Translational control of ERK signaling through miRNA/4EHP-directed silencing

MicroRNAs (miRNAs) exert a broad influence over gene expression by directing effector activities that impinge on translation and stability of mRNAs. We recently discovered that the cap-binding protein 4EHP is a key component of the mammalian miRNA-Induced Silencing Complex (miRISC), which mediates gene silencing. However, little is known about the mRNA repertoire that is controlled by the 4EHP/miRNA mechanism or its biological importance. Here, using ribosome profiling, we identify a subset of mRNAs that are translationally controlled by 4EHP. We show that the Dusp6 mRNA, which encodes an ERK1/2 phosphatase, is translationally repressed by 4EHP and a specific miRNA, miR-145. This promotes ERK1/2 phosphorylation, resulting in augmented cell growth and reduced apoptosis. Our findings thus empirically define the integral role of translational repression in miRNA-induced gene silencing and reveal a critical function for this process in the control of the ERK signaling cascade in mammalian cells.

Introduction mRNA translation commences with the binding of the eukaryotic initiation factor 4F (eIF4F) to the mRNA 5´cap structure. eIF4F is a three-subunit complex composed of eIF4E, the m 7 GpppN (cap)interacting factor; eIF4G, a scaffolding protein, and eIF4A, a DEAD-box RNA helicase (Sonenberg and Hinnebusch, 2009). eIF4G also interacts with eIF3, through which it recruits the pre-initiation complex, comprised of the 40S ribosomal subunit and associated factors, to the mRNA. Binding of the mRNA 5´cap by the 4E Homologous Protein (4EHP, encoded by Eif4e2), in contrast to eIF4E, impairs translation initiation (Cho et al., 2005;Morita et al., 2012;Rom et al., 1998). 4EHP shares 28% sequence identity with eIF4E (Rom et al., 1998) and is ubiquitously expressed, although it is 5-10 times less abundant than eIF4E in most cell types (Joshi et al., 2004). 4EHP binds the cap with 30-to 100-fold weaker affinity than eIF4E, but its affinity is increased by interactions with other proteins such as 4E-T or post-translational modification Okumura et al., 2007). 4EHP is involved in translational repression directed by miRNAs Chen and Gao, 2017). The miRNA-Induced Silencing Complex (miRISC) recruits the CCR4-NOT complex to effect mRNA translational repression and decay (Jonas and Izaurralde, 2015). CCR4-NOT in turn recruits DDX6, 4E-T (eIF4E-Transporter; a conserved 4EHP/ eIF4E-binding protein) and 4EHP to suppress cap-dependent mRNA translation Jonas and Izaurralde, 2015;Kamenska et al., 2014;Kamenska et al., 2016;Ozgur et al., 2015). However, which cellular mRNAs are targeted by 4EHP remains unknown.
The Extracellular signal-Regulated Kinases (ERK1/2) are important effectors of the highly conserved Mitogen-Activated Protein Kinase (MAPK) signaling pathway . ERK signaling is controlled by the RAS GTPase, which activates RAF, a serine/threonine kinase. RAF phosphorylates and activates the kinase MEK, which in turn phosphorylates and activates the effector serine/ threonine kinases ERK1/2. Activated ERK signaling elicits multiple outcomes, including transcriptional programs that control cellular functions such as cell proliferation (Aktas et al., 1997;Samatar and Poulikakos, 2014), apoptosis (Xia et al., 1995) and mRNA translation (Fukunaga and Hunter, 1997).
Here, we employed ribosome profiling to identify a subset of mRNAs that are regulated by 4EHP. We discovered that Dusp6 mRNA translation is repressed by a 4EHP/miRNA-dependent mechanism, which impacts on ERK1/2 phosphorylation, cell proliferation, and apoptosis. Our results underscore the biological importance of this translation repression mechanism, which is jointly orchestrated by miRNAs and 4EHP.

