Research Article. Key words: DISC, c-flip, JNK, Mitochondria, Caspase

Transcription

1 Research Article 6459 p38α, but not p38β, inhibits the phosphorylation and presence of c-flip S in DISC to potentiate Fas-mediated caspase-8 activation and type I apoptotic signaling Leon Tourian Jr, Hong Zhao and Coimbatore B. Srikant* Fraser Laboratories, Department of Medicine, McGill University Health Centre and Royal Victoria Hospital, Montreal, Quebec, H3A 1A1, Canada *Author for correspondence ( coimbatore.srikant@mcgill.ca) Accepted 30 September 2004 Journal of Cell Science 117, Published by The Company of Biologists 2004 doi: /jcs Summary Pharmacological inhibitors of JNK (SP600125) and p38 (PD169316) sensitize tumor cells to Fas-mediated apoptosis. PD is less potent than SP and diminishes its effect when present together. Because the p38 isoforms that promote (p38α) or inhibit (p38β) apoptosis are both suppressed by PD169316, we investigated their regulatory involvement in Fas-signaling. We report here, that p38α, but not p38β, exerts its proapoptotic effect by inhibiting the phosphorylation and presence of c-flip S, but not c-flip L, in the DISC to promote caspase-8 activation and type I signaling in Fas-activated Jurkat cells. Its effect was enhanced by enforced expression of Flagtagged p38α and was attenuated by its inactive mutant (p38α-agf) or by translational silencing. By contrast, type II signaling was facilitated by p38α-dependent mitochondrial presence of tbid and inhibition of Bcl-2 (Ser70) phosphorylation as well as by p38α/β-dependent mitochondrial localization of Bax and inhibition of phosphorylation of Bad (Ser112/Ser155). Potentiation of Fas-mediated apoptosis by the inhibition of JNK1/2 correlated with the loss of Bad (Ser136) phosphorylation and was dependent on the stimulatory effect of p38α on DISC and the downstream effects of both p38α and p38β. These data underscore the need to reassess the findings obtained with pan-p38 inhibitors and suggest that activation of p38α coupled with targeted inhibition of p38β and JNK1/2 should optimally sensitize tumor cells to Fasmediated apoptosis. Key words: DISC, c-flip, JNK, Mitochondria, Caspase Introduction Fas (Apo-1/CD95) is an important death receptor that belongs to the tumor necrosis factor receptor family. Binding of Fas to its ligand FasL triggers the sequential activation of cysteineaspartate-specific proteases (caspases) that play an essential role in apoptosis. The autoproteolytic generation of caspase-8 from procaspase-8 is the initial event which then induces downstream caspases directly (type I signaling) or via the mitochondrial pathway leading to the intermediate activation of caspase-9 (type II signaling) (Fulda et al., 2001; Hengartner, 2000). Fas interacts through the death domain in its intracellular segment with death-effector-domain-containing proteins such as FADD/MORT1 (Boldin et al., 1996; Chinnaiyan et al., 1995) and procaspase-8 (FLICE, MACH, Mch5) (Kischkel et al., 1995; Medema et al., 1997) to form the death-inducing signal complex (DISC) both in the native and activated states (Mundle and Raza, 2002; Muzio et al., 1996; Siegel et al., 2000; Srinivasula et al., 1996). Autoproteolytic activation of procasapse-8 occurs within DISC in the presence of ligated Fas, a process that is inhibited by the cellular FLICE-inhibitory protein (c-flip). Expressed in long (c-flip L ) and short (c- FLIP S ) variants, c-flip contains two death effector domains and displays homology to the N-terminal region of procaspase- 8 (Irmler et al., 1997; Shu et al., 1997). A caspase-like domain that lacks the catalytic activity is also present in c-flip L. Recruitment of procasapase-8 to Fas-DISC and its autoproteolysis is inhibited by c-flip S. By contrast, c-flip L does not completely exclude procaspase-8 from DISC and permits its partial cleavage into a p43 form (Krueger et al., 2001). Proteins of the Bcl-2 family act distal to caspase-8 to regulate apoptotic signals at the mitochondria. The antiapoptotic proteins Bcl-2 and Bcl-X L contain four Bcl-2 homology (BH) domains and a C-terminal membraneanchoring domain. Homologues of Bcl-2 that possess up to three BH domains but lack the membrane-anchoring domain are proapoptotic. Of these, Bax and tbid (a cleavage product of Bid) migrate to the mitochondria in response to proapoptotic stimuli (Eskes et al., 2000; Wei et al., 2001; Wolter et al., 1997). The regulated presence and interactions of Bcl-2 proteins at the mitochondria precedes the disruption of its integrity and the release of cytochrome c (cyt c) into the cytoplasm leading to the activation of caspase-9 (Scaffidi et al., 1998; Stennicke et al., 1998). For instance, tbid facilitates proapoptotic activity by inducing Bax-Bak oligomerization and

2 6460 Journal of Cell Science 117 (26) by sequestering Bcl-2 and Bcl-X L (Cheng et al., 2001). Bad, whose function is governed by serine phosphorylation (residues 112, 136 and 155), acts by inhibiting Bcl-2 and Bcl-X L (Gross et al., 1999; Virdee et al., 2000; Zha et al., 1996). Stress-activated protein kinases p38 and JNK are induced during apoptosis (Kyriakis and Avruch, 1996) but their roles in regulating cell death have remained controversial. Apoptosis is facilitated by p38 and JNK (Aoshiba et al., 1999; Bae and Song, 2003; Vivo et al., 2003; Yoshino et al., 2001) in some models but inhibited in others (Assefa et al., 1999; Nemoto et al., 1998; Porras et al., 2004; Zhang et al., 2002). These kinases have been reported to influence the expression Fas and/or FasL (Hsu et al., 1999; Ivanov and Ronai, 2000; Zhang et al., 2000), interaction of FADD with procaspase-8 or c-flip (Park et al., 2001), mitochondrial proapoptotic activity (Holmstrom and Eriksson, 2000; Tournier et al., 2000; Yoshino et al., 2001) and phosphorylation of Bcl-2 congeners (Deng et al., 2000; Deng et al., 2001; Haldar et al., 1994; Ito et al., 1997). In the present study, we observed that