Báo cáo khoa học: Characterization of scorpion a-like toxin group using two new toxins from the scorpion Leiurus quinquestriatus hebraeus doc

Eur J Biochem 269, 3920–3933 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03065.x Characterization of scorpion a-like toxin group using two new toxins from the scorpion Leiurus quinquestriatus hebraeus ` Alain Hamon1, Nicolas Gilles2, Pierre Sautiere3, Arlette Martinage3, Charles Kopeyan4, Chris Ulens5, Jan Tytgat5, Jean-Marc Lancelin6 and Dalia Gordon7 Laboratoire de Neurophysiologie, UPRES EA-2647, Universite´ d’Angers, France; 2CEA, DIEP, Saclay, Gif-sur-Yvette, France, Laboratoire de Chimie des Biomole´cules, CNRS URA-1309, Institut Pasteur de Lille, France; 4Laboratoire de Biochimie, CNRS URA-1455, Faculte´ de Me´decine Nord, Marseille, France; 5Laboratory of Toxicology, University of Leuven, Belgium; Laboratoire de RMN Biomole´culaire, CNRS UMR-5078, Lyon, France; 7Department of Plant Sciences, Tel Aviv University, Ramat-Aviv, Tel Aviv Israeăl Two novel toxins, Lqh6 and Lqh7, isolated from the venom of the scorpion Leiurus quinquestriatus hebraeus, have in their sequence a molecular signature (8Q/KPE10) associated with a recently defined group of a-toxins that target Na channels, namely the a-like toxins [reviewed in Gordon, D., Savarin, P., Gurevitz, M & Zinn-Justin, S (1998) J Toxicol Toxin Rev 17, 131–159] Lqh6 and Lqh7 are highly toxic to insects and mice, and inhibit the binding of a-toxins to cockroach neuronal membranes Although they kill rodents by intracerebroventricular injection, they not inhibit the binding of antimammal a-toxins (e.g Lqh2) to rat brain synaptosomes, not even at high concentrations Furthermore, in voltage-clamp experiments, rat brain Na channels IIA (rNav1.2A) expressed in Xenopus oocytes are not affected by Lqh6 nor by Lqh7 below lM In contrast, muscular Na channels (rNav1.4 and hNav1.5) expressed in the same cells respond to nanomolar concentrations of Lqh6 and Lqh7 by slowing of Na current inactivation and a leftward shift of the peak conductance–voltage curve The structural and pharmacological properties of the new toxins are compared to those of other scorpion a-toxins in order to re-examine the hallmarks previously set for the a-like toxin group Natural toxins have been widely used to investigate the localization, biophysical properties and structure–function relationships of voltage-gated Na channels [1,2] Until now, however, most of the toxins did not provide efficient tools for investigating the functional roles of particular channel subtypes due to their poor selectivity In mammals, at least 10 distinct pore-forming a-subunit subtypes can be distinguished on the basis of their primary structure, biophysical properties, tissue distribution and sensitivity to tetrodotoxin [3] In the brain, the four main subunits, named Nav1.1 (Brain I), Nav1.2 (BII), Nav1.3 (BIII) and Nav1.6 (NaCh6), are all highly sensitive to tetrodotoxin The somatodendritic concentration of types Nav1.1, 1.3 and 1.6 suggests a potential role in the integration of synaptic inputs, whereas the tentative axonal localization of subtype Nav1.2 implies a role in the conduction of action potentials in unmyelinated fibers [4] In the periphery, Nav1.6 is also highly expressed at nodes of Ranvier [5], reflecting its involvement in conduction along myelinated axons Three additional subtypes are abundant in dorsal root ganglion (DRG) neurons: the tetrodotoxin-sensitive Nav1.7 (PN1) and the tetrodotoxinresistant Nav1.8 (PN3, SNS) and Nav1.9 (NaN, SNS2), the two latter being involved in nociception Lastly, the tetrodotoxin-sensitive Nav1.4 (SkM1) and the tetrodotoxin-resistant Nav1.5 (H1) are primarily expressed in skeletal muscles and heart, respectively Other subtypes have been isolated but remain to be functionally characterized As the mammalian brain expresses a great variety of Na channels whose functional roles are poorly understood, the discovery of toxins that discriminate between neuronal subtypes would be of high interest With that objective in mind, we have concentrated our efforts on a particular group of scorpion a-toxins, which has been recently shown to discriminate between neuronal Na channel subtypes [6,7] All scorpion a-toxins are long polypeptides (60–70 aminoacid residues) stabilized by four disulfide bridges and their 3D structure shows a dense core comprising an a helix and a three-stranded b sheet motif [8,9] Their main effect is to prolong action potentials by slowing the inactivation of Na currents through binding to the so-called receptor site on the a subunit [2] When administered by a subcutaneous route, all a-toxins reveal a quite similar high toxicity to mice mostly via their effects on skeletal muscle Na channels [10,11] However, they can be divided into three functional Correspondence to D Gordon, Department of Plant Sciences, Tel Aviv University, Ramat-Aviv, Tel Aviv 69978, Israel Fax: + 972 640 6100; E-mail: dgordon@post.tau.ac.il, or A Hamon, Laboratoire de Neurophysiologie, UPRES-EA 2647, ´ Universite d’Angers, Bd Lavoisier, 49045 Angers cedex 01, France Fax: + 33 241 735 215; E-mail: alain.hamon@univ-angers.fr Abbreviations: Aah2, alpha toxin II from the venom of the scorpion Androctonus australis hector, also called AaH; ATX II, toxin II of the sea anemone Anemonia sulcata; Bom3,4, a-like toxins from the venom of the scorpion Buthus occitanus mardochei, LD50, 50% lethal dose; LqhaIT, Lqh2 and Lqh3, alpha toxin highly active on insects, alphatoxin highly active on mammals and alpha-like toxin, respectively, from the venom of the scorpion Leiurus quinquestriatus hebraeus; Lqq3, alpha toxin from the venom of the scorpion Leiurus quinquestriatus quinquestriatus (Received 30 April 2002, accepted 20 June 2002) Keywords: scorpion toxin; sodium channel; oocyte; insect; mammal Ó FEBS 2002 groups according to their preferential toxicity to mammals or insects and their differential binding properties The classical a-toxins or antimammal a-toxins (e.g Aah2 and Lqh2) are highly toxic to mammals and very poorly active on insects, whereas the anti-insect a-toxins (e.g LqhaIT and Lqq3) show a very high binding affinity to insect Na channels Studies with Bom3 and Bom4 from Buthus occitanus mardochei and Lqh3 from Leiurus quinquestriatus hebraeus have led to the characterization of a third group, the a-like toxins [12] These toxins kill both mammals and insects but bind to insect channels with a lower affinity than the antiinsect LqhaIT In addition, they neither bind nor compete for binding with classical a-toxins in rat brain synaptosomes, although they are lethal to rodents by intracerebroventricular (i.c.v.) injection This toxic effect in the brain is mediated (at least partly) by tetrodotoxin-sensitive Na channels located mainly on neuronal cell bodies but not on nerve terminals [6] Identification of the precise channel subtype that is targeted by the a-like toxins in brain neurons as well as clarification of the molecular basis of their unique selectivity represents an important challenge of future research One possible approach to the latter issue is to compare the structural features that distinguish these toxins from antimammal and antiinsect a-toxins Here, we report the biochemical and pharmacological characterization of two new toxins, designated Lqh6 and Lqh7, from the venom of the yellow scorpion Leiurus quinquestriatus hebraeus and show that both represent new members of the a-like group In addition, using a comparative approach and structural modeling, we re-examine the criteria that may be used to characterize this group of scorpion toxins EXPERIMENTAL