A chimeric scorpion ��-toxin displays de novo electrophysiological properties similar to those of ��-like toxins: Design of chimeric scorpion ��-toxins

A chimeric scorpion a-toxin displays de novo electrophysiologicalproperties similar to those of a-like toxins

Balkiss Bouhaouala-Zahar1, Rym Benkhalifa1, Najet Srairi1, Ilhem Zenouaki1, Caroline Ligny-Lemaire2,Pascal Drevet2, Francois Sampieri3, Marcel Pelhate4, Mohamed El Ayeb1, Andre Menez2, Habib Karoui1

and Frederic Ducancel2

1Laboratoire des Venins et Toxines, Institut Pasteur de Tunis, Tunisia; 2Departement d’Ingenierie et d’Etude des Proteines,

CEA, Saclay, France; 3UMR 6560, Universite de la Mediterranee, CNRS, Ingenierie des Proteines, Laboratoire de Biochimie,

IFR Jean-Roche, Faculte de Medecine Nord, Marseille, France; 4Laboratoire de Neurophysiologie UPRES EA 2647,

UFR Sciences, Angers, France

BotXIV and LqhaIT are two structurally related long chainscorpion a-toxins that inhibit sodium current inactivation inexcitable cells. However, while LqhaIT from Leiurusquinquestriatus hebraeus is classified as a true and stronginsect a-toxin, BotXIV from Buthus occitanus tunetanus ischaracterized bymoderate biological activities. To assess thepossibility that structural differences between these twomolecules could reflect the localization of particular func-tional topographies, we compared their sequences. Threestructurally deviating segments located in three distinct andexposed loops were identified. They correspond to residues8–10, 19–22, and 38–43. To evaluate their functional role,three BotXIV/LqhaIT chimeras were designed by transfer-ring the corresponding LqhaIT sequences into BotXIV.

Structural and antigenic characterizations of the resultingrecombinant chimera show that BotXIV can accommodatethe imposed modifications, confirming the structural flexi-bility of that particulara/b fold. Interestingly, substitution ofresidues 8–10 yields to a new electrophysiological profile ofthe corresponding variant, partially comparable to that oneof a-like scorpion toxins. Taken together, these resultssuggest that even limited structural deviations can reflectfunctional diversity, and also that the structure–functionrelationships between insect a-toxins and a-like scorpiontoxins are probably more complex than expected.

Keywords: chimeric scorpion toxin; insect sodium channel;sodium current kinetics; molecular modelling.

Long-chain scorpion toxins isolated from Androctonusaustralis hector and Buthus occitanus tunetanus scorpions[1,2] are responsible for human envenomation, a publicheath problem in Tunisia [3]. These small and basicpolypeptides, composed of a globular core compacted byfour disulfide bridges, bind and modulate sodium channelsin excitable cells [4,5]. They have been divided into a- [6,7]and b-toxins [8,9] according to their mode of action andbinding properties [10,11].Whereas b-toxins interfere with the current activation

stage, a-toxins inhibit sodium current inactivation inexcitable cells (reviewed in [10,11]). a-Toxins have beeninstrumental in functional mapping of voltage gated sodiumchannels [12,13] and display a wide array of preferences oninteraction with sodium channels of different animal phyla[10].Scorpion a-toxins are classically divided into the follow-

ing three groups: (a) mammal a-toxins that are highly active

on mammals, and display very low toxicity to insects,e.g. AahII toxin from the venom of the scorpion Androct-onus australis hector; (b) insect a-toxins that are highly toxicto insect and shows weak activity in mammalian centralnervous system, e.g. LqhaIT from Leiurus quinquestriatushebraeus scorpion; (c) a-like toxins, that display similar hightoxicity to both mammals and insects, e.g. BomIII andBomIV [11,14] fromButhus occitanus mardochei and LqhIIIfrom Leiurus quinquestriatus hebraeus scorpions [15].Classically, binding of scorpion a-toxins to the receptor

site 3 on the extracellular surface of sodium channelsinduces prolongation of action potentials due to selectiveinhibition or slowing of the fast inactivation process of thesodium current in vertebrate and insect electrophysiologicalpreparations [12,16,17]. Interestingly, and despite somedifferences in their primary structure, all scorpion a-toxinsthat compete for binding to receptor site 3 on sodiumchannel reveal similar effects of inhibition [11,12,14,18].However, some comparative studies make uncertain thestrict assignment of many toxins to a particular pharmaco-logical group [11,17]. Thus, scorpion a and a-like toxinsshare similar and competitive binding activities towardsinsect sodium channels, when this is not the case in rat brainsynaptosomes [17]. These observations support the existenceof two distinct receptor sites for a and a-like toxins onsodium channels, the latter being more or less related inmammals or insects. On the other hand, LqhaIT one of themost studied insect a-toxins, seems to share some pharma-cological properties with a-like toxins [17], suggesting thatthe receptor sites recognized by both families of toxins either

Correspondence to B. Bouhaouala-Zahar, Laboratoire des Venins

et Toxines, Institut Pasteur de Tunis, 13 Place Pasteur, Belvedere,

Tunis, 1002 Tunisia.

Fax: + 216 71 791 833, Tel.: + 216 71 1843 755,

E-mail: [email protected]

Abbreviations: BotXIV, a-toxin from the venom of Buthus occitanustunetanus; LqhaIT, a-toxin from the venom of Leiurus quinquestriatushebraeus; TSB, tryptic soy broth; AP, action potential.

(Received 26 November 2001, revised 13 March 2002,

accepted 5 April 2002)

Eur. J. Biochem. 269, 2831–2841 (2002) � FEBS 2002 doi:10.1046/j.1432-1033.2002.02918.x

partially overlap, or are closely localized on insect sodiumchannels.Furthermore, recentdatamentioned thepossibilitythat LqhaIT could have a weak effect on sodium channelsactivation, an activity classically attributed to a-like toxinseven if such activity is not very significant [17]. Clearly,additional data are necessary to tentatively elucidate thestructure–function relationship of a-toxins in general, andbetween insect a-toxins and a-like toxins in particular.Recently, we cloned and characterized a new insect

a-toxin from the venom of the scorpion Buthus occitanustunetanus called BotXIV [19]. We showed that BotXIVshows 49.25 and 52.23% identities with LqhaIT andLqqIII, respectively, and is not toxic on mice even at highconcentration [up to 2.5 lg per 20 g of body weight atintracerebroventricularly (i.c.v.) route]. However, unlikeother insect a-toxins, BotXIV displays a weak anti-insectactivity, a moderate toxicity to cockroach and slows onlythe inactivation process of insect sodium channels. Also,comparison of BotXIV with BomIV, a classical member ofa-like toxins [11] revealed 57% of amino-acid identity, ascompared to the 73% existing identity between LqhaIT andBomIV. These data suggest that a-like toxins form a largefamily of structurally related compounds, displaying similarbasic biological properties, but susceptible also of expressingparticular activities.The aim of this paper was to tentatively explore the

