(Received for publication, August 19, 1996, and in revised form, October 7, 1996)
From the Nitrogen Fixation Laboratory, John Innes
Centre, Norwich NR4 7UH and the ¶ Department of Biochemistry,
University of Sussex, Brighton BN1 9QG, United Kingdom
Mutations have been introduced at
residues Arg-38 or His-42 in horseradish peroxidase isoenzyme C (HRPC)
in order to probe the role of these key distal residues in the reaction
of ferrous HRPC with dioxygen. The association and dissociation rate
constants for dioxygen binding to His-42 Leu, His-42
Arg,
Arg-38
Leu, Arg-38
Lys, Arg-38
Ser, and Arg-38
Gly
variants have been measured using stopped-flow spectrophotometry.
Replacement of His-42 by Leu or Arg increases the oxygen binding rate
constant by less than an order of magnitude, whereas changing the polar distal Arg-38 causes increases of more than 2 orders. These results demonstrate that His-42 and Arg-38 impede the binding of dioxygen to
ferrous HRPC, presumably by steric and/or electrostatic interactions in
the distal heme cavity. Recombinant HRPC oxyperoxidase reverted slowly
to the ferric state with no spectrophotometrically detectable intermediates and with an apparent first-order rate constant of 9.0 × 10
3 s
1, which is essentially
the same as that for the native, glycosylated enzyme. This reaction was
accelerated when His-42 was replaced by Leu or Arg
(kdecay = 0.10 and 0.07 s
1,
respectively) presumably due to the loss of the hydrogen bond between
the His-42 imidazole and the bound dioxygen. Substitution of Arg-38 by
Leu, Lys, or Gly also produced a less stable oxyperoxidase (kdecay = 0.22, 0.20, and 0.58 s
1, respectively). However, with the Arg-38
Ser
variant, a transient intermediate, proposed to be a ferric-superoxide
complex, was detected by rapid-scan stopped-flow spectrophotometry
during the conversion of oxyperoxidase to the ferric state. This
variant also exhibits an unusually high affinity for dioxygen. It is
proposed that Arg-38 interacts with the bound dioxygen to promote
superoxide character, thereby stabilizing the oxyperoxidase state and
making the binding of dioxygen to ferrous HRPC essentially
irreversible. We conclude that Arg-38 and His-42 not only promote the
heterolytic cleavage of bound hydrogen peroxide to form compound I but
also decrease the lability of the ferrous enzyme-dioxygen complex in order to suppress the formation of the inactive ferrous state.
Horseradish peroxidase (HRP1;
donor:H2O2 oxidoreductase) is a member of the
plant peroxidase superfamily (1). It is able to utilize hydrogen
peroxide to catalyze the one-electron oxidation of a wide range of
aromatic substrates. Peroxidase action involves 2e
oxidation of the enzyme by hydrogen peroxide to give an intermediate known as compound I. This then reverts to the resting enzyme via two
successive 1e
reactions with reducing substrate
molecules, the first yielding a second enzyme intermediate, compound II
(2). A third enzyme intermediate, compound III, is also observed in the
reaction of HRP with an excess of hydrogen peroxide (3). This compound, the peroxidase analogue of oxymyoglobin and oxyhemoglobin, does not
normally participate in the peroxidatic cycle of HRP (2). The spectrum
and reactivity pattern of compound III correspond to those expected for
an Fe(II)-O2 adduct or oxyperoxidase (4). Oxyperoxidase can
be formed in four different ways: (i) addition of a large excess of
H2O2 to the native ferric enzyme, (ii)
reduction of native ferric enzyme to ferroperoxidase followed by the
addition of dioxygen, (iii) reaction of H2O2
with compound II, and (iv) addition of superoxide anion to native
ferric peroxidase (2). The formation of oxyperoxidase from ferrous
peroxidase and dioxygen follows second-order kinetics with a calculated
rate constant of 5.8 × 104
M
1 s
1 at pH 7.0 and 20 °C
(4). The mechanisms of oxy-HRPC (5, 6) and lignin-oxyperoxidase (7)
decomposition have also been studied. In both cases the oxyperoxidase
form reverted slowly to the native ferric state with no
spectroscopically detectable intermediates. This contrasts with the
reactivity of ferrous yeast cytochrome c peroxidase
(CcP) with dioxygen, which differs dramatically from that of
other peroxidases. Photolysis of the CcP(II)-CO complex in
the presence of oxygen converts the enzyme to a product with an
absorption spectrum and an EPR radical signal at g = 2.00 that were identical to those of compound I formed by the reaction
of ferric CcP with hydrogen peroxide (8, 9).
