Purification and properties of prostaglandin D synthetase from rat brain.

The prostaglandin D synthetase system was isolated from rat brain. Prostaglandin endoperoxide synthetase solubilized from a microsomal fraction catalyzed the conversion of arachidonic acid to prostaglandin H2 in the presence of heme and tryptophan. Prostaglandin D synthetase (prostaglandin endoperoxidase-D isomerase) catalyzing the isomerization of prostaglandin H2 to prostaglandin D2 was found predominantly in a cytosol fraction and was purified to apparent homogeneity with a specific activity of 1.7 mumol/min/mg of protein at 24 degrees C. The enzyme also acted upon prostaglandin G2 and produced a compound presumed to be 15-hydroperoxy-prostaglandin D2. Glutathione was not required for the enzyme reaction, but the enzyme was stabilized by thiol compounds including glutathione. The enzyme was inhibited by p-chloromercuribenzoic acid in a reversible manner. The purified enzyme was essentially free of the glutathione S-transferase activity which was found in the cytosol of brain.     

Levuglandinyl adducts of proteins are formed via a prostaglandin H2 synthase-dependent pathway after platelet activation

The product of oxygenation of arachidonic acid by the prostaglandin H synthases (PGHS), prostaglandin H(2) (PGH(2)), undergoes rearrangement to the highly reactive gamma-ketoaldehydes, levuglandin (LG) E(2), and LGD(2). We have demonstrated previously that LGE(2) reacts with the epsilon-amine of lysine to form both the levuglandinyl-lysine Schiff base and the pyrrole-derived levuglandinyl-lysine lactam adducts. We also have reported that these levuglandinyl-lysine adducts are formed on purified PGHSs following the oxygenation of arachidonic acid. We now present evidence that the levuglandinyl-lysine lactam adduct is formed in human platelets upon activation with exogenous arachidonic acid or thrombin. After proteolytic digestion of the platelet proteins, and isolation of the adducted amino acid residues, this adduct was identified by liquid chromatography-tandem mass spectrometry. We also demonstrate that formation of these adducts is inhibited by indomethacin, a PGHS inhibitor, and is enhanced by an inhibitor of thromboxane synthase. These data establish that levuglandinyl-lysine adducts are formed via a PGHS-dependent pathway in whole cells, even in the presence of an enzyme that metabolizes PGH(2). They also demonstrate that a physiological stimulus is sufficient to lead to the lipid modification of proteins through the levuglandin pathway in human platelets.     

Acrolein: a potent modulator of lung macrophage arachidonic acid metabolism

Resting rat pulmonary alveolar macrophages exposed to acrolein were stimulated to synthesize and release thromboxane B2 and prostaglandin E2 in a dose-dependent manner. Zymosan-activated pulmonary alveolar macrophages released approximately twice as much prostaglandin E2 as thromboxane B2, whereas acrolein-activated pulmonary alveolar macrophages released 4-5 times less prostaglandin E2 than thromboxane B2. In the zymosan-stimulated pulmonary alveolar macrophages, acrolein also induced a reversal in the relative amounts of prostaglandin E2 and thromboxane B2 synthesized and released into the culture medium. This reversal was achieved by a dose-dependent reduction in prostaglandin E2 synthesis. Although phagocytosis was also inhibited in a dose-dependent manner, the reduction in prostaglandin E2 appeared to be partially independent of particle ingestion since thromboxane B2 synthesis was not affected by low doses of acrolein. In fact, high doses induced a slight enhancement in thromboxane B2 synthesis. These results suggest that acrolein selectively inhibited the enzyme, prostaglandin endoperoxide E isomerase, necessary for the conversion of the endoperoxide to prostaglandin E2. Sulfhydryl reagents such as N-ethylmaleimide and 5,5'-dithiobis (2-nitrobenzoic acid) mimicked acrolein's effects, and reduced glutathione afforded protection against the effects of acrolein. These results indicated the possible involvement of acrolein's sulfhydryl reactivity in the inhibition of the isomerase enzyme. Propionaldehyde had no effect on macrophage arachidonic acid metabolism whereas crotonaldehyde mimicked the effects of acrolein. Pulmonary macrophages were unable to reverse the acrolein effects on arachidonate metabolite synthesis after 6 h in an acrolein-free environment. These data indicated the necessity of the unsaturated carbon bond for the acrolein effects on arachidonic acid metabolism and the relative irreversibility of acrolein's reaction with the macrophage.     