the JNK inhibitor SP potentiated Fas-mediated apoptosis to a greater extent than the p38 inhibitor PD169316, distal to caspase-8 in the absence but not in presence of the latter in several tumor cells. Given that PD inhibits p38 isoforms that promote (p38α) or suppress (p38β) apoptosis (Jiang et al., 1996; Kaiser et al., 2004; Nemoto et al., 1998) we analyzed their involvement in regulating Fas-mediated apoptosis in Jurkat T lymphocytes. Here, we demonstrate that p38α, but not p38β, facilitates Fas-mediated activation of caspase-8 by inhibiting the phosphorylation and presence of c-flip S, but not c-flip L, in the DISC. Inhibition of Fas-mediated caspase-8 activation in p38α-lacking cells correlated with the expected inhibition of tbid formation, mitochondrial presence of tbid and Bax and dephosphorylation of Bad at Ser112 and Ser155. By contrast, only mitochondrial presence of Bax and Bad dephosphorylation were suppressed in cells lacking p38β. Inhibition of JNK prevented the phosphorylation of Bad at Ser136 and enhanced type II signaling but only in the presence of p38α-dependent caspase-8 activation and downstream effects of p38α and p38β. Materials and Methods Materials Specialty reagents and antibodies were obtained from commercial sources as indicated. Polyclonal or monoclonal antibodies against p38, p38α, JNK, caspase-9, Bid, PARP, FADD, phospho-ser70 Bcl-2, Bad, site-specific phospho-bad (pser112, pser136 and pser155) and phospho-p38 (pthr180/p-tyr182, mab28b10), phospho-jnk (pthr183/ptyr185, mabg9), caspase-8 (1C12), cleaved caspase-3 (5A1) and MAPK assay kits (Cell Signaling, Beverly, MA); agonistic anti-fas antibody (Fas-mAb, 2R2, monoclonal), annexin-v-fluos apoptosis detection kit and the complete protease inhibitor cocktail tablets (Roche Diagnostics, Montreal, QC); antibodies against c-flip (C-19) c-flip S/L (H-202), Fas (non-agonistic, B10), Bax (pabp19) and Protein A/G PLUS-Agarose (Santa-Cruz Biotechnology, Santa- Cruz, CA); human recombinant Fas ligand plus, anti-bcl-2 (Ab1) and anti-cyt c (Ab2) (Oncogene, San Diego, CA); p38 inhibitors PD169316, SB203580, SB and the nonfunctional derivative SB and the JNK inhibitor SP (Calbiochem, San Diego, CA); potential sensitive mitochondrial dye DiOC 6 (3), Sypro Ruby protein gel stain and ProQ Diamond phosphoprotein gel stain (Molecular Probes, Eugene, OR); aminomethylcoumarin (AMC) derivatives of caspase substrates IETD, DEVD and LEHD and caspase inhibitors IETD-CHO and DEVD-CHO (BIOMOL, Plymouth Meeting, PA); polyclonal anti-p38β antibody (Zymed, San Francisco, CA); TransIT Jurkat small interfering RNA (sirna) transfection reagent (Mirus, Madison, WC). Double-stranded sirnas that selectively target p38α (5 -GGUCUCUGGAGGAAUUCAAtt-3, 3 - ttccagagaccuccuuaaguu-5 ), p38β (5 -GGACUUCAGCG- AAGUGUACtt-3, 3 -ctccugaagucgcuucacaug-5 ), JNK1 (5 -GGAGCUCAAGGAAUAGUAUtt-3, 3 -gtccucgaguucc- UUAUCAUA-5 ) or JNK2 (5 -GGGAUUGUUUGUGCUGCAUtt-3, 3 -ttcccuaacaaacacgacgua-5 ) as well as a double stranded negative control sirna were purchased from Ambion (Austin, TX). Plasmids encoding the Flag-tagged wild-type p38α (pcmv-flag-p38α and the inactive mutant p38α-agf (pcmv-flag-p38α-agf) were generously provided by Roger J. Davis (University of Massachusetts Medical School, Worcester, MA). Cell culture and transfection Jurkat cells were grown in RPMI 1640 medium (Invitrogen Canada, Burlington, ON) supplemented with 10% fetal bovine serum, 1% antibiotic-antimycotic, 2 mg/ml glucose, 10 mm HEPES, 1mM sodium pyruvate and 2 mm L-Glutamine at 37 C in a humidified atmosphere containing 5% CO 2. To achieve targeted ablation of individual kinase isoforms, cells were transfected with annealed dssirna oligonucleotides. Briefly, 3 µl of TransIT reagent was added to 200 µl of serum-free medium maintained at room temperature, mixed with ds-sirna (60 pmol) and incubated for 30 minutes before adding it to the cells. The efficiency of sirna-mediated translational silencing of each kinase was monitored by immunoblot analysis as a function of time with maximal suppression being observed at 36 to 48 hours (details not shown). To generate Jurkat T cells that transiently overexpress p38α and its dominant negative variant p38-α-agf, we transfected the cells with the plasmids pcmv-flag-p38-α and pcmv-flag-p38-α-agf. Experiments were carried out 48 hours after transfection when maximal expression of the transfected proteins was observed. Detection of apoptosis Cells were incubated with 150 ng/ml of Fas-mAb and were pretreated or not with PD and/or SP for 30 minutes. In parallel experiments, we assessed the effect of these inhibitors on apoptosis induced by FasL plus (0.1 µg/ml). Cells were washed and labeled with annexin-v-fitc and propidium iodide (PI) using the annexin-v- FLUOS staining kit according to the manufacturer s instructions. Apoptotic cells were detected by flow cytometry by annexin-v-fitc labeling in the absence of PI uptake. At least 20,000-gated events were counted for each sample and analyzed by WinList software (Verity Software House, ME). Measurement of mitochondrial membrane potential ( Ψ m ) The potential sensitive dye DiOC 6 (3) (10 µm) was added to the cells 30 minutes before the treatment with Fas-mAb was finished. Cells were washed with PBS, excited at 360 nm and the fluorescence emission was captured at 560 nm in a flow cytometer (Liu et al., 2000). Caspase activity measurements Activities of caspase-3, caspase-8 and caspase-9 in cell extracts were measured using AMC derivatives of DEVD, IETD and LEHD respectively (Liu et al., 2000). The generation of the fluorogenic product was monitored by exciting the samples at 380 nm and measuring the emission at 460 nm, and quantified against a standard curve generated by using nm AMC. To minimize cross reactivity, inhibitors of caspase-3 (DEVD-CHO) or caspase-8 (IETD- CHO) were included at a concentration of 1.5 µm when measuring the activities of caspase-8 and caspase-3, respectively.

