crystal structure of a highly acidic neurotoxin from scorpion buthus tamulus at 2.2 Ǻ resolution...

Journal of

www.elsevier.com/locate/yjsbi

Journal of Structural Biology 155 (2006) 52–62

StructuralBiology

Crystal structure of a highly acidic neurotoxin from scorpion

Buthus tamulus at 2.2 �A0

resolution reveals novel structural features

Madhu Sharma, A.S. Ethayathulla, Talat Jabeen, Nagendra Singh, K. Sarvanan,Savita Yadav, Sujata Sharma, A. Srinivasan, Tej P. Singh *

Department of Biophysics, All India Institute of Medical Sciences, New Delhi 110029, India

Received 25 October 2005; received in revised form 2 December 2005; accepted 6 December 2005Available online 9 January 2006

Abstract

The crystal structure of a highly acidic neurotoxin from the scorpion Buthus tamulus has been determined at 2.2 �A0

resolution. Theamino acid sequence determination shows that the polypeptide chain has 64 amino acid residues. The pI measurement gave a valueof 4.3 which is one of the lowest pI values reported so far for a scorpion toxin. As observed in other a-toxins, it contains four disulphidebridges, Cys12–Cys63, Cys16–Cys36, Cys22–Cys46, and Cys26–Cys48. The crystal structure reveals the presence of two crystallograph-ically independent molecules in the asymmetric unit. The conformations of two molecules are identical with an r.m.s. value of 0.3 A fortheir Ca tracings. The overall fold of the toxin is very similar to other scorpion a-toxins. It is a babb protein. The b-sheet involvesresidues Glu2–Ile6 (strand b1), Asp32–Trp39 (strand b3) and Val45–Val55 (strand b4). The single a-helix formed is by residuesAsn19–Asp28 (a2). The structure shows a trans peptide bond between residues 9 and 10 in the five-membered reverse turn Asp8–Cys12. This suggests that this toxin belongs to classical a-toxin subfamily. The surface features of the present toxin are highly charac-teristic, the first (A-site) has residues, Phe18, Trp38 and Trp39 that protrude outwardly presumably to interact with its receptor. There isanother novel face (N-site) of this neurotoxin that contains several negatively charged residues such as, Glu2, Asp3, Asp32, Glu49 andAsp50 which are clustered in a small region of the toxin structure. On yet another face (P-site) in a triangular arrangement, with respectto the above two faces there are several positively charged residues, Arg58, Lys62 and Arg64 that also protrude outwardly for a poten-tially potent interaction with other molecules. This toxin with three strong features appears to be one of the most toxic molecules report-ed so far. In this sense, it may be a new subclass of neurotoxins with the largest number of hot spots.� 2005 Elsevier Inc. All rights reserved.

Keywords: Binding site; Crystal structure; Sequence determination; Sodium channel; Toxin

1. Introduction

Scorpion venoms contain a large number of neurotoxins.These toxins are single chain polypeptides composed of60–70 amino acids with four disulphide bridges. Theyessentially belong to two broad classes identified asa- and b-toxins according to their physiological effects andbinding properties (Couraud et al., 1982; Gordon et al.,1998; Jover et al., 1980; Wheeler et al., 1983). These toxinsare known to interact specifically with Na+ channels. The

1047-8477/$ - see front matter � 2005 Elsevier Inc. All rights reserved.

doi:10.1016/j.jsb.2005.12.005

* Corresponding author. Fax: +91 11 2658 8663.E-mail address: [email protected] (T.P. Singh).

a-toxins prolong the action potential by slowing theinactivation of Na+ flow with no direct effect on activationwhereas b-toxins shift the activation voltage to morenegative potentials and bind to the different site on the recep-tor. According to their binding specificities, thea-toxins have been further divided into three subgroups,classical a, a-like and insect a-toxins (Gordon et al., 1996).The classical a-toxins primarily act against mammal targetswhereas insect a-toxins are toxic to insects. Yet another sub-group known as a-like toxins can act against both mammalsand insects (Gordon et al., 1996). These variations in thebinding specificities of toxins imply heterogeneity in thefunctional surfaces as well as their receptor sites (Catterall,

mailto:[email protected]

M. Sharma et al. / Journal of Structural Biology 155 (2006) 52–62 53

1980; Garcia et al., 1997; Gordon et al., 1998; Valdivia et al.,1992). Therefore, the elucidation of their three-dimensionalstructures and correlationship with their binding propertieswill provide valuable information for the design of selectivedrugs and insecticides. Although crystal structures of severalscorpion toxins have been determined (Cook et al., 2002; Heet al., 1999, 2000; Huang et al., 2003; Housset et al., 1994; Liet al., 1996; Oren et al., 1998; Pinheiro et al., 2003; Polikar-pov et al., 1999; Szyk et al., 2004; Zhao et al., 1992), it hasnot yet been possible to identify unique features that couldbe linked to their specific pharmacological properties. There-fore, new structures with novel features may be needed toestablish the precise structure–function relationships.

