NMR-based Binding Screen and Structural Analysis of the Complex Formed between α-Cobratoxin and an 18-Mer Cognate Peptide Derived from the α1 Subunit of the Nicotinic Acetylcholine Receptor fromTorpedo californica *

The α18-mer peptide, spanning residues 181–198 of the Torpedo nicotinic acetylcholine receptor α1 subunit, contains key binding determinants for agonists and competitive antagonists. To investigate whether the α18-mer can bind other α-neurotoxins besides α-bungarotoxin, we designed a two-dimensional1H-15N heteronuclear single quantum correlation experiment to screen four related neurotoxins for their binding ability to the peptide. Of the four toxins tested (erabutoxin a, erabutoxin b, LSIII, and α-cobratoxin), only α-cobratoxin binds the α18-mer to form a 1:1 complex. The NMR solution structure of the α-cobratoxin·α18-mer complex was determined with a backbone root mean square deviation of 1.46 Å. In the structure, α-cobratoxin contacts the α18-mer at the tips of loop I and II and through C-terminal cationic residues. The contact zone derived from the intermolecular nuclear Overhauser effects is in agreement with recent biochemical data. Furthermore, the structural models support the involvement of cation-π interactions in stabilizing the complex. In addition, the binding screen results suggest that C-terminal cationic residues of α-bungarotoxin and α-cobratoxin contribute significantly to binding of the α18-mer. Finally, we present a structural model for nicotinic acetylcholine receptor-α-cobratoxin interaction by superimposing the α-cobratoxin·α18-mer complex onto the crystal structure of the acetylcholine-binding protein (Protein Data Bank code 1I9B).

The nicotinic acetylcholine receptor (nAChR) 1 mediates excitatory transmission at the neuromuscular junction and in the central and peripheral nervous systems. As the prototype for the superfamily of ligand-gated ion channels, it has been intensively studied. The ligand-gated ion channel family includes the glycine, ␥-aminobutyric acid, type A, ␥-aminobutyric acid, type C, and 5-hydroxytryptamine type 3 receptors. All are apparently pentamers composed of either identical or homologous transmembrane subunits. The skeletal muscle-type nAChR contains two ␣1 subunits and one each of the ␤1, ␥(⑀), and ␦ subunits. The ␣␥(⑀) and ␣␦ interfaces form the sites for ligand binding (see Ref. 1 for review). Recently, a natural homologue of the extracellular domain of the nAChR, the acetylcholine-binding protein (AChBP), was discovered as a soluble, secreted pentamer in the central nervous system of the snail, Lymnaea stagnalis. The x-ray crystal structure of this AChBP at 2.7 Å reveals the putative conformational architecture of the extracellular domain of the nAChR (2).
The snake venom-derived ␣-neurotoxins are classic competitive antagonists of the muscle-type nAChRs, and they can be divided into two related groups as follows: short and long ␣-neurotoxins. X-ray crystal structures and NMR solution structures of several different ␣-neurotoxins reveal a conserved structural pattern consisting of a core of disulfides with three fingers or loops (3)(4)(5)(6)(7)(8)(9)(10)(11)(12). Short ␣-neurotoxins have four core disulfide bonds, whereas the long ␣-neurotoxins have, in addition, a fifth disulfide in the tip of loop II.
Mutagenesis studies involving the nAChR and several ␣-neurotoxins (13)(14)(15)(16)(17) have helped elucidate the interactions between the nAChR and the ␣-neurotoxins and have led to new models based on the docking of a rigid structural model of an ␣-neurotoxin to the proposal ligand-binding site (17).
It binds muscle-type nAChR at the neuromuscular junction and causes paralysis by preventing acetylcholine binding to the nAChR. LSIII, from the venom of Laticauda semifasciata, is also classified as a long ␣-neurotoxin because of its fifth disulfide bond and is composed of 66 amino acid residues (8). Erabutoxin a (Ea) and erabutoxin b (Eb) are both closely related short ␣-neurotoxins, with 62 amino acid residues, that are also found in the snake venom from L. semifasciata.
