ABSTRACT
Farnesoid X receptor α (FXR) is highly expressed in the liver and regulates the expression of various genes involved in liver repair. In this study, we demonstrated that activated poly(ADP-ribose) polymerase 1 (PARP1) promoted hepatic cell death by inhibiting the expression of FXR-dependent hepatoprotective genes. PARP1 could bind to and poly(ADP-ribosyl)ate FXR. Poly(ADP-ribosyl)ation dissociated FXR from the FXR response element (FXRE), present in the promoters of target genes, and suppressed FXR-mediated gene transcription. Moreover, treatment with a FXR agonist attenuated poly(ADP-ribosyl)ation of FXR and promoted FXR-dependent gene expression. We further established the CCl4-induced acute liver injury model in wild-type and FXR-knockout mice and identified an essential role of FXR poly(ADP-ribosyl)ation in CCl4-induced liver injury. Thus, our results identified poly(ADP-ribosyl)ation of FXR by PARP1 as a key step in oxidative-stress-induced hepatic cell death. The molecular association between PARP1 and FXR provides new insight into the mechanism, suggesting that inhibition of PARP1 could prevent liver injury.
INTRODUCTION
Farnesoid X receptors (FXRs; also known as the bile acid receptors [BAR]), the receptors for bile acids, were originally cloned as an orphan nuclear receptor (1). Two FXR genes are known and are referred to as the FXRα (NR1H4) and FXRβ (NR1H5) genes, although the FXRβ gene is a pseudogene in humans and other primates (2, 3). FXRα (now known as FXR) can be activated by a natural ligand or a synthetic agonist, such as GW4064 or chenodeoxycholic acid (CDCA). When activated, FXR forms heterodimers with retinoid X receptor α (RXRα; also designated NR2B) and then modulates the expression of FXR target genes via direct binding to specific response elements, called FXREs (FXR response elements), within their promoters (4). FXR has been reported as a metabolic regulator and cell protector in the liver (2, 5, 6). Recent research has shown that FXR activation suppresses liver injury by modulating the expression of FXR target genes (6–8).
The gene encoding the nuclear protein poly(ADP-ribose) polymerase 1 (PARP1) belongs to a family of 18 identified genes that transcribe poly(ADP-ribose) polymerases, enzymes that catalyze the covalent transfer of poly(ADP) units from NAD+ to receptors to affect their transcriptional activity (9–13). PARP1 can be activated by DNA strand breaks due to reactive oxygen species (ROS). Under oxidative stress, ROS, including superoxide and hydrogen peroxide (H2O2), are generated at high levels, leading to cellular damage and cell death (14). Increasingly, studies have shown that PARP1 plays a crucial role in the regulation of cell death by modulating various transcription factors (15).
Therefore, we investigated the effects of PARP1 on the transcriptional activation of FXR in oxidative-stress-induced liver cell death. Here we demonstrate that PARP1-mediated poly(ADP-ribosyl)ation acts as an important regulator in the activation of FXR signaling in liver cells. Since FXR is a protector of liver cells, we hypothesize that the modification of FXR by PARP1 participates in oxidative-stress-induced liver injury.
MATERIALS AND METHODS
Ethics statement.All animal work in this study was approved by the Ethics Committee of Tongji Medical College, Huazhong University of Science and Technology (HUST), and conformed to the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals (16).
Preparation of cell cultures.HepG2 cells were purchased from the American Type Culture Collection (ATCC) and were maintained in Eagle's minimal essential medium (MEM) supplemented with 10% fetal bovine serum (FBS) and 1% nonessential amino acids.
Primary mouse hepatocytes were isolated from C57BL/6 mice (obtained from Tongji Medical College, HUST, Wuhan, China) as described previously (17). Briefly, the livers of anesthetized mice were perfused first with freshly prepared Leffert's buffer (10 mM HEPES, 3 mM KCl, 130 mM NaCl, 1 mM NaH2PO4, 10 mM d-glucose [pH 7.4]) with 0.5 mM EGTA via the portal vein for 5 min and then with digestion buffer (Leffert's buffer plus 0.33 mg of collagenase I/ml and 0.0279% CaCl2) for 12 min. The perfused livers were removed from the mice and were dissociated in ice-cold wash buffer (Leffert's buffer plus 0.0279% CaCl2) by agitation. The hepatocytes were washed with ice-cold wash buffer 3 times by centrifugation, seeded onto Primaria tissue culture plates (BD Biosciences) at 400,000 to 600,000 cells per well in culture medium (RPMI 1640 plus 10% fetal calf serum [FCS], 100 nM dexamethasone, 10 μg/ml insulin, and 5 μg/ml transferrin), and cultured overnight at 37°C under 5% CO2.
The preparation of whole-cell extracts and nuclear extracts has been described previously (18, 19). The protein concentrations of these extracts were determined by the Bradford assay. The cell extracts obtained were stored at −80°C until use.
Cell death assay.Cell death experiments were performed as described previously (20). HepG2 cells and primary mouse hepatocytes were grown to confluence in 60-mm dishes, serum starved overnight, and treated as indicated in the figure legends. Cells were stained with trypan blue and were counted in a hemocytometer under a fluorescence microscope. Cells were counted in 4 fields, and the average cell number was calculated.
PARP1 activity assay.PARP1 activity was assayed using the universal colorimetric PARP assay kit (Trevigen), based on the incorporation of biotinylated ADP-ribose into histone proteins. Cell lysates containing 50 μg of protein were loaded into a 96-well plate coated with histones and biotinylated poly(ADP-ribose), allowed to incubate for 1 h, treated with horseradish peroxidase (HRP)-conjugated streptavidin, and read at 450 nm in a spectrophotometer.
RNA interference and transfection.Small interfering RNAs (siRNAs) were synthesized by RiboBio Co. Ltd. (China). The cultured cells were transfected in 6-well plates at 70% confluence. Transfection of siRNA was performed at a final concentration of 50 nM using Lipofectamine 2000 (Invitrogen). The sequence of PARP1 siRNA was 5′-GGAUGAUCUUCGACGUGGA-3′; the sequence of FXR siRNA was 5′-GAGGAUGCCUCAGGAAAUA-3′; and the sequence of unrelated siRNA was 5′-UUCUCCGAACGUGUCACGU-3′. The efficiency of PARP1 siRNA or FXR siRNA was determined by a real-time reverse transcription (RT)-PCR assay (see Fig. S2 in the supplemental material).
