Knockin mouse models demonstrate differential contributions of synaptotagmin-1 and -2 as receptors for botulinum neurotoxins.

Botulinum neurotoxins (BoNTs) are the most potent toxins known and are also utilized to treat a wide range of disorders including muscle spasm, overactive bladder, and pain. BoNTs' ability to target neurons determines their specificity, potency, and therapeutic efficacy. Homologous synaptic vesicle membrane proteins synaptotagmin-1 (Syt1) and synaptotagmin-2 (Syt2) have been identified as receptors for BoNT family members including BoNT/B, DC, and G, but their contributions at physiologically relevant toxin concentrations in vivo have yet to be validated and established. Here we generated two knockin mutant mouse models containing three designed point-mutations that specifically disrupt BoNT binding in endogenous Syt1 or Syt2, respectively. Utilizing digit abduction score assay by injecting toxins into the leg muscle, we found that Syt1 mutant mice showed similar sensitivity as the wild type mice, whereas Syt2 mutant mice showed reduced sensitivity to BoNT/B, DC, and G, demonstrating that Syt2 is the dominant receptor at skeletal neuromuscular junctions. We further developed an in vivo bladder injection assay for analyzing BoNT action on bladder tissues and demonstrated that Syt1 is the dominant toxin receptor in autonomic nerves controlling bladder tissues. These findings establish the critical role of protein receptors for the potency and specificity of BoNTs in vivo and demonstrate the differential contributions of Syt1 and Syt2 in two sets of clinically relevant target tissues.

Characterization of a membrane binding loop leads to engineering botulinum neurotoxin B with improved therapeutic efficacy.

Botulinum neurotoxins (BoNTs) are a family of bacterial toxins with seven major serotypes (BoNT/A-G). The ability of these toxins to target and bind to motor nerve terminals is a key factor determining their potency and efficacy. Among these toxins, BoNT/B is one of the two types approved for medical and cosmetic uses. Besides binding to well-established receptors, an extended loop in the C-terminal receptor-binding domain (HC) of BoNT/B (HC/B) has been proposed to also contribute to toxin binding to neurons by interacting with lipid membranes (termed lipid-binding loop [LBL]). Analogous loops exist in the HCs of BoNT/C, D, G, and a chimeric toxin DC. However, it has been challenging to detect and characterize binding of LBLs to lipid membranes. Here, using the nanodisc system and biolayer interferometry assays, we find that HC/DC, C, and G, but not HC/B and HC/D, are capable of binding to receptor-free lipids directly, with HC/DC having the highest level of binding. Mutagenesis studies demonstrate the critical role of consecutive aromatic residues at the tip of the LBL for binding of HC/DC to lipid membranes. Taking advantage of this insight, we then create a "gain-of-function" mutant HC/B by replacing two nonaromatic residues at the tip of its LBL with tryptophan. Cocrystallization studies confirm that these two tryptophan residues do not alter the structure of HC/B or the interactions with its receptors. Such a mutated HC/B gains the ability to bind receptor-free lipid membranes and shows enhanced binding to cultured neurons. Finally, full-length BoNT/B containing two tryptophan mutations in its LBL, together with two additional mutations (E1191M/S1199Y) that increase binding to human receptors, is produced and evaluated in mice in vivo using Digit Abduction Score assays. This mutant toxin shows enhanced efficacy in paralyzing local muscles at the injection site and lower systemic diffusion, thus extending both safety range and duration of paralysis compared with the control BoNT/B. These findings establish a mechanistic understanding of LBL-lipid interactions and create a modified BoNT/B with improved therapeutic efficacy.

Massively parallel de novo protein design for targeted therapeutics.

De novo protein design holds promise for creating small stable proteins with shapes customized to bind therapeutic targets. We describe a massively parallel approach for designing, manufacturing and screening mini-protein binders, integrating large-scale computational design, oligonucleotide synthesis, yeast display screening and next-generation sequencing. We designed and tested 22,660 mini-proteins of 37-43 residues that target influenza haemagglutinin and botulinum neurotoxin B, along with 6,286 control sequences to probe contributions to folding and binding, and identified 2,618 high-affinity binders. Comparison of the binding and non-binding design sets, which are two orders of magnitude larger than any previously investigated, enabled the evaluation and improvement of the computational model. Biophysical characterization of a subset of the binder designs showed that they are extremely stable and, unlike antibodies, do not lose activity after exposure to high temperatures. The designs elicit little or no immune response and provide potent prophylactic and therapeutic protection against influenza, even after extensive repeated dosing.

