Target specificity in the ADAR editing landscape explains the unique immunoregulatory function of the interferon-inducible isoform ADAR1-p150

Abstract

The adenosine deaminase acting on RNA (ADAR) family includes three enzymes, ADAR1-p150, ADAR1-p110, and ADAR2 that bind endogenous and pathogen-derived double-stranded RNA (dsRNA) and convert adenosine (A) to inosine (I). This process called A-to-I editing disrupts dsRNA structures and counteracts activation of innate immune sensors MDA-5 and PKR, preventing autoimmune disorders like Aicardi-Goutieres-Syndrome. We and others have shown that the interferon-inducible ADAR1-p150, but not the other ADARs, protects cells and organisms from aberrant innate immune activation by self-derived dsRNA, even though A-to-I editing is performed by all three ADARs. We hypothesized that ADAR1-p150 edits specific transcripts that cannot be edited by the other enzymes. Here we present a novel approach to quantify RNA editing in transcriptomic data. Our tool TSniffer detected more than 40,000 regions of A-to-I editing distributed over the entire human transcriptome, affecting 30% of protein-coding transcripts and 10% of long non-coding RNAs. Differential editing frequencies in cell lines with knock-outs for the different ADARs revealed transcripts exclusively edited by ADAR1-p150. In contrast, editing by ADAR1-p110 and ADAR2 was cooperative and more redundant. Our analyses fill gaps of knowledge about the A-to-I editing landscape, highlight the dominant immunoregulatory function of ADAR1-p150, and provide new insights in the role of aberrant A-to-I editing in disease development.