Structural Elements in α-Conotoxin ImI Essential for Binding to Neuronal α7 Receptors*

The neuronal-specific toxin α-conotoxin ImI (CTx ImI) has the sequence Gly-Cys-Cys-Ser-Asp-Pro-Arg-Cys-Ala-Trp-Arg-Cys-NH2, in which each cysteine forms a disulfide bridge to produce a constrained two-loop structure. To investigate the structural basis for bioactivity we mutated individual residues in CTx ImI and determined bioactivity. Bioactivity of the toxins was determined by their competition against 125I-labeled α-bungarotoxin binding to homomeric receptors containing α7 sequence in the major extracellular domain and 5HT-3 sequence elsewhere. The results reveal two regions in CTx ImI essential for binding to the α7/5HT-3 receptor. The first is the triad Asp-Pro-Arg in the first loop, where conservative mutations of each residue diminish affinity by 2–3 orders of magnitude. The second region is the lone Trp in the second loop, where an aromatic side chain is required. The overall results suggest that within the triad of the first loop, Pro positions the flanking Asp and Arg for optimal interaction with one portion of the binding site, while within the second loop, Trp stabilizes the complex through its aromatic ring.

Specificity of a particular ␣-conotoxin likely reflects structural differences in the various AChR binding sites. Binding sites of nicotinic AChRs are formed at interfaces between pairs of ␣ and non-␣ subunits (reviewed in Ref. 8). In the muscle AChR, the binding sites are formed by ␣ 1 -␦, ␣ 1 -␥, and ␣ 1 -⑀ subunit pairs, whereas in the homomeric ␣ 7 AChR the binding sites are formed by pairs of identical subunits, ␣ 7 -␣ 7 . Thus across the various AChR subtypes, the different binding site interfaces contribute different residues which are recognized by the various ␣-conotoxins.
All ␣-conotoxins contain two disulfide bridges, proline in the first loop, and basic and aromatic residues in the second loop. However, each ␣-conotoxin targets a particular binding site through differences in its number and type of residues. For example, CTx ImI contains four residues in the first loop and three in the second, whereas muscle-specific conotoxins contain three residues in the first loop and five in the second. Furthermore, unlike muscle-specific conotoxins, CTx ImI contains both positively and negatively charged residues in the first loop. Thus structural differences in ␣-conotoxins likely reflect structural differences at the various AChR binding site interfaces.
Competitive antagonists are potential probes of binding site structure that can identify residues of close approach which are distant in the linear sequence or contained in different protein subunits. When the antagonist is structurally constrained it can also serve as a molecular caliper for estimating distances between these residues. For example, previous work showed that the conformationally restricted antagonist dimethyl-d-tubocurarine bridges the ␣ and ␥ subunits in the muscle receptor through interaction between its two quaternary nitrogens and tyrosines in each subunit (9,10). Given the distance between quaternary nitrogens in dimethyl-d-tubocurarine, the two tyrosines in the ␣ and ␥ subunits are estimated to be 11 Å apart. Similarly, ␣-conotoxins are potential probes of the binding sites of muscle and neuronal AChRs. Solution and crystal structures of ␣-conotoxins reveal a triangular structure with positive charges at two vertices separated by 15 Å (11-13). Thus, by identifying active residues in CTx ImI and the ␣ 7 binding site interface, residues of close approach can be identified, and their separation can be estimated.
We recently identified residues of the ␣ 7 binding site that confer neuronal specificity of CTx ImI (23). The present paper continues our work characterizing the ␣ 7 binding site by constructing a series of CTx ImI mutants and measuring binding affinity of the mutant toxins. The results reveal two key regions in CTx ImI essential for bioactivity.
Synthesis and Purification of Conotoxin ImI-Mutant and wild type ␣-conotoxin ImI were synthesized by standard Fmoc (N-(9-fluorenyl) methoxycarbonyl) chemistry on an Applied Biosystems 431A peptide synthesizer. During synthesis, cysteine protecting groups (S-triphenylmethyl) were incorporated at cysteines 3 and 12, and acetamidomethyl protecting groups (ACM) were incorporated at cysteines 2 and 8. The linear peptide was purified by reversed phase high performance liquid chromatography using a Vydac C18 preparative column with trifluoroacetic acid/acetonitrile buffers. The two intramolecular disulfide bridges were formed as follows: the cysteine S-triphenylmethyl protecting groups of cysteines 3 and 12 were removed during trifluoroacetic acid cleavage of the linear peptide from the support resin, and the peptide was oxidized by molecular oxygen to form the 3-12 disulfide by stirring in 50 mM ammonium bicarbonate buffer, pH 8.5, at 25°C for 24 h. The peptide was lyophilized prior to formation of the second disulfide bridge. The ACM protecting groups on cysteine 2 and 8 were removed oxidatively by iodine as described (16) except the peptide/ iodine reaction was allowed to progress 16 h prior to carbon tetrachloride extraction. Residual iodine was separated from the pure product by high performance liquid chromatography, and the product was verified by mass spectrometry (Table I). The CTx ImI mutants are named as follows: the first letter and number refers to the wild type residue and position, and the following letter is the substituted residue at that position.
