Integration of exogenous DNA into mouse embryonic stem cell chromosomes shows preference into genes and frequent modification at junctions.

Chromosomal integration of exogenous DNA in mammalian cells allows stable gene expression for a variety of biological applications. Although it is presumably mediated by DNA repair machinery, little is known regarding site preferences and other characteristics. We isolated and analyzed 256 chromosomal-plasmid DNA integration junctions from 158 plasmid integrants after electroporation in mouse embryonic stem (ES) cells. The frequency of integrations in transcription units (40%) showed a slight but significant increase over the frequency estimated by computer simulation of random events (30%), suggesting preferential integration into genes. Microarray analysis revealed preference into genes, which are expressed in mouse ES cells. In contrast, bias toward integrations around transcriptional start sites, CpG islands and repeat elements was not observed. Furthermore, all host chromosome sequences as well as the majority of plasmids (96%) at the integration junctions were modified by deletions and/or insertions of additional nucleotides. Detailed analyses revealed frequent stem loop/hairpin formation mediated by weak homologies near plasmid ends before integration. Our study sheds light on a natural fate of exogenous DNA, which preferentially integrates into transcriptionally active chromosomal sites and by an imprecise end-joining pathway, associated with highly frequent modification of the end sequences.