Hemoproteins show great diversity in their biological functions while
retaining an essentially unaltered prosthetic group, iron
protoporphyrin IX. Although key catalytic residues in the active site
of peroxidases are highly conserved (10), in the context of the present
paper it is useful to compare HRPC with CcP. The complete
three-dimensional structure of CcP is known from
crystallographic data at 1.7-Å resolution (11). Recombinant native
CcP and several variants have also been available for
structural and mechanistic studies (12, 13). Although the primary
structure of CcP has only 18% identity with that of HRPC
(14), over 80% of this sequence identity is centered around two highly
conserved residues, the distal and the proximal histidines. An arginine residue in the distal heme pocket is also conserved in peroxidases. Distal and proximal histidines are also conserved in the globins, which
react with hydrogen peroxide much slower than do peroxidases. This
differential reactivity with hydrogen peroxide has been ascribed to the
distal arginine, which has no equivalent in the globins (15). Recent
studies with CcP showed that the guanidinium side chain of
Arg-48 is not absolutely required for the heterolytic cleavage of
peroxide (12, 13, 16), although we have shown that in HRPC this residue
does play an important role in facilitating the binding of peroxide and
modulating its subsequent reactivity (17, 18). Miller et al.
(19) determined the crystal structure of the Fe(II)-O2
complex formed by a yeast CcP mutant (Trp-191 Phe). The
refined structure showed that dioxygen forms a hydrogen bond with the
imidazole side chain of the conserved distal histidine, but not with
the guanidinium group of the conserved distal arginine. The recent
availability of site-directed mutants of HRPC has now allowed us to
investigate the role of Arg-38 and His-42 in the formation of oxy-HRPC.
The results also provide further insights into the origins of the
reactivity differences of HRP and CcP oxyperoxidases (9, 19)
and the relationship between oxyperoxidases and oxyglobins (15,
19).
HRPC (type 4B) was purchased from Biozyme Ltd.,
United Kingdom (Blaenavon, Gwent, UK) and used without further
purification. Construction of the gene, expression and purification of
HRPC*, H42L, H42R, R38L, and R38K have been described previously (17, 20, 21, 22). Construction of R38S and R38G genes was carried out as
described previously for the R38L mutant (17), except that a Ser (TCA)
or Gly (GGT) codon was introduced at position 126. The spectral
properties of the enzymes used in this study are given in Table
I. The specific activities (units mg1) of
the HRP preparations used were 920 for HRPC, 930 for HRPC*, 0.2 for
H42L, 5.0 for H42R, 5.2 for R38L, 47.7 for R38G, 7.1 for R38S, and 2.9 for R38K. Activities were measured as described previously with 0.3 mM ABTS (2,2
-azinobis(3-ethylbenzothiazolinesulfonic acid)
and 2.5 mM H2O2 (20). Reagent grade
H2O2 (30% v/v) was obtained from BDH/Merck
(Poole, UK) and its concentration determined by iodide titration with
HRPC (23). All other chemicals were of analytical grade and supplied by
BDH/Merck.
|
Fully reduced native HRPC and its variants were prepared by incubation of enzyme with sodium dithionite for several minutes in a buffer of 0.1 M sodium phosphate at pH 7.0 until no further change in the reduced Soret band could be detected. Since ferroperoxidase reacts rapidly with dioxygen, the following procedures were carried out in an anaerobic glove box operating under N2 with less than 1 ppm of O2. The excess sodium dithionite was removed from the ferroperoxidase by passing the solution through a Sephadex G-25 column that had been equilibrated with deoxygenated buffer. Anaerobic rapid-scan stopped-flow studies were performed using a Hi-Tech SF-51 stopped-flow spectrophotometer equipped with a xenon lamp, 360 nm filter, and a diode-array detection system (MG-6000, Hi-Tech Scientific, Salisbury, UK). Single wavelength experiments were carried out with the same instrument equipped with a tungsten lamp. Enzyme stock and buffer solutions (0.1 M sodium phosphate, pH 7.0) were deoxygenated for 30 min by repeated evacuation and back filling with nitrogen on a vacuum line before transfer into the glove box. Air- or oxygen-saturated buffers were transferred into the anaerobic box in hermetically sealed serum vials. Oxygen concentrations were varied from 20 to 600 µM by premixing the air- or oxygen-saturated buffer with the deoxygenated buffer in syringes with no gas head space to avoid loss of O2 from the solution. Data were recorded through a RS232 interface on a microcomputer and analyzed by fitting absorbance versus time curves to exponential functions using a least-squares minimization program supplied by Hi-Tech Scientific Ltd. Temperature was controlled at 25 °C using a Techne C-400 circulating bath with a heater-cooler also installed in the anaerobic box.
