Purification, Properties, Sequencing, and Cloning of a Lipoprotein-Associated, Serine-Dependent Phospholipase Involved in the Oxidative Modification of Low-Density Lipoproteins
Abstract A novel LDL-associated phospholipase A2 (LDL-PLA2) has been purified to homogeneity from human LDL obtained from plasma apheresis. This enzyme has activity toward both oxidized phosphatidylcholine and platelet activating factor (PAF). A simple purification procedure involving detergent solubilization and affinity and ion exchange chromatography has been devised. Vmax and Km for the purified enzyme are 170 μmol·min−1·mg−1 and 12 μmol/L, respectively. Extensive peptide sequence from LDL-PLA2 facilitated identification of an expressed sequence tag partial cDNA. This has led to cloning and expression of active protein in baculovirus. A lipase motif is also evident from sequence information, indicating that the enzyme is serine dependent. Inhibition by diethyl p-nitrophenyl phosphate and 3,4-dichloroisocoumarin and insensitivity to EDTA, Ca2+, and sulfhydryl reagents confirm that the enzyme is indeed a serine-dependent hydrolase. The protein is extensively glycosylated, and the glycosylation site has been identified. Antibodies to this LDL-PLA2 have been raised and used to show that this enzyme is responsible for >95% of the phospholipase activity associated with LDL. Inhibition of LDL-PLA2 before oxidation of LDL reduces both lysophosphatidylcholine content and monocyte chemoattractant ability of the resulting oxidized LDL. Lysophosphatidylcholine production and monocyte chemoattractant ability can be restored by addition of physiological quantities of pure LDL-PLA2.
- Received July 5, 1995.
- Accepted December 13, 1995.
Oxidation of LDL is widely recognized as the key early event in the development of atherosclerosis.1 2 3 Oxidation of LDL involves oxidation of the polyunsaturated fatty acid component of phospholipids. It is known that the oxidation of LDL ultimately leads to the conversion of PtdCho to lyso-PtdCho. Indeed, during the oxidation of LDL, as much as 40% of its PtdCho can be converted to lyso-PtdCho.4 The lyso-PtdCho thus formed shows selective loss of the fatty acid moiety at the 2 position, implying that LDL must have a phospholipase activity associated closely with it.5 Lyso-PtdCho has many biological activities and therefore may well play an important pathophysiological role in atherosclerosis. For example, the increased lyso-PtdCho content of oxidized LDL has been shown to be chemoattractant for human monocytes but not neutrophils6 and has been demonstrated to selectively induce the expression of endothelial leukocyte adhesion molecules.7 The raised lyso-PtdCho level of oxidized LDL has also been directly related to the ability of oxidized LDL to induce endothelial dysfunction in various arteries.8 9 10
LDL is known to possess PAF-AH activity.11 PAF-AH has been shown to be capable of hydrolyzing PtdCho’s that possess short and/or polar fatty acid substituents in the 2 position.12 These PtdCho’s with short polar 2 substituents are thought to be similar to oxidatively truncated PtdCho. Consequently, it has been proposed that PAF-AH is the phospholipase activity responsible for the production of lyso-PtdCho in oxidized LDL. Although it has been suggested that the apoprotein B component of LDL possesses a PAF-AH activity itself,5 the recent cloning and cDNA sequencing of PAF-AH13 argue strongly against this proposal. The protein sequence data reported here, along with our cloning of this enzyme, confirm that the enzyme responsible for this LDL-PLA2 activity and PAF-AH are one and the same enzyme.
Although the purification of PAF-AH has been reported previously,14 it is apparent both from the present work and from the recently reported cloning of PAF-AH13 that previous purifications were incomplete. The present study provides a simple, rapid, high-yielding purification for LDL-PLA2 (PAF-AH) with specific activity ≈10-fold greater than any previously reported. In this study we also provide novel data that clearly demonstrate that the LDL-PLA2 we have purified and cloned is completely responsible for the increased lyso-PtdCho content of oxidized LDL. In addition, we have found evidence to suggest that the gene for this enzyme is part of a translocation reported in a family with SVAS.15 This finding may provide a fascinating insight into the origins of this genetic disorder.
Blue Sepharose 6FF, blue Sepharose HiTrap, Mono Q HR 5/5, and PhastGel 10 to 15 gradient gels were obtained from Pharmacia. CHAPS, MES, MOPS, HEPES, Tris, bovine serum albumin, and protein cleavage reagents were obtained from Sigma. Polyvinylidenedifluoride membranes (Immobilon) were obtained from Millipore. PNGase F, digoxigenin glycan detection kit, and DIG glycan differentiation kit were obtained from Boehringer Mannheim. PAF was obtained from Cascade Biochemicals. Trypsin was obtained from Promega. DNGP was prepared as described previously.16 Reagent-grade solvents and HPLC-grade water were used for all HPLC procedures. Glass capillary tubes for column construction were obtained from Camlab. Plastic nuts and unions were from Alltech Associates, and column support frits (1/16-in outer diameter) were from Upchurch (Anachem). Poros RI column packings were obtained from Perceptive Biosystems. Vydac material was from HighChrom.
