nAChRs activated by the endogenous neurotransmitter acetylcholine belong to the superfamily of ligand-gated ion channels that also includes GABA_A, 5-hydroxytryptamine-3, and glycine receptors (Changeux, 1993). These different ligand-gated ion channels show considerable sequence and structural homology. Each of the subunits has a relatively hydrophilic amino terminal half (∼200 amino acids) that constitutes an extracellular domain. This is followed by three hydrophobic transmembrane domains, a large intracellular loop, and then a fourth hydrophobic transmembrane span.

A large number of genes have been cloned that encode subunits of nAChRs. It has been proposed that these subunits may be divided into subfamilies on the basis of both gene structure and mature protein sequence. The subunits α2, α3, α4, and α6 belong to subfamily III, tribe 1; β2 and β4 belong to tribe III-2; and the putative structural subunits α5 and β3 belong to tribe III-3 (Corringer et al., 2000). Within tribe III-1, subunits α3 and α6 show considerable sequence identity (∼80% in the ligand-binding extracellular domain). Thus, designing ligands to distinguish between α3^*1 and α6^* is particularly challenging.

α-Conotoxin MII is a 16 amino acid peptide originally isolated from the venom of the marine snail Conus magus. This peptide potently targets neuronal in preference to the muscle subtype of nicotinic receptor with high affinity for both α3β2 and α6^* nAChRs. Unfortunately, α-conotoxin MII may not distinguish well between α3^* and α6^* nAChRs (Kuryatov et al., 2000). In an effort to remedy this situation and produce a selective ligand for α6^* nAChRs, we have generated a series of α-conotoxin MII analogs.

The α6 subunit is expressed in catecholaminergic neurons and in retina (Le Novère et al., 1996, 1999; Vailati et al., 1999). In striatum, α6^* nAChRs seem to play a central role in the modulation of dopamine release. Recently, homozygous null mutant (α6-/-) mice were generated. Receptor autoradiography studies in these animals indicate that the α6 nAChR subunit is a critical component of [¹²⁵I]α-conotoxin MII binding in the central nervous system (Champtiaux et al., 2002). Studies using mice with nAChR subunit deletion indicate that α3 does not participate in most [¹²⁵I]α-conotoxin MII binding sites but does influence expression in the habenulo-peduncular tract (Whiteaker et al., 2002). Thus, α6-selective ligands would be useful to distinguish the α6^* majority form from the α3^* minority of such sites.

Materials and Methods

Chemical Synthesis. Peptides were synthesized on a Rink amide resin, 0.45 mmol/g [Fmoc-Cys(Trityl)-Wang; Novabiochem, San Diego, CA] using N-(9-fluorenyl)methoxycarboxyl chemistry and standard side chain protection except on cysteine residues. Cysteine residues were protected in pairs with either S-trityl on the first and third cysteines or S-acetamidomethyl on the second and fourth cysteines. Amino acid derivatives were from Advanced Chemtech (Louisville, KY). The peptides were removed from the resin and precipitated, and a two-step oxidation protocol was used to selectively fold the peptides as described previously (Luo et al., 1999). Briefly, the first disulfide bridge was closed by dripping the peptide into an equal volume of 20 mM potassium ferricyanide and 0.1 M Tris, pH 7.5. The solution was allowed to react for 30 min, and the monocyclic peptide was purified by reverse-phase HPLC. Simultaneous removal of the S-acetamidomethyl groups and closure of the second disulfide bridge was carried out by iodine oxidation. The monocyclic peptide and HPLC eluent was dripped into an equal volume of iodine (10 mM) in H₂O/trifluoroacetic acid/acetonitrile (78:2:20 by volume) and allowed to react for 10 min. The reaction was terminated by the addition of ascorbic acid diluted 20-fold with 0.1% trifluoroacetic acid and the bicyclic product purified by HPLC.

