Generation of Novel Nucleases with Extended Specificity by Rational and Combinatorial Strategies

After the discovery of DNA as the molecule carrying the hereditary information of all living organisms, DNA processing enzymes, in particular restriction endonucleases, led to the development of recombinant DNA technology. Today, in the postgenomic era, a steadily growing number of entire genome sequences is available. The emerging genome-engineering technology requires more sophisticated nucleases with very high specificity, that ideally target a unique sequence in a complex genome. The design and control of these rare-cutting endonucleases by using rational and evolutionary strategies is the focus of this Minireview. Designed rare-cutting endonucleases recognize specific target sequences that are >14 bp long and are the prime tools for precise gene targeting and therapy. Their high specificity allows the introduction of a single sequence-specific double-strand break (DSB) in a large genome. The DSB is repaired mainly by either homologous recombination (HR) or error-prone nonhomologous end joining (NHEJ), depending on cell type and cell cycle. Providing a homologous repair template during DSB repair by HR allows the correction of a mutated gene in its natural context without effecting the genome elsewhere (gene correction or gene surgery; Figure 1 A) or targeted gene addition in an appropriate genome locus, a socalled “safe harbor” (gene addition; Figure 1 B). The latter approach relies on the complementation of the defective gene by the new transgene. Without the donor template the DSB is repaired preferentially by error-prone NHEJ and small deletions or insertions are introduced at the targeted locus, thereby inactivating an endogenous gene or an integrated viral gene (gene disruption; Figure 1 C). These possible therapeutic applications explain the great efforts of various scientific groups and companies to generate tailor-made highly specific endonucleases. Today our toolbox for genome engineering contains various nuclease architectures to facilitate targeted double-strand breaks: heterodimeric zinc finger nucleases (ZFNs), engineered homing endonucleases (meganucleases), TAL effector nucleases (TALENs), conjugates of chemical nucleases or restriction enzymes with triple-helix forming oligonucleotides (TFOs), and other fusions of a highly specific DNA binding domain or protein with a monomeric or dimeric nuclease.