LDL-PLA2 activity was measured using either DNGP or [3H]PAF as a substrate. All assays were performed at 37°C in 50 mmol/L HEPES and 150 mmol/L NaCl, pH 7.4, unless stated otherwise. DNGP was prepared as a 10-mmol/L stock solution in methanol and diluted into buffer as required. LDL-PLA2 was added to 50 μmol/L DNGP in buffer at 37°C. The absorbance increase was followed at 400 nm, using either a diode array spectrophotometer (Hewlett-Packard) or a 96-well plate reader (Molecular Devices, Tmax) running in kinetic mode. Product was quantified using the published extinction coefficient, ε400=15 000·L·mol−1·cm−1.16 For PAF-AH activity, [3H]PAF and LDL-PLA2 were incubated in a final volume of 200 μL for 10 minutes at 37°C. The reaction was stopped by vortexing with 600 μL of CHCl3/MeOH (2:1), and the CHCl3 and aqueous layers were separated by centrifugation. The aqueous layer was removed (250 μL) and vortexed with 250 μL of CHCl3. The aqueous layer was again removed and the [3H]acetate determined by scintillation counting. Protein was determined using the Pierce bicinchonic acid assay kit, according to the manufacturer’s instructions.
To determine the inhibitory effect of various compounds, purified LDL-PLA2 and the compound were preincubated at 37°C for 10 minutes before running the enzyme assay as described above.
LDL was obtained after apheresis and stored at −20°C. LDL was dialyzed at 4°C against two volumes of 50 mmol/L MES and 0.5 mol/L NaCl, pH 6.0, overnight. Any precipitate was removed by centrifugation. Solid CHAPS was added to 10 mmol/L final volume and the solution stirred for 10 minutes to effect solubilization. The resulting solution was applied to a preequilibrated blue Sepharose 6FF column (5×15 cm). The column was washed with 50 mmol/L MES, 0.5 mol/L NaCl, and 10 mmol/L CHAPS at pH 6.0 and then 50 mmol/L MOPS, 0.5 mol/L NaCl, and 10 mmol/L CHAPS at pH 7.5, each time until the absorbance of the eluate at 280 nm stabilized. LDL-PLA2 enzyme was eluted with 50 mmol/L Tris, 1.5 mol/L NaCl, and 10 mmol/L CHAPS at pH 8.0. Active fractions were pooled, concentrated by ultrafiltration (Amicon YM30) and dialyzed overnight against 50 mmol/L Tris and 10 mmol/L CHAPS at pH 8.0.
After dialysis, the LDL-PLA2 was loaded onto a preequilibrated (50 mmol/L Tris, 10 mmol/L CHAPS at pH 8.0) Mono Q HR5/5 column and a gradient from 0 to 1 mol/L NaCl applied. LDL-PLA2 eluted as a broad peak at ≈0.2 mol/L NaCl. Active fractions were pooled, concentrated by ultrafiltration, and dialyzed against 50 mmol/L MES, 0.5 mol/L NaCl, and 10 mmol/L CHAPS at pH 6.0.
LDL-PLA2 activity was loaded onto a preequilibrated (50 mmol/L MES, 0.5 mol/L NaCl, 10 mmol/L CHAPS at pH 6.0) HiTrap blue cartridge (1 mL) and the same washing and elution procedure as for the blue Sepharose 6FF column were used. Active fractions were pooled and concentrated (Amicon YM30) before storage at 4°C. LDL-PLA2 was found to be stable under these conditions for several months.
For the preparation of antibodies to LDL-PLA2, protein prepared as above was subjected to size exclusion chromatography on a Superose 12 column using 20 mmol/L sodium phosphate and 150 mmol/L NaCl at pH 7.4.
Preparation of Antibodies to LDL-PLA2 and Immunoadsorption Studies
Polyclonal antibodies against human LDL-PLA2 were prepared in rabbits (2.5 kg) by using standard protocols. Briefly, a 1:1 emulsion of purified LDL-PLA2 (200 μg) in saline and Freund’s complete adjuvant was injected subcutaneously into four sites. Identical booster injections of antigen (50 to 100 μg) with incomplete adjuvant were given 4 weeks later. Excellent antibody titer was observed at 6 to 7 weeks after the initial immunization. Immunoadsorption analysis using solid-phase reagents (protein A agarose–based) coupled to rabbit serum raised against purified human LDL-PLA2 was carried out as detailed previously.17
Deglycosylation of LDL-PLA2
Deglycosylation using PNGase F18 was carried out as follows. Five microliters of LDL-PLA2 (≈1 mg/mL) was mixed with 5 μL of 0.2 mol/L Tris, 0.1 mol/L EDTA, 2% β-mercaptoethanol at pH 8.0. Two microliters of PNGase F (0.4 U) was added and the mixture incubated overnight at 37°C. For deglycosylation of larger quantities of LDL-PLA2, eg, for sequencing, the above procedure was scaled up directly.
