Molecular Cloning, Characterization, and Expression of a Novel Human Neutral Sphingomyelinase*

Neutral sphingomyelinase (N-SMase) has emerged as an important cell membrane-associated enzyme that participates in several signal transduction and cell regulatory phenomena. Using expression cloning, we have identified a 3.7-kilobase pair cDNA transcript for N-SMase whose open reading frame predicts a 397-amino acid polypeptide. Transfection of COS-7 cells with cDNA for N-SMase resulted in a marked increase in N-SMase activity. Recombinant N-SMase (r-N-SMase) had the following physical-chemical properties. Mg2+ activated and Cu2+ and glutathione inhibited the activity of r-N-SMase. In contrast, dithiothreitol did not alter the activity of the enzyme. Of several phospholipids examined, sphingomyelin was the preferred substrate for r-N-SMase. The apparent molecular mass of r-N-SMase derived from COS-7 cells was ∼90 kDa, similar to the native neutral sphingomyelinase prepared from human urine. However, upon expression inEscherichia coli, the apparent molecular mass of the recombinant enzyme was ∼45 kDa. We speculate that this apparent difference in recombinant enzymes derived from COS-7 and E. coli cells may be due to extensive post-transcriptional changes. r-N-SMase has numerous post-transcriptional modification sites such as phosphorylation sites via protein kinase C, casein kinase II, tyrosine kinase, and cAMP- and cGMP-dependent protein kinases as well as sites for glycosylation and myristoylation. Amino acid sequence alignment studies revealed that r-N-SMase has some similarity to acid sphingomyelinase and significant homology to the death domains of tumor necrosis factor-α receptor-1 and Fas/Apo-I. We believe that the molecular cloning and characterization of N-SMase cDNA will accelerate the process to define its role as a key regulator in apoptosis, lipid and lipoprotein metabolism, and other cell regulatory pathways.

Type C sphingomyelinases (sphingomyelin phosphodiesterase, EC 3.1.4.12) are a group of phospholipases that catalyze the hydrolytic cleavage of sphingomyelin to ceramide and phosphocholine (1). Neutral sphingomyelinase from human urine and cultured human kidney proximal tubular cell membranes has an apparent molecular mass of 92 kDa and neutral pH optima and is heat-unstable. This enzyme is associated with the cell membrane in tissues and cultured cells (1)(2)(3)(4).
In cultured mammalian cells, the addition of diverse agonists, i.e. vitamin D 3 , tumor necrosis factor-␣ (TNF-␣), 1 interferon-␥, and nerve growth factor, results in the activation of N-SMase and the consequent production of ceramide. Ceramide and its higher homologs have been shown to serve as lipid second messengers that lead to diverse cell regulatory phenomena, such as differentiation, proliferation, and programmed cell death or apoptosis (1, 4 -6). Activation of N-SMase by TNF-␣ in human skin fibroblasts results in not only the hydrolytic cleavage of sphingomyelin, residing on the cell surface, but also the mobilization of cholesterol to the interior of the cell. Such cholesterol is esterified by the action of fatty-acyl-coenzyme A acyltransferase to form cholesteryl esters (1,7). In a human hepatocyte cell line, TNF-␣-induced N-SMase activation and ceramide production led to the maturation of sterol regulatory element-binding protein-1 and a subsequent increase in LDL receptor mRNA expression (8). Interestingly, this phenomenon is not accompanied by apoptosis and occurs in a sterol-independent fashion. Thus, LDL receptors may be up-regulated via this cascade of reactions involving N-SMase independent of the presence of sterols in the culture medium. These studies suggest that N-SMase may be involved in the regulation of lipid and lipoprotein metabolism and sterol influx (7,8). Rabbit skeletal muscle has been shown to contain at least two kinds of N-SMase. These are the classical 92-kDa Mg 2ϩ -dependent N-SMase and a Mg 2ϩ -independent 53-kDa protein. The localization of N-SMase in skeletal muscle transferase tubule membrane is in agreement with the production of the sphingomyelin-derived second messenger, sphingosine, in such tubules. Since sphingosine has been shown to modulate calcium release from sarcoplasmic reticulum membranes, these studies imply that N-SMase/sphingosine signaling systems may be a physiologically relevant mechanism of regulation of Ca 2ϩ levels in skeletal muscle and may well be involved in muscle contraction (9). In additional studies, we have shown that Sindbis virus entry into cells triggers apoptosis by activating sphingomyelinase and the release of ceramide. 2 Collectively, these studies imply that N-SMase may play a central role in diverse cell regulatory phenomena. Nevertheless, the evidence accumulated so far in support of this view has been largely indirect and has recently been the subject of rigorous debate (11), particularly in regard to the controversial role of ceramide in apoptosis (12).
