1 Introduction

The botulinum neurotoxins (BoNT) are a group of proteins that include seven immunologically distinct serotypes, designated A through G, which are produced by different strains of the anaerobic bacterium Clostridium botulinum. The type A neurotoxin complex (BoNTA complex) purified from C. botulinum is recognized for its toxicity as well as its therapeutic value. Ingestion of the BoNTA complex can lead to the severe disease state of botulism, which is caused by BoNTA-mediated blockage of acetylcholine release at neuromuscular junctions and results in flaccid paralysis. Paradoxically, the high specificity and extended duration of action of the BoNTA complex make it a remarkably effective therapeutic agent used for modulating overactive neurotransmission [1, 2].

Following its original isolation from the C. botulinum culture [3, 4], the biophysical characteristics of BoNTA were soon reported. The first of these studies were carried out under acidic conditions (pH = 4.4–4.5) in sodium acetate buffer. Based on electrophoretic, sedimentation, and diffusion characteristics, Kegeles reported a calculated molecular mass for the neurotoxin complex of 1,130,000 Da [5]. Soon after, Putnam and coworkers reported a molecular mass of 900,000 Da based on an analysis by analytical ultracentrifugation [6]. The latter value of 900,000 Da became the accepted molecular mass for several decades.

It is now known that the BoNTA complex is a multimeric protein comprised of a neurotoxic polypeptide and several neurotoxin associated proteins (NAPs) consisting of four hemagglutinating polypeptides (HA17, HA23, HA34, and HA48; named according to their molecular mass values in kDa) and a non-toxic non-hemagglutinating polypeptide (NTNH) [7, 8]. Six open reading frames of the BoNTA gene cluster have been sequenced [9]. These genes encode the neurotoxin polypeptide, three HA polypeptides (HA17, HA34, and HA70), the NTNH polypeptide, and the regulatory component botR/OrfX. The neurotoxin chain is nicked by an unidentified protease to form a 98 kDa heavy chain (HC) and a 51 kDa light chain (LC) that remain linked through a disulfide bond [10]. Similarly, the HA70 protein is cleaved to produce the polypeptides HA23 and HA48 [9, 11]. While the exact function of the NAPs remains unclear, studies suggest that they play a role in protecting the neurotoxin in the harsh environment of the gastro-intestinal tract [12, 13].

Using a combination of gel filtration chromatography and sucrose density gradient centrifugation, Sugii and Sakaguchi [14] reported that three different forms of the neurotoxin complex could be isolated from the C. botulinum type A culture. These compounds have approximate calculated molecular masses of 900 kDa (19 S), 500 kDa (16 S) and 300 kDa (12 S). Based on subsequent densitometric analysis of Coomassie-stained SDS-PAGE gels coupled with N-terminal sequencing, the 300 kDa form was reported to be comprised of the 150 kDa neurotoxin protein and a 120 kDa fragment of the NTNH protein [11]. The same paper reported that the 500 kDa and 900 kDa complexes contain the neurotoxin, the full-length NTNH, and all of the HA subunits. Since the two forms were found to differ only in the relative amounts of HA subunits, the authors speculated that the 19 S form of the neurotoxin complex may be a dimer of the 16 S form coupled through additional HA34 subunits [11].

Despite years of inquiry, the composition of the BoNTA complex is not well understood. To obtain additional insight into the stoichiometry of the neurotoxin complex, we performed a comprehensive study using denaturing non-gel sieving capillary electrophoresis, size exclusion HPLC chromatography, and multi-angle laser light scattering (MALLS). The results of these studies are described herein.

2 Materials and Methods

2.1 Preparation of the Clostridium botulinum Type A Neurotoxin Complex

Growth of Clostridium botulinum and purification of the type A neurotoxin complex was performed using the process described by Schantz and Johnson [15]. The purified neurotoxin complex is stored as an ammonium sulfate suspension.