Determination of pH Dependence of Activity
The pH dependence of Vmax was determined using both DNGP and [3H]PAF as substrates. For both substrates, the velocity was measured at both 100 and 200 μmol/L substrate. The fact that the measured velocity was the same for both substrate concentrations was taken to indicate that Vmax was being measured. The buffers used (50 mmol/L each) over the different pH ranges were pH 5.6 to 6.6, MES; pH 6.8 to 8.2, HEPES; and pH 8.4 to 9.0, AMPSO. All buffers contained 150 mmol/L NaCl. When DNGP was substrate, product formation was corrected for the pKa of p-nitrophenol and background was corrected for buffer-catalyzed substrate hydrolysis at high pH.
Protein, 100 to 300 μg, was collected by preparative RP-HPLC for chemical cleavage. For methionine cleavages by CNBr, the protein was resuspended in 50 μL of 70% TFA, a small crystal of CNBr added under nitrogen, and the sample left overnight at room temperature in the dark. The reaction mixture was removed by vacuum centrifugation and 50 μL of 6 mol/L guanidine HCl added to ensure peptide solubilization before RP-HPLC separation. A tryptophan cleavage using 2-(2′-nitrophenylsulfenyl)-3-methyl-3′-bromoindolamine was adapted from Crimmins et al,19 with the inclusion of two extractions with ether (300 μL) to reduce excess reagent before RP-HPLC separation. Tryptic peptides were prepared by the addition of trypsin at a ratio of 1:50 with substrate directly to the deglycosylation mixture described above.
A Beckman System Gold HPLC was used, fitted with a microbore mixer and model 167 dual-wavelength UV detector. Microbore HPLC columns were prepared from glass capillary tubes as described by Southan,20 except for the substitution of a newer type of 1/16-in frit in the lower fitting (Upchurch part No. C407). Columns were assembled without a top frit to reduce blocking by particulate material and dry packed with Poros RI, a reverse-phase material suitable for high-speed separations.21 These were used for the analysis of protein fractions and the isolation of large peptides from the tryptophan cleavage mixture. Solvent A was 0.08% TFA in water, and solvent B 0.06% TFA in 80% acetonitrile. Typical conditions for a 0.65×50-mm Poros 10-μm RI column were a flow rate of 0.5 mL/min and gradients of 0% to 80% solvent B in 8 minutes. At several stages of purification, we used the area under the protein peak detected at 215 nm for protein quantification by comparing this area with a 1-μg BSA standard injected under identical conditions.22 For preparative collection of protein, Poros RI was packed into a 2.1-mm standard PEEK column. For CNBr and tryptic peptide separations, a 0.96×80-mm glass capillary was slurry packed at 2000 psi with Vydac C18 10-μm material. This was run with the solvents described above at 0.1 mL/min, and samples were separated with a gradient of 0% to 70% solvent B in 35 minutes. For the trypsin digest, a cleaning gradient rising from 70% to 100% solvent B in 5 minutes was used to elute hydrophobic peptide fractions.
Protein- or peptide-containing HPLC peaks were reduced in volume down to 10 to 20 μL by vacuum centrifugation, taking care to avoid drying. Samples were stored at −20°C before Edman sequencing with a Perkin-Elmer/Applied Biosystems 477a protein sequencer using standard protocols. Sample disks were loaded with polybrene for all peptide samples. To minimize background signals, these disks were used consecutively for up to six peptide runs before a fresh disk was applied. Data were processed using the 900 data system and checked by manual inspection of the chromatograms.
Mass spectrometry was performed with a matrix-assisted laser desorption instrument (Lasermat, Finnigan MAT). Sinapinic acid was used as a matrix at 10 mg/mL in 70% acetonitrile with 0.1% TFA. Sample (0.5 μL) and matrix (0.2 μL) were dried on the sample slides.
