Evidence of a Cyclooxygenase-related Prostaglandin Synthesis in Coral

Certain corals are rich natural sources of prostaglandins, the metabolic origin of which has remained undefined. By analogy with the lipoxygenase/allene oxide synthase pathway to jasmonic acid in plants, the presence of (8R)-lipoxygenase and allene oxide synthase in the coral Plexaura homomallasuggested a potential metabolic route to prostaglandins (Brash, A. R., Baertshi, S. W., Ingram, C.D., and Harris, T. M. (1987)J. Biol. Chem. 262, 15829–15839). Other evidence, from the Arctic coral Gersemia fruticosa, has indicated a cyclooxygenase intermediate in the biosynthesis (Varvas, K., Koljak, R., Järving, I., Pehk, T., and Samel, N. (1994) Tetrahedron Lett. 35, 8267–8270). In the present study, active preparations of G. fruticosa have been used to identify both types of arachidonic acid metabolism and specific inhibitors were used to establish the enzyme type involved in the prostaglandin biosynthesis. The synthesis of prostaglandins and (11R)-hydroxyeicosatetraenoic acid was inhibited by mammalian cyclooxygenase inhibitors (indomethacin, aspirin, and tolfenamic acid), while the formation of the products of the 8-lipoxygenase/allene oxide pathway was not affected or was increased. The specific cyclooxygenase-2 inhibitor, nimesulide, did not inhibit the synthesis of prostaglandins in coral. We conclude that coral uses two parallel routes for the initial oxidation of polyenoic acids: the cyclooxygenase route, which leads to optically active prostaglandins, and the lipoxygenase/allene oxide synthase metabolism, the role of which remains to be established. An enzyme related to mammalian cyclooxygenases is the key to prostaglandin synthesis in coral. Based on our inhibitor data, the catalytic site of this evolutionary early cyclooxygenase appears to differ significantly from both known mammalian cyclooxygenases.

and PGE 2 esters in P. homomalla amounts to 2-3% of dry weight (2,3). There have been proposals over the years that biosynthesis of prostaglandins in coral may occur by a route totally different from the vertebrate cyclooxygenase pathway (4 -6). Currently, despite the intensive studies and widespread occurrence of prostaglandins in corals, the mechanism of biosynthesis of these compounds has remained unresolved. This is mainly attributable to the inability of the coral preparations studied to form prostaglandins in vitro (4 -7).
P. homomalla and other Caribbean corals readily metabolize exogenous arachidonic acid via the (8R)-lipoxygenase pathway (6 -8). The product, (8R)-HPETE, is further converted into an allene oxide, a highly unstable epoxide (7,9). As one of the key properties of allene oxides is their ability to undergo cyclization to give cyclopentenone rings, this compound appeared to be a promising intermediate in the potential lipoxygenase pathway from arachidonic acid to prostaglandins (Scheme 1, left side) (6,7,10). This type of transformation has been observed in plants where allene oxide is an intermediate of biosynthesis of the five-membered carbon ring of the growth hormone, jasmonic acid (11). In preparations of P. homomalla, however, the enzymatic cyclization of allene oxide to prostaglandins has not been realized (6 -8).
In 1993, we reported that the Arctic soft coral Gersemia fruticosa (Octocorallia, Alcyonacea, Nephtheidae) contains optically active natural prostaglandins PGD 2 , PGE 2 , PGF 2␣ , and 15-keto PGF 2␣ (12). The content of prostaglandins of the Arctic coral (0.008% of dry weight) is several orders of magnitude lower than that of P. homomalla, but G. fruticosa has a significant ability to convert the exogenous arachidonic acid into optically active prostaglandins in vitro (12). Our further work led to the isolation of the prostaglandin endoperoxide intermediate, PGG 2 , from G. fruticosa incubations (13). It is notable that it was the 15-hydroperoxy compound PGG 2 , not PGH 2 , that accumulated in short incubations even in the presence of peroxidase reducing substrates. The nature of the enzymes involved in this pathway is the subject of the present study.
