Introduction

The restoration of the blood supply to an ischemic organ and its subsequent reoxygenation is frequently associated with tissue injury and severe inflammatory response, called reperfusion injury. Reperfusion injury is involved in many diseases such as myocardial infarction, ischemic stroke, acute kidney injury, trauma, circulatory arrest, sickle cell disease and sleep apnea (Eltzschig and Eckle, 2011). In these diseases, the reactive oxygen species (ROS) produced by the reperfused tissues or invading immune cells induce tissue damage, which in turn promotes uncontrolled inflammation and tissue injury by inducing the accumulation and activation of inflammatory and immune cells (Iadecola and Anrather, 2011).

Nuclear factor-κB (NF-κB) transcription factors play a pivotal role in the regulation of cell survival, immune cell maturation and inflammation in various cell types (Hayden and Ghosh, 2008; Vallabhapurapu and Karin, 2009). NF-κB activation during the ischemic and inflammatory stages of reperfusion injury has been implicated in both the prevention of tissue damage and the exacerbation of inflammation and immune reactions, depending on the tissue type and the timing of its activation (Van der Heiden et al., 2010; Gordon et al., 2011). In unstimulated cells, NF-κB is associated with an inhibitory IκB protein, which inhibits the nuclear localization and thus DNA binding activity of NF-κB. In response to stimuli, including proinflammatory cytokines such as TNFα and IL-1β and endogenous ligands, such as high-mobility group box 1 protein and RNA released upon tissue damage, the IκBs are phosphorylated at Ser-32 and Ser-36 by IκB kinase (IKK); this is one of the major steps of regulation in the NF-κB signaling pathway (Hayden and Ghosh, 2008; Vallabhapurapu and Karin, 2009). The phosphorylation of IκB induces its ubiquitination by the E3 ligase βTrCP (E3RSIκB) and subsequent degradation by the 26S proteasome, releasing NF-κB, which then enters the nucleus and binds DNA to induce the expression of specific target genes (Ciechanover, 1998; Karin and Ben-Neriah, 2000).

Previous studies have shown that ROS-induced oxidative stress plays various inhibitory or stimulatory roles in NF-κB signaling (Morgan and Liu, 2011). It has been suggested that oxidative stress induces NF-κB activation in that hydrogen peroxide (H2O2) can induce NF-κB activation in lymphocytes and monocytes, and various antioxidants inhibit NF-κB activation in cells stimulated with TNFα, IL-1, LPS and phorbol esters (Schreck et al., 1991, 1992). Early studies implicated the H2O2-induced phosphorylation of Tyr-42 on IκBα instead of IKK-mediated Ser-32/36 phosphorylation in NF-κB activation, whereas recent reports suggested that H2O2 also activates NF-κB by modulating IKK activity (reviewed by Gloire et al., 2006; Morgan and Liu, 2011). Although these results suggest that a shift in the cellular redox equilibrium to an oxidized state promotes NF-κB activation, it was also shown that exposure to H2O2 and other oxidizing agents suppressed TNFα-induced NF-κB activation in various cell types by inactivating upstream kinase IKK (Korn et al., 2001; Byun et al., 2002; Levrand et al., 2005; Loukili et al., 2010). This inhibitory effect was suggested to occur through the oxidation of Cys-179 within the β subunit of IKK (Reynaert et al., 2004, 2006). ROS are also known to directly modify NF-κB subunits p50 and p65, inhibiting their DNA-binding abilities (Toledano et al., 1993). Moreover, it has been reported that the sustained exposure of lens epithelial cells to H2O2 inhibited NF-κB activation by inhibiting the proteasomal peptidase and thus stabilizing IκBα (Wu et al., 2009).

In a previous study transient oxidative stress was used to test the protective effect of ischemic preconditioning (Zahler et al., 2000). Thus, we investigated the effect of transient oxidative stress on NF-κB activation in human embryonic kidney (HEK) 293 cells stimulated with TNFα. Pre-exposure to H2O2 inhibited TNFα-induced NF-κB activation by delaying the degradation of IκBα, but it did not inhibit TNFα-induced IκBα phosphorylation and IKK activation. Examination of the ubiquitin-proteasome pathway revealed that H2O2 inhibits the polyubiquitination of IκBα, whereas proteasomal peptidase activities were not changed, indicating that transient oxidative stress temporally inhibits NF-κB activation by blocking IκBα ubiquitination in TNFα-stimulated HEK293 cells.

Results

Transient H2O2 delays TNFα-induced NF-κB activation

To determine the effect of transient oxidative stress on TNFα-induced NF-κB activity, HEK293 cells were incubated with H2O2 for 20 min. Then, the medium was exchanged with medium containing TNFα, and NF-κB-dependent gene expression was measured and calibrated with β-galactosidase expressed under the β-actin promoter (Figure 1). TNFα induced a 6.5-fold increase of NF-κB reporter expression after 4 h, whereas the pre-incubation of cells with 0.1 mM H2O2 resulted in an approximately 50% decrease in reporter expression. No further decreases were observed at higher doses of H2O2 (Figure 1A). Cell viability was decreased by approximately 25% in cells pre-exposed to H2O2 after 4 h of TNFα stimulation, and it was further decreased to approximately 50% after 24 h (Figure S1). To understand the time-dependent effects of H2O2 on NF-κB activation, we measured reporter expression at various time points after the addition of TNFα to cells pretreated with 0.5 mM H2O2 (Figure 1B). The suppressive effect of H2O2 on TNFα-induced NF-κB activation appeared at up to 4 h of TNFα stimulation, and inhibition was no longer significant after 8 h of TNFα stimulation. When we measured the mRNAs of inflammatory proteins, IL-1β and intercellular adhesion molecule-1 (ICAM1), which are known NF-κB transcriptional targets (Karin and Ben-Neriah, 2000; Van der Heiden et al., 2010), we observed similar delays in expression induced by pre-exposure to H2O2 (Figure 1C). Interestingly, while the suppression of IL-1β expression by H2O2 was fully recovered after 4 h, ICAM1 expression remained reduced compared to that in control cells even after 12 h of TNFα treatment. Overall, these results suggested that the transient exposure of cells to oxidative stress temporally suppresses NF-κB activation and NF-κB-dependent gene expression.

Figure 1
figure 1

The transient inhibition of TNFα-induced NF-κB activation by H2O2. (A) HEK293 cells were transfected with an NF-κB reporter construct (IgκB-Luc) together with a control expression vector (β-actin/β-galactosidase) and incubated for 2 days. The cells were incubated with various doses of H2O2 in for 20 min in DPBS/D-glucose, washed and stimulated with TNFα (20 ng/ml) in DMEM/10% FBS for 4 h. The luciferase activity in the cell lysate was determined by luminometry and normalized to the β-gal activity (n = 3). (B) The cells were incubated with or without H2O2 (0.5 mM) for 20 min and then stimulated with TNFα for various times, and the luciferase activity was measured (n = 3). (C) The cells were treated as in (B), and cellular total RNA was isolated. cDNA was synthesized using MMLV reverse transcriptase, and quantitative real-time PCR was performed with primer and probe sets for IL-1β (upper panel) and ICAM-1 (lower panel) to calculate the mRNA level, which was then normalized to GAPDH (n = 3).

Full size image

H2O2 does not inhibit IκBα phosphorylation but delays NF-κB nuclear localization

Previous studies have shown that the IKK complex, which catalyzes the signal-induced phosphorylation of IκBα on specific serine residues, is susceptible to inactivation by ROS and reactive nitrogen species (Korn et al., 2001; Byun et al., 2002; Reynaert et al., 2004; Levrand et al., 2005; Loukili et al., 2010). To determine whether transient exposure to H2O2 inhibited NF-κB activation by blocking IKK activity, the IKK complex was isolated from HEK293 cells that had been sequentially treated with H2O2 and TNFα, and its kinase activity was measured in a reaction mixture containing [γ-32P]ATP and GST-IκBα (Figure 2). Our results showed that TNFα-induced IKK activity was not inhibited by the transient exposure of cells to up to 1 mM H2O2 (Figure 2A). The time course for the activation of IKK by TNFα stimulation in cells pre-exposed to H2O2 was also not different from that of non-treated control cells, and the IKK kinase activity at 15 min was somewhat elevated by H2O2 pre-treatment (Figure 2B). Because a previous study showed that the inhibitory effect of ROS on IKK activity depended on the concentration of the reducing agent in the enzyme reaction mixture (Korn et al., 2001), we lysed the cells, immunoprecipitated IKK complex, and determined its kinase activity in buffers containing different concentrations of dithiothreitol (DTT) (Figure 2C). Although an H2O2-induced decrease in kinase activity was observed in IKK prepared in buffers without added DTT, even 0.03 mM DTT was sufficient to recover IKK activity in cells pre-treated with H2O2 to the level of activity in untreated control cells.

