HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 21, 18785-18792, May 24, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/21/18785    most recent M111778200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Holden, J. H. Articles by Czajkowski, C. Search for Related Content PubMed PubMed Citation Articles by Holden, J. H. Articles by Czajkowski, C. Social Bookmarking                      What's this?

Different Residues in the GABA_A Receptor ₁T60-₁K70 Region Mediate GABA and SR-95531 Actions^*

Jessica H. Holden and Cynthia Czajkowski

From the Department of Physiology and Molecular and Cellular Pharmacology Program, University of Wisconsin, Madison, Wisconsin 53706

Received for publication, December 10, 2001, and in revised form, March 13, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Although -aminobutyric acid type A receptor agonists and antagonists bind to a common site, they produce different conformational changes within the site because agonists cause channel opening and antagonists do not. We used the substituted cysteine accessibility method and two-electrode voltage clamping to identify residues within the binding pocket that are important for mediating these different actions. Each residue from ₁T60 to ₁K70 was mutated to cysteine and expressed with wild-type ₂ subunits in Xenopus oocytes. Methanethiosulfonate reagents reacted with ₁T60C, ₁D62C, ₁F64C, ₁R66C, ₁S68C, and ₁K70C. -Aminobutyric acid (GABA) slowed methanethiosulfonate modification of ₁F64C, ₁R66C, and ₁S68C, whereas SR-95531 slowed modification of ₁D62C, ₁F64C, and ₁R66C, demonstrating that different residues are important for mediating GABA and SR-95531 actions. In addition, methanethiosulfonate reaction rates were fastest for ₁F64C and ₁R66C, indicating that these residues are located in an open, aqueous environment lining the core of the binding pocket. Positively charged methanethiosulfonate reagents derivatized ₁F64C and ₁R66C significantly faster than a negatively charged reagent, suggesting that a negative subsite important for interacting with the ammonium group of GABA exists within the binding pocket. Pentobarbital activation of the receptor increased the rate of methanethiosulfonate modification of ₁D62C and ₁S68C, demonstrating that parts of the binding site undergo structural rearrangements during channel gating.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Few studies of ligand-gated ion channels (LGICs)¹ have addressed the question of how the binding of compounds with divergent structure leads to dramatic functional differences. For example, agonist binding induces conformational changes that result in channel opening, whereas binding of competitive antagonists does not. Distinguishing the specific amino acid residues involved in the binding of agonists and antagonists will help to elucidate the structural rearrangements that govern the pharmacological effects of these compounds. In this paper, we examined the molecular determinants important for the binding of the agonist GABA and the competitive antagonist SR-95531 to the -aminobutyric acid type A (GABA_A), receptor and explored the conformational changes that occur within the GABA-binding site during channel activation by a barbiturate.

GABA_A receptors are heteropentameric chloride channels that mediate fast synaptic inhibition in the brain and are members of an evolutionarily related superfamily of LGICs that also includes nicotinic acetylcholine, glycine, and serotonin-type 3 receptors (1). To date, 16 different GABA_A receptor subunit isoforms (_1-6, _1-3, _1-3, , , , and ) have been cloned (2-7). Most native receptors are thought to contain , , and  subunits (8) in a 2:2:1 stoichiometry (9), although functional channels that lack benzodiazepine modulation can be formed without the  subunit (10, 11).

The neurotransmitter recognition site, where agonists such as GABA and muscimol and antagonists such as SR-95531 and bicuculline bind, is located at the interface between the  and  subunits because residues have been identified on both subunits that are important for ligand recognition. On the ₁ subunit, residues identified include Phe⁶⁴ (12, 13), Arg⁶⁶, Ser⁶⁸ (14), Arg¹¹⁹, and Ile¹²⁰ (15, 16). On the ₂ subunit, residues Tyr¹⁵⁷, Thr¹⁶⁰ (17), Thr²⁰², Ser²⁰⁴, Tyr²⁰⁵, Arg²⁰⁷, and Ser²⁰⁹ (17, 18) have been identified. Based on work on the related nicotinic acetylcholine receptor, residues that contribute to forming the binding site are located in at least six different non-contiguous extracellular N-terminal regions of the  and  subunits. These regions have been designated loops A-F (19). Residues within these loops likely have different functional roles. Some residues may directly contact ligand, some may be important for maintaining the structural integrity of the binding site, and others may mediate local conformational movements within the site.

In the present study, we examined the binding site region surrounding ₁F64 (loop D) of the GABA_A receptor. In the homologous region of the serotonin-type 3 receptor, White and colleagues (20) used alanine-scanning mutagenesis and determined that different amino acid residues contribute to the binding of agonists and antagonists. We hypothesized that the region surrounding ₁F64 of the GABA-binding site also contains unique residues important for agonist and antagonist binding, and we tested this hypothesis by using the substituted cysteine accessibility method (SCAM).