Previous work (18) indicates that residues 173-204 from the ␣1 subunit of the nAChR form a major binding determinant for ␣-bungarotoxin (Bgtx), a long ␣-neurotoxin from Bungarus multicinctus. A peptide fragment (␣18-mer) with a sequence corresponding to residues 181-198 (␣1-181 YRGWKHWVYYTC-CPDTPY 198 ), from the Torpedo californica nAChR, binds Bgtx with an apparent K d of ϳ65 nM (19). We have described previously three NMR solution structures of Bgtx in complex with this ␣18-mer (10) and with two other ␣-subunit derived peptides, an ␣12-mer (Torpedo ␣1, 185-196) (20) and an ␣19-mer (Chick ␣7, 178 -196) (12). The structures of these Bgtx⅐peptide complexes provide valuable information about the contribution of the contact residues involving in Bgtx binding to the native nAChR. We now report an NMR-based binding assay using the ␣18-mer to screen for interaction with four ␣-neurotoxins related to Bgtx. We also report an NMR structural analysis of the Cbtx⅐␣18-mer complex. Our goal is to further define the contribution of individual amino acid residues in the ␣1 subunit to the interaction between the nAChR and the ␣-neurotoxins.

EXPERIMENTAL PROCEDURES
Expression of the ␣18-Mer Peptide-The ␣18-mer was expressed and purified as described previously (10). Briefly, the peptide was expressed as a ketosteroid isomerase fusion protein using plasmid pET 31b(ϩ) in standard M9 medium and Escherichia coli strain BL21 (DE3) (Novagen) with the replacement of 15 NH 4 Cl for NH 4 Cl to uniformly enrich the peptide with 15 N. The peptide was cleaved from the fusion protein by reaction with cyanogen bromide, and it was then purified by reverse phase-high performance liquid chromatography. As before, the C-terminal homoserine lactone form of the ␣18-mer was used for the NMR studies described below.
NMR Sample Preparation-For the HSQC screen, the 15 N-enriched ␣18-mer was resuspended in 50 mM of per-deuterated potassium acetate buffer (pH 4.0) with 5% D 2 O and 0.05% sodium azide. Cbtx, Ea, Eb, and LSIII (from Sigma) were prepared in the same buffer at a concentration of 1 mM. Toxins from their stock solutions were added to the ␣18-mer in a 1:1 stoichiometry. The final concentration of all mixtures was 0.2 mM. For the structure analysis of the Cbtx⅐␣18-mer complex, the 15 N labeled ␣18-mer was resuspended in the buffer described above. Cbtx, from a 5 mM stock solution, was then added to the ␣18-mer to form a 1:1 Cbtx⅐␣18-mer complex. The final concentration of the Cbtx⅐␣18mer complex was 1.6 mM.
NMR Experiments-All NMR spectra were recorded on a Bruker Avance 600 MHz NMR spectrometer at a temperature of 35°C. Chemical shifts at this temperature were calibrated with respect to internal 3-(trimethylsilyl) tetradeutero sodium propionate (0.0 ppm). The various toxin/␣18-mer samples were analyzed by a two-dimensional 15 N heteronuclear single quantum correlation ( 1 H-15 N HSQC) experiment (21)(22)(23). The formation of the Cbtx⅐␣18-mer complex was confirmed in a mole ratio titration series using the HSQC protocol for analysis. Amino acid spin systems were identified by two-dimensional total correlation spectroscopy (TOCSY) (21,22,24) and three-dimensional TOCSY-HSQC experiments (24 -28) using a 60-ms MLEV-17 spin-lock sequence. The assignments of the H N protons and H ␣ protons of the amino acid spin systems of the peptide were further confirmed by a threedimensional HNHA experiment (29,30). Nuclear Overhauser effect (NOE) correlations (sequential, medium range, and long range NOEs) were identified by two-dimensional nuclear Overhauser enhancement spectroscopy (NOESY) (21,22) and three-dimensional NOESY-HSQC experiments (25-28) with a mixing time of 120 ms. NH exchange experiments in D 2 O were performed to identify slowly exchanging amide protons presumably involved in stable hydrogen bonds. The Cbtx⅐␣18-mer complex sample was lyophilized, and immediately after resuspending the complex in 99.99% D 2 O, an HSQC and 12 sequential TOCSY spectra (3-h acquisition time for each) were collected. All NMR spectra were processed and analyzed with XwinNmr (Bruker), NMRPipe (31), and Sparky (32).
