Purification and Characterization from Rat Kidney Membranes of a Novel Platelet-activating Factor (PAF)-dependent Transacetylase That Catalyzes the Hydrolysis of PAF, Formation of PAF Analogs, and C2-ceramide*

We have previously identified two enzyme activities that transfer the acetyl group from platelet-activating factor (PAF) in a CoA-independent manner to lysoplasmalogen or sphingosine in HL-60 cells, endothelial cells, and a variety of rat tissues. These were termed as PAF:lysoplasmalogen (lysophospholipid) transacetylase and PAF:sphingosine transacetylase, respectively. In the present study, we have solubilized and purified this PAF-dependent transacetylase 13,700-fold from rat kidney membranes (mitochondrial plus microsomal membranes) based on the PAF:lysoplasmalogen transacetylase activity. The mitochondria and microsomes were prepared and washed three times, then solubilized with 0.04% Tween 20 at a detergent/protein (w/w) ratio of 0.1. The solubilized fractions from mitochondria and microsomes were combined and subjected to sequential column chromatographies on DEAE-Sepharose, hydroxyapatite, phenyl-Sepharose, and chromatofocusing. The enzyme was further purified by native-polyacrylamide gel electrophoresis (PAGE) and affinity gel matrix in which the competitive inhibitor of the enzyme, 1-O-hexadecyl-2-N-methylcarbamyl-sn-glycero-3-phosphoethanolamine was covalently attached to the CH-Sepharose. On SDS-PAGE, the purified enzyme showed a single homogeneous band with an apparent molecular mass of 40 kDa. The purified enzyme catalyzed transacetylation of the acetyl group not only from PAF to lysoplasmalogen forming plasmalogen analogs of PAF, but also to sphingosine producing N-acetylsphingosine (C2-ceramide). In addition, this enzyme acted as a PAF-acetylhydrolase in the absence of lipid acceptor molecules. These results suggest that PAF-dependent transacetylase is an enzyme that modifies the cellular functions of PAF through generation of other diverse lipid mediators.

Platelet-activating factor (PAF), 1 1-O-alkyl-2-acetyl-sn-glyc-ero-3-phosphocholine (alkylacetyl-GPC), is a potent lipid mediator with a wide variety of biological activities related to physiological and pathological phenomena (1,2). PAF is produced by either de novo or remodeling pathway (1). In addition, we found the PAF could also be metabolized by a novel pathway catalyzed by membrane-associated transacetylase that transfers the acetate group of PAF to lysoplasmalogen in HL 60 cells (3). This enzyme is CoA-independent and transfers the acetyl group from PAF to a variety of lysophospholipids acceptors in the order of radyl-GPC Ͼ radyl-glycerophosphoethanolamine (GPE) Ͼ acyl-glycerophosphoserine Ͼ acyl-glycerophosphoinositol Ͼ acyl-glycerophosphate Ͼ alkyl-glycerophosphate Ͼ fatty alcohol, whereas alkylglycerol, acylglycerol, or cholesterol are inactive as acceptors. This PAF-dependent transacetylase is participated in the biosynthesis of acyl analog of PAF (4), which is the predominant molecular species of PAF in hematopoietic cells including endothelial cells, mast cells, and basophils, etc. (5). In endothelial cells, PAF-dependent transacetylase activity is regulated by phosphorylation/ dephosphorylation (4).
Recently, we have demonstrated that a similar PAF-dependent transacetylase transfers the acetyl group of PAF to sphingosine in HL 60 cells (6,7). This enzyme activity appears to be responsible for the presence of acetylsphingosine (C 2 -ceramide) in the biological systems. For instance, C 2 -ceramide occurred in the micromolar range in undifferentiated and differentiated HL-60 cells (6). This is the concentration range that C 2 -ceramide has been shown to exert as a second messenger and a lipid mediator by many investigators (8,9). Since C 2 -ceramide has many biological activities that differ from PAF and sphingosine, therefore, this enzyme may serve as a modifier for the functions of PAF and sphingosine (6). Both enzyme activities are also found in rat tissues (6). Rat kidney membrane fractions have the highest PAF:sphingosine transacetylase activity, while both rat kidneys and lung have the highest PAF: lysoplasmalogen transacetylase activity.