Mass Spectrometry. Measurements were performed at the Salk Institute for Biological Studies (San Diego, CA) under the direction of Jean Rivier. Matrix-assisted laser desorption ionization time-of-flight mass spectrometry and liquid secondary ionization mass spectrometry were used.

Preparation of nAChR Subunit cRNA. Attempts to express the rat nAChR α6 subtype in Xenopus laevis oocytes consistently failed; that is, no ACh-gated currents were detected. To improve functional expression, we created a chimeric receptor of the rat α6 and α3 subtypes. The chimeric receptor consists of amino acids 1 to 237 of the rat α6 subunit protein linked to amino acids 233 to 499 of the rat α3 subunit protein. The chimeric junction is located at the paired-RR- residues immediately preceding the M1 transmembrane segment of the α3 subunit. The resulting chimeric receptor represents the extracellular ligand-binding domain of the α6 subunit linked to membrane-spanning and intracellular segments of the α3 subunit. The α6/α3 cDNA was constructed by the introduction of BspEI sites at the chimeric junction into the α6 and α3 cDNA sequences using mutagenic primers to introduce the restriction sites through silent codon changes. The α6 and α3 segments were generated by polymerase chain reaction of rat brain cDNA using primers in the 5′ and 3′ untranslated regions of the corresponding cDNAs along with the internal mutagenic primers. The polymerase chain reaction products were digested with BspEI and ligated to generate the chimeric construct. The final chimeric construct was cloned and completely sequenced to confirm the correct cDNA sequence. To further improve expression levels, all of the 5′ and 3′ untranslated regions of the nAChR cDNA were deleted, and the chimeric construct was cloned into the X. laevis expression vector pT7TS, placing X. laevis globin 5′ and 3′ untranslated regions around the nAChR cDNA. The expression construct pT7TS/rα6α3 was transcribed with T7 RNA polymerase to generate sense-strand RNA for oocyte expression.

Electrophysiology and Data Analysis. Clones of rat nAChR subunits were used to produce cRNA for injection into X. laevis oocytes as described previously (Cartier et al., 1996). The rat α6 and β3 subunits were a generous gift from S. Heinemann (Salk Institute, San Diego, CA) (Deneris et al., 1989). To express nAChRs in oocytes, 5 ng of each nAChR subunit was injected. In the case of α6β4, 50 ng of each subunit was injected because of absent expression when using 5 ng of cRNA. Likewise, 20 ng was used for the α6/α3β2 combination that expresses poorly without the β3 subunit. A 30-μl cylindrical oocyte recording chamber fabricated from Sylgard was gravity-perfused with ND96A (96.0 mM NaCl, 2.0 mM KCl, 1.8 mM CaCl₂, 1.0 mM MgCl₂,1 μM atropine, and 5 mM HEPES, pH 7.1-7.5) at a rate of ∼2 ml/min (Luo et al., 1998). All toxin solutions also contained 0.1 mg/ml bovine serum albumin to reduce nonspecific adsorption of peptide. Toxin was preapplied for 5 min. ACh-gated currents were obtained with a 2-electrode voltage-clamp amplifier (model OC-725B; Warner Instrument, Hamden, CT), and data were captured as described previously (Luo et al., 1998). The membrane potential of the oocytes was clamped at -70 mV. To apply a pulse of ACh to the oocytes, the perfusion fluid was switched to one containing ACh for 1 s. This was done automatically at intervals of 1 to 5 min. The shortest time interval was chosen such that reproducible control responses were obtained with no observable desensitization. The concentration of ACh was 10 μM for trials with α1β1δϵ and 100 μM for all other nAChRs. Toxin was bath-applied for 5 min, followed by a pulse of ACh. Thereafter, toxin was washed away, and subsequent ACh pulses were given every 1 min, unless otherwise indicated. All ACh pulses contain no toxin, for it was assumed that little if any bound toxin washed away in the brief time (less than the 2 s it takes for the responses to peak). In our recording chamber, the bolus of ACh does not project directly at the oocyte but rather enters tangentially, swirls, and mixes with the bath solution. The volume of entering ACh is such that the toxin concentration remains at a level >50% of that originally in the bath until the ACh response has peaked (<2 s). When longer than 5 min of toxin application was needed to reach maximum block, toxin was applied by continuous perfusion to the oocytes as described previously (Luo et al., 1994), except that ACh was applied once every 2 min.