2.2 Capillary Electrophoresis of the C. botulinum Type A Neurotoxin Complex

Capillary electrophoresis (CE) was performed using a Beckman Coulter P/ACE MDQ CE system (Beckman Coulter, Fullerton, CA). Approximately 25 μg of the BoNTA complex was collected by centrifugation of the ammonium sulfate suspension at 10,000g at 4 °C for 10 min using an Eppendorf microcentrifuge (VWR, Westchester, PA). The supernatant was discarded and the pellet was dissolved in 60 μL water. To this solution 60 μL of the CE sample buffer containing sodium dodecyl sulphate (SDS) (BioRad, Hercules, CA) and 2 μL of 2-mercaptoethanol (Pierce, Rockford, IL) were added. The solution was mixed, briefly centrifuged to collect the contents at the bottom of the tube, and incubated at 95 °C for 10 min. The sample was subsequently cooled at room temperature for 10 min and briefly centrifuged to collect the contents at the bottom of the tube. The sample was transferred to an Ultrafree MC centrifugal filter unit containing a 0.22 μm membrane (Millipore, Billerica, MA) and centrifuged at 2000g in an Eppendorf centrifuge. The filtrate was transferred to a CE sample tube (Beckman Coulter, Fullerton, CA) and placed in the sample tray of the P/ACE MDQ instrument. The sample compartment was maintained at 10 °C during the sample analysis. An uncoated capillary (Polymicro Technologies, Phoenix, AZ) with an effective length of 31 cm and an inner diameter of 50 μm was used for resolving the toxin subunits. The subunit separation was monitored at 214 nm.

2.3 SDS-CE Method Parameters

The uncoated capillary was preconditioned with 0.1 N NaOH at 10 psi for 2 min followed by 0.1 N HCl at 10 psi for 2 min. The capillary was rinsed with water and subsequently filled with the CE-SDS run buffer (BioRad, Hercules, CA) at 12 psi for 10 min. The sample was injected electrokinetically at 0.323 kV/cm or hydrodynamically at 2 psi for 20 s and separated at 0.355 kV/cm for 26 min.

2.4 Recovery of the Neurotoxin Complex During SDS-CE Analysis

To ensure that the analysis did not give rise to a bias due to differential adsorption of some of the subunits to the capillary, the recovery of the toxin complex was assessed using bovine serum albumin (BSA) and carbonic anhydrase (CA) as internal standards. The concentrations of BSA and CA were determined by UV absorbance at 278 nm using an absorptivity for 1 mg/mL solutions equal to 0.66 and 1.9 mL mg−1 cm−1, respectively [16]. A solution of the toxin complex was prepared by centrifugation of a 100 μL aliquot of the toxin complex suspension and dissolving the pellet in 625 μL water. The concentration of this solution was established using the UV absorbance at 278 nm and an absorptivity for 1 mg/mL solutions equal to 1.65 mL mg−1 cm−1 (Allergan unpublished results, [17]). Separate samples were subsequently prepared containing toxin complex spiked with either CA or BSA at two different concentrations and analyzed using the established procedure.

2.5 Size Exclusion High Performance Liquid Chromatography (SE-HPLC) and Multi-Angle Laser Light Scattering (MALLS)

Size-exclusion high performance liquid chromatography was performed on an Agilent 1100 HPLC system controlled using Waters Empower software and equipped with a Tosoh Bioscience TSK-GEL G4000SWxl column (7.8 mm ID × 30 cm L). Isocratic elution was performed at a flow rate of 0.4 mL/min using a mobile phase of 50 mM sodium citrate (pH 6.0), 100 mM sodium chloride, and 0.02% sodium azide. Test samples were prepared as follows: 100 μL of C. botulinum type A neurotoxin complex in an ammonium sulfate suspension was transferred to a 2.0 mL microcentrifuge tube and centrifuged at 10,000g for 10 min. The supernatant was removed and the pellet was reconstituted into 320 μL of mobile phase, yielding a final toxin concentration of approximately 0.8 μg/μL. A mixture of molecular weight standards containing thyroglobulin (Mr = 670 kDa), gamma globulin (Mr = 158 kDa), ovalbumin (Mr = 44 kDa), and myoglobin (Mr = 17 kDa) (BioRad, Hercules, CA) was prepared according to the manufacturer’s instructions. The test samples and the mixture of molecular weight standards were analyzed using an injection volume of 45 μL.

