Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Benzodiazepines are used to treat anxiety disorders and epilepsy as well as in the induction of general anesthesia. The -aminobutyric acid (GABA) type A (GABA_A) receptors are the molecular targets for these clinical effects. GABA_A receptors are the major inhibitory neurotransmitter receptors in the central nervous system. Five homologous subunits arranged pseudosymmetrically around a central pore form the receptor-channel complex (Macdonald and Olsen, 1994; Nayeem et al., 1994; Karlin and Akabas, 1995). Each subunit has an ~200-amino acid extracellular N-terminal domain, four membrane-spanning segments (M1, M2, M3, and M4), a large intracellular loop between M3 and M4, and a short extracellular C terminus (Schofield et al., 1987; Macdonald and Olsen, 1994). Each subunit contributes to the channel lining that is largely formed by residues from the M2, and possibly from the M3, membrane-spanning segments (Xu and Akabas, 1996; Williams and Akabas, 1999). The subunit stoichiometry varies in different brain regions, but a common stoichiometry is two -, two -, and one -subunits (Chang et al., 1996; McKernan and Whiting, 1996; Tretter et al., 1997; Farrar et al., 1999).

Benzodiazepine binding allosterically potentiates GABA binding. This increases the currents induced by submaximal GABA concentrations (Macdonald and Olsen, 1994; Rabow et al., 1995; Sieghart, 1995; Hevers and Luddens, 1998). At the single channel level, benzodiazepines increase channel-opening rates (Study and Barker, 1981), whereas closing and desensitization rates are not altered (Rogers et al., 1994; Ghansah and Weiss, 1999). In hippocampal cells, the GABA_A receptor single channel conductance was reportedly increased by benzodiazepines (Eghbali et al., 1997), but in other cells this has not been observed (Study and Barker, 1981; Rogers et al., 1994).

The -subunit is critical for benzodiazepine potentiation of GABA-induced currents (Pritchett et al., 1989; Gunther et al., 1995). The main site of photoaffinity labeling by flunitrazepam, however, is in the -subunit at position ₁His101 (Stephenson et al., 1990; Duncalfe et al., 1996). This and other evidence suggests that the benzodiazepine-binding site is located in the extracellular domain at the interface between the - and -subunits (Duncalfe et al., 1996; Amin et al., 1997; Sigel and Buhr, 1997; Boileau et al., 1998). The GABA-binding sites are located in similar regions of the - and -subunit interface (Amin and Weiss, 1993; Smith and Olsen, 1995; Boileau et al., 1999).

Several classes of drugs interact with the benzodiazepine site (Macdonald and Olsen, 1994; Rabow et al., 1995; Sieghart, 1995; Hevers and Luddens, 1998). At low concentrations, flumazenil is a benzodiazepine antagonist and -carbolines are inverse agonists (Chan and Farb, 1985; Sigel and Baur, 1988). At higher concentrations, however, flumazenil and -carbolines potentiate GABA-induced currents. This potentiation presumably involves interactions at other sites on the GABA_A receptors (Chan and Farb, 1985; Sigel and Baur, 1988; Mehta and Ticku, 1989; Yakushiji et al., 1989; Stevenson et al., 1995). In some mutant GABA_A receptors, low concentrations of flumazenil and/or -carbolines cause benzodiazepine-like potentiation (Dunn et al., 1999).

In contrast to GABA and many general anesthetics, benzodiazepines do not directly open the channel gate (Macdonald and Olsen, 1994; Rabow et al., 1995), which is located in the membrane-spanning domain (Xu and Akabas, 1996). It is not known whether the benzodiazepine-induced conformational changes are confined to the extracellular domain or whether they extend into the membrane-spanning, channel-forming domains. Mutation of three residues in the extracellular loop connecting the M2 and M3 segments altered allosteric modulation by benzodiazepines but did not affect their binding (Boileau and Czajkowski, 1999). This and other evidence suggest that the transduction of benzodiazepine binding may involve conformational changes in the membrane-spanning domain.

