Establishing Site-Directed A-to-I RNA Editing in Cell Culture

Abstract

Tools to manipulate genetic information without interfering at the DNA level are highly desirable in medicine and the life sciences. Recently, our group introduced the first engineered, RNA-guided deaminase. The approach relies on the in situ covalent bond formation between a benzylguanine-modified guide RNA (BG-gRNA) and a SNAP-tagged deaminase (SNAP-ADAR). Once the gRNA-deaminase conjugate is formed, it enables specific adenosine-to-inosine (A-to-I) substitutions in target RNAs. Since inosine is biochemically interpreted as guanosine by the cellular machinery, site-directed A-to-I editing provides the possibility to manipulate RNA and protein function. In this PhD project, it was aimed at elucidating the potential of the SNAP-ADAR approach for future applications. Therefore, the performance of the editing system in mammalian cells was comprehensively characterized. It could be shown that efficient site-directed RNA editing with SNAP-ADAR enzymes in cell culture requires the chemical modification of the BG-gRNA. A strong performance in the editing of endogenous transcripts was demonstrated in engineered cell lines stably expressing SNAP-ADAR enzymes. Editing yields up to 90% were achieved and remained stable even when several transcripts or multiple sites on a single transcript were concurrently targeted. Maximum editing was reached after 3 hours of BG-gRNA transfection and stayed unchanged for several days. Additionally, low concentrations (≥ 1.25 pmol/96-well) of the BG-gRNA were sufficient to obtain highest editing levels. The SNAP-ADAR approach holds great promise for the recoding of many functionally important amino acid residues as 11 out of the 16 adenosine-containing 5’-NAN triplets were editable between 50% and 90%. First evidence was provided that the editing system might be a valuable tool for the correction of disease-causing mutations. Moreover, the possibility of manipulating entire signaling networks was highlighted by the efficient and concurrent editing of two disease-relevant transcripts, KRAS and STAT1. Photo-controlled A-to-I editing was applied to direct protein localization within the cell by introducing alternative start or stop codons which allowed the expression of signals for nuclear and membrane translocation. NGS-based analysis revealed that wild-type SNAP-ADAR enzymes are highly precise editing machines. Their more active versions (SNAP-ADARQ enzymes) produced some off-target edits into the transcriptome, but the observed off-target activity appeared to be reducible by lowering the SNAP-ADAR protein amounts without great inhibition of the on-target editing. Nevertheless, these enzymes were one order more precise than editing machines applied by competing approaches. The chemical modification of the BG-gRNA was shown to suppress the off-target editing within a duplex formed by the BG-gRNA and the target RNA. The SNAP-ADAR approach outcompetes all well-characterized approaches for site-directed RNA editing due the best balance between efficiency and specificity.