Dynamic changes in eIF4F-mRNA interactions revealed by global analyses of environmental stress responses
© The Author(s). 2017
Received: 22 August 2017
Accepted: 13 October 2017
Published: 27 October 2017
Translation factors eIF4E and eIF4G form eIF4F, which interacts with the messenger RNA (mRNA) 5′ cap to promote ribosome recruitment and translation initiation. Variations in the association of eIF4F with individual mRNAs likely contribute to differences in translation initiation frequencies between mRNAs. As translation initiation is globally reprogrammed by environmental stresses, we were interested in determining whether eIF4F interactions with individual mRNAs are reprogrammed and how this may contribute to global environmental stress responses.
Using a tagged-factor protein capture and RNA-sequencing (RNA-seq) approach, we have assessed how mRNA associations with eIF4E, eIF4G1 and eIF4G2 change globally in response to three defined stresses that each cause a rapid attenuation of protein synthesis: oxidative stress induced by hydrogen peroxide and nutrient stresses caused by amino acid or glucose withdrawal. We find that acute stress leads to dynamic and unexpected changes in eIF4F–mRNA interactions that are shared among each factor and across the stresses imposed. eIF4F–mRNA interactions stabilised by stress are predominantly associated with translational repression, while more actively initiating mRNAs become relatively depleted for eIF4F. Simultaneously, other mRNAs are insulated from these stress-induced changes in eIF4F association.
Dynamic eIF4F–mRNA interaction changes are part of a coordinated early translational control response shared across environmental stresses. Our data are compatible with a model where multiple mRNA closed-loop complexes form with differing stability. Hence, unexpectedly, in the absence of other stabilising factors, rapid translation initiation on mRNAs correlates with less stable eIF4F interactions.
Eukaryotic messenger RNAs (mRNAs) bear a 5′ m7GpppN cap structure that is added co-transcriptionally in the nucleus and is important for their translation and stability in the cytoplasm. eIF4E is the major cytoplasmic cap-binding protein that, together with eIF4G, plays important roles enhancing translation rates via recruiting 40S ribosomes and associated factors for translation initiation events. eIF4G binds to eIF4E via conserved regions on the surfaces of each factor [1, 2]. The poly A tail at the 3′ UTR also enhances translation rates, acting synergistically with the 5′ cap . Multiple lines of evidence support the formation of an mRNA ‘closed loop’ wherein the poly A binding protein interacts with both the poly A tail and eIF4G that in turn simultaneously binds to 5′ cap-bound eIF4E. Such a structure has been demonstrated to circularise mRNAs and is proposed to explain the enhanced translation rates of capped and polyadenylated mRNAs. However, the mechanisms of translational enhancement are not yet clear [4, 5].
eIF4E and eIF4G function at the 5′ cap to help recruit active eIF2-ternary complex (TC)–bound small ribosome subunits (40S) to the mRNA 5′ end. This complex contains additional translation factors including eIF1, eIF1A, eIF3 and eIF5 and is termed the 43S pre-initiation complex . Interactions between eIF4G and eIF3 or eIF5 serve as a bridge to the 40S subunit [6, 7]. 40S recruitment is also facilitated by the RNA helicase eIF4A, which can unwind mRNA secondary structures close to the 5′ cap. By promoting this cap-dependent initiation pathway for ribosome recruitment, the eIF4F components may also help prevent aberrant competing pathways, such as eIF3-led mRNA recruitment .
Translation initiation occurs at different rates on individual mRNAs and contributes to the discrepancies seen between protein and mRNA levels when measured globally [9–11]. Mechanistically, how the dynamics of translation factor binding to individual mRNAs influences initiation frequencies is not yet understood, as very few mRNAs and/or factors have been studied intensely. However, a global computational analysis suggested that the time taken for an individual initiation event can vary over 100-fold between individual yeast mRNAs (in the range of approximately 1 s to > 130 s) . This is in agreement with multiple studies indicating that initiation is rate-limiting for the overall translation process .
We recently employed an RNA-immune precipitation and next-generation sequencing approach (RIP-seq) to quantify the abundance of mRNAs associating with different translation factors and RNA-binding proteins in budding yeast, making use of isogenic strains bearing individual TAP-tagged proteins [13–15]. We examined interactions between mRNAs and eIF4E, the two isoforms of eIF4G (4G1 and 4G2), Pab1 and the two yeast eIF4E-binding proteins (4E-BPs) Caf20 and Eap1 that can compete with eIF4G for binding to eIF4E . While there was relatively little variation in the enrichment of mRNAs with Pab1, which can bind multiple times to the poly A tail, there were marked differences in the association of the other proteins examined and some results were unexpected. We found that eIF4E, eIF4G1 and eIF4G2 were co-enriched or depleted across many hundreds of mRNAs, such that there were almost no differences in the enrichment patterns for all three proteins across thousands of mRNAs. Similarly, the 4E-BPs shared similar enrichment patterns to each other that were distinct from the eIF4F proteins. By clustering the mRNAs, we grouped nearly 3000 mRNAs into four broad mRNA classes that were either relatively co-enriched (Groups III and IV) or co-depleted (Groups I and II) for eIF4E and eIF4G and also either relatively co-enriched (Groups II and IV) or co-depleted (Groups I and III) for the 4E-BPs . In accord with predictions from the closed-loop model, Group III mRNAs (bound by eIF4E and eIF4G, but depleted for the 4E-BPs) encode abundant proteins and have high translation efficiencies (TEs) as inferred from ribosome footprinting studies, while the 4E-BP enriched mRNA classes had lower TEs .
Subsequent analyses showed that Group III mRNAs, also called the ‘strong closed-loop’ set, had short ORFs [16, 17] and that they were also preferentially bound by ribosomes bearing the Asc1/RACK1 protein  leading to the idea that Asc1/RACK1 helps direct ribosomes to bind and promote translation on Group III mRNAs with short ORFs [5, 17]. Further genome-wide analyses indicated that Group III mRNAs were translated largely independent of eIFs 4A, 4B and Ded1, as translation of these mRNAs was relatively unaffected by mutants inactivating each factor . Two unexpected groups were Groups I and II. Group II was depleted for eIF4F and enriched for the 4E-BPs. An explanation for this observation was found subsequently, as both yeast 4E-BPs were found to interact with translating ribosomes engaged with specific mRNAs independently of 4E-BP–eIF4E interactions . In contrast, Group I mRNAs were relatively depleted for closed-loop proteins but have high TEs and encode abundant proteins. This group included mRNAs encoding many glycolytic enzymes. It was suggested that these mRNAs may recruit ribosomes via an alternative mechanism, possibly involving Pab1. Indeed, earlier observations suggested that Pab1 can stimulate translation by multiple means [3, 19]. In addition, there is evidence that eIF3 and eIF2 can promote binding of some mRNAs to 40S ribosomes independently of eIF4G . However, the mechanism of mRNA selection operating on Group I RNAs remains to be determined.
