Noxa is necessary for hydrogen peroxide-induced caspase-dependent cell death

2.1. Reagents 

An anti-Noxa monoclonal antibody (clone 114C307) was obtained from Abcam (Cambridge, UK). An anti-Tp53 monoclonal (DO-7), anti-Bim polyclonal, anti-Mcl-1 monoclonal, and anti-cytochrome c monoclonal antibodies (clone 7H8.2C12) were purchased from BD Pharmingen (San Diego). Anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH) monoclonal antibody and anti-Mcl-1 polyclonal antibody were obtained from Chemicon International (Temecula) and from Santa Cruz Biotechnology (Santa Cruz), respectively. Anti-cAMP response element-binding protein (CREB) monoclonal antibody was from Cell Signaling Technology (Danvers). Hydrogen peroxide (H2O2) was obtained from Wako Co. (Osaka, Japan) and z-VAD.fmk was from the Peptide Institute (Osaka, Japan).

2.2. Cell culture and DNA transfection 

Human Jurkat cells (T leukemia) and HeLa/D98 cells were grown at 37°C in RPMI 1640 medium supplemented with 10% fetal bovine serum, 100U/ml penicillin, and 100μg/ml streptomycin. DNA encoding FLAG-tagged human Noxa in the pUC-CAGGS expression vector or pZsProSensor-1 vector (Clontech) were used to transfect Jurkat cells with the Amaxa electroporation system (Nucleofector kit V, program A-17). The transfection efficiency was more than 50% as assessed by transfection with DNA expressing GFP. The siRNAs for cytochrome c were produced by Dharmacon Research (Chicago) and had the following nucleotide sequences: 5′-AAGCAUUAAGAAGAAGGAAGA-3′ (#54) and 5′-AAGGAAGAAAGGGCAGACUUA-3′ (#60). siRNAs for other human genes, including Noxa as well as negative control siRNA, were purchased from Qiagen (Hilden, Germany). Cells (1×106) were transfected twice on alternate days with 3–13μg of siRNA using the Amaxa electroporation system. At 48h after the transfection, cells were used for the experiments.

2.3. Cell viability 

After staining with propidium iodide (1μM), FITC-conjugated Annexin V (1μM), or Hoechst 33342 (1μM) for 10min, cell death was assessed by a flow cytometer (BD Biosciences, FACS Canto II) or by a fluorescence microscope (Olympus, BX50).

2.4. Preparation of whole-cell lysates and cytoplasmic fractions 

For the detection of cytochrome c released from the mitochondria, cytoplasmic fractions were collected from Jurkat cells after incubation with 0.1mg/ml digitonin for 5min at 37°C in isotonic buffer [20mM potassium-Hepes (pH 7.4), 10mM KCl, 1.5mM MgCl2, 250mM sucrose, and 1mM Na2+-EDTA]. After centrifugation, aliquots of the supernatant (cytoplasmic fraction) and the pellet containing the mitochondria were analyzed by Western blotting. In some experiments, cells were lysed in 50mM Tris–HCl (pH 8.0), 0.1% SDS, 1% Nonidet P-40, 0.5% deoxycholate sodium salt, and 150mM NaCl to obtain whole-cell lysates.

2.5. Luciferase assay 

The Firefly luciferase plasmid pGL4.18 (Promega) was used to generate plasmids carrying different regions of the Noxa promoter (−5619 to +153, −925 to +153, −198 to +153, −134 to +153, and −74 to +153): pNoxa5619-luc, pNoxa925-luc, pNoxa198-luc, pNoxa134-luc, and pNoxa74-luc, respectively. pNoxa198+1-luc was a derivative of pNoxa198-luc, lacking +1 to +136. pNoxa198mt-luc carried a mutation of the p53-binding site [10]. pNoxa134CB-luc had two mutations (CTACTCCA) within the cAMP response element (CRE) site (CTACGTCA). Cells were co-transfected with 1μg of control Renilla luciferase pRL-SV40 (promega) and with 10μg of Firefly luciferase reporter constructs by Amaxa electroporation. Luciferase activity was measured by using a luminometer and the Dual-Glo luciferase assay system (Promega).

2.6. Other methods 

Immunoprecipitation was done with anti-Mcl-1 antibody as described [14] and proteolytic activities were measured as described [17].

3.1. Induction of mitochondria-dependent apoptosis by H2O2 in Jurkat cells 

H2O2 induced death of Jurkat cells as assessed by Annexin V staining (Fig. 1A and Fig. S1), which was inhibited by the pan-caspase inhibitor z-VAD.fmk or by Bcl-2 overexpression (Fig. 1A and Fig. S1). Staining with Hoechst 33342 revealed typical apoptotic nuclear morphological changes (chromatin condensation and fragmentation), which were also inhibited by z-VAD.fmk (Fig. 1B). These results demonstrated that H2O2 induced apoptosis of Jurkat cells.

