PRe-ART (Predictive Reagent-Antibody Replacement Technology):Engineering and analysis of randomised DNA libraries encoding designed armadillo repeat proteins

Abstract

PRe-ART (Predictive Reagent- Antibody Replacement Technology) aims to replace reagent monoclonal antibodies with designed armadillo repeat proteins (dArmRPs), made from sequence-defined modular subunits capable of specific and conserved dipeptide recognition and binding. These modular units joined in a ‘Lego-brick’ fashion, will generate proteins capable of binding a user-defined target peptide, removing the costly and timely traditional immunisation process and the associated issue of unreproducible results. This project contributed to PRe-ART by generating randomised DNA libraries targeting key binding resides of the two pockets within the armadillo repeat, aiming to alter the unit’s specificity. Successful saturation of seven positions using MAX randomisation, produced a randomised DNA library of the dArmRP pocket that originally bound arginine. Computational designs provided by collaborators in the Höcker group (University of Bayreuth), directed the engineering of a specific DNA library, aiming to engineer an improved threonine-binder. Separately, to accommodate for saturating contiguous codons in the second, Lysine-binding pocket (not possible with MAX randomisation) a new saturation mutagenesis technology, ParaMAX randomisation, was invented. This MAX randomisation derivative was implemented on an adapted dArmRP sequence, generating a region of four contiguous randomised codons. Subsequently, novel Next Generation Sequencing (NGS) analysis techniques were developed to assess the success of positional saturation as a quality control stage before protein expression and screening by collaborators in the Plückthun group (University of Zurich). Analyses of the DNA libraries engineered using MAX randomisation showed successful target saturation and were therefore used in protein production and screening. The proof of concept ParaMAX library analysis revealed further optimisation of the ParaMAX process was required to prevent deletions interfering with amino acid representation. This analysis also revealed limitations in existing alignment technologies when processing such unique DNA libraries. Alternative stratagems for ParaMAX and the processing of NGS data are considered in light of these results.