JBC

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Originally published In Press as doi:10.1074/jbc.M314219200 on February 20, 2004

J. Biol. Chem., Vol. 279, Issue 19, 20108-20117, May 7, 2004
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
279/19/20108    most recent
M314219200v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Cullinan, S. B.
Right arrow Articles by Diehl, J. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Cullinan, S. B.
Right arrow Articles by Diehl, J. A.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

PERK-dependent Activation of Nrf2 Contributes to Redox Homeostasis and Cell Survival following Endoplasmic Reticulum Stress*

Sara B. Cullinan and J. Alan Diehl{ddagger}

From the Leonard and Madlyn Abramson Family Cancer Research Institute and Cancer Center, Department of Cancer Biology, University of Pennsylvania Cancer Center, Philadelphia, Pennsylvania 19104

Received for publication, December 28, 2003 , and in revised form, February 19, 2004.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
The accumulation of unfolded proteins elicits a cellular response that triggers both pro-survival and pro-apoptotic signaling events. PERK-dependent activation of NF-E2-related factor-2 (Nrf2) is critical for survival signaling during this response; however, the mechanism whereby Nrf2 confers a protective advantage to stressed cells remains to be defined. We now demonstrate that Nrf2 activation contributes to the maintenance of glutathione levels, which in turn functions as a buffer for the accumulation of reactive oxygen species during the unfolded protein response. The deleterious effects of Nrf2 or PERK deficiencies could be attenuated by the restoration of cellular glutathione levels or Nrf2 activity. In addition, the inhibition of reactive oxygen species production attenuated apoptotic induction following endoplasmic reticulum stress. Our data suggest that perturbations in cellular redox status sensitize cells to the harmful effects of endoplasmic reticulum stress, but that other factors are essential for apoptotic commitment.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
Mammalian cells possess a signaling network that senses the accumulation of misfolded proteins within the endoplasmic reticulum (ER).1 This signaling pathway, termed the unfolded protein response (UPR), is induced by agents that affect calcium homeostasis and protein glycosylation or through physiological stresses such as hypoxia and glucose deprivation; all of these stresses share a common capacity to disrupt the folding of proteins within the ER (1–3). The signaling network consists of three related transmembrane protein kinases (Ire1{alpha}, Ire1{beta}, and PERK), a transmembrane transcription factor (ATF6), and a transmembrane protease (caspase-12), which together coordinate cellular responses following UPR activation (1, 4–8). Ire1-dependent signaling in concert with ATF6 activation promotes expression of ER-localized chaperones, which facilitate the restoration of proper protein folding within the ER (6). UPR-mediated PERK activation impedes protein translation via phosphorylation-dependent inhibition of eukaryotic translation initiation factor-2{alpha} (eIF2{alpha}) (7, 9). Independent of its translational regulatory capacity, PERK-dependent signals elicit the activation of the prosurvival transcription factor NF-E2-related factor-2 (Nrf2) via site-specific phosphorylation (10).

PERK belongs to an eIF2{alpha} kinase family that includes the interferon-inducible, RNA-dependent protein kinase PKR, the heme-regulated kinase HRI, and GCN2 (11, 12). Of these, PERK function appears to be specifically required for the cellular response to ER stress (13). Following UPR activation, PERK-dependent phosphorylation of eIF2{alpha} at Ser51 attenuates translation of a majority of cellular proteins (7, 9) while paradoxically promoting increased translation of select target proteins, including ATF4 and the pro-apoptotic transcription factor CHOP/Gadd153 (14). The fact that mice harboring either a homozygous knock-in of a non-phosphorylatable eIF2{alpha} allele or a targeted deletion of PERK exhibit defects in glucose homeostasis and postnatal lethality (15–17) demonstrates that this translational regulation program is critical for organismal and cellular homeostasis.

Nrf2 activity has also been implicated in the promotion of cell survival following ER stress (10). Under basal conditions, Nrf2 is localized in the cytoplasm through an interaction with the BTB (broad complex, Tramtrack, and Bric-a-Brac) domain-containing protein Keap1 (18). PERK-dependent phosphorylation leads to the nuclear accumulation of Nrf2 and increased transcription of Nrf2 target genes (10). Previous work has revealed that cells deficient in Nrf2 (Nrf2-/-) display increased apoptosis and decreased survival in response to tunicamycin treatment (10), demonstrating that Nrf2 function is critical for cell survival following ER stress. Although the importance of Nrf2 signaling in response to ER stress is established, the mechanism whereby Nrf2 promotes cell survival remains undefined.

Nrf2 regulates transcriptional induction of Phase II detoxifying enzymes following exposure of cells to agents that promote the accumulation of reactive oxygen species (ROS) (19). Loss of Nrf2 through targeted gene deletion decreases cell survival in response to oxidative stress (20). In addition, glutathione levels are markedly decreased in Nrf2-/- cells compared with wild-type counterparts under conditions of stress and even under normal growth conditions (21, 22). Notably, cells deficient in either PERK or an Nrf2 heterodimeric partner, ATF4, also bear an increased oxidative stress burden (23). These observations suggest that perturbations in redox homeostasis are associated with ER stress and that PERK-dependent signals, including those that promote Nrf2 activation, function to buffer the increased oxidative stress burden associated with UPR activation.

We now demonstrate that ROS accumulation and an increased oxidative stress burden sensitize Nrf2-/- cells to apoptosis following UPR induction. Interference with ROS production via treatment of cells with ROS scavengers or restoration of glutathione levels delayed apoptosis, but was not sufficient to ablate ER stress-mediated cell death. These results suggest a model wherein ROS production sensitizes cells to ER stress, thereby contributing to apoptosis in cells deficient in components involved in signaling from the stressed ER.


   EXPERIMENTAL PROCEDURES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
RESULTS
 DISCUSSION
 REFERENCES
 
Tissue Culture Conditions and Plasmids—Cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, antibiotics, nonessential amino acids, {beta}-mercaptoethanol, and glutamine (Mediatech, Inc.). To derive {rho} zero cells, wild-type and Nrf2-/- murine embryo fibroblasts (20) immortalized via a standard 3T9 protocol (24) were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, antibiotics, nonessential amino acids, glutamine, 2 mM sodium pyruvate, 50 µM uridine, and 100 ng/ml ethidium bromide (25). Transfections were performed using LipofectAMINE Plus reagent (Invitrogen). During glucose deprivation, cells were cultured in Dulbecco's modified Eagle's medium without glucose (Invitrogen) and 10% dialyzed fetal calf serum (Invitrogen). All chemicals were purchased from Sigma. Plasmids encoding hemagglutinin (HA)-Nrf2, Myc-Keap1, and MafG have been described previously (10).

Immunofluorescence—Cells proliferating on glass coverslips were transfected and treated as described in text. Cells were fixed in 3% paraformaldehyde and permeabilized in 0.1% Triton X-100, in phosphate-buffered saline. The HA epitope was detected with monoclonal antibody 12CA5. Cells were stained with fluorescein isothiocyanate-conjugated IgG (Vector Labs, Inc.). DNA was detected with Hoechst 33258 (Sigma). Cells were visualized using a Nikon microscope fitted with appropriate filters.

Electrophoretic Mobility Shift Assays—10 µg of cell extract (lysed in 50 mM Tris (pH 7.5), 250 mM NaCl, 5 mM EDTA, 0.1% Triton X-100, and 4 mM dithiothreitol) was incubated for 4 h at 4 °C in a reaction mixture containing 20 mM HEPES (pH 8.0), 1 mM EDTA, 20 mM KCl, 4 mM MgCl2, 4% glycerol, 1 µg of poly(dI-dC), 5 mM dithiothreitol, and the indicated antibody. Following the incubation, a consensus antioxidant response element oligomer (10) end-labeled with [{gamma}-32P]ATP was added, and the reaction mixture was incubated at 37 °C for 30 min. Reactions were resolved on native gels, dried, and visualized by autoradiography.

Annexin V Staining and Clonogenic Survival Assays—Cells proliferating on glass coverslips were left untreated, were treated with tunicamycin (5 µg/ml), or were cultured in glucose-free medium in the absence or presence of 5 mM N-acetylcysteine or 1 mM cysteine for the indicated intervals. Cells were then washed with phosphate-buffered saline and stained with fluorescein isothiocyanate-conjugated anti-Annexin V antibody and propidium iodide (Pharmingen) and examined by fluorescence and light microscopy. Annexin V-positive/propidium iodide-negative cells are expressed as the percentage of total cells. For colony outgrowth assays, cells were plated at 2 x 104/60-mm dish; and after 24 h, they were left untreated, were treated with tunicamycin (5 µg/ml), or were cultured in glucose-free medium for the indicated intervals. Following treatment, cells were washed with phosphate-buffered saline, refed with complete medium, and allowed to grow for 7 days. Cells were visualized with Giemsa stain (Sigma).

Immunoblotting, Cellular Fractionation, and Immunoprecipitation— For detection of phosphorylated eIF2{alpha} and total eIF2{alpha}, cells were lysed in 1% SDS, and proteins were resolved by SDS-PAGE. Following transfer to a nitrocellulose membrane (Osmonics, Inc.), total eIF2{alpha} and phosphorylated eIF2{alpha} (Cell Signaling) were detected by immunoblotting. For detection of CHOP, lysates from wild-type and Nrf2-/- fibroblasts left untreated or cultured in glucose-free medium for 4 h or from NIH-3T3 cells were prepared in EBC buffer, transfected as indicated, and treated with tunicamycin. Proteins were resolved by SDS-PAGE. Following transfer to a nitrocellulose membrane, CHOP was detected using an anti-CHOP polyclonal antibody (F-168, Santa Cruz Biotechnology). To collect nuclear extracts, wild-type and PERK-deficient (PERK-/-) fibroblasts were cultured in glucose-free medium for the indicated intervals. Nuclei were collected in buffer containing 10 mM HEPES (pH 8.0), 10 mM KCl, 1.5 mM MgCl2, 0.1 mM EDTA, 0.1 mM EGTA, and 0.1% IGEPAL and in buffer containing 420 mM NaCl, 1.5 mM MgCl2, 1.0 mM EDTA, 1.0 mM EGTA, and 20% glycerol and stored in buffer containing 20 mM HEPES (pH 8.0), 50 mM KCl, 0.2 mM EDTA, 0.2 mM EGTA, and 20% glycerol. To detect total Nrf2 levels, wild-type and PERK-/- fibroblasts were lysed in EBC buffer. Nrf2 was precipitated with an anti-Nrf2 polyclonal antibody (H-300, Santa Cruz Biotechnology), and precipitated proteins were resolved by SDS-PAGE, transferred to nitrocellulose membranes, and blotted with an anti-Nrf2 polyclonal antibody. All buffers were supplemented with 0.1 mM phenylmethylsulfonyl fluoride, 1 µg/ml aprotinin, 1 µg/ml leupeptin, 1 µg/ml pepstatin, and 25 mM {beta}-glycerol phosphate. Purity of nuclear extracts was routinely monitored via immunoblotting of cAMP-responsive element-binding protein (nuclear) and {beta}-tubulin (cytoplasmic).

Northern and Southern Blotting—RNA was extracted using TRIzol reagent (Invitrogen). RNA immobilized on nylon membranes (Osmonics, Inc.) was hybridized with 32P-labeled probes specific for the glutamate-cysteine ligase catalytic subunit (GCLC; {gamma}-glutamylcysteine synthetase), NADPH-quinone oxidoreductase-1 (NQO1), CHOP, or {gamma}-actin. Signals were detected by phosphorimaging. To verify {rho} zero status, genomic DNA was collected from parental cells and from cells cultured in the presence of ethidium bromide and digested with SpeI. Digested DNA immobilized on nylon membranes was hybridized with a 32P-labeled probe specific for cytochrome c oxidase-2 (primers 5'-CCTACCCATTCCAACTTGGTC and 5'-CCACAAATTTCAGAGCATTGG). Signals were detected by phosphorimaging.

Detection of ROS and Glutathione—To measure total glutathione and GSSG levels, cells were grown in 6-well plates and treated as indicated. Cells were scraped into glutathione assay buffer (100 mM sodium phosphate and 1 mM EDTA (pH 7.5)) with 0.1% Triton X-100. Total protein was measured using the BCA assay kit (Pierce), and all results were normalized according to protein concentration. Proteins were precipitated with 5% 5-sulfosalicylic acid, and samples were assayed following dilution of 5-sulfosalicylic acid to <0.1%. Glutathione measurements were completed in a reaction mixture containing cell lysates in 15 mM 5,5'-dithiobis(2-nitrobenzoic acid), 0.2 mM NADPH, and 1.0 unit/ml glutathione reductase (26). Reactions were carried out in 96-well plates and measured at 405 nm over 5 min using SoftMax software and an automated plate reader. In the rescue experiments, Nrf2-/- or PERK-/- fibroblasts were transfected with an empty vector or with a vector encoding HA-Nrf2 or Nrf2 containing a mutant nuclear localization signal, and wild-type fibroblasts were transfected with an empty vector. Total glutathione levels were then measured as described above. To measure oxidized glutathione levels, reduced glutathione was derivatized following protein precipitation using 2-vinylpyridine, and reactions were then carried out as described above. All experiments were carried out in duplicate and measured in triplicate. All chemicals were purchased from Sigma.

To detect endogenous peroxide production, cells were left untreated or were treated with tunicamycin (5 µg/ml) for the indicated intervals in the presence of 2',7'-dichlorofluorescin diacetate (DCF) fluorescent dye (10 µM; Molecular Probes, Inc.). Cells were harvested, and fluorescence emission at 520 nm was measured using a fluorometer with an automated plate reader. All experiments were carried out in triplicate and measured in duplicate.

The presence of carbonyl-modified cellular proteins was detected using the Oxyblot kit (Chemicon International, Inc.) according to the manufacturer's instructions. Cells were treated as indicated and lysed in EBC buffer containing protease and phosphatase inhibitors. Carbonyl groups were derivatized to 2,4-dinitrophenylhydrazone (DNP) via the addition of 2,4-dinitrophenylhydrazine (DNPH). Proteins were resolved by SDS-PAGE; and following transfer to a nitrocellulose membrane, the presence of DNP was detected using an antibody specific for DNP. Control reactions were carried out in parallel using non-DNPH-treated extracts to confirm the specificity of the antibody.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
DISCUSSION
 REFERENCES
 
Activation of the Nrf2 Transcription Factor and Nrf2 Target Genes in Response to Glucose Deprivation—We previously demonstrated that Nrf2 is activated following a pharmacologically induced UPR via PERK-mediated phosphorylation (10). To determine whether Nrf2 participates in a physiological response to ER stress, we examined PERK-dependent regulation of Nrf2 in cells deprived of glucose. Although glucose starvation affects global cellular metabolism, a more proximal effect is the interference with proper protein glycosylation in the ER (2). Initially, we examined the kinetics of UPR activation in wild-type and Nrf2-/- mouse fibroblasts (20) following glucose starvation. UPR activation of PERK results in phosphorylation of both eIF2{alpha} and Nrf2 and a concomitant increase in expression of the CHOP transcription factor (7, 9, 10, 14). eIF2{alpha} phosphorylation was evident in both wild-type and Nrf2-/- cells by 1–2 h of glucose starvation (Fig. 1A, lanes 1–3 and 6–8). These data demonstrate that the UPR and, more specifically, PERK activation occur following glucose restriction in Nrf2-/- cells.



View larger version (48K):
[in this window]
[in a new window]
  		FIG. 1.
Glucose deprivation activates Nrf2. A, wild-type (WT; lanes 1–5) and Nrf2-/- (lanes 6–10) cells were cultured in glucose-free medium for the indicated intervals. Proteins were resolved by SDS-PAGE and immunoblotted with antibodies specific for Ser51-phosphorylated eIF2{alpha} (peIF2{alpha}; upper panels) and total eIF2{alpha} (lower panels). B, total cell lysates from wild-type and Nrf2-/- cells cultured in complete or glucose-free medium for 4 h were resolved by SDS-PAGE and immunoblotted with an anti-CHOP antibody. C, RNA was collected from wild-type (lanes 1–5) and Nrf2-/- (lanes 6–10) fibroblasts deprived of glucose for the indicated intervals. Expression of CHOP and actin was determined by Northern blot analysis. D, total cell lysates from NIH-3T3 cells mock-transfected (lanes 1–4) or transfected with a plasmid encoding HA-Nrf2 (lanes 5–8) or with plasmids encoding HA-Nrf2 and MafG (lanes 9–12) and treated with tunicamycin (Tun; 5 µg/ml) for the indicated intervals were resolved by SDS-PAGE and immunoblotted with an anti-CHOP antibody and monoclonal antibody 12CA5. E, NIH-3T3 cells proliferating on glass coverslips were co-transfected with plasmids encoding HA-Nrf2 and Myc-Keap1. Transfected cells were glucose-deprived for the indicated intervals and fixed, and Nrf2 was visualized by fluorescence microscopy with an anti-HA epitope antibody. DNA was stained with Hoechst dye. F, nuclear extracts were prepared from wild-type (lanes 1–4) and PERK-/- (lanes 5–8) cells deprived of glucose for the indicated intervals, and Nrf2 levels were assessed by immunoprecipitation followed by immunoblot analysis with the Nrf2 antiserum. Lanes 4 and 8 are control immunoprecipitations carried out with normal rabbit serum (NRS). G, total cell lysates were prepared from wild-type (lanes 1–4) and PERK-/- (lanes 5–8) fibroblasts deprived of glucose for the indicated intervals, and Nrf2 levels were assessed by immunoprecipitation followed by immunoblot analysis with the Nrf2 antiserum. Lanes 4 and 8 are control immunoprecipitations.

 
Strong CHOP accumulation was also noted in both cell types following glucose limitation (Fig. 1B). We noted that the CHOP protein was present under basal conditions in Nrf2-/- cells (compare lanes 1 and 3), albeit at levels lower than in stressed cells. This result suggests that Nrf2 might function directly or indirectly as a negative regulator of CHOP expression. Consistent with this, we noted increased kinetics of CHOP mRNA accumulation in Nrf2-/- fibroblasts following ER stress relative to that in wild-type fibroblasts (Fig. 1C, compare lanes 1–5 and 6–10). In addition, although Nrf2 expression alone had a minimal impact on CHOP protein levels, we found that expression of Nrf2 along with its heterodimeric partner MafG (27) greatly attenuated CHOP induction following treatment of cells with tunicamycin (Fig. 1D). Similar results were observed when Nrf2 was co-expressed with ATF4, another Nrf2 binding partner (28) (data not shown). These results suggest that Nrf2 negatively regulates CHOP expression both basally and in response to ER stress. Additional work will be required to assess whether Nrf2 can directly regulate CHOP gene expression.

UPR-dependent activation via glycosylation inhibitors or perturbations in ER calcium concentration promotes PERK-dependent phosphorylation and directs the nuclear accumulation of transcriptionally active Nrf2 (10). Under homeostatic conditions, Nrf2 is sequestered in the cytoplasm by virtue of its interaction with Keap1; following stress, Nrf2 dissociates from Keap1, whereupon it translocates into the nucleus (18). To assess Nrf2 regulation in cells deprived of glucose, NIH-3T3 cells were co-transfected with plasmids encoding HA-Nrf2 and Myc-Keap1. In unstressed cells, HA-Nrf2 was exclusively cytoplasmic (Fig. 1E, panel a). Within 1 h of glucose deprivation, a significant fraction of HA-Nrf2 was nuclear (panel d) and remained nuclear for at least 2 h of glucose deprivation (panel g). We also assessed endogenous Nrf2 nuclear localization in wild-type and PERK-/- fibroblasts (13). Nuclear extracts were prepared from cells cultured in complete or glucose-free medium for the indicated intervals. Nrf2 nuclear accumulation was assessed by immunoprecipitation, followed by immunoblot analysis. Nrf2 nuclear accumulation was observed upon glucose restriction in wild-type fibroblasts (Fig. 1F, lanes 2 and 3), but not in PERK-/- fibroblasts (lanes 6 and 7). As a control, total Nrf2 levels in wild-type and PERK-/- fibroblasts were measured during the course of glucose deprivation (Fig. 1G). Basal Nrf2 levels were comparable in the two cell lines (compare lanes 1 and 5), and Nrf2 levels were constant in the PERK-/- cells (lanes 5–7). We did note slightly higher Nrf2 levels in wild-type cells after 1 h of glucose treatment (compare lanes 1 and 2). This change may be the result of reported increased Nrf2 accumulation following stress induction (29, 30).

We next assessed mRNA accumulation of two documented Nrf2 target genes, GCLC (21, 27) and NQO1 (31), following glucose restriction. Both GCLC and NQO1 were induced following glucose limitation in wild-type cells (Fig. 2, compare lanes 1 and 4), but not in Nrf2-/- fibroblasts (lanes 6–10). Consistent with PERK being a requisite proximal activator of Nrf2, induction of neither GCLC nor NQO1 was observed in PERK-/- fibroblasts (lanes 11–15). The increased levels of GCLC and NQO1 seen in PERK-/- fibroblasts may be a result of transient Nrf2 activation in response to normal cellular stresses (32).



View larger version (35K):
[in this window]
[in a new window]
  		FIG. 2.
PERK- and Nrf2-dependent expression of Nrf2 target genes. RNA was collected from wild-type (WT; lanes 1–5), Nrf2-/- (lanes 6–10), and PERK-/- (lanes 11–15) fibroblasts deprived of glucose for the indicated intervals. Expression of GCLC, NQO1, and actin was determined by Northern blot analysis.

 
Increased Sensitivity to ER Stress Correlates with Increased Oxidative Stress Burden—To determine the impact of Nrf2 deficiency on cell survival following ER stress induced by transient glucose restriction, wild-type and Nrf2-/- fibroblasts were plated at low density, cultured in normal or glucose-free medium for 8 or 16 h, and then refed with complete glucose-containing medium. Cells were allowed to grow for 7 days, at which point colonies were visualized with Giemsa stain. Nrf2 deficiency greatly diminished the ability of cells to survive glucose restriction (Fig. 3A). This result demonstrates that Nrf2 deficiency increases cell sensitivity to the pro-apoptotic effects of nutrient restriction. Consistent with this, Nrf2-/- fibroblasts also exhibited increased levels of Annexin V staining following glucose restriction (Fig. 3, B and C, triangles).



View larger version (32K):
[in this window]
[in a new window]
  		FIG. 3.
Redox sensitivity to ER stress in Nrf2-/- cells. A, wild-type (WT) and Nrf2-/- cells plated at low density were glucose-starved for the indicated intervals and then refed complete medium. After 7 days, colonies were visualized with Giemsa stain. Wild-type ({diamondsuit}) and Nrf2-/- ({blacktriangleup}) cells proliferating on glass coverslips were deprived of glucose for the indicated intervals in the absence or presence of cysteine (B) or N-acetylcysteine (C). Wild-type ({blacksquare}) and Nrf2-/- (•) cells grown in supplemented medium are indicated. Cells were stained with propidium iodide and fluorescein isothiocyanate-conjugated Annexin V and visualized by light and fluorescence microscopy. Annexin V-positive cells are shown as the percentage of total cells examined. Error bars indicate S.D. from a minimum of three independent experiments.

 
We next considered the possibility that Nrf2-/- fibroblasts might exhibit reduced levels of glutathione following ER stress given that they are deficient in GCLC and NQO1 production. Nrf2-/- fibroblasts exhibited lower basal levels of glutathione compared with wild-type cells (Fig. 4, A and B, compare black and hatched bars). In addition, a rapid decrease in glutathione levels (to nearly 50% of that in untreated cells) was observed in response to glucose deprivation in Nrf2-/- cells (Fig. 4A). A similar trend was noted for GSSG levels in Nrf2-/- cells (data not shown). In contrast, glucose deprivation resulted in a transient increase in glutathione levels in wild-type cells (Fig. 4A, black bars). These results suggest that Nrf2 activation contributes to the maintenance of intracellular glutathione levels following ER stress.



View larger version (39K):
[in this window]
[in a new window]
  		FIG. 4.
Decreased intracellular glutathione and increased ROS accumulation in Nrf2-/- fibroblasts. A, wild-type (black bars) and Nrf2-/- (hatched bars) fibroblasts were glucose-restricted for the indicated intervals, and total intracellular glutathione levels were measured. Basal glutathione levels in wild-type cells were set at 1. Error bars represent S.D. from at least three independent experiments. *, p < 0.05 using paired Student's t test. t tests compare the changes within genotypes during the course of glucose deprivation. B, mock-transfected wild-type cells (WT; black bars), mock-transfected Nrf2-/- cells (hatched bars), and Nrf2-/- cells transfected with a plasmid encoding Nrf2 (gray bars) were left untreated or were glucose-deprived for 4 h. Total intracellular glutathione levels were measured, and basal glutathione levels in mock-transfected wild-type cells were set at 1. Error bars indicate S.D. from at least three independent experiments. *, p < 0.05; **, p < 0.01 (paired Student's t test). t tests compare the glutathione levels of mock-transfected Nrf2-/- cells with those of Nrf2-/- cells expressing ectopic Nrf2 at matched time points. C, wild-type (black bars) and Nrf2-/- (hatched bars) cells were treated with 5 µg/ml tunicamycin (Tun) and 10 µM DCF for the indicated intervals. Cells were collected, and DCF fluorescence was measured. Measurements obtained from untreated wild-type cells were set at 1. Error bars represent S.D. from three independent experiments. *, p < 0.05 (paired Student's t test). t tests compare DCF readings within a genotype over the course of tunicamycin treatment. D, cell lysates were collected from wild-type (upper panel) and Nrf2-/- (lower panel) cells cultured in glucose-free medium for the indicated intervals. Carbonyl groups were derivatized to DNP via the addition of DNPH. Proteins were separated by SDS-PAGE, and the presence of DNP was detected by immunoblotting using an anti-DNP antibody.

 
To confirm this finding, Nrf2-/- cells were transfected with a plasmid encoding HA-Nrf2, and total glutathione levels were measured following glucose deprivation. The introduction of Nrf2 into the Nrf2-/- background restored cellular glutathione levels to levels comparable with those seen in wild-type cells (Fig. 4B, compare black and gray bars). Glutathione levels remained elevated in Nrf2-transfected cells even after 4 h of glucose deprivation (Fig. 4B), indicating that the loss of Nrf2 contributes to the decreased glutathione levels observed in Nrf2-/- cells. In addition, we were able to restore glutathione levels in Nrf2-/- cells via supplementation of the cell culture medium with cysteine (data not shown).

The reduction in cellular glutathione levels in Nrf2-/- cells suggested that Nrf2 deficiency might sensitize cells to perturbations in redox homeostasis. In fact, UPR activation is associated with oxidative stress, and cells containing a targeted deletion of PERK exhibit increased levels of intracellular ROS in response to ER stress (23). Previous studies indicate that Nrf2-/- cells are subject to impaired redox homeostasis (22). To assess the accumulation of ROS in response to the UPR in Nrf2-/- cells, wild-type and Nrf2-/- cells were left untreated or were treated with tunicamycin (5 µg/ml). Following treatment, ROS levels were measured using the fluorescent dye DCF. Basal ROS levels were higher in Nrf2-/- cells compared with their wild-type counterparts (Fig. 4C), and high ROS levels were evident throughout the time course. In wild-type cells, ROS levels steadily increased though 8 h of stress exposure.

The accumulation of ROS in cells results in the modification of cellular proteins through the addition of carbonyl groups (33). Therefore, as a complementary approach for assessing ROS, we measured the abundance of carbonyl-modified proteins following glucose restriction. Although untreated wild-type cells contained relatively low levels of carbonyl-modified proteins, Nrf2-/- cells contained a significantly higher level of modified proteins (Fig. 4D, compare lanes 1 in upper and lower panels). A steady increase in the amount of carbonyl-modified proteins was noted in wild-type cells in response to glucose starvation, which by 8 h approached that detected in Nrf2-/- fibroblasts (upper panel). Nrf2-/- cells displayed a consistently high level of carbonyl-modified proteins in the absence or presence of glucose deprivation (lower panel).

Based on the reduced levels of intracellular glutathione in Nrf2-/- cells and their rapid accumulation of ROS, we postulated that providing intermediates that feed into the glutathione biosynthetic pathway downstream of Nrf2 target enzymes should rescue sensitivity of Nrf2-/- fibroblasts to ER stress. A similar strategy was utilized to rescue cells deficient in GCLC (34), an Nrf2 target gene that is normally induced in response to ER stress (Fig. 2) (10). Wild-type and Nrf2-/- fibroblasts were glucose-restricted for the indicated intervals in the absence or presence of cysteine (1 mM) (Fig. 3B) or N-acetylcysteine (5 mM) (Fig. 3C) and assessed for Annexin V staining. Glucose starvation elicited a rapid induction of apoptosis in Nrf2-/- cells, but not in wild-type cells (Fig. 3, B and C). However, the addition of either cysteine or N-acetylcysteine to the culture medium suppressed apoptosis in Nrf2-/- cells to levels comparable with those in wild-type cells (Fig. 3, B and C). Similar results were observed when Nrf2-/- cells were challenged with tunicamycin, but supplemented with N-acetylcysteine (data not shown). Likewise, the ER stress-dependent cell death of PERK-/- cells can also be reduced through the addition of cysteine to the culture medium (23). Taken together, the above results suggest that Nrf2-/- cells are highly sensitive to ER stress because of their inability to increase expression of Nrf2 target genes that combat oxidative stress during the UPR. Collectively, our data demonstrate PERK-dependent nuclear translocation and activation of Nrf2 following cellular glucose restriction and provide evidence that Nrf2 activity contributes to the maintenance of redox homeostasis in cells experiencing ER stress.

Loss of Nrf2 Signaling Contributes to ER Stress Sensitivity in PERK-/- Cells—During the UPR, PERK signaling activates Nrf2, leading to increased expression of Nrf2 target genes (Fig. 2) (10). As PERK-/- cells are deficient in Nrf2 activation following UPR activation (10), we reasoned that they should experience severe decreases in intracellular glutathione levels. Indeed, although PERK-/- cells are able to maintain basal glutathione levels, glucose restriction resulted in a rapid decrease in intracellular glutathione levels (Fig. 5A, hatched bars). By 8 h of glucose restriction, glutathione levels decreased by 50% in PERK-/- cells (hatched bars), whereas at the same time point, wild-type cells, in which Nrf2 is activated in response to ER stress (10), maintained cellular glutathione levels (black bars). To directly assess the importance of Nrf2 in the maintenance of glutathione levels, PERK-/- cells were transfected with a plasmid encoding HA-Nrf2, and total glutathione levels were measured following glucose restriction. Under these conditions Nrf2 will be constitutively active, as exogenous Keap1 is not provided (18). Ectopic expression of Nrf2 in PERK-/- cells resulted in the maintenance of glutathione levels during ER stress conditions in a manner similar to that in wild-type cells (gray bars). To confirm that the rescue is dependent upon active Nrf2, we also performed the experiment using an Nrf2 mutant that is unable to enter the nucleus. Expression of this mutant did not increase cellular glutathione levels in the PERK-/- background (data not shown), indicating that Nrf2 activity is responsible for the maintenance of cellular glutathione levels during the UPR.



View larger version (39K):
[in this window]
[in a new window]
  		FIG. 5.
Oxidative stress in PERK-/- cells. A, mock-transfected wild-type cells (WT; black bars), mock-transfected PERK-/- cells (hatched bars), and PERK-/- cells transfected with a plasmid encoding HA-Nrf2 (gray bars) were cultured in glucose-free medium for the indicated intervals, and total intracellular glutathione levels were measured. Within each cell type, basal glutathione levels were set at 1, and the data represent the change in glutathione levels within each cell type over time. Error bars represent S.D. from three independent experiments. B, whole cell extracts were collected from wild-type and PERK-/- cells left untreated (lanes 1 and 5), treated with 2.5 µg/ml tunicamycin (tun; lanes 2 and 6) with 10 µM 2-deoxyglucose (2-DOG; lanes 3 and 7), or cultured in glucose-free medium (lanes 4 and 8) for 6 h. Carbonyl groups were derivatized to DNP via the addition of DNPH. Proteins were separated by SDS-PAGE, and the presence of DNP was detected by immunoblot analysis using an anti-DNP antibody.

 
We next assessed the level of oxidative stress experienced by PERK-/- cells in response to ER stress by measuring the level of carbonyl-modified proteins. Little difference in total oxidative stress was noted in untreated cells of either genetic background (Fig. 5B, compare lanes 1 and 5), suggesting that PERK does not contribute to redox homeostasis in the absence of ER stress. However, in response to various ER stress conditions, PERK-/- cells displayed a rapid increase in the amount of carbonyl-modified proteins (lanes 6–8). This finding is in agreement with previous data demonstrating increased DCF fluorescence in PERK-/- cells following tunicamycin treatment (23). In contrast, in wild-type cells, a more modest increase in the level of carbonyl-modified proteins was evident following ER stress treatments (lanes 2–4). These data suggest that the increased sensitivity to ER stress experienced by PERK-/- cells is due, at least in part, to the lack of Nrf2 signaling in these cells.

Mitochondrial ROS Contribute to UPR-associated Oxidative Stress—Increased oxidative stress is evident in PERK-/- and Nrf2-/- cells. The increased oxidative burden is likely to contribute to the increased sensitivity of both cell types to ER stress-induced apoptosis (23). Although ROS are produced primarily by the mitochondria as a by-product of oxidative phosphorylation (35), the depletion of cellular glutathione resulting from perturbations in the ER environment would likely contribute to an increased oxidative mitochondrial burden and thereby increase intracellular ROS levels. We therefore set out to assess the role of mitochondrial ROS in sensitization of cells to ER stress-induced apoptosis through generation of wild-type and Nrf2-/- fibroblasts lacking a mitochondrial genome and therefore a functional mitochondrial respiratory chain ({rho} zero cells) (25). If the mitochondria are the source of ROS in Nrf2-/- fibroblasts, then their {rho} zero derivatives should be resistant to ER stress. To verify that {rho} zero cells were deficient in mitochondrial DNA, we assessed the genomic status of cytochrome c oxidase-2, a component of the mitochondrial electron transport chain. Cytochrome c oxidase-2 DNA was present in wild-type and Nrf2-/- parental cells, but was absent in their {rho} zero derivatives (Fig. 6A). Additionally, we confirmed that {rho} zero cells did not accumulate ROS in response to ER stress. Because Nrf2-/- cells constitutively produced high ROS levels (Fig. 4), we initially compared the Nrf2-/- parental and {rho} zero cells during a time course of tunicamycin treatment. We were unable to conduct this experiment under conditions of glucose deprivation given that {rho} zero cells are dependent on glucose for all cellular metabolism. In contrast to the parental cells, which produced high ROS levels throughout the experiment, Nrf2-/- {rho} zero cells did not accumulate ROS in response to tunicamycin treatment (5 µg/ml), as judged by monitoring of carbonyl-modified proteins (Fig. 6B). Likewise, wild-type {rho} zero cells did not accumulate high levels of carbonyl-modified proteins following tunicamycin treatment (data not shown).



View larger version (45K):
[in this window]
[in a new window]
  		FIG. 6.
Generation of UPR-competent {rho} zero cells. A, DNA was collected from wild-type (WT) and Nrf2-/- cells cultured in the absence or presence of ethidium bromide and probed for the presence of cytochrome c oxidase-2 (Cox2) DNA by Southern blot analysis. B, whole cell extracts were collected from Nrf2-/- cells cultured in the absence (lanes 1 and 2) or presence (lanes 3 and 4) of EtBr and were left untreated (lanes 1 and 3) or were treated with 5 µg/ml tunicamycin (Tun; lanes 2 and 4) for 6 h. Carbonyl groups were derivatized to DNP via the addition of DNPH. Proteins were resolved by SDS-PAGE, and the presence of DNP was detected by immunoblot analysis. C, lysates were collected from wild-type and Nrf2-/- cells cultured in the absence (lanes 1–4 and lanes 9–12, respectively) or presence (lanes 5–8 and lanes 13–16, respectively) of EtBr and treated with 5 µg/ml tunicamycin for the indicated intervals. Proteins were resolved by SDS-PAGE and immunoblotted with antibodies specific for Ser51-phosphorylated eIF2{alpha} (peIF2{alpha}; upper panel) and total eIF2{alpha} (lower panel). D, cell lysates from wild-type and {rho} zero cells treated with tunicamycin (5 µg/ml) were mixed with a radiolabeled antioxidant response element probe in electrophoretic mobility shift assay reactions. Protein/DNA adducts were visualized by autoradiography. NRS, normal rabbit serum.

 
To confirm that the UPR remains intact in {rho} zero cells, eIF2{alpha} phosphorylation was monitored during a time course of tunicamycin treatment. The induction of phosphorylated eIF2{alpha} was similar in both {rho} zero and parental cell lines (Fig. 6C). Induction of CHOP, an ER stress-inducible transcription factor (36), was also noted in both {rho} zero and parental cell lines (data not shown). To ensure that interference with mitochondrial signaling did not affect UPR-dependent Nrf2 signaling, we assessed UPR-dependent induction of Nrf2/DNA binding activity in {rho} zero cells. Extracts were prepared from untreated and tunicamycin-treated (5 µg/ml) wild-type and {rho} zero cells. Using a radiolabeled probe containing a consensus antioxidant response element, electrophoretic mobility shift assays were performed in the absence or presence of an anti-Nrf2 antibody to identify Nrf2-specific DNA binding; the addition of an anti-Nrf2 antibody ablates Nrf2/DNA complex formation (10). Basal DNA binding activity was detected in both cell lines, but this activity was not competed by the Nrf2 antiserum (Fig. 6D, lanes 3 and 7). This activity could be competed by an anti-Nrf1 antibody (data not shown), consistent with Nrf1 being a constitutively active member of the Cap "N" Collar family (37). Treatment of either wild-type or {rho} zero cells with tunicamycin resulted in increased binding to the antioxidant response element oligonucleotide; this activity was reduced by the addition of an anti-Nrf2 antibody (lanes 5 and 9), but not normal rabbit serum (lanes 4 and 8) or an anti-Nrf1 antibody (data not shown). We noted increased Nrf2/DNA binding activity in the stressed {rho} zero cells; however, the nature of this difference is unknown. We also noted endogenous Nrf2 nuclear import in these cells during tunicamycin treatment (data not shown), consistent with the increased formation of Nrf2/DNA adducts.

To assess the effect that the loss of mitochondrial respiration would have on {rho} zero cells during the UPR, cells were treated with tunicamycin (5 µg/ml) and then evaluated for levels of apoptotic commitment via Annexin V staining. Similar to previous results (10), Nrf2-/- cells displayed increased sensitivity to tunicamycin compared with their wild-type counterparts (Fig. 7A). In striking contrast, Nrf2-/- {rho} zero cells displayed levels of increased Annexin V staining that were comparable with wild-type cells (Fig. 7A), suggesting that mitochondrially produced ROS are responsible for the heightened sensitivity of Nrf2-/- cells to ER stress. Similarly, recent data demonstrate that PERK-/- {rho} zero cells are able to evade ER stress-dependent cell death (23). To confirm that {rho} zero cells can undergo cell death following enforced increases in intracellular ROS, we treated wild-type and {rho} zero cells with H2O2 and noted similar cell death profiles (data not shown).



View larger version (41K):
[in this window]
[in a new window]
  		FIG. 7.
ROS production sensitizes cells to apoptosis following UPR induction. A, wild-type ({diamondsuit}), wild-type {rho} zero ({blacksquare}), Nrf2-/- ({blacktriangleup}), and Nrf2-/- {rho} zero (•) cells proliferating on glass coverslips were treated with 5 µg/ml tunicamycin (Tun) for the indicated intervals. Cells were stained with propidium iodide and fluorescein isothiocyanate-conjugated Annexin V and examined by fluorescence and light microscopy. Annexin V-positive cells are shown as the percentage of total cells examined. Error bars indicate S.D. from a minimum of three independent trials. B, the conditions were the same as described for A, except that wild-type ({diamondsuit}), wild-type {rho} zero ({blacksquare}), and wild-type cells with 1 mM cysteine added to the medium ({blacktriangleup}) were treated with tunicamycin for the indicated intervals. C, wild-type (WT) and {rho} zero cells were plated at low density and were left untreated or were challenged with tunicamycin (5 µg/ml) for 8 h. Following tunicamycin treatment, cells were refed complete medium. After 7 days, colonies were visualized with Giemsa stain. D, wild-type (black bars), Nrf2-/- (dotted white bars), wild-type {rho} zero (hatched bars), and Nrf2-/- {rho} zero (dotted black bars) cells were treated with tunicamycin (5 µg/ml) for the indicated intervals, and total intracellular glutathione levels were measured. Glutathione levels are expressed relative to those in wild-type cells and were arbitrarily set at 1. Error bars represent S.D. from at least three independent experiments. *, p < 0.05; **, p < 0.01 (paired Student's t test). t tests compare the difference in parental and {rho} zero cells at matched intervals.

 
Wild-type cells do succumb to apoptotic cell death in response to ER stress (1, 38), and recent data suggest that increased ROS production following ER stress contributes to the onset of apoptosis (23). However, it is unclear whether ROS production is the key trigger for ER stress-induced cell death (Fig. 4) (23, 39) or whether increased ROS production in cells deficient in PERK-dependent activation of antioxidant signaling sensitizes these cells to pro-apoptotic signals initiated by the UPR (this study and Refs. 10, 15, 23, 39, and 40). We reasoned that if increased ROS production is the critical signal for cell death following UPR induction, wild-type {rho} zero cells should not be susceptible to UPR-triggered cell death. To test this idea, wild-type and {rho} zero cells were challenged with tunicamycin (5 µg/ml), and apoptosis was assessed by Annexin V staining. As expected, ER stress-dependent increases in apoptosis were observed in the parental cells (Fig. 7B). Tunicamycin treatment also promoted apoptosis in {rho} zero cells and in cells grown in cysteine-supplemented medium, but to a lesser degree than in the parental cells (Fig. 7B). Additionally, in a colony outgrowth assay, no appreciable difference was observed between wild-type and {rho} zero cells challenged with tunicamycin (Fig. 7C). These data suggest that ROS production sensitizes cells to ER stress, but is not essential for UPR-dependent cell death.

We next considered the possibility that elimination of mitochondrial ROS might also eliminate glutathione depletion in Nrf2-/- fibroblasts. The disruption of the mitochondrial respiratory chain increased glutathione levels in wild-type and Nrf2-/- cells both basally and in response to ER stress (Fig. 7D). This finding supports the idea that {rho} zero cells, by virtue of decreasing ROS production, maintain elevated thiol levels within the cell. Collectively, these data suggest that the decreased survival of Nrf2-/- cells in response to ER stress is a direct consequence of failure to combat elevated mitochondrially produced ROS through the production of glutathione.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
REFERENCES
 
Nrf2 and UPR Signaling—Mammalian cells possess a sophisticated signaling network that both senses and reacts to stress emanating from the ER. The production of molecular chaperones is rapidly induced (6); and in a parallel pathway, global translation is decreased to reduce the protein folding requirements within the stressed ER (7). We previously identified a new component of this signaling pathway: PERK-mediated activation of the pro-survival transcription factor Nrf2 (10). Of note, Nrf2 activity increases the survival of cells challenged with tunicamycin (10). In this study, we set out to determine, mechanistically, how PERK-dependent activation of Nrf2 contributes to cell survival following induction of the UPR via a physiological stress. Induction of Nrf2-dependent transcription contributes to increased transcription of Phase II detoxifying enzymes, providing cells with a buffer to increased oxidative stress (19, 27, 31, 41, 42). Nrf2 promotes expression of enzymes that mediate the production of glutathione, which, in addition to scavenging free radicals, also maintains proper intracellular redox balance through the reduction of stress-induced disulfide bonds and the maintenance of protein integrity (35).

ER stress rapidly induces the transcription of GCLC and NQO1, precursors in the glutathione biosynthetic pathway (Fig. 2) (10). Induction of both enzymes is dependent upon PERK and Nrf2. The capacity of Nrf2 to rescue glutathione levels in both Nrf2- and PERK-deficient cells is consistent with Nrf2 functioning downstream of PERK in the regulation of glutathione biosynthesis following induction of ER stress. Our results strongly suggest that Nrf2 activation at the onset of ER stress provides a protective mechanism by which cells increase glutathione production in an effort to counteract the deleterious effects that ER stress inflicts on cellular redox balance. The transient nature of Nrf2 activation during the UPR suggests that, although this response initially functions to protect the cell from ER stress-induced damage, a prolonged stress response is ultimately so harmful that cells must commit to an apoptotic fate.

Nrf2 and CHOP Accumulation—We have identified Nrf2 as an ER stress-activated transcription factor (this study and Ref. 10). Although many Nrf2 target genes have been identified through candidate approaches and microarray technology, these studies have focused on Nrf2 signaling following oxidative stress, not ER stress. CHOP expression is up-regulated in response to a variety of cellular stresses, including during the UPR (36, 43, 44). We have demonstrated here that CHOP expression correlates negatively with the presence of Nrf2; in Nrf2-/- cells, CHOP expression is constitutively higher than in wild-type cells, and Nrf2 overexpression attenuates CHOP accumulation during the UPR. This effect is more pronounced in conjunction with overexpression of other transcription factors that cooperate with Nrf2. Although the mechanism by which Nrf2 affects CHOP induction is not yet known, it is tempting to speculate that Nrf2 acts as a direct transcriptional repressor at the CHOP promoter. Other possibilities include the following: 1) Nrf2 precludes the binding of other transcription factors to the CHOP promoter, 2) and Nrf2 target genes directly affect CHOP expression.

Enforced CHOP expression can also result in the depletion of cellular glutathione stores (39). We have noted that Nrf2-/- fibroblasts express elevated basal CHOP levels as well as constitutively low glutathione levels, which correspond with low expression of crucial intermediates in the glutathione signaling pathway (21, 22, 31). Although Nrf2 loss contributes to decreased expression of rate-limiting enzymes in the glutathione biosynthetic pathway and thereby low glutathione levels, how increased CHOP levels contribute to glutathione depletion remains to be established. The interplay between CHOP and Nrf2 signaling is an area that requires further investigation.

Physiological Role for Nrf2 Signaling—Many of the previous studies that have focused on the elucidation of the molecular basis of Nrf2 activation have relied mainly on overexpression studies as well as pharmacological agents to activate Nrf2. Therefore, one concern is that these approaches might elicit pleiotropic effects that alter the true physiological role of Nrf2 signaling. We have demonstrated here that Nrf2 is activated under physiological conditions that promote ER stress. Glucose restriction elicits a rapid induction of Nrf2-dependent gene transcription, including that of GCLC, a requisite precursor for glutathione biosynthesis (30). That this response is PERK-dependent is consistent with the model that PERK signaling plays an important role in promoting survival during the UPR (13, 23).

Tumor cells may be particularly susceptible to nutrient limitations such as oxygen and glucose (45, 46). Tumor cells are highly glycolytic and thus require a high concentration of glucose as their carbon source (47, 48). The ability of tumors to proliferate under conditions of limiting nutrient availability suggests that these cells have adapted and initiated a mechanism whereby they circumvent the UPR signaling pathway. Just how this is accomplished remains to be determined, but alterations in the PERK-dependent arm of the UPR, including Nrf2 overexpression, are attractive candidates. Indeed, enforced Nrf2 expression up-regulates the transcription of many pro-survival genes (27, 31, 41, 49–51).

ROS Production Sensitizes Cells to Stress-induced Cell Death—The underlying cause of UPR-induced cell death has been under considerable scrutiny. Activation of caspases is ultimately responsible for cell death (8). However, the more proximal events leading up to caspase activation remain unclear. Increased oxidative burden is noted in cells deficient in UPR signaling, as evidenced by high ROS production and low intracellular glutathione levels (this study and Refs. 23 and 39). In Nrf2-/- (but not PERK-/-) cells, the perturbation in the redox environment is evident even in the absence of cellular stress, and death ensues only after increased stress is incurred. Likewise, pharmacological glutathione depletion greatly enhances UPR-dependent cell death (39), demonstrating the importance of cellular redox balance for survival following ER stress. Because wild-type cells also succumb to ER stress-induced apoptosis, it is important to assess whether the accumulation of ROS is sufficient for cell death signaling or whether other events are required. To address this question, we examined ER stress signaling in Nrf2-/- cells. Although these cells accumulate ROS at levels beyond those of their wild-type counterparts, marked apoptosis occurs only following the addition of a stress agent (this study and Ref. 10). To clearly assess the role of ROS production in cell death, we utilized a genetic system in which we eliminated the mitochondrial genome and therefore also eliminated mitochondrial ROS production ({rho} zero cells). The elimination of ROS production effectively reduced the sensitivity of Nrf2-/- cells to ER stress-induced apoptosis. It is important to note that this merely delayed, but did not eliminate, apoptosis. A similar result was obtained when examining the contribution of oxidative stress to ER stress-induced apoptosis in wild-type cells. Therefore, a model in which ROS accumulation contributes to, but is not the sole effector of, cell death must be considered. In this model, we envision that ROS production poises cells for ensuing cell death, thereby functioning as a potential signal to initiate the apoptotic pathway that eliminates severely damaged cells. What signal ultimately triggers cell death during ER stress remains unresolved, but good candidates would be UPR target genes induced independently of the PERK signaling pathway. Such targets to consider include PUMA, a BH3 domain-only protein that was recently shown to be induced during ER stress conditions (52), and other BH3 domain-only family proteins that are known to reside in both the mitochondria and ER (53–55).


   FOOTNOTES
 
* This work was supported by the Abramson Family Cancer Research Institute. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Back

{ddagger} To whom correspondence should be addressed: Abramson Family Cancer Research Inst., 454 BRB II/III, 421 Curie Blvd., Philadelphia, PA 19104. Tel.: 215-746-6389; Fax: 215-746-5511; E-mail: adiehl{at}mail.med.upenn.edu.

1 The abbreviations used are: ER, endoplasmic reticulum; UPR, unfolded protein response; ATF, activating transcription factor; eIF2{alpha}, eukaryotic translation initiation factor-2{alpha}; Nrf2, NF-E2-related factor-2; CHOP, CCAAT/enhancer-binding protein homologous protein; ROS, reactive oxygen species; HA, hemagglutinin; GCLC, glutamate-cysteine ligase catalytic subunit ({gamma}-glutamylcysteine synthetase); NQO1, NADPH-quinone oxidoreductase-1; DCF, 2',7'-dichlorofluorescin diacetate; DNP, 2,4-dinitrophenylhydrazone; DNPH, 2,4-dinitrophenylhydrazine; BH3, Bcl2-homology domain 3. Back


   ACKNOWLEDGMENTS
 
We thank Drs. Y. W. Kan and J. Y. Chan for providing the Nrf2-/- murine embryo fibroblasts and the Nrf2 cDNA, Drs. H. P. Harding and D. Ron for the PERK-/- murine embryo fibroblasts, Dr. M. Hannink for the mutant Nrf2 expression vector, K. Mansfield for assistance with the DCF assays and generation of {rho} zero cells, and R. Woolery for technical assistance.



   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Kaufman, R. J. (1999) Genes Dev. 13, 1211-1233[Free Full Text]
  2. Kaufman, R. J., Scheuner, D., Schroder, M., Shen, X., Lee, K., Liu, C. Y., and Arnold, S. M. (2002) Nat. Rev. Mol. Cell. Biol. 3, 411-421[CrossRef][Medline] [Order article via Infotrieve]
  3. Koumenis, C., Naczki, C., Koritzinsky, M., Rastani, S., Diehl, A., Sonenberg, N., Koromilas, A., and Wouters, B. G. (2002) Mol. Cell. Biol. 22, 7405-7416[Abstract/Free Full Text]
  4. Haze, K., Yoshida, H., Yanagi, H., Yura, T., and Mori, K. (1999) Mol. Biol. Cell 10, 3787-3799[Abstract/Free Full Text]
  5. Li, M., Baumeister, P., Roy, B., Phan, T., Foti, D., Luo, S., and Lee, A. S. (2000) Mol. Cell. Biol. 20, 5096-5106[Abstract/Free Full Text]
  6. Tirasophon, W., Welihinda, A. A., and Kaufman, R. J. (1998) Genes Dev. 12, 1812-1824[Abstract/Free Full Text]
  7. Harding, H. P., Zhang, Y., and Ron, D. (1999) Nature 397, 271-274[CrossRef][Medline] [Order article via Infotrieve]
  8. Nakagawa, T., Zhu, H., Morishima, N., Li, E., Xu, J., Yankner, B. A., and Yuan, J. (2000) Nature 403, 98-103[CrossRef][Medline] [Order article via Infotrieve]
  9. Shi, Y., Vattem, K. M., Sood, R., An, J., Liang, J., Stramm, L., and Wek, R. C. (1998) Mol. Cell. Biol. 18, 7499-7509[Abstract/Free Full Text]
  10. Cullinan, S. B., Zhang, D., Hannink, M., Arvisais, E., Kaufman, R. J., and Diehl, J. A. (2003) Mol. Cell. Biol. 23, 7198-7209[Abstract/Free Full Text]
  11. Berlanga, J. J., Santoyo, J., and De Haro, C. (1999) Eur. J. Biochem. 265, 754-762[Medline] [Order article via Infotrieve]
  12. Wek, R. C. (1994) Trends Biochem. Sci. 19, 491-496[CrossRef][Medline] [Order article via Infotrieve]
  13. Harding, H. P., Zhang, Y., Bertolotti, A., Zeng, H., and Ron, D. (2000) Mol. Cell 5, 897-904[CrossRef][Medline] [Order article via Infotrieve]
  14. Harding, H. P., Novoa, I., Zhang, Y., Zeng, H., Wek, R., Schapira, M., and Ron, D. (2000) Mol. Cell 6, 1099-1108[CrossRef][Medline] [Order article via Infotrieve]
  15. Scheuner, D., Song, B., McEwen, E., Liu, C., Laybutt, R., Gillespie, P., Saunders, T., Bonner-Weir, S., and Kaufman, R. J. (2001) Mol. Cell 7, 1165-1176[CrossRef][Medline] [Order article via Infotrieve]
  16. Harding, H. P., Zeng, H., Zhang, Y., Jungries, R., Chung, P., Plesken, H., Sabatini, D. D., and Ron, D. (2001) Mol. Cell 7, 1153-1163[CrossRef][Medline] [Order article via Infotrieve]
  17. Zhang, P., McGrath, B., Li, S., Frank, A., Zambito, F., Reinert, J., Gannon, M., Ma, K., McNaughton, K., and Cavener, D. R. (2002) Mol. Cell. Biol. 22, 3864-3874[Abstract/Free Full Text]
  18. Itoh, K., Wakabayashi, N., Katoh, Y., Ishii, T., Igarashi, K., Engel, J. D., and Yamamoto, M. (1999) Genes Dev. 13, 76-86[Abstract/Free Full Text]
  19. Hayes, J. D., and McMahon, M. (2001) Cancer Lett. 174, 103-113[CrossRef][Medline] [Order article via Infotrieve]
  20. Chan, K., and Kan, Y. W. (1999) Proc. Natl. Acad. Sci. U. S. A. 96, 12731-12736[Abstract/Free Full Text]
  21. Chan, J. Y., and Kwong, M. (2000) Biochim. Biophys. Acta 1517, 19-26[Medline] [Order article via Infotrieve]
  22. Leung, L., Kwong, M., Hou, S., Lee, C., and Chan, J. Y. (2003) J. Biol. Chem. 278, 48021-48029[Abstract/Free Full Text]
  23. Harding, H. P., Zhang, Y., Zeng, H., Novoa, I., Lu, P. D., Calfon, M., Sadri, N., Yun, C., Popko, B., Paules, R., Stojdl, D. F., Bell, J. C., Hettmann, T., Leiden, J. M., and Ron, D. (2003) Mol. Cell 11, 619-633[CrossRef][Medline] [Order article via Infotrieve]
  24. Todaro, G. J., and Green, H. (1963) J. Cell. Biol. 17, 299-313[Medline] [Order article via Infotrieve]
  25. King, M. P., and Attardi, G. (1996) Methods Enzymol. 264, 304-313[Medline] [Order article via Infotrieve]
  26. Baker, M. A., Cerniglia, G. J., and Zaman, A. (1990) Anal. Biochem. 190, 360-365[CrossRef][Medline] [Order article via Infotrieve]
  27. Wild, A. C., Moinova, H. R., and Mulcahy, R. T. (1999) J. Biol. Chem. 274, 33627-33636[Abstract/Free Full Text]
  28. He, C. H., Gong, P., Hu, B., Stewart, D., Choi, M. E., Choi, A. M., and Alam, J. (2001) J. Biol. Chem. 276, 20858-20865[Abstract/Free Full Text]
  29. Stewart, D., Killeen, E., Naquin, R., Alam, S., and Alam, J. (2003) J. Biol. Chem. 278, 2396-2402[Abstract/Free Full Text]
  30. Nguyen, T., Sherratt, P. J., Huang, H. C., Yang, C. S., and Pickett, C. B. (2003) J. Biol. Chem. 278, 4536-4541[Abstract/Free Full Text]
  31. Itoh, K., Chiba, T., Takahashi, S., Ishii, T., Igarashi, K., Katoh, Y., Oyake, T., Hayashi, N., Satoh, K., Hatayama, I., Yamamoto, M., and Nabeshima, Y. (1997) Biochem. Biophys. Res. Commun. 236, 313-322[CrossRef][Medline] [Order article via Infotrieve]
  32. Hayes, J. D., Chanas, S. A., Henderson, C. J., McMahon, M., Sun, C., Moffat, G. J., Wolf, C. R., and Yamamoto, M. (2000) Biochem. Soc. Trans. 28, 33-41[Medline] [Order article via Infotrieve]
  33. Stadtman, E. R., and Levine, R. L. (2000) Ann. N. Y. Acad. Sci. 899, 191-208[Abstract/Free Full Text]
  34. Shi, Z. Z., Osei-Frimpong, J., Kala, G., Kala, S. V., Barrios, R. J., Habib, G. M., Lukin, D. J., Danney, C. M., Matzuk, M. M., and Lieberman, M. W. (2000) Proc. Natl. Acad. Sci. U. S. A. 97, 5101-5106[Abstract/Free Full Text]
  35. Richter, C., Gogvadze, V., Laffranchi, R., Schlapbach, R., Schweizer, M., Suter, M., Walter, P., and Yaffee, M. (1995) Biochim. Biophys. Acta 1271, 67-74[Medline] [Order article via Infotrieve]
  36. Wang, X. Z., Lawson, B., Brewer, J. W., Zinszner, H., Sanjay, A., Mi, L. J., Boorstein, R., Kreibich, G., Hendershot, L. M., and Ron, D. (1996) Mol. Cell. Biol. 16, 4273-4280[Abstract]
  37. Kwong, M., Kan, Y. W., an