Glucose is the preferred carbon source for most eukaryotes and so it is important that cells can sense and react rapidly to fluctuations in glucose levels. It is becoming increasingly clear that the regulation of gene expression at the post-transcriptional level is important in the adaptation to changes in glucose levels, possibly as this could engender more rapid alterations in the concentrations of key proteins, such as metabolic enzymes. Following the removal of glucose from yeast cells a rapid inhibition of translation is observed. As a consequence, mRNPs (messenger ribonucleoproteins) relocalize into cytoplasmic granules known as P-bodies (processing bodies) and EGP-bodies. mRNA decay components localize into P-bodies, and thus these assemblies are likely to represent sites where mRNAs are targeted for degradation. In contrast, EGP-bodies lack any decay components and contain the eukaryotic translation initiation factors eIF4E, eIF4G and Pab1p, as well as other RNA-binding proteins. Therefore EGP-bodies probably constitute sites where mRNAs are earmarked for storage. So, it is possible that cells distinguish between transcripts and target them to either P-bodies or EGP-bodies depending on their functional value. The localization of mRNAs into these granules following glucose starvation may serve to preserve mRNAs that are involved in the diauxic shift to ethanol growth and entry into stationary phase, as well as to degrade mRNAs that are solely involved in glucose fermentation.
- glucose derepression
- mRNA localization
- processing body (P-body)
- stress granule
- translation control
Nutrient signalling is intrinsically linked to cellular growth and the ability to sense fluctuations in nutrient levels allows cells to adapt to changing environments to ensure cell survival. Glucose is the most abundant monosaccharide in nature and is a major energy source for most eukaryotes. For the yeast Saccharomyces cerevisiae glucose is the preferred carbon source and, like most organisms, yeast possess signalling pathways that monitor glucose levels and alter gene expression accordingly (reviewed in ).
Adapting to the depletion of glucose is a vital process for cells. Under such nutrient-limiting conditions yeast cells stop proliferating and enter a G0-phase of the cell cycle, commonly known as the stationary phase . This allows the maintenance of cell viability during times of stress until more favourable conditions arise. A switch from normal vegetative growth to a filamentous form of growth has also been described in haploid yeast under glucose-limiting conditions . This growth pattern is believed to be a mechanism for nutrient foraging. In addition, glucose starvation has an impact on the cytoskeleton in yeast causing a rapid depolarization of actin structures .
However, responses to limiting glucose levels most notably culminate in profound changes at the transcriptional level. For instance, following periods of glucose depletion, an estimated 30% of yeast genes become transcriptionally derepressed . These genes have functions in the utilization of alternative carbon sources, gluconeogenesis and respiratory metabolism. Moreover, a major down-regulation of most housekeeping genes has been observed . Nevertheless, it is becoming increasingly clear that cells also utilize an arsenal of post-transcriptional control mechanisms in response to alterations in glucose levels.
Effect of glucose starvation on translation initiation
Polyribosome analysis shows that withdrawal of glucose leads to the disassembly of polysomes and an accumulation of monosomes in a profile characteristic of the inhibition of translation initiation  (Figure 1B). Upon re-addition of glucose, translation has been shown to resume as normal. The inhibition of translation initiation following glucose starvation represents one of the most severe and rapid global down-regulatory effects on protein synthesis that has thus far been studied in yeast .
A conserved mechanism for translation control has previously been described involving eIF2α (eukaryotic translation initiation factor 2α) kinases. Indeed the higher eukaryotic eIF2α kinase PERK [PKR (double-stranded-RNA-dependent protein kinase)-like endoplasmic reticulum kinase] is activated in low-glucose conditions in pancreatic β-cells . In yeast, amino acid starvation activates the only yeast eIF2α kinase Gcn2p by increasing the level of uncharged tRNAs. Gcn2p phosphorylates the α-subunit of the translation initiation factor eIF2 to ultimately lead to an inhibition of translation initiation. An outcome of this mechanism is an increase in the translation of particular mRNAs, such as ATF4 (activating transcription factor 4) in mammalian cells  and GCN4 in yeast , via complex mechanisms involving upstream open reading frames. However, the inhibition of translation initiation in response to glucose starvation in yeast does not operate via this pathway. Cells with a mutation in eIF2α rendering it non-phosphorylatable are still completely inhibited at the translational level following glucose starvation, even though these strains are translationally resistant to the effects of amino acid starvation . Another mechanism by which translation initiation can be globally down-regulated is via the activation of 4E-BPs (eIF4E-binding proteins), which inhibit the eIF4E–eIF4G interaction to prevent the formation of pre-initiation complexes. Yeast harbours two characterized 4E-BPs, Caf20p and Eap1p. However, strains mutant in either of these genes still show complete inhibition of translation initiation after glucose depletion [M. Ashe, unpublished work]. Furthermore, the integrity of the eIF4E–eIF4G complex is not altered by glucose starvation .
Investigation of a number of mutants in the glucose signalling pathways suggested that elements of these pathways may play a role in glucose-mediated translational control . Glucose repression is one of the major signal transduction pathways in S. cerevisiae . The Snf1p protein kinase complex plays a central role in controlling glucose repression as it regulates a number of transcription repressors that inhibit expression of a large number of genes (Figure 1A). The Snf1p kinase complex itself is regulated by a protein phosphatase complex, Glc7p–Reg1p  and the hexokinase Hxk2p, which inhibits Snf1p activity in the presence of glucose . Therefore strains mutant in REG1 or HXK2 harbour constitutively active Snf1p kinase and are constitutively derepressed in terms of the transcriptional glucose repression pathway. These strains also do not exhibit an inhibition of translation in response to glucose starvation  (Figure 1B). Furthermore, mutation of SNF1 in these mutants restores the inhibitory effects of glucose depletion on translation. In addition, mutation of the genes for the transcription repressors Mig1p and Mig2p, generates constitutively glucose-derepressed yeast, that are also resistant to the translational inhibition caused by glucose depletion (Figure 1B). Therefore these results suggest that the glucose-repressed state is a prerequisite for the control of translation in response to glucose starvation.
The cAMP-dependent PKA (protein kinase A) pathway also plays a role in the response to different glucose levels in yeast . Mutant strains with low PKA activity continue translation following glucose starvation. Therefore, it appears that high PKA activity is also a prerequisite for the inhibition of translation caused by glucose withdrawal. The precise mechanism for signalling to the translation apparatus and for the dramatic inhibition of translation are currently unknown, although it does appear that glucose starvation targets the formation of the 48S pre-initiation complex on the mRNA .
P-bodies (processing bodies) form as a result of glucose starvation
Another consequence of glucose depletion is that following the rapid inhibition of translation initiation, cytoplasmic granules known as P-bodies form (Figure 2). P-bodies are highly conserved structures seen in both lower and higher eukaryotes. Although their composition varies between organisms, they contain a core set of proteins involved in the 5′-to-3′ mRNA degradation pathway. These include the decapping enzymes Dcp1p and Dcp2p, the activators of decapping Dhh1p, Edc3p, Lsm1p and Pat1p, and the 5′-to-3′ exoribonuclease Xrn1p [15,16]. Factors of the nonsense-mediated mRNA decay pathway have also been found to cycle into P-bodies in yeast , whereas in higher eukaryotes components of the ARE (adenine-uracil-rich element)-mediated mRNA decay and miRNA (microRNA) pathways have also been identified in P-bodies [18,19]. More recently it has been shown that catalytic subunits of PKA, a key component in the glucose signalling response, also accumulate in P-bodies .
The core components of P-bodies are involved in the decapping and degradation of mRNA. Stabilization of mRNAs in polysomes reduces the pool of mRNA available for decapping and leads to the disappearance of P-bodies, whereas trapping mRNA in an intermediate state of decay in mRNA decay mutants, e.g. in xrn1 mutants, leads to increased P-body formation . These results are consistent with the idea that the size and abundance of P-bodies is dependent on the flux of mRNA.
Since stresses, such as glucose starvation, lead to the formation of P-bodies, it is thought that mRNPs (messenger ribonucleoproteins) are targeted to these foci following translational arrest. This idea is supported by the observation that not only do mRNAs accumulate in P-bodies [11,21], but P-body assembly is dependent on RNA, as shown by a loss of P-bodies in response to RNase A treatment . In addition, regions within the P-body factors Edc3p and Lsm4p have been shown to account for the aggregation of mRNPs that result in P-bodies . Many factors found in P-bodies have been shown to repress translation, such as the activators of decapping, Dhh1p and Pat1p . Furthermore, when cells are restored to unstressed conditions, some mRNAs can re-enter the translation pool by returning to polysomes from P-bodies .
Interestingly, mutants in a variety of components of the 5-to-3′ mRNA decay pathway exhibit resistance to the translation regulation that is normally observed following glucose starvation . However, these strains are also resistant to translation inhibition caused by amino acid starvation, which does not lead to P-body formation. Hence we have speculated previously that the translational resistance of the mRNA decay mutants relates to increased mRNA levels, which overwhelm the translational control mechanisms acting by mass action.
EGP-bodies are sites of mRNA storage
Initially it was believed that P-bodies lacked any translation initiation factors ; however, subsequent studies found that following longer periods of glucose depletion the translation initiation factors eIF4E, eIF4G and the poly(A)-binding protein Pab1p formed in P-bodies (Figure 2) [11,25]. eIF4E and eIF4G are components of the eIF4F complex that associate with the 5′ cap of mRNA and, in conjunction with Pab1p, forms a closed-loop mRNA complex . mRNA degradation and translation are in constant competition and thus formation of a closed-loop complex appears to protect the mRNA from decay, as well as promote translation initiation [8,27]. Therefore it is probable that the localization of eIF4E, eIF4G and Pab1p into P-bodies may aid the short-term storage of mRNAs that have been released from polysomes.
These closed-loop complex factors not only accumulate into P-bodies, but they also form into spatially distinct granules that have been called both ‘EGP-bodies’  and ‘yeast stress granules’  (Figure 2). Given the compositional distinction between these granules and stress granules from other organisms, and the fact that yeast stress granules with a similar composition to mammalian stress granules have recently been described following heat-stress , we use the name EGP-bodies for the granules induced by glucose starvation. EGP-bodies appear in response to longer periods of perturbation and lack any mRNA degradation factors. The fact that these granules form after a more prolonged period of stress could mean that EGP-bodies are sites for the longer term storage of mRNA . The kinetics of eIF4E, eIF4G and Pab1p localization suggests that following glucose starvation at least some mRNAs stay associated with polysomes and are only released much later to P-bodies and EGP-bodies . However, the precise source of the mRNAs that accumulate early in P-bodies compared with those that enter both P-bodies and EGP-bodies at later time points are currently unknown. It is possible that the early mRNAs represent those recently exported from the nucleus or somehow dissociated from translation initiation factors, whereas the later mRNA relocalization, along with that of eIF4E, eIF4G and Pab1p, represents those mRNAs that were very actively translated prior to the glucose starvation.
Until recently the only known components of EGP-bodies were eIF4E, eIF4G, Pab1p and mRNA. However, investigations into a number of RNA-binding proteins have found that Pbp1p, Pub1p and Ngr1p also accumulate in P-bodies and EGP-bodies . Interestingly these RNA-binding proteins are orthologues of the mammalian stress granule components ataxin-2, TIA-1 (T-cell-restricted intracellular-antigen-1) and TIAR (TIA-1-related).
Stress granules and EGP-bodies
Stress granules are similar structures that function in mRNA storage following cellular perturbations and have been well characterized in mammalian cells. They arise in response to translational arrest as a result of environmental stresses and they contain a core set of proteins, which includes the translation initiation factors eIF3, eIF4E and eIF4G, the 40S ribosomal subunit and PABP [poly(A)-binding protein] . Consequently stress granules are believed to be sites where stalled 43S pre-initiation complexes or the remnants of these complexes are localized following translational arrest . Similar to P-bodies, stabilizing or inducing the disassembly of polysomes has a direct effect on the formation of stress granules, which is indicative that mRNAs can move between polysomes and stress granules . Moreover, time-lapse microscopy illustrated that stress granules and P-bodies dynamically interact with each other . These studies therefore led to the model that stress granules are sites where mRNAs are sorted to be stored, re-enter translation or enter P-bodies for degradation .
Mammalian stress granules also contain a number of RNA-binding proteins, such as TIA-1 and TIAR, which are responsible for their assembly due to their prion-like properties . Similarly the yeast counterparts Pub1p and Ngr1p also appear to be involved in the assembly of EGP-bodies . However, EGP-bodies lack some of the core components of mammalian stress granules and cannot be viewed as sites where stalled 43S complexes accumulate due to the absence of the 40S ribosome from EGP-bodies. Recently, heat-shock granules that are compositionally more similar to the mammalian stress granules have been identified in S. cerevisiae . In response to severe temperatures, eIF3a and the 40S ribosomal subunit form foci that co-localized with other core mammalian stress granule components, such as eIF4G and Pab1p. In addition, these granules also co-localized with Dcp1p foci, which is concurrent with the mammalian stress granule model in which stress granules and P-bodies dock and interact with one another .
Potential for regulation by mRNA triage into different cytoplasmic granules
The widespread mechanisms that cells employ to regulate gene expression following glucose starvation are evident. For the vast majority of mRNAs, a rapid inhibition of translation initiation is observed following glucose depletion [34,35]. However, specific mRNAs, such as SDH2 and SUC2 are stable at low glucose. At higher glucose concentrations these mRNAs are rapidly degraded [36–38]. In contrast, gluconeogenic mRNAs, which are also stable in low glucose, are rapidly degraded following further reductions in glucose levels . mRNA translation can also be activated in response to nutrient depletion, via cap-independent mechanisms . For instance, IRESs (internal ribosomal entry sites) account for the translation of mRNAs encoding proteins involved in invasive growth and translation factors following glucose depletion .
P-bodies and EGP-bodies arise in response to glucose depletion and appear to primarily function in mRNA degradation and storage respectively. Therefore this raises the question as to whether it is possible that cells employ a mechanism that distributes mRNAs into these granules depending on the potential future value of these transcripts. As glucose is depleted, yeast cells enter the diauxic shift where they rapidly employ mechanisms to conserve energy and prepare the cell for growth on alternative carbon sources, such as ethanol. During this period, it would be beneficial for cells to protect valuable transcripts in cytoplasmic granules, where they are primed for translation, yet degrade transcripts that are either excessively abundant or are not required during growth on other carbon sources. Precisely how cells would control the differential localization of mRNAs to these granules is still unclear. It is possible that different combinations of mRNA-binding proteins and mRNA features could predispose an mRNA to a particular fate.
A variety of cellular stresses lead to a rapid inhibition of translation initiation. As well as preserving energy, such an inhibition provides cells with a period where existing mRNAs and proteins can be degraded and new programmes of gene expression can be initiated at the transcriptional level. The discovery of EGP-bodies opens up the possibility that this respite period can be used to select and preserve valuable mRNAs, while allowing those mRNAs that are surplus to requirements to be degraded in P-bodies.
This work was supported by the Wellcome Trust [grant number 088141/Z/09/Z (to M.P. A.)]. J.L. was supported by a Biotechnology and Biological Sciences Research Council doctoral training award.
We thank G. Pavitt and L. Castelli for their comments on the manuscript prior to submission, and M. Carlson and F. Winston for yeast strains.
RNA UK 2010: An Independent Meeting held at The Burnside Hotel, Cumbria, U.K., 22–24 January 2010. Organized and Edited by Jeremy Brown and Nick Watkins (Newcastle, U.K.).
Abbreviations: eIF, eukaryotic translation initiation factor; 4E-BP, eIF4E-binding protein; mRNP, messenger ribonucleoprotein; P-body, processing body; PKA, protein kinase A; TIA-1, T-cell-restricted intracellular-antigen-1; TIAR, TIA-1-related
- © The Authors Journal compilation © 2010 Biochemical Society