An Alternative Splicing Form of Phosphatidylserine-specific Phospholipase A1 That Exhibits Lysophosphatidylserine-specific Lysophospholipase Activity in Humans*

Phosphatidylserine-specific phospholipase A1 (PS-PLA1), which acts specifically on phosphatidylserine (PS) and 1-acyl-2-lysophosphatidylserine (lyso-PS) to hydrolyze fatty acids at the sn-1 position of these phospholipids, was first identified in rat platelets (Sato, T., Aoki, J., Nagai, Y., Dohmae, N., Takio, K., Doi, T., Arai, H., and Inoue, K. (1997) J. Biol. Chem. 272, 2192–2198). In this study we isolated and sequenced cDNA clones encoding human PS-PLA1, which showed 80% homology with rat PS-PLA1 at the amino acid level. In addition to an mRNA encoding a 456-amino acid product (PS-PLA1), an mRNA with four extra bases inserted at the boundary of the exon–intron junction was detected in human tissues and various human cell lines. This mRNA is most probably produced via an alternative use of the 5′-splicing site (two consensus sequences for RNA splicing occur at the boundary of the exon–intron junction) and encodes a 376-amino acid product (PS-PLA1ΔC) that lacks two-thirds of the C-terminal domain of PS-PLA1. Unlike PS-PLA1, PS-PLA1ΔC hydrolyzed exclusively lyso-PS but not PS appreciably. Any other phospholipids such as phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidic acid (PA), and their lyso derivatives were not hydrolyzed at all. These data demonstrated that PS-PLA1ΔC exhibits lyso-PS-specific lysophospholipase activity and that the C-terminal domain of PS-PLA1 is responsible for recognizing diacylphospholipids. In addition, human PS-PLA 1 gene was mapped to chromosome 3q13.13–13.2 and was unexpectedly identical to the nmdgene, which is highly expressed in nonmetastatic melanoma cell lines but poorly expressed in metastatic cell lines (van Groningen, J. J., Bloemers, H. P., and Swart, G. W. (1995) Cancer Res. 55, 6237–6243).

Phosphatidylserine (PS) 1 in cell membranes is known to be an essential cofactor for the activation of protein kinase C (1) and for blood coagulation (2). More recently, PS has been shown to regulate the activity of various enzymes, such as c-Raf-1 protein kinase (3), nitric oxide synthase (4), Na ϩ /K ϩ -ATPase (5), dynamin GTPase (6), and diacylglycerol kinase (7). PS is predominantly located on the inner leaflet of plasma membranes in various types of cells (8) but appears on the outer leaflet after stimulation by various factors such as cytokines (9,10), inflammatory reactions, and platelet activation (8,(11)(12)(13). Surface-exposed PS has also been shown to act as a signal for the removal of damaged or aged cells by the reticuloendothelial system and is observed in cells undergoing apoptosis (14,15). Thus, the exposure of PS on the cell surface must be tightly regulated.
Another serine-containing phospholipid, lyso-PS, is implicated to act as a lipid mediator under pathophysiological conditions (16). For example, lyso-PS is demonstrated to interact with local mast cells (17), producing specific and stereoselective activation (18). It also induces transient increases in cytosolic free Ca 2ϩ ([Ca 2ϩ ] i ) in ovarian and breast cancer cells (19) and lyso-PS; 2-acyl-1-lyso-PS with unsaturated fatty acids especially inhibits mitogen-induced T cell activation (20). Lyso-PS is present in human serum, the aqueous humor and the lachrymal gland fluid of the eye (21). It is likely to be produced from PS by phospholipase A 1 or A 2 , but the precise mechanisms of lyso-PS production and elimination in vivo remain to be clarified.
