CRISPR needs to anchor itself near a short sequence called PAM to do its job. In the book “Modern Prometheus” (Cambridge University Press) James Kozubek says a PAM is like a shoehorn, where the Cas9 nuclease begins to clasp down to recognize the right site and cut. In order to fit every gene, a super-adjustable shoehorn would be needed. Think of it as the equivalent of a bump key that can open any door. A Broad Institute group led by David Liu has almost reached the goal with xCas9, the new super-adjustable Cas9 variant described in Nature this week.

CRISPR was a bacterial immune system before it became a biotech tool and needed a targeting code to distinguish self from non-self DNA, i.e., to attack the viral genome while preserving the bacterial genome. PAM sequences are the code (the acronym stands for “Protospacer adjacent motif”). If they are not there, the molecular scissors cannot cut. Therefore reasearchers can design their CRISPR tool to edit virtually any gene in any species with one caveat: a PAM must be present at the target site. The most commonly used nuclease, Cas9 from Streptococcus pyogenes, needs a three-letters PAM sequence in which any nucleotide is followed by two guanine bases (NGG) and this greatly limits the fraction of genomic DNA that can be targeted with CRISPR. Therefore several research groups are looking for natural or engineered nucleases with different PAM requirements. The best result obtained so far is xCas9: it works fine in human cells and recognizes a wide range of PAMs including NG, GAA and GAT. The new enzyme, evolved using an approach called PACE, is not only more PAM-compatible than the classic Cas9, but also more specific. The good news, in short, is that we don’t have to choose between efficiency, compatibility and specificity, we can probably have them all.