Enhanced association of platelet-activating factor acetylhydrolase with lipoprotein (a) in comparison with low density lipoprotein.

Paired samples of human Lipoprotein (a) (Lp(a)) and low density lipoprotein (LDL) were assayed for their platelet-activating factor (PAF) acetylhydrolase activity. Lp(a) displayed markedly enhanced PAF acetylhydrolase activity (approximately 7-fold based on equal particle concentrations) in comparison to LDL isolated from the same individual. Lp(a)-associated acetylhydrolase exhibited properties observed for LDL-associated acetylhydrolase as well as for the purified enzyme; significant inhibition was obtained by treatment with diisopropylfluorophosphate (1 mM, 90%) and phenylmethanesulfonyl fluoride (5 mM, 50%). Furthermore, the hydrolytic activity of both lipoproteins was abolished with paraoxon (6 mM, IC50 0.9 mM) and with the fluorescent and active site-directed probe 4-hexyl-(6'-O-butyl-(4'-pyrenyl))-benzoic estersulfonyl fluoride (2) (KI(inact) = 525 microM), a novel irreversible inhibitor of PAF acetylhydrolase. Treatment with 2 and subsequent quantitation of protein-bound fluorescence suggests an increased concentration of enzyme associated to Lp(a) rather than alterations of kinetic constants due to the additional apolipoprotein apolipoprotein (a). Exposure of Lp(a) to Cu2+ (20 microM, 37 degrees C) was followed by a concomitant decrease of hydrolytic activity. A reduction of the basal activity by 91% was found after 15 h. Whereas immunoprecipitation with anti-apoB antiserum could remove enzymatic activity of Lp(a) regardless of a reductive treatment with dithiothreitol, precipitation with anti-apolipoprotein (a)-antibodies was accompanied by a minor reduction (approximately 30%) of the PAF-hydrolyzing ability. These results suggest that PAF acetylhydrolase exhibits an enhanced association with Lp(a) due to an increased affinity to Lp(a) apolipoprotein B.

Paired samples of human Lipoprotein (a) (Lp(a)) and low density lipoprotein (LDL) were assayed for their platelet-activating factor (PAF) acetylhydrolase activity. Lp(a) displayed markedly enhanced PAF acetylhydrolase activity (approximately 7-fold based on equal particle concentrations) in comparison to LDL isolated from the same individual. Lp(a)-associated acetylhydrolase exhibited properties observed for LDL-associated acetylhydrolase as well as for the purified enzyme; significant inhibition was obtained by treatment with diisopropylfluorophosphate (1 mM, 90%) and phenylmethanesulfonyl fluoride (5 mM, 50%). Furthermore, the hydrolytic activity of both lipoproteins was abolished with paraoxon (6 mM, IC 50 0. 9

mM) and with the fluorescent and active site-directed probe 4-hexyl-(6-O-butyl-(4-pyrenyl))-benzoic estersulfonyl fluoride (2) (K I(inact)
‫؍‬ 525 M), a novel irreversible inhibitor of PAF acetylhydrolase. Treatment with 2 and subsequent quantitation of protein-bound fluorescence suggests an increased concentration of enzyme associated to Lp(a) rather than alterations of kinetic constants due to the additional apolipoprotein apolipoprotein (a). Exposure of Lp(a) to Cu 2؉ (20 M, 37°C) was followed by a concomitant decrease of hydrolytic activity. A reduction of the basal activity by 91% was found after 15 h. Whereas immunoprecipitation with anti-apoB antiserum could remove enzymatic activity of Lp(a) regardless of a reductive treatment with dithiothreitol, precipitation with anti-apolipoprotein (a)-antibodies was accompanied by a minor reduction (approximately 30%) of the PAF-hydrolyzing ability. These results suggest that PAF acetylhydrolase exhibits an enhanced association with Lp(a) due to an increased affinity to Lp(a) apolipoprotein B.