To elucidate the relationship of both enzyme activities, we attempted to purify the transacetylase from rat kidney membranes, in which the highest enzyme activity toward both lipid acceptors (6,7). In the present report, we have achieved purification of the transacetylase to homogeneity and shown that this enzyme possesses three catalytic activities, namely, PAFacetylhydrolase, PAF:lysophospholipid transacetylase, and PAF:sphingosine transacetylase.
Enzyme Assays-PAF:lysoplasmalogen transacetylase and PAF: sphingosine transacetylase activities were determined according to our previously described methods (3,6). The assay system of PAF:lysoplasmalogen transacetylase consisted of 50 M 1-O-hexadecyl-2-[ 3 H]acetyl-GPC (0.3 Ci), 300 M lysoplasmalogen (suspended in 0.1% bovine serum albumin (BSA)-saline), 100 mM Tris-HCl (pH 7.4), 2 mM sodium acetate, and 10 mM EDTA in a total volume of 250 l. Incubations were carried out at 37°C for 15 min. The lipids were extracted by the method of Bligh and Dyer (10). The lipids were separated by thin layer chromatography (TLC) using a solvent system of CHCl 3 /CH 3 OH/CH 3 COOH/ H 2 O (50: 25:8:4, v/v/v/v), and the radioactivities of the areas corresponding to PAF and alk-1-enylacetyl-GPE were counted using liquid scintillation fluid. The assay system of PAF:sphingosine transacetylase contained 15 M 1-O-hexadecyl-2-[ 3 H]acetyl-GPC (1 Ci), 50 M sphingosine (suspended in equal molar ratio of BSA), 100 mM Tris-HCl (pH 7.4), 2 mM sodium acetate, and 10 mM EDTA in a total volume of 250 l. Incubations were carried out at 37°C for 30 min. The lipids were separated by TLC using a solvent system of CHCl 3 /CH 3 OH (90:10, v/v). The radioactivities of the areas corresponding to PAF and C 2 -ceramide were measured. PAF-AH activity was assayed according to the method previously described (11). The assay system of PAF-AH was composed of 20 M 1-O-hexadecyl-2-[ 3 H]acetyl-GPC (0.1 Ci), 1 mM EDTA, and 100 M potassium phosphate (pH 8.0) in a total volume of 500 l. Incubations were carried out at 37°C for 10 min. The reaction was stopped by sequential additions of 1 ml of CHCl 3 , 1 ml of CH 3 OH, and 0.5 ml of 10% sodium bicarbonate. The upper phase was washed with 1 ml of CHCl 3 three times, and the radioactivities in an aliquot of 0.4 ml were determined.
Preparation of Rat Kidney Membranes-The kidneys were dissected out from male rats weighing 150 -250 g and homogenized with four volumes of 0.25 M sucrose, 20 mM Tris-HCl (pH 7.4), 1 mM DTT, 1 mM EDTA, and 1 g/ml leupeptin. The homogenates were centrifuged at 440 ϫ g for 10 min; the supernatant was collected as postnuclear fraction. The postnuclear fraction was centrifuged at 15,000 ϫ g for 10 min to isolate the mitochondrial pellets. The postmitochondrial fractions were centrifuged at 100,000 ϫ g for 1 h to obtain the microsomal fractions. Both mitochondrial and microsomal fractions were washed with the same buffer solution twice, and the washed mitochondria and microsomes were used as a source for enzyme purification.
Solubilization of the Enzyme-Mitochondrial fractions were suspended in 20 mM Tris-HCl (pH 7.4) containing 0.04% Tween 20, 1 g/ml leupeptin, 1 mM EDTA, and 1 mM DTT at a detergent/protein (w/w) ratio of 0.1. The mixture was gently stirred at 0°C for 1 h and then centrifuged at 100,000 ϫ g for 1 h. The enzyme activity was twice extracted from microsomal fractions by detergent in the same way as described for the mitochondria. High salt wash (0.5 M NaCl) without the presence of detergent (data not shown) could not further elute the enzyme activity from mitochondria or microsomes.
Column Chromatographies-We have observed that the mitochondrial fraction had the highest specific activity and the microsomal fraction contained the highest total activity of transacetylase among the subcellular fractions isolated from HL-60 cells (7). Therefore, we decided to combine both solubilized mitochondria and microsomes as the starting materials for column chromatography. The solubilized enzyme was applied onto a column of DEAE-Sepharose (2.5 ϫ 6.11 cm, 30 ml), which was equilibrated with 20 mM Tris-HCl (pH 7.4), 1 mM DTT, 1 g/ml leupeptin, and 0.02% Tween 20. The column was washed with the same buffer solution and was then eluted with 0.15 M NaCl in 20 mM Tris-HCl (pH 7.4), 1 mM DTT, and 0.02% Tween 20. During initial experiments, the solubilized enzyme bound to the DEAE-Sepharose column was eluted by a linear gradient of NaCl. There was only one major eluted peak of enzyme activity that started to come out at 0.15 M NaCl from the column. Based on this information, stepwise elution of the enzyme activity with 0.15 M NaCl from the DEAE-Sepharose column was performed subsequently in order to speed up the purification process.
Active fractions of DEAE-Sepharose were applied onto a column of hydroxyapatite (1.5 ϫ 8.49 cm, 15 ml), which was equilibrated with 20 mM Tris-HCl (pH 7.4), 1 mM DTT, 1 g/ml leupeptin, and 0.02% Tween 20. The column was washed with the same buffer solution, and the enzyme activity was eluted with 50 mM potassium phosphate (pH 7.0) containing 1 mM DTT, 1 g/ml leupeptin, and 0.02% Tween 20. The rationale to use the stepwise elution from hydroxyapatite with 50 mM potassium phosphate was similar to that of the experiments with DEAE-Sepharose column chromatography. We had previously found that the most of the enzyme activities was eluted from hydroxyapatite column starting at 50 mM potassium phosphate during a linear gradient run of the chromatography.
The active fractions pooled from phenyl-Sepharose were applied onto a column of PBE94 (1.0 ϫ 6.37 cm, 5 ml), which was equilibrated with 20 mM Tris-HCl (pH 7.4), 1 mM DTT, 1 mM EDTA, 1 g/ml leupeptin, and 0.02% Tween 20. The column was washed with 25 mM histidine-HCl (pH 6.2), 1 mM DTT, 1 mM EDTA, 1 g/ml leupeptin, and 0.02% Tween 20, and the enzyme activity was eluted by decreasing pH with Polybuffer 74 (pH 4.0) in 1 mM DTT and 0.02% Tween 20. Five-ml fractions were collected into the tube containing 0.7 ml of 1 M Tris-HCl (pH 7.4) to neutralize the pH. The enzyme activity was unstable at pH 5.0, and 88% of enzyme activity was lost 16 h thereafter. It was necessary to neutralize the pH of the sample solution soon after the enzyme was eluted from the column. The active fractions from chromatofocusing were concentrated into 4.5 ml by using a small size of hydroxyapatite column (1 ϫ 0.27 cm, 1 ml).
Native-PAGE-The enzyme was further purified by native-PAGE according to the method of Ornstein and Davis (12). Separating gel (2 ml) consisting of 7.5% polyacrylamide, 0.375 M Tris-HCl (pH 8.8), and 0.02% Tween 20 was prepared in a glass tube (15 ϫ 0.5 cm). Stacking gel (0.1 ml) composed of 5% polyacrylamide, 0.125 M Tris-H 3 PO 4 (pH 6.8), and 0.02% Tween 20 were overlaid to the separating gel. The enzyme solution (0.4 ml) was added to each tube. Electrophoresis was carried out at a constant voltage of 80 V and 4°C until the dye migrated toward the end of the gel. The gels were then removed from the glass tubes and horizontally sliced into fragments (2 mm distance). The individual pieces were transferred into microtiter plate wells, soaked in 50 l of 25 mM Tris-HCl (pH 7.4), 1 mM DTT, 0.02% Tween 20 for 16 h, and the supernatants with the highest enzyme activities were collected.