3 Regulation of Fas-induced apoptosis by p38α and p38β 6461 Isolation of mitochondria Cells were washed with PBS, lysed on ice for 10 minutes with the cytosol extraction buffer (4.3 mm Na 2 HPO 4, 1.4 mm KH 2 PO 4, 250 mm sucrose, 70 mm KCl, 137 mm NaCl, 200 µg/ml digitonin and protease inhibitors) and centrifuged at 3000 g for 5 minutes to obtain the cytosolic fraction. The pellet was washed thrice with cytosol extraction buffer and lysed with 50 mm Tris-HCl buffer (ph 7.5) containing 150 mm NaCl, 2 mm EDTA, 2 mm EGTA, 0.2% v/v Triton X-100, 0.3% NP40, 0.5% sodium deoxycholate and protease inhibitors), incubated on ice for 15 minutes and the mitochondrial fraction was recovered in the supernatant following centrifugation at 10,500 g for 10 minutes at 4 C. Immunoblot analysis Cells were lysed in 50 mm Tris-HCl buffer (ph 8.0) containing 1% NP40, 150 mm NaCl, 1 mm EDTA 0.05% SDS and protease inhibitors. Aliquots of lysates containing 40 µg of protein were subjected to immunoblot analysis with antibodies as indicated and the appropriate horseradish peroxidase-conjugated secondary antibodies, and the signals were captured in an Alpha Innotech Imager (San Leandro, CA). were detected by in vitro phosphorylation of ATF-2 (by p38) and JUN (by JNK). Results Fas-signaled apoptosis in Jurkat cells activates p38 and JNK upstream of caspase-8 Jurkat cells treated with Fas-mAb underwent apoptosis in a time-dependent manner, evidenced by the progressive increase of the apoptosis marker annexin-v from 11.4±0.6% at 4 hours to >90% at 24 hours (Fig. 1A). Fas-ligation activated both p38 and JNK in a time-dependent manner, evidenced by the increase in phosphorylated (Fig. 1B, blots 1 and 3), but not total, proteins (blots 2 and 4). The increase in phosphorylation reflected in the increase in the ability of p38 to phosphorylate ATF-2 (Fig. 1C, blot 1) and of JNK to phosphorylate JUN (Fig. 1C, blot 2) in vitro. Both kinases were induced upstream of caspase-8 activation because presence of cleaved caspase-8 fragments p44/43 and p16 was detectable only beyond 1 hour (Fig. 1D). Immunoprecipitation of Fas-associated proteins Cells were lysed in 10 mm Tris-HCl buffer (ph 7.5) containing 1% v/v Nonidet P-40, 150 mm NaCl and 0.4 mm EDTA and protease inhibitors as described previously (Daigle et al., 2002) with slight modifications. Aliquots of lysates containing 1 mg total protein were pre-cleared using normal mouse IgG and incubated with 2 µg of the non-agonistic anti-fas antibody in presence of 80 µl of Protein A/G Sepharose beads. To minimize the interference of heavy and light chains of primary antibody during the IP, we employed Seize Classic A&G Immunoprecipitation Kit from Pierce (Rockland, IL). Briefly, cells lysates were incubated with either anti-c-flip L or anti-c-flip S (Santa Cruz) overnight after which immobilized Protein A suspension, containing a cross linking agent (DSS), was added. Immune complexes were washed, electrophoresed and immunoblotted for the detection of Fas, FADD, c-flip L/S and procaspase-8. For detecting c-flip phosphorylation, c-flip was immunoprecipitated, electrophoresed, probed with ProQ Diamond phosphoprotein gel stain and scanned at 590 nm in a Typhoon 9200 Imager (Amersham Biosciences). Metabolic labeling and detection of 32 P labeled c-flip L and c-flip S Endogenous phosphate was first depleted from cells by incubation in phosphate-free minimum essential medium (MEM) for 48 hours. The cells were then exposed to fresh medium supplemented with [ 32 P]sodium o-phosphate (37.5 µci/well) for 24 hours. The cells were washed and incubated for 4 hours in phosphate-free medium before lysis. Cell lysates containing 1 mg protein aliquots in 0.5 ml lysis buffer was pre-cleared by incubation with 30 ml of Sepharose-CL 4B for 3 hours at 4 C. Labeled c-flip L and c-flip S were then immunoprecipitated with anti-c-flip S/L antibody (10 µg) using the Seize Classic A&G Immunoprecipitation kit. Immune complexes were pelleted by centrifugation for 5 minutes at 14,000 g, washed five times with the lysis buffer and resolved by SDS-PAGE. The gels were dried and the radioactive bands were visualized using the Cyclone Storage Phosphor System (Perkin Elmer, Torrance, CA). In vitro kinase activity assays Aliquots of cell extracts containing equal amounts of protein were immunoprecipitated with phospho-p38 or phospho-jnk antibodies. The activities of the immunoprecipitated phosphorylated-kinases Potentiation of Fas mediated cell death by pharmacological inhibition of p38 and JNK We first established the specificity of the inhibitors of p38 and JNK by demonstrating that PD inhibited the ability of p38 in Fas-activated cells to phosphorylate ATF-2, whereas SP selectively inhibited JNK-mediated phosphorylation of JUN (Fig. 2, panels A and B, compare lanes 3 and 4 with lane 2). Both kinases were inhibited in cells that had been simultaneously exposed to the two inhibitors (Fig. 2, lane 5). Their activities in native Jurkat cells were below the detection limit under these conditions (Fig. 2, lane 1). The concentration of the inhibitors (15 µm) used in this experiment was determined to be the lowest concentration that effectively inhibited Fas-mediated activation of the respective kinases for up to 24 hours (details not shown). The sensitizing effect of these inhibitors on Fas-signaled apoptosis was best observed following a 4-hour treatment with Fas-mAb. A 2.5- and 4-fold increase in apoptosis occurred in the presence of PD and SP (26.9±0.5 and 45.3±0.7%, respectively) compared with 11.4±0.6% observed in their absence (Fig. 3A, left panel, lanes 2-4). When added together, PD reversed the greater potentiating effect SP (lane 5). Following treatment with Fas-mAb, a significant increase in the number of cells with reduced ψ m was observed in the presence of SP (58±4 and 37±2% for presence and absence, respectively) (Fig. 3B). PD inhibited SP induced increase in the number of cells with decreased ψ m, (28±5%) but not the effect of Fas-mAb itself. On its own, neither inhibitor induced apoptosis or decreased ψ m (Fig. 3A,B, right panels). Apoptosis and the reduction in ψ m induced by ligating Fas with its natural ligand FasL were also sensitized to a greater extent by SP in the absence, but not presence, of PD (Fig. 3C,D). The effects of inhibition of p38 on Fas-signaled apoptosis and its sensitization by SP were confirmed with two other pharmacological inhibitors of p38 SB and SB and by the lack of effect of the nonfunctional derivative SB (Fig. 3, panels E and F). The ability of SP and PD to potentiate Fas-mediated apoptosis

