1 PROSTAGLANDIN H SYNTHASE : EFFECTS OF PEROXIDASE COSUBSTRATES ON CYCLOOXYGENASE VELOCITY *

Many cosubstrates for the peroxidase activity of prostaglandin H synthase-1 (PGHS-1) have been reported to produce a large (2-7-fold) increase in the cyclooxygenase velocity in addition to a substantial increase in the number of cyclooxygenase catalytic turnovers. The large stimulation of cyclooxygenase velocity has become an important criterion for evaluation of putative PGHS reaction mechanisms. This criterion has been a major weakness of branched-chain tyrosyl radical mechanisms, which correctly predict many other cyclooxygenase characteristics. Our computer simulations based on a branched-chain mechanism indicated that the uncorrected oxygen electrode signals commonly used to monitor activity can seriously overestimate the effects of cosubstrate on cyclooxygenase velocity. The simulation results prompted re-examination of the effect of several cosubstrates (phenol, acetaminophen, N,N,N',N'-tetramethylphenylenediamine, and Trolox) on PGHS-1 cyclooxygenase velocity. Cyclooxygenase kinetics were examined at reduced temperature or elevated pH, where the oxygen electrode signal can be corrected to provide reliable oxygen consumption trajectories. The cosubstrates produced only a slight (10-60%) stimulation of the cyclooxygenase velocity. Peroxidase cosubstrates thus have a much smaller stimulatory effect on cyclooxygenase velocity than previously reported. This corrects a longstanding misperception of cosubstrate effects, provides more realistic kinetic constraints on PGHS mechanisms, and removes what was a major deficiency of branched-chain tyrosyl radical mechanisms.

The cyclooxygenase activity of prostaglandin H synthase isoforms 1 and 2 (PGHS-1 and -2) 1 is a key control point in the biosynthesis of all prostanoid lipid mediators (1,2).Besides the cyclooxygenase activity, both PGHS-1 and -2 have a heme-dependent peroxidase activity (1).The PGHS peroxidase catalytic cycle resembles that of other heme-dependent peroxidases: the ferric heme of the resting enzyme is oxidized to Intermediate I (Compound I) by reaction with peroxide, and electron-donating cosubstrates complete the peroxidase catalytic cycle by reducing Intermediate I to Compound II and then to resting enzyme (3).Endogenous peroxidase cosubstrates, such as uric acid, are present in cell cytosol (4).
Cosubstrates for PGHS peroxidase also have been reported to affect PGHS cyclooxygenase activity, attenuating self-inactivation and increasing the catalytic velocity (5)(6)(7)(8)(9)(10)(11)(12)(13).The protective action of cellular reductants greatly increases the number of catalytic turnovers before cyclooxygenase selfinactivation, thereby increasing the capacity for synthesis of potent lipid mediators (14).For its part, the large stimulation of cyclooxygenase velocity by cosubstrate has become a defining characteristic of catalytic behavior and an important criterion for evaluation of potential PGHS reaction mechanisms (11,13,15).Many mechanistic schemes have been devised to account for stimulation of cyclooxygenase velocity.One strategy has been to suggest that the reductive action of cosubstrates retrieves enzyme from catalytically unproductive oxidized intermediates (11,13,16), but plausible, mechanisms adopting this approach have generally not been tested for their quantitative kinetic predictions.Another approach proposed a catalytic role for cosubstrate in the cyclooxygenase active site (17), but simultaneous binding of cosubstrate and arachidonate seems at odds with spatial constraints at the top of the cyclooxygenase channel in crystallographic models (18)(19)(20).
A tyrosyl radical at residue 385 (ovine PGHS-1 numbering) has emerged as a key cyclooxygenase catalytic intermediate in both isoforms (21)(22)(23)(24)(25)(26), and branched-chain reaction mechanisms based on the tyrosyl radical successfully account for many aspects of peroxidase and cyclooxygenase kinetics (15,27,28).However, one perceived deficiency of branched-chain mechanisms has been an inability to predict significant stimulation of cyclooxygenase velocity by cosubstrates without ad hoc assumptions (15,17).Recent studies have extended the basic branched-chain mechanism for PGHS to include steps for peroxidase self-inactivation (29).In the present work, we used kinetic simulations based on an extended branched-chain mechanism to evaluate the possibility that cosubstrates stimulate cyclooxygenase velocity via rescue of intermediates in the peroxide-induced self-inactivation pathway.
The kinetic simulations predicted that peroxidase cosubstrates greatly increase the number of cyclooxygenase catalytic turnovers before self-inactivation, but with only a small increase in the cyclooxygenase rate.The simulations also predicted that uncorrected oxygen electrode measurements can greatly exaggerate the stimulatory effects of cosubstrate on cyclooxygenase velocity.
Oxygen electrode measurements have been the mainstay of previous detailed studies of cosubstrate effects on cyclooxygenase velocity; because the potential problems were not suspected, these studies did not correct the electrode data for dampening effects (11)(12)(13).This situation prompted a detailed re-examination of the actions of several cosubstrates on PGHS-1 cyclooxygenase kinetics under conditions permitting reliable correction of oxygen electrode signals.The results indicate that cosubstrates actually increase the peak cyclooxygenase velocity much less than previously reported.This surprising observation greatly simplifies the kinetic behavior that putative reaction mechanisms must account for and removes what had been a major weakness of branched-chain mechanisms.