Enrichment for miRNA-binding sites in 4EHP-regulated mRNAs
We recently discovered that 4EHP acts as a key component of the translational repression machinery, which is mobilized by miRNAs . To identify mRNAs that are translationally controlled by 4EHP, we carried out ribosome profiling (Ingolia et al., 2011) in wild-type (WT) and 4EHP knockout (4EHP-KO) mouse embryonic fibroblasts (MEFs) (Figure 1-figure Supplement 1A and B). This assay measures the ribosome occupancy of each mRNA by deep sequencing of ribosome-protected mRNA fragments (ribosome footprints; RFPs) (Ingolia et al., 2011). We used the Babel tool Stumpf et al., 2013) to detect significant changes in translation efficiency (abundance of RFPs independently of changes in the levels of their corresponding mRNAs). Translation was up-regulated for 117 mRNAs (hereafter referred to as upregulated mRNAs) in 4EHP-KO in comparison to WT cells, while translation was down-regulated for 167 mRNAs ( Figure 1A and Supplementary file 1). Whereas the translational up-regulation of the mRNAs can be explained by the activity of 4EHP as translational suppressor, translational downregulation may be the result of indirect adaptation effects following 4EHP loss.
We next analyzed the upregulated mRNAs for the presence of common sequence features in their UTRs or coding sequences. A significant positive correlation was observed between the length of the 3 0 UTR and increased translation of the upregulated mRNAs in the 4EHP-KO cells (average of analysis of the association of eIF4E with 4EHP targets in 4EHP-KO MEFs. eIF4E was immunoprecipitated using a monoclonal antibody against eIF4E from WT and 4EHP-KO MEFs. Levels of the indicated mRNAs (normalized to b-actin mRNA) in the inputs and eIF4E-bound mRNAs were analyzed by RT-qPCR. Data are mean ± SD (n = 3). The p-value was determined by two-tailed Student's t-test: (ns) non-significant, (*) p<0.05; (**) p<0.01; (***) p<0.001. DOI: https://doi.org/10.7554/eLife.35034.002 The following figure supplement is available for figure 1: 2838.6, 2325.2, and 2016 nt for the up-regulated, unchanged and down-regulated mRNAs, respectively; p-value<2.2e-16; Figure 1B). We also found a less significant correlation (p=1.742e-05;  between the length of the 5´UTR and increased mRNA translation efficiency in the 4EHP-KO cells. This indicates that mRNAs with longer 3´UTR are more likely to be translationally repressed by 4EHP. mRNAs with long 3 0 UTR generally contain more miRNA-binding sites (Cheng et al., 2009). We examined the number of miRNA-binding sites in the 3 0 UTR of the up-regulated mRNAs (Agarwal et al., 2015). mRNAs which exhibit increased translation in 4EHP-KO cells, contained significantly more predicted miRNA-binding sites (642.8, 518.4, and 442.6 for the up-regulated, unchanged and down-regulated mRNAs, respectively; p-values: 0.0004, Figure 1C). We also calculated the density of miRNA-binding sites per 100-nucleotide of 3´UTR and found 22.9, 22.1, and 21.1 for the up-regulated, unchanged and down-regulated mRNAs, respectively (p-values: 0.0063, Figure 1D), indicating a greater density of miRNA-binding sites in 3´UTR of up-regulated mRNAs. These findings are in agreement with our previous report showing that 4EHP contributes to the translational silencing of miRNA targets by displacing eIF4E from the mRNA cap . To verify that this mechanism affects the up-regulated mRNAs, we performed RNA immunoprecipitation (RIP) with an anti-eIF4E antibody in WT and 4EHP-KO MEFs. IP resulted in specific recovery of eIF4E (Figure 1-figure Supplement 1D). We examined the enrichment of the top three most translationally up-regulated mRNAs in 4EHP-KO cells (Tmed7, Slc35e1 and Klhl21; Supplementary file 1) among the eIF4E-bound mRNAs ( Figure 1E). Slc35e1 and Klhl21 but not Tmed7 mRNAs were significantly enriched in eIF4E IP in 4EHP-KO cells in comparison with WT ( Figure 1E). Lyar and Iqgap1, which were among the most significant translationally down-regulated mRNAs, were not enriched in eIF4E IP as a consequence of 4EHP loss ( Figure 1F). These data show increased binding of eIF4E to the up-regulated mRNAs in 4EHP-KO cells, and indicate that 4EHP blocks the physical association of its target mRNAs with eIF4E.

