaCl₂, and 1.1 mM EGTA, pH 7.2, osmolarity 315 mosmol. Pipette-to-bath resistance was 4 to 6 M. 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.

Data Analysis. 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 predrug and postdrug GABA-induced currents). Concentration response data were fitted (KaleidaGraph, Reading, PA) with the equation: I/I_max = 100 × [drug]^n_H/([drug]^n_H + (EC₅₀)^n_H), where I/I_max is the percentage of the maximum obtainable response, EC₅₀ is the concentration producing a half-maximal response, and n_H is the Hill coefficient. Pooled data are presented throughout as mean ± S.E. Statistical significance was determined by one-way analysis of variance with Dunnett's post hoc test, unless otherwise specified.

Drugs. Stock solutions of GABA (Sigma) and the propofol analogs were diluted into extracellular solution daily before use. With the exception of phenol, the propofol analogs were all prepared as stock solutions in dimethyl sulfoxide as carrier before being dissolved in the extracellular medium. The maximum final concentration of dimethyl sulfoxide was 0.2% (v/v), which was determined during control experiments to have no significant effect on GABA-induced currents in the recombinant or neuronal GABA_A receptors analyzed in this study.

Calculated Physicochemical Properties of the Propofol Analogs. Log P (where P = octanol/water partition coefficient) and the molecular volume for each molecule were calculated using the QSAR Properties program version 1.6, used in conjunction with HyperChem 5.0 software (Hypercube Inc., Gainesville, FL).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Potencies for the Propofol Analogs in Producing Loss of Righting Reflex in Tadpoles. Potencies for producing loss of righting reflex in X. laevis tadpoles were determined for propofol and related analogs (Table 1). The concentration-response curve for the loss of righting reflex elicited by propofol and three analogs is shown in Fig. 2A. A previous study has determined the EC₅₀ for propofol-induced loss of righting reflex in X. laevis tadpoles to be 1.9 ± 0.1 µM (Tonner et al., 1997), identical to the EC₅₀ value described here. The EC₅₀ for loss of righting reflex in tadpoles differs by severalfold from the general anesthetic potency of propofol in mammals, although the potency in mammals is complicated by extensive binding to serum proteins (Franks and Lieb, 1994).

View this table: [in this window] [in a new window]   TABLE 1 Physicochemical properties of the propofol analogs and potencies of the analogs for loss of righting reflex in X. laevis tadpoles and for potentiation of GABA responses and direct activation at GABAA 122s receptors expressed in HEK 293 cells Molecular volume and log P are calculated parameters (under Materials and Methods). The EC50 and maximal effect values are presented as mean ± S.E. n = 5-12 for all data points in the potentiation of GABA responses and direct activation experiments. The LC50 values in tadpoles are given as mean ± S.E., when possible, or otherwise as an approximate range of concentrations.

View larger version (23K): [in this window] [in a new window]   Fig. 2.   A, concentration-response curves for loss of righting reflex in X. laevis tadpoles for propofol, 2,6-diethylphenol, 2,6-dimethylphenol, and 2,6-di-tert-butylphenol. 2,6-Di-tert-butylphenol did not produce loss of righting reflex at any concentration. All curves are for 30 min of exposure. Curves for propofol, 2,6-diethylphenol, and 2,6-dimethylphenol are fit by quantal dose-response relationships as described under Materials and Methods, with the curve fit parameters listed in Table 1. B, time course for loss of righting reflex in tadpoles for propofol, 4-iodopropofol, and 2,4-di-sec-butylphenol. The ordinate depicts the estimate of the EC50 for loss of righting reflex at a particular time point of drug exposure.

