Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Mitochondrial manganese superoxide dismutase (MnSOD) is critical in the protection of respiratory-chain proteins from superoxide (O₂) and its reaction products. MnSOD appears responsible for the ability of bacteria to grow in aerobic conditions because Escherichia coli, lacking the MnSOD and FeSOD genes, are unable to grow in aerobic minimal medium. Anaerobic growth is, however, normal (1). MnSOD is also essential to survival of eukaryotic organisms in aerobic environments. Homozygous knockout mice (SOD2^m1BCM/SOD2^m1BCM), in which the MnSOD gene has been inactivated, die within 3 wk of birth because of overwhelming oxidant stress. Their phenotype includes severe anemia, degeneration of neurons in basal ganglia and brainstem, and degeneration of cardiac myocytes (2). Li and colleagues described a similar line of MnSOD knockout mice that die within the first 10 d of life from dilated cardiomyopathy, lipid accumulation, and metabolic acidosis (3). Surfactant protein-C (SP-C) promoter-driven MnSOD activity in mitochondria of lung epithelial cells increases the resistance of MnSOD-transgenic mice to pulmonary oxygen toxicity (4). MnSOD-transgenic mice were highly protected from lung injury during exposure to 95% oxygen, surviving significantly longer than nontransgenic littermates. Pulmonary pathology showed less hemorrhage, hyaline membrane formation, and edema in the transgenic animals (4).

In situ hybridization of control and 85% oxygen-exposed rat lungs has revealed MnSOD transcripts in alveolar type II cells, arterioles, endothelial cells, and pleural mesothelial cells (5). Quantitative immunocytochemistry of rat lung has shown the MnSOD protein in alveolar epithelial cells, where it is both abundant and induced by hyperoxia (6). MnSOD in the mitochondrial matrix increased significantly after 7- or 14-d exposures to 85% oxygen (6).

Ho and others have characterized the structure and function of the rat MnSOD gene promoter (7). A 507-bp region located 5' to the transcriptional start site directs expression of a reporter in transgenic mice. Cis-acting elements responsive to hyperoxia and tumor necrosis factor (TNF)- exist in a 218-bp region between base pairs 507 and 405 and 405 and 289 of the promoter, respectively. The 5' flanking region is guanidine plus cytosine (G + C) rich and contains copies of the promoter-selective transcription factor (SP)-1 binding motif, the simian virus-40 (SV40) core enhancer, an activator protein (AP)-1 binding motif, and a nuclear factor-B (NF-B)-like site (7).

MnSOD regulation is complex. Regulation occurs at both pre- and posttranslational levels. The rate of MnSOD protein synthesis decreases more than does overall protein synthesis in anoxia (8). The enzymatic activity of lung MnSOD also decreases after exposure of rats to 95% oxygen for 48 h (9). Induction of lung MnSOD protein and activity by TNF- or interleukin (IL)-1 in rats is associated with increased oxygen tolerance (10). Phorbol ester simultaneously activates both protein kinase C and NF-B in cells in which it induces MnSOD (11). MnSOD gene expression is also importantly regulated by alterations in messenger RNA (mRNA) stability. Message stability increases in oxygen-exposed rat lungs. The half-life of MnSOD mRNA was found to increase significantly, from 8.2 ± 0.8 h to 19.0 ± 2.8 h (mean ± SE) in adult rats exposed to oxygen for 48 h (9).

In contrast, both MnSOD activity and protein expression consistently decrease in hypoventilated and hypoperfused (i.e., hypoxic) rabbit lungs (12). In our studies, this has occurred without a change in MnSOD specific activity, since MnSOD protein and activity decrease in concert. Freshly isolated, cultured rabbit alveolar type II cells also showed decreased MnSOD gene expression (both mRNA and protein) after in vitro exposure to 2.5% or 1% oxygen (13).

Rats develop increased lung MnSOD activity and oxygen tolerance after exposure to hypoxia (14). This response is not typical of other mammals. Predictably, the absence of induction of lung MnSOD in mice exposed to 10% O₂ is not associated with increased oxygen tolerance (14). Other mechanisms accounting for increased oxygen tolerance of rats preexposed to hypoxia include induction of lung glutathione redox cycle activity (15).