Here, we report the crystal structure of a novel neurotoxin

from scorpion Buthus tamulus at 2.2 �A0

resolution. The struc-ture revealed unique sites having a series of aromatic residueson one face of the molecule while on the other face a very highconcentration of acidic residues pointing in the oppositedirection has been observed. There is another face with pos-itively charged residues that are protruding in the third direc-tion. Despite an overall similar scaffolding, the presentstructure of B. tamulus neurotoxin (BTN) presents novel fea-tures that are highly suggestive of its potential binding prop-erties to receptors and these have not yet been observed in theprevious structures reported so far (Cook et al., 2002; Heet al., 1999, 2000; Housset et al., 1994; Huang et al., 2003;Li et al., 1996; Oren et al., 1998; Pinheiro et al., 2003; Polikar-pov et al., 1999; Szyk et al., 2004; Zhao et al., 1992).

Table 1Data collection and processing statistics

Resolution range (A) 15.0–2.2Space group P21

Cell dimensions (�A0)

a 50.8b 21.0c 52.4b (�) 94.4Vm (A3/Da) 2.0Number of molecules in the unit cell 4Solvent content (%) 38Number of unique reflections 5286Overall completeness 88Completeness in the highest resolution shell (2.28–2.20 A) 63Rsym 9.4Rsym in the highest resolution shell (2.28–2.20 A) 18.6Mean I/r (I) 8.2I/r (I) in the highest resolution shell (2.28–2.20 A) 2.0

2. Materials and methods

2.1. Purification

The samples of B. tamulus venom (red scorpion) wereobtained from Irula cooperative Snake Farm in TamilNadu, India. Two hundred and fifty milligrams of venomwas dissolved in 20 ml of 25 mM Tris–HCl buffer, pH 8.0.It was centrifuged at 8000g for 15 min to remove insolublematerial. The supernatant was loaded on a DEAE–Sephacelcolumn (30 · 2 cm) which was equilibrated with 25 mMTris–HCl, pH 8.0. The bound protein was eluted with a con-tinuous gradient of NaCl (0.0–1.0 M NaCl) in 25 mM Tris–HCl buffer, pH 8.0. The peak eluted at 0.6 M NaCl waspooled and gel filtered using Sephadex G-50 column(100 · 1.5 cm). The column was equilibrated with 25 mMTris–HCl, pH 8.0. The flow rate was adjusted to 5 ml/h.As reported earlier (Tu, 1977; Zlotkin et al., 1978) the pres-ence of toxin was confirmed by an activity bioassay. Thesamples of toxin were dialyzed against distilled water andlyophilized. The pI of BTN was estimated using isoelectricfocusing instrument that indicated a value of 4.35.

2.2. Complete sequence determination

The venom glands of B. tamulus were obtained fromIrula Snake Farm, Mahabalipuram, Tamil Nadu, India

with the permission of the Forest Department of theTamil Nadu government, Chennai, India and the com-plete cDNA sequence was determined. The venom glandswere removed from 25 Indian red scorpions (B. tamulus)after 2 days of devenomizing them. These were homoge-nized in 4 M GITC buffer (pH 5.0) in ice-cooled condi-tions and stored at �70 �C. The total RNA wasextracted by the phenol/chloroform method (Chomczyn-ski and Sacchi, 1987). The poly(A)+ mRNAs were isolat-ed from the total RNA using an oligo(dT)–cellulosecolumn (Amersham-Pharmacia). The small syringe col-umn packed with oligo(dT) cellulose was washed with10 ml of high salt buffer (1 M NaCl, 1 mM Na2–EDTA,and 40 mM Tris–HCl, pH 7.4). The total RNA wasmixed with an equal volume of high salt buffer andwarmed to 65 �C and cooled immediately by placing itin the ice. The chilled RNA was passed through columnpacked with oligo(dT)–cellulose. The column was washedwith 3 ml of low salt buffer (0.1 M NaCl, 1 mM Na2–EDTA), which was pre-warmed to 65 �C. The poly(A)+

RNA was used for cDNA synthesis which was carriedout with Moloney murine leukemia virus (MMLV)-reverse transcriptase using oligo(dT) primers. A portion(2 ll) of the reverse transcriptase polymerase chain reac-tion (RT-PCR) was used for the PCR amplification ofthe gene. The conserved nucleotide sequences from scor-pion venom toxins and N-terminal sequence obtainedusing Edman degradation were used for the design ofprimers. The sequences 5 0-CTATCCTGTCGAGGCCAAAGGA-3 0 and 5 0-AATTTATTGGACCTTCTGGCC-3 0 were used as forward and reverse primers, respec-tively. The PCRs were carried out using Taq polymerase(Promega, USA) using MJ research thermal cycler modelPTC-100. The nucleotide sequencing was carried outusing the cloned double-stranded DNA (pGEM-T) onan automatic sequencer model ABI-377. Both strandswere used for amplification.

