![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
Vol. 19, Issue 7, 2995-3007, July 2008
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||
The University of Manchester, Faculty of Life Sciences, Manchester M13 9PT, United Kingdom
Submitted November 26, 2007;
Revised March 31, 2008;
Accepted April 9, 2008
Monitoring Editor: Thomas Fox
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
subunit of eukaryotic initiation factor (eIF)2 by the Gcn2 kinase. Activation of Gcn2 is mediated by uncharged tRNA because mutation of its HisRS domain abolishes regulation in response to hydroperoxides. Furthermore, Gcn4 is translationally up-regulated in response to H2O2, and it is required for hydroperoxide resistance. We used transcriptional profiling to identify a wide range of genes that mediate this response as part of the Gcn4-dependent H2O2-regulon. In contrast to hydroperoxides, regulation of translation initiation in response to cadmium and diamide depends on both Gcn2 and the eIF4E binding protein Eap1. Thus, the response to oxidative stress is mediated by oxidant-specific regulation of translation initiation, and we suggest that this is an important mechanism underlying the ability of cells to adapt to different oxidants. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
; Gutteridge, 1994). To protect against oxidant damage, cells contain effective defense mechanisms, including antioxidant enzymes and free radical scavengers (Temple et al., 2005). It is now well-established that yeast cells can adapt to an oxidative stress by altering global transcription patterns, including genes encoding antioxidants and other metabolic enzymes (Gasch et al., 2000; Causton et al., 2001). Additionally, they respond by invoking complex regulatory mechanisms that include global inhibition of translation. For example, we have shown that exposure of yeast cells to hydrogen peroxide results in a rapid and reversible inhibition of protein synthesis (Shenton et al., 2006). This reduction in protein synthesis may allow time for the turnover of existing mRNAs and proteins while gene expression is reprogrammed to deal with the stress.
The initiation phase of protein synthesis is the main target of regulation, and it represents a key control point for eukaryotic gene expression (Hinnebusch, 2000). Phosphorylation of eukaryotic initiation factor (eIF) 2 is important for this control in response to diverse stress conditions. eIF2 is a guanine nucleotide binding factor, which in its GTP-bound form interacts with the initiator methionyl-tRNA (Met-tRNAiMet) to form a ternary complex that is competent for translation initiation. After each round of initiation, eIF2 is released from the ribosome as a binary complex with guanosine diphosphate (GDP). GDP is replaced by GTP in a guanine-nucleotide exchange reaction catalyzed by eIF2B. Met-tRNAiMet can only bind the eIF2/GTP complex, so translational control can be regulated by the activity of eIF2B. In both yeast and mammals, this is achieved by phosphorylation of the subunit of eIF2 at a conserved serine (Ser51) residue (Pavitt et al., 1998; Harding et al., 2000). Phosphorylation converts eIF2 from a substrate to a competitive inhibitor of the guanine nucleotide exchange factor eIF2B and the resulting decrease in eIF2B activity leads to reduced ternary complex levels, which inhibits translation initiation (Pavitt et al., 1998).
Four mammalian kinases have been identified that inhibit translation initiation by phosphorylating eIF2. These eIF2 kinases are regulated independently in response to various different cellular stresses (Dever, 2002; Proud, 2005). For example, PKR-like endoplasmic reticulum eIF2 kinase (PERK) has been found in all multicellular eukaryotes, and it is a component of the unfolded protein response. Consistent with its central role in the endoplasmic reticulum (ER) stress response, cells lacking PERK fail to phosphorylate eIF2, and they do not down-regulate protein synthesis during ER stress conditions (Bertolotti et al., 2000). Attenuating protein synthesis may act to reduce the burden of newly synthesized ER client proteins on the ER folding machinery. Additionally, eIF2 phosphorylation induces translation of specific mRNAs, such as that encoding the metazoan activating transcription factor 4 (ATF4) (Lu et al., 2004; Vattem and Wek, 2004). ATF4 mediates the "integrated stress response" whose targets include genes encoding proteins involved in amino acid metabolism and resistance to oxidative stress, ultimately protecting against the deleterious consequences of ER oxidation (Harding et al., 2003).
In the yeast Saccharomyces cerevisiae, Gcn2 is the sole eIF2 kinase (Dever, 2002). It is activated in response to a variety of conditions, including nutrient starvation (amino acids, purines, and glucose) and exposure to sodium chloride, rapamycin, and volatile anesthetics (Hinnebusch, 2005; Palmer et al., 2005). Depletion of amino acids leads to an accumulation of uncharged tRNA, which activates the Gcn2 protein kinase via its HisRS-related domain. It is likely that other stress conditions ultimately affect the levels of uncharged tRNA in the cell. For example, volatile anesthetics inhibit amino acid uptake (Palmer et al., 2005), and Gcn2 is activated by glucose starvation partly through an effect on vacuolar amino acid pools (Yang et al., 2000). Phosphorylation of eIF2 by Gcn2 reduces global protein synthesis, but it also enhances translation of the GCN4 mRNA. Translation of the GCN4 mRNA is activated in response to low ternary complex levels in a mechanism involving four short upstream open reading frames (Hinnebusch, 2005). Gcn4 is itself a transcription factor that activates gene expression of many targets, including amino acid biosynthetic genes (Natarajan et al., 2001). Thus, analogous to the mammalian integrated stress response, activation of Gcn4 serves to overcome the imposed amino starvation, which initially led to the translational induction of GCN4 expression.