Decomposition of OxyperoxidasesOxyperoxidase solutions were prepared as described above. The disappearance of wild-type and mutant oxyperoxidases and the appearance of native ferric enzymes were followed at their respective Soret maximum indicated in Tables II and I, respectively.
|
Addition of molecular oxygen to ferrous HRPC resulted in the rapid formation of oxyperoxidase with absorption maxima at 417, 543, and 577 nm (Table II) as described previously by Wittenberg et al. (4). The reaction of ferrous HRPC with oxygen was followed at 417 nm. The observed first-order rate constant was directly proportional to the initial oxygen concentration (results not shown). At pH 7.0 and 25 °C, the oxyperoxidase subsequently reverted over a period of minutes to the native ferric state (Reaction R1). To study the kinetics of this reaction, the disappearance of oxy-HRPC and the appearance of native HRPC were followed at 417 and 403 nm, respectively. The rate constant (kdecay) was determined by fitting the absorbance versus time curves to a single exponential function and was found to be independent of oxygen concentration (Table III).
![]() |
![]() |
|
The distal histidine of heme proteins such as myoglobin and hemoglobin is thought to influence the affinity of diatomic ligands bound at the sixth coordination position of the heme. Structural and kinetic studies indicate that CO and O2 binding are destabilized by the steric hindrance of the distal histidine (15, 24, 25). Recently, we have reported a significant increase in the second-order rate constant for CO binding to HRPC mutants in which the distal histidine had been replaced by leucine or arginine (22). However, a distal histidine residue can also stabilize adducts by forming a hydrogen bond, as shown by neutron diffraction studies with oxymyoglobin (26). This interaction is critical for globins because it causes O2 to bind sufficiently tightly for efficient storage and transport while discriminating against the binding of carbon monoxide in order to decrease its toxicity (15). The crystal structure of oxy-CcP also shows a hydrogen bond between His-52 and bound dioxygen (19). We have therefore prepared the H42L and H42R variants of HRPC* and studied the effect on the kinetics of dioxygen binding to the ferrous forms of the enzyme.
The spectrum of the oxyperoxidase form of H42L was similar to that of
the wild-type enzyme with absorbance maxima at 416, 541, and 575 nm
(Table II). This compound autoxidized to the ferric form with
isosbestic points at 408, 467, 524 and 598 nm (Fig. 1)
and with a decay constant approximately 12 times faster than for the
wild-type oxyperoxidase. The decay constant was also independent of the
oxygen concentration (Table III). Similar results were obtained for the
decomposition of the H42R oxyperoxidase (Tables II and III). The
reactions of ferrous H42L and H42R with oxygen gave single exponential
absorbance changes with pseudo-first-order rate constants that
exhibited a linear dependence on the initial oxygen concentration (Fig.
2). The bimolecular rate constants for H42L and H42R
were 6- and 2-fold higher, respectively, than that for wild-type enzyme (Table III). Unlike the data for the native enzyme forms, the ordinate intercepts for the mutants were statistically significant due to higher
koff values (five determinations at each of the
oxygen concentrations). This allowed the values of
koff and the equilibrium constant for oxygen
binding to both mutants to be calculated (Table III). The substitution
of His-42 by Leu or Arg produced an increase in both the rates of
binding and dissociation, resulting in a net decrease in the affinity
for oxygen, i.e. less stable oxyperoxidases.