4 6462 Journal of Cell Science 117 (26) Fig. 2. Specificity of the p38 and JNK inhibitors. (A) Cells were incubated with Fas-mAb for 1 hour in the absence (lane 2) or presence of PD (lane 3), SP (lane 4) or both (lane 5). Fas-induced increase in p38 activity was inhibited by PD but not SP (B) Phosphorylation of JUN by JNK was inhibited by SP but was unaffected by PD Inhibition of both kinases was observed only in cells treated with both inhibitors. Fig. 1. Phosphorylation and activation of p38 and JNK precedes caspase-8 induction in Fas-mediated apoptosis in Jurkat cells. Following incubation with Fas-mAb, cells were either stained with annexin-v to detect apoptosis or were lysed and subjected to immunoblot analysis. (A) Fas-induced apoptosis was detectable by 4 hours, increased with time and was maximal by 24 hours. (B) Timedependent increase occurred in the phosphorylation of p38 (blot 1) and JNK (blot 3) in the absence of any change in the total levels of p38 (blot 2) and JNK (blot 4). (C) The in vitro activity of each kinase was assessed following immunoprecipitation with the respective phosphospecific antibodies. Increased phosphorylation of p38 correlated with its ability to phosphorylate ATF-2 in vitro (blot 1) and of JNK with its ability to phosphorylate JUN (blot 2). (D) The formation of significant amounts of cleaved fragments of caspase-8 (p44/43) and p16 was detectable only beyond 2 hours. was not unique to Jurkat cells and was also observed in other tumor cell types such as HeLa, MCF-7 and T47D (Table 1). Sensitizing effect of SP is abrogated by PD distal to caspase-8 Fas-mediated activation of caspase-8 was greater in cells treated with SP than with PD (0.51±0.02 and 0.30±0.01 nmol/µg/minute) compared with 0.19±0.006 nmol/µg/minute without either inhibitor (Fig. 4A). PD did not, however, suppress the effect of SP on caspase-8 activation. By contrast, caspase-9 activity was lower in cells incubated with PD than without, but was higher in cells treated with SP (0.30±0.003 vs 0.17± ±0.006 nmol/µg/ minute) (Fig. 4B). In the presence of both inhibitors, Fasinduced caspase-9 activation was comparable to that seen with PD alone (0.17±0.04 vs 0.24±0.04 nmol/µg/minute). Likewise, caspase-3 activity was greater in the presence of SP than PD (0.53±0.007 and 0.40±0.02, respectively compared with 0.19±0.005 nmol/µg/min in their absence, Fig. 4C). The observed differences in the activities of these caspases correlated with the increase in the amount of cleaved fragments of procaspase-8 in cells treated with these inhibitors either individually or together (Fig. 4D, blot 1), and of cleaved fragments of caspase-9 and caspase -3 in cells treated with SP in the absence, but not presence of PD (Fig. 4D, blots 3 and 5); they correlated with the generation of tbid, the cleavage product of the caspase-8 substrate Bid, cytosolic accumulation of cyt c and the cleaved fragment of PARP (Fig. 4D, blots 2, 4 and 6, respectively). As shown in these blots, the potentiating effect of SP on the activation of caspase-9 and caspase-3 in Fas-ligated cells was abrogated by PD (Fig. 4D, compare lanes 4 and 5) to the level seen in the presence of PD alone (Fig. 4D, lane 3). Both p38α and p38β isoforms are activated in Fasligated cells Because PD inhibits both α and β isoforms of p38, we confirmed that both these isoforms are activated in response to Fas-ligation in Jurkat cells. Total p38 was immunoprecipitated with an anti-p38 antibody that recognized both isoforms and immunoblotted with antibodies specific for p38α and p38β (Fig. 5A, blots 2 and 3). Fas-activation did not alter the levels of either p38α or p38β. The increase in their phosphorylation in Fas-activated cells was confirmed by immunoprecipitation with specific antibodies against p38α and p38β and subsequent immunoblot analysis with phospho-p38 antibody (Fig. 5B, blots 1 and 2). Effect of translational silencing of p38α and p38β on Fas-mediated apoptosis To elucidate the regulation of apoptotic signaling by these p38

5 Regulation of Fas-induced apoptosis by p38α and p38β 6463 Fig. 3. Effect of the inhibition of p38 and JNK on Fas-mediated apoptosis. Jurkat cells were incubated with Fas-mAb (150 ng/ml) in the absence and presence (15 µm) of SP and/or p38 inhibitors for 4 hours. (A) Fas-signaled apoptosis (lane 2) was enhanced 2.8-fold by PD (lane 3) and 4-fold by SP (lane 4). (B) The reduction in mitochondrial membrane potential ( ψ m ) in cells undergoing Fasmediated apoptosis (lane 2) was enhanced by SP (lane 4) but not PD (lane 3). PD suppressed the sensitizing effect of SP (lane 5). These inhibitors by themselves (individually or together) did not trigger apoptosis or decrease of ψ m (lanes 6-8). The sensitizing effect of SP on Fas-induced apoptosis (panel C) and the reduction in ψ m (panel D) was abrogated by two other p38 inhibitors SB and SB but not by the inactive derivative SB (mean±s.e.m., n=12). isoforms, we inhibited their expression by using sirna. Expression of p38α, but not that of p38β, was abrogated in sirna-p38α (si-p38α) cells (Fig. 5C, compare lane 2 in blots 1 and 2) whereas a significant inhibition of the expression of p38β, but not p38α, was seen in sirna-p38β (si-p38β) cells (lane 4). As shown in these blots, transfection of a control sirna (si-c) did not affect the expression of both isoforms of p38 (lane 3). In parallel, we generated cells lacking JNK isoforms 1 and 2 [sirna-jnk1/2 (si-jnk1/2)] in which the expression of both JNK1 and JNK2 was suppressed (Fig. 5D). Fas-mediated apoptosis was substantially inhibited in sip38α cells and enhanced in si-p38β and si-jnk1/2 cells (5.4±0.4%, 41.3±1% and 46±2.8%, respectively, compared Table 1. Sensitization of Fas-mediated apoptosis by SP is attenuated by PD Cell line Apoptotic cells (%) Fas-mAb PD SP Jurkat 5.4± ± ± ± ±1.1 HeLa 1.0± ± ± ± ±0.9 MCF-7 3.1± ± ± ± ±1.1 T47D 4.0± ± ± ± ±0.5 The number of apoptotic cells was quantitated as annexin-v positive, PInegative cells (mean±s.e.m. from three separate experiments in triplicate, n=9). Jurkat cells were analyzed after 4-hour treatment. HeLa, MCF-7 and T47D cells were analyzed after 24-hour treatment. with 12±1.2% in si-c cells, Fig. 6A). The reduction in Ψ m in response to Fas-activation was also much lower in si-p38α cells (11.3±0.6%), but not in si-p38β cells (36±0.9%), compared with the value of 33±2% in si-c cells (Fig. 6B). Interestingly, the reduction in Ψ m was much greater in si- JNK1/2 cells (74±1.4%). Fas-induced generation of caspases- 8, -9 and -3 was diminished in si-p38α cells but was enhanced in si-p38β cells (Fig. 6C, blots 1, 2 and 3). Abrogation of expression of p38α, but not p38β, inhibited cyt c release (blot 4). In si-jnk1/2 cells, enhanced presence of cleaved caspases- 8, -9 and -3 and cytosolic cyt c was observed (Fig. 6D). Ectopically introduced p38α and its inactive mutant p38α-agf exert opposing effects on Fas-signaling The selective requirement of p38α for Fas-mediated apoptosis was confirmed by comparing the effects of ectopically introduced HA-tagged wild-type p38α and its dominant negative mutant p38α-agf. The expression of the Flag-tagged proteins was verified by immunoblot analysis using anti-p38α antibody (Fig. 7, lanes 1-3) and anti-flag antibody (lanes 4-6). Four-hour treatment with agonistic anti-fas antibody induced greater apoptosis in cells expressing Flag-p38α compared with native Jurkat cells (31±1.3 vs 11.4±0.6%) and reduction in ψ m (64±2.1 vs 37±2%, Fig. 7C). By contrast, Fas-ligation failed to induce apoptosis and reduction in ψ m in cells expressing FLAG-p38α-AGF. The basal level of apoptosis in Jurkat cells was unaffected by the ectopically expressed wild type p38α or its mutant p38α-agf.

