Bile Acids Up-regulate Death Receptor 5/TRAIL-receptor 2 Expression via a c-Jun N-terminal Kinase-dependent Pathway Involving Sp1*

Bile acids up-regulate death receptor 5 (DR5)/TRAIL-receptor 2 (TRAIL-R2) expression thereby sensitizing hepatocytes to TRAIL-mediated apoptosis. However, the precise mechanism by which bile acids enhance DR5/TRAIL-R2 expression is unknown. Although several bile acids enhanced DR5/TRAIL-R2 expression, deoxycholic acid (DCA) was the most potent. DCA stimulated JNK activation and the JNK inhibitor SP600125 blocked DCA-induced DR5/TRAIL-R2 mRNA and protein expression. Reporter gene analysis identified a 5′-flanking region containing two Sp1 binding sites within the DR5/TRAIL-R2 promoter as bile acid responsive. Sp1 binding to one of the two sites was enhanced by DCA treatment as evaluated by electrophoretic mobility shift assays and chromatin immunoprecipitation studies. JNK inhibition with SP600125 also blocked binding of Sp1 to the DR5/TRAIL-R2 promoter. Finally, point mutations of the Sp1 binding site attenuated promoter activity. In conclusion, Sp1 is a bile acid-responsive transcription factor that mediates DR5/TRAIL-R2 gene expression downstream of JNK.

In virtually all human liver diseases, hepatocytes undergo cell death by apoptosis (1,2). This is also true for cholestatic liver diseases, pathophysiologic syndromes characterized by impaired hepatocellular secretion of bile acids into bile. In cholestasis, the intracellular accumulation of toxic bile acids within hepatocytes promotes cellular injury and the subsequent development of hepatic cirrhosis and liver failure (3). Numerous studies have shown that bile acids mediate their cytotoxicity by inducing hepatocellular apoptosis (4 -6).
Bile acid-triggered apoptosis involves death receptor activation. Consistent with their known effects on gene expression, bile acids stimulate transcription of the death receptor 5/tumor necrosis factor-related, apoptosis-inducing ligand-receptor 2 (DR5/TRAIL-R2) expression and aggregation, promoting a death receptor-dependent apoptosis (10). The importance of bile acid-mediated DR5/TRAIL-R2 induction has been amply demonstrated by in vitro and in vivo observations (10,24). DR5/TRAIL-R2 is a highly inducible receptor whose expression is regulated by both p53-dependent and -independent pathways (10,(25)(26)(27)(28). In previous studies examining bile acidmediated DR5/TRAIL-R2 expression and cell death, a cell line with a defective p53 mutant was utilized, thereby, excluding a role for p53 in bile acid-DR5/TRAIL-R2 expression (10). An FXR agonist also failed to enhance expression of this death receptor. The observations make it likely that bile acid enhances DR5/TRAIL-R2 expression by a MAPK signaling pathway.
The overall objective of this study was to test the hypothesis that bile acid enhances DR5/TRAIL-R2 transcription by a MAPKdependent process. To address this hypothesis, the following two questions were formulated: (i) Do MAPK inhibitors block bile acid-mediated DR5/TRAIL-R2 transcription? (ii) What is the transcription factor responsible for bile acid-associated DR5/TRAIL-R2 expression? The results indicate that DCAmediated JNK1/2 activation contributes to DR5/TRAIL-R2 expression. The DCA-stimulated JNK1/2 pathway was found to be associated with the specificity protein 1 (Sp1) transcription factor. Thus, Sp1 appears to be a bile acid-responsive transcription factor that contributes to DR5/TRAIL-R2 gene expression downstream of JNK1/2.
Immunoblot Analysis-Cells were lysed by incubation on ice for 30 min in lysis buffer containing 20 mM Tris-HCl (pH 7.5), 1% Triton X-100, 150 mM NaCl, 10% glycerol, 1 mM Na 3 VO 4 , 50 mM NaF, 100 mM phenylmethylsulfonyl fluoride, and a commercial protease inhibitor mixture (Complete Protease Inhibitor Mixture; Roche Applied Science, Mannheim, Germany). After insoluble debris was pelleted by centrifugation at 14,000 ϫ g for 15 min at 4°C, the supernatants were collected. Samples were resolved by 12.5% SDS-PAGE, transferred to nitrocellulose membrane, and blotted with appropriate primary antibodies at dilution of 1:1,000. Horseradish peroxidase-conjugated secondary antibodies (BIOSOURCE International, Camarillo, CA) were incubated at a dilution of 1:2,000 to 1:10,000. Bound antibody was visualized using a chemiluminescent substrate (ECL; Amersham Biosciences, Arlington Heights, IL) and exposed to Kodak X-Omat film.
