Amino Acid Residues That Confer High Selectivity of the α6 Nicotinic Acetylcholine Receptor Subunit to α-Conotoxin MII[S4A,E11A,L15A]*

Nicotinic acetylcholine receptors (nAChRs) containing α3 and β2 subunits are found in autonomic ganglia and mediate ganglionic transmission. The closely related α6 nAChR subtype is found in the central nervous system where changes in its level of expression are observed in Parkinson's disease. To obtain a ligand that discriminates between these two receptors, we designed and synthesized a novel analog ofα-conotoxin MII, MII[S4A,E11A,L15A], and tested it on nAChRs expressed in Xenopus oocytes. The peptide blocked chimeric α6/α3β2β3 nAChRs with an IC50 of 1.2 nm; in contrast, its IC50 on the closely related α3β2 as well as non-α6 nAChRs was three orders of magnitude higher. We identified the residues in the receptors that are responsible for their differential sensitivity to the peptide. We constructed chimeras with increasingly longer fragments of the N-terminal ligand binding domain of the α3 subunit inserted into the homologous positions of the α6 subunit, and these were used to determine that the region downstream of the first 140 amino acids was involved. Further mutagenesis of this region revealed that the α6 subunit residues Glu-152, Asp-184, and Thr-195 were critical, and replacement of these three residues with their homologs from the α3 subunit increased the IC50 of the peptide by >1000-fold. Conversely, when these key residues inα3 were replaced with those fromα6, the IC50 decreased by almost 150-fold. Similar effects were seen with other α6-selective conotoxins, suggesting the general importance of theseα6 residues in conferring selective binding.

␣6 can form receptors in combination with ␤2 through ␤4, leading to a large assortment of different receptor subtypes.
The ␣6 subunit has a limited distribution within the brain, largely being expressed by catecholaminergic neurons (3,4) at presynaptic endings where its activity modulates release of dopamine (5,6) and norepinephrine (7). Chronic exposure to nicotine has been shown to selectively affect the expression and function of nAChRs on striatal dopaminergic terminals containing this subunit (8 -10). Furthermore, ␣6 nAChRs appear to play a role in the pathophysiology of Parkinson's disease, a disease involving the loss of dopaminergic neurons. In animal models of Parkinson's disease, 1-methyl 4-phenyl 1,2,3,6-tetrahydropyridine-induced injury of dopaminergic neurons is associated with a selective loss of ␣6* 3 nAChRs in both rodents and monkeys (10 -12). In addition, there is selective loss of ␣6 nAChRs in human Parkinson's disease. Thus, ligands that selectively target ␣6␤2* nAChRs potentially represent a novel class of therapeutic agents (13).
However, it is difficult to pharmacologically distinguish ␣3␤2 from ␣6␤2* nAChRs due to the close structural similarities of the ␣ subunits. ␣3␤2* nAChRs are found on peripheral autonomic ganglia where they modulate cardiac and enteric functions (14,15). Therapeutic agents targeting CNS ␣6␤2* nAChRs would need to be devoid of activity on ␣3␤2 nAChRs to avoid cardiovascular and intestinal side effects. ␣-Conotoxin (CTX) MII, a 16-amino acid peptide, is the signature ligand for ␣6␤2* nAChRs; however, ␣-CTX MII also blocks ␣3␤2 nAChRs with high affinity (16). This is not surprising, given the overall high sequence homology between ␣6 and ␣3 subunits and the conservation of residues responsible for interaction with ␣-CTX MII (17). In this study, we created a novel ligand, based on ␣-CTX MII, that discriminates between ␣6/␣3␤2␤3 and ␣3␤2 nAChRs by a factor of 1000 (␣6/␣3 is a chimera that contains the N-terminal binding region of the ␣6 subunit and the remaining fragment of the ␣3 subunit, see "Experimental Procedures"). We constructed receptor chimeras and mutants and determined the non-homologous residues near the ligand binding sites of ␣6 and ␣3 that are responsible for this selectivity. The results of the present study provide insight into the nAChR-ligand interaction and may aid in development of ␣6* nAChR-selective therapeutics.
