Genome-wide de novo L1 Retrotransposition Connects Endonuclease Activity with Replication

Abstract

L1 retrotransposon-derived sequences comprise approximately 17% of the human genome. Darwinian selectivepressures alter L1genomicdistributionsduring evolution, confounding the ability to determine initial L1 integration preferences. Here, we generated high-confidence datasets of greater than 88,000 engineered L1 insertions in human cell lines that act as proxies for cells that accommodate retrotransposition in vivo. Comparing these insertions to a null model, in which L1 endonuclease activity is the sole determinant dictating L1 integration preferences, demonstrated that L1 insertions are not significantly enriched in genes, transcribed regions, or open chromatin. By comparison, we provide compelling evidence that the L1 endonuclease disproportionately cleaves predominant lagging strand DNA replication templates, while lagging strand 3’-hydroxyl groups may prime endonuclease-independent L1 retrotransposition in a Fanconi anemia cell line. Thus, acquisition of an endonucleasedomain, in conjunctionwith the ability to integrate into replicating DNA, allowed L1 to become an autonomous, interspersed retrotransposon.