Current Biology, Vol. 12, 1279–1286, August 6, 2002, 2002 Elsevier Science Ltd. All rights reserved. PII S0960-9822(02)01037-0

Activation of GCN2 in UV-Irradiated CellsInhibits Translation

vated protein (MAP) kinase signaling modules to elicitadditional transcriptional responses to UV irradiation[6–8].

Jing Deng,1 Heather P. Harding,2 Brian Raught,1,5

Anne-Claude Gingras,1,5 Juan Jose Berlanga,1

Donalyn Scheuner,3 Randal J. Kaufman,3

David Ron,2 and Nahum Sonenberg1,4 Less well understood are the mechanisms coupling1Department of Biochemistry UV irradiation to translational repression [2, 9, 10]. The

and McGill Cancer Centre phosphorylation of the � subunit of translation initiationMcGill University factor 2 (eIF2�) is a common response of eukaryotic3655 Promenade Sir William Osler cells to different types of stress [11, 12]. eIF2 forms aMontreal, Quebec H3G 1Y6 ternary complex with Met-tRNAi

Met and GTP. An earlyCanada step in translation initiation involves the binding of the2 Skirball Institute ternary complex to the 40S ribosomal subunit. The 40SDepartments of Medicine and Cell Biology ribosomal subunit, in conjunction with the eIF2 ternary

and the Kaplan Cancer Center complex and other initiation factors, subsequentlyNew York University School of Medicine scans the mRNA until an initiation codon in the properNew York, New York 10016 sequence context is detected. Following pairing be-3 Howard Hughes Medical Institute tween the Met-tRNAi

Met anticodon and the AUG startUniversity of Michigan Medical Center codon, the eIF2-associated GTP is hydrolyzed to GDP.Ann Arbor, Michigan 48109 eIF2-GDP is then released, and translation proceeds.

eIF2-GDP must be converted to eIF2-GTP to catalyzesubsequent rounds of translation initiation.

Summary The eIF2 GDP-GTP exchange reaction is stimulatedby the guanine nucleotide exchange factor eIF2B [11].

Background: Mammalian cells subjected to ultraviolet Phosphorylation of eIF2� on Ser 51 inhibits the GDP-(UV) irradiation actively repress DNA replication, tran- GTP exchange reaction by stabilizing the eIF2�-GDP-scription, and mRNA translation. While the effects of UV eIF2B interaction. Since eIF2 is present at higher levelsirradiation on DNA replication and transcription have than eIF2B, phosphorylation of only a fraction of eIF2been extensively studied, the mechanism(s) responsible can lead to significant eIF2B sequestration and a resul-for translational repression are poorly understood. tant dramatic translation inhibition. eIF2� is phosphory-Results: Here, we demonstrate that UV irradiation elic- lated on Ser 51 in response to a variety of stress condi-its phosphorylation of the � subunit of eukaryotic trans- tions, including suboptimal levels of amino acids, serum,lation initiation factor 2 (eIF2�) by activating the kinase or glucose; exposure to heat, heavy metals, or arsenite;GCN2 in a manner that does not require SAPK/JNK or or hypoxic or hyperosmotic conditions [11, 13–15]. Thisp38 MAP kinase. GCN2�/� cells, and cells expressing mode of translational control is evolutionarily conservednonphosphorylatable eIF2� as their only source of eIF2� and is effected by at least four different eIF2� kinases.protein, fail to repress translation in response to UV GCN2 (general control non-derepressible-2), an eIF2�irradiation. kinase conserved from yeast to mammals, is activatedConclusions: These results provide a mechanism for in response to amino acid starvation, purine limitation,translation inhibition by UV irradiation and identify a

or methyl methane sulfonate, a DNA-damaging agenthitherto unrecognized role for mammalian GCN2 as a

[11, 16–19], whereas PERK (also known as PEK) is anmediator of the cellular response to UV stress.

eIF2� kinase found in metazoans that couples proteinfolding in the endoplasmic reticulum to translation initia-Introductiontion rates in the cytoplasm [20, 21]. The vertebrate ki-nases PKR (double-stranded RNA activated protein ki-Ultraviolet (UV) irradiation can severely damage nucleicnase) and HRI (heme-regulated inhibitor) are activatedacids [1, 2]. Damage to DNA, mRNA, rRNA, or tRNA canby double-stranded RNA during viral infection and he-dramatically affect cellular physiology. Eukaryotic cellsmin deprivation, respectively [22, 23]. While GCN2,have thus evolved an array of protective mechanismsPERK, and PKR share a broad tissue distribution, HRIthat are activated in response to this and other typesis found mainly in erythrocytes.of environmental stresses. The best-characterized UV-

Here, we report that eIF2� phosphorylation in mam-induced protective mechanism is the DNA damage-cellmalian cells is significantly increased in response tocycle checkpoint pathway, activation of which resultsUVC (ultraviolet-C; 200–280 nm) irradiation, and that thisin cell cycle arrest and the induction of DNA repair genesconserved mechanism is required for translation repres-[3–5]. A second, less-characterized pathway is cyto-sion in response to UV stress. Using mouse embryonicplasmic and is dependent on Ras, which activates thefibroblasts (MEFs) harboring targeted deletions of thec-Jun N-terminal kinase (JNK) and p38 mitogen-acti-eIF2� kinases PKR, PERK, or GCN2, we demonstratethat GCN2 is responsible for eIF2� phosphorylation and4 Correspondence: nahum.sonenberg@mcgill.cathe inhibition of translation, in response to UVC irradi-5 Present address: Institute for Systems Biology, 1441 North 34th

Street, Seattle, Washington 98103. ation.

