Cell Reports - Cell Press

40 downloads 551 Views 3MB Size Report
May 28, 2015 - 3Lunenfeld-Tanenbaum Research Institute, Mount Sinai Hospital, 600 ..... 3C). In contrast, a 4E-T C-terminal fragment lacking amino acids.
Article

The eIF4E-Binding Protein 4E-T Is a Component of the mRNA Decay Machinery that Bridges the 50 and 30 Termini of Target mRNAs Graphical Abstract

Authors Tamiko Nishimura, Zoya Padamsi, ..., Anne-Claude Gingras, Marc R. Fabian

Correspondence [email protected]

In Brief In this study, Nishimura et al. demonstrate that the human eIF4Ebinding protein 4E-T is a component of the mRNA decay machinery. 4E-T promotes the turnover of mRNAs targeted by the CCR4-NOT deadenylase complex.

Highlights d

4E-T is a component of the mRNA decapping and decay machinery

d

4E-T promotes the decay of mRNAs targeted by the CCR4NOT deadenylase complex

d

4E-T-mediated mRNA decay requires that it bind to eIF4E

Nishimura et al., 2015, Cell Reports 11, 1425–1436 June 9, 2015 ª2015 The Authors http://dx.doi.org/10.1016/j.celrep.2015.04.065

Cell Reports

Article The eIF4E-Binding Protein 4E-T Is a Component of the mRNA Decay Machinery that Bridges the 50 and 30 Termini of Target mRNAs Tamiko Nishimura,1,2 Zoya Padamsi,1,2 Hana Fakim,1,2 Simon Milette,1,2 Wade H. Dunham,3,4 Anne-Claude Gingras,3,4 and Marc R. Fabian1,2,* 1Department

of Oncology, McGill University, Montreal, QC H3A 0G4, Canada Cancer Centre, Jewish General Hospital, Lady Davis Institute for Medical Research, Montreal, QC H3T 1E2, Canada 3Lunenfeld-Tanenbaum Research Institute, Mount Sinai Hospital, 600 University Avenue, Toronto, ON M5G 1X5, Canada 4Department of Molecular Genetics, University of Toronto, Toronto, ON M5S 1A8, Canada *Correspondence: [email protected] http://dx.doi.org/10.1016/j.celrep.2015.04.065 This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). 2Segal

SUMMARY

Eukaryotic mRNA degradation often initiates with the recruitment of the CCR4-NOT deadenylase complex and decay factors to the mRNA 30 terminus. How the 30 -proximal decay machinery interacts with the 50 -terminal cap structure in order to engender mRNA decapping and 50 –30 degradation is unclear. Human 4E-T is an eIF4E-binding protein that has been reported to promote mRNA decay, albeit via an unknown mechanism. Here, we show that 4E-T is a component of the mRNA decay machinery and interacts with factors including DDX6, LSM14, and the LSM1-7-PAT1 complex. We also provide evidence that 4E-T associates with, and enhances the decay of, mRNAs targeted by the CCR4-NOT deadenylase complex, including microRNA targets. Importantly, we demonstrate that 4E-T must interact with eIF4E to engender mRNA decay. Taken together, our data support a model where 4E-T promotes mRNA turnover by physically linking the 30 -terminal mRNA decay machinery to the 50 cap via its interaction with eIF4E. INTRODUCTION All nuclear transcribed eukaryotic mRNAs contain at their 50 end the structure m7GpppN (where N is any nucleotide) termed the ‘‘cap’’ and, with the exception of histone mRNAs, a 30 poly(A) tail (Hershey et al., 2012). Recognition of these cis-acting elements is mediated by the poly(A)-binding protein PABP and by eIF4F, a three-subunit complex composed of (1) eIF4E, a 24kDa polypeptide that specifically interacts with the 50 -cap structure; (2) eIF4A, a DEAD-box RNA helicase; and (3) eIF4G, a large scaffolding protein that binds to eIF4E, eIF4A, and PABP (Sachs and Davis, 1989; Tarun et al., 1997), but also with the multi-subunit factor eIF3, which in turn recruits the ribosome to the mRNA. The interaction between eIF4G and PABP serves to bridge the 50

and 30 regions of the mRNA, thus circularizing the transcript (Wells et al., 1998). mRNA degradation is a key process that post-transcriptionally regulates gene expression. One of the major mRNA degradation pathways initiates with the recruitment of the CCR4-NOT deadenylase complex to the 30 UTRs of mRNAs targeted by microRNAs (miRNAs) and by a number of RNA-binding proteins, including tristetraprolin (TTP). Once recruited, the CCR4-NOT displaces PABP and converts the poly(A) tail into an oligoadenylated form (Zekri et al., 2013). This acts as a docking platform for the mRNA decapping machinery, including the LSM1-7 complex, and PATL1, a scaffold for the DCP1-DCP2 decapping enzyme complex and the 50 –30 exonuclease XRN1 (Chowdhury and Tharun, 2009). Prior to decapping, eIF4G is displaced from the mRNA via an unknown mechanism of action (Zekri et al., 2013). Moreover, we currently do not understand how the mRNA decapping machinery, which is recruited to the mRNA 30 terminus, interfaces with the eIF4E-bound 50 -terminal cap structure. Many proteins that are involved in mRNA decay, such as CCR4-NOT deadenylase components, DCP2, PATL1, EDC3, DDX6, and LSM14, are enriched in dynamic cytoplasmic foci called processing bodies (P-bodies) (Kedersha and Anderson, 2009; Sheth and Parker, 2006). A key protein that promotes P-body formation is the eIF4E-transporter protein, 4E-T (Ferraiuolo et al., 2005). Human 4E-T is a member of the eIF4E-binding protein family, which includes proteins such as eIF4G and the small 4E-BPs, and contains an evolutionarily conserved canonical eIF4E-binding motif (4E-BM [YXXXLF], where X is any amino acid and F is hydrophobic) located close to its N terminus. 4E-T has been linked to translational repression and mRNA decay, albeit via unknown mechanism(s) of action (Cargnello et al., 2012; Dostie et al., 2000; Ferraiuolo et al., 2005; Kamenska et al., 2014). Here, we explored the function of 4E-T in post-transcriptional gene silencing. Using both affinity purification and proximal biotinylation (BioID), as well as purified recombinant proteins, we demonstrate that human 4E-T is a component of the mRNA decapping machinery. In addition to binding eIF4E, 4E-T interacts with multiple mRNA decapping factors (DDX6, PATL1, LSM14, Cell Reports 11, 1425–1436, June 9, 2015 ª2015 The Authors 1425

Figure 1. 4E-T Is a Component of the mRNA Decay Machinery (A) Schematic representation of the wild-type 4E-T, an eIF4E-binding mutant (4E-TD4E-BM), and the N-terminal 100 aa of 4E-T (4E-T1–100). (B) Schematic representation of the BioID method (adapted from Roux et al., 2012). Expression of a BirA fusion protein in vivo and subsequent addition of biotin to cell culture leads to the selective biotinylation of proximal proteins. Biotinylated proteins are purified under harsh conditions using streptavidin resin, resolved by SDS-PAGE, and analyzed by western blotting. (C) Western blot analysis of lysates derived from HEK293 cells expressing either FLAG-BirA or FLAG-BirA-4E-T and probed with anti-FLAG or b-actin antibodies. (D) Streptavidin pull-downs of biotinylated proteins from benzonase-treated lysates outlined in (C). Precipitated proteins were subjected to SDS-PAGE and probed with antibodies against the indicated proteins. (E) Western blot analysis of lysates derived from HeLa expressing FLAG-4E-T, FLAG-4E-TD4E-BM, or FLAG-4E-T1–100 and probed with anti-FLAG or b-actin antibodies. (F) Immunoprecipitation (IP) of transiently transfected FLAG-4E-T proteins from benzonasetreated HeLa cell extracts using anti-FLAG antibody. Immunoprecipitated complexes were separated by SDS-PAGE and probed with antibodies against the indicated proteins.

