Molecular Immunology, Vol. 30, No. 12, pp. 1061-1068, 1993 0161-5890/93 $6.00 + 0.

00
Printed in Great Britain. 0 1993PergamonPress Ltd

FINE MOLECULAR ANALYSIS OF THE ANTIGENICITY OF THE

ANDROCTONUS AUSTRALIS HECTOR SCORPION NEUROTOXIN II: A
NEW ANTIGENIC EPITOPE DISCLOSED BY THE PEPSCAN METHOD

CHRISTIANEDEVAUX,*MARIANICKJUIN, PASCALMANSUELLE and CLAUDEGRANIER

Laboratoire de Biochimie, CNRS URA 1455, Facultt de MCdecine Secteur Nord,
Boulevard P. Dramard, 13916 Marseille Cedex 20, France

(First received 23 January 1993; accepted in revised form 21 March 1993)

Abstract-A set of 58 overlapping rod-bound peptides was used to map the antigenic reactivity
pattern of a 64-residue neurotoxin (AaH II) from the venom of the scorpion Androctonus austrafis
hector. Five anti-toxin rabbit antisera were assayed serially for their capacity to bind to each peptide
in the set. Six regions of antigenic reactivity were thus identified (sequences: l-8,4-12, 27-35, 3945,
52-58 and 5541). When positioned on a 3-D model of the toxin, these regions appeared to
correspond to either p-turn or extended parts of the molecule. The antigenic regions revealed by this
technique agree fairly well with those previously mapped on the same toxin by different methods.
One discrepancy was, however, that the present study shows the N-terminus to be strongly reactive
with anti-toxin antibodies. The antigenicity of this region was confirmed, since rabbit antibodies
raised against a synthetic peptide mimicking the sequence 1-8 of the toxin were found to bind the
toxin with high efficiency. A fine analysis of the recognition of this region was performed.
Alanine-containing analogs of the sequence l-7 and peptides mimicking the N-terminal of the four
main toxins of AaH were probed with anti-toxin and anti-peptide antibodies. Lysine 2, aspartic acid
3 and glycine 4 were shown to be key residues in the recognition of the N-terminal region of the
AaH II toxin by anti-toxin antibodies. In contrast, a loose specificity of recognition was shown by
one anti-peptide serum which was, in addition, able to recognize the N-termini of all four AaH
toxins.

INTRODUCTION AaH IV belong to the same serotype (El Ayeb et al.,

The noxious and, sometimes, fatal effect of scorpion
stings are due to the neurotoxic properties of small
proteins present at low concn in the venom. These
molecules constitute a family of structurally related
proteins [60-70 amino acids (Rochat et al., 1979)] exert-
ing neurotoxic effects by interacting specifically with the
voltage-dependent sodium channels of excitable cells
(Catterall, 1977). There is thus a great medical and
scientific interest in developing strategies to neutralize
efficiently these toxins, after they have entered the body.
Understanding the antigenic structure of scorpion toxins
could lead to improvement of antibody therapy of
envenomation.
The Androctonus australis hector scorpion (AaH) is
one of the most venomous species in North Africa,
causing hundreds of deaths a year (Goyffon, 1982; Jeddi
et al., 1988). Four toxins (AaH I-IV) active in mammals
have been characterized. One of them, AaH II, is
particularly potent (Martin and Rochat, 1986). These
four molecules can be divided into two different groups
on the basis of antigenic reactivity. AaH I, AaH III and
*Author to whom correspondence should be addressed.
Abbreviations: AaH, Androctonus australis hector; ELISA,
enzyme linked immunosorbent assay; PBS, phosphate
buffered saline; TMB, 3,3’,5,5’-tetramethylbenzidine.
1061
1983; Mansuelle et al., 1992), AaH II to a different one.
Five antigenic regions of the AaH II molecule have been
previously mapped using synthetic peptides (review:
Granier et al., 1989). Efforts have been made to mimic
some of the conformational features of the toxin with
synthetic peptides: three of the five antigenic regions
appeared to be dependent on the local conformation of
AaH II (El Ayeb et al., 1984; Bahraoui et al., 1987;
Fourquet et al., 1988).
The exact delineation of a conformational epitope can
only be deduced from the X-ray crystallographic struc-
ture of the antigen-antibody complex (Amit et al., 1986;
Davies et al., 1990). Moreover, conformational epitopes
are expected to be difficult to mimic with a synthetic
peptide. Epitope mapping using short synthetic peptides
to identify linear or continuous epitopes is well devel-
oped, but also controversial (Jemmerson, 1987; Laver
et al., 1990). Nevertheless, this approach is expected to
be extremely useful in the development of peptide-based
vaccines and/or peptide-based immunoassays (Lerner,
1982; Arnon and Horwitz, 1992). In the present studies,
we used a series of short overlapping heptapeptides
corresponding to the entire sequence of AaH II to
systematically search for linear epitopes of AaH II using
the approach described by Geysen et al. (1984, 1987).
The purpose was to assess the capacity of the technique
to identify known epitopes of the toxin and to indicate
new antigenic regions.
1062 C. DEVAUX et al.