54 M. Sharma et al. / Journal of Structural Biology 155 (2006) 52–62

2.3. Crystallization

The lyophilized samples of neurotoxins were usedfor crystallization experiments. Seven milligrams permilliliter protein solution was prepared in 25 mMTris–HCl, 5 mM CaCl2, pH 8.0. Crystals wereobtained at room temperature using the sitting dropvapour diffusion method. The drops were made bymixing 15 ll of protein solution with 5 ll of the reser-voir solution containing 80–85% of 2-methyl-2,4-pan-tanediol (MPD), 25 mM Tris–HCl, pH 8.0, and5 mM CaCl2. These drops were equilibrated against areservoir filled with 1 ml reservoir solution. Crystalsappeared after a week and grew to dimensions of0.5 · 0.1 · 0.05 mm3.

Fig. 1. Nucleotide and deduced amino acid sequences of acidic toxin from sc

Table 2Refinement statistics

PDB code 2A7TResolution limits (A) 15.0–2.2Number of reflections 5286Rcryst 20.6Rfree 22.5Protein atoms 980Water molecules 117Bond length (A) 0.007Bond angles (�) 1.4Dihedral angles (�) 24.0Overall G factor 0.15

Ramachandran plotResidues in the most favoured regions (%) 90.2Residues in the additionally allowed regions (%) 9.8

Average B factors (A2)From Wilson plot 15.0For main chain atoms 11.0For side chain atoms and water molecules 16.1For all atoms 14.0

2.4. Data collection and processing

The crystals belong to monoclinic space group P21 with

unit cell dimensions a = 50.9, b = 21.0, c = 52.4 �A0

andb = 94.4�. The Matthews coefficient was 2.0, assumingthe molecular weight to be 6.5 kDa and that there aretwo molecules in the asymmetric unit. Intensity data were

collected to a resolution of 2.2 �A0

from one crystal on aMAR Research imaging plate scanner, using CuKa mono-chromated radiation produced by a Regaku RU-200 rotat-ing anode X-ray generator. A total of 12426 reflectionswere collected, scaled and merged using the program SCA-LEPACAK (Otwinowski and Minor, 1997) into a set of5462 unique reflections with an Rsym of 9.4%.The complete-

ness was 88% to 2.2 �A0

resolution. The Rsym and complete-

ness in the resolution shell (2.28–2.20 �A0) were 18.6 and

63%, respectively. The data collection and processingdetails are given in Table 1.

2.5. Structure determination and refinement

The structure of was solved by molecular replacementusing the program AmoRe (Navaza, 1994). The neurotoxinBmK-M8 having sequence identity of 46% with BTNstripped of water molecules (PDB code: 1SNB, Li et al.,1996) was used as the initial search molecule. The solutiongave two distinct peaks indicating the presence of two mol-ecules in the asymmetric unit of space group P21. After afew rounds of rigid-body refinement, the models were com-pleted using ARP/wARP and REFMAC (Murshudovet al., 1997; Perrakis et al., 1999). Five percent of theobserved reflections were flagged for calculation of the freeR factor, Rfree. The model was improved by manually

orpion B. tamulus. The amino acid residues are given in three-letter code.

Table 3Sequence alignment of B. tamulus scorpion toxin with (a) classical a-toxins and (b) a-like toxins with secondary structure elements

(a) Classical-a toxins

(b) a-like toxins

The acidic amino acids are shaded in red, basic in blue and cysteines in yellow.


adjusting residues with program O (Jones et al., 1991) intothe electron densities calculated using |2Fo � Fc| and|Fo � Fc| maps. Several rounds of model building andrefinement allowed the model to converge with good geom-etry that well fitted into the electron density. The final mod-el contains 980 protein atoms from two molecules in the

Fig. 3. Ribbon representation of the averaged coordinates of BTNmolecule drawn with PyMol (DeLano, 2002). Three b-strands (b1, b3 andb4) and one a-helix (a2) of babb motif are indicated with starting andending residue numbers. Disulfide bridges are shown in yellow. Theimportant loops are indicated in red.

Fig. 2. Two crystallographically independent molecules A and B in theasymmetric unit. The two molecules are rotated by 84� with respect toeach other. The dotted line indicates an intermolecular hydrogen bondThe overall conformations of two molecules are identical.

.

asymmetric unit and 117 water molecules. The final R fac-tor is 0.206 and Rfree is 0.225. The final refinement statisticsis given in Table 2.

3. Results and discussion

3.1. Sequence analysis

The nucleotide and the deduced amino acid sequencesare given in Fig. 1. The polypeptide contains 64 amino acid

Fig. 4. Stereoview of the reverse turn formed by residues 8–12 which is stabilized by a hydrogen bond network between main chain atoms as well asbetween main chain and side chain atoms. It is important for the proper location of Cys12 which forms an S–S bridge with Cys63.

