Content uploaded by David J. Adams
Author content
All content in this area was uploaded by David J. Adams on Mar 05, 2018
Content may be subject to copyright.
Research Overview
Conotoxins and Their Potential Pharmaceutical
Applications
David J. Adams,1* Paul F. Alewood,2 David J. Craik,2 Roger D. Drinkwater,3 and
Richard J. Lewis1,2
1Department of Physiology and Pharmacology, University of Queensland, Brisbane, QLD, Australia
2Centre for Drug Design and Development, and 3CSIRO, Gehrmann Laboratories, University of Queensland,
Brisbane, QLD, Australia
ABSTRACT The neurotoxins isolated from cone shell venoms are a diverse group of small, disulfide-
rich peptides. Most of the active peptides isolated to date have been shown to specifically target various
components of neural transmission, and have generally demonstrated high specificities for ion channel and
receptor types and subtypes. The specificity of conotoxins is one of the attributes that make them valuable
diagnostic tools in the characterisation of neural pathways, as therapeutic agents in medicine, and poten-
tially as biodegradable toxic agents in agroveterinary applications. The number of novel, active peptides
within the numerous Conus species is considered to be enormous. Currently, however, relatively few pep-
tides have been characterised. In this article, we review current research on conotoxins with a focus on
drug potential being developed at the University of Queensland, Australia. Drug Dev. Res. 46:219–234,
1999. © 1999 Wiley-Liss, Inc.
Key words: ion channels; sodium channel; acetylcholine; nicotinic receptor; synaptic transmission; peptide; gene cloning;
NMR spectroscopy; crystal structure
Strategy, Management and Health Policy
Venture Capital
Enabling
Technology
Preclinical
Research
Preclinical Development
Toxicology, Formulation
Drug Delivery,
Pharmacokinetics
Clinical Development
Phases I-III
Regulatory, Quality,
Manufacturing
Postmarketing
Phase IV
© 1999 Wiley-Liss, Inc.
DRUG DEVELOPMENT RESEARCH 46:219–234 (1999)
INTRODUCTION: ION CHANNELS AS
DRUG TARGETS
Voltage-dependent ion channels are intrinsic mem-
brane proteins that play an important role in fast com-
munication in excitable cells. A short stretch of amino
acids, the pore region, is the sole determinant of cation
selectivity and also forms the binding site for many chan-
nel blockers. Toxins that interact intimately with this re-
gion can be used as structural templates to deduce the
spatial organisation of the pore region of the ion chan-
nels. These models of pore structure are valuable for
understanding the mechanisms of ion permeation, and
ultimately may be useful for the rational design of drugs
that modify the function of ion channels in clinical con-
ditions such as stroke, pain, or epilepsy.
Broadly, ion channels have structural and functional
similarities, but even within a class of ion channels there
are significant differences that can be targeted in drug
applications. The diversity and distribution of ion chan-
nel types and subtypes being uncovered through the use
of molecular biology and toxin probes present an excit-
ing opportunity for the discovery of new therapeutics
which are specific for channel subtypes involved in dis-
ease states. The various ion channels to be considered
will be examined briefly in turn.
Nicotinic Acetylcholine Receptor-Channels
The nicotinic acetylcholine receptor (nAChR) is
part of the ligand-gated ion channel superfamily, which
includes the GABAA, serotonin, and glutamate (NMDA,
Contract grant sponsors: DIST and the Australian Research
Council; Contract grant number: 96/ARCL244G.
*Correspondence to: Professor David J. Adams, Department
of Physiology and Pharmacology, University of Queensland, St. Lucia,
QLD 4072, Australia. E-mail: dadams@plpk.uq.edu.au
220 ADAMS ET AL.
AMPA, kainate) receptors. All ligand-gated ion channels
are large, membrane-bound pentamers with various sub-
unit compositions. These receptors have several conserved
features. Ligand-gated ion channels are pentamers, with
each subunit containing four transmembrane helices (M1
to M4), with the M2 helix lining the ion channel lumen
and providing it selectivity. Binding of an endogenous
ligand to a large, extracellular domain remote to the M2
helix brings about a conformational change in the M2
helices that causes the pore to open. Due to the size of
these receptors (~290 kDa), the only direct structure
determinations have been of low resolution (~9 Å) using
electron microscopy [Unwin, 1998].
Nicotinic ACh receptors are found throughout the
central and peripheral nervous systems, with distinct
genes encoding the nAChR subunits which form a
heteropentameric ion channel complex selective for cat-
ions. The muscle-subtype nAChR has been well
characterised due to the availability of specific probes
(e.g., α-bungarotoxin) and has the subunit composition
(α1)2β1δγ or ε in mature muscle. In mammalian central
and autonomic neurones and adrenal medulla, the neu-
ronal nAChRs are composed of α and β subunits only. At
least seven different α subunits (α2–α7 and α9) and three
β subunits (β2–β4) have been identified and it has been
shown that α2, α3, and α4 can combine with β2 or β4 to
form functional channels in the Xenopus oocyte expres-
sion system [McGehee and Role, 1995]. In addition, α7
and α9 subunits can be expressed as functional homo-
oligomers in this system, with the α7 gene product be-
ing α-bungarotoxin-sensitive and highly permeable to
Ca2+ [Colquhoun and Patrick, 1997]. Although these neu-
ronal nAChR subunits are homologous with one another,
each functional subunit combination is physiologically
and pharmacologically distinct. This may account for the
diversity of neuronal nAChRs observed in vivo. For ex-
ample, the α5 subunit appears to participate in nAChRs
expressed in heterologous systems and primary neurones
and contributes to the pore lining of functionally unique
nAChRs. Recent studies using single cell RT-PCR analy-
sis of nAChR gene transcripts indicate that multiple
nAChR subtypes are expressed by individual rat intrac-
ardiac neurones and that the combination of subtypes
expressed varies among cells [Poth et al., 1997]. The de-
velopment of specific pharmacological probes for neu-
ronal nAChR subunits will provide new insight into the
structural composition and functional role of the differ-
ent neuronal nAChRs subtypes.
Activation of distinct subtypes of these presynaptic
nAChRs by nicotinic agonists can selectively regulate the
release of different neurotransmitters, including dopam-
ine, norepinephrine, glutamate, and acetylcholine [Kulak
et al., 1997; Kaiser et al., 1998; Picciotto et al., 1998].
Such receptors have also been implicated in the patho-
physiology of several neuropsychiatric disorders, includ-
ing schizophrenia, Alzheimer’s disease, Parkinson’s dis-
ease, and Tourette’s syndrome [Kulak et al., 1997].
Despite their importance, few of the nicotinic receptor
antagonists identified to date are highly selective between
the multiple neuronal nAChR subtypes. Thus, the abil-
ity of recently discovered α-conotoxins — small (12–19
amino acids), rigid, highly disulfide-bonded peptides iso-
lated from marine snails of the genus Conus — to target
neuronal nAChR subunits with high specificity has con-
siderable significance for both basic neuroscience and
potential drug development.
Sodium Channels
Sodium channels consist of three separate and bio-
chemically separable protein subunits, the α plus β-1
and β-2 auxiliary subunits, which comprise the channel
in a 1:1:1 stoichiometry. The α-subunit is a transmem-
brane glycoprotein of approximately 260 kDa molecular
weight that binds a diverse range of neurotoxins at spe-
cific positions on its surface (six or seven sites are cur-
rently identified). The two β-subunits have smaller
molecular weights (~30 kDa each) and are integral mem-
brane glycoproteins. Numerous models of sodium chan-
nel α-subunit structure have appeared, based on primary
sequence data [Noda et al., 1984]. All show four highly
homologous regions of sequence domains, labeled I–IV,
with each domain containing six transmembrane heli-
ces, denoted S1–S6. The S5 and S6 segments of each
domain are highly nonpolar, the S1, S2, and S3 segments
are relatively nonpolar, with just a few charged
sidechains, but the S4 segments within each domain have
the distinctive feature that every third residue is posi-
tively charged (mostly arginines). The S4 segments are
believed to move upward on depolarisation to open the
activation gate (m gate) and allow the selective influx of
sodium ions. In the process, voltage-dependent move-
ment of an IFM particle to interact with adjacent intra-
cellular loops is facilitated and inactivation occurs,
blocking the further flow of ions (Fig. 1).
There is considerable structural homology among
the three types of brain Na+ channel α-subunits (I, II,
and III), the µ1-sodium channel α-subunit from adult
skeletal muscle, and the h1 sodium channel α-subunit
from heart and denervated muscle. Despite these simi-
larities, considerable pharmacological diversity exists. For
example, tetrodotoxin (TTX) blocks the brain types I, II,
and III at nanomolar concentrations, and the h1 form
from the heart at micromolar concentrations.
