DAP kinase activity is critical for C 2 -ceramide-induced apoptosis in PC12 cells Mutsuya Yamamoto, Takeshi Hioki, Takehisa Ishii, Sadayo Nakajima-Iijima and Shigeo Uchino* Pharmaceuticals Discovery Laboratory, Yokohama Research Center, Mitsubishi-Tokyo Pharmaceuticals Inc., Aoba-ku, Yokohama, Japan Exposure o f PC12 cells to C 2 -ceramide results in dose- dependent apoptosis. Here, we investigate the involvement of death-associated protein (DAP) kinase, initially identi®ed as a positive mediator of the interferon-c-induced apoptosis of HeLa cells, i n t he C 2 -ceramide-induced apoptosis of PC12 cells. DAP kinase is endogenously expressed i n these cells. On exposure of PC12 cells to 30 l M C 2 -ceramide, both the total (assayed in the presence of Ca 2+ /calmodulin) and Ca 2+ / calmodulin-independent (assayed in the presence of EGTA) DAP kinase activities were transiently increased 5.0- and 12.2-fold, respectively, at 10 min, and t hen decreased to 1.7- and 3.4-fold at 90 min. After 10 min exposure to 30 l M C 2 -ceramide, the C a 2+ /calmodulin independent activity/ total activity ratio increased f rom 0.22 to 0.60. These eects were dependent on the C 2 -ceramide concentration. C 8 -cera- mide, another active ceramide analog, also induced apoptosis and activated DAP kinase, while C 2 -dihydroceramide, an inactive ceramide analog, failed to induce apoptosis and increase DAP kinase activity. Furthermore, transfection studies revealed that overexpression of wild-type DAP kinase enhanced the sensitivity to C 2 -andC 8 -ceramide, while a catalytically inactive DAP kinase mutant and a construct containing the death domain and C-terminal tail of DAP kinase, which act in a dominant-n egative manner, rescued cells from C 2 -, and C 8 -ceramide-induced apoptosis. These ®ndings demonstrate that DAP kinase is an important component of the apoptotic machinery involved in ceramide- induced apoptosis, and that the intrinsic DAP kinase activity is critical for ceramide-induced apoptosis. Keywords: apoptosis; DAP kinase; ceramide; PC12; CaM kinase. Using a functional gene selection approach based on random inactivation of gene expression by antisense cDNA libraries, death-associated protein kinase (DAP kinase) was cloned as a positive mediator of interferon-c-ind uced HeLa cell apoptosis [1±3]. DAP k inase c ontains a calmodulin- binding region and phosphorylates both itself and other substrates in a C a 2+ /calmodulin-dependent fashion [1,4]. DCaM-DK, a mutant lacking the c almodulin-binding domain, is constitutively active, while mutant K42A-DK, in which a lysine residue within the kinase domain is replaced by an alanine residue, is inactive. The death- promoting function of DAP kinase requires t he kinase activity; therefore overexpression of the DCaM-DK mutant leads to cell death without any external stimuli, whereas overexpression of the K42A-DK mutant is not cytotoxic [4]. In addition, DAP kinase contains unique domains and motifs, including eight ankyrin repeats, two potential ATP/ GTP binding sites (P-loops), a cytoskeleton binding domain, a d eath domain, and a serine rich C-terminal region [4,5], which are thought to be involved in various biological processes and to modulate DAP kinase function [4,6,7]. We recently found that DAP kinase mRNA is predom- inantly expressed in the brain and lung. In the developing brain, a high level of DAP kinase m RNA expression is detected in proliferative and postmitotic regions of the cerebral cortex and hippocampus and in cerebellar Purkinje cells, suggesting that DAP kinase may p lay a pivotal role in neurogenesis; furthermore, the expression of DAP kinase mRNA is increased prior to cell death induced by transient forebrain ischemia [8]. These ®ndings indicate a possible relationship between DAP kinase and physiological or pathological neuronal cell death. However, it is not known whether DAP kinase activity is critical for neuronal cell death. Ceramide, generated by sphingomyelin hydrolysis, has been proposed as a pleiotropic lipid second messenger regulating cell cycle arrest, d ifferentiation and apoptosis [9,10]. Several reports have shown that exposure to a cell-permeable ceramide analog, N-acetylsphingosine (C 2 -ceramide), induces apoptotic cell death in many types of cells, including neuronal cells, such as primary cultures of mesencephalic neurons [11], sensory neurons [12], immature cerebellar granule cells [13], and the rat pheochromocytoma cell line, PC12 [14]. Correspondence to M. Yamamoto, Pharmaceuticals Discovery Laboratory, Yokohama Research Center, Mitsubishi-Tokyo Pharmaceuticals Inc., 1000 Kamoshida, Aoba-ku, Yokohama 227-8502, Japan. Fax: + 81 45 963 3992, Tel.: + 81 45 963 4340, E-mail: mut u@phy.med.kyoto-u.ac.jp Abbreviations: DAP kinase, death-associated protein kinase; MLC kinase, myosin light chain protein kinase; CaM kinase, Ca 2+ /cal- modulin-dependent protein kinase; ZIP kinase, zipper-interacting protein kinase; DMEM, Dulbecco's modi®ed Eagle's medium; C 2 -ceramide, N-Acetylsphingosine; C 8 -ceramide, N-octanoylsp- hingosine C 2 -dihydroceramide, N-acetylsphinganine; EGFP, enhanced green ¯uorescent protein. *Present address: Department of Neurochemistry, National Institute of Neuroscience, Kodaira, Tokyo, Japan. (Received 2 J uly 2001, accepted 24 October 2001) Eur. J. Biochem. 