Reduction of hydroperoxides in the prostaglandin biosynthetic pathway by a microsomal peroxidase.

In the cascade of reactions leading to prostaglandins, a peroxidatic enzyme reduces the hydroperoxides formed by oxygenation of unsaturated fatty acids. This manuscript describes the enzymology of a peroxidase which utilizes prostaglandin Gz, Ui-hydroperoxyprostaglandin E, 15-hydroperoxy-5,8,11,13-eicosatetraenoic acid, 12-hydroperoxy-5,8,10,14-eicosatetraenoic acid, or HzO, as substrate. A variety of radical-scavenging, reducing co-substrates stimulated the reaction and were oxidized in conjunction with hydroperoxide reduction. Enzyme-dependent co-substrate oxidation was studied qualitatively with phenol and quantitatively with cis-5-fluoro-2-methyl-l-[p-(methylthio)benzilidinelindene-3-acetic acid (sulindac sulfide). Molar equivalence between sulindac sulfide oxidation and 15hydroperoxyprostaglandin E, reduction was observed under certain conditions. The kinetics of the peroxidase reaction indicated substrate-dependent, irreversible self-deactivation. Furthermore, EPR measurements demonstrated the generation of radicals during enzymatic hydroperoxide reduction. These signals were quenched by each stimulatory co-substrate. Consequently, autodeactivation of peroxidase activity is interpreted as oxidative attack on the enzyme, a mechanism consistent with co-substrate oxidation by a highly reactive species. Stimulation of the peroxidase would, therefore, reflect the protective effect of radical-scavenging, reducing agents, leading to increased substrate turnovers prior to destructive attack. In duo, this system could provide oxidizing equivalents either for essential metabolic processes or a self-destruct mechanism in response to excessive hydroperoxide biosynthesis.

A variety of radical-scavenging, reducing co-substrates stimulated the reaction and were oxidized in conjunction with hydroperoxide reduction.
Enzyme-dependent co-substrate oxidation was studied qualitatively with phenol and quantitatively with cis-5-fluoro-2-methyl-l-[p-(methylthio)benzilidinelindene-3-acetic acid (sulindac sulfide). Molar equivalence between sulindac sulfide oxidation and 15hydroperoxyprostaglandin E, reduction was observed under certain conditions. The kinetics of the peroxidase reaction indicated substrate-dependent, irreversible self-deactivation. Furthermore, EPR measurements demonstrated the generation of radicals during enzymatic hydroperoxide reduction.
These signals were quenched by each stimulatory co-substrate. Consequently, autodeactivation of peroxidase activity is interpreted as oxidative attack on the enzyme, a mechanism consistent with co-substrate oxidation by a highly reactive species. Stimulation of the peroxidase would, therefore, reflect the protective effect of radical-scavenging, reducing agents, leading to increased substrate turnovers prior to destructive attack.
In duo, this system could provide oxidizing equivalents either for essential metabolic processes or a self-destruct mechanism in response to excessive hydroperoxide biosynthesis.
A number of enzymatic reactions are involved in the oxygenation of arachidonic acid. Two such mammalian enzymes of importance are prostaglandin cyclooxygenase (endoperoxide synthetase) (1) and the lipoxygenases (2). The former catalyzes a double dioxygenase reaction which leads to prostaglandin GB (PGG2) ' and PGHz (3,4). These endoperoxides * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ To whom correspondence and reprint requests should be addressed.
In response to substrate oxygenation (18), the cyclooxygenase undergoes natural autodeactivation, which has been interpreted as irreversible attack on the enzyme by a highly, yet indiscriminately reactive oxidant released during hydroperoxide reduction (6). During the metabolism of PGG2 by seminal vesicle microsomes, formation of a radical was detected by EPR measurements (6). It was, therefore, postulated that the oxidant was either a radical itself or was capable of inducing radical formation. Hence, attack by this oxidant may have been the process which caused irreversible enzyme deactivation, and radical-scavenging reducing agents would have stimulated the reaction by protecting the enzyme. This concept has led to the detailed studies of the peroxidase described in this paper.
The effects of these stimulatory cofactors have also been examined in cell culture systems. With mouse 3T3 fibroblasts and human platelets, cresols and phenol depressed prostaglandin biosynthesis (19). On the other hand, prostaglandin formation was stimulated by benzo(a)pyrene (a stimulator of prostaglandin cyclooxygenase (5)) in canine kidney cells (20,21) or by isoproterenol in rabbit kidney medullary cells (22). Furthermore, the anti-inflammatory activity of phenols has been postulated to depend upon their ability to scavenge oxidizing species (23).
