The Role of Platelet-activating Factor-dependent Transacetylase in the Biosynthesis of 1-Acyl-2-acetyl-sn-glycero-3-phosphocholine by Stimulated Endothelial Cells*

Acyl analogs of platelet-activating factor (PAF) (1-acyl-2-acetyl-sn-glycero-3-phosphocholine, acylacetyl -GPC) are the predominant products synthesized during thrombin or ionophore A23187-mediated activation of endothelial cells. However, the biosynthetic pathway responsible for the production of acylacetyl-GPC is not well understood. In the present investigation, we have demonstrated that the acyl analogs of PAF are also the major products from calf pulmonary artery endothelial cells in response to a time-dependent stimulation of ATP (10−3 m), bradykinin (10−8 m), or ionophore A23187 (2 μm). In addition, we have found that the CoA-independent PAF:acyllyso-GPC transacetylase recently identified by us is concurrently and transiently induced with maximal 4-fold enhancement at 5 min and returned to near basal level by 10 min treatment of endothelial cells with ATP. Acid phosphatase reduces the increased PAF:acyllyso-GPC transacetylase activity from the homogenates of ATP-activated endothelial cells. Reduced PAF:acyllyso-GPC transacetylase activity can be restored by incubating the acid phosphatase-treated homogenates with ATP (5 mm) and Mg2+ (10 mm). Furthermore, okadaic acid, a protein phosphatase 1 and 2A inhibitor, incubated with endothelial cells in a dose-dependent manner (1–100 nm) for 10-min potentiates and sustained the stimulation of PAF:acyllyso-GPC transacetylase activity by ATP. On the other hand, genistein, tyrphostin-25 (inhibitors of tyrosine-specific protein kinase), and calphostin C (an inhibitor of protein kinase C) block the activation of PAF:acyllyso-GPC transacetylase by ATP. These results are consistent with the notion that ATP regulates the transacetylase activity by reversible activation and inactivation via the phosphorylation and dephosphorylation cycle. ATP also augments the activities of alkyllyso-GPC/acyllyso-GPC:acetyl-CoA acetyltransferase. However, the activation of the acetyltransferases precedes that of the transacetylase with peak activation occurring at 1–2 min of the ATP treatment. In addition, sodium vanadate, also an inhibitor of protein phosphatase, stimulates the increase in the incorporation of [3H]acetate into acyl[3H]acetyl-GPC of the ATP-treated endothelial cells. Collectively, our data show that both acetyltransferases and transacetylase participate in and contribute to the biosynthesis of acyl analogs of PAF in a coordinate fashion in endothelial cells.

Platelet-activating factor (PAF) 1 is a potent lipid mediator with diverse pathophysiological responses such as inflammation, allergic reactions, and reproduction (see reviews Refs. [1][2][3]. Different cell types from various species activated by specific stimuli are able to produce PAF (4). The chemical structure of PAF is originally elucidated as 1-alkyl-2-acetyl-snglycero-3-phosphocholine (alkylacetyl-GPC, 5-7). However, it recently becomes apparent that acylacetyl-GPC can be the predominant product in certain cell types. For instance, the major compound produced by human umbilical vein endothelial cells activated with thrombin and ionophore A23187 (8 -11) or bovine pulmonary artery endothelial cells stimulated by ionophore A23187 (12) is acylacetyl-GPC. Also, depending on the agonist used, either alkylacetyl-GPC or acylacetyl-GPC can be synthesized within the same cell type. For example, the human basophils generate mainly acylacetyl-GPC in response to anti-IgE and mostly alkylacetyl-GPC in response to ionophore A23187 (13).
Although the biological activities of acylacetyl-GPC are Ͼ500-fold less potent than alkylacetyl-GPC in lowering the blood pressure of spontaneous hypertensive rats, in releasing serotonin from rabbit platelets (14), in increasing the intracellular Ca 2ϩ concentrations, and in activating microtubule-associated protein-2 kinase activity (15), however, acylacetyl-GPC can act as a specific noncompetitive inhibitor of alkylacetyl-GPC-induced activation of the human neutrophils and leukotriene C 4 release from the human leukocytes (10,16). Furthermore, 1-acyl analogs of PAF at nM concentrations can prime polymorphonuclear leukocytes for enhanced O 2 Ϫ production after stimulation with fMet-Leu-Phe or human recombinant C5a (17). Additionally, acylacetyl-GPC has the ability to decrease the susceptibility of the low density lipoprotein particles to oxidative modification mediated by copper ions, monocytes, or endothelial cells, whereas PAF has no effect (18).
