Contributions of Torpedo Nicotinic Acetylcholine Receptor γTrp-55 and δTrp-57 to Agonist and Competitive Antagonist Function

Results of affinity-labeling studies and mutational analyses provide evidence that the agonist binding sites of the nicotinic acetylcholine receptor (nAChR) are located at the α-γ and α-δ subunit interfaces. For Torpedo nAChR, photoaffinity-labeling studies with the competitive antagonistd-[3H]tubocurarine (dTC) identified two tryptophans, γTrp-55 and δTrp-57, as the primary sites of photolabeling in the non-α subunits. To characterize the importance of γTrp-55 and δTrp-57 to the interactions of agonists and antagonists, Torpedo nAChRs were expressed inXenopus oocytes, and equilibrium binding assays and electrophysiological recordings were used to examine the functional consequences when either or both tryptophans were mutated to leucine. Neither substitution altered the equilibrium binding of dTC. However, the δW57L and γW55L mutations decreased acetylcholine (ACh) binding affinity by 20- and 7,000-fold respectively. For the wild-type, γW55L, and δW57L nAChRs, the concentration dependence of channel activation was characterized by Hill coefficients of 1.8, 1.1, and 1.7. For the γW55L mutant, dTC binding at the α-γ site acts not as a competitive antagonist but as a coactivator or partial agonist. These results establish that interactions with γ Trp-55 of theTorpedo nAChR play a crucial role in agonist binding and in the agonist-induced conformational changes that lead to channel opening.

Results of affinity-labeling studies and mutational analyses provide evidence that the agonist binding sites of the nicotinic acetylcholine receptor (nAChR) are located at the ␣-␥ and ␣-␦ subunit interfaces. For Torpedo nAChR, photoaffinity-labeling studies with the competitive antagonist d-[ 3 H]tubocurarine (dTC) identified two tryptophans, ␥Trp-55 and ␦Trp-57, as the primary sites of photolabeling in the non-␣ subunits. To characterize the importance of ␥Trp-55 and ␦Trp-57 to the interactions of agonists and antagonists, Torpedo nAChRs were expressed in Xenopus oocytes, and equilibrium binding assays and electrophysiological recordings were used to examine the functional consequences when either or both tryptophans were mutated to leucine. Neither substitution altered the equilibrium binding of dTC. However, the ␦W57L and ␥W55L mutations decreased acetylcholine (ACh) binding affinity by 20-and 7,000-fold respectively. For the wild-type, ␥W55L, and ␦W57L nAChRs, the concentration dependence of channel activation was characterized by Hill coefficients of 1.8, 1.1, and 1.7. For the ␥W55L mutant, dTC binding at the ␣-␥ site acts not as a competitive antagonist but as a coactivator or partial agonist. These results establish that interactions with ␥ Trp-55 of the Torpedo nAChR play a crucial role in agonist binding and in the agonist-induced conformational changes that lead to channel opening.
The nicotinic acetylcholine receptor (nAChR) 1 from Torpedo electric organ and vertebrate skeletal muscle is a pentameric transmembrane protein composed of four homologous subunits with a stoichiometry of ␣ 2 ␤␥␦ (reviewed in Refs. [1][2][3]. The nAChR contains two binding sites for agonists and competitive antagonists, located at the ␣-␥ and ␣-␦ subunit interfaces (4,5). The two sites are nonequivalent, and many competitive antagonists bind with high affinity to only one of the sites (6,7). Affinity labeling and mutational analyses provide evidence that amino acids from three discrete regions of ␣ subunit primary structure and from three (or more) regions of the ␥ (or ␦) subunit contribute to the structure of the binding sites (reviewed in Refs. 8 and 9).
In this report we examine the contributions of ␥Trp-55 and ␦Trp-57 as determinants of agonist binding and channel gating as well as determinants of competitive antagonist function. ␥Trp-55 and ␦Trp-57 were mutated to leucine, and the interaction of agonists and antagonists were examined using both binding assays and electrophysiological recording. Concentration-dependent inhibition of 125 I-␣-bungarotoxin (␣-BgTx) binding by agonists and antagonists was studied using wildtype and mutant Torpedo nAChRs in membranes isolated from Xenopus oocyte homogenates. In parallel experiments, we studied the activation of wild-type and mutant nAChRs by ACh using a two-electrode voltage clamp. Our results establish that the mutation ␥W55L has no effect on dTC equilibrium binding affinity. However, the mutation has a profound effect on ACh binding and on the gating of the ion channel. Furthermore, for the ␥W55L mutant nAChR, dTC acts not as an antagonist when bound to its high affinity site but as a coactivator with ACh.
In Vitro Transcription and Expression in Xenopus Oocytes-SP64based plasmids (pMXT) with cDNAs encoding wild-type ␣, ␥, ␦, and the ␥W55L subunits were gifts from Dr. Michael M. White, and the cDNA (in plasmid SP64) encoding the wild-type ␤ subunit was from Dr. Henry Lester. Sequence analysis of the ␦W57L mutant cDNA (from Dr. White) revealed a point deletion 353 bases 3Ј to the ␦W57L mutation point that resulted in a truncated form of the subunit. Therefore, the full-length ␦W57L mutant was prepared by subcloning a NheI (present in the vector) and BstXI fragment that included the desired mutation (␦W57L) but excluded the point deletion into the wild-type ␦ subunit cDNA from which the corresponding region had been excised. cDNAs were linearized with either XbaI (for wild-type ␣, ␥, ␦, ␥W55L, and ␦W57L mutant subunits) or FspI (for the wild-type ␤ subunit). In vitro transcription reactions were carried out in transcription buffer (Promega) containing 40 mM Tris (pH 7.5), 6 mM MgCl 2 , 2 mM spermidine, and 10 mM NaCl. Linear cDNAs (5-10 g) were incubated with 10 mM dithiothreitol, NTPs (1 mM each except GTP, which was 0.2 mM), 0.6 mM diguanosine triphosphate (Amersham Pharmacia Biotech), 100 units of RNasin (Promega), and 40 units of SP6 RNA polymerase (Promega) in transcription buffer at 37°C for 1 h. An additional 40 units of SP6 RNA polymerase was added after 1 h, and the incubation was continued for another hour. In some experiments, [ 3 H]UTP was included in the reaction mixture as a tracer for quantitation of reaction yields. RNAs were extracted with phenol/chloroform and chloroform, precipitated from isopropanol, and then resuspended in water at a concentration of Ϸ3 g/l. Isolated, follicle-free oocytes were microinjected with 0.5-10 ng of subunit-specific RNAs in a molar stoichiometry of (␣ 2 ␤␥␦) for wild-type and mutant receptors. For ␥-less and ␦-less receptors, nAChR subunit RNAs were mixed in a molar ratio of (␣ 2 ␤␦ 2 ) and (␣ 2 ␤␥ 2 ), respectively. Oocytes were injected with 10 ng of subunit RNA for assays of 125 I-␣-BgTx binding in intact oocytes or in membrane fractions and for all electrophysiological assays with the exception of full agonist-dose response relations for wild-type receptors. Because of the high currents produced by activation of wild-type receptors, maximal responses were determined for oocytes injected with 0.5 ng of RNA. Oocytes were maintained in ND96 buffer containing 96 mM NaCl, 2 mM KCl, 1.8 mM CaCl 2 , 1 mM MgCl 2 , 5 mM HEPES (pH 7.6), and 50 g/ml gentamicin for at least 48 h before use.
