An Antisense Oligonucleotide to 1-cys Peroxiredoxin Causes Lipid Peroxidation and Apoptosis in Lung Epithelial Cells*

1-cys peroxiredoxin (1-cysPrx), a member of the peroxiredoxin superfamily, reduces phospholipid hydroperoxides as well as organic peroxides and H2O2. To determine the physiological function(s) of 1-cysPrx, we have used an antisense strategy to suppress endogenous 1-cysPrx in L2 cells, a rat lung epithelial cell line. A 25-base antisense morpholino oligonucleotide was designed to bind a complementary sequence overlapping the translational start site (−18 to +7) in the rat 1-cysPrx mRNA, blocking protein synthesis. Treatment with an antisense oligonucleotide for 48 h resulted in approximately 60% suppression of the 1-cysPrx protein content as measured by immunoblot analysis and an approximately 44% decrease of glutathione peroxidase activity as compared with random oligonucleotide treated and control (vehicle only) cells. Accumulation of phosphatidylcholine hydroperoxide in plasma membranes was demonstrated by high pressure liquid chromatography assay for conjugated dienes (260 pmol/106cells for antisense versus 70 pmol/106 cells for random oligonucleotide and control cells) and by fluorescence of diphenyl-1-pyrenylphosphine, a probe for lipid peroxidation. The percentage of cells showing positive staining for annexin V and propidium iodide after antisense treatment was 40% at 28 h and 80% at 48 h. TdT-mediated dUTP nick end labeling assay at 48 h indicated DNA fragmentation in antisense-treated cells that was blocked by prior infection with adenovirus encoding 1-cysPrx or by pretreatment with a vitamin E analogue. The results indicate that 1-cysPrx can function in the intact cell as an antioxidant enzyme to reduce the accumulation of phospholipid hydroperoxides and prevent apoptotic cell death.


1-cys peroxiredoxin
, a member of the peroxiredoxin superfamily, reduces phospholipid hydroperoxides as well as organic peroxides and H 2 O 2 . To determine the physiological function(s) of 1-cysPrx, we have used an antisense strategy to suppress endogenous 1-cysPrx in L2 cells, a rat lung epithelial cell line. A 25-base antisense morpholino oligonucleotide was designed to bind a complementary sequence overlapping the translational start site (؊18 to ؉7) in the rat 1-cysPrx mRNA, blocking protein synthesis. Treatment with an antisense oligonucleotide for 48 h resulted in approximately 60% suppression of the 1-cysPrx protein content as measured by immunoblot analysis and an approximately 44% decrease of glutathione peroxidase activity as compared with random oligonucleotide treated and control (vehicle only) cells. Accumulation of phosphatidylcholine hydroperoxide in plasma membranes was demonstrated by high pressure liquid chromatography assay for conjugated dienes (260 pmol/10 6 cells for antisense versus 70 pmol/10 6 cells for random oligonucleotide and control cells) and by fluorescence of diphenyl-1-pyrenylphosphine, a probe for lipid peroxidation. The percentage of cells showing positive staining for annexin V and propidium iodide after antisense treatment was 40% at 28 h and 80% at 48 h. TdT-mediated dUTP nick end labeling assay at 48 h indicated DNA fragmentation in antisense-treated cells that was blocked by prior infection with adenovirus encoding 1-cysPrx or by pretreatment with a vitamin E analogue. The results indicate that 1-cysPrx can function in the intact cell as an antioxidant enzyme to reduce the accumulation of phospholipid hydroperoxides and prevent apoptotic cell death.
The peroxiredoxins (Prx) 1 constitute a recently described and widely distributed superfamily of nonselenium glutathione per-oxidases. 1-cysPrx, also named peroxiredoxin VI, is a member of this superfamily that can reduce phospholipid hydroperoxides (PLOOH) as well as organic peroxides and H 2 O 2 (1-3). 1-cysPrx has been isolated from bovine ciliary epithelium (4,5), rat olfactory epithelium (6,7), and rat and bovine lungs (8,9); Northern and immunoblot analyses of mammalian tissues show the presence of 1-cysPrx in all organs with especially high levels of 1-cysPrx protein and mRNA expression in lung (8,10). 1-cysPrx is a bifunctional enzyme with both glutathione peroxidase (GPx) and phospholipase A 2 (PLA 2 ) activities (3). Despite multiple reports on 1-cysPrx activity in vitro (1-5, 11, 12), little is known about its physiological function and its potential role in protection of cells against reactive oxygen species (ROS)mediated damage.
