ABSTRACT
Clostridium botulinum types C and D cause animal botulism by the production of serotype-specific or mosaic botulinum neurotoxin (BoNT). The D/C mosaic BoNT (BoNT/DC), which is produced by the isolate from bovine botulism in Japan, exhibits the highest toxicity to mice among all BoNTs. In contrast, rats appeared to be very resistant to BoNT/DC in type C and D BoNTs and their mosaic BoNTs. We attempted to characterize the enzymatic and receptor-binding activities of BoNT/DC by comparison with those of type C and D BoNTs (BoNT/C and BoNT/D). BoNT/DC and D showed similar toxic effects on cerebellar granule cells (CGCs) derived from the mouse, but the former showed less toxicity to rat CGCs. In recombinant murine-derived vesicle-associated membrane protein (VAMP), the enzymatic activities of both BoNTs to rat isoform 1 VAMP (VAMP1) were lower than those to the other VAMP homologues. We then examined the physiological significance of gangliosides as the binding components for types C and D, and mosaic BoNTs. BoNT/DC and C were found to cleave an intracellular substrate of PC12 cells upon the exogenous addition of GM1a and GT1b gangliosides, respectively, suggesting that each BoNT recognizes a different ganglioside moiety. The effect of BoNT/DC on glutamate release from CGCs was prevented by cholera toxin B-subunit (CTB) but not by a site-directed mutant of CTB that did not bind to GM1a. Bovine adrenal chromaffin cells appeared to be more sensitive to BoNT/DC than to BoNT/C and D. These results suggest that a unique mechanism of receptor binding of BoNT/DC may differentially regulate its biological activities in animals.
INTRODUCTION
Seven immunologically distinct types of botulinum neurotoxin (BoNT/A to G) cause botulism in humans and animals. This neurological disorder results from the inhibition of acetylcholine release at the neurotransmitter junctions of motor neuron endings. BoNT is produced by Clostridium botulinum as single-chain peptides with a molecular mass of about 150 kDa which are proteolytically activated into a light chain (approximately 50 kDa) and a heavy chain (approximately 100 kDa) linked by a disulfide bond. The heavy chain is further divided into two domains, an amino-terminal (HN) and a carboxyl-terminal (HC) domain. The light chain acts as a zinc-dependent endopeptidase and cleaves soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) protein family members (30, 37). Synaptosomal-associated protein with a molecular mass of 25 kDa (SNAP-25) and syntaxin, which predominantly reside on the plasma membrane, are cleaved by type A, C, and E BoNTs (BoNT/A, C, and E) and BoNT/C, respectively. Type B, D, F, and G BoNTs (BoNT/B, D, F, and G) act on the vesicular protein, synaptobrevin (vesicle-associated membrane protein, VAMP) (11). The HN is responsible for the translocation of the light chain from the lumen of an acidic intracellular compartment into the cytosol (10, 34). The HC exhibits serotype-specific binding activity to presynaptic membrane. Synaptic membrane proteins and/or gangliosides, sialic acid-containing glycosphingolipids, are the candidate receptors for BoNT. Synaptic vesicle glycoprotein SV2C and synaptotagmin I or II are considered to be the protein receptors for BoNT/A and B, respectively (8, 23). Meanwhile, for BoNT/C, gangliosides GD1b and GT1b have been identified as binding components (42). Although BoNT/D lacks the SXWY ganglioside-binding motif that is present in other BoNT serotypes, it has recently been reported that the toxin binds to b-series gangliosides (14).
In C. botulinum type C and D strains, which are associated with animal botulism, some varieties produce mosaic-formed BoNT. D/C mosaic BoNT (BoNT/DC) is comprised of two-thirds BoNT/D, including light chain and HN, and one-third BoNT/C, corresponding to the HC portion, while C/D mosaic BoNT (BoNT/CD) consists of the reverse structure from BoNT/DC (Fig. 1) (20). We previously reported that BoNT/DC is a pathogenic agent causing bovine botulism which exhibits the highest toxic activity to mice and possesses immunological specificity (21). However, the question still remains as to whether BoNT/DC possesses the characteristics in regard to high toxicity to bovines and mice. In this study, we further investigated the biological activities of BoNT/DC and compared the properties of the toxin with those of the other type C and D BoNTs.