The pathways leading to the accumulation of acylacetyl-GPC are not completely understood. It is suggested that the remodeling pathway (3) is responsible for the biosynthesis of alkylacetyl-GPC/acylacetyl-GPC during short term stimulation by agonists in endothelial cells (19). The first step of the remodeling pathway involved the conversion of alkylarachidonoyl-GPC/acylarachidonoyl-GPC to alkyllyso-GPC/acyllyso-GPC by a putative phospholipase A 2 or/and a CoA-independent transacy-lase (20 -22). The alkyllyso-GPC/acyllyso-GPC formed is subsequently acetylated by the alkyllyso-GPC/acyllyso-GPC:acetyl-CoA acetyltransferase (19,23,24). Increases in intracellular calcium concentration and protein kinase C activate the phospholipase A 2 that is involved in the synthesis of PAF (25). The alkyllyso-GPC:acetyl-CoA acetyltransferase is regulated by phosphorylation and dephosphorylation (26 -29).
Alkylacetyl-GPC and acylacetyl-GPC have different catabolic routes. In most cell types, alkylacetyl-GPC is hydrolyzed by acetylhydrolase to form alkyllyso-GPC and then the alkyllyso-GPC is reacylated to produce alkylacyl-GPC (see reviews herein). On the other hand, the major metabolic pathway for the acylacetyl-GPC is the removal of acyl group at the sn-1 position by a yet to be characterized 1-acylhydrolase (30). The acetyl-GPC formed from this reaction is further degraded to glycerophosphocholine, choline, or phosphocholine (30). However, it is reported that lysophospholipase can also catalyze the stoichiometric conversion of acylacetyl-GPC to acetyl-GPC (31,32). In addition, the activity of lysophospholipase may control the level of acyl analogs of PAF, because peritoneal macrophages that preferentially synthesize acyl analogs of PAF contain only one-sixth of the lysophospholipase activity in comparison to that of alveolar macrophages which primarily make PAF (33). Besides, the amounts of acylacetyl-GPC produced within a specific cell type can be increased by blocking the phenylmethylsulfonyl fluoride-sensitive phospholipase A 1 or lysophospholipase activity (31,34,35).
We have previously identified a novel CoA-independent and PAF-dependent transacetylase that transfers the acetate group from PAF to a variety of lysophospholipids; acyllyso-GPC is the most active acceptor that converts to acylacetyl-GPC by the transacetylase (36). In the present studies, we investigate the possibility that PAF-dependent transacetylase may be involved in the biosynthesis of acyl analogs of PAF. Our results demonstrate that the PAF:acyllyso-GPC transacetylase is induced severalfold during agonist-coupled activation of the endothelial cells and is most likely to be regulated via covalent modification of the enzyme through phosphorylation and dephosphorylation. Based on the temporal relationships that exist between the synthesis of acylacetyl-GPC and induction of the transacetylase activity in stimulus-activated cells, our findings suggest that the PAF-dependent transacetylase participates in the production and controls the level of acylacetyl-GPC in endothelial cells. , and with or without agonist, thrombin (2 units/ml), bradykinin (10 Ϫ8 M), ATP (10 Ϫ3 M), or ionophore A23187 (2 M) at 37°C for various times as indicated in the figures. At the end of incubations, the media were removed, and the cells were rinsed twice with 5 ml of HBSS, 10 mM Hepes (pH 7.4) before scraping into 3 ml of methanol. The cellular lipids were extracted by the method of Bligh and Dyer (37) except that 2% of acetic acid was included in the methanol. The thin layer chromatographic solvent system of chloroform/methanol/acetic acid/H 2 O (50:25: 8:6, v/v/v/v) was used to isolate the radyl[ 3 H]acetyl-GPC fraction. Radioactivity in lipid fractions separated by thin layer chromatography (TLC) was determined by area or zonal scraping of the silica gel into vials for liquid scintillation counting.