Confirmation of Disulfide Bond Synthesis by Ellman's Analysis-To confirm disulfide bond formation, we measured the colorimetric reaction of DTNB (Ellman's Reagent) with linear, non-oxidized CTx ImI and compared it with that of commercially available CTx MI and all of our synthetic CTx ImI mutants. 100 g of each conotoxin was dissolved in 200 l of 0.1 mM phosphate buffer, 4 l of DTNB was added, the mixture was incubated at room temperature for 30 min for color development, and absorbance at 405 nm was measured. Reactivity of each synthetic CTx ImI mutant is expressed relative to that obtained for 100 g of non-oxidized CTx ImI (Table I).
Construction of ␣ 7 /5HT-3 Chimera and Expression in 293 HEK Cells-Acetylcholine receptor subunit cDNAs were subcloned into the cytomegalovirus-based expression vector pRBG4 (10). The ␣ 7 /5HT-3 chimera (␣ 7 200/5HT-3) was constructed by bridging a 58-base pair synthetic oligonucleotide from a TfiI site in human ␣ 7 to an EcoRV site in rat 5HT-3. The nucleotide sequence of the chimera was confirmed by dideoxy sequencing. HEK cells were transfected with muscle or ␣ 7 / 5HT-3 cDNAs using calcium phosphate precipitation as described (10). Two days after transfection, intact cells were harvested by gentle agitation in phosphate-buffered saline containing 5 mM EDTA for ligand binding measurements.
Ligand Binding Measurements-Ligand binding to intact cells was measured by competition against the initial rate of 125 I-␣-Bgt binding (10). The cells were briefly centrifuged, resuspended in potassium Ringer's solution, and divided into aliquots for ligand binding. Potassium Ringer's solution contains: 140 mM KCl, 5.4 mM NaCl, 1.8 mM CaCl 2 , 1.7 mM MgCl 2 , 25 mM HEPES, and 30 mg/liter bovine serum albumin, adjusted to a pH of 7.4 with 10 mM NaOH. Specified concentrations of ligand were added 30 min prior to addition of 3.75 nM 125 I-␣-Bgt, which was allowed to bind 15 min to occupy approximately half of the surface receptors. Binding was terminated by addition of 2 ml of potassium Ringer's solution containing 600 M d-tubocurarine chloride. All experiments were performed at 24 Ϯ 2°C. Cells were harvested by filtration through Whatman GF-B filters using a Brandel Cell Harvester and washed three times with 3 ml of potassium Ringer's solution. Prior to use, filters were soaked in potassium Ringer's solution containing 4% skim milk. Nonspecific binding was determined in the presence of 10 nM ␣-Bgt and was typically 1% of the total number of binding sites. The total number of binding sites was determined by incubation with toxin for 120 min. The initial rate of toxin binding was calculated as described (17) to yield the fractional occupancy of competing ligand. Binding measurements were analyzed according to either the monophasic Hill equation (Equation 1) or the sum of two distinct binding sites (Equation 2), where Y is fractional occupancy of the competing ligand, K app is the apparent dissociation constant, n H is the Hill coefficient, K A and K B are intrinsic dissociation constants for two binding sites, and fract A is the fraction of sites with dissociation constant K A . Fitted parameters and standard errors were obtained using UltraFit (BIOSOFT). For multiple experiments, means of the individual fitted parameters and standard deviations are presented (Tables II and III).

RESULTS
␣ 7 /5HT-3 Receptor Assay-Studies of the homomeric ␣ 7 binding site have been limited by low expression of native ␣ 7 in mammalian cells (18). To increase expression, we and others constructed ␣ 7 /5HT-3 chimeras containing the extracellular domain of ␣ 7 joined to the M1 junction of the rat 5HT-3 receptor (19,20,23). We showed that the extracellular domain of human ␣ 7 maintained the ligand recognition properties of wild type human ␣ 7 , but that the presence of 5HT-3 sequence greatly enhanced expression (23). Thus, to determine affinities of the CTx ImI mutants described in the present study, we measured binding to ␣ 7 /5HT-3 receptors expressed in 293 HEK cells.
