Fas- or Ceramide-induced Apoptosis Is Mediated by a Rac1-regulated Activation of Jun N-terminal Kinase/p38 Kinases and GADD153*

In the present study, we show that Fas receptor ligation or cellular treatment with synthetic C6-ceramide results in activation or phosphorylation, respectively, of the small G-protein Rac1, Jun N-terminal kinase (JNK)/p38 kinases (p38-K), and the transcription factor GADD153. A signaling cascade from the Fas receptor via ceramide, Ras, Rac1, and JNK/p38-K to GADD153 is demonstrated employing transfection of transdominant inhibitory N17Ras, N17Rac1, c-Jun, or treatment with a specific p38-K inhibitor. The critical function of this signaling cascade is indicated by prevention of Fas- or C6-ceramide-induced apoptosis after inhibition of Ras, Rac1, or JNK/p38-K.

Programmed cell death or apoptosis has been identified as a conserved and fundamental active cellular mechanism occurring under a range of physiological and pathological conditions (1)(2)(3). Programmed cell death is characterized by distinct morphological changes of the cell, including nucleus condensation and fragmentation, membrane blebbing, or formation of apoptotic bodies (4). A variety of stimuli have been identified to induce apoptosis in different cell types, for example stimulation of cells via the TNF, 1 CD95/Fas/ApoI, or nerve growth factor receptor, treatment of cells with UV irradiation, reactive oxygen intermediates, heat shock, ceramides or cytotoxic drugs, infection by some viruses, or withdrawal of growth factors (5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17).
Apoptosis in mature lymphocytes can be induced by cellular stimulation via the Fas receptor (11,13) belonging to the family of the nerve growth factor/TNF receptors (5), which are important in the regulation of apoptosis, proliferation or differentiation (5,14). The major function of the Fas receptor/Fas ligand system seems to be the regulation of the peripheral immune response (18,19). Thus, mutations of the Fas receptor or its ligand result in the defects of lpr and gld mice, respectively, characterized by lymphadenopathy, lympho-accumulation, and autoimmune organ failure (20,21). Recently, mutations of the Fas receptor have been suggested as a mechanism for some human immunodeficiencies and the T-cell deficiency of human immunodeficiency virus might be due to a pathological stimulation via the Fas receptor (22)(23)(24). Several components of the death machinery have been identified; in particular, proteases of the ICE/Ced-3-family have been shown to be important in Fas triggered apoptosis (25,26). Recently, it was demonstrated that the Fas receptor associates via its death domain, which exhibits significant homology with an intracellular domain of the TNF receptor and is required for induction of apoptosis (27), with FADD, TRADD, or RIP (11,28,29). FADD in turn binds to FLICE/MACH-1 (30,31). FLICE/MACH-1 contains an ICE/Ced-3-like protease domain, which seems to be activated upon ligation and trimerization of the Fas receptor by a conformational change of FLICE/MACH-1 (30,31). FLICE/ MACH-1 then transmits the activation signal to ICE and CPP32, finally triggering cell death (30,31). Recent studies suggest that caspases regulate the release of ceramide (32), a molecule that has been shown by us (9) and others (7) to be released upon Fas receptor stimulation implying that sphingomyelinases are downstream targets of caspases. Ceramides are known stimuli of apoptosis and a function of Jun N-terminal kinase (JNK) activation in the apoptotic effect of ceramides has been recently demonstrated (16). Several other molecules have been shown to be involved in apoptosis, in particular the proto-oncogenes Myc, Abl, p120 GAP Fos, or Ras (9,(33)(34)(35)(36)(37)(38)(39)(40). For example, knock-out mice of p120 GAP show dramatic apoptosis of neurons in the developing brain (35), inhibition of Ras blocks Fas-or ceramide-induced programmed cell death in Jurkat T-lymphocytes (9), and expression of Myc triggers cell death in T-lymphocytes or fibroblasts (34,37). Further molecules activated upon Fas receptor triggering are neutral and acidic sphingomyelinases (41), phospholipase A 2 (41), NFB (42), Jun kinases (43)(44)(45), tyrosine and serine/threonine kinase (46,47), the tyrosine phosphatase FAP (48), and Rac proteins (49).
In the present study, we investigated downstream effector molecules of Ras and Rac proteins upon Fas receptor stimulation. Activation of Jurkat cells via the Fas receptor or with synthetic ceramides resulted in a Ras-and Rac1-dependent stimulation of JNK and p38-K. The functional significance of the Fas-and ceramide-initiated signaling pathway from Ras via Rac to JNK/p38-K is indicated by an almost complete inhibition of Fas-or C 6 -ceramide-mediated programmed cell death after transfection of the cells with transdominant inhibitory Ras, Rac1, or an inhibitory Jun construct (Tam67) and simultaneous treatment with a specific pharmacological blocker (SB 203580) of p38-K suggesting an important function of the described signaling pathway in the regulation of Fas-mediated programmed cell death.