Until recently, there was no hard evidence to indi-
cate that pharmacologically distinct forms of neuronal
sodium channels are expressed in sensory neurons, and
thus no evidence that a specific Na+ channel pathway
could be modulated to control particular diseases. The
PHARMACEUTICAL APPLICATIONS OF CONOTOXINS 221
newly discovered TTX-insensitive sodium channel,
named PN3 or SNS [Sangameswaran et al., 1996], which
is located specifically in sensory neurons, represents one
of a number of potential Na+ channel targets for drug
discovery. Additional neuronal pathways for therapeutic
intervention may also be uncovered using conopeptides
such as µ-conotoxin PIIIA, the first conopeptide to dis-
tinguish amongst neuronal TTX-sensitive Na+ channels
[Shon et al., 1998; Watson et al., 1998].
Calcium Channels
Structurally, the calcium channels are closely re-
lated to sodium channels, with the main difference be-
ing the positioning and nature of the residues that line
the selectivity filter in the pore of the channel. There
are at least six pharmacologically distinct calcium chan-
nels types, including L-, N-, P/Q-, T, and R-type cal-
cium channels, and within each group are multiple
subtypes that are presently less easy to distinguish. In
the nervous system, several types of ion channels may
contribute to processes such as neurotransmitter release,
with the ratio and role for each type varying among dif-
ferent nervous tissues [Olivera et al., 1994]. This situa-
tion provides the possibility for selective modulation of
nerve function with type and subtype selective modula-
tors that may allow the selective treatment of conditions
such as pain and stroke. The ω-conotoxins have been of
enormous importance as physiological tools, with cur-
rently one peptide (MVIIA or Ziconitide) in clinical tri-
als for pain and stroke.
Potassium Channels
There are numerous types of potassium channel,
each with its own distinctive electrophysiological and
pharmacological properties; what they all have in com-
mon is that they tend to stabilise the membrane poten-
tial at the K+ equilibrium potential. DNA sequencing
reveals that the potassium channels encoded by Droso-
phila and vertebrate genes all resemble a single domain
of the voltage-dependent sodium channel [Jan and Jan,
1997]. Voltage-dependent potassium channels are tet-
rameric homo-oligomers organised in axial fourfold sym-
metry around the K+-selective pore. Analogous to
voltage-dependent sodium and calcium channels, the S4
transmembrane segment carries a cluster of positively
charged residues and is thought to act as the voltage sen-
Fig. 1. Transmembrane segments of the sodium channel and a model
of their construction into the α-subunit that forms the pore of the chan-
nel. The repeating nature of the transmembrane domains is emphasised,
as is the IFM particle that acts as the inactivation gate. The nature of the
residues comprising the activation gate is presently not known. Interest-
ingly, the calcium channel has the same general structure.
222 ADAMS ET AL.
sor for channel activation. Site-directed mutagenesis
studies, coupled with the use of selective toxins, have
proved invaluable in unraveling which residues of the
potassium channel protein are functionally important.
Recently, the crystal structure of a K+ channel has been
determined [Doyle et al., 1998]. The pore structure de-
termined previously from toxin binding interaction stud-
ies has proved to be remarkably predictive [Miller, 1995],
though it lacks the structural detail obtained by X-ray
crystallography. κ-Conotoxin PVIIA is a new structural
class of K+ channel blocking peptide that binds in a volt-
age-sensitive manner to the outer vestibule of the chan-
nel [Scanlon et al., 1997].
DISCOVERY AND CHARACTERISATION OF
NOVEL CONOTOXINS
Tropical waters, especially in coral reef ecosystems,
house an extraordinary diversity of invertebrate species,
many of whom use novel bioactive compounds as part of
defensive or prey capture strategies. The cone shells
which comprise a group of some 500+ predatory mol-
luscs are the most specialised, with venoms that target
fish, worms, and other molluscs. The venom is injected
through a harpoon-like apparatus and contains a com-
plex mix of small, constrained peptides which contain
10–40 amino acids and up to five disulfide bonds [Myers
et al., 1993]. This cocktail of peptides targets a diverse
range of voltage-sensitive sodium, calcium, and potas-
sium channels and N-methyl-d-aspartate, glutamate, va-
sopressin, serotonin, and acetylcholine receptors, which
leads to an immediate and efficient immobilisation of the
prey. The conotoxins present in the venom have been
divided into a number of major classes based on their
pharmacological activity and cysteine frameworks (Table
1). Their high potency and specificity, and convenient
chemical synthesis, also make the conotoxins attractive
leads in drug design programs. In addition to the
conotoxins being among the smallest bioactive peptides,
they are unusual in containing a high density of cysteine
residues and posttranslation modifications, including
hydroxylation, carboxylation, amidation, sulphation, and
bromination. These features often complicate their
chemical characterisation and occasionally their chemi-
cal synthesis.
All major classes of conotoxins have been identified
through initial in vitro or in vivo functional assays [Olivera
et al., 1990]. Screening based on receptor-binding dis-
placement of radiolabeled ligands is also playing a major
role. At the 3D Centre, University of Queensland, sensi-
tive 125I-GVIA and 125I-MVIIC assays have been estab-
lished for rat and human brain preparations to allow for
the isolation of new ω-conotoxins. More recently, other
chemical and molecular biology approaches have facili-
tated the identification and primary structure determina-
tion of new conotoxins. In practice, all of these approaches
are used in concert to discover new conotoxins.
The realisation that most if not all conotoxins were
biologically active led us to establish chemical approaches
to rapidly identify new conotoxins and confirm the pres-
ence of known conotoxins. The starting source of venom
was either from the dissected venom ducts of Conidae or
from the milked venom of captive species. The venom
paste was then extracted with varying amounts of aceto-
nitrile acidified with 0.1% trifluoroacetic acid. This pro-
cedure efficiently extracts most of the conotoxins present.
Early research findings at the 3D Centre revealed con-
siderable inter- and intraspecies variability in the com-
ponents in cone shell venoms and also that most species
contained in excess of 100 different peptides [Bingham
et al., 1996]. This analysis was facilitated by the applica-
tion of Ionspray mass spectrometry, which dramatically
reduced the time and quantity of venom required to
characterise the components of these complex mixtures
[Lewis et al., 1994; Bingham et al., 1996; Jones et al.,
1996]. An example of an LC/MS analysis of the peptides
present in the crude venom from Conus geographus is
given in Figure 2. From analyses of more than 30 spe-
cies, it is evident that the 60+ conotoxins reported to
date represent less than 0.1% of the peptides present in
the venoms of Conidae.
HPLC/electrospray mass spectrometry analysis is
generally complemented with a suite of chemical tech-
niques to rapidly “mass profile” each crude venom. The
tagging of each molecular component has facilitated the
subsequent isolation and characterisation of novel pep-
tides. Fractionation of the venom is often directed by the
mass and number of disulfide bonds present in the pep-
tide. Posttranslational modifications, which are common
in cone shell venoms, are usually identified by MS/MS,
enzymatic degradation/MS studies, amino acid analysis,
and Edman chemistry [Loughnan et al., 1998]. Fortu-
nately, most conopeptides are not N-terminally blocked.
The determination of disulfide bond connectivity
for many conotoxins remains challenging. Classical ap-
proaches using enzymic degradation often fail, as most
conotoxins are resistant to proteolysis, even with high
levels of enzyme present. Success has been achieved us-
ing a reductive alkylation/Edman sequence strategy
[Gray, 1993]. However, this approach occasionally fails,
as the alkylation step is performed under basic condi-
tions where scrambling may occur. Recently, we devel-
oped a more general approach that employs both mass
spectrometry and Edman chemistry [Jones et al., 1996].
Briefly, the conotoxin is sequentially reduced and alky-
lated under acidic conditions with mass spectrometric/
HPLC analysis and Edman sequencing. For smaller pep-
tides (e.g., the α-conotoxins), the differentially alkylated
products need only be subjected to collision-induced dis-
PHARMACEUTICAL APPLICATIONS OF CONOTOXINS 223
sociation to locate the labeled cysteine residues and hence
deduce the disulfide bond connectivity pattern.
Conotoxins are synthesised by cone shells from
mRNA templates derived from toxin genes, and expressed
in the venom ducts as precursor peptides. There are now
numerous gene cloning techniques that can be used to
isolate and characterise the precursor molecules, as a
prelude to predicting the composition of the mature pep-
tide. The mRNA can be isolated and converted to either
single-stranded (ss) or double-stranded (ds) complemen-
tary DNA (cDNA). Cloning of the ds-cDNA produces a
venom duct library, which can be screened with DNA
probes from known toxin mature peptide sequence or
precursor peptide sequence to find closely related clones.