The average peak amplitude of three control responses just preceding exposure to toxin was used to normalize the amplitude of each test response to obtain a “% response” or “% block”. Each data point of a dose-response curve represents the average value ± S.E. of measurements from at least three oocytes. Dose-response curves were fit to the equation %response = 100/{1 + ([toxin]/IC₅₀)^nH}, where n_H is the Hill slope determined with Prism software (Graph-Pad Software, San Diego, CA) on an Macintosh (Apple Computers, Cupertino, CA). For three or fewer data points, n_H was set to 1.0.

Membrane Preparation. Mice were killed by cervical dislocation. Brains were removed from the skulls and dissected on an ice-cold platform. Membranes containing [¹²⁵I]α-conotoxin MII binding sites were prepared from pooled olfactory tubercles, striatum, and superior colliculus. Samples were homogenized in 2× physiological buffer (288 mM NaCl; 3 mM KCl; 4 mM CaCl₂; 2 mM MgSO₄; and 40 mM HEPES, pH 7.5; 22°C) using a glass-polytetrafluoroethylene tissue grinder. Homogenates were then treated with phenylmethylsulfonyl fluoride (final concentration, 1 mM; 15 min at 22°C) to inactivate endogenous serine proteases before centrifugation (20,000g for 20 min at 4°C). Pellets were washed twice by homogenization in distilled deionized water glass-polytetrafluoroethylene tissue grinder, 4°C) and centrifugation (20,000g for 20 min at 4°C). Pooled tissue from a single mouse provided sufficient material for a single 96-well format assay.

Inhibition of [¹²⁵I]α-Conotoxin MII Binding. Inhibition of [¹²⁵I]α-conotoxin MII binding to mouse brain membranes was performed using a modified version of the 96-well plate procedure described previously (Whiteaker et al., 2000a). Assays were performed in triplicate using 1.2-ml siliconized polypropylene tubes arranged in a 96-well format. Membrane pellets were resuspended into distilled deionized water. Total (no drug) and nonspecific (with 1 μM epibatidine) binding determinations were included in each experiment for each drug dilution series. Initial incubations proceeded for 3 h at 22°C in 1× protease inhibitor buffer [1× physiological buffer supplemented with bovine serum albumin (0.1% w/v), 5 mM EDTA, 5 mM EGTA, and 10 μg/ml each of aprotinin, leupeptin trifluoroacetate, and pepstatin A]. Each tube contained 10 μl of membrane preparation, 10 μl of competing ligand (or nonspecific or total determinations) in 1× protease inhibitor buffer, and 10 μl of [¹²⁵I]α-conotoxin MII (1.5 nM in 2× protease inhibitor buffer, giving a final assay radioligand concentration of 0.5 nM). After incubation, each tube was diluted with 1 ml of physiological buffer plus 0.1% (w/v) bovine serum albumin. Tubes were then incubated for a further 4 min at 22°C to reduce nonspecific binding to the membrane preparation. The binding reactions were then terminated by filtration onto a single thickness of GF/F filter paper (Whatman, Clifton, NJ) using a cell harvester (Inotech Biosystems, Rockville, MD). The filters were incubated previously for 15 min with 5% dried skim milk to reduce nonspecific binding. Assays were washed with four changes of physiological buffer supplemented with bovine serum albumin (0.1% w/v). Washes were performed at 30-s intervals, with each lasting approximately 5 s. All filtration and collection steps were performed at 4°C. Bound ligand was quantified for each filter disc by gamma counting using a Cobra II counter (≈85% efficiency) (PerkinElmer Life and Analytical Sciences, Boston, MA).