Using the substituted-cysteine (Cys)-accessibility method, we previously showed that the ability of p-chloromercuribenzenesulfonate (pCMBS) to react with the ₁ M3 segment Cys-substitution mutants ₁F298C, ₁A300C, ₁L301C, and ₁E303C was state dependent, i.e., the residues were accessible only in the presence of GABA (Williams and Akabas, 1999). At a fifth position, ₁F296C, the effect of pCMBS applied in the presence of GABA was equivocal (as discussed below in the statistics section of Materials and Methods). In addition, pCMBS reacted with ₁A291C and ₁Y294C when applied in the presence or in the absence of GABA. We inferred that the GABA-induced conformational change creates a water-filled crevice(s) that extends from the extracellular surface into the interior of the membrane-spanning segments. This crevice allows extracellularly applied sulfhydryl reagents to gain access to these residues. We used the state-dependent accessibility of the M3 substituted Cys to probe for conformational changes induced by benzodiazepine binding. We show that benzodiazepine binding alters the accessibility of a subset of the residues that were reactive only in the presence of GABA.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Mutants and Expression The rat ₁M3 segment Cys-substitution mutants were generated and characterized previously (Williams and Akabas, 1999). In vitro mRNA transcription and Xenopus oocyte preparation and injection were as described previously (Xu and Akabas, 1993; Williams and Akabas, 1999). Oocytes were injected with 50 nl of a 200 pg/nl solution of subunit mRNA in a 1:1:1 ratio of ₁:₁:₂.

Electrophysiology Two-electrode voltage clamp recording from Xenopus oocytes and data acquisition and analysis were performed as described previously (Williams and Akabas, 1999). Oocytes were continuously perfused at 5 ml/min with calcium-free-frog Ringers (CFFR; 115 mM NaCl, 2.5 mM KCl, 1.8 mM MgCl₂, 10 mM HEPES, pH 7.5 with NaOH) at room temperature. The holding potential was 80 mV.

Experimental Protocols. The sulfhydryl-specific reagent used in these experiments was pCMBS; Sigma, St. Louis, MO). After reaction with a Cys pCMBS adds -HgC₆H₄SO₃ onto the reactive sulfhydryl.

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Effect of a 1-min coapplication of 0.5 mM pCMBS and 100 nM diazepam on M3 segment Cys-substitution mutants. Currents were recorded from individual oocytes expressing 1F296C (A), 1F298C (B), 1L301C (C), and 1E303C (D). The initial four traces for each mutant demonstrate that 100 nM diazepam potentiates GABA-induced currents and that diazepam alone does not induce a current. The final four traces for each mutant show the effect of a 1-min application of pCMBS and diazepam on the subsequent GABA-induced currents. For the mutants 1F296C (A), 1F298C (B), and 1L301C (C), subsequent currents were potentiated. For 1E303C (D), the subsequent currents were unchanged. The concentration of the GABA test pulses was 1 µM (A), 30 µM (B), and 30 µM (C). In D, the EC20 GABA concentration was 1 µM and the saturating GABA concentration was 100 µM. Bars above current traces indicate the period of application of the reagent indicated to the left. Traces are separated by 3- to 5-min washes with CFFR buffer to allow recovery from desensitization.

Measurement of Reaction Rates. Rates of reaction of pCMBS with the engineered Cys mutants were determined by the effect of sequential brief applications of pCMBS. A test pulse of GABA was applied to measure the GABA-induced current. GABA plus pCMBS (0.2-0.5 mM) was applied for 15 to 60 s. After washout of pCMBS, a test pulse of GABA was applied, and the GABA-induced current was measured. The effect of five to eight brief, sequential applications of pCMBS was determined. The magnitudes of the GABA test currents were normalized relative to the initial test pulse. The normalized current was plotted as a function of the cumulative duration of pCMBS treatment and fitted with a single exponential function using Prism2 software (GraphPad, San Diego, CA). The second order rate constant was calculated by dividing the pseudo-first order rate constant obtained from the exponential fit by the pCMBS concentration.