Current Biology1280

Results

eIF2� Phosphorylation Is Responsible for Inhibitionof Translation by UV IrradiationUV irradiation inhibits translation in eukaryotic cells. Acommon response to various types of environmentalstresses is the phosphorylation of eIF2�. To characterizethe effects of UV irradiation on eIF2� phosphorylationin mammalian cells, dose response and time courseexperiments were conducted. A significant increase inthe phosphorylation of eIF2� was detected when cellswere irradiated at a UV dose of 60 J/m2 (Figure 1A). Thisdose was found not to cause significant apoptosis infibroblasts [24] (and our data not shown) and is the samedose at which the phosphorylation of the JNK and p38MAP kinases is detected [25, 26] (Figure 1A). eIF2� isphosphorylated rapidly (within 1 min) following UV irradi-ation (80 J/m2 ) and peaks at 20 min post-UV exposure(Figure 1B). Temporally, peak eIF2� phosphorylation co-incides with JNK and p38 MAP kinase phosphorylation,which are early events in the cellular response to UVirradiation (Figure 1B).

To determine if the observed phosphorylation of eIF2�on Ser51 plays an important role in translational repressionin UV-irradiated cells, we compared the rates of proteinsynthesis in UV-irradiated wild-type (wt) MEFs with thoseof MEFs derived from homozygous eIF2�Ser51Ala/Ser51Ala

“knockin” mouse embryos [27]. eIF2� with the Ser51Alasubstitution can no longer function as a substrate forkinases that control translation initiation rates but other-wise provides normal ternary complex function. In wtcells, UV strongly repressed translation initiation,whereas in the irradiated eIF2�Ser51Ala/Ser51Ala cells, transla-tion proceeded unabated (Figure 1C), which correlateswith the lack of accumulation of eIF2� phosphorylationin those knockin cells (Figure 1D). To exclude the possi-bility that the reduction of 35S-methionine incorporationinto newly synthesized proteins is due to an effect ofirradiation on the uptake of 35S-methionine, we deter-mined the total amount of radioactivity taken up by thecells. At a dose (e.g., 80 J/m2 ) that causes strong eIF2�phosphorylation and inhibition of translation in wt cells,there was no change in the total amount of radioactivityin eIF2�Ser51Ala/Ser51Ala cells (data not shown).

PKR and PERK Are Not Required for UV-InducedeIF2� PhosphorylationHaving determined that eIF2� phosphorylation plays anessential role in translational repression in UV-irradiated

(B) The time course of phosphorylation. 3T3 cells were harvestedat the indicated times following UV irradiation (80 J/m2 ). Cell lysateswere analyzed for phosphorylated and total protein as in (A).(C) Protein synthesis rates were measured in MEFs expressing wild-

Figure 1. Dose Response and Time Course Analyses of the Phos- type or the eIF2�Ser51Ala/ Ser51Ala knockin mutant. Cells were irradiatedphorylation of eIF2�, JNK, and p38 MAP Kinase Following UV Irradia- with UV light or were treated with thapsigargin (Tg, 500 nM, 1 hr),tion; Expression of the eIF2� Ser51Ala Mutant Prevents the Attenua- then subjected to a 30-min pulse-labeling in the presence of 35S-tion of Protein Synthesis by UV Irradiation. methionine. Cell lysates were precipitated by trichloroacetic acid(A) UV dose response of eIF2�, JNK, and p38 MAP kinase phosphor- (TCA) and were measured by scintillation counting. Percentage val-ylation. 3T3 cells were irradiated with increasing doses of UV light ues were obtained by normalizing the counts of treated samplesand were allowed to recover in complete medium for 30 min before against the controls (the results from three independent experimentsharvest. Cell lysates were analyzed for the phosphorylation of eIF2�, are shown).JNK, and p38 MAP kinase with phospho-specific antibodies as indi- (D) Western blot analysis of eIF2� phosphorylation in the wt andcated. eIF2�Ser51Ala/ Ser51Ala cells shown in (C). C: control.

