The genome-wide sequence preference of ionising radiation-induced cleavage in human DNA.

For ionising radiation (IR)-induced cellular toxicity, DNA cleavage is thought to be a crucial step. In this paper, the genome-wide DNA sequence preference of gamma radiation-induced cleavage was investigated in purified human DNA. We utilised Illumina short read technology and over 80 million double-strand breaks (DSBs) were analysed in this study. The frequency of occurrence of individual nucleotides at the 50,000 most frequently cleaved sites was calculated and C nucleotides were found to be most prevalent at the cleavage site, followed by G and T, with A being the least prevalent. 5'-CC and 5'-CC dinucleotides (where * is the cleavage site) were found to be the present at the highest frequency at the cleavage site; while it was 5'-CCC for trinucleotides and 5'-GCCC and 5'-CCCC for tetranucleotides. The frequency of occurrence of individual nucleotides at the most frequently cleaved sites was determined and the nucleotides in the sequence 5'-GGCMH (where M is A or C, H is any nucleotide except G) were found to occur most frequently for DNA that was treated with endonuclease IV (to remove blocking 3'-phosphoglycolate termini); and 5'-GSC*MH (where S is G or C) for non-endonuclease IV-treated DNA. It was concluded that GC-rich sequences were preferentially targeted for cleavage by gamma irradiation. This was the first occasion that an extensive examination of the genome-wide DNA sequence preference of IR-induced DSBs has been performed.