Natureより
In late 2018, chemical biologist David Liu of the Broad Institute of MIT and Harvard in Cambridge, Massachusetts, received an e-mail from across the country: in Seattle, a team led by microbiologist Joseph Mougous at the University of Washington had discovered a strange enzyme. It was a toxin made by the bacterium Burkholderia cenocepacia — and when it encountered the DNA base C, it converted it to a U. Because U, which is not commonly found in DNA, behaves like a T, the enzymes that replicate the cell’s DNA copy it as a T, effectively converting a C in the genome sequence to a T.
Liu had harnessed similar enzymes in base editing, which allows researchers to use components of CRISPR–Cas9 to change one DNA base to another. But those enzymes, called cytidine deaminases, normally act only on single-stranded DNA. DNA in human cells consists of two strands wound together and, in the past, Liu had to rely on the Cas9 enzyme to break the DNA and create a region of unwound, single-stranded DNA for his enzymes to act on. Because of its reliance on the strand of RNA that guides Cas9, this technique wouldn’t be able to reach the mitochondrial genome.
But the enzyme that Mougous’s team had found, called DddA, could act directly on double-stranded DNA without relying on the Cas9 enzyme to break it. This, Liu and Mougous reasoned, could make DddA suitable for reaching the mitochondrial genome.
But to turn DddA into a genome-editing tool, Liu first needed to “tame the beast” — the ability to modify double-stranded DNA also makes the enzyme deadly because, if set loose, it would mutate every C it came across. To prevent this, the team split the enzyme into two pieces that would change DNA only when brought together in the right orientation. And to control which DNA sequence the enzyme modified, the team then linked each half of DddA to proteins that were engineered to bind to specific sites in the genome.