This strategy was successfully used to isolate and define
the precursor structure of the ω-conotoxin GVIA from a
C. geographus library [Colledge et al., 1992]. An alterna-
tive approach is to make use of polymerase chain reac-
tion (PCR) technologies. In this method, oligonucleotide
primers homologous to known mature peptide sequence
TABLE 1. Six Major Classes of Conopeptides and Their Disulfide Connectivity
α-conopeptides (2 loop framework peptides that inhibit nicotinic acetylcholine receptors)
GI E CCN– P A CGR HY S – – C*
GIA E CCN– P A CGR HY S – – CGK*
GII E CCH– P A CGK HF S – – C*
MI G R CCH– P A CGK NY S – – C*
SI I CCN– P A CGP K Y S – – C*
SIA Y CCH– P A CGK NF D– – C*CC..C...C
SII G CCCNOACGP B Y G– – CGT S CS
PnIA G CCSLPPCA A NNP DY C*
PnIB G CCSLPPCAL S NPDYC*
ImI G CCSDPRCAWR– –– –C*
EI R D O CCYHPTCNMS NP QI C*
MII G CCSNPVCHL E HS NL C*
EpI G CCSDPRCNMNNP DY (SO
4
)C*
AuIB G CCSYPPCFATNPD–C
µ-conopeptides (3 loop framework that block sodium channels)
GIIIA R D CCTOOKKCKDRQCKOQRCCA* CC.C...C.CC
GIIIB R D CCTOORKCKDRRCKOMKCCA*
GIIIC R D CCTOOKKCKDRRCKOLKCCA*
PIIIA R L CCGF OK S CRS RQCKOHRCC*
ω-conopeptides (4 loop framework peptides that block calcuim channels)
GVIA CKSOGSS CSOTSYNCC–RSCNOY T K R CY
GVIB CKSOGSS CSOTSYNCC–RSCNOY T K R CYG*
GVIC CKSOGSS CSOTSYNCC–RSCNOY T K R C*
SVIA CRSSGSOCGV T S I – CCGR – C––YRGKCT*
SVIB CKLKGQSCRKTS Y DCCSGSCGR S – GK C*C...C..CC.C..C
GVIIA CKSOGTOCSRGMRDCC––SCLLYSNKCRRY*
GVIIB CKSOGTOCSRGMRDCCT–SCLSYSNKCRRY*
MVIIA CKGKGAKCSRLMYDCCTGSCRS––GKC*
MVIIB CKGKGAS CHR T S Y DCCTGSCNR––GKC*
MVIIC CKGKGAPCRKTMYDCCSGSCGR R – GK C*
MVIID CQGR G A S CRKTMY NCCSGSCNR––GRC*
TVIA CLSOGSSCSOTSYNCC–RSCNOY S R K CY*
δ-conopeptides (4 loop framework peptides that delay inactivation of sodium channels)
TxVIA W CKQS GEMCNLLDQNCCDGY – CIVLVCT
TxVIB W CKQSGEMCNLLDQNCCDGY – CIVFVCTC...C..CC.C..C
GmVIA V K P CRKE GQL CDP I F QNCCRGWNC–VLFCV
NgVIA S K CFSOGTFCGI K OGL CCSVR–CFSLFCISFE
PVIA E A CYAPGTFCGI K OGL CCSEF–CLPGVCFG*
κ-conopeptides (4 loop framework pepetide that blocks Shaker potassium channels)
PVIIA CR I ONQK CFQHLDDCCS RKCNR F NKCVC...C..CC.C..C
conantokins (helical peptides that inhibit the NMDA-glutamate receptor)
Con-G G E Z Z L Q Z N Q Z L I R Z K S N
Con-T GEZZYQKML ZNLRZAE VKKNA
Amino acid sequences shown with cysteines (bold) aligned within each structural framework.
*Processed carboxyl terminal; O = hydroxyproline residue, Z = γ-carboxyglutamic acid residue. Letter prefixes indicate conopeptides from the fish
hunters C. magus (M), C. geographus (G), C. tulipa (T), C. striatus (S), C. purpurascens (P), and C. ermineus (E); the mollusc hunters C. textile (Tx), C.
episcopatus (Ep), C. gloriamaris (Gm), C. nigropunctatus (Ng), and C. aulicus (Au); and the worm hunter C. imperialis (Im).
224 ADAMS ET AL.
can be used to derive 5′ leader propeptide and untrans-
lated sequence using adaptor ligated ds-cDNA (5′RACE).
This sequence can then be used to identify conserved
regions in the precursor leader sequences in which to
position oligonucleotide primers that are specific to
conopeptide families. PCR using these specific primers
in conjunction with a 3′ anchor primer on venom duct
ss-cDNA will produce amplified copies of the expressed
peptides in that particular family (3′ RACE). Cloning and
sequencing will produce the full peptide sequence, from
which the mature peptide region can be predicted. Apart
from the targeted approaches of the library and PCR strat-
egies, the complete screening of venom duct cDNA li-
braries in a manner similar to an EST (expressed
sequence tag) strategy is quite feasible and very produc-
tive. Most conopeptide sequences are less than 1,000
nucleotides, allowing complete sequencing of each pep-
tide gene simply by using primer sites based on the vec-
tor sequence of the clones. While the molecular cloning
methods of conotoxin isolation does have a number of
distinct benefits in comparison to assay directed fraction-
ation of whole venom, they have the disadvantage of not
being able to predict posttranslational modifications of
the mature peptides. Conopeptides can be highly and
unusually modified, such as the alpha peptide EpI, which
has a sulphated tyrosine [Loughnan et al., 1998]. These
modifications provide chemical alterations that may well
be important in the activity of the conopeptide at the re-
ceptor target. At present, the gene structures that com-
bine to produce a toxin peptide precursor mRNA
transcript are not known. The identification of these genes
and the mRNA splicing pathways that ultimately pro-
duce the highly variable toxin peptides will provide a
much better understanding of toxin peptide evolution in
the Conus species, and will undoubtedly lead to more
effective strategies for library-based and PCR-based toxin
peptide isolation.
CONOTOXIN SYNTHESIS, FOLDING, AND
PURIFICATION
All conotoxins described to date, with the excep-
tion of the conantokins, contain multiple disulfide bonds.
Unlike studies on other animal toxins (e.g., snakes, scor-
pions), both the complexity of the venom and the small
quantities available (usually micrograms) preclude in-
depth studies on the native material. Solid phase pep-
Fig. 2. LC-MS chromatogram of Conus geographus venom. Reverse-
phase HPLC/mass spectrometry profile of the crude venom from Conus
geographus collected on the Great Barrier Reef, Australia. Conotoxins
which target sodium channels (GIIIA), N-type calcium channels (GVIA),
and the NMDA receptor (conantokin G) are indicated.
PHARMACEUTICAL APPLICATIONS OF CONOTOXINS 225
tide synthesis has been the most successful approach in
providing significant quantities of these peptides for bio-
logical and structural studies. Most often this has been
achieved through synthesis of the fully reduced polypep-
tide before “folding” under oxidative conditions. Although
this approach yields the desired peptide, in many in-
stances it is present in a mixture of other “wrongly” or
partially folded isomers. The nonnative isomers differ
solely in the connectivity of their disulfide bridges and
can be difficult to separate from the native material, lead-
ing to reduced yields of pure conopeptide.
Directed Folding
Conotoxin GI is part of the α-conotoxin family and
contains 13 residues with two intramolecular disulfide
bridges. Various oxidative techniques on fully reduced
α-conotoxin GI yield mixtures of all three potential iso-
mers with the native isomer α-CTX GI(2-7;3-13) gener-
ally predominating. We recently described an on-resin
“directed-disulfide” strategy to gain access to each iso-
mer [Alewood, 1998]. This is illustrated in the directed
synthesis of the native isomer (Fig. 3). The orthogonal
protecting groups acetomidomethyl (Acm) and fluorenyl-
methyl (Fm) were chosen to allow stepwise regiospecific
disulfide formation on the resin. Chain assembly was
performed using standard Boc chemistry [Schnölzer et
al., 1992] on p-methylbenzhydrylamine resin. The Fm
group was removed and oxidised with piperidine-DMF.
Deprotection and oxidation of the Acm group by iodine
in DMF led to the formation of the second disulfide bond.
Final HF cleavage led to deprotected “crude” conotoxin
containing minor amounts of polymer. Reversed-phase
HPLC analysis confirmed that only the native isomer was
formed. The two nonnative isomers of α-conotoxin GI
were made employing a similar strategy.
Rapid Solid Phase Peptide Synthesis (SPPS)
A bottleneck in structure–function studies of the
conotoxins has been the availability of the desired mutants
within a reasonable time frame. The small number of such
studies reflects, in part, the difficulties in the synthesis and
folding of these cysteine-rich frameworks. As such there is
a pressing need to develop faster, more efficient chemistry.
In recent years, there have been efforts [Schnölzer
et al., 1992; Alewood et al., 1997] by several groups to
improve the speed and efficiency of SPPS. The intro-
duction of HBTU/in situ neutralisation chemistry has al-
lowed routine synthesis where three residues per hour
are incorporated in the growing peptide chain. The fur-
ther development of improved acylating agents such as
HATU has opened up the possibility of more rapid syn-
thetic procedures using HATU/Boc in situ neutralization
[Alewood et al., 1997]. This is illustrated by the rapid
chain assembly of the A10L mutant of PnIA conotoxin
from Conus pennaceus, which blocks the nicotinic ace-
tylcholine receptor. The conotoxin was assembled in a
little over 1 h, worked up, and oxidised to give fully folded
homogeneous material within a day.