Plasmid construction.The full-length cDNA of human PARP1 was cloned by RT-PCR from HepG2 cells. Full-length PARP1 and fragments A (amino acids [aa] 1 to 214), B (aa 215 to 372), C (aa 373 to 476), D (aa 477 to 524), E (aa 525 to 656), and F (aa 657 to 1014) were constructed in the mammalian expression vector p3flag-CMV (Sigma-Aldrich). A catalytically inactive mutant of PARP1 (mut-PARP1) in which lysine 893 is replaced by isoleucine (K893I) was generated as described previously by using the QuikChange site-directed mutagenesis kit (Stratagene) (21). Human cDNAs encoding FXR were cloned by RT-PCR from HepG2 cells. The expression plasmids for enhanced green fluorescent protein (EGFP)-tagged FXR fragments A (aa 1 to 243), B (aa 1 to 186), C (aa 125 to 477), D (aa 187 to 477), and E (aa 244 to 477) or glutathione S-transferase (GST)-tagged FXR fragments A (aa 1 to 124), B (aa 125 to 186), C (aa 187 to 243), and D (aa 244 to 477) were constructed by inserting PCR fragments from FXR cDNA into the pEGTP-N1 (Clontech) and pGEX-4T-1 (Amersham Biosciences) vectors. Primers are listed in Tables S1, S2, and S3 in the supplemental material.
Real-time RT-RCR assays.Total RNA from cultured cells was isolated using TRIzol reagent (TaKaRa) according to the manufacturer's instructions. One microgram of total RNA was reverse transcribed using an RNA PCR kit (TaKaRa), and the resulting cDNA was used as a PCR template. The mRNA levels were determined by real-time PCR with an ABI Prism 7900 sequence detection system (Applied Biosystems) according to the manufacturer's instructions. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as the endogenous control. The PCR mixture contained SYBR green I (TaKaRa), cDNA, and the primers. The relative gene expression level (the amount of the target gene, normalized to the amount of the endogenous-control gene) was calculated using the comparative threshold cycle (CT) method formula 2−ΔΔCT. Primers are listed in Table S4 in the supplemental material.
Luciferase assays.The oligonucleotides used for the construction of p3FXRE-luc (with three copies of FXRE) and mutant p3FXRE-luc (mut-FXRE) were as follows: FXRE sense, 5′-GCCCTTAGGGACATTGATCCTTAGGCAA-3′; mut-FXRE sense, 5′-TTGCCTAAGAAATCAATGTTTCTAAGGGC-3′. The complementary oligonucleotides were inserted into the pRL-TK luciferase expression vector (Promega). Cells grown to 60 to 80% confluence were trypsinized and seeded in 24-well culture plates (1 × 105 cells/well). Cells were transiently transfected with 500 ng of plasmid p3FXRE-luc or p3FXRE-luc by use of Lipofectamine 2000 (Invitrogen) in accordance with the manufacturer's instructions. After incubation for 24 h, cells were harvested, lysed, and assayed for luciferase activity with the Dual-Luciferase reporter assay kit (Promega) according to the manufacturer's instructions.
Western blot assays.Western blot assays were performed as described previously (22). After denaturation and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), separated proteins were transferred to nitrocellulose membranes. Membranes were then blocked with 5% nonfat milk in TBST (50 mM Tris-HCl [pH 7.6], 150 mM NaCl, 0.2% Tween 20) for 3 h. After incubation with primary antibodies (anti-PARP1 [1:1,000; R&D], anti-PAR [1:1,000; Trevigen], anti-histone H1 [1:200; Santa Cruz], anti-β-actin [1:500; Santa Cruz], anti-GAPDH [1:500; Santa Cruz], and anti-FXR [1:1,000; R&D]) in an antibody diluent (Abcam) at 4°C overnight, membranes were washed 3 times with TBST and were incubated with a peroxidase-conjugated secondary antibody in an antibody diluent at room temperature for 2 h. Specific bands were detected by chemiluminescence assays (ECL detection reagents; Pierce) and were recorded on X-ray film. Bio-Rad's Quantity One software (version 4.4) was used for quantification.
In vitro protein-protein interaction assays (far-Western blot assays).Far-Western blot assays were performed by resolving protein samples by SDS-PAGE and transferring them to polyvinylidene difluoride membranes. Membranes were then incubated overnight at 4°C with Hyb-75 buffer (20 mM HEPES [pH 7.6], 75 mM KCl, 0.1 mM EDTA, 2.5 mM MgCl2, 0.005% Nonidet P-40, 1 mM dithiothreitol [DTT]) supplemented with 5% nonfat milk powder. Recombinant PARP1 protein (1 μg/ml; Trevigen) was incubated with NAD+, active DNA, and PARP buffer (from the PARP assay kit) at room temperature for 1 h to obtain autopoly(ADP-ribosyl)ated PARP1 protein (AP-PARP1). Membranes were washed briefly with Hyb-75 buffer and were then incubated with 1 μg/ml recombinant PARP1 protein (Trevigen), AP-PARP1 protein, recombinant FXR protein (Abnova), or recombinant β-actin (Abnova) at room temperature for 1 h. After three washes with Hyb-75 buffer, membranes were incubated with the indicated antibodies at 4°C overnight. After washing, membranes were incubated with HRP-conjugated secondary antibodies for 2 h. Specific bands were detected using the ECL detection system (Pierce).
In vitro poly(ADP-ribosyl)ation assay.GST-fused proteins immobilized on glutathione-Sepharose were incubated with purified PARP1 (Trevigen) in poly(ADP-ribosyl)ation assay buffer (50 mM Tris-HCl [pH 8.0], 10 mM MgCl2, 1 mM DTT, 200 mM NAD+, 3 ng/ml single-stranded DNA [ssDNA]) for 1 h at 37°C. Poly(ADP-ribosyl)ation of FXR was detected by a Western blot assay using an anti-PAR antibody.
Immunoprecipitation (IP) assays.Briefly, 500 μg of nuclear extracts was incubated with the indicated antibodies against PARP1, FXR, or PAR, or with nonspecific IgG, at 4°C for 1 h, and with protein G-agarose at 4°C for 12 h. The immunoprecipitates were pelleted by centrifugation at 5,000 × g for 1 min and were washed 4 times with lysis buffer. The pellets were suspended in SDS gel loading buffer, boiled for 10 min, and subjected to Western blot assays. Nonspecific IgG was used as a negative control.