Based on these yet unreconciled observations, we hypothesized that the rat and human MnSOD promoters might respond differently to hypoxia. In particular, we wished to address whether the rat MnSOD promoter increased MnSOD gene transcriptional activity in hypoxia.

This study investigated the response of the rat MnSOD gene promoter to hypoxia in vitro, using MnSOD promoter deletions coupled to luciferase reporters. We tested effects of hypoxia on the complete (2,505-bp) and fragmentary (1,089-, 507-, 405-, and 289-base pairs [bp]) MnSOD promoters. Several cell lines, including human lung adenocarcinoma cells (A549), rabbit lung fibroblasts (R9ab), and rat alveolar type II cell lines (E1A-T2 and L2) were transfected under identical conditions with the MnSOD promoter-luciferase reporters. Consistently increased MnSOD promoter activity occurred only in transfected A549 adenocarcinoma cells. The response was unrelated to consensus sequences for hypoxia-inducible factor-1 (HIF-1) binding sites in the 5' flanking region of the rat MnSOD gene. In separate experiments, we compared endogenous human MnSOD gene expression in A549 cells as reflected by Northern blot assay and reverse transcription-polymerase chain reaction (RT-PCR).

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Experimental Design

The experiments in the study investigated the expression of the MnSOD promoter in transiently transfected lung epithelial cells, with a promoter-deletion series cloned from the 5' flanking region of the rat MnSOD gene. Cells in monolayer culture were exposed to air/5% CO₂ ("air") or 2.5% O₂/5% CO₂/balance N₂ ("hypoxia") for 24 h. Luciferase expression driven by the transiently transfected rat MnSOD promoters was tested in hypoxia and compared with expression of the same fragment in air. Azide (10 µm) was used in some experiments to test effects of mitochondrial inhibition per se. The function of two apparent HIF-1 sites within the smallest promoter (289 bp) was tested through site-directed mutagenesis.

In separate experiments, we also examined expression of the endogenous MnSOD gene (i.e., controlled by the human promoter). We assessed A549 cells' MnSOD gene expression using Northern blotting and RT-PCR after air or hypoxia for 1 or 3 d.

Cell Culture

Lung epithelial cells from a human bronchogenic carcinoma (A549 cells, line CCL 185; American Type Culture Collection [ATCC], Rockville, MD) were the primary model (16). Cells were maintained in Dulbecco's modified Eagle's medium (DMEM)/F-12 medium (Cellgro, Herndon, VA) and 10% fetal bovine serum (Sigma, St. Louis, MO) with 1× penicillin-streptomycin (GIBCO BRL, Gaithersburg, MD). Cells were passaged weekly and were uniformly 80- 90% confluent (1-2 d after splitting at 1:4) when used.

Air exposures were done in a tissue culture CO₂ incubator (Model 1358; Forma Scientific, Marietta, OH). Hypoxia exposures took place in a Plexiglas chamber (Plas Labs, Lansing, MI) kept in a tissue culture incubator at 37°C. Gas composition in the chamber was controlled by infusing 95% N₂/5% CO₂ intermittently. Anoxic gas inflow was controlled automatically by a solenoid valve (Peter Paul Electronic Co., New Britain, CT) activated by a digital oxygen controller (Model 1630; Engineered Systems and Designs, Newark, DE) that monitored the chamber's internal oxygen level continuously. The controller kept the oxygen concentration  2.5% (PO₂ ~ 19 mm Hg).

MnSOD Promoter Deletion Constructs

We used a rat MnSOD gene promoter-deletion series consisting of fragments of 3,303- (full length), 2,505-, 1,064-, 507-, 405-, and 289-bp (bp numbered in the 5' direction from the transcription initiation [CAP] site). The plasmid contained the choline acetyltransferase (CAT) gene, SV40 splice, and poly A sites. We removed sections by restriction digestion and subcloned them into a promoterless firefly luciferase vector, pGL2-Basic (Promega, Madison, WI). The 2,505-bp promoter was subcloned in sense and antisense orientations. Other fragments were cloned in the sense orientation only. Promoter constructs are shown in Figure 1.

