Particle.news
Download on the App Store

Scientists Unveil AI-Driven Workflow for Designing Bespoke Enzymes

A collaborative effort by UCSB, UCSF, and the University of Pittsburgh has produced highly efficient, stable, and selective enzymes, published in Science, with implications for green chemistry and industrial innovation.

Overview

  • Researchers have developed a novel de novo enzyme design workflow integrating AI, X-ray crystallography, and chemical intuition for precise protein engineering.
  • This approach enables the creation of enzymes that catalyze challenging reactions, including carbon-carbon and carbon-silicon bond formations, which are inefficient with natural enzymes.
  • The designed enzymes exhibit exceptional stability in thermal and organic solvent conditions, expanding their utility in diverse industrial and pharmaceutical applications.
  • A second round of optimization using a loop-search algorithm and expert refinement significantly improved enzyme activity and stereoselectivity.
  • Future research aims to create simpler, smaller enzymes and explore catalytic mechanisms beyond those found in natural systems, further advancing sustainable biocatalysis.