N. Shimamoto, Laboratory of Pharmacology, Faculty of Pharmaceutical Sciences at Kagawa, Tokushima Bunri University, 1314-1, Shido, Sanuki, Kagawa 769-2193, Japan Fax: +81 87 894 0181 Tel: +81 87 894 5111 ext. 6513 E-mail: email@example.com
We previously reported that the inhibition of catalase and glutathione peroxidase activities by treatment with 3-amino-1,2,4-triazole (ATZ) and mercaptosuccinic acid evoked sustained increases in the levels of reactive oxygen species and apoptosis in rat primary hepatocytes. Apoptosis was accompanied by increased expression of BimEL, following activation of extracellular signal-regulated kinase. The aim of this study was to characterize the mechanism underlying hepatocyte apoptosis by identifying the transcription factor that induces BimEL expression. The bim promoter region was cloned into a promoterless-luc vector, and promoter activity was monitored by a luciferase assay. The luciferase activity increased in the presence of ATZ + mercaptosuccinic acid. Pretreatment with a MEK inhibitor, U0126, or an antioxidant, vitamin C, suppressed the promoter activity. Furthermore, ATZ + mercaptosuccinic acid-induced luciferase activity was attenuated by mutation of the activator protein-1 binding site in the bim promoter region. The amounts of total and phosphorylated c-Fos increased over time in the presence of ATZ + mercaptosuccinic acid, whereas the amounts of total and phosphorylated c-Jun remained unchanged. Chromatin immunoprecipitation revealed that both c-Fos and c-Jun localized to the activator protein-1-binding site in the bim promoter region. BimEL expression and hepatocyte apoptosis were suppressed by knockdown of c-Fos and c-Jun, respectively. These results indicate that increases in c-Fos following extracellular signal-regulated kinase activation are critical for BimEL upregulation and apoptosis.
Apoptosis has several morphological features, including cell shrinkage, nuclear condensation, and nucleosomal DNA fragmentation. Extensive studies to uncover the mechanisms underlying the induction of apoptosis have yielded the generally accepted theory that mitochondria play a fundamental role in the process. Apoptotic stimuli activate the mitochondrial permeability transition pore and the release of apoptosis-promoting molecules such as cytochrome c, apoptosis-inducing factor, and endonuclease G . The pathways upstream of the mitochondria for apoptotic signal transduction have recently been identified. Several molecules that are known to be involved in proliferation and/or differentiation have been reported to induce apoptosis [2,3].
Extracellular signal-regulated kinase (ERK) is a classic mitogen-activated protein kinase that is activated by growth factors and induces cell cycle progression via cyclin transcription. However, increasing evidence shows that ERK is activated by reactive oxygen species (ROS), and that this is followed by the induction of apoptosis [4–6]. ERK-dependent apoptosis induced by ROS has been recognized in several pathological conditions, such as alcoholic liver injury [7,8], lung hyperoxia , and cisplatin-induced renal toxicity . However, little is known about the mechanism responsible for apoptotic signaling elicited by active ERK, and this process therefore needs to be investigated.
The mechanism responsible for ERK activation by ROS is well understood. The phosphorylation of ERK or its upstream kinases is regulated by phosphatases such as PTP1B , MKP3 , and LMW-PTP . The cysteines in the active sites of these phosphatases are easily inactivated by ROS, resulting in activation of the ERK pathway . However, factors that act on the mitochondria downstream of ERK have been rarely reported. Recently, we showed that ROS-activated ERK increased the transcriptional expression of BimEL, a major isoform among the bim gene products, leading to apoptosis in rat primary hepatocytes .
