| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
B pathway via disulfide Bcl-3/p50 complex
,1
* Laboratoire de Biologie Moléculaire et Cellulaire du Cancer, Fondation Recherche sur le Cancer et les Maladies du Sang, Hôpital Kirchberg, Luxembourg;
Dipartimento di Biologia, Università di Roma "Tor Vergata," Rome, Italy; and
Istituto di Genetica Molecolare Consiglio Nazionale delle Ricerche, Pavia, Italy
1 Correspondence: L.G., Dipartimento di Biologia, Università di Roma "Tor Vergata," Via della Ricerca Scientifica, 00133 Rome, Italy. E-mail: ghibelli{at}uniroma2.it; M.D., Laboratoire de Biologie Moléculaire et Cellulaire du Cancer, Fondation Recherche sur le Cancer et les Maladies du Sang, Hôpital Kirchberg, 9, rue Edward Steichen, L-2540 Luxembourg. E-mail: marc.diederich{at}lbmcc.lu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
B- and p65-independent NF-B pathway via a multistep process leading to the up-regulation of Bcl-2. The slow BSO-induced GSH depletion allows separation of two redox-related phases, namely, early thiol disequilibrium and late frank oxidative stress; each phase contributes to the progressive activation of a p50-p50 homodimer. The early phase, coinciding with substantial thiol depletion, produces a cytosolic preparative complex, consisting of p50 and its interactor Bcl-3 linked by interprotein disulfide bridges. The late phase, coinciding with reactive oxygen species production, is responsible, probably via p38 activation, for nuclear targeting of the complex and trans-activation of Bcl-2.—Cristofanon, S., Morceau, F., Scovassi, A. I., Dicato, M., Ghibelli, L., Diederich, M. Oxidative, multistep activation of the noncanonical NF-B pathway via disulfide Bcl-3/p50 complex.
Key Words: glutathione • Bcl-2 • ROS • p38
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
B IS A TRANSCRIPTION FACTOR FAMILY that responds to cytokine receptor stimulation to bind a common DNA sequence motif, the B box (1)
, thus promoting transcription of genes involved in many cellular dynamics (activation, proliferation, and apoptosis) that control the inflammatory and immune response (2)
. Five different proteins (RelA or p65, RelB, cRel, p50, and p52) constitute the NF-B family; they are differently assembled to produce different active homo- or heterodimers (3
, 4)
displaying selective promoter targeting, thus accounting for the differential set of genes trans-activated by each dimer (5)
.
NF-B is present in unstimulated cells as complexes kept sequestered in the cytosol by ankyrin-containing proteins: inhibitory NF-B (IB) (for p65, RelB, and cRel), which is phosphorylated and degraded on stimulation (6)
, or the precursor forms of p50 or p52 themselves, whose C-terminal moiety is an ankyrin repeat-rich cytosolic anchor that is processed away, releasing the mature form (6)
. Thus, NF-B activation consists mainly of the phosphorylation and degradation of the sequestering proteins, occurring via a complex chain of kinases (7)
, that releases an active NF-B dimer, which can move to the nucleus.
Three major pathways, each characterized by a specific dimer, are presently described. The best studied is the so-called canonical pathway, involved mainly in primary inflammatory responses, which is carried on by the heterodimer p65-p50, released by very rapid IB protein phosphorylation and degradation (8)
. An alternative pathway, involved in tissue homeostasis, which is carried on by p52-relB, is activated by the receptor-controlled NF-B-inducing kinase (NIK), processing of p52 precursor (p100), and translocation of the heterodimers to the nucleus (9)
. Recently, a "noncanonical" pathway has been proposed, which is atypical because 1) the active dimer is p50-p50 (10)
, and 2) no receptor whose stimulation may trigger this pathway has so far been identified. p50, unlike p65, does not possess a transcription-activating domain; for this reason, the p50-p50 dimer is considered to be an inhibitor, rather than an activator, of transcription (11
, 12)
. However, evidence is emerging that this dimer may turn into a transcriptional activator on binding with Bcl-3 (13
, 14)
, the coding product of a B-cell leukemia oncogene, which belongs to the IB family. Bcl-3 is a nuclear protein (15)
able to form heterocomplexes with p50 and p52, thus acting as a downstream regulator of NF-B functions (13)
; the molecular mechanisms of the interactions with p50 and p52 are still a matter of investigation. However, in some cell systems (hepatocytes, erythroblasts, and cylindroma keratinocytes), Bcl-3 is also present in the cytosol and requires activation before nuclear translocation (10
, 16
, 17)
; whether this may provide the ability for upstream regulation of NF-B responses is presently unknown.
