Global landscape of RNA editing in hematopoiesis

Adenosine-to-inosine (A-to-I) RNA editing plays a vital role in many biological processes. Our previous study has demonstrated that the hematopoietic stem and progenitor cells (HSPCs) deficient in RNA editing cannot reconstitute the irradiated recipients in vivo. However, the overall profile of RNA editome in hematopoiesis has not been established and the underlying mechanism is poorly understood. Here, we sorted 12 murine adult hematopoietic cell populations, and performed deep RNA and DNA sequencing. We depicted the first landscape of RNA editing in hematopoietic cells and identified 30,796 editing sites in total. Although total editing frequency was stable across the 12 hematopoietic populations, we identified stage-specific and group-specific editing genes, and uncovered multiple famous hematopoietic regulators were dynamically edited. By charactering group specific editing genes, we found that the Azin1 (antizyme inhibitor 1) was highly edited in HSPCs, which was required for HSPCs differentiation. A-to-I editing of Azin1 transcript resulted in serine-to-glycine substitution at residue 367 of AZIN1, induced its translocation from cytoplasm to nucleus, enabled its higher binding affinity with DDX1 (DEAD box polypeptide 1) in nucleus further to alter the chromatin distribution of DDX1, resulting in multiple hematopoietic regulators expression change. Collectively, this study provided the first landscape of RNA editing in hematopoietic cascade and uncovered the mechanism of A-to-I editing of Azin1 in regulating HSPCs function. (1)LSK+ HSPC DNA-seq were generated by deep sequencing, in duplicate. (2) mRNA profiles of LT-HSC / ST-HSC / MPP / CMP / MEP / GMP / MK / Ery-A / Ery-B / monocyte / macrophage / granulocyte were generated by deep sequencing, in duplicate. (3) mRNA profiles of overexpression of Azin1-TC/Azin1-G in c-Kit+ cell were generated by deep sequencing, in duplicate. (4)mRNA profiles of Ddx1 knock down / normal control in c-Kit+ cell were generated by deep sequencing, in duplicate. (5)Ddx1 ChIP-seq with input in 32D cells were generated by deep sequencing, in duplicate.