Improved super-resolution ribosome profiling reveals prevalent translation of upstream ORFs and small ORFs in Arabidopsis

Authors: HSIN-YEN, LARRY WU - Michigan State University; AI, QIAOYUN - Michigan State University; TEIXEIRA, RITA - Michigan State University; NGUYEN, PHONG - Michigan State University; SONG, GAOYUAN - Iowa State University; MONTES, CHRISTIAN - Iowa State University; Elmore, James - Mitch; WALLEY, JUSTIN - Iowa State University; YINGSHAN HSU, POLLY - Michigan State University

Submitted to: The Plant Cell
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 10/17/2023
Publication Date: 11/24/2023
Citation: Hsin-Yen, L., Ai, Q., Teixeira, R., Nguyen, P., Song, G., Montes, C., Elmore, J.M., Walley, J.W., Yingshan Hsu, P. 2023. Improved super-resolution ribosome profiling reveals prevalent translation of upstream ORFs and small ORFs in Arabidopsis. The Plant Cell. 36(3):510-539. https://doi.org/10.1093/plcell/koad290.
DOI: https://doi.org/10.1093/plcell/koad290

Interpretive Summary: Accurately defining gene models and determining translated open reading frames (ORFs) from genomic sequence data is fundamental for studying gene functions and monitoring cellular activity in all living organisms. Despite extensive efforts, our understanding of the translational landscape remains incomplete. It has become increasingly clear that a substantial fraction of the transcriptomes in diverse plant species has prevalent unannotated translated ORFs. These unannotated translated ORFs include small ORFs (sORFs) encoded by presumed non-coding RNAs, as well as short upstream ORFs (uORFs) present in the 5' untranslated regions (5' UTRs) of protein-coding RNAs (i.e., mRNAs). Despite their importance, short ORFs are commonly excluded by computational genome annotations. Thus, comprehensive identification of these hidden short ORFs is the first step toward understanding and characterizing their functions. We present an improved Ribo-seq dataset with enhanced ribosome-protected mRNA fragment (RPF) coverage in Arabidopsis seedlings. Our new data significantly improves the identification of translated uORFs and sORFs, and provides evidence for supporting noncanonical translation associated with various gene regulation events. Additionally, we developed a new computational method to address the issues of out-of-frame mapping at translation termination and overlapping between uORFs for uncovering relatively short uORFs. Combining our new data and new computational approach has identified over 7,000 unannotated translation events. Thus, our enhanced Arabidopsis translational landscape facilitates the discovery of translated ORFs and offers valuable resources for studying gene functions.

Technical Abstract: A crucial step in functional genomics is identifying actively translated Open Reading Frames (ORFs) that link biological functions. The challenge lies in identifying short ORFs, as they are greatly impacted by data quality and depth. Here, we improved the coverage of super-resolution Ribo-seq in Arabidopsis, revealing uncharacterized translation events in nucleus-, chloroplast-, and mitochondria-encoded genes. We identified 7,751 unconventional translation events, including 6,996 upstream ORFs (uORFs) and 209 downstream ORFs on annotated protein-coding genes, as well as 546 ncORFs on presumed non-coding RNAs. Proteomics data confirmed the production of stable proteins from some of the unannotated translation events. We presented evidence of active translation on primary transcripts of tasiRNAs (TAS1-4) and microRNAs (pri-miR163, pri-miR169), and periodic ribosome stalling supporting co-translational decay. Additionally, we developed a method for identifying extremely short uORFs, including 370 minimum uORF (AUG-stop), and 2,984 tiny uORFs (2-10 aa), as well as 681 uORFs that overlap with each other. Remarkably, these short uORFs exhibit strong translational repression as longer uORFs. We also systematically discovered 594 uORFs regulated by alternative splicing, suggesting widespread isoform-specific translational control. Finally, these prevalent uORFs are associated with numerous important pathways. In summary, our improved Arabidopsis translational landscape provides valuable resources to study gene expression regulation.