MATERIALS AND METHODS
Arachidonic acid was purchased from NuChek Preps, Inc (Elysian, MN).Heme, phenol, TMPD, and acetaminophen were from Sigma (St. Louis, MO).Trolox was from Aldrich Chemical Co.
(Milwaukee, WI).PGHS-1 was purified to homogeneity from sheep seminal vesicles and reconstituted with an equimolar amount of heme before use (30).
Oxygen consumption during the cyclooxygenase reaction was monitored with a Model 53 oxygen electrode system (YSI, Inc., Yellow Springs, OH) connected to a Model ADC-1 A/D converter (Remote Measurement Systems, Seattle, WA) and an Apple Macintosh SE computer (31).A "High Sensitivity" membrane was used to cover the electrode.The reaction buffer (3 ml of either 0.1 M potassium phosphate, pH 7.2, or 0.1 M borate, pH 9.8) contained 100 µM arachidonate and 1 µM heme.Reactions were initiated by injection of enzyme.When desired, digitized oxygen concentration values from the electrode were corrected for the dampening effects of the membrane using the electrode response kinetics determined from injection of buffer saturated with nitrogen gas (15,32).
For measurement of oxygen electrode cuvette mixing kinetics, the apparatus was set up just as for cyclooxygenase reaction measurements, except that the water circulator used to regulate temperature was turned off and the water bath surrounding the cuvette was drained.The cuvette was filled with 3 ml of 0.1 M potassium phosphate, pH 7.2, the electrode/probe assembly inserted, and stirring established at 900 rpm.The beam from a laser pointer (Radio Shack LX 3000) was directed horizontally through the center of the liquid column in the cuvette, striking a photovoltaic cell (Radio Shack 0.75 x 1.5 in) on the other side.The photovoltaic cell output voltage was fed to a chart recorder (Kipp & Zonen Model BD111) set at 0.5 V / full scale and 2.0 cm / s.A purple dye solution was made by oxidizing TMPD with potassium ferricyanide.When injected into the cuvette, the purple chromophore strongly absorbed the red laser light, decreasing the voltage produced by the photocell.Recorder traces were obtained for several injections of oxidized TMPD, the traces digitized (Data Thief, by Kees Huyser and Jan van der Laan, Computer Systems Group, Nuclear Physics Section, National Institute for Nuclear Physics and High Energy Physics, The Netherlands), and the data fitted to a single-exponential decay equation.The resulting mixing rate constant was 5.7 s -1 , indicating that mixing was essentially complete within about 0.6 s after injection.
Computer simulations of the effect of cosubstrate on cyclooxygenase kinetics were carried out by numerical integration using the SCoP program (Simulation Resources, Redlands, CA).Calculations were based on the PGHS mechanism shown in Fig. 1, which is derived from our earlier mechanism (28), with addition of a two-step inactivation process (steps k 7 and k 9 ) involving Intermediate III to parallel the peroxidase inactivation mechanism described by Wu et  (conversion of Intermediate III to E(inact)), 1 s -1 and 0.05 s -1 , respectively, reflect observed rates during PGHS-1 peroxidase self-inactivation (29).Reduction of Intermediate III by cosubstrate (k 8 ) was set up as a saturable process, with a maximum rate of 0.002 s -1 and a K m for cosubstrate (K mAH ) of 0.1 mM based on the phenol concentration reported to give half-maximal stimulation of the cyclooxygenase (12,13).Two sets of cyclooxygenase reaction time courses are generated in each simulation run.The first time course is that for changes in arachidonate concentration ([AA]) in the bulk reaction (Eq.9).The second time course is for changes in arachidonate concentration that would be determined from the output of an oxygen electrode immersed in the bulk reaction ([AAe]) (Eq.10).The latter reflect attenuation by a membrane diffusion factor (k Diff ); the value of k Diff , 0.25 s -1 , approximates the response rate of the YSI oxygen electrode covered by a "Standard" membrane (15,27,31,32).The predicted timecourses are given in terms of arachidonate consumption.Because of the fixed substrate stoichiometry (27), multiplication by a factor of two provides the corresponding value for oxygen consumption.