View larger version (11K): [in this window] [in a new window]   Figure 1.   Rat MnSOD promoter restriction-deletion luciferase constructs. The full-length rat MnSOD promoter (3.3 kb) was cut at the indicated restriction sites and ligated into a promoterless firefly luciferase vector (pGL2-Basic). Constructs are identified by the number of base pairs 5' to the transcription initiation site (CAP site). The shortest restriction fragment, of 289 bp, harbored two HIF-1 binding site-like sequences (5'-CGTG-3') at 197 and 179 (5' position relative to the CAP site).

Transient Transfection of Lung Epithelial Cells

Cells (A549 and other cells) were transiently transfected by lipofection with MnSOD promoter-luciferase constructs (with or without the herpes simplex thymidine kinase-driven Renilla luciferase reporter [pRL-TK] plasmid). Cells were transfected using DOTAP-DOPE (Avanti Polar Lipids, Birmingham, AL) consisting of 1, 2-diolyeyl-3-trimethylammonium propane (DOTAP) and dioleylphosphatidylethanolamine (DOPE) (17) in water. Cells were 80-90% confluent in six-well Primaria plates (Becton Dickinson Labware, Lincoln Park, NJ) when transfected with DNA:DOTAP-DOPE in a ratio of 1:2 (2 µg DNA and 4 µg lipid per well).

pRL-TK (Promega) (18) was used in cotransfection experiments to compare transfection efficiency. The final ratio of pGL-MnSOD DNA to pRL-TK DNA in cotransfection experiments was 10:1.

Luciferase Assays

Firefly and Renilla luciferases were measured with the Dual-Luciferase Reporter Assay system (Promega) according to the manufacturer's protocol. Cells from each well of a six-well plate were lysed in passive lysis buffer (400 µl/well) and transferred into microcentrifuge tubes. A 20-µl aliquot was assayed sequentially for firefly luciferase activity and Renilla luciferase activity. Assay background activity was determined and subtracted. Data are expressed as arbitrary relative light units (RLU).

RNA Preparation

Epithelial cells were lysed in RNAzol (Biotecx Laboratories, Houston, TX) according to the manufacturer's protocol. Total cellular RNA was isolated. Pellets were washed with 75% ethanol and dried. RNA was resuspended in 22 µl diethylpyrocarbonate-treated distilled water and stored at 20°C.

Northern Blotting

Total cellular RNA was denatured with glyoxal, and 20 µg was loaded in each lane. RNA was electrophoresed at 100 V for ~ 2 h in 1.2% agarose gels and transferred to nylon membranes. RNA on the membranes was fixed by UV light before prehybridization. Blots were hybridized subsequently with [³²P]-labeled complementary DNA (cDNA) probes for human MnSOD (HMS probe) and 18S RNA.

Blots of total RNA prepared from transfected (289-bp construct), air- and hypoxia-exposed A549 cells were hybridized with a cDNA probe for firefly luciferase. The blots were then rehybridized with the 18S probe.

RT-PCR Estimation of MnSOD and CuZnSOD mRNAs

Two micrograms of total RNA were combined with distilled water and random hexamers (100 pmol) in a 0.65-ml, thin-walled PCR tube. The mixture was heated at 68°C for 5 min and then cooled to 37°C in a thermal cycler (Stratagene RoboCycler Gradient 40; Stratagen, La Jolla, CA). The mixture was then reverse-transcribed (RT) by adding 1× transcription buffer, dithiothreitol (10 mM), deoxynucleotide triphosphates (dNTPs) (0.5 mM), and 400 units of Superscript II (ribonuclease H reverse transcriptase; GIBCO BRL). The mixture was incubated at 37°C for 60 min and 95°C for 5 min. The reaction was diluted 5-fold with sterile distilled water.

Ten microliters of RT product was amplified by PCR by adding 1× PCR buffer, dNTPs (1.5 mM), MgCl₂ (1.5 mM), 100 pmol of primer specific for either MnSOD or CuZnSOD, and 2.5 units of Taq DNA polymerase (GIBCO BRL). The mixture was heated to 95°C for 5 min, followed by 28 cycles of PCR at 95°C for 1 min, 62°C for 1 min, and 72°C for 1 min. Additional extension was done at 72°C for 7 min. PCR products were visualized by ethidium bromide staining after electrophoresis in 1% agarose gels.