Quantitative Real-time PCR-Total RNA was isolated from HuH-BAT cells using TRIzol reagent (Invitrogen, Carlsbad, CA), and cDNA was prepared using an random primer and murine leukemia virus reverse transcriptase as previously described in detail (10). The cDNA product was amplified by PCR with Taq DNA polymerase using standard protocols (10). PCR primers for human TRAIL-R2/DR5 were as follows: forward 5Ј-TGC AGC CGT AGT CTT GAT TG-3Ј and reverse 5Ј-GCA CCA AGT CTG CAA AGT CA-3Ј. For an internal control, primers for 18 S ribosomal RNA were purchased from Ambion Inc. (Austin, TX). After electrophoresis in 1% low melting temperature agarose gel, the expected base pair PCR products were identified, and the PCR products were eluted into Tris-HCl using a DNA elution kit (Qiagen, Valencia, CA). Extracted PCR products were prepared as standards at the concentrations of 10 9 , 10 8 , 10 7 , 10 6 , 10 5 , 10 4 , and 10 3 copies/l. The identity of the PCR products was verified using dye terminator technology. Real-time PCR was performed with the Roche LightCycler using SYBR green as the fluorophore (Molecular Probes) (10). The linear relationship between copy number and cycle number was then determined. For the quantitation of each mRNA, real-time PCR was performed using the cDNA samples obtained from the HuH-BAT cells. The standard curve was used to calculate the copy number in the experimental sample. The copy number of TRAIL-R2/DR5 mRNA in each sample was normalized using the copy number of 18 S RNA. All PCR conditions and primers were optimized to produce a single product of the correct base pair size.
Reporter Plasmid Construction-The plasmid pGL2-DR5/TRAIL-R2-Full containing 1.6 kb upstream of the ATG site through intron 2 in human DR5/TRAIL-R2 genomic DNA has been described previously, and was a kind gift from El-Deiry and co-workers (University of Pennsylvania) (10,26). The firefly luciferase-based reporter plasmids, pGL3basic and pGL3-promoter, containing 195 bp of the SV40 minimal early promoter (pGL3-SV), were obtained from Promega. The 5Ј-flanking lesion (corresponding nucleotide position at Ϫ1220/Ϫ2, numbers from the ATG site) and the downstream lesion containing exon 1 through intron 1 (nucleotide position at Ϫ1/ϩ416) were amplified by PCR using primers: forward 5Ј-GTC CAT GCT AGC AGC TTC ACT CCT GAG CCA GT-3Ј/reverse 5Ј-GCA CTC AGA TCT GCG GTA GGG AAC GCT CTT ATA G-3Ј for Ϫ1220/Ϫ2 lesion, and 5Ј-AC GTC CAT GCT AGC CAT GGA ACA ACG GGG ACA G-3Ј and 5Ј-AC GCA CTC AGA TCT CGT GCT TCA CGC AGC TTA CT-3Ј for Ϫ1/ϩ416 lesion. The PCR fragments were purified by 1% agarose gel electrophoresis and gel extraction and digested with NheI and BglII. The Ϫ1223/Ϫ2 fragment was subcloned into the pGL3-basic vector between NheI and BglII sites to make pGL3-5Ј-1223. The Ϫ1/ϩ416 fragment was subcloned into the pGL3-basic vector or pGL3-SV vector to make pGL3-Int1ϩ416 and pGL3-SV-Int1ϩ416, respectively. Deletion mutants of the pGL3-5Ј-1220 were generated by using an Erase-Base system (Promega) according to the manufacturer's instruction. Point mutations were introduced to the pGL3-5Ј-1220 by PCR-based site-directed mutagenesis technology using the primers: 5Ј-GGA TCT GAT TCG CCA AGT TCC GAA TGA CGC C-3Ј and 5Ј-GGC GTC ATT CGG AAC TTG GCG AAT CAG ATC C-3Ј for Sp1 binding site 1; and 5Ј-GAA AGT ACA GCC GCG AAG TTC CAA GTC AGC CTG-3Ј and 5Ј-CAG GCT GAC TTG GAA CTT CGC GGC TGT ACT TTC-3Ј for Sp1 binding site 2. The sequences of all constructs were confirmed by nucleotide-sequencing analysis.
Reporter Gene Assay-HuH-BAT cells cultured in 24-well plates were co-transfected with 5 ng of TK-Renilla-CMV and 250 ng of firefly luciferase-based reporter plasmids described above. Twelve hours after the transfection, cells were incubated with bile acids or media (control) for 8 h, and then cell lysates were prepared as previously described (10,30). Both firefly and Renilla luciferase activities were quantitated using the dual luciferase reporter assay system (Promega) according to the manufacturer's instructions, and firefly luciferase activity was normalized using Renilla luciferase activity.