Schematic representation of type C and D BoNTs. The homologies of amino acid sequences for L chain, HN, and HC are shown between bars. The regions that show more than 70% homology are indicated as the same pattern.
MATERIALS AND METHODS
Purification of neurotoxins and determination of toxicity to murine species.C. botulinum type C and D strains CB-19, 003-9, 1873, and OFD05, which produce BoNT/C, BoNT/CD, BoNT/D, and BoNT/DC, respectively, were used to purify each BoNT according to a previously described method (21). The toxicities of BoNTs were titrated by serial 2-fold dilution before intraperitoneal (i.p.) injection into SD rats (female, 8 weeks old, approximately 200 g; Japan SLC, Inc., Shizuoka, Japan), and ddY and C57BL/6J mice (male, 4 weeks old, approximately 20 g; Japan SLC, Inc.) to obtain the mean 50% lethal dose (LD50) by the Reed and Muench calculation (29). The toxicity of BoNT/DC to GM3 synthase-deficient (knockout) mice (42) was assayed by the intravenous injection method (21). The Ethics Committee of the Graduate School of Life and Environmental Sciences, Osaka Prefecture University, approved all animal tests.
Plasmid constructions and recombinant proteins.DNA fragments encoding the full-length sequence of the vamp1, vamp2, or vamp3 gene derived from mouse and rat were amplified by PCR. The primers used for construction of the vamp genes and rat isoform 1 VAMP (VAMP1) mutants are listed in Table 1. Gateway technology (Invitrogen, Carlsbad, CA) was used for their construction. VAMP homologue coding regions were cloned in the destination vector pDEST15 to generate the expression construct glutathione S-transferase (GST)-tagged proteins. The recombinant plasmids were introduced into Escherichia coli BL21 CodonPlus (DE3)-RIL (Stratagene, La Jolla, CA). Cultures were grown at 37°C in 300 ml of Luria-Bertani broth containing 100 μg/ml ampicillin until an optical density of 0.4 at 600 nm was reached. After the addition of isopropyl-β-d-thiogalactoside (Wako Pure Chemical Industries, Kyoto, Japan) at a final concentration of 1 mM, growth was continued at 16°C for an additional 24 h. The cells were collected, suspended in 12 ml of phosphate-buffered saline (140 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 1.8 mM KH2PO4, pH 7.3), and then solubilized on ice by sonication and centrifuged at 27,000 × g for 10 min at 4°C. The supernatant was subjected to purification of GST-tagged VAMP with glutathione-Sepharose 4B (GE Healthcare, Buckinghamshire, United Kingdom) as recommended by the instruction manual. Site-directed mutations were introduced into the rat vamp1 gene in pDEST15 by using a QuikChange site-directed mutagenesis kit (Stratagene) as instructed by the manufacturer. All mutated genes were sequenced and expressed in E. coli BL21 CodonPlus (DE3)-RIL in a manner similar to the procedure for the wild-type rat VAMP1. Recombinant HCs (HC from strain CB-19 BoNT [HC/CB-19 {HC/C}], amino acids [aa] 863 to 1291; HC/003-9, aa 863 to 1280; HC/1873, aa 859 to 1276 [HC/D]; and HC/OFD05, aa 857 to 1285 [HC/DC]), cholera toxin B subunit (CTB), and a site-directed mutant (G33D) of CTB (mCTB) were prepared according to previously reported methods (2, 7, 21).
Specific primers used for preparation of the VAMP homologues and mutant rat VAMP1
Cleavage assay of neurotoxin to recombinant VAMP homologues.Each concentration of BoNT/D and DC was reduced for 30 min at 37°C in 10 mM Tris-HCl buffer, pH 7.6, containing 5 mM dithiothreitol (DTT) and 20 mM NaCl (5). GST-tagged VAMP homologues from murine and rat VAMP1 mutants (1 μM) were incubated with neurotoxin for 1 h at 37°C in the same buffer without DTT. The reaction was terminated by the addition of SDS-PAGE sample buffer. The hydrolyzed samples were separated by SDS-PAGE on a 12.5% gel under reducing conditions by the method of Laemmli (15) and stained with Coomassie brilliant blue. Sample hydrolysis was monitored by analysis of densitometry scans (Scion Image; Scion Corporation, Frederick, MD).
