In the post-genomic era, highly specific nucleases became important tools for directed gene targeting and gene therapy. They are used to specifically cleave genomic sequences, involved in for instance diseases and by creating this double-strand break, homologous recombination, an otherwise rarely occurring event in the cell, will be stimulated. Most commonly used are zinc-finger nucleases, consisting of specific zinc-finger modules responsible for recognizing DNA, fused to the catalytic domain of the restriction enzyme FokI. Some of these zinc-finger nucleases are already in clinical trial phases. As alternative in the past three years so-called TALE-nuclease came into the focus of research. Their advantage might be the high predictability of specificity and their possibly lower cytotoxic effects due to off-site target-cleavage, which have to be confirmed in the future.In this work, two other approaches for generating highly specific nucleases were presented. The first one is based on directed evolution, where the specificity of a naturally occurring meganuclease was to be changed. The investigated enzyme belonging to the LAGLIDADG family of homing endonucleases was PI-SceI. Mutant libraries of this enzyme were assayed in a two-plasmid selection system, selecting for variants cleaving a target site differing from the natural PI-SceI target in up to seven bp. Since structural and biochemical data for PI-SceI are available, only residues known to be involved in DNA-binding were used for mutagenesis. The obtained variants, showing in vivo an up to 100 - fold increase in specificity did not show the same result in vitro, under all conditions tested. This showed that results obtained in vivo and in vitro do not necessarily correspond to each other and that the developed selection assay needs further optimization.As second approach, a more rational strategy similar to zinc-finger or TALE nucleases was pursued, by fusion of a catalytically inactive variant of the LAGLIDADG homing endonuclease I-SceI to the type IIP restriction enzyme PvuII. With this, an alternative to the often used catalytic domain of FokI as cleavage module was presented. After optimizing the linker between the two proteins and several mutations in I-SceI to abolish any catalytic activity (D44N, D145A) and mutations in PvuII to either decrease its activity (T46G, H83A or Y94F) or disturb the dimerization interface (L12E, P14G or H15D) a final variant, namely P(P14G)-L(6)-Ss*, was obtained. This enzyme showed in vitro an over 1000 - fold preference for the addressed (PvuII site flanked by two I-SceI sites) over an unaddressed site (PvuII site alone). Furthermore, this protein had an up to 10 - fold preference for the distance of 6 bp between the respective I-SceI and PvuII sites and showed no cleavage of genomic DNA (bacteriophage &
#955;-DNA). The next step would be in vivo testing of this enzyme variant, ideally in planta, since there are already I-SceI target sites in several crop species available that could serve as landing platforms for gene targeting.
Verknüpfung zu Publikationen oder weiteren Datensätzen
