Journal list menu
The regulation of Bax by c-Jun N-terminal protein kinase (JNK) is a prerequisite to the mitochondrial-induced apoptotic pathway
Abstract
The signaling mechanism by which JNK affects mitochondria is critical to initiate apoptosis. Here we show that the absence of JNK provides a partial resistance to the toxic effect of the heavy metal cadmium. Both wild type and jnk−/− fibroblasts undergoing death exhibit cytosolic cytochrome c but, unlike wild type cells, the JNK-deficient fibroblasts do not display increased caspase activity and DNA fragmentation. The absence of apoptotic death correlates with a specific defect in activation of Bax. We conclude that JNK-dependent regulation of Bax is essential to mediate the apoptotic release of cytochrome c regardless of Bid and Bim activation.
1 Introduction
Apoptosis is a physiological process by which cells undergo a programmed cell death (PCD) [1]. The characteristic morphological changes displayed by apoptotic cells reflect the activation of a tightly regulated intrinsic cell signaling machinery that leads to the activation of caspases [2]. The stress-induced caspase cascade is triggered by the mitochondrial release of cytochrome c that initiates the formation of an oligomeric multiprotein complex, the apoptosome, that includes the apoptotic protease-activating factor 1, and dATP or ATP [3, 4]. How cytochrome c is released from the inter membrane space of the mitochondria remains a matter of intense debate [5]. The consensus model implicates the selective permeabilization of the outer mitochondrial membrane (OMM) by specific channels that involve Bcl-2 proteins [5, 6]. This model offers the advantage of maintaining functional mitochondria and distinguishes apoptosis from necrosis where the release of cytochrome c occurs via the rupture of the OMM caused by the sustained opening of the permeability transition pore (PTP) [7].
The Bcl-2 family includes proteins that inhibit (e.g., Bcl-2, Bcl-xL) and promote (e.g., Bax, Bak, Bid, Bim) apoptosis. Bcl-2, Bcl-xL and Bak reside predominantly in the mitochondria, whereas Bax, Bid, and Bim reside mainly in the cytosol of healthy cells [8]. The accumulation of Bax at the mitochondria associated with a change in Bax conformation constitutes a critical checkpoint in the release of cytochrome c [9, 10]. Activation of Bid correlates with its cleavage by caspase 8 and the generation of the death promoting fragment tBid [11, 12]. Unlike Bid, Bim becomes hyperphosphorylated upon apoptotic stress and dissociates from the microtubule-associated dynein motor complex where it is normally sequestered [13].
Evidence suggests that the c-Jun NH2-terminal protein kinase (JNK) cascade participates in the post-translational modifications of Bid and Bim [14-17]. Its requirement in mediating cytochrome c release was first demonstrated by the genetic analysis of homozygous jnk1 and jnk2 null (jnk−/−) mouse embryonic fibroblasts (MEFs) [17]. The essential role of Bax and Bak in mediating apoptosis via JNK is exemplified by the inability of a dominant active JNK to induce the death of cells deficient in Bax and Bak [18]. Whether JNK controls Bax/Bak function via the post-translational modification of Bid and Bim remains unresolved.
To further understand how cytochrome c is released from mitochondria we analyzed the effect of jnk gene deletion on the coordinated regulation of Bcl-2 proteins in response to cadmium (Cd). Our data demonstrate that the regulation of Bax by JNK is: (i) independent of Bim and Bid activation; and (ii) essential for mediating the apoptotic release of cytochrome c.
2 Materials and methods
2.1 Tissue culture and preparation of lysates
MEFs were cultured as previously described [17]. The cells were placed in 2% serum for 24 h prior to being treated with Cd chloride (CdCl2). Cytosolic proteins were extracted in triton lysis buffer [17]. For detecting the cytosolic release of cytochrome c and apoptosis inducing factor (AIF), the cells were incubated at 4 °C for 15 min in mannitol buffer (10 mM HEPES, pH 7.2, 210 mM mannitol, 70 mM sucrose, 5 mM sodium succinate, 0.2 mM EGTA) containing 80 μg/ml digitonin. Supernatants were subsequently concentrated ∼10-fold using Amicon Bioseparations centrifugal filters.
