NMR Solution Structure of Butantoxin

Archives of Biochemistry and Biophysics Vol. 379, No. 1, July 1, pp. 18 –27, 2000 doi:10.1006/abbi.2000.1858, available online at http://www.idealibrary.com on NMR Solution Structure of Butantoxin 1 S. Kent Holaday, Jr.,* Brian M. Martin,† Paul L. Fletcher, Jr.,‡ and N. Rama Krishna* ,2 *Department of Biochemistry and Molecular Genetics, University of Alabama at Birmingham, Birmingham, Alabama 35294-2041; †Clinical Neuroscience Branch, NIMH, National Institutes of Health, Bethesda, Maryland 20892-4405; and ‡Department of Microbiology and Immunology, East Carolina University School of Medicine, Greenville, North Carolina 27858 Received December 22, 1999, and in revised form April 7, 2000 The NMR structure of a new toxin, butantoxin (BuTX), which is present in the venoms of the three Brazilian scorpions Tityus serrulatus, Tityus bahiensis, and Tityus stigmurus, has been investigated. This toxin was shown to reversibly block the Shaker B potassium channels (K d ; 660 nM) and inhibit the proliferation of T-cells and the interleukin-2 production of antigen-stimulated T-helper cells. BuTX is a 40 amino acid basic protein stabilized by the four disulfide bridges: Cys2-Cys5, Cys10-Cys31, Cys16-Cys36, and Cys20-Cys38. The latter three are conserved among all members of the short-chain scorpion toxin family, while the first is unique to BuTX. The three-dimensional structure of BuTX was determined using 1HNMR spectroscopy. NOESY, phase sensitive COSY (PH-COSY), and amide hydrogen exchange data were used to generate constraints for molecular modeling calculations. Distance geometry and simulated annealing calculations were performed to generate a family of 49 structures free of constraint violations. The secondary structure of BuTX consists of a short 221 turn a-helix (Glu15-Phe23) and a b-sheet. The b-sheet is composed of two well-defined antiparallel strands (Gly29-Met32 and Lys35-Cys38) connected by a type-I* b-turn (Asn33-Asn34). Residues Cys5-Ala9 form a quasi-third strand of the b-sheet. The N-terminal C2-C5 disulfide bridge unique to this toxin does not appear to confer stability to the protein. © 2000 Academic Press 1 This study was supported in part by NSF Grant MCB-9630775, NCI Grant CA13148 (NMR and Mass Spectrometry Facilities), and the Biotechnology Program Fund of the ECU School of Medicine. The atomic coordinates (PDB codes 1C55 and 1C56) have been deposited in the Protein Data Bank, Research Collaboratory for Structural Bioinformatics, Rutgers University, New Brunswick, NJ (http://www.rcsb.org/). The chemical shifts were deposited in the BioMagResBank (Accession Code: 4443). 2 To whom correspondence and reprint requests should be addressed. Fax: (205) 934-6475. E-mail: nrkrishna@bmg.bhs.uab.edu. 18 Key Words: butantoxin; 2D-NMR; solution structure; potassium channels. Scorpion venom is a complex mixture of small proteins, histamine, serotonin, enzymes, enzyme inhibitors, mucus, and other poorly characterized compounds. These small proteins interact specifically with the Na 1, K 1, and Cl 2 channels of mammals and insects, and are responsible for the neurotoxic activity of scorpion venom (1). The Na 1 channel toxins are responsible for the majority of toxicity associated with envenomation in humans (2). The K 1 channel toxins are less important medically. However, they are valuable as tools for studying K 1 channel function and localization. The K 1 channel toxins are small basic proteins with 31– 41 residues and 3 or 4 disulfide bonds. Structural studies on K 1 channel toxins from scorpions have revealed a conserved three-dimensional (3D) scaffold composed of an a-helix connected by two disulfide bonds to one of the strands in a b-sheet in the abDB motif (3). A third disulfide bond connects the b-sheet to an extended section in the amino-terminal portion of the toxin. These conserved secondary structural elements display considerable primary sequence variability (Table I) that affects the affinity of the toxins for the various K 1 channels. The K 1 channel toxins have been grouped into subfamilies based upon sequence similarities (Table I). Charybdotoxin (ChTX) is the most studied of the K 1 channel toxins. Eight residues (* residues in Table I) of ChTX were found to be critical for binding to the K 1 channel. A phenylalanine at position 2 and a tryptophan at position 14 characterize the ChTX family (subfamily 1). Noxiustoxin (NTX) was the first K 1 channel toxin isolated. The NTX family (subfamily 2) is char0003-9861/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved. STRUCTURE OF BUTANTOXIN acterized by the addition of a small residue (Ala or Gly) at position 27. The agitoxin (AgTX) family (subfamily 3) is characterized by a single point deletion after position 21. Tityus toxin Ka (TsTX Ka) is the only member of the subfamily 4 due to a lack of sequence homology with the other subfamilies. The leirutoxin (LeTX) family displays several point deletions and a shorter amino-terminal