Báo cáo khoa học: Synthesis and characterization of Pi4, a scorpion toxin from Pandinus imperator that acts on K+ channels doc

Synthesis and characterization of Pi4, a scorpion toxin
from
Pandinus imperator
that acts on K
+
channels
Sarrah M’Barek
1
, Amor Mosbah
1
, Guillaume Sandoz
2
, Ziad Fajloun
1
, Timoteo Olamendi-Portugal
3
,
Herve
´
Rochat
1
, Franc¸ois Sampieri
1
,J.In
˜
aki Guijarro
4
, Pascal Mansuelle
1
, Muriel Delepierre
4

,
Michel De Waard
2
and Jean-Marc Sabatier
1
1
Laboratoire International Associe
´
d’Inge
´
nierie Biomole
´
culaire et Laboratoire de Biochimie CNRS UMR 6560, IFR Jean Roche,
Faculte
´
de Me
´
decine Nord, Marseille, France;
2
Laboratoire Canaux Ioniques et Signalization, Equipe mixte INSERM 99–31,
CEA, Grenoble, France;
3
Institute of Biotechnology, National Autonomous University of Mexico, Cuernavaca, Mexico;
4
Unite
´
de RMN des Biomole
´
cules, De
´

pt. de Biochimie Structurale et Chimie, Institut Pasteur, CNRS URA 2185, Paris, France
Pi4 is a 38-residue toxin cross-linked by four disulfide bridges
that has been isolated from the venom of the Chactidae
scorpion Pandinus imperator. Together with maurotoxin,
Pi1, Pi7 and HsTx1, Pi4 belongs to the a KTX6 subfamily of
short four-disulfide-bridged scorpion toxins acting on K
+
channels. Due to its very low abundance in venom, Pi4 was
chemically synthesized in order to better characterize its
pharmacology and structural properties. An enzyme-based
cleavage of synthetic Pi4 (sPi4) indicated half-cystine pair-
ings between Cys6–Cys27, Cys12–32, Cys16–34 and Cys22–
37, which denotes a conventional pattern of scorpion toxin
reticulation (Pi1/HsTx1 type). In vivo, sPi4 was lethal after
intracerebroventricular injection to mice (LD
50
of 0.2 lg
per mouse). In vitro, addition of sPi4 onto Xenopus laevis
oocytes heterologously expressing various voltage-gated K
+
channel subtypes showed potent inhibition of currents from
rat Kv1.2 (IC
50
of 8 p
M
)andShaker B(IC
50
of 3 n
M
)

channels, whereas no effect was observed on rat Kv1.1 and
Kv1.3 channels. The sPi4 was also found to compete with
125
I-labeled apamin for binding to small-conductance Ca
2+
-
activated K
+
(SK) channels from rat brain synaptosomes
(IC
50
value of 0.5 l
M
). sPi4 is a high affinity blocker of the
Kv1.2 channel. The toxin was docked (
BIGGER
program) on
the Kv channel using the solution structure of sPi4 and a
molecular model of the Kv1.2 channel pore region. The
model suggests a key role for residues Arg10, Arg19, Lys26
(dyad), Ile28, Lys30, Lys33 and Tyr35 (dyad) in the inter-
action and the associated blockage of the Kv1.2 channel.
Keywords: Pi4; scorpion toxin; K
+
channels; half-cystine
pairings; molecular docking.
Pi4 is a K
+
channel-acting toxin that was isolated from the
venom of scorpion Pandinus imperator [1]. It is a 38-mer

peptide cross-linked by four disulfide bridges, and therefore
belongs to the aKTX6 subfamily [2] of short-chain four-
disulfide-bridged scorpion toxins acting on K
+
channels.
The highest sequence identities of Pi4 are shared with
members of this structural subfamily, i.e. 68% with Mauro-
toxin (Scorpio maurus palmatus) [3,4], 66% with Pi7
(P. imperator)[1],61%withPi1(P. imperator) [5–7] and
45% with HsTx1 (Heterometrus spinnifer) [8,9]. The pri-
mary structure of Pi4 also contains a variant of the
consensus sequence of scorpion toxins (i.e. […]C
1
[…]
C
2
XXPC
3
[…]C
4
[…](A/S/G)XC
5
[…]C
6
XC
7
[…]C
8
instead
of […]C

1
[…]C
2
XXXC
3
[…](G/A/S)XC
4
[…]C
5
XC
6
[…]for
three-disulfide-bridged toxins) that is representative of
toxins from the a KTX6 structural subclass. Both the
classical and variant consensus sequences are thought to be
responsible for folding of toxins according to a common a/b
scaffold [10–13] independent of the toxin size and pharma-
cology (except for Ca
2+
channel-acting toxins which fold
according to an Ôinhibitor cystine knotÕ motif [14,15]).
Recent
1
H-NMR analysis showed that synthetic Pi4 (sPi4)
indeed exhibits the a/b scaffold [16]. This motif, from which
arises the great functional diversity of scorpion toxins, is
mainly composed of an a-helix connected to a b-sheet
structure (two or three strands) by two disulfide bridges.
The first report on native Pi4 by Olamendi-Portugal et al.
[1] provided some preliminary data on its pharmacology:

(a) it blocked completely and reversibly voltage-gated
Shaker B channel expressed in Sf9 insect cells, at 100 n
M
toxin concentration (IC
50
value of 8 n
M
), and (b) it
competed with
125
I-labeled noxiustoxin (Centruroides
noxius) for binding on rat brain synaptosomal membranes,
with an IC
50
value of approximately 10 n
M
.
Here, we report the first chemical synthesis of Pi4 in order
to better characterize its pharmacology on the various K
+
channel subtypes generally recognized by toxins of the
a KTX6 subfamily, i.e. insect Shaker B, rat SK, Kv1.1,
Kv1.2 and Kv1.3. We also investigated the disulfide bridge
Correspondence to J M. Sabatier, Laboratoire de Biochimie CNRS
UMR 6560, et Laboratoire International Associe
´
d’Inge
´
nierie
Biomole

´
culaire, IFR Jean Roche, Faculte
´
de Me
´
decine Nord,
Bd Pierre Dramard, 13916 Marseille Cedex 20, France.
Fax: +33 491 657595, Tel.: + 33 491 698852,
E-mail:
Abbreviations: HsTx1, toxin 1 from the scorpion Heterometrus spin-
nifer; Kv channel, mammalian voltage-gated K
+
channel; Pi1, Pi4,
Pi7, toxin 1, 4 or 7 from the scorpion Pandinus imperator; SK channel,
small-conductance Ca
2+
-activated K
+
channel; sPi4, synthetic Pi4.
(Received 7 April 2003, revised 11 June 2003, accepted 8 July 2003)
Eur. J. Biochem. 270, 3583–3592 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03743.x
organization of sPi4 as two distinct patterns of reticulation
have been described for these toxins hitherto, which are
herein referred to as the maurotoxin (C
1
–C
5
,C
2
–C

6
,C
3
–C
4
and C
7
–C
8
) [3] and Pi1/HsTx1 (C
1
–C
5
,C
2
–C
6
,C
3
–C
7
and
C
4
–C
8
) [7,9] types. Because sPi4 was found to be active at
a picomolar concentration range on rat Kv1.2 channel, we
detailed the interaction of the toxin with the latter channel
by in silico docking experiments. The