MATERIALS AND METHODS proceed overnight. The reaction mixture was then
aliquoted and frozen. Four rabbits (80, 81, 82 and 83)
Peptide synthesis received 100 pg of peptide at 3-week intervals first by
Synthesis in the Pepscan format. The 58 overlapping intradermal, then by subcutaneous injections. Rabbits
heptapeptides covering the entire 64-residue sequence of 80 and 81 received the immunogen in Freund’s adjuvant
AaH II (Rochat et al., 1972) were simultaneously syn- (Difco, Detroit MI, U.S.A.), rabbits 82 and 83 in
thesized by the Pepscan method (Geysen et al., 1984, TiterMax (Cyt Rx, Atlanta GA, U.S.A). Bleedings were
1987) using the commercial Epitope Scanning Kit (CRB, performed 10 days after each injection. The resulting
Cambridge, U.K). Two additional sets of peptides were sera were prepared, aliquoted and stored at -20°C.
also prepared in the same experimental conditions: one
corresponded to amino acid sequences l-7 of toxins
AaH I, AaH II, AaH III and AaH IV, the other to Immunoassays
analogs of the sequence l-7 of the toxin AaH II in which Pepscan ELISA. ELISA reactions on rod-bound pep-
each residue was in turn replaced by an alanine. A tides were performed as previously described (Yahi et al.,
control peptide (sequence 47-53 of toxin AaH II: 1992) with minor improvements. Briefly, the assembled
YCYKLPD) was prepared which was used to assess rods were saturated for 1 hr at room temp with 1%
non-specific fixation. At the end of the synthesis the casein in PBS containing 0.01% Tween (PBS-T), washed
Ncc-terminal residue of every peptide was acetylated, with 0.9% NaCl, 0.01% Tween (washing buffer), incu-
except peptides corresponding to the N-terminal part of bated overnight at 4°C with an appropriate dilution of
the toxin. All synthetic peptides were prepared in dupli- rabbit serum (1:400 to 1: 15000) and washed again.
cate on polyethylene rods arranged in the format of a Bound antibodies were detected by incubating blocks of
96-well ELISA plate. All reagents and solvents were of rods for 1 hr either with: (i) goat anti-rabbit IgG coupled
the best available quality and used according to the to alkaline phosphatase. After washing, p-nitrophenyl
manufacturers’ instructions. phosphate was used as substrate and the absorbance at
Preparation of the peptide (l-8)-SH. The peptide was 405 nm was measured after 60-90 mn on a Multiscan
prepared by Fmoc chemistry on a polyacrylamide resin Titertek plate reader, or (ii) peroxidase-conjugated goat
(Expansinfm, Expansia, Aramon, France) that had been anti-rabbit IgG (H + L) (Diagnostics Pasteur, France).
modified to allow cleavage of the final, deprotected After washing the TMB (3,3’,5,5’-tetramethylbenzidine)
peptide by a mercaptan (Mery et al., 1993). The peptide- microwell peroxidase substrate system (Kirkegaard and
resin (ca 50 mg) was dissolved at 20 mg/ml in degassed Perry, Gaithersburg, MD) was used. Ten minutes later
0.1 M borate buffer pH 9 containing a lo-fold molar the absorbance at 620 nm was measured, the reaction
excess (over the peptide) of dithioerithritol (Pierce stopped immediately by 1 M phosphoric acid and the
Chemical Company). The screw-capped tube containing absorbance at 450 nm measured. In the assays using
the reaction mixture was allowed to react overnight, with the set of 58 heptapeptides the experimental results
gentle rotation. The free peptide, containing a C-termi- were expressed as described by Trifilieff et al. (1991): the
nal cysteamine residue, was then separated from the mean absorbance value of the 29 pins with the lowest
resin by filtration on a Spin-X’” system (Costar, Cam- absorbance was calculated and taken as the noise (N) of
bridge MA, U.S.A.) and further purified by chromatog- the experiment; then the S/N ratio (the absorbance value
raphy on a 50 x 1.2 cm Sephadex G15 column in 0.1 N S of each peptide divided by the noise N) was calcu-
acetic acid. The amino acid composition and HPLC lated. The results were plotted against the peptide num-
profile of the peptide were consistent with a purity ber (the peptide number was defined as the position of
greater than 95% (Mery et al., 1993). its N-terminal residue in the sequence of AaH II). In the
assays testing the analogs of peptide (l-7) the back-
ground was the absorbance value of the control peptide
Antisera and immunization (sequence 47-53). The results were expressed relatively
Five rabbit anti-AaH II sera described by Delori et al. to the reactivity of the parent peptide (l-7) of AaH II.
(1981) were used in this work under their original Between each assay, bound antibodies were removed
designations: S114, S182, S184, S185 and S187. The from the rods by sonication for 30 min at 60°C in a
anti-(l-8) peptide sera were obtained by injection into 0.1 M phosphate buffer pH 7.2 containing 1% sodium
rabbits of the peptide linked to Keyhole limpet hemo- dodecyl sulfate and 0.1% 2-b mercaptoethanol. Rods
cyanin (KLH). Briefly, 100 p 1 of 10 mg/ml of m-maleimi- were then extensively washed in water at 60°C.
dobenzoyl-N-hydroxysuccinimide ester (MBS, Pierce) in Direct ELBA. The sera from various bleedings of
acetonitrile were added to 5 mg of KLH (Calbiochem) in rabbits 80,8 1,82 and 83 were checked for their reactivity
565 pl of 10mM potassium phosphate buffer pH 7. with both the peptide l-8 and the toxin AaH II in an
After 30 mn (at room temp) with agitation, the mixture ELISA assay. Nunc Maxisorb plates were coated
was desalted on a Pharmacia PD 10 column in 50 mM overnight with 100 ~1 of a 4 pg/ml solution (in carbonate
potassium phosphate buffer, pH 6. The void-volume buffer pH 9.6) of either the peptide or the toxin.
fractions were collected and added to the lyophilized Preimmune and immune sera of the four rabbits were
peptide (5 pmoles). The pH was adjusted to 7.5 with tested at a 1:2500 dilution (2 hr, 37’C). The second
0.1 M NaOH and the coupling reaction was allowed to antibody used was the peroxidase-conjugated anti-rabbit
 Peptide epitopes of a scorpion toxin 1063

