The Activation of the Rat Copper/Zinc Superoxide Dismutase Gene by Hydrogen Peroxide through the Hydrogen Peroxide-responsive Element and by Paraquat and Heat Shock through the Same Heat Shock Element*

Copper/zinc superoxide dismutase (SOD1) protects cells against oxidative hazards by the dismutation of superoxide radicals. The promoter activity of theSOD1 gene was increased 3–5-fold by hydrogen peroxide, paraquat (PQ) and heat shock. Functional analyses of the regulatory region of the SOD1 gene by deletions, mutations, and heterologous promoter systems confirmed the induction of theSOD1 gene by H2O2 through the hydrogen peroxide-responsive element (HRE) (between nucleotides −533 and −520). Gel mobility shift assays showed that the existence of an H2O2-inducible protein bound to the oligonucleotide of the HRE. Similar analyses showed that the heat shock activated the SOD1 promoter through the heat shock element (HSE) (between nucleotides −185 and −171). A strong specific far-shifted complex with the oligonucleotide of the HSE was observed by the treatment of heat shock. When cells were treated with PQ, a strong far-shifted complex with the HSE was observed and was competed out by the cold HSE probe, indicating that PQ also activated theSOD1 promoter through the same HSE site. It is very interesting to note that chemical and physical stresses, such as PQ and heat shock, respectively, activated the SOD1 promoter through the same cis-element HSE. These results indicate that the SOD1 was inducible by H2O2through the HRE and by PQ and heat shock through the same HSE to protect cells from oxidative hazards.

Copper/zinc superoxide dismutase (SOD1) 1 is one of the major cellular defense enzymes that perform a vital role in protecting cells against the toxic effect of superoxide radicals. It catalyzes the dismutation of superoxide radicals (O 2 . ) to oxygen and hydrogen peroxide (1). SOD not only prevents the Fenton reaction and DNA nicking in vitro but also protects against toxicity by H 2 O 2 in vivo without O 2 . being directly involved in the generation of DNA damage (2). The production and/or removal of superoxides has been observed to play a significant role in a variety of critical homeostatic mechanisms both at the cellular and organismic levels. Because biological macromolecules are targets for the damaging action of abundant oxygen radicals, it is assumed that these increased superoxides should be initially eliminated by SOD. Therefore, the regulation and induction mechanism of the SOD1 gene would be of great interest (3). It has also been reported that SOD1 could prevent oncogenesis and tumor promotion (4), reduce the cytotoxic and cardiotoxic effects of anticancer drugs (5), and protect against reperfusion damage of ischemic tissue (6). A recent report suggested that overexpression of SOD1 and catalase could increase the average lifespan of the fly (7). Lutropin, Ca 2ϩ , and reactive oxygen seemed to induce SOD1 in rats (8 -10). Stress conditions (i.e. oxidative stresses, heat shock, osmotic stress, and toxic metals) are deleterious to normal cellular function. To survive those environmental and physiological stresses, all organisms possess specific defense systems to protect themselves from various stresses. Aerobic organisms are continuously exposed to oxygen, which renders them prone to damage generated by oxygen-derived free radicals. Oxidative stress is largely mediated by reactive oxygen species, including superoxide anion (O 2 . ), H 2 O 2 , and hydroxyl radical (OH ⅐ ), which are intermediates of oxygen reduction generated by metal ion catalyzed redox reactions, metabolism of chemicals, and normal physiological activities, including respiration and inflammatory responses to infection (11). The free radicals generated by these mechanisms cause severe damage to critical cellular macromolecules, including nucleic acids, proteins, and lipids (12). Oxidative damage has been strongly correlated with aging and a number of diseases, including Parkinson's disease, Lou Gehrig's disease (amyotrophic lateral sclerosis), rheumatoid arthritis, and cancer (11,13). Therefore, free radical levels must be carefully monitored under both physiological conditions and when generated by environmental stress. Cells employ a number of defense mechanisms to sense and respond appropriately to oxidative stress. Enzymes such as superoxide dismutases and catalases play critical roles in oxidative stress protection through catalyzing the conversion of reactive oxygen species to less harmful products (12,14). As has been demonstrated in yeast, cells lacking a functional gene encoding SOD1 are highly sensitive to dioxygen and redoxcycling drugs, fail to grow on respiratory carbon sources, and exhibit increased spontaneous mutagenesis rates and a number of other phenotypes during aerobic growth (15). Free radical scavenging activities are also exacerbated by small antioxidant molecules, including glutathione, thioredoxin, and ascorbic acid (12). In addition to the prevention of oxidative damage, repair mechanisms are employed by cells to remove or repair damaged cellular components, including exo-and endo-* This work was supported by a grant from Korea Science and Engineering Foundation through the Research Center for Cell Differentiation. 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.