Table 4Classification of neurotoxins based on unique structural features

Source PDB ID Type Peptide 9–10:Cis/Trans

(a)Buthus martensii karsch 1SNB Classical a-toxin Trans

Buthus martensii karsch 1T7A Classical a-toxin Trans

Androctonus australis 1AHO Classical a-toxin Trans

Androctonus australis 1PTX Classical a-toxin Trans

(b)Buthus martensii karsch 1T7E a-like toxin Cis His10Buthus martensii karsch 1T7B a-like toxin Cis His10Buthus martensii karsch 1SN1 a-like toxin Cis His10Buthus martensii karsch 1DJT a-like toxin Cis His10Buthus martensii karsch 1KV0 a-like toxin Cis His10Buthus martensii karsch 1SN4 a-like toxin Cis His10Buthus martensii karsch 1OMY a-like toxin Cis Tyr11Androctonus australis 1SEG a-like toxin Cis Tyr10

(c)Centruroides sculpturatus 1JZA b-Toxin Cis Pro59Tityus serrulatus 1NPI b-Toxin Cis Pro40Centruroides exilcauda 2SN3 b-Toxin Cis Pro59

(d)Butholus judaicus 1BCG Insect toxin Trans

Buthus martensii karsch 1T0Z Insect toxin Trans


residues. The sequence of BTN contains the highest num-ber of acidic residues among the sequences of a-toxinsknown so far. As a result it has a highly acidic pI of4.35. It shows a sequence identity of approximately 50%with both classical a and a-like toxins (Table 3). All ofthem contain eight cysteine residues at identical locations.Several critical residues of BTN such as Asp8, Asp10,Thr13, Ile15, Asp28, Asp32, Trp38, Pro41, and Asp50are similar to classical a-toxins. Overall, the comparisonsof amino acid sequence of BTN show closer resemblanceto the sequences of classical a-toxins than those of a-liketoxins. A critical examination of amino acid sequence inBTN reveals clusters of acidic (Glu2, Asp3, Asp8, Asp10and Asp28, Asp32, Asp37), basic (Lys29, Lys30 andArg58, Lys62, Arg64) and aromatic/hydrophobic (Trp38,Trp39, Val40, Pro41, Tyr42, Gly43, Val44, Val45, Trp47)residues. It is noteworthy that a high concentration of neg-atively charged residues occurs at the N-terminal regionwhile positively charged residues are located on the C-ter-minal region. Yet another notable feature in the sequenceof BTN is the presence of several proline residues withinthe segment 51–57 giving rise to novel structural featureswhich may be important in the binding to specific receptors(Pintar et al., 1999).

3.2. Quality of the model

The final model consists of two crystallographically inde-pendent molecules with 980 protein atoms from 128 aminoacid residues and 91 water molecules. The protein structure

has a good geometry with r.m.s. deviations of 0.007 �A0

and1.4� from standard values of bond lengths and angles, respec-tively. The final crystallographic R factor for all the reflec-

tions in the resolution range 15.0–2.2 �A0

was 20.6% and

Rfree was 22.5%. The overall mean B factor was 14.0 �A0

2.

The final |2Fo � Fc| map shows that the structure is welldefined without any break in the backbone as well as in theside chains. A Ramachandran plot (Ramachandran and Sas-isekharan, 1968) of the main chain torsion angles (/,w)shows that 90.2% of the residues are in the most favouredregions as defined in the program PROCHECK (Laskowskiet al., 1993). The refinement statistics are given in Table 2.

3.3. Overall structure

The crystal structure determination reveals that thereare two crystallographically independent molecules in theasymmetric part of the unit cell (Fig. 2). The two molecules


will be referred hereafter to as molecules A and B. The con-formations of molecules A and B are identical with anr.m.s. shift of 0.3 �A

0for the positions of their Ca atoms.

The two molecules are oriented at 84� with respect to eachother. These molecules in the asymmetric unit interactpoorly with each other making just one direct hydrogenbond involving Trp39 O (molecule A) and Asn20 Nd1

(molecule B). There are a few more hydrogen bondsthrough solvent molecules together with a small numberof hydrophobic interactions. Since the two crystallograph-ically independent molecules have identical conformations,the analysis in the subsequent discussion of molecular

Fig. 5. The molecule of BTN toxin is characterized by three unique sites having(P-site) and hydrophobic/aromatic residues (A-site). The three sites are separa

Fig. 6. Distribution of the electrostatic charge at the surface of (A) BTN andregions are red. (C) The corresponding superimposed global fold of BTN and

structure will be presented using one BTN molecule (mole-cule A) only. The molecular structure of BTN toxin con-tains one a-helix (a2) involving residues 19–28 and threeanti-parallel b-strands (b1: residues 2–6, b 3: residues32–39 and b4: residues 45–55) resulting in a babb motifas observed in other a-toxins (Fig. 3). In addition to awell-ordered babb motif three distinct loops of functionalsignificance consisting of residues 8–12, 38–44 and 55–64have been observed. These loops protrude outwardly fromthe molecular scaffold (Fig. 3). The most characteristic ofthem is the segment 8–12 that forms a typical reverse turn(Fig. 4, Table 4). Although it does not belong to any

clusters of negatively charged residues (N-site), positively charged residuested widely in a triangular arrangement.