A controversial issue is the ability of the NF-B signal to respond to redox alterations and oxidative stress (18)
. In early studies, it was noted that NF-B was able to respond to prooxidant stimuli. However, because most such stimuli are of an inflammatory nature, it was necessary to disentangle the actual role played by oxidation or inflammation in NF-B activation. Now the evidence points to inhibition, rather than activation, of NF-B signaling by oxidative stress. Critical cysteines, which have been found to undergo oxidation, nitrosylation, and S-glutathionylation, are present in the DNA binding site of p50; site-directed mutagenesis against cysteine 62 indeed caused loss of regulation of the complex (19)
. Glutathione (GSH), the main intracellular endogenous antioxidant, was found to be necessary for NF-B functioning (20
, 21)
. Moreover, the kinases responsible for IB degradation are inhibited by oxidative stress (22)
. The latter finding may imply that the inhibition of NF-B may be limited to those NF-B pathways that depend on IB activation, namely the canonical one. Whether the noncanonical pathway may instead be stimulated by oxidation is not clear, even though some studies do report oxidation-dependent activation, which is slower than the canonical, cytokine-mediated activation, thus being compatible with the slower-reacting noncanonical pathways (18)
.
We have recently described a set of "responder" tumor cell models, wherein the redox imbalance and oxidative stress consequent to the pharmacological depletion of GSH with buthionine sulfoximine (BSO) stimulate the activation of survival pathways, culminating in the up-regulation of the antiapoptotic protein Bcl-2 (23)
. We also provided evidence of possible involvement of NF-B in promoting transcription of Bcl-2 (23)
. The promoters of Bcl-2 contain three canonical B boxes (24)
; however, Bcl-2 is not a universally recognized target of NF-B trans-activation. However, it has been reported that Bcl-2 may be transcribed by a p50-p50 homodimer (24
, 25)
, thus possibly belonging to the noncanonical NF-B target genes.
For this reason, we decided to investigate whether GSH depletion-induced transcription of Bcl-2 might be mediated by a noncanonical NF-B pathway. We describe here recruiting by GSH depletion, in a redox-dependent fashion and via a multistep process, of a disulfide Bcl-3/p50/p50 complex to the nuclei of U937 cells without involving p65, thus promoting Bcl-2 up-regulation.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. All of the experiments were performed in complete medium on log-phase cells at a cell density of 7 x 105 cells/ml. Apoptosis was quantified as the fraction of apoptotic nuclei on staining with the DNA-specific dye Hoechst 33342 (Sigma-Aldrich Corp., St. Louis, MO, USA) as described previously (27)
. GSH was depleted by inhibiting GSH neosynthesis with 1 mM BSO (Sigma-Aldrich Corp.), as described previously (26)
. The vitamin E analog Trolox C (Fluka AG, Buchs, Switzerland) was used as a radical scavenger (28)
; it was added to the cell culture at the final concentration of 500 µM (30 min before BSO when they were used together). In the experiments involving the use of compounds known to inhibit kinase activity, U937 cells were incubated for 1 h before the BSO treatment in a medium containing either 5 µM U0126 (Promega Benelux, Madison, WI, USA), 25 µM SB 203580 (Promega Benelux), 25 µM PD 98059 (Promega Benelux), or 10 µM SP600125 (Calbiochem Novabiochem Corp., San Diego, CA, USA) or 20 µM LY 294002 (Calbiochem Novabiochem Corp.). To induce the NF-B canonical pathway, 20 ng/ml tumor necrosis factor-
(TNF-) (Sigma-Aldrich Corp.) was added to the culture medium for 2 h.
Western blot analysis
Whole-cell extracts were prepared with M-PER Mammalian Protein Extraction Reagent, according to the manufacturer’s instruction (Pierce Biotechnology, Rockford, IL, USA). In brief, 107 U937 cells were washed in PBS, and the pellet was resuspended in 400 µl of the solution containing protease inhibitor, vortexed horizontally for 15 min at 4°C, and then centrifuged at 13,000 rpm for 15 min. Nuclear and cytosolic proteins were extracted from 107 cells in different experimental conditions according to Schreiber et al. (29)
. During this procedure, the cells were lysed in a hypertonic detergent medium containing protease inhibitors. The extraction was performed on ice to avoid denaturation of the proteins. For samples separated in nonreducing gels, the same procedure as above was followed, except that cells were lysed in nonreducing hypotonic buffer [without dithiothreitol (DTT)], and the loading buffer did not contain β-mercaptoethanol.
Equivalent amounts of proteins were separated through 10% SDS-PAGE and electrotransferred to nitrocellulose membranes. Immunoblotting was performed by blocking overnight aspecific binding with 5% nonfat milk in PBS-0.1% Tween. Blots were then incubated 1 h with 0.5–1 µg/ml of the following antibodies diluted in a solution of PBS-0.1% Tween and 5% milk: anti-NF-B-p50 rabbit polyclonal (NLS), anti-NF-B-p65 rabbit polyclonal (A), anti-IB rabbit polyclonal (C-21), anti-Bcl-3 rabbit polyclonal (C-14), anti-IBβ rabbit polyclonal, anti-IB
rabbit polyclonal, anti-IB
rabbit polyclonal (all obtained from Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA), anti-Bcl-2 mouse monoclonal (Calbiochem Novabiochem Corp.), anti-tubulin mouse monoclonal (Calbiochem Novabiochem Corp.), anti-active-p38 rabbit polyclonal (Promega Benelux), anti-p38 rabbit polyclonal (Cell Signaling Technology, Danvers, MA, USA), anti-actin mouse monoclonal (Sigma-Aldrich), and anti-lamin A/C mouse monoclonal (VisionbioSystems Novocastra, Newcastle, UK). The membranes were washed and incubated for 1 h with a secondary antibody, horseradish peroxidase (HRP)-conjugated goat anti-rabbit or goat anti-mouse (all obtained from Santa Cruz Biotechnology, Inc.). The specific immunoreactive protein bands were identified using the ECL system (Amersham, Little Chalfont, UK).