View Large Image Download to PowerPoint

Fig. 1. Induction of mitochondria-dependent apoptosis in Jurkat cells by H2O2. (A) Inhibition of H2O2-induced cell death by overexpression of Bcl-2 and by a pan-caspase inhibitor. Parental cells (open circles), two Bcl-2-overexpressing derivatives [HB2-8 (red crosses) and HB2-10 (red open squares)], and two vector-transfected derivatives [V1 (open triangles) and V2 (filled triangles)] were treated with H2O2 (90μM). Parental cells were also treated with H2O2 in the presence of 50μM z-VAD.fmk (blue open circles). The extent of cell death was determined by Annexin V staining followed by FACS analysis. Representative dot plots for Annexin V staining were shown in Fig. S1. Data are shown as the means±S.D. (n=3). (B) Apoptotic nuclear changes induced by H2O2. Nuclei were stained with Hoechst 33342. (C) H2O2-induced cytochrome c release. Jurkat cells were treated with H2O2/z-VAD.fmk and subjected to subcellular fractionation, followed by Western blotting. Note pellet fractions contained the mitochondria and unbroken cells, with the latter showing signals for GAPDH. (D and E) Inhibition of H2O2-induced cell death by silencing of Cyt. c. (D) Jurkat cells were treated with two different Cyt. c siRNAs, after which expression of Cyt. c and GAPDH was analyzed. (E) Jurkat cells with silencing of Cyt. c were treated with H2O2, after which cell death was assessed (the means±S.D., n=3). (F and G) Inhibition of H2O2-induced death by ablation of Bax/Bak. Jurkat cells were treated with 5μg of Bak siRNAs, and expression of Bak was analyzed by Western blotting. (G) Jurkat cells with Bak siRNA or control siRNA with or without z-VAD.fmk were treated with H2O2. Cell death was assessed (the means±S.D., n=3).

Cytochrome c (Cyt. c) was released from the mitochondria of H2O2-treated Jurkat cells (Fig. 1C). Silencing of Cyt. c (Fig. 1D) strongly suppressed H2O2-induced apoptosis (Fig. 1E). To determine whether H2O2-induced apoptosis was dependent on Bax/Bak, Jurkat cells which lacks Bax expression [18] were transfected with Bak siRNA (Fig. 1F). Silencing of Bak markedly reduced H2O2-induced apoptosis (Fig. 1G). These results indicated that the mitochondrial pathway was involved in H2O2-induced apoptosis of Jurkat cells.

3.2. Crucial role of Noxa in H2O2-induced apoptosis of Jurkat cell 

BH3-only members of the Bcl-2 family either activate multidomain pro-apoptotic Bax/Bak or inhibit anti-apoptotic members, such as Bcl-2 [19], [20]. In response to apoptotic stimuli, BH3-only proteins are activated by either transcriptional upregulation or by posttranslational modifications [10], [21], [22], [23]. Toward determining which BH3-only proteins were involved in H2O2-induced apoptosis of Jurkat cells, we analyzed levels of various BH3-only proteins. As shown in Fig. 2A and Fig. S2, there was an increase of Noxa, Bim, Hrk, Bnip3, and Bnip3L protein levels, whereas Bmf, Bad, and Bik proteins were decreased.

View Large Image Download to PowerPoint

Fig. 2. Role of Noxa in H2O2-induced apoptosis. (A) H2O2-induced accumulation of Noxa protein. Jurkat cells were treated with H2O2/z-VAD.fmk. The cells were harvested, lysed, and subjected to Western blotting. GAPDH was used as loading control. (B and C) Inhibition of H2O2-induced apoptosis in Jurkat cells by silencing of Noxa. Cells were transfected with 13μg of siRNA for Noxa or negative control siRNA. Expression of Noxa and GAPDH was analyzed by Western blotting (B). (C) After 48h of transfection, cells transfected with the indicated siRNA were treated with H2O2. Then cell death was assessed by Annexin V staining. Data are shown as the means±S.D. (n=3) (∗P<0.0001). (D) Jurkat cells were transfected with 13μg each of the indicated siRNAs. Expression of Noxa, Bid, Bim, and GAPDH after 48h of transfection was analyzed by Western blotting. After transfection, cells with silencing of the indicated proteins were treated with H2O2 for 12h. Dead cells were counted (the means±S.D., n=3). (E) Increased association of Noxa with Mcl-1 and decreased association of Bim with Mcl-1 in response to H2O2. Jurkat cells were treated with H2O2 for 12h and cell extracts were immunoprecipitated with anti-Mcl-1 antibody. Normal rabbit IgG served as the negative control. An aliquot of whole-cell lysate (input) and IP samples were subjected to Western blotting analysis. A representative result of three independent experiments that yielded similar findings is shown.

To determine which BH3-only proteins were involved in H2O2-induced apoptosis, Jurkat cells were transfected with siRNAs for various BH3-only proteins, and the influence on apoptosis was assessed. Silencing of the respective BH3-only proteins was verified by Western blot analysis or reverse transcription polymerase chain reaction (RT-PCR) (Fig. 2B and D and data not shown). Among the proteins tested, silencing of Noxa was found to most significantly protect Jurkat cells against H2O2-induced apoptosis (Fig. 2C and Fig. S3). To investigate a role of Noxa in H2O2-induced apoptosis in other cell line, we examined with HeLa/D98 cells and found silencing of Noxa provided cells a significant level of resistance to H2O2-induced apoptosis (Fig. S4).

Silencing of either Bid or Bim also significantly protected Jurkat cells against H2O2-induced apoptosis, although to lesser extents (Fig. 2D). Two models have been proposed with regard to BH3-only protein function. According to one model, all BH3-only proteins act by inhibiting anti-apoptotic members of the Bcl-2 family [16]. According to the other model, Bim and Bid (or possibly Puma) directly activate Bax and Bak, while other members (including Noxa) inactivate the anti-apoptotic Bcl-2 family members, suggesting Noxa’s role upstream of Bid/Bim [14], [15]. Noxa is known to inhibit Mcl-1 (and Bfl-1/A1), but not Bcl-2 and Bcl-xL [14], [15]. Consistent with the second model, silencing of Noxa/Bim did not significantly increase protection against apoptosis compared with silencing of Noxa alone, and that of Noxa/Bid provided no more than slight resistance (Fig. 2D). Also consistent with the second model, Mcl-1-bound Noxa increased and Mcl-1-bound Bim showed a significant decrease in response to H2O2 (Fig. 2E). These results might support the possibility that increased Noxa in response to H2O2 binds with Mcl-1 to release Bim (Bid), which activates Bax/Bak.

3.3. Accumulation of Noxa by inhibition of proteasomal degradation after exposure to H2O2 

Noxa protein could be upregulated after exposure to H2O2 either by increased protein production or by inhibition of degradation. Since H2O2 reportedly affects the proteasomal degradation of proteins [2], [17], we investigated this possibility for Noxa by using cycloheximide and the proteasome inhibitor MG132. Since we found that treatment of Jurkat cells with these inhibitors resulted in an increase of endogenous Noxa mRNA, whereas Flag-tagged Noxa mRNA was not, which was produced from the transiently transfected pUC-CAGGS expression plasmid, we used the Flag-tagged Noxa for this purpose. The experiments with cycloheximide demonstrated that Flag-tagged Noxa turnover was more rapid in healthy cells than in H2O2-treated cells, with the t1/2 being 9.43 in H2O2-treated cells versus 2.08 in the control (Fig. 3A). Addition of MG132 restored Noxa levels in untreated cells (Fig. 3A). These results indicated that Noxa protein was subjected to proteasome-dependent degradation in healthy cells, whereas its degradation was inhibited by H2O2.

View Large Image Download to PowerPoint

Fig. 3. H2O2-induced stabilization of Noxa protein. (A) Jurkat cells expressing Flag-tagged Noxa from the pUC-CAGGS plasmid were treated with cycloheximide (1μg/ml) with or without H2O2 (90μM) or with MG132 (10μM), and Flag-tagged Noxa levels were determined by Western blotting. % values represent the Noxa signal intensity compared to the 0h time point. The t1/2 values are shown below. (B and C) Cells were treated with H2O2 at the indicated concentrations and whole-cell lysates were analyzed for ATP-stimulated (B) and ATP-independent (C) MG132-sensitive proteolytic activity, which represented total proteasome (26S and 20S) activity and 20S-proteasome activity, respectively. A representative result of three independent experiments yielding similar findings is shown. (D) Jurkat cells were transfected with 1μg of pZsProSensor-1 vector for 48h. Then the cells were treated with H2O2 or MG132, and cells positive for ZsGreen were counted by FACS analysis. Data are shown as the means±S.D. (n=3).

Inhibition of Noxa degradation might result from either H2O2-induced inhibition of ubiquitin processing or inhibition of proteasome. H2O2 treatment has been reported to directly inhibit 26S-proteasome activity in K562 cells at 400μM concentration [17]. Therefore, we measured the activity of 20S and 26S proteasomes in lysates of Jurkat cell treated with not only 90μM of H2O2 but also up to 500μM of H2O2 as described in Section 2. As shown in Fig. 3B and C, neither 20S-proteasome nor 26S-proteasome activity was significantly affected by H2O2, suggesting that the proteasome machinery itself was not affected by H2O2.

We then examined whether H2O2-induced inhibition of protein degradation was specific for Noxa by employing the pZsProSensor-1 vector, in which mouse ornithine decarboxylase (mODC) d410 was fused to the C-terminus of a green fluorescent protein (ZsGreen). mODC d410 contains several PEST sequences [24], which render the ZsGreen-mODC d410 fusion protein highly susceptible to proteasomal degradation. As shown in Fig. 3D, ZsGreen fluorescence gradually accumulated after H2O2 treatment, as it did with MG132. These results indicated that H2O2-induced accumulation of Noxa was at least partly due to the general inhibition of proteasome-dependent degradation.

3.4. Transcriptional upregulation of Noxa by exposure to H2O2 

In addition to inhibition of the proteasome-dependent degradation of Noxa, other processes (such as the accumulation of Noxa mRNA) might also contribute to the H2O2-induced increase of Noxa protein. In fact, the reverse transcription real-time PCR showed gradual accumulation of Noxa mRNA after H2O2 treatment (Fig. 4A). In general, mRNA accumulation is either due to its stabilization or to an increase of de novo synthesis. Since Noxa gene is known to be activated by different transcriptional factors in response to cellular stresses [10], [11], [12], [13], we examined the possible contribution of transcriptional activation of Noxa promoter to H2O2-induced mRNA accumulation by employing the luciferase assay. The Noxa promoter contains binding sites for various transcription factors, including E2F-1, Myc, p53, and HIF, which are located at positions +130, +87, −198, and −1110, respectively, from the transcription start site [10], [11], [12], [13]. As shown in Fig. 4B and C, pNoxa5619, the longest Noxa promoter and some shorter versions (pNoxa925, pNoxa198, pNoxa134, and pNoxa198+1) all exhibited a similar response to H2O2, indicating that transcriptional activation induced by H2O2 was mediated by the region of the Noxa promoter from −134 to +1 and +137 to +153, which excluded a role of p53. Consistently, a mutation of the p53-binding site (pNoxa198mt) did not inhibit H2O2-induced Noxa activation (Fig. 4B). Furthermore, silencing of p53 had no effect on the accumulation of Noxa or on cell death induced by H2O2 (Fig. S5), excluding a role of p53 in H2O2-induced Noxa transactivation.

View Large Image Download to PowerPoint

Fig. 4. Transcriptional activation of Noxa promoter by H2O2. (A) Noxa mRNA accumulation in Jurkat cells with H2O2. Jurkat cells were treated with H2O2/z-VAD.fmk. Noxa mRNA was quantified by RT-PCR. (B–D) Transcriptional activation of the Noxa promoter in response to H2O2. At 72h after transfection of reporter constructs, cells were treated without (open bars) or with H2O2/z-VAD.fmk (closed bars) for 6h, after which Firefly- and Renilla-dependent luciferase activity were determined. Firefly-dependent luciferase activity was normalized by the Renilla signals. Data are shown as the means±S.D. (n=3). (E and F) Jurkat cells were transfected with 13μg of the indicated siRNA for 48h. Then the cells were treated with H2O2/z-VAD.fmk. Expression of ATF4 (E) and Noxa (F) mRNA was analyzed by RT-PCR. GAPDH was used as an internal control. Data are shown as the means±S.D. (n=3).

Computer-based analysis of the Noxa promoter from −134 to +1 identified the CRE. Mutation within the CRE significantly reduced transcriptional activation of the Noxa promoter by H2O2 (Fig. 4D). It has been reported that CRE-binding proteins are involved in apoptotic cell death induced by various stresses [25], [26], [27], [28]. To investigate which CRE-binding proteins were involved in the transcriptional activation of Noxa after stimulation by H2O2, we transfected Jurkat cells with siRNAs for five CRE-binding proteins (CREB1, c-Jun, activating transcription factor (ATF)2, ATF3, and ATF4). Silencing of ATF4 significantly inhibited H2O2-induced accumulation of Noxa mRNA (Fig. 4E and F), whereas silencing of the other CRE-binding proteins did not (Fig. S6 and data not shown). These results suggested that ATF4 protein was involved in the H2O2-induced transcriptional activation of Noxa.