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
1

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;
6
Laboratoire de RMN Biomole´culaire, CNRS UMR-5078, Lyon, France; 7Department of Plant Sciences, Tel Aviv University,
Ramat-Aviv, Tel Aviv Israeăl
3

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 do 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 3 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 3 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 3 640 6100; E-mail: , or
A. Hamon, Laboratoire de Neurophysiologie, UPRES-EA 2647,
´
Universite d’Angers, 2 Bd Lavoisier, 49045 Angers cedex 01, France.
Fax: + 33 241 735 215; E-mail:
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 3 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 4 and 5
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 min at 70 °C under
argon in 0.1 M ammonium bicarbonate pH 8.3 containing
6 M guanidinium chloride and 0.1 M dithiothreitol. After
cooling the reaction mixture to 4 °C, 6 M iodoacetamide
was added to a 2 M final concentration and alkylation was
performed for 45 min 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
0 to 30% for 90 min 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 1 h using an
enzyme/substrate ratio of 1 : 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 ± 2 g body
weight). Anti-insect activity was evaluated in cockroaches
(Blatella germanica, 50 ± 2 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 6 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
5 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 2 mgỈmL)1 BSA. Wash buffer
composition was (in mM): Cl, 140; CaCl2, 1.8; KCl, 5.4;
MgSO4, 0.8; Hepes, 50; 5 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 min (for rat brain synaptosomes) or 60 min (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 2 mgỈmL)1 collagenase (type
IA, Sigma) in calcium-free SOS for 10–15 min. 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 2 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 5 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 4 and 5 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. 1. 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 do 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 3 and not site 4 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 6 ± 3 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 1. 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
5
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. 2. 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 min at 22 °C with 180 pM 125I-labeled Lqh3 and increasing concentration of the indicated toxins. Nonspecific binding, determined in the presence of 1 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 2. (B) Competition curves for 125I-labeled Lqh6 binding inhibition by various toxins. Cockroach neuronal membranes
(16.3 lgỈmL)1) were incubated for 60 min at 22 °C with 200 pM of 125I-labeled Lqh6 and increasing concentration of the indicated toxins. Nonspecific binding, determined in the presence of 1 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 3. 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 6 ± 3 pmolỈmg)1 of protein, n ¼ 3.

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
1 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. 3. 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 min 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 3 [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. 4. 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. 5. 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 1 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 1 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 1 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 4 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 0 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 min 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 1 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 5 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 1 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 4 ) 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 do not follow this correlation,
suggesting that they do 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 do 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. 5 [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. 6. 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. 7. 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 7 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. 8. 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 3 of
neuronal Na channels. As mutant Nav1.4 channels with
Nav1.6 receptor site 3 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. 8. 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.

1. 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.
`
2. Cestele, S. & Catterall, W.A. (2000) Molecular mechanisms of
neurotoxin action on voltage-gated sodium channels. Biochimie
82, 883–892.
3. Novakovic, S.D., Eglen, R.M. & Hunter, J.C. (2001) Regulation
of Na+ channel distribution in the nervous system. Trends Neurosci. 24, 473–478.
4. 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.
5. 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.
6. 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)

7.

8.

9.

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,
423–432.
`
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
3 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, 361371.
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, 186195.
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, 1 (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.