Abstract
Tetracaine (N,N-dimethylaminoethyl-4-butylaminobenzoate) and related N,N-dialkylaminoethyl substituted benzoic acid esters have been used to characterize the high-affinity binding site for aromatic amine noncompetitive antagonists in theTorpedo nicotinic acetylcholine receptor (nAChR). [3H]Tetracaine binds at equilibrium to a single site with a K eq value of 0.5 μM in the absence of agonist or presence of α-bungarotoxin and with aK eq value of 30 μM in the presence of agonist (i.e., for nAChR in the desensitized state). Preferential binding to nAChR in the absence of agonist is also seen forN,N-DEAE andN,N-diethylaminopropyl esters, both binding with 10-fold higher affinity in the absence of agonist than in the presence, and for the 4-ethoxybenzoic acid ester ofN,N-diethylaminoethanol, but not for the 4-amino benzoate ester (procaine). Irradiation at 302 nm of nAChR-rich membranes equilibrated with [3H]tetracaine resulted in covalent incorporation with similar efficiency into nAChR α, β, γ, and δ subunits. The pharmacological specificity of nAChR subunit photolabeling as well as its dependence on [3H]tetracaine concentration establish that the observed photolabeling is at the high-affinity [3H]tetracaine-binding site. Within α subunit, ≥95% of specific photolabeling was contained within a 20-kilodalton proteolytic fragment beginning at Ser173 that contains the M1 to M3 hydrophobic segments. With all four subunits contributing to [3H]tetracaine site, the site in the closed channel state of the nAChR is most likely within the central ion channel domain.
The muscle nicotinic acetylcholine receptor (nAChR) consists of four homologous subunits (α2βγδ) arranged pseudosymmetrically around a central axis that is a cation-selective ion channel. Each subunit has a common primary structure motif: a hydrophilic, extracellular N-terminal half containing amino acids of the agonist-binding sites, followed by three hydrophobic membrane spanning segments (M1–M3), a cytoplasmic domain, a fourth transmembrane segment (M4), and short extracellular C-terminal tail. Affinity labeling studies, site-directed mutagenesis, and low-resolution (9 Å) cryoelectron microscopy provide considerable information about nAChR structure (reviewed in Karlin and Akabas, 1995;Hucho et al., 1996; Unwin, 1998). The two agonist-binding sites, which are located extracellularly at the α-γ and α-δ subunit interfaces, are composed of multiple loops of primary structure from α and γ (or δ) subunits (reviewed in Prince and Sine, 1998). M2 domains from each subunit line the pore of the ion channel (Imoto et al., 1988; Unwin, 1995), with additional contributions from the extracellular ends of the M1 segments (Zhang and Karlin, 1997), whereas M3 and M4 segments are more peripheral and in contact with lipid (Blanton and Cohen, 1994).
Noncompetitive antagonists (NCAs) block the nAChR permeability response without preventing the binding of ACh (acetylcholine). A structurally diverse group of drugs act as NCAs, including many aromatic amines, general anesthetics, fatty acids, steroids, and neuropeptides such as Substance P (reviewed in Arias, 1998). Studies of the binding of the aromatic amines [3H]meproadifen and [3H]phencyclidine (PCP) or of the spiropiperidine [3H]histrionicotoxin ([3H]HTX) to nAChR-rich membranes fromTorpedo electric organ establish that each binds with high affinity (K ∼ μM) to one site per nAChR and to additional lower-affinity sites (Heidmann et al., 1983). The high-affinity site is linked allosterically to the ACh site, with most aromatic amines binding at equilibrium with highest affinity in the presence of agonist [i.e., to the desensitized state of the nAChR (Cohen et al., 1985; Arias, 1996; Lurtz and Pedersen, 1999)].
Affinity-labeling studies have identified homologous residues near the cytoplasmic end of each M2 segment that contribute to the high-affinity binding site in desensitized Torpedo nAChRs for the aromatic amine NCAs [3H]chlorpromazine (Revah et al., 1990) and [3H]trimethylphenylphosphonium (Hucho et al., 1986). Mutational analyses also implicate these amino acids as affinity determinants for the aromatic amine QX-222 (Charnet et al., 1990) and for aliphatic alcohols (Forman, 1997) acting as open channel blockers. However, in each conformational state, there may be distinct, nonoverlapping sites for positively charged NCAs, and different regions within the ion channel may contribute to the binding site for a single ligand in different conformational states. In the desensitized nAChR, [3H]meproadifen mustard reacts with αGlu262 at the extracellular end of M2 (Pedersen et al., 1992), and based on fluorescence energy transfer, the binding site for ethidium is located above the level of the bilayer in the vestibule of the channel (Johnson and Nuss, 1994, but see Lurtz et al., 1997). In the open channel state, [3H]quinacrine azide is photoincorporated into amino acids within αM1 (DiPaola et al., 1990), and mutations within αM1 affect quinacrine potency as an NCA but not chlorpromazine (Tamamizu et al., 1995). Photoaffinity labeling studies with 3-(trifluoromethyl)-3-(M-[125I]iodophenyl) diazirine ([125I]TID; White and Cohen, 1992) and [3H]diazofluorene (Blanton et al., 1998), uncharged, hydrophobic NCAs, have identified a binding site in the M2 domain in the absence of agonist (closed channel), as well as changes in structure of the M2 domain between resting and desensitized states. However, the TID site in the closed channel appears distinct from the binding site for PCP because PCP does not inhibit [125I]TID photoincorporation and TID does not inhibit [3H]PCP binding (White et al., 1991).
Tetracaine (dimethylaminoethyl-p-butylaminobenzoate) is an unusual aromatic amine NCA because it is 100-fold more potent as an inhibitor of [3H]HTX binding (K≈ 1 μM) in the absence of agonist than in the presence (Blanchard et al., 1979), and it stabilizes the resting state rather than the desensitized state of the nAChR (Boyd and Cohen, 1984). Here, we characterize the binding properties of several structural analogs of tetracaine to further define the requirements for preferential binding to the closed channel state, and we use [3H]tetracaine itself as an intrinsic photoaffinity reagent to define the structure of its high-affinity binding site in the nAChR in the absence of agonist. [3H]Tetracaine is specifically photoincorporated with similar efficiency into each nAChR subunit. In the following report (Gallagher and Cohen, 1999), we identify the homologous amino acids in the M2 segment of each subunit that contribute to this high-affinity [3H]tetracaine site.
Experimental Procedures
Materials.
[ring-3,5-3H]Tetracaine ([3H]tetracaine, 36 Ci/mmol) and [3H]HTX (60 Ci/mmol) were prepared at New England Nuclear Research Products (Boston, MA) by tritium gas catalytic reduction of 3,5-dibromotetracaine anddl-decahydro(pentenyl)histrionicotoxin (H10-HTX), respectively. For binding experiments, [3H]tetracaine was purified to >95% by silica thin-layer chromatography (5:4:1 cyclohexane/chloroform/diethylamine;R f = 0.17). When stored in ethanol, [3H]tetracaine decomposed at ∼10% per month, forming tritiated degradation products that did not partition into nAChR-rich membranes or bind to glass filters. Also, these degradation products did not appear to photoincorporate into nAChR-rich membranes because the same [3H]tetracaine photolabeling patterns were seen with [3H]tetracaine of 95 or 50% radiochemical purity (not shown).
H10-HTX anddl-perhydrohistrionicotoxin (H12-HTX) were kindly provided by Dr. Y. Kishi (Harvard University, Cambridge, MA). Proadifen was obtained from SmithKline Beecham (Philadelphia, PA). Piperocaine was obtained from Eli Lilly (Indianapolis, IN). Tetracaine, procaine, carbamylcholine chloride, d-tubocurarine chloride, oxidized glutathione (GSSG), and endoglycosidase H were obtained from Sigma Chemical Co. (St. Louis, MO). α-Bungarotoxin was purchased from Biotoxins Inc. (St. Cloud, FL). PCP was obtained from Alltech Associates. Staphylococcus aureus V8 protease (V8 protease) was obtained from ICN Biochemicals (Costa Mesa, CA).
The 2-(diethylamino)ethyl- and 3-(diethylamino)propyl esters ofp-butylaminobenzoic acid were synthesized by coupling with diisopropylcarbodiimide using chemicals from Aldrich Chemical (Milwaukee, WI). For synthesis of the DEAE ester, 2 g (0.01 mol) of 4-(butylamino)benzoic acid, 1.8 ml (0.011 mol) of diisopropylcarbodiimide, and 0.2 g (0.0013 mol) of 4-pyrrolidinopyridine were stirred for 5 min at room temperature in 10 ml of methylene chloride, and then 3 equivalents (6 ml) ofN,N-diethylethanolamine were added and allowed to react overnight. After removal of methylene chloride under vacuum, the residue was resuspended in anhydrous ether, with the insoluble diisopropyl urea removed by filtration. The filtrate was washed with 5% NaHCO3 and water, and the organic layer was dried over anhydrous Na2SO4. HCl gas was bubbled into the ethereal solution to precipitate the dihydrochloride salt that was collected by filtration and then recrystallized three times from ethanol ether. Two grams of product (55% yield) was obtained. A similar synthesis starting with 3-diethylamino-1-propanol yielded 3.4 g of product (60% yield). Appropriate elemental analyses and NMR spectra were obtained for each compound. Meproadifen iodide was prepared by reaction of proadifen base with methyl iodide.
nAChR-rich membranes were isolated from the electric organs ofTorpedo californica (Marinus, Inc., Westchester, CA) as described (Pedersen and Cohen, 1990). The final membrane pool was stored in 38% sucrose, 0.02% NaN3 at −80°C under argon and contained 1 to 2 nmol of acetylcholine-binding sites/mg of protein, as measured by a [3H]ACh centrifugation assay (Pedersen et al., 1986).
Radioligand Binding Assays.
The equilibrium binding of [3H]tetracaine to Torpedo nAChR-rich membranes in Torpedo physiological saline (TPS; 250 mM NaCl, 5 mM KCl, 3 mM CaCl2, 2 mM MgCl2, 5 mM sodium phosphate, pH 7.0) was assayed by centrifugation. Membrane suspensions (≈400 nM ACh sites) were equilibrated at 4°C for 4 h with varying concentrations of [3H]tetracaine (0.5 Ci/mmol). Membranes were also preequilibrated with agonist or competitive antagonist, or with excess nonradioactive tetracaine (50–100 μM) to define nonspecific binding. Aliquots (0.1 ml) were then centrifuged in a Beckman Airfuge at 100,000g for 10 min at ∼8°C. After removal of the supernatants, membrane pellets were resuspended in 0.1 ml of 10% SDS, and pellet and supernatant 3H were determined by liquid scintillation counting. This centrifugation assay was also used to quantify the concentration of high-affinity [3H]tetracaine sites compared with the concentration of [3H]ACh and/or [3H]HTX sites measured simultaneously with the same membrane suspension. For studies with [3H]ACh, membrane suspensions were pretreated with 0.3 mM diisopropylphosphofluoridate to inhibit cholinesterase activity. [3H]HTX was diluted isotopically with H12-HTX to produce a final radiochemical specific activity of 2 Ci/mmol for binding assays.
[3H]HTX and [3H]tetracaine binding were also determined by a filtration assay using glass-fiber filters (2.5 cm, No. 32; Schleicher & Schuell, Keene, NH) that had been pretreated with an organosilane (1.0% Prosil; Lancaster Synthesis, Windham, NH). Membrane aliquots (typically 100–200 μl) were applied to the filters in the absence of vacuum to achieve a uniform spread of the suspension over the filter surface; then, vacuum was applied, and the filter was washed with 5 ml of TPS at 4°C.
Photoaffinity Labeling of nAChR-Rich Membranes with [3H]Tetracaine.
Membrane suspensions (30 μl, 1.5–1.8 mg protein/ml) in TPS were equilibrated at room temperature with [3H]tetracaine (4 Ci/mmol) and cholinergic ligands and then irradiated for 30 min with a 302-nm lamp (Spectroline EB-280C, 1150 μW/cm2) at a distance of 12 cm as described (Pedersen and Cohen, 1990). Unless indicated otherwise, suspensions also included 50 mM GSSG as an aqueous photochemical scavenger. The 96-well microtiter plate containing the samples was placed in a water bath during photolysis, so the temperature increased <2°C during 30 min of irradiation. After photolysis, samples were usually prepared for SDS-polyacrylamide gel electrophoresis (PAGE) by the direct addition of 10 μl of 4× sample loading buffer to reaction mixtures. To quantify the dependence of 3H incorporation as a function of free [3H]tetracaine, after photolysis the reaction mixtures were centrifuged, with the supernatants assayed to determine free [3H]tetracaine, and the pellets dissolved in sample loading buffer. In preliminary experiments, the efficiency of [3H]tetracaine photoincorporation into nAChR-rich membranes was found to be 10 times higher for the 302-nm lamp than for lamps of 254 or 365 nm of similar radiant flux density (not shown).
Gel Electrophoresis.
Polypeptides were resolved by SDS-PAGE on 8% acrylamide gels with a modified Laemmli buffer system (White and Cohen, 1988), and the incorporation of [3H]tetracaine was determined by fluorography or quantified by scintillation counting of gel slices as described previously (Middleton and Cohen, 1991). Proteolytic mapping of the [3H]tetracaine-labeled α subunit withS. aureus V8 protease was performed according to the procedure of Cleveland et al. (1977) as described by White and Cohen (1988). Membrane suspensions (540-μg aliquots) were photolabeled with [3H]tetracaine and then electrophoresed on an 8% acrylamide gel. After a brief staining with Coomassie brilliant blue and destaining, the band containing α subunit was excised and transferred to the well of a mapping gel (15% acrylamide) to which V8 protease (3 μg) was added. 3H incorporation into the proteolytic fragments of [3H]tetracaine-labeled α subunit was analyzed by fluorography and by scintillation counting of gel slices.
Data Analysis
The equilibrium binding of [3H]tetracaine and the concentration dependence of3H incorporation into nAChR subunits were fit by nonlinear least-squares (SigmaPlot; Jandel Scientific, San Rafael, CA) to the function {B = A/[1 + (K eq/L)] +m * L}, where B is the bound [3H]tetracaine (or cpm incorporated),A is the maximum specific binding (or cpm incorporated),K eq is the dissociation constant (or apparent dissociation constant, K AP),L is the measured free [3H]tetracaine concentration, and m is the slope of nonspecific binding (or labeling), which was usually determined in parallel experiments performed in the presence of excess nonradioactive tetracaine or H10-HTX. For labeling experiments, the measured free3H cpm was adjusted to account for the ≈50% radioimpurities known to be present in the [3H]tetracaine used for photolabeling.
The concentration dependence of the inhibition of reversible binding of [3H]HTX or [3H]tetracaine and of [3H]tetracaine photolabeling of nAChR subunits was fit to a function {B = A/[1 + (I/K I n)] + NSP}, where B is the 3H cpm bound (or incorporated) in the presence of inhibitor at a total concentration, I; A is the specific3H cpm bound (or incorporated) in the absence of inhibitor; K I is the apparent inhibitor dissociation constant; n is the Hill coefficient, and NSP is the observed nonspecific binding or labeling, which was not treated as an adjustable parameter. The parameter n Hwas treated as adjustable only if inhibition curves deviated from a single-site model (n H = 1). Under the assay conditions used, with the concentration of free [3H]HTX or [3H]tetracaine much less than theirK eq value, K iwill be close to the inhibitor equilibrium dissociation constant whenn H = 1 and the total inhibitor concentration is a good approximation of the free concentration.
Results
Inhibition of [3H]HTX Binding by Tetracaine Analogs.
Because tetracaine’s preferential binding to the closed channel state of the nAChR appeared unusual for aromatic amine NCAs, we first examined other benzoic acid esters as inhibitors of [3H]HTX binding to identify structural features responsible for tetracaine’s binding properties (Table1). All drugs at high concentrations completely inhibited specific binding of [3H]HTX, with the concentration dependence of inhibition for each drug consistent with competition at a single site. With the exception of procaine (V), all bound with highest affinity in the absence of agonist. For the esters ofp-butylaminobenzoic acid, replacement of the dimethylaminoethyl of tetracaine (I) by DEAE (II) resulted in a ∼6-fold increase in binding affinity in both the absence and presence of agonist. Comparison of procaine (V) with compound II established that aliphatic substitution at the aryl nitrogen increased affinity for the closed channel conformation by 4000-fold but by only 300-fold for the desensitized conformation. However, high-affinity binding in the absence of agonist did not depend uniquely on the butylamino group because thep-ethoxybenzoic acid ester (IV) also bound preferentially to the resting state (K I = 1.6 μM), and piperocaine (VI), an unsubstituted benzoic acid ester, was bound with 50-fold higher affinity in the absence of agonist. Thus, selective, high-affinity binding in the absence of agonist appeared to be the rule for many simple benzoic acid esters, with procaine a notable exception.
Equilibrium Binding of [3H]Tetracaine.
The equilibrium binding of [3H]tetracaine to nAChR-rich membranes in TPS at 4°C was determined by centrifugation and filtration assays (Fig. 1A). For [3H]tetracaine concentrations to 4 μM, in each assay the total binding was well fit by a hyperbolic binding function with a linear, nonspecific component. The nonspecific binding calculated from the fit of the total binding function was approximately the same as that measured in the presence of 50 μM nonradioactive tetracaine. [3H]Tetracaine bound to the same number of high-affinity sites in both assays, with the ratio of sites determined by filtration and centrifugation equal to 1.2 ± 0.2 in three experiments. The same value of the dissociation constant was determined by centrifugation (K eq = 0.5 ± 0.1 μM) and filtration (K eq = 0.6 ± 0.1 μM). The nonspecific binding determined by filtration, with a brief (5 s) wash, was only 20% that determined by centrifugation. The partition coefficient, determined as the ratio of the nonspecifically bound to the free tetracaine concentration normalized to the membrane protein concentration was 0.08 ± 0.02 (mg/ml)−1 by centrifugation and 0.014 ± 0.005 by filtration. Thus, in the absence of agonist, specific [3H]tetracaine binding to the nAChR-rich membranes was characterized by a singleK eq, and a filtration assay can be used to quantify equilibrium binding of [3H]tetracaine to its high-affinity site.
The filtration assay was used to examine the effects of an agonist (carbamylcholine) and competitive antagonists (α-bungarotoxin andd-tubocurarine) on the binding to nAChR-rich membranes of [3H]tetracaine at concentrations to 20 μM (Fig. 1B). None of the ligands altered the nonspecific binding determined in the presence of 100 μM tetracaine (not shown), and [3H]tetracaine was bound with similar affinity (K eq = 0.5 ± 0.1 μM) and to the same number of sites (ratio = 1.1 ± 0.2, n = 19) in the absence and presence of α-bungarotoxin. Partial desensitization of the nAChR with d-tubocurarine resulted in [3H]tetracaine binding to the same number of sites but with a K eq value of 1.1 ± 0.6 μM (n = 23). In the presence of carbamylcholine, which converts nAChRs fully to the desensitized state, [3H]tetracaine binding affinity was decreased dramatically, with only ∼80% of the expected tetracaine-binding sites occupied at 20 μM [3H]tetracaine. When the data were fit using the site concentration determined in the presence of α-bungarotoxin or d-tubocurarine,K eq was 29 ± 7 μM (n = 4).
To determine the number of [3H]tetracaine-binding sites per nAChR, [3H]tetracaine binding was measured in parallel with binding assays of [3H]HTX (in the presence of 100 μM carbamylcholine) and [3H]ACh. The ratio of [3H]tetracaine sites to [3H]ACh sites was 0.43 ± 0.07 (n = 19), and the ratio of [3H]HTX to [3H]ACh sites was 0.5 ± 0.1. Thus, [3H]tetracaine and [3H]HTX each bound to the same number of sites, and this number was half the number of [3H]ACh sites.
Inhibition of [3H]Tetracaine Binding by NCAs.
The desensitizing NCAs PCP, meproadifen, proadifen, and H10-HTX were examined as inhibitors of the equilibrium binding of [3H]tetracaine at 4°C in the absence of other cholinergic ligands or in the presence of either α-bungarotoxin or d-tubocurarine (Fig.2). Although HTX binds with similar affinity (K eq = 0.3 μM) in the absence or presence of agonist at 20°C (Blanchard et al., 1979), at 4°C it binds with high affinity (K eq = 0.3 μM) in the presence of agonist but only weakly (K eq = 8 μM) in the absence (Cohen et al., 1985). The four NCAs inhibited [3H]tetracaine binding in a concentration-dependent manner, with high concentrations inhibiting [3H]tetracaine binding by the same extent as 100 μM tetracaine. K I values for inhibition of [3H]tetracaine binding were 6- to 10-fold lower in the presence of d-tubocurarine than in its absence, consistent with the preferential binding of these NCAs to desensitized nAChR (Heidmann et al., 1983; Cohen et al., 1985). The concentration dependence of inhibition by H10-HTX (Fig. 2A) and PCP (Fig. 2B) was well fit by Hill coefficients of 1 in both the absence and presence of d-tubocurarine or α-bungarotoxin, consistent with simple competitive inhibition of [3H]tetracaine binding. For proadifen (Fig.2C), the Hill coefficients for inhibition were unitary except in the presence of α-bungarotoxin (n H = 1.6). For meproadifen (Fig. 2D), the dose dependence of inhibition was characterized by n H = 1 in the presence ofd-tubocurarine but by n H = 1.5 in the absence of ACh site ligand and by n H= 1.8 in the presence of α-bungarotoxin.
Photoincorporation of [3H]Tetracaine into nAChR-Rich Membranes.
To determine whether [3H]tetracaine could be specifically photoincorporated into its high-affinity binding site, nAChR-rich membranes were equilibrated with 5 μM [3H]tetracaine and then irradiated at 302 nm for 30 min in the presence or absence of the NCA H10-HTX (30 μM). When the proteins were resolved by SDS-PAGE and the gel was processed for fluorography, it was seen (Fig. 3, lanes 2 and 3) that H10-HTX decreased [3H]tetracaine incorporation not only in the nAChR subunits but also into another protein of the nicotinic postsynaptic membrane [rapsyn/43 kilodaltons (kDa) protein] and into polypeptides of contaminating membrane fragments such as the α subunit of the Na+,K+-ATPase (90 kDa;White and Cohen, 1988). Photoincorporation into nonreceptor polypeptides was also decreased by the agonist carbamylcholine, which allosterically inhibits [3H]tetracaine binding (see Figs. 6 and 8). It was improbable that HTX and carbamylcholine actually inhibited [3H]tetracaine binding to specific sites on the Na+,K+-ATPase. A more likely explanation was that a reactive [3H]tetracaine intermediate generated primarily when [3H]tetracaine was bound to its high-affinity site in the nAChR could dissociate from its NCA site and then react with other membrane proteins. This pseudospecific photoincorporation should be sensitive to aqueous compounds that could quench the free reactive species.
GSSG was examined as a potential aqueous scavenger, and indeed high concentrations (50 mM) were effective in dramatically reducing the pseudospecific [3H]tetracaine photoincorporation into nonreceptor polypeptides (Fig. 3, lanes 4–7). With 50 mM GSSG, the H10-HTX-sensitive labeling was primarily associated with the four nAChR subunits. There also remained specific photolabeling of an nAChR β subunit proteolytic fragment, which appears as a band migrating with slightly greater mobility than the α subunit (Pedersen and Cohen, 1990), as well as of bands migrating between the α and β subunits in a gel region known to contain proteolytic fragments of nAChR γ and δ subunits (Pedersen and Cohen, 1990).
The effects of GSSG at concentrations up to 100 mM on specific and nonspecific [3H]tetracaine photoincorporation into nAChR subunits and nonreceptor polypeptides were quantified by analyzing 3H incorporation in gel slices excised from stained gels (Fig. 4). For nAChR subunits and nonreceptor polypeptides, the effects of GSSG were most pronounced at concentrations up to 10 mM. For nAChR subunits, the [3H]tetracaine photolabeling inhibitable by H10-HTX (the difference between3H incorporation for samples labeled in the absence and presence of HTX) was essentially constant at GSSG concentrations of >10 mM. In the presence of 50 mM GSSG, the H10-HTX-sensitive 3H incorporation into the nAChR α, β, γ, and δ subunits was 792, 730, 301, and 715 cpm, respectively, with only 21 and 54 cpm in the nonreceptor polypeptides of 37 kDa (calelectrin) and 90 kDa (Na+,K+-ATPase α subunit), respectively. GSSG reduced the pseudospecific incorporation into nonreceptor polypeptides by ≈85%, and all subsequent experiments included 50 mM GSSG in the reaction mixture.
[3H]Tetracaine photoincorporation into nAChR subunits was quantified by measuring 3H incorporation as a function of free [3H]tetracaine concentration (Fig.5). The nonspecific labeling of each nAChR subunit in the presence of 50 μM H10-HTX increased linearly with free tetracaine (Fig. 5, open symbols), and for each subunit, the total 3H incorporation (Fig. 5, filled symbols) was well fit by a simple hyperbolic binding function plus a linear, nonspecific component. For each subunit,K AP was ∼1.4 μM, a value close to theK eq value of 0.5 μM determined directly for the reversible, equilibrium binding of [3H]tetracaine (Fig. 1A) and well below the affinity of [3H]tetracaine for the agonist site (K ∼ 800 μM; Blanchard et al., 1979). For the conditions of photolabeling used, the maximum specific cpm incorporated was equivalent to labeling of 0.6, 1.1, 0.9, and 1.0% of the α, β, γ, and δ subunits, respectively, which was increased by 50 to 75% when the time of irradiation was increased from 30 to 60 min (not shown).
Effects of Cholinergic Ligands on [3H]Tetracaine Photoincorporation into nAChR-Rich Membranes.
[3H]Tetracaine photoincorporation was examined in the presence of agonists, competitive antagonists, and NCAs (Fig.6) for comparison with the known pharmacology of [3H]tetracaine binding to the NCA site. The NCAs proadifen, PCP, and H10-HTX each inhibited the photoincorporation of [3H]tetracaine into all four nAChR subunits (Fig. 6, lanes 2 versus 3–5). PCP was the least potent, consistent with its relative potency as an inhibitor of [3H]tetracaine equilibrium binding in the absence of agonist (Fig. 2). [3H]Tetracaine photoincorporation into the nAChR subunits was not inhibited by excess α-bungarotoxin (Fig. 6, lanes 2 versus 9 and 16), and the three NCAs still inhibited subunit photolabeling in the presence of α-bungarotoxin (lanes 6–8). Carbamylcholine (10 and 300 μM) reduced [3H]tetracaine incorporation into all four nAChR subunits by 70 to 80% (Fig. 6, lanes 2 versus 17 and 18, and counting of excised gel bands), but this inhibition was not seen when α-bungarotoxin was present, preventing carbamylcholine from binding to the agonist site (lanes 14 and 15). Althoughd-tubocurarine at 2 μM had little effect on [3H]tetracaine photoincorporation (lane 13),d-tubocurarine at 50 μM reduced labeling by 25 to 40% (lane 12), an inhibition not seen in the presence of α-bungarotoxin (lane 10). Thus, the inhibition of [3H]tetracaine photolabeling by the NCAs was consistent with a competitive inhibition of high-affinity [3H]tetracaine binding, whereas the inhibition of photolabeling by carbamylcholine and d-tubocurarine resulted from the allosteric inhibition of [3H]tetracaine binding when the nAChR was desensitized by drugs binding to the agonist site.
The concentration dependence for the inhibition of [3H]tetracaine photolabeling of nAChR subunits by H10-HTX (Fig. 7) and carbamylcholine (Fig. 8) was determined by quantification of 3H incorporation in gel slices. For H10-HTX, the inhibition data for each subunit were well fit by a single-site inhibition function with IC50 values ranging from 3 to 4 μM and maximal inhibition of 81, 94, 65, and 93% for the α, β, γ, and δ subunits, respectively. The observed IC50values exceeded the directly measured K eqvalue of 0.3 μM at 20°C (Heidmann et al., 1983), but they were consistent with that value for the assay conditions used: 1) the site concentration (1.3 μM) exceeds the K eqvalue for H10-HTX, 2) a significant percentage of the H10-HTX is bound nonspecifically [partition coefficient = 0.3 (mg/ml)−1], and 3) the free concentration of [3H]tetracaine was greater than its K eq value. For the known nAChR concentration, the approximate free [3H]tetracaine concentration, and the known partition coefficient for H10-HTX, the observed IC50 value leads to a calculated dissociation constant of ≈0.8 μM. For carbamylcholine, the subunit inhibition data were fit by IC50 values between 1.7 and 2.6 μM, with maximal inhibition of labeling of 90, 91, 61, and 91% for the α, β, γ, and δ subunits. Again, the observed total concentration of carbamylcholine for 50% inhibition exceeded the directly measured K eq value of 0.1 μM (Boyd and Cohen, 1980), which was not surprising for an assay using 2.7 μM ACh sites. When the inhibition curves were fit to determine the free concentration of carbamylcholine associated with 50% inhibition (K AP; White and Cohen, 1988), theK AP values for each subunit were between 0.3 and 0.9 μM.
Mapping the [3H]Tetracaine-Labeled Site in nAChR α Subunit with S.aureus V8 Protease.
Digestion of the nAChR α subunit by S. aureus V8 protease using the procedure of Cleveland et al. (1977) produces four nonoverlapping fragments of 20 (αV8-20), 18 (αV8-18), 10 (αV8-10), and 4 kDa (αV8-4) that are readily resolved by SDS-PAGE (Pedersen et al., 1986). The N termini of αV8-20, αV8-18, αV8-10, and αV8-4 begin at αSer173, αVal46, αAsn339, and αSer1, respectively (Pedersen et al., 1986;White and Cohen, 1988). αV8-18 contains the only Asn-linked carbohydrate α subunit, and this fragment is shifted to a band of 12 kDa (αV8-12) when the carbohydrate is removed by digestion with endoglycosidase H (Pedersen et al., 1986). nAChR-rich membranes were labeled with [3H]tetracaine in the absence or presence of 50 mM GSSG and with or without 30 μM H10-HTX. After treatment of the labeled membranes with endoglycosidase H, α subunits were isolated by SDS-PAGE and then digested with V8 protease as described in Experimental Procedures to determine the 3H incorporation within each α subunit proteolytic fragment (Fig.9). In the presence of 50 mM GSSG (lanes 5–8), αV8-20 was the only fragment labeled specifically. When gel pieces corresponding to the proteolytic fragments were excised and the3H incorporation was quantified by scintillation counting, it was found that 96% of the specific incorporation was in αV8-20, with specific labeling in αV8-18, αV8-10, and αV8-4 each <2% that of αV8-20. The principal effect of 50 mM GSSG was to reduce the nonspecific incorporation into αV8-20 by 80%, whereas the specific labeling was not decreased at all. In contrast, H10-HTX-sensitive labeling of αV8-18, αV8-10, and αV8-4 was reduced by 78, 69, and 97%, respectively.
Discussion
The study presented here concerns the nature of the binding site for amine NCAs that bind selectively to the nAChR in the absence of agonist (i.e., to the closed channel state of the nAChR). In contrast to the selectivity for the desensitized state seen for most bulky amine NCAs containing fused aromatic rings or multiple aromatic or aliphatic rings, we found that most N,N-substituted ethanolamine esters of benzoic acid actually bind selectively in the absence of agonist (Table 1). With the exception of procaine, which binds weakly and with similar affinity in the absence and presence of agonist, 4-butylamino and 4-ethoxybenzoate esters bind preferentially to the resting state, as does piperocaine, an unsubstituted benzoate. The selectivity of tetracaine for the resting state results from contributions both from the 4-butylamino substitution and from the presence of the N,N-dimethyl rather thanN,N-diethyl. With reference to procaine (compoundV, Table 1), the 4-butylamino substitution (compoundII) enhanced binding to the resting state by 10-fold more than to the desensitized state, with the further methyl replacements in tetracaine (compound I) weakening binding to the desensitized state by slightly more than to the resting state. In addition to the benzoic acid esters, at least one phenylacetic acid ester of N,N-diethylaminoethanol binds preferentially in the absence of agonist (Cohen et al., 1986). Although proadifen (the ester of 2,2-diphenylpropionic acid) bound with 10-fold higher affinity to the desensitized state and adiphenine (the ester of diphenylacetic acid) bound with similar affinity (K I = 4 μM) in the presence or absence of agonist, butethemate (the 2-phenylbutyryl ester) bound with 8-fold higher affinity to the resting state. The only other drug known to have high resting state selectivity is amobarbital, which binds with 500-fold higher affinity to the resting state, whereas other barbiturates bind preferentially to the desensitized or open channel states (Cohen et al., 1986; de Armendi et al., 1993).
In the absence of agonist, [3H]tetracaine was bound at equilibrium with high affinity (K eq = 0.5 μM) to one site per nAChR monomer, and [3H]tetracaine bound with the same high affinity when ACh sites were occupied with α-bungarotoxin. [3H]Tetracaine also bound to a single site in the desensitized state of the nAChR but with 60-fold lower affinity. In contrast to other, more hydrophobic aromatic amine NCAs that bind to as many as 30 low-affinity sites as well as to the high-affinity site (Heidmann et al., 1983), there was no evidence that [3H]tetracaine (up to 20 μM) binds to additional low-affinity sites in the nAChR-rich membranes. [3H]Tetracaine bound to the same number of high-affinity sites as [3H]HTX, and as seen previously for [3H]PCP and [3H]HTX (Heidmann et al., 1983), this number was half the number of [3H]ACh sites in the nAChR-rich membranes. Because [3H]ACh binds to two sites per nAChR (Neubig and Cohen, 1979), these NCAs each bind with high affinity to one site per nAChR.
HTX, PCP, proadifen, and meproadifen each completely inhibited the specific [3H]tetracaine binding to nAChR-rich membranes in the absence as well as in the presence of the competitive antagonists d-tubocurarine and α-bungarotoxin. For HTX and PCP, in each condition the concentration dependence of inhibition was well fit by a simple, single-site model (n H= 1). In the presence of d-tubocurarine, the concentration dependence of inhibition by meproadifen or proadifen was also consistent with competition at a single site (n H = 1). However, at the higher concentrations (>10 μM) of these drugs necessary to inhibit [3H]tetracaine binding in either the presence of α-bungarotoxin or the absence of other agonist or competitive antagonists, the inhibition curves were characterized by Hill coefficients between 1.5 and 2.0. The steep concentration dependence seen in the presence of α-bungarotoxin suggests that high concentrations of meproadifen and proadifen desensitize the nAChR by a mechanism unrelated to their binding to the high-affinity NCA site or to the ACh sites (Heidmann et al., 1983; Boyd and Cohen, 1984). Of the four NCAs studied, meproadifen binds with highest affinity to the ACh site (K eq = 50 μM; Heidmann et al., 1983), so in the absence of α-bungarotoxin ord-tubocurarine, it may also inhibit [3H]tetracaine binding by binding to the agonist site and desensitizing the nAChR.
To characterize the structure of the aromatic amine NCA site in the closed channel state of the nAChR, we tested whether [3H]tetracaine itself could be used as a photoaffinity probe. Irradiation of nAChR-rich membranes equilibrated with [3H]tetracaine resulted in covalent incorporation into membrane polypeptides. However, the observed photolabeling in the absence of aqueous scavenger was surprising, with photoincorporation into all polypeptides in the membrane suspensions inhibited by the presence of the NCA H10-HTX (Figs. 3 and 4) or the agonist carbamylcholine, drugs that inhibit [3H]tetracaine binding to its high-affinity site in the nAChR. Because the H10-HTX-sensitive [3H]tetracaine incorporation into polypeptides other than the nAChR subunits was reduced by at least 85% in the presence of the aqueous scavenger GSSG (50 mM), this labeling probably occurred after photoactivated [3H]tetracaine dissociated from its nAChR binding site. GSSG was also an effective aqueous scavenger of the nonspecific photolabeling by [3H]d-tubocurarine (Pedersen and Cohen, 1990) and [3H]nicotine (Middleton and Cohen, 1991). However, the specific component of [3H]tetracaine photolabeling was insensitive to GSSG concentrations to 50 mM, whereas for the photolabeling at the agonist site, concentrations of GSSG of more than 1 mM inhibited the specific as well as the nonspecific components. Thus, it appears that the [3H]tetracaine binding site in the closed channel state of the nAChR is substantially more protected from the aqueous environment than the ACh site. Another difference between the photolabeling by [3H]tetracaine and that by [3H]d-tubocurarine or [3H]nicotine is the wavelength dependence, with tetracaine photoincorporation more efficient for irradiation above 300 nm, whereas the other drugs required irradiation below 300 nm. Because tetracaine’s long wavelength absorption maximum is 300 nm in benzene and 310 nm in water, it is tetracaine and not the nAChR that is photoactivated.
In the presence of 50 mM GSSG, nearly all [3H]tetracaine photoincorporation sensitive to H10-HTX was restricted to the four nAChR subunits. The specific photolabeling of the four subunits resulted from activation of [3H]tetracaine bound to its high-affinity NCA site, as evidenced by the dependence of this specific photolabeling on free [3H]tetracaine concentration (Fig. 5) and the concentration dependence of inhibition by H10-HTX (Fig. 7) or carbamylcholine (Fig. 8). Also, as expected from the equilibrium binding data, the subunit photolabeling was inhibitable by other aromatic amine NCAs but not by α-bungarotoxin (Fig. 6). Thus, [3H]tetracaine bound to its single high-affinity binding site in the nAChR in the absence of agonist is photoincorporated into all four subunits with similar efficiency, which indicates that the binding site is located within the central ion channel domain at the interface of all five subunits. Within α subunit, the amino acids specifically labeled by [3H]tetracaine are restricted to a 20-kDa fragment containing the M1–M3 hydrophobic segments. As described inGallagher and Cohen (1999), amino acids from the M2 segments of each subunit are specifically photolabeled by [3H]tetracaine, and the binding site for [3H]tetracaine in the closed state of the ion channel overlaps and extends beyond the binding site for [125I]TID in the M2 ion channel domain (White and Cohen, 1992).
Acknowledgments
We thank Dr. Wu Schyong Liu for synthesis of thep-butylaminobenzoic acid esters and Drs. Benjamin White and Steen Pedersen for helpful suggestions during the course of these experiments.
Footnotes
- Received March 23, 1999.
- Accepted May 18, 1999.
-
Send reprint requests to: Dr. Jonathan B. Cohen, Department of Neurobiology, Harvard Medical School, 220 Longwood Ave., Boston, MA 02115. E-mail: jonathan__cohen{at}hms.harvard.edu
-
↵1 Present address: Merck Research Laboratories, Rahway, New Jersey 07065.
-
This work was supported in part by U.S. Public Health Service Grant NS19522.
Abbreviations
- nAChR
- nicotinic acetylcholine receptor
- ACh
- acetylcholine
- NCA
- noncompetitive antagonist
- V8 protease
- Staphylococcus aureus glutamyl endopeptidase
- GSSG
- oxidized glutathione
- [3H]HTX
- [3H]histrionicotoxin
- [125I]TID
- 3-(trifluoromethyl)-3-(M-[125I]iodophenyl)diazirine
- HTX
- dl-perhydrohistrionicotoxin
- H10-HTX
- dl-decahydro(pentenyl)histrionicotoxin
- PAGE
- polyacrylamide gel electrophoresis
- PCP
- phencyclidine
- TPS
- Torpedo physiological saline
- The American Society for Pharmacology and Experimental Therapeutics