Team:Queens Canada/Design

Design Overview

We have developed a novel antibody-based biosensor for THC detection. The device is based on a recombinantly expressed anti-THC antibody fragment, which is linked to a fluorescent protein. The design approach of the device started with the creation of an antibody library, which was assisted by computational models of the antibodies.

Antibody Library Design

We will be using protein sequences of an anti-THC scFv, and an anti-THC Fab that are available in the public domain (1, 2). Both antibody sequences were used in our design approach, as conjugating antibody fragments to fluorescent proteins can sometimes lead to a loss of affinity for the antigen. Hence, we created a library of fluorescently conjugated antibodies to overcome affinity issues. For the Fab, the location for fluorescent protein ligation was based on the known crystal structure (PDB: 3ls4). Based on the structure of the anti-THC Fab we designed a linker large enough to bind the fluorescent protein. We chose to attach the fluorescent protein to the N-terminal of the heavy chain, rather than the light chain, as the heavy chain is usually attached to another domain (Fig. 1A). Additionally, we aimed to develop a truncated version of the Fab, by using the fluorescent protein as a linker between the N-terminal of the light chain and the C-terminal of the heavy chain (Fig. 1B). It has been recorded that using a fluorescent protein as a linker between the heavy and light chains can increase the solubility of the recombinantly expressed antibody (3).

Figure 1. A) Structure of anti-THC ScFv, as predicted by ABodyBuilder. B) Structure of anti-THC ScFv linked to mNG at the N-terminal of the light chain.

Computer Assisted Antibody Library Design

On the other hand, creating a ScFv library was more challenging, as there is no known crystal structure. In order to attach a fluorescent protein to this antibody fragment, we had to be sure that it is unlikely to interfere with the complimentary determining regions (CDRs) of the antibody; therefore, we modelled the predicted structure. ABodyBuilder was used for structure determination, which is a structure determination software designed specifically for nanobodies (4). Structures are based on orientation prediction, CDR modelling, and side-chain prediction, and results are given a confidence score based on the root-mean square deviation threshold. From the ABodyBuilder model we used the generated PDB file to determine optimal linkage to a fluorescent protein (Fig. 2A). Hence, the optimal binding sites to connect the fluorescent protein to are the N-termini of the LC and HC. However, the N-terminus of the HC is occupied, as it links to the C-terminal of the LC, via a GGGGSGGGGSGGS linker. Therefore, the only optimal linkage point for the fluorescent protein is the N-terminal of the LC (Fig. 2B).

Figure 2. A) Structure of anti-THC ScFv, as predicted by ABodyBuilder. B) Structure of anti-THC ScFv linked to mNG at the N-terminal of the light chain.

Developing a Novel Membrane Assay

To test the antibodies' ability to bind THC we developed a novel one-step immunoassay (Protocol: Novel Immunoassay for Lipophilic Antigens). The protocol uses a lipophilic membrane to isolate THC, and relies on the antibody for the specificity (Fig. 3). To determine the validity of the assay, we first tested it with commercially available anti-THC antibodies, with a fluorescent secondary antibody (Fig. 4).

Figure 3. Explanation of the theory behind the novel immunoassay for THC detection. Note that this concept could be adapted to other lipophilic antigens.

Figure 4. THC membrane assay using commercial goat anti-THC and varying concentrations of THC. The signal was detected via an anti-goat fluorescent secondary antibody.

Customizable Biosensor

Additionally, the modular design of fluorescently linked antibody fragment allows for the creation of libraries of antibodies, specific to various antigen, which exert distinct fluorescent signals (Fig. 5).

Figure 5. THC membrane assay using various antibodies conjugated to fluorescent proteins with distinct fluorescent signals.

References

1. Niemi, M. H., Turunen, L., Pulli, T., Nevanen, T. K., Höyhtyä, M., Söderlund, H., Rouvinen, J., and Takkinen, K. (2010) A Structural Insight into the Molecular Recognition of a (−)-Δ9-Tetrahydrocannabinol and the Development of a Sensitive, One-Step, Homogeneous Immunocomplex-Based Assay for Its Detection. Journal of Molecular Biology. 400, 803–814
2. Brennan, J. (2005) The production of recombinant single chain antibody fragments for the detection of illicit drug residues. doctoral thesis, Dublin City University, [online] http://doras.dcu.ie/17319/ (Accessed March 12, 2019)
3. Markiv, A., Beatson, R., Burchell, J., Durvasula, R. V., and Kang, A. S. (2011) Expression of recombinant multi-coloured fluorescent antibodies in gor -/trxB- E. coli cytoplasm. BMC Biotechnol. 11, 117
4. ABodyBuilder: Automated antibody structure prediction with data–driven accuracy estimation: mAbs: Vol 8, No 7 [online] https://www.tandfonline.com/doi/full/10.1080/19420862.2016.1205773?scroll=top&needAccess=true (Accessed October 9, 2019)