We rationalized that to unambiguously demonstrate the functional role of N-SMase in the various cell regulatory phenomena discussed above, it was essential to first clone this protein. Therefore, we employed a monospecific polyclonal an-tibody against human N-SMase and a human kidney cDNA library to pursue expression cloning. In this work, we describe the molecular cloning and characterization of recombinant N-SMase (r-N-SMase) expressed in Escherichia coli as well as in cultured COS-7 cells. In work to be published elsewhere, we will show that overexpression of N-SMase results in spontaneous apoptosis in human aortic smooth muscle cells independent of the presence of agonists. 3
The multiple tissue Northern blot and human kidney library were purchased from CLONTECH (Palo Alto, CA). Restriction endonucleases were purchased from Amersham Pharmacia Biotech. The transient expression vector pSV-SPOT-1, LipofectAMINE TM , and cell culture medium were purchased from Life Technologies, Inc.
Measurement of Sphingomyelinase Activity-Bacterial cells and COS-7 cells transfected with N-SMase cDNA were homogenized in Tris/glycine buffer (pH 7.4) containing 0.1% cutscum. The samples were mixed vigorously and sonicated for 10 s. Next, the samples were transferred to a 4°C incubator and shaken for ϳ2 h. Every hour, the samples were sonicated again on ice and further shaken. Subsequently, the samples were centrifuged at 10,000 ϫ g for 10 min. The supernatants were collected, and the protein content was measured using bovine serum albumin as a standard and subjected to N-SMase assay using [ 14 C]sphingomyelin as a substrate (2).
Activity Gel Assay for the Measurement of Sphingomyelinase Activity-COS-7 cells transfected with pSV-SPOT-1 and pHH1 were subjected to detergent extraction as described above. The supernatants were subjected to sodium lauryl sarcosine-polyacrylamide (7.5%) gel electrophoresis at 4°C as described previously (13). The gel was calibrated with prestained proteins of known molecular mass (Bio-Rad). Subsequently, the gel was sliced, and pieces were transferred to a glass test tube and subjected to N-SMase assay at pH 7.4 using [ 14 C]sphingomyelin as a substrate.
Expression Cloning of N-SMase in E. coli and Screening of cDNA Library-The human kidney library was screened using the anti-N-SMase antibody available in our laboratory according to the manufacturer's protocol. Briefly, gt11 phage was plated at 3 ϫ 10 4 plaqueforming units/150-mm plate on a lawn of E. coli strain y1090r. Incubation was carried out at 42°C for 3.5 h to allow lytic phage growth. Then, a filter saturated with 10 mM isopropyl-␤-D-thiogalactopyranoside was placed on top of the plate and incubated overnight at 37°C. Next, the filter was blocked with a solution of 5% nonfat dry milk for 1 h at room temperature. Subsequently, the filter was incubated with antibody against N-SMase at 1:200 dilution at room temperature overnight, and signal was detected by the enhanced chemiluminescence technique (ECL, Amersham Pharmacia Biotech). Sixty-three clones were obtained by screening 1 ϫ 10 6 gt11 phage clones. The most intense clones of cells were subjected to secondary and tertiary screening. All positive clones were subjected to PCR to identify their insert size. Finally, a clone containing the longest insert called 32-1 was used for further analysis by subcloning, sequencing, and expression. We followed the protocol as described by the manufacturer. In this experiment, the host growing in the logarithmic phase was infected with phage and incubated for 2 h at 30°C. Next, isopropyl-␤-D-thiogalactopyranoside (10 mM) was added, and incubation was continued at 37°C. The cell cultures were removed at 0, 15, 30, and 45 min and 1, 2, 4, and 24 h. The cells were centrifuged, washed with phosphate-buffered saline, and stored frozen for further biochemical experimentation.
