Activation of Acid Sphingomyelinase by Protein Kinase Cδ-mediated Phosphorylation*

Although important for cellular stress signaling pathways, the molecular mechanisms of acid sphingomyelinase (ASMase) activation remain poorly understood. Previous studies showed that treatment of MCF-7 mammary carcinoma cells with the potent protein kinase C (PKC) agonist, phorbol 12-myristate 13-acetate (PMA), induces a transient drop in sphingomyelin concomitant with an increase in cellular ceramide levels (Becker, K. P., Kitatani, K., Idkowiak-Baldys, J., Bielawski, J., and Hannun, Y. A. (2005) J. Biol. Chem. 280, 2606-2612). Here we show that PMA selectively activates ASMase and that ASMase accounts for the majority of PMA-induced ceramide. Pharmacologic inhibition and RNA interference experiments indicated that the novel PKC, PKCδ, is required for ASMase activation. Immunoprecipitation experiments revealed the formation of a novel PKCδ-ASMase complex after PMA stimulation, and PKCδ was able to phosphorylate ASMase in vitro and in cells. Using site-directed mutagenesis, we identify serine 508 as the key residue phosphorylated in response to PMA. Phosphorylation of Ser508 proved to be an indispensable step for ASMase activation and membrane translocation in response to PMA. The relevance of the proposed mechanism of ASMase regulation is further validated in a model of UV radiation. UV radiation also induced phosphorylation of ASMase at serine 508. Moreover, when transiently overexpressed, ASMaseS508A blocked the ceramide formation after PMA treatment, suggesting a dominant negative function for this mutant. Taken together, these results establish a novel direct biochemical mechanism for ASMase activation in which PKCδ serves as a key upstream kinase.


inhibition and RNA interference experiments indicated that the novel PKC, PKC␦, is required for ASMase activation. Immunoprecipitation experiments revealed the formation of a novel PKC␦-ASMase complex after PMA stimulation, and PKC␦ was able to phosphorylate ASMase in vitro and in cells.
Using site-directed mutagenesis, we identify serine 508 as the key residue phosphorylated in response to PMA. Phosphorylation of Ser 508 proved to be an indispensable step for ASMase activation and membrane translocation in response to PMA. The relevance of the proposed mechanism of ASMase regulation is further validated in a model of UV radiation. UV radiation also induced phosphorylation of ASMase at serine 508. Moreover, when transiently overexpressed, ASMase S508A blocked the ceramide formation after PMA treatment, suggesting a dominant negative function for this mutant. Taken together, these results establish a novel direct biochemical mechanism for ASMase activation in which PKC␦ serves as a key upstream kinase.
Hydrolysis of membrane sphingomyelin (SM) 2 constitutes a major route for ceramide formation in mammalian systems. In contrast to the multistep de novo pathway of ceramide biosynthesis, breakdown of SM is catalyzed by a single class of enzymes, the sphingomyelinases (SMases). Currently, the NCBI data base contains five entries for human SMases that are grouped according to their optimal pH as acid (ASMase or SMPD1), neutral (NMase or SMPD2, -3, and -4), and alkaline sphingomyelinases (Alk-SMase or ENPP7).
Acid sphingomyelinase (ASMase) was the first SMase to be purified and cloned (2)(3)(4). Although encoded by the same gene and mRNA, two forms of the enzyme have been described: an endolysosomal form (or Zn 2ϩ -independent form) and a secretory form (or Zn 2ϩ -dependent form) (5). Post-translational protein glycosylation plays an important role in both trafficking and stability of ASMase within the lysosomal milieu (6,7). In addition, cysteine 629 situated toward the C terminus of ASMase has been proposed as an important residue for regulation of the enzyme such that modification (mutation or deletion) of this residue results in 5-fold activation of the enzyme (8).
In humans, mutation of the SMPD1 gene results in Niemann-Pick disease, an autosomal recessive neurovisceral disease of high lethality. The severity of the Niemann-Pick disease phenotype correlates well with the deficiency in the in vitro ASMase activity, which has become a standard test for classification of patients with the disease. Indeed, the knock-out mouse in the SMPD1 gene recapitulates the Niemann-Pick disease phenotype as described by multiple investigators (9,10).
Biologically, ASMase has been implicated in a variety of physiological and pathophysiologic processes (reviewed in Refs. [11][12][13][14]. In particular, ASMase has been shown to be activated in response to various stress stimuli including 1) ligation of death receptors (tumor necrosis factor-␣, CD95, and TRAIL), 2) radiation (UV-C and ionizing radiation), 3) chemotherapeutic agents (cisplatin, doxorubicin, paclitaxel, and histone deacetylase inhibitors), 4) viral, bacterial, and parasitic pathogens (rhinoviruses, Neisseria gonorrhea, Staphylococcus aureus, Pseudomonas aeruginosa, and Cryptosporidium parvum), and 5) cytokines (e.g. IL-1␤). Evidence from studies conducted in cell culture and animal models suggested that activation of ASMase by the above agents is a key step during cellular differentiation, growth arrest, apoptosis, and immune defense mechanisms. Although most of these studies describe rapid activation of the enzyme (most likely via a post-translational modification), transcriptional change is yet another regulator of ASMase activity. For instance, differentiation of monocytic cells to macrophages requires up-regulation of the enzyme at a transcriptional level (15). Transcription factors SP1 and AP-2 mediate the ASMase response during monocytic differentiation. Later studies demonstrated that AP-2 also up-regulates ASMase in leukemia cells treated with retinoic acid (16). More recently, ASMase was implicated in maintenance of calcium homeostasis in mouse and cell culture models (17,18). In particular, cerebellar deficits in the Niemann-Pick mouse were reported to be associated with defects in the calcium transport machinery in the endoplasmic reticulum.
