If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
To whom correspondence should be addressed: Center for Cardiovascular Diagnostics and Prevention, Cleveland Clinic Foundation, 9500 Euclid Ave., NE-10, Cleveland, OH 44195. Tel.: 216-445-9763; Fax: 216-636-0392
* This work was supported by National Institutes of Health Grants HL70621, P01 HL076491, and P01 HL077107 (to S. L. H.) and HL53315, EY016813, and GM21249 (to R. G. S.). 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. 1 Both authors contributed equally to this work.
An essential feature of the innate immune system is maintaining cellular homeostasis by identifying and removing senescent and apoptotic cells and modified lipoproteins. Identification is achieved through the recognition of molecular patterns, including structurally distinct oxidized phospholipids, on target cells by macrophage receptors. Both the structural nature of the molecular patterns recognized and their orientation within membranes has remained elusive. We recently described the membrane conformation of an endogenous oxidized phospholipid ligand for macrophage scavenger receptor CD36, where the truncated oxidized sn-2 fatty acid moiety protrudes into the aqueous phase, rendering it accessible for recognition. Herein we examine the generality of this conformational motif for peroxidized glycerophospholipids within membranes. Our data reveal that the addition of a polar oxygen atom on numerous peroxidized fatty acids reorients the acyl chain, whereby it no longer remains buried within the membrane interior but rather protrudes into the aqueous compartment. Moreover, we show that neither a conformational change in the head group relative to the membrane surface nor the presence of a polar head group is essential for CD36 recognition of free oxidized phospholipid ligands within membranes. Rather, our results suggest the following global phenomenon. As cellular membranes undergo lipid peroxidation, such as during senescence or apoptosis, previously hydrophobic portions of fatty acids will move from the interior of the lipid bilayer to the aqueous exterior. This enables physical contact between pattern recognition receptor and molecular pattern ligand. Cell membranes thus “grow whiskers” as phospholipids undergo peroxidation, and many of their oxidized fatty acids protrude at the surface.
The fluid mosaic model of membrane architecture, first proposed by Singer and Nicolson (
), is a cornerstone of cell biology. An integral feature of this model is the macromolecular assembly of amphipathic phospholipids into a bilayer structure, with polar head groups directed toward the aqueous phase and hydrophobic aliphatic fatty acid chains of glycerophospholipids extending toward the membrane interior. The bilayer structure posited explains how a cell membrane maintains a critical barrier function while it simultaneously facilitates rapid lateral diffusion of proteins and lipids within the planar membrane surface.
Some biological processes, however, are not compatible with a classic phospholipid orientation as described for a lamellar (bilayer) phase membrane structure and instead involve alternative phospholipid conformations. For example, membrane fusion of two vesicles must be accompanied by a transient non-lamellar phase orientation of the phospholipids at the site of membrane fusion (
). Perhaps less obvious is the recognition of a phospholipid ligand within a membrane bilayer of one cell by an alternative cell's receptor. It is hard to envision how a macrophage pattern recognition receptor like CD36 can identify senescent or apoptotic cells through the presence of low abundance structurally specific oxidized phospholipid ligands interspersed within a target membrane bilayer, particularly if the high affinity motif on the lipid resides within a buried acyl chain.
The ability of macrophages and other phagocytes to specifically recognize modified lipids on the surface of cell membranes and lipoprotein particles is of crucial importance in the innate immune system (
). CD36 also recognizes and binds to oxidized low density lipoprotein via oxidized phospholipids, leading to cholesterol accumulation and macrophage foam cell formation, an early histological marker of atherosclerosis development (
ligands recognized by macrophage pattern recognition receptors. Although there is general consensus that a significant portion of ligand binding activity on oxidized lipoproteins and senescent or apoptotic cells resides in extractable lipids (
), many key questions remain. For example, based upon studies with model oxPL protein or peptide adducts generated in vitro, a role for the choline head group of glycerophospholipids when tethered to proteins has been reported (
). However, within a membrane bilayer, the site(s) of recognition on oxPL is less clear. It has been hypothesized that a conformational alteration in the choline polar head group of phosphatidylcholine concurrent with phospholipid oxidation may occur, unmasking a cryptic epitope that enables recognition (
). Studies with structurally defined pure oxPCCD36 obtained by unambiguous syntheses show that they function as high affinity ligands for CD36, promoting specific binding and phagocytosis of target model vesicles, cellular membranes, and oxidized lipoproteins harboring even only trace levels of the species (
). A CD36 binding moiety on the oxidatively truncated sn-2 acyl groups of oxidized choline glycerophospholipids, namely a terminal γ-hydroxy(or oxo)-α,β-unsaturated carbonyl (oxPCCD36), was delineated (
). Further studies demonstrated both the presence and enrichment of oxPCCD36 species in atherosclerotic plaque tissue, low density lipoproteins recovered from aortic tissues, and senescent cell membranes (
). Finally, recent subsequent studies have also shown that CD36 binds to oxidized phosphatidylserine (oxPSCD36) lipids that are structural and chemical analogues of oxPCCD36 and that oxPSCD36 promotes macrophage recognition and engulfment of apoptotic cells (
To gain further insight into how a pattern recognition receptor such as the scavenger receptor CD36 might recognize the high affinity binding motif at the terminal end of an oxidized fatty acid on oxPL within a membrane, we recently examined the conformation of a prototypic high affinity ligand, 1-palmitoyl-2-(5-keto-6-octene-dioyl)phosphatidylcholine (KOdiA-PC) within a membrane (
). We probed the conformation of the oxPC near the hydrophobic-hydrophilic interface within membrane bilayers by determining multiple critical internuclear distances using nuclear Overhauser enhancement (NOE) spectroscopy. Remarkably, the truncated oxidized sn-2 fatty acid chain of KOdiA-PC within membranes was shown to protrude into the aqueous phase, thus enabling direct physical access to the cell surface macrophage CD36 receptor. In the present studies, we hypothesized that this unusual conformation for a phospholipid within a membrane may not be unique to KOdiA-PC but rather might represent a more global phenomenon of oxidized phospholipids, enabling physical contact between pattern recognition receptor and molecular pattern ligand. Herein we utilize NOE spectroscopy to directly examine the conformation of multiple structurally distinct phospholipid peroxidation products in model membranes. By determining critical internuclear distances within oxPC and native (nonoxidized) PC in perdeuterated membrane bilayers, our results reveal that no significant conformational alterations occur in the head group region during oxidation; rather, for many oxPL examined, a dramatic reorienting of the more polar oxidized fatty acid chain occurs such that it protrudes into the aqueous compartment. Our results are thus consistent with a global model in which peroxidized target membranes signal and engage scanning macrophages through protrusion of diverse sn-2 oxidized lipid fatty acid chains or “whiskers” at the cell surface.
Materials—KOdiA-PC was purchased from Cayman Chemical (Ann Arbor, MI). 1,2-Dimyristoyl-d54-sn-glycerol-3-phosphocholine-1,1,2,2-d4-N,N,N-trimethyl-d9 ((d67)-DMPC), 1-O-1′-(Z)-octadecenyl-2-oleoyl-sn-glycero-3-phosphocholine, and most phospholipids were obtained from Avanti Polar Lipids (Alabaster, AL) unless otherwise indicated. In an effort to remove trace levels of oxidized phosphatidylethanolamine (oxPE) species present in even the best commercial PE sources, the oxidant-sensitive PE species 1-palmitoyl-2-linoleoyl-sn-glycero-3-phosphoethanolamine (PLPE) and 1-palmitoyl-2-arachidonoyl-sn-glycero-3-phosphoethanolamine (PAPE) were purified by preparative HPLC and stored under an atmosphere of argon within amber vials at -80 °C until use. 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate (DiI) was obtained from Molecular Probes, Inc. (Eugene, OR). Anti-mouse CD36 antibody was made as described previously (
). All other chemicals of the highest quality available were purchased from either Sigma or Fisher, unless otherwise specified. Human myeloperoxidase was isolated and quantified as previously described (
Synthesis of Phospholipids—Total syntheses of the γ-hydroxy-α,β-unsaturated aldehydic phospholipids, the 4-keto-(2E)-octenedioic acid, and 4-keto-(2E)-heptenoic acid esters of 1-hexadecanoyl lysophosphatidylethanolamine (KOdiA-PE and KHdiA-PE, respectively), were performed as described elsewhere (
). OV-PC, ON-PC, and 13-HODE-PC, the 5-oxovaleric acid, 9-oxononanoic acid, and (13S)-hydroxy-(9Z,11E)-octadeca-9,11-dienoic acid esters of 2-lyso-PC (Table 1), were synthesized and isolated as previously described (
Cell Culture, Plasmids, and Transfections—Human chronic myelogenous leukemia cells (K562), and primary mouse peritoneal macrophages were cultured in RPMI 1640 supplemented with 10% fetal bovine serum and penicillin/streptomycin antibiotics. African green monkey kidney fibroblasts (COS-7) were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and penicillin/streptomycin antibiotics.
An open reading frame encoding wild-type CD36 protein was subcloned into pCGCG bicistronic expression vector containing GFP under translational control of a picornavirus internal ribosome entry site as previously described (
). Use of this construct results in GFP expression in a constant molar ratio with intact native human CD36 within transfected cells, permitting cellular mean fluorescence to be used as a quantitative measure proportional to CD36 expression. This plasmid is referred to as CD36,GFP and results in co-expression of both full-length CD36 and GFP. Cells were transfected with 20 μg of DNA in media made to 150 mm NaCl supplemented with 10 mm HEPES, pH 7.4, using a Bio-Rad electroporator (GenePulser™) at 950 microfarads and 230 V (K562) or 200 V (COS-7). Time constants were typically in the range of 44-48 ms.
Vesicle Preparations for NMR Experiments—Appropriate amounts of phospholipids were initially dissolved in organic solvent (e.g. CDCl3), mixed, and dried with N2 prior to exhaustive drying under vacuum overnight. To prepare small unilamellar vesicles (SUV), lipids were initially fully hydrated by the addition of buffer (20 mm phosphate buffer, 100 mm NaCl in D2O, pD 7.4) in argon-purged sealed vials above the phase transition temperature. Following vortexing to disperse the hydrated lipids, SUV were generated by extrusion through 0.4-μm (6 times) and then 0.1-μm polycarbonate filters (11 times) using an Avanti Mini-Extruder Set (Avanti Polar Lipids, Alabaster, AL). Throughout all experiments, freshly prepared vesicles were maintained well above the phase transition temperature to avoid vesicle fusion and under both argon atmosphere and protection from light to minimize adventitious oxidation. The hydrodynamic radius of SUV preparations was determined with a DynaPro-801 dynamic light scattering instrument with MicroCell attachment and Dynamics 4.0 software (Protein Solutions Inc., Charlottesville, VA). The performance of the instrument was verified both with polystyrene bead standards and a 2 mg/ml aqueous solution of bovine serum albumin prior to sample measurement. Mean liposome diameter of SUV composed of either 100 mol % DMPC or 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) was 100 ± 5 or 80 ± 2 nm, respectively. Mean liposome diameters of SUV composed of 20 mol % oxPC (either KOdiA-PC, OV-PC, ON-PC, or 13-HODE-PC)/80 mol % DMPC were similar (65-80 ± 5 nm).
Vesicle Preparation and Modification for CD36 Binding Assays—The preparation of small unilamellar vesicles composed of POPC containing low a molar percentage of either native or oxidized ethanolamine glycerophospholipids (PLPE, PAPE, or 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine (POPE)), was similar as described for NMR analysis above except in H2O rather than D2O. For a typical experiment, the relatively nonoxidizable phospholipid POPC was used as carrier, and the specified molecular species was incorporated as a minor component (20 mol % or less as indicated). For direct binding experiments, 1 mol % DiI was dried down together with the indicated phospholipids and incorporated into lipid vesicles by extrusion. CD36 ligands were also formed by oxidation of PAPE- or PLPE-containing vesicles (0.2 mg of lipid/ml) by exposure to myeloperoxidase (30 nm), an H2O2-generating system composed of glucose and glucose oxidase (100 ng/ml) (constant H2O2 flux of 0.8 μm/min), and NaNO2 (0.5 mm) at 37 °C for 8 h. Reactions were stopped by the addition of butylated hydroxytoluene (40 μm) and catalase (300 nm) and stored under argon atmosphere at -80 °C until use in cell binding studies.
