Recombinant expression of receptor binding domains of all eight subtypes of botulinum neurotoxin type A for generation of antitoxins with broad reactivity

Introduction

Botulinum neurotoxins (BoNTs) are the most toxic substances known to the humankind with lethal dose values in the range of nanogram per kilogram body weight scale.¹ Most commonly produced by Clostridium botulinum, these toxins are proteins composed of a 50-kDa light chain (LC) linked to a 100-kDa heavy chain (HC) via a disulfide bond. The LC fragment contains a zinc-protease specific domain, whereas the HC consists of an N-terminal translocation domain (H_N) and a C-terminal receptor-binding domain (H_C).² The mode of action of BoNTs includes three steps. In the first place, the H_C domain of BoNTs bind specifically to peripheral nerve terminals via polysialoganglioside and synaptic vesicle receptors. Subsequently, BoNTs enter into nerve terminals by endocytosis. Under acidic conditions, the H_N domain translocates the LC into the nerve terminal cytosol where the latter cleaves one of three soluble N-ethylmaleimide-sensitive factor attachment protein receptors (SNARE) that are involved in neurotransmitter release, thereby causing nerve paralysis.²

BoNTs are traditionally classified into seven serotypes (BoNT/A–BoNT/G), of which BoNT/A represents a great threat to humans because of its most potent toxicity with the longest duration of paralysis.³ In addition, BoNT/A is categorized into eight subtypes (BoNT/A1–A8) with significant levels of protein sequence differences (up to 12.3%),⁴ which complicates the development of a broadly protective monoclonal antitoxin.

Botulinum intoxication is fatal in 5–10% of cases and requires early treatment with antitoxin. Currently, the only available antitoxins for botulism are the heptavalent botulinum antitoxin (HBAT), which contains fragments of immunoglobulins from horses vaccinated with all seven traditionnal serotypes of BoNTs, and BabyBIG, which consists of polyclonal antibodies from human immunized with recombinant botulinum vaccine for serotypes A and B. However, these antitoxin types have limitations due to adverse side effects, limited availability and exorbitant cost.⁵ To overcome these drawbacks, neutralizing monoclonal antibodies (mAbs) against BoNT/A, B, E, and F, which cause human botulism, have been generated.^6–9 It has been shown that a combination of several mAbs is required to efficiently neutralize subtypes belonging to a BoNT serotype.^8,9 Another strategy to combat botulism is to develop camelid single-domain antibodies (sdAbs), also referred to as VHHs or nanobodies, that can neutralize BoNTs via the interactions with the functional domains of the toxins. Several VHHs with high affinity against H_C domain of a BoNT/A subtype have been shown to display protective activity when challenged with the same toxin in animal models.^10–12 However, it remains unclear if these VHHs can neutralize effectively other subtypes of BoNT/A. This is because of the lack of a comprehensive toxin resource available for all subtypes of BoNT/A. Here, for the first time, the generation of recombinant H_C domains of BoNT/A1-A8 was described. These proteins were then used to characterize a panel of VHHs targeting BoNT/A1 with unknown binding sites in order to identify novel VHHs with broad reactivity against all subtypes of BoNT/A.

Methods

Results and discussion

Due to the extremely high toxicity of BoNTs, the development of antitoxins against them is of major interest for therapeutic applications. The target for antitoxin can be each of the three structural domains of BoNTs: (i) H_C responsible for receptor binding; (ii) H_N for toxin translocation; and (iii) LC for cleavage of SNARE proteins. Among these domains, the H_C fragment is the target of choice for the generation of antidote to BoNT intoxication,²³ as well as for the development of vaccines²⁴ and for intracellular delivery of cargo molecules specifically to neurons.²⁵ In the present study, we focused on the production of the H_C fragments of all subtypes of BoNT/A because of its extremely high toxicity, long persistence, and high sequence divergence among subtypes. However, only bont/A1 gene was available in our group. Consequently, we opted for an approach that involved both site-directed mutagenesis from H_CA1 construct and in-house gene synthesis to generate expression vectors for H_C domains of BoNT/A2-A8.^14–20 These recombinant proteins were expressed in the E. coli Rosetta 2(DE3) and purified by nickel affinity chromatography. SDS-PAGE analysis showed that only one band was observed for all purified samples at the expected molecular mass of 50 kDa ( Figure 1), suggesting that H_C fragments of BoNT/A1-A8 were successfully expressed and prepared. The purified proteins were then evaluated by ELISA for antigenicity using the Rabbit Anti-Botulinum Toxin A and B IgG. All samples were recognized by the commercial polyclonal antibody,²⁶ which indicated that the receptor binding domains of all BoNT/A subtypes were correctly expressed in E. coli Rosetta 2(DE3).