Overexpression and purification of recombinant proteins.GST fusion proteins were expressed in HepG2 cells by induction with 0.1 mM isopropyl-β-d-thiogalactopyranoside at 25°C for 3.5 h. Bacteria were then lysed by sonication in lysis buffer (phosphate-buffered saline [PBS] containing 0.5 mM phenylmethylsulfonyl fluoride and 1 mM DTT) and were centrifuged. Soluble extracts were incubated with glutathione-Sepharose beads (GE Healthcare) for 1 h at 4°C before three washes in lysis buffer. The concentration of proteins immobilized on beads was quantitated by SDS-PAGE by comparison with a titration of bovine serum albumin (BSA; Sigma-Aldrich) after Coomassie blue staining.
EMSAs.DNA-protein interactions were detected using a LightShift chemiluminescent electrophoretic mobility shift assay (EMSA) kit (Thermo Scientific) according to the manufacturer's protocol. The sequences of FXRE consensus oligonucleotides were as follows: forward, 5′-TCAAGAGGTCATTGACCTTTTTG-3′; reverse, 5′-CAAAAAGGTCAATGACCTCTTGA-3′. The oligonucleotides were labeled with biotin at their 5′ ends. In brief, binding reactions with reaction mixtures containing 20 μg of nuclear extracts and 50 fM oligonucleotide were performed for 30 min in binding buffer [2.5% glycerol, 0.05% Nonidet P-40, 50 mM KCl, 5 mM MgCl2, 1 mM EDTA, 10 mM Tris (pH 7.6), and 50 ng of poly(dI-dC)]. The unlabeled probe, at a concentration 200-fold that of the labeled probe, was used as a competitor. Protein-DNA complexes were subjected to 6% native polyacrylamide gel electrophoresis and thereafter were transferred to nylon membranes (Pierce), which were immediately cross-linked on a UV transilluminator. Bands were then detected by the chemiluminescent method according to the manufacturer's protocol.
ChIP.Chromatin immunoprecipitation (ChIP) experiments were performed as described previously (23). HepG2 cells were sonicated, and the lysates were immunoprecipitated using an anti-FXR antibody (Santa Cruz). In re-ChIP assays, chromatin was first immunoprecipitated with an anti-FXR antibody, then eluted with 100 μl of elution buffer with 10 mM DTT at 37°C for 30 min, diluted (25-fold) with dilution buffer (20 mM Tris-HCl [pH 8.0], 150 mM NaCl, 2 mM EDTA, 1% Triton X-100), and finally reimmunoprecipitated with IgG or an antibody against PARP1, PAR, RNA polymerase II (RNA Pol II), CCCTC-binding factor (Ctcf), or RXRα (Santa Cruz). Real-time PCR was performed using 1 μg of template DNA with primers specific for the human bile salt export pump (BSEP) (sense primer, 5′-TTTCCCAAGCACACTCTGTGTTTGGGG-3′; antisense primer, 5′-GAGGAAGCCAGAGGAAATAATGGACTCC-3′). The input chromosomal DNA and ChIP DNA with nonspecific IgG were subjected to the same PCR amplification. PCR products were separated on an ethidium bromide-stained 2% agarose gel.
Animal models.All experimental procedures were approved by the Animal Care Committee of Peking University. FXR−/− mice (stock no. 007214), maintained on an inbred C57BL/6 background, were purchased from The Jackson Laboratory (Bar Harbor, ME, USA).
All mice were kept in a temperature-controlled room on a 12-h light/dark cycle, with 60% humidity, and with food and water ad libitum. The mice were acclimatized to laboratory conditions for 1 week before the study. After injection with a single dose of CCl4 (1 ml/kg of body weight), mice were either injected with PJ34 (20 mg/kg/day) or a vehicle (normal saline), or were gavaged with GW4064 (30 mg/kg), once a day for 3 days. On the 3rd day, all mice were anesthetized with pentobarbital (50 mg/kg; Sigma), and a serum sample was drawn from each individual mouse by retro-orbital puncture for serum biochemical analysis. Serum alanine aminotransferase (ALT) levels were measured at the Union Hospital. Livers were excised and were weighed immediately. The liver samples were used for subsequent experiments.
Histological analyses.Liver samples were fixed in 4% paraformaldehyde and were embedded in paraffin. Tissues were cut with a microtome into sequential 5-μm-thick slices and were stained with hematoxylin and eosin (H&E). Light microscope images were captured with a color video camera (Olympus BX51 microscope) and were analyzed with image analysis software (Qianping Imaging).
Statistical analysis.Values are means ± standard errors of the means (SEM) for at least three independent experiments. The significance of differences was estimated by the independent-samples t test, the Mann-Whitney-Wilcoxon test, one-way analysis of variance (ANOVA), or ANOVA on ranks followed by Student-Newman-Keuls multiple comparison tests. A P value of <0.05 was considered significant. All statistical analyses were performed with SPSS software (version 11.0; SPSS Inc.).
RESULTS
Inhibition of PARP1 prevented H2O2-induced liver cell death through the FXR pathway.ROS have been shown to allow calcium influx, glutathione depletion, mitochondrial dysfunction, caspase-3 activation, and subsequent necrotic and apoptotic cell death (24). In this study, cultured HepG2 cells were exposed to H2O2, the most stable form of ROS, for 12 h. An increase in intracellular ROS levels through H2O2 treatment shortened the life spans of cells in a dose-dependent manner (Fig. 1A). Like the DNA-damaging capacity of ROS, the enzymatic activity of PARP was dramatically increased in HepG2 cells treated with H2O2 (Fig. 1B). To investigate the role of PARP activation in H2O2-induced cell death, HepG2 cells were treated with the PARP inhibitor N-(6-oxo-5,6-dihydrophenanthridin-2-yl)-2-(N,N-dimethylamino)acetamide (PJ34) in the presence of H2O2. Cell death assays confirmed that PJ34 treatment significantly reduced the level of HepG2 cell death from that for the control group (Fig. 1C). Since PARP1 accounts for about 90% of total cellular PARP activity (9, 10), HepG2 cells were then transfected with PARP1 siRNA. The results showed that cell death was dramatically attenuated by PARP1 depletion (Fig. 1C). We then speculated that PARP1 activation participated in H2O2-induced cell death. To confirm our conjecture, we built the wild-type PARP1 (wt-PARP1) expression vector and the mutant PARP1 (mut-PARP1) plasmid expressing an enzymatically inactive PARP1 protein in which lysine 893 was replaced with isoleucine (K893I). The results showed that forced expression of wt-PARP1 in HepG2 cells aggravated H2O2-induced cell death, while mut-PARP1 did not affect it (Fig. 1D). These results indicated that inhibition of PARP1 activation prevented H2O2-induced HepG2 cell death. Similar results were obtained with primary mouse hepatocytes (see Fig. S1A through D in the supplemental material).