Species-specific gene duplication in Clostridia produces variations of cholesterol-dependent cytolysin with different cytotoxicity.

Cholesterol-dependent cytolysin (CDC) is a bacterial toxin that binds to eukaryotic cholesterol-containing membranes, forms oligomeric complexes, and is inserted into the bilayer to create large aqueous pores. Recently, we reported a species-specific duplication of the hemolysin gene in group III Clostridium botulinum. The duplicated genes (bly1 and bly2) encoded two separate CDC proteins (botulinolysins; BLY1 and BLY2). Here, we aimed to investigate whether BLY1 and BLY2 exert differential cytotoxicity. We isolated two bly genes from C. botulinum and evaluated the cytotoxicity of two recombinant BLY proteins (rBLY1 and rBLY2) in HeLa, IEC-6, and NRK cells. rBLYs were cytotoxic to equine erythrocytes. rBLY1 showed higher hemolytic activity than rBLY2. rBLY2 showed no or very weak cytotoxicity to the HeLa, IEC-6, and NRK cells, whereas rBLY1 showed high cytotoxicity to these cells. The comparison of the amino acid sequence of BLYs with those of other CDCs revealed that the already-known amino acid residues involved in cholesterol-containing membrane recognition, oligomerization, and insertion into membranes are well conserved in both BLYs. However, several amino acid substitutions were observed in the conserved regions, particularly in L2 and L3 regions involved in cell binding. These findings suggest that gene duplication in group III C. botulinum evolved distinct functional specializations, and differential cytotoxicity of BLY1 and BLY2 could be due to the amino acid substitution in the conserved regions. However, the structural and functional comparisons of the two BLYs are essential to gain insights into the function of the CDCs.

Complete subunit structure of serotype C and D botulinum progenitor toxin complex induces vacuolation in the specific epithelial cell line.

Clostridium botulinum produces seven botulinum neurotoxin (BoNT) serotypes. In nature, BoNT exists as a part of the progenitor toxin complex (PTC) through associations with neurotoxin associated proteins (NAPs), including nontoxic nonhemagglutinin and hemagglutinin (HA) complex, consists of HA-70, HA-17 and HA-33. Because PTC displays higher oral toxicity than pure BoNTs, NAPs play a critical role in food poisoning. In a previous study, we demonstrated that the NAP complex in mature large-sized PTC (L-PTC) from serotypes C and D concomitantly induced cell death and cytoplasmic vacuolation in the rat intestinal epithelial cell line IEC-6. Here, we found that the serotype D NAP complex induces only cytoplasmic vacuolation in the normal rat kidney cell line NRK-52E without reducing cell viability. NAP complexes from serotype A and B L-PTCs did not affect cell viability or cytoplasmic vacuolation in IEC-6 and NRK-52E cells. Furthermore, we assessed the effect of immature L-PTCs with fewer HA-33/HA-17 trimers (two HA-33 and one HA-17) than mature L-PTCs on cell viability and cytoplasmic vacuolation in IEC-6 and NRK-52E cells. As a result, mature L-PTCs with the maximum number of HA-33/HA-17 trimers displayed the greatest potency. Consequently, the reduction in cell viability and vacuolation induction are related to the number of HA-33/HA-17 trimers in PTC. The discovery of an epithelial cell model where botulinum PTC specifically induces vacuolization may help clarify the unknown cytotoxicity of PTC, which plays an important role in the trans-epithelial transport of the toxin.

Sugar-induced conformational change found in the HA-33/HA-17 trimer of the botulinum toxin complex.

Large-sized botulinum toxin complex (L-TC) is formed by conjugation of neurotoxin, nontoxic nonhemagglutinin and hemagglutinin (HA) complex. The HA complex is formed by association of three HA-70 molecules and three HA-33/HA-17 trimers, comprised of a single HA-17 and two HA-33 proteins. The HA-33/HA-17 trimer isolated from serotype D L-TC has the ability to bind to and penetrate through the intestinal epithelial cell monolayer in a sialic acid-dependent manner, and thus it plays an important role in toxin delivery through the intestinal cell wall. In this study, we determined the solution structure of the HA-33/HA-17 trimer by using small-angle X-ray scattering (SAXS). The SAXS image of HA-33/HA-17 exhibited broadly similar appearance to the crystal image of the complex. On the other hand, in the presence of N-acetylneuraminic acid, glucose and galactose, the solution structure of the HA-33/HA-17 trimer was drastically altered compared to the structure in the absence of the sugars. Sugar-induced structural change of the HA-33/HA-17 trimer may contribute to cell binding and subsequent transport across the intestinal cell layer.