IgG with TMB as substrate, as described above. The (residues l-8: VKDGYVD) were reactive with three
same general protocol was used for titration exper- different sera (S114, S182 and S184) and gave the highest
iments except that the pre-immune and immune sera ELISA signals. Peptides 27, 28 and 29 (residues 27-35:
were serially diluted. TKLKGESGY) were recognized by two sera (Sl14 and
S187). Peptide 55 (residues 55561: VRTKGPG) was
reactive with both serum S185 and S187. In addition,
RESULTS three more peptides or groups of adjacent peptides were
specifically recognized by one serum each: peptides 46
Mapping of the linear epitopes of AaH II dejined by jive
(residues 4-l 2: GYIVDDVN) reacted with S184, peptide
rabbit anti-AaH II sera
39 (residues 3945: ASPYGNA) with S182 and peptide
Hyperimmune sera from five rabbits immunized with 52 (residues 52-58: PDHVRTKG) with Sl14. The
native AaH II were tested for their reactivity to each of l-8/412 and 52-58155-61 antigenic regions overlapped.
the 58 overlapping heptapeptides covering the AaH II The continuous antigenic regions thus identified were
sequence (Fig. 1). Each serum showed a characteristic positioned (Fig. 2B) on the 3D-structure of AaH II
reactivity pattern, but only six regions of the AaH II (Fontecilla-Camps et al., 1988) to assess how they are
molecule were identified as antigenic in this assay: the six related to secondary structural features (Fig. 2A).
regions correspond to peptides 1-2, peptides 46, pep- Region l-8 includes a short b-sheet strand (located at
tides 27-29, peptide 39, peptide 52 and peptide 55. The residues 24) and the regions 4-12, 27-35, 3946 and
assay was very sensitive and reproducible. Signals were 52-58 each contain one of the 4 /?-turns of AaH II
observed using sera diluted 1: 1500 or 1: 3000 and the (located at residues 8-12, 27-30, 4043 and 52-55) with
strongest responses were detectable at dilution as low some extension to two other b-sheet strands (corre-
as 1: 10000 (peptides l-2 and 27-29). Figure 2A shows sponding to residues 32-37 and 45-51). The N-termi-
the location of the antibody binding regions along the nal part of the region 55561 appears to be in an
amino acid sequence of the toxin- Peptides 1 and 2 extended conformation in the 3D-structure of AaH II
(Fontecilla-Camps et al., 1988).
Although the results of the Pepscan method were in
good agreement with previous findings on the antigenic
S/N
structure of AaH II (see Discussion) there was one major
I 6 I, 16 21 26 31 36 41 46 51 56 exception. The N-terminal region (residues l-8) was
strongly recognized by three different anti-AaH II sera
(Fig. 1) but its antigenicity had not previously been
clearly demonstrated (Granier et al., 1989). To confirm
SIN
the antigenicity of the N-terminal region of AaH II it
0
I 6 ,I 16 21 26 31 36 41 46 51 56
was, therefore, necessary to verify whether anti-(l-8)
peptide antibodies recognize the cognate sequence in the
S184
native toxin.
4
SIN

0 Cross-reactivity of antibodies raised against synthetic

16 I, 16 21 26 31 36 41 46 51 56
peptide l-8 with AaH II

Sl85 Anti-(l-8) peptide antibodies were raised in rabbits

using the peptide linked to KLH by its C-terminal
.Sm
residue, as described in Materials and Methods. Four
rabbits were given the same immunogen either in Fre-
I 6 II 16 2, 26 31 36 41 46 51 56
und’s adjuvant (rabbits 80-81) or in TiterMax adjuvant
Sl87
(rabbits 82-83). All four rabbits responded to the pep-
4~ tide. However, rabbits 80 and 81 gave a consistently
SIN
higher anti-peptide antibody response and were there-
I 6 ,I 16 21 26 31 36 41 46 51 56
fore used for investigation of peptide l-8 antigenicity.
When the successive bleedings of rabbits 80 and 81 were
Peptide number
tested for their capacity to bind to plate-coated peptide
Fig. 1. Antibody binding patterns of overlapping heptapeptides corre- l-8 or to the parent AaH II, strong and constant
sponding to the 64-residue sequence of AaH II. Peptides were syn-
thesized on polyethylene rods. The immunoreactivity of rod-bound anti-peptide and anti-protein responses were observed
peptides with anti-AaH II rabbit antisera (Sl14, S182, S184, S185 and (Fig. 3A). Anti-AaH II antibodies were detectable in
S187) was tested in Pepscan ELISA. Sera were diluted 1: 1500 or the first bleedings and their level remained high during
1: 3000. Goat anti-rabbit IgG coupled to alkaline phosphatase (diluted
1: 1000) were used as second antibody. After absorbance reading, the the immunization protocol. The capacity of anti-peptide
signal over noise ratio (S/N) for each peptide was calculated, as antibodies to recognize and bind AaH II was further
described in Materials and Methods. Background values were found demonstrated by titration experiments (Fig. 3B).
to be 0.092,0.175,0.093,0.090,0.121, respectively for sera SI 14, S182,
S184, S185 and S187. A S/N superior to 2 was taken as a threshold Both anti-peptide and anti-toxin antibodies titers were
for significant binding. high: rabbit sera (S80 and S81) diluted 100,000
1064 C. DEVAUX et al.

