Alcohol Induces RNA Polymerase III-dependent Transcription through c-Jun by Co-regulating TATA-binding Protein (TBP) and Brf1 Expression^*

Plasmids, Reagents, and Animals

HepG2-ADH and HepG2-vector cells were kindly provided by Dr. D. L. Clemens (14). Ethanol was from Sigma-Aldrich; cell culture medium (minimum Eagle's medium) and DMEM were obtained from Cellgro. Zeocin, RT-PCR kit, Lipofectin reagent, Lipofectamine 2000, and Opti-MEM were from Invitrogen. All antibodies were obtained from Santa Cruz Biotechnology with the exception of phospho-JNK and phospho Elk-1 antibodies that were from Cell Signaling. Mice expressing the HCV NS5A gene under control of the apoE promoter were obtained from Ratna Ray (St. Louis University). Wild-type and NS5A transgenic mice on the C57BL/6 strain were fed with Lieber-DeCarli diet containing 3.5% ethanol or isocaloric dextrin for long term alcohol feeding (15). All animal experiments were performed with age- and sex-matched mice from same littermates and conducted in accordance with the approved Institutional Animal Care and Use Committee protocol at the University of Southern California.

RT-qPCR and Transient Transfection Assays

Total RNA was isolated from engineered HepG2 cells, primary mouse hepatocytes, or liver tissues of mice fed alcohol using the single-step extraction method TRIzol reagent (Invitrogen). Precursor tRNA^Leu 5S rRNA transcripts were measured as described previously (16).

For transient transfection assays, cells were transfected with DNA or siRNAs as described previously (11). Serum-free medium was added to each dish with Lipofectin-DNA or Lipofectamine 2000-siRNA complexes, and cells were further incubated at 37 °C for 4 h. The medium was changed, and cells were incubated overnight before harvesting. Protein concentrations of the resultant lysates were measured by the Bradford method, and luciferase activity was measured using a Promega luciferase assay system. Resultant luciferase activities were normalized to the amount of protein in each lysate. The -fold change in promoter activity was calculated by measuring the level of luciferase activity with empty vector and without ethanol treatment and setting this value at 1 for each independent experiment. Differences in TBP promoter activity are expressed as the means ± S.E. At least three independent experiments were conducted for each determination.

Immunoblot Analysis

Immunoblot analysis was carried out as described previously (7). Cells were grown to 80% confluency in 10% fetal bovine serum (FBS) in DMEM and then serum-deprived using 0.1% FBS in DMEM for 3 h. Where noted, cells were pretreated with 5 μm of the JNK inhibitor, SP600125, for 1 h and then incubated with or without 50 mm ethanol for 60 min. Lysates (50 μg of protein) were subjected to immunoblot analysis. Membranes were probed with specific antibodies as indicated. Hybond-P membrane was used for protein transfer. Bound primary antibody was visualized using horseradish peroxidase-conjugated secondary antibody (Vector Laboratories) and enhanced chemiluminescence reagents (Amersham Biosciences).

Alcohol Induces RNA pol III-dependent Transcription via the Activation of JNK1

To determine whether alcohol regulates RNA pol III-dependent transcription, we used the hepatocellular carcinoma cell lines HepG2, stably expressing ADH (HepG2-ADH), or vector alone (HepG2-vector). Previous analysis of these cells revealed that although the parental HepG2 cells do not exhibit detectable ADH activity, the HepG2-ADH cells effectively metabolize ethanol to acetaldehyde (14). These cells were treated with ethanol, and the amounts of precursor tRNA^Leu, 5S rRNA, and 7SL RNA transcripts were measured by RT-qPCR. Ethanol treatment resulted in a concentration-dependent increase in transcription where the maximum response for ethanol-mediated induction was observed at the ethanol concentration of 50 mm after 60 min of treatment (data not shown). Therefore, all experiments were performed using these conditions. This effect was dependent on the metabolism of ethanol as transcription induction occurred in cells that expressed the ADH gene but not the vector control cells (Fig. 1A). Primary mouse hepatocytes treated with ethanol also displayed an increase in transcription from all three genes (Fig. 1B). These results indicate that RNA pol III-dependent gene transcription is induced by ethanol through its ability to be metabolized.

Given that the JNKs play an important role in regulating RNA pol III transcription (17), we next examined the role of JNK activation in alcohol-mediated transcription induction. HepG2-ADH cells were treated with ethanol, and immunoblot analysis was performed to measure the amount of phosphorylated JNK1 and JNK2 (Fig. 2A). Ethanol induced a strong activation in JNK1 but a more modest activation of JNK2. No activation of these JNKs was observed in the HepG2-vector cells. Pretreatment of the HepG2-ADH cells with the JNK inhibitor, SP600125, or transfection with siRNA specific for JNK1 to reduce JNK1 expression (Fig. 2C) abrogated the ethanol-mediated increase in tRNA and 5S rRNA gene transcription (Fig. 2B). These results support the idea that alcohol preferentially activates JNK1 and that alcohol-mediated induction of RNA pol III transcription requires the activation of JNK1.

Enhanced Expression of Both TBP and Brf1 Mediates Alcohol Induction of RNA pol III-dependent Transcription

To determine the mechanism by which alcohol enhances RNA pol III transcription, we examined potential changes in the expression of TBP and its associated factor, Brf1. Ethanol treatment induced an increase in both TBP and Brf1 expression but not that of the TFIIIC₆₃ subunit in HepG2-ADH cells (Fig. 2C). These increases also required the activation of JNK1 as treatment with a JNK inhibitor or decreased expression of JNK1 prevented ethanol-mediated increases in TBP and Brf1 expression.

We next assessed whether the alcohol-mediated increase in TBP and/or Brf1 expression were required for enhanced RNA pol III transcription. Inhibiting the alcohol-mediated increase in either TBP (Fig. 3A) or Brf1 (Fig. 3C) abrogated alcohol-mediated induction of RNA pol III transcription. Conversely, enhanced TBP (Fig. 3B) or Brf1 (Fig. 3D) expression had a modest effect on RNA pol III transcription in the absence of alcohol. However, treatment with alcohol induced a robust increase in RNA pol III transcription when either TBP or Brf1 expression was increased. These results support the idea that alcohol mediates RNA pol III-dependent transcription induction through its ability to enhance the expression of both TBP and Brf1.

Alcohol Induces Both Elk-1 and AP-1 Activation, but Only Increased c-Jun Expression Is Required to Induce TBP, Brf1, and tRNA Gene Expression

The TBP promoter is transcriptionally regulated through the recruitment of either Elk-1 (11) or AP-1 (13) to the promoter. Therefore, we measured the affect of alcohol on Elk-1, c-Jun, and c-Fos. Immunoblot analysis indicated that alcohol increased Elk-1 phosphorylation and expression of c-Jun, but no change in c-Fos expression was observed (Fig. 4A). AP-1 activity was assessed by transfecting a human AP-1-dependent promoter-driven luciferase construct, or the same promoter containing a mutated AP-1 site, into HepG2-ADH cells (Fig. 4B). Expression of these constructs in HepG2-ADH revealed that only the promoter containing an intact AP-1 binding site was induced by alcohol. To determine Elk-1 activity, cells were cotransfected with a Gal4-responsive promoter-luciferase construct and an Elk-1-Gal4 DNA binding domain fusion protein expression vector (Fig. 4C). Ethanol treatment induced Elk-1 transactivation function. Together, these results demonstrate that alcohol induces c-Jun expression and Elk-1 phosphorylation and enhances the activities of both AP-1-dependent and Elk-1-dependent promoters.

We next assessed the potential role of Elk-1 and c-Jun in mediating the alcohol-dependent increases in TBP, Brf1, and tRNA gene expression. Cellular Elk-1 expression was decreased, and RT-qPCR was used to measure the amounts of TBP and Brf1 mRNAs as well as pre-tRNA^Leu transcripts (Fig. 4D). No significant change in the expression of these transcripts was observed upon decreased expression of Elk-1. In contrast, down-regulation of c-jun expression abrogated the ethanol-mediated increases in all three of these transcripts (Fig. 4E). This can be compared with the expression of TFIIIC₆₃, which was unaffected by decreased c-Jun expression. These results support the idea that although Elk-1 is activated in alcohol-treated cells, increased expression of c-Jun alone drives the enhanced expression of TBP, Brf1, and tRNA genes.

Alcohol Induces Enhanced Recruitment of c-Jun to TBP, Brf1, and tRNA Gene Promoters

To determine whether the alcohol-mediated increases c-Jun required to stimulate TBP, Brf1, and tRNA gene expression were due to enhanced occupancy of c-Jun at these promoters, chromatin immunoprecipitation (ChIP) assays were performed. Ethanol induced a marked increase in the recruitment of c-Jun to the TBP promoter, whereas a modest increase in the occupancy of both c-Fos and Elk-1 was observed (Fig. 5A). These results support the idea that alcohol mediates an increase in TBP expression primarily through enhanced expression of c-Jun and its recruitment to the TBP promoter.

Sequence analysis revealed both putative AP-1 and putative Elk-1 binding sites in regions 5′ of the transcription start site of the Brf1 gene (Fig. 5B). ChIP analysis revealed that ethanol induced an increase in the occupancy of c-Jun to the region containing an AP-1 binding site, whereas no significant increase in binding of Elk-1 to the Elk-1 binding site was observed. Interestingly, c-Jun, c-Fos, and Elk-1 were also found to occupy the tRNA^Leu gene (Fig. 5C). Upon ethanol treatment, an increase in TBP recruitment to the tRNA^Leu gene was observed, as well as a marked increase in the occupancy of c-Jun. However, alcohol did not induce any apparent change in recruitment of either c-Fos or Elk-1 to the tRNA gene. These results suggest that c-Jun may function directly at sequences near tRNA genes to induce transcription. Together, these results support the idea that alcohol enhances expression of TBP, Brf1, and tRNA gene by facilitating c-Jun recruitment to these promoters.

TBP, Brf1, and RNA pol III-dependent Gene Expression Are Increased in Alcohol-fed Mice

We further assessed whether the observed alcohol-mediated increases in gene expression occur in vivo. C57BL/6 mice were chronically fed with a control diet or 3.5% ethanol in the Lieber-DeCarli liquid diet for 12 months. These mice display mild fatty liver disease (15). RNA was subsequently isolated from the livers of these mice, and RT-qPCR was used to measure the amounts of pre-tRNA^Leu and 5S rRNA (Fig. 6A), as well as TBP, and Brf1 mRNA (Fig. 6B). Alcohol feeding resulted in a modest induction of all of these transcripts.

To further evaluate potential alterations in these transcripts by alcohol, a transgenic mouse model expressing the hepatitis C NS5A protein was examined. Although C57BL/6 mice fed alcohol over a 12-month period do not develop liver tumors, previous studies demonstrated that alcohol feeding of C57BL/6 NS5A transgenic mice for 12 months induces liver tumors in ∼23% of these mice (15). In addition, these mice develop more advanced fatty liver disease and hepatomegaly and have significantly higher serum TNFα levels than wild-type mice fed alcohol. Comparison of non-alcohol-fed wild-type and NS5A transgenic mice revealed that the livers of NS5A transgenic mice exhibited ∼2–3-fold higher levels of pre-tRNA^Leu, 5S rRNA, TBP, and Brf1 transcripts (Fig. 6). Upon alcohol feeding of the NS5A transgenic mice, further induction of RNA pol III transcription as well as TBP and Brf1 expression was observed in the livers of these mice without tumors. More pronounced induction of these genes occurred in liver tumors of the same mice. When compared with the control livers in non-alcohol-fed NS5A transgenic mice, there was a robust increase in pre-tRNA^Leu and 5S rRNA, as well as TBP, and Brf1 mRNA in the liver tumors of the alcohol-fed mice. Together, these results demonstrate that both RNA pol III-dependent transcription and TBP and Brf1 expression are induced in animals that are chronically fed alcohol. In addition, these results support the idea that these gene expression processes are further enhanced in the oncogenic state.