Figure 2
figure 2

IKK activity was not obviously inhibited by H2O2. (A) HEK293 cells were exposed to various doses of H2O2 for 20 min, washed and stimulated with TNFα (20 ng/ml) for 10 min. IKK complexes in the cell lysate were immunoprecipitated with anti-IKKγ antibody. An in vitro kinase assay (KA) was carried out using [γ-32P]ATP and GST-IκBα(1-54) as substrates. The IKKγ in the kinase assay mixture was measured by immunoblot (IB) analysis. (B) The cells were exposed to H2O2 (0.5 mM) for 20 min and stimulated with TNFα for various times. IKK activity was determined as in (A). (C) The cells were exposed to H2O2 (0.5 mM) for 20 min and stimulated with TNFα for 10 min. The cells were lysed in cell lysis buffer containing various concentrations of DTT. Immunoprecipitation and in vitro kinase assays were performed as in (A) using buffers containing the indicated concentrations of DTT.

Full size image

To elucidate the cause of the reduced NF-κB activity in cells pre-exposed to H2O2, we then determined the levels of proteins involved in the NF-κB signaling pathway and their modifications (Figure 3). Immunoblotting analysis of cytosolic IκBα revealed that its degradation upon TNFα stimulation was blocked by pre-exposure to H2O2 (Figure 3A). The detection of IκBα phosphorylated at Ser-32 and Ser-36 using a phospho-specific antibody revealed that TNFα-induced IκBα phosphorylation was not inhibited by pre-exposure to H2O2. Consistent with the H2O2-induced inhibition of IκBα degradation, an increase in the nuclear level of NF-κB subunit p65 was inhibited by pre-exposure to H2O2. When we measured time-dependent changes in NF-κB signaling proteins after TNFα stimulation, pre-exposure of cells to H2O2 significantly delayed IκBα degradation, whereas IκBα phosphorylation was not changed in the same cells (Figure 3B). The increase of nuclear p65 was delayed, again reflecting the delayed degradation of IκBα. Similarly, electrophoretic mobility assays (EMSA) of nuclear extract prepared from cells pre-exposed to H2O2 showed a delay in the appearance of κB-sequence binding activity compared with non-exposed TNFα-stimulated control cells (Figure 3C).

Figure 3
figure 3

The effect of H2O2 on the TNFα-induced phosphorylation and degradation of IκBα, and the nuclear translocation of p65. (A) HEK293 cells were treated with various doses of H2O2 for 20 min and stimulated with TNFα (20 ng/ml) for 10 min. The levels of IκBα and phosphorylated IκBα (p-IκBα) in the cytoplasmic extract (CE) and p65 in the nuclear extract (NE) were determined by immunoblotting analysis using specific antibodies. β-Actin and nucleoporin (Nucln) were used as control proteins for the cytoplasmic and nuclear extracts, respectively. (B) The cells were exposed to H2O2 (0.5 mM) for 20 min and stimulated with TNFα for various times. The changes in the levels of proteins in the cytoplasmic and nuclear extracts were determined by immunoblotting analysis as described in (A). (C) Nuclear extracts prepared from cells treated as in (B) were used to determine NF-κB-DNA binding activity by EMSA. The upper panel shows the retarded NF-κB band, and the lower panel shows unbound free probe.

Full size image

H2O2 inhibits the poly-ubiquitination of phosphorylated IκBα

Our results suggested that transient exposure to H2O2 inhibits the TNFα-induced activation of NF-κB in HEK293 cells by blocking the degradation of phosphorylated IκBα. Because the signal-induced degradation of IκBα is known to occur through the ubiquitin-proteasome pathway, we measured proteasomal enzyme activity and IκBα ubiquitination in cells pre-exposed to H2O2 (Figure 4). The proteasomal chymotrypsin-like, trypsin-like, and peptidylglutamyl peptide hydrolase activities in the cell extract were determined using specific fluorogenic substrates. As shown in Figure 4A, the incubation of cells with 0.5 mM H2O2 did not significantly change the activities of the three proteasomal peptidases evaluated. Both chymotrypsin-like and peptidylglutamyl peptide hydrolase activities were decreased by approximately 30% after 15 min of H2O2 treatment, but these differences were not statistically significant (P > 0.05 by Student's t-test). We then monitored the ubiquitination of IκBα by immunoblotting analysis (Figures 4B and 4C). The stimulation of cells with TNFα in the absence of the proteasomal inhibitor MG-132 induced the time-dependent degradation of IκBα. However, when the cells were stimulated with TNFα in the presence of MG-132, high-molecular-weight poly-ubiquitinated IκBα appeared after 5 min and remained undegraded. When the cells were pre-exposed to H2O2 and stimulated with TNFα in the presence of MG-132, the poly-ubiquitination of IκBα was obvious only after 30 min. Immunoblotting analysis with anti-ubiquitin antibody revealed that the cellular level of total ubiquitinated proteins was largely unchanged after TNFα stimulation. In contrast, pre-exposure to H2O2 resulted in an approximately 20% decrease in total protein ubiquitination, which was not recovered in cells incubated for 1 h in the absence of H2O2.

Figure 4
figure 4

The transient inhibition of IκBα poly-ubiquitination by H2O2. (A) HEK293 cells were incubated with H2O2 (0.5 mM) for various times, and the cells were lysed in hypotonic buffer to prepare cell extract. Chymotrypsin-like, trypsin-like, and peptidylglutamyl peptide hydrolase (PGPH) peptidase activities were measured using fluorogenic peptide substrates (n = 3). (B) HEK293 cells were incubated with or without MG-132 (20 µM) for 1 h, and stimulated with TNFα (20 ng/ml) for up to 20 min. The cells were lysed, and immunoblotting analysis was performed using the anti-IκBα antibody. (C) Cells were incubated with MG-132 for 1 h, and a group of cells was exposed to 0.5 mM H2O2 for 20 min. After stimulation with TNFα, cells were lysed at various times. The cell lysate was analyzed by immunoblotting using anti-IκBα (upper panel), anti-ubiquitin (middle panel) and anti-β-actin (lower panel) antibodies.

Full size image

Discussion

In this study, the transient oxidative stress elicited by exposing cells to H2O2 was found to temporally inhibit NF-κB reporter expression in TNFα-stimulated HEK293 cells. The inhibitory effect of H2O2 appeared even at low concentrations (0.1 mM) of H2O2 (Figure 1A), suggesting that the NF-κB signaling pathway responds sensitively to changes in redox state induced by various pathologic conditions. Pre-exposure to H2O2 inhibited the expression of an NF-κB-reporter genes and IL-1β and ICAM-1 in response to TNFα, but their expression was recovered to the level of non-treated control cells after a lag period (Figures 1B and 1C). The differences in the time required to reach maximal expression between H2O2-treated and non-treated cells were approximately 4 h for the luciferase reporter and approximately 2 h for IL-1β and ICAM-1 mRNA. These results suggest that the inhibitory effect of transient oxidative stress on NF-κB activation is temporary and that the NF-κB signaling molecule that was modified by ROS recovers its original activity via cellular anti-oxidants and anti-oxidant enzymes.

Previous studies, including one from our research group, have shown that IKK in various cell types is susceptible to inactivation by ROS such as H2O2, nitric oxide (NO) and peroxynitrite (Korn et al., 2001; Byun et al., 2002; Reynaert et al., 2004; Levrand et al., 2005; Loukili et al., 2010). It was suggested that the ROS-induced inhibition of IKK is caused by the oxidative modification of Cys-179 of the IKKβ subunit of the IKK complex (Reynaert et al., 2004, 2006), which is located in the activation loop of IKKβ and plays a crucial role in the regulation of enzyme activity and signal-induced enzyme activation (Byun et al., 2006). Interestingly, when we measured the in vitro activity of IKK complexes prepared from HEK293 cells transiently exposed to H2O2, we could not detect a significant change in its activity compared with that of non-treated control cells (Figure 2A). Our results showed that the IKK activity of H2O2-pre-exposed cells at 15 min after TNFα stimulation was elevated compared to that of non-treated control cells (Figure 2B). Recently, Loukili et al. (2010) showed that oxidants induce the hyperactivation of IKK by blocking protein phosphatase 2A, thus stabilizing active phosphorylated IKK, possibly explaining our result. Because the in vitro IKK assay is performed in the presence of exogenously added reducing agents such as DTT, it is possible that the oxidized IKK in the cell is re-reduced during enzyme isolation and assays, and thus not representative of actual intracellular IKK activity. Indeed, we observed H2O2-induced decrease of enzyme activity in IKK prepared in the absence of DTT (Figure 2C), although this also does not reflect its actual intracellular activity. In our immunoblot analysis of IκBα phosphorylated on Ser-32 and Ser-36, pre-exposing cells to H2O2 did not significantly change the level and timing of IκBα phosphorylation (Figures 3A and 3B), suggesting that IKK activity is not modulated by H2O2 pre-exposure in these cells.

Whereas the phosphorylation of IκBα was not changed by pre-exposure of cells to H2O2, the TNFα-induced degradation of IκBα was delayed (Figures 3A and 3B). In the same cells, increases in nuclear NF-κB p65 and κB-binding activity were delayed by H2O2 treatment via similar modes (Figures 3B and 3C). These results indicated that the effect of H2O2 was mediated through blocking the degradation of phosphorylated IκBα, whereas the DNA-binding ability of NF-κB was not changed by H2O2. Because the degradation of phosphorylated IκBα occurs via the ubiquitin-proteasome pathway (Ciechanover, 1998; Karin and Ben-Neriah, 2000), we examined proteasomal peptidase activities and the ubiquitination of IκBα in H2O2-treated cells. Our results showed that although proteasomal peptidase activities were not changed, the TNFα-induced poly-ubiquitination of IκBα was delayed by H2O2 pre-exposure for a time period (10-30 min) similar to the delay in IκBα degradation and p65 nuclear translocation (Figures 3 and 4). These results clearly indicated that the delay of NF-κB activation that occurs in cells exposed to H2O2 is caused by the inhibition of IκBα poly-ubiquitination and suggested two different possible modes for the inhibition of IκBα poly-ubiquitination by H2O2: the transient modification of IκBα or the transient inhibition of IκBα ubiquitinating enzyme(s). In a previous study, exposure of cells to various nitrosative stresses induced S-nitrosylation of the Bcl-2 protein, leading to its decreased ubiquitination and degradation via ubiquitin-proteasome pathway, which in turn inhibited apoptosis (Azad et al., 2010). In contrast, NO-induced S-nitrosylation modulates the ubiquitin E3 ligase activity of parkin, contributing to the accumulation of neurotoxic protein aggregates within neurons and glial cells in Parkinson's disease and other neurodegenerative disorders (Chung et al., 2004; Yao et al., 2004). Although our results do not conclusively determine which mode is responsible for the H2O2-induced inhibition of IκBα ubiquitination, these results suggested that it was caused by a similar modification of IκBα or IκBα ubiquitinating enzyme(s).

The results shown in Figure 4C suggest that the inhibitory effect of H2O2 on IκBα ubiquitination is specific and not due to a general inhibition of protein ubiquitination. IκBα poly-ubiquitination was blocked by H2O2 and then slowly recovered to normal levels over a 30 min period; in the same cells, the level of total ubiquitinated proteins was not changed by H2O2 pre-exposure and slowly decreased thereafter. These results suggest that H2O2 acts via two different modes: immediate and sensitive inhibition for IκBα ubiquitination, and slow and partial inhibition for general protein ubiquitination.

Ischemic preconditioning is a strategy in which brief, transient episodes of ischemia attenuate tissue injury during subsequent ischemia and reperfusion (Eltzschig and Eckle, 2011). Previous studies have revealed that the oxygen-dependent induction of hypoxia-inducible factor-1 and the resulting adenosine receptor signaling play a critical role in ischemic preconditioning (Eckle et al., 2007, 2008). As an in vitro model to test the role of oxidative stress that occurs during preconditioning, Zahler et al. (2000) transiently exposed endothelial cells to 1 mM H2O2 for 5 min. They observed that the TNFα-induced expression of cell-adhesion molecules and inflammatory cytokines was inhibited by this pre-exposure to H2O2 and suggested that this phenomenon occurred due to the inhibition of NF-κB activity. Our results showed that transient pre-exposure to H2O2 inhibits TNFα-induced NF-κB activation by temporarily blocking the ubiquitination of phosphorylated IκBα, whereas IKK activity and IκBα phosphorylation remain largely intact in the same cells. These results suggest that modulation of IκBα ubiquitination by ROS plays a critical role in regulating the NF-κB-dependent expression of inflammatory and immune genes in reperfusion injury and other acute and chronic inflammatory diseases.

Methods

Cells and reagents

HEK293 cells were obtained from the American Type Culture Collection (Manassas, VA), and maintained in DMEM supplemented with 10% heat-inactivated FBS, and antibiotics. Antibodies against IKKγ, p65/RelA, IκBα, and ubiquitin were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Antibodies to phospho-Ser-32 and phospho-Ser-36 IκBα, β-actin, and nucleoporin were obtained from Cell Signaling Technology (Beverly, MA), Sigma (St. Louis, MO), and BD Biosciences (San Jose, CA), respectively. Recombinant glutathione S-transferase (GST)-IκBα containing the N-terminal 54 residues of IκBα and recombinant human TNFα were expressed in Escherichia coli as described previously (Ha et al., 2009). Hydrogen peroxide was purchased from Junsei Chemical (Tokyo, Japan) and diluted fresh in Dulbecco's phosphate-buffered saline (DPBS) containing 5 mM D-glucose prior to use. The fluorogenic substrates for proteasomal peptidase and MG-132 were obtained from Sigma.

IKK assay and NF-κB reporter assay

Cell extract preparation and anti-IKKγ immunoprecipitation were performed as previously described (Byun et al., 2006). Kinase activity was measured in reaction mixtures containing 10 µM ATP, [γ-32P]ATP (3 µCi) and GST-IκBα (2 µg) (Ha et al., 2009). The reaction products were analyzed by SDS-PAGE on a 10% gel and electrophoretically transferred to a PVDF membrane. Phosphorylated GST-IκBα was visualized by autoradiography and quantitated in a phosphor image analyzer (Fujifilm, Tokyo, Japan). The NF-κB reporter gene assay was performed in HEK293 cells transfected with IgκB-Luc, as described previously (Byun et al., 2002). A β-Actin promoter-driven β-galactosidase expression plasmid was used for the normalization of luciferase activity.

Immunoblotting analysis

The proteins in the cell extracts were analyzed by immunoblotting and detected using the WEST-one detection system (Intron Biotechnology, Seongnam, Korea) according to the recommended procedure. For the analysis of ubiquitinated proteins, cells grown on a 6-well plate were rinsed with DPBS and lysed in 60 µl of 1× SDS sample buffer. The lysed cells were scraped off the plate, transferred to a microtube, and boiled for 5 min. The cell lysate was cleared by centrifugation at 10,000 g for 2 min, and 30 µl of the lysate was analyzed via SDS-PAGE on an 8% gel and subsequent immunoblotting.

Electrophoretic mobility shift assay

Nuclear and cytoplasmic extracts were prepared from HEK293 cells as described previously (Byun et al., 2006). An oligonucleotide containing the consensus recognition sequence was obtained from Santa Cruz Biotechnology and end-labeled with T4 polynucleotide kinase and [γ-32P]ATP. The binding reaction was performed with 10 µg of nuclear extract and the reaction products were analyzed by electrophoresis on a 6% polyacrylamide gel in 0.5× TBE buffer (45 mM Tris-HCl, pH 8.5, 45 mM borate and 1 mM EDTA). The gel was dried under a vacuum, and radioactive bands were detected by autoradiography.

Analysis of mRNA

Total cellular RNA was isolated using an RNA isolation kit (RNA Stat-60, amsbio, Abingdon, UK) and cDNA was synthesized using M-MLV reverse transcriptase (Promega, Seoul, Korea). IL-1β, ICAM-1, and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) mRNA levels were measured by quantitative RT-PCR using a kit (qPCR Mix, QARTA Bio, Fremont, CA) and a real-time PCR machine (Applied Biosystems 7300). The PCR primers and fluorescent probes used are presented in Table S1.

Assay of proteasome activity and cell viability

The peptidase activities of three proteasome components were measured using a fluorogenic substrate, as described previously (Wu et al., 2009). Cells were lysed in 25 mM Tris-HCl buffer (pH 7.6) containing 1 mM DTT, and the lysate was cleared by centrifugation at 15,000 g for 10 min. The cell extract was incubated with succinyl-Leu-Leu-Val-Tyr-amidomethylcoumarin (LLVY-AMC), N-t-butyloxycarbonyl-Leu-Ser-Thr-Arg-amidomethylcoumarin (LSTR-AMC), or benzyloxycarbonyl-Leu-Leu-Glu-amidomethylcoumarin (LLE-AMC) to determine chymotrypsin-like, trypsin-like, and peptidylglutamyl peptide hydrolase activity, respectively. Each assay mixture contained 10 µg protein of cell extract, and enzyme activity was measured at 25℃ in a temperature-controlled microplate fluorometric reader (Victor 3, PerkinElmer, Waltham, MA) using excitation/emission wavelengths of 380/440 nm. Cell viability was measured using a kit (Cell Counting Kit-8, Dojindo Lab, Rockville, MD) according to the manufacturer's recommendations.