The Binding of Oxygen to Ferrous HRPC* Arginine 38 Mutants
Two pieces of evidence suggest that the distal arginine
impedes the reaction of ferroperoxidase with diatomic ligands such as
O2 and CO. First, proteins designed to react with oxygen
such as myoglobins and hemoglobins have an apolar residue at the
position corresponding to that of the distal arginine in peroxidases
(27, 28). Second, replacing Arg-38 by Leu in ferrous HRPC increases the
affinity for CO by 3 orders of magnitude relative to native enzyme
(22). In order to better understand the role of Arg-38 in the binding
of oxygen to ferrous HRPC, we have replaced it by other amino acids
with side chains of different lengths and/or polarity and determined
the effect on the kinetics of oxygen binding (Table III). Simple
monophasic time courses were observed for oxygen binding to each of
these mutant enzymes. Replacement of Arg-38 by Leu, Lys, or Gly had
similar effects on both kon and
koff (Fig. 3). There was a
progressive increase in the observed bimolecular rate constant for
oxygen binding in the order wild-type < R38K < R38L < R38G. The value for the R38G mutant was increased by nearly 2 orders of
magnitude relative to the wild-type enzyme. However, a more than
500-fold compensating increase in the dissociation constants
(koff) for each of the mutants resulted in a net
lower affinity for O2 than the native enzyme (Table III).
The spectra of the Arg-38 variant oxyperoxidases were all similar to
that obtained for the native enzyme (Table II).
The R38S mutant bound oxygen 50 times tighter than any one of the other Arg-38 variants (Fig. 3). The low KD results from both an increase in the second-order association rate constant (kon) and a decrease in the dissociation rate constant (koff) (Table III). Although R38S oxyperoxidase has a similar absorbance spectrum to that of native oxy-HRPC, it does exhibit a 6-nm displacement of the Soret band maximum to 411 nm. This suggests that the environment of the bound oxygen has been significantly perturbed by the introduction of a serine residue at position 38.
Fig. 4 shows a linear decrease in the logarithm of the
second-order rate constant for oxygen binding to each Arg-38 variant as
a function of the calculated side chain volumes of each of the residues
substituting for arginine at position 38. This behavior is entirely
consistent with the side chains of the residues at position 38 sterically hindering oxygen binding. However, a purely steric effect
clearly does not explain the deviance of the point for the native
enzyme with an arginine at position 38 and suggests electrostatic
factors associated with this positively charged residue are important.
We discuss below the structural and mechanistic significance of this
observation.
Decomposition of Arginine 38 Mutant Oxyperoxidase
Table III
shows the effect of various substitutions of Arg-38 on the stability of
the oxyperoxidase complex. The R38L oxyperoxidase prepared by the
reaction of ferrous enzyme with O2 reverted directly to the
native ferric enzyme with a half-life of 3 s. Similar results were
observed for the the R38K (half-life 3.5 s) and R38G (half-life 1.2 s) variants. However, the oxyperoxidase formed by the reaction of R38S ferrous enzyme with dioxygen reverted to native ferric enzyme
with the accumulation of a spectroscopically detectable transient
intermediate shown by the single wavelength stopped-flow trace in Fig.
5. Rapid-scan stopped-flow spectrophotometry reveals the
formation of a transient species with a Soret maximum at 406 nm (the
same as that for the resting enzyme) but with a higher extinction
coefficient (406 nm = 160 mM
1
cm
1; cf. 133 mM
1
cm
1 for native enzyme) (Fig.
6A). Moreover, this transient species has
peaks at 499, 573, and 630 nm and a shoulder at 537 nm (Fig. 6B). This new intermediate was converted to ferric enzyme,
as indicated by the decrease in the absorbance in the Soret region (Fig. 6C). Differences in the spectra between this new
intermediate and that of the ferric enzyme are shown in Fig.
6B. A transient species, thought to be a ferric
hydroperoxide complex (Fe(III)-OOH), with a spectrum similar to that
shown in Fig. 6 (A-C) has been previously observed during
the reaction of the R38L variant with hydrogen peroxide (17).