PROCEDURES Materials Lqh6 and Lqh7 were isolated and purified from the ionexchange fractions of the venom of the scorpion Leiurus quinquestriatus hebraeus (Lqh) obtained according to the procedure described previously [13] Lqh2, Lqh3 and LqhaIT, from the same venom, were purchased from Latoxan Toxin II from Androctonus australis hector (Aah2) was a generous gift from H Rochat (University of Marseille, France) In order to unify the nomenclature, all scorpion toxins used in the present work are designated using Arabic numbers, as recently suggested [14] Chymotrypsin (EC 3.4.21.1) treated with tosyllysyl-chloromethane was obtained from Merck Carboxypeptidase A (EC 3.4.17.1) treated with diisoprylfluorphosphate was purchased from Sigma and carboxypeptidase P (EC 3.4.17.16) sequencing grade was obtained from Boehringer (Mannheim, Germany) All reagents and solvents were of the highest purity available Purification, molecular mass and sequence of Lqh6 and Lqh7 Purification procedure Ion-exchange fractions and were submitted to preparative RP-HPLC on a Nucleosil column C18 and the purity of Lqh6 and Lqh7 (polypeptides Lqh4-2 and Lqh5-2, respectively) obtained after preparative RP-HPLC was assessed by capillary electrophoresis according to [15] Two new a-like toxins (Eur J Biochem 269) 3921 Mass spectrometry The molecular masses of the native polypeptides and of their fragments generated from enzymatic cleavage were determined by ion-spray mass spectrometry [15] Reduction and alkylation The polypeptide Lqh6 was reduced, then alkylated as described previously [15] The polypeptide Lqh7 was reduced for 30 at 70 °C under argon in 0.1 M ammonium bicarbonate pH 8.3 containing M guanidinium chloride and 0.1 M dithiothreitol After cooling the reaction mixture to °C, M iodoacetamide was added to a M final concentration and alkylation was performed for 45 under argon in the dark After desalting by RP-HPLC on a C8 column, the alkylated polypeptides were freeze-dried [15] 2-(2¢-Nitrophenylsulfonyl)3-methyl-3¢-bromoindolene–Skatole cleavage Carboxamidomethylated Lqh6 (4.5 nmol) dissolved in 100 lL of 75% acetic acid, was cleaved with a 10-fold molar excess of 2-(2¢-nitrophenylsulfonyl) 3-methyl-3¢-bromoindolene–Skatole (Pierce) at 37 °C for 24 h in the dark The cleavage products were separated by RP-HPLC on a Vydac C18 column (200 · 2.1 mm) using a linear gradient of acetonitrile/0.1% trifluoroacetic acid from to 30% for 90 at a flow rate of 0.1 mLỈmin)1 Chymotryptic hydrolysis Carboxamidoethylated Lqh7 (150 nmol) dissolved in 200 lL of 0.1 M ammonium acetate buffer pH 5.0 was hydrolysed with chymotrypsin and the hydrolysate was fractionated by RP-HPLC [15] Sequence analysis Amino acid analyses were performed on a Beckman 6400 amino acid analyzer Digestion of carboxamidomethylated Lqh6 with carboxypeptidase P was performed as in [15] Carboxamidoethylated Lqh7 was hydrolyzed with carboxypeptidase A in 0.1 M ammonium bicarbonate, pH 8.0, at 37 °C for h using an enzyme/substrate ratio of : 25 (w/w) In both experiments, the released amino acids were analyzed on the amino acid analyzer In vivo bioassays Fifty percent lethal doses (LD50) were established as described previously [16] The antimammal activity was tested by intracerebroventricular (i.c.v.) or subcutaneous (s.c.) injections into C57 BL/6 black mice (20 ± g body weight) Anti-insect activity was evaluated in cockroaches (Blatella germanica, 50 ± mg) using an automatic microsyringe from the Burker Manufacturing Company (Rickmansworth, UK) Binding experiments Neuronal membrane preparations All buffers contained a cocktail of proteinase inhibitors composed of: phenylmethanesulfonyl fluoride (50 lgỈmL)1), pepstatin A (1 lM), iodoacetamide (1 mM) and 1,10-phenanthroline (1 mM) Insect synaptosomes were prepared from whole heads of adult cockroaches of Periplaneta americana, according to a previously described method [17,23] Rat brain synaptosomes were prepared from adult albino Sprague–Dawley rats ( 300 g, laboratory bred), according Ó FEBS 2002 3922 A Hamon et al (Eur J Biochem 269) to the method described by Kanner [18] No loss of binding activity was observed after at least months at )80 °C [26] Membrane protein concentration was determined using a Bio-Rad Protein Assay, using BSA as standard Iodination of Lqh2, Lqh3 and Lqh6 The three toxins were radioiodinated by Iodogen (Pierce, Rockford, USA) using lg toxin and 0.5 mCi carrier-free Na125I (Amersham, U.K) and the monoiodotoxins were purified using an analytical Vydac RP-HPLC C18 column, as previously described [12,19] The concentration of the radiolabeled toxin was determined according to the specific activity of the 125 I corresponding to 2500–3000 d.p.m.Ỉfmol)1 of monoiodotoxin, depending on the age of the radiotoxin and by estimation of its biological activity (usually 50–70%; [19]) Binding assays Standard binding medium composition was (in mM): choline Cl 130, CaCl2 1.8, KCl 5, MgSO4 0.8, Hepes 50; Glucose 10, and mgỈmL)1 BSA Wash buffer composition was (in mM): Cl, 140; CaCl2, 1.8; KCl, 5.4; MgSO4, 0.8; Hepes, 50; mgỈmL)1 BSA, pH 7.5 Membrane preparations (rat brain synaptosomes, cockroach neuronal membranes) were suspended in 0.2 mL binding buffer, containing iodinated toxins After incubation for 20 (for rat brain synaptosomes) or 60 (for insect membranes), the reactions were terminated as previously described [19] Nonspecific toxin binding was determined in the presence of high concentration of the unlabeled toxin, as specified in figure legends, and consisted typically of 5% (for Lqh2) and 15% (for Lqh3 and Lqh6) of total binding Equilibrium competition and cold saturation assays were performed using increasing concentrations of the unlabeled Lqh toxins in the presence of a constant low concentration of 125I-labeled toxins Cold-saturation experiments were analyzed by the iterative program LIGAND (Elsevier Biosoft, Cambridge, UK) using ÔCold SaturationÕ analysis Competition binding experiments were analyzed using the computer program KALEIDAGRAPH (Synergy Software, Reading, USA) using a nonlinear Hill equation (for IC50 determination) The Ki were calculated by the equation Ki ¼ IC50/ (1 + (L*/Kd)), where L* is the concentration of hot toxin and Kd is its dissociation constant [20] Electrophysiological recordings from Xenopus oocytes expressing cloned mammal Na channels Ovarian lobes were surgically removed from adult female Xenopus laevis and thoroughly rinsed in standard oocyte saline (SOS) composed of (in mM): NaCl, 100; KCl, 2; CaCl2, 1.8; MgCl2, 1; Hepes, 5; pH 7.5 Stage V–VI oocytes were isolated by digestion with mgỈmL)1 collagenase (type IA, Sigma) in calcium-free SOS for 10–15 Na channel a-subunits from rat skeletal muscle (rNav1.4 ¼ rSkM1) or rat brain (rNav1.2A ¼ rBIIA), were expressed by injecting the nucleus of defolliculated oocytes with 0.1–0.5 ng of the pGW1H/rNav1.4 construct (gift from P Backx, University of Toronto, Canada) or 0.5–1.5 ng of the pHL/rNav1.2 A construct (gift from R Dunn, Mc Gill University, Montreal, Canada) For expression of the a-subunit from human heart (hNav1.5 ¼ hH1), the cytoplasm of oocytes was injected with 50 ng of cRNA transcribed in vitro from the pSP64T/hNav1.5 construct (gift from R.G Kallen, University of Pennsylvania, Philadelphia, USA) after linearization with SpeI As the main objective of our studies was to compare the effects of toxins on various channel subtypes, all Na currents were mediated by expression of the a-subunit alone After injection, oocytes were stored at 20 °C in a sterile medium consisting of SOS supplemented with gentamycin (50 lgỈmL)1), penicillin (100 mL)1), streptomycin (100 