Pharmacology of GABA_A Receptors from Dissociated Tadpole Spinal Neurons. The pharmacology of tadpole GABA_A receptors was assessed in acutely dissociated spinal neurons. Dissociation of tadpole spinal neurons by the method of Dale (1991) yielded healthy neurons that were suitable for patch-clamp analysis from 4 h after dissociation. Application of GABA elicited current responses in approximately 70% of such neurons. Desensitization during GABA application was especially noticeable at high GABA concentrations (>500 µM), and recovery from desensitization occurred over several minutes following application of 1 mM GABA. GABA concentration-response data for the spinal neurons is depicted in Fig. 3. The maximal response to GABA, for those neurons that responded to GABA, was 488 ± 149 pA (n = 16). Figure 3B also displays the GABA concentration-response curve for the GABA_{A 122s} receptor expressed in HEK 293 cells.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   GABA concentration-response relationships for GABAA receptors of dissociated X. laevis tadpole spinal neurons. A, representative traces from an individual tadpole spinal neuron in response to application of 5, 20, 50, 100, 1000, and 2000 µM GABA. Note that the responses to 1000 and 2000 µM GABA overlap. B, concentration-response curve for GABA at dissociated tadpole spinal neurons with the GABA concentration curve for the GABAA 122s receptor expressed in HEK 293 cells shown for comparison. The EC50 for GABA in the dissociated spinal neurons was 28.3 ± 2.1 µM with a Hill slope of 1.7 ± 0.2 (n = 6). The EC50 for the GABAA 122s receptor expressed in HEK 293 cells was 29.5 ± 2.4 µM with a Hill slope of 1.4 ± 0.1 (n = 7).

View larger version (22K): [in this window] [in a new window]   Fig. 4.   Submaximal (EC20) GABA current responses in tadpole spinal neurons are enhanced by loreclezole (5 µM), the benzodiazepine midazolam (0.5 µM) (A), and propofol (2 µM), but not by 2,6-di-tert-butylphenol (100 µM) (B). In addition, 2,6-di-tert-butylphenol failed to produce any direct activation during preapplication. Traces shown in A and B are individual recordings from tadpole spinal neurons. C, summary of the effects of loreclezole, midazolam, propofol, and 2,6-di-tert-butylphenol (2,6-DTBP) on submaximal GABA responses at tadpole spinal neurons (n = 5 for all points). The ordinate depicts percentage of potentiation of an EC20 test concentration of GABA by coapplication with drug.

Potentiation of GABA Responses and Direct Activation by the Propofol Analogs at GABA_{A 122s} Receptors. Concentration-response curves for potentiation of GABA responses and direct activation by all 27 propofol analogs were determined at GABA_{A 122s} receptors (Table 1; see Fig. 3B for the GABA concentration-response curve for the GABA_{A 122s} receptor). Log P (octanol/water partition coefficients) and molecular volume values for the molecules are also included in Table 1. The compounds in Table 1 are sorted by molecular volume.

View larger version (17K): [in this window] [in a new window]   Fig. 5.   Propofol and 2,6-di-sec-butylphenol potentiate GABA responses and directly activate GABAA 122s receptors, whereas 2,6-di-tert-butylphenol is inactive. A, potentiation of EC20 GABA responses by propofol at GABAA 122s receptors. Propofol (1 and 5 µM) elicits direct activation during preapplication. B, potentiation of EC20 GABA responses by 2,6-di-sec-butylphenol at GABAA 122s receptors. 2,6-Di-sec-butylphenol (5 and 20 µM) elicits direct activation during preapplication. C, in contrast, 2,6-di-tert-butylphenol at 1, 100, and 500 µM does not potentiate responses to 12 µM GABA. The 500 µM application of 2,6-di-tert-butylphenol illustrates the complete lack of direct receptor activation produced by this compound. Traces shown in A through C are individual recordings from HEK 293 cells transfected with cDNAs encoding the GABAA 1, 2, and 2s subunits.

View larger version (29K): [in this window] [in a new window]   Fig. 6.   Concentration-response relationships for potentiation of GABA responses (A) and direct activation (B) at GABAA 122s receptors by propofol, 2,6-diethylphenol, 2,6-dimethylphenol, 2-isopropylphenol, and 2,6-di-tert-butylphenol (n = 5-12 for all points). 2,6-Di-tert-butylphenol failed to potentiate GABA responses or directly activate GABAA 122s receptors at all concentrations tested. See Table 1 for summary data on the curve fits.