Conotoxin Folding
Most reduced forms of native conotoxins are capable
of folding efficiently. The folding/oxidation thus remains a
matter of probing sufficient “folding” space so that the
desired conotoxin forms uniquely or as the predominant
product. Whereas many laboratories have the capacity to
isolate quantities of the reduced purified precursors, their
efforts at the “folding” stage have often been inadequate.
This may be a direct result of not having access to native
material for comparison. This is particularly important in
cases where the disulfide bond connectivity of the
conotoxin has not been unambiguously determined.
More specifically, the folding of ω-conotoxins has
caused difficulties in several laboratories where nonna-
tive isomers have formed a significant proportion of the
oxidised products. This is readily illustrated in the fold-
ing of the N-type neuronal calcium channel blocker, GVIA
(Fig. 4), where the selection of inappropriate though com-
monly used folding conditions (trace E) led exclusively
to nonnative products. Moreover, the selection of “ap-
propriate” folding conditions (trace A) yielded almost
exclusively the correctly folded native conotoxin.
STRUCTURE DETERMINATION OF
CONOTOXINS BY NMR
NMR spectroscopy is now a well-established
method for structure determination of peptides and pro-
Fig. 3. Directed folding of α-conotoxin GI.
226 ADAMS ET AL.
teins. The method relies on the measurement of a large
number of distance restraints between pairs of protons.
These restraints are used in a simulated annealing pro-
tocol to calculate a family of structures consistent with
both the input restraints and with a force-field defining
covalent geometry of atoms. Distance restraints are of-
ten supplemented with restraints on peptide backbone
and sidechain dihedral angles. The distance restraints are
derived from NOESY spectra and the dihedral restraints
from a combination of coupling constant and NOE data.
Depending on the size of the protein being studied and
the complexity of the spectra, 2D, 3D, or 4D NMR meth-
ods may be required. The higher dimensional spectra
(i.e., 3D or 4D) generally require uniform labeling of the
protein with 15N and/or 13C isotopes so that spectral over-
lap may be resolved using the additional frequency di-
mensions associated with these NMR active nuclei, as
well as the usual proton chemical shift axis.
An assumption inherent in the NMR structure de-
termination method is that the peptide or protein adopts
predominantly a single conformation in solution. For lin-
ear peptides comprising fewer than approximately 30
amino acid residues this is often not the case, with such
small peptides being extremely flexible and adopting a
myriad of conformations in solution. Thus, for these pep-
tides only qualitative conclusions can be drawn about
solution conformations. However, peptides which are
cross-linked by disulfide bonds are more restrained in
their conformations and are very suitable for quantita-
tive structure determination by NMR. As indicated
above, the conotoxins are rich in disulfide bonds and are
hence particularly amenable to conformational analysis
by NMR.
An additional advantage of conotoxins is that their
small size (generally less than 30 residues) means that
spectral overlap is generally not a problem, and 2D rather
than 3D or 4D NMR methods are sufficient for spectral
assignment and structure determination. Because isoto-
pic labeling is not required for such studies, it is in prin-
ciple possible to determine structure from native peptides
extracted from venom ducts. However, in practice the
amounts of material required (~1 mg) generally means
that it is more convenient to synthesise the conotoxins
using the methods described above.
Over the last few years we have determined the struc-
tures of more than 30 conotoxins (from all the known classes
and from novel ones as yet unreported) and are using these
structures in several drug design programs. From these
studies, and from studies by colleagues in the literature, it
has become clear that conotoxin structures fall into a lim-
ited number of families. Representatives of these struc-
tural families are summarised in Figure 5.
From these structures it can be seen that the α-
conotoxins adopt a fold such that the N- and C-termini
are brought into close proximity by the internal disul-
fide bonds, and that a short helical segment is present.
By contrast, the structures of the µ-, κ-, and ω-conotoxins
are dominated by a series of loops which are superim-
posed on a core comprising well-defined elements of sec-
ondary structure. For the µ-conotoxins these secondary
structure elements include a helical region and a β-hair-
pin, while the κ- and ω-conotoxins contain a triple-
stranded β-sheet. The conantokins have no disulfide
bonds, but adopt helical structures [Skjaerbaek et al.,
Fig. 4. Folding of ω-conotoxins GVIA under different oxidation con-
ditions.
Fig. 5. Three-dimensional structures of several classes of conotoxins
recently determined in our laboratory. The backbone folds are repre-
sented by tubular ribbons, with sidechains shown in stick form. Nitrogen
sidechain atoms are shown in blue, oxygen atoms are in red, and the
sulfur atoms of cysteine residues are in yellow.
Peptide Conc
Buffer (mM) GSH/GSSG pH
ANH
4
Oac/GnHC1 (0.33M/0.5M) 0.05 100:10 7.8
BNH
4
Oac/GnHC1 (0.33M/0.5M) 0.05 – 7.8
CNH
4
Oac/GnHC1 (0.33M/0.5M) 0.19 100:10 7.8
DNH
4
Oac/GnHC1 (0.33M/0.5M) 0.19 – 7.8
E NaOAc (50mM) 0.05 – 7.5
*Note that the buffer used in A gave the best results and buffer E was
poorest.
PHARMACEUTICAL APPLICATIONS OF CONOTOXINS 227
Figure 5.
228 ADAMS ET AL.
1997]. We return later to a more extensive discussion of
specific details of some of these structures, but emphasise
here that the conotoxins clearly may be regarded as
“mini-proteins” and adopt well-defined solution struc-
tures with all of the features of larger proteins. The de-
fined presentation of amino acids on the surface of the
frameworks in Figure 5 accounts for the specificity of
their binding interactions and the small size of the mol-
ecules makes them valuable lead compounds in drug
design applications.
Conotoxins Which Block the Nicotinic
ACh Receptor
There are several classes of ligands that bind to the
nAChR. These comprise small molecules, such as the
endogenous ligand agonist acetylcholine, small peptides,
including the lophotoxins and α-conotoxins, and large
peptide toxins isolated from snake venoms (e.g., α-
bungarotoxin). The α-conotoxins are widespread in the
venoms of cone snails and have been isolated from pis-
civorous, molluscivorous, and vermivorous species [Gray
et al., 1981; McIntosh et al., 1982, 1994; Zafaralla et al.,
1988; Myers et al., 1991; Ramilo et al., 1992; Fainzilber
et al., 1994; Martinez et al., 1995; Loughnan et al., 1998].
These toxins are valuable ligands for probing structure–
function relationships of various nAChR subtypes as they
are potent antagonists and exhibit marked selectivity
between the peripheral and neuronal forms of the recep-
tor. Typically, the α-conotoxins are 12–18 residues in
length and are characterised by the presence of two con-
served disulfide bonds and two loops in the peptide back-
bone between the cysteines. The number of amino acids
in these two intra-cysteine loops varies, giving rise to the
α3/5, α4/7, and α4/3 subclasses of α-conotoxins.
The affinity of a given α-conotoxin depends both
on the species and the subtype of nAChR present. The
first α-conotoxins discovered were found to bind to the
muscle-type nAChR (e.g., α-conotoxin GI) through a
highly selective interaction at the α–δ over the α–γ sub-
unit interface [Hann et al., 1994; Groebe et al., 1995].
Recent studies have described the isolation and
characterisation of α-conotoxins selective for neuronal
nAChRs and the molecular basis of the interaction of
these α-conotoxins with the nAChR is beginning to be
revealed. For example, α-conotoxin ImI selectively tar-
gets the homomeric α7 and α9 subtypes of neuronal
nAChR [Johnson et al., 1995], whereas α-conotoxin MII
was found to potently block α3β2 nAChRs expressed in
Xenopus oocytes with an IC50 of 0.5 nM [Cartier et al.,
1996; Harvey et al., 1997]. The structure of MII differs
significantly from the other α-conotoxins; however, the
disulfide bonding is conserved and it has the α 4/7 spac-
ing like α-conotoxins PnIA, PnIB, and EpI. α-Conotoxins
PnIA and PnIB from C. pennaceus have been reported
to have a different phylogenic specificity compared to
the other α-conotoxins, in that they block neuronal
nAChRs in molluscs [Fainzilber et al., 1994]. However,
in preliminary experiments on dissociated rat para-
sympathetic neurones, we found that α-conotoxins
PnIA[A10L] and PnIB (0.1–1 µM) inhibit the α-bun-
garotoxin-sensitive component of the ACh-evoked cur-
rent [Hogg et al., 1999], and identified EpI as a new
selective neuronal nAChR antagonist [Loughnan et al.,
1998]. We believe that the growing diversity of α-
conotoxins in terms of their selectivity, not only between
muscle and neuronal nAChR subtypes but between neu-
ronal α subunits, will provide the molecular tools needed
to probe and distinguish between neuronal nAChR sub-
types so that their distinct function(s) can be understood.
