Does the oxidation of nitric oxide by oxyMyoglobin share an intermediate with the metMyoglobin-catalyzed isomerization of peroxynitrite?

Article abstract of DOI:10.1021/ic400697a

The reaction of nitric oxide with oxy-myoglobin (oxyMb) to form ferric myoglobin (metMb) and nitrate, and the metMb-catalyzed isomerization of peroxynitrite to nitrate, have long been assumed to proceed via the same iron-bound peroxynitrite intermediate (metMb(OONO)). More recent research showed that the metMb-catalyzed isomerization of peroxynitrite to nitrate produces detectable amounts of nitrogen dioxide and ferryl myoglobin (ferrylMb). This suggests a mechanism in which the peroxynitrite binds to the metMb, ferrylMb is transiently generated by dissociation of NO₂, and nitrate is formed when the NO₂ nitrogen attacks the ferrylMb oxo ligand. The presence of free NO₂ and ferrylMb products reveals that small amounts of NO₂ escape from myoglobin's interior before recombination can occur. Free NO₂ and ferrylMb should also be generated in the reaction of oxyMb with NO, if the common intermediate metMb(OONO) is formed. However, this report presents a series of time-resolved UV/vis spectroscopy experiments in which no ferrylMb was detected when oxyMb and NO reacted. The sensitivity of the methodology is such that as little as 10% of the ferrylMb predicted from the experiments with metMb and peroxynitrite should have been detectable. These results lead to the conclusion that the oxyMb + NO and metMb + ONOO^- reactions do not proceed via a common intermediate as previously thought. The conclusion has significant implications for researchers that propose a possible role of oxyMb in intracellular NO regulation, because it means that toxic NO₂ and ferrylMb are not generated during NO oxidation by this species.

Full text of DOI:10.1021/ic400697a

Article
pubs.acs.org/IC
Does the Oxidation of Nitric Oxide by oxyMyoglobin Share an
Intermediate with the metMyoglobin-Catalyzed Isomerization of
Peroxynitrite?
†
Karl J. Koebke, Daniel J. Pauly, Leonid Lerner, Xien Liu, and A. Andrew Pacheco*
Department of Chemistry and Biochemistry, University of Wisconsin−Milwaukee, Milwaukee, Wisconsin 53211, United States
*
S Supporting Information
ABSTRACT: The reaction of nitric oxide with oxy-myoglobin (oxyMb) to form
ferric myoglobin (metMb) and nitrate, and the metMb-catalyzed isomerization of
peroxynitrite to nitrate, have long been assumed to proceed via the same iron-
bound peroxynitrite intermediate (metMb(OONO)). More recent research
showed that the metMb-catalyzed isomerization of peroxynitrite to nitrate
produces detectable amounts of nitrogen dioxide and ferryl myoglobin
(
ferrylMb). This suggests a mechanism in which the peroxynitrite binds to the
metMb, ferrylMb is transiently generated by dissociation of NO , and nitrate is
2
formed when the NO nitrogen attacks the ferrylMb oxo ligand. The presence of
2
free NO and ferrylMb products reveals that small amounts of NO escape from
2
2
myoglobin’s interior before recombination can occur. Free NO and ferrylMb
2
should also be generated in the reaction of oxyMb with NO, if the common
intermediate metMb(OONO) is formed. However, this report presents a series
of time-resolved UV/vis spectroscopy experiments in which no ferrylMb was detected when oxyMb and NO reacted. The
sensitivity of the methodology is such that as little as 10% of the ferrylMb predicted from the experiments with metMb and−
peroxynitrite should have been detectable. These results lead to the conclusion that the oxyMb + NO and metMb + ONOO
reactions do not proceed via a common intermediate as previously thought. The conclusion has significant implications for
researchers that propose a possible role of oxyMb in intracellular NO regulation, because it means that toxic NO and ferrylMb
2
are not generated during NO oxidation by this species.
INTRODUCTION
behavior was later detected in the reaction of ferrylMb with
■
NO , and identified as the metMb(ONO ) adduct (Scheme
2
2
Myoglobin’s role as an oxygen storage protein in tissues is
14
1
,2
1). Very recently this assignment was confirmed using rapid
freeze-quench resonance Raman spectroscopic analysis. One
indirect piece of evidence for the intermediacy of metMb-
widely accepted, despite some studies demonstrating that
15
myoglobin-deficient mice have apparently normal muscle
3
function. Nevertheless, in recent years many researchers
(
OONO) is that metMb and related proteins can catalyze the
have suggested that the protein may also serve a variety of
4
−7
isomerization of peroxynitrite to nitrate, presumably through
the same intermediate, as shown in refs 16−19. Model studies
with synthetic porphyrins by Schopfer et al. provide additional
other physiological functions. One such function is as a NO
sink. In this case, oxy-myoglobin (oxyMb) reacts with NO to
release nitrate, leaving the myoglobin in an aquated ferric state
20
4
,5,8
indirect evidence.
known as aquomet-myoglobin (metMb) (Scheme 1).
Nitric
Though a peroxynitrite intermediate is generally assumed in
the reaction of oxyMb with NO, there has been less agreement
on the mechanism by which this intermediate rearranges to the
oxide plays many important physiological roles as a signaling
9
,10
molecule, all of which take advantage of its transient nature.
Thus, the reaction of Scheme 1 could serve to spatially limit the
range of a NO signal within tissues that contain high
concentrations of myglobin, such as cardiac and skeletal
muscles, and to keep the concentration of NO within the cell
nitrate adduct (metMb(ONO )), with some researchers
2
18,21
favoring a concerted rearrangement (Scheme 1, path 1),
and others an O−O bond homolysis to generate a ferryl
4
myoglobin (ferrylMb) intermediate and NO , followed by a
2
below toxic levels.
recombination in which the N of NO binds to the ferryl oxo
2
The first step in the reaction of oxyMb with NO is
commonly accepted to be the formation of a peroxynitrite-
bound intermediate (metMb(OONO)), as shown in Scheme
1
5,17,19,20
group (Scheme 1, path 2).
Over the past decade
though, a series of experiments that followed the reaction of
peroxynitrite with metMb has provided persuasive evidence
that the metMb(OONO) adduct rearranges according to the
1
1,12
1,
though it should be noted that such an intermediate has
never been directly detected. At alkaline pH an intermediate
8
was detected in UV/vis stopped-flow and rapid freeze-quench
1
3
EPR studies, that at first was believed to be the peroxynitrite
Received: March 20, 2013
species. However, an intermediate species with identical kinetic
©
XXXX American Chemical Society
A
dx.doi.org/10.1021/ic400697a | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Scheme 1
Scheme 2
water. Stock solutions of most reagents were made up in 50 mM
HEPES. In the case of ascorbate, a 1.25 M stock solution was made by
dissolving 5.5 g of ascorbic acid and 7.45 g of HEPES sodium salt in 50
mL of water and then adjusting the pH to 7.4 with NaOH or HCl, as
appropriate. Reaction mixtures were made by combining appropriate
volumes of the stock solutions. Routine UV/vis spectra were obtained
using CARY 50 spectrophotometers (Varian); an instrument installed
in the glovebox was available for air-sensitive measurements.