4EHP-depletion impinges on cell viability and ERK1/2 phosphorylation
It was reported that while 4EHP expression is dispensable for growth in cell culture under physiological conditions, it is required under low oxygen conditions (Uniacke et al., 2014). However, at variance with these findings, we found that 4EHP-KO MEFs grew significantly slower than their WT counterparts (48 ± 3% less on day 6; p=0.002) under standard cell culture conditions (5% CO 2 and 20% O 2 ) ( Notably, FACS analysis showed that unlike in MEFs, depletion of 4EHP in U251 cells increased the fraction of cells in sub-G1, which is associated with apoptosis (shCTR: 0.9%, sh4EHP#1: 15.5%, and sh4EHP#2: 11.4; Figure 2C and The signaling pathways RAS/RAF/MEK/ERK and PI3K/mTOR control cell proliferation, growth and apoptosis, either in parallel or by converging on common downstream factors (Cagnol and Chambard, 2010;Laplante and Sabatini, 2012;Mendoza et al., 2011). We determined the phosphorylation levels of ERK1/2 and ribosomal protein S6 (RPS6) as respective markers of RAS/RAF/MEK/ERK and PI3K/mTOR activity by western blot (WB) analysis. While RPS6 phosphorylation remained unchanged, ERK1/2 phosphorylation (Thr202/Tyr204; pERK) was more than 80% reduced in 4EHP-KO MEFs in comparison with WT ( Figure 2D). A similar result was obtained in U251 cells upon 4EHP-knockdown (Figure 2-figure Supplement 1H). However, phosphorylation of MEK, the immediate upstream kinase of ERK1/2, remained unchanged in 4EHP-depleted cells ( Figure 2D  . These results suggest that the expression or activity of a factor upstream of ERK1/2, which is independent of MEK, is deregulated in 4EHP-depleted cells.