Cloning, Baculovirus Expression, and Chromosomal Localization
Extensive peptide data from purified human LDL-PLA2 were used to search a human EST database. After a 100% identity match with an EST, a full-length LDL-PLA2 cDNA was isolated from a human T-cell lymphoma cDNA library, using the EST as a probe. Final matches with peptide data covered 64% of the reading frame. The initial EST clones used in the study were prepared by scientists at the Institute for Genomic Research using established EST methods.23 24 These clones are part of a larger EST project (M.D. Adams et al, unpublished data, 1996). Library screening, DNA sequencing, and general molecular cloning techniques were carried out according to standard protocols.25 26 27 Generation of recombinant virus to infect the cultured insect cell line Spodoptera frugiperda 9 (SF-9) and so express LDL-PLA2 was carried out according to standard procedures.26 28 The LDL-PLA2 cDNA was localized to chromosome 6 by hybridization to a panel of somatic hybrids (BIOS Laboratories). Chromosome 6 was the only human chromosome yielding a value of 0% discordance.
Preparation of DENP-Treated LDL
Human LDL was prepared from fresh EDTA/plasma by density-gradient ultracentrifugation followed by fractionation. LDL-PLA2 activity was irreversibly inhibited by treating 3-mL portions of LDL (1.5 mg protein per milliliter) with 1 mmol/L DENP for 60 minutes at 37°C. DENP and EDTA were then removed by gel filtration on a Superdex 200 (prep grade) column (1.6×35 cm) preequilibrated with phosphate-buffered saline as running buffer. DENP- or vehicle-treated LDL pools were adjusted to 0.2 mg/mL and split in two, with one portion supplemented with purified LDL-PLA2 (50 ng/mL final concentration).
Analysis of the Lyso-PtdCho Content and Monocyte Chemotactic Activity of Copper-Oxidized LDL
After the addition of 5 μmol/L copper, oxidation was allowed to proceed to completion at 37°C by monitoring the rate of conjugated diene formation at 234 nm (ie, 1.8 to 2.0 absorbance units change). No oxidation was observed in samples lacking copper. Incubations were terminated (by the addition of organic solvents) and lipids extracted as previously described.29 Half of this sample was dried down, vortexed in 1 mL Hanks’ balanced salt solution containing 0.1% BSA (wt/vol), and immediately assayed for human monocyte chemotactic activity using a previously described procedure.30 The other half of the lipid sample was spotted onto high-performance TLC plates, developed in chloroform/methanol/25% to 30% methylamine (60/20/5, vol/vol/vol), and phospholipids were visualized by careful spraying with the fluorescence indicator TNS (1 mmol/L in 50 mmol/L Tris-HCl, pH 7.4). Fluorescence was measured using a CAMAG TLC scanner and the lyso-PtdCho content quantified via a standard curve to 0.05 to 0.6 mg synthetic 1-palmitoyl lyso-PtdCho. Each individual TLC plate had its own standard curve to account for slight variations in plate spraying, and these routinely demonstrated excellent linearity (r>.98).
As reported previously,5 a phospholipase activity was found to be associated with LDL after separation of LDL from plasma proteins and other lipoproteins by density-gradient ultracentrifugation. Also, the LDL-PLA2 activity from individuals with familial hypercholesterolemia was significantly higher than that found in the plasma from normal individuals.31 32 As a consequence of this discovery, we decided to use human LDL available to us from patients undergoing plasma apheresis as a starting material. This procedure gave us both a purer starting material relative to whole plasma and a higher total enzyme activity relative to “normal” LDL.
To separate the phospholipase from LDL, we solubilized the LDL with 10 mmol/L CHAPS. This method was found to separate activity from the majority of the LDL protein, as judged by Superose 12 size-exclusion chromatography (data not shown), without affecting enzyme activity.
Blue Sepharose 6FF chromatography was found to be a key step in purifying LDL-PLA2. Loading the protein at pH 6.0 allows most proteins and solubilized lipid to pass through the column, while the LDL-PLA2 binds. The stepwise loading, washing, and elution protocol used was more convenient than a series of gradients varying among the three buffers used and gave good reproducibility. This step alone typically provided between 500- and 1000-fold purification in good yield. The partially purified protein was then subjected to anion exchange HPLC on a Mono Q FPLC column. The LDL-PLA2 activity showed a rather broad elution profile (Fig 1⇓), uncharacteristic of high-resolution ion exchange. Finally, the blue Sepharose chromatography was repeated on a small scale using a HiTrap blue cartridge (Fig 2⇓). This three-step procedure yielded protein of approximately >95% purity, as judged by HPLC (Fig 3⇓). If higher purity is required, the LDL-PLA2 can be subjected to size-exclusion chromatography in the absence of CHAPS. A typical purification is summarized in Table 1⇓.
Although RP-HPLC shows the protein to be ≈95% pure, SDS-PAGE indicates two or three broad bands between the 43- and 67-kD markers (Fig 4⇓). To investigate the possibility that this heterogeneity was due to glycosylation, LDL-PLA2 was treated with the enzyme PNGase F, which is known to hydrolyze N-linked carbohydrates off the asparagine residue to which they are attached.18 SDS-PAGE before and after PNGase F treatment is shown in Fig 4⇓. It is clear that PNGase F treatment of LDL-PLA2 converts the multiplet of bands seen on SDS-PAGE to a single band of molecular weight 47 kD, thus indicating that LDL-PLA2 is extensively N glycosylated. This supposition was confirmed by running SDS-PAGE of LDL-PLA2 both before and after PNGase F treatment, and then blotting onto polyvinylidene difluoride and staining for carbohydrate using a glycan detection kit (Boehringer Mannheim). Native LDL-PLA2 stains strongly for carbohydrate, while LDL-PLA2 treated with PNGase F shows no staining (Fig 5⇓). This suggests that all of the carbohydrate is N linked.
Matrix-assisted laser desorption mass spectrometry of LDL-PLA2 shows a broad range of masses from 45 to 52 kD. Deglycosylation of LDL-PLA2 sharpens the mass spectrum and indicates a molecular weight of 47 kD (data not shown).
When ≈200 pmol LDL-derived protein was applied to the protein sequencer, the following sequences were recorded in quantitative order: FD()QY, KIPRG, and GQT(K)IP. The relative yields of these sequences were variable between preparations, and in some cases overall initial yields were low and background signals high, indicating the possible presence of N-terminally blocked material on the sample disk.
Initial experiments showed the native enzyme to be resistant to solution cleavage with trypsin or lys-C protease, and solubility was problematic after S-carboxymethylation. Effective cleavage was obtained with CNBr on RP-HPLC–purified native protein. A successful tryptic cleavage was obtained by adding protease directly to the deglycosylation, suggesting that glycosylation was at least in part responsible for the protease resistance of native enzyme. After performing cleavages and RP-HPLC separations as described in “Methods,” the peptide sequences were obtained as indicated in Fig 6⇓.
From overlapping peptides, a pair of extended contiguous sequences were assembled (Fig 6⇓), including their N-terminal residues from the known specificity of reagent cleavages. These were used for database searching.
Cloning and Baculovirus Expression of Human LDL-PLA2
The peptide sequence information above enabled the identification of an EST and facilitated cloning of a full-length LDL-PLA2 cDNA encoding a predicted polypeptide of 441 amino acids (Fig 6⇓). The predicted molecular weight of 50 kD is in good agreement with mass spectrometric measurements above, allowing for removal of a predicted signal sequence. Analysis by the PROSITE33 program identified a section of sequence, -GHSFGG-, which conforms to the lipase consensus motif, GXSXXG.34 This finding supports our biochemical characterization of this enzyme as a new form of mammalian lipase and locates the putative essential serine residue involved in the catalytic mechanism. We subsequently expressed the cDNA, which resulted in high levels of active LDL-PLA2 in a baculovirus system. The expressed enzyme was sensitive to DENP treatment and specifically was immunoadsorbed by antiserum against human LDL-PLA2, thus confirming its identity (data not shown).
LDL-PLA2 shows normal Michaelis-Menten kinetics, with both PAF and the synthetic substrate used throughout the purification. This is highly unusual for a phospholipase, as interfacial activation is normally seen for this class of enzyme. The Km for PAF and DNGP is essentially identical, 13 μmol/L and 12 μmol/L, respectively, while Vmax is 170 U/mg and 112 U/mg, respectively. Thus Vmax is greater than an order of magnitude higher than that reported previously for PAF-AH.14 The pH dependence of Vmax is very flat for both substrates over the range 5.5 to 9 (Fig 7⇓). Optimal pH is ≈7 to 7.5, with a noticeable decrease in activity above pH 8.5. Deglycosylation with PNGase F has no significant effect on activity.
In agreement with previous reports for lipoprotein-associated phospholipase activity, LDL-PLA2 activity is insensitive to EDTA, Ca2+, and thiol reagents such as iodoacetic acid (Table 2⇓). The potent serine protease inhibitor DFP is a weak inhibitor of LDL-PLA2, while the related compound DENP is significantly more potent (Table 2⇓). In addition, 3,4-dichloroisocoumarin, a serine protease inhibitor, was found to be a potent LDL-PLA2 inhibitor. However, unlike DFP and DENP, this inhibition was found to be slowly reversible.
The inhibition by DFP, DENP, and 3,4-dichloroisocoumarin, along with the insensitivity to EDTA, Ca2+, and thiol reagents such as iodoacetic acid, suggests that LDL-PLA2 is a serine-dependent hydrolase.
Antibodies raised to purified LDL-PLA2 were able to immunoprecipitate >95% of LDL-PLA2 activity in CHAPS-solubilized LDL (Fig 8⇓), confirming that LDL-PLA2 is the sole phospholipase associated with LDL.
To confirm the key role played by LDL-PLA2 in the formation of lyso-PtdCho during LDL oxidation, DENP was used to inhibit endogenous LDL-PLA2 activity in LDL before oxidation by Cu2+. Because DENP is an irreversible inhibitor of LDL-PLA2, the LDL was gel filtered after inhibition to remove excess DENP and its hydrolysis products. This procedure had the advantage of both minimizing any artifactual effects due to DENP’s directly affecting Cu2+-catalyzed oxidation and allowing pure LDL-PLA2 to be added back to restore LDL-PLA2 activity to the LDL before oxidation. Fig 9A⇓ clearly shows that the lyso-PtdCho produced by Cu2+-catalyzed oxidation of LDL is directly related to the presence of LDL-PLA2 activity whether the activity is endogenous or a result of adding pure LDL-PLA2 back to DENP-treated LDL. In addition, the enhanced monocyte chemotactic activity of oxidized LDL is seen only when LDL-PLA2 activity is present in the LDL sample, Fig 9B⇓.
LDL-PLA2 and SVAS
A search of Genbank with the full LDL-PLA2 nucleotide sequence outlined in Fig 6⇓ revealed a 98% identity of the first 73 bp of the 5′ untranslated region with the 3′ end of a sequence determined for a translocation allele.15 This translocation event forms a novel junction in exon 28 of the elastin gene on chromosome 7, with uncharacterized sequence (≈12.7 kb) from chromosome 6 in a family with autosomal-dominant SVAS. Curran and coworkers15 were unable to identify any significant open reading frames within 1000 bp of the break point. The identification of a new stop codon 6 bp downstream from the translocation break point (leading potentially to production of a truncated elastin) suggested that such a mutation in the elastin gene could cause SVAS in these families. However, the significant identity over the 73 bases between the 5′ end of two independently derived LDL-PLA2 clones, T-cell lymphoma (Fig 6⇓) and activated monocyte (data not shown), and the 3′ region of the translocation sequence provides a potential new interpretation for the molecular basis of SVAS. We propose that the “unknown” sequence from chromosome 6 is in fact LDL-PLA2 (Fig 10⇓). To support this hypothesis, chromosomal localization studies have localized LDL-PLA2 to chromosome 6 (data not shown).
We have purified an LDL-associated PLA2 activity to homogeneity in a rapid, simple three-column procedure from LDL. The final specific activity was 170 μmol·min−1·mg−1. Purity was confirmed by three criteria. (1) Only a single protein peak was observed by RP-HPLC. (2) Several applications of RP-HPLC–purified material to the protein sequencer gave no N-terminal sequence but did show a rising background of PTH signals, indicating the presence of only N-terminally blocked protein on the sample disk. (3) None of the accumulated peptide sequence data gave database matches indicative of contamination with known proteins. Immunoadsorption experiments using antibodies raised to this enzyme confirm that LDL-PLA2 comprises >95% of the phospholipase activity associated with LDL, and this phospholipase activity also accounts for >95% of PAF-AH activity. Thus, our phospholipase accounts for >95% of the activity ascribed previously to PAF-AH. In addition, the specific activity reported here is >10-fold higher than any value reported previously for this enzyme. This underlines the improvements in the purity of the enzyme isolated here compared with previously reported preparations.14
LDL-PLA2 appears to be unusual for a phospholipase in that it shows no signs of interfacial activation. When DNGP was used as a substrate, no deviation from simple Michaelis-Menten kinetics was seen up to 100 μmol/L substrate. Complete insensitivity to Ca2+ and EDTA indicates that this PLA2 is not a member of the family of Ca2+-dependent phospholipases. DFP, a serine hydrolase inhibitor, had a modest inhibitory effect. However, the related compound DENP was 10-fold more potent than DFP as an inhibitor of LDL-PLA2. These findings strongly suggest that LDL-PLA2 is a serine-dependent lipase and is in keeping with the lipase motif34 identified in the primary sequence predicted from the cDNA (Fig 6⇓).
The use of DENP as an LDL-PLA2 inhibitor has allowed us to further underline the important role this phospholipase plays in the oxidation of LDL. Inhibition of LDL-PLA2 activity results in a lack of both lyso-PtdCho production and enhanced monocyte chemotactic activity of oxidized LDL despite the LDL’s being fully oxidized, as judged by conjugated diene formation measured at 234 nm.35 Restoration of LDL-PLA2 activity to DENP-treated LDL by supplementation with physiological quantities of pure LDL-PLA2 (50 ng/mL) restores both lyso-PtdCho production during LDL oxidation and the monocyte chemotactic activity after oxidation. Again, the extent of oxidation is identical to the control, as judged by conjugated diene formation. Restoration of lyso-PtdCho production and enhanced monocyte chemotactic activity confirms that LDL-PLA2 is key to the altered physiological properties of oxidized LDL.
The phospholipase described here is clearly not limited to PAF and PAF analogues as substrates, as evidenced by the use of DNGP as a substrate and its ability to hydrolyze oxidized PtdCho on LDL. Hydrolysis of oxidized PtdCho releases lyso-PtdCho and an oxidized fatty acid moiety. Both of these products are known potential inflammatory mediators. Thus, although this hydrolase clearly has the ability to inactivate PAF, it may well play other physiologically relevant roles. In a study of the correlation between the deficiency of PAF-AH and respiratory symptoms in asthmatic children, a significant proportion of the Japanese population was found to be completely deficient in PAF-AH.36 However, there were no significant differences between the enzyme activities of the patients with and without asthmatic attacks. If PAF plays a role in asthma, LDL-PLA2 (PAF-AH) would not appear to be an influence. Thus, LDL-PLA2 appears to have two quite opposite possible in vivo effects. On the one hand, as PAF-AH, it deactivates PAF and therefore could be an anti-inflammatory enzyme; on the other, its ability to hydrolyze oxidized LDL is chemoattractant for monocytes, and therefore, it may act as a proinflammatory enzyme. The data presented here, which clearly demonstrate that LDL-PLA2 activity is responsible for the lyso-PtdCho content and the monocyte chemoattractant properties of oxidized LDL, are supportive of the idea that this enzyme is proinflammatory. We would suggest, therefore, that a more general name, such as LDL-PLA2, be used, that PAF-AH is a somewhat misleading name because the enzyme has properties beyond the hydrolysis of PAF. This broader identification will hopefully aid in settling the debate on the role of this unusual phospholipase.
We have shown that the first 73-bp section of the 5′ untranslated region of our LDL-PLA2 clone has a 98% identity with the 3′ end of the translocated sequence shown to be present in a family with autosomal-dominant SVAS. Curran and coworkers15 demonstrated that this translocation (from chromosome 6 to 7) disrupted the elastin gene and speculated that mutations in elastin can cause SVAS. We speculate that this translocation may actually be a translocation of the LDL-PLA2 gene. We have located the LDL-PLA2 gene to chromosome 6, in keeping with this idea. We have established that the LDL-PLA2 gene is ≈8 kb long (data not shown) and that there is at least a 12-kb portion of chromosome 6 downstream from the translocation break point identified by Curran et al.15 The 680 bp upstream from the initiation codon for LDL-PLA2 may contain all the regulatory elements necessary for expression; indeed, consensus TATA and CAAT boxes can be identified in this sequence. Thus, it is possible that the translocation identified by Curran et al15 not only disrupts the elastin gene but also leads to inappropriate expression of LDL-PLA2. Inappropriate expression of LDL-PLA2 could lead to the pathological features in SVAS, ie, intimal proliferation of vascular smooth muscle cells and fibroblasts, causing significant narrowing of large elastic arteries,37 since, for example, the major product of LDL-PLA2 action, lyso-PtdCho, is known to selectively increase the cellular mRNA for the potent smooth muscle mitogens platelet-derived growth factor and heparin-binding epidermal growth factor–like protein.38 39 Clearly, in the absence of biochemical data, this is speculation on our part. However, given the similarity of the pathologies of SVAS and atherosclerosis, it would seem to be more than coincidental that the LDL-PLA2 gene is found to be translocated into the site thought to be responsible for SVAS in this case.
Selected Abbreviations and Acronyms
|DENP||=||diethyl p-nitrophenyl phosphate|
|EST||=||expressed sequence tag|
|HPLC||=||high-performance liquid chromatography|
|PNGase F||=||peptide N-glycoside F|
|SDS-PAGE||=||SDS-polyacrylamide gel electrophoresis|
|SVAS||=||supravalvular aortic stenosis|
We wish to thank Dr Gilbert Thompson and Clare Newers, Hammersmith Hospital, London, for supply of LDL, and Kevin Milliner and Kenneth Fantom for excellent technical assistance.
Holvoet P, Collen D. Oxidized lipoproteins in atherosclerosis and thrombosis. FASEB J. 1994;8:1279-1284.
Parthasarathy S, Steinbrecher UP, Barnett J, Witztum JL, Steinberg D. Essential role of phospholipase A2 activity in endothelial cell–induced modification of low density lipoprotein. Proc Natl Acad Sci U S A. 1985;82:3000-3004.
Parthasarathy S, Barnett J. Phospholipase A2 activity of low density lipoprotein: evidence for an intrinsic phospholipase A2 activity of apoprotein B-100. Proc Natl Acad Sci U S A. 1990;87:9741-9745.
Quinn MT, Parthasarathy S, Steinberg D. Lysophosphatidylcholine: a chemotactic factor for human monocytes and its potential role in atherogenesis. Proc Natl Acad Sci U S A. 1988;85:2805-2809.
Kume N, Cybulsky MI, Gimbrone MA. Lysophosphatidylcholine, a component of atherogenic lipoproteins, induces mononuclear leukocyte adhesion molecules in cultured human and rabbit arterial endothelial cells. J Clin Invest. 1992;90:1138-1144.
Murohara T, Kugiyama K, Ohgushi M, Sugiyama S, Ohta Y, Yasue H. LPC in oxidized LDL elicits vasoconstriction and inhibits endothelium-dependent relaxation. Am J Physiol. 1994;267:H2441-H2449.
Stafforini DM, McIntyre TM, Carter ME, Prescott SM. Human plasma platelet-activating factor acetylhydrolase: association with lipoprotein particles and role in the degradation of platelet-activating factor. J Biol Chem. 1987;262:4215-4222.
Stremler KE, Stafforini DM, Prescott SM, McIntyre TM. Human plasma platelet-activating factor acetylhydrolase: oxidatively fragmented phospholipids as substrates. J Biol Chem. 1991;266:11095-11103.
Stafforini DM, Prescott SM, McIntyre TM. Human plasma platelet-activating factor acetylhydrolase: purification and properties. J Biol Chem. 1987;262:4223-4230.
Macphee CH, Harrison SA, Beavo JA. Immunological identification of the major platelet low-Km cAMP phosphodiesterase: probable target for anti-thrombotic agents. Proc Natl Acad Sci U S A. 1986;83:6660-6663.
Tarentino AL, Plummer TH. PNGase F. Methods Enzymol. 1987;138:777-778.
Crimmins DL, McCourt DW, Thoma RS, Scott MG, Macke K, Schwartz BD. In situ chemical cleavage of proteins immobilized to glass-fiber and polyvinylidenedifluoride membranes: cleavage at tryptophan residues with 2-(2′-nitrophenylsulfenyl)-3-methyl-3′-bromoindolenine to obtain internal amino acid sequence. Anal Biochem. 1990;187:27-38.
Southan C. The use of glass capillary tubes as microbore columns for HPLC of proteins and peptides. In: Hugli T, ed. Techniques in Protein Chemistry. San Diego, Calif: Academic Press Inc; 1988:392-398.
Southan C. Fast, flexible, sensitive, and cheap: the use of home made microcolumns for the separation of proteins and peptides. J Protein Chem. 1994;13:461-462.
Adams MD, Kelley JM, Gocayne JD, Dubnick M, Polymeropoulos MH, Xiao H, Merril CR, Wu A, Olde B, Moreno RF, Kerlavage AR, McCombie WR, Venter JC. Complementary DNA sequencing: expressed sequence tags and human genome project. Science. 1991;252:1651-1656.
Church GM, Gilbert W. Genomic sequencing. Proc Natl Acad Sci U S A. 1984;81:1991-1995.
Sambrook J, Fritsch EF, Maniatis T. Molecular Cloning: A Laboratory Manual. New York, NY: Cold Spring Harbor Laboratory Press; 1987.
Bullock WO, Fernandez JM, Short JM. XL1-blue: a high efficiency plasmid transforming recA Escherichia coli strain with beta-galactosidase selection. Biotechniques. 1987;5:376-379.
Summers MD, Smith GE. A Manual of Methods for Baculovirus Vectors and Insect Cell Culture Procedures. Texas Agriculture Experiment Station Bulletin No. 1555; 1987.
Bligh EG, Dyer WJ. A rapid method of total lipid extraction and purification. Can J Biochem Physiol. 1959;37:911-917.
Quinn MT, Parthasarathy S, Fong LG, Steinberg D. Oxidatively modified low density lipoproteins: a potential role in recruitment and retention of monocyte/macrophages during atherogenesis. Proc Natl Acad Sci U S A. 1987;84:2995-2998.
Bairoch A. The PROSITE dictionary of sites and patterns in proteins: its current status. Nucleic Acids Res. 1993;21:3097-3103.
Halliwell B, Gutteridge JMC. Free Radicals in Biology and Medicine. Oxford, England: Clarendon Press; 1989:220-223.
Miwa M, Miyake T, Yamanaka T, Sugatani J, Suzuki Y, Sakata S, Araki Y, Matsumoto M. Characterization of serum platelet-activating factor (PAF) acetylhydrolase: correlation between deficiency of serum PAF acetylhydrolase and respiratory symptoms in asthmatic children. J Clin Invest. 1988;82:1983-1991.
Nakano T, Raines EW, Abraham JA, Klagsbrun M, Ross R. Lysophosphatidylcholine upregulates the level of heparin-binding epidermal growth factor–like growth factor mRNA in human monocytes. Proc Natl Acad Sci U S A. 1994;91:1069-1073.
Kume N, Gimbrone MA Jr. Lysophosphatidylcholine transcriptionally induces growth factor gene expression in cultured human endothelial cells. J Clin Invest. 1994;93:907-911.