Carbene Footprinting Directs Design of Genetically Encoded Proximity-Reactive Protein Binders

Abstract

Genetically encoding proximal-reactive unnatural amino acids (PrUaas), such as fluorosulfate-L-tyrosine (FSY), into natural proteins of interest (POI) confer the POI with the ability to covalently bind to its interacting proteins (IPs). The PrUaaincorporated POIs hold promise for blocking undesirable POI−IP interactions. Selecting appropriate PrUaa anchor sites is crucial, but it remains challenging with the current methodology, which heavily relies on crystallography to identify the proximal residues between the POIs and the IPs for the PrUaa anchorage. To address the challenge, here, we propose a footprinting-directed genetically encoded covalent binder (footprinting-GECB) approach. This approach employs carbene footprinting, a structural mass spectrometry (MS) technique that quantifies the extent of labeling of the POI following the addition of its IP, and thus identifies the responsive residues. By genetically encoding PrUaa into these responsive sites, POI variants with covalent bonding ability to its IP can be produced without the need for crystallography. Using the POI−IP model, KRAS/RAF1, we showed that engineering FSY at the footprint-assigned KRAS residue resulted in a KRAS variant that can bind irreversibly to RAF1. Additionally, we inserted FSY at the responsive residue in RAF1 upon footprinting the oncogenic KRAS G12D / RAF1, which lacks crystal structure, and generated a covalent binder to KRAS G12D . Together, we demonstrated that by adopting carbene footprinting to direct PrUaa anchorage, we can greatly expand the opportunities for designing covalent protein binders for PPIs without relying on crystallography. This holds promise for creating effective PPI inhibitors and supports both fundamental research and biotherapeutics development.