Person:

Dong, Min

Botulinum neurotoxins (BoNTs) include seven bacterial toxins (BoNT/A-G) that target presynaptic terminals and act as proteases cleaving proteins required for synaptic vesicle exocytosis. Here we identified synaptic vesicle protein SV2 as the protein receptor for BoNT/D. BoNT/D enters cultured hippocampal neurons via synaptic vesicle recycling and can bind SV2 in brain detergent extracts. BoNT/D failed to bind and enter neurons lacking SV2, which can be rescued by expressing one of the three SV2 isoforms (SV2A/B/C). Localization of SV2 on plasma membranes mediated BoNT/D binding in both neurons and HEK293 cells. Furthermore, chimeric receptors containing the binding sites for BoNT/A and E, two other BoNTs that use SV2 as receptors, failed to mediate the entry of BoNT/D suggesting that BoNT/D binds SV2 via a mechanism distinct from BoNT/A and E. Finally, we demonstrated that gangliosides are essential for the binding and entry of BoNT/D into neurons and for its toxicity in vivo, supporting a double-receptor model for this toxin.

Botulinum neurotoxins are highly toxic, and bind two receptors to achieve their high affinity and specificity for neurons. Here we present the first structure of a botulinum neurotoxin bound to both its receptors. We determine the 2.3 Å structure of a ternary complex of botulinum neurotoxin type B bound to both its protein receptor Synaptotagmin II and its ganglioside receptor GD1a. We show that there is no direct contact between the two receptors, and that the binding affinity towards Synaptotagmin II is not influenced by the presence of GD1a. The interactions of botulinum neurotoxin type B with the sialic acid 5 moiety of GD1a are important for the ganglioside selectivity. The structure demonstrates that the protein receptor and the ganglioside receptor occupy nearby but separate binding sites, thus providing two independent anchoring points.

Botulinum neurotoxins (BoNT/A-G) are well-known to act by blocking synaptic vesicle exocytosis. Whether BoNTs disrupt additional neuronal functions has not been addressed. Here we report that cleavage of syntaxin 1 (Syx 1) by BoNT/C and cleavage of SNAP-25 by BoNT/E both induce degeneration of cultured rodent and human neurons. Furthermore, although SNAP-25 cleaved by BoNT/A can still support neuron survival, it has reduced capacity to tolerate additional mutations and also fails to pair with syntaxin isoforms other than Syx 1. Syx 1 and SNAP-25 are well-known for mediating synaptic vesicle exocytosis, but we found that neuronal death is due to blockage of plasma membrane recycling processes that share Syx 1/SNAP-25 for exocytosis, independent of blockage of synaptic vesicle exocytosis. These findings reveal neuronal cytotoxicity for a subset of BoNTs and directly link Syx 1/SNAP-25 to neuron survival as the prevalent SNARE proteins mediating multiple fusion events on neuronal plasma membranes.

Botulinum neurotoxins (BoNT/A-G), the most potent toxins known, act by cleaving three SNARE proteins required for synaptic vesicle exocytosis. Previous studies on BoNTs have generally utilized the major SNARE homologues expressed in brain (VAMP2, syntaxin 1, and SNAP-25). However, BoNTs target peripheral motor neurons and cause death by paralyzing respiratory muscles such as the diaphragm. Here we report that VAMP1, but not VAMP2, is the SNARE homologue predominantly expressed in adult rodent diaphragm motor nerve terminals and in differentiated human motor neurons. In contrast to the highly conserved VAMP2, BoNT-resistant variations in VAMP1 are widespread across vertebrates. In particular, we identified a polymorphism at position 48 of VAMP1 in rats, which renders VAMP1 either resistant (I48) or sensitive (M48) to BoNT/D. Taking advantage of this finding, we showed that rat diaphragms with I48 in VAMP1 are insensitive to BoNT/D compared to rat diaphragms with M48 in VAMP1. This unique intra-species comparison establishes VAMP1 as a physiological toxin target in diaphragm motor nerve terminals, and demonstrates that the resistance of VAMP1 to BoNTs can underlie the insensitivity of a species to members of BoNTs. Consistently, human VAMP1 contains I48, which may explain why humans are insensitive to BoNT/D. Finally, we report that residue 48 of VAMP1 varies frequently between M and I across seventeen closely related primate species, suggesting a potential selective pressure from members of BoNTs for resistance in vertebrates.

Antibody treatment is currently the only available countermeasure for botulism, a fatal illness caused by flaccid paralysis of muscles due to botulinum neurotoxin (BoNT) intoxication. Among the seven major serotypes of BoNT/A-G, BoNT/A poses the most serious threat to humans because of its high potency and long duration of action. Prior to entering neurons and blocking neurotransmitter release, BoNT/A recognizes motoneurons via a dual-receptor binding process in which it engages both the neuron surface polysialoganglioside (PSG) and synaptic vesicle glycoprotein 2 (SV2). Previously, we identified a potent neutralizing antitoxin against BoNT/A1 termed ciA-C2, derived from a camelid heavy-chain-only antibody (VHH). In this study, we demonstrate that ciA-C2 prevents BoNT/A1 intoxication by inhibiting its binding to neuronal receptor SV2. Furthermore, we determined the crystal structure of ciA-C2 in complex with the receptor-binding domain of BoNT/A1 (HCA1) at 1.68 Å resolution. The structure revealed that ciA-C2 partially occupies the SV2-binding site on HCA1, causing direct interference of HCA1 interaction with both the N-glycan and peptide-moiety of SV2. Interestingly, this neutralization mechanism is similar to that of a monoclonal antibody in clinical trials, despite that ciA-C2 is more than 10-times smaller. Taken together, these results enlighten our understanding of BoNT/A1 interactions with its neuronal receptor, and further demonstrate that inhibiting toxin binding to the host receptor is an efficient countermeasure strategy.