Bim is a member of the Bcl-2 family of proteins, which play a fundamental role in the induction of mitochondria-driven apoptosis. Under normal conditions, antiapoptotic Bcl-2 family members such as Bcl-2, Bcl-xL and Mcl-1 interact with the proapoptotic Bcl-2 family members Bax/Bak, to inhibit the ability of Bax/Bak to permeabilize the mitochondrial membrane. Bim activates the mitochondrial permeability transition mediated by Bax/Bak through two different mechanisms : (a) Bim binds to antiapoptotic Bcl-2 family proteins to liberate Bax/Bak, leading to mitochondrial permeability transition; and (ii) Bim directly activates Bax/Bak (induces a conformational change), thus leading to pore formation.
The bim gene is a direct target of transcription factors such as FOXO3A, Myb and c-Jun [17–21]. The 5′-end of the bim gene contains binding sites for FOXO, Myb, and activator protein-1 (AP-1) . However, the mechanism underlying the transcriptional activation of BimEL downstream of ERK activation is not known. The aim of this study was to identify the ERK-responsive transcription factor that regulates BimEL expression.
We previously showed that treatment with 3-amino-1,2,4-triazole (ATZ) and mercaptosuccinic acid inhibited catalase and glutathione peroxidase, which are antioxidative enzymes that eliminate hydrogen peroxide, and caused sustained increases in ROS levels and apoptosis in rat primary hepatocytes [22,23]. In addition, we recently reported that ROS-activated ERK induces BimEL transactivation, followed by hepatocyte apoptosis . This study was designed to examine the mechanism of hepatocyte apoptosis, with a particular focus on identifying the transcription factor(s) that activate BimEL transcription downstream of the ERK pathway.
We cloned a 2.9-kb fragment of the rat bim promoter region from rat primary hepatocytes. The bim promoter region included an AP-1-binding site, a FOXO-binding site, and three Myb-binding sites (Fig. 1A). The bim promoter region was subcloned into pGL4.24 (pGL4.24-BimProm). pGL4.24-BimProm mutations were generated at each transcription factor-binding site (mutated points are indicated in Fig. 1A), and bim promoter activity in the presence of ATZ + mercaptosuccinic acid was assessed with a luciferase reporter assay. The mutations at the binding sites used in this study reportedly attenuate the activity of each transcription factor [19,24,25]. When rat primary hepatocytes were transfected with pGL4.24-BimProm and treated with ATZ + mercaptosuccinic acid for 9 h, the luciferase activity increased 3.3 ± 0.3-fold in comparison with untreated cells (Fig. 1B). However, pretreatment with U0126, a potent inhibitor of MEK1, or vitamin C, an antioxidant, largely suppressed ATZ + mercaptosuccinic acid-induced luciferase activity (Fig. 1B). In addition, when rat primary hepatocytes were transfected with a mutated AP-1 (AP-1m) promoter construct, the ATZ + mercaptosuccinic acid-mediated increase in luciferase activity was greatly attenuated (Fig. 1B). Transfection with a promoter construct containing myb1m had no effect on the luciferase activity, whereas transfection with myb2m, myb3m or FOXOm promoters partially suppressed ATZ + mercaptosuccinic acid-induced luciferase activity (Fig. 1B). These results suggest that AP-1 is involved in increasing BimEL expression downstream of ERK activation in response to treatment with ATZ + mercaptosuccinic acid.
The AP-1 transcription factor consists of Fos and Jun proteins . Fos and Jun form a dimer, which in turn binds to AP-1 regulatory elements and enhancer regions of numerous mammalian genes. Jun forms homodimers and heterodimers with Fos proteins, whereas Fos proteins do not form homodimers, and require heterodimerization to bind DNA [27,28]. Active ERK phosphorylates one of the major Fos family proteins, c-Fos, and stabilizes it ; ERK also phosphorylates c-Jun directly, leading to transactivation of AP-1. On the basis of these findings, we next examined the expression and phosphorylation of c-Fos and c-Jun. The total amount of nuclear c-Fos increased over time in the presence of ATZ + mercaptosuccinic acid (Fig. 2). Interestingly, phosphorylation of c-Fos at Ser374 occurred in parallel with increases in nuclear c-Fos levels (Fig. 2). Pretreatment with U0126 or vitamin C largely suppressed the accumulation of total and phosphorylated c-Fos in the presence of ATZ + mercaptosuccinic acid (Fig. 2). In contrast, there were no changes in the levels of total and phosphorylated nuclear c-Jun throughout the 9-h exposure to ATZ + mercaptosuccinic acid (Fig. 2).
To show that AP-1 proteins directly bind to the consensus AP-1 site in the bim promoter region (from −2491 to −2497), a chromatin immunoprecipitation (ChIP) assay was performed. A PCR analysis demonstrated that c-Fos and c-Jun antibodies apparently precipitated the bim promoter region from rat primary hepatocytes treated with ATZ + mercaptosuccinic acid, whereas untreated hepatocytes and those pretreated with U0126 or vitamin C showed only slight DNA binding (Fig. 3). Pretreatment with SP600125, an inhibitor of c-Jun N-terminal kinase, showed no effect on the DNA binding of c-Fos and c-Jun induced by treatment with ATZ + mercaptosuccinic acid, indicating that JNK is not involved in the binding of AP-1 to the bim promoter region. Nonspecific IgG also did not exhibit DNA-binding activity (Fig. 3). These results indicate that the AP-1 proteins bind specifically to the AP-1 cis-regulatory region of the bim promoter in hepatocytes treated with ATZ + mercaptosuccinic acid.
Next, we examined the effect of c-Fos and c-Jun on BimEL transactivation and apoptosis. Transfection with small interfering RNAs (siRNAs) targeted to c-Fos and c-Jun clearly reduced the target protein levels (Fig. 4A). Elevation of BimEL mRNA expression by treatment with ATZ + mercaptosuccinic acid was suppressed by transfection with siRNAs against c-Fos and c-Jun (Fig. 4B). Increases in BimEL levels caused by ATZ + mercaptosuccinic acid were also attenuated by c-Fos or c-Jun knockdown (Fig. 4C). AT Z+ mercaptosuccinic acid-induced cell death, chromatin condensation and DNA fragmentation were all abrogated by knockdown of c-Fos and c-Jun (Fig. 5A–C). Transfection of scrambled siRNAs showed no effects on the expression levels of c-Fos, c-Jun, or BimEL, and did not affect hepatocyte apoptosis (Figs 4 and 5). These results indicate that c-Fos and c-Hun are crucial for BimEL expression and induction of hepatocyte apoptosis.
The bim promoter activity induced by treatment with ATZ + mercaptosuccinic acid was largely attenuated by mutating the AP-1-binding site in the bim promoter region. Whereas the amounts of total and phosphorylated c-Fos increased in the presence of ATZ + mercaptosuccinic acid, there was no change in the levels of total and phosphorylated c-Jun throughout the experimental period. Both c-Fos and c-Jun interacted with the AP-1-binding site in the bim promoter region. Knockdown of c-Fos or c-Jun suppressed not only BimEL transactivation, but also hepatocyte apoptosis.
Pretreatment with U0126 or vitamin C largely abolished ATZ + mercaptosuccinic acid-induced luciferase activity, confirming that the ERK pathway elicited by ROS is involved in Bim transcription in this experimental system [15,23]. In addition, mutation of the AP-1-binding site in the bim promoter region markedly suppressed the luciferase activity induced by ATZ + mercaptosuccinic acid, suggesting that AP-1 is responsible for Bim transcription. Biswas et al. reported that Bim expression was coregulated by three transcription factors – c-Jun, FOXO, and Myb – when PC12 cells were stimulated by nerve growth factor deprivation, and insisted that the bim promoter acts as a coincidence detector . Interestingly, mutation of the Myb-binding and FOXO-binding sites also slightly, but significantly, reduced the luciferase activity in this study. Therefore, the involvement of FOXO and Myb in hepatocyte apoptosis should be examined further.
c-Fos is one of the main components of the AP-1 transcription factor complex . Activated ERK phosphorylates c-Fos at Ser-374, leading to its stabilization [29,31]. Therefore, we examined the expression and phosphorylation of c-Fos in this study. The total and phosphorylated c-Fos levels increased over time in the presence of ATZ + mercaptosuccinic acid, and this increase was suppressed by pretreatment with U0126. Therefore, c-Fos is stabilized by phosphorylation, which is mediated by ERK, allowing c-Fos to accumulate. In contrast, c-Jun, another major component of the AP-1 complex, is reportedly phosphorylated at Ser63 and Ser73 by active ERK, and this is followed by increased c-Jun transcriptional activity [32,33]. However, the total and phosphorylated c-Jun levels in nuclei remained unaffected in the presence of ATZ + mercaptosuccinic acid. Because c-Fos alone cannot bind to DNA, c-Jun is required for transcriptional activation [27,28]. Thus, BimEL expression is dependent on both increased levels of c-Fos and basal levels of c-Jun. This idea is supported by the results of the ChIP assay, which indicated that both c-Fos and c-Jun localize to the AP-1-binding site in the bim promoter region. Furthermore, knockdown of c-Fos or c-Jun attenuated BimEL transactivation and apoptosis, supporting the hypothesis that c-Fos and c-Jun act coordinately to increase the expression of BimEL. Increased c-Fos levels are therefore critical for BimEL expression and apoptosis in this experimental system.
Active ERK is known to phosphorylate BimEL, resulting in the ubiquitination and degradation of BimEL [34,35]. Therefore, ERK activation was expected to reduce the level of BimEL, leading to increased cell survival as long as the proteasome maintains its normal functions. We previously reported that BimEL degradation was suppressed in this experimental system, because ROS generated by treatment with ATZ + mercaptosuccinic acid inhibited the activities of the proteasome . Namely, BimEL was upregulated by both increased expression and decreased degradation in this type of hepatocyte apoptosis. c-Fos was also reported to be degraded by the ubiquitin–proteasome system . In this study, pretreatment with U0126 did not completely abrogate the c-Fos expression induced by treatment with ATZ + mercaptosuccinic acid. Therefore, proteasome inhibition by ROS might be involved in the increased expression of c-Fos in this experimental system. The mechanism(s) underlying the upregulation of c-Fos should be examined in greater detail.
The duration of the ERK signal is reported to be important for c-Fos stability . Transient activation of ERK could increase c-Fos transcription but could not lead to c-Fos phosphorylation, because the ERK signal is inactivated when c-Fos protein is synthesized. Nonphosphorylated c-Fos is rapidly degraded by the ubiquitin–proteasome system [29,36]. In contrast, sustained ERK activation increases c-Fos transcription and phosphorylation, leading to phosphorylated c-Fos accumulation. Therefore, under conditions where ERK is persistently activated, c-Fos could transcriptionally activate several genes, together with c-Jun. In this experimental model, ERK was activated for 9 h after the addition of ATZ + mercaptosuccinic acid, owing to inactivation of protein tyrosine phosphatase caused by sustained increases in intracellular ROS levels . Therefore, we concluded that AP-1-dependent gene expression occurred under the conditions of sustained oxidative stress. This idea is supported by data showing that transient oxidative stress for 3 or 6 h did not induce apoptosis .
In conclusion, ERK activation resulting from sustained oxidative stress increased the amounts of total and phosphorylated nuclear c-Fos. Increased c-Fos and basal c-Jun localized to the AP-1-binding site in the bim promoter region and induced transcription of BimEL mRNA, followed by hepatocyte apoptosis. Therefore, the increase in c-Fos downstream of ERK activation is critical for BimEL upregulation and apoptosis. The duration of exposure to oxidative stress affects c-Fos stability and BimEL expression by changing the duration of the ERK signal. Therefore, the duration of oxidative stress might be a fundamental determinant of cellular fate.
ATZ and mercaptosuccinic acid were from Sigma–Aldrich (St Louis, MO, USA). U0126 was from Promega (Madison, WI, USA). SP600125 was from Bio Mol (Plymouth Meeting, PA, USA). Vitamin C was from Wako Pure Industries (Osaka, Japan). All other chemicals were obtained from Sigma–Aldrich or Wako Pure Industries, and were of the highest quality commercially available.
Preparation of rat primary hepatocytes
All procedures performed on animals were in accordance with the Fundamental Guidelines for Proper Conduct of Animal Experiment and Related Activities in Academic Research Institutions under the jurisdiction of the Ministry of Education, Culture, Sports, Science and Technology, Japan, and the Animal Care and Use Committee of Tokushima Bunri University, Kagawa, Japan.
Rat primary hepatocytes were prepared from male Wistar rats (body weight of 150–200 g) (Nippon CLEA, Osaka, Japan) by collagenase perfusion, as described in our previous report . Cells were plated onto collagen type I-coated dishes in hepatocyte culture medium (Williams’ medium E containing 10% fetal bovine serum, 300 nm insulin, and 100 nm dexamethasone). After a 2-h attachment period, the medium was exchanged and cells were used for experiments.
Cloning and site-directed mutagenesis of the rat bim promoter region
Rat genomic DNA was extracted from rat primary hepatocytes with the DNeasy Blood & Tissue Kit (Qiagen, Valencia, CA, USA). The bim promoter region, including the transcriptional initiation site (2903 bp), was amplified with Platinum Taq DNA polymerase (Invitrogen, Carlsbad, CA, USA) (primers: Fw, 5′-GCCAGGCGAGAAATTTAGTGTC-3′; and Rv, 5′-CAACAAGCTGTTGACCCAGTG-3′), and ligated into pGL4.24 to create pGL4.24-BimProm, which contains a BimProm-luc transcriptional fusion.
Mutation of the binding sites for AP-1, Myb and FOXO in pGL4.24-BimProm was performed by site-directed mutagenesis with the QuikChange kit (Stratagene, Santa Clara, CA, USA) (primers: AP-1 Se, 5′-CCGTCAGCGGTGACTTGGATTCACAGAGAC-3′; FOXO Se, 5′-CAAGTCACTAGGGTACCCACGCCGGGGTGG-3′; Myb1 Se, 5′-GACCAAGATGGTCCATCGGTGGGACGACAG-3′; Myb2 Se, 5′-CTCCCTGGTCTCTCATCTGTCCTTCCCACC-3′; Myb3 Se, 5′-CCTCCTGAGGCTTCCATCTGGCGGCCGCGG-3′). Mutations were confirmed by nucleotide sequencing.
Transfection and luciferase activity assays
Cells were cotransfected with pGL4.24-BimProm or mutant pGL4.24-BimProm and with pRL-RSV, using the Nucleofection system (Amaxa, Koln, Germany), as described previously . Luciferase reporter activity was measured with the Dual-Glo Luciferase Assay System (Promega). Firefly luciferase activity was normalized to Renilla luciferase activity and total protein levels.
Extraction of nuclear proteins and immunoblotting
Nuclear extracts were prepared according to our previous report, with slight modifications . Briefly, cells were suspended in buffer A (10 mm Hepes, pH 7.8, 10 mm KCl, 2 mm MgCl2, 0.1 mm EDTA, 0.5 mm dithiothreitol, and protease inhibitor cocktail) and incubated on ice for 15 min. Nonidet-40 at a final concentration of 0.6% was added to the cell suspension, which was immediately vortexed and centrifuged at 18 000 g for 30 sec. A white pellet was washed with buffer A and used as a nuclear fraction.