Inhibition of Ras, Rac, JNK, or p38-K-Co-transfection of Jurkat cells with transdominant inhibitory pEF-N17ras, pCEV-N17rac1, pRc/ CMV-tam67 (each 50 g/20 ϫ 10 6 cells) or vector control (pEF, pCEV) with an expression vector for CD20 (pRc/CMV-cd20) (10 g) was performed as described previously (9). N17Ras, N17Rac, or Tam67 have been shown previously to inhibit the function of Ras, Rac, or JNK, respectively (17,49,50). Briefly, cells were electroporated using a BTX electroporator at 5 pulses (99 s) and 500 V. 12 h later, viable cells were purified by Ficoll gradient centrifugation and cultured for an additional 24 h. CD20 ϩ cells were then selected by incubation with 50 g/ml anti-CD20 monoclonal antibody (Dianova, Hamburg, Germany) (60 min, 4°C), three washes, and further incubation (60 min, 4°C) with magnetic beads coated with a sheep anti-mouse immunoglobulin (Dynal, Hamburg, Germany) (25). Since the ratio of 5:1 for pEF-N17ras, pCEV-N17rac1, or pRc/CMV-tam67:pRc/CMV-cd20 drives expression of N17Ras, N17Rac1, or Tam67 in any CD20 ϩ cell, the selection of CD20 ϩ cells permits effective sorting for N17Ras-or N17Rac1-expressing cells. The fraction of CD20-positive and thus N17Ras-transfected cells was determined by in vivo labeling of the cells with [ 3 H]thymidine as described previously (9). Aliquots of the labeled cells were counted prior and after sorting via magnetic beads. These experiments showed that approximately 10% of all cells were CD20-positive. Nonspecific binding of the cells to the magnetic beads did not exceed 0.5% of all cells, as determined by incubation of cells with an irrelevant monoclonal mouse antibody (50 g) and magnetic beads. Purified cells were allowed to recover for 30 min at 37°C and then used for determination of JNK, p38-K, Rac1, or Ras activity and apoptosis.
Inhibition of p38-K was achieved by preincubation with the specific kinase inhibitor SB 203580 (10 M, kindly provided by Dr. B. MacKintosh, University of Dundee).
Ras and Rac Assay-Untransfected Jurkat cells or Jurkat cells cotransfected with N17Ras, N17Rac1, pCEV, or pEF and CD20 were washed twice in phosphate-free Dulbecco's modified Eagle's medium, resuspended in the same medium complemented with 10% dialyzed fetal calf serum, and labeled for 4 h with 1 mCi/ml 32 P i . Cells were stimulated with anti-Fas (2 g/ml) or C 6 -ceramide (5 M) for the indicated time or left untreated. Stimulation was terminated by lysis in 25 mM Tris-HCl (pH 7.5), 1% Triton X-100, 0.2% SDS, 0.5% sodium deoxycholate, 20 mM MgCl 2 , 450 mM NaCl, and 100 g each of aprotinin and leupeptin (lysis buffer). Nuclei and cell debris were removed by centrifugation (25,000 ϫ g) at 4°C for 15 min. Ras or Rac1 were immunoprecipitated from the lysates at 4°C for 45 min using the monoclonal anti-Ras antibody Y13-259 or a polyclonal, affinity-purified anti-Rac1 antibody (Santa Cruz Biotechnology Inc., Santa Cruz, CA), respectively. The anti-Rac1 antibody detects amino acids 179 -188 of the C-terminal protein domain and does not cross-react with Rac2, Ras, or other known small G-proteins. It has been shown previously to immunoprecipitate active Rac1 proteins (49). The immunoprecipitates were collected by further incubation with protein A/G-conjugated agarose beads (Santa Cruz). Precipitates were washed seven times in lysis buffer and once in 25 mM Tris-HCl (pH 7.5), 0.1% Triton X-100, 1 mM MgCl 2 , and 125 mM NaCl; bound nucleotides were eluted in 1 mM EDTA at 68°C for 20 min. The samples were then centrifuged at 15,000 ϫ g for 5 min, and the nucleotides in the supernatant were separated on polyethyleneiminecellulose plates (Machery & Nagel, Dueren, Germany) with 0.75 M KH 2 PO 4 (pH 3.5). The TLC plates were analyzed by autoradiography, the spots corresponding to GDP and GTP were scraped from the plate, and radioactivity was determined by liquid scintillation counting. Ac-tivation of Ras or Rac1 is expressed by an increase of the GTP/(GDP ϩ GTP) ratio.
