P58IPK, a novel endoplasmic reticulum stress-inducible protein and potential negative regulator of eIF2alpha signaling.

The unfolded protein response, which is activated in response to the loss of endoplasmic reticulum (ER) Ca(2+) homeostasis and/or the accumulation of misfolded, unassembled, or aggregated proteins in the ER lumen, involves both transcriptional and translational regulation. In the current studies we sought to identify novel ER stress-induced genes by conducting microarray analysis on tunicamycin-treated cells. We identified P58(IPK), an inhibitor of the interferon-induced double-stranded RNA-activated protein kinase, as induced during ER stress. Additional studies suggested that p58(IPK) induction was mediated via ATF6 and that P58(IPK) played a role in down-regulating the activity of the pancreatic eIF2 kinase/eukaryotic initiation factor 2alpha (eIF2alpha)-like ER kinase/activation transcription factor (ATF) 4 pathway. Modulation of P58(IPK) levels altered the phosphorylation status of eIF2alpha, and thereby affected expression of its downstream targets, ATF4 and Gadd153. Overexpression of P58(IPK) inhibited eIF2alpha phosphorylation and reduced ATF4 and Gadd153 protein accumulation, whereas silencing of P58(IPK) expression enhanced pancreatic eIF2alpha-like ER kinase and eIF2alpha phosphorylation and increased ATF4 and Gadd153 accumulation. These findings implicate P58(IPK) as an important component of a negative feedback loop used by the cell to inhibit eIF2alpha signaling, and thus attenuate the unfolded protein response.

The endoplasmic reticulum (ER) 1 is an important organelle in which newly synthesized secretory and membrane-associated proteins are correctly folded and assembled. Perturbations in the ER environment result in a condition known as ER stress, which can threaten cell survival. ER stress can be induced in cells by a variety of treatments including agents known to affect calcium homeostasis, inhibitors of glycosyla-tion, and overloading of the cell with mutant proteins that cannot be properly folded. Such stress triggers the activation of a complex response termed the unfolded protein response (UPR), which in mammalian cells is characterized by coordinate transcriptional up-regulation of a number of proteins including molecular chaperones and folding enzymes, global inhibition of protein synthesis, and activation of apoptotic pathways (1). The first two components serve to reduce the load of client proteins and alleviate the stress, whereas the third functions to eliminate severely damaged cells.
At least three separate mechanisms contribute to the transcriptional response to ER stress in mammalian cells. The first involves IRE1␣ and IRE1␤, transmembrane protein kinases with endoribonuclease function, that sense the presence of unfolded proteins in the ER, leading to their activation (2)(3)(4). In turn, activation of IRE1 results in the splicing of mRNA encoding the transcription factor XBP-1, increasing its efficiency of translation, thereby enhancing its expression (5)(6)(7). A second mechanism contributing to the transcriptional response to ER stresses is mediated through the ATF6 transcription factor, a member of the ATF/cAMP-response element-binding protein family of basic leucine zipper proteins. ATF6 is a type II ER transmembrane protein with its NH 2 -terminal DNAbinding domain facing the cytosol and its COOH terminus in the ER lumen (8). In response to ER stress, the cytosolic domain of ATF6 is cleaved off and translocates to the nucleus to activate transcription of ER stress target genes (7, 8, 10 -15). Overexpression of the active nuclear form of ATF6 is sufficient for transcriptional induction of GRP78 and transcription factors Gadd153 and XBP-1 (7,16). The third mechanism contributing to transcriptional regulation by ER stress involves the transcription factor ATF4, another ATF/cAMP-response element-binding protein family member. Its expression is regulated via the PERK/eIF2␣ pathway as discussed below. Like ATF6, ATF4 plays an important role in the activation of CHOP/ Gadd153 (17,18).
Repression of protein synthesis in response to ER stress is mediated through the increased phosphorylation of eukaryotic initiation factor 2 (eIF2␣), a modification that interferes with the formation of an active 43 S translation-initiation complex (19). Phosphorylation of eIF2␣ during ER stress is carried out by the pancreatic eIF2␣ kinase (PEK or PERK), which is activated as part of the unfolded protein response (20 -22). In addition to PERK, three other eukaryotic protein kinases are known to phosphorylate eIF2␣, each of which responds to distinct stress signals. These include the heme-regulated inhibitory kinase, which phosphorylates eIF2␣ in response to heme depletion (23,24), the general control non-derepression-2-kinase, activated in response to amino acid starvation (25,26), and the interferon-induced protein kinase (PKR), which is activated by dsRNA produced during viral infection (27,28). * 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.
Genome-wide expression analysis using DNA microarrays has revealed that activation of the UPR in yeast results in the up-regulation of more than 350 genes. These include genes involved in various aspects of the secretory pathway, such as protein folding, ER to Golgi vesicular transport, and ER-associated protein degradation (22). Although many of the molecular details of the UPR have been conserved in yeast and mammalian systems, the scope of UPR outputs in the mammalian cell is more complex and diverse. Accordingly, it is likely to involve a greater number of proteins and gene expression changes than seen in yeast.
The present study utilized DNA microarray analysis to search for novel genes induced by ER stress in mouse embryo fibroblasts (MEFs). We report here the identification of P58 IPK , an inhibitor of the interferon-induced double-stranded RNAactivated protein kinase (PKR), as a gene whose expression is up-regulated in response to ER stress. Additional studies provide evidence that ATF6 contributes to the induction of P58 IPK and P58 IPK plays a role in regulating the activity of the PERK/ eIF2␣/ATF4 pathway.

MATERIALS AND METHODS
Cell Culture, Treatments, Plasmid Construction, and Transfection-MEFs, human embryo kidney fibroblasts (HEK-293), and human cervical carcinoma HeLa cells were cultured in Dulbecco's modified essential medium (Invitrogen). All media were supplemented with 10% fetal bovine serum (Hyclone, Logan, UT), 100 units of penicillin/ml, and 100 g of streptomycin (Invitrogen) per ml, and were maintained in a humidified atmosphere containing 5% CO 2 . Recombinant DNA techniques were performed by standard procedures. Plasmid pCGN-atf6 was kindly provided by Dr. R. Prywes (Columbia University) and used to construct pCGN-p50atf6 by inserting a PCR fragment spanning the NH 2 -terminal 1-373 amino acids of atf6, followed by a stop codon into the XbaI and BamHI sites of the pCGN vector. HEK-293 and HeLa cells were grown to 60% confluency in 100-mm plates and transfected using Polyfect reagent (Qiagen, Valencia, CA). Six g of pCGN-p50atf6, pCDNA3-atf4 (kindly provided by Dr. J. Leiden, University of Chicago), RT-PCR and Northern Blot Analysis-To analyze RNA expression by reverse transcription-PCR (RT-PCR), total RNA from each sample was treated with DNase I and used for RT-PCR with SuperScript One-Step RT-PCR with the Platinum Taq system (Invitrogen). The primers for gene-specific RT-PCR analysis were as follows: for P58 IPK , GAGGTTT-GTGTTGGGATGCAG (5Ј) and GCTCTTCAGCTGACTCAATCAG (3Ј); for ATF4, AGGAGTTCGCCTTGGATGCCCTG (5Ј) and AGTGATATC-CACTTCACTGCCCAG (3Ј); for ATF6, ATCAGTTTACAACCTGCAC-CCAC (5Ј) and CTGTCTCCTTAGCACAGCAATATC (3Ј); for Gadd153, CTGAGTCATTGCCTTTCTCTTCG (5Ј) and CTCTGACTGGAATCTG-GAGAGTG (3Ј); GAPDH, ACATCAAGAAGGTGGTGAAGCAGG (5Ј) and CTCTTGCTCTCAGATCCTTGCTGG (3Ј). Equal aliquots of the PCR products were electrophoresed through 2% agarose gels. For Northern blot analysis, 4-g aliquots of total RNA (harvested using Nucleo Spin RNAII kit (Clontech, Palo Alto, CA)) were run on agaroseformaldehyde gels and transferred onto GeneScreen Plus membranes (PerkinElmer Life Sciences). cDNAs corresponding to atf4, gadd153, and grp78 (a generous gift from Amy S. Lee), were labeled by the random primer method and used to detect corresponding mRNAs on Northern blots. An end-labeled 24-bp oligonucleotide complementary to 18 S rRNA (ACGGTATCTGATCGTCTTCGAACC) was used as a probe to verify RNA integrity and loading differences.
cDNA Array Analysis-Total RNA was extracted from all samples using a NucleoSpin RNAII kit (Clontech, Palo Alto, CA). Atlas Human 1.2 K filters (www.Clontech) each containing 1174 genes were used. Total RNA (5 g) was reverse transcribed and labeled with [␣-32 P]dATP using the Clontech cDNA array labeling kit. Hybridizations and washes were performed as recommended by the manufacturer. The cDNA array membranes were visualized for analysis by using a PhosphorImager (Amersham Biosciences), and were quantitated as described (31).
DAPI Staining-DAPI staining was performed as described previously (24). In brief, prior to staining, the cells were fixed with 4% paraformaldehyde for 30 min at room temperature, and then washed with phosphate-buffered saline. DAPI was added to the fixed cells for 30 min, after which they were examined by fluorescence microscopy. Apoptotic cells were identified by condensation and fragmentation of nuclei. Percentage of apoptotic cells was calculated as the ratio of apoptotic cells to total cells counted ϫ100. A minimum of 400 cells were counted for each treatment.
Construction of Small Interference RNA (siRNA) Duplexes and Transfection-Twenty-one nucleotide double-stranded RNAs were transcribed in vitro using the Silencer TM siRNA construction kit according to the manufacturer instructions (Ambion Inc.). The targeting sequence of human P58 IPK (accession number U28424), corresponding to nucleotide positions 137-157 (coding region), was AATTACTTG-CAGCTGGACAGC. An siRNA targeting the luciferase mRNA (accession number X65324) served as a control. Cells were seeded in 6-well plates on the day before transfection at a concentration of 10 5 cells per well. Cells were transfected with OligofectAMINE reagent according to the manufacturer's instructions (Invitrogen). Briefly, Opti-MEM (165 l) was mixed with 20 l of 1 M siRNA duplex. In a separate tube, 12 l of Opti-MEM I was incubated with 3 l of OligofectAMINE for 5 min at room temperature. The two mixtures were combined, gently mixed, and incubated for another 20 min at room temperature. The entire mixture was added to the cells in 0.8 ml of 10% fetal bovine serumcontaining Dulbecco's modified Eagle's medium without antibiotics. Cells were assayed at different time intervals after transfection.
[ 35 S]Methionine Metabolic Labeling-Cells were seeded into 6-well plates at a density of 10 5 cells per well. Twenty-four h following P58 IPK siRNA transfection, cells were placed for 30 min in methionine-free minimal essential medium (BIOSOURCE, Camarillo, CA), and labeled by addition of [ 35 S]methionine (20 Ci/ml; 1,000 Ci/mmol; Amersham Biosciences) to the culture medium for 2 h. Cells were washed twice with ice-cold phosphate-buffered saline and collected in lysis buffer (20 mM Hepes, pH 7.4, 2 mM EGTA, 50 mM ␤-glycerophosphate, 1 mM Na 3 VO 4 , 5 mM NaF, 1% Triton X-100, 10% glycerol, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 5 g/ml leupeptin, 5 g/ml aprotinin). The protein concentration was measured using the Bio-Rad protein DC assay kit. The [ 35 S]methionine incorporation was measured by cold trichloroacetic acid precipitation and analyzed by SDS-PAGE. For trichloroacetic acid precipitation, equal amounts of protein were added to 0.5 ml of 0.1 mg/ml bovine serum albumin containing 0.02% sodium azide and placed on ice. Ice-cold 20% trichloroacetic acid (0.5 ml) was added and samples were vortexed vigorously and incubated for 30 min on ice. Cell suspensions were filtered through glass microfiber discs (Whatman). Discs were washed three times with ice-cold 10% trichloroacetic acid, and twice with 100% ethanol, after which they were air-dried, and the radioactivity was measured by scintillation counting.

Identification of P58 IPK as an ER Stress-induced Gene-We
sought to investigate the UPR stress response in mammalian cells by assessing changes in gene expression profiles after ER stress using the Atlas cDNA Gene Array (Clontech). This array contains cDNAs for 1174 genes involved in apoptosis, cell cycle control, stress responses, transcription, and signaling. Total RNA was isolated from MEFs that were either left untreated or treated for 6 h with 2 g/ml tunicamycin, an agent that causes ER stress by inhibiting protein N-glycosylation. Analysis of the resulting signals on cDNA arrays (carried out as described under "Materials and Methods") revealed 19 candidate genes whose expression was increased in response to tunicamycin treatment. Among these were a number of genes previously shown to play a role in the UPR response, including grp78, gadd153, and erp72. One novel gene whose expression was significantly elevated in MEFs following ER stress was p58 IPK , an established inhibitor of the eIF2␣ kinase PKR, a kinase related to the ER-specific kinase, PERK (Fig. 1A). Hence, we further investigated the regulation of p58 IPK by ER stress and examined its potential role during the UPR response.
To validate the findings obtained by microarray analysis, p58 IPK expression levels in MEFs and HEK-293 cells were examined using RT-PCR after treatment with either tunicamycin or thapsigargin (another ER stress agent that inhibits ER Ca 2ϩ -ATPase, thereby causing ER stress through a different mechanism). As shown, p58 IPK mRNA levels increased in both cell lines in response to each ER stress agent in a time-dependent manner (Fig. 1B). Further characterization of the response was carried out through kinetic analysis of P58 IPK protein expression by Western blot analysis. As depicted in Fig. 1C, P58 IPK protein levels similarly increased following treatment with tunicamycin in both cell lines (Fig. 1C).
Because ATF4 and ATF6 are known to play a role in the transcriptional activation of ER stress-inducible genes, we examined their contribution to the induction of p58 IPK expression. To investigate the potential role of ATF6, we transfected HEK-293 cells with a plasmid that expresses p50 ATF6 (the soluble form of ATF6, capable of translocating into the nucleus). Cells were harvested 48 h after transfection and p58 IPK mRNA was examined using RT-PCR (Fig. 2). Overexpression of p50 ATF6 alone in HEK-293 cells resulted in up-regulation of the p58 IPK mRNA, even in the absence of ER stress. In contrast, overexpression of full-length ATF4 alone did not alter p58 IPK mRNA expression. To ensure that the ATF4 construct was functional in this assay, we also analyzed expression of gadd153, an established transcriptional target of both ATF4 and ATF6. gadd153 mRNA was induced similarly by ectopically overexpressed p50 ATF6 and ATF4, indicating that the ATF4 construct was functional. These observations indicate that p58 IPK is a novel target of the ATF6, but not ATF4 pathway.
Overexpression of P58 IPK Inhibits eIF2␣ Phosphorylation and ATF4 and Gadd153 Induction-To gain a better understanding of the biological role of P58 IPK in the UPR, we tried to generate stable cell lines overexpressing P58 IPK using both HEK-293 and HeLa cells. Despite our best efforts, no stable P58 IPK -overexpressing clones could be obtained (data not shown). As an alternative strategy, we transiently transfected a plasmid expressing P58 IPK into HeLa cells, where high transfection efficiencies could be achieved. ER stress is known to induce translational repression, which is mediated by phosphorylation of eIF2␣. Phosphorylation of eIF2␣ leads to downregulation of translation initiation through a well characterized mechanism involving inhibition of eIF2␤ activity (32). P58 IPK is a known inhibitor of the eIF2␣ kinase PKR, and it has been shown that overexpression of P58 IPK can inhibit dsRNAinduced phosphorylation of eIF2␣ by PKR. Because ER stress leads to elevated P58 IPK expression, we hypothesized that P58 IPK might affect the eIF2␣ phosphorylation during the UPR. To address this possibility HeLa cells were transiently transfected with a P58 IPK expression vector or empty vector (Fig. 3A). Forty-eight hours post-transfection, cells were treated with tunicamycin, and protein lysates were analyzed for both eIF2␣ phosphorylation and total eIF2␣ protein levels  2. Overexpression of ATF6 induces p58 IPK expression. A, HEK-293 cells were transfected with an empty vector or vectors expressing either atf4 or atf6. Cells were harvested 48 h after transfection and total RNA prepared for assessment of the expression levels of mRNAs encoding p58 IPK and gapdh (to control for RNA integrity and loading) by RT-PCR. B, cells were transfected as described in the legend of panel A and total RNA was prepared to assess the expression levels of gadd153, atf4, atf6, and gapdh by RT-PCR. by Western blotting. As shown in Fig. 3B, overexpression of P58 IPK significantly attenuated tunicamycin-induced eIF2␣ phosphorylation.
Given that elevated P58 IPK expression decreased the phosphorylation levels of eIF2␣, we sought to analyze the levels of ATF4, a downstream target of this pathway, and Gadd153, whose expression is in turn regulated (at least in part) by ATF4. P58 IPK overexpression profoundly inhibited ATF4 protein accumulation in tunicamycin-treated HeLa cells (Fig. 3C), consistent with previous reports showing that the production of ATF4 protein requires eIF2␣ phosphorylation (18). Gadd153 protein induction in the P58 IPK overexpressed cells was ϳ70% of that seen in control cells after 4 -6 h treatment with tunicamycin (Fig. 3C).
Previous reports have shown that P58 IPK can protect cells from dsRNA or tumor necrosis factor-␣-induced apoptosis. We were therefore interested in determining whether P58 IPK could protect cells from ER stress-induced apoptosis. HeLa cells transiently transfected with either a p58 IPK -containing expression vector or an empty vector were treated with tunicamycin for different time periods and then analyzed for apoptosis. A hallmark of apoptosis is cleavage of the nuclear 116-kDa PARP (poly(ADP-ribose) polymerase) protein to an 85-kDa inactive polypeptide. Inactivation of PARP through proteolytic cleavage facilitates chromosomal DNA fragmentation as part of the cellular apoptotic program (33). Our results show that the tunicamycin-induced PARP proteolysis was similar in vector and P58 IPK -transfected HeLa cells (Fig. 3D). DAPI staining revealed a similar pattern of condensed and fragmented nuclei for both the control and P58 IPK -transfected cells (Fig. 3E). Taken together, these results suggest that under the conditions utilized here, overexpression of P58 IPK does not alter the apoptotic response to ER stress.
Silencing of P58 IPK by siRNA Induces Apoptosis-Whereas elevated P58 IPK expression failed to alter the cellular outcome following ER stress, it remained possible that a reduction in P58 IPK could influence the response. We decided to employ the RNA interference technique to address this possibility. A small inhibitory double-stranded RNA homologous to a 21-nucleotide sequence unique to the human P58 IPK was used to reduce P58 IPK expression. As shown in Fig. 4A, transfection of HeLa or HEK-293 cells with P58 IPK siRNA resulted in a reduction of P58 IPK protein levels, although control transfections with siRNA specific for luciferase, carried out in parallel, showed no effect on P58 IPK expression. These observations indicate that the P58 IPK siRNA treatment specifically reduced the abundance of P58 IPK protein.
Approximately 28 h after transfection with the P58 IPK siRNA duplex, we observed that many cells showed reduced

FIG. 3. Overexpression of P58 IPK reduces eIF2␣ phosphorylation and expression of ATF4 and Gadd153.
A, 48 h after transfection with either an empty plasmid or a vector expressing p58 IPK , HeLa cells were treated with 2 g/ml tunicamycin for the indicated times and P58 IPK and GAPDH protein levels were assessed by immunoblot. B, cells were transfected as described in panel A, whereupon they were treated with 1 M thapsigargin (Tg) for 1 h and phospho-eIF2␣ and total eIF2␣ levels were assessed by immunoblotting. C, cells were treated as described in panel A, then treated with 2 g/ml tunicamycin (Tn) for the times indicated. Expression of ATF4, Gadd153, and GAPDH proteins was assessed by immunoblot analysis. The values between the panels represent the -fold induction of ATF4 and Gadd153 levels compared with control levels after normalization to GAPDH levels, as determined by densitometric analysis. D, control and P58 IPK -overexpressing HeLa cells were harvested at the indicated times, and PARP cleavage was assessed by Western blot analysis using a monoclonal antibody that recognizes cleaved PARP. E, HeLa cells that were treated with tunicamycin for the times indicated were fixed with 4% paraformaldehyde, stained with DAPI, and subjected to counting of apoptotic nuclei by fluorescence microscopy. Data represent the mean Ϯ S.E.

FIG. 4. Reduction of P58 IPK protein expression by small interference RNA (siRNA) causes apoptosis.
A, reduction of P58 IPK protein expression by RNA interference. An siRNA duplex specific for human P58 IPK as well as a duplex specific for the firefly gene luciferase (described under "Materials and Methods") were transfected into HEK-293 and HeLa cells. Whole cell extracts were prepared 20 h after transfection and were used to detect P58 IPK and GAPDH by immunoblotting. B, induction of cell death by P58 IPK siRNA. HEK-293 and HeLa cells were transfected with siRNA control duplexes and duplexes specific for human P58 IPK . At different time intervals after transfection, cells were counted and assessed for viability by trypan blue dye exclusion. C, induction of caspase 3 and PARP cleavage by P58 IPK siRNA. HeLa cells were transfected with siRNA for human P58 IPK and harvested at the indicated times. Whole cell lysates were subjected to immunoblot analysis to assess the levels of cleaved caspase-3 by using a polyclonal cleaved caspase-3 (Asp 175 ) antibody that detects endogenous levels of the large fragment (17/19 kDa) of activated caspase-3 resulting from cleavage adjacent to (Asp 175 ). PARP cleavage was assessed by using a monoclonal antibody that recognizes cleaved PARP. ability to adhere to the plate and floated in the medium. To determine whether silencing P58 IPK causes cell death, we examined the cell growth and viability of P58 IPK siRNA-transfected cells. Cells transfected with the control siRNA grew very well. By contrast, cells transfected with P58 IPK siRNA showed a marked reduction in viability (Fig. 4B). Caspase-3 is one of the key executioners of apoptosis, being responsible either partially or totally for the proteolytic cleavage of many key proteins such as the nuclear enzyme PARP. Activation of caspase-3 requires the proteolytic processing of its inactive zymogen into activated p17 and p12 subunits. Cleavage of caspase-3 and PARP can be detected by Western blot analysis and was apparent 20 h after transfection of P58 IPK siRNA in HeLa cells, whereas no caspase-3 or PARP cleavage were detected in the control cells (Fig. 4C). These results indicate that silencing P58 IPK decreases viability by causing apoptotic cell death.
siRNA Silencing of P58 IPK Induces eIF2␣ Phosphorylation and Inhibits Protein Synthesis-Because overexpression of P58 IPK can inhibit eIF2␣ phosphorylation, we postulated that silencing of P58 IPK would induce eIF2␣ phosphorylation and inhibit protein translation. To determine the effect of P58 IPK siRNA on eIF2␣ phosphorylation, HeLa and HEK-293 cells were transfected with the control siRNA or P58 IPK siRNA and 20 h post-transfection, protein lysates were prepared and analyzed by Western blotting (Fig. 5A). The results show that P58 IPK siRNA treatment markedly increased the level of eIF2␣ phosphorylation compared with control transfected cells.
To further establish that P58 IPK siRNA inhibits overall protein synthesis, [ 35 S]methionine incorporation was used to quantify the rate of protein synthesis after P58 IPK silencing. HeLa and HEK-293 cells were transfected with the P58 IPK siRNA duplex and metabolically labeled with [ 35 S]methionine at 20 h post-transfection. Analysis by SDS-PAGE and trichlo-roacetic acid precipitation demonstrated that general protein synthesis was reduced to 16.5 and 51% for HeLa and HEK-293, respectively, following P58 IPK silencing (Fig. 5B). Taken together these results strongly suggest that silencing P58 IPK activates the eIF2␣ stress-signaling pathway resulting in the inhibition of overall protein synthesis.
siRNA Silencing of P58 IPK Induces PERK Phosphorylation-Having determined that P58 IPK siRNA induces eIF2␣ phosphorylation we next wished to determine whether any of the known eIF2␣ kinases might be implicated in this process. P58 IPK was originally discovered as an inhibitor of PKR, an interferon-induced, double-stranded RNA-activated kinase that is activated during virus infection (34). Activation of PKR by double-stranded RNA results in PKR dimerization and autophosphorylation at positions Thr 446 and Thr 451 in the activation loop (35). To determine whether PKR is activated by P58 IPK siRNA, we examined its phosphorylation state. P58 IPK siRNA did not markedly induce PKR phosphorylation, as determined by Western blotting with an antiserum reactive to phospho-Thr 446/451 (Fig. 6A). It was interesting to note that total PKR protein levels were slightly increased by P58 IPK siRNA. Although moderate, such increases in PKR expression were seen reproducibly.
PERK, a kinase found in the lumen of the ER, also phosphorylates eIF2␣ in response to various stimuli that induce ER stress. A very recent paper from Yan et al. (36) demonstrates that P58 IPK is associated with the ER and represses PERK activity. Unfortunately, the low expression levels of endogenous PERK and the present unavailability of good anti-PERK antibodies did not allow a clear analysis of the effects of P58 IPK siRNA on PERK phosphorylation. Thus, to gain insight into this potential regulation, we transiently transfected HEK-293 cells using a plasmid that expressed wild-type mouse PERK. PERK activation can occur through autophosphorylation of its cytoplasmic kinase domain, and can be monitored by immunoblotting with an antibody that recognizes PERK only when it is phosphorylated on Thr 980 . HEK-293 cells transiently overexpressing PERK were transfected with the control siRNA or P58 IPK siRNA and, 20 h after transfection, lysates were prepared and subjected to Western blot analysis (Fig. 6B). As  Fig. 4), whole cell protein extracts were subjected to immunoblot analysis to assess the levels of total eIF2␣ and phosphorylated eIF2␣. B, whole cell extracts from HEK-293 and HeLa cells that were transfected with either the control siRNA duplex or the P58 IPK siRNA duplex were subjected to pulse-labeling of total proteins using [ 35 S]methionine (described under "Materials and Methods"). Proteins were resolved by 4 -12% gradient SDS-PAGE and stained with Coomassie Blue R-250 (right panel), or visualized for analysis using a PhosphorImager.
FIG. 6. Silencing of P58 IPK by siRNA leads to PERK phosphorylation. A, 20 h after transfection of HEK-293 cells with P58 IPK siRNA or control siRNA (described in the legend of Fig. 4) whole cell protein extracts were subjected to immunoblot analysis to assess the levels of total PKR and phosphorylated PKR. B, HEK-293 cells were transfected with PERK and 24 h later split into 6-well plates. Twentyfour h later, HEK-293 cells were transfected with P58 IPK siRNA or control siRNA (described in the legend of Fig. 4) and after 20 h, whole cell protein extracts were subjected to immunoblot analysis to assess the levels of total PERK and phospho-PERK. For the total PERK, the membrane was probed with the 9E10 antibody, which recognizes a c-myc epitope tag expressed at the COOH terminus of PERK.
shown, P58 IPK siRNA treatment markedly increased the level of PERK phosphorylation compared with control transfected cells.
P58 IPK Silencing by siRNA Induces ATF4 and Gadd153-Gadd153 is expressed at low or undetectable levels under normal growth conditions (37), but is highly induced following thapsigargin or tunicamycin treatments, and has been implicated in ER stress-induced apoptosis (38). Gadd153 transcription is induced by eIF2␣ phosphorylation through preferential translation of ATF4, as previously described (18). Because P58 IPK siRNA increased the level of eIF2␣ phosphorylation, we predicted that it would likely induce ATF4 and Gadd153 expression. To determine the effect of P58 IPK siRNA on ATF4 and Gadd153 expression, Northern and immunoblot analyses were performed. HEK-293 cells were transfected with control or P58 IPK siRNA and 20 h later were treated with tunicamycin for different time intervals, whereupon total RNA and protein were prepared for analysis (Fig. 7, A and B). ATF4 protein was low in untreated cells, but increased in response to tunicamycin treatment. Transfection to decrease P58 IPK levels in HEK-293 cells, did not significantly affect atf4 mRNA levels, but markedly increased ATF4 protein levels, even in untreated cells. These results are consistent with earlier reports showing that ATF4 expression is regulated via its preferential translation during ER stress in an eIF2␣-dependent manner (18). Gadd153, which was barely detectable in the control transfected cells, was likewise induced after P58 IPK siRNA transfection. This was evident at both the mRNA and protein levels.
Transcriptional induction of gadd153 has been shown to closely parallel that of grp78 under many ER stress-triggering conditions and indeed like gadd153, grp78 mRNA levels were increased at 4 h following tunicamycin treatment. However, unlike gadd153, expression of grp78 was not increased by P58 IPK siRNA. These findings suggest that silencing P58 IPK specifically affects eIF2␣ stress signaling pathways and that the P58 IPK protein plays a critical role in regulating ATF4 and Gadd153 protein levels. DISCUSSION The UPR, initiated in response to perturbations in the ER environment, is characterized by the activation of signaling pathways that transduce stress signals generated and sensed in the ER to other cellular compartments, thereby effecting changes in gene transcription and transiently suppressing translation. The biological objective of the UPR is to regain homeostasis in the ER by reducing demands on the organelle and increasing its capacity to carry out its protein folding and/or modifying functions. So far, three distinct pathways, IRE1/XBP-1, PERK/eIF2␣/ATF4, and ATF6, have been shown to contribute to the transcriptional response, but there is increasing evidence for both cross-talk and redundancy in the pathways. For example, transcriptional up-regulation of XBP-1 through the ATF6 pathway provides more substrate for IRE1, thus potentiating IRE1 pathway signaling. In addition, some genes, such as gadd153, appear to be targets of more than one pathway (i.e. both ATF4 and ATF6 contribute to its transcription) (15,17,40).
In this report, we have identified P58 IPK , an inhibitor of the interferon-induced protein kinase PKR, as a target for transcriptional up-regulation by the UPR. This finding is consistent with a previous report by Kaufman and colleagues (41) in which p58 IPK was listed as one of 67 genes induced upon tunicamycin treatment in MEFs using a different cDNA array. However, the physiologic relevance of p58 IPK induction was not addressed in that study, nor was its up-regulation verified by other methods in that report. In the current studies we have provided evidence suggesting that p58 IPK induction is mediated via ATF6 and that P58 IPK plays a role in down-regulating the activity of the PERK/eIF2␣/ATF4 pathway. These findings identify a novel mechanism contributing to the regulation of the UPR and define a new avenue for cross-talk between ER stress-activated pathways.
That ATF6 contributes to the induction of p58 IPK during the UPR is suggested by the finding that overexpression of its activated nuclear form (p50 ATF6) leads to elevated p58 IPK expression. A 19-bp ER stress responsive element consisting of the consensus sequence CCAAT(N 9 )CCACG has been identified in the promoter region of the p58 IPK gene (36). ER stressresponsive element sequences have been implicated in the transcriptional activation of downstream genes after treatment with agents that activate the mammalian UPR pathway (36), and ATF6 has been shown to interact with the ER stressresponsive element (16). Whereas definitive proof of its role in regulating p58 IPK must await further investigation, a role for ATF6 in regulating p58 IPK expression is consistent with studies of Okada and co-workers (42), in which microarray analysis revealed genes encoding molecular chaperones and folding enzymes as primary transcriptional targets of the ATF6 pathway. As a member of the DnaJ molecular chaperone family, P58 IPK would fit into this category (34).
P58 IPK was first recognized for its ability to inhibit the double-stranded RNA-activated protein kinase, PKR (34). PKR is part of the interferon-induced host defense against viral infection, and functions to repress translation initiation via phosphorylation of eIF2␣ (43,44). The proposed inhibitory mechanism of P58 IPK is as follows. Influenza virus activates the P58 IPK pathway by promoting the dissociation of P58 IPK from its own inhibitor, hsp40. The free P58 IPK represses PKR-mediated eIF2␣ phosphorylation through direct protein-protein interaction, and thereby relieves the PKR-imposed block on mRNA translation (45). Importantly, unlike the situation seen here for ER stress, P58 IPK expression is not altered by viral infection. Available evidence indicates that PKR is not involved in the ER stress response, as PKR knockout cells are unim- FIG. 7. Silencing of P58 IPK induces ATF4 and Gadd153. Western blot and Northern blot analyses of ATF4, Gadd153, and GRP78 expression in cells silenced with P58 IPK siRNA. HEK-293 cells were transfected with P58 IPK siRNA and 20 h later treated with 2 g/ml tunicamycin for different time periods. A, immunoblot analysis of protein lysates from control transfected populations and from P58 IPK siRNAtransfected cells to monitor ATF4, Gadd153, Grp78, and GAPDH protein levels. B, abundance of mRNAs encoding atf4, gadd153, grp78, and gapdh. Total RNA was prepared from transfected HEK-293 cells and analyzed by Northern blotting. paired in their ability to respond to ER stress by attenuating translation rates (20). Because PERK shares many properties with PKR, we investigated the possibility that P58 IPK could modify this pathway. A very recent report by Yan and colleagues (36) provides evidence that P58 IPK interacts with PERK and inhibits its activity. Our data showing that modulation of P58 IPK levels alters the phosphorylation status of PERK and eIF2␣ is in agreement with this report. We found that overexpression of P58 IPK inhibited eIF2␣ phosphorylation and reduced ATF4 protein accumulation, whereas silencing of P58 IPK expression enhanced PERK and eIF2␣ phosphorylation and increased ATF4 accumulation. The current model for regulation of ATF4 protein levels proposes that, under normal conditions, ATF4 mRNA is expressed, but is translated poorly. However, under ER stress conditions, ATF4 protein is rapidly synthesized in a manner that is dependent on eIF2␣ phosphorylation. This then leads to enhanced expression of its target genes. Previous studies implicate both ATF4 and ATF6 in the transcriptional activation of gadd153 (15,17,40). The more modest effect of P58 IPK overexpression on gadd153 expression (relative to its effect on ATF4) likely reflects the contribution of ATF6 to gadd153 induction.
P58 IPK has been demonstrated to possess an anti-apoptotic function, protecting cells against tumor necrosis factor-␣-and dsRNA-induced cell death (39). In our study systems, overexpression of P58 IPK did not alter cell death following ER stress. However, silencing of p58 IPK was associated with a significant increase in cell sensitivity to tunicamycin treatment. Higher levels of Gadd153/CHOP, resulting from increased activity of the PERK/eIF2␣/ATF4 pathway, might contribute to the enhanced sensitivity to undergo apoptosis. Indeed, overexpression of gadd153 has been shown to lead to cell cycle arrest and apoptosis through mechanisms that likely involve down-regulation of Bcl-2 (9).
In summary, our findings are consistent with a model in which P58 IPK functions to regulate PERK and eIF2␣ phosphorylation during normal physiologic conditions and under conditions of ER stress. Phosphorylation of PERK and eIF2␣ is required for translation attenuation, transcriptional induction, and cellular survival in response to ER stress. However, PERK and eIF2␣ phosphorylation and inhibition of protein translation is transient in stressed cells. We propose that enhanced expression of P58 IPK during the UPR serves as an important component of a negative feedback loop used by the cell to attenuate eIF2␣ signaling. This model is reminiscent of that proposed for Gadd34, another gene product induced during the UPR, which has been shown to play an important feedback role in regulating eIF2␣ phosphorylation, through promotion of its dephosphorylation (29).