Construction of Chimeras-Chimeras were made by PCR using the primer sequences shown in Table 1. Because the ␣6 subunit does not form functional receptors, we used a functional surrogate formed by splicing the N-terminal extracellular ligand-binding region of the ␣6 subunit with the remaining fragment of the ␣3 subunit as previously described (19). Hence, all chimeras of ␣6 used in this study had portions of the N-terminal ligand-binding region of the ␣6 subunit replaced with the corresponding region of the ␣3 subunit. The notation for these chimeras is the length of the ␣3 sequence at the N-terminal portion, followed by the remaining ␣6/␣3 sequence, which makes up the C-terminal portion. Primers were designed to amplify each length of the extracellular region of the ␣3 subunit plus a 25-bp 3Ј-overhang homologous to the remaining extracellular region of the ␣6 subunit. Similarly, each ␣6/␣3 subunit fragment was amplified that contained a 25-bp 5Ј-overhang homologous to the ␣3 subunit. Amplifications of the ␣3 and ␣6/␣3 fragments were carried out by using Taq DNA polymerase (Promega, Madison, WI). Each corresponding ␣3 and ␣6/␣3 PCR fragment (containing the overhangs) was first hybridized and subsequently extended by PCR, using primers specific for a 5Ј-segment of the ␣3 sequence (containing an NotI site: 5Ј-AAGGAAAAAAGCGGCCGCGACATGGGT-GTTGTGCTGCTC-3Ј) and a 3Ј-segment of the ␣6/␣3 sequence (containing an XbaI site: 5Ј-GTCCATCTAGACACAGGT-GAGCCTCGATG-3Ј), using PfuTurbo DNA polymerase (Stratagene, La Jolla, CA). The amplified ␣3 1-x ␣6/␣3 PCR products were cloned into pT7TS plasmid (a modified pGEM4Z plasmid containing a 5Ј and a 3Ј Xenopus globin untranslated region), using the NotI and SpeI sites. The ligated products were transformed into either DH10B or HB101 competent cells and grown overnight, and the cDNA was isolated using a miniprep kit (Qiagen, Valencia, CA) and subsequently sequenced.
Construction of Point Mutations-Point mutants were made by PCR using primers shown in Table 1. Primers containing the desired point mutation flanked by 15 bases on either side were synthesized. Using the non-strand displacing action of PfuTurbo DNA polymerase, the mutagenic primers were extended and incorporated by PCR. The methylated, non-mutated parental cDNA was digested with DpnI. The mutated DNA was transformed into DH10B or DH5␣ competent cells and isolated using the Qiagen miniprep kit and sequenced to ascertain the incorporation of the desired mutation.
cRNA Preparation and Injection-Capped cRNA for the various subunits were made using the mMessage mMachine in vitro transcription kit (Ambion, Austin, TX) following linearization of the plasmid. The chimeras/point mutants made from the original ␣6/␣3 chimera were linearized with SalI, and point mutants originally made from the ␣3 subunit were linearized with EcoRI, and transcribed with T7 and SP6, respectively. The cRNA was purified using an RNeasy kit (Qiagen). The concentration of cRNA was determined by absorbance at 260 nm. cRNA of each chimera and point mutant were combined with cRNA of high expressing ␤2 and ␤3 subunits (in the pGEMHE vector) to give 167-500 ng/l of each subunit cRNA. 100 nl of this mixture was injected into each Xenopus oocyte with a Drummond microdispenser (Drummond Scientific, Broomall, PA), as described previously (16), and incubated at 17°C. Oocytes were injected within 1 day of harvesting, and recordings were made 2-4 days post-injection.
Voltage Clamp Recording-Oocytes were voltage-clamped and exposed to ACh and peptide as described previously (16). Briefly, the oocyte chamber consisting of a cylindrical well (ϳ30 l in volume) was gravity-perfused at a rate of ϳ2 ml/min with ND-96 buffer (96.0 mM NaCl, 2.0 mM KCl, 1.8 mM CaCl 2 , 1.0 mM MgCl 2 , 5 mM HEPES, pH 7.1-7.5) containing 1 M atropine and 0.1 mg/ml bovine serum albumin. In the case of the ␣9␣10 subtype, the ND96 contained no Mg 2ϩ and no atropine, and the oocytes were incubated in 1,2-bis(2-aminophenoxy)ethane-N,N,NЈ,NЈ-tetraacetic acid tetrakis (acetoxymethyl ester) for 3-4 h prior to recording. The oocyte was subjected once a minute to a 1-s pulse of 100 M ACh. For screening of receptor chimeras and mutants, for toxin concentrations of 1 M and lower, once a stable baseline was achieved, either ND-96 alone or ND-96 containing varying concentrations of the ␣-conotoxins was perfusion-applied, during which 1-s pulses of 100 M ACh were applied every minute until a constant level of block was achieved. For toxin concentrations of 10 M and higher, the buffer flow was stopped and the toxin was bath-applied and allowed to incubate with the oocyte for 5 min, after which the ACh pulse was resumed.