(B) Bmk-M8. Positively charged regions in blue and negatively chargedBmk-M8 is displayed in the same orientation.


standard class of b turns, it provides a highly correlatedpeptide bond between residues 9 and 10. This peptide hasboth trans as well as cis conformations in the scorpion ven-om toxins and has been considered as one of the most char-acteristic features for the classification of toxins into a- andb-toxins (Table 4). It adopts a trans conformation in clas-sical a-toxins while a-like toxins invariably have it as acis bond. As seen from Table 3, the sequences of toxinsindicate that the residue at position 8 is invariably Asp inclassical a-toxins while it is generally Lys and never Asp(acidic) in a-like toxins. Therefore, an examination of thesequences of toxins can give an indication of it being eithera-toxin or a-like toxin. On the other hand b-toxins are rec-ognized with cis Pro40 or cis Pro59 whereas insect toxinshave trans conformations. In the classical a-toxins the lastresidue of the reverse turn Cys12 is involved in the disul-phide bond with Cys63 of the C-terminal segment. Fourdisulphide bonds, Cys12–Cys63, Cys16–Cys36, Cys22–Cys46 and Cys26 - Cys48 stabilize the structure. The struc-ture of BTN toxin contains three distinct surface featuresconsisting of negatively charged region (N-site), hydropho-bic/aromatic face (A-site) and positively charged segment(P-site) (Fig. 5). These three sites are spaced widely in a tri-angular arrangement and are presumed to be important

Fig. 7. Comparison of acidic residues of N-site of BTN (green) and Bmk-M8 (yellow). The residue numbers are also indicated.

Table 5The clusters of hydrophobic/aromatic residues (A-site) assumed to be involvedknown crystal structures

Source of toxin (PDB code) Residue Nos.

(BTN): Buthus tamulus (1DQ7) Thr17(BmK-M8): Buthus matensii karsch (1SNB) Gly17(BmK-M1): Buthus matensii karsch (1T7A) Ala17(AaH): Androctonus australis (1AHO) Gly17(AaH II): Androctonus australis (1PTX) Gly17

functionally and also determine the relative potencies ofvarious toxins.

3.4. Highly acidic neurotoxin

The predominant part of toxins are basic proteins(Strong, 1990). A few acidic toxins have also beenreported from snake venoms (Francis et al., 1995; Liet al., 1996; Gu et al., 2002). Bmk-M8 was reported tobe the most acidic neurotoxin among them with a pI val-ue of 5.30 (Li et al., 1996). The BTN has extended thislimit further with a pI value of 4.35 and as of now, itis the most acidic neurotoxin known. It shows that thetoxin isoforms can occur over a wide range of pI values(4.35–9.01) (Figs. 6A and B) and cause different degreesof toxicity (Rosso and Rochat, 1985; Wang, 1999).Although attempts have been made to correlate toxicitywith pI values (Possani et al., 1999) but it is difficultto generalize this relationship. Multiple forms of mole-cules, from acidic to basic nature could have beenevolved to modulate a wide spectrum of gating mecha-nism of Na+ channels.

3.5. Negatively charged site (N-site)

As seen from Fig. 5, several acidic residues are clusteredtogether in one part of the folded molecule. It is also strik-ing that these are present at one end of the well definedbabb motif in a highly ordered environment. No other neu-rotoxin structure including the closely related Bmk-M8 hasa comparable feature showing such a high density of nega-tively charged residues (Fig. 7). This patch may also beinvolved in the toxin receptor binding.

3.6. Hydrophobic/aromatic face (A-site)

Scorpion toxins are involved in the binding withvoltage-dependent sodium channels involving aromaticresidues (Fontecilla-Camps et al., 1988; Landon et al.,1997). The face A (Fig. 5) shows three aromatic residuesPhe18, Trp38 and Trp39 lined up in a row with their sidechains protruding outwardly. It is notable that all theseresidues are located at relatively flexible loops providingenough scope for the final adjustment for an effectivebinding with the receptor. This is one of the three strongfeatures of the BTN structure and no other toxinstructure reported so far compares with its potential

in the receptor binding for the classical a-toxins from different species with

Phe18 Trp38 Trp39 Val45Ser18 Trp38 Ala39 Asn45Arg18 Trp38 Val39 Asn45Arg18 Trp38 Ala39 Asn45Arg18 Trp38 Ala39 Asn45


(Table 5). Even the closely related a-toxins, BmK-M8 (Liet al., 1996), Aah-II (Karbat et al., 2004) and Aah (Smithet al., 1997) structures lack considerably in providing sucha strong aromatic environment (Figs. 8A–C). This kind ofstrong structural arrangement suggests a high affinity ofBTN toxin to its receptor.