GCN2 Represses Translation in UV Stress1281

Figure 2. PKR and PERK Are Not Requiredfor UV-Induced eIF2� Phosphorylation

MEFs were irradiated with an increasing doseof UV light, and cell lysates were analyzed byWestern blotting.(A) Phosphorylated and total eIF2�, and theexpression of PKR protein, in wild-type (�/�)and PKR knockout (�/�) cells.(B) Phosphorylated and total eIF2� in �/�and PERK�/� cells. MEFs were also treatedwith thapsigargin (Tg, 500 nM, 1 hr). C:control.

cells, we next wished to determine if any of the known GCN2 in eIF2� phosphorylation in UV-irradiated cells.To determine if GCN2 is activated by UV irradiation, weeIF2� kinases might be implicated in this process. To

this end, we utilized MEFs, or established MEF cell lines, examined its phosphorylation state. GCN2 activationcorrelates with Thr898 phosphorylation in its activationderived from mice in which the genes encoding the

eIF2� kinases PKR or PERK were deleted by gene tar- loop in response to leucine depletion [19]. UV irradiationmarkedly induced GCN2 phosphorylation, as deter-geting. eIF2� phosphorylation in response to UV irradia-

tion was not abolished in PKR�/� or PERK�/� cells mined by Western blotting with an antiserum reactiveto phospho-Thr898 (Figure 3C). Levels of GCN2 phos-(Figures 2A and 2B). This suggests that neither kinase

plays an essential role in coupling UV irradiation to eIF2� phorylation in UV-irradiated cells were comparable tothose found in cells deprived of leucine, indicating thatphosphorylation. As expected, the PKR�/� cells lacked

PKR protein (Figure 2A), and the PERK�/� cells were UV irradiation is a strong activator of GCN2.severely impaired in their ability to elicit eIF2� phosphor-ylation in response to thapsigargin, an agent that causes Deletion of the GCN2 Gene Prevents Inhibition

of Translation by UV Irradiationendoplasmic reticulum stress (Figure 2B). These dataconfirmed the presence of the respective deletions. To further examine the role of GCN2 in translational

inhibition by UV irradiation, we compared protein syn-thesis rates in irradiated wt and GCN2�/� cells. Re-GCN2 Is Required for UV-Induced eIF2�

Phosphorylation duced incorporation of 35S-methionine (in a 30 min pulse-labeling) into newly synthesized proteins was apparentThe third eIF2� kinase tested was GCN2, which is acti-

vated in response to amino acid starvation in mamma- in wt cells at a UV dose of 80 J/m2 (20% decrease,Figures 4A and 4B). Further increasing the UV doselian cells [19]. Leucine deprivation resulted in a strong

induction of eIF2� phosphorylation in isogenic wt MEFs resulted in a dramatic translation inhibition in wt cells.However, the UV-induced inhibition of translation was(Figure 3A). However, leucine deprivation failed to in-

duce eIF2� phosphorylation in MEFs derived from significantly diminished in GCN2�/� cells (e.g., com-pare wt and GCN2�/� cells at 120 J/m2; Figures 4A andGCN2�/� mice [19]. This validates the GCN2-negative

status of these cells. Strikingly, while the wt MEFs were 4B). Even at a dose of 160 J/m2, at which translationwas inhibited in wt MEFs by almost 70%, GCN2�/�responsive to UV stress in that eIF2� became phosphor-

ylated, UV irradiation failed to induce eIF2� phosphory- cell translation rates remained at �80% of unirradiatedvalues (the inhibition observed in the absence of GCN2lation in GCN2�/� MEFs (Figures 3A and 3B). Phos-

phorylation of the JNK protein occurred normally in at higher UV doses may be caused by damage to ribo-somal RNA [2]). Thapsigargin treatment, which activatesGCN2�/� MEFs (Figure 3B). This indicates that these

cells are not defective in all aspects of the UV stress PERK, elicited a similar degree of translation inhibitionin wt and GCN2�/� cells (Figure 4C). The GCN2�/�response. These results establish an essential role for

Current Biology1282

Figure 3. Targeted Deletion of the GCN2Gene Abolishes UV-Induced eIF2� Phos-phorylation; GCN2 Is Activated by UV Irradi-ation

(A and B) UV-induced GCN2-dependentphosphorylation of eIF2�. Wild-type andGCN2 knockout MEFs were irradiated withan increasing dose of UV light, (A) from 80 to200 J/m2 or (B) from 10 to 80 J/m2. MEFs,deprived of leucine, were treated in parallelas a control for GCN2-specific eIF2� phos-phorylation. Cell lysates were analyzed forphosphorylation of eIF2� and JNK as in Fig-ure 1A. The phosphorylated and total proteinsare shown as indicated.(C) The phosphorylation state of GCN2 fol-lowing UV irradiation. Cell lysates preparedfrom UV-irradiated wild-type GCN2 MEFswere immunoprecipitated with anti-GCN2 an-tibody and were blotted with a phospho-Thr898 GCN2 antiserum. Cells deprived ofleucine were analyzed as a control for GCN2activation. Total GCN2 levels were subse-quently blotted with anti-GCN2 antibody. C:control; Leu: leucine.

cells, therefore, remained competent for eIF2� phos- rather increased to some extent in p38 MAP kinase �/�cells (Figure 5B). These results clearly demonstrate thatphorylation in response to other types of stress.p38 MAP kinase is not required for the phosphorylationof eIF2� by GCN2. To demonstrate that JNK is also not

UV-Induced eIF2� Phosphorylation by GCN2 involved in UV-induced eIF2� phosphorylation, 3T3 cellsIs Independent of JNK and p38 MAP Kinase were treated with a specific JNK inhibitor, SP600125UV irradiation activates two important stress signaling [29]. While the phosphorylation of c-Jun was abolishedpathways, the JNK and p38 MAP kinase modules. Acti- by SP600125, eIF2� phosphorylation was not reducedvation of JNK and p38 MAP kinase by phosphorylation by SP600125 following UV irradiation (Figure 5C). Again,leads to transcriptional activation of genes functioning as seen with the p38 MAP kinase �/� cells, SP600125in the cellular stress response [28]. It was therefore im- treatment resulted in a slight increase in eIF2� phos-portant to determine whether GCN2, JNK, and p38 MAP phorylation. Therefore, the JNK and p38 MAP kinasekinase impact on each other following UV irradiation. In signaling modules do not appear to modulate GCN2response to increasing doses of UV irradiation (from 80 function, and GCN2 function is not required for JNK andto 200 J/m2 ), both the JNK and p38 MAP kinases are p38 MAP kinase signaling. The slight increase in eIF2�phosphorylated in GCN2�/� cells, to a similar extent phosphorylation by UV irradiation in p38 MAP kinaseas in wt cells (Figure 5A). Thus, GCN2 is not required �/� cells and in the presence of the JNK inhibitorfor the phosphorylation of JNK and p38 MAP kinases SP600125 could be due to enhanced UV stress in thoseby UV irradiation. To determine whether p38 MAP kinase cells. It is possible that the inhibition of the UV responsemay be involved in eIF2� phosphorylation, p38 MAP mediated by p38 MAP kinase and JNK results in thekinase �/� cells were UV irradiated (from 80 to 200 increased activity of other UV stress responses such as

the DNA damage-cell cycle checkpoint.J/m2 ). eIF2� phosphorylation was not reduced, but

GCN2 Represses Translation in UV Stress1283

Figure 4. Deletion of the GCN2 Gene Pre-vents the Attenuation of Protein Synthesis byUV Irradiation

(A) Wild-type (wt) and GCN2 knockout (ko)cells were subjected to 35S-methionine meta-bolic labeling after UV irradiation. Whole-cellextracts were resolved by SDS-PAGE (12%;each UV dose is presented in duplicate) andexposed to film. Same extracts were used forWestern blotting of actin.(B) A graphic presentation of three indepen-dent experiments as shown in (A). Trichloro-acetic acid (TCA) precipitable material wasmeasured by scintillation counting. Percent-age values were obtained by normalizing thecounts of treated samples against the con-trols.(C) Metabolic labeling was performed in wild-type and GCN2 knockout MEFs treated withthapsigargin (Tg), in parallel with UV irradia-tion shown in (B). C: control.

Discussion of the cellular sensors of environmental stresses, suchas UVC, that operate in other pathways also remainunknown. For example, it is not clear how UVC activatesOur study reveals that the inhibition of translation in

response to UVC irradiation is an active adaptation that JNK or p38 MAP kinases [30, 31]. However, as notedearlier [2], RNA molecules are prime candidates for sens-is not a direct consequence of damage to the translation

machinery, nor is it indirectly due to a block in gene ing UVC irradiation, as the RNA absorption maxima isat 254 nm.transcription or posttranscriptional mRNA processing.

Rather, like other forms of cellular stress, UVC irradiation In yeast, Gcn2p is activated by binding to unchargedtRNAs, which accumulate in amino acid-starved cellsis coupled to translational repression through eIF2�

phosphorylation. Our study establishes a causal rela- [32]. This binding occurs through the HisRS region ofGcn2p, which possesses three conserved motifs foundtionship between UV stress and translation inhibition,

and it goes on to identify the kinase responsible for in class II aminoacyl-tRNA synthetases [33]. The conser-vation of this region in GCN2 kinases across specieseIF2� phosphorylation in UVC-irradiated fibroblasts as

GCN2. We have not identified the cellular sensor of UVC suggests that they too are activated by uncharged tRNA.Indeed, targeted deletion of the murine GCN2 gene ab-that activates GCN2. It is noteworthy that the identity

Current Biology1284

Figure 5. UV-Induced eIF2� Phosphorylationby GCN2 Is Independent of JNK and p38 MAPKinase Pathways

(A) Western blot analysis of JNK and p38 MAPkinase phosphorylation. Cell lysates, pre-pared from UV-irradiated wild-type and GCN2knockout MEFs, were analyzed for phospho-JNK and phospho-p38 MAP kinase. Total lev-els for each protein are shown as a loadingcontrol.(B) Western blot analysis of p38 MAP kinaseand eIF2� phosphorylation. Cell lysates, pre-pared from UV-irradiated wild-type and p38MAP kinase knockout MEFs, were analyzedfor phospho-p38 MAP kinase and phospho-eIF2�. The total levels for each protein areshown as indicated.(C) Effect of JNK inhibitor SP600125 on phos-phorylation. 3T3 cells were pretreated with aspecific JNK inhibitor, SP600125 (25 �M, 15min), before being irradiated with an increas-ing UV dose. Cell lysates were analyzed byWestern blotting for eIF2� and c-Jun phos-phorylation. Total levels of eIF2� are shownas a loading control. C: control; DMSO: di-methyl sulphoxide.

rogated the increase in eIF2� phosphorylation in re- tion [18, 34–38]. However, when eIF2� phosphorylationis only mildly increased, the translation of certainsponse to amino acid starvation ([19] and Figure 3A).

We could detect no evidence for tRNA deacylation in mRNAs is actually increased. This paradoxical phenom-enon has been extensively studied in yeast, in whichUVC-irradiated cells (J.D., unpublished data). An alterna-

tive model for GCN2 activation by UVC is the induction of eIF2� is phosphorylated to low levels under amino acidstarvation conditions, leading to an upregulation of thecovalent crosslinking of GCN2 to aminoacylated tRNAs,

which could result in GCN2 activation. translation of the mRNA coding for the transcriptionfactor GCN4 [11]. A similar mechanism functions whenSignificant eIF2� phosphorylation by PKR, HRI,

GCN2, or PERK results in a general inhibition of transla- PERK is activated in mammalian cells, leading to in-

GCN2 Represses Translation in UV Stress1285

for 3 hr. Following stress treatments, cells were washed in ice-coldcreased translation of the transcription factor ATF4PBS supplemented with phosphatase inhibitors (100 mM NaF, 20mRNA [19]. It is thus conceivable that the translation ofmM Na3MO4, and 20 mM �-glycerophosphate) and 1 mM EDTAsuch mRNAs is also enhanced when GCN2 is activatedand were lysed in 250 �l lysis buffer [45]. Lysates were cleared by

in response to UVC initiating a stress-induced gene ex- centrifugation at 14,000 rpm, and 30 �g of extract from each samplepression program. was resolved by SDS-PAGE. Resolved proteins were transferred

eIF2� phosphorylation may also play a role in cell to nitrocellulose membranes (Schleicher and Schuell). Blots wereincubated with primary antisera (see below) and horseradish peroxi-cycle arrest following UV irradiation [39]. Phosphoryla-dase-conjugated anti-rabbit or anti-mouse antisera (Amershamtion of eIF2� as a consequence of UPR (unfolded proteinPharmacia Biotech, 1:2500). The following primary antibodies wereresponse) results in inhibition of cyclin D1 mRNA transla-used: rabbit anti-eIF2� phospho-Ser51 (Research Genetics,

tion, and this inhibition is responsible for the induction RG1001, 1:1000), monoclonal mouse anti-eIF2� (1:2500) [13], rabbitof a G1 phase arrest [40]. It was suggested that such anti-GCN2 (recognizing total protein or phospho-Thr898, 1:1000)an arrest might be required to provide sufficient time [19], rabbit anti-PKR (1:1000) [42], rabbit anti-JNK and anti-p38 MAPfor reestablishing cellular homeostasis following UPR. kinase (recognizing total proteins, JNK phospho-Thr183/Tyr185 or

p38 MAP kinase phospho-Thr180/Tyr182, Cell Signaling Technol-It is conceivable that cyclin D1 downregulation mightogy, 1:1000), rabbit anti-c-Jun (Cell Signaling Technology, 1:1000),also occur in response to UVC irradiation to arrest cellsand monoclonal mouse anti-actin (ICN, 1:5000). To detect phospho-at G1 to allow sufficient time for DNA repair.GCN2, treated cells were subjected to immunoprecipitation and

After completion of this work, a paper was published Western blotting, as described previously [19].by Wu et al. [41] that demonstrated that UVC elicits adose-dependent decrease in translation that temporally 35S-Methionine Metabolic Labelingcoincides with the phosphorylation of eIF2�. However, Cells were seeded into 6-well plates (1 � 105 cells/well, in triplicate).

Following stress treatment, cells were placed for 15 min in methio-Wu et al. conclude that eIF2� phosphorylation in re-nine-free medium (methionine- and glutamine-free DMEM supple-sponse to UV stress is effected by PERK. This conclu-mented with 10% dialyzed FBS and 2 mM glutamine), then pulse-sion was based primarily on the finding that a carboxy-labeled for 30 min with [35S]Met/Cys labeling mix (New Englandterminal-deleted dominant-negative form of PERK couldNuclear 772; 10 �Ci/well in 500 �l methionine-free medium). Cells

prevent UV-induced eIF2� phosphorylation. Also, eIF2� were washed twice with ice-cold PBS containing 1 mM EDTA andphosphorylation was measured 4 hr (as compared to were lysed in 250 �l/well RIPA buffer. Lysates were clarified byless than 1 hr in our study) following UVC irradiation and centrifugation, and 20 �l lysate was resolved by SDS-PAGE (12%).

Gels were fixed, dried, and exposed for autoradiography. In parallel,in different cell lines (MCF-7, HIT, COS-1). Our results20 �l lysate was subjected to trichloroacetic acid (TCA) precipitation.demonstrate that, under the conditions used here, PERKInsoluble material was measured by scintillation counting to monitoris not necessary for eIF2� phosphorylation in responsethe incorporation of 35S-methionine.to UVC irradiation, because the response remains intact

in PERK�/� cells. However, it is possible that PERK Acknowledgmentscontributes to translational attenuation upon UVC irradi-ation under some circumstances. The authors thank M. Karin for p38 MAP kinase MEFs; A. Hinnebusch

and T. Dever for critical reading of the manuscript; A. Brasey, V.In conclusion, our data demonstrate that UV stressEvdokimova, and F. Poulin for helpful discussions; and C. Lister forelicits an inhibition of general translation in mammaliantechnical assistance. This work was supported by a grant from thecells by activating the eIF2� kinase GCN2, thus identi-National Cancer Institute of Canada to N.S., who is a Canadianfying a mechanism for translation inhibition by UV irradi-Institute of Health Research distinguished scientist and a Howard

ation, and providing GCN2 as a novel mediator of cellular Hughes Medical Institute International Scholar; a grant from theresponse to UV stress. United States Public Health Service-National Institutes of Health to

D.R.; and a postdoctoral fellowship from Canadian Cancer ResearchExperimental Procedures Society to J.D.

Cell Culture Received: May 30, 2002Mouse embryonic fibroblasts (MEFs), or MEF cell lines, established Revised: June 25, 2002from wild-type (wt) and knockout PKR [42], PERK [43], GCN2 [19], Accepted: June 17, 2002and p38 MAP kinase [44] animals, or the eIF2� Ser51Ala mutant

Published: August 6, 2002knockin mouse [27], have been described previously. Unless other-wise specified, cells were maintained in Dulbecco’s modified Eagle’s

Referencesmedium (DMEM; Life Technologies) supplemented with 10% heat-inactivated fetal bovine serum (FBS), 2 mM L-glutamine and penicil-

1. Herrlich, P., Blattner, C., Knebel, A., Bender, K., and Rahmsdorf,lin-streptomycin (100 U/ml). For wt and eIF2� knockin MEFs, me-H.J. (1997). Nuclear and non-nuclear targets of genotoxicdium was supplemented with 1� nonessential amino acids andagents in the induction of gene expression. Shared principles1� essential amino acids. Leucine-free medium was prepared byin yeast, rodents, man and plants. Biol. Chem. 378, 1217–1229.adding 10% dialyzed FBS, 100 U/ml penicillin-streptomycin, and 2

2. Iordanov, M.S., Pribnow, D., Magun, J.L., Dinh, T.H., Pearson,mM L-glutamine to leucine- and glutamine-free DMEM (ICN).J.A., and Magun, B.E. (1998). Ultraviolet radiation triggers theribotoxic stress response in mammalian cells. J. Biol. Chem.Stress Treatment, Immunoprecipitation,273, 15794–15803.and Immunoblotting

3. Zhou, B.B., and Elledge, S.J. (2000). The DNA damage response:For UV irradiation, 4 � 105 cells seeded in 100-mm plates wereputting checkpoints in perspective. Nature 408, 433–439.washed twice with phosphate-buffered saline (PBS), and the PBS

4. Woodgate, R. (2001). Evolution of the two-step model for UV-was removed from the plates and cells were irradiated with themutagenesis. Mutat. Res. 485, 83–92.indicated UV dose (254 nm, UV stratalinker 2400, Stratagene). Cells

5. Norbury, C.J., and Hickson, I.D. (2001). Cellular responses towere then allowed to recover in complete media for the indicatedDNA damage. Annu. Rev. Pharmacol. Toxicol. 41, 367–401.times before harvest. For JNK inhibition, cells were pretreated for

6. Karin, M. (1998). Mitogen-activated protein kinase cascades15 min with a specific JNK inhibitor, SP600125 (25 �M, Calbiochem),as regulators of stress responses. Ann. N Y Acad. Sci. 851,before UV irradiation. Additionally, cells were treated with thapsigar-

gin (500 nM, Sigma) for 1 hr or were placed in leucine-free medium 139–146.

Current Biology1286

7. Davis, R.J. (2000). Signal transduction by the JNK group of MAP 28. Karin, M., and Hunter, T. (1995). Transcriptional control by pro-tein phosphorylation: signal transmission from the cell surfacekinases. Cell 103, 239–252.

8. Ono, K., and Han, J. (2000). The p38 signal transduction path- to the nucleus. Curr. Biol. 5, 747–757.29. Bennet, B.L., Sasaki, D.T., Murray, B.W., O’Leary, E.C., Sakata,way: activation and function. Cell. Signal. 12, 1–13.

9. Favre, A., Hajnsdorf, E., Thiam, K., and Caldeira de Araujo, A. S.T., Xu, W., Leisten, J.C., Motiwala, A., Pierce, S., Satoh, Y.,et al. (2001). SP600125, an anthrapyrazolone inhibitor of Jun(1985). Mutagenesis and growth delay induced in Escherichia

coli by near-ultraviolet radiations. Biochimie 67, 335–342. N-terminal kinase. Proc. Natl. Acad. Sci. USA 98, 13681–13686.30. Rosette, C., and Karin, M. (1996). Ultraviolet light and osmotic10. Liu, G., Zhang, Y., Bode, A.M., Ma, W., and Dong, Z. (2002).

Phosphorylation of 4E–BP1 is mediated by the p38/MSK1 path- stress: activation of the JNK cascade through multiple growthfactor and cytokine receptors. Science 274, 1194–1197.way in repsonse to UVB irradiation. J. Biol. Chem. 277, 8810–

8816. 31. Michelson, R., and Weinert, T. (1999). Sensor-less checkpointactivation? Nat. Cell Biol. 1, E177–E179.11. Hinnebusch, A.G. (2000). Mechanism and regulation of initiator

methionyl-tRNA binding to ribosomes. In Translational Control 32. Dong, J., Qiu, H., Garcia-Barrio, M., Anderson, J., and Hinne-busch, A.G. (2000). Uncharged tRNA activates GCN2 by displac-of Gene Expression, N. Sonenberg, J. Hershey, and M. Mathews,

eds. (Cold Spring Harbor, NY: Cold Spring Harbor Laboratory ing the protein kinase moiety from a bipartite tRNA-bindingdomain. Mol. Cell 6, 269–279.Press), pp. 185–243.

12. Clemens, M.J. (2001). Initiation factor eIF2 alpha phosphoryla- 33. Wek, S.A., Zhu, S., and Wek, R.C. (1995). The histidyl-tRNAsynthetase-related sequence in the eIF-2 alpha protein kinasetion in stress responses and apoptosis. Prog. Mol. Subcell. Biol.

27, 57–89. GCN2 interacts with tRNA and is required for activation in re-sponse to starvation for different amino acids. Mol. Cell. Biol.13. Scorsone, K.A., Panniers, R., Rowlands, A.G., and Henshaw,

E.C. (1987). Phosphorylation of eukaryotic initiation factor 2 15, 4497–4506.34. Chong, K.L., Feng, L., Schappert, K., Meurs, E., Donahue, T.F.,during physiological stresses which affect protein synthesis. J.

Biol. Chem. 262, 14538–14543. Friesen, J.D., Hovanessian, A.G., and Williams, B.R. (1992). Hu-man p68 kinase exhibits growth suppression in yeast and ho-14. Duncan, R.F. (1996). Translational control during heat shock. In

Translational Control, J. Hershey, M. Mathews, and N. Sonen- mology to the translational regulator GCN2. EMBO J. 11, 1553–1562.berg, eds. (Cold Spring Harbor, NY: Cold Spring Harbor Labora-

tory Press), pp. 271–294. 35. Ramirez, M., Wek, R.C., Vazquez de Aldana, C.R., Jackson,B.M., Freeman, B., and Hinnebusch, A.G. (1992). Mutations acti-15. Brostrom, C.O., and Brostrom, M.A. (1998). Regulation of trans-

lational initiation during cellular responses to stress. Prog. Nu- vating the yeast eIF-2 alpha kinase GCN2: isolation of allelesaltering the domain related to histidyl-tRNA synthetases. Mol.cleic Acid Res. Mol. Biol. 58, 79–125.

16. Natarajan, K., Meyer, M.R., Jackson, B.M., Slade, D., Roberts, Cell. Biol. 12, 5801–5815.36. Dever, T.E., Chen, J.J., Barber, G.N., Cigan, A.M., Feng, L.,C., Hinnebusch, A.G., and Marton, M.J. (2001). Transcriptional

profiling shows that Gcn4p is a master regulator of gene expres- Donahue, T.F., London, I.M., Katze, M.G., and Hinnebusch, A.G.(1993). Mammalian eukaryotic initiation factor 2 alpha kinasession during amino acid starvation in yeast. Mol. Cell. Biol. 21,

4347–4368. functionally substitute for GCN2 protein kinase in the GCN4translational control mechanism of yeast. Proc. Natl. Acad. Sci.17. Berlanga, J.J., Santoyo, J., and De Haro, C. (1999). Characteriza-

tion of a mammalian homolog of the GCN2 eukaryotic initiation USA 90, 4616–4620.37. Romano, P.R., Garcia-Barrio, M.T., Zhang, X., Wang, Q., Taylor,factor 2alpha kinase. Eur. J. Biochem. 265, 754–762.

