Content uploaded by Cesare Montecucco
Author content
All content in this area was uploaded by Cesare Montecucco on Nov 14, 2014
Content may be subject to copyright.
Clostridium is a genus of sporulating and anaerobic
Gram-positive, rod-shaped bacteria that includes more
than 150 species. These bacteria are widely distributed
in the environment and in anaerobic regions of the
intestines of several animals, where they are typically
found as spores, which are resistant to physical and
chemical stresses and can persist for long periods of
time until favourable conditions enable germination1,2.
Under appropriate environmental conditions (such as
humidity, nutrients and the absence of oxygen), spores
germinate into vegetative cells; conversely, exposure to
oxygen, as well as water and nutrient deprivation, trig-
ger sporulation. Several clostridia, including Clostridium
difficile, Clostridiumper fringens and Clostridiumsordelli,
are pathogenic, owing to the release of protein toxins,
but only a few species are neurotoxigenic. For example,
Clostridium tetani produces tetanus neurotoxin, which
blocks neurotransmitter release in spinal cord inter-
neurons and causes the spastic paralysis of tetanus3. In
addition, six phylogenetically distinct clostridia produce
more than 40 different botulinum neurotoxins (BoNTs)
(BOX1). BoNTs consist of three primary domains: two
of these domains enable binding to nerve terminals and
translocation of the toxin into the neuronal cytosol, and
the third domain comprises a metalloprotease that inhib-
its the release of neurotransmitter by peripheral nerve
terminals (BOX2), which causes the flaccid paralysis and
autonomic dysfunctions that are typical of botulism2,4.
The neurospecificity and toxic potency of BoNTs make
them the most powerful known toxins, and they are
potential bioterrorism weapons5,6. By contrast, their
absolute neurospecificity has enabled BoNTs to be used
as effective therapeutic agents for human diseases that
are characterized by hyperfunctioning nerve terminals,
as the local injection of minute amounts of these toxins
counteracts hyperactivity of the nerve terminal7.
In the past few years, major advances in our under-
standing of the structures and mechanism of action of
BoNTs have been made. In this Review, we summarize
the life cycle of Clostridiumbotulinum in humans and
animals and discuss the recent structural and mecha-
nistic studies that have advanced our understanding of
BoNT entry into neurons, trafficking in nerve cells and
intoxication.
Botulism
Botulism mostly affects wild and domesticated animals,
and outbreaks of animal botulism can spread rapidly,
leading to the intoxication of hundreds of thousands of
animals in just a few days8. Outbreaks typically occur in
environments that contain C.botulinum spores, which
can germinate in decomposing organic material under
anaerobiosis. Environmental conditions that favour
botulism outbreaks include warm temperatures, shal-
low alkaline waters that contain abundant invertebrate
populations, and decomposing vertebrate carcasses8.
Toxigenic clostridial strains are responsible for out-
breaks, but horizontal gene transfer to non-toxigenic
strains can occur, including the transfer of toxin-
encodin g loci, which causes non-toxigenic strains to
become toxigenic9. The life cycle of toxigenic clostridia
in wildlife begins with the growth of vegetative cells in
decomposing organic material and the release of BoNTs
via autolysis. The infected organic material is ingested
by BoNT-insensitive invertebrates such as worms, mus-
sels and larvae. These invertebrates are consumed by
Neurotransmitter
An endogenous chemical that
transmits signals across a
synapse from a neuron to a
postsynaptic cell.
Metalloprotease
A proteolytic enzyme that is
defined by the presence of an
essential active-site metal ion,
which is most often zinc.
Botulinum neurotoxins: genetic,
structural and mechanistic insights
Ornella Rossetto1,2*, Marco Pirazzini1,2* and Cesare Montecucco1,2
Abstract | Botulinum neurotoxins (BoNTs) are produced by anaerobic bacteria of the genus
Clostridium and cause a persistent paralysis of peripheral nerve terminals, which is known as
botulism. Neurotoxigenic clostridia belong to six phylogenetically distinct groups and
produce more than 40 different BoNT types, which inactivate neurotransmitter release
owing to their metalloprotease activity. In this Review, we discuss recent studies that have
improved our understanding of the genetics and structure of BoNT complexes. We also
describe recent insights into the mechanisms of BoNT entry into the general circulation,
neuronal binding, membrane translocation and neuroparalysis.
1Department of Biomedical
Sciences, University of
Padova.
2National Research Council
Institute of Neuroscience,
University of Padova,
Via Ugo Bassi 58/B,
35131 Padova, Italy.
*These authors contributed
equally to this work.
Correspondence to C.M.
e‑mail: cesare.montecucco@
gmail.com
doi:10.1038/nrmicro3295
Published online 30 June 2104
Nature Reviews Microbiology
|
AOP, published online 30 June 2014; doi:10.1038/nrmicro3295 REVIEWS
NATURE REV IEWS
|
MICROBIOLOGY ADVANCE ONLINE PUBLICATION
|
1
© 2014 Macmillan Publishers Limited. All rights reserved
Nature Reviews | Microbiology
BoNT/F
BoNT/A
BoNT/G
F1
A1
A5
A9
A6
A10
A2
A7
A3
A4
F4
F2
F3
F6
F5
F7
DC
D
C
CD BoNT/C
BoNT/D
E1
E2
E3
E7
E8 BoNT/E
E4
E6
E5
E9
BoNT/B
B1
B5
B2
B3
B6
B8
B7
B4
Mouse lethal dose
Corresponds to the toxin dose
that is required to kill 50% of
exposed mice; it is usually
expressed as the median lethal
dose (LD50). The mouse LD50 of
BoNTs is 0.1–1 ng per kg for
subtype 1 of the seven serotypes.
fish, birds and other terrestrial or aquatic vertebrate
animals that are sensitive to the toxin and become para-
lysed, eventually resulting in death8,10. Higher tempera-
tures and the presence of vertebrate carcasses favour
the deposition of insect eggs on the cadavers, and a
bird–maggot or fish–maggot cycle for the propagation
of botulism is established (FIG.1a). Vertebrate animals
feed on the BoNT-containing maggots that are present
on intoxicated cadavers and become paralysed. Thus, a
new cycle of transmission begins, and such cycles are
amplified by the death of an increasing number of ani-
mals. Importantly, the ingestion of only a small amount
of toxin (that is, much less than the mouse lethal dose that
has been determined in the laboratory; see below) can
Six phylogenetically distinct clostridia (Clostridium
botulinum groups I‑IV and some strains ofClostridium
butyricum and Clostridiumbaratii) produce seven
serotypically distinct botulinum neurotoxins (BoNTs)
(serotypes A–G; see the table). Recent data suggest that
there is an eighth serotype (known as BoNT/H), but this
requires experimental validation16. Each toxin serotype is
categorized into various subtypes on the basis of their
amino acid sequences (see the table). BoNT serotypes
Cand D are closely related to each other, as are serotypes
B and G, and E and F (see the figure). Although most strains
of C.botulinum express a single toxin serotype, some
isolates produce more than one serotype; for example,
some proteolytic C.botulinum group I isolates produce a
mixture of Ab, Af, Ba and Bf subtypes9,133–135, and several
C.botulinum strains that cause food-borne and infant
botulism, and produce type A toxins, have been found to
also encode a silent bontb gene and are denoted A(B). This
finding suggests that the toxin types in these strains are
actively evolving. Furthermore, a C.botulinum strain that
expresses three bont genes (chromosomally encoded
bonta2 and bontf4 genes and a plasmid-borne bontf5 gene)
has recently been described136. BoNT/DC and BoNT/CD
are mosaic toxins; BoNT/DC comprises the L chain and HN
domain of serotype D and the HC domain of serotype C,
whereas BoNT/CD consists of the L chain and HN domain
of serotype C and the HC domain of serotype D.
The number of subtypes has grown in recent years,
owing to the increased use of whole-genome sequencing
and mass spectrometry, as well as the availability of
high-affinity monoclonal antibodies. This has revealed
the striking variety of distinct BoNT subtypes that are
produced by clostridial species (see the table), in
contrast to the production of only one tetanus neurotoxin
by Clostridium tetani (although a comparable genomic effort has not been
made for this species). Notably, C.botulinum strains from groups I and II
produce the widest range of neurotoxins, whereas group III and the other
three clostridial species produce substantially fewer toxin types. We
propose that this difference might result from differences in the usual
physiological state of bacteria in the environment. Specifically, bacteria
that produce only one or a few toxin types might be present in the
environment mostly as spores (in this case, C.tetani, C.botulinum strains
from group III and the three other clostridial species). As spores are in a
non-replicative state, these bacteria have less opportunity to evolve
compared with strains that are mostly in a vegetative state (such as those
from C.botulinum groups I and II). That is, the number of replicating cells
and the ratio of spores/vegetative cells are key factors that could possibly
determine the ‘evolvability’ of bont genes. However, this hypothesis is
speculative and requires experimental validation. Oxygen tension is one
of the main parameters that control the ratio of spores/vegetative cells.
The hypothesis would then predict that C.tetani, C.botulinum group III
and Clostridiumargentinense are strict anaerobes, whereas C.botulinum
group I and II strains are capable of growth in micro-aerophilic conditions.
Given the general interest of explaining the basis of the evolvability of
bont genes, it will be interesting to test this hypothesis. The data on BoNT
subtypes that is presented in the table include the BoNTs whose
sequences are available.
Box 1 | Serotypes and subtypes of BoNTs produced by different classes of neurotoxigenic Clostridium spp.
Clostridial
species
Proteolytic C.botulinum
group I
Non-proteolytic
C.botulinumgroup II
C.botulinum
group III
C.argentinense
(group IV)
C. butyricum C. baratii
Type A; B; F; (H)* B; E; F C; D G E F
Subtype A1; A2; A3; A4; A5; A6; A7; A8;
A9; A10; B1; B2; B3; B5 (Ba); B6;
B7; A(B); Ab; Af; Af84; Bf; F1; F2;
F3; F4; F5
B4; E1; E2; E3; E6; E7; E8;
E9; E10; E11; F6 C; D; CD; DC E4; E5 F7
*Serotype H has been proposed as a novel serotype, but this remains to be experimentally verified.
REVIEWS
2
|
ADVANCE ONLINE PUBLICATION www.nature.com/reviews/micro
© 2014 Macmillan Publishers Limited. All rights reserved
Nature Reviews | Microbiology
VAMP
Early endosome
Reserve pool
Neurotransmitter
uptake
Direct recycling
Endocytosis
Fission
Fusion
Uncoating
Hemifusion
intermediate
Priming
Complexin
Ca2+ channel Ca2+
Docking
Syt
Neurotransmitter
Clathrin
Presynaptic
nerve terminal
Synaptic cle
Postsynaptic cell
SNAP25
Active zone Neurotransmitter release
Munc18
Syntaxin
ATP
H+
ADP
Box 2 | Neurotransmission at synapses
Signalling at chemical synapses is mediated by neurotransmitters, which are released from the presynaptic nerve
terminal and bind to receptors that are located on the postsynaptic cell (such as muscle or exocrine cells; see the figure).