269, 139±147 (2002) Ó FEBS 2002 In this study, we employed P C12 cells, w hic h are extensively used as a model to study mechanisms regulating neu ronal s urvival and a poptotic cell death [15], to study the role of DAP kinase in ceramide-induced apoptosis. DAP kinase was found to be activated in the early response to ceramide exposure, and the activity depended on the ceramide concentration. The proportion of Ca 2+ /calmodulin-independent activity was also increased during this process. Furthermore, overexpression of wild-type DAP kinase made cells more sensitive to ceramide. In contrast, overexpression of the K42A-DK mutant or of DD-DK, a construct containing the death domain a nd the C-terminal tail, which acts as another kind of dominant negative mutant [6], protected cells from ceramide-induced apoptosis. This is the ®rst study to quantitatively measure DAP kinase activity and to show that DAP kinase is an important component of the apoptotic machinery involved in ceramide-induced apoptosis. EXPERIMENTAL PROCEDURES Drugs and reagents N-Acetylsphingosine (C 2 -ceramide) (Sigma Chemical Co., St Louis, MO), N-octanoylsphingosine (C 8 -ceramide) (Calbiochem-Novabiochem Corp., San Diego, CA) and N-acetylsphingan ine (C 2 -dihydroceramide) (ICN Pharma- ceuticals, Inc., Costa Mesa, CA) were dissolved in ethanol. The ®nal ethanol concentration in the culture medium in all experiments was less than 0.25%. [c- 32 P]ATP (6000 Ciámmol )1 ) and ATP were purchased from Amersham Pharmacia Biotech (Little Chalfont Bucking- hamshire, UK) and AMRESCO Inc. (Solon, OH, USA), respectiv ely. Cell culture PC12 cells were maintained at 37 °C on collagen-coated dishes (Iwaki Glass Co., Ltd, Tokyo, Japan) in normal growth medium composed of Dulbecco's modi®ed Eagle's medium (DMEM) (Life Technologies, Inc., Grand Island, NY, U SA) supplemented with 5% fetal bovine serum (Life Technologies, Inc.), 5% horse serum (Life Technologies, Inc.), 15 m M Hepes (pH 7.4; Life Technologies, Inc.), and 50 lgámL )1 of gentamycin (Sigma) in a humidi®ed atmo- sphere containing 5% CO 2 . DNA fragmentation assay PC12 ce lls were collected by centrifugation and lysed at 50 °C for 2 h in 200 m M Tris/HCl (pH 8.0), 100 m M EDTA,1%SDS,and0.1mgámL )1 of proteinase K (Nacalai tesque, Inc., Kyoto, Japan). Potassium acetate (5.0 M ) was then added to a ®nal concentration of 1.0 M and the lysate centrifuged at 15 000 g for 15 min at 4 °C. The DNA in the supernatant was extracted with an equal volume of phenol/chloroform, precipitated with ethanol, and s uspended i n 1 0 m M Tris/HCl (pH 8.0) and 1 m M EDTA containing 20 lgámL )1 of RNaseA (Nacalai tesque, Inc.). The DNA (2 lg) was then electrophoresed on a 2% agarose gel, and visualized under UV light after staining with ethidium bromide. Kinase assay PC12 cells were collected by centrifugation, washed twice with ice-cold NaCl/P i , then lysed in lysis buffer [50 m M Tris/HCl (pH 7.5), 70 m M NaCl, 1% Triton X-100, 5 m M EDTA, 5 m M EGTA, 2 m M phenylmethylsulfonyl ¯ uo- ride (Sigma), 1 lgámL )1 aprotinin (Roche Diagnostics GmbH, Mannheim, Germany), 1 lgámL )1 leupeptin (Roche Diagnostics GmbH), 100 n M okadaic acid (Calbiochem-Novabiochem Corp.), 50 m M NaF (Nacalai tesque, Inc.), 10 m M Na 3 VO 4 ,30l M mycalolide B (Wako Pure Chemical Industries, Ltd, Osaka, Japan)] [16]. After insoluble material was removed by centrifugation (15 000 g for 5 min a t 4 °C), the lysates were immuno- precipitated with monoclonal anti-(DAP kinase) Ig (Sigma) absorbed on p rotein G±Sepharose 4FF (Amer- sham Pharmacia Biotech). After three washes with kinase assay buffer [50 m M Tris/HCl (p H 7.5), 8 m M MgCl 2 , 0.01% BSA, 100 n M okadaic acid, 0.5 m M dithiothreitol], the beads were incubated for 10 min at 30 °Cinkinase assay buffer containing 100 l M skeletal muscle myosin light chain kinase (MLC k inase) substrate peptide (AKRPQRATSNVFS) ( ICN Pharmaceuticals, I nc.), 200 l M [c- 32 P]ATP (1 Ciámmol )1 ) and either 0.5 m M CaCl 2 and 1 l M bovine calmodulin (Sigma) (total activity) or 1 m M EGTA (Ca 2+ /calmodulin ind ependent activity) [17]. Incorporation of 32 P into MLC kinase substrate peptide w as measured essentially as described by DeR- emer et al. [18]. After incubation, aliquots of reaction mixture were spotted onto P81 phosphocellulose paper (Whatman Inc., Clifton, NJ, USA), which was then washed three times with 75 m M H 3 PO 4 , rinsed with ethanol, dried, and subjected to liquid scintillation count- ing. Kinase activity was normalized to the amount of DAP kinase in the immunoprecipitants, determined by immunoblotting using monoclonal anti-(DAP kinase) Ig (1 : 250 dilution; Transduction Laboratories, Lexington, KY, USA) as primary antibody and horseradish peroxi- dase-conjugated goat anti-(mouse IgG) Ig ( 1 : 1000 dilu- tion; Santa Cruz B iotechnology, Inc., Santa Cruz, CA, USA) as secondary antibody with quanti®cation on an Imaging Scanner (ES-8000; Epson, Tokyo, Japan) with NIH IMAGE 1.61 software. DNA transfection PC12 cells were plated at a density of 2.6 ´ 10 4 cells per cm 2 in normal growth medium in collagen-coated culture dishes and incubated a t 37 °C for 24 h. The cells were cotransfected with 0.04 lgácm )2 of pQBI25 containing enhanced green ¯uorescent protein (EGFP) cDNA (Quantum Biotechnologies Inc., Montreal, Canada) and with 0.36 lgácm )2 of mock expression plasmid (pcDNA3; Invitrogen Corp., Carlsbad, CA, USA) or the DAP kinase expression plasmids, pDK-wt, pDK-DCaM, pDK-K42A, or pDK-DD, carrying, respectively, cDNA coding for wild-type DAP kinase, th e DCaM-DAP kinase mutant, the K42A-DAP kinase mutant, or the DD-DAP kinase mutant (aspartic acid 1299 to arginine 1430), all with an hemagglutinin (HA)-tag, in pcDNA3 [5,6] using 2 lLácm )2 of SuperFect reagent (QIAGEN Inc., Valencia, CA, USA) according to the manufacturer's instructions. 