Redox Regulation of Adenovirus-Induced AP1 Activation by Overexpression of Manganese-Containing Superoxide Dismutase

  2002, 76(1):355. DOI: 10.1128/JVI.76.1.355-363.2002. J. Virol. 

Frederick E. Domann, Larry W. Oberley and Kevin C. KregelHannah J. Zhang, Victoria J. Drake, Linjing Xu, Jianfang Hu, DismutaseManganese-Containing SuperoxideAP-1 Activation by Overexpression of Redox Regulation of Adenovirus-Induced

http://jvi.asm.org/content/76/1/355Updated information and services can be found at:

These include:

REFERENCEShttp://jvi.asm.org/content/76/1/355#ref-list-1at:

This article cites 36 articles, 11 of which can be accessed free

CONTENT ALERTS more»articles cite this article),

Receive: RSS Feeds, eTOCs, free email alerts (when new

http://journals.asm.org/site/misc/reprints.xhtmlInformation about commercial reprint orders: http://journals.asm.org/site/subscriptions/To subscribe to to another ASM Journal go to:

on Novem

ber 13, 2013 by guesthttp://jvi.asm

.org/D

ownloaded from

on N

ovember 13, 2013 by guest

http://jvi.asm.org/

Dow

nloaded from

JOURNAL OF VIROLOGY,0022-538X/02/$04.00�0 DOI: 10.1128/JVI.76.1.355–363.2002

Jan. 2002, p. 355–363 Vol. 76, No. 1

Copyright © 2002, American Society for Microbiology. All Rights Reserved.

Redox Regulation of Adenovirus-Induced AP-1 Activation byOverexpression of Manganese-Containing

Superoxide DismutaseHannah J. Zhang,1 Victoria J. Drake,1 Linjing Xu,1 Jianfang Hu,2 Frederick E. Domann,2

Larry W. Oberley,2 and Kevin C. Kregel1*Department of Exercise Science1 and Free Radical and Radiation Biology Program, Department of

Radiation Oncology,2 The University of Iowa, Iowa City, Iowa 52242

Received 9 August 2001/Accepted 20 September 2001

Adenovirus gene therapy is a promising tool in the clinical treatment of many genetic and acquired diseases.However, it has also caused pathogenic effects in organs such as the liver. The redox-sensitive transcriptionfactors AP-1 and NF-�B have been implicated in these effects. To study the mechanisms of adenovirus-mediated AP-1 and NF-�B activation and the possible involvement of oxidative stress in adenovirus trans-duction, rats were injected with either replication-defective recombinant adenovirus with DNA containing thecytomegalovirus promoter region only (AdCMV), adenovirus containing human manganese-containing super-oxide dismutase (MnSOD) cDNA (AdMnSOD), or vehicle. Compared to vehicle and AdCMV transduction,MnSOD gene transfer yielded a fivefold increase in liver MnSOD activity 7 days postinjection. Gel shift assayshowed that AdCMV transduction induced DNA binding activity for AP-1 but not NF-�B. MnSOD overex-pression abolished this activation. Western blotting analysis of c-Fos and c-Jun suggested that up-regulationof c-fos and c-jun gene expression does not directly contribute to the induction of AP-1 activation. Glutathione/glutathione disulfide ratios were decreased by adenovirus transduction and restored by MnSOD overexpres-sion. The AP-1 binding activity that was induced by AdCMV was decreased by immunoprecipitation of Ref-1protein. Ref-1 involvement was confirmed by restoration of AP-1 binding activity after the immunoprecipitatedRef-1 protein had been added back. AP-1 DNA binding activity was also elevated in control and AdMnSOD-injected rats after addition of the immunoprecipitated Ref-1 protein. These data indicate that cellular trans-duction by recombinant adenovirus stimulates AP-1 DNA binding activity. Furthermore, our results suggestthat MnSOD overexpression decreases AP-1 DNA binding activity by regulating intracellular redox status, withthe possible involvement of Ref-1 in this redox-sensitive pathway.

Gene therapy is a promising tool for the clinical treatment ofmany genetic and acquired diseases. The success of gene ther-apy relies largely on the delivery systems that transfer targetgenes into cells and lead to gene expression. Recombinantadenoviruses have been developed as one of these deliverysystems. These recombinant adenoviruses are in general rep-lication defective, because a large portion of the genes (such asE1 and E3 genes) in these viruses have been replaced byforeign genes. This system provides many advantages overother conventional delivery systems, including (i) the ability toproduce extremely efficient gene transduction with high levelsof recombinant gene expression in a variety of cellular targets,including both quiescent and dividing cells (22), (ii) the possi-bility of large-scale production, and (iii) the ability of the virusto be engineered to accommodate a broad range of transgenesizes.

However, in recent years, problems associated with recom-binant adenovirus gene therapy have arisen (21, 30, 34). One ofthe major problems is cytotoxicity following injection with ad-enovirus in vivo. For example, systemic application of the firstgeneration of adenovirus resulted in liver damage and necrosis

(20). The exact mechanisms by which infection with the repli-cation defective virus can cause cytotoxicity are not clear. How-ever, systemic symptoms that have been observed after in vivotransduction of recombinant adenovirus, such as shock, fever,and inflammation, are similar to the in vivo stress responsenoted in many other pathological conditions. Therefore, it isreasonable to speculate that recombinant adenovirus infectioncan result in a stress response at both the systemic and cellularlevels. Importantly, these stress responses may play a role inthe cytotoxicity observed with adenoviral administration.

Eukaryotic organisms respond to stress by increasing stressresponse gene expression. Several signal transduction cascadesare usually involved in the activation of stress response pro-teins. NF-�B and AP-1 are widely recognized as two of theearly-response transcriptional factors that participate in thesesignal transduction cascades (19, 31). NF-�B and AP-1 aresensitive to changes in cell environment and activate theirtarget genes by binding to specific motifs on the regulatoryregions of stress response genes. Therefore, it is tenable topostulate that DNA binding activity of NF-�B and AP-1 can beinduced by adenovirus transduction. In fact, a recent studydemonstrated that NF-�B and AP-1 were up-regulated by re-combinant adenovirus transduction (24). However, the mech-anisms responsible for induction of NF-�B and AP-1 adenovi-rus have not been delineated.

The NF-�B DNA binding complex is composed of homo-

* Corresponding author. Mailing address: Integrative PhysiologyLaboratory, 532 FH, The University of Iowa, Iowa City, IA 52242.Phone: (319) 335-7596. Fax: (319) 335-6966. E-mail: [email protected].