Fig. 2. Analysis of the linear epitopes disclosed by the Pepscan method. (A) Immunoreactive regions revealed
by each serum are positioned on the sequence of AaH II. Only the peptides which reacted with serum with a
S/N>3(D)or >2( ) are indicated. The secondary features deduced from X-ray crystallographic data
(Fontecilla-Camps et al., 1988) are represented below: cc-helix (m), p-sheets (0) or p-turns (m).
(B) Immunoreactive regions (m) are superimposed on the three-dimensional structure of AaH II, as
determined by Fontecilla-Camps et al., (1988). All residues of the six immunoreactive regions (sequences:
l-8, 4-12, 27735, 39-45, 52-58 and 5561) are indicated (a). The symbol (a*) indicated the residues of the
l-8/4-12 and 52258155561 epitopes which overlap.

times still give a detectable ELISA signal (Fig. 3B). The residue of the epitope l-7 to antigenicity was further
specificity of the antibody binding to plate-coated pep- analyzed.
tide was shown by inhibition experiments using the free
peptide l-8 as competitor (&,, = 4 x lo-(‘M; data not
Role of individual amino acids in the antibody binding
shown). Thus, region l-8 of the AaH II toxin was found
region l-8
to be antigenic on the basis of the reactivity of anti-toxin
antibodies with peptides l-7 and 228 assembled on From the Pepscan data the two N-terminal hep-
polyethylene rods and of the capacity of anti-(l-8) tapeptides (VKDGYIV and KDGYIVD) gave a strong
peptide antibodies to recognize the toxin. Since this response with three different anti-AaH II sera. The
newly identified antigenic region corresponds to a part minimal epitope recognized was then K2D3G4Y516V7. To
of the sequence which is quite well conserved between determine which residues are essential to the antibody
different scorpion toxins, the contribution of each recognition of this region, a set of seven analogs were
 Peptide epitopes of a scorpion toxin
A Recognition analysis of dlgerent sequences l-7 from four
E 24
1 2 3 synthesized on rods and probed with anti-toxin and
Bleedings
I 2 3 4 5 6
Log [dilution]
Fig. 3. Reactivity of anti-(l-8) peptide antibodies with the peptide
and the toxin. (A). Sera from the three bleedings (day 10, 31 and 52)
of two rabbits were tested by ELISA for binding to peptide l-8
(B rabbit 80 and q rabbit 81) or AaH toxin II (0 rabbit 80 and
q rabbit 81). Sera were diluted 1:2500. (B) Titration of preimmune
(squares) and immune (triangles) sera from bleeding 2 of rabbit
80 using peptide l-8 (white symbols) or AaH II (black symbols) as
coated antigens (4pg/ml). Antibody binding was revealed with goat
anti-rabbit IgG peroxidase conjugate.
synthesized on rods: in these analogs each amino acid of
the sequence l-7 was replaced in turn by an alanine.
Rod-bound peptides were tested both with anti-toxin
sera (S114, S182 and S184, which reacted with peptides
l-7 and 2-8; see Fig. 1) and with anti-(l-8) peptide sera
(S80 and S81). Three amino acids at positions 2, 3 and
4 seemed to be critical to the antigen-antibody inter-
action (Fig. 4 A and B). Substitution of these three
residues gave the largest decreases (of >80%) in
immunoreactivity. In contrast residues 1, 5, 6 or 7
could be replaced by alanine without such large effects
on the binding of any serum. There were differences in
the fine specificity patterns of each serum due to the
individual polyclonal responses of each rabbit. This is
consistent with the experiment described above (Fig. 1).
The lysine (position 2) and the aspartic acid (position 3)
were apparently essential for the binding of all three
anti-toxin antibodies. In addition, the binding of S184
seemed greatly affected by the replacement of the glycine
in position 4, whereas the reactivity of S182 slightly
decreased when the isoleucine in position 6 was replaced
(Fig. 4A). The recognition of the peptide l-7 by anti-
(l-8) peptide serum S80 was also strongly affected when
K*, D3 or G4 were replaced by alanine; the recog-
nition pattern displayed by the other anti-(l-8) peptide
serum S81 seemed to be dependent only on the aspartic
acid side-chain (D3) (Fig. 4B), indicating some degener-
acy in the recognition specificity of anti-(l-8) peptide
antibodies.
1065
AaH toxins
The N-terminal sequences (residues l-7) of four tox-
ins from AaH venom are quite conserved (Rochat et al.,
1979; Mansuelle et al., 1992). They contain five con-
served amino acids (positions 3-7) and differ only by the
amino acids in position 1 and 2: KR in AaH I, VK in
AaH II, VR in AaH III and GR in AaH IV. Heptapep-
tides mimicking the four N-termini of these toxins were
anti-peptide antibodies (Fig. 5). The recognition of
peptides l-7 homologous of AaH I, AaH III and AaH
IV N-terminal sequences by the three anti-AaH II sera
(S114, S182 and S184) was 60-90% lower than that of
the parent AaH II peptide l-7 (Fig. 5A). In particular,
the lack of reactivity of anti-toxin antibodies with the
AaH III sequence 1-7 indicates that basic residue KZ
could not be interchanged with another basic amino acid
(R). This is consistent with the results of the alanine
replacement experiments (Fig. 4A), indicating that
residue 2 was not replaceable. When the set of differing
N-terminal peptides was probed with anti-peptide anti-
bodies (Fig. 5B) it was apparent that serum S81 recog-
nized equally well the four different peptides l-7
mimicking each of the N-terminal regions of the four
A
s-, loo
.=:
.z 80
P 60
p!
P 40
._
z 20
a
Y
0
None VI K2 D3 G4 Y5 16 V7
Modified residue
None VI K2 D3 G4 Y5 16 V7
Modified residue
Fig. 4. Anti-toxin (A) and anti-peptide (B) antibody reactivity with an
alanine replacement set of the N-terminal heptapeptide (VKDGYIV).
A set of rod-bound peptides was prepared in which each residue of the
AaH II sequence l-7 was replaced in turn by an alanine residue. The
peptide l-7 corresponding to unmodified N-terminal of AaH II (none)
and the set of seven analogs (identified by the amino acid which was
replaced by an alanine) were tested in Pepscan ELISA with (A)
anti-AaH II sera: Sl14 ( ), S182 (a), Sl84 (0) and (B) antG(l-8)
peptide sera: S8l (m) and S80 (8). Sera were diluted I:4000 (Sl14,
S80 and S81) or 1: 1500 (S182 and S184). Antibody binding was
revealed by goat anti-rabbit Ig peroxidase conjugate. Each absorbance
value was corrected for non specific binding by subtracting the signal
given by the control peptide (sequence 47-53: YCYKLPD) which was
shown to be unreactive in the previous Pepscan analysis (see Fig. 1).
Corrected absorbance value for the parent peptide (sequence l-7 from
AaH II) were taken as 100% reactivity, and the reactivity of alanine
analogs are expressed relatively to this value.
1066 C. DEVAUX et al.

