Downloaded from genome.cshlp.org on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press 1 nanoCAGE reveals 5’UTR features that define specific modes of translation of 2 functionally related MTOR-sensitive mRNAs 3 Valentina Gandin1,2,3,4*, Laia Masvidal5*, Laura Hulea1,2*, Simon-Pierre Gravel4,6, Marie 4 Cargnello1,2, Shannon McLaughlan1,2, Yutian Cai1,4, Preetika Balanathan7, Masahiro 5 Morita4,6, Arjuna Rajakumar1, Luc Furic7, Michael Pollak1,2,3, John A. Porco, Jr.8, Julie 6 St-Pierre4,6, Jerry Pelletier2,4,6,, Ola Larsson5** and Ivan Topisirovic1,2,3,4** 7 1Lady Davis Institute, SMBD Jewish General Hospital, Montreal, Canada; Departments 8 of 2Oncology, 3Experimental Medicine, and 4Biochemistry McGill University, Montreal, 9 Canada; 5Department of Oncology-Pathology, Science for Life Laboratory, Karolinska 10 Institutet, Stockholm, Sweden; 6Goodman Cancer Research Centre, McGill University, 11 Montreal, Canada; 7Cancer Program, Biomedicine Discovery Institute and Department of 12 Anatomy & Developmental Biology, Monash University, VIC, Australia, 8Center for 13 Chemical Methodology and Library Development, Boston University, 590 14 Commonwealth Avenue, Boston, Massachusetts 02215, United States 15 16 *equally contributing 17 18 Correspondence: Ola Larsson, Department of Oncology-Pathology, Karolinska Institutet 19 Tomtebodavägen 23A, 171 65 Solna Sweden Tel: +46 (0)8 524 81 228; Email: 20 [email protected] & Ivan Topisirovic, Lady Davis Institute for Medical Research, Office 21 E-445; 3755 Côte Ste-Catherine Road, Montreal, Quebec H3T 1E2, Canada; Tel: (514) 22 340-8222 ext 3146; Email: [email protected] 23 Running title: 5’UTR features of MTOR-sensitive mRNAs. 1 Downloaded from genome.cshlp.org on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press 24 Key words: nanoCAGE, mRNA translation, MTOR, EIF4E, EIF4A1 25 Abstract 26 The diversity of MTOR-regulated mRNA translation remains unresolved. Whereas 27 ribosome-profiling suggested that MTOR almost exclusively stimulates translation of 28 TOP (terminal oligopyrimidine motif) and TOP-like mRNAs, polysome-profiling 29 indicated that MTOR also modulates translation of mRNAs without 5’TOP motif (non- 30 TOP mRNAs). We demonstrate that in ribosome-profiling studies detection of MTOR- 31 dependent changes in non-TOP mRNA translation was obscured by low sensitivity and 32 methodology biases. Transcription start site profiling using Nano-Cap Analysis of Gene 33 Expression (nanoCAGE) revealed not only do many MTOR-sensitive mRNAs lack 34 5’TOP motif but that 5’UTR features distinguish two functionally, and translationally 35 distinct subsets of MTOR-sensitive mRNAs: i) mRNAs with short 5’UTR enriched for 36 mitochondrial functions, that require EIF4E, but are less EIF4A1-sensitive and ii) long 37 5’UTRs mRNAs encoding proliferation- and survival-promoting proteins, that are both 38 EIF4E- and EIF4A1-sensitive. Selective inhibition of translation of mRNAs harboring 39 long 5’UTRs via EIF4A1 suppression leads to sustained expression of proteins involved 40 in respiration but concomitant loss of those protecting mitochondrial structural integrity, 41 resulting in apoptosis. Conversely, simultaneous suppression of translation of both long 42 and short 5’UTR mRNAs, by MTOR inhibitors, results in metabolic dormancy and a 43 predominantly cytostatic effect. Thus, 5’UTR features define different modes of MTOR- 44 sensitive translation of functionally distinct subsets of mRNAs which may explain the 45 diverse impact of MTOR and EIF4A inhibitors on neoplastic cells. 46 47 48 49 50 51 2 Downloaded from genome.cshlp.org on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press 52 Introduction: 53 Genome-wide gene expression studies have mostly focused on measuring ‘steady-state’ 54 mRNA abundance that reflects alterations in transcription and mRNA stability (Piccirillo 55 et al. 2014). Changes in steady-state mRNA abundance do not, however, completely 56 mirror those occurring in the corresponding proteome (Gygi et al. 1999; de Sousa Abreu 57 et al. 2009; Larsson et al. 2012; Kristensen et al. 2013; Ly et al. 2014). Although still 58 debated (Li and Biggin 2015), these studies implicate post-transcriptional mechanisms, 59 including mRNA translation, as key influencers over the proteome (Schwanhausser et al. 60 2011; Vogel and Marcotte 2012; Kristensen et al. 2013; Li et al. 2014; Jovanovic et al. 61 2015). 62 63 Control of translation largely occurs at the initiation step, during which the mRNA is 64 recruited to the ribosome by the eIF4F complex (Hinnebusch 2014). This complex 65 consists of the mRNA cap-binding protein EIF4E, a large scaffolding protein EIF4G and 66 DEAD box helicase EIF4A (Hinnebusch 2014). eIF4F complex assembly is regulated by 67 the mechanistic/mammalian target of rapamycin (mTOR) complex 1 (mTORC1) which 68 phosphorylates and inactivates 4E-binding proteins (EIF4EBP1, 2 and 3) (von Manteuffel 69 et al. 1996; Hara et al. 1997; Burnett et al. 1998; Gingras et al. 1999; Gingras et al. 2001). 