Optimal guideRNAs for re-directing deaminase activity of hADAR1 and hADAR2 in trans

Adenosine deaminases that act on RNA (ADAR) are a class of enzymes that catalyze the conversion of adenosine to inosine in RNA. Since inosine is read as guanosine ADAR activity formally introduces A-to-G point mutations. Re-addressing ADAR activity toward new targets in an RNA-dependent manner is a highly rational, programmable approach for the manipulation of RNA and protein function. However, the strategy encounters limitations with respect to sequence and codon contexts. Selectivity is difficult to achieve in adenosine-rich sequences and some codons, like 5'-GAG, seem virtually inert. To overcome such restrictions, we systematically studied the possibilities of activating difficult codons by optimizing the guideRNA that is applied in trans. We find that all 5'-XAG codons with X = U, A, C, G are editable in vitro to a substantial amount of at least 50% once the guideRNA/mRNA duplex is optimized. Notably, some codons, including CAG and GAG, accept or even require the presence of 5'-mismatched neighboring base pairs. This was unexpected from the reported analysis of global editing preferences on large double-stranded RNA substrates. Furthermore, we report the usage of guanosine mismatching as a means to suppress unwanted off-site editing in proximity to targeted adenosine bases. Together, our findings are very important to achieve selective and efficient editing in difficult codon and sequence contexts.