Electrophoretic Mobility Shift Assay-Nuclear protein extracts were prepared from the HuH-BAT cells using high salt extraction as described previously (30). The buffer composition of the final nuclear extract was 20 mM HEPES, pH 7.9, 20% glycerol, 140 mM NaCl, 16 mM KCl, 0.2 mM EDTA, 0.5 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, and 1ϫ concentration of the protease inhibitor mix (Roche Applied Science).
Single-stranded complimentary synthetic oligonucleotides were annealed and end-labeled with [␥-32 P]ATP (3000 Ci/mmol) and T4 polynucleotide kinase. Six microliters of nuclear extract containing up to 6 g of protein was incubated on ice in binding buffer (100 mM HEPES, 300 mM KCl, 20% Ficoll, 0.02% Nonidet P-40, 0.5 mg/ml bovine serum albumin, and 0.2 mg/ml poly(dI-dC)).C. O. The binding reaction was initiated by addition of 20,000 cpm of labeled oligonucleotide probe and continued for 15 min at room temperature. Antibodies for supershifting or competing double-stranded oligonucleotides were added to the binding buffer 15 min prior to the addition of labeled oligonucleotides. Samples were loaded onto 4% or 5% polyacrylamide gel containing 0.5ϫ Tris borate/EDTA. The gel was dried and exposed to BioMax-MR films at Ϫ70°C.
Chromatin Immunoprecipitation-ChIP was performed using the ChIP assay kit (Upstate Cell Signaling Solutions, Waltham, MA) according to the manufacturer's instruction. Briefly, cells were incubated with formaldehyde (final concentration, 1%) at room temperature for 20 min, to obtain a cross-linking of DNA and DNA-bound proteins. After washing the cells twice with phosphate-buffered saline, cells were collected and lysed within 200 l of the SDS lysis buffer provided by the kit. The cell extracts were sonicated to shear the DNA into 200-to 1,000-bp fragments, and centrifuged at 13,000 rpm for 10 min. The DNA concentration in the supernatant was quantitated by measuring absorbance at 260 nm. Samples containing 100 ng of DNA were diluted with the ChIP dilution buffer obtained from the kit, to make a final volume of 1,050 l. 50 l of each sample was used for non-immunoprecipitation control experiments (e.g. samples not treated with antisera).
The remaining samples (1 ml) were subjected to the immunoprecipitation by adding rabbit-polyclonal anti-Sp1 antibody (final concentration, 5 g/ml) and 60 l of salmon sperm DNA/Protein A-agarose-50% slurry. After washing the agarose beads according to the manufacturer's instructions, the protein A-agarose⅐immune complexes were eluted with the lysis buffer (1% SDS, 0.1 M NaHCO 3 ). The elution was repeated three times, and 750 l (total volume) of elutes were obtained from each vial. 50 l of each elute was stored for a Sp1 immunoblot analysis to determine the immunoprecipitation efficiency. 20 l of 5 M NaCl was added to the remaining samples, and the samples were incubated at 65C°for 5 h to reverse cross-linking. After the samples were digested with proteinase K, DNA fragments were purified by the phenol/chloroform extraction and ethanol precipitation methods. The DNA pellet was resolved in 20 l of water. The PCR primers were designated to amplify 215-bp fragments containing both Sp1 BS1 and Sp1 BS2. The sequences of the primers were as follows: 5Ј-AGG TTA GTT CCG GTC CCT TC-3Ј, forward; 5Ј-CGC GTG CTG ATT TAT GTG TC-3Ј, reverse. The copy number for precipitated DR5/TRAIL-R2 promoter fragments was quantitated using real-time PCR technique, and the copy number was normalized by the copy number of the DR5/TRAIL-R2 promoter fragments within parallel samples handled identically except for omission of the immunoprecipitation step.
Statistical Analysis-All data represent at least three independent experiments and are expressed as the mean Ϯ S.D. unless otherwise indicated. Differences between groups were compared using a one-way analysis of variance and a post hoc Bonferroni test.

RESULTS
Bile Acids Up-regulate DR5/TRAIL-R2 Expression-Initially, we screened several bile acids for their effects on DR5/ TRAIL-R2 induction to ascertain which bile acids were the most potent in augmenting expression of this death receptor. Among the bile acids tested (100 M), deoxycholic acid (DCA), chenodeoxycholic acid (CDCA), and the glycine-conjugated form of deoxycholic acid (GDCA) were the most effective in up-regulating DR5/TRAIL-R2 protein (Fig. 1A) and mRNA expression (Fig. 1B). Interestingly, the hydrophilic, cytoprotective bile acid ursodeoxycholic acid (UDCA) did not significantly enhance DR5/TRAIL-R2 expression. In contrast to their effects on DR5/TRAIL-R2 expression, bile acids altered neither DR4, DcR1, nor DcR2 protein expression. Thus, bile acid-mediated TRAIL-related receptor up-regulation is DR5/TRAIL-R2-specific. Because DCA was the most potent in inducing DR5/ TRAIL-R2 expression, we choose this bile acid for the remainder of our studies.
Next, the concentration-dependent response for DR5/ TRAIL-R2 up-regulation by DCA was examined. DCA up-regulates DR5/TRAIL-R2 mRNA and protein level in a concentration-dependent manner with concentrations Ն 150 M sufficient to induce maximum DR5/TRAIL-R2 expression (Fig.  1, C and D). In contrast, DCA, at all concentrations tested, did not alter expression of the related DR4/TRAIL-R1 receptor. Thus, DCA selectively up-regulates DR5/TRAIL-R2 in a concentration-dependent manner.
Bile Acid-mediated DR5/TRAIL-R2 Induction Is JNKdependent-Several mitogen-activated protein kinases (MAPKs) are activated by DCA (5,18,31) and can modulate gene expression. Therefore, we formulated the hypothesis that bile acidactivated MAPK activity was responsible for DR5/TRAIL-R2 gene induction by DCA. To test this hypothesis, several different kinase inhibitors, including the p42/44 MAPK inhibitor PD098059, the p38 MAPK inhibitor SB203580, and the JNK1/2 inhibitor SP600125 were screened for the effects on DR5/ TRAIL-R2 protein expression. Among the inhibitors tested, the JNK1/2 inhibitor SP600125 completely abrogated DCA- HuH-BAT cells were cultured in the presence or absence of the indicated bile acids for 12 h. Protein extracts were prepared for immunoblot analysis, and RNA was extracted for real-time PCR experiments as described under "Experimental Procedures." Immunoblot analysis was performed to assess protein expression for the TRAIL receptors (A and C), and real-time PCR was performed to quantitate DR5/TRAIL-R2 mRNA expression (B and D). For the immunoblot analysis, actin was probed to assure consistent loading between lanes. DR5/ TRAIL-R2 mRNA level (copies/l) was normalized to 18 S ribosomal RNA copy numbers. DCA, deoxycholic acid; GDCA, glycodeoxycholic acid; TDCA, taurodeoxycholic acid; CDCA, chenodeoxycholic acid; GCDCA, glycochenodeoxycholic acid; TCDCA, taurochenodeoxycholic acid; and UDCA, ursodeoxycholic acid. induced DR5/TRAIL-R2 protein expression, whereas the p42/44 and p38 MAPK inhibitors failed to block DR5/TRAIL-R2 induction ( Fig. 2A). The JNK1/2 inhibitor SP600125 also inhibited DCA-induced DR5/TRAIL-R2 mRNA induction as evaluated by real-time PCR (Fig. 2B). Finally, we confirmed DCA-mediated JNK1/2 activation in HuH-BAT cells as demonstrated by an increase in phospho-JNK1/2 (Fig. 2C). These data provide a link between bile acid-induced signaling cascade and DR5/TRAIL-R2 gene induction.
DCA Enhances DR5/TRAIL-R2 Promoter Activity via the 5Ј-Flanking Region-Current data indicate that DR5/TRAIL-R2 gene transcription is regulated by the transcription factors p53 (25,26) and NF-B (32,33). These specific transcription factors bind to specific DNA sequences located within the DR5/ TRAIL-R2 gene locus at the intronic promoter region (26,32). In addition, the importance of the 5Ј-flanking promoter region of DR5/TRAIL-R2 gene in regulating its transcriptional activity has also been suggested by other studies (10, 34). Therefore, we first determined whether the 5Ј-flanking or the intronic promoter regions were responsible for bile acid-mediated DR5/ TRAIL-R2 gene transcription. As shown in Fig. 3A, we compared several different reporter constructs: pGL2-DR5/TRAIL-R2-full, which contains 1.6 kb upstream of the ATG site through intron 2 in the human DR5/TRAIL-R2 genomic locus; pGL3-5Ј-1220, which contains 1220 bp of the 5Ј-flanking region; pGL3-Int1ϩ416, which contains exon 1 and part of intron 1, including both p53 and NF-B binding sites; and pGL3-Int1-SV, which contains the same sequences of pGL3-Int1 upstream of SV40 minimal early promoter. DCA strongly enhanced the promoter activity of pGL3-5Ј-1220 containing the 5Ј-flanking region (12-fold elevation). In contrast, minimal activation by DCA was observed with the pGL2-DR5/TRAIL-R2-full, pGL3-Int1ϩ416, or pGL3-Int1ϩ416-SV constructs containing the intronic promoter region with the p53 and NF-B binding sites. These results demonstrate that DCA enhancement of DR5/ TRAIL-R2 promoter activity is dependent upon its 5Ј-flanking region. The fact that DCA-mediated DR5/TRAIL-R2 induction did not require the p53 and NF-B binding regions is supported by additional information. First, p53 is mutated and not functional in this cell line. Second, although bile acids may activate NF-B, adenovirus-mediated transduction with the I-B superrepressor failed to attenuate DCA-mediated DR5/TRAIL-R2 induction (data not shown). This information further strengthens the relationship between bile acid-mediated DR5/ TRAIL-R2 induction and its 5Ј-flanking region.