segment. Subfamilies 6 and 7 are composed of toxins primarily from the scorpion Pandinus imperator and have a small residue substituted for the conserved glycine residue at position 22. Subfamily 7 and the insectotoxins exhibit a fourth disulfide bond, which distinguishes them from the other toxin subfamilies. The new toxin butantoxin (BuTX) was originally isolated from the venom of the Brazilian scorpion Tityus serrulatus. Toxins were subsequently isolated from the venom of the scorpions T. bahiensis and T. stigmurus, which showed identical primary sequences to that of BuTX. This represents the first case of an identical toxin being isolated from more than a single species of scorpion. Previously, toxin sequences varied from species to species, even within a given genus (4). The four-disulfide bridging pattern of BuTX, as determined by limited proteolysis and sequencing of the fragments, showed a novel connectivity that included a C2-C5 disulfide bond (4). Other K 1 channel toxins (Table I) with 4 disulfide bonds include those from Pandinus imperator (Pi1, Pi4, and Pi7) (5), Heterometrus spinnifer (HsTX1) (6), and Scorpio maurus (MTX) (7). In all these toxins the fourth disulfide bond involves a cysteine residue positioned in the carboxyl-terminal part of the molecule. The insectotoxins (Lqh-8/6 (8), P1 BS (9), AmmP2 (10, 11), and I 5A (12)), and chlorotoxin (13) have a fourth disulfide bond involving a cysteine residue located in the amino-terminal part of the molecule. However, the C2–C5 bridging pattern is unique to BuTX and is not found in any other toxins. Last year, Novello and coworkers reported the isolation and primary sequence of tityus toxin IV (TsTX-IV) from Tityus serrulatus venom (14). TsTX-IV blocks the high conductance Ca 21-activated K 1 channels at nanomolar concentrations. TsTX-IV is a 41 amino acid protein which shows 97% sequence homology with BuTX. The only difference in the primary sequence is the additional residue (N41) at the carboxyl-terminus of TsTX-IV. The other 40 residues are identical to BuTX. The reported mass for TsTX-IV from mass spectrometry data is 4518. This is within 10 units of the mass for BuTX (M r 4508) and well below the calculated mass for a 41 residue protein (M r 4622). Therefore, TsTX-IV and 3 Abbreviations used: BuTX, butantoxin; DQF-COSY, double quantum filtered correlation spectroscopy; NOE, nuclear Overhauser effect; NOESY, NOE spectroscopy; TOCSY, total correlation spectroscopy; rmsd, root mean square deviation. 19 BuTX are likely the same toxin and represent a new subfamily of short-chain scorpion toxins (see Table I). The observed differences in the affinities might be due to differences in the targets used in the bioassays. Due to the unique nature of BuTX, nuclear magnetic resonance (NMR) solution studies were undertaken to help determine its structure. The present study reports the BuTX 1H-NMR assignments, the secondary structure determination, and molecular modeling calculation of the 3D structure. The structure obtained is compared to representative structures from the different subfamilies of short-chain scorpion toxins. This comparison suggests some sequence–structure relationships that may provide insight into the mechanism and action of toxin binding to the K 1 channel. EXPERIMENTAL PROCEDURES Purification and sample preparation. The isolation and purification of BuTX (4508 Da) was described elsewhere (4). Briefly, the isolation is accomplished with three column chromatography steps. First, dried whole venom was solubilized in 20 mM ammonium acetate (pH 4.7). The solution was resolved by gel filtration on a Sephadex G-50 column. Fraction-III (the third major component) was then further resolved by ion exchange chromatography on a Whatman CM-32 carboxymethylcellulose column equilibrated with 20 mM ammonium acetate and eluted with a 0 to 1 M sodium chloride gradient. The fraction III-6,7 (a large unresolved peak-pair of the sixth and seventh peaks) was then pooled and dialyzed briefly against distilled water. This fraction was further resolved on the same stationary phase (CM-32), equilibrated with 50 mM sodium phosphate (pH 6.0), and eluted with a 0 to 1 M NaCl gradient. The final peak was collected, pooled, and dialyzed against distilled water. The resulting pure toxin was then lyophilized. Molecular weight was determined by the matrix assisted laser desorption ionization time of flight (MALDI-TOF) method on a Voyager Elite mass spectrometer (Fig. 1). The NMR sample was prepared at ;1.0 mM and pH 5.0 with 50 mM sodium acetate-d 4 and 10% D 2O in a volume of 250 ml for use in a Shigemi microcell. After completion of the experiments in 90% H 2O, the sample was lyophilized to dryness and then brought up in 100% D 2O for the H-D exchange experiment and additional experiments in D 2O. MALDI-TOF mass spectrometry. Samples were analyzed in the positive mode on a Voyager Elite mass spectrometer with delayed extraction technology (PerSeptive Biosystems, Framingham, MA). The