Spectral Properties of the Higher Oxidation States of Prostaglandin H Synthase*

Prostaglandin H (PGH) synthase reacts with organic hydroperoxides and fatty acid hydroperoxides on a millisecond time scale to generate an intermediate that is spectrally similar to compound I of horseradish per- oxidase. Compound I of PGH synthase is converted to compound I1 within 170 ms. Compound I1 decays to resting enzyme in a few seconds. Thus, the peroxidase reaction of PGH synthase appears to involve a cycle of native enzyme, compound I, and compound 11, typical of heme-containing peroxidases. The Soret absorption maximum of compound I appears to occur at 412 nm but a small amount of compound I1 may be present. Soret maxima occur at 420, 433, and 419 for com- pound 11, the ferrous enzyme, and the oxyferrous enzyme (compound 111), respectively. Rapid scan analysis of the reaction of PGH synthase with arachidonic acid reveals the absorbance of compound 11 but no evidence for ferrous or oxyferrous enzyme. the first two reactions in the biosynthesis of prostaglandins. The cyclooxygenase component oxygenates arachidonic acid to PGG2, a hydroperoxy endoperoxide, and the peroxidase component reduces the hydroperoxy group of PGGz to an alcohol, PGHz (Equation (1). PGH synthase is a membrane-bound protein and been purified to apparent electrophoretic

Prostaglandin H (PGH) synthase reacts with organic hydroperoxides and fatty acid hydroperoxides on a millisecond time scale to generate an intermediate that is spectrally similar to compound I of horseradish peroxidase. Compound I of PGH synthase is converted to compound I1 within 170 ms. Compound I1 decays to resting enzyme in a few seconds. Thus, the peroxidase reaction of PGH synthase appears to involve a cycle of native enzyme, compound I, and compound 11, typical of heme-containing peroxidases. The Soret absorption maximum of compound I appears to occur at 412 nm but a small amount of compound I1 may be present. Soret maxima occur at 420, 433, and 419 for compound 11, the ferrous enzyme, and the oxyferrous enzyme (compound 111), respectively. Rapid scan analysis of the reaction of PGH synthase with arachidonic acid reveals the absorbance of compound 11 but no evidence for ferrous or oxyferrous enzyme. I PGH' synthase catalyzes the first two reactions in the biosynthesis of prostaglandins. The cyclooxygenase component oxygenates arachidonic acid to PGG2, a hydroperoxy endoperoxide, and the peroxidase component reduces the hydroperoxy group of PGGz to an alcohol, PGHz (Equation 1) (1). PGH synthase is a membrane-bound protein and has been purified to apparent electrophoretic homogeneity from bovine and ovine seminal vesicle microsomal preparations (2-4). The purified enzyme, a homodimer of 70,000-dalton subunits, contains both cyclooxygenase and peroxidase activities (2, 4). It is a hemeprotein and the heme prosthetic group is required for both activities (2, 5).
The peroxidase component of PGH synthase reduces several different hydroperoxides while oxidizing a variety of reducing substrates (1,6 We have utilized rapid-scan spectrophotometry to record the absolute spectra of peroxidase intermediates generated in the millisecond time range. Species are detectable that are very similar to compounds I and I1 of horseradish peroxidase. Compound I1 is responsible for difference spectra previously attributed to compound I. Experiments with arachidonic acid and PGH synthase indicate that these intermediates are formed as a result of the cyclooxygenase reaction and that no other spectrally detectable species are generated.

MATERIALS AND METHODS
PGH synthase was purified from ram seminal vesicle microsomes as previously described (9). The enzyme preparation was greater than 95% pure as judged by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The purified enzyme was 44-50% holoenzyme based on cyclooxygenase activity in the presence and absence of exogenous hematin and on absorbance at 410 nm using a molar absorptivity of 123 cm" (10). The enzyme was used without reconstitution with hematin. It exhibited cyclooxygenase activity of 22 pmol of arachidonic acid oxidized per mg of protein/min in the absence of exogenous hematin and 49 pmol of arachidonic acid oxidized per mg of protein/min in the presence of 1 p~ hematin. Enzyme concentrations were estimated by absorbance at 410 nm. Protein was estimated using the Bio-Rad assay with bovine serum albumin as the standard (Bio-Rad). Unless otherwise indicated, aliquots of the purified enzyme were desalted on Sephadex G-25 equilibrated in 100 mM phosphate (pH 7.7), 30% glycerol to remove diethyldithiocarbamate present for stabilization during the purification and storage of the enzyme.
15-Hydroperoxyeicosatetraenoic acid was synthesized from arachidonic acid using soybean lipoxygenase by the method of Funk et ai. (11). PPHP was a generous gift from Paul Weller (Department of Chemistry, Wayne State University). Arachidonic acid was purchased from NuChek Prep. Substrate samples were dissolved in 100 mM phosphate (pH 7.7), 30% glycerol. Hydrogen peroxide (30%) was purchased from Fisher, rn-chloroperoxybenzoic acid from Aldrich, and indomethacin from Sigma. Stock solutions of PPHP and indoacid in acetone; both were diluted (at least 100-fold) in glycerol/ methacin were prepared in methanol, and rn-chloroperoxybenzoic phosphate buffer.
Absorption measurements were performed with a Cary 219 spectrophotometer. Rapid scan spectra were recorded with a Union Giken RAG01 Rapid Reaction Analyzer equipped with a 1-cm cell thermostated at 5 "C -C 0.3 "C. The absorption spectra were measured by means of a multichannel photodiode array and memorized in a digital computer system (Sord M200 Mark 111). The analogue replica was plotted on an X-Y recorder. The dead time of the flow apparatus was 4 ms. Regions (96 nm) were scanned from 360 to 680 nm with 10 nm overlaps in all regions.
In experiments with reduced PGH synthase, buffer and enzyme solutions were made anaerobic by flushing with argon (Zero Gas, Canadian Liquid Air) purified by passage through an oxygen trap (Alltech Associates). Enzyme solutions were reduced by adding microliter amounts of a concentrated sodium dithionite solution (0.1 M in 1 mM NaOH) to a final concentration of 100 or 200 p M (12). The reduction was performed immediately before transferring the enzyme to the rapid scan apparatus.

