INTRODUCTION

Top Abstract Introduction Materials & Methods Results & Discussion References

Botulinum neurotoxin is secreted by the anaerobic bacterium Clostridium botulinum as one of seven serotypes, classified A through G (35). The different serotypes proteolytically cleave specific proteins involved in synaptic vesicle docking that are necessary for cellular communication at the neuromuscular junction (12). Serotype A can be purified as a 900-kDa complex (BoNTA-HA) consisting of a toxic component (BoNTA [botulinum neurotoxin serotype A]) and a nontoxic component (HA [hemagglutinin]) (5-7). Studies involving relative oral toxicities (26, 33), intestinal absorption (34), and comparison to tetanus toxin (31) have indicated that the complex, not the botulinum toxin alone, is responsible for the extremely low amount of botulinum neurotoxin required in botulism poisoning.

The most common mechanism of botulism poisoning is through oral ingestion of the complex, which is found in food contaminated with C. botulinum. Ingested spores of the bacteria may also colonize and produce toxin in the intestinal tracts of infants, resulting in infant botulism (36). Previous experiments have demonstrated that the 900-kDa complex protects the toxin during its exposure to harsh conditions. Ohishi and coworkers (26) have demonstrated that the 900-kDa complex has a 360-fold-higher oral toxicity in rats than the 150-kDa botulinum neurotoxin component alone. Most proteins are broken down into peptides and amino acids in the stomach and small intestine during the process of digestion (2). However, the 900-kDa complex enters the stomach and withstands the acidic (pH 2) gastric juice containing the protease pepsin. The complex then enters the small intestine, where it encounters several more proteases (trypsin, chymotrypsin, and carboxypeptidase) that function at pH 7 to 8. Despite these harsh denaturing and proteolytic conditions, active botulinum neurotoxin (13, 22) and nontoxic component (34) can be detected in the lymph and circulatory systems.

The botulinum neurotoxin is comprised of a C-terminal 100-kDa heavy chain and an N-terminal 50-kDa light chain linked by a disulfide bond. From C terminus to N terminus, the protein can be further divided into three 50-kDa functional domains (3, 4, 18)binding, translocation, and catalyticwhich allow the protein to bind a cell surface receptor, pass across the membrane (29), and cleave a protein involved in synaptic vesicle docking (12), respectively. The nontoxic component is composed of several protein subunits and can be separated into nontoxic agglutinating proteins (14, 17, 32) and nontoxic nonagglutinating proteins (11, 14). Several subcomponents of the nontoxic component have been characterized. One of the nontoxic nonagglutinating proteins is a single peptide of 120 kDa and has been sequenced (11). The molecular weights of the subcomponents of the nontoxic agglutinating portion have also been determined (14, 32). However, the physiological role and molecular organization of the nontoxic component are not well understood. In addition, many of the studies of toxin exposure to harsh conditions have been conducted with crude cell supernatant. To understand the stability and equilibrium of the botulinum neurotoxin complex and its separated components at the molecular level, we conducted a series of biophysical experiments using pure preparations of botulinum neurotoxin, botulinum neurotoxin complex, and the nontoxic agglutinating portion.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results & Discussion References

Purification of proteins. For all experiments, proteins were obtained as ammonium sulfate precipitates and were purified by ion-exchange chromatography as previously described (5).

Light scattering. The hydrodynamic radius, estimated molecular weight, and polydispersity of protein samples were determined by using a Dynapro-801 dynamic light scattering instrument (Protein Solutions, Charlottesville, Va.). Samples of botulinum neurotoxin (0.18 mg/ml), nontoxic component (0.20 mg/ml), and 900-kDa complex (0.22 mg/ml) were incubated for 30 min at 4°C over a range of pHs in 10 mM buffers containing 100 mM sodium chloride. Buffers were citric acid (pH 1 to 5) (citric acid has three pK_as at 25°C; pK₁ = 3.128, pK₂ = 4.761, and pK₃ = 6.396), bis-Tris [bis(2-hydroxyethyl)imino-tris(hydroxymethyl)methane; pH 6], HEPES (pH 7 to 8), and CHES [2-(N-cyclohexylamino)ethanesulfonic acid] (pH 9 to 10). Individual samples were recorded for at least 5 min, during which 12 to 13 data points were taken and then analyzed by using a monomodal fitting program. The polydispersity served as the deviation of the size distribution based on the mean hydrodynamic radius of each sample. Experiments were performed in triplicate.

Test for pH effects. Protein samples were incubated at the appropriate pH for 48 h at 4°C. Protein samples were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) on a 12% polyacrylamide gel, and the bands were visualized with silver stain (10).

Proteolysis. Proteins were incubated at various pHs and with several protease mixtures. Each protein sample contained 2 ml of protein (0.42 mg/ml) in dialysis tubing with a 50-kDa pore size. Dialysis tubing containing the protein sample was placed into the appropriate protease mixture. The mixture was incubated with gentle stirring for 2 h at 37°C. Pepsin incubations were performed with 0.4 mg of pepsin per ml in 10 mM hydrochloric acid (pH 2). Intestinal cocktail consisted of trypsin (0.02 mg/ml), chymotrypsin (0.03 mg/ml), and carboxypeptidase A (0.02 mg/ml) in 50 mM sodium bicarbonate-100 mM sodium chloride buffer (pH 8). Protease concentrations used were chosen based on reported values (15, 16), though the physiological protease concentration depends on the feeding state of the organism. After incubation, proteases were quenched with 0.5 ml of pepstatin (1 mg/ml) or phenylmethylsulfonyl fluoride (17.4 mg/ml) and 1 ml of 0.5 M EDTA. Proteins were also exposed to incubation of pepsin followed by intestinal cocktail or incubation at pH 2 followed by intestinal cocktail. Gels were 12% polyacrylamide, and protein bands were visualized with silver stain (10).

Circular dichroism. Circular dichroic spectra were gathered on a model J-600 spectropolarimeter (Jasco, Easton, Md.) at 37 and 25°C. The bandwidth used was 1 nm, and the step resolution was 2 nm. Four scans of each sample were made, using a time constant of 1 s and a scan rate of 50 nm/min. The cell volume was approximately 1 ml, with a path length of 0.1 cm. The cell was jacketed for temperature adjustment and controlled by using a water bath. Proteins at a concentration of 0.025 mg/ml were dialyzed extensively in their buffers before being examined on the spectropolarimeter. The pH 6 and pH 8 buffers were 100 mM sodium phosphate containing 100 mM sodium chloride. The pH 2 buffer was 60 mM hydrochloric acid-potassium chloride buffer containing 100 mM NaCl. No appreciable difference in signal was observed at 37 and 25°C; thus, only 37°C data are shown.

ELISA. Botulinum neurotoxin (0.01 mg/ml) was incubated in pepsin (0.4 mg/ml) and intestinal cocktail (see above) for 2 h at 37°C. The concentration of substrate protein used was the minimum amount of protein necessary to give strong enzyme-linked immunosorbent assay (ELISA) signal. Using single-chain variable fragments of mouse or of human monoclonal antibodies, ELISAs were performed on proteolyzed samples as described by Chen et al. (5).

Surface plasmon resonance. Binding of botulinum neurotoxin and 900-kDa complex to lipid was monitored by using surface plasmon resonance on a BIAcore 2000 (Biosensor, Piscataway, N.J.). Botulinum neurotoxin (84 to 1,300 nM) or 900-kDa complex (52 to 832.5 nM) was injected over a lipid monolayer containing 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC). The lipid monolayer was formed by incubating a solution of POPC liposomes (2 mg/ml) on a BIAcore sensorchip (type HPA) overnight in a moist environment. Liposomes were made by sonicating a solution of POPC in a buffer of 10 mM Tris (pH 7.0) containing 100 mM sodium chloride and 2 mM sodium azide. The experiments were performed at a flow rate of 40 ml/min for 20 min in running buffer of 10 mM Tris (pH 7.0) containing 100 mM sodium chloride and 2 mM sodium azide.

Sugar binding. Agglutination assays were performed as previously described (20). Erythrocytes (RBC) from rabbits were washed twice with saline (68 mM sodium citrate, 146 mM sodium chloride [pH 7.4]), spun at 4,000 × g, and diluted to 0.5% for each assay. Protein solution (0.5 ml) at twice the desired concentration was added to a disposable semimicrocuvette. Then 0.5 ml of the 0.5% RBC was added to the protein, resulting in the desired protein concentration and a 0.25% RBC solution. The mixture was incubated at 37°C for 2 h, and the transmittance was measured at 405 nm in a Shimadzu UV-160 spectrophotometer. The saline control contained 0.25% RBC incubated in saline for 2 h at 37°C. Each sample was performed in triplicate. The 900-kDa complex and nontoxic component were also proteolyzed and then used in an agglutination assay. The saline control was subtracted from the values for the other samples. The agglutination ability of the unproteolyzed HA was scaled to a value of 100. The agglutination abilities of the other samples represent their percent transmittances relative to the percent transmittance by unproteolyzed HA. Previous reports (20) show that the 900-kDa BoNTA-HA complex is capable of agglutination at lower temperatures. When we performed the agglutination assay at a lower temperature (12°C), we observed that the complex showed an increased ability to agglutinate RBC, consistent with the literature but not physiologically relevant.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials & Methods Results & Discussion References

Light scattering. Dynamic light scattering was used to determine the aggregation state and stability of the three species under a variety of conditions simulating the environments in the gastrointestinal tract. This technique measures the diffusion coefficient of particles in solution, allowing a radius and estimated molecular weight to be calculated. A molecular weight calculated to be higher than expected indicates a deviation from a spherical shape. This deviation may be due to nonspherical species, denatured protein, or the presence of large aggregates of many proteins. The distribution of particle sizes is indicated by the polydispersity value. A polydispersity of zero signifies that there is only one particle size in solution. A large polydispersity value indicates the presence of different-size particles in solution. Examination of the radii of the three molecular species between pH 1 and 10 yielded a number of intriguing results. From the hydrodynamic radii the corresponding molecular weights were calculated, assuming an approximate, spherical shape for the protein.

View larger version (25K): [in this window] [in a new window]   FIG. 1.   Dynamic light scattering results, expressed as pH versus apparent radius and molecular weight of botulinum neurotoxin (A), nontoxic component (B), and 900-kDa complex (C) at various pHs. (D) Estimated molecular weights (MW) of botulinum neurotoxin (triangles), nontoxic component (squares), and 900-kDa complex (circles). The bar around each data point represents the average polydispersity value for three separate experiments. The polydispersity is the deviation of size distributions based on the mean radius measurement. Specimens containing aggregates or unfolded proteins have large polydispersity values.

pH effects. The effect of pH on the neurotoxin's structural stability was examined by SDS-PAGE. A polyacrylamide gel of botulinum neurotoxin incubated at pH 2, 6, and 8 for 48 h indicated that the neurotoxin is susceptible to breakdown at pH 2 and 8 (Fig. 2). Table 1 shows that nonincubated neurotoxin runs as a 100-kDa heavy chain and a 50-kDa light chain. After 48 h of incubation at pH 6, we observed primarily the 100-kDa heavy chain and 50-kDa light chain of botulinum neurotoxin. A band at 150 kDa corresponded to a small quantity of unnicked neurotoxin. Although the 100- and 50-kDa bands predominated at pH 2 and 8, we observed a ladder of many lower-molecular-weight fragments. This fragmentation could be a result of acid-catalyzed hydrolysis, base-catalyzed hydrolysis, or proteolysis.

View larger version (31K): [in this window] [in a new window]   FIG. 2.   SDS-PAGE of botulinum neurotoxin incubated at pH 2, 6, and 8. Arrows identify the heavy chain (100 kDa) and light chain (50 kDa) of botulinum neurotoxin. The band appearing at approximately 150 kDa is unnicked neurotoxin.

Proteolysis. Besides pH extremes, the other stresses that the 900-kDa complex must endure come from proteases. The effects of proteases on botulinum neurotoxin, nontoxic component, and 900-kDa complex were observed by performing protease incubations and assaying the results by gel electrophoresis (Table 1). Incubations consisted of pepsin at pH 2, intestinal cocktail (trypsin, chymotrypsin, carboxypeptidase) at pH 8, and pepsin at pH 2 followed by intestinal cocktail at pH 8. When incubated individually, both the neurotoxin and the nontoxic component were susceptible to all of the protease incubations (Table 1). In contrast, the 900-kDa complex demonstrated an amazing resistance to proteolysis. After incubation in pepsin at pH 2, the 900-kDa complex lost none of its bands. When the 900-kDa complex was incubated in intestinal cocktail, the 100- and 50-kDa bands of botulinum neurotoxin were proteolyzed along with the 120- and 106-kDa bands of the nontoxic component. However, incubating the 900-kDa complex in pepsin at pH 2 before exposing the complex to intestinal cocktail inhibited the proteolysis of the bands corresponding to botulinum neurotoxin. The bands which have been previously shown to be responsible for agglutination (14, 19, 21, 35, and 52 kDa) (14, 32) remained intact according to our gel whereas the band previously shown to be nontoxic-nonagglutinating (11) was destroyed.

Circular dichroism. Circular dichroic spectra of the botulinum neurotoxin, the nontoxic component, and the complex were taken at 37 and 25°C to determine the effects of pH on the structure and stability of the proteins. Temperature appeared not to have an effect on the spectra. At pH 2, 6, and 8, the 900-kDa complex shows little change in secondary structure (Fig. 3A). Thus, the increased protease resistance of the 900-kDa complex, upon exposure to pH 2, did not result from a visible conformational change in the secondary structure. When the nontoxic component was studied at pH 6 and 8, it likewise showed minor change in secondary structure (Fig. 3B). Upon exposure to pH 2, the nontoxic component showed a slight decrease in helicity. This loss of secondary structure is not large enough to indicate that the nontoxic component significantly unfolds. Since individual nontoxic components do not unfold, an increase in hydrodynamic radius, which corresponds to an increase in the size of the particle being observed, likely results from the aggregation of multiple nontoxic components. This aggregation explains the size increase of the nontoxic component observed by light scattering (Fig. 1B).

View larger version (26K): [in this window] [in a new window]   FIG. 3.   Circular dichroism (CD) spectra of the 900-kDa complex at pH 8 (top), pH 6 (middle), and pH 2 (bottom) (A), the nontoxic component at pH 2 (top), pH 6 (middle), and pH 8 (bottom) (B), and botulinum neurotoxin at pH 8 (top), pH 6 (middle), and pH 2 (bottom) (C).

ELISA and surface plasmon resonance. The nontoxic component appears to prevent the vulnerable regions of the botulinum neurotoxin from protease attack, acid hydrolysis, and aggregation at low pH. Epitope mapping suggests that in the 900-kDa complex, the nontoxic component covers a large portion of the binding domain of botulinum neurotoxin (5). Previous experiments have shown that the binding domain is highly susceptible to trypsin cleavage (28). Therefore, the nontoxic component may play an important role in protecting the binding domain in the gastrointestinal tract. Furthermore, ELISA performed on botulinum neurotoxin after pepsin and intestinal proteolysis showed that antibodies no longer bind to it, suggesting that cleavage, denaturation, or both have occurred. In addition to being highly susceptible to proteolysis, botulinum neurotoxin appears to stick strongly and irreversibly to lipid monolayers such as POPC, a lipid commonly found in membranes of cells lining the gastrointestinal tract (9). Using surface plasmon resonance, we found that botulinum neurotoxin injected over POPC monolayers adsorbed strongly whereas the 900-kDa complex did not. Similarly, unless the toxic component was in its complexed form, it would probably adhere to lipid membranes along the gastrointestinal tract. Additional evidence for the association of botulinum neurotoxin with phospholipids has been seen by several researchers (17, 24, 25, 27). Thus, without the nontoxic component for protection, botulinum neurotoxin might never leave the gastrointestinal tract but instead be digested like most other proteins.

Sugar binding. The only known ability of the nontoxic component is its capacity to bind certain sugars (1). Upon binding these sugars, the nontoxic component is able to agglutinate, or clump together, cells (19). Whether this property aids the neurotoxin in reaching its target is under investigation. The clumping of RBC was monitored with agglutination assays (Table 1). Previous work has shown that the 14-, 19-, 21-, 35-, and 52-kDa bands are primarily responsible for agglutination (14, 32). Unproteolyzed nontoxic component containing these five bands was able to agglutinate RBC at 37°C (Table 1). After proteolysis by pepsin (pH 2), the nontoxic component experienced a slight decrease in agglutination ability (Table 1). Upon exposure to sequential protease incubations, the nontoxic component showed the largest decrease in the ability to agglutinate RBC (Table 1). This decrease appears to correlate with the loss of the 120- and 106-kDa bands (Table 1). The 106-kDa band was shown to be a proteolyzed portion of the 120-kDa band (14). Therefore, the 120-kDa band appears to be necessary for the nontoxic component to maintain optimal agglutination ability by protecting the other proteins but is not directly involved in agglutination. When unproteolyzed or exposed to pepsin, the complex showed a low ability to agglutinate (Table 1). Upon exposure to a sequential pepsin and intestinal cocktail incubation, the complex increased its ability to agglutinate RBC (Table 1). The sequentially proteolyzed complex contains the 14-, 19-, 21-, 35-, and 52-kDa bands plus the 100- and 50-kDa bands corresponding to botulinum neurotoxin (Table 1). Thus, the presence of botulinum neurotoxin appears to protect the ability of the complex to agglutinate after a sequential protease incubation.