acceleration voltage was set at 25 kV, and 10 –50 laser shots were summed. Sinapinic acid (Aldrich, D13,460-0) dissolved in acetonitrile:0.1% TFA (1:1) was the matrix used. The mass spectrometer was calibrated with insulin. Samples were diluted 1:10 with matrix, and 1 ml was pipetted onto a smooth plate. NMR spectroscopy. The NMR experiments were performed on a Bruker AM-600 spectrometer equipped with an Aspect 3000 computer. Data were collected at 283, 293, and 313 K. The experimental data were processed using Felix 97.2. Chemical shifts were referenced to the water signal at 4.84 ppm (273 K) and at 4.63 ppm (313 K). NOESY (18), DQF-COSY (19), PH-COSY (20), and TOCSY (21) measurements were performed in pure absorption mode using time proportional phase increment and presaturation for water suppression. The 2D 1H NMR spectra were recorded with 256 scans and 512 t 1 points. Zero-filling was applied prior to Fourier transformation, and data were processed with shifted sine bell window functions in both dimensions. Mixing times of 100, 200, and 400 ms for NOESY in H 2O, and 200 and 400 ms in D 2O were employed. Additional 400 ms NOESY data in H 2O were collected using a jump-return read se- 20 HOLADAY ET AL. TABLE I Amino Acid Sequences a of Selected Neurotoxins b Toxin Sequence ChTX IbTX LQ2 C11TX-I C11TX-II MgTX NTX NTX2 KTX-1 KTX-2 AgTX-1 AgTX-2 AgTX-3 TsTX Ka PO5 LeTXI PiTX Ka Pi3 Pi1 Pi4 Pi7 MTX HsTX1 BuTX TsTX-IV LQH-8/6 C1TX Cons. 1 Cons. 2 Cons. 3 Cons. 4 * * * * ** * * ---QFTNVSCTTS-----KECWSVCQRLHNTSRG-KCMNKKCRCYS----QFTDVDCSVS-----KECWSVCKDLFGVDRG-KCMGKKCRCYQ----QFTQESCTAS-----NQCWSICKRLHNTNRG-KCMNKKCRCYS----ITINVKCTSP-----QQCLRPCKDRFGQHAGGKCINGKCKCYP----TVIDVKCTSP-----KQCLPPCKEIYGRHAGAKCMNGKCKC------TIINVKCTSP-----KQCLPPCKAQFGQSAGAKCMNGKCKCYPH---TIINVKCTSP-----KQCSKPCKELYGSSAGAKCMNGKCKCYNN---TIINEKCFAT-----SQCWTPCKKAIGSLQS-KCMNGKCKCYNG--GVEINVKCSGS-----PQCLKPCKDA-GMRFG-KCMNRKCHCTPK---VRIPVSCSKH-----SGCLKPCKDA-GMRFG-KCMNGKCDCTPK--GVPINVKCTGS-----PQCLKPCKDA-GMRFG-KCINGKCHCTPK--GVPINVSCTGS-----PQCIKPCKDA-GMRFG-KCMNRKCHCTPK--GVPINVPCTGS-----PQCIKPCKDA-GMRFG-KCMNRKCHCTPK---VFINAKCRGS-----PECLPKCKEAIGKAAG-KCMNGKCKCYP--------TVCN-L-----RRCQLSCRSL-G-LLG-KCIGVKCECVKH-------AFCN-L-----RMCQLSCRSL-G-LLG-KCIGDKCECVKH------TISCTNP-----KQCYPHCKKETGYPNA-KCMNRKCKCFGR------TISCTNE-----KQCYPHCKKETGYPNA-KCMNRKCKCFGR------LVKCRGT-----SDCGRPCQQQTGCPNS-KCINRMCKCYGC----IEAIRCGGG-----RDCYRPCQKRTGCPNA-KCINRTCKCYGCS ----DEAIRCTGT-----KDCYIPCRYITGCFNS-RCINKSCKCYGCT -------VSCTGS-----KDCYAPCRKQTGCPNA-KCINKSCKCYGC-------ASCRTP-----KDCADPCRKETGCPYG-KCMNRKCKCNRCWCSTCLDLACGAS-----RECYDPCFKAFGRAHG-KCMNNKCRCYT-WCSTCLDLACGAS-----RECYDPCFKAFGRAHG-KCMNNKCRCYTN-----RCSPCFTTDQQMTKKCYDCCGGK----GKGKCYGPQCICAPY-----MCMPCFTTDHQMARKCDDCCGGK----GRGKCYGPQCLCR-----------C----------C---C----------KC----C-C------------C----------C---C-----C----KC----C-C--C------C--C----------C--CC----------KC----C-C----C--C----C----------C---C----------KC----C-C---- Subfamily c 1 2 3 4 5 6 7 8(?) Insect All 7 Insect BuTX Comparative alignment of selected short-chain scorpion neurotoxins representing the different subfamilies (15, 16). Sequences were obtained from the SWISS-PROT database and aligned with Clustal W (17). Sequences were aligned with the six conserved cysteines and the critical lysine (Cons. 1). Gaps were introduced to improve overlap. Additional consensus sequences are listed showing the cysteines involved in a fourth disulfide bridge found in the latter three subfamilies. The eight most important residues involved in charybdotoxin binding to the potassium channel are indicated (*). b This table lists only some representative toxins with in each family. The Brookhaven Protein Data Bank has at least 26 structures for scorpion toxins, not all of which have been included in this table. c The current classification scheme is derived from that initiated by Miller (15) and extended by Possani et al. (48), and more recently by Savarin et al. (49) and Tygat et al. (50). As scorpion toxin structures for those toxins acting at K 1-voltage gated channels become more diverse, the discrimination into clear-cut subgroups becomes less absolute. The use of this scheme preserves the historical lineage in the literature and the relationships among the original subgroups until its use is no longer useful. a quence. For assignment purposes, 52 ms TOCSY experiments were performed in both H 2O and D 2O. The H-D exchange experiment involved a series of seven 1D spectra collected at 5-min intervals and a 54 ms TOCSY with 32 scans and 256 t 1 points. These experiments were run within 5 min of dissolving the protein in D 2O. Molecular modeling. The structure refinement calculations were performed with the XPLOR/QUANTA/CHARMm package (Molecular Simulations, Inc., San Diego, CA) on an Origin 200 (Silicon Graphics, Inc., Mountain View, CA). The solution–state structures were generated using a hybrid method consisting of distance geometry combined with dynamic simulated annealing (22) with some