Thus, although the scientific literature is replete with reports implicating ASMase in various cellular responses, the biochemical and molecular mechanisms of ASMase regulation remain unclear. Interestingly, recent studies suggest regulation of ASMase by redox mechanisms. For instance, the death ligand TRAIL signals ASMase activation through reactive oxygen species production (19). Additional mechanisms proposed to regulate ASMase enzymatic activity include interaction with lysosomal anionic lipids (bismonoacylglycerophosphate) and sphingolipid activator protein SAP-C (20). Nevertheless, the proximal mechanisms involved in direct regulation of ASMase remain undefined.
In a previous study, we identified a novel mechanism of regulation of ceramide formation by protein kinase C (PKC) through activation of the salvage pathway (1). In this model, treatment of MCF-7 breast cancer cells with PMA, a potent PKC agonist, resulted in a significant increase in cellular ceramide concomitant with a drop in sphingomyelin levels. Interestingly, this response was blocked upon treatment with the ceramide synthase inhibitor, fumonisin B1, but not an inhibitor of serine palmitoyltransferase, myriocin, thus ruling out involvement of the de novo pathway.
In this study, we demonstrate that phorbol esters induce activation of ASMase through PKC␦-dependent phosphorylation. Using site-directed mutagenesis, it is shown that PMA induces phosphorylation of ASMase at Ser 508 , and this phosphorylation is indispensable for activation of the enzyme. Moreover, this phosphorylation event appears to be required for translocation of ASMase to the plasma membrane. The significance of Ser 508 phosphorylation and the role of PKC␦-ASMase association are further confirmed in the context of agonist-driven ASMase activation induced by UV radiation.

EXPERIMENTAL PROCEDURES
Materials-Cell culture material, including RPMI medium, fetal bovine serum (FBS), phosphate-free medium, and dialyzed serum, were from Invitrogen. Bovine sphingomyelin, phosphatidylserine, and diacylglycerol were from Avanti polar lipids (Alabaster, AL). [choline-methyl-14 C]Sphingomyelin was kindly provided by Dr. Alicja Bielawska (Medical University of South Carolina, Charleston, SC). Antibodies against PKC␦ (polyclonal) and LAMP-1 were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Monoclonal antibodies against the V5 epitope and GFP were from Invitrogen. Anti-phosphoserine (polyclonal) was from Zymed Laboratories Inc. (San Francisco, CA). Purified ASMase was a kind gift from Dr. Gary Smith (Glaxosmithkline). The PKC antibody sampler kit was purchased from BD Biosciences. The endoglycosidase H digestion kit and all other materials were from Sigma.
Cell Lines and Culture Conditions-MCF-7 cells were originally purchased from ATCC (Manassas, VA). Cells were maintained in RPMI 1640 supplemented with 10% FBS at 37°C in a 5% CO 2 incubator. Where indicated, cells were shifted to a phosphate-free medium. Testing for the presence of mycoplasma infections was performed routinely on a monthly basis.
UV Irradiation-MCF-7 cells grown on 10-cm dishes were coated with a thin layer of phosphate-buffered saline prior to irradiation. Radiation was performed using a GS Gene Linker UV chamber (Bio-Rad), which emits UV-C light ( avg ϭ 254 nm) at a dose of 50 J/m 2 . Cells were then reincubated with RPMI medium and then collected for analysis at the indicated time(s) postirradiation.
ASMase Peptide and Antiserum Production-Development of ASMase antiserum was conducted at the Sigma Genosys facility. The ASMase peptide sequence was analyzed using the PCGENE program. The program selects regions of a protein that are candidates for raising antibodies, based on antigenicity, flexibility, and ␤-turn. Based on this analysis, the following peptide sequence was chosen and synthesized: CVGELQAAEDRGDKV (15 mer, residues 361-374). This sequence received the following scores: antigenicity ϭ 2.5; flexibility ϭ 1.064; ␤-turn ϭ 2.4. The peptide was conjugated with keyhole limpet hemocyanin before immunization of two rabbits. The third bleed was affinity-purified, and specificity of the antibody was verified by enzyme-linked immunosorbent assay, Western blotting, and immunofluorescence.
SMase Assays-In vitro enzymatic assays for acid and neutral sphingomyelinases were performed using [choline-methyl- 14 C]sphingomyelin. The assays were performed as previously reported (21,22).
RNA Interference-Gene silencing of human ASMase and PKC␦ was performed essentially according to a standard protocol (21). Sequence-specific small interfering RNA reagents were purchased from Qiagen (Valencia, CA). The sequences of the sense and antisense small interfering RNAs are shown in Table 1. The specificity of the RNAi was verified by sequence comparison with the human genome data base using the NIH Blast program.
In Vitro Kinase Assays-In vitro phosphorylation reactions were performed as follows: 6 g of purified ASMase were incubated with 25 ng of recombinant PKC␦ (Calbiochem) at 30°C in a kinase buffer (20 mM HEPES, pH 7.4, 10 mM MgCl 2 , 0.1 mM EGTA, 0.1 mM ATP), 0.5 Ci of [␥-32 P]ATP, and where indicated, the buffer was supplemented with lipid-detergent mixed micelles consisting of 100 g/ml phosphatidylserine and 10 g/ml diacylglycerol sonicated in 0.3% Triton X-100 solution. At the indicated time points, reactions were terminated by the addition of SDS sample buffer, and samples were boiled for 5 min. Samples were electrophoresed by SDS-PAGE, and then the gels were dried. ASMase phosphorylation was detected by autoradiography.
