Induction of Human Methionine Adenosyltransferase 2A Expression by Tumor Necrosis Factor α

Two genes (MAT1A and MAT2A) encode for methionine adenosyltransferase (MAT), an essential cellular enzyme responsible for S-adenosylmethionine biosynthesis. MAT1A is expressed mostly in the liver, whereas MAT2A is widely distributed. We showed a switch from MAT1A to MAT2A expression in human hepatocellular carcinoma (HCC), which facilitates cancer cell growth. Using DNase I footprinting analysis, we previously identified a region in the MAT2A promoter protected from DNase I digestion in HCC. This region contains NF-κB and AP-1 elements, and the present study examined whether they regulate MAT2A promoter activity. We found nuclear binding of NF-κB and AP-1 to the MAT2A promoter increased in HCC. Tumor necrosis factor α (TNFα), which activates both NF-κB and AP-1, increased MAT2A expression in a dose- and time-dependent manner, binding of both NF-κB and AP-1 to the MAT2A promoter and MAT2A promoter activity, with the latter effect blocked by site-directed mutagenesis of the NF-κB and AP-1 binding sites. Blocking NF-κB with IκB super-repressor or AP-1 with dominant-negative c-Jun led to decreased basal MAT2A expression and prevented the TNFα-induced increase in MAT2A expression. Although blocking NF-κB had no influence on the ability of TNFα to increase AP-1 nuclear binding, blocking AP-1 with dominant-negative c-Jun prevented the TNFα-mediated increase in NF-κB binding. In conclusion, both NF-κB and AP-1 are required for basal MAT2A expression in HepG2 cells and mediate the increase in MAT2A expression in response to TNFα treatment. Increased trans-activation of these two sites also contributes to MAT2A up-regulation in HCC.

Methionine adenosyltransferase (MAT) 1 is an essential cellular enzyme that catalyzes the formation of S-adenosylmethi-onine (SAMe), the principal biological methyl donor and the ultimate source of the propylamine moiety used in polyamine biosynthesis (1,2). In mammals, two different genes, MAT1A and MAT2A, encode for two homologous MAT catalytic subunits, ␣1 and ␣2 (3)(4)(5). MAT1A is expressed mostly in liver, and it encodes the ␣1 subunit found in two native MAT isozymes, which are either a dimer (MAT III) or tetramer (MAT I) of this single subunit (5). MAT2A encodes for a catalytic subunit (␣2) found in a native MAT isozyme (MAT II), which is associated with a catalytically inactive regulatory subunit (␤) in lymphocytes encoded by yet a third gene (5,6). MAT2A is widely distributed (3)(4)(5). MAT2A also predominates in the fetal liver and is progressively replaced by MAT1A during liver development (7,8). In adult liver, increased expression of MAT2A is associated with rapid growth or de-differentiation of the liver (9 -11). Using a cell line model that differs only in the type of MAT expressed, we demonstrated that a switch in MAT expression in liver cancer (from MAT1A to MAT2A) plays an important pathogenetic role by facilitating liver cancer growth (12). The influence of MAT expression on liver growth and injury was further demonstrated using a MAT1A knockout mouse model (13,14). In this model, absence of hepatic MAT1A is compensated by induction of MAT2A. These animals exhibit chronic hepatic SAMe deficiency, are prone to liver injury and develop spontaneous hepatocellular carcinoma (HCC) (13,14).