1
H-NMR solution
structure of sPi4 [16] and a molecular model of the rat Kv1.2
channel pore region (S5–H5–S6 portion) were used for this
purpose. The data were used to generate a functional map
of Pi4 towards this channel, highlighting some key residues
of this interaction such as those belonging to the toxin
functional dyad (a well-defined pair of residues that is
thought to be crucial for toxin blocking efficacy towards
the voltage-gated K
+
channels).
Experimental procedures
Materials
N-a-Fmoc-
L
-amino acids, 4-hydroxymethylphenyloxy
(HMP) resin, and reagents used for solid-phase chemical
synthesis of Pi4 were purchased from Perkin-Elmer.
Organic solvents were analytical grade products from
SDS. Enzymes (trypsin, and chymotrypsin) were obtained
from Boehringer Mannheim.
Solid-phase synthesis of sPi4
The sPi4 was synthesized by the solid-phase method [17]
using an automated peptide synthesizer (model 433A,
Applied Biosystems Inc.). Peptide chains were assembled
by conventional stepwise synthesis on 0.3 molar equivalents
of HMP resin (1% cross-linked; 0.69 molar equivalents of
amino group g) using 1 mmol of N-a-fluorenylmethyloxy-
carbonyl (Fmoc) amino acid derivatives [7,18]. The side-

chain protecting groups of sPi4 trifunctional residues were:
tert-butyl (t-Bu) for Ser, Thr, Tyr, Asp, and Glu; trityl (Trt)
for Cys, Asn, and Gln; pentamethylchroman (Pmc) for
Arg, and tert-butyloxycarbonyl (Boc) for Lys. N-a-Amino
groups were deprotected by treatments with 18% and 20%
(v/v) piperidine/N-methylpyrrolidone for 3 and 8 min,
respectively. The peptide resin was washed with N-methyl-
pyrrolidone (5 · 1 min), then Fmoc-amino acid derivatives
were coupled (20 min) as their hydroxybenzotriazole active
esters (OBt) in N-methylpyrrolidone (3.3-fold excess). After
complete peptide chain assembly and removal of the
N-terminal Fmoc group, the peptide-resin (c. 2.1 g) was
treated under stirring for 3 h at room temperature with a
mixture of 88% trifluoroacetic acid/5% H
2
O/5% thioani-
sole/2% ethanedithiol (v/v) in the presence of crystalline
phenol (2.5 g) in a final volume of 30 mLÆg
)1
of peptide
resin. The peptide mixture was filtered to remove the resin,
and the filtrate was precipitated and washed twice with cold
diethyl ether. After centrifugation at 2800 g for 12 min, the
supernatant was discarded and the crude peptide was
dissolved in H
2
O and lyophilized. Oxidative folding of the
reduced peptide was performed by dissolving the lyophilized
peptide ( 1m
M

final concentration) in 0.2
M
Tris/HCl
buffer, pH 8.4 and gentle stirring under air for 72 h at 25 °C.
The sPi4 was purified to homogeneity by semipreparative
reverse-phase HPLC (Perkin-Elmer, C
18
Aquapore ODS
20 lm, 250 · 10 mm) by means of a 60 min linear gradient
of 0.08% trifluoroacetic acid/0% to 35% acetonitrile (v/v)
in 0.1% trifluoroacetic acid/H
2
O (v/v) at a flow rate of
6mLÆmin
)1
(k ¼ 230 nm). The correct identity and the
high degree of homogeneity of sPi4 were established by:
(a) analytical C
18
reverse-phase HPLC (Chromolith RP18,
5 lm, 4.6 · 100 mm) using a 40-min linear gradient of
0.08% trifluoroacetic acid/0–60% acetonitrile (v/v) in
0.1% (v/v) trifluoroacetic acid/H
2
O (v/v) at a flow rate
of 1 mLÆmin
)1
; (b) amino acid composition after acidolysis
[6
M

HCl/2% phenol (w/v), 20 h, 118 °C, N
2
atmosphere);
(c) Edman sequencing; and (d) molecular mass analysis by
MALDI-TOF mass spectrometry.
Assignment of sPi4 half-cystine pairings
The sPi4 (800 lg) was incubated with a mixture of trypsin
and chymotrypsin at 10% (w/w) each, in 0.2
M
Tris/HCl
buffer, pH 7.4 (14 h, 37 °C). The resulting peptides were
separated by analytical reverse-phase HPLC (Chromolith
RP18, 5 lm, 4.6 · 100 mm) with a 60-min linear gradient
of 0.08% (v/v) trifluoroacetic acid/0–40% acetonitrile in
0.1% (v/v) trifluoroacetic acid/H
2
O at a flow rate of
1mLÆmin
)1
(k ¼ 230 nm), and freeze-dried prior to ana-
lysis. The peptides were treated by acidolysis (6
M
HCl/
phenol) and their amino acid contents were determined
(Beckman, System 6300 amino acid analyzer). The peptides
were also analyzed by mass spectrometry (RP-DE Voyager,
Perseptive Biosystems), and Edman sequencing using a
gas-phase microsequencer (Applied Biosystems 470A).
In standard HPLC conditions for analyzing phenylthio-
hydantoin (PTH) amino acid derivatives, diPTH-cystine

eluted at a retention time of 9.8 min.
Lethal activity of sPi4 in mice
The sPi4 was tested in vivo for neurotoxicity by determining
the 50% lethal dose (LD
50
) after intracerebroventricular
injection of 20 g C57/Bl6 mice (animal testing agreement
number 006573 granted by the department Sante
´
et
Protection Animales, Ministe
`
re de l’Agriculture et de la
Peˆ che). Groups of four to six mice per dose were injected
with 5 lL sPi4 solution containing 0.1% bovine serum
albumin and 0.9% sodium chloride (w/v).
Binding of sPi4 on SK channels from rat brain
synaptosomes
Rat brain synaptosomes (P2 fraction) were prepared
according to Gray and Whittaker [19]. Protein content
was assayed by a modified Lowry method. [
125
I]Apamin
(2000 CiÆmmol
)1
) was obtained as described [20]. Aliquots
of 50 lL0.1n
M
[
125

I]apamin were added to 400 lLof
synaptosome suspension (0.4 mg proteinÆmL
)1
). The sam-
ples were incubated for 1 h at 0 °Cwith50lL of one of a
series of concentrations of sPi4 or apamin (10
)2
)10
)13
M
)in
500 lL final volume. The incubation buffer was 25 m
M
Tris/HCl, 10 m
M
KCl, pH 7.2. The samples were centri-
fuged and the resulting pellets were washed three times with
1 mL of the same buffer. Bound radioactivity was counted
(Packard Crystal II). Reported values represent the means
3584 S. M’Barek et al. (Eur. J. Biochem. 270) Ó FEBS 2003
of triplicate experiments. Nonspecific binding was evaluated
in the presence of an excess (10 n
M
) of unlabeled apamin
and was shown to be less than 8% of the total binding.
Electrophysiology
Oocyte preparation. Xenopus laevis oocytes (stages V and
VI) were recovered and prepared for cRNA injection and
electrophysiological recordings as described [21]. The oocyte
follicular cell layers were removed by enzyme treatment