Random phage display-based screening of peptides that bind to botulinum neurotoxin binding protein, nontoxic nonhemagglutinin.

Botulinum neurotoxin (BoNT) binds to nontoxic nonhemagglutinin (NTNHA) protein in a pH-dependent manner, and yields the protease-resistant BoNT/NTNHA complex. Here, we screened short peptides that bind to the serotype D NTNHA (NTNHA-D) using random phage display technique. NTNHA was fixed onto electrode of quartz crystal microbalance (QCM) apparatus, and then the phages displaying random heptapeptides were exposed to the NTNHA-D under the acidic condition. After rinsing with acidic buffer, the released phages under the alkaline condition were collected. The binding and release of the phage were monitored by the frequency shift on the QCM. As a result of the screening, 16 were selected as peptides that bind to NTNHA-D. The selected peptides do not share any conserved sequence, but tend to be rich in basic and/or hydrophobic amino acid. This would explain the binding manner of the BoNT to the NTNHA protein.

Characterization of Serotype CD Mosaic Botulinum Neurotoxin in Comparison with Serotype C and A.

Botulinum neurotoxin (BoNT), produced by Clostridium botulinum, cleaves proteins involved in neurotransmitter release, thereby triggering flaccid paralyses, which are responsible for botulism. BoNT is classified into seven serotypes (BoNT/A-G); BoNT/A and BoNT/B are used as medical therapeutics and anti-wrinkle reagents. In this study, we investigated the efficacy of BoNT/CD, a mosaic toxin of BoNT/C and BoNT/D, to assess its potential as a therapeutic alternative for BoNT/A. In a cultured neuron assay, BoNT/CD cleaved syntaxin and SNAP-25 with higher efficacy than BoNT/C and BoNT/A. Intramuscularly administrated BoNT/CD induced dose-dependent muscle paralysis, and the paralysis lasted ~21 days in a mouse digit abduction score assay (BoNT/A-induced paralysis lasted ~30 days). BoNT/C failed to induce local paralysis without systemic toxicity. Multiple alignment analyses of the amino acid sequences of the receptor binding domain (HC) of eight BoNT/CDs and two BoNT/Ds showed sequence clustering in five groups. Comparing BoNT/CD strain 003-9 (BoNT/CD003-9) and strain 6813 (BoNT/CD6813) showed that both BoNT/CDs displayed similar efficacies in cultured neurons, but BoNT/CD003-9 displayed higher efficacy in a mouse model than BoNT/CD6813. These findings suggest that BoNT/CD may be a potential alternative for patients who do not respond to existing BoNT-based therapeutics.

Functional Deimmunization of Botulinum Neurotoxin Protease Domain via Computationally Driven Library Design and Ultrahigh-Throughput Screening.

Botulinum neurotoxin serotype A (BoNT/A) is a widely used cosmetic agent that also has diverse therapeutic applications; however, adverse antidrug immune responses and associated loss of efficacy have been reported in clinical uses. Here, we describe computational design and ultrahigh-throughput screening of a massive BoNT/A light-chain (BoNT/A-LC) library optimized for reduced T cell epitope content and thereby dampened immunogenicity. We developed a functional assay based on bacterial co-expression of BoNT/A-LC library members with a Förster resonance energy transfer (FRET) sensor for BoNT/A-LC enzymatic activity, and we employed high-speed fluorescence-activated cell sorting (FACS) to identify numerous computationally designed variants having wild-type-like enzyme kinetics. Many of these variants exhibited decreased immunogenicity in humanized HLA transgenic mice and manifested in vivo paralytic activity when incorporated into full-length toxin. One variant achieved near-wild-type paralytic potency and a 300% reduction in antidrug antibody response in vivo. Thus, we have achieved a striking level of BoNT/A-LC functional deimmunization by combining computational library design and ultrahigh-throughput screening. This strategy holds promise for deimmunizing other biologics with complex superstructures and mechanisms of action.

Toxic and nontoxic components of botulinum neurotoxin complex are evolved from a common ancestral zinc protein.