Statistics and Curve Fitting. Data are expressed as the percentage changes of current after modification ± S.E.M. The significance of differences between each mutant and wild type was determined by one-way ANOVA using the Student-Newman-Kuels post hoc test (SPSS for Windows, SPSS, Inc., Chicago, IL). Dose-response curves were fit using Prism2 software.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Reactions with pCMBS Applied in the Presence of Diazepam. The EC₅₀ for GABA activation of wild-type ₁₁₂ receptors was 2.1 ± 1.3 µM (n = 6) and for the Cys-substitution mutants ranged between 0.5 ± 0.1 µM for ₁A291C and ₁Y294C and 48 ± 15 µM for ₁L301C as reported previously (Williams and Akabas, 1999). For wild-type receptors, 100 nM diazepam potentiated the currents induced by an EC₃₀ GABA concentration by 26 ± 3% (n = 3). For the Cys-substitution mutants, potentiation by 100 nM diazepam ranged between 21 ± 3% (n = 4) for ₁V297C and 73 ± 24% (n = 4) for ₁Y294C (Table 1). Application of 100 nM diazepam alone did not elicit a current in wild type or any of the Cys-substitution mutants.

View this table: [in this window] [in a new window]   TABLE 2 Percentage change in the subsequent GABA-induced currents due to a 1-min application of 0.5 mM pCMBS in the presence of the indicated reagenta Data given as means ± S.E.M. (n).

Reaction of pCMBS Applied in the Presence of the Inverse Agonist -CCE. Application of 100 nM -CCE alone did not elicit a current in wild type or any of the Cys-substitution mutants; nor did it have an effect on subsequent GABA-induced currents. Coapplication of 100 nM -CCE had no significant effects on the current induced by an EC₃₀ GABA concentration in oocytes expressing either wild type or the Cys-substitution mutants (data not shown).

View larger version (14K): [in this window] [in a new window]   Fig. 2.   Effect of a 1-min application of 0.5 mM pCMBS applied in the presence of -CCE or flumazenil on the subsequent GABA-induced currents for the mutants 1F298C and 1L301C. Currents were recorded from oocytes expressing 1F298C (A and B) and 1L301C (C). The initial four traces for each mutant demonstrate that 100 nM -CCE (A) or 100 nM flumazenil (B and C) did not significantly effect the GABA-induced currents and that the drugs alone did not induce a current. The final four traces for each mutant show the effect of a 1-min application of pCMBS and drug on the subsequent GABA-induced currents. In all three cases the subsequent currents were potentiated. The GABA concentration was 30 µM. Bars above current traces indicate the period of application of the reagent indicated to the left. Traces are separated by 3- to 5-min washes with CFFR buffer to allow recovery from desensitization.

Reaction of pCMBS Applied in the Presence of Flumazenil. At concentrations below 10 nM, flumazenil behaves as a benzodiazepine antagonist; however, at concentrations greater than 1 µM, it potentiates GABA-induced currents (Chan and Farb, 1985; Sigel and Baur, 1988; Mehta and Ticku, 1989; Yakushiji et al., 1989; Stevenson et al., 1995). For wild-type and the Cys-substitution mutants, application of 100 nM flumazenil in the absence of GABA did not elicit any current. After washout of flumazenil, there were no effects on subsequent GABA-induced currents (data not shown). Furthermore, coapplication of 100 nM flumazenil with submaximal GABA concentrations did not significantly alter the magnitude of the GABA-induced currents (data not shown).

Reaction Rates of pCMBS with Accessible Cys-Substitution Mutants. The reaction rates of sulfhydryl reagents with substituted Cys residues depend on multiple factors, including time-averaged water surface accessibility, local steric factors, and local reagent concentration. In the aspartate chemotaxis receptor, reactivity of substituted Cys was well correlated with solvent accessibility calculated from the crystal structure (Danielson et al., 1997). Thus, the pCMBS reaction rates with the reactive Cys-substitution mutants were measured in the presence of GABA, diazepam, flumazenil, and -CCE.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Measurement of rate of reaction of pCMBS applied in the presence of 100 nM diazepam. A, currents recorded from an oocyte expressing 1F298C. Bars indicate period of application of GABA test pulses; arrows indicate application of 0.5 mM pCMBS plus 100 nM diazepam. The first two applications were for 30 s, and the subsequent five applications were for 60 s each. Current traces during application of pCMBS and diazepam are not shown. Each application of GABA and drugs were separated by 3-min washes with CFFR. B, plots of the percentage change in the GABA test peak currents relative to the initial peak currents as a function of the cumulative duration of application of pCMBS plus diazepam. The smooth line is a single exponential fit of the data.