using 2 mgÆmL
)1
collagenase IA (Sigma) in classical Barth’s
medium [in m
M
:88NaCl,1KCl,0.82MgSO
4
,
0.33 Ca(NO
3
)
2
,0.41CaCl
2
, 2.4 NaHCO
3
,
15
N-2-hydroxy-
ethylpiperazine-N¢-ethanesulphonic acid (Hepes), pH 7.4
with NaOH]. The plasmids were cleaved with SmaI
(Shaker B), Not1(ratKv1.1),Xba1 (rat Kv1.2) and EcoR1
(rat Kv1.3). The linearized plasmids were transcribed by
means of a T7 or SP6 mMessage mMachine kit (Ambion).
The cRNAs (1 lgÆlL
)1
) were kept frozen in H
2
Oat)80 °C.
The cells were microinjected 2 days later with 50 nL of

cRNA (0.2 lgÆlL
)1
Shaker B, rat Kv1.1, Kv1.2, or Kv1.3
channels). To favor ion channel expression, cells were
incubated at 16 °C in a defined nutrient oocyte medium [22]
2 to 6 days before current recordings.
Electrophysiological recordings. A standard two-micro-
electrode technique was used to record oocyte currents (20–
22 °C). The current and voltage electrodes were filled with
140 m
M
KCl and had resistance values ranging between 0.5
and 1 MX. The recordings of potassium currents were
performed using a voltage-clamp amplifier (GeneClamp
500, Axon Instruments, Foster City, CA, USA) interfaced
with a 16-bit AD/DA converter (Digidata 1200 A, Axon
Instruments) for acquisition and voltage protocol applica-
tion. Current records were sampled at 10 kHz and low pass-
filtered at 2 kHz using an eight-pole Bessel filter. The data
were stored on a computer for subsequent analysis.
The extracellular recording solution contained (in m
M
):
88 NaCl, 10 KCl, 2 MgCl
2
, 0.5 CaCl
2
, 0.5 niflumic acid,
5 Hepes, pH 7.4 (NaOH). Leak and capacitive currents were
subtracted on-line by a P/4 protocol. Residual capacitive

artifacts were blanked for display purposes. The sPi4
solutions were perfused in the recording chamber at a flow
rate of 2 mLÆmin
)1
using a ValveBank4 apparatus (Auto-
mate Scientific Inc.). Bovine serum albumin (0.1%) was
added to the recording and perfusion solutions to prevent
toxin loss to the plastic chamber and tubules and nonspecific
binding onto the cell. Data analysis was performed using the
software
PCLAMP
6.0.3 (Axon Instruments, Foster City, CA,
USA). Results are presented as mean ± SEM.
Molecular modeling of rat Kv1 channel subtypes
Molecular modeling of rat Kv1 channel subtypes was based
on the crystal structure of KcsA solved at 2.0 A
˚
resolution
(PDB accession number 1K4C) [23]. The amino acid
sequences from residues 323–422 (corresponding to
domains S5–H5–S6) of the rat voltage-gated K
+
channel
a-subunits [24] were aligned with region 26–125 of KcsA.
The sequence identities between the Kv-type ion channel
regions and KcsA are approximately 30%. For the regions
between the two transmembrane segments S5 and S6, the
sequence identities with KcsA are approximately 50%.
Based on the high degree of similarity, the S5–H5–S6
regions of rat Kv1.1, Kv1.2, and Kv1.3 channels (Swiss-Prot

accession numbers P10499, A33814 and P15384, respect-
ively) were modeled by homology methods. Sequences were
aligned with clustal [25], the residue mutations were
introduced in the KcsA channel structure with the
TURBO
-
FRODO
software [26], and the structures of the Kv channels
thus obtained were minimized by using CNS [27]. The final
molecular models of the S5–H5–S6 regions of the rat Kv1.1,
Kv1.2 and Kv1.3 channels adopt 3D structures which are
similar to that of the KcsA channel (not shown). The
transmembrane a-helices are comprised between residues
322–344 (S5) and 383–413 (S6). The pore region contains
one a-helix between residues 360–370. Of note, the ion
conducting pathway region is formed by four S5–H5–S6
domains. The quality of all molecular models was assessed
using the programs
WHAT IF
[28] and
PROCHECK
[29]. In each
case, the stereochemical quality and the Ramachandran [30]
scores were good and similar to that of the template.
Docking of Pi4 on rat Kv1.2 channel
The experimental 3D structure of sPi4 in solution, as
recently solved by
1
H-NMR [16], was used in docking
experiments together with the molecular models we gener-

ated for rat Kv channel subtypes (S5–H5–S6 domains). The
molecular interaction simulations were performed using
BIGGER
, a docking program [31]. The algorithm used by
BIGGER
performs a complete and systematic search for
surface complementarity (both geometry complementarity
and amino acid residue pairwise affinities are considered)
between two potentially interacting molecules, and enables
an implicit treatment of molecular flexibility. In each case,
a population of 1000 candidate protein–protein-docked
geometries was selected by
BIGGER
. In a subsequent step, the
docked structures were ranked using an interaction scoring
function, which combines several interaction terms that are
thought to be relevant for the stabilization of protein
complexes: geometric packing of the surfaces, electrostatic
interactions, desolvation energy, and pairwise propensities
of the residue sidechains to contact across the molecular
interface. In the ab initio simulations, the entire molecular
surface was searched using absolutely no additional infor-
mation regarding the binding sites. Among the 1000
candidate protein–protein-docked geometries selected, the
five best scoring Pi4–Kv1.2 channel complexes were further
treated with the
TURBO
-
FRODO
software, taking also into

account the proposed functional maps of voltage-gated K
+
channel-acting scorpion toxins [32–38]. Finally, a rigid body
minimization was used to minimize the selected complexes.
The best energy solutions, corresponding to the most
favorable Pi4–Kv1.2 channel complexes, were selected. The
de visu analysis was carried out using the
TURBO
-
FRODO
software, and the geometric quality of the structures was
assessed by
PROCHECK
3.3 [39].
Results and discussion
Pi4 (Fig. 1) shares its highest sequence identities with
scorpion toxins belonging to the a KTX6 subfamily:
maurotoxin (68%) [3,4], Pi7 (66%) [1], Pi1 (61%) [5–7],
Ó FEBS 2003 Synthesis and pharmacology of Pi4 (Eur. J. Biochem. 270) 3585
and HsTx1 (45%) [8,9]. This subclass of K
+
channel
blockers contains short-chain toxins, from 34 to 38 amino
acid residues, cross-linked by four disulfide bridges (instead
of the three disulfide bridges generally observed in other K
+
channel-acting scorpion toxins). To better characterize Pi4,
we chemically synthesized the toxin by using the solid-phase
technique [17].
Production of synthetic Pi4 (sPi4)