minor modifications as previously described (3). Distance geometry substructures were first generated using only medium and long range (residues i, i 1 2 or greater) distance constraints and torsion angle constraints. Dynamical simulated annealing followed by energy minimization was then performed using a 4-step protocol (3, 23). This protocol was first run without hydrogen bond constraints. From the prefinal structures obtained, the H-bond constraints were determined by identifying acceptors for the slowly exchanging amide protons. The protocol was repeated with these additional constraints to produce the final structures. From a total of 100 starting structures, a family of 49 structures with nOe distance constraint violations less than 0.2 Å and dihedral angle constraint violations less than 2°, designated as the simulated annealing family (^SA& k), were considered acceptable and included in the statistical analysis (Table III). The simulated annealing average structure (^SA& k) was obtained by taking the averages of the atomic coordinates of these acceptable structures. The final energy minimized average struc- STRUCTURE OF BUTANTOXIN 21 FIG. 1. MALDI-TOF mass spectral analysis of Butantoxin. Mass spectrum showing purified butantoxin (4508 Da) and insulin (5734 Da) run as an internal standard. ture (^SA& kr ) was obtained from ^SA& k by energy minimization first with and then without NMR-derived constraints to remove bond length and bond angle distortions. the D18 –P19 peptide bond is in the trans conformation. The chemical shifts of all protons of BuTX are listed in the Table II. RESULTS Experimental Constraints Sequence-Specific Assignments The distance constraints were grouped as strong (1.8 –2.7 Å), medium (1.8 –3.3 Å), or weak (1.8 –5.0 Å) based upon the peak intensities from the 200 ms NOESY experiment in H 2O at 293 K. Additional strong and weak constraints were taken from the 100 ms NOESY in H 2O and the 400 ms NOESY in H 2O and D 2O, respectively. A total of 353 NOESY contacts were observed: 183 intraresidue, 95 sequential (ui 2 ju 5 1), 25 medium range (2 # ui 2 ju # 4), and 50 long range The assignments according to amino acid type (24) were made through analysis and comparison of TOCSY and COSY data in D 2O and H 2O, and NOESY data in D 2O. The sequential resonance assignments were made using standard methodology (24) based on d aN(i, i 1 1), d NN(i, i 1 1), and d bN(i, i 1 1) NOESY contacts (Fig. 2). The sequential nOe connectivity of the D18 amide proton with the P19 delta protons indicates that FIG. 2. Summary of NMR data for butantoxin. NOESY contacts are shown as filled bars. The strength of sequential NOESY contacts is indicated by the relative size of the bars. Medium-range NOESY contacts are shown as continuous lines. Bars involving proline represent d Nd (i, i 11) and d bd (i, i 1 1) contacts. Slowly exchanging amide protons filled circles; vicinal coupling constants, 3 J Hna . 8 Hz filled squares, 3 J Hna , 7 Hz open squares, and 3 J Hna 5 7– 8 Hz X; torsional angels: x 1 5 180° filled upward triangles, x 1 5 260° filled downward triangles, and x 1 5 160° open triangles. Regions of secondary structure are shown alpha helix (a), beta sheet (b), type I9 turn (t), and a beta-like region (b9). 22 HOLADAY ET AL. TABLE II Resonance Assignments for Butantoxin Residue HN Ha Hba, Hbb W1 C2 S3 T4 C5 L6 D7 L8 A9 C10 G11 Other 2H 7.35; 4H 7.58; 5H 7.13; 6H 7.23; 7H 7.51; NH 10.25; 8.83 7.33 7.32 8.40 8.76 7.33 8.40 8.72 9.55 7.82 4.55 4.10 4.08 4.85 4.65 4.65 4.38 4.92 4.80 4.28, 3.72 4.80 3.31,3.15 3.80,3.95 3.95 3.05,3.18 1.48,1.63 2.72,2.25 1.32,1.35 1.40 2.83,2.38 2.2, 2.26 2.71,3.18 3.18,3.10 2.85,2.81 2.28,2.08 3.23,3.32 3.36,3.02 1.80 0.77 3.42,2.90 A12 S13 R14 E15 C16 Y17 D18 P19 C20 F21 K22 7.94 8.60 8.55 4.08 4.68 4.13 4.82 4.23 4.75 3.98 3.72 A23 F24 7.75 8.08 3.97 4.92 G25 8.01 R26 A27 H28 G29 7.02 9.05 7.73 7.50 K30 C31 M32 N33 N34 K35 C36 R37 C38 Y39 T40 9.30 7.91 8.97 9.53 8.78 7.75 7.90 9.30 8.35 8.05 8.15 4.02, 3.28 4.71 4.68 4.53 4.72, 3.95 4.75 4.90 4.80 4.28 4.38 5.40 5.05 4.82 5.32 4.60 4.25 7.40 8.08 7.32 9.25 g 0.60 g 1.75; d 0.75, 0.83 g 1.68, d 0.92, 0.96 1.50 1.85,1.68 1.15 2.87,3.00 1.75 2.98,2.92 2.07,1.75 3.00,2.85 3.00,3.32 2.00,1.73 2.62,2.70 1.83,1.78 3.45,2.55 2.63,3.14 4.05 g 2.28, 2.46 2,6H 7.32; 3,5H 7.00 g 1.62; d 3.68, 3.58 g 1.45; d 1.60; e 3.08; zNH 3 6.98 2,6H 7.20; 3,5H 7.38; 4H 7.30 g 1.54 2H 7.00; 4H 8.55 g g g g 2.42 NH 7.63, 6.95 NH 7.68, 7.00 1.52, 1.35; d 1.70; e 3.00 g 3.05 2,6H 6.98; 3,5H 6.65 g 1.05 Note. Underlined shifts correspond to Hb3 in IUB-IUPAC convention. (ui 2 ju $ 5). A total of 12 distance constraints were introduced corresponding to the four disulfide bonds. Pseudo atoms were used for the non-stereospecifically assigned atoms (25). Phi angle constraints, 260° 6 30° for 3 J HNa , 7 Hz and 2120° 6 40° for 3 J HNa . 