Studies on the reduction of endogenously generated prostaglandin G2 by prostaglandin H synthase.

Prostaglandin H synthase oxidizes arachidonic acid to prostaglandin G2 (PGG2) via its cyclooxygenase activity and reduces PGG2 to prostaglandin H2 by its peroxidase activity. The purpose of this study was to determine if endogenously generated PGG2 is the preferred substrate for the peroxidase compared with exogenous PGG2. Arachidonic acid and varying concentrations of exogenous PGG2 were incubated with ram seminal vesicle microsomes or purified prostaglandin H synthase in the presence of the reducing cosubstrate, aminopyrine. The formation of the aminopyrine cation free radical (AP.+) served as an index of peroxide reduction. The simultaneous addition of PGG2 with arachidonic acid did not alter cyclooxygenase activity of ram seminal vesicle microsomes or the formation of the AP.+. This suggests that the formation of AP.+, catalyzed by the peroxidase, was supported by endogenous endoperoxide formed from arachidonic acid oxidation rather than by the reduction of exogenous PGG2. In addition to the AP.+ assay, the reduction of exogenous versus endogenous PGG2 was studied by using [5,6,8,9,11,12,14,15-2H]arachidonic acid and unlabeled PGG2 as substrates, with gas chromatography-mass spectrometry techniques to measure the amount of reduction of endogenous versus exogenous PGG2. Two distinct results were observed. With ram seminal vesicle microsomes, little reduction of exogenous PGG2 was observed even under conditions in which all of the endogenous PGG2 was reduced. In contrast, studies with purified prostaglandin H synthase showed complete reduction of both exogenous and endogenous PGG2 using similar experimental conditions. Our findings indicate that PGG2 formed by the oxidation of arachidonic acid by prostaglandin H synthase in microsomal membranes is reduced preferentially by prostaglandin H synthase.

Prostaglandin H synthase oxidizes arachidonic acid to prostaglandin GZ (PGG2) via its cyclooxygenase activity and reduces PGGz to prostaglandin Hz by its peroxidase activity. The purpose of this study was to determine if endogenously generated PGGz is the preferred substrate for the peroxidase compared with exogenous PGGz. Arachidonic acid and varying concentrations of exogenous PGGz were incubated with ram seminal vesicle microsomes or purified prostaglandin H synthase in the presence of the reducing cosubstrate, aminopyrine. The formation of the aminopyrine cation free radical (AP'+) served as an index of peroxide reduction. The simultaneous addition of PGGz with arachidonic acid did not alter cyclooxygenase activity of ram seminal vesicle microsomes or the formation of the AP". This suggests that the formation of AP'+, catalyzed by the peroxidase, was supported by endogenous endoperoxide formed from arachidonic acid oxidation rather than by the reduction of exogenous PGG2. In addition to the AP" assay, the reduction of exogenous versus endogenous PGGz was studied by using [5,6,8,9,11,12,14,15-2H]arachidonic acid and unlabeled PGGz as substrates, with gas chromatography-mass spectrometry techniques to measure the amount of reduction of endogenous versus exogenous PGGz. Two distinct results were observed. With ram seminal vesicle microsomes, little reduction of exogenous PGGz was observed even under conditions in which all of the endogenous PGGz was reduced. In contrast, studies with purified prostaglandin H synthase showed complete reduction of both exogenous and endogenous PGGz using similar experimental conditions. Our findings indicate that PGG2 formed by the oxidation of arachidonic acid by prostaglandin H synthase in microsomal membranes is reduced preferentially by prostaglandin H synthase.