Zinc atoms play an essential role in a number of enzymes. Botulinum neurotoxin (BoNT), the most potent toxin known in nature, is a zinc-dependent endopeptidase. Here we identify the nontoxic nonhemagglutinin (NTNHA), one of the BoNT-complex constituents, as a zinc-binding protein, along with BoNT. A protein structure classification database search indicated that BoNT and NTNHA share a similar domain architecture, comprising a zinc-dependent metalloproteinase-like, BoNT coiled-coil motif and concanavalin A-like domains. Inductively coupled plasma-mass spectrometry analysis demonstrated that every single NTNHA molecule contains a single zinc atom. This is the first demonstration of a zinc atom in this protein, as far as we know. However, the NTNHA molecule does not possess any known zinc-coordinating motif, whereas all BoNT serotypes possess the classical HEXXH motif. Homology modeling of the NTNHA structure implied that a consensus K-C-L-I-K-X(35)-D sequence common among all NTNHA serotype molecules appears to coordinate a single zinc atom. These findings lead us to propose that NTNHA and BoNT may have evolved distinct functional specializations following their branching out from a common ancestral zinc protein.

Small-angle X-ray scattering reveals structural dynamics of the botulinum neurotoxin associating protein, nontoxic nonhemagglutinin.

In cell culture supernatants, the botulinum neurotoxin (BoNT) exists as part of a toxin complex (TC) in which nontoxic nonhemagglutinin (NTNHA) and/or hemagglutinins (HAs) are assembled onto the BoNT. A series of investigations indicated that formation of the TC is vital for delivery of the toxin to nerve cells through the digestive tract. In the assembly process, BoNT binds to NTNHA yielding M-TC, and it then matures into L-TC by further association with the HAs via NTNHA in the M-TC. Here, we report a crystal structure of the NTNHA from Clostridium botulinum serotype D strain 4947. Additionally, we performed small-angle X-ray scattering (SAXS) analysis of the NTNHA and the M-TC to elucidate the solution structure. The crystal structure of D-4947 NTNHA revealed that BoNT and NTNHA share a closely related structure consisting of three domains. The SAXS image indicated that, even though the N-terminal two-thirds of the NTNHA molecule had an apparently similar conformation in both the crystal and solution structures, the C-terminal third of the molecule showed a more extended structure in the SAXS image than that seen in the crystallographic image. The discrepancy between the crystal and solution structures implies a high flexibility of the C-terminal third domain of NTNHA, which is involved in binding to BoNT. Structural dynamics of the NTNHA molecule revealed by SAXS may explain its binding to BoNT to form the BoNT/NTNHA complex.

Crystallization and preliminary X-ray analysis of the Clostridium botulinum type D nontoxic nonhaemagglutinin.

Clostridium botulinum produces botulinum neurotoxin (BoNT) as a large toxin complex assembled with nontoxic nonhaemagglutinin (NTNHA) and/or haemagglutinin components. Complex formation with NTNHA is considered to be critical in eliciting food poisoning because the complex shields the BoNT from the harsh conditions in the digestive tract. In the present study, NTNHA was expressed in Escherichia coli and crystallized. Diffraction data were collected to 3.9 Å resolution. The crystal belonged to the trigonal space group P321 or P3(1)21/P3(2)21, with unit-cell parameters a = b = 147.85, c = 229.74 Å. The structure of NTNHA will provide insight into the assembly mechanism that produces the unique BoNT-NTNHA complex.

Delivery of single-domain antibodies into neurons using a chimeric toxin-based platform is therapeutic in mouse models of botulism.

Efficient penetration of cell membranes and specific targeting of a cell type represent major challenges for developing therapeutics toward intracellular targets. One example facing these hurdles is to develop post-exposure treatment for botulinum neurotoxins (BoNTs), a group of bacterial toxins (BoNT/A to BoNT/G) that are major potential bioterrorism agents. BoNTs enter motor neurons, block neurotransmitter release, and cause a paralytic disease botulism. Members of BoNTs such as BoNT/A exhibit extremely long half-life within neurons, resulting in persistent paralysis for months, yet there are no therapeutics that can inhibit BoNTs once they enter neurons. Here, we developed a chimeric toxin-based delivery platform by fusing the receptor-binding domain of a BoNT, which targets neurons, with the membrane translocation domain and inactivated protease domain of the recently discovered BoNT-like toxin BoNT/X, which can deliver cargoes across endosomal membranes into the cytosol. A therapeutic protein was then created by fusing a single-domain antibody (nanobody) against BoNT/A with the delivery platform. In vitro characterization demonstrated that nanobodies were delivered into cultured neurons and neutralized BoNT/A in neurons. Administration of this protein in mice shortened duration of local muscle paralysis, restoring muscle function within hours, and rescued mice from systemic toxicity of lethal doses of BoNT/A. Fusion of two nanobodies, one against BoNT/A and the other against BoNT/B, created a multivalent therapeutic protein able to neutralize both BoNT/A and BoNT/B in mice. These studies provide an effective post-exposure treatment for botulism and establish a platform for intracellular delivery of therapeutics targeting cytosolic proteins and processes.