Phosphorylation of JNK and p38-K-Cell stimulation was terminated by lysis in 25 mM Hepes (pH 7.4), 0.2% SDS, 0.5% sodium deoxycholate, 1% Triton X-100, 125 mM NaCl, 10 mM each NaF, Na 3 VO 4 , and sodium pyrophosphate and 10 g/ml of each aprotinin and leupeptin (RIPA buffer), the lysates were centrifuged at 25,000 ϫ g for 20 min, and JNK or p38-K were immunoprecipitated from the supernatants overnight at 4°C using 3 g of rabbit polyclonal anti-JNK or anti-p38-K antiserum (Santa Cruz). All control immunoprecipitates were performed with polyclonal rabbit immunoglobulins. After addition of protein A/G-agarose, incubation was continued for at least 60 min. Immunocomplexes were washed six times in lysis buffer and resuspended in SDS-sample buffer (60 mM Tris (pH 6.8), 2.3% SDS, 10% glycerol, 5% ␤-mercaptoethanol). Separation of proteins was performed by 10% SDS-PAGE, followed by an electrophoretic transfer to Immobilon polyvinylidene difluoride filters (Millipore, Eschborn, Germany). Blots were blocked and incubated overnight at 4°C with anti-phospho-JNK or anti-phospho-p38-K antibodies (New England Biolabs, Schwalbach, Germany). Each antibody was diluted to 0.5 g/ml in Tris-buffered saline supplemented with 0.1% Tween 20. Immunoblots were developed by incubation with horseradish peroxidase-conjugated protein G (Bio-Rad) and a chemiluminescence kit (Amersham). All blots were stripped by a 45-min incubation in 20 mM Tris (pH 6.8), 2% SDS, 70 mM ␤-mercaptoethanol at 70°C after primary analysis and reprobed to test for equal amounts of immunoprecipitated protein. Alternatively, an aliquot of the immunoprecipitates was separated by SDS-PAGE and analyzed by Western blotting for equal amounts of protein in all immunoprecipitates.
JNK and p38-K Activity-To measure the activity of JNK and p38-K, cells were lysed in RIPA buffer as above after stimulation via Fas or with C 6 -ceramide. Lysates were centrifuged at 25,000 ϫ g for 20 min, and JNK or p38-K were immunoprecipitated from the supernatants at 4°C for 4 h using polyclonal rabbit anti-human JNK or p38-K antisera. Immunocomplexes were immobilized on agarose-coupled protein A/G; incubated for an additional 60 min at 4°C; washed twice in RIPA buffer; twice in H/S, 1% Nonidet P-40, 2 mM Na 3 VO 4 ; once in 100 mM Tris (pH 7.5), 0.5 M LiCl; and finally twice in kinase buffer 12.5 mM MOPS (pH 7.5), 12.5 mM ␤-glycerophosphate, 0.5 mM EGTA, 7.5 mM MgCl 2 , 0.5 mM NaF, 0.5 mM Na 3 VO 4 . After washing, the immunoprecipitates were resuspended in kinase buffer supplemented with 10 Ci/sample [␥-32 P]ATP (6000 Ci/mmol, NEN Life Science Products), 10 M ATP, and 1 g/ml GST-c-Jun (amino acids 1-79) or GST-ATF-2 (amino acids 1-96). The samples were incubated at 30°C for 15 min and stopped by addition of 5 l of boiling 5 ϫ reducing SDS sample buffer. Samples were separated by 10% SDS-PAGE and analyzed by autoradiography and laser densitometry. The substrates GST-c-Jun and GST-ATF-2 were expressed in DH5␣ bacteria after transformation, growth for 12 h, and incubation with isopropyl-1-thio-␤-D-galactopyranoside (200 M) for 4 h. Bacteria were lysed in 25 mM Hepes (pH 7.4), 0.2% SDS, 0.5% sodium deoxycholate, 1% Triton X-100, 125 mM NaCl, and 10 g/ml each of aprotinin and leupeptin; GST fusion proteins were purified by binding to glutathione-agarose and eluted in kinase buffer with 20 mM glutathione. Purity of the preparations was tested by SDS-PAGE and Coomassie staining.
Phosphorylation of GADD153-To determine the serine/threonine phosphorylation of GADD153, Jurkat cells were metabolically labeled with 32 P i for 4 h, stimulated with anti-Fas or C 6 -ceramide, and lysed in RIPA buffer as above. GADD153 was immunoprecipitated from the post-centrifugation supernatant using a polyclonal rabbit anti-GADD153 antibody (Santa Cruz) as above, the samples were separated by 10% SDS-PAGE and transferred to an Immobilon membrane, and phosphorylation of GADD153 was analyzed by autoradiography. After analysis, the blots were reprobed with anti-GADD153 antibodies to test for equal amounts of protein in all lanes. In addition, GADD153 was immunoprecipitated from unlabeled, stimulated, or unstimulated cells and the immunoprecipitates were separated and blotted as above. The blots were blocked and incubated overnight at 4°C with anti-phosphoserine/phosphothreonine antibodies (0.5 g/ml; Sigma, Deisenhofen, Germany). Immunoblots were developed using horseradish peroxidaseconjugated protein G and a chemiluminescence kit. The blots were stripped after primary analysis and reprobed to test for equal amounts of immunoprecipitated GADD153.
Phosphorylation of GADD153 after cellular stimulation with anti-Fas or C 6 -ceramide was further analyzed by incubation of recombinant GADD153 (0.1 g/sample, Santa Cruz) with cell lysates from stimulated or unstimulated samples. Cells were lysed in 12.5 mM MOPS (pH 7.5), 1% Triton, 12.5 mM ␤-glycerophosphate, 0.5 mM EGTA, 7.5 mM MgCl 2 , 0.5 mM NaF, 0.5 mM Na 3 VO 4 , centrifuged, and recombinant GADD153 and 10 Ci/sample [␥-32 P]ATP were added to the supernatants. The samples were then incubated at 37°C for 20 min, and the reaction was stopped by addition of 5 l of boiling SDS sample buffer and 5% ␤-mercaptoethanol. The samples were separated by 10% SDS-PAGE, transferred to a polyvinylidene difluoride membrane, and analyzed by autoradiography and laser densitometry.
