Isolation and Characterization of the Novel Botulinum Neurotoxin A Subtype 6

Botulinum neurotoxins (BoNTs) have proved to be an effective treatment for a large number of neuropathic conditions. BoNTs comprise a large family of zinc metalloproteases, but BoNT/A1 is used nearly exclusively for pharmaceutical purposes. The genetic inactivation of a second BoNT gene in the native strain enabled expression and isolation of a single BoNT/A6 from cultures. Its characterization indicated that BoNT/A subtype A6 has a long duration of action comparable to A1, while it enters neurons faster and more efficiently and remains more localized after intramuscular injection. These characteristics of BoNT/A6 are of interest for potential use of BoNT/A6 as a novel BoNT-based therapeutic that is effective and has a fast onset, an improved safety profile, and a long duration of action. Use of BoNT/A6 as a pharmaceutical also has the potential to reveal novel treatment motifs compared to currently used treatments.

tration was determined to be 0.09 pM (95% confidence interval: 0.07 to 0.12) for A1 and 0.15 pM (95% confidence interval: 0.12 to 0.2) for A6 by analyzing the averages and standard deviations of each toxin concentration from the three independent assays. Similar data were observed with longer (up to 21 h) incubations (data not shown).
BoNT/A6 has a high potency in neuronal cells. Previous studies have described the potency of BoNT/A subtypes 1 through 5 in various cell models (15,16). Exposure of cultured mouse and rat primary spinal cord cells (MSC and RSC cells) to serial dilutions of BoNT/A6 resulted in a SNAP-25 cleavage pattern similar to that previously observed for BoNT/A2, with 50% SNAP-25 cleavage occurring at ϳ0.03 U/50 l/well (27 fM) after 48 h of toxin exposure (Fig. S4). The EC 50 of BoNT/A6 was also determined in human iPSC-derived neurons. As with other BoNT/A subtypes, exposure of hiPSC neurons resulted in a steep dose-response curve extending from 0% to 100% cleavage within less than 4 logs for cells exposed to toxin for 48 h. The EC 50 was calculated to be ϳ0.03 U/50 l/well (Fig. 3), which is about 20-fold more sensitive than was seen in a parallel experiment for BoNT/A1 (0.6 U/50 l/well). These data indicate that BoNT/A6 is significantly more potent in cultured neurons than BoNT/A1, especially in human neurons.  BoNT/A6 enters cells more efficiently than other BoNT/A subtypes. To examine the cell entry kinetics of BoNT/A6 compared to subtypes BoNT/A1 and /A2, hiPSCderived neurons were exposed to 67 pM of each of the toxins (Fig. 4). Cells were harvested at the indicated time points through 10 h postexposure. Lysates were examined for the amount of cleaved and uncleaved SNAP-25 by Western blotting and densitometry. The onset of SNAP-25 cleavage occurred more rapidly in cells exposed to BoNT/A6 and BoNT/A2 than BoNT/A1 (Fig. 4). Cells exposed to BoNT/A6 also experienced an average of 94% Ϯ 5.3% SNAP-25 cleavage at the final 10-hour time point, while cells exposed to BoNT/A1 and /A2 reached approximately 64% Ϯ 4% and 84% Ϯ 1.2% cleavage, respectively. While equimolar amounts of toxin were used in this assay, the three toxins had similar specific activities (5.6 pg/LD 50 for A1, 4.9 pg/LD 50 for A2, and 5.3 pg/LD 50 for A6); thus, the results also compared similar numbers of biologically active units. This indicates that BoNT/A6 has an earlier onset of activity in cultured neurons compared to BoNT/A1 and similar to or even faster than BoNT/A2. Since catalytic activities of the BoNT/A6 and A1 LCs are similar, the earlier onset of activity in cultured neurons is likely due to faster or more efficient cell entry.
BoNT/A6 has a similarly long persistence of activity in cultured rat and human neurons as BoNT/A1. A previous report has shown that LC activities of BoNT/A1, /A2, /A4, and /A5 persist in cultured primary rat spinal cord neurons for over 10 months, and that the recovery rate for uncleaved SNAP-25 in the intoxicated cells is steady but very low, reaching less than 50% after 10 months (28). To determine the duration of action of BoNT/A6 in cultured primary rodent neurons, primary rat spinal cord (RSC) cells, recovery from SNAP-25 cleavage after a pulse of toxin exposure was monitored over 8 months. Up to 6 months, only the BoNT/A cleavage product of SNAP-25 was detected by Western blotting. At 7 months after initial toxin exposure, about 90% of the detected SNAP-25 was cleaved and about 10% uncleaved, indicating beginning of recovery of the cells. After 8 months, about 79% of SNAP-25 remained cleaved (Fig. 5D). Although signs of SNAP-25 recovery were not seen until later compared to previous work on BoNT/A subtypes 1 to 5, the estimated slope of the line when signs of uncleaved SNAP-25 began was similar to that of BoNT/A1, A2, A4, and A5 previously observed, indicating similar persistence of LC activity within cultured neuronal cells (28).