The Pi4 backbone assembly was achieved stepwise on
0.3 mmol HMP resin by means of an optimized Fmoc/t-Bu
chemical strategy [18]. We found that the amount of target
peptide linked to the resin was 0.21 mmol, which indicates a
70% yield of peptide assembly. Accordingly, a relative
homogeneity of the crude reduced peptide was observed
after final acidolytic treatment, as assessed by analytical C
18
reverse-phase HPLC (Fig. 2). The crude peptide was folded/
oxidized for 72 h in alkaline conditions using a standard
oxidative buffer [7], and the main oxidized species (sPi4) was
purified to >99% homogeneity by preparative C
18
reverse-
phase HPLC (Fig. 2). An amino acid analysis of this
purified product showed amino acid ratios that were
consistent with the values deduced from the Pi4 primary
structure (Fig. 3). The mass spectrometry analysis
(MALDI-TOF technique) of the peptide gave an experi-
mental M
r
(M + H)
+
of 4180.7, which is similar to the M
r
(M + H)
+
of 4180.9 calculated for Pi4 from its sequence.
The identity and homogeneity of sPi4 were also verified by
Edman sequencing (data not shown). The total yield of sPi4

synthesis (which combines yields of peptide assembly, final
acidolytic treatment, oxidative folding and peptide purifi-
cation), was approximately 3% (9 lmol).
Disulfide bridge pattern of sPi4
In order to assign half-cystine pairings, sPi4 was treated by
a mixture of enzymes (trypsin and chymotrypsin). The
resulting peptide fragments were separated by HPLC, then
characterized by amino acid analysis, Edman sequencing
and mass spectrometry (Fig. 4). The results obtained from
these techniques unambiguously map the half-cystine pair-
ings as Cys6–Cys27, Cys12–Cys32, Cys16–Cys34 and
Cys22–Cys37. Therefore, the disulfide bridge organization
of sPi4 is of the conventional type being C
1
–C
5
,C
2
–C
6
,C
3
–
C
7
and C
4
–C
8
, a pattern identical to that of Pi1 or HsTx1

but different from that of maurotoxin. Of note, the disulfide
bridges of natural Pi4 have been determined by NMR and
structure calculations, as well as by Edman sequencing/mass
spectrometry identification of peptides obtained by proteo-
lysis of natural Pi4 [16]. Both approaches have indicated
half-cystine pairings between Cys6–Cys27, Cys12–Cys32,
Cys16–Cys34 and Cys22–Cys37 for the natural toxin,
consistent with the disulfide bridge arrangement found
experimentally for sPi4. As expected, analysis of 2D
1
H-
NMR spectra of natural Pi4 and sPi4 [16] indicates that
both peptides have the same structure.
Biological properties of sPi4
In our bioassays, we tested sPi4 rather than its natural
counterpart as the latter is present in too low abundance in
the venom of scorpion P. imperator to allow a detailed
analysis of its structural and pharmacological properties.
In vivo, sPi4 injected intracerebroventricularly produced
Fig. 1. Amino acid sequence (one-letter code) of Pi4 and comparison
with other related scorpion toxin sequences. The amino acid sequences
of Pi4 (P. imperator) [1], maurotoxin (S. maurus palmatus) [4], Pi7
and Pi1 (P. imperator) [1,5–7], and HsTx1 (H. spinnifer)[8]were
aligned according to the eight half-cystine residues. The positions of
half-cystine residues are highlighted in gray boxes and numbered from
N- to C-terminus. The asterisk indicates a C-terminal carboxylami-
dated extremity.
Fig. 2. Analytical C
18
reverse-phase HPLC

profiles of Pi4 at different stages of its chemical
synthesis. (A) The crude reduced peptide after
final trifluoroacetic acid treatment. (B) The
crude peptide after 72 h folding/oxidation.
(C) The purified folded/oxidized peptide, sPi4.
For conditions, see Experimental procedures.
3586 S. M’Barek et al. (Eur. J. Biochem. 270) Ó FEBS 2003
lethal effects in mice, with an LD
50
value of 0.2 lgper
mouse. This activity is identical to that of Pi1, but 2.5-fold
less potent than that of maurotoxin. The lethal effects found
for sPi4 are about 10-fold less potent than those produced by
K
+
channel-acting scorpion toxins reticulated by three
disulfide bridges [40,41]. The sPi4-induced neurotoxic
symptoms resembled those of other K
+
channel-acting
scorpion toxins suggesting that it targets some K
+
channels.
In vitro, we first tested the ability of sPi4 to compete with
125
I-labeled apamin for binding on rat brain synaptosomes.
Figure 5 illustrates an sPi4-induced, concentration-depend-
ent, inhibition of
125
I-labeled apamin binding, with an IC

50
value of 0.5 ± 0.2 l
M
. At a similar concentration, unlabe-
led apamin produced a complete inhibition (IC
100
)of
125
I-
labeledapaminbinding,withanIC
50
value of 6 ± 3 p
M
.
The half-effect of sPi4 occurred at a 100- or 10 000-fold
higher concentration than that required for maurotoxin
[3,4,42] or Pi1 [7], respectively, indicating a low, but
significant, affinity interaction of sPi4 with rat brain
apamin-sensitive SK channels. Of note, HsTx1 was reported
to be inactive for binding on these SK channels [8], whereas
Pi7 binding capability on the latter has not been described.
The blocking efficacy of sPi4 was also investigated by
electrophysiology on different subtypes of K
+
channels that
were heterologously expressed in Xenopus laevis oocytes.
We focused on K
+
channels recognized by toxins from the
a KTX6 subfamily, i.e. Shaker B, rat Kv1.1, Kv1.2 and

Kv1.3 subtypes, and studied the putative sPi4-induced dose-
dependent inhibition of currents associated with these
channels. Figure 6A shows that the application of 100 n
M
sPi4 blocked over 80% of Shaker B currents. The dose–
response curves for sPi4 current inhibition were obtained for
Shaker B (Fig. 6B), rat Kv1.2 (Fig. 6C), rat Kv1.1 and
Kv1.3 channels (Fig. 6D). The IC
50
values of sPi4 were
3.0 ± 2.2 n
M
for Shaker B(n ¼ 45) and 8 ± 5 p
M
for rat
Kv1.2 (n ¼ 55) channels, whereas it had no detectable effect
at concentrations up to 10 l
M
on rat Kv1.1 (n ¼ 10) and
Kv1.3 (n ¼ 10) channels. It is noteworthy that the maximal
extent of blockage of K
+
currents by sPi4 was approxi-
mately 80% (Shaker B) or 60% (rat Kv1.2) of total currents.
Similar partial current blocks (from 50% to 80% of total
currents) of these channels have been described for several
scorpion toxins and their analogs, including Pi1 and
maurotoxin [7,42–45]. This phenomenon remains difficult
to explain, but an incomplete permeability block, possibly
associated to an imperfect ion channel pore occlusion, can

Fig. 4. Assignment of sPi4 half-cystine pairings. (A) Characterization
of the sPi4-derived peptides that were generated by enzyme-based
cleavage of sPi4 (see Experimental procedures). Retention times in
HPLC (left column) and identified half-cystine pairings (right column)
of the peptides are shown. (B) Complete disulfide bridge organization
of sPi4 as experimentally established by proteolysis of the synthetic
toxin. The half-cystine residues are numbered according to their
positions in the Pi4 amino acid sequence. The half-cystine connections
are represented by solid lines.
Fig. 5. Binding of sPi4 on apamin-sensitive SK channels from rat brain
synaptosomes. Concentration-dependent inhibition of binding of
[
125
I]apamin to rat brain synaptosomes by either unlabeled apamin (d)
or sPi4 (s) in a competition assay. B
0
is the binding of [
125
I]apamin
without any other ligand, and B is the binding in the presence of the
indicated concentrations of competitor. Abscissa axis is the logarithm
of the molar concentration of competitor. Nonspecific binding, less
than 8% total binding, was subtracted for the calculation of ratios.
When absent, error bars are within symbol size. The data were fitted to
the equation y ¼ y
o
+ a/[1+exp(–(x ) IC
50
)/b)]. The resulting
IC