We have previously analyzed the regulation of protein synthesis in response to oxidative stress induced by exposure to hydrogen peroxide (Shenton et al., 2006). H2O2 induces a dose-dependent inhibition of protein synthesis. This inhibition primarily occurs at the level of translation initiation, and it is regulated by Gcn2-mediated phosphorylation of eIF2. In this current study, we have extended this analysis to include reagents that induce the formation of different ROS. This is important because an oxidative stress can be caused by many different ROS with differing reactivities. We induced oxidative stress by exposing cells to cumene hydroperoxide (CHP), cadmium sulfate (hereafter referred to as cadmium), and diamide. Our data show that phosphorylation of eIF2 is a common response that inhibits translation initiation in response to oxidative stress caused by diverse ROS, but there are additional oxidant-specific regulatory mechanisms. Furthermore, Gcn4 is specifically up-regulated in response to H2O2 and the Gcn4-dependent H2O2 regulon is required for resistance to hydroperoxides, but not to other oxidants.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
) at 30°C and 180 rpm. Media were solidified by the addition of 2% (wt/vol) agar. Stress sensitivity was determined by growing cells to stationary phase and spotting onto agar plates containing various concentrations of oxidants.
|
). Plasmids containing EAP1 and eap1m3 have been described in Ibrahimo et al. (2006).
Western Blot Analysis
Protein extracts were electrophoresed under reducing conditions on SDS-polyacrylamide gel electrophoresis minigels and electroblotted onto polyvinylidene difluoride membrane (GE Healthcare, Chalfont St. Giles, United Kingdom). Blots were probed using eIF2 and phosphospecific eIF2 antibodies as described previously (Holmes et al., 2004). Myc-tagged proteins were detected using anti-myc antibodies (Roche Applied Science, Indianapolis, IN).
Analysis of Protein Synthesis
The rate of protein synthesis was measured in exponential phase cells treated with various concentrations of cumene hydroperoxide, diamide, or cadmium sulfate. Cells were treated with oxidants for 15 min and pulse-labeled for the last 5 min of the treatment with 85 µM L-[35S]cysteine/methionine as described previously (Shenton and Grant, 2003). For the analysis of ribosome distribution on sucrose density gradients, yeast cultures were grown to exponential phase and treated with oxidants for 15 min. Extracts were prepared in 100 µg of cycloheximide/ml and layered onto 15–50% sucrose gradients. The gradients were sedimented via centrifugation at 40,000 rpm in a Beckman ultracentrifuge for 2.5 h, and the A254 measured continuously to give the traces shown, as described previously (Ashe et al., 2000). Monosome and polysome peaks were quantified using the National Institutes of Health ImageJ software (http://rsb.info.nih.gov/ij/).
Microarray Hybridizations and Data Analysis
Yeast cells were grown in triplicate to midexponential phase in minimal SD media. The preparation of RNA, probes, and hybridization to whole yeast genome microarrays (Yeast Genome 2.0 array) was performed as described previously (Wishart et al., 2005). The complete data sets are publicly available at ArrayExpress (http://www.ebi.ac.uk/arrayexpress; accession number E-MEXP-998). Arrays that passed outlier data-quality assessment using dChip software (Li and Wong, 2001) were normalized with robust multichip average (Bolstad et al., 2003). Statistical tests were performed with limma by using the lmFit and eBayes functions (Smyth, 2004). Gene lists of differentially expressed genes were controlled for false discovery rate (fdr) errors by using the method of QVALUE using default settings (Storey and Tibshirani, 2003). Microarray statistical procedures were performed with Bioconductor (Gentleman et al., 2004). Principal components analysis was performed using a covariance dispersion matrix with Partek Genomics Suite, version 6.3 (Partek, St. Charles, MO). Clustering was performed using a k-means clustering algorithm ("Slope" similarity metric "Super Grouper" plugin of maxdView software (available from http://bioinf.man.ac.uk/microarray/maxd/). Clustering was performed on the means of each sample group (log 2) that had been z-transformed (for each probe set the mean set to zero, SD to 1). For each cluster, functional enrichment was determined using FunSpec (Robinson et al., 2002). Data were verified by real-time reverse transcription-polymerase chain reaction (RT-PCR) by using the MyIQ single-color real-time PCR detection system and iScript SYBR Green Supermix (Bio-Rad Laboratories).
We compared our data set of genes that were differentially expressed in response to H2O2 in wild-type versus the gcn4 deletion mutant with amino acid starvation microarray data published by Natarajan et al. (2001). Because the input for QVALUE is simply a list of p values, it was possible to calculate q values for the microarray data of Natarajan et al. (2001) by using their own p values. Given the array platform differences, we only considered data where the gene is present on both array platforms.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
1 mM (Shenton et al., 2006). Exposure to diamide caused a dose-dependent inhibition of protein synthesis at concentrations between 1 and 4 mM (Figure 1A). Exposure to the heavy metal cadmium also inhibited protein synthesis with 50% inhibition observed at concentrations between 0.1 and 1.0 mM (Figure 1A). In contrast to the other oxidants, inhibition with cadmium was not dose dependent. The reason for this is unclear, but it may reflect differences in the way in which cadmium is taken up by cells or is detoxified, compared with other oxidants. For example, treatments with different doses of cadmium may result in similar intracellular concentrations of cadmium because cadmium requires active uptake, compared with other oxidants such as hydrogen peroxide, which are freely diffusible.
|
Oxidative Stress Induces Gcn2-dependent eIF2 Phosphorylation
Oxidative stress caused by exposure to H2O2 induces Gcn2-mediated phosphorylation of eIF2 (Shenton et al., 2006). To test whether the translational inhibition observed in response to other oxidants relies upon this pathway, we examined eIF2 phosphorylation by immunoblot analysis. Oxidant concentrations were chosen that caused a substantial redistribution of polysome profiles (0.1 mM CHP, 0.2 mM cadmium, and 4.0 mM diamide). An increase in phosphorylation was observed in response to all three oxidants compared with the untreated control (Figure 2A). No phosphorylation of eIF2 was observed in response to any of the oxidants in a gcn2 mutant (Figure 2B), confirming that Gcn2-mediated phosphorylation of eIF2 is a common response to oxidative stress.
|
10% in the wild-type strain to 35% in the gcn2 mutant after treatments with H2O2 or CHP. Protein synthesis was, however, still inhibited in the gcn2 mutant, despite their being no inhibition of translation initiation detected by polysome analysis. This is consistent with our previous observation that H2O2 inhibits translation at both the initiation and postinitiation (elongation or termination) phases of protein synthesis (Shenton et al., 2006). These data indicate that Gcn2 is required for the inhibition of translation initiation in response to peroxides, but translation is still inhibited at the postinitiation phase in a gcn2 mutant.
|
90% in the wild-type strain compared with an approximate 33% reduction in the gcn2 mutant in response to cadmium or diamide (Figure 3A). These data indicate that Gcn2 is the major factor that inhibits translation initiation in response to cadmium or diamide, but translation initiation can still be inhibited in response to these oxidants in a gcn2 mutant. Furthermore, little or no restoration of protein synthesis was observed in the gcn2 mutant treated with cadmium or diamide compared with the untreated condition (Figure 3B). These data indicate that in contrast to hydroperoxides, cadmium and diamide inhibit translation initiation via both a Gcn2-dependent and Gcn2-independent mechanism.
Activation of Gcn2 in Response to Oxidative Stress Conditions
Gcn2 can be activated in response to diverse stress conditions. Depletion of amino acids leads to an accumulation of uncharged tRNA, which activates the Gcn2 protein kinase via its HisRS-related domain. Alternatively, rapamycin stimulates eIF2 phosphorylation by Gcn2 via dephosphorylation of Ser577 in Gcn2 (Hinnebusch, 2005). To determine whether oxidants activate Gcn2 via these same pathways, we analyzed translational activity in strains containing a mutation in the HisRS domain that abolishes tRNA binding (gcn2-m2; Garcia-Barrio et al., 2002) or containing a mutation in Ser577 (gcn2-S577A; Cherkasova and Hinnebusch, 2003). The Ser577 mutant has been shown to dampen the effects of rapamycin on eIF2 phosphorylation, suggesting that Gcn2 activation by rapamycin involves Ser577 dephosphorylation. Translation inhibition in response to H2O2 or CHP was unaffected by mutation of Ser577 (Figure 4). In contrast, no inhibition was observed in response to hydroperoxides in the gcn2-m2 mutant, indicating that hydroperoxides activate Gcn2 via binding of uncharged tRNA. Similarly, immunoblot analysis revealed that the phosphorylation of eIF2 observed in response to hydroperoxides is dependent on the m2 motif but is unaffected by mutation of Ser577 (Figure 5A). The experiments shown in Figures 4 and 5A were performed using a prototrophic strain (SCY51), ruling out any amino acid starvation arising as a result of inhibiting the uptake of an auoxtrophic amino acid.
|
|
background) strain used for these experiments is generally more resistant to oxidative stress than the W303 strain used for previous experiments; hence, higher concentrations of cadmium (1.0 mM) and diamide (10 mM) have been used. Quantification of polysome traces revealed that cadmium and diamide decreased the ratio of polysomes:monsomes (p:m) in the gcn2-m2 mutant by 33%, similar to the reduction in the gcn2 deletion mutant (Figure 2A). Phosphorylation of eIF2 was observed in the gcn2-S577A mutant, but not in the gcn2-m2 mutant, confirming that cadmium and diamide inhibit translation initiation independently of eIF2 phosphorylation (Figure 5B). Furthermore, inhibition of translation initiation was observed in response to cadmium or diamide in a mutant containing a nonphosphorylatable allele of eIF2 (sui2-S51A), whereas, no inhibition was observed in response to H2O2 (data not shown). These data further indicate that cadmium and diamide can inhibit translation initiation via a mechanism that does not involve Gcn2-mediated phosphorylation of eIF2.
EAP1 Is Required for Inhibition of Translation Initiation in Response to Cadmium and Diamide
Translation initiation can be controlled by regulating the availability of the cap-binding factor eIF4E. We therefore examined whether cadmium and diamide inhibit translation initiation through a mechanism that depends on the eIF4E-binding proteins (4EBPs) Caf20 and Eap1. No requirement was found for Caf20 to mediate translation in response to cadmium or diamide (data not shown). In contrast, Eap1 is partially required for the inhibition of translation initiation in response to both oxidants. The inhibition of translation initiation in response to H2O2 was unaffected in the eap1 mutant (data not shown). Loss of EAP1 or GCN2 partially abrogated the inhibition of translation initiation in response to cadmium or diamide (Figure 6A). Furthermore, inhibition was abolished in a gcn2 eap1 double mutant, indicating that Gcn2 and Eap1 mediate the inhibition of translation initiation in response to both cadmium and diamide (Figure 6A). Protein synthesis was still inhibited in the eap1 gcn2 mutant (Figure 3B), further confirming that oxidants inhibit translation at both the initiation and postinitiation (elongation or termination) phases of protein synthesis (Shenton et al., 2006).
|
). A plasmid containing this allele was transformed into an eap1 null and translational activity examined after treatment with cadmium (Figure 6B). Inhibition of translation initiation was similar in the eap1m3mutant to the vector control (see Figure 6B p:m ratios) consistent with the idea that Eap1-eIF4E binding is required for the control of translation initiation in response to cadmium. To further confirm this finding we examined a gcn2 eap1m3double mutant (Figure 6B). Inhibition of translation initiation was largely abrogated in response to cadmium stress, comparable with the gcn2 eap1 null mutant (Figure 6A).
Translational Induction of GCN4 Is Not a General Response to Oxidative Stress
To further characterize the role of translational control in oxidative stress, we examined the sensitivity of general control mutants to oxidants. Mutants lacking GCN4 were sensitive to H2O2 and CHP, indicating that Gcn4 is required for tolerance to hydroperoxides (Figure 7A). The sensitivity of the gcn4 mutant to peroxides is comparable with that of other yeast antioxidant mutants in this same strain background such as those lacking the YAP1 transcription factor or unable to synthesize glutathione (Grant et al., 1996). In contrast, the gcn4 mutant was unaffected in resistance to cadmium or diamide. The gcn2 mutant was no more sensitive than the wild type to ROS in spot tests.
|
). This result is complicated, because the reporter construct only contains the first 56 codons of GCN4 fused to a large heterologous reporter gene (lacZ). GCN4-lacZ therefore may not accurately reflect translational control of GCN4, particularly for oxidants, which inhibits translation at a postinitiation phase (Shenton et al., 2006). To overcome this problem, we measured GCN4 expression by using an epitope-tagged version of GCN4, which contains the entire coding region of GCN4-fused to a small Myc-tag (Lipford et al., 2005). Cells were treated with various concentrations of H2O2, cadmium and diamide for 2 h and Gcn4 levels determined by immunoblot analysis. Elevated GCN4 expression was observed in response to H2O2 at concentrations of 1.0 and 2.0 mM (Figure 7B). Induction in response to amino acid starvation is included as a positive control, and it indicates that Gcn4 is induced to higher levels than for peroxide treatments (Figure 7B). To confirm that the induction of GCN4 expression seen in response to hydrogen peroxide is a translational control response, expression was examined in a mutant lacking GCN2 (Figure 7B). No induction was observed in the gcn2 mutant in response to amino starvation or hydrogen peroxide indicating that phosphorylation of eIF2 is required for this response. In contrast to peroxide, no induction was observed in response to cadmium or diamide and Gcn4 seemed to be somewhat lower compared with the untreated control after treatment with these oxidants (Figure 7C). Lowered levels of Gcn4-myc protein are unlikely to arise due to degradation in the cell extracts isolated from oxidant-treated cells because trichloroacetic acid extraction was used to denature proteases before cell breakage. Together, these data indicate that Gcn4 is specifically required for the response to hydroperoxides.
Identification of the Gcn4-dependent H2O2 Regulon
To further understand the role of the Gcn4 pathway in the response to oxidative stress, we analyzed the global transcriptional response to H2O2. Wild-type, gcn2, and gcn4 mutant cells were analyzed after treatment with 0.5 mM H2O2 for 15 min. This concentration of H2O2 was chosen because it causes maximal phosphorylation of eIF2 after a 15-min treatment (Shenton et al., 2006). Microarray analysis revealed that this H2O2 treatment caused a significant effect on the transcriptome, with 476 genes up-regulated greater than twofold and 161 genes down-regulated greater than twofold in the wild-type strain. We identified which genes show gcn2- or gcn4-dependent H2O2 expression by measuring the interaction term in an analysis of variance model. This was done in two separate tests for wild-type versus gcn2 mutant and for wild-type versus gcn4 mutant. Differentially expressed genes picked for subsequent analysis met the criterion of q value (a fdr-corrected p value) <0.05 for either the interaction of wild-type versus gcn2 or wild-type versus gcn4 (see Materials and Methods). This produced a data set of 376 genes whose expression was significantly different in the gcn4 mutant, and 357 genes whose expression was significantly different in the gcn2 mutant, compared with the wild-type, respectively. This enriched data set was segregated into eight clusters based on similarity of expression profile across the data set using a k-means clustering algorithm (Figure 8). The microarray data were validated for a range of genes by using real-time RT-PCR analysis (Figure 9). Representative genes are shown from cluster B (YOR1), cluster D (GCY1), cluster E (ATP3, SDH1, COR1), and cluster G (FRE1). The RT-PCR analysis confirmed the general trends that were detected from the microarray analysis. For each of the clusters shown in Figure 8, significant overrepresentation of genes belonging to a particular classification was determined using FunSpec (Table 2).
|
|
|
). Several genes encoding stress defenses or antioxidants were identified in this cluster. Most prominently, a catalase gene (CTT1) was up-regulated that functions to detoxify hydrogen peroxide (Izawa et al., 1996). Several osmotic stress (RVS161, GPD1, HOR2, CAP1, and GCY1), heat stress (SSE2 and HSP30), and stationary phase (YGP1)-related genes were identified, emphasizing the physiological link between diverse environmental stress conditions (Gasch et al., 2000). Cluster E included genes in which expression was constitutively elevated in gcn2 and gcn4 mutants, indicating that Gcn4 normally acts to repress these genes. This cluster encompassed a significant number of mitochondrial genes affecting energy generation and respiration in agreement with previous reports suggesting that mitochondrial function is particularly important for resistance to H2O2 (Grant et al., 1996; Thorpe et al., 2004). Cluster F-included genes where expression was moderately down-regulated in response to peroxide in the gcn4 mutant and moderately up-regulated in the gcn2 mutant, but no significantly overrepresented functional classes were identified in this cluster (Table 2).
Genes Down-Regulated by Hydrogen Peroxide in the Wild Type
Clusters A and C included genes that are down-regulated in response to H2O2 exposure. Cluster C was the largest cluster identified, and it included genes in which down-regulation was dampened in both the gcn2 and gcn4 mutants consistent with regulation via translational control of GCN4 expression. This cluster included a large number of gene involved in transcription and RNA metabolism. This presumably indicates a requirement to down-regulate energy-consuming processes during stress conditions as noted previously during the environmental stress response (Gasch et al., 2000). Similarly, several genes affecting the cell cycle were down-regulated, consistent with previous observations indicating that the cell cycle is arrested as part of the defense response against H2O2 stress (Flattery-O'Brien and Dawes, 1998). Surprisingly, the SKN7 transcription factor was identified as a component of cluster C. Skn7 is a two-component system response regulator that cooperates with the Yap1 transcription factor for the induction of many oxidative stress genes (Ikner and Shiozaki, 2005). Regulation of Yap1 activity via down-regulation of Skn7 may therefore represent a previously unrecognized response to hydrogen peroxide. Cluster A included genes that were down-regulated in both the wild-type and gcn2 mutants, but they were moderately up-regulated in the gcn4 mutant. The main functional classes identified in this cluster predominantly affected lipid metabolism.
Genes That Are Unaffected in the Wild Type but Show Altered Expression in Response to Hydrogen Peroxide in gcn Mutants
Many genes were identified that showed little change in gene expression in the wild-type strain, but they were altered in response to hydrogen peroxide in gcn mutants. Gene expression was strongly up-regulated in the gcn4 mutant in both clusters B and G. The gcn2 mutant was similar to the wild-type strain for cluster G, whereas expression was constitutively elevated in the gcn2 mutant for cluster B. Significantly, genes affecting heavy metal uptake and use were overrepresented in both clusters. This included several genes involved in iron (FRE1, FRE2, FRE3, FRE4, FRE5 SIT1, ARN1, ARN2, SMF3), copper (CCC2), and zinc (COT1) transport. Increased iron uptake during oxidative stress is somewhat surprising given that it can potentially lead to the generation of the hydroxyl radical (·OH) via the Fenton reaction. However, we have previously noted that a number of iron transport genes are translationally up-regulated in response to oxidative stress and suggested that this may indicate a requirement to restore iron homeostasis after ROS exposure (Shenton et al., 2006). Cellular iron is found largely complexed in cells, for example in iron–sulfur (Fe/S) clusters. Oxidation of these clusters causes release of the iron, resulting in enzyme inactivation; hence, there may be a requirement to replace lost iron. Cluster H included genes that were unaffected in the wild type, but they were down-regulated in response to H2O2 stress in both the gcn2 and gcn4 mutants. This cluster included several metabolic genes, including those affecting amino acid (lysine, arginine, isoleucine, and tryptophan) and vitamin (BIO2) biosynthesis. These genes have been identified previously as genes that are regulated by Gcn4 in response to amino acid starvation (Natarajan et al., 2001). It is unclear why these genes are down-regulated in gcn mutants, but it may indicate that Gcn4regulated genes may be disrupted in response to oxidative stress in strains lacking general control.
Comparison with the Gcn4-dependent Response to Amino Acid Starvation
Given that Gcn4 has best been characterized as a transcriptional regulator that responds to amino acid starvation conditions, we compared our expression data with previous amino acid starvation expression data. Many genes have been identified that are induced or repressed in response to starvation for histidine by treatment with 3-aminotriazole (3-AT) (Natarajan et al., 2001). The scatter plots in Figure 10 show a comparison of changes in gene expression due to H2O2 exposure and 3-AT treatment in wild-type and gcn4 mutant strains. Although there is not a strong correlation between peroxide and amino acid starvation-regulated genes on a genome-wide scale, many genes were similarly induced or repressed in response to both treatments. Furthermore, loss of GCN4 dampened this effect, and we were able to identify 64 genes that show Gcn4-dependent expression in both data sets (Figure 10, genes marked with a square symbol; Supplemental Table 2). The 44 up-regulated genes included significant overrepresentation of genes affecting metabolism (carbon, amino acid, and energy) and the stress response. The 22 down-regulated genes included significant overrepresentation of genes affecting transcription.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). This seems to be an evolutionarily conserved response that inhibits global translation while promoting continued translation of defense genes that protect against and/or repair the resulting oxidant damage (Harding et al., 2003; Shenton et al., 2006). H2O2 inhibits translation initiation via a mechanism that depends on the Gcn2 protein kinase, which phosphorylates the subunit of eIF2 (Shenton et al., 2006). In this current study, we have shown that eIF2 is also phosphorylated in response to diverse oxidants. This is somewhat surprising given the differing mechanisms of toxicity mediated by these oxidants, and it suggests that translation inhibition mediated via phosphorylation of eIF2 is a general response to oxidative stress. However, diamide and cadmium also inhibit translation initiation via a Gcn2-independent mechanism, which may provide further oxidant-specific regulation of gene expression.
Hydrogen peroxide is a ubiquitous molecule formed as a by-product of aerobic respiration and after exposure to diverse biological and environmental factors. It can damage cells by promoting oxidative stress, but it also plays important roles as a signaling molecule in the regulation of many biological processes (reviewed in Veal et al., 2007). As well as being both freely diffusible and reactive, H2O2 also must also be removed from cells to avoid Fenton and Haber–Weiss reactions leading to the formation of highly reactive hydroxyl radicals (reviewed in Temple et al., 2005). Cumene hydroperoxide is a lipid-soluble hydroperoxide that is widely used as an intracellular source of ROS (Thorpe et al., 2004). It can generate highly reactive free radicals such as the alkoxy radical, resulting in high mutagenicity and toxicity (Simic et al., 1989). Cadmium is a highly toxic metal and a well-established human carcinogen. It is capable of entering cells via the same transport systems used by the essential heavy metals. Once inside the cell, the main mechanism for toxicity is through the depletion of glutathione and binding to sulfydryl groups (Valko et al., 2005). It can also displace iron and copper from various cytoplasmic and membrane proteins, increasing the levels of unbound free or chelated copper and iron ions contributing to oxidative stress via Fenton reactions (Valko et al., 2006). Diamide is a membrane-permeable, thiol-specific oxidant, which promotes the formation of disulfides (Kosower and Kosower, 1995). It reacts rapidly and specifically with glutathione; thus, it causes oxidative stress by inhibiting antioxidant defenses. Hence, these diverse oxidants can generate ROS through both direct and indirect mechanisms.
Hydroperoxides, diamide, and cadmium all seem to activate Gcn2 via a similar mechanism. Mutations in the m2 motif of the HisRS-like domain (Y1119L and R1120L) that abolish tRNA binding by Gcn2 (Cherkasova and Hinnebusch, 2003) abrogate (hydroperoxides) or reduce (cadmium and diamide) inhibition. These results suggest that ROS activate Gcn2 via a mechanism that elevates uncharged tRNA levels. However, it should be noted that we cannot rule out that some as yet undefined signal requires this same domain for Gcn2 activation. Gcn2 is known to be activated in response to a variety of conditions, including nutrient starvation (amino acids, purines, and glucose) and exposure to sodium chloride, rapamycin, ethanol and volatile anesthetics (reviewed in Hinnebusch, 2005; Palmer et al., 2005). Depletion of amino acids leads to an accumulation of uncharged tRNA, which activates the Gcn2 protein kinase via its HisRS-related domain. It is likely that other stress conditions ultimately affect the levels of uncharged tRNA in the cell. For example, volatile anesthetics inhibit amino acid uptake (Palmer et al., 2005), and Gcn2 is activated by glucose starvation partly through an effect on vacuolar amino acid pools (Yang et al., 2000). Our data indicate that ROS activate Gcn2 in a prototrophic strain (a strain that does not require exogenous amino acids from the growth media), ruling out any effects on amino acid uptake. This raises the question as to how ROS cause an amino acid starvation, effect cellular uncharged tRNA levels, or both. Oxidative stress may affect the transport and storage of amino acids within cells, similar to the effect of glucose starvation. Additionally, oxidative stress may conceivably cause an accumulation of uncharged tRNA through a variety of mechanisms. Free amino acids and amino acids in proteins are highly susceptible to oxidation by ROS, which may cause an imbalance in amino acid pool sizes (Stadtman and Levine, 2003). Alternatively, the proteins and nucleic acids which are required for tRNA-aminoacylation may be susceptible to oxidation, resulting in an accumulation of uncharged tRNA and activation of Gcn2.
The signals activating Gcn2 in response to rapamycin and NaCl are not well understood. Rapamycin seems to work by blocking Tor-mediated phosphorylation of Gcn2 at Ser577 (Cherkasova and Hinnebusch, 2003). However, activation of Gcn2 by rapamycin and NaCl still requires the HisRS-related domain of Gcn2, as well as Gcn1 and Gcn20, which are thought to mediate the activation of Gcn2 by uncharged tRNA (Narasimhan et al., 2004). ROS act differently to rapamycin or NaCl because a Gcn2-S577A mutant can still be activated to phosphorylate eIF2. Heterologous protein production also activates Gcn2 through an effect on uncharged tRNA levels, although it is unclear how heterologous expression affects amino acid levels (Steffensen and Pedersen, 2006). This is particularly interesting because heterologous protein production has previously been shown to induce an oxidative stress (Bannister and Wittrup, 2000); hence, Gcn2 may be activated via a ROS-dependent mechanism during heterologous expression. It is tempting to speculate that Gcn2 activation is a defense mechanism that is invoked in response to any condition that elevates the cellular concentrations of ROS.
A key control point for regulating eukaryotic translation initiation is via binding of initiation factors to the mRNA cap (Richter and Sonenberg, 2005). The initiation factor eIF4E recognizes the mRNA cap structure and also interacts with the "multi-adaptor" protein eIF4G. This process is critical for the recruitment of ribosomes to the 5' end of mRNAs and translation initiation can be regulated by the competitive binding of eIF4E-binding proteins (4E-BPs) to eIF4E. Two such proteins have been identified in yeast, Caf20 and Eap1 (Altmann et al., 1997; Cosentino et al., 2000). These 4E-BPs have broad functions in cell growth, proliferation, and development (Ibrahimo et al., 2006; Park et al., 2006). Disruption of EAP1 (but not CAF20) was found to impair cadmium- and diamide-induced regulation of translation initiation. In contrast, peroxide-induced inhibition was unaffected in the eap1 mutant, highlighting the different translational control mechanisms invoked by these different oxidants. Simultaneous loss of EAP1 and GCN2 abrogated the translation initiation inhibition mediated by cadmium and diamide. Similarly, yeast cells respond to a membrane stress by attenuating translation initiation via a mechanism that is mediated by Gcn2 and Eap1 (Deloche et al., 2004). This response may serve to prevent the mislocalization of proteins after disruption of vesicular transport pathways. Interestingly, diamide readily oxidizes the ER and mutants defective in vacuolar protein-sorting functions are particularly sensitive to diamide (Cuozzo and Kaiser, 1999; Thorpe et al., 2004). Furthermore, diamide-stress induces the expression of many cell wall biosynthesis genes and genes involved in protein secretion and processing in the ER, suggesting that diamide causes defects in secretion (Gasch et al., 2000). Diamide may therefore, invoke a membrane transport defect that promotes Gcn2- and Eap1-mediated attenuation of translation initiation. Cadmium toxicity is less well understood, and it is unclear at present whether cadmium particularly affects protein secretion or trafficking. Unlike diamide, cadmium does not directly oxidize glutathione, but it may affect the redox state of secretory processes via an indirect effect on protein and low- molecular-weight thiols.
Gcn4 is a transcriptional activator of amino acid biosynthetic genes. It is translationally up-regulated in response to phosphorylation of eIF2 (Hinnebusch, 2005). However, despite the finding that phosphorylation of eIF2 is a general response to oxidative stress, not all oxidants induce translational expression of GCN4. Gcn4 is up-regulated in response to H2O2, and it is specifically required for hydroperoxide resistance. We have proposed that translation of the GCN4 mRNA is specifically able to continue during peroxide stress, a condition where global translation is inhibited at both the initiation and postinitiation phases (Shenton et al., 2006). GCN4 is induced by the well-known Gcn2-dependent, eIF2 phosphorylation mechanism, but its translation must also be resistant to the postinitiation block in protein synthesis. In contrast, Gcn4 protein levels are somewhat reduced after diamide or cadmium stress. One possibility to explain these findings is that GCN4 expression is inhibited in response to cadmium and diamide by Eap1-mediated inhibition of translation initiation. However, additional translational regulatory mechanisms must exist because loss of both GCN2 and EAP1 does not restore the rate of protein synthesis to uninhibited levels after diamide or cadmium stress. Presumably, GCN4 expression is inhibited by other translational control mechanisms which act at the postinitiation phase in response to these oxidants. Loss of both GCN2 and EAP1 abrogates the inhibition of translation initiation in response to cadmium and diamide, but it does not affect this postinitiation block. Gcn2 and Eap1 do not affect the postinitiation block; hence, any inhibition of ribosomal transit induced by diamide or cadmium should be comparable in wild-type and gcn2 eap1 mutant cells inhibiting protein synthesis. Induction of GCN4 expression correlates with the finding that Gcn4 is required for resistance to peroxides but not to diamide and cadmium. The sensitivity of the gcn4 mutant was determined using spots tests, which measure the tolerance of strains to a lethal dose of oxidant. This sensitivity therefore suggests that basal levels of expression of Gcn4-traget genes are required for peroxide-tolerance. In contrast, the gcn2 mutant was unaffected in sensitivity to peroxides in this assay. The requirement for Gcn2 is apparent during adaptive conditions when it is required to inhibit translation initiation and to induce the expression of GCN4 in response to low concentrations of hydrogen peroxide stress in actively growing cells.
Previous transcriptional profiling studies have shown that 10% of the yeast genome is regulated by Gcn4 in response to amino acid starvation (Natarajan et al., 2001). The broad transcriptional response controlled by Gcn4 suggests that Gcn4 acts as a master regulator of gene expression. Our analysis of a gcn4 mutant indicates that Gcn4 is required for a more limited set of genes in response to H2O2-stress. Comparison of peroxide and amino acid starvation-regulated genes revealed that there is not a strong correlation at the genome-wide level. Nevertheless, several genes are similarly regulated by both stress conditions, including 64 genes (altered by greater than twofold), which require Gcn4 for regulation in response to both stress conditions. Hydrogen peroxide stress damages many intracellular targets and affects diverse cellular processes. Our data indicate that part of the response to peroxide is mediated through an amino acid starvation. However, additional regulatory inputs must control Gcn4-target gene expression in response to ROS, as seen previously for glucose and heterologous protein production (Yang et al., 2000; Steffensen and Pedersen, 2006). Similarly other transcription factors including the yeast coactivator protein Mbf1, and Hac1, which regulates the unfolded protein response, are known to mediate Gcn4 transcriptional activity (Takemaru et al., 1998; Patil et al., 2004).
Cluster analysis revealed a complex pattern of transcriptional regulation in gcn mutants. The requirement for translational regulation of GCN4 expression in response to hydrogen peroxide stress was most clearly shown by the identification of genes that are up-regulated (clusters D–F) or down-regulated (cluster C) dependent on the presence of both GCN2 and GCN4. It is also apparent from our data that loss of GCN4 alters the peroxide-regulated expression of many genes independent of Gcn2. These include genes that are similarly regulated in the wild-type and gcn2 mutant, but that are up-regulated (clusters A, B, and G) or down-regulated (cluster H) in the gcn4 mutant. The sensitivity of the gcn4 mutant to hydrogen peroxide may mean that regulation of these genes is somehow preadapted such that they strongly respond to stress conditions in the absence of normal Gcn4-mediated control. It is now well established that the response to oxidative stress requires extensive reprogramming of transcription and translation. Our data indicate that translational control of GCN4 expression and transcriptional control of Gcn4 target genes are key components of this adaptive response.
|
ACKNOWLEDGMENTS |
|---|
|
Footnotes |
|---|
Address correspondence to: Chris M. Grant (chris.grant{at}manchester.ac.uk)
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Ashe, M. P., De Long, S. K., and Sachs, A. B. (2000). Glucose depletion rapidly inhibits translation initiation in yeast. Mol. Biol. Cell 11, 833–848.
Bannister, S. J., and Wittrup, K. D. (2000). Glutathione excretion in response to heterologous protein secretion in Saccharomyces cerevisiae. Biotechnol. Bioeng 68, 389–395.[CrossRef][Medline]
Bertolotti, A., Zhang, Y., Hendershot, L. M., Harding, H. P., and Ron, D. (2000). Dynamic interaction of BiP and ER stress transducers in the unfolded-protein response. Nat. Cell Biol 2, 326–332.[CrossRef][Medline]
Bolstad, B. M., Irizarry, R. A., Astrand, M., and Speed, T. P. (2003). A comparison of normalization methods for high density oligonucleotide array data based on variance and bias. Bioinformatics 19, 185–193.
Causton, H. C., Ren, B., Koh, S. S., Harbison, C. T., Kanin, E., Jennings, E. G., Lee, T. I., True, H. L., Lander, E. S., and Young, R. A. (2001). Remodeling of yeast genome expression in response to environmental changes. Mol. Biol. Cell 12, 323–337.
Cherkasova, V. A., and Hinnebusch, A. G. (2003). Translational control by TOR and TAP42 through dephosphorylation of eIF2alpha kinase GCN2. Genes Dev 17, 859–872.
Cosentino, G. P., Schmelzle, T., Haghighat, A., Helliwell, S. B., Hall, M. N., and Sonenberg, N. (2000). Eap1p, a novel eukaryotic translation initiation factor 4E-associated protein in Saccharomyces cerevisiae. Mol. Cell. Biol 20, 4604–4613.
Cuozzo, J. W., and Kaiser, C. A. (1999). Competition between glutathione and protein thiols for disulphide-bond formation. Nat. Cell Biol 1, 130–135.[CrossRef][Medline]
Deloche, O., de la Cruz, J., Kressler, D., Doere, M., and Linder, P. (2004). A membrane transport defect leads to a rapid attenuation of translation initiation in Saccharomyces cerevisiae. Mol. Cell 13, 357–366.[CrossRef][Medline]
Dever, T. E. (2002). Gene-specific regulation by general translation factors. Cell 108, 545–556.[CrossRef][Medline]
Flattery-O'Brien, J. A., and Dawes, I. W. (1998). Hydrogen peroxide causes RAD9-dependent cell cycle arrest in G2 in Saccharomyces cerevisiae whereas menadione causes G1 arrest independent of RAD9 function. J. Biol. Chem 273, 8564–8571.
Garcia-Barrio, M., Dong, J., Cherkasova, V. A., Zhang, X., Zhang, F., Ufano, S., Lai, R., Qin, J., and Hinnebusch, A. G. (2002). Serine 577 is phosphorylated and negatively affects the tRNA binding and eIF2alpha kinase activities of GCN2. J. Biol. Chem 277, 30675–30683.
Gasch, A. P., Spellman, P. T., Kao, C. M., Carmel-Harel, O., Eisen, M. B., Storz, G., Botstein, D., and Brown, P. O. (2000). Genomic expression programs in the response of yeast cells to environmental changes. Mol. Biol. Cell 11, 4241–4257.
Gentleman, R. C. et al. (2004). Bioconductor: open software development for computational biology and bioinformatics. Genome Biol 5, R80.[CrossRef][Medline]
Godon, C., Lagniel, G., Lee, J., Buhler, J. M., Kieffer, S., Perrot, M., Boucherie, H., Toledano, M. B., and Labarre, J. (1998). The H2O2 stimulon in Saccharomyces cerevisiae. J. Biol. Chem 273, 22480–22489.
Grant, C. M., MacIver, F. H., and Dawes, I. W. (1996). Glutathione is an essential metabolite required for resistance to oxidative stress in the yeast Saccharomyces cerevisiae. Curr. Genet 29, 511–515.[Medline]
Gutteridge, J. M. (1994). Biological origin of free radicals, and mechanisms of antioxidant protection. Chem. Biol. Interact 91, 133–140.[CrossRef][Medline]
Halliwell, B., and Gutteridge, J.M.C. (1989). Free Radicals in Biology and Medicine, Oxford, New York: Oxford University Press.
Harding, H. P., Novoa, I., Zhang, Y., Zeng, H., Wek, R., Schapira, M., and Ron, D. (2000). Regulated translation initiation controls stress-induced gene expression in mammalian cells. Mol. Cell 6, 1099–1108.[CrossRef][Medline]
Harding, H. P. et al. (2003). An integrated stress response regulates amino acid metabolism and resistance to oxidative stress. Mol. Cell 11, 619–633.[CrossRef][Medline]
Hinnebusch, A. G. (2000). Mechanism and Regulation of Initiator Methionyl-tRNA Binding to Ribosomes. In: Translational Control of Gene Expression, N. Sonenberg, J.W.B. Hershey, and M. B. Mathews, Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press.
Hinnebusch, A. G. (2005). Translational regulation of GCN4 and the general amino acid control of yeast. Annu. Rev. Microbiol 59, 407–450.[CrossRef][Medline]
Holmes, L. E., Campbell, S. G., De Long, S. K., Sachs, A. B., and Ashe, M. P. (2004). Loss of translational control in yeast compromised for the major mRNA decay pathway. Mol. Cell. Biol 24, 2998–3010.
Ibrahimo, S., Holmes, L. E., and Ashe, M. P. (2006). Regulation of translation initiation by the yeast eIF4E binding proteins is required for the pseudohyphal response. Yeast 23, 1075–1088.[CrossRef][Medline]
Ikner, A., and Shiozaki, K. (2005). Yeast signaling pathways in the oxidative stress response. Mutat. Res 569, 13–27.