140 M. Yamamoto et al. (Eur. J. Biochem. 269) Ó FEBS 2002 Immunoblotting Forty-eight hours after transfection, the PC12 cells were washed with ice-cold NaCl/P i , harvested, and lysed with a buffer consisting of 50 m M Tris/HCl (pH 7.5), 1 50 m M NaCl, 1% Triton X-100, 5 m M EDTA, 2 m M phen- ylmethanesulfonyl ¯uoride, 1 lgámL )1 aprotinin, 1 lgámL )1 leupeptin, and 30 l M mycalolide B. After the insoluble material was removed by centrifugation (15 000 g for 5 min at 4 °C), the protein concentration of the supernatant was determined using a protein assay kit (Bio-Rad Laboratories, Hercules, C A, USA). Equal amounts of protein (2 lg) were then separated by electro- phoresis on 10% or 15% SDS/polyacrylamide gels, and electrophoretically transferred onto poly(vinylidene di¯uo- ride) (PVDF) membranes (Millipore C orp., Bedford, MA, USA). To detect the exogenous DAP kinase or actin used as the internal control, the membranes were blocked by incubation for 30 min at ro om temperature with 2% dried skimmed milk in NaCl/P i containing 0.1% Tween-20, (NaCl/P i /Tween) then incubated for 1.5 h at room temper- ature with rat monoclonal anti-HA Ig (1 : 5000 dilution; Roche Diagnostics GmbH) or polyclonal anti-actin Ig (1 : 5000 dilution; Sigma Chemical C o.). After washing with NaCl/P i /Tween, the blots were incubated with horseradish peroxidase-conjugated goat anti-(rat IgG) Ig ( 1 : 1000 dilution; Santa Cruz Biotechnology, Inc.) or horseradish peroxidase-conjugated goat anti-(rabbit IgG) Ig (1 : 5000 dilution; Santa Cruz Biotechnology, Inc.) and the reactive bands visualized using a n enhanced chemiluminescence system (ECL plus; Amersham Pharmacia Biotech.) as indicated in the manufacture's protocol. RESULTS Apoptosis induced by ceramide Recent reports indicate that exposure to high concentrations (10±50 l M ) of exogenous C 2 -ceramide results in apoptosis of neuronal cells and that the ap optotic effect depends on the concentration o f C 2 -ceramide and the cell plating density [11±14]. We therefore initially determined the optimal experimental conditions leading to apoptosis in PC12 cells. PC12 cells were plated at a density of 5.0 ´ 10 4 cells per cm 2 in normal growth medium, then, after 2 days, the culture medium was changed to DMEM containing 1% horse serum, 50 lgámL )1 gentamycin (serum reduced medium), and 0, 3, 10, 20, 30, or 50 l M C 2 -ceramide. After incubation at 37 °C f or 12 h, DNA was extracted and chromatin fragmentation examined by electrophoresis. No DNA ladder was seen in PC12 cells treated with 0, 3, 10, or 20 l M C 2 -ceramide, but a typical DNA ladder was detected in t he p re sence of 30 o r 50 l M C 2 -ceramide (Fig. 1A), showing that concentrations of C 2 -ceramide greater than 30 l M induced apoptosis in PC12 cells under these condi- tions. Similar results were obtained using C 8 -ceramide, an active ceramide analog with a longer 8-carbon fatty acid chain (Fig. 1B). In contrast, an inactive ceramide analog, C 2 -dihydroceramide, which lacks the C4±5 trans double bond in th e sphingolipid backbone that is required for the biological effects of ceramide [19], failed to induce apoptosis (Fig. 1C). Typical apoptotic molphological changes, such as shrinkage, rounding up, and loss of adherence, and staining of the nuclei by propidium iodide, were seen in PC12 cells exposed to 30 l M C 2 - ( Fig. 1D) and C 8 -ceramide (data not shown), but not in those e xposed to 30 l M C 2 -dihydrocera - mide (Fig. 1E). Ceramide-induced apoptosis is accompanied by DAP kinase activation DAP kinase, a Ca 2+ /calmodulin-regulated serine/threonine kinase, is a positive mediator of apoptosis [1±3]. As Western blotting showed that DAP kinase was endogenously expressed in PC12 cells (Fig. 2 , lane 1), we examined whether the ceramide-induced apoptosis in PC12 cells was accompanied by DAP kinase activation. The physiological substrate of D AP kinase has not been identi®ed. In a previous report [4], puri®ed myosin light chain was used as a substrate, as the catalytic domain of DAP kinase is highly h omologous to that of MLC kinase. Zipper-interacting protein kinase (ZIP kinase), with a sequence 81% identical to that of the kinase domain of DAP kinase [20±22], a lso phosphorylates the regulatory light chain of myosin II the phosphorylation sites being threonine 18 and serine 1 9 [22]. In this study, the synthetic peptide, skeletal myosin light chain kinase substrate peptide, which contains the regulatory light chain of myosin II site phosphorylated by ZIP kinase, was used as substrate t o measure DAP kinase activity. DAP kinase activity was measured 0, 3, 10, 3 0, and 90 min a fter changing to serum-reduced medium containing 30 l M C 2 -ceramide, C 8 -ceramide, or C 2 -dihydroceramide in the presence (total activity) or in the absence (Ca 2+ / calmodulin independent activity) of Ca 2+ and calmodulin [17]. When the cells were exposed to C 2 -andC 8 -ceramide, the Ca 2+ /calmodulin-independent DAP kinase activity was increased 3.6- and 2.9-fold, respectively, at 3 min, showed a maximal increase of 12.2- and 11.6-fold at 10 min, then reduced to 5.6- and 6.4-fold at 30 min; a 3.4- and 3.6-fold increase was still apparent after 90 min (Fig. 3A). As shown in Fig. 3B, in the presence of C 2 -andC 8 -ceramide, the total DAP kinase