Abstract
RNA base editing has the capability to rewrite genetic codes while offering improved safety due to its reversibility and temporal-spatial tunability. The development of cytosine-to-uridine editors faces issues of DNA mutagenic activity of APOBECs and the large-size anchoring domains associated with immunogenicity due to their microbial origin. To address these issues, we took an approach by returning to the original formula of adenosine deaminase acting on RNA (ADAR), i.e., to convert ADAR2 to a cytosine deaminase acting on RNA by introducing 17 mutant substitutions derived from RESCUE-S-it is named ADAR2-mimic base editor for C-to-U RNA editing (AMBER). AMBER displays similar sequence contexts preference to that of RESCUE, and exhibits robust performance across various tested cell lines. By applying AMBER to correct several pathogenic transcript mutations, we achieved substantial editing with efficiency ranging from 8 to 38%. For example, for the Pah mutant (c.788T>C) that caused phenylketonuria in mice, AMBER had an editing efficiency of 19.7% in HEK293T cells. Additionally, concurrent bystander editing of target transcripts was mitigated by two alternate strategies. To demonstrate its effectiveness in vivo, we applied AMBER in C57BL/6 mice, and observed gRNA-dependent editing with ~21% on-target efficiencies at 72 h post injection, which remained detectable up to 1 wk. RNA-sequencing analysis of the mouse liver transcriptome revealed minimal off-target effects and no significant changes to endogenous RNA editing events. These results highlight AMBER as a promising therapeutic tool for repairing T-to-C mutations, offering great potential for the treatment of genetic diseases.