4EHP represses Dusp6 mRNA translation
We interrogated the 4EHP-KO MEF ribosome profiling data to identify candidate genes that could explain the strong impact of 4EHP on ERK1/2 phosphorylation. Interestingly, the mRNA encoding DUSP6, a potent and specific ERK1/2 phosphatase (Caunt and Keyse, 2013), was among the most translationally up-regulated transcripts in 4EHP-KO MEFs as compared to WT MEFs, with no significant change in its mRNA levels (Supplementary file 1). As expected, depletion of DUSP6 by shRNAs in U251 cells elicited ERK1/2 phosphorylation (Figure 2-figure Supplement 1I). To determine whether increased translation of Dusp6 mRNA in 4EHP-KO MEFs is because of enhanced initiation, which is the rate limiting step in translation, we performed polysome profiling, which resolves mRNAs on a sucrose gradient according to the number of ribosomes with which they associate ( . Importantly, restoring 4EHP expression in 4EHP-KO MEFs significantly reduced DUSP6 protein levels (~3 fold repression; Figure 2G). Taken together, these data demonstrate that 4EHP controls expression of the ERK1/2 phosphatase DUSP6 at the level of mRNA translation initiation.
Dusp6 3´UTR confers translational sensitivity to 4EHP To determine whether 4EHP regulates Dusp6 translation by displacing eIF4E from the cap Cho et al., 2005), we examined the association of Dusp6 mRNA with eIF4E in WT versus 4EHP-KO MEFs, using RIP. While Dusp6 mRNA levels were not significantly different reporter. Bottom; CTR, CNOT1, 4E-T, or 4EHP-knockdown cells were co-transfected with psiCHECK-FL-Dusp6 3´UTR reporter or the psiCHECK reporter (as control) in HEK293T cells. Luciferase activity was measured 24 hr after transfection. Firefly (F-Luc) values were normalized against Renilla (R-Luc) levels, and repression fold was calculated for the psiCHECK-FL-Dusp6 3´UTR reporter relative to psiCHECK reporter level for each condition. Data are mean ± SD (n = 3). (C) The psiCHECK reporter (control) or psiCHECK-RL with truncated fragments of the Dusp6 3´UTR were transfected into the HEK293T cells. Luciferase activity was measured 24 hr after transfection. R-Luc values were normalized against F-Luc levels, and repression fold was calculated for the psiCHECK-RL-Dusp6 3´UTR reporter relative to psiCHECK reporter level for each condition. Data are mean ± SD (n = 3). (D) WB for the indicated proteins in U251 cells transfected with si4EHP or the indicated miRNA inhibitors. (E) The psiCHECK reporter (control) or psiCHECK-FL-Dusp6 3´UTR were co-transfected along with the mock or miR-145 inhibitor in the control (CTR) or 4EHP-KO HEK293 cells. Luciferase activity was measured 24 hr after transfection. F-Luc values were normalized against R-Luc levels, and repression fold was calculated relative to the psiCHECK reporter/control inhibitor for each condition. Data are mean ± SD (n = 3). The p-values was determined by two-tailed Student's t-test: (ns) nonsignificant, (*) p<0.05; (**) p<0.01; (***) p<0.001. DOI: https://doi.org/10.7554/eLife.35034.006 The following figure supplement is available for figure 3: between the WT and 4EHP-KO cells ( Figure 3A; for corresponding WB analysis, see Figure 1-figure Supplement 1D), an 8-fold enrichment of Dusp6 mRNA was detected in eIF4E IP from 4EHP-KO MEF lysates, as compared to WT ( Figure 3A). As control, Dusp7 mRNA was not enriched in eIF4E IP from 4EHP-KO MEFs lysates. These data lend further support to our model of displacement of eIF4E from the cap by 4EHP, and demonstrate that this mechanism causes translational repression of Dusp6 mRNA.
3´UTRs effect mRNA translation through trans-acting factors such as RNA-binding proteins (RBPs) and miRNAs (Szostak and Gebauer, 2013). DUSP6 expression is regulated by miRNAs including miR-145 (Gu et al., 2015), miR-181a (Li et al., 2012), and the RBP PUM2 (Bermudez et al., 2011), a homolog of Drosophila pumilio. We thus sought to study the role of the 3´UTR of Dusp6 mRNA in translational repression by 4EHP. To this end, 3´rapid amplification of cDNA ends (3´RACE) analysis was performed to amplify the 3´UTR of Dusp6 mRNA in U251 cells. A 1192-nucleotides segment was amplified (Supplementary file 2) and cloned into the psiCHECK-2 luciferase reporter vector. The resulting construct was transfected into HEK293T cells along with control siRNA (siCTR) or siRNA against 4EHP (si4EHP), or its partners CNOT1 (siCNOT1) and 4E-T (si4E-T). In the siCTR-transfected cells, the 3´UTR of Dusp6 mRNA caused a 3-fold repression in comparison with the backbone reporter alone ( Figure 3B). However, knockdown of 4EHP or its partners CNOT1 and 4E-T significantly de-repressed the psiCHECK-Dusp6-3´UTR reporter (38%, 49%, and 44% respectively as compared to siCTR; Figure 3B), thus supporting the role of CCR4-NOT/4E-T/4EHP pathway in Dusp6 mRNA translational repression. Consistent with the latter results, knockdown of CNOT1 and CNOT9, two critical subunits of the CCR4-NOT complex, also led to an increase of DUSP6 protein amounts in U251 cells (1.4 and 2.2-folds, respectively; Figure 3-figure Supplement 1A).
We next mapped the repressive activity of 4EHP to elements of the 3´UTR of Dusp6 mRNA. To this end, we sub-cloned six~200 nt fragments of the 3´UTR into the psiCHECK-2 luciferase reporter ( Figure 3-figure Supplement 1B). A segment harbouring both miR-145 and miR-181a binding sites exerted the strongest repression on the reporter (1.5 fold; p=0001, Figure 3C), which was alleviated upon 4EHP knockdown ( Figure 3C). To identify which miRNA is responsible for repression of Dusp6 mRNA, we used specific inhibitors to block miR-145, miR-181a, and miR-124 in U251 cells. While blocking miR-124 and miR-181a did not affect DUSP6 expression, a miR-145 inhibitor increased DUSP6 accumulation to a similar degree as knockdown of 4EHP ( Figure 3D), without affecting the stability of the Dusp6 mRNA (Figure 3-figure Supplement 1C). We further investigated the effect of miR-145 inhibitor on a luciferase reporter with the full-length Dusp6 3´UTR. Unlike the control reporter, the expression of the reporter containing Dusp6 3´UTR was significantly de-repressed in the presence of miR-145 inhibitor (1.25 fold repression compared with 2.09 for mock inhibitor; Figure 3E). Consistent with our observation that siRNA depletion of 4EHP in HEK293T cells derepressed the Dusp6 3´UTR reporter ( Figure 3B), silencing of the same reporter was fully reversed in a 4EHP-KO HEK293 cells ( Figure 3E). No de-repression by the miR-145 inhibitor was observed in 4EHP-KO HEK293 cells ( Figure 3E). This confirms the requirement for 4EHP in miR-145-induced translational silencing of Dusp6 mRNA. Taken together, these data demonstrate that the Dusp6 mRNA translation is controlled by its 3´UTR through the miRNA/CCR4-NOT/4E-T/4EHP pathway.

