Zinc finger nuclease (ZFN) is composed of an engineered site-specific Cys

2

-His

2

zinc finger domain and the nonspecific restriction enzyme Fok I cleavage domain

which is able to cut at a specified genomic locus to generate double-strand break (DSB) of DNA. The DSBs induced by ZFN are subsequently repaired through two different DNA repair mechanisms

either non-homologous endjoining (NHEJ) or homology-directed recombination (HDR). NHEJ is prone to introduce sequence insertions or deletions (indels)

and can therefore produce frameshifts in open reading frames and gene loss of function. HDR requires a donor template with the sequence similar to the genome to mend a lesion. By introducing a DNA donor with desired modifications

precise genomic modifications can be achieved at a frequency improved 10

2

-10

4

fold as compared to the traditional gene targeting method. Currently

most studies have focused on screening ZFN with higher activity

and improving the delivery efficiency of ZFN and donor into host cells

less studies have investigated the relationship between the homologous arm length with the efficiency of ZFN induced homologous recombination. Here

we constructed a pair of ZFN plasmids targeting to EGFP and verified its cutting activity. Then we designed a series of donors with different lengths of homologous arms. By introducing individual donor with ZFN into CHO cells harboring a frame-shift GFP gene

we measured the homologous recombination efficiencies through the flow cytometric analysis. We found that a 50 bp short homology arm was capable to support ZFN-mediated homologous recombination. Increasing the length of the homologous arms could improve the efficiency of ZFNmediated homologous recombination. A dramatic improvement (10

4

fold higher than traditional method) requires a homology arm longer than 1 000 bp.