8 Hz, were obtained from the PH-COSY in H 2O. For residues not giving rise to COSY peaks, phi angle constraints were calculated from NOESY and TOCSY data (26). A phi angle constraint of 295° 6 65° was used for residues with 3 J HNa 5 7– 8 Hz, or when 3 J HNa could not be determined. This constraint limits the phi angle to the allowed regions of the Ramachandran plot. A total of 19 stereospecific assignments for the spin systems were made using 3 J abb, coupling constants and the 200 ms NOESY data. From this data, Chi constraints were given as the optimum angle (180°, 60°, or 260°) 6 30°. A total of 37 phi and 19 chi constraints were obtained. An omega angle constraint of 180° 6 5° was used to keep P19 in the trans conformation. Figure 2 shows a summary of sequential and medium range NOESY contacts and the regions of secondary structure identified from these contacts. The H/D exchange experiments identified 17 amide peaks slowly exchanging with solvent. These data were used to identify hydrogen bonding within the structure when appropriate hydrogen bond acceptors could be identified in the prefinal structures. The backbone NH group of K22 showed equal probability of hydrogen bonding to the carbonyl of D18 or P19. Both constraints were used. Each hydrogen bond had two constraints, yielding a total of 36 constraints. Secondary Structure Elements The establishment of the secondary structural elements was made using the unique NOESY contacts, coupling constants, and slowly exchanging amide protons characteristic of these structures (24). Two regions of secondary structure were observed: a 221 turn a-helix (residues 15–23) and a two-strand b-sheet (residues 29 to 32, 35 to 38). A type I9 b-turn spanning residues 32 to 35 was identified on the basis of coupling constants and NOESY contacts. A potential third strand of the b-sheet was identified in residues 5 to 9. This string of residues showed large 3J HNa couplings characteristic of a b-strand. Residues 6 and 8 showed side-chain nOe contacts to residues in the b-sheet, and both had slowly exchanging amides. However, backbone nOes (HA i–HA j and HA i–HN j) across the strand were only observed for residue A9 to K35. This lack of nOe contacts prevents the conclusive assignment of residues 5 to 9 as part of the b-sheet. Similar pseudo third b-strand of this type has been observed in the NMR structures of the toxins Lq2 (27), Pi1 (28), AgTX2 (29), and KTX (30). Modeling Calculations A total of 458 constraints (353 NOESY contacts, the 57 torsional angles, 36 constraints for the hydrogen bonds, and 12 constraints for the 4 disulfide linkages) were used in the calculations (see Fig. 3 bottom panel). The average number of constraints per residue was 11.5. From the initial 100 distance geometry struc- 23 STRUCTURE OF BUTANTOXIN 34 that have positive phi values characteristic of a type I9 b-turn. DISCUSSION General Features of the Butantoxin Solution Structure Figure 4 shows 35 representative ^SA& k backbone structures resulting from the simulated annealing calculations. The protein is compact with a well-defined core consisting of an a-helix and b-sheet connected by two disulfide bridges. The a-helix stretches from E15 to A23. Residues A12 to R14 appear to form a turn of 3 10 helix. The a-helix is slightly bent due to the presence of a proline residue at position 19. This proline residue is conserved in most of the short-chain scorpion toxin subfamilies. Residues C5 to A9 form a pseudo b-strand interrupted by a bulge at D7, which may explain the lack of cross-strand nOes for this segment. Residues A9 and C10 are in close contact with residues N34 and K35 in the tight turn. The overall structure is similar to other short-chain scorpion toxins and is well defined. The rmsd for the 49 backbone structures with respect to the average structure is 0.79 Å, and it is 0.55 Å when the residues W1, C2, and T40, which are disordered, are excluded. Some secondary structural regions are very well defined, e.g., the a-helix (rmsd 5 0.18 Å) and the b-sheet (rmsd 5 0.26 Å). In general, the rmsd for the backbone and side-chain atoms show inverse correlation with the number of constraints (Fig. 3). The two positive phi values for the nonglycine residues arise from the i 1 1 and i 1 2 positions of the type I9 turn. Type I9 turns are FIG. 3. Constraints and atomic rmsd per residue. Distribution of experimental constraints per residue (bottom); distribution of atomic rmsd per residue of the ^SA& k structures with respect to the ^SA& kr structure for the backbone atoms (middle) and for the side-chain atoms (top). tures, the simulated annealing protocol produced 49 acceptable structures (with no dihedral violations . 2° and no NOESY violations . 