AAAS Research Volume 2018, Article ID 7089174, 15 pages https://doi.org/10.1155/2018/7089174 Research Article A Coding Sequence-Embedded Principle Governs Translational Reading Frame Fidelity ⋆ Ji Wan,1 Xiangwei Gao,1 Yuanhui Mao,1 Xingqian Zhang,1 and Shu-Bing Qian1,2, 1 Division of Nutritional Sciences, Cornell University, Ithaca, NY 14853, USA 2Graduate Programs in Genetics Genomics and Development, Biochemistry Molecular and Cellular Biology, CornellUniversity,Ithaca,NY14853,USA ⋆ Correspondence should be addressed to Shu-Bing Qian; [email protected] Received 28 May 2018; Accepted 28 August 2018; Published 20 September 2018 Copyright © 2018 Ji Wan et al. Exclusive Licensee Science and Technology Review Publishing House. Distributed under a Creative Commons Attribution License (CC BY 4.0). Upon initiation at a start codon, the ribosome must maintain the correct reading frame for hundreds of codons in order to produce functional proteins. While some sequence elements are able to trigger programmed ribosomal frameshifing (PRF), very little is known about how the ribosome normally prevents spontaneous frameshif errors that can have dire consequences if uncorrected. � Using high resolution ribosome profling data sets, we discovered that the translating ribosome uses the 3 end of 18S rRNA to scan the AUG-like codons afer the decoding process. Te postdecoding mRNA:rRNA interaction not only contributes to predominant translational pausing, but also provides a retrospective mechanism to safeguard the ribosome in the correct reading frame. Partially eliminating the AUG-like “sticky” codons in the reporter message leads to increased +1 frameshif errors. Remarkably, mutating the highly conserved CAU triplet of 18S rRNA globally changes the codon “stickiness”. Further supporting the role of “sticky” sequences in reading frame maintenance, the codon composition of open reading frames is highly optimized across eukaryotic genomes. Tese results suggest an important layer of information embedded within the protein-coding sequences that instructs the ribosome to ensure reading frame fdelity during translation. 1. Introduction Previous attempts to investigate the mechanisms under- lying reading frame maintenance have focused on the inter- Te importance of the translation machinery and its fdelity action between mRNA and tRNA at the decoding center have been apparent since the discovery of the genetic code [4]. Although codon:anticodon pairing is central to decoding [1]. During translation, the ribosome reads nucleotide triplets accuracy, there is an extensive network of interaction between in a successive, nonoverlapping fashion. For a given coding mRNA, rRNA, and ribosomal proteins [5]. Te stepwise, sequence, it is the start codon (most likely AUG) that codon by codon, progression of the ribosome along the establishes the reading frame. Since there is no punctuation � � between codons for successive amino acid residues, all three mRNA from the 5 to 3 direction is primarily driven by reading frames give rise to diferent codon sequences. How- translocation [6]. During this step, the ribosome undergoes ever, only one reading frame primarily serves as the coding massive conformational rearrangement, allowing movement template for functional proteins because a high frequency of of P- and A-site tRNAs into the E- and P-sites, respectively stop signals exists in alternative frames [2]. Tis feature repre- [7–10]. It is thought that the coordinated rotation of large and sents a strong evolutionary optimization of codon sequences small subunits ensures one codon precision in the movement that minimizes the consequence of frameshif errors [3]. of mRNA and the attached tRNAs afer peptidyl transfer Unlike missense errors, which are not necessarily destructive [11]. However, tremendous fexibility exists. In bacteria, while to proteins, frameshif errors are deleterious as premature ter- elongation factor G (EF-G) mediates one codon forward mination ofen leads to nonfunctional translational products. movement of mRNA:tRNA complexes, elongation factor 4 Terefore, the ribosome must keep the correct reading frame (EF4) catalyzes back-translocation [12]. Lack of EF4 has been for hundreds to thousands of codons during translation to shown to reduce translation fdelity under stress conditions ensure proper protein production. [13], suggesting that translocation is error-prone. 