Solution structure of an M-1 conotoxin with a novel
disulfide linkage
Wei-Hong Du
1,2,
*, Yu-Hong Han
3,4,
*, Fei-juan Huang
1,
*, Juan Li
2
, Cheng-Wu Chi
3,4
and Wei-Hai Fang
2
1 Department of Chemistry, Renmin University of China, Beijing, China
2 Department of Chemistry, Beijing Normal University, China
3 Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Graduate School
of the Chinese Academy of Sciences, Shanghai, China
4 Institute of Protein Research, Tongji University, Shanghai, China
Over their 50 million years of evolution, cone snails
have developed a series of small disulfide-rich peptides
(conotoxins) in their venoms. Each peptide can selec-
tively target a specific isoform of ion channel or mem-
brane receptor [1,2]. Although it is estimated that each
species of cone snail possesses 50–200 conotoxins in its
arsenal, and there are more than 50 000 known cono-
toxins, the majority belong to several gene super-
families and only several structural motifs are widely
shared.
M-superfamily conotoxins form a group with a typ-
ical cysteine arrangement of (-CC-C-C-CC-) and par-

solution structure
Correspondence
W H. Fang, Department of Chemistry,
Beijing Normal University, 19 Xin Jie Kou
Wai St., Beijing 100875, China
Fax: +86 10 5880 2075
Tel: +86 10 5880 5382
E-mail:
C W. Chi, Shanghai Institute of
Biochemistry and Cell Biology, Chinese
Academy of Sciences, 320 YueYang Road,
Shanghai 200031, China
Fax: +86 21 5492 1011
Tel: +86 21 5492 1165
E-mail:
*These authors contributed equally to this
study
(Received 10 December 2006, revised 7
March 2007, accepted 16 March 2007)
doi:10.1111/j.1742-4658.2007.05795.x
The M-superfamily of conotoxins has a typical Cys framework (-CC-C-C-
CC-), and is one of the eight major superfamilies found in the venom of
the cone snail. Depending on the number of residues located in the last
Cys loop (between Cys4 and Cys5), the M-superfamily family can be divi-
ded into four branches, namely M-1, -2, -3 and -4. Recently, two
M-1 branch conotoxins (mr3e and tx3a) have been reported to possess a
new disulfide bond arrangement between Cys1 and Cys5, Cys2 and Cys4,
and Cys3 and Cys6, which is different from those seen in the M-2 and M-4
branches. Here we report the 3D structure of mr3e determined by 2D
1

5
[3]. In addition, BtIIIB, another M-2 branch
conotoxin from the venom of a vermivorous cone snail
Conus betulinus, has been proven to have the same
disulfide linkage as mr3a and tx3c [8].
Recently, a third disulfide bond arrangement within
M-superfamily conotoxins has been characterized. Two
M-1 branch conotoxins, mr3e (Fig. 1) and tx3a were
found to have a new disulfide linkage (C
1
–C
5
,C
2
–C
4
,
C
3
–C
6
), which differs from those seen in M-2 and M-4
branch conotoxins [9]. Here we report the 3D structure
of mr3e, a novel M-1 branch conotoxin with the above
new disulfide connectivity.
Results
Sequence-specific resonance assignments
2D NMR spectroscopy was used to investigate the 3D
structure of conotoxin mr3e in aqueous solution at
pH 3. Proton resonances for conotoxin mr3e were

geometry calculations. Figure 2 shows the sequential
d
aN(i,i+1)
connectivities on the CaH-NH fingerprint
region of the NOESY spectrum with a mixing time of
200 ms. All chemical shifts are listed in Table 1.
Structure calculation and evaluation
NMR experiments provided enough distance and angle
constraints to calculate the structure of mr3e. The con-
straints for structure elucidation were determined from
a survey of NMR data using the traditional visual ana-
lysis method developed by Wuthrich [10]. In total, 169
distance constraints were obtained from the 200-ms
NOESY spectrum. Six u angle constraints and three
disulfide bonds from Cys2 to Cys14, Cys3 to Cys12,
and Cys8 to Cys15 were added to the distance
constraints for primary structure determination. A set
of 20 structures was generated with a mean global
rmsd of 1.88 A
˚
using the dyana (v. 1.5) [11] software
package. The lowest energy structure was then dis-
played, and ambiguous NOESY signals were evaluated
Fig. 1. Conotoxin mr3e sequence and its disulfide linkage.
3.5
E10
V1
G6
F5
Y13

mean global rmsd of 0.69 A
˚
.
dyana was used to provide hydrogen-bond informa-
tion during the minimization. Deuterium-exchange
studies indicated that hydrogen bonds might form
exCysist between the amide protons of Gly7, Cys8 and
Cys12 and nearby oxygen or nitrogen atoms. The reso-
nances of amide protons in these residues were not
diminished after 3 h in D
2
O at 294 K in a 1D proton
time course experiment. dyana provided hydrogen
bond acceptor oxygen and nitrogen atoms for each of
the amide protons from these four residues. The
hydrogen bonds for Gly7, Cys8 and Cys12 were used
as constraints. Thus, six upper and six lower distance
constraints were added for the hydrogen-bond interac-
tions, and another round of minimization was per-
formed. The result was a final set of 20 structures with
a mean global backbone rmsd of 0.56 ± 0.16 A
˚
and a
mean global heavy atom rmsd of 1.30 ± 0.28 A
˚
.
Finally, refinement of the structure was carried out
using amber 5 [12] for energy minimization. An
ensemble of 20 structures with lower energy and better
Ramachandran plots was chosen to represent the 3D

Cys12 is apparently stabilized by a hydrogen bond
between the carbonyl oxygen of His9 and the amide
proton of Cys12. The interaction is characteristic of a
type II b-turn.
Table 1. Proton resonance assignments (p.p.m.) for mr3e.
Residue HN abOther
Val1 3.88 2.25 c: 1.06 ± 1.03
Cys2 8.79 4.92 2.72, 2.51
Cys3 8.49 4.49 3.72, 3.47
Pro4 4.6 2.30, 2.03 c: 2.11, 1.81
d: 3.77, 3.61
Phe5 8.75 4.2 3.05, 2.98 d: 7.26
e: 7.31
f: 7.54
Gly6 8.65 4.03, 3.53
Gly7 8.18 4.58, 3.31
Cys8 8.42 4.7 3.06
His9 7.3 4.89 3.53, 3.26 d: 7.29
e: 8.67
Glu10 8.92 4.15 2.13 c: 2.55
Leu11 8.45 4.08 2.01, 1.72 c: 1.61
d: 0.94 ± 0.89
Cys12 7.57 4.46 3.22, 3.17
Tyr13 9.08 4.34 3.23, 3.04 d: 7.17
e: 6.88
Cys14 7.96 4.64 3.79, 3.16
Cys15 9.44 5.07 3.42, 3.03
Asp16 8.76 4.49 2.69, 2.52
Table 2. Structural statistics for the family of 20 structures of cono-
toxin mr3e.

back views of the surface of the peptide. The double-
turn conformation in conotoxin mr3e produces an over-
all shape of a ‘flying bird’ when viewed from the front.
Discussion
M-superfamily conotoxins, one of the major groups of
disulfide-rich peptides, are widely distributed in the
venoms of all three feeding types of cone snails.
Depending on the number of residues located in the
last Cys loop, M-superfamily conotoxins have been
provisionally divided into four branches, namely M-1,
-2, -3, -4. Interestingly, to the best of our knowledge,
three different disulfide linkages can be found in
M-1 (1–5, 2–4, 3–6), M-2 (1–6, 2–4, 3–5) and M-4
(1–4, 2–5, 3–6) branch conotoxins, respectively.
mr3e is an M-1 branch conotoxin purified from the
venom of a mollusk-hunting cone snail, C. marmoreus;
it has 16 amino acids in its mature peptide. Previously,
we have shown that mr3e is characterized by its dis-
tinctive disulfide connectivity (C
1
–C
5
,C
2
–C
4
,C
3
–C
6

molecule (Fig. 6C).
In contrast to the typical excitory symptoms, such
as circular movements, barrel rolling and convulsions,
elicited by cranial injection of mr3a [3], mr3e has no
obvious effect on mice [9]. Therefore, it is most likely
that these two conotoxins have different physiological
functions, and this is not surprising considering that
they have completely different backbone scaffolds.
Although M-1 and M-2 branch conotoxins are similar
in size and cysteine framework, and are all abundant
in mollusk- and worm-hunting cone snails, more evi-
dence has emerged that they are phylogenetically diver-
gent groups. These two groups of M-superfamily
conotoxins differ with respect to signal peptide
sequence, disulfide linkage, backbone scaffold and
most likely molecular target.
It seems to be a favored strategy of cone snails to
generate different backbone scaffolds within conotox-
ins by introducing different disulfide linkages into
conotoxins that share the same cysteine framework.
For instance, a-conotoxin and v-conotoxin share the
same ‘-CC-C-C-’ cysteine framework, but differ greatly
in disulfide linkage, backbone scaffold and conse-
quently molecular target [16–18]. Such a strategy,
which yields more structural and functional diversity
in the conotoxins, will help cone snails to survive
severe environmental pressures.
Experimental procedures
Peptide synthesis and refolding
mr3e was chemically synthesized as described previously [9].

O ⁄ D
2
O(v⁄ v) were acquired
with the transmitter set at 4.80 p.p.m. and a spectral win-
dow of 6000 Hz. All 2D NMR spectra were acquired in a
phase-sensitive mode using time-proportional phase incre-
mentation for quadrature detection in the t
1
dimension.
Presaturation during the relaxation delay period was used
to solvent resonance. A series of NOESY spectra was
acquired with mixing times of 400, 200, 150 100 and 50 ms.
TOCSY spectra under both solvent conditions were
acquired with a mixing time of 120 ms.
Spectra were processed using xwinnmr or topspin soft-
ware. Phase-shifted sine-squared window functions were
applied before Fourier transformation, with shifts of 60 or
90 ° in both dimensions. Final matrix sizes were usually
2048 · 2048 real points. To identify the slow exchange of
backbone amide protons, the sample lyophilized from a
H
2
O solution was redissolved in D
2
O. 1D
1
H spectra were
measured after 5 min, and subsequently every 0.5 h up to
20 h. Chemical shifts were referenced to the methyl reson-
ance of 4,4-dimethyl-4-silapentane-1-sulfonic acid used as

of a high resolution 1D proton spectrum of conotoxin mr3e.
The u angle constraints were set to )120 ± 40° for
3
J
NHa
> 8.0 Hz (Gly7, Glu10) and to )65±25° for
3
J
NHa
< 5.5 Hz (Cys3, Phe5, Leu11, Cys12). Backbone dihedral
constraints were not applied for
3
J
NHa
values between 5.5
and 8.0 Hz. After the initial calculation, hydrogen-bonds
constraints were added as target values of 2.2 A
˚
for NH(i)–
O(j) and 3.2 A
˚
for N(i)–O(j), respectively.
One thousand random structures were generated by
dyana (v. 1.5) that fit the primary sequence and covalent
and spatial requirements of mr3e. A total of 190 distance
constraints, six u angle restraints and three hydrogen bonds
constraints were input for the molecular modeling protocol
for the dyana algorithm. The outcome was a set of 20
structures with a mean global rmsd of 0.56 ± 0.16 A
˚

GIIIA, a peptide ligand for muscle sodium channels:
chemical synthesis, radiolabeling, and receptor charac-
terization. Biochemistry 28, 3437–3442.
5 Shon KJ, Grilley M, Jacobsen R, Cartier GE, Hopkins
C, Gray WR, Watkins M, Hillyard DR, Rivier J,
Torres J et al. (1997) A noncompetitive peptide
inhibitor of the nicotinic acetylcholine receptor from
Conus purpurascens venom. Biochemistry 36, 9581–
9587.
6 Ferber M, Sporning A, Jeserich G, De La Cruz R,
Watkins M, Olivera BM & Terlau H (2003) A novel
Conus peptide ligand for K
+
channels. J Biol Chem
278, 2177–2183.
7 Al-Sabi A, Lennartz D, Ferber M, Gulyas J, Rivier JE,
Olivera BM, Carlomagno T & Terlau H (2004)
Kappa M-conotoxin RIIIK, structural and functional
novelty in a K
+
channel antagonist. Biochemistry 43,
8625–8635.
8 Zhao TY, Cao Y, Dai XD, Fan CX & Chen JS (2005)
Purification, sequence and disulfide bonding pattern of
a novel conotoxin BtIIIB. Acta Chimica Sinica 63,
163–168.
9 Han YH, Wang Q, Jiang H, Liu L, Xiao C, Yuan DD,
Shao XX, Dai QY, Cheng JS & Chi CW (2006) Charac-
terization of novel M-superfamily conotoxins with new
disulfide linkage. FEBS J 273, 4972–4982.

tion from the venom of Conus marmoreus. J Biol Chem
275, 39516–39522.
17 McIntosh JM, Corpuz GO, Layer RT, Garrett JE,
Wagstaff JD, Bulaj G, Vyazovkina A, Yoshikami D,
Cruz LJ & Olivera BM (2000) Isolation and characteri-
W H. Du et al. Solution structure of an M-1 conotoxin
FEBS Journal 274 (2007) 2596–2602 ª 2007 The Authors Journal compilation ª 2007 FEBS 2601
zation of a novel conus peptide with apparent antinoci-
ceptive activity. J Biol Chem 275, 32391–32397.
18 Sharpe IA, Gehrmann J, Loughnan ML, Thomas L,
Adams DA, Atkins A, Palant E, Craik DJ, Adams DJ,
Alewood PF et al. (2001) Two new classes of conopep-
tides inhibit the alpha1-adrenoceptor and noradrenaline
transporter. Nat Neurosci 4, 902–907.
Solution structure of an M-1 conotoxin W H. Du et al.
2602 FEBS Journal 274 (2007) 2596–2602 ª 2007 The Authors Journal compilation ª 2007 FEBS