Lack of Negative Allosteric or Null Modulation by the Propofol Analogs. We next looked for evidence that any of the propofol analogs functioned as negative allosteric or null modulators (i.e., a "propofol antagonist" that blocks the effects of propofol but has no intrinsic action at the GABA_A receptor). The model for this is the impressive range of ligands that has been shown to compete for a common benzodiazepine "binding site" at GABA_A receptors. Positive, negative, and null benzodiazepine modulators have all been described (Sigel and Buhr, 1997).

View larger version (21K): [in this window] [in a new window]   Fig. 7.   2,6-Di-tert-butylphenol does not antagonize the loss of righting reflex by propofol in X. laevis tadpoles, nor does 2,6-di-tert-butylphenol antagonize the potentiation of GABA responses or direct activation by propofol at GABAA 122s receptors. A, concentration-response curves for loss of righting reflex in tadpoles for propofol alone and for propofol in the presence of 200 µM 2,6-di-tert-butylphenol. Both curves are for 20 min of exposure. Curves in A are fit by quantal dose-response relationships as described under Materials and Methods. Concentration-response curves for potentiation of GABA responses (B) and direct activation (C) by propofol at GABAA 122s receptors in the presence and absence of 500 µM 2,6-di-tert-butylphenol. See Results for summary data on the curve fits in A-C.

Correlation Analysis of the Potencies of the Propofol Analogs. We next evaluated the correlations between the potencies of the propofol analogs in producing loss of righting reflex in tadpoles and for producing potentiation of GABA responses and direct activation of GABA_{A 122s} receptors. These potencies were also compared with two physicochemical parameters, molecular volume and log P. The aim was to assess the impact of molecular size and lipophilicity of the propofol analogs on biological activity.

View larger version (23K): [in this window] [in a new window]   Fig. 8.   Correlation of propofol analog log P and molecular volume values with the potency for producing loss of righting reflex in tadpoles (A and B), potentiation of GABA responses at GABAA 122s receptors (C and D), and direct activation of GABAA 122s receptors (E and F). Potencies are expressed as log(EC50), with EC50 in molar units. The inactive analogs () have been arbitrarily given a log(EC50) value of 1. The lines drawn through the data points on the graphs indicated linear regression of the relationship between the potencies of the active compounds () and the independent variable (log P or molecular volume). The correlations of the active compounds and the independent variables were subjected to statistical testing (under Materials and Methods), the results of which are indicated on the graphs. Statistical significance is indicated by *p < 0.05, **p < 0.01, or ***p < 0.001.

View larger version (18K): [in this window] [in a new window]   Fig. 9.   Cross-correlation of the potencies of the propofol analogs for producing loss of righting reflex in tadpoles, potentiation of GABA responses at GABAA 122s receptors, and direct activation of GABAA 122s receptors. Potencies are expressed as log(EC50), with EC50 in molar units. Only compounds for which EC50 values could be determined are included in these graphs (i.e., the active compounds). In each graph, the dashed line indicates the line of unity, whereas the solid line is from linear regression analysis. The correlations were subjected to statistical testing, the results of which are indicated on the graphs. Statistical significance is indicated by *p < 0.05, **p < 0.01, or ***p < 0.001.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

This is the first study to make rigorous comparisons over a chemically diverse series of propofol analogs between their anesthetic potency in vivo and their electrophysiological actions at GABA_A receptors. The SAR data for the propofol analogs reported here are clearly compatible with the hypothesis that GABA_A receptors play a significant role in the anesthetic actions of propofol and related analogs. The results also are consistent with previous in vivo studies in mice (James and Glen, 1980).