355

on Novem

ber 13, 2013 by guesthttp://jvi.asm

.org/D

ownloaded from

dimers or heterodimers of the NF-�B family members (i.e.,p50 and p65). The activation of NF-�B is controlled by itsinhibitory protein, I�B. In most cells, NF-�B is sequestered inan inactive, cytoplasmic complex by binding to I�B. Manystress factors can stimulate I�B kinase (IKK)-mediated phos-phorylation of I�B, leading to its ubiquitination and degrada-tion (27). The removal of I�B allows the NF-�B complex to betranslocated to the nucleus and act as a transcriptional activa-tor. In a different pathway, AP-1 activation takes place at boththe transcriptional and posttranscriptional levels. First, thetranscription of AP-1 family members, Jun and Fos, can beup-regulated after stress (9). A second mechanism for AP-1activation includes redox regulation of Fos and Jun DNA bind-ing activity (1). Redox factor 1 (Ref-1) is thought to function inthis pathway (10). Ref-1, which was first found to be a DNArepair enzyme, apurinic/apyrimidinic endonuclease (12), is alsonow known to be a redox-sensing signal transduction proteinand to participate in the modulation of several transcriptionfactors, such as NF-�B (12) and AP-1 (36).

NF-�B and AP-1 are two transcriptional factors known torespond directly to oxidative stress. Oxidative stress has beenlinked to pathological cell death from many insults, such asischemia-reperfusion, trauma, and hyperthermia, concomitantwith induction of NF-�B and AP-1 activation in these condi-tions (3). Therefore, oxidative stress may also participate in theadenovirus-induced stress response. To study this possibility,we manipulated the in vivo levels of manganese-containingsuperoxide dismutase (MnSOD) and modified the activation ofAP-1 and NF-�B. MnSOD, which is found in mitochondria, isone of the primary antioxidant enzymes that catalyze the dis-mutation of superoxide to hydrogen peroxide. Hydrogen per-oxide is further reduced to water by catalase or one of theperoxidases. Thus, overexpression of MnSOD creates a mech-anism for reactive oxygen species (ROS) removal and, in turn,modulates intracellular redox status.

The overall purpose of the present study was to test thehypothesis that AP-1 and NF-�B DNA binding activity can beinduced by in vivo recombinant adenovirus transduction. Wealso speculated that if oxidative stress is involved in this induc-tion, AP-1 and NF-�B DNA binding activity could be modu-lated by MnSOD overexpression through a redox regulatorymechanism. Finally, we postulated that Ref-1 may be part ofthis regulation pathway. Our results indicated that in vivo ad-ministration of adenovirus stimulated AP-1 but not NF-�BDNA binding activity in the liver of rats. This activation doesnot appear to be due to the up-regulation of c-fos and c-jungene expression. Instead, the induction of oxidative stress byadenovirus most likely contributed to the activation of AP-1DNA binding. Furthermore, our findings suggest that overex-pression of MnSOD suppresses adenovirus-induced AP-1DNA binding activity by regulating intracellular redox status,with the possible involvement of Ref-1 in this redox-sensitivepathway.

MATERIALS AND METHODS

Animals. Male Fischer 344 rats (350 to 450 g; 24 months old; National Instituteon Aging) were used in these experiments. Rats were housed in The Universityof Iowa Animal Care Facility, and all experimental procedures conformed toinstitutional animal care guidelines. Animals were maintained at 22 to 24°C on

a 12-h:12-h light-dark cycle and provided food (standard rat chow) and water adlibitum.

Adenovirus gene transfer. Replication-defective recombinant adenovirus type5 with the E1 region replaced with DNA containing the cytomegalovirus (CMV)promoter region only (AdCMV) or with a specific cDNA (e.g., MnSOD) afterthe CMV promoter were provided by the Gene Transfer Vector Core at TheUniversity of Iowa. For all experiments, the adenovirus concentration was 1012

particles per ml in phosphate-buffered saline (PBS) buffer with 3% sucrose. Thenumber of infectious units of the adenovirus was typically 1010 PFU/ml.

To examine the efficiency of recombinant adenovirus transduction and geneexpression in the liver, recombinant adenovirus containing �-galactosidase (�-Gal) cDNA was used as a reporter gene. Fischer 344 rats anesthetized withsodium pentobarbital (40 mg/kg intraperitoneally) were given either recombi-nant adenovirus containing LacZ cDNA (n � 12) or vehicle (3% sucrose in PBS;n � 3) via carotid catheter injection and then housed for 5, 7, 10, or 14 days. Onthe designated day, tissues were harvested from both the small and large lobes ofthe liver and processed for X-Gal (5-bromo-4-chloro-3-indolyl-�-D-galactopy-ranoside) staining.

In subsequent experiments, rats were given either AdCMV (empty vector; n �3), an adenovirus containing human MnSOD cDNA (AdMnSOD; n � 3), orvehicle alone (3% sucrose in PBS; n � 3) via carotid catheter injection. Each ratreceived 1 ml of adenovirus (1010 infectious units) and was returned to itsindividual cage. On the designated experimental day, animals were given anoverdose of sodium pentobarbital (80 mg/kg intraperitoneally). Livers were col-lected, rinsed in PBS, and processed immediately for an experiment or frozen inliquid nitrogen.

X-Gal staining for �-Gal activity. X-Gal staining for �-Gal activity was per-formed as previously described (6). Tissues (0.125 cm3) were harvested fromboth the small and large lobes of the liver, fixed in 4% paraformaldehyde fixativesolution, incubated at 4°C in PBS containing 2 mM MgCl2 and 30% sucrose, andembedded in tissue-freezing medium. Cryostat sections (10 �m) were cut andmounted on glass slides. The slides were then stained with X-Gal solution andcounterstained with fast red to permit calculation of the percentage of cellsexpressing the �-Gal protein as a result of LacZ cDNA uptake and expression.The numbers of cells expressing �-Gal (blue) and not expressing �-Gal (red)were counted by microscopic viewing. Several fields from each slide werecounted in order to obtain an accurate estimation of cell number. Final resultswere expressed as the percent �-Gal-positive cells from an average of threeanimals in each group.

Immunohistochemistry. Formalin-fixed tissues were processed for immuno-histochemistry as previously described (14, 15). Sections (4 �m) cut from paraffinblocks were evaluated for immunoreactive MnSOD proteins using immunoper-oxidase staining techniques and antibodies specific for MnSOD. Two slides wereprepared from both a large and small lobe of each liver sample for the staining.The specificity of the antibodies has been described elsewhere (37). Normalrabbit serum was substituted for the MnSOD antibody as a negative control.

Tissue preparation. Liver samples were prepared for Western blotting andenzyme activity measurements as previously described (14). Frozen liver sampleswere ground in liquid nitrogen, homogenized in 50 mM phosphate buffer with ahomogenizer, and sonicated with a Vibra cell sonicator (Sonics & Materials).Total protein concentrations were determined by the Bradford assay (Bio-Rad).