Calculations. Data from individual [¹²⁵I]α-conotoxin MII inhibition binding experiments were processed using a single-site fit using the nonlinear least-squares fitting algorithm of GraphPad Prism. Values of K_i were derived for each experiment by the method described by Cheng and Prusoff (1973), K_i = IC₅₀/1 +(L/K_D), where K_i for [¹²⁵I]α-conotoxin is 0.32 nM.

Results

Peptide Synthesis. The sequence of native α-conotoxin MII is GCCSNPVCHLEHSNLC. Peptide analogs were synthesized by substituting one or more residues with alanine. These peptides are named according to the residue(s) substituted; for example, MII[E11A] has the glutamic acid in position 11 substituted with alanine. Cysteine residues were orthogonally protected to direct the formation of disulfide bonds in the configuration found in α-conotoxin MII, that is cysteine 1 to cysteine 3 and cysteine 2 to cysteine 4. The first and third cysteine residues were protected with acid-labile groups that were removed first after a cleavage from the resin; ferricyanide was used to close the first disulfide bridge. The monocyclic peptides were purified by reverse-phase HPLC. Then the acid-stable acetamidomethyl groups were removed from the second and fourth cysteines by iodine oxidation that also closed the second disulfide bridge. The fully folded peptides were again purified by HPLC. Mass spectrometry was used to confirm synthesis. The observed molecular mass for each peptide was within 0.1 Da of the expected mass.

Peptide Effects on α6^* and α3^* nAChRs. Injection of rat α6 subunits into oocytes either alone or in combination with β2 and/or β3 subunits yields few or no functional nAChRs. Using a previously reported strategy for human α6 (Kuryatov et al., 2000), we joined the extracellular domain of the rat α6 subunit to the transmembrane and intracellular portion of the closely related rat α3 subunit. Alanine analogs were then tested against α3β2 or α6/α3β2β3 subunit combinations heterologously expressed in oocytes. The β3 subunit was used with α6/α3β2, for without it there was generally little or no functional expression. In addition, β3 is associated with native α6β2^*-containing nAChRs (Zoli et al., 2002; Cui et al., 2003). Results are shown in Fig. 1 and Table 1. Substitution of alanine for Asn5, Pro6, or His12 resulted in substantially decreased activity compared with native MII at both α6^* and α3^* nAChRs, whereas substitution for Val7 had the most pronounced effect on the α6/α3β2β3 nAChR. Substitution for Ser4, His9, Leu10, Glu11, Ser13, Asn14, and Leu15 had only modest effects on α6/α3β2β3; however, mutations at Ser4, His9, Glu11, Asn14, and Leu15 resulted in substantially lower activity on α3β2 nAChRs. Thus, these mutations are analogs that preferentially block α6/α3β2β3 versus α3β2 nAChRs. We note that for certain analogs, including S4A, L10A, E11A, S13A, and N14A, the t_1/2 for recovery from toxin block of α6/α3β2β3 nAChRs was long (>25 min). For these analogs, 10- to 15-min toxin incubations were used to achieve maximum block at 10 nM concentration, and 20- to 35-min incubations were used to achieve maximum block at 1 nM concentration. A slow off-rate but similar affinity to other analogs implies that the analogs with a slow off-rate also have slower on-rates and thus the need for longer application times. MII[H9A] and MII[L15A] failed to block α4β2 nAChRs; at 10 μM peptide concentration, the ACh-evoked current was 105.8 ± 2.4 and 102.3 ± 5.3% of control, respectively (data from six oocytes).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