Enzyme Assays-PAF:acyllyso-GPC transacetylase and PAF:lysoplasmalogen transacetylase were determined as reported earlier by us (36) with minor modification. Briefly, standard incubations consisted of 50 M PAF (0.5 Ci in 50 l of 0.1% bovine serum albumin/saline), 300 M palmitoyllyso-GPC or lysoplasmalogen (in 50 l of 0.1% bovine serum albumin/saline), 5 mM EDTA, 1 mM sodium acetate, 100 mM Tris-HCl (pH 7.4), and 100 g of homogenate protein in a final volume of 0.50 ml. The incubations were carried out at 37°C for 15 and 30 min to ensure that the enzyme activity is measured under zero order kinetic. The cell homogenates were prepared by scraping agonist-treated or nontreated cells from the surface of the flasks into 2 ml of homogenizing buffer (0.25 M sucrose, 100 mM Tris-HCl (pH 7.3), 1 g/ml leupeptin, 1 mM dithiothreitol, 50 mM NaF). The cell suspensions were homogenized by using an ultrasonic cell disruptor (Microson) at 30% power output for 15 s ϫ 5 times. The extracted lipid product (Ͼ95% as radyl[ 3 H]acetyl-GPC) was treated by phospholipase C. The resulting radyl[ 3 H]acetylglycerols were converted to their respective benzoate derivatives and analyzed by a combination of TLC and liquid scintillation counting as described in the above section.
Alkyllyso-GPC/acyllyso-GPC:acetyl-CoA acetyltransferase and PAF acetylhydrolase activities were measured according to the methods we previously described (39,40) except that the cell homogenates were used as the enzyme sources in most of the experiments. The protein content of the enzyme preparations was determined by the method of Lowry et al. (41).

Incorporation of [ 3 H]Acetate into Radyl[ 3 H]acetyl-GPC and
Its Subclasses in Agonist-activated Endothelial Cells-To study the possible involvement of PAF-dependent transacetylase in the biosynthesis of acylacetyl-GPC of endothelial cells, it was necessary to establish in our laboratory that the stimulustreated endothelial cells produced acyl analogs of PAF as the predominant product. We chose bovine pulmonary artery endothelial cells as a model system for the convenience. Data in Fig. 1A showed that the incorporation of [ 3 H]acetate into radyl[ 3 H]acetyl-GPC was increased in a time-dependent fashion with noticeable rise at 1 min and approaching plateau at 10 min, when the endothelial cells were incubated with ATP (10 Ϫ3 M). Treatment of endothelial cells with bradykinin (10 Ϫ8 M) or ionophore A23187 (2 M) yielded similar time course responses (data not shown). However, the increase in radylacetyl-GPC synthesis was minimal in thrombin-treated (2 units/ml) endothelial cells (data not shown). These results were consistent with the findings demonstrated by Garcia et al. (12). They showed that 1 M bradykinin or 10 M ionophore A23187 enhanced the [ 3 H]acetate incorporation into radylacetyl-GPC in bovine pulmonary artery endothelial cells. Besides, our data were similar to that reported by Whatley et al. (42) in that cultured endothelial cells from various bovine blood vessels responded to ATP, bradykinin, or ionophore A23187 but not to thrombin stimulation in generating an increase in PAF production.
Garcia et al. (12) described that when calf pulmonary artery endothelial cells were treated with ionophore A23187 (10 M), the primary product formed under the conditions was acyl analogs of PAF rather than PAF. We found that acyl analogs of PAF were also the major product when calf pulmonary artery endothelial cells were stimulated with ATP (approximately 85% as acylacetyl-GPC) (Fig. 1B), bradykinin (about 67% as acylacetyl-GPC), or ionophore A23187 (nearly over 80% as acylacetyl-GPC) (data not shown). The kinetics of the time-dependent incorporation of [ 3 H]acetate into the subclasses of radylacetyl-GPC were similar among the three agonists tested. On the other hand, the kinetic of the time-dependent incorporation of [ 3 H]acetate into the subclasses of radylacetyl-GPC treated with thrombin was different from that treated with ATP, bradykinin, or ionophore A23187. It is possible that the kinetics of the time-dependent incorporation of [ 3 H]acetate in thrombintreated endothelial cells reflected the turnover of the endogenous subclasses of radylacetyl-GPC, since bovine pulmonary artery endothelial cells did not respond to thrombin activation.
Induction of PAF:Acyllyso-GPC Transacetylase by ATP and Its Mechanism-Since ATP caused the highest stimulation on the incorporation of [ 3 H]acetate into radyl[ 3 H]acetyl-GPC among the physiological agonists we tested, ATP was chosen as the agonist in the subsequent experiments. Fig. 2 depicted that ATP induced a rapid and transient increase in PAF:acyllyso-GPC transacetylase activity with maximal activation at 5 min. The activity of PAF:acyllyso-GPC transacetylase returned to near basal levels after 10 min of ATP treatment. The increase in transacetylase activity at 5 min incubation of ATP with endothelial cells was about 4-fold over that in the untreated cells.
It was unlikely that the increase in PAF:acyllyso-GPC transacetylase activity by ATP was due to de novo protein synthesis, because the induction of the transacetylase activity by ATP was a rapid process (within 2 min). Therefore, we investigated the possibility that the activity of PAF:acyllyso-GPC transacetylase may be regulated by ATP through covalent modifications, specifically through phosphorylation and dephosphorylation. When the homogenates of ATP-stimulated endothelial cells were treated with an acid phosphatase from potato for 15 min, a decrease of 61% in transacetylase activity was observed (Fig. 3, 2nd column versus 1st column). The percent of decrease in transacetylase activity depended on the concentrations of acid phosphatase used and the lengths of the preincubation times of the homogenates with acid phosphatase (data not shown). Furthermore, preincubation of the homogenates with boiled acid phosphatase (at 100°C for 10 min) had no effect on the transacetylase activity (Fig. 3, 3rd column versus 1st column). On the other hand, addition of ATP (5 mM) plus Mg 2ϩ (10 mM) to the homogenates of ATP-treated endothelial cells (Fig. 3, 4th column versus 1st column) or to the homogenates of ATP-stimulated endothelial cells that were pretreated with acid phosphatase (Fig. 3, 5th column versus 2nd column) could further potentiate the transacetylase activity. These data from in vitro experiments suggested that the transacetylase activity is activated/inactivated through phosphorylation/dephosphorylation.
To further substantiate the fact that the transacetylase activity is regulated through phosphorylation and dephosphorylation, several following experiments using whole cell systems were performed.
Endothelial cells were preincubated with various concentrations of okadaic acid (1 to 100 nM) for 30 min and then incubated with ATP (10 Ϫ3 M) for an additional 10 min. The activity of transacetylase was closed to near basal levels when endothelial cells were treated with ATP for 10 min (see Fig. 2). Okadaic acid, a potent inhibitor of protein phosphatase 1 and 2A (43), induced a 2-fold increase in PAF:acyllyso-GPC transacetylase activity over that of the 10-min ATP-treated control (Fig. 4). Optimal concentrations of okadaic acid that potentiated the ATP-augmented transacetylase activity occurred between 10 and 100 nM (Fig. 4). Therefore, these results were consistent with the notion that by blocking the dephosphorylation of the proteins with okadaic acid, the transacetylase activity is amplified. A different protein phosphatase inhibitor, sodium orthovanadate (1 mM), likewise sustained and prolonged the activation of transacetylase in calf pulmonary artery endothelial cells treated with ATP for 10 min (data not shown).

FIG. 1. Time-dependent incorporation of [ 3 H]acetate into radyl[ 3 H]acetyl-GPC (A) and subclasses of radyl[ 3 H]acetyl-GPC of ATP-stimulated endothelial cells (B). Monolayers of calf pulmonary artery endothelial cells were incubated with ATP and 25 Ci of
When endothelial cells were preincubated with a selective inhibitor of tyrosine-specific protein kinase, tyrphostin-25 (44,45) for 10 min, and incubated with ATP for additional 5 min, tyrphostin-25 blocked the induction of PAF:acyllyso-GPC transacetylase activity by ATP in a dose-dependent manner with close to total inhibition at 50 M (Fig. 5). Similar results were obtained when another inhibitor of tyrosine-specific protein kinase, genistein (46), was used (data not shown). These data provide added support that PAF:acyllyso-GPC transacetylase activity is controlled by reversible activation and inactivation through phosphorylation and dephosphorylation.
To gain insight as to which of the protein kinase(s) may be involved in the activation of PAF:acyllyso-GPC transacetylase besides tyrosine-specific protein kinase, we tested the effects of inhibitors for three different kinds of protein kinases, namely protein kinase C, protein kinase A, and casein kinase II, on both PAF:lysoplasmalogen transacetylase and PAF:acyllyso-GPC transacetylase activities (Table I). Based on the results we obtained previously from the mixed substrate experiments, the transfer of the acetate group from PAF to either lysoplasmalogen or acyllyso-GPC is catalyzed by a single enzyme (47). Therefore, it is not surprising that the effects of various protein kinase inhibitors are similar for both PAF:lysoplasmalogen transacetylase and PAF:acyllyso-GPC transacetylase activities (Table I). Only calphostin C, a specific inhibitor for protein kinase C, exerted inhibitory effect on both transacetylase activities, whereas the inhibitors for protein kinase A (KT 5720 or H-89) and casein kinase II (DRB) did not block neither ATPinduced transacetylase activities. Thus, our data suggest that tyrosine protein kinase and protein kinase C are directly or indirectly involved in the activation of the transacetylase activity through protein phosphorylation.
To determine which of the amino acids (i.e. tyrosine, serine, and/or threonine) in the transacetylase is/are phosphorylated by the protein kinase(s), we carried out several experiments similar to that described by Whitman et al. (48) to show the phosphorylation of tyrosine in the phosphatidylinositol 3-kinase. When ATP-treated or control untreated endothelial cell lysates were prepared and immunoprecipitated with polyclonal antibody against either phosphotyrosine, phosphoserine, or phosphothreonine (Zymed Laboratories Inc.) using a combined method specified by Whitman et al. (48), Domin et al. (49), and manufacturer's instructions, our preliminary data showed that the PAF:lysoplasmalogen transacetylase activity in ATPtreated endothelial cell lysates was enriched in part in the immunocomplex precipitated by anti-phosphotyrosine antibody in comparison to that of the untreated control endothelial cell lysates. Similarly, the PAF:lysoplasmalogen transacetylase activity was higher in anti-phosphotyrosine-treated immunoprecipitate of the ATP-stimulated endothelial cell lysates than that in antigen affinity purified IgG antibody precipitated immunocomplex of the same cell lysates (a kind gift from Dr. Steve Kennel, Oak Ridge National Laboratory). These results suggest that PAF:lysoplasmalogen transacetylase is phosphorylated by a tyrosine kinase. However, we are currently attempting to purify the transacetylase and prepare the antibody against the purified protein to conclusively address the questions concerning the specific protein kinase(s) and the phosphorylation site of the amino acids involved in the activation of the transacetylase.
A Protein Phosphatase Inhibitor, Sodium Vanadate, Potenti- 3. Effect of acid phosphatase and ATP on the PAF:acyllyso-GPC transacetylase activity from the homogenates of ATPstimulated endothelial cells. Endothelial cells were treated with ATP (10 Ϫ3 M) for 5 min, and the homogenates were isolated as described under "Experimental Procedures." The isolated homogenates were preincubated without (1st column) or with 5 g of acid phosphatase (2nd column), 5 g of boiled acid phosphatase (3rd column), and ATP (5 mM) plus Mg 2ϩ (10 mM) (4th column) for 15 min at room temperature before the transacetylase activities were assayed according to the methods described under "Experimental Procedures." In the 5th column, the isolated homogenates were first preincubated with and without 5 g of acid phosphatase for 15 min at room temperature. The acid phosphatase was removed by centrifuging the homogenates at 100,000 ϫ g for 60 min, and the isolated membrane fraction was then assayed for the transacetylase activity in the presence of ATP (5 mM) plus Mg 2ϩ (10 mM) and supernatants obtained from the homogenates that were not treated with acid phosphatase. Data were expressed as means Ϯ ranges and were a representative of five similar experiments.

ates the Incorporation of [ 3 H]Acetate into Radyl[ 3 H]acetyl-GPC in ATP-stimulated Endothelial
Cells-It is reasonable to assume that if PAF:acyllyso-GPC transacetylase activity is reversibly activated/inactivated through phosphorylation/dephosphorylation and is involved in the synthesis of acyl analogs of PAF, then a protein phosphatase inhibitor that potentiates the activation of transacetylase activity by ATP (see Fig. 4 and above) should also augment the incorporation of [ 3 H]acetate into acylacetyl-GPC. Results in Table II indeed showed that the incorporation of [ 3 H]acetate into radyl[ 3 H]acetyl-GPC was increased severalfold by vanadate in both ATP-treated and untreated endothelial cells. However, vanadate had no effect on the percent of [ 3 H]acetate incorporated into each subclass of radyl[ 3 H]acetyl-GPC (data not shown).