TLC overlay assay.Lipids were extracted from lyophilized brain from C57BL/6J and GM3 synthase knockout mice. Thin-layer chromatography (TLC) and immunostaining were performed by a previously described method with slight modifications (12). Extracted lipids (corresponding to 20 μg of dry tissue weight) and gangliosides (GM3, GM1a, GD1a, GD1b, and GT1b; 50 pmol each) were developed on plastic-coated TLC plates (Marchery-Nagel, Duran, Germany) in chloroform–methanol–0.5% CaCl2 in water (55:45:10, vol/vol/vol), followed by blocking with HBS (3 mM HEPES, 150 mM NaCl, 2.5 mM KCl, 2 mM MgCl2, 2 mM CaCl2, pH 7.4), containing 1% bovine serum albumin. For the overlay assay, the plates were incubated with each BoNT (10 nM) in HBS containing 3% polyvinylpyrrolidone (HBS-PVP) for 2 h at room temperature. The bound BoNTs were probed with each affinity-purified polyclonal antibody against BoNT (1 μg/ml) and peroxidase-conjugated goat anti-rabbit IgG (Bio-Rad, Hertfordshire, United Kingdom) in HBS-PVP. Peroxidase on the plate was visualized with 4-chloro-1-naphthol (Wako) as a substrate.
Treatment of murine cerebellar granule, PC12, and bovine adrenal chromaffin cells with BoNT.Cerebellar granule neurons from 8-day-old C57BL/6J mice and SD rats were prepared according to a previously reported method (42). The cells were plated on polyethylenimine-coated 4-well culture dishes (Nalge Nunc, Rochester, NY) at a density of 2.6 × 105 cells/cm2 in HK-MEM (minimal essential medium containing 25 mM K+, 10 mM glucose, 30 nM Na2SeO4, 80 μg/ml gentamicin, pH 7.4) with 5% horse serum and 5% fetal calf serum and maintained overnight at 37°C in 5% CO2. The medium was then replaced with HK-MEM containing 2% B-27 supplement (Invitrogen) and 5 μM cytosine arabinofuranoside (serum-free HK-MEM), and the cells were incubated for 6 days without a further change of medium. BoNT/D or DC at various concentrations was added to cultures and incubated at 37°C for 20 min. After washing with serum-free HK-MEM, the cultures were incubated in the same medium for 18 h. Extracellular glutamic acid (Glu) release experiments were performed as described previously (42). The Glu content was determined using a high-performance liquid chromatography (HPLC)-electron capture detector (HTEC-500; Eicom, Kyoto, Japan), which consists of a GU-GEL matrix separation column, a Glu oxidase-immobilized column, and a platinum electrode for electrochemical detection. The concentration giving 50% inhibition (IC50) was calculated using Graph-Pad Prism software (GraphPad Software, San Diego, CA).
Rat adrenal pheochromocytoma PC12 cells were seeded into polyethylenimine-coated 12-well culture dishes (Asahi Glass Co., Chiba, Japan) at a density of 2.0 × 104 cells/cm2 in RPMI 1640 medium (RPMI 1640 [Wako], 0.3 μM Na2SeO4, and penicillin-streptomycin solution [Sigma, St. Louis, MO]) with 5% horse serum, 5% fetal calf serum at 37°C in 5% CO2 overnight. The medium was changed to RPMI 1640 medium, containing 50 ng/ml nerve growth factor (Sigma), 10 μg/ml transferrin (Sigma), 10 μg/ml insulin (Sigma), and 0.8 μM progesterone (Sigma) with or without ganglioside GM1a or GT1b (8 μg each; Sigma). After washing the cells with the same medium, BoNT/C or D or each of the two mosaic BoNTs (20 nM) was added to the culture. After incubation for 18 h, the cells were harvested. For the detection of syntaxin or VAMP2, cells were solubilized with Dulbecco's phosphate-buffered saline containing 0.5% Triton X-100, 0.05% SDS, and protease inhibitor cocktail (Sigma). The samples were separated by SDS-PAGE (12.5% acrylamide gel) and subjected to immunoblotting. After electrophoresis, proteins were transferred onto a polyvinylidene difluoride membrane, and the membrane was blocked with Blocking One (Nacalai Tesque, Kyoto, Japan). The membrane was then incubated with monoclonal anti-syntaxin (22), polyclonal anti-VAMP2 (N-terminal peptides 1 to 20), and anti-β-actin (Sigma) antibodies followed by peroxidase-conjugated goat anti-rabbit or anti-mouse IgG (Bio-Rad). The reactive bands were visualized by chemiluminescence using SuperSignal West Femto maximum sensitivity substrate (Thermo Fisher Scientific) and quantified with densitometry software (Scion Image).
