2024 Nobel Laureate Professor David Baker (University of Washington), along with 20 of his coworkers, and Professor Ken Houk and his former UCLA graduate student, Dr. Cooper Jamieson (Ph.D. ’21, now at Gilead), have reported the first computational design of functional serine hydrolases that have folds different from natural serine hydrolases. Their work, recently published in Science, shows that it is now possible to design a new protein from scratch, rather than by modifying known proteins.
Houk and Professor Donald Hilvert (ETH Zürich) and coworkers had previously shown (J. Am. Chem. Soc., 130, 15222 (2008)) that the quantum mechanical computational of the 3D arrangement of catalytic groups necessary for the 9-step reaction of binding of ester, hydrolysis and release of acid and alcohol products predicted a geometry of catalytic groups that overlayed very closely with the arrangement that Nature had evolved for many hydrolase enzymes in three different fold families.
In the same year, the Houk and Baker groups published the successful computational design of enzymes for three unnatural reactions, a retro-aldolase, the Kemp elimination, and a Diels-Alder reaction, but those all involved redesigning the active sites of known enzymes. The new work is a major leap forward since the designs gave active enzymes based on entirely new folds not present in natural enzymes. This comes very close to the ultimate goal of designing enzymes for any desired reaction, for therapy or for synthesis, independent of what Nature has done before.
The work depended upon the Houk group quantum mechanical modeling, but to a very great degree on the new AI methods, RFdiffusion and PLACER, for protein design from Baker’s group at the University of Washington.
For more information, contact Professor Ken Houk (houk@chem.ucla.edu).
