Transcriptional Activation of the Human Manganese Superoxide Dismutase Gene Mediated by Tetradecanoylphorbol Acetate*

Transcriptional activation of human manganese superoxide dismutase (MnSOD) mRNA induced by a phorbol ester, 12-O-tetradecanoylphorbol-13-acetate (TPA), was examined to identify the responsive transcriptional regulator. The effect of various deletions and mutations within the 5′-flanking region of the human MnSOD gene promoter was evaluated using the luciferase reporter system in A549 human lung carcinoma cells. Deletion of a region between −1292 and −1202 nucleotides upstream of the transcription start site abolished TPA-responsive induction, whereas deletion of the putative binding sequence for NF-κB or AP-1 did not. The region between −1292 and −1202 contains a cAMP-responsive element-like sequence, TGACGTCT, which we identified as the manganese superoxide dismutase TPA-responsive element, MSTRE. Site-specific mutation of the MSTRE abolished the TPA-responsive induction, validating the critical role of this sequence. We detected specific MSTRE activity from nuclear extracts and demonstrated by antibody supershift assay that this activity is closely related to CREB-1/ATF-1. TPA treatment rapidly induced phosphorylation of the CREB-1/ATF-1-like factor via the protein kinase C pathway. These results led us to conclude that the human MnSOD gene having the promoter construct used in this study is induced by TPA via activation of a CREB-1/ATF-1-like factor and not via either NF-κB or AP-1. In addition, we found that this induction was blocked by inhibitors of flavoproteins and NADPH oxidases, indicating involvement of enhanced generation of superoxide radical anion as an upstream signal.

In eukaryotes, manganese superoxide dismutase (MnSOD) 1 is located in the mitochondrial matrix, which is encoded by a nuclear gene. Mitochondria are particularly prone to oxidative DNA damage because they metabolize over 95% of the oxygen (2), but they lack histones and have a poor ability for DNA repair (3). Thus, MnSOD is regarded as the primary defensive enzyme against oxidative stress within mitochondria. Loss of MnSOD is implicated in causing various human tumors (4 -6), whereas its overexpression suppresses tumorigenicity, presumably by decreasing the concentration of O 2 . , which promotes cancer (7)(8)(9)(10).
A number of studies demonstrated the induction of MnSOD in various cell lines and tissues following oxidative stress, such as treatments with TNF-␣ (11,12), interleukin-1 (12), lipopolysaccharide (13), interferon-␥ (14), 12-O-tetradecanoylphorbol-13-acetate (TPA) (15)(16)(17), and x-irradiation or hyperoxia (18,19). Most of these treatments could generate reactive oxygen species in cells that, in turn, could activate transcription factors, mainly NF-B and AP-1 (20,21), to allow their nuclear translocation and allow them to bind to genes involved in the stress responses. In accordance, the involvement of NF-B (22)(23)(24) and AP-1 (25) in activating MnSOD genes in response to TNF-␣, TPA, H 2 O 2 , or thiol-reducing agents has been suggested in human and rat cells. These studies correlated the conditions that changed the DNA binding activity of either NF-B or AP-1 with the induction of MnSOD, and thus concluded that the activation of the MnSOD gene is mediated by either of these factors. However, the responsive cis-acting elements, as well as the changes in or modification of the binding factor, have not been directly demonstrated. Recently, Jones et al. (26) have shown that the mouse MnSOD gene is regulated by TNF-␣ and interleukin-1 through a complex intronic enhancer located in intron 2 involving binding of C/EBP and NF-B. They also found that the 5Ј-flanking promoter region of the MnSOD gene is not involved in the TNF-mediated MnSOD mRNA induction. However, relatively little information is available regarding the molecular mechanism for modulating the expression of the MnSOD gene, SOD2, by TPA.
In this study, we present evidence that the TPA-responsive element in the induction of MnSOD gene expression is a cAMPresponsive element (CRE)-like element bound to a CREB-1/ ATF-1-like factor, which is located in the 5Ј-flanking DNA sequence. The activation of the CREB-1/ATF-1-like factor is caused by rapid phosphorylation via the protein kinase C (PKC) pathway. In addition, NADPH oxidase or similar flavoproteins are also involved in the induction of the MnSOD gene, which may indicate that superoxide radical anion is an upstream signal for this induction.