Superoxide dismutase plays a key role in protection from the damage caused by oxygen radicals. The SOD1 expression is highly induced during environmental stresses. However, little is known about how the environmental signal is transmitted to a transcriptional regulator in cells during oxidative stress such as hydrogen peroxide, paraquat (PQ) and heat shock. In this study, the induction of the SOD1 gene by hydrogen peroxide, PQ, and heat shock was investigated by the transfection of CAT expression vectors containing the rat SOD1 promoter. The deletion, mutation, and heterologous promoter systems were used to identify cis-elements that respond to hydrogen peroxide, PQ, and heat shock for minimizing hazardous effects. A specific interaction of the hydrogen peroxide-responsive element (HRE) binding protein and heat shock element (HSE) binding protein (heat shock factor) was also investigated by gel mobility shift assays. The results demonstrated that the SOD1 was inducible by H 2 O 2 through the HRE and by PQ and heat shock through the same HSE to protect cells from oxidative hazards.
Cell Culture and Transfection-Human HepG2 hepatoma cells were grown in Dulbecco's modified Eagle's medium/10% fetal calf serum/ penicillin G sodium at 100 units/ml/streptomycin sulfate at 100 g/ml/ amphotericin B at 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 (19). 5 g of pRSV␤-gal plasmid (20) was introduced in all experiments to correct the variations of transfection efficiency.
Chemical and Heat Shock Treatment and Nuclear Extract Preparation-Chemicals were added to and heat shock treatment was carried out on the culture medium at 36 h after transfection, and the cells were maintained for appropriate times. The various chemicals were evaluated for cytotoxicity over a concentration range of 1-500 M by monitoring cell death. Nuclear extracts were prepared by a modified procedure of Andrews and Faller (21).
CAT Assay-The CAT assay was performed as described previously (22). The transfected cells were washed twice with phosphate-buffered saline 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 removing 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 chlor-amphenicol to acetylated form was calculated by measuring radioactivities. The relative CAT activities were calculated from the percentage conversion. Results are the average of three independent experiments.
Mobility Shift Assay-The oligonucleotides for the HRE and HSE sites were synthesized and labeled with [␥-32 P]ATP and polynucleotide kinase (23). An equal amount (10 g) of nuclear extract from each sample was mixed with labeled 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). The binding reaction mixtures were electrophoresed in 6% acrylamide gels in 0.5ϫ TBE (44 mM Tris, 44 mM boric acid, and 1 mM EDTA). After electrophoresis, the gels were dried and exposed to x-ray film. For competition assays, a binding reaction was performed with an excess of cold probe or competitor DNA. In the supershift assay, the specific polyclonal antibody against Elk1 (Santa Cruz Biotech. Inc) was added to the reaction mixture after the binding reaction.

RESULTS
Activation of the SOD1 Promoter by Hydrogen Peroxide, PQ, and Heat Shock-SOD1 is one of the major cellular defense enzymes protecting cells against oxidative stress. It was hypothesized that hydrogen peroxide, PQ, and heat shock are involved in the activation of the SOD1 promoter. H 2 O 2 was used as a membrane-permeable reagent, which allows studies of the effects of oxygen radicals in living cells. If hydrogen peroxide is inefficiently removed or if an excess is present, it can react with superoxide to generate the highly toxic hydroxyl radical. PQ is a redox-active compound that is photoreduced and subsequently reoxidized by transfer of its electrons to oxygen, generating superoxide (O 2 . H 2 O 2 , a 3-fold induction of CAT activity was observed. No induction was observed in cells bearing the plasmid pRSP-412 (Fig. 3C). These results mean that the activation of the SOD1 gene by H 2 O 2 may be due to the existence of DNA elements between Ϫ576 and Ϫ412. The induction by PQ was observed in both transfection experiments with pRSP-576 and pRSP-412 (data not shown). It seems that the cis-element corresponding to induction by PQ was located downstream of Ϫ412 of the SOD1 gene. When the SOD-CAT construct (pRSP-305) was transiently expressed in HepG2 cells, 4-and 3-fold induction of CAT activity by heat shock (42°C) and PQ was observed, respectively (Fig. 3D). No induction was observed in cells bearing the plasmid pRSP-55 (Fig. 3D). These results mean that the activation of the SOD1 promoter by heat shock and PQ may be due to the existence of DNA elements between Ϫ305 and Ϫ56.
