Transcriptome-wide investigation of stop codon readthrough in Saccharomyces cerevisiae

Translation of mRNA into a polypeptide is terminated when the release factor eRF1 recognizes a UAA, UAG, or UGA stop codon in the ribosomal A site and stimulates nascent peptide release. However, stop codon readthrough can occur when a near-cognate tRNA outcompetes eRF1 in decoding the stop codon, resulting in the continuation of the elongation phase of protein synthesis. At the end of a conventional mRNA coding region readthrough allows translation into the mRNA 3’-UTR. Previous studies with reporter systems have shown that the efficiency of termination or readthrough is modulated by cis-acting elements other than stop codon identity, including two nucleotides 5’ of the stop codon, six nucleotides 3’ of the stop codon in the ribosomal mRNA channel, and stem-loop structures in the mRNA 3’-UTR. It is unknown whether these elements are important at a genome-wide level and whether other mRNA features proximal to the stop codon significantly affect termination and readthrough efficiencies in vivo. Accordingly, we carried out ribosome profiling analyses of yeast cells expressing wild-type or temperature-sensitive eRF1 and developed bioinformatics strategies to calculate readthrough efficiency, and to identify mRNA and peptide features which influence that efficiency. We found that the stop codon (nt +1 to +3), the nucleotide after it (nt +4), the codon in the P site (nt -3 to -1), and 3’-UTR length are the most influential features in the control of readthrough efficiency, while nts +5 to +9 and mRNA secondary structure in the 3’-UTR had milder effects. Additionally, we found low readthrough genes to have shorter 3’-UTRs compared to high readthrough genes in cells with thermally inactivated eRF1, while this trend was reversed in wild-type cells. Together, our results demonstrated the general roles of known regulatory elements in genome-wide regulation and identified several new mRNA or peptide features important for translation termination and readthrough.当释放因子eRF1（eRF1）识别核糖体A位点（ribosomal A site）中的UAA、UAG或UGA终止密码子（stop codon）并刺激新生肽链释放时，mRNA翻译为多肽的过程即宣告终止。然而，当近同源tRNA（near-cognate tRNA）在解码终止密码子时竞争胜出eRF1，便会发生终止密码子通读（stop codon readthrough），进而使蛋白质合成的延伸阶段得以继续。在常规mRNA编码区末端，通读可使翻译延伸至mRNA 3'-非翻译区（3'-UTR）。既往基于报告系统（reporter systems）的研究表明，翻译终止效率或通读效率可受终止密码子身份以外的顺式作用元件（cis-acting elements）调控，包括终止密码子5'端的两个核苷酸、位于核糖体mRNA通道内的终止密码子3'端的六个核苷酸，以及mRNA 3'-UTR中的茎环结构（stem-loop structures）。目前尚不清楚这些元件在全基因组（genome-wide）层面是否具有重要作用，以及终止密码子近端的其他mRNA特征是否会在体内（in vivo）显著影响翻译终止与通读效率。据此，我们对表达野生型或温度敏感型eRF1的酵母细胞开展了核糖体谱分析（ribosome profiling），并开发了生物信息学（bioinformatics）策略以计算通读效率，同时鉴定影响该效率的mRNA及肽段特征。我们发现，对通读效率调控影响最为显著的特征包括：终止密码子（核苷酸+1至+3）、其下游首个核苷酸（nt +4）、P位点（ribosomal P site）密码子（核苷酸-3至-1）以及3'-UTR长度；而核苷酸+5至+9区域与3'-UTR内的mRNA二级结构则仅产生较弱影响。此外，我们还观察到，在eRF1热失活的细胞中，低通读效率基因的3'-UTR相较于高通读效率基因更短；而这一趋势在野生型细胞中恰好相反。综上，我们的研究证实了已知调控元件在全基因组翻译终止与通读调控中的普遍作用，并鉴定出若干对翻译终止及通读具有重要意义的新型mRNA或肽段特征。