Mapping in vivo protein-RNA interactions at single-nucleotide resolution from HITS-CLIP data

Mammalian RNA complexity is regulated through interactions of RNA-binding proteins (RBPs) with their target transcripts. High-throughput sequencing together with UV-crosslinking and immunoprecipitation (HITS-CLIP) is able to globally map RBP-binding footprint regions at a resolution of ~30-60 nucleotides. Here we describe a systematic way to analyze HITS-CLIP data to identify exact crosslink sites, and thereby determine protein-RNA interactions at single-nucleotide resolution. We found that reverse transcriptase used in CLIP frequently skips the crosslinked amino-acid-RNA adduct, resulting in a nucleotide deletion. Genome-wide analysis of these crosslinking-induced mutation sites (CIMS) in HITS-CLIP data for Nova and Argonaute (Ago) proteins in mouse brain tissue revealed deletions in ~8-20% of mRNA tags, which mapped to Nova and Ago binding sites on mRNA or miRNA. CIMS analysis provides a general and more precise means of mapping protein-RNA interactions than currently available methods and insight into the biochemical properties of such interactions in living tissues.