Báo cáo Y học: Variations in receptor site-3 on rat brain and insect sodium channels highlighted by binding of a funnel-web spider d-atracotoxin pdf

Variations in receptor site-3 on rat brain and insect sodium channels
highlighted by binding of a funnel-web spider d-atracotoxin
Nicolas Gilles
1
, Greg Harrison
1
, Izhar Karbat
2
, Michael Gurevitz
2
, Graham M. Nicholson
3,
*
and Dalia Gordon
2
1
CEA, De
`
partement d’Inge
`
nierie et d’Etudes des Prote
`
ines, Gif-sur-Yvette, France;
2
Department of Plant Sciences, Tel Aviv
University, Israel;
3
Department of Health Sciences, University of Technology, Sydney, Australia
d-Atracotoxins (d-ACTXs) from Australian funnel-web
spiders differ structurally from scorpion a-toxins (ScaTx)
but similarly slow sodium current inactivation and compete

for their binding to sodium channels at receptor site-3.
Characterization of the binding of
125
I-labelled d-ACTX-
Hv1a to various sodium channels reveals a decrease in
affinity for depolarized (0 mV; K
d
¼ 6.5 ±1.4 n
M
) vs.polar-
ized ()55 mV; K
d
¼ 0.6±0.2n
M
) rat brain s ynaptosomes.
The in creased K
d
under depolarized conditions correlates
with a 4 .3-fold reduction in the association rate a nd a 1 .8-
increase in the dissociation rate. In comparison, ScaTx
binding affinity decreased 33-fold under depolarized condi-
tions due to a 48-fold reduction in the association rate. The
binding of
125
I-labelled d-ACTX-Hv1a to rat brain syna-
ptosomes is inhibited c ompetitively by classical ScaTxs and
allosterically by brevetoxin-1, similar to ScaTx binding.
However, in contrast with classical ScaTxs,
125
I-labelled d-

ACTX-Hv1a binds with high affinity to cockroach Na
+
channels (K
d
¼ 0.42 ± 0.1 n
M
) and is displaced by the
ScaTx, LqhaIT, a well-defined ligand of insect sodium
channel receptor s ite-3. However, d-ACTX-Hv1a exhibits a
surprisingly low binding affinity to locust sodium channels.
Thus, unlike ScaTxs, which are capable of differentiating
between mammalian and insect sodium channels, d-ACTXs
differentiate between various insect sodium channels but
bind with similar high affinity to rat brain a nd cockroach
channels. Structural comparison of d-ACTX-Hv1a to
ScaTxs suggests a similar putative bioactive surface but a
ÔslimmerÕ overall shape of the spider toxin. A slimmer shape
may ease the interacti on w ith t he cockroach and mammalian
receptor site-3 and facilitate its association with different
conformations of the rat brain receptor, correlated with
closed/open a nd slow-inactivated channel states.
Keywords: brevetoxin; sodium channel; spider toxin; syna-
ptosomes; v oltage-dependent binding.
Australian funnel-web spiders (Araneae: Hexathelidae:
Atracinae) are M ygalomorph spiders confined to the
south-eastern seaboard of Australia. A number o f neuro-
toxins, named atraco toxins (ACTXs) t hat display various
pharmacological properties, have been isolated from the
venom of the funnel-web spider subfamily, Atracinae [1–4].
Several lethal atracotoxins that modulate sodium channel

gating have b een assigned to the d-ACTX group because of
their ability to induce spontaneous repetitive firing in
neuronal cells, accompanied by plateau action potentials
[5–8]. The d-ACTXs, d-ACTX-Hv1a (formerly versutoxin
[9]), the vertebrate-selective toxin d-ACTX-Hv1b [8] from
the venom of Hadronyche versuta,andd-ACTX-Ar1
(formerly robustoxin [10]) from the venom of the male
Sydney funnel-web spider Atrax robustus,arehighly
homologous 42-residue polypeptides. T hese toxins contain
a h igh p roportion of basic residues and show no significant
sequence homology with any presently known neurotoxin.
They are tightly folded molecules constrained by four
conserved intramolecular d isulfide bonds, arranged in a
unique formation. The solution s tructures of d-ACTX-
Hv1a and -Ar1 have been determined by NM R spectro-
scopy [11,12] and constitute a small triple-stranded
antiparallel b-sheet and a Ôcystine knotÕ motive [13].
d-ACTXs slow tetrodotoxin (TTX)-sensitive s odium chan-
nel inactivation and produce modest shifts in the voltage-
dependence of sodium channel activation in i nsect and
mammalian n eurons [5–8] in a manner similar to scorpion
a-toxins and sea anemone toxins [14,15]. Despite the similar
effect on sodium current inactivation kinetics, d-ACTXs
have a distinct three-dimensional structure, which differs
greatly from those of other toxins interacting with recep-
Correspondence to D. Gordon, Department of Plant Sciences,
Tel-Aviv University, Ramat-Aviv, Tel Aviv 69978, Israel.
Fax: +972 3 640 6100, E-mail:
Abbreviations: Aah-II, antimammalian a-toxin II from the venom of
the scorpion Androctonus australis hector; d-ACTX-Ar1, d-atraco-

toxin-Ar1 ( formerly robustoxin) from Atrax robustus; d-ACTX-Hv1a,
d-atracotoxin-Hv1a (formerly versutoxin) from Hadronyche versuta;
ATX-II, toxin II from the sea anemone Anemonia sulcata;
[
3
H]BTX,[
3
H]batrachotoxinin, A-20a-benzoate; IC
50
, median inhibi-
tory concentration; K
d
, dissociation constant; K
i
, inhibitory constant;
k
off
, dissociation rate constant; k
on
, association rate constant; Lqh-II,
Lqh-III, LqhaIT, a classical a-toxin, an a-like toxin, and an a-toxin
highly active on insects, respectively, from the venom of the scorpion
Leiurus quinquestriatus hebraeus; PbTx-1, brevetoxin-1 from the
dinoflagellate Ptychodiscus brevis.
Enzyme: lactoperoxidase (EC 1 .11.1.7).
*Presen t address: Department o f Health S ciences, University of
Technology, Sydney P O Box 123, Broadway NSW 2007, Au stralia.
Fax: +61 2 9514 2 228, E-mail:
(Received 2 5 September 2001, revised 15 J anuary 2002, accepted
21 January 2002)

Eur. J. Biochem. 269, 1500–1510 (2002) Ó FEBS 2002
tor site-3 {e.g. scorpion a-toxins, Aah-II (toxin II from
Androctonus australis hector [16]), LqhaIT (from Leiurus
quinquestriatus hebraeus [17]), and the sea anemone toxin,
anthopleurin-B [18]}.
At least seven neurotoxin receptor sites h ave been
identified on the voltage-gated sodium channel by radio-
labelled toxin binding studies [19]. Scorpion a-andsea
anemone toxins such as ATX-II bind to neurotoxin
receptor site-3 (for reviews s ee [19–22]). The binding of
classical s corpion a-toxins, such as Aah-II [23] and Lqh-II
[24,25], to receptor site-3 on rat brain s odium channels is
voltage d ependent and allosterically modulated by lipid-
soluble sodium channel activators such as brevetoxin,
veratridine and batrachotoxin [21,22,27–30]. Notably, at
nanomolar con centrations d-ACTX-Ar1 a nd -Hv1a com-
pletely inhibit the binding of classical scorpion a-toxins
(e.g. Aah-II and Lqh-II) to rat brain synaptosomes as well
as the binding of LqhaIT to insect sodium channels
[29,30]. Thus, d-ACTXs constitute a unique group of
polypeptides capable of high affinity binding presumably
to receptor site-3 on both mammalian and insect voltage-
gated sodium channels. Indeed, they enhance
3
H-batra-
chotoxin binding similarly to scorpion a-toxins; however,
they differ from scorpion toxins in that they inhibit, rather
than enhance, the activation o f sodium channels by
batrachotoxin [30]. Thus, clarification of d-ACTXs recep-
tor sites on sodium channels requires a detailed analysis of