Oxidative stress-induced HepG2 cell death was attenuated by PARP1 inhibition. (A) HepG2 cells were exposed to H2O2 (3, 30, or 300 μM) for 12 h. Cells were trypsinized and were then counted in 3 fields by a hemocytometer with trypan blue dye (n = 3). Asterisks indicate significant differences (*, P < 0.05; **, P < 0.01) from the control group. (B) HepG2 cells were treated with H2O2 (3, 30, 300 or μM) for 12 h, and cellular PARP activity was assayed as described in Materials and Methods. Data are expressed as means ± SEM. Asterisks indicate significant differences (*, P < 0.05; **, P < 0.01) from the control group. (C) HepG2 cells were pretreated either with PJ34 (15 μM) for 24 h, with PARP1 siRNA (50 nM) for 48 h, or with GW4064 (1 μM) for 24 h and were then exposed to H2O2 (3, 30, or 300 μM) for 12 h. Cells were trypsinized and were counted in 3 fields by a hemocytometer with trypan blue dye (n = 3). Asterisks indicate significant differences (**, P < 0.01) from the control group. (D) HepG2 cells were transfected with either an empty vector (p3flag-CMV), wild type PARP1 (wt-PARP1), or the enzymatic mutant PARP1 (mut-PARP1) at 1 mg/liter for 48 h and were then exposed to H2O2 (3, 30, or 300 μM) for 12 h. Cells were trypsinized and were counted in 3 fields by a hemocytometer with trypan blue dye (n = 3). Asterisks indicate significant differences (*, P < 0.05) from the control group. (E) After transfection with FXR siRNA (50 nM) for 72 h, HepG2 cells were treated either with PJ34 (15 μM) for 24 h, with PARP1 siRNA (50 nM) for 48 h, or with GW4064 (1 μM) for 24 h and were then exposed to H2O2 (3, 30, or 300 μM) for 12 h. Cells were trypsinized and were counted in 3 fields by a hemocytometer with trypan blue dye (n = 3). Asterisks indicate significant differences (*, P < 0.05) from the unrelated siRNA group. The efficiencies of PARP1 siRNA, FXR siRNA, and wt- or mut-PARP1 transfection were determined by a real-time RT-PCR assay (shown in Fig. S2 in the supplemental material).
FXR plays an important role in protecting the liver from injury (25). Here we found that the FXR agonist GW4064 also prevented H2O2-induced HepG2 cell death (Fig. 1C). To explore the role of FXR in the protective effect of PARP inhibition against cell death, HepG2 cells were transfected with FXR siRNA. We observed that the protective effects of FXR agonists, PARP inhibitors, or PARP1 siRNA against HepG2 cell death were abrogated in FXR-knockdown cells (Fig. 1E). Similar results were observed for primary mouse hepatocytes (see Fig. S1E in the supplemental material), indicating that FXR mediated the protective effects of PARP1 inhibition against liver cell death.
Inhibition of PARP1 enhanced FXR-dependent gene transcription.FXR plays a crucial role in liver repair by regulating its target genes (6–8). Real-time RT-PCR assays were performed in order to investigate the role of PARP1 in the expression of FXR target genes involved in liver repair, including genes expressing the bile salt export pump (BSEP), fibroblast growth factor 19 (FGF19), Forkhead box M1b (Foxm1b), and small heterodimer partner (SHP), in HepG2 cells. The results showed that administration of the FXR agonist GW4064 or the PARP inhibitor PJ34 increased the mRNA expression of these genes under basal conditions (Fig. 2A and B). We also found that H2O2 treatment resulted in decreased expression of these genes, while treatment with GW4064 or PJ34 reversed the inhibitory effect of H2O2 (Fig. 2A and B). Similar transcriptional enhancement of FXR target gene expression was observed when PARP1 was knocked down by PARP1 siRNA (Fig. 2C). These results suggested that inhibition of PARP1 enhanced FXR-dependent gene expression. In line with this notion, forced expression of wt-PARP1 in HepG2 cells significantly reduced the expression of FXR-dependent genes, while transfection with the mut-PARP1 vector did not influence their expression (Fig. 2D).
Inhibition of PARP1 promoted the transcription of FXR-dependent hepatoprotective genes. (A and B) Real-time RT-PCR assays of BSEP, FGF19, Foxm1b, and SHP in HepG2 cells. After treatment with 1 μM GW4064 (A) or 15 μM PJ34 (B) for 12 h, cells were treated or not treated with H2O2 (300 μM; 12 h). Data are expressed as means ± SEM. Significant differences from the control group (**, P < 0.01) or from the H2O2 group (#, P < 0.05) are indicated. (C) Real-time RT-PCR assays of BSEP, FGF19, Foxm1b, and SHP in HepG2 cells transfected with either 50 nM PARP1 siRNA or 50 nM unrelated siRNA for 48 h, with or without H2O2 treatment (300 μM; 12 h). Data are expressed as means ± SEM. Significant differences from the control group (*, P < 0.05; **, P < 0.01) or from the H2O2 group (#, P < 0.05) are indicated. (D) Real-time RT-PCR assays of BSEP, FGF19, Foxm1b, and SHP in HepG2 cells transfected with either an empty vector (p3flag-CMV), wt-PARP1, or mut-PARP1 at 1 mg/liter for 48 h. Data are expressed as means ± SEM. Asterisks indicate significant differences (*, P < 0.05) from the control group. (E and F) Relative FXRE- or mutant (mut-FXRE)-driven luciferase reporter activity in HepG2 cells treated with 1 μM GW4064 (E) or 15 μM PJ34 (F) for 24 h, with or without H2O2 treatment (300 μM; 12 h). Data are expressed as means ± SEM. Significant differences from the control group (**, P < 0.01) or from the H2O2 group (#, P < 0.05) are indicated. (G) Relative FXRE- or mut-FXRE-driven luciferase reporter activity in HepG2 cells transfected with 50 nM PARP1 siRNA or 50 nM unrelated siRNA for 48 h, with or without H2O2 treatment (300 μM; 12 h). Data are expressed as means ± SEM. Significant differences from the control group (**, P < 0.01) or from the H2O2 group (#, P < 0.05) are indicated. (H) Relative FXRE- or mut-FXRE-driven luciferase reporter activity in HepG2 cells treated with an empty vector (p3flag-CMV), wt-PARP1, or mut-PARP1 at 1 mg/liter for 48 h. Data are expressed as means ± SEM. Asterisks indicate significant differences (**, P < 0.01) from the control group.