Covalent Modifiers of Botulinum Neurotoxin Counteract Toxin Persistence.

Botulinum neurotoxins (BoNTs) are the most potent toxins known to man and a significant threat as weapons of bioterrorism. BoNTs contain a metalloprotease domain that blocks neurotransmitter release in nerve terminals, resulting in a descending, flaccid paralysis with a 5-10% mortality rate. Existing treatment options cannot access or neutralize the toxin following its endocytosis, so there is a clear need to develop novel therapies. Numerous substrate-based and zinc-chelating small-molecule inhibitors have been reported; however, none have progressed to the clinic. This is likely due to the difficulty that reversible inhibitors have in achieving sustained inhibition of the toxin, which has a half-life of months in vivo. An alternative strategy for mitigating BoNT persistence is covalent, irreversible inhibition of toxin function. However, few examples of covalent BoNT inhibitors have been reported. Here, we describe a competition-based screen to identify covalent modifiers of the conserved active-site-adjacent cysteine C165 in the BoNT/A serotype. We found that compounds containing cysteine-reactive electrophiles designed to target cysteine proteases failed to bind C165 while selenide compounds were efficient covalent binders of this cysteine. Importantly, covalent modification at C165 resulted in sustained, irreversible inhibition of BoNT/A protease activity. Covalent selenide inhibitors were nontoxic and protective in a neuronal assay of intoxication, making them promising new scaffolds for the study of the BoNT/A toxin as well as for the design of novel therapy agents.

Atomic force microscopic image data of botulinum neurotoxin complexes with different molecular sizes.

This data article provides atomic force microscopy (AFM) amplitude images of botulinum toxin complex (TC) molecules produced by Clostridium botulinum serotype D strain. C. botulinum produces different-sized TC molecules, such as a complex of botulinum neurotoxin and nontoxic nonhemagglutinin proteins (M-TC) and complex of M-TC and hemagglutinin subcomplex (L-TC). In this data article, the M and L-TC produced by serotype D strain 4947 were imaged by AFM. The M-TC molecule had a globular structure with a 30.5-nm diameter and a 2.1-nm height, while the L-TC molecule had a distinct structure in which several spheres were connected to a globular structure that was 40.7 nm in diameter and 3.5 nm in height.

Genetically encoded fluorescent indicators for imaging intracellular potassium ion concentration.

Potassium ion (K+) homeostasis and dynamics play critical roles in biological activities. Here we describe three genetically encoded K+ indicators. KIRIN1 (potassium (K) ion ratiometric indicator) and KIRIN1-GR are Förster resonance energy transfer (FRET)-based indicators with a bacterial K+ binding protein (Kbp) inserting between the fluorescent protein FRET pairs mCerulean3/cp173Venus and Clover/mRuby2, respectively. GINKO1 (green indicator of K+ for optical imaging) is a single fluorescent protein-based K+ indicator constructed by insertion of Kbp into enhanced green fluorescent protein (EGFP). These indicators are suitable for detecting K+ at physiologically relevant concentrations in vitro and in cells. KIRIN1 enabled imaging of cytosolic K+ depletion in live cells and K+ efflux and reuptake in cultured neurons. GINKO1, in conjunction with red fluorescent Ca2+ indicator, enable dual-color imaging of K+ and Ca2+ dynamics in neurons and glial cells. These results demonstrate that KIRIN1 and GINKO1 are useful tools for imaging intracellular K+ dynamics.

Sulfated glycosaminoglycans and low-density lipoprotein receptor contribute to Clostridium difficile toxin A entry into cells.