Given the importance of MAT expression in liver disease and cancer, we have been interested in understanding transcriptional regulation of MAT genes. We characterized the promoter region of both human MAT genes (15,16) and showed previously that in human HCC, both promoter hypomethylation (17) and increased expression of c-Myb and Sp1 with subsequent trans-activation of the MAT2A promoter contribute to transcriptional up-regulation of MAT2A in HCC (18). In the latter work we described increased protein binding to the MAT2A promoter region (Ϫ354 to Ϫ312) in HCC as compared with normal liver (18). Two other consensus elements in this DNase I protected region are nuclear factor kappa B (NF-B) and activator protein 1 (AP-1). We had previously observed that nuclear binding of NF-B and AP-1 to the promoter of human glutamate-cysteine ligase catalytic subunit was increased in HCC (19). In the current study, we examined whether there is increased nuclear binding of NF-B and AP-1 to the MAT2A promoter in HCC and whether these two transcription factors regulate MAT2A expression by using tumor necrosis factor alpha (TNF␣) as an inducing agent. * This work was supported by National Institutes of Health Grants DK51719, AA12677, and AT1576 (to S. C. L.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. All other reagents were of analytical grade and were obtained from commercial sources.
Source of Normal and Cancerous Liver Tissue-Normal liver tissue was obtained from normal liver included in the resected liver specimens of five patients with metastatic colon or breast carcinoma. Cancerous liver tissue was obtained from five patients undergoing surgical resec-tion for primary HCC. Written informed consent was obtained from each patient. The contamination of HCC samples with noncancerous tissue was less than 5% as determined by histopathology. These tissues were immediately frozen in liquid nitrogen for subsequent isolation of nuclear proteins for electrophoretic mobility shift assay (EMSA) as described below. The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki as reflected in a priori approval by Keck School of Medicine University of Southern California's human research review committee.
Cell Culture and TNF␣ Treatment-HepG2 cells were obtained from the Cell Culture Core of the USC Liver Disease Research Center and grown according to instructions provided by the American Type Culture Collection (Rockville, MD) in Earle's minimal essential medium supple-FIG. 1. A, consensus binding sites present in the DNase I protected region of MAT2A in HCC (18). B, electrophoretic mobility shift and supershift assay for the probe that spans Ϫ354 to Ϫ326 of the human MAT2A gene. EMSA and supershift were done as described under "Experimental Procedures." In HCC, there is increased binding (two bands) to the probe as compared with normal liver. Supershift analysis shows that the top band is due to c-Myb binding, whereas the bottom band is likely to be NF-B binding. C, electrophoretic mobility shift assay for the probe that spans the AP-1 site at Ϫ327 of the human MAT2A gene. In HCC, there is increased binding to the AP-1 probe as compared with normal liver, which is confirmed by supershift analysis. Representative EMSAs are shown from at least three experiments. mented with 10% fetal bovine serum, 2 mM L-glutamine, and 1% penicillin-streptomycin mixture. Prior to treatment with TNF␣, medium was changed to withhold serum overnight. Cells were then treated with TNF␣ (1-25 ng/ml) for 15 min to 16 h for various assays as described below.
Recombinant Plasmids and Adenoviral Vectors-The human MAT2A promoter constructs were previously described (15,17) and subcloned in the sense orientation upstream of the luciferase coding sequence of the pGL-3 enhancer vector (Promega). Recombinant, replication-defective adenovirus expressing dominant-negative c-Jun, TAM67, was kindly provided by Dr. David Brenner (20). Recombinant, replication-defective adenovirus expressing IB super-repressor (IBSR), which expresses a mutant IB that cannot be phosphorylated and therefore irreversibly binds NF-B and prevents its activation was kindly provided by Ebrahim Zandi (21).
Infection of HepG2 Cells with Adenoviral Vectors-Recombinant adenoviruses encoding TAM67, IBSR, or empty vector were amplified in 293 cells. HepG2 cells were infected with unpurified recombinant adenovirus encoding for TAM67, IBSR, or empty vector at multiplicity of 20 plaque-forming units/cell for 24 h. After 24 h of infection, the viruses were removed and replaced with fresh medium for transfection analysis as described below.
Effect of TNF␣ on the Expression of MAT2A in HepG2 Cells-HepG2 cells were treated with 0 -25 ng/ml TNF␣ for 8 h or 15 ng/ml for 0 -16 h. In some experiments, cells had been infected with adenovirus encoding IBSR, TAM67, or adenovirus vector alone prior to TNF␣ treat- ment. At the end of the TNF␣ treatment, total RNA was extracted, and Northern hybridization analysis was performed using specific MAT2A cDNA probe as described previously (18). Northern hybridization analysis was also performed using a specific c-Jun cDNA probe as we described (22), and a p65 cDNA probe that corresponds to nucleotides 481-1032 of the published human p65 sequence (23). The p65 cDNA probe was obtained by reverse transcription and PCR using a one step RT-PCR kit (Clontech). To ensure equal loading of RNA samples and transfer in each of the lanes, prior to hybridization, membranes were rinsed with ethidium bromide and photographed, and the same mem-branes were also rehybridized with a 32 P-labeled ␤-actin cDNA probe as described (18). Autoradiography and densitometry (Gel Documentation System, Scientific Technologies, Carlsbad, CA, and National Institutes of Health Image 1.60 software program) were used to quantitate relative RNA. Results of Northern blot analysis were normalized to ␤-actin.
Analysis of Promoter Constructs in HepG2 Cells-To study the relative transcriptional activities of the MAT2A promoter fragments, HepG2 cells (1 ϫ 10 6 cells in 2 ml of serum-free medium) were transiently transfected with 2.5 g of MAT2A promoter luciferase gene construct or promoterless pGL3-enhancer vector (as negative control) FIG. 3. Effect of TNF␣ treatment on luciferase activity driven by the human MAT2A promoter. Progressive 5Ј deletions of the MAT2A promoter extending from Ϫ1329 to ϩ60 bp were generated and fused to the promoterless luciferase pGL-3 enhancer vector as described under "Experimental Procedures." HepG2 cells were transfected transiently with these promoter constructs or pGL-3 enhancer vector alone and treated with TNF␣ (15 ng/ml for 8 h) or vehicle control. Cells were co-transfected with Renilla luciferase for control of transfection efficiency. Results represent mean Ϯ S.E. from four independent experiments performed in triplicates. Data are expressed as relative luciferase activity to that of pGL-3 enhancer vector control, which is assigned a value of 1.0. *, p Ͻ 0.05 versus respective control.

FIG. 4. Effect of TNF␣ treatment on
DNase I footprinting analysis of the ؊494 to ؊152 region of the human MAT2A promoter. DNA fragment was end-labeled on either strand and digested with DNase I in the absence (0) or presence of 5-15 g of nuclear protein extracts from HepG2 cells treated with TNF␣ (15 ng/ml for 8 h) or vehicle control. Position of the protected region is indicated at the right of the figure. Lane GϩA represents a Maxam-Gilbert sequencing reaction in the same fragment. Representative DNase I footprinting is shown.
using the SuperFect Transfection Reagent (Qiagen, Valencia, CA) as we described previously (22). To control for transfection efficiency, cells were co-transfected with the Renilla phRL-TK vector (Promega, Madison, WI), which is a plasmid containing Renilla luciferase gene driven by HSV-TK promoter. After 10 h, cells were harvested and lysed in 200 l of reporter lysis buffer (Luciferase Assay System, Promega). Aliquots of the cell lysates were sequentially assessed for firefly and Renilla luciferase activities using a TD-20/20 Luminometer (Promega). The luciferase activity driven by the MAT2A promoter constructs was normalized to Renilla luciferase activity. Each experiment was done with triplicate samples.
The effect of TNF␣ on MAT2A promoter activity was examined by measuring luciferase activity driven by MAT2A promoter luciferase gene constructs in transfected HepG2 cells treated with TNF␣ (15 ng/ml) during the last 8 h of the transfection. To confirm the importance of NF-B and AP-1 binding in mediating the effect of TNF␣ on MAT2A promoter activity, human MAT2A promoter constructs 571/ϩ60-LUC mutated in either the NF-B and/or AP-1 binding sites were generated using PCR. MAT2A promoter constructs mutated in either the putative NF-B site (Ϫ334 to Ϫ326) (from 5Ј-GGGAGGTGC-3Ј to 5Ј-GGATA-ATGC-3Ј), AP-1 site (Ϫ328 to Ϫ318) (from 5Ј-TGCCATGTCAC-3Ј to 5Ј-TGCCTCTACAC-3Ј), or both were subcloned into the pGL-3 promoter-luciferase vector (Promega). HepG2 cells transfected with either wild type or mutant MAT2A promoter constructs were treated with TNF␣ (15 ng/ml) during the last 8 h of the transfection.
DNase I Footprinting Analysis-The 32 P-end-labeled fragment of the 5Ј-flanking region of human MAT2A strongly induced by TNF␣ treatment was generated by digestion with restriction endonucleases. DNase I footprinting analysis was performed with 0 -15 g of nuclear protein from HepG2 cells treated with TNF␣ (15 ng/ml for 8 h) or vehicle, and labeled double-stranded fragments corresponding to nucleotides Ϫ494 to Ϫ152 of the human MAT2A gene as we described before (18).
To see if TNF␣ treatment of HepG2 cells resulted in increased NF-B and AP-1 binding to the MAT2A promoter, EMSA with supershift analysis was performed using the NF-B and AP-1 probes after treating cells with TNF␣ (15 ng/ml for 8 h). To assess the effect of IBSR or dominant-negative c-Jun, cells were previously infected with IBSR, TAM67, or empty viral vector and subsequently treated with TNF␣ (15 ng/ml for 8 h) and processed for EMSA and supershift analysis.
Measurement of Apoptosis-Apoptosis was assessed by both DNA fragmentation analysis and Hoechst staining. To assess DNA fragmen- were obtained from HepG2 cells previously infected with IBSR or adenoviral vector alone (Adv) and subsequently treated with TNF␣ (15 ng/ml for 8 h). EMSA (to NF-B probe) with supershift was done using anti-p65 antibodies as described under "Experimental Procedures." B, AP-1 binding. Nuclear protein extracts (15 g) were obtained from HepG2 cells previously infected with TAM67 or adenoviral vector alone (Adv) and subsequently treated with TNF␣ (15 ng/ml for 8 h). EMSA (to AP-1 probe) with supershift was done using anti-c-Jun antibodies as described under "Experimental Procedures." The arrows to the right point to specific complexes that were supershifted in the presence of specific antibodies.
tation, after treating HepG2 cells with TNF␣ with or without prior infection with adenoviral vector expressing IBSR or TAM67, cells were lysed in 500 l of lysis buffer (10 mM Tris-HCl, pH 8.0, 10 mM EDTA, 1% Nonidet P-40, and 0.5 mg/ml proteinase K) and incubated at room temperature for 1 h. The lysates were centrifuged at 20,800 ϫ g for 10 min at 4°C. The supernatant was then incubated with 50 g/ml RNase at 37°C for 1 h, and DNA was extracted with phenol and chloroform (1:1), precipitated by ethanol, and resuspended in Tris-EDTA buffer. DNA samples (5 g each) were separated electrophoretically on 2% agarose gel containing ethidium bromide (0.5 g/ml). The gel was destained in water for 20 min and photographed.
To perform Hoechst staining, HepG2 cells were grown on coverslips and treated as above. Following treatment, cells were fixed with paraformaldehyde and stained with 8 g/ml Hoechst 33258 dye for 30 min. Cells with bright, fragmented, condensed nuclei were identified as apoptotic cells using the Nikon Eclipse TE300 fluorescence microscope. At least five random fields (at ϫ300) were counted.
Statistical Analysis-Data are given as mean Ϯ S.E. Statistical analysis was performed using analysis of variance followed by Fisher's test for multiple comparisons. For changes in mRNA levels, ratios of MAT2A, c-jun, and p65 to ␤-actin densitometric values were compared by analysis of variance. Significance was defined by p Ͻ 0.05.