In nuclei, FXR binds directly to FXRE in the promoters of target genes to regulate their transcription. To explore the influence of PARP1 on FXR-dependent transcriptional responses, luciferase assays with the wild-type FXRE-driven luciferase reporter and the mutant FXRE construct (mut-FXRE)-driven luciferase reporter in HepG2 cells were performed. The results showed that treatment with the FXR agonist GW4064, or inhibition of PARP1 by PJ34 or PARP1 siRNA promoted the transcriptional output of the FXRE-driven luciferase reporter but not that of the mut-FXRE-driven luciferase reporter (Fig. 2E, F, and G). We also found that H2O2 treatment resulted in decreased luciferase activity of the FXRE-driven reporter, while treatment with GW4064, PJ34, or PARP1 siRNA reversed the inhibitory effect of H2O2 (Fig. 2E, F, and G). In addition, HepG2 cells transfected with the mut-PARP1 vector showed a level of FXRE-driven luciferase activity much higher than that for cells transfected with the wt-PARP1 vector (Fig. 2H). These results indicated that inhibition of PARP1 activation promoted FXR-dependent transcriptional activation.
PARP1 bound to FXR.The finding that PARP1 regulates FXR-mediated transcriptional activation led us to explore the interaction between PARP1 and FXR in detail. We screened for a nuclear protein interacting with PARP1 by far-Western blot assays using PARP1 protein as a probe. Both unpoly(ADP-ribosyl)ated recombinant human PARP1 (UP-PARP1) and autopoly(ADP-ribosyl)ated recombinant human PARP1 (AP-PARP1) were found to bind specifically to a 55-kDa protein in nuclear extracts from HepG2 cells (Fig. 3A). Since FXR is a 55-kDa protein, we speculated that this 55-kDa nuclear protein might be FXR. Thereafter, siRNA-mediated knockdown of FXR effectively attenuated the binding of PARP1 to the 55-kDa nuclear protein (Fig. 3A). To further determine if PARP1 and FXR were in the same nuclear complex, we performed coimmunoprecipitation (co-IP) assays. Nuclear extracts from HepG2 cells were subjected to IP with an antibody specific for FXR or PARP1. Western blot assays revealed that FXR was coprecipitated with PARP1, and vice versa (Fig. 3B and C). Similar results were obtained for primary hepatocytes (see Fig. S3A and B in the supplemental material). In a cell-free system, far-Western blot assays also showed that recombinant FXR protein (56 kDa) could bind directly to either UP-PARP1 or AP-PARP1 (Fig. 3D).
PARP1 bound directly to FXR. (A) Far-Western blot assays of nuclear extracts from HepG2 cells treated with an unrelated siRNA or FXR siRNA. Unpoly(ADP-ribosyl)ated PARP1 (UP-PARP1), autopoly(ADP-ribosyl)ated PARP1 (AP-PARP1), or β-actin protein (negative control) was used as a probe. Histone H1 (his1) served as a loading control (bottom panels). (B) Coimmunoprecipitation assays of FXR-bound proteins from HepG2 cells, followed by Western blot assays using an anti-PARP1 antibody. Nonspecific IgG served as a negative control. (C) Coimmunoprecipitation assays of PARP1-bound proteins or poly(ADP-ribosyl)ated proteins from HepG2 cells, followed by Western blot assays using an anti-FXR antibody. Nonspecific IgG served as a negative control. (D) Far-Western blot assays of recombinant FXR protein. UP-PARP1 or AP-PARP1 was used as a probe. β-Actin protein served as a negative control. (E) Diagram of Flag-tagged human PARP1 with its domains: DNA-binding domain (DBD), nuclear localization signal (NLS), BRCA1 C terminus (BRCT)/automodification domain (AMD), and catalytic domain (CD). Fragments A to F with their amino acid coordinates are listed. HepG2 cells were transfected with EGFP-tagged full-length FXR and Flag-tagged PARP1 mutants. Coimmunoprecipitation assays demonstrated the specific binding of FXR to the BRCT/AMD of PARP1. (F) Diagram of EGFP-tagged human FXR with its domains. LBD, ligand-binding domain. Fragments A to E with their amino acid coordinates are listed. HepG2 cells were transfected with Flag-tagged full-length PARP1 and EGFP-tagged FXR mutants. Coimmunoprecipitation assays demonstrated the specific binding of PARP1 to the LBD of FXR.
To define which domain of PARP1 mediates protein-protein interaction with FXR, we made a series of PARP1 deletion mutants. We then observed that exogenous FXR bound exclusively to the BRCA1 C terminus (BRCT)/automodification domain (AMD) (aa 477 to 524) of PARP1 (Fig. 3E). By using FXR deletion mutants, we confirmed that PARP1 interacted with the region of FXR (aa 244 to 477) containing the C-terminal ligand-binding domain (LBD) (Fig. 3F). Thus, we concluded that the LBD of FXR associated directly with the central BRCT/AMD of PARP1.
FXR could be poly(ADP-ribosyl)ated by PARP1.PARP1 is known to be a poly(ADP-ribosyl)ating enzyme (11). We therefore sought to determine whether FXR is poly(ADP-ribosyl)ated by PARP1. Nuclear extracts of HepG2 cells were subjected to IP with an antibody specific for the poly(ADP-ribose) polymer (PAR). Western blot assays with an anti-FXR antibody revealed that FXR could be poly(ADP-ribosyl)ated (Fig. 3C). An IP assay with an anti-FXR antibody followed by Western blot assays with an anti-PAR antibody demonstrated that treatment with H2O2 significantly increased the poly(ADP-ribosyl)ation level of FXR, while the stimulatory effect of H2O2 was reversed by treatment with the PARP inhibitor PJ34 or PARP1 siRNA (Fig. 4A and B). In a cell-free system, incubation of recombinant FXR protein with recombinant PARP1 protein, NAD+, and active DNA resulted in strong poly(ADP-ribosyl)ation of FXR (Fig. 4C), which was inhibited by the addition of 3AB (Fig. 4C). These results indicate that FXR is a bona fide substrate for PARP1-dependent poly(ADP-ribosyl)ation.