Replacement of the distal His-42 by Leu or Arg produced a 6- and 2-fold increase in the association rate constants, respectively, whereas the replacement of the distal Arg-38 caused more dramatic increases of up to 100-fold (Table III). Since there do not appear to be any previous reports describing the role of His-42 and Arg-38 in the binding of oxygen to ferrous HRPC, it is only possible to discuss our data in the context of other oxygen-binding heme proteins. In pig and sperm whale myoglobins, substitutions of the distal histidine (His-64) by leucine produced a 7-fold increase in kon and a greater than 100-fold increase in koff, resulting in a dramatic loss of oxygen affinity in these variants (25, 29). The equivalent substitution in glycera hemoglobin induces significantly weaker oxygen binding (15, 30). Substitution of His-42 by Leu in HRPC* produces a similar increase in kon to that for myoglobin and hemoglobin variants. These observations suggest that the access of oxygen to the heme is hindered by the distal histidine of HRPC in a similar manner to that which occurs with the globins. Since a hydrogen bond between the bound oxygen and the distal histidine has been demonstrated in oxy-CcP (19) and oxymyoglobin (26), the instability of the His-42 oxyperoxidase variants is best explained by the presence of a similar interaction in oxy-HRPC. However, other residues in the protein must also contribute to the stabilization of oxy-HRPC because a more dramatic decrease in both oxygen affinity and oxyperoxidase stability was observed when Arg-38 was replaced. The high polarity of the distal pocket created by the invariant distal arginine has been proposed to be a major determinant of ligand binding rates for peroxidases (15). We have not only confirmed this in the present studies with oxygen but also recently for CO binding to HRPC* variants (22). Arg-38 is clearly both a steric and an electrostatic impediment to the binding of oxygen and CO to the ferrous iron in HRPC.
Mechanism of Oxyperoxidase DecayThe mechanism of
oxyperoxidase decay has been studied previously in the presence of
ferroperoxidase, which accelerates the formation of ferriperoxidase
(6). Oxyperoxidase is relatively stable when there is no
ferroperoxidase present with a half-time for decay to ferriperoxidase
of many minutes. The present experiments were performed under these
conditions. The mechanism of oxy-HRPC decay most likely involves the
dissociation of a ferric-superoxide complex to yield ferric-HRPC and
O2. Peisach et al. (31) originally proposed that
oxygen binding to the ferrous heme induces partial oxidation of the
iron with electron density migrating to the oxygen. Resonance Raman
studies have shown that oxygen bound to oxy-HRPC and
lignin-oxyperoxidase have substantial superoxide character (32, 33). A
hydrogen bond between the
-oxygen of oxyperoxidase and the
N
H of His-42 and an electrostatic interaction between the guanidinium group of Arg-38 and the bound oxygen would be expected
to stabilize such a charge distribution, thereby increasing the
affinity for oxygen (19). The increased rates of decay of the R38K,
R38L, and R38G variants (kdecay, Table III) must
mean that the decreased affinity for oxygen (KD) is
more than compensated for by an increase in the rate of dissociation of superoxide anion from the ferric enzyme.
In the absence of a high
resolution crystal structure of HRPC (34), the data presented above
allow some predictions on the structure of the oxyperoxidase form. (i)
The NH of the imidazole group of the distal His-42 most
likely hydrogen bonds (
3.0 Å) with the bound oxygen. Such an
interaction has been observed in the crystal structure of
oxy-CcP (19) and in oxymyoglobin (26). (ii) The guanidinium
side chain of Arg-38 is close enough to the
-oxygen of oxy-HRPC for
hydrogen bonding and/or an electrostatic interaction. This contrasts
with the situation in oxy-CcP, in which the guanidinium
group of Arg-48 is some distance away from the bound oxygen with an
intervening water molecule (water 648) (19). These structural
differences explain the different reactivities of the oxyperoxidase
forms of native CcP and HRPC. (iii) An interaction between
the
-amino group of Lys in the R38K variant and a group from the
heme or from the protein is the most likely explanation for the 20-fold
increase in dioxygen binding rates. The crystal structure of the
corresponding CcP mutant shows that the side chain of Lys-48
is positioned between His-52 and the heme iron with the
-amino group
of Lys-48 2.7 Å distant from the iron occupying the position of
water-595 in the wild type (13). The different positions and/or
interactions of Lys in HRPC and CcP Arg
Lys variants can
also explain their different reactivities with hydrogen peroxide (13,
35, 36). (iv) We expected a decrease in the second-order rate constant
for O2 association in the H42R mutant, but instead we
observed a 2-fold increase similar to the 6-fold increase for the H42L
variant. The presence of two arginines in the distal pocket should
have increased the polarity of the oxygen access channel and caused a
decrease in the rate of oxygen binding. Our data suggest that in the
H42R mutant, the new arginine (Arg-42) adopts a conformation with the
charged side chain pointing out of the distal pocket toward the
solvent, thereby creating a less hindered pathway for oxygen binding. A
similar structural adjustment was postulated to occur in
-chains of
hemoglobin Zurich, in which the distal histidine was
replaced by arginine. The guanidinium group swings out of the active
center and interacts with one of the heme propionates, thereby removing
a hydrogen bonding interaction with the sixth coordination position and
facilitating oxygen binding (37).