Western blot analysis. Equal amounts of total protein or nuclear extracts wereseparated by sodium dodecyl sulfate–12.5% polyacrylamide gel electrophoresisaccording to the method of Laemmli (23). The separated proteins were thentransferred onto nitrocellulose membranes and blocked with 5% dry milk inTris-buffered saline with 0.1% Tween 20 (TTBS). The nitrocellulose membraneswere incubated overnight at 4°C with an antibody (1:1,000 for c-Fos and c-Jun,1:25 for Ref-1) specific for the proteins of interest. After three washings withTTBS, the membranes were incubated with a secondary antibody conjugatedwith horseradish peroxidase (1:10,000) in TTBS for 1 h at room temperature.Blots were stained by the chemiluminescent ECL method (Amersham LifeSciences), and immunoreactive signals were visualized by exposure to an X-rayfilm (Kodak).

MnSOD activity assay. SOD activity was measured by the modified nitrobluetetrazolium (NBT) method described by Spitz and Oberley (35). This is anindirect assay based on a competition reaction between SOD and the superoxideindicator molecule NBT. The rate of increase in the absorption at 560 nm overa 5-min period indicates the reduction of NBT by superoxide. The competitiveinhibition of this reaction is an indicator of total SOD activity. In the assay, thexanthine-xanthine oxidase system was used to generate superoxide. Variousamounts of total protein were added to the reaction until maximal inhibition wasobtained, as determined by spectrophotometry. Total SOD activity was deter-mined by the amount of protein necessary for half-maximal inhibition of the

356 ZHANG ET AL. J. VIROL.

on Novem

ber 13, 2013 by guesthttp://jvi.asm

.org/D

ownloaded from

NBT reaction. MnSOD activity was quantified in the presence of 5 mM NaCN,which inhibits only copper- and zinc-containing SOD (CuZnSOD) activity. Oneunit of activity was defined as the concentration of SOD that reduced the NBTreaction to one-half of the maximum.

SOD activity gel assay. The native activity gel assay described by Beauchampand Fridovich (5) was used. Fifty micrograms of total protein from each samplewas separated in a native 12% polyacrylamide gel. The gel was then stained byincubation with 2.43 mM NBT, 28 mM riboflavin, and 28 mM TEMED(N,N,N�,N�-tetramethylethylenediamine) for 20 min in the dark. After the gelhad been exposed to a fluorescent light, the achromatic bands corresponding toSOD activity appeared on a dark blue background.

Preparation of nuclear extracts. Nuclear extracts from rat livers were preparedusing a modification of the method of Hattori et al. (16). All procedures werecarried out at 4°C. Freshly harvested liver tissues were minced on ice, weighed,and placed in homogenizing buffer (0.3 M sucrose, 10 mM HEPES [pH 7.6],10 mM KCl, 0.74 mM spermidine, 0.15 mM spermine, 0.1 mM EDTA, 0.1 mMEGTA, 1 mM dithiothreitol [DTT], 0.5 mM phenylmethylsulfonyl fluoride, andone complete protease inhibitor cocktail tablet [Boehringer Mannheim] per50 ml of buffer). The buffer-tissue mixture (2 ml of buffer per g of tissue) washomogenized with a motor-driven Teflon pestle tissue homogenizer and thenmixed with 2 volumes of cushion buffer (same components as the homogenizingbuffer except that the sucrose concentration was 2.2 M). The homogenate-cushion buffer solution was then layered over 5 ml of cushion buffer in 38.5-mlpolyallomer tubes. Homogenates were centrifuged in a Beckman ultracentrifugeat 105,000 � g for 150 min. The supernatant, containing the cytoplasmic fraction,was decanted, and buffer was removed. Each nuclear pellet was resuspended in500 �l of nuclear lysis buffer (20 mM HEPES [pH 7.9], 25% glycerol, 420 mMNaCl, 1.5 mM MgCl2, 0.2 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride,and 0.5 mM DTT), briefly homogenized in a Dounce homogenizer, and gentlyrotated for 40 min at 4°C. Nuclear membranes were then pelleted by spinning thenuclear lysate in an ultracentrifuge at 35,730 � g for 11 min. The resultantsupernatants were collected as nuclear fractions. Their total protein contentswere determined by using the Bradford assay.

EMSA. Nuclear extract (5 �g; 5 �l) was combined with electrophoretic mo-bility shift assay (EMSA) buffer (250 mM KCl, 100 mM HEPES [pH 7.9], 25%glycerol, 5 mM EDTA, 5 mM DTT), 1 �l of bovine serum albumin (1 �g/�l), 1 �lof poly(dI-dC) oligonucleotide (1 �g/�l), and 6.6 �l of water to form a reactionmixture. For supershift assays, 5 �l of antibody specific for AP-1 (c-Jun) orNF-�B (p50 and p65) was incubated with the nuclear extracts 2 h before theEMSA. Oligonucleotides (2.5 pmol) with the consensus AP-1 binding sequence(5�-CGCTTGATGACTCAGCCGGAA-3�) or the consensus NF-�B binding se-quence (5�-AGTTGAGGGGACTTTCCCAGG-3�) were radioactively end la-beled with [�-32P]dATP (50 Ci/mmol; Amersham) by T4 polynucleotide kinase(New England Biolabs). For each sample, 2 �l of the radiolabeled probe wascombined with the reaction mixture. After a 30-min incubation at room temper-ature, bound and free oligonucleotides were separated on a 6% nondenaturingpolyacrylamide gel. The extent of AP-1 or NF-�B DNA binding activity wasdetected by autoradiography.

Immunodepletion of Ref-1 protein. Five microliters of a specific anti-Ref-1antibody and 25 �l of protein A-conjugated agarose beads were added to 20 �lof one nuclear extract from the vector samples. After overnight gentle shaking at4°C, the supernatant was separated from the agarose beads by centrifugation at12,000 � g in a microcentrifuge for 5 min. The supernatant was then collected asimmunodepleted nuclear extract for EMSA. The pellet was further washed eighttimes with TTBS to remove the free residual chemicals from the immunopre-cipitation reaction mixture before it was either reacted directly with oligonucle-otides containing the consensus AP-1 binding sequence or added to variousnuclear extracts for EMSA.

GSH and GSSG measurement. The method for glutathione (GSH) measure-ment was adapted from previously described techniques (2, 13). Total GSH ofsamples was measured based on the colorimetric reaction of DTNB (5,5�-dithio-bis 2-nitrobenzoic acid) with GSH to form TNB (5-thio-2-nitrobenzoic acid). Therate of TNB formation, which was proportional to GSH concentration, wasmonitored spectrophotometrically at 412 nm. Cellular glutathione disulfide(GSSG) was reduced to GSH by the specific GSSG reductase (GR). To prepareliver samples, ground tissues were homogenized in ice-cold 5% 5-sulfosalicylicacid. The measurement of total GSH was started by mixing 700 �l of 2 mMNADPH with 100 �l of 6 mM DTNB in phosphate buffer (pH 7.4). The sampleswere diluted with water to make a total volume of 150 �l and added to thereaction solution. The assay was initiated by addition of 50 �l of GR solution.The rate of TNB formation was monitored at 412 nm. The total GSH of a samplewas calculated from a standard curve of GSH concentration versus rate (changein absorbance/time). The GSSG level was determined by the same DTNB assay

when GSH was masked by 2-vinylpyridine. In brief, 2 �l of 2-vinylpyridine wasadded to 100 �l of sample prepared as noted above for total GSH measurement.After a 90-min incubation on ice, GSSG was measured as described for the totalGSH assay.

RNase protection assay. Total RNA was isolated from liver tissues harvestedas indicated above by phenol-guanidium extraction methods. RNase protectionassays were performed using a RiboQuant multiprobe RNase protection assaykit (PharMingen, San Diego, Calif.) according to the manufacturer’s instructions.Steady-state mRNA levels of a panel of 11 liver cytokines were determined.These cytokines included interleukin 1 (IL-1), IL-1�, IL-2, IL-3, IL-4, IL-5,IL-6, IL-10, tumor necrosis factor alpha (TNF-), TNF-�, and gamma interferon.

Statistical analysis. Data are presented as means standard errors of themeans. Analysis of variance-Tukey’s multiple comparison was used to determinethe statistical significance of the data at a P level of �0.05.

RESULTS

Adenovirus gene transfer efficiency. The efficiency of recom-binant adenovirus transduction and gene expression in the liverwas evaluated by injecting rats with adenovirus containing�-Gal cDNA as a reporter gene. Microscopic analysis of livertissue sections revealed that at day 5 posttransduction, over96% of hepatocytes were expressing �-Gal protein (Fig. 1).This level of expression lasted through day 10 posttransduc-tion. By day 14, approximately 50% of the cells were stillexpressing the protein. These results were consistent for boththe large and small lobes of the liver.

Verification of MnSOD overexpression. Seven days afterrecombinant adenovirus transduction, liver samples were har-vested from rats (three rats per group) in the following threegroups: (i) vehicle injection (control), (ii) transduction withAdCMV, and (iii) transduction with AdMnSOD. MnSOD lev-els were examined by Western blot analysis and MnSOD ac-tivity assays. All three animals in the AdMnSOD group had alarge increase in MnSOD protein expression compared to thecontrol and AdCMV groups (Fig. 2A), with small variations inMnSOD protein levels among individual rats. However, onaverage, no significant differences were noticed between thecontrol and the AdCMV groups. Densitometry analysis dem-onstrated an approximately fourfold increase in MnSOD ex-pression in the AdMnSOD group compared with the controland AdCMV groups (data not shown).

The overexpression of MnSOD was further measured usingboth activity assay and activity gel techniques. Figure 2B sum-marizes the results of the MnSOD activity gel assays. All threeanimals transduced with AdMnSOD had large increases inMnSOD activity levels. The MnSOD bands of control andAdCMV groups were not detectable, due to the fact that ratMnSOD does not produce bands under the conditions used forthe activity gel assays. Only the exogenous human MnSODexpressed from adenovirus transduction was detectable. Therewere no differences among samples from the three treatmentgroups for the CuZnSOD activity as demonstrated by activitygel assays (Fig. 2B). The levels of catalase and GSH peroxidasewere also measured by activity gel assays. No changes in theactivities of these antioxidant enzymes were found among thesamples (data not shown).

MnSOD activity was also determined by the NBT competi-tion assay (Fig. 2C) to further examine whether AdMnSODtransduction would affect endogenous MnSOD activity. Ani-mals transduced with AdMnSOD had a fourfold increase inMnSOD activity compared to the control group (P � 0.05).This level of increase in MnSOD activity was consistent with

VOL. 76, 2002 AP-1 ACTIVATION BY ADENOVIRUS 357

on Novem

ber 13, 2013 by guesthttp://jvi.asm

.org/D

ownloaded from

the increase in MnSOD protein levels as measured by Westernblotting. The AdCMV group had a MnSOD activity level sim-ilar to that of the control group.

Immunohistochemical staining for MnSOD protein by spe-cific MnSOD antibody was also performed in all rats to exam-ine the zonal distribution of MnSOD in hepatocytes of theliver. Figure 3 shows a representative slide of MnSOD im-munostaining from each group. The immunostaining ofMnSOD was predominantly intensified in hepatocytes of theAdMnSOD group (Fig. 3C). Sections from control andAdCMV animals had only light staining of MnSOD, represent-ing endogenous levels of rat MnSOD. These immunohisto-chemistry findings were consistent with the results obtained inWestern blotting experiments.

Response of AP-1 and NF-�B DNA binding activities toadenovirus transduction and MnSOD overexpression. DNAbinding activities of AP-1 and NF-�B were determined in nu-clear extracts from adenovirus-transduced livers and controlanimals using EMSA. EMSA resulted in different DNA bind-ing patterns for AP-1 and NF-�B (Fig. 4). The DNA bindingactivity of AP-1 was greatly induced in the AdCMV groupcompared with the basal level in the control group. However,this elevated AP-1 DNA binding activity was strongly sup-pressed in livers from rats overexpressing MnSOD (Fig. 4A).In contrast to the AP-1 response, there were large variationsamong all animals in NF-�B DNA binding activity (Fig. 4B),which may reflect the endogenous difference of individual an-imals. Adenovirus transduction did not change the DNA bind-

FIG. 1. Intravenous injection of adenovirus was highly efficient at delivering transgenes to rat hepatocytes. Rats were transduced withadenovirus, and livers were harvested at designated time points after transduction. Tissues from large and small lobes of the liver were fixed, cutinto sections, and stained for �-Gal activity and counterstained with fast red. The number of cells expressing �-Gal (blue) and the number of cellsnot expressing �-Gal (red) were counted. The graph shows the ratio of �-Gal-positive to �-Gal-negative cells from an average of three animalsper group, calculated as percent gene expression. (A) Representative liver section at 7 days after transduction with adenovirus without �-GalcDNA; (B to E) representative liver sections at 5, 7, 10, and 14 days, respectively, after transduction with adenovirus containing �-Gal cDNA. n �3 for each group. Magnification, �50.

358 ZHANG ET AL. J. VIROL.

on Novem

ber 13, 2013 by guesthttp://jvi.asm

.org/D

ownloaded from

ing activity of NF-�B, nor did MnSOD overexpression haveany effect on NF-�B DNA binding activity (Fig. 4B).

A supershift experiment was then performed to verify thatthe bands shown in the EMSA were the AP-1 and NF-�Bcomplexes. Two nuclear extracts from control animals (rats 1and 3) were incubated with either an antibody specific to c-Junor antibodies specific to NF-�B p50 and p65 before theseextracts were subjected to EMSA. As shown in Fig. 5A, a cleardecrease in intensity in the band for normal AP-1 complex wasobserved. Simultaneously, a new band migrated higher thanthe normal AP-1 binding complex, representing the AP-1 com-plex supershifted by anti-c-Jun antibody. Similar results wereobserved in the case of antibody to p50 for the NF-�B super-shift experiment. However, p65 antibody failed to supershiftthe NF-�B complex, indicating that the observed NF-�B com-plex band consisted only of p50-p50 homodimer.

Response of c-Jun and c-Fos protein expression after ade-novirus transduction and MnSOD expression. The signalingpathways that may regulate the response of AP-1 to adenovirustransduction were evaluated by measuring c-Fos and c-Junprotein levels by Western blot analysis, since the Fos and Junfamilies are two major components of AP-1 protein. Resultsindicated that the overall expression of both c-Fos and c-Junproteins varied greatly among individual animals in all three

groups (Fig. 6). Therefore, it is unlikely that the up-regulationof c-jun and c-fos gene expression was responsible for the in-duction of AP-1 DNA binding activity by AdCMV transductionand restoration of its activity by overexpression of MnSOD.

FIG. 2. AdMnSOD, but not AdCMV, yielded high-level expressionof human MnSOD. Rats were infected with adenovirus (1010 infectiousunits), and livers were harvested at day 7 after the infection. (A)Western blotting analysis demonstrated an increase in MnSOD pro-tein in all three AdMnSOD-transduced rats. (B) Activity gel assayshowed intense exogenous MnSOD activity bands in AdMnSOD-transduced rats but not control or AdCMV-transduced rats. Each lanewas loaded with samples from individual rats. (C) NBT activity assaydemonstrated a significant increase of MnSOD activity in AdMnSOD-transduced rats but no differences in control and AdCMV-transducedrats. n � 3 for each group. �, P � 0.05

FIG. 3. Photomicrographs of immunohistochemical prepara-tions demonstrating increased staining for MnSOD in the livers ofAdMnSOD-transduced rats. Livers were harvested at day 7 after trans-duction. (A) A representative liver section from control rats stainedwith an antibody specific for MnSOD shows a very light brown color,indicating a low level of constitutive MnSOD protein. (B) A represen-tative liver section from an AdCMV-transduced rat shows low-levelMnSOD staining that is close to the intensity of controls. (C) A rep-resentative liver section from an AdMnSOD-transduced rat shows adramatic increase in MnSOD staining, which is predominantly locatedin hepatocytes. Liver sections from AdMnSOD-transduced rats werealso stained with normal rabbit serum and showed no brown staining(not shown). Three animals from each group were examined, and themicrographs shown are representative of the whole group.

VOL. 76, 2002 AP-1 ACTIVATION BY ADENOVIRUS 359

on Novem

ber 13, 2013 by guesthttp://jvi.asm

.org/D

ownloaded from

Redox regulation of AP-1. The ratio of GSH to GSSG isconsidered an important parameter for evaluation of intra-cellular redox status (32). Therefore, GSH/GSSG ratios weremeasured for animals in the three groups (Fig. 7). AdCMVtransduction caused a fivefold decrease in GSH/GSSG ratio,indicative of a significantly increased intracellular oxidizingenvironment, whereas MnSOD overexpression significantly re-versed the GSH/GSSH ratio from the level of adenovirustransduction to control levels. These changes in the GSH/GSSG ratio were related to the changes in AP-1 DNA bindingactivity, suggesting that a cellular redox component may haveparticipated in activation of AP-1 DNA binding in the liverof animals injected with adenovirus but not overexpressingMnSOD.

Ref-1 is involved in the regulation of AP-1 DNA bindingactivity. It has been reported that Ref-1 is associated with the

redox regulation of AP-1 activation (17, 38). We examinedwhether Ref-1 is involved in the pathway of redox regulation ofAP-1 DNA binding activity by adenovirus transduction andMnSOD overexpression. Ref-1 protein was immunodepletedby an antibody specific to Ref-1 from nuclear extracts ofAdCMV-injected animals, and then these nuclear extractswere subjected to EMSA. AP-1 DNA binding ability was de-creased after Ref-1 depletion (Fig. 8A, rat 5, lane 6) comparedto the assay without Ref-1 depletion (lane 4). When Ref-1 wasadded back to the EMSA, the AP-1 DNA binding ability wasrestored (Fig. 8, lane 7). When the immunoprecipitated Ref-1was added to nuclear extracts from MnSOD-overexpressinganimals (rats 8 and 7), increases in AP-1 DNA binding activ-ities (Fig. 8, lanes 9 and 11) from low binding levels (lanes 8and 10) were observed. Addition of immunoprecipitated Ref-1

FIG. 5. AP-1 and NF-�B DNA binding activities were supershiftedby specific c-Jun or p50 antibodies reacting with nuclear extracts fromtwo control animals. Each nuclear extract (5 �g) from the experimentrepresented in Fig. 4 was incubated with 5 �l of antibody (ab) specificfor AP-1 (c-Jun) (A) or NF-�B (p50 and p65) (B) 2 h before EMSAanalysis. NS, nonspecific DNA binding; FP, free probe. The experi-ments were repeated three times, and representative pictures are shown.

FIG. 4. DNA binding activity of AP-1, but not NF-�B, was acti-vated by adenovirus transduction and suppressed by MnSOD. Ratswere transduced with either AdCMV or AdMnSOD cDNA and com-pared to vehicle-treated controls. Livers were harvested at day 7 aftertransduction. Each nuclear extract (5 �g) was analyzed by gel mobilityshift assay using AP-1-specific (A) or NF-�B-specific (B) 32P-labeledoligonucleotides. NS, nonspecific binding; FP, free probes. Three an-imals were examined for each group, and each lane was loaded withsamples from individual rats. The experiments were repeated threetimes, and representative pictures are shown.

360 ZHANG ET AL. J. VIROL.

on Novem

ber 13, 2013 by guesthttp://jvi.asm

.org/D

ownloaded from

to a nuclear extract from a control animal (rat 3) also inducedAP-1 activation (Fig. 8, lane 3) from its basal level (lane 2).Immunoprecipitated Ref-1 alone did not bind to the oligonu-cleotides with the consensus AP-1 binding sequence (Fig. 8,lane 5), and its binding was similar to that of the free probe(lane 1). These results were repeated using another nuclearextract in the AdCMV group (rat 6) (data not shown). In orderto eliminate the possibility that c-Jun and c-Fos may be copre-cipitated with Ref-1 protein, the presence of c-Jus and c-Fos inimmunoprecipitated pellets was checked by Western blottingusing c-Jun- and c-Fos-specific antibodies. No detectable c-Junor c-Fos protein was noted (data not shown). To confirm thatRef-1 was indeed immunoprecipitated, the presence of Ref-1protein in nuclear extract, pellet, and supernatant (rat 5) wasexamined by Western blotting. As shown in Fig. 8B, an intenseRef-1 protein band was present in the nuclear extract but dis-appeared in the supernatant after immunoprecipitation. ThisRef-1 band was also in the pellet of the immunoprecipitate.These data suggest that Ref-1 is involved in the pathway of

redox regulation of AP-1 binding activation by adenovirustransduction and MnSOD overexpression.

An inflammatory response was not involved in the up-reg-ulation of AP-1 activation. To examine whether inflammationplays a role in the up-regulation of AP-1 activation, RNaseprotection assays were used to determine steady-state mRNAlevels of a panel of 11 liver cytokines. These cytokines includedIL-1, IL-1�, IL-2, IL-3, IL-4, IL-5, IL-6, IL-10, TNF-,TNF-�, and gamma interferon. All the cytokines were belowdetection limits except IL-1 and -�, which had high steady-state levels of mRNA. However, no differences were observedamong AdCMV-transfected, AdMnSOD-transfected, and con-trol animals (data not shown).

FIG. 6. Response of c-Fos and c-Jun protein expression after ade-novirus transduction and MnSOD expression. Livers were harvested atday 7 after adenovirus transduction. Thirty micrograms of total liverprotein was separated by sodium dodecyl sulfate–12.5% polyacryl-amide gel electrophoresis, transferred onto nitrocellulose membranes,and processed with antibodies specific for c-Jun and c-Fos. Threeanimals were examined for each group, and each lane was loaded withsamples from individual rats. The experiments were repeated threetimes, and representative pictures are shown.

FIG. 7. The increase in oxidation status caused by adenovirustransduction is relieved by MnSOD overexpression. Rats were infectedwith adenovirus, and livers were harvested at day 7 after the infection.The GSH content of cells was measured based on the colorimetricreaction of DTNB with GSH to form TNB. The GSSG level wasdetermined by masking GSH with 2-vinylpyridine. Three animals wereexamined for each group, and each bar shows the averaged ratio forthree individual animals. �, P � 0.05 versus the control; ��, P � 0.05versus AdCMV.

FIG. 8. Ref-1 is involved in the induction of AP-1 DNA bindingactivity by adenovirus transduction. Ref-1 was immunoprecipitatedfrom one of the nuclear extracts in the AdCMV group (rat 5) by theaddition of a specific anti-Ref-1 antibody and protein A agarose beads.After overnight shaking at 4°C, the supernatant was separated fromthe agarose beads by centrifugation at 12,000 � g in a microcentrifugefor 5 min. (A) The supernatant (equivalent to 5 �g of nuclear extract)was then used as immunodepleted Ref-1 nuclear extract for EMSA.The pellet was further washed eight times with TTBS to remove thefree residual chemicals from the immunoprecipitation reaction mix-ture. The pellet was then resuspended in 350 �l of water, and 10 �l ofthis solution (equivalent to 5 �g of nuclear extract) was added backeither to the Ref-1-depleted supernatant or to nuclear extracts from anAdMnSOD-transfected and control animals prior to EMSA. AP-1DNA binding activity was measured by EMSA for the nuclear extractsbefore Ref-1 depletion (lane 4), the pellet after Ref-1 precipitation(lane 5), the supernatant after Ref-1 depletion (lane 6), and the su-pernatant to which precipitated Ref-1 protein was added back (lane 7).EMSA was also performed on nuclear extracts from two MnSODoverexpressing animals (lane 8 and 10) and one control animal (lane 2)and after addition of immunoprecipitated Ref-1 protein to these sam-ples (lanes 3, 9, and 11). Minus signs indicate Ref-1-depleted nuclearextract; plus signs indicate that Ref-1 was added to the nuclear extract.NS, nonspecific binding; FP, free probes. (B) Twenty micrograms ofnuclear extract protein (NE) from rat 5 and equal amounts of agarosebeads containing immunoprecipitated Ref-1 protein (pellet) and Ref-1-depleted supernatant (supernatant) were subjected to Western im-munoblotting using an antibody specific to Ref-1.

VOL. 76, 2002 AP-1 ACTIVATION BY ADENOVIRUS 361

on Novem

ber 13, 2013 by guesthttp://jvi.asm

.org/D

ownloaded from

DISCUSSION

Virus-induced stress responses are an important issue inmany types of viral infections. For example, there is evidencethat cellular heat shock proteins are induced during adenovirustransductions (18, 29). In the present study, we demonstratedthat replication-defective recombinant adenovirus induced asignificant increase in AP-1 DNA binding activity. This resultclearly indicates that a stress response can be induced in vivo,even though the adenovirus used is replication defective.

The mechanisms of virus-induced activation of transcriptionfactor responses are still under active investigation. Our resultsindicate that AP-1 DNA binding activity was up-regulated, butnot NF-�B DNA binding activity, suggesting that the regula-tion of a specific signal transduction pathway is stimulatedafter recombinant adenovirus transduction, rather than just ageneralized stress response.

It is known that AP-1 activation is controlled at severaldifferent levels (11). Our data indicate that adenovirus trans-fection did not induce the expression of c-fos and c-jun genes.Therefore, redox regulation of AP-1 activation was hypothe-sized. The fact that MnSOD, a primary antioxidant enzyme,suppressed induction of AP-1 activation by adenovirus clearlydemonstrated ROS involvement with both recombinant ade-novirus transduction and the nature of AP-1 redox regulation.

Importantly, GSH/GSH ratio measurements showed thatadenovirus transfection increased the oxidation of GSH anddecreased the GSH/GSSG ratios in the livers of rats, whileMnSOD overexpression restored the GSH/GSSG ratio to nor-mal levels. GSH is believed to be the most abundant redoxbalance buffer in vivo (32). Therefore, the shift from reducedGSH to GSSG indicates the presence of oxidative stress and anoxidative environment in vivo. The change of GSH/GSSG ra-tios by adenovirus transduction and MnSOD overexpressionstrongly suggests the presence of a redox regulation mecha-nism in the activation of AP-1 DNA binding by recombinantadenovirus transfection.

It is known that redox regulation of AP-1 activation involvesconserved cysteine residues (Fos Cys-154 and Jun Cys-272) inits DNA binding domain (1), and investigators have suggestedthe involvement of Ref-1 protein in the stimulation of AP-1DNA binding through a redox regulation pathway (17, 38).

Ref-1 is a bifunctional protein, serving as a DNA repairenzyme and as a regulator protein for several transcriptionfactors, including AP-1, in a redox-dependent manner. Geneticanalysis has identified a cysteine residue at position 65 in theredox regulatory domain of Ref-1 protein that is critical for theredox activation of AP-1 DNA binding (38). This residue isrequired for the direct interaction between Ref-1 and c-Junthrough Cys-272 on c-Jun protein in vitro (36). We observed adecrease in AP-1 DNA binding activity after immunodepletionof Ref-1. Also, this reduction was restored by adding back theimmunoprecipitated Ref-1 protein. However, immunopre-cipitated Ref-1 alone did react with the oligonucleotides con-taining the consensus AP-1 binding sequence. These resultsdemonstrate that Ref-1 must be physically present with AP-1protein in the stimulation of AP-1 activation after adenovirustransfection. In addition, AP-1 activation was also elevatedafter the nuclear extracts from the control and MnSOD-over-expressing animals had been supplemented with the immuno-

precipitated Ref-1 protein from adenovirus-transfected ani-mals. These data indicate that adenovirus transfection alteredRef-1 protein so that it could react with AP-1 protein andelevate AP-1 DNA binding ability in the samples that had lowAP-1 binding. Overall, our results strongly demonstrate thatredox regulation is one important mechanism for the adeno-virus induction of AP-1 activation in vivo, with the possibleinvolvement of Ref-1 protein.

In contrast to AP-1 activation, we did not observed induc-tion of NF-�B by adenovirus transfection at 7 days postinjec-tion. This observation conflicts with reports of NF-�B induc-tion by recombinant adenovirus from other groups (8, 24, 26).One major reason for the observed difference in transcriptionfactor induction may be that in all cases where NF-�B was up-regulated by adenovirus transduction, it was also accompaniedby induction of cytokines such as TNF- and IL-1. Thesecytokines induce NF-�B activation by phosphorylation of IKKs(27). When IKKs are phosphorylated, they can lead to thedegradation of the NF-�B inhibitor, I�B. This in turn causesthe components of NF-�B, p50 and p65, to be dissociated fromI�B and translocate into the nucleus. The translocations of p50and p65 are key steps during NF-�B activation. Therefore, thefact that we did not observe an induction of cytokines at the7-day point after adenovirus transduction could well be whyNF-�B was not activated at the same time point. Secondly, wespeculate that the duration of NF-�B activation may be tran-sient. A biphasic induction of NF-�B activation is usually ob-served after adenovirus infection (25). The first early-phaseinduction is often only minutes after the infection, probablyrelated to interaction of the viral particles and cell receptors. Asecond phase is usually seen approximately 3 days postinfec-tion and is thought to be due to foreign gene expression. In thepresent study, we measured NF-�B activity only at 7 days afterinfection. If the NF-�B gene is a rapid early-responding gene,it is very possible that its activation had returned to controllevels in our experiments.

NF-�B is also believed to be directly regulated by oxidativestress. However, the precise steps by which ROS mediatesactivation of NF-�B are not well defined. It is thought that thecontrol of I�B phosphorylation is the key step in NF-�B redoxactivation, which indicates that the direct reaction of ROS witheither IKK or a protein upstream of IKK is essential for redoxregulation. The specific ROS that can induce NF-�B activationneed to be elucidated as well. While research has shown thatH2O2 can activate the NF-�B pathway (28), one study indi-cated that this activation is cell dependent (3). Shi et al. (33)recently demonstrated that hydroxyl radical, rather than H2O2,is the primary signal for the activation of NF-�B in macro-phages and Jurkat cells. These observations provide addi-tional support for our results, which suggest that there aredifferences in the mechanisms between NF-�B and AP-1activation in vivo.

Unlike some other reports suggesting that an inflammatoryresponse is the direct cause of the stress response after recom-binant adenovirus infection (4, 7), our results from the RNaseprotection assay, which has a high level of sensitivity, did notshow increased cytokine expression at 7 days after administra-tion of adenovirus. It is very possible that cytokines were ele-vated at early time points and had returned to normal 7 dayspostinjection. Therefore, a kinetic analysis of cytokine regula-

362 ZHANG ET AL. J. VIROL.

on Novem

ber 13, 2013 by guesthttp://jvi.asm

.org/D

ownloaded from

tion is warranted in future studies. Besides cytokine regulationof transcription factors, Lieber et al. (24) have demonstratedthat the induction of NF-�B activation immediately after re-combinant adenovirus transduction is independent of Kupffercell presence in the liver. Our study has also identified a sig-naling mechanism other than inflammatory cytokine stimula-tion, which is often associated with the recombinant adenovi-rus transduction, that can lead to liver toxicity and total bodystress responses. Our results suggest that oxidative stress maybe partly responsible for the toxicity of adenovirus transduc-tion. Therefore, maintenance of redox balance during and af-ter adenovirus transduction may be beneficial for adenovirus-mediated gene therapy. The present study contributes to anunderstanding of the signaling regulation mechanisms relatedto recombinant adenovirus transduction.

ACKNOWLEDGMENTS

We thank associates of the Vector Core Facility of The University ofIowa Center for Gene Therapy of Cystic Fibrosis and Other GeneticDiseases, supported by NIH/NIDDK P30 DK 54759 and the CysticFibrosis Foundation, for providing adenoviruses.

This work was supported by Public Health Service grants AG 12350and AG 14687 from the National Institute on Aging.

REFERENCES

1. Abate, C., L. Patel, F. J. Rauscher III, and T. Curran. 1990. Redoxregulation of Fos and Jun DNA-binding activity in vitro. Science 249:1157–1161.

2. Anderson, M. E. 1985. Tissue glutathione, p. 317–323. In R. A. Greenwald(ed.), CRC handbook of methods for oxygen radical research. CRC Press,Boca Raton, Fla.

3. Baeuerle, P. A., R. A. Rupec, and H. L. Pahl. 1996. Reactive oxygen inter-mediates as second messengers of a general pathogen response. Pathol. Biol.44:29–35.

4. Barr, D., J. Tubb, D. Ferguson, A. Scaria, A. Lieber, C. Wilson, J. Perkins,and M. A. Kay. 1995. Strain related variations in adenovirally mediatedtransgene expression from mouse hepatocytes in vivo: comparisons betweenimmunocompetent and immunodeficient inbred strains. Gene Ther. 2:151–155.

5. Beauchamp, C., and I. Fridovich. 1971. Superoxide dismutase: improvedassays and an assay applicable to acrylamide gels. Anal. Biochem. 44:276–287.

6. Bosch, A., P. McCray, S. Chang, T. Ulich, W. Simonet, D. Jolly, and B.Davidson. 1996. Proliferation induced by keratinocyte growth factor en-hances in vivo retroviral-mediated gene transfer to mouse hepatocytes.J. Clin. Investig. 98:2683–2687.

7. Boucher, R. C. 1996. Current status of CF gene therapy. Trends Genet. 12:81–84.

8. Clesham, G. J., P. J. Adam, D. Proudfoot, P. D. Flynn, S. Efstathiou, andP. L. Weissberg. 1998. High adenoviral loads stimulate NF-�B-dependentgene expression in human vascular smooth muscle cells. Gene Ther. 5:174–180.

9. Dalton, T. P., H. G. Shertzer, and A. Puga. 1999. Regulation of gene expres-sion by reactive oxygen. Annu. Rev. Pharmacol. Toxicol. 39:67–101.

10. Diamond, D. A., A. Parsian, C. R. Hunt, S. Lofgren, D. R. Spitz, P. C.Goswami, and D. Gius. 1999. Redox factor-1 (Ref-1) mediates the activationof AP-1 in HeLa and NIH 3T3 cells in response to heat shock. J. Biol. Chem.274:16959–16964.

11. Engelhardt, J. F. 1999. Redox-mediated gene therapies for environmentalinjury: approaches and concepts. Antioxid. Redox Signal. 1:5–27.

12. Gorman, M. A., S. Morera, D. G. Rothwell, E. de La Fortelle, D. Mol, J. A.Tainer, I. D. Hickson, and P. S. Freemont. 1997. The crystal structure of thehuman DNA repair endonuclease HAP1 suggests the recognition of extra-helical deoxyribose at DNA abasic sites. EMBO J. 16:6548–6558.

13. Griffith, O. W. 1980. Determination of glutathione and glutathione disulfideusing glutathione reductase and 2-vinylpyridine. Anal. Biochem. 106:207–212.

14. Hall, D. M., T. D. Oberley, P. L. Moseley, G. R. Buettner, L. W. Oberley, R.Weindruch, and K. C. Kregel. 2000. Caloric restriction improves thermo-tolerance and reduces hyperthermia-induced cellular damage in old rats.FASEB J. 14:78–86.

15. Hall, D. M., L. Xu, V. J. Drake, L. W. Oberley, T. D. Oberley, P. L. Moseley,and K. C. Kregel. 2000. Aging reduces adaptive capacity and stress proteinexpression in the liver after heat stress. J. Appl. Physiol. 89:749–759.

16. Hattori, M., A. Tugores, A. Veloz, M. Karin, and D. A. Brenner. 1990. Asimplified method for the preparation of transcriptionally active liver nuclearextracts. DNA Cell Biol. 9:777–781.

17. Hirota, K., M. Matsui, S. Iwata, A. Nishiyama, K. Mori, and J. Yodoi. 1997.AP-1 transcriptional activity is regulated by a direct association betweenthioredoxin and Ref-1. Proc. Natl. Acad. Sci. USA 94:3633–3638.

18. Kao, H. T., and J. R. Nevis. 1983. Transcription activation and subsequentcontrol of the human heat shock gene during adenovirus infection. Mol. Cell.Biol. 3:2058–2065.

19. Karin, M., Z. Liu, and E. Zandi. 1997. AP-1 function and regulation. Curr.Opin. Cell Biol. 9:240–246.

20. Kay, M. A., Q. Li, M. Finegold, L. D. Stratford-Perricaudet, and S. L. C.Woo. 1993. Recombinant adenoviral vectors for hepatic gene therapy J. Cell.Biochem. 17:E207.

21. Knowles, M. R. 1995. A controlled study of adenoviral vector-mediated genetransfer in the nasal epithelium of patients with cystic fibrosis. N. Engl.J. Med. 333:823–831.

22. Kozarsky, K. F., and J. M. Wilson. 1993. Gene therapy: adenovirus vectors.Curr. Opin. Genet. Dev. 3:499–503.

23. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly ofthe head of bacteriophage T4. Nature (London) 227:680–685.

24. Lieber, A., C. Y. He, L. Meuse, D. Schowalter, I. Kirillova, B. Winther, andM. A. Kay. 1997. The role of Kupffer cell activation and viral gene expressionin early liver toxicity after infusion of recombinant adenovirus vectors. J. Vi-rol. 71:8798–8807.

25. Lieber, A., M.-J. Vrancken-Peeters, L. Meuse, N. Fausto, J. Perkins, andM. A. Kay. 1995. Adenovirus mediated urokinase gene transfer induces liverregeneration and allows for efficient retrovirus transduction of hepatocytesin vivo. Proc. Natl. Acad. Sci. USA 92:6210–6214.

26. Loeser, P., G. S. Jennings, M. Strauss, and V. Sandig. 1998. Reactivation ofthe previously silenced cytomegalovirus promoter in the mouse liver: in-volvement of NF-�B. J. Virol. 72:180–190.

27. Mercurio, F., H. Zhu, B. W. Murry, A. Shevchenko, B. L. Bennett, J. Li, D. B.Young, M. Barbosa, M. Mann, A. Manning, and A. Rao. 1997. IKK-1 andIKK-2: cytokine-activated I�B kinases essential for NF-�B activation. Sci-ence 278:860–866.

28. Meyer, M., R. Scheck, and P. A. Baeuerle. 1993. H2O2 and antioxidants haveopposite effects on activation of NF-�B and AP-1 in intact cells: AP-1 assecondary antioxidant-responsive factor. EMBO J. 12:2005–2015.

29. Nevins, J. R. 1982. Induction of the synthesis of a 70,000 dalton mam-malian heat shock protein by the adenovirus E1A gene product. Cell 29:913–919.

30. Newman, K. D. 1995. Adenovirus-mediated gene transfer into normal rabbitarteries results in prolonged vascular cell activation, inflammation and neo-intimal hyperplasia. J. Clin. Investig. 96:2955–2965.

31. Piette, J., B. Piret, G. Bonizzi, S. Schoonbroodt, M. P. Merville, S. Legrand-Poels, and B. Bours. 1997. Multiple redox regulation in NF-�B transcriptionfactor activation. Biol. Chem. 378:1237–1245.

32. Powis, G., M. Briehl, and J. Oblong. 1995. Redox signalling and the controlof cell growth and death. Pharmacol. Ther. 68:149–173.

33. Shi, X., Z. Dong, C. Huang, W. Ma, K. Liu, J. Ye, F. Chen, S. S. Leonard, M.Ding, V. Castranova, and V. Vallyathan. 1999. The role of hydroxyl radicalas a messenger in the activation of nuclear transcription factor NF-�B. Mol.Cell. Biochem. 194:63–70.

34. Simon, R., J. Engelhardt, Y. Yang, M. Zepeda, S. Weber-Pendleton, M. Gross-man, and J. Wilson. 1993. Adenovirus-mediated transfer of the CFTR geneto lung of nonhuman primates: toxicity study. Hum. Gene Ther. 4:771–780.

35. Spitz, D., and L. Oberley. 1989. An assay for superoxide dismutase activity inmammalian tissue homogenates. Anal. Biochem. 179:8–18.

36. Xanthoudakis, S., G. Miao, F. Weng, Y. E. Pan, and T. Curran. 1992. Redoxactivation of Fos-Jun DNA binding activity is mediated by a DNA repairenzyme. EMBO J. 11:3323–3335.

37. Yan, T., X. Jiang, H. J. Zhang, S. Li, and L. W. Oberley. 1998. Use ofcommercial antibodies for detection of the primary antioxidant enzymes.Free Radic. Biol. Med. 25:688–693.

38. Yao, K., S. Xanthoudakis, T. Curran, and P. J. O’Dwyer. 1994. Activation ofAP-1 and of a nuclear redox factor, Ref-1, in the response of HT29 coloncancer cells to hypoxia. Mol. Cell. Biol. 14:5997–6003.

VOL. 76, 2002 AP-1 ACTIVATION BY ADENOVIRUS 363

on Novem

ber 13, 2013 by guesthttp://jvi.asm

.org/D

ownloaded from