Binding of 125 I-␣-BgTx to Intact Oocytes and to Oocyte Membranes-To measure binding to nAChRs expressed on the surface, oocytes were incubated with 2.5 nM 125 I-␣-BgTx for 2 h in a final volume of 100 l of low Ca 2ϩ ND96 buffer containing 96 mM NaCl, 2 mM KCl, 0.3 mM CaCl 2 , 1 mM MgCl 2 , and 5 mM HEPES (pH 7.6). Oocytes were washed 3 times with 1 ml of ice-cold low Ca 2ϩ ND96 buffer containing 1% bovine serum albumin and counted in a ␥ counter. Nonspecific binding was determined using uninjected oocytes. Oocyte membranes were prepared by homogenizing oocytes in ice-cold homogenization buffer (HB, 0.1 ml/oocyte) containing 140 mM NaCl, 20 mM sodium phosphate, 1 mM EDTA, 1 mM EGTA, 1 mM phenylmethysulfonylfluoride, and 0.1 units of aprotinin/ml (pH 7.6). Membranes were isolated as described (12) by differential and sucrose density centrifugation. Membranes were resuspended in HB at 5 l per oocyte, frozen in liquid nitrogen, and stored at Ϫ80°C. A centrifugation assay (12) was used to measure the initial rate of 125 I-␣-BgTx binding to oocyte membranes in low Ca 2ϩ ND96 assay buffer (supplemented with the cholinesterase inhibitor diisopropylphosphofluoridate (0.1 mM) for ACh binding experiments). Binding assays were initiated by adding 125 I-␣-BgTx (a final concentration of 2.5 nM) to triplicate samples (2-3 oocytes/sample) in a total volume of 100 l. Membranes were preequilibrated for 15 min with appropriate concentrations of dTC, ACh, or other ligands before the addition of 125 I-␣-BgTx, and then after a 20-min incubation, the reactions were stopped by adding 1 M unlabeled ␣-BgTx. Under these conditions, in the absence of competing ligands 125 I-␣-BgTx binding was Ϸ30% of equilibrium. Nonspecific binding of 125 I-␣-BgTx to wild-type and mutant receptors was determined either in the presence of 100 M dTC (24 Ϯ 6% of total binding, n ϭ 8 experiments) or 100 mM ACh (27 Ϯ 15% of total binding, n ϭ 8 experiments) for dTC and ACh inhibition, respectively.
Biosynthetic Labeling and Immunoprecipitation of nAChRs-15-30 injected oocytes were incubated in 2 ml of low Ca 2ϩ ND96 (plus 50 g/ml gentamicin) containing 0.1-0.5 mCi/ml [ 35 S]methonine/cysteine (PerkinElmer Life Sciences) in 24-well plates at 18°C for at least 48 h before use. Ten to fifteen healthy oocytes were selected and washed five times with 2 ml of low Ca 2ϩ ND96 to remove excess [ 35 S]Met/Cys before the addition of ␣-BgTx (2.5-5 nM) to label surface receptors in a final volume of 1 ml. After a 2-h incubation, unbound ␣-BgTx was removed by washing oocytes three times with 1 ml of low Ca 2ϩ ND96. The oocytes were then used to prepare membranes containing ␣-BgTx-labeled surface receptors. To characterize internal receptors, membranes containing nonradioactive ␣-BgTx-prelabeled surface nAChRs were incubated with ␣-BgTx (2.5-5 nM) for another 2 h followed by washing. The membranes prepared in this manner contain both ␣-BgTx-labeled surface and internal nAChRs and are designated as "total receptors." The membranes containing either surface or total receptors were then resuspended in HB buffer supplemented with 1% Triton X-100. After solubilization for 30 min, the extracts were treated with 1% Immunoprecipitin (Life Technologies, Inc.) at 4°C for 20 min and centrifuged for 1 min in an Eppendorf microcentrifuge. Immunoprecipitin-pretreated supernatants were then collected and incubated with an excess amount of rabbit anti-␣-BgTx antibody at 4°C overnight. After overnight incubation, Immunoprecipitin (1%) was added to the nAChR-antibody complexes and incubated at 4°C for 30 min. The samples were then spun in a microcentrifuge for 1 min, and the supernatants were discarded. The immunoprecipitate was washed four times with HB buffer containing 1% Triton X-100, 0.1% SDS, and 0.5% bovine serum albumin and once with HB buffer containing 0.1% SDS and 0.05% Triton X-100 (without bovine serum albumin). After the last wash, the pellets were resuspended in 50 l of SDS-polyacrylamide gel electrophoresis sample buffer and incubated at room temperature for 20 -30 min. The samples were electrophoresed on an 8% SDS-polyacrylamide gel. After staining with Coomassie Blue and destaining, the gels were soaked in Amplify (Amersham Pharmacia Biotech) for 30 min, dried at 65°C for 2 h, and exposed to x-ray film at Ϫ80°C for 24 h.
