An enzyme cuts DNA (artist’s impression). Synthetic protein-cleaving enzymes were designed quickly with the help of artificial intelligence.Credit: Artur Plawgo/Science Photo Library
Scientists have harnessed artificial-intelligence models to create synthetic CRISPR proteins that edit the genome more efficiently than their naturally occurring counterparts. Such synthetic CRISPR systems could one day power discoveries in fields from medicine to agriculture.
The results1 were published on 16 July in Science.
“Much like CRISPR democratized the ability to edit DNA at will, AI-based protein design promises to allow anyone to create totally novel properties in the protein space,” says Soeren Lienkamp, a molecular biologist at the University of Zurich in Switzerland who was not involved in the research. He adds that the paper “marries two transformative fields”: AI-guided design and enzymes called RNA-guided nucleases, which can cut DNA and RNA strands.
Snip, snip
These nucleases form the backbone of the gene-editing system known as CRISPR, which uses a ‘guide RNA’ to direct the nuclease to a target DNA sequence. The nuclease then acts like molecular scissors and snips out the targeted material, enabling scientists to edit, delete or add genetic information. CRISPR systems are based on the machinery that bacteria use to defend themselves against viruses. The most common CRISPR nucleases, such as Cas9 and Cas12, are co-opted from bacteria.
As powerful a tool as gene editing is, the process is complex. Nucleases must complete a carefully orchestrated series of steps, making it difficult to move beyond what evolution has already produced, says Jennifer Doudna, a biochemist at the University of California, Berkeley, and lead author of the paper, who shared the 2020 Nobel Prize in Chemistry for her work on CRISPR systems. “Once you start tweaking things, you realize pretty quickly that while you can make changes, they ultimately produce something that isn’t functional.”
Ex-Meta scientists debut gigantic AI protein design model
AI tools offer the opportunity to supercharge the process of identifying promising candidates for new, functional nucleases. Rather than performing hundreds or even thousands of exploratory experiments, researchers could, in theory, ask machine learning to do it for them. To test this idea, Doudna and her collaborators focused on creating synthetic versions of a group of tiny nucleases called TnpBs, which are evolutionary precursors to the commonly used Cas12. The scientists wanted to know how far they could alter the proteins’ sequences while retaining the proteins’ ability to edit genes.
For a protein to function, it often needs to assume a certain shape, or conformation. The team began by providing an AI model with the final conformation of a type of TnpB and asking it to reverse-engineer changes to the underlying DNA templates that would nevertheless maintain the protein’s final shape. This approach produced thousands of potential changes, but said nothing about whether the resulting protein would be active.