In Vivo Labeling of ASMase-MCF-7 cells plated in 10-cm dishes (250 ϫ 10 3 cells/plate) were transiently transfected with V5-ASMase. Cells were incubated in phosphate-free RPMI medium supplemented with 5% dialyzed FBS for 1 h. Then cells were shifted to the labeling medium containing 0.2 mCi/ml [ 32 P]orthophosphate and incubated for an additional 4 h. Treatment with either PMA or Me 2 SO was performed in phosphate-free medium for 1 h. ASMase was immunoprecipitated via the V5 epitope and loaded on SDS-PAGE. Incorporation of 32 P into ASMase was detected by autoradiography after gel drying.
Confocal Microscopy-Approximately 5 ϫ 10 5 cells were seeded to a 2-cm poly-L-lysine-coated confocal plate (MatTek Corp.). After treatment, cells were fixed with 4% paraformaldehyde for 10 min and permeabilized with methanol for 5 min. Blocking was performed in 2.5% FBS solution. Primary antibodies were diluted in a solution of 1.5% FBS, 0.15% saponin and incubated for 3 h. The samples were then washed with 1.5% FBS solution three times. This was followed by incubation with secondary antibodies for 1 h at room temperature. Samples were stored at 4°C until image acquisition. Images were acquired using a Zeiss laser-scanning confocal microscope (LSM 510). Excitation wavelengths 488, 543, and 633 were used. Images were acquired at the equatorial plane of monolayer cells guided by DRAQ5 (Alexis) nuclear stain. For in vivo imaging, cells were fitted in a special chamber under controlled temperature and CO 2 settings.
Mass Spectroscopy for Ceramide and Sphingomyelin-Sphingolipid analysis was performed using electrospray ionization/ tandem mass spectrometry on a Thermo Finnigan TSQ 7000 triple quadrupole mass spectrometer, operating in a multiple reaction-monitoring positive ionization mode. This method has been recently described (23).
Plasmid Constructs and Overexpression-Original ASMase cDNA was a kind gift from Dr. Ed Schuchman (Mount Sinai School of Medicine). The cDNA (2.37 kb) encoding human ASMase was excised from the original PSVK3 plasmid using EcoRI and EcoRV restriction enzymes. After amplification by conventional PCR, the cDNA was ligated into pEF6/V5-His-TOPO vector (Invitrogen). The V5 epitope is situated at the carboxyl terminus of ASMase. Complementary DNA for PKC␦ was ligated into pEGFP-C3 vector (Clontech). The restriction sites used were XhoI and KpnI. The GFP epitope is situated at the amino terminus of PKC␦. The integrity of the new plasmids was verified by DNA sequencing. For transient overexpression experiments, endotoxin-free plasmids were transfected into MCF-7 cells using Effectene (Qiagen) according to the manufacturer's recommendations.
Real Time PCR Analysis (Reverse Transcription-PCR/Quantitative PCR)-DNA-free messenger RNA from MCF-7 cells was isolated using the Qiagen minikit for mRNA extraction. Complementary DNA was synthesized from 5 g of total RNA using the reverse transcriptase kit from Promega (Madison, WI). Real time reverse transcription-PCR was performed on an iCycler system from Bio-Rad. The sample real time reverse transcription-PCR reaction volume was 25 l, including 12.5 l of SYBR Green PCR reagents (Qiagen, Valencia, CA), 7.5 l of cDNA template, and 5 l of the diluted forward and reverse primers (3 M). First steps of reverse transcription-PCR were 2 min at 50°C followed by a 10-min hold at 95°C. Cycles (n ϭ 40) consisted of a 15-s melt at 95°C, followed by a 1-min annealing/ extension at 60°C. The final step was a 60°C incubation for 1 min. Reactions were performed in triplicate. The threshold for cycle of threshold (Ct) analysis of all samples was set at 0.15 relative fluorescence units. The data were normalized to levels of an internal control gene, ␤-actin. The reverse transcription primers used in this study were as follows: ␤-actin (forward, 5Ј-ATTGGCAATGAGCGGTTCC-3Ј; reverse, 5Ј-GGTAGT-TTCGTGGATGCCACA-3Ј) and ASMase (forward, 5Ј-TGG-CTCTATGAAGCGATGGC-3Ј; reverse, 5Ј-TTGAGAGAGA-TGAGGCGGAGAC).
Deglycosylation with Endoglycosidase H-Treatment of immunoprecipitated ASMase with endoglycosidase H (Sigma) was performed according to the instructions of the manufacturer. Briefly, samples were boiled in denaturation buffer containing 50 mM ␤-mercaptoethanol and 0.1% SDS for 5 min. Samples were cooled on ice for another 5 min. Then endoglycosidase was added (0.01 units/sample), and the reaction was carried at 37°C for 12 h in a reaction buffer consisting of 50 mM sodium phosphate, pH 5.5.
Statistical Analysis-Mann-Whitney or Student's t tests were performed between control and treated states and/or between treatment and treatment plus RNAi-mediated inhibition states in a minimum of three independent experiments. A p value of 0.05 or less is considered as statistically significant and marked in the figures with an asterisk.