ROS are generated in all aerobic organisms during normal cellular metabolism, during anti-microbial defense, and associated with environmental insults. Exposure to increased levels of ROS, referred to as oxidative stress, leads to oxidation of proteins, lipid peroxidation, and DNA base modification and strand breakage (13,14). Oxidative damage caused by ROS can also induce uncontrolled cellular proliferation, differentiation, aging, and apoptosis (15,16). Living organisms have evolved both nonenzymatic and enzymatic antioxidants to cope with ROS-mediated stresses.
Phospholipids form the major structural component of biological membranes, the lipid bilayer, thus providing the environment for the function of membrane-associated proteins such as ion channels, receptors, and enzymes through lipid-lipid and lipid-protein interactions. Peroxidation of phospholipids perturbs the structure of the bilayer, thus altering the active conformation of membrane-associated proteins resulting in the inability of cells to maintain the nonequilibrium state between cytosol and the extracellular milieu (17). As a consequence, avoidance of cellular toxicity requires specific mechanisms to protect against an accumulation of PLOOH in plasma membrane. Because cytosolic GPx (type I) cannot reduce PLOOH (2), it has been proposed that membrane repair of peroxidative damage requires the sequential action of three enzymes: phospholipase A to cleave the peroxidized acyl chain, GPx to reduce the peroxidized fatty acid, and acyl transferase to regenerate the phospholipid (18). More recently, phospholipid hydroperoxide GPx, a seleno enzyme, has been shown to directly reduce PLOOH (19) and to protect cells against oxidant stress (20). Glutathione S-transferase can also reduce PLOOH, although at a relatively low efficiency (21,22). On the other hand, 1-cysPrx can reduce PLOOH with a rate constant similar to that of phospholipid hydroperoxide GPx (2). However, its effectiveness in protection against cellular phospholipid peroxidation has not been established.
In a previous study, we showed that overexpression of 1-cysPrx in a cell line protects against peroxidative injury associated with extracellularly generated ⅐ OH (23). In the pres-ent study, we used an antisense strategy to examine the effect of suppression of 1-cysPrx expression on viability of L2 cells, a rat lung epithelial cell line. The L2 line was chosen because these cells express 1-cysPrx mRNA and protein (24). This cell line allowed us to test the hypothesis that decreased 1-cysPrx expression would result in lipid peroxidation and damage to cell membranes.

EXPERIMENTAL PROCEDURES
Materials-Glutathione, glutathione reductase, NADPH, tert-butyl hydroperoxide, and H 2 O 2 were purchased from Sigma. All cell culture medium components were from Invitrogen. Trolox, a water-soluble form of vitamin E, was obtained from Aldrich. Restriction enzymes and T4 DNA ligase were purchased from New England BioLabs (Beverly, MA). Components for the PCR were purchased from PE Applied Biosystems (Foster City, CA). Diphenyl-1-pyrenylphosphine (DPPP) was purchased from Dojindo Molecular Technologies (Kumamoto, Japan). High titer adenoviral stocks were produced by the Vector Core of the Institute of Human Gene Therapy at the University of Pennsylvania.
Cell Culture-L2 cells (American Type Culture Collection, Manassas, VA) were cultured in minimal essential medium containing L-glutamine and supplemented with 5% bovine serum and 1% penicillin/streptomycin at 37°C in a 5% CO 2 humidified atmosphere. The cells, 3 ϫ 10 5 for a 60-mm culture dish or 2 ϫ 10 4 cells for a two chamber slide, were seeded 24 h prior to oligonucleotide treatment. Alveolar type II epithelial cells were isolated from rat lungs by digestion with elastase and maintained in primary culture for 24 h as previously described (8).