Neurotransmitters are synthesized in the neuronal cytosol and are stored in the presynaptic nerve terminal inside small
synaptic vesicles137. The accumulation of neurotransmitters in the lumen of synaptic vesicles is mainly driven by the
electrochemical proton gradient that is generated by the vesicular ATPase proton pump, which is located in the synaptic
vesicle membrane and pumps protons into the synaptic vesicle using the energy that is released by ATP hydrolysis. The
synaptic vesicles form a reserve pool of neurotransmitters within the nerve terminal or bind to specialized sites of the
presynaptic membrane that are known as active zones138,139 , in a process known as docking115,140. The large set of proteins
that regulate synaptic vesicle docking115,137,140 are not depicted in the figure for simplicity. Two synaptic vesicle
integral membrane proteins, VAMP (also known as synaptobrevin) and synaptotagmin (Syt); two proteins in the
presynaptic membrane, SNAP25 and syntaxin; and cytosolic proteins, including complexin and Munc18, are involved in
the subsequent step, which is known as priming and enables the synaptic vesicle to fuse rapidly with the presynaptic
membrane in response to Ca2+ influx (see the figure). Syt interacts with presynaptic membrane inositol phospholipids,
whereas VAMP forms a coiled-coil complex with SNAP25 and syntaxin, which is known as the SNARE (soluble
N-ethylmaleimide-sensitive factor attachment protein receptor) complex, in a process that is regulated by Munc18 and
other proteins. After docking, fusion is prevented by complexin, which functions as a brake and, together with Munc18,
also promotes the assembly of several SNARE complexes to form a radial super-SNARE complex116,120. This is the core
ofthe nanomachine that mediates neurotransmitter release. The carboxy terminus of SNAP25 has an essential role in
protein–protein interactions between the SNARE complexes within the super-SNARE complex116,120. It is likely that the
synaptic vesicle and presynaptic membrane are hemifused in the primed state141, which would account for the ultrafast
(milliseconds or less) release of neurotransmitter at the neuromuscular junction (NMJ)140. Depolarization of the
nerveterminal results in the opening of the Ca2+ channels and the influx of Ca2+ ions that induce release of the primed
synaptic vesicle following binding to Syt. Such binding triggers a rapid conformational change that leads to
completesynaptic vesicle–presynaptic membrane fusion and the formation of a pore through which neurotransmitter
isreleased into the synaptic cleft (see the figure). Neurotransmitter diffuses out of the nerve terminal and binds to a
postsynaptic receptor, which triggers signalling in the postsynaptic cell. In the case of the NMJ, acetylcholine is released
and binds to the acetylcholine receptor, which results in depolarization of the muscle plasma membrane, leading to Ca2+
entry and muscle contraction. During neurotransmitter release, the lumen of the synaptic vesicle is transiently opened to
the outside, but itis later internalized into the nerve terminal by endocytosis70,137. The exocytosis and endocytosis of
neurotransmitter are strictly coupled: inhibition of one process leads to inhibition of the other70. Most endocytosis of
synaptic vesicles atthe NMJ is mediated by a clathrin coat. After internalization and uncoating, the synaptic vesicle is
refilled with neuro trans mitter and the next cycle of neurotransmission begins.
REVIEWS
NATURE REV IEWS
|
MICROBIOLOGY ADVANCE ONLINE PUBLICATION
|
3
© 2014 Macmillan Publishers Limited. All rights reserved
Cholinergic nerve terminals
Axonal terminals that use
acetylcholine for
neurotransmission.
lead to physiological dysfunctions that make the intoxi-
cated animal unfit; it therefore becomes either prey or a
cadaver in the wild. For example, only a minimal amount
of BoNT is required to cause visual impairment, which
reduces survival in the wild but not in the laboratory.
Furthermore, several vertebrates can carry neurotoxi-
genic C.botulinum as part of their intestinal microbiota,
which can invade the cadaver post-mortem. Thus, the
consumption of contaminated carcasses is particularly
relevant on farms, as it facilitates the spread of botulism
among livestock8.
Human botulism is much rarer than animal botu-
lism and is mostly caused by BoNT type A (BoNT/A),
BoNT/B, BoNT/E, and rarely by BoNT/F2,8 (BOX1). There
are five different forms of the disease, which are classi-
fied according to the route of entry of the toxin (FIG.1b).
Food-borne botulism occurs after the ingestion of
BoNT-contaminated food (typically canned food that
contains the pre-formed toxin)2, and the toxin must sur-
vive the proteolytic environment of the gastrointestinal
tract to reach the intestines, where it is absorbed. Simi-
larly, infant botulism is typically caused by the consump-
tion of food that is contaminated with neurotoxigenic
spores that germinate in the intestine11,12. The coloni-
zation of infants is facilitated by a lack of competition
from the resident microbiota, as infants tend to have a
less robust bowel microbiota compared with adults2,12.
In this case, BoNTs are produced and released in the
intestines for prolonged periods of time unless the infant
is treated with antibiotics. Although rare, an adult form
of infant botulism has been documented in individuals
that have anatomical or functional bowel abnormalities,
or following antibiotic therapy, both of which might
protect clostridial species from being outcompeted by
the intestinal microbiota2,4. In food-borne and infant
botulism, BoNTs cross the intestinal mucus layer and
the polarized intestinal epithelial monolayer from the
apical to the basolateral side to reach the general circu-
lation2,13. BoNTs then reach peripheral cholinergic nerve
terminals and paralyse the nerve terminals, which causes
botulism (FIG.1b). Wound botulism results from tissue
contamination with spores and is almost exclusively
associated with injection drug users14. Iatrogenic botu-
lism occurs as a result of excessive exposure to BoNTs for
cosmetic or therapeutic purposes15. In wound and iatro-
genic botulism, BoNTs bypass intestinal absorption and
directly enter the general circulation. It should be noted
that BoNTs are much more toxic when injected (the LD50
ranges from 0.1 ng per kg to 1 ng per kg in laboratory
mice) than when administered orally (which is >100–
1000 times less toxic)8. Finally, in inhalational botulism,
the toxin enters via the respiratory tract; however, deliv-
ery via aerosols is inefficient2,6. Food-borne and infant
botulism are the predominant forms of the disease in
humans, and the other forms are rarely encountered.
Diversity and structures of BoNTs
As summarized in BOX1, six phylogenetically dis-
tinct clostridial groups (C. botulinum groups I–III,
Clostridiumargentinense and some strains of Clostridiu m
baratii and Clostridium butyricum) produce seven
serotypically distinct BoNTs (which are denoted
BoNT/A–BoNT/G)1,9. An additional serotype (known
as BoNT/H) has been proposed, but its confirmation
as a novel toxin serotype requires further experimental
validation16. BoNT serotypes are divided into subtypes
on the basis of their amino acid sequences (BOX1). The
bont genes are encoded by mobile genetic elements that
enable horizontal transfer among different isolates,
which is thought to contribute to evolution of the bont
loci and thereby to the large number of distinct BoNTs
that are currently known1,9 (BOX1).
BoNT proteins are initially synthesized as single
polypeptide chains of ~150 kDa, which are cleaved by
proteases at a loop that is formed by a disulphide bond
to yield the mature toxin, which consists of a light chain
(L chain; which is 50 kDa) and a heavy chain (H chain;
which is 100 kDa). The L chain and H chain are held
together by a long peptide belt, non-covalent interac-
tions and a single inter-chain disulphide bond (FIG.2a).
The crystallographic structures of the entire BoNT/A1,
BoNT/B1 and BoNT/E1 are available17–19, in addition to
some individual domains and L chain–substrate com-
plexes. Similarly to all bacterial exotoxins that have intra-
cellular targets, BoNTs consist of multiple domains that
fulfil different functions during the intoxication process:
the L chain encodes the toxic moiety, which is a metal-
loprotease domain that specifically cleaves the SNARE
(soluble N-ethylmaleimide-sensitive factor attachment
protein receptor) proteins that are necessary for neuro-
transmitter exocytosis; the HN domain (the Nterminus
Figure 1 | Animal and human botulism. a | Botulism
mainly affects wild and domesticated animals and begins
with the growth of toxigenic clostridia in decaying
anaerobic material, followed by release of the toxin. This
infected material is consumed by botulinum neurotoxin
(BoNT)-insensitive invertebrates (such as maggots), which
disseminate the bacterium and the toxin to vertebrates.
The cadavers of intoxicated animals provide an anaerobic
environment that enables the bacterium to proliferate
and release the toxin. The deposition of insect eggs (for
example, from flies) leads to the growth of many intoxicated
larvae, which are eaten by birds (or fish), generating a
self-amplifying cycle that may rapidly involve many birds
and/or fish. b | There are five forms of human botulism. The
two most common forms are food-borne botulism (which
occurs following the ingestion of BoNT-containing foods —
typically canned foods) and infant botulism, which is
caused by the ingestion of food contaminated with spores
that germinate into neurotoxigenic clostridia in the
gastrointestinal tract. In the infant gut, the bacterium has
the potential to proliferate, owing to a lack of competition
from the resident microbiota, which tends to be less robust
in infants. The other three forms of human botulism are
much rarer and include inhalational botulism (owing to
inhalation of BoNT-containing aerosols), iatrogenic botulism
(which is caused by the injection of excessive clinical doses
of BoNT) and wound botulism (which is almost exclusively
associated with drug injection). Following transcytosis
across the intestinal epithelium and subsequent entry
into the general circulation, the toxin eventually enters
peripheral cholinergic nerve terminals, which causes the
flaccid paralysis of botulism.
▶
REVIEWS
4
|
ADVANCE ONLINE PUBLICATION www.nature.com/reviews/micro
© 2014 Macmillan Publishers Limited. All rights reserved
of the H chain) is required for translocation of the
Lchain across the membrane of endocytic vesicles into
the neuronal cytosol; and the HC domain (the C termi-
nus of the H chain) is responsible for presynaptic bind-
ing and endocytosis and consists of two sub domains that
have different folding and binding propertie s (FIG.2a).
A unique feature of BoNTs is the presence of a belt that
encircles the L chain and attaches it tightly to the HN
domain. The level of amino acid sequence similarity
and the available single-domain structures suggest that
BoNT/C, BoNT/D and BoNT/G are similar to BoNT/A
and BoNT/B, whereas BoNT/F is more similar to
BoNT/E (BOX1).
The bont gene is located next to the non-toxic non-
haemagglutinin gene (ntnha), which encodes a protein
that forms a heterodimer with BoNT; for example,
BoNT/A1 and NTNHA/A1 adopt a similar fold and
bind to each other as interlocking hands20 (FIG.2b).
This arrangement, which results in extensive protein–
protei n contacts, effectively decreases the exposure of
the BoNT to external damaging agents. It also suggests
that the two genes resulted from a duplication event and
that the ntnha gene evolved to have a protective func-
tion1,9,20. Considering that the toxin is mostly produced
in decaying biological material, where it remains active
for months to years8, we propose that the primary role of
NTNHA is to protect BoNT from pH denaturation and
the many proteases and protein-modifying agents that
are present in this material8, rather than the previously
suggested primary role of protection in the gastrointes-
tinal tract21,22. As passage through the gastrointestinal
tract is relatively rapid (it occurs in minutes to hours),
protection against the hostile environment of this invivo
compartment is likely to be a secondary role of NTNHA.
The bont and ntnha genes are in close proximity to
genes that encode either haemagglutinin or OrfX pro-
teins; these proteins associate with the BoNT–NTNHA
heterodimer and are also thought to have a protective
role. The orfX locus is present in genomes that encode
BoNT subtypes A1 (strain NCTC2916), A2, A3, A4,
E1–E11 and F1–F6 (REF. 9). The haemagglutinin operon
is present in strains that produce BoNT/A1 (strain Hall),
A5, B1–B7, C, D and G (REF. 9). The protein products of
the haemagglutinin operon (which are HA17, HA33 and
HA70) form large complexes (known as haemagglutinin
complexes) that interact with the NTNHA–BoNT/A
heterodimer, generating large oligomers that are known
as progenitor toxin complexes (PTCs)23,24 (FIG.2c). The
corresponding PTCs of BoNT/B and BoNT/E have also
been structurally characterized24. The overall structure
of the PTC resembles that of a λ phage, and the haem-
agglutinin proteins show only little protein–protein
contact with NTNHA and no contact with BoNT/A1
(FIG.2c). The haemagglutinin proteins of PTCs provide
nine potential carbohydrate-binding sites23, and these
structural features, as well as recent experiments25–27,
suggest that the main role of the haemagglutinin com-
plex is to facilitate trans-epithelial absorption of the
toxin23 (rather than the previously suggested role in
protection21,22). Thus, it is possible that haemagglutinins
function as adhesins and attach to the mucus layer, epi-
thelial cells or other cells in the intestinal layer, such as
M cells and neuroendocrine crypt cells27. A complex entry
route has been suggested, in which the PTC is proposed
to cross the epithelial barrier, followed by its release on
the basolateral side. The haemagglutinin complex then
dissociates from PTC and disrupts the epithelial barrier
by loosening E-cadherin-mediated cell–cell adhesion,
which opens the paracellular route to the toxin25,26.