PMA Induces Activation and Membrane Translocation of
ASMase-Previously, we reported that PMA treatment induces an elevation in ceramide levels that persisted after blocking the de novo pathway with myriocin but was inhibited by fumonisin B1, suggesting operation of the salvage pathway (resynthesis of ceramide from sphingosine derived from the breakdown of complex sphingolipids but not from de novo synthesis) (1). The concomitant drop in sphingomyelin levels induced by PMA treatment, therefore, prompted us to investigate the potential involvement of SMases. As shown in Fig. 1A, treatment of MCF-7 cells with 100 nM PMA resulted in transient activation of ASMase, with the maximum in vitro activity detected after 60 min (Fig. 1A). Since diacylglycerol has been suggested to activate ASMase directly (24), the effect of direct addition of PMA into the enzyme reaction buffer (post-cell lysis) was evaluated; however, PMA failed to activate ASMase directly (data not shown), suggesting that the stimulation process was indirect. Moreover, no change in total cellular neutral SMase activity was detected within 60 min of PMA treatment (Fig. 1A).
Next, we examined whether activation of ASMase was associated with a change in its subcellular localization. For these studies, a novel polyclonal antibody that recognizes ASMase was employed. After affinity purification of third bleed serum, the specificity of the antibody was tested. Western blot analysis of liver homogenates derived from WT and ASMase Ϫ/Ϫ mice revealed that the new antibody detected the 72-kDa form of ASMase (supplemental Fig. 1A). Moreover, the antibody proved useful for immunofluorescence applications, as indi-cated by colocalization of its signal with the lysosomal marker LAMP-1 (lysosomal-associated membrane protein 1) (supplemental Fig. 1B).
As expected, under basal conditions, most of cellular ASMase was concentrated within acidic compartments, as evidenced by colocalization with the acidophilic dye lysotracker red. There was no change in ASMase staining acutely after PMA treatment (5 min). However, confocal microscopy imaging revealed translocation of ASMase from endolysosomes to the plasma membrane after sustained PMA treatment (100 nM, 60 min) (Fig. 1B). Therefore, these data demonstrate that the stimulatory effect of PMA on ASMase is associated with relocation of the enzyme to the plasma membrane.   Table 1. ␤-Actin was used as internal control, and results were normalized to the amount of ␤-actin. B, in vitro ASMase activity after knockdown using RNAi technology. Cells were transfected with a 10 nM concentration of either SCR or ASMase RNAi (sequence 1) for 48 or 72 h. Lysates (100 g) from each treatment group were subject to an ASMase activity assay. C, effect of ASMase knockdown on PMA-induced ceramide formation. MCF-7 cells plated at 60% confluence were transfected with either ASMase RNAi or SCR sequence for 48 h. Consequently, cells were treated with either PMA (100 nM, 1 h) or Me 2 SO. The different treatment groups were subjected to mass spectrometric analysis as indicated under "Experimental Procedures." Ceramide measurements were normalized to total phospholipids. Quantitative PCR measurements are averages of two experiments. Ceramide and SMase activity results were performed in triplicate. Shown are averages Ϯ S.E. (p Ͻ 0.05). ns, not significant; CN, control.
Requirement for ASMase in PMA-induced Ceramide Formation-The contribution of ASMase to the observed accumulation of ceramide following PMA treatment was consequently investigated. Although tricyclic amines (and other closely related compounds) have been routinely used to block ASMase, our recent data showed that the action of these compounds is not restricted to ASMase (21). Therefore, knockdown of ASMase was attempted using two different RNAi sequences. Real time PCR analysis was employed to analyze levels of ASMase mRNA (with ␤-actin as internal control) 48 h after transfection of two different RNAi sequences. As shown in Fig.  2A, successful knockdown of ASMase mRNA was achieved with one of the two RNAi sequences (sequence 1) at a concentration as low as 5 nM. Knockdown of ASMase was further confirmed by in vitro ASMase enzymatic assay. Transfection of sequence 1 (5 nM) resulted in about 70% reduction in total ASMase activity observed at 48 and 72 h post-transfection (Fig.  2B). Therefore, cells were transfected with either scrambled (SCR) or ASMase RNAi (5 nM for 48 h) prior to treatment with PMA. Mass spectrometric analysis showed a doubling in total cellular ceramide levels in response to PMA in cells pretreated with SCR sequence. This response was significantly blocked in cells where ASMase was knocked down (Fig. 2C). Taken together, these results suggest that ASMase is required for PMA-induced ceramide formation in the salvage pathway.
Role of PKC␦ in Activation of ASMase-Phorbol esters signal primarily through proteins with C1 domains, of which the classical and novel PKCs are the most studied. Therefore, to investigate the possible involvement of PKCs in ASMase activation, a panel of pharmacologic inhibitors that block different classes of PKCs was employed. In vitro SMase assays showed that pretreatment with either bisindoleilamide, an inhibitor of classical and novel PKCs, or rottlerin, an inhibitor of novel PKC␦, blocked the ASMase activation in response to PMA. However, the classical PKC inhibitor GÖ 6976 did not significantly affect ASMase activity (Fig. 3A). Based on the above results and on published studies on the role of PKC␦ in mediating stress responses, we hypothesized that the novel PKC␦ may mediate activation of ASMase. RNAi knockdown of PKC␦ was therefore employed. Transfection of a specific PKC␦ RNAi sequence (Table 1) at a dose of 5 nM for 48 h resulted in substantial down-regulation of PKC␦ at the protein level without influenc- SMase assay was conducted in an acidic buffer (pH 4.5), and results are normalized to protein levels. B, knockdown of PKC␦ using RNA interference. Cells were transfected with a 5 nM concentration of either SCR or PKC␦-specific RNAi for 48 h. Cells were collected and lysed, and 30 g of lysates from each sample were loaded to 10% SDS gels. PKC␦, -␣, -, -⑀, and -were detected at 78, 75, 79, 90, and 68 kDa, respectively, using specific polyclonal and monoclonal antibodies. C, effect of PKC␦ knockdown on ASMase activity. After transfection of either SCR or PKC␦ RNAi sequences (48 h, 5 nM), cells were treated with PMA (100 nM, 1 h). After cell lysis, 100 g of each lysate were used for determination of ASMase activity. D, cells were treated as indicated in C. After Bligh and Dyer lipid extraction, ceramide levels were determined by mass spectroscopy. Results are normalized to total phospholipids. Shown are averages of three independent measurements of SMase activity and ceramide levels Ϯ S.E. (p Ͻ 0.05). ns, not significant. E, cells were double transfected with V5-ASMase and PKC␦ RNAi (or SCR) for 48 h. After treatment with either PMA (100 nM) or Me 2 SO for 48 h, cells were stained with a V5 monoclonal antibody (green channel). Nuclei were visualized via DRAQ-5 nuclear stain (red). Results shown are representative of images obtained in three independent experiments. CN, control.