Oligonucleotide Treatment-A 25-base morpholino oligonucleotide (5Ј-CGGGCATGGCGGCAGTGGTGACGCT-3Ј) complementary to 18 nucleotides before the start of translation through the first two amino acid codons in rat 1-cysPrx cDNA (8) was prepared and purified at GeneTools, LLC (Philomath, OR). A randomly generated 25-base morpholino oligonucleotide with the sequence 5Ј-CCTCTTACCTCAGTTA-CAATTTATA-3Ј was used as a control for antisense specificity. The oligonucleotides were resuspended in sterile water and stored at Ϫ20°C. Cellular treatment with oligonucleotides was performed using ethoxylated polyethylenimine in a delivery solution provided by Gene-Tools, LLC. Morpholino oligonucleotide (0.47 M), delivery solution (0.19 M ethoxylated polyethylenimine), and serum-free minimal essential medium were incubated at room temperature for 20 min, and then 2.85 ml was added to each 60-mm culture dish. The same solution with an equivalent volume of H 2 O substituted for oligonucleotide was used as a vehicle control. After 3 h of incubation at 37°C, the transfection solution was replaced with fresh serum containing minimal essential medium. In some experiments, 150 M Trolox was added to the culture medium at that time.
Construction of a Recombinant Adenovirus Encoding Rat 1-cysPrx-The assembly and production of recombinant adenovirus was performed using the Adeno-X Expression System (Clontech, Palo Alto, CA). The coding region of rat 1-cysPrx was generated by PCR using fulllength rat 1-cysPrx cDNA as a template. An XbaI restriction site was introduced into the upstream primer (5Ј-CCTCTAGAATGCCCG-GAGGGCTGCTT-3Ј), and a KpnI site was introduced into the downstream primer (5Ј-CCGGTACCTTAAGGCTGGGGCGTATAACG-3Ј). Purified 675-bp PCR products were double digested with XbaI and KpnI, ligated to the similarly digested pShuttle mammalian expression vector, and transformed into competent TOP10FЈ Escherichia coli cells (Invitrogen). The plasmids containing the insert of 1-cysPrx were verified by DNA sequencing. The expression cassette was transferred to the adenoviral genome by means of an in vitro ligation reaction; the entire expression cassette was flanked by unique PI-SceI and I-CeuI cleavage sites so that it could be excised and ligated to the adenoviral DNA. The adenoviral plasmid DNA from transformed colonies was checked for the presence of the insert DNA and its orientation by restriction enzyme digestion. The newly tailored recombinant adenoviral plasmid was linearized with PacI and then transfected into human embryonic kidney 293 cells using FuGENE 6 (Roche Molecular Biochemicals) as recommended by the supplier. After 4 -7 days, the plaques became visible, and subsequently most of the cells detached from the plate. The cell suspension was cleared by centrifugation (1500 ϫ g) for 5 min, and the supernatant containing ϳ10 3 -10 5 viral particles (a low titer viral stock) was frozen at Ϫ70°C until use. This low titer viral stock was used for the production of a high titer viral stock (Ad.1-cysPrx). L2 cells were plated 12-24 h before infection with adenovirus at a density of 5 ϫ 10 5 cells/60-mm culture dish. After 1-2 h of incubation, the virus-containing medium was replaced by fresh complete growth medium and placed in a 37°C humidified CO 2 incu-bator for 48 h. As a control for efficiency of infection and viral toxicity, some dishes were infected with the same number of particles of adenovirus containing green fluorescence protein (GFP) gene (Ad.GFP).
Immunoblot Analysis-The cells were washed twice with PBS and disrupted with lysis buffer (25 mM Tris-HCl, pH 8, 1 mM EDTA, 1 mM EGTA, 2 mM dithiothreitol, 10% glycerol, 1% Triton X-100, complete protease inhibitor mixture). The cell extracts were homogenized by sonication and centrifuged at 10,000 ϫ g for 20 min. The protein concentration was determined using the Coomassie Blue assay (Bio-Rad). Supernatant containing 30 g of protein was mixed with 2ϫ sample buffer, boiled for 3 min, and applied to a 12% SDS-PAGE gel. After electrophoresis, the proteins were transferred onto nitrocellulose membrane. The blot was probed with 1:3000 diluted rabbit anti-1-cysPrx polyclonal antibody (24) and 1:5000 diluted horse-radish peroxidase-conjugated goat anti-rabbit IgG (Jackson ImmunoResearch Laboratory, West Grove, PA) with detection by ECL and quantitated by densitometric scanning of x-ray film using the FluorS multi-imager (Bio-Rad). The membrane then was washed with stripping solution (62.5 mM Tris-HCl, pH 6.8, 100 mM ␤-mercaptoethanol, 2% SDS) at 55°C and reprobed with 1:500 diluted rabbit anti-actin polyclonal antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) to normalize for protein loading.
