Neutral Sphingomyelinase 2 (nSMase2) Is a Phosphoprotein Regulated by Calcineurin (PP2B)*

We previously reported that exposure of human airway epithelial cells to oxidative stress increased ceramide generation via specific activation of neutral sphingomyelinase2 (nSMase2). Here we show that nSMase2 is a phosphoprotein exclusively phosphorylated at serine residues. The level of nSMase2 phosphorylation can be modulated by treatment with anisomycin or phorbol 12-myristate 13-acetate (PMA/12-O-tetradecanoylphorbol-13-acetate), suggesting that p38 mitogen-activated protein kinase (MAPK) and protein kinases Cs are upstream of nSMase2 phosphorylation. Oxidative stress enhances both the activity and phosphorylation of nSMase2. Strikingly, we show here that nSMase2 is bound directly by the phosphatase calcineurin (CaN), which acts as an on/off switch for nSMase2 phosphorylation in the presence or absence of oxidative stress. Specifically, CaN is being inhibited/degraded and therefore does not bind nSMase2 under oxidative stress, and a mutant nSMase2 that lacks the CaN binding site exhibits constitutively elevated phosphorylation and increased activity relative to wild type nSMase2. Importantly, the phosphorylation and activity of the mutant no longer responds to oxidative stress, confirming that CaN is the critical link that allows oxidative stress to modulate nSMase2 phosphorylation and function.

We previously reported that exposure of human airway epithelial cells to oxidative stress increased ceramide generation via specific activation of neutral sphingomyelinase2 (nSMase2). Here we show that nSMase2 is a phosphoprotein exclusively phosphorylated at serine residues. The level of nSMase2 phosphorylation can be modulated by treatment with anisomycin or phorbol 12-myristate 13-acetate (PMA/12-O-tetradecanoylphorbol-13-acetate), suggesting that p38 mitogen-activated protein kinase (MAPK) and protein kinases Cs are upstream of nSMase2 phosphorylation. Oxidative stress enhances both the activity and phosphorylation of nSMase2. Strikingly, we show here that nSMase2 is bound directly by the phosphatase calcineurin (CaN), which acts as an on/off switch for nSMase2 phosphorylation in the presence or absence of oxidative stress. Specifically, CaN is being inhibited/degraded and therefore does not bind nSMase2 under oxidative stress, and a mutant nSMase2 that lacks the CaN binding site exhibits constitutively elevated phosphorylation and increased activity relative to wild type nSMase2. Importantly, the phosphorylation and activity of the mutant no longer responds to oxidative stress, confirming that CaN is the critical link that allows oxidative stress to modulate nSMase2 phosphorylation and function.
Ceramide is synthesized through either a de novo pathway involving serine palmitoyl-CoA transferase and ceramide synthase, or from breakdown of membrane sphingomyelin (N-acylsphingosine-1-phosphocholine) (Fig. 1A), a phospholipid preferentially concentrated in the plasma membrane of mammalian cells (9). Sphingomyelin catabolism occurs via the action of sphingomyelinases (SMases), 3 which are sphingomyelin-specific forms of phospholipase C that hydrolyze the phos-phodiester bond of sphingomyelin, yielding ceramide and phosphorylcholine. Ceramide then serves as a second messenger, leading to apoptotic DNA degradation.
We suggested that reactive oxidants up-regulate ceramide generation and cause elevated apoptosis in human airway epithelial (HAE) cells, thereby leading to lung injury pathologies. However, the mechanisms and target molecules of reactive oxidants affecting the HAE cells are not fully understood. Therefore, we proposed that increased oxidative stress and elevated ceramide generation are coupled at the molecular level by an unknown SMase that generates ceramide by hydrolysis of sphingomyelin (Fig. 1B). To proceed from the cellular to the molecular level, we searched for the specific SMase that is modulated by reactive oxygen species in lung epithelial cells, which led to our isolation of the novel nSMase2 from monkey lung tissue and HAE cells (1). This nSMase2 was previously found in the brain (10).
We then demonstrated that nSMase2 is the only member of the sphingomyelin phosphodiesterases family that is up-regulated and responsible for ceramide generation in HAE cells exposed to cigarette smoke (CS) or to H 2 O 2 (1,2). Moreover, we ascertained that CS exposure generates H 2 O 2 in the medium of HAE cells in a dose-dependent manner (2,11) and that pretreatment with glutathione (GSH) prevented both H 2 O 2 and ceramide generation (2,8).