Preparation of GST-N-SMase Fusion Protein-To prepare GST-N-SMase fusion protein, a pJK2 expression plasmid, pBC32-2, was digested with BssHII and EcoRI. A 2793-bp insert representing the N-SMase open reading frame that is missing 18 bp of N-terminal sequence was ligated with a phosphorylated BamHI-BssHII linker containing 18 bp of N-terminal sequence (synthesized at the Core facility at The Johns Hopkins University) and a pGEX4T-1 vector double-digested with BamHI and EcoRI (Amersham Pharmacia Biotech).
To express and purify GST-N-SMase fusion protein, plasmid pJK2 was transformed into E. coli HB101 cells. A single colony of HB101[pJK2] was grown in 2ϫ yeast extract/tryptone agarose medium at 30°C until appropriate cell density (A 600 ϭ 1.5) was achieved. Isopropyl-␤-D-thiogalactopyranoside (0.1 M) was added to induce fusion protein expression for 2 h. Cells were harvested, and the fusion protein was purified using glutathione-Sepharose 4B chromatography according to instructions provided by the manufacturer (Amersham Pharmacia Biotech). N-SMase was released from the fusion protein by thrombin digestion. Such preparations were subjected to activity measurements and Western immunoblot assays.
Transient Expression of N-SMase in COS-7 Cells-To put N-SMase cDNA into a transient expression vector, pBC32-2 was double-digested with restriction endonucleases NotI and SalI. The 3.6-kb insert containing N-SMase was gel-purified and inserted into the transient expression vector pSV-SPOT-1. A plasmid called pHH1 was thus constructed. To transfect COS-7 cells with pHH1 and mock vector, we seeded 3 ϫ 10 5 COS-7 cells/plate in a p100 plate in 8 ml of Dulbecco's modified Eagle's medium with 10% dialyzed fetal calf serum. Cells were incubated in a 10% CO 2 37°C incubator until they were 80% confluent. The cells were then transfected with 1.0 g/ml purified pHH1 using LipofectAMINE TM in medium. The medium was changed after overnight incubation. Finally, the cells were harvested at various time points (16,24,36, and 48 h post-transfection) by centrifugation at 1500 ϫ g for 10 min, washed with phosphate-buffered saline, and stored frozen at Ϫ20°C. r-N-SMase derived from COS-7 cells was purified as described previously (2) and was employed for detailed characterization studies with regard to pH optima, substrate specificity, requirement for metal ions for activation, and other previously described agonists/antagonists of N-SMase as well as immunoprecipitation with anti-N-SMase antibody and preimmune rabbit serum IgG. The immunoprecipitates were subjected to Western immunoblot assays, and the activity of N-SMase was measured in the supernatants.
Northern Blot Assays-We selected one set of sequence primers from pBC32-2 (T3-2/4T3R5) as the reverse transcription-PCR primer to carry out reverse transcription-PCR as described above. A 465-bp specific product was obtained and gel-purified; 50 ng of this product was labeled with 25 Ci of [␣-32 P]dATP and 25 Ci of [␣-32 P]dCTP using random hexamer primers (Life Technologies, Inc.). The specific activity of this probe was 1.88 ϫ 10 9 cpm/g. Next, the multiple tissue Northern blot was prehybridized in 7 ml of hybridization-prehybridization solution at 42°C for 5 h with continuous agitation. Then, 10 g of fresh denatured salmon sperm was added. Next, 10 ml of hybridization solution was added and agitated continuously at 50°C. The blots were washed twice in 2ϫ SSC and 0.05% SDS at room temperature for 30 -40 min and then in 0.1ϫ SSC and 0.05% SDS for 40 min at 50°C. Finally, the blot was exposed overnight to a x-ray film at Ϫ70°C using two intensifying screens. A plasmid containing a ␤-actin insert was used as a control in these experiments.

RESULTS
Isolation of Human N-SMase cDNA-Our strategy for the molecular cloning of cDNA for N-SMase involved the use of monospecific antibodies against human urinary N-SMase (2). This rationale stemmed from our previous findings: (i) in human leukemic cells (HL-60), TNF-␣-induced activation of N-SMase was abrogated by such antibodies; (ii) in HL-60 cells and aortic smooth muscle cells, TNF-␣-and oxidized LDL-induced apoptosis, respectively, was abrogated by anti-N-SMase antibody; and (iii) both TNF-␣-induced maturation of sterol regulatory element-binding protein-1 in human liver cells and TNF-␣-induced cholesteryl ester synthesis in human skin fibroblasts were abrogated by preincubation of cells with such antibodies (7). We used this antibody to screen the human kidney gt11 cDNA library. Sixty-three positive clones were obtained by screening 1 ϫ 10 6 clones. The most intense clones were subjected to secondary and tertiary screening. All positive clones were subjected to PCR to identify their insert size. Finally, a clone containing the longest insert (3.7 kb) called 32-1 was subjected to subcloning into pBluescript II SK to prepare plas-mid pBC32-2. The insert was sequenced with Sequenase using T7 and T3 primers by automatic sequencing.
Amino Acid Sequence of N-SMase-The 3.7-kb nucleotide sequence of cDNA revealed an open reading frame size of 1197 base pairs, which predicts a 397-amino acid polypeptide. The deduced amino acid sequence of N-SMase is shown in Fig. 1. There are several potential modification sites in this protein.
Hydropathy plot analysis of N-SMase indicated that there are a few hydrophobic stretches, i.e. residues 75-100, indicating the presence of a transmembrane domain. The N-glycosylation site at position 353, the tyrosine phosphorylation site at position 238, and several other phosphorylation sites are presumably located on the exterior. Such sites may be subject to further glycosylation and phosphorylation. The alignment of the death domain of N-SMase with the death domains of TNF-␣ 55-kDa receptor-1 (TNF-␣-R1) and Fas/Apo-I is shown in Fig. 2. N-SMase shows closer homology to the TNF-␣-R1 death domain (ϳ25%) than to the Fas/Apo-I (ϳ19%) death domain (14). Indeed, N-SMase and TNF-␣-R1 have complete conservation of the ATL peptide at the C terminus of the death domain. Moreover, we found that over a 52-amino acid overlap, N-SMase had 26.9% identity and 71% similarity to acid sphingomyelinase (15) (data not shown). Since there is no cysteine residue in our deduced amino acid sequence, this may suggest multiple posttranslational modifications of N-SMase. To assess this hypothesis, we expressed N-SMase in both E. coli and COS-7 cells.
Expression of N-SMase in E. coli and COS-7 Cells-To confirm that the cDNA does encode N-SMase with an apparent molecular mass of 45.4 kDa, we made constructs of the cDNA-coding region fused with glutathione S-transferase. We then transfected this plasmid into E. coli HB101 to express and to purify GST-N-SMase fusion protein. The expression of the fusion protein was induced by isopropyl-␤-D-thiogalactopyranoside (0.1 M) for 2 h. The crude fusion protein (Fig. 3, lane 2) was purified using glutathione-Sepharose 4B chromatography. After N-SMase was released from the fusion protein by glutathi-one-Sepharose 4B chromatography, it resolved as a single band with a molecular mass of 46 kDa (Fig. 3, lane 3), and the estimated pI is 4.93. The urinary N-SMase reported by us previously has an apparent molecular mass of ϳ92 kDa and a pI of ϳ6.55 (2). This difference may be due to multiple posttranscriptional modifications of the mammalian N-SMase compared with the bacterial N-SMase (see below). The bacterial N-SMase was recognized by antibody against human N-SMase (Fig. 3, lane 4). Affinity-purified r-N-SMase expressed in E. coli had an activity of 3.9 mol/mg of protein/h. The deduced amino acid sequence of N-SMase indicates the lack of cysteine residues.
To further confirm the cDNA encoding N-SMase, we inserted it into a mammalian cell transient expression vector and transfected it into COS-7 cells. Fig. 4A shows that cells transfected with N-SMase cDNA (pHH1) exhibited a 10-fold increase in N-SMase activity compared with cells transfected with mock vector (pSV-SPOT-1) 24 h post-transfection. Measurement of N-SMase activity following sodium lauryl sarcosine gel electrophoretic separation of r-N-SMase showed optimal activity corresponding to an ϳ90 -100-kDa protein (Fig. 4B). This observation was confirmed further by Western immunoblot assay of r-N-SMase, which revealed that the apparent molecular mass of r-N-SMase is ϳ90 kDa, and this was similar to the native N-SMase derived from human urine (data not shown).
Characterization of r-N-SMase from COS-7 Cells-As shown in Fig. 5A, r-N-SMase derived from cells overexpressing the enzyme had a pH optimum of ϳ7.4. r-N-SMase did not have substantial enzyme activity in the acidic or alkaline range. Substrate specificity studies revealed that, of several phospholipids investigated, sphingomyelin appeared to be the most preferred substrate (data not shown). Studies on metal ion requirement revealed that r-N-SMase was activable with Mg 2ϩ (2.5 mM) (Fig. 5B). In contrast, Cu 2ϩ (5 M) markedly inhibited the activity of the enzyme. Additional studies revealed that r-N-SMase was insensitive to pretreatment with dithiothreitol, but the activity was inhibited significantly by glutathione (2.5 mM) (Fig. 5B).
Additional characterization studies using antibody against N-SMase and rabbit preimmune serum IgG as a control revealed the following. When the immunoprecipitates were subjected to Western immunoblot assays, an anti-N-SMase antibody concentration-dependent increase in the mass of N-SMase protein was revealed (Fig. 6A). On the other hand, measurement of N-SMase activity in the supernatants obtained following immunoprecipitation with anti-N-SMase antibody revealed a progressive decrease (Fig. 6B). Collectively, these data indicate that anti-N-SMase antibody could quantitatively immunoprecipitate r-N-SMase. In contrast, preimmune serum IgG did not immunoprecipitate the Mg 2ϩ -dependent N-SMase activity, and most of the N-SMase activity was associated with the supernatant.
Expression of N-SMase in Various Human Tissues-The 465-bp reverse transcription-PCR fragment amplified at the 5Ј-end of the N-SMase cDNA-coding region was used to probe N-SMase mRNA in various human tissues. N-SMase was expressed in all the human tissues investigated; the transcript size and copy numbers varied from one tissue to another (Fig.  7). For example, the size of N-SMase transcripts in human pancreas, liver, lung, and placenta ranged from 2.4 to 9.5 kb in an ascending order. However, Northern blot analysis revealed that the major transcript size of N-SMase expressed in all of the eight human tissues investigated is 1.7 kb. The 1.7-kb transcript may be derived either from different genes or from alternative splicing. A ␤-actin Northern blot shown at the bottom of Fig. 7 was run to serve as a positive control. DISCUSSION Using a human N-SMase monospecific polyclonal antibody, we have screened a human kidney cDNA library and have identified a 3.7-kb cDNA transcript. The open reading frame of the cDNA predicts a 397-amino acid polypeptide. When expressed in E. coli, it encodes a protein of ϳ45 kDa based upon SDS gel electrophoretic analysis. When expressed in COS-7 cells, the cDNA conferred to r-N-SMase a 10-fold higher activity relative to mock cDNA-transfected cells and had an apparent molecular mass of ϳ90 kDa, suggesting multiple posttranscriptional modifications. The presence of numerous phosphorylation sites via the action of protein kinase C, casein kinase II, cAMP-and cGMP-dependent protein kinases, and tyrosine kinase is predictive of the sensitivity of this enzyme to inhibitors of protein kinase C and serine, tyrosine, and casein II kinases (4,16,17). The deduced amino acid sequence of N-SMase indicates the lack of cysteine residues. This may explain its insensitivity to reducing agents, i.e. dithiothreitol (Fig. 5B) and ␤-mercaptoethanol (18,19). Moreover, inhibition of enzyme activity by glutathione is consistent with previous studies implicating a potential regulatory mechanism for this enzyme (19).