Northern Blot Analysis-Total RNA was isolated from adenovirusinfected cells using the RNeasy mini kit (Qiagen, Valencia, CA) according to the manufacturer's instructions. Total RNA (5 g) was separated by electrophoresis on a 1% agarose gel containing formaldehyde. The size-fractionated RNA was then transferred onto a nylon membrane (Schleicher & Schuell) by capillary action and hybridized to 32 P-labeled rat 1-cysPrx cDNA probe generated by random priming (Amersham Biosciences). After a high stringency wash, the membrane was exposed to Kodak x-ray film with an intensifying screen at Ϫ70°C for 14 h and quantitated as described above for immunoblots. To normalize for loading, the membrane was stripped of the 1-cysPrx probe, rehybridized with 32 P-labeled glyceraldehyde-3-phosphate dehydrogenase cDNA probe and re-exposed to x-ray film.
Enzyme Assay-GPx activity was determined by coupled NADPH/ glutathione reductase assay in the presence of GSH with PCOOH or tert-butyl hydroperoxide as substrate, and lysosomal-type PLA 2 activity was measured at pH 4 with a liposome-based assay using radiochemical detection as described previously (3).
Analysis of Phospholipids by HPLC-L2 cells from the three experimental conditions (control, antisense and random oligonucleotidetreated) that were adherent to 60-mm culture dishes were extracted by the Bligh and Dyer procedure (25). The extract was dried under a stream of N 2 , and the pellet of phospholipids was redissolved in 50 l of methanol for analysis by HPLC (Waters, Milford, MA). The samples were injected onto a reverse phase high resolution C 18 column (3.9 ϫ 150 mm) (NovaPack), eluted isocratically with methanol, 50 mM ammonium acetate (60:40 v/v, pH 5.0) at a flow rate of 1 ml/min using an Alliance 2690 separation unit, and detected with PDA 996 UV and RI 442 detectors. The data were processed using Millenium 32 software. Phosphatidylcholine hydroperoxide (PCOOH) was detected by absorbance of conjugated dienes at 234 nm. A PCOOH standard was prepared using 15-lipoxygenase (Cayman Chemical Company Inc., Ann Arbor, MI) and 1-palmitoyl-2-linolenoyl-sn-glycero-3-phosphocholine (Sigma) as reported previously (2).
Phospholipid Hydroperoxide Detection in Real Time-DPPP was used as a membrane-localized fluorescent probe that reacts specifically with hydroperoxides and becomes highly fluorescent when oxidized (26). Control, antisense and random oligonucleotide-treated L2 cells adherent to 12 ϫ 25-mm plastic slides (Aclar; Allied Signal, Morristown, NJ) were grown in 60-mm culture dishes (two identical slides/ dish) in complete medium. The slides were washed twice with PBS and placed in a standard quartz cuvette (10 ϫ 10 mm) containing 50 M DPPP in PBS. After incubation at 37°C for 10 min, the slides were washed twice with fresh PBS and placed in a standard quartz cuvette at a 45°angle to the excitation beam of a spectrofluorometer (PTI, Bricktown, NJ). After recording the initial emission spectrum, the increase in fluorescence (excitation, 351 nm; emission 380 nm, excitation slit, 1 nm; emission slit, 3 nm) was measured continuously before and after the addition of 10 M Cu 2ϩ and 0.5 mM ascorbate (Asc).
Detection of Cell Damage-To detect the early stages of injury, translocation of phosphatidylserine in the cell membrane was examined by the annexin V assay according to the manufacturer's instructions (Clontech, Palo Alto, CA). At 0, 28, or 48 h after antisense or random morpholino oligonucleotide treatment, the cells were labeled in situ with annexin V-FITC and propidium iodide (PI), and six randomly selected fields were counted for fluorescent cells using a Nikon Diaphot microscope. Subsequent image capture and analysis was performed using Metamorph software (Universal Imaging, Downingtown, PA). The number of stained cells was normalized to the total number of cells as counted by phase microscopy of the same field.