2.2 Biochemical studies
Lysates (50 μg) were analyzed by immunoblot [17]. Antibodies to cytochrome c (Pharmingen), AIF (Santa Cruz), caspase 3 (Cell Signaling Technology), poly (ADP-ribose) polymerase (PARP; Cell Signaling Technology), Bid (R&D Systems), Bim (Calbiochem), Bax (Upstate), Bak (Upstate), actin (Calbiochem), or tubulin (Sigma) were used. Immunocomplexes were detected by enhanced chemiluminescence (Amersham-Pharmacia). JNK activity was measured in cell lysates (30 μg) by pull down protein kinase assay [17].
2.3 Cell viability and caspase assays
Cell viability was measured by crystal violet (CV) staining [17] and MTT assay. For the MTT assay, cells were incubated with 1 μg/μl of 3-(4,5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide and lysed in DMSO. Caspase activity was measured by spectrofluorometer using the DEVD-AMC caspase 3 specific fluorogenic substrate.
2.4 Fluorescence microscopy
TUNEL was performed using the in situ Cell Death Detection kit (fluorescein) according to the manufacturer's instructions (Roche). For immunofluorescence the cells were fixed in methanol prior to being incubated with an anti-Bax antibody (clone 6A7, Sigma). Immune complexes were detected with a Texas red-conjugated anti-mouse immunoglobulin. Nuclei were stained with Hoechst 33342 (1 g/l). Fluorescence images were viewed with an Olympus Widefield microscope.
3 Results
The heavy metal cadmium (Cd) is a widespread industrial and environmental pollutant. Its toxicity is, at least in part, associated with its ability to cause necrotic and apoptotic cell death. Increased JNK activity was detected in wild type (wt) but not in jnk1 and jnk2 null (jnk−/−) MEFs incubated with 10 and 25 μM CdCl2 with a maximum of 35- and 75-fold at 18 and 6 h, respectively (Fig. 1 A). Control experiment showed no marked difference in the ability of Cd to increase p38 mitogen-activated protein kinase (MAPK) activity in both cell types (data not shown).
Consistent with the requirement of JNK in mediating apoptosis [17], jnk−/− MEFs were resistant to the toxic effect of 10 μM CdCl2 that resulted in 50% death of the wt MEFs at 20 h (Fig. 1B, left panel). In contrast, wt and jnk−/− fibroblasts displayed similar sensitivity to 25 μM CdCl2 resulting in 25% survival 15 h post-treatment (Fig. 1B, right panel). The finding that jnk−/− MEFs display distinct sensitivity to Cd toxicity was confirmed by MTT assay (Fig. 1C). Mitochondrial activity of wt but not jnk−/− cells was significantly impaired following incubation with 10 μM CdCl2 for 18 h (Fig. 1C, left panel). In contrast, both cell types displayed a similar decrease in mitochondrial activity when incubated with 25 μM CdCl2 (Fig. 1C, right panel). These results demonstrate that the protective effect of jnk gene deletion against Cd toxicity is dose dependent. High concentrations of Cd can trigger cell death via a JNK-independent mechanism.
To determine whether Cd triggers apoptotic death of fibroblasts, we examined its ability to induce DNA fragmentation, one hallmark of PCD [1]. Fragmented DNA was detected in wt fibroblasts treated with either 10 or 25 μM CdCl2 (Fig. 2 A). Concomitantly, the nuclei appeared shrunken with condensed chromatin (Fig. 2A). A similar change in the nuclear size was observed in the jnk−/− MEFs incubated with 25 μM but not 10 μM CdCl2 (Fig. 2A). However, no TUNEL positive JNK-deficient cells were detected (Fig. 2A).