Sucrose Density Gradient Analysis of nAChRs in Oocytes-To characterize surface nAChRs, 15-30 injected intact oocytes were incubated with 2.5 nM 125 I-␣-BgTx for 2 h followed by a wash, and oocyte membranes were then prepared as described above. The membranes were then solubilized with 1% Triton X-100 in a final volume of 200 l of HB buffer and centrifuged in a Ti 42.2 rotor at 35,000 rpm for 20 min. The supernatants (Ϸ180 l) were saved and used for sedimentation analysis. For total nAChRs, oocyte membranes containing surface nAChRs prelabeled by 125 I-␣-BgTx were then incubated with 2.5 nM 125 I-␣-BgTx for another 2 h. After a 2-h incubation, membranes were washed and solubilized as described above. 180-l extracts from 10 -15 oocytes containing either surface or total receptors were sedimented on a 10-ml 3-30% sucrose gradient (150 mM NaCl, 5 mM EDTA, 50 mM Tris pH 7.6, 1% Triton X-100, and 1 mM dithiothreitol) at 40,000 rpm in a SW 45 rotor (Beckman) for 22 h at 4°C. 200-l extracts from Torpedo membranes prelabeled with 125 I-␣-BgTx were used as the control for monomeric (9.5 S) and dimeric nAChR (13 S), and free 125 I-␣-BgTx (1.7 S), alkaline phosphatase (6.1 S), and ␤-galactosidase (15.9 S) were used as standards for sedimentation coefficients. Fractions (200 l) were collected and counted on a ␥ counter.
Electrophysiology-Currents elicited by ACh were measured using a standard two-electrode voltage clamp (Oocyte Clamp OC-725B, Warner Instrument Corp.) at a holding potential of Ϫ70 mV. Electrodes were filled with 3 M KCl and had resistances of 0.5-1.5 megaohms. The recording chamber (about 150 l in volume) was perfused continually by gravity with low Ca 2ϩ ND96 (plus 1 M atropine, pH 7.6). Appropriate concentrations of ACh (or other agonists) in the absence or presence of antagonists were applied through solenoid valves into the recording chamber for 3-5 s. For some experiments, oocytes were preincubated with dTC by perfusing the oocytes for ϳ1 min with dTC in low Ca 2ϩ ND96 before application for 5 s of solution containing ACh with the same concentration of dTC.
Data Analysis-The concentration-dependent inhibition of 125 I-␣-BgTx binding by agonists and antagonists was fit according to two models as follows.
f ϭ 100/͑1 ϩ ͓͑X]/IC 50 ͒ n ͒ (Eq. 1) where [X] is the concentration of inhibitor, n is the Hill coefficient, and IC 50 is the inhibitor concentration reducing the initial rate of 125 I-␣-BgTx binding by 50% and where [X] is the concentration of competing ligand, K H and K L are the ligand affinities for the high and low affinity binding sites, respectively. This equation is based on the assumption that ␣-BgTx binds at equal rates to the two sites.
For receptor activation, concentration-response curves for ACh and other agonists were fit to the following equation.
where I and I max are the currents at a given concentration of ACh and the maximal value, respectively, and K ap is the concentration of ACh required for half-maximal current. Because high concentrations of ACh do not result in a concentration-independent maximal response (due to desensitization and/or channel block), the Hill coefficients (n H ) for the agonist dose-response relations were estimated from the slope of plots of log I versus log[agonist] at currents less than 20% of the maximal response for each agonist. The dose-dependent inhibition of ACh-induced currents by antagonists was fit according to Equation 1 above. SigmaPlot (Jandel Scientific) was used for nonlinear least squares fit of the data, and the S.E. of the parameter fits are indicated in the Tables.

RESULTS
Influence of ␥W55L and ␦W57L on the Binding of d-Tubocurarine and Acetylcholine-Torpedo nAChRs were expressed in Xenopus oocytes by injection of cRNAs encoding ␣, ␤, ␥, and ␦ subunits. We characterized the equilibrium binding of antagonists and agonists by their inhibition of 125 I-␣-BgTx binding to Torpedo nAChRs in membranes isolated from oocyte homogenates. dTC inhibited 125 I-␣-BgTx binding to wild-type (␣ 2 ␤␥␦) nAChRs in a concentration-dependent manner with an IC 50 of 440 Ϯ 30 nM (Hill coefficient, n H ϭ 0.50 Ϯ 0.01) (Fig. 1A, see "Experimental Procedures"). These data were well fit by a two-site model with K H ϭ 50 nM (high affinity site) and K L ϭ 4 M (low affinity site) (see Table I), and the data were consistent with the equilibrium binding of [ 3 H]dTC to nAChRs in Torpedo membranes (5,6) and with the assumption that 125 I-␣-BgTx binds to the two agonist sites with equal association rate constants. dTC binds to ␦-less (␣ 2 ␤␥ 2 ) receptors with an IC 50 of 140 nM and to ␥-less (␣ 2 ␤␦ 2 ) receptors with an IC 50 of 8 M, consistent with the notion that for dTC the high and low affinity sites are formed at the ␣-␥ and ␣-␦ interfaces, respectively ( Fig.  1A, inset, and Table I). We determined the effects of the ␥W55L and ␦W57L mutations on 125 I-␣-BgTx binding and on dTC competition with 125 I-␣-BgTx binding. Control experiments indicated that ␥W55L and ␦W57L had no effect on 125 I-␣-BgTx binding affinity (data not shown). Despite the decrease of dTC potency as an inhibitor of nAChRs containing ␥W55L (Ref. 13 and see below), nAChRs containing either a ␥W55L (␥ m ), a ␦W57L (␦ m ) subunit, or both mutant subunits (␣ 2 ␤␥ m ␦ m ) had the same binding affinities for dTC as wild-type receptors (Fig.  1A). The concentration dependence for dTC inhibition of 125 I-␣-BgTx binding to mutant receptors was clearly different from that observed for either ␥-less or ␦-less nAChRs ( Fig. 1, inset), which indicates that the lack of effect of the mutations on dTC binding is not due to omission of either ␥ m or ␦ m subunit. Thus, substitution of ␥Trp-55 or ␦Trp-57 by leucine has no effect on dTC binding affinity. These tryptophan residues are, however, located near the agonist binding sites because they can be affinity-photolabeled by [ 3 H]dTC and [ 3 H]nicotine (10,11). The discrepancy between binding and functional antagonism will be explained below.