DNA cleavage, a characteristic of apoptosis, was detected by the TdT-mediated dUTP nick end labeling (TUNEL) assay as described by the manufacturer (Roche Molecular Biochemicals). At 5, 28, or 48 h after antisense or random oligonucleotide treatment, L2 cells cultured on two chamber slides were washed twice with PBS and fixed with 4% paraformaldehyde in PBS at room temperature for 20 min. The fixed cells were permeabilized with 0.1% Triton X-100, 0.1% sodium citrate at 4°C for 2 min and incubated for 1 h at 37°C with TUNEL reaction mixture containing the terminal deoxynucleotidyl transferase. 4Ј6ЈDiamidino-2-phenylindole dihydrochloride (DAPI, 300 nM) was used to counterstain double-stranded DNA in nuclei after TUNEL staining. The cells were air-dried, and coverslips were applied with 3 l of Mowiol (Calbiochem, San Diego, CA). Images of nuclear fluorescence typical of apoptotic cells were collected by fluorescence microscopy as described above.
Data Analysis-The data are expressed as the means Ϯ S.E. Statistical significance was evaluated by Mann Whitney t test or one-way analysis of variance, followed by comparison of the treated versus untreated by the Bonferroni procedure using SigmaStat software (Jandel, San Rafael, CA). The level of significance was taken as p Ͻ 0.05.

RESULTS
Cell Growth-L2 cells treated with antisense oligonucleotide grew poorly compared with control and random oligonucleotide-treated cells and exhibited a round cell shape by microscopy with a high rate of detachment from the dishes (data not shown). Cellular morphology and cell attachment in antisense oligonucleotide-treated cells was improved significantly by the addition of Trolox to the culture medium.
Effect of Antisense Oligonucleotide on 1-cysPrx Expression-Control L2 cells expressed 1-cysPrx at ϳ35% of the level observed in primarily isolated rat type II alveolar epithelial cells after 24 h in culture as determined by Northern and immunoblots (Fig. 1). Control (vehicle only) and random oligonucleotide-treated L2 cells showed no change in immunoreactive 1-cysPrx during 48 h of culture (Fig. 2). By contrast, treatment of L2 cells with the antisense oligonucleotide for 48 h decreased the immunodetectable 1-cysPrx protein level by ϳ60% ( Fig.  2A). Antisense oligonucleotide treatment had no significant effect on 1-cysPrx protein expression in the primarily isolated rat type II cells (data not shown). The reason for the difference in efficacy was evaluated by the use of fluoroscein-labeled 1-cysPrx antisense oligonucleotide, which showed a 5-fold higher fluorescence in the L2 cells versus type II alveolar epithelial cells (data not shown), indicating resistance to transfection in the primarily isolated cells.
Enzymatic activities were evaluated in crude cell extracts from vehicle only (control) and antisense and random oligonucleotide-treated L2 cells at 48 h of culture. Antisense oligonucleotide-treated cells showed decreased activity by 44% for GPx and 48% for PLA 2 relative to control, whereas random oligonucleotide had no effect on activities (Fig. 2B).
Detection of Cellular Phospholipid Peroxidation-Analysis of lipid extracts of antisense and random oligonucleotide-treated L2 cells by HPLC showed similar levels of PCOOH at the start of incubation (Fig. 3A). After 48 h of culture, antisense oligonucleotide-treated cells showed a marked increase in the phospholipid hydroperoxide peak on the HPLC tracing, whereas the corresponding peak in random oligonucleotide-treated cells was unchanged (Fig. 3A). The increase in PCOOH at 48 h of culture was ϳ3.5-fold (p Ͻ 0.05) in antisense oligonucleotide compared with random oligonucleotide-treated L2 cells (Fig.  3B). Accumulation of PCOOH in antisense oligonucleotidetreated cells was significantly reduced by the addition of Trolox to the culture medium (Fig. 3B). Lipid peroxidation was analyzed further by use of the specific fluorescent probe, DPPP. Control (vehicle only) L2 cells cultured for 48 h, then labeled with DPPP, and analyzed in a fluorometer showed low level fluorescence emission with a peak at 380 nm (Fig. 4A). Random oligonucleotide-treated cells showed a similar level of DPPP fluorescence. There was approximately four times more oxidized DPPP fluorescence in antisense oligonucleotide-treated L2 cells after 48 h of incubation as compared with random oligonucleotide-treated and control cells (Fig. 4A). These results indicate lipid peroxidation of antisense oligonucleotidetreated cells during growth in normal culture medium.

FIG. 1. Expression of 1-cysPrx in rat type II cells and L2 cells detected by northern (A) and immunoblot (B) analysis.
Total RNA (5 g) from rat lung alveolar type II epithelial (Type II) or L2 cells were hybridized with a 32 P-labeled rat full-length 1-cysPrx cDNA probe. Total soluble protein (30 g) from each cell lysate was immunoblotted with an anti-1-cysPrx polyclonal antibody. The levels of 1-cysPrx expression were normalized with 18 S rRNA for the mRNA or actin for the proteins. The values for L2 cells are indicated as percentages of the corresponding value in type II cells (100%).

FIG. 2. Treatment of L2 cells with antisense or random morpholino oligonucleotides or with vehicle (ethoxylated polyethylenimine) alone (control).
A, cells were harvested 24 or 48 h posttreatment, and 30 g of total soluble protein from each lysate were immunoblotted with an anti-1-cysPrx polyclonal antibody. Expression of 1-cysPrx protein was normalized to that for actin as indicated by the ratio below each lane. B, GPx and lysosomal-type phospholipase A 2 (aiPLA 2 ) activities were measured from the same cell homogenate at 48 h post-treatment. Each data set represents the mean Ϯ S.E. of three independent experiments. *, p Ͻ 0.05 compared with the corresponding activity for random oligonucleotide or control.
The Cu 2ϩ -Asc chemical reaction that results in the generation of ⅐ OH (27) and DPPP were used to evaluate lipid peroxidation in antisense oligonucleotide-treated cells in response to oxidative stress. DPPP fluorescence of cells was normalized to background fluorescence before the Cu 2ϩ -Asc treatment. Cu 2ϩ -Asc treatment of L2 cells showed an instantaneous increase in DPPP fluorescence with a rate of fluorescence change that was ϳ60% greater in antisense oligonucleotide-treated cells as compared with that of random oligonucleotide-treated or vehicle only (control) cells (Fig. 4B). This suggests that antisense oligonucleotide-treated cells with decreased 1-cysPrx have a diminished capacity to reduce plasma membrane phospholipid hydroperoxides generated by oxidant stress.
Detection of Cell Injury-To determine whether antisense oligonucleotide treatment resulted in translocation of phosphatidylserine to the outer leaf of the plasma membrane, we examined cellular binding of annexin V as an index of perturbation of the membrane bilayer. PI staining was used as a marker of increased plasma membrane permeability. Approximately 10% of cells showed annexin V and PI fluorescence at zero time (prior to treatment with oligonucleotide) (Fig. 5A). No change in annexin V binding or PI uptake was seen at 5 h of incubation, but a significant increase in annexin V-and PIlabeled cells (ϳ40%) was detected at 28 h after antisense treatment with a further increase to ϳ80% at 48 h (Fig. 5). In addition, there was a marked decrease in the number of cells that remained attached to the plate as shown in the phase contrast image (Fig. 5B). By contrast, control (vehicle only; not shown) or random oligonucleotide-treated cells did not show any increase in annexin V-or PI-labeled cells during the 48-h incubation, and the number of attached cells did not appear to be affected significantly (Fig. 5).
The development of apoptosis was assessed by DNA fragmentation through TUNEL assay. The appearance of fluorescence in the nuclei of antisense-treated cells shows evidence of apoptosis at 48 h, whereas only minimal DNA fragmentation was observed in cells treated with random oligonucleotide (Fig.  6) or in control (vehicle only) cells (not shown). There is a marked decrease in the number of antisense oligonucleotide-treated cells on the culture dish compared with the random oligonucleotide-treated cells as shown by phase contrast and DAPI staining (Fig. 6). Pretreatment with Trolox prevented the development of apoptosis and cell loss in antisense oligonucleotide-treated cells (Fig. 6).
Ad.1-cysPrx-infected L2 Cells-To further investigate whether the decrease in 1-cysPrx expression was linked to apoptosis, 1-cysPrx was overexpressed in cells by adenovirusmediated transfection. The efficiency of L2 cell infection with adenovirus was examined with Ad.GFP. At a multiplicity of infection of 2, 5, or 10 plaque forming units/cell, transfection efficiency as determined by fluorescence microscopy was ϳ50, ϳ80, and ϳ98% at 24 h post-infection, respectively (data not shown). The cells were then treated with adenovirus harboring the coding region of 1-cysPrx cDNA. Northern blot analysis of total RNA using a full-length 1-cysPrx cDNA probe revealed the presence of a more rapidly migrating 1-cysPrx message only in Ad.1-cysPrx-infected L2 cells and not in Ad.GFP-infected or noninfected (control) L2 cells (Fig. 7A). This truncated message (about ϳ0.7 kb) is consistent with the vector design of Ad.1-cysPrx, which does not contain the 3Ј-untranslated region. Immunoblot analysis showed increased expression of 1-cysPrx protein in cells infected with Ad.1-cysPrx, confirming that the 1-cysPrx mRNA detected by Northern hybridization was correctly translated in the cells (Fig. 7B). The 1-cysPrx protein content was not increased by infection of cells with Ad.GFP.
To study whether 1-cysPrx overexpression could "rescue" L2 cells from the effect of antisense treatment, cells were incubated with Ad.1-cysPrx at an multiplicity of infection of 8 plaque forming units/cell for 2 h prior to exposure to antisense or random oligonucleotides. For comparison, L2 cells were infected with the same multiplicity of infection of Ad.GFP or were left untreated (control). Immunoblot analysis of antisense oligonucleotide-treated cells showed a significant increase in expression of 1-cysPrx at 48 h after infection with Ad.1-cysPrx to levels that were comparable with vehicle only and random oligonucleotide-treated cells that were not infected with adenovirus (Fig. 7C). Increases in 1-cysPrx expression after Ad.1-cysPrx infection also were observed in vehicle only and random oligonucleotide-treated cells (Fig. 7C). There was no change in 1-cysPrx expression compared with control in cells infected with Ad.GFP (Fig. 7C).
The effect of Ad.1-cysPrx infection on lipid peroxidation was determined by DPPP fluorescence. Control (noninfected) cells treated with antisense oligonucleotide showed a marked increase in DPPP fluorescence of the cell extracts compared with vehicle only and random oligonucleotide-treated cells (Figs. 4A and 8). Antisense oligonucleotide-treated L2 cells infected with Ad.1-cysPrx demonstrated levels of DPPP fluorescence that were similar to random oligonucleotide-treated cells (Fig. 8).
Both DNA fragmentation as measured by the TUNEL assay and loss of cells from the plate as shown by phase contrast and DAPI staining were attenuated in the antisense-treated cells in response to Ad.1-cysPrx infection (Fig. 6). Ad.1-cysPrx infection had no effect on the results of TUNEL assay in random oligonucleotide-treated cells (not shown). These results suggest that a certain level of 1-cysPrx is essential to protect L2 cells from apoptosis mediated by ROS generated during normal cellular metabolism.

DISCUSSION
Prx have been classified into two major groups, one containing a single conserved cysteine in the N-terminal region (1-cysPrx) and the other (2-cysPrx) containing an additional conserved cysteine near the C-terminal region, which is separated from the first cysteine by roughly 120 amino acids (28,29). The presence or absence of the second cysteine is correlated with the consensus sequence neighboring the first cysteine: Pro-Val-Cys-Thr for 1-cysPrx and Phe-Val-Cys-Thr for 2-cysPrx. The

FIG. 7. Analysis of L2 cells infected with adenovirus expressing GFP (Ad.GFP) or 1-cysPrx (Ad.Prx), or noninfected control (Con). A, Northern blot of L2 cells treated with adenovirus compared
with control (no adenovirus). Total RNA (5 g) from the cells was hybridized with a 32 P-labeled full-length rat 1-cysPrx cDNA probe. The same membrane was rehybridized with 32 P-labeled rat full-length glyceraldehyde-3-phosphate dehydrogenase cDNA as a control for RNA loading. The positions of the mRNA for endogenous 1-cysPrx (Endo) and exogenous 1-cysPrx encoded by the adenovirus (Exog) are indicated. B, soluble protein (30 g) was loaded onto each lane for 12% SDS-PAGE, followed by immunoblot analysis with an anti-1-cysPrx polyclonal antibody. Immunoblot for actin is shown as a control for protein loading. C, effect of Ad.1-cysPrx on 1-cysPrx expression. L2 cells were infected with Ad.GFP or Ad.1-cysPrx (Ad.Prx) prior to treatment with antisense or random oligonucleotides or vehicle alone. Cells from each plate were harvested, and 30 g of total soluble protein were immunoblotted using an anti-1-cysPrx polyclonal antibody. Actin was used as a control for protein loading.