together, these observations indicate that 4E-T is a central component of the mRNA decapping machinery and suggest that it facilitates mRNA decay by physically linking 30 -terminal decapping factors to the mRNA 50 cap via its interaction with eIF4E. RESULTS

and LSM2) via a network of protein-protein interactions. For example, our data establish that 4E-T directly binds the DDX6 C-terminal RecA domain. We also show that the N-terminal Sm-fold of LSM14 directly contacts two conserved regions of 4E-T. 4E-T associates with, and contributes to the decay of mRNAs targeted by both TTP and the miRNA-induced silencing complex (miRISC). Importantly, the ability of 4E-T to promote target mRNA decay requires that it contact eIF4E. Taken 1426 Cell Reports 11, 1425–1436, June 9, 2015 ª2015 The Authors

4E-T Interacts with mRNA Decapping and Decay Factors 4E-T is a large (985-aa) protein with no characterized regions, except for the short eIF4E-binding motif at its N terminus, and nuclear localization and export signals (Figure 1A) (Dostie et al., 2000). 4E-T was originally characterized as a shuttling protein that imports eIF4E to the nucleus and recruits eIF4E to cytoplasmic P-bodies (Dostie et al., 2000; Ferraiuolo et al., 2005). 4E-T was also reported to promote mRNA decay (Ferraiuolo et al., 2005), although its mechanism of action remains unknown. We utilized conventional affinity purification and proximal biotinylation (BioID) technologies to investigate whether 4E-T interacts with any mRNA decay factors that localize to P-bodies. BioID utilizes a variant of the E. coli biotin ligase BirA (R118G), which is fused to a protein of

interest and covalently biotinylates proximal proteins in living cells (Figure 1B) (Couzens et al., 2013; Kim et al., 2014; Roux et al., 2013; Roux et al., 2012). BioID experiments therefore allow one to generate a history of proteins that proximally associated with the ‘‘bait’’ protein fused to BirA in vivo. HEK293 cells expressing either FLAG-BirA or FLAG-BirA-4E-T proteins were incubated with biotin for 24 hr and lysed under harsh conditions to disrupt protein-protein interactions (Figure 1C). Biotinylated proteins were purified from benzonase-treated lysates with streptavidin agarose, eluted, and analyzed by western blotting with antibodies against a number of P-body proteins, including deadenylation and decapping factors, as well as eIF4E (Figure 1D). This method was validated by the ability of FLAGBirA-4E-T to biotinylate eIF4E, a protein that directly binds 4E-T (Dostie et al., 2000), but not PABP or actin. FLAG-BirA4E-T also biotinylated CNOT1, the central subunit of the CCR4-NOT complex, as well as the decapping factors PATL1, LSM14, and DDX6 (Figure 1D). In contrast, FLAG-BirA, which was expressed at 10-fold higher levels than FLAG-BirA-4E-T, biotinylated these proteins to a much lesser extent. We next carried out FLAG-4E-T co-immunoprecipitation experiments to ascertain whether these 4E-T-proximal proteins physically interact with 4E-T. Plasmids encoding wild-type FLAG-tagged 4E-T, a 4E-T mutant that lacks the canonical eIF4E-binding motif (4E-BM) (30YTKEELL36 [4E-TD4E-BM]), or an N-terminal 4E-T fragment that contains the 4E-BM (4E-T1–100) were transiently transfected into HeLa cells (Figures 1A and 1E), and FLAG-4E-T proteins were immunoprecipitated from benzonase-treated lysates with FLAG antibody. Co-immunoprecipitating proteins were resolved by SDS-PAGE and analyzed by western blotting (Figure 1F). FLAG-4E-T co-precipitated PATL1, LSM14, DDX6, and eIF4E, but not actin. In contrast, FLAG-4E-TD4E-BM failed to interact with eIF4E, as expected. However, it could still interact with PATL1, LSM14, and DDX6 as efficiently as wild-type FLAG-4E-T. Conversely, FLAG-4E-T1–100, which maintains the eIF4E binding motif but lacks all downstream sequence, coprecipitated eIF4E but failed to interact with PATL1, LSM14, or DDX6. Interestingly, while FLAG-BirA-4E-T biotinylated CNOT1 in vivo (Figure 1D), FLAG-4E-T could not co-precipitate CNOT1 (Figure 1F). Taken together, these data demonstrate that 4E-T physically interacts with the mRNA decapping and decay machinery in a manner that is independent of its binding to eIF4E. They also suggest that 4E-T proximally associates with the CCR4-NOT deadenylase complex; however, this association is transient and/or not as stable as its interaction with decapping factors. DDX6 Directly Binds the 4E-T Cup Homology Domain Comparing human 4E-T to homologs of several insect species identified multiple conserved stretches of amino acids (Figure S1). Based on BioID and co-precipitation data (Figure 1), we hypothesized that 4E-T, through unidentified sequence(s), may directly bind decapping machinery components. One potential candidate was the decapping activator DDX6 (also known as RCK/p54, Me31B, or Dhh1), which can directly bind CNOT1, as well as the decapping activators PATL1, EDC3, and LSM14, and was identified alongside 4E-T as a 50 -cap-associated protein (Cargnello et al., 2012; Chen et al., 2014; Mathys et al., 2014;

Rouya et al., 2014; Tritschler et al., 2009). To test this hypothesis we purified a series of glutathione-S-transferase (GST)-fused 4ET fragments (Figure 2A; we were unable to express and purify the full-length protein) and incubated these fragments with purified maltose binding protein (MBP)-tagged DDX6. A 4E-T fragment encompassing residues 105–263 (GST-4E-T105–263) efficiently bound MBP-DDX6, whereas all other 4E-T fragments did not (Figure 2B). The 4E-T fragment that precipitated DDX6 (GST-4E-T105– 263 ) contains the Cup homology domain (CHD), an evolutionarily conserved patch of amino acids (219–240) that is present in 4ET proteins and in the Drosophila protein Cup, a translational repressor that also binds to eIF4E (Figures S1 and 2C) (Wilhelm et al., 2003). The function of the CHD in both 4E-T and Cup has not been established. The CHD sequence was necessary for the interaction between 4E-T and DDX6 as a GST-4E-T105–263 fragment lacking the CHD (GST-4E-TDCHD; Figure 2A) failed to pull down MBP-DDX6 (Figure 2B). Moreover, the CHD sequence was also sufficient to mediate this interaction, as a GST-4E-T fragment containing only the CHD (GST-4E-TCHD; Figure 2A) efficiently bound MBP-DDX6 (Figure 2B). DDX6 is a member of the DEAD-box family of proteins and maintains a bi-lobal architecture of N- and C-terminal RecA domains (Figure 2D) (Henn et al., 2012). To determine which part of DDX6 was involved in this interaction, we incubated GST-4E-TCHD with either the DDX6 N-terminal (MBP-DDX61–302) or C-terminal (MBP-DDX6303–483) lobes (Figure 2D). GST-4E-TCHD could not precipitate MBP-DDX61–302, whereas it efficiently pulled down MBP-DDX6303–483 (Figure 2E). Taken together, the above results demonstrate that the 4E-T CHD is a DDX6-interaction motif. 4E-T Middle and C-Terminal Regions Directly Bind LSM14 Based on 4E-T BioID and co-precipitation experiments, we surmised that LSM14, which was also identified as a 50 -cap-associated protein (Cargnello et al., 2012), may also directly bind 4E-T. To this end, GST-fused 4E-T fragments were incubated with purified MBP-tagged LSM14 (Figures 3A and 3B). In contrast to MBP-DDX6, MBP-LSM14 did not precipitate with GST-4ET105–263 but rather was pulled down with both the middle and C-terminal GST-4E-T fragments (GST-4E-T255–675 and GST-4E-T675–985, respectively; Figure 3C). We next assessed how the middle and C-terminal regions of 4E-T interact with LSM14. Our comparative sequence analysis identified multiple conserved stretches of amino acids downstream of the CHD (Figure S1), several of which we hypothesized may bind LSM14. MBP-LSM14 bound a 4E-T C-terminal mutant lacking residues 694–704 (GST-4E-T675–985-MUT1; Figures 3A and 3C). In contrast, a 4E-T C-terminal fragment lacking amino acids 948–985 (GST-4E-T675–985-MUT2) failed to bind MBP-LSM14 (Figures 3A and 3C). Based on these results, we conclude that the C-terminal 37 aa of 4E-T form an LSM14-interaction motif. The middle region of 4E-T contains multiple conserved stretches of amino acids spanning residues 335–443 (Figure S1). This region was both necessary and sufficient for mediating the interaction: MBP-LSM14 failed to bind a 4E-T fragment lacking these conserved amino acids (GST-4E-T255–675-MUT1), while a 4E-T fragment containing only these elements (GST-4E-T335–490) efficiently bound MBP-LSM14 (Figures 3A and 3D). Cell Reports 11, 1425–1436, June 9, 2015 ª2015 The Authors 1427