Lp(a) 1 is a lipoprotein from human plasma differing from low density lipoprotein by apo(a) an additional unique glycoprotein (Ref. 1; for review see Ref. 2). Apo(a) is linked to apoB, the main protein component of LDL, by a single disulfide bond (3,4) and exhibits a homology to plasminogen (5). In clinical studies, elevated Lp(a) plasma concentrations have been shown to correlate with the incidence of stroke and coronary heart disease (6 -8), and an accumulation in atherosclerotic lesions has been demonstrated (9), features that provide evidence for elevated Lp(a) levels as a substantial risk factor. Moreover, an overexpression of apo(a) promotes the development of lesions in transgenic animals (10). Although Lp(a) has also been suggested to interfere in fibrinolysis (11,12), the actual proatherogenic mechanisms remain unknown. Amidolytic activity was found to be associated with apo(a) some time ago (13). More recently, it has been reported that apo(a) contains a trypsin-like protease domain (14) and is capable of catalyzing the cleavage of fibronectin (15).
LDL as well as high density lipoprotein contains an associated hydrolytic activity characterized as lipoprotein-associated platelet-activating factor (1-O-alkyl-2-acetyl-sn-glycero-3phosphocholine) acetylhydrolase (PAF-AH) (16 -19). This unique type of enzyme utilizes a serine residue to catalytically cleave phospholipid ester bonds and is also likely to control the PAF concentration in human plasma (20). The enzyme also exhibits a marked selectivity for phospholipids with a short acyl group (21). Hence, the enzyme is capable of hydrolyzing compounds generated by lipid peroxidation (22), which abolishes the PAF-like activity of these substances (23). In conjunction with hydrolytic activity, it has become increasingly apparent that the platelet-activating factor acetylhydrolase exhibits a protective role in terms of oxidative LDL-modification (24,25) and cell-cell interactions induced by minimally modified LDL (26). Furthermore, an anti-inflammatory effect could be demonstrated in animal experiments (20). These results are in line with our previous data concerning the uptake of oxidatively modified LDL by macrophages, a process found to be affected by the susceptibility of oxidizable phospholipids to undergo enzymatic hydrolysis (27). Uncontrolled receptor-mediated uptake leading to the accumulation of lipids in macrophages and to their conversion to foam cells, however, is considered to be a critical feature of atherosclerosis (28). It follows that the level of the hydrolytic activity of PAF-AH could be one parameter determining the atherogenic potential of oxidizable lipoproteins. In this context, it is of interest that we most recently described that Lp(a) contains a phospholipase A 2 activity (29); however, to our present knowledge nothing is known about a PAF hydrolyzing activity of Lp(a).
In this study we have quantitated and characterized the PAF  [12][13][14][15][16]1995. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This study is dedicated to Prof. R. Neidlein, Heidelberg, on the occasion of his 65th birthday.
Analytical Procedures-After thin layer chromatography, the percentage of PAF hydrolysis was determined with a two-dimensional TLC scanner (Berthold Digital Autoradiograph, Berthold, Wildbad, Germany); proton nuclear magnetic resonance spectra were obtained on a Bruker WM 300 spectrometer (Bruker Physik AG, Karlsruhe, Germany), and proton chemical shifts are relative to tetramethylsilane as internal reference. Multiplicities are reported as singlet (s), doublet (d), triplet (t), or multiplet (m); the mass spectrum was recorded on a MAT 311 A (Varian, Bremen, Germany). A spectrofluorometer was used for fluorescence measurement (Kontron SFM 23). A Foss Heraeus vario EL analyzer was used for elemental analysis. The term "satisfactory" designates elemental analyses with deviations that are Յ0.4%.
