Botulinum neurotoxin type A: cleavage of the heavy chain into two halves and their partial sequences

Purification and Characterization of Recombinant Botulinum Neurotoxin Serotype FA, Also Known as Serotype H

We have purified and characterized recombinant botulinum neurotoxin serotype FA (BoNT/FA). This protein has also been named as a new serotype (serotype H), but the classification has been controversial. A lack of well-characterized, highly pure material has been a roadblock to study. Here we report purification and characterization of enzymatically active, and of inactive nontoxic, recombinant forms of BoNT/FA as tractable alternatives to purifying this neurotoxin from native Clostridium botulinum. BoNT/FA cleaves the same intracellular target proteins as BoNT/F1 and other F serotype BoNTs; the intracellular targets are vesicle associated membrane proteins (VAMP) 1, 2 and 3. BoNT/FA cleaves the same site in VAMP-2 as BoNT/F5, which is different from the cleavage site of other F serotype BoNTs. BoNT/FA has slower enzyme kinetics than BoNT/F1 in a cell-free protease assay and is less potent at inhibiting ex vivo nerve-stimulated skeletal muscle contraction. In contrast, BoNT/FA is more potent at inhibiting neurotransmitter release from cultured neurons.

High Yield Preparation of Functionally Active Catalytic-Translocation Domain Module of Botulinum Neurotoxin Type A That Exhibits Uniquely Different Enzyme Kinetics

Botulinum neurotoxins (BoNTs) are the most toxic proteins known to cause flaccid muscle paralysis as a result of inhibition of neurotransmitter release from peripheral cholinergic synapses. BoNT type A (BoNT/A) is a 150 kDa protein consisting of two major subunits: light chain (LC) and heavy chain (HC). The LC is required for the catalytic activity of neurotoxin, whereas the C and N terminal domains of the HC are required for cell binding, and translocation of LC across the endosome membranes, respectively. To better understand the structural and functional aspects of BoNT/A intoxication we report here the development of high yield Escherichia coli expression system (2-20-fold higher yield than the value reported in the literature) for the production of recombinant light chain-translocation domain (rLC-TD/A) module of BoNT/A which is catalytically active and translocation competent. The open reading frame of rLC-TD/A was PCR amplified from deactivated recombinant BoNT/A gene (a non-select agent reagent), and was cloned using pET45b (+) vector to express in E. coli cells. The purification procedure included a sequential order of affinity chromatography, trypsinization, and anion exchange column chromatography. We were able to purify > 95% pure, catalytically active and structurally well-folded protein. Comparison of enzyme kinetics of purified LC-TD/A to full-length toxin and recombinant light chain A suggest that the affinity for the substrate is in between endopeptidase domain and botulinum toxin. The potential application of the purified protein has been discussed in toxicity and translocation assays.

Botulinum neurotoxin types A, B, and E: fragmentations by autoproteolysis and other mechanisms including by O-phenanthroline-dithiothreitol, and association of the dinucleotides NAD(+)/NADH with the heavy chain of the three neurotoxins

The first evidence of autoproteolytic activity of the approximately 50-kDa light chain of the clostridial neurotoxins (NT) is traceable to the observations that the light chains of botulinum NT serotypes A and E, separated from their approximately 100-kDa heavy chain conjugate, were found cleaved at the amino side of Tyr250 and Arg244, respectively [DasGupta and Foley (1989). Biochimie 71: 1183-1200]. Specific cleavages of the recombinant light chain of NT type A, including at Tyr249-Tyr250, firmly established that the cleavages reported earlier were due to autoproteolysis [Ahmed et al. (2001). J. Protein Chem. 20: 221-231; Ahmed et al. (2003). Biochemistry 42:12539-12549] and not by contaminating proteases or non-enzymatic. We now report many cleavages in the NT types A, B and E and also in their separated light and heavy chains, and identification of several of the peptide bonds cleaved. None of the identified cleaved bonds (-P1-P1' -) in one serotype (except Asp-Pro) was found common in other serotypes or cleaved within itself at a second site. After separation from the heavy chain self-cleavages of the light chains of type A, B and E at Tyr249-Tyr250, Gln258-Ser259 and Ile243-Arg244, respectively indicate an intriguing feature (in the aligned sequences these bonds of type A and B are 2 and type A and E are 4 peptide bonds apart) that may have some role in the NT's structure-function relationship yet to be understood. We point out that autoproteolysis of a single peptide bond (Phe418-Thr419 or Phe422-Glu423) in NT type A reported by Ahmed et al. (2001) can potentially generate proteolytically active light chain freed of the heavy chain; this is an efficient pathway, that by-passes nicking by a trypsin-like protease(s) inside the intrachain disulfide bridge and its reductive cleavage. We offer probable explanations for the observed cleavages such as acid- and metal-mediated (non-catalytic and non-stoichiometric) reactions in addition to autoproteolysis but cannot predict which mechanism(s) of cleavage occur or prevail following NT's entry in the body as poison or therapeutic agent. The metal chelator O-phenanthroline (above critical miceller concentration) in the presence of dithiothreitol cleaved type E NT at limited sites generating discrete 114-, 87-, 49-, 42-, and 31-kDa fragments but degraded NTs type A and B extensively. The limited cleavage of type E NT was dependent on the presence of metal ion(s) bound to the protein and its native (urea sensitive) conformation. The self-cleavage of the NTs at specific sites prompted us to search for specific binding sites on the NTs analogous to SNARE-motifs-the 9-residuelong motifs present on the NT's natural substrates (SNAP-25, syntaxin, VAMP/synaptobrevin); such putative binding motifs (sites) noted on all clostridial NTs are reported here. Their relationship to the observed autoproteolysis remains to be determined experimentally. The dinucleotide NAD(+)/NADH associated with the NTs type A, B and E (2-3 NADH per protein molecule) via their H-chains, and a portion of the H-chain (toward the C-terminus) appears to exhibit limited amino acid sequence homology with lactate dehydrogenase-a representative NAD(+)/NADH binding protein.

Covalent structure of botulinum neurotoxin type E: location of sulfhydryl groups, and disulfide bridges and identification of C-termini of light and heavy chains

Botulinum neurotoxin (NT) serotype E is synthesized by Clostridium botulinum as an approximately 150-kDa single-chain polypeptide of 1252 amino acid residues of which 8 are Cys residues [Puolet et al. (1992); Biochem. Biophys. Res. Commun. 183, 107-113]. The posttranslational processing of the gene product removes only the initiating methionine. A very narrow segment of this 1251-residue-long mature protein--at one-third the distance from the N-terminus (between residues Lys 418 and Arg 421)--is highly sensitive to proteases, such as trypsin. The single-chain NT easily undergoes an exogenous posttranslational modification by trypsin; residues 419-421 (Gly-Ile-Arg) are excised. The proteolytically processed NT is a dichain protein in which Pro 1-Lys 418 constitute the approximately 50-kDa light chain, Lys 422-Lys 1251 constitute the approximately 100-kDa heavy chain; Cys 411-Cys 425 and Cys 1196-Cys 1237 form the interchain and intrachain disulfide bonds, respectively; the other four Cys residues at positions 25, 346, 941, and 1035 remain as free sulfhydryl groups. The approximately 150-kDa dichain NT, and separated light and heavy chains, were fragmented with CNBr and endoproteases (pepsin and clostripain); some of these fragments were carboxymethylated with iodoacetamide (with or without 14C label) before and after fragmentation. The fragments were separated and analyzed for amino acid compositions and sequences by Edman degradation to determine the complete covalent structure of the dichain type E NT. A total of 208 amino acid residues, i.e., 16.5% of the entire protein's sequence deduced from nucleotide sequence, was identified. Direct chemical identification of these amino acids was in complete agreement with that deduced from nucleotide sequence.

Botulinum neurotoxin type A: limited proteolysis by endoproteinase Glu-C and alpha-chymotrypsin enhanced following reduction; identification of the cleaved sites and fragments

Botulinum neurotoxin (NT) serotype A is a approximately 150-kDa dichain protein. Posttranslational nicking of the single-chain NT (residues Pro 1-Leu 1295) by the protease(s) endogenous to Clostridium botulinum excises 10 residues, leaving Pro 1-Lys 437 and Ala 448-Leu 1295 in the approximately 50-kDa light (L) and approximately 100-kDa heavy (H) chains, respectively, connected by a Cys 429-Cys 453 disulfide and noncovalent bonds [Krieglstein et al. (1994), J. Protein Chem. 13, 49-57]. The L chain is a metalloprotease, while the amino- and carboxy-terminal halves of the H chain have channel-forming and receptor-binding activities, respectively [Montecucco and Schiavo (1995), Q. Rev. Biophys. 28, 423-472]. Endoproteinase Glu-C and alpha-chymotrypsin were used for controlled digestion at pH 7.4 of the approximately 150-kDa dichain NT and the isolated approximately 100-kDa H chain (i.e., freed from the L chain) in order to map the cleavage sites and isolate the proteolytic fragments. The dichain NT appeared more resistant to cleavage by endoproteinase Glu-C than the isolated H chain. In contrast, the NT with its disulfide(s) reduced showed rapid digestion of both chains, including a cleavage between Glu 251 and Met 252 (resulting in approximately 30- and approximately 20-kDa fragments of the L chain) which was not noted unless the NT was reduced. Interestingly, an adjacent bond, Tyr 249-Tyr 250, was noted earlier [DasGupta and Foley (1989), Biochimie 71, 1193-1200] to undergo "self-cleavage" following reductive separation of the L chain from the H chain. The site Tyr-Tyr-Glu-Met (residues 249-252) appears to become exposed following reduction of Cys 429-Cys 453 disulfide. Identification of Glu 669-Ile 670 and Tyr 683-Ile 684 as protease-susceptible sites demonstrated for the first time that at least two peptide bonds in the segment of the H chain (residues 659-684), part of which (residues 659-681) is thought to interact with the endosomal membranes and forms channels [Oblatt-Montal et al., (1995), Protein Sci, 4, 1490-1497], are exposed on the surface of the NT. Two of the fragments of the H chain we generated and purified by chromatography are suitable for structure-function studies; the approximately 85- and approximately 45-kDa fragments beginning at residue Leu 544 and Ser 884, respectively (both extend presumably to Leu 1295) contain the channel-forming segment and receptor-binding segments, respectively. In determining partial amino acid sequences of 10 fragments, a total of 149 amino acids in the 1275-residue NT were chemically identified.

Antibody mapping to domains of botulinum neurotoxin serotype A in the complexed and uncomplexed forms

The domain organization of the botulinum neurotoxin serotype A was studied by using antibody mapping of 44 monoclonal single-chain variable fragments. The analysis was carried out on (i) the individual domains of botulinum neurotoxin holotoxin (binding, translocation, and catalytic), (ii) botulinum neurotoxin holotoxin, (iii) the botulinum neurotoxin holotoxin in complex with the nontoxic portion, and (iv) botulinum neurotoxin holotoxin and nontoxic portion of the complex recombined in vitro. All 44 antibodies mapped to individual domains of botulinum neurotoxin. Forty of the 44 single-chain variable fragments bound the botulinum neurotoxin holotoxin relative to the isolated domains, suggesting that 4 epitopes are covered when the individual domains are in the holotoxin form. Only 20 of the antibodies showed a positive reaction to the toxin while in complex with the nontoxic portion. All of the covered epitopes were mapped to the binding domain of botulinum neurotoxin, which suggested that the binding domain is in direct contact with the nontoxic portion in the complex. Based on the antibody mapping to the different domains of the botulinum neurotoxin holotoxin and the entire complex, a model of the botulinum neurotoxin complex is proposed.

Covalent structure of botulinum neurotoxin type A: location of sulfhydryl groups, and disulfide bridges and identification of C-termini of light and heavy chains

Botulinum neurotoxin Type A is synthesized by Clostridium botulinum as a approximately 150 kD single chain polypeptide. The posttranslational processing of the 1296 amino acid residue long gene product involves removal of the initiating methionine, formation of disulfide bridges, and limited proteolysis (nicking) by the bacterial protease(s). The mature dichain neurotoxin is made of a approximately 50-kD light chain and a approximately 100-kD heavy chain connected by a disulfide bridge. DNA derived amino acid sequence predicted a total of 9 Cys residues (Binz et al., 1990, J. Biol. Chem. 265, 9153-9158; Thompson et al., 1990, Eur. J. Biochem. 189, 73-81). Treatment of the dichain neurotoxin, dissolved in 6 M guanidine. HCl, with 4-vinylpyridine converted 5 Cys residues into S-pyridylethyl cysteine residues; but alkylation after mercaptolysis converted all 9 Cys residues in the S-pyridylethylated form. After confirming the predicted number of Cys residues by amino acid analysis, the positions of the 5 Cys residues carrying sulfhydryl groups and the 4 involved in disulfide bridges were determined by comparing the elution patterns in reversed-phase HPLC of the cyanogen bromide mixtures of the exclusively alkylated and the mercaptolyzed-alkylated neurotoxin. The chromatographically isolated components were identified by N-terminal amino acid sequence analysis. The HPLC patterns showed characteristic differences. The Cys residues predicted in positions 133, 164, 790, 966, and 1059 were found in the sulfhydryl form; Cys 429 and 453 were found disulfide-bridge connecting the light and heavy chains, and Cys 1234 and 1279 were found in an intrachain disulfide-bridge near the C-terminus in the heavy chain.(ABSTRACT TRUNCATED AT 250 WORDS)

Botulinum type A neurotoxin digested with pepsin yields 132, 97, 72, 45, 42, and 18 kD fragments

Botulinum neurotoxin (NT) serotype A is a dichain protein made of a light and a heavy chain linked by at least one interchain disulfide; based on SDS-polyacrylamide gel electrophoresis their molecular masses appear as 147, 52, and 93 kD, respectively. Digestion of the NT with pepsin under controlled pH (4.3 and 6.0), time (1 and 24 hr), and temperature (25 and 30 degrees C) produced 132, 97, 42, and 18 kD fragments. The three larger fragments were isolated by ion-exchange chromatography. The 132 and 97 kD fragments are composed of 52 kD light chain and 72 and 45 kD fragments of the heavy chain, respectively. The sequences of amino terminal residues of these fragments were determined to identify the pepsin cleavage sites in the NT, which based on nucleotide sequence has 1295 amino acid residues (Binz et al., J. Biol. Chem. 265, 9153, 1990). The 42 kD fragment, beginning with residue 866, is the C-terminal half of the heavy chain. The 18 kD fragment, of which the first 72 residues were identified beginning with residue 1147, represents the C-terminal segment of the heavy chain. The 132 kD fragment (residue 1 to approximately 1146) is thus a truncated version of the NT without its 18 kD C-terminal segment. The 97 kD fragment (residue 1 to approximately 865) is also a truncated NT with its 42 kD C-terminal segment excised. These peptic fragments contain one or two of the three functional domains of the NT (binds receptors, forms channels, and intracellularly inhibits exocytosis of the neurotransmitter) that can be used for structure-function studies of the NT. This report also demonstrates for the first time that of the six Cys residues 453, 790, 966, 1059, 1234, and 1279 located in the heavy chain the later four do not form interchain disulfide links with the light chain; however, Cys 1234 and 1279 contained within the 18 kD fragment form intrachain disulfide. The electrophoretic behaviors of type A NT and its fragments in native gels and their comparison with botulinum NT serotypes B and E as well as tetanus NT suggest that each NT forms dimers or other aggregates and the aggregation does not occur when the 42 kD C-terminal half of the heavy chain is excised. Thus, the C-terminal half of the heavy chain appears important in the self-association to form dimers.

Adherence of botulinum and tetanus neurotoxins to synaptosomal proteins

The ability of 125I-labeled botulinum type A and tetanus neurotoxins to adhere to blots of synaptosomal proteins separated by SDS-polyacrylamide gel electrophoresis was studied. Both neurotoxins appeared to adhere preferentially to an approximately 80 kDa and to a lesser extent to an approximately 116 kDa protein(s). Adherence of the neurotoxins to these proteins was enhanced by preincubation of the neurotoxins with GT 1b. The approximately 100 kDa heavy chain segment of BTxA adhered to the same proteins. The carboxy terminal half of the heavy chain adhered primarily to the approximately 80 kDa protein(s) while the amino terminal portion bound most intensely to the approximately 116 kDa protein(s). The ability of the approximately 80 and approximately 116 kDa proteins to stain positively with the periodic acid-Schiff reagent and to bind 125I-labeled wheat germ lectin suggests that they are glycosylated. Both neurotoxins appear to adhere to the same approximately 80 and approximately 116 kDa proteins because tetanus neurotoxin preincubated with GT 1b was able to reduce binding of radiolabeled botulinum type A neurotoxin to both proteins. Neither neurotoxin adhered to blots of proteins from liver, spleen, or kidney, suggesting that the proteins adhered to are neural components.

Binding of botulinum and tetanus neurotoxins to ganglioside GT1b and derivatives thereof

The ability of fragments derived from botulinum neurotoxin (BTx) serotype A to bind to GT1b-coated plastic wells was investigated and compared with the binding characteristics of the parent approximately 150-kDa protein. Although the approximately 50-kDa light chain of BTxA had a marginal binding capacity, the predominant adherence to GT1b-coated wells was exhibited by the approximately 50-kDa carboxy-terminal half of the approximately 100-kDa heavy chain of BTxA; the amino-terminal half of the heavy chain lacked the ability to bind. Binding to GT1b by BTxA and its fragments was compared with that of tetanus neurotoxin (TTx) and the carboxy-terminal half of its heavy chain. Binding of BTxA and the C-terminal half of the heavy chain was optimal in buffers of low ionic strength (mu less than or equal to 0.04 and 0.06, respectively), whereas the heavy chain bound GT1b best at mu greater than or equal to 0.10. TTx and the approximately 50-kDa C-terminal half of its approximately 100-kDa heavy chain bound GT1b at ionic strengths similar to those of BTxA. Comparison of the binding of BTx serotypes A, B, and E to GT1b (using conditions that were found to be optimal for binding by BTxA) indicated differences in the interaction of the three serotypes with GT1b. Compared with BTxA, adherence to GT1b by serotypes B and E was reduced by approximately 60 and approximately 90%, respectively.(ABSTRACT TRUNCATED AT 250 WORDS)

The complete amino acid sequence of the Clostridium botulinum type A neurotoxin, deduced by nucleotide sequence analysis of the encoding gene

A 26-mer oligonucleotide probe was synthesized (based on the determined amino acid sequence of the N-terminus of the Clostridium botulinum type A neurotoxin, BoNT/A) and used in Southern blot analysis to construct a restriction map of the region of the clostridial genome encompassing BoNT/A. The detailed information obtained enabled the cloning of the structural gene as three distinct fragments, none of which were capable of directing the expression of a toxic molecule. The central portion was cloned as a 2-kb PvuII-TaqI fragment and the remaining regions of the light chain and heavy chain as a 2.4-kb ScaI-TaqI fragment and a 3.4-kb HpaI-PvuII fragment, respectively. The nucleotide sequence of all three fragments was determined and an open reading frame identified, composed of 1296 codons corresponding to a polypeptide of 149 502 Da. The deduced amino acid sequence exhibited 33% similarity to tetanus toxin, with the most highly conserved regions occurring between the N-termini of the respective heavy chains. Conservation of Cys residues flanking the position at which the toxins are cleaved to yield the heavy chain and light chain allowed the tentative identification of those residues which probably form the disulphide bridges linking the two toxin subfragments.

Changes in the molecular topography of the light and heavy chains of type A botulinum neurotoxin following their separation

Botulinum neurotoxin serotype A, an approx. 150 kDa protein, is composed of two subunits, the light and heavy chains (approximately 50 and approximately 100 kDa, respectively). The neurotoxin's mode of action is believed to depend on coordinated but independent actions of the two subunit chains. The molecular environments of the aromatic amino acid residues of the dichain neurotoxin and the two isolated subunit chains were analyzed using near-ultraviolet circular dichroism (CD) (between 250 and 320 nm) and second-derivative ultraviolet absorption spectroscopy (between 240 and 320 nm) to investigate the conformational variations of the subunit chains in separated and conjugated forms. The mean residue weight ellipticities showed virtually no change (i.e., 1.7%) in the vicinities of Phe (268 nm), and only a small change (11%) around Tyr (279 nm) residues following dissociation of the subunit chains. However, significant changes (23-26%) at 286 nm as well as at 292 nm were noted, suggesting considerable alteration in the conformation of the subunits. Second-derivative ultraviolet absorption spectra indicated the degree of Tyr exposure in the dichain neurotoxin, isolated heavy and light chains at 70.7, 81.5 and 46.4%, respectively. A weighted mean of the degree of exposed Tyr residues in the separated heavy and light chains was 69.6%, virtually same as the 70.7% exposed Tyr residues observed in the intact dichain neurotoxin, indicating no difference in their Tyr exposure upon separation of the two chains. This was corroborated by the CD data which revealed only small changes in the CD signals of Tyr residues, and no alteration in those of the Phe residues following separation of the subunit chains. However, a change in the CD signal at 292 nm suggested that the conformations of Trp-containing segments of the two chains were significantly influenced upon their separation. The heavy and light chains of the neurotoxin therefore appear to exist as two semi-independent domains, in spite of being linked by disulfide and noncovalent bonds, and at least part of their conformations depends on interactions between them.