SCAM has been used on a variety of ion channels to elucidate channel lining and binding site residues, to determine the location of channel gates and selectivity filters, and to identify regions of the protein that are involved in conformational rearrangements during state changes (21). In this method, individual amino acid residues are mutated to cysteine, and the ability of sulfhydryl-specific reagents to modify covalently each introduced cysteine is assessed by observing the effect of the reagent on receptor function. We measured the rates of sulfhydryl modification of accessible introduced cysteine residues in the presence and absence of GABA and SR-95531. We identified a subsite important for agonist binding that includes ₁F64, ₁R66, and ₁S68 and an antagonist-binding subsite that includes ₁D62, ₁F64, and ₁R66. In addition, we used sulfhydryl-specific reagents of different charge and determined that a negative subsite exists within the binding pocket. Finally, we measured rates of sulfhydryl modification in the presence of pentobarbital (a GABA_A receptor modulator that opens the channel), and we identified conformational changes that occur within the GABA-binding site during channel activation.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Site-directed Mutagenesis-- The ₁ cysteine mutants were engineered using the Altered Sites II® in vitro Mutagenesis Systems (Promega Corp., Madison, WI) or by recombinant PCR as described previously (14, 22). Cysteine substitutions were made in the rat ₁ subunit at positions Tyr⁵⁹, Thr⁶⁰, Ile⁶¹, Asp⁶², Val⁶³, Phe⁶⁴, Phe⁶⁵, Arg⁶⁶, Gln⁶⁷, Ser⁶⁸, Trp⁶⁹, and Lys⁷⁰, where the number reflects the position in the mature ₁ subunit protein. The cysteine mutants were subcloned into pGH19 (23, 24) for expression in Xenopus laevis oocytes. The presence of the mutations was verified by restriction endonuclease digestion and double-strand cDNA sequencing. The mutants have been named, using the single letter amino acid code, as wild-type residue, residue number, and mutated residue.

Expression in Oocytes-- X. laevis oocytes were prepared as described previously (25). cRNA transcripts were generated using the mMessage T₇ kit (Ambion, Austin, TX). GABA_A receptor rat ₁ or ₁ mutants were expressed with wild-type rat ₂ subunits by injection of cRNA into oocytes (0.3 ng of cRNA/subunit/oocyte, except for ₁F64C₂ and ₁R66C₂ that were injected at 7 ng of cRNA/subunit to ensure high levels of receptor expression). Mean maximal responses to GABA ranged from 1 to 10 µA. The oocytes were stored in ND96 medium (in mM: 96 NaCl, 2 KCl, 1 MgCl₂, 1.8 CaCl₂, and 5 HEPES, pH 7.4) supplemented with 100 µg/ml gentamicin and 100 µg/ml bovine serum albumin for 2-14 days and used for electrophysiological recordings.

Voltage Clamp Analysis-- Oocytes under two-electrode voltage clamp (V_hold of 80 mV) were continuously perfused with ND96 at a rate of 5 ml/min. The bath volume was 200 µl. GABA, SR-95531 (Sigma), pentobarbital (Research Biochemicals, Natick, MA), and methanethiosulfonate (MTS) reagents (Toronto Research Chemicals, Toronto, Ontario, Canada) were dissolved in ND96. Standard two-electrode voltage clamp recording was carried out using a GeneClamp 500 amplifier (Axon Instruments, Foster City, CA) interfaced to a computer with a Digidata 1200 (Axon Instruments). Electrodes were filled with 3 M KCl and had resistances of 0.5-2.0 megaohms in ND96. Data acquisition and analysis were performed using pClamp 6 (Axon Instruments).

Pulse Protocol for Measuring MTS Effects-- The sulfhydryl-specific reagents used were derivatives of MTS obtained from Toronto Research Chemicals (Toronto, Ontario, Canada). The reagents used were 2-aminoethyl methanethiosulfonate (MTSEA), 2-(trimethylammonium)ethyl methanethiosulfonate (MTSET), and 2-sulfonatoethyl methanethiosulfonate (MTSES). All oocytes were stabilized before addition of MTS reagent by application of GABA (5 s) at 10-min intervals until the GABA-activated peak currents (I_GABA) varied by <10%. GABA concentrations used were EC₄₀-EC₆₀ for each mutant. After the GABA response stabilized, freshly diluted MTS reagent was applied for 2 min; the cell was washed for 5 min, and then GABA was applied at the same concentration used before the MTS treatment. MTSEA (2 mM), MTSET (2 mM), or MTSES (5 mM) were used. The effect of the MTS reagent was calculated as (I_GABA-post/I_GABA-pre)  1, where I_GABA-post is the current elicited by GABA after MTS application, and I_GABA-pre is the current elicited by GABA before MTS application.