Flow Cytometry and Fluorescent Microscopy Analysis—Flow cytometry studies were performed on a BD Biosciences FacsScan instrument. For binding studies, the indicated cells (2 × 105 cells in 200 μl) were incubated with 500 μg/ml DiI-labeled liposomes at 4 °C for 30 min to 1 h. Parallel studies monitored binding and uptake of liposomes by incubation at 37 °C for 2 h. Because cells incubated with DiI-labeled liposomes at 37 °C for 2 h showed similar results, we only present binding (4 °C) data. Following incubation, cells were pelleted, washed three times with phosphate-buffered saline containing 0.1% bovine serum albumin, and resuspended in 400 μl for immediate analysis. For fluorescent microscopy, cells were grown on glass coverslips and incubated with DiI-labeled oxidized low density lipoprotein or liposomes at 4 °C for 30 min. Following incubation, cells were fixed with 3.7% paraformaldehyde for 20 min and mounted on glass slides with VectaShield (Vector Laboratories, Burlingame, CA) mounting medium. Fluorescent microscopy studies were performed on a Leica DMR microscope and QImaging EX1 CCD camera, and images were processed using Image-Pro software.
Mouse Peritoneal Macrophages—CD36 knock-out mice back-crossed >10 generations on C57Bl/6 background, along with C57Bl/6 parent colonies, were used for these studies. Mice were injected intraperitoneally with 1 ml of thioglycollate medium (4% in H2O), and cells were harvested by peritoneal lavage with sterile phosphate-buffered saline/RPMI 1640 (1:1) at 4 days postinjection to maximize CD36 surface expression. Approximately 105 cells were plated on sterile glass coverslips in complete RPMI 1640 medium. All animal studies were performed using approved protocols from the Animal Research Committee of the Cleveland Clinic Foundation.
NMR Experiments—NOE studies were performed as recently described (
). Briefly, all lipids and lipid dispersions were freshly prepared and kept under inert atmosphere (either argon or N2) in sealed NMR tubes. All 1H NMR experiments were carried out on either Bruker Avance 600- or 800-MHz spectrometers equipped with cryogenic probes at 30 °C. TDNOE experiments were performed as described by Wager and Wüthrich (
), the correlation time can be assumed to remain constant. The approximate distances (r) between protons were obtained from the initial build-up rate (σ) of the NOE according to the following equation.
The known distance between α- and β-vinyl ether protons of 1-O-1′-(Z)-octadecenyl-2-oleoyl-sn-glycero-3-phosphocholine (2.3 Å) was used as an internal reference with which to estimate internuclear distances by comparisons of their first order NOE build-up rates. Because lateral diffusion effects in lipid dispersions are inevitable, we used initial build-up rates of NOE over periods of time that are short relative to the rate of spin diffusion of the liposome system, as reported previously (
). Confirmation that NOE signals monitored represented intramolecular (as opposed to intermolecular) interactions was achieved by demonstrating comparable build-up rates in NOE signals for OV-PC-containing membranes where <5 mol % incorporation of the synthetic oxPC into the perdeuterated DMPC membrane reduced potential nearest neighbor interactions (data not shown).
KOdiA-PC (for structure, see Table 1), a model oxPCCD36, is a stable oxidation product of PAPC that promotes CD36-specific recognition, macrophage cholesterol accumulation, and foam cell formation when present at low mol % within membrane bilayers or modified lipoproteins (
). Recent NOE-based interrogations of KOdiA-PC in model membranes showed the close spatial proximity between protons on the oxidatively truncated sn-2 acyl group and the polar choline head group, indicating that the distal end of the sn-2 acyl chain protrudes into the aqueous phase (Fig. 1a) (
). In the present studies, we sought to further interrogate the conformation of KOdiA-PC and a variety of other oxPC and nonoxidized PC within model membrane bilayers. NOE studies focused on answering two questions, the first being whether or not detection of an NOE interaction between protons on the distal end of the sn-2 oxidized fatty acids of alternative oxPC and the -N(CH3)3 protons on the choline head group were observable. Such a finding on other oxPC would indicate that lipid peroxidation of membranes is accompanied by protrusion of multiple distinct oxidized fatty acids into the aqueous phase and thus establish the generality of the phenomenon. Second, we hypothesized that NOE studies could serve as a direct experimental approach to test whether the polar head group of oxPC within a membrane has a distinct conformation compared with native (nonoxidized) PC, a hypothesis suggested to account for how pattern recognition receptors like CD36 can detect oxPC compared with native PC within a membrane. We reasoned that NOE-based quantification of the interatomic distances between the -N(CH3)3 protons on the choline head group and the protons of the glycerol backbone at the sn-1-, sn-2-, and sn-3-positions of a phospholipid would be sensitive to the “tilt” of the head group (Fig. 1, b and c) and thus may serve as a direct and quantitative means of discriminating whether a global conformational alteration exists between the head group of oxPC versus native PC within a membrane.
Nuclear Overhauser Effect Spectroscopy of oxPC Species within Membranes Demonstrates That Many Possess sn-2 Acyl Groups That Protrude into the Aqueous Phase, as Indicated by Detection of NOE Signals between Protons of the Choline Head Group and the Distal End of sn-2-Oxidized Fatty Acids—Multiple synthetic oxPC species (OV-PC, ON-PC, and 13-HODE-PC) or nonoxidized PC (DMPC and POPC) were individually incorporated at low molar percentages within SUV preparations composed of carrier perdeuterated (d67)-DMPC, and then the vesicles were analyzed by one-dimensional NOE. Protons in close spatial proximity with the irradiated -N(CH3)3 proton of the choline head group were identified by the presence of an NOE. Remarkably, both OV-PC and ON-PC demonstrated prominent NOE signals between the -N(CH3)3 proton and their characteristic down-field aldehydic proton, a unique signature with which to look for an NOE signal with distal protons on the sn-2-oxidized aliphatic chains. As shown in Fig. 2, a prominent NOE signal is observed between the irradiated -N(CH3)3 protons of the choline head group and the terminal aldehydic proton of OV-PC (note the NOE build-up between protons 1 and 7 upon irradiation of -N(CH3)3 protons (proton 1); Fig. 2c). Since a NOE is extremely sensitive to distance (proportional to 1/r6; see Equation 1) within short irradiation times, it is only detected when the two protons are in close spatial proximity (<5 Å). Thus, these results indicate that the relatively short sn-2 appendage of the oxidized fatty acid in OV-PC (and ON-PC) protrudes into the aqueous phase. In contrast, no NOE signal was detected between protons from the head group and the sn-2 acyl chain in 13-HODE-PC, consistent with the 18 carbon sn-2 chain remaining embedded within the hydrophobic membrane bilayer (note the absence of NOE build-up between protons 1 and 8-10 upon irradiation of the -N(CH3)3 protons (proton 1; Fig. 2f, short irradiation times). As expected, no NOE signals were detectable between the irradiated -N(CH3)3 protons of the choline head group and the terminal aliphatic protons (-CH3) of the sn-1 and sn-2 acyl chains of either POPC or DMPC (data not shown), consistent with the aliphatic chains possessing the anticipated conformation of remaining buried within the hydrophobic membrane interior.
Determination of Critical Internuclear Distances between Protons of Glycerol Backbone and Choline Head Group in Model Membranes Reveals That the Choline Head Group Does Not Undergo Significant Conformational Change Following PC Oxidation—We next performed studies to directly quantify distances between critical protons on an array of synthetic oxPC and native (nonoxidized) PC species within perdeuterated DMPC or DPPC membranes (Table 2). Our goal was to directly quantify distances between choline head group protons and protons in oxidatively truncated sn-2 acyl groups of oxPC as well as detect potential oxidation-induced conformational changes incurred by the choline head group. For these studies, we chose to compare and contrast nonoxidizable species, DMPC and POPC, with a prototypic oxPCCD36 ligand, KOdiA-PC, and other ox-PC ligands, OV-PC, ON-PC, and 13-HODE-PC. Initial rates of NOE build-up were monitored within oxPC incorporated at low mol % within SUV and compared with NOE build-up rates in vesicles containing internal standard synthetic phospholipid (plasmenyl choline) with known distance between the proton pair (α- and β-vinyl protons; see “Experimental Procedures”). Protons monitored included the aforementioned terminal aldehydic protons on oxidized truncated sn-2 acyl groups of OV-PC and ON-PC and the glycerol protons (g1, g2, and g3, corresponding to the C1, C2, and C3 carbons of the PC glycerol backbone) within membranes. Using this approach, even modest changes in the overall conformation of the choline head group relative to the membrane surface would be detected, since it would significantly alter one or more of the distances measured between the choline N-methyl protons and the glycerol (g1, g2, or g3) protons (Fig. 2, b and c).
TABLE 2NOE-derived critical internuclear distances of oxPC and nonoxidized PC species within membrane bilayers
Initially, control studies were performed for determining relative intramolecular distances. Synthetic 1-O-1′-(Z)-octadecenyl-2-oleoyl-sn-glycero-3-phosphocholine (20 mol %) was introduced into the perdeuterated DMPC (60 mol %) SUV containing KOdiA-PC (20 mol %) to serve as an internal reference with which to calibrate distances (
). Irradiation of these vesicles at 4.3 ppm, which corresponds to the β-vinyl ether proton of the plasmalogen, resulted in subsequent build-up of resonance corresponding to the sn-1 α-vinyl ether protons. Using the known distance (2.3 Å) between the α- and β-vinyl protons in the rigid planar sn-1 vinyl ether linkage of plasmenyl choline as an internal standard, the distances between other protons were next deduced by internal comparisons of their respective initial rate for NOE build-up during NOE experiments (Equation 1). These results indicate that the choline -NCH2-protons were on average ∼3.5 Å from the choline -N(CH3)3 protons, consistent with published studies (
). Next, as a further control, 20 mol % DMPC was incorporated into 80 mol % (d67)-DMPC vesicles. Upon irradiation of vesicles at 3.26 ppm (Fig. 3f), the chemical shift corresponding to the choline terminal methyl (-N(CH3)3) tance between choline -N(CH3)3 and NCH2 protons (3.4 Å) as a reference distance to calculate other distances between protons relative to the -N(CH3)3 protons in the glycerol backbone by comparing the initial build-up rate of the indicated pairs (Table 2) according to Equation 1.
In subsequent studies, multiple individual structurally specific oxPC and native PC were examined within (d67)-DMPC membranes. NOE intensity build-up curves between different proton pairs within individual oxPC and nonoxidized PC were similarly generated from build-up rates of NOE difference spectra following irradiation of protons in distinct regions of individual PC species (Fig. 3, a-g). These permitted calculation of multiple intramolecular distances between the choline terminal methyl (-N(CH3)3) protons and other glycerol backbone protons of the molecule (Table 2). Similar measurements were performed for the oxPC lipid species KOdiA-PC, OV-PC, ON-PC, and 13-HODE-PC as well as the native PC molecular species DMPC and POPC. As shown in Table 2, comparisons of the distances between the -N(CH3)3 protons of the choline head group and the glycerol backbone (g1, g2, and g3) protons revealed no significant difference between head group-glycerol backbone proton distances among the various oxPC and native (nonoxidized) PC molecular species examined within the DMPC membranes. In a separate series of confirmatory studies using a more physiological membrane composition, intramolecular distances between protons within KOdiA-PC (20 mol %) in SUV composed of perdeuterated 50 mol % DPPC, 30 mol % cholesterol were performed. Similar NOE build-up rates (Fig. 3) and thus interproton distances (Table 2) were observed for the oxPC species. Collectively, these results indicate that no significant overall conformational change or “tilt” occurs within the choline head group relative to the membrane aqueous interface upon oxidation of PC.
1H-1H NOE are mediated by through-space dipolar interactions and thus theoretically might not distinguish between inter- and intramolecular nuclear spin exchange in vesicles. Performance of the one-dimensional NOE experiments within ≥75 mol % perdeuterated DMPC dramatically diluted potential nearest neighbor interactions, making the NOE distances measured representative of intra-versus intermolecular nuclear spin exchange measurements. This was confirmed in separate studies, since no differences in the observed kinetics of resonance enhancement were noted with perdeuterated DMPC SUV containing either 5 or 20 mol % KOdiA-PC (not shown), confirming their intramolecular character. The intramolecular distances calculated for the six lipid species examined (DMPC, POPC, KOdiA-PC (in DMPC or DPPC/cholesterol), 13-HODE-PC, OV-PC, and ON-PC) are summarized in Table 2. Fig. 4 illustrates the structural conformations within a membrane bilayer for the oxPC species examined.