Preparation of Affinity Gel Matrix-1-O-Hexadecyl-2-N-methylcarbamyl-GPC (5 mg, 9.28 mol) dissolved in 0.5 ml of 0.16 M acetate (pH 5.6) containing 80 mM CaCl 2 was combined with 1 ml of cabbage phospholipase D and 0.5 ml of 20% ethanolamine. The reaction was carried out at room temperature for 16 h (similar to that described in Ref. 13), and the lipid was extracted by the method of Bligh and Dyer (10). The reaction product, 1-O-hexadecyl-2-N-methylcarbamyl-GPE, was purified by TLC using a solvent system of CHCl 3 /CH 3 OH/ CH 3 COOH/H 2 O (50:25:8:4, v/v/v/v). The purified phospholipid (8.36 mol) was dissolved in 2 ml of 50 mM borate (pH 8.0) in CH 3 OH and reacted with activated CH-Sepharose (2 ml) suspended in the same solution for 16 h at 22°C. The resulting gel was washed with 50 mM borate (pH 8.0) in CH 3 OH thoroughly, and the remaining reactive sites were blocked with 1 M Tris-HCl (pH 8.0) for 16 h at room temperature. Based on results from phosphorus determinations (14), 0.98 mol of ligand was bound to 1 ml of CH-Sepharose Purification of the Enzyme by Affinity Gel Matrix-The active fractions of native-PAGE were combined with 50 l of affinity gel matrix in a microcentrifuge tube, which was equilibrated with 20 mM Tris-HCl (pH 7.4) containing 0.1 M NaCl, 1 mM DTT, and 0.02% Tween 20 and mixed gently at 4°C for 30 min. The gel was washed with the same buffer and followed by the same solution without NaCl. The washed gel was combined with 5 mM alkylacetyl-GPC dissolved in 20 mM Tris-HCl (pH 7.4) in 1 mM DTT, and 0.02% Tween 20 and incubated for 1 h at 4°C. The mixture was centrifuged for 5 min at 10,000 ϫ g, and the supernatant with enzyme activity was transferred into another tube. PAF in this enzyme preparation was removed by hydroxyapatite (0.5 ml) column chromatography. The resulting enzyme solution was dialyzed against 20 mM Tris-HCl (pH 7.4) containing 40% glycerol, 1 mM DTT, and 0.02% Tween 20. The purified preparation was stored at Ϫ20°C, and no significant decrease in enzyme activity was observed at least for 1 month.
SDS-PAGE-SDS-PAGE was carried out according to the method of Laemmli (15) using 10% polyacrylamide gel. The proteins were visualized by silver staining using a silver staining kit for protein (Pharmacia Fine Chemicals).
Sequencing-The sequencing of the protein was carried out at Harvard Microchemistry Facility (Boston, MA) by tandem mass spectrometry and Edman degradation analysis. The purified protein was subjected to electrophoresis on 10% SDS-PAGE gel. The protein band on the gel was visualized by Coomassie Brilliant Blue. The sliced gel was subjected to digestion with trypsin and microsequencing.
Protein Determination-In most instances, protein was measured by a protein assay kit (Bio-Rad) using BSA as the standard. However, when the protein concentrations in the samples were low (e.g. Table III, steps 3-7), a series of one to one (1:1) dilutions of 1 g/ml BSA (ranging from 31.25 ng to 1 g per spot) and the samples in duplicates were spotted on a 3MM filter paper. The spots were allowed to air-dry and were stained with 0.25% Coomassie Brilliant Blue R-250 (Bio-Rad), 45% CH 3 OH, 10% CH 3 COOH for 30 min. The paper was destained with 45% CH 3 OH, 10% CH 3 COOH until the background was nearly white. Then, the protein amount was determined by comparing the intensities of the spots from samples with those from BSA standards. The amount of protein in the final enzyme preparation (Table III, step 8) was estimated by comparing the intensity of the protein band with those of standard proteins after SDS-PAGE and silver staining.

RESULTS
Purification of PAF-dependent Transacetylase-An essential step in purifying the membrane enzyme is to first solubilize the enzyme from the membrane. Although there are assortments of detergents (ionic and non-ionic) and chaotropic agents available commercially, the detergent selected will be the one that does not inhibit or inactivate enzyme activity, does not present the enzyme in an aggregated form, and has a high critical micellar concentration. Additionally, the ratio of protein to detergent concentration is important. In preliminary experiment, among the several ionic and non-ionic detergents we tested (such as CHAPSO, octylglucoside, and Triton X-100), PAF:lysoplasmalogen transacetylase activity was most effectively solubilized by Tween 20. Enzyme activity was solubilized from rat kidney membrane fractions (mitochondria plus microsomes, 100,000 ϫ g pellets) by Tween 20 at a concentration of 0.02-0.05%, and the maximal specific activity was observed at a detergent/protein (w/w) ratio of 0.1 (Table I).
When the efficiency of the solubilization was tested on mitochondrial and microsomal fractions separately instead of using total membrane fraction, it was found that substantial amounts of enzyme activity still remained in the 100,000 ϫ g pellets, especially with that of the microsomal fractions (Table  II). Therefore, a second attempt on the solubilization of the remaining pellets was carried out. Around 8% and 10% of transacetylase activities were obtained from the second solubilization of the mitochondrial and microsomal pellets, respectively. However, the second solubilization from the mitochondrial fraction decreased the specific activity of the enzyme because relatively large amounts of unspecific proteins were concomitantly released. Thus, solubilization procedures were performed twice for the microsomal fraction and once for the mitochondrial fraction hereon.
Overall, Tween 20 (0.02%) was included in all purification steps because it maintained the stability of enzyme activity. The apparent isoelectric point of the enzyme was estimated to be 5.0 by chromatofocusing (data not shown). Native-PAGE was the crucial step for the purification of this enzyme. The specific activity was increased 17-fold, as compared with that in the previous step (Table III). It should be noted that the enzyme activity was easily extracted from the polyacrylamide gel by soaking the gel slice in the buffer solution without squeezing the gel or the aid of electricity, as commonly performed as a method to extract the proteins from the polyacrylamide gel.
1-O-Hexadecyl-2-N-methylcarbamyl-sn-glycero-3-phosphocholine (data not shown) and 1-O-hexadecyl-2-N-methylcarbamyl-GPE, which are structurally related to PAF, competitively inhibit PAF:lysoplasmalogen transacetylase activity in Tween 20 (0.05%) solubilized membrane fraction from rat kidneys with K m ϭ 9.1 M, and K i ϭ 10.1 M for hexadecylacetyl-GPC and hexadecyl-N-methylcarbamyl-GPE, respectively (Fig.  1). Thus, hexadecyl-N-methylcarbamyl-GPE is expected to be useful as a specific ligand for the affinity gel matrix. This lipid has the advantage of being resistant to enzymatic hydrolysis, while acetyl ester of PAF is easily degraded by PAF-acetylhydrolase(s) in the cells. Additionally, the amino group of hexadecylmethylcarbamyl-GPE can be linked to the carboxyl-terminal group of CH-Sepharose covalently. All the enzyme activity was adsorbed to the resulting gel matrix, while enzyme activity was not adsorbed to the CH-Sepharose without ligand at all (data not shown). The enzyme activity was not eluted from the gel matrix by 0.5% CHAPS, 1% Triton X-100, 50% ethylene glycol, or 0.5 M NaCl. In order to elute the enzyme activity from the column with a reasonable recovery, more than 5 mM PAF TABLE I Effects of Tween 20 concentrations and Tween 20: protein ratios on the solubilization of PAF:lysoplasmalogen transacetylase from the membrane fraction of rat kidneys Rat kidneys were homogenized in buffer and the membrane fraction was isolated as described (13). The membrane fraction suspended in 0.25 M sucrose and 10 mM Hepes, (pH 7.0) was mixed with 0.2% Tween 20 to attain the desired Tween 20 concentration and Tween 20/protein ratio. The mixture was stirred at 4°C for 30 min and then centrifuged at 100,000 ϫ g for 60 min to obtain the supernatant and pellet fraction. Transacetylase activities and protein concentrations in both fractions were measured to ensure proper recovery. Data represent the average of at least duplicate determinations. Tween  was required. The concentration of the ligand in the gel matrix was estimated to be 0.98 mM; therefore, a relatively high concentration of PAF was needed for replacement. These results also indicate that the enzyme specifically recognizes the ligand at the site, which is necessary for the interaction with the substrate. Starting from rat kidney membrane, 13,700-fold purification of the enzyme was achieved with a yield of about 4%. A typical result of the purification procedures is summarized in Table III.
Purity and Molecular Weight of the Enzyme-Different fractions obtained during the enzyme purification were examined by SDS-PAGE (Fig. 2). The enzyme obtained after the final step of purification showed a single homogeneous band. By comparing the mobility of the enzyme band on SDS-PAGE with molecular size markers, a molecular mass of 40 kDa was deduced for the enzyme (Fig. 3). In addition, the intensity of the band of 40 kDa on SDS-PAGE was correlated with enzyme activity in the individual purification steps including chromatofocusing, native-PAGE, and affinity gel matrix (data not shown), suggesting that the band of 40 kDa is PAF-dependent transacetylase. Although we have attempted to estimate the molecular weight of natural enzyme on gel filtration column of Sephacryl S-200, the enzyme activity was eluted near the void volume due to the formation of mixed micelles with Tween 20 (data not shown).
Partial Amino Acid Sequences-The purified protein was digested with trypsin, and five of the peptide fragments were sequenced as described under "Experimental Procedures." Amino acid sequences of five peptide fragments were: GTLDPYEGQEVMVR, AML*AF*L*QK, LFSSGTR, IKEG-EKEFHVR, and L*PVSWNGPF*K* (* indicate that isobaric amino acid residues cannot be unambiguously differentiated in mass spectrometric sequence, but such residues are displayed by alignment with a known or homology to a known, and with reference to enzyme specificity). A search using a protein sequence data bank indicated that these sequences had homology with the sequences present in bovine PAF-acetylhydrolase II (GenBank TM /EBI Data Bank accession number D87559; 78.6%, 100%, 71.4%, 81.8%, and 100%, respectively).
Substrate Specificity-A different type of transacetylase that transfers the acetate of PAF to sphingosine forming C 2 -ceramide was described in HL 60 cells by us (6). The ratios of specific activity toward lysoplasmalogens/sphingosine were 49.7 and 80.8 in crude mitochondrial and microsomal membranes, respectively. In the final purified preparation, it also contains PAF:sphingosine transacetylase activity. Nevertheless, the sphingosine transacetylase activity was more labile during purification and storage than that of lysoplasmalogen transacetylase. Thus, we could not accurately assess the ratio of lysoplasmalogen/sphingosine transacetylase activities. Notwithstanding, these data indicate that a single enzyme can catalyze both kinds of transacetylase activities. The higher ratio toward lysoplasmalogens as the substrate may partly be explained by the fact that, in order to assay PAF:sphingosine transacetylase activity, we had to include equal molar ratio of BSA in the incubations, and BSA severely inhibited sphingosine transacetylase reaction (6). However, incorporation of BSA as the assay medium for sphingosine transacetylase was nec-  essary in order to maintain zero order kinetics and reduce nonenzymatic generation of the product (6). Lysoplasmalogen transacetylase did not show the same BSA requirement as that of sphingosine transacetylase and was measured in the presence of 0.1% BSA-saline (3).
The finding that the enzyme shared the sequence homology with bovine PAF-acetylhydrolase II led us to ask the question of whether the enzyme carried the activity of hydrolyzing PAF. We found that the PAF-dependent transacetylase also contains PAF hydrolyzing activities in crude membrane fractions and purified preparations; the ratios of lysoplasmalogen transacetylase/acetylhydrolase were 0.74, 0.84, and 1.04 (average of two experiments with Ͻ10% variation) in the mitochondria, microsomes, and purified enzyme, respectively.
The reason for the slight increase in ratios of lysoplasmalogen transacetylase/acetylhydrolase activities during purifica-tion is not clear at present. It is possible that some of the contaminating acetylhydrolase activities were removed during purification procedures or the presentation of substrates might differ between the purified enzyme and crude membranebound enzymes due to the possible interaction of membrane phospholipids with the lipid substrates.
Kinetic Properties-The dependence of enzyme activity on substrate concentration is shown in Fig. 4. PAF:lysoplasmalogen transacetylase activity displayed typical Michaelis-Menten kinetics, with an apparent K m of 227 M for lysoplasmalogen and a V max of 81 mol/min⅐mg (Fig. 4A). Interestingly, the K m for lysoplasmalogens was 106.4 M in the membranes of HL-60 cells (3) and 80 M in the homogenates of calf pulmonary artery endothelial cells (data not shown). These results suggest that the K m values are varied with different tissues and cell types. Nevertheless, these values are well below the physiological concentration of lysoplasmalogens in the cells. Calculation from the data obtaining by Tessner et al. (16) and Nieto et al. (17) indicated that the lysoplasmalogen concentration could reach 0.5-0.7 mM when human neutrophils were stimulated with ionophore A23187 for 5 min.
In contrast, inhibition of sphingosine transacetylase activity (Fig. 4B) was resulted when sphingosine concentrations were higher than 50 M. If the points during the ascending portion of the curve were used, the calculated K m for sphingosine was 4.1 M. This value is similar to the K m reported for sphingosine of the purified rat kidney sphingosine kinase (5 M) (18).