Photochemical fragmentation of species 1 was initiated with 10 ns,
5
00 nm laser pulses generated using an OPO tunable laser (Opotec
Rainbow Vis). An OLIS RSM-1000 spectrophotometer was used to
monitor the absorbance changes induced in the reaction mixture by
the laser pulse. The configuration of the laser and spectrophotometric
2
7,28
equipment was as described in general terms elsewhere,
except
that the laser pulse was transmitted via an optical fiber instead of by
using mirrors. Reactions with lifetimes of less than 2 ms were
monitored with the OLIS RSM-1000 in fixed wavelength mode. Data
arrays that tracked the changes in the complete spectra as a function of
time were then constructed from multiple absorbance vs time traces
collected for the same reaction at varying wavelengths. For reactions
with lifetimes higher than 2 ms, data were collected with the OLIS
RSM-1000 in rapid-scanning average 1 mode, which provides 62
scans/s (1 scan every 16 ms). After verifying that no significant
changes were occurring on the millisecond time scale, the raw spectra
were further averaged to obtain spectra spaced 160 ms apart. For all
experiments, the solutions were held in 1.5 mm × 1.5 mm fluorometer
submicro cuvettes (Starna). The laser pulse was focused as a spot of
path 2 homolysis model. The culmination came in a recent
paper where Su and Groves were able to directly detect the
putative ferrylMb intermediate. This and earlier papers also
19
reported the nitration of a specific tyrosine within Mb
(
Tyr103), presumably by NO generated from the peroxynitrite
2
fragmentation and lost through the cage escape pathway
(
17,19
Scheme 1, lower-right).
With the chemistry of the
metMb(OONO) species now fairly well established, this
paper revisits the question of whether this species in fact
represents a common intermediate when oxyMb reacts with
NO, and when metMb reacts with peroxynitrite.
∼
5 mm diameter which irradiated the entire cuvette window. The laser
pulse energies were measured before and after each day’s experiments
MATERIALS AND METHODS
All reagents were purchased from Sigma-Aldrich or Fisher Scientific
unless otherwise specified. The light-activated NO precursor [Mn-
■
using a Scientech AC2501 bolometer; typical pulse energy intensities
2
at the cuvette were ∼15−18 mJ/cm .
All data were analyzed using programs written within the
commercially available software packages Origin version 6.0 (Microcal
(
PaPy Q)NO]ClO4 (Scheme 2, species 1, where PaPy Q is the
2
2
pentadentate ligand N,N-bis(2-pyridylmethyl)-amine-N-ethyl-2-quino-
Software) or Mathcad 13 (PTC Software). The analysis strategies used
line-2-carboxamide) and the pH-jump activated NO precursor 1-(N,N-
22,28−31
22
in our laboratories have been previously described;
details
diethylamino)diazene-1-ium-1,2-diolate (DEANO) were prepared
23
24,25
specific to this work are provided as Supporting Information.
by the methods of Eroy-Reveles et al. and Drago and Paulik,
respectively. Horse-heart myoglobin as obtained from Sigma was in
the metMb form. This was reduced in a glovebox using an excess of
RESULTS
■
2
6
electrolytically generated methyl viologen monocation radical to
generate ferromyoglobin (Mb), which could in turn be converted to
oxyMb by exposure to ambient air. FerrylMb was generated by
exposing metMb to excess hydrogen peroxide, which was then
In this study, NO was generated photochemically by irradiating
the photoactive manganese nitrosyl complex 1 (Scheme 2).
23
When species 1 is irradiated with a 500 nm, 10 ns laser pulse, it
fragments in less than 1 μs to release free NO and an aqua-
substituted product of 1, as shown in Scheme 2. This is a very
convenient way of generating NO in an aerobic solution of
oxyMb, because under the reaction conditions free NO is in
contact with free oxygen for less than 0.5 ms before it is all
−
8
19
removed by adding ∼10 M catalase, as previously described.
Nitrosylated metMb (metMbNO) and Mb (MbNO) were prepared
by exposing metMb and Mb, respectively, to NO gas generated in situ
2
4,25
from DEANO.
Extinction coefficient spectra were obtained for all
the above-mentioned forms of myoglobin.
All experiments described herein were carried out in solutions
buffered at pH 7.4 using HEPES and made using 18 MΩ purified
8
consumed in the reaction with oxyMb. Thus, unwanted side-
reactions, particularly the formation of NO2 from direct
B
dx.doi.org/10.1021/ic400697a | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
interaction between NO and O , are easily avoided. For the
If one accepts the hypothesis that the same metMb(OONO)
intermediate is generated in the reaction of oxyMb with NO
and in the metMb-catalyzed isomerization of peroxynitrite
(Scheme 1), then it is possible to estimate the expected final
product distribution using only kinetic parameters available in
2
purposes of the present study, species 1 is superior to other
known NO photoprecursors for two reasons. First, the NO can
be generated in large quantities (as high as 150 μM in the
present study) using long-wavelength light (500 nm or longer);
other high-yield NO photoprecursors that were used in the past
8
,14,19,34−36
the literature.
The relevant reactions, together with
28
by this research group require UV laser pulses. Second, the
aquated photoproduct of 1 is not very reactive on the time
scales being investigated, as is verified in experiments described
below. Many other NO photoprecursors fragment to yield
the associated kinetic parameters and the source references, are
shown in Scheme 3. Su and Groves estimated that the rate for
a
Scheme 3
2
8
products that are either strong oxidants
reductants,
or strong
3
2,33
which can take part in side-reactions on the
time scales of interest, thus complicating their use in
experiments.
Experiments with No Ascorbate Added. Figure 1 shows
the difference spectrum obtained immediately after irradiating a
Figure 1. Difference spectrum obtained immediately after irradiating a
solution of 316 μM oxyMb and 1.66 mM 1 with a 500 nm laser pulse.
Solid blue trace: experimentally obtained data. Dashed red trace:
calculated by least-squares fit using the extinction coefficient spectra of
oxyMb, metMb, ferrylMb, and 1. Dashed green trace: same as red
trace but using the Mb extinction coefficient spectrum instead of the
ferrylMb one. The fitted experimental spectrum was the average of the
first 20 raw spectra collected every 16 ms using the Olis
spectrophotometer (absorbance remained essentially constant
throughout this time period, which allowed for signal averaging).
a
References for the rate constants are as follows: (a) refs 8 and 19; (b)
ref 35; (c) ref 36.; (d) ref 14; (e) ref 34.
in-cage recombination of NO and ferrylMb to form metMb-
2
(
ONO ) is ∼10× that of cage escape by NO (k /k ∼ 10 in
2
2
r
e
19
Scheme 1). For the reaction that gave rise to the Figure 1
results, in which ∼86 μM of oxyMb reacted with an equivalent
amount of photogenerated NO, this would mean that initially
about 8 μM ferrylMb, 8 μM free NO , and 78 μM metMb
2
solution of 316 μM oxyMb and 1.66 mM 1 with a 500 nm laser
pulse. The Figure 1 difference spectrum can be fitted
reasonably well using the independently known extinction
coefficients of metMb, oxyMb, 1, and ferrylMb (Figure 1, red
dashed trace). On the basis of this fit, the changes in
concentrations of the various species are calculated to be
should have been generated. This estimate of ferrylMb
generated is about twice that calculated in the red trace
analysis of Figure 1 but does not take into account the
subsequent fate of the NO , which can itself react with oxyMb,
2
1
4,34
ferrylMb, and NO as shown in Scheme 3.
Using the
literature values of the rate constants for the Scheme 3
reactions, it is straightforward to predict the expected product
distributions as a function of time (see Supporting
Information). Figure 2 shows the calculated time evolution of
all the species in Scheme 3, assuming initial concentrations
ΔC
= 81 μM, ΔC
= ΔC = −86 μM, and ΔC
=
metMb
oxyMb
1
ferrylMb
5
μM. Note however that this method is quite insensitive, and a
similar fit can be obtained using the extinction coefficients of
metMb, oxyMb, 1, and Mb (Figure 1, green dashed trace). In
this case, the calculated concentrations of the species are
[MbO ] = 316 μM and [NO] = 86 μM, where the NO
2
ΔC
= −ΔC = 82 μM, ΔC
= −86 μM, and ΔC = 4
concentration is taken as equal to the amount of 1 photolyzed,
as calculated from the Figure 1 fit. Notice that the Figure 2
calculations predict that, under the given experimental
metMb
1
oxyMb
Mb
μM. Neither fit is improved by adding in contributions from the
extinction coefficient spectrum of MbNO. The presence of
either ferrylMb or Mb is mechanistically plausible: if a
peroxynitrite intermediate is formed from the reaction of
conditions, the reactions of NO with the various Mb species
2
will themselves result in a net increase in the ferrylMb
concentration, so that after 10 ms the ferrylMb will reach
∼16 μM. Thus, the 5 μM ferrylMb calculated in the red trace
analysis of Figure 1 is less than a third of that predicted for the
reaction of oxyMb with NO, if one assumes that the reaction
19
oxyMb with NO, then ferrylMb is an expected product. On
the other hand, the rapid reaction of oxyMb with NO also
consumes O , thus perturbing the equilibrium between oxyMb,
2
free Mb, and dissolved O2.
C
dx.doi.org/10.1021/ic400697a | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
ΔC
= 4 μM or ΔCmetMb = −ΔC = 97 μM, ΔCoxyMb=
ferrylMb
1
−
100 μM, and ΔC = 3 μM (see Supporting Information).
Mb
The experiment was then followed for an additional 45 after the
initial laser pulse, and double-difference spectra were generated
by subtracting the first difference spectrum from every
subsequent one. Figure 3a shows a number of these double-
Figure 2. Time evolution of reactant and product concentrations that
is expected following the reaction of 85.6 μM NO with 316 μM oxyMb
(
initial concentrations in the Figure 1 data), if one assumes a
metMb(OONO) intermediate in common with the metMb +
−
ONOO reaction. Concentrations were calculated using numerical
integration of the corresponding rate equations, and the known rate
constants given in Scheme 3, as described in the Supporting
Information. Panel a shows major species and panel b, minor ones.
goes via the same intermediate as the metMb-catalyzed
peroxynitrite isomerization.
Figure 3. Spectral changes that follow irradiation of a solution initially
containing 272 μM oxyMb, 1.72 mM 1, and 200 mM ascorbate with a
500 nm laser pulse. (a) Double-difference spectra obtained by
subtracting the first difference spectrum (equivalent to the Figure 1
spectrum) from all subsequent difference spectra that followed the
laser pulse. The representative spectra shown here were collected 0, 2,
19
Experiments in the Presence of Ascorbate. The result
presented above places an upper bound on the amount of
ferrylMb produced when NO is photogenerated from 1 in the
presence of a large excess of oxyMb but leaves open the
question of whether any ferrylMb is in fact produced. This is
because the Figure 1 spectrum is equally well fitted if one
assumes that Mb is a minority species instead of ferrylMb.
Furthermore, the minor features in the residual trace of Figure
4
, 6, 8, 10, 20, 30, and 45 s after the initial spectrum. Theoretical traces
(dashed red lines) were calculated as described in the Supporting
Information. Arrows indicate regions where spectral changes are
monotonic, while stars indicate regions in which ΔA first increases and
then decreases. (b) ΔA vs t slices at selected wavelengths,
corresponding to the data shown in a. The short orange trace shows
the absorbance change at 596 nm that would accompany the reduction
of 4 μM ferrylMb to metMb by 0.2 M ascorbate.
1
allow for the possibility that minority species other than Mb
or ferrylMb are also contributing to the Figure 1 spectrum, and
one cannot assess the significance of such species from analysis
of single spectral traces such as Figure 1.
In order to more precisely quantify the amount of ferrylMb
produced in the reaction of photogenerated NO with excess
oxyMb, the Figure 1 experiment was repeated in the presence
of varying amounts of ascorbate, with continued monitoring of
the reaction solution over tens of seconds. FerrylMb reacts with
ascorbate in a second-order process governed by a rate constant
difference spectra, collected at selected times after the laser
pulse and refined using singular value decomposition (see
38,39
Supporting Information).
Figure 3b shows corresponding
ΔA vs time traces at selected wavelengths. The complete set of
λ
double-difference spectra, collected at varying times and
wavelengths, can be fitted empirically using a model that builds
the spectra from the known extinction coefficients of Mb,
metMb, oxyMb,and species 1 and accounts for the time
dependence of ΔA at each wavelength using eq 1. The detailed
fitting procedure is described in the Supporting Information.
Equation 1 shows that at any given wavelength two
components contribute to the total absorbance change, one
of 2.7 ± 0.8 M− s , which should generate a readily
detectable, ascorbate concentration-dependent kinetic finger-
print with a many-second lifetime. When a solution containing
1
−1 37
2
72 μM oxyMb, 1.66 mM 1, and 200 mM ascorbate was
irradiated with a 500 nm laser pulse, an analysis analogous to
that of Figure 1 above showed initial product distributions of
either ΔCmetMb = 96 μM, ΔCoxyMb = ΔC = −100 μM, and
1
D
dx.doi.org/10.1021/ic400697a | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
varying linearly and the second exponentially with time. The
analysis analogous to that which produced the red trace of
Figure 1 (Supporting Information), it should then have reacted
with ascorbate to generate metMb, giving rise to a detectable
spectral signature with a maximum amplitude at 596 nm. Figure
parameter m represents the rate of change
λ
ΔA = m t + ΔAinf (1 − e−kt)
λ
λ
(1)
λ
3
b compares the absorbance change expected from the
of the linear component (in ΔAU/s), while ΔAinf and k
λ
reduction of ferrylMb to metMb with that actually observed
at 596 nm. Though the predicted change is small, it should have
been clearly resolvable given the low noise of the data obtained
on this time scale (a conservative estimate suggests that 1−2
μM ferrylMb should have been detectable by the applied
method; see Supporting Information for a statistical analysis).
Furthermore, the theoretical analysis of Scheme 3 and Figure 2
predicts a spectral change that is ∼3−4 times larger. On the
basis of the Figure 3 analysis, one can therefore conclude that
ferrylMb is not present in detectable quantities (>1−2 μM) by
the time the first postlaser flash spectra are collected, 16−160
ms after the laser pulse. Other minor species, such as Mb (e.g.,
green trace, Figure 1), may be present at this point.
represent the final amplitude and rate constant, respectively, for
the exponential component. The calculated spectra and time
traces are overlaid on the experimental data of Figure 3a and b.
From the fitting procedure, the changes in concentration with
time of Mb, metMb, oxyMb, and NO (indirectly obtained from
the concentration change of 1) were calculated; these are
plotted in Figure 4. In addition, the value of the rate constant
governing the exponential part of eq 1 was calculated to be k =
−1
0
.133 ± 0.005 s .
The results presented in Figures 3 and 4 can be qualitatively
explained on the basis of NO formed by the photolysis of 1 in
the spectrophotometer probe beam and of the known reactivity
of ascorbate with oxyMb and metMb that is outlined in Scheme
1
9,37,40
4
.
The linear increase in NO concentration shown in
a
Scheme 4
aHere, AH− is ascorbate, and A−• is the ascorbyl radical. Two
equivalents of H O are generated in step 3, one of which binds to
2
Fe(III) in the metMb.
Figure 4 (ΔC /Δt = 4.7 × 10− Ms ) arises from the
7
−1
NO
photolysis of 1 by the spectrophotometer probe beam, as can
be verified by submitting solutions of pure 1 to the same
treatment (Supporting Information). The NO so generated will
immediately react with oxyMb to produce metMb, thus
contributing to the concentration changes of these two species.
Step 3 of Scheme 4 is a second process that results in a net
40
oxidation of oxyMb to metMb. In this process, 1 equiv of
ascorbate first reacts with oxyMb to generate metMb and
4
0
H O . The metMb and H O then react to give the ferrylMb
2
2
2
2
+
•
40,41
cation radical (ferrylMb , formally Mb with Fe(V)),
which
is rapidly reduced by 2 equiv of ascorbate back to metMb. The
net process of step 3 in Scheme 4 results in the four-electron
reduction of O to give 2 equiv of H O (one of which binds to
2
2
Figure 4. (a) Changes in concentration of the myoglobin species that
gave rise to the spectral changes of Figure 2, calculated as described in
the Supporting Information. Also shown is the increase in [NO] with
time that was effected by photolysis of 1 in the spectrophotometer
probe beam. This was calculated from the absorbance change that
accompanied the denitrosylation of 1. (b) Similar to a, this plot shows
the total myoglobin reduced from the Fe(III) to the Fe(II) state as a
function of time.
the Fe of oxidized myoglobin), with three of the necessary
electrons coming from ascorbate and 1 equiv from the
myoglobin Fe(II). The 3 equiv of ascorbyl radicals generated
3
7,40
in the process subsequently disproportionate.
Counter-
acting the net oxidation of oxyMb to metMb by NO and by
step 3 of Scheme 4 is the reduction of metMb by ascorbate
(
step 1, Scheme 4). From experiments with metMb and
ascorbate under anaerobic conditions, the second-order rate
Given the 2.7 M−
1
s
−1
rate constant for the reaction of
constant for this reaction is calculated to be 0.053 ± 0.002 M
−1
37
−1
ferrylMb with ascorbate, any ferrylMb present immediately
s
at pH 7.4 (Supporting Information). This step initially
after the laser pulse should have subsequently been reduced by
generates Mb, but in the presence of O2 the Mb is in
equilibrium with oxyMb (step 2, Scheme 4).
Figure 4b shows that over the time period monitored there
was a net reduction of Fe(III) (metMb) to Fe(II) (oxyMb and
Mb) but that the rate of reduction slowed over time.
Immediately after the laser pulse, the reaction mixture contains
the 200 mM ascorbate in a pseudo-first-order process with k
obs
−1
=
0.54 s , which translates to a half-life of 1.3 s. The Figure 3
kinetic data show no evidence for such a kinetic process. If 4
μM of ferrylMb had in fact been present immediately after
irradiating the reaction solution, as suggested by spectral
E
dx.doi.org/10.1021/ic400697a | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
roughly 97 μM metMb, 3 μM Mb, and 172 μM oxyMb (as
determined by analysis analogous to that which produced the
green trace of Figure 1; Supporting Information). Presumably,
as metMb is subsequently consumed and reduced myoglobin is
generated, the rates of the net oxidizing and net reducing
processes move toward a dynamic equilibrium, beyond which
no further changes in concentration would be observed. The
distribution of reduced species also changes with time as seen
in Figure 4a. Step 3 of Scheme 4 consumes oxygen, as does the
reaction of oxyMb with NO, and in the small cuvette employed
for the experiments the oxygen is not rapidly replaced by
diffusion from the surroundings. Thus, over time, the
concentration of oxyMb decreases and that of Mb increases.
Experiments with solutions containing only oxyMb and
ascorbate in the same type of cuvette showed the same general
reactivity (Supporting Information), though the exact reaction
conditions could not be reproduced in the absence of the laser
pulse that rapidly consumed an initial quantity of O2.
Figure 5. Spectral changes observed at 409 nm after mixing 4.2 μM
ferrylMb with 238 μM 1. Blue circles: experimental data. Red curve:
theoretical fit to an exponential function.
In addition to the experiment described above (Figures 3 and
4
1
), similar experiments were also performed in the presence of
00 μM and 300 μM ascorbate (see Supporting Information).
between ferrylMb and NO photogenerated from 1 in the
+
spectrometer probe beam. Alternatively, the Mn(PaPy Q)H O
2
2
In all of these experiments, the time-resolved spectra lacked the
kinetic fingerprint that would be expected if ferrylMb had been
generated in the reaction of oxyMb with NO but were
otherwise consistent with the expected reactivity of the reaction
mixtures, as outlined in Scheme 4. Finally, to ensure that the
reaction of ferrylMb with ascorbate is not altered under the
specific reaction conditions employed in this study, authentic
ferrylMb was reacted with ascorbate in the presence and
absence of 1, oxyMb, or Mb. In all cases, the same rate constant
was obtained, which in turn agreed with the literature values
that is cogenerated from the photolysis of 1 could also
contribute to ferrylMb reduction. These possibilities were
addressed in kinetic and stoichiometric studies. In the kinetic
experiments, species 1 was photolyzed in the presence of
varying amounts of ferrylMb, and the subsequent reactions
were monitored on the microsecond time scale (Figure 6).
These experiments were carried out under pseudo-first-order
conditions, such that the concentration of ferrylMb greatly
exceeded the amount of NO photogenerated in every
1
9,37
(see Supporting Information).
Stability of ferrylMb in the Presence of Species 1,
with and without Laser Irradiation. The experiments
presented above demonstrate conclusively that no ferrylMb is
detectable 16 ms after irradiating solutions of oxyMb and 1
with a 500 nm laser pulse. However, they do not preclude the
possibility that detectable amounts of ferrylMb were generated
on the microsecond time scale but consumed prior to 16 ms in
reactions specific to the conditions employed. In order to test
for this possibility, several experiments were performed with
authentic ferrylMb and 1.
Figure 5 shows the spectral changes observed at 409 nm after
mixing 4.2 μM ferrylMb with 238 μM 1. The absorbance at 409
increases exponentially, indicating that the ferrylMb is being
−
3
reduced to metMb with an apparent rate constant of 1.7 × 10
−1
s . When the experiment was repeated in the presence of half
the concentration of 1, the apparent rate constant also
decreased by half, showing that the reduction is first order in
[1] and second order overall. The value of the true second-
order rate constant for reduction of ferrylMb in the presence of
−1 −1
1
1
1
is therefore ∼7 M s . In the experiments with oxyMb and
described above, the concentration of 1 was typically between
.5 and 2 mM; therefore one expects any ferrylMb generated
under those conditions to be reduced with a half-life of ∼50−
6
5 s. This reaction is too slow to eliminate ferrylMb before it
could be detected in the experiments of the previous sections;
in particular it would not compete with the reaction of ferrylMb
with 200 mM ascorbate, which is predicted to occur with a half-
life of 1.3 s.
Figure 6. (a) Spectral changes observed at 409 nm after irradiating a
solution containing 50 μM ferrylMb and 40 μM 1 with a 500 nm laser
pulse. Blue trace: experimental data. Red trace: data fitted to a single
exponential. (b) Dependence of the observed rate constant on
As shown in Scheme 3, NO reacts rapidly with ferrylMb.
Thus, the spectral changes seen in solutions containing
ferrylMb and 1 (Figure 5) could simply be due to the reaction
[
ferrylMb] when the experiment of part a was repeated with varying
initial ferrylMb concentrations. In all cases, [NO]generated ≪
[ferrylMb] .
0
F
dx.doi.org/10.1021/ic400697a | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
experiment. The change in absorbance at 409 nm with time was
exponential at all ferrylMb concentrations (Figure 6a), and
plots of kobs vs [ferrylMb] were linear (Figure 6b), showing that
the reaction was first-order in [ferrylMb] and second-order
overall. The true second-order rate constant for the reaction
outcome of the simulations, given the reaction conditions being
modeled. This leaves k , k , k , k , and k as parameters that
4
5
8
−8
9
could significantly affect the predicted time evolution of
[ferrylMb].
The parameters k , k , k , k , and k all govern the fate of the
4
5
8
−8
9
7
−1 −1
was calculated to be (1.4 ± 0.1) × 10 M
s
from the slope
NO that putatively escapes the protein cage. Any increase in
2
of the Figure 6b line, a value that is close to that previously
reported in the literature for the reaction between NO and
the ratio k /k will lead to greater net recombination of
5
4
ferrylMb with NO , thus decreasing the final amounts of both
2
35
ferrylMb (Scheme 3).
The agreement between the rate constant calculated from
species. However, in practice it would take a huge error in the
literature derived value of this ratio (0.267 from Scheme 3) to
have any significant effect on ferrylMb survival under the
reaction conditions being modeled, where at all times the
oxyMb concentration far exceeded any ferrylMb that might be
35
Figure 6b and that reported earlier by Herold and Rehman
demonstrates that NO is the fastest reducing agent present in
photolyzed mixtures of ferrylMb and 1. If it were not, the rate
constant calculated from Figure 6b would have been higher
than Herold and Rehman’s. The close agreement does not,
generated. Thus, for example, if the value of k is decreased to 5
4
6
−1 −1
8
−1 −1
× 10 M
s
and that for k
is increased to 10 M
s
(both
5
+
however, exclude the possibility that Mn(PaPy Q)H O also
these numbers are well outside the reported uncertainties for
the respective parameters), a simulation still predicts the
2
2
reduces ferrylMb at rates comparable to or lower than NO. To
address this possible alternative, the amount of ferrylMb
reduced and the amount of NO generated by photolysis of a
given amount of 1 were determined independently, in a set of
experiments that are described in Supporting Information. In
these experiments, only one equivalent of ferrylMb was reduced
per equivalent of 1 photolyzed, which shows that NO is the
presence of ∼10 μM ferrylMb 10 ms after the laser pulse. A k
5
8
−1 −1
value of 10 M
s
for the ferrylMb reaction with NO
would
2
make it by far the largest known rate constant for the reaction
1
4
6
−1
of NO
s
with a biomolecule, while the k
would be at the low end of values reported for the
interaction of small gas molecules with either oxyMb or
value of 5 × 10 M
2
4
−1
8
,14,42,43
sole reducing agent generated in the photolysis. If Mn-
Mb.
And yet, much greater changes in the input rate
+
(
PaPy Q)H O were also capable of reducing ferrylMb on
constants would be needed to decrease the ferrylMb
concentration below 1−2 μM (the estimated detection limit
for these experiments).
2
2
the subsecond time scale, two equivalents of ferrylMb would
have been reduced per equivalent of 1 photolyzed in these
experiments. Therefore, in the experiments in which solutions
of oxyMb and 1 were irradiated to generate NO, free ferrylMb
generated by homolytic fragmentation of a metMb(OONO)
intermediate should have still been present 16 ms after
photolysis, and thus detectable.
The processes governed by k
at which NO is consumed independently of ferrylMb. The
value of k for the reaction of NO with NO is at the diffusion
limit and so is unlikely to be underestimated. However, if k−8 is
, k−8, and k determine the rate
8
9
2
8
2
overestimated and/or k is underestimated in Scheme 3, then
9
all else being equal the Figure 2 analysis would be
overestimating the amount of ferrylMb that is eventually
Evaluation of the Scheme 3 Model. Given that ferrylMb
was not generated in detectable quantities in the present
experiments, contrary to the predictions of earlier reports as
outlined in Scheme 3, it is now necessary to assess the
sensitivity of the Scheme 3 conclusions to possible errors in the
published kinetic parameters. One question that immediately
arises is whether Su and Groves could have overestimated the
amount of cage escape (Scheme 1) that occurred in their
experiments; the values of k and k12 in Scheme 3 are based
generated by reaction of NO with oxyMb. Note however that
2
even in the extreme case whereby none of the NO resulted in
2
additional ferrylMb formation, one would still see the 10%
ferrylMb generated by the initial NO cage escape, if this were
2
occurring.
To summarize, even if one assumes much larger uncertainties
in the kinetic parameters of Scheme 3 than are provided with
the literature values, ferrylMb is predicted to accumulate to
concentrations 5−10 times above the detection limit in the
experiments presented herein, if the reactions of oxyMb with
NO and metMb with peroxynitrite share a common metMb-
1
1
19
directly on this estimate. When the numerical simulation that
they used to obtain the k /k estimate of 10 was repeated, this
r
e
estimate was found to be very insensitive to even large
variations in the inputted initial concentrations of metMb and
peroxynitrite. Therefore, Su and Groves’ estimate for initial
ferrylMb generation via cage escape cannot be very far off.
Next to be examined were the possible effects that
uncertainties in the Scheme 3 rate constants might have on
the predicted time evolution of the ferrylMb concentration. Of
all the parameters in Scheme 3, the one with the greatest
(
OONO) intermediate, as proposed in Scheme 1.
DISCUSSION
■
Investigations by the Groves group and others over the past
decade have shown fairly convincingly that, during metMb-
catalyzed isomerization of peroxynitrite, some nitration of
reported uncertainty is k , for which the literature value is given
Tyr103 from the metMb takes place, and ferrylMb builds up as
6
4
−1 −1
7
−1 −1 14
12,17,19
as a range (10 M
that yielded Figure 2, the upper bound was arbitrarily chosen
Supporting Information), but a smaller value of k is unlikely
s
to 10 M s ). For the calculations
a transient but detectable intermediate.
These results are
consistent with the mechanism of Scheme 1 path 2, whereby
peroxynitrite first binds the Fe heme center to produce a
metMb(OONO) intermediate, which then rapidly cleaves
homolytically to form ferrylMb and caged NO . The NO
(
6
+
•
to change the conclusions; if ferrylMb reacts more slowly
with oxyMb, then it is almost certainly still converted to
ferrylMb by other competing processes, such as reduction by
ascorbate.
2
2
can then either recombine with the ferrylMb to release nitrate,
or escape the cage; the latter path gives rise to the ferrylMb that
is transiently detectable, and to the observed Tyr103 nitration.
In an early report, the extent of cage escape was estimated to be
The fact that (k + k ) and k are within 20% of the
11
12
2
8
,35
reported values was independently verified herein (Support-
ing Information and Figure 6). Of the remaining parameters, k3
17
as high as 20%, but this was later revised downward to
19
and k prove to be too small to have any real impact on the
∼10%.
7
G
dx.doi.org/10.1021/ic400697a | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
It has been widely assumed until now that the reaction
between oxyMb and NO proceeds through the same
metMb(OONO) intermediate as the reaction between
metMb and peroxynitrite (Scheme 1); however, if this
assumption is correct, then under appropriate conditions one
should be able to detect the ferrylMb side product due to 10%
Scheme 5
cage escape of NO , as Groves’ group did when reacting
2
19
peroxynitrite with metMb. The fact that this intermediate was
not detected in any of the experiments presented herein calls
the assumption into question. In this regard, the key
experiments are those performed in the presence of ascorbate
Kurtikyan and Ford recently reported careful studies in which
synthetic complexes with the formula Fe(Por)(NH )(O ),
3
2
(Figures 3, 4 and Supporting Information). Single spectrum
sublimed in vacuo onto liquid nitrogen-cooled CaF or KBr
2
44
analyses such as that given in Figure 1 are of limited use
because all of the components contributing to such a spectrum
may not be known. However, the kinetic analyses of the
ascorbate experiments are substantially more sensitive to
ferrylMb concentration and relatively insensitive to the effects
of unknown parameters. No matter what species are
contributing to the overall spectra, any spectral changes due
to conversion of ferrylMb to metMb should have been readily
detectable if ferrylMb concentrations greater than 1−2 μM had
been produced from the reaction between oxyMb and
photogenerated NO. This is clear from inspection of the
short orange trace in Figure 3b, which shows the spectral
change at one representative wavelength that would be
associated with the conversion of 4 μM ferrylMb to metMb
in the presence of 200 mM ascorbate. However, 4 μM is only
surfaces, were exposed to NO gas. Even in these low-
temperature studies, the first detectable intermediate as the
surface warmed up was invariably the nitrate adduct. From
these studies, Kurtikyan and Ford concluded that if
peroxynitrite intermediates actually formed, subsequent barriers
for their isomerization to nitrate species were at most ∼7 kcal/
44
mol. By contrast, Lee et al. had shown earlier that some
water-soluble Fe(III)porphyrins react with peroxynitrite to
generate very long-lived oxoFe(IV) species and NO , that only
2
45
slowly recombine. The Kurtikyan and Lee experimental
conditions are very different (hydrophobic porphyrins adsorbed
on surfaces vs water-soluble porphyrins in aqueous solution),
and this alone could explain the different product distributions.
Alternatively it is also possible that the Kurtikyan and Ford
reactions proceeded via a mechanism akin to Scheme 5, while
those of Lee et al. followed path 2 of Scheme 1. It should be
noted that Kurtikyan and co-workers did, very recently, detect
the cobalt equivalent of a peroxynitrite-bound species and later
the largest amount of ferrylMb consistent with analysis of the t0
spectrum for the data set (the analogue of the Figure 1
spectrum; see Supporting Information). If the reaction of
oxyMb with NO had produced 10% ferrylMb, as predicted
from the experiments with metMb and peroxynitrite, the
amplitude of the Figure 3 short orange trace should have been
at least twice as large, and according to the extended analysis
based on Scheme 3 it could have been as much as 4 times
greater. Furthermore, the short orange trace of Figure 3 is only
for one wavelength. The analysis presented herein was a global
one, and the expected kinetic signature for ferrylMb reduction
to metMb was absent from the entire monitored spectral region
NO , when they reacted synthetic oxy-coboglobin models with
2
NO, using the surface deposition methodology summarized
46
above. Also, Schopfer and co-workers have reported indirect
evidence that synthetic Fe(Por)(NH )(O ) complexes can
3
2
react with NO via a pathway akin to that shown in Scheme 1,
path 2. When tetrahydrofuran (THF) solutions containing
tyrosine mimic 2,4-di-tert-butylphenol, and oxyMb mimic
III
+•
(THF)(F )Fe (O ) [where F8 = tetrakis(2,6-difluoro-
8
2
phenyl)porphyrinate(2-)], were exposed to NO, the nitrated
product 2,4-di-tert-butyl-6-nitrophenol was recovered, as would
(
480−620 nm). The same was true in experiments with 100
mM and 300 mM ascorbate (see Supporting Information).
The experiments with ferrylMb and 1 (Figures 5 and 6 and
Supporting Information) show that neither 1 nor its photo-
be expected if free NO was being formed from fragmentation
2
20
of a peroxynitrite intermediate. Schopfer et al. did not directly
observe such an intermediate, but their experiments show that
the question is far from settled.
+
product Mn(PaPy Q)H O will react with ferrylMb in less than
2
2
1
s, so if this species had been generated from oxyMb and NO
Scheme 6 suggests an alternative to Scheme 5 that would also
allow the reaction mechanism for oxyMb + NO to deviate from
that of metMb + peroxynitrite but would not preclude a
metMb(OONO) transient species in either case. In this
scenario, the metMb(OONO) moiety itself is formed in both
it should have persisted long enough to react with ascorbate on
the time scale of seconds. Furthermore, the experiments with
ferrylMb and ascorbate show that the kinetics of this crucial
reaction were unaffected by the specifics of the reaction
conditions (see Supporting Information). Together, the
experiments with ferrylMb confirm that this transient species
should have been detectable if it had been generated in the
quantities predicted from experiments with metMb and
reactions and fragments homolytically to release NO into the
2
immediate surroundings of the heme, but the conformation of
the protein surrounding the heme center is assumed to be
significantly different in the two cases. This is depicted in
Scheme 6 by representing the Fe(II)-derived protein
conformation by a rectangle surrounding the heme moiety
and that for Fe(III) by an ellipse. If the conformation of the
protein derived from metMb were “leakier” than that derived
from oxyMb (k ≫ k in Scheme 6), this would explain why
1
7,19
peroxynitrite,
via path 2 of Scheme 1.
Scheme 5 proposes one reaction mechanism for oxyMb +
NO that would yield a product distribution different from that
obtained in the reaction between metMb and peroxynitrite. In
this scenario, NO directly inserts into the oxyMb O−O bond,
thus completely bypassing the metMb(OONO) intermediate.
Some model studies with synthetic iron porphyrin complexes
support the hypothesis that different mechanisms may operate
in reactions of ferric porphyrins with peroxynitrite and in
reactions of oxyferrous porphyrins with NO. For example,
e2
e1
ferrylMb and NO are produced in the former case and not the
2
latter. Using the most conservative estimate of the detection
limit in the ascorbate experiments (∼2 μM), this mechanism
would predict that k in Scheme 6 is at least 10 times greater
e2
than ke1.
H
dx.doi.org/10.1021/ic400697a | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Scheme 6
AUTHOR INFORMATION
■
*
Corresponding Author
Phone: (414)-229-4413. E-mail: apacheco@uwm.edu.
Present Address
Department of Material Science and Engineering, Penn State
University, 322 Steidle Building, University Park, PA 16802.
†
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by grants UWM RGI 101X157,
UWM RGI 101X225, and NSF 0843459. D.J.P. gratefully
acknowledges support from the University of Wisconsin’s
Office of Undergraduate Research. The authors gratefully
acknowledge Prof. Graham Moran for the use of his stopped-
flow apparatus.
REFERENCES
■
(
1) Wittenberg, J. B.; Wittenberg, B. A. Annu. Rev. Biophys. Biophys.
Chem. 1990, 19, 217.
(2) Takahashi, E.; Endoh, H.; Doi, K. Biophys. J. 2000, 78, 3252.
(3) Garry, D. J.; Ordway, G. A.; Lorenz, J. N.; Radford, N. B.; Chin,
E. R.; Grange, R. W.; Bassel-Duby, R.; Williams, R. S. Nature 1998,
3
95, 905.
(4) Brunori, M. Trends Biochem. Sci. 2001, 26, 209.
(
5) Frauenfelder, H.; McMmahon, B. H.; Austin, R. H.; Chu, K.;
The fact that the reaction of oxyMb with NO gives different
product distributions than that of metMb with peroxynitrite has
important implications for researchers investigating the possible
physiological role of myoglobin and related molecules as NO
sinks. Path 2 of Scheme 1, which has now been identified as the
likely mechanism for metMb-catalyzed isomerization of
Groves, J. T. Proc. Natl. Acad. Sci. U. S. A. 2001, 98, 2370.
6) Hendgen-Cotta, U. B.; Merx, M. W.; Shiva, S.; Schmidz, J.;
(
Becher, S.; Klare, J. P.; Steinhoff, H.-J.; Goedecke, A.; Schrader, J.;
Gladwin, M. T.; Kelm, M.; Rassaf, T. Proc. Natl. Acad. Sci. U. S. A.
2
008, 105, 10256.
(
7) Cossins, A.; Berenbrink, M. Nature 2008, 454, 416.
1
9
peroxynitrite, allows a significant amount of free NO to
(8) Herold, S.; Exner, M.; Nauser, T. Biochemistry 2001, 40, 3385.
(9) Moncada, S.; Palmer, R. M. J.; Higgs, E. A. Pharmacol. Rev. 1991,
43, 109.
2
escape the protein and concomitantly results in formation of
ferrylMb. If NO dioxygenation by oxyMb were to follow exactly
the same path, then removal of the reactive NO would be
(
10) Nitric Oxide, Principles and Actions; Lancaster, J., Ed.; Academic
Press: San Diego, CA, 1996.
11) Ford, P. C.; Lorkovic, I. M. Chem. Rev. 2002, 102, 993.
(12) Su, J.; Groves, J. T. Inorg. Chem. 2010, 49, 6317.
13) Olson, J. S.; Foley, E. W.; Rogge, C.; Tsai, A. L.; Doyle, M. P.;
Lemon, D. D. Free Radic. Biol. Med. 2004, 36, 685.
14) Goldstein, S.; Merenyi, G.; Samuni, A. J. Am. Chem. Soc. 2004,
26, 15694.
15) Yukl, E. T.; de Vries, S.; Moenne-Loccoz, P. J. Am. Chem. Soc.
2009, 131, 7234.
16) Gardner, P. R.; Gardner, A. M.; Martin, L. A.; Salzman, A. L.
Proc. Natl. Acad. Sci. U. S. A. 1998, 95, 10378.
17) Bourassa, J. L.; Ives, E. P.; Marqueling, A. L.; Shimanovich, R.;
Groves, J. T. J. Am. Chem. Soc. 2001, 123, 5142.
accompanied by generation of NO and ferrylMb, which in
2
(
many ways are at least as reactive. In contrast, the results of this
investigation show that oxyMb converts NO to NO3− without
generating any detectable ferrylMb side product. This makes a
physiological role for myoglobin’s NO dioxygenase activity that
much more plausible.
(
(
1
(
ASSOCIATED CONTENT
Supporting Information
■
(
*
S
Determination of the rate constant for the reaction of metMb
with ascorbate. The reaction of oxyMb with ascorbate. The
reaction of ferrylMb with ascorbate. Laser photoinitiated
fragmentation of species 1 in the absence of protein. Expected
result of mixing 100 μM NO with 272 μM oxyMb, if the
reaction of oxyMb with NO shared an intermediate with the
metMb-catalyzed isomerization of peroxynitrite: a numerical
simulation of the product distribution’s evolution during the
first 3 ms. Laser photoinitiated reaction of NO with excess
oxyMb in the presence and absence of ascorbate: analysis of the
data. Monitoring the reaction of oxyMb with NO on the
microsecond time scale. Stability of ferrylMb in the presence of
species 1 after laser irradiation: stoichiometric analysis.
Statistical analysis of ferrylMb concentration in the reaction
of oxyMb with NO at low ascorbate concentrations and in
control experiments. This material is available free of charge via
the Internet at http://pubs.acs.org.
(
(18) Herold, S.; Shivashankar, K. Biochemistry 2003, 42, 14036.
(19) Su, J.; Groves, J. T. J. Am. Chem. Soc. 2009, 131, 12979.
(20) Schopfer, M. P.; Mondal, B.; Lee, D.-H.; Sarjeant, A. A. N.;
Karlin, K. D. J. Am. Chem. Soc. 2009, 131, 11304.
(21) Herold, S.; Shivashankar, K.; Mehl, M. Biochemistry 2002, 41,
13460.
(
22) Purwar, N.; McGarry, J. M.; Kostera, J.; Pacheco, A. A.; Schmidt,
M. Biochemistry 2011, 50, 4491.
23) Eroy-Reveles, A.; Leung, Y.; Beavers, C. M.; Olmstead, M. M.;
Mascharak, P. K. J. Am. Chem. Soc. 2008, 130, 4447.
(
(
(
24) Drago, R. S.; Paulik, F. E. J. Am. Chem. Soc. 1960, 82, 96.
25) Maragos, C. M.; Morley, D.; Wink, D. A.; Dunams, T. M.;
Saavedra, J. E.; Hoffman, A.; Bove, A. A.; Isaac, L.; Hrabie, J. A.;
Keefer, L. K. J. Med. Chem. 1991, 34, 3242.
(26) Atkinson, S. J.; Mowat, C. G.; Reid, G. A.; Chapman, S. K. FEBS
Lett. 2007, 581, 3805.
I
dx.doi.org/10.1021/ic400697a | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
(
27) Cabail, M. Z.; Lace, P. J.; Uselding, J.; Pacheco, A. A. J.
Photochem. Photobiol. A 2002, 152, 109.
28) Cabail, M. Z.; Moua, V.; Bae, E.; Meyer, A.; Pacheco, A. A. J.
Phys. Chem. A 2007, 111, 1207.
29) Kostera, J.; Youngblut, M. D.; Slosarczyk, J. M.; Pacheco, A. A. J.
Biol. Inorg. Chem. 2008, 13, 1073.
30) Kostera, J.; McGarry, J. M.; Pacheco, A. A. Biochemistry 2010,
9, 8546.
31) Youngblut, M.; Judd, E. T.; Srajer, V.; Sayyed, B.; Goelzer, T.;
Elliott, S. J.; Schmidt, M.; Pacheco, A. A. J. Biol. Inorg. Chem. 2012, 17,
47.
(
(
(
4
(
6
(
32) DeRosa, F.; Bu, X.; Ford, P. C. Inorg. Chem. 2005, 44, 4157.
(
33) Bourassa, J.; DeGraff, W.; Kudo, S.; Wink, D. A.; Mitchell, J. B.;
Ford, P. C. J. Am. Chem. Soc. 1997, 119, 2853.
(34) Goldstein, S.; Czapski, G. J. Am. Chem. Soc. 1995, 117, 12078.
(35) Herold, S.; Rehmann, F.-J. K. J. Biol. Inorg. Chem. 2001, 6, 543.
(36) Laverman, L. E.; Wanat, A.; Oszajca, J.; Stochel, G.; Ford, P. C.;
van Eldik, R. J. Am. Chem. Soc. 2001, 123, 285.
37) Kroger-Ohlsen, M.; Skibsted, L. H. J. Agric. Food Chem. 1997,
5, 668.
38) Press, W. H.; Teukolsky, S. A.; Vetterling, W. T.; Flannery, B. P.
(
4
(
Numerical Recipes: The Art of Scientific Computing, 3rd ed.; Cambridge
University Press: New York, 2007; p 65.
(
39) Henry, E. R.; Hofrichter, J. In Meth. Enzymol.; Brand, L.,
Johnson, M. L., Eds.; Academic Press: San Diego, CA, 1992; Vol. 210,
p 129.
(
40) Giulivi, C.; Cadenas, E. FEBS 1993, 332, 287.
(
41) Fenwick, C.; Marmor, S.; Govindaraju, K.; English, A. M.;
Wishart, J. F.; Sun, J. J. Am. Chem. Soc. 1994, 116, 3169.
(
42) Jameson, G. B.; Ibers, J. A. In Biological Inorganic Chemistry;
Bertini, I., Gray, H. B., Stiefel, E. I., Valentine, J. S., Eds.; University
Science Books: Sausalito, CA, 2007; p 363.
(
43) Hoshino, M.; Maeda, M.; Konishi, R.; Seki, H.; Ford, P. C. J.
Am. Chem. Soc. 1996, 118, 5702.
(
(
44) Kurtikyan, T. S.; Ford, P. C. Chem. Commun. 2010, 46, 8570.
45) Lee, J.; Hunt, J. A.; Groves, J. T. J. Am. Chem. Soc. 1998, 120,
7
493.
(
46) Kurtikyan, T. S.; Eksuzyan, S. R.; Hayrapetyan, V. A.;
Martirosyan, G. G.; Hovhannisyan, G. S.; Goodwin, J. A. J. Am.
Chem. Soc. 2012, 134, 13861.
J
dx.doi.org/10.1021/ic400697a | Inorg. Chem. XXXX, XXX, XXX−XXX