Design

Overview

This section will outline how a recombinantly modular protein was developed and how it was implemented into a lateral assay. The modular protein consisted of an ScFv, linked to either a GFP or an alkaline phosphatase which were both bound by glycine linkers. Initial designs began with the sequencing of the antibody followed by computational analysis.

Antibody Design

To create the ScFv, IgG ScFv fragments were sourced for both the VH and VL chains from Ghosh and Huber which were taken from Genbank (EF028212 and EF028213) (1). The 3-24 variant was selected over the 8-10 variant as the former was proven to have stronger binding to OspA in the reviewed paper. An attempt was made to fold these chains together using SAbPred's ABodyBuilder a free online software that predicts the structural folding of antibodies (2). Structures were 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.

Before folding could occur, complementary determining region (CDR) anchor regions had to be determined. These regions were determined by conducting a BLAST analysis of sequence homologies of the VH and VL sequences. The BLAST indicated the top 100 similar sequences which were then aligned using Seaview to analyze sequences similarities (Figure 1) (3).

Figure 1 - Seaview sequence alignment of homologous sequences from BLAST search of generateed ScFv. Linker region between VH and VL shown as (GGGS)₃ linkers. Our engineered ScFv noted as AAS111111.1.

The CDR anchor regions were then determined based on the most conserved residues from the aligned sequences. With CDR anchor regions determined, folding was predicted through ABodyBuilder. The following sequences were submitted for folding prediction of the ScFv:

Heavy Chain (VH)

EVQLVQSGAEVKKPGASVKVSCKASGYTFTDYYLHWVRQAPGQGLEWLGRINPSSGATYSPQRFQGRVTMTTDTSISTAYMELSSLRSDDTAVYFCATLTTFNIWGFDYWGQGTLVSS

Light Chain (VL)

DIQMTQSPSSLSASVGDRVTITCRASQSISTYLNWYQQKPGKAPKLLIFTASSLQSGVPSTFSGSGSGTDFTLTISSLQPEDFATYYCQQSYSATFTFGGGTKVEIKR

The determined CDR anchor regions were bolded while the CDR regions were underlined for distinction.

The results from ABodyBuilder indicated VH and VL confidence scores of 0.89 and 0.99 respectively (Table 1).

Fv Region	Template PDB (chain)	Selection Method	Score
VH frameword	5nh3 (H)	Sequence Identity	0.89
CDR H1	5nh3 (H)	CDR Specific Fread	74
CDR H2	5nh3 (H)	CDR Specific Fread	39
CDR H3	1oar (J)	CDR Specific Fread	33
VL framework	5b71 (C)	Sequence Identity	0.99
CDR L1	5b71 (C)	CDR Specific Fread	47
CDR L2	4qth (L)	CDR Specific Fread	26
CDR L3	4llw (B)	CDR Specific Fread	27
VH-VL Orientation	5f7e (HL)	Chosen by ABangle	0.81

Table 1 - ABodyBuilder folding prediction confidence values. VH and VL predicted with confidence values of 0.89 and 0.99 respectively.

When comparing the predicted CDR regions from ABodyBuilder to that of Ghosh and Huber, all CDRs were identical except VH CDR3 (Table 2). In the predicted model, VH CDR3 was extended by an additional four residues however, as discussed later, these additional residues were shown to have little to no impact on OspA binding. From the ABodyBuilder model, the pdb file was downloaded for later analysis of binding to OspA to test the affinity of the generated ScFv.

>

Ghosh and Huber (1)	ABodyBuilder
VH CDR1: DYYLH	VH CDR1: DYYLH
VH CDR2: RINPSSGATYSPQRFQG	VH CDR2: RINPSSGATYSPQRFQG
VH CDR3: LTTFNIW	VH CDR3: LTTFNIWGFDY
VL CDR1: RASQSISTYLN	VL CDR1: RASQSISTYLN
VL CDR2: TASSLQS	VL CDR2: TASSLQS
VL CDR3: QQSYSATFTF	VL CDR3: QQSYSATFTF

Table 2 - Complimentary determining regions (CDRs) as predicted by Ghosh and Huber and ABodyBuilder. Note that the independently determiend CDRs are in complete agreement with each other, except for an additional four C-terminal amino acids in the ABodyBuilder VH CDR3.