2 Research While ribosomes must maintain the correct reading insertion site, can the shifed ribosome reach frame 1 coding frameinordertoproducefunctionalproteins,somecis- region of Rluc (Figure 1(a), highlighted region). Compared to acting signals located in mRNAs are able to trigger pro- thein-framecontrol,weobservedareproducible+1FSrate grammed ribosomal frameshifing (PRF) [14]. Such recoding of 0.52% in HEK293 cells and 0.37% in MEF cells (Figure 1(a), events have been exploited by viruses to expand the genetic right panel). Afer adjustment by the size of the FS window, coding capacity within their limited genome size [15, 16]. the average +1 FS rate per codon in HEK293 and MEF cells is Te typical signals for –1 PRF include a slippery sequence, 0.04% and 0.03%, respectively. We next constructed a –1 FS a downstream secondary structure (usually a pseudoknot), reporter by inserting two additional bases (CA) between fag and a short spacer in between. With an increasing amount and Rluc. Once the ribosome skips two bases or slips back one of PRF signals identifed, it is clear that the ribosome bears base within the FS window, it will reach frame 2 Rluc coding tremendous fexibility in decoding mRNAs. Te rich com- region for translation. Compared to the +1 FS reporter, we munication between the ribosome and the message is thus observed a relatively lower –1 FS rate of 0.03% in HEK293 and equally, if not more, important as codon:anticodon pairing in 0.02% in MEF afer adjustment by the size of the FS window maintaining translation fdelity. However, our understanding (Figure 1(a)). Tis value is in the similar range of spontaneous of individual PRF cases provides limited insights on how the FS rates reported previously [23]. ribosome normally maintains the correct reading frame [17– Te seemingly low FS rates measured from the reporter 20]. Te fundamental question centers on how the ribosome assay could have striking efects on ribosome engagement if is able to achieve remarkable precision for the majority the FS rate remains constant over the entire coding region. of mRNA coding sequences while permitting of-track in Even with the lower bound FS rate (0.03% for +1 FS and certain regions. 0.02% for –1 FS), the theoretical exponential decay predicts Here we set out to investigate mechanistic features of cod- that nearly 20% of ribosomes will be derailed from the mRNA ing sequences in translational reading frame fdelity. Using before reaching the stop codon of Rluc (Figure 1(b)). To high resolution ribosome profling data sets obtained from experimentally determine the pattern of ribosome occupancy eukaryotic cells, we uncovered universal “sticky” sequences along the Rluc coding region, we conducted ribosome profl- within the coding region that retains the ribosome by pairing ing in HEK293 cells transfected with the control bicistronic with 18S rRNA. Analogous to the internal Shine-Dalgarno plasmid [24, 25]. In spite of the nonuniform read distribution, (SD) sequences in prokaryotic mRNAs [21], the eukaryotic the average trend of ribosome occupancy along the Rluc cod- “sticky” sequences contribute to pervasive translational paus- ing region does not parallel with the predicted exponential ing. Importantly, the interaction between the mRNA “sticky” decay (Figure 1(b), bottom panel). � sequence and the 3 end of 18S rRNA ofers a gripping We next assessed global ribosome drop-of across mechanism afer decoding to guide the ribosome in the endogenous transcripts by comparing the average ribosome correct reading frame during translation. In addition, we density for the frst and second half of each coding region found that mRNA coding sequences are subject to multilevel (CDS) in HEK293 cells (Figure 1(c)) [26]. Te strong cor- optimization during evolution to minimize frameshif errors. relation (R = 0.955) between the two regions indicates that Our work illustrates new layers of information embedded there is minimal drop-of in ribosome occupancy for the within protein-coding sequences that extends our under- vast majority of genes (Table S1). If spontaneous FS errors standing of the genetic code. occur in a continuous manner, the overall ribosome drop-of is expected to increase as function of CDS length. However, 2. Results the ratio of ribosome density between the frst and second half of CDS largely maintains irrespective of the CDS size 2.1. Measuring Spontaneous Frameshif Rates during Trans- (Figure 1(d)). Te same pattern also holds true in MEF cells lation. To understand how ribosomes normally maintain (Figure S1A and S1B). Interestingly, a few genes that display translational reading frame fdelity, it is necessary to deter- large drop-of are known examples of PRF. For instance, mine the spontaneous frameshif (FS) rate under physio- both OAZ1 and OAZ2 undergo controllable +1 PRF, allowing logical conditions. We adopted a luciferase-based bicistronic synthesis of ornithine decarboxylase (ODC) antizymes [27] reporter assay, in which both the Renilla luciferase (Rluc) (Figure 1(c) and S1C). and the Firefy luciferase (Fluc) are synthesized from the same mRNA (Figure 1(a)) [22]. A fag-tag was added at the