18. Sood, R., Porter, A.C., Olsen, D.A., Cavener, D.R., and Wek, D.R., Zhang, F., Herring, C., Mathews, M.B., Qin, J., and Hinne-busch, A.G. (1998). Autophosphorylation in the activation loopR.C. (2000). A mammalian homologue of GCN2 protein kinase

important for translational control by phosphorylation of eukary- is required for full kinase activity in vivo of human and yeasteukaryotic initiation factor 2alpha kinases PKR and GCN2. Mol.otic initiation factor-2alpha. Genetics 154, 787–801.

19. Harding, H.P., Novoa, I.I., Zhang, Y., Zeng, H., Wek, R., Schapira, Cell. Biol. 18, 2282–2297.38. Sood, R., Porter, A.C., Ma, K., Quilliam, L.A., and Wek, R.C.M., and Ron, D. (2000). Regulated translation initiation controls

stress-induced gene expression in mammalian cells. Mol. Cell (2000). Pancreatic eukaryotic initiation factor-2alpha kinase(PEK) homologues in humans, Drosophila melanogaster and6, 1099–1108.

20. Shi, Y., Vattem, K.M., Sood, R., An, J., Liang, J., Stramm, L., Caenorhabditis elegans that mediate translational control in re-sponse to endoplasmic reticulum stress. Biochem. J. 346,and Wek, R.C. (1998). Identification and characterization of pan-

creatic eukaryotic initiation factor 2 alpha-subunit kinase, PEK, 281–293.39. Lakin, N.D., and Jackson, S.P. (1999). Regulation of p53 in re-involved in translational control. Mol. Cell. Biol. 18, 7499–7509.

21. Harding, H.P., Zhang, Y., and Ron, D. (1999). Protein translation sponse to DNA damage. Oncogene 18, 7644–7655.40. Brewer, J.W., and Diehl, J.A. (2000). PERK mediates cell-cycleand folding are coupled by an endoplasmic-reticulum-resident

kinase. Nature 397, 271–274. exit during the mammalian unfolded protein response. Proc.Natl. Acad. Sci. USA 97, 12625–12630.22. Kaufman, R.J. (1999). Double-stranded RNA-activated protein

kinase mediates virus-induced apoptosis: a new role for an old 41. Wu, S., Hu, Y., Wang, J., Chatterjee, M., Shi, Y., and Kaufman,R.J. (2002). Ultraviolet light inhibits translation through activa-actor. Proc. Natl. Acad. Sci. USA 96, 11693–11695.

23. Chen, J.J. (2000). Heme-regulated eIF2� kinase. In Translational tion of the unfolded protein response kinase PERK in the lumenof the endoplasmic reticulum. J. Biol. Chem. 277, 18077–18083.Control of Gene Expression, N. Sonenberg, J.W.B. Hershey,

and M.B. Mathews, eds. (Cold Spring Harbor, NY: Cold Spring 42. Abraham, N., Stojdl, D.F., Duncan, P.I., Methot, N., Ishii, T.,Dube, M., Vanderhyden, B.C., Atkins, H.L., Gray, D.A., McBur-Harbor Laboratory Press), pp. 529–546.

24. Kennedy, S.G., Kandel, E.S., Cross, T.K., and Hay, N. (1999). ney, M.W., et al. (1999). Characterization of transgenic micewith targeted disruption of the catalytic domain of the double-Akt/protein kinase B inhibits cell death by preventing the release

of cytochrome c from mitochondria. Mol. Biol. Cell 19, 5800– stranded RNA-dependent protein kinase, PKR. J. Biol. Chem.274, 5953–5962.5810.

25. Devary, Y., Gottlieb, R.A., Smeal, T., and Karin, M. (1992). The 43. Harding, H.P., Zhang, Y., Bertolotti, A., Zeng, H., and Ron, D.(2000). Perk is essential for translational regulation and cell sur-mammalian ultraviolet response is triggered by activation of Src

tyrosine kinases. Cell 71, 1081–1091. vival during the unfolded protein response. Mol. Cell 5, 897–904.44. Tamura, K., Sudo, T., Senftleben, U., Dadak, A.M., Johnson, R.,26. Engelberg, D., Klein, C., Martinetto, H., Struhl, K., and Karin, M.

(1994). The UV response involving the Ras signaling pathway and Karin, M. (2000). Requirement for p38alpha in erythropoietinexpression: a role for stress kinases in erythopoiesis. Cell 102,and AP-1 transcription factors is conserved between yeast and

mammals. Cell 77, 381–390. 221–231.45. Savinova, O., and Jagus, R. (1997). Use of vertical slab isoelec-27. Scheuner, D., Song, B., McEwen, E., Liu, C., Laybutt, R., Gilles-

pie, P., Saunders, T., Bonner-Weir, S., and Kaufman, R.J. (2001). tric focusing and immunoblotting to evaluate steady-state phos-phorylation of eIF2 alpha in cultured cells. Methods. 11,Translational control is required for the unfolded protein re-

sponse and in vivo glucose homeostasis. Mol. Cell 7, 1165– 419–425.1176.