H9A and L15A analogs of α-MII discriminate between α6/α3β2β3 and α3β2 nAChRs. Rat nAChR subunits were heterologously expressed in X. laevis oocytes. Concentration-response analysis of the peptide block of ACh-induced current was performed as described under Materials and Methods. α-Conotoxin MII blocked α3β2 and α6/α3β2β3 with IC₅₀ values of 2.18 and 0.39 nM, respectively. See also Table 1 for confidence intervals. The Hill slopes were 0.75 ± 0.13 and 0.53 ± 0.04, respectively. MII[H9A] blocked α3β2 and α6/α3β2β3 nAChRs with IC₅₀ values of 59.0 and 0.79 nM, respectively, and with Hill slopes of 0.83 ± 0.08 and 0.73 ± 0.08. MII[L15A] blocked α3β2 and α6/α3β2β3 nAChRs with IC₅₀ values of 34 and 0.92 nM, respectively, and with Hill slopes of 0.58 ± 0.08 and 0.75 ± 0.08, respectively. The data are from three to six separate oocytes; ± value is the standard error of the mean.

Selectivity of MII[E11A]. The single alanine substitution MII[E11A] has ∼50-fold preference for α6/α3β2β3 versus α3β2 nAChRs and seems to be the most potent analog on α6/α3β2β3 nAChRs. We therefore tested its effects on additional nAChR subtypes. The apparent on-rate for α6/α3β4 nAChRs is slow; at concentrations of toxin ≤ 10 nM, 60 to 70 min of toxin application was required to reach a steady-state level of nAChR block. Concentration-response curves are shown in Fig. 2 and IC₅₀ values are shown in Table 2.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Concentration-response analysis of α-conotoxin MII[E11A] on nAChR subtypes expressed in X. laevis oocytes. Peptide was perfusion-applied at concentrations ≤100 nM and bath-applied at higher concentrations as described under Materials and Methods. A, block by MII[E11A] of β2-containing nAChRs. B, block by MII[E11A] of β4-containing and α7 nAChRs. Data are from three to five oocytes. Error bars are S.E.M. Results are summarized in Table 2. Note the strong preference for α6/α3^* nAChRs.

Double Mutants. A series of double alanine-substituted mutations was also constructed. These mutations were tested with respect to their activity at α6/α3β2β3 and α3β2 nAChRs. As seen in Table 3, each of these double mutants preferentially blocks the α6/α3β2β3 receptor versus the α3β2 receptor. The IC₅₀ of the MII[H9A;L15A] analog was approximately 2000-fold lower for α6/α3β2β3 versus α3β2, and this analog was selected for further characterization (Fig. 3).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

The [H9A,L15A] analog of α-MII discriminates between α6^* and α3^* nAChRs. Peptide was applied to oocytes expressing the indicated nAChR subunit combinations as described under Materials and Methods. A, the peptide blocked rat α3β2 with an IC₅₀ of 4.8 μM (CI = 3.5-6.6 μM) and n_H of 0.48 ± 0.04. The peptide blocked rat α6/α3β2β3 with an IC₅₀ of 2.4 nM (CI = 1.7-3.4 nM) and n_H of 0.72 ± 0.09. B, the peptide blocked rat α3β4 with an IC₅₀ of 7.8 μM (CI = 5.3-11.5 μM) and n_H of 0.75 ± 0.1. The peptide blocked rat α6β4 with an IC₅₀ of 269 nM (CI = 153-476 nM) and n_H of 0.60 ± 0.09; ± values are standard error of the mean.

Kinetics of Block by MII[H9A;L15A]. α-Conotoxin MII is slowly reversible on α3β2 nAChRs and very slowly reversible on α6/α3β2β3 nAChRs (Fig. 4). Substitution of Ala for His9 or Leu15 leads to more rapid recovery from block for both receptor subtypes. In the case of the double mutant MII[H9A;L15A] recovery from toxin block is rapid. The magnitude of the change of recovery rate is greater than the magnitude of change in IC₅₀ at the α6/α3β2β3 receptor. This implies that changes in the peptide that lead to a rapid off-rate also lead to a faster on-rate of binding. We note that α-conotoxin GIC, a more rapidly reversible homolog of α-conotoxin MII, also has an alanine rather than the histidine found in position 9 of α-MII (McIntosh et al., 2002).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Kinetics of block. MII, MII[H9A], MII[L15A], and MII[H9A;L15A] were applied to X. laevis oocytes heterologously expressing rat α6/α3β2β3 and α3β2 nAChRs. Peptide at the indicated concentrations was bath-applied for 5 min and then washed out. Kinetics of unblock were monitored by applying a 1-s pulse of ACh every 1 min.