Apoptosis-Fas-induced cell death was determined by metabolic labeling of Jurkat cells for 12 h with 10 Ci/ml [ 3 H]thymidine (8.3 Ci/mmol, NEN Life Science Products) (9). Cells were washed, aliquoted, and incubated with anti-Fas (200 ng/ml) or left untreated. Cell death was determined after 3 h by DNA fragmentation and trypan blue staining. Briefly, cells were disrupted by one cycle of freezing at Ϫ20°C and thawing, after which unfragmented genomic DNA was collected by filtration through glass fiber filters (Pharmacia, Freiburg, Germany) and counted by liquid scintillation (9). Results are expressed as % DNA fragmentation Ϯ S.D. compared with control samples. Experiments were done in triplicate and repeated three times. In all experiments an aliquot of the cells was analyzed for cell death by trypan blue. Apoptosis was identified by the typical morphological changes, in particular membrane blebbing, condensation, and fragmentation of nuclei. Comparison of the DNA fragmentation method and the morphologic changes per-mits exact determination of apoptosis versus necrosis. In all experiments initial rates of apoptosis did not exceed 5% of all cells.

RESULTS
We recently demonstrated an activation of the Ras signaling pathway upon Fas receptor triggering or cellular treatment with ceramides (9). To identify downstream targets of this signaling pathway initiated by Fas or ceramide, we tested the activation of JNK/p38-K. Fas receptor triggering resulted in an approximately 15-fold stimulation of JNK-and p38-K -activity (Fig. 1, A and B). The stimulation of JNK or p38-K after Fas was comparable to the activation of the kinases upon cellular treatment with C 6 -ceramide (Fig. 1, A and B). The activity of JNK or p38-K was determined by phosphorylation of the substrates GST-c-Jun or GST-ATF-2, respectively. The concentration of 5 M C 6 -ceramide used in the present study has been shown to result in an intracellular concentration of 10 -100 nmol/nmol of lipid (51). These concentrations are physiologically relevant, since they are also obtained upon Fas receptor ligation or serum deprivation (51). The activation of JNK/ Washed immunoprecipitates were subjected to in vitro kinase assays using GST-c-Jun or GST-ATF-2 fusion proteins as substrates. Samples were separated by 10% SDS-PAGE, blotted, and analyzed by autoradiographies. The blots show an approximately 15-fold stimulation of JNK and p38-K after Fas receptor ligation or treatment of the cells with C 6 -ceramide. After primary analysis, all blots were reprobed with anti-JNK or anti-p38-K antibodies to determine the level of JNK or p38-K protein in the samples (small blots). The analysis shows similar protein levels in all lanes. C, cellular treatment with heat shock, sorbitol, or PMA (known stimuli of JNK or p38-K) induces a similar activation of JNK and p38-K as observed after stimulation with anti-Fas (2 g/ml) or C 6 -ceramide (5 M). The inactive stereoisomer dihydro-C 2 -ceramide (5 M) does not induce the stimulation of JNK or p38-K showing the specificity of the observed activation event. Cells were treated at 42°C for 30 min, with 400 mM sorbitol, 10 ng/ml PMA, anti-Fas (2 g/ml), and C 6 -ceramide or dihydro-C 2 -ceramide (each 5 M). The activity of JNK and p38-K was determined as described above. The reprobed blots show similar amounts of JNK or p38-K in all lanes. D, activation of JNK and p38-K correlates with a phosphorylation of the kinases after Fas receptor ligation or cellular treatment with C 6 -ceramide (5 M). Cells were stimulated as indicated, and JNK or p38-K were immunoprecipitated from the lysates. The immunoprecipitates were blotted and analyzed for phosphorylation using a phospho-JNK or phospho-p38-K specific antibody, followed by ECL-development. The blots were stripped and reprobed with an anti-JNK or anti-p38-K antibody showing similar amounts of all proteins in all lanes. Control (c) obtained from Fas-or ceramide-stimulated samples using an unspecific, irrelevant rabbit antibody did not show any cross-reacting kinase activity. All experiments were repeated three times. p38-K by Fas or C 6 -ceramide was similar to the stimulation of JNK or p38-K by prevoiusly reported stimuli, in particular heat shock, sorbitol, or PMA (Fig. 1C) (16, 52Ϫ55). The inactive stereoisomer dihydro-C 2 -ceramide did not induce stimulation of JNK or p38-K, showing the specificity of the activation.
The increase of JNK or p38-K activity correlated with a phosphorylation of the two kinases upon Fas receptor ligation or cellular treatment with C 6 -ceramide (Fig. 1D). Phosphorylation was determined by incubation of JNK or p38-K immunoprecipitates with a phospho-JNK or phospho-p38-K specific antibody and development using the ECL technique. Reprobing the blots with anti-JNK or anti-p38-K antibodies revealed similar amounts of proteins in all lanes (shown in the smaller blots below).