Nature Reviews | Microbiology
Ingestion of toxic maggots
Growth of C. botulinum,
toxin production and
deposition of insect eggs
Transcytosis across
intestinal epithelium
BoNT
Spread in general circulation Entry into peripheral
nerve terminals
Infant botulism
Adult botulism
Death of intoxicated vertebrate
a
b
Decomposing organic matter
containing C. botulinum
Iatrogenic
botulism
Food-borne
botulism
Inhalational
botulism
Wound
botulism
Ingestion
of spores
Fly
REVIEWS
NATURE REV IEWS
|
MICROBIOLOGY ADVANCE ONLINE PUBLICATION
|
5
© 2014 Macmillan Publishers Limited. All rights reserved
M cells
Specialized epithelial cells of
the follicle-associated
epithelium of the
gastrointestinal tract that are
involved in the rapid uptake
and presentation of particular
antigens and microorganisms
to immune cells of the
lymphoid follicle, thereby
inducing an effective immune
response.
Neuroendocrine crypt cells
Cells that are distributed
throughout the intestinal
epithelium and that secrete
peptide hormones in an
endocrine or paracrine manner
from dense core or
neurosecretory granules.
Synaptic vesicles
Neuronal vesicles that store
and release neurotransmitters
or neuropeptides at the
synapse.
Presynaptic receptor
A receptor that is localized on
the surface of the presynaptic
membrane; it is either protein
or lipid in nature.
By contrast, other studies indicate that BoNT alone is
capable of crossing epithelial cells13,28, although with
lower efficiency. Additional studies are clearly required
to determine the role of the PTC complex invivo. Future
research should investigate the role of the range of acces -
sory proteins that are associated with different BoNTs
and should use representative models of the intestine
that are derived from different animal species and that
include the mucus layer.
Entry into the circulation
After breaching the intestinal epithelial barrier, BoNTs
disperse in extracellular fluids and enter the lymphatic
system and then the blood circulation13. The mecha-
nism that is involved in crossing lymphatic and blood
vessels is unknown, but BoNTs are unable to cross the
blood–brain barrier and therefore cannot enter the cen-
tral nervous system (CNS) using this route29. BoNTs are
robust and can remain in the circulation for many days
in humans30,31 and rodents13.
The specificity of BoNTs is surprising as, among
the hundreds of different cell types that are present
in the body of vertebrates, BoNTs only bind to periph-
eral nerve terminals, particularly those of skeletal and
autonomic cholinergic nerves32, the surfaces of which
are only a small proportion of the total cell surface area
that is exposed to extracellular fluids. This is even more
remarkable considering that the mouse lethal dose
corresponds to a BoNT concentration of ~10−15 M in
circulating fluids.
Dual receptor binding
To selectively target the presynaptic membrane of
peripheral nerve terminals, BoNTs have evolved a
unique binding mode that is based on the use of two
independent receptors: a polysialoganglioside (PSG) and
a protein receptor in the lumen of synaptic vesicles33–37
(FIG.3). It is also likely that additional, low affinity but
selective interactions contribute to neurospecificity38–40.
This unique binding mode may have evolved to simul-
taneously overcome several physiological obstacles, such
as the low BoNT concentration in circulating fluids, the
high velocity of movement of extracellular fluids around
cells and the reduced surface area of peripheral nerve
terminals compared with that of other cells that are
exposed to extracellular fluids.
Initial binding. The ‘evolutionary choice’ of PSG as
the first presynaptic receptor that BoNT contacts onthe
nerve terminal41,42 seems to be ideal, as PSG molecules
are present at a high density on the presynaptic mem-
brane, are organized in microdomains that also include
glycoproteins, and their oligosaccharide portion (which
is the BoNT-binding moiety) is flexible and projects far
beyond the membrane surface43,44. In addition, PSGs
form a large family of glycolipids with chemically com-
plex oligosaccharides that can generate very specific
interactions with target proteins. PSGs also influence
transmembrane signalling, endocytosis and vesicle
trafficking43,44. Thus, PSGs are perfectly equipped to
function as ‘antennae’ that capture BoNTs as they pass
in close proximity and thereby concentrate them on the
nerve terminal surface. Indeed, BoNTs bind to the most
distal part of the PSG sugar head via a PSG-binding site
that is located in the HC domain of the BoNT molecule
(FIG.3). That PSG alone functions as the first and major
presynaptic receptor of BoNTs is also supported by the
fact that some autoimmune PSG-specific antibodies
bind to PSG and recruit complement on the presynaptic
membrane, causing entry of Ca2+ ions45,46.
The binding of BoNT to the negatively charged
PSG molecule is probably rapid, as it is likely to be
controlled only by the rate of diffusion. In fact, BoNTs
are dipoles, with their positively charged end located
close to the binding site on PSG (FIG.2a). Thus, PSG
and other anionic lipids might be involved in reorient-
ing the BoNT dipole as it approaches the membrane,
which would make almost any PSG-binding attempt
productive47. This effect may contribute to the rapid
binding of BoNTs to the nerve terminal invivo13. In
terms of binding density, studies of the rat NMJ have
shown that hundreds of BoNT/A or BoNT/B molecules
can bind per square micrometre of the presynaptic
membrane48.
Figure 2 | Structure of isolated BoNT molecules and
BoNT complexes. a | Crystal structure of botulinum
neurotoxin A1 (BoNT/A1)17, showing its associated
electrical dipole and the organization of individual toxin
domains, each of which has a specific function in cell
intoxication: the HC domain binds specifically to nerve
terminals; the HN domain translocates the L chain into the
nerve terminal cytosol; and the L chain is a metalloprotease
that cleaves and inactivates specific SNARE (soluble
N-ethylmaleimide-sensitive factor attachment protein
receptor) proteins that are involved in neurotransmitter
release, thereby causing nerve paralysis. A peptide belt
(shown in dark blue), which surrounds the L domain and the
inter-chain disulphide bond (orange), links the L chain to
the HN domain. This unique feature is also present in
BoNT/B1 and BoNT/E1 (Protein Data Bank (PDB) accessions
3BTA, 1EPW and 3FFZ). b | Crystal structure of BoNT/A1 in
complex with the NTNHA/A1 protein20 (PDB accession
3VOC). NTNHA/A1 has the same domain organization
as BoNT/A1, and the two proteins form an interlocking
complex, which suggests that NTNHA/A1 protects
BoNT/A1 from proteases and other damaging agents that
the toxin encounters in the exvivo environment and in the
gastrointestinal tract. The lower bar shows the schematic
organization of the two proteins. c | Structure of the
precursor toxin complex (PTC), which contains the
NTNHA/A1–BoNT/A1 heterodimer complexed to PTC/A1
(REF.23), in which the NTNHA/A1–BoNT/A1 heterodimer
occupies the central position and the haemagglutinin
proteins (HA17, HA33 and HA70) of the PTC are shown
underneath. There are six HA33 proteins, three HA17
proteins and three HA70 proteins in each NTNHA/A1–
BoNT/A1 complex. PTC/A1 forms three spider-like legs
that have little protein–protein contact with NTNHA/A1
and no contact with the toxin. This structure suggests that
the haemagluttinin proteins function in binding to the
intestinal epithelium to facilitate absorption of the toxin,
rather than in the protection of BoNT from protease attack.
Structure in part c courtesy of R. Jin, University of
California, Irvine, USA.
▶
REVIEWS
6
|
ADVANCE ONLINE PUBLICATION www.nature.com/reviews/micro
© 2014 Macmillan Publishers Limited. All rights reserved
The PSG-binding site is located on the surface of the
carboxy-terminal subdomain of the HC domain (HC-C
domain) (FIG.2a) and has been characterized in detail34.
In BoNT/A, BoNT/B, BoNT/E, BoNT/F and BoNT/G,
the PSG-binding site is defined by the motif E(or D or
Q)---H(or K or G)---SXWY---G (where X is any amino
acid and --- denotes a variable number of residues). The
PSG-binding site of BoNT/C, BoNT/DC and BoNT/D is
located in a similar position, but the binding residues are
different. This group of BoNTs also has a second PSG-
binding site within the HC domain, which includes the
W-Y(or F) residues. These residues are known to be
involved in binding to carbohydrate receptors49–51, and
this PSG-binding site is involved in binding to neu-
rons in culture35,49–54. In addition, there is evidence that
fibroblast growth factor receptor 3 (FGFR3) might be
involved in the binding of BoNT/A1 to nerve cells55.
Such binding could account for the high specificity and
affinity of BoNT/A1 for peripheral nerve terminals.
Functional binding. Following attachment to PSG,
BoNT/B1, BoNT/DC and BoNT/G bind to segment
40–60 of the synaptic vesicle luminal domain of
synaptotagmin (Syt) via a binding site in the HC-C
domain that is close to the PSG-binding site56–59 (FIG.3).
However, the two binding sites are structurally sepa-
rated, and binding interactions with PSG and Syt are
independent of each other35–37,60. By contrast, BoNT/A1
and BoNT/E1 bind specifically to two different segments
of the fourth luminal loop of the synaptic vesicle trans-
membrane protein SV2 (REFS 61–64). Although isoform
SV2C seems to be the main receptor that is involved in
BoNT/A1 binding invitro63,65 — via an interaction with
the N-terminal and C- terminal subdomains of the HC
domain65 (FIG.3) — both SV2A and SV2B can also medi-
ate BoNT/A1 entry, and all three isoforms are expressed
on motor nerve terminals61,62. Glycosylated residues are
present in the toxin-binding site of SV2 (REF. 65) and
are potentially clinically relevant, but this requires
further investigation. In fact, a different pattern of
glycosylation among individuals would provide a sim-
ple explanation for the variable sensitivity of different
patients to BoNT/A1 injection, which is often observed
in clinical settings. Clearly, this variability might also be
applicable to different vertebrate species.
Syt and SV2 are integral proteins of the synaptic vesi-
cle membrane and expose their BoNT-binding sites to
the synaptic vesicle lumen (FIG.3). Therefore, unlike PSG,
these protein receptors are not exposed on the nerve
terminal surface and are not accessible to BoNT. How-
ever, they become available following the fusion of the
synaptic vesicle with the presynaptic membrane, which
exposes the synaptic vesicle lumen to the extracellular
environment (BOX2). Accordingly, BoNT binding to
protein receptors occurs only after fusion of the synap-
tic vesicle to the presynaptic membrane, and this seems
to facilitate the subsequent step of intoxication, which
requires the endocytosis of BoNT (FIG.3). However, it is
possible that some Syt molecules might be present on the
presynaptic membrane following complete merging of
the synaptic vesicle with the plasma membrane66 (FIG.3).
Nature Reviews | Microbiology
a BoNT/A1
b BoNT/A1–NTNHA/A1 complex
c PTC/A1 complex
110 Å
L chain
NC
+–
S S HNL HC
Catalytic
domain Translocation
domain Binding
domain
H chain
130 Å
110 Å
120 Å
260 Å
NC
NC
S S HNL
nHC nHN nL
HC
BoNT/A1
NTNHA/A1
HA70
HA70
HA17 HA17
HA33 HA33 HA33
REVIEWS
NATURE REV IEWS
|
MICROBIOLOGY ADVANCE ONLINE PUBLICATION
|
7
© 2014 Macmillan Publishers Limited. All rights reserved
Synaptotagmin
(Syt). A protein that spans the
membrane of synaptic vesicles
and binds to Ca2+ to trigger the
fusion of synaptic vesicles with
the plasma membrane of the
neuron.
SV2
A protein that spans the
membrane of synaptic vesicles
and has an unknown function.
Following fusion of the synaptic
vesicle to the plasma
membrane, the luminal domain
of SV2 becomes exposed to
the extracellular medium and
functions as a receptor for
botulinum neurotoxins.
Figure 3 | Binding and trafficking of botulinum neurotoxins inside nerve terminals. The first step in intoxication
involves the binding of the carboxy-terminal end of the HC domain (the HC-C domain) to a polysialoganglioside (PSG)
receptor that is present on the presynaptic membrane, followed by binding to a protein receptor (either synaptotagmin
(Syt) or SV2) that is located either inside the exocytosed synaptic vesicle or on the presynaptic membrane (step 1).