Activity of MII[H9A;L15A] on Other nAChR Subtypes. MII[H9A;L15A] has highest affinity for the α6/α3β2β3 subunit combination and ∼100-fold less activity on the α6/α3β4 combination (Fig. 3). MII[H9A;L15A] has low or no activity on the remaining neuronal subunit combinations tested, including α2β2, α2β4, α3β4, α4β2, α4β4, and α7 (Fig. 5 and Table 4). Thus, MII[H9A;L15A] selectively blocks α6^* nAChRs, with preference for the α6/α3β2β3 versus α6/α3β4 subunit combination.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

Effects of MII[H9A,L15A] on additional nAChR subtypes. Peptide at 100 nM (α6/α3β2β3) and 1 μM (all other subtypes) was bath-applied for 5 min to X. laevis oocytes expressing the indicated rat nAChR subunits. Traces are representative of experiments on three to five oocytes. C, control response to ACh. Second response in each trace pair is the response to ACh in the presence of peptide.

Effect of β3 Subunit. Occasional expression of α6/α3β2 was seen without coinjection of the β3 subunit. MII[H9A; L15A] blocked α6/α3β2 nAChRs with an IC₅₀ of 8.21 (6.36-10.6) nM compared with 2.4 nM (1.68-3.43) for α6/α3β2β3. As indicated above (Fig. 2 and Table 2), MII[E11A] blocked α6/α3β2 nAChRs with an IC₅₀ of 0.154 nM (0.134-0.178) compared with 0.16 nM (0.135-0.189) on α6/α3β2β3 nAChRs. Numbers in parentheses are 95% confidence intervals.

Activity of Analogs at Native Mouse Brain nAChRs. A concentration-response analysis was performed on four of the analogs with α6/α3^* versus α3^* selectivity—MII[H9A], MII[E11A], MII[L15A], and MII[H9A;L15A]—using inhibition of [¹²⁵I]α-conotoxin MII binding (Whiteaker et al., 2000b) to mouse brain homogenates. Results are shown in Fig. 6. The values obtained for these analogs correlate well with values obtained on α6/α3β2β3 rather than α3β2 nAChRs as expressed in X. laevis oocytes (Table 5).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6.

Concentration-response analysis of α-conotoxin MII analogs on native nAChRs. Analogs were assessed for their ability to displace [¹²⁵I]α-conotoxin MII binding on mouse brain homogenates as described under Materials and Methods. Nonspecific binding was defined with 1 μM epibatidine. K_i values are shown in Table 5. The Hill slope was 0.95 ± 0.13, 0.89 ± 0.14, and 1.1 ± 0.14 for MII[H9A], MII[L15A], and MII[H9A; L15A], respectively. ± values are standard error of the mean. Data are from three to seven experiments.

Discussion

Although the sequence of the coding region for the α6 gene has been known for many years (Lamar et al., 1990), its functional significance has been challenging to elucidate because of difficulties in heterologously expressing α6 and because of a lack of subtype-specific ligands. Indeed, originally it was not entirely certain that the α6 gene encoded a nicotinic receptor subunit. The α6 subunit has relatively discrete localization, with expression in catacholaminergic nuclei including the locus coeruleus, the ventral tegmental area, and the substantia nigra (Le Novère et al., 1996; Göldner et al., 1997; Han et al., 2000; Quik et al., 2000; Azam et al., 2002). It is also found in trigeminal ganglion and olfactory bulb (Keiger and Walker, 2000). In addition, α6 complexes have been reported in chick retina (Vailati et al., 1999). The α6 mRNA expression pattern overlaps extensively with that of the α3 subunit, leading to initial confusion over the composition of [¹²⁵I]α-conotoxin MII-binding nAChRs (Whiteaker et al., 2000b).

