Expanding beyond the natural protein repertoire to engineer targeted vaccines and diagnostics

Proteins have evolved over millions of years to carry out the vast majority of biological functions that are fundamental to life. Their three-dimensional structures and functions have been in the focus of biomedical research for many decades, and substantial progress has been made in understanding pro-tein folding and structure-function relationships. The inverse problem of protein folding is called protein design ¿ the design of novel protein sequenc-es that fold into predefined three-dimensional structures. Towards this aim, computational tools have emerged as a powerful tool for the de novo design of proteins with structural and biophysical proper-ties that are not found in nature. To date, however, the vast majority of de novo designed proteins has been deprived of biological functions. Recently, the de novo design of proteins with customized mo-lecular and biological functions has gained momentum, aiming to exploit them to tackle outstanding biotechnological and biomedical challenges of the 21st century. One of these grand biomedical challenges that could be transformed by de novo protein design is the design of novel and more effective vaccines, especially for pathogens where traditional approaches for vaccine development have failed. Among these pathogens is the respiratory syncytial virus (RSV), which causes severe lower respiratory tract infections in young children and the elderly. Recently, numerous broadly protective, RSV neutralizing antibodies (nAbs) have been isolated from humans, and their structural characterization in complex with their target epitopes has greatly improved our mo-lecular understanding of an effective nAb response. A remaining challenge is the design of immuno-gens that effectively spotlight these antigenic sites, and elicit targeted nAb responses in vivo. My thesis work leverages de novo protein design for the design of epitope-focused immunogens that induce nAbs in vivo. Strikingly, we show how a cocktail of three de novo designed immunogens pre-senting selected neutralization epitopes elicit RSV nAbs in non-human primates. Furthermore, the designed immunogens bear unique potential as boosting immunogens in non-naïve subjects, allowing the focusing of nAbs onto defined antigenic sites. Together, these represent a substantial step forward for the use of immunogens based on computationally designed proteins, and offers a roadmap to em-ploy computational protein design in the engineering of precision immunogens for other pathogens. From a protein design perspective, my work introduces a `bottom-up¿ approach towards the de novo design of functional proteins. The bottom-up approach presents a general computational protocol to build de novo proteins with embedded binding motifs, including those that are structurally irregular or discontinuous, i.e. consist of multiple segments. Beyond immunogens, we exploit the designed proteins as biosensors to detect and quantify epitope-specific antibody responses, providing a pra