Figure 1. SDS-PAGE analysis of purified recombinant BoNT/A H_C domains.

To our knowledge, only three VHHs neutralizing BoNT/A by binding to the H_C domain have been described in the literature,^10–12 of which ciA-C2 have been extensively characterized for the inhibition mechanism on the BoNT/A1.¹¹ In a separate report, 18 VHHs have been identified for their specific recognition of BoNT/A1²¹ but their binding sites were unknown. We therefore performed a systematic screening of a panel of five most promising VHHs from this report and ciA-C2, on the H_C fragments of BoNT/A1-A8 by ELISA. All the VHHs were successfully expressed and purified.²² Screening results clearly indicated that apart from ciA-C2, two nanobodies (VHH-A1 and VHH-A3) could recognize the H_C domains of all BoNT/A subtypes with VHH-A3 exhibiting significantly higher affinity than VHH-A1.²⁷ Consequently, only ciA-C2 and VHH-A3 were characterized in subsequent experiments. According to the EC₅₀ values calculated using ELISA ( Table 2),²⁸ both ciA-C2 and VHH-A3 exhibited the highest affinity for the H_CA1 (EC₅₀ = 11.0 and 24.0 nM respectively). This is not surprising, because these two VHHs were generated based on the selection with BoNT/A1.^11,21 VHH-A3 displayed an intermediate affinity for H_CA4 (EC₅₀ = 46.0 nM) and comparably low affinity for the remaining subtypes. In comparison to ciA-C2, VHH-A3 displayed similar EC₅₀ values for H_C domains of BoNT/A2, A3, A5, A6, A7, and A8. Concerning ciA-C2, the binding mechanism of this VHH to H_CA1 involves a cation-π interaction and multiple hydrogen bonds between CDR1 and residues K289, N318 and D419 of H_CA1. In addition, CDR2, CDR3, FR2, FR3, and FR4 of ciA-C2 also participate in the binding to H_CA1 through hydrogen bonds with residues T193, H194, Y242, T276, E423 and a hydrophobic interaction with P425 of the domain.¹¹ Consistent with these structural observations, the affinity of ciA-C2 was least affected for H_CA4 (EC₅₀ = 26.2 nM) with only a T193 to P193 replacement,¹⁶ whereas it was most affected for H_CA2, H_CA3 and H_CA8 (EC₅₀ ≥ 80.4 nM) containing three major substitutions T193P, H194R, and P425S.^14,15,20 These data underline the importance of the structural studies of VHHs in order to generate antitoxins with a broad protection to BoNTs. Furthermore, considering the sequence divergence among H_C domains of BoNTs²⁹ and most studies so far use BoNT/A1 as the selection agent to generate VHHs, it would be of interest to include a divergent H_C domain, for instance, H_CA2, H_CA3 or H_CA8, during the selection steps in order to obtain VHHs having high affinity against these domains. Similarly, these recombinant fragments could be combined with H_CA1 for the development of vaccines or polyclonal antitoxins with broad potency compared to conventional approach using only one BoNT/A subtype for immunization.

In summary, this study provided recombinant H_C domains of all BoNT/A subtypes, which could be used for the development of antitoxins and vaccines against BoNTs. This study also identified two new nanobodies, VHH-A1 and VHH-A3, capable of binding to all BoNT/A H_C domains. However, one question remains unsolved in this study, whether the VHH-A1, VHH-A3 and ciA-C2 would bind to a distinct, non-overlapping epitope. Further research is on-going to resolve this question and to improve neutralizing activity of ciA-C2 through the generation of heterodimers.

Ethics and consent

Ethical approval and consent were not required.