Rate of MTS Modification-- The rate of MTS reagent covalent modification of introduced cysteines was determined by measuring the outcome of sequential applications of MTS reagents on I_GABA. The protocol was as follows: EC₂₀-EC₆₀ GABA was applied for 5 s; the cell was washed for 30 s; MTS reagent was applied for 5-20 s; the cell was washed for 2.5 min; and the procedure was repeated until I_GABA no longer changed indicating that the reaction was complete. Before the rate of MTS modification was measured, GABA was applied every 3 min until I_GABA stabilized to within 3% demonstrating that the observed changes in I_GABA after application of MTS reagent were due to the effects of the MTS reagent. Concentration of MTS reagent and time of application varied as follows: ₁D62C: MTSEA, 1 mM, 20 s; ₁F64C: MTSES, 10 µM, 5 s; MTSET, 10 nM, 5 s; MTSEA, 100 nM, 5 s; ₁R66C: MTSES, 500 µM, 20 s; MTSET, 1 µM, 5 s; MTSEA, 10 µM, 10 s; ₁S68C: MTSES: 150 µM, 10 s; MTSET: 100 µM, 5 s; MTSEA: 100 µM, 5 s. The effects of agonists and antagonists on the rate of MTS modification were tested by co-applying GABA (EC₈₅-EC₉₅), SR-95531 (IC₉₀-IC₉₅), or pentobarbital (500 µM) with MTSES for all mutants except ₁D62C, in which case they were co-applied with MTSEA. For these studies, I_GABA was stabilized before the rate of MTS reaction was measured as follows: apply GABA (EC_20-60) for 5 s, wash for 30 s, apply GABA, SR-95531, or pentobarbital at high concentration for 5-20 s, wash for 2.5 min, and repeat the procedure. This procedure was repeated until the peak of the GABA (EC_20-60) current was within 3% of the previous GABA (EC_20-60) current peak.

For all rate experiments, the decrease in current was plotted versus cumulative time of MTS exposure. We assume that the concentration of MTS reagent does not change significantly during the reaction, and thus, we can determine a pseudo first-order rate constant from the rate of decrease in I_GABA. Peak current at each time point was normalized to the initial peak current, and a pseudo first-order rate constant (k₁) was determined by fitting the data with a single exponential decay equation: y = span·e^kt + plateau. Because the data are normalized to I_GABA at time 0, span = 1  plateau. The second-order rate constant (k₂) for MTS reaction was determined by dividing the calculated pseudo first-order rate constant by the concentration of MTS reagent used (26). To verify the accuracy of this protocol, second-order rate constants were determined using at least two different concentrations of MTS reagents for several mutants.

EC₅₀ Analysis-- Concentration-response experiments were performed as described previously (14). In brief, these trials used a low concentration of GABA (EC₂-EC₇) immediately before the test concentration of agonist to correct for any slow drift in GABA responses that may occur during the experiment. Currents elicited by each test concentration were normalized to the corresponding low concentration current before curve fitting. Concentration-response data were fit to the following equation: I = I_max/(1 + (EC₅₀/[A])ⁿ), where I is the peak response to a given concentration of GABA; I_max is the maximum amplitude of current; EC₅₀ is the concentration of GABA that produces a half-maximal response; [A] is the concentration of GABA; and n is the Hill coefficient.

IC₅₀ Analysis-- IC₅₀ values were measured as described previously (18). SR-95531 IC₅₀ values were measured by applying a fixed concentration of GABA (EC₂₀-EC₆₀) immediately followed by co-application of the same concentration of GABA and a test concentration of SR-95531. Inhibition was calculated as I_{GABA + SR-95531}/I_GABA. Data were fit to the following equation: inhibition = 1  1/(1 + (IC₅₀/[Ant])ⁿ), where IC₅₀ is the concentration of antagonist that blocks half of I_GABA; [Ant] is the concentration of antagonist, and n is the Hill coefficient. K_I values were calculated using the Cheng-Prusoff/Chou equation (27, 28): K_I = IC₅₀/(1 + [A]/EC₅₀), where [A] is the concentration of GABA used, and EC₅₀ is the concentration of GABA that elicits a half-maximal response.

Statistical Analysis-- Data analysis was carried out using nonlinear regression analysis included in the GraphPad Prism software package (San Diego, CA; www.graphpad.com). Statistical analysis was conducted using a one-way analysis of variance, followed by a post hoc Dunnett's test.

Measurement of Length of MTS Reagents, GABA, and SR-95531-- All compounds were measured after energy minimization (<0.5 kcal/Å; Chemsketch, ADC, Toronto, Ontario, Canada). All MTS regents were measured from the sulfur to the center of the base of the tetrahedron formed by the terminal tertiary group. GABA was measured from the nitrogen to the base of the tetrahedron formed by the carboxyl group. SR-95531 was measured from the carbon of the methyl group to the center of the base of tetrahedron formed by the carboxyl group.

Structural Modeling-- The mature protein sequences of the rat ₁ and ₂ subunits were homology modeled with a subunit of the acetylcholine-binding protein (AChBP) (29). The crystal structure of the AChBP was downloaded from the RCSB Protein Data Bank (code 1I9B) and loaded into Swiss Protein Data bank Viewer (SPDBV, ca.expasy.org/spdbv). The ₁ protein sequence from Thr¹²-Ile²²⁷ and the ₂ protein sequence from Ser¹⁰-Leu²¹⁸ were aligned with the AChBP sequence using the alignment function of SPDBV. The aligned sequences of the ₁ and ₂ subunits were threaded onto an AChBP subunit using the "Interactive Magic Fit" function of SPDBV. The threaded subunits were imported into SYBYL (Tripos, Inc., St. Louis, MO) where energy minimization was carried out with the first 100 iterations carried out using Simplex minimization followed by 10,000 iterations using the Powell method.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Modification of Cysteine Mutants by MTS Reagents-- We reported previously that when mutated to cysteine, alternating residues from ₁T60 to ₁S68 are accessible to covalent modification by MTSEA-biotin, suggesting that this region of the GABA-binding site forms a -strand (14). We also determined that the presence of GABA inhibits the reaction of MTS compounds at ₁F64C, ₁R66C, and ₁S68C, indicating that these residues may face into the agonist-binding pocket. In the present study, we used MTS reagents of different size and charge to explore the physicochemical environment of this region of the GABA-binding site. The MTS reagents used were MTSEA (3.7 Å long), which covalently adds a positively charged ethyl-ammonium group, MTSET (4.5 Å), which adds a positively charged ethyl-trimethylammonium group, and MTSES (4.8 Å), which adds a negatively charged ethyl-sulfonate group (Fig. 1).