In vertebrates, prostaglandins are synthesized by prosta-glandin endoperoxide synthase (PGHS) known also as cyclooxygenase. PGHS is a hemoprotein with two distinct catalytic activities: the cyclooxygenase activity involved in the formation of PGG 2 from arachidonic acid, and the peroxidase activity, which catalyzes the reduction of PGG 2 to PGH 2 (14). There are two PGHS isozymes called PGHS-1 (constitutive) and PGHS-2 (inducible) (15). NSAIDs, such as aspirin, ibuprofen, and indomethacin, directly target cyclooxygenase activity of both isozymes (14,16). Several selective PGHS-2 inhibitors have been developed to reduce the unfavored gastrointestinal side effects associated with the inhibition of PGHS-1 by traditional NSAIDs (17,18). Cyclooxygenase type enzymes have not been identified in nonvertebrate species, despite the detection of prostaglandins in various organisms. The existence of prostaglandins in plant tissue has also been reported (reviewed in Ref. 19), but there is evidence for the catalytic oxidation of arachidonic acid into PGF 2␣ by one of the plant lipoxygenases (a non-heme iron dioxygenase) (20). It therefore remains unclear whether the lipoxygenase-or cyclooxygenase-related enzyme catalyzes prostaglandin synthesis in G. fruticosa. In the course of the present study, we discovered that preparations of G. fruticosa metabolize arachidonic acid to a number of additional metabolites, including those of the (8R)-lipoxygenase/allene oxide pathway. This enabled us to elucidate the possible relation of these arachidonic acid metabolites to the production of optically active prostaglandins in coral.

Materials and Reagents
Unlabeled arachidonic acid and biosynthetic prostaglandin standards were obtained as a generous gift from Kevelt (Tallinn, Estonia). [ 14 C]Arachidonic acid was purchased from Amersham Pharmacia Biotech; it was diluted with unlabeled arachidonic acid to the specific activity needed. Racemic HETE standards were prepared by vitamin E-controlled autooxidation (21). The (8R,15S)-diHETE standard was prepared by incubation of arachidonic acid with the acetone powder of the coral P. homomalla followed by incubation of the resulting (8R)-HETE with soybean lipoxygenase (6). Tolfenamic acid and nimesulide were bought from Cayman Chemical. Aspirin, indomethacin and SnCl 2 were purchased from Sigma. N,O-Bis-(trimethylsilyl)trifluoroacet-amide, methoxylamine hydrochloride, and pyridine were obtained from Serva. The coral G. fruticosa was collected in the Gulf of Kandalaksha at a depth of 25-30 m and water temperature about 4°C. The collected coral was frozen in liquid nitrogen within 5 min, transported on dry ice, and stored below Ϫ70°C.

Enzyme Preparation
The acetone powder of G. fruticosa was prepared by homogenization of 20 g of frozen coral with a Warning blender for 1 min at full speed in cold acetone (Ϫ20°C). The mixture was centrifuged at 3000 ϫ g at 4°C for 5 min, the supernatant was decanted, and the procedure was repeated three times on the residual solids. The skeletal elements were removed from the mixture by swirling and decanting with cold acetone. The fine solids obtained were filtered and dried in a stream of argon at room temperature. The yield was 1-1.5 g of powder. The acetone powder was divided into aliquots and stored under argon at Ϫ70°C. The enzyme preparations remained active for several months.
Incubations-15 mg of the coral acetone powder was blended with 3 ml of 50 mM Tris buffer, pH 9, and preincubated for 5 min. The reaction was initiated with 0.4 Ci of [ 14 C]arachidonic acid (final concentration, 100 M) in ethanol solution. Incubations were performed at room temperature for 10 min, after which the incubation mixture was treated with SnCl 2 (2 mg/ml), acidified, and the reaction products were extracted with three portions of ethyl acetate. When hydroperoxy compounds were studied, SnCl 2 was omitted. The combined extracts were washed with brine, dried over potassium sulfate, and evaporated at reduced pressure to dryness. The residue was subjected to TLC analysis, and the radioactivity was monitored. The total recovery of radioactivity was about 60%.