Site-Directed Mutagenesis

Two apparent HIF-1 consensus sequences (5'-RCGTG-3') within the 289-bp construct, at 197 and 179 (numbered as bp 5' from the transcription initiation site), were mutated with the QuikChange (Stratagene, La Jolla, CA) site-directed mutagenesis kit. Each pyrimidine was exchanged for a noncomplementary purine and vice versa. A construct, containing the 5-bp mutations at each of the two apparent HIF-1 consensus sequences at nucleotides 179 (2,309-2,313) and 197 (2,326-2,330), was generated by PCR (pGL-R62HIF). Mutations were verified by automated sequencing (University of Alabama at Birmingham core facility). They contained the desired 5-bp mutations and no others.

Data Analysis

Data are expressed as arithmetic means ± SD or SE as indicated. Multiple comparisons were made by two-way analysis of variance (ANOVA) followed by t tests modified for multiple comparisons (Bonferroni's procedure) (19).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Hypoxic Modulation of Endogenous (Human) MnSOD Transcript Expression

Northern blots. Northern blots hybridized with a cDNA probe for human MnSOD (HMS probe) showed a decrease in MnSOD message expression. The 4.0-kb MnSOD mRNA transcript decreased, as predicted on the basis of our prior studies (12, 13), after exposure of A549 cells to hypoxia for 1 or 3 d. Blots were rehybridized with a probe for 18S RNA that confirmed equivalent loading. These data are shown in Figure 2.

View larger version (92K): [in this window] [in a new window]   Figure 2.   Northern blot showing decreased A549-cell 4-kb MnSOD message after hypoxia exposure. The 4.0-kb MnSOD transcript appeared to decrease slightly after 1 d in hypoxia (2.5% O2), and it decreased appreciably after 3 d in hypoxia. Rehybridization of the blot with a cDNA probe for 18S RNA showed nearly equivalent lane loading. Twenty micrograms of total A549 cellular RNA were loaded in each lane. The autoradiogram for MnSOD was exposed for 6 d at 80°C. The 18S autoradiogram was exposed overnight at 80°C. Densitometry readings of the 18S signals were: 1a = 0.35; 1h = 0.33; 3a = 0.29; 3h = 0.39. Abbreviations: 1a, 1 d (24 h) air; 1 h, 1 d (24 h) hypoxia; 3a, 3 d (72 h) air; 3 h, 3 d (72 h) hypoxia.

View larger version (54K): [in this window] [in a new window]   Figure 3.   RT-PCR data confirming decreased A549-cell MnSOD message expression after hypoxia exposure. The bar graph shows the mean ratio (n = 2) of the MnSOD-to-CuZnSOD RT- PCR products. The height of the bars is proportional to message abundance. The bar height represents absorbance of the photographic negative as determined by densitometry (arbitrary units). The ratio appeared to decrease slightly (8%) after 1 d in hypoxia, and it decreased significantly (62%) after 3 d in hypoxia. The CuZnSOD RT-PCR product remained constant in hypoxia: air = 0.71; 1 d hypoxia = 0.70; 3 d hypoxia = 0.65 (data are arbitrary absorbance units). These data show that expression of the human MnSOD gene decreases in hypoxia.

Hypoxic Modulation of Reporter (pRL-TK) Expression

In some experiments, we transfected cells with pRL-TK alone, the thymidine kinase-driven luciferase reporter, to test transfection efficiency. Cells were exposed later to air or hypoxia, and luciferase expression assayed as described.

Hypoxia did not significantly change the expression of pRL-TK when it alone was transfected into A549 cells. Luciferase expression in cells transfected only with pRL-TK was: air, 272 ± 96 RLU in air and 237 ± 60 RLU with hypoxia (mean ± SD) (P > 0.05 by unpaired t test; n  12 for each group).

Effects of Hypoxia on MnSOD Promoter Activity in A549 Cells

Using the human A549 cell line, we tested effects of hypoxia on the rat promoter deletions transfected alone (data expressed as total RLU per volume lysate) or cotransfected with the pRL-TK reporter to normalize for transfection efficiency (expressed as ratio of MnSOD promoter-driven firefly luciferase to TK-driven Renilla luciferase). To account for effects of hypoxia on the number of cells, we also normalized the MnSOD-driven firefly luciferase activity to monolayer protein content.

