Μake us preferred on Google Artificial intelligence is moving gene editing beyond the limits of natural evolution.
A team led by Nobel laureate Jennifer Doudna has developed an AI-based method for designing new DNA-editing enzymes from scratch, creating proteins with sequences that have never been observed in nature. The advance could expand the toolkit available for gene-editing research and, eventually, the development of new therapies.
Doudna and Emmanuelle Charpentier received the 2020 Nobel Prize in Chemistry for developing CRISPR-Cas9, the technology that transformed genetic engineering by allowing scientists to make targeted changes to DNA.
The new research, reported in Science, focuses on TnpB, an unusually small enzyme that could be particularly attractive for therapeutic applications because compact proteins are easier to deliver into cells.
Rather than making small, incremental changes to the naturally occurring enzyme, as scientists traditionally do, the researchers used AI to design entirely new versions. The model was trained to preserve parts of the protein that evolution has shown to be essential to its function while allowing other regions to be radically redesigned.
The result was dozens of synthetic enzymes, dubbed SynTnpBs, whose amino acid sequences do not exist in nature. Researchers produced the proteins in the laboratory and tested them in bacteria, plant cells and human cells.
Many of the synthetic enzymes proved highly active, while some outperformed the natural TnpB enzyme.
The findings suggest that AI could provide a new route to engineering biological molecules, allowing scientists to explore proteins beyond those produced by billions of years of evolution.
If the approach can be extended to other enzymes, it could open a much larger design space for gene-editing tools and accelerate the search for new biomedical applications.