The requirement of JNK in mediating apoptosis was confirmed by the lack of caspase 3 activation in jnk−/− MEFs treated with 10 or 25 μM CdCl2 (Fig. 2B). Immunoblot analysis demonstrated the proteolytic cleavage of caspase 3 in wt but not in jnk−/− MEFs following Cd treatment (Fig. 2C). The reduction of the inactive p32 proenzyme correlated with the appearance of the active p17 product and the cleavage of the poly (ADP-ribose) polymerase (PARP). Consistent with these results, the pancaspase inhibitor zVAD-fmk protected the wt but not the jnk−/− MEFs against Cd toxicity (data not shown). Altogether, these results provide strong genetic evidence that JNK is essential for mediating the apoptotic response of cells to Cd. In the absence of JNK, Cd triggers cell death via a caspase-independent mechanism.
To determine the relative importance of JNK1 and JNK2 in mediating the apoptotic response of fibroblasts to 10 μM CdCl2, we tested the effect of jnk1 and jnk2 deletion (Fig. 3 ). The tissue specific JNK3 isoform is not detected in MEFs [17]. No marked difference was observed between the ability of Cd to increase caspase 3 activity in wt and jnk2 null (jnk1+/+jnk2−/−) fibroblasts. In contrast, cells carrying homozygous ablation of the jnk1 gene (jnk1−/−jnk2+/+ or jnk1−/−jnk2+/−) were resistant to Cd treatment. A greater role for JNK1 in mediating Cd-stimulated caspase 3 activation is consistent with the partial restoration of the wild type phenotype following heterozygous deletion of the jnk1 gene in absence of JNK2 expression (jnk1+/−jnk2−/−).
To further establish the apoptotic defect caused by jnk gene deletion we examined the requirement of JNK in promoting the mitochondrial release of cytochrome c, one of the key events of apoptosis [3, 4] (Fig. 4 ). 10 μM CdCl2 promoted the release of cytochrome c and AIF in wt but not in jnk−/− MEFs (Fig. 4A, left panel). In contrast, a similar amount of cytochrome c and AIF was detected in the cytosolic fractions of wt and jnk−/− MEFs treated with 25 μM (Fig. 4A, right panel). These data indicate that the requirement of JNK for mediating Cd-induced cytochrome c release depends on the level of stress.
Next we demonstrated that neither zVAD-fmk (25 μM) nor cyclosporin A (CsA; 0.5 μM), an inhibitor of the PTP opening, blocked the ability of Cd to induce cytochome c release in fibroblasts (Fig. 4B and C). Similar results were obtained with 10 μM CsA (data not shown). Consistent with these data the decrease in cell viability following Cd treatment was not affected by the pretreatment of the cells with CsA (data not shown). We concluded that Cd-induced cytochrome c release does not implicate caspase activation or permeability transition, leading us to test the hypothesis that Cd initiates apoptosis via the regulation of Bcl-2 proteins.
The basal level of Bid, Bax, Bak and Bim expression was similar in wt and jnk−/− fibroblasts (Fig. 5 ). Cleavage of Bid in wt cells incubated with Cd correlated with loss of the full-length protein (25 kDa) and the concomitant appearance of Bid fragments at 20 and 15 kDa (tBid) (Fig. 5A). tBid was detected in the jnk−/− MEFs incubated with Cd. However, unlike in wt cells, this correlated with the accumulation of the full-length protein and the appearance of a 22 kDa fragment (Fig. 5A). We concluded that JNK is not essential for the appearance of tBid in response to stress. In addition to inducing the apoptotic cleavage of Bid, Cd increases Bid expression by a mechanism that is inhibited by JNK.
Bim is another BH3-only protein implicated in the intrinsic cell death signaling pathway. Incubation of the wt cells with 10 or 25 μM CdCl2 caused distinct electrophoretic mobility shifts indicative of Bim being phosphorylated (Fig. 5A). Similar phosphorylation pattern of Bim was detected in the jnk−/− cells but only in response to 25 μM CdCl2. These results indicate that JNK is the primary kinase to phosphorylate Bim, but its requirement can be overturned by increasing the level of stress.
No change in the level of Bak was detected in MEFs after Cd treatment (Fig. 5B). In contrast, a significant increase in Bax expression was detected in wt cells in response to 10 and 25 μM CdCl2 after 12 and 6 h stimulation, respectively (Fig. 5B). This effect was prevented by jnk gene deletion. Immunofluorescence analysis using a conformation specific anti-Bax antibody [19] showed the presence of active Bax at the mitochondria in wt but not in jnk−/− MEFs stimulated with Cd (Fig. 5C). Previous studies have shown that this characteristic change in Bax conformation precedes the apoptotic release of cytochrome c in the cytosol [10]. Consistent with the requirement of Bax in Cd-induced apoptosis, we found that Bax was essential for mediating increased cytochrome c release and caspase 3 activation in response to 10 μM CdCl2 treatment (Fig. 6 ).
These observations suggest that, in absence of JNK, cytochrome c is released from the mitochondria via a mechanism independent of Bax. Altogether these experiments clearly identify JNK as an essential regulator of Bax activity.
4 Discussion
Consistent with a previous study [18] our data provide direct genetic evidence that JNK is critical for mediating Bax activation in response to stress. Bax is not a JNK substrate (data not shown). The dissociation of the 14-3-3/Bax complex via JNK-dependent phosphorylation of 14-3-3 [20] constitutes a possible mechanism by which JNK promotes the accumulation of active Bax at the mitochondria. In addition, we demonstrate that JNK controls the expression of Bax. A previous study has shown that increased AP-1 activity by JNK induces the transcription of the bax gene in response to butyric acid [21]. However, real time PCR analysis indicates that Cd does not regulate the levels of bax mRNA (data not shown). Consistent with the ability of JNK to control the apoptotic response of MEFs to UV without requiring de novo gene expression [17], Cd induces Bax expression via a post-transcriptional mechanism that includes the stabilization of the protein by JNK.
In contrast to Bax, increased Bid expression following stress appears to be inhibited by JNK. This may be a consequence of a decrease in the efficiency of Bid cleavage in cells lacking JNK. A recent study showed that the activation of JNK by TNFα resulted in the generation of a distinct Bid cleavage product termed jBid (21 kDa) [14]. jBid promotes the specific release of Smac, thereby neutralizing inhibitors of apoptosis proteins and leading to increased caspase 8 activity. The indirect activation of caspase 8 by JNK via jBid provides an elegant mechanism to explain how JNK contributes to the efficient cleavage of Bid. Whether the 20 kDa Bid cleavage product identified in wild type MEFs corresponds to jBid remains to be confirmed.
Previous studies have shown that phosphorylation of Bim by JNK promotes its dissociation from the microtubule-associated dynein motor complex, thereby enhancing its pro-apoptotic activity [15, 16]. Consistent with these findings we identified Bim as a downstream target of JNK. However the requirement of JNK can be bypassed by increasing the levels of stress. p38 MAPK which activity is upregulated by Cd (data not shown) may be the protein kinase responsible for Bim phosphorylation in the absence of JNK.
Altogether these studies provide a molecular link between JNK and the initiation of the mitochondrial apoptotic pathway (Fig. 7 ). The data support the hypothesis that Bax-dependent release of cytochrome c via JNK is a pre-determining step in the apoptotic process. The defect in Bax regulation associated with homozygous deletion of jnk genes prevents the formation of specific channels in the OMM [6]. Consequently, the release of cytochrome c in the jnk−/− MEFs results from the rupture of the OMM caused by the sustained opening of the PTP and subsequently mitochondrial matrix swelling, a mechanism implicated in necrotic cell death induced by high concentrations of Cd [7].
Acknowledgments
We are indebted to Roger Davis, Richard Flavell, and Stanley Korsmeyer for kindly providing the jnk−/− and bax−/− fibroblasts. We thank Andrew Gilmore and Alan Whitmarsh for critically reviewing the manuscript. We are grateful to Anthony Valentijn for helpful technical advice. This work was supported in part by the AICR and the Royal Society, and principally by the MRC and a Lister Institute of Preventive Medicine Research Fellowship to C.T. E.P. was supported by the MRC and Merck Sharp and Dhome.
