Proteases are enzymes that cut other proteins, and designing new ones has long been one of the hardest problems in protein design. Researchers at UW Medicine's Institute for Protein Design report two successes in the journal Nature.
In the first study, led by Hojae Choi, deep-learning methods were used to build cysteine proteases around an activated cysteine "blade." About 19 percent of the designs worked, cutting their target sequence up to 30 million times faster than the uncatalyzed reaction, and with folds unlike any natural protease. In the second, led by Anqi Chen and Kejia Wu, metalloproteases that use a bound metal ion as the blade showed a 36 percent success rate. They cut amyloid-beta, a protein linked to Alzheimer's disease, 100 million times faster than the uncatalyzed reaction.
Because one enzyme can cut many targets, designed proteases might one day work at much lower doses than drugs that simply bind their targets.
Source: UW Medicine Newsroom, October 14, 2026 — https://newsroom.uw.edu/news-releases/molecular-scissors-open-path-to-new-therapies/