DCA Enhances Sp1⅐DNA Binding to the DR5/TRAIL-R2 Promoter-Next, we further evaluated the bile acid-responsive region in the DR5/TRAIL-R2 promoter by a luciferase reporter gene analysis using various deletion mutants of the pGL3-5Ј-1220 plasmid. Partial deletion of the DR5/TRAIL-R2 promoter 5Ј-flanking region upstream of Ϫ531 did not reduce DCA-mediated luciferase activity. In contrast 5Ј-deletion at Ϫ240 significantly reduced luciferase activity (Fig. 4A), suggesting that the nucleotide region downstream of Ϫ531 contains bile acid response elements. By more detailed deletion analysis of this region, we observed that deletions of the Ϫ509/Ϫ437 and Ϫ243/ Ϫ137 regions markedly reduced DR5/TRAIL-R2 promoter activity (Fig. 4B).
Next we performed an electrophoretic mobility shift assay using the probes containing the above base pair regions for the 5Ј-flanking region of the DR5/TRAIL-R2 open reading frame. Fig. 5A illustrates the location of potential transcription factor binding sites for the DR5/TRAIL-R2 promoter, including the Ϫ509/Ϫ437 and Ϫ243/Ϫ137 regions. The Ϫ509/Ϫ437-bp region contains two c-Myb binding sites, and the Ϫ243/Ϫ137 region contains two Sp1 binding sites as previously described (34). In addition, we noted that both the Ϫ509/Ϫ437 and Ϫ243/Ϫ137 regions contain TGACG sequences, which are the core binding elements for the JNK1/2-inducible transcription factors AP-1 and/or CREB/ATF (the computed score for these regions using the TRANSFAC MatInspector search were as follows: Ϫ449/ Ϫ428: core similarity 1.000 and matrix similarity 0.831 for AP-1, core similarity 1.000 and matrix similarity 0.918 for CREB/ATF; Ϫ191/Ϫ170: core similarity 1.000 and matrix sim- ilarity 0.885 for AP-1, core similarity 1.000 and matrix similarity 0.906 for CREB/ATF). Therefore, we performed EMSA experiments using ␥-32 P-labeled oligonucleotide DNA probes or unlabeled competitor oligonucleotide-DNAs as shown in Fig.  5A. When the region Ϫ510/Ϫ411 was used as a probe, we found that there were several DNA⅐protein complexes identified, however, the formation of these DNA⅐protein complexes was not enhanced by DCA treatment (Fig. 5B). In contrast, when the Ϫ235/Ϫ136 region was used as a probe, we observed three specific DNA⅐protein complexes (indicated as C1, C2, and C3 in Fig. 5B), with complex C2 formation dramatically enhanced by DCA treatment. Thus, the Ϫ235/Ϫ136 region appears to contain the DCA-responsive transcription factor binding sites.
Because the Ϫ235/Ϫ136 region contains two Sp1 binding sites (locate at Ϫ198/Ϫ189 and Ϫ152/Ϫ143), we next determined if the bile acid-modulated DNA⅐protein complexes were  Sp1-related (Fig. 5, C and D). The formation of DNA⅐protein complexes C1, C2, and C3 was completely abolished by 10-, 50-, and 100-fold amounts of competitor oligonucleotide-DNA containing Sp1 consensus binding sequences (Sp1 cons., Fig. 5C,  left). Consistently, incubation with another competitor Sp1 BS1 (containing Sp1 binding site 1 at Ϫ198/Ϫ189) also abrogated the formation of protein⅐DNA binding complexes (Fig. 5C,  right). Interestingly, the Sp1 BS2 oligonucleotide (corresponding to the Sp1 binding site 2 at Ϫ152/Ϫ143 lesion) was less effective in these competition studies as compared with the Sp1 BS1 oligonucleotide. These results suggest that the bile acidresponsive DNA⅐protein binding is likely Sp1-related. Among the two Sp1 binding sites on the DR5/TRAIL-R2 promoter Ϫ235/Ϫ136 region, BS1 appears to be the only one which is bile acid-responsive.