To elucidate the mechanism of JNK and p38-K activation by Fas and C 6 -ceramide, we tested whether Ras and Rac proteins, which have been shown to be upstream regulators of JNK/ p38-K (52,56), regulate the stimulation of JNK/p38-K after anti-Fas or C 6 -ceramide treatment. Rac proteins have been implied as intermediates transmitting signals from Ras to Jun N-terminal kinase kinase or MAP-kinase kinase, which then activate JNK or p38-K, respectively (52,56). A stimulation of Ras by Fas receptor ligation or cellular treatment with ceramides has been previously shown by us (9). Further studies from our group showed a Ras-dependent activation of Rac1 and Rac2 upon Fas receptor triggering (49). Here, we provide a detailed analysis of the kinetics of Rac activation upon Fas receptor ligation and show that Rac is rapidly stimulated by Fas receptor triggering peaking 10 min after stimulation (Fig. 2A). The activity of Rac1 after Fas remained increased for more than 40 min after anti-Fas-treatment.
The activation of Rac1 upon Fas receptor ligation was mimicked by cellular treatment with C 6 -ceramide (Fig. 2B); Rac1 activation peaked approximately 10 min after addition of C 6ceramide and remained stimulated for more than 40 min after addition of C 6 -ceramide (Fig. 2B). The specificity of C 6 -ceramide-mediated Rac1 stimulation is indicated by the finding that the biologically inactive stereoisomer dihydro-C 2 -ceramide does not trigger Rac1 activation (data not shown).
To analyze the involvement of Ras and Rac proteins in JNK/ p38-K activation by Fas or C 6 -ceramide, Jurkat cells were transiently co-transfected with an expression vectors for transdominant inhibitory N17Ras or N17Rac1 and CD20. The mutants have a very high affinity to GDP (50) and bind endogenous guanine nucleotide exchange factors, preventing the activation of endogenous Ras or Rac1. N17Rac1 inhibits the activation of both Rac1 and Rac2 after Fas receptor stimulation as demonstrated previously (49). Expression of the B-lymphocyte antigen CD20 in transfected cell permits efficient purification of N17Ras-or N17Rac1-expressing cells (9). Inhibition of endogenous Ras by N17Ras or Rac by N17Rac1 almost completely prevented activation of JNK or p38-K after Fas receptor triggering (Fig. 3, A and B) or cellular treatment with ceramides (Fig. 3, C and D). The crucial role of Ras and Rac proteins for Fas-or C 6 -ceramide-initiated JNK and p38-K activation is also indicated by an inhibition of the phosphorylation of JNK and p38-K upon Fas receptor or C 6 -ceramide triggering by transfection of N17Ras or N17Rac1 (data not shown). The efficiency of N17Ras or N17Rac1 expression and the purification process has been shown previously (49) and is demonstrated in Fig. 3E. These data indicate an approximately 90% inhibition of Ras or Rac1 activation after anti-Fas triggering ( Fig. 3E and Ref. 49). In summary, these data suggest that Ras and Rac proteins are upstream regulators of JNK/p38-K upon Fas receptor stimulation or treatment with C 6 -ceramides.
To determine the significance of the observed signaling cascade for Fas-or ceramide-induced programmed cell death, we measured the effect of an inhibition of endogenous JNK/p38-K, Rac1, or Ras by transfection of transdominant inhibitory constructs or treatment with pharmacological inhibitors on apoptosis after anti-Fas or C 6 -ceramide (Fig. 4). Cells were cotransfected with either Tam67, N17Rac1, or N17Ras and CD20 or treated with the specific p38-K inhibitor SB 203580, metabolically labeled with [ 3 H]thymidine, purified by sorting via CD20, and stimulated with anti-Fas or C 6 -ceramide, and apoptosis was determined as described above. In all experiments apoptosis was also detected by typical morphological changes after trypan blue staining. The experiments show that inhibition of JNK and p38-K, endogenous Rac1, or Ras prevents Fasand C 6 -ceramide-induced apoptosis by more than 80%, indicating the significance of the signaling cascade from Ras via Rac proteins to JNK/p38-K for Fas-and ceramide-triggered programmed cell death. Single inhibition of JNK or p38-K did not change Fas-or ceramide-induced apoptosis. Control experiments with SB 203580 showed that the inhibitor reduced p38-K activity after Fas receptor stimulation by approximately 90% (data not shown). The efficiency of Tam67 transfection was tested by measuring the up-regulation of c-Jun and IL-2 production upon treatment of the cells with PMA (10 nM) and ionomycin (500 ng/ml). Both activation markers have been previously shown to be inhibited by expression of Tam67 (45). Induction of c-Jun or release of IL-2 were almost completely inhibited in cells transfected with Tam67 and sorted via CD20, whereas control transfected cells responded normally with c-Jun expression and IL-2 synthesis (data not shown). These results indicate a sufficient transfection and expression of Tam67 in CD20-sorted Jurkat cells.
To gain insight into the function of JNK/p38-K activation upon Fas receptor stimulation, we measured the phosphorylation of the transcription factor GADD153. GADD153 is a known target of JNK and has been implied in the regulation of gene expression and/or cell cycle regulation (57)(58)(59). Fas receptor ligation (Fig. 5A) or ceramide treatment (Fig. 5B) induced a   FIG. 2. Fas (A) or C 6 -ceramide (B) induce an activation of Rac1. Jurkat cells were metabolically labeled for 4 h and stimulated with anti-Fas (2 g/ml) or C 6 -ceramide (5 M) for the indicated time. Rac1 was immunoprecipitated from the lysates, and bound guanine nucleotides were eluted from the immunoprecipitates, separated by TLC, and analyzed by autoradiography. The results show a long lasting activation of Rac1 upon ligation of the Fas receptor, which is comparable with the stimulation of Rac1 after cellular treatment with C 6 -ceramide. Activation of Rac1 was repeated three times. Immunoprecipitates using an irrelevant polyclonal control rabbit antiserum did not reveal any unspecific immunoprecipitation of a small G-protein (c ϭ control immunoprecipitates). strong phosphorylation of GADD153, which was almost completely inhibited by transfection of transdominant inhibitory N17Rac1 or of the transdominant inhibitory c-Jun construct Tam67 and simultaneous pretreatment with the p38-K inhibitor SB 203580 (Fig. 5, C and D). Similar results were obtained using an anti-phosphoserine antibody showing an approximately 10-fold stimulation of GADD153 phosphorylation (data not shown).
The in vivo phosphorylation of GADD153 could be mimicked by incubation of recombinant GADD153 with cell lysates obtained from cells stimulated via the Fas receptor or with C 6 ceramides (Fig. 5E).
The data show that the phosphorylation of GADD153 is regulated by a signaling cascade via Rac proteins and JNK/p38-K. DISCUSSION The results of the present and previous studies (9,49) show a signaling cascade from the Fas receptor via the small Gproteins Ras and Rac to JNK/p38-K and the transcription factor GADD153. Activation of this cascade seems to be necessary for Fas-or C 6 -ceramide-mediated cell death, since genetic or pharmacological inhibition of Ras, Rac, or JNK/p38-K prevents the induction of apoptosis. Furthermore, the results identify GADD153 as a new, Ras-, Rac-, and JNK/p38-K-regulated downstream effector of Fas and synthetic ceramide. This notion is supported by the finding that incubation of recombinant GADD153 with cell lysates in vitro from Fas or C 6 -ceramide stimulated cells results in a phosphorylation of GADD153.
In the present study, inhibition of Ras and Rac proteins was achieved by transient transfection of transdominant inhibitory stimulation pointing to a regulation of the two kinases by Ras and Rac. Jurkat cells were co-transfected with transdominant inhibitory N17Ras or N17Rac-1 and CD20; CD20-positive cells were sorted and stimulated via the Fas receptor (2 g/ml) or with C 6 -ceramide (5 M); and JNK or p38-K activity was determined as described above. Transfection of the vector controls (pEF or pCEV) did not affect Fas-or C 6 -ceramide-induced stimulation of JNK or p38-K. E, transfection of N17Rac1 prevents the activation of Rac1 after Fas receptor ligation, whereas transfection of the vector did not alter the stimulation of Rac1. Transfection of N17Rac1 also prevents the activation of Rac2 by Fas as demonstrated previously (9). The autoradiographies were also analyzed by laser scanning and the results are presented as % increase of the GTP/(GDP ϩ GTP) ratio. The data show the efficiency of the transfection and sorting process almost completely blocking the activation of endogenous Rac1. Control immunoprecipitates (c) using a rabbit antiserum do not show the precipitation of a cross-reacting kinase. All experiments were repeated at least two times.
constructs. These mutants, N17Ras or N17Rac1, have a very high affinity to GDP and therefore bind the endogenous GDP/ GTP-exchange factors, preventing the activation of endogenous Ras or Rac1/2 (49,50). To identify transfected cells, we performed a co-transfection with an expression vector for CD20, which is a B-cell antigen not expressed in T-lymphocytes. This co-transfection technique permits efficient sorting of transfected cells as described previously (9). The blockade of JNK/ p38-K stimulation by genetic inhibition of Ras and Rac indi-cates that Fas receptor-induced JNK and p38-K activation is mediated by a sequential activation of Ras and Rac proteins. Our data show that Rac1 is longer (more than 40 min) active than JNK/p38-K. Since Rac1 is able to interact with different downstream effector molecules linking Rac1 to cytoskeleton regulation (60,61), JNK/p38 activation (52,56), or reactive oxygen release (62,63), it is likely that JNK/p38-K are not the only target molecules of Rac1 upon Fas receptor ligation. Thus, it might be possible that Rac determines in a temporally organized way the activity of different downstream molecules, and therefore the activation time of these molecules might be shorter than of Rac1.
Inhibition of JNK was achieved by transfection of a transdominant inhibitory c-Jun construct, TAM67, which has been used previously to inhibit cell death triggered by the nerve growth factor-p75 receptor (17). The Tam67 c-Jun mutant lacks the N-terminal transactivation domain, which includes the JNK binding site, and functions as a transdominant interfering mutant. p38-MAP kinase was inhibited by incubation of the cells with the pharmacological inhibitor SB 203580, which is considered to be a specific inhibitor of p38-K (64). Tam67 transfection prevented the up-regulation of c-Jun or release of IL-2 after T-cell stimulation. Likewise, SB 203580 almost completely inhibited the activity of p38-K upon Fas receptor triggering. This demonstrates the efficiency of the Tam67 transfection and the kinase inhibitor treatment. Since, in accordance with other studies (45,65), single inhibition of JNK or p38-K did not prevent Fas-or ceramide-triggered programmed cell death, whereas combined inhibition of JNK and p38-K significantly reduced Fas-or ceramide-mediated apoptosis, the activation of both kinases seems to be required for induction of apoptosis in Jurkat cells. Likewise, ceramide-initiated JNK activation has been demonstrated to be essential for the induction of apoptosis after ionizing radiation, UV light, heat shock, ceramide, TNF␣, or H 2 O 2 (16). However, in this study, Tam67transfected U937 cells were almost completely resistant to induction of apoptosis by the mentioned stimuli. Therefore, it might be possible that the expression levels of JNK versus p38-K determine whether both kinases are required for induction of apoptosis or only one of the two kinases. Thus, in some cell types Tam67 might be sufficient to prevent apoptosis by ceramide, Fas, or stress, whereas in other cells, e.g. Jurkat, inhibition of both JNK and p38-K is required to prevent apoptosis. JNK/p38-K do not seem to be important in all forms of apoptosis, since transfection with Tam67, treatment with SB 203580, or combination of the two inhibitors did not prevent apoptosis in Jurkat cells incubated with thapsigargine (1 M), a microsomal Ca 2ϩ ATPase inhibitor. 2 Our notion of an activation of JNK and p38-K by Fas receptor ligation is supported by previous studies showing that Fas stimulates these kinases (16, 43-45, 65, 66). The time course of JNK activation after Fas seems to depend on the dose of the antibody used for stimulation, since treatment with 1-2 g/ml anti-Fas results in JNK activation after already 10 min (Ref. 43 and present study), whereas lower concentrations (30 -100 ng/ ml) require longer incubation times (60 -90 min) to induce JNK stimulation.
In summary, the studies suggest an important function of JNK/p38-K for Fas-or ceramide-mediated programmed cell death.
The activation of a signaling cascade from the Fas receptor via Ras or Rac proteins and JNK/p38-K to GADD153 may have several functions important for Fas-or ceramide-triggered pro- FIG. 4. Inhibition of Rac proteins, JNK, and p38-K prevents Fas-or C 6 -ceramide-induced programmed cell death. Cells were labeled for 12 h with [ 3 H]thymidine prior to induction of apoptosis. Rac proteins or JNK were inhibited by transient co-transfection of transdominant inhibitory N17Rac1 or Tam67 and CD20. CD20 ϩ cells were sorted and treated with anti-Fas (2 g/ml) or C 6 -ceramide (5 M) for 4 h. p38-K was inhibited by preincubation with the specific p38-K inhibitor SB 203580 (10 M). DNA fragmentation was determined by binding of intact DNA to glass fiber filters, followed by liquid scintillation counting (9). Apoptosis was also measured by staining an aliquot of the cells with trypan blue and determining typical morphological changes. The two methods measuring apoptosis or DNA fragmentation showed very similar results. grammed cell death; JNK/p38-K-regulated phosphorylation of GADD153 may result in a change of GADD153 activity and in an altered regulation of gene expression and/or the cell cycle. GADD153, also known as CHOP or growth arrest and DNA damage-inducible gene 153, has been recently shown to be phosphorylated upon stress via a pathway involving p38-K (57,59). GADD153 forms heterodimers with members of the C/EBP family transcription factors (57)(58)(59), which either inhibit C/EBP binding to DNA or direct the GADD153-C/EBP heterodimers to other DNA sequences resulting in a change of gene expression or of the cell cycle (58,59). The function of GADD153 phosphorylation upon Fas receptor triggering is unknown. However, since several reports indicate an important role of the cell cycle in the regulation of cellular apoptosis and cells may undergo apoptosis only during a certain phase of the cell cycle (36,(67)(68)(69), it might be possible that regulation of the cell cycle is the major role of Fas-initiated GADD153 regulation via Ras and Rac1. In this context, the phosphorylation of GADD153 may have one of the following functions. First, if phosphorylation of GADD153 upon Fas receptor ligation inhibits the activity of GADD153 to induce cell cycle arrest, the cells will progress faster through the cell cycle. If cells are sensitive to Fas only during a specific phase of the cell cycle, a faster cell cycle will result in a higher percentage of Fas sensitive cells per time, which might be the function of GADD153. This hypothesis is supported by the effects of N17Ras or N17Rac on Fas-or ceramide-induced death. Ras or Rac may be necessary for cell cycle progress and inhibition of the cell cycle by N17Ras or N17Rac transfection may therefore block the cell cycle preventing Fas-or ceramide-induced cell death. Second, the phospho-rylation of GADD153 may promote its ability to induce cell cycle arrest. In this case, the cell cycle might be arrested during a Fas-sensitive phase permitting Fas receptor ligation to trigger apoptosis. Finally, it might be possible that ceramide and Fas induce growth arrest and/or DNA damage, which may result in an altered activity of GADD153. Thus, the phosphorylation of GADD153 would be secondary to the induction of apoptosis. This possibility seems to be unlikely since the phosphorylation of GADD153 was also observed in cell lysates lacking nuclei. Therefore, the phosphorylation of GADD153 does not depend on the presence of a nucleus. Furthermore, GADD153 phosphorylation is already observed 10 min after Fas receptor ligation or C 6 -ceramide treatment. This time seems to be rather short to mediate a significant DNA damage resulting in the regulation of GADD153. However, since the physiological function and regulation of GADD153 is only poorly characterized, an alternative function is certainly possible and elucidation of the exact role of GADD153 in Fasinduced apoptosis has to be performed in future studies.