Conformation Calculation and Analysis-The distance constraints were obtained from the cross-peak volumes in the two-dimensional NOESY spectra of the Cbtx⅐␣18-mer complex by integration using the Gaussian fitting protocol in Sparky. The cross-peaks, according to their volumes, were manually classified into three categories: strong, medium, and weak with corresponding distance ranges of 1.8 -3.0, 1.8 -4.0, and 1.8 -5.0 Å, respectively. Those amide protons, involved in ␤-sheetlike H ␣ -H ␣ and H ␣ -H N NOEs, which could be identified in NH exchange experiments as stable to exchange after 21 h in D 2 O were classified as "hydrogen bond" protected amide protons and as hydrogen bond donors. For the structural incorporation of hydrogen bonds, the distance con-straint of H N -O was assigned a value of 1.6 -2.5 Å and that of N-O was constrained to 2.5-3.3 Å. The H ␣ -H N 3 J coupling constants of the ␣18mer peptide were derived from the three-dimensional HNHA experiments (33). The 3 J coupling constants were converted to dihedral angle restraints using methods described previously (34). For 3 J Ͻ 6 Hz, the dihedral angle restraint () was assigned to Ϫ60 Ϯ 30°; for 3 J Ͼ 8 Hz, the was Ϫ120 Ϯ 40°. All structures were calculated from random models with distance geometry and by applying the simulated annealing protocols in the Crystallography and NMR System (CNS) software package (35). The potential energy function used in these calculations was the sum of the van der Waals repulsion term whose force constant varied from 0.003 to 4 kcal mol Ϫ1 Å Ϫ4 during the cooling stage, the NOE distance constraints using a square-well potential with a force constant of 50 kcal mol Ϫ1 , the dihedral angles with a force constant of 200 kcal mol Ϫ1 rad Ϫ2 , the bond length, and the bond angles. Pseudoatoms were used for protons that could not be stereospecifically assigned. The pseudoatom correction feature of CNS was used to adjust the NOE distance constraint range automatically. Each round of calculation was initiated with a random seed number. Resulting structural models with no NOE violation larger than 0.5 Å were regarded as acceptable structures. The 10 lowest energy structures of 100 acceptable structures were selected to form an ensemble to represent the final structural model of the Cbtx⅐␣18-mer complex. The intermolecular contact surface areas of all the final individual Cbtx⅐␣18-mer complex structures were calculated using Contacts of Structural Units software (36). The analysis of energetically significant cation-interactions within the structure models of the Cbtx⅐␣18-mer complex was performed using CaP-TURE (37). The visualization of the structures was carried out in MOLMOL (38). All structure coordinates of the Cbtx⅐␣18-mer complex have been deposited at the Research Collaboratory for Structural Bioinformatics Protein Data Bank. The entries for the ensemble structures and the minimized average structure are 1LXG and 1LXH, respectively. (10), the ␣18-mer can form a stoichiometric complex with Bgtx that is amenable for NMR structure analysis. To investigate whether the ␣18-mer can also be recognized by other ␣-neurotoxins, we designed an NMR-based screen to test for interactions with ␣-neurotoxins related to Bgtx (Cbtx, Ea, Eb, and LSIII). The screen was based on the use of an 15 N-enriched ␣18-mer and a two-dimensional 1 H-15 N HSQC experiment that is designed to acquire signal only from 15 N-attached protons of the peptide (i.e. amide protons and the side chain protons of Arg and Trp). Comparison of the spectra of the ␣18-mer⅐␣neurotoxin samples with that of free ␣18-mer was used to determine whether there is an interaction between the ␣18mer and these toxins. A binding interaction between ␣18-mer and the toxin is indicated in those cases where H N resonances of the ␣18-mer undergo chemical shift changes. An HSQC titration is necessary to determine whether this binding is in the realm of slow, medium, or fast exchange. If none of the H N resonances of the ␣18-mer undergo any chemical shift changes upon toxin addition, an interaction in the millimolar to submillimolar affinity range (the NMR sample concentration range) can be effectively ruled out.