possibility that subtle structural deviations between BotXIVand LqhaIT, could reflect the localization of a particularfunctional topography. Such a relation was recentlyevidenced in the case of three fingered toxins from snakes[20,21]. To investigate this hypothesis, we compared indetailed the amino-acid sequences of BotXIV and LqhaIT,and searched for significant differences. Thus, we identifiedthree different stretches of amino acid residues located inthree distinct exposed areas on the surface of the toxins: thefirst five-residue turn, the N-terminal part of ahelix and theb turn between the two last b strands. Using site-directedmutagenesis, three BotXIV variants were constructed byreplacement of residues 8–10, 19–22, or 38–43 with thosefound in LqhaIT. These three BotXIV/LqhaIT chimerastogether with the native BotXIV recombinant toxin wereexpressed in Escherichia coli. Their overall structural anddetailed electrophysiological properties were studied andcompared. Interestingly, in this paper we show thatsubstitution of residues 8–10 is associated with de novoelectrophysiological properties partially comparable withthose of BomIV, an a-like scorpion toxin. Implications ofthat particular functional anatomy elucidation regardingthe classification of a-toxins of scorpions will be discussed.

M A T E R I A L S A N D M E T H O D S

Materials

Enzymes were purchased from Boehringer–Mannheim andBiolabs. Oligonucleotides were synthesized by Genset.HPLC separation procedure was performed using a Mercksystem (L-4250 UV/Vis detector and L-6200 intelligentpump). N-Terminal sequencing was carried out using anApplied Biosystem sequencer (477A protein sequencer) online with a phenylthiohydantoin analyser (120A analyser).Dichroic spectra were recorded at 20 �C on a Jobin–YvonCD6 using toxin solutions which concentrations were

determined by spectrometry. SDS/PAGE was performedusing the Phast System from Pharmacia.

Bacterial strains and plasmids

The E. coli strain MV1190 was used as the host strain fortransformations by M13-derived vectors. The strain CJ236[dut–, ung–, thi1, relA1/pCJ105 (Cmr)] was used to preparesingle-stranded template DNA for mutagenesis, as des-cribed by Kunkel [22]. The bacterial host used for expres-sion was E. coli HB101 [F–D(mcrC-mrr) leu supE44 ara14galK2 lacY1 proA2 rpsL20 (Strr) xyl-5 mtl-1 recA13 23].Expression vector pEZZ18 was obtained from Pharmacia.

Molecular biology

Manipulations of DNA were performed according topublished procedures [24]. Single- and double-strandedDNA sequencing procedures were performed by thedideoxynucleotide method [25] using the T7 sequencing kitfrom Pharmacia and [35S]dATP (Amersham). Site-directedmutagenesis assays were performed according to Kunkelet al. [22] using a Bio-Rad kit. The cDNA encoding theprecursor of BotXIV [19] was modified as follows: a KpnI/BamHI fragment carrying the sequence encoding BotXIVwas inserted into the corresponding restriction sites ofM13mp19 to produce M13mp19-BotXIV template formutagenesis. Phages from an individual lysis plaque wereused to re-infect fresh host cells to produce high-titer phagestock. This stock was then passed through two rounds ofinfection of E. coli CJ236 dut– ung– host. Single-strandedphage DNA was then isolated from a large volume ofphage-containing supernatant. Substitutions of BotXIVamino-acid stretches: Q8-P-H10, I19-S-S-G22 or G38-H-K-S-G-H43 by the corresponding sequences in LqhaIT wereperformed using the following oligonucleotides: 5¢-GGTTATATTGCCAAGAACTATAACTGTGCATAC-3¢, 5¢-CATTGTTTAAAAATCTCCTCAGGCTGCGACACTTTA-3¢, and 5¢-ACGAGTGGCCACTGCGGACATAAATCTGGACACGGAAGTGCCTGCTGG-3¢, respectively.

Production of recombinant chimera toxins in E. coli

Bacteria E. coli HB101 transformed by the expressionvectors pEZZ-M8-10, pEZZ-M19-22 or pEZZ-M38-43were grown in a 5-L fermentor (LSL Biolafitte, SaintGermain en Lay, France) with an initial culture volume of4-L of tryptic soy broth (TSB) medium (Difco) supple-mented with 5 gÆL)1 of glucose and 200 lgÆmL)1 ampicillin.Conditions of production were performed as previouslydescribed [19]. Hybrid recombinant proteins contained inextracted periplasmic fractions and in the culture mediumwere purified by affinity chromatography on an IgG-Sepharose column according to Ducancel et al. [26], then,lyophilized. The procedure followed to cleave the fusionproteins by CNBr treatment was previously described byBoyot et al. [27]. Purification of cleaved recombinantchimeras was performed as previously described [19].

Electrophysiological techniques

Adult male cockroaches (Periplaneta americana) were used.A segment (1.5–2.5 mm) of one giant axon was isolated

2832 B. Bouhaouala-Zahar et al. (Eur. J. Biochem. 269) � FEBS 2002

from a connective linking the fourth and fifth abdominalganglia. The preparation was immersed in paraffin oil andan �artificial node of Ranvier� was created [28]. Activemembrane area of 0.01–0.02 mm2 (node) was superfusedwith saline or test solutions. Membrane potentials andtransmembrane currents of this small surface of axonalmembrane were recorded in current-clamp or voltage-clampusing the double oil-gap single fiber technique as describedin detail earlier [29,30]. Normal physiological saline had thefollowing composition (in mM): NaCl, 200; KCl, 3.1; CaCl2,5.4; MgCl2, 5.0; Hepes buffer, 1.0; pH 7.2. LyophilizedM8-10 and BotXIV were dissolved in the saline solution to finalconcentrations of 0.5 or 2.0 · 10)6 M, in the presence ofbovine serum albumin (0.25 mgÆmL)1) before tests. Potas-sium currents were blocked by 0.5 · 10)3 M 3,4-diamino-pyridine (Sigma Chemical, France), and when neededsodium currents were blocked by 5 · 10)7 M tetrodotoxin(Sigma Chemical, France).Sodium conductance (gNa) can be calculated as a function

of the membrane potential according to the equation:

gNa ¼ INa=ðEm � ENaÞ

where Em and ENa are the membrane potential, and thereversal potential for Na+ current, respectively. Smoothcurves correspond to the best fit through the mean datapoints according to the Boltzmann distribution:

gNa=gNa max ¼ 1=f1þ exp½ðE0:5 þ EmÞ=k�g

where E0.5 is the potential at which 50% of the maximalsodium conductance are reached, k is the slope factor.Voltage-dependence of steady-state inactivation of Na+

channels was determined using a conventional two-pulsesprotocol: the test pulse to )10 mV is preceded by long(40 ms) prepulses from )80 to +30 mV, and the relativeamplitude of the peak Na+ current during the test pulse isplotted according to the prepulse value. Smooth curvescorrespond to the best fit through the mean data pointsaccording to the Boltzmann distribution:

INa=INa max ¼ 1=f1þ exp½ðEm þ E0:5Þ=k�g

where E0.5 is the potential at which 50% of the sodiumchannels are inactivated, k is the slope factor.