3.7. Positively charged site (P-site)

As seen from Table 1, this is the only neurotoxin whichdoes not contain negatively charged residue beyond the

Fig. 8. Comparison of aromatic residues of A-site of BTN (green)

position 50, while it has three positively charged Arg/Lysresidues placed at the end of a well-ordered rigid stretchcontaining several prolines and a cysteine residue. Thiscondition is comparable with other neurotoxins neithersequence-wise nor in terms of the conformation of theC-terminal fragment beyond residue 50. Although C-termi-nus is away from the highly structured babb motif, it issupported by one of the most rigidifying amino acidsequences (Table 3). This segment is further rigidified bya disulphide bridge, Cys12–Cys63. Overall, the positivecharge on this site may be responsible for a strong antico-

with (A) Bmk-M8, (B) AaH-II (yellow) and (C) AaH (yellow).

Fig. 9. Molecular packing of BTN toxin in the unit cell. The molecules generate an infinite helical column. The interior of the helical columns is filled witharomatic residues while charged residues are on the periphery.


agulant activity as has been reported in phospholipase A2

(Boffa et al., 1976; Carredano et al., 1998; Gowda et al.,1994; Jabeen et al., 2005; Kini and Evans, 1987).

3.8. Molecular packing

The molecules of toxin are packed in the unit cell in theform of helical columns. The interiors of these columns arefilled with aromatic residues while the periphery is occupiedby charged residues presenting negative and positive surfacesalternately (Fig. 9). The molecular columns interact witheach other through opposite charges, hydrogen bonds andhydrogen bonds through solvent molecules. In this arrange-ment of packing the clustering of aromatic residues andregions of charged residues exist distinctly. As seen fromFig. 9, the molecular packing is very tight and does not leavelarge voids. As a result the observed solvent content is only38% which one of the lower side of protein crystals.

4. Conclusions

More than a dozen crystal structures of scorpion tox-ins have been determined. The majority of them belongto a-like toxins. The present toxin from B. tamulus ischaracterized as classical a-toxin and is the most acidictoxin reported so far. The a- and b-toxins modify theNa+ permeability by modulating the gating mechanism

of the channel, rather than blocking the channel poreitself. Inspite of sequence information on more than 70scorpion toxins and crystal structures of more than 15toxin molecules, the relationships between the sequencedeterminants and structural features with toxicity arenot yet fully understood. However, some hot spots havebeen identified in the structures of toxins. The clusters ofnegatively charged residues (N-site), the positivelycharged residues (P-site) and hydrophobic/aromatic resi-dues (A-site) form important elements for direct interac-tions with the sodium channel and are responsible forthe formation of an electric potential involved in the rec-ognition of the sodium channel. The high charge densi-ties at N- and P-sites offer stronger electrostaticpotential and hence can alter the recognition levels ofsodium channels. These two sites in the BTN toxin mol-ecule contain the highest charge densities reported so farmaking the BTN toxin as the most potent classicala-toxin characterized so far. Similarly, at the A-site inthe BTN toxin structure, the number of hydrophobic/aromatic residues and their orientations make it as themost potent toxin to interact with the receptors. Theseresults indicate clearly that the BTN toxin has uniquestructural determinants that are responsible for itstoxicity. Indeed, the crystal structure of BTN toxin pro-vides an ideal framework with the formation of threewidely separated potent functional sites.


5. Data banks accession numbers

The nucleotide sequence of BTN has been submitted toGenBank with accession code of DQ116789. The finalatomic coordinates have been deposited with the RCSBProtein Data Bank and are available under accession code2A7T.

Acknowledgments

The authors acknowledge the Department of Scienceand Technology (DST), New Delhi for financial supportunder the Funds for Infrastructure in Science and Technol-ogy (FIST) programme as well as Department of Biotech-nology (DBT) for the infrastructure grant. T.J. and N.S.thank DST for financial grants under Fast-Trackprogramme in Life Sciences.

References

Boffa, G.A., Boffa, M.C., Winchenne, J.J., 1976. A phospholipase A2 withanticoagulant activity. Isolation from Vipera berus venom andproperties. Biochim. Biophys. Acta 429, 828–838.

Carredano, E., Westerlund, B., Persson, B., Saarinen, M., Ramaswamy,S., Eaker, D., Eklund, H., 1998. The three-dimensional structures oftwo toxins from snake venom throw light on the anticoagulant andneurotoxic sites of phospholipase A2. Toxicon 36, 75–92.

Catterall, W.A., 1980. Neurotoxins that act on voltage-sensitive sodiumchannels in excitable membranes. Annu. Rev. Pharmacol. Toxicol. 20,15–43.

Chomczynski, P., Sacchi, N., 1987. Single-step method of RNA isolationby acid guanidinium thiocyanate–phenol–chloroform extraction. Anal.Biochem. 162, 156–159.

Cook, W.J., Zell, A., Watt, D.D., Ealick, S.E., 2002. Structure of variant 2scorpion toxin from Centruroides sculpturatus Ewing. Protein Sci. 11,479–486.

