Computational design of an epitope-specific Keap1 binding antibody using hotspot residues grafting and CDR loop swapping

Abstract

Therapeutic and diagnostic applications of monoclonal antibodies often require careful selection of binders that recognize specific epitopes on the target molecule to exert a desired modulation of biological function. Here we present a proof-of-concept application for the rational design of an epitope-specific antibody binding with the target protein Keap1, by grafting pre-defined structural interaction patterns from the native binding partner protein, Nrf2, onto geometrically matched positions of a set of antibody scaffolds. The designed antibodies bind to Keap1 and block the Keap1-Nrf2 interaction in an epitope-specific way. One resulting antibody is further optimised to achieve low-nanomolar binding affinity by in silico redesign of the CDRH3 sequences. An X-ray co-crystal structure of one resulting design reveals that the actual binding orientation and interface with Keap1 is very close to the design model, despite an unexpected CDRH3 tilt and V_H/V_L interface deviation, which indicates that the modelling precision may be improved by taking into account simultaneous CDR loops conformation and V_H/V_L orientation optimisation upon antibody sequence change. Our study confirms that, given a pre-existing crystal structure of the target protein-protein interaction, hotspots grafting with CDR loop swapping is an attractive route to the rational design of an antibody targeting a pre-selected epitope.

Figure 1. Design and characterisation of anti-Keap1 antibodies targeting Nrf2 binding site by hotspots graft.

(a) Workflow scheme of antibody design using hotspots-guided scaffold graft approach. Three hotspots residues from Nrf2 are identified as the anchors, onto which the antibody Fv scaffold structures from PDB are aligned. The interfacial residues in CDR loops are mutated to reduce the clashes and improve the computed binding energies with Keap1. The designed models with different scaffolds hosting the hotspots are ranked by binding energies and the top rank ones are selected for synthesis and binding kinetics test using SPR. (b) SPR kinetics profiles for G54.1/Keap1 and G85/Keap1 complexes with designed antibody Fabs immobilized on the chips, which shows that these two example grafts bind to Keap1 and titrate well, with affinities of 126 nM and 236 nM respectively. (c) Competitive SPR kinetics profiles for G54.1/Keap1 and G85/Keap1 complexes by titration with cognate high affinity Nrf2 peptide segment, which shows that the two example grafts cross-block Nrf2 in a concentration-dependent manner as expected, indicating that the epitopes of the antibodies overlap the Nrf2 binding site on Keap1. The amount of G85 Fab was reduced to observe competition.

Figure 2. Design and characterisation of affinity improvement of G54.1 antibody by CDRH3 loop swap.

(a) Workflow scheme of rational design of affinity improved G54.1 using CDRH3 loop swap approach. The original CDRH3 loop in the modelled Keap1-G54.1 structure was replaced by other CDRH3 fragments collected from PDB. The residues of CDRH3 in resulted new complex structures are subjected to computational design to improve the computed binding energies with Keap1. The top rank designs are selected for synthesis and binding kinetics test using SPR. (b) Sequence alignment of CDRH3 loop in the CDRH3 swap variants of G54.1 tested in this work, which shows low sequence identities between the CDRH3 loops of parental G54.1 and designed CDRH3 swap variants. (c) Experimentally determined binding affinity improvements of designed CDRH3-swap variants over parental G54.1 Fab. LS171, LS145, LS168 and LS146 show most prominent affinity improvement over G54.1, with the affinities of 4.1, 5.4, 9.5 and 19.6 nM respectively. (d) Computationally modelled CDRH3 conformations and interaction modes of G54.1, LS171 and LS168 with Keap1, highlighting the different loop conformations and interactions of the CDRH3 loops in the affinity-improved antibodies. The key contact residues in CDRH3 loops are depicted as sticks.

Figure 3. Crystal structure of LS146-scFv/Keap1 complex.

(a) LS146-scFv /Keap1 crystal complex structure. b-e, Close-ups of interactions between Keap1 and individual loops in LS146-scFv: CDRH2 (b), CDRH3 (c), CDRH1 (d), and V_H framework 3 (FR3, e), respectively, with the key interactional residues depicted as sticks, and hydrogen bonds depicted as red dot lines. It shows that the interactions between antibody and Keap1 are dominantly mediated by heavy chain, and CDRH2 as the hotspots acceptor makes the most intense interactions.

Figure 4. Comparison of LS146-scFv/Keap1 crystal complex structure with design model shows the precision of the computational design.

(a) Comparison of binding modes of crystal LS146-scFv with modelled LS146-Fab by superimposing onto the Keap1 side, showing the antibody bind in the manner as modelled, though an obvious disengagement of V_L chain from the binding interface with Keap1 is observed. The coloured dots represent the centres of mass (COM) for crystal (green) and modelled (red) Fv structures; the blue vectors represent the axis around which the crystal structure rotates to superpose on the design model. The rigid body deviation between crystal and modelled Fv structures are characterised by the distances between the COMs and rotational angles. (b) Crystal packing between CDRH3, L2 loops and blade 4 of Keap1 from the neighbouring crystal unit, which leads to rigid distortion of both heavy and light chains. The key interactional residues of Keap1 from neighbouring crystal unit and the ones on the antibody side that undergo apparent conformational change from prediction are depicted as sticks. The respective rigid body deviations between crystal and modelled V_H and V_L domain are characterised by the distances between the COMs and rotational angles. (c) Individual rigid-body deviation of the CDRH3 loop (left) and V_L chain (right) by superposing the V_H chains of crystal structure on that of design model. The coloured dots represent the COMs of CDRH3 (left) and V_L (right); the blue vectors represent the axis around which the crystal structure rotates to superpose on the design model. The rigid body deviation between crystal and modelled structures are characterised by the distances between the COMs and rotational angles. (d) Close-up comparison of residues packing at V_H/V_L interface from crystal and modelled structures when they are superposed on V_H chain. The key packing residues that undergo apparent conformational change from prediction are depicted as sticks. (e–g) Close-up comparison of backbone conformations and sidechain orientations of interfacial CDR loops from crystal and modelled structures: CDRH2 (e), CDRH3 (f), and CDRH1 and V_H framework 3 (g). The key contacting residues are depicted as sticks.