RESULTS AND DISCUSSION
When 11 eq of PPHP are added to native PGH synthase, a good yield of compound I is observed (Fig. 1) M-' s-')' as judged by the comparable behavior of horseradish peroxidase with several reactive peroxides. Compound I of PGH synthase then decays spontaneously to a species with a compound 11-like spectrum (Fig. 2), again in accord with the behavior of many other peroxidases ( k = 1.2 X lo6 "' s-').
The visible region spectrum is particularly relevant, where the O( and ( 3 bands are located close to those found for compound TI of horseradish peroxidase and distinct from those of horseradish peroxidase compound I11 (oxyferrous peroxidase). Spectral data for PGH synthase with comparisons to horseradish peroxidase are summarized in Table I. Spontaneous decay of compound I1 of PGH synthase back to the native enzyme is then observed with a small loss in absorbance indicating some heme destruction (data not shown) ( k < 0.1 s-'). The rate of disappearance of the compound I1 spectrum increases with increasing [PPHP], which may indicate that PPHP can also act as a reducing substrate.
The enzyme can be recycled many times by subsequent addition of equimolar amounts of PPHP. A few per cent destruction of heme is observed upon completion of each cycle.
The addition of a 10-fold excess of 15-hydroperoxyeicosatetraenoic acid leads to a smaller yield of compound I of PGH synthase and on a slower time scale than the reaction of an equimolar amount of PPHP (data not shown) ( k = 1.5 X lo6 poorer oxidizing substrate than PPHP. When either hydrogen peroxide or m-chloroperoxybenzoic acid is added to native PGH synthase, very little change is recorded on the millisecond or longer time scale in the Soret spectral region. Thus, two excellent oxidizing substrates for horseradish peroxidase do not produce spectral intermediates with PGH synthase, indicating profound differences in the distal active site regions. Upon addition of a 10-fold molar excess of arachidonic acid to an aerated solution of PGH synthase, no compound I formation is detected. Rather over the time interval of 2 to 26 s after mixing, the conversion of native enzyme to compound I1 is observed (Fig. 3)   zyme (14). In the visible region we observe a single peak at 559 nm with a shoulder at 590 nm. Upon mixing the ferrous enzyme with aerated buffer, a compound I11 (oxyferrous PGH synthase) spectrum is obtained (Fig. 4 and Table I). The lack of an isosbestic point and subsequent loss of 16% absorbance indicates that heme destruction is occurring in this lime interval. Independent measurements3 indicate that excess dithionite is being consumed in a reaction with molecular oxygen during this period. Heme destruction with excess dithionite and molecular oxygen has been utilized through several cycles to prepare the apoenzyme (15). After 800 ms, heme destruction ends and an isosbestic point is observed between the oxyferrous PGH synthase (tl,z = 4.5 s) and its decay product, the native enzyme (Fig. 5). When 1.8 PM ferrous enzyme is mixed with air-saturated buffer containing 17 ~L M arachidonic acid, no change in the rate of compound I11 formation or decay is observed.
The studies summarized in Figs. 1 and 2 represent the first complete characterization of the peroxidase reaction of PGH synthase. Intermediates are detected that are spectrally analogous to those of other heme peroxidases, which indicates that, in terms of its reaction cycle, PGH synthase is a typical heme peroxidase. The similarities to horseradish peroxidase are striking (Table I)   Experimental conditions as in Fig. 4. The spectra shown are 1.5 and 6 s after mixing and the final (ferric) spectrum upon completion of the decay process. The arrows indicate the direction of the absorbance change with increasing time.
synthase has been determined to be iron protoporphyrin IX by the pyridine hemochromogen method* (16). The position of the visible absorbance maxima of compound I1 then may indicate that imidazole is the fifth ligand to the heme iron; however, differences in native enzyme and compound I spectra remain to be explained.
The slower rate of reaction of the enzyme with arachidonic acid relative to PPHP is expected because of the reduced rate of hydroperoxide formation at 5 "C. Under these conditions, the inability to detect compound I arises because the rate of its formation from arachidonic acid (reflecting the slow rate of PGG, formation) is slower than the rate of its consumption. An intriguing possibility is that arachidonate reacts with compound I, generating compound I1 and an arachidonate radical which reacts with O2 to form PGG2. What is very clear from these experiments is that ferrous and oxyferrous forms of the enzyme are not detectable during either the peroxidase or cyclooxygenase catalytic cycles. Thus, experimental evidence does not support the hypothesis that the enzyme cycles between Fe3+ and Fez+ during cyclooxygenase catalysis (17).