A novel subclass of exon-derived microRNA is widespread in the mammalian genome

Abstract: Argonaute-2, the primary effector molecule of the RNA Induced Silencing Complex, loads both canonical miRNAs and a wide variety of other non-canonical miRNA-like species that enable base pair mediated transcript regulation. Here, using ultra-deep RNA immunoprecipitation and sequencing of nuclear and cytoplasmic hAGO2 in human breast cancer-derived MCF-7 cells, we have identified over one thousand novel AGO2-associated small RNAs. These include a wide variety of isomiRs, in excess of 100 mirtron loci and H/ACA snoRNA and HY3 repeat-derived small RNAs. Unexpectedly, we also discovered 298 AGO2-loaded exon-derived miRNAs (emiRs). These conserved small RNAs are encoded within protein-coding exons and 3’UTRs, have a 5’ end adenosine bias (in contrast to the 5’ U observed in canonical miRNAs), are DICER-1 dependent, processed from mature mRNAs post-splicing and derived from a wide variety of genes involved in differentiation and development including SOX4, JUN, c-MYC and AGO2 itself. We find evidence of emiR expression in human brain and other peripheral tissues, and demonstrate that human emiRs are conserved and detected in murine Ago2 small RNA libraries. Five human emiR loci coincide with known Mendelian disease-causing mutations, and the expression of emiRs is altered in human breast cancer, which may indicate that these species can play a role in disease etiology and progression. We propose that emiRs are the small RNA complement to the competing endogenous RNA (ceRNA) system. Small RNAs and AGO2-bound small RNAs and in whole cell, nuclear and cytoplasmic compartments of MCF-7 cells.