the structure and mode of action of different botulinum toxins
TRANSCRIPT
ORIGINAL ARTICLE
The structure and mode of action of different botulinum toxins
J. O. Dollya and K. R. AokibaInternational Centre for Neurotherapeutics, Dublin City University, Glasnevin, Dublin, Ireland; and bNeurotoxin Research Program,
Department of Biological Sciences, Allergan Inc., Irvine, CA, USA
Keywords:
botulinum toxin, sensitive
factor attachment protein-
25, sensitive factor
attachment protein recep-
tor
Received 30 August 2006
Accepted 15 September 2006
The seven serotypes (A–G) of botulinum neurotoxin (BoNT) are proteins produced by
Clostridium botulinum and have multifunctional abilities: (i) they target cholinergic
nerve endings via binding to ecto-acceptors (ii) they undergo endocytosis/transloca-
tion and (iii) their light chains act intraneuronally to block acetylcholine release. The
fundamental process of quantal transmitter release occurs by Ca2+-regulated exocy-
tosis involving sensitive factor attachment protein-25 (SNAP-25), syntaxin and syn-
aptobrevin. Proteolytic cleavage by BoNT-A of nine amino acids from the C-terminal
of SNAP-25 disables its function, causing prolonged muscle weakness. This unique
combination of activities underlies the effectiveness of BoNT-A haemagglutinin
complex in treating human conditions resulting from hyperactivity at peripheral
cholinergic nerve endings. In vivo imaging and immunomicroscopy of murine muscles
injected with type A toxin revealed that the extended duration of action results from
the longevity of its protease, persistence of the cleaved SNAP-25 and a protracted time
course for the remodelling of treated nerve–muscle synapses. In addition, an appli-
cation in pain management has been indicated by the ability of BoNT to inhibit
neuropeptide release from nociceptors, thereby blocking central and peripheral pain
sensitization processes. The widespread cellular distribution of SNAP-25 and the
diversity of the toxin’s neuronal acceptors are being exploited for other therapeutic
applications.
Introduction
Botulinum toxin was first identified in 1897 as the
product of the anaerobic bacterium Clostridium botuli-
num and the causative agent of botulism food poisoning
[1]. The development of a purification method for the
toxin began in 1944 and led to crystallization of the type
A serotype [2]. The toxin’s ability to disrupt neuro-
transmitter release at the neuromuscular junction,
thereby reducing muscle activity, resulted in the initi-
ation of investigations on the potential clinical appli-
cations for botulinum toxin by A. Scott in 1968. These
studies culminated in Food and Drug Administration
approval being granted in 1989 for the type A toxin
complex (specifically BOTOX�, Allergan Inc., Cali-
fornia, USA) treatment of individuals with involuntary
muscle disorders, such as strabismus, blepharospasm
and hemifacial spasm. Subsequently, botulinum toxin
has been used for a wide range of conditions charac-
terized by involuntary muscle contraction. More recent
findings have shown that, in addition to its effects at the
neuromuscular junction, botulinum toxin also has
additional clinical applications as a result of its ability
to block the responses of autonomic and sensory
nerves.
The botulinum neurotoxin family
Seven immunologically distinct forms of botulinum
neurotoxin (BoNT) exist, designated as serotypes A, B,
C1, D, E, F and G [3]. All toxin serotypes are synthes-
ized as a single-chain polypeptide of molecular mass
approximately 150 kDa. When produced by the bac-
terium, the toxin molecule associates with additional
non-toxic proteins to form a range of macromolecular
complexes between 300 and 900 kDa. To gain maximum
biological activity, the 150 kDa toxin polypeptide must
be cleaved into a 100 kDa heavy chain and a 50 kDa
light chain, which remain connected by a disulphide
bond and non-covalent interactions. All toxin serotypes
act by preventing acetylcholine release at the peripheral
nerve endings, thus inducing temporary denervation
and relaxation of muscles.
All BoNT serotypes are highly specific in their
targeting of cholinergic nerves, as a result of the
C-terminal region of the toxin binding with high affinity
to ecto-acceptors on the pre-synaptic motor nerve
endings [4]. Application of radiolabelled BoNT to
nerve–muscle preparations showed discrete localization
Correspondence: J. Oliver Dolly, International Centre for Neuro-
therapeutics, Dublin City University, Glasnevin, Dublin 9, Ireland
(tel.: +353 (0) 1 700 7757; fax: +353 (0) 1 700 7758;
e-mail: [email protected]).
� 2006 EFNS 1
European Journal of Neurology 2006, 13 (Suppl. 4): 1–9
of toxin to the motor neurones [5]. After binding,
BoNT is taken up into the motor neurone, via the
process of acceptor-mediated endocytosis [6]. Following
internalization, the light chain is translocated into
the cytoplasm of the motor neurone and mediates the
toxin’s activity. BoNT selectively blocks acetylcholine
release at the neuromuscular junction, thereby abol-
ishing the motor end-plate potential [7]. This effect is
highly specific and non-cholinergic neurotransmission is
not compromised [8].
Recent observations that BoNT may prevent
migraine in some patients prompted re-evaluation of
the neuronal selectivity of its actions, as it was accepted
that the mechanism of the headache pain was inde-
pendent of any muscle contractions [3]. Subsequent
studies have identified a potential antinociceptive
mechanism for BoNT-A on sensory pain nerves [9]. The
biochemical mechanism in sensory nerves (e.g. cleavage
of a SNARE protein target) is identical to that seen
with motor neurones, with the peripheral terminal
being the site of action for both nerves. The difference
in the pharmacological effect is related to the peripheral
release of neuropeptides, which act to sensitize the pain
nerve endings, leading to increased sensitivity of the
terminal to activation [e.g. the direction of pharmaco-
logical effect is periphery to the central nervous system
(CNS)]. In contrast, the action of BoNT on the motor
nerve is to block neurotransmission from the CNS to
the muscle (e.g. direction of pharmacological effect is
CNS to the periphery). In summary, the intracellular
biochemical mechanism and target of BoNT is the same
for a motor nerve or a sensory pain nerve and is
localized to the peripheral terminal of the nerve.
The administration of BoNT has also been shown to
block acetylcholine release in post-ganglionic fibres of
the parasympathetic division of the autonomic nervous
system. This suggests possible clinical roles for BoNT in
the treatment of conditions such as achalasia, hyper-
hidrosis and hypersalivation. In vitro studies have
shown that BoNT-A blocks cholinergic excitatory
transmission in the guinea pig ileum [10] and that radio-
labelled toxin binds to cholinergic nerves in the mouse
ileum [4].
Although the effects of BoNT are mediated by dis-
ruption of acetylcholine release, early studies showed
that introduction of the toxin directly into any nerve
cell could block the release of any neurotransmitter.
These findings indicate that BoNT acts on ubiquitous
intracellular targets required for neurotransmitter
release and that the mechanism of the process is com-
mon to all tissues [8,11]. The in vivo cholinergic specif-
icity of the toxin was solely mediated by the presence of
suitable ecto-acceptors at the motor nerve ending that
mediate toxin binding. Recent reports have indicated
that the synaptic vesicle protein SV2 is responsible for
the binding of BoNT to neurones [12,13].
The molecular basis of botulinum toxinactivity at the neuromuscular junction
Characterization of the molecular basis of BoNT action
was facilitated by the finding that introduction of the
isolated toxin light chain directly into the nerve ending
was sufficient to block acetylcholine release [14]. BoNT
light chains were identified as having zinc-dependent
protease activity and to specifically target a group of
proteins known as SNAREs, involved in mediating
neurotransmitter release from the motor nerve ending
[15,16].
The release of neurotransmitter involves a complex,
multistep process in which stimulation of the nerve
leads to membrane depolarization, which activates
voltage-dependent calcium channels, effecting an influx
of calcium into the nerve terminal. A resultant
increased concentration of calcium causes synaptic
vesicles to fuse with the plasma membrane, thereby
releasing neurotransmitter at the neuromuscular junc-
tion [17]. The SNARE proteins mediate the fusion of
the synaptic vesicle with the plasma membrane and
are classified as the vesicle-associated SNAREs
(v-SNAREs) and the target membrane SNAREs
(t-SNAREs) located at the plasma membrane
[Fig. 1(a)]. The v-SNARE is a protein known as synap-
tobrevin or vesicle-associated membrane protein
(VAMP) that is attached to the synaptic vesicle mem-
brane via its C-terminal region. The t-SNAREs com-
prise two proteins both located at the plasmalemma:
synaptosomal protein with a molecular weight of
25 kDa (SNAP-25) and syntaxin. When the synaptic
vesicle approaches the plasma membrane, the three
proteins interact to form a ternary complex, in which
Syntaxin SNAP-25
PM
Transmitter-containing vesicle
C
N
N
C
N
C
Synaptobrevin
4-Helix bundle
v-SNARE
t-SNAREs
(a)
(b)
Figure 1 The site of action of botulinum toxin light chains within
the SNARE complex. (Adapted from Dolly [17].)
2 J. O. Dolly and K. R. Aoki
� 2006 EFNS European Journal of Neurology, 13 (Suppl. 4): 1–9
two domains from SNAP-25 and one domain each
from synaptobrevin and syntaxin form a four-helix
bundle [Fig. 1(b)]. This complex serves as the SNARE
and plays an essential role in exocytosis, by bringing the
vesicle membrane into close apposition with the plasma
membrane. Once the SNARE complex has formed, it
facilitates avid binding of soluble N-ethylmaleimide
SNAP, which, in turn, allows the association of N-
ethylmaleimide sensitive factor. The latter is an ATPase
enzyme and the cleavage of ATP provides the energy to
dissociate the SNARE complex and allows exocytic
processes to continue.
The light chains of the different BoNT serotypes all
act to disable the SNARE complex by selectively clea-
ving the different SNARE components. BoNT-A
cleaves SNAP-25 by removing nine amino acids from
the C-terminus, whereas BoNT-E cleaves 26 amino
acids from the C-terminus of SNAP-25. The light chains
of the other BoNT serotypes cleave syntaxin (BoNT-C1)
and synaptobrevin (BoNT-B, BoNT-D, BoNT-F and
BoNT-G) at various locations [Fig. 1(b)] [17]. Proteo-
lytic cleavage of the SNARE components serves to
inactivate the individual proteins and disrupt the func-
tioning of the SNARE complex, thereby preventing
exocytosis. However, the ability of the various BoNT
light chain serotypes to act at different cleavage sites
within the SNARE components confers subtle differ-
ences in the resulting inhibition of neurotransmitter
release and clinical applications of the toxins [18].
The clinical utility of botulinum toxinserotype A
One of the major differences seen between the effects of
the toxin serotypes is the duration of the blockade of
neurotransmitter release. The clinical muscle relaxa-
tion seen with type A BoNT complex persists for
4–6 months, which is longer than is seen with other
serotypes. BoNT-A light chain effects cleavage of nine
C-terminal amino acids from SNAP-25 resulting in a
197-amino-acid product. In an animal model, the
inhibition of neurotransmitter release by BoNT-A
persisted for approximately 28 days [19]. In contrast,
deletion of an additional 17 amino acids from SNAP-25
by BoNT-E (giving a 180-amino-acid product) results
in transient paralysis of approximately 5 days duration.
Understanding of the molecular basis of action of the
different serotypes has helped to elucidate the mecha-
nisms behind BoNT-A’s prolonged clinical effect.
A major factor in the persistence of the blockade of
neurotransmitter release by BoNT-A is the intracellular
stability of the toxin’s proteolytic activity. In an
experimental model, the treatment of chromaffin cells
with BoNT-A resulted in a blockade of exocytosis over
at least 3 weeks. In an attempt to overcome this
blockade, the gene-encoding SNAP-25 was over-
expressed in chromaffin cells but could not restore
exocytosis, indicative of persistent intracellular protease
activity [20]. Only over-expression of a mutant form of
SNAP-25, resistant to cleavage by BoNT-A light chain,
was able to rescue exocytic function. Further, pulse-
chase labelling experiments confirmed that BoNT-A
proteolytic activity persists in cerebellar granular neu-
rones for over 31 days [18]. In contrast, the proteolytic
activity of the other toxin serotypes showed shorter
durations: BoNT-C1 for 25 days, BoNT-B for 10 days,
BoNT-F for 2 days and BoNT-E for 0.8 days. Studies
have shown that the BoNT-A light chain has motifs
that promote sequestration at the plasma membrane,
which seemingly protects it from normal degradative
processes [21,22].
The BoNT-A cleavage product, SNAP-25A, is also
very stable and persists for extended periods at the
motor nerve terminal. Confocal microscopy was used to
image the motor nerve and the presence of SNAP-25Aand SNAP-25E using labelled antibodies [19]. Figure 2
images mouse nerve terminals using a motor end-plate
marker (red) and SNAP-25A/E (green), following
treatment with either BoNT-A [Fig. 2(a)] or BoNT-E
[Fig. 2(b)]. Following treatment with BoNT-A, SNAP-
25A accumulated in the motor nerve terminals and
persisted for at least 40 days after toxin administration.
In contrast, the accumulation of SNAP-25E was con-
siderably more short-lived, being apparent 2 days after
toxin administration but not at 7 days. In addition, it
has been shown that the blockade of neurotransmitter
release is maintained by an antagonistic effect mediated
by SNAP-25A on the function of full-length SNAP-25.
Expression of a gene-encoding-truncated SNAP-25
(amino acids 1–197) in chromaffin cells is sufficient to
block exocytosis, indicating that SNAP-25A is able to
antagonize the SNARE function of native SNAP-25
[20].
These properties of the BoNT-A toxin serotype make
it ideally suited for clinical use, as the toxin persists at
the plasma membrane of the motor nerve terminal
affecting long-term proteolysis of SNAP-25. Not only
does the cleavage of SNAP-25 blocks the function of
SNARE complex but also the truncated SNAP-25Aproduct is antagonistic and inhibits the function of
unproteolysed SNAP-25.
Synaptic plasticity associated with botulinumtoxin treatment
A key clinical factor when considering the action of
BoNT-A at the neuromuscular junction is that the
blockade of neurotransmitter release occurs without
Structure and mode of action of botulinum toxins 3
� 2006 EFNS European Journal of Neurology, 13 (Suppl. 4): 1–9
atrophy of the nerve endings. In vivo imaging of the
neurological changes occurring in the mouse following
application of BoNT-A has shown that affected motor
nerve endings retain viability and undergo extensive
nerve sprouting and end-plate remodelling [23].
Figure 3 images the nerve terminals in the mouse
sternomastoid muscle following intramuscular injection
of BoNT-A complex. The body of the nerve was char-
acterized by the uptake of the fluorescent vital dye
4-(4-diethyl aminostyryl)-N-methylpyridinium iodide
(green) and vesicle recycling was demonstrated by
labelling with the exo-endocytic marker N-(3-triethyl
ammonium propyl)-4-(4-(dibutylamino)styryl) pyridi-
nium dibromide (FMI-43, red). Prior to the adminis-
tration of the toxin (day 0), the green vital dye staining
and FMI-43 co-localize in the motor nerve endings,
resulting in an overall yellow colouration. At day 14,
only the green colouration was associated with the
motor nerve endings, indicating that exocytic activity
had been lost. However, by day 14, long �nerve sprouts�were apparent extending from the original motor nerve
and these stained with both the green and the red (the
exoendocytic marker) dyes. Later, acetylcholine recep-
tors clustered on the muscle membrane adjacent to
where the red dye was localized on the nerve plasma
membrane. At day 28, the original motor end plate
remained inactive, but the nerve sprouts had extended
further and were able to transmit nerve impulses,
causing muscle twitch and showing low-level recovery
of function at the neuromuscular junction. By day 63,
the nerve sprouts had stopped growing but retained
exocytic activity. At 91 days after the toxin injection,
the original nerve terminal had regained exocytic
function and the extensive network of nerve sprouts
had retracted, with these findings correlating with full
recovery of nerve-evoked muscle contraction. The
motor end-plate remodelling and formation of fun-
ctional extra-junctional synapses is thought to be
important in allowing recovery of the original motor
end plate and provides evidence of the synaptic plasti-
city associated with the neuromuscular junction.
Sensory effects of botulinum toxin
During early investigations, it was noted that admin-
istration of BoNT-A complex often resulted in a
considerable reduction in pain and that pain relief
often preceded or exceeded the clinical reduction in
muscle tone. Various pre-clinical studies have subse-
quently shown that the toxin is able to inhibit neu-
ropeptide release from primary nociceptive afferent
C fibres and possibly from lightly myelinated A delta
fibres [24]. As these nociceptors are primarily located
in the skin and the toxin is administered in the peri-
phery, a localized effect is postulated for the pain relief
mechanism.
Ove
rlay
SN
AP
-25 A
Ove
rlay
SN
AP
-25 E
(b)
(a)
Figure 2 The persistence of SNAP-25A at
the mouse motor end plate. (Reproduced
from Meunier [19].)
4 J. O. Dolly and K. R. Aoki
� 2006 EFNS European Journal of Neurology, 13 (Suppl. 4): 1–9
The proposed mechanism for action of the toxin on
nociceptors is summarized in Fig. 4 [3]. Peripheral pain
stimulus is detected by the nociceptive C fibres and the
resulting action potential transmits a signal to the CNS.
The signal provokes release of neurotransmitters such
as glutamate and peptides in the CNS, within the dorsal
root ganglion. There is most probably an antidromic
activation signal, which travels down one of the per-
ipheral branches of the sensory nerve to cause auto- or
heterosensitization. The antidromic signal initiates the
release of glutamate and peptides by the peripheral
nerve terminals, which further stimulate all nerve ter-
minals in the vicinity. This localized stimulation of
nociceptors results in peripheral sensitization and in
additional afferent signals being transmitted to the
CNS, whereby more glutamate and peptides are
released, affecting central sensitization. These processes
result in a cyclical exacerbation of the pain response,
and in a chronic state, the CNS misinterprets external
signals, such as touch and elevated temperatures, as
d0
d14
d28 Bar =10 µm
Bar =10 µm
d 0
d 63
d 91Figure 3 BoNT-A-induced nerve sprout-
ing reproduced from de Paiva [23].
Figure 4 Model of the neurological basis
of the pain response.
Structure and mode of action of botulinum toxins 5
� 2006 EFNS European Journal of Neurology, 13 (Suppl. 4): 1–9
pain. It is proposed that BoNT acts by blocking neu-
rotransmitter release at the peripheral nerves by dis-
rupting SNARE-dependent exocytic processes. The
blockade of glutamate and peptide release in the
periphery prevents the development of peripheral
sensitization. As this disrupts transmission of the
additional activation signals to the CNS, the toxin
indirectly blocks the process of central sensitization.
Various experimental models support this hypothesis
for the toxin’s pain reduction.
The formalin-induced pain model
In the formalin-induced pain model, a subcutaneous
injection of formalin into the rat footpad elicits quan-
tifiable behavioural responses indicative of pain, such as
shaking and licking the treated paw [25]. With this
model, the pain response occurred as an initial acute
phase (phase I) and a prolonged second phase (phase
II), characterized by localized inflammation and central
sensitization. The pre-treatment of the footpad with
BoNT-A complex resulted in a dose-dependent reduc-
tion in pain responses, at toxin doses too low to cause
muscle weakness [26]. This effect was only apparent
during the phase II pain response, indicating that the
toxin was not exerting a localized anaesthetic effect on
the nociceptors. Introduction of a dialysis probe into
the footpad showed that the toxin blocked the forma-
lin-induced release of glutamate during phase II of the
pain response.
In addition to limiting peripheral sensitization, there is
also evidence that type A toxin reduces subsequent CNS
sensitization. In the formalin-induced pain model, the
spinal wide dynamic range neurones show increased
activity during the phase I and II pain responses. The
phase II increase in the activity of wide dynamic range
neurones can be blocked by pre-treatment of the footpad
with BoNT-A complex [27,9]. In addition, this pre-
treatment resulted in a dose-dependent reduction in the
proportion of dorsal horn neuronal cells expressing the
c-fosgene, amarkerof second-orderneuronal stimulation.
The capsaicin model
In a second model, capsaicin was injected into the
plantar rat footpad surface and pain was assessed
according to the magnitude of pressure stimulus or
temperature elevation required to induce a paw with-
drawal response [28,29]. Enhanced sensitivity is thought
to reflect secondary thermal hyperalgesia or mechanical
allodynia. Blood flow measurement and mapping of the
wide dynamic range neurone receptive field within the
footpad can also be assessed. The injection of capsaicin
to the rat footpad affected an expansion in the wide
dynamic range neurone’s receptive field that was almost
completely blocked by pre-treatment with BoNT-A
complex [30]. This correlated with a reduced frequency
of foot withdrawal in response to pressure and an
increased latency period before foot withdrawal in
response to elevated temperature [31]. The ability of the
toxin to abolish the enhanced pressure and temperature
sensitivity following capsaicin injection or sciatic nerve
transaction has been confirmed by other investigators
[32,33]. A reduction in blood flow in the capsaicin-
treated footpad was also seen following pre-treatment
with BoNT-A complex, and occurred at doses consid-
erably lower than those required to inhibit mechanical
allodynia.
Capsaicin binds to the vanilloid/heat receptor
TRPV1 located on sensory nerve endings. The
TRPV1 receptor transduces noxious chemical and
thermal signals and stimulates the local release of
substance P. The Ferrer-Montiel laboratory [34] has
shown that TRPV1 surface expression in vitro is
increased with inflammatory mediator stimulation.
The mechanism by which TRPV1 is transferred to the
plasma membrane involves SNARE-dependent exo-
cytosis and is inhibited by BoNT-A treatment. As
the peripheral sensitization mediated by capsaicin
involves upregulation of the number of TRPV1
receptors on the surface of sensory neurones, the
ability of BoNT-A to block exocytosis may contribute
to the reduced pain responses seen in animal models.
These findings are supported by reports that patients
with intractable detrusor over activity show elevated
levels of TRPVI receptors in bladder biopsy samples
and that these levels are reduced by treatment with
BoNT-A complex [35].
The diabetic neuropathy model
The experimental diabetic neuropathy model uses
streptozocin to induce mechanical allodynia in rats,
with this model thought to act via nerves other than the
large myelinated afferent fibres [36–38]. Once allodynia
is established, determined by a paw withdrawal
threshold of less than 6 g, the ability of type A toxin to
reverse the effect is determined. The administration of
BoNT-A reverses this established allodynia in the ipsi-
lateral but not in the contralateral paw (Fig. 5) [31,38].
In contrast, administration of gabapentin reversed
allodynia in both paws and required daily treatments
for a sustained effect.
Clinical pain models
A number of human volunteer studies have considered
the use of BoNT in pain management, but with
6 J. O. Dolly and K. R. Aoki
� 2006 EFNS European Journal of Neurology, 13 (Suppl. 4): 1–9
conflicting results [39–43]. Careful consideration needs
to be given as to how such studies are conducted to
optimize the timing of (i) the challenge event; (ii) toxin
treatment; (iii) the dose employed; (iv) the pain model
used and (v) the coincidence of pain stimuli and toxin
treatment areas. Negative results in a human clinical
pain model may not represent the toxin’s effectiveness
in a patient if (i) the treated areas and the �pain� area do
not coincide; (ii) the peripheral pain source is predom-
inately driven by the sensory neurone and (iii) the
timing between the toxin treatment and the stimulus are
inappropriate.
Conclusions
Botulinum neurotoxin is a potent neurotoxin that
inhibits neurotransmitter release by disrupting the
functioning of the SNARE complex, required for exo-
cytosis of synaptic vesicles. The different BoNT sero-
types exhibit differences in their toxic properties,
depending on how they disrupt SNARE complex for-
mation. The characteristics of BoNT-A mean that it is
ideally suited to clinical use, as it exerts a long-acting
blockade of acetylcholine release at the neuromuscular
junction. This, in turn, effects long-lasting muscle
paralysis, over a period of 4–6 months in humans. This
optimal blockade of vesicle exocytosis is the result of
the multifunctional activity of the toxin, which persists
for extended periods at the neuronal plasma membrane,
inactivates SNAP-25, thereby disabling the exocytosis
machinery and producing a truncated form of SNAP-
25A that inhibits the formation of functional SNARE
complexes.
The ability of BoNT-A to inhibit neuromuscular
transmission is well characterized; however, a role for
the toxin in blocking neurotransmitter release by
sensory neurones is increasingly being recognized.
Pre-clinical studies have shown BoNT to inhibit glu-
tamate and neuropeptide release in the periphery,
which disrupts both peripheral and central sensitiza-
tion in response to chronic pain stimuli. The
molecular basis of the pain-inhibiting properties of
type A toxin is in common with the mechanism by
which the toxin affects muscle relaxation. As yet the
clinical relevance of these pre-clinical results remains
to be established, but the results tie in with obser-
vations concerning pain relief.
Further understanding of the molecular basis of the
action of both BoNT-A and other toxin serotypes is
likely to unveil novel clinical applications for this family
of molecules.
Acknowledgement
The production of this article was supported by an
unrestricted educational grant from Allergan.
Declaration of interest
Prof. J. Oliver Dolly received an honorarium from
Ashley Communications for lecturing on botulinum
toxin and has a research contract from Allergan Inc. for
basic science studies on botulinum toxins. Dr K. Roger
Aoki is an employee of Allergan Inc., which manufac-
tures and sells BOTOX�.
References
1. Schantz EJ, Johnson EA. Botulinum toxin: the storyof its development for the treatment of human disease.Perspectives in Biology and Medicine 1997; 40: 317–327.
2. Schantz EJ, Johnson EA. Properties and use of botulinumtoxin and other microbial neurotoxins in medicine.Microbiological Reviews 1992; 56: 80–99.
Figure 5 Inhibition of established allody-
nia in the diabetic neuropathy model.
Structure and mode of action of botulinum toxins 7
� 2006 EFNS European Journal of Neurology, 13 (Suppl. 4): 1–9
3. Aoki KR. Evidence for antinociceptive activity of botul-inum toxin type A in pain management. Headache 2003;43(Suppl. 1): S9–S15.
4. Black JD, Dolly JO. Selective location of acceptors forbotulinum neurotoxin A in the central and peripheralnervous systems. Neuroscience 1987; 23: 767–779.
5. Dolly JO, Black J, Williams RS, Melling J. Acceptorsfor botulinum neurotoxin reside on motor nerve terminalsand mediate its internalization. Nature 1984; 307: 457–460.
6. Black JD, Dolly JO. Interaction of 125I-labeled botuli-num neurotoxins with nerve terminals. II. Autoradio-graphic evidence for its uptake into motor nerves byacceptor-mediated endocytosis. Journal of Cell Biology1986; 103: 535–544.
7. Dolly JO, Lande S, Wray DW. The effects of in vitroapplication of purified botulinum neurotoxin at mousemotor nerve terminals. Journal of Physiology 1987; 386:475–484.
8. Poulain B, Tauc L, Maisey EA, Wadsworth JD, MohanPM, Dolly JO. Neurotransmitter release is blockedintracellularly by botulinum neurotoxin, and this requiresuptake of both toxin polypeptides by a process mediatedby the larger chain. Proceedings of the National Academyof Sciences of the United States of America 1988; 85: 4090–4094.
9. Aoki KR. Review of a proposed mechanism for theantinociceptive action of botulinum toxin type A.Neurotoxicology 2005; 26: 785–793.
10. MacKenzie I, Burnstock G, Dolly JO. The effects ofpurified botulinum neurotoxin type A on cholinergic,adrenergic and non-adrenergic, atropine-resistant auto-nomic neuromuscular transmission. Neuroscience 1982; 7:997–1006.
11. Ashton AC, Dolly JO. Characterization of the inhibitoryaction of botulinum neurotoxin type A on the release ofseveral transmitters from rat cerebrocortical synapto-somes. Journal of Neurochemistry 1988; 50: 1808–1816.
12. Dong M, Yeh F, Tepp WH et al. SV2 is the proteinreceptor for botulinum neurotoxin A. Science 2006; 312:592–596.
13. Mahrhold S, Rummel A, Bigalke H, Davletov B, Binz T.The synaptic vesicle protein 2C mediates the uptake ofbotulinum neurotoxin A into phrenic nerves. FEBS Let-ters 2006; 580: 2011–2014.
14. de Paiva A, Dolly JO. Light chain of botulinum neuro-toxin is active in mammalian motor nerve terminals whendelivered via liposomes. FEBS Letters 1990; 277: 171–174.
15. Schiavo G, Benfenati F, Poulain B et al. Tetanus andbotulinum-B neurotoxins block neurotransmitter releaseby proteolytic cleavage of synaptobrevin. Nature 1992;359: 832–835.
16. Schiavo G, Rossetto O, Santucci A, DasGupta BR,Montecucco C. Botulinum neurotoxins are zinc proteins.The Journal of Biological Chemistry 1992; 267: 23479–23483.
17. Dolly O. Synaptic transmission: inhibition of neuro-transmitter release by botulinum toxins. Headache 2003;43(Suppl. 1): S16–S24.
18. Foran PG, Mohammed N, Lisk GO et al. Evaluation ofthe therapeutic usefulness of botulinum neurotoxin B, C1,E and F compared with the long lasting type A. Basis fordistinct durations of inhibition of exocytosis in centralneurons. The Journal of Biological Chemistry 2003; 278:1363–1371.
19. Meunier FA, Lisk G, Sesardic D, Dolly JO. Dynamics ofmotor nerve terminal remodeling unveiled using SNARE-cleaving botulinum toxins: the extent and duration aredictated by the sites of SNAP-25 truncation. Molecularand Cellular Neurosciences 2003; 22: 454–466.
20. O’Sullivan GA, Mohammed N, Foran PG, LawrenceGW, Oliver Dolly J. Rescue of exocytosis in botulinumtoxin A-poisoned chromaffin cells by expression of clea-vage-resistant SNAP-25. Identification of the minimalessential C-terminal residues. The Journal of BiologicalChemistry 1999; 274: 36897–36904.
21. Fernandez-Salas E, Ho H, Garay P, Steward LE, AokiKR. Is the light chain subcellular localization animportant factor in botulinum toxin duration of action?Movement Disorders 2004; 19(Suppl. 8): S23–S34.
22. Fernandez-Salas E, Steward LE, Ho H et al. Plasmamembrane localization signals in the light chain of bot-ulinum neurotoxin. Proceedings of the National Academyof Sciences of the United States of America 2004; 101:
3208–3213.23. de Paiva A, Meunier FA, Molgo J, Aoki KR, Dolly JO.
Functional repair of motor endplates after botulinumneurotoxin type A poisoning: biphasic switch of synapticactivity between nerve sprouts and their parent terminals.Proceedings of the National Academy of Sciences of theUnited States of America 1999; 96: 3200–3205.