Binding Modelling

To analyze and predict binding between OspA Genbank (M57248.1) and our ScFv, Cluspro protein-protein docking supercomputer (4–7) was used. Specifically, the antibody mode predicted the binding between the receptor (ScFv) and the ligand (OspA)(8). The Cluspro supercomputer is designed to test billions of docking confirmations between two proteins with lowest-energy structures undergoing root-mean square deviation and energy minimization (Table 3))(7).

Cluster	Members	Representative	Weighted Score
0	118	Center	-323.9
0	118	Lowest Energy	-330.9
1	81	Center	-306.9
1	81	Lowest Energy	-320.1
2	40	Center	-275.4
2	40	Lowest Energy	-306.0
3	40	Center	-273.2
3	40	Lowest Energy	-306.0
4	37	Center	-337.2
4	37	Lowest Energy	-337.2
5	28	Center	-268.3
5	28	Lowest Energy	-298.1

Table 3 - Showcases the top six lowest docking energy confirmations depicted by Cluspro supercomputer. The model that was selected for analysis was Cluster 0 as this model had the lowest energy conformation.

From the models generated, the lowest energy conformation binding model was selected and the CDR binding interactions within PyMol (Table 1) was analyzed to determine how effectively the ScFv binds to OspA. To effectively analyze binding, all residues within the CDR region that were shown to have polar interactions with OspA residues were highlighted as stick structures.

Figure 2A - Overview of OspA and ScFv docking. OspA and ScFv shown in cyan and green respectively with VH and VL CDRs highlighted in blue and magenta respectively.

Figure 2B - Enhanced view of interaction between VH and OspA. Polar interactions between VH and OspA shown with yellow dashes.

Figure 2C - Enhanced view of interaction between VL and OspA. Polar interactions between VL and OspA shown with yellow dashes.

Fusion Proteins

Once we were confident that the ScFv would theoretically bind to OspA, Chimera in combination with PyMol was used to build and model the resulting fusion proteins (9). The first of these fusion proteins was the combination of the 3-24 ScFv bound via C-terminal through a (GGGS)3 linker region to a GFP (Figure 3).

Figure 3 - 3D rendering of 3-24 ScFv + GFP modular protein binding to OspA ligand.

References

1. Ghosh, S., and Huber, B. T. (2007) Clonal diversification in OspA-specific antibodies from peripheral circulation of a chronic Lyme arthritis patient. J. Immunol. Methods. 321, 121–134

2. Dunbar, J., Krawczyk, K., Leem, J., Marks, C., Nowak, J., Regep, C., Georges, G., Kelm, S., Popovic, B., and Deane, C. M. (2016) SAbPred: a structure-based antibody prediction server. Nucleic Acids Res. 44, W474–W478

3. Gouy, M., Guindon, S., and Gascuel, O. (2010) Sea view version 4: A multiplatform graphical user interface for sequence alignment and phylogenetic tree building. Mol. Biol. Evol. 27, 221–224

4. Desta, I. T., Porter, K. A., Xia, B., Kozakov, D., and Vajda, S. (2020) Performance and Its Limits in Rigid Body Protein-Protein Docking. Structure. 28, 1071-1081.e3

5. Vajda, S., Yueh, C., Beglov, D., Bohnuud, T., Mottarella, S. E., Xia, B., Hall, D. R., and Kozakov, D. (2017) New additions to the ClusPro server motivated by CAPRI. Proteins Struct. Funct. Bioinforma. 85, 435–444

6. Kozakov, D., Hall, D. R., Xia, B., Porter, K. A., Padhorny, D., Yueh, C., Beglov, D., and Vajda, S. (2017) The ClusPro web server for protein-protein docking. Nat. Protoc. 12, 255–278

7. Kozakov, D., Beglov, D., Bohnuud, T., Mottarella, S. E., Xia, B., Hall, D. R., and Vajda, S. (2013) How good is automated protein docking? Proteins Struct. Funct. Bioinforma. 81, 2159–2166

8. Brenke, R., Hall, D. R., Chuang, G.-Y., Comeau, S. R., Bohnuud, T., Beglov, D., Schueler-Furman, O., Vajda, S., and Kozakov, D. (2012) Structural bioinformatics Application of asymmetric statistical potentials to antibody-protein docking. 28, 2608–2614

9. Pettersen, E. F., Goddard, T. D., Huang, C. C., Couch, G. S., Greenblatt, D. M., Meng, E. C., and Ferrin, T. E. (2004) UCSF Chimera - A visualization system for exploratory research and analysis. J. Comput. Chem. 25, 1605–1612

Overview Antibody Design Binding Modelling Fusion Proteins References Go to Top

Team:Queens Canada/Design