Preparation, Analysis, and Treatment of Lipoproteins-Paired samples of LDL and Lp(a) (all S-1 isoform) were obtained from individuals with elevated Lp(a) plasma levels (Ͼ45 mg/dl). Lipoproteins were isolated from plasma containing EDTA (1 mg/ml) and butylated hydroxytoluene (20 M) as described earlier (33) by sequential flotation in an ultracentrifuge at densities of 1.025-1.060 g/ml for LDL and 1.070 -1.120 g/ml for Lp(a). Butylated hydroxytoluene and EDTA (at the above concentrations) were added to all buffers used during isolation. Further purification of Lp(a) was performed by chromatography over Biogel A 5 m (30). LDL and Lp(a) were Ն97% pure as judged by immunodiffusion and polyacrylamide gel electrophoresis. Lipoprotein concentrations were measured by protein quantitation according to Lowry (50) and by cholesterol quantitation; for LDL a protein content of 22.5 and a cholesterol content of 40% (w/w) was assumed. The corresponding values for Lp(a) were 34 and 36%, respectively. Protein molecular masses of 500 kDa were used for calculation of LDL concentration (particle concentrations) and 850 kDa for Lp(a) (molecular mass of apo(a) S-1 ϭ 350 kDa) (34). A 10 mM DTT solution was used for reductive cleavage of the disulfide apoB-apo(a) linkage (2 h at 37°C), and the reaction was controlled by agarose gel electrophoresis (0.5%) in Tris borate (pH 8.3, 8 mM, EDTA 0.2 mM); protein staining was carried out with Coomassie Blue R250. For determination of PAF-AH activity of Lp(a) subsequent to electrophoresis, each lane was sliced (0.5 cm), pressed through a syringe (0.5 mm), and dispersed into 1.5 ml of Tris-HCl buffer (100 mM, pH. 7.8, 0.1% Tween 20, EDTA 5 mM). The sample was extracted overnight at 4°C and centrifuged (30 min at 10,000 ϫ g), and enzymatic activity and protein content were determined in the supernatant. Prior to oxidation, the lipoproteins were dialyzed against Tris-HCl buffer (100 mM, pH 7.4) or passed through Econopac ® columns (Bio-Rad). Oxidation of LDL was initiated by the addition of CuSO 4 (final concentration, 20 M), and the samples were kept at 37°C for up to 24 h. Aliquots were taken, and the reaction was terminated by the addition of EDTA (final concentration, 2 mM).
Isolated lipoproteins (final concentration, 0.12 mg/ml) were treated with the inhibitors paraoxon, PMSF, and DFP (100 mM stock solutions (7:7:2:4, v/v/v/v); furthermore, a 100-l aliquot of the reaction mixture was extracted with toluene or methylene chloride (2 ϫ 250 l), the organic phase was concentrated to approximately 50 l, and a 10-l aliquot was analyzed by HPTLC (hexane/ethyl acetate (3:1, v/v)). (2)-The kinetics of PAF-AH inactivation were determined using lipoproteins (0.5 ml, 120 g/ml Lp(a) or LDL) or partially purified enzyme (0.5-ml samples, 0.12 g/ml, approximately 1340-fold purified from human plasma according to Ref. 17 up to the DEAE-column step, the preparation produced 687 nM lyso PAF/min/mg) as a source of enzymatic activity. The reaction (at 37°C) was started by the addition of 2 (from a 100 mM stock solution in Me 2 SO) to the enzyme in incubation buffer. At selected times, aliquots were removed and diluted 20 -80-fold (increasing with inhibitor concentration) into an assay mixture, and residual activity was assayed as described above.

Inactivation of PAF-AH by 4-Hexyl-(6Ј-O-butyl-(4Ј-pyrenyl))-benzoic Ester Sulfonyl Fluoride
Chemical Nature of Inactivation Process-The irreversibility of inhibition was tested in two ways. First, the inactivated enzyme (partially purified PAF-AH, 60 ng or LDL, 30 g and 2, 450 M as described) was separated from excess inhibitor by dialysis (incubation buffer for 24 h at 4°C with three changes using 2000 times the sample volume each time), and the sample was assayed for return of activity. Second, the sample was subjected to anion-exchange chromatography (1.5 ml of Whatman DE 52 equilibrated with the above buffer in disposable mini columns (Bio-Rad)). The column was washed with 8 ml of dialysis buffer, and the sample (active enzyme as determined in control experiments) was then eluted with 2 ml of the same buffer containing 0.25 M NaCl. After dialysis and concentration, the sample was assayed for activity. Neither procedure restored enzymatic activity.

