(Received for publication, January 22, 1997)
From the Division of Hematology, Department of
Internal Medicine, University of Texas Health Science Center at
Houston, Houston, Texas 77030 and the ¶ Department of Chemistry
and Institute of Molecular Biophysics, Florida State University,
Tallahassee, Florida 32306
Prostaglandin H synthase (PGHS) is a
heme protein that catalyzes both the cyclooxygenase and peroxidase
reactions needed to produce prostaglandins G2 and
H2 from arachidonic acid. Replacement of the heme group by
mangano protoporphyrin IX largely preserves the cyclooxygenase
activity, but lowers the steady-state peroxidase activity by 25-fold.
Thus, mangano protoporphyrin IX serves as a useful tool to evaluate the
function of the heme in PGHS. A detailed kinetic analysis of the
peroxidase reaction using 15-hydroperoxyeicosatetraenoic acid
(15-HPETE), EtOOH, and other peroxides as substrates has been carried
out to compare the characteristics of PGHS reconstituted with mangano
protoporphyrin IX (Mn-PGHS) to those of the native heme enzyme
(Fe-PGHS). The rate constant describing the reaction of Mn-PGHS with
15-HPETE to form the oxidized, Mn(IV) intermediate with absorption at
420 nm, exhibits saturable behavior as the 15-HPETE concentration is
raised from 10 to 400 µM. This is most likely due to the
presence of a second, earlier intermediate between the resting enzyme
and the Mn(IV) species. Measurements at high substrate concentrations
permitted resolution of the absorbance spectra of the two oxidized
Mn-PGHS intermediates. The spectrum of the initial intermediate,
assigned to a Mn(V) species, had a line shape similar to that of the
later intermediate, assigned to a Mn(IV) species, suggesting that a
porphyrin -cation radical is not generated in the peroxidase
reaction of Mn-PGHS. The rate constant estimated for the formation of
the earlier intermediate with 15-HPETE is 1.0 × 106
M
1 s
1 (20 °C, pH 7.3). A
rate constant of 400 ± 100 s
1 was estimated for the
second step in the reaction. Thus, Mn-PGHS reacts considerably more
slowly than Fe-PGHS with 15-HPETE to form the first high-valent
intermediate, but the two enzymes appear to follow a similar overall
reaction mechanism for generation of oxidized intermediates. The
difference in rate constants explains the observed lower steady-state
peroxidase activity of Mn-PGHS compared with Fe-PGHS.
Prostaglandin H synthase (PGHS)1 catalyzes the first irreversible step in the biosynthesis of many important prostanoids. PGHS exhibits two enzymatic activities: a cyclooxygenase activity which converts arachidonic acid to PGG2, and a peroxidase activity which transforms PGG2 to PGH2. Lambeir et al. (1) observed that intermediates similar to horseradish peroxidase Compounds I and II are formed during the peroxidase reaction of PGHS. The subsequent discovery of a tyrosyl radical intermediate (2-5) suggested that the PGHS peroxidase reaction resembles a hybrid of the reactions of horseradish peroxidase and cytochrome c peroxidase. In this hybrid mechanism, PGHS first interacts with a peroxide substrate to generate Intermediate I, equivalent to Compound I of horseradish peroxidase. Intermediate I then converts to Intermediate II, equivalent to Complex ES of cytochrome c peroxidase, through an intramolecular electron transfer from a tyrosine residue to the oxidized porphyrin. Generation of this tyrosyl radical has been proposed to be necessary for cyclooxygenase activity (2-5).
Replacement of the heme group of PGHS by mangano protoporphyrin IX was originally reported to result in an almost complete loss of peroxidase activity, whereas the cyclooxygenase activity was largely preserved (6). This finding suggested that there might be a peroxidase-independent cyclooxygenase reaction mechanism. However, later studies showed that Mn-PGHS has true peroxidase activity and formed a high-valence peroxidase intermediate, but at a much slower rate than Fe-PGHS (7-9). Interestingly, only one intermediate was observed in the Mn-PGHS peroxidase cycle (7-9), suggesting that the mechanism might be fundamentally different from Fe-PGHS. The present work describes a detailed comparison of peroxidase kinetics of Fe- and Mn-PGHS. The results demonstrate that the peroxidase reaction of Mn-PGHS proceeds through two intermediates that are analogous to Intermediates I and II found with Fe-PGHS. Thus, the overall Fe-PGHS peroxidase reaction mechanism appears to be conserved in Mn-PGHS, despite the change in prosthetic group.