activities a lso increased 2.1- and 1.8-fold at 3 min, showed a maximal increase of 5.0- and 4.8- fold at 10 min, then gradually decreased to 1.7- and 1.8-fold at 90 min The Ca 2+ /calmodulin independent activity/total activity ratio is shown in Fig. 3C. This ratio, which was 0.22 in the absence of ceramide (0 min), increased to 0.60 and 0.64 after 30 min exposure to C 2 -andC 8 -ceramide, respectively, then declined gradually. In contrast, 30 l M C 2 -dihydroceramide caused no increase in either total or Ca 2+ /calmodulin-independent DAP kinase activity, and no change in the ratio (Fig. 3A±C). We next determined whether the increase in DAP k inase activity was dependent on the ceramide concentration by exposing P C12 cells to 0, 10, 20, 30, or 50 l M C 2 -ceramide, C 8 -ceramide, or C 2 -dihydroceramide f or 10 min. As shown in Fig. 4A, Ca 2+ /calmodulin-independent DAP kinase activity was slightly stimulated by 1.4- and 1.3-fold by exposure to 10 l M C 2 -andC 8 -ceramide, and greater activity was seen with increasing concentrations of these reagents, theincreaseat50l M C 2 -andC 8 -ceramide being 12.2- and 11.6-fold, respectively. Total DAP kinase activity also increa- sed with increasing concentration of ceramide, the increase at 50 l M C 2 -andC 8 -ceramide b eing 6.5- and 6.1-fold, respectively (Fig. 4B). The Ca 2+ /calmodulin-independent Ó FEBS 2002 DAP kinase as ceramide-induced apoptotic component (Eur. J. Biochem. 269) 141 activity/total activity ratio a lso increased with ceramide concentration, reaching a p lateau of 0.54 and 0.55 at 30 l M C 2 -andC 8 -ceramide, respectively (Fig. 4C). In contrast, C 2 -dihydroceramide was ineffective in activating DAP kinase and in changing the Ca 2+ /calmodulin-independent activity/total activity ratio at any c oncentration t ested (Fig. 4A±C). Approximately equal amounts of DAP kinase were immunoprecipitated irrespective of the test exposure time or the concentration of C 2 -ceramide, C 8 -ceramide, or C 2 -dihydroceramide. Figure 2 shows the results of a typical immunoblotting experiment after 10 min exposure to 30 l M C 2 -ceramide, C 8 -ceramide, or C 2 -dihydroceramide. Protection from ceramide-induced apoptosis by DAP kinase inhibition To determine whe ther DAP kinase activity was critical for C2- and C 8 -ceramide-induced apoptosis, we investigated the viability of PC12 cells overexpressing the wild-type DAP kinase, t he DCaM-DAP kinase m utant (a c onstitutively active kinase mutant), the K 42A-DAP kinase mutant (a catalytically inactive mutant displaying dominant negative features), or a construct encompassing the death domain and the serine-rich C-terminal tail (another kind of Fig. 1. C 2 -ceramide and C 8 -ceramide induce apoptosis in PC12 cells. PC12 cells were cultured in serum-reduced medium containing 0 l M (lane 1), 3 l M (lane 2), 10 l M (lane 3), 20 l M (lane 4) , 30 l M (lane 5 ), or 50 l M (lane 6 ) C 2 -ceramide (A), C 8 -ceramide (B ), or C 2 -dihydroceramide (C). After 12 h , 2 lg of DNA was prepared and examined for the presence of a DNA ladder as described in Experimental procedures. The size markers are indicated on the left. Morp hologic al appearance of PC12 c ells exposed to 30 l M C 2 -ceramide (D) or 30 l M C 2 -dihydroceramide (E). PC12 cells were stainedwith3 lgámL )1 propidium iodide (PI) (D ojindo, Kumamoto, Japan) and o bserved under a phasecontrast (BF) or a ¯uorescence microscope (PI) (Carl Zeiss, Esslingen, Germany). Fig. 2. Immunoblot analysis of DAP kinase in immunoprecipitates. Monoclonal anti-(DAP kinase) Ig (Sigma) was used to immunopre- cipitate endogenous DAP kinase from extracts of P C12 cells for 10 m in in the absence of ceramide (lane 1) or in the presence of 30 l M C 2 -ceramide ( lane 2), 30 l M C 8 -ceramide ( lane 3), or 30 l M C 2 -dihy- droceramide ( lane 4). The immunoprecipitates were analyzed b y immunoblotting using another monoclonal anti-(DAP kinase) Ig (Transduction Laboratories). 142 M. Yamamoto et al. (Eur. J. Biochem. 269) Ó FEBS 2002 Fig. 3. Time course of ceramide-induced DAP kinase activation. PC1 2 cells were exposed to 30 l M C 2 -ceramide, C 8 -ceramide, or C 2 -dihy- droceramide for 0, 3, 10, 30, or 90 min, then the cells wer e solubilized in lysis b uer, and DAP kinase immunoprecipitated with mono- clonal anti-(DAP kinase) Ig. The kinase reaction was performed in the absence (Ca 2+ / calmodulin-independent activity, A) o r pres- ence (total activity, B) of Ca 2+ and calmodu- lin. The relative DAP k inase activity was determined by comparison with the value measured in the absence of ceramide (0 min). The data are presented as the mean  SD of four measurements. (C) Shows the ratio obtained by dividing the actual value for Ca 2+ /calmodulin-independent activity by that for the total activity. Fig. 4. Concentration dependency of DAP kinase activation by ceramide. PC12 cells were exposed for 10 min to 0, 10, 20, 3 0, or 50 l M C 2 -ceramide, C 8 -ceramide, or C 2 -dihydro- ceramide, then solubilized in lysis buer, and DAP kinase immunoprecipitated with mono- clonal anti-(DAP kinase) Ig. The kinase reaction was performed in the absence (Ca 2+ / calmodulin-independent activity, A) o r pres- ence (total activity, B) of Ca 2+ and calmodu- lin. The relative activity of DAP kinase w as determined by comparison with that in the absence of ceramide (0 l M ). The data a re the mean  SD of four measurements. (C) Shows the ratio obtained by dividing the actual value for Ca 2+ /calmodulin-indepen- dent activity by that for t he total a ctivity. Ó FEBS 2002 DAP kinase as ceramide-induced apoptotic component (Eur. J. Biochem. 