Improvements to the technique could pave the way for its use in patients with genetic diseases.
The CRISPR/Cas9 technique relies on a bacterial protein (Cas9) that is associated with a short piece of RNA. When Cas9 encounters a DNA sequence that matches up with that of its RNA guide, it cuts the DNA molecule at that location.
While the technique is very accurate, it is not infallible, and occasionally causes 'off-target effects' – cuts to the genome at sites where the DNA sequence is very close to the target sequence.
In the study, which was published in Nature, the scientists made changes to parts of the Cas9 which contact DNA molecules, creating a variant called SpCas9-HF1. While this version of the protein did not bind so effectively to DNA, it cut the genomes of lab-grown human cells in the correct places nearly as efficiently as the standard version of Cas9.
It also made far fewer 'off-target' mistakes: with no mistakes made with six out of seven test sequences, and only one made with the seventh. With the same guide sequences, the standard version of Cas9 made between 2 and 25 errors.
'We expected it would be improved in its specificity,' said Dr J. Keith Joung of Massachusetts General Hospital in Boston, lead author of the study, to GenomeWeb. 'What was striking was just how good it was. In many cases, all of the off-target effects caused by wild-type Cas9 drop to undetectable levels.'
Similar improvements to the accuracy of the CRISPR/Cas9 technique have recently been made by researchers at the Broad Institute (see BioNews 831). In that study, published in Science, changes were made to a different part of the Cas9 protein.
These improvements in accuracy are a step closer towards clinical applications of CRISPR. Unwanted damage to the genome could cause cell death or uncontrolled cell division and cancer, and this represents an important challenge.
Dr Joung is hopeful that the improved technique will be taken up straight away. 'We envision that our high-fidelity variant will supplant the use of standard Cas9 for many research and therapeutic applications,' he said.
However, the end point of these advances is not yet clear. Dr Charles Gersbach of Duke University in Durham, North Carolina, who was not involved in the research, told Nature News: 'At some point everyone needs to decide how specific is specific enough.'