RESULTS
The tyrosyl radical at Tyr385 in Intermediate II is thought to be a key intermediate in cyclooxygenase catalysis (33).Recently, self-inactivation of PGHS-1 peroxidase activity was found to involve a pathway originating with Intermediate II (29).It thus seemed possible that reduction of an oxidized intermediate in the self-inactivation pathway by cosubstrates might decrease self-inactivation and thereby increase the cyclooxygenase catalytic rate.This possibility was tested by computer simulation of cyclooxygenase reaction kinetics at various cosubstrate levels.The mechanism used for the computer simulations (Fig. 1) was adapted from an earlier mechanism (28) by adding an additional step in the selfinactivation process (k 7 in Fig. 1) to reflect the observation of Intermediate III (29), and by allowing reduction of Intermediate III species by cosubstrate to restore the cyclooxygenase center to ground state (k 8 step in Fig. 1), thereby attenuating self-inactivation.Computer simulations were run to predict the peak cyclooxygenase velocity (also termed the optimal velocity, or V opt ) both for the bulk reaction (cuvette value) and as detected by an oxygen electrode (electrode value) at a variety of cosubstrate concentrations (Fig. 2).The simulations predicted that the cuvette V opt increases slightly at low cosubstrate concentrations, peaking at about 20% over control at 0.2 mM cosubstrate, and then decreases at higher cosubstrate levels.Surprisingly, the electrode V opt value was predicted to increase markedly at lower cosubstrate levels, peaking at about three times the control value, and then decrease at higher by guest on January 21, 2018 http://www.jbc.org/Downloaded from cosubstrate levels.This predicted bimodal response of uncorrected oxygen electrode V opt to cosubstrate level is just the pattern observed in PGHS-1 reactions in the presence of a variety of cosubstrates (12,13,34).
The difference in the predicted electrode and cuvette V opt responses to cosubstrate level seen in Fig. 2 can be traced to the effect of cosubstrate on the cyclooxygenase reaction trajectory and in the response characteristics of the oxygen electrode (Fig. 3).In the absence of cosubstrate, cyclooxygenase catalysis in the cuvette is predicted to occur in a sharp burst, with the velocity peaking within 1 s after beginning the reaction.The predicted oxygen electrode response (a value of 0.25 s -1 was used to simulate the behavior with a standard membrane) is attenuated and spread out; the predicted electrode V opt is thus smaller and is registered later than the cuvette V opt (Fig. 3).Addition of cosubstrate is predicted to slow and prolong cyclooxygenase catalysis in the cuvette (Fig. 3).Slower reactions are less attenuated by the electrode response than fast reactions, so the electrode V opt at 1 mM cosubstrate is predicted to be larger than the control value, even though the cuvette V opt is predicted to be somewhat lower than the corresponding control (Fig. 3).
These computer simulations suggested that uncorrected oxygen electrode observations may give a distorted picture of the effects of reducing cosubstrates on the cyclooxygenase catalytic velocity.This prompted a detailed re-examination of the actions of cosubstrate on cyclooxygenase kinetics.The oxygen electrode is particularly valuable to kinetic analyses because it allows convenient and continuous monitoring of cyclooxygenase catalysis.Once corrected for membrane dampening, electrode values for cyclooxygenase activity in relatively slow reactions track