Cyclo-oxygenases-1/2 (COX-1/2) catalyse the oxygenation of AA (arachidonic acid) and related polyunsaturated fatty acids to endoperoxide precursors of prostanoids. COX-1 is referred to as a constitutive enzyme involved in haemostasis, whereas COX-2 is an inducible enzyme expressed in inflammatory diseases and cancer. The fungus Dipodascopsis uninucleata has been shown by us to convert exogenous AA into 3(R)-HETE [3(R)-hydroxy-5Z,8Z,11Z,14Z-eicosatetraenoic acid]. 3R-HETE is stereochemically identical with AA, except that a hydroxy group is attached at its C-3 position. Molecular modelling studies with 3-HETE and COX-1/2 revealed a similar enzyme–substrate structure as reported for AA and COX-1/2. Here, we report that 3-HETE is an appropriate substrate for COX-1 and -2, albeit with a lower activity of oxygenation than AA. Oxygenation of 3-HETE by COX-2 produced a novel cascade of 3-hydroxyeicosanoids, as identified with EI (electron impact)–GC–MS, LC–MS–ES (electrospray) and LC–MS–API (atmospheric pressure ionization) methods. Evidence for in vitro production of 3-hydroxy-PGE2 (3-hydroxy-prostaglandin E2) was obtained upon infection of HeLa cells with Candida albicans at an MOI (multiplicity of infection) of 100. Analogous to interaction of AA and aspirin-treated COX-2, 3-HETE was transformed by acetylated COX-2 to 3,15-di-HETE (3,15-dihydroxy-HETE), whereby C-15 showed the (R)-stereochemistry. 3-Hydroxy-PGs are potent biologically active compounds. Thus 3-hydroxy-PGE2 induced interleukin-6 gene expression via the EP3 receptor (PGE2 receptor 3) in A549 cells, and raised cAMP levels via the EP4 receptor in Jurkat cells. Moreover, 3R,15S-di-HETE triggered the opening of the K+ channel in HTM (human trabecular meshwork) cells, as measured by the patch–clamp technique. Since many fatty acid disorders are associated with an ‘escape’ of 3-hydroxy fatty acids from the β-oxidation cycle, the production of 3-hydroxyeicosanoids may be critical in modulation of effects of endogenously produced eicosanoids.

INTRODUCTION

More than a decade ago, we, together with Dr Kock's group in South Africa, discovered that a South African fungus of the Lipomycetaceae family, called Dipodascopsis uninucleata, produced a novel metabolite of AA (arachidonic acid), 3R-HETE [3(R)-hydroxy-5Z,8Z,11Z,14Z-eicosatetraenoic acid], when fed with exogenous AA. The compound carried a hydroxy group at C-3 position [1]. Unlike lipoxygenase-derived eicosanoids occurring in mammalian and plant cells, such as 5-, 9-, 11-, 12-, and 15-HETEs, 3R-HETE resulted from an incomplete β-oxidation of AA [2,3]. Without utilizing any exogenous substrate, D. uninucleata was capable of synthesizing a 3-hydroxy fatty acid, identified as 3R-HTDE (3R-hydroxy-5Z,8Z-tetradecadienoic acid), from the endogenous linoleic acid after 3 cycles of β-oxidation [2]. Formation of 3R-HTDE was also observed in AA-fed Mucor genevensis, whereby AA underwent a retroconversion into linoleic acid prior to β-oxidation [4]. Interestingly, the formation of 3R-HETE was suppressed by aspirin. As salicylates also suppressed 3R-HETE formation, COX (cyclo-oxygenase)-mediated formation of 3R-HETE could be ruled out. Since polyclonal antibodies raised against 3R-HETE cross-reacted with all 3-hydroxylated fatty acids [5], we termed all short- or long-chain 3-hydroxy fatty acids, in general, as 3-hydroxyoxylipins. Moreover, we uncovered 3-hydroxyoxylipins of different chain lengths in many lipomycetaceous species, such as Lipomyces doorenjongii, Lipomyces yarrowi, Lipomyces starkeyi etc. [6]. Immunofluorescence microscopic studies of the life cycle of D. uninucleata revealed that 3-hydroxyoxylipins serve as growth regulators during the sexual phase [5], and that aspirin prevented the onset of the reproductive cycle. 3-Hydroxyoxylipins have also been found to be associated with the ascospore ornamentation in D. uninucleata [7] and as a flocculating agent in Saccharomyces cerevisiae [8]. Recently, we showed the presence of 3-hydroxyoxylipin, identified as 3,18-di-HETE (3,18-dihydroxy-5Z,8Z,11Z,14Z-eicosatetraenoic acid), in Candida albicans isolated from the vaginas of patients suffering from recurrent candidosis [9,10].

COX-1 and COX-2 [also known as PG (prostaglandin) endoperoxide synthases 1 and 2] have been shown to react with AA consecutively via two distinct activities: (i) an oxygenase activity, which is responsible for oxygenation and cyclization of AA within the COX active site to produce PGG2; and (ii) a peroxidase activity, which causes reduction of the released PGG2 to PGH2 at the peroxidase active site [11]. PGH2 is the essential precursor of various eicosanoids, including PGs, thromboxane and prostacyclin [12]. Although COX-1 and COX-2 have 60% identical amino acid sequences, the amino acids constituting the active sites of both isoforms are fully conserved [13] and their X-ray crystallographic structures are almost superimposable. COX-1 is a constitutive enzyme responsible for PG and thromboxane production in different cells and tissues to maintain homoeostasis [14], whereas COX-2 is an inducible enzyme predominantly expressed in hyperalgesia, inflammation and cancer [15]. COX-2 is stimulated in response to various stimuli, including growth factors, cytokines and other lipid mediators [1618]. Recently, we reported the selective up-regulation of COX-2 by C. albicans [19]. Homozygous COX-1 mutant mice exhibit, in general, reduced platelet aggregation, AA-induced inflammatory responses and indomethacin-induced gastric ulceration [20] as compared with homozygous COX-2 mutant mice, which have been shown to possess a normal inflammatory reaction to AA and phorbol esters, but develop severe nephropathy [21].

The distinct biochemical actions of COX-1 and COX-2 prompted several investigators to perform the structural analysis of both enzyme isoforms with the help of different COX inhibitors [2225]. Recently, several groups have used site-directed mutagenesis and supplied significant information for a deeper understanding of AA–COX interactions. It was reported that Arg-120 in COX-1 or COX-2 binds to the carboxy group of AA [2629], with subsequent abstraction of the 13-pro-S hydrogen by Tyr-385 in the oxygenation step [30,31]. Here, the ω-end of AA was positioned in a ‘kinked’ or ‘L'-form above Ser-530, in a so-called ‘top channel region’ [32]. Moreover, it was found that the top channel does not function as a product exit route, and that the oxygenated substrate leaves the COX active site in a direction proximal to the membrane [32]. In a recent report, Marnett and co-workers [33] showed that 2-AG (2-arachidonoylglycerol) is also an appropriate substrate for COX-2, and that its metabolism is similar to that of AA.

Since the structure of 3-HETE, although basically different from that of mammalian lipoxygenase- or COX-derived mono-hydroxyeicosatetraenoic acids [2], is stereochemically almost identical with AA, we report in the present study that 3-HETE is an excellent substrate for COX-1 and COX-2, and leads to the formation of a cascade of novel bioactive 3-hydroxyeicosanoids. Based on the observations of biological actions of 3-hydroxy-PGE2 in the present study, it can be proposed without reservation that these 3-hydroxyeicosanoids do exhibit strong biological activities comparable with that of endogenous eicosanoids in various organs and tissues. The reaction of 3-HETE with COX-2 was found to be slightly stronger than with COX-1, but weaker than that of AA with COX-2 and COX-1. Since the activities of 3R- and 3S-HETE isomers towards COXs hardly differed, we performed all experiments in the present study with racemic 3-HETE, unless stated otherwise.

MATERIALS AND METHODS

Reagents

Synthetic 3R- and 3S-HETEs were synthesized according to procedures described previously [34,35]. Ovine COX-1 (specific activity >40000 units/mg) and COX-2 (specific activity >3000 units/mg) and AA were purchased from Cayman Chemicals (Ann Arbor, MI, U.S.A.). Human COX-2 (specific activity >10000 units/mg) and Gly-533 mutant COX-2 proteins (G533A, G533V and G533L) were kindly given by Dr Lawrence J. Marnett (Department of Biochemistry, Vanderbilt University Medical School, Nashville, TN, U.S.A.). [14C]AA (specific radioactivity 40–60 mCi, or 1.48–2.22 GBq/mmol) was supplied by NEN (Dreieich, Germany). Chemicals for making methyl- and silyl-derivatives of products obtained from 3-HETE–COX-1- and 3-HETE–COX-2 interactions were purchased from Sigma (Munich, Germany). Blocking peptides for EP receptors (PGE2 receptors) were purchased from Cayman Chemicals. Primers for IL-6 (interleukin 6) gene analysis were synthesized by TIB Biomol (Berlin, Germany). All other chemicals were supplied by Sigma, unless otherwise stated.

Molecular modelling

The modelling of 3R- and 3S-HETE isomers with COX-1 and COX-2 was performed analogously to that of AA and COX enzymes, as described previously by others [2229,32,33]. After conversion of Cartesian co-ordinates into internal co-ordinates, the C-3 hydroxy group was added to create 3R- and 3S-HETE using MODEL software [36]. HTR and HTS (3R-HETE and 3S-HETE when in complexes with COX respectively) were placed in active cavities of COX-1 and COX-2 with C-13 in close proximity (within 4Å) from Tyr-385 using MOLMOL software [37]. The 3-HETEs were allowed to translate and rotate around the cavity axis, and simultaneously translate along the x, y and z directions. Thus we obtained four models: HTR–COX-2 (HTR2), HTS–COX-2 (HTS2), HTR–COX-1 (HTR1) and HTS–COX-1 (HTS1). Data obtained for complexes of AA with COX-1 (ACD1) and with COX-2 (ACD2) were used for purposes of comparison. In all models of 3-HETE isomers with COXs, Tyr-385 was found in close proximity to the C-13 group to facilitate the abstraction of the 13-pro-S hydrogen, which is a prerequisite for the oxygenation of AA to PGG2.

Cell culturing

A549 lung adenocarcinoma cells were cultured in DMEM (Dulbecco's modified Eagle's medium) supplemented with 10% fetal calf serum. Cells were treated with trypsin, and 2×105 cells were seeded into each well of a 24-well plate in the same medium and allowed to adhere overnight. Jurkat T-cells were cultivated in RPMI 1640 medium supplemented with 100 units/ml penicillin and 100 μg/ml streptomycin.

COX activity

Oxygen consumption was measured at 37 °C with a Strathkelvin Oxymeter 781 (Edinburgh, Scotland, U.K.) equipped with an electrode and a thermostat-controlled respiration chamber. Aliquots of COX enzymes (20 μg) were added to 100 mM Tris/HCl buffer, pH 8.0, containing 500 μM phenol and 1 μM haematin in a final volume of 1.2 ml. Oxygen uptake was triggered by the addition of 100 μM 3-HETE or AA. The initial reaction rate was determined from the linear part of the oxygen-uptake curve.

LDH (lactate dehydrogenase) activity assay

Following incubations of Jurkat and A549 cells with 3-hydroxy-PGE2, supernatants were collected at various time points and assayed for LDH using a commercial kit (Sigma) according to the manufacturer's instructions. LDH released in the supernatant was used as an index of necrotic death.

Peroxidase activity assay

Peroxidase activity of wild-type and mutant COX-1 and COX-2 proteins was assayed spectrophotometrically (Shimadzu UV 160U; Kyoto, Japan) by measuring the initial rates of guaiacol oxidation at 436 nm. Aliquots of enzymes were added to 100 mM Tris/HCl, pH 8.0, containing 1 μM haematin and 500 μM guaiacol in a total volume of 1 ml. The reaction was started by the addition of 400 μM H2O2.

Assay of COX products

Aliquots of COX enzymes containing 1 μM haematin and 500 μM phenol were incubated for 2 min at 37 °C with 100 μM AA or 3-HETE. After termination of the reaction, products were extracted as described previously [38]. The extracted compounds were separated by RP (reverse-phase)-HPLC (Shimadzu) and analysed by GC–MS (Varian, Darmstadt, Germany).

For assaying production of PGE2 in intact cells, HeLa cells were incubated with AA or 3-HETE, as described in the Figure legends. Products were extracted as described in [38], and subjected to HPLC and GC–MS analyses [39].

RP-HPLC system

HPLC separation was performed on a Shimadzu system equipped with two pumps coupled with a Novapack column (C18 column, 250 mm×4 mm, 4 μm particle size; Macherey-Nagel, Germany) using a gradient system of acetonitrile/water/0.05% acetic acid. The acetonitrile concentration was increased from an initial concentration of 20% to 90% (final) in 20 min, and was then kept stable for a further 15 min. The flow rate was 1.0 ml/min. UV detection was performed on a diode-array detector (Shimadzu SPD-M10A) using 195 nm for PGs, 235 nm for HETEs and 278 nm for PGB2. The latter served as a standard. The fractions showing relevant peaks were collected and subjected to GC–MS or LC–MS analysis.

Chiral-phase HPLC analysis

Chiral-phase analysis of HETEs and dihydroxy-HETEs was performed on a Daicel Chiralpak AD column (Daicel, Nagoya, Japan; 4.6 mm×250 mm, 5 μm) using a solvent system of n-hexane:propan-2-ol (96:4, v/v). Authentic standards of 3R- and 3S-HETE, 11R- and 11S-HETE and 15R- and 15S-HETE were used as reference compounds.

GC–MS analysis

GC–MS analysis was performed on a Varian tandem apparatus SATURN 2000 as described previously [9]. Briefly, fractions containing relevant compounds in HPLC were dried under a nitrogen stream and reconstituted in methanol. The products were first methylated with freshly prepared diazomethane and then silylated with BSTFA [bis-(trimethylsilyl)-trifluoroacetamide] at 50 °C for 10 min. The sample was then reconstituted in n-hexane and analysed by GC–MS. For locating double bonds, methylated compounds were first hydrogenated by hydrogen gas on platinum oxide as a catalyst, and then silylated for GC–MS analysis.

LC–MS analysis

LC–MS analysis was performed on a Waters/Micromass Masslynx instrument using a Novapack C18 column (4 μm, 150 mm×4 mm; Macherey-Nagel) using the gradient mobile phase described above at a flow rate of 50 μl/min. API (atmospheric pressure ionization) was performed using nitrogen as the desolvation gas (flow rate 215 l/h) and cone gas (flow rate 212 l/h). The cone temperature was kept at 520 °C; the cone voltage was kept at 46 V. The desolvation temperature was 450 °C, and the extractor voltage was 10 V. For LC–MS–ES (electrospray) experiments, the spray voltage was −2800 V, with the nebulizer kept at 0.6 ml/min. The orifice voltage was kept at −50 V, with a collisional offset potential of 15 eV. Argon was used in the collision cell at a pressure equivalent to 150×1013 molecules/cm2.

RT (reverse transcriptase)-PCR

IL-6 cDNA was reverse-transcribed from the total RNA (1 μg) extracted using the RNAeasy extraction kit (Qiagen, Hilden, Germany) from A549 cells without or with pre-treatment of 50 nM PGE2 or 3-hydroxy-PGE2 for 24 h. PCR amplification was performed for 35 cycles using the human IL-6-gene-specific primers 5′-ATGACTTCCAAGCTGGCCGTGGCT-3′ (sense) and 5′-GAAGAGCCCTCAGGCTGGACTG-3′ (antisense). The cycling parameters were as follows: initial denaturation for 4 min at 95 °C; denaturation for 30 s at 94 °C; annealing for 60 s at 60 °C; extension for 60 s at 75 °C; and a final extension for 5 min at 75 °C. For normalization of IL-6 mRNA, amplification of GAPDH (glyceraldehyde-3-phosphate dehydrogenase) was performed using the specific primers 5′-TCGGAGTCAACGGATTTGGTCGTA-3′ (sense) and 5′-ATGGACTGTGGTCATGAGTCCTTC-3′ (antisense). The PCR products were separated on a 2% (w/v) agarose gel containing ethidium bromide, and visualized. Blocking peptide for the EP3 receptor was used to selectively antagonize the IL-6 gene expression. PCR data are representative of two separate experiments.

The density of the bands of RT-PCR was measured by scanning densitometry, and expressed as normalized values relative to GAPDH. Values represent the means of two separate PCR experiments.

Assay of cAMP release by 3-hydroxy-PGE2 from Jurkat cells

For cAMP assay, Jurkat T-cells without or with modulators for the EP2 (agonist butaprost) or EP4 (inhibitor L161982) receptor were pre-incubated in a six-well plate (106 cells/well) with 50 μM IBMX (isobutylmethylxanthine) for 30 min, and then with increasing concentrations of 3-hydroxy-PGE2 and PGE2 for 5 min. The reaction was stopped by addition of 100 μl of 70% TCA (trichloroacetic acid). Cells were then sonicated for 10 s, followed by centrifugation at 14000 rev./min (17500 g) for 10 min. The supernatant subsequently collected was extracted three times with 5 ml of ethyl ether (saturated previously with water) to remove TCA. Aliquots (50 μl) of supernatant after dilution (1:10) with PBS were used for assay with a commercial ELISA kit (Cayman Chemicals), according to the manufacturer's instructions.

Measurement of current induced by 3-HETE, 3,15-di-HETE (3,15-dihydroxy-HETE) and AA using the patch–clamp technique in the HTM (human trabecular meshwork) of the eye

HTM cells were isolated from the tissue obtained from donor eyes or from eyes enucleated for melanoma. Cells were cultured in Dulbecco's modification of DMEM supplemented with 20% fetal-calf serum. Patch–clamp recordings were made at room temperature, as described previously [40]. Briefly, a coverslip with cells was introduced under continuous superfusion with Ringer's solution at 37 °C under an inverted microscope (Axiovert 35; Zeiss, Oberkochen, Germany). Superfusion with calcium-free solution was achieved using EGTA solution containing 1 μM thapsigargin, a Ca2+-ATPase inhibitor. Iberiotoxin (100 nM) was used as a blocker of ‘maxi-K’ channels (high-conductance, calcium-activated K+ channels). The input resistance was 4–5 Ω. Potentials were referenced to the bath, so that a negative potential corresponds to a negative pipette potential. Membrane currents with and without AA, 3-HETE, 3,15- or 3,9-di-HETE were recorded using an EPC-9 (Heka Electronic, Lambrecht, Germany) patch–clamp amplifier. TIDA software for Windows was employed for electrical stimulation, data storage and data processing. Values of increases in outward current are given as multiplication (fold) of the initial current level.

RESULTS

Molecular modelling

The energy minimization of complexes HTR–COX-2, HTS–COX-2, HTR–COX-1 and HTS–COX-1 revealed that HTR and HTS complexes with COX-2 have an energetically more favourable orientation than those with COX-1 (−5740 and −5719 kcal/mol for HTR–COX-2 and HTS–COX-2 compared with −5621 and −5463 kcal/mol for HTR–COX-1 and HTS–COX-1 respectively; 1 cal≡4.184 J). This energy is indicative of the stability of the complex.

As far as conformational features of 3-HETEs are concerned, Figures 1(A) and 1(B) describe HTR and HTS in complex with COX-1 and COX-2 in an EM model. For comparison, complexes of both enzymes with AA, i.e. ACD1 and ACD2, have also been shown. C-1, C-2 and C-3 atoms in complexes of HTR and HTS with COX-1 and COX-2 were found to overlap, as in the case of AA (Figures 1A and 1B). At the ω-end, however, significant differences between 3-HETE isomers and AA were observed.

Conformational aspects of HTR and HTS with COX-2 and COX-1 in an EM model

Figure 1
Conformational aspects of HTR and HTS with COX-2 and COX-1 in an EM model

For comparison, also shown are conformations of AA in complex with COX-1 designated as ACD1 in crystal structure IDIY, and AA in complex with COX-2 designated as ACD2 in crystal structure ICVU. Four models for 3-HETEs were thus obtained: HTR–COX-1 (HTR1), HTS–COX-1 (HTS1), HTR–COX-2 (HTR2) and HTS–COX-2 (HTS2). In all cases, atoms C-1, C-2 and C-3 of each model are overlapped using the juxtaposition algorithm of IMF1. Significant conformational differences between 3-HETEs–COX-1 complexes and 3-HETEs–COX-2 complexes can be seen.

Figure 1
Conformational aspects of HTR and HTS with COX-2 and COX-1 in an EM model

For comparison, also shown are conformations of AA in complex with COX-1 designated as ACD1 in crystal structure IDIY, and AA in complex with COX-2 designated as ACD2 in crystal structure ICVU. Four models for 3-HETEs were thus obtained: HTR–COX-1 (HTR1), HTS–COX-1 (HTS1), HTR–COX-2 (HTR2) and HTS–COX-2 (HTS2). In all cases, atoms C-1, C-2 and C-3 of each model are overlapped using the juxtaposition algorithm of IMF1. Significant conformational differences between 3-HETEs–COX-1 complexes and 3-HETEs–COX-2 complexes can be seen.

Oxygenation of 3-HETE by purified COX-2 enzymes

Uptake of oxygen for the oxygenation of 3-HETE by ovine COX-1 and COX-2 was measured using a Clark electrode and compared with AA oxygenation. The rate of O2 uptake for the 3-HETE/COX-2 reaction was a little bit higher than that for 3-HETE/COX-1, but much less than that observed with AA as sub-strate (Figure 2, inset). Taking the values of O2 consumption for reactions of AA with human COX-2, ovine COX-2 and ovine COX-1 as 100%, the oxygenation rates of 3-HETE with the respective COX enzymes were comparatively lower, being 40%, 61% and 33% respectively (Figure 2). Previously, Rowlinson et al. [32] have shown that Gly-533 is crucial for conversion of AA into PGG2, which in turn is converted into PGH2 at the peroxidase active site. Thus incubation of AA with a G533A mutant enzyme led to reduced AA binding and PGG2 formation, whereas G533L and G533V mutants were totally unable to convert AA into PGG2. Analogous reactions of these mutants with 3-HETE displayed comparable activities, albeit to a far less extent (Figure 2). The peroxidase activity of the wild-type enzyme and the G533A mutant for 3-HETE as substrate was, however, similar to that for AA [32] (results not shown). Chiral-phase analysis of the oxygenated products of 3-HETE showed that both R- and S-isomers of 3-HETE were equally active as substrates (results not shown). Also, oxygenation of 3-HETE was abrogated when the incubation mixture was pre-treated with 1 mM aspirin (results not shown). Kinetic analyses of 3-HETE with ovine COXs, using 100 μM substrate, revealed Km values that were comparable to those of AA, albeit the Vmax values were only approx. 10% those of AA (Table 1). The kcat/Km values for oxygenation of 3-HETE by COX-2 were higher than for COX-1. Moreover, these were 9-fold higher for AA than for 3-HETE. Owing to a lack of double bonds in their structures, saturated 3-hydroxy-16:0 and 3-hydroxy-14:0 were not appropriate substrates for COXs (results not shown). Introduction of a methoxy group in place of the hydroxy group at the C-3 position of AA rendered 3-methoxy-AA an inappropriate substrate for oxygenation (results not shown).

Oxygenation of 3-HETE by purified ovine COX-2 and COX-1 and human COX-2 mutants

Figure 2
Oxygenation of 3-HETE by purified ovine COX-2 and COX-1 and human COX-2 mutants

Aliquots of COXs (20 μg) were added to 100 mM Tris/HCl buffer containing phenol (500 μM) and haematin (1 μM) in a final volume of 1.2 ml. Oxygen uptake was measured by the addition of 3-HETE or AA at 37 °C in a thermostat-controlled respiration chamber. The linear part of the oxygen consumption curve was used for calculating the initial reaction rate. Data for human COX-2 and ovine COX-1/2 represent three independent experiments, each performed in duplicate. The values are expressed as means±S.E.M. For human COX-2 mutants, the values are expressed as means of two independent experiments. Inset: representative original tracings of oxygen uptake in experiments of AA or 3R-HETE reacting with COX-1 and COX-2 are shown. The horizontal bar represents 10 s; the vertical bar represents 2 μM O2.

Figure 2
Oxygenation of 3-HETE by purified ovine COX-2 and COX-1 and human COX-2 mutants

Aliquots of COXs (20 μg) were added to 100 mM Tris/HCl buffer containing phenol (500 μM) and haematin (1 μM) in a final volume of 1.2 ml. Oxygen uptake was measured by the addition of 3-HETE or AA at 37 °C in a thermostat-controlled respiration chamber. The linear part of the oxygen consumption curve was used for calculating the initial reaction rate. Data for human COX-2 and ovine COX-1/2 represent three independent experiments, each performed in duplicate. The values are expressed as means±S.E.M. For human COX-2 mutants, the values are expressed as means of two independent experiments. Inset: representative original tracings of oxygen uptake in experiments of AA or 3R-HETE reacting with COX-1 and COX-2 are shown. The horizontal bar represents 10 s; the vertical bar represents 2 μM O2.

Table 1
Kinetic analysis of AA and 3R-HETE interactions with COX-2 and COX-1

Km and Vmax values were calculated from Lineweaver–Burk plot analysis (1/[S] against 1/V). Parameters used for both experiments (COX-1 and COX-2): the temperature was 37 °C; reactions were allowed to proceed in a volume of 610 μl (600 μl of 100 mM Tris/HCl, pH 8, and 10 μl of enzyme corresponding to 15 μg). Four concentrations were used for AA and 3-HETE: 1.7 μM (0.1 μl); 4.1 μM (0.25 μl); 7.0 μM (0.425 μl); and 12.3 μM (0.75 μl). Each concentration was assayed in duplicate.

 Kinetic parameters 
Enzyme used/substrate Km (μM) kcat (s−1kcat/Km (s−1·μM−1Vmax (μM O2·min−1·mg protein−1
COX-2/AA 2.1±0.2 14.7±1 12.2±0.8 
COX-2/3-HETE 1.9±0.2 1.5±0.2 0.8 1.2±0.2 
COX-1/AA 8.3±0.2 15±2 1.8 13±2 
COX-1/3-HETE 11.0±1.1 2.1±0.2 0.2 1.8±0.6 
 Kinetic parameters 
Enzyme used/substrate Km (μM) kcat (s−1kcat/Km (s−1·μM−1Vmax (μM O2·min−1·mg protein−1
COX-2/AA 2.1±0.2 14.7±1 12.2±0.8 
COX-2/3-HETE 1.9±0.2 1.5±0.2 0.8 1.2±0.2 
COX-1/AA 8.3±0.2 15±2 1.8 13±2 
COX-1/3-HETE 11.0±1.1 2.1±0.2 0.2 1.8±0.6 

Analysis of oxygenated products of the 3-HETE/COX-2 reaction

Products obtained from the oxygenation of 3-HETE with COX-2 were characterized by HPLC, EI–GC–MS–MS, LC–MS–ES and LC–MS–API methods. Purified wild-type COX-2, and mutant enzymes with or without pre-treatment of 1 mM aspirin (for 1 h), were incubated for 30 min at 37 °C with 100 μM AA and processed as described in the Materials and methods section. Samples were analysed on HPLC for eicosanoids, with standard prostanoids serving as controls. Relevant fractions of peaks related to parent compounds were collected and analysed by GC–MS or LC–MS methods. Since the hydroxy group at the C-3 position is not a phenolic hydroxy group, we found that the transformation of 3-HETE by COX-2 was identical with that of AA, except that the eicosanoids produced were now 3-hydroxyeicosanoids. Figures 3(A)–3(C) describe the HPLC profiles of spiked eicosanoids PGE2, PGF and PGD2, and novel eicosanoids 3-hydroxy-PGE2, 3-hydroxy-PGF, 3-hydroxy-PGD2, 3,11-di-HETE and 3,15-di-HETE obtained from the incubation of 3-HETE with ovine COX-2 (left panels). The ratio of 3-hydroxy-PGs to di-HETEs as measured by HPLC peak quantification was approx. 0.3. Being more polar due to the additional hydroxy group, 3-hydroxyeicosanoids displayed shorter retention times on the RP-HPLC column than the parent eicosanoids. Alkaline hydrolysis of 3-hydroxy-PGE2 with NaOH resulted in the dehydration/isomerization product 3-hydroxy-PGB2, which could be easily detected due to its UV absorption at 280 nm (Figure 3B). The corresponding EI–GC–MS and LC–MS–ES/API spectra, together with characteristic mass fragments of 490 (M+−2×OTMSi), 407 (M+−OTMSi-CH·OTMSi·C5H11) and 317 (M+−2×OTMSi-CH·OTMSi·C5H11) (where OTMSi represents the trimethylsilyl ester) are shown in Figures 3(D)–3(F), and in Table 2. Being an unstable compound, 3-hydroxy-PGH2, an intermediate and precursor of the aforementioned prostanoids, could not be detected. However, when treated with endoperoxide-reducing agents, such as sodium thiosulphate, 3-hydroxy-PGH2 was converted into 3-hydroxy-PGF, which was identified by LC–MS–ES analysis. Characteristic mass fragments were 420 (M++1), 400 (M+−1−H2O), 384 (M++1−2H2O) and 442 (M++Na+), as shown in Table 2.

Characterization by HPLC and GC– and LC–MS of 3-hydroxyeicosanoids formed following 3R-HETE–ovine-COX-2 interaction

Figure 3
Characterization by HPLC and GC– and LC–MS of 3-hydroxyeicosanoids formed following 3R-HETE–ovine-COX-2 interaction

Aliquots of COX enzyme (20 μg) containing 1 μM haematin and 300 μM phenol were incubated for 2 min at 37 °C with 100 μM 3R-HETE. After termination of the reaction with methanol/acetic acid (3:2, v/v) on ice, products were extracted as described in the Materials and methods section. The extracted compounds were separated by RP-HPLC. The peaks of interest were collected and analysed by GC–MS. Similar experiments were performed with AA as substrate to check the enzyme activity. For LC–MS, the extracted compounds after 3R-HETE–COX-2 interaction were analysed by LC–API- or LC–ES–MS. (A) HPLC chromatogram taken with a DAD (diode array detector) at a wavelength of 195 nm of ‘spiked’ eicosanoids (1 μg) PGE2 and PGD2, as well as of transformed 3-hydroxy-PGE2 and 3-hydroxy-PGD2. (B) HPLC chromatogram taken with DAD at a wavelength of 278 nm of spiked eicosanoid (1 μg) PGB2, the hydrolysis product of PGE2, as well as of 3-hydroxy-PGB2. (C) HPLC chromatogram taken with DAD at 235 nm wavelength of spiked eicosanoids (1 μg) 11-HETE and 5,15-di-HETE, as well as of transformed 3,11- and 3,15-di-HETEs. (D, E) LC–MS–APCI spectra of 3-hydroxy-PGE2 (D) with M++1 at 367 and 3-hydroxy-PGB2 (E) with M++1 at 351. (F, G) GC–MS spectra of 3,15-di-HETE (F) and 3,11-di-HETE (G) with M+−15 at 479.

Figure 3
Characterization by HPLC and GC– and LC–MS of 3-hydroxyeicosanoids formed following 3R-HETE–ovine-COX-2 interaction

Aliquots of COX enzyme (20 μg) containing 1 μM haematin and 300 μM phenol were incubated for 2 min at 37 °C with 100 μM 3R-HETE. After termination of the reaction with methanol/acetic acid (3:2, v/v) on ice, products were extracted as described in the Materials and methods section. The extracted compounds were separated by RP-HPLC. The peaks of interest were collected and analysed by GC–MS. Similar experiments were performed with AA as substrate to check the enzyme activity. For LC–MS, the extracted compounds after 3R-HETE–COX-2 interaction were analysed by LC–API- or LC–ES–MS. (A) HPLC chromatogram taken with a DAD (diode array detector) at a wavelength of 195 nm of ‘spiked’ eicosanoids (1 μg) PGE2 and PGD2, as well as of transformed 3-hydroxy-PGE2 and 3-hydroxy-PGD2. (B) HPLC chromatogram taken with DAD at a wavelength of 278 nm of spiked eicosanoid (1 μg) PGB2, the hydrolysis product of PGE2, as well as of 3-hydroxy-PGB2. (C) HPLC chromatogram taken with DAD at 235 nm wavelength of spiked eicosanoids (1 μg) 11-HETE and 5,15-di-HETE, as well as of transformed 3,11- and 3,15-di-HETEs. (D, E) LC–MS–APCI spectra of 3-hydroxy-PGE2 (D) with M++1 at 367 and 3-hydroxy-PGB2 (E) with M++1 at 351. (F, G) GC–MS spectra of 3,15-di-HETE (F) and 3,11-di-HETE (G) with M+−15 at 479.

Table 2
Characteristic mass fragments of various 3-hydroxyeicosanoids obtained from the oxygenation of 3R-HETE by COX-2

OTMSi, trimethylsilyl ester.

3-Hydroxyeicosanoid Ionization mode Characteristic mass fragment 
3R-HETE API 367 (M+−1); 349 (M+−1−H2O); 333 (M++1−2H2O); 
  315 (M++1−3H2O); 391 (M++Na+); 415 (M++2Na+
3-Hydroxy-PGE2 EI 580 [M+−(OTMSi)]; 490 [M+−2(OTMSi)]; 407 [M+−(OTMSi)−CH·(OTMSi)·C5H11]; 
  317 [M+−2(OTMSi)−CH·(OTMSi)·C5H11
3-Hydroxy-PGD2 API 369 (M++1); 351 (M++1−H2O); 333 (M++1 −2H2O); 
  315 (M++1 −3H2O); 391 (M++Na+
3-Hydroxy-PGB2 API 351 (M++1); 333 (M++1−H2O); 315 (M++1−2H2O); 
  297 (M++1−3H2O); 373 (M++Na+
3-Hydroxy-PGF API 420 (M++1); 400 (M+−1−H2O); 
  384 (M++1−2H2O); 442 (M++Na+
3,15-Di-HETE EI 479 (M+−CH3); 389 [M+−CH3−(OTMSi)]; 225 [C4H4·CH(OTMSi)·C5H11]+
  175 [CH(OTMSi)·CH2COOCH3]+; 173 [CH(OTMSi)·C5H11]+ 
3,11-Di-HETE EI 479 (M+−CH3); 225 [C4H4·CH(OTMSi)·C5H11]+
  175 [CH(OTMSi)·CH2COOCH3]+ 
3-Hydroxyeicosanoid Ionization mode Characteristic mass fragment 
3R-HETE API 367 (M+−1); 349 (M+−1−H2O); 333 (M++1−2H2O); 
  315 (M++1−3H2O); 391 (M++Na+); 415 (M++2Na+
3-Hydroxy-PGE2 EI 580 [M+−(OTMSi)]; 490 [M+−2(OTMSi)]; 407 [M+−(OTMSi)−CH·(OTMSi)·C5H11]; 
  317 [M+−2(OTMSi)−CH·(OTMSi)·C5H11
3-Hydroxy-PGD2 API 369 (M++1); 351 (M++1−H2O); 333 (M++1 −2H2O); 
  315 (M++1 −3H2O); 391 (M++Na+
3-Hydroxy-PGB2 API 351 (M++1); 333 (M++1−H2O); 315 (M++1−2H2O); 
  297 (M++1−3H2O); 373 (M++Na+
3-Hydroxy-PGF API 420 (M++1); 400 (M+−1−H2O); 
  384 (M++1−2H2O); 442 (M++Na+
3,15-Di-HETE EI 479 (M+−CH3); 389 [M+−CH3−(OTMSi)]; 225 [C4H4·CH(OTMSi)·C5H11]+
  175 [CH(OTMSi)·CH2COOCH3]+; 173 [CH(OTMSi)·C5H11]+ 
3,11-Di-HETE EI 479 (M+−CH3); 225 [C4H4·CH(OTMSi)·C5H11]+
  175 [CH(OTMSi)·CH2COOCH3]+ 

Furthermore, LC–MS–ES and LC–MS–API spectra revealed the formation of two major products of the 3-HETE/COX-2 incubations, i.e. 3,15- and 3,11-di-HETEs (Table 2), whereby 3,11-di-HETE was predominantly formed. The ratio of 3,11-di-HETE to 3,15–di-HETE was 0.6. However, aspirin-treated COX-2 altered the ratio of formation of 3,15- to 3,11-di-HETEs to 3.0, indicating an increase of 500% in favour of 3,15-di-HETE. Since acetylated mouse or human COX-2 has been shown previously to convert AA into 15R-HETE [41], we might have expected, from chiral-phase analysis of 3,11- and 3,15-di-HETEs without or with aspirin-treated COX-2, resultant (R)-stereochemistry for these compounds (results not shown). This would have been in line with several reports [4143] in which conversion of AA by wild-type human COX-2 delivered 99% of 11-HETE in the (R)-configuration and approx. 85% of 15-HETE in the (S)-configuration. However, the conversion of AA by aspirin-treated COX-2 (acetylation of Ser-530) yielded 13% of 11-HETE and 14–99% 15-HETE in the (R)-configuration, the latter depending on mutations of human COX-2 of S530T or S530V [41]. However, no formation of 11S-HETE was observed.

1 summarizes the cascade of bioactive prostanoids obtained so far from the conversion of 3-HETE by ovine COX-2.

Cascade of 3-hydroxyeicosanoids produced from the 3-HETE–COX-2 reaction

Scheme 1
Cascade of 3-hydroxyeicosanoids produced from the 3-HETE–COX-2 reaction

The compound shown in square brackets was not detected.

Scheme 1
Cascade of 3-hydroxyeicosanoids produced from the 3-HETE–COX-2 reaction

The compound shown in square brackets was not detected.

Production of 3-hydroxy-PGE2 by C. albicans-infected HeLa cells in the presence of AA

C. albicans has been reported to produce 3-HETE when incubated with exogenous AA [9]. This led us to investigate whether co-incubation of HeLa cells and C. albicans, which simulates the host–pathogen interaction in vulvovaginal candidiasis, would produce 3-hydroxyeicosanoids in vivo. Thus HeLa cells were treated with 30 μM AA for 5 min prior to infection with C. albicans for 6 h at an MOI (multiplicity of infection) of 100. After separation from C. albicans, cells were lysed and extracted according to the method of Bligh and Dyer [44]. The lipid fraction was reconstituted in n-hexane and subjected to HPLC analysis. Fractions containing 3-hydroxy-PGE2 were collected, re-extracted and treated with diazomethane and BSTFA to convert 3-hydroxy-PGE2 into its Me-TMS derivative. Upon analysis with GC–MS, we could clearly identify 3-hydroxy-PGE2 by comparing the mass spectrum with that of the parent PGE2 (Figure 4) and the characteristic mass fragments (Table 2). Further confirmation was achieved by HPLC analysis of 3-hydroxy-PGB2 obtained after alkaline hydrolysis of 3-hydroxy-PGE2, as described above (results not shown).

Evidence for the conversion of AA into 3-hydroxy-PGE2 by C. albicans-infected HeLa cells

Figure 4
Evidence for the conversion of AA into 3-hydroxy-PGE2 by C. albicans-infected HeLa cells

HeLa cells (1×107) were treated with 30 μM AA for 5 min, before infecting with C. albicans for 6 h at a MOI of 100. Cell lysates were then prepared and lipids were extracted according to the method of Bligh and Dyer [43]. Extracted compounds were separated by RP-HPLC. The peak corresponding to 3-hydroxy-PGE2 was collected and analysed by GC–MS, as described in the Materials and methods section. Upper panel: EI–mass spectrum of 3-hydroxy-PGE2 as obtained from C. albicans-infected HeLa cells. For characteristic mass fragments, see Table 2. Lower panel: EI–mass spectrum of synthetic PGE2. Characteristic mass fragments are 567 (M+−15), 492 (M+−90), 409 (M+−173) and 319 (M+−90−173). OTMS, trimethylsilyl ether.

Figure 4
Evidence for the conversion of AA into 3-hydroxy-PGE2 by C. albicans-infected HeLa cells

HeLa cells (1×107) were treated with 30 μM AA for 5 min, before infecting with C. albicans for 6 h at a MOI of 100. Cell lysates were then prepared and lipids were extracted according to the method of Bligh and Dyer [43]. Extracted compounds were separated by RP-HPLC. The peak corresponding to 3-hydroxy-PGE2 was collected and analysed by GC–MS, as described in the Materials and methods section. Upper panel: EI–mass spectrum of 3-hydroxy-PGE2 as obtained from C. albicans-infected HeLa cells. For characteristic mass fragments, see Table 2. Lower panel: EI–mass spectrum of synthetic PGE2. Characteristic mass fragments are 567 (M+−15), 492 (M+−90), 409 (M+−173) and 319 (M+−90−173). OTMS, trimethylsilyl ether.

Biological activities of 3-hydroxy-PGE2, 3-HETE and 3,15-di-HETE

Activation of IL-6 gene expression by 3-hydroxy-PGE2 from lung adenocarcinoma A549 cells

The biological actions of PGE2 are mediated by four different G-protein-coupled receptors: EP1, EP2, EP3 and EP4 [45]. Whereas EP2 and EP4 receptors signal through coupling to Gαs proteins and increasing intracellular cAMP levels, the signal transduction pathway for the EP3 receptor is mediated by Gαi protein, leading to decreased intracellular cAMP levels [46]. Human A549 lung adenocarcinoma cells express EP3 as well as EP4 receptors, whereas alveolar type II cells express only the EP4 receptor [47]. Activation of A549 cells by PGE2 also stimulates cytokine release. Thus we next determined whether 3-hydroxy-PGE2 could modify IL-6 gene expression, and whether this effect was mediated by the EP3 or the EP4 receptor. As described above, A549 cells were incubated with either vehicle or 50 nM PGE2 or 50 nM 3-hydroxy-PGE2. After a further 24 h, RNA was extracted and IL-6 mRNA expression was determined by RT-PCR. IL-6 mRNA expression was significantly enhanced by 3-hydroxy-PGE2, whereas PGE2-mediated IL-6 activation was insignificantly altered (Figures 5A and 5B). No further significant enhancement of IL-6 mRNA was observed when higher concentrations of 3-hydroxy-PGE2 (1 μM) were applied (results not shown). To identify the EP receptor subtype that mediates the IL-6 activation, we compared the effects of PGE2 or 3-hydroxy-PGE2 in cells pre-treated with EP3- or EP4-blocking peptide. Blocking peptides for the EP3 and EP4 receptors were used because, in separate immunohistochemical experiments, these peptides were found to inhibit immunostaining of the EP3 and EP4 receptors respectively (results not shown). Whereas the enhanced response to 3-hydroxy-PGE2 and feeble response to PGE2 was completely abrogated in EP3-blocked cells (Figures 5A and 5B), no difference in the response of EP4-blocked cells to PGE2 or 3-hydroxy-PGE2 was observed (results not shown). These results demonstrate that IL-6 activation by 3-hydroxy-PGE2 is mediated by EP3 receptors.

Effect of 3-hydroxy-PGE2 on IL-6 mRNA expression in A549 cells
Figure 5
Effect of 3-hydroxy-PGE2 on IL-6 mRNA expression in A549 cells

(A) A549 cells were treated with 50 nM PGE2 or 3-hydroxy-PGE2 for 24 h. Total RNA (1 μg) was extracted from A549 cells without or with pre-treatment of 50 nM PGE2 or 3-hydroxy-PGE2 for 24 h. PCR amplification using the human IL-6-gene-specific primers was performed as described in the Materials and methods section. Amplification of GAPDH was performed as a control to ensure that RNA amounts were equal. PCR data are representative of two separate experiments. (B) The density of the bands of RT-PCR was measured by scanning densitometry and expressed as normalized values to GAPDH. Values represent the means of two separate PCR experiments.

Figure 5
Effect of 3-hydroxy-PGE2 on IL-6 mRNA expression in A549 cells

(A) A549 cells were treated with 50 nM PGE2 or 3-hydroxy-PGE2 for 24 h. Total RNA (1 μg) was extracted from A549 cells without or with pre-treatment of 50 nM PGE2 or 3-hydroxy-PGE2 for 24 h. PCR amplification using the human IL-6-gene-specific primers was performed as described in the Materials and methods section. Amplification of GAPDH was performed as a control to ensure that RNA amounts were equal. PCR data are representative of two separate experiments. (B) The density of the bands of RT-PCR was measured by scanning densitometry and expressed as normalized values to GAPDH. Values represent the means of two separate PCR experiments.

Release of cAMP by 3-hydroxy-PGE2 from Jurkat cells

In a recent study, PGE2 has been shown to stimulate cAMP in Jurkat cells via EP4 receptors [48]. To investigate whether 3-hydroxy-PGE2 also enhances the release of cAMP by EP4 receptors, Jurkat cells were incubated with increasing concentrations of PGE2 and 3-hydroxy-PGE2. Although the release of cAMP was hardly affected by 10 nM 3-hydroxy-PGE2, in comparison with a 5-fold augmentation by 10 nM PGE2, a concentration of 100 nM 3-hydroxy-PGE2 raised the cAMP level almost to the same extent as did PGE2 (Figure 6). This accumulation of cAMP, however, was significantly inhibited by 100 nM EP4 inhibitor L161982 (5.44 compared with 1.76 pmol/ml for PGE2 and 6.60 compared with 2.51 pmol/ml for 3-hydroxy-PGE2), whereas stimulation of the EP2 receptor with 100 nM butaprost, or blocking of the receptor with EP2-specific peptide, did not lead to any alterations in the cAMP level (R. Ciccoli, S. Sahi, R. Deva, G. Ishdorj, S. Singh, M. P. Zafiriou and S. Nigam, unpublished work). This suggests that EP4 receptors are responsible for Gαs stimulation of adenylate cyclase and cAMP up-regulation.

Effect of 3-hydroxy-PGE2 on cAMP release from Jurkat cells
Figure 6
Effect of 3-hydroxy-PGE2 on cAMP release from Jurkat cells

Jurkat T-cells (106 cells/well) were pre-incubated with 50 μM IBMX for 30 min, followed by treatment (5 min) with 0.01 μM 3-hydroxy-PGE2 or 1 μM PGE2. After stopping the reaction with TCA, cells were sonicated for 10 s and then centrifuged at 14000 rev./min (17500 g) for 10 min. After removing the TCA by diethyl ether treatment, the supernatant (50 μl) was assayed for cAMP using a commercial ELISA kit. The values are expressed as means±S.E.M. for three independent experiments, each performed in triplicate. cAMP production by 3-hydroxy-PGE2 at a concentration of 1 μM was identical with that observed for PGE2, and highly significant over the control value (*P<0.001).

Figure 6
Effect of 3-hydroxy-PGE2 on cAMP release from Jurkat cells

Jurkat T-cells (106 cells/well) were pre-incubated with 50 μM IBMX for 30 min, followed by treatment (5 min) with 0.01 μM 3-hydroxy-PGE2 or 1 μM PGE2. After stopping the reaction with TCA, cells were sonicated for 10 s and then centrifuged at 14000 rev./min (17500 g) for 10 min. After removing the TCA by diethyl ether treatment, the supernatant (50 μl) was assayed for cAMP using a commercial ELISA kit. The values are expressed as means±S.E.M. for three independent experiments, each performed in triplicate. cAMP production by 3-hydroxy-PGE2 at a concentration of 1 μM was identical with that observed for PGE2, and highly significant over the control value (*P<0.001).

Effects of 3-HETE and 3,15-di-HETE on maxi-K channels in HTM cells

To investigate the effect of 3-HETE metabolites on ion channels using the patch–clamp technique, we applied 1 μM 3,15-di-HETE or 3-HETE or 3,9-di-HETE to HTM cells. Both 3,15-di-HETE and 3-HETE were potent, and stimulated maxi-K channels. The increase in outward current was almost 8-fold for 3,15-di-HETE and 5-fold for 3-HETE (Figure 7), the latter being comparable with that elicited by 1 μM AA. This effect could be abrogated by 100 nM iberiotoxin, a specific inhibitor of maxi-K channels. To determine whether the stimulation of maxi-K channels by 3-HETE or 3,15-di-HETE is caused by a rise in cytosolic calcium, cells were superfused with EGTA solution containing 1 μM thapsigargin (a Ca2+-ATPase inhibitor) for depletion of intra- and extra-cellular calcium. However, depletion of calcium had no effect on the stimulated outward current (results not shown). Surprisingly, 3,9-di-HETE, a product obtained by the interaction of 3-HETE and 5-lipoxygenase, did not activate maxi-K channels (Figure 7).

Enhancement by 3-hydroxyeicosanoids of current through maxi-K channels of HTM cells as measured by the patch–clamp technique
Figure 7
Enhancement by 3-hydroxyeicosanoids of current through maxi-K channels of HTM cells as measured by the patch–clamp technique

Dispersed cells were allowed to settle on coverslips, which were then superfused with Ringer's solution. AA, 3-HETE, 3,15- or 3,9-di-HETE (all at 1 μM concentration) were added without or with the maxi-K channel blocker iberiotoxin (100 nM), and outward current was recorded as described in the Materials and methods section. An initial outward current at 80 mV was taken as 100%. Results are expressed as means±S.D. of n=7 measurements. Significance: for 3-HETE compared with the control, *P<0.001; and for 3,15-di-HETE compared with AA, **P<0.01.

Figure 7
Enhancement by 3-hydroxyeicosanoids of current through maxi-K channels of HTM cells as measured by the patch–clamp technique

Dispersed cells were allowed to settle on coverslips, which were then superfused with Ringer's solution. AA, 3-HETE, 3,15- or 3,9-di-HETE (all at 1 μM concentration) were added without or with the maxi-K channel blocker iberiotoxin (100 nM), and outward current was recorded as described in the Materials and methods section. An initial outward current at 80 mV was taken as 100%. Results are expressed as means±S.D. of n=7 measurements. Significance: for 3-HETE compared with the control, *P<0.001; and for 3,15-di-HETE compared with AA, **P<0.01.

DISCUSSION

The present study describes the oxygenation by COX-2 of 3-HETE, a mimetic of AA obtained by incomplete β-oxidation of AA, discovered by us in the fungus D. uninucleata [1]. Considering the stereochemical structure of 3-HETE, which is almost identical to that of AA, it was not surprising that 3-HETE served as an excellent substrate for COX-2, but less so for COX-1. This finding was in line with the enzyme/substrate oxygenation pattern described earlier for AA/COX-2 or 2-AG/COX-2 [32,33,49]. The molecular modelling studies revealed that 3-HETE also binds in an L-shaped configuration as shown for AA, leading to the smallest C-1–C-20 and C-10–C-20 distances and ϕ angle, which demonstrates the strong folding at C-10. All complexes of 3-HETE with COX-2 demonstrated that Arg-120 and Tyr-355 in COX-2 interact with the carboxy-group end of 3R-HETE (Figure 1) by placing the 13-pro-S hydrogen next to Tyr-385 in a way similar to that described for AA [32], albeit the latter is much closer in the AA–COX-2 complex.

Analogously to previous studies of COX-2 interaction with AA or 2-AG, we found that 3-HETE was oxygenated by various COXs, albeit to a lesser extent than AA (Figure 2). Although it is also a substrate for COX-1, the activity for transformation to 3-hydroxyprostanoids was significantly diminished (results not shown). Since the hydroxy group at the C-3 position is not a phenolic hydroxy group, we found the transformation of 3-HETE by COX-2 to occur in an identical manner as of AA, except that the eicosanoids produced were now 3-hydroxy-eicosanoids. The conversion of the endoperoxide product of the 3-HETE–COX-2 interaction into various prostaglandins (Figure 3, Table 2 and 1) results in the production of a novel family of eicosanoids, which are as diverse as those produced from AA. The production of 3-hydroxy-PGE2 in the lysate of C. albicans-infected HeLa cells pre-treated with AA (Figure 4) gives clear-cut evidence of the following mechanism in candidiasis: AA released from the infected or inflamed host tissue/cell is converted by C. albicans into 3R-HETE [9,10], which serves as substrate for C. albicans-stimulated COX-2 [19] in the host tissue/cell to produce pro-inflammatory 3-hydroxy-PGE2. Strikingly, our preliminary experiments with 3-HETE and human platelets or human neutrophils have shown, instead of 3-hydroxy-thromboxane A2 and 3-hydroxy-leukotriene B4 respectively, different novel di- and trihydroxy-metabolites of 3-HETE (R. Ciccoli, R. Deva, S. Singh, M. P. Zafiriou, S. Sahi, R. Roux, G. Ishdorj and S. Nigam, unpublished work). Whether the hydroxy group at the C-3 position near the carboxy group influenced the product pattern is unclear. Molecular modelling experiments showing alignment of C-1, C-2 and C-3 groups in the large side pocket of the COX-2 enzyme argue against this assumption. Reaction of 3-HETE with COX-2 pre-treated with aspirin produced largely 3,15R-di-HETE and a minor quantity of 3,11-di-HETE. The mechanism for a shift in the product pattern 3,11-di-HETE:3,15S-di-HETE (ratio 0.6) to 3,15R-di-HETE:3,11R-di-HETE (ratio 3) is related to one described by Rowlinson et al. [32] for an AA–acetylated-COX-2 interaction. Thus 3-HETE, with a carboxy group, should bind to Arg-120 of acetylated COX-2, and the rest of the carbon chain then slips sinuously (‘like a snake’) into the enzyme cavity, having a kink at the C-13 position for pro-hydrogen abstraction by Tyr-385. As a consequence, the conformation of 3-HETE should change in order to access the top channel of acetylated COX-2. Owing to an increased volume near Arg-513, a stable conformation of the active site is developed, which facilitates 3-HETE binding to produce favourably 3,15R-di-HETE.

To investigate the biological activities of these novel eicosanoids, we isolated and purified 3-hydroxy-PGE2 as a representative compound for its biological actions on mammalian cells. RT-PCR studies on the induction of IL-6 mRNA in 24 h showed that 3-hydroxy-PGE2 is more potent in stimulating the IL-6 mRNA expression than PGE2, which hardly affects the IL-6 mRNA expression (Figures 5A and 5B). A high basal level of IL-6 found in these experiments may be due to the action of some leakage of endogenous PGE2 during the long incubation time of 24 h, albeit that no LDH was detected in the supernatant of A549 cells up to 18 h (results not shown). The differential activity of PGE2 and 3-hydroxy-PGE2 with respect to EP3 receptors, as shown by utilizing the EP3-specific blocking peptide, is in agreement with a recent study [50] that has highlighted the different affinities of PGE2 receptors for PGE2 and PGE2 metabolites in signalling pathways located downstream of adenylate cyclase up-regulation. Thus we consider that 3-hydroxy-PGE2 desensitizes the EP3 receptor to a lesser extent than does PGE2, and that the rapid desensitization by PGE2 of the EP3 receptor may thus account for significantly lower release of PGE2-stimulated IL-6 expression.

Furthermore, the stimulatory effect of 3-hydroxy-PGE2 on cAMP via EP4, but not via EP2, receptors in Jurkat cells demonstrated that this hydroxyeicosanoid imitates the action of PGE2. However, the difference in action of PGE2 and 3-hydroxy-PGE2 with respect to receptor subtype may be attributed to the pattern and level of expression of these two receptors. These differences can vary dramatically, as seen above in the case of EP3 receptor, which underwent apparently rapid desensitization. Nevertheless, it remains to be established whether the cAMP activation is accompanied by a rise in intracellular calcium, another indispensable second messenger, for eliciting biological effects of PGE2-like compounds. The strong activity of 3-hydroxy-PGE2 with respect to IL-6 and cAMP release from A549 and Jurkat cells respectively pinpoints that, in a large number of patients with genetically distinct inherited mitochondrial fatty acid β-oxidation disorders [51,52], 3-hydroxyeicosanoids may aggravate the effect of endogenous eicosanoids. We believe that the in vitro synthesis of 3-hydroxyeicosanoids shown in the present study may reflect the real situation in patients suffering from recurrent candidosis. A marked effect of a 3-HETE metabolite (3,18-di-HETE) produced by C. albicans isolated from patients with vaginal candidiasis has recently been shown to enhance the production of PGE2 in HeLa cells, which in turn up-regulates the proliferation of C. albicans [10]. However, it should be stressed that the biological effects of a large number of possible 3,n-HETEs, exemplarily shown for 3,15-di-HETE on maxi-K channels of HTM cells in the present study (Figure 7), may have diverse effects in various organisms. Moreover, 3-hydroxyeicosanoids may also compete effectively for endogenous eicosanoids in eliciting different effects.

In conclusion, we have shown that 3-HETE is an appropriate substrate for COX-2, which is almost as effective as AA and produces novel 3-hydroxyeicosanoids (1). In addition, we have shown that 3-hydroxyeicosanoids, especially 3-hydroxy-PGE2, possess strong biological activities comparable with endogenous eicosanoids, in some cases even more potent than parent compounds. The molecular structures of (R)- and (S)-isomers of 3-HETE with COX-1 and COX-2 in modelling studies showed hardly any preference for 3R-HETE over 3S-HETE with COX-2. These studies may, therefore, be effectively used for designing and developing selective COX-2 inhibitors.

This study was generously supported under the “Schwerpunktprogram” by the Deutsche Forschungsgemeinschaft, Bonn, Germany (Ni242/33-1). We thank Dr Igor Ivanov and Dr Galina Myagkova of the Lomonossov Institute for Fine Chemical Technology, Moscow, Russia for the chemical synthesis of racemic and 3R- and 3S-HETEs, as well as of other 3-HETE-derived fatty acids, Dr Lawrence J. Marnett for kindly giving human COX-2, mouse COX-2 and the G533A, G533V and G533F mutants, Merck-Frosst (Quebec, Canada) for providing L161982, Drs Rita Rosenthal and Peter Strauss (Department of Clinical Physiology, University Medical Centre Benjamin Franklin, Free University of Berlin, Germany) for patch–clamp measurements and Dr Baude (Schering AG, Berlin, Germany) for identification of 3-hydroxyeicosanoids and di-HETEs using LC–MS–ES and LC–MS–APC techniques. We also thank Dr Werner Skuballa (Schering, Berlin, Germany) for helpful discussions. R.C. is recipient of a PhD scholarship of the Free University Berlin. S. Sa is recipient of a Senior Research Fellowship from the Council of Scientific and Industrial Research (CSIR), India. We also thank Viola Claus and Agnieszka Rybak for their assistance.

Abbreviations

     
  • AA

    arachidonic acid

  •  
  • 2-AG

    2-arachidonoylglycerol

  •  
  • API

    atmospheric pressure ionization

  •  
  • COX

    cyclo-oxygenase

  •  
  • ACD

    AA–COX complex

  •  
  • BSTFA

    bis-(trimethylsilyl)-trifluoroacetamide

  •  
  • DMEM

    Dulbecco's modified Eagle's medium

  •  
  • EI

    electron impact

  •  
  • ES

    electrospray

  •  
  • GAPDH

    glyceraldehyde-3-phosphate dehydrogenase

  •  
  • 3(R)-HETE

    3(R)-hydroxy-5Z,8Z,11Z,14Z-eicosatetraenoic acid

  •  
  • 3R-HTDE

    3R-hydroxy-5Z,8Z-tetradecadienoic acid

  •  
  • 3,15-di-HETE

    etc., 3,15-dihydroxy-HETE, etc

  •  
  • HTM

    human trabecular meshwork

  •  
  • HTR

    3R-HETE when in a complex with COX

  •  
  • HTS

    3S-HETE when in a complex with COX

  •  
  • IBMX

    isobutylmethylxanthine

  •  
  • IL-6

    interleukin 6

  •  
  • LC

    liquid chromatography

  •  
  • LDH

    lactate dehydrogenase

  •  
  • MOI

    multiplicity of infection

  •  
  • PG

    prostaglandin

  •  
  • EP1/2/3/4

    PGE2 receptor 1, 2, 3 or 4 respectively

  •  
  • RP

    reverse phase

  •  
  • RT-PCR

    reverse transcriptase-PCR

  •  
  • TCA

    trichloroacetic acid

  •  
  • TMS

    trimethylsilyl

References

References
1
Van Dyk
M. S.
Kock
J. L. F.
Coetzee
D. J.
Augustyn
O. P. H.
Nigam
S.
Isolation of a novel arachidonic acid metabolite 3-hydroxy-5,8,11,14-eicosatetraenoic acid (3-HETE) from the yeast Dipodascopsis uninucleata UOFs-Y128
FEBS Lett.
1991
, vol. 
283
 (pg. 
195
-
198
)
2
Venter
P.
Kock
J. L. F.
Kumar
G. S.
Botha
A.
Coetzee
D. J.
Botes
P. J.
Bhatt
R. K.
Falck
J. R.
Schewe
T.
Nigam
S.
Production of 3R-hydroxy-polyenoic fatty acids by the yeast Dipodascopsis uninucleata
Lipids
1997
, vol. 
32
 (pg. 
1277
-
1283
)
3
Fox
S.
Hamberg
M.
Friend
J.
Ratledge
C.
Evidence for the mitochondrial biosynthesis of 3R-hydroxy-5Z,8Z,11Z,1 4Z-eicosatetraenoic acid in the yeast Dipodascopsis uninucleata
Lipids
2000
, vol. 
35
 (pg. 
1205
-
1214
)
4
Pohl
C.
Botha
A.
Kock
J. L. F.
Coetzee
D. J.
Botes
P. J.
Schewe
T.
Nigam
S.
Oxylipin formation in fungi: biotransformation of arachidonic acid to 3-hydroxy-5,8-tetradecadienoic acid by Mucor genevensis
Biochem. Biophys. Res. Commun.
1998
, vol. 
253
 (pg. 
703
-
706
)
5
Kock
J. L. F.
Venter
P.
Linke
D.
Schewe
T.
Nigam
S.
Biological dynamics and distribution of 3-hydroxy fatty acids in the yeast Dipodascopsis uninucleata as investigated by immunofluorescence microscopy. Evidence for a putative regulatory role in the sexual reproductive cycle
FEBS Lett.
1998
, vol. 
427
 (pg. 
345
-
348
)
6
Kock
J. L. F.
Strauss
C. J.
Pohl
C. H.
Nigam
S.
The distribution of 3-hydroxy-oxylipins in fungi
Prostaglandins Other Lipid Mediators
2003
, vol. 
71
 (pg. 
85
-
96
)
7
Kock
J. L. F.
Venter
P.
Smith
D. P.
Van Wyk
P. W. J.
Botes
P. J.
Coetzee
D. J.
Pohl
C. H.
Botha
A.
Riedel
K. H.
Nigam
S.
A novel oxylipin-associated ‘ghosting phenomenon’ in yeast flocculation
Antonie van Leeuwenhoek
2000
, vol. 
77
 (pg. 
401
-
406
)
8
Strauss
C. J.
Kock
J. L. F.
Van Wyk
P. W. J.
Viljoen
B. C.
Botes
P. J.
Hulse
G.
Nigam
S.
Inverse flocculation patterns in Saccharomyces cereviseae UOFS Y-2330
J. Inst. Brew.
2003
, vol. 
109
 (pg. 
3
-
7
)
9
Deva
R.
Ciccoli
R.
Schewe
T.
Kock
J. L. F.
Nigam
S.
Arachidonic acid stimulates cell growth and forms a novel oxygenated metabolite in Candida albicans
Biochim. Biophys. Acta.
2000
, vol. 
1486
 (pg. 
299
-
311
)
10
Deva
R.
Ciccoli
R.
Kock
L.
Nigam
S.
Involvement of aspirin-sensitive oxylipins in vulvovaginal candidiasis
FEMS Microbiol. Lett.
2001
, vol. 
198
 (pg. 
37
-
43
)
11
Smith
W. L.
Marnett
L. J.
Prostaglandin endoperoxide synthase: structure and catalysis
Biochim. Biophys. Acta
1991
, vol. 
1083
 (pg. 
1
-
17
)
12
Hamberg
M.
Samuelsson
B.
Detection and isolation of an endoperoxide intermediate in prostaglandin biosynthesis
Proc. Natl. Acad. Sci. U.S.A.
1973
, vol. 
70
 (pg. 
899
-
903
)
13
Wong
E.
Bayly
C.
Waterman
H. L.
Riendeau
D.
Mancini
J. A
Conversion of prostaglandin G/H synthase-1 into an enzyme sensitive to PGHS-2-selective inhibitors by a double His513→Arg and Ile523→Val mutation
J. Biol. Chem.
1997
, vol. 
272
 (pg. 
9280
-
9286
)
14
Vane
J. R.
Bakhle
Y. S.
Botting
R. M.
Cyclooxygenases 1 and 2
Annu. Rev. Pharmacol. Toxicol.
1998
, vol. 
38
 (pg. 
97
-
120
)
15
Dubois
R. N.
Abramson
S. B.
Crofford
L.
Gupta
R. A.
Simon
L. S.
Van den Putte
B. A.
Lipsky
P. E.
Cyclooxygenases in biology and disease
FASEB J.
1998
, vol. 
12
 (pg. 
1063
-
1073
)
16
Herschman
H. R.
Xie
W. L.
Reddy
S.
Inflammation, reproduction, cancer and all that. The regulation and role of the inducible prostaglandin synthase
BioEssays
1995
, vol. 
17
 (pg. 
1031
-
1037
)
17
Herschman
H. R.
Prostaglandin synthase-2
Biochim. Biophys. Acta
1996
, vol. 
1299
 (pg. 
125
-
140
)
18
Smith
W. L.
DeWitt
D. L.
Prostaglandin endoperoxide H synthases-1 and -2
Adv. Immunol.
1996
, vol. 
62
 (pg. 
167
-
215
)
19
Deva
R.
Ciccoli
R.
Nigam
S.
Candida albicans induces selectively transcriptional activation of cyclooxygenase-2 (COX-2) in HeLa cells. Pivotal roles of toll-like receptors (TLR), p38 mitogen-activated protein kinase (p38MAPK) and nuclear factor-kappaB (NF kappa B)
J. Immunol.
2003
, vol. 
171
 (pg. 
3047
-
3055
)
20
Langenbach
R.
Morham
S. G.
Tiano
H. F.
Loftin
C. D.
Ghanayem
B. I.
Chulada
P. C.
Mahler
J. F.
Lee
C. A.
Goulding
E. H.
Kluckman
K. D.
, et al. 
Prostaglandin synthase I gene disruption in mice reduces arachidonic acid-induced inflammation and indomethacin-induced gastric ulceration
Cell
1995
, vol. 
83
 (pg. 
483
-
492
)
21
Morham
S. G.
Langenbach
R.
Loftin
C. D.
Tiano
H. F.
Vouloumanos
N.
Jenette
J. C.
Mahler
J. F.
Kluckman
K. D.
Ledford
A.
Lee
C. A.
, et al. 
Prostaglandin synthase I gene disruption in mice reduces arachidonic acid-induced inflammation and indomethacin-induced gastric ulceration
Cell
1995
, vol. 
83
 (pg. 
473
-
482
)
22
Picot
D.
Loll
P. J.
Garavito
R. M.
The X-ray crystal structure of the membrane protein prostaglandin H2 synthase-1
Nature (London)
1994
, vol. 
367
 (pg. 
243
-
249
)
23
Luong
C.
Miller
A.
Barnett
J.
Chow
J.
Ramesha
C.
Browner
M. F.
Flexibility of the NSAID binding site in the structure of human cyclooxygenase-2
Nat. Struct. Biol.
1996
, vol. 
3
 (pg. 
927
-
933
)
24
Kurumbail
R. G.
Stevens
A. M.
Gierse
J. K.
McDonald
J. J.
Stegman
R. A.
Pak
J. Y.
Gildehaus
D.
Miyashiro
J. M.
Penning
T. D.
Seibert
K.
, et al. 
Structural basis for selective inhibition of cyclooxygenase-2 by anti-inflammatory agents
Nature (London)
1996
, vol. 
384
 (pg. 
644
-
648
)
25
Loll
P. J.
Picot
D.
Ekabo
O.
Garavito
R. M.
Synthesis and use of iodinated nonsteroidal antiinflammatory drug analogs as crystallographic probes of the prostaglandin H2 synthase cyclooxygenase active site
Biochemistry
1996
, vol. 
35
 (pg. 
7330
-
7340
)
26
Picot
D.
Garavito
R. M.
Prostaglandin H synthase: implications for membrane structure
FEBS Lett.
1994
, vol. 
346
 (pg. 
21
-
25
)
27
Mancini
J. A.
Riendeau
D.
Falgueyret
J. P.
Vickers
P. J.
O'Neill
G. P.
Arginine 120 of prostaglandin G/H synthase-1 is required for the inhibition by nonsteroidal anti-inflammatory drugs containing a carboxylic acid moiety
J. Biol. Chem.
1995
, vol. 
270
 (pg. 
29372
-
29377
)
28
Bhattacharyya
D. K.
Lecomte
M.
Rieke
C. J.
Garavito
R. M.
Smith
W. L.
Involvement of arginine 120, glutamate 524, and tyrosine 355 in the binding of arachidonate and 2-phenylpropionic acid inhibitors to the cyclooxygenase active site of ovine prostaglandin endoperoxide H synthase-1
J. Biol. Chem.
1996
, vol. 
271
 (pg. 
2179
-
2184
)
29
Rieke
C. J.
Mulichak
A. M.
Garavito
R. M.
Smith
W. L.
The role of arginine 120 of human prostaglandin endoperoxide H synthase-2 in the interaction with fatty acid substrates and inhibitors
J. Biol. Chem.
1999
, vol. 
274
 (pg. 
17109
-
17114
)
30
Shimokawa
T.
Kulmacz
R. J.
DeWitt
D. L.
Smith
W. L.
Tyrosine 385 of prostaglandin endoperoxide synthase is required for cyclooxygenase catalysis
J. Biol. Chem.
1990
, vol. 
265
 (pg. 
20073
-
20076
)
31
Goodwin
D. C.
Gunther
M. H.
His
L. H.
Crews
B. C.
Eling
T. E.
Mason
R. P.
Marnett
L. J.
Nitric oxide trapping of tyrosyl radicals generated during prostaglandin endoperoxide synthase turnover. Detection of the radical derivative of tyrosine 385
J. Biol. Chem.
1998
, vol. 
273
 (pg. 
8903
-
8909
)
32
Rowlinson
S. W.
Crews
B. C.
Lanzo
C. A.
Marnett
L. J.
The binding of arachidonic acid in the cyclooxygenase active site of mouse prostaglandin endoperoxide synthase-2 (COX-2)
J. Biol. Chem.
1999
, vol. 
274
 (pg. 
23305
-
23310
)
33
Kozak
K. R.
Rowlinson
S. W.
Marnett
L. J.
Oxygenation of the endocannabinoid, 2-arachidonoylglycerol, to glyceryl prostaglandins by cyclooxygenase-2
J. Biol. Chem.
2000
, vol. 
275
 (pg. 
33744
-
33749
)
34
Bhatt
R. K.
Falck
J. R.
Nigam
S.
Enantiospecific total synthesis of a novel arachidonic acid metabolite 3-hydroxyeicosatetraenoic acid
Tetrahedron Lett.
1998
, vol. 
39
 (pg. 
249
-
252
)
35
Groza
N.
Ivanov
I. V.
Romanov
S. G.
Myagkova
G. I.
Nigam
S.
A novel synthesis of 3(R)-HETE, 3(R)-HTDE and enzymatic synthesis of 3(R),15(S)-DiHETE
Tetrahedron
2002
, vol. 
58
 (pg. 
9859
-
9863
)
36
Kothekar
V.
MODEL, Molecular Building Program, Dept. of Biophysics
2001
New Delhi, India
All India Institute of Medical Sciences
37
Konradi
R.
MOLMOL, A molecule analysis and display program, Inst. fuer Molekularbiologie und Biophysik
1997
Zürich, Switzerland
Spectrospin AG, Fallenden, ETH
38
Hamberg
M.
Stereochemistry of oxygenation of linoleic acid catalyzed by prostaglandin-endoperoxide H synthase-2
Arch. Biochem. Biophys.
1998
, vol. 
349
 (pg. 
376
-
380
)
39
Marnett
L. J.
Rowlinson
S. W.
Goodwin
D. C.
Kalgutkar
A. S.
Lanzo
C. A.
Arachidonic acid oxygenation by COX-1 and COX-2
J. Biol. Chem.
1999
, vol. 
274
 (pg. 
22903
-
22906
)
40
Steinhausen
K.
Stumpff
F.
Strauss
O.
Thieme
H.
Wiederholt
M.
Influence of muscarinic agonists and tyrosine kinase inhibitors on L-type Ca2+ Channels in human and bovine trabecular meshwork cells
Exp. Eye Res.
2000
, vol. 
70
 (pg. 
285
-
293
)
41
Schneider
C.
Boeglin
W. E.
Prusakiewicz
J. J.
Rowlinson
S. W.
Marnett
L. J.
Samel
N.
Brash
A. R.
Control of prostaglandin stereochemistry at the 15-carbon by cyclooxygenases-1 and -2
J. Biol. Chem.
2002
, vol. 
277
 (pg. 
478
-
485
)
42
Lecomote
M.
Laneuville
O.
Ji
C.
DeWitt
D. L.
Smith
W. L.
Acetylation of human prostaglandin endoperoxide synthase-2 (cyclooxygenase-2) by aspirin
J. Biol. Chem.
1994
, vol. 
269
 (pg. 
13207
-
13215
)
43
Rowlinson
S. W.
Crews
B. C.
Goodwin
D. C.
Schneider
C.
Gierse
J. K. Sp.
Marnett
L. J.
Spatial requirements for 15-R-hydroxy-5Z,8Z,11Z,13E-eicosatetraenoic acid synthesis within the cyclooxygenase active site of murine COX-2
J. Biol. Chem.
1999
, vol. 
275
 (pg. 
6586
-
6591
)
44
Bligh
E. J.
Dyer
W. J.
A rapid method of total lipid extraction and purification
Can. J. Med. Sci.
1959
, vol. 
37
 (pg. 
911
-
917
)
45
Narumiya
S.
Sugimoto
S.
Ushikubi
F.
Prostanoid receptors: structures, properties, and functions
Physiol. Rev.
1999
, vol. 
79
 (pg. 
1193
-
1226
)
46
Nguyen
M. T.
Solle
M.
Audoly
L. P.
Tilley
S. L.
Stock
J. L.
McNeish
J. D.
Coffman
T. M.
Dombrowicz
D.
Koller
B. H.
Receptors and signaling mechanisms required for prostaglandin E2-mediated regulation of mast cell degranulation and IL-6 production
J. Immunol.
2002
, vol. 
169
 (pg. 
4586
-
4593
)
47
Yano
T.
Zissel
G.
Müller-Quernheim
J.
Jae-Shin
S.
Satoh
H.
Ichikawa
T.
Prostaglandin E2 reinforces the activation of ras signal pathway in lung adenocarcinoma cells via EP3
FEBS Lett.
2002
, vol. 
518
 (pg. 
154
-
158
)
48
Gerlo
S.
Verdood
P.
Gellersen
B.
Hooghe-Peters
E. L.
Kooijman
R.
Mechanism of prostaglandin (PG)-induced prolactin expression in human T cells: cooperation of two PGE2 receptor subtypes, E-prostanoid (EP)3 and EP4, via calcium- and cyclic adenosine 5′-monophosphate-mediated signaling pathways
J. Immunol.
2004
, vol. 
173
 (pg. 
5952
-
5962
)
49
Garavito
R. M.
DeWitt
D. L.
The cyclooxygenase isoforms: structural insights into the conversion of arachidonic acid to prostaglandins
Biochim. Biophys. Acta.
1999
, vol. 
1441
 (pg. 
278
-
287
)
50
Fujino
H.
West
K. A.
Regan
J. A.
Phosphorylation of glycogen synthase kinase-3 and stimulation of T-cell factor signaling following activation of EP2 and EP4 prostanoid receptors by prostaglandin E2
J. Biol. Chem.
2002
, vol. 
277
 (pg. 
2614
-
2619
)
51
Roe
C. R.
Coates
P. M.
Scriver
C. R.
Beaudet
A. L.
Sly
W. S.
Valle
D.
Mitochondrial fatty acid oxidation disorders
The Metabolic and Molecular Basis of Inherited Disease
1995
7th edn
New York
McGraw-Hill
(pg. 
1501
-
1533
)
52
Hale
D. E.
Bennet
M. J.
Fatty acid oxidation disorders: a new class of metabolic diseases
J. Pediatr.
1992
, vol. 
121
 (pg. 
1
-
11
)