Adrenal chromaffin cells were isolated from fresh bovine adrenal glands by collagenase digestion and cultured in collagen-coated 4-well dishes (Nalge Nunc) at a density of 2.6 × 105 cells/cm2 for at least 4 days (31). The cells were treated for 18 h with BoNT/C, D, or DC (10 nM each). Assays of total catecholamine (CA) and CA release from BoNT-treated cells were performed as previously described (31). The cells were washed twice with 2 ml of Lock's solution and then stimulated with acetylcholine (30 μM) for 5 min.
Protein concentrations.Protein concentrations were determined by the method of either Bradford (3) or Lowry et al. (18) with bovine gamma globulin or bovine serum albumin as the standard, respectively.
Statistical analysis.Significant differences were determined using the two-tailed paired Student's t test. Statistical significance was set at a P value of <0.05.
RESULTS
Toxicities of type C, type D, and mosaic BoNTs to mice and rats.The toxicities of BoNT/C and D and the two mosaic BoNTs to ddY mice and SD rats were titrated, and the specific activities per kilogram of body weight were calculated (Table 2). Upon comparison of these toxicities, BoNT/C and D exhibited lower levels of lethal activity in rats. The toxicity of BoNT/DC in rats was 2.2 × 102 i.p. LD50/mg protein, which was approximately 105-fold lower in terms of toxic activity than in mice.
Toxicity of type C, type D, and mosaic BoNTs to murines
To evaluate the different toxic activities of BoNT/DC in the two murines in vitro, we then prepared cerebellar granule cells (CGCs) from C57BL/6J mice and SD rats and examined the toxic effects of BoNT/DC and D on the CGCs. The two toxins showed similar levels of inhibition of the exocytotic release of glutamate to the CGCs of mice, with IC50 (50% inhibition concentration) values that were equivalent (Fig. 2, IC50s of 0.014 and 0.031 pM for BoNT/DC and D, respectively). In contrast, the inhibitory effect of BoNT/DC on CGCs of rats differed from that of BoNT/D (IC50s of 43 and 0.11 pM, respectively). When the amount of VAMP2 in CGCs from both murine species was detected by immunoblotting, both toxins also digested the intracellular VAMP2 of mouse CGCs in the same manner, while BoNT/DC-induced hydrolysis of VAMP2 in rat CGCs did not parallel that of BoNT/D (data not shown).
Inhibition of high-K+-induced exocytosis in BoNT-treated cerebellar granule cells from mice and rats. The cultured cerebellar granule cells were treated with increasing concentrations of BoNT/DC (open circle) or D (open triangle) at 37°C for 20 min. Glutamic acid (Glu) content, which was released from cells under low-K+ (5 mM) or high-K+ (50 mM) conditions, was determined by HPLC equipped with an electrochemical detector. Glu release under the high-K+ condition was adjusted for the release under the low-K+ condition and is presented as the percentage of the control value. Values are the means ± standard deviations from three independent experiments.