Alkyllyso-GPC/Acyllyso-GPC:Acetyl-CoA Acetyltransferases in Activated Endothelial Cells-To determine the role of alkyllyso-GPC/acyllyso-GPC:acetyl-CoA acetyltransferases in the synthesis of acylacetyl-GPC, we measured the K m and V max values, and the time course of induction by ATP for the acetyltransferases in the membrane fraction of the calf pulmonary artery endothelial cells. The V max was only slightly less when palmitoyllyso-GPC was the substrate in comparison with that when hexadecyllyso-GPC was the substrate (7.6 nmol/min⅐mg protein versus 9.2 nmol/min⅐mg protein). However, the acetyltransferase had a slightly higher substrate affinity for the palmitoyllyso-GPC than that of hexadecyllyso-GPC (8.3 versus 9.1 M). Additionally, both alkyllyso-GPC:acetyl-CoA acetyltransferase and acyllyso-GPC:acetyl-CoA acetyltransferase in the homogenates of calf pulmonary artery endothelial cells responded to a time-dependent ATP activation with maximal increase within 1-2 min after the addition of ATP (Fig. 6). The stimulation in activity by ATP was higher for acyllyso-GPC: acetyl-CoA acetyltransferase (190% over basal at 1 min) than that for alkyllyso-GPC:acetyl-CoA acetyltransferase (147% over basal at 2 min). Because the induction of alkyllyso-GPC/ acyllyso-GPC:acetyl-CoA acetyltransferase by ATP (peak increase in activity occurred between 1 and 2 min) preceded that of PAF:acyllyso-GPC transacetylase (maximal increase in activity occurred at 5 min) (also see Fig. 3 versus Fig. 6), our data implied that both acetyltransferases and transacetylase participated in the synthesis of acylacetyl-GPC. Holland et al. (19) have shown that the alkyllyso-GPC:acetyl-CoA acetyltransferase was transiently induced in thrombin-exposed human endothelial cells and in bradykinin-treated bovine pulmonary artery endothelial cells and concluded that the synthesis of PAF in intact cells was regulated by the alkyllyso-GPC:acetyl-CoA acetyltransferase. Furthermore, they found that the alkyllyso-GPC was a 2-fold superior substrate than acyllyso-GPC in both control and thrombin-activated human endothelial cells (19).
Effect of ATP on the PAF Acetylhydrolase Activity in Endothelial Cells-Since the incubations of [ 3 H]acetate and ATP with endothelial cells were short durations (1-10 min), the rates of incorporation of [ 3 H]acetate into radyl[ 3 H]acetyl-GPC that we measured should reflect mostly as the rates of synthesis of radyl[ 3 H]acetyl-GPC. Nevertheless, we did assess the possible effects ATP may have on the PAF acetylhydrolase activity. Results (data not shown) indicated that ATP had minimal effect on the PAF acetylhydrolase activity. DISCUSSION We have established that ATP (10 Ϫ3 M), bradykinin (10 Ϫ8 M), and ionophore A23187 (2 M) stimulate the incorporation of  6. Time-dependent activation of alkyllyso-GPC/acyllyso-GPC:acetyl-CoA acetyltransferase in ATP-treated endothelial cells. Monolayers of calf pulmonary artery endothelial cells were incubated with ATP (10 Ϫ3 M) for the times indicated and were assayed for the alkyllyso-GPC/acyllyso-GPC:acetyl-CoA acetyltransferase activities as described under "Experimental Procedures" except that in A hexadecyllyso-GPC was used as the substrate and in B palmitoyllyso-GPC was the substrate. Data were expressed as means Ϯ ranges and were a representative of duplicate experiments. p values were Ͻ0.05 when the acetyltransferase activities from the ATP-treated samples at 1 and 2 min were compared with that of untreated control at 2 min.