A range of acute stresses causes widespread reprogramming of translation initiation in different cell types [21, 22]. We, and others, have studied environmental stress responses in yeast on a global scale . Relevant to this present study are glucose and amino acid nutritional starvations as well as oxidative stress induced by hydrogen peroxide addition. Both amino acid and peroxide stresses promote phosphorylation of eIF2. This impairs translation initiation globally by inhibiting the activity of eIF2B, the guanine nucleotide exchange factor that normally activates eIF2 to promote TC formation [23–25]. Thus, eIF2 phosphorylation lowers global translation by interfering with TC formation. Both stresses also activate GCN4 translation (a transcriptional activator of amino acid biosynthetic genes) as upstream ORFs inversely couple eIF2B activity and GCN4 translation levels . Thus, translational control activates stress-responsive regulatory networks. Despite these common events, each stress impacts distinctly on the translation of some individual mRNAs, indicating that other controlling elements are important to both repress bulk translation and promote translation of specific stress-responsive mRNAs to promote recovery [26–29]. In contrast, short-term glucose withdrawal does not promote phosphorylation of eIF2 and translational repression therefore operates via a distinct mechanism . It was shown that loss of eIF4A from translating ribosomes occurs early following glucose removal  and so inhibition of eIF4A activity may contribute to glucose-mediated translational repression. Following glucose depletion, many mRNAs become unstable , while others are localised within cytoplasmic granules called P bodies that contain both eIF4E and RNA-decay proteins. In addition, ‘stress’ granules (also termed ‘EGP bodies’) form that lack RNA-decay proteins but contain mRNAs and some translation factors including eIF4E, eIF4G and Pab1 [33–35]. mRNAs translationally activated following glucose starvation include pentose-phosphate pathway mRNAs . It was also found that heat shock mRNAs transcribed from promoters bound by Hsf1 were translationally activated rather than being localised to P bodies or stress granules .
As we unexpectedly found that eIF4E and eIF4G were differently associated with mRNAs , and because the role of these 5′ cap associated factors in the translational responses to stress in yeast had not been evaluated, we have now investigated how eIF4E/G1/G2–mRNA interactions are changed in response to these three acute stresses that reprogram translation. Following a previously used RIP-seq approach, , we find that each stress does alter the pattern of association of mRNAs with each factor in a coordinated way. Many of the changes observed are not stress-specific, suggesting that there is a common early general stress response that alters eIF4F–mRNA interactions. Our analyses indicate that the direction of change in eIF4F interactions with many mRNAs opposes both changes in transcript levels and changes in ribosome occupancy following stress. We find that the Group I mRNAs, as described above, are particularly sensitive to stress and interpret our data as compatible with a model where multiple mRNA closed-loop complexes can form with differing stability.
Stresses inhibit translation initiation but do not promote large changes in closed-loop factor protein–protein interactions
Transcript abundance changes are stress specific and factor independent
We proceeded to examine the mRNAs bound to each factor by affinity capture following stress and next-generation RNA sequencing (RNA-seq). As controls and to monitor stress-induced changes in mRNA abundance, rRNA-depleted total RNA was sequenced (Fig. 1c). All samples were sequenced in triplicate and mapped to the S. cerevisiae genome using standard tools (see ‘Methods’). Per gene counts are shown in Additional file 1: Supplementary Source Data 1 and summarised in Additional file 2: Figure S1A. First, we compared total RNA samples to examine changes in transcript abundance following stress. All replicates correlated well with each other; pairwise Pearson correlations between replicates exceeded 0.95 (except eIF4G2 –glu > 0.935). In comparisons between tagged strains, variations are no greater between control unstressed samples and those exposed to the same stress. Only Pearson correlations for –aa samples differ greatly from the other stresses (Additional file 2: Figure S1B). The replicate data were analysed with edgeR to calculate the fold changes in transcript abundance (presented as log2 ratios of cpm, here termed ∆T; Additional file 3: Supplementary Source Data 2). Following these short treatment periods, pairwise ∆T plots show that –aa stress has a large impact on the transcriptome (the differences in transcript abundance are spread along a wider range), while –glu and + H2O2 do not (Additional file 2: Figure S2A). –aa stress targets expected gene categories, significantly downregulating ribosome and protein synthesis genes, while upregulating transport and carbohydrate metabolism in line with previous reports [26, 28]. Hence while each stress has a similar impact on the global polysome profiles, they have different effects on the transcriptomes at these early time points following stress application. As transcription changes more robustly when longer stress periods are used [37, 38], these data indicate that translational repression largely precedes transcriptional reprogramming (Additional file 2: Figure S2B).
Stress induces changes in relative association of mRNAs with the eIF4F proteins
As outlined in the introduction, our prior analysis of eIF4E and eIF4G1/2 association with mRNAs revealed greater variation between mRNAs than expected . Our stress TAP-IP sequencing samples have Pearson correlations between replicates and across strains subjected to the same stress similar to those for unstressed samples (Additional file 2: Figure S3A). In line with our prior analyses of unstressed TAP-IP samples, edgeR data analyses indicate that there are hundreds of mRNAs significantly differentially enriched or depleted with eIFs 4E, 4G1 and 4G2 compared with total RNA (termed IP/T) following each stress (false discovery rate [FDR] < 0.05; Additional file 2: Figure S3B and Additional file 4: Supplementary Source Data 3).
Anti-correlated changes in transcription (∆T) and eIF4F association (∆IP)
Gene Ontology (GO) comparisons of significantly changing mRNAs in both ∆T and ∆IP using yeast GO Slim categories (Additional file 2: Figure S5) also highlight the opposing natures of ∆T and ∆IP. ∆T changes are largely stress-specific and as expected from previous comparable studies [26, 27, 31]; –aa upregulates carbohydrate metabolism and membrane transport functions and downregulates translation, –glu upregulates carbohydrate transport systems and + H2O2 induces oxidative stress-response mRNAs. In contrast, ∆IP changes are more typically shared across the stresses and proteins. There are very few GO Slim terms over-represented among the eIF4F-depleted mRNAs (blue cells in Additional file 2: Figure S5, right panel), instead almost all coordinated GO term enrichments appear among those mRNAs whose eIF4F interactions are increased following stress. Thus, we conclude that ∆T and ∆IP changes in eIF4F mRNA association are negatively correlated across the stresses investigated and that decreased eIF4F-mRNA binding following stress is not apparently targeted to specific GO categories.
The current understanding of the role of eIF4F is to promote translation initiation. Frequently, stress changes promote ‘potentiation’ where enhanced translation accompanies and reinforces changes in transcription  and this has been observed previously for –aa starvation  and glucose depletion . So, it was highly unexpected that the changes in eIF4E and eIF4G association with mRNAs that we observed should oppose the changes in transcript abundance. However, the responses seen are consistent across all three independent stresses applied.
Anti-correlated changes in ribosome occupancy (∆TE) and eIF4F association (∆IP)
To compare our ∆IP data with changes in translation we were able to use literature-sourced ribosome footprinting datasets for each of our chosen stresses. When sequencing of ribosome-protected RNA fragments was first described as a powerful measure of protein synthesis, Ingolia et al. used the technique to address the role of amino acid starvation . Similarly, Gerashchenko et al. reported the impact of hydrogen peroxide  and Zid and O’Shea studied glucose depletion . All experimental data we analysed here employed the same yeast strain and stress protocols were similar, but not identical, to our own (e.g. cells grown at 30 °C and short time-point stresses: –aa 20 min, –glu 15 min, +H2O2 30 min). A commonly accepted way to summarise the ribosome footprinting data is to calculate a measure of ribosome engagement with mRNAs by summing the footprints mapped to a gene and dividing by total transcript counts (each normalised to read depth and ORF length). This is widely termed TE. Because bulk polysomes are depleted in stressed cells (Fig. 1), absolute TEs of most RNAs are reduced by stress. However, these effects are dampened by the normalisation of total footprint counts, such that changes in TE or ∆TE calculations are relative rather than absolute. Therefore, genes with a negative ∆TE have a greater than average reduction in translation, while those with a positive ∆TE represent both mRNAs more resistant to repression than average and those mRNAs with enhanced translation under the applied stress.
Many mRNAs have been found to accumulate in P bodies and/or stress granules during the glucose starvation response, while others remain cytoplasmically diffuse. Granules formed during stress are believed to be sites where non-translating mRNAs accumulate. For example, PGK1, RPS16A, PDC1 and HXK1 have all been shown to rapidly localise to P bodies following glucose depletion using fluorescence tagging experiments [35, 36]. All these mRNAs become relatively eIF4F-associated in our experiments and tend to have lower ∆TEs (black spots in Fig. 4b, right panels). In contrast, several HSP mRNAs remained diffusely cytoplasmic in mRNA localisation studies and maintained high ∆TEs . These mRNAs become relatively depleted of eIF4F proteins, particularly HSP30 and PHM8. Phm8 is a nucleotidase functioning in the ribose salvage pathway that is induced following glucose depletion .
Taken together these data indicate that eIF4E and the eIF4G isoforms respond in a coordinated fashion to these different stresses, changing their affinity to a portion of the mRNAs in the cell. The changes in eIF4F binding go against prior expectations, such that those mRNAs enriched with the eIF4F proteins (positive ∆IP) are among those mRNAs that become relatively depleted of ribosomes (negative ∆TE); while those mRNAs that become relatively depleted of eIF4F following stress are among those that are comparatively better engaged with ribosomes. One explanation for the increased stability of eIF4F with translationally repressed mRNAs is the formation of a stable mRNA closed loop enriched in eIF4E and eIF4G on translationally inactive mRNAs. When glucose is depleted these enter P bodies and/or stress granules. In contrast, mRNAs that become relatively depleted in eIF4F proteins are a sub-set of those mRNAs that remain or become better engaged with ribosomes following stress. These mRNAs are reminiscent of a class of mRNA identified in our previous study of closed-loop factor–mRNA interaction in actively growing cells . As outlined in the ‘Introduction’, we previously found a cluster of mRNAs (termed Group I) that were well translated but depleted for the eIF4F proteins. So, to gain further insight into how stress remodels eIF4F–mRNA interactions we examined the fate of the different cluster mRNAs to each stress.
Group I mRNAs become enriched in eIF4F following stress
The impact of stress on factor associations indicates that Group I mRNAs and, to a lesser extent, Group II mRNAs (the clusters most depleted of eIF4F proteins in unstressed cells) become relatively enriched with eIF4E in response to all three stresses (Fig. 5b, red and blue boxes). Equivalent trends are observed for eIF4G1 and 4G2 (Fig. 5a and Additional file 2: Figure S6B). At the same time, these mRNA groups become relatively translationally repressed according to ∆TE measurements (Fig. 5a and c). This suggests that the reciprocal effects seen in Fig. 4 on ∆TE and ∆IP associated with translational repression following multiple stresses are linked with changes in eIF4F associations with mRNAs in Group I, and to a lesser extent Group II. In contrast, Groups IIIB, IVA and IVB that were more stably associated with eIF4F in unstressed IPs tend to be the most depleted groups following stress; though the degree of change for these groups is typically quite modest. At least some of these changes are stress-specific. For example, when the scatterplots shown in Fig. 4 are re-plotted to highlight cluster Group I and Group IVA mRNAs, it is clear that there are reciprocal changes in both eIF4F factor associations and ∆TE for –aa and + H2O2 (red and lilac spots in Fig. 5d and Additional file 2: Figure S6C). Group IVA changes in ∆TE are more varied in the –glu ∆TE data from Zid and O’Shea  (bottom panels). This may be because –glu does not induce translation initiation repression by eIF2 phosphorylation, unlike the other two stresses examined [27, 30]. Finally, Group III mRNAs (especially IIIB) were originally classified as relatively enriched in eIF4F with high TE and were later shown to have short ORFs (Additional file 2: Figure S6A). Group III mRNAs appear largely resistant to change in IP (Fig. 5b) and only have low ∆TE following –glu (Fig. 5c). Therefore, each of the previously identified closed-loop group mRNAs not only have altered factor associations in unstressed cells, they also each exhibit different responses to stress.
Anti-correlation between eIF4F association and calculated initiation times
Our experiments have addressed how interactions between mRNAs and the eIF4F proteins (eIF4E and the two eIF4G forms) are altered in response to three acute stresses that each cause a global translation initiation inhibition, specifically nutrient withdrawal (–glu and –aa) or response to a toxic oxidant (+H2O2) (Fig. 1). We used a relatively simple metric of the degree of mRNA association with TAP-tagged proteins captured on magnetic IgG coupled beads. Many mRNAs change in their associations with the eIF4F proteins (Fig. 2). In accord with prior studies there appears to be a large overlap and broad agreement between the mRNAs binding each eIF4G isoform [13, 40]. While there may be some gene-specific differences between the eIF4G isoform interactions with some mRNAs, the global changes in RNA–protein association that we observe are generally shared between the factors and stresses, suggesting that the eIF4E–eIF4G complex is responding globally to stress and/or translation inhibition as a single unit rather than as individual proteins (Fig. 2). This common global response to environmental stress that changes eIF4F–mRNA interactions likely precedes the well-known common transcriptional response to multiple environmental stresses described previously [37, 38]. This is because at the short times following stress examined here (10–20 min), cells have not yet induced fully their transcriptional response programs.
When we compared a range of mRNA features across the ∆IP significantly enriched (P < 0.05) and depleted gene sets, including 5′ UTR and 3′ UTR lengths, we reported secondary structure and for enrichments for known RNA-binding proteins. No clear enrichments were found (not shown). This is perhaps unexpected, but likely reflects that eIF4E binds primarily to the 5′ cap structure in yeast and that eIF4G-binding to eIF4E enhances this cap affinity (KD = < 15–20 nM) [8, 41, 42] to promote 40S ribosome recruitment and protein synthesis initiation. Many studies show that eIF4E and eIF4G interaction with mRNA promote its translation. It was therefore against expectations that the changes in factor–mRNA associations induced in response to stress are anti-correlated with both transcription (Fig. 3) and translation changes (Fig. 4). By these measures of gene expression, it appears that mRNAs which become more associated with both eIF4E and eIF4G are among those mRNAs that are translationally repressed by stress. Our data specifically point to classes of mRNA that are normally well translated, but depleted of eIF4F in unstressed cells, that become relatively more enriched in eIF4F and relatively poorly translated in stressed cells (Group I and to a lesser extent Group II mRNAs; Fig. 5). At the same time mRNAs that become less associated with eIF4F are among those mRNAs relatively better translated in response to stress (Figs. 4 and 5). Calculated overall initiation times are also lower for mRNAs depleted for eIF4F in unstressed cells (Fig. 6). Hence, these results appear to run counter to widespread prior observations that eIF4E and eIF4G function to promote translation initiation [1, 8, 43].
In thinking of how to explain these observations, it is important to consider both eIF4E-dependent and -independent mechanisms that may account for our observed eIF4F depletion on specific mRNAs and the reciprocal changes induced by stress and translation inhibition. eIF4E-independent mechanisms include the use of alternative cap-binding proteins. For example, it has recently been found that human eIF4E has differential mRNA specificity dependent upon the 5′ terminal nucleotide of the mRNA . A test RNA beginning with a C residue had lower affinity for eIF4E than RNAs beginning with A or G. In addition, transcripts beginning with C had lower ribosome occupancy . However, C-initiating transcripts, including 5′ TOP mRNAs, can bind preferentially to the 5′ cap-C–binding DM15 domain within human LARP1 . These studies indicate that specific mRNAs may interact with cap-binding proteins other than eIF4E. However, such a mechanism appears unlikely to explain our findings, because (1) both yeast LARP proteins, Slf1 and Sro9, lack DM15 related regions ; and (2) the defined mRNA groups I–IVC show no altered preferences in mapped 5′ end nucleotides (analysis not shown). In yeast, other known 5′ cap interacting proteins include the predominantly nuclear cap-binding complex (CBC) of Cbc1/Sto1 and Cbc2, equivalent to mammalian Cbp80 and Cbp20, respectively. The CBC is involved in mRNA splicing, transcription termination, mRNA export and the pioneer round of translation . In addition, deletion of CBC1 was shown to enhance translational repression and delay the cellular response to hyperosmotic stress , although there is no report of a specific role for the CBC in any of the stresses studied here. Second, the Dcp1 and Dcp2 decapping complex that removes the 5′ cap structure from mRNAs during mRNA decay also binds the cap before its cleavage [49, 50]. PGK1 is a Group I mRNA and is one of the key relatively stable mRNAs upon which the general mRNA-decay model is derived. This model indicates that poly A tail shortening precedes decapping and 5′ - 3′ exonucleolytic decay [49, 50]. It follows therefore that deadenylated mRNAs should have lower Pab1-binding, but we found Pab1 binding enhanced on Group I mRNAs (Additional file 2: Figure S6A; ). So, as a whole, our data are not compatible with current understanding of known alternative cap-interacting proteins explaining Group I mRNA behaviour, so they are not discussed further.
It was demonstrated recently that the d subunit of eIF3 possesses 5′ cap binding activity and it was further shown that eIF3-led 43S recruitment promotes c-Jun mRNA translation . Thus, a similar means could provide an alternative explanation for the behaviour of Group I mRNAs. However, yeast eIF3 does not have an eIF3d orthologue  and so a directly analogous mechanism appears unlikely. However, because yeast eIF3 possesses RNA-binding domains  and can enhance recruitment of 40S ribosomes to mRNAs independently of the eIF4F complex [8, 20], eIF3-promoted mRNA recruitment remains one possible explanation for low eIF4F retention on Group I mRNAs. How yeast eIF3 could protect the 5′ cap to resist RNA decay remains unresolved.
If not initiation via an eIF4E-independent route, then how could an eIF4F-dependent mechanism account for our findings? There is evidence that there are multiple forms of closed-loop complex and that some states are more readily disrupted than others (see below). Canonical translation initiation typically has several main steps: (1) eIF4F–mRNA interactions; (2) recruitment of 43S complexes; (3) scanning; (4) AUG codon recognition; and (5) 60S subunit joining. Only one initiation event is thought to occur on an individual mRNA at any one time, so highly translating mRNAs likely cycle rapidly between successive rounds of initiation.
Previous, mainly in vitro, studies support the idea of distinct eIF4F–mRNA complexes during the initiation cycle. In vitro translation reactions in yeast identified distinct 48S and 80S forms of closed-loop complexes containing eIF4E, eIF4G and Pab1 . In addition, evidence for a distinct 60S joining ‘closed-loop’ complex was found using reporter RNAs electroporated into cells containing various mutated translation factors . Similarly, rabbit reticulocyte extract translation of 5′ capped poly A– RNA recently showed that eIF4F was bound to 48S complexes in sucrose gradients, but released from 80S complexes . These data support the idea that the closed-loop complex is remodelled upon 60S subunit joining. Mechanistically, this may entail disruption of 40S-eIF4G interactions that would promote both 60S joining and free eIF4G to recruit a new 40S complex. To explain our data, we propose that one of these states is less stable to capture in our RIP-seq experiments. Thus, rapidly initiating mRNAs become relatively depleted in eIF4F in our experimental system, which therefore provides a convenient tool to identify these RNAs (Fig. 6b).
In this model, in unstressed cells, Group I mRNAs are actively translated and cycle between closed-loop states 2 and 3 (Fig. 7). During acute stress, these mRNAs become translationally repressed and spend more time in state 1 (red arrow in Fig. 7), which is stable to capture in our assay. Likewise, there are other RNAs such as those in Group IVA that are relatively poorly translated in unstressed actively growing cells, but relatively enhanced in eIF4F interactions. In our model, these RNAs are spending longer in state 1 in unstressed cells, but have relatively enhanced translation during stress and lower eIF4F association (Fig. 6c and Additional file 2: Figure S6C). Hence these mRNAs become activated from state 1 into state 2–3 during stress (green arrow in Fig. 7). Presumably recovery from stress and resumed translation and cell growth would reverse these changes. This is different to a previous proposal for differential enrichment of eIF4F/Pab1 with various mRNAs where the mRNA closed loop was suggested to be transient and only prevalent primarily during mRNA activation and becoming less widespread during steady state translation .
Importantly, not all mRNAs fit this model. Many mRNAs appear resistant to reciprocal changes in TE and eIF4F IP. Most notable are those in RNA classes IIIA and IIIB. These short mRNAs are relatively stably bound to eIF4F and are well translated in unstressed cells. As indicated in the introduction, these mRNAs are enriched in ribosomes bearing the 40S associated factor Asc1/RACK1 suggesting that Asc1/RACK1 helps direct ribosomes to bind these mRNAs to promote their translation . As eIF4F interactions for Group IIIA and B mRNAs appear relatively unaffected by stress (Fig. 6), it is possible that Asc1/RACK1 acts to maintain the closed loop on these mRNAs. As Asc1 has been shown to recruit other RNA-binding proteins to ribosomes, such as Scp160 , further networks of RNA–protein interactions will likely contribute to the modulation of eIF4F–mRNA interactions and translational responses to stress.
We show that 5′ cap-associated eIF4F proteins change mRNA interactions as part of a coordinated common early translational response to environmental stresses. The eIF4F–mRNA interactions are dynamic and, unexpectedly, generally oppose changes in relative translation and transcription. The data are compatible with a model where multiple mRNA-eIF4F complexes can form with differing stability. Hence, in the absence of other eIF4F stabilizing factors, rapid translation initiation on mRNAs correlates with less stable eIF4F interactions, while the converse is seen for repressed mRNAs.
Strains and growth conditions
eIF4G1 and eIF4G2 TAP-tagged His+ strains in the BY4741 background were obtained from Open Biosystems. An untagged HIS3 BY4741 control strain (GP6001) was used as a control for all TAP experiments . Because the eIF4E-TAP strain altered eIF4E levels a new strain was made using standard cre-lox procedures in GP6001 creating an eIF4E-TAP-tagged strain (GP6312) retaining its native 3′ UTR and a lox scar, as described previously . Strains were grown at 30 °C in synthetic complete dextrose media lacking histidine [SCD–his] grown to A600 = ~0.6. Stress treatments included harvesting cells by rapid centrifugation in a warm centrifuge and resuspending in pre-warmed SD minimal medium (–aa, for 20 min) or SC–his medium lacking glucose (–glu for 10 min) or by addition of 0.4 mM H2O2 (+H2O2 for 15 min) before cell harvest.
S. cerevisiae was grown to an OD600 = 0.6 and, where appropriate, cells were stressed as described above. Cycloheximide was added to a final concentration of 0.1 mg/mL immediately before cell harvest by centrifugation. Cells were lysed into polyribosomal buffer containing cycloheximide and 5 OD260 units were loaded onto a 15–50% sucrose gradient, poured as previously described .
TAP-affinity purifications were performed as described previously [13, 57]. Briefly, yeast cultures were grown to A600 = ~0.6, pelleted, snap frozen in liquid nitrogen and ground in Buffer A (20 mM Tris-HCl [pH 8], 140 mM NaCl, 1 mM MgCl2, 0.5% NP40, 0.5 mM DTT, 1 mM PMSF, EDTA free Protease Inhibitor cocktail tablet (Roche), 100 μM NaV3O4, 5 mM NaF and 40 units/mL RNAsin) using liquid nitrogen and a 6870 Freezer Mill (Spex). Lysates were cleared through two centrifugation steps (15,000 x g at 4 °C) and quantified using Bradford Reagent. Ten milligrams of mg total protein were loaded onto Rabbit IgG coupled Tosyl-activated Dynabeads M-280 magnetic beads (Dynal) to ensure maximum depletion of the tagged protein from each extract. Coupling of Rabbit IgG to Tosyl-activated Dynabeads M-280 magnetic beads and TAP affinity purification was performed as previously described . After the final wash, the beads were re-suspended in 270 μL Buffer A. A 20-μL aliquot of the sample was set aside for western blot analysis, and RNA was purified from the remaining 250 μL for RNA-seq.
Western blot analyses
Protein samples were mixed with 2 x SDS loading dye and heated to 95 °C for 10 min to dissociate protein complexes from the IgG Tosyl-activated Dynabeads M-280 magnetic beads. IP samples were resolved by SDS–PAGE, electroblotted onto nitrocellulose membrane and probed using the relevant primary antibody. TAP tagged proteins were detected using an HRP-conjugated primary antibody to Protein A (Abcam, Cambridge, MA, USA). All other primary antibodies were detected with horseradish peroxidase (HRP)-conjugated rabbit secondary antibody, except Pab1, which was detected using HRP-conjugated mouse secondary antibody as previously described .
Triplicate samples of TAP-associated RNA were isolated on M-280 magnetic beads as described above. For RNA target identification, after the final wash, the beads were resuspended in 270 μL Buffer A. A 20-μL aliquot of the sample was set aside for western blot analysis and RNA was purified from the remaining 250 μL. Total (T) RNA and paired immune precipitated (IP) RNA were isolated using a standard Trizol Reagent (Life Technologies, Carlsbad, CA, USA) protocol and re-suspended in 10 μL diethylpyrocarbonate (DEPC) treated water. RNA was quantified using a Nanodrop 8000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). rRNA was then depleted from both total and IP RNA samples using the Ribominus™ Eukaryote Kit for RNA-seq (Life Technologies, Carlsbad, CA, USA). Depleted samples were ethanol precipitated, washed twice with 70% ethanol and resuspended in 10 μL DEPC water. rRNA depletion was assessed on a 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA) using a RNA nanochip and the remaining RNA stored at –80 °C. Sequencing libraries were generated using the whole Transcriptome Library Preparation protocol provided with the SOLiD® Total RNA-Seq Kit (Life Technologies, Carlsbad, CA, USA) as described by the manufacturer. DNA libraries were deposited on slides and sequenced using the SOLiD v4 sequencing system (Life Technologies, Carlsbad, CA, USA) at the University of Manchester FBMH Genomic technologies core research facility or at BGI Genomics (Shenzhen, China).
Sequencing reads were mapped to the S. cerevisiae genome (genome assembly EF4 downloaded from ENSEMBL) using Bowtie  and counts for each transcript were calculated using HTseq . Counts were imported into edgeR where we used the Generalised Linear Model (GLM) in order to test for statistical differences between stresses . We performed three different statistical analyses: (1) ∆T represents differences in transcript abundance between stressed (s) and unstressed (u) samples (Ts/Tu;); (2) IP/T represent target-identification experiments, which were performed comparing the abundance of pull-down mRNAs with that of the total transcripts (IPs/Ts and IPu/Tu); and (3) ∆IP represents changes in target binding following stress (IPs/Ts/IPu/Tu). Unstressed IPu/Tu results were reported previously . Statistical tests involving corresponding IP and T samples were performed using a paired design. All gene lists were generated using FDR < 0.05 as a significance threshold for inclusion. Sequencing data of stress experiments have been deposited at ArrayExpress (E-MTAB-5836) . Data from unstressed experiments are deposited in ArrayExpress (E-MTAB-2464) .
GO analyses of RNA gene lists from edgeR files were done using the web-based GeneCodis tools (http://genecodis.cnb.csic.es/) to assess the yeast GO Slim functional enrichments of eIF4F associated mRNAs against a reference list of 5348 common yeast ORFs [64, 65]. All reported P values were corrected for multiple testing using the Benjamini–Hochberg FDR method. A FDR cut off of 0.05 was used.
Published ribosome profiling data [28, 29, 36] were obtained from the appropriate database deposited files and processed to determine TE, defined as the normalised mean number of ribosome protected fragments mapped to an ORF, divided by the number of transcripts per million reads for that ORF, and ∆TE defined as the change in TE in stress conditions divided by TE under unstressed conditions from the same datasets. The –aa and + H2O2 datasets were from duplicate experiments, while only a single –glu replicate was available as other samples in this study used either a different yeast strain or poly A selection rather than RNA-depletion. Poly A selection has been shown to have sequencing bias with yeast samples  so cannot be readily compared with our rRNA depleted data.
Initiation times calculated  for 4531 mRNAs were compared with unstressed IP/T data for eIF4E and eIF4G. Box-and-whisker plots were drawn using the BoxPlotR online tool (http://shiny.chemgrid.org/boxplotr/). Centre lines show the medians; box limits indicate the 25th and 75th percentiles as determined by R software; whiskers extend to 5th and 95th percentiles, outliers are represented by dots. The notches are defined as ± 1.58 × interquartile range/square root (n). This represents the 95% confidence interval for each median. Non-overlapping notches give roughly 95% confidence that two medians differ.
We thank Andrew Hayes and staff within the FBMH Genomic technologies core research facility for their assistance with RNA-seq data collection.
This work was supported by grant BB/G012571/1 from the UK Biotechnology and Biological Sciences Research Council (www.bbsrc.ac.uk). Additional support for core research facilities was provided by Wellcome Trust Institutional Strategic Support Fund awards (www.wellcome.ac.uk). The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.
Availability of data and materials
All data generated or analysed during this study are included in this published article and its supplementary information files. Sequencing data of stress experiments are available at ArrayExpress (E-MTAB-5836) . Data from unstressed experiments are at ArrayExpress (E-MTAB-2464) .
PFGS, MPA, CMG, SJH and GDP designed and led the study; JLC, CJK, LMC performed RIP-seq, polysome and western blot experiments; WR mapped seq reads and performed initial data analyses; DT and GDP performed main data analyses; GDP wrote the manuscript; all authors were involved in intellectual aspects of the study, and they edited and approved the final manuscript. All authors read and approved the final manuscript.
Ethics approval and consent to participate
This study did not require ethical approval.
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- Hinnebusch AG. The scanning mechanism of eukaryotic translation initiation. Annu Rev Biochem. 2014;83:779–812.View ArticlePubMedGoogle Scholar
- Gruner S, Peter D, Weber R, Wohlbold L, Chung MY, Weichenrieder O, et al. The structures of eIF4E-eIF4G complexes reveal an extended interface to regulate translation initiation. Mol Cell. 2016;64:467–79.View ArticlePubMedGoogle Scholar
- Preiss T, Hentze MW. Dual function of the messenger RNA cap structure in poly(A)-tail-promoted translation in yeast. Nature. 1998;392:516–20.View ArticlePubMedGoogle Scholar
- Dever TE, Kinzy TG, Pavitt GD. Mechanism and regulation of protein synthesis in Saccharomyces cerevisiae. Genetics. 2016;203:65–107.View ArticlePubMedPubMed CentralGoogle Scholar
- Thompson MK, Gilbert WV. mRNA length-sensing in eukaryotic translation: reconsidering the “closed loop” and its implications for translational control. Curr Genet. 2017;63:613–20.View ArticlePubMedGoogle Scholar
- Villa N, Do A, Hershey JW, Fraser CS. Human eukaryotic initiation factor 4G (eIF4G) protein binds to eIF3c, -d, and -e to promote mRNA recruitment to the ribosome. J Biol Chem. 2013;288:32932–40.View ArticlePubMedPubMed CentralGoogle Scholar
- Singh CR, Watanabe R, Chowdhury W, Hiraishi H, Murai MJ, Yamamoto Y, et al. Sequential eukaryotic translation initiation factor 5 (eIF5) binding to the charged disordered segments of eIF4G and eIF2beta stabilizes the 48S preinitiation complex and promotes its shift to the initiation mode. Mol Cell Biol. 2012;32:3978–89.View ArticlePubMedPubMed CentralGoogle Scholar
- Mitchell SF, Walker SE, Algire MA, Park EH, Hinnebusch AG, Lorsch JR. The 5′-7-methylguanosine cap on eukaryotic mRNAs serves both to stimulate canonical translation initiation and to block an alternative pathway. Mol Cell. 2010;39:950–62.View ArticlePubMedPubMed CentralGoogle Scholar
- Schwanhausser B, Busse D, Li N, Dittmar G, Schuchhardt J, Wolf J, et al. Global quantification of mammalian gene expression control. Nature. 2011;473:337–42.View ArticlePubMedGoogle Scholar
- Vogel C, Marcotte EM. Insights into the regulation of protein abundance from proteomic and transcriptomic analyses. Nat Rev Genet. 2012;13:227–32.PubMedPubMed CentralGoogle Scholar
- Lawless C, Holman SW, Brownridge P, Lanthaler K, Harman VM, Watkins R, et al. Direct and absolute quantification of over 1800 yeast proteins via selected reaction monitoring. Mol Cell Proteomics. 1800;2016(15):1309–22.Google Scholar
- Siwiak M, Zielenkiewicz P. A comprehensive, quantitative, and genome-wide model of translation. PLoS Comput Biol. 2010;6:e1000865.View ArticlePubMedPubMed CentralGoogle Scholar
- Costello J, Castelli LM, Rowe W, Kershaw CJ, Talavera D, Mohammad-Qureshi SS, et al. Global mRNA selection mechanisms for translation initiation. Genome Biol. 2015;16:10.View ArticlePubMedPubMed CentralGoogle Scholar
- Kershaw CJ, Costello JL, Castelli LM, Talavera D, Rowe W, Sims PF, et al. The yeast La related protein Slf1p is a key activator of translation during the oxidative stress response. PLoS Genet. 2015;11:e1004903.View ArticlePubMedPubMed CentralGoogle Scholar
- Kershaw CJ, Costello JL, Talavera D, Rowe W, Castelli LM, Sims PF, et al. Integrated multi-omics analyses reveal the pleiotropic nature of the control of gene expression by Puf3p. Sci Rep. 2015;5:15518.View ArticlePubMedPubMed CentralGoogle Scholar
- Castelli LM, Talavera D, Kershaw CJ, Mohammad-Qureshi SS, Costello JL, Rowe W, et al. The 4E-BP Caf20p mediates both eIF4E-dependent and independent repression of translation. PLoS Genet. 2015;11:e1005233.View ArticlePubMedPubMed CentralGoogle Scholar
- Thompson MK, Rojas-Duran MF, Gangaramani P, Gilbert WV. The ribosomal protein Asc1/RACK1 is required for efficient translation of short mRNAs. Elife. 2016;5:e11154.PubMedPubMed CentralGoogle Scholar
- Sen ND, Zhou F, Harris MS, Ingolia NT, Hinnebusch AG. eIF4B stimulates translation of long mRNAs with structured 5′ UTRs and low closed-loop potential but weak dependence on eIF4G. Proc Natl Acad Sci U S A. 2016;113:10464–72.View ArticlePubMedPubMed CentralGoogle Scholar
- Otero LJ, Ashe MP, Sachs AB. The yeast poly(A)-binding protein Pab1p stimulates in vitro poly(A)-dependent and cap-dependent translation by distinct mechanisms. EMBO J. 1999;18:3153–63.View ArticlePubMedPubMed CentralGoogle Scholar
- Jivotovskaya AV, Valasek L, Hinnebusch AG, Nielsen KH. Eukaryotic translation initiation factor 3 (eIF3) and eIF2 can promote mRNA binding to 40S subunits independently of eIF4G in yeast. Mol Cell Biol. 2006;26:1355–72.View ArticlePubMedPubMed CentralGoogle Scholar
- Spriggs KA, Bushell M, Willis AE. Translational regulation of gene expression during conditions of cell stress. Mol Cell. 2010;40:228–37.View ArticlePubMedGoogle Scholar
- Simpson CE, Ashe MP. Adaptation to stress in yeast: to translate or not? Biochem Soc Trans. 2012;40:794–9.View ArticlePubMedGoogle Scholar
- Pavitt GD, Ramaiah KV, Kimball SR, Hinnebusch AG. eIF2 independently binds two distinct eIF2B subcomplexes that catalyze and regulate guanine-nucleotide exchange. Genes Dev. 1998;12:514–26.View ArticlePubMedPubMed CentralGoogle Scholar
- Jennings MD, Zhou Y, Mohammad-Qureshi SS, Bennett D, Pavitt GD. eIF2B promotes eIF5 dissociation from eIF2*GDP to facilitate guanine nucleotide exchange for translation initiation. Genes Dev. 2013;27:2696–707.View ArticlePubMedPubMed CentralGoogle Scholar
- Jennings MD, Kershaw CJ, Adomavicius T, Pavitt GD. Fail-safe control of translation initiation by dissociation of eIF2alpha phosphorylated ternary complexes. Elife. 2017;6:e24542.View ArticlePubMedPubMed CentralGoogle Scholar
- Smirnova JB, Selley JN, Sanchez-Cabo F, Carroll K, Eddy AA, McCarthy JE, et al. Global gene expression profiling reveals widespread yet distinctive translational responses to different eukaryotic translation initiation factor 2B-targeting stress pathways. Mol Cell Biol. 2005;25:9340–9.View ArticlePubMedPubMed CentralGoogle Scholar
- Shenton D, Smirnova JB, Selley JN, Carroll K, Hubbard SJ, Pavitt GD, et al. Global translational responses to oxidative stress impact upon multiple levels of protein synthesis. J Biol Chem. 2006;281:29011–21.View ArticlePubMedGoogle Scholar
- Ingolia NT, Ghaemmaghami S, Newman JR, Weissman JS. Genome-wide analysis in vivo of translation with nucleotide resolution using ribosome profiling. Science. 2009;324:218–23.View ArticlePubMedPubMed CentralGoogle Scholar
- Gerashchenko MV, Lobanov AV, Gladyshev VN. Genome-wide ribosome profiling reveals complex translational regulation in response to oxidative stress. Proc Natl Acad Sci U S A. 2012;109:17394–9.View ArticlePubMedPubMed CentralGoogle Scholar
- Ashe MP, De Long SK, Sachs AB. Glucose depletion rapidly inhibits translation initiation in yeast. Mol Biol Cell. 2000;11:833–48.View ArticlePubMedPubMed CentralGoogle Scholar
- Castelli LM, Lui J, Campbell SG, Rowe W, Zeef LA, Holmes LE, et al. Glucose depletion inhibits translation initiation via eIF4A loss and subsequent 48S preinitiation complex accumulation, while the pentose phosphate pathway is coordinately up-regulated. Mol Biol Cell. 2011;22:3379–93.View ArticlePubMedPubMed CentralGoogle Scholar
- Arribere JA, Doudna JA, Gilbert WV. Reconsidering movement of eukaryotic mRNAs between polysomes and P bodies. Mol Cell. 2011;44:745–58.View ArticlePubMedPubMed CentralGoogle Scholar
- Hoyle NP, Castelli LM, Campbell SG, Holmes LE, Ashe MP. Stress-dependent relocalization of translationally primed mRNPs to cytoplasmic granules that are kinetically and spatially distinct from P-bodies. J Cell Biol. 2007;179:65–74.View ArticlePubMedPubMed CentralGoogle Scholar
- Buchan JR, Muhlrad D, Parker R. P bodies promote stress granule assembly in Saccharomyces cerevisiae. J Cell Biol. 2008;183:441–55.View ArticlePubMedPubMed CentralGoogle Scholar
- Simpson CE, Lui J, Kershaw CJ, Sims PF, Ashe MP. mRNA localization to P-bodies in yeast is bi-phasic with many mRNAs captured in a late Bfr1p-dependent wave. J Cell Sci. 2014;127:1254–62.View ArticlePubMedPubMed CentralGoogle Scholar
- Zid BM, O’Shea EK. Promoter sequences direct cytoplasmic localization and translation of mRNAs during starvation in yeast. Nature. 2014;514:117–21.View ArticlePubMedPubMed CentralGoogle Scholar
- Gasch AP, Spellman PT, Kao CM, Carmel-Harel O, Eisen MB, Storz G, et al. Genomic expression programs in the response of yeast cells to environmental changes. Mol Biol Cell. 2000;11:4241–57.View ArticlePubMedPubMed CentralGoogle Scholar
- Causton HC, Ren B, Koh SS, Harbison CT, Kanin E, Jennings EG, et al. Remodeling of yeast genome expression in response to environmental changes. Mol Biol Cell. 2001;12:323–37.View ArticlePubMedPubMed CentralGoogle Scholar
- Preiss T, Baron-Benhamou J, Ansorge W, Hentze MW. Homodirectional changes in transcriptome composition and mRNA translation induced by rapamycin and heat shock. Nat Struct Biol. 2003;10:1039–47.View ArticlePubMedGoogle Scholar
- Clarkson BK, Gilbert WV, Doudna JA. Functional overlap between eIF4G isoforms in Saccharomyces cerevisiae. PLoS One. 2010;5:e9114.View ArticlePubMedPubMed CentralGoogle Scholar
- Gross JD, Moerke NJ, von der Haar T, Lugovskoy AA, Sachs AB, McCarthy JE, et al. Ribosome loading onto the mRNA cap is driven by conformational coupling between eIF4G and eIF4E. Cell. 2003;115:739–50.View ArticlePubMedGoogle Scholar
- O’Leary SE, Petrov A, Chen J, Puglisi JD. Dynamic recognition of the mRNA cap by Saccharomyces cerevisiae eIF4E. Structure. 2013;21:2197–207.View ArticlePubMedGoogle Scholar
- Park EH, Zhang F, Warringer J, Sunnerhagen P, Hinnebusch AG. Depletion of eIF4G from yeast cells narrows the range of translational efficiencies genome-wide. BMC Genomics. 2011;12:68.View ArticlePubMedPubMed CentralGoogle Scholar
- Tamarkin-Ben-Harush A, Vasseur JJ, Debart F, Ulitsky I, Dikstein R. Cap-proximal nucleotides via differential eIF4E binding and alternative promoter usage mediate translational response to energy stress. Elife. 2017;6:e21907.View ArticlePubMedPubMed CentralGoogle Scholar
- Lahr RM, Fonseca BD, Ciotti GE, Al-Ashtal HA, Jia JJ, Niklaus MR, et al. La-related protein 1 (LARP1) binds the mRNA cap, blocking eIF4F assembly on TOP mRNAs. Elife. 2017;6:e24146.View ArticlePubMedPubMed CentralGoogle Scholar
- Bousquet-Antonelli C, Deragon JM. A comprehensive analysis of the La-motif protein superfamily. RNA. 2009;15:750–64.View ArticlePubMedPubMed CentralGoogle Scholar
- Gonatopoulos-Pournatzis T, Cowling VH. Cap-binding complex (CBC). Biochem J. 2014;457:231–42.View ArticlePubMedGoogle Scholar
- Garre E, Romero-Santacreu L, De Clercq N, Blasco-Angulo N, Sunnerhagen P, Alepuz P. Yeast mRNA cap-binding protein Cbc1/Sto1 is necessary for the rapid reprogramming of translation after hyperosmotic shock. Mol Biol Cell. 2012;23:137–50.View ArticlePubMedPubMed CentralGoogle Scholar
- Parker R. RNA degradation in Saccharomyces cerevisae. Genetics. 2012;191:671–702.View ArticlePubMedPubMed CentralGoogle Scholar
- Huch S, Nissan T. Interrelations between translation and general mRNA degradation in yeast. Wiley Interdiscip Rev RNA. 2014;5:747–63.View ArticlePubMedPubMed CentralGoogle Scholar
- Lee AS, Kranzusch PJ, Doudna JA, Cate JH. eIF3d is an mRNA cap-binding protein that is required for specialized translation initiation. Nature. 2016;536:96–9.View ArticlePubMedPubMed CentralGoogle Scholar
- Amrani N, Ghosh S, Mangus DA, Jacobson A. Translation factors promote the formation of two states of the closed-loop mRNP. Nature. 2008;453:1276–80.View ArticlePubMedPubMed CentralGoogle Scholar
- Searfoss A, Dever TE, Wickner R. Linking the 3′ poly(A) tail to the subunit joining step of translation initiation: relations of Pab1p, eukaryotic translation initiation factor 5b (Fun12p), and Ski2p-Slh1p. Mol Cell Biol. 2001;21:4900–8.View ArticlePubMedPubMed CentralGoogle Scholar
- Sinvani H, Haimov O, Svitkin Y, Sonenberg N, Tamarkin-Ben-Harush A, Viollet B, et al. Translational tolerance of mitochondrial genes to metabolic energy stress involves TISU and eIF1-eIF4GI cooperation in start codon selection. Cell Metab. 2015;21:479–92.View ArticlePubMedGoogle Scholar
- Archer SK, Shirokikh NE, Hallwirth CV, Beilharz TH, Preiss T. Probing the closed-loop model of mRNA translation in living cells. RNA Biol. 2015;12:248–54.View ArticlePubMedPubMed CentralGoogle Scholar
- Baum S, Bittins M, Frey S, Seedorf M. Asc1p, a WD40-domain containing adaptor protein, is required for the interaction of the RNA-binding protein Scp160p with polysomes. Biochem J. 2004;380:823–30.View ArticlePubMedPubMed CentralGoogle Scholar
- Rowe W, Kershaw CJ, Castelli LM, Costello JL, Ashe MP, Grant CM, et al. Puf3p induces translational repression of genes linked to oxidative stress. Nucleic Acids Res. 2014;42:1026–41.View ArticlePubMedGoogle Scholar
- Taylor EJ, Campbell SG, Griffiths CD, Reid PJ, Slaven JW, Harrison RJ, et al. Fusel alcohols regulate translation initiation by inhibiting eIF2B to reduce ternary complex in a mechanism that may involve altering the integrity and dynamics of the eIF2B body. Mol Biol Cell. 2010;21:2202–16.View ArticlePubMedPubMed CentralGoogle Scholar
- Langmead B, Trapnell C, Pop M, Salzberg SL. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol. 2009;10:R25.View ArticlePubMedPubMed CentralGoogle Scholar
- Anders S, Pyl PT, Huber W. HTSeq--a Python framework to work with high-throughput sequencing data. Bioinformatics. 2015;31:166–9.View ArticlePubMedGoogle Scholar
- Robinson MD, McCarthy DJ, Smyth GK. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics. 2010;26:139–40.View ArticlePubMedGoogle Scholar
- Costello J, Castelli LM, Kershaw CJ, Talavera D, Rowe W, Sims PF, et al. Effect of stress on the binding affinity of yeast eIF4F components to their target RNAs. In: ArrayExpress. 2017. https://www.ebi.ac.uk/arrayexpress/experiments/E-MTAB-5836.Google Scholar
- Costello J, Castelli LM, Rowe W, Kershaw CJ, Talavera D, Mohammad-Qureshi SS, et al. RIP-seq of Saccharomyces cerevisiae to identify RNAs which associate with members of the closed loop complex (eIF4E, eIF4GI, eIF4GII, Pab1p) and inhibtors of the complex (Caf20p, Eap1p). In: ArrayExpress. 2015. https://www.ebi.ac.uk/arrayexpress/experiments/E-MTAB-2464.Google Scholar
- Carmona-Saez P, Chagoyen M, Tirado F, Carazo JM, Pascual-Montano A. GENECODIS: a web-based tool for finding significant concurrent annotations in gene lists. Genome Biol. 2007;8:R3.View ArticlePubMedPubMed CentralGoogle Scholar
- Nogales-Cadenas R, Carmona-Saez P, Vazquez M, Vicente C, Yang X, Tirado F, et al. GeneCodis: interpreting gene lists through enrichment analysis and integration of diverse biological information. Nucleic Acids Res. 2009;37:W317–22.View ArticlePubMedPubMed CentralGoogle Scholar
- Weinberg DE, Shah P, Eichhorn SW, Hussmann JA, Plotkin JB, Bartel DP. Improved ribosome-footprint and mRNA measurements provide insights into dynamics and regulation of yeast translation. Cell Rep. 2016;14:1787–99.View ArticlePubMedPubMed CentralGoogle Scholar