Light scattering analysis was performed using an Agilent 1100 HPLC system consisting of a binary pump with automatic degasser, thermostated auto-sampler, and diode array detector (Agilent Technologies, Palo Alto, CA) and connected in-line to a DAWN-EOS static light scattering detector equipped with a K5 flow cell (Wyatt Technology, Santa Barbara, CA). The Agilent 1100 HPLC system was controlled using Chemstation software (Agilent Technologies, rev B.01.01). The light scattering detector was controlled using the ASTRA V software (Wyatt Technology, version 5.1.9.1). Prior to use, the light scattering detector was calibrated with toluene according to the manufacturer’s instructions. Calibration of the light scattering detector was subsequently verified using a bovine serum albumin standard (Pierce, Rockford, IL). BoNTA complex sample preparation and size exclusion chromatography were performed as described above using a Shodex Protein KW−804 size exclusion column (8.0 mm ID × 300 mm L, 7 μm particle size) (Phenomenex, Torrance, CA). The concentration of the BoNTA neurotoxin complex during the light scattering experiments was monitored by UV absorbance at 278 nm using an absorptivity of 1.65 mL mg−1 cm−1 as described above. The molecular mass for the BoNTA complex was calculated from the light scattering data using a specific refractive index increment (dn/dc) value of 0.180 mL/g.

2.6 Molecular Mass Values of the Subunits

The actual molecular mass values of the different subunits used for the stoichiometry calculations are listed in Table 1. The values were calculated based on the translated nucleotide sequences of the neurotoxin complex genes as determined by Zhang et al. [9] for the C. botulinum Hall A strain (GenBank accession numbers AF488745, AF488746, AF488747, AF488748, AF488749, and AF488750). The N-terminal amino acid sequence of each subunit was used to confirm the actual start of the amino acid sequence and identify any post-translational nicking [11].

Table 1 Overview of molecular mass values of each of the subunits used for the stoichiometry calculations. The molecular mass values were calculated based on the translated DNA sequence for the different neurotoxin complex subunits combined with N-terminal sequencing information
Full size table

3 Results

3.1 Capillary Electrophoresis of the C. botulinum Type A Neurotoxin Complex

Two modes for injecting sample into the capillary; electrokinetic and hydrodynamic, were evaluated to assess the impact of the injection mode on the relative peak areas for the different subunits. The two modes were subsequently compared by determining the ratio of the relative peak areas obtained for each of the components using the electrokinetic injection mode over the area ratios obtained using the hydrodynamic injection mode. The results are listed in Table 2 and suggest that the electrokinetic injection mode causes a bias towards the smaller subunits. Since this bias will make it difficult to obtain an accurate assessment of the relative abundance of the various subunits, the hydrodynamic injection mode (pressure injection) was selected to determine the stoichiometry of the neurotoxin complex. Figure 1 shows a representative electropherogram obtained for the separation of the BoNTA complex with the individual components labeled. The peak identities of the HC and LC peaks were assigned based on a comparison of non-reduced electropherograms, in which the two subunits are covalently linked, to reduced electropherograms where each of the subunits are resolved as two discrete peaks. The remaining subunits were identified based on their molecular mass analogous to the assignments reported for SDS-PAGE [11]. Due to the small difference in their molecular mass, the HA48 and neurotoxin LC components are not resolved.

