Combined experimental and theoretical study on the reactivity of Compounds I and II in horseradish peroxidase biomimetics

Article abstract of DOI:10.1002/chem.201402347

For the exploration of the intrinsic reactivity of two key active species in the catalytic cycle of horseradish peroxidase (HRP), Compound I (HRP-I) and Compound II (HRP-II), we generated in situ [Fe^IV=O(TMP⁺?)(2-MeIm)] + and [Fe^IV=O(TMP)(2-MeIm)]⁰ (TMP = 5,10,15,20-tetramesitylporphyrin; 2-MeIm = 2-methylimidazole) as biomimetics for HRP-I and HRP-II, respectively. Their catalytic activities in epoxidation, hydrogen abstraction, and heteroatom oxidation reactions were studied in acetonitrile at -15 °C by utilizing rapid-scan UV/Vis spectroscopy. Comparison of the secondorder rate constants measured for the direct reactions of the HRP-I and HRP-II mimics with the selected substrates clearly confirmed the outstanding oxidizing capability of the HRP-I mimic, which is significantly higher than that of HRP-II. The experimental study was supported by computational modeling (DFT calculations) of the oxidation mechanism of the selected substrates with the involvement of quartet and doublet HRP-I mimics (^2,4Cpd I) and the closed-shell triplet spin HRP-II model (³Cpd II) as oxidizing species. The significantly lower activation barriers calculated for the oxidation systems involving ^2,4Cpd I than those found for ³Cpd II are in line with the much higher oxidizing efficiency of the HRP-I mimic proven in the experimental part of the study. In addition, the DFT calculations show that all three reaction types catalyzed by HRP-I occur on the doublet spin surface in an effectively concerted manner, whereas these reactions may proceed in a stepwise mechanism with the HRP-II mimic as oxidant. However, the high desaturation or oxygen rebound barriers during C-H bond activation processes by the HRP-II mimic predict a sufficient lifetime for the substrate radical formed through hydrogen abstraction. Thus, the theoretical calculations suggest that the dissociation of the substrate radical may be a more favorable pathway than desaturation or oxygen rebound processes. Importantly, depending on the electronic nature of the oxidizing species, that is, ^2,4Cpd I or ³Cpd II, an interesting region-selective conversion phenomenon between sulfoxidation and H-atom abstraction was revealed in the course of the oxidation reaction of dimethylsulfide. The combined experimental and theoretical study on the elucidation of the intrinsic reactivity patterns of the HRP-I and HRP-II mimics provides a valuable tool for evaluating the particular role of the HRP active species in biological systems.

Full text of DOI:10.1002/chem.201402347

DOI: 10.1002/chem.201402347
Full Paper
&
Biomimetics
Combined Experimental and Theoretical Study on the Reactivity
of Compounds I and II in Horseradish Peroxidase Biomimetics
Li Ji,[a, b] Alicja Franke,[a] Małgorzata Brindell,[c] Maria Oszajca,[c] Achim Zahl,[a] and
Rudi van Eldik*[a, c]
Abstract: For the exploration of the intrinsic reactivity of
two key active species in the catalytic cycle of horseradish
peroxidase (HRP), Compound I (HRP-I) and Compound II
proven in the experimental part of the study. In addition,
the DFT calculations show that all three reaction types cata-
lyzed by HRP-I occur on the doublet spin surface in an effec-
tively concerted manner, whereas these reactions may pro-
ceed in a stepwise mechanism with the HRP-II mimic as oxi-
dant. However, the high desaturation or oxygen rebound
barriers during CꢀH bond activation processes by the HRP-II
mimic predict a sufficient lifetime for the substrate radical
formed through hydrogen abstraction. Thus, the theoretical
calculations suggest that the dissociation of the substrate
radical may be a more favorable pathway than desaturation
or oxygen rebound processes. Importantly, depending on
the electronic nature of the oxidizing species, that is, 2,4Cpd I
or 3Cpd II, an interesting region-selective conversion phe-
nomenon between sulfoxidation and H-atom abstraction
was revealed in the course of the oxidation reaction of di-
methylsulfide. The combined experimental and theoretical
study on the elucidation of the intrinsic reactivity patterns of
the HRP-I and HRP-II mimics provides a valuable tool for
evaluating the particular role of the HRP active species in
biological systems.
IV
+
+
C
(HRP-II), we generated in situ [Fe =O(TMP )(2-MeIm)] and
[FeIV=O(TMP)(2-MeIm)]0
(TMP=5,10,15,20-tetramesitylpor-
phyrin; 2-MeIm=2-methylimidazole) as biomimetics for
HRP-I and HRP-II, respectively. Their catalytic activities in ep-
oxidation, hydrogen abstraction, and heteroatom oxidation
reactions were studied in acetonitrile at ꢀ158C by utilizing
rapid-scan UV/Vis spectroscopy. Comparison of the second-
order rate constants measured for the direct reactions of the
HRP-I and HRP-II mimics with the selected substrates clearly
confirmed the outstanding oxidizing capability of the HRP-I
mimic, which is significantly higher than that of HRP-II. The
experimental study was supported by computational model-
ing (DFT calculations) of the oxidation mechanism of the se-
lected substrates with the involvement of quartet and dou-
blet HRP-I mimics (2,4Cpd I) and the closed-shell triplet spin
HRP-II model (3Cpd II) as oxidizing species. The significantly
lower activation barriers calculated for the oxidation systems
involving 2,4Cpd I than those found for 3Cpd II are in line
with the much higher oxidizing efficiency of the HRP-I mimic
Introduction
to their ability to catalyze a wide range of important oxidation
reactions. An oxo-iron(IV) porphyrin p cation radical species,
Compound I (Cpd I), has been characterized and accepted as
the most reactive species in the catalytic cycle of heme-con-
taining mono-oxygenases.[5] Even though the Cpd I intermedi-
ates of the above-mentioned enzymes possess the same heme
cofactor, the ability of these strongly oxidizing species to cata-
lyze various types of oxidation reactions is accomplished by
the utility of different proximal axial ligands such as the thio-
late ligand in P450, the phenolate ligand in catalases, or the
imidazole ligand in HRP. In particular, it was revealed that the
presence of the thiolate proximal ligand in cytochrome P450
substantially increases the oxidizing power of P450 Cpd I
toward the hydroxylation of unactivated CꢀH bonds, epoxida-
tion of C=C bonds, and N-, O-, and S-oxidation, as well as de-
hydrogenation reactions.[6] On the other hand, the electron-
pushing effect of the thiolate ligand in P450 Cpd I results in
a poorer electron acceptor ability of its heme cofactor com-
pared with the histidine-coordinated heme in Cpd I of HRP.
Therefore, the latter is thought to operate as an electron sink
Heme-containing mono-oxygenases such as cytochrome P450
(P450),[1] chloroperoxidase,[2] catalase,[3] and horseradish perox-
idase (HRP)[4] are widely distributed in living organisms owing
[a] Dr. L. Ji, Dr. A. Franke, Dr. A. Zahl, Prof. Dr. R. van Eldik
Inorganic Chemistry
Department of Chemistry and Pharmacy
University of Erlangen-Nuremberg
Egerlandstrasse 1, 91058 Erlangen (Germany)
E-mail: vaneldik@chemie.uni-erlangen.de
[b] Dr. L. Ji
College of Environmental and Resource Sciences
Zhejiang University
Yuhangtang Road 866, Hangzhou 310058 (China)
[c] Dr. M. Brindell, Dr. M. Oszajca, Prof. Dr. R. van Eldik
Faculty of Chemistry
Jagiellonian University
Ingardena 3, 30-060 Krakow (Poland)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201402347.
Chem. Eur. J. 2014, 20, 14437 – 14450
14437
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
to participate in many electron-transfer as well as oxygen-
transfer reactions.[7]
N-MeIm and Imꢀ ligands on the vFe frequency, which may
=
O
suggest that such effects should be reflected in the reactivities
of Cpd I and Cpd II mimics.[15] However, in the case of various
substituted models for Cpd I, no correlation between the Fe=O
bond energy and oxidation capability of Cpd I could be
found.[12g,14,16] An explanation for this phenomenon offered by
Gross et al.[12 g,14] is based on the multistep oxidation mecha-
nism, according to which only intramolecular electron transfer
HRP belongs to one of the first heme-containing enzymes
for which an X-ray structure is available.[8] HRP utilizes hydro-
gen peroxide for the catalytic oxidation of a wide range of or-
ganic substrates. The classical HRP cycle involving oxidation of
an inert substrate (AH) occurs as depicted in Equations (1)–
(3).[9] The first stable intermediate in the catalytic cycle of HRP,
Cpd I (HRP-I), is formed by the transfer of one oxygen atom
from H2O2 to the ferric iron of the heme in the resting state
[Eq. (1)]. HRP-I performs the oxidation of AH to form another
stable intermediate, the one-electron reduced form of Cpd I,
Compound II (Cpd II; HRP-II) [Eq. (2)]. The latter undergoes
one-electron reduction to the resting state by oxidizing the
second molecule of substrate [Eq. (3)].
III
IV
from iron to the porphyrin radical, (P+ )Fe (O)!(P)Fe (O), can
be affected by the axial ligand, the nature of which may lower
the energy barrier. The experimental results are in line with the
conclusions drawn from theoretical studies on the influence of
neutral nitrogen axial ligands on model compounds for Cpd I
and Cpd II.[17] They have shown that although axial ligands en-
large the Fe=O bond length and decrease the polarity of the
Fe=O bond for both Cpd I and Cpd II, they cannot switch the
reactivity of these two oxidants. Thus, only a meaningful effect
of axial ligands on the reaction selectivity of Cpd I could be re-
vealed, which is in agreement with previous experimental
studies showing that axial ligands in iron-porphyrin catalyst
models of Cpd I have a prominent influence on the distribu-
tion of epoxidation and hydroxylation products.[13a,18] On the
other hand, the combined experimental and theoretical studies
performed by Kang et al.[12q] demonstrated that neutral axial li-
gands can affect the reactivity of Cpd I, although their effect
on the rate constants of substrate oxidation is significantly
smaller than that observed for anionic axial ligands. Recently,
a systematic electrochemical study on Cpd I models with vari-
ous porphyrin structures and axial ligands was performed.[12v]
C
Both HRP-I and HRP-II are stable enough to have been char-
acterized by a variety of spectroscopic techniques.[10]
HRP þ H2O2 ! HRP-I þ H2O
HRP-I þ AH ! HRP-II þ A
HRP-II þ AH ! HRP þ A
ð1Þ
ð2Þ
ð3Þ
Although HRP can catalyze mono-oxygenation reactions, its
catalytic ability was shown to be poorer than that of P450.[11]
One of the obvious factors influencing the oxidation power of
P450 and HRP is the protein architecture of their distal pockets
and the accessibility of the latter for the substrates. In contrast
to HRP, the active site of P450 is highly hydrophobic and lacks
acid-base residues such as the imidazole side chain of the
distal histidine in HRP, which can operate as a catalytic inter-
mediate in the proton transfer reactions. Besides the apparent
dissimilarity in the heme environments of P450 and HRP, it
seems that the different proximal ligation of HRP-I/HRP-II and
P450-I/P450-II plays a crucial role in their catalytic abilities, and
is responsible for the observed difference in the intrinsic prop-
erties of these active species. Biomimetic studies involving the
application of synthetic iron-porphyrin models to generate
Cpd I and Cpd II mimics provide valuable insights into the fun-
damental properties of heme active species, providing infor-
mation on different factors that influence the reaction mecha-
nism and chemoselectivity of the reactive intermediates in oxy-
genation reactions. Most attention has been devoted to Cpd I
owing to its very high reactivity,[12] whereas only a few biomim-
etic studies have focused on the properties of heme Cpd II
species.[12o,u,13] The ability to catalyze different types of reac-
tions is associated with the structural differences in the pros-
thetic group and its environment, so much effort was devoted
Subsequently,
a
correlation between the reactivity of
(TMP+ )Fe =O(L) and the redox potential of FeII/FeIII in
(TMP)FeIII(L) was proposed by Takahashi et al.,[12w] who suggest-
ed that the reactivity of model Cpd I complexes with various
anionic axial ligands depends on the thermodynamic stability
of the final reaction product, that is, (TMP)FeIII(L). The influence
of axial ligation on the reactivity of model complexes for Cpd II
has been studied to a significantly lesser extent. Theoretical
calculations showed only an insignificant axial ligand influence
on the reaction barriers of substrate oxidation by (Por)FeIV=
O.[13d,17] Similarly, a recent study by Rosa et al.[13c] demonstrated
that even a strong s donor such as hydrosulfide does not
affect the H-abstraction barrier on the ground-state triplet sur-
face of model Cpd II, but has a pronounced increasing effect
on the energy barrier on the quintet surface.
IV
C
The results from our recent kinetic studies and DFT compu-
tations involving substrate oxidation by (TMPS)FeIV=O(OH) and
(TMPS)FeIV=O(N-MeIm) showed no pronounced effect of the N-
MeIm axial ligand on the reactivity of the Cpd II mimic in com-
parison with the OHꢀ-substituted analogue.[13e] The main goal
to the elucidation of the effect of the axial ligands on the FeIV= of the present work was to explore the intrinsic reactivity of
O bond strength and the reactivity of Cpd I models. Experi-
mental studies by Gross et al. clearly demonstrated the signifi-
cant effects of various anionic axial ligands on the rate of sty-
rene epoxidation.[12g,14] Similarly, results reported by Fujii et al.
showed a significant enhancement of the reactivity of Cpd I
with axially coordinated imidazole and phenolate.[12r] Reso-
nance Raman investigations show a pronounced influence of
the key active species of HRP, that is, Cpd I and Cpd II, with the
application of functional models for this enzyme. In this con-
IV
+
+
C
text, we generated in situ [Fe =O(TMP )(2-MeIm)]
and
[FeIV=O(TMP)(2-MeIm)]0 (TMP: 5,10,15,20-tetramesitylporphyr-
in), as mimics for HRP-I and HRP-II, respectively, with the use of
FeIII(TMP)(2-MeIm) (2-MeIm=2-methylimidazole) and the oxi-
dants m-CPBA (meta-chloroperoxobenzoic acid) and iodosyl-
Chem. Eur. J. 2014, 20, 14437 – 14450
www.chemeurj.org
14438
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Despite the usefulness of the spectroscopic and kinetic stud-
ies, the overall mechanistic picture of substrate oxidation reac-
tions catalyzed by heme enzymes and their mimics still re-
mains unclear. Especially, the intrinsic difference in the elec-
tronic structure of the HRP-I and HRP-II biomimetics, which is
responsible for the catalytic ability of these two active species,
still awaits further clarification. Computational chemistry has
become a useful tool for studying reaction mechanisms, as it
delivers the structures and energies of all relevant transition
and intermediate states to compose a complete mechanistic
picture. It was shown in previous computational studies that
the two close-lying spin states of Cpd I of P450, viz. the quartet
high-spin and doublet low-spin states, contribute differently to
various reaction types.[19] For instance, they give comparable
activation energies for the aliphatic hydroxylation in both spin
states,[20] whereas the addition of Cpd I to the aromatic ring or
heteroatoms is most favorable only via the doublet spin sur-
face.[21] It is generally believed that the H-atom abstraction
ability of Cpd II is weaker than that of Cpd I;[13b] however, the
oxidizing capability of Cpd II in P450 reactions has been dem-
onstrated previously.[13c,22] Moreover, in our earlier experimental
work, the porphyrin mimic for Cpd II turned out to be the
most efficient catalyst for hydride transfer reactions.[12u] Thus,
there is a need to clarify the fundamental factors influencing
the reactivity of Cpd I and Cpd II of HRP from the computed
electronic features.
In the current work, we used density functional theory (DFT)
to investigate all the reaction paths of the experimental work
mentioned above. The combined experimental and theoretical
study presented herein demonstrates the difference in the
overall course of the substrate oxidation reactions with the ap-
plication of HRP-I and HRP-II biomimics.
Results and Discussion
Scheme 1. A) Structure of FeIII(TMP)(2-MeIm) together with a schematic rep-
resentation of the formation of HRP-I and HRP-II models using m-CPBA and
PhIO as oxidants, respectively. B) Structures of the selected substrates and
their main oxidation products formed in the oxidation reactions catalyzed
by Cpd I and Cpd II models (see text and Experimental section for informa-
tion on the side products obtained during the substrate oxidation reactions
involving the Cpd II mimic).
Generation and identification of the HRP-I and HRP-II
mimics
Addition of a small excess of 2-MeIm (3:1) to an acetonitrile so-
lution of [FeIII(TMP)(Cl)] resulted in the formation of mono- and
bis-2-methylimidazole derivatives of FeIII(TMP), viz. [FeIII(TMP)(2-
MeIm)]+ and [FeIII(TMP)(2-MeIm)2]+, respectively (see the UV/
Vis spectra recorded for this reaction at ꢀ158C in Figure S1,
Supporting Information). This is in line with the conclusion
drawn from other studies involving substitution reactions of
synthetic iron(III) porphyrins with different nitrogen base
ligand derivatives.[23] As shown, the formation of the mono-imi-
dazole derivative of iron(III) porphyrins as a sole species in the
reaction medium is almost impossible, even at a very low imi-
dazole concentration. This is a consequence of the high equi-
librium constant for the second imidazole ligation, which is
often much higher than that for the formation of the mono-
imidazole derivative. Thus, [FeIII(TMP)(2-MeIm)]+ and [FeIII(TMP)-
(2-MeIm)2]+ species exist in equilibrium under the selected re-
action conditions. As shown in Figure 1, the addition of an
excess of m-CPBA (1:6) to the solution of [FeIII(TMP)(2-MeIm)]+/
[FeIII(TMP)(2-MeIm)2]+ leads to the formation of iron(IV)-oxo
benzene (PhIO) (see Scheme 1A). With the application of
a low-temperature rapid-scan technique, we were able to per-
form UV/Vis characterization of the HRP-I and HRP-II mimics,
and to study their reactivities toward the selected inert sub-
strates. To cover the common oxidation reaction types that
occur in nature, we focused on C=C bond epoxidation, CꢀH
bond activation, and heteroatom oxidation reactions catalyzed
by both Cpd I and Cpd II models for HRP.
Several typical substrates were chosen, such as cyclohexe-
ne[12k] and bicyclo[2.2.1]hept-2-ene (BICY)[12b] for epoxidation,
9,10-dihydroanthracene (DHA)[12o] for CꢀH bond activation, and
dimethylsulfide (DMS)[12n] for heteroatom oxidation (see
Scheme 1B for the structures of the studied substrates and
their main oxidation products).
Chem. Eur. J. 2014, 20, 14437 – 14450
www.chemeurj.org
14439
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 1. Top: Spectral changes accompanying the formation of the Cpd I
mimic in acetonitrile at ꢀ158C. Experimental conditions: [FeIII(TMP)(Cl)] =
5.3ꢁ10ꢀ6 m, [2-methylimidazole] = 1.6ꢁ10ꢀ5 m, [m-CPBA] = 3.1ꢁ10ꢀ5 m.
Bottom: Kinetic trace recorded for this reaction at 670 nm.
Figure 2. A) Spectral changes recorded for the formation of the Cpd II mimic
in acetonitrile at ꢀ158C. B) Kinetic trace recorded for this reaction at the
Soret band. Experimental conditions: [FeIII(TMP)(Cl)]=5.5ꢁ10ꢀ6 m, [2-methyl-
imidazole]=1.6ꢁ10ꢀ5 m, [PhIO]=6.3ꢁ10ꢀ6 m.
porphyrin cation radical species (model for Cpd I) characterized
by a significant absorbance decrease in the Soret band and
a substantial absorbance increase in the range 550–700 nm. In-
terestingly, the kinetic trace recorded for this reaction at
670 nm (see bottom part of Figure 1) clearly indicates that the
Cpd I mimic is generated in two reaction steps. This reactivity
behavior can be explained in view of the fact that under the
selected conditions, the reactive species [FeIV(TMP+·)(O)(2-
MeIm)]+ can be formed from both the mono- and bis-2-meth-
ylimidazole derivatives of FeIII(TMP). As known from other relat-
ed studies,[13e,24] the formation of the oxo-iron(IV) species from
bis-imidazole-substituted iron(III) porphyrins proceeds signifi-
cantly more slowly than from mono-imidazole derivatives.
Thus, the fast and slow reaction steps observed in the present
study can be ascribed to the formation of the same reactive
species, that is, [FeIV(TMP+·)(O)(2-MeIm)]+, from [FeIII(TMP)(2-
MeIm)]+ and [FeIII(TMP)(2-MeIm)2]+, respectively.
trile solution of [FeIII(TMP)(Cl)] at ꢀ158C. The formation of [FeIV-
(TMP)(O)(2-MeIm)] was accompanied by a small but significant
increase in absorbance at the Soret band and at 565 nm
(Figure 2), characteristic of the formation of iron(IV)-oxo por-
phyrin species. Notably, the spectral characteristics of the ob-
tained species reveal a Q-band maximum that is significantly
redshifted in comparison to that known for [FeIV(TMP)(O)(Cl)]ꢀ,
which can be regarded as an indication of 2-MeIm ligation in
the trans position to the oxo axial ligand. Similar shifts in the
Soret and Q bands to longer wavelengths have already been
observed following the addition of N-MeIm to (TMP)FeIV=O in
organic solvents[15,25] or to (TMPS)FeIV=O(OH) in aqueous solu-
tion,[13e] and have been ascribed to the formation of N-methyli-
midazole-substituted oxo-iron(IV) porphyrin.
The spectroscopic identification of both oxidizing species,
that is, [FeIV(TMP+·)(O)(2-MeIm)]+ and [FeIV(TMP)(O)(2-MeIm)],
along with their relatively long lifetimes at low temperatures in
acetonitrile, provided the opportunity to test their reactivities
toward various inert organic substrates.
For the generation of the Cpd II mimic, an equivalent
amount of the nonacidic oxidizing agent iodosylbenzene
(PhIO) was used to oxidize [FeIII(TMP)(2-MeIm)]+/[FeIII(TMP)(2-
MeIm)2]+ produced by the addition of 2-MeIm to an acetoni-
Chem. Eur. J. 2014, 20, 14437 – 14450
www.chemeurj.org
14440
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Oxidation of inert organic substrates by the HRP-I
and HRP-II mimics
Oxidation of the selected organic substrates by the
Cpd I and Cpd II mimics generated in situ was per-
formed under pseudo-first-order conditions by using
different excesses of the substrates. Upon addition of
the selected substrate to solutions of Cpd I or Cpd II,
the decomposition reaction of the reactive intermedi-
ate back to the starting iron(III) porphyrin complex
could be monitored by UV/Vis absorption spectrosco-
py. Typical spectral changes and the kinetic trace re-
corded during the reduction of [FeIV(TMP+·)(O)(2-
MeIm)]+ by BICY are shown in Figure 3.
The substrate oxidation reactions were followed ki-
netically by the depletion of Cpd I (decay at 670 nm)
or Cpd II (decay at 417 nm) upon the addition of vari-
ous concentrations of the substrate. Kinetic traces
were fitted to a pseudo-first-order decay function
and used to determine the observed rate constants,
kobs. In all cases, plots of kobs versus substrate concen-
tration showed a linear dependence without a signifi-
cant intercept (see Figure 4A and B for the oxidation
reactions of the Cpd I and Cpd II mimics, respective-
ly). The second-order rate constants determined by
linear fits of the obtained data are listed in Table 1.
As seen from Table 1, the HRP-I mimic appears to
be a considerably more effective catalytic species
than its one-electron reduced form (HRP-II model) for
all reactions considered in the present study. The out-
standing oxidative ability of the Cpd I model is espe-
cially clear in the DMS oxidation reaction, for which
this reactive species outperforms the Cpd II model by
several orders of magnitude on the basis of the
second-order rate constants. Moreover, the oxidation
of DMS by the Cpd I model appears to be the fastest
of all the oxidation reactions studied. In contrast, the
rate of DMS oxidation by Cpd II is of the same order
of magnitude as found for other types of oxidation
reactions involving this reactive species (see, for ex-
ample, the CꢀH bond activation of DHA by Cpd II).
The trend shown in Table 1 suggests that the epoxi-
dation rate constants determined for cyclohexene
and BICY by the Cpd I and Cpd II models correlate
with the C=C bond lengths of these substrates (BICY:
1.340 ꢂ; cyclohexene: 1.337 ꢂ). The C=C bond length
represents the p bond strength of the olefin, so it is
reasonable to assume that the energy required for
the conversion of the double bond of the olefin into
a single bond of the epoxide product could deter-
mine the overall reactivity pattern of the studied ep-
oxidation reactions. Notably, an experimental study
by Crestoni et al.[26] and theoretical studies by de
Visser et al.[27] demonstrated that the trend observed
Figure 3. Top: Spectral changes accompanying the re-formation of the starting iron(III)
porphyrin complex from Cpd I after addition of BICY in acetonitrile at ꢀ158C. Experimen-
tal conditions: [FeIII(TMP)(Cl)]=5.5ꢁ10ꢀ6 m, [2-methylimidazole]=1.61ꢁ10ꢀ5 m,
[BICY]=9.1ꢁ10ꢀ4 m. Bottom: Kinetic trace recorded for this reaction at the Soret band.
Table 1. Values of the second-order rate constants for the direct oxidation of selected
organic substrates by the HRP-I and HRP-II mimics produced in situ in MeCN at
ꢀ158C.
Substrate
Reaction
kCpd I
[Mꢀ1 sꢀ1
kCpd II
[Mꢀ1 sꢀ1
]
]
cyclohexene
BICY
DHA
epoxidation
epoxidation
CꢀH bond activation
heteroatom oxidation/CꢀH bond activation
48ꢁ1
0.70ꢁ0.02
1.30ꢁ0.03
3.40ꢁ0.05
4.0ꢁ0.1
166ꢁ3
161ꢁ3
14680[a]
DMS
[a] Owing to the very fast reaction we were unable to measure the kobs dependence
on [DMS] at ꢀ158C; the reported second-order rate constant for DMS oxidation by
Cpd I was calculated from the pseudo-first-order rate constant measured at [DMS]=
4.7ꢁ10ꢀ4 m.
in the epoxidation reactions correlates well with the ionization
potentials of the studied olefins, that is, the rate of oxygen
atom transfer increases (activation enthalpy for the oxidation
processes decreases) as the ionization energy of the olefin de-
creases.
Chem. Eur. J. 2014, 20, 14437 – 14450
www.chemeurj.org
14441
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
environmental properties of the distal active site of HRP, addi-
tional calculations were performed for the oxidation routes in-
IV
+
+
volving [Fe =O(TMP )(5-MeIm)] and [FeIV=O(TMP)(5-MeIm)]0
derivatives (see Scheme S1 in Supporting Information for an
example of energy potential profiles constructed for DMS oxi-
dation by such oxidant models). However, no meaningful dif-
ference in the reactivity pattern or energy barriers could be
observed for the oxidation routes with the participation of the
2-MeIm- and 5-MeIm-substituted Cpd I/Cpd II mimics.
C
Epoxidation mechanism
The potential energy surface with key geometric parameters
obtained for the epoxidation of cyclohexene by the HRP-I
mimic is depicted in Scheme 2A. The epoxidation reaction
starts from the doublet and quartet states of Cpd I, 4,2Cpd I,
which are very close in energy (within 0.46 kcalmolꢀ1). The
spin-state measurements performed for the HRP-I mimic using
1
the H NMR Evans method (see Experimental Section for a de-
tailed description of the 1H NMR experiments) revealed that
the spin of this intermediate is closer to S=3/2 than to S=1/2.
The initial and rate-determining step in the reaction mecha-
nism is CꢀO bond formation via a transition state TSO. Impor-
tantly, both spin routes are effectively concerted with direct
barrier-free ring closure to the epoxidation products 2,4P. As
shown in Figure S4 (Supporting Information), the geometry
scan along the CꢀO bond length yielded the epoxidation prod-
uct directly even on the quartet pathway. The transition-state
geometries reveal that the doublet spin-state pathway has an
early transition state with a CꢀO length of 2.275 ꢂ, whereas
the quartet transition state is later with a CꢀO length of
2.073 ꢂ. As a result of the geometric difference in the transi-
Figure 4. Determination of the second-order rate constants for substrate oxi-
dation by Cpd I (A) and Cpd II (B) with cyclohexene (squares), BICY (trian-
gles), DHA (circles), and DMS (inverted triangles). Experimental conditions:
[FeIII(TMP)(Cl)]=5.5ꢁ10ꢀ6 m, [2-methylimidazole]=1.61ꢁ10ꢀ5 m in acetoni-
trile at ꢀ158C.
4
tion state, the barrier gap between TSO (17.51 kcalmolꢀ1) and
2TSO (13.64 kcalmolꢀ1) is 3.87 kcalmolꢀ1, which means that the
reaction proceeds predominantly through the doublet reaction
path.
Computational modeling of the reaction mechanism
Figure S2 in the Supporting Information shows the optimized
geometries of [FeIV=O(TMP+·)(2-MeIm)]+ (HRP-I model) and
[FeIV=O(TMP)(2-MeIm)]0 (HRP-II model), and the corresponding
key spin density data. It is seen from the spin density distribu-
tion that one-electron reduction of the quartet and doublet
HRP-I mimics (2,4Cpd I) fills the singly occupied a2u orbital of
the porphyrin ring to give the triplet spin state HRP-II (3Cpd II).
Thus, the Cpd II model is a closed-shell species with a spin
density of only 0.04 on the porphyrin ring, which is in line with
the results of previous DFT and QM/MM studies on the triplet
ground state of the oxo-ferryl species for HRP-II.[28] The geome-
try of the Cpd I and Cpd II mimics reveals that the FeꢀN bond
of the coordinated 2-MeIm axial ligand in Cpd II is significantly
As shown in Scheme 2B, the epoxidation of cyclohexene by
the HRP-II mimic starts from the triplet spin-state surface of
3Cpd II. The reactivity picture for epoxidation by 3Cpd II ap-
pears to differ completely from that seen for the reaction in-
volving 4,2Cpd I as the oxidizing species. It is described by
a stepwise mechanism with a significant ring-closure barrier
leading to the epoxide product. The initial reaction is the rate-
3
determining step to form a radical intermediate, INT, via a Cꢀ
3
O bond-formation transition state TSO (20.75 kcalmolꢀ1). The
2
latter is 7.57 kcalmolꢀ1 higher in energy than TSO (13.18 kcal
molꢀ1) for the HRP-I mimic as oxidant. This finding is in line
with the experimental results obtained for cyclohexene epoxi-
dation, showing that the ratio of the rate constants for cyclo-
hexene oxidation by HRP-I and HRP-II is about 70 (48.4 vs.
0.7mꢀ1 sꢀ1). Moreover, product analysis of the reaction mixtures
containing the HRP-I and HRP-II mimics as oxidants (see Experi-
mental Section for detailed description of the product analysis
experiments) revealed that in the former case cyclohexene
oxide was formed as the major product of cyclohexene oxida-
tion, whereas cyclohexen-1-one was formed in the case of cy-
clohexene oxidation by the HRP-II mimic. The formation of the
2
longer than that of the Cpd I species (Cpd II: 2.343 ꢂ; Cpd I/
4Cpd I: 2.238/2.232 ꢂ), whereas the length of the Fe=O bond is
almost the same for both reactive species. On the basis of
such optimized geometries, the DFT calculations were per-
formed for the three reaction systems involving epoxidation,
CꢀH bond activation, and heteroatom oxidation catalyzed by
both Cpd I and Cpd II mimics. Furthermore, since the 5-MeIm-
substituted Cpd I and Cpd II models most reliably reflect the
Chem. Eur. J. 2014, 20, 14437 – 14450
www.chemeurj.org
14442
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 2. Potential energy profiles for cyclohexene epoxidation by the HRP-I mimic via quartet and doublet states (A), and cyclohexene epoxidation by the
HRP-II mimic via the triplet state (B), along with optimized geometries of the key reaction species. Relative energies are given at the solvated environment
(UB3LYP/LACVP**//LACVP**+Bulk Polarity+ZPE) level. Geometric parameters (lengths in ꢂ) are given in the order for the Cpd I system: doublet state (quartet
state). TSO =transition state of CꢀO bond formation, INT=oxide intermediate complex with CꢀO bond formation by Cpd II mimic, TSrc =ring-closure transition
state, P=epoxide product + Fe(Por)(2-MeIm).
latter product can be the result of the significant ring-closure
the epoxidation of BICY listed in Table 1 (Cpd I: 165.6mꢀ1 sꢀ1
Cpd II: 1.3mꢀ1 ꢀ1). Although the steric difference between cy-
;
3
barrier for a radical intermediate, INT, which leads to its pro-
s
longed lifetime, and therefore, to the formation of the side
products. In addition, the overall reaction thermodynamics for
the system involving the HRP-II mimic (ꢀ10.16 kcalmolꢀ1) are
less exothermic by 21.38 kcalmolꢀ1 than obtained for HRP-I in
the doublet spin-state route (ꢀ31.54 kcalmolꢀ1).
clohexene and BICY cannot be negligible, an important reason
for the larger rate constant for BICY than for cyclohexene may
be the difference in reactivity. It is seen that the computed Cꢀ
O bond-formation barrier difference between cyclohexene and
BICY for Cpd I along the doublet pathway (0.46 kcalmolꢀ1) and
Cpd II along the triplet pathway (2.77 kcalmolꢀ1) is in line with
the experimentally observed difference in rate constants. It
should, however, be kept in mind that the computed energy
barriers reflect only the enthalpy barrier, whereas the experi-
mental rate constants are determined by both the enthalpy
and entropy barriers. This is especially important in the case of
the entropy barrier, which could be significantly different for
cyclohexene and BICY also in terms of steric factors.
For the epoxidation of BICY by HRP-I and HRP-II, the rate-de-
termining step was found to involve CꢀO bond formation. The
obtained doublet/quartet CꢀO bond-formation barrier for
Cpd I is 13.18/15.67 kcalmolꢀ1, and the corresponding value
for Cpd II in the triplet spin state is 17.98 kcalmolꢀ1. Thus, ep-
oxidation of BICY catalyzed by HRP-I should mainly proceed
through the doublet spin pathway, and this reaction should be
significantly faster than that involving HRP-II. This finding is in
good agreement with the experimental results obtained for
Chem. Eur. J. 2014, 20, 14437 – 14450
www.chemeurj.org
14443
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 3. Potential energy profiles for DHA CꢀH bond activation by the HRP-I mimic via quartet and doublet states (A), and by the HRP-II mimic via the trip-
let state (B), along with optimized geometries of the key reaction species. Relative energies are given at the solvated environment (UB3LYP/LACVP**//
LACVP**+Bulk Polarity+ZPE) level. Geometric parameters (lengths in ꢂ) are given in the order for the Cpd I system: doublet state (quartet state).
TSH =transition state of H-abstraction, INT=intermediate complex with H abstracted by Cpd II mimic, TSreb =rebound transition state, PH =hydroxylated prod-
uct + Fe(Por)(2-MeIm), Pd =desaturated product + Fe(Por)(2-MeIm).
CꢀH bond activation mechanism
as the oxidizing species, the rate-determining step is described
by CꢀH bond activation through an H-abstraction transition
The computed potential energy profiles for the CꢀH bond acti-
vation of DHA by the HRP-I and HRP-II mimics are shown in
Scheme 3A and B, respectively. In the system involving HRP-I
state (2,4TSH), followed by either a barrier-free second H-atom
transfer to form the desaturation product complex 2,4Pd or
a barrier-free rebound to form the hydroxylation product com-
Chem. Eur. J. 2014, 20, 14437 – 14450
www.chemeurj.org
14444
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
plex 2,4PH. The geometry scans (see Scheme S5–S7, Supporting
Information), reveal that a hydroxo-iron species interacting
through H-bonding with the DHA radical appears as an inter-
mediate (2,4INTH). However, the latter disappears during the
final refinement of the structures and jumps directly to the hy-
droxylation and desaturation product complexes. As seen from
Scheme 3A, the desaturation process is thermodynamically
preferred over the hydroxylation reaction route, which is in ac-
cordance with the results of product analysis experiments
showing that anthracene is the major product of DHA oxida-
tion by HRP-I models (see Experimental section). The energy
tion (left side) of DMS by the HRP-I mimic. The sulfoxidation
mechanism involves O-addition to the S lone pair of DMS via
transition states 2,4TSS to form the oxidized products 2,4PS. The
sulfoxidation transition state in the doublet state (10.24 kcal
molꢀ1) is more favorable by 4.86 kcalmolꢀ1 than in the quartet
state (15.10 kcalmolꢀ1), indicating that sulfoxidation is more
likely to occur along the doublet route, which is in line with
other heteroatom oxidation reactions.[21c–f] The O-addition bar-
rier gap between the two spin states is also reflected in the
transition-state geometries: the doublet route has an early
transition state with a SꢀO distance of 2.582 ꢂ, whereas the
quartet route has a later transition state with a SꢀO distance
of 2.156 ꢂ. CꢀH bond activation of DMS by the HRP-I mimic
(left side of Scheme 4A) involves the formation of the oxidized
product 2,4PH. The latter is formed in a concerted manner
through an H-abstraction transition state, which exhibits an
energy splitting with an LS state (16.64 kcalmolꢀ1) lower than
the HS state (19.14 kcalmolꢀ1) by 2.5 kcalmolꢀ1. Therefore, it is
suggested that the CꢀH bond activation reaction of DMS cata-
lyzed by the HRP-I mimic is more likely to occur on the dou-
blet spin-state surface. On the other hand, in view of the fact
that the activation energy for the CꢀH bond activation of DMS
difference between TSH (14.20 kcalmolꢀ1) and TSH (16.68 kcal
molꢀ1) is about 2.5 kcalmolꢀ1, which means that the reaction
should proceed only through the doublet spin pathway, as
was also found for the epoxidation reaction.
2
4
The energy profile for the CꢀH bond activation of DHA by
3
the HRP-II mimic starts from the TSH state (Scheme 3B), and
involves an H-abstraction step with an energy barrier of
16.42 kcalmolꢀ1. This leads to the formation of a DHA radical
coordinated to the FeIII-OH species, the 3INTH intermediate,
which may go through a high-energy OH rebound transition
state (3TSreb: 11.69 kcalmolꢀ1) to form the FeII hydroxylation
product complex 3PH. However, the presence of a very flat
energy surface suggested by the geometry scan for the DHA
2
is larger than for sulfoxidation (by 6.39 kcalmolꢀ1 for Cpd I), it
can be concluded that the HRP-I mimic will perform rapid sul-
foxidation of DMS with a barrier of only 9.79 kcalmolꢀ1
through the doublet-state pathway. This finding is supported
by the experimental results, which show that the rate constant
for DMS oxidation by model HRP-I is too high to be measured
at ꢀ158C, indicating a very low energy barrier for this reaction.
The energy profile for the sulfoxidation (left side) and CꢀH
bond activation (right side) of DMS by the HRP-II mimic in the
triplet state is depicted in Scheme 4B. The barrier for the sul-
3
radical rotation from INTH (see Figure S8 in Supporting Infor-
mation) suggests the possibility of the formation of a lower-
3
energy intermediate, INTd, in which the FeIII-OH species inter-
acts with the DHA radical through H-bonding of the hydrogen
atom in the CH2 group of DHA. The latter may be followed by
the second H-atom transfer (3TSd: 10.91 kcalmolꢀ1) to form the
FeII desaturation product complex 3Pd. However, the high-
energy second H-atom transfer barrier or OH rebound barrier
predicts a long lifetime of the DHA radical. On the basis of our
further calculations, the dissociation of the DHA radical is fa-
vored by 6.2 kcalmolꢀ1 (see Supporting Information for a de-
tailed description of the dissociation energy calculation), which
3
foxidation of DMS by Cpd II (25.54 kcalmolꢀ1) is significantly
2
higher than for the sulfoxidation reaction involving Cpd I as
oxidant. The CꢀH bond activation of DMS by the HRP-II mimic
displays a similar mechanism to that seen for the DHA CꢀH
bond activation process, that is, the reaction starts from an ini-
tial rate-determining H-abstraction step with the transition
state 3TSH (18.34 kcalmolꢀ1) to form an intermediate 3INT.
Owing to the existence of a significant radical rebound transi-
tion state, 3TSreb (8.57 kcalmolꢀ1), 3INT can dissociate more
easily to the DMS radical and FeIII-OH species with a dissocia-
tion energy of 1.3 kcalmolꢀ1. Importantly, the energy barrier
for the initial H-abstraction reaction step lies 7.2 kcalmolꢀ1
lower than the energy barrier found for the DMS sulfoxidation
pathway (see left side of Scheme 4B). Thus, the CꢀH bond acti-
vation path should dominate the mechanism of DMS oxidation
catalyzed by the HRP-II mimic. This is in contrast with the
mechanism of DMS oxidation by the 2,4Cpd I species, for which
the energy barrier for the sulfoxidation pathway was consider-
ably lower for the H-abstraction step of the CꢀH bond activa-
tion process. Interestingly, the magnitude of the energy barrier
3
means that the DHA radical in INTH is relatively weakly bound
to the FeIII-OH species. Therefore, it is reasonable to propose
that dissociation of the DHA radical with formation of the FeIII
product may be more preferable than the desaturation or hy-
droxylation pathways. Similar conclusions were drawn for the
mechanism of CꢀH bond activation by non-heme FeIVO[29] and
MnIVO[30] complexes. This finding is confirmed by product anal-
ysis experiments (see Experimental Section) showing that oxi-
dation of DHA by the HRP-II mimic results in the formation not
only of anthracene, but also of a significant amount of 9,10-di-
hydro-9-anthracenol as a side product.
For the rate-determining H-abstraction step, the computed
barrier for Cpd I in the doublet state is 2.51 kcalmolꢀ1 lower
than for Cpd II in the triplet state, which is in line with the ex-
perimentally observed rate constants (Cpd I: 160.6mꢀ1 sꢀ1
Cpd II: 3.4mꢀ1 ꢀ1) shown in Table 1.
;
s
3
for CꢀH bond activation of DMS by Cpd II approaches that of
the energy barrier for the epoxidation of cyclohexene and CꢀH
bond activation of DHA by this oxidizing species. This finding
is nicely reflected by the experimental results obtained for the
oxidation reactions of the HRP-II mimic, which show that the
Heteroatom oxidation reaction
Scheme 4A displays the potential energy profile and critical
species for the sulfoxidation (right side) and CꢀH bond activa-
Chem. Eur. J. 2014, 20, 14437 – 14450
www.chemeurj.org
14445
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 4. Potential energy profiles of DMS oxidation by HRP-I on quartet and doublet state (A) and by HRP-II on triplet state (B) along with optimized geo-
metries of the key reaction species. Relative energies are given at the solvation environment (UB3LYP/LACVP**//LACVP**+Bulk Polarity+ZPE) level. Geomet-
ric parameters (lengths in ꢂ) are given in the order for the Cpd I system: doublet state (quartet state). TSS =transition state of S-O formation, TSH =transition
state of H-abstraction, INT=intermediate complex with H abstracted by Cpd II, TSreb =rebound transition state, PS =sulfoxidation product + Fe(Por)(2-MeIm),
PH =CꢀH bond activation product + Fe(Por)(2-MeIm).
Summary of the theoretical approach and correlation with
the experimental study
values of the oxidation rate constants found for all three reac-
tion types are of the same order of magnitude. Unfortunately,
the HPLC analysis of the oxidation products of DMS involving
the Cpd I and Cpd II mimics did not furnish any additional in-
formation regarding the proposed difference in the oxidation
mechanism of DMS by these oxidants (see Experimental sec-
tion for more details). This results mainly from the fact that
both the sulfoxidation product (DMSO) and hydroxylation
product (or other side products formed because of the dissoci-
ation of the DMS radical) are relatively small, hydrophilic mole-
cules, which are very difficult to separate with the applied
method. Moreover, it was observed that DMS also reacts spon-
taneously with m-CPBA and PhIO to form mainly DMSO as an
oxidation product. Thus, spontaneous oxidation of DMS, al-
though slower than in the presence of the catalyst, makes the
identification of the obtained oxidation products and the cal-
culation of their yields even more problematic.
DFT calculations carried out for the three reaction systems in-
volving epoxidation, CꢀH bond activation, and heteroatom oxi-
dation catalyzed by Cpd I and Cpd II mimics of horseradish
peroxidase clearly reveal that the HRP-I mimic is a significantly
better oxidant than its one-electron reduced form, HRP-II. This
is in accordance with reports dealing with theoretical studies
on the reactivities of various model ferryl complexes of P450
(QM/MM method)[31] and oxo-iron porphyrin systems with chlo-
ride as an axial ligand (DFT method).[13d] As previous theoretical
work demonstrated, Cpd I of HRP, intrinsically, should also be
an efficient oxidant in various types of oxidation reactions
compared with Cpd I of P450,[32] which is proven by our cur-
rent experimental and theoretical study. Moreover, the HRP-II
mimic can perform oxidation reactions from a barrier size
lower than 20 kcalmolꢀ1, although these processes are not as
Chem. Eur. J. 2014, 20, 14437 – 14450
www.chemeurj.org
14446
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
efficient as those involving Cpd I as oxidant. All three reaction
types catalyzed by the HRP-I mimic represent an effective con-
certed process via the doublet state, whereas the oxidation re-
actions with the HRP-II model proceed according to a stepwise
mechanism. Note that the ring-closure barrier for the epoxida-
tion reaction and especially the radical rebound barrier for the
CꢀH bond activation reaction and heteroatom oxidation are
significant for the HRP-II catalyzed system. This implies that
Table 2. Comparison of the free energy of activation determined experi-
mentally with the energy barriers for the rate-determining step calculated
by the DFT method.
Cpd I
Energy barrier DG
Cpd II
Energy barrier
¼[a]
¼[a]
Substrate
DG
[kcalmolꢀ1
3
]
for 2,4Cpd I[b]
[kcalmolꢀ1
[kcalmolꢀ1
]
for Cpd II
[kcalmolꢀ1
]
]
cyclohexene 13.1ꢁ0.1
13.6 (17.5)
13.2 (15.7)
14.2 (16.7)
10.2 (15.1)[c]
16.6 (19.1)[d]
15.2ꢁ0.1
14.9ꢁ0.1
14.4ꢁ0.1
14.3ꢁ0.1
20.8
18.0
3
the intermediates INT formed during the course of the Cpd II
BICY
DHA
12.4ꢁ0.1
12.5ꢁ0.1
16.4
oxidation reaction should have relatively long lifetimes. As
shown previously, the very high rebound or desaturation barri-
ers allow fast dissociation of these intermediates to form the
substrate radical and FeIII-OH species.
25.6[c]
18.3[d]
DMS
10.1
[a] Free energy of activation calculated from the second-order rate con-
stants at ꢀ158C. [b] Energy barrier calculated by DFT for 2Cpd I and
4Cpd I (in parentheses). [c] Energy barriers calculated for S-oxidation path-
way. [d] Energy barriers calculated for CꢀH abstraction pathway.
Figure S3 in the Supporting Information presents spin densi-
ty data (1) characterizing the electronic structures of the transi-
tion states for cyclohexene epoxidation (CꢀO bond formation),
DHA CꢀH bond activation (H-abstraction), and DMS oxidation
(SꢀO bond formation and H-abstraction) processes. For all
three reaction types, the formation of the transition state from
¼
DG values determined experimentally with the energy barri-
2
the corresponding substrate and Cpd I species is associated
ers computed theoretically can sometimes be misleading and
lead to inappropriate conclusions. This results from the fact
with a considerable decrease in the spin density on the por-
phyrin ring. This means that one of the electrons from the sub-
strate is transferred into the a2u orbital of the porphyrin ring in
the transition state, whereas iron remains in its initial oxidation
state, that is, FeIV. In contrast, the formation of the transition
state with the involvement of 3Cpd II does not result in a signifi-
cant change in the spin density on the porphyrin ring in the
transition state, implying that one of the electrons from the
corresponding substrate enters into one of the higher-lying
iron dp* orbitals (dp*xz or dp*yz) to change the iron valence
state to FeIII. The spin density data presented for the transition
states involving the epoxidation, CꢀH bond activation, and
heteroatom oxidation reactions confirm very well the funda-
mental reason why Cpd I is a more efficient oxidant than
Cpd II. As mentioned in earlier reports,[13d,33] the porphyrin’s
“hole” (singly occupied a2u orbital) of the Cpd I species plays
a very important role as an electronic sink, favoring the oxida-
tion reactions. The lower a2u orbital of the porphyrin ring is al-
ready fully filled in Cpd II, so the transfer of an electron from
the substrate into the higher-energy iron dp* orbital will re-
quire more energy than is needed for the formation of the
transition state from 2Cpd I possessing only the low-lying
singly occupied a2u porphyrin orbital (porphyrin “hole” serving
as an electronic sink). This feature is reflected by the signifi-
cantly higher energy barriers obtained for the HRP-II reaction
system compared with those found for HRP-I as the oxidizing
species.
¼
that DG values calculated from the second-order rate con-
¼
stants are determined by both enthalpy (DH ) and entropy
¼
¼
(ꢀTDS ) terms, and thus, also depend on the value of DS
and the selected temperature. In the theoretical approach, the
entropic contribution to the free energy of activation is as-
sumed to be negligible, and therefore, the computed activa-
tion energies reflect only the enthalpy barriers. The above con-
siderations are in line with the conclusions drawn from studies
by Takahashi et al.,[34] as well as from our very recent mechanis-
tic investigations involving temperature and pressure effects
on CꢀH abstraction reactions by Cpd I and II mimics.[35]
The results of theoretical calculations on the oxidation of di-
methyl sulfide reveal an interesting region-selective conversion
phenomenon between sulfoxidation and CꢀH bond activation
depending on the chemical nature of the oxidized species.
Namely, the sulfoxidation path dominates the oxidation mech-
2
anism if Cpd I species are used as the catalyst, whereas CꢀH
bond activation becomes a favorable process in the case of
3
DMS oxidation by the Cpd II mimic. Importantly, this finding is
¼
nicely reflected by the trend observed in the values of DG
determined experimentally. As shown in Table 2, the low value
¼
of DG determined for DMS oxidation by Cpd I corresponds
very well to the low energy barrier calculated for the sulfoxida-
2
tion pathway involving the Cpd I species as oxidant. The CꢀH
activation path in this case is associated with a much higher
energy barrier, which is even somewhat higher than the
energy barrier calculated for CꢀH bond activation in DHA by
the Cpd I model. On the contrary, the experimental value of
Table 2 summarizes the values of the free energy of activa-
tion calculated for the substrate oxidation reactions based on
the second-order rate constants determined experimentally
¼
DG found for DMS oxidation by the Cpd II mimic correlates
¼
(DG at ꢀ158C) together with the DFT energy barriers related
much better with the energy barrier calculated for the CꢀH ac-
tivation pathway, as the energy barrier for the pathway involv-
ing oxidation of the sulfur atom is significantly higher (ca.
7.3 kcalmolꢀ1) than that of the alternative oxidation route.
to the rate-determining steps of particular oxidation routes. In
general, a good correlation can be found between the values
¼
of DG and the computed energy barriers, that is, the increase
¼
in the values of the free energy of activation (decrease in the
rate of oxidation reaction) is accompanied by an increase in
the calculated energy barriers. However, a direct comparison of
Moreover, the values of DG determined for the CꢀH activa-
tion in DHA and for DMS oxidation by the Cpd II mimic are
almost the same, and correlate well with the energy barriers of
Chem. Eur. J. 2014, 20, 14437 – 14450
www.chemeurj.org
14447
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
the C-H activation pathways calculated for the oxidation of
these substrates by the 3Cpd II mimic (the rate-determining
step of both oxidation routes, that is, desaturation of DHA and
hydroxylation of DMS, is the same, and involves a hydrogen
abstraction reaction). Thus, the theoretical and experimental
results clearly demonstrate that in some cases, the chemical
nature of the oxidant (its electron or proton affinity) can cause
a change in the region-selectivity of the substrate oxidation. It
seems that the formation of a SꢀO bond with the involvement
of the Cpd I mimic, characterized by high electron affinity, is
a more favorable process than the sulfoxidation reaction with
the participation of its one-electron reduced form, the Cpd II
model. The higher basicity (proton affinity) of the latter species
triggers an alternative oxidation pathway in which the forma-
tion of the OꢀH bond (formal transfer of one electron and one
proton from the methyl group of DMS to the FeIV=O entity),
appears to be energetically more beneficial than the transfer
of oxygen to the sulfur atom of the substrate. These considera-
tions resemble the conclusions drawn from experimental[36]
and theoretical[37] studies, which clearly showed that the
events accompanying the formation of the transition state in
the course of Cpd I/Cpd II-catalyzed oxidation reactions are not
only determined by the electron affinity of the oxidant, but are
also related to the proton affinity (basicity, pKa) of the hypo-
thetical one-electron reduced species formed after electron
transfer.
tions. However, the different distribution of oxidation products
obtained for the reactions catalyzed by the HRP-I and HRP-II
mimics must not always be an aftermath of rearrangement
processes that occurred in the reaction medium. This finding is
nicely illustrated by the results of computational modeling of
the reaction mechanism performed in this study for the oxida-
3
tion of DMS by 2,4Cpd I and Cpd II species. They show an inter-
esting region-selective conversion phenomenon between sul-
foxidation and CꢀH bond activation already occurring at the
level of the initial rate-determining step, which may also be re-
sponsible for the observed diverse product distribution involv-
ing HRP-I- and HRP-II-catalyzed oxidation reactions.
Experimental Section
Materials
All solutions were prepared in acetonitrile (99.9% AMD) purchased
from Sigma–Aldrich. m-Chloroperoxybenzoic acid (m-CPBA) was
purchased from Acros Organics and purified before use by recrys-
tallization in hexane. [FeIII(TMP)(2-MeIm)] was generated in situ
after addition of an appropriate amount of 2-methylimidazole (2-
MeIm, Merck) to the solution of [FeIII(TMP)Cl] (Frontier Scientific
Porphyrin Product). Iodosylbenzene (PhIO) was synthesized from
iodobenzene diacetate (98%, Acros Organics) according to the pro-
cedure described in the literature.[38] Owing to the very poor solu-
bility of PhIO in acetonitrile, the stock solution of this oxidant was
prepared in methanol (99.9% AMD CHROMASOLV, Sigma–Aldrich).
All organic substrates, including cyclohexene, BICY, DHA, and DMS,
were purchased from Sigma–Aldrich and used without further pu-
rification.
Conclusions
A combined experimental and theoretical study on the oxidiz-
ing capabilities of Compound I and Compound II of HRP biomi-
metics in three types of substrate oxidation reactions, that is,
olefin epoxidation, alkane CꢀH bond activation, and heteroa-
tom oxidation, strongly confirm the competence of the
iron(IV)-oxo porphyrin p-cation radical intermediate as a highly
reactive and catalytic active species in the HRP cycle. The barri-
er gaps calculated for the rate-determining steps of the oxida-
Instrumentation
A quartz glass dip-in detector (Spectralytics, Aalen, Germany) cou-
pled to a J&M TIDAS 500–3 (TSPEC) diode array spectrophotometer
(J&M Analytik, Esslingen, Germany) was used to record the time-re-
solved UV/Vis spectra. The reactions were performed in a 20 mL
double-wall reaction vessel, and the temperature was kept at
ꢀ158C by using cold methanol circulation (Colora WK 14–1 DS,
Lorch, Germany) and a 2 III 800 W heating unit. Complete spectra
were recorded between 300 and 726 nm with the integrated soft-
ware program J&M TIDAS-DAQ 2.3.7.4.
3
tion processes involving 2,4Cpd I and Cpd II species nicely re-
flect the reactivity trend explored experimentally for the HRP-I
and HRP-II mimics from kinetic measurements. Although the
HRP-II mimic possessing a closed-shell a2u porphyrin orbital
was shown to be able to perform oxidation reactions on the
selected substrates, the preferred stepwise reaction mecha-
nism involving relatively high energy barriers makes this oxidiz-
ing species a considerably less efficient catalyst than its HRP-I
mimic. Importantly, the crucial difference between the HRP-I-
and HRP-II-catalyzed oxidation reactions concerns not only the
efficiency of the overall catalytic processes, but also refers to
the chemical nature of the obtained oxidation products (viz.
product distribution). In contrast to effective concerted pro-
Kinetic measurements
The kinetics of the reaction of the high-valent iron species with se-
lected substrates were studied under pseudo-first-order conditions.
In a typical experiment, a small excess of 2-MeIm was added to an
acetonitrile solution of [FeIII(TMP)(Cl)] at ꢀ158C. The formation of
mono- and di-2-methylimidazole derivatives of FeIII(TMP), viz. [FeIII-
(TMP)(2-MeIm)]+ and [FeIII(TMP)(2-MeIm)2]+, respectively, was fol-
lowed by UV/Vis spectrophotometry. A small excess of m-CPBA or
an equivalent amount of PhIO was added to the prepared solution
containing a mixture of [FeIII(TMP)(2-MeIm)]+ and [FeIII(TMP)(2-
MeIm)2]+ to produce [Fe (TMP )(O)(2-MeIm)] or [FeIV(TMP)(O)(2-
MeIm)], respectively. The characteristic spectral changes accompa-
nying the formation of the latter species and the completeness of
the oxidation reaction were studied with the UV/Vis diode array
spectrophotometer. Upon addition of different excesses of the se-
lected substrate to solutions of the Cpd I and Cpd II mimics gener-
ated in situ, the decomposition reaction of the reactive intermedi-
2
cesses performed by Cpd I, the oxidation mechanism of the
IV
+
+
C
3Cpd II species involves a significant ring-closure barrier for ep-
oxidation, and especially rather high rebound/desaturation
barriers for CꢀH bond activation. This implies the formation of
relatively long-lived intermediates, which can dissociate to the
substrate radical and FeIII-OH species, and therefore, result in
other oxidation products, as seen for the Cpd I oxidation reac-
Chem. Eur. J. 2014, 20, 14437 – 14450
www.chemeurj.org
14448
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
ate back to the starting iron(III) porphyrin complex was monitored
kinetically at 670 nm (depletion of Cpd I) or at 417 nm (depletion
of Cpd II). Kinetic traces were fitted to a pseudo-first-order decay
function with the use of the SPECFIT global analysis program (Ver-
sion 2.10).
insert tube. The obtained difference in frequency between the two
measured signals was relatively small, but meaningful. It results
mainly from the fact that the applied concentration of the para-
magnetic substance (HRP-I mimic) was low because of the low sol-
ubility of the FeIII(TMP) complex in acetonitrile (highest obtained
concentration was 6ꢁ10ꢀ5 m). The magnetic moment was calculat-
ed from Equation (4), in which Df is the difference in frequency
(Hz) between two measured signals, f is the oscillator frequency
(Hz) of the superconducting spectrometer, m is the mass (g) of the
paramagnetic substance per cm3 of solution, c0 is the mass sus-
ceptibility of acetonitrile (cm3gꢀ1), M is the molecular weight of the
paramagnetic substance (g), and T is the absolute temperature.
Product analysis experiments
For the analysis of the oxidation products of selected substrates,
m-CPBA and PhIO (6ꢁ10ꢀ4 m) were added to a 6ꢁ10ꢀ5 m porphyrin
complex in acetonitrile for the generation of Cpds I and II, respec-
tively. Subsequently, cyclohexene, BICY, or DHA (150 mm) was
added to the solutions containing the formed oxo-iron(IV) species.
After completion of the oxidation reactions, the obtained reaction
mixtures were analyzed with a PerkinElmer high-performance
liquid chromatography (HPLC) Chromera system. A Brownlee Vali-
dated AQ C18 5 mm, 250ꢁ4.6 mm column was employed for HPLC
separation. For separation of the products in the reaction with cy-
clohexene, a gradient of 0–100% CH3CN over 30 min with H2O as
the second eluent was used at a flow rate of 1 mLminꢀ1. The
chemical identities of the obtained oxidation products were deter-
mined by comparison of their characteristic UV spectra and/or re-
tention times with those of known authentic samples. The result-
ing product yields were calculated on the basis of the amount of
added oxidant. It was revealed that cyclohexene oxide, exo-2,3-ep-
oxynornbornane, and anthracene (33%) were formed as major
products in the oxidation reactions of Cpd I with cyclohexene,
BICY, and DHA, respectively. The oxidation reactions with Cpd II re-
sulted in the formation of 2-cyclohexene-1-one (14%) as the major
product of cyclohexene oxidation, exo-2,3-epoxynornbornane as
the major product of BICY oxidation, and anthracene (23%) and
9,10-dihydro-9-anthracenol as the major products of DHA oxida-
tion.
1=2
ð4Þ
m ¼ 2:83 ꢂ ½ð3Df=4pfm þ c0Þ ꢂ M ꢂ Tꢃ
The value of m was calculated to be 4.9ꢁ2.5 MB, which results in
the number of unpaired electrons (N=ꢀ1+(m2 +1)1/2) equal to
N=4.0ꢁ1.5. This means that the spin of the HRP-I intermediate is
close to S=3/2 in acetonitrile at ꢀ158C.
Computational methodology
We used six-coordinate oxo-iron species FeC20H12N4O(2-MeIm) as
Cpd I and Cpd II models, in which all side chains of the porphyrin
ring were abbreviated by hydrogen atoms, and the overall charge
was +1 for Cpd I and 0 for Cpd II. Subsequently, cyclohexene,
BICY, DHA, and DMS were added to the chemical system to study
the reaction mechanisms. Full geometry optimization was per-
formed by using the unrestricted B3LYP hybrid density functional
method[40] in combination with the Los Alamos effective core po-
tential plus double-x basis set (Lanl2dz)[41] on iron, and a 6-31G**
basis set on the remaining atoms (UB3LYP/LACVP** level). The ge-
ometry scanning method was used to first obtain the approximate
transition states, in which one of the structural parameters was
chosen as the internal reaction coordinate, and all other geometri-
cal degrees of freedom were fully optimized. Then, the geometry
of the approximate transition states was fully optimized without
symmetry constraints to obtain the exact transition states. Vibra-
tional frequencies were computed systematically to confirm that
an optimized geometry corresponded correctly to a local minimum
with only real frequencies or a transition state with only one imagi-
nary frequency. Bulk polarity effects were evaluated by single-point
calculations at the UB3LYP/LACVP** level with the PCM solvation
model using a polar solvent, acetonitrile (e=35.688). ZPE correc-
tions were based on UB3LYP/LACVP** frequency calculations. All
relative energies reported were at the UB3LYP/LACVP** level and
included solvation and ZPE corrections. All the above calculations
were carried out with the Gaussian 09 package.[42]
For analysis of the oxidation products of DMS by Cpd I and Cpd II
mimics, 0.01% trifluoroacetic acid (TFA) was used as an eluent at
a flow rate of 1 mLminꢀ1 for 8 min. Then, a gradient of CH3CN
(50%) and TFA (50%) was applied for 15 min, a gradient to 100%
CH3CN for 10 min, followed by 5 min of 100% CH3CN, and a gradi-
ent of 0–100% CH3CN for 10 min. Subsequently, only CH3CN was
used for the last 5 min. The major product of DMS oxidation by
both Cpd I and Cpd II models was identified as dimethyl sulfoxide
(DMSO). However, it should be noted that both DMSO and the hy-
droxylation product (alcohol or other side products formed owing
to dissociation of the DMS radical) were relatively small, hydrophil-
ic molecules, which were very difficult to separate through the ap-
plied method. Moreover, it was observed that DMS reacts also
spontaneously with m-CPBA and PhIO to form mainly DMSO as the
oxidation product. Although the oxidation of DMS without a cata-
lyst was slower than that in the presence of the catalyst, the spon-
taneous reaction of the substrate with oxidants made it difficult to
determine the actual nature and yield of the obtained products of
DMS oxidation.
Acknowledgements
LJ gratefully acknowledges the support from the National Nat-
ural Science Foundation of China (Grant No. 21307107). AF and
RvE gratefully acknowledge financial support from the Deut-
sche Forschungsgemeinschaft and the National Science Centre
of Poland (Grant No. 2012/05/B/ST5/00389). The China National
Supercomputing Center in Shenzhen (Shenzhen Cloud Com-
puting Center) is acknowledged for providing the Gaussian 09
software and high-performance computing clusters (Contract
No. S13037 and S14011). This work was in part carried out with
equipment purchased through financial support from the Eu-
Spin-state measurements for the HRP-I mimic
The magnetic moment of the HRP-I intermediate was determined
by using the modified 1H NMR method of Evans[39] at ꢀ158C in
acetonitrile. A coaxial insert tube containing the blank acetonitrile
solvent was inserted into the normal NMR tube containing 6ꢁ
10ꢀ5 m HRP-I mimic in acetonitrile. Owing to the high reactivity of
the oxo-iron(IV) porphyrin cation radical intermediate, no addition-
al reference was added to either of the tubes. The chemical shift of
the solvent peak in the presence of the paramagnetic metal com-
plex was compared to that of the solvent peak in the inner coaxial
Chem. Eur. J. 2014, 20, 14437 – 14450
www.chemeurj.org
14449
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[17] Y. Sun, X. Hu, H. Li, A. F. Jalbout, Comput. Theor. Chem. 2011, 966, 62–
69.
[18] W. J. Song, Y. O. Ryu, R. Song, W. Nam, J. Biol. Inorg. Chem. 2005, 10,
294–304.
ropean Regional Development Fund within the framework of
the Polish Innovation Economy Operational Program (contract
no. POIG.02.01.00-12-023/08).
[19] S. Shaik, D. Kumar, S. P. de Visser, A. Altun, W. Thiel, Chem. Rev. 2005,
105, 2279–2328.
[20] a) F. Ogliaro, N. Harris, S. Cohen, M. Filatov, S. P. de Visser, S. Shaik, J. Am.
Chem. Soc. 2000, 122, 8977–8989; b) S. P. de Visser, F. Ogliaro, P. K.
Sharma, S. Shaik, J. Am. Chem. Soc. 2002, 124, 11809–11826; c) L. Ji, G.
Schꢃꢃrmann, J. Phys. Chem. B 2012, 116, 903–912.
[21] a) S. P. de Visser, S. Shaik, J. Am. Chem. Soc. 2003, 125, 7413–7424;
b) C. M. Bathelt, L. Ridder, A. J. Mulholland, J. N. Harvey, J. Am. Chem.
Soc. 2003, 125, 15004–15005; c) C. Li, L. Zhang, C. Zhang, H. Hirao, W.
Wu, S. Shaik, Angew. Chem. 2007, 119, 8316–8318; Angew. Chem. Int.
Ed. 2007, 46, 8168–8170; d) P. Rydberg, U. Ryde, L. Olsen, J. Chem.
Theory Comput. 2008, 4, 1369–1377; e) C. S. Li, W. Wu, K. B. Cho, S.
Shaik, Chem. Eur. J. 2009, 15, 8492–8503; f) L. Ji, G. Schꢃꢃrmann,
Angew. Chem. 2013, 125, 772–776; Angew. Chem. Int. Ed. 2013, 52,
744–748.
[22] a) Y. Wang, C. Yang, H. Wang, K. Han, S. Shaik, Chembiochem 2007, 8,
277–281; b) K. Sen, J. C. Hackett, J. Am. Chem. Soc. 2010, 132, 10293–
10305.
[23] a) F. A. Walker, M. W. Lo, M. T. Ree, J. Am. Chem. Soc. 1976, 98, 5552–
5560; b) Y. O. Kim, H. M. Goff, Inorg. Chem. 1990, 29, 3907–3908;
c) G. N. La Mar, F. A. Walker, J. Am. Chem. Soc. 1972, 94, 8607–8708;
d) V. L. Balke, F. A. Walker, J. T. West, J. Am. Chem. Soc. 1985, 107, 1226–
1233.
Keywords: compound I
peroxidase · kinetics · UV/Vis spectroscopy
·
compound II
·
horseradish
[1] I. G. Denisov, T. M. Makris, S. G. Sligar, I. Schlichting, Chem. Rev. 2005,
105, 2253–2277.
[2] M. Sundaramoorthy, J. Terner, T. L. Poulos, Structure 1995, 3, 1367–
1377.
[3] P. Chelikani, I. Fita, P. C. Loewen, Cell. Mol. Life Sci. 2004, 61, 192–208.
[4] M. Gajhede, D. J. Schuller, A. Henriksen, A. T. Smith, T. L. Poulos, Nat.
Struct. Biol. 1997, 4, 1032–1038.
[5] J. Rittle, M. T. Green, Science 2010, 330, 933–937.
[6] B. Meunier, S. P. de Visser, S. Shaik, Chem. Rev. 2004, 104, 3947–3980.
[7] S. P. deVisser, S. Shaik, P. K. Sharma, D. Kumar, W. Thiel, J. Am. Chem. Soc.
2003, 125, 15779–15788.
[8] B. Chance, L. Powers, Y. Ching, T. Poulos, G. R. Schonbaum, I. Yamazaki,
K. G. Paul, Arch. Biochem. Biophys. 1984, 235, 596–611.
[9] W. D. Hewson, L. P. Hager, J. Biol. Chem. 1979, 254, 3182–3186.
[10] D. L. Harris, G. H. Loew, J. Am. Chem. Soc. 1996, 118, 10588–10594.
[11] P. R. Ortiz de Montellano, Annu. Rev. Pharmacol. Toxicol. 1992, 32, 89–
107.
[12] a) J. T. Groves, R. C. Haushalter, M. Nakamura, T. E. Nemo, B. J. Evans, J.
Am. Chem. Soc. 1981, 103, 2884–2886; b) J. T. Groves, T. E. Nemo, J. Am.
Chem. Soc. 1983, 105, 5786–5791; c) J. T. Groves, D. V. Subramanian, J.
Am. Chem. Soc. 1984, 106, 2177–2181; d) J. T. Groves, Y. Watanabe, J.
Am. Chem. Soc. 1986, 108, 507–508; e) J. P. Collman, T. Kodadek, J. I.
Brauman, J. Am. Chem. Soc. 1986, 108, 2588–2594; f) D. Ostovic, T. C.
Bruice, J. Am. Chem. Soc. 1989, 111, 6511–6517; g) Z. Gross, S. Nimri,
Inorg. Chem. 1994, 33, 1731–1732; h) H. Fujii, T. Yoshimura, H. Kamada,
Inorg. Chem. 1997, 36, 6142–6143; i) W. Nam, Y. M. Goh, Y. J. Lee, M. H.
Lim, C. Kim, Inorg. Chem. 1999, 38, 3238–3240; j) Y. Goto, T. Matsui, S.
Ozaki, Y. Watanabe, S. Fukuzumi, J. Am. Chem. Soc. 1999, 121, 9497–
9502; k) M. H. Lim, S. W. Jin, Y. J. Lee, G. J. Jhon, W. Nam, C. Kim, Bull.
Korean Chem. Soc. 2001, 22, 93–96; l) W. Nam, Acc. Chem. Res. 2007, 40,
522–531; m) N. Hessenauer-Ilicheva, A. Franke, D. Meyer, W. D.
Woggon, R. van Eldik, J. Am. Chem. Soc. 2007, 129, 12473–12479; n) A.
Franke, C. Fertinger, R. van Eldik, Angew. Chem. 2008, 120, 5316–5320;
Angew. Chem. Int. Ed. 2008, 47, 5238–5242; o) Y. J. Jeong, Y. Kang, A. R.
Han, Y. M. Lee, H. Kotani, S. Fukuzumi, W. Nam, Angew. Chem. 2008,
120, 7431–7434; Angew. Chem. Int. Ed. 2008, 47, 7321–7324; p) D.
Kumar, L. Tahsini, S. P. de Visser, H. Y. Kang, S. J. Kim, W. Nam, J. Phys.
Chem. A 2009, 113, 11713–11722; q) Y. Kang, H. Chen, Y. J. Jeong, W.
Lai, E. H. Bae, S. Shaik, W. Nam, Chem. Eur. J. 2009, 15, 10039–10046;
r) A. Takahashi, T. Kurahashi, H. Fujii, Inorg. Chem. 2009, 48, 2614–2625;
s) N. Hessenauer-Ilicheva, A. Franke, D. Meyer, W. D. Woggon, R. van El-
dik, Chem. Eur. J. 2009, 15, 2941–2959; t) N. Hessenauer-Ilicheva, A.
Franke, M. Wolak, T. Higuchi, R. van Eldik, Chem. Eur. J. 2009, 15, 12447–
12459; u) C. Fertinger, N. Hessenauer-Ilicheva, A. Franke, R. van Eldik,
Chem. Eur. J. 2009, 15, 13435–13440; v) A. Takahashi, T. Kurahashi, H.
Fujii, Inorg. Chem. 2011, 50, 6922–6928; w) A. Takahashi, D. Yamaki, K.
Ikemura, T. Kurahashi, T. Ogura, M. Hada, H. Fujii, Inorg. Chem. 2012, 51,
7296–7305; x) D. Kumar, G. N. Sastry, S. P. de Visser, Chem. Eur. J. 2011,
17, 6196–6205; y) S. Shaik, Y. Wang, H. Chen, J. S. A. Song, R. Meir, Fara-
day Discuss. 2010, 145, 49–70.
[24] M. J. Beck, E. Gopinath, T. G. Bruice, J. Am. Chem. Soc. 1993, 115, 21–29.
[25] a) R. D. Arasasingham, C. R. Cornman, A. L. Balch, J. Am. Chem. Soc.
1989, 111, 7800–7805; b) A. L. Balch, Y.-W. Chan, R.-J. Cheng, G. N.
La Mar, L. Latos-Grazynski, M. W. Renner, J. Am. Chem. Soc. 1984, 106,
7779–7785.
[26] M. E. Crestoni, S. Fornarini, F. Lanucara, Chem. Eur. J. 2009, 15, 7863–
7866.
[27] a) D. Kumar, B. Karamzadeh, G. N. Sastry, S. P. de Visser, J. Am. Chem.
Soc. 2010, 132, 7656–7667; b) D. Kumar, R. Latifi, S. Kumar, E. V. Rybak-
Akimova, M. A. Sainna, S. P. de Visser, Inorg. Chem. 2013, 52, 7968–
7979.
[28] E. Derat, S. Shaik, J. Am. Chem. Soc. 2006, 128, 8185–8198.
[29] a) K. B. Cho, X. Wu, Y. M. Lee, Y. H. Kwon, S. Shaik, W. Nam, J. Am. Chem.
Soc. 2012, 134, 20222–20225; b) D. Janardanan, D. Usharani, H. Chen, S.
Shaik, J. Phys. Chem. Lett. 2011, 2, 2610–2617.
[30] X. Wu, M. S. Seo, K. M. Davis, Y. M. Lee, J. Chen, K. B. Cho, Y. N. Pushkar,
W. Nam, J. Am. Chem. Soc. 2011, 133, 20088–20091.
[31] A. Altun, S. Shaik, W. Thiel, J. Am. Chem. Soc. 2007, 129, 8978–8987.
[32] a) D. Kumar, S. P. de Visser, P. K. Sharma, H. Hirao, S. Shaik, Biochemistry
2005, 44, 8148–8158; b) D. Kumar, S. P. de Visser, P. K. Sharma, E. Derat,
S. Shaik, J. Biol. Inorg. Chem. 2005, 10, 181–189.
[33] M. Filatov, N. Harris, S. Shaik, J. Chem. Soc. Perkin Trans. 2 1999, 399–
410.
[34] A. Takahashi, T. Kurahashi, H. Fujii, Inorg. Chem. 2007, 46, 6227–6229.
[35] M. Oszajca, A. Franke, A. Drzewiecka-Matuszek, M. Brindell, G. Stochel,
R. van Eldik, Inorg. Chem. 2014, 53, 2848–2857.
[36] a) K. A. Prokop, S. P. de Visser, D. P. Goldberg, Angew. Chem. 2010, 122,
5217–5221; Angew. Chem. Int. Ed. 2010, 49, 5091–5095; b) M. T. Green,
J. H. Dawson, H. B. Gray, Science 2004, 304, 1653–1656.
[37] D. Kumar, G. N. Sastry, S. P. de Visser, J. Phys. Chem. B 2012, 116, 718–
730.
[38] H. Saltzmann, J. G. Sharefkin, Org. Synth. 1963, 43, 60–61.
[39] D. F. Evans, Proc. Chem. Soc. 1958, 115–116; b) D. F. Evans, J. Chem. Soc.
1959, 2003–2005.
[40] A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652.
[41] a) P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 270–283; b) P. J. Hay,
W. R. Wadt, J. Chem. Phys. 1985, 82, 299–310; c) W. R. Wadt, P. J. Hay, J.
Chem. Phys. 1985, 82, 284–298.
[13] a) J. T. Groves, Z. Gross, M. K. Stern, Inorg. Chem. 1994, 33, 5065–5072;
b) W. Nam, S. E. Park, I. K. Lim, M. H. Lim, J. K. Hong, J. Kim, J. Am. Chem.
Soc. 2003, 125, 14674–14675; c) A. Rosa, G. Ricciardi, Inorg. Chem.
2012, 51, 9833–9845; d) L. Tahsini, M. Bagherzadeh, W. Nam, S. P. de
Visser, Inorg. Chem. 2009, 48, 6661–6669; e) M. Oszajca, A. Drzewiecka-
Matuszek, A. Franke, D. Rutkowska-Zbik, M. Brindell, M. Witko, G. Sto-
chel, R. van Eldik, Chem. Eur. J. 2014, 20, 2328–2343.
[14] a) Z. Gross, J. Biol. Inorg. Chem. 1996, 1, 368–371; b) Z. Gross, S. Nimri,
C. M. Barzilay, L. Simkhovich, J. Biol. Inorg. Chem. 1997, 2, 492–505.
[15] H. Ishimaru, H. Fujii, T. Ogura, Chem. Lett. 2010, 39, 332–333.
[16] L. A. Andersson, J. H. Dawson, Struct. Bonding 1991, 74, 1–40.
[42] Gaussian 09, Revision C.01, M. J. Frisch et al., Gaussian, Inc., Wallingford,
CT, 2010, see Supporting Information.
Received: February 25, 2014
Published online on September 12, 2014
Chem. Eur. J. 2014, 20, 14437 – 14450
www.chemeurj.org
14450
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim