Hydrogen Peroxide Prolongs Nuclear Localization of NF-κB in Activated Cells by Suppressing Negative Regulatory Mechanisms*

NF-κB transcription factors induce pro-inflammatory molecules (e.g. IL-8) in response to cytokines (e.g. TNFα, IL-1β) or other stimuli. In the basal state, they are sequestered in the cytoplasm by inhibitory IκB proteins. Pro-inflammatory signaling triggers polyubiquitination of intermediaries (e.g. RIP1), which activate IκB kinases that trigger Ser phosphorylation and degradation of IκBα, thereby promoting nuclear translocation of NF-κB. A negative feedback loop exists whereby NF-κB drives resynthesis of IκBα, which promotes export of NF-κB from the nucleus to the cytoplasm. This process relies on Cezanne, a deubiquitinating cysteine protease that stabilizes resynthesized IκBα by removing polyubiquitin from modified intermediaries. H2O2 is generated during inflammation. Here we examined the effects of H2O2 on NF-κB dynamics and pro-inflammatory activation in cultured cells co-stimulated with TNFα or IL-1β. Quantitative reverse transcription-PCR and enzyme-linked immunosorbent assay revealed that H2O2 enhanced the induction of IL-8 by TNFα or IL-1β. We demonstrated by using assays of NF-κB nuclear localization and by imaging of live cells expressing a fluorescent form of NF-κB that H2O2 prolonged NF-κB nuclear localization in cells co-stimulated with TNFα or IL-1β by suppressing its export from the nucleus. We provide evidence that H2O2 suppresses NF-κB export by prolonging polyubiquitination of signaling intermediaries, which promotes Ser phosphorylation and destabilization of newly synthesized IκBα proteins. Finally, we observed that the catalytic activity of Cezanne and its ability to suppress RIP1 polyubiquitination and NF-κB transcriptional activity were inhibited by H2O2. We conclude that H2O2 prolongs NF-κB activation in co-stimulated cells by suppressing the negative regulatory functions of Cezanne and IκBα.

Pro-inflammatory cytokines (e.g. TNF␣, 2 IL-1␤) drive inflammation by activating the redox-sensitive transcription factor NF-B, which induces pro-inflammatory molecules including chemokines (e.g. IL-8). In unstimulated cells, NF-B is sequestered in the cytoplasm through binding to inhibitory IB molecules, which mask its nuclear localization sequence (1,2). TNF␣ or IL-1␤ activate distinct pathways that converge to activate IB kinases (IKK), which phosphorylate IB at Ser residues (2)(3)(4)(5)(6)(7). This process is regulated by ubiquitin, a protein that can be covalently attached to Lys residues of other cellular proteins in the form of isopeptide-linked polyubiquitin chains (8). Signaling through TNFR or IL-1/Toll-like receptor relies on modification of RIP1 or TRAF6 signaling intermediaries, respectively, with a "non-classical" form of polyubiquitin linked through Lys-63 that triggers the activation of IKK␤ (9 -13). In addition, polyubiquitin chains linked through Lys-48 play a well recognized role during NF-B activation by targeting phosphorylated IB proteins for proteasomal degradation, thus liberating NF-B for nuclear entry (2, 14 -17).
The resolution of inflammatory responses is regulated by multiple negative feedback mechanisms that act in concert to suppress NF-B activity. A key step in this process is the induction of IB␣ by NF-B in activated cells. Newly synthesized IB␣ proteins translocate to the nucleus, where they target activated NF-B complexes for export to the cytoplasm (18 -21). NF-B induces other negative regulators of pro-inflammatory signaling including A20 (22) and its sister molecule Cezanne (23), which belong to the OTU family of deubiquitinating cysteine proteases that can cleave ubiquitin monomers from modified proteins (13, 24 -26). Both of these proteins can be recruited to activated TNFRs, where they suppress the activity of pro-inflammatory signaling intermediaries by removing polyubiquitin chains from them (13,26). Thus suppression of TNFR signaling by A20 and Cezanne stabilizes newly synthesized IB␣, which inhibits NF-B by removing it from the nucleus.
Reactive oxygen species (ROS) such as hydrogen peroxide are generated in inflamed tissues and may contribute to the pathogenesis of chronic inflammatory diseases including atherosclerosis, rheumatoid arthritis, and chronic obstructive pulmonary disease (27)(28)(29). Although numerous studies have identified a role for ROS in regulating signaling to NF-B (30), their potential effects on negative regulators of NF-B have received little attention. Nevertheless, emerging reports suggest that ROS may enhance cellular activation by suppressing the activity of anti-inflammatory enzymes such as histone deacetylases (31) and mitogen-activated protein (MAP) kinase phosphatases (MKPs) (32). Here we demonstrate that H 2 O 2 can suppress the negative regulatory functions of IB␣ and Cezanne, thus prolonging NF-B nuclear localization and pro-inflammatory transcriptional responses in activated cells.
Cell Lines and Transfection-HeLa and A549 cells were cultured using Dulbecco's modified Eagle's medium, 10% fetal-calf serum, supplemented with 100 units/ml penicillin G, 100 g/ml streptomycin. Transient transfection of cells was achieved using Lipofectamine (Invitrogen) according to the manufacturer's instructions.
Measurement of Intracellular ROS-Intracellular ROS were measured in live cells using CM-H 2 DCFDA, a probe that forms a fluorescent adduct (dichlorofluorescein) when oxidized. Cells were incubated with 5 M CM-H 2 DCFDA for 15 min and then washed with phosphate-buffered saline prior to experimentation. Fluorescence was quantified by flow cytometry.
ELISA to Detect IL-8-The levels of IL-8 proteins in cell culture supernatants were quantified using a commercial enzyme-linked immunosorbent assay (ELISA, human IL-8 Duo Set; R&D Systems, Minneapolis, MN) according to the manufacturer's instructions. Absorbance (OD) was measured at 450 nm on a microtiter plate reader. Concentrations were obtained by interpolation from standard curves as described previously (34).
Comparative Real-time PCR-Transcript levels were quantified by comparative real-time PCR using gene-specific primers for IL-8 (sense, 5Ј-TGCCAAGGAGTGCTAAAG-3Ј; antisense, 5Ј-CTCCACAACCCTCTGCAC-3Ј), IB␣ (sense, 5Ј-CTGATGTCAACAGAGTTACCTACCAG-3Ј; antisense, 5Ј-CGTGAACTCTGACTCTGTGTCATAG-3Ј), and ␤-actin (sense, 5Ј-CTGGAACGGTGAAGGTGACA-3Ј; antisense, 5Ј-AAGGGACTTCCTGTAACAATGCA-3Ј) purchased from Sigma-Aldrich. Quantitation of GNB2L1 was carried out using primers obtained from Applied Biosystems (Foster City, CA). Total RNA was extracted and reverse-transcribed as described previously (35). Real-time PCR was carried out using the iCycler system (Bio-Rad) and SYBR green master mix (Bio-Rad) according to the manufacturer's instructions. Reactions were incubated at 95°C for 3 min before thermal cycling at 95°C for 10 s and 56°C for 45 s. Reactions were performed in triplicate. Relative gene expression was calculated by comparing the number of thermal cycles that were necessary to generate threshold amounts of product (CT) as described previously (35). CT was calculated for the genes of interest and for the housekeeping gene ␤-actin. For each cDNA sample, the CT for ␤-actin was subtracted from the CT for each gene of interest to give the parameter ⌬CT, thus normalizing the initial amount of RNA used. The amount of each target was calculated as 2 Ϫ⌬⌬CT , where ⌬⌬CT is the difference between the ⌬CT of the two cDNA samples to be compared.
Assays of NF-B Intracellular Localization-Intracellular localization of endogenous RelA was assessed by immunostaining of paraformaldehyde-fixed cells using anti-RelA antibodies and Alexa Fluor 568-conjugated secondary antibodies followed by laser-scanning confocal microscopy (LSM 510 META; Zeiss, Oberkochen, Germany) following Ref. 35. Image analysis was performed using Velocity software (Improvision, Coventry, UK) to calculate the ratio of RelA present in the nucleus when compared with the cytoplasm. Alternatively, levels of endogenous RelA were measured in cytosolic or nuclear lysates prepared using the NucBuster kit (Novagen, San Diego, CA) by Western blotting using anti-RelA antibodies, horseradish peroxidase-conjugated secondary antibodies, and chemiluminescent detection.
The intracellular localization of RelA was also assessed in real time using cells cultured on glass coverslips in 60-mm dishes. They were transfected with 0.5 g of mammalian expression vector containing RelA-dsRed and allowed to recover for 48 h. The localization of RelA-dsRed in live cells was assessed using a laser-scanning confocal microscope with a specialized stage that maintained cultures at 37°C in a 5% CO 2 environment (LSM 510 META; Zeiss, Oberkochen, Germany). Images were made every 3-5 min following stimulation. Image analysis was performed using Velocity software (Improvision, Coventry, UK) to calculate the proportion of RelA-dsRed in the nucleus at multiple time points following stimulation. The duration of nuclear localization was defined as the time interval at which the proportion of nuclear RelA exceeded 20%.
Assay of NF-B Transcriptional Activity-NF-B transcriptional activity was measured using an NF-B reporter (pGL2) as described previously (26). Cells were co-transfected with pNFluc and pRL-TK (encoding Renilla luciferase to normalize transfection efficiency) using Lipofectamine and incubated for 16 h. Cells were then treated with TNF␣ (10 ng/ml) for 16 h before measurement of NF-B activity. Firefly and Renilla luciferase activity was assessed using the Dual-Luciferase reporter assay kit (Promega, Madison, WI) and luminescence counter (Topcount microplate scintillation; Packard).
Assay of Cezanne Catalytic Activity-The catalytic activity of Cezanne was assessed by measuring its capacity to bind to a ubiquitinderived probe (HAUbVME), which contained a thiol-reactive vinylmethyl ester group at the C terminus (36). It was chosen because of its ability to bind covalently to the catalytic cysteine of Cezanne (25). The probe is tagged with HA epitope to facilitate detection and is ϳ10 kDa in size. Cytosolic lysates were made from cells expressing GFP-tagged Cezanne using 50 mM Tris (pH 7.6), 0.2% Nonidet P-40, 150 mM NaCl, 0.5 mM EDTA, 0.5 mM 4-(2-aminoethyl)benzenesulfonyl fluoride. Probe was applied to cytosolic lysates, and reactions were incubated at 37°C for 1 h. Probe-Cezanne conjugates were detected by Western blotting using anti-HA epitope antibodies (1:1000; Roche Applied Science), horseradish peroxidase-conjugated secondary antibodies, and chemiluminescent detection.
Statistics-Differences between samples were analyzed using a paired Student's t test and analysis of variance (*, p Ͻ 0.05, **, p Ͻ 0.01, ***, p Ͻ 0.001, as shown in the figure legends).

H 2 O 2 Enhanced Transcriptional Activation of Pro-inflammatory
Genes in Response to TNF␣ or IL-1␤-We examined whether H 2 O 2 could regulate pro-inflammatory activation of HeLa or A549 cells, either alone or in combination with IL-1␤ or TNF␣. We observed by comparative real-time PCR that the application of  However, H 2 O 2 significantly enhanced the production of IL-8 transcripts in a concentration-dependent manner in cells that were co-treated with IL-1␤ (Fig. 1A, compare lane 2 with lanes 3-5) without affecting cell proliferation or survival (data not shown). Thus we used H 2 O 2 at 100 M in subsequent experiments. In time course experiments, we observed that co-stimulation with H 2 O 2 significantly enhanced the induction of IL-8 transcripts by IL-1␤ or TNF␣ at both early and late time points (Fig. 1B, compare lanes 4, 6, 8, and 10 with lanes 5, 7, 9, and 11; Fig. 1C, compare lanes 3 and 5 with lanes 4 and 6). Similarly, we observed by ELISA of cell culture supernatants that the application of H 2 O 2 significantly enhanced the production of IL-8 protein in cells that were co-treated with IL-1␤ (Fig. 1D, compare lanes 3 and 4 with lanes 7 and 8) or TNF␣ (Fig. 1E, compare lanes 5 and 7).
We next examined whether the effects of H 2 O 2 were mediated via elevated levels of intracellular ROS. Using a fluorescent probe, it was demonstrated that levels of intracellular ROS were elevated rapidly by the application of exogenous H 2 O 2 (either in the presence or in the absence of IL-1␤) and were sustained for at least 360 min (Fig. 1F). In contrast, treatment of cells with IL-1␤ alone did not elevate intracellular ROS (Fig. 1F). Moreover, we observed that the capacity of H 2 O 2 to enhance IL-8 induction in response to TNF␣ was suppressed in cultures that were pretreated with the anti-oxidant N-acetyl-cysteine (NAC) (10 mM, Fig. 1E, compare lanes 7 and 8), confirming that H 2 O 2 exerts its modulatory effects on pro-inflammatory signaling by altering intracellular redox status.
The potential effects of H 2 O 2 on IL-8 mRNA stability were assessed using actinomycin D to block de novo transcription in activated cells. We observed that H 2 O 2 did not alter the rate of decay of IL-8 transcript levels in cells co-treated with TNF␣ (Fig. 1G) Fig. 2A, upper panel, and Fig. 2B, compare lanes 1 and 2) or TNF␣ (Fig. 2A, compare lanes 6 and 7) but subsequently localized to the cytoplasm following 120 min of treatment ( Fig. 2A, compare lanes 2 and 3 with lanes 7 and 8;  Fig. 2B, compare lanes 2 and 3). Co-stimulation with H 2 O 2 and IL-1␤ or with H 2 O 2 and TNF␣ led to increased nuclear localization of NF-B at late (120 -240 min) time points (Fig. 2A,  compare lanes 3 and 5 and lanes 8 and 10; Fig. 2B, compare lanes 3 and 5) without significantly affecting nuclear RelA expression at early time points (30 min) (Fig. 2A, compare  lanes 2 and 4 and lanes 7 and 9; Fig. 2B, compare lanes 2 and  4). Thus we conclude that co-stimulation with H 2 O 2 significantly enhanced nuclear localization of NF-B and binding to DNA sequences in cells that were co-stimulated with IL-1␤ or TNF␣ for 120 -240 min. These findings were supported by in vitro assays, which revealed that H 2 O 2 altered the kinetics of NF-B activation in response to IL-1␤ by enhancing DNA binding following 120 or 240 min of cotreatment (data not shown).
H 2 O 2 Suppresses Export of RelA from the Nucleus in IL-1␤treated Cells-To assess directly the potential effects of H 2 O 2 on NF-B dynamics, we used laser-scanning confocal microscopy to monitor the intracellular localization of a fluorescent form of RelA (RelA-dsRed) in real time in living cells. Analysis of multiple cells revealed that RelA-dsRed translocated from the cytoplasm to the nucleus in response to IL-1␤ and that peak nuclear fluorescence occurred following an average of 48 min of treatment (Fig. 3A, compare panels 1 and 3; Fig. 3B, treatment 1; Fig. 3C, left panel, lane 1). This was followed by export of RelA-dsRed from the nucleus to the cytoplasm (Fig. 3A, compare panels 3 and 6; Fig. 3B). Thus RelA-dsRed was localized to the nucleus in IL-1␤-treated cells for an average of 146 min (Fig.  3C, right panel, lane 1). Treatment with H 2 O 2 alone did not stimulate nuclear translocation of RelA-dsRed (data not shown). However, H 2 O 2 altered NF-B dynamics in response to IL-1␤ by modestly delaying import of RelA-dsRed in a concentration-dependent manner. Thus peak nuclear fluorescence occurred at later time points in cells co-treated with IL-1␤ and H 2 O 2 (Fig. 3A, compare panels 3 and 9; Fig. 3B; Fig. 3C, left panel, compare lane 3 with lane 1). We observed that H 2 O 2 also suppressed the export of RelA-dsRed from the nucleus (Fig. 3A,  compare panels 6 and 12; Fig. 3B). Thus the average duration of RelA-dsRed nuclear localization in response to IL-1␤ was significantlyelevatedbyco-treatmentwithH 2 O 2 inaconcentrationdependent manner (Fig. 3C, right panel). We conclude that H 2 O 2 enhances and prolongs pro-inflammatory activation in response to IL-1␤ or TNF␣ by delaying the export of NF-B from the nucleus. A, cytosolic lysates were tested by Western blotting using anti-IB␣ antibodies or by using anti-tubulin antibodies to normalize protein levels (upper panels). IB␣ levels were quantified by densitometry and normalized by measuring tubulin protein levels. Mean values calculated from triplicate measurements were pooled from two experiments and are shown with standard deviations (lower panels). B, IB␣ transcript levels were quantified by comparative real-time PCR and were normalized by measuring ␤-actin transcript levels. Mean values calculated from triplicate wells were pooled from three experiments and are shown with standard deviations. C, the effect of H 2 O 2 on Ser-32/Ser-36 phosphorylation of newly synthesized IB␣ was assessed. HeLa cells were stimulated with IL-1␤ (20 ng/ml) for varying times either in the presence or in the absence of H 2 O 2 or remained untreated as a control. Cytosolic lysates were tested by Western blotting using anti-phosphoIB␣ (pIB␣) Ser-32/Ser-36 or anti-IB␣ antibodies or by using antitubulin antibodies to normalize protein levels (left panels). Levels of phosphorylated IB␣ were quantified by densitometry and normalized by measuring total IB␣ levels (right panel).

H 2 O 2 Promotes Ser-32/Ser-36 Phosphorylation and Degradation of Newly Synthesized IB␣ Proteins-We reasoned that H 2 O 2 may
prolong NF-B nuclear localization by suppressing the expression or stability of newly synthesized IB␣ molecules, which normally inhibit pro-inflammatory activation by exporting NF-B complexes from the nucleus to the cytoplasm. Western blotting of cytosolic lysates revealed that IB␣ was degraded rapidly (15 min) in response to IL-1␤ (Fig. 4A, compare lanes 1 and  2) or TNF␣ (compare lanes 6 and 7) and that this process was not suppressed by co-treatment with H 2 O 2 (Fig. 4A, compare lanes 2 and 4 and  lanes 7 and 9). After 60 -120 min, IB␣ protein levels were restored fully in cells treated with IL-1␤ or TNF␣ (Fig. 4A, compare lanes 2 and  3 with lanes 7 and 8) but only partially restored in cells that were cotreated with H 2 O 2 (Fig. 4A, compare lanes 3 and 5 with lanes 8 and  10).
Comparative real-time PCR revealed that co-treatment with H 2 O 2 enhanced the induction of IB␣ transcripts by IL-1␤ or TNF␣ (Fig. 4B). We conclude therefore that H 2 O 2 suppresses newly synthesized IB␣ at a post-transcriptional level. We next examined whether H 2 O 2 regulates phosphorylation of resynthesized IB␣ proteins at Ser-32/Ser-36, a modification that is known to target IB␣ for proteasomal degradation. We observed that serine phosphorylation of newly synthesized IB␣ was significantly greater in cells co-treated for 1 h with H 2 O 2 and IL-1␤ when compared with cells treated with IL-1␤ alone (Fig. 4C, compare lanes 5 and  7). Thus we conclude that H 2 O 2 suppresses export of RelA from the nucleus in activated cells by promoting Ser phosphorylation and destabilization of newly synthesized IB␣ proteins.

H 2 O 2 Prolongs TNFR Signaling by Inhibiting an Anti-inflammatory Deubiquitinating
Enzyme-We examined whether phosphorylation and destabilization of newly synthe-

FIGURE 5. H 2 O 2 prolongs TNFR signaling by inhibiting an anti-inflammatory deubiquitinating enzyme.
A, RIP1 ubiquitination was measured by precipitation of TNF␣⅐TNFR complexes from HeLa cells using biotinylated TNF␣ (200 ng/ml). Cells were transfected with an expression vector encoding Cezanne (pHM6-Cez) or remained untransfected. Cultures were subsequently stimulated with biotinylated TNF␣ for varying times either in the presence or in the absence of 100 M H 2 O 2 before preparation of cytosolic lysates. Alternatively, cytosolic lysates were made from untreated cells as a control and then supplemented with biotinylated TNF␣. Streptavidin-coated beads were then used to precipitate TNF␣⅐TNFR complexes, which were tested by Western blotting using anti-RIP1 antibodies (upper panel). Cytosolic lysates were tested by Western blotting using anti-RIP1 or anti-Cezanne antibodies (center panels). Images shown are representative of those from three closely similar experiments. Levels of polyubiquitinated RIP1 (polyub.-RIP) were quantified by densitometry. Mean values calculated from triplicate measurements are shown with standard deviations (lower panel). monoub.-RIP, monoubiquitinated RIP1. B, reporter gene assays were performed to assess the effects of H 2 O 2 on NF-B suppression by Cezanne. HeLa cells were co-transfected with pHM6-Cezanne (or empty pHM6), pGL2 (NF-B reporter), and pRL-TK (Renilla luciferase control). After 16 h, cells were stimulated with TNF␣ (10 ng/ml) for 16 h either in the presence or in the absence of 100 M H 2 O 2 . Cell lysates were analyzed, and the ratio of firefly/Renilla luciferase activity was calculated, which is a measure of NF-B activity normalized for transfection efficiencies. Mean values of NF-B activity calculated from triplicate wells were pooled from three experiments and are shown with standard deviations. C and D, a thiol-reactive probe (HAUbVME) was used to assess the effects of H 2 O 2 on the catalytic activity of Cezanne. C, HeLa cells expressing GFP-Cezanne were treated with TNF␣ (10 ng/ml) either in the presence or in the absence of 100 M H 2 O 2 or remained untreated as a control. Cytosolic lysates were incubated with HAUbVME probe or were incubated in the absence of probe as a control. Cezanne linked covalently to probe was detected by Western blotting using anti-HA epitope antibodies (lower panel). Levels of Cezanne in cytosolic lysates were normalized by Western blotting using anti-Cezanne antibodies (anti-Cez., lower panel). Results are representative of two independent experiments. D, cytosolic lysates made from cells expressing GFP-tagged Cezanne were incubated with HAUbVME probe in the presence of varying concentrations of H 2 O 2 . Alternatively, lysates were incubated with probe in the presence of a reducing agent (dithiothreitol (DTT)). Cezanne linked covalently to probe was detected by Western blotting using anti-HA epitope antibodies. sized IB␣ by H 2 O 2 in activated cells was associated with prolonged activity of upstream signaling intermediaries. The effects of H 2 O 2 on the recruitment of RIP1 to the TNFR and subsequent polyubiquitination were assessed following purification of TNFR complexes using biotinylated TNF␣. Western blotting revealed that TNF␣ induced transient polyubiquitination of RIP1 in TNFR complexes, which was observed following 2 or 5 min of stimulation but had declined by 20 min (Fig. 5A,  compare lanes 1, 2, 6, and 10). Co-treatment of cells with H 2 O 2 altered the kinetics of RIP1 polyubiquitination in response to TNF␣ by significantly reducing RIP1 modification following 2 or 5 min of co-treatment (compare lanes 2 and 3 with lanes 6  and 7) and significantly enhancing it at 20 min (compare lanes  10 and 11). Thus although TNFR signaling to NF-B was initially retarded, it was significantly prolonged in cells that were co-treated with H 2 O 2 .
The export of NF-B from the nucleus to the cytoplasm in chronically stimulated cells relies on the induction of Cezanne, a deubiquitinating cysteine protease that suppresses TNFR signaling to IB␣ by cleaving ubiquitin from modified RIP1 (26). Given the exquisite sensitivity of cysteine proteases to oxidative stress, we reasoned that H 2 O 2 may prolong RIP1 polyubiquitination by inhibiting the catalytic activity of Cezanne. Western blotting revealed that overexpression of Cezanne significantly reduced the buildup of polyubiquitinated RIP1 in cells treated with TNF␣ at 5 min but not 20 min (Fig. 5A, compare lanes 6 and 8 with lanes 10 and 12) but had little or no effect in cells that were co-treated with TNF␣ and H 2 O 2 at both 5 and 20 min (compare lanes 7 and 9 and lanes 11 and 13). In addition, we observed by reporter gene assay that H 2 O 2 significantly reduced the capacity of overexpressed Cezanne to suppress NF-B transcriptional activity in response to TNF␣ (Fig. 5B,  compare lanes 3 and 6). We conclude therefore that H 2 O 2 suppresses the capacity of Cezanne to inhibit TNFR signaling to NF-B in cultured cells.
We assessed the effects of H 2 O 2 on the catalytic activity of Cezanne using an assay employing a ubiquitin-derived thiolreactive probe (HAUbVME), which is known to bind specifically to the catalytic cysteine of deubiquitinating enzymes (36). Incubation of probe with lysates from cells transfected with GFP-Cezanne revealed a prominent band corresponding to ubiquitin-modified Cezanne (Fig. 5C, compare lanes 1 and 2;  Fig. 5D). We confirmed that probe binding observed in these experiments was due to Cezanne sequences by demonstrating that it was not observed using lysates from untransfected cells (data not shown). We observed that probe binding to Cezanne was not influenced by treatment of cells with TNF␣ alone (Fig.  5C, compare lanes 4 and 8 with lane 2) but was reduced by co-treatment of cells with TNF␣ and H 2 O 2 (compare lanes 6 and 10 with lane 2). In addition, probe binding to Cezanne was suppressed by the addition of H 2 O 2 to reaction mixtures in vitro in a concentration-dependent manner (Fig. 5D, compare  lanes 2-5 with lane 6). Taken together, our findings indicate that the catalytic activity of Cezanne can be inhibited by H 2 O 2 . We conclude that the ability of H 2 O 2 to prolong RIP1 polyubiquitination in TNF␣-treated cells can potentially be explained by its capacity to inhibit the catalytic activity of Cezanne.

DISCUSSION
Chronically inflamed tissues are characterized by a complex milieu of pro-inflammatory cytokines (e.g. TNF␣ and IL-1␤) and ROS (e.g. H 2 O 2 ). A potential role of ROS in the pathogenesis of chronic inflammatory diseases has been suggested from numerous in vivo studies in which inflammatory processes have been suppressed by administration or overexpression of antioxidant enzymes (37)(38)(39)(40)(41) or have been enhanced by deletion of genes encoding antioxidants (42,43).
Studies of cultured cells have revealed that ROS have profound effects on numerous physiological activities including pro-inflammatory activation, which can be altered by ROS through poorly defined mechanisms. Numerous reports have indicated that NF-B-dependent transcription is a redox-sensitive process (30). However, the effects of oxidative stress on cellular activation are complex and vary between particular types of cells, ROS, and co-activating stimuli. For example, H 2 O 2 has been shown to be an direct inducer of NF-B transcriptional activity and pro-inflammatory activation in some cell types (e.g. T-cells, skeletal muscle myotubes, and human breast MCF-7 cells) but not others (e.g. human microvessel endothelial cells and A549 epithelial cells) (44 -46). In addition, H 2 O 2 can enhance or prolong cellular activation in response to pro-inflammatory cytokines in several cell types (47,48).
The molecular mechanisms by which ROS alter signaling to NF-B remain uncertain and are likely to operate at multiple levels. Previous studies have suggested that H 2 O 2 stimulates IB phosphorylation at Tyr-42 via Syk protein-tyrosine kinase, which leads to disassociation of NF-B, which can subsequently enter the nucleus (45). Others have shown that H 2 O 2 can activate IKK by triggering phosphorylation of serine residues in the activation loop (49), a process that leads to serine phosphorylation and destabilization of IB␣. In addition, NF-B proteins are themselves sensitive to oxidation, a modification that decreases DNA binding (50).
Here we examined the effects of H 2 O 2 on cellular activation in response to TNF␣ or IL-1␤. We chose to treat cells with up to 100 M H 2 O 2 because we have previously detected this concentration in supernatants from cultures of activated primary human macrophages. 3 We observed that H 2 O 2 and pro-inflammatory cytokines (IL-1␤, TNF␣) induced IL-8 in synergy in cultured epithelial cells. We also describe a novel mechanism for the pro-inflammatory effects of H 2 O 2 , which significantly altered NF-B dynamics in co-stimulated cells by causing it to be retained in the nucleus for prolonged periods. Prior reports suggest that timely NF-B export from the nucleus is essential for suppression of pro-inflammatory processes (18 -21). The corollary is that retention of nuclear NF-B in cells exposed to ROS may lead to persistent cellular activation and chronic inflammation.
Activation of NF-B leads to the de novo production of IB␣, which participates in a negative feedback mechanism by binding nuclear NF-B and exporting it from the nucleus to the cytoplasm (18,21). Our study suggests that H 2 O 2 alters NF-B dynamics in activated cells by triggering Ser-32/Ser-36 phos-phorylation and subsequent destabilization of newly synthesized IB␣ proteins. This concept is consistent with previous observations that co-stimulation with H 2 O 2 can prolong IKK activation in response to TNF␣ (49). Thus sustained pro-inflammatory signaling in activated cells exposed to H 2 O 2 may suppress newly synthesized IB␣ proteins and prevent NF-B export.
To define the level at which H 2 O 2 prolongs pro-inflammatory signaling, we assessed its effects on activation of RIP1, which is a key component of the TNFR complex. Recent studies have revealed that TNFR signaling relies on modification of RIP1 with polyubiquitin chains that trigger the activation of IKK␤ and TAK1 (11,13). We demonstrate for the first time that the kinetics of RIP1 polyubiquitination at the TNFR in response to TNF␣ treatment was altered by co-treatment with H 2 O 2 . Analysis of precipitated TNFR complexes revealed that modification of RIP1 was initially delayed by co-treatment with H 2 O 2 . The underlying molecular mechanism is uncertain, but we speculate that it could be due to the inhibition by H 2 O 2 of ubiquitin-activating enzymes that are known to rely on redoxsensitive Cys residues for catalysis (8). However, the physiological relevance of delayed pro-inflammatory signaling in cells co-treated with H 2 O 2 remains uncertain as we did not detect a delay in subsequent transcriptional activity (Fig. 1B). Importantly, we observed that RIP1 polyubiquitination in response to TNF␣ was significantly prolonged in cells that were co-treated with H 2 O 2 . Thus it is likely that H 2 O 2 prolongs NF-B nuclear localization by prolonging TNFR signaling in co-treated cells, which promotes Ser phosphorylation and degradation of newly synthesized IB␣.
Our group has recently identified a cysteine protease called Cezanne that functions as a negative regulator of TNFR signaling by cleaving ubiquitin monomers from polyubiquitinated RIP1, thus stabilizing IB␣ for NF-B suppression (23,24,26). Thus Cezanne inhibits the induction of IL-8 by TNF␣ by shortening the duration of TNFR signaling (26). Given the sensitivity of cysteine proteases to oxidative stress, we reasoned that H 2 O 2 may prolong signaling to NF-B by inhibiting the activity of Cezanne. Consistent with this hypothesis, we observed that the catalytic activity of Cezanne and its ability to suppress RIP1 polyubiquitination and NF-B transcriptional activity were inhibited by H 2 O 2 . We conclude that the ability of H 2 O 2 to prolong TNFR signaling can potentially be explained by its inhibitory effects on Cezanne activity.
In summary, we describe a novel mechanism for the proinflammatory effects of H 2 O 2 , which prolonged nuclear localization of NF-B in activated cells by suppressing the negative regulatory functions of IB␣ and Cezanne. Our novel findings that ROS can inhibit an anti-inflammatory deubiquitinating enzyme are consistent with an emerging theme that oxidative stress can promote chronic inflammation by inhibiting the catalytic activity of anti-inflammatory enzymes (31,32).