The crystal structure of botulinum neurotoxin B (BoNT/B) bound to Syt and PSG is shown on the lower left-hand side
and the crystal structure of BoNT/A bound to PSG and to SV2 is shown on the lower right-hand side. The BoNT is then
endocytosed inside synaptic vesicles (step 2) as it exploits the vesicular ATPase proton pump, which drives the re-uptake
of neurotransmitter. Owing to the acidification of the vesicle, the BoNT becomes protonated, which results in
translocation of the L chain across the synaptic vesicle membrane (step 3) into the cytosol. Translocation can also occur
across the endosomal membrane following the fusion of a synaptic vesicle with an endosome (which seems to occur in
cultured neurons68). The L chain is released from the HN domain, owing to the action of the thioredoxin reductase–
thioredoxin system (TrxR–Trx), which cleaves the inter-chain disulphide bond (S–S; shown in orange). The L-chain
metalloproteases of BoNT/B, BoNT/D, BoNT/F and BoNT/G cleave VAMP, the L-chain metalloproteases of BoNT/A and
BoNT/E cleave SNAP25 and the L-chain metalloprotease of BoNT/C cleaves both SNAP25 and syntaxin (step 4), all of
which result in the inhibition of neurotransmitter release and consequent neuroparalysis.
Nature Reviews | Microbiology
H+ Trx
VAMP
BoNT/A, C, E
PSG
Syt or SV2
L chain
ATPase
proton
pump
Neurotransmitter
BoNT/C
BoNT/B, D, F, G
SNAP25
Syntaxin
ATP
2
3
4
1
ADP
SH
HN domain
HC-C domain
HC-N domain
Syt
SV2
Presynaptic membrane
Cytosol
Nerve terminal surface PSG
S–S bond
PSG
REVIEWS
8
|
ADVANCE ONLINE PUBLICATION www.nature.com/reviews/micro
© 2014 Macmillan Publishers Limited. All rights reserved
Phrenic nerve
hemidiaphragm
An ex vivo preparation that
includes a portion of the
diaphragm, as well as the axon
and nerve terminal of the
phrenic nerve. This nerve
contains motor, as well as
sensory and sympathetic,
fibres and controls the
contraction of the diaphragm
muscle via the release of
acetylcholine. Its inhibition by
botulinum neurotoxins blocks
respiration, which causes
death.
Patch clamp technique
An electrophysiological
technique that is based on
microelectrodes that are
sealed on the plasma
membrane of a cell, which
enables the measurement of
electrical activity and the
properties of ion channels.
The protein receptors of other BoNTs have not been
characterized in comparable detail so far, and conflicting
results have been reported, which indicates that further
characterization is needed.
Entry into nerve terminals
The second step of nerve terminal intoxication involves
BoNT internalization (FIG.3). The dual binding interac-
tion with PSG and synaptic vesicle receptors (Syt or SV2,
depending on the toxin serotype, as discussed above)
increases the strength of BoNT interactions with the
membrane, which is the product of the two binding
affinities33.
In both cultured neurons and invivo, BoNT/A1
rapidly enters the synaptic vesicle lumen67,68, and the
number of toxin molecules (either one or two67,68)
correlates with the number of SV2 molecules in the
synaptic vesicle membrane69. The rate of entry for
BoNT/A1 correlates with the rate of synaptic vesicle
endocytosis70 and with the rate of paralysis of the mouse
phrenic nerve hemidiaphragm, which is the standard
NMJ that is used to test the potency of BoNTs71,72. The
mechanism of internalization of other BoNTs remains
to be established, but their ability to rapidly paralyse
the mouse phrenic nerve hemidiaphragm suggests that
they all use the synaptic vesicle as a ‘Trojan horse’ to
enter motor neuron terminals invivo. By contrast, in
cultured CNS neurons, other vesicles and trafficking
routes might contribute to entry68, particularly at the
very high toxin concentrations that are frequently used
in the laboratory73.
Release into the cytosol
In order to reach their target SNARE proteins in the
cytosol of nerve cells, the catalytically active L chain
must be translocated from the synaptic vesicle lumen
into the cytosol. The main driving force for L-chain
translocation is the transmembrane pH gradient that
is generated by the vesicular ATPase proton pump,
which drives the re-entry of neurotransmitter into the
synaptic vesicle (along with H+ ions) after exocytosis74
(BOX2; FIG.3). This is supported by the observation that
specific ATPase inhibitors completely block nerve ter-
minal intoxication by all BoNTs52,67,73,75–77. Thus, BoNTs
of neurotoxigenic clostridia have evolved to exploit two
major physiological events that occur at nerve terminals:
synaptic vesicle endocytosis (to enter nerve terminals)
and neurotransmitter refilling of the synaptic vesicle (to
deliver the L chain metalloprotease into the cytosol).
The molecular aspects of BoNT translocation across the
synaptic vesicle membrane into the cytosol have been
only partially elucidated, but studies that have been car-
ried out in the past decade have provided considerable
insights and have led to the proposal of a molecular
model for this process78,79.
Translocation across the synaptic vesicle membrane. It
has long been known that BoNTs form ion channels of
low conductance in planar lipid bilayers at low pH80–82,
and this process is associated with translocation of the
L chain and the cleavage of its target SNARE proteins83.
A major advance in understanding the mechanisms that
are involved was made using the patch clamp technique
in Neuro2A cells78,79,84,85 and PC12 cells77,86. This experi-
mental approach mimics invivo conditions and ena-
bles events that occur at the single-molecule level to
be resolved79,85. Collectively, these studies suggest that
lowering the pH at the cis side of the membrane (that
is, the side that faces the synaptic vesicle lumen) induces
the L chains of BoNT/A1 and BoNT/E1 to cross the
membrane through a channel that is 15–20 Å in diam-
eter87 (FIG.4a). These channel dimensions enable the
passage of α-helices but not of tertiary structural ele-
ments, which suggests that the L chain must unfold to
pass through the channel. Stabilization of the L chain
tertiary structure with antibodies prevents channel for-
mation88, which highlights the importance of unfolding
for translocation and also suggests that this unfolding
is linked to channel formation. This conclusion is also
supported by the finding that cargo molecules, which
are capable of unfolding at low pH, are transported
into the neuronal cytosol when they are attached to
the N terminus of BoNT89. Further studies have sug-
gested that the HN domain alone is sufficient to form
the transmembrane channel and that the peptide belt
that links the L chain and the H chain regulates the
formation of the HN channel78,79,90,91. Residues that are
present in all three BoNT domains are responsible for
the pH sensitivity of translocation52,92. The release of the
L chain on the trans side (that is, the cytosolic side) of
the membrane requires the inter-chain disulphide bond
to be reduced84. The crucial role of cytosolic disulphide
bond reduction is highlighted by the fact that BoNTs
that have a reduced inter-chain disulphide bond do
not form channels84. These data are consistent with
the finding that only reduced BoNTs can hydrolyse
their substrates93 and also explain why this disulphide
bridge is essential for neurotoxicity52,94,95. On the basis
of these data, a model for translocation has been pro-
posed (FIG.4a). This model posits that the low pH of the
synaptic vesicle lumen induces a conformational change
in the HN domain, which then inserts into the mem-
brane and forms a translocation channel that chaper-
ones the passage of the partially unfolded L chain from
the luminal side to the cytosolic side of the synaptic
vesicle membrane. The L chain remains attached to the
synaptic vesicle until the inter-chain disulphide bond
is reduced, which occurs at the end of this process78,84.
More recent data (discussed below) have clarified
the molecular events that are involved in the interaction
between BoNT and the membrane (which are induced
at low pH) and in the reduction of the inter-chain disul-
phide bond (FIG.4b). BoNT/B1, and the L chain and
HNdomains of BoNT/A1, do not change conforma-
tion at low pH in solution90,96, whereas they do change
conformation in the presence of PSG or PSG-containin g
membranes77,90,97,98. Using a protocol that bypasses
the synaptic vesicle internalization step and enablesthe
Lchain to be translocated from the cell surface into the
cytosol, it was found that BoNTs must be anchored to
the membrane by two receptors52 and that translocation
occurs within minutes at 37 °C in the pH range 4.5–6
REVIEWS
NATURE REV IEWS
|
MICROBIOLOGY ADVANCE ONLINE PUBLICATION
|
9
© 2014 Macmillan Publishers Limited. All rights reserved
pH sensor
In the context of this Review;
amino acid residues that
change protonation state
according to variations in pH.
A change in protein structure
may consequently occur, owing
to altered hydrogen bonding
and electrostatic interactions.
(REFS 77,99), which is consistent with the pH inside the
synaptic vesicle100,101. It was also noted that one face of
the BoNT contains several conserved high pKa carbox-
ylates, the inter-chain disulphide bond and a segment
that has a high propensity for membrane insertion52,97.
Replacement of three of the carboxylate residues with
the corresponding amides in BoNT/B1 eliminates the
requirement for their protonation and causes the L chain
to enter the cytosol more rapidly, thus increasing toxic-
ity92. These data suggest that there is no single pH sensor
in BoNTs, but instead, several carboxylates that have
high pKa values have a role in the low pH-driven release
of the L chain into the cytosol.
Updated model of BoNT translocation. By consider-
ing the findings that are described above, an updated
model for BoNT translocation can now be presented. It
should be noted that this model requires further experi-
mental studies to determine whether the steps that are
outlined below are indeed correct. BoNT initially binds
to its two receptors (PSG and SV2 or Syt) inside the
synaptic vesicle lumen, which has a neutral pH, imme-
diately after endocytosis (FIG.4b). The vesicular ATPase
then pumps protons into the synaptic vesicle and the
luminal pH becomes progressively more acidic. Notably,
protons and other cations are attracted to the anionic
membrane surface of the synaptic vesicle and their local
Nature Reviews | Microbiology
pH 5
a
b
Synaptic vesicle lumen
Cytosol
HN domain
S–S bond
-SH
HS-
Unfolded LC
L chain
15–20 Å
L chain α-helices
pH 5.5–6.0
pH 7
PSG Syt or SV2
H+
pH 7
Synaptic vesicle membrane
Figure 4 | Model for the molecular events that occur during L-chain translocation across the synaptic vesicle
membrane. a | On the basis of a series of biophysical studies, a model for membrane translocation of the L chain has been
proposed78. Owing to the action of the vesicular ATPase proton pump, the pH in the lumen of synaptic vesicles becomes
acidic, which causes a conformational change in the HN domain, enabling it to penetrate the lipid bilayer. This leads to the
formation of a channel that chaperones the partially unfolded L chain across the membrane. The inter-chain disulphide
bond (S–S bond) is proposed to cross the membrane at a late stage during translocation, and its reduction on the cytosolic
side of the synaptic vesicle membrane releases the L chain into the cytosol. b | Proposed steps that are suggested to occur
before membrane insertion of botulinum neurotoxins (BoNTs) (BoNT/B, Protein Data Bank accession 1EPW, is shown). The
process begins with the binding of BoNT to its two receptors (polysialoganglioside (PSG) and Syt or SV2), which is a
prerequisite for subsequent steps (left-hand structure). Cationic and anionic residues in BoNT/B are shown in blue and
red, respectively. The acidification of the synaptic vesicle lumen (owing to the action of the ATPase proton pump) causes
protonation of some conserved glutamic acid and aspartic acid residues that have high pKa values (circled in yellow in the
left-hand structure), which are clustered on one face of the toxin. This face of the toxin also contains the inter-chain
disulphide bond52 (orange) and a segment that is predicted to have high membrane insertion propensity96 (brown; centre
structure). The partially protonated, positively charged face of BoNT is attracted by the anionic membrane surface, which
is rich in negatively charged lipids, and the toxin eventually collapses onto the membrane. The ensuing events are unclear,
but it is likely that the long α-helices of the HN domain break into shorter helices (shown in right-hand panel with different
colours), which then insert into the lipid bilayer to form the ion channel. The L chain is predicted to become a molten
globule, which is a protein conformation that is capable of inserting into the lipid bilayer of the membrane.
REVIEWS
10
|
ADVANCE ONLINE PUBLICATION www.nature.com/reviews/micro
© 2014 Macmillan Publishers Limited. All rights reserved
Neuroexocytosis
nanomachine
A molecular machine of
nanometre dimensions that is
used for the release of
neurotransmitters.
concentration reduces the pH near the membrane to
1–1.5 units below that of the lumen100–103 (FIG.4b).
The amino acids histidine, glutamate and aspartate
become protonated within the pH range (4.5–6) and
are predicted to be involved in L-chain translocation.
However, the actual pKa values of these residues depend
on their molecular surroundings. BoNTs lack conserved
histidine residues, except those that are in the active
site, but they do contain conserved carboxylate residues
that are predicted to have high pKa values52. Assum-
ing that the residues that are important for the low-
pH-driven process are conserved, seven conserved
carboxylates that have high pKa values are located in
the HN domain, three are located in the L chain and
one is located in the HC domain52 (FIG.4b). The spatial
distribution of these residues reinforces the suggestion
that BoNTs contain more than a single pH sensor92.
The model posits that these carboxylates become proto-
nated — partially or entirely — in a sequential manner
(depending on their pKa values) as the pH of the syn-
aptic vesicle lumen decreases (FIG.4b). Even a partially
protonated BoNT has a net positive charge that favours
its interaction with the anionic membrane surface52,103–105.
The BoNT surface that is involved in membrane inter-
actions is suggested to be the surface that contains the
inter-chain disulphide bond and the membrane-inserting
segment (residues 637–688) (FIG.4b); the opposite side
of the BoNT molecule lacks carboxylates of appropriate
pKa values52.
The predicted collapse of BoNT onto the membrane
surface is not prevented by receptor interactions, as
either binding is weakened by the low pH65 or the two
receptors are flexible35,36. BoNT is suggested to undergo
a gross structural change that involves both the L chain
and the HN domain and is facilitated by simultaneous
changes in the conformation and organization of mem-
brane lipids (FIG.4). Such changes are caused by the acidic
pH of the lumen, but other factors that might contribute
include ionic strength, the high Ca2+ concentration and
the high negative curvature of the luminal synaptic vesi-
cle membrane. The ensuing molecular events are cur-
rently unknown, but, on the basis of previous studies,
we suggest that the L chain becomes a ‘molten globule’,
which is a protein variant that retains native secondary
structure but has increased hydrophobicity, to enable
membrane insertion98,106–109. The α-helices of the HN
domain contain amphipathic segments and residues that
have a low propensity to form a helical structure, which
suggests that the long α-helices of the HN domain might
break into shorter protein segments that insert into the
membrane and thereby form an ion channel. However,
whether this actually occurs is currently unknown, and
clearly, more studies are needed to clarify this essential
step of the BoNT intoxication process.
Importantly, the reduction of the inter-chain disul-
phide bond at any stage before its exposure to the cyto-
sol prevents L-chain translocation, so this domain must
emerge on the cytosolic side before reduction takes
place84. The reduction of protein disulphide bonds is
catalysed in the cell cytosol by different enzymatic sys-
tems, including glutaredoxins, thioredoxins and other
systems110–112. Using a discriminating pharmacologi-
cal approach, the redox system NADPH–thioredoxin
reductas e (TrxR)–thioredoxin (Trx) was found to have
a major role in release of the L chain into the neuronal
cytosol113. Following Trx-mediated reduction of the
disulphide bond, L-chain translocation is irreversible and
the toxin is now free to interact with its target proteins
(FIG.3). The Trx tertiary fold is similar to that of ancestral
chaperonins, so it is also possible that Trx functions as
a chaperonin for L-chain translocation112,114.
Mechanism of BoNT-induced neuroparalysis
The L chains of all known BoNTs are metalloproteases
that are specific for one of the SNARE proteins: VAMP
(vesicle-associated membrane protein; also known as
synaptobrevin), SNAP25 (synaptosomal-associated
protei n of 25 kDa) or syntaxin (FIG.3). BoNT/C cleaves
both SNAP25 and syntaxin, BoNT/B, BoNT/D, BoNT/F
and BoNT/G only target VAMP and BoNT/A and
BoNT/E cleave SNAP25. The fact that inactivation of
any one of these three proteins inhibits neurotransmitter
release is the strongest evidence that these three proteins
form the core of the neuroexocytosis nanomachine115,116
(BOX2). The SNARE family of proteins includes many
isoforms of VAMP, SNAP25 and syntaxin, which are
differentially expressed in many non-neuronal cells and
tissues. Although several of these isoforms can be
cleaved by BoNTs, these substrates are not accessible
invivo, as non-neuronal cells lack appropriate receptors
for the toxin3,117.
The molecular basis of the neuroparalytic activity
of BoNTs has recently been reviewed in depth116,117,
and only the more recent findings are discussed here.
With the exception of BoNT/A and BoNT/C, all BoNTs
cleave isolated SNARE proteins by removing large
cytosolic segments, which prevents the formation of
the SNARE complex118,119 (BOX2). BoNT/A and BoNT/C
remove only a few residues from the Cterminus of
SNAP25 (REFS 13,11 6 ,11 7 ), and this truncated form
of SNAP25 can form a stable SNARE complex118; thus,
the molecular mechanism of BoNT/A- and BoNT/C-
induced neuroparalysis remains to be elucidated. It is
possible that the core of the nanomachine is comprised
of a SNARE supercomplex that is formed by several
SNARE complexes and that the C terminus of SNAP25
is involved in protein–protein interactions among
the individual SNARE complexes116,120. An alternative
explanation is that BoNT/A cleaves another protein (or
proteins) that is (or are) essential for neurotransmitter
release. However, such protein substrates have not yet
been found, despite extensive searches, and they are
unlikely to exist, owing to the unique mode of recog-
nition of VAMP, SNAP25 and syntaxin by the L-chain
metalloprotease116,117. In fact, the SNARE-binding site
of the metalloprotease is a long channel that is occu-
pied by the peptide belt in the intact protein (FIG.2a);
however, when the L chain is released, this channel is
vacated and the substrate can then insert into the chan-
nel. The L chain interacts extensively with the substrate
and contacts several exosites of the protein in addition
to the cleavage site116,117.
REVIEWS
NATURE REV IEWS
|
MICROBIOLOGY ADVANCE ONLINE PUBLICATION
|
11
© 2014 Macmillan Publishers Limited. All rights reserved
Serum sickness
An illness of humans that is
caused by a hypersensitive
reaction to proteins in
antiserum derived from a
non-human source; it usually
occurs 4–10days after
exposure.
Camelid-like antibodies
Single-domain antibodies
that are derived from the
heavy-chain antibodies of
camelids; they are a new
generation of therapeutic
agents and immunoreagents.
The substrates of only a few of the BoNT subtypes
have so far been determined. In addition, the rate of
substrate turnover for the metalloproteases inside
nerve terminals is unknown, and this parameter has
a profound effect on the onset of paralysis. Although
it seems unlikely that novel BoNT substrates will be
found, it is probable that novel cleavage sites in the
SNARE proteins will be revealed, as indicated by
the recent report that BoNT/F5 cleaves VAMP121 at
a different peptide bond compared with BoNT/F1
(REF.122). The available evidence suggests that BoNT/A
subtypes have different enzymatic rates123,124, and it is
therefore possible that individual subtypes are highly
variable in their potency, onset and duration of action.
Understanding the enzyme kinetics of BoNTs invivo is
also essential for the design of novel inhibitors of
BoNTs (BOX3).
Duration of action. One special feature of BoNTs is
the reversibility of their action invivo. Patients with
botulism fully recover if death by respiratory paralysis
is prevented by mechanical ventilation2,4. This is prob-
ably because BoNTs paralyse the nerve terminal but do
not kill the neuron, the cell body of which is frequently
located a substantial distance away from the paralysed
nerve terminal. However, although BoNTs do not cause
axonal degeneration at the doses that cause botulism,
reversible muscle atrophy is evident.
There is a remarkable diversity in the duration of
BoNT-induced neuroparalysis. The lifetime of the met-
alloprotease within the nerve terminal cytosol is the
predominant, but not the only, factor that contributes to
the duration of paralysis116,125,126. In general, the duration
of paralysis in mice is in the following order (with the
longest duration of action first): BoNT/A1, BoNT/C 1,
BoNT/B1, BoNT/D, BoNT/F1, BoNT/G, BoNT/E1.
However, the duration of paralysis also varies with the
vertebrate species, the activity of the affected muscle and
the toxin dose. Moreover, the paralysis of autonomic
human cholinergic nerve terminals lasts 3–4 times
longer than that of the NMJ127. However, as data for less
than one-third of the BoNTs subtypes are available, it
seems likely that paralytic activity may vary considerably
among different subtypes.
Duration of action is the main factor that contributes
to the biological action of BoNTs, as it determines the
severity of human botulism (type A toxins cause more
severe botulism than type B toxins, which cause much
more severe botulism than type E toxins)2–4. Knowing
the duration of action should also provide information
about the mechanisms of SNARE protein inactivation
and turnover inside nerve terminals125,128 as well as the
assembly of the SNARE complex116. In addition, the per-
sistence of BoNT activity is important for their therapeu-
tic application, as a toxin that has a longer duration of
action requires fewer injections of lower doses. Never-
theless, for certain conditions, such as facial lacerations
or disjointed bone fractures, a toxin that has a short
duration of action might be more useful in ameliorating
the course and outcome of the illness. There is a grow-
ing area of research that aims at changing the binding
specificity, affinity and the duration of BoNT action in
order to obtain tailor-made therapeutic agents and more
sophisticated tools to be used in cell biology studies129–132.
Conclusions and future perspectives
The toxic potency of BoNTs is the result of targeting
a physiological function that is essential for life in all
vertebrates. The discovery of many novel BoNTs, which
vary in potency and duration of action, raises questions
regarding the evolutionary advantage that is associated
with the production of such a large number of diverse
neurotoxins that have the potential to kill the host. How-
ever, this seems to be an obligatory survival strategy for
an anaerobic organism that can multiply only within a
non-oxygenated medium, such as a cadaver. In turn, this
strategy must be coupled to an alternative lifestyle, such
as sporulation, which enables the anaerobic organism
to survive the complete consumption of nutrients and
Box 3 | Vaccines, antibodies and chemical inhibitors
There is currently no approved pharmacological treatment for botulinum neurotoxin
(BoNT) intoxication, but the growing concern for the potential use of BoNTs as
biological weapons and the need to prevent botulism outbreaks has stimulated
research aimed at developing a range of agents to prevent and/or treat botulism.
Vaccines
Early attempts to create botulism-specific vaccines involved the treatment of partially
purified BoNTs with formalin (to inactivate the BoNT) and the addition of aluminium
hydroxide as an adjuvant142. However, after it was shown that injection of a
recombinant version of the HC domain of tetanus neurotoxin was sufficient to induce
a protective immune response against tetanus143, various BoNT HC domains were
expressed in Pichia pastoris and were shown to induce protective antibodies in
animals144,145. A recombinant vaccine composed of the HC domains from BoNT/A1 and
BoNT/B1 has shown promising results in clinical trials, and vaccines for other
serotypes are now under development. Other BoNT domains have been tested in
animals146, but there is currently no licensed vaccine available for human use. However,
several animal vaccines are on the market, which are used to prevent botulism
outbreaks.
Antibodies
Specific antitoxin antibodies can be used to prevent and treat botulism by eliminating
circulating BoNTs. Antibodies are also often used in research laboratories to identify
BoNT serotypes. However, BoNT-specific antibodies have difficulty entering neurons
so their use for the treatment of overt disease is limited. BoNT-specific antibodies have
traditionally been produced in animals — mainly horses142 — however, although these
polyclonal antibodies are efficient at BoNT neutralization, they are usually rapidly
eliminated from the human body and can also lead to serum sickness. To overcome
these obstacles, BoNT/A-specific polyclonal human antibodies have been isolated from
the sera of human volunteers that have been immunized with BoNT toxoids and are
used to treat infant botulism147. However, sophisticated biotechnologies to produce
high-affinity humanized monoclonal antibodies are now available and have been used
to produce BoNT/A-, BoNT/B-, BoNT/E- and BoNT/F-specific antibodies148–150, and
workto extend this approach to all serotypes is ongoing. Another promising approach
is to generate single-chain toxin-binding camelid-like antibodies, which have the
potential for intracellular use151,152.
Small-molecule inhibitors
Therapeutic inhibitors for post‑intoxication treatment must block L‑chain
metalloprotease activity inside nerve terminals. Such inhibitors must be non-toxic and
capable of crossing the plasma membrane of nerve cells. However, the development of
such agents is complicated owing to the complex mode by which the L chain binds to
its substrate, which involves several interaction sites116,117. Thus, despite intensive efforts
in several laboratories, including the screening of large chemical and natural compound
libraries, structure-based molecular design and several chemical synthesis approaches,
few molecules have passed the stage of inhibition of the toxins in cultured neurons and
have therefore not yet reached the level of testing in animals153.
REVIEWS
12
|
ADVANCE ONLINE PUBLICATION www.nature.com/reviews/micro
© 2014 Macmillan Publishers Limited. All rights reserved
the ensuing exposure to oxygen. Thus, toxin production
and sporulation are essential aspects of the life cycle of
toxigenic clostridia.