Here we show that nSMase2 is a phosphoprotein in which the level of phosphorylation is modulated by oxidative stress, which also controls nSMase2 function. Furthermore, we demonstrate here that nSMase2 phosphorylation is regulated by a specific phosphatase that is modulated by oxidative stress.
Calcineurin (CaN) phosphatase (also known as protein phosphatase 2B (PP2B)) interacts directly with nSMase2, but not under the exposure to H 2 O 2 -induced oxidative stress. CaN is a Ca 2ϩ /calmodulin-dependent serine/threonine phosphatase, which can be inhibited by H 2 O 2 that modifies 2 Cys residues through the oxidative formation of a disulfide bridge, and eventually leads to CaN degradation (12). This phosphatase is known to bind to its substrates via a PXIXIT motif as described for its binding to nuclear factor of activated T cells (13). Indeed, we found that nSMase2 contains a PQIKIY sequence, and we show here unequivocally that deletion of this 6-amino acid PXIXIT-related sequence blocks CaN phosphatase binding to nSMase2. Furthermore, the nSMase2 mutant, which does not bind CaN, is much more phosphorylated and activated than the wild type (WT) nSMase2. This validates that the function of nSMase2 is modulated via its de-phosphorylation by CaN, which does not interact with nSMase2 under oxidative stress.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfections-We are using immortalized human bronchial epithelial (HBE1) cells (from Dr. Reen Wu, University of California, Davis, CA) and A549 adenocarcinoma cells, which behave similarly to primary airway epithelial cells with respect to reactive oxidant regulation of ceramide generation and apoptosis induction. Therefore, we routinely use HBE1 and A549 cells for most of our studies and then verify key findings in the primary lung epithelial cells. Culture conditions are identical for both HBE1 and primary cells, as described (14,15): cells are grown in serum-free medium supplemented with insulin (5 g/ml), transferrin (5 g/ml), epidermal growth factor (5 ng/ml), dexamethasone (0.1 M), cholera toxin (20 ng/ml), and bovine hypothalamus extract (15 g/ml). A549 cells were grown in F12-K medium (Invitrogen) supplemented with 10% fetal bovine serum and 5% penicillin/streptomycin. The cells were incubated at 37°C in a humidified atmosphere of 5% CO 2 and cultured as previously described (6,16). Transient transfections of pFLAG-nSMase2 and V5-nSMase2 were performed using Lipofectamine 2000 Transfection Reagent (Invitrogen) according to the manufacturer's protocol. Transfected cells were treated 24 h post-transfection.
The mutant (MT) that lacks the putative CaN binding site was prepared by site-directed mutagenesis, deleting the nucleotide sequence "CCTCAGATCAAGATCTAC" corresponding to nSMase2 (SMPD3) residues PQIKIY. Wild type-or MT-nSMase2 were tagged at 3Ј with the V5 epitope by using the pEF6/V5-His-TOPO vector (Invitrogen) for transfection; FLAG-nSMase2 was transfected in cells using vector pCMV-Tag2 (Stratagene).
Immunoprecipitation-Cells in Triton X-100 lysis buffer were incubated for 30 min at 4°C in an orbital shaker. Cell lysates were centrifuged at 14,000 ϫ g for 5 min at 4°C and the postnuclear supernatant was collected. Aliquots of the postnuclear supernatant containing equal amounts of protein (400 g) were immunoprecipitated either 2 h or overnight at 4°C with anti-(␣)V5 (Invitrogen) and anti-calcineurin (Millipore) antibodies, conjugated with protein A-Sepharose (RepliGen).
Alkaline Phosphatase Treatment-After immunoprecipitation of V5-nSMase2, the treatment was done by incubating the proteins in a mixture containing alkaline phosphatase enzyme for 30 min at 37°C, according to the manufacturer's instruction (New England Biolabs). Partial Acid Hydrolysis of 32 P-labeled V5-nSMase2-In vivo radiolabeled V5-nSMase2 was immunoprecipitated, discriminated on 9% acrylamide SDS-PAGE, transferred to nitrocellulose membrane, and stained by immunoblotting against the V5 epitope. Then, the relative cut radiolabeled bands were rinsed in water and acid hydrolyzed in 5.8 N HCl at 110°C for 1.5 h. The samples were dried under vacuum and resuspended in a standard mixture of phosphoamino acids (1 mg/ml of o-phospho-DL-serine, threonine, and tyrosine dissolved in H 2 O). Counts per minute were determined by scintillation counting and a sample volume corresponding to 300 counts/min was resolved by thin-layer electrophoresis as described by Hardin and Wolniak (17) at 250 V for 1.5 h. Radiosensible intensifying screen was used to visualize the radioactive amino acids, and staining with 0.5% ninhydrin solution in acetone was used to detect the standards.

