Deep sequencing reveals global patterns of mRNA recruitment during translation initiation

www.nature.com/scientificreports OPEN received: 19 February 2016 accepted: 28 June 2016 Published: 27 July 2016 Deep sequencing reveals global patterns of mRNA recruitment during translation initiation Rong Gao1,*, Kai Yu1,*, Jukui Nie1, Tengfei Lian1, Jianshi Jin1, Anders Liljas2 & Xiao-Dong Su1 In this work, we developed a method to systematically study the sequence preference of mRNAs during translation initiation. Traditionally, the dynamic process of translation initiation has been studied at the single molecule level with limited sequencing possibility. Using deep sequencing techniques, we identified the sequence preference at different stages of the initiation complexes. Our results provide a comprehensive and dynamic view of the initiation elements in the translation initiation region (TIR), including the S1 binding sequence, the Shine-Dalgarno (SD)/anti-SD interaction and the second codon, at the equilibrium of different initiation complexes. Moreover, our experiments reveal the conformational changes and regional dynamics throughout the dynamic process of mRNA recruitment. Translational control is an important type of posttranscriptional regulation in determining weather, when, and how much protein will be synthesized with a given mRNA1. Regulation at the translational level accounts for large variations in expression among different genes2. Among the many aspects that can regulate translation efficiency2–5, translation initiation is usually considered a key determinant of the translational yield for most mRNAs6–9. The translation initiation efficiency of a given mRNA is determined by its translation initiation region (TIR)1. TIRs have varied sequences with some preferred bases, but the sequences are non-unique9,10. A combination of multiple elements in this region is usually thought to contribute to the mRNA recruitment11, which includes the initiation codon, the Shine-Dalgarno (SD) sequence, the availability of the SD sequence for binding to the anti-SD (ASD) sequence near the 3′end of the 16S rRNA8, the distance between the SD sequence and the initiation codon, and the specific enhancer sequences (A/U-rich elements) recognized by protein S1 of the 30S ribosomal subunit7,11. The efficiency of mRNA recruitment is the cooperative and cumulative result of the multiple elements at the translation initiation region9,10. Translational initiation is a dynamic process and is often referred to as the rate-limiting step of translation3,12,13. In bacteria, the binding of mRNAs and formation of the translational initiation complex mainly proceed in three stages: assembly of the 30S pre-initiation complex (30S PIC), transition into a mature 30S initiation complex (30S IC) with structural rearrangements, and final formation of the 70S initiation complex (70S IC) that is ready for elongation11,13–15. Initial binding of mRNA has been reported independent of any initiation factors and can take place at any moment during 30S PIC assembly16. This rapid binding step facilitates the swift recruitment of mRNAs to the 30S subunit of the ribosome, but with poor specificity8,17. To guarantee that translation starts at the proper site, mRNA binding to the ribosome depends on kinetic control based on multiple checkpoints as the initiation complex proceeds from the 30S PIC to the 30S IC and finally to the 70S IC11,16. Although extensive research has been carried out to study the effect of the initiation elements, the focus has mainly been on the manipulation of protein expression by the 70S IC, the final stage of the initiation complex. Moreover, although many experiments have been performed with a single RNA sequence to study the kinetics of mRNAs at different stages of initiation1,7,8,11,16, such experiments cannot easily reveal the entire range of sequence possibilities for mRNA recruitment. In this study, we attempted to elucidate the dynamic process of mRNA recruitment and to study the selection of mRNAs at a truly global level. We developed techniques based on high-throughput sequencing (deep sequencing) to clarify the behavior of the potential initiation elements at the equilibrium state in different stages of initiation complexes (Supplementary Fig. S1). To achieve this goal, artificial mRNA libraries were prepared with a randomized region at desired positions (Supplementary Fig. S2). 1 Biodynamic Optical Imaging Center (BIOPIC), and State Key Laboratory of Protein and Plant Gene Research, School of Life Sciences, Peking University, Beijing 100871, China. 2Department of Biochemistry and Structural Biology, Lund University, Lund, Sweden. *These authors contributed equally to this work. Correspondence and requests for materials should be addressed to R.G. (email: ) or X.-D.S. (email: ) Scientific Reports | 6:30170 | DOI: 10.1038/srep30170 1 www.nature.com/scientificreports/ Figure 1. Overall view of the base preference for mRNA library N20U with 20 randomized nucleotides in the TIR upstream of AUG. (a,b) Results from two independent repeats. In each repeat, the base preference was illustrated in three translation initiation stages: 30S, 30SIC and 70SIC. These mRNA libraries were selected by the 30S ribosome, the 30S IC, and the 70S IC individually (referred to as 30S, 30SIC, and 70SIC from here on). Our results provide a comprehensive view of the dynamic process of mRNA recruitment during translation initiation. Results As shown in Supplementary Figs S1 and S2, we used different mRNA libraries. Our experiments were carried out at equilibrium conditions. Briefly, each mRNA library was incubated with ribosomes (30S, 30SIC, or 70SIC) at 37 °C for 30 minutes to form mRNA-ribosome complexes. The bound RNAs were then separated from the unbound RNAs by binding His-tagged ribosomes to Ni-NTA resin with additional incubation at 4 °C for 16 hours. Overview of the sequence properties in the TIR upstream of AUG. To elucidate the sequence properties in the TIR upstream of AUG (−20 to −1), we synthesized the mRNA library N20U, which contains 20 randomized nucleotides upstream of the AUG codon (Supplementary Fig. S2). The mRNA library was selected by the 30S subunit, the 30SIC or the 70SIC, followed by deep sequencing with a non-selected mRNA library as the blank control (Supplementary Table S2). The results from two independently repeated experiments are provided in Fig. 1. The two repeats presented similar base preference in the TIR upstream of AUG. After determining the influence of the background library (with the relative value normalized to 1 for the background), we detected striking base preferences at different positions during the process of forming different initiation complexes. Obviously, this is a dynamic process with the initiation signals accumulated to different extents at different stages. A significant enrichment in G-bases in the middle region of the randomized 20-base region (−15 to −5) was observed during 30SIC and 70SIC formation, but not so obviously in 30S (...truncated)