Abstract
DNA transposons and retroviruses are versatile tools in functional genomics and gene therapy. To facilitate their application, we conducted a genome-wide insertion site profiling of the piggyBac (PB), Tol2 and Sleeping Beauty (SB) transposons and the murine leukemia virus (MLV) in mouse embryonic stem cells (ESCs). PB and MLV preferred highly expressed genes, whereas Tol2 and SB preferred weakly expressed genes. However, correlations with DNase I hypersensitive sites were different for all vectors, indicating that chromatin accessibility is not the sole determinant. Therefore, we analysed various chromatin states. PB and MLV highly correlated with Cohesin, Mediator and ESC-specific transcription factors. Notably, CTCF sites were correlated with PB but not with MLV, suggesting MLV prefers smaller promoter-enhancer loops, whereas PB insertion encompasses larger chromatin loops termed topologically associating domains. Tol2 also correlated with Cohesin and CTCF. However, correlations with ESC-specific transcription factors were weaker, suggesting that Tol2 prefers transcriptionally weak chromatin loops. Consistently, Tol2 insertions were associated with bivalent histone modifications characteristic of silent and inducible loci. SB showed minimum preference to all chromatin states, suggesting the least adverse effect on adjacent genes. These results will be useful for vector selection for various applications.