Persistence of BoNT LC activity in human iPSC-derived neurons has previously been shown to be shorter than in primary rat spinal cord neurons, with dose-dependent recovery of the neurons within 3 to 5 months (16,28). This allowed for estimation of the half-life of activity in human neurons by exposing the neurons to serial dilutions of either BoNT/A1 or BoNT/A6 and determining the EC 50 at different time points posttoxin exposure. Western blotting and densitometry analyses of triplicate sets of cells with each dilution series harvested at days 3, 39, and 70 post-BoNT exposure resulted in EC 50 values for BoNT/A6 of ϳ0.04 U/50 l/well (day 3), 0.8 U/50 l/well (day 39), and 1 U/50 l/well (day 70) (or 28, 495, and 707 fM, respectively). EC 50 values of BoNT/A1 were ϳ0.42 U/50 l/well (day 3), 4.0 U/50 l/well (day 39), and 17 U/50 l/well (day 70) (or 314, 2,990, and 12,690 fM, respectively) ( Fig. 5A and B). The half-life of activity of BoNT/A1 and /A6 in these hiPSC-derived neurons was estimated from the EC 50 values over time and was similar for both BoNT/A1 and /A6, at approximately 12 days and 14 days, respectively (Fig. 5C). Taken together, these results indicate that the BoNT/A6 LC persists and is enzymatically active in neurons for a similarly long time as BoNT/A1 LCs.
Onset and duration of action of BoNT/A6 is similar to BoNT/A1 in mice. To determine the onset and duration of action of BoNT/A6 compared to BoNT/A1, mice were injected with dilutions of either BoNT ranging from 0.2 to 0.6 U into the right hind gastrocnemius muscle. DAS scores (to measure local paralysis) and Rotarod times (to measure overall motor neuron deficiency) were recorded for each mouse at several time points throughout the first 48 h and then once each day through 16 days postinjection. The injected doses of each toxin were confirmed by IP LD 50 assay using the same dilutions (data not shown). As previously seen for other BoNTs, the DAS scores and Rotarod recovery times were both dose dependent (Fig. 6). The peak DAS score appeared between 36 and 48 h, and it started to decrease at day 3 to 4, similar to previous reports for BoNT/A2 and slightly earlier than BoNT/A1 (Fig. 6A). By day 16, the DAS score had dropped to ϳ1.5 for mice injected with the highest toxin dilution (Fig. 6A), which is similar as previously observed with BoNT/A1, /A2, and /A5 (17). Overall motor neuron deficiency, as measured by Rotarod, was similar for both BoNT/A1 and /A6 (Fig. 6B) and similar to that previously observed for other BoNT/A subtypes (17). These data indicate similar onset and duration of action of BoNT/A6 compared to FIG 5 EC 50 and duration of action of BoNT/A6 and /A1 in neurons. Human iPSC-derived neurons were exposed to serial dilutions of BoNT/A6 (A) or /A1 (B) for 72 h, followed by BoNT removal. Cells were incubated further in toxin-free medium and harvested on days 3, 39, and 70 after exposure. Graphs generated in PRISM6 software depicting the average and standard deviation of triplicate samples are shown (A and B). The EC 50 values were determined in PRISM6 using a nonlinear regression (four parameters). The half-lives of BoNT/A1 and /A6 in hiPSCs were estimated by plotting the EC 50 values versus time and using the formula t 1/2 ϭ ln(2)/slope of regression line (C). The half-lives were similar, being ϳ12 and ϳ14 days for A1 and A6, respectively. The duration of action was also evaluated in RSCs (D). RSCs were exposed to 8 pM BoNT/A6 for 3 days, followed by complete toxin removal. Cells were incubated for a longer duration in toxin-free medium, and cell lysates were prepared and analyzed for cleaved/uncleaved SNAP-25 by Western blotting and densitometry monthly until 8 months post-toxin exposure. Averages and standard deviations of quadruplicate samples are shown.