Cotransfection experiments. Luciferase expression driven by the MnSOD deletion constructs increased significantly (for the experiments shown, F = 36.23 and P < 0.0001 by two-way ANOVA for the effect of hypoxia). Hypoxia- induced increases in promoter activity of the rat MnSOD promoter-driven firefly luciferase (expressed as RLU/ml cell lysate) were +20% (2,505 bp), +29% (1,064 bp), +39% (507 bp), +57% (405 bp), and +51% (289 bp). Data from the luciferase and Renilla cotransfection experiments showing relative luciferase expression of the promoters and individual air and hypoxia comparisons are summarized in Figure 4A. Promoter activities for each of the individual fragments were compared in air and hypoxia, using t tests modified for multiple comparisons. These results are shown as asterisks in Figures 4A and 5. 

View larger version (33K): [in this window] [in a new window]   View larger version (34K): [in this window] [in a new window]   View larger version (29K): [in this window] [in a new window]   Figure 4.   Hypoxia increased MnSOD promoter activity of transiently transfected luciferase constructs in A549 cells. These data are from the firefly and Renilla cotransfection experiments. The format is the same in each panel. The height of the bars represents mean (n = 6) promoter activity for each construct in air (solid bars) or hypoxia (hatched bars). One SD is indicated above the bars in A and B. The results of two-way ANOVA comparing promoter activity in air and hypoxia are shown in A and B. Asterisks above the bar indicate P < 0.05 for each individual comparison of air with hypoxia (Bonferroni's modification of the t test). A: hypoxia increased MnSOD promoter activity in hypoxia, when expressed as RLU per volume cell sonicate. B: hypoxia decreased pRL-TK activity slightly when Renilla luciferase was expressed as RLU per volume cell sonicate. However, no individual comparison reached statistical significance. C: hypoxia increased MnSOD promoter activity when expressed as the ratio of firefly luciferase (MnSOD promoter) to Renilla luciferase (TK promoter) activity. Data shown are the ratios of the firefly and Renilla luciferase acitivities shown individually in A and B.

Transfection of MnSOD promoter constructs alone. We also transfected MnSOD promoter restriction constructs alone into A549 cells. In these experiments, we normalized MnSOD-driven luciferase expression to total cellular protein. These experiments confirmed hypoxia-induced increases in MnSOD promoter-driven luciferase expression (expressed as firefly RLU/mg cellular protein) (for the experiment shown, F = 40.7 and P < 0.0001 by two-way ANOVA for the effect of hypoxia): of +250% (2,505 bp), +26% (1,089 bp), +105% (507 bp), +92% (405 bp), and +44% (289 bp). The data and individual comparisons are summarized in Figure 5.

View larger version (22K): [in this window] [in a new window]   Figure 5.   Hypoxia increased MnSOD promoter activity of transiently transfected luciferase constructs in A549 cells. The format is identical to that in Figure 4. Firefly luciferase activity (driven by the rat MnSOD promoter) was normalized to cellular protein. These data show that the increase in luciferase activity was not due to changes in cell numbers occurring in hypoxia.

Mitochondrial inhibition with azide. We also investigated whether the effects of hypoxia were due to inhibition of mitochondrial respiration. Transfected (2,505-bp and 289-bp promoter constructs) lung epithelial cells were incubated with sodium azide (10 µM), an inhibitor of cytochrome oxidase, for 24 h before lucerifase was assayed. Azide slightly decreased expression of the MnSOD promoter-lucerifase reporter constructs, in contrast to the stimulatory effects of hypoxia on the MnSOD promoters.

Effect of Hypoxia on MnSOD Promoter-Driven Luciferase Expression in L2, R9ab, and E1A-T2 Cells

To determine whether the apparent increases in MnSOD promoter activity were cell type specific, we also transiently transfected two rat lung epithelial cell lines and rabbit lung fibroblasts. Cells were exposed to hypoxia (2.5% O₂ for 24 h) in protocols identical to those described for the A549 cells.

L2 cells. MnSOD promoter-driven firefly luciferase activity did not increase in L2 cells exposed to hypoxia. (One restriction fragment [507 bp] showed decreased activity in hypoxia when expressed in absolute terms or as the firefly-to-Renilla luciferase ratio.) Hypoxia did not change pRL-TK expression significantly in the L2 cells.