We previously demonstrated that a serine phospholipid-specific phospholipase A 1 /lysophospholipase is secreted from rat platelets when they are activated (22,23). Very recently we purified this enzyme from rat platelets and cloned its cDNA (24). Although this novel phospholipase, named PS-PLA 1 , has a similar structure to members of the lipase family such as hepatic, pancreatic, and lipoprotein lipases, PS-PLA 1 does not hydrolyze triacylglycerol but specifically acts on PS or 1-acyl-2-lyso-PS to hydrolyze fatty acids at the sn-1 position of these phospholipids (24). Thus, PS-PLA 1 mediates two types of reactions, producing 2-acyl-1-lyso-PS from PS and eliminating 1-acyl-2-lyso-PS. It is secreted from cells and has an affinity for heparin, like other members of the lipase family (24). It is likely that PS-PLA 1 is involved in regulating PS/lyso-PS-dependent reactions under physiological conditions. Indeed, PS-PLA 1 can hydrolyze PS in rat platelets when the cell is activated (23). The physiological function of PS-PLA 1 , however, is still unknown and requires further investigation. As a part of our continuing study of the physiological role of PS-PLA 1 , we have isolated the cDNA for human PS-PLA 1 . In the course of the study, we detected an alternative splicing form of human PS-PLA 1 and identified it as a lyso-PS-specific lysophospholipase.

MATERIALS AND METHODS
cDNA Cloning of Human PS-PLA 1 -With the polymerase chain reaction (PCR), we amplified cDNA using a human liver gt11 cDNA library (CLONTECH, Palo Alto, CA) as a template, and the following primers, based on the EST clone: 5Ј-CACGAGGATTTCCAGCTCAG-3Ј and 5Ј-CCCTGTACATGTTTCTATTG-3Ј.
The resulting 1.7-kb PCR fragment was used as a DNA probe to screen the human liver gt11 cDNA library as described previously (24). One positive clone was identified, and its insert cDNA was subcloned into the pBluescript II SK phagemid vector (Stratagene, La Jolla, CA). The DNA sequence was determined by the dideoxynucleotide chain termination method using a Dye Terminator Cycle Sequencing FS Ready Reaction Kit (Perkin-Elmer) and an ABI PRISM 377 DNA sequencer (Perkin-Elmer).
Southern Blot Analysis-Human and rat genomic DNAs (12 g each) were digested with the appropriate restriction enzymes (50 units each) for 12 h then run on a 0.8% SeaKem Gold Agarose gel (FMC BioProducts, Rockland, ME). The DNA was then transferred under capillary pressure to a Hybond-N nylon hybridization transfer membrane (Amersham Pharmacia Biotech). The relevant DNA probes were labeled by random priming with [␣-32 P]dCTP. Hybridization was carried out at 65°C for 4 h in a rapid-hybridization buffer (Amersham Pharmacia Biotech). The blot was rinsed in 2ϫ SSC (1ϫ SSC ϭ 0.15 M NaCl and 0.015 M sodium citrate) at room temperature for 5 min, then washed twice in 0.5ϫ SSC, 0.1% SDS at 65°C for 40 min. The blot was exposed to Kodak X-Omat AR film at Ϫ80°C with an intensifying screen for 12 h.
Northern Blot Analysis-Human multiple tissue Northern blots were purchased from CLONTECH. Total RNA from human platelets (10 g/lane) was separated by 1% agarose-formaldehyde gel electrophoresis and transferred onto Hybond-N. Hybridization with the human PS-PLA 1 probe and washing were carried out as described for the Southern blot analysis. The blots were rehybridized with a glyceraldehyde-3-phosphate dehydrogenase cDNA probe (CLONTECH) as an internal standard.
Amplification and Restriction Enzyme Digestion of Human PS-PLA 1 cDNA-Total RNA (1 g) obtained from various human tissues and cell lines was used as a template in a reverse transcriptase (RT)-PCR. A Ready-To-Go T-Primed First-Strand Kit (Amersham Pharmacia Biotech) was used according to the manufacturer's protocol to synthesize the first-strand cDNA. DNA fragments were amplified by PCR using 3 l of the first-strand reaction product as a template and the following set of primers: 5Ј-ACAAGGACACCAACATCGAGGTTACCTTCC-3Ј (nucleotide positions 1055-1084 of the cDNA encoding human PS-PLA 1 ) and 5Ј-CAGTCACACTTGCTTGTAAGTTCACTGG-3Ј (nucleotide positions 1312-1339). The RT-PCR products (285 or 289 bp) were digested with BsrGI for 4 h at 60°C then subjected to 4% NuSieve 3:1 Agarose (FMC BioProducts) gel electrophoresis.
