Phosphatidate phosphohydrolase catalyzes the hydrolysis of ceramide 1-phosphate, lysophosphatidate, and sphingosine 1-phosphate.

A Mg2+-independent phosphatidate phosphohydrolase was purified from rat liver plasma membranes in two distinct forms, an anionic protein and a cationic protein. Both forms of the enzyme dephosphorylated phosphatidate, ceramide 1-phosphate, lysophosphatidate, and sphingosine 1-phosphate. When assayed at a constant molar ratio of lipid to Triton X-100 of 1:500, the apparent Km values of the anionic phosphohydrolase for the lipid substrates was 3.5, 1.9, 0.4, and 4.0 μM, respectively. The relative catalytic efficiency of the enzyme for phosphatidate, ceramide 1-phosphate, lysophosphatidate, and sphingosine 1-phosphate was 0.16, 0.14, 0.48, and 0.04 liter (min·mg)−1, respectively. The hydrolysis of phosphatidate was inhibited competitively by ceramide 1-phosphate, lysophosphatidate, and sphingosine 1-phosphate. The Ki(app) values were 5.5, 5.9, and 4.0 μM, respectively. The hydrolysis of phosphatidate by the phosphohydrolase conformed to a surface dilution kinetic model. It is concluded that the enzyme is a lipid phosphomonoesterase that could modify the balance of phosphatidate, ceramide 1-phosphate, lysophosphatidate, and sphingosine 1-phosphate relative to diacylglycerol, ceramide, monoacylglycerol, and sphingosine, respectively. The enzyme could thus play an important role in regulating cell activation and signal transduction.

Stimulation of cells with a wide variety of agonists including neurotransmitters and hormones activates phospholipases that generate glycerolipids and sphingolipids that are putative second messengers in signal transduction. Diacylglycerol (DAG) 1 formed by agonist-stimulated hydrolysis of phosphatidylinositol bisphosphate activates protein kinase C which in turn phosphorylates numerous target proteins. In many situations a second and larger increase in DAG mass is generated directly from phosphatidylcholine by a phospholipase C or via phospholipase D and PAP (1,2). Such changes in DAG are involved in the induction of DNA synthesis (1), oocyte maturation (3), and morphological changes in fibroblasts (4). PA has potent mitogenic effects in several cell lines (2,(5)(6)(7); it stimulates the respiratory burst in neutrophils independently of DAG (8,9), and it activates monoacylglycerol acyltransferase (10), phospholipase C-␥ (11), p21 ras (12), and phosphatidylinositol 4-phosphate kinase (13). PA can also serve as a precursor for LPA, which is released from activated platelets, and this process is thought to be involved in local wound repair (14). LPA is a potent mitogen for fibroblasts (2,14). Production of LPA by secretory phospholipase A 2 action on membrane microvesicles may represent a novel pro-inflammatory pathway (15). When added externally to cells LPA activates tyrosine kinases, the Ras-Raf-mitogen-activated protein kinase pathway, focal adhesion kinase, arachidonate production, Ca 2ϩ mobilization, and phospholipase D activity, and it decreases cAMP concentrations (14,16). Many of these effects are similar to those of PA (2,7,14).
Agonist-induced stimulation of sphingomyelinase also plays an important role in signal transduction (17)(18)(19)(20). The generation of intracellular ceramides or addition of cell-permeable ceramides to cells inhibits cell division and can also induce apoptosis (7,(17)(18)(19). Ceramides also play a role in inflammation by modulating the secretion of prostaglandin E 2 (21) and stimulating the secretion of interleukin-2 by lymphocytes (22). Ceramides stimulate specific protein kinases (18) and phosphoprotein phosphatases (17), and they inhibit the activation of phospholipase D (7,23). Deacylation of ceramide produces sphingosine which inhibits protein kinase C (20) and Mg 2ϩ -dependent and -independent PAP activities (24,25). Sphingosine may also activate phospholipase D (26) and DAG kinase (27). Sphingosine also stimulates cell proliferation in fibroblasts (23,28) and induces intracellular Ca 2ϩ mobilization (29,30). When added exogenously to cells sphingosine is converted to SPP which may mediate some, but not all, of the effects of sphingosine (31). SPP is a potent stimulator of cell division and phospholipase D activation (23,31), and it causes Ca 2ϩ mobilization independently of inositol lipid hydrolysis (32). SPP may be a second messenger involved in signal transduction via plateletderived growth factor (33). Ceramide can also be phosphorylated in cells (34), and CerP can be degraded by a phosphohydrolase (35,36). It is not known whether CerP has a physiological role in cell signaling, but exogenous short-chain CerP stimulates DNA synthesis and cell division. This occurs without stimulating mitogen-activated protein kinase and phospholipase D or decreasing cAMP concentrations (36).