Structures of Wild Type and Mutant CTx ImI-Wild type and mutants of CTx ImI were synthesized as described under "Experimental Procedures." Molecular weights of each toxin were determined by mass spectrometry and compared with calculated molecular weights (Table I). The close agreement between measured and calculated molecular weights supports the amino acid compositions and formation of the two intramolecular disulfide bonds. To further confirm that both disulfide bonds formed, we assayed for free sulfhydryls using the colorimetric reaction of DTNB. Whereas the linear, nonoxidized CTx ImI reacted strongly with DTNB, neither the commercially available CTx MI nor any of the CTx ImI mutants reacted (Table I). Thus the mass spectrometry data combined with DTNB assay confirm that the wild type and mutant conotoxins are fully oxidized.
Neuronal Specificity of CTx ImI-To establish that CTx ImI shows the correct neuronal specificity for our ␣ 7 /5HT-3 chi-  mera, we compared binding of CTx ImI with that of the musclespecific conotoxins by competition against the initial rate of 125 I-labeled ␣-Bgt binding. As observed in other expression systems (21), CTx ImI binds with much higher affinity to ␣ 7 / 5HT-3 receptors than CTx MI, GI, or SI (Fig. 2, top panel and Table II). We further examined neuronal specificity of CTx ImI by comparing it to the muscle-specific ␣-conotoxins in binding to human adult muscle receptors. Whereas CTx MI, GI, and SI bind with high affinity to muscle receptors, CTx ImI binds with much lower affinity (Fig. 2, bottom panel and Table II). In addition, conotoxins GI and SI select between the two sites of the muscle receptor, with the two-site fit revealing dissociation constants different by 70-and 100-fold, respectively (Table II). By contrast, the monophasic binding of conotoxins ImI and MI indicate similar affinities for both binding sites of the muscle receptor; the two-site fit reveals dissociation constants different by only 5-and 10-fold for CTx ImI and MI, respectively.
Mutagenic Scan of CTx ImI-We introduced conservative substitutions for each non-cysteine residue in CTx ImI, and measured binding of each mutant toxin to ␣ 7 /5HT-3 receptors. The results reveal two key regions in CTx ImI essential for high affinity binding (Fig. 3 and Table III). The first region is the triad Asp-Pro-Arg in the first loop, where individual mutations decrease affinity by 70 -500-fold. The second region is the single tryptophan in the second loop, which when mutated to threonine decreases affinity by 30-fold. On the other hand, mutating the four remaining non-cysteines in CTx ImI does not significantly alter affinity for ␣ 7 /5HT-3 receptors. These mutations include acetylation of the amino-terminal glycine and neutralization of arginine in the second loop (R11Q). Thus mutation of four of the eight non-cysteines in CTx ImI alters affinity for ␣ 7 /5HT-3 receptors.
Side Chain Specificity of the Active Residues-To determine the chemical nature of the contributions of each of the four essential residues in CTx ImI, we introduced a systematic series of side chains at each position and measured binding of each mutant toxin to ␣ 7 /5HT-3 receptors (Fig. 4). Beginning with aspartic acid at position 5, neutralization by substituting asparagine decreases affinity by 100-fold, as described above (Fig. 3). However, replacement with glutamic acid, which lengthens the side chain but maintains the negative charge, decreases affinity even more. Introducing the positively charged lysine produces the greatest decrease of affinity at position 5. These results demonstrate that a negative charge and correct side chain length are required at position 5, suggesting an interaction with a focal electron acceptor in the receptor binding site.
We next examined the remaining charged residue of the triad, arginine at position 7. Surprisingly, maintaining the positive charge by mutation to lysine decreases affinity 100fold, similar to the decrease observed by mutation to the neutral glutamine (R7Q in Fig. 3). Introducing the negatively charged glutamic acid produces the greatest decrease of affinity at position 7. Thus the structural requirements at position 7 are analogous to those at position 5. A positive charge with particular size is required, suggesting interaction with an electron-rich subsite in the receptor.