Enzymatic activities of type D and D/C mosaic BoNT on VAMP homologues from murine species.The light chain of BoNT/D was reported to exhibit cleavage activity to isoforms 1, 2, and 3 of VAMP (VAMP1, VAMP2, and VAMP3, respectively) (11), with cleavage at a single Lys-Leu peptide bond in rat VAMP1 and VAMP2 (33). After reduction of BoNT/DC and D with 5 mM DTT, we examined the proteolytic activities with recombinant GST-VAMP homologues prepared from murine cDNA. All GST-VAMP homologues were hydrolyzed to the same extent by BoNT/DC and D (data not shown). In the quantitative analysis of VAMP homologues that were cleaved by various concentrations of BoNT/DC, the toxin showed a level of enzymatic activity to rat VAMP1 that was lower than that for mouse VAMP1 (Fig. 3A and B). There were no significant differences between the amounts of mouse and rat VAMP2 or VAMP3 cleaved by the toxin. In the amino acid sequences of VAMP1 from mouse and rat, substitutions were produced in three residues. According to the sequences of mouse VAMP1, three single-point mutations of rat VAMP1 (T28M, I48M, and V78Q) were generated using the QuikChange kit. Among the three amino acid substitution variants, I48M was associated with a marked increase in the sensitivity, to a level similar to that of mouse VAMP1 (Fig. 3C and D).
Cleavage of murine VAMP homologues and mutant rat VAMP1 by BoNT/DC. (A, C) VAMP homologues and rat VAMP1 mutants were treated with the indicated concentrations of BoNT/DC. The hydrolyzed samples were separated by SDS-PAGE and stained with Coomassie brilliant blue. Representative cleavage patterns of VAMPs by BoNT/DC from a single independent experiment are shown. (B, D) Each intact VAMP from rat (solid line) and mouse (dashed line) was determined by densitometry scanning. The cleaved VAMP densities are presented as percentages of the control value. Values are the means ± standard deviations from three independent experiments.
Binding activity of D/C mosaic to gangliosides.In a previous study (42), we reported that HC/CB-19 and HC/1873 (HC/C and HC/D) bound to gangliosides GD1b and GT1b and to phosphatidylethanolamine, respectively. In addition, GM3 synthase knockout mice, which cannot synthesize GM3 ganglioside, the precursor for the biosynthesis of a-series and b-series gangliosides, were resistant to BoNT/C. We evaluated the toxicity of BoNT/DC to knockout mice and its binding activity to ganglioside with TLC-immunoblotting. The specific toxicity of BoNT/DC to C57BL/6J (wild-type) mice was equivalent to that obtained in ddY mice (data not shown). When BoNT/DC was injected intravenously into knockout mice, the survival time was increased compared with that in wild-type mice, resulting in a decrease in toxicity to 31% of the LD50 for the wild type (Table 3). In TLC immunostaining with the same amount of individual ganglioside (50 pmol), BoNT/DC seemed to bind strongly to ganglioside GM1a and exhibited low levels of affinity to GD1a, GD1b, and GM3 (Fig. 4A). BoNT/C reacted weakly to ganglioside GD1a, in addition to gangliosides GD1b and GT1b. The major brain gangliosides present in knockout mice have been expected to be gangliosides GM1b and GD1α of the α-series (44), whereas GD1a appears to be a major ganglioside in the whole brain of mammals (24). When we performed the same experiments using crude lipid extracts from mouse brain, BoNT/DC bound to lipids in wild-type or knockout mouse brain corresponding to the migration positions of gangliosides GM1a and GD1a or GM1b, respectively (Fig. 4B), whereas BoNT/C reacted to two lipids only in wild-type mouse brain, whose migration positions were identical to those of gangliosides GD1b and GT1b.
Sensitivities of wild-type and knockout mice to BoNT/DC
Direct binding of BoNT/DC and C to gangliosides. (A) Gangliosides (50 pmol) were developed on plastic-coated TLC plates in chloroform–methanol–0.5% CaCl2 in water (55:45:10, vol/vol/vol). The spots were stained with resorcinol reagent (left) or visualized by TLC-immunostaining with 10 nM BoNT/DC (center) or BoNT/C (right) and polyclonal antibody against each BoNT. (B) The lipid extracts from mouse brain were chromatographed and detected with BoNT by TLC-immunoblotting assay or resorcinol reagent. Mixtures of gangliosides GM1a, GD1a, and GT1b or GM1b and GD1b were used as standards. Each independent experiment was performed in triplicate, and representative binding patterns of BoNT/DC and C are shown.