To trap the intermediates a mild reducing agent (SnCl 2 ; final concentration, 0.5 mM) was added to the incubation mixture directly before the substrate. Incubations in the presence of SnCl 2 were performed in 50 mM Tris buffer, pH 8.
For inhibition studies, the coral acetone powder was preincubated at room temperature for 5 min with various amounts of inhibitors added in a few microliters of ethanol. An equal amount of ethanol was added to the control. The reaction was initiated with 0.3 Ci of [ 14 C]arachidonic acid (final concentration, 50 M). The incubation was carried out for 10 min, followed by treatment with SnCl 2 . The products were extracted with ethyl acetate and analyzed by TLC, and the radioactivity was monitored. All experiments were carried out in triplicate.

Chromatographic Methods
Thin Layer Chromatography-TLC was performed using precoated plates (Silica Gel 60, 0.25 mm) from E. Merck (Darmstadt, Germany). A solvent system (A) of benzene:dioxane:acetic acid (10:5:0.25, v/v/v) for separation of polar products and a solvent system (B) of diethyl ether: hexane:acetic acid (3:3:0.05, v/v/v) for separation of nonpolar products were used. For the product quantification, the TLC plates were cut into zones and extracted with methanol, and the radioactivity was measured with a liquid scintillation counter. Authentic eicosanoid standards, used as markers, were visualized with an anisaldehyde spray reagent (22).
High Performance Liquid Chromatography-For identification of the mono-and dihydroxy acids formed, products were separated initially by RP-HPLC using a Beckman Ultrasphere 5-m ODS column (25 ϫ 0.46 cm) with a solvent system of methanol:water:acetic acid (80:20:0.01, v/v/v) at a flow rate of 1 ml/min. The uv spectra and the chromatograms at 205, 220, 235, and 270 nm were recorded using a Hewlett-Packard 1040A diode array detector, and the radioactivity was monitored on-line using a Radiomatic Instruments Flo-One detector. For structural characterization, the main radiolabeled peaks were methylated with ethereal diazomethane and purified by SP-HPLC using a Beckman Ultrasphere 5-m silica column (25 ϫ 0.46 cm) and the solvent system of hexane:isopropyl alcohol:acetic acid (100:2:0.1, v/v/v) at a flow rate of 1 ml/min. The steric analysis of the hydroxy products was carried out with the methyl ester derivatives by chiral-phase HPLC using a Chiralcel OB column (23).
Gas Chromatography-Mass Spectrometry-Samples were methylated with diazomethane and purified by SP-HPLC or by TLC. The keto groups were derivatized with a 2% solution of methoxylamine hydrochloride in pyridine at 60°C for 1 h. The hydroxyl groups were silylated by treatment with bis-(trimethylsilyl)trifluoroacetamide:pyridine (2:1, v/v) at room temperature for 1 h, the solvent was removed, and the residue was dissolved in hexane. Positive-ion electron impact spectra were obtained using a Hitachi M 80-B mass spectrometer operated at 70 eV. Samples were run on an RSL fused silica capillary column (10 m ϫ SCHEME 1 0.53 mm) with temperature programming from 180 to 300°C at 15°/min. [ 14 C]Arachidonic acid was metabolized by G. fruticosa acetone powder to a complex mixture of oxygenated compounds. The products were separated by TLC in solvent systems A and/or B, followed by monitoring of the radioactivity. Typically, the reaction product pattern obtained with solvent system A consisted of the four well separated polar bands (about 19 -25% of recovered radioactivity) co-migrated with natural mammalian prostaglandin standards and gave color reactions with anisaldehyde spray reagent characteristic of the corresponding prostaglandins (Fig. 1A). The prostaglandins PGD 2 , PGE 2 , PGF 2␣ , and 15-keto PGF 2␣ formed from the exogenous arachidonic acid with the G. fruticosa acetone powder were characterized earlier by HPLC, GC-MS, 13 C NMR, and optical rotation measurements in comparison with authentic standards (12). The main nonpolar products 8-KETE, 8-HPETE, and 8hydroxy-9-ketoeicosatrienoic acid (20 -22%), and unchanged arachidonic acid (30 -35%) unresolved in the first solvent system were separated using solvent system B (Fig. 1B). When [ 14 C]arachidonic acid was incubated with the coral acetone powder in the presence of a mild reducing agent to trap peroxy intermediates to the corresponding hydroxy compounds, TLC analysis with solvent system A showed three main bands (Fig.  1C). These products were completely separated by RP-HPLC (see Fig. 2 and text below). When identical oxygenations were carried out using a heat-inactivated coral enzyme (at 80°C for 40 min), no detectable radioactive products were obtained.