We considered the possibility that Asp-5 and Arg-7 form an intramolecular salt bridge essential for activity of CTx ImI. Thus we switched the positions of the two charged residues with the mutation D5R ϩ R7D, with the goal of maintaining the salt bridge of the native toxin. The double mutant toxin decreases affinity by 2000-fold, the greatest decrease observed, indicating that either the salt bridge is not formed, or a salt  Table II. Bottom panel, 293 HEK cells were transfected with human muscle ␣ 2 ␤⑀␦ subunit cDNAs and binding of CTx ImI, GI, MI, and SI was determined as described under "Experimental Procedures." The curves through the data are fits to either the Hill equation (Equation 1) or the two-site equation (Equation 2). Mean and S.E. of the fitted parameters are given in Table II.  Fig. 2. K app is the apparent dissociation constant, n H is the Hill coefficient, and n is the number of independent experiments. K A and K B are the dissociation constants for each site obtained from a two-site fit with the fractional contribution of each site set to 0.5.  bridge between Asp-5 and Arg-7 is not the basis for high affinity.
The third member of the triad, proline at position 6, is conserved in all ␣-conotoxins described to date. The large decrease in affinity produced by the mutation P6G (Fig. 3) suggests that proline orients the side chains of Arg-7 and Asp-5 for optimal interaction with the receptor binding site. Because glycine contributes only a hydrogen side chain, we constructed P6A and P6V to increase side chain size to approach that of proline. Both P6A and P6V mutations decrease affinity similar to P6G, suggesting that conformational restriction by proline at position 6 is required for high affinity binding to the ␣ 7 receptor.
The second essential region is the lone tryptophan in the second loop at position 10, which when mutated to threonine decreases affinity by 30-fold (Fig. 3). To determine whether aromaticity is required at position 10, we exchanged tryptophan for phenylalanine. Affinity of the W10F mutant decreases only 3-fold compared with CTx ImI, indicating that an aromatic side chain is required at position 10.
The overall results reveal active residues in both loops of CTx ImI. The first loop contains a triad of residues that requires specific charge, side chain length, and conformational restriction. The second loop contains a tryptophan which contributes its aromatic ring to stabilize the ␣ 7 -CTx ImI complex.

DISCUSSION
To investigate the basis of specificity of CTx ImI for neuronal ␣ 7 receptors, we used ␣ 7 /5HT-3 chimeras to express ␣ 7 binding sites and to measure binding affinity of a series of CTx ImI mutants. The results reveal two regions of CTx ImI that confer specificity for human neuronal ␣ 7 receptors. The first region is the conformationally sensitive triad Asp-Pro-Arg within the first loop of CTx ImI. Subtle changes in side chain lengths of the aspartic acid and arginine reduce affinity, and their side chains appear to be held in place by the intervening proline. Thus, the triad must maintain a specific conformation to fit properly into a specific and focal counterpart in the ␣ 7 binding site. The second region is the single tryptophan within the second loop of CTx ImI. Studies of side chain specificity at FIG. 3. Mutagenic scan of CTx ImI. Dissociation constants of the CTx ImI mutants are expressed as the log ratio relative to wild type CTx ImI. The affinity of wild type CTx ImI for ␣ 7 /5HT-3 receptors is shown by the vertical bold line, and the error bars indicate Ϯ S.D. The drawings to the right are schematic representations of the mutant toxins, with the mutant residues highlighted. Mean and S.E. of the fitted parameters are given in Table III. FIG. 4. Side chain specificity of the determinants of CTx ImI affinity. For each CTx ImI mutant, affinity for ␣ 7 /5HT-3 receptors is expressed as in Fig. 3. The affinity of wild type CTx ImI for ␣ 7 /5HT-3 receptors is shown by the vertical bold line, and the error bars indicate Ϯ S.D. Mean and S.E. of the fitted parameters are given in Table  III. In the schematic representations to the right, X indicates the mutant residue. position 10 indicate the requirement of an aromatic ring. Mutagenesis and site directed labeling studies establish that each ligand binding site in muscle and neuronal AChRs contains contributions from both ␣ and non-␣ subunits. Residues of the ␣ portion of the binding site, termed the (ϩ) face, are located in three regions well separated in the primary sequence, suggesting a three-loop model of the (ϩ) face of the binding site (reviewed in Refs. 8 and 22). Similarly, residues of the non-␣ portion of the binding site, termed the (Ϫ) face, are located in four separate regions of the primary sequence, suggesting a four loop model for the (Ϫ) face of the binding site (8,22). Unlike the two binding sites of the muscle AChR, which are formed at interfaces between ␣ 1 -␦ and either ␣ 1 -␥ (fetal) or ␣ 1 -⑀ (adult) subunit pairs, binding sites of the homo-oligomeric ␣ 7 receptor are formed at interfaces between pairs of identical subunits, ␣ 7 -␣ 7 . Consequences of a homo-oligomeric pentamer include the potential for five binding sites and formation of both the (ϩ) and (Ϫ) faces by a single ␣ 7 subunit.
␣-CTx ImI is a competitive antagonist of neuronal ␣ 7 receptors. It contains two disulfide bonds which hold the toxin in a constrained two-loop structure. The solution and crystal structures of members of the ␣-conotoxin family reveal a compact triangular structure with two positive charges separated by 15 Å, similar to the rigid structure and 11 Å separation of quaternary nitrogens of curariform antagonists. Previous work showed that curariform antagonists bridge the interface between the ␣ and ␥ subunits of the muscle receptor through quaternary-aromatic interactions (9). Similarly, the two loops of CTx ImI likely bridge the (ϩ) and (Ϫ) faces of the ␣ 7 binding site. Because CTx ImI is small enough for structural determination at atomic resolution, it may be used as a molecular caliper to estimate distances between points of ligand contact at the ␣ 7 binding site.
Our studies reveal two distinct regions in CTx ImI essential for binding to neuronal ␣ 7 receptors. The first region is within the first loop of CTx ImI, the conformationally-sensitive triad Asp-Pro-Arg. Previous studies with CTx MI demonstrated the importance of proline in the first loop, where the mutation P6G reduced biopotency approximately 100-fold (14). Because our results with CTx ImI reveal a similar decrease in affinity with P6G, and because Pro-6 is conserved in all ␣-conotoxins, the proline likely contributes to structural rigidity along with the two disulfide bridges.
Aside from Pro-6, no other essential residues have been reported in the first loop of the ␣-conotoxins. For CTx ImI, we show that aspartic acid at position 5 and arginine at position 7 are essential for high affinity binding. Surprisingly, conservative substitutions of Asp-5 and Arg-7 markedly diminish activity of CTx ImI. The substitutions D5E and R7K, which alter side chain length by one methyl group while maintaining charge, reduce affinity more than 100-fold. In addition, introducing opposite charges at Asp-5 and Arg-7 reduces affinity approximately 1000-fold, suggesting repulsion by residues at the ␣ 7 binding site. The overall results of the Asp-5 and Arg-7 side chain experiments suggest that both charge and side chain length are important for the activity of CTx ImI. Together with residue Pro-6, Asp-5, and Arg-7 form a conformationally sensitive triad essential for CTx ImI affinity. Our results do not distinguish whether P6 contributes directly to CTx ImI binding or structurally stabilizes the first loop. However, the presence of a glycine mutation at position 6 likely allows rotational freedom around its ␣ carbon that would affect the orientation of the side chains of Asp-5 and Arg-7.
The second essential region in CTx ImI is the single Trp at position 10 of the second loop. Mutation of the two remaining non-cysteines in the second loop fails to affect CTx ImI affinity. Similar to CTx ImI, the muscle-specific conotoxins contain a conserved aromatic residue in the penultimate position of the second loop (Fig. 1). Tryptophan of CTx ImI occupies the position equivalent to that of tyrosine in CTx MI, GI, and SI. Replacement of L-tyrosine with D-tyrosine in CTx MI decreases bioactivity, indicating that the conformation of the tyrosine is essential (14). Our results show that converting tryptophan to threonine in CTx ImI reduces affinity 30-fold. However, converting tryptophan to phenylalanine, which maintains the aromatic side chain, decreases affinity only 3-fold. Thus an aromatic side chain at position 10 stabilizes the CTx ImI-␣ 7 receptor complex.
Surprisingly, mutations similar to those that affect affinity of muscle ␣-conotoxins do not affect affinity of CTx ImI. For example, a cationic side chain in the second loop is critical for activity of the muscle-specific ␣-conotoxins MI and SI (6, 7). Our results with CTx ImI show no effect of the mutation R11Q. In addition, previous work suggested that the N-terminal amide stabilizes the toxin-receptor complex through a -cation interaction (12). By contrast, acetylation of the amino-terminal glycine does not affect affinity of CTx ImI for ␣ 7 receptors.
The overall results reveal two structural motifs in CTx ImI that confer high affinity binding to ␣ 7 receptors. Knowledge of the precise contacts between CTx ImI and ␣ 7 awaits experiments that mutate residues in both the toxin and the receptor.