BoNT/A1 for local paralysis in mice after local intramuscular injection, in addition to similar overall motor neuron deficiency.
BoNT/A6 remains more localized at the injection site after local intramuscular injection. BoNT/A2, which enters cultured neuronal cells faster and more efficiently, similar to BoNT/A6, has previously been suggested to remain more localized after intramuscular injection into rodent gastrocnemius muscle than BoNT/A1 (15,19). To determine if the faster cell entry of BoNT/A6 also leads to less toxin spread away from the injection site after local intramuscular injection, a relative IM LD 50 (relative to the IP LD 50 of the same toxin dilutions) in mice was determined for injection of BoNT/A1, /A2, and /A6 into the right gastrocnemius muscle. Toxin dilutions of each BoNT/A subtype were injected either intraperitoneally or intramuscularly in parallel, and mice were scored for survival. Most mice injected with lethal doses IP were found dead the day after injection, with a few dead at 2 days postinjection. However, the mice injected with lethal doses IM died slower, with deaths still occurring over a wide range of time postinjection. Using the Reed and Muench calculation methods, the LD 50 values were calculated to be 5.6, 4.9, and 5.3 pg for the mice injected IP and 7.9, 9.9, and 10.6 pg for mice injected IM with BoNT/A1, /A2, and /A6, respectively. Thus, the IM LD 50 for BoNT/A2 and BoNT/A6 both was ϳ2 IP LD 50 units whereas the IM LD 50 for BoNT/A1 was ϳ1.4 IP units (Table 1).

DISCUSSION
BoNT/A6 is naturally expressed in a dual-toxin-producing strain, CDC41370, isolated from a food-borne botulism case in 1996 (23). In addition to BoNT/A6, this strain also  expresses BoNT/B2, complicating the isolation and characterization of the BoNT/A6 toxin. As has been previously accomplished for the isolation of BoNT/FA from a bivalent strain (24), strain CDC41370 was modified to eliminate expression of BoNT/B2. BoNT/A6 was expressed at high levels in the resulting strain CDC41370B2 tox-, and purification using standard protocols existing for other BoNT/A subtypes resulted in isolation of an Ն95% pure BoNT/A6. As has been previously observed for BoNT/A2, BoNT/A6 had a specific activity similar to BoNT/A1 when injected IP in mice but had about 20-fold higher potency in all tested cell models, including primary rat and mouse spinal cord cultures and hiPSC-derived neurons (Fig. 3). BoNT/A6 had similar in vitro LC enzymatic activity as BoNT/A1 in the BoTest but showed significantly faster SNAP-25 cleavage in cultured neurons ( Fig. 2 and 4). This earlier onset of intracellular SNAP-25 cleavage combined with similar enzymatic activity of the LC indicates faster cell entry by BoNT/A6 than BoNT/A1 and similar entry to that previously observed for BoNT/A2 (5,20). These data are consistent with the previous observation that BoNT/A6 may be derived from a genetic recombination event of BoNT/A2 and other BoNT/A subtypes (23) and indicate that there may be common amino acid sequence and structural elements in the HC of BoNT/A6 and /A2 underlying the faster cell entry kinetics.
We have previously observed that BoNT/A3 has a significantly shorter duration of action than BoNT/A1. The duration of action of BoNT/A6 was examined both in cultured neurons and in mice as previously performed for other BoNT/A subtypes (17,28). Both the cell culture and in vivo mouse data indicate similar duration of action of BoNT/A6 and /A1 ( Fig. 5 and 6) (17,28). Interestingly, while persistence of BoNT/A1 and /A6 LC activity in cultured primary rodent neurons was very long, with only about 50% recovery after 10 months (Fig. 5D) (28), persistence of BoNT LC activity in cultured human iPSC-derived neurons was much shorter, with dose-dependent recovery of the neurons within 3 to 5 months, as has previously been shown for other BoNTs (16). This recovery time is similar to the recovery of paralysis of human muscle after local injection with pharmaceutical BoNT/A1. While the reasons for the difference in recovery time between the two cell models are not known, likely explanations include differences in differentiation stages of the cell models, differences in cell composition (Ͼ98% pure forebrain-like neurons versus a mixture of spinal cord neurons and glial cells), and the presence of glial cells in the primary neurons versus the pure neuronal population of the hiPSC-derived neurons. Furthermore, the different BoNTs may show different persistence because of different reactivities or mechanisms of BoNT degradation in cells such as ubiquitin-mediated protein degradation. Regardless of the underlying mechanism, this shorter recovery time enabled the estimation of the half-lives of BoNT/A1 and /A6 in human neurons, with the data indicating similar half-lives of ϳ12 and ϳ14 days for A1 and A6, respectively (Fig. 5C), which is consistent with a similar duration of action in primary rat spinal cord cells and in mice. Together, these data demonstrate a similar duration of action of BoNT/A6 and BoNT/A1, which is consistent with the high degree of amino acid identity of the A6 and A1 LCs, which differ by only one residue (T414A).