As a result of recent studies in our laboratories, X-
ray crystal structures have been determined for GI
[Guddat et al., 1996], PnIA [Hu et al., 1996], and PnIB
[Hu et al., 1997], EpI [Hu et al., 1998], ImI and SII (un-
published structures). Combined with NMR solution
structures of a similar range of α-conotoxins, they pro-
vide initial insights into the putative binding surface of
these peptides. Comparison of the published structures
of the α-conotoxins indicates that their backbones are
superimposable. This structural consensus allows us to
model differences in specificity and potency for AChRs
with differences in the position of exposed sidechains in
the α-conotoxins. From this analysis we will identify resi-
dues important for selectivity, allowing us to design new
selective α-conotoxins. To illustrate recent work from our
laboratory in this class of conotoxins, we describe stud-
ies on conotoxin GI, the first and one of the smallest α-
conotoxins to be discovered, and on MII, a recently
discovered member of the family with selectivity of neu-
ronal nAChRs. The sequence and characteristics of these
peptides are given in Table 2.
Conotoxin GI
Our interest in GI has focused on its use as a model
to explore conformational diversity resulting from disul-
fide bond engineering. As already noted, conotoxins are
characterised by their particularly high content of cys-
teine, with the cysteine residues almost invariably con-
nected in pairs to form disulfide bonds. In peptide toxins,
even more so than in larger proteins, these disulfide bonds
have a crucial bearing on three-dimensional structure and
function. As the number of cysteine residues in a pep-
tide increases, the number of ways of connecting the cys-
teines in disulfide bonds increases dramatically, leading
to a large number of potential isomers. It is interesting
and highly significant that invariably only one of the pos-
sible isomers occurs naturally, i.e., venoms do not nor-
PHARMACEUTICAL APPLICATIONS OF CONOTOXINS 229
TABLE 2. α-Conotoxin Sequences and Specificities
Loop nAChR
Name Sequence sizes Species Prey target Reference
GI ECCNPACGRHYSC-NH23:5 C. geographus fish α/δGray et al., 1981; Groebe et al., 1997
GIA ECCNPACGRHYSCGK-NH23:5 C. geographus fish Gray et al., 1981
GII ECCHPACGKHFSC-NH23:5 C. geographus fish Gray et al., 1981
MI G RCCHPACGKNYSC-NH23:5 C. magus fish α/δMcIntosh et al., 1982; Groebe et al., 1995
SIA YCCHPACGKNFDC-NH23:5 C. striatus fish Ramilo et al., 1992
SI ICCNPACGPKYSC-NH23:5 C. striatus fish α/δZafaralla et al., 1988; Groebe et al., 1995
SII GCCCNPACGPNYGCGTSCS 3:5:3 C. striatus fish Ramilo et al., 1992
ImI GCCSDPRCAWRC-NH24:3 C. imperialis worm α7 McIntosh et al., 1994
MII GCCSNPVCHLEHSNLC-NH24:7 C. magus fish α3β2 Cartier et al., 1997; Harvey et al., 1997
EI RD OCCYHPTCNMSNPQIC-NH24:7 C. ermineus fish Martinez et al., 1995
AuIA GCCSYPPCFATNSDYC-NH24:7 C. aulicus mollusc α3β4 Luo et al., 1998
AuIB GCCSYPPCFATNPDC-NH24:6 C. aulicus mollusc α3β4 Luo et al., 1998
AuIC GCCSYPPCFATNSGYC-NH24:7 C. aulicus mollusc α3β4 Luo et al., 1998
EpI GCCSDPRCNMNNPDY (SO4)C-NH24:7 C. episcopatus mollusc α3β; α3β4 Loughnan et al., 1998
PnIA* GCCSLPPCAANNPDYC-NH24:7 C. pennaceus mollusc Fainzilber et al., 1994
PnIB* GCCSLPPCALSNPDYC-NH24:7 C. pennaceus mollusc α7 Fainzilber et al., 1994
αA-PIVA GCCGSYONAACHOCSCKDROSYCGQ-NH27:2:1:6 C. purpurascens fish α/δ; α/γHopkins et al., 1995
αA-EIVA GCCGPYONAACHOCGCKVGROOYCDROSGG-NH27:2:1:7 C. ermineus fish α/δ; α/γJacobsen et al., 1997
αA-EIVB GCCGKYONAACHOCGCTVGROOYCDROSGG-NH27:2:1:7 C. ermineus fish α/δ; α/γJacobsen et al., 1997
*PnIA and PnIB have been shown to target neuronal nAChRs of molluscs and, more recently, PnIA[A10L] and PnIB have been reported to block the mammalian α7 nAChR subunit [Hogg et al., 1999].
230 ADAMS ET AL.
mally contain different isomers of the same conotoxins
with different connections of the disulfide bonds. How-
ever, using solid phase chemical methods it is possible to
selectively produce each of the individual disulfide bond
isomers. As noted in a section above, we used this ap-
proach to synthesise all three possible disulfide bond iso-
mers of the α-conotoxin GI and have determined their
structures [Gehrmann et al., 1998]. We refer to the three
isomers as GI(2-7;3-13), GI(2-13;3-7), and GI(2-3;7-13).
The structural findings may be summarised by not-
ing that the native connectivity of the four constituent
cysteine residues produces a significantly more stable and
well defined structure than either of the two alternative
arrangements of the disulfide bonds [Gehrmann et al.,
1998]. A single solution conformation was detected for
the native isomer, GI(2-7;3-13), which consists primarily
of a distorted 310 helix from residues 5 to 11. The two
nonnative forms exhibit multiple conformations in solu-
tion, with the major populated forms being different in
structure both from each other and with the native form.
We concluded that the disulfide bonds in GI play a ma-
jor role in determining both the structure and stability of
the peptide. A trend for increased conformational flex-
ibility was observed in the order GI(2-7;3-13) < GI(2-
13;3-7) < GI(2-3;7-13).
Interest in making nonnative isomers arises because
peptide analogues are widely regarded as valuable drug
leads, and in recent years there has been much effort di-
rected towards the development of peptide libraries. It
has been of particular interest to develop methods to in-
crease the surface variability of peptides because the di-
versity of peptide libraries are, to some extent, limited by
the use of the 20 natural amino acids. The study described
above shows that the use of alternative disulfide bond
connectivities provides another way of altering molecular
conformations without modifying the sequence.
Conotoxin MII
The recently identified α-conotoxin MII from C.
magus belongs to the α4/7 subclass and is a potent and
highly specific blocker of mammalian neuronal nAChRs
composed of α3β2 subunits. MII was first reported by
Cartier et al. [1996] following the electrophysiological
screening of RP-HPLC fractions of duct venom against
cloned nAChRs expressed in Xenopus oocytes. We inde-
pendently isolated and characterised MII as part of a
comprehensive study of the milked venom of C. magus
and recently reported its three-dimensional structure
[Hill et al., 1998].
The molecule folds into a highly compact globu-
lar structure consisting of a central region of α-helix
and a series of overlapping β-turns at the N- and C-
termini. The α-helix comprising residues 6–12 exhibits
two turns and is amphipathic, with Cys8, His9, Glu11,
and His12 on one side and Pro6, Val7, and Leu10 on
the other. Remarkably, the hydrophobic residues of the
α-helix are more exposed to the solvent than the
charged/hydrophilic residues. However, this is consis-
tent with the fact that MII is more hydrophobic when
oxidised than in the reduced form. Hydrophilic resi-
dues on the surface include Ser4, Asn5, and residues
Glu11, His12, Ser13, and Asn14. The latter patch, com-
prising residues with both polar and charged groups
(Glu11-Asn14), may be responsible for initial recogni-
tion by the nAChR, with further stabilisation of bind-
ing provided by the proximal hydrophobic residues.
Analysis of the solvent accessibility of individual residues
provides support for Pro6, Val7, Leu10, Glu11, and Asn
14 as potential residues for interaction with the nAChR
as they are highly solvent exposed [Hill et al., 1998].
Sodium Channel Binding Conotoxins
The piscivorous cone snail, C. geographus, produces
polypeptide neurotoxins that specifically inhibit skeletal
muscle and eel electroplax sodium channels [Sato et al.,
1983; Cruz et al., 1985; Yanagawa et al., 1988; Moczyd-
lowski et al., 1986]. These toxins, the µ-conotoxins and
conotoxin GS, are attractive probes of sodium channel
structure because of their high binding affinity and abil-
ity to discriminate between the skeletal muscle and neu-
ronal and cardiac channel isoforms [Yanagawa et al., 1988;
Moczydlowski et al., 1986; Ohizumi et al., 1986; Chen et
al., 1992]. It is remarkable that while these peptides be-
long to the same pharmacological class they have differ-
ent structural frameworks, as illustrated in Table 1.
The µ-conotoxins, a family of highly basic 22-resi-
due polypeptides (GIIIA, GIIIB, and GIIIC), contain six
cysteine residues which are paired in a 1–4, 2–5, 3–6
pattern to form three intramolecular disulfide bonds and
a three-loop framework. Conotoxin GS has a strikingly
different sequence and is 50% larger than the µ-
conotoxins. This polypeptide contains six cysteine resi-
dues arranged in a similar 1–4, 2–5, 3–6 pattern [Nakao
et al., 1995]; however, differences in the spacings between
cysteine residues results in a four-loop framework rather
than a three-loop framework. Despite the low sequence
identity, conotoxin GS binds competitively with µ-
conotoxin GIIIA, suggesting overlapping binding sites
on the extracellular surface of skeletal muscle and eel
electroplax sodium channels [Yanagawa et al., 1988].
Conotoxin GIIIB
GIIIB adopts a compact structure [Hill et al., 1996]
consisting of a distorted 310-helix, a small β-hairpin, a cis-
hydroxyproline, and several turns. The molecule is
stabilised by three disulfide bonds, two of which con-
nect the helix and the β-hairpin, forming a structural core
with similarities to the CSαβ motif [Cornet et al., 1995].
PHARMACEUTICAL APPLICATIONS OF CONOTOXINS 231
This motif is common to several families of small pro-
teins, including scorpion toxins and insect defensins.
Other structural features of GIIIB include the presence
of eight arginine and lysine sidechains that project into
the solvent in a radial orientation relative to the core of
the molecule. These cationic sidechains form potential
sites of interaction with anionic sites on sodium chan-
nels. The global fold is similar to that reported for µ-
conotoxin GIIIA, and together the structures provide a
basis for further understanding of the structure–activity
relationships of the µ-conotoxins and for their binding to
skeletal muscle sodium channels.
Conotoxin GS
The three-dimensional structure of conotoxin GS
[Hill et al., 1997] consists of a compact, disulfide-bonded
core from which several loops and the C-terminus project.
The main element of secondary structure is a double-
stranded antiparallel β-sheet comprising residues 17–20
and 26–29 connected by a turn involving residues 21–25
to give a β-hairpin structure. A further peripheral β-strand
involving residues 7–9 is almost perpendicular to the β-
hairpin, with only Ser7 hydrogen-bonded to the central
β-strand forming an isolated β-bridge.
GS is unusual in that it contains the posttrans-
lationally modified residue γ–carboxy glutamic acid. To
investigate the role of Gla32 in this polypeptide, an ana-
log [Glu32]conotoxin GS was synthesised and the NMR
spectra compared with those of conotoxin GS. The chemi-
cal shift differences for the backbone Hα and NH pro-
tons of conotoxin GS and [Glu32]conotoxin GS were small
(≤0.05 ppm), suggesting that the backbone conformation
of the two peptides is essentially identical. Several other
parameters, including the observed NOEs, 3JNH-Hα cou-
pling constants and amide exchange rates are similar,
providing further evidence of conserved structure in
these peptides. This suggests that the Gla residue does
not play a role in modulating the three-dimensional struc-
ture of conotoxin GS.
As the sequence and structure of conotoxin GS is
quite different from the µ-conotoxins, it provides a valu-
able new probe for further characterisation of sodium chan-
nel geometry. The structure of conotoxin GS will facilitate
the design of analogues to define the binding surface and
to undertake complementary mutagenesis on the sodium
channel to identify the interacting residues. These experi-
ments with conotoxins may prove as useful in modeling
the outer vestibule of sodium channels as the peptide tox-
ins from scorpions have been for potassium channels.
Calcium Channel Blocking Conotoxins
The ω-conotoxins are a set of structurally related
peptides that have a wide range of specificities for differ-
ent subtypes of the voltage-sensitive calcium channel
(VSCC). To understand their VSCC subtype differentia-
tion, we studied the structure of two naturally occurring
ω-conotoxins, MVIIA (specific to N-type VSCCs) and
SVIB (specific to P/Q-type) and a synthetic hybrid, SNX-
202, which has altered specificities to both VSCC sub-
types [Nielsen et al., 1996]. The secondary structures of
the three peptides are almost identical, consisting of a
triple-stranded β-sheet and several turns. The three-di-
mensional structures of SVIB and MVIIA are likewise
quite similar, but some subtle differences are manifested
as orientational differences between two key loops.
A remarkable feature of the six cysteine / four-loop
framework exemplified by the ω-conotoxins is the pres-
ence of a cystine knot within the structures. This motif
consists of an embedded loop in the structure formed by
two of the disulfide bonds and their connecting back-
bone segments. This loop is penetrated by the third dis-
ulfide bond in a remarkable example of Nature’s
engineering designs.
Although the structural rigidity of the core of
MVIIA is apparently assured by the knotted disulfide
structure, we used NMR to probe for possible conforma-
tional flexibility in the exposed loops. As indicated above,
it is important be aware of potential conformational
changes that might affect receptor binding. In the case
of MVIIA, the Hα shifts were found to be similar in a
range of solvents, indicating that there are no solvent-
induced changes in structure.
From the above structural studies and a large num-
ber of other studies of molecules within this family it is
apparent that the ω-conotoxins form a consensus struc-
ture despite differences in sequence and VSCC subtype
specificity. This indicates that the ω-conotoxin macrosites
for the N/P/Q-subfamily of VSCCs are related, with speci-
ficity for receptor targets being conferred by the positions
of functional sidechains on the surface of the peptides.
As mentioned earlier, the ω-conotoxins have at-
tracted the most interest for potential pharmaceutical
applications. Indeed, conotoxin MVIIA is currently in
clinical trial for the treatment of chronic pain. Structural
studies of the type described above are likely to lead to
the development of second-generation analogues which
may overcome some of the side effects of MVIIA itself.
Potassium Blocking Conotoxins
κ-PVIIA is a 27-residue polypeptide isolated from
the venom of C. purpurascens and is the first member of a
new class of conotoxins that block potassium channels. By
comparison to other ion channels of eukaryotic cell mem-
branes, voltage-sensitive potassium channels are relatively
simple and methodology has been developed for mapping
their interactions with small peptide toxins. PVIIA, there-
fore, is a valuable new probe of potassium channel struc-
ture. In a recent study, we determined the solution
232 ADAMS ET AL.
structure and mode of channel binding of PVIIA [Scanlon
et al., 1997] and this forms the basis for mapping the inter-
acting residues at the conotoxin–ion channel interface.
The three-dimensional structure of PVIIA re-
sembles the triple-stranded β-sheet / cystine knot motif
formed by a number of toxic and inhibitory peptides, in-
cluding the ω-conotoxins and conotoxin GS, as described
above. Subtle structural differences, however, predomi-
nantly in loops 2 and 4, are observed between PVIIA
and other conotoxins with similar structural frameworks.
Electrophysiological binding data suggest that PVIIA
blocks K+ channel currents by binding in a voltage-sen-
sitive manner to the external vestibule and occluding the
pore. Comparison of the electrostatic surface of PVIIA
with that of the well-characterised potassium channel
blocker charybdotoxin suggested a likely binding orien-
tation for PVIIA. Although the structure of PVIIA is con-
siderably different from that of the αK scorpion toxins, it
has a similar mechanism of channel blockade. On the basis
of a comparison of the structures of PVIIA and
charybdotoxin, we suggested that Lys 19 of PVIIA is the
residue responsible for physically occluding the pore of
the potassium channel.
Common Structural Frameworks.
From the studies described above it has become
clear that conotoxins with the six cysteine / four-loop
framework are the most abundant group of peptides iso-
lated from Conus venoms so far. This structural class en-
compasses at least five known pharmacological classes:
ω-conotoxin calcium channel blockers, δ-conotoxins
which inhibit the inactivation of sodium channels, κ-
conotoxin PVIIA which blocks potassium channels, the
sodium channel blocker conotoxin GS, and two peptides
recently found in C. marmoreus that affect both sodium
and calcium currents [Myers et al., 1993; Cruz, 1996;
Terlau et al., 1996]. The solution structures of several of
these classes have now been determined, including the
ω-conotoxins, κ-conotoxin, and GS, and all contain a
triple-stranded antiparallel β-sheet with +2x, -1 topol-
ogy and cystine knot motif common to that observed in a
number of toxic and inhibitory peptides [Pallaghy et al.,
1994; Narasimhan et al., 1994].
Thus, there are now many examples where one
structural framework is associated with different phar-
macological activities. Interestingly, the converse also
occurs; that is, the same pharmacological activity may be
associated with completely different structural frame-
works, as demonstrated, for example, with the studies
described above on the µ-conotoxins and conotoxin GS.
CONCLUSIONS
Conotoxins provide a vast library of peptides with
unique abilities to discriminate among types and sub-
types of ion channels in a manner that is unmatched by
the typical small molecule drugs which dominate the
pharmaceutical industry. In addition, cone venom pep-
tides are small and inherently stable, making them ideal
leads for peptide therapeutics, especially ion channel
therapeutics. The high structural resolution now obtained
with modern NMR spectroscopy and X-ray crystallogra-
phy provides emerging opportunities to use conotoxins
as templates for the design of smaller peptidomimetics
that incorporate the selectivity and potency of conotoxins.
Because of its selectivity and potency, ω-conotoxin
MVIIA (Ziconotide) is being developed as a drug for the
treatment of chronic pain. Conotoxins continue to be dis-
covered that define new pharmacological targets. With
improvement in methods of delivering peptides, it is an-
ticipated that conopeptides can be modified for effective
oral delivery.
ACKNOWLEDGMENTS
DJC is an Australian Research Council Professo-
rial Fellow. We thank our colleagues at the University of
Queensland who were involved in various aspects of these
studies: Michael Dooley, John Gehrmann, Justine Hill,
Kathy Nielsen, Martin Scanlon, Marion Loughnan, Trudy
Bond, Linda Thomas, Alun Jones, Denise Adams, Elka
Palant (3D Centre), Javier Cuevas, Ron Hogg, and
Michael Watson (Dept. of Physiology and Pharmacology).
REFERENCES
Alewood PF. 1998. Conotoxins as molecular templates for drug de-
sign. In: Ramage R, Epton R, editors. Peptides. Bodwin, UK: May-
flower Scientific. p 183–186.
Alewood PF, Alewood D, Miranda L, Love S, Meutermans W, Wilson
D. 1997. Rapid in situ neutralization protocols for Boc and Imoc
solid phase chemistries. Methods Enzymol 289:14–29.
Bingham J-P, Jones A, Lewis RJ, Andrews PR, Alewood PF. 1996. Co-
nus venom peptides (conopeptides): inter-species, intra-species and
within individual variation revealed by ionspray mass spectrometry.
In: Lazarovici P et al., editors. Biochemical aspects of marine phar-
macology. Alaken Inc., Colorado. p 13–27.
Cartier GE, Yoshikami D, Gray WR, Luo S, Olivera BM, McIntosh
JM. 1996. A new α-conotoxin which targets α3β2 nicotinic acetyl-
choline receptors. J Biol Chem 271:7522–7528.
Chen L-Q, Chahine M, Kallen RG, Barchi RL, Horn R. 1992. Chi-
meric study of sodium channels from rat skeletal and cardiac muscle.
FEBS Lett 309:253–257.
Colledge CJ, Hunsperger JP, Imperial JS, Hillyard DR. 1992. Precur-
sor structure of ω-conotoxin GVIA from a cDNA clone. Toxicon
30:111–1116.
Colquhoun LM, Patrick JW. 1997. Pharmacology of neuronal nico-
tinic acetylcholine receptor subtypes. Adv Pharmacol 39:191–220.
Cornet B, Bonmatin J-M, Hetru C, Hoffman JA, Ptak M, Vovelle F.
1995. Refined three-dimensional solution structure of insect defensin
A. Structure 3:435–448.
Cruz LJ. 1996. Primary structural motifs of Conus peptides. In: Singh
BR, Tu AT, editors. Natural toxins II. New York: Plenum Press. p
155–167.
PHARMACEUTICAL APPLICATIONS OF CONOTOXINS 233
Cruz LJ, Gray WR, Olivera BM, Zeikus RD, Kerr L, Yoshikami D,
Moczydlowski E. 1985. Conus geographus toxins that discriminate
between neuronal and muscle sodium channels. J Biol Chem
260:9280–9288.
Doyle DA, Cabral JM, Pfuetzner RA, Kuo A, Gulbis JM, Cohen SL,
Chait BT, MacKinnon R. 1998. The structure of the potassium chan-
nel: molecular basis of K+ conduction and selectivity. Science
280:69–77.
Fainzilber M, Hasson A, Oren R, Burlingame AL, Gordon D, Spira
ME, Zlotkin E. 1994. New mollusc-specific alpha-conotoxins block
Aplysia neuronal acetylcholine receptors. Biochemistry 33:9523–
9529.
Gray WR. 1993. Disulfide structures of highly bridged peptides: a new
strategy for analysis. Protein Sci 2:1732–1748.
Gray WR, Luque A, Olivera BM, Barrett J, Cruz LJ. 1981. Pep-
tide toxins from Conus geographus venom. J Biol Chem 256:
4734–4740.
Gehrmann J, Alewood PF, Craik DJ. 1998. Structure determination
of the three disulfide bond isomers of α-conotoxin GI: a model
for the role of disulfide bonds in structural stability. J Mol Biol
278:401–415.
Groebe DR, Dumm JM, Levitan ES, Abramson SN. 1995. α-
Conotoxins selectively inhibit one of the two acetylcholine binding
sites of nicotinic receptors. Mol Pharmacol 48:105–111.
Guddat LW, Martin JL, Shan L, Edmundson AB, Gray WR. 1996.
Three-dimensional structure of the α-conotoxin GI at 1.2 Å resolu-
tion. Biochemistry 35:11329–11335.
Hann RM, Pagon OR, Eterovic VA. 1994. The α-conotoxins GI and
MI distinguish between the nicotinic acetylcholine receptor ago-
nist sites while SI does not. Biochemistry 33:14058–14063.
Harvey SC, McIntosh JM, Cartier GE, Maddox FN, Luetje CW. 1997.
Determinants of specificity for α-conotoxin MII on α3β2 neuronal
nicotinic receptors. Mol Pharmacol 51:336–342.
Hill JM, Alewood PF, Craik DJ. 1996. Three-dimensional solution struc-
ture of µ-conotoxin GIIIB, a specific blocker of skeletal muscle so-
dium channels. Biochemistry 35:8824–8835.
Hill JM, Alewood PF, Craik DJ. 1997. Solution structure of the sodium
channel antagonist conotoxin GS: a new molecular caliper for prob-
ing sodium channel geometry. Structure 5:571–583.
Hill JM, Oomen CJ, Miranda LP, Bingham J-P, Alewood PF, Craik DJ.
1998. Three-dimensional solution structure of α-conotoxin MII by
NMR spectroscopy: effects of solution environment on helicity. Bio -
chemistry 37:15621–15630.
Hogg RC, Lewis RJ, Adams DJ. 1999. α-Conotoxin analogue,
PnIA[A10L], and α-bungarotoxin block the same component of nico-
tinic ACh-induced current in rat parasymapthetic neurons. Proc Aust
Neurosci Soc 10:152.
Hopkins C, Grilley M, Miller C, Shon K-J, Cruz LJ, Gray WR, Dykert
J, Rivier J, Yoshikami D, Olivera BM. 1995. A new family of Conus
peptides targed to the nicotinic acetylcholine receptor. J Biol Chem
270:22361–22367.
Hu S-H, Gehrmann J, Guddat LW, Alewood PF, Craik DJ, Martin JL.
1996. The 1.1 Å crystal structure of the neuronal acetylcholine re-
ceptor antagonist, α-conotoxin PnIA from Conus pennaceus. Struc-
ture 4:417–423.
Hu S-H, Gehrmann J, Alewood PF, Craik DJ, Martin JL. 1997. Crystal
structure at 1.1 Å resolution of α-conotoxin PnIB: comparison with
α-conotoxins PnIA and GI. Biochemistry 36:11323–11330.
Hu S-H, Loughnan M, Miller R, Weeks CM, Blessing RH, Alewood
PF, Lewis RJ, Martin JL. 1998. The 1.1 Å resolution crystal struc-
ture of [Tyr15]EpI, a novel α-conotoxin from Conus episcopatus,
solved by direct methods. Biochemistry 37:11425–11433.
Jacobsen R, Yoshikami D, Ellison M, Martinez J, Gray WR, Cartier
GE, Shon K-J, Groebe DR, Abramson SN, Olivera BM, McIntosh
JM. 1997. Differential targeting of nicotinic acetylcholine receptors
by novel αA-conotoxins. J Biol Chem 272:22531–22537.
Jan LY, Jan YN. 1997. Cloned potassium channels from eukaryotes
and prokaryotes. Annu Rev Neurosci 20:91–123.
Johnson DS, Martinez J, Elgoyhen AB, Heinemann SF, McIntosh JM.
1995. α-Conotoxin Im1 exhibits subtype-specific nicotinic acetyl-
choline receptor blockade: preferential inhibition of homomeric α7
and α9 receptors. Mol Pharmacol 48:194–199.
Jones A, Bingham J-P, Gehrmann J, Bond T, Loughnan M, Atkins A,
Lewis RJ, Alewood PF. 1996. Isolation and characterisation of
conopeptides by high-performance liquid chromatography com-
bined with mass spectrometry and tandem mass spectrometry. Rapid
Commun Mass Spectrom 10:138–143.
Kaiser SA, Soliakov L, Harvey SC, Luetje CW, Wonnacott S. 1998.
Differential inhibition by α-conotoxin-MII of the nicotinic stimula-
tion of [3H]dopamine release from rat striatal synaptosomes and
slices. J Neurochem 70:1069–1076.
Kulak JM, Nguyen TA, Olivera BM, McIntosh JM. 1997. α-Conotoxin
MII blocks nicotine-stimulated dopamine release in rat striatal syn-
aptosomes. J Neurosci 17:5263–5270.
Lewis RJ, Bingham J–P, Jones A, Alewood PF. 1994. Analysis of Co-
nus venoms by ionspray mass spectrometry. Australasian Biotech
4:298–300.
Loughnan M, Bond T, Atkins A, Cuevas J, Adams DJ, Broxton NM,
Livett BG, Down JG, Jones A, Alewood PF, Lewis RJ. 1998. α-
Conotoxin EpI, a novel sulfated peptide from Conus episcopatus
that selectively targets neuronal nicotinic acetylcholine receptors. J
Biol Chem 273:15667–15674.
Luo S, Kulak JM, Cartier GE, Jacobsen RB, Yoshikami D, Olivera BM,
McIntosh JM. 1998. α-Conotoxin AuIB selectively blocks α3β4 nico-
tinic acetylcholine receptors and nicotine-evoked norepinephrine
release. J Neurosci 18:8571–8579.
Martinez JS, Olivera BM, Gray WR, Craig AG, Groebe DR, Abramson
SN, McIntosh JM. 1995. α-Conotoxin EI, a new nicotinic acetyl-
choline receptor antagonist with novel selectivity. Biochemistry
34:14519–14526.
McGehee DS, Role LW. 1995. Physiological diversity of nicotinic ace-
tylcholine receptors expressed by vertebrate neurons. Annu Rev
Physiol 57:521–546.
McIntosh JM, Cruz LJ, Hunkapiller MW, Gray WR, Olivera BM. 1982.
Isolation and structure of a peptide toxin from the marine snail Co-
nus magus. Arch Biochem Biophys 218:329–334.
McIntosh JM, Yoshikami D, Mahe E, Nielsen DB, Rivier JE, Gray
WR, Olivera, BM. 1994. A nicotinic acetylcholine receptor ligand
of unique specificity, α-conotoxin ImI. J Biol Chem 269:16733–
16739.
Miller C. 1995. The charybdotoxin family of K+-channel blocking pep-
tides. Neuron 15:5–10.
Moczydlowski E, Olivera BM, Gray WR, Strichartz GR. 1986. Dis-
crimination of muscle and neuronal Na-channel subtypes by bind-
ing competition between [3H]saxitoxin and µ-conotoxins. Proc Natl
Acad Sci USA 83:5321–5325.
Myers RA, Zafaralla GC, Gray WR, Abbott J, Cruz LJ, Olivera BM.
1991. α-Conotoxins, small peptide probes of nicotinic acetylcho-
line receptors. Biochemistry 30:9370–9377.
Myers RA, Cruz LJ, Rivier JE, Olivera BM. 1993. Conus peptides as
234 ADAMS ET AL.
chemical probes for receptors and ion channels. Chem Rev 93:1923–
1936.
Nakao M, Nishiuchi Y, Nakata M, Watanabe TX, Kimura T, Sakakibara
S. 1995. Synthesis and disulfide structure determination of conotoxin
GS, a γ-carboxyglutamic acid-containing neurotoxic peptide. Lett
Pept Sci 2:17–26.
Narasimhan L, Singh J, Humblet C, Guruprasad K, Blundell T. 1994.
Snail and spider toxins share a similar tertiary structure and ‘cystine
motif.’ Nat Struct Biol 1:850–852.
Nielsen KJ, Thomas L, Lewis RJ, Alewood PF, Craik DJ. 1996. A con-
sensus structure for ω-conotoxins with different selectivities for
voltage-sensitive calcium channel subtypes: comparison of MVIIA,
SVIB and SNX-202. J Mol Biol 263:297–310.
Noda M, Shimizu S, Tanabe T, Takai T, Kayano T, Ikeda T, Takahashi
H, Nakayama H, Kanaoka Y, Minamino N, Kangawa K, Matsuo H,
Raftery MA, Hirose T, Inayama S, Hayashida H, Miyata T, Numa S.
1984. Primary structure of Electrophorus electricus sodium channel
deduced from cDNA sequence. Nature 312:121–127.
Ohizumi Y, Nakamura H, Kobayashi J, Catterall WA. 1986. Specific
inhibition of [3H]saxitoxin binding to skeletal muscle sodium chan-
nels by geographutoxin II, a polypeptide channel blocker. J Biol
Chem 261:6149–6152.
Olivera BM, Rivier J, Clark C, Ramilo C, Corpuz GP, Abogadie FC,
Mena EE, Woodward SR, Hillyard DR, Cruz LJ. 1990. Diversity of
Conus neuropeptides. Science 249:257–263.
Olivera BM, Miljanich G, Ramachandran J, Adams ME. 1994. Cal-
cium channel diversity and neurotransmitter release: the ω-
conotoxins and ω-agatoxins. Annu Rev Biochem 63:823–867.
Pallaghy PK, Nielsen KJ, Craik DJ, Norton RS. 1994. A common struc-
tural motif incorporating a cystine knot and a triple-stranded β-sheet
in toxic and inhibitory polypeptides. Protein Sci 3:1833–1839.
Picciotto MR, Zoli M, Rimondini R, Léna C, Marubio LM, Pich EM,
Fuxe K, Changeux J-P. 1998. Acetylcholine receptors containing the
β2 subunit are involved in the reinforcing properties of nicotine.
Nature 391:173–177.
Poth K, Nutter TJ, Cuevas J, Parker MJ, Adams DJ, Luetje CW. 1997.
Heterogeneity of nicotinic receptor class and subunit mRNA ex-
pression among individual parasympathetic neurons from rat intra-
cardiac ganglia. J Neurosci 17:586–596.
Ramilo CA, Zafaralla GC, Nadasdi L, Hammerland LG, Yoshikami D,
Gray WR, Kristipati R, Ramachandran J, Miljanich G, Olivera BM,
Cruz LJ. 1992. Novel α- and ω-conotoxins from Conus striatus
venom. Biochemistry 31:9919–9926.
Sangameswaran L, Delgado SG, Fish LM, Koch BD, Jakeman LB,
Stewart GR, Sze P, Hunter JC, Eglen RM, Herman RC. 1996. Struc-
ture and function of a novel voltage-gated, tetrodotoxin-resistant
sodium channel specific to sensory neurons. J Biol Chem 271:5953–
5956.
Sato S, Nakamura H, Ohizumi Y, Kobayashi J, Hirata Y. 1983. The
amino acid sequences of homologous hydroxyproline-containing
myotoxins from the marine snail Conus geographus venom. FEBS
Lett 155:277–280.
Scanlon MJ, Naranjo D, Thomas L, Alewood, PF, Lewis RJ, Craik DJ.
1997. The structure of novel potassium channel toxin, κ-conotoxin
PVIIA, from the venom of Conus purpurascens: implications for the
mechanism of channel block. Structure 5:1585–1597.
Schnölzer M, Alewood P, Jones A, Alewood D, Kent SBH. 1992. In
situ neutralization in Boc-chemistry solid phase peptide synthesis.
Int J Pept Protein Res 40:180–193.
Shon K-J, Olivera BM, Watkins M, Jacobsen RB, Gray WR,
Floresca CZ, Cruz LJ, Hillyard DR, Brink A, Terlau H, Yoshi-
kami D. 1998. µ-Conotoxin PIIIA, a new peptide for discrimi-
nating among tetrodotoxin-sensitive Na channel subtypes. J
Neurosci 18:4473–4481.
Skjaerbaek N, Nielsen KJ, Lewis RJ, Alewood P, Craik DJ. 1997. Determi-
nation of the solution structures of conantokin-G and conantokin-T by
CD and NMR spectroscopy. J Biol Chem 272:2291–2299.
Terlau H, Shon K-J, Grilley M, Stocker M, Stühmer W, Olivera BM.
1996. Strategy for rapid immobilisation of prey by a fish-hunting
marine snail. Nature 381:148–151.
Unwin N. 1998. The nicotinic acetylcholine receptor of the Torpedo
electric ray. J Struct Biol 121:181–190.
Watson MJ, Lewis RJ, Adams DJ. 1998. Effects of synthetic µ-
conotoxins on voltage-dependent sodium currents in rat sensory
neurones. Proc Aust Physiol Pharmacol Soc 29:326P.
Yanagawa Y, Abe T, Satake M, Odani S, Suzuki J, Ishikawa K. 1988. A
novel sodium channel inhibitor from Conus geographus: purifica-
tion, structure, and pharmacological properties. Biochemistry
27:6256–6262.
Zafaralla GC, Ramilo C, Gray WR, Karlstrom R, Olivera BM, Cruz LJ.
1988. Phylogenetic specificity of cholinergic ligands: α-conotoxin
SI. Biochemistry 27:7102–7105.