Nuclear Binding of NF-B and AP-1 to the MAT2A Promoter in HCC-
We showed previously that Ϫ354 to Ϫ312 of the human MAT2A promoter was protected from DNase I digestion in HCC but not in normal liver (18). Consensus elements present in this region include nuclear factor 1 (NF1), c-Myb, nuclear factor-erythroid 2 related factor (Nrf-2), NF-B, and AP-1 (Fig.  1A). We demonstrated increased expression of c-Myb and trans-activation of its binding site in HCC (18). To see if there is increased binding to the other potential binding sites, EMSA and supershift analysis were performed using specific antibod-ies against NF1, Nrf-2, NF-B subunits, and c-Jun. Fig. 1B shows that in addition to c-Myb, there is also increased p65 binding to the EMSA probe Ϫ354 to Ϫ326 but not NF-1 or Nrf-2. To see if there is increased AP-1 binding, EMSA was done using probe Ϫ328 to Ϫ316, which span mostly the AP-1 binding site. Fig. 1C shows that there is also increased AP-1 binding to this site in HCC.
Effect of TNF␣ on MAT2A Expression and Promoter Activity-TNF␣ is a pleiotropic cytokine that can elicit a diversity of cellular responses, depending on the cell type and concentration used (24,25). It is well know to induce both AP-1 and NF-B (24,25) and can serve as an ideal agent to examine whether these two transcription factors modulate MAT2A promoter activity. We first examined whether TNF␣ treatment affects MAT2A expression in HepG2 cells. Fig. 2A shows that TNF␣ increased MAT2A expression in HepG2 cells in a doseand time-dependent manner. It also induced c-jun and p65 expression in a similar manner (Fig. 2, B and C). Maximal induction of MAT2A was seen with 15 ng/ml TNF␣ for 8 h (300 Ϯ 34% of control, results represent mean Ϯ S.E. from four experiments, p Ͻ 0.05 versus control), the condition used for all subsequent experiments.
We next investigated the effect of TNF␣ treatment on the promoter activity of MAT2A. Fig. 3 shows that TNF␣ treatment induced the reporter activity driven by the MAT2A promoter, particularly the promoter construct Ϫ571 to ϩ60, where TNF␣ led to a nearly 2-fold increase in activity. It had no influence on the promoter activity driven by the construct Ϫ270 to ϩ60, suggesting that the region responsible for this stimulatory effect lie between Ϫ571 and Ϫ270.
DNase I Footprinting Analysis of MAT2A 5Ј-Flanking Region-To delineate the cis-acting elements that may be involved in mediating the effect of TNF␣ on the MAT2A promoter, DNase I footprinting analysis of the Ϫ494 to Ϫ152 5Ј-flanking region of the MAT2A gene was carried out. Fig. 4 shows that TNF␣ treatment led to an increase in nuclear protein-dependent DNase I-protected region from Ϫ352 to Ϫ314 on both strands. Both NF-B and AP-1 consensus binding sites are present in this region.
EMSA and Supershift Analysis-We next examined whether TNF␣ treatment led to an increase in both NF-B and AP-1 binding to their respective sites on the MAT2A promoter (Fig.  5). As expected, nuclear binding to the NF-B and AP-1 sites increased following TNF␣ treatment. Note that cells previously infected with IBSR had a lower basal NF-B binding and blunted the increase in NF-B binding following TNF␣ treatment (Fig. 5A). Similarly, cells previously infected with TAM67 also had a lower basal AP-1 binding and prevented the increase in AP-1 binding following TNF␣ treatment (Fig. 5B). These treatments did not induce apoptosis in HepG2 cells as determined by DNA fragmentation and Hoechst staining (data not shown).

Role of NF-B and AP-1 in Basal and TNF␣-induced
Increase in MAT2A Expression-To evaluate the importance of NF-B and AP-1 in both the basal expression and in the TNF␣mediated increase in MAT2A expression, HepG2 cells were infected with IBSR, TAM67, or empty vector and the effect of blocking NF-B or AP-1 activity on TNF␣-mediated changes was examined. Fig. 6A shows that the basal expression of MAT2A fell when NF-B was blocked (43% lower), and although the TNF␣-mediated induction in MAT2A was lower, it was still much higher than IBSR alone. This suggests that TNF␣-induced MAT2A expression by both NF-B-dependent and -independent mechanisms. Fig. 6B shows that the basal expression of MAT2A also depends on AP-1, because the expression fell (50% lower) when AP-1 was blocked. Furthermore, dominant negative c-Jun was able to prevent the increase in MAT2A following TNF␣ treatment more completely than IBSR.
The importance of NF-B and AP-1 binding in mediating the effect of TNF␣ on the MAT2A promoter was confirmed using promoter constructs mutated in these binding sites. We first determined the number of bases that need to be mutated to obliterate nuclear binding completely. Fig. 7A shows that nuclear binding was completely obliterated only when four bases are mutated. These were then used in subsequent transfection experiments. Fig. 7B shows that when the NF-B or AP-1 site is mutated, the effect of TNF␣ is significantly blunted, and when both sites are mutated, the effect of TNF␣ is completely prevented.

Interactions between NF-B and AP-1-Previous studies
show that the NF-B p65 can physically interact with c-Fos and c-Jun and potentiate each other's biological function (26). To see if p65 and c-Jun physically interact in the modulation of MAT2A promoter activity, we first examined whether p65 and c-Jun can bind to each other's binding sites. Fig. 8A shows that c-Jun does not bind to the NF-B site, and p65 does not bind to the AP-1 site. We next examined the effect of blocking either NF-B or AP-1 on each other's nuclear binding activity. Although blocking NF-B had no influence on the ability of TNF␣ to induce AP-1 binding to the MAT2A promoter (Fig. 8B), blocking AP-1 with dominant negative c-Jun abolished the ability of TNF␣ to induce NF-B binding (Fig. 8C).  LUC (B). A, EMSA was performed with 10 g of nuclear protein from control or TNF␣-treated HepG2 cells (15 ng/ml for 8 h) using wild type AP-1 and NF-B probes or probes mutated by 1-4 bp as described under "Experimental Procedures." Note that nuclear binding was obliterated only when four bases are mutated. B, MAT2A promoter-luciferase construct Ϫ571/ϩ60, wild type or mutated in the putative AP-1 site (Ϫ328 to Ϫ318) (from 5Ј-TGCCATGTCAC-3Ј to 5Ј-TGCCTCTACAC-3Ј), the putative NF-B site (Ϫ334 to Ϫ326) (from 5Ј-GGGAGGTGC-3Ј to 5Ј-GGATAATGC-3Ј), or both were subcloned into pGL-3 enhancer vector and used for transient transfection analysis in HepG2 cells. Cells were treated with TNF␣ (15 ng/ml for 8 h) or vehicle control prior to luciferase assays. Results represent mean Ϯ S.E. from three independent experiments performed in triplicates. Data are expressed as relative luciferase activity to that of pGL-3 enhancer vector control, which is assigned a value of 1.0. *, p Ͻ 0.05 versus Ϫ571/ϩ60-LUC construct; †, p Ͻ 0.05 versus respective mutated Ϫ571/ϩ60-LUC constructs; **, p Ͻ 0.05 versus Ϫ571/ϩ60-LUC construct and constructs mutated in either AP-1 or NF-B binding sites.

DISCUSSION
MAT is a critical cellular enzyme because it catalyzes the only reaction that generates SAMe. The MAT gene is one of 482 genes absolutely required for survival of an organism (1). In mammalian liver, MAT1A is a marker for the differentiated or mature liver phenotype, whereas MAT2A is a marker for rapid growth and de-differentiation (27). Although the two MAT genes are highly homologous, the enzymes they encode for are different in the kinetic profiles and regulatory properties (27). Due to these differences, the type of MAT expressed by a cell can influence the steady-state SAMe level of the cell and methylation status. Using a cell line model that differs only in the type of MAT expressed, cells that expressed MAT1A had the highest intracellular SAMe level and DNA methylation, whereas cells that expressed MAT2A had the opposite (12). Interestingly, cells that expressed MAT2A grew faster than cells that expressed MAT1A (12). Thus, the switch in MAT expression in liver cancer is pathogenetically important, because it offered the cancerous cell a growth advantage. The importance of MAT expression on SAMe level and liver pheno-type was recently confirmed in the MAT1A knockout mouse model where replacement of MAT1A with MAT2A resulted in chronic hepatic SAMe depletion and eventual development of HCC (13,14).
Despite the importance of MAT2A, studies addressing its transcriptional regulation have only begun to appear recently. MAT2A gene expression is influenced by the cell cycle, as evident by its induction during liver regeneration, malignant liver transformation and T-lymphocyte activation (10,17,28). It has been speculated that the induction in MAT2A and MAT II may be a mechanism for the cell to provide an increased supply of SAMe, the precursor to polyamine synthesis that is required for cell growth (28). Our laboratory has been interested in studying transcriptional regulation of MAT2A and, in particular, the molecular mechanism for its up-regulation in human liver cancer. Our previous studies (17,18) described promoter hypomethylation and increased trans-activation of the MAT2A promoter by c-Myb and Sp1 as mechanisms that contribute to its up-regulation in HCC. Two other consensus elements present in the DNase I-protected region of the FIG. 8. Interactions between NF-B and AP-1 in the TNF␣-mediated increase in nuclear binding to MAT2A promoter. A, nuclear protein extracts (15 g) were obtained from HepG2 cells treated with TNF␣ (15 ng/ml for 8 h) or vehicle control, and EMSA with supershift was done as described under "Experimental Procedures" using both NF-B and AP-1 probes. Note that supershift occurred only with anti-p65 antibodies to the NF-B probe and anti-c-Jun antibodies to the AP-1 probe. In B, binding to the AP-1 probe was assessed in cells previously infected with IBSR and subsequently treated with TNF␣ (15 ng/ml for 8 h). In C, binding to the NF-B probe was assessed in cells previously infected with TAM67 and subsequently treated with TNF␣ (15 ng/ml for 8 h). Note that TAM67 treatment blocked the increase in NF-B binding following TNF␣ treatment. The arrows to the right point to specific complexes that were supershifted in the presence of specific antibodies.
MAT2A promoter in HCC are NF-B and AP-1. In the current work we tested the hypothesis that MAT2A is also regulated by NF-B and AP-1.
We first established that in HCC there is also increased binding of NF-B and AP-1 to the MAT2A promoter. On the other hand, two other transcription factors, NF1 and Nrf-2, did not bind to this region. We next used TNF␣ as an inducing agent to see if it affects MAT2A expression. TNF␣ was chosen for several reasons. First is the fact that it is well known to induce both NF-B and Jun kinase (24,25). Second, MAT2A expression is induced in many conditions, where TNF␣ is also known to be induced. Specifically, liver regeneration after partial hepatectomy and after ethanol feeding (10,29). Finally, TNF␣ is known to act as a mitogen to stimulate cell proliferation (24) and MAT2A is known to be induced under a proliferative state (9,10,27). In addition to being a mitogen, TNF␣ can also elicit a cytotoxic response (25,30). Hepatocytes are normally resistant to TNF␣ cytotoxicity but undergoes cell death in the setting of transcriptional or translational arrest, or inactivation of NF-B (25,30). Thus, NF-B activation is thought to turn on several antiapoptotic proteins to resist TNF␣ toxicity. The role of Jun kinase activation in TNF␣ cytotoxicity has been more controversial, because both antiapoptotic and proapoptotic effects have been described (25,30). In our studies we have utilized adenoviral vectors that block either NF-B or AP-1. We have made certain that under our experimental conditions, there was no difference in apoptosis to complicate interpretation of the results.