50
values are 0.5 ± 0.2 l
M
(sPi4) and 6 ± 3 p
M
(unlabeled
apamin).
Fig. 3. Physicochemical characterization of sPi4. Aminoacidcontent
(uncorrected values) of sPi4 after hydrolysis (118 °C, 20 h, N
2
atmo-
sphere) with 6
M
HCl in the presence of 2% (w/v) phenol. The deduced
amino acid composition is shown in parenthesis. Deduced and
experimental relative molecular masses are indicated.
Ó FEBS 2003 Synthesis and pharmacology of Pi4 (Eur. J. Biochem. 270) 3587
tentatively be proposed. Similar phenomena have also been
reported in a number of cases [42,43,45]. From the
experimental data obtained both in vivo and in vitro,sPi4
appears to be pharmacologically more closely related to Pi1
than to maurotoxin, Pi7 or HsTx1. Indeed, the two toxins
apparently share the same lethal effects and selectivity profile
towards the tested K
+
channel subtypes (SK, Shaker Band
rat Kv1.2 channels), although their binding properties or
blocking efficacies towards these channels are clearly
distinct. At the structural level, this should obviously rely
on marked differences of amino acid sequences between

both toxins, which guide the number and/or spatial
positioning of key functional residues that participate in
the interaction with the ion channel pore protein. These data
strengthen the idea of a multipoint interaction between
scorpion toxins and their target ion channels. As sPi4
blocked rat Kv1.2 channel at low picomolar concentrations,
we examined the interaction between Pi4 and this channel at
the molecular level, using the docking program
BIGGER
[31].
Docking experiments
To perform computed docking experiments, the structure of
sPi4 in solution recently solved from
1
H-NMR data [16] was
used, and specific models of rat voltage-gated K
+
channels
(S5–H5–S6 domains) [24] were generated. According to
the docking simulation (Fig. 7A–C), the toxin–ion channel
complex is stabilized by four salt bridges between the
sidechains of Glu332 of each rat Kv1.2 a-subunit
(Kv1.2 channel is composed of four a-subunits) and
Arg10, Arg19, Lys30 and Lys33 of Pi4. The Lys26 sidechain
of Pi4 enters into the ion channel pore and is surrounded by
the four Asp357 carbonyl oxygen atoms of the P-domain
selectivity filter. Residue Tyr35 of Pi4 is involved in a
hydrophobic cluster of aromatic residues consisting of
Trp344, from one of the four Kv1.2 a-subunits, and Trp345
and Tyr355 of an adjacent a-subunit. The phenol ring of

Tyr35 additionally forms an hydrogen bond with the N
e
of
the Trp344 indole ring. Some hydrophobic interactions are
also likely to occur between Ile28 of Pi4 and Val361 of the
Kv1.2 a-subunit. For comparison, Pi1 was also docked on
the Kv1.2 channel (data not shown). Similar types of low-
energy interactions were found but involving Arg5, Arg12,
Lys24 (dyad), Ile26, Arg28, Lys31 and Tyr33 (dyad)
residues of Pi1. However, in the case of Pi1, toxin
positioning over the channel was different, with a slight
rotation over the channel as compared to Pi4, its Tyr33
being in contact with the cluster of aromatic residues
belonging to the same Kv1.2 a-subunit. Of note, Pi4 (or Pi1)
did not give good scores when assayed for docking on rat
Kv1.1 and Kv1.3 channels (data not shown), in agreement
with its lack of bioactivity on these channels.
Functional maps of Pi4 and Pi1 towards rat Kv1.2
channels
Results from docking experiments allow us to propose
functional maps for both Pi4 and Pi1 regarding their
Fig. 6. Blocking efficacy of sPi4 towards the voltage-gated K
+
channel subtypes. (A) Oocytes expressing Shaker BK
+
channels were recorded under
two-electrode voltage clamp. K
+
currents were obtained by depolarization from a holding potential of )90 mV to +70 mV. Left panel: Shaker B
K

+
control currents during superfusion of 100 n
M
of sPi4 illustrating over 80% block; right panel: K
+
currents during superfusion of 100 n
M
sPi4,
illustrating over 60% block. The dose–response curves for sPi4 current inhibition were performed for: (B) Shaker B, (C) rat Kv1.2, and (D) rat
Kv1.1 (s)andKv1.3(d) channels. The solid lines through the data are obtained from the equation y ¼ y
o
+ a/[1+exp(–(x ) IC
50
)/b)]. The
IC
50
values of sPi4 were 3.0 ± 2.2 n
M
for Shaker B(n ¼ 45) and 8 ± 5 p
M
for rat Kv1.2 (n ¼ 55) channels. No significant effects on rat Kv1.1
(n ¼ 10) and Kv1.3 (n ¼ 10) channels were detected at sPi4 concentrations up to 10 l
M
. Data points are the mean ± SEM. When absent, error
bars are within symbol size. All inhibitions were determined by inducing currents by depolarizations at +70 mV.
3588 S. M’Barek et al. (Eur. J. Biochem. 270) Ó FEBS 2003
recognition and blockage of rat Kv1.2 channels Fig. 8A,B).
These maps suggest an important contribution of Arg10,
Arg19, Lys26, Ile28, Lys30, Lys33 and Tyr35 residues for
Pi4, as well as of Arg5, Arg12, Lys24, Ile26, Arg28, Lys31

and Tyr33 residues for Pi1. The functional dyads are
attributed to Lys26 and Tyr35 for Pi4 [1], and Lys24 and
Tyr33 for Pi1 [7]. Therefore, as mentioned by Olamendi-
Portugal et al. [1], the substitution of the functional Lys26
in Pi4 for an Arg26 in the structurally homologous Pi7
might be a key natural Ôpoint mutationÕ responsible for the
lack of activity of Pi7 on Kv channels. We suggest a two-
step pictorial view of Pi4 binding in which the toxin ring of
basic residues (ring composed of Arg10, Arg19, Lys30 and
Lys33) plays a crucial role (via electrostatic forces) in the
recognition, interaction and correct positioning of Pi4 on
the Kv1.2 channels, and then a tighter interaction takes
place through both hydrophobic forces and hydrogen
bonding between Tyr35 (dyad) and the aromatic cluster
consisting of Trp344, Trp345* and Tyr355* (Fig. 7B,
legend), and between Ile28 and Val361. The Lys26 (dyad)
sidechain enters the ion channel pore and is stabilized by
the four Asp357 carbonyl oxygen atoms of the Kv1.2
a-subunits; the Lys sidechain presumably acts by blocking
K
+
ion flux through the pore, and might thus be involved
in the toxin blocking efficacy.
Presence of the ring of basic residues in other toxins
active on Kv1.2 channels
To examine the potential importance of the ring of basic
residues in the recognition and interaction of Pi4 with Kv1.2
channels, we focused on two scorpion toxins, Pi2 (P. im-
perator) [6,46] and TsTXa (Tityus serrulatus) [47], that are
also classified as high affinity blockers of Kv1.2 channels