The best correlation was between the potencies for potentiation of GABA responses and loss of righting reflex in tadpoles, particularly because the 13 analogs that did not potentiate submaximal GABA currents also failed to produce loss of righting reflex in the tadpoles. Despite the overall correlation between the potencies for potentiation of GABA responses and loss of righting reflex in tadpoles (Fig. 9A), there are certainly compounds that deviate significantly from perfect linear correspondence. There are several possible explanations for this. The experimental design used in this study cannot, obviously, model all of the complexities of the actions of general anesthetics at intact neuronal circuits (Mody et al., 1994; Antkowiak, 1999). In addition, the distinction between potentiation of GABA responses and direct activation by propofol, which can be easily made in this in vitro system, is, in fact, somewhat artificial. In a living animal, both effects are expected to occur with some qualitative differences between the two effects. For instance, direct receptor activation would be expected to occur continuously at all GABA_A receptors, even those located extrasynaptically, whereas potentiating actions would only occur at synapses when GABA is present.

In addition, this study cannot address whether molecular targets other than GABA_A receptors contribute to the anesthetic effects of propofol and its analogs in tadpoles. So far, convincing effects of propofol at clinically relevant concentrations have only been demonstrated at GABA_A and strychnine-sensitive glycine receptors (for review, see Krasowski and Harrison, 1999; although see discussion about "clinically relevant concentrations" by Eckenhoff and Johansson, 1999). It is currently not known whether other molecular targets, such as the two-pore domain potassium channels recently shown to be sensitive to volatile anesthetics (Patel et al., 1999; Sirois et al., 2000), will also be shown to be sensitive to propofol.

The electrophysiological studies of native GABA_A receptors in dissociated tadpole spinal neurons are in agreement with previous reports demonstrating that the pharmacology of GABA_A receptors in frog neurons (Akaike et al., 1985) closely parallels that of mammalian GABA_A receptors. This manuscript is the first published report of the effects of propofol on GABA_A receptors expressed in neurons from tadpoles or frogs. The potentiation of GABA-evoked responses and direct activation by propofol, but not by 2,6-di-tert-butylphenol, in the tadpole spinal neurons is also consistent with the hypothesis that alteration of GABA_A receptor function is responsible, at least in part, for the anesthetic actions (or lack thereof) of propofol and related analogs in tadpoles.

One of the most striking observations from the SAR is that the presence of bulky or nonplanar alkyl substituents at positions 2 and 6 of the aromatic ring abolished activity of the 2,6-dialkylphenol analogs (Fig. 10). This is illustrated by the complete inactivity of 2,6-di-tert-butylphenol, 2,6-dicyclopentylphenol, and 2,6-dicyclohexylphenol. An intriguing finding is that 2,6-di-sec-butylphenol (an isomer of 2,6-di-tert-butylphenol) has high potency at GABA_A receptors and for loss of righting reflex in tadpoles. Similar findings were originally noted by James and Glen (1980), who demonstrated that 2,6-di-tert-butylphenol, 2,6-dicyclopentylphenol, and 2,6-dicyclohexylphenol were all ineffective as intravenous anesthetics in mice, whereas 2,6-di-sec-butylphenol was highly potent. To explain the inactivity of 2,6-di-tert-butylphenol, James and Glen (1980) suggested that the tert-butyl moieties crowd the phenol hydroxyl and interfere with a critical interaction between the phenol hydroxyl and the target receptor. Cyclopentyl and cyclohexyl groups would similarly "sterically hinder" the phenol hydroxyl more than a sec-butyl or isopropyl group.

View larger version (21K): [in this window] [in a new window]   Fig. 10.   Summary of the pattern of activity for the propofol analogs. Active here refers to analogs for which EC50 values could be determined for loss of righting in tadpoles, potentiation of GABA responses at GABAA 122s receptors, and direct activation of GABAA 122s receptors. Note that direct activation of GABAA 122s receptors was not detected for phenol, although phenol produced loss of righting reflex in tadpoles and potentiated GABA responses at GABAA 122s receptors. Phenol was classified as an active compound.

Although bulky alkyl groups at positions 2 and 6 of the aromatic ring interfere with the activity of the propofol analogs, the potencies of the propofol analogs clearly increase with the size of alkyl groups up to a point, as illustrated by the following series: phenol, 2,6-dimethylphenol, 2-isopropylphenol, 2,6-diethylphenol, propofol, and 2,6-di-sec-butylphenol. In fact, the potencies of the compounds within this small group increase almost linearly with molecular volume and log P, a finding also noted by James and Glen (1980). James and Glen (1980) considered only simple alkylphenols in their structure-activity studies, which probably accounts for their observation that log P correlated very strongly with in vivo anesthetic potency for the alkylphenols.