To further determine if Sp1 binds to the region, a supershift analysis was performed using a Sp1-specific antisera (Fig. 5D). Indeed, the density of the C2 band dramatically decreased in the presence of anti-Sp1 antibody, and two additional bands (indicated as *1 and *2 in Fig. 5D) were observed. In contrast, the density of the other complexes (C1 and C3) was not altered by anti-Sp1 antisera. The additional band at the lower position (*2) appears to be a small molecular weight supershifted band, whereas the higher band (*1) likely represents the formation of larger molecular weight complexes because it increases in relationship to the anti-Sp1 antibody concentration. Taken together, DCA enhances C2 complex formation through the Sp1 binding site at Ϫ198/Ϫ189 (BS1).
Finally, as a control, we also tested the contribution of the known JNK1/2-regulated transcription factors, AP-1 and CREB/ATF, in the formation of these DNA⅐protein complexes. However, neither addition of the AP-1 nor the CREB/ATF consensus sequences (both 100-fold the amount of the probe) altered DNA⅐protein complex formation (Fig. 5E). Thus, neither the AP-1 nor the CREB/ATF transcription factor appears to bind to the Ϫ235/Ϫ136 region of the DR5/TRAIL-R2 promoter.
DCA-mediated Sp1⅐DNA Binding Is JNK-dependent-Because bile acid-mediated DR5/TRAIL-R2 induction is JNK1/ 2-dependent, we next tested if bile acid-mediated Sp1⅐DNA binding was JNK1/2-dependent. HuH-BAT cells were incu-bated with DCA in the presence or absence of SP600125, and the nuclear protein extracts from those cells were subjected to EMSA. JNK1/2 inhibition abrogated DCA-mediated augmentation of DNA⅐protein complexes C1, C2, and C3 on the Ϫ235/ Ϫ136 DNA probe (Fig. 6A). The DNA⅐protein complex C2 was dramatically reduced by the anti-Sp1 supershifting antibody. Therefore, bile acid-mediated Sp1 binding to the DR5/ TRAIL-R2 promoter is JNK1/2-dependent, likely by an indirect mechanism.
We next compared Sp1⅐DNA binding affinity between the Sp1 consensus sequence, Sp1 binding site 1 and Sp1 binding site 2 (Fig. 6B). Indeed, when the Sp1 consensus sequence was used as a probe (Fig. 6B, left), DCA dramatically enhanced Sp1⅐DNA binding as compared with control nuclear extract. More importantly, the JNK1/2 inhibitor SP600125 dramatically reduced formation of Sp1⅐DNA complex. The additional bands observed below the Sp1 bands were likely Sp3⅐DNA complexes, because these bands were eliminated when an anti-Sp3-supershifting antibody was added (data not shown). Therefore, Sp3 binds to the Sp1 consensus sequence. DCA also appears to enhance Sp3⅐DNA binding by a JNK1/2-dependent mechanism. Similar results were observed when Sp1 BS1 sequence was used as a probe (Fig. 6B, center), however, the binding affinity of Sp1 to the Sp1 BS 2 probe was significantly less than its binding to the Sp1 BS1 site (Fig. 6B, right). These results are consistent with the competition studies in that Sp1 BS1 sequence more effectively competed with the Ϫ235/Ϫ136 probe than the Sp1 BS2 sequences.