ing total cellular levels of other detected PKC isoforms (Fig. 3B). Interestingly, knockdown of PKC␦ blocked activation of ASMase after PMA treatment but had no effect on basal enzymatic activity (Fig. 3C). In further support of these results, mass spectrometric analysis revealed a proportional drop in ceramide levels after PKC␦ knockdown (Fig. 3D). Additionally, knockdown of PKC␦ blocked the PMA-induced ASMase translocation to the plasma membrane (Fig. 3E). Therefore, these results indicate that PKC␦ lies upstream of ASMase activation.
Early Translocation of PKC␦ to the Plasma Membrane Followed by Partial Localization to the Lysosome-Similar to most other members of the PKC family, translocation of PKC␦ from the cytosol to the plasma membrane is a hallmark of its activation. However, it is worth mentioning that additional sites of PKC␦ translocation have been recently reported, including mitochondria (25,26), Golgi (27), and the nuclear envelope (28,29). Using confocal microscopy imaging of live cells, time-dependent changes in the subcellular localization of PKC␦ were examined after PMA treatment. To that end, MCF-7 cells were transiently transfected with GFP-tagged PKC␦ and labeled with the acidophilic dye lysotracker red. Initially (t ϭ 0), PKC␦ adopted a diffuse cytosolic pattern, whereas lysotracker red concentrated mainly within compartments of different sizes, consistent with endolysosomes. These compartments were not only undergoing fission and fusion but also in continuous contact with the plasma membrane (see supplemental material for movie). As expected and within 1 min of PMA treatment, PKC␦ translocated to the plasma membrane along with clearing of the GFP signal from the cytosol (Fig. 4). In some instances, translocation of PKC␦ was seen as early as 30 s after PMA addition. This localization pattern persisted up to 15 min. By 20 min, PKC␦ was seen to (a) internalize within endocytic compartments that colocalized with lysotracker red and (b) adopt a concomitant perinuclear distribution ( Fig. 4 and supplemental data). It should be also noted that the intensity of the lysotracker red signal was preserved up to 1 h, suggesting that PMA treatment does not profoundly influence lysosomal integrity. Therefore, the above results raised the possibility that in addition to the plasma membrane, PMA treatment induces redistribution of a portion of PKC␦ to ASMase proximal sites, such as endolysosomes.
Formation of a PKC␦-ASMase Complex upon PMA Treatment-Given that PMA treatment induced ASMase activation as well as PKC␦ relocation to endolysosomes within comparable time frames, it became important to determine whether there is direct or indirect protein-protein interaction between PKC␦ and ASMase. To that end, ASMase was immunoprecipitated from lysates of cells treated with either Me 2 SO or PMA. The isolated immune complexes were then tested for the presence of PKC␦ using a polyclonal antibody. Although PKC␦ was absent (or below detection level) in complexes derived from cells treated with a short course of PMA (5 min), PKC␦ coimmunoprecipitated with ASMase after 60 min of PMA treatment (Fig. 5A), corresponding to the time frame in which maximal ASMase activation was detected. These results were further corroborated by transient overexpression of PKC␦ (GFP-tagged) and ASMase (V5-tagged). Association of PKC␦ and ASMase was again observed to be dependent on PMA stimulation (Fig. 5B). The subcellular site of PKC␦-ASMase association was further examined by immunofluorescence. Consistent with the previous result (Fig. 4, third panel), within 30 min of PMA stimulation, ASMase and PKC␦ co-stained in a perinuclear subset of lysosomes, and only minimal ASMase translocation was observed (Fig. 5C). Taken together, these results suggest that sustained stimulation of PKC␦ with PMA induces its translocation to endolysosomes and further association with ASMase.
Phosphorylation of ASMase in Vitro by PKC␦-The above results prompted us to investigate whether ASMase is a potential substrate for PKC␦. To that end, an in vitro phosphorylation assay was performed using recombinant PKC␦ and ASMase proteins. The purity of the recombinant proteins was evaluated by Coomassie Blue gel staining, which revealed only the major bands for ASMase at 72 kDa and PKC␦ at 78 kDa (data not  shown). In the absence of diacylglycerol and PS, only minor incorporation of 32 P into ASMase could be detected (Fig. 6,  top). However, the addition of diacylglycerol (10 g/ml) and PS (100 g/ml) resulted in ASMase phosphorylation within 10 min of coincubation with PKC␦ and was further enhanced after 60 min (Fig. 6, bottom).
PKC␦ Phosphorylates ASMase at Ser 508 -Based on the data above, it became important to investigate whether ASMase is regulated by phosphorylation and to determine if there are specific phosphorylation sites that respond to PKC␦. The human ASMase primary sequence was analyzed using the NetPhos phosphorylation prediction algorithm. Four potential serine phosphorylation residues were found to be conserved in mouse, rat, and Caenorhabditis elegans homologs of ASMase. Interestingly, two of these sites were also reported to be mutated in Niemann-Pick disease patients (30). To investigate phosphorylation of ASMase at these sites, a site-directed mutagenesis approach was employed, where each of the candidate serines was mutated to alanine. After transient transfection into MCF-7 cells, the activity, expression, and localization of all four mutants were determined and compared with wild type (WT) ASMase. In vitro SMase assays and Western blot analysis revealed similar basal activities and degree of expression among the mutant and WT ASMases (Fig.  7A). Furthermore, an endolysosomal staining pattern was observed for these mutants and confirmed by partial colocalization with LAMP-1 (Fig. 7B).
Next, the response of the four Ser 3 Ala mutants to agonistdriven activation of ASMase was examined. Three of the mutants exhibited a response comparable with that of WT ASMase upon stimulation with PMA (100 nM, 1 h). In contrast, mutation of serine 508 to alanine abrogated PMA-induced ASMase activation (Fig. 8A). Serine phosphorylation of WT and the four Ser 3 Ala ASMase mutants was studied next. As shown in Fig.  8B, although PMA treatment induced serine phosphorylation of WT and three of the mutants, ASMase S508A failed to show any change (Fig. 8B). Interestingly, a basal phosphoserine signal could be detected and remained unchanged in all mutants.
Next, 32 P metabolic labeling studies were employed to corroborate these results. However, there was a major drawback for this approach, since initially a very intense signal was detected by autoradiography for both WT and ASMase S508A that masked any differences in peptide phosphorylation. This signal has been examined previously and found to be due to phosphomannose glycosylation of asparagine residues of ASMase (7). Therefore, digestion with endoglycosidase H was employed to remove this phosphomannose. Indeed, endoglycosidase H abrogated this signal (Fig. 3C), supporting the previous findings that the 32 P labeling was preferentially taking place at mannose 6-phosphate (7). A concomitant decrease in molecular mass of about 10 kDa was also detected. In line with previous studies, basal phosphorylation of the ASMase peptide core was still observed after deglycosylation (31). Importantly, the addition of PMA induced a significant increase in the phosphorylation of WT ASMase but not ASMase S508A (Fig. 8C). Moreover, ClustalW analysis of ASMase sequence revealed that Ser 508 , located within the catalytic domain, is conserved in human, mouse, and monkey homologs (Fig. 8D). Therefore, these results corroborate the phosphorylation of ASMase in response to PMA and provide strong evidence that Ser 508 is involved in this phosphorylation.
Next, the role of PKC␦ in PMA-induced ASMase phosphorylation was investigated. Knockdown of PKC␦ was successfully achieved (more than 80%) via PKC␦-specific RNAi oligonucleotides. Phosphorylation of ASMase in response to PMA was determined 48 h after transfection of either PKC␦ RNAi or control sequence (SCR). As shown in Fig. 8E, loss of PKC␦ impeded the ability of PMA to induce ASMase serine phosphorylation. Therefore, these findings provide genetic evidence for the requirement of PKC␦ in the ASMase response to PMA.
UV Light Triggers the PKC␦-ASMase Pathway-Based on the above biochemical studies, it became important to determine if the phosphorylation of Ser 508 is required for stress-induced activation of ASMase. To this end, MCF-7 cells overexpressing WT ASMase or ASMase S508A were exposed to UV-C light radiation (50 J/m 2 ), a previously established ASMase stimulant (32)(33)(34). Phosphorylation of ASMase was examined 10 min after UV exposure, a time where maximal ASMase activity was detected. In line with previous results, basal serine phosphorylation could be detected in untreated cells overexpressing WT ASMase and ASMase S508A . However, enhanced serine phosphorylation, comparable with the one previously observed with PMA treatment (positive control), was detected in WT ASMase but not ASMase S508A overexpressor following UV exposure (Fig. 9). These results demonstrate that phosphorylation of ASMase at Ser 508 is induced by stress stimuli as represented by UV light.
ASMase S508A Acts as a Dominant Negative ASMase-Several recent reports described that activation of ASMase is associated with its translocation from the endolysosomal compartment to the plasma membrane (35). Since the above results indicated a similar translocation process after PMA treatment, it became important to investigate whether ASMase translocation to the plasma membrane is dependent on Ser 508 phosphorylation. To that end, MCF-7 cells plated on poly-L-lysine-coated confocal dishes were transiently transfected with either WT or S508A ASMase. Although a similar distribution pattern was observed under basal conditions, both PMA and UV light stimulation induced membrane association of only WT and not ASMase S508A (Fig. 10A). Therefore, these findings suggest that phosphorylation of Ser 508 is also required for ASMase translocation to the plasma membrane.
The significance of Ser 508 phosphorylation in the response of MCF-7 cells to PMA was further pursued. Cells transiently overexpressing either WT or ASMase S508A were treated with PMA (1 h). After lipid extraction, samples were subjected to mass spectrometric analysis of different ceramide species. No major differences in the different ceramides were seen under basal conditions. However, upon PMA treatment, a major increase in C16-ceramide and a minor one in C24-ceramide were detected only in WT-and not ASMase S508A -overexpressing cells (Fig. 10B). Therefore, these results demonstrate that phosphorylation of Ser 508 is required for agonist-induced activation of ASMase and subsequent ceramide generation.