Determination of Protein-bound Fluorescence after Reaction with 4-Hexyl-(6Ј-O-butyl-(4Ј-pyrenyl))-benzoic Ester Sulfonyl Fluoride
(2)-To estimate the amount of lipoprotein-associated PAF acetylhydrolase, a sample (0.5 ml, 120 g/ml) of Lp(a) or LDL was incubated with compound 2 (final concentration 450 M) in Tris-HCl incubation buffer for 3 h at 37°C. The solution was dialyzed against the above Tris-buffer, and the excess reagent was removed (and the apoproteins precipitated) by treatment with precooled acetone (5 ml at Ϫ20°C). The mixture was kept at Ϫ20°C, and after 1 h, the precipitate was isolated by centrifugation and resuspended in another 5 ml of solvent. The residue obtained after the third washing was dissolved in 250 l of NaOH (50%, w/w) and kept at 40°C for 24 h. After cooling the solution was neutralized with HCl (final volume, 500 l) and extracted with toluene or methylene chloride (2 ϫ 250 l), and the organic phase was separated by brief centrifugation. Fluorescence in the residual aqueous phase as fluorescence in further extracts (Ͼ2) was below the detection limit. The extract was dried (Na 2 SO 4 ) and concentrated to a volume of 100 l in a stream of nitrogen, and the fluorescence of a 50-l aliqot measured (methylene chloride, exc 344 nm, em 398 nm as found for max of 1); data obtained from lipoproteins are compared on the basis of relative flourescence intensities of samples with equal protein concentrations and are corrected by substraction of fluorescence values obtained for residual unspecific binding as determined by incubation of 2 with DFP-pretreated lipoproteins; approximately 6% of the proteinassociated fluorescence (background) was left after DFP-pretreatment (conditions as described for inactivation of the antisera, treatment with 2 as outlined above). In control experiments, radioiodinated lipoproteins were subjected to the above protocol, and 79 Ϯ 14% (LDL, mean of 4 determinations) and 84 Ϯ 18% (Lp(a), mean of 3 determinations) of the radioactivity was recovered in the pellet after the third acetone washing. Because the N-bromosuccinimide method (35) used for the labeling of the lipoproteins affords a maximum of 18% of lipoproteinbound radioiodine associated with the lipids, these data indicate that the protein fraction of the samples is precipitated completely and is retained almost quantitatively in the pellet. In further experiments, LDL samples labeled with 2, (5 ml, 120 g/ml protein) or samples of plain compound 2 in buffer were processed as per the procedure described above (acetone extraction of the lipoprotein samples only). HPTLC analysis (hexane/ethyl acetate (3:1, v/v)) including authentic standard and fluorescence spectroscopy confirmed complete conversion of protein-bound (or free) compound 2 to the alcohol 1 and the absence of other fluorescent compounds.

PAF-AH Activity of Lp(a) and Susceptibility to Oxidative
Inactivation-We assayed the PAF acetylhydrolase activity of paired Lp(a) and LDL samples isolated from human plasma. In all samples we measured consistently significantly elevated levels of lipoprotein (a)-associated hydrolytic activity. Results based on equimolar concentrations of lipoprotein particles are presented in Fig. 1 and indicate that the Lp(a)-associated acetylhydrolase activity exceeded those found associated with LDL 6.9-fold. Analysis of the data (8.9 nM (lyso-PAF)/min ϫ mg protein for Lp(a) and 2.2 nM/min ϫ mg for LDL) on the basis of equal protein concentrations affords a ratio of the activities of 4 (Lp(a)):1 (LDL).