(both being active at the picomolar concentration range)
and of known 3D structures [48,49]. The two toxins possess
well-defined b-sheet-associated functional dyads, i.e. Lys27
and Tyr36 (TsTXa) and Lys24 and Phe33 (Pi2). Of note, the
usual aromatic Tyr is replaced by an aromatic Phe in the
case of Pi2, which is thought to interact, via its phenyl ring,
with the aromatic cluster of the Kv1.2 a-subunit as well. Pi2
and TsTXa also exhibit a four-membered ring of basic
residues similar to that of Pi4 or Pi1. It is clear that more
structure–activity relationship studies on these toxins are
needed to validate the idea of a possible key role of such a
ring in the toxin binding on Kv1.2 channels.
Conclusions
From a number of previous reports on different scorpion
toxins that act on Kv-type channels, it appears that the toxin
Fig. 7. Docking of sPi4 on rat voltage-gated Kv1.2 channel (pore region). (A) Side view (
TURBO
-
FRODO
software) depicting the interaction of sPi4 [1]
(structure solved by
1
H-NMR) with rat voltage-gated Kv1.2 channel (molecular model of the S5–H5–S6 domains) [24]. For clarity, Ca peptide
backbones of only two out of the four S5–H5–S6 a-subunits of the Kv1.2 channel are presented (deep blue). The Ca peptide backbone of sPi4 is
shown in green. Only the sidechains of amino acid residues that are involved in the sPi4–Kv1.2 channel interaction are displayed. Basic, acidic and
aromatic residues are shown in light blue, red and purple, respectively. The residues are numbered according to their positions within the Pi4 and rat
Kv1.2 a-subunit amino acid sequences [1,24]. (B) Magnified side view showing the interactions of sPi4 with the rat Kv1.2 channel. For sPi4, only the
sidechains of residues involved in this interaction are depicted. Also, only interacting residues from the Kv1.2 a-subunits are pictured in their exact
3D positions, according to the ion channel molecular model (see Fig. 7A for details). The asterisks indicate that the corresponding residues belong
to distinct a-subunits. (C) Top view showing the docking of sPi4 on rat voltage-gated Kv1.2 channel (pore region). Only interacting residues are

presented with their corresponding sidechains (see Fig. 7A for details). The four a-subunits (S5–H5–S6 domains) forming the Kv1.2 channel are
noted from A to D.
Ó FEBS 2003 Synthesis and pharmacology of Pi4 (Eur. J. Biochem. 270) 3589
b-sheet structure plays a premium role in binding to these
channels [32,41,50,51]. Amongst the residues belonging to
the b-sheet, the key contribution of a pair of well-defined
basic and aromatic residues, referred to as the functional
dyad, which we attributed to Lys26 and Tyr35 in the case of
Pi4, has been shown. The docking of Pi4 (or Pi1) on rat
Kv1.2 channels further provides additional insights into the
structural basis of this recognition/interaction. Indeed, an
unexpected contribution of a ring composed of four basic
residues belonging to various faces of the toxin has been
highlighted, which supports the idea of a multipoint
interaction between Pi4 and this ion channel. It is interesting
to note that this ring of basic residues also exists in other
potent Kv1.2 channel-acting scorpion toxins, such as Pi1 [7],
Pi2 [6,46], and TsTXa [47,49]. At the level of rat Kv1.2
channel, a key functional residue appears to be Glu332 of the
a-subunit, a residue absent in Kv1.1 and Kv1.3 a-subunits
[24]. In the context of the channel, the four Glu332 from the
four a-subunits are thought to interact, via salt bridges, with
the four residues from the toxin ring of basic residues. The
production of some selected Pi4 analogs, notably those with
an altered ring of basic residues, will help to test experi-
mentally the Pi4 functional map deduced from the docking
experiments. Because the latter also gave some insights that
might potentially explain the selectivity of the Pi4 action on
voltage-gated Kv1.2 channels, the docking approach will be
used to design Pi4 analogs that exhibit some changes in

pharmacological selectivity or affinity towards the K
+
channel subtypes. The completeness of Kv1.2 pore occlusion
by Pi4 is a parameter that can tentatively be improved by
selective mutation of some Pi4 residues. It is worth noting
that the actual docking simulation of Pi4 is informative but
remains insufficient to reasonably explain, at a molecular
level, the partial Kv1.2 pore occlusion. The present study will
be further extended to other Kv-channel acting scorpion
toxins [6,9,47,52] to shed light on the molecular basis of
the toxin to Kv-type channel recognition. Finally, it will be
interesting to test the sPi4 bioactivity on different classes of
K
+
channels (e.g. Eag, HERG, KCNQ, Slo, IKCa, Kir, and
KCNK) in order to determine the uniqueness or not of toxin
action on Kv1.2 channel.
Acknowledgements
The authors wish to thank Drs J. Van Rietschoten and C. Devaux for
helpful discussions. This work was supported by funds from the CNRS
and Cellpep SA (Paris, France). Dr A. Mosbah is a recipient of a
fellowship from Cellpep SA.
References
1. Olamendi-Portugal, T., Gomez-Lagunas, F., Gurrola, G.B. &
Possani, L.D. (1998) Two similar peptides from the venom of the
scorpion Pandinus imperator, one hightly effective blocker and the
other inactive on K
+
channels. Toxicon 36, 759–770.
2. Tytgat, J., Chandy, G., Garcia, M.L., Gutman, G.A., Martin

Eauclaire, M.F., van der Walt, J.J. & Possani, L.D. (1999) A
unified nomenclature for short-chain peptides isolated from
scorpion venoms: a-KTx. Trends Physiol. Sci. 20, 444–447.
3. Kharrat, R., Mabrouk, K., Crest, M., Darbon, H., Oughideni, R.,
Martin-Eauclaire,M.F.,Jacquet,G.,ElAyeb,M.,Van
Rietschoten, J., Rochat, H. & Sabatier, J.M. (1996) Chemical
synthesis and characterization of maurotoxin, a short scorpion
toxin with four disulfide bridges that acts on K
+
channels. Eur. J.
Biochem. 242, 491–498.
4. Kharrat, R., Mansuelle, P., Sampieri, F., Crest, M., Oughideni,
R., Van Rietschoten, J., Martin-Eauclaire, M F., Rochat, H. &
Fig. 8. Functional map of Pi4 towards rat Kv1.2 channel, and comparison with that of Pi1. Functional maps of Pi4 [1] and Pi1 [5–7] (A and B,
respectively). The structures of scorpion toxins are shown according to a space-filling representation (
TURBO
-
FRODO
software). The highlighted
amino acid residues are those belonging to the toxin functional maps with regard to rat Kv1.2 channel. These residues are numbered according to
the toxin amino acid sequences. The basic and hydrophobic residues of the toxin functional maps are pictured in light blue and green, respectively.
Residues from the functional dyad are shown in purple. Other residues are shown in pink.
3590 S. M’Barek et al. (Eur. J. Biochem. 270) Ó FEBS 2003
El Ayeb, M. (1997) Maurotoxin, a four disulfide bridge toxin from
Scorpio maurus venom: purification, structure and action on
potassium channels. FEBS Lett. 406, 284–290.
5. Olamendi-Portugal, T., Gomez-Lagunas, F., Gurrola, G.B. &
Possani, L.D. (1996) A novel structural class of K
+
-channel