PARP1 poly(ADP-ribosyl)ated the LBD of FXR. (A) Nuclear extracts from HepG2 cells were subjected to an immunoprecipitation assay with an anti-FXR antibody, followed by a Western blot assay using an anti-PAR antibody. Cells were treated with 15 μM PJ34 for 24 h in the presence or absence of H2O2 (300 μM; 12 h). (B) Nuclear extracts from HepG2 cells were subjected to an immunoprecipitation assay with an anti-FXR antibody, followed by a Western blot assay using an anti-PAR antibody. Cells were transfected with PARP1 siRNA or unrelated siRNA at 50 nM for 48 h with or without H2O2 treatment (300 μM; 12 h). (C) Recombinant FXR proteins were incubated either with a vehicle (PBS), with PARP1, NAD+, and active DNA, or with PARP1, NAD+, active DNA, and 3AB, as indicated. Western blot assays were used to detect the poly(ADP-ribosyl)ation levels of FXR. (D) (Top) Diagram of GST-tagged human FXR with its domains. (Center) Purified GST-FXR fragments are shown after Coomassie staining. (Bottom) Bacterially expressed GST-FXR deletion mutants were incubated with recombinant PARP1 protein in the presence of DNA and NAD+. Poly(ADP-ribosyl)ation of GST-FXR mutants was detected by a Western blot assay with an anti-PAR antibody.
To further determine which region of FXR is poly(ADP-ribosyl)ated by PARP1, we built constructs encoding different functional domains of the FXR. Four GST fusion proteins were incubated with recombinant PARP1 protein, NAD+, and active DNA in a cell-free system. A Western blot assay demonstrated that the amino acid residues in the LBD of FXR represent acceptor sites for poly(ADP-ribosyl)ation by PARP1 (Fig. 4D).
Poly(ADP-ribosyl)ation prevented the formation of the FXR-promoter complex.When activated by ligands, FXR bound to FXRE present in the promoters of target genes (2). To explore the influence of poly(ADP-ribosyl)ation on the the DNA binding activity of FXR, electrophoretic mobility shift assays (EMSAs) were performed with an oligonucleotide probe containing FXRE. In HepG2 cells, H2O2 treatment dramatically inhibited the formation of the FXR-FXRE complex (Fig. 5A and B; see also Fig. S4A in the supplemental material). Conversely, treatment with the PARP inhibitor PJ34 or knockdown of PARP1 by siRNA enhanced complex formation under either basal or H2O2-treated conditions (Fig. 5A and B), suggesting that inhibition of poly(ADP-ribosyl)ation promoted the binding of FXR to DNA. Thereby, the direct effect of poly(ADP-ribosyl)ation on the formation of the FXR-FXRE complex was investigated. Incubation of nuclear extracts from untreated HepG2 cells with NAD+ and active DNA inhibited the formation of the FXR-FXRE complex in a NAD+ dose-dependent manner (Fig. 5C). In a cell-free system, we also observed similar results (see Fig. S4B in the supplemental material). All these results established that poly(ADP-ribosyl)ation prevented the binding of FXR to FXRE.
Poly(ADP-ribosyl)ation inhibited the binding of FXR to FXRE in the target promoter. (A) HepG2 cells were treated with 15 μM PJ34 for 24 h with or without H2O2 treatment (300 μM; 12 h). Binding of FXR to FXRE was detected by EMSA. (B) HepG2 cells were transfected with either PARP1 siRNA or an unrelated siRNA at 50 nM for 48 h, with or without H2O2 treatment (300 μM; 12 h). Binding of FXR to FXRE was detected by EMSA. (C) Nuclear extracts from untreated HepG2 cells were incubated with active DNA and NAD+ (1, 10, or 100 μM) and were then subjected to EMSA. (D and E) ChIP-PCR assays using an anti-FXR antibody for amplification of BSEP promoters in HepG2 cells treated with 15 μM PJ34 for 24 h (D) or transfected with PARP1 siRNA or an unrelated siRNA at 50 nM for 48 h (E), with or without H2O2 treatment (300 μM; 12 h). Data are expressed as means ± SEM. Significant differences from the control group (**, P < 0.01) or from the H2O2 group (#, P < 0.05) are indicated. (F) In re-ChIP assays, chromatin was first immunoprecipitated with an anti-FXR or anti-Ctcf (positive-control) antibody and was then reimmunoprecipitated with an anti-PAR antibody, an anti-PARP1 antibody, IgG, or an anti-RNA Pol II antibody. IgG served as a negative control. (G) Nuclear extracts from HepG2 cells were incubated with an anti-FXR, anti-PARP1, or anti-PAR antibody or with nonspecific IgG (negative control) and were then subjected to EMSA. (H) Recombinant FXR proteins were incubated with recombinant PARP1 and/or RXRα as indicated and were then subjected to EMSA.
Then we further examined the influence of poly(ADP-ribosyl)ation on the recruitment of FXR to the promoter of BSEP, a target gene of FXR, by chromatin immunoprecipitation (ChIP) assays. Inhibition of PARP1 by PJ34 or by PARP1 siRNA assisted the recruitment of FXR to the BSEP promoter in HepG2 cells that were either left untreated or treated with H2O2 (Fig. 5D and E), suggesting that poly(ADP-ribosyl)ation inhibited the recruitment of FXR to its target promoter. To test this conjecture, re-ChIP assays with an anti-FXR antibody, in which chromatin was reprecipitated using an anti-PAR antibody, were performed. The results showed that poly(ADP-ribosyl)ated FXR was undetectable in the BSEP promoter (Fig. 5F). In line with this result, a supershift assay indicated that incubation of nuclear extracts from HepG2 cells with an anti-PAR antibody failed to abrogate the band of the FXR-FXRE complex (Fig. 5G). We thus concluded that poly(ADP-ribosyl)ation inhibited the formation of the FXR-promoter complex.
