Proteins are chains of amino acids held together by peptide bonds, which are remarkably stable on their own, with estimates suggesting a half-life of several hundred years. In nature, however, most proteins are broken down much faster by specialized enzymes called proteases, which have evolved to catalyze the cleavage of peptide bonds and have diversified into countless variants.
Now, researchers have used AI-based protein design tools to take the catalytic core of two types of protease and replace the entire remainder of the protein. This effectively short-circuits the natural evolutionary process, allowing the creation of synthetic enzymes that can be tuned to cut specific target proteins.
The resulting enzymes were demonstrated to cleave proteins associated with neurological diseases, suggesting potential applications in research and, eventually, medicine. The work highlights how AI can accelerate protein engineering beyond the capabilities of traditional evolutionary approaches.