Couraud, F., Jover, E., Dubois, J.M., Rochat, H., 1982. Two types ofscorpion receptor sites, one related to the activation, the other tothe inactivation of the action potential sodium channel. Toxicon 20,9–16.

DeLano, W.L., 2002. The PyMol molecular Graphics System. DeLanoScientific, San Carlos, CA. Available from: <http://www.pymol.org>.

Fontecilla-Camps, J.C., Habersetzer-Rochat, C., Rochat, H., 1988.Orthorhombic crystals and three-dimensional structure of the potenttoxin II from the scorpion Androctonus australis Hector. Proc. Natl.Acad. Sci. USA 85, 7443–7447.

Francis, B., Coffield, J.A., Simpson, L.L., Kaiser II, 1995. Amino acidsequence of a new type of toxic phospholipase A2 from the venom ofthe Australian tiger snake (Notechis scutatus scutatus). Arch. Biochem.Biophys. 318, 481–488.

Garcia, M.L., Hanner, M., Knaus, H.G., Koch, R., Schmalhofer, W.,Slaughter, R.S., Kaczorowski, G.J., 1997. Pharmacology of potassiumchannels. Adv. Pharmacol. 39, 425–471.

Gordon, D., Martin-Eauclaire, M.F., Cestele, S., Kopeyan, C., Carlier, E.,Khalifa, R.B., Pelhate, M., Rochat, H., 1996. Scorpion toxins affectingsodium current inactivation bind to distinct homologous receptor siteson rat brain and insect sodium channels. J. Biol. Chem. 271,8034–8045.

Gordon, D., Savarin, P., Gurevitz, M., Zinn-Justin, S., 1998. Functionalanatomy of scorpion toxins affecting sodium channels. J. Toxicol.Toxin Rev. 17, 131–159.

Gowda, V.T., Schmidt, J., Middlebrook, J.L., 1994. Primary sequencedetermination of the most basic myonecrotic phospholipase A2 fromthe venom of Vipera russelli. Toxicon 32, 665–673.

Gu, L., Zhang, H., Song, S., Zhou, Y., Lin, Z., 2002. Structure of an acidicphospholipase A2 from the venom of Deinagkistrodon acutus. ActaCrystallogr. D 58, 104–110.

He, X.L., Deng, J.P., Wang, M., Zhang, Y., Wang, D.C., 2000. Structureof a new neurotoxin from the scorpion Buthus martensii Karsch at1.76 �A

0. Acta Crystallogr. D 56, 25–33.

He, X.L., Li, H.M., Zeng, Z.H., Liu, X.Q., Wang, M., Wang, D.C., 1999.Crystal structures of two alpha-like scorpion toxins: non-proline cis

peptide bonds and implications for new binding site selectivity on thesodium channel. J. Mol. Biol. 292, 125–135.

Housset, D., Habersetzer-Rochat, C., Astier, J.P., Fontecilla-Camps, J.C.,1994. Crystal structure of toxin II from the scorpionAndroctonus australis Hector refined at 1.3 �A

0resolution. J. Mol. Biol.

238, 88–103.Huang, Y., Huang, Q., Chen, H., Tang, Y., Miyake, H., Kusunoki, M.,

2003. Crystallization and preliminary crystallographic study of rBmKalpha IT1, a recombinant alpha-insect toxin from the scorpion Buthus

martensii Karsch. Acta Crystallogr. D 59, 1635–1636.Jabeen, T., Singh, N., Singh, R.K., Ethayathulla, A.S., Sharma, S.,

Srinivasan, A., Singh, T.P., 2005. Crystal structure of a novelphospholipase A2 from Naja naja sagittifera with a strong anticoag-ulant activity. Toxicon 46, 865–875.

Jones, T.A., Zou, J., Cowan, S.W., Kjeldgaard, M., 1991. Improvedmethods for building models in electron density maps and the locationof errors in these models. Acta Crystallogr. A 47, 110–118.

Jover, E., Couraud, F., Rochat, H., 1980. Two types of scorpion neurotoxinscharacterized by their binding to two separate receptor sites on rat brainsynaptosomes. Biochem. Biophys. Res. Commun. 95, 1607–1614.

Karbat, I., Frolow, F., Froy, O., Gilles, N., Cohen, L., Turkov, M.,Gordon, D., Gurevitz, M., 2004. Molecular basis of the highinsecticidal potency of scorpion alpha-toxins. J. Biol. Chem. 279,31679–31686.

Kini, R.M., Evans, H.J., 1987. Structure–function relationships ofphospholipases. The anticoagulant region of phospholipases A2. J.Biol. Chem. 262, 14402–14407.

Landon, C., Sodano, P., Cornet, B., Bonmatin, J.M., Kopeyan, C.,Rochat, H., Vovelle, F., Ptak, M., 1997. Refined solution structure ofthe anti-mammal and anti-insect LqqIII scorpion toxin: comparisonwith other scorpion toxins. Proteins 28, 360–374.