Enzymatic formation of prostaglandin F2 alpha from prostaglandin H2 and D2. Purification and properties of prostaglandin F synthetase from bovine lung

Prostaglandin F synthetase from bovine lung was purified 540-fold to apparent homogeneity, as assessed by polyacrylamide gel electrophoreses and ultracentrifugation. The purified enzyme proved to be a monomeric protein with a molecular weight of about 30,500. The enzyme catalyzed not only the reduction of the 11-keto group of prostaglandin D2 but also the reduction of 9,11-endoperoxide of prostaglandin H2 and various carbonyl compounds (e.g. phenanthrenequinone). Experiments using column chromatography, polyacrylamide gel electrophoreses, immunotitration using antibody against the purified enzyme, and heat treatment indicated that three enzyme activities resided in a single protein. Although phenanthrenequinone and prostaglandin D2 competitively inhibited the prostaglandin D2 and phenanthrenequinone reductase activities, respectively, these two substrates were all but ineffective on the prostaglandin H2 (at the Km value) reductase activity up to 14-fold of those Km values. These results suggest that a single enzyme protein purified from the bovine lung catalyzes the reduction of prostaglandin D2, prostaglandin H2, and various carbonyl compounds and that prostaglandin D2 and prostaglandin H2 are metabolized at two different active sites, yielding prostaglandin F2 alpha as the reaction product.     

Arachidonic acid metabolism by perfused ram testis

1. Arachidonic acid was metabolized by lipoxygenase and prostaglandin synthetase enzymes systems in the perfused ram testis. 2. The major product of the prostaglandin synthetase was 6-keto-PGF1 alpha (6KF). 3. Addition of testosterone resulted in a significant increase in the 6KF. 4. Arachidonic acid (AA) as well as testosterone penetrated the perfused testis. 5. Both 15-HPETE and 15-HETE, the products of the 15-lipoxygenase enzyme, were detected. 6. Addition of 0.1% BSA changed the pattern of the oxidized arachidonic acid metabolism.     

Arachidonic acid is preferentially metabolized by cyclooxygenase-2 to prostacyclin and prostaglandin E2

The two cyclooxygenase isoforms, cyclooxygenase-1 and cyclooxygenase-2, both metabolize arachidonic acid to prostaglandin H2, which is subsequently processed by downstream enzymes to the various prostanoids. In the present study, we asked if the two isoforms differ in the profile of prostanoids that ultimately arise from their action on arachidonic acid. Resident peritoneal macrophages contained only cyclooxygenase-1 and synthesized (from either endogenous or exogenous arachidonic acid) a balance of four major prostanoids: prostacyclin, thromboxane A2, prostaglandin D2, and 12-hydroxyheptadecatrienoic acid. Prostaglandin E2 was a minor fifth product, although these cells efficiently converted exogenous prostaglandin H2 to prostaglandin E2. By contrast, induction of cyclooxygenase-2 with lipopol- ysaccharide resulted in the preferential production of prostacyclin and prostaglandin E2. This shift in product profile was accentuated if cyclooxygenase-1 was permanently inactivated with aspirin before cyclooxygenase-2 induction. The conversion of exogenous prostaglandin H2 to prostaglandin E2 was only modestly increased by lipopolysaccharide treatment. Thus, cyclooxygenase-2 induction leads to a shift in arachidonic acid metabolism from the production of several prostanoids with diverse effects as mediated by cyclooxygenase-1 to the preferential synthesis of two prostanoids, prostacyclin and prostaglandin E2, which evoke common effects at the cellular level.     

Solubilization and partial purification of prostaglandin endoperoxide synthetase of rabbit kidney medulla.

The microsomes of rabbit kidney medulla converted arachidonic acid into prostaglandin E2 in the presence of hemoglobin, tryptophan and glutathione as activators. When themicrosomal suspension was treated with 1% Tween 20, a solubilized enzyme was obtained which catalyzed the conversion of arachidonic acid to prostaglandins G2 and H2. The solubilized enzyme was adsorbed to and then eluted from an omega-aminooctyl Sepharose 4B column, resulting in about 10-fold purification over the microsomes. The partially purified enzyme produced predominantly prostaglandin G2 in the presence of hemoglobin, while prostaglandin H2 was produced in the presence of both hemoglobin and tryptophan. The stimulation of prostaglandin endoperoxide formation was also observed with other heme and aromatic compounds. Prostaglandin H2 synthesis was inhibited by a variety of compounds including non-steroidal anti-inflammatory drugs, thiol compounds and prostaglandin analogues with a thiol group(s).     

Prostaglandin generation from gastroduodenal mucosa: regional and species differences

Regional and species differences in prostaglandin synthesis from gastroduodenal mucosa were assessed radiometrically. In the presence of excess added arachidonic acid substrate, corporal mucosa generated more prostanoid product per DNA than did antral or duodenal mucosa whether the whole homogenate or the microsomal fraction was used as an enzyme source. This appeared to be secondary to variability in cyclooxygenase activity and could not be explained by regional differences in the activity of enzymes competing for arachidonic acid substrate, in free endogenous arachidonic acid levels, in prostaglandin catabolizing activity, or in homogenate inhibitors. The qualitative product profile differed between species but not between regions within a species.     

Characterization of the lysyl adducts of prostaglandin H-synthases that are derived from oxygenation of arachidonic acid

These investigations characterize the covalent binding of reactive products of prostaglandin H-synthases (PGHSs) to the enzyme and to other molecules. The intermediate product of oxygenation of arachidonic acid by the PGHSs, prostaglandin (PG) H2, undergoes rearrangement to the highly reactive gamma-keto aldehydes, levuglandin (LG) E2 and D2. We previously have demonstrated that LGE2 reacts with the epsilon-amine of lysine to form both the lysyl-levuglandin Shiff base and the pyrrole-derived lysyl-levuglandin lactam adducts. We now demonstrate that these lysyl-levuglandin adducts are formed on the PGHSs following the oxygenation of arachidonic acid; after reduction of the putative Schiff base, proteolytic digestion of the enzyme, and isolation of the adducted amino acid residues, these adducts were identified by liquid chromatography-tandem mass spectrometry. The reactivity of the LGs is reflected by the finding that virtually all of the LG predicted to be formed from PGH2 can be accounted for as adducts of the PGH-synthase and that oxygenation of arachidonic acid by PGH-synthases also leads to the formation of adducts of other proteins present in the reaction solution. The reactivity of the PGH-synthase adducts themselves is demonstrated by the formation of intermolecular cross-links.     

Isolation and characterization of thromboxane synthase from human platelets as a cytochrome P-450 enzyme

Thromboxane synthase from human platelets was purified to apparent homogeneity by conventional chromatographic techniques. A 423-fold enrichment over the specific content in the 100,000 X g sediment from platelet homogenates was obtained. The enzyme gave a single band on sodium dodecyl sulfate-gel electrophoresis corresponding to a monomeric molecular weight of 58,800. One heme per polypeptide chain was present, and by optical and EPR spectroscopy a close analogy to the group of cytochrome P-450 proteins was established. From its substrate prostaglandin H2, the stable end product thromboxane B2 is formed with a specific activity of 24.1 mumol min-1 mg of protein-1 which corresponds to a molecular activity of 1628 min-1. The enzyme formed 12L-hydroxy-5,8,10-heptadecatrienoic acid together with thromboxane B2 in a 1:1 ratio. Both products were identified by gas chromatography-mass spectrometry analysis. As reported previously for platelet microsomes (Ullrich, V., and Haurand, M. (1983) Adv. Prostaglandin Thromboxane Leukotriene Res. 11, 105-110), the pure hemoprotein spectrally interacts with pyridine- or imidazole-based inhibitors and for the potent inhibitor imidazo-(1,5-a)pyridine-5-hexanoic acid a stoichiometric binding to the heme was shown. Substrate analogs with a methylene group replacing the oxygen in either the 9- or 11-position caused difference spectra showing spectral shifts towards 387 and 407 nm, respectively. The identification of thromboxane synthase as a P-450 protein suggests that the heme-thiolate group of the enzyme is required to split and activate the endoperoxide bond of prostaglandin H2.     