Figure 2. Human 4E-T Cup Domain Directly Binds DDX6

Homolog

(A) Schematic diagram of full-length 4E-T and 4E-T fragments used in co-precipitation experiments in (B) and (E). Dashed lines indicated the region required for DDX6 binding in (B). (B) Coomassie-stained SDS-PAGE gel of input (left lane) and GST pull-down of GST alone or fused to 4E-T fragments immobilized on glutathione-agarose beads incubated with MBP-tagged full-length DDX6. MW, molecular weight. (C) Sequence alignment of conserved amino acids within the 4E-T Cup homolog domain that binds DDX6 of human (Hs), D. melanogaster (Dm), and C. elegans (Ce) 4E-T and CUP proteins. (D) Schematic diagram of human DDX6 protein fragments used in co-precipitation experiments. Conserved elements are indicated within diagram. DEAD/H, DEAD/DEAH-box helicase domain; N-term domain; HELIC, helicase C-terminal domain. Dashed lines indicated the region required for 4E-T binding in (E). (E) Coomassie-stained SDS-PAGE gel of input (left lane) and GST pull-down of GST-4E-TCHD immobilized on glutathione-agarose beads incubated with MBP-tagged full-length, N-terminal or C-terminal DDX6 fragments. MW, molecular weight.

How does LSM14 contact 4E-T? LSM14 has an N-terminal Sm-like motif (LSM), as well as a central region, termed an FDF motif, that binds DDX6 (Figure 3B) (Marnef et al., 2009). To determine how LSM14 contacts 4E-T, we incubated GST-4E-T MID and CTD fragments with MBP fused to the LSM14 N-terminal domain (NTD), middle domain (MID), or C-terminal domain (CTD). Strikingly, both the GST-4E-T255–675 and GST-4ET675–985 proteins pulled down MBP-LSM14NTD, whereas neither bound MBP-LSM14MID or MBP-LSM14CTD (Figure 3E). Thus, the LSM14 N-terminal Sm domain directly interacts with both the 4E-T middle and C-terminal regions. A Network of Protein-Protein Interactions Connect 4E-T with the mRNA Decapping Machinery Our data show that DDX6 can directly bind 4E-T via the CHD. However, as 4E-T also interacts with LSM14 and PATL1, proteins that can directly contact DDX6 (Tritschler et al., 2009), 4E-T may also indirectly interact with DDX6 via these latter proteins. To test this, we immunoprecipitated FLAG-4E-T or FLAG-4E-T mutants lacking the CHD and/or the regions that bind LSM14 (FLAG-4E-TDCHD, FLAG-4E-TDLSM14, and FLAG4E-TDCHD+LSM14, respectively; Figure 4). Unexpectedly, FLAG1428 Cell Reports 11, 1425–1436, June 9, 2015 ª2015 The Authors

4E-TDCHD co-precipitated DDX6, PATL1, and LSM14 as well as wild-type FLAG4E-T, even though it lacks the CHD. Both FLAG-4E-TDLSM14 and FLAG-4ETDCHD+LSM14 failed to pull down LSM14, as expected, yet could still interact with PATL1. However, neither of these 4E-T mutants co-precipitated DDX6 as well as wild-type FLAG-4E-T. Interestingly, a 4E-T protein lacking its N-terminal 220 aa (FLAG-4E-T220–985), which lacks both the 4E-BM and an additional stretch of conserved amino acids (Figure S1), co-precipitated LSM14 but failed to pull down PATL1 and did not interact with DDX6 as well as wild-type 4E-T. PATL1 has been previously shown to interact with the decapping enzyme subunit DCP1 and the LSM1-7 complex (Ozgur et al., 2010). Wild-type 4E-T also pulled down both DCP1 and the LSM2 subunit of the LSM1-7 complex, whereas FLAG-4E-T220–985 did not interact with these proteins. Thus, these data suggest that LSM14 and PATL1 contribute to the stable association of DDX6 with 4E-T in vivo. In addition, these results indicate that 4E-T interacts with PATL1, DCP1, and the LSM1-7 complex via a stretch of amino acids between the 4E-BM and the CHD. 4E-T Associates with the AU-Rich Element Binding Protein TTP Our results demonstrate that 4E-T (1) interacts with PATL1, DDX6, LSM14, DCP1, and LSM2 and (2) proximally associates with the CNOT1 subunit of the CCR4-NOT complex. The CCR4-NOT complex is recruited to mRNAs via a number of gene-silencing proteins, including the AU-rich element binding

Figure 3. Human 4E-T Directly Binds Sm Domain of LSM14 (A) Schematic diagram of full-length 4E-T and 4E-T fragments used in co-precipitation experiments in (C)–(E). Dashed lines indicated the region that contacts LSM14 (C and D). (B) Schematic diagram of human LSM14, and LSM14 protein fragments used in co-precipitation experiments. Conserved elements are indicated within diagram. Dashed lines indicated the region required for 4E-T binding in (E). (C and D) Coomassie-stained SDS-PAGE gel of input (left lane) and GST pull-down of GST alone or fused to 4E-T fragments immobilized on glutathione-agarose beads incubated with MBP-tagged full-length LSM14. MW, molecular weight. (E) Coomassie-stained SDS-PAGE gel of input (four left lanes) and GST pull-down of GST-4E-T fragments immobilized on glutathione-agarose beads incubated with MBP-tagged full-length, N-terminal, middle (MID), or C-terminal (C-term) LSM14 fragments. MW, molecular weight.

Cell Reports 11, 1425–1436, June 9, 2015 ª2015 The Authors 1429

lates known interactions of TTP with deadenylation and decapping components. Importantly, it also shows that TTP associates with 4E-T via its N- and C-terminal activation domains that it utilizes to silence target mRNAs.