cis-Element Responsible for the Activation of SOD1 Promoter by H 2 O 2 -When the plasmids pRSP-576 and pRSP-412 were transfected into HepG2 cells and exposed to H 2 O 2 , the induction with H 2 O 2 was observed only in the case of transfection with pRSP-576 (Fig. 3C). However, the induction with PQ was observed in both cases (data not shown). These results might suggest that H 2 O 2 and PQ regulate the SOD1 gene by different regulatory elements and factors. The region between Ϫ576 and Ϫ413 was previously identified as a positive regulatory element (PRE) (24). As an initial trial to identify the transcriptional enhancer element that was involved in H 2 O 2 induction of the SOD1 gene, transient assays using CAT constructs were carried out. The plasmid pPREtk, which has the PRE linked to a herpes simplex virus-thymidine kinase promoter connected to the CAT gene, and pRSP-576 as a positive control were introduced into HepG2 cells, and the cells were treated with H 2 O 2 for 2 h. The CAT activity was measured, and typical results are shown in Fig. 4B. The results clearly indicate that the CAT activity of pPREtk increased about 5-fold with H 2 O 2 . The PRE was thought to be a major target for H 2 O 2 stimuli to induce the SOD1 gene. From the sequence analysis of transcription factor binding sites in the PRE, the sequence of Elk1, which is the serum response factor accessory protein (25), was located between Ϫ533 and Ϫ520 of the SOD1 gene (24). Induction by H 2 O 2 of c-fos and possibly other early response genes was mediated through the activation of the serum response element for the binding site of serum response factor (26). The Ϫ533 to Ϫ520 sequence named as the HRE was tested for induction ability by H 2 O 2 . Double-stranded oligonucleotide containing the HRE sequence was inserted upstream of the thymidine kinase promoter connected to the CAT gene (Fig.  5A). CAT activity was induced by H 2 O 2 in cells transfected with pHREtk by 3-fold, but CAT activity was not induced by H 2 O 2 when pmHREtk containing a mutated HRE site was trans-fected (Fig. 5C). These results confirmed that the HRE is responsible for the activation of the SOD1 promoter by H 2 O 2 .
Heat Shock and PQ Activation of the SOD1 Promoter through the Same HSE-The activation of the SOD1 gene by heat shock and PQ was due to the DNA sequences between Ϫ305 and Ϫ56 (Fig. 3D). From the sequence analysis of transcription factor binding sites in this region, the heat shock element (NGAAN) (27), was found to be located between Ϫ185 and Ϫ171 of the SOD1 gene (Fig. 1A). Heat shock treatment of murine macrophage resulted in an enhanced capacity to release superoxide anion (O 2 . ) (28). Also, Omar and Pappolla (29) showed that heat shock protein synthesis was induced by high levels of superoxide anion. Therefore, it was assumed that the activation of the SOD1 promoter by heat shock and PQ may be mediated through the heat shock element. To confirm this possibility, the heat shock element was synthesized and ligated to the herpes simplex virus minimal thymidine kinase promoter proximal to the CAT gene (Fig. 5B). The CAT activity of this construct was analyzed after exposure of cells to either heat shock or PQ (Fig.   FIG. 2. Time- 5D). When pHSEtk was transfected into HepG2 cells, CAT activity was induced by heat shock or PQ by about a 5-fold. But the CAT activity was not induced by heat shock or PQ when pmHSEtk containing the mutated HSE site was transfected. These findings indicate that the HSE is sufficient to confer the PQ response and that this inducibility is mediated by HSF. These results establish that the HSF play some kind of role in eukaryotic oxidative stress protection and suggest that heat and oxidative stress signals communicate with the HSF to activate gene transcription.