Log P and molecular volume did not correlate well with the biological potencies of the active compounds as an entire group (i.e., beyond the simple 2,6-dialkylphenols). For example, 2,6-di-sec-butylphenol and 2,4-di-sec-butylphenol have nearly identical log P values and molecular volumes yet have biological potencies that differ by more than one order of magnitude. Log P and molecular volume also do not explain the lack of activity exhibited by the remaining compounds such as 2,6-di-tert-butylphenol. These observations appear difficult to reconcile with a "nonspecific" or lipid-based mechanism of action for the propofol analogs, which would predict that compounds such as 2,6-di-tert-butylphenol would have high anesthetic potency. Compounds such as 2,6-di-tert-butylphenol and 2,6-dicyclopentylphenol may be analogous to the volatile "nonimmobilizers" (Koblin et al., 1994), in that these compounds have high lipid solubility, yet do not produce loss of righting reflex in tadpoles.

The SAR studies suggest a critical interplay between substituents at positions 1, 2, and 6 of the aromatic ring (Fig. 10). Removal of the phenolic hydroxyl group of propofol abolished activity (1,3-diisopropylbenzene), whereas bromide, isocyanate, and isothiocyanate could all substitute for the hydroxyl group, although this usually resulted in a drop in potency at GABA_A receptors or for producing tadpole loss of righting reflex (Fig. 10; Table 1). An interesting observation was that 2,6-diisopropylphenyl isocyanate and 2,6-diisopropylphenyl isothiocyanate were completely inactive both at the GABA_{A 122s} receptor and in tadpoles. The inactivity of 2,6-diisopropylphenyl isocyanate (molecular volume = 681 Å³) and 2,6-diisopropylphenyl isothiocyanate (713 Å³) is especially striking compared with the activities of propofol (641 Å³), 2,6-diethylphenyl isocyanate (602 Å³), and 2,6-diethylphenyl isothiocyanate (638 Å³). A key influence on the interactions of propofol with the target site, thus, appears to be the size and shape of the alkyl groups at positions 2 and 6 of the aromatic ring relative to the substituent at position 1.

Three other studies have examined the activity of propofol analogs at GABA_A receptors (Trapani et al., 1998; Sanna et al., 1999; Lingamaneni et al., 2001; for review, see Trapani et al., 2000). Most of the analogs analyzed by Trapani et al. (1998) involved groups added to position 4 of the propofol aromatic ring (i.e., the para-position). Two compounds analyzed in this study, 4-iodopropofol and 2-hydroxy-3-isopropylbenzoic acid, were also studied by Trapani et al. (1998). Most substitutions at the para-position are remarkably well tolerated, even groups as large as -CO(Phenyl). One possible explanation for this is that the substituents at the para-position do not form critical interactions with the target receptor and, thus, diverse chemical groups may be added to the para-position without abolishing activity.

We have previously reported that mutation of a methionine residue at position 286 of the GABA_{A 1}-subunit to tryptophan abolishes potentiation of GABA responses by propofol at GABA_A receptors (Krasowski et al., 1998). This methionine residue, thought to be near the interface of transmembrane domain 3 with the extracellular fluid (Williams and Akabas, 1999), is also necessary for the actions of volatile ether anesthetic and alcohols at GABA_A receptors (Mihic et al., 1997; Krasowski and Harrison, 2000). A potential synthesis of the available data is that propofol and related analogs interact with a binding site formed by the extracellular region of the transmembrane domains of the -subunit. Substituents at the para-position may "hang off" into the extracellular space, whereas positions 1, 2, and 6 of the propofol molecule form the major interactions with a conformationally restricted binding pocket formed in part by methionine 286 of the -subunit. This hypothesis will be tested in subsequent studies.