lgỈmL)1), sodium pyruvate (2.5 mM) and horse serum (1–5%) One to eight days after injection, oocytes were tested for Na channel expression using a two-electrode voltage clamp amplifier (Geneclamp 500, Axon Instruments) Each oocyte was retained by fine pins in a 100 lL chamber superfused with SOS and impaled with two glass microelectrodes filled with M KCl Electrode resistance was 2–8 MW for voltage recording electrodes and 0.7–1.2 MW for current passing electrodes The voltage dependence of Na currents, before and after application of toxins, was studied by eliciting 50 ms voltage pulses (from )60 to +30 mV in or 10 mV increments) from a holding potential (HP) of )100 mV at a frequency of 0.1 Hz (rNav1.2A, rNav1.4) or from a HP of )90 mV at a frequency of 0.2 Hz (hNav1.5) Peak current amplitudes were measured after digital subtraction of leak and capacitive currents Toxins were dissolved in SOS in the presence of BSA (0.25 mgỈmL)1) Due to the small amounts available, all toxins (20 lL) were added directly to the bathing medium (80 lL); BSA was also added to the saline (SOS) at a concentration of 0.25 mgỈmL)1 in order to reduce the nonspecific adsorption of toxins to the walls of the recording chamber Molecular modeling Lqh6 and Lqh7 were modeled using version 4.0 of the program [21] and the closest structure to the average of the NMR ensemble of Lqh3 (Protein Data Bank entry 1BMR [17]); as a template structure, using the sequence alignment shown in Fig 1A Hydrogens were added using the MOLMOL program [22] Histidine side chains were considered as fully positively charged (for a model of Lqh3 with all histidines considered as neutral; [19]) The simple charge electrostatic potentials associated to the water-accessible surfaces of the different scorpion toxins were calculated and displayed using MOLMOL MODELLER RESULTS Purification and sequence analysis of two new toxins from L quinquestriatus hebraeus By fractionation on C18 Nucleosil column, the ionexchange fractions and from the venom yielded the new toxins Lqh6 (molecular mass: 6793 Da) and Lqh7 (molecular mass: 6821 Da), respectively The toxins were obtained with a high degree of purity as assessed by capillary electrophoresis, mass spectrometry, amino acid analysis and direct sequencing of the proteins They are characterized by a high content of glycine and a low amount of aromatic residues The data obtained from the automated Edman degradation of the alkylated toxins allowed the positive identification of the first 53 and 50 amino acid residues of Lqh6 and Lqh7, respectively The remainder of Ó FEBS 2002 Two new a-like toxins (Eur J Biochem 269) 3923 Fig Comparison of the amino-acid sequence of Lqh6 and Lqh7 with other scorpion toxins that affect sodium current inactivation (A) Sequences are aligned with cysteine residues and deletions are introduced for maximum accuracy The toxins are classified in three groups according to their structural homologies and phylogenic preference The first group corresponds to classical a-toxins that are highly active on mammals, the second group to a-toxins highly active on insects and the third group to the a-like toxins that are similarly active on mammals and insects The secondary structures are indicated on top (BB, b sheet; TT, turn; HH, a helix) Arrows show residues conserved in all toxins presented Note that the underlined 8–10 residues within the 8–12 turn are very different in the three groups of toxins The numbers at the right of toxin names correspond to the global electrostatic charge calculated from the number of charged residues in the sequence (K and R ¼ +1, H ¼ +0.5, E and D ¼ )1) Asterisks indicate amidation at the C-terminus (B) Percentage of identical residues calculated for maximal homology between each pair of protein sequences Light and dark grey backgrounds indicate 45–60% and 65–95% identical residues, respectively the sequence of Lqh6 was established from the sequence data provided by the C-terminal peptide obtained by cleavage at the tryptophan residue (position 45) with 2-(2¢-nitrophenylsulfonyl)3-methyl-3¢-bromoindolene–Skatole The C-terminal sequence of Lqh7 was determined from the sequence analysis of the chymotryptic peptide obtained by hydrolysis of the carboxamidoethylated toxin with chymotrypsin at pH 5.0 At this pH, the specificity of chymotrypsin is restricted to the C-terminus of aromatic residues Ó FEBS 2002 3924 A Hamon et al (Eur J Biochem 269) (Lqh3 [15], and Bj-xtrIT [23,24], respectively) was studied in the presence of increasing concentrations of the two new toxins Up to 10 lM, Lqh6 and Lqh7 not displace the binding of 125I-labeled Bj-xtrIT, an excitatory anti-insect selective toxin (site-4) [23] On the other hand, the binding of the a-like toxin 125I-labeled Lqh3 (site 3), was inhibited by Lqh6 and Lqh7 as well as by the insect a-toxin, LqhaIT, and by the a-like toxins, Bom3 and Bom4 (Fig 2A, see also [17,19]) These results show that the new toxins, Lqh6 and Lqh7, bind receptor site and not site on cockroach Na channels These data were confirmed by experiments with radio-labeled Lqh6 (Fig 2B) The binding of 125I-labeled Lqh6 is competitively inhibited by LqhaIT (Ki ¼ 0.98 ± 0.04 nM), Lqh7 (Ki ¼ 1.61 ± 0.2 nM) and Lqh3 (Ki ¼ 0.42 ± 0.08 nM) These values are very close to those found for competition with the binding of 125I-labeled Lqh3 The Scatchard representation (Fig 2B, inset) derived from the cold saturation of Lqh6 indicates a single, high affinity binding site with a Kd of 3.44 ± 0.5 nM (n ¼ 3) and a Bmax of ± pmolỈmg)1 of protein The affinity of Lqh6 is in accordance with its Ki (4.35 ± 0.8 nM) and the capacity of its binding sites is in the range of values reported previously for a-toxins in cockroach neuronal membranes [12,19,25] The complete amino acid sequence of the two toxins (Fig 1A, bottom) is in good agreement with the data provided by amino acid analysis and mass spectrometry The difference between the measured mass of Lqh6 and the calculated molecular mass indicates amidation of the C-terminal amino acid residue The absence of a free a-carboxyl group was confirmed by the failure of carboxypeptidase P to release any amino acid from this toxin In contrast, Lqh7 has a free a-carboxyl group as shown by the release of histidine upon digestion of the toxin with carboxypeptidase A Comparison of primary structures of representative a-mammal, a-insect and a-like scorpion toxins with the sequence of the new toxins reveals that 18 amino-acid residues (vertical arrows at the bottom of Fig 1A) are present at identical positions Lqh6 and Lqh7 share eight additional residues with a-insect and a-like toxins but the highest sequence similarity (> 80%) is with Lqh3, a representative a-like toxin [15,17] The 8–10 residues, which differ strikingly in the three groups of a-toxins, are QPE and KPE in Lqh6 and Lqh7, respectively, as in other a-like toxins (underlined residues in Fig 1A) Lqh6 and Lqh7 exhibit four histidines at the same positions as in Lqh3 (in bold in Fig 1A), but Lqh7 possesses an additional histidine at position 14 As the binding of Lqh3 to insect Na channels is strongly influenced by protonation or deprotonation of its histidine residues upon pH variations [19], the same is to be expected for the two new toxins Lack of binding of Lqh6 and Lqh7 to rat brain synaptosomes despite toxicity in mice brain When injected to mice brain (i.c.v.), toxicity of Lqh6 and Lqh7 is quite comparable to that of anti-insect a-toxins but notably lower than that of the other a-like toxins (Table 1, Fig 3) However, Lqh6 and Lqh7 are not able to compete for the high affinity binding of the classical a-toxin, Lqh2, to