View larger version (29K): [in this window] [in a new window]   Fig. 1.   Effects of MTS reagents on wild-type and mutant GABAA receptors. A, structures and lengths of MTS reagents. Shown are the portions of the MTS reagents that covalently modify an introduced cysteine. Lengths were measured after energy minimization (<0.5 kcal/Å; Chemsketch, ADC, Toronto, Ontario, Canada). B, amino acid residues 1T60-1K70 were individually mutated to cysteine and expressed with 2 subunits in Xenopus oocytes. By using two-electrode voltage clamping, the accessibility of the introduced cysteines to MTSEA, MTSET, and MTSES was examined. The absolute change in GABA-mediated current after MTS treatment is plotted for wild-type (WT) and mutant receptors (percent effect = ( 1 (IGABA, after/IGABA, before)) × 100 ). , percent effect reflects inhibition of current for all mutants except T60C where currents were increased after application of MTS reagents. *, cysteine substitution was not tolerated at these positions.

In order to determine the ability of the MTS reagents to react with each introduced cysteine, mutant ₁ and wild-type ₂ subunits were co-expressed in Xenopus oocytes, and I_GABA (EC_40-60) was measured before and after a 2-min MTS application. Because the MTS reagents did not affect the amplitude of I_GABA at wild-type ₁₂ receptors, we assumed that current changes observed in mutant receptors were due to covalent modification of the introduced cysteine residues (Fig. 1). In general, the residues that were reported previously (14) to be modified by MTSEA-biotin (₁T60C, ₁D62C, ₁F64C, ₁R66C, and ₁S68C) were also accessible to modification by MTSEA, MTSET, and MTSES (Fig. 1). Reaction with MTS reagents reduced GABA current by 14 (₁K70C, MTSET) to 96% (₁F64C, MTSEA). In contrast to all other mutants, covalent modification of ₁T60C caused an increase in I_GABA suggesting that the GABA EC₅₀ value for this mutant receptor decreases following covalent modification.

At any given position, the magnitude of the MTS effect on I_GABA was dependent on the specific MTS reagent used (Fig. 1). The observed differences in MTS effects may be due to the charge and/or size of the functional group tethered within the binding site. However, it is also possible that the MTS reactions did not go to completion due to their varied intrinsic reactivities (21). To test this possibility, we measured the rate at which each MTS reagent modified ₁D62C, ₁F64C, ₁R66C, and ₁S68C, making sure that each reaction was followed to completion.

MTS Reaction Rate Constants-- The rates of covalent modification of an introduced cysteine were obtained by measuring the effect of successive subsaturating applications of each MTS reagent on I_GABA (Fig. 2A). The decrease in I_GABA was plotted versus cumulative duration of MTS exposure and fit with a one-phase exponential decay curve, which yields a pseudo first-order rate constant (k₁). To correct for the concentration dependence of the rate, a second-order rate constant (k₂, Table I) was calculated by dividing k₁ by the concentration of MTS used ("Experimental Procedures"). In general, the maximal effects of the MTS reagents observed in the rate experiments were consistent with those measured in the 2-min pulse protocol (mutant: MTSEA maximal inhibition/MTSEA 2-min inhibition; ₁D62C: 64/54%; ₁F64C: 92/96%; ₁R66C: 32/33%; ₁S68C: 34/32%). Because the reactions went to completion, these data indicate that tethering groups of different size and charge to the mutant receptors differentially affects I_GABA.

View larger version (11K): [in this window] [in a new window]   Fig. 2.   Measurement of MTS reaction rates. A, representative current trace recorded while measuring the reaction rate of MTSES at 1R66C2 receptors. Downward deflections represent inward currents elicited by 6-s applications of 1 mM GABA (~EC25). Bars indicate the time of application of GABA or MTSES. MTSES (500 µM) was applied for 20 s each time. B, the logarithms of the ratio of the second-order rate constants (k2) of the positively charged MTSET over the negatively charged MTSES for 2-ME, 1F64C, 1R66C, and 1S68C are shown. All data represent the mean ± S.E. of at least 4 experiments. The k2 values are listed in Table I. Cationic MTSET reacts significantly faster than anionic MTSES at 1F64C and 1R66C. The rates for 2-ME were reported by Karlin and Akabas (21).