EXPERIMENTAL PROCEDURES
Cell Lines-The human lung adenocarcinoma cell line A549 was purchased from American Type Culture Collection (Rockville, MD). Cells were cultured in RPMI 1640 supplemented with 10% fetal bovine serum (Life Technologies, Inc.) at 37°C under 5% CO 2 .
Cloning and Nucleotide Sequencing of the 5Ј-Flanking Region of the Human MnSOD Gene-A human lymphocyte genomic library constructed in Dash TM vector (Stratagene, La Jolla, CA) was screened with the human MnSOD cDNA probe according to the standard procedure (27). One positive clone was further analyzed. The nucleotide sequence of the 3.6-kilobase pair BamHI fragment containing 3.3 kilobase pairs of the promoter upstream region was determined.
Construction of Luciferase Reporter Plasmids-The coding region of the firefly luciferase gene was ligated with the 5Ј-flanking region of the human MnSOD gene in pGL3-Basic vector (Promega, Madison, WI). The 3.6-kilobase pair BamHI fragment containing the 5Ј-flanking region, exon 1, and a part of the intron 1 (from Ϫ3340 to ϩ260 nucleotide residues) of the MnSOD gene was ligated with pGL3-Basic vector linearized with BglII to generate pSODLUC-3340. A series of 5Ј deletions were generated by digestion with exonuclease III of pSODLUC-3340 cut with KpnI and XhoI located at multiple cloning sites. The end point of each deletion was determined by dideoxy chain-termination sequencing.
Transfection and Luciferase Assay-A549 cells were transfected with pSODLUC plasmids by lipofection using LipofectAMINE reagent (Life Technologies, Inc.) as recommended by the manufacturer. The pRL-TK (Promega, Madison, WI) plasmid containing Renilla luciferase driven by thymidine kinase promoter was co-transfected as a transfection control. The pGL3-Control (Promega) containing SV40 promoter-enhancer and promoter-less pGL3-Basic plasmids were transfected as positive and negative controls of promoter activity, respectively. Twenty-four hours after transfection, TPA (20 ng/ml) was added and cells were incubated for another 12 h. The firefly and Renilla luciferase activities were measured using the Dual Luciferase assay system (Promega). The pSODLUC-3340 plasmid yielded an 8-fold enhancement in luciferase activity relative to that of the pGL3-Basic vector. It represents 1.6% of the luciferase activity observed with the SV40 promoterenhancer in the pGL3-Control plasmid.
Site-directed Mutagenesis-A CRE-like element (TGACGTCT; Mn-SOD TPA-responsive element, MSTRE) at Ϫ1258 was mutated using mutagenic primers as described by Yim et al. (28). Polymerase chain reaction with the two internal mutagenic primers and flanking primers generated mutagenized DNA fragments encompassing the SacI site at Ϫ1993 and the SpeI site at Ϫ825 with the MSTRE sequence changed to TGTGGTCT (MSTREmt, with mutated residues underlined). The internal BstxI/SpeI fragment of pSODLUC-1517 was replaced with the mutated polymerase chain reaction fragment to generate pSODLUC-1517mt.
Immunoblotting-Proteins in A549 nuclear extracts (10 g/lane) were electrophoresed on 4 -20% gradient polyacrylamide gel containing SDS and transferred onto polyvinylidene difluoride membrane. Blots were incubated with rat antibodies against either CREB, phospho-CREB, or phospho-ATF-1 (New England Biolabs, Beverly, MA) overnight at 4°C. Anti-rabbit IgG secondary antibody conjugated with horseradish peroxidase (New England Biolabs) was then incubated and visualized using a Phototope-HRP Western blot detection kit (New England Biolabs).
Northern Analysis-Total cellular RNA was prepared from A549 cells using Trizol reagent (Life Technologies, Inc.) (30). The total RNA (10 -20 g) were run on 1% agarose gel containing 2.2 M formaldehyde. The size-fractionated RNA was transferred to Nytran Plus (Schleicher & Schuell). After hybridization with 32 P-labeled probe at 68°C in the presence of Quickhyb hybridization solution (Stratagene, La Jolla, CA), the membrane was washed twice at room temperature in 2ϫ saline sodium citrate (SSC) and 0.1% SDS for 10 min each, and then washed at 60°C in 0.1ϫ SSC and 0.1% SDS for 15 min. The Kodak X-Omat AR films were exposed for 2-12 h with an intensifying screen at Ϫ70°C. The intensity of a band was quantitated by the PhosphorImager Storm 860 (Molecular Dynamics, Sunnyvale, CA).

Nucleotide Sequence Analysis of the Promoter Region-
The newly determined 5Ј-flanking region sequence of the human MnSOD gene was deposited in GenBank/EMBL data base under accession number AF059197. This promoter region of the human MnSOD gene up to Ϫ3340 residues was searched against the transcription factor data base for putative transcriptional regulatory elements using SIGSCAN programs (version 4.05) (31). The analysis did not reveal any TATA or CAAT boxes consistent with previous observations with bovine, rat, and human MnSOD genes (32)(33)(34)(35)(36). It contains a GC-rich (75%) region within 500 base pairs of the transcription start site, which is preceded by nine potential binding sites for Sp1 and six sites for AP-2. Several other potential regulatory elements were also detected. Two copies of the AP-1 binding sequence were found at positions Ϫ2883 and Ϫ820, two copies of the NF-B binding sequence were found at positions Ϫ3215 and Ϫ1551, and one copy of the CREB/ATF family-binding sequence was found at position Ϫ1258.
The TPA-induced MnSOD Promoter Activity-In order to identify the cis-acting elements responsible for TPA-induced transcription of the MnSOD gene, a series of 5Ј deletions of the promoter fragment from Ϫ3340 were subcloned to the pGL3-Basic vector containing the coding sequence of the firefly luciferase gene (Fig. 1) as described under "Experimental Procedures." The resultant pSODLUC plasmids were transfected into A549 cells and the promoter activities were measured by luciferase assay. The results depicted in Fig. 1A show that cells transfected with pSODLUC-3340 induced MnSOD gene expression by 2-fold upon treatment with TPA (20 ng/ml). However, similar treatment of these cells with TNF-␣ (50 ng/ml) did not show any increase in the luciferase activity. This result indicates that the regulation of the human MnSOD gene by TNF-␣ involves parts of the gene other than the Ϫ3340 base pairs 5Ј-flanking region. This finding is in agreement with previous results obtained in the mouse MnSOD gene (26), demonstrating that the gene is regulated by TNF-␣ through intronic enhancer, but not through the 5Ј-flanking region. Deletion of sequences between Ϫ3340 and Ϫ1396 (Fig. 1B), which includes two NF-B and one AP-1 binding sequences, did not significantly affect TPA-induced MnSOD transcription in A549 cells in terms of folds induced. Further deletion from Ϫ1292 to Ϫ1202 resulted in abrogation of the induction by TPA stimulation. The induction ratio decreased from 2.8 to 1.3, suggesting the presence of the TPA-responsive element within this region. The region contains a putative CRE-like sequence (5Ј-TGACGTCT-3Ј motif) between Ϫ1258 and Ϫ1251.
Deletion of sequences between Ϫ3340 to Ϫ263 did not significantly affect the basal activity of the MnSOD promoter in A549 cells. Further 5Ј deletions to nucleotides Ϫ106 and Ϫ1 resulted in a 50% and 90% decrease, respectively, in the basal activity of the MnSOD promoter. This result may indicate the presence of a minimal promoter between Ϫ262 and Ϫ1, the region that contains six SP1 and three AP2 binding sites.

Identification of a TPA-responsive Element in the MnSOD
Promoter-To test whether this CRE-like sequence is responsible for the induction by TPA treatment, two nucleotides of this motif were mutated in the pSODLUC-1517 to create the plasmid pSODLUC-1517mt. Expression from this construct was analyzed by luciferase assay (Fig. 2). The mutation resulted in a substantial decrease in TPA-promoted induction with little effect on basal expression, indicating that this CRElike element is a TPA-responsive element in MnSOD gene (MSTRE).
The binding activity of the MSTRE in a nuclear extract of A549 cells was examined by electrophoretic mobility shift assay using the oligonucleotide MSTRE. Formation of several prominent complexes was observed (Fig. 3A, lane 1), among which complexes I, II, and III were suggested to arise from specific binding to the MSTRE motif, since they were not competed off by the mutated MSTRE probe (Fig. 3A, lane 3). Competition with the CREB-binding consensus sequence (TGACGTCA) abolished these complexes (Fig. 3A, lane 4). The AP-1-binding consensus sequence (TGAGTCA) competed partially (lane 5), whereas the Oct-1-binding sequence (TGCAAAT) did not (lane 6). These results suggest that the specific binding to MSTRE is due to a CREB-like factor, which can also recognize the AP-1 consensus sequence with lower affinity.
In order to identify the specific transcription factor that binds to MSTRE, we performed antibody supershift analysis. As shown in Fig. 3B, the antibody against both CREB-1 and ATF-1 specifically abolished complexes I and II, decreased the intensity of complex III, and created slower-moving supershifted complexes (lane 2). More specific antibody that recognized CREB-1 but not ATF-1 completely abolished complexes I and II (lane 3), while the antibody against ATF-1 abolished complex II and decreased the intensity of complex III (lane 4). A longer exposure of the gel also showed the supershifted bands in lanes 3 and 4, although they are much weaker than that in lane 2 (data not shown). In contrast, antibodies against ATF-2, c-Jun, and c-Fos, did not affect the binding pattern (lanes 5-7). Therefore, it is most likely that the MSTRE-specific binding factors are the homodimers of CREB-1 and ATF-1 (for complex I and complex III, respectively), heterodimer of CREB-1/ATF-1 (for complex II), or their very close relatives. AP-1 complexes composed of c-Jun and c-Fos are not likely to be the MSTREspecific binding factors.
CREB Phosphorylation Versus TPA-mediated MnSOD mRNA Induction-To investigate the mechanism whereby MSTRE mediates TPA-induced transcription of the MnSOD gene, we examined the changes in MSTRE binding activity upon treating A549 cells with TPA for different lengths of time (Fig. 4A). The retarded band pattern did not change at all following TPA treatment, clearly demonstrating that TPA does not alter the binding activity of a CREB-1/ATF-1-like factor to MSTRE.
Phosphorylation of transcription factors is one of the major mechanisms for regulating their activities (37). It can regulate the activity of transcription factors by affecting nuclear transport, dimerization, DNA binding, and/or transcriptional activation (37). CREB is known to be regulated by phosphorylation of its Ser 133 by protein kinase A (PKA) upon stimulation of the adenylyl cyclase pathway, resulting in a dramatic increase in its transactivating potential (38). To examine whether TPA also stimulates phosphorylation of CREB at Ser 133 in A549 cells, we probed with an antibody specific for the phosphoryl- ated Ser 133 of CREB. As shown in Fig. 4B (upper panel), little if any phospho-CREB is present in untreated A549 cells (lane 2). However, treatment with TPA (20 ng/ml) increased CREB phosphorylation at Ser 133 within 15 min (lane 3), which slowly disappeared afterward. The antibody against CREB phosphorylated at Ser 133 also reacted with a 38-kDa protein (marked as p-ATF-1 in Fig. 4B) that co-migrates with ATF-1, a related transcription factor that shares amino acid sequence identity surrounding Ser 133 and cross-reacts with the antibody used. The total amount of CREB protein in the same nuclear extracts did not change, as judged by blotting with antibody that recognized CREB, regardless of their phosphorylation status at Ser 133 (Fig. 4B, lower panel).
Two kinds of specific protein kinase inhibitors were used to examine whether the phosphorylation of the CREB-1/ATF-1-like factor is directly related to the TPA-mediated MnSOD mRNA induction. The H-89 (K i for PKA ϭ 48 nM; K i for PKC ϭ 32 M) and bisindolylmaleimide (BIM) (K i for PKC ϭ 10 nM; K i for PKA ϭ 2 M) are highly selective and cell-permeable inhibitors of protein kinase A and C, respectively (39 -41). Preincubation of A549 cells with BIM (1 M) inhibited the TPAmediated MnSOD mRNA induction by 80% (Fig. 5A). In contrast, H-89 (up to 4 M) had little effect on the induction (Fig. 5B). These results suggest that TPA treatment induces MnSOD mRNA through the PKC rather than the PKA pathway.  (Fig. 6A). AEBSF, a serine protease inhibitor, was reported to prevent the activation of NADPH oxidase in both intact macrophases and in cell-free systems by interfering with the interaction of p47 phox and/or p67 phox with cytochrome b 558 (46). Treatment of the cells with AEBSF also blocked TPAmediated induction of MnSOD mRNA (Fig. 6B). Similar effects were also observed in MnSOD promoter activity determined by luciferase assay (data not shown). These results together suggest that an upstream signal for TPA-mediated induction of MnSOD mRNA may be related to O 2 . generated by NADPH oxidase or similar oxidases activated by TPA treatment.

DISCUSSION
The results obtained in this study allowed us to identify a responsive element for TPA-induced MnSOD mRNA, the MSTRE, at the 5Ј-flanking region of MnSOD gene (Figs. 1 and  2). The TPA-responsive induction of MnSOD mRNA requires binding of the MSTRE to the CREB-1/ATF-1-like transcription factor (Fig. 3), which is activated by PKC-catalyzed phosphorylation (Figs. 4 and 5). This induction was blocked by inhibi- A, effect of TPA on the MSTRE binding activity. Nuclear extracts were prepared from A549 cells treated with 20 ng/ml TPA for 0, 15, 30, 60, and 120 min. Gel mobility shift assay was done as in Fig. 3. B, effect of TPA on phosphorylation of CREB at Ser 133 . The same nuclear extracts used in panel A were analyzed by Western blotting using antibody against phosphorylated Ser 133 of CREB (P-CREB), which can cross-react with phosphorylated Ser 63 of ATF-1 (P-ATF-1). The amount of total CREB in each sample was detected by Western blotting using antibodies that recognized CREB regardless of phosphorylation status.  1 and 4) or presence of BIM (lanes 2, 3, 5, and 6) for 1 h. Total RNA was prepared from cells after incubation with medium only (lanes 1-3) or 20 ng/ml TPA (lanes 4 -6) for 4 h. Northern analysis of the RNA was carried out using 32 P-labeled MnSOD or glyceraldehyde-3-phosphate dehydrogenase probes. B, effect of a PKA-specific inhibitor, H-89. A549 cells were incubated in the absence (lanes 1 and 4) or presence of H-89 (lanes 2, 3, 5, and 6) for 1 h. Other procedures are identical to those in A. tors of NADPH oxidase or related flavoproteins (Fig. 6), suggesting the involvement of enhanced generation of superoxide radical anion as an upstream signal for the TPA-mediated MnSOD induction.