The effects of CerP, SPP, PA, and LPA in stimulating DNA synthesis in rat fibroblasts are all antagonized by cell-permeable ceramides, and this is accompanied by an increased dephosphorylation of these compounds (7,23,36 the dephosphorylation of SPP is catalyzed by an enzyme that has properties similar to that of Mg 2ϩ -independent and Nethylmaleimide-insensitive PAP that is located in plasma membranes (23,24). The present work employed PAP purified from rat liver and establishes that a single enzyme hydrolyzes PA, LPA, CerP, and SPP. This enzyme could therefore regulate signaling by these four bioactive lipids and putative second messengers.

EXPERIMENTAL PROCEDURES
Preparation of PA, LPA, CerP, and SPP-The sources of most materials have been described previously (7,23,36,37). Dolichol, dolichol monophosphate, and dioleoylglycerol were purchased from Sigma. Monooleoylglycerol was purchased from NuChek Prep. Long-chain ceramide, sn-1,2-dioleoylglycerol, or monooleoylglycerol (approximately 2 mg) was solubilized in 5 mM cardiolipin, 7.5% octyl-␤-glucopyranoside, 1 mM diethylenetriaminepentaacetic acid by sonication and resuspended in a mixture containing 50 mM imidazole, pH 6.6, 50 mM NaCl, 100 mM MgCl 2 , 1 mM EGTA, and 0.38 units of DAG kinase/ml. The reaction was started with 25 mM ATP. After 17 h at 37°C the reaction was stopped by extracting the lipids with a modified Bligh and Dyer method (38) and by using 2 M KCl and 0.2 M phosphoric acid to separate the phases (36). The chloroform phase was dried under N 2 , and the lipid was applied to thin layer plates of silica gel G in chloroform/methanol (1:1, by volume). For the isolation of CerP and LPA, plates were developed sequentially with chloroform/methanol/ammonia (65:35:7.5, by volume) and chloroform/acetone/acetic acid/methanol/water (50:20:15:10:5, by volume). PA was isolated by developing the plates twice with chloroform/methanol/ ammonia (65:35:7.5, by volume). Radiolabeled phospholipids were prepared in the same manner but with limiting [␥-32 P]ATP (1 mCi, 0.2 nmol) to generate phospholipid with high specific activity. Phospholipid products were identified by using a phosphate-specific spray or autoradiography and co-migration with authentic standards (7,23,36). Phospholipids were eluted from the silica with 5 washes of 3 ml of chloroform/methanol/acetic acid/water (50:39:1:1, by volume). Silica was removed by centrifugation and filtration. After drying under N 2 , the purified phospholipids were resuspended in a known volume of chloroform/methanol (9:1, by volume), quantitated by phosphate analysis (39) or by liquid scintillation counting, and then stored in chloroform at Ϫ20°C. SPP was prepared and purified after acid hydrolysis of CerP (23) on the day before it was used.