ACh Dose Response-To acquire ACh dose-response data, the conventional oocyte chamber was replaced by a chamber

Sequences of primers used to construct chimeras and point mutants
For the chimeras, the numbering reflects the length of the ␣3 N-terminal fragment substituted into the corresponding region of the ␣6 subunit. For the point mutants, the first amino acid designates the wild-type residue at the numbered location that is replaced with the second amino acid, which is found at the location in the opposite subunit (␣3 in case of ␣6/␣3 and ␣6/␣3 in case of ␣3).

Chimera/mutant Primer
constructed from a disposable 200-l polypropylene pipette tip with a length of 50 mm and an internal diameter of 0.5 mm at the upstream or intake end and 5 mm at the downstream or exhaust end. The chamber was mounted horizontally with its intake end connected to the perfusion supply, while its exhaust end had a vertical meniscus whose location was dictated by the tip of a sipper made from a 27-gauge hypodermic needle connected to a vacuum line. The chamber had two apertures in its dorsal wall: 1) a 1.5-mm circular hole centered 13 mm downstream from the intake, and 2) a 2.5 ϫ 5 mm (longitudinal) oval centered 14 mm downstream from the hole (i.e. a total of 27 mm from the intake end). The oocyte was introduced into the chamber through the oval aperture and secured against the chamber floor by the two voltage clamp glass microelectrodes that impaled the oocyte. The chamber was perfused at a rate of ϳ2 ml/min. To introduce ACh into the chamber, the perfusion was halted and 20 l of ACh was manually applied to the chamber via the small circular hole upstream from the oocyte. This volume was too small for ACh to reach the oocyte unless the perfusion was resumed. Upon resumption of perfusion (which was started immediately following the introduction of ACh into the chamber), the bolus of ACh rapidly engulfed the oocyte and washed past it in a matter of seconds, as judged by the time course of ACh response. This process was repeated with different concentrations of ACh with a time interval between applications long enough to avoid desensitization. Data Analysis-For the baseline response, at least three ACh responses were averaged. To determine the percent block induced by toxin, two to three ACh responses, obtained after a steady-state block had been achieved, were averaged, and the value was divided by the pre-toxin baseline value to yield a % response. The dose-response data were fit to the equation, Y ϭ 100/(1 ϩ 10 ∧ ((log EC 50 Ϫ log[Toxin]) ϫ n H )), where n H is the Hill coefficient, by non-linear regression analysis using Graph-Pad Prism (GraphPad Software, San Diego, CA). Each data point is mean Ϯ S.E.M. from at least three oocytes. For ACh dose-response curves, the response to a given ACh concentration was normalized to the response to 100 M ACh, which served as an internal control.

RESULTS
␣-CTX MII[S4A,E11A,L15A] Is a Selective Antagonist of ␣6* nAChRs-␣-CTX MII is a two-disulfide bridge peptide with the sequence, GCCSNPVCHLEHSNLC-amide. It has high, but similar, potency in blocking both ␣6␤2and ␣3␤2-containing nAChRs (IC 50 values of 0.39 and 2.2 nM, respectively (19)). Structure-function studies of ␣-conotoxin MII indicate that several of the non-Cys residues are less important for activity against ␣6 versus ␣3 nAChRs (19). We replaced three of these residues, Ser-4, Glu-11, and Leu-15, with Ala and tested them against heterologously expressed ␣3␤2 and ␣6/␣3␤2␤3 nAChRs. Because the ␣6 subunit does not form functional receptors, we used a functional surrogate formed by splicing the N-terminal extracellular ligand-binding region of the ␣6 subunit with the remaining fragment of the ␣3 subunit. The pharmacology of this chimera has previously been shown to match that of native ␣6␤2* nAChRs for ␣-CTX MII binding (19). The combined Ala substitutions resulted in a peptide with Ͼ600-fold lower activity against ␣3␤2 nAChRs but only 3-fold lower activity against ␣6/␣3␤2␤3 nAChRs compared with native ␣-CTX MII (Table 2). Thus, the resulting analog has Ͼ1000fold preference for ␣6/␣3␤2␤3 over ␣3␤2 nAChRs ( Table 2). The rate of recovery from block was slower for ␣6/␣3␤2␤3 than for ␣3␤2 nAChRs (Fig. 1). A 1-min wash was sufficient for recovery of ␣3␤2 nAChRs, whereas full recovery of ␣6/␣3␤2␤3 nAChRs required a 5-min wash.  ␣3␤2 nAChRs (bottom series) were voltage clamped at Ϫ70 mV and subjected to a 1-s pulse of 100 M ACh every minute as described under "Experimental Procedures". In each series, the first three responses are controls, following which the oocyte was exposed to the toxin as indicated. The perfusion and ACh pulses were then resumed to monitor the recovery from block during washout of toxin. The toxin was more potent in blocking ␣6/␣3␤2␤3 than ␣3␤2 nAChRs (note 10 4 -fold difference in toxin concentration) (see also Fig.  3). Likewise, the block of ␣6/␣3␤2␤3 nAChRs was more slowly reversible than that of ␣3␤2 nAChRs. nAChR-The N-terminal binding region (NTR) of ␣6 and ␣3 nAChR subunits displays ϳ80% homology (Fig. 2). To determine the residues responsible for the differential interaction with the peptide, a series of chimeras were constructed, each with a progressively longer fragment of the NTR of the ␣6 subunit replaced by the corresponding region of the ␣3 subunit.
Replacing the first 140 amino acids of the ␣6 subunit with the ␣3 subunit (␣3 1-140 ␣6/␣3) did not affect the IC 50 (Table 3). However, replacing the first 160 residues (␣3 1-160 ␣6/␣3) decreased the potency of the toxin by ϳ50-fold (Table 3), suggesting interaction of the toxin with amino acids located between residues 140 and 160 of the ␣6 NTR. Replacing an additional 10 amino acids of the ␣6 subunit with the ␣3  (Fig. 2). To determine which residues of the ␣3 subunit should be substituted into the ␣6 subunit, the NTR of both ␣3 and ␣6 subunits were aligned with that of the ␣4 subunit, which does not show affinity for the ␣-CTX MII analog (Fig. 2). Of the five residues that differ between the ␣6 and ␣3 subunits in the first 140 -160 amino acids, three are shared by both ␣3 and ␣4 subunits (Met-141, Lys-152, and Val-156) and one is different in all three (Leu-157 in ␣3, which is a Ser in ␣4 and an Ile in ␣6). When each of these ␣3 residues was systematically substituted for the corresponding residue in ␣6, the ␣6Glu-152 substitution with Lys created the most substantial rightward shift in the dose-response curve, decreasing the potency of the ␣-CTX MII[S4A,E11A,L15A] for the ␣6/␣3␤2␤3 mutant receptor by 8-fold relative to that of the WT receptor (Table 4).
The next stretch of amino acids that affected potency of ␣-CTX MII[S4A,E11A,L15A] was located in the region downstream of the first 170 amino acids (Table 3). Once again, residues that differed between the ␣6 and the ␣3 subunits, but were common to the ␣3 and the ␣4 subunits, were exchanged. There are 8 amino acids that differ between the ␣6 and ␣3 subunits in the region between residues 170 and 204 (Fig. 2). Although ␣6Val-176 (Ile in ␣3), ␣6Asp-177 (Lys in ␣3) and ␣6Phe-201 (Leu in ␣3) differ in the two subunits, they are shared by the ␣6 and ␣4 subunits and therefore were not examined. Of the eight different residues, only one is shared by ␣3 and ␣4 subunits (Gly-171). However, substituting this residue for Ser found in the ␣6 NTR did not markedly affect the susceptibility of the receptor to toxin. As for the remaining residues, only substitutions of ␣6Asp-184 to Glu and ␣6Thr-195 to Gln decreased the receptor's susceptibility to ␣-CTX MII[S4A,E11A,L15A], by 8-fold and 9-fold, respectively (Table 4). When the three resi-   (Table 4 and Fig. 3A).
To ascertain whether the decreased toxin susceptibility of the ␣6E152KD184ET195Q/␣3 mutant involved changes in ACh affinity, ACh dose-response curves were obtained for the WT and mutant receptors. ACh activated the WT and mutant receptors with similar potencies, with an EC 50 of 7.7 M (95% CI: 5-12 M) and 14 M (95% CI: 12-16 M) for the WT and mutant receptor, respectively (Fig. 4).