The intriguing observations of our work are the induction of PAF:acyllyso-GPC transacetylase activity (Fig. 2) and the modulation of PAF:acyllyso-GPC transacetylase activity by ATP through covalent modification of phosphorylation and dephosphorylation (Figs. [3][4][5]. In addition, the maximal activation of PAF:acyllyso-GPC transacetylase activity by ATP (occurred at 5 min, Fig. 2) lags behind that of alkyllyso-GPC/acyllyso-GPC: acetyl-CoA acetyltransferase activities (maximum at 1-2 min, Fig. 6). These results indicate that the activities of acetyltransferases and transacetylase are stimulated by ATP in a coordinated fashion. It is reported that PAF stimulates the alkyllyso-GPC:acetyl-CoA acetyltransferase activity in neutrophils (56). Nevertheless, PAF:acyllyso-GPC transacetylase is not induced or augmented by incubating the endothelial cells with PAF (1 M) for 3 min or PAF (1 M) plus ATP for 5 min, respectively (data not shown). Furthermore, the increase of PAF:acyllyso-GPC transacetylase was returned to near basal level after 10 min treatment of the endothelial cells with ATP (Fig. 2), whereas the incorporation of [ 3 H]acetate into acyl[ 3 H]acetyl-GPC remained elevated at 10 min (Fig. 1). These data suggest the possibility that other catabolic enzymes (i.e. 1-acylhydrolase and/or lysophospholipase) may be involved in regulating the level of acyl analogs of PAF; however, these issues are beyond the scope of the present investigation.
It is logical that the stimulation of alkyllyso-GPC:acetyl-CoA acetyltransferase activity by ATP precedes prior to the increase in the transacetylase activity, because alkyllyso-GPC:acetyl-CoA acetyltransferase supplies PAF which is one of the cosubstrates for the PAF:acyllyso-GPC transacetylase. Also, Prescott et al. (57) have shown that essentially all of the newly synthesized PAF remains associated with the endothelial cells so that PAF:acyllyso-GPC transacetylase will have easy access to PAF.
Our in vitro experiments showed that the PAF:acyllyso-GPC transacetylase activity was reduced by treatment of the homogenates of the ATP-stimulated endothelial cells with acid phosphatase. Reduction in PAF:acyllyso-GPC transacetylase activity could be reversed by incubating the acid phosphatasetreated and reconstituted (acid phosphatase-treated membrane fraction plus acid phosphatase-untreated soluble fraction) homogenates with ATP (Fig. 3). These results indicate that PAF:acyllyso-GPC transacetylase activity is activated and inactivated through reversible phosphorylation and dephosphorylation. It is not clear at present which types of protein kinase(s) and phosphatase(s) are involved in the direct phosphorylation and dephosphorylation of the transacetylase. Since both tyrosine kinase inhibitors (genistein and tyrphostin-25 (Fig. 5)) and protein kinase C inhibitor (calphostin C (Table I)) inhibit the stimulation of PAF:acyllyso-GPC transacetylase by ATP, these results suggest that the phosphorylation of the tyrosine residue(s), and/or serine and threonine residue(s) of the proteins, is involved in the signal transduction transmitted by ATP for the activation of the transacetylase activity. However, we cannot totally rule out the possibility that phosphorylation/dephosphorylation of another protein is affecting the transacetylase activity. Investigations are currently underway in our laboratory to determine the type(s) of protein kinase(s) and phosphatase(s) that mediate the direct activation and inactivation of the transacetylase activity.
We and others (19,23,24,58) have shown that alkyllyso-GPC:acetyl-CoA acetyltransferase can use acyllyso-GPC as a substrate, but at a reduced rate (50%), to generate acylacetyl-GPC. This is consistent with the findings reported in this work except that the differences in V max between alkyllyso-GPC: acetyl-CoA acetyltransferase and acyllyso-GPC:acetyl-CoA acetyltransferase are less in calf pulmonary artery endothelial cells than in other cell systems previously observed. Further studies are needed to discern the possibility for the existence of two isoforms of acetyltransferase. Importantly, based on the kinetic patterns of the [ 3 H]acetate incorporation into acyl-[ 3 H]acetyl-GPC and the inductions of acetyltransferases and transacetylase activities by ATP, and the potentiation effect of sodium vanadate on the increased synthesis of acylacetyl-GPC by ATP (Figs. 1, 2, and 6 and Table II), our studies establish that activations of alkyllyso-GPC:acetyl-CoA acetyltransferase, acyllyso-GPC:acetyl-CoA acetyltransferase, and PAF:acyllyso-GPC transacetylase by ATP contribute to the stimulated synthesis of acylacetyl-GPC in calf pulmonary artery endothelial cells.