Indirect immunofluorescence
After treatments, U937 cells were fixed and permeabilized with a BD Cytofix/Cytoperm Kit according to the manufacturer’s instruction (BD Biosciences, San Jose, CA, USA). After 20 min the cells were washed with 1x BD Perm/Wash solution according to the manufacturer’s instruction. Primary antibody incubation was performed with 10 µg/ml of the following antibodies diluted in BD Perm/Wash solution for 1 h at room temperature: anti-tubulin mouse monoclonal, anti-p50 mouse monoclonal (E-10) (Santa Cruz Biotechnology, Inc.), and anti-Bcl-3 rabbit polyclonal (H-146) (Santa Cruz). After two washes with PBS, the cells were incubated for 30 min at room temperature with 8 µg/ml anti-rabbit secondary Alexa Fluor 488 and anti-mouse secondary Alexa Fluor 568 (Molecular Probes Inc., Junction City, OR, USA). After washing 2 times with PBS again, the cells were counterstained with Hoechst 33342 and observed with a DM IRB microscope (Leica, Wetzlar, Germany), and the images were analyzed with ImageJ software.
Flow cytometric analysis
Bcl-2 determination
Cells were fixed, permeabilized, and stained with anti-Bcl-2 monoclonal antibody (Calbiochem Novabiochem Corp.) according to the manufacturer’s instruction (23)
. Detection was performed with anti-mouse secondary Alexa Fluor 568 and processed in a FACSCalibur (BD Biosciences) flow cytometer.
Determination of ROS production
Cells were incubated with 10 µM 2,7-dichlorofluorescein diacetate (Molecular Probes Inc.), which fluoresces only when oxidized (excitation 490 nm; emission 520 nm) at 37°C in the dark for 20 min. Then cells were washed and resuspended in PBS and analyzed with the FACSCalibur flow cytometer.
Statistics were elaborated in 10,000 events/sample by CellQuest software. Mean values given by this analysis were used for further elaboration. For comparison between different experiments, the value of each treated cell sample was compared with the value of the control cell sample as fold increase (control=1).
Immunoprecipitation
For immunoprecipitation experiments, aliquots of 107 U937 cells with or without BSO treatment were lysed and prepared as described previously (30)
. The DNA-free nuclear fractions of each sample were incubated for 1 h at room temperature and then overnight at 4°C with 2 µg of anti-p50 monoclonal antibody (E-10) or anti-Bcl-3 polyclonal antibody (H-146). Anti-mouse immunoglobulin G (IgG; Santa Cruz Biotechnology, Inc.) was used as the negative control in a parallel immunoprecipitation.
On the next day, samples were incubated for 1 h at room temperature with 2 mg/ml protein A-Sepharose (Amersham Biosciences Corp., Piscataway, NJ, USA) and then, after centrifugation and washing, were analyzed by SDS-PAGE and Western blot. Immunoblot analysis was performed with anti-protein Bcl-3 polyclonal antibody (H-146) or anti-p50 monoclonal antibody (E-10).
For samples separated in nonreducing gels, the same procedure as above was followed, except that cells were lysed in nonreducing hypotonic buffer (without DTT), and the loading buffer did not contain β-mercaptoethanol.
To control original protein input, 20 µl of the initial lysate was loaded in the gels.
Electrophoretic mobility shift assay (EMSA)
Nuclear extracts were prepared after lysis in high-salt extraction buffer as described previously (31)
and stored at –80°C. Aliquots of nuclear extracts (20 µg of protein) were incubated with a 32P-labeled B DNA probe, consensus NF-B site 5'-AGTTGAGGGGACTTTCCCAGGC-3' (Eurogentec, Liege, Belgium) followed by analysis of DNA binding activities by EMSA. The binding reaction was performed as described (31)
. Complexes were analyzed by electrophoresis on a nondenaturing 5% polyacrylamide gel. DNA-protein complexes migrate more slowly than the unbound radioactive probe and consequently they are visualized as distinct bands of radioactivity. The acrylamide gel image was detected by autoradiography. In immunodepletion experiments, before incubation with labeled probes the nuclear extracts were incubated for 15 min on ice with 2 µg of antibodies (anti-p50 and anti-p65).