View this table: [in this window] [in a new window]   TABLE 3 Second order reaction rates (l/mol-s) of pCMBS applied in the presence of the indicated reagent with the Cys-substitution mutants Data given as means ± S.E.M. (n).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In the presence of diazepam, pCMBS reacted with Cys substituted for three residues in the M3 membrane-spanning segment, ₁Phe296, ₁Phe298, and ₁Leu301. In the absence of diazepam or GABA, pCMBS did not react at a measurable rate with any of these Cys-substitution mutants. This implies that diazepam induced a conformational change in the GABA_A receptor ₁-subunit membrane-spanning domain to increase the reactivity of the engineered Cys at these positions. The residues accessible in the presence of diazepam are a subset of the M3 segment residues that became accessible in the presence of GABA (Table 4). In the presence of GABA, pCMBS reacted with the three M3 segment, diazepam-accessible residues plus two others, ₁Ala300 and ₁Glu303 (Williams and Akabas, 1999).

View this table: [in this window] [in a new window]   TABLE 4 Summary of the effect on subsequent GABA-induced currents of pCMBS application in the presence of various reagents

Several pieces of evidence suggest that pCMBS reaction in the diazepam-bound conformation is not occurring because of a diazepam-induced increase in the spontaneous open probability of the receptor. First, at the three reactive residues the reaction rates were similar in the presence of diazepam and GABA. This suggests that the reactive engineered Cys residues were spending a similar percentage of time on the water-accessible surface with diazepam and GABA. Second, the residue that reacted fastest in the presence of GABA, ₁E303C, did not react in the presence of diazepam. If reaction were occurring in a short-lived state structurally similar to those occurring in the presence of GABA this residue should have reacted. Third, the effect of pCMBS modification of ₁F296C was different in the presence of GABA and diazepam. In the presence of GABA, covalent modification of ₁F296C caused inhibition of subsequent GABA-induced currents, whereas in the presence of diazepam, modification caused potentiation of subsequent currents. Thus, pCMBS modification appeared to trap the receptor in distinct conformations depending on whether pCMBS was applied in the presence of diazepam or GABA. We infer that the position of ₁F296C relative to the rest of the protein is different in the presence of the two reagents. A similar phenomenon, opposite effects on subsequent currents, was observed when the nicotinic acetylcholine receptor mutant S252C was covalently modified in the presence and in the absence of acetylcholine (Akabas et al., 1994). Thus, we conclude that diazepam binding induces a new conformation of the receptor that is structurally distinct from the conformations induced by GABA binding.

Our previous results suggested that GABA-induced gating produced changes in the conformations of both the M2 and M3 membrane-spanning segments (Xu and Akabas, 1996; Williams and Akabas, 1999). In addition to opening the channel, this GABA-induced conformational change appears to create a water-filled crevice into the interior of the membrane-spanning domain to expose the residues in the M3 segment and allow pCMBS penetration into the interior of the membrane-spanning domain. Creation of this water-filled crevice suggests that the membrane-spanning segments are less tightly packed in the presence of GABA and diazepam compared with the closed state of the receptor. Whether this crevice extends from the extracellular surface or from the channel lumen is uncertain (Williams and Akabas, 1999). The diazepam-induced conformational change creates a more limited water-filled crevice that also extends into the interior of the membrane-spanning domain in the region of the ₁ M3 segment, thus allowing pCMBS access to only three of the M3 residues. This more limited diazepam-induced conformational change is insufficient to open the channel gate. The diazepam-bound conformation may represent an intermediate state between the closed and open states: the conformation of the GABA-binding site has been shifted to the higher affinity state and the membrane-spanning domain has undergone a portion of the changes induced by GABA activation. These conformational changes would permit easier activation by GABA and suggests that diazepam may increase the spontaneous open probability of the channel.

