AI Brings Synthetic Enzyme Design to Natural-Level Efficiency

Scientists at the Weizmann Institute of Science in Israel have successfully designed highly efficient synthetic enzymes using a computer algorithm, virtually eliminating the need for labor-intensive laboratory fine-tuning. These artificial enzymes can catalyze chemical reactions that natural proteins cannot, and they operate with efficiencies comparable to those of natural enzymes. The research recently published in Nature marks a significant milestone in ushering in a new era of rapid, customized enzyme design.

Traditionally, enzymes created through computational methods have suffered from low catalytic efficiency, requiring extensive experimental optimization—a time-consuming process. To overcome this limitation, the Weizmann research team adopted a novel approach, using the non-natural reaction "Kemp elimination"—which involves removing a proton from a specific carbon atom in the substrate—as a test case. By analyzing natural enzyme data, the team computationally deconstructed and recombined protein sequences, then applied atomic-level behavior models to identify optimal structures.

Contrary to conventional understanding that enzyme active sites require cyclic amino acids, the algorithm surprisingly predicted that non-cyclic amino acids could yield higher catalytic efficiency. This adjustment led to a significant enhancement in performance. The final enzyme design differed from its closest natural counterpart by more than 140 amino acids and demonstrated catalytic efficiency up to 100 times greater than previously designed AI enzymes.

While the researchers acknowledge that natural enzymes can perform complex, multi-step reactions through dynamic structural shifts—capabilities still lacking in current synthetic enzymes—the potential applications are vast. The team is now exploring whether this algorithmic system can improve the performance of Rubisco, the key enzyme involved in photosynthesis.