It has been shown that Fas-induced cell death does not require new mRNA or protein synthesis (70). Our experiments do not exclude the possibility that cells already in a certain state of the cell cycle are able to undergo apoptosis upon Fas receptor ligation, whereas in a cell population with heterogenous cell cycle status the cell cycle progression regulated via Ras, Rac, JNK/p38-K, and GADD153 might be necessary to achieve complete induction of apoptosis by Fas.
Alternatively, JNK and p38-K may have cytosolic targets involved in the apoptosis of lymphocytes upon Fas receptor ligation. In particular, p38-K has been implied in the phospho- 5. A and B, GADD153 is phosphorylated after stimulation via the Fas receptor ligation (A) or treatment with C 6 -ceramide (B). Jurkat cells were metabolically labeled with 32 P i , stimulated with anti-Fas (2 g/ml) (A) or C 6 -ceramide (5 M) (B) for the indicated time, GADD153 was immunoprecipitated, samples were separated by 10% SDS-PAGE and blotted, and phosphorylation was analyzed by autoradiography. Reprobing of the blots revealed similar amounts of GADD153 in all lanes. C and D, inhibition of Rac or JNK and p38-K prevents phosphorylation of GADD153 upon Fas receptor (C) triggering or C 6 -ceramide (D) application. Jurkat cells were cotransfected with N17Rac1 or Tam67 and CD20 and sorted for CD20 ϩ cells. Tam67-transfected cells were also incubated with the p38-K inhibitor SB 203580. Cells were stimulated with anti-Fas (2 g/ml) (C) or C 6 -ceramide (5 M) (D), and the phosphorylation of GADD153 was determined as above. E, cell lysates from Fas-or ceramide-activated cells phosphorylate GADD153 in vitro. Jurkat cells were stimulated via the Fas receptor (2 g/ml) or with C 6 -ceramide (5 M) and lysed, and the lysates were incubated with recombinant GADD153 in the presence of [␥-32 P]ATP. Lysates from stimulated cells contain a kinase activity phosphorylating GADD153. Control immunoprecipitates (c) using irrelevant rabbit antibodies show the specificity of GADD153 immunoprecipitation. The experiments were repeated three times. rylation of heat shock protein HSP 25/27 (71,72). HSP 25/27 is phosphorylated by a MAP kinase-activated protein kinase, which is a substrate of p38-K during cellular stress responses (72,73). Since constitutive expression of HSP 25/27 has been shown to prevent Fas-induced apoptosis (73), it might be possible that phosphorylation of HSP 25/27 inhibits the function of HSP 25/27, permitting the cell to undergo apoptosis after Fas receptor ligation.
The effects of Fas receptor ligation on Ras, Rac, and JNK/ p38-K are mimicked by synthetic ceramides, and inhibition of the signaling cascade from Ras to JNK also prevents Fas-as well as ceramide-triggered programmed cell death. Further data from our group show that the consumption of sphingomyelin, the release of ceramide, activation of the acidic sphingomyelinase, and stimulation of JNK/p38-K are prevented by inhibition of caspases by transfection with CrmA or after treatment with pharmacological caspase inhibitors. 3 This notion is supported by recent findings showing that the activation of JNK and p38-K upon Fas receptor ligation requires caspases (45,65). The data suggest a signaling cascade from Fas via caspases, the acidic sphingomyelinase, release of ceramide, and activation of Ras and Rac to a stimulation of JNK/p38-K. However, a recent study using transfection of FADD, RIP, or TRAF-2 into MCF-7 cells proposed that JNK activation and apoptosis after TNF receptor ligation are separate responses (74). Therefore, JNK/p38-K activation may not be sufficient by itself to trigger apoptosis; however, under physiological conditions it might be necessary as a positive signal for the pathway signaling programmed cell death. In situations with a very high activation level of the pathway triggering apoptosis, e.g. after transfection of genes inducing cell death, this positive feedback may not be absolutely required.
Ras, Rac, and JNK/p38-K are not only involved in apoptosis but are also important in the regulation of cell proliferation. For example, Ras and JNK are also activated upon CD28 or CD40 receptor stimulation, which induce cell proliferation or differentiation (75)(76)(77)(78). A similar dual function for the regulation of cell proliferation/differentiation or programmed cell death has been shown for c-Myc (34,37), c-Fos (33), Cdc-2 (36), or p120 GAP (35). Therefore, the actual function of Ras, Rac, JNK/p38-K, and GADD153 may depend on co-signals provided from, e.g., growth factor receptors, oncogenes, or receptors inducing programmed cell death. In this concept, Ras and Rac proteins and JNK/p38-K may control apoptosis and may be necessary for but not sufficient to induce apoptosis.