their binding properties.
Here we provide a detailed characterization of radio-
labelled d-ACTX-Hv1a direct binding to rat brain and
insect sodium channels. We present evidence that
d-ACTX-Hv1a acts similarly to scorpion a-t oxins in terms
of its interaction with sodium channels at nanomolar
affinities, similar voltage dependence and allosteric inter-
action with brevetoxin-1 on rat brain sodium channels.
Nevertheless, d-ACTX-Hv1a differs f rom scorpion
a-toxins in its lower voltage dependency and ability to
differentiate between receptor site-3 of cockroach and
locust sodium channels rather than between rat brain and
cockroach.
EXPERIMENTAL PROCEDURES
Materials
The scorpion a-toxin, Lqh-II, f rom the venom of the
scorpion L. q. hebraeus, was purchased from Latoxan (A.P.
1724, 05150 Rosans, France) and, in part, w as also a
generous gift from Dr P. Sautie
¨
re (Insitut Pasteur, Lille,
France) [24]. LqhaIT, an a-insect toxin from the scorpion
L. q. hebraeus, was produced in Escherichia coli as described
previously [31]. Lactoperoxidase (EC 1.11.1.7) was pur-
chased from Sigma. Carrier-free Na
125
I was purchased from
Amersham (Buckinghamshire, UK). All other chemicals
were of analytical grade. Filters for binding assays were GF/
C glass fibre discs (Whatman) preincubated in 0 .3%

polyethylenimine (Sigma).
Purification of d-ACTX-Hv1a and d-ACTX-Ar1
Crude venom was ÔmilkedÕ b y direct aspiration from the
chelicerae of live spiders maintained in a colony, using
silanized (Coatasil; Ajax Chemicals, Australia) glass
pipettes. d-ACTX-Hv1a was obtained from a dult male or
female H. versuta spiders while d-ACTX-Ar1 was obtained
from adult male A. robustus spiders. Crude venom was
washed from pipettes with 0.1% (v/v) trifluoroacetic acid
and d-ACTX-Hv1a and d-ACTX-Ar1 isolated a nd purified
by RP-HPLC. Purification was achieved using a Pharmacia
HPLC system using a Vydac analytical rpHPLC column
(C
18
, 250 · 4.6 mm, 300 A
˚
,5lm p article size). Pooled
venom was applied to the column and venom components
eluted at a flow rate of 1 mLÆmin
)1
using a linear gradient of
5–25% acetonitrile/0.1% trifluoroacetic acid over 22 min,
followed by a gradient of 25–50% acetonitrile/0.1%
trifluoroacetic acid over 48 min. Fractions containing
d-ACTX-Hv1a or d-ACTX-Ar1 were then purified further
using a linear gradient of 23–32% acetonitrile/0.1%
trifluoroacetic acid over 20 min at a flow rate of 1Æml min
)1
.
Toxin quantification was performed using a bicinchoninic

acid Protein Assay K it (Pierce) using BSA as a standard.
Absorbance was read at 570 nm on a BIO-RAD
Model 450 microplate reader. The molecular mass was
determined by electrospray ionization MS. The fractions
containing d-ACTX-Hv1a (M
r
¼ 4852) or d-A CTX-
Ar1 (M
r
¼ 4854) were stored lyophilized at )20 °Cin
5–10-nmol aliquots. When required, spider toxins were
dissolved in 10 m
M
Hepes (N-2-hydroxyethylpiperazine-N-
2-ethanesulfonic acid)/Tris buffer (pH 6.0) and an aliquot
of this stock solution was diluted in the binding solution.
Any unused d-ACTX stock solution was kept a t 4 °C
andusedwithin2weeks.
Neuronal membrane preparation
All buffers used for preparation of neuronal membranes
contained a cocktail of proteinase inhibitors composed of:
phenylmethylsulphonyl fluoride (50 lgÆmL
)1
), pepstatin A
(1 l
M
), iodoacetamide (1 m
M
) and 1,10-phenantroline
(1 m

M
). All m embrane preparation steps were performed
on ice. Rat brain synaptosomes were prepared from adult
albino Sprague-Dawley r ats ( 300 g, laboratory bred),
according to the method described by Kanner [ 32]. The
synaptosomes, w hich were concentrated at the 1 2 a nd 16%
Ficoll gradie nt interface, were washed and aliquoted into
Eppendorf tubes and s tored at )80 °C. Before experiments,
the synaptosomes were rapidly defrosted for 30 s in a 37 °C
water bath, placed on ice and used immediately (for
polarized synaptosomes). For experiments carried out
under depolarized conditions, s ynaptosomes were defrosted
and i ncubated at 37 °C for 30 min to facilitate ion gradient
dissipation, and then kept on ice until used. Insect
synaptosomes were prepared from entire heads of a dult
cockroaches Periplaneta a mericana according to a pre-
viously described method [33]. Locust synaptosomes were
prepared from dissected brains and ventral nerve cords of
adult locusts, Locusta migratoria, as described previously
[34,35]. Frozen synaptosomes were used within 6 months.
Membrane protein c oncentration was determined by a Bio-
Rad Protein Assay Kit, using BSA as a s tandard.
Radioiodination of d-ACTX-Hv1, LqhaIT and Lqh-II
Toxins (5 lg) were radioiodinated for 1 min using 0.7 IU of
lactoperoxidase (EC 1.11.1.7) from bovine milk and 0.5 mCi
carrier-free Na
125
I, in 10 lLH
2
O

2
(diluted 1 : 50 000) and
Ó FEBS 2002 Binding of d-atracotoxins on sodium channels (Eur. J. Biochem. 269) 1501
50 lL20m
M
phosphate buffer pH 7.2. The mono-iodo-
toxins were purified using a Vydac analytical C
18
rpHPLC
column and an acetonitrile gradient from 45 to 60% B
(A ¼ aqueous 0.1% trifluoroacetic acid, B ¼ 0 .085% tri-
fluoroacetic acid, 50% acetonitrile, 0.2% B per min) at a
flow rate of 1 mlÆmin
)1
. The peak of the mono-iodo LqhaIT
toxin eluted just a fter the peak of unmodified toxin as
described previously [36]. T he unmodified d-ACTX- Hv1a,
which eluted at 26% acetonitrile, was followed by t wo
radiolabelled fractions, eluting at 26.5 and 27.1% a cetonit-
rile. As d-ACTX-Hv1a co ntains only two Tyr residues
(positions 22 and 25), we determined the iodinated r esidue of
each mono-iodo
125
I-labelled d-ACTX-Hv1a fraction, using
approximately 150 000 c.p.m. of each fraction in the
presence of unlabelled toxin by Edman degradation and
sequencing (Applied Biosystem, 477 A p rotein sequencing),
as described in detail f or
125
I-labelled LqhaIT [36]. The

N-terminal sequence analysis indicated that the first fraction
of
125
I-labelled d-ACTX-Hv1a was labelled on tyrosine 22,
while the second mono-iodo
125
I-labelled d-ACTX-Hv1a
was labelled on t yrosine 25. The concentration of t he radio-
labelled toxins were determined according to the specific
activity of
125
I corresponding to 3000–2500 d.p.m.Æfmol
)1
mono-iodotoxin, depending on the age of the radiotoxin
and by est imation of its biological activity as described
previously ([36], usually 60–70% for
125
I-labelled L qhaIT
and 35–55% for
125
I-labelled d-ACTX-Hv1a).
125
I-labelled
LqhaIT was use d within 2 weeks whereas
125
I-labelled
d-ACTX-Hv1awasusedwithin7days.
Competition binding studies
For competition binding experiments using
125

I-labelled
d-ACTX-Hv1a on rat brain sodium channels, synapto-
somes were thawed at 37 °C (for 30 s) and suspended in
0.2 or 0.6 mL binding buffer c ontaining a low concentra-
tion of radiolabelled toxins (see figure legends). Standard
binding medium composition w as (in m
M
): choline Cl, 130;
CaCl
2
,1.8;KCl,5;MgSO
4
, 0.8; Hepes, 50;
D
-glucose 10;
and 2 mgÆmL
)1
BSA. Following incubation for the
designated time periods the reaction was terminated by
dilution with 2 ml ice-cold wash buffer of the following
composition (in m
M
): choline Cl, 140; CaCl
2
,1.8;KCl,5.4;
MgSO
4
,0.8;Hepes,50;pH7.2and5mgmL
)1
BSA.

Separation of free from bound toxin was achieved by rapid
filtration under vacuum using Whatman GF/C filters
preincubated with 0.3% polyethylenimine. The filter discs
were then rapidly washed twice with 2 mL buffer.
Termination of the reaction and washing lasted 10 s.
Nonspecific toxin binding was determined in the presence
of a high concentration o f t he unlabelled toxin, as specified
in figure legends, and comprised t ypically 5–10% o f total
binding for
125
I-labelled LqhaIT and 30–50% for
125
I-
labelled d-ACTX-Hv1a. Competition binding experiments
using
125
I-labelled L qhaIT and
125
I-labelled d-A CTX-Hv1a
on insect neuronal membranes were performed according
to established methods [35,36] under c onditions specified in
the figure legends.
Equilibrium and kinetic analysis of binding
The median inhibitory concentration (IC
50
) values for the
inhibition of toxin binding were determined by nonlinear
regression analysis using t he Hill equation using a Hill
coefficient (n
H

) of 1. Mathematical curve fitting was
accomplished using
KALEIDAGRAPH
(Synergy Software,
USA) f or I C
50
determination and the K
i
values were
calculated [37]. Cold saturation a ssays were performed
using increasing concentrations of the unlabelled toxin in
the presence of a constant low concentration of the mono-
iodinated toxin. H ot saturation assays were performed
using increasing concentrations of the radiolabelled toxin,
with the s ame amount of membranes. Data were analysed
using the iterative program
LIGAND
(Elsevier B iosoft) u sing
Ôcold saturationÕ or Ôhot saturationÕ analysis. The kinetic
data for ligand association and dissociation rates were
subjected to analysis by
LIGAND
,usingÔkinetic analysisÕ.
Each curve was subjected to multislope analysis to detect the
presence of one or two slopes. Toxin dissociation curves
were initiated by the addition of excess unlabelled toxin and
the dissociation rate constant (k
off
) was determined directly
from a first order plot of ligand dissociation v s. time. The

rate of toxin association (k
on
) was determined from the
equation:
k
on
¼ k
obs
½RL
e
½L½RL
max

where [ L] is the c oncentration of ligand, [RL]
e
is the
concentration o f t he complex at equilibrium, [RL]
max
is the
maximum number of receptors present (determined in a
parallel saturation experiment) and k
obs
is the slope of the
pseudo-first order plot of l n ([RL]
e
/{[RL]
e
–[RL]
t
}) vs. time

[38]. The conce ntration of labelled ligand in association
kinetic determinations was adjusted t o keep the reaction at
pseudo-first order conditions and varied according to the K
d
values of the toxin under polarized or depolarized co ndi-
tions [38]. Results were compared using a Student’s t-test
and all data are expressed as the mean ± SEM from the
number of experiments (n) indicated. The corresponding
affinity (K
d
) can be calculate from t he k inetics parameters
according to the equation K
d
¼ k
off
/k
on
.
RESULTS
Both d-ACTX-Hv1a and d-ACTX-Ar1 inhibit c ompletely
the binding of
125
I-labelled Aah-II [28] and
125
I-labelled
Lqh-II [29] to rat brain sodium channels. In order to
characterize the receptor binding site fo r d-ACTXs on
sodium channels, w e analysed the binding properties of
radioiodinated d-ACTX-Hv1a to rat brain synaptosomes.
We have first examined the ability of the classical scorpion

a-toxin, Lqh-II, to displace
125
I-labelled d-A CTX-Hv1a
from its binding site and found a K
i
value of
0.25 ± 0.03 n
M
(n ¼ 4; data not shown), supporting the
notion that d-ACTXs share r eceptor site-3 with scorpion
a-t oxins.
Evidence that d-ACTX-Hv1a binding to rat brain
synaptosomes is voltage dependent
Binding of scorpion a-toxins depends on polarization of
the synaptosome membrane and therefore is a useful
measure in m onitoring membrane p otential. Indeed, a 90%
decrease of the i nitial maximal binding between polarized
and depolarized synaptosomes has been shown [39–42].
1502 N. Gilles et al. (Eur. J. Biochem. 269) Ó FEBS 2002
The resting membrane potential of rat brain synaptosomes
is approximately )55 mV (a t 5 m
M
[K
+
]
o
) due mainly to a
high intracellular c oncentration of K
+
, which diffuses

passively through the membrane [39,41,43]. Although
depolarization of synaptosomes by elevating [K
+
]
0
is often
used for measuring the influence o f m embrane potential on
scorpion a-toxin binding [40,44], we have found that high
concentrations of K
+
in the binding buffer p erturb the
binding of
125
I-labelled d-ACTX-Hv1a (data not shown).
Unlike the situation with the binding of the a-toxin,
125
I-labelled L qh-II, where non specific b inding did not
change with increasing external K
+
concentration (bet-
ween 5 and 135 m
M
[42]), the level of nonspecific binding
of
125
I-labelled d-A CTX-Hv1a v aried great ly, posing
difficulties for data analysis. Therefore, in order to
depolarize the membrane without affecting o ther binding
conditions, the synaptosomes were incubated for 30 min at
37 °C i n normal b inding b uffer (containing 5 m

M
K
+
see
Experimental procedures) prior to addition of the labelled
toxin (under such conditions the K
+
gradient dissipates,
the membrane potential approaches 0 mV and scorpion
a-toxin binding is decreased by 90% [42]). The time-course
of
125
I-labelled d-ACTX-Hv1a binding to polarized rat
brain synaptosomes w as performed at 22 °Ctomaintain
the r esting membrane potential for longer duration
(Fig. 1 A [42]). Maximal binding was achieved after
10–15 min and was maintained for an additional 10 m in
before an apparent decrease could be observed. A similar
decrease in saturable binding was observed with the
scorpion a-toxin Lqh-II (Fig. 1, i nset), suggesting depend-
ency of binding on membrane depolarization for both
toxins. T herefore, all subsequent experiments on polarized
synaptosomes were performed at 22 °Cwitha20min
incubation time to reach equilibrium binding conditions.
Despite the similar effect of membrane potential on
binding of both t oxins, the ratio between
125
I-labelled
d-ACTX-Hv1a (F ig. 1B, left bars) maximal binding to
polarized (emp ty bar) vs. depolarized (gray bars) synapto-

somes was substantially different from that measured for
125
I-labelled Lqh-II (Fig. 1B, right bars). This difference
necessitated analysis of t he binding affinity d-ACTX-Hv1a
to rat brain synaptosomes under polarized and depolarized
conditions.
Affinity of d-ACTX-Hv1a for polarized and depolarized
rat brain synaptosomes
To study the influence of synaptosome membrane potential
on
125
I-labelled d-ACTX-Hv1a affinity,
125
I-labelled
d-ACTX-Hv1a was incubated with polarized or depolarized
rat brain synaptosomes in the presence of i ncreasing
concentrations of unlabelled toxin (cold saturation). The
dissociation constant (K
d
)ofd-ACTX-Hv1a increased
11-fold between polarized (K
d
¼ 0.5 7 ± 0.20 nM; n ¼ 3)
and depolarized (K
d
¼ 6.5 ± 1.4 n
M
; n ¼ 5) synapto-
somes, whereas the maximum number of receptor sites
(B

max
) increased 1.8-fold ( P <0.05; B
max
¼ 1.24 ±
0.17 pmol Æmg protein
)1
; n ¼ 3and2.26±0.05pmolÆmg
protein
)1
; n ¼ 5, respectively; Fig. 2). To assure the signi-
ficance of the change in B
max
, the experim ents under
polarized and depolarized conditions were performed in
parallel using the same batch of rat brain synaptosomes. It is
Fig. 1. Time-course of
125
I-labelled d-ACTX-Hv1a and
125
I-labelled Lqh-II binding to rat brain syna ptosomes at 22 °C. (A) Typical association
kinetics of
125
I-labelled d-ACTX-Hv1a (75 p
M
) to p olarized synaptosomes (20 lgÆmL
)1
membrane protein). Non-specific binding, determined in
thepresenceof1l
M
Lqh-II, was t ime-invariant, and was subtrac ted f rom t he e xperime ntal data points. Maximal bind ing o f

125
I-labelled d-ACTX-
Hv1a rema ined stable for 10 min before decreasing due to spontaneous depolarization of synaptosomes. The time-course of 60 p
M
125
I-labelled
Lqh-II binding to p o larized synaptosomes (20 lgÆmL
)1
) is p resent ed in the i nset as pe r cent of maximal s pecific binding. (B) Comparison of t he
maximal binding of
125
I-labelled d-ACTX-Hv1a (left bars) and
125
I-labelled Lqh-II (right bars) to polarized (empty bars) and to depolarized (gray
bars) synaptosomes.
125
I-labelled d-ACTX-Hv1a binding was performed as described for panel (A). Maximal binding under polarized
(5.5 ± 0.3 p
M
and 5 .1 ± 0.2 p
M
bound
125
I-labelled d-ACTX-Hv1a or
125
I-labelled Lqh-II, respectively) and depolarized synaptosome s pre-
treated at 37 °C for 30 m in (3 .8 ± 0.6 p
M
and 0.75 ± 0.15 p
M