Previous studies have shown that PARP1 modulates the transactivation of several transcription factors through physical interaction. To investigate whether the physical interaction with PARP1 influenced the formation of the FXR-FXRE complex, a supershift assay was performed. The results showed that incubation of nuclear extracts from HepG2 cells with an anti-PARP1 antibody failed to abrogate the band of the FXR-FXRE complex (Fig. 5G). Moreover, re-ChIP assays with an anti-FXR antibody, in which chromatin was reprecipitated using an anti-PARP1 antibody showed that the FXR-PARP1 complex could not be detected in the BSEP promoter (Fig. 5F). In a cell-free experiment, incubation of FXR protein with PARP1 protein did not affect the formation of the FXR-FXRE complex (Fig. 5H). These results indicated that the physical interaction with PARP1 did not influence FXR transactivation.
Ligand induced FXR transactivation was mediated through the inhibition of poly(ADP-ribosyl)ation.To further determine whether poly(ADP-ribosyl)ation was involved in ligand-induced FXR transactivation, IP assays with anti-FXR antibody, followed by Western blotting with an anti-PAR antibody, were performed. The results showed that treatment of HepG2 cells with the FXR agonist GW4064 significantly reduced the amount of poly(ADP-ribosyl)ated FXR under either basal conditions or H2O2-treated conditions (Fig. 6A and B), suggesting that the FXR ligand inhibited FXR poly(ADP-ribosyl)ation. However, a Western blot assay revealed that the protein levels of PARP1 in HepG2 cells remained unchanged under GW4064 treatment (Fig. 6C). These results suggested that poly(ADP-ribosyl)ation, but not PARP1 protein, might mediate ligand-induced FXR activation. To confirm our conjecture, EMSAs and ChIP assays were performed. EMSAs showed that treatment with the FXR ligand GW4064 or CDCA, which inhibited FXR poly(ADP-ribosyl)ation, restored FXR-FXRE complex formation in HepG2 cells transfected with the wt-PARP1 vector (Fig. 6D). ChIP assays showed that the inhibitory effect of wt-PARP1 overexpression on the recruitment of FXR to its target promoter was reversed by FXR ligand treatment (Fig. 6E). All these data indicated that inhibition of FXR poly(ADP-ribosyl)ation promoted ligand-induced FXR transactivation.
Ligand-induced FXR transactivation was mediated by the inhibition of poly(ADP-ribosyl)ation. (A) Nuclear extracts from HepG2 cells were subjected to an immunoprecipitation assay with an anti-FXR antibody, followed by a Western blot assay using an anti-PAR antibody. Cells were treated with GW4064 (0.5, 1, or 2 μM) or a vehicle (dimethyl sulfoxide) for 24 h. (B) Nuclear extracts from HepG2 cells were subjected to an immunoprecipitation assay with an anti-FXR antibody, followed by a Western blot assay using an anti-PAR antibody. Cells were treated with 1 μM GW4064 for 24 h, with or without H2O2 (300 μM; 12 h). (C) The protein expression of PARP1 or FXR was determined by a Western blot assay. HepG2 cells were treated with GW4064 (0.5, 1, or 2 μM) or a vehicle (dimethyl sulfoxide) for 24 h. (D) EMSAs were used to detect the FXR-FXRE complex in HepG2 cells. After treatment with an empty vector (p3flag-CMV), wt-PARP1, or mut-PARP1 at 1 mg/liter for 24 h, cells were treated with nicotinic acid (50 μM; 24 h), GW4064 (1 μM; 24 h), or CDCA (50 μM; 24 h) as indicated. Nicotinic acid served as a negative control. (E) ChIP-PCR assays using an anti-FXR antibody for the amplification of BSEP promoters. HepG2 cells were treated with an empty vector (p3flag-CMV), wt-PARP1, or mut-PARP1 at 1 mg/liter for 48 h in the absence or presence of nicotinic acid (50 μM; 24 h), GW4064 (1 μM; 24 h), or CDCA (50 μM; 24 h) as indicated. Data are expressed as means ± SEM. Significant differences from the control group (*, P < 0.05) and from the wt-PARP1 transfection group (#, P < 0.05) are indicated.
Poly(ADP-ribosyl)ation of FXR mediated CCl4-induced liver injury.CCl4 has been used extensively to study liver injury induced by ROS in the mouse model. In this study, wild-type (WT) and FXR−/− mice were injected with CCl4 in corn oil or with the same volume of corn oil alone as a vehicle control. Histological analysis of liver sections with hematoxylin and eosin (H&E) staining was performed to assess injury to the liver. The results showed that FXR−/− mice exhibited dramatically lower numbers of hepatocytes than WT mice after CCl4 treatment (Fig. 7A). Of interest, the FXR agonist GW4064 or the PARP inhibitor PJ34 significantly inhibited CCl4-induced liver injury in WT mice but not in FXR−/− mice (Fig. 7A). Consistently, 3 days after CCl4 injection, FXR−/− mice had significantly higher levels of serum alanine aminotransferase (ALT) than WT mice. Both GW4064 and PJ34 dramatically suppressed CCl4-induced increases in the ALT levels of WT mice but not those for FXR−/− mice (Fig. 7B). These results indicated that inhibition of poly(ADP-ribosyl)ation reduced CCl4-induced liver injury through the FXR pathway. To determine whether poly(ADP-ribosyl)ation of FXR mediated CCl4-induced liver injury, an IP assay with an anti-FXR antibody, followed by a Western blot assay with an anti-PAR antibody, was performed. The results showed that CCl4 promoted poly(ADP-ribosyl)ation of FXR and that the FXR agonist GW4064 or the PARP inhibitor PJ34 decreased the elevated levels of FXR poly(ADP-ribosyl)ation in the livers of CCl4-treated WT mice (Fig. 7C). To further investigate the underlying mechanism, nuclear extracts of liver tissues were subjected to EMSAs. The results revealed that treatment with CCl4 inhibited the binding of FXR to FXRE in WT mice but not in FXR−/− mice (Fig. 7D). Administration of GW4064 or PJ34 reversed the inhibitory effect of CCl4 on FXR binding to FXRE (Fig. 7D). In agreement with these results, both GW4064 and PJ34 promoted the expression of FXR-dependent genes, including BSEP, FGF19, and SHP, in CCl4-treated WT mice but not in FXR−/− mice (Fig. 7E). All these results identified an essential role of FXR poly(ADP-ribosyl)ation in CCl4-induced liver injury.