Figure 4. 4E-T Interacts with mRNA Decapping Machinery via a Network of Protein-Protein Interactions (A) Schematic representation of the FLAG-tagged wild-type 4E-T and 4E-T deletion mutants used in co-immunoprecipitation experiments. Dashed lines indicate the region required for 4E-T to interact with PATL1, DCP1, and LSM2. (B) Immunoprecipitation (IP) of transiently transfected FLAG-4E-T proteins from benzonase-treated HeLa cell extracts using anti-FLAG antibody. Immunoprecipitated complexes were separated by SDS-PAGE and probed with antibodies against the indicated proteins.

protein TTP and the miRISC (Chen et al., 2014; Fabian and Sonenberg, 2012; Leppek et al., 2013; Mathys et al., 2014; Rouya et al., 2014; Sandler et al., 2011; Tan et al., 2014; Van Etten et al., 2012). Based on these data, we wished to utilize BioID to ascertain if TTP proximally associates with 4E-T. HeLa cells were transfected with plasmids coding for myc-tagged BirA alone, myc-BirA-TTP, or myc-BirA fused to the TTP RNA-binding domain (myc-BirA-RBD), the latter protein being able to bind to, but not deadenylate, target mRNAs (Figures 5A and 5B) (LykkeAndersen and Wagner, 2005). myc-BirA-TTP biotinylated CNOT1, which directly binds TTP (Fabian et al., 2013) but not actin. Myc-BirA-TTP also biotinylated PATL1, DDX6, LSM14, and 4E-T. Notably, CNOT1, 4E-T, PATL1, LSM14, and DDX6 were not biotinylated in cells expressing myc-BirA (Figure 6, lane 2). Moreover, myc-BirA-RBD, which lacks the TTP N- and C-terminal activation domains that TTP utilizes to deadenylate target mRNAs (Lykke-Andersen and Wagner, 2005; Sandler et al., 2011), did not biotinylate these factors nearly as well as myc-BirA-TTP, even though it was expressed at over 3.5-fold higher levels (Figures 5B and 5C). In summary, BioID recapitu1430 Cell Reports 11, 1425–1436, June 9, 2015 ª2015 The Authors

4E-T Contributes to the Turnover of a TTP-Targeted mRNA As 4E-T binds the mRNA decapping machinery and associates with TTP and the CCR4-NOT complex, we surmised that 4E-T might play an active role in promoting the decay of mRNAs targeted by TTP. To this end, we depleted 4E-T and assessed if this impacted the stability of TTP-targeted mRNAs. We used the lN-BoxB tethering system (Baron-Benhamou et al., 2004) in HeLa cells to assess mRNA decay mediated by TTP. This was accomplished by fusing TTP or LacZ (control) to hemagglutinin (HA) tags and the bacteriophage lN-peptide, which recognizes and binds to the two BoxB hairpins in the 30 UTR of a Renilla-luciferase-encoding reporter mRNA (RL-2BoxB; Figures 6A and 6B). HeLa cells were transfected with small interfering RNAs (siRNAs) against GFP (control), 4E-T, or the decapping factor PATL1 (Figure 6C) and re-transfected 48 hr later with plasmids encoding a lNHA-tagged protein and the RL-2BoxB reporter. RL-2BoxB mRNA stability was assessed in control and 4E-T-knockdown cells by treating transfected cells with actinomycin D (Act-D) to inhibit de novo transcription. Importantly, knocking down 4E-T did not perturb the stability of its binding partners (Figure S2A). Total RNA was isolated at specific time points following Act-D treatment, and RL-2BoxB mRNA levels were quantified via random hexamer-primed qRT-PCR. Tethering lN-HA-LacZ to RL-2BoxB RNA did not impact its stability over time (data not shown), whereas tethering lN-HA-TTP markedly destabilized RL-2BoxB mRNA (Figure 6D). Interestingly, while a TTP-targeted RL-2BoxB mRNA decayed with half-life of 2.3 hr in control-depleted cells, it was significantly stabilized in 4E-Tknockdown cells (6 hr half-life). Importantly, the reporter mRNA displayed near-identical decay kinetics in cells depleted of the decapping factor PATL1. In addition, defects in TTPmediated mRNA decay in 4E-T-depleted cells were abolished in HeLa cells rescued with siRNA-resistant wild-type FLAG4E-T (Figures 6D and S2B). We next examined the poly(A) tail length of the TTP-targeted reporter in control and 4E-T-knockdown cells over Act-D time courses to ascertain whether 4E-T promotes mRNA deadenylation. This was accomplished using an extension poly(A) test (ePAT) procedure (Ja¨nicke et al., 2012) followed by semiquantitative PCR analysis (Figure S3). RL-2BoxB RNA maintained a short (20 nt) poly(A) tail in control cells, as compared to GAPDH mRNA that maintained longer (50–100 nt) poly(A) tails. However, knocking down 4E-T did not increase the length of the poly(A) tail but rather stabilized the reporter mRNA. These data therefore suggest that 4E-T impacts mRNA decapping and/or 50 -30 mRNA decay rather than mRNA deadenylation. 4E-T/eIF4E Interaction Promotes the Decay of mRNAs Targeted by the CCR4-NOT Complex In stark contrast to HeLa cells rescued with wild-type 4E-T, a 4E-T mutant lacking its canonical eIF4E-binding motif (30YTKEELL36 [4E-TD4E-BM]), which interacts with mRNA decay factors, but not

Figure 5. 4E-T Proximally Associates with the AU-Rich Element-Binding Protein TTP (A) Schematic representation of the myc-tagged BirA enzyme alone, fused to full-length TTP containing N-terminal and C-terminal domains (NTD and CTD), or fused to the TTP RNA-binding domain (RBD). (B) Western blot analysis of lysates derived from HeLa cells transfected with myc-BirA, myc-BirA-TTP, or myc-BirA-TTP-RBD and probed with anti-myc or b-actin antibodies. (C) Streptavidin pull-downs of biotinylated proteins from benzonase-treated lysates outlined in (B). Precipitated proteins were subjected to SDS-PAGE and probed with antibodies against the indicated proteins.

eIF4E (Figure 1F), failed to stimulate TTP-dependent mRNA decay (Figures 6E and S2B). Importantly, myc-BirA-TTP biotinylates 4ETD4E-BM equally well as compared to wild-type 4E-T (Figures S4A and S4B), suggesting that 4E-TD4E-BM proximally associates with TTP even though it cannot promote TTP-mediated mRNA decay. We also wished to determine if 4E-T promotes the decay of cellular mRNAs targeted by the CCR4-NOT complex. To this end, we depleted 4E-T and assessed if this impacted the stability of the HMGA2 mRNA, a bona fide let-7 miRNA target. The HMGA2 mRNA contains multiple let-7 miRNA target sites in its 30 -UTR, and knocking down 4E-T has been reported to increase HMGA2 protein levels (Kamenska et al., 2014; Lee and Dutta, 2007; Mayr et al., 2007). Importantly, inhibiting the let-7 miRNA with antisense oligonucleotides has been reported to stabilize the HMGA2 mRNA steady-state levels 3.7-fold (Clancy et al., 2011). HMGA2 mRNA was destabilized by 60% 6 hr post-Act-D treatment in control-depleted cells (Figure 6F). However, HMGA2 mRNA remained stable in 4E-T-knockdown cells, even after transcription was shut down for 6 hr. HMGA2 mRNA decay was rescued in cells complemented with wild-type FLAG-4E-T (Figure 6F), but not in cells expressing 4E-TD4E-BM (Figure 6G). These data demonstrate that 4E-T must be able to bind eIF4E in order to promote the turnover of both TTP- and miRNA-targeted mRNAs. DISCUSSION A Network of Interactions Link 4E-T with the mRNA Decay Machinery Several observations presented in this work suggest that 4E-T functions as a component of the mRNA decay machinery, and