This communication describes the enzymology of a microsomal peroxidase which preferentially reduces prostaglandin and other lipid hydroperoxides.
The kinetics, radical generating characteristics, requirements for stimulatory cofactors, and the metabolism of these reducing co-substrates are also discussed in terms of a proposed mechanism for the enzyme. by an enzyme in microsomes prepared from ram seminal vesicles has been studied using l-"'C-labeled substrates. The upper traces of Fig. 1 are radiochromatograms of the products and unreacted substrate in ether extracts from the three reaction mixtures. The PGGz was reduced to PGH2, the 15-HPE, to PGE,, and the 15-HPETE to two products, one of which was 15-hydroxy-5,8,11,13-eicosatetraenoic acid (15-HETE).
In each case, comparable incubations showed less than 5% of the enzymatic reaction when conducted either with denatured enzyme (up to 90°C for 30 min) or in the absence of enzyme.
Panel A displays the metabolism of PGG2, the purity of which was established by chromatographic characterization against a standard sample (6). In the absence of additive, 11% of the PGGZ was converted to PGH2. Both materials produced positive Fe(SCN)2 peroxide tests (26), as did 15-HPEI and 15-HPETE. Although reaction conditions leading to only minimal PGG2 reduction are displayed here in order to demonstrate the dramatic stimulatory effect of phenol (lower trace of Panel A), higher protein levels elicited greater yields of PGH2. Furthermore, with higher protein levels or longer incubation times, products of PGH2 metabolism (&keto-PGF,,, , PGEZ, PGFZ,,) were observed near the origin but were not resolved in this chromatography system. The enzymatic reduction of 15-HPEI to PGE, is demonstrated in Fig. 1B. As in Panel A, these peaks were characterized by chromatographic comparison against standards. In the reaction with no additive, about 40% of the 15-HPEI was converted to PGE,. The last traces (Panel C) show the metabolism of 15-HPETE.
The peaks designated 15-HPETE and 15-HETE co-chromatographed with samples prepared and characterized as described under "Methods." In the upper trace, 32% of the 15-HPETE was converted to 15-HETE. However, a significant quantity of another more polar product was also formed. It was established that this second product was not a peroxide. Since no 15-HETE metabolism was detected under these conditions, it evolved enzymatically from 15-HPETE rather than from secondary breakdown of 15-HETE.
By conducting all three reactions under identical conditions for 30 s at 24 pM substrate and 1.2 mg of protein/ml, the relative reactivities of the hydroperoxides have been established. Based on a value of 100% for PGGZ, 15-HPE,, and 15-HPETE reacted 77% and 105%, respectively. Stimulation of the Peroxidase by Co-substrates-The effects of several materials were examined on the hydroperoxide reduction reactions shown in Fig. 1. For example, the outcome of incubations containing phenol are depicted in the three bottom traces of that figure. Increased conversion does not necessarily reflect a change in initial rate, since, due to deactivation of the enzyme (see Fig. 2), reactions were complete within the 30-s incubation period. The conversion of PGG:! to PGHz was increased about 7-fold by 500 pM phenol. Likewise, 15-HPEI reduction to PGEl was increased about 4-fold by 50 pM phenol. On the other hand, 500 pM phenol altered the distribution as well as the amount of products from 15-HPETE metabolism. Although the utilization of 15-HPETE increased as expected (from 60 to lOO%), the amount of polar product also decreased from 15 to 5%. Hence, I5-HETE was also generated at the expense of polar material, apparently A.  Table I is a series of materials whose influences on these enzymatic reactions have been established in a fashion similar to phenol. The effect of each co-substrate was concentration-dependent.
However, the concentration differences listed in the table preclude quantitative comparison, a comparison which would not contribute significantly to the present discussion anyway. The percentage changes listed in the last three columns were based on extent of reaction and were calculated relative to controls without any additive. Under these conditions, none of the hydroperoxides were susceptible to nonenzymatic reduction. Those compounds listed in the top half of the table, phenol, aminopyrine, diethyldithiocarbamate, promethazine, sulindac sulfide, lipoic acid, methional, and tryptophan, stimulated the reduction of each hydroperoxide. On the other hand, those in the bottom half of Table I, anisole, salicylic acid, reduced glutathione, methionine, sulindac, and indomethacin had either no significant effect or were mild inhibitors.
