Elizabeth Evans, Health Protection Agency, Centre for Emergency Preparedness & Response, Porton Down, Salisbury, Wilts SP4 0JG, UK. E-mail: email@example.com
Aim: To develop a novel assay technique for the botulinum neurotoxin family (BoNTs) which is dependent on both the endopeptidase and receptor-binding activities of the BoNTs and which is insensitive to antigenic variation with the toxin family.
Methods and Results: An endopeptidase activity, receptor-binding assay (EARB assay) has been developed which captures biologically active toxin from media using brain synaptosomes. After capture, the bound toxin can be incubated with its substrate, and cleavage detected using serotype-specific antibodies raised against the cleaved product of each toxin serotype. The EARB assay was assessed using a range of BoNT serotypes and subtypes. For BoNT/A, detection limits for subtypes A1, A2 and A3 were 0·5, 3 and 10 MLD50 ml−1, respectively. The limit of detection for BoNT/B1 was 5 MLD50 ml−1 and a novel antibody-based endopeptidase assay for BoNT/F detected toxin at 0·5 MLD50 ml−1. All these BoNTs can be captured from media containing up to 10% serum without loss of sensitivity. BoNT/A1 could also be detected in dilutions of a lactose- containing formulation similar to that used for clinical preparations of the toxin. Different serotypes were found to possess different optimal cleavage pHs (pH 6·5 for A1, pH 7·4 for B1).
Conclusions: The EARB assay has been shown to be able to detect a broad range of BoNT serotypes and subtypes from various media.
Significance and Impact of the Study: The EARB assay system described is the first convenient in vitro assay system described which is requires multiple functional biological activities with the BoNTs. The assay will have applications in instances where it is essential or desirable to distinguish biologically active from inactive neurotoxin.
The botulinum neurotoxins (BoNTs) are a group of seven antigenically distinct toxins produced by Clostridium botulinum. They are implicated in food poisoning, and are considered a significant bioterrorism threat. All the BoNTs exert their biological effects on their target nerve cell via a similar mechanism (Niemann 1991; Schiavo et al. 2000). Initially the toxins bind to ganglioside and protein receptors on the presynaptic cell surface (Dong et al. 2006; Rummel et al. 2007). The toxin is then internalized and translocated within the nerve ending. Once inside the nerve terminal each toxin serotype mediates its effect via a highly specific endopeptidase activity directed at small proteins involved in the fusion and release of synaptic vesicles (Schiavo et al. 2000). Each of the BoNTs is highly specific in its proteolytic action. Serotypes A and E cleave the 25 kDa synaptosomal associated protein (SNAP25), types B, D, F and G cleave vesicle associated membrane protein (VAMP or synaptobrevin) and serotype C acts on both SNAP25 and syntaxin.
The seven BoNT serotypes (A–G) are increasingly being subdivided into subtypes. These subtypes are not fully antigenically distinct, but do vary in their antibody binding profiles (Smith et al. 2005). There are currently four BoNT/A subtypes, four BoNT/B subtypes and six BoNT/E subtypes. It seems likely that the final number of toxins within the BoNT family may well exceed 30.
Most of the currently available in vitro assays for the detection of botulinum neurotoxins are in the form of fluorescence assays (Schmidt and Stafford 2003), conventional ELISA systems or other immunoassays (Shone et al. 2006; Attrée et al. 2007; Bagramyan et al. 2008; Rasooly et al. 2008). The discovery of the unique endopeptidase activities of each neurotoxin serotype has led to the development of a new generation of assays in which a biological activity of the neurotoxin itself is used to amplify the assay signal (Hallis et al. 1996; Wictome et al. 1999; Jones et al. 2008). A key feature of these endopeptidase assays is the use of antibodies which recognize only the peptide substrate after cleavage by a particular BoNT and not the intact substrate. Although several BoNTs cleave the same substrate, they cleave at different sites on the protein and it has therefore been possible to define assays specific for a particular toxin serotype based on the use of different cleavage product-specific antibodies in the assays. A single cleavage product-specific antibody is required to detect the BoNTs within each serotype (Shone et al. 2006; Jones et al. 2008).
One difficulty with the above endopeptidase assays is that, unless a large dilution of the test sample can be made, some form of capture system is required for the BoNT to allow the endopeptidase reaction to occur in the optimal buffer environment. Previously, this has been achieved through an antibody capture step (Wictome et al. 1999; Bagramyan et al. 2008; Rasooly et al. 2008). However, given the antigenic variation within each toxin serotype, defining antibody reagents which capture all toxin subtypes with equal efficacy (Arndt et al. 2006) is a major problem. Polyclonal anti-toxin antibodies will not recognize all subtypes equally well, and if different monoclonal antibodies are used for each subtype, a capture phase capable of extracting all subtypes of several BoNT serotypes may require tens of different antibodies.
Here, we describe a novel assay with a toxin capture step that utilizes the naturally occurring toxin receptors, rather than toxin antibodies. Rat brain synaptosomes contain all the machinery required for toxin binding in vivo (Shone et al. 1985; Kozaki et al. 1998; Maruta et al. 2004), and should therefore be able to capture any toxin subtype that causes disease. Capturing the toxin via its native binding activity thus circumvents the problems of antigenic variation within the BoNTs. In the EARB assays system described here, specific receptor-mediated binding is combined with a measurement of the serotype specific endopeptidase activity of the BoNTs. With two biological activities of the toxin required the detection of inactive toxin is likely to be very low, as is the probability of false positives in this assay system.
Brain synaptosomes were prepared from rat brains as described previously (Shone et al. 1985). The brains were dissected from 12 rats and the white matter discarded. Four brains were added to 30 ml of ice cold 0·32 mol l−1 sucrose made in Kreb’s buffer (1 l made using 200 ml 5 × Kreb’s buffer (0·78 mol l−1 NaCl, 0·03 mol l−1 KCl, 8·1 mmol l−1 MgSO4, 7·8 mmol l−1 KH2PO4), 30 ml 31 mmol l−1 CaCl2, 12·5 ml 1 mol l−1 glucose, with 100 ml 0·25 mol l−1 Na2HPO4 added slowly while stirring). Each portion of brains was homogenized in a Teflon homogenizer system (900 rpm; 6 up/down strokes) then centrifuged at 2000 g at 4°C for 10 min. The supernatant was decanted into fresh tubes and centrifuged again at 20 000 g at 4°C for 20 min. The supernatant from this step was discarded, and the precipitates resuspended in a total volume of 40 ml ice cold 0·32 mol l−1 sucrose. The following discontinuous gradient was set up in six ultracentrifuge tubes: add 8 ml 0·8 mol l−1 sucrose; carefully add 8 ml of 1·2 mol l−1 sucrose to the bottom of the tube, under the first layer; carefully layer 7–8 ml of homogenate on the top. These tubes were centrifuged at 100 000 g at 4°C for 1 h. The synaptosomes form a layer at the 0·8/1·2 mol l−1 interface, and were removed with a Pasteur pipette. 1·2 volumes of distilled water was added to the pooled synaptosomes to bring the final sucrose concentration to approximately 0·45 mol l−1, and then they were centrifuged at 20 000 g at 4°C for 20 min. The precipitates were resuspended and briefly re-homogenized (one up/down stroke as above) in 12 ml 50 mmol l−1 HEPES, 150 mmol l−1 NaCl; pH 7·4 to give a final protein concentration of 2 mg ml−1, aliquoted, and stored at −80°C until required.
BoNT/A synaptosome capture assay
Toxin samples (5 ml, diluted from stocks with the following potency: A1 1·5 × 108 MLD50 (mouse lethal dose 50%) mg−1; A2 7 × 107 MLD50 mg−1; A3 2 × 106 MLD50 mg−1 to the required MLD50 ml−1) containing 10% spiked human serum in HZ buffer (50 mmol l−1 HEPES, 20 μmol l−1 ZnCl2; pH 7·4) were incubated with 20 μl rat brain synaptosomes for 30 min at 37°C in a rotating mixer. The samples were centrifuged at 4000 rpm at 4°C for 20 min and the supernatant discarded. The synaptosome pellet was washed with twice with 1 ml HZ buffer by gentle pipetting up and down and centrifuged at 10 000 g at room temperature for 1 min in a 1·5 ml tube.
Capture in lactose buffer was performed by making a 1000 MLD50 ml−1 stock of toxin in 0·1 mol l−1 Tris, 10 mg ml−1 lactose, 0·05% BSA; pH 7·4. 1 ml serial dilutions of this stock were performed in HZ buffer and incubated with synaptosomes as described above.
Capture of LHN/A was performed in 1 ml volumes of HZ buffer. LHN/A was produced as described in Shone et al. (1985), and had a potency of less than 103 MLD50 mg−1.
For the cleavage reaction, the synaptosome pellet was resuspended by gentle pipetting up and down in 100 μl of HZ buffer containing 1% foetal calf serum, 1 mmol l−1 DTT (pH 6·5) and 10 μg ml−1 synthetic N-terminal biotinylated SNAP25 72mer (SNAP25 132–204, N-terminal biotin, Thistle Peptides) and incubated for 2 h at 37°C with gentle agitation. This reaction was diluted with 400 μl cleavage buffer (to give a final SNAP25-72mer concentration of 2 μg ml−1) immediately prior to application of quadruplicate 100 μl samples to a Maxisorp (Nunc) plate previously coated with 10 μg ml−1 neutravidin (Sigma) in PBS (Sigma) for 1 h at 37°C with gentle agitation. Control reactions were set up in HZ buffer with the same toxin concentration per ml as in the synaptosome binding step, with 2 μg ml−1 SNAP25-72mer in a reaction volume of 500 μl, which was applied in quadruplicate to the plate as described above. The cleavage reaction was incubated on the plate for 5 min at 37°C with gentle agitation to allow binding of the biotinylated substrate to the neutravidin.
The plate was then blocked with 200 μl per well 5% foetal calf serum in PBS–0·1% Tween-20 (blocking buffer) for 1 h at 37°C with gentle agitation. The plate was washed three times with PBS-Tween20. Rabbit anti-DEANQ antibody (Sigma-Genosys, CTRIDEANQ peptide conjugated to KLH used to vaccinate rabbits according to Sigma-Genosys standard protocol) at a concentration of 1 μg ml−1 in blocking buffer was applied to the plate at 100 μl per well and incubated as above. The plate was washed as above and anti-rabbit antibody–HRP conjugate (Sigma) added, diluted 1 in 1000 in blocking buffer and incubated as above. The plate was then washed and TMB (3,3′,5,5′-tetramethylbenzidine) substrate (BioFX) added at 100 μl per well and the colour allowed to develop. The reaction was stopped with TMB stop solution (BioFX) and the absorbance read at 450 nm. The quoted assay detection limit is the minimum concentration at which an absorbance reading of twice background was obtained.
BoNT/B synaptosome capture assay
Essentially as the BoNT/A assay, except the binding HZ buffer was pH 6·5 and the cleavage HZ buffer was pH 7·4. The toxin was diluted from a stock with a potency of 2 × 108 MLD50 mg−1. The optimal pH for the binding and cleavage steps was determined empirically for each serotype. The substrate was 100 μg ml−1 hV35 peptide (VAMP 59–94, C-terminal biotin; Sigma-Genosys), which when diluted for application to the plate gives a final hV35 concentration of 20 μg ml−1. The primary antibody was anti-FESSAAKC, produced as described above.
BoNT/F synaptosome capture assay
Essentially as the BoNT/A assay, except the binding HZ buffer was pH 6·5 and the cleavage HZ buffer was pH 7·4. The toxin was diluted from a stock with a potency of 2 × 107 MLD50 mg−1. The substrate was 500 μg ml−1 VAMP (VAMP 1–94, C-terminal biotin), which when diluted for application to the plate gives a final VAMP concentration of 100 μg ml−1. The primary antibody was anti-KLSELAAC, produced as described above.
Variable pH cleavage assays
Cleavage reaction were performed in 500 μl volumes of 50 mmol l−1 HEPES, 20 μmol l−1 ZnCl2, 1 mmol l−1 DTT, 1% FCS at either pH 6·5 or 7·4. Substrates were 2 μg ml−1 SNAP25-72mer for BoNT/A1, 20 μg ml−1 hV35 for BoNT/B and 100 μg ml−1 VAMP (VAMP 1–94) for BoNT/F. The cleavage reactions were incubated for 2 h at 37°C with gentle agitation, and then applied to a neutravidin coated plate as described above. The primary and secondary antibodies for each serotype were applied as described above. The plate was then washed and the colour allowed to develop as described above.
Expression and purification of recombinant VAMP(1–94)
A rat VAMP1 cDNA clone was obtained from Professor C. Montecucco (University of Padova). The fragment was modified by PCR to introduce BamHI and XhoI sites at the 5′ and 3′ ends respectively. The gene was cloned into the expression vector pGEX-6P (GE Healthcare) digested BamHI–XhoI. All clones were checked by sequencing to confirm the insertion of the correct fragment. The clones were transformed into the BL21 expression strain (Promega, UK) before expression and purification.
Fifty millilitre cultures of BL21 pGEX 6P-VAMP in Terrific Broth containing 100 μg ml−1 ampicillin and 0·5% glucose were incubated at 30°C overnight. Twenty millilitres of this starter culture was diluted into 1 l Terrific Broth as above and incubated at 30°C for 3 h, then induced with 500 μmol l−1 IPTG at 25°C for 3 h. Cells were harvested by centrifugation at 3000 g at 4°C for 20 min, and resuspended in 35 ml PBS (pH 7·4). The resuspended cells were disrupted by sonication and centrifuged at 20 000 g at 4°C for 30 min. The supernatant was diluted with 35 ml of PBS and applied slowly to a 5 ml glutathione sepharose GSTrap FF column (AP Biotech), previously equilibrated with five column volumes of PBS. The column was washed with 10 column volumes of PBS.
The GST-VAMP was cleaved on the column with 500 units of Precision protease (GE Healthcare) according to the manufacturer’s instructions overnight at 4°C. Free VAMP was eluted from the column after cleavage and buffer exchanged against PBS. Fresh PEO-Maleimide Activated Biotin was prepared at 10 mmol l−1 in PBS. Hundred microlitres of the biotin solution was added to 2·5 ml of approx. 1 mg ml−1 GST-VAMP in PBS. This was incubated at room temperature for 4 h, and then buffer exchanged, to remove free biotin, into 50 mmol l−1 HEPES (pH 7·4).
Synaptosome capture assays for BoNT/A1, A2 and A3
The EARB assay design incorporates a concentration step (Fig. 1), whereby toxin from a sample (up to 5 ml) is captured by the synaptosomes, and the cleavage reaction is performed in a smaller volume (100 μl). This concentration step allowed the sensitivity of the assays system to be increased. Figure 2 shows the endopeptidase activity of BoNT/A toxin, in solution and captured from 10% human serum by synaptosomes. For all subtypes tested (BoNT/A1, A2, A3) the sensitivity of the assay for synaptosome bound toxin is very similar to that for the non-bound toxin in each case (see Fig. 2b–d). The detection limit of the synaptosome assay for A1 is 0·5 MLD50 ml−1 (linear dose response starts at 3 MLD50 ml−1; Fig. 2a), for A2, 3 MLD50 ml−1 (linear dose response starts at 10 MLD50 ml−1; Fig. 2b), and for A3, 10 MLD50 ml−1 (linear dose response starts at 30 MLD50 ml−1; Fig. 2c). These differences in sensitivity, particularly for A3 (see Fig. 2a), are interesting given the normalization of the assay on mouse toxicity, and were also seen when using full-length SNAP25 as a substrate as opposed to the 72mer (data not shown).
This assay is unable to distinguish between the different BoNT/A subtypes since they all cleave the same substrate at the same residue. It can, however, distinguish between serotypes because of the different substrates and cleaved-end antibodies required. This assay is unable to detect the essentially non-toxic LHN/A, since this lacks the cell-binding HC domain (see Fig. 3a).
Synaptosome capture assay for BoNT/A in media
Synaptosomes were found to capture BoNT/A efficiently in concentrations of serum up to 10%. Above that serum concentration however, the assay sensitivity was markedly reduced which may in part be due to the increased ionic strength of media. The potential of the EARB assay to detect BoNT/A1 in a formulation buffer similar to that used in clinical preparations of BoNT/A was also assessed. In a formulation containing 1000 MLD50 ml−1 in 0·1 mol l–1 Tris, 10 mg ml−1 lactose, 0·05% BSA; pH 7·4, the assay detected toxin with a similar sensitivity to buffer controls (Fig. 3b). The EARD assay could therefore have applications in the assessment of the biological activity of clinical formulations of BoNT.
Synaptosome capture assays for BoNT/B and /F
An assay for BoNT/B was developed in a similar manner for BoNT/A subtypes. The sensitivity of the B1 assay was found to be 5 MLD50 ml−1 (linear dose response starts at 10 MLD50 ml−1; Fig. 4a). Interestingly, the sensitivity of the B1 assay was slightly higher when the toxin was captured compared to toxin in solution. This is probably due to the concentration effect provided by the synaptosome capture. Figure 4b shows the first published antibody-based endopeptidase assay for BoNT/F, with a sensitivity of 0·3 MLD50 ml−1. The peptide used to generate the primary antibody is slightly modified from the native VAMP sequence. Antibodies to the native VAMP sequence for the BoNT/F cleavage product (KLSELDD; VAMP59–65) recognized the intact substrate (data not shown) and as such were unsuitable for this assay. A modified peptide sequence (KLSELAA) in which the two aspartate residues were changed to alanine was assessed. Antibodies raised against this modified peptide were found to recognize the BoNT/F cleavage product but not intact VAMP, and therefore can be used in this assay format. As for BoNT/A and B, the capture of the toxin by synaptosomes, and presence of up to 10% human serum does not affect the assay sensitivity.
BoNT/A1 is more sensitive to pH than BoNT/B1 or BoNT/F. Figure 5a shows that BoNT/A1 cleaves at a higher rate at pH 6·5 than at pH 7·4. BoNT/B1 (Fig. 5b), conversely, shows improved cleavage at pH 7·4, but only at low toxin concentrations. BoNT/F cleavage is the same at both pH 6·5 and 7·4 (Fig. 5c). Optimal cleavage pH needs to be determined empirically for each toxin serotype. Toxin subtypes A2 and A3 show the same pattern of improved cleavage at pH 6·5 as A1 (data not shown). This variation in pH sensitivity between the serotypes is unexpected, and may warrant further investigation.
The expression of toxicity by the BoNTs is dependent on the concerted biological action of several domains within its structure concerned with receptor-binding, translocation and intracellular effect via endopeptidase activity (Montecucco and Schiavo 1994; Poulain 1994). Disruption of any one of these activities leads to a significant reduction or ablation of specific toxicity (Schiavo et al. 2000). Previously developed in vitro assays for the BoNTs either measure none of these biological activities, e.g. polyclonal or monoclonal antibody-based ELISA assays, or the activity of just one domain, e.g. endopeptidase assays (Shone et al. 2006). As such these assays may not correlate well with the biological activity of the neurotoxin. In the case of endopeptidase assays for example, the endopeptidase activity of the native toxin may be the same as that of the virtually non-toxic LHN fragment from which the C-terminal receptor binding domain has been lost (Chaddock et al. 2002). In the present study, we report the development of the EARB assay system which has a similar sensitivity to the mouse bioassay and which is dependant upon functional receptor-binding and endopeptidase domains. The receptor binding activity of the BoNTs has been shown to be mediated via a C-terminal sub-domain (HCC) of the HC fragment and studies on BoNT/A have shown that loss of a short peptide from the C-terminus of the molecule leads to concomitant loss of synaptosome binding and mouse toxicity (Shone et al. 1985; Rummel et al. 2007). Thus, receptor binding activity is an indicator of the integrity of the extreme C-terminal region of the BoNT molecule. Similarly, endopeptidase activity provides an indicator of correct folding of the N-terminal light chain domain (Chaddock et al. 2002). The EARB assay which is dependent on both these domains being functional is therefore likely to be a good indictor of overall BoNT integrity. While the assay does not directly measure the activity of the translocation domain, it seems unlikely that a BoNT molecule that has correctly folded N- and C-terminal domains would have a mis-folded, non functional HN domain. The EARB assay is predicted to correlate well with toxicity of the neurotoxin, though further studies using BoNT stressed by a variety of different methods would be needed to confirm this.
Synaptosomes offer several advantages as a capture phase in the EARB assay system in that they are relatively easy to prepare and can be stored frozen over long periods without loss of activity. A key advantage of synaptosomes as a capture phase is that binding is mediated via their receptor components to active site regions on the BoNTs which are relatively highly conserved structural regions (Maruta et al. 2004). This is in contrast to antibody based capture assays systems which are prone to false negative assay readings due to antigenic variation between BoNT subtypes (Smith et al. 2005). The EARB assay was able to detect the three principal subtypes (A1, A2 and A3) of BoNT/A, although the detection sensitivity for the A3 subtype was lower than that for A1. Sequence alignment of the A1 and A3 subtypes shows that their predicted ganglioside binding domains are highly conserved (Arndt et al. 2006), suggesting that their binding to synaptosomes will be similar. However, the light chains of A1 and A3 are more variable, particularly in the α-exosite, suggesting that the differences in cleavage rates in the endopeptidase assay may be due to the decreased SNAP25 binding affinity of A3 postulated by Arndt et al. (2006).
A limitation of the EARB assay is in the detection of BoNT in complex media such as food stuffs and undiluted human serum in which the assay sensitivity is reduced principally by the higher ionic strength of the media compared to the standard binding buffer used for the assay system. The EARB assay was able to detect toxins in a 10% human serum and also in a formulation similar to that used for clinical BoNT products. Potential applications of the assay are detection of biologically active toxin in compatible defined media, notably the assessment of toxins sample during the manufacture BoNT/A and other BoNT serotypes for clinical use. In this capacity, the EARB assay may allow a significant reduction in animal usage in such efficacy testing.
The present study also describes the development of a novel, antibody-based endopeptidase assay for BoNT/F. Based on the strategy described by Hallis et al. (1996), the assay relies upon antibody reagents which specifically recognize the cleavage products of VAMP produced by BoNT/F and not the intact substrate. In the previously described method, such product-specific antibodies were raised against short peptides (8–10 residues) representing either the newly exposed N- or C-terminus of VAMP. This strategy, however, proved unsatisfactory for the BoNT/F assay since antibodies raised against the appropriate peptides (e.g. KLSELDDC) also bound strongly to the intact VAMP substrate. The most likely explanation for this is that the short peptide adopts an antigenic conformation that is also present on the intact substrate, perhaps imparted be the ‘DD’ motif. Antibodies raised against a modified peptide, KLSELAAC, were found provide product-specific antibodies which allowed the development of an EARB assay for BoNT/F with sensitivity greater than the mouse bioassay.
In conclusion, a sensitive assay system for BoNTs serotypes A, B and F is described which is dependent on two biological activities within the neurotoxin molecule. The assay will have applications in instances where it is essential or desirable to distinguish biologically active from inactive neurotoxin.