HRE Binding Protein and Increased Binding Activity Because of H 2 O 2 -To determine the nuclear factors that interact with the HRE sequence, gel mobility shift assays were carried out by using a double-stranded oligonucleotide corresponding to HRE as a probe (Fig. 6). When a nuclear extract prepared from HepG2 cultures was incubated with a 32 P-labeled SOD1 HRE probe, one prominent DNA-protein complex was observed. An HRE oligonucleotide-protein complex disappeared when unlabeled oligonucleotide (50ϫ, 100ϫ) was added to the reaction mixture (Fig. 6A, lanes 3 and 4). The formation of the protein-DNA complex was not inhibited with a mutated HRE oligonucleotide, indicating the specificity of this protein-DNA interaction. The binding of Elk1 to the HRE was further confirmed by the result that the protein-DNA complex was supershifted by the addition of the anti-Elk1 antibody (Fig. 6A, lane  6). These results imply that Elk1 is the binding protein of HRE. When nuclear proteins prepared from the H 2 O 2 -treated HepG2 cells (2 h) were used, the intensity of the specific complex increased (Fig. 6B). The binding activity of the HRE-binding factor was also investigated in the cell-free system (Fig. 6C). Nuclear extract prepared from nontreated HepG2 cells was incubated with H 2 O 2 (200 M) to examine the variation of its binding activity to the HRE site. The incubation of the nuclear extract with H 2 O 2 resulted in increasing the intensity of the retarded complex (Fig. 6C, lane 3) and disappearance by competition with the cold probe (Fig. 6C, lane 4). Because the intensity of the specific HRE-protein complex was increased by H 2 O 2 treatment in the cell-free system (Fig. 6C, lane 3), it was also assumed that hydrogen peroxide treatment increased the binding activity of the HRE binding protein.
HSE Binding Protein (Heat Shock Factor) and Change of Binding-Gel mobility shift assays were used to detect se-quence-specific binding of factors to the SOD1 HSE sequence in extracts from HepG2 cells. Nuclear extracts were prepared from cells that had been heat treated at 42°C for 1 h or maintained at the control temperature of 37°C. These extracts were incubated with a 5Ј end-labeled oligonucleotide containing an HSE sequence (between sequence Ϫ185 and Ϫ171 of the SOD1 gene). In reactions with extracts from the control cells, two sets of retarded bands were observed (Fig. 7). Bands produced by sequence-specific factor binding were identified by competition assays with unlabeled DNA fragments. Unlabeled HSE fragments competed efficiently with the formation of a faster-migrating protein-DNA complex (Fig. 7, lanes 3 and 4). No such competition was observed with the mHSE oligonucleotide (Fig. 7, lane 5). The formation of the more slowly migrating complex was not in competition with the HSE fragment. Thus, the latter complexes did not result from sequence-specific factor binding to the HSE sequences. When nuclear proteins prepared from the heat-treated cells were used, a strong specific far-shifted complex was observed. This strong new complex disappeared by self-competition (Fig. 7, lanes 9 and 10). The HSF in eukaryotes acquires high affinity DNA binding activity upon heat shock activation by conversion from a monomer to homotrimer (30). Therefore, the new far-shifted complex may be a homotrimer HSF-DNA complex. Analogous results were obtained from experiments with nuclear extracts from cells treated with PQ for 1 h (Fig. 8A). A far-shifted complex disap- peared by self-competition, demonstrating a specific interaction (Fig. 8A, lane 5). These results indicate that the SOD1 gene protects cells from PQ toxicity in an HSF-dependent manner and that heat and oxidative stress signals communicate with HSF to activate gene transcription. The far-shifted protein complex disappeared after reequilibration to 37°C for 1 h following a heat shock (Fig. 8B, lane 4). These results also showed that the transcription of the SOD1 gene is induced by heat shock through the same binding of the HSF to the HSE site. DISCUSSION In transient expression experiments the activation of the SOD1 promoter by H 2 O 2 was mediated through the HRE sequence. To analyze the target of transcriptional activation of the SOD1 gene, CAT plasmids were constructed that contain these elements in the upstream region of the thymidine kinase promoter. From these results, it was shown that the HRE sequence is one of the targets of transcriptional activation by H 2 O 2 . Also, we have found that superoxides could activate the SOD1 gene. Because the steady state concentration of H 2 O 2 in the cell is so low (1 ϫ 10 Ϫ8 M) and specific enzymes (i.e. catalase, glutathione peroxidase, and myeloperoxidase) actively convert H 2 O 2 to H 2 O in a fast speed, the resulting hydrogen peroxide by SOD1 is thought not to activate the SOD1 gene by the positive feedback mechanism. In this experiment, the promoter activation was observed at the concentration around 100 M of H 2 O 2 . Active oxygens are suggested to be involved in inflammatory responses; the activations by H 2 O 2 of c-fos, c-jun, and egr-1, which activate the expression of cellular genes, are probably essential for these responses. DNA damaging agents have been reported to activate c-fos and AP1-inducible genes such as those for collagenase and metallothionein (31). Active oxygens such as O 2 . or H 2 O 2 are produced in mitochondria during oxidative electron transport and in the endoplasmic reticulum, peroxisomes, and nuclear and plasma membranes. The active oxygens generated in these locations may regulate certain cellular genes by inducing early response genes.