Subunit-specific antibodies have been used to immunoprecipitate α6^* receptors from chick retina. When reconstituted in lipid bilayers, these receptors formed cationic channels characteristic of nAChRs, thus establishing a functional role for native α6^* nAChRs (Vailati et al., 1999). Antibodies have also been used recently to demonstrate the presence of α6β2^* nAChRs in striatal dopaminergic terminals in rat. β3 and/or α4 subunits are also present in a proportion of these nAChRs (Zoli et al., 2002). Subunit knockout mice suggest that the high-affinity binding site of [¹²⁵I]α-conotoxin MII is predominately composed of α6^* rather than α3^* nAChRs (Champtiaux et al., 2002; Whiteaker et al., 2002). It has been hypothesized recently that putative α6^* nAChRs in the striatum may participate in the pathophysiology of Parkinson's disease, a neurodegenerative disorder characterized by progressive loss of dopamine neurons. Treatment of primates with 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (a dopaminergic neurotoxin) leads to selective decline of putative α6β2^* nAChRs (Quik et al., 2001; Kulak et al., 2002). Thus, there is a significant need for ligands that selectively act at α6^* nAChRs.

We demonstrate in this report that certain analogs of α-conotoxin MII exhibit preferential loss of activity at α3β2 versus α6/α3β2β3 nAChRs. Additionally, at concentrations tested, the MII[H9A;L15A] analog has little or no activity at α2^*, α4^*, or α7^* nAChRs. Indeed, MII[H9A;L15A] is the most selective α6 ligand thus far reported.

A number of native α-conotoxins have been characterized that target various subtypes of nAChRs. Despite their differences in primary sequence, NMR and X-ray crystallography studies show a high conservation of the structural peptide backbone (Hu et al., 1996, 1997, 1998; Shon et al., 1997; Hill et al., 1998; Cho et al., 2000; Park et al., 2001; Nicke et al., 2003). It seems that this backbone serves as a scaffold that presents a variety of amino acid side chains leading to differences in specificity. In the present study, we have systematically replaced the noncysteine residues of α-conotoxin MII with alanine. The predominant effect is to preferentially decrease activity at the α3β2 receptor relative to the α6/α3β2(±β3) subunit combination.

In an attempt to further increase selectivity, double mutations were constructed from the more selective single mutation analogs. Each of these double mutants retains a low nanomolar IC₅₀ for the α6/α3β2β3 nAChR (Table 3). It is particularly noteworthy that the native MII peptide potently blocks both α6/α3β2β3 and α3β2 nAChRs, whereas the MII[H9A;L15A] analog discriminates between these nAChRs by three orders of magnitude. This discrimination is caused by differences in the extracellular region of the α subunit because the transmembrane and intracellular portions of the chimeric α6/α3 and α3 subunits are identical. Also, the addition of the β3 subunit to α6/α3β2 nAChRs has only a 3.4-fold effect on MII[H9A,L15A] block. MII[E11A] also preferentially blocks α6/α3β2 versus α3β2 nAChRs, again implicating the extracellular portion of the α6 subunit. Furthermore, coexpression of the β3 subunit with the α6/α3 and β2 subunits had no effect on the IC₅₀ of MII[E11A]. However, the presence of a β2 versus β4 subunit does seem to influence peptide affinity. MII[E11A] preferentially blocks α6/α3β2 versus α6/α3β4 nAChRs and preferentially blocks α3β2 versus α3β4 nAChRs.