The LDL-associated PAF acetylhydrolase is sensitive to Cu 2ϩ -initiated lipid peroxidation, a process that is accompanied by a concomitant decrease of enzymatic activity (36). To determine the relative sensitivity of the Lp(a) associated enzyme, the time-course was analyzed under pro-oxidative conditions (Fig. 1). We found that the Lp(a)-associated activity decreased somewhat faster than the LDL-associated activity. Although 50% of activity was left in Lp(a) after 8 h, the LDLassociated activity was reduced to 57%. After 15 h, Lp(a) displayed 9% of the initial activity, whereas LDL contained 19% of the basal activity. The relative decrease in hydrolytic activity of Lp(a) was slightly enhanced under pro-oxidative conditions; the absolute activity, however, exceeded the LDL-associated activity at all times monitored.
Inhibition of Lp(a)-associated PAF Acetylhydrolase Activity-To characterize the properties of the enzymatic activity of Lp(a) in comparison to LDL, we used DFP and PMSF, which are known inhibitors of the plasma enzyme (37,38). As shown in Fig. 2, substantial inhibition (reduction by 90%) of both samples was achieved by incubation with 1 mM DFP for 2 h. A parallel reduction of the hydrolytic capacity was observed after treatment with 5 mM PMSF (60% for LDL and 52% for Lp(a)) and with the cholinesterase inhibitor paraoxon (1 mM) (49). We found that paraoxon reduced the activity of both particles equally in a concentration-dependent manner (Fig. 2, inset) with an IC 50 of approximately 0.9 mM. At a concentration of 6 mM, complete inhibition was observed in both samples. As reported for the LDL associated enzyme (38,39), PAF hydrolysis by Lp(a) could not be reduced by treatment with p-bromophenacylbromide and was insensitive to DTT at a concentration of up to 100 mM. In agreement with a previous report (38), C 6 -1-palmitoyl-2-[(7-nitrobenz-2-oxa-1,3-diazol-4-yl)-amino]hexanoyl-sn-glycero-3-phosphocholine was found to be a substrate of the Lp(a) associated activity. As expected, the Lp(a) associated activity failed to cleave the ester bond of 1-palmitoyl-sn-2-glycero- [5,6,8,9,11,12,14, H] -arachidonyl-3-phosphocholine, indicating a preference for substrates with short acyl chains as reported for the isolated protein (20) (data not shown). Hence, the hydrolytic activity of Lp(a) displayed all characteristics known for the LDL-associated enzyme as well as for the purified PAF acetylhydrolase, indicating that this enzyme is the sole source of the Lp(a)-associated PAF hydrolytic activity.
Effects of Glycolytic Enzymes-Because apo(a) is a highly glycosylated protein (3), we examined whether treatment with glucosidases could affect either enzymatic activity or association of the enzyme to Lp(a). Incubation (4 h) of Lp(a) (120 g/ml) with a combination of ␤-glucosidase (200 milliunits/ml) and neuraminidase (100 milliunits/ml) caused a loss of 88% of sialic acid (as determined from apoLp(a) by the thiobarbituric acid procedure according to Ref. 40) and a loss of 47% of total carbohydrate content (determined from apoLp(a) according to Ref. 41, average values of two experiments). However, only 20% of enzymatic activity (relative to an untreated control as determined after reisolation by gel filtration chromatography or by native agarose electrophoresis, data not shown) was lost suggesting that glycosylation is not critical to enzyme-apoprotein interactions, to the access of substrate and/or to the release of products from the active site.
Apoprotein Association of PAF Acetylhydrolase Activity-To probe the association of the enzyme to apo-Lp(a) we used antibodies for immunoprecipitation. Treatment of Lp(a) with rabbit or sheep anti-apo(a)-antiserum and subsequent centrifugation resulted in a significant loss of enzymatic activity (8 and 15% residual activity, respectively, Table I). Because apo(a) can be dissociated from apoB under reducing conditions (42), we investigated the enzyme-apoprotein association after disulfide reduction. Lp(a) was treated with DTT, and apo(a) was precipitated by use of the antisera. A substantial portion of PAF acetylhydrolase activity (67 and 74% of control activity, respectively) remained in the supernatant. Incubation and precipitation with anti-apoB antibodies, however, removed the activity almost completely, a result that was unaffected by a preceeding DTT treatment. The loss of acetylhydrolase activity found in DTT-pretreated Lp(a) after incubation with anti-apo(a) antiserum may be explained by partial coprecipitation of apoB with apo(a) immunocomplexes due to noncovalent interactions. (4ЈЈ-pyrenyl))-benzoic ester sulfonyl fluoride (2) (Fig. 3) was designed to achieve nonradioactive labeling of PAF-AH, allowing the determination of enzyme bound compound or vice versa, the quantitation of enzyme. We reasoned that the poor inhibitory potency of the PAF-AH inhibitor PMSF may be improved by the introduction of an aliphatic chain attached to the phenyl group carrying the reactive methylenesulfonyl fluoride moiety. This variation promised an increase of the structural resemblance to the original substrate PAF. Moreover, the reactivity of the sulfonylfluoride with nucleophilic amino acid residues should be enhanced by a direct phenyl-sulfonyl fluoride linkage. We therefore synthesized the pyrene-labeled compound 2 and characterized its properties and in particular its effects on lipoprotein associated as well as on partially purified PAF acetylhydrolase. First, the behavior in the absence of PAF acetylhydrolase activity was examined. To estimate the chemical stability of compound 2, the reagent was incubated in plain buffer up to 3 h. No hydrolysis was observed by TLC analysis. We next added peptide 1 (a hydrophylic fragment of apoB (3120 -3137) (43) containing several nucleophilic amino acid residues including one serine) to a solution of 2 and repeated the experiment. Using TLC analysis we found that for up to 3 h, no labeling of the peptide occurred, a result that indicates that unspecific reactions between the fluorescent probe 2 and peptides are unlikely.

Characterization of the Fluorescent Probe 4-Hexyl-(6Ј-O-butyl-(4Ј-pyrenyl))-benzoic Ester Sulfonyl Fluoride (2)-The novel fluorescent agent 4-hexyl-(6Ј-O-butyl-
Incubation with Lp(a) or other PAF-AH-containing samples resulted in a concomitant time-and concentration-dependent loss of enzymatic activity. The inhibitory activity of 2 is provided in Fig. 4. Because excessive inhibitor was employed, no recovery of enzymatic activity was observed, and inactivation followed first-order kinetics, inactivation was analyzed according to the reaction outlined in Equation 1. Reaction of the enzyme (E) with the inhibitor (I) affords the irreversible inhibitor enzyme complex E-I upon equilibration with the reversible enzyme complex E‫ء‬I. 1/k obs ϭ K I͑inact͒ /k 2 ͓I͔ ϩ 1/k 2 (Eq. 3) (Fig. 4, inset) provided the K I(inact) value ϭ 525 M and the k 2 value 0.03 min Ϫ1 . Consistent with the notion that covalent modification of the hydrolase has taken place, extensive dialysis of the inhibitorenzyme reaction mixture as anion-exchange chromatography failed to restore enzymatic activity.
Correlation between the Content of Active PAF Acetylhydrolase and Protein-bound Fluorescence-Given a defined ratio between active PAF acetylhydrolase and protein-bound fluorescence after inactivation with compound 2, the fluorescent probe should enable the comparison of the amounts of LDL-and Lp(a)-associated enzyme. To analyze the relationship between protein-bound fluorescence and the relative amount of active enzyme, Lp(a) samples were inactivated with DFP and recombined with untreated fractions of the same lipoprotein. The combined samples then were treated with compound 2 at conditions providing complete inactivation of residual enzymatic activity, and protein-bound fluorescence was determined (Fig.  5). A linear relation (r 2 ϭ 0.98) was observed between the fraction of enzymatically active lipoprotein and the fluorescence associated with the apoprotein. Similar data were obtained in analog experiments with recombined LDL samples (r 2 ϭ 0.96, agreement of data points with the presented values of Lp(a) within 19% when samples with the same enzymatic activity prior to treatment with 2 were used) as well as in experiments with partially inactivated purified enzyme (r 2 ϭ 0.98, data not shown). These results indicate that the analysis of of apoproteins Residual PAF acetylhydrolase activity after treatment with anti-apoB and anti-apo(a) antibodies. Lipoproteins (0.12 mg/ml) were incubated with limiting amounts of antibodies to apo(a) and apoB (16 h at 4°C) followed by precipitation with protein A-Sepharose. The remaining activity is given relative to an equally treated (except for the antibodies) control sample. Antisera were inactivated with DFP as described under "Experimental Procedures." Except for sample size, conditions of PAF-AH activity determination were as described in the legend to Fig. 1.  (44). Data points indicate a mean of three experiments that agreed within 16% at each data point. Similar data were obtained with samples of LDL (0.5 ml, 120 g/ml) and with partially purified acetylhydrolase (0.5 ml, 0.12 g/ml).
protein-bound fluorescence can be utilized as a measure of the amount of active enzyme. Consistently, pretreatment of Lp(a) with DFP to complete enzymatic inactivation precludes fluorescence labeling almost completely (Fig. 5, inset), suggesting that both agents occupy the same enzymic site. Moreover, the presence of the unhydrolyzable PAF analog N-acetyl-PAF (K Irev with purified PAF-AH about 370 M in our system) 2 during incubation with compound 2 as well as the presence of excessive substrate (3 mM PAF, not shown) block protein labeling (Fig. 5, inset) indicating an active site-specific reaction of inhibitor 2 with the enzyme.
Comparison of Enzyme Content of LDL and Lp(a)-The increased level of enzymatic activity found with Lp(a) could be a reflection of the binding of a greater amount of enzyme. Inactivation of lipoprotein-associated enzyme with compound (2) was carried out to estimate the relative content of active enzyme in both lipoproteins. Using equal protein concentrations, we compared protein-bound fluorescence measured upon treatment of both lipoproteins, LDL and Lp(a), with the sulfonyl fluoride (2). We found that Lp(a) (the apoproteic part) released 3.4-times more fluorescent compound 1 (Fig. 5, inset), a value that compares favorably with the ratio (4:1) of the lipoproteinassociated PAF-hydrolyzing activities. DISCUSSION In the experiments presented we have characterized the Lp(a)-associated PAFacetylhydrolase activity and studied the differences to the LDL-associated one. We demonstrate for the first time that Lp(a) exhibits severalfold enhanced PAF acetyl-hydrolase activity compared with LDL. Data analyzed on a molar basis show that the Lp(a)-associated activity exceeds the LDL-associated activity approximately 7-fold. The enzyme associated with Lp(a) exhibits a number of known properties like substrate specificity and sensitivity to different inhibitors found in the LDL-associated PAF acetylhydrolase as well as in the purified protein. These results confirm that the Lp(a) associated PAF acetylhydrolase is identical to the enzyme found associated with LDL.
Because Lp(a)-associated PAH acetylhydrolase is likely to affect oxidative modification of the Lp(a) particle, we determined its susceptibility to oxidative inactivation in comparison to the LDL-associated activity. We found that despite Lp(a)'s greater stability toward oxidation (45), the relative decrease in hydrolytic activity of Lp(a) was slightly enhanced under prooxidative conditions. This suggests that the enzyme, when it is associated with Lp(a), may be more susceptible to oxidative damage than the enzyme associated with LDL. The absolute activity (per particle or per mg protein), however, exceeded the activity shown by the LDL-associated enzyme at all times monitored.