is recruited to mRNAs targeted by the CCR4-NOT deadenylase complex. First, 4E-T co-immunoprecipitates PATL1, LSM14, DDX6, DCP1, and LSM2 in a manner that is independent of its binding to eIF4E (Figures 1 and 4). Second, 4E-T BioID experiments demonstrate that it proximally associates with the CNOT1 subunit of the CCR4-NOT complex (Figure 1). Third, BioID experiments also show that the AU-rich element-binding protein TTP, which recruits the CCR4-NOT complex to target mRNAs, associates with 4E-T via it N- and C-terminal activation domains (Figure 5). Finally, knocking down 4E-T stabilizes both TTP- and miRNA-targeted mRNAs without impacting their poly(A) tail lengths (Figures 6 and S3, respectively). Interestingly, while 4E-T BioID data demonstrate that 4E-T proximally associates with CNOT1, it does not co-precipitate this deadenylase subunit from nuclease-treated lysates. Taken together, these data suggest that the association of 4E-T with the CCR4-NOT complex is indirect and/or is a transient interaction that occurs when 4E-T is recruited to target mRNAs in complex with PATL1, such as those bound by TTP or by the miRISC. DDX6 interacts with the CNOT1 subunit of the CCR4-NOT complex, as well as multiple decapping factors (PATL1, LSM14, and EDC3) (Chen et al., 2014; Mathys et al., 2014; Rouya et al., 2014; Tritschler et al., 2009). In this report, we establish 4E-T as a novel DDX6-interacting protein. Our data demonstrate that the 4E-T CHD is a platform that directly binds to the C-terminal lobe of DDX6 (Figure 2). As this sequence is also conserved in the Drosophila translational repressor protein Cup (Wilhelm et al., 2003), it is likely that Cup also binds the Drosophila DDX6 homolog, Me31b, in a similar manner. Notwithstanding this direct contact point, Cell Reports 11, 1425–1436, June 9, 2015 ª2015 The Authors 1431

Figure 6. eIF4E/4E-T Interaction Contributes to the Decay of miRNA- and TTP-Targeted mRNAs (A) Schematic depiction of lNHA-TTP and -LacZ tethered to RL-2BoxB reporter mRNA. (B) Western blot analysis of lysates derived from HeLa cells transfected with lNHA-LacZ or -TTP and probed with HA or anti-b-actin antibodies. (C) Western blot analysis of lysates derived from HeLa cells transfected with siRNAs targeting GFP, PATL1, or 4E-T and probed with antibodies against 4E-T and PATL1. (D and E) HeLa cells transfected with siRNA targeting GFP, PATL1, or 4E-T or HeLa cell lines stably expressing siRNA-resistant wild-type 4E-T or 4E-TD4E-BM were subsequently cotransfected with plasmids encoding lNHA-tagged proteins (B), along with RL-2BoxB. 24 hr after transfections, the stability of RL-2BoxB mRNA was assessed by using actinomycin D (5 mg/ml) for the indicated amount of time. Total RNA was isolated, reverse transcribed with random hexamer oligonucleotide primers, and RL-2BoxB RNA was quantified by qPCR. RL-2BoxB mRNA decay rates were normalized to GAPDH mRNA levels with the zero time point set at 1. Error bars represent the SEM of multiple independent experiments. (F and G) HMGA2 mRNA stability from Act-D time course analysis. HMGA2 mRNA decay rates were normalized to GAPDH mRNA levels with the zero time point set at 1. Error bars represent the SEM of multiple independent experiments.

co-immunoprecipitation experiments demonstrate that the CHD is not critical for 4E-T to stably interact with DDX6. Instead, DDX6 is recruited to 4E-T in vivo via its interactions with both LSM14 and PATL1, proteins that also directly contact the DDX6 C-terminal RecA domain. Thus, the CHD serves as but one of multiple ways by which 4E-T interacts with DDX6. The function of the CHD/DDX6 interaction therefore remains unclear. One possibility is that the direct interaction between the CHD and DDX6 may serve as a secondary binding site upon 4E-T interacting with DDX6 via LSM14 and PATL1. We also show that 4E-T directly contacts with LSM14. This is accomplished via two non-contiguous regions of 4E-T that contact the N-terminal Sm motif of LSM14 (Figure 3). Interestingly, the N-terminal Sm domain of LSM14 is structurally similar to that of EDC3/LSM16, a decapping enhancer protein that also contains a DDX6-interacting FDF motif (Figure S5A) (Fromm et al., 2012). Indeed, GST-tagged DDX6 could bind both MBP-LSM14 and MBP-EDC3 (Figure S5B). However, in contrast to MBP-LSM14, MBP-EDC3 failed to co-precipitate with any GST-4E-T fragment (Figure S5C). These observations suggest that, while structurally similar, the Sm domains of LSM14 and EDC3 are functionally divergent. How the LSM14 Sm fold structurally contacts 4E-T remains to be established. 1432 Cell Reports 11, 1425–1436, June 9, 2015 ª2015 The Authors

4E-T/eIF4E Complex Enhances the Decay of mRNAs Targeted by the CCR4-NOT Complex Previous studies have generated conflicting reports with respect to the function of 4E-T. Cells depleted of 4E-T partially stabilize a beta globin reporter mRNA containing the Granulocyte macrophage colony-stimulating factor (GMCSF) 30 UTR (Ferraiuolo et al., 2005) via an unknown mechanism. In contrast, artificially tethering 4E-T to a reporter mRNA represses mRNA translation while not impacting mRNA stability (Kamenska et al., 2014). In this work, we show using complementation assays that 4E-T contributes to the decay of mRNAs targeted by the CCR4-NOT complex. This conclusion is supported by the fact that knocking down 4E-T markedly stabilizes mRNAs targeted by TTP or by the let-7 miRNA, two gene-silencing mechanisms that recruit the CCR4-NOT complex. Importantly, our data also indicate that 4E-T must contact eIF4E in order to promote mRNA decay, as a 4E-T protein lacking the canonical 4E-BM cannot. Interestingly, the mRNA decapping/decay machinery is also recruited to histone mRNAs, which lack poly(A) tails (Mullen and Marzluff, 2008). Whether 4E-T also plays a role in degrading histone mRNAs is not known. In stark contrast to 4E-T, the Drosophila eIF4E-binding protein Cup has been reported to stabilize a target mRNA by binding to eIF4E and preventing mRNA decapping (Igreja and Izaurralde,

Figure 7. Model for a Role of 4E-T in Mediating Decapping of a mRNA Targeted by the CCR4-NOT Complex (A) mRNA circularization during translation initiation. eIF4G promotes mRNA circularization by interacting with eIF4E and PABP, proteins that are bound to the 50 cap and 30 poly(A) tail, respectively. (B) The miRISC interacts with a miRNA target site and recruits the CCR4-NOT complex to displace PABP and convert the poly(A) tail into an oligoadenylated sequence. (C) The mRNA decapping machinery, including 4E-T, is recruited to the oligoadenylated 30 terminus and the CCR4-NOT. 4E-T-mediated circularization, via its interaction with eIF4E, brings decapping (DCP1 and DCP2) factors and the XRN1 exonuclease into proximity with the 50 cap to initiate its hydrolysis and mRNA decay.

