The ability to make controlled changes in the genome is a very powerful tool. Several technologies exist for this purpose, but most have drawbacks. For instance, transgenes introduced into mammalian cells with viruses are subject to position effects. Homologous recombination does not work well in some systems, such as human cells.

ZFNs target specific sites in the genome.

Zinc-finger nucleases (ZFNs) can be used to create targeted double-stranded breaks in the genome and have thus generated much excitement. These enzymes consist of a DNA-binding portion, which can in principle be designed to target a particular location in the genome, and a Fok1 nuclease domain, which introduces the break. A donor DNA can be exogenously supplied to effect repair of the break by a homology-driven process, resulting in gene replacement with a desired sequence. Alternatively, mutations are introduced during an error-prone repair process.

ZFNs have by now been used to target the genome in several different systems: in mouse and human cells, in zebrafish, in plants, in fruit flies and in the rat. ZFNs can be used to create deletions of up to 15 megabase pairs in human cells and have been used recently to disrupt or tag genes in human pluripotent stem cells, a system of tremendous interest for genome engineering. No doubt the application of these tools will continue.

One of the ongoing difficulties is that it is not trivial to design ZFNs. Current methods are either relatively inefficient at generating enzymes that function well, or they are laborious and require dedicated screening systems. Many users depend on commercial services, at relatively high cost. Methods developers, however, continue to both improve approaches to design ZFNs (for instance, Nat. Methods 8, 67–69; 2011) and to engineer enzymes with better properties (for instance, Nat. Methods 8, 74–79; 2011). An efficient and user-friendly method to generate ZFNs that perform well is likely to spread rapidly among researchers in many fields.