Expression of Human PS-PLA 1 in Sf9 Cells-DNA fragments encoding the normal and truncated forms of human PS-PLA 1 were amplified by PCR using two synthetic oligonucleotides. The 5Ј oligonucleotide, 5Ј-CAGCGGATCCATGCCCCCAGGTCCCTGGGA-3Ј, contained the BamHI site in addition to the human PS-PLA 1 sequence, whereas the 3Ј oligonucleotide, 5Ј-AAAAAAGCTTGCAGGGAGATGTGTCCTGCCCA-GG-3Ј, contained the HindIII site and a complementary human PS-PLA 1 sequence. The amplified cDNA fragments were subcloned into the BamHI and HindIII sites of the pFASTBAC1 expression vector (Life Technologies, Inc.) to generate donor plasmids. Recombinant viruses were then prepared using the BAC-TO-BAC baculovirus expression system (Life Technologies, Inc.) according to the manufacturer's protocol. The resulting recombinant baculovirus was used to infect Sf9 cells. Four days after infection, the culture supernatant of the infected cells was collected, and PS-PLA 1 activity was determined as described below.
Column Chromatography-The culture supernatant of Sf9 cells or human plasma was loaded onto a HiTrap Heparin (1 ml) affinity column (Amersham Pharmacia Biotech) that had been preequilibrated with 100 mM NaCl containing 10 mM Tris-HCl (pH 7.4) using the fast protein liquid chromatography system (Amersham Pharmacia Biotech). The column was washed thoroughly with this buffer and eluted with a linear gradient of NaCl (0.1-1.1 M). The PS-PLA 1 activity in each fraction was determined as described above.
Analysis of the Gene Structure of Human PS-PLA 1 -PCR was performed using human genomic DNA as a template and the primers 5Ј-GAGAAACAAGGACACCAACATCGAGGTTACC-3Ј and 5Ј-CATTG-TAGGGTGGCATGGGCTATGATTCC-3Ј, based on the sequence around the insertion. The DNA sequence of the resulting 3.5-kb fragment was determined by direct sequencing using the above oligonucleotides as sequence primers.
Chromosomal Preparation and in Situ Hybridization-We used the direct R-banding fluorescence in situ hybridization method to determine the human chromosomal location of the PS-PLA 1 gene. We prepared R-banded chromosomes and performed fluorescence in situ hybridization as described previously (25). A mitogen-stimulated lymphocyte culture was synchronized by thymidine blockade, with 5-bromodeoxyuridine incorporated during the late replication stage to allow differential replication staining after removing excessive thymidine. R-band staining was performed by exposing chromosome slides to UV light after staining with Hoechst 33258.
The chromosome slides were hardened at 65°C for 2 h, then denatured at 70°C in 70% formamide in 2 ϫ SSC and dehydrated in a 70, 85, 100% ethanol series at 4°C. The 1.75-kb human cDNA fragment inserted into pBluescript II SK was used as a probe. The cDNA fragments were labeled by nick translation with biotin 16-dUTP (Boehringer Mannheim) following the manufacturer's protocol. The labeled DNA fragment was ethanol-precipitated with salmon sperm DNA and Escherichia coli tRNA, then denatured at 75°C for 10 min in 100% formamide. The denatured probe was mixed with an equal volume of hybridization solution to produce a final concentration of 50% formamide, 2ϫ SSC, 10% dextran sulfate, and 2 mg/ml bovine serum albumin (Sigma). Twenty l of this mixture containing 250 ng of labeled DNA was placed onto the denatured slide, covered with Parafilm, and incubated overnight at 37°C. The slides were washed in 50% formamide in 2ϫ SSC at 37°C, then in 2ϫ SSC, and finally in 1ϫ SSC (for 20 min each time) at room temperature. After rinsing in 4ϫ SSC, they were incubated under a coverslip with goat anti-biotin antibody (Vector Laboratories) at a dilution of 1:500 in 1% bovine serum albumin, 4ϫ SSC, then with 4ϫ SSC (for 5 min each time). The slides were then stained with fluorescein-anti-goat IgG (Nordic Immunology) at a dilution of 1:500 for 1 h at 37°C. After washing with 4ϫ SSC, then 0.1% Nonidet P-40 in 4ϫ SSC, and again with 4ϫ SSC (for 10 min each time) on a shaker, the slides were rinsed with 2ϫ SSC and stained with 0.75 g/ml propidium iodide. Excitation at a wavelength of 450 -490 nm (Nikon filter set B-2A) and near 365 nm (UV-2A) was then measured. Kodak Ektachrome ASA100 films were used for microphotography.