rat brain synaptosomes, not even at 30 lM (Lqh6) or 60 lM (Lqh7) (Table 1) In order to see whether the new toxins could bind on another site than that targeted by Lqh2, we have tried to measure the direct binding of 125I-labeled Lqh6 to rat brain synaptosomes, under polarized and depolarized Binding of Lqh6 and Lqh7 to site-3 on insect Na channels The toxicity to insects of Lqh6 and Lqh7 was investigated in adult cockroaches (Blatella germanica) LD50 values are 34.3 and 28.7 pmolỈg)1 of insect for Lqh6 and Lqh7, respectively, which is in the range found for other a-like toxins (19.7–52.6 pmolỈg)1; Table 1) To determine the receptor site that is targeted by Lqh6 and Lqh7 on insect Na channels, the binding of toxins representing a- or b-classes Table Activity of some scorpion a-toxins on mammals and insects LD50 (pmolỈg)1) Ki (nM) Mice (s.c) d Aah2 Lqh2 Lqq5 1.7 8.8 3.4 a Lqh3 Bom3 Bom4 Lqh6 Lqh7 23 b 19.7 5.5 14.2 42.9 Toxins Classical a-toxins a-Like toxins a-Insect toxins LqhaIT Lqq3 8.3 6.9 b,g a a a a c (i.c.v) e 0.004 0.014 0.018 0.36 0.16 0.16 24 7.9 7.9 b a b a a Cockroach a b a 897 280 2317 c 0.2 0.4 1a b c 28 b 52.6 19.7 34.3 28.7 2.5 8.3 Rat brain j ([125I]a-toxin) c c c c a 58.8 16 k 120 c b 2004 >10 >10 >30 >60 Cockroach l ([125I]a-toxin) f,i 000 000 000 000 2760 f,i 700 h a a f f c 0.43 b 29.3 c 4.6 c 12 k 6.4 k 0.02 0.03 c c a [47] b [15] c [12] d Subcutaneous injection to mice e intracerebroventricular injection to mice f Competition for 125I-labeled Lqh2 binding to rat brain synaptosomes g LD50 determined in Swiss white mice h [48] I [6] j Competition for 125I-labeled Aah2 or 125I-labeled Lqh2 binding to rat brain synaptosomes k Competition for 125I-labeled Lqh3 binding to insect neuronal preparation l Competition for 125 I-labeled LqhaIT binding to insect neuronal preparation Ó FEBS 2002 Two new a-like toxins (Eur J Biochem 269) 3925 Fig Binding interaction of a-toxins in cockroach neuronal membranes (A) Competition curves for 125I-labeled Lqh3 binding inhibition by a-like toxins Cockroach neuronal membranes (16.3 lg/mL) were incubated for 60 at 22 °C with 180 pM 125I-labeled Lqh3 and increasing concentration of the indicated toxins Nonspecific binding, determined in the presence of lM of Lqh3, was subtracted The amount of 125I-labeled Lqh3 bound is expressed as the percentage of the maximal specific binding without additional toxin Competition curves are fitted by the nonlinear Hill equation (with a Hill coefficient of 1) to determine the IC50 values (See Experimental procedures) The Ki values are (in nM, n ¼ number of experiments): LqhaIT, 0.7 ± 0.5, n ¼ 3; Lqh3, 1.93 ± 0.90, n ¼ 5; Bom4, 5.3 ± 1, n ¼ 2; Lqh7, 6.4 ± 0.8, n ¼ 3; Lqh6, 11.9 ± 3, n ¼ 3; Bom3, 12.3 ± 4, n ¼ (B) Competition curves for 125I-labeled Lqh6 binding inhibition by various toxins Cockroach neuronal membranes (16.3 lgỈmL)1) were incubated for 60 at 22 °C with 200 pM of 125I-labeled Lqh6 and increasing concentration of the indicated toxins Nonspecific binding, determined in the presence of lM LqhaIT, was subtracted The amount of 125I-labeled Lqh6 bound is expressed as the percentage of the maximal specific binding without additional toxin The competition curves were fitted by the nonlinear Hill equation (with a Hill coefficient of 1) to determine the IC50 values (See Experimental procedures) The Ki values are, in nM: LqhaIT, 0.98 ± 0.04, n ¼ 2; Lqh3, 0.42 ± 0.08, n ¼ 3; Lqh7, 1.61 ± 0.2, n ¼ 2; Lqh6, 4.35 ± 0.8, n ¼ Inset, Scatchard transformation of the competition curve of 125I-labeled Lqh6 by increasing concentration of Lqh6 (cold saturation) The equilibrium binding parameters were calculated by the program Ligand and are Kd ¼ 3.44 ± 0.5 nM, Bmax ¼ ± pmolỈmg)1 of protein, n ¼ membrane potential conditions No specific binding was detected, even when the conditions were set for optimal binding of a-toxins (not illustrated) [26] The discrepancy between toxicity in whole brain and lack of binding to synaptosomes is in agreement with previous data obtained with other a-like toxins, such as Bom3, Bom4 [25,27] and Lqh3 [6,10] To further examine the channel targeted by Lqh6 and Lqh7 in rat brain, we examined directly their effects on rNav1.2 A, one of the most abundant Na channel subtypes expressed in rat brain [28,29] Effects of Lqh6 and Lqh7 on rat brain Na channels (rNav1.2 A) The functional effects of Lqh6 and Lqh7 on rNav1.2A channels expressed in Xenopus oocytes were compared to those of Aah2, a classical a-toxin [30] Addition of Aah2 to the bath medium at a saturating concentration (0.1 lM) induced a progressive slowing of the inactivation kinetics of Na currents (Fig 4A) After stabilization (2–5 min), the current measured at the end of 50 ms pulses to )10 mV 3926 A Hamon et al (Eur J Biochem 269) Ó FEBS 2002 could reach 25–30% of the peak Current peak amplitudes were also markedly increased; for depolarization to )10 mV, the increase was typically 130–170% in the presence of 0.1 lM Aah2 In contrast, no effect could be observed with Lqh6, even at very high concentrations (10 lM) Lqh7 was also inactive at concentrations up to lM, but at higher concentrations (3–10 lM), a slight dosedependent effect on the inactivation rate and on the peak current amplitude was observed (Figs 4B,C) Re-examination of the Lqh7 samples by mass-spectrometry confirmed that no contaminants were present, thus confirming the differences in effects between Lqh6 and Lqh7 Effects of Lqh6 and Lqh7 on muscle Na channels (rNav1.4 and hNav1.5) Fig Binding interaction to rat brain synaptosomes Competition for 125 I-labeled Lqh2 binding by Lqh2, Lqh6 and Lqh7 toxins Rat brain synaptosomes (64.8 lg proteinỈmL)1) were incubated 20 at room temperature with 110 pM 125I-labeled Lqh2 and increasing concentration of the indicated toxins Non-specific binding, determined in the presence of 200 nM Lqh2, was subtracted The amount of 125I-labeled Lqh2 bound is expressed as the percentage of the maximal specific binding without additional toxin The competition curves were fitted by the nonlinear Hill equation (with a Hill coefficient of 1) to determine the IC50 values (See Experimental procedures) The Ki of Lqh2 is 0.18 ± 0.06 nM, n ¼ [6], but neither Lqh6 nor Lqh7 can displace 125 I-labeled Lqh2, even at 30 lM (Lqh6) or 60 lM (Lqh7) The effects of classical a- and a-like scorpion toxins by s.c injection are partly mediated by muscular Na channels [6, 10] To clarify whether Lqh6 and Lqh7 affect these channels, the two toxins were tested on rat skeletal muscle (rNav1.4) and human heart (hNav1.5) sodium channels expressed in Xenopus oocytes Effects on action potentials One or two days after nuclear injection of the cDNA encoding rNav1.4, action potentials (APs) could be recorded from most oocytes by removing the voltage-clamp (Vh ¼ )100 mV) and allowing the cell membrane to recover its spontaneous resting potential Only oocytes expressing at least 1–1.5 lA of peak Na current were able to generate APs under these conditions These APs had (a) a threshold close to )40 mV; (b) an amplitude up to 90 mV, which varied among cells (about 80 mV in Fig 5A); (c) a duration at mid-peak of Fig Compared effects of a-like toxins and Aah2 on rNav1.2 A channels expressed in Xenopus oocytes (A,B) Families of currents recorded before (left traces) and after (right traces) application of 0.1 lM Aah2 (A) or 10 lM Lqh7 (B) Currents were evoked by depolarizing test pulses of 50 ms duration from a holding potential (HP) of )100 mV Traces displayed are from )40 to +30 mV (C) Semi-logarithmic dose–response curves of the maintained Na+ current induced by Aah2 and a-like toxins Currents were evoked using test pulses to )10 