6 6464 Journal of Cell Science 117 (26) Effect of p38α and p38β on c-flip S phosphorylation and DISC assembly Depletion of p38α, but not p38β or JNK1/2, diminished Fasmediated caspase-8 activation. Hence, we compared the effects of chemical inhibitors and sirna-mediated knock down of these kinases on DISC assembly in the native and Fas-activated Jurkat cells. An increase in Fas-associated FADD and procaspase-8 and a concomitant decrease in the presence of c- FLIP S was evident in the native cells treated with PD and/or SP (Fig. 8A, top left panel). These changes became more pronounced in cells incubated with Fas-mAb for 1 hour (when there was no detectable formation of active caspase-8 fragments) in the presence of PD and SP (top right panel). Strikingly, no change in the level of DISC-associated c-flip L was seen in presence of either inhibitor alone or both inhibitors together. In parallel experiments, we observed that a marked reduction in the presence of c-flip S and a concomitant increase in FADD and procaspase-8 was seen in si-p38β cells (Fig. 8B, lane 3, top panel) compared with native Jurkat cells (lane 1) and si-p38α cells (lane 2). Increased presence of FADD and procaspase-8 was also observed in si-jnk1/2 cells (lane 4). Treatment of the cells with PD and SP or silencing of these kinases did not alter the levels of Fas (Fig. 8A and B, top panels), or of total FADD, procaspase-8, c-flip L and c-flip S (bottom panels). Phosphorylation of c-flip S, but not c-flip L, is increased in the cells lacking p38α To determine whether the striking difference in the presence of c-flip S in the DISC is because of an altered phosphorylation status in Jurkat cells lacking p38α, we immunoprecipitated c-flip S, electrophoresed and probed the gel with Pro-Q Diamond stain for the detection of phosphorylated proteins. Phosphorylation of c-flip S was found to be greater in si-p38α than in si-c cells (Fig. 9A, bottom panel, compare lanes 1 and 2). By contrast, phosphorylation of c-flip S was undetectable in si-p38β or si-jnk1/2 cells (lanes 3 and 4). The levels of immunoprecipitated c-flip S in these cells were comparable as demonstrated by Sypro Ruby protein staining (Fig. 9A, bottom panel). These findings were further confirmed by monitoring the phosphorylation of c-flip in Jurkat cells metabolically labeled with radioactive 32 P. The increase in phosphorylation of c-flip S seen in cells transfected with sip38α was higher than that in cells transiently expressing Flagp38α-AGF (Fig. 9B, compare lanes 3 and 4). Enforced expression of p38α suppressed c-flip S phosphorylation (lane 5) to the level seen in the native and in control sirna Fig. 4. Effect of p38 and JNK inhibitors on Fas-mediated activation of caspases. (A) Fas-mediated activation of caspase-8 (lane 2) was increased to a greater extent in cells treated with SP in the absence (lane 4) or presence of PD (lane 5) than that seen with the latter alone (lane 5). (B) Caspase-9 activation was potentiated by SP (lane 4). PD inhibited caspase-9 activation in Fas-activated cells both in the absence (lane 2) and in presence of SP (lane 5). (C) Caspase-3 activity was higher in Fas-activated cells in presence of SP (lane 4) than PD (lane 3). PD blunted the potentiating effect of SP (lane 5). (D) Immunoblot analysis demonstrating the effects of PD and SP on the formation of cleaved fragments of caspase-8 (blot 1), caspase-9 (blot 3) and caspase-3 (blot 5). Increased presence of cleaved caspase-9 and caspase-3 was detected in cells treated with SP (lane 4), but not PD alone (lane 3) or in the presence of the former (lane 5). The differential effects of these inhibitors occurring distal to caspase-8 were confirmed by the differences in cyt c release (blot 4) and PARP cleavage (blot 6) and the lack of a difference in the extent of cleavage of Bid into tbid (blot 2). (Data are representative of three experiments).

7 Regulation of Fas-induced apoptosis by p38α and p38β 6465 Fig. 5. Fas-mediated apoptosis is associated with the activation of p38α and β isoforms. (A) A pan anti-p38 antibody was used to immunoblot immunoprecipitated p38α and p38β, by using their respective specific antibodies. Fas-activation did not affect the cellular level of these kinases. (B) Fas induced phosphorylation of p38α and p38β was detected by immunoprecipitating phosphorylated p38 using phospho-specific anti-p-38 antibody and immunoblotting with antibodies specific for p38α (blot 1) or p38β (blot 2). (C) The expression of p38α is selectively abrogated in si-p38α cells (top panel) whereas p38β was suppressed only in si-p38β cells (bottom panel) (D) Expression of JNK1 and JNK2 was decreased in si-jnk1/2 cells. Fig. 6. Effect of translational silencing of p38α, p38β and JNK1/2 on Fas-mediated apoptotic signaling. (A) Dot-plot analysis demonstrates the potentiation of Fas-mediated apoptosis in si-p38β and si-jnk1/2 cells, and its suppression in si-p38α cells compared with that in si-c cells. Apoptotic cells are identified by an increase in annexin-v-fitc positive, PI negative cells (bottom right quadrants). (B) The percentage of cells with decreased Ψ m (region L) compared with that of cells with high or normal Ψ m (region H) was enhanced to a greater extent in si-jnk1/2 cells than in si-p38β cells, but was decreased in si-p38α cells. (C and D) Fas-induced cleavage of procaspase-8, procaspase-9 and procaspase-3 was evident in si-c cells and was inhibited in si-p38α cell but enhanced in si-p38β cells (compare lanes 3 and 4 with lane 2, blots 1,2,3, panel C) and in si-jnk cells (compare lanes 3 and 2, panel D). The changes in caspase-9 cleavage correlate with parallel differences in the cytosolic presence of cyt c.

8 6466 Journal of Cell Science 117 (26) transfected Jurkat cells (lanes 1 and 2). By contrast, such manipulation of p38α activity did not affect the phosphorylation status of c-flip L. Potentiating effect of SP on Fas-mediated apoptosis in sirna transfected cells In si-p38α cells, treatment with PD enhanced the apoptotic responsiveness to Fas-ligation fivefold (30.3±2.5 vs 5.4±0.9%) and 2.5-fold in si-c cells (30.5±1.1 vs 12±1.2%). By contrast, it decreased the extent of apoptosis by 50% in sip38β cells (22.9±1.5 vs 41.3±1.1%). Likewise, the proportion of cells with reduced ψ m was lower in si-p38α cells (11.3±0.6) but was higher in si-p38β cells (36±1%) compared with that in si-c cells (32.9±2%, Fig. 10B). Fas-mediated apoptosis was higher in presence of SP than in its absence in si-c cells (48.7±0.8% vs 12±1.2%) and si-p38α cells (18.5±0.8% vs 5.4±0.4%), but not si-p38β cells (45.3±1.1% vs 41.3 ±1%, Fig. 10A). Similarly, the reduction in ψ m was higher in the presence of SP than in its absence in si-c (68±1.9% vs 33±2.1%) and si-p38α (27.9±1.8% vs 11.3±0.6%), but not in si-p38β cells (31.2±2.8% vs 36±0.9%). A significantly higher activation of caspase-8 (0.53±0.02 vs 0.2±0.02 nmol/µg/min) and caspase-3 (0.55±0.03 vs 0.21±0.02 nmol/µg/min) but not caspase-9 (0.25±0.03 vs 0.34±0.03 nmol/µg/min) was seen in si-p38β cells compared with si-c cells, whereas these caspases were only minimally induced in si-p38α cells. Fas-mediated activation of caspase-8 was potentiated by PD in si-c and si-p38α, but not p38β, cells. The effect of PD on Fas-mediated activation of caspase-3 revealed a similar pattern, being higher in si-c and p38α cells and lower in si-p38β cells compared with the values observed following Fas-activation in the absence of this inhibitor. By contrast, caspase-9 was activated in si-c and sip38β, but not in si-p38α cells. Of particular interest was the finding that PD completely abrogated Fas-mediated activation of caspase-9 in si-p38β cells (from 0.25±0.04 to 0.06±0.01 nmol/µg/minute) but only partially in si-c cells (0.21±0.01 from 0.34±0.04 nmol/µg/minute). As expected, SP did not exert any effect in si-jnk cells. Moreover, PD inhibited the sensitizing effect of SP on Fasmediated activation of all three caspases in si-p38α cells and si-jnk1/2 cells but not in si-p38β cells. Fig. 7. Fas-induced apoptosis in Jurkat cells is potentiated by ectopically introduced Flag-tagged p38α and suppressed by the dominant negative variant p38α-agf. (A) Immunoblot analysis using anti-p38α antibody (lanes 1-3) and anti-flag antibody (lanes 4-6) demonstrate the expression of transfected Flag-p38α (lanes 2,5) and Flag-p38α-AGF (lanes 3,6). (B) Cells expressing Flag-p38α displayed threefold greater sensitivity to Fas-mediated apoptosis than the native Jurkat cells (30.6±1.1 vs 11.4±0.6% apoptotic cells, lanes 3 and 2). By contrast, Flag-p38α-AGF expression rendered the cells resistant to Fas-activation (lane 4). (C) The decrease in mitochondrial membrane potential ( ψ m ) in Jurkat cells (37±2.6%) was increased in cells expressing Flag-p38α (66±1.2%) but was suppressed by Flag-p38α-AGF (10.7±0.7%). The transfected proteins did not affect the basal levels of apoptosis or ψ m (lanes 1,5 and 6). Influence of the inhibition of p38α, p38β and JNK on Fas-induced alterations in the biochemical and cellular localization properties of Bcl-2 family proteins The differential regulatory effects of p38α, p38β and JNK1/2 on mitochondrial function in apoptosis prompted us to investigate their effects on the biochemical and subcellular localization properties of Bcl-2 family of proteins. Immunoblot analysis of the mitochondrial and cytosolic fractions revealed that PD169316, but not SP600125, prevented the mitochondrial localization of Bax. The ability of PD to inhibit the mitochondrial presence of Bax was not affected by SP (Fig. 11A, blots 1 and 2, compare lanes 2-5). The presence of tbid in the mitochondria was seen in Fas-ligated cells in the absence and presence of SP600125, but not PD (Fig. 11A, blot 3). However, PD did not inhibit the mitochondrial presence of tbid in presence of SP The presence of Bak in mitochondrial fraction was unaffected under these experimental conditions (Fig. 11A, blot 4). In Fas-activated cells, there was a loss of Bad phosphorylation at Ser112 and Ser155, but not at Ser136. Whereas PD reversed the loss of Bad phosphorylation at Ser112 and Ser155, SP promoted the loss of Ser136 phosphorylation (Fig. 11A, blots 5, 6 and 7, compare lanes 2,3 and 4). These changes in Bad phosphorylation occurred in the absence of any change in the total levels of Bad (Fig. 11A, blot 8). Ser70 phosphorylation of mitochondrial Bcl-2 was

9 Regulation of Fas-induced apoptosis by p38α and p38β 6467 Fig. 8. DISC composition is altered in p38 and JNK inhibited cells. (A) Lysates of native and Fas-activated cells in the absence and presence of PD and/or SP were immunoprecipitated with a nonagonistic anti-fas antibody and analyzed for the presence of Fasassociated proteins. The increased presence of FADD in the DISC and the concomitant decrease in the presence of c- FLIP S, but not c-flip L, is seen in PD treated native (lane 2) and Fas-activated cells (lane 6) (top panel). Similar changes were observed in presence of SP in the native (lane 3) and Fas-activated (lane 7) cells. These inhibitors did not exert additive effects either in the native (lane 4) or in Fas-activated (lane 8) cells. (B) Presence of c-flip S, but not c-flip L, in DISC is increased and induced reciprocal changes in DISC-associated FADD and procaspase-8 in si-p38α cells but not in si-p38β and si-jnk1/2 cells (top panel) in the absence any change in protein in the levels of the all the proteins tested (bottom panel). Data are representative of three separate experiments. inhibited in the presence of PD169316, but not SP (Fig. 11A, blot 9, compare lanes 2,3 and 4). These inhibitors did not affect the mitochondrial presence of Bcl-2 (blot 10). The efficacy of subcellular fractionation was confirmed by the detection of the mitochondrial marker TOM-20 exclusively in mitochondrial fractions (Fig. 11A, blot 11). Mitochondrial translocation of Bax failed to occur in Fasactivated si-p38α and si-p38β cells (Fig. 11B, blots 1 and 2, compare lanes 3, 4 with lane 2). However, the mitochondrial localization of Bax in si-jnk1/2 cells was higher than in si-c cells (Fig. 11B, blots 1 and 2, compare lanes 2 and 5). Cleavage of Bid into tbid (Fig. 11B, blot 3) and translocation of tbid to the mitochondria occurred in si-c, si-p38β and si-jnk1/2 cells, but not in si-p38α cells (Fig. 11B, blot 4). Fas-induced loss of phosphorylation of Bad at Ser 112 and Ser 155 was prevented in both si-p38α and si-p38β cells (Fig. 11B, blots 5 and 7, compare lanes 2, 3 and 4). Bad phosphorylation at Ser 136 was absent uniquely in si-jnk1/2 cells (Fig. 11B, blot 6, lane 5). Total Bad levels were comparable in these cells (Fig. 11B, blot 8). In the case of Bcl-2, Ser 70 phosphorylation was abrogated in si-p38α, but not si-p38β or si-jnk1/2 cells (Fig. 11B, blot 9) without affecting its presence at the mitochondria (Fig. 11B, blot 10). Discussion In this study, we demonstrated that the presence and activity in the DISC of c-flip S, but not c-flip L, is controlled by p38αdependent dephosphorylation. Inhibition of Fas-mediated caspase-8 activation in p38α-lacking cells correlated with the expected inhibition of tbid formation and its mitochondrial presence. Mitochondrial presence of Bax and the loss of Bad phosphorylation at Ser112 and/or Ser155 were regulated by both p38α and p38β and were abrogated in cells lacking either of these p38 isoforms. Enhanced mitochondrial proapoptotic activity in response to the inhibition of JNK1/2 correlated with the loss of Bad (Ser136) phosphorylation only when there was concomitant, p38α/β-dependent decrease in Ser112 and/or Ser155 phosphorylation. Collectively, these findings reveal that in Jurkat cells Fas-mediated caspase-8 activation and type I Fig. 9. p38α, but not p38β is essential for maintaining c-flip S in non-phosphorylated state. Phosphorylation of c-flip S was determined by Pro Q diamond phosphoprotein gel stain following immunoprecipitation with anti-c-flip antibody. Phosphorylation of c-flip S is enhanced in si-p38α, but inhibited in si-p38β or si- JNK1/2 cells compared with that seen in native Jurkat cells (top panel). No change in the total c-flip S levels was observed in any of the cells (bottom panel). (B) Jurkat cells expressing control (si-c) or p38α sirna (si-p38α) and flag-p38α or flag-p38α-agf were metabolically labeled with sodium ortho[ 32 P]phosphate. Phosphorylated c-flip was immunoprecipitated from the lysates with anti-c-flip antibody that recognizes both c-flip L and c-flip S, electrophoresed on SDS-polyacrylamide gels and subjected to autoradiographic analysis. Phosphorylation of c-flip S was higher in (si-p38α, lane 4) than in cells expressing the dominant negative mutant flag-p38α-agf (lane 3) whereas minimal phosphorylation was seen in native Jurkat (lane 1) si-c (lane 2) and Flag-p38α (lane 5) cells. By contrast, c-flip L phosphorylation was unaffected by p38α.

10 6468 Journal of Cell Science 117 (26) Fig. 10. Effect of PD on SP induced sensitization of Fas-mediated apoptosis in Jurkat cells lacking p38α, p38β and JNK1/2. (A) Fas-mediated apoptosis was sensitized by PD but inhibited by SP in si-38α cells but only by SP in si-p38β cells, SP600125, but not PD exerted the sensitizing effect. Both inhibitors (SP > PD169316) potentiated Fasmediated apoptosis in si-c cells. Fas-mediated apoptosis in si- JNK1/2 cells was comparable to that induced by SP in si-c cells and was inhibited by PD (B) The potentiating effect of SP on the reduction in ψ m in response to Fas-activation was abrogated in both si-p38α cells and si-p38β cells, but not in si-c cells. The reduction in ψ m in Fas-activated si-jnk1/2 cells was attenuated by. PD (C,D,E) Fas-mediated activation of the three caspases in si-p38α cells was observed in presence, but not in absence of PD and/or SP Caspases 8 and 3 were activated to a greater extent in si-p38β and si-jnk1/2 cells. Caspase- 9 activation was lesser in si-p38β cells and greater in si-jnk1/2 cells than that seen in si-c cells, and was suppressed by PD The sensitizing effect of SP on caspase-9 was markedly lower in si-p38α and si-p38β cells than in si-c cells. signaling depend on p38α-regulated inhibition of c-flip S phosphorylation and its exclusion from the DISC, whereas activation of type II signaling depends on the regulation of mitochondrial proapoptotic activity by both p38α and p38β. Caspase-8 is generated by the autoproteolytic cleavage of procaspase-8 within the DISC in the presence of ligandactivated Fas and the adapter protein FADD (Kischkel et al., 1995; Medema et al., 1997). c-flip S and c-flip L bind to FADD through its DED domains and thus block the recruitment and/or the autoproteolysis of procaspase-8 (Krueger et al., 2001). Here, we demonstrated that the phosphorylation and DISC localization of c-flip S but not c- FLIP L is attenuated by p38α, an effect that is enhanced by ectopically introduced Flag-tagged wild-type p38α and inhibited by the depletion of p38α by sirna-mediated translational silencing and by the dominant negative effect of Flag-tagged inactive mutant p38α-agf. This, to our knowledge, constitutes the first evidence for selective posttranslational modification of c-flip S by p38α-dependent phosphorylation. The identity of putative, p38α-regulated kinase(s) that act solely on c-flip S remains to be discovered. In previous reports, calcium-calmodulin-dependent kinase and PKC were identified as capable of phosphorylating c-flip. Calcium/calmodulin-dependent protein kinase II was shown to phosphorylate c-flip L but not c-flip S, increase its presence in the DISC and inhibit Fas-induced apoptosis in glioma cells (Yang et al., 2003). However, PKC-dependent increase in the phosphorylation and presence of both c-flip S and c-flip L in TRAIL-DISC was shown to promote TRAIL-mediated apoptosis in bile-acid-treated hepatocytes (Higuchi et al., 2003). The presence of multiple phosphorylation sites in the DED domains necessitates further studies aimed at the identification of the sites of phosphorylation that are sensitive to different kinases, and in particular the sites targeted in a p38α-dependent manner. The site-specific phosphorylation of the DED domain of c-flip variants may account for the differential regulation of DISC activation by different deathreceptor ligands. To date, regulation of c-flip S/L phosphorylation by phosphatase(s) has not been shown. The possibility that p38α protects c-flip S phosphorylation by inhibiting phosphatase(s) cannot be excluded and remains to be examined. Fas-mediated activation of caspase-8 was potentiated by SP in si-c cells and si-p38β cells but not si-p38α cells. However, enhanced activation of caspase-9 in the presence of SP was seen in Fas-activated si-c cells but not si-p38α cells or si-p38β cells. Moreover, PD diminished Fasmediated caspase-9 activation in si-jnk1/2 cells. Taken together, these