Peptide-binding antibodies are widely used for protein purification and detection, perform scaffolding functions, and can serve as versatile crystallization chaperones. Engineering antibody specificity requires randomization of the complementarity determining regions (CDRs), six loops that extend from the protein framework to form the antigen binding site. However, current methods are inefficient compared to the potential of in silico protein engineering. OptCDR is a computational method that, using rules derived from 883 antibody structures, designs CDR sequences de novoto bind a given antigen. Here, we perform for the first time experimental validation of the OptCDR modeling method by engineering a FLAG peptide-binding antibody fragment.90 single chain antibody fragments (scFvs) predicted by OptCDR to bind the FLAG peptide were displayed on the surface of M13 phage for a functional ELISA test. The fraction with specificity for the FLAG peptide ranged from 8-73%, depending on the scFv framework used. Functional protein was expressed in E. coli after conversion to the Fab format (addition of the constant light and heavy chains) to improve stability and solubility. The dissociation constant (KD) was estimated to be >1 �M, so affinity maturation was performed after PCR-based randomization of the variable heavy chain. Selection using a combination of phage and E. colidisplay is expected to yield high-affinity variants.This is the first time to our knowledge that de novo CDR modeling from canonical structure rules has been used for the basis of antibody engineering. The best initial designs possessed lower affinity, but comparison to failed designs and affinity-matured variants will yield insights to improve future predictions. These results suggest that de novo CDR modeling is a viable strategy for antibody engineering, which could eliminate costly and time-consuming steps required by conventional methods.

<i>De novo</i> antibody design by complementarity determining region modeling

247th National Meeting (2014)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.

De novo antibody design by complementarity determining region modeling