Role of ganglioside GM1a in the entry of D/C mosaic BoNT into PC12 and cultured cerebellar granule cells.To determine the requirement for ganglioside as the functional receptor for BoNT/DC, we first used rat adrenal pheochromocytoma PC12 cells as a model. When PC12 cells were differentiated by nerve growth factor in serum-free medium containing transferrin, insulin, and progesterone, we found no morphological change and cleavage of the substrates in cells by the treatment with BoNT/C and D and the two mosaic BoNTs (20 nM each). Since BoNT/DC binds to ganglioside GM1a integrated into liposome in the liquid phase (25), BoNT was added after washing the ganglioside-loaded culture to avoid the binding of the toxin to exogenous ganglioside in the medium. BoNT/DC but not C affected the hydrolysis of substrates by the exogenous addition of ganglioside GM1a (Fig. 5). In contrast, ganglioside GT1b was required for the entry of BoNT/C but not DC into the cell. BoNT/D and/CD failed to cleave the intracellular substrate in the presence of either ganglioside GM1a or GT1b (data not shown).
Entry of BoNT/DC and C into ganglioside-treated PC12 cells. (A) Representative immunoblotting data show the effects of BoNTs on the substrate in PC12 cells. Rat adrenal pheochromocytoma PC12 cells were cultured in RPMI 1640 medium containing nerve growth factor, 10 μg/ml transferrin, 10 μg/ml insulin, and 0.8 μM progesterone with or without ganglioside GM1a or GT1b. After washing, BoNT/DC or C (20 nM each) was added to the culture in the medium. Cells were further incubated for 18 h and solubilized. Syntaxin, VAMP2, and β-actin in solubilized sample were detected by immunoblotting. β-Actin served as a sample loading control. (B) The chemiluminescence intensities of syntaxin, VAMP2, and β-actin in the PC12 cells were quantitated, and the percent densities compared to that of the control (no BoNT or ganglioside; white bar) were plotted as means ± standard deviations of values from three independent experiments. ***, P < 0.001.
We then examined whether the receptor for BoNT/DC corresponds to the cholera toxin receptor by using a competition assay with various receptor-binding domains. Glu release from mouse CGCs was inhibited by the treatment with BoNT/DC and D (Fig. 2). Various recombinant proteins, type C and D HCs (HC/C, HC/D, and HC/OFD05 [HC/DC]), CTB, and mCTB, which is known to lose affinity to GM1a (19), were used as competitors for the toxic effect on the CGCs (Fig. 6). The toxic effect of BoNT/DC was reduced only in the presence of CTB (P < 0.001). These results suggest that ganglioside GM1a is a receptor component for BoNT/DC. The HC/C was found to compete with the effect of BoNT/D on the CGCs (P < 0.05).
Inhibition of the effect of BoNT on cerebellar granule cells (CGCs) of mice by recombinant HC, cholera toxin B-subunit (CTB), or site-directed mutant (G33D) of CTB (mCTB). Recombinant HC/CB-19, HC/1873, and HC/OFD05 (HC/C, HC/D, and HC/DC), CTB, and mCTB (10 nM each) were added to cultured CGCs, followed 60 min later by treatment with BoNT/DC (white bars) or D (gray bars) at a final concentration of 2 pM in the presence of these competitors. Glutamic acid (Glu) content from each cell culture was determined by the method described in the legend to Fig. 2. The data presented are the means ± standard deviations of five determinations from individual wells in three individual experiments. Asterisks indicate statistical significance (*, P < 0.05; **, P < 0.01; ***, P < 0.001) compared with the group that was treated with each BoNT.
Sensitivity of bovine adrenal chromaffin cells to type C, type D, and D/C mosaic BoNT.Intact bovine adrenal chromaffin cells were exposed to BoNT/C, D, and DC, and then catecholamine release from the cells was measured in conjunction with the monitoring of the cleavage of the substrates by immunoblotting. The treatment with BoNT/DC and C significantly decreased the exocytotic release of CA evoked by acetylcholine (Fig. 7; P < 0.001 and P < 0.05, respectively). The toxic effect of BoNT/CD on the cells was not observed (data not shown). As expected from the results of sequence alignment, recombinant GST-tagged bovine VAMP1 and VAMP2 showed sensitivities to BoNT/DC and D similar to those derived from mouse (data not shown). These data suggested that the neuronal cells derived from bovines show higher sensitivity to BoNT/DC than to other BoNTs.