0.2 Å). The structural statistics for the ^SA& k family and the energy minimized average structure, ^SA& kr, are shown in Table III. These data show well-converged structures with good stereochemistry. This family of ^SA& k structures had rms deviations of 0.79 and 1.41 Å for the backbone and all heavy atoms, respectively, relative to the average structure, ^SA& k. The distributions of the average atomic rms deviations by residue for the backbone and side-chain atoms are shown in Fig. 3. The Ramachandran plot of 35 representative ^SA& k structures (not shown) shows that the phi and psi angles occupy the allowed regions, with the exception of Asn 33 and Asn TABLE III Structural Statistics for Butantoxin Quantity Rms deviations from experimental distance constraints (Å) All (389) Sequential (ui 2 ju 5 1) (95) Medium-range (1 , ui 2 ju , 5) (25) Long-range (ui 2 ju . 4) (50) Intra-residue (183) H-bonds (36) Rms deviations from experimental dihedral angle constraints (deg.) Energies (kcal mol 21) E (NOE) E (tor) E (repel) Deviations from idealized covalent geometry Bonds (Å) Angles (deg.) Impropers (deg.) ^SA& k ^SA& kr 0.009 0.012 0.011 0.010 0.003 0.015 0.031 0.021 0.062 0.014 0.032 0.084 0.103 0.569 1.811 0.043 3.256 0.373 0.001 0.071 0.0016 0.511 0.361 0.0010 0.466 0.324 24 HOLADAY ET AL. structure to which all others are compared. Alignment of the BuTX residues 5– 40 and ChTX residues 2–37 (Fig. 6) yielded a backbone rmsd of 1.61 Å. Six of the eight residues critical for ChTX interaction with the K 1 channel (see Table I) are conserved in BuTX. The other two (ChTX W14 and R25) show conservative replacement in BuTX (Y17 and H28). BuTX displays a greater amount of twist in the b-sheet and a slightly different turn between the a-helix and the b-sheet. This turn causes a slight mismatch of the b-sheets when the a-helices are matched. Most of the side chains have a similar spatial orientation with the following exceptions. The W1 residue of BuTX is oriented in a similar orientation as F2 of ChTX despite being four residues out of register, restoring the aromaticity of this position. However, the W1 ring is located ;5.5 Å farther away from the face of the molecule and may play a role in BuTX’s reduced affinity for the Shaker B K 1 channel. FIG. 4. Two stereoviews of the best fit superposition of 35 ^SA& k structures. somewhat rare, but they are preferentially found in b-hairpins. Presumably this is because they fit the twist of the b-sheet, and cross-strand hydrogen bonding stabilizes the unfavorable torsional angles (31, 32). While the backbone is well defined in these structures, two of the disulfide bonds (C10 –C31 and C16 – C36) show conformational disorder. This disorder is conistent with a lack of specific HB i–HB j nOe contacts across the bond, as reported previously for BPTI (33). Comparison with Other K 1 Channel Toxins The final energy minimized average structure, ^SA& kr, of BuTX (Fig. 5) was compared with several members of the short-chain scorpion toxin family. Using information from the sequence alignment in Table I, the following toxin NMR structures were matched pairwise with BuTX in QUANTA: charybdotoxin (2CRD) (34), noxioustoxin (1SXM) (35), kaliotoxin (1KTX) (36), PO5-NH 2 (1PNH) (37), P. imperator toxin K-a (2PTA) (38), maurotoxin (1TXM) (39), Lqh-8/6 (8), and chlorotoxin (1CHL) (40). Coordinates were taken from the Protein Data Bank. Key structural differences, which may play a role in toxin binding affinity, will be summarized. Being the most studied member of the short-chain K 1 channel toxins, charybdotoxin is generally the FIG. 5. Ribbon diagram of Butantoxin ^SA& kr structure. The disufide bridges and the aromatic side chains are shown. 25 STRUCTURE OF BUTANTOXIN FIG. 6. Comparison of backbone (Ca trace) of Butantoxin to Charybdotoxin and to Noxiustoxin. A thick line indicates BuTX. Thin lines represent ChTX and NXT in the figures. Noxiustoxin was the first K 1 channel toxin isolated (41). NTX has a single residue insertion at the end of the a-helix (S23) and an extra carboxyl-terminal residue (N39). BuTX residues 4 –25 and 26 – 40 were aligned with NTX residues 1–22 and 24 –38 separately (Fig. 6). The first noticeable difference was a lack of aromatic residues on the surface of NTX (BuTX residues W1, Y17, F21, and H28). NTX also displays a slightly smaller twist of the b-sheet and a different ab loop region. Kaliotoxin has a shorter a-helix by one turn, causing noticeable differences in the length of the ab loop region. The amino-terminal segment makes contact with the opposite side of the b-sheet for KTX, making direct comparison difficult. This b-sheet is highly pleated and twists in the opposite direction. PO5-NH 2 is 10 residues shorter than BuTX at its amino-terminus. PO5-NH 2 lacks aromatic residues in the helical region (BuTX Y17 and F21; PO5-NH 2 Q9 and R13). Lysine 22 in BuTX has no equivalent charged residue in PO5-NH 2. The b-sheet from both toxins displays a very similar twist. The pdb file for Pandinus imperator toxin Ka is numbered according to ChTX and that numbering will be used here. The backbone of PiTX-Ka is very similar to BuTX, but displays a lesser degree of twist in the b-sheet. Six nonconservative amino acid substitutions were noted between the two structures, located primarily on the last 121 turns of the a-helix and the ab loop region. Most notably, BuTX F21 and R26 become K18 and Y23 in