Molecular
Cloning of Human PS-PLA 1 -Southern blot analysis showed the existence of a single gene copy in humans and rats (data not shown). We used the BLASTN program to search the protein data base for sequences similar to that of rat PS-PLA 1 and found two human EST sequences with homology to rat PS-PLA 1 on the GenBank™ DNA data base (GenBank™ accession numbers T96213 (5Ј region) and T96131 (3Ј region)). We prepared one set of PCR primers corresponding to the 5Ј-untranslated region and the 3Ј-untranslated region based on the DNA sequences of these EST clones. Because the EST clones were derived from a human fetal liver and spleen cDNA library, we performed PCR using human liver cDNA as a template. We then used the resulting 1.7-kb DNA fragment to screen a human liver cDNA library, and one positive clone was isolated. DNA sequence analysis revealed that this clone has a sequence highly homologous with that of rat PS-PLA 1 , covering the whole region corresponding to the open reading frame (Fig.  1a). This cDNA clone contained a 1368-bp open reading frame that encoded 456 amino acids, starting with an initiation codon His-236, which may make up a catalytic triad in rat PS-PLA 1 (24), are conserved between the rat and human enzymes. The amino acid sequences around these three amino acids were also highly conserved between the two species. A putative lid, which is composed of 22 or 23 amino acid residues in most of the other lipases and suggested to be involved in both substrate recognition and surface activation, is composed of 12 residues, as for the rat enzyme.
When the coding region of PS-PLA 1 mRNA from various human tissues and cell lines was amplified using RT-PCR and analyzed by DNA sequencing, an aberrant mRNA was frequently detected in addition to PS-PLA 1 mRNA. These clones, designated PS-PLA 1 ⌬C, had four additional bases (GTAC) inserted at nucleotide position 1122 of PS-PLA 1 , and the other nucleotide sequence of this clone was identical to that of PS-PLA 1 . As the insertion of these four bases caused a frameshift (Fig. 1b), PS-PLA 1 ⌬C loses about 80 amino acid residues that are present in C terminus of PS-PLA 1 , but the amino acids forming a catalytic triad including Ser-142, Asp-166, and His-236 are all preserved, suggesting that PS-PLA 1 ⌬C is catalytically active (Fig. 1c).
Production of Two Isoforms by Alternative Splicing-We analyzed the structure of the gene and nucleotide sequences around the insertion point, as described under "Materials and Methods." The four extra bases, GTAC, inserted in the PS-PLA 1 ⌬C cDNA were found at the beginning of the intron (Fig.  2a). Two separate consensus sequences for RNA splicing (gt(a/ g)agt) (Fig. 2b) can be found at this exon-intron boundary. It is likely from this information that mRNA for PS-PLA 1 is produced when the first consensus site (gtacgt) is used, and mRNA for PS-PLA 1 ⌬C is produced when the second (gtaagt) is used (Fig. 2c). Thus, we concluded that PS-PLA 1 and PS-PLA 1 ⌬C arise from alternative use of the 5Ј-splicing donor sites (GT).
PS-PLA 1 ⌬C as a Lyso-PS-specific Lysophospholipase-To determine whether the products encoded by the two mRNAs have PS-or lyso-PS-hydrolyzing activity, recombinant PS-PLA 1 and PS-PLA 1 ⌬C were produced using a baculovirus system and examined for activities. As shown in Fig. 3a, we detected appreciable lyso-PS-hydrolyzing activity in the supernatant of Sf9 cells infected with either PS-PLA 1 or PS-PLA 1 ⌬C recombinant viruses. This shows that both PS-PLA 1 and PS-PLA 1 ⌬C are secreted from these cells and possess enzymatic activity to hydrolyze lyso-PS. By contrast, no PS-hydrolyzing activity was observed in the supernatant of cells infected with the PS-PLA 1 ⌬C recombinant virus (Fig. 3a). We also examined the substrate specificity using lyso-PS, 1-oleoyl-2-lysophosphatidylcholine, 1-oleoyl-2-lysophosphatidylethanolamine, and 1oleoyl-2-lysophosphatidic acid. As shown in Fig. 3b, both PS-PLA 1 and PS-PLA 1 ⌬C hydrolyzed lyso-PS, but neither produced appreciable hydrolysis of other lysophospholipids. Thus, PS-PLA 1 hydrolyzes both PS and lyso-PS, but PS-PLA 1 ⌬C exhibits lyso-PS-specific lysophospholipase activity.
We have previously demonstrated that rat PS-PLA 1 possesses a heparin binding site, because this enzyme absorbed to a heparin-Sepharose column (24). We found that both human PS-PLA 1 and PS-PLA 1 ⌬C were also absorbed to the column and eluted with increasing concentrations of NaCl. The elution profile of PS-PLA 1 ⌬C on heparin-Sepharose column chromatography was almost identical to that of PS-PLA 1 (Fig. 3c), indicating that the heparin binding site of PS-PLA 1 is located in N-terminal region of this enzyme. The capital letters represent the exon, and the small letters represent the intron for PS-PLA 1 . The four bases, gatc (in italics), were found at the 5Ј-end of the exon-intron boundary. The nucleotide numbers in Fig. 1 were shown for PS-PLA 1 and PS-PLA 1 ⌬C, respectively. b, possible exon-intron structure of the human PS-PLA 1 gene. The 5Ј-splicing donor sites are underlined. Consensus sequences for RNA splicing are shown in the lower panel. The nucleotide numbers in Fig. 1 were shown for PS-PLA 1 and PS-PLA 1 ⌬C, respectively. c, model for alternative splicing of PS-PLA 1 . PS-PLA 1 mRNA is produced when the first 5Ј-splicing donor site (gtacgt) is used; mRNA for the PS-PLA 1 ⌬C is produced when the second site (gtaagt) is used. The nucleotide numbers in Fig. 1 were shown for PS-PLA 1 and PS-PLA 1 ⌬C, respectively.

Expression of PS-PLA 1 and PS-PLA 1 ⌬C in Human Tissues
and Cells-First we used Northern blot analysis to examine the expression of PS-PLA 1 mRNA (PS-PLA 1 and PS-PLA 1 ⌬C) using the cDNA corresponding to the coding region as a probe (Fig. 4). Transcripts of 1.9-kb messages were seen in most of the tissues examined, with the highest expression in the liver and prostate gland. PS-PLA 1 mRNAs were not detected in leukocytes or platelets (Fig. 4). Because PS-PLA 1 was first identified in rat platelets, we analyzed the expression of PS-PLA 1 in human platelets by RT-PCR analysis and an assay for PS-PLA 1 activity. No appreciable PS-PLA 1 activity or transcripts were detected in human platelets (data not shown).