View this table: [in this window] [in a new window]   Table I Rates of reaction of MTSES, MTSET, and MTSEA at 1D62C2, 1F64C2, 1R66C2, and 1S68C2 receptors Rates of covalent modification of cysteine-containing receptors were measured as described under "Experimental Procedures." k2 values represent mean second-order rate constants ± S.D. of at least three experiments. The free solution (free sol.) rates were reported by Karlin and Akabas (21) and reflect the rate at which each MTS compound reacts with 2-mercaptoethanol, in solution.

The rate of reaction of MTS modification of a binding site engineered cysteine depends on several factors as follows: 1) the movement of the MTS reagent from bulk solution to the substituted cysteine in the binding pocket (permeability of the pathway); 2) the intrinsic electrostatic potential within the pocket and along the pathway; 3) the ionization (acid dissociation) of the substituted cysteine's sulfhydryl group; and 4) steric restrictions in forming an activated complex between the thiolate of the substituted cysteine and the MTS reagent. MTS reagents react preferentially with the ionized thiolate (RS) form of cysteine (30, 31). Of the residues tested, covalent modification of ₁F64C was the fastest, indicating that this is the most accessible residue in loop D. For example, MTSET modified ₁F64C with a k₂ of ~5,500,000 M¹ s¹, which was about 340-fold faster than reaction at ₁R66C. Reaction at ₁R66C, in turn, was about 40-fold faster than modification at ₁S68C (Table I). At ₁D62C, MTSEA was the only reagent tested that significantly altered I_GABA. There are two possible explanations for this result. Either MTSET and MTSES do not react with ₁D62C or covalent modification by these reagents does not change I_GABA, implying that any apparent modification is functionally silent. To test these possibilities, we measured the ability of MTSEA to modify covalently ₁D62C after application of MTSET or MTSES. If MTSET or MTSES modified ₁D62C, then reaction with MTSEA should not occur, and no change in I_GABA should be observed. Application of MTSES or MTSET prior to MTSEA had no effect on the ability of MTSEA to inhibit I_GABA (data not shown), indicating MTSET and MTSES do not react with ₁D62C. It should be noted that the reaction rate of MTSEA with ₁D62C was very slow (k₂ = 16 M¹ s¹) indicating that ₁D62C has limited accessibility.

In free solution, the rates of MTSEA, MTSET, and MTSES with 2-mercaptoethanol (2-ME) are 76,000, 212,000, and 17,000 M¹ s¹, respectively (21) (Table I). The rate constants depend on the charges of the reactants. Because the net charge of 2-ME is 1, positively charged MTS reagents react faster than negatively charged MTS reagents with this compound (31). Interestingly, the reaction rate constants of MTSET and MTSEA with ₁F64C were ~30-fold faster than the their rates of reaction with 2-ME in free solution (Table I). The rate of reaction of the MTS reagents with ₁F64C is influenced by the intrinsic electrostatic potential of the GABA-binding site, which arises from fixed charges and dipoles in the protein. The faster rates of MTSET and MTSEA modification of ₁F64C compared with the rates of modification of a simple thiol in solution are likely due to these intrinsic properties of the protein and suggest that the short range interactions of MTSET and MTSEA with the GABA-binding site are stronger than those with a simple thiol. Similar fast rates were measured for MTSET and MTSEA reaction with the acetylcholine-binding site cysteines, C192/193, in reduced, wild-type Torpedo nicotinic acetylcholine receptors, k₂ ~3 × 10⁶ M¹ s¹ (31).

Intrinsic Negative Electrostatic Potential in the GABA-binding Site-- The intrinsic electrostatic potential at a substituted cysteine can be examined by determining the rate of reactions of MTS reagents that differ in charge (26, 31, 32). We examined the electrostatic potential at ₁F64C, ₁R66C, and ₁S68C by comparing the rates of reaction of the positively charged MTSET and the negatively charged MTSES (Fig. 2B). Because MTSET and MTSES are approximately equivalent in size and have a common reaction mechanism, differences in their respective rates of reaction at a given residue are likely due to their opposite charges. The second-order rate constants for MTSET modification of ₁F64C, ₁R66C, and ₁S68C were 235-, 2320-, and 10.4-fold faster than that for MTSES, respectively (Table I and Fig. 2B). In comparison, the second-order rate constant for MTSET modification of 2-ME is 12.5-fold faster than that for MTSES. To factor out the intrinsic differences in the reactivities of the two MTS reagents and the extent of ionization of the respective thiols, we divided the ratio of the rates of the two reagents at an introduced cysteine by the ratio of the rates for the two reagents with 2-ME (31-33). For ₁F64C, the ratio of ratios is  = 235/12.5 = 18.8. For ₁R66C and ₁S68C,  = 185.6 and 0.84, respectively. A ratio of ratios that is significantly larger than one indicates that there is a negative potential experienced by that thiol. A ratio of ratios of ~1 indicates that there is no charge selectivity for the reaction with this residue. We can estimate the effective electrostatic potential at an introduced cysteine as shown in Equation 1,

(Eq. 1)

where z is the charge of the MTS reagent; R is the gas constant; T is absolute temperature, and F is Faraday's constant (31, 32). The calculated electrostatic potential at ₁F64C is 37 mV and the potential at ₁R66C is 66 mV. The negative potential may be smaller at ₁F64C because of the nearby positively charged arginine at position 66. The data indicate that there is a substantial negative potential experienced by ₁F64C and ₁R66C and that a negative subsite exists within the GABA-binding pocket that interacts with the positive charge on MTSET and MTSEA during their reaction with ₁F64C or ₁R66C.