Physical-chemical characterization of r-N-SMase revealed properties similar to those of the human urinary/kidney N-SMase described by us previously (1, 2, 18). These were a pH optimum of 7.4, requirement for Mg 2ϩ , heat stability (control activity of 3223 cpm/h; activity after heat treatment at 60°C for 30 min of 330 cpm/h), and detergent for activation and inhibition by Cu 2ϩ . The amino acid sequence and hydropathy curve analysis collectively suggest this enzyme to be a membranebound Mg 2ϩ -sensitive N-SMase. Collectively, the physicalchemical properties of r-N-SMase derived from COS-7 cells, including molecular mass, are identical to those of the native enzyme derived from human urine (2). Although nucleotide sequence alignment of N-SMase revealed some similarity to acid SMase (15), it was insensitive to treatment with 5Ј-adenosine monophosphate (data not shown), a compound that has been shown to inhibit acid SMase activity (2). Although r-N-SMase was recognized (immunoprecipitated) by antibody against N-SMase, but not by preimmune serum IgG, the activity of the enzyme in the immunoprecipitate could not be recovered. Similar observations have been made previously using ␤-galactosidase and the corresponding antibody. 4 This may be explained on the basis that the polyclonal antibody bound to the active site in N-SMase, rendering it inactive. Application of conditions or procedures such as high salt concentration (4 M KCl), heating, and lowering the pH to 4 to dissociate the enzyme from the immunoprecipitate did not facilitate the recovery of activity, as the enzyme itself was highly sensitive to such treatments. Nevertheless, we demonstrated a dose-dependent increase in the immunoprecipitation of the antigen by anti-N-SMase antibody and a decrease in the activity of N-SMase in the supernatant. In the future, the availability of a monoclonal antibody directed against N-SMase (exclusive of the active site) to immunoprecipitate the enzyme may help retain the activity of the enzyme. The physiological effects of r-N-SMase (data not shown) were also similar to those of human urinary/kidney 4 D. Usher, personal communication.

FIG. 4. Expression of N-SMase activity in COS-7 cells. N-SMase
constructs were prepared with the cDNA-coding region fused with mammalian cell transient expression vector (pHH1) and transfected into COS-7 cells using LipofectAMINE TM . Cells were also transfected with mock cDNA (pSV-SPOT-1). A, 24 h post-transfection, cells were harvested, and the activity of neutral sphingomyelinase was measured using [ 14 C]sphingomyelin as a substrate (2). B, COS-7 cells transfected with pSV-SPOT-1 and pHH1 were subjected to sodium lauryl sarcosinepolyacrylamide (7.5%) gel electrophoresis at 4°C. The gel was calibrated with prestained proteins of known molecular mass. Following electrophoresis, the gel was sliced, and N-SMase activity was measured in gel slices using [ 14 C]sphingomyelin as a substrate.
N-SMase, such as a dose-dependent increase in cholesteryl ester synthesis and stimulation of the maturation of sterol regulatory element-binding protein-1 and LDL receptor mRNA expression in a continuous line of human hepatocytes (8). Fi-nally, overexpression of r-N-SMase in human aortic smooth muscle cells resulted in apoptosis and augmented oxidized LDL-induced apoptosis in these cells, 3 and antibody against N-SMase abrogated this phenomenon. 5 The cytoplasmic domain of two cell-surface receptors (Fas/ Apo-I/CD95 and TNF-␣-R) that share significant homology has been termed the "death domain." These two proteins contain cysteine-rich repeats that are also found in the nerve growth factor family of proteins such as TNF-␣-R2, CD30, CD26, CD40, Ox-4, and H-1BB. The ligands for Fas and TNF-␣-R are FasL and TNF-␣, respectively, and they can induce apoptosis in certain target cells (22,23). The death domains of TNF-␣-R and Fas have been found to interact with several interactive and adaptive proteins that have homology to the death domain and that provide a link to caspase activation and apoptosis (14). Previous studies have shown that overexpression of Fas/Apo-I or TNF-␣-R1 can lead to apoptosis in the absence of ligands due to multi-oligomerization of the death domain (10). Therefore, it was not surprising to find spontaneous apoptosis in cells overexpressing N-SMase 5 that was ligand-independent. This tenet 5 H. Han, T. C. Wan, and S. Chatterjee, submitted for publication. FIG. 6. Immunoprecipitation of sphingomyelinase from COS-7 cells transiently transfected with pHH1 and measurement of N-SMase activity in the supernatant. COS-7 cells transiently transfected with pHH1 