E1A-T2 cells. Expression of MnSOD promoter-driven firefly luciferase did not increase significantly in transiently transfected E1A-T2 cells after hypoxia exposure, when expressed as absolute luciferase activity or when normalized to pRL-TK expression. pRL-TK expression did not change significantly in transiently transfected E1A-T2 cells exposed to hypoxia.

R9ab cells. MnSOD promoter-driven firefly luciferase expression did not change significantly in transiently transfected rabbit lung fibroblasts. These data are summarized in Table 1.

Site-Directed Mutagenesis of HIF-1 Sites

We identified two 5-bp sequences within the 289-bp rat MnSOD promoter that resembled the HIF-1 consensus sequence (5'-RCGTG-3'). These were located within the 289-bp restriction fragment at nucleotides 197 and 179 (nucleotides numbered 2,309-2,313 and 2,326-2,330 according to the genomic sequence [7]).

Both sites were altered by site-directed mutagenesis, so that they did not resemble known transcription-factor binding sites. The mutated sequences were 5'-TATGT-3'. Mutated plasmid DNA was then amplified by PCR and transfected into XL-1 Blue Supercompetent Cells. Elimination of the putative HIF-1 sites did not abrogate the increase in MnSOD promoter-driven luciferase activity in hypoxia. The data are summarized in Figure 6.

View larger version (54K): [in this window] [in a new window]   Figure 6.   Mutation of HIF-1 sites in the rat MnSOD gene promoter did not eliminate induction of promoter activity by hypoxia. The bars represent the mean (n = 6) luciferase activity of the mutated MnSOD 289-bp promoter normalized to Renilla luciferase. One SD is indicated above the bars. Mutation of two apparent HIF-1 sites in the 289-bp promoter did not eliminate the increase in transcriptional activity that occurred in hypoxia.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

We had previously found that MnSOD expression in epithelial cells (specifically, rabbit alveolar type II epithelial cells) decreased after lung tissue hypoxia in vivo or hypoxic culture in vitro (12, 13, 20). Hypoxia-induced decreases in MnSOD, the enzyme that protects mitochondrial proteins from oxidative inactivation, may be important in the pathophysiology of reoxygenation. At the other extreme, MnSOD gene expression has been implicated as a mechanism of adaptation to hyperoxia (especially in the rat model) (9) and in protection of cultured cells from the effects of TNF- (21).

The experiments in our study tested the response of the rat MnSOD gene promoter to hypoxia, and compared this with expression of the human MnSOD gene. On the basis of previously reported, apparently divergent adaptive responses (12, 20), we hypothesized that responses of the rat and human MnSOD promoters might differ in hypoxia. Both Northern blotting and RT-PCR showed a decrease in endogenous MnSOD transcripts in human lung epithelial cells (A549 cells) cultured in hypoxia (Figures 2 and 3). In contrast, transiently transfected rat MnSOD gene promoter constructs consistently increased A549 cell luciferase expression in hypoxia. The increase occurred only in A549 cells and not in other rat lung epithelial lines or rabbit fibroblasts. Hypoxia-induced increases in promoter activity occurred with all promoter fragments, including the shortest (289 bp). Two sequences resembling minimal HIF-1 sites exist within the 289-bp promoter. Site-directed mutagenesis of these apparent consensus sequences did not, however, extinguish the hypoxic stimulation of MnSOD promoter activity. Thus, hypoxia-sensitive elements appear to be present in this region of the promoter.

Endogenous Human MnSOD Gene Expression in Hypoxia

We initially studied expression of the human MnSOD gene in hypoxia, using cultured A549 cells. As predicted from experiments in vivo (22) and with cultured lung alveolar type cells in vitro (13), hypoxia caused a significant decrease in MnSOD gene expression. Decreased MnSOD gene expression in hypoxic A549 cells corresponds to the response we reported previously in mice exposed to hypoxia in vivo (22), in hypoxic and hypoperfused lung tissue in vivo (12), and in freshly isolated alveolar type II cells and lung fibroblasts cultured in hypoxia in vitro (13). These studies established MnSOD downregulation as a general response to hypoxia.