Table 2 Preferential loading of the low molecular mass subunits versus the high molecular mass subunits observed when using the electrokinetic injection mode compared to the hydrodynamic injection mode during capillary electrophoresis of the Clostridium botulinum type A neurotoxin complex. The ratios were obtained by dividing the relative peak areas obtained for each of the neurotoxin complex subunits using the electrokinetic injection mode by the relative peak areas obtained for the subunits using the hydrodynamic injection mode
Full size table
Fig. 1
figure 1

SDS-CE electropherogram of the reduced Clostridium botulinum type A neurotoxin complex. The sample was prepared and analyzed as described in the text

Full size image

The precision of the established procedure was subsequently estimated by performing independent runs of a single BoNTA complex sample and calculating the relative peak areas for each of the subunits. Based on the results, the RSD of the relative peak area for each of the individual subunits was determined to be less than 9%. In addition, the recovery of the neurotoxin complex analyzed according to the established procedure was verified using either BSA or CA as an internal standard. Compared to these standards, the recovery of the neurotoxin complex was determined to be 131% relative to BSA and 84% relative to CA, respectively. This suggests that the electropherograms obtained for the neurotoxin complex are representative of the composition of the complex and can be used to assess its stoichiometry.

3.2 Stoichiometry of the C. botulinum Type A Neurotoxin Complex

Calculation of the BoNTA complex stoichiometry from the relative peak areas obtained for the different subunits of the reduced neurotoxin complex was based on the following two assumptions: (1) the per-residue absorbance at 214 nm is the same for each of the subunits (i.e., the absorbance of each of the subunits at 214 nm is only dependent on the length of the peptide chain, and (2) the light chain and the heavy chain, which together constitute the 150 kDa neurotoxin component, are present in equimolar amounts.

The stoichiometry was subsequently calculated as follows: The corrected peak areas for each of the subunits were determined from the electropherogram using Beckman 32 Karat software. The expected peak area for the light chain was calculated based on the peak area obtained for the heavy chain after correction for the difference in molecular mass.

$$ \hbox{LC\,peak\,area}=\hbox{HC\,peak\,area}\times\hbox{(LC\,molecular\,mass/HC\,molecular\,mass)} $$

The peak area for the HA48 was determined by subtracting the calculated peak area for the light chain from the peak area obtained for the (HA48 + LC) peak.

$$ \hbox{HA48\,peak\,area}=\hbox{(HA48}+\hbox{LC)\,peak\,area}-\hbox{LC\,peak\,area} $$

To account for the difference in molecular mass and its impact on the absorbance at 214 nm, the peak areas obtained for each of the subunits were divided by the molecular mass for each of the subunits to obtain the adjusted peak area.

$$ \hbox{Adjusted\,peak\,area}=\hbox{Subunit\,peak\,area/subunit\,molecular\,mass} $$

Finally, the adjusted peak areas obtained for each of the subunits were normalized against the adjusted peak area of the heavy + light chain (HC + LC).

$$ \hbox{Normalized\,subunit\,ratio}=\hbox{Adjusted\,peak\,area/adjusted\,peak\,area\,for\,(HC}+\hbox{LC)} $$

The average complex stoichiometry, obtained by analyzing seven different C. botulinum type A neurotoxin complex batches, is listed in Table 3.

Table 3 Results for the subunit composition of the Clostridium botulinum type A neurotoxin complex. The results represent the average of seven neurotoxin complex batches. The values in parentheses reflect the calculated standard deviation
Full size table

3.3 Size Exclusion HPLC with Light Scattering

The BoNTA complex elutes as a single peak when analyzed using SE-HPLC chromatography (Fig. 2). The estimated molecular mass of the complex under these conditions, as determined by comparison of its retention time to a calibration curve created using the BioRad molecular weight standards mixture, is 880 kDa. The average molecular mass determined using MALLS is 925 ± 45 kDa.

Fig. 2
figure 2

Size exclusion-MALLS HPLC chromatogram of the Clostridium botulinum type A neurotoxin complex. The line across the peak represents the average molar mass distribution determined using multi-angle laser light scattering detection

Full size image

4 Discussion

The molecular mass and subunit stoichiometry of the BoNTA complex has received considerable attention since its discovery. Despite this effort, definitive characterization of these properties has remained elusive. To obtain insight into the subunit stoichiometry of the neurotoxin complex, a denaturing capillary electrophoresis procedure was established for the C. botulinum type A neurotoxin complex, which provides a quantitative means of evaluating the relative ratios of the subunits in the complex. Denaturing capillary electrophoresis uses a sieving medium inside a fused silica capillary to separate SDS-denatured proteins by size and molecular mass under a high voltage electric field. The technique has been extensively used for the characterization of recombinant therapeutic proteins. Although the separation of analytes by SDS capillary electrophoresis utilizes the same separation principles as traditional slab gel techniques, the use of a capillary allows on-line UV or fluorescence detection. Therefore, unlike previous studies where the stoichiometry of the complex was assessed using denaturing SDS-PAGE [11, 18], the current method is not affected by potential differences in the dye binding affinity of each subunit.

The results of the CE analysis indicate that the BoNTA complex is composed of 5–6 HA17, 4–5 HA23, 8–9 HA34, 3–4 HA48, and 1 NTNH subunit per 150 kDa neurotoxin subunit; resulting in a molecular mass of 934 ± 63 kDa for the complex. The weight average molecular mass of the neurotoxin complex determined by MALLS is 925 ± 45 kDa. The MALLS analysis also indicates that the population of the BoNTA complex is homogenous in terms of the molecular mass distribution, since the molecular mass calculated for the complex is constant throughout the peak (see Fig. 2). The molecular mass of the neurotoxin complex estimated from its elution time during size exclusion HPLC analysis when compared to external protein standards is 880 kDa. The elution of a protein during size-exclusion HPLC is dependent on several factors in addition to its molecular mass (e.g., molecular shape and interaction with column matrix) while the weight average molecular mass determined by MALLS is independent of these factors. Therefore, the value determined using MALLS is expected to be more accurate. Both values are, however, consistent with the molecular mass calculated for the neurotoxin complex based on the subunit composition determined by capillary electrophoresis.

With the exception of the value obtained for the HA34 subunit, the stoichiometry results obtained in this study are consistent with a previous report in which the relative abundance of the different subunits was determined by densitometric analysis of SDS-PAGE gels [18]. However, the results are not consistent with the molar ratios determined in another SDS-PAGE based study that evaluated the subunit molar ratios in the purified 16 S and 19 S BoNTA complexes [11]. Based on their results, the authors speculated that the 19 S complex is a dimer of the 16 S complex with additional HA34 subunits acting as a cross-linker. This model is not supported by a recent study describing the crystal structure of the HA34 subunit [19]. In the latter study the authors concluded that, although the asymmetric unit of crystallized HA34 (referred to as HA33 in the study) contained two HA34 molecules, the limited surface interactions identified between two HA34 molecules do not support its function as a cross linker.

The gene sequence information available for the neurotoxin complex indicates that the HA48 and HA23 subunits are expressed as a single 70 kDa polypeptide, which is subsequently cleaved to form the two separate subunits. The relative ratios observed for the two subunits in this study as well as in the previous studies show that the HA48 and the HA23 subunits are not present in equimolar amounts. The reason for this is not known and it remains to be established whether this reflects a true difference in the presence of these subunits in the neurotoxin complex.

In summary, the molecular mass determined for the intact neurotoxin complex by SE-HPLC and SE-HPLC/MALLS experiments is 880 kDa and 925 kDa, respectively. These values are consistent with the calculated molecular mass of 934 kDa for the complex based on the proposed subunit composition as determined by CE analysis. The close agreement between these orthogonal approaches suggests that the stoichiometry proposed in this study is a reasonable model to describe the composition of the C. botulinum type A neurotoxin complex.