De-repression of DUSP6 impedes ERK activity and proliferation in 4EHP-depleted cells
We next sought to determine the consequences of DUSP6 de-repression on ERK signaling and functions in 4EHP-KO MEFs. We used a selective small molecule inhibitor of DUSP6, 2-benzylidene-3-(cyclohexylamino)À1-Indanone hydrochloride (BCI) (Molina et al., 2009;Shojaee et al., 2015). Treatment of 4EHP-KO cells with BCI increased pERK1/2 to levels comparable with untreated WT cells within 30 min ( Figure 4A). Similar results were obtained with U251 cells expressing an shRNA against 4EHP (Figure 4-figure Supplement 1A). These data confirm that reduced ERK1/2 phosphorylation in 4EHP-depleted cells is due to increased DUSP6 activity. Next, we examined the consequence of DUSP6 inhibition on proliferation of 4EHP-depleted cells by using shRNAs to knockdown DUSP6 in WT and 4EHP-KO cells (Figure 4- Figure 4B). This result demonstrates that the reduced proliferation of 4EHP-KO cells is at least partially due to de-repression of DUSP6.
Extracellular signals or mutations in Ras or Raf, which occur frequently in cancers, activate a phosphorylation cascade that results in phosphorylation and activation of ERK signaling (Samatar and Poulikakos, 2014). We examined whether 4EHP-depletion and the resulting increased DUSP6 expression could interfere with ERK1/2 phosphorylation in response to upstream activation of RAS. To this end, we expressed a constitutively active mutant KRAS (G12V) (Prior et al., 2012) and monitored ERK signaling by WB and proliferation assays. While ERK1/2 phosphorylation was increased by forced KRAS activity in WT MEFs, pERK levels remained unchanged in 4EHP-KO MEFs ( Figure 4C). Consistent with these results, WT MEFs proliferation was slightly increased upon enforced KRAS activity, but remained unaffected in 4EHP-KO MEFs (Figure 4-  Taken together, the data demonstrate that 4EHP up-regulates ERK1/2 phosphorylation by effecting the miRNA-induced translational repression of Dusp6 mRNA, and that depletion of 4EHP limits ERK activation by upstream signaling (Figure 4D, model).