2011). It will therefore be interesting to determine why one eIF4Ebinding protein promotes mRNA decay whereas the other prevents it. The eIF4E/4E-T complex has recently been reported to regulate neurogenesis by repressing key gene expression programs in neuronal precursors (Yang et al., 2014). 4ET is also a component of cytoplasmic polyadenylation-binding protein (CPEB) repression complex in Xenopus oocytes, which silences maternal mRNAs during oocyte maturation (Minshall et al., 2007; Standart and Minshall, 2008). Interestingly, the CPEB complex also includes Pat1a, LSM14, DDX6, and the ovary-specific cap-binding protein eIF4E1b. Whether 4E-T plays a role in other biological processes remains to be determined. An Integrated Model for 4E-T Promoting mRNA Turnover The data presented in this work, and additional results from other groups, allow us to develop an integrated model for 4E-T gene

silencing engendered by the CCR4-NOT complex (Figure 7). In this model, the CCR4-NOT complex is recruited to the 30 UTR of a target mRNA by the miRISC, TTP, or other 30 UTR-binding proteins. The CCR4-NOT complex proceeds to mediate PABP displacement from the 30 poly(A) tail (Zekri et al., 2013) and convert it into an oligoadenylated form. In combination with the CCR4-NOT complex, this oligoadenylated tail serves as a binding platform for the LSM1-7/PATL1 complex (Chowdhury and Tharun, 2009), which interacts with 4E-T, DDX6, XRN1, and other decapping factors. How might the 4E-T/ eIF4E interaction function to enhance mRNA decay? It is conceivable that 4E-T may juxtapose the 30 -terminal mRNA decay machinery with the mRNA 50 terminal cap structure via its interaction with eIF4E. This would, in turn, increase the local concentration of DCP2 with the cap, thereby allowing it to compete with eIF4E for binding and promote eIF4E dissociation. It has been reported that, in addition to PABP, eIF4G dissociates from a miRISC-silenced reporter mRNA following its deadenylation but prior to mRNA decapping (Zekri et al., 2013). However, in contrast to PABP, eIF4G does not dissociate from a reporter if the poly(A) tail remains intact, suggesting that PABP dissociation alone does not lead to eIF4G displacement. 4E-T and eIF4G both directly and mutually exclusively bind to the dorsal surface of eIF4E through a conserved canonical 4E-BM (Igreja et al., 2014). However, unlike eIF4G, 4E-T also interacts with the lateral surface of eIF4E via its secondary non-canonical 4E-BM (Igreja et al., 2014). Moreover, this secondary 4E-BM, in combination with the canonical 4E-BM, is critical for a 4E-T peptide to compete with eIF4G for binding to eIF4E in vitro (Igreja et al., 2014). We envision that recruitment of 4E-T to the 30 terminus of deadenylated mRNAs promotes its binding to the lateral Cell Reports 11, 1425–1436, June 9, 2015 ª2015 The Authors 1433

surface of 50 -cap-bound eIF4E, which is still in complex with eIF4G. This would allow for 4E-T to gain a foothold on eIF4E even as the dorsal surface of eIF4E is occupied by eIF4G. Subsequent to this initial binding, 4E-T would then use its canonical 4E-BM to compete with eIF4G for the dorsal surface of eIF4E and promote eIF4G dissociation. Although intriguing, it is important to note that these possibilities await future experimental validation. EXPERIMENTAL PROCEDURES Antibodies Antibodies against PATL1, DCP1, and DDX6 were from Bethyl Laboratories. LSM14 and LSM2 antibodies were from Millipore. CNOT1 antibody was from Proteintech. 4E-T antibody was from Abcam. b-actin and eIF4E antibodies were from Cell Signaling. FLAG, HA, and MYC antibodies were from Sigma-Aldrich, Covance, and Bioshop, respectively. DNA Constructs pcDNA3.1-myc-BirA(R118G) was obtained from Addgene. Full-length TTP and TTP RBD were amplified and cloned into the EcoRI and HindIII sites to generate myc-BirA-TTP wild-type and myc-BirA-TTP-RBD. For expression in HeLa cells, full-length TTP, and PUM1 were cloned into the EcoRI and NotI sites in pCI-lNHA plasmid in order to generate pCI-lNHA-TTP. pCIlNHA-LacZ and reporter plasmids (RL-2BoxB and FL) were described previously (Fabian et al., 2011). For stable expression of FLAG-4E-T, 4E-T was amplified and cloned into BamHI and SalI sites of pBABE-FLAG-puro. 4E-T mutations were generated by site-directed mutagenesis using Phusion Hotstart II polymerase (Thermo Scientific). Transfections and Co-immunoprecipitations All plasmid transfections were carried out with PEI reagent (Polysciences). siRNA transfections were carried out using Lipofectamine 2000 (Invitrogen). For co-immunoprecipitation experiments, HeLa cells expressing FLAG-4E-T proteins were pelleted, frozen on dry ice, and resuspended in assay buffer (50 mM HEPES [pH 7.3], 100 mM KCl, 2 mM EDTA, 10% glycerol, 1 mM DTT, and 0.1% NP-40) supplemented with protease inhibitors. Lysates were clarified by centrifuging once at 20,000 3 g for 30 min. Lysates were pre-cleared with protein G Sepharose (GE Life Sciences) for 1 hr at 4 C with gentle rocking. 1,000 mg pre-cleared lysates (2 mg/ml) were incubated with benzonase (Sigma-Aldrich) and 20 ml of packed FLAG M2 agarose (Sigma-Aldrich) at 4 C for 5 hr with gentle rocking. The beads were washed five times with 1 ml of assay buffer, and precipitated proteins were eluted by heating the beads with 40 ml of Laemmli sample buffer at 95 C for 8 min. BioID Assay After an overnight incubation in 50 mM biotin, cell pellets were either lysed on ice for 1 hr in 500 ml of RIPA buffer containing 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1% NP-40, 1 mM EDTA, 1 mM EGTA, 0.1% SDS, and 0.5% sodium deoxycholate, to which a protease inhibitor cocktail (aprotinin, leupeptin, and pepstatin) and benzonase (50 U) were added immediately before use, or cell pellets were lysed in 500 ml of 13 Laemmli at room temperature. Lysates were sonicated and incubated with streptavidin-coupled agarose (Millipore) for 3 hr at 4 C to capture biotinylated proteins. The beads were then pelleted, and if the samples were initially lysed in RIPA, the beads were washed with 3 3 1 ml with RIPA buffer (not containing protease inhibitors, sodium deoxycholate, and benzonase) and 2 3 1 ml lysis buffer containing 50 mM HEPESKOH (pH 7.5), 0.1 M KCl, 10% glycerol, 2 mM EDTA, and 0.1% NP-40. If the samples were initially lysed in 13 Laemmli, the beads were washed 5 3 1 ml RIPA buffer containing 1% Triton X-100, 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 0.1% SDS, to which PMSF (1 mM) and sodium deoxycholate (0.5%) were added immediately before use. After the final washes, 40 mL of 23 Laemmli buffer was added to the samples, which were then boiled at 100 C for 8 min to elute.