In contrast to what was observed for dTC binding, ␥W55L and ␦W57L dramatically increased the IC 50 for ACh competition against 125 I-␣-BgTx binding to nAChRs. ACh inhibited 125 I-␣-BgTx binding to the wild-type Torpedo nAChR expressed in oocytes in a concentration-dependent manner with an IC 50 of 340 nM and a Hill coefficient of 0.6 ( Fig. 1B). ␥W55L and ␦W57L increased the IC 50 for ACh binding by 62-and 8-fold, respectively, and for the double mutant (␣ 2 ␤␥ m ␦ m ), the IC 50 was 530-fold higher than for wild-type receptor (Fig. 1B). The IC 50 values for ACh binding to either ␣ 2 ␤(␥ m ) 2 or ␣ 2 ␤(␦ m ) 2 receptors were each increased by about 1,300-fold (Fig. 1B). The concentration dependence of ACh inhibition of 125 I-␣-BgTx binding reflects the binding of both ACh and 125 I-␣-BgTx to two agonist binding sites. For the wild-type receptor this concentration dependence was well fit by a two-site model with K H of 55 nM and K L of 2.7 M (Table I) The effects of the ␥W55L and ␦W57L mutations on ACh binding were clearly different than that observed for nAChRs lacking either ␥ subunit (␥-less, ␣ 2 ␤␦ 2 , IC 50 ϭ 0.8 M) or ␦ subunit (␦-less, ␣ 2 ␤␥ 2 , IC 50 ϭ 0.3 M), which were characterized by inhibition curves similar to wild-type (Fig. 1B, inset and Table I). As judged by inhibition of 125 I-␣-BgTx binding, ACh binding to Torpedo nAChRs expressed in oocytes also differed from the binding to native nAChRs in Torpedo membranes (IC 50 ϭ 25 nM and a Hill coefficient of 0.9. (Fig. 1B)).
The observed equilibrium binding reflects the binding affinity of the ACh sites in the desensitized nAChR and the conformational equilibrium between resting and desensitized states. To test whether the leucine substitution had a predominant effect on the latter parameter, we examined the effect of proadifen, a desensitizing noncompetitive antagonist (24), on the ACh equilibrium binding function. For wild-type and ␣ 2 ␤␥ m ␦ m , FIG. 1. Effects of ␥W55L and ␦W57L mutations on the equilibrium binding of dTC (A) and ACh (B) to Torpedo nAChRs. Equilibrium binding was determined by the inhibition of 125 I-␣-BgTx binding to Torpedo nAChRs in membranes isolated from oocyte homogenates for wild-type nAChR (q, ␣ 2 ␤␥␦) or mutant receptors con- . For ACh, binding was also determined for ␣ 2 ␤(␦ m ) 2 (‚) and ␣ 2 ␤(␥ m ) 2 (ƒ) mutant nAChRs as well as for native Torpedo membranes (छ). Assay aliquots (100 l) of Torpedo membranes contained 50 fmol of nAChR and membranes from 3 uninjected oocytes and were treated with 100 M diisopropylphosphofluoridate to inactivate cholinesterase. Insets, dTC (A) and ACh (B) binding to ␥-less (Ⅺ, ␣ 2 ␤␦ 2 ) and ␦-less (‚, ␣ 2 ␤␥ 2 ) nAChRs compared with wild-type (E). For each data set, the data points represent the mean Ϯ S.D. of triplicate samples from a single experiment representative of 2-4 experiments. Solid curves are calculated from the parameters of Table I for  proadifen produced only a modest (less than 3-fold) left shift of the ACh equilibrium binding function (data not shown), as it does for nAChRs in membranes from Torpedo electric organ (24), but in contrast to the 100-fold enhancement of ACh affinity seen for mouse muscle nAChR in the presence of proadifen (20).
Composition and Assembly of Subunits in Mutant nAChRs-To rule out the possibility that the perturbation of ACh binding resulted from nAChRs of altered subunit composition, we examined both the size and subunit composition of the wild-type and mutant nAChRs formed in our expression system. nAChR biosynthetic assembly intermediates can form high affinity binding sites for 125 I-␣-BgTx (␣ subunit alone) or agonists and competitive antagonists (␣/␥ or ␣/␦ subunit pairs) (4). The binding studies described above were performed with membranes isolated from oocyte homogenates that contain both surface receptors and receptors from intracellular membranes. This internal pool may contain agonist binding sites that are very different than those found in the pentameric nAChRs, since the internal ␣-BgTx binding sites may include nAChR assembly intermediates. We therefore compared surface and internal receptors in oocyte membranes by examining both the size(s) of the 125 I-␣-BgTx binding components by sedimentation analysis and the subunit composition by biosynthetic labeling and immunoprecipitation.
Sucrose density gradient analysis was carried out to determine the size(s) of the receptors expressed in oocytes. To label surface nAChRs, intact oocytes were incubated with 125 I-␣-BgTx before isolation of oocyte membranes, and isolated membranes were reincubated in 125 I-␣-BgTx to label all sites made accessible after homogenization of the oocytes (total receptors). Membranes were extracted in 1% Triton X-100, and the sedimentation properties of these nAChRs were compared (Fig. 2) with 125 I-␣-BgTx-labeled native Torpedo nAChRs (monomer, 9.5 S; dimer, 13 S) ( Fig. 2A). In the membranes isolated from oocyte homogenates, there were 3-5 times more total 125 I-␣-BgTx sites (Fig. 2, B-E, closed circles) than surface sites (Fig. 2,  B-E, open circles). However, the major populations of both surface and total 125 I-␣-BgTx binding sites in oocyte membranes were pentameric nAChRs, as revealed by a characteristic large peak of 125 I at 9.5S for wild-type as well as mutant nAChRs (␥ m , ␦ m , ␥ m ␦ m , Fig. 2, B-E). In some experiments (Figs. 2, B and D), total receptors in oocyte membranes also yielded a much smaller peak of 125 I at 5.0 S, which has been shown previously to be subunit pairs of either ␣␥ or ␣␦ subunits (25). The peak at 1.7 S in all our experiments represented free FIG. 2. Sucrose density gradient characterization of nAChRs containing ␥W55L and/or ␦W57L mutant subunits. Membranes were prepared from oocytes preincubated with 125 I-␣-BgTx to label surface nAChRs (E), and isolated membranes were further incubated with 125 I-␣-BgTx to label surface and internal ␣-BgTx binding sites (q, total receptors). Triton X-100 extracts of labeled membranes prepared from pools of 10 -15 injected oocytes were sedimented on 3-30% sucrose gradients. Fractions (200 l) were collected from the top of the gradient and counted in a ␥ counter. 125 I-␣-BgTx-labeled native Torpedo nAChRs in Triton X-100 extracts, which sedimented primarily as pentameric monomer (9.5 S) along with 13 S dimers, were used as controls (Panel A). For nAChRs expressed in oocytes (panels B-H), the total receptors (q) were 3-5 times more abundant than surface receptors (E) for When the mutant ␥ subunit was coexpressed with wild-type ␣ and ␤ subunits without the ␦ subunit (␣ 2 ␤(␥ m ) 2 , panel F), as seen for wild-type ␦-less receptor (␣ 2 ␤␥ 2 ) (data not shown), no stable subunit complexes were detected in Triton X-100. For nAChRs containing four different subunits, including either ␥ m , ␦ m , or ␥ m ␦ m , total as well as surface receptors sedimented as 9.5 S pentamers (panels C-E), whereas for receptors lacking the ␥ subunit but containing either ␦ or ␦ m , a 5.0 S peak of 125 I-␣-BgTx binding was also prominent (panels G and H), consistent with the expected sedimentation of dimeric forms containing ␣-␦ subunits. The peak of 125 I found at 1.7 S in all experiments resulted from the presence of free 125 I-␣-BgTx.