Differential effects of BoNT/DC, C, and D on bovine adrenal chromaffin cells. Bovine adrenal chromaffin cells were incubated with each BoNT (10 nM) for 18 h and then stimulated with acetylcholine buffer (30 μM). The amounts of CA released are expressed as the percentages of total cellular CA content. Data are presented as the means ± standard deviations of 12 determinations from individual wells in three independent experiments. Asterisks indicate statistical significance (*, P < 0.05; ***, P < 0.001) compared with the group that was not treated with each BoNT.
DISCUSSION
Type A, B, E, and F strains are known to cause human botulism, whereas types C and D are the most common sources of botulism in animals. We have reported previously that BoNT/CD and DC are related to avian and bovine botulism, respectively (21, 40). However, there has been less information about the toxicities of type C, D, and mosaic BoNTs to various animal species. BoNT/DC exhibited significantly different toxicities in mice and rats (Table 2). We first examined the enzymatic activity of the toxin to murine-derived substrates to reveal the causes of differential toxicities. The poor cleavability of the toxin to rat VAMP1 was shown by a single amino acid exchange (Fig. 3). Regarding the toxicities of BoNT/C, CD, and D to mice and chickens, the previous data indicated that BoNT/D exhibited higher toxicity to mice than BoNT/C and/CD, but chickens showed resistance to BoNT/D (41). The substrate of type C toxin is syntaxin and SNAP-25, whereas type D cleaves synaptobrevin. Moreover, the amino acid sequences of chicken VAMP1 and VAMP2 are quite different from those of mouse and rat VAMPs, although Met48 in the sequence of mouse VAMP1 is conserved in that of chicken VAMP1. The low sensitivity of rat VAMP1 to BoNT/D cleavage has already been reported (43), but the present results indicate a potential relationship between the cleavage of VAMP1 and flaccid paralysis by BoNT/DC. VAMP1 and VAMP2 are two highly homologous isoforms of VAMP and exhibit different expression patterns in neurons. VAMP2 is the abundant synaptic vesicle protein in the brain, whereas VAMP1 is expressed in the spinal cord (16, 28). VAMP2 mediates Ca2+-triggered vesicle exocytosis in hippocampal neurons, and VAMP3 rescues synaptic transmission in VAMP2-deficient ones (6). On the other hand, a null mutation in VAMP1 has been characterized by a general lack of movement and wasting, eventually leading to death before weaning (26). Additionally, it has recently been reported that VAMP1 plays a critical role in synaptic vesicle exocytosis at the mouse neuromuscular synaptic junction (17). When the effects of neuromuscular transmission blockage by different types of BoNT were investigated quantitatively by the compound muscle action potential (CMAP) test with SD rats, this method was shown to be more sensitive than the mouse bioassay for BoNT/A, C, CD, and E but not for BoNT/B, D, and F (41). Since rat VAMP1 is also resistant to BoNT/B (32), the toxicity of BoNT, at least BoNT/B, D, and DC, in the two murine species may depend on the hydrolysis activity of VAMP1 rather than that of VAMP2. The cleavage of VAMP1 inside neuronal cells needs to be confirmed in further studies to clarify the importance of VAMP1 as a target of the toxic action.
In spite of the equal cleavage activities of BoNT/DC and D to VAMP homologues derived from murine and bovine species, differential toxicities of these BoNTs to rat CGCs and bovine chromaffin cells were observed (Fig. 2 and 7). Recent analyses based on the crystal structures of type C and D HCs have revealed the presence of ganglioside-binding sites in each BoNT (13, 14, 25). We investigated the role of ganglioside as the functional receptor for BoNT/C and D and the two mosaic BoNTs. BoNT/C required GT1b for the entry into PC12 cells (Fig. 5), and HC/C inhibited the effect of BoNT/D on mouse CGCs (Fig. 6). However, the treatment of PC12 by GT1b did not facilitate cleavage of VAMP2 in PC12 cells by BoNT/D. Although previous findings have demonstrated the need for ganglioside for toxicity of BoNT/D (27, 39) and BoNT/D shares the same ganglioside receptor requirement as BoNT/C in mouse CGCs, other components may be needed for the binding of BoNT/D to presynaptic membrane.