findings suggest that p38α promotes caspase-3 activation via type I signaling and that p38α and p38β both promote type II signaling, whereas JNK1/2 selectively inhibits type II signaling. The greater reduction in Ψ m, cyt c release and caspase-9 activation induced by Fas-activation in JNKinhibited cells correlated with the loss of Ser136- phosphorylation in Bad and depended on p38α-p38β-regulated loss of phosphorylation at Ser112 and Ser155 and also mitochondrial localization of Bax and tbid. Fas-induced targeting of Bax and tbid and the loss of phosphorylation at Ser112 and Ser155 in Bad were prevented by inhibition of p38, irrespective of the presence or absence of JNK activity. The ability of SP to sensitize mitochondrial proapoptotic activity was seen in si-p38β cells, but not si-p38α cells. Moreover, the degree of sensitizing effect of JNK inhibition was significantly lower in si-p38β cells compared with that seen in si-jnk1/2 or si-c cells (Fig. 9). The stimulatory effect of p38α and the inhibitory effect of p38β on Fas-mediated caspase-8 activation largely account for the observed differences in the proapoptotic events distal to caspase-8. As expected, Fas-induced mitochondrial presence of tbid and Bax, and loss of phosphorylation of Ser70 in Bcl-2 and, the

11 Regulation of Fas-induced apoptosis by p38α and p38β 6469 Fig. 11. Effect of inhibition of p38 and JNK on cellular localization and/or biochemical properties Bcl-2 family of proteins in Jurkat cells undergoing Fas-mediated apoptosis. Aliquots containing equal amounts of protein from cytosolic (C) and mitochondrial (M) fractions or whole cell lysates (L) were subjected to SDS-PAGE and immunoblot analysis. (A) Fas-induced translocation of Bax from the cytosol to the mitochondria (compare blots 1 and 2) and tbid (blot 3) is inhibited by PD (lane 3) but enhanced by SP (lane 4). Mitochondrial targeting of Bax, but not tbid, was inhibited in presence of both inhibitors (lane 5). The presence of Bak in the mitochondria (blot 4) was unaffected under all experimental conditions. The loss of Bad phosphorylation at Ser112 (blot 5) and Ser155 (blot 7) in response to Fas-activation is inhibited by PD169316, but not SP By contrast, Ser136 phosphorylation was inhibited in presence, but not absence of, SP (blot 6). Total Bad remained unchanged (blot 8). In cells treated with both inhibitors there was a significant, but not complete reversal of Bad phosphorylation at Ser112, Ser136 and Ser155 (lane 5, blots 5, 6 and 7). Bcl-2 phosphorylation at Ser70 was decreased in Fas-activated cells in presence of PD169316, but not SP (blot 9) in the absence of a change in total Bcl-2 level (blot 10). The quality of mitochondrial preparations was established by immunoblot analysis for the marker protein Tom-20 (blot 11). (B) Mitochondrial presence of Bax in response to Fas-activation was inhibited in si-p38α and si-p38β, but not si-jnk1/2 cells (compare blots 1 and 2). Cleavage of Bid into tbid (blot 3) and the mitochondrial presence of tbid (blot 4) occurred in si- p38β and si-jnk1/2 cells, but not si-p38α cells. Fas-induced reduction in Bad Ser112 and Ser155 phosphorylation was inhibited in both si-38α and si-p38β cells (blots 5 and 7) whereas Ser136 phosphorylation was decreased uniquely in si-jnk1/2 cells (blot 6). No change in the total level of Bad was seen in these cells (blot 8). Bcl-2 Ser70 phosphorylation was inhibited in uniquely in si-p38α cells (blot 9) in the absence of any change in total Bcl-2 level (blot 10). Data are representative of three experiments. phosphorylation of Ser112 and Ser155 in Bad were not seen in si-p38α cells. Likewise, in si-p38β cells, which were more sensitive to Fas-signaling because of the presence and activation of p38α, we observed the expected presence of tbid at the mitochondria, reduction in Ψ m, cyt c release and, activation of caspase-9 and caspase-3. However, Bax was excluded from the mitochondria despite the presence of tbid at this organelle in si-p38β cells. The failure of Bax to localize to the mitochondria despite the presence of tbid in cells lacking p38β may, therefore, be owing to the lack of Bad dephosphorylation at Ser112 and Ser155 (Yusta et al., 2000). In its unphosphorylated form Bad can interact with Bcl-2 or Bcl-X L (Bae et al., 2001; Harada et al., 1999) and dephosphorylation of all three Ser moieties is essential for the ability of Bad to sequester Bcl-2 and Bcl-X L (Bae and Song, 2003; Datta et al., 2002) and to facilitate tbid-induced molecular interaction between the proapoptotic partners Bax and Bak (Wei et al., 2001). Ser155 phosphorylation is known to promote the dissociation of Bad from Bcl-X L (Klumpp and Krieglstein, 2002), suggesting that p38α and/or p38β may influence mitochondrial dysfunction by regulating the molecular interaction between pro- and anti-apoptotic Bcl-2 proteins. Direct regulation of Bad phosphorylation by p38α, p38β and JNK has not been shown, raising the possibility that it may be mediated via other kinases including p21-activated kinase, PKA, PKC, Raf-1, Rsk and Akt/PKB (Gnesutta et al., 2001; Jones et al., 2002; Schurmann et al., 2000; Wolf et al., 2001), and phosphatases such as PP1, PP2A and PP2B (Ayllon et al., 2000; Chiang et al., 2001). An additional mechanism may involve Bcl-2 dependence of Bax recruitment. Such a regulation has been reported to result from altered Bcl-2 phosphorylation (Ishikawa et al., 2003). Absence of Bax at the mitochondria correlated with reduced Bcl-2 (Ser70) phosphorylation in cells treated with PD and in si-p38α cells, but not si-p38β cells. Hence, p38α-dependent phosphorylation of Bcl-2 may be necessary to facilitate mitochondrial localization of Bax in cells undergoing apoptosis. Recently, p38 was shown to sequester in the mitochondria in apoptotic cells, raising the possibility that it directly regulates the disruption of this organelle (Tikhomirov and Carpenter, 2004). It remains to be determined, whether p38α, p38β or both localize to this organelle and affect its integrity through event(s) that are independent of their effects on events upstream of mitochondria. PD treatment enhanced Fas-mediated apoptosis more than fivefold from 5.4% to 30% in si-p38α cells and by 2.5- fold from 12% to 30.5% in si-c cells. By contrast, its effect was inhibitory in si-p38β and si-jnk1/2 cells. Caspase-8 activation was unaffected by PD in si-jnk1/2 cells, enhanced in si-c cells and si-p38α cells, and inhibited in sip38β cells. SP was able to potentiate caspase-8 activation and apoptotic responsiveness in si-p38β cells but not