Discussion
We previously demonstrated that the cap-binding protein 4EHP acts as an effector of translational repression instigated by miRNAs. Here, we identify Dusp6 mRNA as a functionally critical target of this silencing mechanism, which occurs in the absence of mRNA decay. Translational repression of Dusp6 mRNA by the combined action of 4EHP and miR-145 down-regulates the MAPK/ERK signaling cascade and its output in cell proliferation and survival. The 4EHP/miRNA repression mechanism thus engenders important biological consequences in homeostasis and disease.
The relative contributions of translational repression and mRNA decay in miRNA-mediated silencing are in dispute. Several large-scale studies reported that mammalian miRNAs predominantly act by decreasing target mRNA levels (Baek et al., 2008;Eichhorn et al., 2014;Guo et al., 2010), while others showed that miRNAs affect the expression of target genes by translation inhibition (Jin et al., 2017;Selbach et al., 2008;Yang et al., 2009). It was convincingly demonstrated in in vitro and in vivo studies that translational repression precedes target mRNA decay (Bazzini et al., 2012;Béthune et al., 2012;Djuranovic et al., 2012;Fabian et al., 2009;Mathonnet et al., 2007). Because of their intricate nature, the exact contribution of either aspect of miRNA-mediated silencing in biological decisions has remained elusive. Our data demonstrate that 4EHP effects miRNAmediated translational repression of Dusp6 mRNA, but not mRNA stability. The relative contribution of translational repression and mRNA degradation to miRNA-mediated silencing may thus depend on the target mRNAs and on the cellular context. Expression of miRISC core and accessory components, post-translational modifications, translation efficiency, RNA structure within a 3´UTR, or interactions with RNA-binding proteins (RBPs) may interfere or promote miRISC activities (Cottrell et al., 2018;Cottrell et al., 2017;Kedde et al., 2010;Kundu et al., 2012;Long et al., 2007). The RBPs PUM2 and TTP were implicated in the post-transcriptional repression of Dusp6 mRNA, presumably in a CCR4-NOT-dependent mechanism (Bermudez et al., 2011;Galgano et al., 2008). Since the abundance of RBPs varies in tissues and under pathological conditions, it is conceivable that the potency and the nature of the miRNA-mediated silencing mechanism are modulated by such RBPs.
Our study underscores the importance of translational control in regulation of the ERK signaling pathway. Indirect up-regulation of ERK1/2 phosphorylation by 4EHP, via repression of Dusp6 translation, explains the diminished cell proliferation in 4EHP-KO MEF cells and apoptosis observed in 4EHP-depleted U251 and U87 cells. A notable observation in our study is the impairment of the RAS/RAF/MEK/ERK pathway in 4EHP-depleted cells. Specifically, constitutively active RAS fails to increase ERK1/2 phosphorylation in 4EHP-KO MEFs. This can be explained by increased DUSP6 expression in 4EHP-KO cells, which effectively impairs phosphorylation of ERK1/2 downstream of RAS. Interestingly, over-expression of constitutively active RAS (Park et al., 2014), or BRAF (Agrawal et al., 2014), also induces DUSP6 expression constituting a negative feedback loop. The feedback loop restrains the activity of the RAS/RAF/MEK/ERK pathway upon induction by stimuli (e.g. growth factors). Thus, increasing DUSP6 expression by inhibiting 4EHP can potentially repress ERK pathway activation. While several pharmacological approaches have been described for targeting eIF4E (Fischer, 2009;Graff et al., 2007), to date no specific inhibitor of 4EHP has been discovered. The elucidation of the crystal structures of 4EHP in association with its binding partners (Peter et al., 2017;Rosettani et al., 2007) may prove useful for this purpose.
Our ribosome profiling data strongly suggest that translational repression through miRNA/4EHP impacts on many other mRNAs. An interesting miRNA to revisit in light of this mechanism is let-7, which suppresses tumorigenesis by directly silencing RAS expression (Johnson et al., 2005). We had previously shown that 4EHP contributes to the translational repression activity of a reporter mRNA by let-7 miRNA , but let-7 miRNA can also clearly instigate mRNA deadenylation and decay. The relative contributions of translation repression and mRNA decay in the function of miRNA/mRNA pairs may be further revealed by systematically addressing their epistasis with 4EHP in the relevant cellular context.
Lentivirus production 8 Â 10 6 HEK293 FT (Thermo Fisher Scientific, R70007) cells were cultured in a 10 cm dish for 24 hr in high glucose DMEM supplemented with 10% v/v FBS. Medium was replaced by OptiMEM (Thermo Fisher Scientific, 51985091) 30 min before transfection. Lentivirus particles were produced by transfecting the HEK293FT cells using Lipofectamine 2000 and 10 mg shRNA plasmid, 6.5 mg psPAX2 (Addgene, plasmid 12260) and 3.5 mg pMD2.G (Addgene, plasmid 12259) packaging plasmids. 5 hr post-transfection, the medium was replaced with fresh high glucose DMEM supplemented with 10% v/v FBS. Supernatant was collected at 48 hr post-transfection, replaced with fresh medium and collected after 24 hr. Viral particles were cleared by filtration (45 mm; Fisher Scientific, 09-720-005) and virus titer was measured by colony formation assay using 293FT cells. The multiplicity of infection (MOI) was adjusted to~5. Virus solution was stored at À80˚C without cryopreservative in 1 ml aliquots or used to infect the cells directly in the presence of 6 mg/ml polybrene (Sigma, H9268).
CRISPR-Cas9 genome engineering for generating 4ehp knockout HEK293 cell line CRISPR-Cas9-mediated genome editing of Flp-In T-REx HEK293 cells was performed as previously described (Ran et al., 2013). Two small guide RNAs (sgRNAs) cognate to the coding region of 4EHP gene: 5´-CAACAAGTTCGACGCGTGAG and 5´-TGAGCTCGTGGGACGGCCGG were designed. The top and bottom strands of each designed sgRNA were annealed creating overhangs for cloning of the guide sequence oligos into pSpCas9(BB)À2A-GFP (Addgene, PX458, Plasmid #48138) by BbsI digestion. To generate gene knockout Flp-In T-REx HEK293 cells, we transfected 130.000 cells with the corresponding guide sequence containing pSpCas9(BB)À2A-GFP plasmid. 24 hr after transfection, GFP-positive single cells were sorted by FACS into 96-well plates and cultivated until colonies were obtained. Clonal cell lines were analyzed by WB for protein depletion as well as by PCR-genotyping. The following primers were used for the PCR-genotyping: sense primer1, 5´-GCCGCCC TGAGCTGGCGTCCC; anti-sense primer1, 5´-CGGCACAGCCACCCCTCCCCC; sense primer2, 5´-GCAGAATCTTTGGCACATTGCAGATAGTTGAGG; anti-sense primer2, 5´-GCCCTTCTGATCAACTC TACAATTCTCATATTTGTTGATACC. PCR products were cloned using the Zero Blunt PCR Cloning Kit (Thermo Fisher Scientific, K270040) and 10 clones sequenced per cell line.