A number
of attempts have been made to detect and characterize the transient
hydroperoxide intermediate (Fe(III)-OOH) formed in the reaction of
ferric enzyme with hydrogen peroxide. Baeck and Van Wart (38, 39),
working in methanol media at 26 °C, detected an intermediate with
a hyperporphyrin optical spectrum that they called compound 0. A
structure of oxy-CcP at 2.2-Å resolution has been proposed
as a model for the transient ferric enzyme-peroxide complex (19). We
have observed a new intermediate with neither a hyperporphyrin nor a
oxyperoxidase spectra formed transiently during the reaction of the
HRPC-R38L variant with hydrogen peroxide at 10 °C (17). The spectrum
obtained using rapid-scan stopped-flow spectrophotometry was similar to
that of the ferric enzyme but with peaks at 397, 487, and 580 nm, a
shoulder at 530 nm, and an increased extinction coefficient in the
Soret region (17). We have now detected a similar spectrum to that
proposed for the Fe(III)-OOH complex in the reaction of ferrous
HRPC-R38S with oxygen (Fig. 6). This suggests a stabilization of the
ferric-superoxide complex through a new hydrogen bond between the
serine introduced at position 38 and the
-oxygen in oxy-HRPC. This
structure is analogous to the Fe(III)-OOH intermediate, and the
similarity of the absorption spectra is to be expected.
Oxyperoxidase is analogous in
many ways to the oxymyoglobins and oxyhemoglobins. They all contain low
spin protoheme with histidine and dioxygen as the fifth and sixth
ligands, respectively. However oxyperoxidases do not bind oxygen
reversibly and are less photolabile. Oxyperoxidases because of their
inherent instability are more difficult to study (33). Clearly the
structures of both the distal and proximal pockets of oxyperoxidase and
oxymyoglobins are optimal for their respective functions. The
differences between oxyperoxidases and oxymyoglobins have been
explained by the different character of the -bond to the proximal
histidine (33). A stronger
-bond to the proximal histidine in
oxyperoxidase may in part explains the greater degree of oxygen
activation in oxyperoxidase compared to that for oxyglobins. This
should weaken the O-O bond and activate toward reduction. The
different reactivities of these two groups of proteins have also been
ascribed to differences in the polarities of the distal heme pockets
(15, 19). Our new data on HRPC variants provide additional evidence
that these different polarities are essential for oxyperoxidases and
oxyglobins to perform their specific functions. The mutations we have
made at Arg-38 in HRPC cause the behavior of the oxyperoxidase forms to
more closely resemble those of oxyglobins. The association rate
constants for oxygen binding (kon) are 2 orders
of magnitude higher for the Arg-38 variants and closer to those
reported for myoglobin or hemoglobin (15). The dissociation rate
constants (koff) are also of the same order as
those of the oxyglobins (15). We have created peroxidase variants that
closely resemble globins with respect to their oxygen binding kinetic
parameters but with enhanced rates of decay to yield ferric enzyme and
superoxide. Clearly other features of the protein environment on the
proximal side of the heme group must contribute to the relative
stability of oxyglobins.
The reversible binding of oxygen to myoglobin and hemoglobin is central to the metabolism of vertebrates and some invertebrates, providing a reservoir of O2 that is available for respiration. In contrast, peroxidases are designed to react with peroxide and reducing substrates and they are involved in various biosynthetic pathways. The active center of peroxidases not only provides optimum reaction rates with hydrogen peroxide and their reducing substrates but also induces irreversible decay of oxyperoxidase to yield ferric enzyme, the active form of the enzyme in the peroxidic cycle. This is a physiologically important reaction in that, under aerobic conditions in vivo, it removes ferrous enzyme and/or compound III that have accumulated due to an excess of hydrogen peroxide or superoxide anion, or following peroxidase catalyzed indole-3-acetic acid oxidative decarboxylation (40, 41).