Enzyme-linked-immuno-sorbent-assays

ELISAs were used to assess cross antigenicity of eachpurified recombinant BotXIV mutant towards differentpolyclonal antibodies. Some were raised against toxicfractions BotG-50 and AahG-50 from Buthus occitanustunetanus and Androctonus australis hector venoms, respect-ively; or against BotI and AahII purified toxins. For thispurpose, optimization of the previously described proce-dures [19,31] was carried out.

In vivo insect and mammal toxicities (biological assays)

For LD50 determination, groups of four female C57/B16mice (22 ± 0.2 g) were individually i.c.v. injected underlight diethyl ether anesthesia with 20 ng to 2.5 lg ofrecombinant proteins. Toxicity of purified BotXIV mutantswere assessed on four Blatella germanica males per dose(50 mg body weight). A volume of 0.5–2 lL was injected in

the abdominal segments, and the lethality was monitoredafter 1 h. For all injections, the solvent used was 0.15 MNaCl containing 1 mg BSA per mL. The LD50 values werecalculated according to Reed & Muench method [32].

Molecular modelling of BotXIV and m8–10 mutant

Molecular modelling of both BotXIV and its 8–10 mutantwere based on the experimentally determined three-dimen-sional structures of two templates: toxin II of Androctonusaustralis hector (AahII) solved at 1.3 A (PDB entry 1ptx[33]), and toxin V (CsV; PDB entry 1nra [34]), of Centruro-ides sculpturatus, by using the program MODELLER 3 [35],running on a silicon Graphics Indigo R3000 workstation.The first set of models was obtained from manual sequencealignment. To limit the problems with backbone dihedralsof nonGly residues, we avoided to aligning the Gly residueswith nonGly residues. The 40 first models (20 in eachprotein) were screened with the programs PROCHECK [36],PROSA I [37] and INSIGHT II (Molecular Simulation Inc.),from which only four models for each protein were selectedand then used as templates in the final subsequent series.

Electrostatic properties

The electrostatic potential and outside isopotential gradientson the molecule surfaces were computed with the programGRASP [38]. The ionic strength was 0.145 M and the proberadius was 1.4 A. The dielectric constant was 2 inside and 80outside the solute molecules. Except the His residues, allacidic and basic residues were set in their ionized form.

R E S U L T S

Identification of divergent regions

To tentatively clarify the structure–function relationshipsexisting between a-like and a-insect scorpion toxins, wecompared in details the primary structures of BotXIV andrepresentative a-toxins. Thus, LqqIII, BotIT1, BomIV andBom III display 49, 53, 54, 57 and 73% identities withLqhaIT, respectively. These data suggest that BotXIVoccupies an intermediate position between strictly insectici-dal a-toxins (LqhaIT, LqqIII, and BotIT1) and typicala-like toxins (BomIV and BomIII). This confirms ourprevious experimental results, which established thatBotXIV was inactive towards mammals and weakly toxicfor Blatella cockroaches [19]. Thus, a precise comparison ofBotXIV and LqhaIT primary sequences essentially revealedthree divergent regions containing most of the amino acidvariations noticed between these two functionally unrelatedmolecules. It is noteworthy, that these divergent regionsmostly correspond to three exposed to solvent loopsconnecting conserved ahelix and b sheet elements (Fig. 1).Thus, segment 8–10 is part of the b turn (8–12) following

the first b strand (residues 1–5). Interestingly, the sequencefound in BotXIV (Q8-P-H10) is similar to the correspondingones in BomIII and BomIV toxins (Q8-P-E10) two typicala-like toxins, when totally different from those ones of trueinsecticidal toxins (LqhaIT, LqqIII, and BotIT1), K/Q8-N-Y10. The second main divergent region is constituted of theloop preceding the unique ahelix. Classically, it displaysvariable length and amino acid content from one toxin to

� FEBS 2002 Design of chimeric scorpion a-toxins (Eur. J. Biochem. 269) 2833

another. Thus, the sequence I-S-S-G(19–22) found inBotXIV is replaced by D-A-Y in LqhaIT. Finally, the thirdvariable segment corresponds to residues 38–43, andincludes three amino acids of the second LqhaIT b strand(34–39) and the following b turn. Here also, BotXIV anda-insect toxins display totally different sequences:GHKSGH(38–43) vs. QWAGKY in LqhaIT for instance.Based on these observations, we built three BotXIV/LqhaIT chimeric molecules corresponding to the individualsubstitution of deviating and exposed to the solvent regions8–10, 19–22 or 38–43 in BotXIV by the equivalent sequencesfound in LqhaIT (Fig. 1). The latter should be noted:BotXIVM8-10, M19-22, and M38-43.

Production, purification and characterizationof recombinant chimera

Mutated DNA fragments surrounded by KpnI (5¢) andBamHI (3¢) restriction site sequences were excised fromM13mp19-BotXIV vectors and inserted into the corres-ponding sites in the pEZZ18 expression vector [39]. Thethree BotXIV variants were produced as recombinant ZZfusion proteins as previously described [19]. Briefly, fusionproteins were mainly found in the culture medium ofbacterial suspensions, as expected from pEZZ-18 expressionvector [39]. Affinity chromatography performed on IgG–Sepharose allowed as expected, recovery of recombinanthybrid proteins having an apparent 22-kDa molecular mass(not shown). We noticed also the presence of few lowmolecular weight fragments resulting probably from pro-teolytic degradation events of the toxin moiety, as previ-ously and classically observed for such compounds[19,26,40,41]. The three IgG-purified fractions revealedsimilar proteic profiles and overall yields were estimated

between 16 and 18 mg of fusion proteins per litre of culture.Recombinant variant fused molecules were treated bycyanogen bromide as previously described [27]. The averageefficiency of the CNBr cleavage was estimated to 35%(Fig. 2, lanes 3 and 4), and the products were purified bycation HPLC (data not shown). Recombinant chimeradisplayed an apparent 7 kDa molecular mass as showed inthe case of BotXIVM8-10 variant (Fig. 2, lane 2). Thethree recombinant chimera comigrated with recombinantBotXIV on a 20% SDS/polyacrylamide gel, and shown theexpected amino acid composition and N-terminal amino-acid sequences (data not shown). The circular dichroicspectra of the recombinant and chimeric BotXIV proteinsrevealed similar overall profiles (Fig. 3), associated however,

Fig. 2. SDS/PAGE (20% Phast-gel) of a cleaved M8-10 variant. Lane

2 represents the cleaved and HPLC-purified recombinant M8-10

variant. Chimeric protein BotXIVM8-10 appears as a proteic band of

7 kDa. Lanes 3 and 4 correspond to crude cleavage mixtures ofHPLC-purified hybrid fractions. Lanes 1 and 5, correspond tomedium

and low molecular mass markers, respectively.

Fig. 1. Alignment of principal scorpion a-neurotoxins amino acid sequences. The amino-acid sequences are aligned according to their cysteine

residues (orange) and their three-dimensional structures (in italic for known three-dimensional structures). Disulfide bridges are indicated in dashed

lines. Positively charged residues (K/R) are indicated in blue, negatively charged residues (D/E) in red, and aromatic residues in green. Consensus

numbering is displayed under the sequences. The secondary structures are indicated in the top line. Deletions are indicated by (–). The three main

divergent regions involved in the building of BotXIV/LqhaIT chimeric molecules are in boxes. LqhaIT, Leiurus quinquestriatus hebraeus a-insecttoxin; LqqIII, Leiurus quinquestriatus quinquestriatus toxin III; BotI, IT1, and XIV, Buthus occitanus tunetanus toxin I, insect toxin I, and XIV,

respectively; BomIII and BomIV, Buthus occitanus mardochei toxins III and IV.

2834 B. Bouhaouala-Zahar et al. (Eur. J. Biochem. 269) � FEBS 2002

to a weak increase in the positive 190-nm band coupled to amore significant one in the negative 205-nm band of the CDspectra. Finally, about 1 mg of each recombinant BotXIVvariants was obtained from 20 mg of initial fusion protein.

Cross antigenicity and biological toxicity of purifiedrecombinant BotXIV mutants

To establish the antigenic profiles of the three HPLC-purified BotXIV mutants, we performed ELISA usingvarious scorpion antitoxins sera [42]. Thus, BotXIVM8-10,BotXIVM19-22 and BotXIVM38-43 variants, were simi-larly recognized by the polyclonal antibodies raised againstBotG-50 (a partially purified mixture of Buthus occitanustunetanus venom) or BotI toxin. On the contrary, very lowcross antigenicity was observed with anti-AahII toxin orAahG-50 toxic fraction (data not shown). A similar resultwas initially observed for recombinant BotXIV [19].Together, these data indicate that the three substitutionsintroduced within BotXIV did not modify significantly itsoverall antigenic profile [19]. Furthermore, they indicatethat these three BotXIV mutants as BotXIV all belong tothe same antigenic group related to Buthus occitanustunetanus (Bot) scorpion a-toxins, which is different fromthose of Androctonus australis hector [42].I.c.v. injections in C57/Black 6 mice of purified Bot-

XIVM8-10, M19-22, or M38-43 variant ranging from 2 ngto 2.5 lg did not cause any toxic effect. These results clearlyestablished that the three BotXIV/LqhaIT chimeric mole-cules are devoided of any toxicity towardsmammal, becausea weakly toxic LD50 value in mammals corresponds on anaverage to 100 ng. This result was not surprising, becausethe starting toxin BotXIV was already inactive towardsmammals [19], and the three chimeras were obtained bylimited substitution of equivalent regions between BotXIV(as starting molecule) and LqhaIT (as donor molecule),which is classically reported as a potent anti-insect toxin[14,15,43]. Unexpectedly however, injection to Blatella of

1 lg of each of the three recombinant chimera did notinduce any additional insect lethality, when in the particularcase of BotXIVM8-10 variant, a contractive effect ofinjected Blatella was noticed. Thus, when substitution of19–22 and 38–43 regions completely affected the initialinsect toxicity of BotXIV, BotXIVM8-10 variant is charac-terized by an intermediate biological mode of actiontowards insects.

Electrophysiological results

To compare the effects of native and mutated BotXIVmolecules on the insect channels, we carried out standardcurrent- and voltage-clamp experiments on cockroachaxonal preparations, as described inMaterials andmethods.As previously reported, BotXIV has typical a-toxins effectsas it was shown to slow down the sodium currentinactivation [19]. First of all, both BotXIVM19-22 andM38-43 did not show any effect neither on action potentialor sodium current decay as compared to BotXIV (data notshown). These data confirm the absence of insect toxicitynoticed previously. On the other hand, in the case ofBotXIVM8-10 variant, modifications of the initial electro-physiological properties of BotXIV were observed incurrent- and voltage-clamp conditions.

Current-clamp conditions. When BotXIV (1.4 · 10)6 M)was devoid of any effect on the action potential (AP)amplitude, or on the membrane resting potential value,BotXIVM8-10 was applied at the same concentration;BotXIVM8-10 caused an unusual and slight membranedepolarization of about 5–8 mV (Fig. 4A,B). Furthermore,BotXIV and BotXIVM8-10 both increase the actionpotential duration and lead to a progressive prolongationof the evokedAPwith amore drastic effect in the case of thenative toxin. In fact, for the same toxin application time(14 min) BotXIV induced a plateau potential lasting for110 ms followed by a long lasting and a very slowly decayeddepolarization. In the case of the BotXIVM8-10, the evoked�plateau� is only 3–6 ms duration with a normal repolariza-tion phase (Fig. 4B). In addition, an artificial re-polarizationby-passing a constant hyperpolarizing current, did restoreneither the PA amplitude nor its initial duration.

Voltage-clamp conditions. Membrane sodium currentswere measured under voltage-clamp conditions in responseto a step depolarization from a holding potential(Eh) ¼ )60 mV to a membrane potential (Em) ¼)10 mV, after suppression of the potassium current with3.4-DAP 0.5 · 10)3 M. In normal conditions, and as shownin Fig. 4C,D less than 3 ms were necessary to obtain acomplete INa inactivation. Application of BotXIVM8-10induced a progressive and dose-dependent development of amaintained sodium current which was about 6% of thepeak at 0.7 · 10)6 M (Fig. 4D), and reached 12% at1.4 · 10)6 M. Besides, this mutant slightly decreased theinitial peak amplitude of the inward sodium current byabout 20–24%. At the end of the voltage pulse, themaintained current returned to 0 with a slightly sloweddeactivation. Finally, a 10-lM tetrodotoxin post-applicationgenerated a constant but slight inward current of about )20to )26 nA at the holding potential (Fig. 4D). In anotherseries of experiments, and after selective suppression of INa

Fig. 3. UV circular dichroism spectra of BotXIV toxin and its three

chimeric variants. The cell path length is 0.05 cm and the recording

temperature is 20 �C. The different CD spectra are shown as following:native BotXIV toxin (––), M8-10 (ÆÆÆÆ), M19-22 (- - -), and M38-43(- Æ - Æ -) BotXIV/LqhaIT chimera.

