Aminophospholipid molecular species asymmetry in the human erythrocyte plasma membrane.

The transbilayer distribution of the molecular species of aminophospholipids in human red blood cell plasma membrane has been investigated using a covalent labelling technique. Separation and quantitative analysis of the molecular species of phosphatidylethanolamine (PE) and phosphatidylserine (PS) was performed using high-performance liquid chromatography with UV detection of the trinitrophenyl derivatives obtained after reaction with trinitrobenzenesulfonic acid (TNBS). When the molecular species distribution obtained with intact cells was compared to that of the whole membrane, a molecular species asymmetry was evident. This phenomenon was most clearly evident when the reaction was performed at low temperatures (0 degrees C) and was obscured by the excessive labelling or probe permeation associated with higher temperatures or longer incubation times. The monoene species were enriched in the outer leaflet, they comprised about 30% of the PE species in this leaflet. The polyunsaturates were preferentially localized in the inner leaflet and this was true of the arachidonyl species in particular as they represented up to 35% of this pool. The w-3 polyunsaturated fatty acids displayed a preferential localization in the plasmalogen subclass in comparison to the diacyl fraction, i.e., they comprised about 58 of the former and 42% of the latter subclass of cellular PE w-3 species. Data concerning the separation, identification and quantification of PS molecular species in human erythrocytes is also presented. The internal localization of the polyunsaturated species as well as the compartmentalization of the w-3 and w-6 pools will have metabolic, structural and physical implications for membrane function.

Pathways of arachidonic acid liberation in thrombin and calcium ionophore A23187-stimulated human endothelial cells: respective roles of phospholipids and triacylglycerol and evidence for diacylglycerol generation from phosphatidylcholine.

Cultured endothelial cells from human umbilical vein were incubated for 20 h at 37 degrees C in the presence of [U-14C]arachidonic acid. Around 60-70% of the radioactive fatty acid was incorporated into cell lipids and was predominantly found in phosphatidylcholine, phosphatidylethanolamine, phosphatidylinositol and triacylglycerol (39%, 33%, 13% and 6.5% of total incorporated radioactivity, respectively). Stimulation of the cells with human thrombin (2 U/ml) or calcium ionophore A23187 (5 microM) promoted the release into supernatants of arachidonic acid, 6-ketoprostaglandin F1 alpha, prostaglandins E2 and F2 alpha, in decreasing order of importance. The amount of secreted material was 4-fold higher with A23187, compared to thrombin. Parallel to the liberation process, phosphatidylcholine underwent a rapid decrease of radioactivity with both agonists, suggesting the involvement of a Ca2+-dependent phospholipase A2. Phosphatidylethanolamine displayed a minor decrease with A23187, whereas some reacylation was observed at 10 min with thrombin. Phosphatidylinositol was non-significantly affected in thrombin-stimulated cells, whereas A23187 promoted an early but minor decrease, followed by resynthesis. In contrast to A23187, thrombin was also able to promote a significant hydrolysis of triacylglycerol, which might thus be implicated in the process of arachidonate liberation. Finally, radioactive phosphatidic acid and diacylglycerol appeared in endothelial cells, in response to the two agonists. However, diacylglycerol formation did not parallel that of phosphatidic acid, especially with A23187. Determination of the 14C/3H ratio of the different lipids upon cell labelling with both [14C]arachidonic acid and [3H]palmitic acid revealed that diacylglycerol and phosphatidic acid are hardly derived from inositol-phospholipid breakdown by phospholipase C. Other possible pathways involving for instance phospholipase C splitting of phosphatidylcholine are discussed.

Binding of 13-HODE and 5-, 12- and 15-HETE to endothelial cells and subsequent platelet, neutrophil and tumor cell adhesion.

Some studies report that endothelial cells preferentially take up the lipoxygenase-derived arachidonic acid metabolite, 5-hydroxyeicosatetraenoic acid (5-HETE), released from stimulated leukocytes (polymorphonuclear leukocytes, PMNs), whereas others report that endothelial cells preferentially take up 12-HETE released from platelets. The biological relevance of these observations, however, is unknown. Recently, we and others have found that, under basal conditions, endothelial cells, PMNs and tumor cells metabolize linoleic acid via the lipoxygenase enzyme to 13-hydroxyoctadecadienoic acid (13-HODE). We propose that endogenous levels of these metabolites regulate blood-vessel wall cell adhesion. In this study, we have measured (1) the relative binding of 5-, 12- and 15-HETE, and 13-HODE to endothelial cell monolayers, and (2) their effects on endothelial cell adhesivity with platelets, PMNs and tumor cells. There was a dose-related and specific binding of 5-[3H]HETE to endothelial cells but no binding of 12- or 15-HETE or 13-HODE. Platelet or PMN adhesion to endothelial cells was unaffected by the 5-HETE binding, but tumor cell adhesion was blocked by 40% (P less than 0.01). Interestingly, preincubation of endothelial cells with 13-HODE, 12-HETE or 15-HETE decreased platelet adhesion to endothelial cells (P less than 0.05), even though these metabolites did not bind to the endothelial cells. We conclude that 5-HETE preferentially binds to endothelial cells and interferes with a specific receptor for tumor cells, whereas the other metabolites neither bind to cells nor affect cell adhesion.

Uptake of HDL unesterified and esterified cholesterol by human endothelial cells. Modulation by HDL phospholipolysis and cell cholesterol content.

Human HDL (1.070-1.210), doubly labelled with 3H/14C-labelled unesterified cholesterol and 3H-labelled esterified cholesterol were incubated for 1-5 h with monolayer cultures of human endothelial cells. HDL were preincubated for 60-120 min the presence of albumin and with/without purified phospholipase A2 (control HDL, phospholipase A2 HDL) before dilution in the cell culture medium. Average phosphatidylcholine (PC) degradation was 62.10% +/- 2.57% (range 45-80%). A purified lipase/phospholipase A1 from guinea pig pancreas was used in some experiments (range of PC hydrolysis: 16-70%). (1) 3H/14C-labelled unesterified cholesterol and 3H-labelled esterified cholesterol appeared in cells during 0-5 h incubations. Trypsin treatment allowed a simple adsorption of HDL onto the cell surface to be avoided, and most of the 3H-labelled esterified cholesterol transferred to cells was hydrolysed. Cell uptake of radioactive cholesterol increased as a function of HDL concentration but no saturation was achieved at the highest lipoprotein concentration used (200 micrograms cholesterol/ml). Flux of 3H/14C-labelled unesterified cholesterol was related to the cell cholesterol content, suggesting that it might partly represent an exchange process. The cell cholesterol content was slightly increased after 5 h incubation with HDL (+16%). (2) Pretreatment of HDL with purified phospholipase A2 doubled on average the amount of cell recovered 3H-labelled esterified cholesterol, while the flux of 3H/14C-labelled unesterified cholesterol was enhanced by 15-25%. Both transfer and cell hydrolysis of 3H-labelled esterified cholesterol were increased. A stimulation was also observed using purified lipase/phospholipase A1, provided that a threshold phospholipid degradation was achieved (between 27 and 45%). (3) Endothelial cells were conditioned in different media so as to modulate their charge in cholesterol. The uptake of 3H-labelled esterified cholesterol was found to be significantly higher in cholesterol-enriched cells compared to the sterol-depleted state. Finally, movements of 3H-labelled esterified cholesterol from HDL to endothelial cells were essentially unaffected by cell density or by the presence of partially purified cholesterol ester transfer protein. The possible roles of the transfer of HDL esterified cholesterol to endothelial cells and its modulation by phospholipases are discussed.

Evidence for a two-step process in prostaglandin secretion. Intracellular accumulation of prostacyclin precedes its release from human endothelial cells in culture.

Cultured endothelial cells (EC) from human umbilical vein were incubated with [U-14C]arachidonic acid (AA) followed by a challenge with thrombin (2 units/ml) or calcium ionophore A23187 (5 microM) for 0.5-10 min at 37 degrees C. In both cases, AA was rapidly liberated from phospholipids and converted into prostaglandin I2 (PGI2), as determined by the radioactivity of the stable derivative 6-keto-PGF1 alpha. Maximal liberation of AA and synthesis of PGI2 were achieved within 2 min, but the two compounds first accumulated in EC prior to their release into supernatants. This finding, which was never reported before, raises the question of the mechanism of AA and PG release through the cell membranes and offers a convenient model to investigate this still obscure process.

Stimulation of 12-HETE production in human platelets by an immunomodulator, LF 1695. Evidence for activation of arachidonate liberation coupled to cyclo-oxygenase inhibition.

Upon incubation with human platelets previously labelled with [14C]arachidonic acid, a new immunomodulator, LF 1695, induced the accumulation of [14C]-12-(S)-hydroxy-5,8,10,14-eicosatetraenoic acid (12-HETE). Although the time course of [14C]HETE accumulation was identical with 60 microM LF 1695 and calcium ionophore A23187, the latter compound also promoted the formation of 14C-labelled thromboxane B2 and 12-(S)-hydroxy-5,8,10-heptadecatrienoic acid (HHT), whereas 12-HETE was the only arachidonic acid metabolite generated under the action of LF 1695, suggesting that the drug inhibited cyclo-oxygenase. This was further confirmed by the fact that LF 1695 inhibited the second wave of platelet aggregation induced by ADP as well as arachidonic acid effects. Cell lipid analysis revealed that arachidonic acid was liberated from both triacylglycerol and phosphatidylcholine. The effect was observed in the concentration range 15-90 microM, with a half-maximal effect at 30 microM for HETE production, 15 microM for triacylglycerol hydrolysis and 45 microM for phosphatidylcholine deacylation. Incubation of platelets with [14C]arachidonic acid in the presence of 60 microM LF 1695 resulted in a strong inhibition of arachidonic acid incorporation into the various cell lipids, indicating that arachidonic acid mobilization might be due to inhibition of reacylation processes. It is concluded that LF 1695 displays an original and complex effect on platelet lipid metabolism, resulting in the specific generation of lipoxygenase metabolites.

Increased content of annexin II (p36) and p11 in human placenta brush-border membrane vesicles during syncytiotrophoblast maturation and differentiation.

Annexins are a group of proteins abundant in placental membranes where they may play diverse functional roles. Annexins are expressed in high levels in mature placenta but little is known about their presence at very early stages of gestation and later. We used the model of brush-border membrane vesicles (BBMV) at different stages of gestation to assess precise localization of some of these proteins in syncytiotrophoblast apical membrane and to determine their appearance along the maturation process of placenta. Here we describe annexins type I, II, IV, V and VI which are present all along gestation in BBMV. Annexin II (p36) is present with the S100 like calcium-binding protein p11 in BBMV, where they can constitute heterotetrameric forms of annexin II linked to cytoskeleton structures. No variation of annexins I, IV and VI content was observed in BBMV along pregnancy. Annexin V undergoes significant decrease after 12th week, which could be related to local anticoagulant activity. Levels of annexin II and p11 increased progressively during gestation suggesting that heterotetrameric forms of annexin II play a role in the differentiation process of placenta and in function of the mature microvilli.

Phosphatidylcholine is the major phospholipid providing arachidonic acid for prostacyclin synthesis in thrombin-stimulated human endothelial cells.

Upon incubation for 24 hours with [3H]arachidonic acid (AA, 1 mu Ci/ml), cultured endothelial cells from human umbilical vein incorporated one half of the added radioactivity, mostly into phospholipids (83% of the total cell radioactivity). Distribution of the label between the various phospholipid classes was found to reflect the distribution of endogenous AA. Stimulation with human thrombin (2 U/ml) promoted a rapid release of radioactive material into supernatants, which contained essentially 6-keto-prostaglandin F1 alpha and non-converted AA. This process levelled off at 10 min, at which time phosphatidylcholine displayed a decrease accounting for 3.7% of the total cell radioactivity. Phosphatidylinositol also appeared significantly diminished, but this decrease was almost 2.5 fold less than that observed in phosphatidylcholine. It is concluded that AA availability for prostacyclin biosynthesis is mostly regulated by a phospholipase A2.

Inhibition of intraocular fibrin formation with annexin V.

Annexin V is a member of the calcium- and phospholipid-binding proteins, known to have an antithrombotic effect. For the first time, we have tested its ability to prevent intraocular postoperative fibrin formation in a standardised rabbit model and compared its effect with that of heparin. Annexin V, 20 micrograms and 60 micrograms, injected in the anterior chamber post-operatively, significantly reduced the area of the fibrin clot and its time to clearing. Annexin V appeared to be as efficient as heparin. It probably acts by preventing phospholipids from playing their role in the coagulation cascade which leads to fibrin formation. Furthermore, annexin V has an anti-inflammatory effect by protecting phospholipids from phospholipase A2 activity. Therefore, annexin V might be considered as a new therapeutic agent acting both on fibrin formation and inflammatory processes.

Isolation and properties of a novel phospholipase A from rat brain that hydrolyses fatty acids at sn-1 and sn-2 positions.

A Ca(2+)-independent phospholipase A that releases various fatty acids from sn-1 and sn-2 positions was partially purified from rat brain soluble fraction. The enzyme showed an approximate molecular mass of 300 kDa on gel filtration column chromatography. Its enzymatic properties are distinct from those of well characterized phospholipase A2 enzymes; by using a series of synthetic phosphatidylcholines, the enzyme cleaved oleic, linoleic, and arachidonic acids like phospholipase A2, and released palmitic and stearic acids like phospholipase A1. Phosphatidylcholine, phosphatidylethanolamine, phosphatidylinositol, and phosphatidic acid were hydrolysed with almost equal efficiencies by this enzyme. These results indicate that the enzyme isolated is a novel Ca(2+)-independent intracellular phospholipase A that might be responsible for production of various fatty acids from membrane phospholipids.

Asymmetric incorporation of dietary n-3 fatty acids into membrane aminophospholipids of human erythrocytes.

Dietary supplementation with different classes of polyunsaturate fatty acids is known to result in their incorporation into cell membranes, but the effects of this on eicosanoid formation and other cell functions frequently does not correspond to the degree of alteration in total membrane fatty acids. This phenomenon may be related to the compartmentalization of polyunsaturate fatty acids both within the organelles and within membranes. Aminophospholipids are asymmetrically distributed across the membrane bilayers of most human cells. These phospholipids are highly enriched in polyunsaturated fatty acids, and are known to have specific interactions with a number of membrane proteins. To determine whether dietary n-3 fatty acids are preferentially incorporated into membrane lipids in a particular spatial pattern, we have utilized the nonpermeant aminophospholipid probe, trinitrobenzenesulfonic acid, to study the transmembrane molecular species distribution of human erythrocyte ethanolamine phospholipids and phosphatidylserines before and at the end of 4 weeks of dietary supplementation with n-3 fatty acids. Selective incorporation of n-3 fatty acids occurred in the inner membrane leaflet ethanolamine phospholipids, particularly into the alkenyl-acyl species. The n-3 species in phosphatidylserines, particularly 18:0 and 22:6 n-3 (sn-1 and sn-2, respectively), replaced n-6 and n-9 species. These data may provide a basis for different cell responses to n-3 fatty acid enrichment, and for different degrees of diet-induced alteration in responses involving inner and outer membrane leaflet functions.

Analysis of aminophospholipid molecular species by high performance liquid chromatography.

A new method is described for the separation of individual molecular species of the aminophospholipids, phosphatidylethanolamine and phosphatidylserine. Trinitrobenzene-sulfonic acid was used to derivatize both aminophospholipids and the derivatives were purified by thin-layer chromatography. A reversed-phase high performance liquid chromatography technique was developed to separate and quantify individual molecular species based upon ultraviolet detection of the attached chromophore. The retention times of the molecular species on the C18 reversed-phase column were longer with increasing carbon chain length and decreasing degree of unsaturation of fatty acyl chain. The overall procedure allowed a quantitative recovery of the aminophospholipid species. The lower limit of detection was about 10 pmol and a linear response was observed in the range of 0.1-10 nmol of phospholipid. Using this method, we were able to separate and quantify trinitrophenyl-phosphatidylethanolamine molecular species of both subclasses (diacyl and alkenyl) from human red blood cells and rat brains. Separation of species was confirmed by gas-liquid chromatographic analysis of the fatty acid content of each peak and by thermospray liquid chromatography-mass spectrometry. This new method provides a convenient and sensitive technique for studies of aminophospholipid molecular species composition. Furthermore, it appears to be a useful tool for the analysis of asymmetric distribution of these species in biological membranes.

Annexin 5 as a potential regulator of annexin 1 phosphorylation by protein kinase C. In vitro inhibition compared with quantitative data on annexin distribution in human endothelial cells.

In vitro phosphorylation of annexin 1 by purified rat brain protein kinase C (PKC) has been studied in the presence of annexin 5, which is not a substrate for PKC. Annexin 5 promoted a dose-dependent inhibition of annexin 1 phosphorylation, which could be overcome by increasing the concentration of phosphatidylserine (PtdSer). In addition, a close relationship was found between the amount of PtdSer uncovered by annexin 5 and the residual phosphorylation of annexin 1. These data fit with the 'surface depletion model' explaining the antiphospholipase activity of annexins. In order to check the possibility that the in vitro effect of annexin 5 could be of some physiological relevance, annexins 1, 2, and 5, as well as the light chain of calpactin 1 (p11), have been quantified in human endothelial cells by measuring the radioactivity bound to the proteins after Western blotting with specific antibodies and 125I-labelled secondary antibody. Our data indicate that annexins 1 and 5, PKC and PtdSer are present in human endothelial cells in relative amounts very similar to those used in vitro under conditions permitting the detection of the inhibitory effect of annexin 5. Since annexin 1 remained refractory to PKC-dependent phosphorylation in intact cells, we suggest that annexin 5 might exert its inhibitory effect towards PKC in vivo, provided that its binding to phospholipids can occur at physiological (micromolar) concentrations of Ca2+. This was previously shown to occur in vitro using phosphatidylethanolamine/phosphatidic acid vesicles [Blackwood and Ernst (1990) Biochem. J. 266, 195-200]. Using identical assay conditions, which also allowed expression of PKC activity, annexin 5 again inhibited annexin 1 phosphorylation without interfering with PKC autophosphorylation. These data suggest that annexins 1 and 5 might interact with each other on the lipid surface, resulting in a specific inhibition of annexin 1 phosphorylation by PKC. Whether a similar mechanism also occurs in vivo remains to be determined.

Annexin I as a potential inhibitor of insulin receptor protein tyrosine kinase.

Insulin receptor (IR) purified from human placenta by wheat germ agglutinin affinity chromatography was incubated in the presence of insulin, [gamma-32P]ATP and annexin I. In parallel to its own tyrosine phosphorylation, annexin I promoted a dose-dependent inhibition of IR autophosphorylation (IC50 0.5 microM). This effect was specific for insulin-stimulated tyrosine kinase activity and required the N-terminal end of the protein containing the phosphorylatable Tyr21 residue. A pentadecapeptide encompassing residues 16-30 of human annexin I displayed a similar activity, but at higher concentrations. These data underscore a specific interaction of IR with annexin I, which should be considered as a potential physiological regulator of the effects of insulin on its target tissues.

Specialization in membrane structure and metabolism with respect to polyunsaturated lipids.

The principal features of the hypotheses presented here are presented in a model (Figure 8). Polyunsaturated phospholipids are preferentially localized on the cytosolic leaflet of the plasma membrane and around proteins. They influence the physical state of these microenvironments and thereby help to regulate cellular functions. Losses of these species due to phospholipase attack, dietary deprivation or other pathological stimuli (such as alcohol), may lead to drastic alterations in some of these functions. The PUFA's released, including 22:6 omega 3, can be metabolised by a lipoxygenase enzyme in the brain and platelet and the products formed are biologically active.

Effect of dexamethasone on prostaglandin synthesis and on lipocortin status in human endothelial cells. Inhibition of prostaglandin I2 synthesis occurring without alteration of arachidonic acid liberation and of lipocortin synthesis.

Glucocorticoids have been shown to decrease prostaglandin I2 synthesis in human endothelial cells, suggesting the possible involvement of lipocortin in the inhibition of arachidonic acid liberation achieved by phospholipase A2 (De Caterina, R., and Weksler, B. B. (1986) Thromb. Haemostasis 55, 369-374). To test this hypothesis, human endothelial cells labeled with [14C]arachidonic acid were stimulated with thrombin (2 units/ml, 10 min), resulting in the secretion of free arachidonic acid together with various 14C-labeled metabolites, mainly 6-keto-prostaglandin F1 alpha, the stable derivative of prostaglandin I2. Under conditions where prior incubation of cells with dexamethasone reduced by 51% 6-keto-prostaglandin F1 alpha production, phospholipid hydrolysis induced by thrombin remained unaffected. Using three rabbit polyclonal antibodies directed against endonexin I, lipocortin I, and lipocortin II, evidence was obtained for the presence in human endothelial cells of equivalent amounts of lipocortin I and an immunologically unrelated 33-kDa protein, together with lower quantities of 67-kDa calelectrin/calcimedin. These Ca2+- and phospholipid-binding proteins were selectively extracted with [ethylene-bis(oxyethylene-nitrilo)]tetraacetic acid (EGTA) from cell membranes precipitated in the presence of Ca2+, and they displayed an inhibitory activity against pig pancreas phospholipase A2. However, the amounts of the three proteins were not changed by cell treatment with 2.5 microM dexamethasone, as detected upon polyacrylamide gel electrophoresis by silver staining, immunoblotting, or autoradiography following [35S]methionine in vivo labeling. Since the antiphospholipase A2 activity of EGTA extracts was hardly modified, it was concluded that an increased synthesis of lipocortin cannot account for the inhibition of prostaglandin synthesis brought about by dexamethasone, suggesting other biological functions for these proteins.

Morphological and biochemical evidence for partial nuclear localization of annexin 1 in endothelial cells.

Using immunofluorescence, an affinity-purified anti-annexin-1 polyclonal antibody showed both cytoplasmic and nuclear staining, whereas antibodies against annexins 2, 5 and 6 labelled almost exclusively the cytoplasm of cultured endothelial cells. This was further confirmed by immunogold labelling and electron microscopy using a monoclonal antibody, annexin 1 being detected close to the plasma membrane, in the cytoplasm, as well as inside the nucleus. Finally, using immunoblotting, purified nuclei were shown to contain annexin 1, which was not removed by EDTA treatment. These data open some new perspectives in the understanding of annexin function, including possible involvement in nucleoskeleton dynamics and regulation of proliferation through cell signalling.