Clostridium difficile toxin A (TcdA) is a major exotoxin contributing to disruption of the colonic epithelium during C. difficile infection. TcdA contains a carbohydrate-binding combined repetitive oligopeptides (CROPs) domain that mediates its attachment to cell surfaces, but recent data suggest the existence of CROPs-independent receptors. Here, we carried out genome-wide clustered regularly interspaced short palindromic repeats (CRISPR)-CRISPR-associated protein 9 (Cas9)-mediated screens using a truncated TcdA lacking the CROPs, and identified sulfated glycosaminoglycans (sGAGs) and low-density lipoprotein receptor (LDLR) as host factors contributing to binding and entry of TcdA. TcdA recognizes the sulfation group in sGAGs. Blocking sulfation and glycosaminoglycan synthesis reduces TcdA binding and entry into cells. Binding of TcdA to the colonic epithelium can be reduced by surfen, a small molecule that masks sGAGs, by GM-1111, a sulfated heparan sulfate analogue, and by sulfated cyclodextrin, a sulfated small molecule. Cells lacking LDLR also show reduced sensitivity to TcdA, although binding between LDLR and TcdA are not detected, suggesting that LDLR may facilitate endocytosis of TcdA. Finally, GM-1111 reduces TcdA-induced fluid accumulation and tissue damage in the colon in a mouse model in which TcdA is injected into the caecum. These data demonstrate in vivo and pathological relevance of TcdA-sGAGs interactions, and reveal a potential therapeutic approach of protecting colonic tissues by blocking these interactions.

Identification of a Botulinum Neurotoxin-like Toxin in a Commensal Strain of Enterococcus faecium.

Botulinum neurotoxins (BoNTs), produced by various Clostridium strains, are a family of potent bacterial toxins and potential bioterrorism agents. Here we report that an Enterococcus faecium strain isolated from cow feces carries a BoNT-like toxin, designated BoNT/En. It cleaves both VAMP2 and SNAP-25, proteins that mediate synaptic vesicle exocytosis in neurons, at sites distinct from known BoNT cleavage sites on these two proteins. Comparative genomic analysis determines that the E. faecium strain carrying BoNT/En is a commensal type and that the BoNT/En gene is located within a typical BoNT gene cluster on a 206 kb putatively conjugative plasmid. Although the host species targeted by BoNT/En remains to be determined, these findings establish an extended member of BoNTs and demonstrate the capability of E. faecium, a commensal organism ubiquitous in humans and animals and a leading cause of hospital-acquired multi-drug-resistant (MDR) infections, to horizontally acquire, and possibly disseminate, a unique BoNT gene cluster.

Identification and characterization of a novel botulinum neurotoxin.

Botulinum neurotoxins are known to have seven serotypes (BoNT/A-G). Here we report a new BoNT serotype, tentatively named BoNT/X, which has the lowest sequence identity with other BoNTs and is not recognized by antisera against known BoNTs. Similar to BoNT/B/D/F/G, BoNT/X cleaves vesicle-associated membrane proteins (VAMP) 1, 2 and 3, but at a novel site (Arg66-Ala67 in VAMP2). Remarkably, BoNT/X is the only toxin that also cleaves non-canonical substrates VAMP4, VAMP5 and Ykt6. To validate its activity, a small amount of full-length BoNT/X was assembled by linking two non-toxic fragments using a transpeptidase (sortase). Assembled BoNT/X cleaves VAMP2 and VAMP4 in cultured neurons and causes flaccid paralysis in mice. Thus, BoNT/X is a novel BoNT with a unique substrate profile. Its discovery posts a challenge to develop effective countermeasures, provides a novel tool for studying intracellular membrane trafficking, and presents a new potential therapeutic toxin for modulating secretions in cells.

Reversible Association of the Hemagglutinin Subcomplex, HA-33/HA-17 Trimer, with the Botulinum Toxin Complex.

Botulinum neurotoxin (BoNT) associates with nontoxic proteins, either a nontoxic nonhemagglutinin (NTNHA) or the complex of NTNHA and hemagglutinin (HA), to form M- or L-toxin complexes (TCs). Single BoNT and NTNHA molecules are associated and form M-TC. A trimer of the 70-kDa HA protein (HA-70) attaches to the M-TC to form M-TC/HA-70. Further, 1-3 arm-like 33- and 17-kDa HA molecules (HA-33/HA-17 trimer), consisting of 1 HA-17 protein and 2 HA-33 proteins, can attach to the M-TC/HA-70 complex, yielding 1-, 2-, and 3-arm L-TC. In this study, the purified 1- and 2-arm L-TCs spontaneously converted into another L-TC species after acquiring the HA-33/HA-17 trimer from other TCs during long-term storage and freezing/thawing. Transmission electron microscopy analysis provided evidence of the formation of detached HA-33/HA-17 trimers in the purified TC preparation. These findings provide evidence of reversible association/dissociation of the M-TC/HA-70 complex with the HA-33/HA-17 trimers, as well as dynamic conversion of the quaternary structure of botulinum TC in culture.