The transduction of benzodiazepine binding in the extracellular domain to the conformational changes in the vicinity of the M3 segment may involve interactions between the extracellular domain and the M2-M3 loop. Consistent with this hypothesis, Boileau and Czajkowski (1999) reported that benzodiazepine potentiation, but not binding, was reduced by mutation of three residues in the ₂-subunit M2-M3 loop. Similar interactions also appear to be important for agonist efficacy. Mutations of M2-M3 loop residues altered agonist efficacy in the GABA, glycine, and acetylcholine receptors (Campos-Caro et al., 1996; Lynch et al., 1997; Sigel et al., 1999). A further indication of the importance of interactions between the M2 and M3 segments in channel gating is that mutation of residues near the extracellular ends of the GABA_A receptor M2 and M3 segments, aligned with ₁Ser270 and ₁Ala291, alter the ability of general anesthetics to potentiate GABA-induced currents (Belelli et al., 1997; Mihic et al., 1997). These two residues are on the water-accessible surface of the protein (Williams and Akabas, 1999), although ₁Ser270 is not on the channel-lining face of the M2 segment (Xu and Akabas, 1996). Thus, these two residues may form a portion of the lining of a water-filled crevice that extends into the interior of the protein. Anesthetics may act by intercalating into the water-filled crevice between these two segments, destabilizing interactions between the M2 and M3 segments, thereby initiating a conformational change similar to that induced by GABA (Belelli et al., 1997; Mihic et al., 1997). Alternatively, these residues may be part of the machinery that transduces the effects of anesthetic binding rather than part of an anesthetic binding site. Further studies will be necessary to determine whether potentiating concentrations of general anesthetics induced conformational changes similar to those induced by diazepam or whether the molecular basis of potentiation by different drugs is different.

It was surprising that in the presence of both flumazenil and -CCE pCMBS reacted with some of the M3 substituted Cys residues that were accessible in the presence of diazepam. In the presence of 100 nM flumazenil, pCMBS reacted with ₁F298C and ₁L301C. In the presence of a 10-fold lower flumazenil concentration (10 nM) pCMBS still reacted with ₁L301C but not with ₁F298C (Table 2). At low concentrations, below 10 nM, flumazenil acts as a benzodiazepine antagonist, but at higher concentrations (>1 µM), it acts as a weak potentiator of GABA-induced currents (Chan and Farb, 1985; Sigel and Baur, 1988; Mehta and Ticku, 1989; Yakushiji et al., 1989). In addition, some experiments have suggested that even at low concentrations flumazenil is not a strict antagonist (Chiu and Rosenberg, 1983). The potentiating activity of flumazenil presumably results from binding at a site distinct from the benzodiazepine-binding site. The location of the potentiating binding site has not been identified. It is uncertain whether the conformational changes induced by flumazenil occur because of binding at the benzodiazepine site or at the other site; however, we did not observe an effect of 100 nM flumazenil on the GABA-induced current, although this concentration was sufficient to increase the accessibility of the two residues. Similarly, 100 nM -CCE increased the accessibility of ₁F298C to react with pCMBS, whereas 10 nM -CCE did not (Table 2). -CCE is a benzodiazepine inverse agonist at low concentrations (>10 nM) and potentiates GABA-induced currents at higher concentrations (>1 µM), consistent with binding at two sites of different affinity (Sigel and Baur, 1988; Yakushiji et al., 1989; Stevenson et al., 1995). Thus, for flumazenil and -CCE, it is not possible to determine which binding site is mediating their actions to increase the accessibility of these M3 residues. Further complicating the interpretation of the flumazenil and -CCE results is the possibility that some of the Cys mutations might alter the affinity or the efficacy of action at one binding site but not at the other. Alternatively, flumazenil binding at the benzodiazepine-binding site may induce conformational changes in the membrane-spanning domains that are not sufficiently extensive to result in alteration of GABA-induced currents but are sufficient to alter the accessibility of some of the M3 substituted Cys mutants. This raises the question as to whether flumazenil is a strict benzodiazepine antagonist causing no conformational change but blocking the binding site, or whether flumazenil binding does alter the receptor conformation but the energy imparted to the receptor by flumazenil binding is insufficient to cause the total conformational change necessary for potentiation of GABA-induced currents.

In summary, diazepam binding induces a conformational change in the GABA_A receptor. This conformational change extends into the membrane-spanning domain of the receptor. Diazepam does not simply induce a state similar to one of the predominant GABA bound states, open or desensitized. Rather, our results indicate that the diazepam bound state is a distinct structural state of the receptor. This has important implications for potential kinetic models of the actions of benzodiazepines on GABA_A receptors.