Laskowski, R.A., Macarthur, M., Moss, D., Thornton, J., 1993. PRO-CHECK: a program to check stereochemical quality of proteinstructures. J. Appl. Crystallogr. 26, 283–290.

Li, H.M., Wang, D.C., Zeng, Z.H., Jin, L., Hu, R.Q., 1996. Crystalstructure of an acidic neurotoxin from scorpion Buthus martensii

Karsch at 1.85 �A0

resolution. J. Mol. Biol. 261, 415–431.Murshudov, G.N., Vagin, A.A., Dodson, E.J., 1997. Refinement of

macromolecular structures by the maximum-likelihood method. ActaCrystallogr. D 53, 240–255.

Navaza, J., 1994. AMoRe: an automated package for molecular replace-ment. Acta Crystallogr. A 50, 157–163.

Oren, D.A., Froy, O., Amit, E., Kleinberger-Doron, N., Gurevitz, M.,Shaanan, B., 1998. An excitatory scorpion toxin with a distinctivefeature: an additional alpha helix at the C terminus and its implicationsfor interaction with insect sodium channels. Structure 6, 1095–1103.

Otwinowski, Z., Minor, W., 1997. Processing of X-ray diffraction datacollected in oscillation mode. Methods Enzymol. 276, 307–326.

Perrakis, A., Morris, R., Lamzin, V.S., 1999. Automated protein modelbuilding combined with iterative structure refinement. Nat. Struct.Biol. 6, 458–463.

Pinheiro, C.B., Marangoni, S., Toyama, M.H., Polikarpov, I., 2003.Structural analysis of Tityus serrulatus Ts1 neurotoxin at atomicresolution: insights into interactions with Na+ channels. ActaCrystallogr. D 59, 405–415.

Pintar, A., Possani, L.D., Delepierre, M., 1999. Solution structure of toxin2 from Centruroides noxius Hoffmann, a beta-scorpion neurotoxinacting on sodium channels. J. Mol. Biol. 287, 359–367.

Polikarpov, I., Junior, M.S., Marangoni, S., Toyama, M.H., Teplyakov,A., 1999. Crystal structure of neurotoxin Ts1 from Tityus serrulatus

http://www.pymol.org


provides insights into the specificity and toxicity of scorpion toxins. J.Mol. Biol. 290, 175–184.

Possani, L.D., Becerril, B., Delepierre, M., Tytgat, J., 1999. Scorpiontoxins specific for Na+-channels. Eur. J. Biochem. 264, 287–300.

Ramachandran, G.N., Sasisekharan, V., 1968. Conformation of polypep-tides and proteins. Adv. Prot. Chem. 23, 283–438.

Rosso, J.P., Rochat, H., 1985. Characterization of ten proteins from thevenom of the Moroccan scorpion Androctonus mauretanicus maureta-

nicus, six of which are toxic to the mouse. Toxicon 23, 113–125.Smith, G.D., Blessing, R.H., Ealick, S.E., FontecillaCamps, J.C., Hau-

ptman, H.A., Housset, D., Langs, D.A., 1997. Ab initio structuredetermination and refinement of a scorpion protein toxin. ActaCrystallogr. D 53, 551–557.

Strong, P.N., 1990. Potassium channel toxins. Pharmacol. Ther. 46,137–162.

Szyk, A., Lu, W., Xu, C., Lubkowski, J., 2004. Structure of the scorpiontoxin BmBKTtx1 solved from single wavelength anomalous scatteringof sulfur. J. Struct. Biol. 145, 289–294.

Tu, A.T., 1977. Scorpions venoms. In: Venoms: Chemistry and MolecularBiology. Wiley, New York, pp. 459–483.

Valdivia, H.H., Kirby, M.S., Lederer, W.J., Coronado, R., 1992. Scorpiontoxins targeted against the sarcoplasmic reticulum Ca(2+)-releasechannel of skeletal and cardiac muscle. Proc. Natl. Acad. Sci. USA 89,12185–12189.

Wang, D.C., 1999. Series of bioactivity-variant neurotoxins from scorpionButhus martensii Karsch: X-ray crystal structure and functionalimplications. J. Nat. Toxins 8, 309–325.

Wheeler, K.P., Watt, D.D., Lazdunski, M., 1983. Classification of Na+

channel receptors specific for various scorpion toxins. Pflugers Archiv.(Eur. J. Physiol.) 397, 164–165.

Zhao, B., Carson, M., Ealick, S.E., Bugg, C.E., 1992. Structure ofscorpion toxin variant-3 at 1.2 A resolution. J. Mol. Biol. 227, 239–252.

Zlotkin, E., Miranda, F., Rochat, H., 1978. Chemistry and pharmacologyof Buthinae scorpion venoms. In: Bettini, S. (Ed.), Arthropod Venoms.Springer, Berlin, pp. 317–369.

crystal structure of a highly acidic neurotoxin from scorpion buthus tamulus at 2.2 Ǻ resolution...

Documents