Transcriptional activation of the Cu,Zn-superoxide dismutase gene through the AP2 site by ginsenoside Rb2 extracted from a medicinal plant, Panax ginseng.

We report here that the ginseng saponins induce the transcription of Cu,Zn-superoxide dismutase gene (SOD1), which is one of the major antioxidant enzymes. Total saponins and panaxatriol did not elevate the level of SOD1, but panaxadiol significantly increased SOD1. Among the panaxadiol fractions, ginsenoside Rb2 was a more specific and more remarkable inducer of the SOD1 gene than ginsenoside Rb1. Deletion analyses of the SOD1 promoter revealed that the proximal promoter is responsible for this induction. Mobility shift assays with cis-elements in the proximal promoter region showed that specific binding of the AP2 transcription factor was significantly increased by treatment with ginsenoside Rb2. Mutations of the AP2 binding sites in the heterologous promoter and natural context systems abolished the transcriptional activation by ginsenoside Rb2. These results suggest that the SOD1 gene was greatly activated by ginsenoside Rb2 through transcription factor AP2 binding sites and its induction.

We report here that the ginseng saponins induce the transcription of Cu,Zn-superoxide dismutase gene (SOD1), which is one of the major antioxidant enzymes. Total saponins and panaxatriol did not elevate the level of SOD1, but panaxadiol significantly increased SOD1. Among the panaxadiol fractions, ginsenoside Rb 2 was a more specific and more remarkable inducer of the SOD1 gene than ginsenoside Rb 1 . Deletion analyses of the SOD1 promoter revealed that the proximal promoter is responsible for this induction. Mobility shift assays with cis-elements in the proximal promoter region showed that specific binding of the AP2 transcription factor was significantly increased by treatment with ginsenoside Rb 2 . Mutations of the AP2 binding sites in the heterologous promoter and natural context systems abolished the transcriptional activation by ginsenoside Rb 2 . These results suggest that the SOD1 gene was greatly activated by ginsenoside Rb 2 through transcription factor AP2 binding sites and its induction.
Cu,Zn-superoxide dismutase (SOD1) 1 is a key enzyme in the metabolism of oxygen free radicals. It catalyzes the dismutation of superoxide radicals (O 2 . ) to oxygen and hydrogen peroxide (1). Generation and/or removal of superoxides have been observed to play significant roles in a variety of critical homeostatic mechanisms both at the cellular and the organismic level. Since biological macromolecules would be a target for the damaging action of abundant oxygen radicals, the regulation mechanism of the SOD1 gene would be of great interest. It has been also reported that SOD1 could prevent oncogenesis and tumor promotion (2), reduce the cytotoxic and cardiotoxic effects of anticancer drugs (3), and protect against reperfusion damage of ischemic tissue (4). A recent report suggested that overexpression of SOD1 and catalase could increase the average life span of the fly (5).
In yeast, the expression of SOD1 is regulated by copper at the level of transcription (6). Lutropin, Cu 2ϩ ion, and reactive oxygen seemed to induce SOD1 in rat (7)(8)(9). The rat SOD1 promoter region has recently been studied for possible regulatory elements that could account for its induction by various agents (10). Functional heat shock elements within the human SOD1 promoter have been identified, which suggests that heat shock factors mediate the activation of SOD1 gene transcription under a heat shock stress condition (11). The diversity of SOD1 inducers implies that there are multiple regulatory elements for this gene. Indeed, the present study identified additional regulatory elements, binding sites for AP2, which would be an important factor mediating the role of SOD1 in the response to oxidative stress. Panax ginseng C. A. Mayer (Araliaceae) is one of the most popular natural tonics that has been used in oriental countries. Ginseng showed antitumor activities in slow growing tumors but not in rapidly growing tumors (12). Ginseng also inhibited tumor angiogenesis and metastasis (13). In an epidemiological study, the intake of ginseng reduced the incidence of human cancer (14). These antitumor effects of ginseng support the suggestion of Brekhman that ginseng may increase the nonspecific resistance of the organism (15). However, the identity of the active substance and its mechanism of action have not been elucidated yet. Ginsenosides can be classified into panaxadiol (PD) and panaxatriol (PT) saponins according to their structures (16). More than 30 types of ginsenosides, differing in the sugar moiety of the molecule, have been isolated and identified to date (17). In this study, we examined the effect of ginseng saponins on the transcription of the rat SOD1 gene, which is very similar to that of humans. They have an almost identical proximal part of the promoter region (11). We identified the active fraction of ginseng saponin and further characterized ginsenoside Rb 2 (Rb 2 ) as a strong transactivator of the SOD1 gene through AP2 binding sites and its induction.