nAChR Agonist Binding Determinant 125 I-␣-BgTx. Thus, as for wild-type nAChRs, binding of 125 I-␣-BgTx to each of these mutants in membrane homogenates will reflect binding to assembled, pentameric nAChRs, and partial assembly cannot account for the altered ACh binding seen for the mutant nAChRs. When the mutant ␥ subunit was coexpressed with wild-type ␣ and ␤ subunits without the ␦ subunit (␣ 2 ␤(␥ m ) 2 , Fig. 2F), stable assembly of nAChR subunits was not observed in Triton X-100. This result, which was also seen for wild-type ␦-less receptor (␣ 2 ␤␥ 2 , data not shown), is consistent with previous observations (25,26). For the mutant ␦ subunit coexpressed with wild-type ␣ and ␤ subunits without the ␥ subunit ␣ 2 ␤(␦ m ) 2 , in addition to the 9.5 S peak of 125 I, there was a prominent 5 S component in the total homogenate, as was seen for wild-type ␥-less receptor (␣ 2 ␤␦ 2 ) (Figs. 2, G and H). These results indicated that partial assembly intermediates are not prominent for the mutant nAChRs containing all four subunits, but they may be more significant for wild-type or mutant nAChRs lacking the ␥ or ␦ subunit.
Subunit compositions of expressed wild-type and mutant nAChRs were characterized by immunoprecipitation of ␣-BgTx-labeled receptors (surface and total) extracted from oocyte homogenates after biosynthetic labeling with [ 35 S]Met/ Cys. Fig. 3 shows that for both surface (A) and total receptors (B), the presence of ␥W55L, ␦W57L, or both mutant subunits had no effect on subunit assembly, and the mutations had no effect on subunit glycosylation as judged by the mobilities of mutant compared with wild-type subunits. Furthermore, the mobilities of these subunits in nAChRs expressed in Xenopus oocytes were the same as the native Torpedo nAChR subunits (not shown). However, for nAChRs lacking the ␥ subunit, mutant ␦ subunit had the same mobility as wild type, but both had slightly higher mobility than ␦ subunits in nAChRs also containing the ␥ subunit. Similarly, omission of the ␦ subunit resulted in ␥ subunits of enhanced mobility (data not shown).
␥W55L and ␦W57L Alter nAChR Activation-We examined the effects of ␥W55L and ␦W57L on the activation of nAChRs expressed in oocytes using a two-electrode voltage clamp. When holding the oocyte membrane potential at Ϫ70 mV, there was a concentration-dependent activation of inward current upon 5-s application of ACh for wild-type (Fig. 4A) and mutant nAChRs (␣ 2 ␤␥ m ␦, Fig. 4B). For wild-type nAChR, preincubation with dTC produced a dose-dependent inhibition of ACh currents characterized by an IC 50 of 40 nM (Figs. 4, C and D), as expected when binding to its high affinity site results in functional antagonism. However, as reported previously (13) for nAChRs containing ␥W55L, dTC inhibited ACh currents only at higher concentrations (IC 50 ϭ 4 M, Fig. 4D), despite the presence of the high affinity dTC binding site (K H ϭ 40 nM, Fig. 1A). Based upon the peak transient current observed for each ACh concentration, the ACh concentration for half-maximal response (K ap ) for wild-type nAChRs was 20 Ϯ 2 M, whereas the K ap values for nAChRs containing ␥W55L or ␦W57L subunits were 170 Ϯ 30 M and 86 Ϯ 2 M, respectively (Fig. 5A). The double mutant receptor (␣ 2 ␤␥ m ␦ m ) was activated by ACh with a K ap of 340 Ϯ 40 M, about 15-fold higher than the wild-type receptor. The consequences of the leucine substitutions were clearly different from that seen for ␥ or ␦ subunit omission, since for ␥-less (␣ 2 ␤␦ 2 ), K ap ϭ 21 Ϯ 4 M and for ␦-less (␣ 2 ␤␥ 2 ), K ap ϭ 14 Ϯ 1 M (not shown). Fig. 5B shows a comparison of the number of 125 I-␣-BgTx binding sites on the oocyte surface and the current amplitudes for mutant AChRs activated by saturating concentrations of ACh. nAChRs containing either ␥W55L, ␦W57L, or both mutant subunits were expressed at nearly the same levels as the wild-type nAChR, as indicated by the number of surface 125 I-␣-BgTx binding sites. In contrast, the maximal currents for the mutant nAChRs were much lower than for wild type. When equal amounts of subunit cRNAs were injected for wild-type and mutant nAChRs, the maximal currents for the ␦W57L . Triton X-100 extracts containing either surface or total receptors labeled by ␣-BgTx were incubated with affinity-purified rabbit antibody against ␣-BgTx and immunoprecipitated with 1% Immunoprecipitin. Immunoprecipitates were extracted in sample buffer and analyzed by SDS-polyacrylamide gel electrophoresis. Gels were stained with Coomassie Blue, destained, and prepared for fluorography (36-h exposure). For both surface (A) and total (B) receptors (1, ␣ 2 ␤␥␦; 2, ␣ 2 ␤␥ m ␦; 3, ␣ 2 ␤␥␦ m ; 4, ␣ 2 ␤␥ m ␦ m ; 5, uninjected oocytes) ␥W55L and ␦W57L mutant subunits were expressed and assembled with other wild-type subunits as efficiently as seen for wild-type subunits.

FIG. 4. Electrophysiological recordings from wild-type and mutant nAChRs expressed in oocytes.