We have shown that, using mixed substrate experiments and crude membrane fractions from HL-60 cells (6), either CoAindependent transacetylase has a higher substrate affinity for sphingosine than any of the other substrate analogs, or the possibility of two isoforms of the transacetylase might be involved in the transfer of acetate from PAF to sphingosine and lysophospholipids. When mixed substrate experiments were performed in the present studies using purified enzyme, similar results to those using crude membranes from HL-60 cells were obtained (6). Sphingosine at 15 and 50 M inhibited the lysoplasmalogen transacetylase activity by 32% and 80%, respectively. On the other hand, lysoplasmalogens at 15 and 50 M had no effect on the sphingosine transacetylase activity. Thus, we further confirmed that transacetylase has a higher substrate affinity for sphingosine than lysophopholipids.
Effect of Inhibitors-PAF:lysoplasmalogen transacetylase activity was inhibited by diisopropyl fluorophosphate in a dose- The log molecular weight of standard proteins versus mobility on the SDS-PAGE was plotted. Molecular weight protein markers were the same as described in Fig. 2. The molecular mass of the purified transacetylase from rat kidney membranes is shown by the arrow. dependent manner (Table IV and data not shown). These data suggest that serine residue is involved in the active site of the enzyme. The specific inhibitor of plasma PAF-AH and serine esterase (19), p-aminoethyl benzenesulfonyl fluoride (Pefabloc), completely inhibited PAF:lysoplasmalogen transacetylase at a concentration of 0.1 mM. These results are in good agreement with the fact that the transacetylase partially shared structural homology with PAF-AH II, and human PAF-AH II exhibited 43% predicted amino acid identity to human plasma PAF-AH (20). Both 5,5Ј-dithiobis(2-nitrobenzoic acid) (DTNB) and N-ethylmaleimide (NEM), which react with the sulfhydryl group of cysteine, partially block the enzyme activity, indicating that sulfhydryl group is important for enzyme activity. Diethyl pyrocarbonate, a modifier of the histidine, at 1 mM partially inhibited the PAF:lysoplasmalogen transacetylase activity. In parallel, the PAF:sphingosine transacetylase activities were also affected by these inhibitors except with different sensitivities especially with the sulfhydryl reagents (Table IV). DISCUSSION PAF-dependent transacetylase was purified from rat kidney membrane 13,700-fold. The final preparation, which was judged to be nearly homogeneous by SDS-APGE, yielded a single protein band of 40 kDa. DTT was included throughout the purification because enzyme activity was rapidly decreased in the absence of DTT. DTT is thought to be necessary for protecting the sulfhydryl group of cysteine from oxidation because DTNB and NEM partially abolish the enzyme activity (Table IV). NEM and DTNB inhibit PAF:lysoplasmalogen and PAF:sphingosine transacetylases with different inhibitory potencies. For example, PAF:sphingosine transacetylase activity (22.6% of enzyme activity remained after 30 min of 0.1 mM NEM treatment) was more susceptible to NEM than PAF: lysoplasmalogen transacetylase (which retained 90.7% of enzyme activity after 0.1 mM NEM treatment for 30 min). These results are consistent with our previous work using HL-60 cell membranes as the enzyme source (6). This differential effect of NEM and DTNB on enzyme activities suggests that the sulfhydryl group of cysteine play a more important role in the interaction of the enzyme with sphingosine, as compared with that of lysoplasmalogen.
In the present investigation, PAF-dependent transacetylase was isolated from mitochondrial and microsomal membrane fractions. However, PAF-dependent transacetylase activity is also present in cytosolic fraction (7). It is well known that some of the enzymes are distributed in both membrane and cytosols, and translocation of the enzymes is regulated by modification of the protein, such as phosphorylation (21) and myristoylation (22). For example, phosphorylation of cytosolic phospholipase A 2 causes translocation from the cytosol to the membrane (21). We have demonstrated that PAF:acyllyso-GPC transacetylase is regulated through phosphorylation/dephosphorylation (4). It is not known presently whether phosphorylation/dephosphorylation is involved in determining the cellular localization of PAF-dependent transacetylase or PAF-AH II is regulated through phosphorylation/dephosphorylation. The availability of a homogeneous preparation of transacetylase protein will facilitate the development of specific antibodies against this enzyme. Specific antibodies against PAF-dependent transacetylase will be useful in clarifying the relationships of the enzyme activities located in membrane and cytosol as well as the mechanism of translocation of this enzyme between these two fractions.