Hydrogen peroxide, Fe-PPIX, aspirin, and indomethacin were purchased from Sigma. Mn-PPIX was obtained from Porphyrin Products (Logan, UT). Arachidonic acid was from NuChek Preps (Elysian, MN). [1-14C]Arachidonic acid was from Amersham Corp. Ethyl hydroperoxide was purchased as a 5% aqueous solution from Polysciences Inc. (Warrington, PA) and trans-5-phenyl-4-pentenyl hydroperoxide (PPHP) was the product of Cayman Chemical Co. (Ann Arbor, MI).
15-Hydroperoxyeicosatetraenoic acid (15-HPETE) was prepared from [1-14C]arachidonic acid with soybean lipoxygenase as described by Graff et al. (10) and purified by silica gel chromatography. The purity of 15-HPETE was assessed chromatographically (10) and quantitated from oxidation of TMPD in the presence of excess Fe-PGHS as described previously (9).
PGH synthase was purified from ram seminal vesicles as described
previously (11) except that 5 mM glutathione was included during isoelectric focusing to increase heme depletion. Protein was
determined either by the Lowry method (12) or by absorbance at 279 nm
(116 mM1 cm
1). Residual heme
was removed by treatment with 5 mM glutathione and
DEAE-cellulose (9). Fe-PGHS was prepared by addition of a slight excess
of heme to the apoenzyme and incubation at room temperature for 30 min.
Nonspecifically bound heme was then removed by treatment with
DEAE-cellulose and gel filtration through a Bio-Rad 10DG column. The
resulting Fe-PGHS was typically >90% in the holoenzyme form, as
assayed by cyclooxygenase activity in the presence and absence of 1 µM hemin. Mn-PGHS was prepared by adding a stoichiometric
amount of Mn-PPIX and incubation for 30 min at room temperature. No
DEAE treatment was applied to Mn-PGHS as this procedure removed both
functional and nonspecifically found Mn-PPIX. The reconstituted
holoenzymes were checked for their enzymatic stability and found to be
comparable to the apoenzyme when 0.02% octyl
-D-glucoside and 0.1% Tween 20 were included in the
buffer. Acetylation of Fe- and Mn-PGHS by aspirin was performed as
described previously (13). Briefly, the holoenzyme was incubated with
0.3 mM aspirin at room temperature for at least 1 h,
followed by gel filtration to remove excess aspirin and salicylate.
Residual cyclooxygenase activity was less than 3% of the control
value. Cyclooxygenase activity was assayed polarographically at
30 °C, and peroxidase activity was measured using spectroscopic
detection of the change at 611 nm in the presence of excess TMPD at
25 °C as described previously (11).
Stopped-flow kinetic studies were conducted with two separate
instruments. Kinetic measurements below 0 °C were carried out in a
custom-built low-temperature stopped-flow machine (14). The entire flow
channels were encased in brass blocks which were cooled with 50%
methanol circulated by a Neslab ULT-80 cooling unit. All valves and
gaskets were spring loaded to prevent leakage at low temperature.
Custom software was used for control of data collection and analysis.
All other kinetic data were collected with a Bio-Sequential DX-17MV
stopped-flow instrument (Applied Photophysics, Leatherhead, United
Kingdom). Temperature-dependent kinetic measurements were
conducted over the range of 10 °C to 35 °C, limited by the
freezing point of the buffer and the enzyme thermostability. For single
wavelength kinetic data, the built-in software was used for rate
analysis. For data obtained in the kinetic scan mode at serial
wavelengths, the data sets were first manipulated by the singular value
decomposition (SVD) method (15), and then fitted to kinetic models
using the Global analysis package (16). The Global package uses a
Marquardt-Levenberg algorithm for least square analysis. The SCoP
program (Simulation Resources Inc., Berrien Springs, MI) was also used
for simulation and for fitting of single wavelength kinetic data to the
same mechanistic model. The default numerical integrator used for
solving ordinary differential equations in our kinetic simulations,
Adrunge, is a Runge-Kutta algorithm with automatic adjustment of time
increment. The "GRID" method and principal axis method, or PRAXIS,
were used to carry out fitting procedures.