The human MnSOD promoter has been characterized (23, 24). A transcription start site has been mapped by primer extension analysis to a guanine residue 67 bp upstream of the translation start site. The human promoter lacks TATA and CAAT boxes, as does the rat promoter. The transcription initiation site is preceded by a G + C-rich region (84% of the first 200 bp), which contains multiple SP-1 sites, and AP-1 and AP-2 consensus sequences. A minimal promoter (at bp 34 to +38) exists within the G + C-rich region, adjacent to the transcription start site (24). SP-1 sites appear critical for transcriptional activity of the minimal promoter. The 3' flanking region of the human MnSOD gene also contains an NF-B sequence of unknown regulatory significance (24). Elements within the 3' flanking region are not relevant to the 5' deletion constructs we studied, and they may account for differences in endogenous and transfected gene responses to hypoxia.

Response of Transfected MnSOD Promoters to Hypoxia

We used a restriction deletion series cloned from the full-length rat MnSOD gene promoter to investigate the response of A549 epithelial cells to hypoxia. In the first case, we expressed MnSOD promoter activity as light-emission units per volume of cell sonicate. The resulting data showed an increase in transfected promoter activity with all restriction fragments in hypoxia, including the 289-bp fragment. The increased promoter activity was consistent and highly significant (by two-way ANOVA) for the overall effect of hypoxia.

In other experiments, we cotransfected cells with a TK-driven Renilla luciferase reporter (pRL-TK) to monitor transfection efficiency. Hypoxia slightly decreased the expression of Renilla luciferase (in the cotransfection experiments), showing that hypoxic activation of the transfected rat MnSOD promoter construct was specific. When rat- gene MnSOD promoter activity (i.e., firefly luciferase) was normalized to Renilla luciferase, MnSOD promoter activity in hypoxia likewise increased significantly. Therefore, differences in transfection efficiency did not account for the results.

Similar results were obtained when rat MnSOD promoter-driven firefly luciferase activity was normalized to monolayer protein. Monolayers cultured in hypoxia for 24 h after transfection contained less cellular protein (i.e., fewer cells) than those cultured in air. When normalized to cellular protein, the transfected promoter fragments significantly increased (by two-way ANOVA) transcriptional activity as evidenced by greater firefly luciferase activity (RLU/mg protein). These data independently confirmed the increase in promoter activity observed when normalized to TK-driven Renilla luciferase.

MnSOD Promoter Activity in Other Lung Epithelial Cells and Fibroblasts

Although we detected a consistent increase in rat MnSOD promoter activity in A549 cells transiently transfected before hypoxia, we did not find an increase in transfected MnSOD promoter activity in other cell types. No significant increases in transfected MnSOD promoter activity were detected in lung epithelial cell lines L2 and E1A-T2 or in the R9ab fibroblast line. Functionally, L2 cells do not produce surfactant, nor do they transport fluid, as do typical epithelial cells. The cells do share specific antigens with alveolar type II cells (25) and are rich in phosphatidate phosphohydrolase activity (26). E1A-T2 cells were derived originally from neonatal rat lungs (27). Although they express lectin-binding glycoproteins (Maclara pomifera agglutinin [MPA]-gp200), as is consistent with epithelial cell differentiation, they lack lamellar bodies, SP-A, and saturated phosphatidylcholine.

R9ab cells respond to hypoxia with a decrease in MnSOD message expression (13). One known mechanism of decreased expression is destabilization of the MnSOD message, the half-life of which decreases significantly in hypoxia (13). A decrease in the transcription rate of the MnSOD gene remains an additional possible mechanism.

In attempting to confirm transcriptional activation of the MnSOD promoter fragments by hypoxia, we used Northern blotting and RT-PCR to detect luciferase mRNA transcripts in hypoxia-exposed, transfected A549 cells (i.e., 36-48 h after transfection). Luciferase transcripts were not detectable with Northern blotting or RT- PCR in air- or hypoxia-exposed cells transfected 36-48 h earlier. RT-PCR easily detected the luciferase-containing plasmid included as a positive control. These results suggest that luciferase mRNA is short-lived, whereas the protein is relatively stable. Supporting this notion was our consistent detection of luciferase activity at this later time point. Because we did not conduct nuclear run-on experiments, increased transcriptional activity is suggested but was not demonstrated directly.