Phosphohydrolase Assay-Kinetic assays were performed with aliquots of anionic PAP, purified as described (37), which were stored at Ϫ70°C and thawed once, prior to use. The reaction was started by adding 18 ng of anionic PAP to a warmed mixture (final assay volume 100 l) buffered with 100 mM Tris maleate and 1 mM sodium phosphate, pH 6.5, that contained substrate/Triton X-100 mixed micelles. After incubation at 37°C, the hydrolysis of 32 P-labeled PA, CerP, or LPA was terminated by addition 0.5 ml of 0.1 M HCl in methanol, and the sample was extracted as described above. Radiolabeled P i in the aqueous phase was quantitated by scintillation counting. For the hydrolysis of [ 32 P]SPP, reactions were stopped with 10 l of concentrated formic acid, and 32 P i was separated from SPP by thin layer chromatography in butan-1-ol/acetic acid/water (3:1:1, by volume). All reaction rates were proportional to enzyme protein and time and were performed in triplicate for 10 -15 min such that no more than 15% of the substrate was hydrolyzed. Measurements of phosphohydrolase activity during its purification were made in the presence of 5 mM N-ethylmaleimide with 100 M PA, LPA, CerP, or SPP and 10 mM Triton X-100. Protein was quantitated using the BCA colorimetric method using bovine serum albumin as a standard.

RESULTS AND DISCUSSION
An N-ethylmaleimide-insensitive phosphohydrolase was purified from rat liver and characterized on the basis of its ability to hydrolyze PA (37). Using a similar assay, the anionic form of PAP was also shown to hydrolyze CerP, LPA, and SPP with typical Michaelis-Menten kinetics (Fig. 1, A and B). These assays were performed with mixed micelles consisting of a constant 1:500 molar ratio of lipid to Triton X-100. Analysis of the results using a Hanes-Woolf plot (Fig. 1, C and D) yielded K m(app) for the bulk concentrations of PA, CerP, LPA, and SPP of 3.5, 1.9, 0.4, and 4.0, M respectively. V max values under these conditions were 0.55, 0.26, 0.19, and 0.15 mol (min⅐mg) Ϫ1 , and the relative catalytic efficiency (V max /K m ) of the enzyme toward the various substrates was 0.16, 0.14, 0.48, and 0.04 liter (min⅐mg) Ϫ1 , respectively. These results indicate that LPA may be a preferred substrate since it is dephosphorylated three times more efficiently than PA and CerP. Although the maximal rate of SPP hydrolysis was similar to the other lipids, the relative efficiency of SPP hydrolysis was 3.5-4-fold less than that of PA or CerP. This might be explained partly because the hydrolysis of SPP gives rise to sphingosine which can inhibit PAP (24). Sphingosine is a mixed type inhibitor of PAP. 2 In the present experiments the hydrolysis of SPP was limited to less than 15% which therefore minimized the extent of product inhibition.
When various concentrations of CerP, LPA, or SPP were included in the assay with different concentrations of [ 32 P]PA, hydrolysis of PA was decreased in a dose-dependent manner by each of the phospholipids (Fig. 2, A-C). A linear secondary plot of the slopes from the inhibition curve versus the concentration of unlabeled phospholipid (Fig. 2, D-F) indicates competitive inhibition (40). Under the conditions employed, the K i(app) for CerP, LPA, and SPP were 5.5, 5.5, and 4.0 M. However, sphingosine formed from SPP hydrolysis could have contributed to the some of the inhibition that was seen at the lower [ 32 P]PA concentrations (Fig. 2C). The inhibitory constants indicate that the different phospholipids have similar affinities for the enzyme, which correlates well with the respective K m(app) , except for that of LPA. The reason for this latter difference is not clear.
The kinetics of substrate hydrolysis by anionic PAP were examined further using PA as a substrate. The maximal velocity of PA hydrolysis increased with the bulk concentrations of PA until it plateaued between 100 and 200 M (Fig. 3A). When surface dilution kinetics (41, 42) were evaluated at 100 M PA, a Hanes-Woolf plot of the results was linear (Fig. 3B)  PA when compared with other interfacial enzymes (42).
