Full text: Download
Abstract Nucleus-encoded circular RNAs (ncircRNAs) have been widely detected in eukaryotes, and most circRNA identification algorithms are designed to identify them. However, using these algorithms, few mitochondrion-encoded circRNAs (mcircRNAs) have been identified in plants, and the role of plant mcircRNAs has not yet been addressed. Here, we developed a circRNA identification algorithm, mitochondrion-encoded circRNA identifier, based on common features of plant mitochondrial genomes. We identified 7,524, 9,819, 1,699, 1,821, 1,809, and 5,133 mcircRNAs in maize (Zea mays), Arabidopsis (Arabidopsis thaliana), rice (Oryza sativa), tomato (Solanum lycopersicum), cucumber (Cucumis sativus), and grape (Vitis vinifera), respectively. These mcircRNAs were experimentally validated. Plant mcircRNAs had distinct characteristics from ncircRNAs, and they were more likely to be derived from RNA degradation but not intron backsplicing. Alternative circularization was prevalent in plant mitochondria, and most parental genomic regions hosted multiple mcircRNA isoforms, which have homogenous 5′ termini but heterogeneous 3′ ends. By analysis of mitopolysome and mitoribosome profiling data, 1,463 mcircRNAs bound to ribosomes were detected in maize and Arabidopsis. Further analysis of mass spectrometry-based proteomics data identified 358 mcircRNA-derived polypeptides. Overall, we developed a computational pipeline that efficiently identifies plant mcircRNAs, and we demonstrated mcircRNAs are widespread and translated in plants.