A detailed analysis of the reaction of Fe-PGHS with several
peroxide substrates was conducted as a benchmark for comparison with
Mn-PGHS. Reactions with a lipid hydroperoxide substrate, 15-HPETE, and
a water-soluble peroxide, EtOOH, were carried out at reduced
temperatures (0 or 10 °C) to slow down the overall rates of
reaction and maximize the range of substrate concentrations amenable to
stopped-flow observation. Reaction progress was monitored at two
wavelengths. Absorbance changes at 410 nm (the Soret maximum of resting
PGHS) were used to monitor the conversion of the resting, ferric enzyme
to Intermediate I. Observations at 424 nm (the reported isosbestic
point for the resting enzyme and Intermediate I (17)) were used to
measure formation of Intermediate II.
The reaction of 0.5 µM Fe-PGHS with excess 15-HPETE in
30% glycerol buffer solution was very rapid even at 0 °C (data not shown). The rate of decrease in A410 was faster
than that of the increase in A424, with the
latter exhibiting a brief lag phase. The changes in
A410 and A424 (after the
initial lag phase) were fitted by a single exponential function to
provide estimates of the pseudo first-order rate constants,
k1,obs and k2,obs, which describe the rates of formation of Intermediates I and II,
respectively. The dependence of k1,obs and
k2,obs on the 15-HPETE concentration is shown in
Fig. 1 (top panels). Similar data were
acquired using EtOOH as the substrate (Fig. 1, bottom
panels).
The values of k1,obs and k2,obs are much lower at equivalent concentrations of EtOOH compared with 15-HPETE. The value of k1,obs increases linearly with 15-HPETE concentration over the range of 0-40 µM concentration range studied (Fig. 1, top left). The y intercept is very close to zero, indicating that the formation of Intermediate I is essentially irreversible. This behavior is consistent with the following mechanism,
![]() |
(Eq. 1) |
![]() |
(Eq. 2) |
The value of k2,obs also increases with increasing concentrations of both 15-HPETE and EtOOH (Fig. 1, right panels). However, the value starts to level off at high concentrations of both peroxides, indicating that the conversion of an earlier intermediate to Intermediate II becomes rate-limiting. Two interpretations for this curvilinear dependence of k2,obs on peroxide concentration are considered here. The first interpretation is based on a two-step, sequential formation of Intermediate II, where the second step is unimolecular,
![]() |
(Eq. 3) |
Alternatively, the curvilinear dependence on peroxide concentration could be a saturation kinetic phenomenon associated with a pre-equilibrium complex formation between Intermediate I and a second molecule of peroxide prior to reduction of the heme center of Intermediate I to Intermediate II, as proposed by Bakovic and Dunford (18),
![]() |
(Eq. 4) |
![]() |
![]() |
(Eq. 5) |
In this mechanism, the formation of the Intermediate I-ROOH complex from Intermediate I and ROOH is fast. Thus, k1,obs essentially reflects the formation of the complex, whereas k2,obs reflects conversion of the complex to Intermediate II. At high [ROOH], it is the complex, and not Intermediate I, that accumulates. It is also important to note that Intermediate II in this mechanism is only one oxidizing equivalent above the resting enzyme, and in this sense, is equivalent to peroxidase Compound II, which lacks an amino acid radical. This is distinct from Intermediate II in the two-step mechanism (Equation 3), which, as an oxyferryl tyrosyl radical, remains two oxidizing equivalents above the resting enzyme, making it analogous to Compound ES of cytochrome c peroxidase.
The k2,obs data were fitted empirically to a
hyperbolic function to estimate k2 (Equation 3)
or k3 (Equation 4) to be 110 s1
for the reaction with 15-HPETE and 25 s
1 for the reaction
with EtOOH (Fig. 1, right panels). Validation of the use of
a hyperbolic function for estimation of k2 in
the first mechanism (Equation 3) is detailed below. Equation 5 directly predicts a hyperbolic response to [ROOH] for the second mechanism. An
average value of 39 ± 0.5 s
1 for
k2 (or k3 in the second
mechanism) was obtained for EtOOH in three separate experiments at
1 °C, considerably lower than that observed with 15-HPETE,
indicating that the rate of conversion from Intermediate I to II was
dependent on the identity of the peroxide.