269) 143 dominant negative mutant) (Fig. 5A). PC12 cells were cotransfected with pcDNA3, pDK-wt, pDK-DCaM, pDK- K42A, or pDK-DD, and pQBI25 containing EGFP cDNA as a marker to visualize the transfected cells. The transfec- tion ef®ciency was determined by counting the number of EGFP-positive cells at 8 h after DNA transfection and was veri®ed to be about 20%, similar in all experiments. After 48 h, ectopic expression of the wild-type and mutant DAP kinase proteins was con®rmed by immunoblotting using monoclonal anti-HA Ig (Fig. 5B). A s expected, the speci®c band of DAP kinase with an apparent molecular mass of  165 KDa was seen in PC12-wtDK and PC12-K42A-DK cells, a slightly lower molecular mass band (160 KDa) was detected in PC12-DCaM-DK cells, a nd a band at  16 KDa was s een in PC12-DD-DK cells. The band intensity in PC12-DCaM-DK cells was weaker than that in the PC12- wtDK and PC12-K42A-DK cells, suggesting that overex- pression of the DCaM-DAP kinase mutant resulted in some growth disadvantage for PC12 cells. No band was seen in PC12 cells transfected w ith pcDNA3 vector alone . Ectopic expression of DAP kinase in PC12 cells (PC12- wtDK) did not cause cell death in the absence of stimuli. This is in agreement with the results o f previous studies using COS cells [1,4], murine Lewis (3LL) and CMT64 lung carcinoma cells [23], but con¯ icts with results in HeLa cells and SV80 human ®broblasts [1,4]. The transfected PC12 cells were cultured for 48 h, then e xposed for 2 4 h to 30 l M C 2 -ceramide, C 8 -ceramide or C 2 -dihydroceramide. T o determine cell viability, we counted the number o f mor- phologically intact EGFP-positive cells under a ¯uorescent microscope (Fig. 6). The viability of PC12-wtDK cells was markedly decreased by exposure to 30 l M C 2 -and C 8 -ceramide, b ut not by exposure t o 30 l M C 2 -dihydro- ceramide. Furthermore, PC12-wtDK cells were more sensitive to C 2 -andC 8 -ceramide than PC12 cells transfected with p cDNA3 mock vector. For PC12-DCaM-DK cells, viable cell numbers were much lower even i n the absence of ceramide or in the presence of 30 l M C 2 -dihydroceramide. In contrast, PC12-K42A-DK and PC12-DD-DK cells were signi®cantly resistant to 30 l M C 2 -andC 8 -ceramide- Control wt-DK ∆CaM-DK K42A-DK anti-HA DD-DK anti-actin B ∆CaM-DK K42A-DK DD-DK wt-DK A HA-tag Kinase domain Ankyrin repeats Death domain CaM binding Cytoskelton binding 0 200 400 600 800 1000 1200 1400 Fig. 5. Expression of wild-type and mutated DAP kinase p roteins i n transiently transfected P C12 c ells. (A) Schematic presentation of the human DAP kinase mutant proteins u se d in these ex pe rimen ts. The DCaM-DK mutant lack s 47 a mino acid residues (pro line 277 to leucine 323) containing the calmodulin-binding domain. In the K42A-DK mutant, a lysine residue at position 42 in t he kinase domain (asterisk) is replaced by an a lanine. The D D-DK construct contains the death domain a nd the serine-rich tail (aspartic acid 1299 to arginine 1430). All expressed proteins have an HA-tag sequence (cross) at th e N -terminal. ( B) PC12 cells were transfected with pcDNA3 (control), pDK-wt (wt-DK) , pDK- DCaM (DCaM-DK), pDK-K42A (K42A-DK), or pDK-DD (D D-DK) and proteins we re prepared as described in Experimental procedures. Approximately 2 lgof each preparation was separated on a 10% SDS-polyacrylamide gel for control, wt-DK, DCaM-DK and K42A-DK, and on a 15% SDS/ polyacrylamide gel for DD-DK and actin. After electrophoretic transfer onto poly( vinylidene di¯uoride) membrane, the blot was reacted with anti- HA Ig to detect ectopic expression of DAP k inase and th e mutants (up per panel) and with a nti-actin Ig ( lower panel) t o quantitate th e loaded protein a mount s. In the left panel, the arrow on the right indicates the w t-DK and K42A-DK bands at  165 K Da and the arrowhead indicates the DCaM-DK band with a slightly lower molecular mass (160 KDa). I n the righ t panel, the arro w indicates the DD-DK band at  16 KDa. 144 M. Yamamoto et al. (Eur. J. Biochem. 269) Ó FEBS 2002 induced apoptosis. These ®ndings clearly demonstrate that DAP k inase activity is critical f or C 2 -andC 8 -ceramide- induced apoptosis in PC12 cells. DISCUSSION DAP kinase is a Ca 2+ /calmodulin-regulated serine/threo- nine kinase that participates in apoptosis induced by a variety of signals [1,6]. We recently showed that the expression of DAP kinase mRNA in brain is increased prior to cell death induced by transient forebrain ischemia [8]. These ®ndings indicate a possible relationship between DAP kinase and neuronal cell death, including apoptosis and necrosis. However, it is not known whether the k inase activity is involved in apoptosis. In the present study, we employed PC12 cells, extensively used as a model to study mechanisms regulating apoptosis [15], to investigate the role of DAP kinase activity in ceramide-induced apoptosis. The results of the DNA fragmentation assay and morphological observation of cells using propidium iodide showed that, in agreement with a previous report [14], C 2 -, and C 8 -ceramide both induced apoptosis in PC12 cells. Although DAP kinase was expressed in PC12 cells, both the total a nd Ca 2+ /calmodulin-independent DAP kinase activ- ities were very low under normal growth conditions. However, in the presence of C 2 -orC 8 -ceramide, the total and Ca 2+ /calmodulin-independent activities were increased in a c oncentration-dependent manner b y a maximum of 13-fold and ®vefold, r espectively, after 10 min of exposure, then gradually decreased, although s till showing a twofold to fourfold increase after 90 min o f exposure. In addition, the Ca 2+ /calmodulin-independent activity/total activity ratio also increased from 0.2 to 0.6 following exposure to C 2 -orC 8 -ceramide, these increases being maintained at 0.4± 0.5 even after 90 min. Similar results have been obtained with Ca 2+ /calmodulin-dependent protein kinase IV ( CaM kinase IV), one of the member of the CaM kinase family [24]. It is thought that CaM kinase IV requires phospho- rylation by a CaM kinase kinase, which is Ca 2+ /calmod- ulin-responsive, for full activation. A previous study using Jurkat T lymphocytes showed that both the total and Ca 2+ / calmodulin-independent CaM kinase IV activities were increased by treatment with ant i-CD3 Ig, which triggers an intracellular Ca 2+ increase, a nd that the Ca 2+ /calmodulin- independent activity/total activity ratio was increased. These results suggest that CaM kinase IV is activated through a CaM kinase kinase cascade triggered by a n increase in intracellular Ca 2+ levels, and as CaM kinase IV activated by CaM kinase k inase has signi®cant Ca 2+ /calmodulin- independent activity, it has the potential for prolonged activation. It is tempting to speculate that ceramide- activated DAP kinase, which shows signi®cant Ca 2+ / calmodulin-independent activity, has the potential for prolonged activation and subsequently results in apoptosis in PC12 cells. This idea was supported by the results of transfection s tudies showing that the K42A-DAP kinase mutant failed to induce apoptosis in the presence of ceramide, that the DCaM-DAP kinase mutant induced apoptosis even in the absence of ceramide, and that PC12 cells overexpressing DAP kinase showed higher sensitivity to ceramide than untransfected PC12 cells. Thus, the intrinsic kinase activity of DAP kinase is critical for apoptosis. The mechanism of DAP kinase activation has not been clear. DAP kinase undergoes phosphorylation and the intrinsic activity is stimulated by Ca 2+ /calmodulin [1,4]. However, in accor dance with a p revious observation in C 6 -ceramide-treated U937 cells [25], permeant-ceramide- treated Jurkat T cells [26] and foreskin ®broblasts [27], we could not detect any C 2 -orC 8 -ceramide-induced increase in intracellular Ca 2+ level using calcium ¯uorometry and a fura-2 ¯uorescent probe (data not shown). T his is also consistent with the result that c eramide-treated PC12 cells had signi®cant Ca 2+ /calmodulin-independent DAP kinase Fig. 6. Viability of PC12 cells expressing wild-type and DAP kinase mutants after ceramide exposure. Re sistance to ceramide-induced apoptosis was quanti®ed by c ounting the number of morphologically intact EGFP-positive cells u nder a ¯uorescence m icroscop e. PC12 cells were transfected with expression plasmids. The transfection eciency, about 20%, is con®rmed to be similar i n all experime nts by cou nting the nu mber of EGF P-positive cells at 8 h after DNA transfection. Then, after 48 h, cells were incubated for 24 h in the absenc e of ceramide (white column) or in the prese nce of 30 l M C 2 -ceramide (black column), C 8 -ceramide (grey column), or C 2 -dihydroceramide (striped column) and the number of the surviving EGFP- positive cells were counted. The percentage of cell viability is caluculated b y dividing t he number of EGFP-positive cells in each experimental conditions by the number of EGFP-positive cells transfected with pcDNA3 alone in t he absence of ceramide. The data a re the mean  SD of seven randomly selected ®elds in t hree independent experiments. Ó FEBS 2002 DAP kinase as ceramide-induced apoptotic component (Eur. J. Biochem. 269) 145 activity. Thus, one possible mechanism of DAP kinase activation is a change in phosphorylation state (i.e. phos- phorylation and dephosphorylation). In the ceramide- induced apoptosis pathway, other molecules t hat are Ca 2+ /calmodulin-nonresponsive and activate DAP kinase in a ceramide-dependent fashion may exist. Further exper- iments will be required to elucidate the DAP kinase activation mechanism. In addition, the apoptotic function of DAP kinase i s thought to be modulated t hrough other functional domains. DAP kinase contains eight ankyrin repeats, two potential ATP/GTP binding sites (P-loops), a cytoskeleton-binding domain, a death domain, and a serine-rich C-terminus tail [4,5]. A previous study showed that deletion of the death domain abrogates the apoptotic function and t hat overex- pression of the death domain protects HEK293 and HeLa cells from TNF-a-, Fas-, and FADD/MORT1-induced apoptosis [6]. In th e present study, e ctopic expression of the death domain or the K42A-DAP kinase mutant in PC12 cells suppressed ceramide-induced apoptosis; however, the protection effect of the death domain was weaker than that of the K42A-DAP kinase mutant. This result is in agreement with previous observations using HEK293 and HeLa cel ls [6]. Treatment with ceramide results in ac tivation of t he CD95 (APO-1/Fas) signaling cascade [28], the stress- activated p rotein kinase (S APK/JNK) signaling cascade [29], and the p53 signaling cascade [30,31]. However, the mechanism i nvolved in C 2 -ceramide-induced apoptosis is controversial. One possible explanation is as follows. As C 2 -ceramide is cell-permeable, it