Real-Time RT-qPCR
1 mg of DNase I-treated total RNA, purified using the TRI-Reagent, was reverse-transcribed using 100 ng of random primers following the Superscript III (Invitrogen) protocol. Real time PCR was performed with SYBR Green master mix (iQ; Biorad) in a real-time PCR detection system (Mastercycler Realplex, Eppendorf). Mean values of triplicate measurements were calculated according to the -DD Ct quantification method, and were normalized against the expression of the indicated mRNA. Specificity was confirmed by analyzing the melting curves of PCR products. RT-qPCR results were repeated at least three times in independent experiments and representative data sets are shown. Sequences of the used primers are listed in the Supplementary file 3.
3´rapid amplification of cDNA ends (3´RACE) 3 0 RACE was performed with the SMARTer RACE 5 0 /3 0 kit (Cat # 634858, Clontech, Mountain View, CA). 1 mg of total RNAs extracted from U251 cells was treated with DNase I (Fermentas) and cDNA was generated by the SMARTScribe Reverse Transcriptase (Clontech), according to the manufacturer's instructions. The resultant cDNA was used for PCR amplification using the human DUSP6 gene-specific forward primers (GSPs) (Supplementary file 2) together with a common Universal Reverse Primer (UPM), provided by the manufacturer. PCR products were resolved by agarose gel electrophoresis and all visible bands were excised and digested by restriction enzymes followed by cloning into the PUC19 vector provided by the manufacturer and sequenced by Sanger sequencing.

RNA immunoprecipitation (RIP)
RIP was performed as described previously (Thoreen et al., 2012) with minor modifications. WT and 4EHP-KO MEFs were seeded in 3 Â 15 cm plates (at 10 Â 10 6 cells per plate) and incubated overnight. Cells were lysed in lysis buffer A (50 mM HEPES-KOH (pH: 7.4), 2 mM EDTA, 10 mM pyrophosphate, 10 mM beta-glycerophosphate, 40 mM NaCl, 1% Trition X-100 and one tablet of EDTAfree protease inhibitors (Roche)) containing 40 U/ml SuperaseIn. Insoluble material was removed by centrifugation at 20,000xg for 5 min at 4˚C. Protein concentration was measured by Bradford assay and 2 mg of lysate was pre-cleared by incubating with 50 ml of 50% protein G agarose fast flow beads (EMD Millipore, for 2 hr at 4˚C with gentle agitation. The cleared lysates were collected by centrifugation at 3000xg for 1 min at 4˚C and collecting the supernatant. In parallel 2 mg of anti-eIF4E antibody was incubated with 50 ml of 50% protein G agarose fast flow beads for on an end-over-end rotator for 2 hr at 4˚C. For IP, the pre-cleared lysates were incubated with the antibody + bead mixture, in 1 ml total volume on an end-over-end rotator for 2 hr at 4˚C. The precipitated beads were then washed 3x with 1 ml buffer A, twice with buffer B (15 mM HEPES-KOH (pH 7.4), 7.5 mM MgCl2, 100 mM KCl, 2 mM DTT and 1.0% Triton X-100), and resuspended in 100 ml buffer B. 10 ml of the final mix was used for WB and the remaining was used for RNA extraction.

Collection of ribosome footprints (RFPs)
Ribosome profiling was performed as described (Ingolia et al., 2012), with minor modifications. Briefly, 500 mg of the ribonucleoproteins were subjected to ribosome footprinting by RNase I treatment at 4˚C for 45 min with end-over-end rotation. Monosomes were pelleted by ultracentrifugation in a 34% sucrose cushion at 70,000xrpm for 3 hr and RNA fragments were extracted twice with acid phenol, once with chloroform, and precipitated with isopropanol in the presence of NaOAc and Gly-coBlue. Purified RNA was resolved on a denaturing 15% polyacrylamide-urea gel and the section corresponding to 28-32 nucleotides containing the RFPs was excised, eluted, and precipitated by isopropanol.
Random RNA fragmentation and mRNA-Seq 100 mg of cytoplasmic RNA was used for mRNA-Seq analysis. Poly(A)+ mRNAs were purified using magnetic oligo-dT DynaBeads (Invitrogen) according to the manufacturer's instructions. Purified RNA was eluted from the beads and mixed with an equal volume of 2X alkaline fragmentation solution (2 mM EDTA, 10 mM Na2CO3, 90 mM NaHCO3, pH 9.2) and incubated for 20 min at 95˚C. Fragmentation reactions were mixed with stop/precipitation solution (300 mM NaOAc pH 5.5 and GlycoBlue), followed by isopropanol precipitation. Fragmented mRNA was size-selected on a denaturing 10% polyacrylamide-urea gel and the area corresponding to 35-50 nucleotides was excised, eluted, and precipitated with isopropanol.

Library preparation and sequencing
Fragmented mRNAs and RFPs were dephosphorylated using T4 polynucleotide kinase (New England Biolabs). Denatured fragments were resuspended in 10 mM Tris (pH 7) and quantified using the Bio-Analyzer Small RNA assay (Agilent). 10 pmol of RNA was ligated to the 3 0 -adaptor with T4 RNA ligase 1 (New England Biolabs) for 2 hr at 37˚C. Reverse transcription was carried out using oNTI223 adapter (Illumina) and SuperScript III reverse transcriptase (Invitrogen) according to the manufacturer's instructions. Products were separated from the empty adaptor on a 10% polyacrylamide Tris/ Borate/EDTA-urea (TBE-urea) gel and circularized by CircLigase (Epicentre). Ribosomal RNA amounts were reduced by subtractive hybridization using biotinylated rDNA complementary oligos (Ingolia et al., 2012). The mRNA and ribosome-footprint libraries were amplified by PCR (12 cycles) using indexed primers and quantified using the Agilent BioAnalyzer High-Sensitivity assay. DNA was then sequenced on the HiSeq 2000 platform with read length of 50 nucleotides (SR50) according to the manufacturer's instructions, with sequencing primer oNTI202 (5CGACAGGTTCAGAGTTC TACAGTCCGACGATC).

Analysis of ribosome profiling data
Prior to alignment, linker and polyA sequences were removed from the 3´ends of reads. Bowtie v0.12.7 (allowing up to two mismatches) was used to perform the alignments. First, reads that aligned to rRNA sequences were discarded. All remaining reads were aligned to the mouse genome (mm10). Finally, still-unaligned reads were aligned to the mouse known canonical transcriptome that includes splice junctions. Reads with unique alignments were used to compute the total number of reads at each position. Footprints and mRNA densities were calculated in units of reads per kilobase per million (RPKM) to normalize for gene length and total reads per sequencing run. The expression patterns were examined for genes that had more than 150 uniquely aligned reads of mRNA and footprints in one of the samples. The Babel computational framework was used to quantitatively evaluate if there are genes that are differently translated in KO cells. The 5´and 3´UTRs were obtained from the UCSC Genome Browser. For translationally induced or repressed genes the length of 5´and 3´UTRs were calculated and compared using Welch Two Sample t-test. Predicted miRNA sites were retrieved from TargetScanMouse. Both conserved and non-conserved sites were taken into account. The number of miRNA sites per 100 bp of 3´UTR was calculated using the 3Ú TR lengths published on TargetScanMouse. The GEO accession numbers for the sequencing data reported in this paper is GSE107826.
RNA stability assay 300,000 cells were plated in 6-well plates and 5 mg/ml actinomycin D (Sigma) was added to the culture medium at the indicated times. RNA was isolated by using Tri Reagent (Sigma-Aldrich, St. Louis, Missouri), according to the manufacturer's protocol and the stability of the indicated transcript was measured by RT-qPCR with the primers indicated in Supplementary file 3.

Preparation of reporter constructs
To generate luciferase reporter plasmids, a modified version of psiCHECK-2 (Promega, Madison, WI) containing the Gateway cassette C.1 (Invitrogen) at the 3´end of the firefly luciferase (F-Luc) gene was used as described before (Suffert et al., 2011). The 3 0 UTR sequence of Dusp6 mRNA inserted in the PUC19 vector was obtained from the U251 cells by 3´RACE assay. The attB-Dusp6 fragment was obtained by PCR with the primers indicated in Supplementary file 3, cloned into pDONR/Zeo (Invitrogen) and recombined in the modified psiCHECK-2 vector by Gateway cloning. The fragments of the 3 0 UTR of Dusp6 were obtained by PCR from the psiCHECK-Dusp6 3´UTR vector and inserted as XhoI-NotI fragments into the psiCHECK-2 vector at the 3 0 -end of the Renilla luciferase gene (R-Luc). Sequences of the used primers are listed in the Supplementary file 3.

Luciferase reporter assay
HEK293T and U251 cells (150,000 cells/well) were co-transfected in a 24-well plate with 10 ng psi-CHECK-Dusp6 3´UTR. For 4EHP knockdown, 4 Â 10 6 cells were plated in a 10 cm culture dish and transfected with a final concentration of 25 nM of siRNA duplexes using Lipofectamine 2000 according to the manufacturer's instructions. After 24 hr, cells were plated in a 24-well plate and transfected a second time with the psiCHECK vectors as described above. Cells were lysed 24 hr after transfection. Luciferase activities were measured with the Dual-Luciferase Reporter Assay System (Promega) in a GloMax 20/20 luminometer (Promega). For experiments with miRNA inhibitors, HEK293 cells were co-transfected in a 24-well plate with 10 ng psiCHECK-Dusp6 3´UTR and miRNA inhibitors were added to the transfection mixture at a final concentration of 50 nM.

Major datasets
The following dataset was generated: