|Home | About | Journals | Submit | Contact Us | Français|
Shiga-toxigenic Escherichia coli (STEC) produces subtilase cytotoxin (SubAB), which cleaves the molecular chaperone BiP in the endoplasmic reticulum (ER), leading to an ER stress response and then activation of apoptotic signaling pathways. Here, we show that an early event in SubAB-induced apoptosis in HeLa cells is mediated by RNA-dependent protein kinase (PKR)-like ER kinase (PERK), not activating transcription factor 6 (ATF6) or inositol-requiring enzyme 1(Ire1), two other ER stress sensors. PERK knockdown suppressed SubAB-induced eIF2α phosphorylation, activating transcription factor 4 (ATF4) expression, caspase activation, and cytotoxicity. Knockdown of eIF2α by small interfering RNA (siRNA) or inhibition of eIF2α dephosphorylation by Sal003 enhanced SubAB-induced caspase activation. Treatment with proteasome inhibitors (i.e., MG132 and lactacystin), but not a general caspase inhibitor (Z-VAD) or a lysosome inhibitor (chloroquine), suppressed SubAB-induced caspase activation and poly(ADP-ribose) polymerase (PARP) cleavage, suggesting that the ubiquitin-proteasome system controls events leading to caspase activation, i.e., Bax/Bak conformational changes, followed by cytochrome c release from mitochondria. Levels of ubiquitinated proteins in HeLa cells were significantly decreased by SubAB treatment. Further, in an early event, some antiapoptotic proteins, which normally turn over rapidly, have their synthesis inhibited, and show enhanced degradation via the proteasome, resulting in apoptosis. In PERK knockdown cells, SubAB-induced loss of ubiquitinated proteins was inhibited. Thus, SubAB-induced ER stress is caused by BiP cleavage, leading to PERK activation, not by accumulation of ubiquitinated proteins, which undergo PERK-dependent degradation via the ubiquitin-proteasome system.
Subtilase cytotoxin (SubAB) is an AB5 toxin identified as a product of Shiga-toxigenic Escherichia coli (STEC) strain O113:H21, which caused an outbreak of hemolytic uremic syndrome (HUS) (38). The SubAB 35-kDa A subunit (SubA) shares sequence homology with a subtilase-like serine protease (37, 38), and its pentamer of 13-kDa B subunits (SubB) recognizes receptors on target cells (5, 53). After binding to cell receptors, SubAB is translocated into the cell via an endocytosis pathway and cleaves the molecular chaperone BiP in the endoplasmic reticulum (ER) (37, 49). BiP belongs to the heat shock protein 70 (Hsp70) family and plays an important role in cell homeostasis (7). BiP cleavage by SubAB triggers an ER stress response, leading to activation of various signaling and metabolic pathways, resulting in transient inhibition of protein synthesis (30, 31), induction of stress-inducible C/EBP-homologous protein (CHOP) as a result of transcription factor-4 (ATF4) activation (49), and downregulation of cyclin D1 with cell cycle arrest in G1 phase (30). In addition, SubAB-induced ER stress signals lead to Bax/Bak conformational changes, cytochrome c release, and caspase activation, resulting in apoptosis through a caspase-dependent pathway (28, 52).
SubAB is lethal for mice; necropsy shows extensive microvascular thrombosis (48), severe inflammatory response and hemorrhage of the small intestine (8), and reduction in the numbers of splenic lymphocytes (47). These findings are similar to the histopathologic, biochemical, and hematologic changes seen in human cases of hemorrhagic colitis and hemolytic-uremic syndrome (32). Catalytically inactivated SubA(S272A)B did not exhibit cytotoxicity, suggesting that BiP cleavage is essential for the actions of SubAB (31, 37).
ER stress has been associated with a variety of human diseases and conditions, including hypoxia, neurodegeneration, stroke, bipolar disorder, cancer, diabetes, muscle degeneration, and viral infection (18). ER stress induces a cellular stress response called the unfolded protein response (UPR), intended to protect cells against toxic aggregated proteins (41). The UPR response is mediated by three primary ER signal transducer molecules: RNA-dependent protein kinase (PKR)-like ER kinase (PERK), inositol-requiring enzyme 1 (Ire1α and -β), and transmembrane activating transcription factor 6 (ATF6) (41, 43). If stress is severe, these UPR sensor proteins induce activation of apoptotic pathways to eliminate cells that may be unnecessary or harmful for the organism (39).
Activation of PERK results in phosphorylation of eukaryotic translation initiation factor 2 subunit α (eIF2α), which leads to suppression of general protein translation. ATF4, which is a member of the cAMP-responsive element-binding protein (CREB) family of basic zipper-containing proteins (17), has upstream open reading frames (ORFs) in its 5′ untranslated region. These upstream ORFs, which prevent translation of active ATF4 under normal conditions, are bypassed when eIF2α is phosphorylated, resulting in ATF4 translation (10, 46). ATF4 induces expression of downstream target genes such as GADD34 (26), CHOP/GADD153 (25), and ATF3 (16) and genes encoding proteins involved in amino acid transport, glutathione biosynthesis, and resistance to oxidative stress (13).
In this study, we demonstrate a central role for PERK signaling in SubAB-induced activation of apoptosis as an early event in toxin action. Transient protein synthesis inhibition by SubAB-induced eIF2α phosphorylation via PERK is a pivotal step in initiating apoptosis. Further, our findings here suggest that the ubiquitin-proteasome pathway in association with SubAB-induced protein synthesis inhibition plays a pivotal role in regulation of some antiapoptotic proteins, leading to Bak/Bax conformational changes, cytochrome c release, caspase activation, and apoptosis.
Escherichia coli producing recombinant His-tagged wild-type subtilase cytotoxin (SubAB) or catalytically inactivated mutant SubA(S272A)B (mSubAB) was used as the source of toxins after purification according to a published procedure (31).
Antibodies against cleaved caspase-7, cleaved caspase-9, cleaved PARP, PERK, phospho-eIF2α, eIF2α (Ser51), Ire1α, Mcl-1, Bax, and Bak were purchased from Cell Signaling. Antibodies against GAPDH (FL335), CREB2/ATF4 (C-20), cytochrome c (7H8), ubiquitin (P4D1), c-Myc (9E10), and EIF2AK3/PERK (H300) were purchased from Santa Cruz Biotechnologies, and mouse monoclonal antibodies against BiP/GRP78 and Bax (clone 3) were from BD Biosciences. Anti-Bak (Ab-2) antibody was purchased from Calbiochem, anti-cyclin D1 (DCS6) antibody from NeoMakers, anti-ATF6 antibody from ANA Spec Inc., anti-p97/VCP antibody from Biolegend, Z-VAD-FMK (Z-VAD) from BD Biosciences, MG132, chloroquine, Sal003 and anti-α-tubulin antibody from Sigma-Aldrich, and lactacystin from Enzo Life Science.
HeLa cells were cultured in Eagle's modified essential medium (EMEM) (Sigma) containing 10% fetal bovine serum (FBS). G401, a Wilms' human kidney tumor cell line, and mouse embryonic fibroblast (MEF) cells were grown in Dulbecco's modified essential medium (DMEM) (Sigma) containing 10% FBS. Cells were plated in 24-well dishes (5 × 104 cells/well) or 12-well dishes (1 × 105 cells/well) in EMEM containing 10% FBS. RNA interference-mediated gene knockdown was performed using validated Qiagen HP small interfering RNAs (siRNAs) for PERK (SI02223718), ATF6 (SI03019205), and Ire1α (SI00605255). Mouse PERK siRNA (mPERK-1) was designed and validated as described by Panaretakis et al. (35). ATF4 siRNA was purchased from Sigma-Aldrich as reported previously (15). VCP siRNA was designed and validated as described by Ballar et al. (2). The eIF2α siRNA pool and negative-control siRNA pool were purchased from Santa Cruz Biotechnology. Negative-control siRNAs were purchased from Sigma. siRNA transfections were performed with the Lipofectamine RNAiMax transfection reagent or Lipofectamine 2000 transfection reagent (Invitrogen), according to the manufacturer's instructions.
Coimmunoprecipitation of conformationally changed Bax or Bak was carried out as described previously (29). Briefly, the indicated siRNA-transfected HeLa cells were treated with SubAB or mSubAB for 3 h. After being washed with ice-cold phosphate-buffered saline (PBS), cells were solubilized with lysis buffer (10 mM HEPES, 150 mM NaCl, 1.5 mM MgCl2, 1 mM EGTA, 2% CHAPS [pH 7.4]) containing protease inhibitor cocktail (Roche Diagnostics) and incubated for 30 min on ice. After centrifugation at 17,400 × g for 15 min at 4°C, solubilized extracts (100 μg/200 μl) were collected and incubated with conformation-specific anti-Bax antibody (clone 3) (BD Bioscience) or anti-Bak antibody (Ab-2) (Calbiochem) at 4°C for 3 h. Immunoprecipitates were collected by incubation with protein G-Sepharose (Invitrogen) for 1 h, followed by centrifugation for 1 min at 4°C. After immunocomplexes were washed with lysis buffer three times, and proteins were dissolved in sodium dodecyl sulfate (SDS) sample buffer, subjected to SDS-polyacrylamide gel electrophoresis (PAGE) in 15% gels, transferred to polyvinylidene difluoride (PVDF) membranes, and then analyzed by Western blotting using anti-Bax or anti-Bak antibodies (Cell Signaling).
To detect cytochrome c release from mitochondria into cytosol, cytosolic fraction was collected, following the method described previously (6). Briefly, cells were treated with SubAB (0.2 μg/ml) for 3 h, collected with a cell scraper, resuspended in 50 μl of homogenate buffer (75 mM KCl, 1 mM NaH2PO4, 8 mM Na2HPO4, 250 mM sucrose, 1 mM EDTA) containing 50 μg/ml digitonin and protease inhibitor cocktail (Roche Diagnostics), and incubated on ice for 5 min. Following centrifugation at 10,000 × g for 10 min, the supernatant was collected as cytosolic fractions, and cytochrome c was detected by Western blotting.
To assay SubAB-induced cell death, the indicated siRNA-transfected HeLa cells were plated on 96-well culture dishes (1 × 104 cells/well). After 24 h, control PBS or PBS containing 0.8 μg/ml of SubAB or mSubAB was added to the cell culture medium, and incubation continued for another 24 h. Cell viability was measured with a cell counting kit (Dojindo Chemicals), according to the instruction manual.
First, we investigated which stress sensor proteins in ER are involved in activation of caspases, leading to apoptosis. Previously, we demonstrated that SubAB induced activation of caspase-3, -7, and -9 with similar kinetics, while caspase-8 activation was noted 6 to 18 h after intoxication (data not shown). Therefore, in this study, we chose caspase-7 to investigate signaling through the apoptotic pathway. After 48 h transfection with a nonsilencing control (NC), PERK, ATF6, or Ire1α siRNA, HeLa cells were treated with SubAB or the active-site mutant SubA(S272A)B (mSubAB). The expression amount of PERK, ATF6, or Ire1α was clearly decreased in the siRNA-transfected cells. BiP cleavage by SubAB in these siRNA-transfected cells was observed at the same level. mSubAB treatment did not affect caspase-7 activation and BiP cleavage. SubAB-induced caspase-7 activation was inhibited in PERK-siRNA transfected cells but not in cells transfected with NC, ATF6, or Ire1α siRNA. In addition, densitometric analysis also showed that SubAB-induced caspase-7 activation was significantly suppressed in PERK siRNA-transfected cells (Fig. 1a). Since SubAB-induced apoptosis in HeLa cells was caused by Bax/Bak conformational changes and cytochrome c release, followed by caspase activation (28, 52), we next examined the effect of PERK, ATF6, or Ire1α siRNA transfection of HeLa cells on SubAB-induced cytochrome c release and Bax/Bak conformational changes with conformation-specific anti-Bax (cBax) or -Bak (cBak) monoclonal antibodies. Densitometric analysis demonstrated that SubAB-induced Bax/Bak conformational changes, Bax/Bak oligomerization, and cytochrome c release were dramatically reduced in PERK knockdown cells but not ATF6 or Ire1α knockdown cells (Fig. 1b). Since SubAB induces HeLa cell death after 72 to 96 h of incubation (data not shown), we investigated cell viability in PERK siRNA-transfected cells after 24 h of incubation with toxin. SubAB-induced cell death was inhibited in these cells compared with NC siRNA-transfected cells (Fig. 1c). The human kidney cell line G401 was susceptible to SubAB intoxication, as previously shown (53). We examined the effect of PERK or Ire1α siRNA transfection of G401 cells on SubAB-induced caspase-7 activation. In PERK knockdown G401 cells, similar to HeLa cells, caspase-7 activation was dramatically inhibited, but not Ire1α knockdown (Fig. 1d). Furthermore, we investigated the effect of mouse PERK (mPERK-1) siRNA transfection of MEF cells on SubAB-induced caspase-7 activation. The densitometric analysis demonstrated that suppression of endogenous PERK by mPERK-1 siRNA resulted in a significant downregulation of SubAB-induced caspase-7 activation (Fig. 1e). These results suggest that PERK is an initiator and regulates activation of SubAB-induced apoptosis.
PERK, activated by ER stress, phosphorylates eukaryotic translation initiation factor 2α (eIF2α), which represses protein translation of cellular mRNA (12). Although BiP cleavage by SubAB was same in both NC and PERK siRNA-transfected cells, reduction of PERK expression by siRNA resulted in dramatic inhibition of SubAB-induced eIF2α phosphorylation and capase-7 activation (Fig. 2a). We next investigated the role of eIF2α phosphorylation in SubAB-induced apoptosis. HeLa cells were treated with Sal003, a selective blocker of eIF2α dephosphorylation, in the presence of SubAB or mSubAB. SubAB treatment induced eIF2α phosphorylation and caspase-7 activation in the absence of Sal003 (Fig. 2b, lane 3), while Sal003 pretreatment promoted eIF2α phosphorylation and caspase-7 activation (Fig. 2b, lane 4). Caspase-7 activation and eIF2α phosphorylation by SubAB treatment was slightly enhanced in the presence of Sal003 (Fig. 2c, lane 6). As expected, mSubAB gave a result similar to Sal003 alone (Fig. 2b, lane 5).
To determine if eIF2α and ATF4 signaling are involved in SubAB-induced apoptosis, eIF2α or ATF4 siRNA-transfected HeLa cells were treated with SubAB or mutant SubAB. Although endogenous eIF2α expression and SubAB-induced ATF4 expression were attenuated by eIF2α siRNA transfection, densitometric analysis showed that eIF2α knockdown induced both caspase-7 activation and cytochrome c release even in the presence of mSubAB, and those changes were enhanced by SubAB (Fig. 3a). Next, we examined the effect of eIF2α knockdown on SubAB-induced Bax/Bak conformational changes and oligomerization by immunoprecipitation using conformational-change-specific Bak or Bax antibodies. As shown in Fig. 3b, eIF2α knockdown induced Bax/Bak conformational changes and their oligomerization, which was promoted in the presence of SubAB. In agreement with these results, cell viability was dramatically decreased by eIF2α knockdown alone, and the decrease in viability was enhanced in the presence of SubAB but not mSubAB (Fig. 3c).
To investigate whether the expression of ATF4 by SubAB was mediated through PERK, PERK, ATF6, or Ire1α siRNA-transfected HeLa cells were treated with SubAB or mSubAB. As shown in Fig. 4a, SubAB-induced ATF4 expression was inhibited in PERK knockdown cells but not ATF6 or Ire1α knockdown cells, consistent with previous studies using a chemical ER stress inducer (46).
Although SubAB-induced ATF4 expression in a time-dependent manner was clearly suppressed 24 h after ATF4 siRNA transfection, densitometric analysis showed that caspase-7 activation by SubAB was not inhibited (Fig. 4b). In addition, densitometric analysis demonstrated that suppression of endogenous level of ATF4 by prolonged treatment with ATF4 siRNA in the presence of mSubAB resulted in a significant upregulation of caspase-7 activation and cytochrome c release, which were enhanced by SubAB treatment (Fig. 4c). However, ATF4 knockdown did not induce Bax/Bak conformational changes, suggesting that cytochrome c release caused by ATF4 knockdown is likely to be independent of Bax and Bak and also results in caspase activation (Fig. 4d). Cell viability in ATF4 knockdown cells with mSubAB or PBS as a control was also significantly decreased; however, it was enhanced by SubAB (Fig. 4e). Taken together, these findings indicate that eIF2α and ATF4 play essential roles in HeLa cell viability.
Our previous studies demonstrated that MG132 pretreatment inhibited early events in SubAB-induced caspase activation (54). MG132 is not only an inhibitor of proteasome and lysosome but also an ER stress inducer (21, 27). To determine which pathways are associated with SubAB-induced caspase activation, cells were pretreated with MG132, lactacystin (proteasome inhibitor), chloroquine (lysosome inhibitor), or Z-VAD (general caspase inhibitor) and then incubated with SubAB or mSubAB. Densitometric analysis demonstrated that SubAB-induced activation of caspase-7 and caspase-9 and PARP cleavage was significantly inhibited by MG132, lactacystin, or Z-VAD but not by chloroquine or DMSO as a control. SubAB-induced BiP cleavage was not affected by these inhibitors (Fig. 5a). To investigate whether SubAB treatment activates the ubiquitin-proteasome system, we examined the effect of SubAB on c-Myc, p53, and Mcl-1, which are known to undergo rapid proteolysis by the 26S proteasome in vivo (9, 56). Our previous report showed that Mcl-1 content was decreased by SubAB treatment in a time-dependent manner (52). Cells were treated with SubAB or mSubAB in the presence or absence of MG132 or Z-VAD. In the presence of MG132, but not Z-VAD, toxin-induced degradation of c-Myc, p53, and Mcl-1 and cytochrome c release were inhibited. Moreover, in the presence of SubAB or mSubAB, MG132 treatment also induced eIF2α phosphorylation and ATF4 expression. Pretreatment of HeLa cells with a general caspase inhibitor (Z-VAD) did not suppress SubAB-induced eIF2α-phosphorylation, ATF4 expression, or cytochrome c release (Fig. 5b). These data show that, via the ubiquitin-proteasome pathway, BiP cleavage by SubAB promotes proteolysis of antiapoptotic (i.e., Mcl-1) or proapoptotic proteins (i.e., c-Myc, p53) or possibly Bax/Bak conformational-change-related proteins, resulting in cytochrome c release.
We next investigated the level of polyubiquitinated proteins in SubAB-treated cells. After HeLa cells were incubated with SubAB or mSubAB in the presence or absence of MG132, lactacystin, or chloroquine for 3 h, the levels of polyubiquitinated proteins in lysate were analyzed by immunoblotting using antipolyubiquitin antibody by short or long exposure (Fig. 5c). Densitometric analysis also showed that similar amounts of ubiquitinated proteins from SubAB- and mSubAB-treated cells were observed in MG132- and lactacystin-treated cells. However, compared with mSubAB-treated cells, the amounts of low-molecular-weight polyubiquitinated proteins in control- or chloroquine-incubated cells were significantly decreased by SubAB treatment (Fig. 5c, lanes 2 and 8).
To examine whether PERK controls the level of polyubiquitinated proteins, PERK knockdown cells were incubated with SubAB or mSubAB for 3 h, and then the amounts of polyubiquitinated proteins were analyzed as described above. In PERK knockdown cells, compared with SubAB-treated control siRNA-transfected cells, densitometric analysis showed that the amounts of low-molecular-weight polyubiquitinated proteins were not decreased, even in the presence of SubAB (Fig. 5d, lanes 2 and 4).
Next, we investigated the effect of SubAB on rapid-turnover proteins such as c-Myc and cyclin D1 in PERK knockdown cells. It was reported that a truncated form of PERK, lacking its kinase domain, acted as a dominant negative when overexpressed in cells, attenuating both cyclin D1 loss and cell cycle arrest during the UPR, without compromising induction of ER chaperones (4). In control siRNA-transfected cells, BiP cleavage was observed after about 30 min incubation with SubAB, followed by transient reduction in c-Myc level, a moderate reduction in cyclin D1, and an increase in caspase-7 activation; mSubAB was inactive. In contrast, after SubAB-induced BiP cleavage in PERK knockdown cells, degradation of c-Myc and cyclin D1 and caspase-7 activation were not detected (Fig. 5e). These findings support the assertion that PERK-mediated events regulate rapidly turning-over proteins, which are associated with cell death.
An early event in SubAB-induced apoptosis involves short-lived antiapoptotic or proapoptotic proteins that play an important role in Bax/Bak conformational changes and oligomerization, leading to cytochrome c release. Mcl-1, an antiapoptotic member of the family, is known to undergo rapid degradation by the proteasome, with rapid turnover following inhibition of protein synthesis but preceding Bax/Bak activation and onset of apoptosis (1). Recently, Xu et al. reported that proteasomal degradation of Mcl-1 is required for action of p97/VCP, which mediates disparate cellular functions, including ERAD, via the ubiquitin-proteasome system (51). In HeLa cells, SubAB treatment induced activation of ERAD (20). To address whether ERAD is associated with SubAB-induced apoptosis, we carried out p97/VCP knockdown by siRNA and then determined the effect of SubAB on caspase activation. Expression of p97/VCP was clearly decreased in p97/VCP siRNA-transfected cells. Densitometric analysis demonstrated that SubAB-induced caspase-7 activation was not suppressed in p97/VCP knockdown cells, suggesting that neither Mcl-1 nor ERAD is involved in SubAB-induced apoptosis (Fig. 6).
When unfolded proteins become excessive, ER stress-induced apoptosis is mediated by three primary signal transducers, i.e., PERK, Ire1, and ATF6 (41, 43). These proteins act by several mechanisms, including direct activation of caspases, kinases, transcription factors, and Bcl-2-family proteins (50). Other studies showed that in ER stress, JNK activation, coupled with Ire1α-TRAF2-ASK1 signaling, induced apoptotic cell death (23, 33, 45). We have demonstrated that SubAB-induced apoptosis did not follow a pathway involving ER stress sensor Ire1α activation in HeLa cells (52). A recent study showed that in NRK-52E rat renal tubular epithelial cells, dominant negative inhibition of PERK enhanced SubAB-induced apoptosis and reduced phosphorylation of ERK and Akt (44). We show here that in HeLa cells, PERK knockdown by siRNA triggers a significant attenuation of SubAB-induced cytotoxicity with decreased Bax/Bak conformational changes and oligomerization, reduced cytochrome c release, and thus decreased caspase activation. We also observed that, in PERK knockdown G401 cells and mPERK knockdown MEF cells, similar to the findings in HeLa cells, SubAB-induced caspase-7 activation was inhibited. There are cases in which PERK activation promotes or protects against ER stress-induced apoptosis (11, 19, 22, 36). These studies imply that PERK plays different roles in regulation of cell proliferation or apoptosis, depending on the cell type or species.
Here we show that proteasome inhibitor, e.g., lactacystin, MG132, pretreatment of HeLa cells inhibited toxin-induced proteosomal degradation of ubiquitinated proteins and apoptotic signaling at an early stage. The ubiquitin-proteasome pathway plays an essential role in protein homeostasis, which is fundamental to the control of cell survival (14). In addition, we demonstrated in PERK knockdown HeLa cells that SubAB-induced loss of ubiquitinated proteins, including c-Myc and cyclin D1, was inhibited, resulting in suppression of apoptosis. Consistent with our results, previous studies found a functional cross talk between PERK-mediated eIF2α phosphorylation and the proteasomal degradation of cyclin D1 and that this degradation was dependent upon eIF2α phosphorylation (4, 40). They observed that PERK serves as a critical effector of ER stress-induced growth arrest, linking stress in the ER to control of cell cycle progression rather than cell death. Thus, our study demonstrates that in HeLa cells, SubAB-induced activation of PERK leads to protein degradation by the ubiquitin-proteasome pathway, resulting in activation of apoptosis.
ER stress-induced eIF2α phosphorylation represses translation initiation of most transcripts, with loss of short-lived proteins (12). Our previous data showed that after treatment of cells with SubAB, protein synthesis was transiently suppressed for 2 h and then moderately recovered; these events were probably regulated by a pathway downstream of eIF2α phosphorylation (30). A recent study showed that eIF2α kinases may mediate UV light-induced apoptosis via eIF2α-dependent or -independent signaling pathways (36). In contrast, studies of β-amyloid (Aβ)-mediated cell death in SK-N-SH human neuroblastoma cells demonstrated that maintaining levels of phosphorylation of eIF2α inhibited cell death (22). Thus, eIF2α kinase plays different roles in regulation of apoptosis due to various stimuli or in different cell types.
In HeLa cells, SubAB-induced eIF2α phosphorylation was decreased in PERK knockdown cells (Fig. 2). Further, pretreatment with Sal003, an eIF2α phosphatase inhibitor, with or without SubAB enhanced eIF2α phosphorylation and slightly promoted SubAB-induced caspase activation. eIF2α phosphatase inhibitor is known not only to enhance cell survival even when apoptosis is induced by ER stressors (3, 22) but also to restrict cell growth (42). In HeLa cells, protein synthesis inhibition by toxin-induced eIF2α phosphorylation may trigger caspase activation.
Although phosphorylation of eIF2α by ER stress leads to inhibition of general protein translation, transcription factor ATF4 content is selectively enhanced. ATF4 regulates the expression of downstream target genes of both antiapoptotic and apoptotic pathways (43). Our results show that transient suppression of ATF4 by siRNA did not inhibit SubAB-induced caspase activation; however, prolonged ATF4 suppression enhanced SubAB-induced cytochrome c release via a Bax/Bak-independent pathway and promoted caspase activation, leading to apoptosis (Fig. 4). These data suggest that ATF4 targets, e.g., CHOP (43), do not contribute to SubAB-induced apoptosis; rather, ATF4 probably plays an important role in HeLa cell survival. Consistent with our data, Ye and collaborators demonstrated that ATF4 knockdown inhibited HT1080 cell proliferation and enhanced apoptosis (55).
Our previous study showed that signaling pathways associated with functional SubAB receptors may be required for activation of SubAB-dependent apoptotic pathways (54). A recent study showed that in chondrocytes, the presence of extracellular matrix alters the response to ER stressors through delayed Grp78 expression and inhibition of ER stress-induced apoptosis, suggesting that signaling by the extracellular matrix affects the chondrocyte ER stress response (34). Since SubAB receptors, e.g., α2β1 integrin and NG2, are known to associate with extracellular matrix (24), we propose that knockdown of SubAB receptors affects the ER stress response, resulting in inhibition of SubAB-induced apoptosis. Further studies are necessary to determine how SubAB-induced PERK-mediated apoptotic signals are linked to signals from SubAB receptors.
In conclusion, we identified a new role for PERK signaling in the regulation of SubAB-induced apoptosis (Fig. 7). BiP cleavage by SubAB in the ER causes PERK activation, followed by eIF2α phosphorylation, leading to transient inhibition of protein synthesis. During transient shutdown of protein synthesis, certain antiapoptotic or proapoptotic proteins, e.g., Mcl-1, p53, and c-Myc, are rapidly turned over; they are ubiquitinated and then degraded in the ubiquitin-proteasome system, leading to Bak/Bax conformational changes and their oligomerization, followed by cytochrome c release, and caspase activation. PERK knockdown by siRNA blocked SubAB-induced apoptosis, with inhibition of eIF2α phosphorylation, ATF4 expression, Bak/Bax conformational changes, cytochrome c release, caspase activation, and ubiquitinated protein degradation. In addition, proteasome inhibitors (MG132 and lactacystin) could suppress SubAB-induced cytochrome c release, followed by caspase activation. Thus, SubAB-induced PERK-mediated Bax/Bak conformational changes are probably regulated by certain ubiquitinated factors, leading to activation of apoptosis. Knockdown of eIF2α induces Bax/Bak conformational changes and oligomerization, followed by cytochrome c release, resulting in caspase activation, a pathway that is enhanced by SubAB-induced BiP cleavage. In addition, increased ATF4 expression is likely to contribute to resistance to cell death. Further studies are necessary to determine the ubiquitinated factors involved in SubAB-induced PERK-mediated Bax/Bak conformational changes, followed by activation of apoptosis.
This work was supported by grants in aid for Scientific Research from the Ministry of Education, Science and Culture of Japan and Improvement of Research Environment for Young Researchers from the Japan Science and Technology Agency. Joel Moss was supported by the Intramural Research Program, National Institutes of Health, National Heart, Lung, and Blood Institute.
We thank T. Hirayama (Institute of Tropical Medicine, Nagasaki University) and N. Morinaga (Chiba University) for helpful discussions and suggestions. We acknowledge the expert technical assistance of C. Noritake and A. Kiuchi.
Published ahead of print 21 February 2012