probably perturbs the membrane structure and increases membrane permeability [32±34]. Changes in membrane permeability, particularly in mitochondria, are involved in the triggering of cell death. In mitochondria, C 2 -ceramide triggers the formation of reac- tive oxygen species, the disruption of electron transport and energy metabolism, and the release of caspase-activating proteins, c ytochrome c, a nd apoptosis-inducing factor, and then activates the caspase family of proteases [35]. Recent studies have demonstrated that C 2 -ceramide-induced apop- tosis is required for caspase activation [36±38]. It has also emerged that C 2 -ceramide-induced apoptosis in PC12 cells is prevented not only by a caspase inhibitor, but also by neurotrophic factors, such as nerve growth factor (NGF) and basic ®broblast growth factor (bFGF). These protective effects exerted by NGF an d bFGF are independent of either the extracellular signal-regulated k inase (ERK) or phos- phatidylinositol 3 k inase (PI3 k inase) cascade, both of which are known to be cell survival-promoting signal pathways [36]. Thus, it is f easible that an as yet undescribed apoptotic pathway, which may include DAP kinase, is antagonized in neurotrophic factor-dependent r escue from C 2 -ceramide- induced apoptosis. Mitocondrial permeability transitions and the subse - quent activation of caspases also t ake place under ischemic conditions [39,40]. We have previously demonstrated that the expression of DAP kinase mRNA is increased prior to cell death resulting from transient forebrain ischemia [8]. The apoptotic function of the DCaM-DAP kinase mutant is blocked by overexpression of natural caspase inhibitors, i.e. CrmA and p35, but not by overexpression of a dominant n egative caspase-8 mutant, indicating that DAP kinase functions downstream of caspase-8 and upstream of some members of the caspase family other than caspase-8 [6]. A recent study by Raveh et al. [41] demonstrated DAP kinase activates p53 via p19 ARF and suppresses oncogenic transformation. Taking together these results and those of the present study, it is thus conceivable that DAP kinase participates in a novel cascade involved not only in C 2 -ceramide-induced apoptosis, but also in physiologically induced apoptosis, as a result of caspase activation, and that DAP kinase activity is a key element leading to apoptosis. ACKNOWLEDGEMENTS We thank Dr Adi Kimchi (Weizmann Institute of Science , Israe l) for providing helpful advice and DAP kinase expression plasmids, Dr Xiaofen Sun and Mr Christopher Booth for assistance in cell viability assay. REFERENCES 1. Deiss,L.,Feinstein,E.,Berissi,H.,Cohen,O.&Kimchi,A.(1995) Identi®cation of a nove l serine/threonine kinase and a novel 15-kD protein as potential mediators of the c interferon-induced cell death. Genes Dev. 9, 15±30. 2. Kimchi, A. (1998) D AP genes: novel apoptotic genes isolated by a functional approach to gene cloning. Biochim. Biop hys. Acta. 1377, F13±F33. 3. Kissil, J.L. & Kim chi, A. (1998) D eath-associate d proteins: from gene ide nti®cation to the an alysis of their a poptotic and tumour suppres sive functions. Mol. Med. Today 4, 268±274. 4. Cohen, O., Feinstein, F. & Kimchi, A. (1997) DAP-kinase is a Ca 2+ /calmodulin-dependent, cytoskeletal-associated protein kinase, with cell death-inducing functions that depend on its catalytic activity. EMBO J. 16, 998±1008. 5. Feinstein, E., Wallach, D., Boldin, M., Varfolomeev, E. & Kimchi, A. (1995) The death domain: a module shared by proteins with diverse cellular functions. Trends Biochem. Sci. 20 , 342±344. 6. Cohen, O., Inbal, B ., Kissil, J.L., Raveh, T., Berissi, H., Spivak- Kroizaman,T.,Feinstein,E.&Kimchi,A.(1999)DAP-kinase participa tes in TNF-a- and Fas-induced apoptosis and its function requires the death domain. J. Ce ll Biol. 146, 141±148. 7. Raveh,T.,Berissi,H.,Eisenstein,M.,Spivak,T.&Kimchi,A. (2000) A functional genetic screen identi®es regions at the C -ter- minal tail and death-domain of death-associated p rotein kinase that are critical for its proapoptotic activity. Proc.NatlAcad.Sci. USA 97, 1572±1577. 8. Yamamoto, M., Takahashi, H., Nakamura, T., Hioki, T., Nagayama, S., Ooashi, N., S un, X., Ishii, T., Kudo, Y ., Nakajima- Iijima, S., Kimchi, A. & Uchino, S. (1999) Developmental changes in distribution of death-associated protein kinase mRNAs. J. Neurosci . Res. 58, 674±683. 9. Hannun, Y.A. & O beid, L.M. (1995) Ceramide: an intracellular signal for apoptosis. Trends Biochem. Sci. 20, 73±77. 10. Spiegel, S., Foster, D. & Kolesnick, R. (1996) Signal transduction through lipid second messengers. Curr. Opin. Cell Biol. 8, 159±167. 11. Brugg, B., Michel, P.P., Agid, Y. & Ruberg, M. (1996) Ceramide induces apoptosis in cultured mesencephalic neurons. J. Neuro- chem. 66, 7 33±739. 12. Ping, S.E. & Barrett, G.L. (1998) Ceramide can induce cell death in sensory neurons, whereas ceramide analogues and sphingosine promote survival. J. Neurosci. Res. 54 , 206±213. 13. Taniwaki,T.,Yamada,T.,Asahara,H.,Ohyagi,Y.&Kira,J. (1999) Ceramide induces a poptosis to immature cerebellar granule cells in culture. Neurochem. Res. 24 , 685±690. 14. Hart®eld, P.J., Mayne, G.C. & Murray, A.W. (1997) Ceramide induces apoptosis in PC12 cells. FEBS L ett. 401, 1 48±152. 146 M. Yamamoto et al. (Eur. J. Biochem. 269) Ó FEBS 2002 15. G reene, L.A. & Tischler, A.S. (1 982) PC12 pheochromocytom a cultures in neurobiological re search. Adv. Cell Neurobiol. 3, 373±414. 