EXPERIMENTAL PROCEDURES
Materials-Restriction endonuclease and DNA-modifying enzymes were obtained from New England Biolabs. The synthetic oligonucleotides for transcription factor binding site were purchased from Promega. [␥-32 P]ATP and [␣-32 P]dATP were from Amersham Corp. All ginsenosides were from Korea Ginseng and Tobacco Research Institute. The monoclonal antibody against AP2 (SE4) was a generous gift of Dr. T. Williams (Yale University). All other chemicals were of analytical or molecular grade and were purchased from Sigma.
Vector Constructions-The 1.7-kilobase pair BamHI/SmaI fragment (nucleotides Ϫ1633 to ϩ85) from the rat SOD1 gene (10) was inserted into the pBLCAT3 (18). Unidirectional 5Ј deletion mutants were produced by cutting 5Ј of the SOD1 promoter with SphI and BamHI, followed by subsequent treatment with exonuclease III (19). Deletion end points were confirmed by DNA sequencing with the Sequenase kit (U.S. Biochemical Corp.). For the construction of pAP2w, the oligonucleotide of the AP2 binding sequence, which corresponded to the SV40 AP2 binding site, was cloned into the BamHI site of pBLCAT2⌬, which is derived from pBLCAT2 (18). The plasmid pBLCAT2⌬ has the mini-mal region (Ϫ80 to ϩ51) of the herpes simplex virus thymidine kinase promoter. Two copies of the AP2 consensus oligonucleotide were introduced. The plasmid pAP2m is a mutant of pAP2w with two copies of the mutated AP2 site (from CCCGCGGC to CCATATGC). Insertion of the AP2 consensus and mutant sequences was confirmed by DNA sequencing. The plasmid pRSP-305AP2m was constructed as follows; DNA fragment (HaeII-NheI) from pRSP-305 (see Fig. 2) was excised and ligated with synthetic oligonucleotide containing the mutated AP2 binding site. The AP2 binding sequences located at Ϫ134 and Ϫ118 were replaced with mutated sequence (at Ϫ134, from CCCCGCCC to CCATATCC; at Ϫ118, from CCCCGCGG to CCATATGG).
Cell Culture, Transfection, and Treatment of Ginseng Saponins and Chemicals-Human HepG2 hepatoma cells were grown in Dulbecco's modified Eagle's medium, 10% fetal calf serum, penicillin G sodium (100 units/ml), streptomycin sulfate (100 g/ml), and amphotericin B (250 ng/ml). Cells were seeded into 60-mm plastic culture dishes (30 -50% confluence) for 24 h prior to transfection. An equal amount (3.0 pmol) of the various constructs was transfected to the cells by the calcium phosphate DNA coprecipitation method (20). Five g of pRSV␤gal plasmid (21) was introduced in all experiments to correct the variations of transfection efficiency. Ginseng saponins were added to culture medium at 36 h after transfection, and the cells were maintained for an additional 22 h. To determine the maximum induction time of Rb 2 , the growth medium was removed, and the Rb 2 was added to the cells at 50 M in phosphate-buffered saline (PBS). Treatments were maintained for 30 min at 37°C. After treatments the growth medium was added back to the cells, and incubation was continued for the intervals indicated in Fig. 1. In order to determine the new RNA and protein synthesis required, cultures of HepG2 cells were treated with either actinomycin D (2.5 g/ml) or cyclohexamide (10 g/ml), or left untreated. After 1 h, cultures were treated with ginsenoside Rb 2 (50 M) for 12 h.