DISCUSSION
ASMase is a well characterized lipid phospholipase with suggested roles in cellular stress responses, yet its biochemical and molecular mechanisms of regulation remain largely unknown. The results from the present study show that activation of ASMase proceeds through phosphorylation at Ser 508 . In addition to its activation, translocation of ASMase from the endolysosomal compartment to the plasma membrane was found to be dependent on this phosphorylation step. The study also indicates that PKC␦ is an important kinase that could function upstream of ASMase in cell signaling pathways.
A major conclusion from this study relates to the role of ASMase in the regulated salvage pathway of ceramide accumu- lation. It was shown previously that the ceramide response to PMA stimulation requires ceramide synthase activity (inhibited by fumonisin B1) but not de novo synthesis (not inhib-ited by myriocin). Using a metabolic labeling approach, the same study indicated that the ceramide response originated from breakdown of complex sphingolipids and not from de novo synthesis (1). Therefore, these results indicated the operation of the "salvage pathway," whereby complex sphingolipids are broken down to ceramide and then to sphingosine, which is then reacylated to ceramide.
In preliminary results, it was found that PMA stimulated breakdown of SM; therefore, it became important to determine the enzyme responsible for the initiation of the salvage pathway. The results from this study implicate ASMase in the ceramide response to PMA. This is supported by two lines of evidence. First, RNAi targeting of ASMase significantly attenuated ceramide accumulation in response to PMA. Second, the dominant negative ASMase S508A mutant also inhibited the PMA response. Therefore, in addition to the classical de novo and SM breakdown pathways, these results highlight a third "hybrid" salvage pathway of ceramide formation.
Interestingly, this salvage pathway, by combining ASMase activation with fumonisin-inhibitable ceramide formation, may explain previously seemingly contradictory results. For example, the chemotherapeutic agents fenretinide and anandamide were shown to induce ceramide production through activation of ASMase (36,37). However, other reports showed that ceramide generated by these agents can be blocked by fumonisin B1 treatment (38,39). In some cases, there were different results reported for the same cell line. Although radiation of MOLT-4 cells induced breakdown of sphingomyelin to ceramide through activation of ASMase, it was shown in a recent study that this ceramide response can be blocked by fumonisin in the same cell line (40,41). Also, a previous study showed an important role for ceramide synthase in the PMA response of LNCaP prostate cancer cells (42); however, the effects on SM turnover were not investigated. . Characterization of ASMase serine to alanine mutants. A, expression and activity of ASMase Ser 3 Ala mutants. Cells grown on 10-cm dishes (60% confluence) were transiently transfected with WT or mutant ASMase plasmids (V5 tagged). After 24 h, cells were collected and lysed, and lysates were used for ASMase activity assays. Alternatively, lysates were electrophoresed by SDS-PAGE, and ASMase overexpression was detected by Western blotting (via V5 epitope). B, localization of ASMase mutants. MCF-7 cells transiently overexpressing the indicated ASMase mutants (V5-tagged) were analyzed by confocal microscopy. ASMase was visualized by staining for V5 epitope (green channel), and LAMP-1 (red channel) was used as a lysosomal marker. Nuclei were visualized via DRAQ-5 nuclear stain (blue). Scale bar, 5 m. Data are expressed as mean Ϯ S.E. from three SMase activity assays.
The effects of PKC agonists on sphingolipid metabolism have been studied previously. In GH3 pituitary cells, it was shown that diacylglycerol but not phorbol esters directly trigger SM breakdown independent of PKC (24). However, using the same metabolic labeling approach with radioactive sphingomyelin as in the aforementioned study, Tettamanti et al. (43) came to a different conclusion. The authors found that treatment of the human neuroblastoma cell line SH-SY5Y with PMA induces SM breakdown, a critical step in the differentiation of this cell line. In agreement with the latter study, we previously reported that stimulation of the MCF-7 breast cancer cell line with PMA triggered the sphingomyelin breakdown pathway (1). In addi- FIGURE 8. Role of Ser 508 in ASMase activation. Cells transiently overexpressing WT ASMase or the indicated Ser 3 Ala mutants were treated with either vehicle (Me 2 SO) or PMA for 1 h. A, lysates (100 g each) were used for the ASMase enzymatic assay described under "Experimental Procedures." B, ASMase was immunoprecipitated (IP) (via V5 epitope) from each sample. Immunoprecipitates were electrophoresed by SDS-PAGE and probed using anti-phosphoserine antibody. C, in vivo phosphorylation of WT ASMase and ASMase S508A . Cells transiently transfected with either WT ASMase or ASMase S508A were labeled with [ 32 P]orthophosphate for 1 h, followed by treatment with PMA for 1 h. After immunoprecipitation with a V5 monoclonal antibody, ASMase was deglycosylated by overnight incubation with endoglycosidase H (Endo H) (except the first lane sample). D, ClustalW algorithm sequence alignment of ASMase sequences from different organisms in the region of Ser 508 of human ASMase. E, cells overexpressing WT-ASMase were transfected with PKC␦ RNAi or SCR (5 nM) for 48 h. ASMase was immunoprecipitated 10 min after treatment with PMA (100 nM, 1 h), using a V5 monoclonal antibody. Serine phosphorylation was evaluated by immunoblotting using a phosphoserine polyclonal antibody. Data are from one of three independent experiments with similar results (p Ͻ 0.05). ns, not significant. CN, control. FIGURE 9. Role of Ser 508 in UV light-induced ASMase activation. MCF-7 cells overexpressing WT-ASMase or ASMase S508A were seeded in 100-mm dishes (5 ϫ 10 5 cells/plate). ASMase was immunoprecipitated (IP) 10 min after exposure to UV light or treatment with PMA (100 nM, 1 h), using a V5 monoclonal antibody. Serine phosphorylation was evaluated by immunoblotting using a phosphoserine polyclonal antibody. Blots shown are representative of at least three independent experiments. CN, control. A, ASMase S508A fails to translocate to the plasma membrane. Cells transiently transfected with WT ASMase or ASMase S508A were analyzed by confocal microscopy prior to and after treatment with PMA for 1 h or 10 min after exposure to UV light ( ϭ 254 nm, 50 J/m 2 ). ASMase was stained using a V5 antibody (green channel), and nuclei were visualized using a nuclear dye (red channel). Scale bar, 5 m. B, overexpression of ASMase S508A blocks ceramide formation after PMA treatment. Cells transiently overexpressing WT ASMase or ASMase S508A were treated with either Me 2 SO or PMA for 1 h. Lipids were extracted according to the Bligh and Dyer method, and ceramide species were analyzed by mass spectroscopy. Ceramide measurements are expressed as mean Ϯ S.E. from three experiments (p Ͻ 0.05). ns, not significant. Measurements were normalized to total cellular phospholipids. CN, control.