Effect of Mutation of HIF-1-like Sites on MnSOD Induction in Hypoxia

HIF-1 is a basic helix-loop-helix transcription factor composed of HIF-1 and HIF-1 protein subunits. Functionally defined HIF-1 binding sites include the consensus sequence (5'-RCGTG-3'), and may require adjacent sequence conservation (28). Transcription of a number of genes, including those for erythropoietin, vascular endothelial growth factor, heme oxygenase-1, and glycolytic enzymes are regulated in hypoxia by HIF-1 protein binding (29). The glucose transporter (GLUT-1) is also transcriptionally regulated by activation of HIF-1 (30). The rat MnSOD gene 5' promoter region contains at least two HIF-1 sites, but these are not required for hypoxic induction. Another HIF-1 site exists in the 3' region of the rat MnSOD gene, but would not be relevant to regulation of the luciferase constructs used in our study. Putative HIF-1 sites are not present in the human or mouse MnSOD gene promoters.

Because of the important role of HIF-1 in regulation of hypoxia-inducible genes, and the presence of apparent HIF-1 sites in the rat promoter, we tested whether these might be functionally active. Mutation of the sites to nonfunctional sequences had no inhibitory effect on MnSOD gene expression in hypoxia. Although activation of rat MnSOD gene transcription by hypoxia appears paradoxical on the basis of the enzyme's central role in oxygen metabolism, other examples of organ-specific, hypoxia-induced gene expression exist (28).

Significance

The experiments done in the present study reveal a unique property of the rat MnSOD promoter region: apparent transcriptional activation by hypoxia. In contrast, human MnSOD gene expression decreases in hypoxia, as predicted from previous experiments. These divergent responses provide an opportunity to examine species-specific gene expression in hypoxia, and transcriptional activators that might be expressed in a cell-specific fashion.

Whole-animal studies found an apparent species-specific response of rat lung MnSOD to hypoxia (10% oxygen for 7 d) (14). However, we previously found that male mice in hypoxia (10% oxygen for 7 d) had decreased lung MnSOD expression (22), and that rabbit lungs made hypoxic (PO₂ ~ 20-30 mm Hg) by nonventilation and nonperfusion also had decreased MnSOD activity and protein expression. It appears that species react differently, vis à vis lung MnSOD expression, to hypoxia. Mouse and rabbit lungs, as well as A549 cells, appear to display a generalized response of decreased MnSOD expression in hypoxia. This appears teleologically appropriate, since MnSOD regulation may depend on oxidant tone (1). The mitochondrial respiratory chain would produce less superoxide in hypoxia, requiring less MnSOD to detoxify it. Postulated species-specific responses derive from in vivo experiments in which rats were exposed to moderate hypoxia (11). Experiments with cultured cells, although utilizing a simpler system, are complicated by possible effects of culture conditions and dedifferentiation.

Since MnSOD is partly controlled by transcriptional mechanisms, responses of specific cell types with regard to this enzyme may depend on the presence of specific transcriptional activators or regulatory proteins. A549 cells appear to be uniquely endowed with factors that activate rat MnSOD promoter activity. Alternately, A549 cells may be uniquely able to process gene transcription products after hypoxia. The hypoxic activation of MnSOD promoter activity appears not to be due to HIF-1 binding to the 5' flanking region, because elimination of the HIF-1 binding sites had no inhibitory effect.

Many transcriptional activators are controlled or affected by redox mechanisms, and these may be expressed differentially (31). The zinc-finger transcription factor, SP-1, may be especially important in MnSOD regulation, and SP-1 is regulated by redox state. The MnSOD 5' flanking region is rich in G + C and harbors binding sites for SP-1 (23). SP-1 protein binding appears to be inhibited by an increase in oxidation potential of the cell, suggesting a possible role of hypoxia in MnSOD activation (32). We have not specifically sought to identify this effect through gel shift or DNA footprinting assays. Both NF-B and AP-1 are expressed in A549 cells, although we do not have data on relative levels of their expression in other cell types. AP-1 activity is regulated by redox state at both transcriptional and posttranscriptional levels. AP-1 can be activated by antioxidants and by c-jun/c-fos gene expression. Although we have observed strong induction of NF-B in A549 cells stimulated by TNF-, we have not sought NF-B translocation in hypoxia (unpublished observations). Although NF-B is activated by oxidation in cell culture, its activation has not been linked causally to MnSOD induction.