Global analysis of LARP1 translation targets reveals tunable and dynamic features of 5′ TOP motifs

• DOI: 10.1073/pnas.1912864117
• Publication Date: 2020-02-25T01:25:28Z
• Authors: Lucas Philippe; Antonia M. G. van...; Maegan J. Watson; Carson C. Thoreen
• Keywords: 0303 health sciences; Mechanistic Target of Rapamycin Complex 1; Autoantigens; SS-B Antigen; 03 medical and health sciences; HEK293 Cells; Pyrimidines; Ribonucleoproteins; Protein Biosynthesis; Humans; RNA, Messenger; Nucleotide Motifs; Transcriptome; Polypyrimidine Tract-Binding Protein

Terminal oligopyrimidine (TOP) motifs are sequences at the 5′ ends of mRNAs that link their translation to the mTOR Complex 1 (mTORC1) nutrient-sensing signaling pathway. They are commonly regarded as discrete elements that reside on ∼100 mRNAs that mostly encode translation factors. However, the full spectrum of TOP sequences and their prevalence throughout the transcriptome remain unclear, primarily because of uncertainty over the mechanism that detects them. Here, we globally analyzed translation targets of La-related protein 1 (LARP1), an RNA-binding protein and mTORC1 effector that has been shown to repress TOP mRNA translation in a few specific cases. We establish that LARP1 is the primary translation regulator of mRNAs with classical TOP motifs genome-wide, and also that these motifs are extreme instances of a broader continuum of regulatory sequences. We identify the features of TOP sequences that determine their potency and quantify these as a metric that accurately predicts mTORC1/LARP1 regulation called a TOPscore. Analysis of TOPscores across the transcriptomes of 16 mammalian tissues defines a constitutive “core” set of TOP mRNAs, but also identifies tissue-specific TOP mRNAs produced via alternative transcription initiation sites. These results establish the central role of LARP1 in TOP mRNA regulation on a transcriptome scale and show how it connects mTORC1 to a tunable and dynamic program of gene expression that is tailored to specific biological contexts.