To examine further the influence of hydroperoxide structure on the rate
of Intermediate II formation, the rate constants
k1 and k2
(k3 in Equation 4) were also evaluated for
reaction of Fe-PGHS with PPHP and HOOH. As summarized in Table
I, the values of k1 and
k2 (k3) for PPHP were
9 × 106 M1 s
1
and 280 s
1, respectively, at 1 °C. These values are
similar to those obtained with 15-HPETE. With HOOH, the rates of change
at 410 and 424 nm were slow and equal, indicating that the first step
remained rate-limiting even at the highest substrate
concentration tested. Because of this, only a rough estimate for
the value of k1, 1 × 103
M
1 s
1 at
10 °C, could be
obtained (Table I).
|
The effects of two cyclooxygenase inhibitors, indomethacin and aspirin, on the individual steps in Fe-PGHS peroxidase catalysis were also examined. The values of k1 and k2 (k3 in Equation 4) for the reaction of Fe-PGHS pretreated with aspirin and indomethacin with 15-HPETE and EtOOH were measured (Table I). These values were very similar to those for the untreated enzyme, indicating that these two nonsteroidal anti-inflammatory agents do not affect these steps in the peroxidase reaction of Fe-PGHS.
Only one optical intermediate, with an absorbance maximum near 420 nm, was observed earlier in the reaction of peroxides with Mn-PGHS (7, 8). Thus, the reaction was monitored at this wavelength. The lower peroxidase activity of Mn-PGHS permitted kinetic measurements with a wide range of substrate concentrations and at higher temperatures than was possible with Fe-PGHS. Surprisingly, the observed rate constant for the absorbance change at 420 nm, kobs, was found to be a saturable function of the 15-HPETE concentration (Fig. 2, left). Similar results were obtained with two other batches of Mn-PGHS (data not shown), showing that the saturable response at higher levels of peroxide is a reproducible characteristic of the Mn-PGHS peroxidase reaction. This behavior points to the existence of at least one earlier intermediate in the Mn-PGHS peroxidase pathway prior to the formation of the intermediate whose formation is being monitored at 420 nm. The recognition that there are two intermediates in the Mn-PGHS peroxidase reaction pathway suggests that it closely parallels that for Fe-PGHS. Thus, we refer to these two intermediates as I and II, respectively. As with Fe-PGHS, two variants of the mechanism for Mn-PGHS were considered. The first is a simple, two-step sequential mechanism analogous to that shown in Equation 3 for Fe-PGHS,
![]() |
(Eq. 6) |
![]() |
(Eq. 7) |
![]() |
![]() |
(Eq. 8) |
The reaction of EtOOH with Mn-PGHS was much slower than with Fe-PGHS.
The observed rate was linearly dependent on EtOOH concentration up to
about 20 mM but showed signs of being saturable at higher levels (Fig. 2, right). The second-order rate constant for
the formation of Intermediate I (k1 in Equations
6 and 7) calculated from the data in Fig. 2 (inset in
right panel) is 6.8 × 102
M1 s
1. Estimation of the values
of k2 (Equation 6) and k3
(Equation 7) for reaction with EtOOH was not attempted because of the
potential for Mn-PPIX destruction at the high EtOOH concentrations
needed to approach saturation. As observed with Fe-PGHS, pretreatment of Mn-PGHS with indomethacin or aspirin to inhibit the
cyclooxygenase activity had only marginal effects on the rate
of the peroxidase reaction with either 15-HPETE or EtOOH
(Table II).
|
Stopped-flow kinetic measurements at multiple wavelengths were used to
resolve the electronic absorbance spectra of Mn-PGHS peroxidase
Intermediates I and II. The analytical approach was first validated
with Fe-PGHS, where the intermediate spectra are known. The Fe- or
Mn-PGHS and 15-HPETE concentrations were carefully selected to maximize
the transient concentration of Intermediate I, and yet minimize the
fraction of the reaction occurring in the instrumental dead time (~3
ms). Under the experimental conditions used, computer modeling with the
rate constants described above predicts a large difference between Fe-
and Mn-PGHS in the peak accumulation of Intermediate I (~60% of
total enzyme for Fe-PGHS and ~8% for Mn-PGHS) (data not shown).