PiTX-Ka. Maurotoxin is a four-disulfide bridged neurotoxin. MTX shows relatively good agreement with the overall fold of BuTX (rmsd 2.0 Å for mainchain residues) and even better agreement for the helix residues (rmsd 0.54 Å). The ab loop region is considerably different, resulting in the b-sheets being oriented nearly 3 Å apart when the helices are aligned. The conserved tyrosine residue in the a-helix of both toxins shows different x1 values (MTX Y10 x1 5 298, BuTX Y17 x1 5 174). MTX has an unusual disulfide bridging pattern with respect to the other K 1 channel toxins (C13–C19 and C31– C34). It is interesting to note that rearrangement of these linkages would restore the conserved disulfide bridging pattern observed in other toxins. Lqh-8/6 and chlorotoxin are four-disulfide bridged neurotoxins active against insect potassium and chloride channels, respectively. Alignment of the shortened helix in Lqh-8/6 and ClTX with BuTX yields backbone rmsd values of 1.04 and 1.24 Å. These toxins have a 4-residue loop insertion just before the a-helix. The fourth disulfide bridge in Lqh-8/6 and ClTX connects the amino-terminus tightly to the a-helix. An alternative sequence alignment (Table IV) of BuTX with Lqh8/6 and ClTX shows that they share a X-C-X-X-C amino-terminus motif and that only a single cysteine is out of alignment (BuTX C10, Lqh-8/6, and ClTX C19). This may represent a link between the K 1 channel toxins and the insect or chloride channel toxins. Several members of the charybdotoxin family of K 1 channel toxins have a conserved tyrosine residue at the carboxyl-terminus that is critical for interaction with the K 1 channel. In ChTX, NTX, PiTX-Ka, and MTX, the x1 value ranges from 268° to 2122°, while this value is 171° for BuTX. This conformational difference in BuTX results from unfavorable steric interactions in the g 2 t 3 (180°) and t 2g 3 (260°) conformations. In the g 2 t 3 conformation, the tyrosine ring and hydroxyl group interact with the carboxyl-terminal CO 2 group and the carbonyl of T40. In the t 2 g 3 conformation there are unfavorable interactions with the backbone carbonyl of T4 and the side-chain oxygen atoms of S3 and TABLE IV Alternative Sequence Alignment Toxin Sequence BuTX LQH-8/6 C1TX Cons. WCSTCLDLACGASRECYDPCFKAFGRAHGKCMNNKCRCYTRCSPCFTTDQQMTKKCYDCCG---GKGKGKCYGPQCICAPY MCMPCFTTDHQMARKCDDCCG---GKGRGKCYGPQCLCR--C--C----------C---C---------KC----C-C--- 26 HOLADAY ET AL. T4. The g 2g 3 (160°) conformation removes all steric interactions and presents the possible hydrogen bonding of T39 ring oxygen to the H28 ring nitrogens. This conformational change may contribute to BuTX’s weakened affinity for the Shaker B K 1 channel. CONCLUSIONS The NMR solution-state structure of a new toxin, BuTX, which is present in the venoms from Tityus serrulatus, T. bahiensis, and T. stigmurus is presented. This structure was calculated using distance geometry and simulated annealing refinement from constraints derived from 1H NMR data. The structures are well converged and display the conserved abDB motif found in all scorpion toxins. The novel feature of this toxin, an amino-terminal disulfide bridge, presents interesting questions in terms of the structural stability and receptor specificity. This bridge does not appear to confer any structural stability as observed for the other disulfide bonds. However, it may play a role in receptor specificity. In 1998, MacKinnon et al. solved the structure of the KcsA potassium channel (42) and studied its interaction with the scorpion toxin AgTX2 (43, 44). This study utilized mutant cycle analysis of the Shaker K 1 channel-AgTX2 interaction (45, 46) to identify residues critical to the interaction (G10, R24, F25, and K27). In BuTX, the first two corresponding residues are off by only one position in the sequence. The F25 is substituted by H28, and the critical lysine residue is conserved in all the K 1 channel toxins. Thus it may be likely that these residues in BuTX might be similarly important. Our NMR structure thus provides a basis for undertaking a detailed analysis of Shaker K 1-channel-BuTX interactions and for further unraveling of the molecular mechanism of action of these K 1-channels. ACKNOWLEDGMENT We appreciate the assistance of Dr. Mike Jablonsky with the NMR measurements. REFERENCES 1. Martin-Euclaire, M. F., and Couranud, F. (1995) in Handbook of Neurotoxicology (Chang, L. W., and Dyer, R. S., Eds.), pp. 683– 716, Marcel Dekker, New York. 2. Gordon, D., Savarin, P., Gurevitz, M., and Zinnjustin, S. (1998) J. Toxicol.-Toxin Rev. 17, 131–159. 3. Lee, W., Moore, C. H., Watt, D. D., and Krishna, N. R. (1994) Eur. J. Biochem. 219, 89 –95. 4. Fletcher, P. L., Jr., Fletcher, M. D., Bartolotti, L., OlamendiPortugal, T., Possani, L. D., Gomez-Lagunas, F., Mannie, M., and Martin, B. M. (1999) submitted. 5. Olamendi-Portugal, T., Gomez-Lagunas, F., Gurrola, G. B., and Possani, L. D. (1996) Biochem. J. 315, 977–981. 6. Lebrun, B., Romi-Lebrun, R., Martin-Eauclaire, M. F., Yasuda, A., Ishiguro, M., Oyama, Y., Pongs, O., and Nakajima, T. (1997) Biochem. J. 328, 321–327. 7. Rochat, H., Kharrat, R., Sabatier, J. M., Mansuelle, P., Crest, M., Martin-Eauclaire, M. F., Sampieri, F., Oughideni, R., Mabrouk, K., Jacquet, G., Van Rietschoten, J., and El Ayeb, M. (1998) Toxicon 36, 1609 –1611. 8. Adjadj, E., Naudat, V., Quiniou, E., Wouters, D., Sautiere, P., and Craescu, C. T. (1997) Eur. J. Biochem. 246, 218 –227. 9. Fazal, A., Beg, O. U., Shafqat, J., Zaidi, Z. H., and Jornvall, H. (1989) FEBS Lett. 257, 260 –262. 10. Zlotkin, E., Miranda, F., and Rochat, H. (1978) in Arthropod Venoms (Bettini, S., Ed.), pp. 317–369, Springer-Verlag, Berlin. 11. Rochat, H., Bernard, P., and Couraud, F. (1979) Adv. Cytopharmacol. 3, 325–334. 12. Grishin, E. V., Soldatov, N. M., Tashmuchamedov, B. A., and Atakuviev, B. U. (1978) Bioorg. Khim. 4, 450 – 461. 13. DeBin, J. A., Maggio, J. E., and Strichartz, G. R. (1993) Am. J. Physiol. 264, C361–9. 14. Novello, J. C., Arantes, E. C., Varanda, W. A., Oliveira, B., Giglio, J. R., and Sergio, M. (1999) Toxicon 37, 651– 660. 15. Miller, C. (1995) Neuron 15, 5–10. 16. Delepierre, M., Prochnicka-Chalufour, A., and Possani, L. D. (1998) Toxicon 36, 1599 –1608. 17. Thompson, J. D., Higgins, D. G., and Gibson, T. J. (1994) Nucleic Acids Res. 22, 4673– 4680. 18. Jeener, J., Meier, B. H., Bachmann, P., and Ernst, R. R. (1979) J. Chem. Phys. 72, 4546 – 4553. 19. Rance, M., Sørensen, O. W., Bodenhausen, G., Ernst, R. R., and Wüthrich, K. (1983) Biochem. Biophys. Res. Commun. 117, 479 – 485. 20. Marion, D., and Wüthrich, K. (1983) Biochem. Biophys. Res. Commun. 113, 967–974. 21. Braunschweiler, L., and Ernst, R. R. (1983) J. Magn. Reson. 53, 521–528. 22. Driscoll, P. C., Gronenborn, A. M., Beress, L., and Clore, G. M. (1989) Biochemistry 28, 2188 –2198. 23. Lee, W., Jablonsky, M. J., Watt, D. D., and Krishna, N. R. (1994) Biochemistry 33, 2468 –2475. 24. Wüthrich, K. (1986) NMR of Proteins and Nucleic Acids, Wiley, New York. 25. Wüthrich, K., Billeter, M., and Braun, W., (1983) J. Mol. Biol. 169, 949 –961. 26. Wang, Y., Nip, A. M., and Wishart, D. S. (1997) J. Biomolec. NMR 10, 373–382. 27. Renisio, J. G., Lu, Z., Blanc, E., Jin, W., Lewis, J. H., Bornet, O., and Darbon, H. (1999) Proteins 34, 417– 426. 28. Delepierre, M., Prochnicka-Chalufour, A., and Possani, L. D. (1997) Biochemistry 36, 2649 –2658. 29. Krezel, A. M., Kasibhatla, C., Hidalgo, P., MacKinnon, R., and Wagner, G. (1995) Protein Sci. 4, 1478 –1489. 30. Gairi, M., Romi, R., Fernandez, I., Rochat, H., Martin-Eauclaire, M. F., Van Rietschoten, J., Pons, M., and Giralt, E. (1997) J. Pept. Sci. 3, 314 –319. 31. Wilmot, C. M., and Thornton, J. M. (1988) J. Mol. Biol. 203, 221–232. 32. Sibanda, B. L., and Thornton, J. M. (1985) Nature 316, 161–171. 33. Berndt, K. D., Guntert, P., Orbons, L. P. M., and Wüthrich, K. (1992) J. Mol. Biol. 227, 757–775. 34. Bontems, F., Roumestand, C., Gilquin, B., Menez, A., and Toma, F. (1991) Science 254, 1521–1523. STRUCTURE OF BUTANTOXIN 35. Dauplais, M., Gilquin, B., Possani, L. D., Gurrola-Briones, G., Roumestand, C., and Menez, A. (1995) Biochemistry 34, 16563– 16573. 36. Fernandez, I., Romi, R., Szendeffy, S., Martin-Eauclaire, M. F., Rochat, H., Van Rietschoten, J., Pons, M., and Giralt, E. (1994) Biochemistry 33, 14256 –14263. 37. Meunier, S., Bernassau, J. M., Sabatier, J. M., Martin-Eauclaire, M. F., Van Rietschoten, J., Cambillau, C., and Darbon, H. (1993) Biochemistry 32, 11969 –11976. 38. Tenenholz, T. C., Rogowski, R. S., Collins, J. H., Blaustein, M. P., and Weber, D. J. (1997) Biochemistry 36, 2763–2771. 39. Blanc, E., Sabatier, J. M., Kharrat, R., Meunier, S., el Ayeb, M., Van Rietschoten, J., and Darbon, H. (1997) Proteins 29, 321–333. 40. Lippens, G., Najib, J., Wodak, S. J., and Tartar, A. (1995) Biochemistry 34, 13–21. 41. Possani, L. D., Martin, B. M., and Svendsen, I. (1982) Carlsberg Res. Commun. 47, 285–289. 42. Doyle, D. A., Cabral, J. M., Pfuetzner, R. A., Kuo, A., Gulbis, J. M., Cohen, S. L., Chait, B. T., and MacKinnon, R. (1998) Science 280, 69 –77. 27 43. Garcia, M. L., Garcia-Calvo, M., Hidalgo, P., Lee, A., and MacKinnon, R. (1994) Biochemistry 33, 6834 – 6839. 44. MacKinnon, R., Cohen, S. L., Kuo, A., Lee, A., and Chait, B. T. (1998) Science 280, 106 –109. 45. Hidalgo, P., and MacKinnon, R. (1995) Science 268, 307–310. 46. Ranganathan, R., Lewis, J. H., and MacKinnon, R. (1996) Neuron 16, 131–139. 47. Legros, C., Feyfant, E., Sampieri, F., Rochat, H., Bougis, P. E., and Martin-Eauclaire, M. F. (1997) FEBS Lett. 417, 123–129. 48. Possani, L. D., Selisko, B., and Gurrola, G. B. (1999) in “Perspectives in Drug Discovery and Design” Vol. 15/16, pp. 15– 40, KLUWER/Academic Press. 49. Savarin, P., Romi-Lebrun, R., Zinn-Justin, S., Lebrun, B., Nakajima, T., Gilquin, B., and Menez, A. (1999) Protein Sci. 8, 2672–2685. 50. Tygat, J., Chandy, K. G., Garcia, M. L., Gutman, G. A., MartinEauclaire, M. F., van der Walt, J. J., and Possani, L. D. (1999) Trends Pharmacol. Sci. 20, 444 – 447.