The PS-PLA 1 transcript present in human tissues was detected as a single band on Northern blot analysis (Fig. 4). To determine the relative amounts of PS-PLA 1 and PS-PLA 1 ⌬C mRNAs in human tissues, RT-PCR analysis was performed using total RNAs derived from various human tissues. Because the insertion of the four extra bases generates a new restriction enzyme site, BsrGI (TGTACA) (Figs. 1 and 2), we quantified the frequency of the two mRNAs by digesting the RT-PCR products with BsrGI, as described under "Materials and Methods." As shown in Fig. 5, both PS-PLA 1 and PS-PLA 1 ⌬C transcripts were detected in various human tissues, including skeletal muscle, kidney, small intestine, spleen, and testis. The amount of PS-PLA 1 ⌬C in these tissues was about 10 to 20% that of the PS-PLA 1 level (Fig. 5). We also examined the expression of PS-PLA 1 and PS-PLA 1 ⌬C in several human cell lines. mRNAs for both PS-PLA 1 and PS-PLA 1 ⌬C were detected in human fibroblast, keratinocyte, melanoma, HepG2, and HeLa cells. Thus, human tissues and cell lines express both PS-PLA 1 and PS-PLA 1 ⌬C, although the relative amount of PS-PLA 1 ⌬C is lower than that of PS-PLA 1 .
Chromosome Mapping of PS-PLA 1 -The location of the PSbaculovirus were applied onto a HiTrap heparin fast protein liquid chromatography column, and PS-PLA 1 and PS-PLA 1 ⌬C were eluted using a linear gradient of NaCl. The lysophospholipase activity of each fraction was measured using lyso-PS as a substrate. PLA 1 gene on the human chromosome was assigned by direct R-banding fluorescence in situ hybridization using a human cDNA fragment as a probe. The PS-PLA 1 gene was localized to human chromosome 3q13.13-13.2 (Fig. 6). DISCUSSION PS-PLA 1 ⌬C as a Lyso-PS-specific Lysophospholipase-So far it has been reported that the various phospholipases for which the cDNA structure is already known, cytosolic PLA 2 (26,27), hepatic lipase (28), guinea pig pancreatic phospholipase (29), phospholipase B (30), Campylobacter coli PLA 2 (31), and PS-PLA 1 (24), show both PLA and lysophospholipase activity. Among these phospholipases, hepatic lipase (32), guinea pig phospholipase (33), phospholipase B, and PS-PLA 1 exhibited both PLA 1 and lysophospholipase activity. Saito et al. (30) reported that the PLA 1 and PLA 2 activities of Penicillium notatum phospholipase B are lost completely by limited proteolysis, whereas its lysophospholipase activity remains unchanged (30), although the protein structures required for each activity have not been characterized yet. Separating lyso-PSspecific lysophospholipase activity from PS-phospholipase A 1 is the first example of determining a protein structure required for lysophospholipase and PLA 1 activity.
The lipases (lipoprotein and pancreatic lipase) are reported to be composed of two domains (the N-terminal and C-terminal domains) (34,35). PS-PLA 1 is also predicted to possess the similar two domains (24), but in PS-PLA 1 ⌬C, two-thirds of the C-terminal domain is lost (Fig. 1). From our results (Fig. 3, b  and c), we conclude that the N terminus of PS-PLA 1 , which is conserved between PS-PLA 1 and PS-PLA 1 ⌬C (Fig. 1c), carries a structure that recognizes the serine residues of serine phospholipids (PS and lyso-PS) and is responsible for heparin binding. We showed in this study that PS-PLA 1 ⌬C fails to hydrolyze PS. In addition, our preliminary experiment shows that when we incorporated lyso-PS as a substrate of PS-PLA 1 ⌬C into PC liposomes, its enzymatic activity was effectively inhibited (data not shown). This indicates that PS-PLA 1 ⌬C may not be able to recognize the lipid surface of PS micelles and that the C terminus of PS-PLA 1 , which is missing in PS-PLA 1 ⌬C, plays an important role in recognizing diacyl-PS in the lipid bilayers. Several studies have in fact suggested an important role of the C terminus of lipases in recognizing lipid surfaces (36,37); however, as PS-PLA 1 does not show any amino acid homology with other lipases in the C-terminal domain (30), it is also possible that this domain of PS-PLA 1 has another specialized function(s).
It is generally known that a lid is involved in the interfacial activation (38) and/or substrate recognition (39). The lid of PS-PLA 1 is composed of 12 amino acid residues (Fig. 1), whereas those of many lipases are composed of 22 or 23 residues. This short lid may not play a role in the interfacial activation, because it was demonstrated that pancreatic lipase loses the ability to surface activation when its lid, composed of 23 amino acids, was shortened by site-directed mutagenesis (38). Rather this short lid may be involved in substrate recognition. In fact, both PS-PLA 1 and PS-PLA 1 ⌬C possess the same lid structure.
Expression of PS-PLA 1 and PS-PLA 1 ⌬C in Human Tissues-We first purified PS-PLA 1 from rat platelets. In this study, we showed that PS-PLA 1 is expressed in various human tissues. However, human platelets, leukocytes (Fig. 4), and red blood cells do not express PS-PLA 1 appreciably. We could not detect PS-PLA 1 activity in platelets from rabbits, cattle, or pigs. Thus, PS-PLA 1 is expressed in platelets in a speciesspecific manner. It is not yet known whether the alternative form (PS-PLA 1 ⌬C) exists in species other than humans, although the nucleotide sequence of the rat PS-PLA 1 gene around the exon-intron boundary is identical to that of the human PS-PLA 1 gene (Fig. 2a). Thus, we expect that the alternative form will be present in the rat and in other species.
Possible Roles of PS-PLA 1 and PS-PLA 1 ⌬C-The physiological significance of PS-PLA 1 is not yet known. The chromosomal location of the human PS-PLA 1 gene does not suggest a possible link to human disease. It is noteworthy, however, that the nucleotide sequence of human PS-PLA 1 is identical to that of the nmd gene product (40). This gene is predominantly expressed in nonmetastatic human melanoma cell lines, with a lower expression level in metastatic cell lines (40). Any molecular link between PS and the metastasis of tumor cells is totally unknown at present, but identifying PS-PLA 1 as the nmd gene product suggests that serine phospholipids (PS or lyso-PS) are involved in the metastatic process of tumor cells. Calderon et al. (4) reported that some cancer cells secrete PS, and it impairs macrophage cytotoxicity by inhibiting the production of nitric oxide from macrophages. Thus they speculated that cancer cells escape from macrophage recognition by secreting PS. PS-PLA 1 may hydrolyze such PS and increase the activity of macrophages in self-defense against tumor cells.
PS-PLA 1 mediates three types of reaction to eliminate PS and 1-acyl-2-lyso-PS and to produce 2-acyl-1-lyso-PS. PS- The PCR products were subjected to BsrGI digestion, as the BsrGI site only occurs in PS-PLA 1 ⌬C. In the control experiment, plasmids carrying PS-PLA 1 or PS-PLA 1 ⌬C were used as a template. After BsrGI digestion, the PS-PLA 1 plasmid gave rise to a single 285 bp band, whereas the PS-PLA 1 ⌬C plasmid gave rise to two bands of 220 and 69 bp. Bands of 220 and 69 bp (derived from PS-PLA 1 ⌬C) were detected in addition to a band of 285 bp (derived from PS-PLA 1 ) in various tissues and cell lines.  1 ⌬C, by contrast, has only an ability to eliminate 1-acyl-2lyso-PS. Once 2-acyl-1-lysophospholipids are produced by the PLA 1 reaction, the fatty acid at the sn-2 position readily migrates to the sn-1 position, which results in production of 1-acyl-2-lysophospholipids (41). PS-PLA 1 ⌬C might eliminate such 1-acyl-2-lyso-PS in vivo, or it might eliminate 1-acyl-2lyso-PS, which is produced by the PLA 2 reaction. Because serine phospholipids are suggested to be involved in various pathophysiological conditions (16), both PS-PLA 1 and PS-PLA 1 ⌬C might be key enzymes that regulate the production and elimination of serine phospholipids. Further study is required to clarify the physiological functions of PS-PLA 1 and PS-PLA 1 ⌬C.