To further determine if Sp1 binds to the endogenous DR5/ TRAIL-R2 promoter, we performed chromatin immunoprecipitation experiments (Fig. 7, A and B). Indeed, Sp1-DR5/ TRAIL-R2 promoter binding was dramatically enhanced by DCA treatment. The samples from DCA-treated HuH-BAT cells contained 9.50 Ϯ 3.71 ϫ 10 3 copies/l of DR5/TRAIL-R2 promoter fragments, whereas 1.86 Ϯ 1.30 ϫ 10 3 copies/l of the DNA fragments were co-precipitated from control cells. The number of co-precipitated DR5/TRAIL-R2 promoter fragments were significantly lower in cells treated with the JNK1/2 inhibitor SP600125 alone and SP600125 plus DCA (0.74 Ϯ 0.31 ϫ 10 3 copies/l and 0.52 Ϯ 3.71 ϫ 10 3 copies/l, respectively). Although minimal nonspecific binding between DNA frag-FIG. 5-continued ments and the agarose beads was detected by the real-time PCR (Ͻ10 2 copies/l), this was negligible by agarose gel electrophoresis. The Sp1 immunoprecipitation efficiency was equivalent in all immunoprecipitated samples as demonstrated by immunoblot analysis. The enhanced binding of Sp1 to the DR5/TRAIL-R2 promoter region after bile acid treatment cannot be explained by increased Sp1 protein expression as Sp1 protein expression was not enhanced by bile acids (Fig. 7C). Thus, DCA enhances endogenous Sp1-DR5/TRAIL-R2 promoter binding by a JNK1/2-dependent mechanism without alterations in Sp1 expression. HuH-BAT cells were incubated with or without DCA (100 M, 6 h) in the presence or absence of SP600125 (20 M). A, cells were fixed, lysed, sonicated, and subjected to Sp1 immunoprecipitation as described under "Experimental Procedures." One microliter of the immunoprecipitated samples and control samples not exposed to the antisera was subjected to real-time PCR to quantitate DR5/TRAIL-R2 sequences. The copy number of each sample was normalized to the DR5/TRAIL-R2 promoter copy number of samples not incubated with the antisera. Data were expressed as mean Ϯ S.D. from three independent experiments. p Ͻ 0.01 for DCA-treated group versus either Control, SP alone, or DCA plus SP. B, PCR products from the ChIP experiments were analyzed by 1% agarose gel electrophoresis (top and middle panels). Immunoblot analysis was performed using anti-Sp1 antisera (bottom panel) to demonstrate equivalent immunoprecipitation efficiency. C, immunoblot analysis was performed after treatment of HuH-BAT with deoxycholate (100 M) to analyze Sp1 cellular protein levels.

Mutation of DR5/TRAIL-R2 Promoter at Sp1 Binding Sites
Abrogates Transcriptional Activity-To further examine the function Sp1 binding in regulating DR5/TRAIL-R2 promoter activity, we introduced point mutations within the Sp1 binding sites of the DR5/TRAIL-R2 promoter. DCA-mediated DR5/ TRAIL-R2 promoter activation was abrogated when the Sp1 BS1 was mutated (Fig. 8). In contrast, mutation of the Sp1 BS2 was less effective in attenuating luciferase activity. Thus, the Sp1 BS1 appears to be crucial for DCA-mediated DR5/ TRAIL-R2 promoter activation. Taken together, DCA up-regulates DR5/TRAIL-R2 promoter activity by enhancing Sp1 binding to the DR5/TRAIL-R2 promoter at the Sp1 BS1.

DISCUSSION
The principal findings of this study relate to the molecular mechanisms by which bile acids up-regulate DR5/TRAIL-R2 expression. The results demonstrate that: (i) a JNK1/2 inhibitor SP600125 effectively suppresses DCA-mediated DR5/ TRAIL-R2 mRNA and protein up-regulation; (ii) the transcription factor Sp1 binds to the 5Ј-flanking region of the DR5/ TRAIL-R2 promoter/enhancer; (iii) DCA enhances Sp1⅐DNA binding activity to specific genomic sequences by a JNK1/2-dependent mechanism; and (iv) mutation of the Sp1 binding sequences abrogates DCA-mediated enhancement of DR5/ TRAIL-R2 promoter activity. These results suggest bile acidmediated cytotoxicity is, in part, mediated by a JNK1/2 signal transduction pathway resulting in Sp1 associated DR5/ TRAIL-R2 up-regulation.
There are three genes for JNK, each with specific functions. JNK1 and -2 are ubiquitously expressed, whereas JNK3 is neuronal specific (35,36). Bile acid activation of JNK1/2 has previously been implicated in gene regulation. For example, taurocholate has been shown to suppress cholesterol 7␣ hydroxylase by a JNK1/2-dependent process (37,38). JNK1/2 has also been implicated in the regulation of the death ligands Fas ligand and tumor necrosis factor-␣ (39 -42). The present study suggests JNK1/2 can up-regulate death receptors, namely DR5/ TRAIL-R2. Whether JNK1 or -2 mediates bile acid-induced DR5/TRAIL-R2 expression is unclear and will require further examination. However, given the recent observation that JNK1 is cytotoxic in the liver (35,43), toxic bile acids may up-regulate DR5/TRAIL-R2 potentially by a JNK1-mediated signaling cascade.