It has previously been demonstrated that TPA activates NF-B and AP-1 via the PKC pathway (20,21,47,48). These previous reports, as well as the presence of consensus binding motifs for NF-B and AP-1 in the 5Ј-flanking region of the MnSOD gene, suggested that induction of the MnSOD gene by TPA treatment might be mediated by NF-B and/or AP-1. The fact that these factors are also activated by pro-oxidants or anti-oxidants (22)(23)(24)(25) provided a conceptual link between the TPA treatment that could generate reactive oxygen species and the mechanism of MnSOD induction that converts one form of reactive oxygen species (O 2 . ) into another (H 2 O 2 ). To our surprise, however, deletion of putative binding sites for NF-B and AP-1 within 3340 nucleotides of the 5Ј-flanking region did not affect the TPA-induced MnSOD gene. Instead, our results indicate that a CRE-related sequence, MSTRE, bound to the CREB-1/ATF-1-like factor and controlled by phosphorylation, is responsible for the TPA-induced MnSOD gene in this promoter construct. The similarity between the CRE (TGACGTCA) and the TPA-responsive element, TRE (TGAGTCA), sequences may allow cross-talk between components of the two major signal transduction pathways. Indeed, it has been shown that AP-1 can bind to both TRE and CRE, which leads to the activation of CRE-dependent transcription in vivo (49)(50)(51). Although we observed enhanced AP-1 binding to its consensus TRE-binding motif following TPA treatment of A549 cells and a partial competition of MSTRE-bound complex by excess cold TRE-containing probes, our data obtained by antibody supershift analysis also clearly demonstrated that the AP-1 complex composed of Jun/Fos does not bind to MSTRE to mediate TPA-induced MnSOD gene (Fig. 3B). The ubiquitously expressed CREB protein is a well characterized transcription factor regulated by phosphorylation, in most cases, via the PKA and PKC pathways (52)(53)(54). Phosphorylation in response to cAMP stimulation has no effect on the ability of CREB to bind CRE, but it enhances transcriptional efficiency 20-fold in vitro (51,55). In addition, it has been shown that the T cell antigen receptor engagement resulted in the rapid phosphorylation of CREB on Ser 133 via PKC pathway and its concomitant activity (53). Our results also clearly demonstrated that there is no change in the abundance or binding activity of CREB-1/ATF-1 complexes following TPA treatment of A549 cells (Fig. 4). This result is consistent with the previously observed behavior of activated CREB (43). This transcription factor can bind to DNA as a homodimer or can form a heterodimer with other bZip proteins, such as ATF-1, which share many structural properties with CREB (56). Our results demonstrated that ATF-1 as well as an CREB-1-like factors bound to MSTRE become phosphorylated upon TPA treatment and suggest that heterodimeric as well as homodimeric complexes of ATF-1 and CREB-1-like factors are involved in inducing the MnSOD gene. In addition, our results demonstrated that the specific inhibitor for PKC, BIM, rather than the PKA inhibitor, blocked the TPA-mediated induction of MnSOD mRNA (Fig. 5). The previous report (53), as well as our preliminary results, showed that BIM also inhibits TPA-responsive phosphorylation of CREB. These findings together suggest that the TPA-responsive MnSOD mRNA induction is mediated mainly by the phosphorylation of CREB-1/ATF-1-like factors via the PKC pathway.
Superoxide radical anions are generated by NADPH oxidase in phagocytes as a host defense. In a resting state, NADPH oxidase assembly in phagocytes consists of several proteins: p47 phox , p67 phox , and Ras-related GTP binding protein, in cytosol, and gp91 phox and p22 phox , which are two components of a flavoprotein cytochrome b 558 , located in plasma membrane. During activation, p47 phox is phosphorylated and translocated together with other cytosolic components to the plasma membrane, where they interact with a flavoprotein cytochrome b 558 to form an active complex that catalyzes the formation of O 2 .
from oxygen (42,57). In other type of cells, such as fibroblasts, the NADPH oxidase-like enzyme is also implicated in the generation of superoxide radical anions in a similar manner, which is probably involved in signaling processes (42,43). Our results showed that both DPI, an inhibitor of flavoproteins, and AEBSF, an inhibitor for the interaction of p47 phox and/or p67 phox with cytochrome b 558 (46), block the induction of Mn-SOD mRNA. It suggests, therefore, that an upstream signal for TPA-mediated induction of MnSOD mRNA may be O 2 . generated by NADPH oxidase or similar oxidases activated by TPA. MSTRE may not be the only TPA-responsive element mediating activation of human MnSOD gene in vivo. The level of MnSOD mRNA increases about 6-fold following treatment of A549 cells with TPA for 4 h. The TPA-induced luciferase expression from pSODLUC exhibits, however, only about 2-3-fold stimulation over the untreated cells. The difference in the extent of induction may imply that other TPA-responsive elements exist in the region outside of the promoter sequences examined in this work. Nevertheless, a mechanism for the TPA-mediated MnSOD induction is likely to include the activation of a NADPH oxidase-like enzyme to generate O 2 . as an upstream signal and phosphorylation of a CREB-1/ATF-1-like factor needed for MSTRE-mediated transcriptional activation of the MnSOD gene. This or a similar mechanism involving O 2 . as an upstream signal for the induction of MnSOD may be congruent with the role of the enzyme against O 2 . toxicity in vivo.