Prx are distinct from other peroxidases in that they do not utilize cofactors such as metals or prosthetic groups. Members of both groups of Prx exist as homodimers with two monomers oriented in a head-to-tail manner (30,31). The 2-cysPrx, also referred to as thioredoxin peroxidases, reduce peroxides with electrons provided by thioredoxin (28,29). 1-cysPrx, also named Prx VI, is capable of reducing peroxides in the presence of small thiols such as dithiothreitol (1,2,7) or glutathione (2,3,4,6). We have recently demonstrated that depletion of glutathione diminishes activity of the enzyme in intact cells, indicating that it can serve as a physiological electron donor (23).
Recombinant human 1-cysPrx has peroxidase activity against H 2 O 2 , small organic hydroperoxides, fatty acid hydroperoxides, and PLOOH in vitro (2). The present study provides evidence that 1-cysPrx functions as an essential antioxidant enzyme to detoxify PLOOH as a result of ROS generated in intact cells. This result complements our previous demonstration that overexpression of 1-cysPrx in NCI-H441 cells protects against ⅐ OH-mediated phospholipid peroxidation (23). Additional evidence for an antioxidant role for 1-cysPrx is its induction in human umbilical vein endothelial cells by hydroperoxides (32), in rat lungs immediately after birth, (24,33) and in kidney following intraperitoneal injection of chloroform (33). Thus, 1-cysPrx is induced by oxidative stress, and its expression appears to protect against oxidative-mediated cell injury.
We chose an antisense strategy to decrease 1-cysPrx expression in L2 cells, an epithelial cell line that was derived originally through clonal isolation from rat lung alveolar epithelial cells in primary culture (34). A morpholino antisense oligonucleotide, 25 bases in length, was designed to target the region overlapping the translation start site, thus inhibiting 1-cysPrx synthesis. This mechanism is different from the more commonly used phosphorothioate oligomers in which the RNA-DNA duplex that is formed becomes accessible to RNase H, leading to mRNA degradation. Advantages of morpholino-substituted oligonucleotides include complete resistance to nucleases, good targeting predictability, excellent sequence specificity, minimal nonantisense activity, and less toxicity (35,36). A relatively minor disadvantage is that an effect of the antisense oligomer cannot be detected by Northern blot analysis.
Treatment with 1-cysPrx antisense oligonucleotides, but not with control or random oligonucleotides, led to a significant reduction of 1-cysPrx expression and a proportional decrease in both peroxidase and PLA 2 activities in L2 cells at 48 h post treatment. The decrease in the 1-cysPrx protein level was associated with diminished cell viability and increased cell membrane hydroperoxide content as indicated by HPLC analysis and DPPP fluorescence. The increased hydroperoxide content was most likely the result of decreased reduction of PLOOH formed in cellular membranes because of endogenous oxidants.
To show that the presence of increased PLOOH perturbed the lipid bilayer, we used the annexin V-FITC test for phosphatidyl serine translocation to the outer leaf of the bilayer (37). Annexin V-FITC-associated fluorescence was detected after 28 h of incubation with antisense oligonucleotide, and nearly all cells were positive at 48 h. These cells also showed increased plasma membrane permeability by the PI test, indicating damage to cell membranes. A positive TUNEL assay indicating an apoptotic death was detected at 48 h of incubation but not at the earlier time point. These data suggest that peroxidative injury to the plasma membrane is followed by the development of apoptosis in antisense oligonucleotide-treated L2 cells incubated for 28 -48 h in cell culture medium. Pretreatment of cells with Trolox, a soluble form of vitamin E, prevented both lipid peroxidation and apoptosis, providing evidence for their linkage. Ad.1-cysPrx infection restored 1-cysPrx expression to control levels and also prevented cellular apoptosis, confirming that the effects of antisense treatment were due to the decreased cellular expression of 1-cysPrx. Thus, 1-cysPrx appears to function as an important cellular antioxidant enzyme in this lung epithelial cell line. 1-cysPrx infected (ϩAd.Prx) and noninfected cells (Control) that were then treated with antisense or random oligonucleotide or vehicle only as described in the legend to Fig. 7. Lipid peroxidation was measured by incubation of cells with DPPP followed by extraction and assay of fluorescence. Fluorescence was normalized for total protein concentration in the cell extracts. Each data set represents the means Ϯ S.E. of three independent experiments. *, p Ͻ 0.05 for corresponding value of random oligonucleotide-treatment. Adenovirus infection of control cells was not evaluated. #, p Ͻ 0.05 for uninfected compared with the corresponding value for adenovirus-infected cells.