Expression and Functional Analysis of ₁R66 Mutations-- Because GABA is zwitterionic, it is plausible that both a positively and a negatively charged subsite are involved in its binding. One residue within loop D that could be part of a positive subsite is ₁R66. Previously, we determined that cysteine substitution at ₁R66 increased the GABA EC₅₀ value 320-fold (14). We mutated ₁R66 to other residues including alanine, histidine, leucine, serine, glutamine, and the positively charged lysine. In each case, the GABA EC₅₀ values were increased by more than 2 orders of magnitude compared with wild-type receptors (R66A, 1000 ± 510 µM; R66H, 6200 ± 1500 µM; R66L, 1400 ± 400 µM; R66S, 2700 ± 375 µM; R66Q, 5300 ± 1700 µM; R66K, 5600 ± 350 µM; WT, 8.2 ± 0.4 µM). We calculate that the change in free energy due to ₁R66 mutation is ~3-4 kcal/mol. These data suggest that the positively charged arginine at position 66 may play a critical role in agonist binding.

We also measured the K_I values for the antagonist SR-95531 in oocytes expressing ₁D62C, ₁F64C, ₁R66C, or ₁S68C. Whereas mutations at ₁F64 and ₁R66 altered GABA EC₅₀ values, SR-95531 K_I values were only altered by cysteine substitution at ₁F64 (180-fold, Table II and Fig. 3). Because SR-95531 is a larger molecule than GABA, it is likely to be stabilized by different amino acid residues within the binding pocket. These data indicate that within the GABA-binding site there are distinct residues important for agonist and antagonist binding. In this paper, the term "GABA subsite" refers to residues within the overall binding pocket that are important for GABA binding. Similarly, the term "SR-95531 subsite" refers to residues within the pocket that contribute to SR-95531 binding. There may be residues within the binding pocket that are involved in both GABA and SR-95531 binding, and thus the subsites may overlap.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   GABA concentration-response curves (top), and SR-95531 competition curves (bottom) were measured for 12 (), 1D62C2 (), 1F64C2 (), 1R66C2 (), and 1S68C2 () receptors expressed in Xenopus oocytes. Data points were normalized to Imax for GABA concentration-response curves and to IGABA in the absence of blocker for SR-95531 competition curves. Points represent the mean ± S.D. from at least three experiments. Data were fit by nonlinear regression as described under "Experimental Procedures." EC50 and KI values are shown in Table II. GABA concentration-response curves are from data reported in Boileau et al. (14).

Effects of GABA, SR-95531, and Pentobarbital on MTS Reaction Rate Constants-- To identify potential agonist and antagonist subsites, we measured the rates of MTS covalent modification of ₁D62C, ₁F64C, ₁R66C, and ₁S68C in the presence and absence of GABA or SR-95531 (Table III and Fig. 4). We reasoned that if these residues line the binding pocket, then the presence of SR-95531 or GABA should slow the rate of MTS reaction due to steric hindrance. Both GABA and SR-95531 significantly slowed the rate of covalent modification at ₁F64C and ₁R66C, suggesting that these residues line a common agonist/antagonist-binding region. However, the rate of covalent modification of ₁D62C was only slowed by SR-95531, whereas the rate of modification of ₁S68C was only slowed by GABA. These results suggest that ₁D62C may form part of an antagonist subsite, whereas ₁S68C appears to form part of an agonist subsite.

View this table: [in this window] [in a new window]   Table III Rates of MTS covalent modification of 1D62C2, 1F64C2, 1R66C2, and 1S68C2 receptors in the presence and absence of GABA, SR-95531, and pentobarbital (PB) Rates of MTS reaction for all mutants except 1D62C2 were measured using MTSES. 1D62C2 does not react with MTSES, so MTSEA was used for this mutant. Numbers reflect means of three or more experiments ± S.D.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   A, structures of GABA and SR-95531. Lengths were measured after energy minimization (<0.5 kcal/Å; Chemsketch, ADC, Toronto, Ontario, Canada). B, rate of sulfhydryl modification of 1F64C2 receptors in the presence and absence of SR-95531. Representative GABA (EC40-60) current traces were recorded while applying MTSES (10 µM) in the presence (bottom) and absence (top) of the antagonist SR-95531 (10 µM, EC95). C, decreases in IGABA were plotted versus cumulative MTS exposure at 1D62C, 1F64C, 1R66C, and 1S68C containing receptors. Data obtained from individual experiments were normalized to the GABA current measured at t = 0 and are presented as mean ± S.D. from at least three independent experiments. Single exponential curve fits of the data reveal the effects of GABA and SR-95531 on the MTS reaction rates (, MTS alone; , MTS + GABA; , MTS + SR-95531). k2 values are shown in Table III. For 1D62C receptors MTSEA was used for covalent modification, whereas for 1F64C, 1R66C, and 1S68C MTSES was used, as described under "Experimental Procedures."