mV from a HP of )100 mV and measured at the end of 50 ms pulses After subtraction of the current obtained in control conditions (when present), the toxin-modified current was plotted as a percentage of the peak current Points are mean values for 4–8 cells Ó FEBS 2002 Two new a-like toxins (Eur J Biochem 269) 3927 Fig Functional effects of a-like toxins on rNav1.4 (A–D) and hNav1.5 (E,F) expressed in Xenopus oocytes (A) Modification of the spiking activity of oocytes expressing rNav1.4 by lM Lqh6 The cell membrane was depolarized by interrupting the voltage-clamp (HP ¼ )100 mV) The toxin induced a very slow repetitive activity and a prolongation of the repolarizing phase of spikes (B) Modification of rNav1.4 currents by lM Lqh6 The current was evoked by depolarizing test pulses to )10 mV from a HP of )100 mV Note the slowing of the late phase of inactivation and the increase in peak amplitude (C) Compared semilogarithmic dose–response curves of rNav1.4-maintained currents induced by Aah2 and a-like toxins The experimental procedure was the same as in Fig 4C SEM have been omitted for clarity on the curves relative to the a-like toxins (D) Normalized peak conductance-voltage relationship (gNa/gNa max vs Vm) before (empty circles) and after (filled circles) modification by lM Lqh6 gNa was calculated according to the equation: gNa ¼ INa/(Vm ) ENa) where INa is the peak current amplitude, Vm the test pulse potential and ENa the equilibrium potential for Na+ ions The half-activation voltage was shifted by about mV in the presence of Lqh6 The data represent the mean ± SEM of five experiments (E) hNav1.5 currents were elicited by step depolarizations from )90 to mV and recorded in control (d) and after different times of incubation with 0.2 lM Lqh3 (5, 10 and 35 min) Note that the initial fast phase of inactivation was not affected by the toxin (F) Modification of the slow inactivation time constant (tau 2) of hNav1.5 currents by a-like toxins tested at 0.2 lM Currents were elicited as in (E) Ó FEBS 2002 3928 A Hamon et al (Eur J Biochem 269) 0.48 ± 0.1 s (n ¼ 9) and they were completely blocked by 0.5 lM tetrodotoxin (data not shown) To our knowledge, this is the first report on the occurrence of APs in these cells in the absence of any Na channel modifier Addition of a classical a-toxin (Aah2) or a-like toxin to the bath solution induced a dose-dependent increase in the duration of APs due to prolongation of their repolarizing phase (illustrated for Lqh6 in Fig 5A) Moreover, cells that fired only one spike in controls exhibited repetitive activity in the presence of toxins All effects were slowly reversible upon washing (data not shown) These results show that the actions of drugs and toxins on Na channels reconstituted in oocytes can be studied not only on currents but also on potentials as in neuronal and muscular preparations Effects on Na currents Upon addition of 1–3 lM a-like toxin, stable effects on rNav1.4 currents were observed within 10–15 min, while only 2–3 of exposure were required with 0.1 lM Aah2 All tested toxins induced a slowing of the late phase of inactivation (Fig 5B), which explains the prolongation of action potentials and the repetitive activity observed during depolarizations It is important to note that the three a-like toxins (Lqh3, Lqh6 and Lqh7) caused a typical a-toxin effect, as was also observed for Lqh3 on frog axon [31] but in contrast to observations made with cockroach axon, where inward Ôholding currentÕ was observed in the presence of Bom4 and Lqh3 [17, 25] The toxins also increased the peak current amplitude At )10 mV, in the presence of lM Lqh6, the increase was 30–50% (Fig 5B), as compared with 100– 180% increase induced by a 10-fold lower concentration of Aah2 This effect does not result from an increase in single channel conductance [11], but may be caused by the slowing of inactivation As activation and inactivation are overlapping processes, inhibition of the latter may allow a greater peak current to be attained Oocytes expressing the cardiac channel, hNav1.5, were also responsive to all tested a-like toxins but the effects developed even more slowly than with rNav1.4 (Fig 5E) The falling phase of currents was prolonged as for rNav1.4 but the peak amplitude was not consistently modified (Fig 5E); the peak increased only for test pulses ranging between )50 and )30 mV, due to a negative shift in the voltage-dependence of activation (see below) Modification of hNav1.5 currents by Lqh3, Lqh6 and Lqh7 demonstrates that a-like toxins bind well to various muscular sodium channels while they discriminate strongly among neuronal channel subtypes As Lqh3 does not interact with tetrodotoxin-resistant Na channels in rat dorsal root ganglion (DRG) neurons [10], hNav1.5 is the first tetrodotoxinresistant Na channel that is affected by a-like toxins [32] Concentration dependence of toxin effects Quantification of the differential ability of toxins to modify the activity of rNav1.4 channels was performed using dose–response curves of the relative maintained current (i.e I50ms/Ipeak) measured for pulses to )10 mV Comparison of Figs 4C and 5C shows that Aah2 acted with nearly the same potency on rNav1.4 and rNav1.2A For both channels, the threshold for toxin effects was slightly below 0.1 nM, the EC50 value was nearly the same (rNav1.2A, 2.7 ± 0.2 nM; rNav1.4, 2.9 ± 0.3 nM) and the curves reached a maximum at a toxin concentration close to 0.1 lM These results are very similar to those obtained with native channels in rat skeletal muscles [33] All tested a-like toxins (Lqh3, Lqh6 and 7) were less potent than Aah2: their threshold was about two orders of magnitude higher and the maintained current reached only 7–12% of the peak at high concentration (3 lM) instead of > 25% for Aah2 Entire dose–response curves were not constructed for hNav1.5 but it was observed that in the presence of a-like toxins at 0.2 lM, the relative amplitude of the maintained current was always higher for hNav1.5 (10–19% of the peak) than for rNav1.4 (2–6% of the peak) The order of potency of toxins was the same for the two muscular channels (Lqh6 > Lqh3 > Lqh7) By comparison with rNav1.4 expressed in mammalian cells [11], it appears that Na channels expressed in oocytes are less sensitive to a-like toxins Indeed, in HEK cells expressing rNav1.4, the EC50 value for removal of inactivation is about nM for Lqh3, while the same concentration of toxin did not induce an effect in oocytes Effects on the voltage dependence of activation The activation curves were shifted to more negative potentials in the presence of any of the tested toxins At mid-activation, the shift for Nav1.4 ranged between 3.3 ± 0.1 mV (Lqh7) and 6.4 ± 0.3 mV (Aah2) As an example, Fig.5D illustrates the effects of Lqh6 Although Nav1.5 channels were challenged at lower concentrations of toxins than Nav1.4 (0.2 lM instead of lM), the shift was more accentuated and comprised between 6.5 mV (Lqh7) and 13 mV (Lqh6) Effects on the kinetics of inactivation The inactivation phase of hNav1.5 and rNav1.4 currents was best fitted by the sum of two decaying exponentials As already reported by several authors using the oocyte expression system, rNav1.4 currents showed slower inactivation kinetics than hNav1.5 currents, due to a shift in the equilibrium between fast and slow gating modes [34–37] Incubation of oocytes expressing hNav1.5 with 0.2 lM a-like toxins induced no consistent modification of the initial phase of inactivation (Fig 5E), while the time constant of the slow component of inactivation (tau 2) was always increased (Fig 5F) Similar effects were observed with rNav1.4: the