tion to the difference in cell lines used, it is worth indicating that these studies employed different time frames of PMA stimulation (1, 12, and 24 h) as well as different methods for SM detection (mass spectroscopy versus metabolic labeling). Taken together, it is likely that the effects of PKC agonists on the sphingomyelin/ceramide pathway vary according to the cell line and pattern of PKC expression.
Another major conclusion from this study emanates from defining a molecular/biochemical mechanism for activation of ASMase. Phosphorylation assays and site-directed mutagenesis revealed that phosphorylation of ASMase at Ser 508 is essential for activation of the enzyme. Moreover, the results show that this phosphorylation is required for translocation of ASMase to the plasma membrane. Indeed, this finding potentially explains previous reports describing ASMase translocation to the plasma membrane. In response to UV radiation and Fas ligand (CD 95), ASMase has been reported to externalize to cholera toxin-positive membrane domains (so-called "lipid rafts") (32,33,44). This process seemed to require the integrity of vesicular trafficking in addition to microtubules and actin cytoskeleton. Hence, it is tempting to speculate that phosphorylation of ASMase facilitates its interaction with membrane transport and/or docking machinery. This hypothesis is further supported by several studies documenting a central role for PKC␦ in stress responses elicited by UV radiation and Fas ligand (45)(46)(47). However, further studies are required in order to identify specific mechanisms by which the phosphorylated form of ASMase translocates to the plasma membrane.
Although determining the exact mechanism by which PKC␦ interacts with ASMase requires further study, it is clear that this kinase is critical for ASMase activation at least in response to PMA and UV light. This conclusion may begin to tie in previous results on stress signaling involving both PKC␦ and ASMase. First, generation of reactive oxygen species has been recently shown to be an obligatory step for ASMase activation by UV radiation and TRAIL (19,33). Interestingly, there are several reports highlighting the role of PKC␦ both as a key target and effector during oxidative stress (48). Second, the PKC␦ Ϫ/Ϫ mouse shows a phenotype of suppressed apoptosis in response to tumor necrosis factor-␣, UV, and ␥-radiation (49). Importantly, the aforementioned stress agents signal ceramide generation primarily through ASMase activation; this raises the question of whether the PKC␦ Ϫ/Ϫ mouse has a defective sphingomyelin/ceramide stress signaling pathway. Although the results shown in this study indicate that PKC␦ can associate and phosphorylate ASMase in cells and in vitro, we cannot rule out the potential involvement of other kinases at this point.
The identification of key substrates for PKC␦ has emerged as an important goal in deciphering mechanisms of action of this kinase. Although activation of PKC␦ has been associated with growth arrest and in some cases initiation of an apoptotic cascade, only few PKC␦ substrates have been identified. For instance, PKC␦-dependent phosphorylation of phospholipid scramblase 1 (plasma membrane) (50), phospholipid scramblase 3 (mitochondrion) (51), and LaminB (nucleus) (52) are critical events during certain apoptotic cascades. Given the involvement of PKC␦ in apoptotic pathways, perhaps it is not surprising that ASMase, which in turn is critical in several cellular stress responses, is a substrate and an effector of PKC␦.
The novel PKC␦ and the sphingolipid ceramide are well established as intermediates in signal transduction pathways, and recent studies have begun to explore the cross-talk between the two (see commentary by Grant and Spiegel (53)). Saito and co-workers (27) showed that ceramide induces translocation of PKC␦ to the Golgi apparatus in HeLa cells. In a later study, the authors reported that ceramide activation of Src kinase induces phosphorylation of PKC␦ inside the Golgi and that this process is indispensable for ceramide-mediated apoptosis (54). Furthermore, de novo synthesis of ceramide caused cytochrome c release by targeting PKC␦ to the mitochondrion in prostate cancer cells treated with DNA-damaging agents (55). The current results further extend this area of research to show that PKC␦ can also function upstream of ceramide generation through regulation of ASMase. With this in mind, one can hypothesize the existence of a positive feedback loop between activation of PKC␦ and downstream ceramide formation.