24. Aoki KR. Pharmacology and immunology of botulinumtoxin serotypes. Journal of Neurology 2001; 248(Suppl. 1):3–10.
25. Wheeler-Aceto H, Porreca F, Cowan A. The rat pawformalin test: comparison of noxious agents. Pain 1990;40: 229–238.
26. Cui M, Khanijou S, Rubino J, Aoki KR. Subcutaneousadministration of botulinum toxin A reduces formalin-induced pain. Pain 2004; 107: 125–133.
27. Cui M, Li Z, You S, Khanijou S, Aoki KR. Mechanismsof the antinociceptive effect of subcutaneous Botox:inhibition of peripheral and central nociceptive process-ing. Naunyn Schmiedeberg’s Archives of Pharmacology2002; 365(Suppl. 2): R17.
28. Wu J, Fang L, Lin Q, Willis WD. Nitric oxide synthase inspinal cord central sensitization following intradermalinjection of capsaicin. Pain 2001; 94: 47–58.
29. Hargreaves K, Dubner R, Brown F, Flores C, Joris J. Anew and sensitive method for measuring thermal noci-ception in cutaneous hyperalgesia. Pain 1988; 32: 77–88.
30. Cui M, Li Z, Khanijou S, Rubino J, You S, Aoki R. Sub-cutaneous administration of BOTOX� inhibits capsaicin-induced thermal hyperalgesia and expansion of hornneuronal receptive field area.Society ofNeurosciencePosterF101. Program No. 812.13. 2003.
31. Francis J, You S, Satorius A et al. Analgesic propertiesand mechanism of action of botulinum toxin type A(BOTOX�). Toxins Meeting (Online) 2005, Abstract 30.
32. Bach-Rojecky L, Lackovic Z. Antinociceptive effect ofbotulinum toxin type a in rat model of carrageenan andcapsaicin induced pain. Croatian Medical Journal 2005;46: 201–208.
33. Bach-Rojecky L, Relja M, Lackovic Z. Botulinum toxintype A in experimental neuropathic pain. Journal ofNeural Transmission 2005; 112: 215–219.
34. Morenilla-Palao C, Planells-Cases R, Garcia-Sanz N,Ferrer-Montiel A. Regulated exocytosis contributes toprotein kinase C potentiation of vanilloid receptor
8 J. O. Dolly and K. R. Aoki
� 2006 EFNS European Journal of Neurology, 13 (Suppl. 4): 1–9
activity. The Journal of Biological Chemistry 2004; 279:25665–25672.
35. Apostolidis A, Popat R, Yiangou Y et al. Decreasedsensory receptors P2X3 and TRPV1 in suburothelialnerve fibers following intradetrusor injections of botuli-num toxin for human detrusor overactivity. Journal ofUrology 2005; 174: 977–983.
36. Dixon WJ. Efficient analysis of experimental observa-tions. Annual Review of Pharmacological Toxicology 1980;20: 441–462.
37. Chaplan SR, Bach FW, Pogrel JW, Chung JM, YakshTL. Quantitative assessment of tactile allodynia in the ratpaw. Journal of Neuroscience Methods 1994; 53: 55–63.
38. You S, Satorius A, Ardila C et al. Subcutaneous botuli-num neurotoxin type A inhibits streptozotocin (STZ)-induced chronic pain in a rat model of peripheral diabeticneuropathy. Society of Neuroscience 2005, Abstract514.28.
39. Voller B, Sycha T, Gustorff B et al. A randomized, dou-ble-blind, placebo controlled study on analgesic effects ofbotulinum toxin A. Neurology 2003; 61: 940–944.
40. Blersch W, Schulte-Mattler WJ, Przywara S, May A,Bigalke H, Wohlfarth K. Botulinum toxin A and thecutaneous nociception in humans: a prospective, double-blind, placebo-controlled, randomized study. Journal ofthe Neurological Sciences 2002; 205: 59–63.
41. Kramer HH, Schmidt K, Leis S, Schmelz M, Sommer C,Birklein F. Angiotensin converting enzyme has an inhib-itory role in CGRP metabolism in human skin. Peptides2006; 27(4):917–920.
42. Sycha T, Samal D, Chizh B et al. A lack of antinocicep-tive or antiinflammatory effect of botulinum toxin A in aninflammatory human pain model. Anesthesia and Anal-gesia 2006; 102: 509–516.
43. Gazerani P, Staahl C, Drewes AM, Arendt-Nielsen L. Theeffects of botulinum toxin type A on capsaicin-evokedpain, flare, and secondary hyperalgesia in an experimentalhuman model of trigeminal sensitisation. Pain 2006; 122:315–325.
Structure and mode of action of botulinum toxins 9
� 2006 EFNS European Journal of Neurology, 13 (Suppl. 4): 1–9