Positive Charge at Position 152 Disfavors Interaction of the ␣3 Subunit with ␣-CTX MII[S4A,E11A,L15A]-Among the residues that affected the potency of ␣-CTX MII [S4A,E11A,L15A], there is a charge reversal, from Glu-152 in ␣6 to a Lys-152 in ␣3; the residue exchange caused an ϳ8-fold reduction in the potency of the peptide for the ␣6/␣3 subunit. To determine whether this change in potency was due to a charge reversal or to a change in the side-chain length, ␣6Glu-152 was systematically replaced with an Arg, Gln, or Met. The Arg mutation retains the charge reversal, but has a longer side chain, whereas FIGURE 3. Differential susceptibility of ␣6/␣3␤2␤3 versus ␣3␤2 nAChRs to ␣-CTX MII[S4A,E11A,L15A] is due largely to three residues. Oocytes expressing one of four different nAChRs were voltage clamped and subjected to ACh pulses as described in Fig. 1 and "Experimental Procedures." A, combined mutations of the three identified residues in the extracellular binding region of the ␣6 subunit decreased sensitivity to ␣-CTX MII[S4A,E11A,L15A] by Ͼ1000-fold. B, substitution of the three ␣3 residues with the homologous ␣6 residues yielded a receptor that was 150-fold more sensitive to the ␣-CTX MII analog compared with WT ␣3␤2 nAChR. Values are mean Ϯ S.E.M. from three to five separate oocytes.  Gln replacement retains the side-chain length of Glu, but eliminates the negative charge. Met has a straight side chain, similar to Lys, but is not charged. As with the ␣6E152K/␣3 mutant, the ␣6E152R/␣3 mutant had a lower sensitivity to ␣-CTX MII[S4A,E11A,L15A] (IC 50 : 23 nM (95% CI: 19 -29 nM)) ( Fig. 6). Both ␣6E152Q/␣3 and ␣6E152M/␣3 mutations only slightly changed the IC 50 of the analog relative to the WT receptor, with IC 50 values of 5.9 nM (95% CI: 5.0 -7.0 nM) and 4.3 nM (95% CI: 3.5-5.5 nM), respectively (Fig. 6).
Amino acid residues that determine the high potency of ␣-CTX MII for the ␣3 subunit were previously characterized and are Lys-185 and Ile-188 (17). However, both of these residues are conserved between the ␣3 and the ␣6 subunits, which helps explain the similar high affinity of ␣-CTX MII for ␣6/␣3␤2␤3 and ␣3␤2 nAChRs. In this study, receptor mutagenesis was used to assess the residues in the extracellular region of the ␣6 versus the ␣3 subunit that confer selectivity of the new peptide ␣-CTX MII[S4A,E11A,L15A] for the ␣6 subunit. These amino acids include Glu-152, Asp-184, and Thr-195. Receptors with ␣6 subunits in which all three residues . D-F, the reciprocal replacement of ␣3 residues with the corresponding ␣6 residues led to an increase in sensitivity to block by the MII analogs, but to a lesser extent to ␣-CTX PIA block. Values are mean Ϯ S.E.M. from three to five separate oocytes. were simultaneously exchanged with analogous residues of the ␣3 subunit had a Ͼ1000-fold higher IC 50 for ␣-CTX MII[S4A,E11A,L15A]. Conversely, replacing residues in ␣3 with the three key ␣6 residues increased sensitivity for ␣-CTX MII[S4A,E11A,L15A] by ϳ150-fold. Recent crystallographic structures of ACh binding proteins (AChBP) from Lymnea stagnalis, Aplysia californica, and Bulinus truncatus, which are used to model the nAChR in its ligandbound state, as well as the high resolution structure of the Torpedo nAChR in the closed state, have aided greatly in understanding the ligand binding domain of nAChRs, and the structural changes that follow agonist and/or antagonist binding (20 -23). The structures consist of an N-terminal ␣-helix, two-short 3 10 helices, and a core of 10 ␤-sheets, ␤1-␤10. An "aromatic cage" that forms the ACh binding pocket consists of loops A-F, with loops A-C belonging to the principal part (or the ␣ subunit in case of muscle and neuronal nAChRs) and loops D-F belonging to the complementary part (the ␤ subunit, in neuronal heteromeric nAChRs) (20,23). It has been proposed that upon agonist (e.g. ACh) binding, loop C moves in a counterclockwise motion, closing on the agonist like a "hinge," thus capping and trapping the agonist (23)(24)(25)(26). However, binding of an antagonist would lock the C-loop in an open (i.e. resting) conformation and prevent opening of the channel. The change in the C-loop conformation has been confirmed in high resolution crystal structures of ␣-CTX PnIA[A10L,D14K] and ␣-CTX ImI bound to AChBP from A. californica (21,27). Lys-185 and Ile-188, the two residues in the ␣3 subunit that interact with ␣-CTX MII (17), both reside in the ␤9 strand that precedes the C-loop, as determined from sequence comparison with both the Lymnea and Aplysia AChBPs (22,27). Similarly, two of the residues that confer high selectivity of ␣-CTX MII[S4A,E11A,L15A] for the ␣6 subunit, Asp-184 and Thr-195, are in the ␤9 and ␤10 strands, respectively, the former precedes and the latter follows the C-loop (Fig. 7). Therefore, subtle differences in the geometry of the C-loop may affect toxin binding. ␣3Gln-195 has previously been shown to interact with ␣-CTX PnIA and -bungarotoxin, which block ␣3␤2 nAChRs (28,29). In addition, an Ile at this position in the Aplysia AChBP is involved in interaction with ␣-CTX ImI (27). The present study, however, indicates that ␣-CTX  MII[S4A,E11A,L15A] prefers a Thr at this position, as found in the ␣6 ligand binding domain. Gln and Thr both have polar, but uncharged side chains. However, the side chain of Gln is longer than that of Thr. Therefore, it is possible that the slightly longer side chain of Gln-195 sterically hinders the binding of ␣-CTX MII[S4A,E11A,L15A] to the ␣3␤2 nAChR. Similarly, Glu at position 184 of the ␣3 subunit may sterically hinder binding of the ␣-CTX MII analog, because the shorter Asp at this position, as found in the ␣6 subunit, favors the binding of the toxin.
The other residue in the ␣6 subunit that was found to influence the binding of ␣-CTX MII[S4A,E11A,L15A], Glu-152, although not directly in the ␤9 -␤10 sheets, lies in close proximity (Fig. 7). Additionally, this residue, which is in a homologous position to that of Glu-151 in the A. californica AChBP, has been shown to be critical for binding of the PnIA variant ␣-CTX PnIA[A10L,D14K] (21). Both the ␣3 and the ␣4 subunits, which are not sensitive to ␣-CTX MII[S4A,E11A,L15A], have a Lys at the homologous position. When the negatively charged ␣6Glu-152 is replaced with positively charged Lys or Arg, the resulting mutant is ϳ8 and 19-fold, respectively, less sensitive to ␣-CTX MII[S4A,E11A,L15A], suggesting that either the negative charge might be important for conferring high selectivity to the toxin, or that introduction of a positive charge might be unfavorable to binding of the toxin. To address the former possibility, ␣6Glu-152 was replaced with the isosteric, but uncharged Gln. This substitution caused an ϳ5-fold decrease in sensitivity to ␣-CTX MII[S4A,E11A,L15A], thus suggesting that the negative charge may have a role in conferring the high affinity of the toxin to the ␣6 subunit. To address the possibility that the introduction of a positive charge may be unfavorable, ␣6Glu-152 was replaced with Met, which is structurally similar to Lys but lacks the positive charge. This substitution caused a 3.5-fold decrease in the potency of the analog, as opposed to 8-and 19-fold decrease when either a Lys or Arg is present (see above), thus suggesting that the positive charge may contribute to the unfavorable interaction with the ␣-CTX MII[S4A,E11A,L15A].
In conclusion, we have created a high affinity MII analog that, unlike the parent peptide, is highly selective for ␣6/␣3␤2␤3 over ␣3␤2 nAChRs. Through receptor mutagenesis we have identified amino acid residues in the ␣6 subunit that are responsible for the differences in toxin affinity between these receptors. Considering the location of these residues with respect to residues that interact with ACh (Fig. 7B), it is likely that ␣-CTX MII[S4A,E11A,L15A] acts competitively to antagonize ACh binding. These residues are distinct from those previously determined as important to ␣-CTX MII interaction with the ␣3 subunit (17,29) and provide further information for distinguishing among these structurally related nAChRs that have distinct roles in the peripheral and central nervous systems.