Hence, the action of a given compound was the same for each hydroperoxide.
The response to stimulators was also examined under identical conditions using 25 ( The kinetics of the microsome-catalyzed reactions of PGG2, 15-HPE,, and 15-HPETE were studied under the conditions described in Fig. 2, where the percentage of substrate remaining is plotted against reaction time. In order to simplify the interpretations, the PGG2 and 15-HPETE reactions were followed in the presence of 500 pM phenol. Under these conditions, PGHZ was the sole product of PGGa reduction and only 15-HETE resulted from 15-HPETE.
The reaction of 15-HPE, exclusively to PGE, was performed without phenol.
As depicted in Fig. 2, addition of PGGa to the buffer con- The kinetic data with 15-HPE, and 15-HPETE were qualitatively identical with the PGG2 case. About 40% of the 15-HPE, was converted to PGE, in about 15 s. No significant reaction occurred in the ensuing 165 s, even following a second addition of 15-HPEI at Point A (open triangle). In contrast, a second protein addition at this point reduced another 30% of the hydroperoxide.
15-HPETE metabolism also proceeded rapidly for the fist 15 s, giving about 25% conversion to 15-HETE. Although a period of inactivity, likewise, succeeded this initial reaction and a second 15-HPETE addition at Point A had no effect (open square), the second protein addition elicited about 20% metabolism.
The rapid secondary reaction was followed by the plateau of inactivity.
Enzyme-dependent cessation of reaction was due neither to product inhibition since addition of fresh enzyme evoked further conversion nor to substrate depletion since the reaction ceased in the presence of excess substrate. The reactions with PGGz were shorter in duration to avoid nonenzymatic degradation which occurred due to its innate instability in " All values normalized to 100 for 100 PM arachidonic acid without phenol.
This difficulty did not arise within 3 min for 15HPE, or 15-HPETE.

Formation
of Radicals during Hydroperoxide Reduction- The enzymatic generation of an EPR signal was monitored following the addition of several hydroperoxides or their reduction products to 300 ~1 of microsomal suspension. Table II lists these substrates along with their concentrations.
As described previously (6), the enzymatic metabolism of arachidonic acid (actually the PGG2 formed during its oxygenation) produced an indomethacin-sensitive EPR absorption. This signal, whose peak-to-peak intensity has been arbitrarily designated 100, was depressed 93% by conducting the reaction in the presence of 500 pM phenol. In contrast, the first nonhydroperoxide in the prostaglandin-forming cascade, PGH2, gave virtually no signal even at 500 pM. Likewise, metabolism of 15-HPE, and 15-HPETE produced signals of 63 and 84, respectively, whereas their reduction products PGEl and 15-HETE gave none even at 1000 pM. Phenol also depressed the signals from these hydroperoxides greater than 90%. The primary product of platelet lipoxygenase (2), 12-HPETE, generated a signal 94% as intense as that for arachidonic acid; 95% depression occurred with 500 PM phenol. Hydrogen peroxide, the only non-lipid hydroperoxide studied, also formed a phenol-sensitive signal about 60% as intense as arachidonic acid. On the other hand, ascaridole, another endoperoxide like PGH2, gave no EPR signal. In each instance, the signals had a single peak, a line width around 25 G and no discernible hyperfine structure. Each value is the average of at least three trials. Control incubations with either enzyme alone, substrate alone, or denatured enzyme (70°C for 10 min) and substrate showed no signals.
In addition to phenol, the other co-substrates listed in Table  I were tested as quenchers of the EPR signal. All those materials which stimulated hydroperoxide reduction, significantly quenched the EPR absorption with each hydroperoxide. In contrast, those materials which did not stimulate did not depress the signals.

Peroxidase-catalyzed
Oxidation of Phenol and Sulindac Sulfide-The metabolism of phenol and of sulindac sulfide has been studied to determine whether stimulation of hydro-peroxide reduction was coincident with oxidation of co-substrate. The products of ['4C]phenol metabolism which were generated in response to incubation with microsomal protein and arachidonic acid are shown in Fig. 3A. Of the several peaks on this radiochromatogram, phenol migrated most rapidly. The material at the origin has been identified as a polymer; it would not traverse standard gas chromatography columns and could not be volatilized in the mass spectrometer chamber.
Furthermore, portions of this product adhered tightly to microsomal protein even following heat denaturation or perchloric acid treatment. Consequently, much of the radioactivity (>50%) was not extracted into ether. Neither the phenol nor the arachidonic acid reactions occurred with denatured enzyme (75°C for 15 min) and phenol metabolism did not proceed with either enzyme alone or arachidonic acid alone. In addition, preincubation of the enzyme for 2 min with 5 pM indomethacin eliminated both arachidonic acid and phenol reactions. In accord with this well documented preincubation requirement for effective inhibition (1 l), failure to preincubate allowed both reactions to proceed normally.
15-HPEI also elicited the same phenol reaction as arachidonic acid. Due to the binding and polymeric nature of the priniciple phenol product, the identities of the other peaks were not determined, and another stimulator of hydroperoxide reduction, sulindac sulfide, was selected for quantitative studies (structure above the small peak in Fig. 3B). This co-substrate was available in both tritiated and nonradioactive forms and, as indicated in Table I, was a potent stimulator of the peroxidase. The radiochromatogram of the reaction mixture from the incubation of ["Hlsulindac sulfide with 15-HPE, in the presence of microsomal protein is shown in Fig. 3B. The sulfide was converted exclusively to sulindac, its analogous sulfoxide (structure adjacent to the larger peak). This product was identified by comparison with chemically prepared sulindac (31) in several chromatography systems and by mass spectral analysis on an LKB 9000 using direct probe insert of the solid (major peaks at 356 (mass ion), 341, and 233). The same enzymatic reaction was elicited by PGG2 and 15 HPETE.
Since sulindac sulfide inhibits arachidonic acid oxygenation (IDso = 0.1 PM), it was not possible to study arachidonic acid-induced oxidation and consequent indomethacin inhibition.
The similar kinetic patterns for the enzymatic reaction of 15-HPE1 and for sulindac sulfide are shown in Fig. 4

Quantitative
Evaluation of Co-substrate Oxidation-Since single products were formed during both 15-HPE, reduction and sulindac sulfide oxidation, this simple set of reactions was ideally suited for quantitative studies of the peroxidase. Consequently, two sets of tandem incubations with 15-['*C]HPE1, ["Hlsulindac sulfide, and their nonradioactive counterparts were conducted as described above. Following the work-up, greater than 95% of the radioactive component was recovered in each case.
Shown in Table III is a quantitative comparison of 15-HPE1 reduction and sulindac sulfide oxidation as oxidant, reductant, or enzyme was varied. Each value represents the average of at least three trials. The molar ratios of sulindac sulfide oxidized/ 15-HPE1 reduced are listed in the last column. For example, a ratio of 0.5 would indicate that 1 molecule of sulindac sulfide was oxidized for every two 15-HPEI molecules reduced.
As the 15-HPE1 concentration was increased from 150 to 500 pM at 100 pM sulindac sulfide and 0.56 mg of protein, the molar ratio decreased from 0.79 to 0.60. On the other hand, as sulindac sulfide increased from 50 to 150 pM at constant 15-HPE, and protein, the ratio increased from 0.52 to 0.95. Finally, as protein level was increased from 0.20 to 1.11 mg, the ratio decreased from 1.00 to 0.69, DISCUSSION A natural and irreversible substrate-dependent deactivation process occurs with prostaglandin cyclooxygenase (endoperoxide synthetase) (18), and a mechanism has been suggested which may explain this phenomenon (6). The proposed destructive event involved indirect attack on the oxygenase by oxidants released during a subsequent peroxidase step. This communication describes the enzymology of this microsomal peroxidase activity in order to elucidate more fully the peroxidase-mediated hydroperoxide reduction reaction and the possible effects of the resulting oxidizing equivalents on the rate, amount and direction of prostaglandin biosynthesis. During prostaglandin formation, peroxidase activity is required to reduce the 15-hydroperoxy group on PGGz or 15-HPE, to the hydroxyl moiety present on subsequent prostaglandins. Due to the mandatory cofactor requirements of the purified cyclooxygenase (9, 10) and the consequent implications to both EPR experiments and quantitative studies of cosubstrate oxidation, a microsomal preparation of the enzymes