� FEBS 2002 Design of chimeric scorpion a-toxins (Eur. J. Biochem. 269) 2835

with 10)6 M of tetrodotoxin, it was demonstrated thatBotXIVM8-10 (1.4 · 10)6 M) did not modify potassiumconductance (data not shown).

Effects on sodium activation and inactivation voltagedependence. Sodium peak current measurements duringvoltage pulses from )80 to 10 mV allowed us to calculatesodium conductance (gNa) as a function of the membranepotential as indicated in Materials and methods. Figure 5Bshows that, BotXIVM8-10 shifted the relative sodiumconductance by about 5 to 10 mV towards negativepotentials, whereas BotXIV had no effect on activation(Fig. 5A). Figure 5 illustrates also the voltage dependenceof steady-state inactivation of Na+ channels. Under normalconditions, the sodium current inactivation was complete at0 to )10 mV, whereas with BotXIVM8-10 the inactivationwas almost complete but the curve is shifted by about 10–20 mV towards negative potentials (Fig. 5B). On the otherhand, under BotXIV application the sodium currentinactivation remained partial and the curve was shifted byabout 10 mV towards more positive potentials (Fig. 5A).It is important to notice that when using BotXIVM8-10,

the voltage dependence was decreased both during theactivation and inactivation mechanisms, in presence ofBotXIV the sodium current voltage dependence wasdecreased only in the case of the inactivation process. These

data imply that the mutated toxin opens Na+ channels atvery negative potential values, suggesting that sodiumcurrent is activated and inactivated earlier than in normalconditions. To summarize, BotXIVM8-10 variant affects inthe same time the sodium current activation and inactiva-tionmechanisms, when BotXIV only affects the inactivationprocess. Together, our results clearly show that the substi-tution of segment 8–10 in BotXIV by its equivalent inLqhaIT, yields to the acquisition by BotXIVM8-10 variantof original electrophysiological properties.Molecular mode-ling and electrostatic potential studies were then carried outto tentatively explain the biological properties displayed byBotXIVM8-10 variant.

Molecular modeling of native BotXIVand BotXIVM8-10 variant

Figure 6 illustrates the best a-carbon BotXIV and M8-10models we obtained. They are similarly oriented (face Aview) as the published experimental structures of CsEV3[34] and AahII [44]. Clearly, the selected models predictthree-dimensional structures similar to the related a-toxins.The main differences affect the C-terminal region especially

Fig. 5. Effects on sodium activation and inactivation voltage dependence.

Sodium peak current measurements during voltage pulses from )80 to10 mV allowed us to calculate sodium conductance (gNa) as a function

of the membrane potential (see Method paragraph). Control meas-

urements are represented by open symbols. (A) 10 min after BotXIV

application, the sodium current steady state inactivation curve (d) is

slowed and is shifted to more positive potentials while the sodium

current activation curve (m) remains unchanged. (B) Ten minutes after

M8-10 application, both sodium inactivation and activation curves

(d,m) are shifted to more negative potentials.

Fig. 4. Effects of native and M8-10 chimeric BotXIV proteins on the

action potential and the sodium current of isolated cockroach axons.

(A,B) Current-clamp experiments: superimposed records of action

potentials evoked by a short (0.5 ms) depolarizing current pulse of 18

to 20 nA, at initial time: control (C), after BotXIV (A) and M8-10 (B)

application. Note the progressive evolution of the action potential

durations under BotXIV (A) and the axonal membrane depolarization

under M8-10 associated to a slight prolongation of AP duration and a

AP amplitude decrease. An artificial repolarization (AR) does not

restitute the initial AP (B). (C,D) Voltage-clamp experiments: cock-

roach axonal sodium current recordings are evoked by voltage pulses

of )10 mV from a holding potential Eh ¼ )60 mV. In normal con-ditions, INa completely inactivates in less than 3 ms (C). BotXIV

application induces a maintained inward sodium current without

affecting the inward peak sodium current (C). M8-10 application

induces a decrease of the peak sodium current amplitude associated

with a maintained inward sodium current (D). Tetrodotoxin applica-

tion reveals a slight holding inward sodium current (D).

2836 B. Bouhaouala-Zahar et al. (Eur. J. Biochem. 269) � FEBS 2002

for BotXIV, the latter being predicted larger in the case ofBotXIVM8-10 variant. It is noteworthy, that in the 40BotXIV and M8-10 variant models, the C-terminusextremity displayed the homogenous organization shownon Fig. 6.

Electrostatic properties of the models

The existence onto the surface of scorpion a-neurotoxins ofan electrostatic potential is suspected to contribute to therecognition of the receptor site [45]. By resolving theBoltzman–Poisson equation, the program GRASP is capableof calculating the electrostatic potential at any point of agrid in the space surrounding a protein molecule [38].Applied to BotXIV and BotXIVM8-10 mutant, theprogram allowed the visualization of several charged aminoacid residues susceptible to form an electrostatic potentialgradient on the surface of the molecule in the solvent space(Fig. 7). The positively and negatively charged residues arein blue and red, respectively, whereas the neutral side-chainsare indicated in white. Figure 7A shows the face A(hydrophobic face) of BotXIV and BotXIVM8-10, com-paratively to the published RMN structure of LqhaIT [46].On the contrary of the clear dipolar charge repartition ofLqhaIT (with positive charges localized in C-terminalregion: R3, K42, K62 R65 and negative charges: D4,D20, D54), BotXIV and BotXIVM8-10 display a differentand less homogenous repartition of charges. However, theglobal charge (+1) of the face A of these three moleculesremains unchanged. BotXIVM8-10 and BotXIV display aconserved charge distribution except for the C-terminalregion, the orientation of which was modified upon substi-tution of segment 8–10, as predicted frommodeling. Finally,analysis of the B faces (Fig. 7B) revealed that the maindifferences between LqhaIT, BotXIV and BotXIVM8-10molecules, rely upon the substitutions K28/E29 in BotXIV

and BotXIVM8-10, together with K9/Q8 in the particularcase of BotXIVM8-10 (Fig. 7B).