,for
125
I-labelled d-ACTX-Hv1a or
125
I-labelled Lqh-II, respectively), corresponds
to the l evel of
125
I-labelled toxin b inding after 2 0 and 60 min of incubation for polarized and d epolarized conditions, respectively.
Ó FEBS 2002 Binding of d-atracotoxins on sodium channels (Eur. J. Biochem. 269) 1503
noteworthy, that in order to maintain a pseudo-first order
reaction conditions, toxin and r eceptor (membrane protein)
concentrations wer e adjusted as a function of the change in
125
I-labelled d-ACTX-Hv1a affinity (Fig . 2 [38]).
Kinetic constants of d-ACTX-Hv1a binding to rat brain
synaptosomes
125
I-Labelled d-ACTX-Hv1a was incubated with synapto-
somes and the association b inding kinetics were monitored
until equilibrium had been reached (Fig. 3A, closed symbols
for polarized, and open symbols for depolarized synapto-
somes, respectively). After 10 min incubation with polarized
synaptosomes, toxin dissociation was initiated by adding
1 l
M
unlabelled d-ACTX-Hv1a (Fig. 3B, closed symbols).
The calculated association and dissociation rate constants,
k
on
and k

off
, under polarized membrane conditions were
1.84 ± 0.2 · 10
6
Æ
M
)1
Æs
)1
(n ¼ 3) and 1.1 ± 0.1 · 10
)3
Æs
)1
(n ¼ 3), respectively. The corresponding K
d
(0.6 ± 0.1 n
M
)
calculated from the k inetic values, was comparable with the
values obtained a t equilibrium (F ig. 2). Equilibrium of
d-ACTX-Hv1a binding to depolarized synaptosomes was
achieved after longer incubation (Fig. 3A, open symbols),
and dissociation was induced after 60 min of association by
adding 1 l
M
unlabelled d-ACTX-Hv1a (Fig. 3B, open
symbols). The calculated k
on
and k
off

under d epolarized
membrane conditions were 0.43 ± 0.13 · 10
6
Æ
M
)1
Æs
)1
(n ¼ 3) and 2.0 ± 0.3 · 10
)3
s
)1
(n ¼ 3), respectively. The
corresponding calculated K
d
was 4.7 ± 2.1 n
M
,which
fitted the value obtained at equilibrium. These results
indicate that d-ACTX-Hv1a binding i s dependent on the
membrane potential of synaptosomes. Interestingly, syna-
ptosome depolarization had a minute effect on k
off
but
decreased fourfold th e k
on
. These results support our recen t
studies using t he classical s corpion a-toxin, Lqh-II [42], a nd
seem to provide a different interpretation to that suggested
previously for scorpion a-toxins, which attributed the

change in binding affinity under depolarized conditions
mainly to an increase in the d issociation rate
[15,26,39,40,44–47].
Allosteric modulation of d-ACTX-Hv1a binding site
on rat brain synaptosomes
Brevetoxin-1 (PbTx-1) from a marine dinoflagellate, inhibits
allosterically the b inding of the s corpion a-tox in, Aah-II, to
rat brain synaptosomes [27,28]. To examine the similarity in
binding to receptor site-3 between scorpion a-toxins and
d-ACTXs, we analysed t he effect of PbTx-1 on
125
I-labelled
d-ACTX-Hv1a binding to rat brain synaptosomes. Simi-
larly to the effect of PbTx-1 on Aah-II binding [27,28],
this brevetoxin substantially inhibited the binding of
125
I-labelled d-ACTX-Hv1a with an IC
50
of 50 n
M
(data
not shown). We a lso analysed the effect of deltamethrin, a
pyrethroid insecticide known to modulate sodium channels,
and like scorpion a-toxins [48] found that it had no allosteric
effect on d-ACTX-Hv1a binding.
d-ACTX-Hv1a differentiates between cockroach
and locust sodium channels
d-Atracotoxins are unique in t heir potency to displace
scorpion a-toxins from their binding sites on both rat brain
and cockroach sodium channels [29,30]. Therefore, the

interaction of
125
I-labelled d-ACTX-Hv1a with scorpion
a-toxins on binding to cockroach neuronal membranes was
examined. Competition bind ing experiments using increas-
ing concentrations of the scorpion toxins, Lqh-II, LqhaIT,
and Lqh-III as well as the related spider toxin, d-ACTX -
Ar1, revealed complete inhibition of
125
I-labelled d-ACTX-
Hv1a binding by all toxins tested (Fig. 4, main panel).
Scatchard transformation of the competition b inding
curve of d-ACTX-Hv1a to c ockroach neuronal m embranes
(cold saturation, Fig. 4, inset) provid ed a K
d
value of
0.42 ± 0.1 n
M
(n ¼ 3), which was highly similar to t he
affinity of d-ACTX-Hv1a binding to polarized rat brain
synaptosomes (Table 1). The receptor site capacity (B
max
¼
2.1±0.5pmolÆmg protein
)1
; n ¼ 3) was similar to that
obtained previously for LqhaIT binding to cockroach
neuronal membranes [35,36].
Fig. 2. Scatchard plots of
125

I-labelled d-ACTX-Hv1a binding to rat
brain synaptosomes. (A) 92 p
M
125
I-labelled d-ACTX-Hv1a incubated
at 22 °C for 20 min with p olarized synaptosomes (28.8 lgÆmL
)1
)and
(B) 168 p
M
125
I-labelled d-ACTX-Hv1a incubated at 22 °Cfor60min
with depolarized synaptosomes (36.5 lgÆmL
)1
), in the presence of
increasing concentrations of unlabelled toxin (cold saturation) (see
Experimental procedures). Analysis of a typical experiment is presen-
ted. Nonspecific binding, determined in the presence of 0.2 l
M
(A) or
1 l
M
(B) d-ACTX-Hv1a, was subtracted. Equilibrium binding
parameters were calculated using the program
LIGAND
(see Experi-
mental procedures). The dissociation constants (K
d
)were
0.57 ± 0.2 n

M
and 6.5 ± 1.4 n
M
and the maximum num ber of
binding sites (B
max
) were 1.24 ± 0.17 pmolÆmg p rote in
)1
and
2.26 ± 0.05 pmolÆmg protein
)1
, under polarized (n ¼ 3) and depo-
larized (n ¼ 5) conditions, respectively.
1504 N. Gilles et al. (Eur. J. Biochem. 269) Ó FEBS 2002
Unexpectedly, however, no specific binding of
125
I-labelled d-ACTX-Hv1a to locust neuronal membranes
could be d etected. As w e have recently demonstrated that
iodination of one Tyr residue in the a-like toxin, Lqh-III,
impairs binding to locust but not cockroach sodium
channels [36], we identified which Tyr residue was
iodinated on
125
I-labelled d-ACTX-Hv1a. Amino acid
sequence analysis of the two radiolabelled peaks obtained
during toxin radioiodination (see Experimental proce-
dures) identified an iodinated Tyr22 in the first peak and a n
iodinated Tyr25 in the second peak. Both i odinated
derivatives did not differ in their binding properties to
cockroach neuronal membranes (data not shown). In order

to eliminate the possibility that the lack of
125
I-labelled
d-ACTX-Hv1a binding to locust neuronal membranes was
consequent on its iodination per s e, we e xamined the
binding of d-ACTX-Hv1a to locust sodium channels
indirectly, by its ability to compete for
125
I-labelled
LqhaIT binding (Fig. 5). Interestingly d-ACTX-Hv1a
competed for LqhaIT binding only at h igh concentrations
(K
i
¼ 67 ± 17 n
M
; n ¼ 3) with a K
i
value 160-fold higher
than the K
d
for cockroach sodium channels (Fig. 4). Thus,
in co ntrast with the s corpion a-like t oxin Lqh-III [36] and
despite i ts high binding affinity for cockroach sodium
channels, d-ACTX-Hv1a is a weak ligand on l ocust
sodium channels (Fig. 5 ).
DISCUSSION
Effect of membrane depolarization on kinetics
of toxin binding to receptor site-3
The binding properties o f the spider toxin, d-ACTX-Hv1a,
to rat brain synaptosomes resemble those of scorpion