GABA not only binds to the receptor but also gates the channel. Therefore, the slowing of the rate of reaction at ₁S68C in the presence of GABA could be due to conformational changes that occur when the channel opens and desensitizes rather than to a direct physical block of position ₁S68C by GABA. To distinguish between these possibilities, we measured the rate of covalent modification of ₁S68C in the presence of pentobarbital, which directly activates the channel (34, 35) by binding to a site distinct from GABA (17).

The rates of covalent modification of ₁F64C and ₁R66C were not altered in the presence of a directly activating concentration of pentobarbital (500 µM), suggesting that the opening of the channel does not change the ability of the MTS reagents to modify these residues. However, the rates of covalent modification of ₁D62C and ₁S68C were significantly accelerated in the presence of a high concentration of pentobarbital (500 µM) (Fig. 5 and Table III). Because opening of the channel with pentobarbital accelerated covalent modification and GABA slowed covalent modification of ₁S68C, we predict that GABA sterically blocks access to this residue.

View larger version (10K): [in this window] [in a new window]   Fig. 5.   Rates of covalent modification in the presence and absence of pentobarbital. The second-order rate constants (k2) of MTS modification in the presence (500 µM pentobarbital, black bars) and absence of pentobarbital (white bars) at 1D62DC, 1F64C, 1R66C, and 1S68C were calculated as described under "Experimental Procedures." For each mutant, the rate constants were normalized to the MTS reaction rate measured in the absence of pentobarbital (control). Data represent mean ± S.D. from at least three experiments. * indicates values significantly different from control MTS values with p < 0.001.

Covalent modification of ₁D62C by MTSEA was slowed in the presence of SR-95531 and accelerated in the presence of pentobarbital. Although the data are consistent with SR-95531 causing a steric block and ₁D62C lining part of an antagonist subsite, it is feasible that SR-95531 could induce a conformational change in the receptor that leads to a slowing of MTSEA reaction at ₁D62C.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Recently, the crystal structure of the AChBP was solved (29). The AChBP is a homologue of the extracellular N-terminal domain of the nicotinic acetylcholine receptor and binds several ligands of this receptor. The nicotinic acetylcholine receptor and the GABA_A receptor are related proteins and are members of a LGIC superfamily of receptors. Thus, by using the AChBP structure as a template, we can begin to model the GABA-binding site (Fig. 6). Our secondary structure prediction of loop D (Fig. 1) (14) agrees with the crystal structure of the AChBP (29), where residues aligned with this region of the GABA_A receptor form a -strand (Fig. 6).

View larger version (31K): [in this window] [in a new window]   Fig. 6.   Structural model of the GABA-binding site. The extracellular N-terminal regions of the GABAA receptor 1 and 2 subunits were threaded onto the crystal structure of acetylcholine-binding protein (29) and energy-minimized as described under "Experimental Procedures." A, model of the / interface of the GABAA receptor. Domains believed to contribute to the GABA-binding site are highlighted in red and are labeled A-F. B and C, amino acid residues 1D62, 1F64, 1R66, and 1S68 in loop D are shown. 1F64 and 1R66 lie within the core of the GABA-binding pocket.

In this study, we used SCAM to examine the physicochemical environment of cysteine mutants in loop D. A residue in a relatively open, aqueous environment will have a faster rate of reaction than a residue in a relatively restrictive, nonpolar environment (26). In loop D, the fastest MTS reaction rate occurs at ₁F64C, followed by ₁R66C, ₁S68C, and ₁D62C (Table I). In addition, MTS reagents produce the largest inhibition of I_GABA at ₁F64C, followed by ₁R66C, ₁S68C, and ₁D62C (Fig. 1). Our data indicate that ₁F64C and ₁R66C are located in an aqueous and sterically unrestricted environment. Such an environment is thought to exist within the core of the binding pocket, and we predict that these residues lie within that core. Our prediction agrees with the AChBP structure where aligned residues are located in the center of the acetylcholine-binding pocket (Fig. 6). In contrast, we predict that ₁D62C is located in a sterically confined region and/or its sulfhydryl chain is in a relatively hydrophobic environment that is poorly ionized, because its rate of covalent modification by MTSEA (k₂ = 16 M¹ s¹) is ~150,000-fold slower than that of ₁F64C. Again, this is consistent with the structure of the AChBP where the aligned position is at the periphery of the binding site on a region of the ₂ strand that is twisting away and below the binding site.

We measured rates of reaction of differently charged MTS reagents to identify charge-specific regions of the binding pocket (Table I). Positively charged MTSET reacts significantly faster than negatively charged MTSES at ₁F64C and ₁R66C (Fig. 2B). We conclude that the difference in rates is due to a negative electrostatic potential located within the binding pocket. Based on the AChBP structure, residues in the binding pocket that could potentially form this negative subsite include ₁E182, ₁D183 in the loop F region of the GABA-binding site. Negatively charged residues in the homologous region of the muscle nicotinic acetylcholine receptor (D174/D180) have been identified that are important for acetylcholine binding (36, 37). Alternatively, the negative subsite could be formed by  electrons of aromatic amino acid side chains (38). Several aromatic residues have been identified that are important for GABA binding (₂Y97 (39), ₂Y157, ₂Y205, (17)). Experiments are in progress to test these hypotheses.

