Designing monoclonal antibody fragment-based affinity resins with high binding capacity by thiol-directed immobilisation and optimisation of pore/ligand size ratio

Abstract

Monoclonal antibody (mAb) based affinity resins usually suffer from low binding capacity, most probably as a result of steric hindrance by the large 150kDa size of the mAb and a random immobilisation approach. The present work investigates the influence of a variety of factors on dynamic binding capacity (DBC) such as pore/ligand size ratio, accessibility of ligand and ligand density. The effect of pore/ligand size ratio was investigated using Fab and scFv fragments on various resins with different pore sizes. The accessibility of the ligand was investigated by a site-directed immobilisation approach, where three C-terminal tags, PPKPPK, FLAG™ and Cys, were introduced into the Fab fragments for immobilisation on resins via amino-, carboxyl- and thiol-groups, respectively. The scFv fragments were tagged at the C-terminal only with FLAG™ to enable a straight forward purification procedure, and were immobilised to resins via amino- and carboxyl-groups. The target protein had a molecular weight (MW) of 50kDa. A 3-fold higher dynamic binding capacity at 100% breakthrough (DBC100%) was observed for Fab wild-type (wt) on CNBr-activated Sepharose 4 FF relative to mAb on same resin at the same ligand density. However, no major difference in DBC100% was observed between Fab wt and scFv immobilised on CNBr-activated Sepharose 4 FF at the same ligand density. Thus, further increase of pore/ligand size ratio from Fab to scFv on a resin with average pore size of 300Å, did not seem to be beneficial. Among the tested tags, only the C-terminal Cys tag proved to site-direct the ligands during immobilisation as it allowed the DBC100% to increase 1.6-fold as compared to Fab wt immobilised via amino-groups on CNBr-activated Sepharose 4 FF and Actigel ALD Superflow at the same ligand density. The influence of ligand density was investigated by selecting immobilised Fab Cys on Sulfhydryl-reactive resin. Increasing ligand density from 0.103 to 0.202μmol/mL resulted in the same utilisation yield (82–85%), whereas a further increase in ligand density from 0.202 to 0.328μmol/mL resulted in a 20%-unit decrease in utilisation yield and less steep breakthrough curve, suggesting steric hindrance in the pores of the resin. In addition, site-directed affinity ligands resulted in a more pronounced, sigmoid-shaped breakthrough curve, leading to more efficient use of capacity. The highest DBC100% was obtained for Fab Cys on Sulfhydryl-reactive resin and scFv on Actigel ALD Superflow; 11mg/mL and 10mg/mL, respectively, as compared to the DBC100% of 0.8mg/mL for mAb on CNBr-activated Sepharose 4 FF. Pore/ligand size ratio of 3, which was achieved for Fab ligands on the studied resins, was shown to be feasible for capturing a protein in MW of 50kDa. Totally, a 13.8-fold improvement in DBC100% was achieved with the Fab-based affinity resin coupled via the C-terminal Cys as compared to the mAb-based affinity resin.