Identification of Lipoxygenase Products
8-HPETE-In order to detect hydroperoxides, the TLC plates eluted in solvent system B were sprayed with the ferrous thiocyanate spray reagent. One compound with R F 0.52 showed a positive reaction. This compound was scraped from the plate, extracted with methanol and reduced with NaBH 4 to get a slightly more polar product. The uv spectra of both unreduced and reduced compound had max at 235 nm, which is typical of a conjugated diene chromophore. The identity of the reduced compound as 8-HETE was confirmed by GC-MS analysis of the methyl ester Me 3 Si ether derivative. The mass-spectrum had characteristic ions at m/z 265 (base peak, represents cleavage of the C7-C8 bond); 375 (MϪ31); 391 (MϪ15), and a molecular ion at m/z 406.
8-Hydroxy-9-ketoeicosatrienoic Acid (␣-Ketol)-This arachidonic acid metabolite was identified in the band with R F 0.44 in solvent system B. The compound was scraped from the TLC plate and extracted with methanol. An aliquot of the extract was derivatized and analyzed by GC-MS. The mass spectrum of the methyl ester methoxime Me 3 Si ether derivative was identical to that of the 8-hydroxy-9-ketoeicosa-5,11,14-trienoic acid formed from arachidonic acid in P. homomalla incubations (6). This mass spectrum has characteristic ion fragments at , which are characteristic of the spectra of the same derivative of 8,9-dihydroxyeicosa-5,11,14-trienoic acid. The isolation of ␣-ketol (the major product of nonenzymatic hydrolysis of unstable allene oxide), which shared 7-9% of total radioactivity, is indicative of the occurrence of an active allene oxide synthase in G. fruticosa.
8-KETE-In addition to 8-HPETE and ␣-ketol, one more arachidonic acid metabolite, having about 10% of total radioactivity, was detected in the nonpolar region (R F 0.62 in solvent system B). The uv spectrum of this compound had max at 280 nm, which is characteristic of the conjugated dienone. The compound was identified as 8-KETE by its conversion into 8-HETE on treatment with NaBH 4 . The reduced compound

Identification of Peroxy Intermediates
In order to simplify product quantification in enzyme inhibition studies, the number of labeled oxygenated metabolites was diminished by the immediate reduction of the peroxy intermediates formed with a mild reducing agent, SnCl 2 , to the corresponding hydroxy compounds. Additionally, the prevention of hydroperoxy acids from further transformations allowed us to isolate several minor HETEs not detectable in the absence of SnCl 2 . The incubation products were examined by TLC and RP-HPLC. Typical profiles of the labeled metabolites formed are shown in Figs. 1C and 2.
Prostaglandin PGF 2␣ -The only prominent peak in the prostaglandin region had the retention time of 5.0 min, and it was co-chromatographed with PGF 2␣ standard. The structural and stereochemical identity of the coral PGF 2␣ with the mammalian's has been shown previously (12). The enantiomeric composition of the products was determined by chiral-phase HPLC of their methyl ester derivatives (23). The uv chromatograms monitored at 235 nm showed that 15-HETE was predominantly in the 15S conformation (93.6%), 11-HETE consisted mainly of R-enantiomer (89%), and 8-HETE was mostly R-enantiomer (96.5%) (Fig. 3).

Effect of Inhibitors on Arachidonic Acid Metabolism in G. fruticosa
The effect of typical mammalian cyclooxygenase inhibitors (aspirin, indomethacin, and tolfenamic acid) was tested against the product formation in coral G. fruticosa. Usually, 0.5 mM SnCl 2 was included in the incubation mixture. Experiments with indomethacin and tolfenamic acid indicated that the influence of SnCl 2 on the enzyme inhibition was not considerable.
Tolfenamic acid was found to be the most potent inhibitor of prostaglandin synthesis among the drugs tested. The IC 50 values were estimated to be 0.45 M in the presence of the reducing agent and 0.65 M when SnCl 2 was omitted. 5 M tolfenamic acid decreased the prostaglandin formation by 90% in both cases. The IC 50 values for indomethacin were determined to be between 10 and 20 M. Dose-response curves for inhibition of coral enzymes by tolfenamic acid and indomethacin are shown in Fig. 4, A and B, respectively. The irreversible cyclooxygenase inhibitor, aspirin, was found to inhibit the prostaglandin forming activity in coral with the IC 50 values of 0.6 mM. As in case of mammalian cyclooxygenases, aspirin inhibited the coral enzyme in a time-dependent manner. For inhibition of the enzyme with 200 M aspirin, the half-life (t 1/2 ) was approximately 25 min; with 300 M aspirin, t 1/2 was about 10 min. The dose-response curves and the time course for inhibition of prostaglandin synthesis by aspirin are shown in Fig. 4, C and D, respectively. Furthermore, NSAIDs blocked the (11R)-HETE synthesis in a dose-dependent manner to 50% of initial activity. The amount of (15S)-HETE was too low to determine whether this product was inhibited or not, but the accumulation of 15-HETE following aspirin treatment of the coral enzyme was not observed. The selective inhibitor of PGHS-2, nimesulide, was found to be a poor inhibitor of prostaglandin catalyzing enzyme in coral and did not affect the prostaglandin synthesis even at very high concentrations up to 1 mM. In side by side incubations, these high concentrations of nimesulide completely blocked the activity of mammalian PGHS-1 (ram seminal vesicle cyclooxygenase) for which it is a less potent inhibitor than it is for PGHS-2. The (8R)-lipoxygenase/allene oxide synthase pathway was not blocked by NSAIDs. Otherwise, the formation of 8-HETE even increased with increasing concentrations of NSAIDs (Fig. 4, A-C). DISCUSSION The results of this study show that the acetone powder of G. fruticosa converts exogenous arachidonic acid into a complex mixture of oxygenation products. One dominating group of metabolites is formed via the (8R)-lipoxygenase/allene oxide synthase and the other via the cyclooxygenase pathway. This is the first report of both biosynthetic capabilities in the same organism. Products of the (8R)-lipoxygenase pathway consist of (8R)-HPETE, 8-KETE, and ␣-ketol. When incubations were performed in the presence of a mild reducing agent, only (8R)-HETE was formed, confirming that 8-KETE and ␣-ketol are formed from 8-HPETE as a common hydroperoxide intermediate. Notably, although the co-incubation with a reducing agent blocked the allene oxide pathway (␣-ketol synthesis), it did not eliminate the prostaglandin biosynthesis in the same incubations. We also isolated and identified 8-HETE as an endogenous component of G. fruticosa. Thus, G. fruticosa is different from P. homomalla and the other gorgonians where none of these lipoxygenase products have been detected in fresh coral extracts. Otherwise, the occurrence of 8-HETE and 8-HEPE in the other Gersemia species, Gersemia rubiformis, has been documented (24).
Unlike P. homomalla and all the other corals studied before, G. fruticosa has a unique ability to convert exogenous arachidonic acid into optically active prostaglandins in vitro. The typical mammalian prostaglandins, PGE 2 , PGD 2 , PGF 2␣ , and 15-keto PGF 2␣ , isolated and identified in fresh coral extracts, were also synthesized from labeled arachidonic acid in a high total yield of 19 -25%. Neither PGA 2 nor its derivatives were detected in the incubation mixtures or coral extracts. The esters of PGA 2 are the most prominent prostaglandins in P.
powder was preincubated at room temperature for 5 min with various amounts of inhibitors added in a few microliters of ethanol. An equal amount of ethanol was added to the control. The reaction was initiated with 0.3 Ci of [ 14 C]arachidonic acid (final concentration, 50 M). The incubation was carried out for 10 min, followed by treatment with SnCl 2 . The products were extracted with ethyl acetate and analyzed by TLC as described in Fig. 1. When the inhibition of intermediates was studied, SnCl 2 (final concentration, 0. homomalla, but the latter also contains esters of PGE 2 and PGF 2␣ (25,26), and it now seems to be much more likely that the PGA derivatives are formed from the PGE derivatives rather than the other way around. The isolation of PGG 2 from short incubations of G. fruticosa gave a strong evidence that prostaglandins are synthesized via the endoperoxide pathway (13). F-type prostaglandins, the most abundant among the incubation products, can be formed from PGG 2 enzymatically or by the non-enzymatic two-electron reduction with a natural agent. However, the similarity of the product patterns obtained with active and thermally inactivated coral preparations makes the latter version more favorable. The high in vitro potency of prostaglandin synthesis of G. fruticosa is all the more striking when taking into account that its endogenous prostaglandin content is 2-3 orders lower than that of the well studied prostaglandin-rich coral P. homomalla.
The effect of specific inhibitors on the product formation was investigated. In our earlier studies reporting prostaglandin biosynthesis in G. fruticosa, we had not detected the significant inhibition by NSAIDs (12). In retrospect, this can be attributed to the high concentrations of protein required to measure activity. In the present work, we used the highly active partially purified enzyme preparations of the freshly collected G. fruticosa. This enabled us to perform inhibition studies at lower protein concentrations, and therefore minimize the loss of inhibitors due to the absorption on membrane proteins. These studies revealed inhibition patterns similar to those obtained for mammalian PGHS. NSAIDs were found to block the production of prostaglandins in G. fruticosa, with tolfenamic acid being the most potent among the NSAIDs tested (IC 50 0.45-0.65 M). In general, the IC 50 values are higher compared with those reported for purified mammalian cyclooxygenases. This may be attributed to some differences in the enzymes and also to the protein binding of the inhibitors in our experiments, which could still be a factor in the crude coral preparations employed.
The synthesis of prostaglandins in coral proved to be completely resistant to the selective PGHS-2 inhibitor, nimesulide. The highest concentrations of nimesulide (1 mM) were sufficient to cause the complete inhibition of sheep PGHS-1, yet the coral enzyme was not affected. These results indicate that the evolutionary early cyclooxygenase in coral could have significant structural differences from both mammalian PGHS-1 and PGHS-2.
While the cyclooxygenase inhibitors blocked the formation of prostaglandins, these agents actually led to an increase in the 8-lipoxygenase pathway products, presumably due to the shunting of the substrate through the lipoxygenase pathway (27). This provides another line of evidence that the (8R)lipoxygenase/allene oxide synthase pathway is not involved in the formation of prostaglandins.
In the incubations performed in the presence of SnCl 2 to trap peroxy-intermediates by the rapid reduction into the corresponding hydroxy products, the main products were PGF 2␣ and (8R)-HETE. Besides these compounds, three more products, (11R)-HETE, (15S)-HETE, and (8R,15S)-diHETE, were detected. In principle, the hydroperoxy precursors of monohydroxy acids, (11R)-HPETE and (15S)-HPETE, could be synthesized by lipoxygenases (28 -30) or, alternatively, be formed as by-products of a cyclooxygenase reaction (28,(31)(32)(33)(34). In both cases, the oxygenation is initiated by the pro-S hydrogen abstraction at C-13, followed by addition of the molecular oxygen at C-11 or C-15. Biosynthesis of (11R)-HETE has been reported in several marine invertebrates where it was not inhibited by indomethacin (28,35). In our experiments, different NSAIDs caused substantial dose-dependent inhibition of (11R)-HETE biosynthesis, suggesting that this compound was produced as a by-product during the cyclooxygenase-catalyzed prostaglandin biosynthesis. The same inhibitory effect of NSAIDs on the production of (11R)-HETE is seen with mammalian cyclooxygenases (32,34). The finding of NSAID inhibition of (11R)-HETE formation in coral provides an additional support for the conclusion that the enzyme related to mammalian cyclooxygenases accounts for the prostaglandin biosynthesis. While the (11R)-HETE synthesis in coral was inhibited by NSAIDs, its formation did not decrease as much as that of the prostaglandins, leaving open the possibility that part of this metabolite could be formed by a specific lipoxygenase. The amount of (15S)-HETE was too small for reliable determination of the effect of inhibitors. However, it was clear that, like mammalian PGHS-1, aspirin treatment of coral cyclooxygenase did not result in accumulation of 15-HETE. Whereas aspirin completely inhibits bis-oxygenation of arachidonic acid by PGHS-1, in contrast, aspirin-acetylated PGHS-2 metabolizes arachidonic acid primarily to (15R)-HETE (16,36).
Regarding the other by-products seen in incubation with SnCl 2 , the origin of (8R,15S)-diHETE is a matter of speculation. As both 8-and 15-monohydroxy acids of the correct stereochemistry are formed in the incubation, a simple explanation is that arachidonic acid is transformed to diHETE by a subsequent attack of these two lipoxygenases. This scheme is supported by the finding that indomethacin and tolfenamic acid did not affect the formation of diHETE, thus excluding the involvement of the PGHS enzyme in the oxygenation at C-15. It was shown that 15-HPETE and 15-HETE are poor substrates for the P. homomalla (8R)-lipoxygenase (37). Thus, a plausible version is that the (8R)-HETE accumulated in high concentration in the presence of SnCl 2 is further oxidized by the (15S)lipoxygenase to give the diHETE product. Alternatively, 8,15-diHETE could be formed via the double oxygenation of arachidonic acid by (8R)-lipoxygenase. This is analogous to reactions of soybean lipoxygenase (38), except that in this case both hydroxy groups would have the R-stereochemistry. The occurrence of several lipoxygenase activities with different biological functions has been reported in various marine invertebrates. P. homomalla is known to synthesize minor 11-HETE (10), but no 15-HETE formation has been found in this coral. A low 15-HETE synthesizing activity has been detected in Pseudoplexaura porosa in addition to its main (8R)-lipoxygenase activity (39).
In summary, we conclude that two distinct fatty acid metabolic routes exist in coral, i.e. both lipoxygenase and cyclooxygenase enzymes are involved in the oxidative metabolism of arachidonic acid. Prostaglandin biosynthesis proceeds similarly to that of mammals via the cyclooxygenase pathway, except that the coral enzyme exhibits very low peroxidase activity. The characterization of the coral enzyme is in progress; in polymerase chain reaction experiments with the coral cDNA using primers based on highly conserved mammalian PGHS sequences, we have detected a PGHS-related sequence with about 50% amino acid identity to each mammalian PGHS-1 and PGHS-2, confirming the existence of an enzyme related to mammalian cyclooxygenases. The allene oxide pathway is not involved in the formation of coral prostaglandins. Nonetheless, there remains the possibility that some of the prostanoids found in other species of coral such as clavulones of Clavularia viridis, could be formed via the lipoxygenase/allene oxide pathway. There remain many interesting possibilities for further study of biosynthetic pathways from arachidonic acid to unusual cyclic oxylipins in coral.
to the collection of G. fruticosa is gratefully acknowledged. We thank William E. Boeglin, Aleksander-Mati Mü ü risepp, and Milana Liiv for assistance with the mass spectrometry and Aino Vahemets for technical assistance.