A, inward currents evoked by 5 s applications of ACh to an oocyte expressing wild-type nAChRs. The oocyte was injected with diluted RNA (2.5 ng, a molar ratio of 2:1:1:1 ␣:␤:␥:␦) to limit the current amplitudes at high ACh concentrations. The recordings were made 48 h after injection, and the oocyte membrane potential was held at Ϫ70 mV. B, concentration-dependent activation of ␥W55L mutant by ACh. The oocyte was injected with 100 ng RNA (2:1:1:1 ␣:␤:␥ m :␦), which resulted in currents of similar magnitude as seen for wild-type nAChR after injection of lower amounts of RNA. C, dTC inhibition of ACh-induced currents for wild-type nAChR. Currents evoked by 3 M ACh (3 s) from an oocyte injected with 10 ng of RNA were inhibited by preincubation with various concentrations of dTC. D, upon preincubation, dTC inhibited 3 M ACh-induced currents from wild-type (WT) nAChRs (q) with an IC 50 of 40 Ϯ 4 nM (n H ϭ 0.9 Ϯ 0.1). For ␥W55L mutant receptors (f) after preincubation, dTC inhibited 100 M ACh-induced currents with an IC 50 of 3.7 Ϯ 0.6 M (n H ϭ 0.9 Ϯ 0.1). The data represent the mean Ϯ S.D. of three measurements. mutant were twice as large as the currents seen for wild-type nAChR at 3 M ACh, a concentration producing ϳ5% of maximal currents. Thus, for nAChRs containing the ␦W57L subunit, the maximal currents were ϳ10% that of wild-type, whereas for receptors containing ␥W55L or both mutant subunits, the maximal currents were only 6 Ϯ 1 and 0.6 Ϯ 0.2% that seen for ␦W57L. These results indicate that mutation of ␥Trp-55 and ␦Trp-57 alters the activation of these channels by ACh.
The Hill coefficients (n H ) characterizing the ACh dose-response relations were estimated from log-log plots of current amplitude (I) versus [ACh] at concentrations of ACh producing currents less than 20% of maximal responses. This is a reliable method of determining n H for ACh responses without reference to the experimentally determined maximal currents, which are limited by desensitization and/or channel block. For the wildtype and ␦W57L mutant receptors, the concentration-response relationship had slope values (n H ) of 1.8 Ϯ 0.1 and 1.7 Ϯ 0.1, respectively (Fig. 5C). In contrast, the slope for the ␥W55L mutant was one (n H ϭ 1.1 Ϯ 0.1) (Fig. 5C).
dTC Potentiation of Agonist-induced Activation in ␥W55L nAChRs-For the wild-type nAChR, upon preincubation, dTC acted as a competitive antagonist characterized by an IC 50 of 40 nM (Fig. 4D), and without preincubation, when coapplied with 3 M ACh for 5 s, dTC inhibited with an IC 50 of 250 nM (Fig. 6C). Similar results were obtained for nAChRs containing ␦W57L (data not shown). For nAChRs containing ␥W55L, despite the fact that dTC binds with high affinity (K H ϭ 40 nM) to one of the sites (Fig. 1A), upon preincubation, dTC inhibited ACh currents only at high concentrations (IC 50 ϭ 4 M, Fig.  4D), as reported by O'Leary et al. (13). We therefore examined in greater detail the interactions of dTC with nAChRs containing ␥W55L. When applied alone, dTC produced no detectable whole cell currents (Ͻ10 nA, data not shown). However, we observed that at low concentrations dTC (10 nM to 1 M) actually potentiated currents activated by ACh (10 -100 M) when both ligands were applied simultaneously to individual oocytes (Fig. 6). The magnitude of the potentiation was dependent on . The maximal currents for the ␦W57L mutant were twice as large as the currents seen for wild-type receptors at 3 M ACh (hatched bar), a concentration producing only 5% of maximal currents. The maximal currents (closed bars) induced by saturating concentrations of ACh for ␥W55L (␣ 2 ␤␥ m ␦, 1.5 Ϯ 0.3 A, n ϭ 6 oocytes) and for ␥W55L/ ␦W57L (␣ 2 ␤␥ m ␦ m , 0.16 Ϯ 0.04 A, n ϭ 6 oocytes) mutants were only 6 and 0.6% that seen for the ␦W57L mutant (␣ 2 ␤␥␦ m , 25 Ϯ 4 A, n ϭ 5 oocytes). The data represent the means Ϯ S.D. C, Hill coefficients for ACh activation of wild-type and mutant nAChRs, determined from the concentration-response relationship at concentrations of ACh producing less than 20% maximal currents. ACh-induced currents (A) recorded from wild-type (q), ␥W55L (f), and ␦W57L (OE) receptors were plotted logarithmically against the concentration of ACh, with each data point the mean Ϯ S.D. of three recordings. The data for a single oocyte were fit by linear regression, and the slope of the line (n H ) for the wild-type receptor (q) was 1.85, with the same analysis for data from three oocytes characterized by a slope of 1.8 Ϯ 0.1. For the ␥W55L mutant (f), the slope for this representative experiment was 1.04, and the average slope from three experiments was 1.10 Ϯ 0.05. The ␦W57L mutation, however, had no effect on the slope of the concentrationresponse relationship of ACh for receptor activation. The slope of the line for the ␦W57L mutant (OE) presented here was 1.71, and the average value of the slope from three experiments was 1.66 Ϯ 0.14. the concentration of both dTC and ACh (Fig. 6C), with the concentration dependence of potentiation by dTC for 10 M ACh consistent with dTC binding to its high affinity binding site. The largest dTC potentiation was observed at low concentrations of ACh, and the magnitude of the potentiation decreased with higher concentrations of dTC. Concentrations above 10 M dTC only inhibited currents activated by any concentration of ACh (Fig. 6C). For ␥W55L nAChRs preincubated with dTC, no potentiation was ever seen for responses to ACh at concentrations between 0.3 and 100 M.
For the mutant receptor containing ␥W55L, potentiation by dTC was also observed for responses to two other agonists, carbamylcholine and suberyldicholine (Fig. 7). As for ACh, for these agonists the Hill coefficient (n H ) characterizing the doseresponse relation was reduced from 1.6 for wild-type nAChRs to ϳ1 for the mutant receptor (Fig. 7A). dTC at concentrations between 10 nM and 1 M produced a dose-dependent enhancement of responses for agonist concentrations producing submaximal responses. Higher dTC concentrations produced a progressive inhibition of the currents. For carbamylcholine, dTC enhanced currents as much as 3-fold (Fig. 7B), whereas for suberyldicholine, currents were increased by about 2-fold (Fig.  7C). For ␣ 2 ␤␥ m ␦, potentiation of agonist-induced currents was not limited to dTC, since metocurine, the 7Ј,12Ј-dimethoxy-2methyl dTC analog, also potentiated activation of the mutant receptor by ACh (Fig. 7D), However, neither 13Ј-iodo-dTC nor two other competitive antagonists, pancuronium and gallamine, potentiated ACh responses (Fig. 7D). These results establish that when dTC (or metocurine) first binds to nAChRs containing ␥W55L, it does not act as an antagonist and, in contrast, acts as a coactivator or weak partial agonist, as evidenced by the potentiation of responses seen when agonists bind to the site at the ␣-␦ subunit interface.