peptides. These variations probably reflect the individual

immune responses of outbred animals which depend on
host factors including MHC diversity and the available
B cell repertoire.
These regions of antibody binding certainly do not
properly describe the complete antigenic structure of the
toxin, as most epitopes are an array of amino acids that
N-terminal heptapeptide from:
need the native conformation of the protein to be
immunoreactive (Benjamin et al., 1984; Davies et al.,

.?i
AaHI AaH II AaH 111 AaH IV
1990). However, it is frequently observed that short
segments of the protein exhibit some antigenicity, prob-
ably by cross-reacting with protein-specific, confor-
x loo mation-dependent, antibodies (Van Regenmortel, 1987).
.z
2 80 In the Pepscan format of immunoassay this type of
e 60 cross-reactivity is probably the rule. Some structural
9 40 motifs of protein architecture may facilitate the detection
._
1 20 of peptide-reactive anti-protein antibodies. Using the
z
c4 n 3D-model of AaH II (Fontecilla-Camps et al., 1988) the
N-terminal heptapeptide from:
six regions disclosed by the Pepscan technique may be
AaHl AaH II AaH 111 AaH IV mapped to the four b-turns of the AaH II (sequences
Fig. 5. Anti-toxin (A) and anti-peptide (B) antibody reactivity with 412, 27-35, 3945 and 52-58) to a P-sheet strand
a set of heptapeptides corresponding to the N-terminii of the four (sequence l-8) and to a non-typical extended region
AaH toxins. Peptides, mimicking the sequences 1-7 of four toxins
from AaH (I, KRDGYIV; II, VKDGYIV; III, VRDGYIV and IV, (sequence 55561). Using a similar technique to map
GRDGYIV), were assembled on rods and tested in ELISA with (A) linear epitopes of transforming growth factor o! (a mol-
anti-AaH II sera: S114 (m). S182 (m), S184 (0) and (B) anti-(l-8) ecule similar in size and overall shape to a scorpion
peptide sera: S81 (W) and S80 (m). Assay conditions were as in
Fig. 4. Corrected absorbance values for the sequence l-7 from AaH toxin), Hoeprich et al. (1989) also found a P-sheet and
II were taken as 100% reactivity. a B-turn to be enclosed in one of the two immunodom-
inant epitopes. All short linear epitopes of a completely
different protein (apolipoprotein A-I) also perfectly co-
AaH toxins. The pattern of recognition of these four incide with regions of the protein folded in B-turns
peptides by the serum S81 was essentially unchanged (Marcel et al., 1991) while Tan et al., (1990) found that
by serum dilution (1: 100&l : 15000). These results are most of the peptide epitopes of dihydrosulfide reductase
consistent with the alanine replacement experiments reside within, or at the boundaries of, p-sheets. Thus, the
(Fig. 4B) indicating a loose specificity of the serum. In experimental assay format is possibly biased for the
contrast, serum S80 recognized the AaH II N-terminal detection of epitopes situated in extended parts of the
peptide more efficiently than peptides derived from other protein backbone (p-sheets) or in locally folded parts
toxins (Fig. 5B). However, some cross-reactivity was (p-turns). To explain how anti-protein antibodies could
detected at a S80 dilution of 1: 400 which was not found bind to short peptides corresponding to these regions we
at 1:4000 (data not shown). Nonetheless, serum S80 have to assume that a proportion of each rod-bound
seemed to be quite specific for the AaH II sequence 1-7, peptide adopt an extended or p-turn-like conformation
a result that agrees with the specificity displayed in (Dyson et al., 1985).
alanine replacement experiments (Fig. 4A). The validity of the multiple synthesis approach to
map antigenic epitopes of toxins was assessed by com-
paring results from this study with previous results
DISCUSSION obtained by more conventional methods. Most of the
regions identified as immunoreactive by the Pepscan
The antigenicity of a set of 58 peptides derived from method are part of larger domains previously described
the amino acid sequence of the scorpion toxin AaH II as antigenic (i.e. they were both able to bind anti-AaH
was assessed by the so-called Pepscan method described II antibodies and to elicit rabbit anti-peptide antibodies
by Geysen et al. (1984, 1987). Six continuous antibody that recognize the native toxin either in solid and/or
binding sites were thereby identified by their capacity to liquid phase immunoassays): sequences 5-14 (El Ayeb
be recognized by one or several hyperimmune rabbit et al., 1984) 28-39 (Bahraoui et al., 1986a), 3646
anti-AaH II sera (Figs 1 and 2). The strongest reactivities (Fourquet et al., 1988) and 5&59 (Bahraoui et al.,
(S/N > 4) were displayed by peptides corresponding to 19866; Ait-Amara et al., 1993). Consistent with previous
sequences l-8,27-35 and 55-61 which were each recog- findings, two anti-AaH II monoclonal antibodies, that
nized by several sera. Three other sequences (residues recognize conformation-dependent epitopes on the toxin
4-12, 39-45, 52-58) were revealed by one serum with a (Bahraoui et al., 1988) did not bind to any rod-bound
S/N > 2. Consistent with previous studies (Geysen et al., heptapeptide (Yahi et al., 1992). Three differences how-
1984; Trifilieff et al., 1991) each rabbit serum showed a ever were observed between present and previous results.
specific profile of immunoreactivity towards the set of One is the absence of reactivity of any peptide in the
 Peptide epitopes
region 19-28 which has previously been found to be
antigenic (Bahraoui et al., 1986a; 1987). This region is
folded into a rigid a-helix in the toxin (Fontecilla-Camps
et al., 1988) and any of our heptapeptides were probably
too short to assume a stable u-helix conformation. A
second difference is that the linear epitope identified by
the Pepscan technique in the sequence 5561, only
partially overlaps with the antigenic region (50-59)
defined by Bahraoui et al. (19866). Finally, a major
difference concerns the N-terminal part of the molecule.
The Pepscan results indicated that the region l-8 is
strongly and frequently recognized by anti-toxin anti-
bodies (Figs 1 and 2). However, previous studies do not
clearly indicate antigenicity of region l-8: a synthetic
peptide l-8 has been found to bind anti-toxin antibodies
but was unable to induce antibodies which recognize
AaH II (Granier et al., 1989). It was thus necessary to
re-examine the antigenicity of the N-terminal region of
the toxin. Synthetic peptide l-8 linked to the KLH
protein carrier by its C-terminal residue was able to elicit
rabbit antibodies which efficiently recognized the parent
AaH II in ELISA (Fig. 3). A similar cross-reactivity was
observed with individual sera of C57 BL/6 mice immu-
nized with the same immunogen and also in a liquid-
phase radioimmunoassay using “‘1-Aah II as the ligand
(results not shown). Previous attempts to obtain anti-
(l-8) antibodies cross-reactive with the toxin pre-
sumably failed because the immunogens were not linked
via the C-terminal residue of the peptide. The orien-
tation of the peptide on the carrier protein may thus be
critical in eliciting a population of antibodies able to
recognize the l-8 domain in the native toxin. Thus, the
Pepscan method helped us to pinpoint a region of
antigenic reactivity which had previously been over-
looked. The characterization of the region l-8 as an
epitope was not surprising because ends of molecules are
frequently found antigenic (Palfreyman et al., 1984),
possibly because of their surface exposure and solvent
accessibility (Thornton and Sibanda, 1983). The avail-
ability of both anti-toxin and anti-peptide antibodies
able to recognize the parent toxin led us to investigate
their capacities to bind to peptide analogs of the N-ter-
minal sequence of AaH II (Fig. 4) and to peptides
mimicking each of sequences l-7 from the three other
toxins of the venom (Fig. 5). These experiments showed
that residues 2, 3 and 4 play a critical role in the
recognition of the sequence l-7. In addition, one of the
anti-peptide rabbit sera (S81) exhibited a loose specifi-
city, recognizing all but one alanine analog of the
(I-7)-AaH II sequence (Fig. 4B) and bound equally well
to N-terminal peptides of all AaH toxins (Fig. 5B). This
degenerate recognition was confirmed, by the capacity of
these antibodies to bind to various plate-coated toxins in
ELISA (data not shown). Thus anti-peptide antibodies
raised against the N-terminal sequence of one toxin
could bind to other toxins, in contrast to anti-toxin
antisera (Fig. 5A). Such an enlarged recognition capacity
may be useful to by-pass the lack of antigenic cross-
reactivity which is frequently observed when polyclonal
anti-toxin antibodies are used (Delori et al., 1981; El
of a scorpion toxin 1067
Ayeb et al., 1983; Mansuelle et al., 1992; De Lima et al.,
1993).
In conclusion, Pepscan analysis of AaH II epitopes
gave results generally in good agreement with previous
results. The main difference suggested the immunoreac-
tivity of the N-terminal part of the molecule. This was
confirmed by the capacity of anti-(l-8) peptide anti-
bodies to recognize the parent toxin. Additional work
indicated that some of these antibodies cross-react with
toxins which display the conserved N-terminal sequence
but which are normally not recognized by anti-toxin
antibodies. This observation could be extremely fruitful
for the preparation of peptide-based polyclonal sera,
able to recognize many different toxins sharing only a
limited set of conserved residues.
Acknowledgements-The authors are indebted to M. Alvitre
for animal care and to J. Mery and J. Brugidou for the
synthesis of peptide l-8. They thank M. El Ayeb and H.
Rochat for advice and encouragement.
REFERENCES
Ait-Amara D., Chavez-Olortegui C., Romi R., Mery J.,
Brugidou J., Albericio F., Devaux C. and Granier C. (1993)
Antibodies cross-reactive with the scorpion-toxin II from
Androctonus australis Hector elicited in mice by a synthetic
peptide. Natural Toxins 1, 255-262.
Amit A. G., Mariuzza R. A., Phillips S. E. V. and Poljak
R. J. (1986) Three-dimensional structure of an antigen-
antibody complex at 2.8 A resolution. Science 233, 747-753.
Arnon R. and Horwitz R. J. (1992) Synthetic pcptides as
vaccines. Current Opinion in Immunology 4, 449454.
Bahraoui E., El Ayeb M., Van Rietschoten J., Rochat H. and
Granier C. (1986a) Immunochemistry of scorpion tl -toxins:
study with synthetic peptides of the antigenicity of four
regions of toxin II of Androctonus australis Hector. Molec.
Immun. 23, 357-366.
Bahraoui E., Granier C., Van Rietschoten J., Rochat H. and
El Ayeb M. (1986b) Specificity and neutralizing capacity of
antibodies elicited by a synthetic peptide of scorpion toxin.
J. Immun. 136, 3371-3377.
Bahraoui E., El Ayeb M., Van Rietschoten J., Granier C. and
Rochat H. (1987) Immunochemistry of scorpion toxins.
Immunogenicity of peptide 19-28 a model of an accessible
and relatively rigid region. Eur. J. Biochem. 167, 371-375.
Bahraoui E., Pichon J., Muller J.-M., Darbon H., El Ayeb M.,
Granier C., Marvaldi J. and Rochat H. (1988) Monoclonal
antibodies to scorpion toxins. Characterization and molecu-
lar mechanisms of neutralization. J. Immun. 141, 214-220.
Benjamin D. C., Berzofsky J. A., East I. J., Hannum C.,
Sydney J. L., Margoliash E., Michael J. G., Miller A.,
Prager E. M., Reichlin M., Sercarz E. E., Smith-Gill S. J.,
Todd P. E. and Wilson A. C. (1984) The antigenic structure
of proteins: a reappraisal. A. Rev. Immun. 2, 67-101.
Catterall W. A. (1977) Membrane potential dependent binding
of scorpion toxins to the action potential Na+ ionophore.
J. biol. Chem. 252, 866&8668.
Davies D. R., Padlan E. A. and Sheriff S. (1990) Anti-
body-antigen complexes. A. Rev. Biochem. 59, 439473.
De Lima M.-E., Martin-Eauclaire M.-F., Chavez-Olortegui
C., Diniz C. R and Granier C. (1993) Tityus serrulatus
1068 C. DEVAUXet al.
scorpion venom toxins display a complex pattern of
antigenic reactivity. Toxicon 31, 223-227.
Delori P., Van Rietschoten J. and Rochat H. (1981) Scorpion
venoms and neurotoxins: an immunologica study. Toxicon
19, 393407.
Dyson H. J., Cross K. J., Houghten R. A., Wilson I. A., Wright
P. E. and Lemer R. A. (1985) The immunodominant site of
synthetic immunogen has a conformational preference in
water for a type-II reverse turn. Nature 318, 480-483.
El Ayeb M., Delori P. and Rochat, H. (1983) Immun~hem-
istry of scorpion u-toxins: antigenic homologies checked
with radioimmunoassays (RIA). Toxicon 21, 709-716.
El Ayeb M., Bahraoui E., Granier C., Delori P., Van Rietscho-
ten J. and Rochat H. (1984) Immunochemistry of scorpion
x-toxins: purification and characterization of two function-
ally inde~ndent IgG popdations raised against toxin II of
Androctonus australis Hector. Molec. Immun. 21, 223-232.
Fontecilla-Camps J.-C., Habersetzer-Rochat, C. and Rochat
H. (1988) Orthorhombic crystals and three-dimensional
structure of the potent toxin II from the scorpion Androc-
tonus australis Hector. Proc. natn. Acad. Sci. U.S.A. 85,
7443-7447.
Fourquet P., Bahraoui E., Fontecilla-Camps J.-C. Van Ri-
etschoten J., Rochat H. and Granier C. (1988) Immuno-
chemistry of scorpion toxins: synthesis and antigenic
properties of a model of a loop region specific to a-toxins.
Int. J. Peptide Protein Res. 32, 81-88.
Geysen H. M., Meloen, R. H. and Barteling S. J. (1984) Use
of peptide synthesis to probe viraf antigens for epitopes to
a resolution of a single amino acid. Proc. natn. Acad. Sci.
U.S.A. 81, 3998-4002.
Geysen H. M., Rodda S. J., Mason T. J., Tribbick G. and
Schoofs P. G. J. (1987) Strategies for epitope analysis using
peptide synthesis. .I. Immun. Meth. 102, 259-274.
Goyffon M., Vachon M. and Brogho N. (1982) Epidemiology
and clinical characteristics of scorpion envenomation in
Tunisia. Toxicon 20, 337-344.
Granier C., Novotny J., Fontecilla-Camps J.-C., Fourquet P.,
El Ayeb M. and Bahraoui E. (1989) The antigenic structure
of a scorpion toxin. Motet. Immun. 26, 5033513.
Hoeprich P. D., Langton B. C., Zhang J.-W. and Tam J. P.
(1989) Identification of immunodominant regions of trans-
forming growth factor a. J. biol. Chem. 264, 19,08619,091.
Jeddi H. M., Ben Hamida C., Damak J., Jeddi M. and Ayadi
N. (1988). In Piqlire de Scorpion Chez L’Homme en Tunisie:
ktude kpidemiologique (Fact&C de Medecine de Sfax).
Jemmerson R. (1987) Antigenicity and native structure of
globular proteins: low frequency of peptide reactive
antibodies. Proc. natn. Acad. Set. U.S.A. 84, 9180-9184.
Laver W. G., Air G. M., Webster R. G. and Smith-Gill S. J.
(1990) Epitopes on protein antigens: misconceptions and
realities. Cell 61, 553-556.
Lerner R. A. (1982) Tapping the immunological repertoire
to produce antibodies of predetermined specificity. Nature
299, 592-596.
Mansuelle P., Martin-Eauciaire M.-F. Rochat H. and Granier
C. (1992) The amino acid sequence of toxin IV from the
Androctonus australis scorpion. Differing effects of natural
mutations in scorpion a-toxins on their antigenic and toxic
properties. Natural Toxins, 1, 61-69.
Marcel Y. L., Provost P. R., Koa H., Raffai E., Vu Dac N.,
Fruchart J.-C. and Rassart E. (1991) The epitopes of
apolipoprotein A-I define distinct structural domains
including a mobile middle region. J. hiol. Chem. 266,
3644-3653.
Martin M.-F. and Rochat H. (1986) Large scale purification of
toxins from the venom of the scorpion Androctonus australis
Hector. Toxicon 24, 1131-I 139.
Mery J., Granier C., Juin M. and Brugidou J. (1993) Disulfide
linkage to polyacrylic resin for automated Fmoc peptide
synthesis. Immunological applications of peptide-resins and
mercapto amide peptides. Int. J. Peptide Protein Res. in
press.
Paifreyman J. W., Aitcheson T. C. and Taylor P. (1984)
Guidelines for production of polypeptide specific antisera
using small synthetic oiigopeptides as immunogens. 1.
immun. Meth. 75, 383-393.
Rochat H., Rochat C., Sampieri F. and Miranda F. (1972)
The amino acid sequence of neurotoxin II of Androctonus
australis Hector. Eur. J. Biochem. 28, 381-388.
Rochat H., Bernard P. and Couraud F. (1979) Scorpion toxins:
chemistry and mode of action. In Advances in Cytopharma-
coiogy. Neurotoxins: Tools in Neurobiology (Ceccarelli B.
and Clementi F., Eds). Vol. 3, pp. 325--334. New York:
Raven Press.
Tan X., Ratman M., Huang S., Smith P. L. and Freisheim
J. H. (1990) Mapping the antigenic epitopes of human
dihydrofolate reductase by systematic synthesis of peptides
on solid supports. J. moiec. Bioi. 265, 8022-8026.
Thornton J. M. and Sibanda B. L. (1983) Amino and carboxy-
terminal regions in globular proteins. J. molec. Biol. 167,
443460.
Trifilieff M. C., Dubs M. C. and Van Regenmortel M. H. V.
(1991) Antigenic cross-reactivity potential of synthetic pep-
tides immobilized on polyethylene rods. Molec. Immun. 28,
889-896.
Van Regenmortel M. H. V. (1987) Antigenic cross-reactiv-
ity between proteins and peptides: new insights and
applications. Trends Biochem. Sci. 12, 237-240.
Yahi N., Devaux C., ~ansueIle P., Defendini M.-L. and
Granier C. (1992) Monoclonal antibodies to toxin II from
the scorpion Androctonus australis Hector: further charac-
terization of epitope specificities and neutralizing capacities.
Toxicon 30, 723-73 1.