blocking toxin from the scorpion Pandinus imperator. Biochem. J.
315, 977–981.
6. Rogowski, R.S., Collins, J.H., O’Neill, T.J., Gustafson, T.A.,
Werkman, T.R., Rogawski, M.A., Tenenholz, T.C., Weber, D.J.
& Blaustein, M.P. (1996) Three new toxins from the scorpion
Pandinus imperator selectively block certain voltage-gated K
+
channels. Mol. Pharmacol. 50, 1167–1177.
7. Fajloun, Z., Carlier, E., Lecomte, C., Geib, S., di Luccio, E.,
Bichet, D., Mabrouk, K., Rochat, H., De Waard, M. & Sabatier,
J.M. (2000) Chemical synthesis and characterization of Pi1, a
scorpion toxin from Pandinus imperator active on K
+
channels.
Eur. J. Biochem. 267, 5149–5155.
8. Lebrun, B., Romi-Lebrun, R., Martin-Eauclaire, M F., Yasuda,
A., Ishiguro, M., Oyama, Y., Pongs, O. & Nakajima, T. (1997) A
four-disulphide-bridged toxin, with high affinity towards voltage-
gated K
+
channels, isolated from Heterometrus spinnifer (Scor-
pionidae) venom. Biochem. J. 328, 321–327.
9. Savarin, P., Romi-Lebrun, R., Zinn-Justin, S., Lebrun, B.,
Nakajima, T., Gilquin, B. & Menez, A. (1999) Structural and
functional consequences of the presence of a fourth disulfide
bridge in the scorpion short toxins: solution structure of the
potassium channel inhibitor HsTX1. Protein Sci. 8, 2672–2685.
10. Bontems, F., Roumestand, C., Gilquin, B., Me
´
nez, A. & Toma, F.

(1991) Refined structure of charybdotoxin: common motifs in
scorpion toxins and insect defensins. Science 254, 1521–1523.
11. Blanc, E., Sabatier, J M., Kharrat, R., Meunier, S., El Ayeb, M.,
Van Rietschoten, J. & Darbon, H. (1997) Solution structure of
maurotoxin, a scorpion toxin from Scorpio maurus, with high
affinity for voltage-gated potassium channels. Proteins 29,321–
333.
12. Delepierre, M., Prochnicka-Chalufour, A. & Possani, L.D. (1997)
A novel potassium channel blocking toxin from the scorpion
Pandinus imperator:a
1
H NMR analysis using a nano-NMR
probe. Biochemistry 36, 2649–2658.
13. Delepierre, M., Prochnicka-Chalufour, A. & Possani, L.D. (1998)
1
H NMR structural analysis of novel potassium blocking toxins
using a nano-NMR probe. Toxicon 36, 1599–1608.
14. Fajloun,Z.,Kharrat,R.,Chen,C.,Lecomte,C.,diLuccio,E.,
Bichet,D.,ElAyeb,M.,Rochat,H.,Allen,P.D.,Pessah,I.N.,De
Waard, M. & Sabatier, J.M. (2000) Chemical synthesis and char-
acterization of maurocalcine, a scorpion toxin that activates Ca
2+
release channel/ryanodine receptors. FEBS Lett. 469, 179–185.
15. Mosbah, A., Kharrat, R., Fajloun, Z., Renisio, J.C., Blanc, E.,
Sabatier, J.M., El Ayeb, M. & Darbon, H. (2000) A new fold
in the scorpion toxin family, associated with an activity on a
ryanodine-sensitive calcium channel. Proteins 40, 436–442.
16. Guijarro, J.I., M’Barek, S., Gomez-Lagunas, F., Garnier, D.,
Rochat,H.,Sabatier,J.M.,Possani,L.D.&Delepierre,M.(2003)
Solution structure of Pi4, a short four-disulfide-bridged scorpion

toxin specific of potassium channels. Prot. Sci. in press.
17. Merrifield, R.B. (1986) Solid phase synthesis. Science 232,341–
347.
18. Sabatier, J.M. (1999) Handbook of Toxinology, Animal Toxins:
Tools in Cell Biology (Rochat, H. & Martin-Eauclaire, M.F., eds),
pp. 198–218. Birkha
¨
user-Verlag, Basel, Switzerland.
19. Gray, E.G. & Whittaker, V.P. (1962) The isolation of nerve end-
ings from brain: an electron microscopic study of cell fragments
derived by homogenization centrifugation. J. Anat. 96, 79–88.
20. Seagar, M., Granier, C. & Couraud, F. (1984) Interactions of the
neurotoxin apamin with a Ca
2+
-activated K
+
channel in primary
neuronal cultures. J. Biol. Chem. 259, 1491–1495.
21. De Waard, M. & Campbell, K.P. (1995) Subunit regulation of the
neuronal alpha 1A Ca
2+
channel expressed in Xenopus oocytes.
J. Physiol. 485, 619–634.
22. Eppig, J.J. & Dumont, J.N. (1976) Defined nutrient medium for
the in vitro maintenance of Xenopus laevis oocytes. In vitro 12,
418–427.
23. Morais Cabral, J.H., Lee, A., Cohen, S.L., Chait, B.T., Li, M. &
Mackinnon, R. (1998) The structure of the potassium channel:
molecular basis of K
+

conduction and selectivity. Science 280,
69–77.
24. Christie, M.J. (1995) Molecular and functional diversity of K
+
channels. Clin. Exp. Pharmacol. Physiol. 22, 944–951.
25. ClustalW WWW Service at the European Bioinformatics Insti-
tute, />26. Roussel, A. & Cambillau, C. (1989) Silicon Graphics Geometry
Partner Directory, pp. 77–78. Silicon Graphics, Mountain View,
CA.
27. Bru
¨
nger, A.T., Adams, P.D., Clore, G.M., DeLano, W.L., Gros,
P., Grosse-Kunstleve, R.W., Jiang, J.S., Kuszewski, J., Nilges, M.,
Pannu,N.S.,Read,R.J.,Rice,L.M.,Simonson,T.&Warren,
G.L. (1998) Crystallography and NMR system (CNS): a new
software system for macromolecular structure determination. Acta
Crystallogr. D54, 905–921.
28.Hooft,R.W.W.,Vriend,G.,Sander,C.&Abola,E.E.(1996)
Errors in protein structures. Nature 381,272.
29. Laskowski, R.A., MacArthur, M.W., Moss, D.S. & Thornton,
J.M. (1993) PROCHECK: a program to check the stereochemical
quality of protein structures. J. Appl. Cryst. 26, 283–291.
30. Morris,A.L.,MacArthur,M.W.,Hutchinson,E.G.&Thornton,
J.M. (1992) Stereochemical quality of protein structure
coordinates. Proteins 12, 345–364.
31. Palma, P.N., Krippahl, L., Wampler, J.E. & Moura, J.J.G. (2000)
BiGGER: a new (soft) docking algorithm for predicting protein
interactions. Proteins Struct. Funct. Gen. 39, 372–384.
32. Dauplais, M., Lecoq, A., Song, J., Cotton, J., Jamin, N., Gilquin,
B., Roumestand, C., Vita, C., de Meideros, C.L., Rowan, E.G.,