Analysis of the partial amino acid sequences of five trypticdigested peptide fragments revealed that this enzyme shared sequence homology with the previously reported bovine cytosolic PAF-AH II (20). Cytosolic PAF-AH II was found to be Nmyristoylated and could be translocated from cytosol to mem-  branes in oxidative stress-induced cells (23). In addition, purified PAF-dependent transacetylase from rat kidney membranes contained PAF-AH activity in the absence of a lipid acceptor. The activities of both PAF-dependent transacetylase and PAF-AH II (24) were inhibited by serine esterase inhibitors and sulfhydryl reagents. Preliminary results (unpublished data in collaboration with Dr. K. Inoue's laboratory) also indicated that purified transacetylase and transacetylases present in the mitochondria and microsomes cross reacts with antihuman PAF-AH II monoclonal antibody. These results suggest that transacetylase and PAF-AH II share similar structural requirement for the active site and the same recognition site for the immunological epitope. However, we have to wait for the result of cDNA cloning of the transacetylase in order to compare the sequence of the transacetylase with that of the PAF AH II. Information obtained from the amino acid sequences of the analyzed peptides of the PAF-dependent transacetylase should provide the necessary means to identify the specific cDNA clones from the rat kidney cDNA library. It is possible that transacetylase and PAF-AH II share the same amino acid sequences, but are differentially posttranslationally modified.
Regardless of structural similarity or difference between transacetylase and PAF-AH II, our finding that transacetylase has three distinct catalytic activities raise important questions concerning the regulation, biological implications, and consequences of this unique enzyme. It is possible that PAF may exert some of its biological effects through transacetylase without the need or presence of participation of intracellular PAF receptor (s).
In addition to the PAF-dependent transacetylase, there are other examples that an enzyme with transesterification activity also possesses hydrolytic activity. Lecithin-cholesterol acyltransferase (LCAT), which normally transfers the acyl group of PC to cholesterol, also hydrolyzes PAF (25). PAF-AH activity of LCAT plays the role of detoxification of oxidized PC, especially when the PAF-AH is absent or inactivated (25). Thus, PAF-dependent transacetylase is similar to LCAT and likely to be an enzyme that catalyzes diverse reactions and thus leads to different biological functions.
In conclusion, we have demonstrated that a single enzyme catalyzes three kinds of reactions, namely PAF:lysoplasmalogen (lysophospholipids) transacetylase, PAF:sphingosine transacetylase, and PAF-acetylhydrolase. Depending on the differences of acceptor lipid molecules, different lipid products such as PAF analogs of plasmalogens (also acyl analogs of PAF and etc.), C 2 -ceramide, and lyso-PAF will be generated by the transacetylase through these three catalytic activities. These different lipid products possess different cellular functions when compared with that of the original substrates (i.e. PAF and sphingosine). For instance, sphingosine can induce a potent and reversible inhibition of protein kinase C in vitro and in cell cultures (26 -29). Recently, Rodriguez-Lafrasse (30) provided the first evidence to show that protein kinase C inhibition is directly related to the sphingosine accumulation in vivo. On the other hand, C 2 -ceramide has been extensively used by many investigators as an unnatural, cell permeable analog of long-chain acyl-ceramides to mimic the effects of various inducers of sphingomyelin-ceramide signal pathways (8,9). Multiple experimental approaches suggest that long-chain acyl ceramides and C 2 -ceramide play an important role in regulating cell cycle arrest, apoptosis, differentiation, and cell senescence (8). However, our previous data (6) demonstrated that PAF-dependent transacetylase is the enzyme responsible for the biosynthesis of C 2 -ceramide, and the cellular concentration of C 2 -ceramide is in the range (micromolar) that could exert significant biological effects. In addition, the signaling pathways utilized by the sphingomyelinase differ from those of cell-permeable ceramide anologs (31). Additionally, no detectable biological effects are observed in that neutral sphingomyelinase overexpressed cells (32). These results suggest that C 2 -ceramide should be classified as a naturally occurring novel lipid mediator. Furthermore, acyl analogs of PAF have biological characteristics distinct from that of PAF (4). Future research in this laboratory is directed toward the elucidation of the mechanism(s) of how a single enzyme controls and regulates three different kinds of enzymatic reactions. Additionally, progress is being made to generate cDNA and polyclonal antibodies against the transacetylase.