Sulfation in vitro of mucus glycoprotein by submandibular salivary gland: effects of prostaglandin and acetylsalicylic acid

Enzymatic sulfation of mucus glycoprotein by rat submandibular salivary gland and the effect of prostaglandin and acetylsalicylic acid on this process were investigated in vitro. The sulfotransferase enzyme which catalyzes the transfer of sulfate ester group from 3'-phosphoadenosine-5'-phosphosulfate to submandibular gland mucus glycoprotein has been located in the detergent extracts of Golgi-rich membrane fraction of the gland. Optimum enzyme activity was obtained at pH 6.8 with 0.5% Triton X-100, 25 mM NaF and 4 mM MgCl2, using the desulfated glycoprotein. The enzyme was also capable of sulfation of the intact mucus glycoprotein, but the acceptor capacity of such glycoprotein was 68% lower. The apparent Km of the submandibular gland sulfotransferase for salivary mucus glycoprotein was 11.1 microM. The 35S-labeled glycoprotein product of the enzyme reaction gave in CsCl density gradient a 35S-labeled peak which coincided with that of the glycoprotein. This glycoprotein upon reductive beta-elimination yielded several acidic 35S-labeled oligosaccharide alditols which accounted for 75% of the 35S-labeled glycoprotein label. Based on the analytical data, the two most abundant oligosaccharides were identified as sulfated tri- and pentasaccharides. The submandibular gland sulfotransferase activity was stimulated by 16,16-dimethyl prostaglandin E2 and inhibited by acetylsalicylic acid. The rate of enhancement of the glycoprotein sulfation was proportional to the concentration of prostaglandin up to 2.10(-5) M, at which point a 31% increase in sulfation was attained. The inhibition of the glycoprotein sulfation by acetylsalicylic acid was proportional to the drug concentration up to 2.5.10(-4) M at which concentration a 48% reduction in the sulfotransferase activity occurred. The apparent Ki value for sulfation of salivary mucus glycoprotein in presence of acetylsalicylic acid was 58.9 microM. The results suggest that prostaglandins may play a role in salivary mucin sulfation and that this process is sensitive to such nonsteroidal anti-inflammatory agents as acetylsalicylic acid.     

Biosynthesis of 18(RD)-hydroxyeicosatetraenoic acid from arachidonic acid by microsomes of monkey seminal vesicles. Some properties of a novel fatty acid omega 3-hydroxylase and omega 3-epoxygenase

Microsomes of seminal vesicles of the cynomolgus monkey were incubated with [14C]5,8,11,14-eicosatetraenoic (arachidonic) acid and NADPH for 40 min at 37 degrees C and the products were characterized. Prostaglandins F2 alpha and E2 were the two main metabolites (approximately 52% of radioactivity), while 18(R)-hydroxy-cis-5,8,11,14-eicosatetraenoic acid (18(R)-HETE) was identified as the main, less polar product (approximately 13%). Significant biosynthesis of the 19-hydroxy or 20-hydroxy metabolites of arachidonic acid could not be detected. The formation of 18(R)-HETE was further investigated in the presence of a prostaglandin synthesis inhibitor, diclofenac sodium. The omega 3-hydroxylation was only partly supported by substituting NADH for NADPH. The hydroxyl oxygen of 18(R)-HETE was derived from the atmosphere and the omega 3-hydroxylation was inhibited by proadifen and partly inhibited by carbon monoxide. These findings suggest that 18(R)-HETE is formed by a cytochrome P-450 (P-450 omega 3). Linoleic acid and 8,11,14-eicosatetraenoic acid were also substrates of the enzyme, but stearic acid was not metabolized. 5,8,11,14,17-Eicosatetraenoic acid was oxygenated under these conditions mainly to 17,18-dihydroxy-5,8,11,14-eicosatetraenoic acid, presumably formed from 17(18)-epoxy-5,8,11,14-eicosatetraenoic acid by hydrolysis. The seminal microsomes thus seem to possess both omega 3-hydroxylase and omega 3-epoxygenase activity. These seminal vesicles also contain prostaglandin E 19-hydroxylase (Oliw, E.H., Kinn, A.-C., and Kvist, U. (1988) J. Biol. Chem. 263, 7222-7227). The presence of arachidonate omega 3-hydroxylase and prostaglandin E 19-hydroxylase was assessed in microsomes of adult and juvenile monkey livers. Arachidonic acid was metabolized extensively to diols (via epoxides), but 18-HETE could not be detected. In contrast, prostaglandin E1 was slowly hydroxylated mainly to 19-hydroxyprostaglandin E1 by both adult male and female juvenile hepatic microsomes. The results indicate that P-450 omega 3 of seminal vesicles might be a tissue-specific enzyme.     

Activation mechanism of prostaglandin endoperoxide synthetase by hemoproteins

The mechanism of the activation of prostaglandin endoperoxide synthetase by hemeproteins was investigated using the enzyme purified from bovine seminal vesicle microsomes. At pH 8, the maximal enzyme activities with methemoglobin (2 microM), indoleamine 2,3-dioxygenase (2 microM), and metmyoglobin (2 microM) were 70%, 42%, and 15% of that with 1 microM hematin. Apomyoglobin and apohemoglobin inhibited the enzyme activities caused by hemoproteins as well as that caused by hematin. The inhibition was removed by the addition of excess hematin. The dissociation of heme from hemoproteins was demonstrated by trapping the free heme with human albumin or to a DE-52 column. The dissociation of heme from methemoglobin was facilitated by increasing concentrations of arachidonic acid. The amount of heme dissociated from hemoproteins (methemoglobin, metmyoglobin, and indoleamine 2,3-dioxygenase) in the presence of arachidonic acid correlated with their stimulatory effects on the prostaglandin endoperoxide synthetase activity. Horseradish peroxidase and beef liver catalase, the hemes of which were not dissociated in the presence of arachidonic acid, were ineffective in activating prostaglandin endoperoxide synthetase. Spectrophotometric titration of prostaglandin endoperoxide synthetase with hematin demonstrated that the enzyme bound hematin at the ratio of 1 mol/mol with an association constant of 0.6 x 10(8) M-1. From these results, we conclude that hemoproteins themselves are ineffective in activating prostaglandin endoperoxide synthetase and free hematin dissociated from the hemoproteins by the interaction of arachidonic acid is the activating factor for the enzyme.     

Regional distribution of prostaglandin endoperoxide synthase studied by enzyme-linked immunoassay using monoclonal antibodies

Prostaglandin endoperoxide synthase transforms arachidonic acid to prostaglandin H2 via prostaglandin G2. The enzyme purified from bovine vesicular gland was given to mice as antigen, and monoclonal antibodies were raised by the hybridoma technique. Two species of the monoclonal antibody recognizing different sites of the enzyme were utilized to establish a peroxidase-linked immunoassay of prostaglandin endoperoxide synthase. Fab' fragment of one of the antibodies was prepared and conjugated to horseradish peroxidase. The conjugate was then bound to prostaglandin endoperoxide synthase, and the labeled enzyme was precipitated by the addition of the other antibody. The peroxidase activity of the immunoprecipitate correlated linearly with the amount of prostaglandin endoperoxide synthase. This sensitive and convenient method to determine the enzyme amount rather than the enzyme activity was utilized to extensively screen the amount of prostaglandin endoperoxide synthase in various bovine tissues. In addition to vesicular gland, platelets and kidney medulla previously known as rich enzyme sources, the immunoenzymometric assay demonstrated a high content of the enzyme in various parts of alimentary tract and a low but significant amount of enzyme in some parts of brain.     

Prostaglandin F2α produced by rabbit renal slices is not a metabolite of prostaglandin E2

Slices of rabbit renal medulla and rabbit renal papilla were incubated with a mixture of [1-¹⁴C]-arachidonic acid and [5,6,8,9,11,12,14,15-³H]-arachidonic acid. In both tissues, comparison of the isotope ratios of the radioactive products with the isotope ratio of the added arachidonic indicated that: (a) there was no discernable isotope effect in the biosynthesis of prostaglandin E₂; (b) prostaglandin F₂α was formed by reduction of prostaglandin H₂ and not by reduction of prostaglandin E₂; and (c) most of the radioactive product arose from arachidonic acid that had been incorporated into the tissue and not from the direct action of cyclooxygenase on arachidonic acid in the medium.     

A heterologous enzyme immunoassay of prostaglandin E2 using a stable enzyme-labeled hapten mimic

A sensitive heterologous enzyme immunoassay for prostaglandin E2 was developed using 9-deoxy-9-methylene-prostaglandin F2 alpha as a stable prostaglandin E2 mimic. beta-Galactosidase was conjugated to the hapten mimic. Anti-prostaglandin E2 IgG was bound to a polystyrene tube. The enzyme-labeled hapten mimic mixed with unlabeled prostaglandin E2 was allowed to react in a competitive manner with the immobilized antibody. Then, the beta-galactosidase specifically bound to the antibody was assayed fluorometrically, and the enzyme activity was correlated with the amount of unlabeled prostaglandin E2. According to the calibration curve thus obtained, prostaglandin E2 could be determined in a range of 1.2-430 fmol. Prostaglandin E2 was extracted from human urine by the use of an octadecylsilyl silica column. The crude extract contained a substance(s) which disturbed the enzyme immunoassay and gave an apparently high content of prostaglandin E2. The interfering substance was separated from prostaglandin E2 by reverse-phase high-performance liquid chromatography. The purified urinary extract was examined by the enzyme immunoassay for prostaglandin E2, and the validity of the results was confirmed by gas chromatography-selected ion monitoring.     

The effect of i-mu-ts'ao on a partially purified prostaglandin E 9-ketoreductase from swine kidney

The enzyme system, prostaglandin E 9-ketoreductase, which catalyzes the reduction of prostaglandin E to form prostaglandin F has been partially purified from swine kidney. This NADPH-linked enzyme is studied spectrophotometrically. The KM of this enzyme for prostaglandin E2 was found to be 180 microM. Studies with the partially purified enzyme indicate that prostaglandin E 9-ketoreductase is affected by a Chinese herbal medicine i-mu-ts'ao (leonurus heterophyllus sweet). An increase in the concentration of i-mu-ts'ao aqueous extract may influence the conversion of prostaglandin E2 into prostaglandin F2 alpha. This finding offers a possible explanation for the physiological role of i-mu-ts'ao when it is treated as a medicine.     

Radioimmunologic identification of prostaglandins produced by serum-stimulated mouse embryo palate mesenchyme cells

High-performance liquid chromatography and radioimmunoassay were used to identify the prostaglandins synthesized by mouse embryo palate mesenchyme cells. Serum stimulated the release of several different metabolites of arachidonic acid including 6-ketoprostaglandin F1 alpha (the stable product of prostacyclin, prostaglandin I2), prostaglandin E2 and prostaglandin F2 alpha. Compared to control cells, the serum-stimulated cells produce elevated levels of prostaglandin E2 (36-fold), 6-ketoprostaglandin F1 alpha (15-fold) and prostaglandin F2 alpha (7-fold). The acetylenic analogue of arachidonic acid, 5,8,11,14-eicosatetraynoic acid prevented this accelerated synthesis.     

The oxidation of arachidonic acid by the cyclooxygenase activity of purified prostaglandin H synthase: spin trapping of a carbon-centered free radical intermediate

The ESR spin trapping technique was used to study the first detectable radical intermediate in the oxidation of arachidonic acid by purified prostaglandin H synthase. The holoenzyme and the apoenzyme, reconstituted with either hematin or Mn2+ protoporphyrin IX, were investigated. Depending on the different types of enzyme activity present, arachidonic acid was oxidized to at least two free radicals. One of these radicals is thought to be the first ESR detectable radical intermediate in the conversion of arachidonic acid to prostaglandin G2 and was detected previously in incubations of ram seminal vesicle microsomes, which are rich in prostaglandin H synthase. The ESR findings correlated with oxygen incorporation into arachidonic acid and prostaglandin formation, where the spin trap inhibits oxygen incorporation and prostaglandin formation by apparently competing with oxygen for the carbon-centered radical. Substitution of arachidonic acid by octadeuterated (5, 6, 8, 9, 11, 12, 14, 15)-arachidonic acid confirmed that the radical adduct contained arachidonic acid that is bound to the spin trap at one of these eight positions. An attempt was made to explain the apparent time lag between the metabolic activity observed in the oxygraph measurements and the appearance of the trapped radical signals.     

A lung type prostaglandin F synthase is expressed in bovine liver: cDNA sequence and expression in E. coli.

Using the cDNA of bovine lung prostaglandin F synthase (EC 1.1.1.2) as a probe, we isolated a clone from a bovine liver cDNA library which differed in only eleven nucleotides from the probe. The corresponding protein contained three amino acid substitutions, including a leucine residue which is conserved throughout all aldo-keto reductases. We inserted the liver cDNA into expression vector pUC19 and expressed the recombinant liver enzyme in E.coli. The purified liver enzyme reduced prostaglandin H2 as well as prostaglandin D2 and various carbonyl compounds. The high relative activity against prostaglandin H2 in combination with a high Km value for prostaglandin D2 identified this liver enzyme as a lung type prostaglandin F synthase. However, the binding constant for NADPH of the liver enzyme was 3.5 fold higher than that of lung prostaglandin F synthase.