Prostaglandins are formed from arachidonic acid by the bifunctional enzyme prostaglandin H synthase. The initial step of this process is a peroxidation of arachidonic acid a t carbon 11 and the subsequent formation of the endoperoxide * 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 should be addressed.  (1, 2). The peroxidase activity of prostaglandin H synthase then reduces PGGz to prostaglandin Hz (PGH,) by reducing the hydroperoxide on carbon 15 to a hydroxyl group (3-6). PGHz is then converted to a variety of compounds, including prostaglandins and thromboxanes. The cyclooxygenase and peroxidase activities copurify (7) from seminal vesicle microsomes, and considerable data indicate that a structural and functional interrelationship exists between the two enzymatic activities within prostaglandin H synthase. However, the nature of this interrelationship is unclear and remains the subject of considerable speculation and controversy (for a review, see Ref. 8 Cyclooxygenase Assay-Cyclooxygenase activity was measured by monitoring oxygen incorporation into arachidonic acid at 37 "C (final volume, 1.5 ml) with a Gilson oxygraph equipped with a Clark-type oxygen electrode. Aminopyrine in ethanol or phenol in aqueous stock solution was added 2 min before the addition of arachidonic acid in ethanol. In some experiments the PGG, was added 1 min prior to the addition of arachidonic acid or added simultaneously with arachidonic acid from a previously prepared stock solution in ethanol.
Prostaglandin Peroxidase-The peroxidase activity of prostaglandin H synthase was quantitated by measurement of the oxidation of aminopyrine to the aminopyrine cation free radical, AP'+ (12,13). This radical has a maximum absorbance at 565 nm ( 6 = 2.23 mM" cm") and can be conveniently measured spectrophotometrically (13). Solubilized RSV microsomes (0.15 mg/ml) were preincubated in 100 mM phosphate buffer, pH 7.8, with varying concentrations of aminopyrine in a final volume of 1.5 ml. The reaction was started by the addition of 100 PM arachidonic acid or peroxide at various concentrations. The formation of the AP'+ was measured as function of time with a Hewlett-Packard 8450A spectrophotometer, and the peak absorbance was used to estimate the amount of radical formed, as described in detail in the table and figure legends.
Conditions were established where the peroxidase activity of prostaglandin H synthase was able either to convert all of the PGG, formed by the cyclooxygenase activity to PGH, or able to convert only about 50% of the PGG, to PGH,. To establish these conditions, RSV microsomes (0.15 mg of protein/ml of reaction mixture) were incubated in 100 mM phosphate buffer, pH 7.8, a t 37 "C with phenol or aminopyrine at varying concentrations in a final volume of 1 ml for 1 min. Reactions were initiated by the addition of 100 p~ &arachidonic acid containing about 0.1 pCi of [3H]arachidonic acid. After 90 s reactions were terminated by the addition of 4 ml of icecold ethanol, and mixtures were allowed to stand for 10 min. Under these conditions, PGG, and PGH, decompose almost exclusively to 15-hydroperoxy-PGE2 and PGE,, respectively (14). Samples were evaporated to dryness and then reconstituted in a small amount of (50/50/1) methanol/H,O/acetic acid, and products were separated by HPLC on a Beckman Altex reverse-phase column as described previously (15). Radiation was detected as samples eluted from the column by a Radiomatic Flo-One detector with Ecolume scintillation mixture. PGE, eluted a t about 22 min, and 15-hydroperoxyprostaglandin E, (15-HP-PGE,) eluted a t 27 min.
Reduction of Endogenous PGG, Formed from dB-Arachidonic Acid in Situ-&-Arachidonic acid is metabolized by prostaglandin H synthase to &-PGG, and eventually &-PGH2. The decomposition of PGH, to PGE, results in the loss of the deuterium at carbon 9 (Scheme 1). The formation of the resulting [5,6,8,11,12,14,15-2H] PGE, (d,-PGE,) from &-arachidonic acid was measured by GC-MS techniques. RSV microsomes (0.15 mg of protein/ml) or purified prostaglandin H synthase (700 units) and 1 p~ hematin were incubated as described above in the presence of either phenol or aminopyrine. Reactions were initiated by the addition of &-arachidonic acid (final concentration, 100 p~) .