D I S C U S S I O N

The aim of this paper was to explore the possibility thatsubtle structural deviations between BotXIV and LqhaIT,could reflect the localization of a particular functionaltopography. Based on sequence analysis and the identifica-tion of three divergent regions between these two a-toxins,we explored the biological implications of these segments bybuilding corresponding chimeric molecules. From a generalpoint of view and as compared to rBotXIV, the substitu-tions performed did not modify the yield of production ofthe three recombinant chimera. This suggests that thevariants display a structural stability and a proteolyticsensitiveness similar to those of the starting recombinantmolecule. This is partially corroborated by the CD spectraof the three purified recombinant variants that revealsimilar, but not identical, profiles as compared to rBotXIV.This observation suggests that the three recombinantBotXIV variants probably adopt a similar overall spatialarrangement as the unmodified compound, characterizedby a similar overall secondary structure content; however,the local structural features may be different. Furthermore,each chimeric molecule shares an identical antigenic profileand cross-reactivity pattern with BotXIV. Together, theseresults strongly suggest that the transfer of residues 8–10,19–22 or 38–43 has a limited effect on the overall three-dimensional structure adopted by the three recombinantmolecules generated and tested in the present study. Thisfurther illustrates the stability and the structural flexibility ofthe a/b scorpion motif [47].When the three segments were substituted to form the

BotXIV variants, M8-10, M19-22 and M38-43 form ahomogenous area largely overlapping the putative toxic

Fig. 6. Homology molecular modeling of

BotXIV and M8-10 mutant. Ribbon

representations of experimentally determined

structures of AahII (cyan), CsV (red), and

best models of BotXIV (orange) and its

M8-10 variant (pink).

� FEBS 2002 Design of chimeric scorpion a-toxins (Eur. J. Biochem. 269) 2837

surface of scorpion a-neurotoxins affecting sodium channelgating [45]. Unexpectedly, however, when the three substi-tutions were performed independently in BotXIV, the weaklethality initially observed after abdominal injection toBlatella cockroaches was totally abolished. More surpris-ingly, when BotXIVM19-22 and M38-43 mutants arecharacterized by an absolute lack of electrophysiologicaleffects on cockroach giant axons, M8-10 variant showscontroversies effects. Assuming the absence of majorstructural change as discussed above, the loss of toxicitysuggests that subtle structural deviations might haveaffected the structural integrity of the �minimal toxic� surface

characterizing BotXIV. Numerous studies are consistentwith a multipoint receptor recognition site onto the surfaceof scorpion a-neurotoxins including residues at positions: 8,10, 17, 18, 58, 59, 62, and 64 in interaction with the receptor[45,48–50]. In addition, the spatial arrangement of the toxinpolypeptide chain together with the formation of anelectrostatic potential are also predicted to participate tothe capacity of these compounds to interact specifically andwith high affinity with voltage-sensitive sodium channels. Inthis respect, our results are not surprising, because thesubstitutions we performed are only partial, and thus toolimited to yield to the design of a LqhaIT-like toxic site

Fig. 7. Electrostatic gradient potential obtained with GRASP program. The gradient potential surfaces were computed from the modeled structures of

BotXIV (top left), M8-10 mutant (top right), comparatively to that experimentally determined of recombinant LqhaIT (bottom). Faces A and B ofthe toxins are shown in (A) and (B), respectively.

2838 B. Bouhaouala-Zahar et al. (Eur. J. Biochem. 269) � FEBS 2002

respecting the structural integrity of the transferred region.Recently, we have shown the importance of such astructural respect in the case of the successful design of arecombinant fasciculin-like molecule obtained by transfer-ring the structural-deviating segments that exist betweenfasciculins and short-chain neurotoxins from snakes [20,21].Clearly, substitution of residues QPH(8–10) in BotXIV

by the segment KNY (LqhaIT), which was recentlyreported as playing a major role in the biological activityof LqhaIT [45], results in de novo electrophysiologicalproperties of BotXIVM8-10 variant as compared toBotXIV. Indeed, and as classically observed with scorpiona-toxins, BotXIVM8-10 variant induces a prolongation ofthe action potential duration on cockroach giant axons.Furthermore under voltage-clamp conditions, the voltagedependence is decreased both during the activation andinactivation mechanisms when in the same conditionsBotXIV toxin only slows the sodium current inactivationprocess. Thus, the sodium current activation and theinactivation mechanisms, are both and uniquely modifiedin presence of BotXIVM8-10. Such results were sometimesobserved after LqhaIT application, but in a much lesssignificant manner [51]. Sodium current activation is alsoaffected in the case of a-like toxins. Thus, it was recentlyshown that BomIV tested on the same insect preparationsinhibits the sodium current inactivation process withadditional effects on the sodium current activation [17]. Inthe presence of BotXIVM8-10 mutant, as with BomIV, weobserved a resting depolarization due to the inducedconstant holding current, which is also responsible of theshift of the sodium current voltage dependence to negativepotentials. Furthermore, this shift induces an �early� inacti-vating sodium current, which has not been seen until nowwith a and a-like neurotoxins. Thus, a part of the affectedsodium channels might be unable to participate in the fasttransient current. This hypothesis could explain, at leastpartially, the decrease of the peak sodium current.For the first time, a shift of inactivation curve towards

negative potentials can be unambiguously attributed to thesubstitution of residues 8–10 within the first b turn of ascorpion a-neurotoxin. That result provides a secondexample of how a structural deviation can be associatedto a particular functional topography. It also emphasizes theunique analytical power of a �positive� functional mappingstrategy based on the identification of a particular biologicalproperty absent in the host scaffold, i.e. BotXIV. However,the reasons for the dual effect observed upon transfer of thesegment 8–10 from LqhaIT to BotXIV (i.e., the loss ofinsect toxicity and the de novo acquisition of particularelectrophysiological properties) remains unclear. Interest-ingly, it was recently proposed that the pharmacologicalversatility displayed by scorpions a-neurotoxins might berelated to the structural configuration of the C-terminal tail[50]. Indeed, based on structure comparisons and bioactivesurface identifications, it was hypothesized that the highlyvariable and dynamic C-terminal tail together with thespatially vicinal residues, form the interacting areas onto thesurface of scorpion a-neurotoxins. Thus, the C-termini ofscorpion a-neurotoxins in general, and of LqhaIT inparticular, is positioned between the five-residue turn(8–12) and the b turn formed by residues 40–43. If the firstone is identical between BotXIVM8-10 and LqhaIT, thesecond turn displays several differences; AGK(40–43) vs.

KSG(40–43) in LqhaIT and BotXIVM8-10, respectively.Furthermore, Lys58, which is conserved within scorpiona-neurotoxins and predicted to form a network stabilizingthe C-terminus, is replaced by an isoleucine in BotXIVM8-10. Finally, the length and amino-acid composition ofthe C-termini of LqhaIT and BotXIVM8-10 are signifi-cantly different; RVPGKCR(58–66) for LqhaIT vs.IVHGEKCHR(59–67) in BotXIVM8-10. Thus, the com-position and length differences between native BotXIV andM8-10 variant vs. LqhaIT C-terminal tails, together with itspeculiar environment in BotXIVM8-10 vs. BotXIV, couldbe related to the biological properties expressed by thesedifferent molecules.Finally, the fact that a limitedmodification can directly or

indirectly be responsible for a new biological propertyshared between a-insect and a-like neurotoxins, suggeststhat BotXIV probably occupies an intermediate positionwithin the evolutionary scheme of these molecules. It alsosupports the hypothesis that the acquisition of suchelectrophysiological properties might constitute an earlybiological event on the way of the molecular design ofpotent sodium channel gated ligands.