Recent research has unravelled the molecular basis
of BoNT action, including neurospecific binding and
the mechanisms that are involved in the catalytic
cleavage of the core proteins of the neuroexocytosis
nanomachine. However, several outstanding questions
remain, particularly regarding the mechanistic details
of toxin endocytosis into synaptic vesicles and the pro-
cess of L-chain translocation across the synaptic vesicle
membrane and its subsequent release into the cytosol.
Another major challenge is to establish methods for the
reliable comparison of the more than 40 distinct BoNTs
that have been identified so far, and of those that are yet
to be identified, which might reveal novel therapeutic
BoNTs that have increased potencies and durations of
1. Popoff,M.R. & Bouvet,P. Genetic characteristics of
toxigenic Clostridia and toxin gene evolution. Toxicon
75, 63–89 (2013).
2. Johnson,E.A. & Montecucco,C. in Handbook of
Clinical Neurology Vol. 91, 333–368 (ed. Engel, A. G.)
(Elsevier, 2008).
3. Schiavo,G., Matteoli,M. & Montecucco,C.
Neurotoxins affecting neuroexocytosis. Physiol. Rev.
80, 717–766 (2000).
4. Cherington,M. Clinical spectrum of botulism. Muscle
Nerve 21, 701–710 (1998).
5. Centers for Disease Control and Prevention,
Department of Health and Human Services.
Possession, use, and transfer of select agents and
toxins; biennial review. Final rule. Fed. Regist. 77,
61083–61115 (2012).
6. Arnon,S.S. etal. Botulinum toxin as a biological
weapon: medical and public health management.
J.Am. Med. Ass. 285, 1059–1070 (2001).
7. Lim,E.C. & Seet,R.C. Use of botulinum toxin in the
neurology clinic. Nature Rev. Neurol. 6, 624–636
(2010 ).
8. Smith,L.D.S. & Sugiyama,H. Botulism: the
Organism, its Toxins, the Disease (Charles C. Thomas
Publisher, 1988).
9. Hill,K.K. & Smith,T.J. Genetic diversity within
Clostridium botulinum serotypes, botulinum
neurotoxin gene clusters and toxin subtypes. Curr. Top.
Microbiol. Immunol. 364, 1–20 (2013).
10. Rocke,E.T. & Samuel,M.D. Water and sediment
characteristics associated with avian botulism
outbreaks in wetlands. J.Wildl. Management 63,
1249–1260 (1999).
11. Aureli,P. etal. Two cases of type E infant botulism
caused by neurotoxigenic Clostridium butyricum in
Italy. J.Infect. Dis. 154, 207–211 (1986).
This is the first report of botulism caused by a
clostridial species other than C.botulinum.
12. Koepke,R., Sobel,J. & Arnon,S.S. Global occurrence
of infant botulism, 1976–2006. Pediatrics 122,
e73–e82 (2008).
13. Simpson,L.L. The life history of a botulinum toxin
molecule. Toxicon 68, 40–59 (2013).
14. Wenham,T.N. Botulism: a rare complication of
injecting drug use. Emerg. Med. J. 25, 55–56 (2008).
15. Chertow,D.S. etal. Botulism in 4 adults following
cosmetic injections with an unlicensed, highly
concentrated botulinum preparation. J.Am. Med. Ass.
296, 2476–2479 (2006).
16. Dover,N., Barash,J.R., Hill,K.K., Xie,G. &
Arnon,S.S. Molecular characterization of a novel
botulinum neurotoxin type H gene J.Infect. Dis. 209,
192–202 (2014).
17. Lacy,D.B., Tepp,W., Cohen,A.C., DasGupta,B.R. &
Stevens,R.C. Crystal structure of botulinum
neurotoxin type A and implications for toxicity. Nature
Struct. Biol. 5, 898–902 (1998).
This study reports the first crystal structure of a
BoNT and provides the molecular basis for
understanding the mechanism of neuron
intoxication.
18. Swaminathan,S. & Eswaramoorthy,S. Structural
analysis of the catalytic and binding sites of
Clostridium botulinum neurotoxin B. Nature Struct.
Biol. 7, 693–699 (2000).
19. Kumaran,D. etal. Domain organization in Clostridium
botulinum neurotoxin type E is unique: its implication
in faster translocation. J.Mol. Biol. 386, 233–245
(2009).
20. Gu,S. etal. Botulinum neurotoxin is shielded by
NTNHA in an interlocked complex. Science 335,
977–981 (2012).
This study reports the unexpected finding that
NTNHA adopts a similar fold to BoNT and that
together, the two proteins form an interlocked
complex, which suggests that NTNHA stabilizes
BoNT and protects the toxin against proteolytic
cleavage.
21. Bonventre,P.F. Absorption of botulinal toxin from the
gastrointestinal tract. Rev. Infect. Dis. 1, 663–667
(1979).
22. Ohishi,I. & Sakaguchi,G. Oral toxicities of Clostridium
botulinum type C and D toxins of different molecular
sizes. Infect. Immun. 28, 303–309 (1980).
23. Lee,K. etal. Structure of a bimodular botulinum
neurotoxin complex provides insights into its oral
toxicity. PLoS Pathog. 9, e1003690 (2013).
24. Benefield,D.A., Dessain,S.K., Shine,N., Ohi,M.D.
& Lacy,D.B. Molecular assembly of botulinum
neurotoxin progenitor complexes. Proc. Natl Acad. Sci.
USA 110, 5630–5635 (2013).
25. Sugawara,Y. etal. Botulinum hemagglutinin disrupts
the intercellular epithelial barrier by directly binding
E-cadherin. J.Cell Biol. 189, 691–700 (2010).
26. Fujinaga,Y., Sugawara,Y. & Matsumura,T. Uptake of
botulinum neurotoxin in the intestine. Curr. Top.
Microbiol. Immunol. 364, 45–59 (2013).
27. Couesnon,A., Molgo,J., Connan,C. & Popoff,M.R.
Preferential entry of botulinum neurotoxin A H domain
through intestinal crypt cells and targeting to
cholinergic neurons of the mouse intestine. PLoS
Pathog. 8, e1002583 (2012).
28. Maksymowych,A.B. etal. Pure botulinum neurotoxin
is absorbed from the stomach and small intestine and
produces peripheral neuromuscular blockade. Infect.
Immun. 67, 4708–4712 (1999).
29. Restani,L. etal. Botulinum neurotoxins A and E
undergo retrograde axonal transport in primary motor
neurons. PLoS Pathog. 8, e1003087 (2012).
30. Sheth,A.N. etal. International outbreak of severe
botulism with prolonged toxemia caused by
commercial carrot juice. Clin. Infect. Dis. 47,
1245–1251 (2008).
31. Fagan,R.P., McLaughlin,J.B. & Middaugh,J.P.
Persistence of botulinum toxin in patients’ serum: Alaska,
1959–2007. J.Infect. Dis. 199, 1029–1031 (2009).
32. Dolly,J.O., Black,J., Williams,R.S. & Melling,J.
Acceptors for botulinum neurotoxin reside on motor
nerve terminals and mediate its internalization. Nature
307, 457–460 (1984).
This study provides the first evidence that BoNTs
bind specifically to the presynaptic membrane
before entering the nerve terminal.
33. Montecucco,C. How do tetanus and botulinum toxins
bind to neuronal membranes? Trends Biochem. Sci.
11, 314–317 (1986).
This paper proposes that dual receptor binding
could account for the high specificity and affinity of
tetanus toxin and BoNTs for the presynaptic
membrane.
34. Rummel,A. Double receptor anchorage of botulinum
neurotoxins accounts for their exquisite
neurospecificity. Curr. Top. Microbiol. Immunol. 364,
61–90 (2013).
35. Chai,Q. etal. Structural basis of cell surface receptor
recognition by botulinum neurotoxin B. Nature 444,
1096–1100 (2006).
36. Jin,R., Rummel,A., Binz,T. & Brunger,A.T. Botulinum
neurotoxin B recognizes its protein receptor with high
affinity and specificity. Nature 444, 1092–1095 (2006).
37. Berntsson,R.P., Peng,L., Dong,M. & Stenmark,P.
Structure of dual receptor binding to botulinum
neurotoxin B. Nature Commun. 4, 2058 (2013).
References 35, 36 and 37 describe the
crystallographic structure of BoNT/B in complex
with both its protein receptor and glycolipid
receptor, which provides experimental evidence for
the dual receptor binding model.
38. Montecucco,C., Rossetto,O. & Schiavo,G.
Presynaptic receptor arrays for clostridial neurotoxins.
Trends Microbiol. 12, 442–446 (2004).
39. Muraro,L., Tosatto,S., Motterlini,L., Rossetto,O. &
Montecucco,C. The N-terminal half of the receptor
domain of botulinum neurotoxin A binds to
microdomains of the plasma membrane. Biochem.
Biophys. Res. Commun. 380, 76–80 (2009).
40. Zhang,Y. etal. Structural insights into the functional
role of the Hn sub-domain of the receptor-binding
domain of the botulinum neurotoxin mosaic serotype
C/D. Biochimie 95, 1379–1385 (2013).
41. Van Heyningen,W.E. Tentative identification of the
tetanus toxin receptor in nervous tissue. J.Gen.
Microbiol. 20, 310–320 (1959).
This paper provides the first experimental evidence
that a ganglioside is involved in the neurospecific
binding of a clostridial neurotoxin.
42. Simpson,L.L. & Rapport,M.M. The binding of
botulinum toxin to membrane lipids: sphingolipids,
steroids and fatty acids. J.Neurochem. 18,
1751–1759 (1971).
43. Simons,K. & Toomre,D. Lipid rafts and signal
transduction. Nature Rev. Mol. Cell Biol. 1, 31–39
(2000).
44. Prinetti,A., Loberto,N., Chigorno,V. & Sonnino,S.
Glycosphingolipid behaviour in complex membranes.
Biochim. Biophys. Acta 1788, 184–193 (2009).
45. Chiba,A., Kusunoki,S., Shimizu,T. & Kanazawa,I.
Serum IgG antibody to ganglioside GQ1b is a possible
marker of Miller Fisher syndrome. Ann. Neurol. 31,
677–679 (1992).
46. Bullens,R.W. etal. Complex gangliosides at the
neuromuscular junction are membrane receptors for
autoantibodies and botulinum neurotoxin but
redundant for normal synaptic function. J.Neurosci.
22, 6876–6884 (2002).
47. Fogolari,F., Tosatto,S.C., Muraro,L. &
Montecucco,C. Electric dipole reorientation in the
interaction of botulinum neurotoxins with neuronal
membranes. FEBS Lett. 583, 2321–2325 (2009).
action. We expect that the number of novel BoNTs will
increase in the coming years, following the analysis of
more clinical and soil isolates. It will then be impor-
tant to determine the evolutionary processes that are
responsible for the generation of such diversity. From a
practical perspective, it is essential to develop stocks of
defined, highly specific antibodies and effective small-
molecule inhibitors to be used in the prevention and
treatment of botulism.
Note added in proof
A recent paper154 provides robust molecular evidence
that the BoNT/A complex trancytoses polarized epi-
thelial cells, disrupts E-cadherin cell–cell adhesion at
adherens junctions and so opens the paracellular route
of toxin entry into the body previously found for the
BoNT/B complex25.
REVIEWS
NATURE REV IEWS
|
MICROBIOLOGY ADVANCE ONLINE PUBLICATION
|
13
© 2014 Macmillan Publishers Limited. All rights reserved
48. Black,J.D. & Dolly,J.O. Interaction of 125I-labeled
botulinum neurotoxins with nerve terminals. II.
Autoradiographic evidence for its uptake into motor
nerves by acceptor-mediated endocytosis. J.Cell Biol.
103, 535–544 (1986).
49. Strotmeier,J. etal. Botulinum neurotoxin serotype D
attacks neurons via two carbohydrate-binding sites in
a ganglioside-dependent manner. Biochem. J. 431,
207–216 (2010).
50. Karalewitz,A.P., Fu,Z., Baldwin,M.R., Kim,J.J. &
Barbieri,J.T. Botulinum neurotoxin serotype C
associates with dual ganglioside receptors to facilitate
cell entry. J.Biol. Chem. 287, 40806–40816 (2012).