To mount an appropriate oxidative stress defense, cells must harbor oxidative stress sensors. A number of antioxidant responses in bacteria have been elegantly studied, and it has been established that reactive oxygen species are directly sensed by key regulatory molecules that activate the expression of genes encoding antioxidant proteins at the level of transcription (32)(33)(34) (34). The SoxRS regulon is controlled in a two-stage process. First an iron-sulfur protein, SoxR, is activated by increases in intracellular superoxide anion levels and triggers transcription of the soxS gene. The SoxS protein in turn induces transcription of other genes of the regulon, including those encoding manganese superoxide dismutase, the DNA repair endonuclease IV, and glucose-6-phosphate dehydrogenese (32,33,35). The oxidative stress response and its relationship to heat shock phenomena are also being intensely investigated in eukaryotic systems. Unlike the bacterial systems described earlier, little is known about how eukaryotic cells co-ordinate gene expression in response to oxidative stress (34). Heat shock proteins are known to protect cells from thermal and oxidative injuries as well as other types of injuries (36). In this study, we have found that the paraquat, O 2 . -generating agent, could activate the SOD1 gene through the heat shock element. Heat shock may cause activation of the membraneassociated oxidase system directly or indirectly through HSP, resulting in increased O 2 . production. Oxygen free radicals induce heat shock protein synthesis in cultured human neuroblastoma cells (29). These findings suggest a common mechanism by which various forms of injury, such as hyperthermis, cause HSP induction, that is, via oxidative stress or increased production of oxygen free radicals.
The results presented in this study showed that the rat SOD1 gene was inducible by hydrogen peroxide, paraquat, and heat shock. It was shown that the HRE sequence is one of the targets of transcriptional activation by H 2 O 2 . Also, it was found that paraquat and heat shock induced the SOD1 gene through the HSE. These results establish that heat and PQ are communicated to HSF to activate gene transcription. The SOD1 enzyme could function as an inducible protective response against oxidative stresses. The SOD1 gene was induced by nontoxic ginsenoside Rb 2 through the AP-2 site (3). The diver-  8 -12). The DNA-protein complex disappeared by cold competitor (lanes 3, 4, 9, and 10) but did not disappear with mutated HRE (mHRE) (lanes 5 and 11) and nonspecific competitor (lanes 6 and 12). Using nuclear extracts prepared from heat-treated cells, a far-shifted DNA-protein complex appeared. The location of specific DNA-protein complex is indicated by an arrowhead. Note that the induced strong bands disappeared with the cold probe.
FIG. 8. The change of binding of HSE binding protein (HSF) by PQ and heat shock. A, a gel mobility shift assay was performed using nuclear extracts prepared from untreated and PQ-treated HepG2 cells. The DNA-protein complex disappeared by cold competitor (lane 3). When nuclear extracts from the cells treated with PQ (50 M) for 1 h were used, a far-shifted band appeared (lane 4). B, the induction of a far-shifted band by heat shock in vitro. The ϩ/Ϫ represents that reequilibration to 37°C following a heat shock (42°C). The location of specific DNA-protein complex is indicated by an arrowhead. sity of SOD1 inducers implies that there are multiple regulatory elements for the proper adjustment to various conditions.