Several bile acid response elements have been identified in a variety of genes (44). In the current study, we found that the DR5/TRAIL-R2 promoter also contains bile acid response elements. By reporter gene analysis we initially identified two regions of the DR5-TRAIL-R2 promoter, Ϫ235/Ϫ136 and Ϫ510/ Ϫ411, that appeared to be important in the regulation of DR5/ TRAIL promoter activity by DCA. However, when used as probes in electrophoretic mobility gel shift assays, only probes from the Ϫ235/Ϫ136 region displayed evidence for bile acidmediated DNA⅐protein complex formation. Although probes from the Ϫ510/Ϫ411 region displayed protein⅐DNA complexes, the complex formation was not enhanced following bile acid treatment. The mechanism by which this region potentially contributes to DR5/TRAIL-R2 promoter activity therefore remains unknown and is a subject for future investigations.
We further examined the region Ϫ235/Ϫ136. This 5Ј-flanking sequence contains two putative Sp1 binding sites. The first putative binding site bound Sp1 more avidly than the second site, a finding consistent with their nucleotide sequences. The first binding region contains the complete Sp1 consensus binding sequence, whereas one guanine is replaced with a cytosine in the second sequence. The first sequence was also the only site responsive to DCA in a reporter gene assay. Consistently, reporter gene analysis revealed that mutations in the first site completely abrogated DR5/TRAIL-R2 promoter activity in response to the bile acid. Taken together, bile acids appear to activate DR5/TRAIL-R2 gene transcription by facilitating Sp1-DR5/TRAIL-R2 promoter binding at the first Sp1 binding site in the Ϫ235/Ϫ136 5Ј-flanking region of this death receptor gene.
Sp1 is the first identified member of Sp/Krü ppel-like family proteins that binds to G-rich elements such as GC-box (GGGGCGGGG) and GT-box (GGTGTGGGG) (45). Although Sp1 was initially identified as a basal transcription factor supporting transcription of so-called housekeeping genes (46), it is now well established that Sp1 is involved in inducible gene expression (45,(47)(48)(49). Sp1 participation in inducible gene expression occurs by post-translational modifications of this transcription factor (45,46,50,51). Currently, several protein kinases, including MAPKs, have been implicated to modulate Sp1 transcription factor activity (45,52). In our present study, DCA enhanced Sp1⅐DNA binding activity, and a JNK1/2 inhibitor SP600125 blocked DCA-mediated Sp1 activation. The results indicate that Sp1 is a bile acid-regulated transcription factor, and JNK1/2 is an intermediary kinase in this relationship. These data provide insight into the p53-independent mechanisms regulating DR5/TRAIL-R2 expression. Given the potential importance of this death receptor in cancer therapy (53), these data maybe important in developing strategies to enhance DR5/TRAIL-R2 expression in other cell types.
The precise mechanism by which JNK1/2 regulates Sp1 function was not elucidated in these studies. Because Sp1 contains Ser/Thr-rich regions and is a known phosphoprotein (45), it is possible that JNK1/2 may directly phosphorylate Sp1, thereby, altering its DNA binding activity. On the other hand, Sp1⅐DNA binding activity may be regulated by interaction with other co-activating proteins. For instance, binding between Sp1 and Egr-1 is regulated by Egr-1 phosphorylation, and Egr-1 phosphorylation results in an increase in Sp1⅐DNA binding (54). Other studies have also implicated yet to be identified cointeracting proteins in the regulation of Sp1 transcriptional activity (45,55,56). AP-1, a classic downstream target of JNK1/2 (35), has also been implicated in a cooperative role with Sp1 in gene induction (57)(58)(59)(60). Indeed, in DCA-treated HuH-BAT cells, AP-1 activation was observed by EMSA (not shown). Thus, it is likely that AP-1 contributes to DCA-mediated DR5/ TRAIL-R2 induction by co-operating with Sp1.
In summary, these data suggest bile acids sensitize HuH-BAT cells to TRAIL-mediated apoptosis by up-regulating DR5/ TRAIL-R2 expression via a JNK1/2/Sp1-dependent cascade. Given the number of genes regulated by bile acids, this paradigm likely pertains to many other bile acid-regulated genes (44). The results are also germane to cholestatic liver injury and cancer therapy (2). In cholestasis, elevated hepatic bile acid concentrations increase DR5/TRAIL-R2 expression and sensitize the liver to TRAIL cytotoxicity (24). Because JNK1/2 appears to be responsible for this up-regulation of DR5/TRAIL-R2, selective JNK1/2 inhibitors could potentially prove useful for treating human cholestatic liver diseases. Conversely, tumor-selective activation of JNK1/2 and/or Sp1 may also be an attractive approach for enhancing TRAIL-mediated cancer therapy, especially because bile acid derivatives have already been developed for pre-clinical use as anti-neoplastic agents (61,62).