HSQC Screen and Titration-As described previously
The results shown in Fig. 1 indicate that Ea (Fig. 1A), Eb (Fig. 1B), and LSIII ( Fig. 1C) do not bind to the ␣18-mer, whereas Cbtx does form an apparent stoichiometric complex with the ␣18-mer (Fig. 1D). We next carried out HSQC-based mole ratio titration studies to determine whether the Cbtx⅐␣18mer complex was indeed stoichiometric and whether the interaction is in slow exchange (Fig. 2). From the HSQC titration, it is clear that the free peptide ( Fig. 2A) is largely unstructured. Upon binding to Cbtx, the chemical shift positions of the H N resonances for the bound peptide are altered due to complex formation but remain fixed in position and do not vary as a function of the Cbtx concentration (Fig. 2, B-D). These observations indicate that the ␣18-mer forms a 1:1 complex with Cbtx, that the ␣18-mer acquires structure upon binding, and that the Cbtx⅐␣18-mer complex is in slow exchange (i.e. time scale in the millisecond to second range).
NMR Assignment-Three-dimensional TOCSY-HSQC, threedimensional NOESY-HSQC, and three-dimensional HNHA experiments were performed to assign the resonances of the ␣18-mer in the Cbtx⅐␣18-mer complex. From these three-dimensional NMR experiments, we assigned the observable resonances for all of the amino acid residues in the ␣18-mer except for the N-terminal Tyr 181 , which has an exchangeable H N , and the two prolines that lack amide protons. These assignments of the peptide resonances were used in the analysis of the two-dimensional NMR data to distinguish peptide resonances from Cbtx-associated proton resonances, greatly facilitating the assignment of the Cbtx resonances. Guided by the two-dimensional NMR assignments of free Cbtx (6), we completed the two-dimensional NMR assignments for free Cbtx and for the Cbtx⅐␣18-mer complex with additional homonuclear two-dimensional NOESY and two-dimensional TOCSY experiments. By comparing the chemical shift assignments for free and bound Cbtx, we found that the region from Asp 27 to Val 37 of Cbtx is characterized by significant chemical shift changes upon complex formation ( Table I).
Structure of the Cbtx⅐␣18-Mer-Distance constraints were obtained from the intensities of the NOE peaks in two-dimensional NOESY spectra of the complex as described under "Experimental Procedures," whereas hydrogen bonds important for secondary structure were identified by NH exchange experiments. In addition, the dihedral angle restraints for the peptide were calculated from H ␣ -H N 3 J couplings obtained through HNHA experiments. All of these constraints (Table II) were then incorporated into CNS (35) to calculate structural models of the Cbtx⅐␣18-mer complex. The CNS calculations incorporate both distance geometry and simulated annealing protocols. After multiple rounds of calculation using different random seed numbers, a pool of 100 structures with no NOE violation larger than 0.5 Å was obtained. From the pool of acceptable structures, 10 structures with the lowest potential energy were selected to represent a structural ensemble of the Cbtx⅐␣18-mer complex. The overall backbone atomic root mean square deviation (r.m.s.d.) between the individual structures and the mean structure of the Cbtx⅐␣18-mer complex is 1.69 Å. We identified a highly defined region composed of residues 2-70 of Cbtx and residues 187-196 of the ␣18-mer. The remaining residues (residues 1 and 71 of Cbtx, 181-186 and 197-198 of the ␣18-mer) lacked long range NOE constraints. In this more highly defined region, the backbone r.m.s.d. is 1.46 Å (Table II). The structural ensemble of the complex is shown in Fig. 3. For clarity, only the more highly defined regions are shown. The ␣18-mer-bound Cbtx is oriented to show the concave surface, loop I on the left, loop III on the right, and the tip of the loop II at the bottom. The overall three-finger-like motif In free Cbtx, this central ␤-sheet is a triple-stranded one with two strands within loop II and the third strand contributed by loop III (6). In our spectra of free Cbtx, we also observed long range H ␣ -H ␣ and H ␣ -H N NOEs within loop II between Val 19 -Trp 25 and Arg 36 -Ala 42 region as described previously (6). However, those NOEs and accompanying slow-exchanging amide protons were not seen in the spectra of the Cbtx⅐␣18-mer complex. We think that the binding induced conformational changes involving the tip of loop II weaken this intra-loop ␤-sheet to a considerable extent, so the typical evidence for a ␤-sheet was not observed within loop II of the peptide-bound Cbtx. Many of the large changes (Ͼ0.15 ppm) in chemical shift in bound Cbtx (Table I) indicate structural alterations in the tip of loop II (Asp 27 -Val 37 ) consistent with extensive contacts between the Ile 32 -Arg 36 region of Cbtx and the peptide (Fig. 3).