a-like toxins (1 lM) produced a twofold to fourfold increase in the value of tau2 (control: 16.1 ± 0.4 ms, n ¼ 5), while tau1 (4.8 ± 0.9 ms, n ¼ ) was not consistently modified DISCUSSION The scorpion a-like toxins have been initially defined by their ability to kill insects and mice by both central (i.c.v.) and peripheral (s.c.) injection and to present a specific binding on insect Na channels but not on rat brain synaptosomes [12,27] The present work establishes that Lqh6 and Lqh7 can be classified into this group of toxins A comparison of their pharmacological and structural properties with those of other a-toxins offers the opportunity to re-examine the criteria that characterize the a-like toxins as a group Pharmacological characteristics of a-like toxins A major difference between a-like toxins and other scorpion a-toxins appears upon comparison of their toxicity to mice by i.c.v injection with their ability to inhibit the binding of Ó FEBS 2002 labeled Lqh2 or Aah2 to rat brain synaptosomes (Fig 6) For antimammal and anti-insect a-toxins, the data points showing the relationship between LD50 and Ki values appear roughly aligned on a logarithmic scale, suggesting that their toxic effects in whole brain is in correlation with their binding to brain nerve terminals In contrast, the Ki determined for Lqh3 is higher by about two orders of magnitude than the value expected from its toxicity by i.c.v The other a-like toxins, including Lqh6 and Lqh7, are unable to displace Lqh2 at the highest concentration tested (Ki values out of the graph) As rat brain synaptosomes express mainly rNav1.2 and rNav1.1 Na channels [28], it can be concluded that these two channel subtypes are not targeted by the a-like toxins, which is confirmed for subtype II by electrophysiological recordings of rNav1.2A expressed in Xenopus oocytes or HEK293 cells ([6,10]; present study) The toxicity of a-like toxins in the brain is mediated by other Na channel subtypes, which are expressed presumably on neuronal cell bodies [6] It may be tempting to classify a-like toxins in the same group as anti-insect a-toxins, as presented in some recent reviews [9,38] Indeed, toxins of these two groups are toxic to both mammals and insects and Lqh3, the most studied a-like toxin, exhibits about the same potency as LqhaIT in competing for Aah2 binding to rat brain synaptosomes (Table 1) However, an important difference precludes the amalgamation of the two groups: whereas anti-insect and antimammal toxins show the same correlation between toxicity and binding inhibition properties in rat brain (Fig 6), a-like toxins not follow this correlation, suggesting that they not target the same channel subtype in the brain as anti-insect toxins In addition, toxins of the two groups also differ by the potency of binding to insect Na channels All anti-insect toxins inhibit the binding of [125I]LqhaIT at low concentrations, whereas Two new a-like toxins (Eur J Biochem 269) 3929 14–1400 higher concentrations of a-like toxins are required to obtain the same level of inhibition This is correlated with the higher toxicity of anti-insect toxins to cockroaches (Table 1) Whereas a-like toxins not affect the brain Na channel Nav1.2 (in contrast to the antimammal a-toxins), they are most active on the cardiac Na channel subtype (more than the antimammal and anti-insect a-toxins, Lqh2 and LqhaIT, respectively) [39] and also affect the skeletal muscle Na channel (Fig [11]) Despite the similar effect on channel inactivation, the association and dissociation kinetics were shown to differ among the three a-toxin groups [11], further supporting the differences among them Structural characteristics of a-like toxins The existence of the a-like toxins as a distinct group is further strengthened by structural analysis All scorpion toxins affecting Na current inactivation are a homologous class of polypeptides with highly similar 3D structure ([17, 40–43]; Fig 7A) A sequence comparison between toxins isolated from the venom of Leiurus quinquestriatus hebraeus show that Lqh6 and Lqh7 share more than 80% identical residues with Lqh3, the representative a-like toxin, whereas the percentage of identity is only about 50% with LqhaIT and less than 40% with classical a-toxins such as Lqh2 or Aah2 (Fig 1B) Among the several loops and turns connecting the conserved secondary structure elements (Fig 7A), the fiveresidue turn (8–12) and the C-terminal tail have been shown by site-directed mutagenesis of LqhaIT to be functionally important [44,45] The five-residue turn contains two residues (N11 and C12) that are shared by all a-toxins, while the three others (8–10) are suggested to Fig Relationship, on logarithmic scale, between the toxicity of various scorpion a-toxins by i.c.v injection in mice and their Ki values determined from competitive binding assays to rat brain synaptosomes, for binding of classical a-toxins (125I-labeled Aah2 or 125I-labeled Lqh2) Note that for classical and insect a-toxins, a straight line can be drawn through the data points Only a-like toxins are up to this line, showing much more activity in the brain than on synaptosomes The vertical arrows mean that Ki values are up to the data points representing the highest concentrations tested on synaptosomes without any displacement of the classical a-toxins Ó FEBS 2002 3930 A Hamon et al (Eur J Biochem 269) form a molecular signature typical of each of the three groups of a-toxins The 8–10 residues of Lqh6 and Lqh7 are Q/KPE, similar to the motif found in other a-like toxins and in contrast with the signature found in classical a-(DDV/K) and insect a-toxins (KNY) The uniqueness of the Q/KPE/H sequence is further emphasized by the presence of P9-E/H10cis peptide bond, identified in the crystal structure of the a-like toxins Bmk M1, M2 and M4 [42, 43] We have also found that Lqh3 in solution is in two conformers, which are compatible with a major form containing a P9-E10 cis peptide bond (PDB entry 1FH3, I Krimm, X Trivelli & J.M Lancelin, unpublished results), which is in slow exchange with a minor form due to P9-E10 trans peptide bond (PDB entry 1BMR [17]) It is very likely, but not yet demonstrated, that the cis–trans peptide bond isomerism may also apply to Lqh6 and Lqh7, but its functional role in the a-like toxins is yet to be demonstrated Diversity of a-like toxins Although all a-like toxins share important structural and functional properties that differentiate them from classicalor anti-insect a-toxins, they also differ from each other in their pharmacological profile (Table 1) These differences further indicate that members of the a-toxins class demonstrate gradual functional and structural diversification, attributed to nonhomologous substitutions located on toxin Fig Structure and electrostatic potentials of the water accessible surfaces of three a-like toxins (A) Structural skeleton of Lqh3, the most studied a-like toxin (1BRM and 1FH3 PDB entries [17], and unpublished results) Elements of secondary structures and the five-residue turns 8–12 where a cis-trans isomerism of the peptide bond between P9 and E10 takes place, are indicated Due to the local structural differences between the two isomers, this does not affect the simple charge electrostatic gross representation of the a-like toxins as shown in (B–F) Ct indicates the C-terminus (B–D) Simple charge potentials associated to the water-accessible molecular surfaces of a-like toxins calculated using MOLMOL [22] Surfaceassociated electrostatic potentials are represented from electronegative to electropositive by a yellow to blue continuous color range, respectively Molecules are shown in