␤-Galactosidase and CAT Assays-The CAT assay was performed as described previously (22). The transfected cells were washed twice with PBS and harvested. The pelleted cells were resuspended in 100 l of 0.25 M Tris-Cl (pH 7.9) and lysed by three cycles of freezing and thawing. After removal of cell debris by centrifugation, cell extracts were first assayed for ␤-galactosidase activity (23). Equal quantities of proteins were assayed for CAT activity on the basis of ␤-galactosidase activity. Extracts were incubated with 0.025 Ci of [ 14 C]chloramphenicol, 0.25 M Tris-Cl, pH 7.6, 0.4 mM acetyl coenzyme A for 1 h at 37°C. The enzyme assay was terminated by adding ethyl acetate. The organic layer was analyzed by TLC with chloroform/methanol (95:5). After autoradiography, both acetylated and unacetylated forms of [ 14 C]chloramphenicol were scraped from the plate, and the conversion of chloramphenicol to acetylated form was calculated by measuring radioactivities. The relative CAT activities were calculated from the percentage of conversion. Results are the average of three independent experiments.
Mobility Shift Assay (MSA)-The double-stranded oligonucleotides corresponding to the metal binding factor, heat shock factor, Sp1, AP2, CCAAT-enhancer binding protein ␣, TATA box binding protein, and NF-B binding sites were purchased from Promega and were synthesized from Genosys Biotechnologies, Inc. (The Woodland, Texas). The oligonucleotides were labeled with [␥-32 P]ATP and polynucleotide kinase (23). Nuclear extracts were prepared from control and Rb 2 -treated cells by the method of Dignam et al. (24). An equal amount (10 g) of nuclear extract from each sample was mixed with AP2 oligonucleotide for 20 min at 20°C in a 15-l solution containing 10 mM HEPES, 100 mM KCl, 5 mM MgCl 2 , 1 mM dithiothreitol, 1 mM EDTA, 10% glycerol, and 2 g of poly(dI-dC). After the binding reaction, a specific monoclonal antibody against AP2 was added to the reaction mixture. For competition assays, the binding reaction was performed with a 50-fold molar excess of cold probe or competitor DNA. DNA fragments from the SOD1 promoter were used as competitors. The numbers in Fig. 3 indicate the end points of competitor DNA fragments. Labeled DNA fragment (nucleotides Ϫ305 to Ϫ74 of the rat SOD1 promoter) was made by labeling with [␣-32 P]dATP and Klenow fragment. Ten ng of purified AP2 protein (Promega Corp.) was added to each binding reaction with or without competitor DNA. The binding reaction mixtures were electrophoresed in 6% acrylamide gels in 0.5 ϫ TBE (44 mM Tris, 44 mM boric acid, and 1 mM EDTA). After electrophoresis, gels were dried and exposed to x-ray film.
Western Blot Analysis-Nuclear extracts were made from Rb 2treated HepG2 cells as described by Dignam et al. (24), and 10 g of protein was subjected to 10% SDS-PAGE and transferred to Hybond C-extra membrane (Amersham). The blot was incubated with 1:1000 dilution of AP2 antibody (SE4) and visualized with the ECL system (Amersham).