Influence of ␥W55L/␦W57L on the Binding of Nicotinic Agonists and Antagonists-We also examined the effects of the double mutation (␣ 2 ␤␥ m ␦ m ) on the binding affinities of several agonists and antagonists to learn more about the effects of these substitutions on the binding of structurally diverse agonists and antagonists (Fig. 8). Most of the agonists tested, including tetramethylammonium, phenyltrimethylammonium, suberyldicholine, and epibatidine, inhibited 125 I-␣-BgTx binding to double mutant receptors with IC 50 values that were ϳ50 -500-fold larger than for the wild-type receptor. Thus, mutation of ␥Trp-55 and ␦Trp-57 influences the binding of most nicotinic agonists, even very small agonists like tetramethylammonium. It is interesting that the agonists with highest affinity for wild-type receptors were affected the most by the double mutation. One exception to the general pattern was nicotine, which bound to the wild-type and double mutant receptors with very similar affinity (Fig. 8F).  H). The double mutation ␥W55L/ ␦W57L (E, ␣ 2 ␤␥ m ␦ m ) had no effect on nicotine binding (F) but increased the IC 50 of the other agonists by 50 -500 fold (see Table II). The double mutation had no effect on gallamine (G) binding but decreased the binding affinity of pancuronium (H) at the high affinity site. The data represent the mean Ϯ S.D. (three samples/point).
The effects of the double mutation (␣ 2 ␤␥ m ␦ m ) on the binding affinities of competitive antagonists are more diverse and complex than that seen for the agonists. The double mutation had no effect on the binding affinity of gallamine (Fig. 8G), similar to what was observed with dTC ( Fig. 1). Pancuronium was bound with 10,000-fold selectivity by wild-type nAChRs (K H ϭ 3 nM, K L ϭ 20 M), with binding at the high affinity site weakened by 100-fold in the double mutant and binding at the low affinity site weakened by less than 3-fold (Table II). The binding affinity of dihydro-␤-erythroidine was decreased by about 10-fold at each site by the double mutation. The results of the binding experiments shown in Fig. 8 are summarized in Table II. DISCUSSION A wide body of evidence establishes that the two agonist binding sites in muscle-type nAChRs are positioned at the interfaces of ␣-␥ and ␣-␦ subunits, with amino acids contributed from three distinct regions of ␣ subunit primary structure and at least three regions of ␥ (or ␦) subunit (reviewed in Refs. 3, 8, and 9). We set out to study the functional contribution of two tryptophans, ␥Trp-55 and ␦Trp-57, at homologous positions of the ␥ and ␦ subunits. Tryptophans ␥Trp-55 and ␦Trp-57 are within or near the agonist/competitive antagonist binding sites, since they are the principle sites of specific photoincorporation by [ 3 H]dTC within those subunits (10), and ␥Trp-55 is the amino acid in the ␥ subunit specifically photolabeled by [ 3 H]nicotine (11). In addition, they appeared likely to contribute to dTC binding affinity, because replacement of ␥Trp-55 by leucine resulted in a decrease of dTC potency as an inhibitor of ACh-induced currents for Torpedo nAChRs expressed in Xenopus oocytes (13). Although we confirmed the observation that for the ␥W55L mutant nAChR, the IC 50 for dTC inhibition was increased 100-fold compared with wild type (Fig. 4), we were surprised to find that the mutation of either or both tryptophans had no effect on dTC equilibrium binding affinity, based upon the inhibition of binding of 125 I-␣-BgTx (Fig. 1). Instead, our results indicate that the shift in dTC potency as an inhibitor of ACh-induced currents was a secondary consequence of a dramatic reduction in the affinity of ACh binding at the ␣-␥ site.
Although the leucine substitutions had no effect on dTC binding, on nAChR subunit assembly as judged by sucrose density gradient velocity sedimentation (Fig. 2) and immunoprecipitation (Fig. 3), or on the level of surface expression of nAChRs (Fig. 5B), the substitutions clearly altered agonist interactions as evidenced by the perturbation of the ACh-induced currents (Figs. 4 and 5). The ACh concentrations producing half-maximal currents (K ap ) were shifted to the right 8-and 20-fold for the ␥W55L and for ␥W55L/␦W57L mutant nAChRs, and although there were no significant changes in the levels of surface expression, the maximal currents were only 0.5 and 0.05%, respectively, that seen for wild-type nAChR. In addition, for the ␥W55L mutant, there was a clear shift in the concentration dependence of the response compared with the wild-type or the ␦W57L mutant receptor. For the ␥W55L mutant the observed Hill coefficient (n H ), determined at ACh concentrations producing Ͻ20% maximal responses, was close to 1 (n H ϭ 1.1 Ϯ 0.1), whereas it was close to 2 for wild-type (n H ϭ 1.8 Ϯ 0.1) and for the ␦W57L mutant (n H ϭ 1.7 Ϯ 0.1).
Although dTC binding affinity at the ␣-␥ site was unaltered by the replacement of ␥Trp-55 by leucine, concentrations of dTC sufficient to occupy the ␣-␥ site did not inhibit AChinduced currents. After pre-equilibration with nAChRs, dTC did not inhibit the ACh response until present at concentrations sufficient to occupy the ␣-␦ site (IC 50 ϭ 4 M, Fig. 4D). When applied simultaneously with ACh, dTC at concentrations sufficient to occupy the ␣-␥ site actually potentiated the ACh responses seen at concentrations less than 100 M (Fig. 6). Therefore, replacement of ␥Trp-55 by leucine allows dTC to function as a partial agonist or coactivator; dTC binding to the ␣-␥ site in the absence of ACh did not produce measurable currents, but when coapplied with ACh, dTC potentiated the ACh response by as much as 5-fold. These functional consequences of the replacement of ␥Trp-55 by leucine were not limited to ACh, since the concentration dependence for the current responses seen for carbamylcholine and suberyldicholine were also characterized by Hill coefficients of 1, and those responses were potentiated by dTC (Fig. 7). Not all competitive antagonists act as ACh coactivators of the ␥W55L mutant nAChR. Although potentiation was seen for dTC and its close structural analog metocurine, pancuronium and gallamine remained antagonists (Fig. 7D).