Faster cell entry of BoNT/A2 has been suggested to correlate with faster onset of local paralysis after intramuscular injection (15,(17)(18)(19)(20). Similar to what was previously observed for BoNT/A2, the onset of local paralysis as measured by DAS appeared slightly earlier for BoNT/A6 than for BoNT/A1 (Fig. 6A), although these observational data rely on qualitative and relatively small differences and will need to be confirmed in larger animals or in a clinical setting. Mice are relatively small animals compared to humans and other animals such as wildlife showing botulism in nature. Therefore, mice and other small animal models require local BoNT doses relatively close to lethal doses for observation of significant local paralysis (0.2 to 0.75 U), with 1 to 2 units being lethal after IM injection (Table 1). Due to the resulting relatively small effective dose range in mice, it may be difficult to quantitatively discern differences in the dose-dependent onset of paralysis time. For botulism in larger animals or in pharmaceutical treatments of humans, smaller relative doses can be injected locally and lead to paralysis, likely due to the larger size of the injected muscle and whole organisms being less prone to systemic distribution of toxin. Therefore, future studies in larger animals or clinical studies will be required to definitively determine whether BoNT/A6 and BoNT/A2 result in a significantly faster onset of pharmaceutical action than BoNT/A1. The faster and more efficient cell entry and possibly faster onset of action by BoNT/A6 and /A2 than BoNT/A1 raise the questions whether BoNT/A6 and /A2 have the potential to remain more localized within the injection site after in vivo intramuscular injection, which would result in fewer side effects during pharmaceutical use due to toxin spread away from the injection site. In vivo studies in mice comparing a pharmaceutical preparation of purified BoNT/A2 (with the excipients unknown) to pharmaceutical BoNT/A1 complex indicated a significant difference in the IM LD 50 for the two toxins (19). Here, we directly compared purified preparations of BoNT/A1, /A2, and /A6 for their lethality in mice after intraperitoneal and intramuscular injection. Importantly, for each subtype the same toxin dilutions and relative concentrations of active BoNT were used for the IM and IP injections, enabling determination of a relative IM LD 50 dose for each toxin. Based on the calculated LD 50 determined from a standard MBA with IP injections, IM injections of BoNT/A6 and BoNT/A2 required 2.0 times as much toxin to be lethal as the amount of toxin injected IP, while BoNT/A1 required only 1.4 times as much toxin (Table 1). These data support the previous observation that BoNT/A2 remains more localized after intramuscular injection than BoNT/A1 and indicate that, similar to BoNT/A2, BoNT/A6 also results in less systemic toxin distribution than for BoNT/A1 (19).
Taken together, our results demonstrate several properties of BoNT/A6 that would be beneficial if used as a new BoNT-based pharmaceutical, including faster and more efficient neuronal cell entry than BoNT/A1, similarly long duration of action, and less systemic spread after local injection. Treatments with BoNT/A6 may relieve targeted symptoms faster with fewer risks for side effects and without compromising the longlasting duration. This study is based on one batch of purified BoNT/A6, and future work with additional batches and pharmacologic preparations of BoNT/A6 will be required to assess safety and effectiveness of BoNT/A6 as a new pharmaceutical. These studies showing different phenotypes in cells and animals for subtypes also provide biological systems for understanding the molecular, cellular, and organismal basis for these different and important properties of BoNTs. Construction of a BoNT/B2 insertional mutant from the native BoNT/A6-producing C. botulinum strain. Strain CDC41370 expresses both BoNT/B2 and BoNT/A6 (23). The production of two BoNTs complicates the isolation of pure BoNT/A6. The gene encoding botulinum neurotoxin B2 (GenBank accession number FJ981697) in C. botulinum strain CDC41370 was inactivated with a ClosTron mutagenesis system by insertion of a mobile group II intron between nucleotide 381 and 382 on a sense strand using plasmid pMTL007C-E2::Cbo:bontbvb-381s as previously described for inactivation of the BoNT/B2 gene in C. botulinum strain CDC69016 (15,(24)(25)(26)(27)29). Inactivation of the BoNT/B2 gene was confirmed by PCR, DNA sequencing, and Southern hybridization as previously described using the same primers and probes (24). This resulted in strain CDC41370B2 tox-, which expressed exclusively BoNT/A6 at high levels. BoNT/A6 (ϳ150-kDa protein) was then purified from this modified strain, yielding highly pure toxin using the method previously described to produce BoNT/FA (15,26,27).