Studies with Structurally Specific Synthetic Oxidized Ethanolamine Glycerophospholipids Confirm That the Choline Head Group of Oxidized Phospholipids Is Not Required for CD36 Recognition—The above studies strongly indicate that although the sn-2 acyl chain of oxPC species often undergoes a dramatic conformational change, protruding into the aqueous phase, there is no major reorientation of the polar head group relative to membrane dialectic upon PC oxidation, since NOE-derived distances of choline methyl protons relative to the g1, g2, and g3 protons of the glycerol backbone are similar (on average to within ±0.1 Å) for each proton pair internuclear distance examined. These studies therefore suggested that it is the sn-2-oxidized fatty acid that confers CD36 recognition within free oxPL in a membrane. To test this hypothesis, we performed additional experiments that involved either altering or removing the choline head group. CD36 binding activity of various molecular species of PE versus oxPE were examined using fluorescent vesicles in which uniformity of presentation of the PE species was maintained by incorporation of a low molar percentage of PE within a lamellar phase-preferring carrier lipid, POPC, under the experimental conditions employed. SUV with 1 mol % DiI and 20 mol % POPE, PLPE, or PAPE (either as native form or following oxidation by the myeloperoxidase- system) was incubated with cells expressing CD36 versus GFP only as control. A dramatic increase in binding to CD36-expressing cells upon incubation with oxPAPE or oxPLPE vesicles is seen (Fig. 5a), as compared with control cells expressing GFP alone or with SUV containing the nonoxidized PE precursor lipid species PAPE and PLPE. Simultaneous quantification of CD36 surface expression and SUV binding revealed that oxPE in a membrane bilayer serves as a preferred ligand for CD36 compared with nonoxidized PE across multiple orders of magnitude of CD36 surface expression on the target cell (Fig. 5b). These results are consistent with our previous studies that show higher CD36 binding activity for oxidized phospholipids containing choline or serine head groups, compared with their nonoxidized precursor (
) molecular species (i.e. those that contain an oxidatively truncated sn-2 acyl group with a terminal γ-hydroxy(or oxo)-α,β-unsaturated carbonyl). KOdiA-PE and KHdiA-PE were individually synthesized, and their structures were confirmed by NMR and mass spectrometry, as described under “Experimental Procedures.” SUV were then made with carrier POPC (94 mol %), diI (1 mol %), and the indicated molecular species of PE (5 mol %). As shown in Fig. 5, c and d, KOdiA-PE and KHdiA-PE, as well as oxPAPE (but not PAPE), bind to CD36-expressing cells but not to control cells. In parallel studies employing COS-7 fibroblasts transfected with either CD36 or GFP alone, binding of diI-labeled SUV containing oxPE was only observed in cells expressing CD36 (Fig. 6a). Finally, additional studies were performed to determine whether our observations in transfected COS-7 cells were representative of endogenous CD36 binding behavior within macrophages. Mouse peritoneal macrophages from wild type versus CD36 knock-out mice were incubated with DiI-labeled SUV containing low mol % PE versus oxPE species, followed by immunolocalization of macrophage CD36 and monitoring of con-comitant fluorescent vesicle binding. Marked increases in oxPE binding to wild type (wt) versus CD36 knock-out (ko) peritoneal macrophages were observed (Fig. 6b) as well as between oxPE and PE binding to wild type cells (Fig. 6b), consistent with CD36 serving as an oxPE receptor expressed in murine macrophages. Collectively, these results strongly support a critical role for the oxidized sn-2 acyl group, and not the polar head group, as critical determinants of CD36 binding activity with free oxPL in model membranes.
The Polar Head Group Is Not Required for CD36-specific Binding of Free Oxidized Phospholipids within Membranes—As a final series of experiments, we examined whether phosphatidic acid species lacking a polar head group altogether would demonstrate CD36-specific binding upon oxidation of the sn-2 acyl group. Although oxidized PLPA (5 mol %) within SUV composed of POPC strongly and preferentially bound to CD36-expressing cells in a dosedependent manner, no specific binding was observed with either GFP (control)-expressing cells, or other SUV containing either non-oxidized PLPA or alternative (nonoxidized) phosphatidic acid molecular species (Fig. 7, a and b).
Our studies suggest a novel generalized characteristic of membrane architecture in senescent or apoptotic cells and oxidized lipoproteins that we refer to as the “lipid whisker model”; as cell membranes and lipoproteins age and undergo lipid peroxidation, they will apparently “sprout whiskers” composed of an assortment of protruding oxidized fatty acids of varied structures into the aqueous phase (Fig. 8). Lipid peroxidation is an inexorable process within aerobic respiring organisms. Oxidative stress through myriad physiologic pathways will thus apparently induce membranes to sprout lipid appendages, or whiskers, as fatty acid chains are oxidatively modified and assume novel conformations whereby they protrude into the aqueous, hydrophilic milieu. Surrounding the cell, they are anatomically positioned to better engage cell surface receptors of immune surveillance cells.
Indeed, it appears that evolution has exploited these biophysically driven conformational changes whereby innate immune cells recognize specific structural patterns as they germinate at the membrane surface. In the case of macrophage (
) interaction with free oxPL within a membrane bilayer, a terminal γ-hydroxy(or oxo)-α,β-unsaturated carbonyl on the end of the oxidatively truncated acyl group of phospholipids facilitates binding and phagocytosis (
). No doubt, alternative pattern recognition receptors recognize structurally distinct oxPL species among a different assortment of oxidatively truncated acyl groups, triggering various downstream signals. For example, in addition to platelet-activating factor (PAF), an alkylglycerophosphatidylcholine molecular species with an sn-2 acetate group, the PAF receptor is reported to also bind and signal with “PAF-like” oxPC that contain oxidatively truncated sn-2 acyl chains (
). Protrusion of the truncated oxidized sn-2 fatty acid moiety into the aqueous phase may produce a critical conformation in these lipids required for PAF receptor recognition. Similar conformational alterations may facilitate recognition of oxidized lipids by members of the toll receptor family. These receptors are known to recognize pathogen-associated molecular patterns, including foreign and oxidatively modified lipids (
). Finally, in an analogous manner, oxidation-induced conformational changes in phospholipids may underlie the preferential substrate selectivity of some phospholipases for oxidatively modified phospholipids, such as with PAF-acetyl hydrolase (
). Selective cleavage of oxPL by such a pathway might indeed be akin to “getting a shave,” since the whiskers are enzymatically clipped, allowing for rapid reacylation of the lysophospholipid and membrane remodeling.
Interestingly, our data also indicate that simply adding oxygen at the terminal end of the sn-2 chain of an oxPL may not always be sufficient to buoy the acyl group into the aqueous phase. For example, NOE analysis of the longer (18-carbon) 13-HODE-PC demonstrated a traditional buried acyl chain orientation within the membrane interior (Table 2 and Fig. 4). In contrast, for PC possessing oxidatively truncated fatty acid species (e.g. KOdiA-PC, OV-PC, and ON-PC), the addition of a single oxygen appears sufficient to promote the partitioning of the more polar oxygenated acyl chain toward the hydrophilic surface, thereby allowing pattern recognition receptors to directly interact with this type of molecular pattern ligand on the membrane surface (Fig. 4).
In this report, we examine oxidized phospholipid ligands of pattern recognition receptor CD36 as a model system. It has been hypothesized that the phosphocholine head group serves as a critical epitope conferring recognition of oxidized phospholipids by the CD36 receptor (
) and that this recognition epitope is made accessible following an oxidation-induced altered conformation, or tilt, of the choline head group relative to the glycerol backbone. This hypothesis is based upon use of oxPC covalently coupled to target proteins or peptides as a ligand, with CD36 and other pattern recognition receptors, such as CRP and specific autoantibodies (
). The scavenger receptor CD36 can thus apparently bind to oxPL either as free species within the fluid phase of the membrane bilayer or as covalent adducts to proteins. Whether the binding sites on CD36 for these disparate oxPL ligands are similar or spatially distinct remains to be examined.
The present studies focused on ascertaining the structural conformation of free oxPL within membranes in general and the structural and conformational requirements for CD36 recognition of free oxPL within model membranes. By measuring multiple critical intramolecular distances between protons of the choline head group and glycerol backbone by NOE NMR spectroscopy, no differences in global conformation were observed within the choline head group relative to the glycerol backbone of oxidized versus nonoxidized forms of phosphatidylcholine within membranes. Further, altering the choline head group (e.g. to either serine (
) or ethanolamine in the present studies) or removing it entirely, as in the case of phosphatidic acid, had no effect on the ability of CD36 to bind the free oxPL within membranes with characteristic high affinity. It should be noted that membranes are dynamic and results of NOE are global “time-averaged” distances that are the sum of many conformations. The polar head groups of both PC and oxPC alike as well as their acyl chains are dynamic in their respective membrane or aqueous compartments.
The fluid mosaic model developed by Singer and Nicolson (
) is a seminal representation of membrane structure. Since then, much additional data have accumulated to suggest revisions to this model, including membrane patchiness, nonrandom distribution, and functionally specialized regions (
). It is possible that individual molecules or patches of exposed oxidized acyl chains in the aqueous phase may aggregate into microdomains within the membrane, perhaps facilitating regulation of signaling pathways. Indeed, CD36 signaling via an alternative pattern recognition receptor, TLR2, has been reported (
). The present studies suggest that the organizational structure of lipids within a membrane is indeed complex and more dynamic than previously anticipated. We propose that the “lipid whisker model” might be a more widespread biomembrane phenomenon as well as a general consequence of lipid peroxidation in biology, such as during senescence, inflammation, and apoptosis.