1434 Cell Reports 11, 1425–1436, June 9, 2015 ª2015 The Authors

Recombinant Protein Purification and GST Pull-Down Experiments His-tagged recombinant proteins were expressed in Rosetta-2(DE3) E. coli cells (EMD Biosciences) and purified by using Ni-NTA resin (QIAGEN). MBPtagged DDX6, LSM14, or EDC3 proteins (35 pmol) were incubated with 20 ml of packed glutathione Sepharose-4B and 35 pmol of GST or various recombinant GST-4E-T fragments in binding buffer (25 mM Tris [pH 7.5], 150 mM NaCl, and 0.1% NP-40) for 2 hr at 4 C with gentle rocking. Beads were then washed five times with 1 ml of binding buffer, and proteins were eluted by boiling the beads with 40 ml of Laemmli sample buffer at 95 C for 8 min. Proteins were separated by SDS-PAGE and visualized by Coomassie staining. siRNA Transfections siRNAs against GFP, 4E-T, and PATL1 were purchased from Dharmacon. siRNAs were transfected with Lipofectamine 2000 (Invitrogen) at a final concentration of 150 nM into HeLa cells in six-well plates. mRNA Stability Assay HeLa cells were transiently transfected with either siRNA targeting GFP (control) or a gene-specific siRNA targeting 4E-T (previously described). 48 hr later, cells were transfected with plasmids coding for RL-2BoxB and FL and various lNHA-expressing proteins using PEI reagent (Polysciences). The following day, cells were treated with Act-D (5 mg/ml) for various amounts of time and lysed, and total RNA was extracted using a total RNA mini-prep kit (Biobasic). RNA was treated with Turbo DNase (Invitrogen) to remove any residual DNA, and 500 ng of purified RNA was reverse transcribed using random hexamer oligonucleotide primers and Maxima RT (Thermo Scientific). qPCR reactions were then carried out on diluted cDNA using primers targeting GAPDH and R-Luc using GoTaq qPCR mastermix (Promega) and a Mastercycler Pro machine (Eppendorf). The expression data were analyzed by the delta-delta Ct method using GAPDH as a reference gene. Poly(A) Tail Length Analysis To measure poly(A) tail lengths of RL-2BoxB and GAPDH mRNAs, we utilized the extension poly(A) test (ePAT) assay as described previously (Ja¨nicke et al., 2012). The anchor primer used in the reverse transcription step is (GCG AGC TCC GCG GCC GCG TTT TTT TTT TTT). The (dT)12VN primer used in the reverse transcription to generate a size marker for the shortest possible LMPAT product was, GCG AGC TCC GCG GCC GCG TTT TTT TTT TTT VN. ePAT PCR products were generated with Phusion polymerase (Thermo Scientific), resolved on a 2% agarose gel pre-stained with SYBR Green II (Invitrogen), and visualized using an Imagequant LAS 4000 imager (GE Healthcare). Retrovirus and Stable Cell Line Generation VSV-G pseudotyped pBABE-FLAG-4E-T retroviruses were produced in HEK293T cells. Supernatants were collected 72 hr post-transfection, passed through a 0.45-mM nitrocellulose filter, supplemented with 5 mg/ml polybrene, and applied to HeLa cells at 20% confluency. Cells were reinfected the following day and selected with puromycin (1 mg/m, Sigma-Aldrich) for 5 days. SUPPLEMENTAL INFORMATION Supplemental Information includes five figures and can be found with this article online at http://dx.doi.org/10.1016/j.celrep.2015.04.065. AUTHOR CONTRIBUTIONS M.R.F., T.N., Z.P., H.F., and S.M. designed experiments. A.C.G. and W.D. generated BirA cell lines and provided technical support. T.N. performed coimmunoprecipitation and mRNA stability assays. Z.P. performed BioID experiments. H.F. and S.M. performed binding assays. M.R.F. conceived and designed the study. All authors contributed to the editing of the manuscript. ACKNOWLEDGMENTS This work was supported by a Canadian Institutes of Health Research (CIHR) grant to M.R.F. (MOP-130425), a Natural Sciences and Engineering Research

Council of Canada (NSERC) Discovery grant to M.R.F. (RGPIN-2015-03712), an NSERC Discovery and Accelerator Supplement to A.C.G. (RGPIN-201406434 and RGPAS 462169), and Fonds de recherche du Que´bec- Sante´ (FRQS) Chercheur-Boursier Junior 1 and CIHR New Investigator awards to M.R.F. Z.P. and S.M. received support from the McGill Integrated Cancer Research Training Program. Received: March 4, 2015 Revised: April 17, 2015 Accepted: April 30, 2015 Published: May 28, 2015 REFERENCES Baron-Benhamou, J., Gehring, N.H., Kulozik, A.E., and Hentze, M.W. (2004). Using the lambdaN peptide to tether proteins to RNAs. Methods Mol. Biol. 257, 135–154. Cargnello, M., Tcherkezian, J., Dorn, J.F., Huttlin, E.L., Maddox, P.S., Gygi, S.P., and Roux, P.P. (2012). Phosphorylation of the eukaryotic translation initiation factor 4E-transporter (4E-T) by c-Jun N-terminal kinase promotes stressdependent P-body assembly. Mol. Cell. Biol. 32, 4572–4584.  lu-O¨ztu¨rk, D., Bawankar, P., Loh, B., Chang, C.T., Chen, Y., Boland, A., Kuzuog Weichenrieder, O., and Izaurralde, E. (2014). A DDX6-CNOT1 complex and W-binding pockets in CNOT9 reveal direct links between miRNA target recognition and silencing. Mol. Cell 54, 737–750. Chowdhury, A., and Tharun, S. (2009). Activation of decapping involves binding of the mRNA and facilitation of the post-binding steps by the Lsm1-7-Pat1 complex. RNA 15, 1837–1848. Clancy, J.L., Wei, G.H., Echner, N., Humphreys, D.T., Beilharz, T.H., and Preiss, T. (2011). mRNA isoform diversity can obscure detection of miRNA-mediated control of translation. RNA 17, 1025–1031. Couzens, A.L., Knight, J.D., Kean, M.J., Teo, G., Weiss, A., Dunham, W.H., Lin, Z.Y., Bagshaw, R.D., Sicheri, F., Pawson, T., et al. (2013). Protein interaction network of the mammalian Hippo pathway reveals mechanisms of kinasephosphatase interactions. Sci. Signal. 6, rs15. Dostie, J., Ferraiuolo, M., Pause, A., Adam, S.A., and Sonenberg, N. (2000). A novel shuttling protein, 4E-T, mediates the nuclear import of the mRNA 50 capbinding protein, eIF4E. EMBO J. 19, 3142–3156. Fabian, M.R., and Sonenberg, N. (2012). The mechanics of miRNA-mediated gene silencing: a look under the hood of miRISC. Nat. Struct. Mol. Biol. 19, 586–593. Fabian, M.R., Cieplak, M.K., Frank, F., Morita, M., Green, J., Srikumar, T., Nagar, B., Yamamoto, T., Raught, B., Duchaine, T.F., and Sonenberg, N. (2011). miRNA-mediated deadenylation is orchestrated by GW182 through two conserved motifs that interact with CCR4-NOT. Nat. Struct. Mol. Biol. 18, 1211–1217. Fabian, M.R., Frank, F., Rouya, C., Siddiqui, N., Lai, W.S., Karetnikov, A., Blackshear, P.J., Nagar, B., and Sonenberg, N. (2013). Structural basis for the recruitment of the human CCR4-NOT deadenylase complex by tristetraprolin. Nat. Struct. Mol. Biol. 20, 735–739. Ferraiuolo, M.A., Basak, S., Dostie, J., Murray, E.L., Schoenberg, D.R., and Sonenberg, N. (2005). A role for the eIF4E-binding protein 4E-T in P-body formation and mRNA decay. J. Cell Biol. 170, 913–924. Fromm, S.A., Truffault, V., Kamenz, J., Braun, J.E., Hoffmann, N.A., Izaurralde, E., and Sprangers, R. (2012). The structural basis of Edc3- and Scd6-mediated activation of the Dcp1:Dcp2 mRNA decapping complex. EMBO J. 31, 279–290. Henn, A., Bradley, M.J., and De La Cruz, E.M. (2012). ATP utilization and RNA conformational rearrangement by DEAD-box proteins. Annu. Rev. Biophys. 41, 247–267. Hershey, J.W., Sonenberg, N., and Mathews, M.B. (2012). Principles of translational control: an overview. Cold Spring Harb. Perspect. Biol. 4. Igreja, C., and Izaurralde, E. (2011). CUP promotes deadenylation and inhibits decapping of mRNA targets. Genes Dev. 25, 1955–1967.

Igreja, C., Peter, D., Weiler, C., and Izaurralde, E. (2014). 4E-BPs require noncanonical 4E-binding motifs and a lateral surface of eIF4E to repress translation. Nat. Commun. 5, 4790. Ja¨nicke, A., Vancuylenberg, J., Boag, P.R., Traven, A., and Beilharz, T.H. (2012). ePAT: a simple method to tag adenylated RNA to measure poly(A)tail length and other 30 RACE applications. RNA 18, 1289–1295. Kamenska, A., Lu, W.T., Kubacka, D., Broomhead, H., Minshall, N., Bushell, M., and Standart, N. (2014). Human 4E-T represses translation of bound mRNAs and enhances microRNA-mediated silencing. Nucleic Acids Res. 42, 3298–3313. Kedersha, N., and Anderson, P. (2009). Regulation of translation by stress granules and processing bodies. Prog. Mol. Biol. Transl. Sci. 90, 155–185. Kim, D.I., Birendra, K.C., Zhu, W., Motamedchaboki, K., Doye, V., and Roux, K.J. (2014). Probing nuclear pore complex architecture with proximity-dependent biotinylation. Proc. Natl. Acad. Sci. USA 111, E2453–E2461. Lee, Y.S., and Dutta, A. (2007). The tumor suppressor microRNA let-7 represses the HMGA2 oncogene. Genes Dev. 21, 1025–1030. Leppek, K., Schott, J., Reitter, S., Poetz, F., Hammond, M.C., and Stoecklin, G. (2013). Roquin promotes constitutive mRNA decay via a conserved class of stem-loop recognition motifs. Cell 153, 869–881. Lykke-Andersen, J., and Wagner, E. (2005). Recruitment and activation of mRNA decay enzymes by two ARE-mediated decay activation domains in the proteins TTP and BRF-1. Genes Dev. 19, 351–361. Marnef, A., Sommerville, J., and Ladomery, M.R. (2009). RAP55: insights into an evolutionarily conserved protein family. Int. J. Biochem. Cell Biol. 41, 977–981. Mathys, H., Basquin, J., Ozgur, S., Czarnocki-Cieciura, M., Bonneau, F., Aartse, A., Dziembowski, A., Nowotny, M., Conti, E., and Filipowicz, W. (2014). Structural and biochemical insights to the role of the CCR4-NOT complex and DDX6 ATPase in microRNA repression. Mol. Cell 54, 751–765. Mayr, C., Hemann, M.T., and Bartel, D.P. (2007). Disrupting the pairing between let-7 and Hmga2 enhances oncogenic transformation. Science 315, 1576–1579. Minshall, N., Reiter, M.H., Weil, D., and Standart, N. (2007). CPEB interacts with an ovary-specific eIF4E and 4E-T in early Xenopus oocytes. J. Biol. Chem. 282, 37389–37401. Mullen, T.E., and Marzluff, W.F. (2008). Degradation of histone mRNA requires oligouridylation followed by decapping and simultaneous degradation of the mRNA both 50 to 30 and 30 to 50 . Genes Dev. 22, 50–65. Ozgur, S., Chekulaeva, M., and Stoecklin, G. (2010). Human Pat1b connects deadenylation with mRNA decapping and controls the assembly of processing bodies. Mol. Cell. Biol. 30, 4308–4323. Roux, K.J., Kim, D.I., Raida, M., and Burke, B. (2012). A promiscuous biotin ligase fusion protein identifies proximal and interacting proteins in mammalian cells. J. Cell Biol. 196, 801–810. Roux, K.J., Kim, D.I., and Burke, B. (2013). BioID: a screen for protein-protein interactions. Curr. Protoc. Protein Sci. 74, 23. Rouya, C., Siddiqui, N., Morita, M., Duchaine, T.F., Fabian, M.R., and Sonenberg, N. (2014). Human DDX6 effects miRNA-mediated gene silencing via direct binding to CNOT1. RNA 20, 1398–1409. Sachs, A.B., and Davis, R.W. (1989). The poly(A) binding protein is required for poly(A) shortening and 60S ribosomal subunit-dependent translation initiation. Cell 58, 857–867. Sandler, H., Kreth, J., Timmers, H.T., and Stoecklin, G. (2011). Not1 mediates recruitment of the deadenylase Caf1 to mRNAs targeted for degradation by tristetraprolin. Nucleic Acids Res. 39, 4373–4386. Sheth, U., and Parker, R. (2006). Targeting of aberrant mRNAs to cytoplasmic processing bodies. Cell 125, 1095–1109. Standart, N., and Minshall, N. (2008). Translational control in early development: CPEB, P-bodies and germinal granules. Biochem. Soc. Trans. 36, 671–676. Tan, D., Zhou, M., Kiledjian, M., and Tong, L. (2014). The ROQ domain of Roquin recognizes mRNA constitutive-decay element and double-stranded RNA. Nat. Struct. Mol. Biol. 21, 679–685.

Cell Reports 11, 1425–1436, June 9, 2015 ª2015 The Authors 1435

Tarun, S.Z., Jr., Wells, S.E., Deardorff, J.A., and Sachs, A.B. (1997). Translation initiation factor eIF4G mediates in vitro poly(A) tail-dependent translation. Proc. Natl. Acad. Sci. USA 94, 9046–9051. Tritschler, F., Braun, J.E., Eulalio, A., Truffault, V., Izaurralde, E., and Weichenrieder, O. (2009). Structural basis for the mutually exclusive anchoring of P body components EDC3 and Tral to the DEAD box protein DDX6/Me31B. Mol. Cell 33, 661–668. Van Etten, J., Schagat, T.L., Hrit, J., Weidmann, C.A., Brumbaugh, J., Coon, J.J., and Goldstrohm, A.C. (2012). Human Pumilio proteins recruit multiple deadenylases to efficiently repress messenger RNAs. J. Biol. Chem. 287, 36370–36383.

1436 Cell Reports 11, 1425–1436, June 9, 2015 ª2015 The Authors

Wells, S.E., Hillner, P.E., Vale, R.D., and Sachs, A.B. (1998). Circularization of mRNA by eukaryotic translation initiation factors. Mol. Cell 2, 135–140. Wilhelm, J.E., Hilton, M., Amos, Q., and Henzel, W.J. (2003). Cup is an eIF4E binding protein required for both the translational repression of oskar and the recruitment of Barentsz. J. Cell Biol. 163, 1197–1204. Yang, G., Smibert, C.A., Kaplan, D.R., and Miller, F.D. (2014). An eIF4E1/4E-T complex determines the genesis of neurons from precursors by translationally repressing a proneurogenic transcription program. Neuron 84, 723–739.  lu-O¨ztu¨rk, D., and Izaurralde, E. (2013). GW182 proteins Zekri, L., Kuzuog cause PABP dissociation from silenced miRNA targets in the absence of deadenylation. EMBO J. 32, 1052–1065.