16. Saito, S., Watabe, S., Ozaki, H., Fusetani, N. & Karaki, H. (1994) Mycalolide B, a novel actin depolymerizing agent. J. Biol. Chem. 269, 29710±29714. 17. Enslen, H., Sun, P., Brickey, D., Soderling, S.H., Klamo, E. & Soderling, T.R . (1994) C haracterization o f Ca 2+ /calmodulin- dependent protein kinase IV. J. Biol. Chem. 269, 15520±15527. 18. DeRemer, M .F., Saeli, R.J. & Edelman, A.M. (1992) Ca 2+ - calmodulin-dependent protein kinases Ia and Ib from Rat Brain. J. Biol. Cehm. 267, 13460±13471. 19. H annun, Y.A . (1994) The sphin gomyelin cy cle and the second messenger function of ceramide. J. Biol. C hem. 269, 3125±3128. 20. Kawai,T.,Matsumoto,M.,Takeda,K.,Sanjo,H.&Akira,S. (1998) ZIP kinase, a n ovel se rine/threo nine kinase wh ich mediate s apoptosis. Mol. Cell. Biol. 18, 1642±1651. 21. K o È gel, D., Plo È ttner, O., Landsb erg, G., C hristian, S . & S cheidt- mann, K.H. (1998) Cloning a nd characterization of Dlk, a novel serine/threonine kinase that is tightly associated with chromatin and phosphorylates core histones. Oncogene 17, 2645±2654. 22. Murai-Hori, M., Suizu, F., Iwasaki, T., Kikuchi, A . & Hosoya, H. (1999) ZIP kinase identi®ed as a novel my osin regulatory light chain kinase in H eLa cells. FEBS Lett. 451, 81±84. 23. Inbal,B.,Cohen,O.,Polak-Charcon,S.,Kopolovic,J.,Vadai,E., Eisenbach,L.&Kimchi,A.(1997)DAPkinaselinksthecontrolof apoptosis to metastasis. Nature 390, 180±184. 24. Park, I K. & Soderling, T.R. (1995) Activation of Ca 2+ / calmodulin-dependent protein kinase (CaM-kinase) IV by CaM-kinase kinase in Jurkat T lymphocytes. J. Biol. Chem. 270, 30464±30469. 25. Quillet-Mary, A ., Jare  zou, J P., Mansat, V., Bordier, C., Naval, J. & Laurent, G. (1997) Implication of mitochondrial hydrogen peroxide ge neration in ceramide-induced apoptosis. J. Biol. Chem. 272, 21388±21395. 26. Breittmayer, J P., Bernard, A. & Aussel, C. (1994) Regulation by sphingomyelinase and sphingosine of Ca 2+ signals elicited by CD3 monoclonal antibody, thapsigargin, or ionomycin in the Jurkat T cell line. J. Biol. Chem. 269, 5054±5058. 27. Chao, C.P., Laulederkind, S.J.F. & Ballou, L.R. (1994) Sphingosine-mediated phosphatidylinositol metabolism and cal- cium mobilization. J. Biol. Chem. 269, 5 849±5856. 28. Herr,I.,Wilhelm,D.,Bo È hler, T., Angel, P. & Debatin, K M. (1997) Activation of CD95 (APO-1/Fas) signaling by ceramide mediates cancer therapy-induced apoptosis. EMBO J. 16 , 6200± 6208. 29. Verheij, M., Bose, R., Lin, X.H., Yao, B., J arvis, W.D., G rant, S., Birrer, M.J., Szabo, E., Zon, L.I., Kyriakis, J.M., Haimovitz- Friedman, A., Fuks, Z. & Kolesnick, R.N. (1996) Requirement for ceramide-initiated SAPK/JNK signalling i n stress-induced apoptosis. Nature 380, 75±79. 30. Pruschy,M.,Resch,H.,Shi,Y Q.,Aalame,N.,Glanzmann,C.& Bodis, S. (1998) Ceramide triggers p53-dependent apoptosis in genetically de®ned ®brosarcoma tumour cells. Br. J. Cancer 80, 693±698. 31. Lo  pez-Marure, R., Ventura, J.L., Sa  nchez, L., Montan Ä o, L.F. & Zentella, A. (2000) Ceramide mimics tumour ne crosis factor-a in the induction of cell cycle arrest in endothelial cells. Eur. J. Biochem. 267, 4325±4333. 32. Ruiz-Argu È ello, M.B., Basa  nez, G., Goni, F.M. & Alonoso, A. (1996) Dierent eects of enzyme-genera ted ceramides and di a- cylglycerols in phospholipid membrane fusion and leakage. J. Biol. Chem. 271, 26616±26621. 33. Scadi, C., Schmitz, I., Zha, J., Korsmeyer, S.J., Krammer, P.H. & P eter, M .E. ( 1999) D ierential modula tion of apoptosis sensi- tivity in CD95 type I and type II cells. J. Biol. Chem. 274, 22532± 22538. 34. Siski nd, L.J. & Colombini, M. (2000) The lipids C 2 -and C 16 -ceramide form large stable channels. J. Biol. Chem. 275 , 38640±38644. 35. Green, D.R. & Reed, J.C. (1998) Mitochondria and apoptosis. Science 281, 1309±1312. 36. Hart®eld, P.J ., Bilney, A.J. & Murray, A.W. (1998) Neurotrophic factors prevent ceramide-induced apoptosis downstream of c-Jun N-terminal kinase activation in PC12 cells. J. Neurochem. 71, 161±169. 37. Mizushima, N., Koike, R., K ohsaka, H., Kushi, Y., Handa, S., Yagita, H. & Miyasaka, N . (1996) Ceramide induces apoptosis via CPP32 activation. FEBS Lett. 395, 267±271. 38. Yoshimura, S., Banno, Y., Nakashima, S., Takenaka, K., Sa kai, H., Nishimura, Y., Sakai, N., Shimizu, S., Eguchi, Y., Tsujimoto, Y. & Nozawa, Y. (1998) Ceramide forma tion leads to caspase-3 activation during hypoxic PC12 cell death. J. Biol. Chem. 273, 6921±6927. 39. Saikumar, P., Dong, Z., Weinberg, J.M. & Venkatachalam, M.A. (1998) Mechanisms of cell death in hypoxia/reoxygenation injury. Oncogene 17, 3341±3349. 40. Krajewski, S., Krajewska, M., Ellerby, L.M., Welsh, K., X ie, Z., Deveraux, Q.L., Salvesen, G.S., Bredesen, D.E., Rosenthal, R.E., Fiskum, G . & Reed, J.C. (1999) Release o f caspase-9 from mito- chondria during neuronal apop tosis and c erebral ischemia. Proc. Natl Acad. Sci. USA 96, 5752±5757. 41. Raveh, T., Droguett, G., Horwitz, M.S., Depinh o, R.A. & Kimchi, A. (2001) DAP kinase activates a p19 ARF /p53-mediated apoptotic checkpoint to suppress oncogenic transformation. Nat. Cell Biol. 3, 1±7. Ó FEBS 2002 DAP kinase as ceramide-induced apoptotic component (Eur. J. Biochem. 269) 147 . involved in ceramide- induced apoptosis, and that the intrinsic DAP kinase activity is critical for ceramide-induced apoptosis. Keywords: apoptosis; DAP kinase; . DAP kinase activity is critical for C 2 -ceramide-induced apoptosis in PC12 cells Mutsuya Yamamoto, Takeshi Hioki, Takehisa Ishii, Sadayo Nakajima-Iijima