51. Strotmeier,J. etal. The biological activity of botulinum
neurotoxin type C is dependent upon novel types of
ganglioside binding sites. Mol. Microbiol. 81, 143–156
(2011).
52. Pirazzini,M., Rossetto,O., Bolognese,P., Shone,C.C.
& Montecucco,C. Double anchorage to the membrane
and intact inter-chain disulfide bond are required for
the low pH induced entry of tetanus and botulinum
neurotoxins into neurons. Cell. Microbiol. 13,
1731–1743 (2011).
53. Kitamura,M., Takamiya,K., Aizawa,S. & Furukawa,K.
Gangliosides are the binding substances in neural cells
for tetanus and botulinum toxins in mice. Biochim.
Biophys. Acta 1441, 1–3 (1999).
54. Yowler,B.C., Kensinger,R.D. & Schengrund,C.L.
Botulinum neurotoxin A activity is dependent upon the
presence of specific gangliosides in neuroblastoma
cells expressing synaptotagmin I. J.Biol. Chem. 277,
32815–32819 (2002).
55. Jacky,B.P.S. etal. Identification of fibroblast growth
factor receptor 3 (FGFR3) as a protein receptor for
botulinum neurotoxin serotype A (BoNT/A). PLoS
Pathog. 9, e1003369 (2013).
56. Nishiki,T. etal. Identification of protein receptor for
Clostridium botulinum type B neurotoxin in rat brain
synaptosomes. J.Biol. Chem. 269, 10498–10503
(1994).
This study is the first to identify a synaptic vesicle
protein receptor for a BoNT by showing that
BoNT/B binds to Syt.
57. Dong,M. etal. Synaptotagmins I and II mediate entry
of botulinum neurotoxin B into cells. J.Cell. Biol. 162,
1293–1303 (2003).
58. Rummel,A. etal. Identification of the protein receptor
binding site of botulinum neurotoxins B and G proves
the double-receptor concept. Proc. Natl Acad. Sci. USA
104, 359–364 (2007).
59. Peng,L. etal. Botulinum neurotoxin D-C uses
synaptotagmin I and II as receptors, and human
synaptotagmin II is not an effective receptor for type B,
D–C and G toxins. J.Cell Sci. 125, 3233–3242 (2012).
60. Berntsson,R.P., Peng,L., Svensson,L.M., Dong,M.
& Stenmark,P. Crystal structures of botulinum
neurotoxin dc in complex with its protein receptors
synaptotagmin I and II. Structure 21, 1602–1611
(2013).
61. Dong,M. etal. SV2 is the protein receptor for
botulinum neurotoxin A. Science. 31 2, 592–596
(2006).
62. Dong,M. etal. Glycosylated SV2A and SV2B mediate
the entry of botulinum neurotoxin E into neurons. Mol.
Biol. Cell 19, 5226–5237 (2008).
63. Mahrhold,S., Rummel,A., Bigalke,H., Davletov,B. &
Binz,T. The synaptic vesicle protein 2C mediates the
uptake of botulinum neurotoxin A into phrenic nerves.
FEBS Lett. 580, 2011–2014 (2006).
References 61, 62 and 63 report that the synaptic
vesicle protein SV2 functions as a protein receptor
for BoNT/A1 and BoNT/E1.
64. Mahrhold,S. etal. Identification of the SV2 protein
receptor-binding site of botulinum neurotoxin type E.
Biochem. J. 453, 37–47 (2013).
65. Benoit,R.M. etal. Structural basis for recognition of
synaptic vesicle protein 2C by botulinum neurotoxin A.
Nature 505, 108–111 (2014).
66. Schiavo,G. Structural biology: dangerous liaisons on
neurons. Nature 444, 1019–1020 (2006).
67. Colasante,C. etal. Botulinum neurotoxin type A is
internalized and translocated from small synaptic
vesicles at the neuromuscular junction. Mol.
Neurobiol. 48, 120–127 (2013).
68. Harper,C.B. etal. Dynamin inhibition blocks
botulinum neurotoxin type A endocytosis in neurons
and delays botulism. J.Biol. Chem. 286,
35966–35976 (2011).
69. Takamori,S. etal. Molecular anatomy of a trafficking
organelle. Cell 127, 831–846 (2006).
This paper provides a landmark analysis of the fine
structure and molecular composition of synaptic
vesicles.
70. Saheki,Y. & De Camilli,P. Synaptic vesicle endocytosis.
Cold Spring Harb. Perspect. Biol. 4, a005645 (2012).
71. Wohlfarth,K., Goschel,H., Frevert,J., Dengler,R. &
Bigalke,H. Botulinum A toxins: units versus units.
Naunyn. Schmiedebergs. Arch. Pharmacol. 355,
335–340 (1997).
72. Rasetti-Escargueil,C., Liu,Y., Rigsby,P., Jones,R.G. &
Sesardic,D. Phrenic nerve hemidiaphragm as a highly
sensitive replacement assay for determination of
functional botulinum toxin antibodies. Toxicon 57,
1008–1016 (2011).
73. Sun,S., Tepp,W.H., Johnson,E.A. & Chapman,E.R.
Botulinum neurotoxins B and E translocate at different
rates and exhibit divergent responses to GT1b and low
pH. Biochemistry 51, 5655–5662 (2012).
74. Ahnert-Hilger,G., Holtje,M., Pahner,I., Winter,S. &
Brunk,I. Regulation of vesicular neurotransmitter
transporters. Rev. Physiol. Biochem. Pharmacol. 150,
140–160 (2003).
75. Simpson,L.L., Coffield,J.A. & Bakry,N. Inhibition of
vacuolar adenosine triphosphatase antagonizes the
effects of clostridial neurotoxins but not phospholipase
A2 neurotoxins. J.Pharmacol. Exp. Ther. 269,
256–262 (1994).
76. Williamson,L.C. & Neale,E.A. Bafilomycin A1 inhibits
the action of tetanus toxin in spinal cord neurons in
cell culture. J.Neurochem. 63, 2342–2345 (1994).
References 75 and 76 show that the acidification
of an intracellular compartment by the vesicular
ATPase proton pump is a necessary step in nerve
intoxication by clostridial neurotoxins.
77. Sun,S. etal. Receptor binding enables botulinum
neurotoxin B to sense low pH for translocation channel
assembly. Cell Host Microbe 10 , 237–247 (2011).
78. Montal,M. Botulinum neurotoxin: a marvel of protein
design. Annu. Rev. Biochem. 79, 591–617 (2010).
79. Fischer,A. Synchronized chaperone function of
botulinum neurotoxin domains mediates light chain
translocation into neurons. Curr. Top. Microbiol.
Immunol. 364, 115–137 (2013).
80. Hoch,D.H. etal. Channels formed by botulinum,
tetanus, and diphtheria toxins in planar lipid bilayers:
relevance to translocation of proteins across membranes.
Proc. Natl Acad. Sci. USA 82, 1692–1696 (1985).
This is the first study to describe the formation of
ion channels by clostridial neurotoxins in planar
lipid bilayers.
81. Donovan,J.J. & Middlebrook,J.L. Ion-conducting
channels produced by botulinum toxin in planar lipid
membranes. Biochemistry 25, 2872–2876 (1986).
82. Blaustein,R.O., Germann,W.J., Finkelstein,A. &
DasGupta,B.R. The N-terminal half of the heavy chain
of botulinum type A neurotoxin forms channels in
planar phospholipid bilayers. FEBS Lett. 226, 115–120
(1987).
83. Koriazova,L.K. & Montal,M. Translocation of
botulinum neurotoxin light chain protease through the
heavy chain channel. Nature Struct. Biol. 10, 13–18
(2003).
84. Fischer,A. & Montal,M. Crucial role of the disulfide
bridge between botulinum neurotoxin light and heavy
chains in protease translocation across membranes.
J.Biol. Chem. 282, 29604–29611 (2007).
This study shows that the disulphide bond that
connects the L chain and H chain of BoNT/A1 and
BoNT/E1 must be reduced on the cytosolic side of
the synaptic vesicle to release the L chain
metalloprotease into the cytosol.
85. Fischer,A. & Montal,M. Single molecule detection of
intermediates during botulinum neurotoxin
translocation across membranes. Proc. Natl Acad. Sci.
USA 104, 10447–10452 (2007).
86. Sheridan,R.E. Gating and permeability of ion
channels produced by botulinum toxin types A and E in
PC12 cell membranes. Toxicon 36, 703–717 (1998).
87. Dalla Serra,M. etal. Conductive properties and gating
of channels formed by syringopeptin 25A, a bioactive
lipodepsipeptide from Pseudomonas syringae pv.
syringae, in planar lipid membranes. Mol. Plant.
Microbe Interact. 12, 401–409 (1999).
88. Fischer,A. etal. Molecular architecture of botulinum
neurotoxin E revealed by single particle electron
microscopy. J.Biol. Chem. 283, 3997–4003 (2008).
89. Bade,S. etal. Botulinum neurotoxin type D enables
cytosolic delivery of enzymatically active cargo proteins
to neurones via unfolded translocation intermediates.
J.Neurochem. 91, 1461–1472 (2004).
90. Galloux,M. etal. Membrane Interaction of botulinum
neurotoxin A translocation (T) domain. The belt region
is a regulatory loop for membrane interaction. J.Biol.
Chem. 283, 27668–27676 (2008).
91. Fischer,A. etal. Bimodal modulation of the botulinum
neurotoxin protein-conducting channel. Proc. Natl
Acad. Sci. USA 106, 1330–1335 (2009).
92. Pirazzini,M. etal. Neutralisation of specific surface
carboxylates speeds up translocation of botulinum
neurotoxin type B enzymatic domain. FEBS Lett. 587,
3831–3836 (2013).
93. Schiavo,G. etal. Tetanus and botulinum-B neurotoxins
block neurotransmitter release by proteolytic cleavage
of synaptobrevin. Nature 359, 832–835 (1992).
This study shows that VAMP has an essential role
in neurotransmitter release and that both tetanus
toxin and BoNT/B cleave the same protein at the
same site, despite the different clinical symptoms
that they cause.
94. Schiavo,G., Papini,E., Genna,G. & Montecucco,C.
An intact interchain disulfide bond is required for the
neurotoxicity of tetanus toxin. Infect. Immun. 58,
4136–4141 (1990).
95. de Paiva,A. etal. A role for the interchain disulfide or
its participating thiols in the internalization of
botulinum neurotoxin A revealed by a toxin derivative
that binds to ecto-acceptors and inhibits transmitter
release intracellularly. J.Biol. Chem. 268,
20838–20844 (1993).
96. Eswaramoorthy,S., Kumaran,D., Keller,J. &
Swaminathan,S. Role of metals in the biological
activity of Clostridium botulinum neurotoxins.
Biochemistry 43, 2209–2216 (2004).
97. Fu,F.N., Busath,D.D. & Singh,B.R. Spectroscopic
analysis of low pH and lipid-induced structural changes
in type A botulinum neurotoxin relevant to membrane
channel formation and translocation. Biophys. Chem.
99, 17–29 (2002).
98. Puhar,A., Johnson,E.A., Rossetto,O. & Montecucco,C.
Comparison of the pH-induced conformational change
of different clostridial neurotoxins. Biochem. Biophys.
Res. Commun. 319, 66–67 (2004).
99. Pirazzini,M. etal. Time course and temperature
dependence of the membrane translocation of tetanus
and botulinum neurotoxins C and D in neurons.
Biochem. Biophys. Res. Commun. 430, 38–42 (2013).
100. Miesenbock,G., De Angelis,D.A. & Rothman,J.E.
Visualizing secretion and synaptic transmission with
pH-sensitive green fluorescent proteins. Nature 394,
192–195 (1998).
101. Sankaranarayanan,S. & Ryan,T.A. Real-time
measurements of vesicle-SNARE recycling in synapses
of the central nervous system. Nature Cell Biol. 2,
197–204 (2000).
102. Eisenberg,M., Gresalfi,T., Riccio,T. & McLaughlin,S.
Adsorption of monovalent cations to bilayer
membranes containing negative phospholipids.
Biochemistry 18, 5213–5223 (1979).
103. Nordera,P., Serra,M.D. & Menestrina,G. The
adsorption of Pseudomonas aeruginosa exotoxin A to
phospholipid monolayers is controlled by pH and
surface potential. Biophys. J. 73, 1468–1478
(1997).