MATERIALS AND METHODS
Botulinum neurotoxins. BoNT/A1 and /A2 were independently purified (ϳ150-kDa proteins) from C. botulinum strains Hall A-hyper and Kyoto-F as previously described (15,26,27). The purity of the toxins was confirmed by spectroscopy and SDS-PAGE. The purified toxins were stored in 0.1 M sodium phosphate buffer, pH 7, with 40% glycerol at Ϫ20°C until use. Specific activities of each subtype preparation were determined using an intraperitoneal mouse bioassay (MBA) as previously described (30)(31)(32)(33). The specific activities of the BoNTs in mice were 5.6 pg/LD 50 (A1), 4.9 pg/LD 50 (A2), and Functional Characterization of BoNT/A6 5.3 pg/LD 50 (A6). Protein concentrations of purified BoNT/A1, /A2, and /A6 were determined by measuring the absorbance at A 278 and an extinction coefficient of 1.63 and by SDS-PAGE gel analysis.
BoTest. The in vitro SNARE fragment-leaving BoTest for A/E BoNTs (Biosentinel Pharmaceuticals) was used according to the manufacturer's instructions. Tenfold serial dilutions of BoNT/A1 and /A6 were prepared in 1ϫ reaction buffer solution containing 5 mM dithiothreitol (DTT) (Sigma) for toxin reduction. The data presented are the average from the results from three independent experiments performed on each subtype, each of which included a control without BoNT present. The plate was incubated at 30°C for 2 h, and emission ratios were determined after addition of reporter to the BoNT dilutions. A BioTek Synergy H1 Hybrid reader recorded the absorbance of each well. The 50% effective concentration (EC 50 ) was determined by analyzing the averages and standard deviations for each toxin concentration from the three independent assays using a nonlinear regression (four parameters) in PRISM 6.
Primary rat (RSC) and mouse spinal cord (MSC) cell assay. Primary rat and mouse spinal cord cells were prepared as previously described (28,34). The cells were plated on a 96-well, flat-bottom plate (Techno Plastic Products [TPP]) plate treated with 0.01% poly-L-ornithine (Sigma) and coated with 8.3 g/cm 2 growth factor-reduced Matrigel (BD Biosciences). The cells were maintained in culture medium (CM) (Neurobasal medium supplemented with B27, GlutaMAX, and penicillin/streptomycin [Life Technologies]) and allowed to mature for a minimum of 2 weeks. Cells were exposed to BoNT in 50 l CM per well and incubated at 37°C in a 5% CO 2 humidified atmosphere for the indicated amount of time. Cells were lysed in 75 l of 1ϫ LDS lysis buffer and analyzed by Western blotting.
To determine the EC 50 , MSCs and RSCs were exposed to serial (3-fold) dilutions of BoNT in CM. Experiments were performed in at least triplicate, and a no-toxin control was included in each replicate. BoNT dilutions remained on cells for 48 h at 37°C in a 5% CO 2 humidified atmosphere. BoNT was then removed, cells were lysed in 75 l of 1ϫ LDS lysis buffer (Life Technologies), and cell lysates were analyzed for the presence of cleaved and uncleaved SNAP-25 by Western blotting as previously described (28,34).
To determine the duration of action of BoNT/A6 in primary neuronal cells, RSCs were exposed for 72 h to 8 pM BoNT/A6, which is the minimum required to achieve 100% SNAP-25 cleavage in RSC cells. Extracellular BoNT was removed by washing the cells in 0.3 ml of CM three times, and cells were further incubated in culture medium without BoNT as previously described for other BoNT/A subtypes (28). Cells were harvested at 3 days after initial BoNT exposure and monthly thereafter until 8 months postexposure. All time points were tested in quadruplicate and included no-toxin controls. Cleaved and uncleaved SNAP-25 was monitored over time by Western blotting and densitometry.
Human iPSC-derived neuron cell assay. iCell GABANeurons (Cellular Dynamics) were stored in liquid nitrogen until use. The cells were plated on a TPP, 96-well, flat-bottom plate treated with 0.01% poly-L-ornithine and 8.3 g/cm 2 Matrigel coated. The cells were maintained in iCell Neuron maintenance medium supplemented with iCell Neuron medium supplement (Cellular Dynamics) and matured for about 1 week until toxin exposure.
To determine the EC 50 , the cells were exposed to serial (3-fold) dilutions of BoNT in CM. Experiments were performed in at least triplicate, and a no-toxin control was included in each replicate. BoNT dilutions remained on cells for 48 h at 37°C in a 5% CO 2 humidified atmosphere. BoNT was then removed, cells were lysed in 50 l of 1ϫ LDS lysis buffer, and cell lysates were analyzed for the presence of cleaved and uncleaved SNAP-25 by Western blotting and densitometry.
To determine the duration of action of BoNT/A6 in hiPSC-derived neurons, cells were exposed to serial dilutions of BoNT/A6 in culture medium for 72 h, after which time extracellular toxin was removed and cells were washed thoroughly. Cells were returned to the incubator and fed every 2 to 3 days. Cells were harvested on days 3, 39, and 70 postexposure, and cell lysates were analyzed by Western blotting and densitometry.
To compare cell entry kinetics of BoNT/A1, /A2, and /A6, hiPSCs were exposed to equal molarity (67 pM) of each BoNT subtype. Cells were harvested at the indicated time points after toxin addition, and cell lysates were analyzed for SNAP-25 cleavage by Western blotting. Entry kinetics were deduced based on the SNAP-25 cleavage determined by Western blotting analysis and densitometry calculations of at least triplicate samples of each toxin at each time point.
Western blot analysis. Western blot analyses were carried out as previously described (28,34). All cell lysates were separated on 12% Novex NUPAGE gels (Life Technologies) using MES running buffer (Invitrogen) and transferred onto a 0.45-m PVDF membrane (Millipore). Membranes were incubated in blocking buffer for 30 min and then in anti-SNAP-25 (Synaptic Systems) primary antibody solution overnight. After five 5-min washes with washing buffer (KPL), the membranes were rinsed with ddH 2 O and incubated in secondary anti-mouse antibody (KPL) for one hour. Five additional washes with washing buffer were done, and membranes were rinsed with ddH 2 O and incubated in chemiluminescent substrate (Phosphaglo, KPL) for ϳ3 min. Images of bands on membranes were obtained using a Fotodyne FOTO/Analyst FX imager and analyzed by densitometry with TotalLab Quant and PRISM 6 software (GraphPad Software Inc.).
Onset and duration of action in vivo. Onset and duration of action were determined in vivo as previously described (17). Using a 0.3-ml insulin syringe with 5-l markings, groups of 5 female ICR mice were injected with the indicated concentrations of BoNT/A1 or BoNT/A6 in 10 l GelPhos (30 mM sodium phosphate, pH 6.3, 0.2% gelatin) into the right gastrocnemius muscle. Local paralysis was measured by the digit abduction score (DAS) for the right hind limb of each mouse at several time points within the first 48 h after injection and every 24 h thereafter on a 0 to 5 scale (17). At the same time points, Rotarod analysis (MED-Associates) was performed on each mouse (after DAS determination). Each mouse attempted to run for a total of 5 min while the Rotarod increased in speed from 4 to 40 rpm. Rotarod analysis was ceased for a group once all mice ran the full 5 min 2 times consecutively.
Determination of intramuscular LD 50 in mice. Groups of 10 female ICR mice (Harlan) were injected with the indicated amounts of BoNT/A1, /A2, or /A6 in 10 l GelPhos buffer using a 0.3-ml insulin syringe (BD) into the right gastrocnemius muscle. In parallel, groups of 5 mice received intraperitoneal injections of the same BoNT dilutions in 0.5 ml GelPhos buffer using a 0.5-ml insulin syringe (BD). Mice were observed through 6 days postinjection, and any deaths were recorded. The LD 50 in pg was calculated for both the intramuscularly (IM) and intraperitoneally (IP) injected mice using the Reed and Muench method (35). Based on the IP data and a unit definition of 1 unit ϭ the amount of BoNT required to result in death of 50% of mice after IP injection within 4 days, the specific activity of each BoNT was determined in units (IP LD 50 ) for the toxin dilutions used in this comparative assay. The IM LD 50 in units was calculated based on the IP LD 50 values.