Harvey,A.L.&Me
´
nez, A. (1997) On the convergent evolution of
animal toxins: conservation of a dyad of functional residues in
potassium channel-blocking toxins with unrelated structures.
J. Biol. Chem. 272, 4302–4309.
33. Gilquin, B., Racape, J., Wrisch, A., Visan, V., Lecoq, A.,
Grissmer, S., Me
´
nez, A. & Gasparini, S. (2002) Structure of the
BgK-Kv1.1 complex based on distance restraints identified by
double mutant cycles. Molecular basis for convergent evolution of
Kv1 channel blockers. J. Biol. Chem. 277, 37406–37413.
34. Alessandri-Haber, N., Lecoq, A., Gasparini, S., Grangier-Mac-
math, G., Jacquet, G., Harvey, A.L., de Medeiros, C., Rowan,
E.G.,Gola,M.,Me
´
nez, A. & Crest, M. (1999) Mapping the
functional anatomy of BgK on Kv1.1, Kv1.2, and Kv1.3: clues to
design analogs with enhanced selectivity. J. Biol. Chem. 274,
35653–35661.
35. Gasparini, S., Danse, J.M., Lecoq, A., Pinkasfeld, S., Zinn-Justin,
S., Young, L.C., de Medeiros, C.C., Rowan, E.G., Harvey, A.L. &
Me
´
nez, A. (1998) Delineation of the functional site of a-dendro-
toxin: the functional topographies of dendrotoxins are different
but share a conserved core with those of other Kv1 potassium
channel-blocking toxins. J. Biol. Chem. 273, 25393–25403.
36. Harvey, A.L., Vatanpour, H., Rowan, E.G., Pinkasfeld, S., Vita,

C., Me
´
nez, A. & Martin-Eauclaire, M.F. (1995) Structure-activity
studies on scorpion toxins that block potassium channels. Toxicon
33, 425–436.
37. Fu,W.,Cui,M.,Briggs,J.M.,Huang,X.,Xiong,B.,Zhang,Y.,
Luo,X.,Shen,J.,Ji,R.,Jiang,H.&Chen,K.(2002)Brownian
dynamics simulations of the recognition of the scorpion toxin
maurotoxin with the voltage-gated potassium ion channels. Bio-
phys. J. 83, 2370–2385.
Ó FEBS 2003 Synthesis and pharmacology of Pi4 (Eur. J. Biochem. 270) 3591
38. Cui,M.,Shen,J.,Briggs,J.M.,Fu,W.,Wu,J.,Zhang,Y.,Luo,
X., Chi, Z., Ji, R., Jiang, H. & Chen, K. (2002) Brownian
dynamics simulations of the recognition of the scorpion toxin P05
with the small-conductance calcium-activated potassium channels.
J. Mol. Biol. 318, 417–428.
39. Koch, R.O., Wanner, S.G., Koschak, A., Hanner, M., Schwarzer,
C., Kaczorowski, G.J., Slaughter, R.S., Garcia, M.L. & Knaus,
H.G. (1997) Complex subunit assembly of neuronal voltage-gated
K
+
channels: basis for high-affinity toxin interactions and phar-
macology. J. Biol. Chem. 272, 27577–27581.
40. Miller, C. (1995) The charybdotoxin family of K+ channel-
blocking peptides. Neuron 15, 5–10.
41. Darbon, H., Blanc, E. & Sabatier, J.M. (1999) Perspectives in Drug
Discovery and Design: Animal Toxins and Potassium Channels
(Darbon, H. & Sabatier, J.M., eds), Vol. 15/16, pp. 40–60.
Kluwer/Escom, Kluwer Academic Publishers, Dordrecht, the
Netherlands.

42. Fajloun, Z., Mosbah, A., Carlier, E., Mansuelle, P., Sandoz, G.,
Fathallah, M., di Luccio, E., Devaux, C., Rochat, H., Darbon, H.,
De Waard, M. & Sabatier, J.M. (2000) Maurotoxin versus Pi1/
HsTx1 toxins: toward new insights in the understanding of their
distinct disulfide bridge patterns. J. Biol. Chem. 275, 39394–39402.
43. Fajloun, Z., Ferrat, G., Carlier, E., Fathallah, M., Lecomte, C.,
Sandoz, G., di Luccio, E., Mabrouk, K., Legros, C., Darbon, H.,
Rochat, H., Sabatier, J M. & De Waard, M. (2000) Synthesis, 1H
NMR structure, and activity of a three-disulfide-bridged mauro-
toxin analog designed to restore the consensus motif of scorpion
toxins. J. Biol. Chem. 275, 13605–13612.
44. Gomez-Lagunas, F., Olamendi-Portugal, T. & Possani, L.D.
(1997) Block of Shaker BK
+
channels by Pi1, a novel class of
scorpion toxin. FEBS Lett. 400, 197–200.
45. Carlier, E., Avdonin, V., Geib, S., Fajloun, Z., Kharrat, R.,
Rochat,H.,Sabatier,J M.,Hoshi,T.&DeWaard,M.(2000)
Effect of maurotoxin, a four disulfide-bridged toxin from the
Chactidae scorpion Scorpio maurus,onShaker K
+
channels.
J. Peptide Res. 55, 417–429.
46. Gomez-Lagunas, F., Olamendi-Portugal, T., Zamudio, F.Z. &
Possani, L.D. (1996) Two novel toxins from the venom of the
scorpion Pandinus imperator show that the N-terminal amino acid
sequence is important for their affinities towards Shaker BK
+
channels. J. Membr. Biol. 152, 49–56.
47. Rogowski, R.S., Krueger, B.K., Collins, J.H. & Blaustein, M.P.

(1994) Tityustoxin Ka blocks voltage-gated noninactivating K
+
channels and unblocks inactivating K
+
channels blocked by
a-dendrotoxin in synaptosomes. Proc. Natl Acad. Sci. USA 91,
1475–1479.
48. Tenenholz, T.C., Rogowski, R.S., Collins, J.H., Blaustein, M.P. &
Weber, D.J. (1997) Solution structure for Pandinus toxin K-a
(PiTX-K alpha), a selective blocker of A-type potassium channels.
Biochemistry 36, 2763–2771.
49. Ellis, K.C., Tenenholz, T.C., Jerng, H., Hayhurst, M., Dudlak,
C.S., Gilly, W.F., Blaustein, M.P. & Weber, D.J. (2001) Inter-
action of a toxin from the scorpion Tityus serrulatus with a cloned
K
+
channel from squid (sqKv1A). Biochemistry 40, 5942–5953.
50. Possani, L.D., Selisko, B. & Gurrola, G.B. (1999) Perspectives in
Drug Discovery and Design: Animal Toxins and Potassium Chan-
nels (Darbon, H. & Sabatier, J.M., eds), Vol. 15/16, pp. 15–40.
Kluwer/Escom, Kluwer Academic Publishers, Dordrecht, the
Netherlands.
51. Srinivasan, K.N., Sivaraja, V., Huys, I., Sasaki, T., Cheng, B.,
Kumar, T.K., Sato, K. & Tytgat, J., YuC., San, B.C., Rangana-
than, S., Bowie, H.J., Kini, R.M. & Gopalakrishnakone, P. (2002)
Hefutoxin1, a novel toxin from the scorpion Heterometrus fulvipes
with unique structure and function: importance of the functional
dyad in potassium channel selectivity. J. Biol. Chem. 277, 30040–
30047.
52. Dudina, E.E., Korolkova, Y.V., Bocharova, N.E., Koshelev,

S.G., Egorov, T.A., Huys, I., Tytgat, J. & Grishin, E.V. (2001)
OsK2, a new selective inhibitor of Kv1.2 potassium channels
purified from the venom of the scorpion Orthochirus scrobiculosus.
Biochem. Biophys. Res. Commun. 286, 841–847.
3592 S. M’Barek et al. (Eur. J. Biochem. 270) Ó FEBS 2003