104. Deutsch,J.W. & Kelly,R.B. Lipids of synaptic vesicles:
relevance to the mechanism of membrane fusion.
Biochemistry 20, 378–385 (1981).
105. Ledeen,R.W., Diebler,M.F., Wu,G., Lu,Z.H. &
Varoqui,H. Ganglioside composition of subcellular
fractions, including pre- and postsynaptic membranes,
from Torpedo electric organ. Neurochem. Res. 18,
1151–1155 (1993).
106. Bychkova,V.E., Pain,R.H. & Ptitsyn,O.B. The
‘molten globule’ state is involved in the translocation of
proteins across membranes. FEBS Lett. 238, 231–234
(1988).
107. Ptitsyn,O.B., Pain,R.H., Semisotnov,G.V.,
Zerovnik,E. & Razgulyaev,O.I. Evidence for a molten
globule state as a general intermediate in protein
folding. FEBS Lett. 262, 20–24 (1990).
108. van der Goot,F.G., Gonzalez-Manas,J.M., Lakey,J.H.
& Pattus,F. A ‘molten-globule’ membrane-insertion
intermediate of the pore-forming domain of colicin A.
Nature 354, 408–410 (1991).
This paper provides the first evidence that a
bacterial toxin adopts a molten globular state
during membrane translocation.
109. Kukreja,R. & Singh,B. Biologically active novel
conformational state of botulinum, the most poisonous
poison. J.Biol. Chem. 280, 39346–39352 (2005).
110 . Meyer,Y., Buchanan,B.B., Vignols,F. &
Reichheld,J.P. Thioredoxins and glutaredoxins:
unifying elements in redox biology. Annu. Rev. Genet.
43, 335–367 (2009).
REVIEWS
14
|
ADVANCE ONLINE PUBLICATION www.nature.com/reviews/micro
© 2014 Macmillan Publishers Limited. All rights reserved
111. Hanschmann,E.M., Godoy,J.R., Berndt,C.,
Hudemann,C. & Lillig,C.H. Thioredoxins,
glutaredoxins, and peroxiredoxins — molecular
mechanisms and health significance: from cofactors to
antioxidants to redox signaling. Antioxid. Redox
Signal. 19, 1539–1605 (2013).
112. Berndt,C., Lillig,C.H. & Holmgren,A. Thioredoxins
and glutaredoxins as facilitators of protein folding.
Biochim. Biophys. Acta 1783, 641–650 (2008).
113. Pirazzini,M. etal. The thioredoxin reductase–
thioredoxin system is involved in the entry of tetanus
and botulinum neurotoxins in the cytosol of nerve
terminals. FEBS Lett. 587, 150–155 (2013).
This study provides the first evidence that the
thioredoxin reductase–thioredoxin protein disul-
phide-reducing system reduces the inter-chain
disulphide bond of clostridial neurotoxins in the
neuronal cytosol.
114. Dekker,C., Willison,K.R. & Taylor,W.R. On the
evolutionary origin of the chaperonins. Proteins 79,
1172–1192 (2011).
115. Sudhof,T.C. & Rizo,J. Synaptic vesicle exocytosis.
Cold Spring Harb. Perspect. Biol. 3, a005637(2011).
116. Pantano,S. & Montecucco,C. The blockade of the
neurotransmitter release apparatus by botulinum
neurotoxins. Cell. Mol. Life Sci. 71 , 793–811 (2014).
117. Binz,T. Clostridial neurotoxin light chains: devices for
SNARE cleavage mediated blockade of
neurotransmission. Curr. Top. Microbiol. Immunol.
364, 139–157 (2013).
118. Hayashi,T. etal. Synaptic vesicle membrane fusion
complex: action of clostridial neurotoxins on assembly.
EMBO J. 13, 5051–5061 (1994).
This study shows that VAMP, SNAP25 and syntaxin
form a tight coiled-coil complex that is resistant to
proteolysis by tetanus and botulinum neurotoxins
and to SDS.
119. Sutton,R.B., Fasshauer,D., Jahn,R. & Brunger,A.T.
Crystal structure of a SNARE complex involved in
synaptic exocytosis at 2.4 Å resolution. Nature 395,
347–353 (1998).
This fundamental paper describes the atomic
coiled-coil structure of the SNARE complex and its
importance for neurotransmitter release.
120. Megighian,A. etal. Evidence for a radial SNARE
super-complex mediating neurotransmitter release at
the Drosophila neuromuscular junction. J.Cell Sci.
126, 3134–3140 (2013).
121. Kalb,S.R. etal. Discovery of a novel enzymatic
cleavage site for botulinum neurotoxin F5. FEBS Lett.
586, 109–115 (2012).
122. Schiavo,G., Shone,C.C., Rossetto,O., Alexander,F.C.
& Montecucco,C. Botulinum neurotoxin serotype F is
a zinc endopeptidase specific for VAMP/synaptobrevin.
J.Biol. Chem. 268, 11516–11519 (1993).
123. Whitemarsh,R.C. etal. Characterization of botulinum
neurotoxin a subtypes1 through 5 by investigation of
activities in mice, in neuronal cell cultures, and invitro.
Infect. Immun. 81, 3894–3902 (2013).
124. Wang,D. etal. Comparison of the catalytic properties
of the botulinum neurotoxin subtypes A1 and A5.
Biochim. Biophys. Acta 1834, 2722–2728 (2013).
125. Shoemaker,C.B. & Oyler,G.A. Persistence of
botulinum neurotoxin inactivation of nerve function.
Curr. Top. Microbiol. Immunol. 364, 179–196 (2013).
126. Whitemarsh,R.C., Tepp,W.H., Johnson,E.A. &
Pellett,S. Persistence of botulinum neurotoxin A
subtypes 1–5 in primary rat spinal cord cells. PLoS
ONE. 9, e90252 (2014).
127. Naumann,M. etal. Evidence-based review and
assessment of botulinum neurotoxin for the
treatment of secretory disorders. Toxicon 67,
141–152 (2013).
128. Wang,J. etal. A dileucine in the protease of botulinum
toxin A underlies its long-lived neuroparalysis: transfer
of longevity to a novel potential therapeutic. J.Biol.
Chem. 286, 6375–6385 (2011).
129. Guo,J., Pan,X., Zhao,Y. & Chen,S. Engineering
clostridia neurotoxins with elevated catalytic activity.
Toxicon 74c, 158–166 (2013).
130. Ma,L. etal. Single application of A2 NTX, a botulinum
toxin A2 subunit, prevents chronic pain over long
periods in both diabetic and spinal cord injury-induced
neuropathic pain models. J.Pharmacol. Sci. 119 ,
282–286 (2012).
131. Chen,S. & Barbieri,J.T. Engineering botulinum
neurotoxin to extend therapeutic intervention. Proc.
Natl Acad. Sci. USA 106, 9180–9184 (2009).
132. Wang,D. etal. Syntaxin requirement for Ca2+-
triggered exocytosis in neurons and endocrine cells
demonstrated with an engineered neurotoxin.
Biochemistry 50, 2711–2713 (2011).
133. Franciosa,G., Ferreira,J.L. & Hatheway,C.L.
Detection of type A, B, and E botulism neurotoxin
genes in Clostridium botulinum and other Clostridium
species by PCR: evidence of unexpressed type B toxin
genes in type A toxigenic organisms. J.Clin. Microbiol.
32, 1911–1917 (1994).
134. Luquez,C., Raphael,B.H. & Maslanka,S.E.
Neurotoxin gene clusters in Clostridium botulinum
type Ab strains. Appl. Environ. Microbiol. 75,
6094–6101 (2009).
135. Carter,A.T., Stringer,S.C., Webb,M.D. &
Peck,M.W. The type F6 neurotoxin gene cluster
locus of group II Clostridium botulinum has evolved
by successive disruption of two different ancestral
precursors. Genome Biol. Evol. 5, 1032–1037
(2013).
136. Dover,N. etal. Clostridium botulinum strain Af84
contains three neurotoxin gene clusters: BoNT/A2,
BoNT/F4 and BoNT/F5. PLoS ONE 8, e61205
(2013).
137. Jahn,R. & Fasshauer,D. Molecular machines
governing exocytosis of synaptic vesicles. Nature 490,
201–207 (2012).
138. Harlow,M.L. etal. Alignment of synaptic vesicle
macromolecules with the macromolecules in active
zone material that direct vesicle docking. PLoS ONE 8,
e69410 (2013).
139. Zhai,R.G. & Bellen,H.J. The architecture of the
active zone in the presynaptic nerve terminal. Physiol.
(Bethesda) 19, 262–270 (2004).
140. Kasai,H., Takahashi,N. & Tokumaru,H. Distinct initial
SNARE configurations underlying the diversity of
exocytosis. Physiol. Rev. 92, 1915–1964 (2012).
141. Chernomordik,L.V. & Kozlov,M.M. Mechanics of
membrane fusion. Nature Struct. Mol. Biol. 15,
675–68 3 (2008).
142. Middlebrook,J.L. & Brown,J.E. Immunodiagnosis
and immunotherapy of tetanus and botulinum
neurotoxins. Curr. Top. Microbiol. Immunol. 195,
89–122 (1995).
143. Fairweather,N.F., Lyness,V.A. & Maskell,D.J.
Immunization of mice against tetanus with fragments
of tetanus toxin synthesized in Escherichia coli. Infect.
Immun. 55, 2541–2545 (1987).
144. Byrne,M.P. & Smith,L.A. Development of vaccines
for prevention of botulism. Biochimie 82, 955–966
(2000).
145. Smith,L.A. Botulism and vaccines for its prevention.
Vaccine 27, D33–D39 (2009).
146. Karalewitz,A.P.-A. & Barbieri,J.T. Vaccines against
botulism. Curr. Opin. Microbiol. 15, 317–324
(2012).
147. Arnon,S.S., Schechter,R., Maslanka,S.E.,
Jewell,N.P. & Hatheway,C.L. Human botulism
immune globulin for the treatment of infant botulism.
N.Engl. J.Med. 354, 462–471 (2006).
148. Garcia-Rodriguez,C. etal. Molecular evolution of
antibody cross-reactivity for two subtypes of type A
botulinum neurotoxin. Nature Biotech. 25, 107–116
(2007).
149. Lou,J. etal. Affinity maturation of human botulinum
neurotoxin antibodies by light chain shuffling via
yeast mating. Protein Eng. Des. Sel. 23, 311–319
(2010).
150. Cheng,L.W., Stanker,L.H., Henderson,T.D., Lou,J.
& Marks,J.D. Antibody protection against botulinum
neurotoxin intoxication in mice. Infect. Immun. 77,
4305–4313 (2009).
151. Conway,J.O., Sherwood,L.J., Collazo,M.T.,
Garza,J.A. & Hayhurst,A. Llama single domain
antibodies specific for the 7 botulinum neurotoxin
serotypes as heptaplex immunoreagents. PLoS ONE 5,
e8818 (2010).
152. Thanongsaksrikul,J. & Chaicumpa,W. Botulinum
neurotoxins and botulism: a novel therapeutic
approach. Toxins (Basel) 3, 469–488 (2011).
153. Li,B. etal. Small molecule inhibitors as
countermeasures for botulinum neurotoxin
intoxication. Molecules 16, 202–220 (2011).
154. Lee, K. et al. Molecular basis for disruption of
E-cadherin adhesion by botulinum neurotoxin A
complex R. Science 344, 1405–1410 (2014).
Acknowledgements
The authors thank T. Binz, G. Franciosa, R. Kammerer, F.
Lista, M. Montal, S. Pellet and G. Schiavo for their comments.
The authors apologize to colleagues whose work could not be
cited owing to space limitations. Research in the authors’
laboratory is supported by University of Padova, Italy, Fon-
dazione CARIPARO, the Axonomics Poject of the Provincia
Autonoma di Trento and the Italian Ministry of Defence.
Competing interests statement
The authors declare no competing interests.
DATABASES
Protein Data Bank (PDB):
http://www.rcsb.org/pdb/home/home.do
3BTA | 1EPW | 3FFZ | 3VOC
ALL LINKS ARE ACTIVE IN THE ONLINE PDF
REVIEWS
NATURE REV IEWS
|
MICROBIOLOGY ADVANCE ONLINE PUBLICATION
|
15
© 2014 Macmillan Publishers Limited. All rights reserved