Effect of Ginseng Saponins on the Expression of Rat Cu,Zn-SOD-
The effect of ginseng saponins on the induction of the SOD1 gene was examined as follows. At first, pRSP-1633, which is the SOD1 promoter region-CAT fusion plasmid (Fig.  2B), was introduced into HepG2 cells (18). Thirty-six h after transfection, ginseng saponins were added to the transfected cells to an appropriate concentration (Fig. 1, B-D). After 22 h, the CAT activity of each transfected cell was determined. Total saponins (TS) and PT treatment of the cell had no effect on the transcription of the SOD1 gene, but PD increased SOD1 about 3-fold (Fig. 1B). PD consisted of ginsenoside Rb 1 (Rb 1 ), Rb 2 and some other minor fractions. Therefore, we tested which subfraction of PD had an effect on the induction of the SOD1 gene. Rb 1 and Rb 2 were tested, and ginsenoside Rg 1 (Rg 1 ), which is a major subfraction of PT, was also tested as a negative control. Fig. 1C showed that Rg 1 had no effect, as expected, whereas Rb 2 was a 2-fold more potent activator than Rb 1 . Treatment of the cell with increased amounts of Rb 2 resulted in a gradual increase in the CAT activity (Fig. 1D). This result confirmed again the Rb 2 -specific activation of the SOD1 gene. To characterize the induction profile, transfected cells were treated with 50 M Rb 2 in PBS for 30 min. After incubation, the cells were washed with PBS, and then fresh growth medium was added. The peak of induction appeared around 2 h after treatment and then declined (Fig. 1E). This phenomenon was similar to the induction profile of heme oxygenase by oxidative stress (25). It is possible that transmission of the Rb 2 signal to induce the SOD1 gene could be mediated through the DNA element located in the promoter region.
Identification of Ginsenoside Response Element in the SOD1 Promoter-To identify the target sequence of Rb 2 in the upstream region of the SOD1 gene, deletion mutants were prepared and transfected into the HepG2 cell with or without Rb 2 in the medium. As outlined in the bar graph of Fig. 2A, deletion of DNA sequences from Ϫ576 to Ϫ412 and Ϫ412 to Ϫ305 resulted in decrease and increase of CAT activity, respectively, suggesting that there were positive and negative responsive elements in the SOD1 promoter (depicted in Fig. 2B as PRE and NRE). In Fig. 2A, Rb 2 -specific induction was observed in every construct except the one that had minimal promoter (pRSP-55). Rb 2 induced SOD1 expression about 3-fold, regardless of its promoter strength, under the appropriate conditions. These results showed that the target sequence for Rb 2 -specific induction was located in the Ϫ305 to Ϫ55 region. This region was the proximal part of the SOD1 gene promoter that consisted of a large number of transcription factor binding sites (10,11). It is possible that induction of SOD1 gene by Rb 2 is mediated by activation of specific transcription factors.
Increase of the AP2 Binding Activity by Ginsenoside Rb 2 -We therefore measured the effect of Rb 2 on the activation of tran-scription factors by MSA. We used double-stranded oligonucleotides corresponding to the possible binding sites of transcription factors including the metal binding factor, heat shock factor, Sp1, AP2, CCAAT-enhancer binding protein ␣, TATA box binding protein, and NF-B binding sites in the SOD1 promoter. The HepG2 cells were incubated with or without Rb 2 in the media for 22 h and harvested for extracting nuclear protein. An equal amount (10 g) of the nuclear protein of the control and Rb 2 -treated cells was used for MSA. A significant increase in AP2-DNA complex was observed upon treatment with Rb 2 in HepG2 cells (Fig. 3A, lanes 1-4). Each of the DNA-protein complexes with other factors was not influenced A, the increasing amount of the AP2 was observed (indicated by filled arrowhead) by the increasing amount of Rb 2 . Increased AP2-DNA complex was supershifted by monoclonal antibody (SE4) against AP2 (indicated by solid circle). B, identification of the increased AP2 protein by Western blot analysis. AP2 protein was visualized by ECL system (Amersham). AP2 protein near 52 kDa was indicated by filled arrowheads. Lane M, protein standard marker (kDa). C, requirement of RNA and protein synthesis for the AP2 induction. HepG2 cells were treated with either actinomycin (act.D) (2.5 g/ml) or cyclohexamide (CHX) (10 g/ml) for 1 h, and then the cells were treated with ginsenoside Rb 2 (50 M) for 12 h. Nuclear extracts were subjected to MSA with AP2 oligonucleotide. D, identification of the AP2 binding site in the promoter region of the SOD1 gene. MSA using the AP2 oligonucleotide with Rb 2 -treated HepG2 nuclear extract was done. Specific DNA-protein complex was observed (indicated by solid arrowhead). DNA fragment from Ϫ305 to Ϫ74 of the SOD1 promoter competed with AP2 oligonucleotide (lane 4). E, MSA using DNA fragment (Ϫ305 to Ϫ74) and purified AP2 protein. Specific complex was observed (indicated by filled arrowhead). F, MSA using AP2 oligonucleotide and purified AP2 protein. Note that the synthetic AP2 oligonucleotide and the SOD1 promoter region (Ϫ305 to Ϫ74) competed with each other. The location of free probe was indicated by blank arrowhead.