TNF␣ induced MAT2A expression in a time-and dose-dependent manner and binding of NF-B and AP-1 to the MAT2A promoter. Early induction of NF-B and AP-1 is likely to occur via activation of existing proteins. However, TNF␣ treatment also induced the gene expression of both p65 and c-jun after 2 h. TNF␣ treatment induced maximally the reporter activity driven by MAT2A promoter construct Ϫ571/ϩ60 but not Ϫ270/ ϩ60, thus narrows the site of action to Ϫ571 to Ϫ270 of MAT2A. DNase I footprinting analysis further narrowed this to Ϫ352 to Ϫ314. The obvious candidates here are NF-B and AP-1.
To establish an essential role for NF-B and AP-1 in modulating MAT2A expression, two approaches were taken. First, we blocked their nuclear binding activity by adenoviral vectors that express either IBSR or dominant negative c-Jun. To our surprise, we found the basal MAT2A expression lower when NF-B or AP-1 was blocked. This suggests that the basal expression requires these transcription factors. TNF␣ was still able to increase MAT2A expression and promoter activity in cells pre-infected with IBSR. This suggests that both NF-Bdependent and -independent mechanisms contribute to TNF␣mediated MAT2A induction. TNF␣ was not able to increase MAT2A expression in cells pre-infected with dominant negative c-Jun, which suggests that AP-1 trans-activation is the other major mechanism that is responsible for TNF␣-mediated MAT2A induction. The second approach was site-directed mutagenesis, which confirmed the importance of these two binding sites in mediating the effect of TNF␣ on the MAT2A promoter. Consistent with results obtained with the dominant negative constructs, mutation of either of these binding sites lowered the basal promoter activity and blunted the TNF␣-mediated increase in promoter activity. When both sites are mutated, the stimulatory effect of TNF␣ is completely prevented. At the present time we cannot exclude the possibility that there may be other unidentified transcription factors involved in this region. However, taken together, our data show that both AP-1 and NF-B sites are functional in this promoter and are largely responsible for the TNF␣-mediated induction in MAT2A expression.
Why would blocking AP-1 prevent TNF␣-mediated MAT2A induction almost completely if NF-B was important? Gel shift analysis provided a major clue to its mechanism. TNF␣ treatment was not able to induce NF-B nuclear binding in cells pre-infected with dominant negative c-Jun. On the other hand, blocking NF-B with IBSR had no influence on TNF␣'s ability to induce AP-1 binding.
Both NF-B and AP-1 belong to families of transcription factors that exert pleiotropic regulation effects. Physical interaction of NF-B subunit p65 with AP-1 subunit c-Jun was shown to synergize and potentiate each other's biological function (26). However, a recent study failed to show physical interaction between these two proteins and in fact showed a reciprocal relationship presumably due to competition for coactivator p300 (31). To complicate this further, another recent study in a rat hepatocyte cell line showed inhibition of NF-B led to a sustained induction of AP-1 (30). In our studies, we did not find binding of p65 to the AP-1 site or c-Jun to the NF-B site. We also did not observe enhanced AP-1 binding activity when NF-B is blocked. However, the fact that dominant negative c-Jun was able to block NF-B activation suggests they do interact in HepG2 cells. One possibility is that the dominant negative c-Jun can bind to p65 and prevent it from binding to NF-B binding sites. This remains to be examined.
In summary, we have identified NF-B and AP-1 to be essential for the basal expression of MAT2A and to mediate the stimulatory effect of TNF␣ on MAT2A promoter activity and expression. This may be one of the key mechanisms for MAT2A up-regulation during rapid growth. In HCC, there is increased nuclear binding of both NF-B and AP-1 to the MAT2A promoter, and it is likely that increased trans-activation by these two transcription factors also contribute to the transcriptional up-regulation of MAT2A in HCC.