1-Oleoyl-2-acetylglycerol stimulates 5-lipoxygenase activity via a putative (phospho)lipid binding site within the N-terminal C2-like domain.

5-Lipoxygenase (5-LO) catalysis is positively regulated by Ca2+ ions and phospholipids that both act via the N-terminal C2-like domain of 5-LO. Previously, we have shown that 1-oleoyl-2-acetylglycerol (OAG) functions as an agonist for human polymorphonuclear leukocytes (PMNL) in stimulating 5-LO product formation. Here we have demonstrated that OAG directly stimulates 5-LO catalysis in vitro. In the absence of Ca2+ (chelated using EDTA), OAG strongly and concentration-dependently stimulated crude 5-LO in 100,000 x g supernatants as well as purified 5-LO enzyme from PMNL. Also, the monoglyceride 1-O-oleyl-rac-glycerol and 1,2-dioctanoyl-sn-glycerol were effective, whereas various phospholipids did not stimulate 5-LO. However, in the presence of Ca2+, OAG caused no stimulation of 5-LO. Also, phospholipids or cellular membranes abolished the effects of OAG. As found previously for Ca2+, OAG renders 5-LO activity resistant against inhibition by glutathione peroxidase activity, and this effect of OAG is reversed by phospholipids. Intriguingly, a 5-LO mutant lacking tryptophan residues (Trp-13, -75, and -102) important for the binding of the 5-LO C2-like domain to phospholipids was not stimulated by OAG. We conclude that OAG directly stimulates 5-LO by acting at a phospholipid binding site located within the C2-like domain.

In the biosynthesis of leukotrienes (LTs) 1 , 5-lipoxygenase (5-LO) catalyzes the initial oxygenation of arachidonic acid (AA) leading to 5-HPETE, which is further metabolized by 5-LO to LTA 4 (for review, see Refs. 1 and 2). The mechanisms leading to activation of 5-LO in the cell are complex, and the enzymatic activity of 5-LO is tightly controlled. At its active site, 5-LO contains an essential nonheme-bound iron (3), which, in the resting state, is in the ferrous (Fe 2ϩ ) form and requires oxidation to the ferric (Fe 3ϩ ) state to enter the catalytic cycle (4,5).
In the absence of stimulating co-factors, 5-LO activity in vitro is low, and it has been found that Ca 2ϩ , ATP, and phospholipids are required for full enzyme activity (for review, see Refs. 1 and 2). Also, a threshold level of lipid hydroperoxides (LOOH) is necessary for initial enzyme activation, converting ferrous to ferric iron (4,6,7). For cellular 5-LO activation, elevated Ca 2ϩ levels (8), phosphorylation events by tyrosine kinases (9) and by members of the mitogen-activated protein kinase family (10,11), an elevated peroxide tone (12)(13)(14), and nuclear membrane association (15,16), including co-localization with the 5-LO-activating protein (17,18), are determinants. Nevertheless, the precise pathway(s) and regulatory mechanisms of 5-LO activation in the cell are incompletely understood.
Based on theoretical models of the tertiary structure, 5-LO consists of a catalytic and an N-terminal C2-like ␤-barrel domain (19 -22). The C2-like domain binds Ca 2ϩ and phosphatidylcholine (PC) (19,21) and targets 5-LO to the nuclear membrane (20,21). Ca 2ϩ and/or phospholipids can strongly augment 5-LO activity in vitro but also may have no stimulatory effect, which depends on the assay conditions (see Refs. 1 and 23 and references therein). Ca 2ϩ has been shown to increase the affinity of 5-LO toward AA (24), to facilitate membrane association and binding to PC vesicles (21,25), and appears to reduce the requirement of 5-LO for activating lipid hydroperoxides (26). For stimulation by phospholipids, apparently only the zwitterionic PC and a non-physiological cationic phospholipid (but not phosphatidylserine, phosphatidylethanolamine, phosphatidylinositol, or diacylglycerol (DAG)) stimulate 5-LO reactions (27,28). It has been proposed that 5-LO catalysis occurs at an interface between lipid and water, and the ratio of AA to phospholipid seems important for its activity (29).
Previous studies show that DAGs similar to OAG up-regulate agonist-induced 5-LO product synthesis in leukocytes, which in part might be due to increased availability of substrate for 5-LO (see Ref. 30 and references therein). Recently, we found that OAG, not only primes leukocytes for enhanced 5-LO product synthesis upon subsequent agonist stimulation, but also functions as a direct agonist for polymorphonuclear leukocytes (PMNL), stimulating, by novel undefined mechanisms (30), 5-LO product formation in the presence of exogenously added AA. Thus, OAG-induced 5-LO activity is not connected with increased nuclear membrane association or * This study was supported by grants from the European Union (QLG1-CT-2001-01521, LSHM-CT-2004-0050333). 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.
Cells-Human PMNL were freshly isolated from leukocyte concentrates obtained at St. Markus Hospital (Frankfurt, Germany). In brief, venous blood was taken from healthy adult donors, and leukocyte concentrates were prepared by centrifugation at 4000 ϫ g for 20 min at 20°C. PMNL were immediately isolated by dextran sedimentation, centrifugation on Nycoprep cushions, and hypotonic lysis of erythrocytes, as described previously (31). Cells (10 8 /ml; purity Ͼ 96 -97%) were finally resuspended in PBS containing 1 mg/ml glucose.
Preparation of 100,000 ϫ g Supernatants and Purification of 5-LO Proteins-For purification of recombinant 5-LO proteins, E. coli was lysed by incubation in 50 mM triethanolamine/HCl, pH 8.0, 5 mM EDTA, soybean trypsin inhibitor (60 g/ml), 1 mM phenylmethylsulphonyl fluoride, and lysozyme (500 g/ml), homogenized by sonication (three times for 15 s each), and centrifuged at 19,000 ϫ g for 15 min. Proteins including 5-LO were precipitated with 50% saturated ammonium sulfate during stirring on ice for 60 min. The precipitate was collected by centrifugation at 16,000 ϫ g for 25 min, and the pellet was resuspended in 20 ml of PBS containing 1 mM EDTA and 1 mM phenylmethylsulphonyl fluoride. For purification of 5-LO from human PMNL, 5 ϫ 10 8 cells were resuspended in 10 ml of PBS containing 1 mM EDTA. After cooling on ice for 10 min, the cells were homogenized by sonication (three times for 10 s each). Whole homogenates of PMNL and the 16,000 ϫ g supernatants from E. coli were centrifuged at 100,000 ϫ g for 70 min at 4°C, and the 100,000 ϫ g supernatant (S100) was applied to an ATP-agarose column. The column was eluted as described previously (32). S100 or purified 5-LOs were immediately used for 5-LO activity assays.
Determination of 5-LO Product Formation in Cell-free Systems-To obtain whole cell homogenates of PMNL and corresponding S100, 10 8 freshly isolated PMNL were resuspended in 1 ml of PBS containing 1 mM EDTA, cooled on ice for 10 min, and sonicated (three times for 10 s each) at 4°C. The whole homogenate was then centrifuged (100,000 ϫ g for 70 min at 4°C) to get the S100. For determination of 5-LO activity in whole homogenates or in S100 from PMNL (7.5 ϫ 10 6 cells) or E. coli, 1 mM ATP, phospholipids, glycerides, and GPx-1/GSH were added as indicated. For determination of the activity of purified 5-LO from PMNL or E. coli, 5-LO (corresponding to 2 ϫ 10 7 PMNL or 0.5 g of 5-LO protein, respectively) was added to 1 ml of a 5-LO reaction mix (PBS, pH 7.4, 1 mM EDTA, and 1 mM ATP). After 5 min at 4°C, the samples containing either homogenates, S100, or partially purified 5-LO were prewarmed for 30 s at 37°C, and AA at the indicated concentrations with or without 2 mM CaCl 2 was added to start the 5-LO reaction. For PMNL-S100 and purified 5-LO enzyme in the presence of Ca 2ϩ or OAG (as well as without additives), maximum amounts of 5-LO products were formed within 6 -8 min, remaining constant until ϳ13-15 min (not shown), implying that a 10-min incubation is a suitable time period to assess 5-LO product formation. Thus, the reaction was stopped after 10 min with 1 ml of methanol. Formed 5-LO metabolites were extracted and analyzed by HPLC, as described previously (33). 5-LO activity in samples derived from PMNL is expressed as ng of 5-LO products/10 6 cells. In samples from E. coli, 5-LO activity is ex-pressed as ng of 5-LO products/g of protein. 5-LO products include LTB 4 (in homogenates and S100 of PMNL) and its all-trans isomers, 5(S),12(S)-DiHETE, 5-HETE, and 5-HPETE. 5-HETE and 5-HPETE co-elute as one major peak, and integration of this peak represents both eicosanoids. Cysteinyl LTs (LTC 4 , -D 4 , and -E 4 ) were not detected, and oxidation products of LTB 4 were not determined.
Determination of Glutathione Peroxidase Activity-GPx activity was measured according to the indirect GSH reductase-coupled method described by Wendel (34). One unit of GPx activity was defined as the conversion of 0.5 mol of NADPH to NADP ϩ /min at 37°C and 1 mM GSH. GPx activity is expressed as milliunits/10 6 cells.
Statistics-The program GraphPad PRISM version 3.0 was used for statistical comparisons. Statistical evaluation of the data was performed using the Student's t test for unpaired observations. A p value of Ͻ0.05 was considered significant.

OAG Stimulates 5-LO in Vitro in the Absence of Ca 2ϩ -
The effects of OAG on 5-LO activity in vitro were determined using whole PMNL homogenates and corresponding S100, as well as for partially purified 5-LO derived from PMNL. Under optimized assay conditions, i.e. the presence of 1 mM ATP and 1 mM Ca 2ϩ and absence of GPx activity, OAG (10 -100 M) caused no enhancement of 5-LO product synthesis, regardless of the substrate concentration (2-40 M AA) and the source of 5-LO (homogenates, S100, or purified 5-LO; Fig. 1A). In contrast, OAG slightly impaired 5-LO activity in a concentration-dependent manner.
To determine whether OAG could activate 5-LO under suboptimal assay conditions, Ca 2ϩ was excluded from the incubations, which resulted in reduced 5-LO activity in homogenates, in S100, and for purified 5-LO in particular at low substrate concentrations (Ͻ20 M AA; compare also Ref. 23). For 5-LO in homogenates, the activity was not increased by OAG (0.3-100 M) in the absence of Ca 2ϩ . However, crude 5-LO in S100 or purified 5-LO was concentration-dependently up-regulated by OAG up to 8-and 14-fold, respectively (Fig. 1B). Significant effects of OAG were apparent already at 0.3 M OAG, and maximal stimulation was observed at 30 M. Simultaneous addition of OAG together with AA caused the same stimulatory effect on 5-LO (in S100 or purified enzyme) as compared with prolonged OAG preincubation for up to 60 min prior to AA addition (not shown). Moreover, OAG was able to up-regulate the activity of 12-LO from human platelets and 15-LO from eosinophilic PMNL (ϳ3-4-fold) in the absence of Ca 2ϩ , although to a lower extent as compared with 5-LO (Fig. 1C). In the presence of Ca 2ϩ , OAG also failed to stimulate 12-and 15-LO (not shown).
In agreement with previous observations (23), the stimulatory effect of Ca 2ϩ on 5-LO activity is most pronounced when the substrate concentration is low, whereas at higher amounts of AA (Ն20 M), Ca 2ϩ fails to increase 5-LO product synthesis (Fig. 1D). Similarly, OAG markedly increased 5-LO product formation at low AA concentrations (0.5-10 M AA) but had no significant effect at AA concentrations Ն20 M (Fig. 1E). In our previous report (26) we suggest that stimulation of 5-LO by Ca 2ϩ is because of an increased affinity toward activating LOOH. In fact, the addition of 5-HPETE to PMNL-S100 concentration-dependently up-regulated the LTA 4 synthase activity of 5-LO in the absence of Ca 2ϩ or OAG (Fig. 1F). However, it should be noted that 5-HPETE could simply serve as the 5-LO substrate for LTA 4 synthesis under these conditions, thereby increasing the amounts of LTA 4 hydrolysis products.
Phospholipids Abolish the 5-LO Stimulatory Effects of OAG-The fact that OAG stimulates 5-LO in S100 but not in homogenates led us to conclude that cellular membrane components of homogenates, i.e. phospholipids, may interfere with the action of OAG on 5-LO. S100, essentially devoid of phos- Homogenates and S100 of isolated PMNL (7.5 ϫ 10 6 cells) or 5-LO isolated from 2 ϫ 10 7 PMNL were incubated in PBS plus 1 mM EDTA, 1 mM ATP, and the indicated concentrations of OAG at 4°C. After 5 min, the samples were prewarmed, and 2 mM CaCl 2 plus 5 M AA was added. After 10 min at 37°C, 5-LO activity was determined as described under "Experimental Procedures." B, effects of OAG in the absence of Ca 2ϩ . Samples were prepared and 5-LO activity was determined as described above, except that CaCl 2 was omitted. 5-LO activity in the absence of OAG (control) was 4.7 Ϯ 1.6, 5.9 Ϯ 1.4, and 0.9 Ϯ 0.3 ng/10 6 cells for homogenates, S100, and purified 5-LO, respectively. C, effects of OAG on the activity of 12-and 15-LO. 12-LO activity was determined in S100 of 10 8 platelets. 5-and 15-LO activities were determined in S100 of isolated PMNL. The activities of the LOs in the absence of Ca 2ϩ were assessed as described above. In the absence of OAG (control), the activities of the LOs were 5.9 Ϯ 1.4, 0.52 Ϯ 0.06, and 4.3 Ϯ 0.9 ng/10 6 cells for 5-, 12-, and 15-LO, respectively. D and E, influence of the AA concentrations on the 5-LO stimulatory effects by Ca 2ϩ or OAG. 5-LO, isolated from PMNL, was diluted in PBS plus 1 mM EDTA and 1 mM ATP. The samples were either prewarmed and 2 mM CaCl 2 plus the indicated concentrations of AA were added (D) or preincubated with 10 M OAG for 5 min at 4°C and then prewarmed, and AA was added (E). After 10 min, 5-LO activity was determined. F, effect of 5(S)-HPETE on 5-LO activity. The S100 of PMNL (7.5 ϫ 10 6 cells) were incubated in PBS plus 1 mM EDTA and 1 mM ATP at 4°C, prewarmed for 30 s at 37°C, and the indicated concentrations of 5(S)-HPETE and 5 M AA were added. After 10 min at 37°C, 5-LO activity (LTB 4 and its all-trans isomers) was determined. All values are given as mean Ϯ S.E., n ϭ 3-5. pholipid membranes, was supplemented with various phospholipids (phosphatidylcholine, phosphatidylinositol, phosphatidylethanolamine, and phosphatidylserine; 30 g/ml each), and 5-LO activity was determined. All of the phospholipids suppressed the effects of OAG without inhibiting 5-LO in the absence of OAG ( Fig. 2A). More detailed analysis showed that PC reduced the effects of OAG in a concentration-dependent manner (IC 50 Ϸ 20 g/ml) in S100 (Fig. 2B). For purified 5-LO (also essentially devoid of endogenous phospholipids), stimulation by OAG (10 M) was almost completely blocked at 3 g/ml of exogenously added PC (Fig. 2C).
Effects of Various Glycerides on 5-LO Activity-Next, other glycerides structurally related to OAG were tested for 5-LO stimulation in PMNL-S100. In the absence of Ca 2ϩ , the monoglyceride OG, the diacylglycerol DOG, as well as the etherlinked O-alkyl-acylglycerol EAG caused 5-LO stimulatory effects in a dose-dependent manner, although the magnitude of 5-LO activation was less pronounced as compared with OAG (Fig. 3). Interestingly, SAG, which is a long chain DAG, failed to stimulate 5-LO. As observed for OAG, no 5-LO stimulation A, PMNL-S100 were incubated with 1 mM ATP and the indicated phospholipids (30 g/ml) with or without 10 M OAG in PBS plus 1 mM EDTA at 4°C for 5 min. The samples were prewarmed, and 5 M AA was added. After 10 min, 5-LO activity was determined as described under "Experimental Procedures." B, PMNL-S100 were incubated with 1 mM ATP and the indicated concentrations of PC with or without 10 M OAG in PBS plus 1 mM EDTA at 4°C for 5 min. The samples were prewarmed, and 5 M AA was added. After 10 min, 5-LO activity was determined. C, 5-LO, isolated from PMNL, was incubated as described for S100, and enzyme activity was determined. Values are given as mean Ϯ S.E., n ϭ 3-4. by OG, EAG, and DOG was detectable in the presence of Ca 2ϩ (not shown).
OAG Renders 5-LO Activity Resistant against GPx-It was found that Ca 2ϩ renders 5-LO activity resistant against inhibition by GPx-1 (26). We attempted to investigate whether OAG could mimic the effect of Ca 2ϩ , rendering 5-LO activity GPx-resistant. To restore GPx activity in cell-free assays, PMNL-S100 were supplemented with defined amounts of exogenously added GPx-1 (isolated from bovine erythrocytes), where traces of endogenous thiols derived from the intracellular environment of the PMNL serve as the co-substrate for GPx-1 to suppress 5-LO activity (26). Alternatively, PMNL-S100 containing intrinsic GPx (ϳ2.8 milliunits GPx/10 6 cells (26)) were supplemented with 1 mM GSH to reconstitute the activity of endogenous GPx enzyme by elevating the co-substrate concentration (compare Ref. 26). In agreement with our previous studies (26), in the absence of Ca 2ϩ , 5-LO activity was dose-dependently suppressed by either exogenously added GPx-1 at low thiol concentrations (Fig. 4A) or by supplementation of GSH as the co-substrate for endogenous GPx (Fig. 4B). As observed previously (26), varying amounts of 5-LO enzyme in the incubations (derived from 1, 2.5, 5, or 10 ϫ 10 6 PMNL) did not alter the degree of 5-LO inhibition by a fixed activity of GPx-1. The GPx-mediated 5-LO inhibition could be reversed by the addition of Ca 2ϩ , but OAG (30 M) also counteracted the suppressive effect of reconstituted GPx activity. It should be noted that 20 M AA was used, and at this substrate concentration, OAG is not capable of up-regulating 5-LO catalysis (compare Fig. 1E). Dose response studies (Fig. 4, C and D) revealed that 3-10 M OAG are effective to partly reverse 5-LO inhibition because of GPx activity, and 30 -100 M OAG almost completely restored 5-LO activity. By means of a GPx activity assay, it was found that OAG (up to 100 M) has no direct inhibitory effect on GPx (100 or 300 units/ml). Also, OAG is no source for LOOH, because no oxidation of the peroxide-sensitive dye 2Ј,7Ј-dichlorofluorescein diacetate, by itself, was detected (not shown). Reversal of GPx-1-mediated 5-LO inhibition was also evident for other glycerides, such as OG, DOG, EAG, and partially also for SAG (Fig. 4, E and F). Again as observed for OAG, these glycerides did not (EAG and SAG) or only slightly (OG and DOG) up-regulate 5-LO activity in the absence of GPx at 20 M AA but effectively reversed GPxmediated suppression of 5-LO.
Mutation of the Phospholipid Binding Site within the C2 Domain Abolishes the Effects of OAG on 5-LO-The C2-like domain of 5-LO binds two Ca 2ϩ ions, essentially via the residues Asn-43, Asp-44, and Glu-46 within the loop 2, and muta-tion of these three residues to alanine (loop2 mut-5LO) leads to diminished Ca 2ϩ binding and a requirement for higher Ca 2ϩ concentrations to stimulate 5-LO activity (19). In addition, the C2-like domain also binds PC and lipid membranes, which is conferred by tryptophan residues (Trp-13, -75, and -102) located in the putative Ca 2ϩ binding loops (20,21), and mutation of Trp-13, -75, and -102 to alanine causes strongly reduced affinity to PC vesicles (21).
It appeared possible that OAG acts via this putative phospholipid binding domain to stimulate 5-LO catalysis. The tryptophan residues Trp-13, -75, and -102 were mutated to alanine, referred to as 3W mut-5LO. In an ongoing study, we found that, in fact, PC is unable to stimulate the 3W mut-5LO under conditions where the activity of wt-5LO was enhanced by PC. 2 wt-5LO, loop2 mut-5LO, and 3W mut-5LO were expressed in E. coli, and stimulation of enzyme activity by OAG was examined. At 5 M AA, the activity of wt-5LO in S100 (Fig. 5A) or of purified 5-LO enzyme (Fig. 5B) was concentration-dependently stimulated by OAG (up to 10-fold) in the absence of Ca 2ϩ , as observed for 5-LO derived from PMNL. In contrast, no such stimulatory effects of OAG (0.1-100 M) were observed for the 3W mut-5LO, nor for crude enzyme in S100, nor for purified 5-LO. In control experiments, Ca 2ϩ substantially increased the enzymatic activity of 3W mut-5LO. Interestingly, the activity of the loop2 mut-5LO was also up-regulated by OAG, showing that the intact phospholipid binding site (rather than the Ca 2ϩ binding residues) is required for enzyme stimulation by glycerides.
Moreover, we assessed the ability of OAG to confer 5-LO enzyme resistance against GPx-1. As shown in Fig. 5C, Ca 2ϩ (but also OAG) restored GPx-1-suppressed product synthesis of wt-5LO. In contrast, Ca 2ϩ could only partially reverse inhibition of 3W mut-5LO, but OAG had no effect at all. Together, OAG seems to mimic the 5-LO-protecting feature of Ca 2ϩ against GPx activity and appears to mediate its effects via the C2-like domain, involving tryptophan residues that are also necessary for phospholipid and membrane binding. DISCUSSION In a previous report, we have presented that OAG, and to a minor degree also EAG, function as agonists for human isolated PMNL, stimulating 5-LO product formation by a yet undefined mechanism (30). Here we showed that OAG and other glycerides directly influence 5-LO catalysis. Thus, OAG mimics the 5-LO stimulatory effects of Ca 2ϩ in vitro, i.e. it enhances 5-LO catalysis, particularly at low substrate concentrations (Ͻ20 M), and it renders 5-LO catalysis resistant against inhibition by GPx activity. Intriguingly, various phospholipids prevent 5-LO stimulation by OAG, possibly through displacement of OAG at the putative 5-LO (phospho)lipid binding site. Along these lines, disruption also of this (phospho)lipid binding site (by replacing Trp-13, -75, and -102 by Ala, each (21)) reduces 5-LO stimulation by OAG. Hence we conclude, that OAG mediates its actions via the C2-like ␤-barrel domain by interacting at a putative (phospho)lipid binding site. Recently it was shown that the OAG-related sphingolipids C 2ceramide or ceramide 1-phosphate can interact directly with the C2 domain of the Ca 2ϩ -regulated cytosolic phospholipase A 2 , which leads to an increase in enzyme activity (35,36).
The regulation of 5-LO catalysis is complex and incompletely understood. Stimulating factors in vitro include Ca 2ϩ , phospholipids or cellular membranes, ATP, LOOH, and yet undefined leukocyte proteins (1). In intact cells, the membrane-bound 5-lipoxygenase-activating protein (18) and phosphorylation events (10,11,37)  To PMNL-S100 diluted in PBS plus 1 mM EDTA and 1 mM ATP, the indicated amounts of GPx-1 (isolated from bovine erythrocytes) were added. The samples were either prewarmed and 20 M AA with or without 2 mM CaCl 2 was added or alternatively were preincubated with 10 M OAG for 5 min at 4°C and then prewarmed, and 20 M AA was added. After 10 min, 5-LO activity was determined. 5-LO activity in the absence of GPx (control) was 25.9 Ϯ 12.5, 22.2 Ϯ 3.8, or 87.1 Ϯ 16.6 ng/10 6 cells for S100 incubated without OAG or Ca 2ϩ , with OAG, or with Ca 2ϩ , respectively. B, OAG protects 5-LO against GSH. PMNL-S100 were diluted in PBS plus 1 mM EDTA and 1 mM ATP with or without 1 mM GSH. The samples were incubated with or without Ca 2ϩ or OAG and 20 M AA, as described above, and 5-LO activity was determined. C, PMNL-S100 were diluted in PBS plus 1 mM EDTA, 1 mM ATP with or without GPx-1 (300 milliunits) and OAG, at the indicated concentrations. After 5 min at 4°C, the samples were prewarmed, 20 M AA was added, and 5-LO activity was determined. D, PMNL-S100 were diluted in PBS plus 1 mM EDTA and 1 mM ATP, with or without 1 mM GSH and the indicated concentrations of OAG. After 5 min at 4°C, the samples were prewarmed, 20 M AA was added, and 5-LO activity was determined. E, PMNL-S100 were diluted in PBS plus 1 mM EDTA and 1 mM ATP, with or without GPx-1 (300 milliunits), and the indicated glycerides (30 M) were added. After 5 min at 4°C, the samples were prewarmed, 20 M AA was added, and 5-LO activity was determined. F, PMNL-S100 were diluted in PBS plus 1 mM EDTA and 1 mM ATP, with or without 1 mM GSH and the indicated glycerides (30 M). After 5 min at 4°C, the samples were prewarmed, 20 M AA was added, and 5-LO activity was determined. Values are given as mean Ϯ S.E., n ϭ 3-4. uct synthesis, depending on the stimulus, the cell type, and the overall assay conditions. When the mechanism of 5-LO activation by OAG in intact PMNL was investigated, well known pathways, including elevation of [Ca 2ϩ ] i , involvement of mitogen-activated protein kinase, increased levels of LOOH, or an enhanced accumulation of 5-LO at the nuclear membrane, were excluded (30). This led us to speculate that OAG could directly stimulate 5-LO catalysis.
In agreement with others (27,28), OAG was unable to augment 5-LO product synthesis in an optimized in vitro activity assay (where all stimulatory co-factors were present), regardless of the purity of the enzyme. However, in the absence of Ca 2ϩ , OAG strongly enhanced 5-LO activity at the physiologically relevant concentrations (0.3-30 M) that are usually ap-plied to investigate high affinity receptors for DAGs (38). As found for Ca 2ϩ (Ref. 23 and Fig. 1D), the stimulatory effect of OAG was evident mainly at AA substrate concentrations of Ͻ20 M. In this respect, it is of interest that Ca 2ϩ lowers the K m value for the enzymatic conversion of AA (24,39). Accordingly, the elevated 5-LO product synthesis by OAG or Ca 2ϩ might be due to an increased affinity of the enzyme to its substrate.
It was suggested that Ca 2ϩ could facilitate the conversion of the active site iron in 5-LO from the inactive ferrous to the active ferric state by increasing the affinity of the enzyme for LOOH at a putative regulatory fatty acid binding site (26). For OAG, a similar mechanism is conceivable. Thus, OAG reduced the requirement of 5-LO for activating LOOH, visualized by its ability to render 5-LO activity resistant against exogenous as FIG. 5. Exchange of the phospholipid binding tryptophan residues (Trp-13, -75, and -102) by alanine prevents stimulation of 5-LO by OAG. wt-5LO, loop2 mut-5LO, and 3W mut-5LO were expressed in E. coli. S100 containing the 5-LO proteins (A) or partially purified 5-LO proteins (B) were diluted in PBS plus 1 mM EDTA, 1 mM ATP, and the indicated concentrations of OAG. After 5 min at 4°C, the samples were prewarmed, and 5 M AA were added. After 10 min, 5-LO activity was determined. 5-LO activity in the absence of OAG (control) was 16.1 Ϯ 3.5, 6.25 Ϯ 1.5, and 4.2 Ϯ 0.8 ng/ml for wt-5LO, loop2 mut-5LO, and 3W mut-5LO, respectively, in S100. For isolated enzymes, the activities of controls were 22.4 Ϯ 2, 3.2 Ϯ 1.2, and 2.8 Ϯ 0.8 ng/ml for wt-5LO, loop2 mut-5LO, and 3W mut-5LO, respectively. C, S100 containing wt-5LO (left panel) or 3W mut-5LO (right panel) were diluted in PBS plus 1 mM EDTA and 1 mM ATP, with (white bars) or without (black bars) GPx-1 (200 milliunits). After 5 min, the samples were prewarmed, and 20 M AA together with or without 2 mM CaCl 2 were added. Alternatively, OAG (30 M) was added, and after 5 min, samples were prewarmed and 20 M AA were added. After 10 min, 5-LO activity was determined. Values are given as mean Ϯ S.E., n ϭ 3-4. Student's t test: *, p Ͻ 0.05; **, p Ͻ 0.01. well as endogenous GPx activity under different experimental settings (Fig. 4, A and B). Such reducing conditions, where GPx activity determines 5-LO activity, actually reflects the situation inside the cell accompanied with a low peroxide tone. Importantly, OAG (as Ca 2ϩ ) did not inhibit GPx-1 in an in vitro assay, excluding the possibility that OAG allows 5-LO product synthesis simply by suppressing GPx activity. Together, it appears that OAG has a similar effect on 5-LO as Ca 2ϩ , stimulating the enzyme by increasing the affinity of 5-LO for LOOH. In fact, LOOH mimicked the effect of OAG (or Ca 2ϩ ), as 13(S)hydroperoxyoctadecadienoic acid reversed GPx-mediated suppression of 5-LO (26), and the addition of 5-HPETE to PMNL-S100 up-regulated 5-LO activity (Fig. 1F).
Despite these similarities, the molecular mechanisms mediating 5-LO activation appear to be different between Ca 2ϩ and OAG. First, for 5-LO stimulation by Ca 2ϩ , PC and cationic lipids further increase AA conversion (25,27,39), whereas for OAG-stimulated 5-LO, phospholipids or membranes prevent the effect of OAG. Note that in the absence of OAG and Ca 2ϩ , all phospholipids tested did not block, but rather increased, 5-LO activity per se, as observed previously by others (28,39,40). Second, the loop2 mut-5LO, lacking the ability to bind Ca 2ϩ (19) and to convert AA in the presence of GPx (26), was still activated by OAG. In contrast, mutation of the Trp-13, -75, and -102 residues, which are necessary for efficient binding to PC (21), led to an enzyme that was not stimulated by OAG and not protected against GPx activity, although Ca 2ϩ mediated both actions. Also, in contrast to Ca 2ϩ , which causes rapid loss of crude 5-LO activity within a few minutes (41), prolonged exposure (60 min) of 5-LO to OAG did not inactivate 5-LO. Finally, we were unable, in our previous study, to detect an increased binding of 5-LO to the nuclear membrane in response to OAG (30), a feature which is typically mediated by Ca 2ϩ (42)(43)(44). Together, OAG may stimulate 5-LO in two ways. First, by providing a scaffold for 5-LO, allowing it to activate the enzyme at low (Յ10 M), but not at high, substrate concentrations. Second, OAG stimulates 5-LO by increasing the affinity for LOOH (which is impaired because of the presence of GPx activity), visualized by stimulating 5-LO at 20 M AA, where OAG (in the absence of GPx) has actually no effect.
DAGs, such as OAG, bind to the so-called C1 domains of a number of proteins including PKC isoenzymes, chimaerins, PKD, RasGRPs, Munc13s, and DAG kinase ␥. These enzymes also bind and respond to phorbol esters (e.g. PMA) (for review, see Ref. 38). The cysteine-rich C1 domains possess zinc fingerlike repeats with the motif HX 12 CX 2 CX 13/14 CX 2 CX 4 HX 2 CX 7 C that bind two Zn 2ϩ ions. Binding of DAGs or phorbol esters at the C1 domain often acts synergistically with Ca 2ϩ -dependent phospholipid binding by the C2 domain, causing activation and/or cellular redistribution to different membranes of the respective enzymes. In analogy to PKC, 5-LO undergoes a Ca 2ϩ -dependent activation and redistribution to the nucleus, where 5-LO binds to the PC-rich nuclear membrane mediated by the N-terminal C2-like ␤-barrel domain. However, in contrast to PKC, 5-LO is not stimulated by PMA in vitro, and inspection of the 5-LO primary sequence revealed no motifs of a C1 domain. Also, it remains to be investigated whether 5-LO indeed is a high affinity receptor of OAG or DAGs. Nevertheless, our data suggest that OAG interacts at a phospholipid binding site within the C2-like domain, favored by the finding that phospholipids prevent the effects of OAG and that OAG fails to affect the 3W mut-5LO, lacking residues important for PC binding. Although phospholipids and OAG may act at a common site, the effects of OAG are considerably different from that of phospholipids. Thus, phospholipids alone stimulate 5-LO much less efficiently and are not able to render 5-LO resistant against GPx. Also, phospholipids (but not OAG) stimulate 5-LO in the presence of Ca 2ϩ (25,27,28,39,40).
C2 domains, first identified in PKC (45), are ϳ130 residues in length and possess the ability to bind Ca 2ϩ and phospholipids, whereas phospholipid binding is regulated by Ca 2ϩ (46). Besides phospholipids, C2 domains can bind other lipophilic molecules, such as retinoic acid and ceramides, which functionally interact with the C2 domains of PKC-␣ (47) and cytosolic phospholipase A 2 (35,36), respectively. 5-LO is built of an N-terminal ␤-barrel domain that shares structural and functional similarity with C2 domains of other enzymes and a catalytic domain where the active site iron resides (19,22). Indeed, it has been shown that the 5-LO C2-like domain has two important functions: (i) it acts as a Ca 2ϩ -dependent membrane anchor allowing the enzyme to co-localize with cytosolic phospholipase A 2 and 5-lipoxygenase-activating protein at the nuclear membrane (20,21), and (ii) it mediates Ca 2ϩ -induced enzyme activation (19), as well as Ca 2ϩ -dependent resistance of 5-LO activity against GPx (26). Also, for the closely related 12and 15-LO, the existence of N-terminal C2-like ␤-barrel domains are obvious from the primary sequence or structure (19,48). However, only the C2 domain of 5-LO has been shown to bind Ca 2ϩ with high affinity (19,21,49), and enzyme catalysis of 12-and 15-LO is much less affected by Ca 2ϩ as compared with 5-LO (50,51). According to their susceptibility for Ca 2ϩ , 12-and 15-LO are activated by OAG to a minor extent, and OAG is less efficient in restoring the GPx-suppressed enzyme activity of 12-and 15-LO.
Certainly, OAG is a synthetic DAG analogue, which, to our knowledge, has not been shown to be endogenously released from stimulated cells. However, one should note that there is no strict specificity of 5-LO for OAG, but other diverse glycerides, such as the monoacylglyceride OG and the diacylglyceride DOG, are also able to stimulate 5-LO and to protect 5-LO against GPx inhibition. Interestingly, the long chain fatty acid diacylglycerol SAG occurring in cells does not stimulate 5-LO but partially protects the enzyme against GPx, indicating that separate mechanisms are operative affecting 5-LO. Approximately 50 different DAGs have been detected in intact cells released by various phospholipases, including PLC, PC-PLC, and PLD/phosphatidate phosphohydrolase, upon the activation of PMNL by many G protein-coupled receptor ligand agonists (e.g. N-formyl-methionyl-leucyl-phenylalanine, zymosan, platelet-activating factor, complement 5a, etc). Stimulation of 5-LO by endogenously released DAGs could be of relevance for the accomplishment of substantial cellular LT synthesis elicited by agonists or various extracellular stimuli that cause only moderate and transient elevations of [Ca 2ϩ ] i (e.g. fMLP, zymosan, PAF, C5a etc). A complex pharmacological approach is currently being conducted in our laboratory to evaluate the role of endogenous (diacyl)glycerides in agonist-induced 5-LO product synthesis and to reveal the physiological relevance of 5-LO stimulation by these lipids. In analogy to PKC or cytosolic phospholipase A 2 , interaction of 5-LO with endogenous glycerides may be a mechanism complementing Ca 2ϩ -induced enzyme activation.