Template and target site recognition by human LINE-1 in retrotransposition.

The Long Interspersed Element-1 (L1) retrotransposon has generated nearly one-third of the human genome and serves as an active source of genetic diversity and human disease 1 . L1 spreads via a mechanism termed target-primed reverse transcription (TPRT), in which the encoded enzyme (ORF2p) nicks the target DNA to prime reverse transcription of its own or non-self RNAs 2 . Here, we purified the full-length L1 ORF2p and biochemically reconstituted robust TPRT with template RNA and target site DNA. We report cryo-electron microscopy structures of the human L1 ORF2p bound to structured template RNAs and initiating cDNA synthesis. The template polyadenosine tract is recognized in a sequence-specific manner by five distinct domains. Among them, a novel RNA-binding domain bends the template backbone to allow engagement of an RNA hairpin stem with the L1 ORF2p C-terminal segment. In addition, structure and biochemical reconstitutions demonstrate a surprising target-site requirement: L1 ORF2p relies on upstream single-stranded DNA to position adjacent duplex in the endonuclease active site for nicking of the longer DNA strand, with a single nick generating a staggered DNA break. Our work provides key insights into the mechanism of ongoing transposition in the human genome and informs the engineering of retrotransposon proteins for gene therapy.