We have visualized the direct binding of BoNT/DC to ganglioside GM1a by TLC overlay assay (Fig. 4A). Previous work has shown that BoNT/DC binds GM1a in addition to several other gangliosides (13, 25). Unlike the toxic effect of BoNT/DC on mouse or rat CGCs, BoNT/DC showed no cleavage activity to the intracellular substrate in PC12 cells, an adrenal medulla-derived cultured cell line (Fig. 5). Furthermore, we found that BoNT/DC elicited no inhibitory effect on high-K+-induced dopamine release from the cells (data not shown). Differential effects of the toxin on the cells were assumed to be due to cell sensitivity, including cell membrane components related to the binding and entry of BoNT into the cell. BoNT/DC cleaved VAMP2 in PC12 cells by exogenous addition of GM1a (Fig. 5). This result provides further information that GM1a not only binds to BoNT/DC but also actually functions as a receptor for BoNT/DC. However, it remains to be elucidated whether some other ganglioside is needed for the entry of the toxin into neuronal cells. For instance, BoNT/DC affected the CA release from intact chromaffin cells (Fig. 7). The major ganglioside in chromaffin granule membranes of bovine adrenal medulla has been characterized as GM3 (36). Moreover, BoNT/DC has been shown to react only to GM1b in the lipid extract from the brains of GM3 synthase knockout mice to whom the toxin was still toxic (Fig. 4B and Table 3). Although it is still unclear whether binding of the toxin to GM3 and GM1b is related to the toxicity, BoNT/DC may be able to enter the chromaffin or neuronal cells via ganglioside-independent binding.
GM1a mediates the entry of cholera toxin, Escherichia coli heat-labile toxin, and tetanus neurotoxin (1, 4). In Fig. 6, it is suggested that BoNT/DC shared the receptor, especially GM1a, on mouse CGCs with cholera toxin. On the other hand, the ganglioside-binding site of HC/DC has little similarity to that of GM1a-binding toxins in terms of amino acid sequences involved directly in binding to GM1a (25). Only partial inhibition of BoNT/DC and D was obtained with the homologous HC fragment. We found that the dissociation rate between HC/DC and GM1a-integrated liposome was quite high in a surface plasmon resonance analysis (data not shown), whereas it has been reported that the affinity between CTB and the liposome is very strong (2), suggesting that there is a difference between the binding activities of these receptor-binding domains. BoNT/DC entry into neurons is not blocked by competition from BoNT/C or D, an indication that different receptors for these three serotypes are utilized in membrane binding. Serotype-specific BoNTs, which possess a common ganglioside-binding motif, are reported usually to recognize GT1b (9, 23, 35, 38). The interaction between BoNT/DC and GM1a is a unique characteristic different from other types of BoNTs.
In this study, the enzymatic and receptor-binding activities of BoNT/DC were characterized. Some new evidence is presented indicating that the cleavage activity of BoNT/DC to VAMP1 may be related to the toxicity to mice and rats and that BoNT/DC utilizes different neuronal surface receptors than BoNT/C or D. These characteristics may provide the key to explaining the different susceptibilities of animals. Since the biological distribution and sugar chain structure of mouse and rat GM1a gangliosides are not clear, it is difficult to explain the relationships between the binding activity of BoNT/DC to GM1a and the toxicities to the two murine species. Determination of the crystal structure of HC/DC suggests the presence of recognition sites for gangliosides and, possibly, other molecules (25). The receptor recognition mechanism of BoNT/DC has yet to be elucidated completely. Additional studies are needed to understand the unique activity of this toxin.
ACKNOWLEDGMENTS
We thank Masao Iwamori (Faculty of Science and Technology, Kinki University) for technical advice about gangliosides and for other useful suggestions.
This work was supported by Grant-in-Aid for Scientific Research, Research Fellowships of the Japan Society for the Promotion of Science (JSPS) for Young Scientists (no. 1181008400) and in part by Grant-in-Aid for Scientific Research from JSPS (no. 21380188).
FOOTNOTES
- Received 22 March 2012.
- Returned for modification 7 May 2012.
- Accepted 23 May 2012.
- Accepted manuscript posted online 4 June 2012.
- Copyright © 2012, American Society for Microbiology. All Rights Reserved.
