Aldose Reductase-catalyzed Reduction of Aldehyde Phospholipids

SUMMARY

Oxidation of unsaturated phospholipids results in the generation of aldehyde side chains that remain esterified to the phospholipid backbone. Such “core” aldehydes elicit immune responses and promote inflammation. However, the biochemical mechanisms by which phospholipid aldehydes are metabolized or detoxified are not well understood. In the studies reported here, we examined whether aldose reductase (AR), which reduces hydrophobic aldehydes, metabolizes phospholipid aldehydes. Incubation with AR led to the reduction of 5-oxovaleroyl, 7-oxo-5-heptenoyl, 5-hydroxy-6-oxo-caproyl, and 5-hydroxy-8-oxo-6-octenoyl phospholipids generated upon oxidation of 1-palmitoyl-2-arachidonoyl-sn-glycero-3-phosphocholine (PAPC). The enzyme also catalyzed the reduction of phospholipid aldehydes generated from the oxidation of 1-alkyl, and 1-alkenyl analogs of PAPC, and 1-palmitoyl-2-arachidonoyl phosphatidic acid or phosphoglycerol. Aldose reductase catalyzed the reduction of chemically synthesized 1-palmitoyl-2-(5-oxovaleroyl)-sn-glycero-3-phosphatidylcholine (POVPC) with a K_m of 10 μM. Addition of POVPC to the culture medium led to incorporation and reduction of the aldehyde in COS-7 and THP-1 cells. Reduction of POVPC in these cells was prevented by the AR inhibitors sorbinil and tolrestat and was increased in COS-7 cells overexpressing AR. Together, these observations suggest that AR may be a significant participant in the metabolism of several structurally diverse phospholipid aldehydes. This metabolism may be a critical regulator of the pro-inflammatory and immunogenic effects of oxidized phospholipids.

INTRODUCTION

Oxidation of cellular lipids is associated with the development of a diverse set of pathological conditions, including atherosclerosis (1–3), diabetes (4, 5), heart failure (6, 7), Alzheimer’s disease (8, 9), and ischemia-reperfusion injury (10–12). In addition, chronic inflammation and toxicity caused by xenobiotic or toxicant exposures cause lipid oxidation, which destroys the structural integrity of membranes, and disrupts metabolic and signaling pathways. Products of lipid peroxidation are highly reactive. They form covalent adducts with proteins and DNA, dysregulate cell growth, and cause apoptosis (2–6, 10–12). Accumulation of protein and DNA adducts derived from the products of lipid peroxidation has been linked to age-induced tissue dysfunction (13, 14) and spontaneous carcinogenesis (15–17).

Lipid peroxidation is a multistep process that involves the abstraction of electrons from unsaturated fatty acids by metals or free radicals and results in the formation of alkoxyl and peroxyl radicals. These radicals in turn oxidize additional substrates, thereby perpetuating an autocatalytic cycle, which is ultimately terminated by radical-radical interactions. The overall process leads to the formation of many non-radical intermediates and end products including peroxides, aldehydes, epoxides, and isoprostanes (18–20). Because most biological phospholipids have unsaturated fatty acids esterified to the sn-2 position, oxidation of these lipids generates an ω fragmentation product (e.g. 4-hydroxalkenals, carboxylic acids) and an abbreviated sn-2 chain. Phospholipids containing oxidation products at the sn-2 position (“core” aldehydes) have been detected in oxidized low density lipoprotein (LDL),¹ atherosclerotic plaques, aged erythrocytes, apoptotic cells, and the plasma of smokers (21–24).

Oxidized phospholipids have high biological activity. Oxidation of phosphoglycerides at the sn-2 position generates products that are structurally similar to the platelet-activating factor (PAF). These and related phospholipids activate the PAF receptors and trigger downstream signaling events leading to smooth muscle cell growth (25), vasoconstriction, monocyte adhesion, and platelet aggregation (21, 23–25). Oxidized derivatives of phosphatidylcholine also stimulate inflammatory responses and promote monocyte binding to endothelial cells (22, 26). Moreover, oxidation of phospholipids results in the appearance of neo-epitopes that serve as recognition sites for the macrophage scavenger receptors and mediate recognition and removal of apoptotic cells by activated macrophages (27), suggesting that oxidized phospholipids may be general ligands for macrophage recognition and phagocytosis.

Aldehydes are the major neoepitopes that appear upon oxidation and fragmentation of unsaturated fatty acids esterified to the sn-2 position (26, 28–31). High levels of such aldehydes are formed when phospholipids containing unsaturated fatty acids are oxidized in vitro (26, 30). Aldehyde phospholipids derived from arachidonic acid oxidation have been detected in minimally modified (mm) LDL and in atherosclerotic lesions of rodents and humans (26, 28–31). Similar aldehydes derived from linolenic or linoleic acids are also present in oxLDL (28). Previous investigations show that oxidized and fragmented phospholipids are hydrolyzed by the plasma enzymes, platelet-activating factor acetyl hydrolase (PAF-AH; Refs. 23, 32–36), and lecithin:cholesterol acyl transferase (LCAT; Refs. 37–39), and by the intracellular PAF-AH (40–42). Nevertheless, it remains unknown whether these phospholipids undergo intracellular oxidation-reduction reactions.

We have previously reported that short to medium (C3–C12) chain aldehydes and their glutathione conjugates are efficiently reduced by the polyol pathway enzyme aldose reductase (AR; Refs. 43–46). These studies indicated that AR has wide substrate specificity and an unusually accommodating active site that efficiently recognizes large hydrophobic and hydrophilic aldehydes (44, 47, 48). Based on these observations, we formulated the hypothesis that AR may be an efficient catalyst for reducing phospholipid aldehydes. Accordingly, we tested the efficacy of the enzyme with aldehyde phospholipids generated from the oxidation of acyl, alkyl, and alkenyl-linked phospholipids and phospholipid-containing different head groups. Our results show that AR efficiently reduces multiple classes of aldehyde phospholipids and participates in the intracellular transformation of phospholipid aldehydes. These results reveal a novel pathway for the metabolism of oxidized phospholipids, which may be a significant regulator of atherogenesis, oxidative injury, inflammation, and apoptosis.