Poly(ADP-ribosyl)ation of FXR mediated CCl4-induced liver injury. After injection with a single dose of CCl4 (1 ml/kg), WT or FXR−/− mice either were injected with PJ34 (20 mg/kg/day) or a vehicle (normal saline) or were gavaged with GW4064 (30 mg/kg) once a day for 3 days. (A) Representative liver sections from WT or FXR−/− mice stained with H&E (original magnification, ×40). (B) Serum ALT levels in WT or FXR−/− mice. Significant differences from the vehicle (WT) group (**, P < 0.01) and from the CCl4 (WT) group (#, P < 0.05) are indicated. (C) Nuclear extracts of liver tissues from WT mice were subjected to immunoprecipitation assays with an anti-FXR antibody, followed by Western blotting with an anti-PAR antibody. (D) Nuclear extracts of liver tissues from WT or FXR−/− mice were subjected to EMSA to detect the formation of the FXR-FXRE complex. (E) A real-time RT-PCR assay was used to detect the mRNA expression of BSEP, FGF19, and SHP in liver tissues from WT or FXR−/− mice. Significant differences from the vehicle (WT) group (**, P < 0.01) and from the CCl4 (WT) group (#, P < 0.05) are shown.
DISCUSSION
PARP1 is a redox-sensitive nuclear enzyme. PARP1 inhibition has been reported to be beneficial in maintaining plaque stability, neuroprotection, and cell homeostasis (26–28). In this study, treatment with a PARP inhibitor or PARP1 siRNA prevented oxidative-stress-induced liver cell death. Moreover, the protective effects were abrogated in FXR-depleted cells. These results indicated that PARP1 participated in cell death by modulating the FXR signaling pathway.
It has been reported that FXR is an established regulator of the apoptosis and growth of liver cells. FXR promotes liver repair by regulating the expression of genes involved in cell cycle progression (25). In this study, we observed that H2O2, a well-established PARP1 activator, inhibited FXR-dependent gene expression. Inhibition of PARP1 using either pharmacological inhibition (PJ34) or siRNA knockdown (PARP1 siRNA) promoted the transcription of FXR target genes under basal or H2O2 treatment conditions. Moreover, the stimulatory effects of PJ34 or PARP1 siRNA were abrogated in FXR-depleted cells. We then concluded that inhibition of PARP1 prevented liver cell injury by promoting the transcription of FXR-dependent genes.
PARP1 has been reported to interact with a variety of transcription factors and thereby to alter their transcriptional activity (9, 18, 19, 29, 30). In this study, we demonstrated that FXR could bind to PARP1 in the absence of any other cellular cofactors or DNA, implying that FXR interacts directly with PARP1. Thus, it was postulated that PARP1 might affect the transactivation of FXR through physical interaction with FXR. However, supershift and re-ChIP assays proved that neither AP-PARP1 nor UP-PARP1 was an intrinsic component of the FXR-FXRE complex. Given that PARP1 is known as a poly(ADP-ribosyl)transferase, we then speculated that the regulatory effects of PARP1 on the transactivation of FXR might be based on the enzymatic activity of PARP1.
Poly(ADP-ribosyl)ation catalyzed by PARP1 suggests important roles for PARP1 in the regulatory function of the receptor proteins. PARP1 can form definitive structures through intramolecular interactions and can change the structure of receptor proteins (11). In this study, the LBD of FXR could be poly(ADP-ribosyl)ated by PARP1 in HepG2 cells. The LBD has been reported to be responsible for FXR activation and signal transduction (31). Given that poly(ADP-ribosyl)ated amino acid residues might change the spatial conformation of the LBD in FXR, we conjectured that poly(ADP-ribosyl)ation might affect the activation of FXR. In line with this notion, oxidative-stress-induced PARP1 activation increased the amount of poly(ADP-ribosyl)ated FXR and thus impeded its transactivation. Moreover, treatment with PARP1 siRNA or a PARP inhibitor reversed the inhibitory effect of H2O2 or CCl4 on the transcriptional activity of FXR. Meanwhile, GW4064 administration inhibited the oxidative-stress-induced poly(ADP-ribosyl)ation of FXR and thus promoted the ligand-induced activation of FXR. Collectively, our data demonstrated that poly(ADP-ribosyl)ation of FXR by PARP1 determined the amount of direct binding of the FXR complex to DNA and thus defined critical thresholds of transcriptional activation and duration.
Therefore, our models have proved that FXR poly(ADP-ribosyl)ation is one of the steps in the termination of FXR-mediated transcription, in addition to other known FXR posttranslational modifications, such as acetylation, phosphorylation, ubiquitination, and sumoylation (32–35). According to our study, the model of the mediation of FXR signaling by PARP1 in ROS-induced liver cell death is as follows. Under oxidative stress, activated PARP1 poly(ADP-ribosyl)ates FXR. FXR undergoes a conformational change in which polymer chains block the “docking” site in the LBD, dislodge the coactivator complexes, and then inhibit the activity of the FXR signaling pathway to aggravate liver cell death. Inhibition of PARP1 or treatment with a FXR agonist, which abrogates the formation of ADP-ribose polymer chains, promotes the formation of the FXR-FXRE complex to stimulate FXR-dependent gene transcription, thus resulting in liver repair.
In summary, the data reveal a novel mechanism underlying FXR transactivation in oxidative-stress-induced liver injury. PARP1 can bind to and poly(ADP-ribosyl)ate FXR. Poly(ADP-ribosyl)ation of FXR by activated PARP1 dissociates the FXR-FXRE complex and inhibits the transcription of FXR-dependent hepatoprotective genes, resulting in liver injury. We believe that PARP1 inhibition will improve hepatoprotective function by activating FXR signaling. Therefore, PARP1 might be a valid therapeutic target for liver diseases.
ACKNOWLEDGMENTS
This work was supported by the National Natural Science Foundation of China (grants 30971245 and 81170239 to Kai Huang and 81000112 to Dan Huang).
FOOTNOTES
- Received 6 February 2013.
- Returned for modification 7 March 2013.
- Accepted 5 September 2013.
- Accepted manuscript posted online 16 September 2013.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/MCB.00160-13.
- Copyright © 2013, American Society for Microbiology. All Rights Reserved.