The leucine substitutions, especially the replacement of ␥Trp-55, caused a major perturbation of the equilibrium binding of ACh, as deduced from the inhibition of the initial rate of 125 I-␣-BgTx binding (Fig. 1B). The equilibrium binding function reflects both the affinity of binding to the desensitized state of the nAChR and the conformational equilibrium between resting and desensitized states. Before considering the effects of the substitutions, it is important to note several aspects of the observed binding to wild-type Torpedo nAChRs expressed in oocytes. The equilibrium binding of ACh by wildtype Torpedo nAChRs expressed in oocytes was well fit by a two-site model with an equal number of high and low affinity sites (K H ϭ 55 nM, K L ϭ 3 M). The parameters for ACh binding Torpedo ␣ 2 ␤␥␦ are similar to those seen for embryonic mouse nAChR expressed in oocytes (16), but they were quite different than those seen for ACh binding to Torpedo nAChR-rich membranes, which was characterized by high affinity (K ϭ 25 nM) binding to a single site (Fig. 1B). We do not know the source of this difference, but it does not appear to result from a shift of the preexisting equilibrium between resting and desensitized states, since the desensitizing noncompetitive antagonist proadifen had similar effects on ACh binding to either the native or expressed Torpedo nAChR. A noteworthy distinction between ACh interactions with Torpedo and embryonic mouse nAChR (␣ 2 ␤␥␦) is that our data indicate that in the Torpedo nAChR, ACh binds with higher affinity to the ␣-␥ than to the ␣-␦ site, whereas it binds with higher affinity at the ␣-␦ site for the mouse nAChR (16,27). The latter conclusion was based upon the observation that mouse ␣ 2 ␤␦ 2 binds ACh with 15-fold higher affinity than ␣ 2 ␤␥ 2 . For Torpedo nAChRs, ACh binds nonequivalently to the two sites even for receptors lacking the ␥ or ␦ subunit, but ACh binds with higher affinity to ␣ 2 ␤␥ 2 than to ␣ 2 ␤␦ 2 (Fig. 1B, Table I). For mouse nAChRs, preferential agonist binding to the ␣-␦ site is not a general rule. For the adult nAChR containing an ⑀ subunit in place of the ␥ subunit, ACh binds nonselectively at the two sites, whereas epibatidine binds with higher affinity at the ␣-␥ (or ␣-⑀) site than at the ␣-␦ site, and for carbamylcholine the rank order is ␣-␦ ϳ ␣-⑀ Ͼ ␣-␥ (28,29).
Replacement of ␥Trp-55 by leucine had a much larger effect on ACh equilibrium binding than did replacement of ␦Trp-57 ( Fig. 1 and Table I). Although the high and low affinity sites characteristic of the ACh equilibrium binding to wild-type nAChRs cannot be assigned unambiguously to binding at ␣-␥ and ␣-␦ sites, the observed binding by wild-type and mutant nAChRs is consistent with a simple interpretation if the high affinity ACh binding is at the ␣-␥ site in wild-type nAChR, and the effects of the mutations are greatest at the binding site containing the mutation. Then ACh binds with high affinity (K H ϭ 50 nM) to the ␣-␥ site in wild-type nAChR, and that nAChR Agonist Binding Determinant agonist, tetramethylammonium, equilibrium binding is weakened by 100-fold for the ␥W55L mutant, for ␥D174N (16), and also by substitutions at ␣Tyr-93, ␣Tyr-190, and ␣Tyr-198 (34). It is unlikely that tetramethylammonium is simultaneously in contact with all of these side chains.
Although the presence of ␥Trp-55 within the agonist binding site was identified on the basis of its photolabeling by [ 3 H]dTC, leucine substitution has no effect on dTC binding affinity. It is likely that replacement by other amino acids will alter dTC binding affinity, because even the leucine substitution weakens the binding of the competitive antagonists dihydro-␤-erythroidine and pancuronium by 10-and 70-fold, respectively. In addition, substitution of the corresponding tryptophan in homooligomeric ␣7 nAChRs (␣7Trp54) by histidine weakened dihydro-␤-erythroidine binding by 10-fold (36).
Our results establish the importance of ␥Trp-55 for agonist binding and channel gating, but they do not establish that ACh interacts directly with this tryptophan. In studies of mutant Torpedo nAChRs containing cysteines within the binding site, there was no evidence that ␥W55C was accessible for reaction with cationic methylthiosulfonates, whereas an adjacent position (␥E57C) as well as ␣Y93C and ␣Y198C were all accessible for modification (37). This result, in conjunction with the fact that ␥Trp-55 is photolabeled by [ 3 H]dTC and [ 3 H]nicotine, suggests that ␥Trp-55 may be within a hydrophobic subdomain of the binding site.
Substitutions at positions equivalent to nAChR ␥Trp-55 also have important functional consequences for other members of the superfamily of ligand-gated ion channels related to the nAChRs. For ␣1␤2␥2 ␥-aminobutyric acid type A receptors, replacement of ␣1Phe-64 by leucine results in 200-fold decrease of ␥-aminobutyric acid potency, whereas substitution of the equivalent positions in the other subunits had no effect on agonist potency (38). In contrast, substitution at the equivalent position in ␥2 subunit (␥2Phe-77) selectively alters binding of ligands at the benzodiazepine site (39). The serotonin 5-HT 3 receptor also contains a tryptophan (Trp-89) at the equivalent position, and substitutions at the position reduce antagonist but not agonist affinity (40).
The results presented here demonstrate that for the Torpedo nAChR a single mutation (␥W55L) within a part of the agonist binding site contributed from the ␥ subunit can effectively prevent ACh binding at that site without altering ACh binding at the ␣-␦ site. In addition, for this mutation, dTC binding to the ␣-␥ site acts as a partial agonist or coactivator of ACh responses. Since dTC binds at equilibrium to the ␣-␥ site in the mutant nAChR with the same high affinity as in wild type, it is most likely that the mutation does not alter dTC binding to any conformational state. Instead, by altering the interactions of ACh with the ␣-␥ site, the mutation reveals an aspect of dTC interaction with wild-type Torpedo nAChR that is not normally seen because of the relative affinities of dTC and ACh for the binding sites.