a-toxins, thereby suggesting a common receptor binding site
on the sodium channel. This resemblance is substantiated by
a similar, yet nonidentical, decrease in binding affinity at
polarized ()55 mV) and depolarized (0 mV) membrane
potentials (K
d
increase of 11.4-fold for d-ACTX-Hv1a and
33-fold for the classical a-toxin, Lqh-II, Table 1 [42,49]).
TheincreaseinK
d
of d-ACTX-Hv1a binding correlates with
a 4 .3-fold l ower association rate a nd a 1.8-fold increase in
the dissociation rate ( Table 1). The more p rofound increase
Fig. 3. Association (A) and dissociation (B) kinetics of
125
I-labelled d-ACTX-Hv1a binding to polarized and depolarized rat brain synaptosomes. Fifty
and 200 p
M
125
I-labelled d-ACTX-Hv1a were incubated at 22 °C (in 200 lL)inthepresenceof37or73lgÆmL
)1
polarized o r d epolariz ed
synaptosomes, respectively, for various periods of time. Nonspecific binding, determined in the presence of 200 n
M
or 1 l
M
Lqh-II (for polarized
and depolarized con ditions, respectively) was time-invariant and was subtracted from the experimental points.
125
I-labelled d-ACTX-Hv1a

dissociation w as initiated by addition of 200 n
M
or 1 l
M
unlabelled t oxin after 10 o r 60 min association u nder polarized and depolarized conditions,
respectively (B). A typical experiment is presented. Kinetic constants, representing the mean of three experiments, were: k
on
¼ 1.84 ±
0.2 · 10
6
s
–1
Æ
M
)1
and k
off
¼ 1.1 ± 0.1 · 10
–3
s
)1
under polarized conditions; k
on
¼ 0.43 ± 0.13 · 10
6
s
)1
Æ
M
)1

and k
off
¼ 2.0 ±
0.3 · 10
–3
s
)1
under depolarized conditions.
Fig. 4. Binding interaction of
125
I-labelled d-ACTX-Hv1a with cock-
roach sodium channels. Competition for
125
I-labelled d-ACTX-Hv1a
(120 p
M
) binding to neuronal membranes (7 lgÆmL
)1
)byvarious
neurotoxins. Nonspecific binding, determined in the presence of
200 n
M
LqhaIT, was subtracted. Bound
125
I-labelled d-ACTX-Hv1a is
expressed as the percentage of maximal specific binding in the absence
of c omp etitor toxins. The competition c urves were fitted by the non-
linear Hill equation (with a Hill coefficient of 1) to determine I C
50
values (see Experimental procedures). Typical curves are presented.

The K
i
values (in n
M
) and the number of experiments (n)are:LqhaIT,
0.12–0.16 (n ¼ 2); Lqh-III, 0.12–0.14 (n ¼ 2); d-ACTX-Hv1a,
2.6–3.0 (n ¼ 2); d-ACTX-Ar1, 1.5–2.5 ( n ¼ 2); L qh-II, 9.5–14.1 n
M
(n ¼ 2). Inset: Scatchard transformation of 55 pM
125
I-labelled
d-ACTX-Hv1a binding to cockroach neuronal membranes
(8.7 lgÆmL
)1
) i n a volume of 600 lL using in creasing concentrations
of unlabelled d-ACTX-Hv1a (Ôcold saturationÕ). The equilibrium
binding parameters were calculated u sing the program
LIGAND
.Data
represents the mean of t wo cold- and two hot-saturation experiments
(see Expe rimental procedures),which s howed no significant differences.
K
d
¼ 0.42 ± 0.1 n
M
; B
max
¼ 2.1±0.5pmolÆmg protein
)1
.

Ó FEBS 2002 Binding of d-atracotoxins on sodium channels (Eur. J. Biochem. 269) 1505
in K
d
of Lqh-II binding under similar steady-state con di-
tions may be related to the 48-fold decrease in its association
rate constant (Table 1). Thus, the conformational alteration
induced by depolarization at receptor site-3 appears to
affect toxin binding by two mechanisms. The first involves
steric (architectural) and/or electrostatic (long-range) chan-
ges, which are unfavourable for d-ACTX-Hv1a access and
even more so for Lqh-II, thus reducing substantially the k
on
.
The d iff ere nce in k
on
suggests that the two toxins bind in a
nonidentical manner to overlapping receptor sites. T he
second mechanism involves a change in the surface of
receptor site-3, which destabilizes its close fit with the bound
toxin, thus increasing the off-rate. Surprisingly, this change
is much s maller a s the off-rate of both t oxins i ncreased less
then twofold b etween polariz ed and d epolarized steady-
state conditions but affects d-ACTX-Hv1a binding affinity
more than that of Lqh-II (Table 1). Hence, depolarization
conditions hinder Lqh-II association to a greater e xtent
than d-ACTX-Hv1a,andmayberelatedtotheirdifferent
structures.
A slow inactivated channel state prevails
in depolarized synaptosomes
d-ACTX-Hv1a binding to rat brain synaptosomes reveals

an increase in K
d
(11.4-fold) a nd in B
max
(1.8-fold) between
polarized and depolarized conditions (Fig. 2). The increase
in B
max
may be attributed to a change in the ratio between
sodium channels at high and low affinity states for toxin
binding. W e assume that at resting membrane potentials,
only  60% of site-3 receptors are in a high affinity
conformation enabling toxin binding (presumably on
sodium channels in closed states [47]), whereas the remain-
ing channels are in a low affinity conformation asso ciated
with the slow-inactivated state. This suggestion is supported
by our study with Lqh-II, using both b inding and electro-
physiological analyses [42] and the study of Smith & Goldin
[50] which implied that, at )55 mV, most rat brain subtype I
(rBI) channels, which comprise  20% o f sodium channels
in synaptosomes [51], were available for activation presum-
ably by being in closed, resting states [50]. Nevertheless they
showed that at identical membrane potential more than
50% of brain subtype IIA (rBIIA) channels that constitute
the ma jority ( 80%) i n s ynaptosomes [51], were in an
inactivated state. Our electrophysiological analysis of rBII
channels expressed in mammalian cells supports this
conclusion [42]. Thus, a substantial fraction of s odium
channels would occupy the slow-inactivated states at
polarized synaptosomes and thus display a low affinity

conformation for t oxin binding [42]. The observed i ncrease
in Aah-II binding to rat brain sodium channels in the
presence of TTX [28] may b e attributed to shifting receptor
site-3 from low to h igh affinity conformation by binding of
TTX to the external vestibule of the slow-inactivated
channel pore. In s o far as the
125
I-labelled d-ACTX-Hv1a
concentrations used in the binding studies were low
compared to the K
d
(Figs 2 and 3 ), the low affinity binding
sites in polarized synaptosomes were undetectable, because
only a small fraction of site-3 receptors were o ccupied.
Conversely, in d epolarized synaptosomes, m ost sodium
channels are in the slow-inactivated states, thus available for
toxin binding only a t low affinity conformations of receptor
site-3. In this situation and under proper ligand concentra-
Table 1. Comparison between equilibrium and kinetic binding parameters of
125
I-labelled d-ACTX-Hv1a and the scorpion a-toxin, Lqh-II. Binding to
polarized (membrane potential of )55 mV) and depolarized (0 mV) synaptosomes is performed as described in Figs 2 and 3 (see text for details).
Data are m eans ± SEM values. n ¼ number of i nde penden t experiments. Data for Lqh-II binding parameters are from G illes et al. [42].
Toxin Synaptosomes
K
d
(n
M
)
k

on
(10
6
M
)1
Æs
)1
)
k
off
(10
–3
Æs
)1
)
d-ACTX-Hv1a Polarized 0.57 ± 0.20 1.84 ± 0.2 1.1 ± 0.1
(n ¼ 3) (n ¼ 3) (n ¼ 3)
Depolarized 6.5 ± 1.4 0.43 ± 0.13 2.0 ± 0.3
(n ¼ 4) (n ¼ 3) (n ¼ 3)
Lqh-II Polarized 0.18 ± 0.04 12.0 ± 4.0 0.82 ± 0.06
(n ¼ 3)(n ¼ 6)(n ¼ 3)
Depolarized 5.85 ± 0.5 0.25 ± 0.03 1.12 ± 0.08
(n ¼ 4)(n ¼ 4)(n ¼ 3)
Fig. 5. Co mpe tit io n of d-ACTX-Hv1a for
125
I-labelled LqhaIT binding
to locust sodium channels.
125
I-labelled LqhaIT (0.2 n
M