were lysed in lysis buffer containing protease inhibitors and centrifuged (10,000 rpm, 10 min, 4°C). The supernatants were precleared with protein A-Sepharose (Amersham Pharmacia Biotech) and centrifuged as described above. Next, the supernatants were subjected to immunoprecipitation with preimmune rabbit serum IgG (Control; 1:200 dilution) and with increasing amounts of monospecific polyclonal antibody against human N-SMase (1:1000, 1:500, and 1:200) at 4°C overnight. Protein A-Sepharose was added to the reaction mixture; incubation was continued for another 2 h at 4°C; and then the samples were centrifuged. A, Western immunoblot assay of immunoprecipitates. The immunoprecipitates were solubilized in buffer and subjected to Western immunoblot assay. First bar, the immunoprecipitate was obtained following treatment with preimmune IgG; second through fourth bars, immunoprecipitates obtained following treatment with antibody against N-SMase at various dilutions (1:1000, 1:500, and 1:200, respectively). B, N-SMase activity in the supernatant. The supernatants were withdrawn to measure sphingomyelinase activity using [ 14  is further supported by the alignment of the death domain of N-SMase with the death domains of Fas/Apo-I and TNF-␣-R1, which showed close homology (Fig. 2). Indeed, N-SMase and TNF-␣-R1 have complete conservation of the ATL peptide at the C terminus of the death domain. Further studies will be necessary to explain the relevance of these preliminary observations.
Our cDNA consists of 3670 bp and contains polyadenylation signals at positions 351 (ATTATT), 805 (ATTAAA), 2562 (AAT-TAA), and 2835 (ATTAAA). These polyadenylation signals vary from the AATAAA consensus sequence, but they were found in 12% (ATTAAA) and 2% (AATTAA and ATTATT) of the mRNAs in vertebrates. Such smaller transcripts exist due to their termination at different locations. RNAs derived from placenta, lung, liver, and pancreas showed strong signals (Fig. 7). Kidney, brain, and heart RNAs showed medium level signals, and skeletal muscle RNA showed the weakest signal. There is apparently no correlation between N-SMase activity and N-SMase mRNA levels in a given tissue. For example, among ϳ40 human tissues surveyed, brain had the highest enzyme activity for N-SMase. Skeletal muscle and kidney had 3-and 10-fold lower N-SMase activities, respectively, as compared with brain, yet the N-SMase mRNA levels in brain and kidney are similar. This may be due to tissue-specific translational regulation at multiple sites (Fig. 1).
While this manuscript was in preparation, another publication appeared in the literature claiming the cloning of a mammalian neutral sphingomyelinase (20) using distal sequence homology with a bacterial sphingomyelinase. The role of this N-SMase in sphingolipid signaling was also assessed. Upon comparison, we noted several physical, chemical, and functional differences between the two cloned neutral sphingomyelinases. First, there is no amino acid sequence homology between the two N-SMases. Second, our N-SMase has multiple sites for protein post-transcriptional modification, and its physical-chemical characteristics are different. Third, overexpression of our N-SMase in human aortic smooth muscle cells stimulated apoptosis without the presence of an agonist such as oxidized LDL. However, oxidized LDL had an additive effect on apoptosis in cells overexpressing N-SMase. In comparison, the overexpression of N-SMase in HEK and U937 cells did not induce apoptosis independent of the presence and/or absence of TNF-␣ (20). There may be several reasons for the discrepancies in these two studies. One may be related to strategies used in the molecular cloning of these enzymes. It is quite likely that spontaneous apoptosis observed in cells overexpressing N-SMase in our study could be due to the presence of novel death domains absent in other reported N-SMases (20). However, given the complexity of the N-SMase molecule and its vital role in regulating diverse cell signaling pathways, we are not surprised to find different molecular isoforms of N-SMase (a). We predict that, in the future, many additional novel N-SMases will be found that will cleave sphingomyelin but may have diverse biological functions relevant to human health and disease. If so, then such proteins may represent ortholog products of the superfamily of N-SMase genes. Orthologs are defined as genes evolving from a common ancestral gene showing common function (21). We believe that the molecular cloning of N-SMase cDNA will accelerate the process to define its role as a key regulator in apoptosis and other cell regulatory pathways.