EXPERIMENTAL PROCEDURES

Materials

PAPC, its alkyl, 1-O-hexadecyl-2-arachidonoyl-PC (alkyl-PAPC), and plasmenyl, 1-O-hexadec-1′-enyl-2-arachidonoyl-PC (p-PAPC) analogs, 1-palmitoyl-2-arachidonoyl-sn-glycerol 3-phosphoglycerol (PAPG), and 1-palmitoyl-2-archidonoyl-sn-glycero-3-phosphatidic acid (PAPA) were purchased from Avanti Polar Lipids (Alabaster, AL). Sephadex G-25 columns were purchased from Amersham Biosciences. Sorbinil and tolrestat were gifts from Pfizer Chemical Company and American Home Products, respectively. Solvents and other analytical grade reagents were obtained from Sigma. The COS-7 cells were obtained from American Type Culture Collection (ATCC) and human monocytoid cells (THP-1) were obtained from Clonetics. Cell culture reagents were purchased from Invitrogen Life Technologies.

Molecular Modeling

The structure of AR was generated from the homologous structure FR-1 (PDB entry: 1FRB.ent; Ref. 49) using MOD-ELER (50). The homology model is consistent with the AR structure deposited in the (PDB entry 1MAR; Ref. 51), which only contains the carbon atoms for AR, and was used for the docked conformational search of POVPC. The NADP⁺ was placed in the same orientation as in FR-1. In this structure the coenzyme binds to the protein in an unusually extended conformation and relative to ribose, both nicotinamide and adenine adopt the anti conformation. For modeling, NADP⁺ was converted to NADPH, and minimized in situ while restraining AR. Although the nicotinamide ring of NADPH is puckered (52), to examine steric interaction a simpler planar conformation of NADPH was used. In AKRs the binding site residues are stronger determinants of substrate orientation than the oxidation state of the cofactor and simulations of substrate binding with planar or bent NADPH show similar conformers (53). The lowest energy starting structure of POVPC was generated and refined using molecular mechanics and dynamics (Macromodel 7.0 Ref. 54, AMBER^* force field, GBSA implicit water solvation, minimization: steepest descents, 1000 steps; conjugate gradient, 1000 steps; 100-ps dynamics at 300 K; time step, 1.5 fs; sampling 20 structures that were subsequently minimized as above, and the lowest energy structure used subsequently). The docking of POVPC with AR was performed using a manual dock, placing POVPC in the active site, followed by simulated annealing (sampling 100 structures over 300 ps using molecular dynamics, 1.5-fs time step, 750 °C, AMBER^* force field, GBSA implicit water solvation, followed by minimization of each structure using steepest descents (1000 steps) conjugate gradient (1000 steps)) of the system with POVPC unrestrained and keeping AR and NADPH coordinates fixed. Distance restraints were added to be consistent with the mechanism previously proposed (55). The distance of the aldehyde oxygen to His¹¹⁰ and, Tyr⁴⁸ and carbon of aldehyde to NADPH was 2.00, 2.53, and 2.59 Å, respectively.

Synthesis of POVPC

POVPC was synthesized as outlined in Scheme 1. Briefly, 5-pentanolide (compound 1, 0.1 mol) in dry methanol (200 ml) was mixed with 2 μmol of concentrated sulfuric acid, and the reaction mixture was refluxed for 5 h and then cooled on an ice/salt bath, after which 1 g of sodium hydrogen carbonate was added. The reaction mixture was stirred for 10 min, filtered, and concentrated under reduced pressure. The methyl-5-hydroxypentanoate (compound 2) generated from the reaction (with a 99% yield) was oxidized with pyridinium chlorochromate to methyl-5-oxopentanoate (compound 3, yield; 70%). The ¹H NMR spectra of methyl-5-hydroxypentanoate and methyl-5-oxopentanoate were similar to those reported earlier (56). The methyl 5-oxopentanoate (18 mmol) was mixed with para-toluene sulfonic acid (p-TSOH·H₂O) (0.53 mmol) and trimethylorthoformate (90 mmol) in dry methanol (30 ml), and the reaction mixture was stirred overnight at room temperature, neutralized with 2% KOH, and concentrated. The residue was dissolved in diethyl ether (50 ml), washed with brine solution (15 ml), dried over anhydrous sodium sulfate, and concentrated under reduced pressure. Methyl 5,5-dimethoxypentanoate (compound 4, yield; 75%) was obtained after purifying the residue on a silica gel column using hexane/ethyl acetate (3:7, v/v). The methyl ester identified by ¹H NMR (54) and 142 mg of the ester was stirred with a solution of NaOH (4 mmol, 5 equivalent) in 4 ml of tetrahydrofuran/water/methanol (3:2:5, v/v/v) at room temperature. After 2 h, the reaction mixture was acidified with cold 0.1 N HCl and extracted with diethyl ether (4 × 20 ml). The combined organic extracts were dried over sodium sulfate and filtered. The solvent was removed under vacuum to obtain 5-dimethoxypentanoic acid (compound 5, 120 mg; 93%), which was identified by its ¹H NMR characteristics. The 5-dimethoxypentanoic acid (compound 5, 0.53 mmol; 3.1 eq) and 1-palmitoyl-2-hydroxy-sn-glycero-3-phosphocholine (0.17mmol; 1 equivalent) were dried by azeotropic distillation with benzene (5 × 4 ml) under reduced pressure, at room temperature, and then at 300 °C. The flask was flushed with argon, and freshly distilled CHCl₃ (10 ml), dicyclohexyl-carbodiimide (0.51 mmol; 3 eq), and p-(N,N-dimethylamino) pyridine (0.51 mmol; 3 equivalent) were added. The reaction mixture was stirred at room temperature for 48 h, and the solvent was removed under vacuum. The residue was flash-chromatographed on a column (20 mm × 180 mm) of silica gel (40 μm Alltech) with chloroform/methanol (9:1) and chloroform/methanol/water (70:26:4 v/v/v) to obtain 1-palmitoyl-2-(5-diemthoxypentanoyl)-PC (compound 7, 185 mg; 86%). The 1-palmitoyl-2-(5-dimethoxypentanoyl)-PC was dissolved in 12 ml of acetone/water (6:1 v/v), and Amberlyst-15 resin (35 mg) was added to the mixture. The reaction mixture was stirred for 6 h and filtered, the solvent was removed under reduced pressure and the residue was purified by flash column chromatography (silica gel 40 μm Alltech, chloroform/methanol/water 70:26:4 v/v/v) to generate POVPC (compound 8, 25 mg, 77%). The ¹H NMR characteristics of the purified compound in CDCl₃ at 500 MHz were: d 9.69 (s, 1H), 5.15 (m, 1H), 4.38–3.76 (m, 8H), 3.29 (s, 9H), 2.61 (t, 2H), 2.41(m, 2H), 2.27(t, 2H), 1.95 (t, 2H), 1.59 (m, 2H), 1.21 (m, 24H), 0.85 (t, 3H), and are in agreement with those reported by Stremler et al. (33).