)wasincubated
for 60 m in at 22 °C with i ncreasin g con centrations of LqhaIT or
d-ACTX-Hv 1a, a nd the binding to loc ust neuronal membranes
(60 lgÆmL
)1
) was determined. Nonspecific binding, determined in the
presence of 200 n
M
LqhaIT, was subtracted. Bou nd
125
I-labelled
LqhaIT is expressed as the percentage of the maximal specific binding
in the absence of c ompet itor . A typical expe riment i s presented. The
competition curves w ere fi tted b y a nonlinear Hill equation (with a Hill
coefficientof1)todetermineIC
50
values (see Experimental proce-
dures). The calculated K
i
values [37] were: LqhaIT, 1 .2–2.2 n
M
(n ¼ 2) ; d-ACTX-Hv1a, 67 ± 17 n
M
(n ¼ 3).
1506 N. Gilles et al. (Eur. J. Biochem. 269) Ó FEBS 2002
tions, maximum binding capacity is observed (Fig. 2, and
see [42]).
The coexistence of (at least) two distinct conformational
states of receptor site-3 among sodium channels in
polarized synaptosomes gains further support from the

1.9-fold increase in receptor site capacity for scorpion
a-toxins b inding in the p resence of b atrachotoxin, an
alkaloid toxin binding to receptor site-2 [41]. This result
suggests that batrachotoxin allosterically affects receptor
site-3 by shifting it from the l ow t o the high affinity state,
increasing both scorpion a-toxin affinity and receptor site
capacity [20,21,28]. Together these results indicate that the
low affinity conformation of receptor site-3 involves
changes in external channel regions, which are affected
by alterations in membrane potential or binding of toxins
to topologically distinct receptor sites on the channel
protein [28]. The m echanisms involved i n this a ffinity
change are different, however, as depolarization affects
mainly the association rate w hereas allosteric modulation
by other toxins affects mainly the dissociation rate
constant [21,27,41].
Resemblance of the putative bioactive surfaces
of LqhaIT, Aah-II, and d-ACTX-Hv1a
Despite the difference in three-dimensional structure,
sequence, and size, the similarity in binding properties of
scorpion a-toxins and d-ACTX-Hv1a suggests some struc-
tural resemblance at the b ioactive surface. However, the
variations in k
on
and unusual binding selectivity of the
toxins to receptor site-3 may result from either variations at
the bioactive surface, or other, yet unidentified, structural
differences. In search for possible resemblance of molecular
exteriors, we co mpared LqhaIT [17], A ah-II (which i s
almost identical to Lqh-II [16,24]) and d-ACTX-Hv1a [1]

focusing on residues s hown in LqhaIT to constitute the
bioactive surface (Fig. 6, left [52,53]). A number of bioactive
residues of LqhaIT appear also on the surface of Aah-II in a
similar position, and interestingly, a lso appear on th e
surface of d-ACTX-Hv1a (Fig. 6). The positively charged
Lys3, Lys4, Arg5, and Lys10 of d-ACTX-Hv1a are oriented
similarly to Lys8, Arg58, Lys62, and Arg18 of LqhaIT. The
aromatic Trp7 and Tyr25 in d-ACTX-Hv1a resemble to
some extent Trp38 and Phe17 in LqhaIT or Trp38 and
Phe15 in Aah-II. The nonpolar Asn6 in d-ACTX-Hv1a
occupies a similar position to Asn44 in LqhaIT or Aah-II.
Finally, the negatively charged Glu12 in d-ACTX-Hv1a
resembles Glu24 in both LqhaIT and Aah-II.
Despite this similarity, the shape of d-ACTX-Hv1a at the
angle presented in Fi g. 6, is slimmer than that of t he
scorpion toxins, which may explain its accessibility, with
high binding affinity, to both r at brain and cockroach
receptor site-3. T his possibility m ay also explain t he smaller
decrease in association rate to r eceptor s ite-3 b etween
polarized and depolarized rat brain synaptosomes (Fig. 2)
compared to Lqh-II. Conformational changes in the sodium
channel s tates that are associated with a s hift from the high
to the low affin ity s tate of r eceptor site-3 b y depolarization
may involve steric hindrance for toxin a ccess, which is l ess
pronounced for the slimmer d-ACTX. This hypothesis is in
concert with the smaller depolarization effect on the
association rate, k
on
(Fig. 3) of d-ACTX compared to that
of Lqh-II (Table 1). In light of the structural resemblance at

the putative bioactive surface between the spider and
scorpion toxins, t he subtle variations in action and b inding
properties [7,30], suggest that d-ACTXs interact with the
sodium channel at a nonidentical, yet overlapping site to
that of scorpion a-toxins.
Differences between cockroach and locust receptor
site-3
All of the site-3 toxins that compete f or classical s corpion
a-toxin binding to rat b rain sodium channels compete for
LqhaIT binding to cockroach sodium channels but with
different potencies [29,30,35]. The s pider d-ACTXs are
unique in that they bind with equally high affinity to
receptor site-3 of both rat brain and cockroach sodium
channels (Figs 2 and 4 [29,30]). Such broad potency for
sodium channels of distant phyla could be related to the
slimmer shape of the spider toxin compared with the bulkier
appearance of scorpion a-toxins (Fig. 6).
d-ACTXs are similar to L qhaIT in toxicity symptoms of
injected insects [30,35,54], and the inactivation of the
sodium current in cockroach neuronal preparations [5,54].
Therefore, the substantial difference in d-ACTX-Hv1a
binding affinity for cockroach vs. locust sodium channels
is surprising. Structural differences between the two i nsect
Fig. 6. Structural comparison of scorpion a-toxins and d-atracotoxin-H v1a. The structures for Aah-II ([16], PDB accession code 1PTX), d-ACTX-
Hv1a ([11]; 1V TX ); and L q haIT ([17]; 1LQH) are presented w ith a similar orientation of their p utative bioactive surfaces . Residues r eported to
participate in the bioactive surface of LqhaIT [52,53] together with t opolo gically related re sidue s in Aah -II and d-ACTX-Hv1a are highlighted and
colour coded: blue, positively charged; green, aromatic; magenta, Asn; yellow, negatively charged (see text for details). Toxin models were prepared
USING SW ISS
-
PDB VIEWER

[57].
Ó FEBS 2002 Binding of d-atracotoxins on sodium channels (Eur. J. Biochem. 269) 1507
sodium channels have been previously inferred from
allosteric modulations of LqhaIT binding. Brevetoxin
(site-5 toxin) and veratridine ( site-2 toxin) increase the
binding of LqhaIT to locust but not t o cockroach sodium
channels [27,35,55]. In a ddition, w e h ave shown that Lqh-
III, which binds to locust and cockroach receptor s ite-3
equally well, lost its binding capacity to locust sodium
channels in its iodinated form [36]. Regardless of iodination,
however, the spider toxin binds with high affinity to
cockroach and rat brain channels and with low affin ity to
locust receptor site-3. Of note is that LqhaIT and Lqh-III
bind very poorly to rat brain sodium channels [49,56] in
contrast with the high affinity binding of d-ACTX- Hv1a
(Table 1). Hence, various toxin probes may expose subtle
differences at receptor site-3. S till, the structural basis for
these selective interactions requires determination of contact
surfaces between the various ligands and their receptor
binding sites.
Concluding remarks
These results suggest that toxin selectivity for r eceptor site-3
may be conferred not only by structural variables at the
direct interacting surface, but also by external channel
elements that may affect toxin access to the binding site. We
have shown that toxins with different shapes or electrostatic
surface potentials are capable of reaching the same or
overlapping receptor sites (this study and see also
[35,42,56]). Thus, structural differences in channel regions
that flank the putative binding site could also be involved in

toxin binding. This assumption is supported by the differ-
ence between the association r ates of Lqh-II and d-ACTX
to polarized synaptosomes (which mainly account for the
difference in their affinity) and the changes detected in k
on
to
different channel states (see Table 1). Also, the differences in
association rates of Lqh-II, Lqh-III and LqhaIT to rat
skeletal muscle channels, expressed in mammalian cells [46],
support this notion. The rational design of su btype-specific
compounds will take into account such architectural
considerations, which may facilitate the design of subtype
specific drugs and insecticides.
ACKNOWLEDGEMENTS
This work was s upported i n part b y a n Australian R esearch C ouncil
research grant and an UTS internal research grant (to G. M. N.), by a
research grant from t he Israeli Science Foundation (508/00, to D. G.)
and by grants from BARD, The United States-Israel Binational
Agricultural Rese arch & D evelopment (IS -2901–9 7C, to M . G . and
IS-3259–01 to D. G.).
REFERENCES
1. Fletc her, J .I., S mith, R., O’Donoghue, S.I., N ilges, M., Connor,
M., Howden, M.E.H., Christie, M.J. & King, G.F. (1997) The
structure o f a novel insecticidal neurotoxin, omega-atracotoxin-
HV1, from the venom of an Australian funnel web spide r. Nat.
Struct. Biol. 4, 559–566.
2. Wang,X H.,Smith,R.,Fletcher,J.I.,Wilson,H.,Wood,C.J.,
Howden,M.E.H.&King,G.F.(1999)Structure–functionstudies
of omega-atracotoxin, a potent antagonist of insect voltage-gated
calcium channels. Eur. J. Biochem. 264, 488–494.

3. Sze to, T.H., Wang, X H., Smith, R., Connor, M., Christie, M.J.,
Nicholson, G.M. & King, G.F. (2000) Isolation of a funnel web
spider polypeptid e with homology to mamb a intestinal toxin 1 and
the e mbryonic head in ducer Dickkopf1. Toxicon 38, 429–442.
4. Wang, X H., Connor, M., Smith, R., Maciejewski, M.W.,
Howden, M.E.H., Nicholson, G.M., Christie, M.J. & King, G.F.
(2000) Discovery and characterization of a family of insecticidal
neurotoxins with a rare vicinal disulfide bridge. Nat. Struct. Biol. 7,
505–513.
5. Grolleau, F., Stankiewicz, M., Birinyi-Strachan, L.C., Wang,
X H.,Nicholson,G.M.,Pelhate,M.&Lapied,B.(2001)Elec-
trophysiological analysis of the neurotoxic action of a funnel-web
spider toxin, d-atracotoxin-Hv1a, on insect vo ltage-gated sodium
channels. J. Exp. Biol. 204, 711–721.
6. Nicholson, G.M., Willow, M., Howden, M.E.H. & Narahashi, T.
(1994) Modification of s odium channel gating and kinetics by
versutoxin from the Australian funnel-web spider Hadronyche
versuta. Pflugers Arch. 428, 400–409.
7. Nicholson, G.M., Walsh, R., Little, M.J. & Tyler, M.I. (19 98)
Characterisation of the effects of robustoxin, the lethal neurotoxin
from the Sydney fun nel- web spider Atrax robustus,onsodium
channel a ctivation and inactivation. Pflug ers Arch. 436, 117–126.
8. Szeto, T.H., Birinyi-Strachan, L.C., Wang, X H., Smith, R.,
Connor, M., Christie, M.J., King, G.F. & Nichols on, G.M. (2000)
Isolation and pharmacological characterisation of d-atracoto xin-
Hv1b, a vertebrate-selective sodium channel toxin. FEBS Lett.
470, 293–299.
9. Brown , M.R., S heumac k, D.D., T yler, M.I. & Howden, M.E.H.
(1988) Amino acid sequence of versutoxin, a lethal neurotoxin
from the ven om of the funnel-web spider Atrax versutus. Biochem.

J. 250, 4 01–405.
10. She umack, D.D., Claassens, R., White ley, N.M. & Howden,
M.E.H. (1985) Complete amin o acid sequence of a new type of
lethal neurotoxin from the venom of the funnel-web spider Atrax
robustus. FEBS Lett. 18 1, 154–156.
11. Fletcher, J.I., Chapman, B. E., Mackay, J.P., Howden, M.E.H. &
King, G.F. (1997) The structure of versutoxin (d-atracoto xin-
Hv1): implications for binding of site-3 to xins to t he voltage-gated
sodium channel. Structure 5, 1525–1535.
12. Pallaghy, P .K., Alewood, D., Alewood, P.F. & Norton, R.S.
(1997) Solution structure of robustoxin, the lethal neurotoxin f rom
the f unnelweb spider Atrax robustus. FEBS Lett. 419, 191–196.
13. Pallaghy, P.K., Neilsen, K.J., Craik, D.J. & Norton, R.S. (1993) A
common structural motif incorporating a cystine k not and a triple-
stranded bet a-sheet in toxic and inh ibitory po lyp eptides. Protein
Sci. 3, 1833–1839.
14. Hanc k, D.A. & Sheets, M.F. (1995) Modification of i nactivation
in cardiac sodium channels: ionic current studies with antho-
pleurin-A toxin. J. Gen. Physiol. 106, 601–616.
15. Strich artz, G.R. & Wang, G.K. (1986) Rapid voltage-dependent
dissociation of scorpion a-toxins c oupled to Na channel inacti-
vation in amphibian myelinated nerves. J. Gen. Physiol. 88 ,
413–435.
16. Fontecilla-Camps, J L., Haberse tzer-Roch at, C. & Rocha t, H.
(1988) Orthorhombic c rystals and three dimensional s tructure of
the poten t t oxin II from the scorpion Androctonus Australis
Hector. Proc. Natl Acad. Sci. USA 85, 7443–7447.
17. Tugarinov, V., Kustanovich, I., Zilberberg, N., Gurevitz, M . &
Anglister, J. (1997) Solution structures of a highly insecticidal
recombinant scorpion alpha-toxin and a mutant with i ncreased

activity. Biochemistry 36 , 2414–2424.
18. Monks, S.A., Pallaghy, P.K., S canlon, M.J. & N orton, R.S. (1995)
Solution structure of the cardiostimulant polypeptide antho-
pleurin-B and comparison with anthopleurin-A. Structure 3,
791–803.
19. Gordon, D. (1997) A new approach to insect-pest control –
combination of neurotoxins interacting with voltage sensitive
sodium channels to increase selectivity and specificity. Invertebr.
Neurosci. 3, 103–116.
1508 N. Gilles et al. (Eur. J. Biochem. 269) Ó FEBS 2002
20. Catterall, W.A. (1986) Molecular properties of voltage-sensitive
sodium ch annels. Annu. Rev. Biochem. 55 , 953–985.
21. C atterall, W.A. (1992) Cellular and molecular biology of voltage-
gated sodium channels. Phys iol. Rev. 72 (Suppl.), S 15–S48.
22. G ordon, D. (1997) S odium 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. Harwood Press,
Amsterdam, th e Netherlands.
23. M artin-Eauclaire, M F. & Couraud, F. (1995) Scorpion neuro-
toxins: Effects and mechanisms. In Handk. Neurotoxicology
(Chang, L.W. & D yer, R.S., eds), pp. 683–716. Marcel Dekker,
New York.
24. Sautie
¨
re, P ., Ceste
¨
le, S., Kopeyan, C., Martinage, A ., Drobecq, H .,
Doljansky, Y. & Gordon, D. (1998) New toxins acting on sodium
channels from the scorpion Leiurus quinquestriatus hebraeus

suggest a clue t o mammalian vs insect selectivity. Toxicon 36,
1141–1154.
25. G ordon, D., S avarin, P., Gurevitz, M. & Zinn-Justin, S. (1998)
Functional anatomy of s corpion toxins affecting s odium channels.
J. Toxicol. Toxin. Rev. 17 , 131–159.
26. C atterall, W.A. (1977) Activation of the action poten tial Na
+
ionophore by neurotoxins. J. Biol. Chem. 252, 8669 –8676.
27. Ceste
¨
le, S., Ben Khalifa, R., Pelhate, M., Rochat, H. & Gordon,
D. (1995) a-Scorpion toxins binding on rat brain and insect
sodium channels rev eal di vergent a llosteric m odulations by
brevetoxin and veratridine. J. Biol. Chem. 270, 15153–15161.
28. Ceste
¨
le, S. & Gordon, D. (1998) Depolarization differentially
affects allosteric modulation by neurotoxins of s corpion alpha-
toxin b inding on voltage-gated s odium channels. J. Neurochem.
70, 1217–1226.
29. L ittle, M.J., Wilson, H., Zappia, C., Ceste
¨
le, S., Tyler, M.I.,
Martin-Eauclaire, M F., Gordon, D . & N icholson, G.M. (1998)
d-Atracoto xins from Australian funnel-web spiders compete with
scorpion a-toxin binding on both rat brain and insect sodium
channels. F EBS Letters 439, 246–252.
30.Little,M.J.,Zappia,C.,Gilles,N.,Connor,M.,Tyler,M.I.,
Martin-Eauclaire, M F., Gordon, D . & N icholson, G.M. (1998)
d-Atracoto xins from Australian funnel-web spiders compete with

scorpion a-toxin b inding but differen tially modulate alkalo id toxin
activation of v oltage-gated sodium channels. J. Biol. Chem. 273,
27076–27083.
31. Zilberberg,N.,Gordon,D.,Pelhate,M.,Adams,M.E.,Norris,
F., Z lotkin, E. & Gurevitz, M. (1996) Functional expression and
genetic m odification o f an alp ha scorpion neurotoxin. Bioc hem-
istry 35, 10215–10222.
32. K anner, B.I. (1978) A ctive transport o f gamma-aminobutyric acid
by membrane vesicles isolated from rat b rain. Biochemistry 17,
1207–1211.
33. K rimm, I., Gilles, N., Sautiere, P., Stankiewicz, M., Pelh ate, M.,
Gordon, D. & Lancelin, J.M. (1999) NMR structures and activity
of a novel alp ha-like t oxin from the scorpion. Leiurus Quinques-
triatus Hebraeus. J. Mol. Biol. 285, 1749–1763.
34. G ordon, D., Mosko witz, H. & Zlotkin, E. (1990) Sodium channel
polypeptides in central nervous systems of various insects identi-
fied with site directed antib odie s. Biochim. Biophy. Acta 1026, 80–
86.
35. G ordon, D., Martin-Eauclaire, M F., Ceste
¨
le, S., Kopeyan, C.,
Carlier, E., Ben-Khalifa, R., Pelhate, M. & Rochat, H. (1996)
Scorpion toxins affecting sodium channel current inactivation
bind to distinct homologous receptor sites on rat brain and insect
sodium ch annels. J. Biol. Chem. 271, 8 034–8045.
36. Gilles,N.,Krimm,I.,Bouet,F.,Froy,O.,Gurevitz,M.,Lancelin,
J.M. & Gordon, D. (2000) Struc tural implications on the inter-
action of scorpion alpha-like toxins with the sodium channel
receptor site inferred from t oxin iodination and pH-dependent
binding. J. Neurochem. 75 , 1735–1745.

37. C heng, Y. & Prusoff, W. (1973) Relationship between inhibition
constant (K
I
) and the concentration of an inhibitor which causes
50% inhibition (IC
50
)ofanenzymaticreaction.Biochem. Phar-
macol. 22, 3099–3108.
38. Weiland, G.A. & Molinoff, P.B. ( 1981) Quantitative analysis of
drug–receptor interactions. I. Determination of kinetic a nd equi-
librium properties. Life S cienc es 29, 313–330.
39. Catterall, W.A. (1976) Purification of a tox ic protein from scor-
pion venom w hich activates the action potential Na
+
ionophore.
J. Biol. Chem. 251, 5528–5536.
40. J over, E ., Courand, F. & Rochat, H. (1980) Two types of scorpion
neurotoxins c haracterized by their binding t o two separate
receptor sites on rat brain synaptosomes. Biochem. Bioph ys. Res.
Comm. 95, 1607–1614.
41. Ray, R. & Catterall, W.A. (1978) Membrane potential dependent
binding of scorpion toxin to the action potential sodium iono-
phore. Studies w ith a 3-(4-hydr oxy 3[
125
I] io dophenyl) propionyl
derivative. J. Neurochem. 31 , 397–407.
42. G illes, 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. B i och emi st ry 40, 14576–14584.

43. Blaustein, M.P. & Goldring, J.M. (1975) Membrane potentials in
pinched-off presynaptic nerve terminals monitored with a fluo-
rescent p robe: evidence that synaptosomes have potassium diffu-
sion potentials. J. Physiol. 247 , 589–615.
44. Ray, R., Morrow, C.S. & Catterall, W.A. (1978) Binding of
scorpion toxin t o receptor s ites associated with voltage-sensitive
sodium channels in synaptic nerve ending particles. J. Biol. Chem.
253, 7307–7313.
45. Catterall, W.A., Morrow, C.S. & Hartshorne, R.P. (1979) Neu-
rotoxin binding to receptor sites a ssociated with v oltage-sensitive
sodium channels in intact, lysed, and detergent-solubilized brain
membranes. J. Biol. Chem. 254, 11379–11387.
46. C hen, H., Gordon, D. & Heinemann, S.H. (2000) Modulation of
cloned skeletal muscle so dium channels by th e scorpion to xins
Lqh-II, Lqh-III, and LqhaIT. Pflugers Arch. 43 9 , 423–432.
47. Rogers,J.C.,Qu,Y.,Tanada,T.N.,Scheuer,T.&Catterall,W.A.
(1996) Molecular d eterminants of h igh affinity b inding of a-scor-
pion toxin and sea anemone toxin in the S3–S4 extracellular loop
in domain IV of the Na
+
channel a subunit. J. Biol. Chem. 271,
15950–15962.
48. T rainer, V.L., McPhee, J.C., Boutelet-Bochan, H., Baker, C.,
Scheuer, T., Babin, D ., Demoute, J.P., G uedin, D. & Catterall,
W.A. (1997) High affi nity binding of pytethrods to the a subunit of
brain sodium channels. Mol. Pharmacol. 51, 651–657.
49. Gilles, N., Chen, H., Wilson, H., Legall, F., Montoya, G., Molgo,
J., Schnˆ herr, R., N ic holson, G., Heinemann, S.H. & Go rd on, D.
(2000) Scorpion a-anda-like toxins differentially interact with
sodium channels in mammalian CNS and periphery. Eur. J.

Neurosci. 12 , 2823–2832.
50. S mith, R.D. & Goldin, A.L. (1998) Functional analysis of
the rat I sodium c hannel in xenopus oocyt es. J. Neurosci. 18,
811–820.
51. G ordon, D., Merrick, D., Auld, V., Dunn, R., Goldin, A.L. ,
Davidson, N. & Catterall, W.A. (1987) Tissue-specific e xpression
of the RI and RII sodium channel subtypes. Proc. Natl Acad. Sci.
USA 84, 8682–8686.
52.Zilberberg,N.,Froy,O.,Cestele,S.,Loret,E.,Arad,D.,
Gordon, D. & Gurevitz, M. (1997) Elucidation of the putative
toxic-surface of a highly insecticidal scorpion a-neurotoxin
affecting voltage-sensitive sodium channels. J. Biol. Chem. 272,
14810–14816.
53. Gurevitz,M.,Gordon,D.,Ben-Natan,S.,Turkov,M.&Froy,O.
(2000) Diversification of neurotoxins by C-tail ÔwigglingÕ –a
scorpion recipe for survival. FASEB J. 15, 1 201–1205.
Ó FEBS 2002 Binding of d-atracotoxins on sodium channels (Eur. J. Biochem. 269) 1509
54. Eitan, M., Fowler, E., Herrmann, R., D uval, A., Pelhate, M. &
Zlotkin, E. (1990) A scorpion venom neurotoxin paralytic to
insects th at affects sodium cu rrent inactivation: purification,
primary structure, a nd mo de o f ac tion. Biochemistry 29, 5941–
5947.
55. Gordon, D. & Z lotkin, E. (1993) Binding of an a-scorpion toxin to
insect sodium c hannels is not d epende nt on membrane pote ntial.
FEBS Lett. 315, 125–128.
56. Gilles, N., Blanchet, B., Shichor, I., Zaninetti, M., Lotan, I.,
Bertrand,D.&Gordon,D.(1999)Ascorpiona-like toxin active
on insects and mamm als reveals an unexpected s pecific ity and
distribution o f so dium channel subtypes i n rat brain neuro ns.
J. Neurosci. 19, 8730–8739.

57. Guex, N. & Peitsch, M.C. (1997) SWISS-MODEL and the Swiss-
PdbViewer: An environment for comparative protein mo deling.
Electrophoresis 18, 2714–2723.
1510 N. Gilles et al. (Eur. J. Biochem. 269) Ó FEBS 2002