In silico design of novel thermostable and alkaline tolerant three-helical proteins for enhanced antibody purification in Industrial settings

Abstract

Monoclonal antibodies (mAbs) have revolutionised therapeutic strategies across oncology, autoimmune diseases, and infectious diseases. Despite their success, large-scale production remains limited by purification bottlenecks, with Protein A affinity chromatography serving as the industry gold standard due to its high specificity and robust performance. However, Protein A-based purifications present key limitations, including high resin costs, sensitivity to alkaline cleaning, ligand degradation, and inefficiencies in purifying complex antibody formats. In response, this thesis applies advanced protein engineering methodologies to improve the stability and resilience of Protein A ligands under industrial conditions. A consensus design approach was employed to develop structurally robust Protein A variants. This led to the creation of a highly thermo- and chemically stable protein, a potential affibody-like scaffold named Thermobody, with a defined three-dimensional structure similar to the immunoglobulin (Ig)-binding domain of Protein A. Further rational engineering of this scaffold enabled the development of novel IgG binders with improved binding kinetics and enhanced thermal and chemical tolerance, offering a promising alternative to current Protein A-based ligands. Additionally, a supercharging protocol was applied to the Thermobody to modify its surface charge, improve solubility, and reduce aggregation, thereby extending the potential applications of this affibody-like scaffold and its variants beyond immunoglobulin purification. A positively supercharged version with membrane permeability was further explored for intracellular delivery applications. These supercharged Thermobody constructs were evaluated for their ability to traverse cell membranes, providing insight into their potential as intracellular targeting agents. The findings presented in this thesis support the development of more cost-effective and scalable purification strategies, addressing the evolving demands of the biopharmaceutical industry while broadening the application scope of engineered affinity ligands. To further optimise ligand performance, machine learning-guided design (Protein MPNN) was employed to predict scaffold binding potential to the Fc region of IgG. The combined effects of these strategies were assessed through biophysical characterisation techniques, including surface plasmon resonance (SPR), circular dichroism (CD), and stability assays under repeated high-pH exposure. Show more