nSMase2 Phosphorylation Regulated by Calcineurin Interaction
Determination of Cellular Ceramide Levels by Diacylglycerol Kinase Assay-Ceramide was quantified by the diacylglycerol kinase assay as previously described (7,16). Briefly, lipids were extracted with methanol, chloroform, 1 N HCl (100:100:1, v/v/ v). The lipids in the organic phase were dried under vacuum and resuspended in 100 l of reaction mixture containing [␥-32 P]ATP and incubated at room temperature for 1 h. The reactions were terminated by extraction of lipids with 1 ml of methanol, chloroform, 1 N HCl, 170 l of buffered saline solution, and 30 l of 0.1 M EDTA. The lower organic phase was dried under vacuum, and the lipids resolved by thin layer chromatography on Silica Gel 60 plates (Whatman) using a solvent of chloroform:methanol:acetic acid (65:15:5, v/v/v). Ceramide 1-phosphate was detected by autoradiography, and incorporated 32 P was quantified by densitometry scanning using a Molecular Dynamics Gel Scanner.
Assays for nSMase Activity-The enzyme activity of nSMase was determined as described (18,19). Briefly, an enzyme preparation of 10 g of total protein (from V5-nSMase2-transfected cells) in 20 mM Tris-HCl, pH 7.4, was mixed with [ 14 C]sphingomyelin (SM) (10 nmol/1,000,000 disintegrations/min) in 0.1% Triton X-100 in 100 mM Tris-HCl, pH 7.4, containing 10 mM MgCl 2 , 10 mM dithiothreitol, 10 nM phosphatidylserine, and 1 mg/ml of bovine serum albumin. The incubation time was 30 min at 37°C. The reaction was terminated by the addition of 1 ml of chloroform:methanol (2:1) followed by 0.2 ml of distilled water. After phase separation, the upper phase was removed and the radioactivity determined by liquid scintillation counting. Alternatively the mixture was incubated with fluorescent C6-NBD-labeled-SM, N-hexanoyl-NBD-sphingosylphosphorylcholine (Matreya). The NBD-ceramide generated in the reaction was solved by thin layer chromatography and quantified by densitometry fluorescent scanning.
Yeast Two-hybrid-A library-scale screen was carried out using full-length hnSMase2-pGBKT7 bait vector to interrogate a human heart Matchmaker cDNA library (Clontech) in the pAct2 prey vector. The initial screen was done at medium stringency and used the activation of just one of the three reporter genes as the basis for candidate selection. Our initial screen gave ϳ400 candidate colonies of a total of 500,000 colonies screened. These (400) colonies were then re-streaked onto agar plates lacking leucine and tryptophan (to maintain selection on both prey and bait plasmids) and lacking histidine, adenine, and supplied with ␣-X-gal (5-bromo-4-chloro-3-indolyl-␤-D-galactopyranoside). Of 400 candidate colonies obtained from the initial medium stringency screen less than 20 clones were able to grow on the full selection plates (-LEU/-TRP/-HIS/-Ade ϩ X-gal). Calcineurin (CaN) phosphatase showed up repeatedly among these clones. All reagents, unless stated otherwise, were from Sigma.
Statistical Analysis-Each time point or dose was analyzed at least in duplicate. Each group of studies was repeated and reproduced, at least three times. The data are reported as mean Ϯ S.D. Statistical significance was determined by Student's t test and p value Ͻ 0.05 was considered statistically significant; n ϭ number of experiments, as indicated.

RESULTS
nSMase2 Is a Phosphoprotein-In an attempt to understand the link between H 2 O 2 oxidative stress generation and nSMase2 activation, we investigated the molecular mechanisms involved in modulation of the nSMase2 function. By using HBE1, human bronchial epithelial cells, and A549 adenocarcinoma cells we showed that nSMase2 is a phosphoprotein (Fig. 2). First, the FLAG-tagged nSMase2 was transiently transfected in HBE1 cells, immunoprecipitated, and then treated (or not) with alkaline phosphatase. A gel shift analysis of a high resolution SDS-PAGE indicated that FIGURE 2. nSMase2 is a phosphoprotein. A, FLAG-nSMase2 was transiently transfected in HBE1 cells, immunoprecipitated with anti-FLAG antibody (Ab), treated (ϩ) or not (Ϫ) with alkaline phosphatase (AP), solved by high resolution SDS-PAGE and then Western immunoblotted (IB) with ␣FLAG antibody, revealing an upper (ϳ78 kDa) and a lower (ϳ74 kDa) band, representing, respectively, the phosphorylated and not phosphorylated forms of FLAG-nSMase2. B, V5-nSMase2 WT or the empty vector (EV) were transfected in A549 cells and labeled in vivo with [ 32 P]orthophosphate for 4 h in Dulbecco's modified Eagle's medium phosphate-free medium: 32 P labeling of nSMase2 was detected on SDS-PAGE by autoradiography after IP with ␣V5 antibody, showing a phosphorylated band at 78 kDa, not detectable in the EV-transfected IP (top panel); this band matched with that of V5-nSMase2 detected by immunoblotting with ␣V5 antibody (bottom panel). C, after 32 P labeling and IP (like in B) V5-nSMase2 was partially acid hydrolyzed and resolved by thin-layer electrophoresis as described under "Experimental Procedures": open black circles indicate the migration distance of ninhydrin-stained phosphoamino acid standards that were run in the same lane.
nSMase2 was phosphorylated ( Fig. 2A). In vivo labeling with [ 32 P]orthophosphate of HBE1 cells transiently transfected with V5-tagged nSMase2 also confirmed that the enzyme was basally phosphorylated (Fig. 2B). Finally, immunoprecipitated nSMase2 from 32 P-labeled cells was hydrolyzed and resolved by thin layer electrophoresis (Fig. 2C), which demonstrated that nSMase2 was phosphorylated on serine residues only.
Studies by others reported that p38 MAPK is upstream of nSMase2 in augmenting its activity during TNF-␣ and PMA treatment of A549 cells. In addition, PKC-␦ was implicated in the translocation of nSMase2 from the Golgi to the plasma membrane during such treatments (20,21). Here we show that treatments with either PMA or anisomycin, the respective stimulators of the "conventional" PKC(s) and p38 MAPK (22,23) elevate nSMase2 phosphorylation (Fig. 3). Anisomycin enhanced the phosphorylation of nSMase2 in a dose-dependent manner, suggesting the involvement of p38 MAPK upstream of nSMase2 phosphorylation (Fig. 3A). At the same time, as shown in Fig. 3B, the p38 inhibitor, SB202190, reduced the phosphorylation of nSMase2.
Interestingly, a 30-min treatment with 20 nM PMA induced a modest increase in nSMase2 phosphorylation, whereas a 4-h treatment with 200 nM PMA caused a substantial decrease in the level of nSMase2 phosphorylation (Fig. 3C). Because, it is well known that PKC activity is stimulated by a short/low concentration PMA treatment, but inhibited after a long/strong treatment (24), these data suggest that PKC(s) may also be involved in maintaining nSMase2 phosphorylation.
nSMase2 Phosphorylation and Activity Are Up-regulated by Oxidative Stress-Next, we carried out experiments to find out whether H 2 O 2 -induced oxidative stress affects the level of nSMase2 phosphorylation. Overexpression of V5-nSMase2 in A549 cells was followed by 32 P in vivo labeling and exposure to 250 M H 2 O 2 for the last 30 min of labeling. After immunoprecipitation the level of nSMase2 phosphorylation was determined by autoradiography and quantified/standardized by immunoblotting the tagged V5 epitope with anti-V5 antibody. As shown in Fig.  4A, H 2 O 2 -induced oxidative stress stimulated a 1.5-fold increase in nSMase2 phosphorylation. As previously reported, we show here again (Fig. 4B) that such exposure to oxidative stress enhances nSMase2 activity, suggesting that nSMase2 phosphorylation may affect its function.
Novel Protein-Protein Interaction: CaN Phosphatase Interacts with nSMase2 and This Interaction Is Down-regulated under Exposure to Oxidative Stress-We used yeast two-hybrid screening to identify potential candidates that may interact physically with nSMase2. Using human nSMase2 as a bait to interrogate a human cDNA library, we searched for candidates that bind nSMase2. In parallel, we ran global mass spectrometry analyses after immunoprecipitation of nSMase2 from primary lung epithelial cells, exposed (or not) to various oxidative stress conditions. Of several possible candidates we report here that CaN phosphatase interacts with nSMase2, but not under H 2 O 2induced oxidative stress.
CaN is a Ca 2ϩ /calmodulin-dependent serine/threonine phosphatase also known as protein phosphatase-2B. It can be inhibited by H 2 O 2 , which modifies 2 Cys residues through the oxidative formation of a disulfide bridge (12). Co-immunoprecipitation analysis of endogenous CaN with overexpressed V5-nSMase2 confirmed the above screenings. The V5 construct was immunoprecipitated from lysates of transiently transfected

nSMase2 Phosphorylation Regulated by Calcineurin Interaction
HBE1 cells, treated (or not) for 30 min with 250 M H 2 O 2 or 5 mM GSH, and analyzed by immunoblotting, using specific antibodies against the V5 epitope and against CaN. As shown in Fig.  5A, CaN indeed interacts with nSMase2 but not when the cells were exposed to H 2 O 2 . Furthermore, when endogenous CaN was immunoprecipitated with anti-CaN (Fig. 5B) the overexpressed nSMase2 could also be pooled down and immunoblotted. But, under H 2 O 2 exposure nSMase2 could not be pulled down and detected by immunoblotting. Because it has been reported (25) that oxidative stress can induce degradation and inactivation of CaN, we treated the HBE1 cells with 5 mM GSH or 250 M H 2 O 2 for the time points indicated in Fig. 5C. We then analyzed samples of cell lysate (50 g of total protein extract) by Western blotting with a specific antibody against CaN. Fig. 5C indeed confirms that H 2 O 2 treatment resulted in a time-dependent down-regulation of CaN. Additionally, Fig. 5D demonstrates that when the endogenous CaN was immunoprecipitated, the endogenous nSMase2 could be clearly pulled down and detected by immunoblotting with the anti-endogenous nSMase2 antibody.
Cyclosporine A (Cycl-A) is a well known inhibitor of CaN (26,27). Therefore, A549 cells transiently transfected with nSMase2 were 32 P in vivo labeled as described above, but exposed to Cycl-A for the last 30 min of labeling. Fig. 6A demonstrates that the Cycl-A treatment indeed triggered an increase of nSMase2 phosphorylation. On the other hand, stim-ulation of CaN activity with ionomycin, a calcium ionophore (28), caused de-phosphorylation of nSMase2 (Fig. 6B). Therefore, we propose that CaN plays a key role in the regulation of nSMase2 phosphorylation: when CaN is inhibited by Cycl-A or activated by ionomycin the phosphorylation of nSMase2 is increased or reduced, respectively. Therefore, CaN may indeed have a critical function in the regulation of nSMase2 phosphorylation under different oxidative stress conditions: CaN is degraded and therefore does not bind nSMase2 during H 2 O 2 treatment, thus allowing nSMase2 to be fully phosphorylated downstream of p38 MAPK and PKCs.
Phosphorylation and Activity of nSMase2 Are Related and Down-regulated by Calcineurin Phosphatase-To further substantiate our findings and in an attempt to define the role of CaN in nSMase2 function we found that the nSMase2 primary sequence presents a region, which may provide a docking site for CaN, PXIXIT (13,29). Therefore, we deleted that putative binding site of CaN, PQIKIY, and generated a MT of nSMase2. Fig. 7A shows that indeed the MT nSMase2 does not bind CaN. In addition, Fig. 7B demonstrates that when endogenous CaN is immunoprecipitated, the MT nSMase2 is hardly pulled down

FIGURE 5. Calcineurin phosphatase interacts with nSMase2 and its interaction is down-regulated under oxidative stress. HBE1 cells were transiently transfected with V5-nSMase2 and treated (or not) for 30 min with 250
M H 2 O 2 or 5 mM GSH. A, V5-nSMase2 was immunoprecipitated with ␣V5 antibody, loaded onto a SDS-PAGE gel, and immunoblotted with ␣V5 antibody and ␣-calcineurin (␣CaN) antibody. B, HBE1 cells were transfected as above and then immunoprecipitated with ␣CaN antibody followed by immunoblotting (IB) with either ␣CaN or ␣V5 antibodies. C, cells were treated as above for the indicated time points. Then, 50 g of the total protein extracts were directly immunoblotted (without IP) with ␣CaN antibody. D, left side, CaN was immunoprecipitated with ␣CaN Ab (from 400 g of total A549 protein lysate), then loaded onto SDS-PAGE and immunoblotted with ␣nSMase2 antibody and ␣CaN antibody, respectively; right side, 30 g of total protein cell lysate (LYS) was directly immunoblotted (without IP) with ␣nSMase2 antibody or ␣CaN antibody. compared with the WT nSMase2. Moreover, Fig. 7C demonstrates that such a mutant is much more phosphorylated in comparison to the WT nSMase2. In fact, it is phosphorylated as much as the WT nSMase2 exposed to H 2 O 2.
