Proteogenomics Reveals Widespread Non-canonical Translational Events in Plants

Abstract

<p>Non-canonical translational events play a crucial role in gene expression regulation, massively expanding proteome diversity. In mammals and microorganisms, various non-canonical translational events, including non-AUG translation initiation and stop codon readthrough (SCR) events, have been extensively studied and demonstrated substantial biological significance. However, the exploration of these events across different plant species remains limited, hindering a comprehensive understanding of non-canonical translational events in plants. In addition, the recording potential of stop codons in SCR events is largely unexplored in Plantae. This thesis employed proteogenomic strategies with customized databases to identify genome-wide nonAUG translation initiation events with N-terminal extensions (NTEs) and SCR events in monocots and dicots, characterizing their features and exploring the biological implications. The key results and findings are as follows: </p><p>1. Using a proteogenomic strategy with 80 customized databases, 1,009 SCR events were identified in two monocots (maize and rice) and two dicots (soybean and <i>Arabidopsis</i>). Plant SCR candidates tended to have shorter transcript lengths, and fewer exons and splice variants than non-SCR transcripts, contrary to the previous reports in animals. Mass spectrometry evidence revealed that stop codons involved in plant SCR events were recoded as 20 standard amino acids, some of which were also supported by suppressor transfer RNA analysis. In addition, multiple functional signals were observed in 34 maize extended proteins, with the structural and subcellular localization changes characterized in the extended protein of basic transcription factor 3. Furthermore, the SCR events in plant species exhibited nonconserved signatures, and the extensions likely underwent protein-coding selection.</p><p>2. A total of 879 transcripts from 508 genes were detected with non-AUG translation initiation events with NTEs in monocots (maize and rice) and dicot soybean using a computational prediction-combined proteogenomic strategy. Over 60% of the non-AUG initiated candidates in these plants had multiple putative upstream translation initiation sites (uTISs), with GUG being the most common uTIS codon, followed by CUG in monocots and AUU in the dicot. Non-AUG initiated candidates had significantly longer 5' untranslated regions (UTRs) compared to the background in all three plant species. Sequence feature analysis showed that putative uTISs in non-AUG translation events tended to have favorable Kozak contexts and high AG content, similar to those of the annotated AUGs in these species. Furthermore, the NTEs of maize non-AUG candidates were predicted to contain diverse N-terminal functional signals, suggesting their novel functionalities. </p><p>Overall, this study reveals the widespread presence of non-canonical translational events in plant species, which contributes to shaping the dynamic composition of plant proteomes, and highlights the unprecedented recoding plasticity of stop codons. It provides important new insights into the complexity and flexibility of decoding eukaryotic genomes.</p>