A total of 20 intermolecular NOEs define the contact zone between the ␣18-mer and Cbtx (Table III). The contact zone  2. The mole ratio titration of the Cbtx⅐␣18-mer complex as detected by HSQC. A, the HSQC spectrum of free ␣18-mer; B, the HSQC spectrum of the ␣18-mer with a half-molar equivalent of Cbtx; C, the HSQC spectrum of the ␣18-mer with a 0.75 molar equivalent of Cbtx; D, the HSQC spectrum of the equimolar mixture of the ␣18-mer and Cbtx. The solid arrow indicates one of the resonances in free ␣18-mer whose intensity decreases upon the incremental addition of Cbtx. Its chemical shift remains constant throughout. The double-headed arrow points to one of the resonances in the Cbtx-bound ␣18-mer. Its intensity increases with the incremental addition of Cbtx, but its chemical shift remains unchanged indicative of slow exchange. has a surface area of ϳ720 Å 2 and involves the following residues: Ile 9 and Thr 10 in loop I of Cbtx; Ile 32 -Arg 36 in loop II of Cbtx; Arg 68 and Lys 69 in the C-terminal tail of Cbtx; and Tyr 189 -Thr 196 in the ␣18-mer. Tyr 189 alone makes multiple contacts with Cbtx residues Ile 9 , Ile 32 , Arg 33 , and Lys 35 (Table  III). The aromatic ring of Tyr 190 interacts extensively with cationic residues Arg 36 , Arg 68 , and Lys 69 (Table III). Intermolecular NOEs between H ␣ and H N of Thr 196 and H ␣ of Ile 9 , between the amide proton of Tyr 189 and ␥ protons of Ile 9 , together with NOEs involving the side chain of Tyr 189 and H ␣ of Ile 9 indicate the formation of a hairpin-like structure in the ␣18-mer upon complexing with Cbtx (Table III) (Table III) imply that these two residues also participate in the contact between the ␣18-mer and Cbtx.