well with results obtained with cumbersome, Decreasing the reaction temperature was the first approach used to slow the cyclooxygenase kinetics so that the bulk of the reaction occurred after completion of mixing, which required about 0.6 s -1 (see Materials and Methods section).The cyclooxygenase kinetics at 25 0 C and 5 0 C in the presence and absence of phenol are compared in Fig. 4. Without phenol at 25 0 C (upper panel in Fig. 4), the uncorrected electrode data registers oxygen consumption beginning after 1 s; correction for membrane dampening makes it apparent that oxygen consumption begins well before 1 sec, within the time required for mixing.Addition of 1 mM phenol slowed the cyclooxygenase reaction, with oxygen consumption beginning after 2 s in the uncorrected electrode data and at around 1 s in the corrected data.As expected, phenol also increased the number of cyclooxygenase turnovers before self-inactivation, with the reaction extent at 30 s increasing from 4.2 µM O 2 in the control to 16.3 µM O 2 with 1 mM phenol.This represents an increase in catalytic turnovers from 110 to 430.
The reaction at 5 0 C without phenol were considerably slower than at 25 0 C, with oxygen consumption apparent only after 4 s in the corrected electrode data (Fig. 4, lower panel).A considerably higher level of PGHS-1 was used for reactions at 5 0 C (47 nM) than at 25 0 C (19 nM) to improve the signal/noise ratio in the early part of the reaction at the lower temperature.Addition of 1 mM phenol at 5 0 C appeared to affect primarily the extent of the cyclooxygenase reaction, which reached a corrected O 2 consumption of 9 µM at 30 s, compared to about 4 µM in the control.It is clear from the results in Fig. 4 that the bulk of the cyclooxygenase reaction at 5 0 C, both with and without phenol present, occurred after mixing in the cuvette was complete.Thus, reducing the reaction temperature allows the cyclooxygenase kinetics to be accurately monitored with the oxygen electrode.
Detailed cyclooxygenase reaction oxygen consumption profiles were determined at reduced temperature (4 o C) for four cosubstrates (phenol, acetaminophen, TMPD, and Trolox) over a range of cosubstrate concentration.The results for selected levels of phenol are presented in Fig. 5.The corrected oxygen consumption traces form a nestled set of curves, with increases in the phenol concentration producing an increase in the ultimate extent of the reaction, but having little effect on the slope of the early part of the reaction.The relative reaction extents and maximal velocities (V opt ) calculated from these corrected oxygen consumption traces are plotted as a function of cosubstrate level in Fig. 6, along with the values for other levels of phenol, and for similar reactions with the other cosubstrates.A consistent pattern was obtained with the four cosubstrates.The reaction extent, reflecting the number of catalytic turnovers before cyclooxygenase inactivation, increased with the cosubstrate level, peaking at a value 2.9-3.6 times that of the control.The corrected V opt values, on the other hand, increased only slightly as the cosubstrate concentration was raised, peaking at 22-61% above the control values; further increases in cosubstrate levels led to progressive declines in the V opt values.Uncorrected V opt values peaked at about twice the control values as the cosubstrate concentrations were raised (data not shown), demonstrating that the membrane dampening exaggerated changes in cyclooxygenase velocity, a similar pattern to that predicted in the computer simulations (Fig. 3).The increases in relative reaction extent with cosubstrate present at 4 o C (Fig. 6) were about 7-fold greater than the increases in corrected V opt , except for TMPD, where the increase in extent was about 4-fold that in V opt .Thus, all of the cosubstrates tested dramatically increased the number of cyclooxygenase catalytic turnovers before self-inactivation, but had only a very modest impact on the peak catalytic rate.
As a second approach to slowing the cyclooxygenase reaction for detailed analysis with the oxygen electrode, reactions were run at elevated pH.PGHS-1 cyclooxygenase kinetics were examined at selected pH values between 7.2 and 10.0.Reaction in borate buffer at pH 9.8 was found to slow the cyclooxygenase kinetics by about an order of magnitude compared to reaction at pH 7.2, so that most of the reaction occurred well after the mixing time of the cuvette (data not shown).Accordingly, this buffer was used for reaction of PGHS-1 with arachidonate in the presence of varying levels of the each cosubstrate at 23 o C, much as described above for the low temperature reactions at pH 7.2.
Cyclooxygenase reaction trajectories for reaction at pH 9.8 with several levels of Trolox are presented in Fig. 7. Addition of Trolox produced marked increases in the plateau level of oxygen cosumption, reflecting increases in the extent of the cyclooxygenase reaction before self-inactivation.
However, Trolox had little effect on the peak slope of the reaction trajectory, with the early part of the reaction traces almost superimposed on one another (Fig. 7).The results of quantitative analysis of the effects of Trolox and the other cosubstrates on the cyclooxygenase V opt and extent are presented in Fig. 8.
The general picture of cosubstrate effects at elevated pH is similar to that for reduced temperature (Fig. 6): cosubstrates produced substantial increases in the extent of the cyclooxygenase reaction (close to three-fold for Trolox), but only small increases in the peak velocity (60% or less for all the cosubstrates).
Maximal actions on both extent and V opt were observed at considerably lower cosubstrate concentrations in reactions at elevated pH than in those at reduced temperature.For example, Trolox, phenol and acetaminophen above 250 µM dramatically depressed the V opt in reactions at pH 9.8 (Fig. 8), whereas strong inhibition of the rate was observed only above 1 mM levels of the same cosubstrates in reactions at 9.8 that the reaction plateau was not reached during the 300 s observation period; the corresponding values for reaction extent are thus underestimates.The actions of TMPD were particularly affected at elevated pH, with the increases in extent and V opt peaking below 10 µM at pH 9.8 (Fig. 8), whereas peak effects required about 100 µM TMPD in reactions at 4 o C (Fig. 6).Maximal stimulation of reaction extent at pH 9.8 was less than at 4 o C for each of the cosubstrates (Figs. 6 and 8), suggesting that the protective actions of the cosubstrates were less effective, or their inhibitory actions more effective, at the elevated pH.The rank order of potency of the cosubstrates (from concentrations needed for half-maximal stimulation) appeared to be the same at pH 9.8 and at 4 o C: TMPD > Trolox > acetaminophen > phenol.
Overall, the results in reactions slowed by elevated pH confirmed those from reactions at decreased temperature: the cosubstrates produced large increases in cyclooxygenase reaction extent, but only small increases in the peak cyclooxygenase rate.

DISCUSSION
Cosubstrates generally exhibit a bimodal effect on the cyclooxygenase velocity, with stimulation at low levels and inhibition at higher levels (34).The balance between the stimulatory and inhibitory modes depends on the cosubstrate structure; inhibition can be the predominant aspect with some cosubstrates (35).As the stimulatory mode was under study here, the four cosubstrates used were selected as compounds reported to give prominent stimulation of PGHS-1 cyclooxygenase velocity (16,(36)(37)(38).
Despite the varied structures, the four cosubstrates produced similar, small increases in the corrected by guest on January 21, 2018 http://www.jbc.org/Downloaded from cyclooxygenase velocity (Figs. 6 and 8).This consistent outcome indicates that the modest velocity increase is a general property, and not dependent on a particular cosubstrate structural features.
Previous studies combining oxygen electrode and optical measurements indicated that cyclooxygenase V opt values calculated from uncorrected oxygen electrode data can significantly underestimate the true peak velocities (27).The present results indicate that the uncorrected electrode data also can introduce a previously unsuspected type of distortion.In this case, a short burst of cyclooxygenase activity with a high peak velocity can produce a lower apparent V opt than a persistent burst of cyclooxygenase activity with a lower peak velocity (Figs. 3 and 4).
Each of the cosubstrates tested allowed more cyclooxygenase catalytic cycles before selfinactivation (Figs. 6 and 8), making the burst of cyclooxygenase activity more persistent, and producing distorted velocity values from the uncorrected electrode data.Thus, V opt values calculated from the raw electrode signals indicated about a two-fold increase in cyclooxygenase V opt for each of the cosubstrates tested 2 (data not shown), when in fact the actual increase was at most 60% above the control (Figs. 6 and 8).Most previous detailed examinations of the effects of cosubstrate on the cyclooxygenase velocity used uncorrected oxygen electrode signals and reported 2-7 fold increases in V opt with various cosubstrates (10)(11)(12)(13).These large apparent increases in cyclooxygenase velocity must be reconsidered in light of the current results.One study used spectrophotometric measurements, which don't require correction, to monitor the effects of Trolox on PGHS-1 oxygenase activity during reaction with 11, 14-eicosadienoic acid and found a 60% increase in peak velocity with this cosubstrate (16).This is quite comparable with the 40-60% increases observed in the present study for Trolox in reactions with arachidonate (Figs. 6 and 8).
For PGHS-1 reconstituted with MnPPIX instead of heme (MnPGHS-1), cosubstrates have been reported to have no effect on the V opt calculated from uncorrected oxygen electrode signals, even though the reaction extent is increased several-fold (11,39).This may be due to the slowly accelerating kinetics of MnPGHS-1 cyclooxygenase even in the absence of cosubstrate, minimizing the dampening effect of the electrode membrane and the associated distortion in velocity profile.The sensitivity of the V opt determined from uncorrected oxygen electrode readings to variations in electrode responsiveness and variations in the accelerative and decelerative aspects of the cyclooxygenase catalytic burst may help explain the large variation in the reported stimulatory effect of phenol on V opt , from 2-3-fold (10,12) to over 6-fold (11).
Earlier PGHS-1 cyclooxygenase studies found that cosubstrates produced roughly coordinate increases in the uncorrected velocity and the overall extent of the reaction (9,12,36,40).It's now clear that the increases in reaction extent are much larger than those in the actual peak velocity (Figs. 6 and 8).The cyclooxygenase reaction extent is accurately reported by the oxygen electrode because changes in oxygen concentration towards the end of the reaction are much slower than the electrode response kinetics.The magnitude of the increases in reaction extent with cosubstrate observed here (2.5-3.5 fold at pH 7.2; Fig. 6) are comparable to those reported previously (9,12,36,41).Increases in cyclooxygenase reaction extent by cosubstrates have been attributed to attenuation of self-inactivation of oxidized intermediates (9).The mechanism in Fig. 1 includes a step for cosubstrate-dependent rescue from a self-inactivation pathway (the k 8 step).This mechanism does predict large increases in reaction extent with increased cosubstrate levels, but the increases are linear throughout the concentration range in Fig. 2, in spite of the saturable form assigned to the k 8 step (data not shown).This differs from the observed saturable effect of cosubstrate on cyclooxygenase extent (4,12,41).This divergence between predicted and observed behaviour indicates that this part of the mechanism needs to be refined.It's worth noting that at pH 9.8 cosubstrate-induced increases in extent tended to be lower, and peak increases came at lower cosubstrate concentrations (Fig. 8).This pH dependence of the stimulatory effectiveness of some cosubstrates may offer clues about the underlying process.
A large stimulation of cyclooxygenase V opt at lower levels of cosubstrate has become an important criterion for testing tentative PGHS reaction mechanisms, and considerable ingenuity has been devoted to devising mechanistic schemes to satisfy this criterion (11,13,15,16), One common strategy has been to postulate that reducing cosubstrates react with oxidized enzyme intermediates which are catalytically inactive or prone to self-destructive reactions (11,13,41).As for cyclooxygenase catalysis itself, there have been several proposals for the key enzyme oxidant in the process, including ferrous and perferryl (Compound I) states of the heme, and a tyrosyl radical (11,21,42).Accumulating spectroscopic and kinetic observations for reactions with arachidonate or eicosadienoic acid (23)(24)(25)(26) have now established the tyrosyl radical as the central cyclooxygenase oxidant.
The original branched-chain tyrosyl radical mechanism postulated that cosubstrates act as reductants in the peroxidase cycle itself and quench the tyrosyl radical in Intermediate II (21).In such a mechanism, increased cosubstrate levels are expected to inhibit cyclooxygenase velocity by depleting Intermediate II and by lowering the peroxide level driving formation of Intermediate II.This inability to account for a large increase in cyclooxygenase velocity has been recognized as weak aspect of the basic branched chain mechanism (13,15).One approach to remedy this deficiency of the branched-chain tyrosyl radical mechanism was to have cosubstrate act in a catalytic manner to speed hydrogen transfer from fatty acid to tyrosyl radical during the cyclooxygenase reaction (17), but this seems inconsistent with the limited space near the radical position (Tyr385) in the cyclooygenase active site of PGHS-1 (18).A more recent modification to the branched chain mechanism (28) incorporated explicit hypotheses for selfinactivation and provision for the observed ability of peroxidase catalysis to operate independently of cyclooxygenase site events (43).However, computer simulations using the mechanism incorporating these changes still predicted a monotonic decrease in cyclooxygenase activity as cosubstrate levels increase (R. Kulmacz, unpublished results).
A recent detailed study of peroxidase self-inactivation found that the inactivation process begins with Intermediate II, the key tyrosyl radical species in cyclooxygenase catalysis, and proceeds via a distinct spectral intermediate, termed Intermediate III (29).With the presumption that cyclooxygenase inactivation coincides with peroxidase inactivation, the two-step inactivation process was incorporated into the mechanistic model (k 7 and k 9 in Fig. 1).To account for the long-recognized ability of cosubstrates to attenuate self-destructive events involving oxidized intermediates (9), a step was added for retrieval of Intermediate III from the inactivation pathway via a saturable, cosubstrate-dependent step (k 8 in Fig. 1).
With these changes, the branched chain reaction mechanism predicts that peroxidase cosubstrates can slightly stimulate the cyclooxygenase V opt , although the predicted increase (Fig. 2) is somewhat less than that observed with the enzyme itself (Figs. 6 and 8).
The overall pattern predicted by the revised branched chain mechanism in Fig. 1, with lower levels of cosubstrates producing a small stimulatory effect on corrected cyclooxygenase velocity and higher itself once the electrode signals are corrected (Figs. 6 and 8).Thus, accounting for the oxygen electrode response characteristics resolves the previous inadequacies of the branched chain mechanism with regard to cosubstrate effects.
With the convincing demonstration that cosubstrates produce only modest increases in cyclooxygenase velocity, the present results correct a longstanding misperception about cyclooxygenase kinetic characteristics and provide more realistic constraints for cyclooxygenase kinetic behavior.This should considerably advance elucidation of the PGHS reaction mechanism, regardless of the particular type of mechanism being considered.

Figure 1 :
Figure 1: Mechanistic model used for kinetic simulations.Redox states are noted in each intermediate for

Figure 2 :
Figure 2: Predicted effects of cosubstrate on cyclooxygenase V opt .Computer simulations based on the

Figure 3 :
Figure 3: Predicted effects of cosubstrate on cyclooxygenase reaction trajectory.Detailed predictions of

Figure 4 :
Figure 4: Effects of decreased temperature on cyclooxygenase kinetics.Cyclooxygenase kinetics were

Figure 5 :
Figure 5: Effects of phenol on cyclooxygenase reaction kinetics at 4 0 C. The cyclooxygenase reaction was

Figure 6 :
Figure 6: Effects of cosubstrates on cyclooxygenase reaction velocity and extent at 4 0 C. Oxygen uptake

Figure 7 :
Figure 7: Effects of Trolox on cyclooxygenase kinetics at elevated pH.The cyclooxygenase reaction was

Figure 8 :
Figure 8: Effects of cosubstrate on cyclooxygenase velocity and extent in reactions at elevated pH.