70 EIF4EBPs bind to EIF4E thereby preventing EIF4E:EIF4G association (Pause et al. 71 1994a). Phosphorylation of EIF4EBPs leads to their dissociation from EIF4E, allowing 72 EIF4E:EIF4G interaction and eIF4F assembly (Gingras et al. 1999; Gingras et al. 2001). 73 Although EIF4E is required for cap-dependent translation of all nuclear-encoded 74 mRNAs, a subset of mRNAs characterized by long and complex 5’UTRs that encode 3 Downloaded from genome.cshlp.org on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press 75 proliferation- (e.g. cyclins), survival- (e.g. BCL2 family members) and tumor-promoting 76 proteins (e.g. MYC) are thought to be particularly sensitive to EIF4E (Koromilas et al. 77 1992; Graff et al. 2008; Roux and Topisirovic 2012; Pelletier et al. 2015). 78 79 UTR features dictate translation efficiency (Sonenberg and Hinnebusch 2009). In 80 mammals, mRNAs with long, complex 5’UTRs exhibit a high EIF4A helicase 81 requirement for scanning of 43S complex towards the initiation codon (Svitkin et al. 82 2001). EIF4A processivity, as a single protein, is low (Pause et al. 1994b). EIF4A:EIF4G 83 interaction dramatically increases EIF4A processivity (Garcia-Garcia et al. 2015). EIF4E 84 recruits EIF4A to the eIF4F complex thereby bolstering EIF4A activity (Feoktistova et al. 85 2013). Recent ribosome-profiling studies indicate that the mTORC1/EIF4EBP/EIF4E 86 pathway almost exclusively regulates translation of mRNAs harboring 5’ terminal 87 oligopyrimidine (5’TOP) and related motifs (Hsieh et al. 2012; Thoreen et al. 2012). 88 5’TOP motif consists of a C immediately after the mRNA cap followed by 4-to-15 89 pyrimidines (Meyuhas and Kahan 2014). TOP mRNAs encode translational machinery 90 components, including ribosomal proteins and elongation factors (Meyuhas and Kahan 91 2014). 92 93 The generality of the conclusions that EIF4EBPs are major regulators of TOP mRNA 94 translation has been questioned since, under a number of conditions, translation of TOP 95 mRNAs, albeit MTOR-dependent, is EIF4EBP-independent (Miloslavski et al. 2014). 96 Consistent with this view, the effects of serum deprivation on TOP mRNA translation 97 also appeared to be EIF4E-independent (Shama et al. 1995). In turn, LARP1 and 4 Downloaded from genome.cshlp.org on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press 98 TIA1/TIAL1 (also known as TIAR) were reported to mediate effects of mTORC1 on 99 TOP mRNAs (Damgaard and Lykke-Andersen 2011; Tcherkezian et al. 2014; Fonseca et 100 al. 2015). Polysome-profiling showed that the mTORC1/EIF4EBP axis also regulates 101 translation of non-TOP mRNAs, including those encoding mitochondrial and survival- 102 and proliferation-promoting proteins (Larsson et al. 2012; Morita et al. 2013). In fact, it 103 appears that ribosome- and polysome-profiling studies provided strikingly different sets 104 of MTOR-sensitive mRNAs. 105 106 Elucidating the precise mechanisms that control mRNA translation also requires accurate 107 knowledge of transcription start sites (TSS). Despite substantial efforts to define TSS 108 using CAGE/nanoCAGE (Consortium et al. 2014) or TSS-seq (Suzuki and Sugano 2003; 109 Suzuki et al. 2015) there are no resources, which provide TSS information for most 110 commonly used model cell lines. Databases such as RefSeq or UCSC are thus routinely 111 used to infer the link between 5’UTR features and translation. These databases are 112 thought to contain numerous errors and for a more accurate understanding of the 113 relationship between the 5’UTR features and translational control both TSS and 114 translation efficiency should be determined in the same cell. 115 116 Results: 117 Simulating the impact of differences in polysome-association on performance of 118 ribosome- and polysome-profiling 119 We compared ribosome- and polysome-profiling for their ability to identify mRNAs 120 undergoing different shifts in translational efficiency. In polysome-profiling, translation 5 Downloaded from genome.cshlp.org on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press 121 efficiency is directly determined by separating efficiently (commonly defined as mRNAs 122 associated with >3 ribosomes) and not efficiently translated mRNAs (defined as mRNAs 123 associated with 3 or less ribosomes) (Gandin et al. 2014). In contrast, with ribosome- 124 profiling translational efficiency is inferred based on the number of RNA sequencing 125 reads corresponding to ribosome-protected fragments (Gandin et al. 2014; Ingolia 2014; 126 King and Gerber 2014). TOP mRNAs are highly expressed (Fig. 1A-B), engage a 127 considerable proportion of cellular ribosomes when MTOR is active (Levy et al. 1991; 128 Meyuhas and Kahan 2014; Miloslavski et al. 2014) and show large shifts in translational 129 efficiency upon MTOR inhibition, as compared to mRNAs that do not contain 5’TOP 130 motif [non-TOP mRNAs; Sup. Fig. 1 (Levy et al. 1991; Miloslavski et al. 2014)]. 131 132 We next determined performance of ribosome- vs. polysome-profiling in identifying 133 differentially translated mRNAs as a function of the size of shifts in translational 134 efficiency using simulation approach.