Scheme 1.

Measurements of AR Activity

Recombinant human AR was prepared as described (57). Before each experiment, the stored enzyme was reduced by incubating with 0.1 M dithiothreitol at 37 °C for 1 h in 0.1 M Tris, pH 8.0. Excess dithiothreitol was removed by gel filtration on a Sephadex G-25 column (PD-10), pre-equilibrated with nitrogen-saturated 0.1 M potassium phosphate, pH 6.0, containing 1 mM EDTA. To prevent enzyme oxidation, all operations were performed at 4 °C. The reduced enzyme, separated from dithiothreitol, was stored under nitrogen and used within 1–3 h. To ensure that the enzyme was fully reduced, the sensitivity of the enzyme to sorbinil was measured. The activity of the fully reduced enzyme is inhibited by 80% in the presence of 10 μM sorbinil (59).

The catalytic activity of the enzyme was measured at 30 °C in a 1-ml reaction system containing 0.1 M potassium phosphate, pH 6.0, with 0.15 mM NADPH and varying concentrations of POVPC. The purity of POVPC was established by ESI⁺/MS and its concentration was determined by measuring inorganic phosphate as described before (60) and by measuring its choline content using the phospholipid kit obtained from Wako Chemicals. The reaction was monitored by measuring the disappearance of NADPH at 340 nm using a Varian spectrophotometer (Cary Bio 50). One unit of AR is defined as the amount of enzyme required to oxidize 1 μmol of NADPH per min. The control cuvette contained all components of the mixture except the substrate. Initial velocity was measured at eight different concentrations of the substrate, and steady-state kinetic parameters were calculated by fitting the general Michaelis-Menten equation to the data using a nonlinear iterative fitting procedure (44).

Oxidation, Reduction, and Analysis of Phospholipids

Phospholipids dissolved in chloroform (1 mg/ml) were transferred into Teflon coated tubes. The chloroform solution (50 μl) was evaporated, and the phospholipids were allowed to autooxidize as a thin film at room temperature in the dark. All phospholipids were oxidized for 24 h, except PAPA, which was oxidized for 72 h. To test the ability of AR in reducing phospholipid aldehydes, the dry air-oxidized extracts were suspended in 500 μl of 0.1 M potassium phosphate, pH 6.0, and 100 μg of recombinant AR and 75 nmol of NADPH were added to the suspension, and the mixture was incubated for 45 min at 30 °C. Additional aliquots of 75 nmol of NADPH were added, and the reaction was allowed to proceed for 3 h.

Phospholipids were extracted by the Bligh and Dyer (61) procedure. Briefly, methanol and chloroform were added to the aqueous phase to a mixture containing methanol/chloroform/water in a 2:1:0.8 (v/v/v) ratio to a total volume of 950 μl. The solution was vortexed for 1 min, the ratio of the mixture was changed to 1:1:0.9 methanol/chloroform/water. The mixture was then centrifuged for 15 min at 800 × g at 4 °C. The chloroform phase was recovered and diluted with an equal volume of methanol containing 1 N HCl to remove bound cations. The solution was mixed with vigorous vortexing and then centrifuged for 1 min at 13,000 × g. After centrifugation, the chloroform phase was recovered and used for mass spectrometry.

Electrospray ionization mass spectrometry (ESI/MS) was performed using the MicroMass ZMD 2000 mass spectrophotometer (Waters-Micromass, Milford, MA) as described previously (60). The standard electrospray capillary attachment was used throughout. The injection solvent was 2:1 methanol/chloroform (v/v) containing 1.0% acetic acid in positive ionization mode and 2:1 methanol/chloroform (v/v) containing 10 mM ammonium hydroxide in the negative ionization mode. Samples were injected into the spectrometer using a Harvard syringe pump at a flow rate of 20 μl per min. The ESI⁺/MS operating parameters were: capillary voltage 3.38 kV, cone voltage 25 V, extractor voltage 9V, RF lens voltage 0.9 V, source block, and desolvation temperatures 100 and 200 °C, respectively. Nitrogen was used as the nebulizer gas at a flow rate of 3.4 liters/h. Spectra were acquired at a rate of 275 atomic mass units per s over the mass range of 2–1000 atomic mass units and were averaged over a period of 5 min or 100 scans (60).

Cell Culture

COS-7 and THP-1 cells were cultured in Dulbecco’s modified Eagle’s medium or M199 medium, respectively, containing 10% fetal bovine serum and 2% penicillin/streptomycin and grown to 80% confluency. For transfection experiments, COS-7 cells grown to 80% confluency were incubated with Lipofectamine^™ reagent alone or with 1 μg of rat lens AR pcDNA (62) for 4 h in serum-free Dulbecco’s modified Eagle’s medium. After incubation, the medium was replaced with fresh Dulbecco’s modified Eagle’s medium containing 10% fetal bovine serum and antibiotics, and the cells were cultured for an additional 48 h to achieve maximal expression. On the day of the experiment, the culture medium was removed from transfected or non-transfected cells and replaced with Hank’s balanced salt solution (HBSS). To inhibit AR, the cells were treated with sorbinil, tolrestat, or vehicle in HBSS. After 45 min of preincubation at 37 °C with the inhibitor, the cells were incubated with10 μM POVPC for the indicated time, and were then washed and harvested in phosphate-buffered saline. The harvested cells were sedimented by centrifugation at 200 × g for 15 min at 4 °C and frozen at −80 °C or used immediately. Phospholipids were extracted by the Bligh and Dyer procedure (61) and analyzed by ESI/MS. The expression of AR in the cells was examined by Western analysis using rabbit polyclonal anti-AR antibodies raised against human recombinant AR in our laboratory as described previously (46, 63, 64).

RESULTS

Molecular Modeling Studies

Molecular studies were performed to determine whether phospholipid aldehydes could bind to the AR active site. Because a crystal structure of the ternary complex of AR·NADPH·aldehyde is not available, AR complexed with NADP⁺ and glucose 6-phosphate (51) was used as a starting point. The NADP⁺ molecule in the crystal structure was converted to NADPH. For substrate binding we used the 5-carbon aldehyde POVPC. Arachidonoyl side chains of phospholipids are particularly prone to oxidation at C-5, and aldehydes with 5 carbons are some of the most abundant products of phospholipid oxidation (34, 65). Manual docking of POVPC with AR was performed to orient the aldehyde group toward the catalytic residues His¹¹⁰ and Tyr⁴⁸. A restrained simulated annealing conformational search of POVPC in the binding site of AR was performed. The lowest energy conformation is shown in Fig. 1. There were 6 conformations with 10 kJ/mol indicting flexibility and conformational freedom of POVPC binding to AR. The model shows that the aldehyde function of POVPC is located near the nicotinamide ring of NADPH, which could accommodate the proposed (55) reaction mechanism. The sn-2 chain is well accommodated in the deep crevice of the β-barrel, with the sn-1 and sn-2 chains oriented in a roughly V-shaped conformation, straddling the lip of the barrel. The contacts of POVPC with AR within 3 Å were with residues Trp²⁰, Val⁴⁷, Tyr⁴⁸, Trp⁷⁹, His¹¹⁰, Trp¹¹¹, Phe¹²¹, Phe¹²², Trp²¹⁹, Leu³⁰⁰. Hydrophilic interactions between the active site of the enzyme and the phospholipid head group, or sn-2 substituents as well as a potential H-bond interaction between phosphate oxygen and N 1ofTrp²⁰ at a distance of 2.8 Å were identified. It should be noted that several orientations of POVPC relatively close in energy are possible indicating flexibility in binding.

Fig. 1. Molecular modeling of POVPC bound to the active site of AR.

Aldose reductase and NADPH are shown as Connolly surfaces, cyan and purple, respectively, with POVPC shown in a ball and stick structure (A). B, shows the interaction of key active site residues: His¹¹⁰ (blue), Tyr⁴⁸ (yellow), and NADPH (gray) with the POVPC molecule, shown in ball and stick. The Trp²⁰ is also shown (orange) in stick representation as this residue forms a hydrogen bond with POVPC.

AR Catalyzes the Reduction of Several Aldehyde Phospholipids Derived from Lipid Peroxidation

To test the predictions of the model and to determine the range of phospholipid aldehydes reduced by AR, we tested whether AR could catalyze the reduction of aldehydes in mixtures of oxidized phospholipids. As shown in Fig. 2A, air oxidation of PAPC (M + H⁺ = 782.9) resulted in the formation of multiple species with m/z values higher and lower than the parent phospholipid. Similar products have been observed before (26). The major ion at m/z 594.9 was ascribed to POVPC, whereas ions at m/z 620.9 and 624.8 were ascribed to 1-palmitoyl-2-(7-oxo-5-heptenyl)-PC (POHPC) and 1-palmitoyl-2-(5-hydroxy-6-oxo-caproyl)-PC (PHOCPC), respectively. The molecule with m/z 650.6 ion has been identified previously as 1-palmitoyl-2-(5-hydroxy-8-oxo-6-octenoyl)-PC (PHOOPC; 31). The PHOOPC and its keto analog 1-palmitoyl-2-(5-keto-8-oxo-6-octenoyl)-PC (PKOOPC; m/z 648.8) have been shown to be ligands for CD36 and to mediate recognition of oxidized LDL by macrophages (66). High molecular weight ions with m/z values of 828.8, 846.8, 862.8, and 878.8 were also observed. Several of these ions (at 846.8, 862.8, and 878.8) arise from the addition of multiple oxygens (+ 4, 5, and 6, respectively) to PAPC. The species with m/z 828.8 has been identified as 1-palmitoyl-2-(5,6-epoxyisoprostane E₂)-PC (PEIPC;67). Incubation of the oxidized PAPC mixture with recombinant AR and NADPH resulted in the appearance of several new phospholipid species with m/z < 800 (Fig. 2B). An increase in the m/z value of an ion by 2 (because of reduction of aldehydes to alcohols) was used as a diagnostic feature for identifying AR-derived products. Incubation with AR led to a decrease in the intensity of the 594.9 ion with the concurrent increase in intensity at m/z 596.9, suggesting that POVPC is reduced to 1-palmitoyl-2-(5-hydroxyvaleroyl)-PC (PHVPC). Similar, two unit increments at m/z 622.9, 626.7, and 652.9, indicate the formation of alcohols 1-palmitoyl-2-(7-hydroxy-5-heptenyl)-PC (PHHPC), 1-palmitoyl-2-(5,6-dihydroxycaproyl)-PC (PDHCPC), and 1-palmitoyl-2-(5,8-dihydroxyoctenoyl)-PC (PDHOPC) derived from the reduction of POHPC, PHOCPC, and PHOOPC, respectively. A decrease in the intensity at 648.8 because of PKOOPC was also observed, but the changes were small and were not examined in detail. No decrease in the relative intensity of m/z 828.8, 846.8, 862.8, or 878.8 ions was observed, indicating that these phospholipids are not AR substrates. Moreover, there was no change in the intensity at m/z 610.2, because of 1-palmitoyl-2-glutaryl-PC (PGPC). Collectively, these data demonstrate that even in a complex mixture, AR catalyzes the reduction of POVPC, and other oxidized phospholipids containing conjugated unsaturated and hydroxyl aldehydes as well as hydroxyl alkenal side chains at the sn-2 position. As listed in Table I, the greatest extent of reduction was observed with PHOOPC (74%) followed by POVPC (64%), POHPC (44%), and POHCPC (24%).

Fig. 2. Aldose reductase catalyzes the reduction of aldehydes generated from oxidation of PAPC.

Aliquots of PAPC (50 μg) were oxidized in air for 24 h and were either left untreated (A) or resuspended in 0.1 m potassium phosphate buffer, pH 6.0 containing 0.15 mm NADPH and 100 μg of reduced, recombinant AR for 3 h at 30 °C (B). The phospholipids were extracted in chloroform/methanol/water and injected into the electrospray using 2:1 methanol/chloroform containing 1.0% acetic acid as the flow injection solvent. The parent PAPC formed a well resolved positive ion with m/z 782.9 as indicated, which was oxidized to POVPC (m/z 594.9), POHPC (m/z 620.9), PHOCPC (m/z 624.8), and PHOOPC (m/z 650.6). High molecular weight oxidation products formed by the addition of oxygen are also evident. Incubation with AR led to an increase in intensity at m/z 596.9 (C), 622.9 (D), 626.7 (D), and 652.6 (E), which represent the reduction products of POVPC, POHPC, PHOCPC, and PHOOPC, respectively.

Table I. Reduction of phospholipid aldehydes by aldose reductase.

PAPC, p-PAPC, PAPA, PAPG, and 1-alkyl PAPC (50 μg) were oxidized in air and then incubated with AR and NADPH as described under “Experimental Procedures.” After incubation, the samples were extracted and analyzed by ESI⁺/MS. Percent reduction was calculated from the decrease in the intensity of the parent aldehyde and the corresponding increase in the intensity of alcohol ion (±S.D.).

sn-2 chain	Parent phospholipid
	PAPC	p-PAPC	PAPA	PAPG	1-alkyl PAPC
	% reduction
5-oxo-Valeroyl	63.6 ± 14.1	50.5 ± 8.5	66.4 ± 6.6	96.2 ± 2.5	65.5 ± 14.5
7-oxo-trans-5-Heptenyl	43.8 ± 1.1	38.3 ± 3.6	70.5 ± 0.5	82.5 ± 4.2	48.6 ± 1.4
5-hydroxy-6-oxo-Caproyl	24.3 ± 1.0		22.6 ± 13.9
5-hydroxy-8-oxo-6-Octenoyl	74.0 ± 4.1	66.3 ± 6.5	62.9 ± 1.8	69.3 ± 4.8	42.3 ± 2.7

To examine whether the nature of the sn-1 linkage affects AR-catalyzed reduction, oxidation products derived from phosphatidylcholine containing either sn-1 plasmenyl or alkyl linkages were tested. Both of these phospholipids contained palmitoyl (sn-1) and arachidonoyl (sn-2) side chains. Upon air oxidation, these phospholipids generated oxidation products similar to those obtained with the acyl-linked phospholipids. Oxidation of plasmenyl PAPC generated predominant ions at m/z 544.8 and 572.9 (Fig. 3A). The molecule with m/z 544.8 has been shown to be caused by a sn-1 lyso generated from the oxidation of the vinyl ether bond at the sn-1 position resulting in the loss of the entire alkenyl chain (68). The ion with m/z 572.9 was assigned to 1-formyl-2-arachidonoyl-PC, which arises from Hoch rearrangement (migration of the vinyl group) followed by Criegee rearrangement (65). Additional ions at m/z 578.9, 604.9, 608.9, and 634.9 were analogous to those generated from the sn-2 oxidation of acyl PAPC: pPOVPC, pPOHPC, pPHOCPC, and pPHOOPC. Incubation of oxidized phospholipids with AR resulted in the appearance of additional species with increments of two m/z values (Fig. 3). Aldose reductase-derived products were observed at m/z 580.9 (pPHVPC), 606.9 (pPHHPC), 610.9 (pPDHCPC), and 636.9 (pPDHOPC), but the extent of reduction of these aldehydes was variable (Fig. 3). As shown in Table I, maximal reduction was observed with pPHOOPC (66%), followed by pPOVPC (50%) and pPOHPC (40%). The change in intensity caused by the reduction of the 5-hydroxy-6-oxohexanoyl-PC (m/z 604.9) was small and could not be reliably quantified. Significantly, 1-formyl-2-arachidonoyl PC (m/z 572) was not reduced by AR.

Fig. 3. Aldose reductase catalyzes the reduction of aldehydes generated from in vitro oxidation of plasmenyl PAPC.

Positive electrospray mass spectra of air oxidized p-PAPC were acquired before (A) and after incubation with AR (B). Aliquots of pPAPC (50 μg) were oxidized in air for 24 h and were either left untreated or resuspended in 0.1 m potassium phosphate buffer, pH 6.0 containing 0.15 mm NADPH and AR for 3 h at 30 °C. The phospholipids were extracted in chloroform/methanol/water and injected into the electrospray using 2:1 methanol/chloroform containing 1.0% acetic acid as the flow injection solvent. The parent pPAPC formed a well resolved positive ion with m/z 766.9, which was oxidized to pPOVPC (m/z 578.9), pPOHPC (m/z 604.9), pPHOCPC (m/z 608.9), and pPHOOPC (m/z 634.9). High molecular weight oxidation products formed by the addition of oxygen are also evident as well as the sn-1 lyso (m/z 544.8) and the formation of a formyl group at the sn-1 position (FAPC; m/z 572.9). Incubation with AR led to an increase in intensity at m/z 580.9 (C), 606.9 (D), 610.9 (D), and 636.9 (E), which represent the reduction products of pPOVPC, pPOHPC, pPHOCPC, and pPHOOPC, respectively. No changes were observed in the intensity of the m/z 572.9 ion (FAPC) upon incubation with AR.

Products obtained from oxidation of 1-alkyl-palmitoyl-2-arachidonyl-PC were also reduced by AR. In this case, the prominent analogs of the oxovaleroyl, oxoheptenoyl, and 5-hydroxy-8-oxo-6-ocentoyl (m/z 580.7, 606.7, and 636.7) derivatives were observed, and they were variably reduced by AR (Table I and Supplemental Fig. 1). Reduction of phospholipid aldehydes was found to be relatively insensitive to the nature of the charged head group (sn-3 substitution). The alkyl, alkenyl, hydroxyl, and hydroxyalkenal aldehydes generated from the oxidation of PAPG (Supplemental Fig. 2) and PAPA (Supplemental Fig. 3) were reduced upon incubation with AR (Table I). The efficacy of AR in reducing phosphatidylethanolamine could not be tested because of inconsistent generation of oxidation products and the formation of Schiff base products derived from inter- or intra-molecular reaction between aldehydes and the head group amine as has been reported before (69). Oxidation of PAPG generated predominant ions at m/z 581.7, 607.8, and 637.9 corresponding to phospholipids containing 5-oxovaleroyl, 7-oxo-heptenyl and 5-hydroxy-8-oxo-octenoyl side chains at the sn-2 position, respectively. Incubation with AR led to an increase in intensity at m/z 583.9, 609.8, and 639.9, which are 2 units higher than the parent phospholipid (Supplemental Fig. 2). Similar ions were observed upon the oxidation of PAPA, which generated predominant ions with m/z 507.5, 533.5, 537.4, and 563.6 corresponding to phospholipids containing 5-oxovaleroyl, 7-oxo-heptenyl, 5-hydroxy-6-oxocaproyl and 5-hydroxy-8-oxo-octenoyl side chains at the sn-2 position, respectively. Incubation with AR led to a 2 m/z shift in the ions with m/z 509.6, 535.7, 539.8, and 565.6 (Supplemental Fig. 3). Together, these data demonstrate that AR catalyzes the reduction of a wide range of phospholipid aldehydes.