Notably, the nSMase2 mutant that does not bind CaN turns out to be much more active constitutively than the WT nSMase2. As shown in Fig. 8A the total nSMase enzymatic activity in cells transiently transfected with empty vector, WT, and MT showed that the MT is about 40% more active than the WT. Furthermore, Fig. 8B confirms that the MT is more active than the WT nSMase2 because when the MT nSMase2 (that misses CaN binding) was transfected into A549 cells that were not exposed to oxidative stress it triggered an increase in the levels of endogenous ceramide that were compatible with the levels of ceramide induced by the transfected WT nSMase2 but only after treatment with H 2 O 2 . Importantly, as shown by the lower panel of Fig. 8B, all the various constructs were transfected to comparable levels.  V5-nSMase2 was immunoprecipitated with ␣V5 antibody, and then immunoblotted with the same antibody and also exposed to autoradiography. C, the figure presents the average value of three independent experiments of nSMase2 phosphorylation detected by densitometry of the autoradiographs and reported as percentage of the untreated (NT), after normalization to the amount of immunoprecipitated V5-nSMase2 determined by immunoblotting (IB) with ␣V5 antibody; standard deviations are indicated; *, p Ͻ 0.05, n ϭ 3.

nSMase2 Phosphorylation Regulated by Calcineurin Interaction
Most importantly, as shown in Figs. 7 and 8, H 2 O 2 could not enhance the phosphorylation or the activity of the mutant, further confirming that interaction between CaN and nSMase2 is not only affecting the phosphorylation of nSMase2 but also affecting its activity. Moreover, this also suggests that the target for oxidative stress modulation of nSMase2 phosphorylation is solely the phosphatase CaN and not any kinase, as discussed below.
All together, and as shown in the model proposed in Fig. 9, these studies demonstrate that nSMase2 phosphorylation and function are inter-connected and that CaN down-regulates nSMase2 activity by de-phosphorylation, whereas in turn the oxidative stress exerts an up-regulation of nSMase2 phosphorylation and activity by abolishing CaN-nSMase2 interaction.

DISCUSSION
The results presented in this work suggest a critical role for PP2B, calcineurin, in the modulation of the phosphorylation and function of nSMase2 during exposure to oxidative stress of human lung epithelial cells, resulting in the increase of cellular ceramide levels.
Ceramide formation in mammalian systems under stress derives mainly from hydrolysis of membrane SM. The break of SM is driven by a single class of enzymes, the SMases. The main forms of SMases are distinguished by their pH optima (18, 19, 30 -33). Human and murine acid sphingomyelinase (aSMase; pH optimum 4.5-5.0) as well as Mg 2ϩ -dependent or -independent neutral SMases (nSMase; pH optimum 7.4) have been cloned and determined to be the products of conserved genes (10,34,35). Interestingly, membrane nSMase does not gain access to the signaling events activated by the lysosomal aSMase and vice versa, indicating that ceramide action may be determined by the subcellular site of its production.
Airway epithelial cells are the first line of defense of the lungs and are thus extensively exposed to reactive oxidants. Over the last few years we initiated studies to address whether these cells FIGURE 8. nSMase2 mutant that lacks the binding site to CaN is more active than the wild type. A, HBE1 cells that were transiently transfected with V5-nSMase2 WT or the MT that lacks the CaN binding site were lysed and assessed for nSMase activity by incubating 10 g of total protein cell lysate in a reaction mixture with 1 nmol of C6-NBD-sphingomyelin. The product generated in the reaction, C6-NBD-ceramide, was visualized by TLC and quantified by densitometry fluorescent scanning; the values shown present the average of four independent experiments, each of them was normalized according to the level of transfection; standard deviations are indicated; *, p Ͻ 0.05, n ϭ 4. B, the level of ceramide generation was determined by the diacylglycerol (as described under "Experimental Procedures"); the transfected cells demonstrated an increase in endogenous ceramide generation when the MT was compared with the WT, and also when the WT was exposed to H 2 O 2 -induced oxidative stress; the values represent the average of three experiments, each of them was normalized according to the level of transfection, which is shown in the same panel (immunoblot (IB) with ␣V5 antibody on total protein cell lysate), and reported as % of ceramide of the WT; standard deviations are indicated; *, p Ͻ 0.05, n ϭ 3. FIGURE 9. Proposed model of nSMase2 phosphorylation, which is modulated by oxidative stress via interaction with calcineurin phosphatase (PP2B). nSMase2 is a protein constitutively phosphorylated on serine residues downstream of p38 MAPK and PKCs; calcineurin binds to nSMase2, dephosphorylates it, and reduces its activity. Oxidative stress (H 2 O 2 ) abolishes the binding of calcineurin to nSMase2 and triggers elevated phosphorylation and activation of nSMase2. APRIL 2, 2010 • VOLUME 285 • NUMBER 14 are capable of entering apoptosis when exposed to micromolar concentrations of H 2 O 2 , and whether the process is mediated by ceramide as a second messenger (7,16). The range of 50 -250 M H 2 O 2 is considered to be the physiological range in which apoptosis can occur depending on the length of exposure to H 2 O 2 . As shown before (2,36), exposure to CS can generate between 100 and 800 M H 2 O 2 . Any concentration above 400 M would be considered pathological (2,8).

nSMase2 Phosphorylation Regulated by Calcineurin Interaction
Our previous studies in HAE cells showed that nSMase2 is activated by both H 2 O 2 -induced oxidative stress and CS exposures (1,2). This elevates the levels of cellular ceramide and augments apoptosis in the exposed lung epithelial cells. Loss of function experiments demonstrated that nSMase2 is the only specific target of H 2 O 2 and CS among the nSMase family members that are activated by oxidative stress and are essential for oxidative stress-induced apoptosis in lung epithelial cells (1,2). Here we show that oxidative stress not only controls nSMase2 function but that nSMase2 is a phosphoprotein in which oxidative stress modulates the level of phosphorylation.
In our yeast two-hybrid screening, CaN showed up repeatedly with multiple positive clones, whereas the number of other positive clones was limited and were viewed as weak candidates. We identified CaN, which is a Ca 2ϩ /calmodulin-dependent serine/threonine phosphatase (also known as PP2B) to directly interact with nSMase2. Moreover, CaN interaction with nSMase2 disappeared under oxidative stress, as confirmed by direct co-IP of endogenous CaN with an overexpressed V5-nSMase2 in HBE1 cells. We found that underexposure to H 2 O 2 CaN is degraded and therefore not available for binding to nSMase2. Moreover, a mutant nSMase2 that lacks the binding site to CaN was found to be constitutively overphosphorylated and activated when compared with the WT nSMase2 (see Figs. 7 and 8). Because CaN is able to down-regulate p38 MAPK signaling (37), whereas PMA can inhibit the activity of CaN (38), it is possible that phosphorylation of nSMase2 is mediated via the p38 and PKC serine/threonine kinases and down-modulated via CaN phosphatase.
It was previously reported that nSMase2 traffics to the plasma membrane in confluent MCF7 human breast adenocarcinoma cells, causing an increase in ceramide levels (39). Moreover, both H 2 O 2 and TNF-␣ exposures could induce nSMase2 translocation to the plasma membrane in A549 cells, suggesting that such a mechanism may be important for regulation of its activity (1,21). Clarke et al. (20) reported that nSMase2 activity is rapidly and transiently up-regulated in A549 cells by TNF-␣ via a p38 MAPK-dependent mechanism. The same group demonstrated that the novel PKC-␦ is upstream of the nSMase2 translocation from the Golgi to the plasma membrane during stimulation with either TNF-␣ or PMA, although PMA did not increase nSMase activity and PKC-␦ did not regulate the TNF-␣-induced increased activity (21). Moreover, PKC-␦ was not found to interact (co-immunoprecipitate) with nSMase2 (20,21). Clearly, additional studies are needed to provide full insight into the modulation of the nSMase2 structure function.