Considering the observations outlined above, it was of particular interest to determine the basis for the differences between lipoprotein-associated activity found with LDL and with Lp(a). Therefore, the fluorescent compound 2 was prepared as a tool to estimate the quantity of lipoprotein-associated enzyme and to analyze the reason for the difference in hydrolytic abilities observed between LDL and Lp(a). This compound irreversibly inhibits lipoprotein-associated or partially purified PAF acetylhydrolase, and the observed progressive development of inhibition is consistent with the hypothesis that it reacts with the enzyme (possibly with the likely active site nucleophile serine 273) (20) to yield a sulfonyl enzyme derivative. Kinetics of 2 is typical for an irreversible inhibitor (44), and the amount of fluorescent compound bound to the protein fractions of the lipoproteins after treatment with 2 is linearly related to the content of active enzyme. This was evident when samples of partially inactivated lipoproteins were combined with enzymatically active fractions, and the protein-bound fluorescence of the mixture was determined. We further found that similar results were also obtained when samples of different classes of partially inactivated lipoproteins (LDL and Lp(a)) were combined (enzymatic activities were found to be additive) and subsequently inactivated with the fluorescent probe 2. 2 Therefore, in the micellar assay system, the kinetics of substrate hydrolysis as well as the reaction with the sulfonyl fluoride (2) appears to be predominantly independent of the presence of other lipoprotein components.
Several experiments provide evidence that fluorescence labeling with 2 is active site-specific. Furthermore, the line obtained by plotting the values of k obs according Equation 3 did not pass through the origin but intercepted the positive y axis indicating the initial formation of reversible complexes. Our data further indicate incorporation of fluorescent label into the enzymatically active enzyme at a defined and constant ratio (label/mol enzyme), thus demonstrating the fulfillment of requirements for enzyme quantitation. The sulfonyl fluoride (2) or related compounds thus may represent sensitive tools to label and quantitate PAF acetylhydrolase, and further structural variations promise compounds with increased inhibitory potency.
The lipoprotein environment alters the catalytic behavior of PAF-AH (46). However, our results suggest that the increased hydrolytic activity of Lp(a) is primarily due to the binding of a greater number of enzyme molecules. Previous studies from our laboratories provide evidence that Lp(a) displays a de-2 C. Blencowe, unpublished observations. FIG. 5. Correlation between protein-bound fluorescence and the fraction of active Lp(a)-associated enzyme. Lp(a) samples were mixed with an increasing fraction of DFP-inactivated Lp(a) (the total volume (0.5 ml) as well as the protein concentration (120 g/ml) was kept constant), the mixtures were treated with 2 (450 M, 3 h 37°C), and the the protein-bound fluorescence intensity ( exc. 344 nm, em. 398 nm) was measured after protein precipitation, basic proteolysis, and extraction with toluene as described under "Experimental Procedures." The labeling of Lp(a) protein with 2 (inset) can be blocked almost completely by preincubation of Lp(a) (120 g/ml) with DFP (10 mM, 37°C, 60 min) or by the presence of N-acetyl-PAF 4 (4 mM) during treatment of Lp(a) with inhibitor (2); 1, fluorescence extracted from Lp(a)-protein; 2, fluorescence extracted from LDL protein after labeling of the lipoproteins (paired samples, both 120 g/ml) with inhibitor (2) creased lipid mobility on the surface (29,47), a feature that may affect the association of apoproteins (48) or could be due to apo-Lp(a) interactions. Hence, differences found in the association of PAF acetylhydrolase with LDL and Lp(a) could be attributed to variations in apoprotein-lipid interactions affecting the binding of the enzyme and/or might be due to alterations in direct interactions between the enzyme and the lipid surface.
Taken together, our data suggest that PAF acetylhydrolase exhibits an enhanced affinity to Lp(a)-apoB, resulting in an increased enzyme/lipoprotein ratio relative to LDL. It remains to be clarified how the potentially protective role of PAF acetylhydrolase is compatible with the apparent atherogenic properties of Lp(a).