Mutational, structural and evolutionary analyses of the human antibody VL domain

Abstract

Human antibodies represent about half of all drugs entering clinical trials. Target interaction is mediated by two key immunoglobulin variable domains (VH and VL) of the antibody molecule. Unfortunately, these domains often display limited stability and a tendency to aggregate. While many different strategies have been proposed to stabilise human VH, methods for the generation of aggregation-resistant human VL domains had so far remained elusive. This thesis outlines a generally applicable strategy for the generation of human VL domains with improved biophysical properties (aggregation resistant, highly expressed, non-‘sticky’). The approach is based on the discovery of a set of ‘hot-spot’ residues in the complementarity determining region of human VL (CDR2: 49, 50-53, 56) that control aggregation propensity. Moreover, I present mutational and X-ray crystallographic data that demonstrate compatibility with antigen binding. This allowed the retrofitting of domains derived from human antibody therapeutics (such as Herceptin) and the generation of intrinsically stable antibody libraries (Garvan-2 phage display libraries). These findings provide a pathway towards the generation of stable and aggregation resistant human antibody therapeutics for biotechnology applications.

The availability of stable human variable domains also opens up novel approaches in basic science. For instance, it has long been proposed that modern day antigen receptors (B- and T-cell receptor) have evolved from a simple primordial ancestor through a process of gene duplication and diversification. However, due to the lack of homologs in phylogenetically ancient species, detailed molecular insights had not been possible. Here, I use laboratory evolution and X-ray crystallography to reconstruct such a primordial receptor. This work underpins an evolutionary pathway for the emergence of immunoglobulin-based adaptive immunity.