We speculate that the amino group of GABA is oriented away from ₁F64 and ₁R66 and faces toward this negative subsite, whereas the carboxylate group of GABA may be stabilized, at least in part, by ₁R66. Consistent with this hypothesis, removal of the bidentate positive charge at ₁R66 increases GABA EC₅₀ values several hundred-fold. In addition, muscimol, a high affinity agonist of the GABA_A receptor, has been shown to photoaffinity label the receptor at ₁F64 (12), and the photochemistry of this reaction indicates that the carboxylate-like part of the muscimol molecule reacts with ₁F64 (40).

In order to elucidate differences between agonist and antagonist binding, we measured rates of covalent modification in the presence and absence of GABA and SR-95531. Covalent modification of ₁D62C is slowed by SR-95531 but not GABA, whereas modification of ₁S68C is slowed by GABA but not SR-95531. In addition, cysteine mutagenesis of ₁R66 causes a change in GABA EC₅₀ values but not SR-95531 K_I values, whereas mutagenesis of ₁F64 causes a change in both GABA EC₅₀ and SR-95531 K_I values. Based on these data, we conclude that different amino acid residues within the loop D region of the binding pocket are important for mediating the effects of GABA and SR-95531. This is most likely due to differences in ligand structure and/or ligand positioning within the site.

Most GABA_A receptor agonists and antagonists contain a positively and a negatively charged functional group ~5 Å apart (41), similar to the GABA molecule. It is possible that these different classes of compounds bind with their intercharge portion in the same orientation. We have provided evidence that the carboxylate group of GABA likely binds near ₁F64 and ₁R66. Thus, like GABA and muscimol, we predict that the negatively charged region of SR-95531 is oriented near ₁R66 and ₁F64.

One problem with this prediction is that mutation of ₁R66 dramatically alters GABA binding but does not affect SR-95531 binding. The larger size of SR-95531 indicates that this molecule likely utilizes more attachment points than GABA within the binding pocket. It is likely that the ring structures of SR-95531 (Fig. 4) provide these additional attachment points. Because the rate of covalent modification of ₁D62C is slowed in the presence of SR-95531 and is not affected by GABA, we hypothesize that the bulky aromatic rings of SR-95531 are located near ₁D62. In lieu of a crystal structure of receptor bound with ligand, it is difficult to identify definitively which residues directly contact a ligand.

We demonstrate that conformational movements occur within the binding pocket during channel gating. Pentobarbital increases the rate of covalent modification of ₁D62C and ₁S68C suggesting that the conformational change induced by pentobarbital channel gating causes ₁D62C and ₁S68C to become more accessible. This may be due to the direct movement of ₁D62C and ₁S68C or due to movement of nearby regions of the protein.

The data presented in this study suggest that although the ₁D62-₁S68 region is clearly involved in forming the ligand-binding site, the specific interactions of GABA and SR-95531 and the local structural movements associated with these interactions are mediated by different amino acid residues. When GABA binds to the binding site, a conformational change occurs that results in channel activation, whereas when an antagonist binds channel gating is prevented. We speculate that the binding of GABA causes a closure or tightening of the binding pocket that leads to the opening of the channel (18, 42), and we hypothesize that the binding of an antagonist like SR-95531, which is larger than GABA, prevents complete closure of the pocket, and thus does not allow for channel gating. Further studies identifying specific residues involved in agonist and antagonist recognition within other parts of the binding pocket will lead to a better understanding of the mechanisms behind ligand agonism and antagonism and the process of channel gating.

	ACKNOWLEDGEMENTS

We thank Dr. Andrew Boileau and Anson Davis for constructing the cysteine mutants and Erin McCarthy, Jeff Malik, and James Seffinga-Clark for treating the oocytes. We also thank James Seffinga-Clark for help in constructing the structural models.

	FOOTNOTES

^* This work was supported by a Pharmaceutical Research and Manufacturers of America Foundation Predoctoral Fellowship (to J. H. H.) and NINDS Grant NS34727 from the National Institutes of Health (to C. C.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence should be addressed: Dept. of Physiology and Molecular and Cellular Pharmacology Program, University of Wisconsin, MSC, 1300 University Ave., Rm. 197, Madison, WI 53706. Tel.: 608-265-5863; Fax: 608-265-5512; E-mail: czajkowski@ physiology.wisc.edu.

Published, JBC Papers in Press, March 14, 2002, DOI 10.1074/jbc.M111778200

	ABBREVIATIONS

The abbreviations used are: LGIC, ligand-gated ion channel; GABA, -aminobutyric acid, GABA_A, -aminobutyric acid type A; SCAM, substituted cysteine accessibility method; MTS, methanethiosulfonate; MTSEA, 2-aminoethyl methanethiosulfonate; MTSET, 2-(trimethylammonium)ethyl methanethiosulfonate; MTSES, 2-sulfonatoethyl methanethiosulfonate; 2-ME, 2-mercaptoethanol; AChBP, acetylcholine-binding protein.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES