Botulism is a disease which can be fatal if untreated and is caused by exposure to any one of the highly toxic protein family known as botulinum neurotoxins (BoNTs). In vivo, BoNT cleaves proteins necessary for nerve signal transmission. This enzymatic cleavage results in the inhibition of the nerve impulse, leading to flaccid paralysis of the victim which can affect the lungs and may necessitate ventilator support. Treatment of the botulism patient involves administration of therapeutic immunoglobulin and is most effective when administered within 24 h of toxin exposure [1]. Due to the extreme toxicity, global availability, and ease of preparation of BoNT, it is considered a likely agent for bioterrorism [2].

Previously, our laboratory reported in several publications on the development of the Endopep-MS method as an assay for BoNT detection and serotype differentiation [3,4]. This method can detect all seven known BoNT serotypes and involves incubating BoNT with a peptide substrate that mimics each toxin's natural in vivo neuronal protein target. The presence of a particular BoNT serotype is demonstrated by mass spectrometric detection of the peptide cleavage products corresponding to their specific toxin-dependent location. Endopep-MS currently uses an antibody-affinity concentration/purification step before the enzymatic reaction with the substrate, and the choice of antibody is critical for the success of this assay [5]. We previously reported that polyclonal anti-BoNT binding could interfere with the activity of BoNT as measured by Endopep-MS [5]. We also reported on the success of using monoclonal (mAb) anti-BoNT/A to detect multiple subtypes of BoNT/A [6,7].

Similar to the other BoNT serotypes, BoNT/B consists of a heavy chain (HC) of approximately 100,000 daltons and a light chain (LC) of about 50,000 daltons. The heavy chain is mainly responsible for both receptor binding by its C-terminal (CT) binding domain [8,9] (H_C) and the delivery of the catalytic light chain (LC) to its target inside the neuron by way of its N-terminal translocation domain (H_N)[10]. Although the LC is responsible for the specific toxicity, it requires the heavy chain to enter the target cell and produce this toxic activity in vivo. As with most of the other BoNT serotypes, BoNT/B exhibits genetic and amino acid variance within the serotype, and this variance is defined as a subtype. BoNT/B is currently defined as consisting of the /B1, /B2, /B3, /B4, /B5, and /B6 subtypes. [11,12]. At the amino acid composition level, the variance among all the BoNT/B is 7% or less, but this degree of variance can affect binding of the toxin to some of the anti-BoNT/B mAbs as shown before [13]. So, it is important to select cross reactive mAbs which are able to detect all toxin subtypes, because an outbreak of BoNT/B botulism may be attributed to more than just the familiar "common" subtype.

Previously, our laboratory demonstrated that the Endopep-MS assay can be used to detect all currently known available subtypes of BoNT/B [7,14]. The goal of this work is to evaluate a panel of mAbs for their inhibitory and extraction abilities, thereby optimizing assay sensitivity with all BoNT/B subtypes available to us for testing. Here, we evaluated a panel of 24 fully human monoclonal anti-BoNT/B mAbs for their ability to inhibit the in vitro light chain activity of BoNT/B1, /B2, /B3, /B4, or /B5. BoNT/B6 was unavailable to us for testing. Additionally, we evaluated the same antibody panel for their ability to extract any of the available subtypes of BoNT/B. Our data show that there were significant differences among those mAbs in their ability to extract different BoNT/B subtypes, and their inhibitory effects on BoNT/B catalytic activity. Surprisingly, some of the mAbs appeared to enhance the light chain enzymatic activity of some subtypes of BoNT/B, a phenomenon that has been reported for BoNT/A, but not yet for BoNT/B [15]. Such differences could be explained in part through analyzing the epitopes of the mAbs and the amino acid sequences of each subtype of BoNT/B. Our results indicate which mAbs have the optimal properties for use in the Endopep-MS detection of BoNT/B.

After examining a panel of BoNT/A [6] and BoNT/B mAbs (this study) we found that there are significant differences in the ability of the mAbs to inhibit or activate BoNT. Most importantly, we did not observe activation of the LC activity by any of the anti-BoNT/A mAbs, whereas we did observe this with a number of BoNT/B mAbs. Interestingly, in this study, the majority of the mAbs that enhanced the enzymatic activity of the BoNT/B subtypes bound the HC portion of the toxin, which is not enzymatically active.

The phenomenon of toxin activation by antibody binding is not a novel one [15], and it is thought that some mAbs induce a conformational change upon the structure of the toxin after binding so that the toxin is induced into an optimal conformation to bind the substrate. It is highly possible that this induced-conformational change was responsible for the increase in BoNT/B LC enzymatic activity with some mAbs, particularly those which did not bind near the enzymatic active site. It is also important that although a mAb could be an activator of LC activity in an enzymatic assay, it does not mean that this mAb would enhance in vivo toxicity.

Not all of the anti-BoNT/B mAbs were LC activators. For example, mAb 1B10.1 inhibited the LC activity of four of the five BoNT/B subtypes. Antibody 1B10.1 bound the LC of the toxin (Table 2, Figure 1), and residues R121, R122, N177, H179, F180, R183, E184, D244, and D245 of BoNT/B contribute to the binding of 1B10.1 as mapped [21] using yeast displayed BoNT/B LC (C. Garcia and J.D. Marks, unpublished data). The active site of BoNT/B was centered around residue 231, which is quite close to the epitope for the binding of mAb 1B10.1. It is highly probable that upon binding of 1B10.1, BoNT/B toxin is no longer able to access the peptide substrate, and is therefore inactive upon the peptide substrate. This explains the ability of 1B10.1 to inhibit the enzymatic activity of most BoNT/B subtypes. Looking at the amino acid similarities of the BoNT/B subtypes in these residues, we found that only one mutation existed--the aspartic acid in position 244, which is negatively charged, is mutated to neutral asparagine in BoNT/B5. This mutation may explain the lower affinity of mAb 1B10.1 for BoNT/B5 and its lack of inhibition.

Another inhibitory antibody is 2B27, which bound the LC of the toxin. The 2B27 epitope overlaps the epitope of mAb 1B10.1 (Table 2 and Figure 1). This proximity to the active site of the toxin explains its ability to inhibit the LC activity of this toxin. In fact, most of the other inhibitory antibodies are also LC binders, presumably because these mAbs interfere with the binding of the substrate to the toxin. It is important, however, that not all LC binders inhibit BoNT/B1 enzymatic activity. mAb 2B23 bound the LC, but had virtually the same response in enzymatic activity as the control with no antibody. Although mAb 2B23 bound the LC (Table 2, Figure 1) the epitope seems to be farther away from the enzymatic active site, so it does not have an inhibitory effect on the toxin's activity.

B6.1 was inhibitory for some BoNT/B subtypes, but acted as an activator of LC activity for other subtypes. mAb B6.1 bound the LC of toxin on epitope 6; however, it did not bind in close proximity to the enzymatic active site (Figure 1). Therefore, it is possible that B6.1 binding inhibits some BoNT/B not by B6.1 itself blocking the peptide substrate from the active site, but rather by the induced conformational change upon that toxin. This would result in an altered quaternary structure of BoNT/B1 and /B4 which has difficulty contacting the peptide substrate, while the changed conformation of BoNT/B2, /B3, and /B5 upon mAb B6.1 binding might have easier access to the substrate, thereby serving as an activator of LC activity for BoNT/B2, /B3, and /B5. This phenomenon could also explain the activity of 2B29 and 2B18.1, which also serve as inhibitors for some subtypes of BoNT/B and LC activators for other subtypes.

The extraction efficiency experiments also shed some light on the interaction of these mAbs with the different BoNT/B subtypes. Three mAbs which had poor extraction efficiency were 1B10.1, 1B22.4, and B1.1. Because the inhibitory experiments showed that the mAbs 1B10.1 and 1B22.4 inhibited the enzymatic activity of BoNT/B, the decreased cleavage products after extraction with these mAbs were likely due to inhibition of activity rather than poor extraction efficiency. In contrast, the mAb B1.1 was a moderate to strong LC activator in the antibody inhibitory study but was one of the worst mAbs for extracting all five BoNT/B subtypes tested. This could be attributed to the relatively low affinity of mAb B1.1 for BoNT/B, or it may reflect either poor coupling of the mAb to the beads or inactivation of the mAb upon coupling.

Additionally, the extraction efficiency experiments demonstrated that some mAbs work very well for most subtypes, but not all of them. For example, mAb B11E8 yielded the highest or second highest response in terms of activity for BoNT/B1, /B2, /B3, and /B5, but was one of the worst choices for BoNT/B4. This mAb was not inhibitory for the activity of BoNT/B4; rather, this mAb has no affinity for BoNT/B4 compared to the other subtypes.

The Endopep-MS assay relies upon mAb extraction of BoNT from a clinical or food sample as a sample preparation step before analysis for the toxin, so high affinity, non-inhibitory mAbs are critical components of the assay. Additionally, mAbs that activate the enzymatic activity of the toxin after binding may further improve the sensitivity of detection in the Endopep-MS assay. Therefore, it is important to assess the binding affinities and potential enzymatic inhibition abilities of mAbs against a variety of BoNT subtypes within the chosen serotype before choosing mAbs for extraction. This assessment of a large panel of monoclonal BoNT/B antibodies should enable us to identify an antibody or a few mAbs that demonstrate strong extraction efficiency for all known BoNT/B, which currently includes the BoNT/B1-/B6 subtypes, without inhibiting the enzymatic activity of the toxin.

After testing a panel of 24 fully human antibodies against all BoNT/B subtypes in our possession, BoNT/B1-/B5, and examining both their inhibitory ability as well as their extraction efficiency, mAbs that had good results with all five subtypes were mAbs 1B18.1, 2B18.1, 2B18.2, 2B18.3, B12.2, and 2B23. Four of these mAbs (1B18.1, 2B18.1, 2B18.2, and 2B18.3) are clonally related, having almost the same HC variable regions and different LC variable regions, and binding to the same H_N epitope. Antibodies interacting with all three domains of the toxin were represented, as B12.2 bound to the H_C and 2B23 bound to the LC. It is known that using multiple mAbs which bind non-overlapping epitopes increases the effective affinity for the toxin by as much as 200-fold over the affinity of the individual antibodies [15]. Use of multiple mAbs binding different epitopes not only increases overall binding affinity, which is important for toxin extraction, but also offers a unique opportunity to design a mixture of mAbs that effectively bind a variety of epitopes, including regions that are conserved across the BoNT/B subtypes and also regions that may represent high-affinity binding sites for only a portion of the toxin subtypes. These multiple antibodies help to ensure that each toxin subtype will be recognized and extracted. As new BoNT/B subtypes are discovered, amino acid mutations affecting binding epitopes may be present. The use of multiple mAbs that recognize a variety of epitopes minimize the impact of these mutations, increasing confidence that all BoNT/B subtypes will be efficiently extracted.

1B18.1, 2B18.1, 2B18.2, and 2B18.3 bound the same epitope in the translocation domain. Examining their results in the extraction assay presented here, we found that 2B18.2 yielded the best overall results for subtypes BoNT/B1-/B5. Because of this, we are opting to use 2B18.2 as one of the mAbs for extraction of BoNT/B1-/B5 from sample matrices before analysis with the Endopep-MS method. Among the final list of mAbs which yielded excellent extraction results with all five subtypes tested, there were two binding unique epitopes which could be used: B12.2 and 2B23. B12.2 had better performance than 2B23, so we are opting to use B12.2, directed against the receptor binding portion of the HC, as a second mAb for extraction of BoNT/B1-/B5 from sample matrices before analysis with the Endopep-MS method. Unfortunately, BoNT/B6, the only other currently known BoNT/B subtype, was not available to us for testing. However, the sequence of BoNT/B6 for the epitope bound by 2B18.2 is completely conserved, so it would be anticipated that 2B18.2 could efficiently extract BoNT/B6 [19].

In addition to determining the best mAbs for sample preparation before Endopep-MS, this work determined in vitro inhibition abilities of a panel of mAbs against BoNT/B1-/B5. Many mAbs showed similar results with BoNT/B1-/B5, but in some cases they differed, indicating differing toxin extraction efficiencies due to differing binding affinities or inhibition of toxin activity. In some cases, activation of toxin LC activity was seen. These findings indicate that mAb choice is crucial to the ability of these types of assays to sensitively detect a diverse range of BoNT/B toxin subtypes, which is a critical first step to providing proper treatment in a timely manner. In addition, these mAb characterizations have the potential to assist with mechanistic studies of BoNT/B protection and treatment, which is important for studying alternative therapeutics for botulism.

SRK conceived of the study and participated in its design and coordination, and helped to draft the manuscript; WIS participated in testing the mAbs with Endopep-MS for extraction and inhibition; JL participated in producing and characterizing the mAbs used in this study and helped to draft the manuscript; ING participated in producing and characterizing the mAbs used in this study; CGR participated in producing and characterizing the mAbs used in this study and in epitope mapping; TJS conceived of the study, participated in its design and coordination, and helped to draft the manuscript; JDM conceived of the study, participated in its design and coordination, and helped to draft the manuscript; LAS participated in the study design and its coordination; JLP participated in the study design and its coordination; JRB conceived of the study and participated in its design and coordination, and helped to draft the manuscript. All authors read and approved the final manuscript.