During the purification of PAP, all fractions that contained catalytic activity toward PA also hydrolyzed the other lipid substrates. When fractions from the purification (37) were assayed under identical conditions the fold purification of phosphohydrolase activity toward each of the substrates remained essentially constant. For example, the relative increase in specific activity from rat liver homogenate for the N-ethylmaleimide-insensitive phosphohydrolase toward CerP, LPA, and SPP in the anionic PAP fraction was 52, 54, and 46%, respectively, of that toward PA. Although highly purified, anionic PAP was not a homogeneous protein sample. Therefore, we evaluated the ability of homogeneous cationic PAP (37) to dephosphorylate the four lipid substrates. Cationic PAP was not used for the extensive kinetic analyses because the activity was relatively low and unstable (37). Freshly prepared cationic PAP, which we concluded to be a desialated form of anionic PAP, also hydrolyzed PA, CerP, LPA, and SPP at similar rates (0.16, 0.13, 0.05, and 0.08 mol (min⅐mg) Ϫ1 , respectively). Furthermore, anionic PAP was purified further using affinity purified antibodies prepared against cationic PAP (37). The immunoprecipitated protein also hydrolyzed PA, CerP, LPA, and SPP at similar rates (0.031, 0.024, 0.021, and 0.009 nmol/min, respectively). These results indicate that homogeneous cationic and anionic PAP (37) dephosphorylated all four phospholipids.
Kanoh et al. (43) reported that LPA did not inhibit the Mg 2ϩ -independent and N-ethylmaleimide-insensitive PAP which they purified from porcine thymus. This result may indicate the existence of an iso-enzyme of PAP that differs in substrate specificity from that isolated from rat liver (37,44). Boudker and Futerman (35) described an N-ethylmaleimideinsensitive CerP phosphohydrolase in plasma membrane fractions of rat liver. This activity was inhibited by PA, but in a noncompetitive manner, and CerP did not inhibit the hydrolysis of PA in their membrane fractions. They therefore suggested the existence of distinct phosphohydrolase activities for PA and CerP. By contrast, the PAP that we purified from rat liver plasma membranes showed competitive inhibitory kinetics toward these two substrates.
It has also been reported that a dolichol monophosphate phosphatase found in the plasma membrane of cells is inhibited by PA and LPA (45,46). To evaluate whether the hydrolysis of PA by anionic PAP was inhibited by dolichol monophosphate, increasing concentrations of the latter lipid were included in an assay of [ 32 P]PA hydrolysis. Dolichol monophosphate inhibited PA hydrolysis but much less so than did an equivalent amount of CerP (Fig. 4A). Dolichol monophosphate was hydrolyzed by anionic PAP but much less efficiently than the other phospholipids. No detectable hydrolysis of dolichol monophosphate was seen in 1-4-h incubations with anionic PAP (results not shown). Hydrolysis by anionic PAP did occur (Fig. 4B) when dolichol monophosphate was incubated for 16 h with 250 times more protein than was used in the kinetic assays. Furthermore, the activity of dolichol monophosphate phosphatase is decreased 70% by 8 mM P i (47), whereas the activity of PAP, which was purified using a variety of buffers including 10 mM sodium phosphate, is not affected by P i . Therefore, we conclude that the PAP that we purified is unlikely to be the dolichol monophosphate phosphatase and that dolichol monophosphate is a poor substrate and inhibitor of anionic PAP. Rat liver PAP does not hydrolyze glycerol 3-phosphate, phosphatidylcholine, phosphatidylethanolamine, phosphatidylserine, phosphatidylinositol, nor diphosphatidylglycerol (24). Additionally, glycerol 3-phosphate (24, 44), phosphatidylcholine, phosphatidylserine, phosphatidylinositol, and diphos- phatidylglycerol (44) do not inhibit PAP activity.
We therefore conclude that the PAP that we purified is a multi-functional lipid phosphomonoesterase that hydrolyzes PA, CerP, LPA, and SPP using the same active site. All of these lipids are bioactive when added externally to cells (2,5,7,23,31,32). PA and SPP are also thought to be produced as second messengers in signal transduction (2,16,36). PAP could therefore play an important role in destroying these signals and counteracting their mitogenic effects. In the case of PA, CerP, and SPP, the products of the PAP reaction are DAG, ceramide, and sphingosine, respectively. The latter lipids are also bioactive (1,2,(17)(18)(19)(20); thus, this enzyme may regulate the balance of important lipid mediators of cell activation and signal transduction. The observation that PA, CerP, LPA, and SPP are mutually competitive for PAP offers a further dimension in which there may be "cross-talk" between the glycerolipid and sphingolipid signaling pathways.