Equal amounts of protein were loaded and separated by SDS/PAGE with a 10% or 12% (w/v) polyacrylamide gel and transferred onto a poly(vinylidene difluoride) membrane. The blocked membranes were incubated with primary antibodies [anti-c-Fos; Rabbit IgG (Cell Signaling Technology, Danvers, MA, USA); anti-c-Fos pSer374; Mouse IgG1 (Calbiochem, Darmstadt, Germany); anti-c-Jun; Rabbit IgG (Cell Signaling Technology); anti-c-Jun pSer63; Rabbit IgG (Cell Signaling Technology); anti-c-Jun pSer73; Rabbit IgG (Cell Signaling Technology); anti-Bim; Rabbit IgG (Cell Signaling Technology); anti-β-actin; Goat IgG (Santa Cruz, CA, USA); anti-histone H1; Mouse IgG2a (Santa Cruz)]. The membranes were incubated with an Alexa680-conjugated secondary antibody (Invitrogen) and visualized.
The cells were fixed in 1% formaldehyde for 10 min at room temperature, and immunoprecipitation was performed with antibodies against c-Fos and c-Jun (Santa Cruz), or control IgG, with the ChIP-IT Express kit (Active Motif, Carlsbad, CA, USA), according to the manufacturer’s instructions. The immunoprecipitates including DNA were analyzed by PCR (primers: Fw, 5′-CCAGACAATCGTCTCGCCCA-3′; and Rv, 5′-GGCTAGGTAACAGTTTAGCGAGGA-3′). Rat genomic DNA extracted from rat primary hepatocytes was used as a positive control. PCR products were analyzed by electrophoresis on 1.5% agarose gels.
Total RNA isolation and real-time PCR
Total RNA extraction from hepatocytes was performed with an RNeasy Mini Kit (Qiagen). First-strand cDNA was synthesized from total RNA with a ThermoScript RT-PCR System (Invitrogen). The level of mRNA for BimEL was measured by real-time quantitative RT-PCR with a 7500 Real-Time PCR System (Applied Biosystems, Foster City, CA, USA), according to our previous report . The sequences of the forward and reverse primers were: Fw, 5′-CCAGATCCCCACTTTTCATC-3′; and Rv, 5′-AAGAGAAATACCCACTGGAGGA-3′. The sequence of the TaqMan fluorogenic probe was 5′-TGCTGTCC-3′ (Universal ProbeLibrary, Roche Diagnostics, Basel, Switzerland). BimEL mRNA levels were corrected by glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA.
Assays for cell death and apoptotic features
Chromatin condensation was assessed with the DNA-binding fluorochrome Hoechst 33342. Nuclei were visualized with a BX51WI fluorescence microscope (Olympus, Tokyo, Japan). To detect DNA fragmentation, an Apoptosis DNA Ladder Kit (Wako) was used.
The siRNA targeted to rat c-Fos was synthesized by Sigma Genosys (Ishikari, Japan) (Se: 5′-CCGAGAUUGCCAAUCUACUTT-3′). The siRNAs targeted to rat c-Jun (siTrio, Cat. No. SRF27A-2035) were purchased from B-Bridge International (Mountain View, CA, USA). Scrambled siRNAs against c-Fos and c-Jun siRNAs were synthesized by Sigma Genosys (scrambled c-Fos siRNA Se, 5′-GUACGCUACCACACUUGAUTT-3′; scrambled c-Jun siRNA1 Se, 5′-GGGAACAGAGCGGAUAGGATT-3′; scrambled c-Jun siRNA2 Se, 5′-GAAAGAUGGCAGAAUAGAATT-3′; and scrambled c-Jun siRNA3 Se, 5′-GAAAGCCUUAAGAAUUGUATT-3′). The transfection of rat primary hepatocytes with siRNA(s) was carried out by electroporation with the Nucleofection system (Amaxa), according to our previous report .
Data for each variable are expressed as the means ± standard error (SE). The data obtained from two groups were compared by the use of Student’s t-test, and data obtained from three or more groups were compared by the use of Dunnett’s test. P-values < 0.05 were considered to be significant.
We thank T. Ohshima for helpful discussions, and T. Shinohara for technical contributions.