The proximity of aromatic residues in the ␣18-mer (Tyr 189 and Tyr 190 ) to several cationic residues in Cbtx (Arg 33 , Lys 35 , Arg 36 , Arg 68 , and Lys 69 ) suggests that cation-interactions may play an important role in the stability of the Cbtx⅐␣18-mer complex. Following cation-interaction analysis of the 10 structural models of the Cbtx⅐␣18-mer complex using CaPTURE (37), we observed that 6 of the 10 structures contain energetically significant intermolecular cation-interactions as defined by the analysis program. The following cation-pairs were observed in the six AChBP⅐Cbtx Model-The backbone r.m.s.d. between the contact residues in the ␣18-mer (Tyr 189 -Thr 196 ) in the complex and the corresponding region from the AChBP (Thr 184 -Ala 191 ) is 1.37 Å. Because of this close fit, we were able to superimpose the structure of the Cbtx⅐␣18-mer complex onto residues 184 -191 in the structure of the AChBP. No additional structural manipulations were needed to obtain a structural model for the predicted AChBP⅐Cbtx (Fig. 4) containing no steric clashes. In this initial model, statically docked loop III of Cbtx is in closer proximity to the ␣ subunit than to the adjoining subunit, and the concave face of Cbtx is closer to the ␣ subunit than the convex face is to the adjoining subunit. This model implies that the bulk of Cbtx is oriented, with respect to the height of the AChBP, such that it is closer to the C-terminal end than the N-terminal end of the AChBP and that it is nearly perpendicular to the long central axis of the AChBP. DISCUSSION By having previously documented that the ␣18-mer can bind Bgtx with high affinity, we wanted to test other related ␣-neurotoxins for ␣18-mer binding. We developed an NMR-based method to test for such binding. We found that the ␣18-mer can bind and form a stable complex with the long ␣-neurotoxin, Cbtx. From the observed intermolecular NOEs and the structural model of the Cbtx⅐␣18-mer complex, we obtained important information on the points of interaction between the ␣18mer and Cbtx.
Our NMR-based screen for ␣-neurotoxin binding to the ␣18mer showed that ␣-neurotoxins Ea, Eb, and LSIII do not appear to interact significantly with the ␣18-mer. An alignment of sequences from the C-terminal tail region (Fig. 5) indicates that Ea, Eb, and LSIII all have relatively shorter C-terminal tail segments than either Cbtx or Bgtx. Importantly, they also lack cationic residues that Cbtx and Bgtx both contain in that region. Previous studies have suggested the importance of the C-terminal tail regions of long ␣-neurotoxins in peptide binding (10,11) and the involvement of this region in Bgtx binding to intact nAChRs (15,39). It is likely therefore that the reason that Ea, Eb, and LSIII fail to bind to the ␣18-mer is due to these differences in the C-terminal tail region. It should be noted that although LSIII is categorized as a long ␣-neurotoxin because of  Table II. The backbone of Cbtx is colored blue, and the backbone of the ␣18-mer is colored red. The N and C termini of Cbtx are colored black. This figure was prepared using the program MOLMOL (38).  the fifth disulfide bond in loop II, its C-terminal tail region, beyond the final Cys residues, is only two amino acids longer than the short ␣-neurotoxins, Ea and Eb. It has been reported previously that short ␣-neurotoxins appear, in general, to have faster on-rates (and faster off-rates) than Bgtx (40). This suggests some fundamental differences in the mode of binding among ␣-neurotoxins. Indeed, the results from a footprinting protection study argue that the short ␣-neurotoxins and Bgtx bind to opposite faces of loop C in the 187-190 region (16). As a consequence of this difference in binding mode to loop C residues, the short ␣-neurotoxins may interact more strongly with additional receptor regions elsewhere (e.g. Trp 149 from loop B (41) or Leu 119 and Glu 176 from the adjoining subunit (42)). The interaction between Cbtx and the ␣18-mer is consistent with an earlier report that Cbtx can bind to a 32-mer peptide (corresponding to ␣1 subunit residues 173-204) with affinity similar to that of Bgtx (43). The HSQC titration (Fig. 2) results clearly demonstrate that the ␣18-mer can bind and form a stable 1:1 complex with Cbtx. This titration further indicated that the Cbtx⅐␣18-mer complex is in slow exchange and that the peptide undergoes a binding induced conformational stabilization. The results obtained with the NMR-based binding screen were consistent with competition-based binding studies using 125 I-Bgtx and surface-immobilized ␣18-mer where significant binding was evident (data not shown).
We conclude that the tip of loop I, together with loop II, and the C-terminal tail region of Cbtx form the peptide-binding pocket. This binding profile is consistent with all available NMR solution and x-ray crystal structures of complexes between Bgtx and nAChR-derived peptides (10,12,20) or mimotope peptides (11,44,45). The contact zone in the ␣18-mer spans from Tyr 189 to Thr 196 and is consistent with earlier struc-tural studies of Bgtx bound to the ␣18-mer (10) or to an ␣7derived 19-mer (12). The identification of this contact region is also well supported by double mutant cycle analyses of recombinant Cbtx and the ␣7 nAChR (17) as described in detail below.