by Rb 2 treatment. 2 The binding of AP2 was further confirmed by the fact that the AP2-DNA complex was supershifted by the addition of anti-AP2 monoclonal antibody (Fig. 3A, lane 5). These results strongly indicated that there was a Rb 2 -specific increase in AP2 factor in a dose-dependent manner. The same result was also observed in MSA with Rb 2 -treated HeLa nuclear extract. 2 We tested whether Rb 2 induction of AP2 was dependent on new RNA and protein synthesis. When actinomycin D, a transcriptional inhibitor, or cyclohexamide, a protein synthesis inhibitor, was treated for 1 h before Rb 2 treatment, the AP2-DNA complex did not appear (Fig. 3C). These results indicate that the induction of AP2 by Rb 2 treatment depends on the synthesis of new RNA and protein. The amount of AP2 protein was determined by Western blot analysis. Using antibody against an AP2 protein, we measured AP2 protein in nuclear extracts of untreated and Rb 2 -treated HepG2 cells. As shown in Fig. 3B, the AP2 protein was considerably increased from the steady state level by Rb 2 treatment.
Presence of AP2 Binding Sites in the Promoter Region of SOD1 Gene-To determine whether the SOD1 gene promoter region contains an AP2 binding site, MSA using an oligonucleotide corresponding to the AP2 binding site was performed with Rb 2 -treated HepG2 nuclear protein. DNA fragments from the SOD1 promoter region were used in MSA as a competitor. A specific AP2 oligonucleotide-protein complex was observed (Fig. 3D, lane 1) and disappeared when the DNA fragment from Ϫ305 to Ϫ74 of the SOD1 promoter was added to the reaction mixture (Fig. 3D, lane 4). No competition was observed with other DNA fragments from the SOD1 promoter as competitors (Fig. 3D, lanes 2, 3, and 5). To further determine whether the transcription factor AP2 can bind to the proximal region of the SOD1 promoter, we performed MSA with purified AP2. Specific DNA-AP2 complexes were observed (Fig. 3E). The doublet formation of the shifted band indicated that there is more than one binding site (10). Labeled AP2 oligonucleotide was also put in competition with the DNA fragment of the SOD1 promoter reciprocally (Fig. 3F). These results suggested that functional AP2 binding sites were located in the proximal part (Ϫ305 to Ϫ74) of the SOD1 promoter region. This was also confirmed by DNA sequence analysis in which two sites of AP2 were located between Ϫ134 and Ϫ111 ( Ϫ134 CCCCGCCC Ϫ127 and Ϫ118 CCCCGCGG Ϫ111 ).
The Transactivation of the SOD1 Gene by Rb 2 through AP2 Binding Sites and Its Induction-We demonstrated that AP2 factor was increased by Rb 2 treatment, and its binding site was in the promoter region of the SOD1 gene. Naturally, we asked whether Rb 2 promotes the activation of the SOD1 promoter, which has two binding sites of AP2. The results showed that Rb 2 induced the CAT activity of SOD1 promoter-CAT constructs ( Figs. 1 and 2). Almost identical results were also obtained from cells that were transfected with the pAP2w plasmid, which has two copies of AP2 binding sites linked to a herpes simplex virus-thymidine kinase promoter (Fig. 4A). The CAT activity of the plasmid pAP2m containing a mutant AP2 site was not affected by Rb 2 (Fig. 4C). We also constructed a plasmid bearing mutant AP2 binding sites in the natural context. Mutated AP2 binding sites were introduced into pRSP-305 (see Fig. 4B). Two sites of AP2 were changed to mutant sequences (from Ϫ134 CCCCGCCC Ϫ127 to Ϫ134 CCATATCC Ϫ127 , from Ϫ118 CCCCGCGC Ϫ111 to Ϫ118 CCATATGC Ϫ111 ). These mutations abolished binding activity with the AP2 protein in vitro. 2 Rb 2 did not affect the level of CAT activity of this mutant plasmid, pRSP-305AP2m, whereas the CAT activity of the wild type pRSP-305 increased about 5-fold (Fig. 4D). These findings strongly confirmed that the Rb 2 transactivation of the SOD1 gene was mediated through AP2 sites and its induction. DISCUSSION In this study, we demonstrated that the activation of the SOD1 gene by ginsenoside Rb 2 was mediated by the transcription factor AP2. Rb 2 is present in relatively large quantities in the saponin fraction of P. ginseng. It has been reported that Rb 2 increased RNA polymerase I and II activity (26) and inhibited tumor angiogenesis and metastasis (13). Rb 2 is a dammaranetype saponin that has been shown to possess various biological activities such as a protein anabolic effect, an anti-diabetes effect, an anti-hyperlipemia effect, and an anti-inflammatory activity (27). Both Rb 1 and Rb 2 have the basic structure of 20(S)-protopanaxadiol. Rb 1 has four molecules of glucose in the sugar moiety, whereas Rb 2 has three molecules of glucose and one molecule of arabinose (Fig. 1A). Although there is only a minor difference in the sugar moiety, it has been suggested that the action mechanism of the two agents is somewhat different (26). In our study, it appeared that Rb 2 was a more potent activator of the SOD1 gene than Rb 1 . As a result, in Fig.  1B PT seemed to somewhat repress the CAT activity, and total saponin did not affect the CAT activity. There is an antagonistic action in the ginseng saponin components. Ginsenosides can stimulate cell growth and inhibit cell proliferation (28,29). Ginsenoside Rg 1 enhances the conversion of arginine to citrulline but not Rb 1 (30). The process of Rb 2 induction appears not to be through the oxidoreduction pathway but rather through the transcription factor AP2. The finding of AP2 binding sites on the SOD1 promoter suggests the importance of SOD1 in the process whereby this transcription factor is activated. Transcriptional activation by AP2 involves the 52-kDa AP2 protein binding to a specific DNA motif found in the cis-regulatory region of the gene (31). AP2 activity is regulated in a cell A, schematic diagram of synthetic AP2 (pAP2w) and mutated AP2 oligonucleotide (pAP2m) in the heterologous promoter (tk) attached to CAT. B, schematic diagram of the SOD promoter from nucleotide Ϫ305 to ϩ85 (pRSP-305) and mutated AP2 sites of the promoter (pRSP-305AP2m) attached to CAT. C, effect of Rb 2 on the synthetic AP2 and mutated AP2 in the heterologous promoter. D, effect of Rb 2 on AP2 and mutated AP2 in the natural context. Note that the mutations in the synthetic AP2 sites and in the natural context abolished the induction activity by Rb 2 . type-specific manner (32) and is induced by phorbol esters, retinoic acids, and cAMP (31)(32)(33). Additionally, mRNA levels of AP2 have been shown to increase dramatically upon differentiation, indicating that the expression of AP2 is regulated during differentiation (33). The promoter of the AP2 gene has a functional AP2 site, and a positive autoregulatory loop has been detected in this promoter (34). AP2 has also a crucial role in the induction of the antioxidant enzyme heme oxygenase 1 by heme (35). These observations suggest that common induction mechanisms exist in the cells for antioxidant enzymes. We interpret these results to mean that Rb 2 can accelerate autoregulation of transcription factor AP2 and that increased AP2 elevates the cellular amount of antioxidant enzymes such as SOD1. These results showed us a novel action mechanism of ginseng saponin on SOD1 transcription and could also provide a molecular link between ginseng saponin intake and its inhibitory effects on aging and mutation by radical oxygen.