Comprehensive Prediction of Target RNA Editing Sites for PLS-Class PPR Proteins in Arabidopsis thaliana.

Abstract

Hundreds of RNA editing events, that is conversion of cytidines (Cs) to uridines (Us), have been observed in the mitochondrial and plastid transcriptome in vascular plants. Defects of C-to-U RNA editing affect a wide variety of physiological processes. These editing sites are recognized by pentatricopeptide repeat (PPR) superfamily proteins. PPR proteins are sequence-specific RNA binding proteins that participate in multiple aspects of organellar RNA metabolism. They are categorized into P and PLS subclasses, where PLS-class proteins are largely identified as RNA editing PPRs. Elucidating the principle involved in PPR-RNA recognition, the so-called PPR code, has enhanced our understanding of the recognition of RNA editing sites, thereby enabling prediction of target RNA editing sites for uncharacterized PLS-class proteins. Computational PPR-RNA prediction in RNA editing can be applied to the study of PPR-deficient plants that are genetically isolated from physiological abnormalities. However, the use of PPR-RNA prediction in RNA editing is still restricted due to ambiguous procedures and prediction reliability. Here, we refined the PPR code dataset, and the reliability of the computational prediction was quantitatively evaluated using known RNA editing PPRs. With this knowledge, a computational analysis was conducted in the 'PPR-to-editing site' and 'editing site-to-PPR' directions, against 199 PLS-class proteins and 499 organelle RNA editing sites in Arabidopsis thaliana. We propose 52 plausible PPR-RNA pairs for uncharacterized proteins and editing sites. The presented data will facilitate the study of organellar RNA editing involved in diverse physiological processes in A. thaliana.