According to the intermolecular NOEs (Table III), Ile 9 and Thr 10 from loop I of Cbtx are involved in ␣18-mer binding. This is in agreement with the structures of the Bgtx⅐␣18-mer and Bgtx⅐␣19-mer complexes in which, Thr 8 and Pro 10 (in the Bgtx⅐␣18-mer complex) and Ala 7 , Ser 9 , and Ile 11 (in the Bgtx⅐␣19-mer complex) have contacts with their respective bound peptides (10,12). The intermolecular NOEs between Ile 9 of loop I in Cbtx and the Tyr 189 in the ␣18-mer (Table III) have counterparts in the Bgtx⅐␣19-mer (12) and in the complexes formed between Bgtx and library-derived mimotope peptides (11,44,45). This finding is in agreement with site-directed mutagenesis studies of Ea where mutations S8T and Q10A have been shown to cause a large reduction in binding affinity to intact nAChR (14). Interestingly, single site mutational analysis of loop I residues in Cbtx failed to detect a significant role for this region in binding to the Torpedo nAChR (14). This may imply that the tip of loop I does not form a major binding determinant for the native receptor or that one or the other of Ile 9 and Thr 10 is sufficient for peptide and receptor binding. The P193A mutation on ␣7 nAChR (Pro 193 in ␣7 corresponds to Thr 196 in the ␣18-mer which is in contact with Ile 9 ) reduces the affinity for Cbtx by only 2-fold (17), suggesting that the Ala substitution mutation at this site is not sufficient to disrupt receptor recognition.
As indicated by the network of contacts between Tyr 189 -Tyr 190 and the Ile 32 -Arg 36 region in Cbtx, the loop II region of Cbtx forms the main contact site for the ␣18-mer. This is consistent with the structures of other Bgtx⅐peptide complexes; in those structures, Tyr 189 and Tyr 190 (or correspondingly, Phe 186 and Tyr 187 , or Tyr 3 and Tyr 4 ) have extensive interactions with Leu 38 -Val 40 of Bgtx, corresponding to Lys 35 -Val 37 of Cbtx (10 -12, 44, 45). This binding profile is also consistent with a recent double-mutant cycle analysis using Cbtx and the ␣7 nAChR. In this study, the R33E mutation reduced the binding affinity by 339-fold; the K35E mutation caused a 144fold reduction in binding affinity, and the R36E mutation gave rise to a 456-fold drop in affinity (17). An earlier mutagenesis study with Cbtx indicated that the R33E mutation decreases binding affinity for Torpedo nAChR by 767-fold, whereas the R36A mutation results in a 7.4-fold reduction in affinity (14). The functional importance of Arg 33 has also been pointed out in mutagenesis studies of Bgtx. The R36A mutation in Bgtx (corresponding to Arg 33 in Cbtx) caused a 90-fold decrease in Bgtx binding to the mouse nAChR (15). On the receptor side, the F186E mutation in ␣7 (corresponding to Tyr 189 in ␣18-mer) causes a 97-fold reduction in affinity, and the Y187H mutation (equivalent to Tyr 190 in ␣18-mer) decreases affinity by 201-fold (17). In addition, chimeric analysis and toxin footprinting studies of nAChR ␣-subunits also support the involvement of Tyr 189 in Bgtx binding. A Bgtx-insensitive ␣3 subunit can be engineered to acquire sub-micromolar affinity for Bgtx with a single K189Y mutation (46). In a footprinting protection study with mouse ␣1 subunit, Bgtx protected the F189C Cys substitution mutation from reacting with a hydrophilic biotinylmaleimide (16). Mutagenesis studies also have identified Tyr 190 as being important in ␣-neurotoxin binding (13,16,41).