We used cloned rat receptor subunits heterologously expressed in X. laevis oocytes to examine the differences between α3^* and α6^* nAChRs. Although difficult, occasional expression of α6 with either β2 or β4 subunits has been described. This expression is enhanced with the addition of the β3 subunit (Kuryatov et al., 2000). Improved efficiency of expression has been achieved by combining the extracellular (putative ligand binding) domain of α6 with the remaining portion of either the α3 or α4 subunit (Kuryatov et al., 2000). We have exploited this technique to screen analogs of α-conotoxin MII. It is possible that there are important differences between this chimeric receptor expressed in oocytes and native nAChRs. To assess this, the α-conotoxin MII analogs were also tested in a radioligand binding assay using native nAChR populations. As can be seen in Table 5, the analogs that potently block the rat α6/α3β2β3 nAChR heterologously expressed in oocytes also potently block the native mouse striatal nAChR bound by radiolabeled MII. This native receptor has been shown in previous studies to contain α6 (rather than α3) and β2 subunits (Champtiaux et al., 2002; Whiteaker et al., 2002; Zoli et al., 2002). Thus, the analogs have high affinity for both native and heterologously expressed α6β2^* nAChRs. The H9A;L15A analog of MII also has a relatively high IC₅₀ for other nAChRs, including α2β2, α2β4, α3β2, α3β4, α4β2, α4β4, and α7. Thus, this peptide represents a novel selective probe for discriminating among numerous nAChR subunit combinations.

The precise mechanism by which the [H9A] and [L15A] mutations cause a selective loss of affinity at α3β2 relative to α6/α3β2^* nAChRs is not addressed by these studies. It has been determined that Lys185 and Ile188 of the α3 subunit are critically important for α-conotoxin MII binding to α3β2 nAChRs (Harvey et al., 1997), and these residues are conserved between the α3 and α6 subunits. The most facile explanation of the results presented here is that the crucial interactions between α-conotoxin MII and the α6 subunit may occur at other subunit side chains. Interestingly, both α-conotoxin PnIA and α-conotoxin MII interact with Ile188 but differ in other important interactions with the α3 subunit. α3 subunit Lys185 is not essential to α-conotoxin PnIA binding, whereas Pro182 and Gln198 are (Everhart et al., 2003). Perhaps significantly, the latter two residues are not conserved between the α3 and α6 subunit. Because all of the above residues are found in the putative “C” loop of the α subunit, it seems possible that interaction in this region may be of particular importance. However, several examples indicate that a more complex explanation may be needed. For instance, α-conotoxin PnIA and its derivative α-conotoxin PnIA[A10L] stabilize different states of the same nAChR (Hogg et al., 2003), presumably by interacting with different sets of subunit residues, whereas α-conotoxin MI has been shown to interact in a different orientation with the same α1 subunit residues, depending on whether it is binding at an α/γ or α/δ interface (Sugiyama et al., 1998). These and a series of mutant-cycle analysis studies (Quiram et al., 1999, 2000; Bren and Sine, 2000) have indicated that toxin/channel interactions may be anchored by a small number of relatively strong interactions and supported by a large number of weaker interactions that strongly determine subtype selectivity (Rogers et al., 2000). If this more multifaceted model is correct, the maintenance of affinity between α-conotoxin MII[H9A;L15A] and the α6/α3β2^* nAChR may reflect either a more prominent role of the “supporting” interactions with the native toxin than is seen for α3β2, which is retained after alteration of the His9 and Leu15 side chains. Alternatively, the orientation of the toxin within the binding pocket may shift after substitution at the His9 and Leu15 positions, but the structure of the α6/α3 binding pocket may be better able to accommodate the new positioning than its α3 counterpart. The fact that several of the alanine mutants exhibit affinities similar to each other and native α-conotoxin MII but have radically different binding kinetics reinforces the idea that different interactions may stabilize the nAChR/toxin complex in each case. It seems likely that an accurate understanding of how the [H9A] and [L15A] mutations produce selectivity between α3β2 and α6/α3β2^* nAChRs will require the performance of a comprehensive set of double mutant-cycle analyses.