Kinetic scan data obtained at individual wavelengths in the reactions
under optimized conditions were processed through a SVD routine to
identify the important kinetic and wavelength components and to remove
the background noise. These SVD-simplified data were then fit to a
sequential model (Equations 3 or 6) by a nonlinear least square
regression in the Global (or Glint) analysis package to obtain resolved
spectra for the individual intermediates, which are shown in Fig.
3.
The spectra of Fe-PGHS and its peroxidase intermediates (Fig. 3) are very similar to those reported by Lambeir et al. (1) and Dietz et al. (17). The Soret band maxima are at 410, 412, and 416 nm for resting Fe-PGHS, Intermediate I and II, respectively. The isosbestic point between resting Fe-PGHS and Intermediate I is 428 nm, that between resting Fe-PGHS and Intermediate II is 422 nm, and those between Intermediate I and Intermediate II are 408 and 448 nm. These isosbestic points are only slightly different from those values reported by Dietz et al. (17). These results with Fe-PGHS confirmed the reliability of the SVD analytical approach.
The spectra obtained for Mn-PGHS and its peroxidase intermediates exhibit quite different line shapes from the corresponding Fe-PGHS intermediates (Fig. 3, right). The maxima for resting Mn-PGHS are at 372 and 468 nm, while the spectra of Intermediates I and II both peak at 420 nm. Mn-PGHS Intermediate II shows two additional peaks at 368 and 464 nm. The relatively high noise level apparent in the spectrum of Mn-PGHS Intermediate I (Fig. 3) is consistent with a smaller contribution to the total absorbance during the kinetic measurements. The isosbestic points between resting Mn-PGHS and Intermediate I are 364, 400, and 446 nm, those between resting enzyme and Intermediate II are 352 and 444 nm, and those between Intermediates I and II are 368, 406, and 433 nm.
The validity of SVD data processing and the fitting method used in the
Global (Glint) package was cross-checked as follows. The SCoP numerical
program was used to conduct a quantitative prediction of the original
kinetic data at individual wavelengths, employing the experimentally
determined rate constants for the first two peroxidase steps shown in
Equations 3 and 6 with 15-HPETE as substrate (Figs. 1 and 2, and Table
I) and the molar absorbance coefficients from the resolved spectra in
Fig. 3. Very good fitting to experimental stopped-flow traces was
obtained at 410 and 420 nm for Fe-PGHS (Fig. 4,
left). Fitting to the stopped-flow data at 430 and 424 was
not as good (not shown), with the major discrepancy occurring in the
deflection area, perhaps reflecting an under-representation of data
points in this time range by the SVD treatment. On the other hand, the
Mn-PGHS kinetic data were fitted well at each of the three wavelengths
examined in detail, 380 (not shown), 420 and 470 nm (Fig. 4,
right).
With these data in hand, we return to the question of whether the
two-step mechanism (Equation 3) can explain the saturation behavior
observed for Intermediate II formation (Fig. 1, right panels; Fig. 2, left panel). To examine this issue for
Fe-PGHS, a set of predicted kinetic data for Intermediate II formation (424 nm) at different 15-HPETE levels was generated by numerical integration based on Equation 3, using the experimental rate constants and the extinction coefficients at 424 nm for the three enzyme species
obtained from the resolved spectra. The predicted kinetic data are
shown in Fig. 5. The portion beyond the initial lag
phase of each simulated kinetic trace was fitted to a
single-exponential function to estimate a rate constant. These pseudo
first-order rate constants were found to approximate a hyperbolic
dependence on the 15-HPETE level (Fig. 5, inset). Thus, the
predicted k2,obs values do indeed show saturable
behavior, and the hyperbolic fit used in Fig. 1 provides a reasonable
empirical estimate for the upper limit of k2. In
addition, the rates obtained from the simulations are similar to those
obtained from actual measurements (compare Fig. 5, inset,
with Fig. 1, top right).
The temperature dependences of k1 and
k2 (or k3 in Equations 4 and 7) for the reaction of Fe-PGHS and Mn-PGHS with 15-HPETE and EtOOH
were measured to calculate thermodynamic activation parameters for
these steps. With Fe-PGHS, the Arrhenius plots were linear for both
k1 and k2 or
k3 (Equations 3 and 4) for both peroxide
substrates (Fig. 6, top panels).
The activation energy, Ea, for each step was calculated from the Arrhenius equation,
![]() |
(Eq. 9) |
![]() |
(Eq. 10) |
|
In the case of Mn-PGHS, the temperature dependence was only examined
for the k1 values. The Arrhenius plots with both
peroxide substrates showed a discontinuity near 4-5 °C (Fig. 6,
bottom panels). The linear dependence of ln k on
1/T above 5 °C was used to calculate the thermodynamic
parameters. As apparent from Table III, both the activation energies
and enthalpies of activation, using either peroxide, were similar to
those found for Fe-PGHS. However, the S+ values for
Mn-PGHS were substantially lower than those for Fe-PGHS. This is seen
most dramatically for the reaction of Mn-PGHS with EtOOH, which had a
S+ of
5.7 e.u., in sharp contrast with the value of
+5.8 e.u. obtained for Fe-PGHS.
The studies reported here enable us to compare in greater detail than reported previously the steps in the peroxidase cycles of Fe- and Mn-PGHS. Both enzymes exhibit a curvilinear dependence of k2,obs on peroxide concentration that has led us to consider two alternative mechanisms, represented by Equations 3 and 4 for Fe-PGHS and Equations 6 and 7 for Mn-PGHS. Both mechanisms can account for the saturation behavior of k2,obs, but differ significantly in other details. These two mechanisms predict the formation of Intermediate II species containing different numbers of oxidation equivalents. However, since an oxyferryl state of the heme is predicted for Intermediate II in both mechanisms, the optical spectrum of this intermediate (Fig. 3) cannot be used to distinguish between the two mechanisms. This would require the use of EPR techniques that could assess whether the Intermediate II formed at different [ROOH] contained a tyrosyl radical.
There are certain observations that favor one mechanism over the other. The three-step mechanism shown in Equation 4 predicts that the values of k2,obs would depend on the identity of the peroxide, since its redox potential should affect the efficiency of this step. On the other hand, the requirement for 2 equivalents of hydroperoxide to form Intermediate II in the three-step mechanism is in conflict with the observation that stoichiometric amounts of EtOOH generate substantial amounts of Intermediate II and tyrosyl radical (19). We have recently found that substoichiometric amounts of peroxide also generate tyrosyl radical in proportion to the amount of peroxide.2 This efficient formation of Intermediate II with limiting levels of hydroperoxide is difficult to reconcile with the stoichiometry of the three-step model.
A possible explanation for these apparently conflicting observations is
that the reaction can proceed by both mechanisms, with the proportion
of Intermediate I decomposing by the two alternate routes being
determined by the hydroperoxide concentration. Specifically, at low
peroxide concentrations, the two-step mechanism described by Equations
3 and 6 would reasonably dominate, leading to formation of the
oxyferryl tyrosyl radical Intermediate II. At high concentrations, the
Intermediate I-ROOH complex is able to form and react intramolecularly to quench the -cation radical and reduce Intermediate I to an oxyferryl intermediate before its conversion to a tyrosyl radical intermediate. This reducing action of peroxide may be a laboratory artifact associated with the high peroxide concentrations needed to
observe saturation.
An important aspect of this study is the establishment that the reaction of Mn-PGHS with peroxides also proceeds through at least two intermediates. The saturation behavior observed for k2,obs strongly supports the existence of an earlier peroxidase intermediate (I) in Mn-PGHS that precedes intermediate II. This finding resolves a puzzle raised by earlier studies which failed to find an intermediate with an oxidation state two equivalents higher than the Mn(III) resting enzyme (7-9). Intermediate I was not observed in previous studies using relatively low 15-HPETE concentrations because k1 is relatively slow and Intermediate I accumulates only when Mn-PGHS is reacted with rather high 15-HPETE concentrations (>100 µM; Fig. 2).
Fe-PGHS exhibits a 100-fold larger value of k1 than Mn-PGHS for reaction with 15-HPETE at 20 °C (Tables I and II). In contrast, the values of k2 for Fe-PGHS and Mn-PGHS differ by only a factor of two at 20 °C (Tables I and II). Therefore, it is clearly the decrease in the rate of Intermediate I formation that is responsible for the marked decrease in steady-state peroxidase activity in Mn-PGHS. Similar results have been found with other heme-dependent peroxidases. Replacement of the heme by Mn-PPIX in cytochrome c peroxidase and horseradish peroxidase decreases the value of kcat (a lower limit estimation for k1) in steady-state assays by 2-3 orders of magnitude (20, 21).
The values of k1 for both Fe- and Mn-PGHS
decrease markedly as the substrate is changed from 15-HPETE to EtOOH to
HOOH (Tables I and II). Although HOOH is still a reasonably reactive
substrate for Fe-PGHS (k1 of 103
M1 s
1 at
10 °C), its
reactivity with Mn-PGHS was too slow to be measured readily. Other
kinetic studies of Fe-PGHS and Mn-PGHS also reveal a clear preference
for lipid peroxides over water-soluble peroxides (1, 7, 9, 18). This
suggests that binding of larger hydrophobic hydroperoxides involves
interaction with a lipid binding subdomain near the heme pocket. As the
aliphatic chain of the substrate shortens, the major binding
contribution shifts from the secondary hydrophobic site to the
metalloporphyrin, revealing the intrinsic difference in reactivity
between Fe-PPIX and Mn-PPIX toward the peroxy function itself. The
x-ray crystallographic data indicate the heme pocket in PGHS is rather
open, without a defined hydroperoxide channel like that found for the
cyclooxygenase site (22), so it is not clear what parts of this region
make up the hydrophobic peroxide-binding site.
It is of interest to consider why Mn-PPIX substitution causes the consistently decreased activity in peroxidases. Water proton relaxation measurements on native and Mn(III)-PPIX-substituted horseradish peroxidase indicate that the heme iron is predominantly penta-coordinate, whereas the metal in Mn-PPIX-substituted horseradish peroxidase is six-coordinate with water or hydroxide as the distal ligand (23). The existence of six-coordinate Mn(III) in horseradish peroxidase has also been supported by resonance Raman data (23). A slowly exchanging water or hydroxyl distal ligand in the Mn(III)-PPIX-substituted peroxidases could substantially decrease activity and even become the rate-limiting step in Intermediate I formation; once this ligand is displaced by hydroperoxide, conversion to Mn(V) proceeds rapidly. This hypothetical reaction sequence for Mn-PGHS is shown below,
![]() |
(Eq. 11) |
![]() |
The Kd of Fe-PGHS for 15-HPETE can be estimated to be submicromolar by extrapolation of the apparent Km values to zero co-substrate (24, 25). A similarly low Kd value is likely for Mn-PGHS (9). These Kd values are much smaller than the 15-HPETE levels observed to produce linear increases in kobs in the kinetic experiments (Fig. 1, top, and Fig. 2, left). The value of k1,obs at the y axis intercept was found to be near zero (Fig. 1, top, and Fig. 2, left), indicating a negligible rate for the reverse reaction. Thus an initial equilibrium binding is not likely to be established between 15-HPETE with either Fe- or Mn-PGHS, similar to the situation with horseradish peroxidase (1, 17, 26).
Formation of Intermediate II is an intramolecular electron transfer
process in the two-step sequential mechanism (Equation 3), so the rate
constant should be independent of the hydroperoxide used. However, as
discussed above, the peroxidase region can interact with hydrophobic
compounds. It is possible that binding of various peroxides or the
product alcohols leads to some persistent perturbation of the structure
of Intermediate I, and influences the rate of electron transfer from
the specific tyrosine in the cyclooxygenase site to the porphyrin
-cation radical. The distance between the porphyrin edge and the
phenyl moiety of Tyr-385 are within 5 Å (22), so that perturbation of
the heme orientation by variations in fatty acid peroxide or alcohol
structure is plausible. Such a perturbation may be reflected in the
much larger entropy change of this reaction step when 15-HPETE was the
substrate as compared with EtOOH (Table III). Perturbation by product
of this early intramolecular redox step would not be detected in
steady-state assays, where the rate-limiting step is the regeneration
of resting enzyme by reducing cosubstrate.
The similarities in the peroxidase kinetics reported here for Fe- and
Mn-PGHS suggest that the two enzymes have similar mechanisms and
intermediates. Turning attention to the intermediates in the Mn-PGHS
reaction, observations from the literature allow us to hypothesize on
their structures. Mn(V) and Mn(IV) metalloporphyrin model compounds
have been studied in aqueous and organic/aqueous solvent systems (27,
28). The optical spectra of Mn(V) model compounds exhibit a Soret band
at 418 nm (27, 28). The spectral line shape of Mn(IV) compounds are
similar, but have a more intense Soret band. Accordingly, the spectra
reported here for Intermediates I and II are consistent with those
expected for Mn(V) and Mn(IV) species, respectively. It is unlikely
that Mn-PGHS Intermediate I contains a porphyrin -cation radical,
since it would be expected to give a broad and undefined spectrum (29),
quite unlike that observed here (Fig. 3). Thus, Intermediate I in the
Mn-PGHS peroxidase reaction is probably a Mn(V)=O species whose formal
+5 charge is localized on the metal, rather than a Mn(IV)
-cation
radical. Intermediate II is most likely the corresponding Mn(IV)=O
species obtained by one-electron reduction at the metal center.
Presumably, since the manganese enzyme has cyclooxygenase activity,
this reduction is intramolecular and produces the tyrosyl radical
needed for activity.
It should be noted that the present studies concern the formation of
Intermediates I and II, and not with the conversion of these
intermediates back to resting enzyme. Although parallel mechanisms were
found for the initial peroxidase steps in Mn-PGHS and Fe-PGHS, it
remains possible that reformation of resting enzyme proceeds by
different mechanisms in the two enzymes. Evidence to support divergent
reductive pathways back to resting enzyme comes from the finding that
Mn-PGHS can oxidize the product alcohol to a keto compound, whereas
Fe-PGHS does not (7, 9, 30). In summary, our observations indicate that
Mn-PGHS follows the same overall peroxidase reaction steps as Fe-PGHS,
as depicted in Fig. 7. Both form an initial high
oxidation state Intermediate I with a formal charge of +5 and a
subsequent Intermediate II with a formal charge of +4. However, the
rate of formation of the first intermediate is much slower in Mn-PGHS,
possibly due to a more tightly bound water molecule as the sixth ligand
in Mn-PGHS. For 6-coordinate Fe-PGHS, in contrast, the weakly bound water or histidine distal ligands are easily displaced by the peroxide
(31). Intermediate I for Mn-PGHS does not appear to be a porphyrin
-cation radical species, unlike the Fe(V) PGHS species, but is more
consistent with a Mn(V) configuration. As with Fe-PGHS, Mn-PGHS forms a
peroxidase-associated radical (9). Kinetic correlation of the initial
radicals with PGG2/PGH2 formation has been
demonstrated for both Fe- and Mn-PGHS (9, 19). The structure of the
Mn-PGHS radical is sensitive to pretreatment with tetranitromethane,
just as found with the Fe-PGHS radical (4, 9). The significant
reduction in the rate of the first step in the peroxidase cycle after
replacement of iron by manganese can account for the 25-fold decrease
of peroxidase activity observed in steady-state assays (9). Taken with
the slower reaction with hydroperoxides, the increased sensitivity to
cyclooxygenase inhibition by glutathione peroxidase after substitution
with Mn-PPIX (7, 9) also fits well with the current hypothesis that the first two steps of PGHS peroxidase catalysis are integral parts of
initiating cyclooxygenase catalysis. This detailed comparison of the
peroxidase kinetics of Mn- and Fe-PGHS has provided valuable insights
into how this physiologically important enzyme harnesses the energy
released in hydroperoxide decomposition to initiate fatty acid
oxygenase catalysis.
We thank Dr. Graham Palmer and Dr. John S. Olson for very helpful discussions.