the same orientation as in (A) Ct, C-terminus Ó FEBS 2002 surfaces, which endow them with the observed pharmacological diversification Among the surface elements implicated in toxin–channel interactions, charged residues have been suggested to play a role in bioactivity of both anti-insect [44] and a-like toxins [19] Figure illustrates the differences in the electrostatic charge distribution of Lqh6, Lqh7 and Lqh3 Lqh7, the most positively charged of the a-like toxins is characterized by a very large electropositive patch including the C-terminus (K64, H66) and residues that are brought close to the C-terminus by the tertiary arrangement of the polypeptide chain (K8, H14, H15, H43) [17,41] In Lqh6, the corresponding positive patch is much less developed and is neutralized by numerous negative residues Lqh3 is also a neutral toxin, but with a very large electropositive surface including the C-terminus, similarly to Lqh7 A major difference between Lqh7 and Lqh3, however, is the presence of two negative residues (E61 and E63) within the C-terminal stretch of the latter toxin The functional significance of these differences in surface electrostatics is presently obscure and significant advances will probably await functional expression of recombinant a-like toxins and subsequent sitedirected modification of the charged residues Another level of structural heterogeneity is revealed by sequence comparisons by which both calculation of percentage of identity (Fig 1B) and computer sequence analysis lead to the conclusion that a-like toxins can be clustered into two distinct subgroups, designated A and B in the consensus phylogenetic tree presented in Fig In subgroup A (Lqh3, Lqh6, Lqh7, Bom3), homology with LqhaIT is only 47–50%, while in subgroup B (Bom4, Bmk M1, M2 and M4) it reaches 73–81% Subgroups A and B can be further discriminated by their differential homology with classical a-toxins (A: 30–38%, B: 50–52%) Moreover, Two new a-like toxins (Eur J Biochem 269) 3931 the degree of sequence homology is much higher within subgroups (A: 65–86%, B: 75–95%) than between subgroups ( 50%) Although the two subgroups are clearly individualized, they share structural peculiarities such as the QPE or KPE/H signature of turn 8–12 and the tentative nonproline cis peptide bond between P9 and E/K10 In summary, we have shown that Lqh6 and Lqh7 can be classified into the a-like group of scorpion a-toxins on the basis of both pharmacological and structural arguments, and used a comparative approach to discuss the features that differentiate this type of toxins from classical- and antiinsect a-toxins Long ignored as a group, a-like toxins are abundant in scorpion venoms and the present study emphasizes the importance of thorough comparative pharmacological characterization of new scorpion toxins Recently, the interest in a-like toxins has been boosted by the discovery of their ability to specifically target a somatic Na channel subtype in the mammalian brain, thus revealing subtle variations in the structure of receptor site of neuronal Na channels As mutant Nav1.4 channels with Nav1.6 receptor site motif are highly sensitive to Lqh3, it has been suggested that Nav1.6 might be the target of a-like toxins in the brain ([46], but see [7]) The present challenge is to identify this target unequivocally, with the perspective to get some information about its physiological roles by using a-like toxins as molecular tools ACKNOWLEDGEMENTS We thank Peter Backx (University of Toronto, Canada) for the pGW1H/rNav1.4 construct, Robert Dunn (McGill University, Montreal, Canada) for the pHL/rNav1.2 A construct and Roland G Kallen (University of Pennsylvania, Philadelphia, USA) for the pSP64T/ ´ hNav1.5 cDNA construct We are also grateful to Herve Rochat ´ (Universite de Marseille, France) for gift of purified Aah2 and to ´ Marcel Pelhate (Universite d’Angers, France) for helpful discussion This work was partly supported by grants from the Israeli Science Foundation (nos 508/00 and 733/01) to D G., and by a grant from BARD, The United States-Israel Binational Agricultural Research & Development (IS-3259-01) to D.G REFERENCES Fig Unrooted phylogenetic tree of scorpion a-toxins Fourteen aligned sequences from Fig 1A were analyzed by a parsimony method, using the PROPARS program of the PHYLIP package A consensus tree was generated following bootstrap analysis and the final diagram was plotted using the DRAWTREE program Gordon, D (1997) Sodium channels as targets for neurotoxins: mode of action and interaction of neurotoxins with receptor sites on sodium channels In Toxins and Signal Transduction (Lazarowici, P & Gutman, Y., eds), pp 119–149, Amsterdam: Harwood ` Cestele, S & Catterall, W.A (2000) Molecular mechanisms of neurotoxin action on voltage-gated sodium channels Biochimie 82, 883–892 Novakovic, S.D., Eglen, R.M & Hunter, J.C (2001) Regulation of Na+ channel distribution in the nervous system Trends Neurosci 24, 473–478 Whitaker, W.R.J., Faull, R.L.M., Waldvogel, H.J., Plumpton, C.J., Emson, P.C & Clare, J.J (2001) Comparative distribution of voltage-gated sodium channel proteins in human brain Mol Brain Res 88, 37–53 Caldwell, J.H., Schaller, K.L., Lasher, R.S., Peles, E & Levinson, S.R (2000) Sodium channel Na(v),1.6 is localized at nodes of Ranvier, dendrites and synapses Proc Natl Acad Sci USA 97, 5616–5620 Gilles, N., Blanchet, C., Shichor, I., Zaninetti, M., Lotan, I., Bertrand, D & Gordon, D (1999) A scorpion a-like toxin that is active in insects and mammals reveals an unexpected specificity of Ó FEBS 2002 3932 A Hamon et al (Eur J Biochem 269) 10 11 12 13 14 15 16 17 18 19 20 21 22 23 distribution of sodium channel subtypes in rat brain neurons J Neurosci 19, 8730–8739 Gordon, D., Gilles, N., Bertrand, D., Molgo, J., Nicholson, G.M., Sauviat, M.P., Benoit, E., Shichor, I., Lotan, I., Gurevitz, M., Kallen, R.G & Heinemann, S.H (2002) Scorpion toxins differentiating among neuronal sodium channel subtypes: nature’s guide for design of selective drugs In Perspectives in Molecular Toxinology, Chapter 12 (Menez, A., ed.), Wiley and Sons, Chichister, England, in press Gordon, D., Savarin, P., Gurevitz, M & Zinn-Justin, S (1998) Functional anatomy of scorpion toxins affecting sodium channels J Toxin Rev 17, 131–159 Possani, L.D., Becerril, B., Delepierre, M & Tytgat, J (1999) Scorpion toxins specific for Na+-channels Eur J Biochem 264, 287–300 Gilles, N., Chen, H., Wilson, H., Le Gall, F., Montoya, G., Molgo, J., Schonherr, R., Nicholson, G., Heinemann, S.H & ă Gordon, D (2000) Scorpion a and a-like toxins differentially interact with sodium channels in mammalian CNS and periphery Eur J Neurosci 12, 2823–2832 Chen, H., Gordon, D & Heinemann, S.F (2000) Modulation of cloned skeletal muscle sodium channels by the scorpion toxins Lqh2, Lqh3 and LqhaIT Phluăgers Arch 439, 423432 ` 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 Marshall, D.L., Vatanpour, I.H., Harvey, A.L., Boyot, P., Pinkasfeld, S., Doljansky, Y., Bouet, F & Menez, A (1994) Neuromuscular effects of some potassium channel blocking toxins from the venom of the scorpion, Leiurus quinquestriatus hebraeus Toxicon 32, 1433–1443 Becerril, B., Marangoni, S & Possani, L.D (1997) Toxins and genes isolated from scorpions of the genus Tityus Toxicon 35, 821–835 ` ` Sautiere, P., Cestele, S., Kopeyan, C., Martinage, A., Drobecq, H., Doljanski, Y & Gordon, D (1998) New toxins acting on sodium channels from the scorpion Leiurus quiquestriatus hebraeus suggest a clue to mammalian insect selectivity Toxicon 36, 1141–1154 Behrens, B & Karber, C (1935) Wie sind reichenversuche fur biologische auswertungen am