A C K N O W L E D G E M E N T S

We wish to thank S. Pinkasfield and Dr F. Bouet for technical

assistance and N-terminal sequencing, Drs D. Gordon, S. Zinn-Justin

for fruitful discussions, and Dr P. Mansuelle for his help in preparation

of Figs 6 and 7. This work was supported in part by an IFS grant

(International Foundation for Science).

R E F E R E N C E S

1. Miranda, F., Kopeyan, C., Rochat, H. & Lissitzky, S. (1970)

Purification of animal neurotoxin. Isolation and characterization

of eleven neurotoxins from the venoms of the scorpions Androc-

tonus australis hector, Buthus occitanus tunetanus and Leiurus

quinquestriatus quinquestriatus. Eur. J. Biochem. 16, 514–523.

2. Gregoire, J. & Rochat, H. (1983) Covalent structure of toxins I

and II from the scorpion Buthus occitanus tunetanus. Toxicon 21,

153–162.

3. Krifi, M.N., El Ayeb, M. & Dellagi, K. (1996) New procedures

and parameters for better evaluation of Androctonus australis

garzonii (Aag) and Buthus occitanus tunetanus (Bot.) scorpion

envenomations and specific serotherapy treatment. Toxicon 34,

257–266.

4. Catterall, W.A. (1986) Molecular properties of voltage-sensitive

sodium channels. Annu. Rev. Biochem. 55, 953–985.

5. Hille, B. (1992) Ionic Channels of Excitable Membranes. Sinauer

Associates, Sunderland MA.

6. Rochat, H., Bernard, P. & Couraud, F. (1979) Scorpion toxins:

chemistry and mode of action. Adv. Cytopharmacol. 3, 325–330.

7. Catterall, W.A. (1984) The molecular basis of neuronal excitabil-

ity. Science 223, 653–661.

8. Jover, E., Couraud, F. &Rochat, H. (1980) Two types of scorpion

neurotoxin characterized by their binding to two separate receptor

sites on rat synaptosomes. Biochem. Biophys. Res. Commun. 95,

1607–1614.

9. Couraud, F., Jover, E., Dubois, J.M. & Rochat, H. (1982) Two

types of scorpion toxins receptors sites, one related to the activa-

tion of the action potential sodium channel. Toxicon 20, 9–16.

10. Martin-Eauclaire, M.F. & Couraud, F. (1995) Scorpion

neurotoxins: effects and mechanisms. In Handle Neurotoxi-

cology (Chang, L.W. & Dyer, R.S eds) p. 683 Marcel Dekker,

New York.

� FEBS 2002 Design of chimeric scorpion a-toxins (Eur. J. Biochem. 269) 2839

11. Gordon, D., Savarin, P., Gurevitz, M. & Zinn-Justin, S. (1998)

Functional anatomy of scorpion toxins affecting sodium channels.

J. Toxicol. Toxin Rev. 17, 131–159.

12. Catterall, W.A. (1992) Cellular and molecular biology of voltage-

gated sodium channels. Physiol. Rev. 72, 15–48.

13. Gordon, D. (1997) A new approach to insect-pest control-com-

bination of neurotoxins interacting with voltage sensitive sodium

channels to increase selectivity and specificity. Invert. Neurosci. 3,

103–116.

14. Gordon, D., Martin-Eauclaire, M.F., Cestele, S., Kopeyan, C.,

Carlier, E., Ben Khalifa, R., Pelhate, M. & Rochat, H. (1996)

Scorpion toxins affecting sodium current inactivation bind to

distinct homologous receptor sites on rat brain and insect sodium

channels. J. Biol. Chem. 271, 8034–8045.

15. Sautiere, P., Cestele, S., Kopeyan, C.,Martinage, A., Drobecq,H.,

Doljansky, Y. & Gordon, D. (1998) New toxins acting on sodium

channels from the scorpion Leiurus quinquestriatus hebraeus sug-

gest a clue to mammalian versus insect selectivity. Toxicon 36,

1141–1154.

16. Gordon, D. & Zlotkin, E. (1993) Binding of scorpion toxin to

insect sodium channels is not dependent on membrane potential.

FEBS Lett. 315, 125–128.

17. Cestele, S., Stankiewicz, M., Mansuelle, P., De Waard, M., Dar-

gent, B., Gilles, N., Pelhate, M., Rochat, H., Martin-Eauclaire,

M.F. & Gordon, D. (1999) Scorpion alpha-like toxins, toxic to

bothmammals and insects, differentially interact with receptor site

3 on voltage-gated sodium channels in mammals and insects. Eur.

J. Neurosci. 11, 975–985.

18. Krimm, I., Gilles, N., Sautiere, P., Stankiewicz, M., Pelhate, M.,

Gordon, D. & Lancelin, J.M. (1999) NMR structures and activity

of a novel alpha-like toxin from the scorpion Leiurus quinques-

triatus hebraeus. J. Mol. Biol. 285, 1749–1763.

19. Bouhaouala-Zahar, B., Ducancel, F., Zenouaki, I., Ben Kha-

lifa, R., Borchani, L., Pelhate, M., Boulain, J.C., El Ayeb,

M., Menez, A. & Karoui, H. (1996) A recombinant insect-

specific alpha-toxin of Buthus occitanus tunetanus scorpion confers

protection against homologous mammal toxins. Eur. J. Biochem.

238, 653–660.

20. Ricciardi, A., Du Le,M.H., Khayati, M., Dajas, F., Boulain, J.C.,

Menez, A. & Ducancel, F. (2000) Do structural deviations

between toxins adopting the same fold reflect functional differ-

ences? J. Biol. Chem. 275, 18302–18310.

21. Le Du, M.H., Ricciardi, A., Khayati, M., Menez, R., Boulain,

J.C., Menez, A. & Ducancel, F. (2000) Stability of a structural

scaffold upon activity transfer: X-ray structure of a three fingers

chimeric protein. J. Mol. Biol. 296, 1017–1026.

22. Kunkel, T.A. (1985) The mutational specificity of DNA poly-

merases alpha and gamma during in vitroDNA synthesis. J. Biol.

Chem. 260, 12866–12874.

23. Boyer, H.W. & Roulland-Dussoix, D. (1969) A complementation

analysis of the restriction and modification of DNA in Escherichia

coli. J. Mol. Biol. 41, 459–472.

24. Sambrook, J., Fritsch, E.F. & Maniatis, T. (1989) Molecular

cloning. A Laboratory Manual. Cold Spring Harbor Laboratory

Press, Cold Spring Harbor NY.

25. Sanger, F., Nicklen, S. & Coulson, A.R. (1977) DNA sequencing

with chain-terminating inhibitors. Biotechnology 24, 104–108.

26. Ducancel, F., Boulain, J.C., Tremeau, O. & Menez, A. (1989)

Direct expression inE. coli of a functionally active proteinA-snake

toxin fusion protein. Protein Eng. 3, 139–143.

27. Boyot, P., Pillet, L., Ducancel, F., Boulain, J.C., Tremeau, O. &

Menez, A. (1990) A recombinant snake neurotoxin generated by

chemical cleavage of a hybrid protein recovers full biological

properties. FEBS Lett. 266, 87–90.

28. Pichon, Y. & Boistel, J. (1967) Current–voltage relations in the

isolated giant axon of the cockroach under voltage-clamp condi-

tions. J. Exp. Biol. 47, 343–355.

29. Pelhate, M. & Sattelle, D.B. (1982) Pharmacological properties of

insects axons: a review. J. Insect Physiol. 28, 889–903.

30. Benkhalifa, R., Stankiewicz, M., Lapied, B., Turkov, M.,

Zilberberg, N., Gurevitz, M. & Pelhate, M. (1997) Refined elec-

trophysiological analysis suggests that a depressant toxin is a

sodium channel opener rather than a blocker. Life Sci. 61,

819–830.

31. Chavez-Olortegui, C., Amara, D.A., Rochat, H., Diniz, C. &

Granier, C. (1991) In vivo protection against scorpion toxins by

liposomal immunization. Vaccine 9, 907–910.

32. Reed, L. &Muench, H. (1938) A simple method of estimating fifty

percent in eucaryotic end point. Am. J. Hyg. 27, 493–497.

33. Housset, D., Habersetzer-Rochat, C., Astier, J.P. & Fontecilla-

Camps, J.C. (1994) Crystal structure of toxin II from the scorpion

Androctonus australis hector refined at 1.3 A resolution. J. Mol.

Biol. 238, 88–103.

34. Jablonski, M.J., Watt, D.D. & Krishna, N.R. (1995) Structure of

an old world-like neurotoxin from the venom of the new world

scorpion Centruroides sculpturatus ewing. J. Mol. Biol. 248,

449–458.

35. Sali, A. & Blundell, T.L. (1993) Comparative protein modeling by

satisfaction of spatial restraints. J. Mol. Biol. 234, 779–815.

36. Laskowski, R.A., MacArthur, M.W., Moss, D.S. & Thornton,

J.M. (1993) PROCHECK – a program to check the stereo

chemical quality of protein structure. J. Appl. Crystallogr. 26,

283–291.

37. Sippl, M.J. (1993) Recognition of errors in three-dimensional

structures of proteins. Proteins 17, 355–362.

38. Nicholls, A., Sharp, K.A. & Honig, B. (1991) Protein folding and

association: insights from the interfacial and dynamic properties of

hydrocarbons. Proteins 11, 281–296.

39. Nilsson, B., Moks, T., Jansson, B., Abrahmsen, L., Elmblad, A.,

Holmgren, E., Henrichson, C., Jones, T.A. & Uhlen, M. (1987)

A synthetic IgG-binding domain based on staphylococcal protein

A. Protein Eng. 1, 107–113.

40. Hodgson, D., Gasparini, S., Drevet, P., Ducancel, F., Bouet, F.,

Boulain, J.C., Harris, J.B. & Menez, A. (1993) Production of

recombinant notechis 11¢2L, an enzymatically active mutant of aphospholipase A2 from Notechis scutatus scutatus venom, as

directly generated by cleavage of a fusion protein produced in

Escherichia coli. Eur. J. Biochem. 212, 441–446.

41. Danse, J.M., Rowan, E.G., Gasparini, S., Ducancel, F., Vatan-

pour-Young, L.C., Poorheidari, G., Lajeunesse, E., Drevet, P.,

Menez, R., Pinskasfeld, S., Boulain, J.C., Harvey, A.L. & Menez,

A. (1994) On the site by which a-dendrotoxin binds to voltage-dependent potassium channels: site directed mutagenesis reveals

that the lysine triplet 28–30 is not essential for binding. FEBS Lett.

356, 153–158.

42. El Ayeb, M., Delori, P. & Rochat, H. (1983) Immunochemistry of

scorpion a-toxins antigenic homologies checked with radio-immuno-assays (RIA). Toxicon 21, 709–716.

43. Chen, H., Gordon, D. & Heinmann, H.S. (2000) Modulation of

cloned skeletal muscle sodium channels by the scorpion toxins

LqhII, LqhIII and LqhaIT. Pflugers Arch. 439, 423–432.44. Fontecilla-Camps, J.C., Habersetzer-Rochat, C. & Rochat, H.

(1988) Orthorhombic crystals and three-dimensional structure of

the potent toxin II from the scorpion.Androctonus australis hector.

Proc. Natl Acad. Sci. USA 85, 7443–7447.

45. Zilberberg, N., Froy, O., Loret, E., Cestele, S., Arad, D., Gordon,

D. &Gurevitz, M. (1997) Identification of structural elements of a

scorpion alpha-neurotoxin important for receptor site recognition.

J. Biol. Chem. 272, 14810–14816.

46. Tugarinov, V., Kustanovitz, I., Zilberberg, N., Gurevitz, M. &

Anglister, Y. (1997) Solution structures of a highly insecticidal

recombinant scorpion alpha-toxin and a mutant with increased

activity. Biochemistry 36, 2414–2424.

2840 B. Bouhaouala-Zahar et al. (Eur. J. Biochem. 269) � FEBS 2002

47. Vita, C. (1997) Engineering novel proteins by transfer of active

sites to natural scaffolds. Curr. Opin. Biotechnol. 8, 429–434.

48. Kharrat, R., Darbon, H., Rochat, H. & Granier, C. (1989)

Structure-activity relationships of scorpion a-toxins. Multipleresidues contribute to the interaction with receptors. Eur. J. Bio-

chem. 181, 381–390.

49. Gordon, D., Moskowitz, H., Eitan, M., Warner, C., Catterall,

W.A. & Zlotkin, E. (1992) Localization of receptor sites for

insect-selective toxins on sodium channels by site directed anti-

bodies. Biochemistry 31, 7622–7628.

50. Gurevitz, M., Gordon, D., Ben-Natan, S., Turkov,M. & Froy, O.

(2001) Diversification of neurotoxins by C-tail �wiggling�: a scor-pion recipe for survival. FASEB J. 15, 1201–1205.

51. Pelhate, M., Stankiewicz, M. & Benkhalifa, R. (1998) Anti-insect

scorpion toxins: historical account, activities and prospects. C. R.

Soc. Biol. 192, 463–484.

� FEBS 2002 Design of chimeric scorpion a-toxins (Eur. J. Biochem. 269) 2841