Speaker
Description
De novo protein enables the creation of novel structures of stable protein structures that exceed those found in nature. However, designing antibody-like scaffolds that combine stability with functional versatility remains a major challenge, constrained by the conserved β-sandwich architecture of immunoglobulin (Ig) domains and designing functional loop regions.
Our group has developed hyperstable, non-natural Ig-domain dimers, validated by X-ray crystallography, that act as robust frameworks for functional loop insertion. Building on these validated scaffolds, we are exploring the possibility of engineering two-domain Ig scaffolds to contain functional cavities at the interface between domains with antigen recognition loops. We hypothesise that the controlled incorporation of small-molecule binding pockets and antigen-binding loops will generate stable and controllable frameworks for performing programmable binding functions.
To address this issue, we are implementing a multi-step computational pipeline for designing a library of Ig-domain pairs with engineered inter-domain pockets and exploring the de novo antigen-binding loop design in those validated two-domain Ig frameworks. AI-based and physic-based tools are used for backbone - sequence design, structure prediction, modelling and interface optimization, followed by in silico evaluation. The most promising candidates will be recombinantly expressed in E. coli and experimentally characterized using SEC-MALS, circular dichroism, SPR/BLI, and X-ray crystallography to validate their structure and function. This will establish a versatile platform for future immunoglobulin-like scaffolds customized to bind protein targets of medical interest.