TransAM
TransAM assays were performed according to the manufacturer’s instructions (Active Motif, Carlsbad, CA, USA). TransAM NF-B family kits are ELISA-based kits designed specifically for the study of NF-B pathways. In brief, 10 µg of nuclear protein samples were incubated for 1 h in a 96-well plate coated with an oligonucleotide that contains the NF-B consensus site (5'-GGGACTTTCC-3') to which activated NF-B factors contained in nuclear extracts specifically bind. By using an antibody directed against an epitope on p50 or p65 that is accessible only when NF-B is activated and bound to its target DNA, the NF-B complex bound to the oligonucleotide is detected. After incubation for 1 h with a secondary HRP-conjugated antibody, specific binding was detected by colorimetric estimation on a spectrophotometer at 450 nm with a reference wavelength of 655 nm.
Small interfering RNA (siRNA) transfection
Two micrograms of siRNA p50 or siRNA p65 (Dharmacon, Lafayette, CO, USA) was transfected into 106 cells using kit V from Amaxa (Cologne, Germany) according to the manufacturer’s protocol and program V-001 with an Amaxa Nucleofector device. Treatments were performed 24 h after transfection. One microgram of enhanced green fluorescent protein (GFP) expression vector (pmaxGFPTM) was cotransfected with the siRNA; the percentage of transfected cells based on the GFP green signal were measured by flow cytometer. All treatments were performed with 60% transfected cells.
Statistical analysis
Data are presented as fold increase ± SD of treated vs. untreated cells. Statistical differences were determined using Student’s t test for unpaired data, and P < 0.05 was considered significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
B activation involves p50 but not p65 and occurs independently of IB degradationB in U937 cells (23)
. Here, we investigate which NF-B pathway/subunits are involved in this activation.
Figure 1A
reports the time course of NF-B activation, showing a time-dependent increase up to 24 h. The specificity of the shifted bands was then investigated by EMSA assays performed after immunodepletion of p50 or p65 (Fig. 1B, C
, respectively), which are the components of the most frequent forms of active NF-B. The band shift promoted by BSO is impaired by anti-p50 antibodies, being instead insensitive to anti-p65, which indicates that p65 might be not implicated in BSO activation of NF-B. Quantification through the TransAM assay confirms this finding, showing p50, but not p65, activation with BSO (Fig. 1D
).
|
The localization of p50 and p65, which are known to translocate to the nucleus on activation, was assessed by Western blot for nuclear and cytosolic fractions. TNF-, used as a control of the canonical pathway, induces early translocation of both subunits, as expected (Fig. 1E
). Instead, BSO only induced nuclear translocation of p50 without affecting the localization of p65.
Because the noncanonical pathways are independent of IB degradation, we analyzed by Western blot the level of IB in cells treated with BSO at different time points. TNF-, as expected, leads to very early IB degradation (10 min) (Fig. 2A
), whereas no IB degradation is induced by BSO, at early (left panel) or late (right panel) time points (Fig. 2B
). To exclude possible involvement of the IB system, other subtypes were analyzed, namely IBβ, IB, and IB; Fig. 2C
shows that they are degraded by BSO. This result indicates that NF-B activation by BSO must occur via other, IB-independent mechanisms. All of this evidence allows us to exclude the possibility that BSO may activate the canonical pathway of NF-B activation.
|
To explore whether some of the other subunits of NF-B (p52, c-Rel, and Rel B) relocalize to the nucleus on BSO treatment, we performed cell fractionation followed by specific Western blot analysis. None of these subunits translocates to the nucleus (Fig. 2D
). This finding allowed exclusion of the possible activation of the alternative (p52-relB) or any other pathway. The only subunit that translocates on BSO treatment is p50, implying that the activated dimer may in fact be the p50-p50 homodimer and thus suggesting that BSO-induced NF-B activation may be of a noncanonical nature.
Nuclear colocalization of p50 and Bcl-3 in BSO-treated cells
The mechanism of activation of p50 was investigated. A reported mechanism is the processing of its precursor p105. Figure 3A
shows the dynamics of p105 and p50 in BSO- vs. TNF--treated U937 cells. TNF- indeed increases p50 while at the same time it decreases p105. Conversely, the p50 increase by BSO treatment is not accompanied by a p105 decrease. In fact, p105 even increased, indicating that activation of p50 must occur via a different mechanism. Many reports indicated that the formation and/or nuclear translocation of the p50-p50 homodimer is controlled by Bcl-3, whose cytosolic vs. nuclear localization may be related with activation. We assessed Bcl-3 intracellular localization by Western blot analysis of nuclear and cytosolic fractions (Fig. 3B
) and by immunofluorescence analysis (Fig. 3C
). Both approaches clearly show that BSO promotes early nuclear translocation of Bcl-3; a quantification of the fraction of cells presenting nuclear localization of Bcl-3 is shown in the right panel of Fig. 3C
.
|
To understand whether the BSO-induced nuclear translocations of p50 and Bcl-3, are independent events or imply the formation of a Bcl-3/p50 complex, we performed an in situ indirect coimmunofluorescence analysis by labeling, on the same cells, p50 and Bcl-3 with fluorescein isothiocyanate- or tetramethylrhodamine B isothiocyanate-linked antibodies, respectively, and identifying the cell nucleus by Hoechst staining (Fig. 4A
). The results shown in the second (p50) and third (Bcl-3) columns confirm that BSO stimulates nuclear translocation of both proteins with similar kinetics. The fourth column (p50+Bcl-3) shows a very early colocalization of the two proteins. The fourth column (merge) indicates that the p50-Bcl-3 colocalization begins in the cytosol (3 h of BSO) and moves to the nucleus at later times. TNF- instead induces a completely different picture, compatible with the canonical activation pathways.
|
BSO promotes early formation and nuclear translocation of a disulfide Bcl-3/p50 complex
To demonstrate a physical interaction between p50 and Bcl-3, we performed immunoprecipitation assays of nuclear cell lysates by means of either anti-p50 or anti-BCL-3 antibodies followed by the evaluation of the presence of both proteins in the immunoprecipitate. BSO greatly increased the cross-precipitated Bcl-3 and p50 bands already after 3 h (Fig. 4B
), showing that Bcl-3 is physically associated with p50 in the nucleus and confirming the results of the image analyses (Fig. 4A
).
We analyzed the nature of the BCL-3/p50 physical association. We know that at 3 h of BSO, 
25% of the intracellular GSH is lost (32)
; we previously demonstrated that under these circumstances critical disulfide bonds form between reactive cysteines (32)
. Because both p50 and Bcl-3 possess exposed cysteines, we investigated whether the binding between p50 and Bcl-3 may be of a disulfide nature. We performed a Western blot for p50 or Bcl-3 under denaturing (SDS) but nonreducing PAGE; interprotein disulfides would slow down migration of the proteins involved, which will migrate as complexes with a mass that is the sum of the single masses. We found that in untreated cells both Bcl-3 and p50 exist as monomers, migrating at 60 or 50 kDa, respectively, or as part of a disulfide-linked complex that migrates, for both proteins, at 160 kDa (Fig. 4C
, top and middle panels). Interestingly, this corresponds to the Mr of a hypothetical dimer Bcl-3-p105 or a p50-p50-Bcl-3 trimer in the cytosol, and a trimer p50-p50-Bcl-3 in the nuclear fraction. The same samples run with regular, reducing PAGE only show the respective monomers (Fig. 4D
; top panel, Bcl-3; bottom panel, p50). This result means that the basal Bcl-3/p50 complex is of a disulfide nature. At 3 h of BSO, the fractions of both proteins present as disulfide complexes vs. monomers increase, especially in the nuclear fractions, indicating that both proteins undergo a redox-mediated interchain disulfide formation (Fig. 4C
, top and middle panels). Interestingly, p65, whose involvement in the BSO response was already ruled out (see Fig. 1
), is not participating in any disulfide complexes but rather is migrating according to its monomeric Mr in the nonreducing gels also. To demonstrate that the coimmunoprecipitated Bcl-3 and p50 are linked by a disulfide, we performed the procedures for immunoprecipitation, omitting all reducing treatments. Fig. 4E
shows that all nuclear Bcl-3 coprecipitated by anti-p50 migrates as a 160-kDa complex, i.e., being part of a disulfide, thus excluding other types of interaction.
BSO-induced ROS are responsible for the late Bcl-3-p50 nuclear translocation and NF-B activation
The events that we have shown to occur at late time points [p50 and Bcl-3 nuclear translocation (Figs. 1E
and 3A
, respectively) and NF-B activation (Fig. 1A
)], may be the result of a slow process; alternatively, they may be the response to an abrupt intracellular alteration. Using 2,7-dichlorofluorescein diacetate to detect intracellular ROS, we discovered that the ROS level does not increase with a steady slope after BSO treatment but shows an abrupt increase at 8 h (Fig. 5A
), indicating that at that moment a deep intracellular change is occurring. To understand whether ROS may be the cause of the late events, we used the radical scavenger Trolox C, an analog of vitamin E, which efficiently blocks BSO-induced ROS production (Fig. 5A
). We discovered that the scavenging of ROS production with Trolox C blocks BSO-induced NF-B activation in an EMSA assay (Fig. 5B
). Accordingly, p50 and Bcl-3 nuclear relocalization also is prevented, even if the two proteins apparently colocalize in the cytosol (Fig. 5C
). Instead, the potentiation of the disulfide Bcl-3/P50 complex induced by BSO is not impaired (Fig. 5D
), consistent with the early occurrence of this event.
|
p50 and ROS are required for BSO-induced Bcl-2 up-regulation and survival
To understand whether the Bcl-3/p50 complex may play a role in the up-regulation of Bcl-2 by BSO, we silenced p50 expression by using siRNA and probed the cells for BSO-dependent Bcl-2 up-regulation. Indeed, p50-silenced U937 cells are no longer able to up-regulate Bcl-2 as a response to BSO (Fig. 6A
). Instead, silencing of p65 does not impair BSO-induced Bcl-2 up-regulation (Fig. 6B
). Next, we probed the role of p50 activation in cell survival by analyzing the effect of BSO on viability of normal or p50-silenced U937 cells. Fig. 6C
shows that BSO turns into a cytotoxic treatment if p50 is silenced.
|
The correlation between the complex Bcl-3/p50 and Bcl-2 was further explored by analyzing the role of ROS in BSO-induced Bcl-2 up-regulation and cell survival. As shown in Fig. 6D, E
, ROS are required to trigger the survival response, i.e., Bcl-2 up-regulation and survival. Thus, BCO-induced Bcl-2 up-regulation and survival depend on p50 and ROS, suggesting that the Bcl-2 promoter may be a target for the p50-p50 homodimer.
Possible involvement of p38 in the BSO-induced NF-B activation
The evidence provided so far points to the requirement of ROS to pass from the preparative to the functional phase of BSO-induced NF-B activation. To understand the molecular switch responsible for the passage, we analyzed the possible involvement of p38, a redox-sensitive mitogen-activated protein (MAP) kinase that seems particularly interesting in this context because 1) it responds to direct oxidations rather than to thiol disequilibrium, and 2) it controls the activation state of cAMP response element-binding protein-binding protein (CBP)/p300, a coactivator of Bcl-3. We determined a time course of p38 activation, by analyzing the nonphosphorylated vs. the phosphorylated form of the protein. Interestingly, the time course of p38 activation overlaps the time course of ROS formation (Fig. 7A
).
|
To assess a possible role of p38 in BSO-induced NF-B activation, we analyzed the end points of the response to BSO, i.e., survival and Bcl-2 up-regulation on p38 inhibition with SB 203580. Figure 7B
shows that the p38 kinase inhibitor sensitizes U937 cells to BSO. At the same time, it prevents BSO-dependent Bcl-2 up-regulation, as results from Western blot (Fig. 7C
) and flow cytometric (Fig. 8D
) analysis indicate. Inhibitors of other critical kinases, such as PD 98059 [extracellular regulated kinase (ERK) inhibitor], U0126 (ERK and HEK inhibitor), LY 294002 (AKT inhibitor), and SP 600125 (c-Jun NH2-terminal kinase inhibitor), although affecting the viability of U937 cells, did not alter the effect of BSO on Bcl-2 expression or cell survival (Fig. 7B-D
), excluding the possibility that these other kinases may play a role in the BSO-induced survival pathway.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. This is not a general cell behavior, because in most cells GSH depletion leads to Bcl-2 down-regulation, perhaps finalizing or favoring a death signal (33
, 34)
; for this reason, GSH depletion is often used as a coadjuvant in antitumor therapies, with the aim of sensitizing tumor cells to chemotherapeutic agents. It is known that tumor cells that are resistant to BSO chemosensitization often possess high basal levels of Bcl-2; alternatively, they may conditionally up-regulate it, being, in fact, responder cells (23)
. Here, we have presented one molecular mechanism at the basis of such a conditional response, involving the multistep activation of a redox-sensitive noncanonical NF-B pathway.
Bcl-2, although possessing the required recognition sequences in its promoter (24)
, is seldom considered to be a target of NF-B transcription, but evidence of a possible role of noncanonical NF-B complexes has been presented, involving the p50-p50 homodimer (24
, 25)
. We have ruled out the possibility that BSO may up-regulate Bcl-2 via the canonical NF-B pathway, as IB and p65 are unaffected. Instead, we provide substantial evidence that the trans-activation is mediated by a noncanonical NF-B, because 1) silencing of p50, but not of p65, prevents BSO-induced Bcl-2 up-regulation; 2) the time course of Bcl-2 transcription coincides with the appearance of the DNA binding activity of the p50-containing complex; and 3) scavenging of the BSO-produced ROS inhibits both the activation of the p50-containing NF-B complex, and the trans-activation of Bcl-2.
The sensitivity to redox imbalance and oxidative stress of the NF-B complexes is a consolidated notion, even though many discrepancies as to whether it plays an activating or repressing role are still unsolved (18)
. In particular, much evidence has been presented as to the requirement of GSH for canonical pathway NF-B activity (20)
. What we demonstrate here is that the noncanonical pathway is activated by redox imbalance and oxidative stress.
This mechanism is rather complex, being in fact a multistep activation. Formation of a Bcl-3/p50 complex occurs in the cytosol and may be a direct response to redox disequilibrium. Such redox imbalance is perceived by other redox-sensitive proteins at early BSO treatment times: we have recently shown that at 3 h of BSO the proapoptotic protein Bax, which possesses two reactive cysteines, forms an interprotein disulfide, thereby undergoing oxidative activation (32)
. Such a mechanism is very important during redox alterations, because the proteins that need to be activated or repressed are themselves a sensor of the redox disequilibrium, rendering the system self-sufficient. Therefore, we wanted to investigate whether a similar activation mechanism might be at the basis of the BSO-induced noncanonical NF-B activation. This is conceivable because both proteins possess several critical cysteine residues that are exposed, thus being theoretically prone to oxidative dimerization: on the one hand, Bcl-3 possesses two exposed cysteines within the fourth and seventh ankyrin repeats (35)
, i.e., the site of interaction with p50; on the other hand, at least one cysteine is present in the p50-p50 dimer binding site and two cysteines (Cys-270 and Cys-259) are present in the putative site of interaction with Bcl-3 (36)
.
It is known that the activity of cytosolic proteins with exposed cysteines is often regulated by oxidative dimerization, but to our knowledge this type of interaction was never searched for in the Bcl-3/p50 system. By performing Western blot analysis in nonreducing conditions, we discovered that the two proteins form a disulfide-mediated, DTT-sensitive complex of 160 kDa, which is recognized by antibodies against both Bcl-3 and p50, suggesting that the interaction between the two proteins is of a disulfide nature. The 160-kDa complex is compatible with a trimer composed of either two p50 and one Bcl-3 molecules (implying either a p50-SS-Bcl-3-SS-p50 or a Bcl-3-SS-p50-SS-p50 asset) or a p50-SS-Bcl-3-SS-PX, with PX being a possible actor with putative nuclear targeting role. If we clearly showed the formation of the disulfide complex, further studies are required to demonstrate its biological relevance. So far, we can speculate that at the functional level the disulfide complex, possessing a Bcl-3 moiety, might give the p50 dimer the possibility of becoming transcriptionally active, as already proposed (13
, 14)
.
When we compared the time courses of the different events described in the Results, it became evident that there are "early" (1–6 h) and "late" (>8 h) events. The early events consist of slight ROS production and slight p38 activation, together with substantial GSH depletion (32)
. This set of events occurs when the Bcl-3/p50 complex is assembled in the cytosol, as results from the immunoprecipitation and the nonreducing Western blot analyses, and migrates to the nucleus.
At late times, 6–8 h, we observed a brisk increase of ROS, together with stronger p38 activation. Thereafter, DNA binding of the activated complex occurs (>8 h of BSO), that is, after ROS production and phospho-p38 reach their plateau levels. p38 is known to be directly activated by ROS (37)
, thus possibly explaining the overlapping time course of ROS and p38 activation as a cause-effect relationship. We showed that inhibition of p38 impairs the ability of BSO to up-regulate Bcl-2 and to preserve cell viability. Does this allow us to hypothesize a direct cause-effect relationship between p38 phosphorylation and transcriptional activation of the p50-Bcl-3 complex? It is known that transcriptional activation of Bcl-3 may be facilitated by several coactivators, among them, CBP/p300 histone acetyltransferase (38
, 39)
, whose DNA unwinding activity is regulated by the p38 MAP kinase pathway (40)
. Oxidative stress and proinflammatory mediators have been suggested to influence histone acetylation and phosphorylation via a mechanism dependent on the activation of the MAP kinase pathway (41
, 42)
. All of this evidence will help in clarifying the role and the direct targets of the p38 activated by BSO. To this purpose, we cannot even exclude a paracrine phenomenon, i.e., that the role of p38 might be the stimulation of the release of biologically active factors, whose binding to cognate cells leads to activation of the already primed Bcl-3/p50 complex. This interpretation may also allow inclusion of a receptor stimulation event in this noncanonical, p50-p50 pathway, in the same guise as it occurs in the other NF-B pathways.
The radical scavenger Trolox C was used to assess the dependence of the noncanonical NF-B activation on ROS formation. Indeed, in its presence all of the late events, i.e., NF-B activation, Bcl-2 up-regulation, and cell survival to BSO, are prevented. Regarding the early events, we show that the formation of the Bcl-3/p50 complex is allowed, even though it does not migrate to the nucleus. It is conceivable that Trolox C, although scavenging ROS, may not affect disulfide formation. Thiol alteration is a milder oxidative condition than the direct presence of ROS; indeed, GSH depletion may differentially affect protein thiols, depending on their spatial configuration, thus allowing the sensitive ones to react even at milder redox imbalances. Stronger and prolonged GSH depletion deprives cells of an important antioxidant defense; thus, ROS produced by the regular aerobic metabolism are less efficiently scavenged, thus accumulating within cells. Trolox C as a ROS scavenger should barely affect thiol equilibrium.
The early (thiol disturbance) and late (ROS formation) events of the noncanonical NF-B pathway could both be turned into very early events by a frank oxidative treatment. Indeed, 1 h of 50 µM H2O2 is sufficient to trigger the entire response in terms of nuclear translocation of Bcl-3 overexpression of Bcl-2 and the requirement of p50 over p65 (data not shown).
Our data indicate that the noncanonical NF-B pathway seems to be active already in steady-state U937 cells, as shown by the high basal levels of p50 and Bcl-3 in the nucleus and by the presence of the 160-kDa disulfide complex; in fact, the redox imbalance and oxidative stress due to BSO seem to strongly potentiate a signal that is already "on." These high steady-state levels may be ascribed to the very strong oxidative metabolism displayed by U937 cells, with high basal levels of ROS; consistently with (and perhaps in response to) this, they possess unusually high levels of GSH (27)
and Bcl-2 (23)
, the two molecules being directly or indirectly implicated in coping with oxidative stress. Thus, Bcl-2 levels in untreated U937 cells may be maintained by mechanisms similar to those described here for BSO, i.e., oxidative signaling. In support of this theory, we observed that accidental cell stress increases basal Bcl-2 levels, which cannot be further increased by BSO challenge (unpublished results).
The molecular bases that allow only some cells to be responders, that is, able to react to oxidations by building up a survival response to GSH depletion, remain an open question. A candidate might be the cytosolic localization of Bcl-3, which characterizes only a subset of cells, because in most cases Bcl-3 is exclusively localized in the nucleus. Bcl-3 is reported to have a cytosolic location in hepatocytes (10)
; interestingly, the hepatocytic cell line HepG2 is able to up-regulate Bcl-2 in response to BSO (23)
. Because Bcl-3 seems to be the primum movens of the response to BSO and the first event is cytosolic, the survival pathway might not be initiated without a cytosolic available Bcl-3.
In conclusion, these findings show that the noncanonical pathway of NF-B may be activated by oxidation (as depicted in Fig. 8
) and suggest the intriguing scenario that oxidations may switch a canonical to a noncanonical NF-B pathway, perhaps allowing the passage to different stages of the inflammatory response (17
, 43)
.
|
ACKNOWLEDGMENTS |
|---|
Received for publication January 28, 2008. Accepted for publication August 14, 2008.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
immunoglobulin enhancer-binding protein NF-B by a posttranslational mechanism. Cell 47,921-928[CrossRef][Medline]B: players, pathways, perspectives. Oncogene 25,6680-6684[CrossRef][Medline]B: a lesson in family values. Cell 80,529-532[CrossRef][Medline]B activity by exchange of dimers. Mol. Cell 11,1563-1574[CrossRef][Medline]B with chromatin: the art of being at the right place at the right time. Nat. Immunol. 6,439-445[CrossRef][Medline]B signaling transduction and transcription. Cell. Mol. Immunol. 1,425-435[Medline]B and cancer. Clin. Oncol. (R. Coll. Radiol.) 19,154-161[Medline]B activity in haematological malignancies. Biochem. Pharmacol. 72,1069-1080[CrossRef][Medline]B pathways. Immunity 17,525-535[CrossRef][Medline]B-inducible BCL-3 expression is an autoregulatory loop controlling nuclear p50/NF-B1 residence. J. Biol. Chem. 276,32080-32093B subunit regulation in nontransformed CD4+ T lymphocytes. Science 256,1452-1456Bs: alternative regulation of NF-B activity. FASEB J. 21,2642-2654B homodimers from the cytoplasmic pool of p50–p105 and nuclear translocation. EMBO J. 16,3609-3620[CrossRef][Medline]B p50 homodimers. Genes Dev. 7,1354-1363B signaling. Cell 125,665-677[CrossRef][Medline]B activation: a reassessment of the evidence in the light of recent discoveries. Biochem. Pharmacol. 59,13-23[CrossRef][Medline]B: a mechanism for redox-induced inhibition of DNA binding. Biochemistry 40,14134-14142[CrossRef][Medline]B activity by intracellular glutathione levels and by variations of the extracellular cysteine supply. FASEB J. 9,246-252[Abstract]B and transcription of human immunodeficiency virus. Proc. Natl. Acad. Sci. U. S. A. 87,9943-9947B activation in HeLa cells. Exp. Mol. Med. 34,332-339[Medline]B1 (p50) homodimers contribute to transcription of the bcl-2 oncogene. J. Biol. Chem. 276,45380-45386B2/p100 induces Bcl-2 expression. Leukemia 17,1349-1356[CrossRef][Medline]B protein family. EMBO J. 20,6180-6190[CrossRef][Medline] B p50 homodimer bound to a B site. Nature 373,303-310[CrossRef][Medline]B/Rel and nuclear co-regulators. Oncogene 18,3316-3323[CrossRef][Medline]B activation and ICAM-1 expression in endothelial cells. Am. J. Physiol. Lung Cell. Mol. Physiol. 287,L1017-L1024B activation pathways and their role in innate and adaptive immunity. Trends Immunol. 25,280-288[CrossRef][Medline]This article has been cited by other articles:
|
|
 |
|
  K. A. Khan, A. Coaquette, C. Davrinche, and G. Herbein Bcl-3-Regulated Transcription from Major Immediate-Early Promoter of Human Cytomegalovirus in Monocyte-Derived Macrophages J. Immunol., June 15, 2009; 182(12): 7784 - 7794. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