The intermolecular NOEs between Tyr 190 and Arg 68 and Lys 69 highlight the involvement of the C-terminal tail region of Cbtx in peptide binding. This finding is consistent with the Bgtx⅐␣18-mer structure (10) and the crystal structure of a complex formed between Bgtx and a library-lead 13-mer peptide (11), where the C-terminal tail region participates in bind- FIG. 4. Structural model of the AChBP⅐Cbtx complex. This structural model was generated by superimposing the hairpin-like structure of the ␣18-mer (Tyr 189 -Thr 196 ) in the Cbtx⅐␣18-mer complex with the corresponding region in the x-ray structure of the AChBP (Thr 184 -Ala 191 ). For clarity, only two adjacent subunits of the AChBP are shown, and the structure of the ␣18-mer is removed from the figure. The AChBP subunit on the left (color blue) corresponds to the ␣-subunit in nAChR and accounts for the major contacts with Cbtx. The AChBP subunit on the right (color green) would correspond to the adjacent non-␣ subunit found in heteromeric nAChRs (e.g. ␥(⑀) or ␦ subunits). The bound-Cbtx is colored red. The figures were prepared using the program MOLMOL (38). ing. The double mutant cycle analysis involving Cbtx and the ␣7 nAChR together with the earlier mutagenesis studies of Cbtx, assessed by binding assays with Torpedo nAChR, also support a role for the C-terminal tail region in Cbtx binding. The F65A mutation has been shown to reduce significantly Cbtx binding affinity for ␣7 nAChR or Torpedo nAChR by 16and 7-fold, respectively (14,17).
We also have focused on the cation-interactions between the positively charged side chains of Arg and Lys and the electronegativity associated with the aromatic cloud in the side chains of Phe, Tyr, and Trp (47). The structure of the Bgtx⅐␣18-mer complex suggested that cation-interactions could be involved in high affinity binding (10). Due to the close contact between aromatic residues of ␣18-mer and cationic residues in Cbtx, it is not surprising that we observed energetically favorable cation-interaction pairs in 60% of the structure models of Cbtx⅐␣18-mer complex. Interestingly, when we used the CaPTURE program (37) to analyze the x-ray crystal structure of a Bgtx⅐mimotope complex (11), we found a cationpairing between Arg 36 and Tyr 4 of the mimotope. This corresponds to the Arg 33 /Tyr 190 pair in the Cbtx⅐␣18-mer complex. The important roles of the cationic residues in Cbtx and of the aromatic residues in the peptide have also been indicated by mutagenesis studies, as reviewed above. Furthermore, chemical modifications of Arg 34 , Arg 37 , Arg 70 , and Arg 72 in toxin a (Ophiophagus hannah) lead to almost complete block of binding activity to nAChR (48). This also points to the importance of cationic residues in binding to the nAChR. We suggest that cation-interactions contribute to the high affinity interaction between Cbtx and the nAChR in a manner similar to the role of cation-interactions in the binding of acetylcholine to the nAChR (49). Additional high resolution structural information focusing on the side chains of Cbtx and Bgtx, together with further biochemical and mutagenesis studies involving the candidate cationic residues, will be required to further evaluate this mechanism.
Because AChBP is a homologue of the extracellular domain of the nAChR (2), we used its x-ray crystal structure to generate a simple, statically docked structural model for the nAChR⅐␣-neurotoxin complex (Fig. 4). The interaction between the nAChR and Bgtx has been modeled previously by superimposing the Bgtx⅐␣19mer or Bgtx⅐mimotope complexes with AChBP (11,12). Here we present a nAChR⅐Cbtx model based on the structure of the Cbtx⅐␣18-mer complex (Fig. 4). The orientation of Cbtx is similar to the two AChBP⅐Bgtx models and the recent model of the AChBP⅐Cbtx complex that was generated by manual docking aimed at incorporating available mutagenesis data (17). Our model, based on the structural information provided by the Cbtx⅐␣18-mer complex, provides strong indications that Cbtx is likely to orient itself at the subunit interface, perpendicular to the long central axis of the receptor. All of these recent models indicate that ␣-neurotoxins approach their binding site from the periphery, rather than from the vestibule.