When PGGZ was to be added, a stock solution containing PGG, and &-arachidonic acid was prepared in ethanol. The reaction was then initiated by the addition of 10 pl of this solution. Oxygen concentrations in reaction mixtures were measured with a Clark-type electrode, and concentrations of endogenously formed PGG, were calculated based on the incorporation of 2 mol of oxygen into 1 mol of arachidonic acid during the formation of PGG, (Scheme 1). After 90 s the reactions were stopped by the addition of 4 ml of ice-cold ethanol, and 0.05 pCi of 15-[3H]hydroxyeicosatetraenoic acid (15-HETE) was added as an internal standard to determine efficiency of extraction from samples before chromatography. Samples were evaporated to dryness by rotary evaporation and chromatographed as described above. Fractions corresponding to the PGE, elution time were collected, and 15-['H]HETE was quantitated by scintillation counting to determine recoveries. In some experiments, radiolabeled arachidonic acid was added to the &-arachidonic acid stock solution to monitor endogenous PGG, reduction. Five ng of dd-PGF,,. was added to the sample prior to HPLC purifications as a n internal standard for derivatization and GC-MS procedures, and samples were evaporated to dryness and reconstituted in a small volume of absolute methanol.
GC-MS Analysis of PGEa-Samples were derivatized sequentially by synthesis of the methyloxime-pentafluorobenzyl ester-trimethylsilyl ether derivatives as described (16-18). Combined capillary gas chromatography-negative ion mass spectrometric analysis of the electron-capture derivatives was performed via multiple ion detection of characteristic fragment ions for &-PGF2, (m/z 573), Q-PGE, (m/z 524), and d7-PGE, (m/r 531). Ratios of peak areas a t m/z 531 and 524 were measured for determinations of substrate transformations catalyzed. &-PGF,, was employed as an internal standard for identification of the 4and d7-species of PGE, and for quantitation of peak areas. Experimental Design-To determine if the PGG, formed from arachidonic acid was the preferred substrate for the peroxidase compared with exogeneous PGG, it was necessary to estimate the amount of PGGa reduced to PGH, in an incubation. We chose to estimate these unstable intermediates by measuring their stable nonenzymatic decomposition products, 15-HP-PGEz and PGE,, respectively, as shown in Scheme 1. These compounds were the major decomposition products, but 12-hydroxyheptadecatrienoic acid and PGD, were also formed in minor amounts at neutral pH (14,19). To distinguish between the reduction of endogenous from exogenous PGG, we used &-arachidonic acid to form endogenous &-PGG* and &-PGG, as the source of exogenous peroxide. The isotopic composition of PGE,, as measured by GC-MS techniques, was used to estimate the competition between endogenous and exogenous PGG,. Furthermore, the conditions of the assay were established such that not all the PGG, was reduced to PGH, so that competition between the PGG, formed from arachidonic acid and added PGG, could occur. Under conditions in which the peroxidase reduces all the PGG, to PGH,, simple dilution of the d,-PGE2 would occur. The peroxidase activity of prostaglandin H synthase can he regulated by the presence of varying concentrations of a reducing cosubstrate such as aminopyrine or phenol. Aminopyrine is an excellent choice since it also provides a direct estimate of the total peroxides reduced by the peroxidase as described under "Experimental Procedures."

RESULTS
Metabolism of Arachidonic Acid by Prostaglandin H Synthase-Under the conditions of our incubations, 15-HP-PGE, and PGE, were the major decomposition products of PGG, and PGH,, respectively (data not shown). The presence of a reducing cosubstrate such as phenol or aminopyrine enhances the peroxidase activity of prostaglandin H synthase and the conversion of arachidonic acid to PGG,; hence, more PGG, is reduced to PGH, and ultimately converted to PGE,. When [3H]arachidonic acid was incubated with ram seminal vesicle microsomes in the presence of 500 pM aminopyrine or phenol and reaction products were analyzed via HPLC, one major peak was detected which coeluted with an authentic PGE, standard ( Fig. 1, top). PGD,, another possible decomposition product of PGH,, coelutes with PGE, on the chromatography system utilized (15); however, analyses of products by GC demonstrated that about 90% of the PGH, was converted to PGE, and about 10% to PGDz (data not shown). In the presence of 50 p~ aminopyrine the decomposition products were distributed about equally between PGE, and 15-HP-PGE, (Fig. 1, middle). Similar results were also obtained with  30 FM phenol rather than aminopyrine (data not shown) although the total oxidation of arachidonic acid to PGG2 as measured by oxygen incorporation was less than with the microsomal preparation used for the aminopyrine experiments (Table I). Thus, in the presence of low concentrations of a reducing cosubstrate a condition existed in which half of the PGG, formed from the oxidation of arachidonic acid by the cyclooxygenase was reduced by the peroxidase to PGH,.
These results indicate that the peroxidase activity was finite and was therefore limited with respect to PGG,. When microsomes were incubated with arachidonic acid in the absence of a reducing cosubstrate the predominant product formed was 15-HP-PGE, (Fig. 1, bottom). This is in agreement with our previous study on arachidonic acid oxidation to PGG, and PGH2 by ram seminal vesicle microsomes (14), which dem-   Table I, the addition of the reducing cosubstrate aminopyrine increased the oxidation of arachidonic acid by prostaglandin H synthase. For example, approximately 44 pM PGGz was formed in the presence of 50 pM aminopyrine compared with 24 p~ formed in the absence of a reducing cosubstrate. The formation of the AP" appeared to be in agreement with the ratio of 15-HP-PGE2 to PGE, as determined by HPLC analysis. However, at 500 p~ aminopyrine the formation of the AP" underestimated the reduction of PGG,. At this aminopyrine concentration, higher concentrations of the AP'+ are achieved, which decays at a faster rate than observed at lower concentrations (13). The rapid speed of the decay results in an underestimation of the AP ' + concentration. Additional experiments were conducted with purified prostaglandin H synthase using varying concentrations of reducing cosubstrate. These results differed significantly from those obtained with microsomal prostaglandin H synthase. Regardless of the amount of phenol (0-500 FM) only PGE, was observed in these incubations as measured by HPLC analysis (data not shown). Maximum enzymatic activity was observed at 500 pM phenol, with approximately 10 pM PGG? being formed at this cosubstrate concentration (Table I). Cyclooxygenase activity was reduced significantly at lower phenol concentrations (data not shown). This is in agreement with previously published reports demonstrating the high sensitivity of purified prostaglandin H synthase to peroxide initiated inactivation in the absence of a reducing cosubstrate (20).

Effect of Exogenous PGG2 on Prostaglandin H Synthme-
The effect of exogenous PGG, on the cyclooxygenase and peroxidase activities of prostaglandin H synthase was examined using RSV microsomes as the source of the enzyme.
Previous data (20) indicated that exogenous PGG, and other peroxides could inactivate prostaglandin H synthase. This inhibition could interfere in competition experiments between endogenous and exogenous PGG,. The addition of PGGz 60 s before the addition of arachidonic acid caused a significant inhibition of prostaglandin H synthase activity as measured by oxygen incorporation with RSV microsomes as seen in Fig.  2. A 5 PM concentration of PGG, produced a 50% inhibition, which is in agreement with published data (20). However, the simultaneous addition of PGG, and arachidonic acid from a single solution did not inhibit the cyclooxygenase activity even at PGG, concentrations as high as 100 p~ (Fig. 2). Thus, under these conditions the simultaneous addition of exogenous PGG, does not contribute to the inhibition of cyclooxy- genase which occurs during arachidonic acid oxidation by prostaglandin H synthase.
To investigate the effect of exogenous PGG, on the peroxidase we measured the formation of the AP'+ by RSV microsomes. Reactions were initiated by the addition of either arachidonic acid, PGG, or a mixture of arachidonic acid and PGG, under conditions described above except that 500 p~ aminopyrine was included. As shown in Fig. 3, the addition of PGG, to solubilized RSV microsomes produced a very rapid appearance of the AP" which then decayed as described previously (13). The peak absorbance value in the curve was used to estimate the total formation of AP" and thus the reduction of PGG, to PGH2 based on a stoichiometry of 1 mol of peroxide12 mol of AP ' + (13). At a PGG, concentration of 50 pM approximately 65 p~ AP" was formed whereas at 100 p~ arachidonic acid approximately 96 p~ AP" was observed. The data presented demonstrated that the simultaneous addition of arachidonic acid and PGG, did not alter the AP" compared with arachidonic acid alone, consistent with the data in Fig. 2. Table I1 shows the quantitative comparison between the formation of the AP" at the various incubation conditions. The simultaneous addition of 100 p~ arachidonic acid and PGG, at concentrations as high as 60 p~ PGG, did not alter the level of AP" formation. This result may be explained by the fact that the enzyme appeared to be operating at near capacity for reduction of PGG, as observed in additional experiments (data not shown). However, the addition of PGG, did not alter the reduction of the PGG, derived from labeled arachidonic acid. As seen in Fig. 3 insets, HPLC analysis indicated essentiallv the same auantitv of PGH,. These results suggest that the AP" formation catalyzed by the peroxidase was supported by PGG, derived from arachidonic acid rather than exogenous PGG,. Analysis of Reduction Products of do-and d7-PGGz-After isolation by HPLC, d7-and Q-PGE2 formed from the decomposition of &and Q-PGH, were analyzed via negative chemical ionization GC-MS as described above and used as an index for estimating PGG, reduction to PGH,. Two peaks at mlz 524 corresponding to the methoxime derivative stereoisomers of Q-PGE, were detected at about 19 min 4 s and 19 min 54 s (Fig. 4, top). Methoxime isomers of d7-PGE2 were detected at mlz 531 as two peaks eluting slightly earlier than those detected at mlz 524 (Fig. 4, middle). The internal standard, Q-PGF,,, was detected as a single peak eluting a t 18 min 40 s at mlz 573 (Fig. 4, bottom).
Reduction of Endogenous and Exogenous PGG2"The isotopic comDosition of PGE, was used to estimate the reduction exogenous and endogenous PGG,. We initially used purified H synthase The formation of the aminopyrine cation free radical (AP.+) was measured spectrophotometrically at 565 nm as described previously (12,13). The quantity of AP'+ formed was calculated from the extinction coefficient and the peak in the absorbance curve as seen in Fig. 3 prostaglandin H synthase to determine if endogenous and exogenous PGG, competed for reduction by the peroxidase. A cosubstrate such as phenol was required for the purified enzyme to oxidize arachidonic acid to the endoperoxides (20). Under these conditions, the percentage of PGEz as d7-PGEz decreased proportionately to the concentration of Q-PGG, added as shown in Fig. 5,panel A . At a Q-PGG, concentration of 10 p~ the percentage was decreased by approximately 50%. Under these conditions approximately 10 p~ &-PGG, was formed from the &-arachidonic acid as estimated from the oxygen incorporation (Table I). Further inspection of the data indicated that the experimental curve paralleled the theoretical dilution curve calculated from the amount of PGGz formed from arachidonic acid and the amount of exogenous PGG, added to the incubation. This theoretical dilution curve would be expected if both the exogenous and endogenous PGGp have equal access for the peroxidase for reduction, a result that was obtained in the experiment shown in panel A . With purified prostaglandin H synthase and 500 p~ phenol, all the endogenous and exogenous PGG, was reduced to PGH,. Hence, simple dilution of the isotope occurred in this experi-ment. Since all the peroxide was reduced, these experimental conditions do not permit us to address accurately whether or not endogenous PGG, is the preferred substrate for the peroxidase, but they do provide evidence for the validity of the techniques and experimental approach.
With the purified prostaglandin H synthase we attempted to establish conditions in which competition between the endogenously generated PGG, and exogenous PGG, could be measured. These attempts were unsuccessful with the purified enzyme but were achieved with ram seminal vesicle microsomes that were washed to reduce any contaminating endogenous cosubstrate for the peroxidase, as described above. Fig.  5 ,  incubations, was not altered by the simultaneous addition of PGG,. Furthermore, the formation of AP", which serves as a measure of the peroxidase activity of prostaglandin H synthase, was not altered by the addition of exogenous PGG2 as compared with incubations with arachidonic acid alone (Table 11). For both 50 and 500 p~ aminopyrine the peroxidase was active. At 50 p~ aminopyrine, approximately 44 p~ &-PGGz was generated in the incubation of 100 p~ &-arachidonic acid with RSV microsomes. With no exogenous Q-PGG, added, all of the PGE, formed was heptadeuterated as expected. As the concentrations of exogenous &-PGGz were increased to 60 p~, a slight dilution of the d7-PGE, (to about 70%) was observed (Fig. 5, panel C). Also shown is the theoretical curve predicted for equal competition between the endogenous and exogenous PGG,. These data demonstrate clearly that less of the exogenous PGG, was reduced than expected under competitive conditions. At 500 p~ aminopyrine, a concentration that supported the formation of 80 pM &-PGG, from 100 pM &arachidonic acid, the percentage of d7-PGEz decreased to only 85% as the concentration of exogenously added Q-PGG, increased to 60 p~ (Fig. 50). A dilution of about 52% was expected for 80 p~ endogenous and 60 p~ exogenous PGGz (Fig. 50, open curves), again indicating less reduction of exogenous PGG, than would be expected if simple dilution occurred. It must be noted that as the data are presented the magnitude of the difference between the experimental and theoretical values was reduced as a larger quantity of endogenous PGG, was formed. Thus, when larger amounts of endogenous PGG, are formed the slope of theoretical dilution curves decreases as can be seen on comparing the panel A with D in Fig. 5 .

DISCUSSION
Prostaglandin H synthase has two enzymatic activities present within a single homodimer: the cyclooxygenase and a peroxidase. The cyclooxygenase catalyzes the oxidation of arachidonic acid to a peroxide, PGG2, and the peroxidase catalyzes the two-electron reduction of the peroxide PGGz to the corresponding alcohol PGHz (Scheme 1). The peroxidase activity of prostaglandin H synthase has the ability to reduce a wide variety of exogenous peroxides at the expense of reducing cosubstrates (1)(2)(3)(4)(5)(6)(7)(8). In this paper we attempted to answer the following question: Is the product of the first enzymatic activity the preferred substrate for the second enzymatic activity?
The data presented here indicate for the microsomal enzyme that endogenously formed PGG, is preferentially reduced by the peroxidase compared with exogenously added PGG,, but some competition occurred between the endogenous and exogenous PGG,. This conclusion is supported by several observations. First, incubations of prostaglandin H synthase with low micromolar concentrations of peroxides such as PGG, inactivated the enzyme (20). However, the simultaneous addition of PGG, at concentrations up to 100 PM with arachidonic acid did not enhance the inhibition of the cyclooxygenase activity which occurs during arachidonic acid oxidation. Since the peroxidase turnover has been linked to inactivation of the cyclooxygenase, this finding indicates that the exogenous PGG, was not accessible to or reduced by the peroxidase in the presence of arachidonic acid.
Second, the addition of PGG, did not alter the formation or reduction of endogenous PGG, as measured by 0, consumption (Fig. 2) or by HPLC analysis of "H-metabolites formed from labeled arachidonic acid. Moreover, measurement of the amount of peroxide reduced by the peroxidase, as estimated by the quantitative formation of the AP'+, was similar for reactions initiated by arachidonic acid alone or by arachidonic acid and PGG, simultaneously (Table 11). These data indicate that the endogenous PGG, produced by the cyclooxygenase from arachidonic acid was supporting the oxidation of the aminopyrine by the peroxidase.
Finally, direct measurement of the reduction of PGG, was accessed by measuring the isotopic composition of the PGH, decomposition product, PGE,, by gas chromatography-mass spectrometry. These results indicate that endogenous PGG, was the preferred substrate for the peroxidase, but some competition between the endogenous and exogenous PGG, occurred. This conclusion was based on the results of two experiments with different preparations of ram seminal vesicle microsomes and using either phenol or aminopyrine as the cosubstrate for the peroxidase. Some variability in the extent to which the exogenous PGG, was able to compete with endogenous PGGZ was observed in these experiments. In one experiment (Fig. 5 , panel B ) the exogenous PGG, was essentially not reduced by the peroxidase in the presence of endogenous PGG,. In the second experiment (Fig. 5 , panels C and D ) with different amounts of the peroxidase cosubstrate present in the incubation some reduction of exogenous PGG, occurred. Clearly in both experiments the experimental results were significantly different from the calculated theoretical values based on the assumption that if the endogenous PGG, dissociated from the enzyme it would freely mix with the exogenous PGG, and then be reduced by the peroxidase to PGH,. The preference of the peroxidase for endogenous