Catalytic Efficiency of AR in Reducing POVPC

To establish the efficiency of AR in catalyzing the reduction of phospholipid aldehydes, we determined the steady-state kinetic parameters for AR with POVPC. For this, POVPC was synthesized as described under “Experimental Procedures.” Upon ESI⁺/MS, reagent POVPC formed a prominent ion at m/z 594.9. Incubation with AR led to an increase in the m/z value of the phospholipid from 594.9 to 596.9, consistent with reduction of the ω-aldehyde to an alcohol (Fig. 4). Aldose reductase-catalyzed the reduction of POVPC with a K_m = 9.7 ± 1.4 μM and a k_cat = 36.2 ± 2.7 min⁻¹, corresponding to a catalytic efficiency of 3747 min⁻¹ mM⁻¹. Thus AR reduces POVPC more efficiently than other lipid peroxidation-derived aldehydes such as acrolein, hexenal, and 4-hydroxy-trans-2-nonenal (44).

Fig. 4. Aldose reductase catalyzes the reduction of POVPC.

Recombinant human AR (100 μg) was incubated with 0.15 mm NADPH and 50 μg POVPC in 0.1 m potassium phosphate, pH 6.0 for 3 h at 30 °C. The samples were extracted by the Bligh and Dyer procedure (59) and injected into the electrospray using a mixture of 2:1 methanol/chloroform and 1.0% acetic acid as the running solvent at a flow rate of 20 μl/min. Note: reagent POVPC corresponds to a positive ion with m/z 594.9 (left panel), which is increased by 2 (596.9) upon reduction by AR (right panel), indicating the formation of PHVPC as shown in the inset. Steady-state kinetic parameters for the POVPC reduction calculated from initial velocity measurements are shown.

Cellular Metabolism of POVPC

To determine cellular metabolism, two phenotypically unrelated cell lines (COS-7 and THP-1) were used. These cells express high levels of AR protein and display high AR activity (data not shown). The ESI⁺/MS of the phospholipid extracts of COS-7 cells incubated with 10 μM POVPC for 60 min showed distinct molecular ions at m/z 594.9 and 596.9. An additional ion at m/z 496.9 was also observed (Fig. 5B). None of these ions were present in extracts from cells that were not treated with POVPC (Fig. 5A). These ions were assigned to POVPC (m/z 594.9), PHVPC (m/z 596.9), and the sn-2 lyso 1-palmitoyl-PC (m/z 496.9). Significantly, no acid metabolite of POVPC with an expected m/z of 610.9 was observed. Similarly, exposure to POVPC led to the appearance of m/z 594.4 and 596.4 ions in the phospholipid extracts from the THP-1 cells (Fig. 6B). In COS-7 cells the intensity of the m/z 596.9 ion represented 51 ± 6% (n = 3) of the parent aldehyde. The corresponding value for THP-1 cells was 37 ± 1% (n = 3). Thus, a significant portion of POVPC incorporated in both these cell type is reduced to PHVPC.

Fig. 5. Cellular metabolism of POVPC in COS-7 cells.

Cells were cultured to 80% confluency, rinsed with HBSS, and incubated with 10 μm POVPC, without (B) or with (C) 1 μm tolrestat for 60 min at 37 °C in HBSS. The tolrestat-treated cells were preincubated with the drug for 45 min before adding POVPC. At the end of the incubation, the medium was aspirated, and the cells were rinsed with phosphate-buffered saline and extracted in chloroform/methanol/water, and their ESI⁺ mass spectra were recorded. A shows the spectrum of the phospholipid extract prepared from COS-7 cells that were not treated with POVPC.

Fig. 6. Metabolism of POVPC in THP-1 cells.

Subconfluent THP-1 cells were removed from the culture medium, resuspended in HBSS and were left untreated (A) or incubated with 1 μm tolrestat (B) for 45 min at 37 °C. Subsequently, 10 μm POVPC was added to the incubation medium, and the incubation was continued for an additional 60 min at 37 °C. At the end of the incubation, the medium was aspirated, and the phospholipids were extracted and examined by ESI⁺/MS.

To probe the role of AR, POVPC metabolism in COS-7 and THP-1 cells was examined in the presence of two structurally unrelated AR inhibitors, tolrestat and sorbinil. Treatment of COS-7 cells with 1 μM tolrestat led to a 49 ± 9% (n = 7) reduction in the intensity of the 596.9 ion (Fig. 5), whereas in the THP-1 cells a 51 ± 2(n = 4) % decrease in the intensity was observed (Fig. 6). Treatment with 50 μM tolrestat resulted in only slightly higher inhibition; PHVPC formation was inhibited by 62 ± 5% (n = 3) in COS-7 cells. Similarly 10 μM sorbinil inhibited PHVPC formation by 34 ± 3% (n = 3) in COS-7 cells and 30 ± 2 (n = 3) %in THP-1 cells. Another AR inhibitor zopolrestat inhibited PHVPC reduction by 29 ± 3% (n = 3) at 1 μM. Collectively, these data suggest that in both COS-7 and THP-1 cells, AR-catalyzed reduction is a significant metabolic fate of POVPC.

To determine whether POVPC and PHVPC in POVPC-exposed cells are localized to the membrane, the cell membranes of POVPC-treated COS-7 cells were isolated by ultracentrifugation. As with the total cell extract, predominant peaks at m/z 594.9 and 596.9 were observed in the membrane fraction. Neither ion was present in the cytosolic extract (data not shown), indicating that addition of exogenous POVPC results in the incorporation of the aldehyde into the cell membrane and that upon reduction POVPC remains associated with the membrane.

POVPC Metabolism in AR-overexpressing COS-7 Cells

To examine how an increase in AR would affect POVPC metabolism, COS-7 cells were transfected with AR cDNA. The transfected cells displayed an 8.6 ± 1.0-fold increase in AR protein, as compared with cells incubated with Lipofectamine alone (Fig. 7). Incubation of Lipofectamine-treated COS-7 cells (wild type) with POVPC resulted in the appearance of POVPC (m/z 594.4) and PHVPC (m/z 596.3) in the phospholipid extract. In these cells, the intensity of PHVPC (m/z 596.3) peak represented 44 ± 7% (n = 4) of POVPC (Fig. 7). Transfection of COS-7 cells with AR cDNA (AR⁺⁺), significantly increased PHVPC formation. In the AR⁺⁺ cells, the intensity of the PHVPC ion (m/z 596.5) represented 85 ± 2% (n = 3) of POVPC (Fig. 7). Treatment with 1 μM tolrestat led to a 34 ± 4% (n = 3) inhibition of PHVPC formation (Fig. 7) in AR⁺⁺ cells, whereas 88 ± 6% inhibition was observed with 50 μM tolrestat. Greater inhibition of PHVPC formation at higher concentration of tolrestat is consistent with a predominant role of AR in POVPC metabolism in AR⁺⁺ cells.

Fig. 7. Metabolism of POVPC in COS-7 cells overexpressing AR.

Subconfluent COS-7 cells were treated either with Lipofectamine (A) or with Lipofectamine + 1 μg of AR cDNA (B and C) and cultured for 48 h. The cells were then rinsed with HBSS and used either for Western analysis or for metabolism experiments. Inset of B shows a representative immunoblot of wild-type (Lipofectamine-treated) and AR-transfected (AR⁺⁺) COS-7 cells developed with anti-AR antibodies. For the metabolism studies, both the wild-type (A) and AR⁺⁺ (B and C) cells were incubated with 10 μm POVPC, without (A and B) or with (C) 1 μm tolrestat for 60 min at 37 °C. The tolrestat-treated cells were preincubated with the drug for 45 min before adding POVPC. At the end of the incubation, the incubation medium was removed, and the cells were rinsed with phosphate-buffered saline and extracted and analyzed by ESI+/MS.

DISCUSSION

This study demonstrates that several structurally different phospholipid aldehydes are efficiently reduced by the enzyme AR, suggesting that AR may be an important component of mechanisms that remove and detoxify these aldehydes when they are generated in oxidized lipids. Aldose reductase was first identified as an enzyme that reduces aldose sugars to polyols (70, 71), and several studies since then have suggested that increased reduction of glucose by AR contributes to the development of secondary diabetic complications (70, 71). However, more recent work indicates that in addition to sugars, AR also reduces lipid peroxidation-derived aldehydes as well as their glutathione conjugates (43–48,72). The role of AR in metabolizing products of lipid peroxidation or their metabolites is further supported by the observations that the expression of AR is enhanced under conditions of oxidative stress both in cell culture systems (46, 73) and in vivo (46, 62, 74), and that inhibition of AR increases the accumulation of lipid peroxidation products during inflammation and ischemia (64, 74). Additionally, it has been shown that cells resistant to oxidative stress express higher levels of AR than the corresponding sensitive cells and that inhibition of AR increases the sensitivity of these cells to cytotoxic aldehydes (75). Taken together, these observations provide firm support to the idea that AR is an antioxidant enzyme that removes and detoxifies products of lipid peroxidation. This view is further reinforced by the observations presented here. The observation that AR reduces phospholipid aldehydes adds a new facet to the antioxidant role of this enzyme and provides novel avenues for understanding and managing the pathological consequences of oxidative stress.

Our studies show that AR catalyzes the reduction of a wide range of phospholipid aldehydes. The enzyme reduced phospholipids with acyl, alkyl, and alkenyl linkages at the sn-1 position and with different head groups at the sn-3 position, suggesting that it recognizes a broad range of phospholipid aldehydes. As indicated by our molecular modeling studies, this promiscuity may be derived in part from the interaction of the aldehydic sn-2 chain with hydrophobic residues flanking both sides of the active site of the enzyme. The specific recognition of sn-2 aldehyde is supported by the observation that sn-1 formyl-PC was not reduced by AR, and the enzyme seems to prefer a short methylene chain preceding the ω-aldehyde group. Previous studies indicate that short C5-C9 aldehydes bind efficiently to the large active site of AR lined by hydrophobic residues and further that the C4-C5 methylene group of these aldehydes may be in contact with Trp²⁰ (44). The W20F mutation, however, decreases the catalytic efficiency of AR with aldehyde-glutathione conjugates, but does not affect the reduction of HNE and other free aldehydes (48). Hence the potential interaction between Trp²⁰ and the sn-2 carbonyl, suggested by the modeling studies, indicates that there may be specific interactions anchoring C5-C9 phospholipid aldehydes at the active site, without a loss of binding efficiency caused by an increase in the bulk of the molecule as compared with simple straight-chained aldehydes.

The catalytic efficiency of AR with pentanal is 2546 M⁻¹·s⁻¹ (44), which is less than that with POVPC (3747 M⁻¹·s⁻¹), indicating a significant contribution of the backbone or side chains of the phospholipid to the stabilization of POVPC at the AR active site. If there were no energetically significant interactions (other than those between the active site and the sn-2 aldehyde), the catalytic efficiency of AR with POVPC should be either equal to pentanal, or more likely, much less than pentanal, since an increase in the aldehyde chain or additional substitutions decrease k_cat/K_m (44). Hence the observation that POVPC is a better substrate than pentanal suggests that AR specifically recognizes the sn-2 aldehyde substituents of oxidized phospholipids and that the active site is less sensitive to the sn-3 head group, which allows phospholipid aldehydes of different structures to be reduced by the enzyme.

Exposure of THP-1 and COS-7 cells to POVPC resulted in the appearance of two major metabolites, PHVPC and the sn-2 lyso PC. This is in contrast to the metabolism of free aldehydes derived from lipid peroxidation, which are metabolized via several different pathways (45, 46). Unsaturated aldehydes, such as 4-hydroxy-trans-2-nonenal, are reduced, oxidized, and conjugated (76). Because of lack of unsaturation, it is unlikely that POVPC forms glutathione conjugates, however, lack of oxidation of POVPC is surprising because aldehyde dehydrogenase-catalyzed oxidation is the primary metabolic fate of most saturated aldehydes (76). Therefore, in the absence of multiple subsidiary metabolic pathways involving oxidation and conjugation, it appears that AR plays a more central role in phospholipid aldehyde metabolism than in the metabolism of free unsaturated aldehydes.

The role of AR in the cellular metabolism of phospholipid aldehydes is supported by our observations that treatment with AR inhibitors significantly prevents POVPC reduction in both COS-7 and THP-1 cells and transfection with AR cDNA increases POVPC reduction in COS-7 cells. Nevertheless, even though AR inhibitors, in concentrations (1–10 μM) at which they inhibit AR and AR-mediated signaling in other cells, significantly decreased POVPC reduction, the inhibition was not complete; even when higher (50 μM) concentrations were used. Persistent reduction of POVPC in the presence of AR inhibitors may be because of incomplete uptake or inhibition of AR inhibitors in these cells or because of the presence of other unidentified reductases such as the alcohol dehydrogenases (76) and hydroxysteroid dehydrogenases (77), which have been shown to reduce hydrophobic aldehydes such as HNE and may also be active with POVPC. Alternatively, because in addition to reduction, POVPC and PHVPC undergo other metabolic transformations (e.g. via phospholipases), their cellular concentration and relative abundance are likely to be a complex function of multiple metabolic transformations. Thus even if the formation of PHVPC is diminished drastically, PHVPC will, in time, accumulate to levels determined by the affinity (K_½) of the processes involved in its subsequent metabolism. Similarly, if inhibition of AR leads to a compensatory increase in POVPC hydrolysis, the ratio of POVPC:PHVPC will appear high despite minimal reduction. At present these metabolic transformations are difficult to quantify since lyso-PC, the product of POVPC or PHVPC hydrolysis, is itself a transient intermediate likely to be re-esterified back into diacyl phospholipids. Clearly further experiments are required to quantify the contribution of AR and other biochemical pathways to the overall metabolism of POVPC and related aldehydes.

Aldehydes are major products of oxidized phospholipids. In agreement with previous work (26, 31, 66), we found that oxidation of PAPC resulted in the generation of several classes of aldehydes including unsubstituted (POVPC) and 2-hydroxy-saturated (PHOVPC) aldehydes, unsaturated aldehydes (PO-HPC), γ-hydroxy-(PHOOPC), and γ-oxo-(PKOOPC) aldehydes. These aldehydes have distinct biological activities. POVPC stimulates smooth muscle cell growth (25) and increases monocyte binding to endothelial cells (26). It also activates PPAR-α thereby inducing the synthesis of monocyte chemotactic protein-1 (MCP-1) and interleukin-8 (IL-8) in monocytes (78). Hence, the observation that AR reduces POVPC suggests that this enzyme may be an important regulator of the inflammatory effects of this and related aldehydes, and that AR could regulate the ability of oxidized phospholipids to impact signaling mechanisms or gene transcription events. Although additional experiments are required to elucidate the role of AR in regulating the metabolism and the bioactivity of individual phospholipids, the observation that chemical reduction by sodium borohydride abolishes the ability of POVPC to promote monocytes binding to endothelial cells (30) suggests that reduction by AR may be an important mechanism for decreasing the biological activity of phospholipid aldehydes. For instance, AR-mediated metabolism of phospholipid aldehydes, by affecting their ability to stimulate PPAR-α or PPAR-γ, could interfere with adaptive metabolic responses to oxidative stress. These responses may be further modified by the ability of AR to regulate TNF-α signaling and NF-κB activation (63). Furthermore, because the structures of the aldehyde (POVPC, PHOVPC) and the non-aldehyde (PGPC, PEIPC) products of phospholipid oxidation are similar, it is possible that the non-aldehyde phospholipids (epoxides and isoprostanes) could be competitive inhibitors of AR and interfere with AR-dependent processes, which in addition to reduction of lipid-derived aldehydes include glucose metabolism (70, 71), biogenic amine detoxification (79), and osmoregulation (80).

In summary, the involvement of AR in metabolizing aldehydes generated during phospholipid oxidation is suggested by multiple lines of evidence showing that: (a) AR is an efficient catalyst for reducing a variety of phospholipid aldehydes; (b) the active site of the enzyme is compatible with phospholipid binding; (c) reduction is a significant fate of phospholipid aldehydes; and (d) at least in the cell types tested, this reduction is catalyzed by AR. Hence, AR may be a significant determinant of the immunological properties of the phospholipid aldehydes and their contribution to atherogenesis and inflammation.

Abbreviations

The abbreviations used are:

LDL

low density lipoprotein

AR

aldose reductase

PAPC

1-palmitoyl-2-arachidonoyl-sn-glycero-3-phosphocholine

alkyl-PAPC

1-O-hexadecyl-2-arachidonoyl-PC

p-PAPC

1-O-hexadec-1′-enyl-2-arachidonoyl-PC

POVPC

y1-palmitoyl-2-(5-oxovaleryl)-sn-glycero-3-phosphocholine

POHPC

1-palmitoyl-2-(7-oxo-5-heptenyl)-PC

PHOCPC

1-palmitoyl-2-(5-hydroxy-6-oxo-caproyl)-PC

PHOOPC

1-palmitoyl-2-(5-hydroxy-8-oxo-6-octenoyl)-PC

PKOOPC

1-palmitoyl-2-(5-keto-8-oxo-6-octenoyl)-PC

PEIPC

1-palmitoyl-2-(5,6-epoxyisoprostane E₂)-PC

PHVPC

1-palmitoyl-2-(5-hydroxyvaleroyl)-PC

PHHPC

1-palmitoyl-2-(7-hydroxy-5-heptenyl)-PC

PDHCPC

1-palmitoyl-2-(5,6-dihydroxycaproyl)-PC

PD-HOPC

1-palmitoyl-2-(5,8-dihydroxyoctenoyl)-PC

PGPC

1-palmitoyl-2-glutaroyl-PC

PAPG

1-palmitoyl-2-arachidonoyl-sn-glycero-3-phosphoglycerol

PAPA

1-palmitoyl-2-arachidonoyl-sn-glycero-3-phosphatidic acid

ESI/MS

electrospray ionization mass spectrometry

PAF

platelet-activating factor

PPAR

peroxisome proliferator-activated receptor

PDB

Protein Data Bank