Indeed, we found that both anisomycin and PMA/TPA (low concentration of 20 nM) enhanced nSMase2 phosphorylation, whereas the p38 inhibitor, SB202190, reduced it, strongly suggesting that nSMase2 may be phosphorylated downstream of p38 ␣MAPK and PKCs. In addition, we observed a 60% decrease of nSMase2 phosphorylation after a prolonged PMA stimulation of lung epithelial cells, a treatment well known to typically inactivate PKCs. However, it seems unlikely that oxidative stress enhances nSMase2 phosphorylation via activation of a PKC or p38 MAPK. Even though others have shown that PKC-␦ and p38 can be activated by exposure to H 2 O 2 -induced oxidative stress (40 -42), our data presented in Figs. 8 and 9 demonstrate unequivocally that once nSMase2 cannot bind CaN its phosphorylation (and function) could not be further enhanced by exposure to H 2 O 2 -induced oxidative stress. Therefore, we suggest (see also model in Fig. 9, below) that only the inhibition of CaN by H 2 O 2 oxidative stress shuts off the phosphatase and enables nSMase2 to be fully phosphorylated and activated downstream of PKC and/or p38 MAPK (38,43,44). Our studies also demonstrated (Fig. 6) that treatment with Cycl-A, a well known inhibitor of CaN (26,27), triggers an increase of nSMase2 phosphorylation, whereas stimulation of CaN activity with ionomycin, a calcium ionophore (28), causes de-phosphorylation of nSMase2, confirming the regulation exerted by CaN on nSMase2 phosphorylation.
Of note is the down-regulation of CaN by Cycl-A. This drug is widely used in post-transplantation procedures because it prevents inflammatory responses generated by cytokine release, which are, in turn, generated by CaN-dependent activation of the NFAT cells (46). At the same time, this drug may also trigger severe adverse symptoms and systemic complications such as renal and neurotoxicity (47). This toxicity of Cycl-A is not well understood. One direction that should be further explored is whether cyclosporine affects sphingolipid metabolism through enhancing nSMase2 function, elevating cellular ceramide levels and thus enhancing cell death (26,48,49).
Recent studies provide a very strong case for cell death having a major role in lung injury in several pulmonary diseases (50 -61). In simple terms, loss of cells by augmented apoptosis would be expected to be involved, or perhaps initiate, the overall tissue destruction responsible for lung injury (53)(54)(55)(56)(57)(58)(59)(60)(61). Although a link between reactive oxidants and epithelial injury in the lung has been established (62), the cellular and molecular mechanisms leading to epithelial dysfunction were poorly defined.
Progress has recently evolved in the understanding of the underlying mechanisms of these destructive lung processes. For example, Gulbins et al. (63) and Worgall et al. (64) have recently reported that ceramide accumulation mediates inflammation, cell death, and infection susceptibility in cystic fibrosis. Several other studies implicate sphingomyelin hydrolysis in acute lung injury (56) and pulmonary edema (65). It was reported that ceramide may be a critical mediator of endothelial and alveolar cell apoptosis in the vascular endothelial growth factor mouse model of chronic obstructive pulmonary disease (56). Also, it has been shown (66) that superoxide dismutase protects against apoptosis and alveolar enlargement induced by ceramide. However, multiple questions still emerge (67), and additional characterization of the ceramide pathway that leads to lung injury is still needed, given that cigarette smoke is a predominant cause of chronic obstructive pulmo-nSMase2 Phosphorylation Regulated by Calcineurin Interaction nary disease. Animal models using smoke exposure would appear to be very useful for investigation, but it is only relatively recent that such models have been created (68).
Our recent data obtained in mouse and rat models 4 showed that when nSMase2 was small interfering RNA-silenced, lung epithelial cells could not generate excess ceramide in response to CS exposure, and indicated that nSMase2 could be a critical target in the reactive oxygen species/CS lung injury model (see Fig. 1B). Furthermore, this suggested that nSMase2 could present a novel target in the prevention of CS-induced lung epithelial cell death in epithelial/alveolar injury of chronic obstructive pulmonary disease patients (1, 2, 7). These recent findings further underscored the importance of elucidating the molecular mechanism of nSMase2 activation under oxidative stress. This study demonstrates that phosphorylation is a regulative mechanism of nSMase2 activation during exposure of HAE cells to oxidative stress, and identifies CaN as the target that interacts with nSMase2 and down-regulates its phosphorylation.
These unique recent findings were not expected. Even though the best characterized ceramide targets are the ceramide-activated serine/threonine protein phosphatases PP1 and PP2A, which are activated downstream of ceramide (45), the interaction of CaN/PP2B with nSMase2 is different. First, it is direct, and second it modulates the phosphorylation and function of nSMase2 thereby leading to augmented ceramide production under oxidant stress in lung epithelia.
As proposed in our model (see Fig. 9) CaN plays a key role in the regulation of nSMase2 phosphorylation under changes in oxidative stress. CaN is degraded during exposure to oxidative stress, and thus does not bind nSMase2, allowing it to be fully phosphorylated downstream of p38 MAPK/PKC.