zweckmassigsten anzwordnen? Arch Exp Path Pharmak 177, 379–388 ` Krimm, I., Gilles, N., Sautiere, P., Stankiewicz, M., Pelhate, M., Gordon, D & Lancelin, J.M (1999) NMR structures and activity of a novel a-like toxin from the scorpion, Leiurus quinquestriatus hebraeus J Mol Biol 285, 1749–1763 Kanner, B.J (1978) Active transport of c-amino butyric acid by membrane vesicles isolated from rat brain Biochemistry 17, 1207– 1211 Gilles, N., Krimm, I., Bouet, F., Froy, D., Gurevitz, M., Lancelin, J.M & Gordon, D (2000) Structural implications on the interaction of scorpion a-like toxins with the sodium channel receptor site inferred from toxin iodination and pH-dependent binding J Neurochem 75, 1735–1745 Cheng, Y.C & Prusoff, W.H (1973) Relationship between the inhibition constant (Ki) and the concentration of inhibitor which causes 50 percent inhibition (IC50) of an enzymatic reaction Biochem Pharmacol 22, 3099–3108 Sali, A & Blundell, T.L (1993) Comparative protein modelling by satisfaction of spatial restraints J Mol Biol 234, 779–815 Koradi, R., Billleter, M & Wuthrich, K (1996) MOLMOL: a ă program for display and analysis of macromolecular structures J Mol Graphics 14, 51–55 Froy, O., Zilberberg, N., Gordon, D., Turkov, M., Gilles, N., Stankiewicz, M., Pelhate, M., Loret, E., Oren, D.A., Shaanan, B 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 & Gurevitz, M (1999) The putative bioactive surface of insectselective scorpion excitatory neurotoxins J Biol Chem 274, 5769–5776 Oren, D.A., Froy, O., Amit, E., Kleinberger-Doron, N., Gurevitz, M & Shaanan, B (1998) An excitatory scorpion toxin with a distinctive feature: an additional a-helix at the C-terminus and its implications for interaction with insect sodium channels Structure 6, 1095–1103 ` Cestele, S., Stankiewicz, M., Mansuelle, P., De Waard, M., Dargent, B., Gilles, N., Pelhate, M., Rochat, H., Martin-Eauclaire, M.F & Gordon, D (1999) Scorpion a-like toxins, toxic to both mammals and insects, differentially interact with receptor site on voltage-gated sodium channels in mammals and insects Eur J Neurosci 11, 975–985 Gilles, N., Leipold, E., Chen, H., Heinemann, S.H & Gordon, D (2001) Effect of depolarization on binding kinetics of scorpion a-toxin highlights conformational changes of rat brain sodium channels Biochemistry 40, 14576–14584 Vargas, O., Martin-Eauclaire, M.F & Rochat, H (1987) Characterization of six toxins from the venom of the Moroccan scorpion, Buthus occitanus mardochei Eur J Biochem 162, 589–599 Gordon, D., Merrick, D., Auld, V., Dunn, R., Goldin, A.L., Davidson, N & Catterall, W.A (1987) Tissue specific expression of RI and RII sodium channel subtypes Proc Natl Acad Sci USA 84, 8682–8686 Beckh, S., Noda, M., Lubbert, H & Numa, S (1989) Differential regulation of three sodium channel messenger RNAs in the rat central nervous system during development EMBO J 82, 3611– 3616 Martin-Eauclaire, M.F & Couraud, F (1995) Scorpion neurotoxins: effects and mechanisms In Handbook of Neurotoxicology (Chang, L.W & Dyer, R.S., eds), pp 683–716 Marcel Dekker, New York Benoit, E & Gordon, D (2001) The scorpion a-like toxin Lqh III specifically alters sodium channel inactivation in frog myelinated axons Neuroscience 104, 551–559 Goudet, C., Huys, I., Clynen, E., Schoofs, L., Wang, D.C., Waelkens, E & Tytgat, J (2001) Electrophysiological characterization of BmK M1, an a-like toxin from Buthus martensi Karsch venom FEBS Lett 495, 61–65 ´ Duval, A., Malecot, C.O., Pelhate, M & Rochat, H (1989) Changes in Na channel properties of frog and rat skeletal muscles induced by the AaH II toxin from the scorpion Androctonus australis Phlu ăgers Arch 415, 361–371 Zhou, J., Potts, J.F., Trimmer, J.S., Agnew, W.S & Sigworth, F.J (1991) Multiple gating modes and the effect of modulating factors on the lI sodium channel Neuron 7, 775–785 Gellens, M.E., George Jr, A.L., Chen, L., Chahine, M., Horn, R., Barchi, R.L & Kallen, R.G (1992) Primary structure and functional expression of the human cardiac tetrodotoxin-insensitive voltage-dependent sodium channel Proc Natl Acad Sci USA 89, 554–558 Chen, C & Cannon, S.C (1995) Modulation of Na+ channel inactivation by the b1 subunit: a deletion analysis Puăgers Arch 431, 186–195 Nuss, H.B., Chiamvimonvat, N., Perez-Garcia, M.T., Tomaselli, G.F & Marban, E (1995) Functional association of the b1 subunit with human cardiac (hH1) and rat skeletal muscle (l1) sodium channel subunits expressed in Xenopus oocytes J General Physiol 106, 1171–1191 Possani, L.D., Merino, E., Corona, M., Bolivar, F & Becerril, B (2000) Peptides and genes coding for scorpion toxins that affect ion-channels Biochimie 82, 861–868 Chen, H & Heinemann, S.H (2001) Interaction of scorpion a-toxins with cardiac sodium channels: binding properties and enhancement of slow inactivation J Gen Physiol 117, 505–518 Ó FEBS 2002 40 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 41 Tugarinov, V., Kustanovich, I., Zilberberg, N., Gurevitz, M & Anglister, J (1997) Solution structures of a highly insecticidal recombinant scorpion a-toxin and a mutant with increased activity Biochemistry 36, 2414–2424 42 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 the sodium channel J Mol Biol 292, 125–135 43 He, X.L., Deng, J.P., Wang, M., Zhang, Y & Wang, D.C (2000) Structure of a new neurotoxin from the scorpion Buthus martensii Karsch at 1.76 Acta Crystallogr Sect D Biol Crystallogr 56, 25–33 ` 44 Zilberberg, N., Froy, O., Loret, E., Cestele, S., Arad, D., Gordon, D & Gurevitz, M (1997) Identification of structural elements of a scorpion a-neurotoxin important for receptor site recognition J Biol Chem 272, 14810–14816 45 Gurevitz, M., Gordon, D., Ben-Natan, S., Turkov, M & Froy, O (2001) Diversification of neurotoxins by C-tail ÔwigglingÕ – a scorpion recipe for survival FASEB J 15, 1201–1205 46 Hansel, A., Lu, S.Q., Leipold, E & Heinemann, S.H (2002) Molecular determinants for the binding of scorpion toxins Lqh2 Two new a-like toxins (Eur J Biochem 269) 3933 and Lqh-3 to mammalian neuronal sodium channels Pu ăgers Arch 443/S02, P022 47 Martin-Eauclaire, M.F., Delabre, M.L., Ceard, B., Ribeiro, A.M., Soogard, M., Svensson, B., Diniz, C.R., Smith, L.A., Rochat, H & Bougis, P.E (1992) Genetics of scorpion toxins In Recent Advances in Toxicology Research, (Gopalakrishnakone, P & Tam, C.K., eds) pp 196–209 Venom and Toxin Research Group, National University of Singapore 48 Kopeyan, C., Mansuelle, P., Martin-Eauclaire, M.F., Rochat, H & Miranda, F (1993) Characterization of toxin III of the scorpion Leiurus quinquestriatus quinquestriatus: a new type of alpha-toxin highly toxic both to mammals and insects Nat Toxins 1, 308–312 SUPPLEMENTARY MATERIAL The following material is available from http:// www.blackwell-science.com/products/journals/suppmat/EJB/ EJB3065/EJB3065sm.htm Table S1 Amino acid composition of Lqh6 and Lqh7 toxins Fig S1 Purification of the toxins Lqh6 (A) and Lqh7 (B) from the venom of the scorpion Leiurus quinquestriatus hebraeus (Lqh) Fig S2 Amino acid sequences of toxins Lqh6 and Lqh7  ... the opportunity to re-examine the criteria that characterize the a-like toxins as a group Pharmacological characteristics of a-like toxins A major difference between a-like toxins and other scorpion. .. them Structural characteristics of a-like toxins The existence of the a-like toxins as a distinct group is further strengthened by structural analysis All scorpion toxins affecting Na current inactivation... they not target the same channel subtype in the brain as anti-insect toxins In addition, toxins of the two groups also differ by the potency of binding to insect Na channels All anti-insect toxins